Environmental photochemistry part II

Hdb Env Chem Vol. 2, Part M (2005): 1–47 DOI 10.1007/b138178 © Springer-Verlag Berlin Heidelberg 2005 Published online:...

Author: Boule P. | Bahnemann D.W. | Robertson P.K.J.

253 downloads 1660 Views 6MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Hdb Env Chem Vol. 2, Part M (2005): 1–47 DOI 10.1007/b138178 © Springer-Verlag Berlin Heidelberg 2005 Published online: 16 September 2005

Basic Concepts of Photochemical Transformations R. P. Wayne Physical and Theoretical Chemistry Laboratory, University of Oxford, South Parks Road, Oxford OX1 3QZ, UK [email protected] 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

2

Concepts of Light . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6

3 3.1 3.2 3.3 3.4 3.5

Electronic Structure of Molecules . . . . . . . . . . . . . . . Diatomic and Linear Polyatomic Molecules . . . . . . . . . . Small, Nonlinear Molecules . . . . . . . . . . . . . . . . . . . Complex Molecules . . . . . . . . . . . . . . . . . . . . . . . Photodissociation: Optical Dissociation and Predissociation Reactions and Isomerizations of Excited Species . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

10 10 11 11 12 13

4 4.1 4.2 4.3

Absorption and Emission of Light . . . Absorption and Emission Processes . . The Beer–Lambert Law . . . . . . . . . Selection Rules for Optical Transitions

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

15 15 18 20

5 5.1 5.2 5.3 5.4

Emission from Excited States . . . . . . . . . . . . . . . . Luminescence . . . . . . . . . . . . . . . . . . . . . . . . . Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . Phosphorescence . . . . . . . . . . . . . . . . . . . . . . . Luminescence Quenching: Kinetics and Radiative Lifetimes

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

23 23 25 28 29

6 6.1 6.2 6.2.1 6.2.2 6.2.3 6.2.4 6.3

Electronic Energy Transfer and Electron Transfer Intramolecular Energy Transfer . . . . . . . . . . Intermolecular Energy Transfer . . . . . . . . . . Radiative Energy Exchange . . . . . . . . . . . . . Collisional Energy Transfer . . . . . . . . . . . . . Coulombic Energy Transfer . . . . . . . . . . . . . Exciton Migration . . . . . . . . . . . . . . . . . . Electron Transfer . . . . . . . . . . . . . . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

32 32 35 35 36 38 39 39

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

42 42 44 45 46

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

47

7

Efﬁciency of Photochemical Processes: Quantum Yields and Photonic Efﬁciencies . . 7.1 Homogeneous Systems: Quantum Yields . . . 7.1.1 Chemical Change . . . . . . . . . . . . . . . . 7.1.2 Emission Processes . . . . . . . . . . . . . . . 7.2 Heterogeneous Systems: Photonic Efﬁciencies 8

. . . . .

. . . . .

2

R.P. Wayne

Abstract Photochemistry is concerned with the interaction between light and matter. The present chapter outlines the basic concepts of photochemistry in order to provide a foundation for the various aspects of environmental photochemistry explored later in the book. Electronically excited states are produced by the absorption of radiation in the visible and ultraviolet regions of the spectrum. The excited states that can be produced depend on the electronic structure of the absorbing species. Excited molecules can suffer a variety of fates; together, these fates make up the various aspects of photochemistry. They include dissociation, ionization and isomerization; emission of luminescent radiation as ﬂuorescence or phosphorescence; and transfer of energy by intramolecular processes to generate electronic states different from those ﬁrst excited, or by intermolecular processes to produce electronically excited states of molecules chemically different from those in which the absorption ﬁrst occurred. Each of these processes is described in the chapter, and the ideas of quantum yields and photonic efﬁciencies are introduced to provide a quantitative expression of their relative contributions.

Abbreviations A B c C d E Ea EPA EPR f (ν) g G h I IC ISC J k k l L n N s S S, S0 , S1 . . ., T1 T u x α ε ε η

Einstein A coefﬁcient Einstein B coefﬁcient Velocity of light Concentration Distance Molar excitation energy Activation energy Mixture of ether, isopentane, and ethanol Electron paramagnetic resonance Spectral distribution Gerade Gibbs free energy Planck’s constant Intensity Internal conversion Intersystem crossing Quantum number (vector) Rate constant Boltzmann constant Quantum number (vector) Quantum number (vector) Number density Avogadro’s number Quantum number (vector) Quantum number (vector) Quantum state Temperature Ungerade Distance Naperian absorption coefﬁcient Decadic absorption coefﬁcient Molecular energy Viscosity

Basic Concepts of Photochemical Transformations λ λ Λ µ ν ν ξ π π Π ρ σ σ Σ τ φ Φ

3

Wavelength Quantum number (vector) Quantum number (vector) Reduced mass Frequency Kinetic chain length Photonic efﬁciency Pi: normal maths Quantum state Quantum state (Radiation) density Quantum state Cross section Quantum state Lifetime Quantum yield (individual) Quantum yield (overall)

1 Introduction That the Sun’s rays inﬂuence matter has been evident to humans from the earliest of times. Photochemistry encompasses the chemical changes brought about by the absorption of electromagnetic radiation, and has become an increasingly important branch of chemistry over the few hundred years since systematic study ﬁrst began. Applications of photochemistry, ranging from photomedicine to photography, make vital contributions to the modern world. Perhaps even more signiﬁcantly, photochemistry is intimately involved in the processes of life itself. The evolution of our atmosphere to its present state depended to a large extent on photochemical processes, and the presence in our atmosphere, apparently uniquely in the Solar system, of oxygen and its photochemical product, ozone, provides an atmospheric shield from ultraviolet radiation from the Sun that would otherwise make life virtually impossible on the surface of our planet because of the sensitivity of nucleic acids and proteins to short-wavelength radiation. At the same time, light is directly involved in biological processes as diverse as photosynthesis and vision. It is the purpose of this chapter to explore the fundamental ideas of photochemistry, especially as applied to environmental processes. Although the word photochemistry obviously implies the chemical change brought about by light, a number of physical processes that do not involve any overall chemical change lie within the province of the photochemist; processes such as ﬂuorescence (in which light is emitted from a species that has absorbed radiation) or chemiluminescence (in which light is emitted as a product of a chemical reaction) must be regarded as of a photochemical

4

R.P. Wayne

nature. The word light is also used loosely, since radiation over a far wider range of wavelengths than the visible spectrum is involved in processes that would be accepted as photochemical. The long-wavelength limit is probably in the near infrared (say at 2000 nm); the region of interest extends into the vacuum ultraviolet, and is limited only formally at the wavelengths where radiation becomes appreciably penetrating (X-rays). The essential feature of photochemistry is probably the way in which excited states (Sect. 5) of atoms or molecules play a part in the processes of interest. The advent of the concept of quantization at the turn of the twentieth century ﬁnally provided the essential background to modern interpretations of photochemical behaviour. Planck’s law (Sect. 2) of 1900 provides the link between the energy of a photon of radiation and its frequency (and thus wavelength). Photons of light in the visible region of the spectrum correspond to energies of a few hundred kilojoules when scaled up to molar numbers, and the long-wavelength limit suggested of 2000 nm corresponds to just under 60 kJ mol–1. These energies are comparable, at the higher end, with chemical bond energies, while even the smaller values are of the same order of magnitude as the activation energies of some chemical processes. Radiation of these wavelengths is thus potentially capable of splitting some bonds, or at least of making reactions more rapid, perhaps to such an extent that processes thought not to occur at all become possible. As it happens, the energies under consideration are also those involved in spectroscopic transitions in atoms and molecules that lead to excitation of upper electronic levels, so that (although a few photochemical processes may involve high levels of vibrational excitation), photochemistry is quite conveniently thought of as the chemistry of electronically excited species. Electronic excitation, in its turn, is of significance for the chemistry, because the new electronic structure of the reactant may give it entirely new chemical properties, quite apart from the energetic aspects that we have emphasized up to now. The essential distinction between thermal and photochemical reactions now needs to be explored more fully. Thermal energy may be distributed about all the modes of excitation in a species: in a molecule these modes will include translational, rotational, and vibrational excitation, as well as electronic excitation. However, for species in thermal equilibrium with their surroundings, the Boltzmann distribution law is obeyed. If we take a typical energy of an electronically excited state equivalent in thermal units to 250 kJ mol–1 , at room temperature a fraction of the species of just 4 × 10–46 would be excited. To achieve a concentration of only 1% of the excited species would require a temperature of around 6800 ◦ C; in the event most molecular species would undergo rapid thermal decomposition from the ground electronic state and it would not be possible to produce appreciable concentrations of electronically excited molecules. In contrast, if molecules absorb radiation at a wavelength of about 500 nm as a result of an electronic transition, then electronic excitation certainly must occur, and the concentration

Basic Concepts of Photochemical Transformations

5

produced depends on several factors, including the intensity of illumination and the rate of loss of the excited species. Thus, we see that photochemical reactions are distinguished from thermal reactions, ﬁrst by the relatively large concentrations of highly excited species, which may react faster than the ground-state species and may even participate isothermally in processes that are endothermic for the latter, and secondly, if the excitation is electronic, by the changes in chemical reactivity that may accompany the new electronic conﬁguration of the species. It is convenient at this stage to examine brieﬂy the various fates of an electronically excited species. Figure 1 represents, in simpliﬁed form, the various paths by which an electronically excited species may lose its energy. Chemical change can come about either as a result of dissociation of the absorbing molecule into reactive fragments (process i), or as a result of direct reaction of the electronically excited species (process ii); electronically excited species may also undergo spontaneous isomerization, as indicated by path iii. Several mechanisms for dissociation are recognized (optical dissociation, predissociation, and induced predissociation. A special case of dissociation is that of ionization, shown as path viii. Energy transfer (Sect. 6), represented by paths iv and v in the ﬁgure, leads to excited species, which can then participate in any of the general processes. Radiative loss of excitation energy (path vi) gives rise to the phenomenon of luminescence: the terms ﬂuorescence and phosphorescence are used to describe particular aspects of the general phenomenon. Luminescence is subject to the laws of radiative

Fig. 1 The several routes to loss of electronic excitation. The use of the symbols ∗ , †, and ‡ is only intended to illustrate the presence of electronic excitation and not necessarily differences in states. One or both of the products of processes i–iii may be excited. (Reproduced from R.P. Wayne, Principles and applications of photochemistry, Oxford University Press, Oxford, 1998. By permission of Oxford University Press]

6

R.P. Wayne

processes, and it is treated brieﬂy in Sect. 6. Path vii indicated in Fig. 1 is physical quenching. In this process an atom or molecule M can relieve AB∗ of its excess energy. Physical quenching differs only formally from intermolecular energy transfer in that M, which must initially take up some excitation energy, does not make its increased energy felt in terms of its chemical behaviour. The electronic excitation of AB∗ is, in fact, frequently converted to translational or vibrational excitation of M.

2 Concepts of Light Since photochemistry is essentially the study of electronically excited species, it is now necessary to see how light can alter the electronic conﬁguration in an absorbing species, or how a change in the conﬁguration can lead to emission of light. From the time of Newton until the advent of quantum theory, the corpuscular (or particle) theory of light lost ground to the wave theory. Phenomena such as diffraction, or more especially interference, were only explicable in terms of a wave theory. However, the actual nature of the wave, and the mechanism of its propagation, was not established until the latter part of the nineteenth century. In the 1860s, James Clerk Maxwell made one of the major contributions to physics: possibly the only earlier work of such stature was that of Newton. Maxwell was attempting to reconcile the laws of electricity with those of magnetism. By powerful mathematical reasoning, Maxwell demonstrated that such reconciliation would be possible if, associated with an oscillating magnetic ﬁeld, there were a similar electric ﬁeld, and vice versa, and if a wave were propagated in a direction perpendicular to a plane containing the electric and magnetic ﬁelds, as illustrated in Fig. 2. The derivation of Maxwell’s equations need not concern us here, but one feature of the equations is of the greatest importance. The velocity of propagation of Maxwell’s electromagnetic waves was shown to be numerically identical to the velocity of light (both in vacuum). This striking result (1865) obviously suggests that light is an electromagnetic wave, but it did not draw much attention until after Hertz had conﬁrmed (1887–1888) Maxwell’s prediction of propagated waves from systems involving oscillating electrical and magnetic ﬁelds. Maxwell’s electromagnetic ﬁeld theory describes radiation in terms of oscillating electric and magnetic ﬁelds. It is one or other of these ﬁelds (usually the electric one) that interacts with the electrons of the chemical species absorbing the radiation. The most signiﬁcant subsequent modiﬁcation of Maxwell’s nineteenthcentury picture of electromagnetic radiation is our awareness that wave motion may have particulate properties associated with it. Planck developed his the-

Basic Concepts of Photochemical Transformations

7

Fig. 2 Oscillation of electric and magnetic ﬁelds in the propagation of light. (Reproduced from R. P. Wayne, Principles and applications of photochemistry, Oxford University Press, Oxford, 1998. By permission of Oxford University Press)

ory of black-body radiation on the basis of a postulate that radiation possessed particulate properties and that the particles, or photons, of radiation of speciﬁc frequency ν had associated with them a ﬁxed energy ε given by the relation ε = hν ,

(1)

where h is called Planck’s constant. This quantum theory of radiation was then used by Einstein to interpret the photoelectric effect. As early as the beginning of the nineteenth century, Grotthus and Draper had formulated a law of photochemistry which stated that only the light absorbed by a molecule could produce photochemical change in the molecule. The development of the quantum theory led to a realization that the radiation would be absorbed in quantized energy packets; Stark and Einstein suggested that one, and only one, photon was absorbed by a single particle to cause its photochemical reaction. It is now appreciated that several processes may compete with chemical reaction to be the fate of the species excited by absorption (see Sect. 1), and a more satisfactory version of the Stark–Einstein law states that if a species absorbs radiation, then one particle is excited for each quantum of radiation absorbed. Although this law might appear trivial, it is of fundamental importance in photochemistry, and the agreement between experiment and predictions based on the law does, in fact, offer substantial evidence in favour of the quantum theory of radiation (and multiphoton processes promoted by high intensities of light, for example from lasers, still require the absorption of light in quantized packets). It is now apparent that the energy of excitation of each absorbing particle is the same as the energy of the quantum given by the Planck relation, and the excitation energy per mole is obtained by multiplying this molecular excitation energy by N, Avogadro’s number. A linear relationship exists between energy and frequency, so that frequency characterizes radiation in a particu-

8

R.P. Wayne

larly direct way. It has been, however, almost universal practice to discuss the visible and ultraviolet regions of the spectrum in terms of the wavelength of the radiation, and it is therefore convenient to express the molar excitation energy, E, as a function of the wavelength, λ: Nhν , (2) λ where c is the velocity of light. Numerical relationships between E and λ may be derived; useful forms are 119 627 1239.34 kJ mol–1 = eV , (3) λ λ where λ is in nanometres. A convenient way of remembering the approximate energies of photochemically active radiation is to recall that the wavelength range is roughly 200–600 nm, while the corresponding energies are in the range 600–200 kJ mol–1 . In the context of environmental photochemistry, electromagnetic radiation from the Sun is clearly of prime importance. The Sun radiates with a total luminosity of 3.8 × 1026 W, and of this power, 1373 W m–2 is incident on the Earth: a quantity known as the solar constant, although it is not in fact constant with time. Over the wavelength range 300 nm (near ultraviolet) to 1 cm (microwave), the solar-radiation spectrum is a close approximation to blackbody radiation (Planck distribution) for a temperature of 5785 K. At shorter and longer wavelengths, the Sun radiates much more energy than the Planck distribution predicts for this temperature, because the emission in both these regions derives from the high-temperature corona of the Sun, rather than from the photosphere, which is the source of the mid-range emission. FigE = Nhν =

Fig. 3 The solar irradiance from λ = 1 nm to λ = 0.1 mm and normalized black-body radiation for T = 5770 K. Solar irradiance data are from SOLAR2000 http://www.spacewx.com/ solar_spectrum.html for 8 February 2002

Basic Concepts of Photochemical Transformations

9

ure 3 illustrates the relative solar irradiance over the wavelength range from 1 nm to 1 cm. The radiation reaching different altitudes within the atmosphere and penetrating to the surface depends, of course, on the absorption by the atmospheric constituents. The most important absorbers are molecular nitrogen, molecular (and atomic) oxygen, and ozone (O3 ). Figure 4 indicates the spectral regions in which each of these species contributes, and shows the altitude at which the optical density (attenuation by a factor 1/e: see Sect. 4) becomes unity. It is evident that, in general, the shorter the wavelength, the higher in the atmosphere the radiation is absorbed. Radiation that penetrates to the stratosphere (below approximately 50 km) is at wavelengths greater than about 200 nm, while radiation that reaches the troposphere (below approximately 20 km) and the Earth’s surface, is of λ 310 nm. At λ 150 nm, ionization phenomena become possible (ﬁrst of NO, then of O2 , and, at λ 100 nm, O, N and N2 ) so that the ionosphere is found at altitudes above approximately 80 km. The absorption in the longer-wavelength regions is mostly due to O2 and its (photochemical) product O3 . Since critical components of living cells, such as nucleic acids and proteins, are rapidly destroyed by short-wavelength ultraviolet radiation, life is only possible on the surface of the planet because of the presence of these atmospheric ﬁlters. In turn, however, virtually all the oxygen—and thus the ozone—present in our atmosphere is a consequence of the biological process of photosynthesis. The relatively high oxygen content (21%) of the Earth’s atmosphere is one feature that distinguishes it from the atmospheres of all other planets in the solar system.

Fig. 4 Depth of penetration of solar ultraviolet radiation through the Earth’s atmosphere. The line shows the altitude at which the optical depth is unity. (Based on ﬁgures presented by H. Friedman in J.A. Ratcliffe, (ed), Physics of the upper atmosphere, Academic Press, 1960, and by P.J. Nawrocki, K. Watanabe and L.G. Smith in The upper atmosphere, GCA Technical Report 61-13-A, 1961)

10

R.P. Wayne

3 Electronic Structure of Molecules Electrons in a molecule may be localized almost entirely on one atom, on one bond, or delocalized essentially over the entire molecule. The atoms that make up a molecule are held together by the forces imposed by the electrons that are shared between atoms, but again these shared electrons may be present in bonding, nonbonding or antibonding atomic orbitals. In the electronic ground state of a stable, bound molecule, the contribution of the electronic energies from the bonding orbitals is greater than that from the antibonding ones: often, but not necessarily, this condition arises because there are more bonding than antibonding electrons. Before extending this discussion, it is essential to provide a very brief indication of the nomenclature used to describe the electronically excited states of molecules, not only because it is convenient to be able to specify the states under discussion, but also because the types of excitation that are possible become revealed. 3.1 Diatomic and Linear Polyatomic Molecules In diatomic and linear polyatomic molecules, the electronic state may be deﬁned in part by the magnitude of the orbital electronic angular momentum resolved along the internuclear axis. This behaviour is similar to that seen in atoms, and the nomenclature follows the general pattern, with Roman letters replaced by Greek characters. Thus l and L for atoms are replaced by λ and Λ for individual orbitals and the whole molecule, respectively, and orbitals with Λ = 0, 1, 2, 3 are the familiar σ, π, δ and φ orbitals, while states with Λ = 0, 1, 2, 3 are Σ, Π, ∆, Φ states. The term symbol consists basically of 2S+1 Λ although a number of other pieces of information may be added on. One of these is the total angular momentum, and one of several possible coupling schemes must be used to derive it. Secondly, some further description of the symmetry properties of the wave function may be possible. A wave function may possess some, or all, of the molecular symmetry. In particular, for a centrosymmetric molecule, the wave function may either remain unchanged or change sign (but not magnitude) on inversion through the centre of symmetry. Such wave functions are called even or odd, respectively: in German the words are gerade and ungerade, and the symbols g or u, given as subscripts after Λ, derive from the initial letters. The wave function for a Σ state (Λ = 0) may remain the same or change sign on reﬂection by a plane of symmetry passing through the line of atomic centres: these two possibilities are represented by the symbols + or – appearing as superscripts after Λ. Some examples of the ground electronic

Basic Concepts of Photochemical Transformations

11

states of familiar diatomic or linear molecules will illustrate the terminology: N2 (1 Σ+g ), CO2 (1 Σ+g ), N2 O(1 Σ+ ), NO(2 Π), O2 (3 Σ–g ). 3.2 Small, Nonlinear Molecules It is not practicable to specify the electronic state of a nonlinear molecule in terms of orbital angular momentum, although spin multiplicity is still meaningful. If the molecule possesses symmetry elements, and the electrons are still sufﬁciently delocalized for the electronic cloud to possess effectively the same symmetry, then it may be possible to describe the electronic state in terms of the effect that the symmetry operations have on the sign of the wave function Symmetry symbols such as A, B, E, and T, with various subscripts and superscripts, are used to classify the behaviour of the wavefunction under the various symmetry operations. As usual, the symbol is preceded by a superscript number giving the spin multiplicity. The numerical subscript may be followed by a g or a u if the molecule has a centre of symmetry. The symbols are, in fact, the labels of the rows of the character tables of group theory, but that does not directly concern us here. It is, however, worth pointing out that a wavefunction of the type A, A1 , or A is symmetric with respect to all the molecular symmetry operations, while those of the type A2 or A are antisymmetric with respect to planes of symmetry present in the molecule. The various B types indicate antisymmetry with respect to rotational operations as well. Typical examples of term symbols for the ground states of simple molecules would be H2 CO(1 A1 ), NO2 (2 A1 ), and C6 H6 (1 A1g ); one excited state in each case is H2 CO(1 A ), NO2 (2 B1 ), and C6 H6 (B2u ). Note that the change from subscript number to superscript primes in the case of excited H2 CO indicates that the symmetry of the molecule has changed slightly on excitation (the molecule becomes slightly nonplanar). 3.3 Complex Molecules Complex molecules may not possess any symmetry elements, or if they do, the localizations of the electrons can so distort the electron cloud that its symmetry bears little relation to the molecular symmetry. In such cases it may be best to revert to a description of states in terms of the individual orbitals. As an example, we will consider formaldehyde, although a molecule as simple as this is probably best described by the group-theoretical term symbol of the last paragraph. The last ﬁlled orbitals in H2 CO can easily be shown to be ...(πCO )2 (nO )2 , where nO represents the nonbonding orbital on the O atom and the two electrons in it are the lone pair. The ﬁrst unﬁlled orbitals in formaldehyde are the π∗CO and σ∗CO antibonding orbitals. Promotion of one

12

R.P. Wayne

electron from, say, the nO orbital to π∗CO leads to excitation, and there is no restriction now on unpairing of spins so both singlet and triplet states are possible. The states are then designated as 3 (n, π∗ ) or T1 (n, π∗ ),1 (n, π∗ ) or S1 (n, π∗ ), and similarly for (pi, π∗ ), (n, σ∗ ), and (π, σ∗ ), states. Spectroscopic transitions are then referred to as n → π∗ , and so on. 3.4 Photodissociation: Optical Dissociation and Predissociation It is immediately apparent that if light is absorbed by a bound molecule in such a way that bonding electrons are excited to nonbonding or antibonding orbitals, and the net energy is no longer bonding, the molecule may dissociate. This is one important mechanism by which molecules may undergo photodissociation (channel i of Fig. 1). In addition, the upper electronic state populated by absorption may be bound at low vibrational levels, but the radiation absorbed may possess sufﬁcient energy to excite the molecule to vibrational levels above the dissociation limit for that state. Both these processes are referred to as optical dissociation. The spectroscopy indicates the detail of the process. A diffuse absorption spectrum is seen at all wavelengths in the ﬁrst case, since an unbound state is being populated, while in the second case, a converging progression of absorption bands is observed, with dissociation being marked by reaching the convergence limit. Predissociation is the term given to another way in which absorption of light can lead to photofragmentation. The excited electronic state ﬁrst populated by optical absorption may itself be bound, but may convert to another unbound state without emission of radiation (an intramolecular energy transfer process, or radiationless transition: channel v of Fig. 1; see Sect. 6), and thus lead to dissociation. The spectroscopic observation is of a normal banded vibrational progression, with a region of rotational diffuseness (resulting from the operation of the uncertainty principle) near the energies at which the radiationless transition occurs, but before the normal optical dissociation limit. It is this early diffuseness that is the origin of the term predissociation. The two kinds of photodissociation can be illustrated by the process O2 + hν → O + O ,

(4)

which is one of the key processes in atmospheric and environmental chemistry. Because O2 is a diatomic molecule, its energy can be represented as a function of the internuclear separation between the two O atoms by a simple potential energy curve. Such curves are shown in Fig. 5 for just three electronic states of the molecule; the symbols associated with these curves are those explained in Sect. 3.1, but need only be regarded as ways of labelling the states. Two of these curves, X3 Σ–g and 5 Πu , are shown to be linked to (correlate with) a pair of ground-state oxygen atoms, O(3 P) + O(3 P), while

Basic Concepts of Photochemical Transformations

13

Fig. 5 Potential energy curves for some of the low-lying states of molecular oxygen. (Reproduced from R.P. Wayne, Chemistry of atmospheres, 3rd edn., Oxford University Press, Oxford, 2000. By permission of Oxford University Press)

the third curve, B3 Σ–u , correlates with one ground-state oxygen atom and one excited atom, O(1 D). (There are, in fact, at least another ﬁve curves that correlate with the pair of ground-state atoms.) Optical dissociation involves absorption from the X3 Σ–g to the B3 Σu curves, when at energies greater than approximately 700 kJ mol–1 (approximately equal to 170 nm) direct formation of the O(3 P) + O(1 D) pair is possible. Close inspection of Fig. 4 shows the presence there of the banded absorption in the wavelength region around 170–200 nm, with the structure disappearing when the convergence limit is reached. Figure 4 also shows clearly that this photodissociation (photolysis) can be effective in the atmosphere only at relatively great altitudes, above say 50 km. At longer wavelengths, and thus at lower altitudes, another optical dissociation process, involving one of the bound states not shown in Fig. 5, is important. For our purposes, however, it is instructive to examine an alternative predissociation that arises from the crossing of the B3 Σu and 5 Πu curves. Radiationless transfer from the ﬁrst to the second of these states occurs at an energy of around 615 kJ mol–1 , so that absorption of radiation at wavelengths approaching 200 nm can lead to the formation of photofragments, but in this case two ground-state oxygen atoms are formed. 3.5 Reactions and Isomerizations of Excited Species Electronic excitation may lead to chemical effects that are subtler than straightforward fragmentation. As stated earlier, the reactivity of excited

14

R.P. Wayne

atoms and molecules may be enhanced because of the excess energy that they carry. The excited oxygen atoms, O(1 D), that we have just mentioned play a central role in several aspects of atmospheric chemistry. In the stratosphere and the troposphere, they can be formed by the photolysis of ozone O3 + hν(λ ≤ 310 nm) → O2 (a1 ∆g ) + O(1 D)

(5)

Note here the relatively long wavelength of the threshold energy, and note also that the O2 fragment is itself formed excited, although that is not the immediate point of interest. The noteworthy feature is that excited atoms are able to participate in processes such as O(1 D) + H2 O → OH + OH O(1 D) + N2 O → NO + NO O(1 D) + CH4 → OH + CH3

(6) (7) (8)

for which the rate constants are large; with ground-state, 3 P, oxygen atoms, the rate constants are exceedingly small at atmospheric temperatures. All three of these reactions are key steps in different aspects of atmospheric chemistry. Not only may the extra energy carried by an electronically excited state of a molecule confer additional reactivity on the species, but the conﬁgurations of the electrons themselves may lead to an increase (or possibly decrease) of the intrinsic reactivity. The requirement, or at least propensity, to conserve spin, angular momentum, or symmetry often inﬂuences the outcome of reactions of electronically excited species. For example, thermal cyclodimerization of substituted ethenes occurs predominantly by cis–trans addition, while photo-cyclodimerization yields mostly cis–cis addition products. The explanation for the opposing behaviour for thermal and photochemical reactions can be sought in terms of the correlations between reactants and products. The energy barrier that makes a ground-state reaction forbidden is a result of the correlation of the reactant and product ground states with upper excited states. Excitation may alter both the sizes and the shapes of molecules. For a particular reaction, the new steric arrangements may then increase (or, indeed, decrease) reactivity. Movement of electrons between bonding, nonbonding, and antibonding orbitals may be expected to change the molecular dimensions. Shapes of molecules may be affected by changes in the nature of the bonding. For example, while the ground state of an alkene such as ethene is planar, the equilibrium structure of the (π, π∗ ) excited state (see Sect. 3.3) has the two CH2 groups lying in perpendicular planes. Promotion of the electron is accompanied by uncoupling of the π orbital to leave only a σ bond between the carbon atoms. The two electrons in the 2p orbitals on the carbon atoms experience minimum electrostatic repulsion with the perpendicular geometry.

Basic Concepts of Photochemical Transformations

15

Isomerization (pathway iii, Fig. 1) is another process that may be induced photochemically as a consequence of the properties of the electronically excited state produced by the absorption of light. One example of simple isomerization is the geometrical (cis–trans) isomerization of unsaturated compounds in which the double bond lends geometrical rigidity. Ethene and its derivatives have lowest excited singlet and triplet states that are (π, π∗ ) in character: an electron is promoted from the highest ﬁlled bonding π orbital to the lowest antibonding π∗ orbital. As just pointed out, it can be shown that the (π, π∗ ) state is most stable if the molecule is twisted, from the planar ground-state conﬁguration, through 90◦ about the double-bond axis. This perpendicular conﬁguration minimizes the overlap between π and π∗ orbitals. It is apparent that, if an alkene is excited to the (π, π∗ ) state, then it will tend to twist to the perpendicular conﬁguration. Subsequent electronic energy degradation to the ground state will then require that the molecule become planar again, both cis and trans isomers being formed. The perpendicular form of the excited state is geometrically equivalent whether it is derived from a cis or a trans ground-state molecule.

4 Absorption and Emission of Light 4.1 Absorption and Emission Processes Three processes must be distinguished: absorption, stimulated emission, and spontaneous emission. Suppose a chemical species possesses two quantized states l and m, of energies εl and εm . If the species is in state l initially, it might be able to interact in some way with electromagnetic radiation and absorb energy in order to reach state m. In a normal process this absorption of energy occurs in a single step, so that the energy difference between ﬁnal and initial levels must be equivalent to the energy of a single photon of radiation. Hence, absorption of radiation can only occur if εm – εl = hν (Bohr condition). The process of absorption has involved the loss of intensity from the electromagnetic radiation and the gain of energy by the absorbing species. The converse process, in which a species in an upper state gives up energy to electromagnetic radiation and increases the intensity, is known as stimulated emission: the word stimulated indicates that it is the interaction between the radiation already present and the energy-rich species that encourages the latter to give up its energy. Although we have not mentioned the nature or magnitude of the interaction between the species and radiation, it is apparent that the rate (intensity) of absorption or stimulated emission is proportional to the rate of collision between photons and the absorber or emitter: that is to say, the intensity change is proportional to the radia-

16

R.P. Wayne

tion density, , and to the concentration of chemical species. The constant of proportionality deﬁnes the so-called Einstein B coefﬁcients. Blm is the coefﬁcient for absorption, while Bml is that for stimulated emission: the principle of microscopic reversibility suggests that Blm = Bml , and this result can also be derived from the complete treatment of radiation theory. The rates of absorption and stimulated emission are Blm nl and Bml nm (= Blm nm ), respectively, where nl and nm are the concentrations of species in lower and upper states. For a system in thermal equilibrium, nm is always less than nl , and absorption is always a more important process than stimulated emission. How much more important depends, of course, on the relation between εm – εl and the temperature, T. The energy levels of signiﬁcance in photochemistry are such that εm – εl kT, and nm nl , so that stimulated emission is rarely important in photochemical processes in which thermal equilibrium is established. However, in nonequilibrium situations, stimulated emission may not be negligible, and if a population inversion (nm > nl ) arises, then the emission process will predominate over absorption, and net emission will result. The LASER (Light Ampliﬁcation by Stimulated Emission of Radiation) depends on the achievement of such population inversions, often by photochemical techniques. In addition to absorption and stimulated emission, a third process, spontaneous emission, is required in the theory of radiation. In this process, an excited species may lose energy in the absence of a radiation ﬁeld to reach a lower energy state. Spontaneous emission is a random process, and the rate of loss of excited species by spontaneous emission (from a statistically large number of excited species) is kinetically ﬁrst-order. A ﬁrst-order rate constant may therefore be used to describe the intensity of spontaneous emission: this constant is the Einstein A factor, Aml , which corresponds for the spontaneous process to the second-order B constant of the induced processes. The rate of spontaneous emission is equal to Aml nm , and intensities of spontaneous emission can be used to calculate nm if Aml is known. Most of the emission phenomena with which we are concerned in photochemistry—ﬂuorescence, phosphorescence, and chemiluminescence—are spontaneous, and the descriptive adjective will be dropped henceforth. Where emission is stimulated, the fact will be stated. We referred in the last paragraph to the calculation of concentrations of excited species from emission intensity measurements. It may, however, not always be possible to determine Aml directly, and some other method of evaluating A factors may be needed. The A coefﬁcient may be calculated from the B coefﬁcient for the same transition by using the relation 8πhν 3 Blm . (9) c3 (For simplicity, the degeneracies of upper and lower states are taken to be identical in this and similar discussions in the present article.) B itself may Aml =

Basic Concepts of Photochemical Transformations

17

be determined from experimental absorption measurements, as will be described in Sect. 4.2. The equation is the ν 3 law to which reference is frequently made by spectroscopists and photochemists. The nature of the interaction between electromagnetic radiation and matter must now be considered. The processes may become clearer if we consider a simple example: the absorption of infrared radiation by a molecule of HCl. The molecule has a permanent dipole moment, so that the energy of the molecule will be affected by the presence of an electric ﬁeld, and the bond will tend to be distorted according to the direction of the ﬁeld. Now consider an oscillating electric ﬁeld, such as that present in electromagnetic radiation. If the frequency of oscillation is equal to the vibration frequency of the H–Cl bond, then the induced motion of the electrons may lead to an increased energy of nuclear motion. The vibrational energy in the molecule will then increase by one quantum, and the intensity of electromagnetic radiation will be depleted by an equivalent amount. This description of the absorption process is obviously just a pictorial representation, but it indicates that the interaction derives from the inﬂuence, via the molecular dipole, of the electric vector of the radiation on the energy of the molecule. A transition occurring through such an interaction is called an electric-dipole transition. Interactions with the magnetic vector of the radiation, or those that involve quadrupoles in the chemical species, give rise to magnetic-dipole, electricquadrupole, and magnetic-quadrupole transitions, etc. All these interactions are, however, much weaker than the electric-dipole interaction and may frequently, but by no means always, be ignored. The absorption or emission of infrared radiation by an oscillating molecule possessing a dipole is readily understood in the pictorial terms of the last paragraph. It is less easy to describe electronic transitions in the same manner. In the classical sense, electronic excitation does not correspond to increasing the energy in an oscillating system, and, in any case, neither the upper nor the lower electronic state may possess a steady dipole (e.g. the electron cloud in an atom is symmetrically disposed about the nucleus in all states, so there is no effective charge separation). However, the general principles of interaction with radiation still apply, and what we need to know is whether an (electric) dipole interaction can occur during transitions between two states. Wave-mechanical techniques provide the only satisfactory method of dealing with this problem: the time-dependent Schrödinger equation can be used to derive the rate at which a system can be changed from one stationary state to another under the inﬂuence of a perturbing effect. If this rate is nonzero for a perturbation of the system involving electric dipole interaction with the electric vector of radiation, then an electric dipole transition can occur. The rate of change between states multiplied by the number of species present in the lower state is, of course, the overall rate of absorption of photons, so that, in principle, solution of the time-dependent Schrödinger equation should lead to prediction of the intensity of the transition. Explicit

18

R.P. Wayne

solutions are, however, rarely available, and in such cases it may be possible only to say whether or not any interaction occurs, rather than to calculate its magnitude. The conditions under which a speciﬁc interaction arises are given as selection rules for that type of transition: for electric-dipole transitions in electronic spectroscopy, they are given in Sect. 4.3. In this section we have distinguished between spontaneous and induced transitions, and we have shown how the probabilities for these processes, the Einstein A and B coefﬁcients, are related to each other. The next section deals with experimental measurements of absorption, and the relation between these measurements and the theoretical quantities is explored. 4.2 The Beer–Lambert Law The fraction of light transmitted through an absorbing system is very frequently found to be represented by the equivalent relations It = e–σnd = e–αCd = 10–Cd . I0

(10)

It and I0 are the transmitted and incident intensities in the system, n or C the concentrations of absorber (measured in units of molecules per cubic centimetre or moles per cubic decimetre, respectively), and d is the thickness of absorber through which the light beam has passed, virtually always measured in centimetres. The quantities σ , α or ε are the constants of proportionality appropriate to the equation employed: σ is known as the absorption cross section, α the Napierian molar absorption coefﬁcient, and ε the decadic molar absorption coefﬁcient (formerly called the decadic extinction coefﬁcient). Physicists and gas-phase chemists tend to use the molecular concentration units, and thus to talk about absorption cross sections for a single molecule, while condensed-phase chemists usually employ molar concentration units and most frequently discuss optical absorption in terms of ε. α is, of course, equal to 2.303ε. Whatever the units, the constant of proportionality is dependent on λ, and the variation with λ represents the quantitative absorption spectrum. Despite the constancy implied by Eq. 10, σ (or α or ε) can, under some circumstances, vary with concentration as well. The law embodied in Eq. 10 was originally known as Lambert’s law; a second law, Beer’s law, stated that if C and d were altered but the product Cd was constant, then the fraction of light transmitted remained the same. Since this latter law follows in any case from Lambert’s law, Eq. 10 is now known as the Beer–Lambert law. A logarithmic form of the equation, for example log10 (I0 /It ) = εCd ,

(11)

is often employed, and the product εCd is called the optical density of the system.

Basic Concepts of Photochemical Transformations

19

A proof of the Beer–Lambert law may be derived if it is assumed that the rate of loss of photons is proportional to the rate of bimolecular collisions between photons and the absorbing species. The decrease, –dI, in intensity I at any point x in the system (Fig. 6) for a small increase in x, dx, is given by – dI = σ In dx ,

(12)

where σ is a constant of proportionality. Integration, with the boundary conditions I = I0 at x = 0, I = It at x = d, yields Eq. 10 in its ﬁrst form. The intensity of radiation absorbed, Iabs , is, of course, I0 – It , so that the fraction absorbed is given by Iabs = 1 – e–σnd = 1 – e–αCd = 1 – 10–Cd . (13) I0 An important approximate expression results when σ nd or αCd is small: expansion of the exponential and rejection of second- and higher-order terms leads to the conclusion that Iabs σ nd = αCd . (14) I0 for σ nd = αCd = 0.01, the approximate value of Iabs /I0 differs from the accurate one by less than 1%, while even for σ nd = αCd = 0.1, the difference is only 5%. Thus, Eq. 14 often gives a sufﬁciently accurate estimate of fractional absorptions less than about 10%. Although the Beer–Lambert law usually offers an adequate description of experimental data, there are some circumstances in which it does not. For example, the width of an absorption band or line depends, in part, on factors such as molecular collision, so that changes in concentration can alter the σ –λ relationship, and hence lead to a breakdown of the law. For oriented systems (e.g. crystals) the value of σ may depend on the plane of polarization of the

Fig. 6 The change in intensity, I, with optical path, x (see Eq. 12). (Reproduced from R.P. Wayne, Principles and applications of photochemistry, Oxford University Press, Oxford, 1998. By permission of Oxford University Press)

20

R.P. Wayne

light. Again, the law holds only if the wavelength range to which the intensity measurements refer is small compared with the width of the absorption band (i.e. if ε is constant over the wavelength range). Thus, experiments in which there is absorption from wide-band incident radiation by a narrow-band absorber will not obey the Beer–Lambert law. More trivially, the concentrations of species that associate or dissociate will not be equal to the concentrations expected on the basis of amounts of material added. The maximum value achieved in an absorption band by the extinction coefﬁcient may be an indicator of the nature of the spectroscopic transition, and especially of whether the transition is allowed or forbidden for electric-dipole interactions (see Sect. 4.3). Although the experimental extinction coefﬁcients σ , α or ε are measured, in principle, at a single wavelength or frequency, a real absorption band spans a range of frequencies. The experimental measure of the probability of the interaction is the absorption coefﬁcient integrated over the band, for example σ dν, rather than the absorption coefﬁcient at a single frequency, the limits of integration being set by the frequency limits of the transition under consideration. The Einstein B coefﬁcient is one measure of the probability that a transition will occur somewhere in the band, and is connected to the integrated absorption coefﬁcient by the relation c σ dν , (15) Blm = hν where the integration is carried out over the entire absorption band. This equation provides the means of estimating B (and hence A: cf. Eq. 9) from absorption measurements. For a given overall probability of transition, there is some kind of inverse relationship between the band width and the absorption coefﬁcient. 4.3 Selection Rules for Optical Transitions Formal rules, known as selection rules, may be used to decide whether or not an electric dipole transition between two states may take place. Perhaps the most important is the rule governing spin multiplicity: spin must not change during an electronic transition. The usual way to write rules of this kind is ∆S = 0 .

(16)

In atoms, for one-electron transitions, we have the selection rules ∆L = ±1 , ∆J = 0, ±1 but J = 0 J = 0 .

(17) (18)

For diatomic and linear polyatomic molecules the orbital momentum rule is ∆Λ = 0, ±1 ,

(19)

Basic Concepts of Photochemical Transformations

21

and, where applicable, the rules governing symmetry are g ↔ u;

+↔+;

–↔– .

(20)

The rules are derived by considering whether or not an interaction with the electric vector of electromagnetic radiation is possible in going from the initial to the proposed ﬁnal states of the chemical species. Frequently, the symmetry properties of the wavefunctions involved sufﬁce to exclude certain transitions, so that the selection rules are a way of representing the possibilities that remain, although they say nothing about the absolute intensities of the interactions. For nonlinear molecules that still possess some symmetry, selection rules can again be derived on the basis of the symmetry properties. If the molecule possesses a centre of symmetry, then g and u states exist, and electricdipole interaction requires that the transition is g ↔ u. Molecules possessing the symmetry of, for example, ground-state H2 CO or NO2 have symmetryallowed transitions A1 ↔ B1 or B2 , but not from A1 to A2 , or B1 to B2 . Similarly, molecules with the symmetry of C6 H6 cannot undergo symmetryallowed electric-dipole transitions from the A1 ground state to a B1 , B2 , or any other A1 state, and since the molecule is centrosymmetric, transitions from the g ground state to any other g state are also excluded. Although the selection rules provide a useful guide to the types of transition that might be expected, they are by no means followed rigidly. Absorption or emission spectra due to forbidden transitions are quite often observed, although they are usually considerably weaker than the allowed transitions. An example that follows directly from the preceding paragraph concerns the 1 B2u ←1 A1g absorption in benzene. According to the rules propounded, the transition is forbidden. However, the symmetry rule is based on the shape of benzene being a perfect hexagon. In reality, molecular vibrations distort the molecule so that the symmetry is reduced and the transition becomes weakly allowed. Such coupling of vibrational and electronic motions (vibronic coupling) is a consequence of a breakdown in the Born– Oppenheimer approximation, which states that all energy modes (e.g. electronic, vibrational, and rotational) can be considered independently, so that the individual wavefunctions can be factorized. Breakdown of the separability of rotational and electronic modes is much less important, although it is still recognized as a source of very weak forbidden transitions. The spin selection rule, ∆S = 0, might be expected to be of universal applicability, since it does not require the molecule under consideration to have any geometrical symmetry. However, spin-forbidden transitions are also frequently observed. The spin rule is based again on the idea of separability of wavefunctions, this time of the spin and spatial components of the electronic wavefunction. However, the electron experiences a magnetic ﬁeld as a result of the relative motion of the positive nucleus with respect to it, and this ﬁeld causes some mixing of spatial and spin components, giving rise to spin–orbit

22

R.P. Wayne

coupling. As a result, the idea of pure spin states must be modiﬁed to allow for exchange of angular momentum with the orbital mode. For example, a state formally described as a singlet may, in reality, have some triplet character, while a formal triplet may have some singlet character. Transitions between singlets and triplets may then be looked on as transitions between the pure singlet and triplet components of the mixed states. Since spin–orbit coupling depends on interaction with the nucleus, its magnitude increases rapidly with increasing nuclear charge (in fact, as the fourth power). Spin-forbidden transitions are thus stronger when heavy nuclei are involved. There are two other fairly common causes of apparent breakdown of the electronic selection rules. First, collisions with other atoms or molecules, or the presence of electric or magnetic ﬁelds, may invalidate selection rules based on state descriptions of the unperturbed species. Secondly, although the transition may be forbidden for an electric-dipole interaction, it may be permitted for the (much weaker) magnetic-dipole or electric-quadrupole transitions. Not only do the selection rules fail to predict the occurrence of forbidden transitions, but they may also fail to predict that a transition will be weak even though it is symmetry-allowed. One well-known example is that of the n → π ∗ transition in pyridine. The transition is of the type A2 → B2 , and the transition is symmetry-allowed. However, the nonbonding electron is on the nitrogen atom in C5 H5 N, while the π orbital involved in the transition is delocalized over the ring. As a result, there is little overlap between the initial and ﬁnal wavefunctions, and electromagnetic radiation cannot excite the electron from the lower to the upper orbital. The transition is therefore weak (about 100 times less intense than would be expected ordinarily) because of the lack of orbital overlap, even though it is allowed on symmetry grounds. So far, this discussion of selection rules has considered only the electronic component of the transition. For molecular species, vibrational and rotational structure is possible in the spectrum, although for complex molecules, especially in condensed phases where collisional line broadening is important, the rotational lines, and sometimes the vibrational bands, may be too close to be resolved. Where the structure exists, however, certain transitions may be allowed or forbidden by vibrational or rotational selection rules. Such rules once again use the Born–Oppenheimer approximation, and assume that the wavefunctions for the individual modes may be separated. Quite apart from the symmetry-related selection rules, there is one further very important factor that determines the intensity of individual vibrational bands in electronic transitions, and that is the geometries of the two electronic states concerned. Relative intensities of different vibrational components of an electronic transition are of importance in connection with both absorption and emission processes. The populations of the vibrational levels obviously affect the relative intensities. In addition, electronic transitions between given vibrational levels in upper and lower states have a speciﬁc probability, determined in part

Basic Concepts of Photochemical Transformations

23

by the electronic transition probability and in part by the probability of ﬁnding a molecule with similar internuclear separations in both states. This last factor is bound up with the Franck–Condon principle, which is yet another aspect of the Born–Oppenheimer approximation. The principle states that there is no change in internuclear separation during the course of an electronic transition. The application of the principle is most readily visualized for a diatomic molecule, for which a line depicting a transition on a potential energy diagram must be drawn vertically. For any molecules, the probability of transition occurring in any small range of internuclear distance will depend on the product of probabilities of a molecule possessing that internuclear distance in each electronic state, and the total transition probability is this probability integrated over all internuclear distances.

5 Emission from Excited States 5.1 Luminescence Luminescence is the general term given to the emission of radiation from excited species, represented by pathway vi of Fig. 1. Luminescent emission provides some of the most reliable information about the nature of primary photochemical processes. Competition exists between emission and other fates of excited species (quenching, reaction, decomposition, etc.), and the dependence of emission intensity on temperature, reactant concentrations, etc., may yield valuable data about the nature and efﬁciencies of the various processes. In particular, quenching by bimolecular collisions, and unimolecular energy degradation by radiationless transitions, is almost always best studied in terms of its effect on the intensity of luminescence. The various individual luminescent phenomena are named according to the mode of excitation of the energy-rich species. Emission from species excited initially by the absorption of radiation is referred to as ﬂuorescence or phosphorescence. The two emission processes were originally distinguished in terms of whether or not there was an observable afterglow. That is, if emission of radiation continued after the exciting radiation was shut off, the emitting species was said to be phosphorescent, while if emission appeared to cease immediately, then the phenomenon was one of ﬂuorescence. The essential problem is what is meant by immediately in this context, since the observation of an afterglow will obviously depend not only on the actual rate of decay of the emission (see Sect. 5.4 for further discussion of emission lifetimes), but also on the techniques used to observe it. In 1935 Jablonski interpreted phosphorescence as being emission from some long-lived metastable electronic state lying lower in energy than

24

R.P. Wayne

the state populated by absorption of radiation, and it was later suggested that the long-lived metastable state was, in many cases, a triplet state of the species; there is now considerable experimental evidence to substantiate this hypothesis. The long lifetime of the emission is a direct consequence of the forbidden nature of a transition from an excited triplet to the ground-state, which is a singlet for most (but certainly not all) molecules. That electricdipole transitions occur at all where ∆S = 0 is due to the inadequacy of S to describe a system in which there is spin–orbit coupling. Extension of this idea to other systems, not necessarily triplet–singlet, in which ∆S = 0 leads to the useful deﬁnition of phosphorescence as a radiative transition between states of different multiplicities: ﬂuorescence is then understood to be a radiative transition between states of the same multiplicity. These deﬁnitions are used almost universally by organic photochemists, although they might be extended to include within the scope of phosphorescence emission processes involving a transition forbidden by any selection rule rather than just the ∆S = 0 rule. Since the distinctions between allowed and forbidden transitions are not sharp, the deﬁnitions lack some precision. Potential energy curves (or, rather, sections through the many-dimensioned potential energy hypersurface) cannot usually be constructed for complex polyatomic molecules in the simple way that they can for diatomic molecules (cf. Fig. 5). Figure 7 shows an alternative diagram in a form devised by Jablonski (Jablonski diagram), and indicates the radiative processes of ﬂuorescence and phosphorescence. The diagram does not attempt to represent the molecular shapes and sizes, and the vibrational levels drawn for each state do not usually correspond to the actual spacings and υ , υ numberings. On the other hand, the energies of the vibrational ground states of each electronic level are shown correctly if the experimental evidence is available. As will be seen, the S0 , Sl ..., Tl ... notation is employed. Wavy lines on the diagram represent radiationless energy conversion. Vertical wavy lines within a particular electronic state indicate degradation of vibrational excitation (probably by a collisional, intermolecular process), while the horizontal wavy lines indicate intramolecular energy exchange (see pathway v, Fig. 1), a process discussed in greater detail in Sect. 6. Formal distinction is drawn between electronic energy exchange permitted by the ∆S = 0 rule and that forbidden by it. The term internal conversion (IC) is applied to radiationless transitions between states of the same spin multiplicity, while intersystem crossing (ISC) refers to energy exchange between states belonging to different (spin) systems. Both IC and ISC are assumed to take place with no change in total electronic plus vibrational energy, and the wavy lines are therefore horizontal (i.e. no translational or rotational energy is released in an intramolecular electronic energy exchange). Absorption of radiation in a singlet-triplet transition is weak, since it is forbidden in the same way as the triplet-singlet phosphorescent emission. It follows that phosphorescence can only be excited inefﬁciently by direct ab-

Basic Concepts of Photochemical Transformations

25

Fig. 7 Jablonski diagram showing absorption, ﬂuorescence, and phosphorescence. (Reproduced from R. P. Wayne, Principles and applications of photochemistry, Oxford University Press, Oxford, 1998. By permission of Oxford University Press)

sorption of radiation, and phosphorescence is much more usually the result of emission from a triplet populated by ISC from an excited singlet formed on absorption. The sequence of events is illustrated in Fig. 7. Absorption populates Sv1 ; vibrational energy, at least in condensed phases, is rapidly degraded and S01 can then lose its energy by radiation, ISC to T1 , or IC to Sv0 . It is, perhaps, surprising that ISC to T1 , which is spin-forbidden by a radiationless transition selection rule, can compete effectively with spin-allowed ﬂuorescence and IC to Sv0 ; phosphorescence is, however, observed in many systems, suggesting that IC from S1 to S0 is relatively inefﬁcient. 5.2 Fluorescence An electronically excited atom must lose its energy either by emission of radiation or by collisional deactivation: chemical decomposition is not possible, and radiationless degradation (involving an increase in translational energy) is extremely improbable. At low enough pressures, therefore, ﬂuorescent emission is expected from all atoms. Many molecular species, however, either do not exhibit ﬂuorescence or ﬂuoresce weakly even when bimolecular reaction or physical deactivation does not occur. Some general principles

26

R.P. Wayne

can suggest whether a polyatomic organic molecule is likely to be strongly ﬂuorescent. First, absorption must occur at a wavelength long enough to ensure that chemical dissociation does not take place. Absorption to an unstable state is clearly very unlikely to result in ﬂuorescence. Further, in many molecules in which the absorption maximum corresponds to an energy greater than the cleavage energy of the least stable bond, no ﬂuorescence is observed. Second, intramolecular energy transfer must be relatively slow compared with the rate of radiation. This appears to mean that ISC from S1 to T1 must be slow; ISC is normally slower from 1 (π, π∗ ) to 3 (π, π∗ ) states than for ISC involving n, π∗ states, and the efﬁciency of the process increases as the energy separation S1 and T1 decreases. Experimental observations of ﬂuorescence are in accord with these ideas: the simpler carbonyl compounds, in which the longest absorption corresponds to an n → π∗ transition, are rarely ﬂuorescent (but often phosphorescent), while aromatic hydrocarbons (π → π∗ absorption) are frequently ﬂuorescent. Increasing conjugation in hydrocarbons shifts the ﬁrst (π → π∗ ) absorption maximum towards longer wavelengths, and thus increases the probability of ﬂuorescence rather than decomposition. Geometrical factors such as rigidity and planarity also affect the efﬁciency of ﬂuorescence. The simplest type of ﬂuorescence is resonance ﬂuorescence, in which the radiation emitted is of the same wavelength as the exciting radiation. Resonance ﬂuorescence is observed only in the gas phase at low pressures, and only with atoms or simple molecules. Transitions also occur from υ to υ levels higher than zero, so that a progression of bands is observed at wavelengths longer than the exciting wavelength in accordance with an empirical observation of Stokes: the lines are called Stokes lines. Irradiation by polychromatic light can obviously excite many υ levels, and ﬂuorescent emission can then be observed from all these levels, up to the dissociation limit. Stepwise collisional relaxation of vibrational excitation is a relatively efﬁcient process, cross sections for single-quantum deactivation being between 1 and 100% of the gas-kinetic cross section for many quenching gases. Resonance ﬂuorescence is not expected, therefore, at pressures at which the kinetic collision frequency greatly exceeds the spontaneous emission rate, and, for A ∼ 108 s–l , observation of resonance emission is conﬁned to pressures at least below about 1 mmHg (and less, if A is smaller than suggested). Lower vibrational levels of the upper electronic state are populated from the level produced on absorption, and at moderate pressures, at which emission and vibrational quenching still compete, emission may be observed from all vibrational levels in the upper state up to υ . At higher gas pressures, at which the collision rate greatly exceeds the rate of emission, vibrational relaxation is essentially complete, and no ﬂuorescence is observed from υ > 0. Vibrational relaxation is extremely probable in solution, and ﬂuorescence from vibrationally excited levels is never observed in the liquid phase. Furthermore, neither the ﬂuorescence spectrum nor de-

Basic Concepts of Photochemical Transformations

27

activation rates are affected by changes in excitation wavelength so long as it lies within the absorption bands. S0 → S1 transitions in organic compounds are often partially forbidden; to obtain sufﬁcient light absorption to render gas-phase ﬂuorescence detectable frequently requires large pressures, which result in vibrational relaxation to υ = 0. This relaxation, together with the probability of radiationless loss in complex species, accounts for the rarity of resonance or vibrationally hot emission phenomena in organic molecules. The intensity of each vibrational emission band depends, in the same way as absorption intensities, on the operation of the Franck–Condon principle. Simple diatomic species frequently have greatly different internuclear separations in ground and excited states: Figure 5 is representative of the type of situation: in the B and X states of O2 , the upper state is larger than the lower state. Absorption in O2 originates almost entirely from υ = 0, and is most probable to υ levels from 7 to 11, and λmax ∼ 185 nm, while ﬂuorescence, at pressures at which vibrational relaxation is complete, is strongest around the (0, 14) band at λ ∼ 340 nm. In contrast, the (0, 0) band is the most intense for many organic molecules, and the maxima of intensity both in absorption and in emission therefore correspond to the same transition. This observation suggests that upper and lower electronic states of such molecules must be of similar size and shape, and it is likely that the vibrational spacings will be the same in both states. In such cases, the absorption and emission spectra may be almost exact mirror images of each other when displayed on an energy (e.g. wavenumber) scale, because, with similar spacings, the (0, 1) emission band will lie at the same energy below the (0, 0) band as the (1, 0) absorption band lies above it, and so on. This mirror-image relationship is of frequent occurrence in the ﬂuorescence of organic substances; assumption of its existence can be useful in sorting out overlapping emission spectra. Whether or not there is a mirror image, the spacing of emission bands indicates the vibrational levels in the ground electronic state, while the spacing of absorption bands depends on vibrational spacing in the upper state. Sometimes, the mirror image has a central gap, because there is a separation between the (0,0) bands in absorption and emission. The separations are caused by energy loss to the solvent environment. The equilibrium interactions with the solvent may be different for ground and excited states of the solute (these are mainly electrical interactions, via the dipole moment of the solute if sizes are similar in both states). Although the species cannot relax to the equilibrium interaction energy during the absorption process, it can do so before ﬂuorescent emission occurs. The magnitude of the separation depends on the dipole moment of the excited state of the emitting species, and also on the polarity of the solvent. Measurement of (0, 0) band separations can be used to estimate dipole moments of excited species. Organic ﬂuorescence usually originates from the lowest excited singlet level, S1 , even though absorption may initially populate a higher singlet (e.g.S2 , S3 , ... Sn ). Apparently there is rapid IC from Sn to S1 , followed by vi-

28

R.P. Wayne

brational degradation. IC from S1 to S0 must be much slower than from Sn to S1 , since emission from S1 can compete with radiationless loss and, perhaps more tellingly, since forbidden ISC from S1 to T1 can compete and lead to phosphorescent emission. The efﬁciency of intramolecular energy transfer usually decreases as the energy difference between two levels increases, and it is possible that the inefﬁciency of the S1 to S0 process, especially compared with that of Sn to S1 , reﬂects the relatively large energy separation between S1 and S0 . 5.3 Phosphorescence Organic molecules trapped in rigid glassy media often show a phosphorescent afterglow following irradiation by light. As explained in Sect. 5.1, phosphorescence in organic molecules is emission of a forbidden band, usually originating from a triplet level. Because of the long radiative lifetime of such transitions, collisional deactivation of the triplet competes effectively with radiation, and visible phosphorescence is not normally observed unless the collisional deactivation rate is sufﬁciently reduced. In rigid media, species are unable to diffuse towards each other, and bimolecular deactivation is slow. The earliest investigations of phosphorescence employed solutions of dyes in gelatin, and subsequently in boric acid glass at room temperature. More satisfactory rigid media are now used: mixtures of ether, isopentane, and ethanol (EPA) frozen at liquid nitrogen temperature (77 K) are frequently employed, and thin ﬁlms of various plastics are becoming popular as rigid matrices. The highest purity of the solvents is necessary to avoid swamping the phosphorescence of the solute by luminescence of the impurities. Although the ﬁrst observations of phosphorescence were conﬁned to rigid glasses, it was soon appreciated that phosphorescence could appear in other phases. Emission from biacetyl vapour is one of the best-known examples of gas-phase phosphorescence. Fluid solutions of species that are phosphorescent in low-temperature glasses also generally show emission, so long as the radiationless transitions from T1 to S0 do not show an increased rate at the higher temperatures. It is, of course, essential that the solvent does not deactivate the triplet, and that quenching impurities are rigorously excluded. Residual impurities may still make the emission intensity weak, and artiﬁcially reduce the luminescence lifetime. Perﬂuoroalkanes make suitable solvents for the study of phosphorescence at room temperature. Conﬁrmation that the emitting species in phosphorescent organic molecules is a triplet has come from several sources. In the 1940s it was discovered that a solution of ﬂuorescein in boric acid glass became paramagnetic under intense irradiation; more recently it has been shown that the paramagnetism and the phosphorescence decay at identical rates when irradiation ceases. The electron paramagnetic resonance (EPR) technique is capable of detect-

Basic Concepts of Photochemical Transformations

29

ing triplet species. The ﬁrst unambiguous detection of a triplet by EPR was of the ∆M = + 1 transition in an irradiated single crystal of naphthalene in durene; ∆M = + 2 transitions have also been observed in irradiated naphthalene. Triplet concentrations, measured by EPR, in solid solutions of certain phosphorescent aromatic ketones decay, after irradiation, at the same rate as the phosphorescence. Optical absorption to a higher triplet has afforded further evidence that the emitting state in phosphorescence is a triplet. Intense irradiation of a boric acid glass containing ﬂuorescein leads to the appearance of a new absorption band due to triplet–triplet absorption. Flash photolysis, in which a sample is exposed to a brief, intense ﬂash of light, can be used to produce high transient concentrations of triplet species: kinetic absorption spectroscopy of the system enables the build-up and decay of several singlet and triplet levels to be followed as a function of time. Phosphorescence most commonly follows population of T1 via ISC from S1 , itself excited by absorption of light. The T1 state is usually of lower energy than S1 , and the long-lived (phosphorescent) emission is almost always of longer wavelength than the short-lived (ﬂuorescent) emission. The relative importance of ﬂuorescence and phosphorescence depends on the rates of radiation and ISC from S1 ; the absolute efﬁciency depends also on intermolecular and intramolecular energy-loss processes, and phosphorescent emission competes not only with collisional quenching of T1 but also with ISC to S0 . 5.4 Luminescence Quenching: Kinetics and Radiative Lifetimes Bimolecular deactivation (pathway vii, Fig. 1) of electronically excited species can compete with the other pathways available for decay of the energy, including emission of luminescent radiation. Quenching of this kind thus reduces the intensity of ﬂuorescence or phosphorescence. Considerable information about the efﬁciencies of radiative and radiationless processes can be obtained from a study of the kinetic dependence of emission intensity on concentrations of emitting and quenching species. The intensity of emission corresponds closely to the quantum yield, a concept explored in Sect. 7. In the present section we shall concentrate on the kinetic aspects, and ﬁrst consider the application of stationary-state methods to ﬂuorescence (or phosphorescence) quenching, and then discuss the lifetimes of luminescent emission under nonstationary conditions. Observable effects in the quenching of ﬂuorescence are usually the result of competition between radiation and bimolecular collisional deactivation of electronic energy, since vibrational relaxation is normally so rapid, especially in condensed phases, that emission derives almost entirely from the ground vibrational level of the upper electronic state. The simplest excitationdeactivation scheme, which does not allow for intramolecular radiationless

30

R.P. Wayne

processes, is ∗

X + hν → X X∗ + M → X + M X∗ → X + hν

rate absorption Iabs quenching kq [X∗ ][M] emission A[X∗ ]

(21) (22) (23)

Solution of the steady-state equations for [X∗ ] (i.e. with d[X∗ ]/dt = 0) provides an expression for the luminescence emission intensity, Ilum , in terms of the intensity of absorbed radiation, Iabs , where A is the Einstein coefﬁcient for spontaneous emission: AIabs . (24) Ilum = A[X∗ ] = A + kq [M] Equation 24 can be inverted to give the Stern–Volmer relation kq [M] 1 1 . 1+ = Ilum Iabs A

(25)

If, therefore, 1/Ilum is plotted as a function of [M], the ratio of slope to intercept provides a value of kq /A, even if Ilum is measured in arbitrary units and Iabs is not determined. Thus, if the Einstein A factor is known, or can be measured, the value of the quenching rate constant can be calculated. The A factor can be calculated from the B factor by use of the ν3 relationship presented as Eq. 9 (and B itself can be calculated from the measured integrated extinction coefﬁcient for the absorption band, as implied by Eq. 15). It is also possible, under suitable conditions, to measure A directly by observation of the decay of emission after suddenly extinguishing the illuminating beam. As will be explained at the end of this section, the ﬂuorescence or phosphorescence lifetime may be shorter than the natural radiative lifetime as a result of intermolecular and intramolecular nonradiative energy degradation, so that due care must be taken in the interpretation of emission decay measurements. Rate constants for quenching can be compared with those predicted by the collision theory of chemical kinetics. According to this theory, a rate constant, k, is given by 8kT 1/2 2 exp(– Ea /RT) (26) k = σcoll πµ (σcoll is the collision cross section, equal to π(rA + rB )2 , where rA , rB are the gas-kinetic collision radii of the reaction partners, and µ is their reduced mass; Ea is the activation energy for the reaction). Ea is expected to be near zero for collisional quenching, so one way of making the comparison is to 2 . calculate the quenching cross section σq2 from kq , and compare it with σcoll 2 corresponds to the familiar P factor of the The ratio of cross sections σq2 /σcoll collision theory (assuming that Ea = 0). Quenching efﬁciencies increase with

Basic Concepts of Photochemical Transformations

31

increasing complexity of M: for polyatomic quenchers the efﬁciency can approach (or even sometimes exceed) unity and even for monatomic species such as He in the gas phase, quenching may occur on between one in a hundred and one in ten collisions. The relatively great rates of the quenching process may mean that in solution the rate is determined more by the rate of diffusion of the quenching and emitting molecules than by the rate of collision. An approximate expression for the diffusion-limited rate constant, kdiff , is given by the Debye equation kdiff =

8RT × 103 dm3 mol–1 s–1 , 3η

(27)

where η is the viscosity of the solvent in newton-seconds per square metre and R, the gas constant, is 8.3 J K–1 mol–1 . For H2 O at room temperature, η = 10–3 N s m–2 , so that kdiff = 1010 dm3 mol–1 s–1 . Note that, since η is temperature-dependent, kq may increase with temperature so that there appears to be an activation energy for the process; however, the true Ea , in the sense of the energy needed for reaction once a collision occurs, can still be zero. Although the quenching rate approaches the diffusion-controlled limit, it is not necessarily true that every molecular collision leads to deactivation. The diffusive process limits the rate at which the excited species and the quencher come together, but prolongs each encounter so that several hundred collisions are possible before the two species diffuse apart. The kinetics of the emission process has been developed in terms of excitation, emission, and collisional deactivation steps. If intramolecular energyloss processes (IC or ISC) occur, then additional ﬁrst-order terms must be added to the denominator of Eq. 24. A similar, but more complex and extended, steady-state treatment can be developed to predict the intensity of phosphorescent emission. It is in the nature of steady-state kinetic calculations that ratios of rate constants are obtained: for example, the expressions for the intensity in Eq. 25, or the parameters extracted from the Stern–Volmer treatment, involve ratios of rate constants to the Einstein A factor for emission. Individual rate constants can often be determined from a comparison of kinetic data obtained under stationary conditions with those obtained under nonstationary conditions. For the present purposes, the nonstationary experiment often involves determination of ﬂuorescence or phosphorescence lifetimes (τf , τp ). If a process follows ﬁrst-order kinetics described by a rate constant k, the mean lifetime, τ (the time taken for the reactant concentration to fall to l/e of its initial value), is given by τ = 1/k .

(28)

If the loss of luminescent species after the exciting radiation is shut off is unimolecular or pseudo-ﬁrst-order, then we may deﬁne mean lifetimes for the decay of emission as the inverse of the sum of all the effective ﬁrst-order rate

32

R.P. Wayne

constants. Thus, for the very simple scheme of the reactions in Eqs. 21–23 τ = (A + kq [M])–1 .

(29)

Measurement of τ as a function of [M] and extrapolation to [M] = 0 then yields a value for A; since A can also be calculated, via B, from extinction coefﬁcients in absorption, the measured value may afford a check on the calculated one. In general, when [M] = 0, the observed lifetime is shorter than the natural radiative lifetime (1/A). Where intramolecular loss processes (radiationless transfer) occur, and are described kinetically by a ﬁrst-order rate coefﬁcient kt , then τ = (A + kt + kq [M])–1 .

(30)

If a reliable value of A may be calculated from B, then kt may be determined explicitly from τ. Most quoted rate constants for IC or ISC processes derive, in fact, from lifetime measurements.

6 Electronic Energy Transfer and Electron Transfer The transfer of electronic energy (and excitation) between two different molecules (intermolecular energy transfer; pathway iv of Fig. 1) or between two different electronic states (intramolecular energy transfer; pathway v) plays a central role in photochemistry. The equivalent processes involving electrons, in which charge is transferred to a different molecule, or is moved in position in one molecule, may also be promoted photochemically. These processes are explored in the present section. 6.1 Intramolecular Energy Transfer Radiationless electronic energy transfer has already been introduced in connection with the IC and ISC processes represented in the Jablonski diagram of Fig. 7 that explain some features of ﬂuorescence and, especially, phosphorescence. As already mentioned, photofragmentation of complex molecules often proceeds via the mechanism of predissociation, in which radiationless energy transfer populates, from the state reached by absorption, a new electronic state above its dissociation limit. Complex organic molecules do not usually undergo optical dissociation in the regions of strongest absorption. The increased number of electronic states, the closer spacing between them, and the large number of vibrational modes all tend to increase the probability of radiationless transitions between states. As for emission processes, the occurrence of transitions with ∆S = 0 is usually a result of spin–orbit coupling in the molecule, and the transition probabilities for ISC follow virtually

Basic Concepts of Photochemical Transformations

33

the same pattern as those discussed in the last section for radiative processes. On the basis of these ideas, El-Sayed has suggested the following rules for spin-forbidden intramolecular energy transfer: 1 or 3

(n, π∗ ) ↔ 3 or 1 (π, π∗ );

3

(n, π∗ ) 1 (n, π∗ );

3

(π, π∗ ) 1 (n, π∗ ) . (31)

The preceding discussion is concerned with the electronic component in the probability of particular types of radiationless transition. Experimental evidence, some of which will be presented later, shows that the probability for energy transfer is also some inverse function of the energy gap between the two states for a given type of electronic transition. This result can be understood in terms of the operation of the Franck–Condon principle in radiationless transitions. The principle (discussed for radiative transitions in Sect. 4.3) proposes that the nuclei of a molecule do not move during the course of an electronic transition: i.e. the transitions are vertical on a potential energy diagram. In an intramolecular radiationless transition, the sum of the electronic and the vibrational energies must remain constant, in distinction to the radiative case where the photon provides or removes the energy difference between starting and ﬁnishing states. In the radiationless case, therefore, the transition is horizontal as well as vertical, so that it is conﬁned to a very small region of a potential energy curve or surface. The overlap in this region between the vibrational probability functions for starting and ﬁnishing states will then determine the efﬁciency of energy transfer for a ﬁxed electronic transition probability. Figure 8 illustrates three possibilities: the curves given may be regarded as potential energy curves for diatomic molecules, or as cross sections through surfaces for more complex species. In Fig. 8a, the two states, X and Y, are of similar geometry, but of widely different energy. The lowest vibrational level, υ = 0, in X has the same total energy as high υ in Y. Because of the nature of the vibrational probability distributions, the overlap is small. In Fig. 8b, however, the energy gap is much smaller, and the difference in vibrational quantum numbers υ and υ is also smaller, with the result that there is far greater vibrational overlap. Thus, the efﬁciency of crossing will increase as υ → 0: i.e. crossing to a state will be favoured if the state can be populated near υ = 0, which means that the electronic energy gap itself must be small. Only if the geometries of X and Y are different, as in Fig. 8c, can there be rapid radiationless conversion between two states of widely separated electronic energies. In general, energy separations lie in the order (S1 – S0 ) > (T1 – S0 ) > (S1 – T1 ), and the rate constants usually lie in the reverse order, even though the second two processes are formally spin-forbidden. Molecular rigidity favours efﬁcient ﬂuorescence of a species. Without that rigidity, changes in geometry may occur, and S1 → S0 IC can depopulate S1 . The rate of spin-forbidden transitions may be perturbed by the external environment. Such an inﬂuence is seen in the effects of the addition

34

R.P. Wayne

Fig. 8 The Franck–Condon principle in radiationless transitions. In a the geometries of states X and Y are similar, but the energy separations are large and the vibrational overlap is small. The separation is smaller in b, and the overlap is larger, while c shows that signiﬁcant overlap is possible for large energy separation if the geometries of X and Y are different. (Reproduced from R.P. Wayne, Principles and applications of photochemistry, Oxford University Press, Oxford, 1998. By permission of Oxford University Press)

of paramagnetic molecules to the solvent. Although O2 and NO decrease phosphorescence yields because of their participation in efﬁcient bimolecular quenching, they increase the rates both of optical transition and of ISC. Absorptions of the T1 ← S0 transition are also increased in intensity when the paramagnetic species is present, and, for example, the T1 ← S0 absorption in benzene (λ ∼ 310–350 nm) practically disappears when the last traces of oxygen are removed. The most dramatic demonstration of the increase in T ← S absorption is afforded by pyrene solutions, which are normally colourless but which turn deep red in the presence of high pressures of oxygen. Heavy atoms in an environment also increase the probability of S ↔ T radiative and radiationless transitions by inducing appreciable spin–orbit coupling in the solute. Thus, solutions of anthracene and some of its derivatives become less ﬂuorescent on addition of bromobenzene, while the triplet–triplet absorption intensity increases as a result of enhanced S1 → T1 ISC. Intramolecular perturbation of transition probabilities is also important. For example, in naphthalene and its monohalogenated derivatives, C10 H7 X, both T1 → S0 + hν radiative emission and T1 → S0 radiationless ISC show increasing rate coefﬁcients in the order of X is H, Cl, Br, I, and with X is I both rate coefﬁcients are 3 orders of magnitude or more larger with X is I than with X is H. One consequence is that phosphorescence, which is weak compared with ﬂuorescence in naphthalene itself, becomes by far the dominant process in iodonaphthalene. Strong intramolecular perturbations may also arise when certain metal ions are chelated to an organic molecule. It is interesting that the natural porphyrins chlorophyll and haemin display markedly different photochemical behaviour: chlorophyll has diamagnetic magnesium as its central ion, while haemin has paramagnetic iron.

Basic Concepts of Photochemical Transformations

35

6.2 Intermolecular Energy Transfer In intermolecular exchange of energy between two discrete partners, an acceptor (A) receives excitation from the donor (D), and A then participates in those processes open to it as an electronically excited species. Photosensitized phenomena, in which the change of interest occurs in a species other than the one that absorbed radiation, are of great signiﬁcance in photobiology; they also provide valuable insights into photophysical processes. Exchange of energy between two different species is not so restricted with regard to exact equivalence of internal energy between initial and ﬁnal states as in the case of intramolecular exchange, since an energy excess can be taken up by translation (or, more rarely, a deﬁciency supplied by the kinetic energy of collision). Ten different types of energy exchange can be classiﬁed according to the modes (electronic, vibration, rotation, and translation) between which the exchange occurs; except in the rare case of exact energy resonance, some energy is always converted to or from translation. The degradation of electronic excitation to vibration, rotation, or translation is responsible for the physical quenching of ﬂuorescence and phosphorescence. In the present section, the emphasis is on electronic–electronic energy exchange, and it will be assumed that excess energy goes into other modes of excitation. Where the absorption spectrum of the acceptor overlaps the emission spectrum of the donor, there are, of course, quantized vibronic levels of A and D for which energy exchange is isoenergetic, and no increase in kinetic energy is needed. For electronic energy exchange between atoms, some translational energy is almost always released. 6.2.1 Radiative Energy Exchange Several different mechanisms of electronic energy transfer are believed to operate under different circumstances. The ﬁrst of these is the so-called trivial mechanism of radiative transfer, which can be represented by the processes D∗ → D + hν , A + hν → A∗ .

(32) (33)

The mechanism is trivial in name and simplicity only, since it is the one energy transfer mechanism that can operate over very large separation of D and A: the interaction necessarily follows the laws of light propagation. Radiative energy transfer is all-important to our existence, because it is how we receive the energy of reactions occurring in the Sun; and the related radiative energy transfer processes occurring in upper and lower atmospheres establish the temperature equilibria and meteorological conditions upon which we

36

R.P. Wayne

depend. The efﬁciency of radiative transfer is a function of the overlap between the emission spectrum of D and the absorption spectrum of A (a factor that appears in all transfer mechanisms), and also of the size and shape of the sample: since D∗ will emit in all directions, the probability of radiative transfer increases with sample volume. It will be obvious that experiments designed to study nonradiative energy transfer must eliminate or make due allowance for the radiative process. 6.2.2 Collisional Energy Transfer Short-range energy transfer arising from exchange interaction occurs over intermolecular or interatomic distances (henceforth referred to as r) not much exceeding the collision diameter; the interaction decreases in a complex fashion with r raised to a high power. Energy transfer by exchange interaction may be thought of as a special kind of chemical reaction in which the chemical identity of the partners A and D does not change, but in which excitation is transferred from one to the other. The transition state is then expected to possess a separation between A and D not greatly different from the sum of the gas-kinetic collision radii, and energy transfer by the exchange mechanism is probably important only for values of r of this order. In common with other chemical processes, energy transfer can be efﬁcient only if the potential energies of reactants and products are connected by a continuous surface that describes the potential energy of the system as a function of the several interatomic distances: a reaction occurring on such a surface is said to proceed adiabatically. In other words, the reactants and products must correlate with each other and with the transition state. Most chemical reactions involving ground-state partners can occur adiabatically, but in processes such as energy exchange, where several electronic states are involved, the requirement for adiabatic reaction may impose some restrictions on the possible states for A, A∗ and D, D∗ if there is to be efﬁcient transfer of excitation. In atoms or small molecules there must be correlation of electron spin, orbital momentum, parity, and so on. However, correlation in complex molecules of low symmetry usually only involves the electron spin. To test for correlation, the possible total spin of the transition state is calculated from the individual spins of the reactants by the addition of the quantum vectors. Thus, for reactants A and B with spins SA and SB , the total spin of the transition state can take magnitudes |SA + SB |, |SA + SB – 1|, ..., |SA – SB |. It is then necessary to see whether the products, X and Y, can also give at least one of the same total spin magnitudes in the transition state. For example, a triplet donor and a singlet acceptor can form adiabatically only a (triplet + singlet) product pair; on the other hand, (triplet + triplet) reactants can yield (singlet + singlet), (triplet + triplet), or even (doublet + doublet) products. Rules of this kind are referred to as the Wigner spin-correlation rules.

Basic Concepts of Photochemical Transformations

37

Although a reaction is likely to be efﬁcient only if it is adiabatic, nonadiabatic reactions can also occur. We may look upon a nonadiabatic reaction as one in which crossing occurs between two intersecting or closely approaching potential energy surfaces. The crossing process is governed by the ordinary selection rules for radiationless transitions. In particular, a spin-forbidden reaction cannot proceed adiabatically because no common spin states can be written for the transition complex, and potential surfaces for transition states derived from reactants and products must be of different multiplicity. Hence, ∆S = 0 for the intramolecular energy transfer, the crossing is of low probability (cf. Sect. 6.1), and the efﬁciency of the nonadiabatic intermolecular energy transfer is small. Energy exchange in molecules occurs efﬁciently if the amount of kinetic (translational) energy that must be liberated is small; thus, rapid exchange of energy is expected between vibronic levels in near-resonance. Excess vibrational energy will be degraded rapidly (at least in condensed-phase systems), and the acceptor molecule left in its ground vibrational state. The apparent energy gap is therefore the electronic difference between D∗ and A∗ , although this is not ∆E for the actual transfer process. Anomalies in the dependence on ∆E of the rate of transfer may arise because of the difference between actual and apparent energy gaps. The efﬁciency of energy transfer occurring by the exchange interaction mechanism is related to whether the process can take place adiabatically, but not to whether optical selection rules permit radiative transitions in donor and acceptor; this behaviour is one way in which exchange interaction and long-range Coulombic interaction (see Sect. 6.2.3) may be distinguished. For example, in the exchange interaction excitation of triplet states by triplet benzophenone, the efﬁciency of energy transfer is roughly the same both for naphthalene and for 1-iodonaphthalene. In Sect. 6.1, it was shown that the T1 → S0 radiative transition is more probable by a factor of at least 1000 in the substituted molecule, so that in this case the optical transition probability in the naphthalene molecule does not seem to affect the probability of energy transfer to it. Energy may be transferred from one excited species, D∗ , to another, A∗ , already possessing some excitation, to raise the latter to a higher electronic state A∗∗ , D∗ + A∗ → D + A∗∗ .

(34)

This process of energy-pooling must be relatively slow at low concentrations of D∗ and A∗ , since bimolecular collision of the excited species will be a rare event. Formation of excited species by energy-pooling has, however, been recognized in several systems, and provides the mechanism for P-type delayed ﬂuorescence. Energy-pooling may also permit the occurrence of chemical reactions that require more energy than is available in a single quantum of

38

R.P. Wayne

radiation, and such energy storage may be a necessary step in several photobiological systems. 6.2.3 Coulombic Energy Transfer For many energy transfer processes, the interaction takes place when the partners are separated by more than the sum of the gas-kinetic collision radii. For example, energy transfer between excited singlet states of hydrocarbons occurs as fast as spontaneous decay at concentrations in benzene corresponding to a distance, r, between exchanging molecules of about 5 nm, or about 10 times the collision diameter. The measured rate constants for transfer of excitation in the hydrocarbons also seem greatly to exceed the diffusionlimited rate, and do not depend on solvent viscosity. Long-range transfer may arise by sequential short-range excitation of many species so that the excitation appears ultimately at a place distant from the original location of excitation. Long-range transfer is, however, predicted to occur also by a direct mechanism sometimes called inductive resonance involving electrical, or Coulombic, interactions between transition dipoles (or higher multipoles). These multipoles are the ones involved in optical interactions with the electric vector of radiation: the usual optical selection rules apply to both the transitions D∗ → D and A → A∗ , and dipole–dipole interactions are stronger than dipole–quadrupole interactions, and so on. For a dipole–dipole interaction, theory predicts that the strength of the interaction should fall off as 1/r6 , and relatively long range energy exchange becomes possible. The Förster equation for the ﬁrst-order rate constant, ke , for energy transfer by the inductive-resonance mechanism can be written in simpliﬁed form 1 ke ∝ 4 6 n τD r

∞ fD (ν)εA (ν) 0

dν , ν4

(35)

where r is the separation between the transferring molecules, n is the refractive index of the solvent, and τ is the mean lifetime of the donor. fD (ν) is the normalized spectral distribution of the donor emission and εA (ν) is the absorption spectrum (coefﬁcient) of the acceptor. This equation reveals several important features about the efﬁciency of energy transfer by the inductiveresonance mechanism. First is the 1/r6 dependence of the rate of transfer. Second is the increased efﬁciency for shorter donor lifetimes. Third is the need for overlap of the donor emission and acceptor absorption spectra (as well as a strong dependence on both ν and n). According to the spin selection rule, ∆S = 0, long-range Coulombic transfer should be impossible for any process involving multiplicity changes, and long-range triplet–triplet energy transfer would then be excluded. However, to the extent that spin–orbit coupling allows electric-dipole optical transi-

Basic Concepts of Photochemical Transformations

39

tions with ∆S = 0 in complex molecules, Coulombic transfer can occur by the dipole–dipole mechanism. Transfer is likely to be slower than for exchange processes in which transitions for donor and acceptor are fully allowed, but, since the actual radiative lifetimes of the triplet states are also long, the longrange energy transfer process may still be important relative to radiation. It follows that the long-range interaction is likely to be demonstrated only in systems in which quenching or ISC is not the predominant loss process for the donor triplet. 6.2.4 Exciton Migration An even longer-range transfer, showing a 1/r3 dependence, may occur in crystals, solid solutions, and some ﬂuids, as a result of exciton migration. The concept of the exciton was introduced by Frenckel to interpret certain crystal spectra; an electron–hole pair was looked upon as an entity that could move about the crystal as a result of interactions between lattice sites. For the present purposes, the electronic excitation in an irradiated species can be regarded as an exciton that is free to wander over a considerable number of lattice sites. 6.3 Electron Transfer Photoionization (pathway viii, Fig. 1) represents the complete removal of an electron from a molecule as a result of the absorption of light, and it is the analogue of photodissociation processes, but where the electron and positive ion are the products. A + hν → A+ + e– .

(36)

In the gas phase, photoionization is usually a high-energy event, occurring only at short wavelengths, but in solution, the energy of solvation can lower the ionization threshold substantially. In aqueous media, the free electrons themselves become hydrated, and can act as strong reducing agents. Charge separation also plays a very important role in photochemistry even where the excited molecule is not fully ionized. An electronically excited molecule is often both a better electron donor (reducing agent) and a better electron acceptor (oxidizing agent) than the corresponding ground-state species, because in many cases an electron has been promoted to an antibonding orbital (thus leaving a hole in a bonding orbital). Processes such as A∗ + B → A· + + B· – A∗ + B → A· – + B· +

(37) (38)

40

R.P. Wayne

are intermolecular electron transfers, with A∗ being the electron donor or acceptor in the reactions in Eqs. 37 and 38, respectively. Electron transfer is widespread in photochemistry, being involved in processes as diverse as photosynthesis and silver halide photography. If A∗ and B are singlet or triplet states, then in the most usual cases, the product ions are doublets (ion radicals). Such radicals may then participate in the secondary reactions typical of such species. For example, they may react with O2 , or dissociate if weakly bound; and radical cations can act as acids or nucleophiles, while radical anions can act as bases or electrophiles. It should be noted, however, that in the bimolecular intermolecular electron transfer process the product ions are formed as a geminate pair, and return electron transfer decreases the efﬁciency of product separation (it is assumed here that the ion-pair states are of lower energy than any locally excited neutral pairs, {A∗ B} or {AB∗ }). Even for the geminate ion pair, two possibilities are envisaged: a tight ion pair, and a loose pair (in which the ions are perhaps separated by solvent). The ion pairs may exist as singlet or triplet states, which are nearly degenerate, especially in the case of the loose pair. The tight singlet pair is a singlet exciplex, and ISC may populate a triplet exciplex. Exciplex ﬂuorescence or phosphorescence may be competitive with radiationless return electron transfer. Intermolecular electron transfer seems normally to require virtually direct contact between the donor and acceptor molecules. According to a widely accepted theory, the rate constant for transfer decreases exponentially with the separation of the donor and acceptor species. In ﬂuids, the internuclear separation ﬂuctuates with time, so that transfer is dominated by the short-distance events. Under such circumstances, the transfer process can be regarded as a normal bimolecular reaction, for which in a transition-state formulation the rate coefﬁcient, kt , is characterized by a free energy of activation, ∆G= , equated somewhat arbitrarily with the activation energy, Ea , of the conventional Arrhenius expression: kt = A exp(– Ea /RT) = A exp(– ∆G= /RT) .

(39)

The change in standard Gibbs free energy, ∆G , for the processes will be the difference between the oxidation potential of the donor and the reduction potential of the acceptor, less the excitation energy of A∗ , but with a term added for the Coulombic (electrostatic) energy of attraction of the newly created ion pair. Rehm–Weller behaviour of the rates of photoinduced electron transfer reactions is based on this description of the free energy of reaction, and its applicability to the free energy of activation, ∆G= , for the transfer process. However, the Marcus–Hush model of electron transfer recognizes that ∆G= depends on more than just ∆G . In the simpliﬁed form of the Marcus model, transfer is presumed to occur at the point of intersection between two parabolic potential curves that are supposed to be representative of the

Basic Concepts of Photochemical Transformations

41

key portion of the free-energy hypersurface for the transferring species. Such behaviour is exactly that described for intramolecular energy transfer controlled by Franck–Condon overlap (Sect. 6.1). The value of ∆G= is thus equal to the (free) energy of this intersection point above the energy of the reactants. The intersection point itself depends on the energy, λ, required to move vertically from the reactant equilibrium geometry to the product curve, which is referred to as the reorganization energy. This reorganization energy reﬂects the difference in equilibrium geometries of initial and ﬁnal states, together with a contribution from the altered polarization of the solvent shell. For the parabolic curves envisaged by Marcus, it may be shown that Ea = ∆G= =

(∆G + λ)2 . 4λ

(40)

This equation predicts normal kinetic behaviour for – ∆G > λ, the rate coefﬁcient increasing with increasing – ∆G (increasing exergonicity), up to a maximum where – ∆G = λ. However, for – ∆G λ, Ea begins to increase again, and the rate coefﬁcient for electron transfer will decrease in what is known as the Marcus inverted region. Where activation energies are near zero (– ∆G = λ), transfer rates are likely to be much faster than the collision rate in solution, so that the kinetics become diffusion-controlled. Within the inverted region, the electron transfer leading to a lower state (e.g. ground state) of the system will become uncompetitive compared with transfer to a higher (excited) state for which – ∆G is smaller. Open-shell species, in which category radical ions fall, generally possess low-lying excited states, so that it has been difﬁcult to demonstrate the expected kinetic behaviour in the inverted region for intermolecular electron transfer, although it is now clear that it is relatively common in photochemical processes. Inverted behaviour is particularly clearly demonstrated for bimolecular return electron transfer within geminate radical-ion pairs. Despite the seeming requirement for what is essentially geometrical contact between the donor and acceptor, mechanisms can be envisaged for long-range photochemical electron transfer. Electron hopping would involve charge transfer to a solvent, with the separated electron possibly being delocalized over several solvent molecules: the subsequent hopping of the electron from one group of solvent molecules to another could occur much more rapidly than a collisionally diffusion limited process. The solvent can also act as a bridge between the donor and acceptor in a process where the donor, acceptor, and bridge form a loose supermolecule over which the electron is delocalized; the process is called superexchange. Since their small mass (momentum) invests electrons with pronounced tunnelling properties, quantummechanical tunnelling also plays a part not only in long-range transfer, but also in affecting the validity of any description of electron transfer that depends on overcoming classical energy barriers.

42

R.P. Wayne

So far, the emphasis in the discussion of electron transfer has been on intermolecular processes. Nevertheless, systems found in nature sometimes operate as a result of intramolecular electron-energy transfer between what are essentially two chromophoric groups attached to the same molecule. Considerable insight into the nature of photoinduced electron transfer has been obtained by deliberately constructing bichromophoric systems with the donor and acceptor groups held apart at a ﬁxed distance in some way. Under these circumstances, the electron transfer process is unimolecular, and more readily susceptible to theoretical interpretation, since the problems of diffusion do not arise. Bichromophoric molecules are known in which the donor and acceptor are separated by a rigid network of covalent spacers that appears to act as an electrical conductor. Intramolecular electron transfer over distances as large as 2 nm can then occur. Donor–acceptor pairs ﬁxed in space by protein frameworks, in rigid frozen media, and by electrostatic complexation are conceptually of the same kind. They all permit testing of theories that attempt to relate the rate of electron transfer to the geometrical separation of the partners; and the kinetic behaviour provides perhaps the most deﬁnitive test of Marcus theory as outlined earlier.

7 Efficiency of Photochemical Processes: Quantum Yields and Photonic Efficiencies The efﬁciency of photochemical processes has been alluded to several times in the preceding sections. It is now appropriate to provide a way of expressing these efﬁciencies quantitatively. 7.1 Homogeneous Systems: Quantum Yields For photochemistry occurring in a single phase (homogeneously), a concept of great value is that of the quantum yield or quantum efﬁciency, φ. As originally understood, φ was the number of molecules of reactant consumed for each photon of light absorbed (or, working in molar units, the number of moles consumed for each einstein—a mole of photons—absorbed). In this form, the quantum yield reﬂects, without distinction, both the efﬁciency of the primary photochemical process in bringing about chemical change and also the extent of secondary reaction. A quantum yield greater than unity suggests the occurrence of secondary reactions, since the Stark–Einstein law indicates that not more than one molecule can be decomposed in the primary step. However, chemical change is not the only consequence of absorption of radiation, as indicated by the several pathways in Fig. 1. The process to which the quantum yield refers should thus be speciﬁed. The quantum yield for any

Basic Concepts of Photochemical Transformations

43

such process is often dependent on the wavelength of radiation absorbed, and the relative contribution of the different pathways will alter as the wavelength is changed. In the discussion that follows, it is always assumed that the initial absorption event is a single-photon process; multiple-photon absorptions can occur with high-intensity light sources, in which case straightforward modiﬁcations must be made to the deﬁnitions and interpretations. The energy of an excited species must go somewhere, so the Stark–Einstein law leads to the conclusion that the sum of the quantum yields for all primary processes, including deactivation, must be unity. Where experimental data are available, this expectation is well substantiated. Measurement of quantum yields in all cases requires a method for measuring absolute numbers of photons or intensities (which are photon ﬂuxes). The conversion of light energy to heat, and the application of Planck’s relation (Eq. 1) is the ultimate connection between photon numbers or ﬂuxes and fundamental quantities. Devices such as thermopiles are employed to determine the heat energy. Secondary standards, such as calibrated carbon-ﬁlament lamps, are often employed; and calibrated photocells and photomultipliers or chemical actinometers are much more convenient for use in normal photochemical experiments. Chemical actinometry, in particular, is a valuable technique. Quantum yields have been determined for some easily and accurately measured photochemical changes using primary standards. A typical example is the photoreduction of Fe3+ to Fe2+ in K3 Fe(C2 O4 )3 (potassium ferrioxalate), in which the Fe2+ may be measured absorptiometrically, and the quantum yield is relatively insensitive to concentrations and other inﬂuences. In subsequent studies, the actinometer material is employed in place of the test substance, and the extent of chemical change is determined as a function of time in order to determine the intensity of absorbed light (and, through the Beer–Lambert law, the intensity of incident light). One great advantage of chemical actinometers is that they can often be used in the same reaction vessel—possibly of unusual shape—as that used for the material whose photochemistry is being studied. In determining quantum yields, the contribution to chemical change of secondary reactions must ﬁrst be eliminated or allowed for, and the absolute efﬁciencies of radiative and nonradiative energy-loss processes must be assessed. It is not always possible even to establish what primary paths exist, so that a full description of the primary processes in terms of quantum yields can be made only in favourable cases. Nevertheless, several indications may be used to gain some insight into the primary process. The nature of the absorption spectrum may suggest the electronic conﬁguration of the excited state and hence the probable fate of the energy. Detection of the intermediates (excited states as well as atoms and radicals) may reveal the products of the primary step. Measurement of overall quantum yields can also give some information about the primary process. If Φ 1, then in all probability little chemical change occurs in the excited absorbing molecule (although cage

44

R.P. Wayne

recombination of radicals in condensed-phase reaction systems is another very common cause of low quantum yields). A search must then be made for radiation emitted from the system; the spectrum will indicate whether the emission is ﬂuorescence of the absorber or whether it is derived from a state populated by intermolecular or intramolecular energy transfer. Study of ﬂuorescence quenching will yield information about physical deactivation processes. Again, if the quantum yield for formation of a speciﬁc product is invariant with experimental conditions, such as reactant concentrations or temperature, then that product probably appears, at the measured efﬁciency, in the primary process. 7.1.1 Chemical Change Where nothing to the contrary is stated, “overall quantum yield” usually implies the yield for removal of reactant, although, if several different secondary paths exist, it may be desirable to quote an overall quantum yield for the formation of a speciﬁc product. The determination of quantum yields for chemical change requires measurement of the numbers of molecules of reactant consumed, or of product formed, using suitable analytical techniques, as well as of the number of quanta of radiation absorbed. Since the photodissociation yield, for example, can be less than unity because of the competition of alternative primary losses of the species excited, secondary processes could be occurring in a photochemical reaction even though the overall quantum yield is less than unity. It is more helpful to consider primary and overall quantum yields, φ and Φ, separately. If, for example, dissociation precedes secondary chemical reactions, the primary quantum yield would be the number of molecules dissociating in the primary step for each quantum of light absorbed, and the ratio of overall to primary quantum yields then indicates the extent of secondary reaction. Quantum yields may also be deﬁned in terms of rates, rather than numbers of molecules and photons. An intensity of radiation, I, refers to an energy or photon ﬂux, and the absorbed intensity, Iabs , of Eq. 13 thus has the same units. Since this intensity is absorbed in a unit area volume of thickness d (see Eq. 10), and the number of molecules that react are contained in the same volume, it follows that Φ=

number of reactant molecules lost rate of reactant loss = number of photons absorbed Iabs

(41)

since a rate is the number of molecules per unit volume per unit time. The ratio of overall to primary quantum yields, Φ/φ, is analogous to the kinetic chain length, ν, determined in studies of thermal chain reactions. The quantities may be expressed in terms of rate constants for the several secondary reactions, and their variation with concentrations of various species may lead

Basic Concepts of Photochemical Transformations

45

to conﬁrmation of a hypothetical reaction mechanism and evaluation of rate constants. In thermal reactions, ν is deﬁned by the relation rate of reactant loss . (42) rate of initiation Furthermore, if we assume that the primary quantum yield, φ, is for formation of reactive intermediates, then φIabs is the rate of initiation in the photochemical system, and ν=

Φ rate of reactant loss = ≡ν. φ rate of initiation

(43)

7.1.2 Emission Processes The quantum yield for an emission process is deﬁned as the number of photons emitted for each photon absorbed, or, in rate terms, the intensity of emitted radiation divided by the absorbed intensity. Thus, for example, Eq. 24 for luminescence intensity can be rewritten φlum =

Ilum A . = Iabs A + kq [M]

(44)

In the more complex situation where intramolecular radiationless processes occur as well as radiative and collisional ones, the equation becomes φlum =

A , A + kt + kq [M]

(45)

where kt is the rate coefﬁcient for the intramolecular process, as introduced in connection with Eq. 30 for the lifetime of emission. A modiﬁed Stern–Volmer plot (Eq. 25) of 1/Ilum as a function of [M] can be extrapolated to [M] = 0 to allow a determination of kt /A. Since Eq. 30 can itself be extrapolated to [M] = 0 to provide a value for (A + kt ), it is evident that a combination of lifetime and quantum yields can be used to provide the quantities A and kt independently. These kinds of measurement do, indeed, lie at the heart of many determinations of the rate parameters for radiative and radiationless transitions. It may be noted, also, that if chemical reaction and decomposition can be discounted, and physical quenching of an excited species does not occur (or an extrapolation can be made to zero quencher concentration), then the sum of the quantum efﬁciencies for all the remaining processes must be unity: φf + φp + φISC + φIC = 1 ,

(46)

where the subscripts refer to ﬂuorescence, phosphorescence, ISC, and IC. Application of this equality can remove the necessity for experimental meas-

46

R.P. Wayne

urement of one of the quantum yields, or provide a valuable check on the consistency of the measurements if all four have been made. 7.2 Heterogeneous Systems: Photonic Efficiencies Many photochemical systems of great practical signiﬁcance involve two phases, such as solids suspended in liquids or other solids, or liquid or solid aerosol particles suspended in gases. Systems of these kinds include biological examples on the one hand and atmospheric clouds on the other. Industrial applications of photocatalysis using, for example, semiconductor particles of TiO2 are widespread. Quite often, the initial absorption event occurs in the suspended material, which may be regarded as the light harvester. The light harvester usually scatters or reﬂects light as well as absorbing it, so that a proper evaluation of the conventional quantum yield, as understood in the discussion of Sect. 7.1, would require an absolute measurement of the intensity of light absorbed by the harvester. Unfortunately, the term quantum yield (efﬁciency) has been used loosely in describing heterogeneous photochemistry. In particular, it has often been taken as the ratio of the number of photochemical events under consideration to the number of incident (rather than absorbed) photons of the active radiation, or, more usually, the ratio of the rate of a process to the incident light intensity. This failure to use the correct number of photons in reporting studies of heterogeneous photochemistry is regrettably sometimes compounded by an accompanying failure to accommodate the variation of the absorption coefﬁcients and (true) quantum yields with wavelength. Quantum yields, especially those for chemical change, often change with the extent of conversion; the use of the term in Sect. 7.1 always implies that there is no interference from product species, and experiments may have to be conducted to study initial amounts or rates of reaction. In heterogeneous photocatalytic studies, at least, one objective may be to obtain very high conversions; an obvious example is the photocatalysed removal of environmental pollutants in water. In such cases, the overall measured quantum yield can differ markedly from the initial yield at low conversions. In order to make it clear that the yields or efﬁciencies in heterogeneous studies refer to incident rather than absorbed intensities, the term photonic efﬁciency (ξ), deﬁned as ξ=

number of reactant molecules lost rate of reactant loss = , number of photons incident I0

(47)

has been adopted by some workers, with the deﬁnition being extended to the measurements being conducted in a cell with ﬂat parallel windows. As for quantum yields, ξ refers to a particular wavelength, which should be speciﬁed.

Basic Concepts of Photochemical Transformations

47

Use of ξ is an improvement over the use of φ in heterogeneous photochemistry, because it does at least make it clear that it is incident rather than absorbed photons or intensities that are under consideration. For all the reasons explained, ξ still does not provide any detailed insight into the underlying photochemistry. It may, however, be useful as a pointer to the efﬁciency of different modiﬁcations of a photocatalytic process, for example. For such purposes, the relative photonic efﬁciency (ξr ) may be preferred. This quantity compares ξ for the test reaction with ξ for some standard, employed much in the same way as a chemical actinometer is used in homogenous studies (see Sect. 7.1). One suggested secondary-actinometer material is phenol, so that ξr is deﬁned in this case as the ratio of the (initial) rate of substrate degradation to the (initial) rate of phenol loss, using an identical reactor, reactor geometry, and light source for the two studies.

8 Conclusion This chapter has attempted to provide a framework for the discussions of photochemical processes that appear in the material that follows in the book. It is concerned with the basic ideas of photochemistry, starting with the way in which light and molecules can interact as a consequence of the properties of electromagnetic radiation itself and the electronic structure of the species that absorb it. This approach places an emphasis on the electronic states that are excited in the absorbers, and photochemistry is seen as the embodiment of the ways in which the excitation brings about chemical and physical change. Molecules can dissociate or isomerize; they can ﬂuoresce or phosphoresce; and they can transfer or lose their excitation in intermolecular processes or rearrange it in intramolecular ones. The many pathways that determine the fates of electronic excitation in molecules are frequently in competition, and a quantitative interpretation of photochemical behaviour must have recourse to the tools of reaction kinetics. The ideas of quantum yields and photonic efﬁciencies are introduced in this context, since they describe the conversion of photon energy into the different available reactive and nonreactive pathways. Photochemistry plays a central part in environmental processes, both those that are natural and those that are altered by man, both biogenic and nonbiogenic, and the evaluation of the impact of light depends on the fundamental concepts presented here.

Hdb Env Chem Vol. 2, Part M (2005): 49–75 DOI 10.1007/b138179 © Springer-Verlag Berlin Heidelberg 2005 Published online: 16 September 2005

Environmental Photochemistry in Heterogeneous Media Mónica C. González1 · Enrique San Román2 (u) 1 Instituto

de Investigaciones Fisicoquímicas Teóricas y Aplicadas (INIFTA), Diagonal 113 y 64, C.C. 16, Suc. 4, 1900 La Plata, Argentina [email protected]

2 Instituto

de Química Física de los Materiales, Medio Ambiente y Energía (INQUIMAE), Ciudad Universitaria, Pab. II, 1428 Buenos Aires, Argentina [email protected]

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

50

2 2.1 2.2

Solid-State and Surface Photochemistry . . . . . . . . . . . . . . . . . . . Metals, Semiconductors and Insulators . . . . . . . . . . . . . . . . . . . . Interaction Between Surface and Adsorbate . . . . . . . . . . . . . . . . . .

51 51 53

3 3.1 3.1.1 3.1.2 3.1.3 3.1.4 3.1.5 3.2 3.3

Heterogeneous Reactions of Environmental Signiﬁcance The Atmosphere . . . . . . . . . . . . . . . . . . . . . . . Interaction Between Light and Aerosols . . . . . . . . . . Reactions in Clouds . . . . . . . . . . . . . . . . . . . . . Reactions on Snow, Ice and Saline Particles . . . . . . . . Reactions over Carbonaceous Particles . . . . . . . . . . Reactions of Polycyclic Aromatic Hydrocarbons . . . . . The Hydrosphere . . . . . . . . . . . . . . . . . . . . . . Soil Reactions . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . .

57 57 57 58 62 63 65 67 69

4

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

72

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

73

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

Abstract Heterogeneous reactions are widespread in all the environmental compartments: soils, hydrosphere, troposphere and at medium altitudes in the stratosphere. In particular, heterogeneous photochemical processes are concerned with the effect of light on interacting molecules and solid surfaces. Sorption, ion exchange, dissolution, precipitation and redox processes bear a decisive role on these reactions, commonly integrated within the concept of “photoinduced surface chemistry.” Photophysical and photochemical reactions in heterogeneous environments differ therefore substantially from analogous homogeneous reactions in respect to reaction rates, product distributions and stereochemistry. In this chapter we review environmental studies involving heterogeneous photochemical processes and introduce some relevant basic concepts needed for their understanding. Selected examples occurring in the atmosphere, waters and soil are presented and discussed in some detail. Whenever possible, emphasis is given to the interrelation between ﬁeld and laboratory studies. Keywords Aerosols · Heterogeneous photochemistry · Hydrosphere · Particulate matter · Soils

50

M.C. González · E. San Román

1 Introduction Heterogeneous reactions are widespread in the environment, not only in soils and in the hydrosphere but even in the troposphere and, at medium altitudes, in the dilute stratosphere. The generation of H2 SO4 and HNO3 as a result of the oxidation of gaseous SO2 and NOx in cloud droplets and the polar cloud depletion of stratospheric O3 were among the ﬁrst multiphase atmospheric processes recognized. Interfacial chemical processes are essential to environmental issues related to air, water and soil pollution, nutrient budgets and long-term global climate changes due to chemical weathering and biogeochemical cycling. The development of steady-state and pulsed irradiation methods and sensitive analytical procedures allowed photochemists to investigate in detail light-induced processes of environmental relevance in homogeneous gaseous and liquid phases. However, only in the last few decades was it fully realized that homogeneous reactions could not account, even qualitatively, for many observed facts. To understand chemical processes occurring in nature, the interaction between different phases has to be considered. Moreover, many relevant chemical and photochemical processes take place at phase boundaries, mostly following pathways entirely different from those commonly found in homogeneous media [1, 2]. Photophysical and photochemical reactions in heterogeneous environments differ substantially from analogous homogeneous reactions in respect to reaction rates, product distributions and stereochemistry. Sorption, ion-exchange, dissolution, precipitation and heterogeneous redox processes bear a decisive role on these reactions. The complexity of the natural environment renders the understanding of photochemical environmental processes through basic laboratory experiments very difﬁcult. The management of chemical systems including different compartments and multiple components, most of them at trace concentrations, is a great challenge for scientists. As an example, despite the present knowledge of gas-phase and heterogeneous chemistry, atmospheric processes are among the most intractable scientiﬁc problems in nature. Research activities on heterogeneous reactions in droplets and solid and semisolid particles increased during the last decade. However, several issues still need deeper investigation. Particle light scattering and the resulting complex radiation ﬁeld require the development of special methods to evaluate the quantity of light absorbed by the system. This is particularly important when heterogeneous processes are compared with their homogeneous homologous. Determination of reaction quantum yields is not as straightforward as in homogeneous systems and, regularly, only photon efﬁciencies related to incident, rather than absorbed light are calculated. It is usually

Environmental Photochemistry in Heterogeneous Media

51

difﬁcult to establish appropriate standards against which the efﬁciency of multiphase processes can be contrasted. Photochemical processes are induced in natural systems by solar radiation; therefore, only heterogeneous processes involving low ﬂuence are of importance. Photon statistics has to be taken into account, as every particle usually acts as an independent reactor. The development of laboratory experimental methods capable of following minute concentrations and concentration changes in the bulk of each phase and at interfaces is required. Sorption and diffusion models need to be included into the kinetic equations. All these factors have an enormous relevance in analyzing reaction mechanisms, which otherwise would be merely speculative. No less important are the interactions of chemical species among each other, with the substrate and with the medium and the effect of the surroundings on the physicochemical properties of particles. The latter issues are relevant in order to evaluate particle photosensitizing and photocatalyzing properties, both in natural environments and in advanced oxidative remediation procedures. It is the aim of the present chapter to review environmental studies involving heterogeneous photochemical processes and to introduce some relevant basic concepts needed for their understanding. Selected examples are presented and discussed in some detail. Whenever possible, emphasis is given to the interrelation between ﬁeld and laboratory studies. General concepts are introduced in Sect. 2 and speciﬁc examples are given in Sect. 3. These examples refer to heterogeneous reactions occurring in the atmosphere, waters and soil. However, evidence found in one compartment can often be used to understand processes occurring in the others.

2 Solid-State and Surface Photochemistry 2.1 Metals, Semiconductors and Insulators Heterogeneous photochemical processes are concerned with the effect of light on interacting molecules and solid surfaces. The concept of photoinduced surface chemistry is commonly used to integrate these processes. As cited earlier, they involve surface phenomena such as adsorption, diffusion, chemical reaction and desorption [3]. Experiments and theoretical calculations make clear that the photochemical behavior of an adsorbed molecule can be very different from that of a molecule in the gas or liquid phase [4]. Photochemical reactions of this type involve molecules and systems of quite different complexity, from species composed of a few atoms in the stratosphere to large chiral organic molecules that presumably were formed in prebiotic systems.

52

M.C. González · E. San Román

Fig. 1 Light absorption in heterogeneous systems. The substrate (S), strongly interacting adsorbate molecules or adducts (A–S) or multilayered adsorbate molecules (A) can be excited

As illustrated in Fig. 1, these processes may be initiated by the absorption of light by the adsorbate (A), the substrate (S) or adducts (A–S) originating from their interaction. Excitation on the adsorbate or the adduct may relax following a reaction pathway in which the bulk solid may either participate actively or remain inactive. Otherwise, the substrate itself could become excited and trigger a reaction within the adsorbate [5]. In order to understand the diverse photochemistry undergone by an adsorbed molecule, it is convenient to analyze the effect of light irradiation on semiconductors, metals and insulators. The electronic structure of semiconductors is characterized by a gap between electronic states populated by valence band (VB) electrons and empty states in the conduction band (CB), as shown in Fig. 2. The former can be promoted to the CB upon excitation with photons carrying energy in excess of Eg , the band-gap energy. This energy is calculated as the difference between the energies at the bottom of the CB and the top of the VB. Such a process yields CB electrons (e–CB ) and VB holes (h+VB ), which initiate redox reactions at the particulate/solution interface. For these reactions to occur the highest

Fig. 2 Interaction of light with a semiconductor. Light promotes a valence band (VB) electron to the conduction band (CB). Electrons are able to reduce suitable oxidants, whereas holes produced in the VB capture electrons from a reductant

Environmental Photochemistry in Heterogeneous Media

53

occupied molecular orbital (HOMO) energy of the molecule to be oxidized has to be higher than that of the VB and the lowest unoccupied molecular orbital (LUMO) energy of the molecule to be reduced has to be lower than that of the CB. A detailed description can be found elsewhere [6]. Populated and empty electronic levels are overlapped in a metal. Illumination results in the promotion of an electron to an unﬁlled level, followed immediately by vibrational relaxation to the ground state with heat transfer to the network. The rate of recombination of an electron–hole pair in a metal is so high that the chance for the photoelectron or the hole to react with a sorbed species is negligible. Insulators are characterized by an energy gap larger than 4 eV, therefore requiring excitation with vacuum UV light, i.e., of wavelengths shorter than 200 nm. Therefore, electron injection to or withdrawal from adsorbates promoted by light absorption by the substrate is not possible at wavelengths typical of the solar spectrum. 2.2 Interaction Between Surface and Adsorbate Heterogeneous photochemical processes depend strongly on the adsorbate and substrate characteristics, such as the electronic properties of the substrate, the nature of the adsorbate, the interaction between the substrate and the adsorbate and that between adsorbate molecules, the presence of surface impurities and the nature of the medium. Modern surface science provides powerful probes for the characterization of the structure and composition of surfaces and the dynamics of reactions occurring on them [3, 7, 8]. Several cases may be distinguished depending on the adsorbate and the substrate properties. When the adsorbate is sensitive to UV–vis light, photoinduced processes may be initiated by direct excitation of the molecule. This is the prevailing mechanism for insulating substrates and may be an important pathway for the decomposition of many organic pollutants [9]. Irradiation may selectively start reactions between adsorbed molecules owing to their proximity and the loss of the degrees of freedom typical of the ﬂuid phase or favored by perturbations caused by neighboring molecules on the potential energy surface on which the photoprocess takes place [10]. Solid surfaces may accommodate and orient molecules at distances close to molecular bonds and reaction rates are inﬂuenced by physical order in the adsorbed layer. Deposition of small molecules on a crystal surface under appropriate temperature and pressure conditions produces ordered molecular monolayers and multilayers. These structures result from the balance of the forces causing adsorption imposed by the surface and the forces between neighboring adsorbed species. Under such conditions, certain reactant

54

M.C. González · E. San Román

geometries will be favored depending upon the relative orientation and alignment of the molecule with respect to its neighbors and to the surface. Surface materials may also control the properties of adsorbate ground and excited states. The excited-state lifetime can be shortened owing to the occurrence of new relaxation channels or increased owing to weakening of vibronic coupling. The ﬁrst effect is important for metals and semiconductors, where a high density of accessible electronic states is available, but does not operate in general for insulators. Vibronic coupling is reduced because of the enhanced rigidity of the adsorbate, which may exclude vibrational modes present in the free molecule. Interaction among adsorbed molecules may cause the formation of molecular aggregates when high local concentrations are involved. Adsorbate aggregation is particularly enhanced when a hydrophobic compound adsorbed on a nonpolar substrate is in contact with water [11]. Upon irradiation of aggregates, new nonradiative decay channels are opened, leading to a lowering of reaction quantum yields. This is evidenced by a reduction of the emission rate when the adsorbate is luminescent [12] or by smaller oxidation rates when the adsorbate is a singlet molecular oxygen photosensitizer [13, 14] or a sensitizer in dye-assisted photocatalysis (vide infra) [15, 16]. Adsorption and aggregation of charged molecules depend on surface charge, which may be sensitive to pH. For example, negatively charged rose bengal molecules remain monomeric on the surface of silica–alumina nanoparticles bearing permanent negative charges, while dimers are formed on positively charged nanoparticles [17]. Moreover, neither monomers nor higher aggregates are found in the latter case at dye loadings ranging from high dilution to surface saturation. Monomeric and dimeric adsorbates coexist on a polar, noncharged surface [18]. In addition to aggregation, excitation energy migration among monomers and trapping by surface defects may be also a cause of radiationless deactivation [19]. For insulating substrates sunlight absorption can only take place within the adsorbate. If the potential energy surface is not perturbed upon adsorption, the electronic spectrum of the adsorbate and that of the molecule in the gas or liquid phase are usually very similar. Under such conditions, data from photophysical studies involving insulator surfaces allow a straightforward interpretation of the adsorbate photodynamics when compared with ﬂuid-phase data [4]. The photoexcitation spectrum of an adsorbed species may be different from that observed in the ﬂuid phase when coupling of the adsorbate energy levels to those of other adsorbate molecules or to surface electronic states is not negligible. Chemisorption requires a proper symmetry and closeness in energy of the adsorbate and surface orbitals [10] and a shift of absorption maxima to higher wavelengths is observed owing to the weakening of the adsorbate bonds. Therefore, the adsorbed species may be excited by light of lower energy than free molecules. Since the intensity of the solar spectrum increases rapidly above 380 nm, this effect may lead to

Environmental Photochemistry in Heterogeneous Media

55

a much higher reaction rate [9]. Adsorbate photochemistry may lead to the formation of grafted radicals. Photoinduced formation of these radicals can be studied at silica–gas and silica–liquid interfaces to test the matrix inﬂuence on reaction and deactivation [20]. In the case of a metal substrate, the experimental evidence shows that metal excitation is dominated by surface photon absorption. Optical radiation excites surface charge carriers, usually free or sub-vacuum-level electrons that can efﬁciently couple to the adsorbate. This often leads to enhanced photolysis cross sections or altered product distributions. Excitation localized on the adsorbed molecule in close proximity to a metallic solid may efﬁciently couple to the electronic states of the surface, leading to excitation quenching. When light-absorbing molecules are separated from the surface by spacer molecules, the inﬂuence of the surface on molecular excitation and relaxation decreases [4, 21]. Excitation of solid substrates, mostly metals and more rarely semiconductors perturbed by adsorbates, leads to new excitation and/or relaxation channels. Irradiation of such systems may result in adsorbate desorption, dissociation or chemical reaction by both thermal and nonthermal processes. Thermal processes occur, usually at high ﬂuence rates, by direct surface phonon excitation or by excitation of adsorbate internal modes, which relax via phonon excitation. Photoinduced surface heating can lead to molecular desorption or chemical reaction. On the other hand, nonthermal or photoprocesses depend, in principle, on the photon energy but also on the dynamics of the electronic excitation/relaxation processes and on nuclear motions. These processes are typical at low ﬂuence rates. Strong electronic coupling between substrate and adsorbate, characterized by a rapid energy exchange, leads to reaction pathways which are independent of where the initial excitation event takes place. If the adduct is excited to a repulsive state, photon energy may be dissipated by adsorbate–substrate bond breaking and subsequent nonthermal photodesorption of the intact molecule. Usually, the photoejected adsorbate is in an excited state. These hot intermediates, absent in thermal reactions, are capable of inducing secondary surface processes such as reaction with coadsorbates [4, 21]. Many metal oxides and sulﬁdes are semiconductors, and may act as photocatalysts (Fig. 2). As already discussed, the initial reaction step in photocatalysis is the creation of an electron–hole pair in the semiconductor or the injection of a charge from an excited adsorbate. Reaction pathways differ from those where energy transfer takes place between substrate and adsorbate. The excited charge carriers can quickly relax to the band edges and become trapped as captured electrons (e–cap ) and hydroxyl radicals (OH• ) as found for TiO2 and shown in Fig. 3. The absence of electronic states within the band gap and the relatively low density of states at the band edges reduce charge transfer back to the ground state; however, carrier recombination is

56

M.C. González · E. San Román

Fig. 3 Energy levels in TiO2 photocatalysis (given in terms of reduction potentials at pH 0): (A) shows the reduction of acceptor A, (B) the direct oxidation of donor D, and (C) the oxidation of D mediated by a sensitizer S

usually the main factor leading to radiationless deactivation. The long-lived charge carriers lead to photoinduced redox reactions. The capability of photoelectrons (holes) to reduce (oxidize) an adsorbate depends on the difference of the CB (VB) energy and the energy of the adsorbate frontier orbitals. Carrier trapping slightly lowers the energy available for reduction or oxidation. The energies of the frontier orbitals are generally unknown but, as a ﬁrst-order approximation, they can be estimated from the standard reduction potential of the aqueous redox couple (Fig. 3). These data are widespread in the chemical and geochemical literature. VB and CB energies are related to the reduction potentials of holes and electrons, respectively. For typical semiconductors, holes have reduction potentials positive enough to oxidize a broad spectrum of adsorbates, either directly or indirectly through the formation of reactive radicals. Electrons are able to reduce suitable compounds such as gaseous or dissolved molecular oxygen. The white arrow on the reduction side of the ﬁgure (path A in Fig. 3) shows the range of reduction potentials required for a compound to be reduced or oxidized in a TiO2 photocatalytic process at pH 0. However, though the photocatalytic process may be thermodynamically feasible, semiconductor impurities, interfacial redox processes, sorption of reactants and desorption of products may control the reaction dynamics. Processes may be diffusion-controlled and reaction rates may depend on the adsorption– desorption equilibrium. Most of these factors are also relevant to other types of heterogeneous reactions. An interesting additional effect of UV irradiation on some semiconductor surfaces is the increase in the extent and rate of adsorption of gases. As an example, in the absence of UV irradiation, SO2 adsorbs on the α-Fe2 O3 surface with a very low sticking coefﬁcient. However, formation of electron–hole pairs by UV irradiation results in the creation of Fe2+ defect sites, enhancing

Environmental Photochemistry in Heterogeneous Media

57

the adsorption and reactivity of SO2 and gases such as O2 , H2 and H2 O over the oxide surface [22]. Direct oxidation by free or trapped holes (path B in Fig. 3) may occur only if the photocatalyst is illuminated with light of energy hv1 , higher than the band-gap energy. For wide-band-gap semiconductors found in nature, e.g., TiO2 and ZnS, the intensity of solar light with energies higher than Eg is low compared with the remainder of the solar spectrum. However, excitation of an adsorbate S with visible light of energy hv2 , lower than the band-gap energy, may indirectly initiate the photocatalytic process. The adsorbate acts in this case as a photosensitizer, injecting an electron into the CB (path C in Fig. 3). It is thought that the accelerated dissolution of iron oxides in the presence of carboxylic acids may be due to this mechanism [23]. It should be recalled that dye-sensitized semiconductors have less oxidizing capacity than unmodiﬁed ones [15], as the oxidant is the dye radical cation, S+ , instead of the hole or the OH• radical. In fact, injection of electrons from the adsorbate singlet state may take place only if the excited electron has an energy higher than the CB. For light of energy smaller than the band gap to be active, this implies in turn that the HOMO of the sensitizer should be above the VB. In both direct and photosensitized photocatalysis, reduction is accomplished by free or trapped CB electrons (path A in Fig. 3). Photocatalysis is mostly thought of in terms of the photodegradation of molecules initiated either by oxidative or by reductive processes. However, photosynthesis may also result from a photocatalytic process, as observed for the reduction of CO2 to formic acid by natural minerals. This reaction is the ﬁrst step in an abiotic pathway for the synthesis of organic molecules and has been proposed to have played a role in the origin of life on earth [9].

3 Heterogeneous Reactions of Environmental Significance 3.1 The Atmosphere 3.1.1 Interaction Between Light and Aerosols The volume fraction of the atmosphere occupied by aerosols in the form of solid or semisolid particles, cloud and fog droplets is always very small. However, aerosols are the sites of a rich chemistry at low and medium altitudes involving ionic, redox and surface reactions that would otherwise not take place [1]. They inﬂuence gas-phase chemistry acting as sinks of reactive species, transforming and sending back trace substances to the gas phase

58

M.C. González · E. San Román

and lowering (by absorption and scattering) or enhancing (by scattering) the actinic ﬂux [2]. The concentrations and lifetimes of several trace gases depend strongly on these processes, which in turn modify the oxidative capacity of the atmosphere. Aerosols also inﬂuence climate via radiative forcing and modiﬁcation of cloud properties [24], as well as the health and well-being of people, principally via inhalation of small particles [25]. These concerns have placed aerosol chemistry at the center of atmospheric research. A wide range of atmospheric particles play a role in the chemistry of the troposphere, ranging from sea salt, soot and mineral dust to condensed organic matter. Mineral dust is composed of oxides of the most prevailing elements in the earth’s crust, approximately 60% SiO2 and 10–15% Al2 O3 and varying percentages of Fe2 O3 , MgO and CaO, depending on the source location [26]. Therefore, mineral dust is thought to play a role in the chemistry of the atmosphere closest to the earth’s surface. Aerosols may serve as efﬁcient photosensitizers and photocatalysts. Conversely, light affects particle growth, nucleation, water uptake, water evaporation and ageing of aerosols. Under ﬁxed meteorological conditions and concentration of trace gases, the formation of ozone and other species in the troposphere is strongly dependent on the loading and optical properties of aerosols [27]. Strongly UV absorbing soot and mineral dust decrease photolysis rates in the gas phase and therefore inhibit ozone production. Purely scattering sulfate aerosols in the boundary layer enhance ozone production in the upper atmosphere. At regular loadings, they increase photolysis rates near the ground but at high loadings their effect is reversed [28]. Measurements of UV-B, O3 , CO and NO2 during an episode of Mexican ﬁre smoke support these conclusions [29]. Heterogeneous chemistry has been reported to be responsible for the majority of ozone destruction observed in polar regions in the winter and in the lowermost stratosphere throughout the year. It has also been invoked to explain anomalous chemistry activation in the arctic boundary layer [1]. Thus, it is important to differentiate between the perturbation effects caused on the atmosphere by light-induced heterogeneous processes on aerosols from those caused by photochemical gas-phase reactions induced or inhibited by aerosol light scattering. In the following subsections several examples of the former type will be discussed. The enhancement of homogeneous photolysis rates and the problem of ozone depletion will not be discussed, as they are outside the scope of the present chapter. 3.1.2 Reactions in Clouds Ageing of aerosols in clouds is accompanied by changes in size distribution, surface structure and water content [30, 31]. Hydrophilic compounds present in the solid matrix begin to dissolve and rain cycles these species again to

Environmental Photochemistry in Heterogeneous Media

59

the earth’s surface [32]. During this process elements change their speciation, increasing their bioavailability. Accordingly, the mobility of nutrients as P and Fe but also that of toxic metals increases. At the same time, dissolved transition metals control the redox chemistry in clouds [33]. Atmospheric chemical simulation models rely on the knowledge of these processes, most of which has been gained through laboratory experiments. As an example, transfer of trace metals between the solid and the liquid phase was demonstrated, as well as the effect of pH on their solubility [34]. There is evidence that metal dissolution is affected by light. However, laboratory results carried out in conditions analogous to those found in clouds are scarce. Direct evidence of metal dissolution is given by the analysis of rain composition [35]. Fe and Cu present in clouds originate in crustal and combustion aerosols. Their concentrations in rural air are in the ranges 0.1–100 and 0.001–0.1 µg m–3 , respectively, while typical concentrations of 0.1–20 µM for Fe and 0.001–0.3 µM for Cu are found in cloud water [1]. Solubility depends strongly on the metal oxidation state. For example, the reduced forms Fe(II) and Mn(II) are more soluble than the oxidized forms Fe(III) and Mn(IV) [36, 37]. Photochemical reactions modify metal oxidation states, as happens with iron species: Fe(III) can be extracted from an aluminosilicate or oxo-hydroxide matrix and photoreduced to the more soluble Fe(II) species [34]. Feedback effects are also possible: Fe(II) ions participate in the oxidation of SO2 to H2 SO4 , decreasing pH and inducing further iron dissolution. Many in-cloud photochemical processes are yet poorly understood, as they involve different radical species and oxidizing or reducing agents. They depend on light exposure time, the microphysics of the cloud and on the availability of transition metals. Speciation of dissolved ions also requires special consideration, as it inﬂuences cycling and bioavailability. As an example, it is uncertain in what form Fe and Cu are dissolved in clouds and wetted aerosols. Most of Fe(II) is thought to be present as aqueous Fe2+ , while Fe(III) is complexed with OH– , SO4 2– and organic ions such as oxalate [38, 39]. Cu is largely bound to organic ligands. Organic components and aerosol photochemistry play a major role in speciation. Dissolved metals affect the concentration of atmospheric trace gases, such as ozone, organics and sulfur compounds. Ozone is formed in the troposphere through a complex series of homogeneous reactions as shown schematically in Fig. 4. The chain represented in the ﬁgure by thick arrows involves the cooperative oxidation of organic molecules and NO with the intermediacy of HO• , formed by subsequent photolysis of NO2 and O3 , and HO2 • radicals. Fe and Cu catalyze, through the cycling of Fe(II)/Fe(III) and Cu(I)/Cu(II) species, respectively, the destruction of superoxide radical (HO2 • ) and its conjugate base, the superoxide anion (O2 •– ), yielding hydrogen perox-

60

M.C. González · E. San Román

Fig. 4 Gas-phase OH• -mediated oxidation of organics. HO2 • and peroxyl radicals (not shown) are oxidation intermediates of these reactions. O3 is formed from the photolysis of NO2 and hydroxyl radicals are photochemically generated from O3 . Oxidation of NO occurs simultaneously to the oxidation of the organic substrates and provides a source for NO2 , closing the OH• formation cycle

ide [38, 39]: Fe(III) + O2 •– (HO2 • ) → Fe(II) + O2 (+ H+ ) , Fe(II) + O2 •– + 2H2 O → Fe(III) + H2 O2 + 2OH– , Fe(II) + HO2 • + H2 O → Fe(III) + H2 O2 + OH– .

(1) (2a) (2b)

As a consequence of the scavenging of the long-lived HO2 • radicals by cloud droplets, oxidation of organics and the production of O3 are slowed down in the gas phase. The hydrogen peroxide formed by reaction 2a or reaction 2b reacts with Fe(II) ions yielding hydroxyl radicals Fe(II) + H2 O2 → Fe(III) + OH– + OH•

(3)

which contribute to the oxidation of water-soluble organic compounds, mainly acids, aldehydes, alcohols and other oxygen-containing molecules. Moreover, scavenging of HO2 • also suppresses the HOx -catalyzed loss of O3 in the aqueous phase, O3 + HO2 • → OH• + 2O2 , O3 + OH• → HO2 • + O2 ,

(4) (5)

thus enhancing the ozone-mediated oxidation of organic compounds in the aqueous phase. As a consequence of all these processes, oxidation of organics is slowed down in the gas phase and enhanced in the aqueous phase. Experiments on model solutions demonstrate that chromophoric compounds, such as Fe(III) oxalato complexes, may act as primary photore-

Environmental Photochemistry in Heterogeneous Media

61

actants. Upon photolysis, these compounds produce superoxide anion and superoxide radicals: Fe(C2 O4 )+ + hv → Fe(II) + CO2 + CO2 •– , CO2 •– + O2 (+ H+ ) → CO2 + O2 •– (HO2 • ) .

(6) (7)

Superoxide radicals react with aqueous ozone according to reaction 4 yielding hydroxyl radicals. As stated before, these radicals participate in the oxidation of organic matter in the droplet. For example, they react with formate ions: HCO2 – + OH• → H2 O + CO2 •– .

(8)

Photolysis of Fe(III) oxalato complexes may therefore lead to the catalytic destruction of ozone. Fe(II)/Fe(III) species also catalyze the oxidation of SO2 to H2 SO4 [33, 38, 40, 41]. Sulfur dioxide is largely in the form of HSO3 – in water at pH values prevailing in clouds. This ion is oxidized by different atmospheric oxidants, particularly dissolved Fe(III) ions. Heterogeneous photocatalysis on iron oxides also plays a role in HSO3 – oxidation to SO4 2– . The band gap of iron oxides ranges between 2.02 and 2.12 eV, which implies that they can be photoactivated by visible light. Efﬁcient F2 O3 charge carriers may also result if light is absorbed by surface ligands capable of undergoing a charge transfer interaction with the iron oxide crystal, as observed in laboratory experiments performed on γ -Fe2 O3 nanocrystals [42].

Fig. 5 Processes involving iron compounds in droplets and their interaction with gasphase chemistry (see text). Both heterogeneous and homogeneous reactions are intimately involved in a complex manifold of reactions

62

M.C. González · E. San Román

Figure 5 exempliﬁes several processes involving iron compounds that may take place in cloud droplets and their interaction with gas-phase chemistry. Long-lived soluble trace gases such as SO2 , H2 O2 and oxygenated organic compounds undergo dynamic exchange with the droplet ﬂuid phase. Oxidative processes take place in the droplet by the different mechanisms already described. 3.1.3 Reactions on Snow, Ice and Saline Particles Snow, ice and ice crystals forming natural clouds (cirrus and convective clouds) are important components of cold ecosystems. They are sinks for atmospheric pollutants and promoters of surface and near-surface reactions that may inﬂuence the fate and behavior of many trace gases [43]. Ice is not completely transparent from 200 to 700 nm owing to the presence of trapped compounds. Photoreactions in ice have received a great deal of attention during the past few years in regard to astrophysical studies and the presence of life in the universe. Concerning the earth’s atmosphere, it has been demonstrated that sunlight illumination of snow releases a host of trace gases to the boundary layer. This effect has been observed in the Arctic, the Antarctic and at mid-latitudes [44, 45]. The gases include NOx (NO and NO2 ), carboxylic acids, aldehydes and other organic compounds that inﬂuence the chemistry and composition of both the atmosphere and the snowpack. Nitrogen oxides are precursors to the production of ozone in the troposphere (Fig. 4). It was recently shown that signiﬁcant amounts of NOx are released by photolysis of nitrate scavenged by cirrus clouds [43]. Hydroxyl radicals are produced during this process. Light affects, therefore, the HNO3 budget in surface snow, the overlaying atmosphere and glacial ice [46]. This fact may be the reason for uncertainties of constrained box models which consider mainly gas-phase reactions, currently used for the simulation of NOx -to-HNO3 ratios in the atmosphere [47]. Formaldehyde, a product of the atmospheric oxidation of methane, has a short lifetime under daylight owing to its photolysis yielding OH• radicals and carbon monoxide. Experiments carried out in the Arctic demonstrated that formaldehyde is also photochemically produced at the air–snow interface and is stripped out of the snowpack, thereby increasing the oxidative capacity of the polar troposphere [48, 49]. Heterogeneous photogeneration of this trace gas accounts for much larger measured concentrations than those predicted from current gas-phase model simulations. Other important heterogeneous photochemical processes taking place in snow or ice packs are the photolysis of chlorine dioxide (OClO) yielding either chlorine peroxy radical (ClOO) or atomic chlorine, and of HBr and brominated organic compounds scavenged by ice crystals which result in Br2 release.

Environmental Photochemistry in Heterogeneous Media

63

Photochemical processes are also relevant in the marine environment and the overlying troposphere. The concentration of halides in sea-salt particles is signiﬁcantly higher (approximately 5 M) than in cloud droplets, and hence different reaction paths should control the overall chemical processes. It has been demonstrated in model studies that marine sea-salt aerosol particles act as a source for halogen species, such as Cl2 , Br2 and BrCl formed by reactions of halides with hypohalogeneous acids, HOCl and HOBr, in pH-dependent reactions: HOX + Y – + H+ → YX + H2 O ,

(9)

where X and Y are Cl or Br atoms [50]. The presence of water in NaCl crystallites enhances the ionic mobility within the crystals and causes enrichment of halogens at the surface. Owing to their limited solubility, molecular halogens evaporate into the gas phase. Such release processes lead to a decrease of the halide concentrations within the particle after enough time. The gasphase halogenated molecules released are photolyzed to form halogen atoms (Cl, Br; about 70% of all Cl sources and more than 99% for Br) and ClO and BrO radicals [1, 51, 52]. The radicals formed react with ozone, dimethyl sulﬁde and gaseous hydrocarbons. Therefore, small particles have an important effect in regulating the concentration of the latter gases in marine areas [53]. Chloride ions from supermicron sea-salt particles are replaced by sulfate, nitrate and minor contributions of oxalate, malonate and succinate. The principal mechanisms causing accumulation of sulfate in sea-salt particles are cloud processing and, to a lesser degree, heterogeneous reactions taking place in deliquescent particles. Mechanisms for the chloride replacement by nitrate are not clear [54]. 3.1.4 Reactions over Carbonaceous Particles Black carbon is a general denomination for materials such as soot, graphitic carbon and carbon black. The incomplete combustion of fossil fuels and biomass releases enormous quantities of these materials into the atmosphere, about 13 Tg year–1 on a global basis [55]. Both natural and anthropogenic processes contribute to this input. The chemical composition of the carbonaceous particles in the atmosphere is highly variable, as it depends on their formation process and ageing. The common feature of all black carbon constituents is the strong absorption within the whole solar spectrum. Black carbon particles are particularly toxic as pollutants; on the other hand, they share the beneﬁcial aspect of shielding efﬁciently harmful UV radiation. Although commercial products such as industrially produced carbon black, impure graphite and activated carbon differ chemically and morphologically from emitted organic aerosols, they are usually used as model particles for screening studies [56]. So far,

64

M.C. González · E. San Román

there have been few studies on the photochemistry of adsorbates on these particles owing to their intractability and their broad and intense light absorption. Soot aerosols have a very large surface area owing to their agglomerate structure; therefore, they react efﬁciently with trace gases in the atmosphere. Graphite has a semimetallic electronic structure, which may perturb the photochemistry of an adsorbed molecule, particularly when there is strong orbital coupling with the adsorbate [56]. Chemical interactions and other properties of black carbon depend on the nature of surface groups, commonly quinones, phenols, carboxyphenols, lactones and nitrogen- and sulfurcontaining molecules (Fig. 6). Oxidized hydrophilic groups may become hydrated and particles act as cloud condensation nuclei. Carbon particles may also function as chain-breaking free-radical donors and acceptors, as decomposers of hydroperoxides and as excited-state quenchers [57]. The photolysis of dibromodiﬂuoromethane adsorbed at monolayer saturation on highly ordered pyrolytic graphite was reported [56]. Studies were carried out at 85 K, as CF2 Br2 desorption occurs at approximately 138–141 K. Naturally, this temperature range is not directly useful for the understanding of atmospheric processes but results may be extrapolated to other possible reactions on black carbon. After UV irradiation, Br• and CF2 Br• radicals are formed and various recombination reactions are observed. The distribution of reaction products is signiﬁcantly different from that observed in the gas phase. Processes like photolysis of adsorbed CF2 Br• radicals and spontaneous scission of hot radicals, together with constraints imposed by the surface, may account for the difference. Electron-induced dissociative-type reactions due to photon absorption by graphite, as observed on irradiation of metallic and semiconducting surfaces, is not observed in spite of the high electron attachment cross section reported for gas-phase CF2 Br2 [58]. As stated earlier (Sect. 3.1.2), aqueous-phase hydroxyl radicals participate in the catalytic depletion of tropospheric ozone. Photochemical studies performed on aqueous suspensions have shown that black carbon particles accelerate this process [59]. Though no detailed mechanistic interpretation was given, it was concluded that hydroxyl radicals formed on the particle surface initiate a chain reaction in the liquid phase leading to the consumption of dissolved ozone. In other studies the interaction of gas-phase ozone

Fig. 6 Surface groups on black carbon

Environmental Photochemistry in Heterogeneous Media

65

with soot has been investigated at room temperature and no thermal catalytic effect was reported [55]. It should be recalled that only catalytic surface reactions are expected to cause signiﬁcant heterogeneous ozone depletion, as ozone mass concentrations are, in general, much higher than those of soot aerosol, unless ozone has been depleted by NO in highly polluted air. As a result, it has to be concluded that ozone cannot be signiﬁcantly removed from the atmosphere by black carbon unless photochemical processes are involved. 3.1.5 Reactions of Polycyclic Aromatic Hydrocarbons Polycyclic aromatic hydrocarbons (PAHs) constitute a group of potentially harmful compounds, many of them suspected to be carcinogenic. Small PAHs such as anthracene and phenanthrene reside primarily in the gas phase in the lower troposphere at moderate temperatures. However, as the temperature is decreased, they begin to sorb onto atmospheric particles [60, 61]. Heavier homologous are adsorbed on atmospheric aerosols even at moderately high temperatures or are included into ice particles and can be transported over long distances [43]. Expectedly, direct photolysis on the surface of particles, promoted by UV sunlight at high altitudes in the troposphere, may control the fate of these anthropogenic pollutants. They can also be formed photochemically in ice from less harmful chemicals as observed for chlorobenzene [43]. Highly polar surfaces such as those of silica and alumina may be used as models for insulators such as SiO2 , Al2 O3 , silicoaluminates and CaCO3 , which account for the 20–30% of the composition of all inorganic atmospheric particulates. A brief account of the heterogeneous photochemistry of PAHs will be given here. A more detailed discussion on the observed mechanisms and reaction products may be found in the chapter by Pagni and Dabestani. Laboratory experiments seem to indicate that two main mechanisms of PAH photodegradation at the solid/air interface may be operative [60, 62]. PAHs behave as energy transfer photosensitizers yielding singlet molecular oxygen (1 O2 ) and some of them (e.g., anthracene, phenanthrene, tetracene) react with this excited species yielding thermally unstable peroxides and/or dioxetanes, which undergo further decomposition. An electron transfer mechanism seems to be involved for those compounds that are insensitive to singlet molecular oxygen (e.g., naphthalene, 1-methylnaphtalene), as the buildup of the corresponding radical cation has been observed. The identity of the electron acceptor is unknown. Molecular oxygen may be a suitable acceptor, as superoxide was formed in photolysis experiments with naphthalene at a silica gel/air interface. Both mechanisms may act simultaneously, as happens with 1-methoxynaphthalene. On the other hand, 1-cyanonaphtalene, a poor singlet molecular oxygen acceptor with a relatively high oxidation potential, does not undergo photolysis at any appreciable rate [60].

66

M.C. González · E. San Román

Fluorene (FL) shows different photolytic half-lives on several substrates under identical conditions: 110 h (silica gel), 62 h (alumina), 37 h (ﬂy ash) and more than 1000 h (carbon black). Its ﬂuorescence is strongly quenched by oxygen at the silica gel/air interface. The effect of increasing surface coverage on ﬂuorescence is consistent with the formation of PAH ground-state dimers and higher-order aggregates [60, 62]. The FL 0 – 0 transition assigned to the monomer on silica is slightly redshifted relative to the value reported in heptane, as expected for a π –π ∗ transition in the more polar environment of the silica surface. It should be taken into account that ground-state pairing may be the consequence of the experimental procedure used for adsorption of PAHs from solution, which may produce a solvent-pooling effect. When dry samples containing aggregates are exposed to water vapor, molecular motion results in the disappearance of the emission of aggregates. Studies carried out on pyrene adsorbed on silica gel, alumina and calcium ﬂuoride allowed the characterization of the potential energy surfaces associated with dimerization. A substantial barrier is found for aggregation, leading to the stabilization of nonequilibrium ground-state pairs. This barrier may arise from the presence of surface silanols, which stabilize pyrene dimers by electronic polarization effects. Photolysis of FL at a silica gel/air interface leads to the generation of 9-ﬂuorenone (FLO) as the only isolable product. No dark reactions were observed and singlet molecular oxygen is not involved in the reaction. Transient spectroscopy shows that both the triplet state and the radical cation of FL are formed, thus indicating that an electron transfer mechanism is involved. Loss of a proton from the radical cation and subsequent reaction with molecular oxygen yields FL peroxide radicals, leading to the formation of 9-hydroxyﬂuorene. The latter readily photolyzes on the silica surface to produce FLO. Adsorbed FLO does not undergo photochemical degradation in the absence of FL, while direct photolysis of FL and FLO coadsorbed on the silica surface leads to the consumption of both species. FLO acts as both an electron acceptor and a proton scavenger in the one-electron oxidation of FL. On the basis of the thermodynamic and kinetic arguments, it has been proposed that oxygen serves as a shuttle for the transport of electrons from FL to FLO. This is in agreement with the fact that coadsorbed FLO does not accelerate FL photolysis [60]. Surface groups may also act as electron acceptors, as observed in transient spectroscopic studies on negatively charged rose bengal adsorbed on the surface of positively charged silica–alumina nanoparticles [17]. Irradiation of the latter system leads to the formation of the radical cation of rose bengal, both in the absence and in the presence of molecular oxygen, thus suggesting that this is not the primary electron acceptor. Instead, Lewis acid sites on the surface (Fig. 7) were proposed as the electron acceptors stabilizing the radical cation over several tens of milliseconds. Subsequent charge trans-

Environmental Photochemistry in Heterogeneous Media

67

Fig. 7 Lewis sites on an alumina surface

fer reactions in the presence of oxygen may be responsible for the oxidation of the organic compound on the surface of the isolating oxide. 3.2 The Hydrosphere There is a general consensus (vide supra) on the environmental importance of catalytic reactions on the surface of many minerals. However, there is limited information in the literature about speciﬁc examples [9]. Systematic studies would allow the understanding of the dependence of the catalytic activity on mineral structure, mineral chemistry and surface reactivity. At the same time, this knowledge would be useful in designing remediation techniques based on minerals instead of synthetic catalysts. For example, sphalerite and ilmenite have been shown to be capable of degrading chlorinated carbon compounds via a photocatalytic mechanism [63]. Photocatalysis may be relevant for the regulation of the concentration of redox-sensitive elements in the photic zones of lakes, streams, hot springs and oceans. Photocatalytic reduction reactions of noble metals (Au, Pt, Rh and also Hg and Pb) may be of importance in hot streams, where semiconductors such as pyrite and stibnite are formed in situ. For example, amorphous As2 S3 and Sb2 S3 precipitated in the Rotokawa spring system (New Zealand) is enriched in metallic gold owing either to adsorption or to photoreduction. On the other hand, photocatalysis is thought to be of importance in regulating Mn and Fe concentrations in surface waters, where dissolved organic matter is excited and injects photoelectrons into solid Fe (hydr)oxides [9]. The photooxidation of several PAHs on aqueous clay suspensions (Ca montmorillonite, laponite, Na montmorillonite, kaolin and Kga-1b, a natural kaolinite from the Clay Mineral Society, USA) was reported [64]. Using chrysene as a model compound, it was found that oxidation rates increased with respect to aqueous solution when Ca montmorillonite, laponite or Na

68

M.C. González · E. San Román

montmorillonite was added. Photooxidation proceeded rapidly using these minerals in fresh water but was entirely quenched in seawater. Reaction rates decreased in the order given and even more when Kaolin or Kga-1b was used. A further example is the reductive reaction of organic compounds at the interface of iron oxides and sulﬁdes, as demonstrated for the degradation of carbon tetrachloride at the surfaces of pyrite and mica [63]. Degradation of phenol, benzoic acid and chlorinated organics was reported using iron oxide, iron hydroxide and iron-containing sand [65]. The Fenton reaction, based on the formation of reactive species from iron complexes and hydrogen peroxide (Fig. 5), can also be conducted in heterogeneous mediua. The photodegradation of the azo dye Mordant Yellow 10 can be accelerated in the presence of goethite (α-FeOOH) and hydrogen peroxide. The reaction mechanism of the heterogeneous photo-Fenton reaction is different from the semiconductor photocatalytic mechanism leading to the photoreductive dissolution of goethite to aqueous Fe(II) and a subsequent homogeneous Fenton reaction [66]. The reaction may involve iron cycling on the surface, initiated by the formation of a surface precursor complex of H2 O2 , > FeO – OH, as also found in the TiO2 /H2 O2 system. UV light irradiation cleaves the O – O bond leading to the formation of an oxoiron species, > FeIV = O, and hydroxyl radical [65]. In solution, the bare high-valent oxoiron species is unstable and may react either with H2 O, forming a second hydroxyl radical, or with organic compounds [67]. If a similar process occurred with the Fe(IV) species on the surface of goethite, the whole process would lead to the catalytic oxidative degradation of the organic substrate. Silica is one of the commonest materials on earth and, therefore, it is of interest to understand its role in the environment. The factor determining the surface properties in amorphous silica is the presence of silanol groups and siloxane bridges (Fig. 8) [68]. Chemical reactions involving these functional groups may lead to modiﬁed surfaces with entirely different properties. A total of 4.6–4.9 OH groups per square nanometer are found for fully hydroxylated silica. Single, (≡ Si – O –)3 Si – OH, and geminal, (≡ Si – O –)2 Si(– OH)2 , silanols present in an irregular arrangement act as

Fig. 8 Silanol and siloxane groups at the silica surface

Environmental Photochemistry in Heterogeneous Media

69

centers for the adsorption of molecules capable of forming hydrogen bonds or of undergoing donor–acceptor interactions. Geminal and single silanols show pKa values of 4.5–6.5 and 8.5–9.0, respectively. Dissociation of surface hydroxyl groups yields ≡ SiO– ions, rendering a negatively charged silica. An experimental approach employed to study liquid/solid interfaces involves colloidal particle dispersions which provide a high concentration of surface sites for reaction. Since colloidal silica has a very high speciﬁc surface area and is optically transparent in the near-UV and visible regions, it is a suitable model surface for investigating interfacial reaction kinetics involving photochemically generated species. Experiments generating sulfate radicals, SO4 •– , by UV photolysis of S2 O8 2– in aqueous suspensions of silica nanoparticles showed a fast disappearance of the aqueous sulfate radicals yielding two transient species with absorption maxima around 320 and 600 nm, respectively [20]. The results indicated that at pH 3–9 SO4 •– radicals build up an adduct on the surface with maximum absorption at 320 nm. This adduct shows similar reactivity to that observed for the sulfate radical in aqueous solution. The transients with absorption maximum at 600 nm were identiﬁed as SiO• surface defects formed from the reaction between the adduct and deprotonated geminal and single silanols. Other less oxidative radicals lead to different radical–silica interactions. For example, thiocyanate radicals react with deprotonated silanols, not involving silanol oxidation. Studies performed on surface reactions on insulators such as silica indicate that a rich chemistry exists on these substrates, which may have unexpected consequences in the environment. A great deal of effort is still needed to evaluate the potential effect of such reactions. 3.3 Soil Reactions Soil is the unconsolidated mineral matter on the surface of the earth that serves as a natural medium for the growth of plants. It is a product of the action of environmental factors such as climate, organisms and topography on parent materials; therefore, soils of different geographical regions have different structures [69]. The soil is a heterogeneous system formed by organic and inorganic solids, gases and water. The inorganic components are sparingly soluble materials, primarily composed of quartz, feldspars, iron and aluminum hydrous oxides, kaolinite, smectites and micaceous minerals. These solids may contain highly reactive charged surfaces that play an important role in immobilizing waste constituents. Organic compounds are present in various stages of decomposition. The term humus refers to solid organic products arising from plant and animal residues. They are the most abundant in the upper 30–60 cm of the earth’s crust, where they interact with

70

M.C. González · E. San Román

air and water. Humus reduces plasticity and cohesion of soils owing to the presence of silicate clay minerals. About 65–75% of soil organic matter consists of humic substances (HS), which are operationally divided according to their solubilities in acid and basic media into humic acids, fulvic acids and humin [70]. HSs are formed by amorphous, dark, hydrophilic, acidic, partly aromatic, chemically complex substances ranging in molecular weight from hundreds to several thousand, with large surface areas (500–800 m2 g–1 ) and high pH-dependent cationexchange capacities. They are composed mainly of carbon, hydrogen and oxygen, likely in the form of polyphenols, polyquinones, polysaccharides, carbohydrates, proteins, peptides, amino acids, fats, waxes, alkanes and low molecular weight organic acids [69]. They are readily decomposed by microorganisms, and thus have a short lifetime. Waste constituents may be immobilized in a soil system mainly by sorption and/or partitioning. Adsorption on soil particles is competitive, pHdependent and, usually, inversely proportional to the solubility of the compound in water. Dry soils are better adsorbents than wet ones. HSs are able to form complexes with metal ions and hydrous oxides and also interact with minerals and a variety of organic compounds, including alkanes, fatty acids, dialkyl phthalates, pesticides and herbicides, and may therefore increase the concentration of these constituents in soil and natural waters. Photochemical degradation has been recognized as an abiotic mechanism that contributes to the loss of waste constituents in soils [71, 72]. It does not often result in complete degradation but allows further action of biotic processes. Phototransformation of soil-adsorbed pollutants takes place through two main pathways. Solar light absorption by the adsorbate may induce chemical changes in the molecule depending on its chemical nature. This process may be an important decomposition route for herbicides showing absorption spectra at λ > 295 nm. Heterocyclic N-containing compounds, phenoxy acids, dinitroanilines, propanil (3 ,4 -dichloropropionanilide) and benzoic acids are very sensitive to photodecomposition. A second mechanism involves photoinduced transformations mediated by components of the soil environment, leading to highly reactive species that readily oxidize organic substrates. HSs act as photosensitizers for the generation of reactive species such as solvated electrons, hydroxyl radicals, singlet molecular oxygen, superoxide ion/hyproperoxyde radical, organic peroxyl radicals, excited triplet states of ketonic moieties and long-lived intermediates [73–77]. They enhance the photodegradation of pesticides, herbicides and other organic pollutants in soils and aquatic environments. Knowledge of soil surfaces is still very poor, as there is limited information on basic parameters such as light intensity proﬁles in the soil, extinction coefﬁcients and reaction quantum yields. Because light intensity decreases

Environmental Photochemistry in Heterogeneous Media

71

rapidly, the apparent photodegradation rate depends on the reaction kinetics and the transport of the compound to the light-exposed zone. Laboratory studies on the degradation of metalaxyl, a transparent pesticide, indicate that photosensitization may be operative and the whole process seems to be favored in the presence of water. In general, photopersistence of soil adsorbates is greater in dry soils and at higher clay content [78]. Formation of dioxins in laboratory irradiation experiments on pentachlorophenol adsorbed on soils may be prevented by adding fulvic acids [79]. The photodegradation of propachlor (2-chloro-N-isopropylacetanilide) and propanil showed faster rates in soils than in water samples [80]. The presence of HSs reduces the photodegradation rate of the latter compounds in water but leads to an enhancement in soil. Hydroxy and dechlorinated derivatives were the major photoproducts observed. Triﬂuralin [2,6-dinitro-N,N-dipropyl-4-(triﬂuoromethyl)-benzenamine], a commonly used herbicide, persists under most environmental conditions [81]. When exposed to light, it degrades to benzimidazols [82]. Triﬂuralin photolysis is strongly inhibited when it is sorbed onto clay particles. Studies on the photolysis of soil-adsorbed atrazine show important effects of soil granularity, pH, humidity, organic and humic acid content and the presence of surface-active agents [83]. Atrazine photocatalytic degradation by natural minerals and environmental particulate matter was also investigated [84]. Experiments on TiO2 and ZnO photocatalysts showed a fast degradation. The degradation rates observed in similar experiments, conducted in the presence of titanium-, zinc- or iron-containing minerals were between 1 and 3 orders of magnitude lower than those with the pure photocatalysts but were still faster than in experiments without particles. On the other hand, particles such as soot, ﬂy ash, sand, road dust and volcanic ash showed no signiﬁcant photocatalytic activity. The photolytic degradation of asulam, triclopyr, aciﬂuorfen and atrazine adsorbed in air-dried soils and in soils maintained at a constant moisture content were independently studied [85]. The half-lives of pollutants in dry soils were 2–7 times longer and, in the case of atrazine, the absence of moisture precluded degradation. The effect of soil composition on the photolysis of soil-adsorbed niclosamide using a xenon lamp was reported [86]. The study was conducted with both moist and air-dried soils fortiﬁed with sodium nitrate (commonly used as a fertilizer), iron (a natural soil component) or humic acid. Degradation in soils fortiﬁed with iron was slower than in that in soils fortiﬁed with sodium nitrate and no extractable degradation products were observed in any of the soils fortiﬁed with humic acid. Soils with higher organic or iron contents or that are exposed to fertilizers do not affect as dramatically the half-life of pesticides as does the presence of moisture in the soil. The maintenance of moisture was found to be crucial to the reliability of soil photolysis studies.

72

M.C. González · E. San Román

A further interesting aspect of heterogeneous photochemistry concerns oil spilled at sea near land, which is subjected to environmental effects such as evaporation, dissolution, photooxidation, dispersion into the water column and biodegradation. The fate of heavy fuel oil stranded on rock was studied under different environmental conditions [87]. Samples exposed to full or reﬂected sunlight showed depletion of the larger and more alkylated aromatic hydrocarbons and formation of resins, in agreement with reported laboratory studies on thin ﬁlms of oil.

4 Conclusions Chemical reactions of environmental interest occur under complex conditions, in which gas- and liquid-phase homogeneous reactions are coupled with heterogeneous processes. Ground water and soil pollution and emission of particulates to the atmosphere enhance the relevance of heterogeneous chemical reactions, thus increasing the complexity and diversity of processes whose impact on the environment need to be addressed. Understanding environmental chemistry requires the input of scientists from several ﬁelds, including, among others, geoscience, surface physics and engineering. The development of highly sensitive analytical techniques able to explore environmental interfacial processes is also of fundamental importance. Despite the laboratory effort made during the last few years to understand environmental heterogeneous processes, there is still a dramatic lack of detailed mechanistic studies on surface photochemical reactions relevant to the environment. In the atmosphere, the involvement of a complex mixture of substrates distributed in different phases and interfaces, low reaction quantum yields and the scarce knowledge of the phase behavior and composition of environmental particulates make the interpretation of observed trends in terms of reaction mechanisms difﬁcult. Further ﬁeld studies on topics related to O3 , NOx and halogen budgets will improve the understanding of the effect of heterogeneous chemistry in clouds, sea salt and sulfate aerosols on ozone atmospheric levels. Evaluation of the importance of heterogeneous photochemistry on soil processes is further complicated by the simultaneous occurrence of biotic pathways and the variability of natural soil moisture. Understanding the impact of HSs in soils and sediments is also a difﬁcult task owing to their variable composition. Field studies, together with the compilation of chemical data, are needed in order to formulate adequate models for pollutant cycles. Such models would allow the evaluation of the transport of pollutants through air and water to regions far from the original emission sources and their possible (photo)transformations in different compartments, resulting either in less

Environmental Photochemistry in Heterogeneous Media

73

toxic, eventually further degradable substances, or in more toxic and persistent pollutants that are subsequently released to the environment.

References 1. Jacob DJ (2000) Atmos Environ 34:2131 2. Heintzenberg J, Raes F, Schwartz S et al. (2001) IGAC integration and synthesis report. Chapter 4: Tropospheric aerosols 3. Anpo M, Yamashita H, Zhang SG (1996) Curr Opin Solid State Mater Sci 1:630 4. Garrett SJ (1998) Proc SPIE 3272:274 5. Al-Shamery K, Freund H-J (1996) Curr Opin Solid State Mater Sci 1:622 6. Serpone N, Salinaro A (1999) Pure Appl Chem 71:303 7. Bauer M, Lei C, Read K, Tobey R, Gland J, Murnane MM, Kapteyn HC (2001) Phys Rev Lett 87:25501 8. Bonn M, Kleyn AW, Kroes GJ (2002) Surf Sci 500:475 9. Schoonen MAA, Xu Y, Strongin DR (1998) J Geochem Explor 62:201, and references therein 10. Franchy R (1998) Rep Prog Phys 61:691 11. Iriel A, Lagorio MG, Dicelio LE, San Román E (2002) Phys Chem Chem Phys 4:224 12. Lagorio MG, Dicelio LE, Litter MI, San Román E (1998) J Chem Soc Faraday Trans 94:419 13. Bourdelande JL, Karzazi M, Marqués Tura G, Dicelio LE, Litter M, San Román E, Vinent V (1997) J Photochem Photobiol A Chem 108:273 14. Amore S, Lagorio MG, Dicelio LE, San Román E (2001) Prog React Kinet Mech 26:159 15. Hodak J, Quinteros C, Litter MI, San Román E (1996) J Chem Soc Faraday Trans 92:5081 16. San Román E, Navío JA, Litter MI (1998) J Adv Oxidat Technol 3:261 17. Daraio ME, San Román E (2001) Helv Chim Acta 84:2601 18. Rodríguez HB, Lagorio MG, San Román E (2004) Photochem Photobiol Sci 3:674 19. Lagorio MG, San Román E, Zeug A, Zimmermann J, Röder B (2001) Phys Chem Chem Phys 3:1524 20. Caregnato P, Mora VC, Carrillo Le Roux G, Mártire DO, Gonzalez MC (2003) J Phys Chem B 107:6131 21. Zhu X-Y (1997) Surf Sci 390:224 22. Toledano DS, Henrich VE (2001) J Phys Chem B 105:3872 23. Litter MI, Navío JA (1996) J Photochem Photobiol A Chem 98:171 24. Andreaevc MO, Crutzen PJ (1997) Science 276:1052 25. Kaiser J (2000) Science 289:22 26. Al-Abadleh HA, Grassian VH (2003) Surf Sci Rep 52:63 27. Dickerson RR, Kondragunta S, Stenchikov G, Ciberolo KL, Doddrige BG, Holben BN (1997) Science 278:827 28. He S, Carmichael GR (1999) J Geophys Res 104:26307 29. Raga GB, Castro T, Baumgardner D (2001) Atmos Environ 35:1805 30. Flossmann AI (1998) J Atmos Sci 55:879 31. Francois F, Maenhaut W, Colin J-L, Losno R, Schulz M, Stahlschmidt T, Spokes L, Jickells T (1995) Atmos Environ 29:837 32. Desboeufs KV, Losno R, Colin JL (2001) Atmos Environ 35:3529 33. Losno R (1999) Phys Chem Earth B 24:281

74

M.C. González · E. San Román

34. 35. 36. 37. 38. 39. 40. 41.

Desboeufs KV, Losno R, Vimeux F, Cholbi S (1999) J Geophys Res 104:21287 Lim B, Jickells TD, Colin JL, Losno R (1993) Global Biogeochem Cycles 8:349 Spokes LJ, Jickells TD, Lim B (1994) Geochim Cosmochim Acta 58:3281 Spokes LJ, Jickells TD (1996) Aquat Geochem 1:355 Faust BC (1994) Environ Sci Technol 28:216 Grgi´c I, Pozniˇc M, Bizjak M (1999) J Atmos Chem 33:89 Graedel TE, Mandich ML, Weschler CJ (1986) J Geophys Res 91:5205 Martin JH, Coale KH, Johnson KS, Fitzwater SE, Gordon RM, Tanner SJ, Hunter CN, Elrod VA, Nowicki JL, Coley TL, Barber RT, Lindley S, Watson AJ, Van Scoy K, Law CS, Liddicoat MI, Ling R, Stanton T, Stockel J, Collins C, Anderson A, Bidigare R, Ondrusek M, Latasa M, Millero FJ, Lee K, Yao W, Zhang JZ, Friederich G, Sakamoto C, Chavez F, Buck K, Kolber Z, Greene R, Falkowski P, Chisholm SW, Hoge F, Swift R, Yungel J, Turner S, Nightingale P, Hatton A, Liss P, Tindale NW (1994) Nature 371:123 Turro NJ, Lakshminarasimhan PH, Jockusch S, O’Brien SP, Grancharov SG, Redl FX (2002) Nanolett 2:325 Klán P, Holoubek I (2002) Chemosphere 46:1201 Zang Q, Anastasio C (2003) Environ Sci Technol 37:3522 Bottenheim JW, Dibb JE, Honrath RE, Shepson PB (2002) Atmos Environ 36:2467 Honrath RE, Lu Y, Peterson MC, Dibb JE, Arsenault MA, Cullen NJ, Steffen K (2002) Atmos Environ 36:2629 Stroud C, Madronich S, Elliot A, Ridley B, Flocke F, Weinheimer A, Talbot B, Fried A, Wert B, Shetter R, Lefer B, Coffey M, Heikes B, Blake D (2003) Atmos Environ 37:3351 Sumner AL, Shepson PB (1999) Nature 398:230 Sumner AL, Shepson PB, Couch TL, Thornberry T, Carroll MA, Sillman S, Pippin M, Bertman S, Tan D, Faloona I, Brune W, Young V, Cooper O, Moody J, Stockwell W (2001) J Geophys Res 106:24387 Herrmann H, Majdik Z, Ervens B, Weise D (2003) Chemosphere 52:485 Pszenny AAP, Moldanová J, Keene WC, Sander R, Maben JR, Martinez M, Crutzen PJ, Perner D, Prinn RG (2004) Atmos Chem Phys 4:147 Sander R, Keene WC, Pszenny AAP, Arimoto R, Ayers GP, Baboukas E, Cainey JM, Crutzen PJ, Duce RA, Hönninger G, Huebert BJ, Maenhaut W, Mihalopoulos N, Turekian VC, Van Dingenen R (2003) Atmos Chem Phys 3:1301 Hemminger JC (1999) Int Rev Phys Chem 18:387–417 Kerminen V-M, Teinilä K, Hillamo R, Pakkanen T (1998) J Aerosol Sci 29:929 Kamm S, Möhler O, Naumann K-H, Saathoff H, Schurath U (1999) Atmos Environ 33:4651 Dorko MJ, Bryden TR, Garrett SJ (2000) J Phys Chem B 104:11695 Peña JM, Allen NS, Edge M, Liauw CM, Roberts I, Valange B (2000) Polym Degrad Stab 70:437 Underwood-Lemons T, Gergel TJ, Moore JH (1995) J Chem Phys 102:119–123 Jans U, Hoigné J (2000) Atmos Environ 34:1069 Barbas JT, Sigman ME, Arce R, Dabestani R (1997) J Photochem Photobiol A Chem 109:229 Venkataraman C, Thomas S, Kulkarni P (1999) J Aerosol Sci 30:759 Thomas JK (1993) Chem Rev 93:301 Kriegman-King MM, Reinhard M (1991) In: Baker R (ed) Organic substances and sediments in water, vol 2, chap 16. Processes and analytical. Lewis, Chelsea, MI p 349 Kong L, Ferry JL (2003) Environ Sci Technol 37:4894 He J, Ma W, He J, Zhao J, Yu JC (2002) Appl Catal B Environ 39:211 Siffert C, Sulzberger B (1991) Langmuir 7:1627

42. 43. 44. 45. 46. 47. 48. 49.

50. 51. 52.

53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66.

Environmental Photochemistry in Heterogeneous Media

75

67. Mártire DO, Caregnato P, Furlong J, Allegretti P, Gonzalez MC (2002) Int J Chem Kinet 34:488 68. Zhuralev LT (2000) Colloids Surf A Physicochem Eng Aspects 173:1 69. Burden DS, Sims JL (1999) EPA/540/S-98/500 April 1999. US Environmental Protection Agency, Ofﬁce of Research and Development, Washington, DC 70. Senn TL, Kingman AR (1973) Research series no. 165. The South Carolina Agricultural Experiment Station, Clemson University, Clemson, SC, USA 71. Sims RC, Sorensen DL, Sims JL, McLean JE, Mahmood RH, Dupont RR (1984) Background information for in-situ treatment, vol 2. US Environmental Protection Agency, Cincinnati, OH, USA 72. Dragun J (1998) The soil chemistry of hazardous material. Dragun Corporation, Farmington Hills, MI, USA 73. Zepp RG, Schlotzhauer PF, Sink RM (1985) Environ Sci Technol 19:74 74. Hoigné J, Faust BC, Haag WR, Scully FE Jr, Zepp RG (1989) Adv Chem Ser 219:363 75. Vaughan PP, Blough NV (1998) Environ Sci Technol 32:2947 76. Aguer J-P, Trubetskaya OE, Trubetskoj OA, Richard C (2001) Chemosphere 44:205 77. Thomas-Smith TE, Blough NV (2001) Environ Sci Technol 35:2721 78. Sukul P, Spiteller M (2001) Chemosphere 45:941 79. Liu P-Y, Zheng M-H, Xu X-B (2002) Chemosphere 46:1191 80. Konstantinou IK, Zarkadis AK, Albanis TA (2001) J Environ Qual 30:121 81. Tor JM, Xu C, Stucki JM, Wander MM, Sims GK (2000) Environ Sci Technol 34:3148 82. Margulies L, Stern T, Rubin B, Ruzo LO (1992) J Agric Food Chem 40:152 83. Gong A, Ye C, Wang X, Lei Z, Liu J (2001) Post Manage Sci 57:380 84. Lackhoff M, Niessner R (2002) Environ Sci Technol 36:5342 85. Graebing P, Frank MP, Chib JS (2003) J Agric Food Chem 51:4331 86. Graebing P, Frank M, Chib JS (2002) J Agric Food Chem 50:7332 87. Jézéquel R, Menot L, Merlin F-X, Prince RC (2003) Mar Pollut Bull 46:983

Hdb Env Chem Vol. 2, Part M (2005): 77–118 DOI 10.1007/b138180 © Springer-Verlag Berlin Heidelberg 2005 Published online: 20 September 2005

Homogeneous and Heterogeneous Photochemistry in the Troposphere Michael R. Hoffmann Environmental Science & Engineering, W. M. Keck Laboratories, California Institute of Technology, Pasadena, CA 91125, USA 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 2.1 2.2 2.3 2.4

Basic Principles – Homogeneous Gas-Phase Photochemistry Photochemical Reaction Rates . . . . . . . . . . . . . . . . . Wavelength Range for Active Photochemistry . . . . . . . . . Quantum Yields . . . . . . . . . . . . . . . . . . . . . . . . . Simpliﬁed Photochemical Kinetics . . . . . . . . . . . . . . .

3 3.1 3.2 3.3 3.4 3.5

Homogeneous Gas-Phase Photochemistry Ozone Photolysis . . . . . . . . . . . . . . Nitrogen Dioxide Photolysis . . . . . . . . Formaldehyde Photolysis . . . . . . . . . . Hydrogen Peroxide Photolysis . . . . . . . Nitrous Acid (HONO) Photolysis . . . . . .

4 4.1 4.2 4.3

Heterogeneous Photochemistry in the Troposphere . . . . . . . . . . . Heterogeneous and Aqueous-Phase Photochemistry . . . . . . . . . . . Photochemical Oxidation of SO2 in Clouds and Fogs . . . . . . . . . . Iron Oxides and Dissolved Iron Species as Chromophoric Reaction Initiators . . . . . . . . . . . . . . . . . . . Photolysis of Hydrogen Peroxide – A Direct Source of Hydroxyl Radical Nitrate and Nitrite Photolysis in Atmospheric Water and Ice . . . . . . Organic Compound Photochemistry in Clouds and Aerosols . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

78

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

79 79 80 82 82

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

82 83 90 94 96 97

. . . . . .

99 99 101

. . . . .

. . . . .

106 108 109 111 112

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

112

4.4 4.5 4.6 4.7

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

Abstract Key features of tropospheric photochemistry are highlighted including both homogeneous gas-phase and heterogeneous reactions that are important in clouds and haze aerosol. Fundamental aspects of photochemical kinetics are reviewed and then extended to the major chromophores present in the multi-phasic, tropospheric atmosphere. Tables of up-to-date absorption cross sections and quantum yields as a function of wavelength range are presented. Primary emphasis is placed on reactions occurring within the troposphere and within clouds. Keywords Absorption cross section · Hydrogen peroxide · Hydroxyl radical · Nitrogen dioxide · Nitrate · Nitrite · Ozone · Photochemistry · Quantum yield · Rate law

78

M.R. Hoffmann

1 Introduction In this overview and review of tropospheric photochemistry, we will examine a limited set of important homogeneous and heterogeneous photochemical reactions of relevance in the troposphere (Table 1). An expanded array of photochemical reactions is considered viable in the upper atmosphere (e.g., stratosphere) due to exposure to actinic radiation at wavelengths below 290 nm. A brief summary of a limited subset of this array of possible photochemical reactions will be provided in this review.

Table 1 A limited subset of photochemical reactions of interest in the troposphere O3 + hν → O2 + O(3 P) O3 + hν → O2 + O(1 D) HO2 + hν → products H2 O2 + hν → OH + OH NO2 + hν → NO + O NO3 + hν → NO2 + O NO3 + hν → NO + O2 N2 O5 + hν → products HONO + hν → OH + NO HO2 NO2 + hν → products CH2 O → H2 + CO CH2 O → H + HCO CH3 OOH + hν → products CH3 C(O)O2 NO2 + hν → products SO2 + hν → SO + O H2 S + hν → HS + H CS2 + hν → CS + S OCS + hν → CO + S

Essential background material and presentation of the fundamentals of atmospheric photochemistry can be obtained from key textbooks such as those by Seinfeld and Pandis [1], Finlayson-Pitts and Pitts [2], Yung and DeMore [3], and Warneck [4]. Ongoing evaluations of kinetic and photochemical data essential for atmospheric chemistry are provided by the Jet Propulsion Laboratory of the California Institute of Technology, by the IUPAC Subcommittee for Gas Kinetic Data Evaluation [5–12].

Homogeneous and Heterogeneous Photochemistry in the Troposphere

79

The current versions of these evaluations can be obtained at the following Web sites: JPL/NASA: http://jpldataeval.jpl.nasa.gov/ IUPAC: http://www.iupac-kinetic.ch.cam.ac.uk/summary/ IUPACsumm_web_latest.pdf

2 Basic Principles – Homogeneous Gas-Phase Photochemistry 2.1 Photochemical Reaction Rates Consider species X, which absorbs light in the appropriate wavelength range and as a consequence undergoes direct photodissociation as follows: X + hν → products

(1)

Where, d[X]/dt, the rate of change of X as a function of time, is given by d[X] =– σX (λ)φi (λ, T)I(λ)dλ [X] (2) dt where σX (λ) = absorption cross section, φ = quantum yield, λ = wavelength, I(λ) = spectral actinic ﬂux, d[X]/dt = photochemical reaction rate, and σX (λ)φi (λ, T)I(λ)dλ = photochemical reaction rate constant. Thus, the photochemical rate of disappearance of homogeneous gas-phase species X can be written as a simple pseudo ﬁrst-order rate constant as follows: d[X] = – kp [X] dt

(3)

k(X, z, χ) =

(4)

σa (X, λ)Φ(X, λ)I(λ, z, χ)dλ

where k(X, z, χ) is the ﬁrst order (s–1 ) rate for species X at altitude z for solar zenith angle χ; σa (X, λ) is the absorption cross section (cm2 ) as a function of λ, Φ(X, λ) is the quantum yield for photo-dissociation, I(λ, z, χ) is the solar actinic ﬂux (photons/cm2 -sec-nm) for wavelength interval dλ. Furthermore, I(λ, z, χ) = I∞ (λ)Tr (λ, z, χ)

(5)

where I∞ (λ) = the solar actinic ﬂux at the edge of the atmosphere and Tr (λ, z, χ) is the transmission function as follows:

Tr (λ, z, χ) = exp – τ O3 , λ, z, χ + τ O2 , λ, z, χ + τaerosol λ, z, χ (6)

80

M.R. Hoffmann

where ∞ τ ≡ “Optical Depth” = sec χ

σa (X, λ, z)n(X, z )dz

(7)

z

In the absence of light scattering, the ﬁrst-order photochemical rate constant can be rewritten as kp (λ, z, χ) = σi (X)φi (X) Ii,∞ Tr,i(z, χ) (8) i

where σi (X)φi (X) ≡ average of σa (X, λ)φ(X, λ) over interval ∆λi and

(9)

Ii,∞ =

q∞ (λ)dλ

(10)

∆λi

The local zenith angle χ varies with time of day, season, geographic location. For example, at latitude φ and a solar declination δ, χ varies with local hour angle, th , according to (th = 0 at local noon). cos χ = sin φ sin δ + cos δ cos φ cos th

(11)

At sunrise/sunset th = ± cos–1 (– tan φ tan δ)

(12)

where, the solar declination is the angle between sun’s direction and the earth’s equatorial plane. 2.2 Wavelength Range for Active Photochemistry In the troposphere, the wavelength range for dynamic photochemistry ranges from 290 nm up to 520 nm, while in the stratosphere the wavelength range is extended down to 200 nm. From the basic relationship between energy and wavelength, E=

hc λ

(13)

where, h is Planck’s constant and c is the speed of light we can determine the energy available in a mole of photons (or per photon) and see that in many cases sufﬁcient energy is available for direct bond cleavage. Given the values of h and c in the appropriate units and for one mole of photons (i.e.,

Homogeneous and Heterogeneous Photochemistry in the Troposphere

1 einstein) we can write Planck’s equation as hc E = 6.023 × 1023 λ

81

(14)

and further reduce it to the following form as follows: kJ 1.196 × 105 E= λ einstein

(15)

where E = energy of a photon (J) h = Planck’s constant (6.6 × 10–34 J s) c = speed of light (3.0 × 108 m s–1 ) λ = wavelength (m) ν = frequency (cycles s–1 or hertz). With this equation, we can readily determine that sufﬁcient energy is available for direct bond breaking over a fairly broad wavelength range. For example, the C – H bond strength in formaldehyde, H2 CO = 368 kJ mol–1 while λ = 290 nm corresponds to 412 kJ mol–1 . However, at λ = 700 nm, the available energy for bond breaking falls to only 171 kJ mol–1. Table 2 shows a range of typical bond energies and the wavelengths that correspond to these respective energies as calculated using the simple formulation of 15. For example, in the case of ozone (O3 ), the O – O bond energy is 6.19 × 10–19 J/bond or 372 kJ mol–1 while in the case of oxygen the corresponding O = O bond energy is 8.18 × 10–19 J/bond or 493 kJ mol–1 . The ﬁrst law of photochemistry states that a molecule must ﬁrst absorb light in order to undergo a photochemical transformation. This law in embodied in the Beer-Lambert law I = e–σNL (16) I0 Table 2 Energy available in one mole of photons at λi and its correspondence to typical chemical bond energies λi (nm)

E (kJ mol–1 )

Bond

257 288 307 332 344

465 415 390 360 348

O–H C–H N–H C–O C–C

353 620

339 193

C – Cl Br – Br

82

M.R. Hoffmann

or ln

I0 = σ NL I

(17)

where σ = “absorption cross section” in cm2 molec–1 , N = number concentration in molecules cm–3 , and L = path length in cm. 2.3 Quantum Yields The quantum yield or photoefﬁciency of a photochemical process of interest is deﬁned as # of molecules reacting via pathway i (18) φi (λ) = total number of photons absorbed by reacting molecule whereas the second law of photochemistry is simply stated as φi (λ) = 1

(19)

i

where the various pathways i include: photodissociation, photoionization, ﬂuorescence, phosphorescence, quenching, chemical reaction, intermolecular energy transfer, and intramolecular energy transfer. 2.4 Simplified Photochemical Kinetics The above treatment of the photochemical reaction rate can be simpliﬁed as follows: d[X] = – kp (λ)[X] (20) dt λ λi d[X] – = φ(λ)Ia (λ)dλ (21) dt reac i 290

Ia (λ) = σX (λ)I(λ)[X]

(22)

3 Homogeneous Gas-Phase Photochemistry A few key (i.e., primary or direct) photochemical reactions are the principal drivers of overall chemical reactions in the troposphere. These reactions involve primarily O3 and NO2 photolysis. Other reactions presented in Table 1 will be discussed later.

Homogeneous and Heterogeneous Photochemistry in the Troposphere

83

3.1 Ozone Photolysis The most important sets of photochemical reactions in the troposphere involve the photolysis of ozone, O3 which occurs over a very broad wavelength Table 3 Absorption cross sections of O3 at 273 K [106] λ (nm)

1020 σ(cm2 ) average λ (nm)

20σ(cm2 ) average

175.439–176.991 176.991–178.571 178.571–180.180

81.1 79.9 78.6

238.095–240.964 240.964–243.902 243.902–246.914

797 900 1000

180.180–181.818 181.818–183.486 183.486–185.185 185.185–186.916 186.916–188.679

76.3 72.9 68.8 62.2 57.6

246.914–250.000 250.000–253.165 253.165–256.410 256.410–259.740 259.740–263.158

1080 1130 1150 1120 1060

188.679–190.476 190.476–192.308 192.308–194.175 194.175–196.078 196.078–198.020

52.6 47.6 42.8 38.3 34.7

263.158–266.667 266.667–270.270 270.270–273.973 273.973–277.778 277.778–281.690

965 834 692 542 402

198.020–200.000 200.000–202.020 202.020–204.082 204.082–206.186 206.186–208.333

32.3 31.4 32.6 36.4 43.4

281.690–285.714 285.714–289.855 289.855–294.118 294.118–298.507 298.507–303.030

277 179 109 62.4 34.3

208.333–210.526 210.526–212.766 212.766–215.054 215.054–217.391 217.391–219.780

54.2 69.9 92.1 119 155

303.030–307.692 307.692–312.5 312.5–317.5 317.5–322.5 322.5–327.5

219.780–222.222 222.222–224.719 224.719–227.273 227.273–229.885 229.885–232.558

199 256 323 400 483

327.5–332.5 332.5–337.5 337.5–342.5 342.5–347.5 347.5–352.5

0.617 0.274 0.117 0.0588 0.0266

232.558–235.294 235.294–238.095

579 686

352.5–357.5 357.5–362.5

0.0109 0.00549

18.5 9.80 5.01 2.49 1.20

84

M.R. Hoffmann

Table 4 Reported quantum yields for O(1 D) production from ozone photolysis λ (nm)

φ - d[O(1 D)]/dt

Reference

308 308 308

0.79 0.79 0.79

Talukdar et al. [30] Takahashi et al. [18, 23] Greenblatt and Wiesenfeld [225]

266 248 248 248

0.88 0.91 0.94 0.85

Brock and Watson [42] Talukdar et al. [30] Greenblatt and Wiesenfeld [225] Amimoto et al. [226]

Fig. 1 Quantum yield for O(1 D) production from ozone photolysis as a function of wavelength [16, 19, 29, 42, 43, 227–229]

range in ﬁve separate channels as follows [13–55]: hν ≤ 310 nm O3 → O(1 D) + O2 a1 ∆g – hν ≤ 411 nm 1 → O( D) + O2 X 3 O3 g + hν ≤ 463 nm O3 → O(1 D) + O2 b1 g

O3 O3

→ O(3 P) + O2 a1 ∆g – hν ≤ 1180 nm 3 → O( P) + O2 X 3 hν ≤ 612 nm

g

(23) (24) (25) (26) (27)

Homogeneous and Heterogeneous Photochemistry in the Troposphere

85

The absorption cross sections for O3 at 273 K as recommended by the 2004 JPL/NASA evaluation are given in Table 3 while measured quantum yields (Fig. 1) for ozone photolysis are given in Table 4. In addition, the JPL/NASA evaluation committee has developed an empirical equation for the estimation of the quantum yield for O(1 D) as a function of wavelength and temperature (Table 5). Perhaps, the most critical atmospheric consequence of the photolysis of ozone is the production of O(1 D) (Eqs. 23–27). O(1 D) reacts with atmospheric water to produce hydroxyl radical · OH, as follows: O(1 D) + H O –→ 2 · OH (28) 2

Hydroxyl radical, · OH, is the principal atmospheric oxidant for a vast array of organic and inorganic compounds in the atmosphere. In addition to being the primary oxidant of non-methane hydrocarbons (representative examples of these secondary reactions are given in Table 6), · OH radical controls the rate of CO and CH4 oxidation. Furthermore, the · OH reaction with ozone also limits the destruction of O3 in the troposphere, it also determines the lifetime of CH3 Cl, CH3 Br, and a wide range of HCFC’s, and it controls the rate of NOx to HNO3 conversion. Concentration proﬁles for hydroxyl radical in the atmosphere are shown in Fig. 2. Table 5 Parameters for the calculation of O(1 D) quantum yields X1 – λ 4 q1 × A1 × exp – lΦ(λ, T) = q1 + q2 ω1 2 q2 T X2 – λ 2 + × A2 × × exp – q1 + q2 300 ω2 1.5 2 T X3 – λ + A3 × +c × exp – 300 ω3 νi where qi = exp – RT and X1–3 , A1–3 , ω1–3 , ν1–2 and c are best-ﬁt parameters, λ is in nm, T is in K, and R = 0.695 (cm–1 /K). The parameter c is assumed to be temperature and wavelength independent. This expression is valid only for the wavelength range 306–328 nm and temperature range 200–320 K. Parameter

i=1

i=2

i=3

Xi (nm) ωi (nm) Ai

304.225 5.576 0.8036

314.957 6.601 8.9061

310.737 2.187 0.1192

825.518 –

– –

νi (cm–1 ) c

0 0.0765

86

M.R. Hoffmann

Fig. 2 Hydroxyl radical production pathways and concentrations as a function of altitude Table 6 Representative OH radical reactions and their 2nd -order rate constants [106]. Units are in cm3 molecule–1 s–1 Reaction

A-Factor

E/R

k(298 K)

O + OH → O2 + H OH + O3 → HO2 + O2 OH + H2 → H2 O + H

2.2 × 10–11 1.7 × 10–12 5.5 × 10–12

– 120 940 2000

3.3 × 10–11 7.3 × 10–14 6.7 × 10–15

OH + OH → H2 O + O OH + HO2 → H2 O + O2 OH + H2 O2 → H2 O + HO2 OH + HONO → H2 O + NO2 OH + HO2 NO2 → products

4.2 × 10–12 4.8 × 10–11 2.9 × 10–12 1.8 × 10–11 1.3 × 10–12

240 – 250 160 390 – 380

1.9 × 10–12 1.1 × 10–10 1.7 × 10–12 4.5 × 10–12 4.6 × 10–12

OH + NH3 → H2 O + NH2 OH + CO → products

1.7 × 10–12 1.5 × 10–13 ×(1 + 0.6Patm )

OH + CH4 → CH3 + H2 O OH + H2 CO → H2 O + HCO OH + CH3 OH → products OH + CH3 OOH → products OH + HC(O)OH → products

2.45 × 10–12 9.0 × 10–12 7.3 × 10–12 3.8 × 10–12 4.0 × 10–13

710 0 1775 0 620 – 200 0

1.6 × 10–13 1.5 × 10–13 ×(1 + 0.6Patm ) 6.3 × 10–15 9.0 × 10–12 9.1 × 10–13 7.4 × 10–12 4.0 × 10–13

Homogeneous and Heterogeneous Photochemistry in the Troposphere

87

Table 6 (continued) Reaction

A-Factor

E/R

k(298 K)

OH + C2 H6 → H2 O + C2 H5

8.7 × 10–12

1070

2.4 × 10–13

OH + C3 H8 → H2 O + C3 H7 OH + CH3 CHO → CH3 CO + H2 O OH + C2 H5 OH → products OH + CH3 C(O)OH → products OH + CH3 ONO2 → products

1.0 × 10–11 5.6 × 10–12 6.9 × 10–12 4.0 × 10–13 5.0 × 10–13

660 – 270 230 – 200 810

1.1 × 10–12 1.4 × 10–11 3.2 × 10–12 8.0 × 10–13 3.3 × 10–14

OH + C2 H5 ONO2 → products OH + 1 – C3 H7 ONO2 → products OH + 2 – C3 H7 ONO2 → products OH + CH3 Cl → CH2 Cl + H2 O OH + CH2 Cl2 → CHCl2 + H2 O

6.8 × 10–13 7.1 × 10–13 1.2 × 10–12 2.4 × 10–12 1.9 × 10–12

320 0 320 1250 870

2.3 × 10–13 7.1 × 10–13 4.1 × 10–13 3.6 × 10–14 1.0 × 10–13

OH + CHCl3 → CCl3 + H2 O OH + CH2 ClCH3 → products OH + CH3 CCl3 → CH2 CCl3 + H2 O OH + CH2 = CHCl → products

2.2 × 10–12 5.4 × 10–12 1.6 × 10–12 1.3 × 10–12

920 800 1520 – 500

1.0 × 10–13 3.7 × 10–13 1.0 × 10–14 6.9 × 10–12

OH + CH2 = CCl2 → products OH + CHCl = CCl2 → products OH + CCl2 = CCl2 → products OH + CCl3CHO → H2 O + CCl3 CO OH + Br2 → HOBr + Br

1.9 × 10–12 8.0 × 10–13 4.7 × 10–12 9.1 × 10–12 4.2 × 10–11

– 530 – 300 990 580 0

1.1 × 10–11 2.2 × 10–12 1.7 × 10–13 1.3 × 10–12 4.2 × 10–11

1.1 × 10–11 2.35 × 10–12 2.0 × 10–12 1.35 × 10–12 1.0 × 10–12

0 1300 840 600 1380

1.1 × 10–11 3.0 × 10–14 1.2 × 10–13 1.8 × 10–13 1.0 × 10–14

OH + CH2 ClBr → CHClBr + H2 O OH + CH3 I → H2 O + CH2 I OH + CF3 I → HOI + CF3 OH + H2 S → SH + H2 O OH + OCS → products

2.4 × 10–12 2.9 × 10–12 2.5 × 10–11 6.0 × 10–12 1.1 × 10–13

920 1100 2070 75 1200

1.1 × 10–13 7.2 × 10–14 2.4 × 10–14 4.7 × 10–12 1.9 × 10–15

OH + CH3 SH → CH3 S + H2 O OH + CH3 SCH3 → H2 O + CH2 SCH3 OH + CH3 SSCH3 → products

9.9 × 10–12 1.2 × 10–11 6.0 × 10–11

– 360 260 – 400

3.3 × 10–11 5.0 × 10–12 2.3 × 10–10

OH + HBr → H2 O + Br OH + CH3 Br → CH2 Br + H2 O OH + CH2 Br2 → CHBr2 + H2 O OH + CHBr3 → CBr3 + H2 O OH + CHF2 Br → CF2 Br + H2 O

88

M.R. Hoffmann

A primary example of the interplay between hydroxyl radical and ozone is provided by the photo-induced oxidation of methane in the troposphere. The overall sequence of stoichiometric reactions can be written as: · OH + CH4 → H2 O + CH3 · (29) · CH3 + O2 + M → CH3 O2 + M CH3 O2 · + NO → NO2 + CH3 O CH3 O + O2 → H2 CO + HO2 · H2 CO + hν → H· + HCO· HCO· + O2 → HO2 · + CO CO + · OH → H· + CO2 2 × H· + O2 + M → HO2 · + M 3 × HO2 · + NO → NO2 + · OH 4 × NO2 + hν → NO + O· 4 × O· + O2 + M → M + O3 · OH + HO · → H O + O 2 2 3 The net result of methane oxidation in the remote troposphere by hydroxyl radical produces 3 molecules of ozone for each molecule oxidized. hν

CH4 + 8 O2 –→ CO2 + 4 H2 O + 3 O3

(30)

Another example of the overall feedback process (Figs. 3 and 4) involving ozone and hydroxyl radical is provided by the oxidation pathway for carbon

Fig. 3 Important reaction cycles and feedback loops for OH radical in the troposphere

Homogeneous and Heterogeneous Photochemistry in the Troposphere

89

Fig. 4 Additional reaction cycles for hydroxyl radical involving hydroperoxyl radical Table 7 Rate constants for termolecular reactions involving photo-generated OH radical [106] Reaction

Low-Pressure Limita –n ko (T) = k300 o (T/300)

High-Pressure Limitb –m k∞ (T) = k300 ∞ (T/300)

k300 o

k300 ∞

n

m

M H· + O2 –→ HO2 ·

(5.7 ± 0.5) × 10–32 1.6 ± 0.5

(7.5 ± 4.0) × 10–11 0 ± 1.0

M HO· + HO· –→ H2 O2

6.9 × 10–31

1.0

(2.6) × 10–11

0

2.6 ± 0.3

(3.6 ± 1.0) × 10–11

0.1 ± 0.5

3.0

2.5 × 10–11

0

0.0 ± 0.2

(8.3 ± 1.0) × 10–13

– 2 ± 0.2

M

HO· + NO· –→ HONO

(7.0 ± 1.0) × 10–31

M HO· + NO2 –→ HONO2

2.0 × 10–30

M

(5.5 ± 2.0) × 10–30

M

HO· + C2 H4 –→ HOC2 H4 ·

(1.0 ± 0.6) × 10–28 0.8 ± 2.0

(8.8 ± 0.9) × 10–12 0 ± 0.2

M

(3.0 ± 1.0) × 10–31 3.3 ± 1.5

(1.5 ± 0.5) × 10–12 0 ± 0.2

HO· + C2 H2 –→ HOC2 H2 · HO· + SO2 –→ HOSO2 · a

The low-pressure-limiting rate constants are given in the form: T –n 6 ko T = k300 cm molecule–2 s–1 , o 300

where k300 o has been adjusted for air as the third body, together with a value of n. b The limiting high-pressure rate constant is given in a similar form: T –m 3 k∞ T = k300 cm molecule–1 s–1 . ∞ 300 The effective second-order rate constant for a given condition of temperature and pressure (altitude = [M]), the following formula is used: 2 –1 1+ log10 kok (T)[M] ko T [M] ∞ (T) 0.6 kf ([M], T) = 1 + kok(T)[M] ∞ (T)

90

M.R. Hoffmann

monoxide, CO, as follows: CO + · OH → H· + CO2 H· + O2 + M → HO2 · + M HO2 · + NO → NO2 + · OH NO2 + hν → O· + NO

(31)

where M, in the normal case, represents the classical third-body in a termolecular reaction. The overall reaction for the sequence of Eq. 31 can be written simply as: hν

CO + 2 O2 –→ CO2 + O3

(32)

3.2 Nitrogen Dioxide Photolysis hν

NO2 –→ NO + O(3 P)

(33)

Nitrogen dioxide photolysis is a key driver of tropospheric atmospheric chemistry since it leads directly to the production and eventual consumption of ozone during the daytime as follows: (34) O· (3 P) + O2 + M –→ O3 + M (a) NO + O3 –→ NO2 + O2

(b)

The absorption cross sections for NO2 and the corresponding quantum yields are given in Table 8 and 9, respectively. The photolysis of NO2 has been investigated intensively over the last 40 years because of its critical role in the formation of ozone in the polluted tropospheric boundary layer [56–63]. The three reactions of Eqs. 33 and 34 form the basis for the photochemical production of ozone. If one considers only these three reactions, then the photo-stationary state (or photochemical steady-state approximation) can be invoked around the oxygen atom as follows: d[O· ] = 0 = khν,33 [NO2 ] – k34a [O· ][O2 ][M] (35) dt where [M] is the concentration of the background (i.e., non-reactive) gases (e.g., O2 and N2 ) functioning as traditional third-body molecules in molecular collisions. The steady-state approximation yields the following relationship for [O· ]: khν,33 [NO2 ] (36) [O· ]ss = k34a [O2 ][M]

Homogeneous and Heterogeneous Photochemistry in the Troposphere

91

Table 8 Absorption cross sections of NO2 [106] λ (nm)

1020 σ; T = 25 ◦ C (cm2 molecule–1 )

λ(nm)

1020 σ; T = 25 ◦ C (cm2 molecule–1 )

202.02 – 204.08 204.08 – 206.19 206.19 – 208.33

41.45 44.78 44.54

273.97 – 277.78 277.78 – 281.69 281.69 – 285.71

5.03 5.88 7.00

208.33 – 210.53 210.53 – 212.77 212.77 – 215.06 215.06 – 217.39 217.39 – 219.78

46.41 48.66 48.18 50.22 44.41

285.71 – 289.85 289.85 – 294.12 294.12 – 298.51 298.51 – 303.03 303.03 – 307.69

8.15 9.72 11.54 13.44 15.89

219.78 – 222.22 222.22 – 224.72 224.72 – 227.27 227.27 – 229.89 229.89 – 232.56

47.13 37.72 39.29 27.40 27.78

307.69 – 312.50 312.5 – 317.5 317.5 – 322.5 322.5 – 327.5 327.5 – 332.5

18.67 21.53 24.77 28.07 31.33

232.56 – 235.29 235.29 – 238.09 238.09 – 240.96 240.96 – 243.90 243.90 – 246.91

16.89 16.18 8.812 7.472 3.909

332.5 – 337.5 337.5 – 342.5 342.5 – 347.5 347.5 – 352.5 352.5 – 357.5

34.25 37.98 40.65 43.13 47.17

246.91 – 250.00 250.00 – 253.17 253.17 – 256.41 256.41 – 259.74 259.74 – 263.16

2.753 2.007 1.973 2.111 2.357

357.5 – 362.5 362.5 – 367.5 367.5 – 372.5 372.5 – 377.5 377.5 – 382.5

48.33 51.66 53.15 55.08 56.44

263.16 – 266.67 266.67 – 270.27 270.27 – 273.97

2.698 3.247 3.785

382.5 – 387.5 387.5 – 392.5 392.5 – 397.5 397.5 – 402.5 402.5 – 407.5

57.57 59.27 58.45 60.21 57.81

407.5 – 412.5 412.5 – 417.5 417.5 – 422.5

59.99 56.51 58.12

92

M.R. Hoffmann

Table 9 Quantum yields for NO2 photolysis [106] λ (nm)

φ

λ (nm)

φ

< 285 290 295 300

1.000 0.999 0.998 0.997

393 394 395 396

0.953 0.950 0.942 0.922

305 310 315 320 325

0.996 0.995 0.994 0.993 0.992

397 398 399 400 401

0.870 0.820 0.760 0.695 0.635

330 335 340 345 350

0.991 0.990 0.989 0.988 0.987

402 403 404 405 406

0.560 0.485 0.425 0.350 0.290

355 360 365 370 375

0.986 0.984 0.983 0.981 0.979

407 408 409 410 411

0.225 0.185 0.153 0.130 0.110

380 381 382 383 384

0.975 0.974 0.973 0.972 0.971

412 413 414 415 416

0.094 0.083 0.070 0.059 0.048

385 386 387 388 389

0.969 0.967 0.966 0.964 0.962

417 418 419 420 421

0.039 0.030 0.023 0.018 0.012

390 391 392

0.960 0.959 0.957

422 423 424

0.008 0.004 0.000

Homogeneous and Heterogeneous Photochemistry in the Troposphere

93

The corresponding steady-state treatment for O3 d[O3 ] = 0 = k34a [O· ][O2 ][M] – k34b [NO][O3 ] dt

(37)

yields the following expression for [O3 ]ss : [O3 ]ss =

k34a [O· ][O2 ][M] khν,33 [NO2 ] = k34b [NO] k34b [NO]

(38)

A typical value [1] for khν,33 at noontime in Los Angeles on June 21 is 8.9 × 10–3 s–1 . khν,33 , of course, is dependent on incident light intensity. As described above, ozone is photolyzed via an alternative channel to produce O(1 D) hν

O3 –→ O(1 D) + O2 khν,39

(39)

where under the same noontime irradiation conditions, khν,39 = 3.7 × 10–5 s–1 . However, the above NOx /O3 chemistry is enhanced by the impact of reactive hydrocarbons, formaldehyde and other reactive species. For example, in the case of generic, reactive hydrocarbon, RH, the following reaction se-

Fig. 5 Reaction pathways initiated by the photolysis of methyl hydroperoxide and formaldehyde

94

M.R. Hoffmann

Fig. 6 Reaction cycles involving NO, NO2 , HONO, and other nitrogen oxide compounds

quence takes place under solar irradiation: RH +· OH –→ H2 O + R· R· + O2 –→ RO2 · RO2 · + NO2 + M → ROONO2 · + M RO2 · + NO → NO2 + RO·

(40) (41) (42) (43)

The feedback loop (Figs. 5 and 6) with respect to ozone involves competition for nitric oxide, NO, as follows: O3 + NO → NO2 + O2

(44)

In the speciﬁc case of propane, the following reactions occur: ˙ CH3 CH2 CH3 + · OH –→ CH3 CHCH 3 + H2 O · ˙ CH3 CHCH 3 + O2 –→ CH3 CH(O2 )CH3 CH3 CH(O2 · )CH3 + NO –→ NO2 + CH3 CH(O· )CH3 CH3 CH(O· )CH3 + O2 –→ CH3 C(O)CH3 + HO2 · HO2 · + NO –→ NO2 + · OH

(45) (46) (47) (48) (49)

3.3 Formaldehyde Photolysis Formaldehyde, HCHO, photolysis [64–76] provides a major source of free radicals in troposphere [2] (Fig. 7). It has a highly structured UV-vis ab-

Homogeneous and Heterogeneous Photochemistry in the Troposphere

95

Fig. 7 Reaction cycles involving oxygen-containing free-radical species (i.e., hydrogen oxide radicals)

sorption spectrum that extends into upward to 400 nm as shown in Table 10. There are two photolysis pathways. One leads to formation of stable products, hydrogen and carbon monoxide as shown as follows hν

HCHO –→ H2 + CO hν

HCHO –→ H· + HCO·

(50) (51)

Both the H· and HCO· radical fragment can in turn react with O2 to yield hydroperoxyl radical as follows: H· + O2 –→ HO2 · HCO· + O2 –→ HO2 · + CO

(52) (53)

In addition, the HCO· radical fragment is also produced by the reaction of HCHO and hydroxyl radical HCHO + · OH –→ HCO + H2 O

(54)

which in turn will lead to the production of a second hydroperoxyl radical, HO2 · . Rogers [77] has estimated that for noontime solar irradiation that 45% of the photolysis of HCHO proceeds via reaction 51 while 55% proceeds via reaction 50. Reaction 51 predominates at shorter wavelengths while reaction 50 becomes more important at longer wavelengths. The corresponding solar noontime values [1] for khν,50 and khν,51 are 3.0 × 10–5 s–1 and 4.3 × 10–5 s–1 , respectively.

96

M.R. Hoffmann

Table 10 Absorption cross sections and quantum yields for photolysis of CH2 O [106] λ (nm)

1020 σ (cm2 ) 223 K 293 K

φ1 (H + HCO)

φ (H2 + CO)

301.25 303.75 306.25 308.75 311.25

1.38 4.67 3.32 2.27 0.758

1.36 4.33 3.25 2.22 0.931

0.749 0.753 0.753 0.748 0.739

0.251 0.247 0.247 0.252 0.261

313.75 316.25 318.75 321.25 323.75

3.65 4.05 1.66 1.24 0.465

3.40 3.89 1.70 1.13 0.473

0.724 0.684 0.623 0.559 0.492

0.276 0.316 0.368 0.423 0.480

326.25 328.75 331.25 333.75 336.25

5.06 2.44 1.39 0.093 0.127

4.44 2.29 1.28 0.123 0.131

0.420 0.343 0.259 0.168 0.093

0.550 0.634 0.697 0.739 0.728

338.75 341.25 343.75 346.25 348.75

3.98 0.805 1.44 0.004 0.009

3.36 0.936 1.26 0.071 0.040

0.033 0.003 0.001 0 0

0.667 0.602 0.535 0.469 0.405

351.25 353.75 356.25

0.169 1.83 0.035

0.235 1.55 0.125

0 0 0

0.337 0.265 0.197

3.4 Hydrogen Peroxide Photolysis Hydrogen peroxide [78–81] is formed in the atmosphere as an indirect, secondary photochemical product from the self-reaction [2] of two hydroperoxyl radicals as follows: 2HO2 ·

M, H2 O

→ H2 O2 + O2

(55)

where k = 5.5 × 10–12 cm3 molec–1 s–1 at 1 atm total pressure and 50% relative humidity. Average concentrations across the continental United States range from 2 to 4 ppbv [1]. The absorption cross sections for H2 O2 as a function of

Homogeneous and Heterogeneous Photochemistry in the Troposphere

97

Table 11 Absorption cross sections of H2 O2 in the gas phase [106] λ (nm)

1020 σ (cm2 ) 298 K 355 K

λ (nm)

1020 σ (cm2 ) 298 K 355 K

190 195 200 205 210

67.2 56.4 47.5 40.8 35.7

270 275 280 285 290

3.3 2.6 2.0 1.5 1.2

3.5 2.8 2.2 1.6 1.3

215 220 225 230 235

30.7 25.8 21.7 18.2 15.0

18.4 15.2

295 300 305 310 315

0.90 0.68 0.51 0.39 0.29

1.0 0.79 0.58 0.46 0.36

240 245 250 255 260

12.4 10.2 8.3 6.7 5.3

12.6 10.8 8.5 6.9 5.5

320 325 330 335 340

0.22 0.16 0.13 0.10 0.07

0.27 0.21 0.17 0.13 0.10

265

4.2

4.4

345 350

0.05 0.04

0.06 0.05

wavelength are given in Table 11. Even though there is a rapid fall off of the absorption cross sections [82–85] above 290 nm, photolysis [86–96] of H2 O2 is still a primary loss pathway in the atmosphere with the corresponding production of hydroxyl radical. hν

H2 O2 –→ 2 · OH

(56)

The quantum yield for photo-dissociation of H2 O2 is close to 1.0 at all wavelengths. 3.5 Nitrous Acid (HONO) Photolysis Nitrous acid, HONO, is considered to be an important species [2] in the photochemistry of the troposphere since it is a source of hydroxyl radical [97]. hν HONO –→ NO· + · OH

(57)

The absorption cross sections [98–103] for HONO are shown in Table 12. At a solar zenith angle of 40◦ the ﬁrst-order photolysis rate constant [104],

98

M.R. Hoffmann

Table 12 Absorption cross sections of HONO [106] λ (nm)

1020 σ (cm2 ) λ (nm)

1020 σ (cm2 )

λ (nm)

1020 σ(cm2 )

310

1.3

339

18.8

368

52.0

311 312 313 314 315

1.9 2.8 2.2 3.6 3.0

340 341 342 343 344

10.0 17.0 38.6 14.9 9.7

369 370 371 372 373

38.8 17.8 11.3 10.0 7.7

316 317 318 319 320

1.4 3.1 5.6 3.6 4.9

345 346 347 348 349

10.9 12.3 10.4 9.1 7.9

374 375 376 377 387

6.2 5.3 5.3 5.0 5.8

321 322 323 324 325

7.8 4.9 5.1 7.1 5.0

350 351 352 353 354

11.2 21.2 15.5 19.1 58.1

379 380 381 382 383

8.0 9.6 11.3 15.9 21.0

326 327 328 329 330

2.9 6.6 11.7 6.1 11.1

355 356 357 358 359

36.4 14.1 11.7 12.0 10.4

384 385 386 387 388

24.1 20.3 13.4 9.0 5.6

331 332 333 334 335

17.9 8.7 7.6 9.6 9.6

360 361 362 363 364

9.0 8.3 8.0 9.6 14.6

389 390 391 392 393

3.4 2.7 2.0 1.5 1.1

336 337 338

7.2 5.3 10.0

365 366 367

16.8 18.3 30.2

394 395 396

0.6 1.0 0.4

khν,54 ≈ 1.3 × 10–3 where φ = 0.92 at 365 nm [105]. In competition with photolysis, the fate of HONO in the atmosphere can be controlled by reaction with · OH [97, 106]. HONO + · OH –→ NO2 + H2 O

(58)

Homogeneous and Heterogeneous Photochemistry in the Troposphere

99

where k298,55 = 4.5 × 10–12 cm3 molec–1 s–1 . In a comprehensive ﬁeld experiment in the region around Berlin, Alicke et al. [107] found that HONO photolysis contributed up to 20% of the total · OH formed in a 24-h period with O3 and HCHO photolysis the other major contributors. However, in the early morning hours, HONO photolysis was clearly the predominant source of · OH.

4 Heterogeneous Photochemistry in the Troposphere Over the last 20 years, the potential role of heterogeneous photochemistry relevant to atmospheric processes has received greater attention. The heterogeneous phases of interest include clouds, fogs, haze aerosol, dry aerosol, ice particles, sea ice and snow. Early interest was focused on illuminated clouds and fogs. In a 1994 review, Faust [108] reviewed some of the important photochemical reactions occurring in clouds, fogs, and aerosols. Important liquid-phase reactions include the photochemical production of H2 O2 , · OH, HO2 · , O2 –· and O2 (1 ∆g ) and the photo-oxidation of dissolved SO2 . 4.1 Heterogeneous and Aqueous-Phase Photochemistry In contrast to gas-phase photochemistry, aqueous-phase and solid-phase photochemistry are treated differently. For example, the Beer-Lambert Law for homogenous liquid systems is expressed simply as: A = ε C

(59)

where A is the dimensionless absorbance, ε is the molar extinction coefﬁcient, M–1 cm–1 (or L mole–1 cm–1 ), C is the concentration of the light-absorbing molecule or chromophore in units of M (or mole L–1 ): and is the path length through the reacting medium in units of cm. In more complete form the BeerLambert law of aqueous systems can be expressed as follows: A ≡ log

I0 (λ) = [α(λ) + ε(λ)C] I(λ)

(60)

where I = light intensity (incident over transmitted), λ = wavelength, α = absorption coefﬁcient of the medium, ε = molar extinction coefﬁcient as before and absorption coefﬁcient of the compound, and = path length. In aqueous solution or another solvent medium d[A] = – kp (λ) [A] dt d[A] = – φr (λ)ka (λ) [A] dt

(61) (62)

100

M.R. Hoffmann

where kp (λ) the apparent ﬁrst-order rate constant (s–1 ) for direct photolysis at wavelength, l; φr (λ) = quantum yield for primary photochemical process of interest (mol ein–1 ), and ka (λ) = light absorption rate (mol ein –1 s–1 ). The quantum yield is again used as a measure of photochemical efﬁciency. φr (λ) (mol ein–1 ) =

moles of compound transformed d[A]/dt = moles of photons absorbed ka (λ)

(63)

In this case, 1 einstein (ein) is equal to 1 mole of photons. Furthermore, the photochemical rate of reaction can be expressed as follows for a wellcontrolled laboratory reactor or vessel with a primary reactant and chromophore, A: d[A] = φ(λ)I0 (λ)(1 – 10–ε [A] ) (64) dt where I0 (λ) is the incident light intensity on the reaction volume and (1 – 10–ε [A] ) is the fraction of incident light that is actually absorbed by the primary chromophore, A. Thus, –

ka (λ)[A] = I0 (λ)(1 – 10–ε [A] )

(65)

In the case of ill-deﬁned photon absorbing volumes such as cloud droplets or arrays of droplet in a well-formed cloud, the speciﬁc light absorption rate can be estimated using the following relationship: ka (λ) = S(λ)W(λ)D0 (λ)ε(λ)

(66)

where S(λ) is a screening factor to account for competitive absorbers, I 0 (λ) is the incident photon ﬂux in units of (einstein cm–2 s–1 nm–1 ), Do (λ) is the effective path length through the irradiated medium (e.g., clouds; ratio of actual path length to direct linear path length – dimensionless), and ε(λ) is the molar extinction coefﬁcient. In addition, as solar light penetrates through a cloud the attenuation of incident light can be approximated by I(zdepth , λ) = 10–αD (λ)zdepth I0 (λ)

(67)

where zdepth is the penetration depth, αD (λ) is the diffuse attenuation coefﬁcient and 1 I0 (λ) (68) αD (λ) = log zdepth I(zdepth , λ) (λ) αD (λ) D(λ) = = (69) zdepth α(λ) S(λ) =

1 – 10–αD (λ)zdepth 2.303 zdepth αD (λ)

where α(λ) is attenuation coefﬁcient of the pure water or the background liquid medium.

Homogeneous and Heterogeneous Photochemistry in the Troposphere

101

Combining terms we obtain ka (λ) =

(1 – 10–αD (λ)zdepth )I0 (λ)ε(λ) zdepth α(λ) 1

(70)

which can be reduced to the following expression after expansion of logarithmic term to give ka (λ) =

(2.3αD (λ)zdepth )I0 (λ)ε(λ) zdepth α(λ) 1

(71)

that further reduces to ka (λ) = 2.3I0 (λ)D(λ)ε(λ)

(72)

In the case of high liquid water clouds, the D(λ) value can be as high as 10 as noted by Madronich [109, 110] and co-workers. For example, they note that “measurements of spectral ultraviolet-B irradiance under optically thick clouds show strongly enhanced attenuation by molecular and particulate absorbers and that the photon path is enhanced due to the presence of the highly scattering medium, leading to an ampliﬁcation of absorption by chromophores. Using discrete ordinate and Monte Carlo model caculations, they [110] showed that “photon paths” (i.e., D(λ)) in realistic water clouds could be enhanced by factors of 10 and more compared to cloudless sky.” 4.2 Photochemical Oxidation of SO2 in Clouds and Fogs Sulfur dioxide is oxidized in the atmosphere mainly within clouds, fogs and other aqueous-phase domains. The primary pathway [111–115] involves oxidation with H2 O2 ; however, other reaction pathways are viable depending on pH. Some of the direct and indirect photochemical reactions of interest related to the fate of SO2 in the atmosphere include: O2 + 2 SO3 2–

Fe(III)

O2 + 2 SO3 2–

Fe(III),Mn(II),hν

O2 + 2 HSO3 –

→ 2SO4 2–

Cu(II),Co(II) hν

→ 2 SO4 2–

→ 2 HSO4 –

α–Fe2 O3

(73) (74) (75)

SO3 2– + O3 –→ SO4 2– + O2

(76)

H2 O2 + HSO3 – –→ HSO4 – + H2 O

(77)

Early studies of the photolysis of aquated sulfur dioxide and its acid dissociation products in water, bisulﬁte (HSO3 – ) and sulﬁte (SO3 2– ) in the presence of

102

M.R. Hoffmann

oxygen were focused on understanding light-induced free-radical chain reactions. For example, in the early part of the 20th century, Backstrom [116–122] studied the light-induced photooxidation of sulﬁte with light centered at 254 nm and concluded that the reaction could be initiated by the direct ejection of an electron to a solvated water molecule to yield the sulﬁte radical anion and an aquated electron as follows: SO3 2– + O2

hν, λ=254 nm

→ SO3 –· + e¯aq

(78)

Further studies on this simple system, were carried out by a variety of investigators [123–128] in the subsequent years. More recently, Warneck and co-workers reexamined the photolysis of sulﬁte [129] and bisulﬁte [130] relevant to atmospheric processes. In their study of the photooxidation of sulﬁte, Deister and Warneck [129] irradiated aqueous solutions of Na2 SO3 with UV light at 254 nm and determined that in the absence of oxygen that the major products were dithionate, S2 O6 2– , and sulfate, SO4 2– in a molar ratio of 0.5 with φ = 0.85. However, in the presence of O2 , dithionate was not observed and the relative quantum yield was found to be φ = 500. Such high apparent quantum efﬁciencies indicate a free-radical chain mechanism is operative. Furthermore, when benzene was added as a radical scavenger phenol was formed indicating the presence of hydroxyl radical; however, further studies allowed them to conclude that sulfate radical was the main free chain carrier. This was conﬁrmed by the addition of t– BuOH, which reacts rapidly with hydroxyl radical but slowly with sulfate radical anion. In this case, the apparent yield of phenol was not affected; this provided strong evidence for SO4 –· radical as the primary chain carrier. After photochemical initiation with the formation of the sulﬁte radical anion, the following reactions constitute the chain propagation steps: SO3 –· + O2 –→ SO5 –· SO5 –· + SO3 2– –→ SO4 2– + SO4 –· SO5 –· + SO3 2– –→ SO5 2– + SO3 –· SO4 –· + SO3 2– –→ SO4 2– + SO3 –· SO5 2– + SO3 2– –→ 2 SO4 2–

(79) (80) (81) (82) (83)

Deister and Warneck [129] reported a rate constant for reaction 82 of k = 5.5 × 108 M–1 s–1 while the branching ratio k81 /k80 = 0.41. The thermal oxidation step (reaction 83) involves reaction of peroxymonosulfate, SO5 2– , with sulﬁte; at pH 7–8, k83 = 350 M–1 s–1 . The catalytic propagation cycle, which has been reported to exceed 50 000 (i.e., φ = 50 000) [116, 117, 119–122], is most often terminated by the self-reaction of two sulﬁte radical anions. SO3 –· + SO3 –· –→ S2 O6 2–

(84)

Homogeneous and Heterogeneous Photochemistry in the Troposphere

103

In a subsequent study, Fischer and Warneck [130] re-examined the photolysis of sulﬁte along with bisulﬁte, HSO3 – , over a broader pH than used by Deister and Warneck [129]. Using Zn arc and Hg vapor lamps, they determined again that the principal reaction products in both species of S(IV) were sulfate and dithionate with measured quantum yields of φ = 0.19 for HSO3 – and φ = 0.39 for SO3 2– . In the presence of N2 O as a scavenger for H atom and aquated electrons, the quantum yields rose to 0.25 and 0.75, respectively, due to suppression of the sulﬁte radical termination step (reaction 84). In the presence of oxygen, the photolysis of HSO3 – led to a shorter chain reaction with sulfate and peroxydisulfate, S2 O8 2– , as observed products. SO4 –· + SO4 –· –→ S2 O8 2–

(85)

In the case of bisulﬁte, the primary propagation reaction involved the peroxymonosulfate radical anion reaction with HSO3 – as follows: SO5 –· + HSO3 – –→ HSO5 – + SO3 –·

(86)

while the major termination step appeared to involve the reaction of the hydroperoxyl radical with SO5 –· , where k87 = 5.5 × 108 M–1 s–1 . SO5 –· + HO2 · –→ SO5 2– + H+ + O2

(87)

Hoffmann and co-workers [131] investigated the kinetics and mechanism of the photo-assisted autoxidation of S(IV) ([S(IV)] ≡ [SO2 · H2 O] + [HSO3 – ] + [SO3 2– ]) over the pH range of 2 to 11 as catalyzed by aqueous colloidal suspensions of nanoparticulate α-Fe2 O3 . O2 + 2 HSO3 –

hν ≤ 520 nm α–Fe2 O3

→ 2 SO4 2– + 2H+

(88)

Quantum yields ranged from 0.08 to 0.3 with a maximum yield found at pH 5.7. The primary initiation pathway involved irradiation at wavelengths equal to or less than the nominal band-gap of hematite which is 2.2 eV or 560 nm. Upon band-gap illumination, conduction-band electrons and valence-band holes are separated; the trapped electrons are transferred either to surface bound dioxygen or to Fe(III) sites on or near the surface while the trapped holes accept electrons from adsorbed SO3 2– to produce surfacebound SO3 –· . The relatively high quantum yields were attributed in part to the desorption of SO3 –· from the α-Fe2 O3 surface and subsequent initiation of a homogeneous aqueous-phase free-radical chain oxidation of S(IV) to S(VI) according to reactions Eqs. 79–84. The following photochemical rate expression was found to describe the observed kinetics over a broad range of conditions. Ks [HSO3 – ] d[S(IV)] = φI0 (1 – 10–ε[α–Fe2 O3 ] ) – (89) dt 1 + Ks [HSO3 – ]

104

M.R. Hoffmann

A similar kinetic expression was found by Hong et al. [132] for the catalytic, photochemical oxidation of S(IV) on TiO2 . In this case, for λ ≤ 385 nm, quantum yields in excess of unity (e.g., 0.5 ≤ φ ≤ 300) were observed and attributed also to desorption of SO3 –· from the TiO2 surface leading to the initiation of homogeneous free-radical chain reactions (i.e., reactions 79 to 84). The observed quantum yields, which ranged between 0.5 and 300, depended on the concentration and nature of free-radical inhibitors present in the heterogeneous suspension. Ansari et al. [133] also studied the effect of light on the catalysis of the oxidation of S(IV) as HSO3 – to S(IV) in the presence of α-Fe2 O3 , α-FeOOH, β-FeOOH and γ -FeOOH. They noted that the rate of HSO3 – oxidation increased signiﬁcantly under irradiation and that in this case, the rate of oxidation of HSO3 – by the four iron(III) oxides follows a kinetic equation of the type, where x is the fraction of HSO3 – oxidized at time t. The rate constant (for an empirical rate law of the form – dXS(IV)/dt = k(1 – XS(IV))/XS(IV) where XS(IV) was the fraction of bisulﬁte remaining at time t) for HSO3 – photo-oxidation per unit of speciﬁc area followed the relative order of αFe2 O3 γ -FeOOH ≈ β-FeOOH α-FeOOH. In addition, Ansari et al. reported that Fe(II) and Fe(III) were detected in solution after irradiation of the iron(III) oxide suspensions indicating that photochemical reductive dissolution as described previously by Faust and Hoffmann [134] had taken place. Leland and Bard [135] investigated the photochemical reactivity of a wider range iron(III) oxide polymorphs (i.e., α-Fe2O3 – Eg = 2.02 eV, γ -FeOOH – Eg = 2.06 eV, β-FeOOH – Eg = 2.12 eV, α-FeOOH – Eg = 2.10 eV, δ-FeOOH – Eg = 1.94 eV, γ -Fe2 O3 – Eg = 2.03 eV) toward the oxidation of sulﬁte and oxalate. Leland and Bard found a larger range of relatively photochemical reactivity than Ansari et al. Their observed that the order of relative photochemical reactivity toward sulﬁte oxidation was γ -FeOOH > α-Fe2 O3 > γ Fe2 O3 > δ-FeOOH > β-FeOOH > α-FeOOH, while in the case of oxalate oxidation the relative order was γ -Fe2 O3 > γ -FeOOH >-Fe2 O3 > α-Fe2 O3 > αFeOOH > β-FeOOH. The nominal process involved in the photoexcitation of the Fe(III) oxides is band-gap excitation to produce transient valence band holes and conduction band electrons (Fig. 8) as follows: α – Fe2 O3

hν

→ h+vb + e–cb

(90)

Photoexcitation is quickly followed by either deep or surﬁcial trapping of the hole and electron or direct interfacial electron transfer. However, in the case of the iron(III) oxides the electron is often trapped by lattice Fe(III) resulting in its reduction [134] to Fe(II), which is then readily released to the surrounding aqueous solution. The trapped hole is then used to oxidize appropriate electron

Homogeneous and Heterogeneous Photochemistry in the Troposphere

105

donors such as low molecular weight organic acids such as oxalate [136]. e–cb + Fe(III)Fe(III)O3 –→ Fe(II)Fe(III)O3 – h+vb + Fe(III)Fe(III)O3 –→ Fe(III)Fe(III)O2 O+· Fe(II)Fe(III)O3 – + 6 H2 O –→ Fe(II)(H2 O)6 2+ + Fe(III)O3 3– Fe(III)Fe(III)O2 O+· + C2 O4 2– –→ C2 O4 –· + Fe(III)Fe(III)O3

(91) (92)

Fig. 8 Molecular orbital depiction of the concept of band-gap energies with corresponding molecular orbital transitions for the Fe(III) oxyhydroxides. The photon action spectra [134, 230] for photochemical reactions [136, 141, 143] of the iron oxyhydroxides (i.e., α-Fe2 O3 , α-FeOOH, β-FeOOH and γ -FeOOH) indicate that the most effective electron transition leading to photocatalysis or photoreduction is the O2– to Fe3+ transition shown schematically above

106

M.R. Hoffmann

4.3 Iron Oxides and Dissolved Iron Species as Chromophoric Reaction Initiators As noted above the polymorphic Fe(III) oxides are potential photochemical catalysts in environmental systems. Many investigations of the photochemical reactivity of these naturally occurring compounds have been carried out over the last 15 years. Hoffmann et al. [137], Waite [138] and Parmon and Zakharenko [139] have provided reviews of these processes. In addition, Hoffmann and co-workers [136, 140–149] have investigated the chemistry and photochemistry of Fe(III) and Fe(II) compounds and aqueous-phase species in relation to chemical transformations in clouds, fogs, and aquated haze aerosols. Siefert et al. [140] developed a procedure to determine the concentration of photochemically reactive Fe in ambient aerosol samples. The collected ambient aerosol samples were suspended in an aqueous solution within a photochemical reactor and irradiated. Under these conditions, which were favorable to the photochemical weathering of aerosol particles, the relative amount of dissolved aqueous Fe(II)(aq) to Fe-total was shown to increase. The extent and rate of Fe(II)(aq) photoproduction was used to characterize the Fe in aerosol samples collected from a variety of locations; photochemically available Fe ranged from 0.07 to 5.52 nmole m–3 while the total Fe concentrations in the aerosol samples ranged from 0.18 to 61 nmole m–3 , and the percentage of photochemically available Fe to Fe-total ranged from 2.8–100%. Calculations based on the experimental observations indicate that the photochemically driven redox reactions of Fe in cloudwater could be an important in situ source of oxidants such as hydroxyl radical, hydroperoxyl radical, and hydrogen peroxide. The estimated oxidant production rate in cloudwater based on these experiments was as high as 60 nM s–1 , with an average value of 16 nM s–1 . This estimated in situ oxidant production rate due to Fe chemistry is approximately equal to previous estimates of the total oxidant ﬂux to cloudwater from the gas phase. In related studies, Hoffmann and co-workers [142, 143, 150–153] examined the rates of photooxidation of selected organic compounds and the production of hydrogen peroxide on a variety of metal oxide semiconductors. In general, electron transfer occurs from the conduction band to dioxygen adsorbed on the surface of the excited state metal oxide as follows: (93) 2 e– + O2 –→ O2 –· cb

pKa =4.8 O2 + H+ ←→ HO2 · 2 HO2 · –→ H2 O2 + O2 –·

Hoffman et al. investigated the photochemical production of H2 O2 on irradiated nanoparticulate ZnO colloids over the wavelength range of 320 ≤ λ ≤ 370 nm in the presence of carboxylic acids and oxygen. Steady-state concentrations up to 2 mM H2 O2 were formed. Maximum H2 O2 concentrations are

Homogeneous and Heterogeneous Photochemistry in the Troposphere

107

obtained only with added electron donors (i.e., valence band hole scavengers). The order of photochemical efﬁciency for H2 O2 production with carboxylic acid hole scavengers was HCO2 – > C2 O4 2– > CH3 CO2 – > citrate. Isotopic labeling of the electron acceptor, O2 , with 18 O veriﬁed that H2 O2 was produced directly by the reduction of adsorbed oxygen by conduction band electrons. Quantum yields were as high as 30% for H2 O2 production at low photon ﬂuxes. At the same time, the quantum yield was shown √to vary –1 with the inverse square root of absorbed light intensity [i.e., φ ∝ Iabs ], with the wavelength –1 of excitation φ ∝ λ , and with the diameter of the Q-sized colloids (i.e., φ ∝ D–1 p ). For example, d[H2 O2 ]/dt is 100–1000 times faster on Q-sized ZnO particles (Dp = 4–5 nm) than with bulk-phase ZnO particles (Dp = 100 nm). As noted above, in situ production of H2 O2 within illuminated clouds and aquated aerosols is most likely a very important process governing the fate of oxidizable inorganic and organic compounds. Hydrogen peroxide is a potent thermal oxidant in its own right (i.e., E0H = + 1.76 V; H2 O2 + 2 e– + 2 H+ 2 H2 O) and a subsequent photochemical source of hydroxyl radical in the condensed or liquid phase. H2 O2

λ ≤ 360 nm

→ 2· OH

(94)

Siefert et al. [136] simulated the chemical conditions of cloudwater using ambient aerosol samples suspended in an aqueous solution. Electron donors that are known to exist in atmospheric cloudwater (oxalate, formate, and acetate) were then added to the simulated cloudwater, and the solution irradiated with UV light at λ ≥ 300 nm. In all cases, H2 O2 and Fe(II)aq were produced as a function of irradiation time. In addition, H2 O2 was also produced without added electron donors simply using ambient aerosols collected from four different sites around the US. In addition, the production of Fe(II)aq showed that Fe from the ambient aerosol was available for photochemical redox reactions. In addition, the simultaneous release of Fe(II) and hydrogen peroxide will result in the indirect photochemical production of hydroxyl radical as follows: (95) H O + Fe(II)(H O) 2+ –→ Fe(III)(H O) 3+ +· OH +– OH 2 2

2

6

2

6

Depending on the pH of the ambient cloudwater, the Fe(III) hexaquo species of reaction 95 will hydrolyze to produce Fe(H2 O)5 OH2+ which is also photochemically active. In this regard Faust and co-workers [154, 155] have investigated the photolysis of Fe(H2 O)5 OH2+ as another potential source of hydroxyl radical in the condensed phases in the atmosphere. Fe(III)(H2 O)5 OH2+ + H2 O

λ ≤ 340 nm

→ Fe(II)(H2 O)6 2+ + · OH

(96)

Joseph et al. [156] determined that the quantum yields, φ, for · OH production via reaction 96 ranged from 0.018 to 0.025 and from 0.014 to 0.018 with UV light and natural sunlight, respectively.

108

M.R. Hoffmann

In an earlier study, Benkelberg and Warneck [157] measured the quantum yields for the production of · OH radical from Fe(H2 O)5 OH2+ and other Fe(III) hydroxyl species in aqueous solution over the wavelength range of 280 to 370 nm at pH 2 and 3. The · OH formation quantum yields from Fe(H2 O)5 OH2+ ranged from φ = 0.07 at 370 nm to φ = 0.31 at 280 nm. Below 300 nm · OH radical production from the photolysis of Fe(III)(H2 O)6 3+ was estimated to have a quantum yield of 0.05. The same authors [157] also determined the quantum yields for the formation of sulfate radical anion, SO4 –· from the photolysis of FeSO4 + complex at pH 2. Fe(III)(H2 O)5 SO4 + + H2 O

λ ≤ 350 nm

→ Fe(II)(H2 O)6 2+ + SO4 –·

(97)

The quantum yields for SO4 –· formation via reaction 97 varied from φ = 0.0013 at 350 nm to φ = 0.0079 at 280 nm. Table 13 Band-gap energies and activation wavelengths for metal oxide and metal sulﬁde semiconductors [137] Compound

Band-gap (eV)

Wavelength (nm)∗

CuO CdO Fe2 O3

1.70 2.10 2.20

729 590 564

TiO2 - Rutile TiO2 - Anatase ZnO MgO Al2 O3

3.00 3.20 3.20 7.20 9.00

564 387 387 172 138

eV = 4.42 × 1014 Hz (s–1 ); λ =

hc ; Eg (eV)ν(Hz eV–1 )

C = 299 792 458 m s–1 ; h = 4.14 × 10–15

4.4 Photolysis of Hydrogen Peroxide – A Direct Source of Hydroxyl Radical Although the gas phase provides major pathway for hydroxyl radical and hydrogen peroxide production in the atmosphere, there is overwhelming evidence [158–168] that aqueous phases in the troposphere also provides a signiﬁcant medium for the photolytic production of these important oxidants. H2 O2

λ ≤ 360 nm

→ 2 · OH

Yu and Barker [169] used chloride ion as an efﬁcient scavenger of hydroxyl radicals to determine the quantum yield of · OH produced from the photolysis H2 O2 in acidic solution and determined that φ = 1 at 248 and 0.8 at

Homogeneous and Heterogeneous Photochemistry in the Troposphere

109

λ = 308 nm. In comparison, the gas-phase quantum yields for the above reaction are φ gas = 2.0 as predicted by the stoichiometry. The lower quantum yields in the condensed aqueous phase are due to geminate recombination in the solvent cage. The UV absorption spectrum [170–172] for H2 O2 in water is a continuum beginning at 380 nm and rising steadily down to 200 nm although the molar extinction coefﬁcients are quite low between 290 and 380 nm. 4.5 Nitrate and Nitrite Photolysis in Atmospheric Water and Ice Nitrate and nitrite, NO3 – and NO2 – respectively, absorb light over the wavelength range of 200 to 400 nm and undergo photolytic decomposition to yield a variety of products [173–176]. The λ > 300 nm photolysis of nitrate in aerated aqueous solutions at pH < 6 proceeds via two predominant pathways: NO3 – + H+ NO3 –

hν

→ NO2 +· OH

hν

→ NO2 – +· O(3 P)

(98) (99)

Peroxynitrous acid, ONOOH, forms in another photochemical channel at shorter wavelengths but is absent at λ > 300 nm. The O-atoms generated in reaction 99 may react with O2 ([O2 ]water ∼ 0.3 mM) via reaction 100 or, preferably, with nitrate via reaction 101. Nitrite (εmax = 22.5 M–1 cm–1 at 360 nm) will undergo secondary photolysis, reaction 102, and oxidation by OH radicals, reaction 103: O2 + O(3 P)

k100

NO3 – + O(3 P) NO2 –

→ O3 k100 = 4 × 109 M–1 s–1 k101

→ NO2 – + O2 k101 = 2 × 108 M–1 s–1

hν

→ NO2 +· OH

NO2 – +· OH

k101

→ NO2 +– OH k103 = 2 × 1010 M–1 s–1

(100) (101) (102) (103)

The excitation of the NO3 – n → π ∗ band in aqueous media, that has a maximum decadic absorption coefﬁcient of εNO3 – = 7.5 M–1 cm–1 at 305 nm, leads to the formation of OH and O(3 P) in relatively low quantum yields: φ98 ∼ 9 × 10–3 , φ98 /φ99 ∼ 9. φ98 increases at shorter and longer wavelengths within the n → π ∗ band, and increases by a factor of 2.2 between 273 K and 308 K. Fischer and Warneck [175] investigated the photolysis of nitrite over the wavelength range of 280 and 390 nm and determined that the quantum yield for hydroxyl radical formation φOH was decreased with increasing wavelength from 0.069 at 280 nm to 0.022 at 390 nm, in agreement with previous

110

M.R. Hoffmann

data [174]. The · OH quantum yield via the photolysis of undissociated nitrous acid, HNO2 , at pH 2 is φOH = 0.35. Zuo and Deng [177] measured molar extinction coefﬁcients of nitrite ion in water between 290 and 400 nm. The maximum extinction coefﬁcient at 354 nm was 22.9 M cm–1 . Using the absorption data over the entire wavelength range and the corresponding quantum yields, they estimated the ﬁrstorder rate constant of 1.0 × 10–4 s–1 for the formation of OH radicals in cloud water via nitrite photolysis at noontime at 40 degrees N in the summer. Hoffmann and co-workers [178, 179] have studied the photolysis of nitrate and nitrite in both water and water ice over broad temperature ranges. In the case of nitrite formation from nitrate photolysis, the photochemical reaction reaches a steady state condition that can be described by the following rate expression: d[NO2 – ] = φNO3 – Iabs,NO3 – + k101 [O(3 P)]ss [NO3 – ] dt – φNO2 – Iabs,NO2 – – k103 [· OH]ss [NO2 – ]

(104)

where [O]ss and [OH]ss are the photo-stationary state concentrations of oxygen atom and hydroxyl radical, respectively. Since the k101 term is quite small under most circumstances the rate of nitrite formation from the photolysis of nitrate can be reduced to the following simpliﬁed equation: d[NO2 – ] = φNO3 – Ia,NO3 – – kdecay [NO2 – ] (105) dt where the last two terms of Eq. 104 are lumped together as a single ﬁrst-order decay term. Integration of Eq. 105 yields φ × Ia,NO3 – [1 – exp(– kdecay t)] (106) [NO2 – ] = kdecay The latter equation predicts the concentration vs. time proﬁle for the attainment of the photo-stationary state for nitrite, since nitrite is both produced and consumed during the photolytic process. On the other hand, the initial loss of nitrate vs. time can be described by the basic ﬁrst-order photolysis equation. – d[NO3 – ] = φNO3 – (λ)I0 (λ)(1 – 10–ε [NO3 ] ) (107) dt – where the term I0 (λ)(1 – 10–ε [NO3 ] ) gives the speciﬁc molar rate of light absorption (i.e., moles [180] L–1 s–1 ).

–

–

Iabs = I0 (λ)(1 – 10–ε [NO3 ] )

(108)

Homogeneous and Heterogeneous Photochemistry in the Troposphere

111

4.6 Organic Compound Photochemistry in Clouds and Aerosols A vast array of organic compounds [181–208] is present in all phases of the atmosphere (i.e., gas, liquid, and solid). Many of these compounds are photochemically active. In addition, it has been recently recognized that a large fraction of the total organic carbon present in the atmosphere has chemical, spectral and photochemical properties similar to humic and fulvic acids found in soils and surface waters [209–221]. In this regard, Faust and co-workers [159, 222] have studied the photochemistry of selected organic compounds in simulated cloudwater. For example, they showed that the aqueous-phase photolysis of biacetyl produces organic acids and peroxides in aqueous aerosols, and fog and cloud drops. The major products of aqueous biacetyl photolysis were acetic acid, peroxyacetic acid, and hydrogen peroxide. Pyruvic acid and methylhydroperoxide were found as minor photoproducts. Common atmospheric reductants such as formate, formaldehyde, glyoxal, and phenol, were found to increase the quantum yields of peroxyacetic acid. Formate also signiﬁcantly increased the quantum yields of hydrogen peroxide. In a related study, Anastasio et al. [159] pointed out that non-phenolic aromatic carbonyl compounds (i.e., benzaldehydes and acetophenones) and various phenols are present in the atmosphere from the combustion of wood and other biomass and probably from the entrainment of terrestrial humic/fulvic substances present in wind-blown soil aerosol. As a consequence, they examined the photolysis of aqueous solutions nonphenolic aromatic carbonyls and phenols and showed that hydrogen peroxide is produced as a by-product of the organic compound photo-oxidation. The quantum yield for H2 O2 was shown to rapidly increase with increasing phenol concentration and decreasing pH. Given the combination of the relatively high photoreactivity and atmospheric prevalence for these chromophores, Anastasio and co-workers suggest that carbonyl compounds, keto-carbonyl, and non-hydroxy aromatic compounds are important sources of H2 O2 in aqueous aerosols, fogs, and clouds. In subsequent work, Anastasio and co-workers [223, 224] investigated a range of possible photochemical transformation of organic compounds in fogs prevalent in the Central Valley of California and showed that hydroxyl radical and singlet molecular oxygen were formed in illuminated winter fog waters. Nitrite photolysis appeared to be the major source of · OH accounting for 47–100% the total hydroxyl radical formed during photolysis. In addition, they showed that · OH radical was a signiﬁcant sink for refractory compounds and a minor sink for reactive trace species. On the other hand, photochemically produced singlet oxygen, 1 O2 appeared be a minor sink for refractory compounds present in fog water but a major sink for electron-rich reactive trace species.

112

M.R. Hoffmann

4.7 Conclusions From the preceding discussions, it should be clear that photochemistry within all phases of the atmosphere is a major driver of chemical transformations in relatively short time scales. With increasing knowledge of the ever-widening array of chromophoric compounds emitted and produced in the atmosphere, there is deﬁnitely room for much more fundamental research into primary and secondary photochemical reactions of relevance. In particular, the role of humic-like substances in aerosol, cloud and ice phases needs to be studied.

References 1. Seinfeld JH, Pandis SN (1998) Atmospheric chemistry and physics. Wiley, New York 2. Finlayson-Pitts BJ, Pitts JN (2000) Chemistry of the upper and lower atmosphere. Academic, New York 3. Yung YL, Demore WB (1999) Photochemistry of planetary atmospheres. Oxford University Press, New York 4. Warneck P (2000) Chemistry of the natural atmosphere. Academic, New York 5. Atkinson R, Baulch DL, Cox RA, Hampson RF, Kerr JA, Rossi MJ, Troe J (2000) J Phys Chem Ref Data 29:167 6. Atkinson R, Baulch DL, Cox RA, Hampson RF, Kerr JA, Troe J (1989) J Phys Chem Ref Data 18:881 7. Atkinson R, Baulch DL, Cox RA, Hampson RF, Kerr JA, Troe J (1989) Int J Chem Kinet 21:115 8. Atkinson R, Baulch DL, Cox RA, Hampson RF, Kerr JA, Troe J (1992) J Phys Chem Ref Data 21:1125 9. Atkinson R, Baulch DL, Cox RA, Hampson RF, Kerr JA, Troe J (1992) Atmos Environ 26:1187 10. Atkinson R, Baulch DL, Cox RA, Hampson RF, Kerr JA, Rossi MJ, Troe J (1997) J Phys Chem Ref Data 26:1329 11. Atkinson R, Baulch DL, Cox RA, Hampson RF, Kerr JA, Rossi MJ, Troe J (1997) J Phys Chem Ref Data 26:521 12. Atkinson R, Baulch DL, Cox RA, Hampson RF, Kerr JA, Rossi MJ, Troe J (1999) J Phys Chem Ref Data 28:191 13. Matsumi Y, Comes FJ, Hancock G, Hofzumahaus A, Hynes AJ, Kawasaki M, Ravishankara AR (2002) J Geophys Res-Atmos 107:4024 14. Matsumi Y, Kawasaki M (2003) Chem Rev 103:4767 15. Shamsuddin SM, Inagaki Y, Matsumi Y, Kawasaki M (1994) Can J Chem Rev Can Chim 72:637 16. Takahashi K, Matsumi Y, Kawasaki M (1996) J Phys Chem 100:4084 17. Takahashi K, Kishigami M, Taniguchi N, Matsumi Y, Kawasaki M (1997) J Chem Phys 106:6390 18. Takahashi K, Taniguchi N, Matsumi Y, Kawasaki M, Ashfold MNR (1998) J Chem Phys 108:7161 19. Smith GD, Molina LT, Molina MJ (2000) J Phys Chem A 104:8916 20. Nishida S, Taketani F, Takahashi K, Matsumi Y (2004) J Phys Chem A 108:2710

Homogeneous and Heterogeneous Photochemistry in the Troposphere

113

21. Takahashi K, Nakayama T, Matsumi Y (2003) J Phys Chem A 107:9368 22. Reisz E, Schmidt W, Schuchmann HP, von Sonntag C (2003) Environ Sci Technol 37:1941 23. Takahashi K, Hayashi S, Matsumi Y, Taniguchi N, Hayashida S (2002) J Geophys ResAtmos 107:4440 24. Koh CI, Lee SJ, Kang JW (2001) Ozone-Sci Eng 23:245 25. Hancock G, Tyley PL (2001) Phys Chem Chem Phys 3:4984 26. Pﬁster G, Baumgartner D, Maderbacher R, Putz E (2000) Atmos Environ 34:4019 27. Green JG, Shi J, Barker JR (2000) J Phys Chem A 104:6218 28. O’Keeffe P, Ridley T, Lawley KP, Maier RRJ, Donovan RJ (1999) J Chem Phys 110:10803 29. Talukdar RK, Longfellow CA, Gilles MK, Ravishankara AR (1998) Geophys Res Lett 25:143 30. Talukdar RK, Gilles MK, BattinLeclerc F, Ravishankara AR, Fracheboud JM, Orlando JJ, Tyndall GS (1997) Geophys Res Lett 24:1091 31. Silvente E, Richter RC, Zheng M, Saltzman ES, Hynes AJ (1997) Chem Phys Lett 264:309 32. Penkett SA, Monks PS, Carpenter LJ, Clemitshaw KC, Ayers GP, Gillett RW, Galbally IE, Meyer CP (1997) J Geophys Res-Atmos 102:12805 33. Denzer W, Hancock G, de Moira JCP, Tyley PL (1997) Chem Phys Lett 280:496 34. Takahashi K, Kishigami M, Matsumi Y, Kawasaki M, OrrEwing AJ (1996) J Chem Phys 105:5290 35. Ball SM, Hancock G (1995) Geophys Res Lett 22:1213 36. Armerding W, Comes FJ, Schulke B (1995) J Phys Chem 99:3137 37. Benderskii AV, Wight CA (1994) Chem Phys 189:307 38. Cooper IA, Neill PJ, Wiesenfeld JR (1993) J Geophys Res-Atmos 98:12795 39. Ball SM, Hancock G, Murphy IJ, Rayner SP (1993) Geophys Res Lett 20:2063 40. Cobos C, Castellano E, Schumacher HJ (1983) J Photochem 21:291 41. Sparks RK, Carlson LR, Shobatake K, Kowalczyk ML, Lee YT (1980) J Chem Phys 72:1401 42. Brock JC, Watson RT (1980) Chem Phys 46:477 43. Brock JC, Watson RT (1980) Chem Phys Lett 71:371 44. Arnold I, Comes FJ (1980) Chem Phys 47:125 45. Arnold I, Comes FJ (1980) J Mol Struct 61:223 46. Kajimoto O, Cvetanovic RJ (1979) Int J Chem Kinet 11:605 47. Bahe FC, Marx WN, Schurath U (1979) Atmos Environ 13:1515 48. Fairchild PW, Lee EKC (1978) Chem Phys Lett 60:36 49. Amimoto ST, Force AP, Wiesenfeld JR (1978) Chem Phys Lett 60:40 50. Philen DL, Watson RT, Davis DD (1977) J Chem Phys 67:3316 51. Arnold I, Comes FJ, Moortgat GK (1977) Chem Phys 24:211 52. Arnold I, Comes FJ, Moortgat GK (1977) Ber Bunsen-Ges Phys Chem Chem Phys 81:1113 53. Martin D, Girman J, Johnston HS (1974) Abstr Pap Am Chem Soc:156 54. Washida N, Mori Y, Tanaka I (1971) J Chem Phys 54:1119 55. Webster H, Bair EJ (1970) J Chem Phys 53:4532 56. Shetter RE, Junkermann W, Swartz WH, Frost GJ, Crawford JH, Lefer BL, Barrick JD, Hall SR, Hofzumahaus A, Bais A, Calvert JG, Cantrell CA, Madronich S, Muller M, Kraus A, Monks PS, Edwards GD, McKenzie R, Johnston P, Schmitt R, Grifﬁoen E, Krol M, Kylling A, Dickerson RR, Lloyd SA, Martin T, Gardiner B, Mayer B, Pﬁster G, Roth EP, Koepke P, Ruggaber A, Schwander H, van Weele M (2003) J Geophys ResAtmos 108:8544

114

M.R. Hoffmann

57. 58. 59. 60. 61. 62.

Morrell C, Breheny C, Haverd V, Cawley A, Hancock G (2002) J Chem Phys 117:11121 Lefer BL, Hall SR, Cinquini L, Shetter RE (2001) Geophys Res Lett 28:3637 Kraus A, Rohrer F, Hofzumahaus A (2000) Geophys Res Lett 27:1115 Hofzumahaus A, Kraus A, Muller M (1999) Appl Optics 38:4443 Beine HJ, Dahlback A, Orbaek JB (1999) J Geophys Res-Atmos 104:16009 Shetter RE, McDaniel AH, Cantrell CA, Madronich S, Calvert JG (1992) J Geophys Res-Atmos 97:10349 Gaedtke H, Troe J (1970) Z Naturforsch A 25:789 Calvert JG, Steacie EWR (1951) J Chem Phys 19:176 Calvert JG, DemerjiaKl, McQuigg RD, Kerr JA (1972) Science 175:751 Badcock CC, Calvert JG, Rabe BR, Damon EK, Sidebottom HW (1972) J Am Chem Soc 94:19 Clark JH, Moore CB, Nogar NS (1978) J Chem Phys 68:1264 Horowitz A, Su F, Calvert JG (1978) Int J Chem Kinet 10:1099 Horowitz A, Calvert JG (1978) Int J Chem Kinet 10:805 Horowitz A, Calvert JG (1978) Int J Chem Kinet 10:713 Moortgat GK, Slemr F, Seiler W, Warneck P (1978) Chem Phys Lett 54:444 Moortgat GK, Warneck P (1979) J Chem Phys 70:3639 Smith GD, Molina LT, Molina MJ (2002) J Phys Chem A 106:1233 Su F, Calvert JG, Shaw JH (1979) J Phys Chem 83:3185 Su F, Calvert JG, Shaw JH, Niki H, Maker PD, Savage CM, Breitenbach LD (1979) Chem Phys Lett 65:221 Tang KY, Fairchild PW, Lee EKC (1979) J Phys Chem 83:569 Rogers JD (1990) J Phys Chem 94:4011 Gunz DW, Hoffmann MR (1990) Atmos Environ 24:1601 Reeves CE, Penkett SA (2003) Chem Rev 103:5199 Lee MH, Heikes BG, O’Sullivan DW (2000) Atmos Environ 34:3475 Jackson AV (1999) Crit Rev Environ Sci Technol 29:175 Brown SS, Wilson RW, Ravishankara AR (2000) J Phys Chem A 104:4976 Nicovich JM, Wine PH (1988) J Geophys Res-Atmos 93:2417 Klee S, Gericke KH, Golzenleuchter H, Comes FJ (1989) Chem Phys 139:415 Molina LT, Schinke SD, Molina MJ (1977) Geophys Res Lett 4:580 Vaghjiani GL, Turnipseed AA, Warren RF, Ravishankara AR (1992) J Chem Phys 96:5878 Vaghjiani GL, Ravishankara AR (1990) J Chem Phys 92:996 Docker MP, Hodgson A, Simons JP (1986) Faraday Discuss 82:25 Cai ZT, Zhang DH, Zhang JZH (1994) J Chem Phys 100:5631 Inagaki Y, Matsumi Y, Kawasaki M (1993) Bull Chem Soc Jpn 66:3166 Schiffman A, Nelson DD, Nesbitt DJ (1993) J Chem Phys 98:6935 Zhang DH, Zhang JZH (1993) J Chem Phys 98:6276 Brouard M, Martinez MT, Milne CJ, Simons JP, Wang JX (1990) Chem Phys Lett 165:423 Schinke R, Staemmler V (1988) Chem Phys Lett 145:486 Gericke KH, Klee S, Comes FJ, Dixon RN (1986) J Chem Phys 85:4463 Docker MP, Hodgson A, Simons JP (1986) Chem Phys Lett 128:264 Atkinson R, Baulch DL, Cox RA, Crowley JN, Hampson RF, Hynes RG, Jenkin ME, Rossi MJ, Troe J (2004) Atmos Chem Phys 4:1461 Stutz J, Kim ES, Platt U, Bruno P, Perrino C, Febo A (2000) J Geophys Res-Atmos 105:14585 Barney WS, Wingen LM, Lakin MJ, Brauers T, Stutz J, Finlayson-Pitts BJ (2000) J Phys Chem A 104:1692

63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99.

Homogeneous and Heterogeneous Photochemistry in the Troposphere 100. 101. 102. 103. 104. 105. 106.

107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143.

115

Brust AS, Becker KH, Kleffmann J, Wiesen P (2000) Atmos Environ 34:13 Pagsberg P, Bjergbakke E, Ratajczak E, Sillesen A (1997) Chem Phys Lett 272:383 Bongartz A, Kames J, Welter F, Schurath U (1991) J Phys Chem 95:1076 Stockwell WR, Calvert JG (1978) J Photochem 8:193 Calvert JG, Yarwood G, Dunker AM (1994) Res Chem Intermed 20:463 Cox RA, Derwent RG (1976) J Photochem 6:23 Sander SP, Friedl RR, Golden DM, Kurylo MJ, Huie RE, Orkin VL, Moortgat GK, Ravishankara AR, Kolb CE, Molina MJ, Finlayson-Pitts BJ (2004), vol Evaluation No 14 JPL/NASA, Pasadena, CA Alicke B, Geyer A, Hofzumahaus A, Holland F, Konrad S, Patz HW, Schafer J, Stutz J, Volz-Thomas A, Platt U (2003) J Geophys Res-Atmos 108:8247 Faust BC (1994) Environ Sci Technol 28:A217 Madronich S (1987) J Geophys Res-Atmos 92:9740 Mayer B, Kylling A, Madronich S, Seckmeyer G (1998) J Geophys Res-Atmos 103:31241 Penkett SA (1986) Nature 319:624 Penkett SA, Jones BMR, Brice KA, Eggleton AEJ (1979) Atmos Environ 13:123 McArdle JV, Hoffmann MR (1983) J Phys Chem 87:5425 Hoffmann MR, Edwards JO (1975) J Phys Chem 79:2096 Jacob DJ, Hoffmann MR (1983) J Geophys Res 88:6611 Backstrom HLJ (1934) Naturwissenschaften 22:170 Backstrom HLJ (1934) Z Physik Chem B25:122 Backstrom HLJ (1934) Z Physik Chem B25:99 Alyea HN, Backstrom HLJ (1929) J Am Chem Soc 51:90 Backstrom HLJ (1927) Medd Vetenskapsakad Nobelinst 6:57 Backstrom HLJ (1927) Medd Vetenskapsakad Nobelinst 6:34 Backstrom HLJ (1927) J Am Chem Soc 49:1460 Waygood SJ, McElroy WJ (1992) J Chem Soc-Faraday Trans 88:1525 Subhani MS, Kauser Z (1978) Revue Roumaine De Chimie 23:1129 Chawla OP, Arthur NL, FessendeRw (1973) J Phys Chem 77:772 Chawla OP, Arthur NL, Fessenden RW (1973) J Phys Chem 77:772 Hayon E, Treinin A, Wilf J (1972) J Am Chem Soc 94:47 Dogliotti L, Hayon E (1968) J Phys Chem 72:1800 Deister U, Warneck P (1990) J Phys Chem 94:2191 Fischer M, Warneck P (1996) J Phys Chem 100:15111 Faust BC, Hoffmann MR, Bahnemann DW (1989) J Phys Chem 93:6371 Hong AP, Bahnemann DW, Hoffmann MR (1987) J Phys Chem 91:6245 Ansari A, Peral J, Domenech X, Rodriquez-Clemente R (1997) Environ Pollut 95:283 Faust BC, Hoffmann MR (1986) Environ Sci Technol 20:943 Leland JK, Bard AJ (1987) J Phys Chem 91:5076 Siefert RL, Pehkonen SO, Erel Y, Hoffmann MR (1994) Geochim Cosmochim Acta 58:3271 Hoffmann MR, Martin ST, Choi WY, Bahnemann DW (1995) Chem Rev 95:69 Waite TD (1990) Rev Miner 23:559 Parmon VN, Zakharenko VS (2001) Cattech 5:96 Siefert RL, Webb SM, Hoffmann MR (1996) J Geophys Res-Atmos 101:14441 Pehkonen SO, Siefert RL, Hoffmann MR (1995) Environ Sci Technol 29:1215 Erel Y, Pehkonen SO, Hoffmann MR (1993) J Geophys Res-Atmos 98:18423 Pehkonen SO, Siefert R, Erel Y, Webb S, Hoffmann MR (1993) Environ Sci Technol 27:2056

116

M.R. Hoffmann

144. 145. 146. 147. 148.

Johansen AM, Hoffmann MR (2003) J Geophys Res-Atmos 108 Johansen AM, Siefert RL, Hoffmann MR (2000) J Geophys Res-Atmos 105:15277 Johansen AM, Siefert RL, Hoffmann MR (1999) J Geophys Res-Atmos 104:26325 Siefert RL, Johansen AM, Hoffmann MR (1999) J Geophys Res-Atmos 104:3511 Siefert RL, Johansen AM, Hoffmann MR, Pehkonen SO (1998) J Air Waste Manage Assoc 48:128 Pehkonen SO, Erel Y, Hoffmann MR (1992) Environ Sci Technol 26:1731 Hoffman AJ, Carraway ER, Hoffmann MR (1994) Environ Sci Technol 28:776 Kormann C, Bahnemann DW, Hoffmann MR (1988) Environ Sci Technol 22:798 Hong AP, Bahnemann DW, Hoffmann MR (1987) J Phys Chem 91:2109 Kormann C, Bahnemann DW, Hoffmann MR (1989) J Photochem Photobiol, A-Chem 48:161 Zepp RG, Faust BC, Hoigne J (1992) Environ Sci Technol 26:313 Faust BC, Hoigne J (1990) Atmos Environ 24:79 Joseph JM, Varghese R, Aravindakumar CT (2001) J Photochem Photobiol A-Chem 146:67 Benkelberg HJ, Warneck P (1995) J Phys Chem 99:5214 Anastasio C, Faust BC, Allen JM (1994) J Geophys Res-Atmos 99:8231 Anastasio C, Faust BC, Rao CJ (1997) Environ Sci Technol 31:218 Arakaki T, Anastasio C, Shu PG, Faust BC (1995) Atmos Environ 29:1697 Arakaki T, Faust BC (1998) J Geophys Res-Atmos 103:3487 Faust BC, Allen JM (1993) Environ Sci Technol 27:1221 Faust BC, Anastasio C, Allen JM, Arakaki T (1993) Science 260:73 Sedlak DL, Hoigne J, David MM, Colvile RN, Seyffer E, Acker K, Wiepercht W, Lind JA, Fuzzi S (1997) Atmos Environ 31:2515 Zuo YG, Hoigne J (1993) Science 260:71 Zuo YG, Holgne J (1992) Environ Sci Technol 26:1014 Zuo YG, Deng YW (1999) Geochim Cosmochim Acta 63:3451 Anastasio C, Jordan AL (2004) Atmos Environ 38:1153 Yu XY, Barker JR (2003) J Phys Chem A 107:1325 Phibbs MK, Giguere PA (1951) Can J Chem-Rev Can Chim 29:490 Taylor RC, Cross PC (1949) J Am Chem Soc 71:2266 Holt RB, McLane CK, Oldenberg O (1948) J Chem Phys 16:225 Mack J, Bolton JR (1999) J Photochem Photobiol A-Chem 128:1 Zellner R, Exner M, Herrmann H (1990) J Atmos Chem 10:411 Fischer M, Warneck P (1996) J Phys Chem 100:18749 Mark G, Korth HG, Schuchmann HP, von Sonntag C (1996) J Photochem Photobiol A-Chem 101:89 Zuo YG, Deng YW (1998) Chemosphere 36:181 Dubowski Y, Colussi AJ, Hoffmann MR (2001) J Phys Chem A 105:4928 Dubowski Y, Colussi AJ, Boxe C, Hoffmann MR (2002) J Phys Chem A 106:6967 Adams SF, DeJoseph CA, Carter CC, Miller TA, Williamson JM (2001) J Phys Chem A 105:5977 Fine PM, Cass GR, Simoneit BRT (2004) Environ Eng Sci 21:705 Fine PM, Cass GR, Simoneit BRT (2004) Environ Eng Sci 21:387 Fine PM, Cass GR, Simoneit BRT (2002) J Geophys Res-Atmos 107 Nolte CG, Schauer JJ, Cass GR, Simoneit BRT (2002) Environ Sci Technol 36:4273 Fine PM, Cass GR, Simoneit BRT (2002) Environ Sci Technol 36:1442 Schauer JJ, Kleeman MJ, Cass GR, Simoneit BRT (2002) Environ Sci Technol 36:1169 Schauer JJ, Kleeman MJ, Cass GR, Simoneit BRT (2002) Environ Sci Technol 36:567

149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187.

Homogeneous and Heterogeneous Photochemistry in the Troposphere 188. 189. 190. 191. 192. 193. 194. 195. 196. 197. 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224.

117

Fine PM, Cass GR, Simoneit BRT (2001) Environ Sci Technol 35:2665 Nolte CG, Schauer JJ, Cass GR, Simoneit BRT (2001) Environ Sci Technol 35:1912 Schauer JJ, Kleeman MJ, Cass GR, Simoneit BRT (2001) Environ Sci Technol 35:1716 Cass GR, Hughes LA, Bhave P, Kleeman MJ, Allen JO, Salmon LG (2000) Philos Trans R Soc Lond Ser A-Math Phys Eng Sci 358:2581 Nolte CG, Schauer JJ, Cass GR, Simoneit BRT (1999) Environ Sci Technol 33:3313 Fine PM, Cass GR, Simoneit BRT (1999) Environ Sci Technol 33:2352 Fraser MP, Cass GR, Simoneit BRT (1999) Atmos Environ 33:2715 Schauer JJ, Kleeman MJ, Cass GR, Simoneit BRT (1999) Environ Sci Technol 33:1566 Schauer JJ, Kleeman MJ, Cass GR, Simoneit BRT (1999) Environ Sci Technol 33:1578 Nolte CG, Fraser MP, Cass GR (1999) Environ Sci Technol 33:540 Simoneit BRT, Schauer JJ, Nolte CG, Oros DR, Elias VO, Fraser MP, Rogge WF, Cass GR (1999) Atmos Environ 33:173 Saxena P, Hildemann LM (1997) Environ Sci Technol 31:3318 Rogge WF, Hildemann LM, Mazurek MA, Cass GR, Simoneit BRT (1997) Environ Sci Technol 31:2726 Rogge WF, Hildemann LM, Mazurek MA, Cass GR, Simoneit BRT (1997) Environ Sci Technol 31:2731 Saxena P, Hildemann LM (1996) J Atmos Chem 24:57 Hildemann LM, Klinedinst DB, Klouda GA, Currie LA, Cass GR (1994) Environ Sci Technol 28:1565 Rogge WF, Hildemann LM, Mazurek MA, Cass GR (1994) Environ Sci Technol 28:1375 Hildemann LM, Mazurek MA, Cass GR, Simoneit BRT (1994) Aerosol Sci Technol 20:303 Rogge WF, Hildemann LM, Mazurek MA, Cass GR, Simoneit BRT (1993) Environ Sci Technol 27:2700 Hildemann LM, Mazurek MA, Cass GR, Simoneit BRT (1991) Environ Sci Technol 25:1311 Rogge WF, Hildemann LM, Mazurek MA, Cass GR, Simonelt BRT (1991) Environ Sci Technol 25:1112 Tuckermann R, Cammenga HK (2004) Atmos Environ 38:6135 Brooks SD, DeMott PJ, Kreidenweis SM (2004) Atmos Environ 38:1859 Hoffer A, Kiss G, Blazso M, Gelencser A (2004) Geophys Res Lett 31 Gysel M, Weingartner E, Nyeki S, Paulsen D, Baltensperger U, Galambos I, Kiss G (2004) Atmos Chem Phys 4:35 Limbeck A, Kulmala M, Puxbaum H (2003) Geophys Res Lett 30 Cappiello A, De Simoni E, Fiorucci C, Mangani F, Palma P, Trufelli H, Decesari S, Facchini MC, Fuzzi S (2003) Environ Sci Technol 37:1229 Andracchio A, Cavicchi C, Tonelli D, Zappoli S (2002) Atmos Environ 36:5097 Zhang Q, Anastasio C (2001) Atmos Environ 35:5629 Kiss G, Varga B, Gelencser A, Krivacsy Z, Molnar A, Alsberg T, Persson L, Hansson HC, Facchini MC (2001) Atmos Environ 35:2193 Subbalakshmi Y, Patti AF, Lee GSH, Hooper MA (2000) J Environ Monit 2:561 Gelencser A, Sallai M, Krivacsy Z, Kiss G, Meszaros E (2000) Atmos Res 54:157 Capel PD, Leuenberger C, Giger W (1991) Atmos Environ 25:1335 Capel PD, Gunde R, Zurcher F, Giger W (1990) Environ Sci Technol 24:722 Faust BC, Powell K, Rao CJ, Anastasio C (1997) Atmos Environ 31:497 Anastasio C, McGregor KG (2001) Atmos Environ 35:1079 McGregor KG, Anastasio C (2001) Atmos Environ 35:1091

118 225. 226. 227. 228. 229. 230.

M.R. Hoffmann Greenblatt GD, Wiesenfeld JR (1983) J Chem Phys 78:4924 Amimoto ST, Force AP, Wiesenfeld JR, Young RH (1980) J Chem Phys 73:1244 Ball SM, Hancock G, Martin SE, deMoira JCP (1997) Chem Phys Lett 264:531 Bauer D, D’Ottone L, Hynes AJ (2000) Phys Chem Chem Phys 2:1421 Trolier M, Wiesenfeld JR (1988) J Geophys Res-Atmos 93:7119 Faust BC, Hoffmann MR, Bahnemann DW (1989) J Phys Chem 93:6371

Hdb Env Chem Vol. 2, Part M (2005): 119–160 DOI 10.1007/b138182 © Springer-Verlag Berlin Heidelberg 2005 Published online: 6 September 2005

Atmospheric Photooxidation of Gas Phase Air Pollutants T. J. Wallington1 (u) · O. J. Nielsen2 1 Ford

Motor Company, Research and Advanced Engineering, Mail Drop SRL-3083, Dearborn, Michigan 48121-2053, USA [email protected]

2 Department

of Chemistry, University of Copenhagen, Universitetsparken 5, DK-2100 Copenhagen, Denmark [email protected]

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

120

2

General Atmospheric Chemistry . . . . . . . . . . . . . . . . . . . . . . . .

122

3

Atmospheric Lifetimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

126

4

Degradation of Alkanes . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

130

5

Degradation of Alkenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

136

6

Degradation of Aromatics . . . . . . . . . . . . . . . . . . . . . . . . . . .

140

7

Degradation of NOx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

143

8

Degradation of SOx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

145

9

Degradation of CFCs and Halons . . . . . . . . . . . . . . . . . . . . . . .

146

10 10.1 10.2 10.3

Degradation of HFCs, HCFCs, and HFEs . . . . . Gas-Phase Chemistry . . . . . . . . . . . . . . . . Reactions of Halogenated Carbonyl Intermediates Heterogeneous and Aqueous Phase Chemistry . .

. . . .

150 151 153 155

11

Research Needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

156

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

157

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

Abstract An overview of the gas-phase reactions responsible for the photooxidation of air pollutants is provided. Starting with an introduction to the ﬁeld of atmospheric chemistry we proceed to discuss the concept, and utility, of atmospheric lifetime. The processes responsible for the degradation of alkanes, alkenes, aromatics, NOx , SOx , CFCs and their replacements are discussed in turn. Areas of uncertainty are highlighted. Finally, the research needs are discussed. Keywords Atmospheric chemistry · Air pollution · Hydrocarbons · NOx

120

T.J. Wallington · O.J. Nielsen

1 Introduction Human activities result in the release of a large quantity and variety of chemical compounds into the atmosphere. These compounds are degraded in a complex series of reactions in which the pollutants are oxidized in successive steps giving increasingly polar and less volatile products. Eventually the pollutant is either completely oxidized, e.g., the conversion of methane into CO2 and H2 O, or it is converted into partially oxidized species, which are removed via wet and/or dry deposition to the Earth’s surface. While the degradation reactions are beneﬁcial because they remove pollutants from the air, they can have unwanted side effects. The degradation products and intermediates can lead directly, or indirectly, to environmental impacts on local, regional, or global scales. Local scale air pollution is responsible for the photochemical smog present in many large-scale metropolitan areas. On a time scale of hours in the presence of sunlight, atmospheric chemical reactions convert reactive hydrocarbons and nitrogen oxides into a mixture of oxidants such as ozone and peroxyacetyl nitrate (PAN— CH3 C(O)O2 NO2 ) that is generically known as urban smog. Consequently, the deleterious effects of smog are experienced 10–100 km from the pollution sources. Regional scale problems are perhaps best exempliﬁed by the phenomenon of acid precipitation. The atmospheric oxidation of SO2 into sulfuric acid (H2 SO4 ) occurs on a time scale of several days and acidic precipitation occurs typically 500–1000 km downwind of the pollution source. Notable examples include the acidic precipitation experienced in Northern Scandinavia originating from emissions in central Europe, acidic precipitation across Japan resulting from emissions from the Asian mainland, and acidic rain in the North Eastern U.S. from emissions in the Ohio river valley. Stratospheric ozone depletion is an example of a global scale problem. Stratospheric ozone loss is caused by the release of chloroﬂuorocarbons (e.g., CF2 Cl2 ) and Halons (e.g., CF3 Br and CF2 ClBr). CFCs and Halons have no signiﬁcant loss mechanisms in the lower atmosphere and within several years of their release they become uniformly distributed in the lower atmosphere on a global scale. CFCs and Halons are transported into the stratosphere as part of the natural air circulation. In the stratosphere the CFCs encounter harsh solar UV irradiation (λ < 250 nm) which is blocked from the lower atmosphere by absorption by the ozone layer. UV irradiation of CFCs and Halons releases Cl and Br atoms which then participate in ozone destruction reactions. To assess the environmental impact of the release of a compound into the atmosphere there are several issues that need to be considered. The ﬁrst step is to determine its atmospheric lifetime. The atmospheric lifetime determines the geographical extent of the possible direct environmental impact. To calculate the atmospheric lifetime of a chemical compound we need information

Atmospheric Photooxidation of Gas Phase Air Pollutants

121

on the kinetics of its reaction with key atmospheric trace species such as OH and NO3 radicals and O3 , the rate of its photolysis, its solubility and hence propensity towards wet deposition, and ﬁnally its rate of dry deposition. If the compound is long lived (> 10 years) and released in substantial quantities it may contribute to global warming and it would be appropriate to measure its infrared spectrum and calculate its potential effect on the radiative balance of the atmosphere. Once the lifetime of the pollutant has been established the next step is to determine the degradation products and intermediates of its atmospheric oxidation and assess whether they pose any environmental threat. The recent search for environmentally acceptable CFC replacements provides a good example of how detailed studies of the atmospheric degradation mechanism of a class of chemical compounds lead to an understanding of their environmental impact. Hydroﬂuorocarbons (e.g., CF3 CFH2 , also known as HFC-134a) and hydrochloroﬂuorocarbons (e.g., CF3 CCl2 H, also known as HCFC-123) are important classes of chemical compounds which are used as replacements for CFCs in a variety of applications. When HFCs and HCFCs were ﬁrst proposed as CFC replacements very little was known about their atmospheric chemistry. Initially there was speculation that ﬂuorine-containing free radical species generated during the atmospheric degradation of HFCs and HCFCs (CF3 Ox , FCOx , and FOx ) might adversely impact stratospheric ozone [1] and that the atmospheric degradation of HFCs and HCFCs might form toxic products which could accumulate in the environment [2]. To assess the environmental acceptability of HFCs and HCFCs the chemical industry sponsored a number of experimental and theoretical studies of their atmospheric chemistry. As a result we now have a good understanding of the atmospheric degradation mechanisms of such compounds. In fact, it can be argued that of all the different classes of chemical compounds emitted into the atmosphere we understand the chemistry of HFCs and HCFCs the best! It is now clear that initial concerns regarding the environmental impact of such compounds were unfounded. The recent work establishing the atmospheric degradation mechanism of HFCs and HCFCs provides a notable example of the successful application of fundamental research to solve a practical environmental problem. In this chapter we will examine the atmospheric degradation mechanisms of the following important classes of anthropogenic molecules: alkanes, alkenes, aromatics, nitrogen oxides, SO2 , CFCs and Halons, and ﬁnally HFCs and HCFCs. Our intent is not to give an exhaustive account of the photochemical oxidation of every man-made chemical species but rather to present examples of the degradation mechanisms of a few representative members of each class of pollutant. First, we need to consider the general features of atmospheric chemistry.

122

T.J. Wallington · O.J. Nielsen

2 General Atmospheric Chemistry The atmosphere is a giant inhomogeneous photochemical reactor in which temperature, pressure, radiation ﬂux, and composition vary widely. All of the energy for atmospheric chemistry comes from solar radiation which is absorbed by various components of the atmosphere. Fig. 1 shows the direct solar ﬂux over the wavelength range 120–360 nm at the top of the atmosphere (labeled solar), and at 40, 20, and 0 km altitude. Molecular oxygen is responsible for the absorption of UV radiation of wavelengths less than 200 nm at altitudes above 40 km. Absorption by ozone in the stratosphere shields the Earth’s surface from UV radiation of wavelengths less than 300 nm. Gas-phase, solution-phase, and heterogeneous reactions all play important roles in atmospheric chemistry. The mean atmospheric composition is given in Table 1. N2 , O2 , and Ar comprise 99.9% of the atmosphere and, for all practical purposes, the relative proportion of these gases is constant in the lower 100 km of the atmosphere. We are concerned here with the fate of pollutants such as CO, volatile organic compounds, halocarbons, sulfur compounds, and nitrogen oxides, which are present in trace amounts and whose concentrations vary signiﬁcantly both spatially and temporally. In discussions of atmospheric processes it is useful to divide the atmosphere into different regions. Temperature proﬁles provide the most convenient basis for this division. Figure 2 shows the altitude proﬁle of atmospheric

Fig. 1 Direct solar ﬂux over the wavelength range 120–360 nm at the top of the atmosphere (labeled solar) and at 40, 20, and 0 km [93]

Atmospheric Photooxidation of Gas Phase Air Pollutants

123

Table 1 Average Composition of Dry Air Gas

N2 O2

Average mixing ratio %

(parts per million)

78.1 20.9

(781 000) (209 000)

Ar CO2 Ne He CH4

0.9 0.038 0.0018 0.00052 0.00017

(9000) (380) (18) (5.2) (1.7)

H2 N2 O CO O3

0.00006 0.00003 0.00001 0.000008

(0.6) (0.3) (0.1) (0.08)

Non-methane hydrocarbons Halocarbons Sulfur compounds Nitrogen oxides (NOy )

0.000001 0.0000002 0.00000008 0.00000005

(0.01) (0.002) (0.0008) (0.0005)

temperature and pressure up to 50 km [3]. The ﬁrst 10–15 km of the atmosphere is characterized by a temperature proﬁle in which colder air overlays warmer air. This situation is caused by the fact that the predominant heat source for this region of the atmosphere is the warm surface of the Earth. The temperature proﬁle of the troposphere is inherently unstable and results in strong vertical convective mixing from which the region derives its name (tropos is Greek for turning). Intense tropical thunderstorm systems can transport molecules from close to the Earth’s surface to the top of the troposphere within a few minutes. However, more typical mixing times are of the order of days to weeks. More than 90% of the mass of the atmosphere is located in the troposphere and it is here that the vast majority of anthropogenic molecules are degraded. In contrast to the troposphere, the stratosphere which lies at approximately 15–50 km altitude is heated principally from above by the absorption of solar UV radiation. In the stratosphere, warm air lies on top of cooler air which is an inherently stable situation resulting in a layered structure which gives the region its name (stratos is Latin for layered). Vertical mixing in the stratosphere proceeds slowly, typically on a time scale of several years. The region, which marks the boundary between the two different regions is called the tropopause. In this chapter we will be concerned mainly with the photochemistry that occurs in the troposphere.

124

T.J. Wallington · O.J. Nielsen

Fig. 2 Temperature and pressure proﬁle of the atmosphere up to 50 km taken from the US Standard Atmosphere [3]

The driving force for most of the chemistry that occurs in the atmosphere is the formation of hydroxyl (OH) radicals via photolysis of ozone to form O(1 D) atoms which react with water vapor. O3 + hν (λ < 320 nm) → O (1 D) + O2 (1 ∆g ) O (1 D) + H2 O → 2 OH

(1) (2)

The ﬂux of UV light, O3 , and H2 O vapor combine to give a potent source of OH radicals. OH radicals react with almost everything emitted into the atmosphere. The atmospheric lifetimes of many pollutants are determined by their reactivity towards OH radicals. The generation of OH radicals is the primary mechanism by which the atmosphere cleanses itself. Only compounds such as CFCs, Halons, and N2 O which are inert towards OH radical attack survive transport through the troposphere into the stratosphere where they can damage the ozone layer. The dominant loss of OH radicals is reaction with CO and organic compounds such as CH4 , both reactions produce peroxy radicals. Peroxy radicals play a key role in atmospheric chemistry. They are intimately involved in the formation and destruction of ozone and in the photooxidation of all organic compounds in the atmosphere [4]. The lifetime of OH radicals with respect to reactions Eq. 3 and Eq. 8 is of the order of a second and in the day-time a steady state condition is established. The OH radical concentration in the atmosphere varies with location, time of day, season, and meteo-

Atmospheric Photooxidation of Gas Phase Air Pollutants

125

rological conditions, the global 24 hour average is typically assumed to be 1.0×106 cm–3 [5–7]. Although only 10% of atmospheric ozone resides in the troposphere (0–15 km altitude) it has a profound impact on tropospheric chemistry. Ozone concentrations in the troposphere vary from typically 20–40 ppb for a remote pristine site to 100–200 ppb in a highly polluted urban environment. Ozone is a reactive molecule, which readily adds to carbon-carbon double bonds [8]. Reaction with ozone provides an important removal mechanism for many unsaturated reactive organic compounds. There are two sources of tropospheric ozone. First, transport from the stratosphere in meteorological events known as “tropospheric folding” in which a layer of stratospheric air is entrained in tropospheric air-ﬂow and mixed into the troposphere. Second, peroxy radical reactions which oxidize NO to NO2 . For example, in the OH radical initiated oxidation of CO: OH + CO → H + CO2 H + O2 + M → HO2 + M HO2 + NO → OH + NO2 NO2 + hν (λ < 420 nm) → NO + O O + O2 + M → O3 + M net: CO + 2O2 → CO2 + O3

(3) (4) (5) (6) (7)

And in the oxidation of CH4 : OH + CH4 → H2 O + CH3 CH3 + O2 + M → CH3 O2 + M CH3 O2 + NO → CH3 O + NO2 CH3 O + O2 → HCHO + HO2 HO2 + NO → OH + NO2 2 × (NO2 + hν (λ < 420 nm) → NO + O) 2 × (O + O2 + M → O3 + M) net: CH4 + 4O2 → H2 O + 2O3 + HCHO

(8) (9) (10) (11) (5) (6) (7)

The formaldehyde formed in the oxidation of CH4 can react with OH or photolyze leading to further NO2 formation and ozone production: HCHO + OH → H2 O + HCO HCHO + hν → H + HCO HCO + O2 → HO2 + CO H + O2 + M → HO2 + M 2 × (HO2 + NO → OH + NO2 )

(12) (13) (14) (15) (5)

126

T.J. Wallington · O.J. Nielsen

Under certain conditions the atmospheric oxidation of one CH4 molecule can lead to the formation of four molecules of ozone. On a global scale, transport from the stratosphere accounts for approximately 10% of tropospheric ozone while peroxy radical chemistry is the source of the remaining 90% [9]. At night, the nitrate radical, NO3 , takes over from the OH radical as the dominant radical species, which initiates the degradation reactions. The nitrate radical is formed by the reaction of NO2 with O3 . NO3 radicals absorb strongly in the visible and are photolyzed rapidly during the day. Hence, NO3 radicals are present in signiﬁcant quantities only at night. The NO3 radical is much less reactive than OH radicals but nevertheless is an important oxidant species for many organic compounds. NO3 radical concentrations can reach levels of the order of 109 cm–3 in polluted nighttime air [10]. NO3 radicals react with a variety of organic species via H atom abstraction and/or addition mechanisms [11–15]. NO3 + HCHO → HNO3 + HCO HCO + O2 → HO2 + CO NO3 + CH3 CH = CH2 + M → CH3 C(•)HCH2 ONO2 + M

(16) (14) (17)

3 Atmospheric Lifetimes The concept of atmospheric lifetime is useful in discussions of the atmospheric degradation of anthropogenic molecules [5]. It can be deﬁned in several ways. Most simply put it can be expressed as the turnover time, which is the atmospheric burden of a given species divided by its rate of emission, assuming a constant emission rate and steady state condition. Alternatively, it can be stated as the reciprocal of the pseudo ﬁrst order rate constant (k ) for its removal: Atmospheric lifetime (τ) =

1 k

Most pollutants are lost from the atmosphere by several routes. For example while 90% of CO is removed from the atmosphere via reaction with OH radicals, about 10% is removed via microbial activity in soils [16]. Similarly, the bulk of CH4 oxidation occurs via OH radical attack however microbial action and reactions Eq. 18 and Eq. 19 also contribute signiﬁcantly [5]. O(1 D) + CH4 → OH + CH3 Cl + CH4 → HCl + CH3

(18) (19)

Atmospheric Photooxidation of Gas Phase Air Pollutants

127

In such cases the atmospheric lifetime is given by Atmospheric lifetime (τ) =

1 = ki i

1 1 i

τi

where, ki and τi are the pseudo ﬁrst order loss rate and lifetime with respect to process i. For example if a compound has a lifetime of 10 years with respect to one loss mechanism and 20 years with respect to another loss mechanism then the overall lifetime will be 7 years. In practice it is often difﬁcult to assign a unique value of atmospheric lifetime of a given species. Difﬁculties arise because of several complicating factors. Firstly, the concept of a single atmospheric lifetime is only applicable to compounds, which are persistent enough to become uniformly mixed throughout the atmosphere. The CFCs provide a good example of such longlived species. Highly reactive compounds like alkenes have lifetimes, which are dependent on the location where they are emitted. Alkenes emitted into air masses with low levels of ozone will survive longer than if emitted into polluted air masses with high ozone levels. Secondly, in cases where the pollutant is released in extremely large amounts, for example CO and CH4 , the lifetime can actually change with the rate of emission. Thus, large-scale emission of CH4 will reduce the OH radical concentration and lead to an increase in the atmospheric lifetime of CH4 and indeed of all other species which are removed via reaction with OH. Thirdly, for some species (most notably CO2 ) there are removal processes in which the species equilibrates with large reservoirs. Atmospheric CO2 equilibrates with CO2 dissolved in the upper layers of the oceans and with the terrestrial biota within approximately 4 years [17]. However, the majority of the CO2 in these reservoirs is returned to the atmosphere within a few years. It is only the relatively small fraction of CO2 that is transferred to the deep ocean that can be considered to be permanently lost from the atmosphere. Loss of CO2 from the atmosphere cannot be represented by a simple exponential decay but is instead is a complex function [18, 19]. As a guide the atmospheric lifetime of CO2 is approximately 50–200 years [17]. Fourthly, the starting point for lifetime estimations is often laboratorygenerated kinetic data for reaction of the compound of interest with OH radicals. The bimolecular rate constants measured in laboratory kinetic experiments need to be converted into a pseudo ﬁrst order rate constant for loss of the compound, k . In principal this conversion is simple, i.e., the bimolecular rate constant merely has to be multiplied by the OH concentration ([OH]). In practice there are difﬁculties associated with the choice of an appropriate value of [OH]. At present we cannot measure the global OH concentration ﬁeld directly. The OH radical concentration varies widely with location, season, and meteorological conditions. To account for such variations requires use of sophisticated 3D computer models of the atmosphere.

128

T.J. Wallington · O.J. Nielsen

In such models the OH concentration ﬁeld is computed using measured or estimated concentration ﬁelds of the precursor molecules and photon ﬂux data. The resulting OH ﬁeld is then tuned such that it correctly predicts the lifetime of methyl chloroform (CH3 CCl3 ) with respect to OH radical attack. From measurements of the atmospheric turnover time of CH3 CCl3 (4.8 years) [20], its lifetime with respect to loss in the stratosphere (45 years), and its lifetime with respect to loss in the oceans (85 years) the tropospheric lifetime of CH3 CCl3 with respect to OH radical attack has been inferred to be 5.7 years [17, 21]. Methyl chloroform is the calibration molecule of choice because it has a long history of precise atmospheric measurements, it has no natural sources, its industrial production is well documented, and because the kinetics of reaction Eq. 20 are well established, k20 = 1.8 × 10–12 exp(– 1500/T) cm3 molecule–1 s–1 [22]. OH + CH3 CCl3 → products

(20)

The use of 3D computer models to calculate atmospheric lifetimes is a rather cumbersome approach and access to such models is limited. A simpler technique to estimate the tropospheric lifetime of compound X with respect to OH attack is to scale the tropospheric lifetime of CCl3 CH3 by the rate constant ratio k(OH+CCl3 CH3 )/k(OH+X). OH + X → products

(21)

If reactions Eq. 20 and Eq. 21 have the same temperature dependence then the rate constant ratio can be evaluated at any given temperature. If not, then the temperature chosen for the comparison is important. Spivakovsky et al. [23] have shown that 272 K is the optimal temperature for comparison. Hence, a reasonable estimate of the atmospheric lifetime of a compound with respect to OH radical attack in the troposphere can be made using: τ (compound)

=

k (OH + CCl3 CH3 ) × 5.7 years k (OH + compound )

Over the past 10–20 years the kinetic data base for reaction of OH radicals with atmospheric pollutants has improved dramatically to a point where uncertainties in kOH are typically in the range 10–20%. Critically evaluated data for OH radical reactions of atmospheric importance are available in several excellent reviews [8, 15, 22, 24, 25]. Likewise there are also extensive databases available for NO3 [8, 13] and O3 reactions [8]. It is normally assumed that the atmospheric lifetime of a pollutant does not change with time. This implies an unchanging OH distribution. As discussed above, the OH concentration decreases with increasing CO and CH4 levels and increases with increasing NOx and O3 levels. Hence, the concentration of tropospheric OH is expected to have changed since the pre-industrial era. However, at present there is no consensus on the magnitude of this change [26].

Atmospheric Photooxidation of Gas Phase Air Pollutants

129

Table 2 Atmospheric lifetimes for selected compounds Compound

Atmospheric lifetime

CO2 CO CH4 – methane

50–200 years a 50 days b 12.2 years

C2 H6 – ethane

60 days b

C3 H6 – propane

12 days b

C8 H18 – i-octane b

3.5 days b

C2 H2 – acetylene

14 days b

C2 H4 – ethene

1.4 days b

C3 H6 – propene

11 hours b

C6 H6 – benzene

10 days b

C6 H5 CH3 – toluene C6 H4 (CH3 )2 – m-xylene

1.9 days b 12 hours

N2 O – nitrous oxide CF2 Cl2 – CFC-12 CFCl3 – CFC-11 CF3 CCl2 H – HCFC-123 CF3 CFH2 – HFC-134a

114 years 100 years 45 years 1.3 years 14 years

CF3 Br – Halon 1301 CF3 H – Fluoroform CF4 – carbon tetraﬂuoride SF6 – sulfur hexaﬂuoride

65 years 270 years 50,000 years 3,200 years

SO2 CF3 I – triﬂuoroiodomethane

12 days b < 1 day

HCHO

6 hours d

a

No single lifetime can be given, see text for details. Estimated using [OH] = 106 cm–3 and the following rate data k(OH + CO) = 2.4 × 10–13 [22], k(OH + C2 H2 ) = 8.1 × 10–13 [25], k(OH + C2 H6 ) = 1.8 × 10–13 [22], k(OH + C3 H8 ) = 9.2 × 10–13 [22], k(OH + C6 H6 ) = 1.2 × 10–12 [25], k(OH + C6 H5 CH3 ) = 6.0 × 10–12 [15], k(OH + C6 H4 (CH3 )2 ) = 2.4 × 10–11 , k(OH + i-octane) = 3.3 × 10–12 [25], k(OH + C2 H4 ) = 8.5 × 10–12 [25], k(OH + C3 H6 ) = 2.6 × 10–11 [25], k(OH + SO2 ) = 9.6 × 10–13 [77]. c i-Octane = 2,2,4-trimethyl pentane, used to deﬁne octane scale of motor fuels. d Approximate lifetime with respect to photolysis, see page 507 of reference [107]. b

Table 2 lists atmospheric lifetimes for a range of anthropogenic molecules. As seen from Table 2 the lifetimes of pollutant molecules range from minutes or hours for photolabile compounds such as CF3 I and HCHO to millennia for perﬂuoro compounds such as CF4 and SF6 . There is discussion as to whether

130

T.J. Wallington · O.J. Nielsen

atmospheric lifetime should be used as a measure of environmental acceptability. It has been argued that it is prudent to restrict the manufacture of compounds, which have excessively long atmospheric lifetimes [5]. The crux of the argument is that because their degradation is so slow any unforeseen adverse environmental impacts of such species would, for all practical purposes, be permanent.

4 Degradation of Alkanes More than 130 different types of alkanes have been identiﬁed in the atmosphere [27] and they comprise a major fraction of the organic pollutants found in urban atmospheres (see Table 3). They are released in large amounts in activities connected with the extraction, reﬁning, distribution, and combustion of fossil fuels. Alkanes are also released during the combustion of organic matter and by microbiological processes associated with the decay

Table 3 The 32 most prevalent non-methane organic compounds (NMOC) in urban air [50] Species

% NMOC a

Species

% NMOC a

2-methyl-butane

7.8

o-xylene

1.5

n-butane toluene propane ethane m, p-xylene

7.2 6.4 4.6 3.9 3.4

3-methylpentane n-hexane 2-methylhexane 2,2,4-trimethylpentane methylcyclopentane

1.5 1.3 1.3 1.2 1.0

2-methyl-propane n-pentane acetylene ethene 1-2-4-trimethylbenzene

3.2 3.1 3.0 2.7 2.3

ethylbenzene formaldehyde acetaldehyde m-ethyltoluene propene

1.0 1.0 1.0 0.9 0.9

4-methyl-nonane 2-methyl-pentane benzene n-decane acetylene

2.1 2.1 2.0 1.9 1.5

2-methyl-propene c-2-pentene 3-methyl-hexane 2,3,3-trimethyl-1-butene n-nonane

0.9 0.8 0.8 0.8 0.7

a

% deﬁned in terms of ppbC, i.e., the % of carbon atoms accounted for by each species

Atmospheric Photooxidation of Gas Phase Air Pollutants

131

of organic matter. Alkanes are relatively unreactive species (the old chemical name for this class of compounds “parafﬁns” is derived from Latin words parum (little) and afﬁnis (having afﬁnity)). Alkanes do not react appreciably with ozone, are not subject to photolysis in the lower atmosphere, and react only slowly with NO3 radicals. The atmospheric oxidation of alkanes is initiated by reaction with OH radicals. The reaction of OH with alkanes proceeds directly via H-atom abstraction and no bound intermediate complex is formed [28]. This reaction produces an alkyl radical which rapidly (within a µs) adds O2 to give an alkyl peroxy radical (RO2 ). OH + RH → R + H2 O R + O2 + M → RO2 + M

(22) (23)

In the atmosphere peroxy radicals react with NO, NO2 , HO2 radicals and other peroxy radicals (R O2 ). The importance of these reactions is dictated by the abundances of NO, NO2 , and HO2 radicals and by the rates of the reactions of RO2 radicals with these species. In the troposphere the concentrations of NO, NO2 , and HO2 vary widely, however, for the present purposes reasonable average concentrations are approximately (2.5–10)×108 cm–3 . Under atmospheric conditions, typical rate constants for the reactions of RO2 radicals with NO, NO2 , and HO2 radicals lie in the ranges (8–20)×10–12 , (5–10)×10–12 , and (5–15)×10–12 cm3 molecule–1 s–1 , respectively [4]. Hence, on average these reactions are of comparable importance in the atmospheric fate of RO2 radicals. On a local scale one reaction may dominate because of variation in the concentrations of NOx (NO and NO2 ) and HO2 radicals. Thus, in remote marine locations with low NOx levels, reaction of RO2 radicals with HO2 will be much more important than in urban air masses with high NOx concentrations. The relative importance of the loss of peroxy radicals via reaction with either HO2 or NOx can be a matter of considerable practical consequence. For example, it has been shown that the degradation of ﬂuorinated aldehydes in the presence of HO2 gives perﬂuorocarboxylic acids [29]. In contrast, degradation of ﬂuorinated aldehydes in the presence of NOx does not give perﬂuorocarboxylic acids [30]. The US Environmental Protection Agency (EPA) has recently stated its concern over the environmental effects of persistent and bioaccumulative long chain PFCAs such as perﬂuorooctanoic acid (PFOA; C7 F15 COOH). There is a current research effort to establish the relative importance of HO2 and NOx as loss mechanisms for the peroxy radicals formed during oxidation of ﬂuorinated aldehydes to assess the potential contribution of ﬂuorinated alcohols and aldehydes to observed PFCA pollution [31]. The reactivity of peroxy radicals towards other peroxy radicals varies over many orders of magnitude depending on the nature of the “R” group. Consequently, it is not possible to provide a simple accounting of the importance of peroxy self and cross reactions compared to the other possible loss mech-

132

T.J. Wallington · O.J. Nielsen

anisms of RO2 radicals. It has been shown by computer modeling studies that cross and self reactions of peroxy radical play an important role in the atmospheric degradation mechanism of organic compounds released into environments containing low levels of NOx , for example in the Amazon Basin [32]. The atmospheric lifetime of RO2 radicals is approximately 1 minute. These reactions are illustrated for the case of methane in Fig. 3. The reaction of peroxy radicals with HO2 radicals can proceed via two reaction channels to give either organic hydroperoxides or carbonyl products. RO2 + HO2 → ROOH + O2 RO2 + HO2 → R CHO + H2 O + O2

(24a) (24b)

The reaction of unsubstituted alkyl peroxy radicals like CH3 O2 and C2 H5 O2 with HO2 leads to the formation of hydroperoxides in essentially 100% yield [33, 34]. In contrast, substituted peroxy radicals (e.g., CH2 FO2 , CH2 ClO2 , HOCH2 O2 , and CH3 OCH2 O2 ) react with HO2 via two pathways to give hydroperoxide and carbonyl products [35–38]. The factors, which determine the branching ratio k24a /k24b for any given RO2 radical are unknown and fur-

Fig. 3 Atmospheric degradation mechanism of methane showing the central role played by the methyl peroxy radical (CH3 O2 ). Values in parentheses are order of magnitude estimates for the lifetimes of the various species

Atmospheric Photooxidation of Gas Phase Air Pollutants

133

ther research is needed in this area. The net effect of reaction Eq. 24 is to convert two reactive radical species into a relatively unreactive hydroperoxide or carbonyl compound. Reaction Eq. 24 slows down the free radical driven photochemical oxidation reactions and reduces the formation of ozone. As indicated in Fig. 3 the hydroperoxides have an atmospheric lifetime of the order of a few days and undergo both photolysis and reaction with OH radicals resulting in the regeneration of ROx radicals. It is well established that the reaction of peroxy radicals with NO2 proceeds by a simple association mechanism in which an alkyl peroxy nitrate is formed. The exothermicity associated with formation of the RO2 – NO2 bond resides initially in the RO2 NO2 molecule. Collisional deactivation is required to remove this excess energy. RO2 + NO2 + M → RO2 NO2 + M

(25)

The RO2 – NO2 bond in alkyl peroxy nitrates is relatively weak (22–25 kcal mol–1 ) and the peroxy nitrate decomposes rapidly to reform the reactants [39]. For example, in 760 Torr of air the lifetimes of CH3 O2 NO2 at room temperature and – 10 ◦ C are 0.6 and 51 s while for C2 H5 O2 NO2 the corresponding lifetimes are 0.3 and 25 s [39]. RO2 NO2 + M → RO2 + NO2 + M

(-25)

Hence, the net effect of the reaction of alkyl peroxy radicals with NO2 is to sequester carbon in the form of peroxy nitrates for a short period of time. Peroxy acyl nitrates (RC(O)O2 NO2 ) such as CH3 C(O)O2 NO2 , C2 H5 C(O)O2 NO2 and C6 H5 C(O)O2 NO2 have RO2 – NO2 bonds which are somewhat stronger than their alkyl counterparts (28 kcal mol–1 for CH3 C(O)O2 NO2 ). This slight increase in bond strength has a dramatic effect on their atmospheric chemistry. The lifetime of CH3 C(O)O2 NO2 at room temperature is 0.6 hours while at – 10 ◦ C it is 13 days. The lifetimes of peroxy acyl nitrates with respect to thermal decomposition are sufﬁcient at the low temperatures characteristic of the upper troposphere to allow long range transport of NOx , and yet short enough at the warmer temperatures of the lower troposphere to release NO2 which can then participate in photochemical reactions [40–42]. While peroxy acyl nitrates are not formed as primary products in the atmospheric oxidation of alkanes, they are important secondary products. For example, ethane is oxidized to give acetaldehyde which can then react to give PAN. Peroxy acetyl nitrate (PAN) is the most abundant peroxy acyl nitrate and is an important component of photochemical smog in urban areas. PAN is largely responsible for the eye irritation experienced in smog episodes and, because of its phytotoxic nature, can damage agricultural crops. Other members of the peroxy acyl nitrate family include peroxy propyl nitrate, C2 H5 C(O)O2 NO2 , and peroxy benzoyl nitrate, C6 H5 C(O)O2 NO2 , which are found at levels ranging from a few percent to 30% of that of PAN [42]. The levels of PAN found in ambient air vary widely; highest levels are found in

134

T.J. Wallington · O.J. Nielsen

polluted urban air in the afternoon and can range up to 30 ppb, while levels as low as a few ppt are observed in cleaner air masses [43]. The reaction of alkyl peroxy radicals with NO is very important in the atmospheric degradation mechanism of alkanes. The reaction proceeds via two pathways. RO2 + NO → RO + NO2 RO2 + NO + M → RONO2 + M

(26a) (26b)

Pathway Eq. 26a is dominant and produces an alkoxy radical (RO) and a molecule of NO2 . Photolysis of NO2 is an important source of tropospheric ozone and is responsible for the formation of photochemical smog in polluted air masses. Pathway Eq. 26b gives alkyl nitrates which are much less reactive than NO or NO2 and so sequester NOx (NO and NO2 ). Reaction channel Eq. 26b represents a loss of both radicals and NOx from the atmosphere and hence slows down the photochemical chain reactions that form ozone. In general, the relative importance of the nitrate producing channel increases with the size of “R” [44] (for example, with R = CH3 , k26b /(k26a + k26b ) < 0.005 [45], while for R = t-butyl, k26b /(k26a + k26b ) = 0.18 [46]). For C1 – C3 hydrocarbons the formation of nitrates is of minor importance while for C4 and above, nitrate formation is signiﬁcant. Barker and coworkers [47, 48] have conducted master equation calculations of the nitrate yield in reaction Eq. 26. It was found that surprisingly small (perhaps physically unreasonable) values of the average energy transfer in each collision were needed to reproduce the experimentally measured pressure dependence of the nitrate yield [48]. The modeling work suggests that there is either a fundamental error in our understanding of the mechanism by which nitrate is formed in RO2 + NO reactions or, there are errors in the limited data base of experimentally determined nitrate yields (particularly the pressure dependence). The alkoxy radicals formed in pathway Eq. 26a have very interesting atmospheric chemistry. The atmospheric fate of alkoxy radicals differs with the nature of the “R” group. Some alkoxy radicals (e.g., CH3 O) are lost solely via reaction with O2 , others undergo rapid decomposition via C – C bond scission. Long chain alkoxy radicals can undergo isomerization via intramolecular H-atom abstraction: CH3 CH2 CH2 CH2 CHO(•)CH3 → CH3 C(•)HCH2 CH2 CH(OH)CH3

(27)

The resulting hydroxy alkyl radical then adds O2 to become a hydroxy alkyl peroxy radical which reacts with more NO to give more NO2 and another alkoxy radical capable of undergoing further isomerization: CH3 CHO(•)CH2 CH2 CH(OH)CH3 → CH3 CH(OH)CH2 CH2 C(•)(OH)CH3

(28)

Atmospheric Photooxidation of Gas Phase Air Pollutants

135

In this fashion the oxidation of long chain alkanes quickly gives multifunctional oxygenated products. These products often have low volatility and form aerosols, which can be removed by wet or dry deposition to the ground. The isomerization reactions given above occur via a six-membered transition state. Isomerization can also occur via ﬁve and seven-membered states. The complexity associated with unraveling the precise degradation mechanism of any given alkane can be appreciated by considering the case of n-hexane. Initial OH radical attack leads to the formation of three different alkoxy radicals, each of which can either react with O2 , decompose, isomerize (via several possible pathways), or undergo a combination of these possible loss processes. Our understanding of the atmospheric chemistry of alkoxy radicals is rather crude at present and this is an area of active research [49]. In polluted air masses typical of continental and urban areas the NO levels are high enough to dominate the loss of peroxy radicals. Alkanes and other volatile organic compounds (VOC) are present at high concentrations in polluted air. Ozone levels in such environments are controlled by the rate at which peroxy radicals are formed, the rate at which they convert NO to NO2 , and the extent to which they form alkyl nitrates and peroxyacylnitrates which sequester NOx . The latter three factors are determined by the quantity and identity of the organic compounds which are present. There is practical interest in establishing the relationship between emissions of alkanes and other volatile organic compounds (VOC) and NOx into urban air masses and the formation of secondary air pollutants such as ozone. Over the past three decades great strides have been made in understanding the fundamental chemistry which is in operation in the formation of oxidants in urban air. It is now recognized that complex non-linear feedback processes relate VOC and NOx emissions to ozone levels [50]. While the complexity of the situation deﬁes easy solution, recognition of the need to reduce VOC and/or NOx emissions has driven the search for new technology and new approaches which reduce the emission of these compounds. The automobile industry provides a good example of the progress made on this front. In the 1960s prior to control, the tailpipe emissions of a typical automobile in California were 8.8 grams of hydrocarbons and 3.6 grams of NOx per mile. With the help of catalyst technology, tailpipe emissions from new super ultra low emission (SULEV) vehicles in California in 2004 have been reduced to less than 0.01 grams of hydrocarbons and 0.02 grams of NOx per mile. Regulations in the U.S. and Europe pertaining to the emission of reactive organic compounds into the atmosphere may move away from standards in which only the total organic mass is considered, to standards, which take into account the different reactivity of different compounds. This has generated interest in the relative ranking of organic compounds in terms of their ability to contribute to ozone formation in urban areas. Calculation of scales of “reactivity factors” [51] or “photochemical ozone creation potentials” (POCPs) [52–54] for organic compounds, have been generated. These

136

T.J. Wallington · O.J. Nielsen

scales reﬂect the rate at which the organic compound gives peroxy radicals together with the efﬁciency with which these peroxy radicals contribute to, or hinder, ozone formation.

5 Degradation of Alkenes More than 140 different alkenes have been identiﬁed in the atmosphere [27]. The sources of alkenes are similar to those for the alkanes with combustion of fossil fuel being a major source. The presence of unsaturated bonds makes these compounds much more reactive than the alkanes. The most persistent member of this class of compounds (ethene) has an atmospheric lifetime of the order of a day, while more typically the lifetimes for alkenes are measured in hours. As a result of their short lifetimes the atmospheric concentrations of alkenes are highly variable and decrease dramatically away from their source locations. The mechanisms of atmospheric oxidation of alkenes have recently been reviewed [55]. As with the alkanes the reaction of OH radicals is an important loss mechanism. This reaction proceeds mainly via addition to the unsaturated bond as illustrated for ethene in Fig. 4. In one atmosphere of air at 298 K the dominant atmospheric fate of the alkoxy radical HOCH2 CH2 O is decomposition via C – C bond scission, while reaction with O2 makes a 20% contribution [56]. The fate of alkoxy radicals resulting from addition of OH to alkenes is generally decomposition via C – C bond scission [8]. Thus, the OH radical initiated oxidation of propene gives acetaldehyde and HCHO, oxida-

Fig. 4 Mechanism of the OH-initiated atmospheric degradation of ethene

Atmospheric Photooxidation of Gas Phase Air Pollutants

137

Table 4 Rate constants and corresponding lifetimes for reaction of selected unsaturated compounds with O3 and NO3 [8] τ(O3 ) b

k(NO3 ) a

τ(NO3 ) b

Compound

k(O3 ) a

ethene propene 1-butene

1.6 × 10–18 1.0 × 10–17 9.6 × 10–18

2 days 11 hours 12 hours

2.1 × 10–16 9.5 × 10–15 1.4 × 10–14

1,3-butadiene cyclohexene limonene 2-methyl-1,3-butadiene (isoprene) α-pinene acetylene

6.3 × 10–18 8.1 × 10–17 2.0 × 10–16 1.3 × 10–17

18 hours 1.4 hours 34 mins 8.7 hours

1.0 × 10–13 5.9 × 10–13 1.2 × 10–11 6.8 × 10–13

11 hours 1.9 hours 6 mins 1.6 days

8.7 × 10–17 1 × 10–20

1.3 hours 470 days

6.2 × 10–12 < 1 × 10–16

1.2 hours > 463 days

220 days 4.9 days 3.3 days

a

Units of cm3 molecule–1 s–1 . Calculated using [O3 ] = 100 ppb and [NO3 ] = 10 ppt. NO3 radicals are only present at night, a calculated lifetime in excess of 12 hours indicates that multiple nights exposure are needed for substantial loss.

b

tion of 1-butene gives propionaldehyde and HCHO, and oxidation of 2-butene gives acetaldehyde [57, 58]. In addition to OH radicals, unsaturated bonds are reactive towards O3 and NO3 radicals and reaction with these species is an important atmospheric degradation mechanism for unsaturated compounds. Table 4 lists rate constants for the reactions of O3 and NO3 radicals with selected alkenes and acetylene. To place such rate constants into perspective we need to consider the typical ambient atmospheric concentrations of O3 and NO3 radicals. Typical ozone concentrations in pristine environments are 20–40 ppb while concentrations in the range 100–200 ppb are experienced in polluted air. The ambient concentration of NO3 is limited by the availability of NOx sources. In remote marine environments the NOx levels are extremely low (a few ppt) and NO3 radicals do not play an important role in atmospheric chemistry. In continental and urban areas the NOx levels are much higher (up to several hundred ppb in polluted urban areas) and NO3 radicals can build up to 5–100 ppt at night (NO3 radicals are photolyzed rapidly and are not present in appreciable amounts during the day). For the purposes of the present discussion we have calculated the atmospheric lifetimes of selected unsaturated compounds in Table 4 in the presence of 100 ppb (2.5 × 1012 cm–3 ) of O3 and 10 ppt (2.5 × 108 cm–3 ) of NO3 . Lifetimes in other environments can be evaluated by appropriate scaling of the data in Table 4. As seen from Table 4, the more reactive unsaturated compounds have lifetimes with respect to reaction with O3 and NO3 radicals of only a few minutes!

138

T.J. Wallington · O.J. Nielsen

Scheme 1

The reaction of ozone with alkenes proceeds via addition to the double bond giving a short lived “energy-rich” ozonide, which decomposes via C – C bond scission to give carbonyl compounds and biradical species known as “Criegee” biradicals: For symmetrical alkenes pathways “a” and “b” give the same products while unsymmetrical alkenes give different products via pathways “a” and “b”. The relative importance of decomposition pathways “a” and “b” are comparable for 1-alkenes [55]. For alkenes of the type R1 R2 = CH2 decomposition via channel “b” to form the disubstituted biradical, R1 R2 C(•)OO(•), is slightly favored (by a ratio of ≈ 65 : 35) over decomposition via channel “a” to give the unsubstituted biradical, (•)CH2 O2 (•). Likewise, ozonides formed in reaction with alkenes of the general type R1 CH = CR3 R4 show a slight preference (≈ 65 : 35) to decompose to give the di-substituted biradical [55]. The reaction of O3 with alkenes is very exothermic. Assuming that the C-H bond dissociation energy in CH3 O2 radicals is the same as that in CH3 OH (94.1 kcal mol–1 ) it can be estimated that reaction Eq. 29 is exothermic by 27 kcal mol–1 . O3 + CH2 = CH2 → HCHO + (•)CH2 O2 (•)

(29)

Criegee biradicals formed in the reaction of ozone with alkenes carry signiﬁcant internal excitation and the majority of them decompose before they can be collisionally stabilized. The biradicals are sufﬁciently excited that many different rearrangements and decomposition pathways are energetically allowed. For example, in the case of the [CH3 C(•)HOO(•)]∗ species formed during propene oxidation the IUPAC panel has recommended the following channels (and yields) [24]: [CH3 C(•)HOO(•)]∗ + M → CH3 C(•)HOO(•) + M∗ 15% [CH3 C(•)HOO(•)]∗ → CH3 + CO + OH 54% [CH3 C(•)HOO(•)]∗ → CH3 + CO2 + H 8.5% [CH3 C(•)HOO(•)]∗ → HCO + CH3 O 8.5% [CH3 C(•)HOO(•)]∗ → CH4 + CO2 14%

(30a) (30b) (30c) (30d) (30e)

Atmospheric Photooxidation of Gas Phase Air Pollutants

139

Channel Eq. 30b is of considerable importance as it produces OH radicals. The OH radical yield varies between 10 and 100% depending on the particular alkene [8, 59]. Although it has been known for many years that OH radicals are produced in the reaction of ozone with alkenes [60] it has only recently been recognized that this could be an important nighttime source of OH radicals in the atmosphere. Channel Eq. 30a gives a stabilized biradical. The atmospheric fate of stabilized biradicals is dominated by reaction with water vapor, which proceeds predominately to give carboxylic acids, e.g., CH3 C(•)HOO(•) + H2 O → CH3 C(O)OH + H2 O

(31)

In summary, the reaction of ozone with alkenes is important in the atmospheric degradation of alkenes. In all cases the reaction leads to rupture of the > C = C < double bond. The double bond is replaced by a carbonyl group on one side and a Criegee biradical on the other. The Criegee biradical is formed energetically excited and decomposes by a variety of different routes to give a complex mixture of oxygenated products (mainly carbonyls). Table 4 lists kinetic data for reactions of NO3 radicals with selected alkenes. As seen in Table 4 reaction with NO3 radicals can play an important role in the atmospheric degradation of alkenes. Reaction proceeds via addition of NO3 to the double bond giving a nitrooxy alkyl radical. The resulting energy-rich nitrooxy radical either decomposes to an oxirane or is stabilized by transferring its excitation during collisions with a third body (M). In the presence of air the thermalized nitrooxy alkyl radical adds O2 rapidly (within a µs) to give a peroxy radical which can then react with NO, NO2 , HO2 , or other peroxy radicals (R O2 ) to give a variety of products.

Scheme 2

Berndt and Böge [61, 62] have studied the relative kinetics in the mechanism given above. For reaction of NO3 radicals with tetramethyl ethylene they report kb /ka = (1.75 ± 0.19) × 10–19 and kd /kc = (1.30 ± 0.34) × 10–16 cm3 molecule–1 . Assuming kd is of the order of 10–12 and that kb approaches the gas kinetic rate limit of 10–10 cm3 molecule–1 s–1 (rates typical for such processes [39, 63]) then kc is of the order of 104 s–1 while ka is of

140

T.J. Wallington · O.J. Nielsen

the order of 109 s–1 . Rate constant ratios kb /ka have been measured for the reaction of NO3 with the following alkenes (in units of 10–19 cm3 molecule): propene (1.53), 1-butene (1.9), trans-2-butene (3.3), 2-methyl propene (7.4), 2-methyl 2 butene (4.4), and 2,3-dimethyl 2-butene (1.8). In the atmosphere the nitrooxy alkyl peroxy radical, > C(ONO2 ) – COO(•) <, behaves like other alkyl peroxy radicals and will react with NO2 , HO2 , and other peroxy radicals. Reaction of nitrooxy alkyl peroxy radical with NO is unlikely because the conditions necessary for the formation of NO3 radicals (high O3 ) are incompatible with the presence of signiﬁcant amounts of NO. For unsymmetrical alkenes the addition of NO3 radicals leads to the formation of two different peroxy radicals, e.g., for propene:

Scheme 3

As discussed in section 4, reaction of the peroxy radicals with NO2 gives thermally unstable peroxy nitrates. Reaction with HO2 gives hydroperoxides and possibly carbonyl compounds. Reaction with other peroxy radicals (R O2 ) gives alkoxy radicals, carbonyls, and alcohols. The alkoxy radicals will then either isomerize, react with O2 , or decompose (see Sect. 3). Thus, the NO3 radical-initiated atmospheric degradation of alkenes leads to oxiranes (generally in small yield), nitrooxy hydroperoxides, nitrooxy carbonyls, and nitrooxyalcohols. For a detailed listing of products from individual alkenes the reader should consult Calvert et al. [55]. In summary, alkenes are reactive compounds and are removed rapidly from the atmosphere by a variety of processes. Reaction with OH radicals, ozone, and NO3 radicals all play important roles. These reactions proceed via addition to the unsaturated bond giving an adduct which decomposes and/or reacts with O2 leading to the generation of a variety of transient radical species which react to form the ﬁrst generation closed-shell products (principally carbonyl compounds).

6 Degradation of Aromatics As illustrated by Table 3 aromatic compounds such as toluene, benzene, o, m, p-xylenes and ethyl benzene are important pollutants in urban air. These compounds comprise a substantial fraction of automotive fuel and its use results in substantial emissions of aromatic compounds into urban atmo-

Atmospheric Photooxidation of Gas Phase Air Pollutants

141

spheres. In addition to the monocyclic aromatic compounds there is considerable interest in the chemistry of polycyclic aromatic hydrocarbons (PAH) such as naphthalene, anthracene and pyrene. The discussion here will be conﬁned to monocyclic aromatics, but the degradation of PAHs has many similarities to that of the simpler one ring systems. Details can be found elsewhere [64]. Reaction of monocyclic aromatics with O3 and NO3 radicals is generally very slow and unimportant. Atmospheric degradation of aromatics is initiated by OH radical attack. The kinetics of the reaction of OH radicals with aromatic compounds are well established [65]. As seen from Table 2 the lifetime of aromatics with respect to reaction with OH is typically a few days or less. Reaction proceeds by addition to the ring and H-atom abstraction from either the substituent groups or possibly from the ring C – H sites. In all cases the addition channel is dominant. For benzene, toluene, o, m, p-xylene, and ethyl benzene the H-atom abstraction pathway accounts for 5–10% of the overall reaction [15], the remaining 90–95% proceeds via addition. For toluene ka /(ka + kb ) = 0.07 [15, 65].

Scheme 4

The benzyl radical formed in channel (a) behaves like an alkyl radical and in air adds O2 rapidly (within a µs) to give a benzyl peroxy radical which either reacts with HO2 radicals to give a hydroperoxide, with NO2 to give a thermally unstable peroxy nitrate, or with NO to give a 10–12% yield of benzyl nitrate and 88–90% yield of an alkoxy radical which then reacts with O2 to produce benzaldehyde [15, 66]. The majority of the reaction proceeds via the addition channel (b) to give a methyl hydroxycyclohexadienyl radical. Cyclohexadienyl radicals are resonance stabilized and are relatively unreactive. Typical alkyl radicals add O2 rapidly with rate constants of the order of 10–12 , in contrast the reaction of cyclohexadienyl and methyl cyclohexadienyl radicals with O2 proceed with rate constants of (2 – 5) × 10–16 cm3 molecule–1 s–1 [67, 68]. It can be calculated that in one atmosphere of air the cyclohexadienyl radicals have a lifetime of

142

T.J. Wallington · O.J. Nielsen

approximately 1 ms with respect to reaction with O2 . Under atmospheric conditions the dominate fate of the OH-aromatic adduct is reaction with O2 . In smog chamber experiments loss of the OH-aromatic adduct via reaction with NO2 can be important. There have been a large number of smog chamber experiments, which have focused on identifying the products formed during the atmospheric degradation of aromatics. This is challenging work as the simultaneous identiﬁcation and quantiﬁcation of the large number of chemically similar products is a difﬁcult task. Table 5 lists the observed major products for benzene, toluene, and o-xylene. The atmospheric degradation of aromatic compounds proceeds via complex pathways, which generate a large number of products. Of all the aromatic compounds, the atmospheric degradation of toluene has been studied most extensively and is best understood. Yet even for this molecule the observed identiﬁed products only account for approximately Table 5 Products of the OH initiated atmospheric degradation of selected aromatic compounds [15] Aromatic

Product

Benzene

Glyoxal Phenol Nitrobenzene Glyoxal Methyl glyoxal

(HC(O)C(O)H) (C6 H5 OH) (C6 H5 NO2 ) (HC(O)C(O)H) (CH3 C(O)C(O)H)

0.21 0.05–0.60 [69] 0.03 a 0.08–0.15 0.08–0.15

Benzaldehyde Benzyl nitrate o-Cresol m, p-Cresol m-Nitrotoluene

(C6 H5 CHO) (C6 H5 CH2 ONO2 ) (C6 H4 (CH3 )OH) (C6 H4 (CH3 )OH) (C6 H4 (CH3 )NO2 )

0.05–0.12 0.008 0.13–0.22 0.05 0.01 a

Glyoxal Methyl glyoxal Biacetyl o-Tolualdehyde 2-Methylbenzyl nitrate

(HC(O)C(O)H) (CH3 C(O)C(O)H) (CH3 C(O)C(O)CH3 ) (C6 H4 (CH3 )CHO) (C6 H4 (CH3 )CH2 ONO2 )

0.03–0.09 0.12–0.25 0.09–0.26 0.05–0.17 0.01 a

2,3-Dimethylphenol 3,4- Dimethylphenol 3-Nitro-o-xylene 4-Nitro-o-xylene Nitro-o-xylenes

(C6 H3 (CH3 )2 OH) (C6 H3 (CH3 )2 OH) (C6 H3 (CH3 )2 NO2 ) (C6 H3 (CH3 )2 NO2 ) (C6 H3 (CH3 )2 NO2 )

0.10 0.06 0.005 0.01 a 0.07

Toluene

o-Xylene

a

in limit of zero [NO2 ]

Molar yield

Atmospheric Photooxidation of Gas Phase Air Pollutants

143

60% of its loss. Experimental studies aimed at establishing/verifying chemical mechanisms for aromatic hydrocarbons should be performed using NOx levels, which are representative of those found in the atmosphere [69]. Despite their recognized importance and a great deal of scientiﬁc study the atmospheric degradation mechanisms of aromatic compounds remains unclear at the present.

7 Degradation of NOx NOx refers to the sum of two of the most reactive nitrogen species NO and NO2 . NOx plays a critical role in determining levels of ozone. Over the last 100 years it is believed that tropospheric background levels of ozone in the Northern hemisphere have increased by approximately a factor of two and that this increase has been strongly inﬂuenced by increased levels of NOx [9]. The major sources of NOx in the troposphere are fossil fuel combustion, biomass burning, microbiological activity in soil, lightning, oxidation of ammonia, and tropopause folding. Each year approximately 53 Tg N of NOx is emitted into the atmosphere. Fossil fuel combustion contributes one half of this total while the contribution from biomass burning is 2.1–5.5 Tg N/yr [70]. N2 O emitted from biological activity in soil is long-lived in the troposphere and is transported to the stratosphere where it reacts with O(1 D) to produce NO. Hence N2 O, which is normally not included in any of the NOx or NOy families, can be a secondary source of NOx . NOy refers to “odd nitrogen” and is the sum of NOx and all oxidized nitrogen species not including N2 O. The chemical role played by NOx in the atmosphere is extremely complex and will not be dealt with in complete detail in this chapter. Sources of NOx have been mentioned brieﬂy above. How then is NOx removed from the atmosphere? A “ﬁrst-order” simpliﬁed atmosphere containing only NO, NO2 and O3 is characterized by the equations: NO + O3 → NO2 + O2 NO2 + hν(λ < 420 nm) → NO + O O + O2 + M → O3 + M

(32) (6) (7)

NO reacts rapidly with O3 giving NO2 . Using k32 = 1.8 × 10–14 cm3 molecule–1 s–1 [22], and assuming a constant O3 concentration of 30 ppbv (7 × 1011 cm–3 ), then NO has a lifetime of 1.3 minutes with respect to reaction Eq. 32. NO2 is a brown-colored gas and absorbs at 400–450 nm. The NO2 photolysis rate, JNO2 , in the troposphere depends on the cloud cover and is typically in the range (0.3–1)×10–2 s–1 , giving a lifetime of NO2 with respect to

144

T.J. Wallington · O.J. Nielsen

photolysis of 2–6 minutes [4]. In one atmosphere of air, reaction Eq. 7 has an effective bimolecular rate constant of k7 = 1.5 × 10–14 cm3 molecule–1 s–1 [22], [O2 ] = 5.2 × 1018 cm–3 , and O atoms have a lifetime of 13 ms with respect to reaction Eq. 7. As a result of the rapidity of reactions Eqs.6, 7, and 32, a photostationary steady state condition is set up where NO and O atoms are created and destroyed continually but maintain a steady concentration, i.e., d[O]/dt = d[NO]/dt = 0. As a result of their rapid interconversion it is convenient to lump NO and NO2 together and discuss the atmospheric degradation of NOx . Reactions Eqs. 6, 7, and 32 form the basis for what is known as the Leighton relationship, after Philip Leighton who wrote one of the ﬁrst textbooks on air pollution in 1961 [71]. Applying steady state conditions one can derive: [NO2 ] k32 [O3 ] = JNO2 [NO] where JNO2 is the ﬁrst order rate constant for photolysis of NO2 . When measurements of JNO2 , [NO2 ], [NO], and [O3 ] are performed in the ﬁeld, the observed concentration ratio [NO2 ]/[NO] is often more than that calculated from the right hand side of the above equation [72, 73]. In other words, to maintain the observed [NO2 ]/[NO] mixing ratio there must be some species, other than O3 , that is capable of oxidizing NO to NO2 which is not accounted for (i.e., missing) in the simple mechanism of reactions Eqs.6, 7, 32. This “missing oxidant” is the peroxy radical. In the presence of peroxy radicals, additional conversion of NO to NO2 occurs via reaction of peroxy radicals with NO. RO2 (or HO2 ) + NO → RO (or OH) + NO2

(26)

When reaction Eq. 26 is included in the steady state analysis the predicted ozone concentration becomes: [NO2 ] k32 [O3 ] k26 [RO2 ] = + JNO2 JNO2 [NO] While reactions Eq. 6 and Eq. 32 interconvert NO and NO2 , they do not remove NOx from the atmosphere. NOx is removed from the atmosphere by the conversion of NO2 into HNO3 and aerosol NO3 – which are then subject to wet and dry deposition [42]. There are a variety of reactions that convert NOx into HNO3 . In the daytime the association reaction of OH radicals with NO2 is important. OH + NO2 + M → HNO3 + M

(33)

Nighttime loss of NOx is tied to N2 O5 which acts as a reservoir. NO2 reacts with O3 to form NO3 which is in equilibrium with N2 O5 : NO3 + NO2 + M → N2 O5 + M

(34)

Atmospheric Photooxidation of Gas Phase Air Pollutants

145

N2 O5 undergoes reaction with water to give HNO3 : N2 O5 + H2 O(aq) → 2HNO3 (aq)

(35)

As discussed in the previous sections reaction of NO3 radicals with organic compounds results in the formation of HNO3 and organic nitrates. In the stratosphere ClO radicals react with NO2 to give ClONO2 which undergoes hydrolysis. ClONO2 + H2 O(aq) → HNO3 (aq) + HCl(aq)

(36)

NO2 can be converted to PAN (CH3 COO2 NO2 ) and PAN-like compounds (RO2 NO2 ): CH3 C(O)O2 + NO2 + M ↔ CH3 C(O)O2 NO2 + M RO2 + NO2 + M ↔ RO2 NO2 + M

(37) (25)

PAN and PAN-like compounds are important reservoirs for NOx and can transport NOx over great distances (see Sect. 4). In the past 10–20 years signiﬁcant progress has been made in understanding the atmospheric chemistry of NOx and other nitrogen species [42]. Measurement of the individual nitrogen species has proven to be a valuable tool for evaluating photochemical processes occurring in an airmass.

8 Degradation of SOx SO2 emitted from human activities amounts to about 140 Mtonnes S yr–1 [74]. There is considerable scatter in the estimates of natural emissions of sulfur compounds, ranging from 34 Mtonnes S yr–1 [75] to 267 Mtonnes S yr–1 [76]. One estimate is 100 Mtonnes per year including SOx from sulﬁdes [74]. Anthropogenic emissions of SO2 are increasing in the developing countries while decreasing in USA, Japan and Western Europe. These anthropogenic emissions are mainly due to fossil fuel combustion. Efforts to reduce SO2 emissions have been mainly through ﬂue-gas desulfurization, use of low-sulfur fuels, and the removal of sulfur from fuels. SO2 is oxidized to sulfuric acid both by homogeneous gas-phase reactions and by multiphase processes when a precursor gas is dissolved in water and then subsequently oxidized. The routes to atmospheric sulfuric acid production have been studied extensively for decades. The most important gas-phase mechanism is oxidation by OH radicals: SO2 + OH + M → HSO3 + M HSO3 + O2 → HO2 + SO3 SO3 + H2 O + M → H2 SO4 + M

(38) (39) (40)

146

T.J. Wallington · O.J. Nielsen

At 298 K in 760 Torr of air k38 = 9.6 × 10–13 cm3 molecule–1 s–1 [77] which leads to a lifetime of SO2 with respect to reaction with OH of approximately 12 days. HO2 produced in reaction Eq. 2 can oxidize NO to NO2 allowing ozone to be formed together with H2 SO4 . There are other gas-phase routes to formation of sulfuric acid from SO2 , but they are of very limited importance in the atmosphere. SO2 is a moderately soluble gas and its solubility is enhanced by hydrolysis: SO2 + H2 O → SO2 H2 O SO2 + H2 O → H+ + HSO3 – HSO3 – → H+ + SO3 2–

(41) (42) (43)

pH inﬂuences the overall solubility in cloud droplets. This has an impact on the efﬁciency of the aqueous phase SO2 oxidation. Since sulfuric acid is a product, the oxidation reaction can become self-limiting. The predominant aqueous phase oxidation mechanism is oxidation by hydrogen peroxide: HSO3 – + H2 O2 → HSO4 – + H2 O HSO4 – + H+ → H2 SO4

(44) (45)

The average pH of cloud and rain water is less than 5, particularly in polluted areas and in remote ocean areas far from large dust concentrations which may neutralize acidity. At a pH between 3 and 5 the sulfuric acid production rate shows a positive response to the presence of H+ ions [78], which counteracts the tendency for SO2 to be less soluble. There is sufﬁcient H2 O2 in clouds to oxidize the SO2 which reaches cloud height and calculations estimate that the multiphase oxidation of SO2 by H2 O2 accounts for most of the sulfuric acid present in rain, and over 70% of the sulfate aerosol observed in the atmosphere [79]. The aerosol is produced after evaporation of cloud droplets. It has been found in experiments carried out in the real atmosphere that the overall rate of oxidation of SO2 in clouds is many times faster than would be predicted from using laboratory-based measurements [80].

9 Degradation of CFCs and Halons Chloroﬂuorocarbons (CFCs) give rise to perhaps the most well-known environmental problem, depletion of the stratospheric ozone layer. CFCs were developed in the late 1920s in the search for more “friendly” refrigerants as substitutes for ammonia and sulfur dioxide. The most widely used CFCs were CFC-11 (CFCl3 ) and CFC-12 (CF2 Cl2 ) which were employed as propellants in aerosol spray cans, foam blowing agents, solvents, and refrigerants. CFC production peaked in 1974 with a global annual production of CFC-11 and

Atmospheric Photooxidation of Gas Phase Air Pollutants

147

CFC-12 of 350 000 tonnes/yr and 400 000 tonnes/yr, respectively. The chemical inertness of CFCs, which is one of their desired properties is also the property that causes them to be environmentally unacceptable. After their use CFCs are released to the atmosphere. Lovelock [81] was the ﬁrst to detect CFCs in the troposphere. The amounts found in the atmosphere were indistinguishable, within experimental error, to the total amount manufactured and released. CFCs do not react with OH or NO3 radicals or ozone and only absorb UV at short wavelengths which do not penetrate into the troposphere; there is no known loss mechanism of CFCs in the troposphere. The UV absorption spectra of CFC-11 and -12 and Halon 1301 (used as a ﬁre extinguisher) are shown in Fig. 5. By comparison of Fig. 5 with the solar ﬂux at different altitudes in Fig. 1 it can be seen that while photolysis of CFCs is not possible in the troposphere, it can occur slowly at 20 km and more rapidly at 40 km. The atmospheric degradation pathway of CFCs is transport to the stratosphere followed by UV photolysis: CF3 Cl + hν → CF3 + Cl

(46)

with photodissociation coefﬁcients at 40 km height of approximately 10–7 s–1 and 10–6 s–1 for CFC-12 and CFC-11, respectively. At 40 km the “local lifetime” of CFC-11 is approximately one month. The local lifetime should not be confused with the “atmospheric lifetime” of CFC-11 which is 45 years (see Table 6). The difference lies in the fact that only a small fraction of the total atmospheric mass resides at 40 km. The rate-determining step in the degradation of CFCs is the time taken for air to pass through the stratosphere as part of the natural circulation. The measured altitude proﬁles of CFCs are con-

Fig. 5 UV absorption spectra (base e) for CFC-11, CFC-12 and Halon 1301 [22]

148

T.J. Wallington · O.J. Nielsen

Table 6 Direct Global Warming Potentials (mass basis) for gases that have adequately characterized lifetimes Industrial designation or common name

Chemical formula

Carbon dioxide Methane Nitrous oxide Chloroﬂuorocarbons

CO2 CH4 N2 O

12.0 114

CFC-11 CFC-12 CFC-13 CFC-113 CFC-114

CCl3 F CCl2 F2 CClF3 CCl2 FCClF2 CClF2 CClF2

45 100 640 85 300

Hydrochloroﬂuorocarbons HCFC-21 HCFC-22 HCFC-123

CHCl2 F CHClF2 CHCl2 CF3

1.7 12.0 1.3

148 1780 76

CHClFCF3 CH3 CCl2 F CH3 CClF2

5.8 9.3 17.9

599 713 2270

HCFC-124 HCFC-141b HCFC-142b Hydroﬂuorocarbons HFC-23

CHF3

HFC-32 HFC-41 HFC-125 HFC-134a HFC-143a

CH2 F2 CH3 F CHF2 CF3 CH2 FCF3 CH3 CF3

HFC-152a HFC-227ea HFC-236cb HFC-236ea HFC-236fa

CH3 CHF2 CF3 CHFCF3 CH2 FCF2 CF3 CHF2 CHFCF3 CF3 CH2 CF3

HFC-245ca HFC-245fa HFC-365mfc HFC-43-10mee

CH2 FCF2 CHF2 CHF2 CH2 CF3 CH3 CF2 CH2 CF3 CF3 CHFCHFCF2 CF3

Lifetime (years)

270

GWP (100 years) 1 23 d 300 4680 10 720 14 190 6030 9880

12 240

4.9 2.4 29 14.0 52

543 90 3450 1320 4400

1.4 34.2 13.6 10.7 240

122 3660 1320 1350 9650

6.2 7.6 8.6 15.9

682 1020 782 1610

Atmospheric Photooxidation of Gas Phase Air Pollutants

149

Table 6 (continued) Industrial designation or common name Fully ﬂuorinated Sulfur hexaﬂuoride Triﬂuoromethylsulfurpentaﬂuoride

Chemical formula

SF6 SF5 CF3 CF4

Lifetime (years)

GWP (100 years)

3200 800

22 450 17 500

50 000

5820

FC-14 Halogenated alcohols ethers HFE-125 HFE-134 HFE-143a

CHF2 OCF3 CHF2 OCHF2 CH3 OCF3

HFE-254cb2 HFE-347mcc3 HFE-356pcf3 HFE-7100 HFE-7200

CH3 OCF2 CHF2 CH3 OCF2 CF2 CF3 CHF2 OCH2 CF2 CHF2 C4 F9 OCH3 C4 F9 OC2 H5

2.6 5.2 3.6 5 0.77

353 566 494 397 56

H-Galden1040x

CHF2 OCF2 OC2 F4 OCHF2

6.3

1840

136 26 4.3

14 670 6220 744

sistent with their photolytic loss in the stratosphere. The radical fragments resulting from CFC photolysis are converted into HF, HCl, and CO2 . HF and HCl are soluble in water and are removed from the atmosphere by rain-out. Molina and Rowland [82] were the ﬁrst to recognize the threat of the released chlorine to the stratospheric ozone layer. The threat originates from the catalytic cycle: Cl + O3 → ClO + O2 ClO + O → Cl + O2 net: O + O3 → O2 + O2

(47) (48)

A chlorine atom may go through this cycle many thousands of times before it is removed in the form of HCl following reaction with methane: Cl + CH4 → HCl + CH3

(49)

Halons such as Halon-1211 (CF2 BrCl) and Halon-1301 (CF3 Br) are brominated CFCs which are used as ﬁre extinguishers. Like CFCs, Halons are chemically inert in the troposphere but photolyze in the stratosphere. Photolysis releases bromine atoms, which can remove stratospheric ozone in a cycle that is analogous to the chlorine-based cycle above. The bromine cycle can couple

150

T.J. Wallington · O.J. Nielsen

with the chlorine cycle in several ways, e.g.: Br + O3 → BrO + O2 Cl + O3 → ClO + O2 BrO + ClO → Br + Cl + O2 net: O3 + O3 → 3O2

(50) (51) (52)

On a per molecule basis bromine is more efﬁcient than chlorine in destroying ozone. Recognition of the adverse impact of CFCs on stratospheric ozone led to legislation in the US, Canada, and Scandinavia in the 1970s to limit the use of such compounds. Further legislation was forthcoming after the discovery of the Antarctic Ozone hole in 1985. In 1987 the so-called “Montreal Protocol” was signed enforcing international reduction in the production and use of CFCs. The Montreal protocol has been further tightened several times since 1987. The production and use of CFCs is now restricted on a global basis. The effect of these restrictions can already seen in the trends in atmospheric concentrations of CFCs. The prompt global action taken on this environmental problem has been extremely successful and should create optimism for solutions to future environmental problems. Within a relatively short time CFC replacement compounds have been identiﬁed, tested, and marketed. The results of a thorough investigation of their atmospheric chemistry are described in the next section.

10 Degradation of HFCs, HCFCs, and HFEs Recognition of the adverse impact of chloroﬂuorocarbons (CFCs) on stratospheric ozone [82] has prompted an international effort to replace CFCs with environmentally acceptable alternatives [83]. Hydroﬂuorocarbons (HFCs), hydrochloroﬂuorocarbons (HCFCs) and hydroﬂuoroethers (HFEs) are three classes of CFC replacements. In contrast to CFCs, HFCs, HCFCs and HFEs contain one or more C – H bonds and so are susceptible to attack by OH radicals in the troposphere. As the result of a substantial research effort to deﬁne the environmental impact of HCFCs, HFCs, and HFEs we have a good understanding of the atmospheric degradation mechanism of such compounds [84, 85]. From a commercial point of view, the most important HFC is CF3 CFH2 (HFC-134a) which has found widespread use as a replacement for CFC-12 (CF2 Cl2 ) in domestic refrigeration and automobile air conditioning units. Annual emissions of CF3 CFH2 have increased rapidly from essentially zero in 1990 to approximately 85 000 tonnes in 2000 [86]. HFEs are used in much smaller quantities than HFCs and ﬁnd applications as cleaning agents

Atmospheric Photooxidation of Gas Phase Air Pollutants

151

for electronic equipment, heat transfer agents in refrigeration systems, and carrier ﬂuids for lubricant deposition. 10.1 Gas-Phase Chemistry Reaction with OH radicals is the dominant loss process for all HFCs, HCFCs and HFEs, accounting for > 90% of the fate of these compounds. In the stratosphere photolysis and reaction with Cl and O (1 D) atoms make minor contributions to the overall loss. A substantial kinetic database exists concerning the reaction of OH radicals with HFCs, HCFCs and HFEs [22]. From this data atmospheric lifetimes can be calculated. Lifetimes range from 1 to 270 years and are listed in Table 6 along with those for CFC-11 and CFC-12 for comparison. Table 6 also lists the global warming potentials for these compounds, which reﬂect their relative ability to contribute to global warming [84, 87]. A generic scheme for the atmospheric oxidation of a C2 haloalkane is given in Fig. 6. Values in parentheses are order of magnitude lifetime estimates. Reaction with OH radicals gives a halogenated alkyl radical which reacts with O2 to give the corresponding peroxy radical (RO2 ). As discussed in previous sections, peroxy radicals can react with three important trace species in the atmosphere: NO, NO2 , and HO2 radicals.

Fig. 6 Generic scheme for the oxidation of a C2 halocarbon, X and Y represent Cl and, or, F. Values in parentheses are order of magnitude estimates for the lifetimes of the various species. Closed-shell species are enclosed in boxes. Radical species are denoted by ellipses

152

T.J. Wallington · O.J. Nielsen

Peroxy radicals react rapidly with NO2 to give alkyl peroxynitrates (RO2 NO2 ). By analogy to the measured rate of reaction of CF2 ClO2 and CF3 CH2 O2 radicals with NO2 [84] the lifetime of RO2 radicals with respect to reaction with NO2 is approximately 10 minutes. Alkyl peroxynitrates are thermally unstable and decompose to regenerate RO2 radicals and NO2 . At room temperature in one atmosphere of air the peroxynitrates derived from HCFC-22 and HFC-134a have lifetimes of 24 seconds and < 90 seconds, respectively. Thermal decomposition dominates the atmospheric chemistry of halogenated alkyl peroxynitrates. The lifetime of CX3 CXYO2 radicals with respect to reaction with HO2 has been estimated to be 2–8 minutes [84]. The reaction of peroxy radicals with HO2 radicals gives hydroperoxides and, in some cases, carbonyl products. Product data are available for two haloperoxy radicals: CH2 FO2 and CF3 CFHO2 . Reaction of CH2 FO2 radicals with HO2 gives 30% yield of the hydroperoxide, CH2 FOOH, and 70% yield of the carbonyl product, HC(O)F [35]. In the reaction of CF3 CFHO2 with HO2 radicals less than 5% of the products appear as the carbonyl CF3 C(O)F and, by inference, > 95% of the reaction proceeds to give the hydroperoxide CF3 CFHOOH or the alkoxy radical CF3 CFHO [88]. The factors which determine the relative importance of the hydroperoxide and carbonyl forming channels are unknown. The hydroperoxide CX3 CXYOOH is expected to be returned to the CX3 CXYOx radical pool via reaction with OH and photolysis. The fate of the carbonyl product CX3 C(O)X produced in the CX3 CXYO2 + HO2 reaction is discussed later. The peroxy radicals derived from HFCs, HCFCs and HFEs react rapidly with NO to give NO2 and an alkoxy radical RO. The lifetime of peroxy radicals with respect to reaction with NO is approximately 3–7 minutes. Numerous product studies of halocarbon oxidation have shown that the atmospheric fate of the alkoxy radical, CX3 CXYO, is either decomposition or reaction with O2 [84]. Decomposition can occur either by C – C bond ﬁssion or Cl atom elimination. Reaction with O2 is only possible when an α-H atom is available (e.g. in CF3 CFHO). In the case of the alkoxy radicals derived from HFC32, HFC-125, and HCFC-22, only one reaction pathway is available. Hence, CHF2 O radicals react with O2 to give C(O)F2 , CF3 CF2 O radicals decompose to give CF3 radicals and C(O)F2 , and CF2 ClO radicals eliminate a Cl atom to give C(O)F2 . The alkoxy radicals derived from HFC-143a, HCFC-123, HCFC124, HCFC-141b, and HCFC-142b all have two or more possible fates, but one loss mechanism dominates in the atmosphere. For HCFCs 123 and 124 the dominant process is elimination of a Cl atom to give CF3 C(O)Cl and CF3 C(O)F, respectively. For HFC-143a, HCFC-141b, and HCFC-142b reaction with O2 dominates, giving CF3 CHO, CFCl2 CHO, and CF2 ClCHO respectively. The case of HFC-134a is the most complex. Under atmospheric conditions, the alkoxy radical derived from HFC-134a, CF3 CFHO, decomposes (to give CF3 radicals and HC(O)F) and reacts with O2 (to give CF3 C(O)F and HO2 radicals) at comparable rates. In the atmosphere 7–20% of the CF3 CFHO rad-

Atmospheric Photooxidation of Gas Phase Air Pollutants

153

icals formed in the CF3 CFHO2 + NO reaction react with O2 to form CF3 C(O)F while the remainder decompose to give CF3 radicals and HC(O)F [89]. Before moving on to consider the fate of the carbonyl products, it is appropriate to discuss the atmospheric fate of CF3 O radicals. The usual modes of alkoxy radical loss are not possible for CF3 O radicals. Reaction with O2 and decomposition via F atom elimination are both thermodynamically impossible under atmospheric conditions. Instead, CF3 O radicals react with NO and hydrocarbons. CF3 O + NO → C(O)F2 + FNO CF3 O + CH4 → CF3 OH + CH3

(53) (54)

Reaction with NO yields C(O)F2 . C(O)F2 does not react with any gas-phase trace atmospheric species and its photolysis is slow [90]. C(O)F2 is removed from the atmosphere by incorporation into water droplets and hydrolysis to give CO2 and HF and by photolysis in the upper stratosphere to give FCO radicals and F atoms. FNO photolyzes to give NO and a F atom [91]. F atoms reversibly form FO2 radicals by combining with O2 , and also react with CH4 and H2 O to give HF which will be rained out of the atmosphere. The reaction of CF3 O radicals with hydrocarbons such as CH4 produces CF3 OH. The CF3 O-H bond is unusually strong (120 kcal mole–1 ). CF3 OH is not attacked by any trace atmospheric radical [92] and is not photolyzed [93, 94]. CF3 OH undergoes heterogeneous decomposition to give C(O)F2 and HF and reaction with atmospheric water droplets to give CO2 and HF [95, 96]. 10.2 Reactions of Halogenated Carbonyl Intermediates Thus far the oxidation of the halocarbons into halogenated carbonyl products has been discussed. While the gas-phase oxidation mechanisms are complex, the carbonyl products are well established and are given in Table 7. The carbonyl products represent a convenient break point in our discussion. The sequence of gas-phase reactions that follow from the initial attack of OH radicals on the parent halocarbon are sufﬁciently rapid that heterogeneous and aqueous processes play no role. In contrast, the lifetimes of the carbonyl products [e.g., HC(O)F, C(O)F2 , CF3 C(O)F] are relatively long. As discussed in the following section, incorporation into water droplets followed by hydrolysis plays an important role in the removal of halogenated carbonyl compounds [97]. In the case of HC(O)F, C(O)F2 , FC(O)Cl, and CF3 C(O)F reaction with OH radicals [98] and photolysis [90] are too slow to be of any signiﬁcance. These compounds are removed entirely by incorporation into water droplets. The gas-phase oxidation mechanism for CX3 C(O)H and CF3 C(O)Cl is shown in Fig. 7. For CX3 C(O)H species reaction with OH radicals is import-

154

T.J. Wallington · O.J. Nielsen

Table 7 Gas-phase atmospheric degradation products of HFCs and HCFCs [84] Compound

Carbon containing degradation products

HFC-23 (CF3 H) HFC-32 (CH2 F2 ) HFC-125 (CF3 CF2 H) HFC-134a (CF3 CFH2 )

C(O)F2 , CF3 OH C(O)F2 C(O)F2 , CF3 OH HC(O)F, CF3 OH, C(O)F2 , CF3 C(O)F

HFC-143a (CF3 CH3 ) HFC-152a (CH3 CHF2 ) HFC-227ea (CF3 CFHCF3 ) HFC-236fa (CF3 CH2 CF3 ) HCFC-22 (CHF2 Cl)

CF3 C(O)H, CF3 OH, C(O)F2 , CO2 COF2 CF3 C(O)F, CF3 OH, C(O)F2 CF3 C(O)CF3 C(O)F2

HCFC-123 (CF3 CCl2 H) HCFC-124 (CF3 CFClH) HCFC-141b (CFCl2 CH3 ) HCFC-142b (CF2 ClCH3 )

CF3 C(O)Cl, CF3 OH, C(O)F2 , CO CF3 C(O)F CFCl2 CHO, C(O)FCl, CO, CO2 CF2 ClCHO, C(O)F2 , CO, CO2

Fig. 7 Mechanism for the degradation of CX3 C(O)H species

ant [99]. The lifetimes of CF3 C(O)H, CF2 ClC(O)H, and CFCl2 C(O)H with respect to OH attack have been estimated to be 24, 19, and 11 days, respectively [99]. Photolysis is probably also an important sink for CF3 C(O)H,

Atmospheric Photooxidation of Gas Phase Air Pollutants

155

CF2 ClC(O)H, and CFCl2 C(O)H [99]. However, while the absorption spectra for these compounds are known, their photolysis quantum yields are unknown and so it is only possible to establish upper limits for their rates of photolysis. Finally, scavenging by water droplets also probably plays a role in the atmospheric fate of these halogenated aldehydes and needs to be investigated. For CF3 C(O)Cl, reaction with OH is not feasible. Photolysis of CF3 C(O)Cl is important [100] and competes with incorporation of CF3 C(O)Cl into water droplets. As shown in Fig. 7, photolysis of CF3 C(O)Cl gives CF3 , CO, and Cl. In addition, trace amounts (< 1% yield) of CF3 Cl were reported. CF3 Cl is a longlived compound that efﬁciently transports chlorine from the lower atmosphere to the stratosphere. However, the low yield of CF3 Cl from CF3 C(O)Cl photolysis renders this pathway of negligible environmental signiﬁcance. Following reaction with OH radicals, CF3 C(O), CF2 ClC(O), and CFCl2 C(O) radicals can either react with O2 , or decompose to give CO and a halogenated methyl radical. Reaction with O2 is essentially the sole atmospheric fate of CF3 C(O) radicals [101, 102] and possibly CF2 ClC(O) and CFCl2 C(O) radicals. The resulting CX3 C(O)O2 radical can react with NO or NO2 . Reaction with NO2 gives a halogenated acetyl peroxynitrate which undergoes thermal decomposition [101, 102] to regenerate CX3 C(O)O2 . Reaction with NO gives a CX3 C(O)O radical which rapidly dissociates to give CX3 radicals and CO2 [102]. 10.3 Heterogeneous and Aqueous Phase Chemistry The ﬁnal step in removal of any species from the atmosphere involves heterogeneous deposition to the Earth’s surface. Removal processes include wet deposition via rain-out (following uptake into tropospheric clouds) and dry deposition to the Earth’s surface, principally to the oceans. The rates of these processes are largely determined by the species’ chemistries in aqueous solution. Heterogeneous lifetimes of the parent HFCs, HCFCs and HFEs are of the order of hundreds of years because of their low aqueous solubility and reactivity. The species listed in Table 8 are degradation products of the parent HFC and HCFC compounds that have removal rates in the gas phase (via reaction or photolysis) that are slow enough (days or longer) that heterogeneous processing might be signiﬁcant. All the halogen containing species are thought to undergo aqueous interactions that are fast enough for efﬁcient wet and dry deposition [103]. Estimates of tropospheric lifetimes for heterogeneous uptake into clouds and into the ocean are given in Table 8. For tropospheric cloud processing, the lower limit of 5 days is indicative of atmospheric transport limitations, i.e., the time taken to transport the species into the clouds.

156

T.J. Wallington · O.J. Nielsen

Table 8 Aqueous-phase atmospheric degradation products of HFCs and HCFCs [84]

Compound C(O)F2 C(O)ClF CF3 C(O)F CF3 C(O)Cl HC(O)F

Lifetime Clouds (days)

Ocean (years)

5–10 5–20

0.3–1.5 0.5–5.0

5–15 5–30 150–1500

0.3–3.0 1.0–9.0 80

Degradation products HF, CO2 HF, HCl, CO2 CF3 C(O)OH, HF CF3 C(O)OH, HCl HF, HCOOH

As seen from Table 8, heterogeneous removal of halocarbonyl species is rapid and is dominated by tropospheric cloud rain-out. A substantial body of data concerning the atmospheric degradation of HFCs, HCFCs and HFEs is available [84]. While some uncertainties exist, the current understanding of the atmospheric degradation of the commercially important HFCs, HCFCs and HFEs is well established. HFCs have no impact on stratospheric ozone. HCFCs have small but non-negligible ozone depletion potentials. The direct global warming potentials of HFCs, HCFCs and HFEs are approximately an order of magnitude less than those of the CFCs they replace. Finally, HFCs, HCFCs and HFEs are sufﬁciently unreactive and are released in such small quantities that they do not contribute to urban smog formation [104].

11 Research Needs There are several areas where further research is needed to better deﬁne the atmospheric degradation mechanisms and hence environmental impact of anthropogenic compounds. In general, there is a fairly complete database concerning the kinetics of the reactions which initiate the oxidation of pollutants. Extensive databases, structure activity relationships, and predictive techniques are available for the reaction of most anthropogenic molecules with OH and NO3 radicals and O3 . When kinetic data are available for other members of the class, the predictive techniques generally provide reliable (within a factor of 2) estimates of kinetic data for new compounds. However, when the compound is a member of a class of compounds for which no kinetic data exist, the predictive techniques provide estimates which are less reliable (uncertain by typically a factor of 5). Our understanding of the subsequent reaction mechanisms and the identity of the oxidation products is

Atmospheric Photooxidation of Gas Phase Air Pollutants

157

much less well understood. Compared to kinetic studies there have been relatively few studies of the products of the atmospheric degradation reactions and there are large uncertainties associated with the atmospheric oxidation mechanism of common compounds, e.g., the aromatics. While there is a growing database concerning peroxy radical atmospheric chemistry, signiﬁcant uncertainties remain. Our understanding of the mechanism of reaction of peroxy with HO2 radicals and the factors determining the nitrate yields in reactions of peroxy radicals with NO are very limited. Also, there are surprisingly few available data concerning factors which inﬂuence the atmospheric fate of alkoxy radicals. The vast bulk of the available data concerning the atmospheric fate of alkoxy radicals is derived from experiments at room temperature and 700–760 Torr. Studies as a function of temperature and pressure are needed. Finally, it is becoming apparent that the formation of particulate matter in the atmosphere can have an adverse impact on human health [105]. Work is needed to better deﬁne the importance of gas-to-particle conversion in polluted urban air and the impact of particulate matter on the photooxidation of gas-phase air pollutants [106, 107].

References 1. 2. 3. 4. 5. 6. 7. 8. 9.

10. 11. 12. 13. 14. 15. 16. 17. 18.

Francisco JS, Goldstein AN, Li Z, Zhao Y (1990) J Phys Chem 94:4791 Tromp TK, Ko MKW, Rodriguez JM, Sze ND (1995) Nature 376:327 Standard Atmosphere, NOAA, NASA, USAF, Washington D.C. 1976 Wallington TJ, Nielsen OJ (1997) Peroxy Radicals and the Atmosphere. In: Alfassi Z (ed) Peroxy Radicals. Wiley, New York Ravishankara AR, Lovejoy ER (1994) J Chem Soc Faraday Trans 90:2159 Dorn H-P, Brandenburger U, Brauers T, Ehhalt DH (1996) Geophys Res Lett 23:2537 Hofzumahaus A, Aschmutat U, Heßling M, Holland F, Ehhalt DH (1996) Geophys Res Lett 23:2541 Atkinson R (1997) J Phys Chem Ref Data 26:215 Crutzen PJ (1995) Ozone in the Troposphere, chapter in: Singh HB (ed) Composition, Chemistry, and Climate of the Atmosphere. Van Nostrand Reinhold, New York Platt U, Perner D, Winer AM, Pitts JN Jr (1980) Geophys Res Lett 7:89 Cantrell CA, Stockwell WR, Anderson LG, Busarow KL, Perner D, Schmeltekopf A, Calvert JG, Johnston HS (1985) J Phys Chem 89:139 Platt U, LeBras G, Burrows JP, Moortgat G (1990) Nature 348:147 Wayne RP, Barnes I, Biggs P, Burrows JP, Canosa-Mas CE, Hjorth J, LeBras G, Moortgat GK, Perner D, Poulet G, Restelli G, Sidebottom H (1991) Atmos Environ 25A:1 Becker E, Rahman MM, Schindler RN (1992) Ber Bunsenges Phys Chem 96:776 Atkinson R (1994) J Phys Chem Ref Data, Monograph 2 Wayne RP (1991) Chemistry of Atmospheres, 2nd edn. Oxford University Press, New York Intergovernmental Panel on Climate Change (IPCC) (1995) The Science of Climate Change. Cambridge University Press, New York Siegenthaler U, Joos F (1992) Tellus 44B:186

158

T.J. Wallington · O.J. Nielsen

19. Joos F, Bruno M, Fink R, Siegenthaler U, Stocker T, Le Quéré C, Sarmiento JL (1996) Tellus 48B:397 20. Prinn RG, Weiss RF, Miller BR, Huang J, Alyea FN, Cunnold DM, Fraser PJ, Hartley DE, Simmonds PG (1995) Science 269:187 21. Intergovernmental Panel on Climate Change (IPCC) (2001) Climate Change 2001 The Scientiﬁc Basis. Cambridge University Press, New York 22. DeMore WB, Sander SP, Golden DM, Hampson RF, Kurylo MJ, Howard CJ, Ravishankara AR, Kolb CE, Molina MJ (1994) JPL Publication 94-26 23. Spivakovsky CM, Logan JA, Montzka SA, Balkanski YL, Foreman-Fowler M, Jones DBA, Horowitz LW, Fusco AC, Brenninkmeijer CAM, Prather MJ, Wofsy SC, McElroy MB (2000) J Geophys Res 105:8931 24. Atkinson R, Baulch DL, Cox RA, Hampson RF, Kerr JA, Troe J (1992) J Phys Chem Ref Data 21:1125 25. Atkinson R (1989) J Phys Chem Ref Data, Monograph No 1 26. Wang YH, Jacob DJ (1998) J Geophys Res 103:31123 27. Graedel TE, Hawkins D, Claxton LD (1986) Atmospheric Chemical Compounds. Academic Press, New York 28. Smith IWM, Ravishankara AR (2002) J Phys Chem A 106:4798 29. Sulbaek Andersen MP, Hurley MD, Wallington TJ, Ball JC, Martin JW, Ellis DA, Mabury SA (2003) Chem Phys Lett 381:14 30. Sulbaek Andersen MP, Hurley MD, Wallington TJ, Ball JC, Martin JW, Ellis DA, Mabury SA, Nielsen OJ (2003) Chem Phys Lett 379:28 31. Martin JW, Smithwick MM, Braune BM, Hekstra PF, Muir DCG, Mabury SA (2004) Environ Sci Technol 38:373 32. Madronich S, Calvert JG (1990) J Geophys Res 95:5697 33. Wallington TJ, Japar SM (1990) Chem Phys Lett 167:513 34. Wallington TJ, Japar SM (1990) Chem Phys Lett 166:495 35. Wallington TJ, Hurley MD, Schneider WF, Sehested J, Nielsen OJ (1994) Chem Phys Lett 218:34 36. Wallington TJ, Hurley MD, Schneider WF (1996) Chem Phys Lett 251:164 37. Burrows JP, Moortgat GK, Tyndall GS, Cox RA, Jenkin ME, Hayman GD, Veyret B (1989) J Phys Chem 93:2375 38. Wallington TJ, Hurley MD, Ball JC, Jenkin ME (1993) Chem Phys Lett 211:41 39. Lightfoot PD, Cox RA, Crowley JN, Destriau M, Hayman GD, Jenkin ME, Moortgat GK, Zabel F (1992) Atmos Environ 26A:1805 40. Crutzen PJ (1979) Ann Rev Earth Planet Sci 7:443 41. Singh HB, Hanst PL (1981) J Geophys Res 8:941 42. Roberts JM (1995) Reactive odd-Nitrogen (NOy ) in the Atmosphere. In: Singh HB (ed) Composition, Chemistry, and Climate of the Atmosphere. Van Nostrand Reinhold, New York 43. Altshuller AP (1993) J Air Waste Manag Assoc 43:1221 44. Atkinson R, Aschmann SM, Carter WPL, Winer AM, Pitts JN Jr (1982) J Phys Chem 86:4563 45. Pate CT, Finlayson BJ, Pitts JN Jr (1974) J Am Chem Soc 96:6554 46. Becker KH, Geiger H, Wiesen P (1991) Chem Phys Lett 184:256 47. Lohr LL, Barker JR, Shroll RM (2003) J Phys Chem A 107:7429 48. Barker JR, Lohr LL, Shroll RM, Reading S (2003) J Phys Chem A 107:7434 49. Orlando JJ, Tyndall GS, Wallington TJ (2003) Chem Rev 103:4657 50. Jeffries HE (1995) Photochemical Air Pollution. In: Singh HB (ed) Composition, Chemistry, and Climate of the Atmosphere. Van Nostrand Reinhold, New York

Atmospheric Photooxidation of Gas Phase Air Pollutants 51. 52. 53. 54. 55.

56. 57. 58. 59. 60. 61. 62. 63. 64. 65.

66. 67. 68. 69.

70. 71. 72.

73. 74.

75. 76. 77. 78. 79. 80.

81. 82. 83.

159

Carter WPL, Atkinson R (1989) Environ Sci Tech 23:864 Derwent RG, Jenkin ME (1991) Atmos Environ 25A:1661 Simpson D (1995) J Atmos Chem 20:163 Derwent RG, Jenkin ME, Saunders SM (1996) Atmos Environ 30:181 Calvert JG, Atkinson R, Kerr JA, Madronich S, Moortgat GK, Wallington TJ, Yarwood G (2000) The Mechanisms of atmospheric oxidation of the alkenes. Oxford University Press, New York Niki H, Maker PD, Savage CM, Breitenbach LP (1981) Chem Phys Lett 80:499 Niki H, Savage CM, Breitenbach LP (1978) J Phys Chem 82:135 Atkinson R, Tuazon EC, Carter WPL (1985) Int J Chem Kinet 17:725 Orzechowska GE, Paulson SE (2002) Atmos Environ 36:571 Niki H, Maker PD, Savage CM, Breitenbach LP, Hurley MD (1987) J Phys Chem 91:941 Berndt T, Böge O (1994) Ber Bunsenges Phys Chem 98:869 Berndt T, Böge O (1995) J Atmos Chem 21:275 Wallington TJ, Dagaut P, Kurylo MJ (1992) Chem Rev 92:667 Atkinson R, Arey J (1994) Environ Health Perspect 102(4):117 Calvert JG, Atkinson R, Becker KH, Kamen RM, Seinfeld JH, Wallington TJ, Yarwood G (2002) The Mechanisms of atmospheric oxidation of aromatic hydrocarbons. Oxford University Press, New York Nozière B, Lesclaux R, Hurley MD, Dearth MA, Wallington TJ (1994) J Phys Chem 98:2864 Zellner R, Fritz B, Preidel M (1985) Chem Phys Lett 121:412 Knispel R, Koch R, Siese M, Zetzsch C (1990) Ber Bunsenges Phys Chem 94:1375 Klotz B, Volkamer R, Hurley MD, Sulbaek Andersen MP, Nielsen OJ, Barnes I, Imamura T, Wirtz K, Becker KH, Platt U, Wallington TJ, Washida N (2002) Phys Chem Chem Phys 4:4399 Crutzen PJ, Andreae MO (1990) Science 250:1669 Leighton P (1961) Photochemistry of Air Pollution. Academic Press, New York Cantrell CA, Lind JA, Shetter RE, Calvert JG, Goldan PD, Kuster W, Fehsenfeld FC, Montzka SA, Parrish DD, Williams EJ, Buhr MP, Westberg HH, Allwine G, Martin R (1992) J Geophys Res 97:20671 Parrish DD, Trainer M, Williams EJ, Fahey DW, Hübler G, Eubank CS, Liu SC, Murphy PC, Albritton DL, Fehsenfeld F (1986) J Geophys Res 91:5361 Ando J (1994) SO2 and NOx Abatement Energy Economics and Inﬂuences on Greenhouse Gases and Acid Rain. In: Calvert JG (ed) The Chemistry of the Atmosphere: Its Impacts on Global Change. Blackwell Scientiﬁc Publications, Oxford Eriksson E (1963) J Geophys Res 68:4001 Granat l, Rohde H, Halberg RO (1976) Ecol Bull SCOPE Report 7, 22:89 Wine PH, Thompson RJ, Ravishankara AR, Semmes DH, Gump CA, Torabi A, Nicovich JM (1984) J Phys Chem 88:2095 Martin LR, Damschen DE (1981) Atmos Environ 15:1615 Charlson RJ, Langner J, Rodhe H, Leovy CB, Warren SG (1991) Tellus 43B:152 Gallagher MW, Downer RM, Chourlarton TW, Gay MJ, Stromber I, Mills CS, Radojevic M, Tyler BJ, Bandy BJ, Penkett SA, Davis TJ, Dollard GJ, Jones BMR (1991) Atmos Environ 25A:2029 Lovelock JE, Maggs RJ, Wade RJ (1973) Nature 241:194 Molina M, Rowland FS (1974) Nature 249:810 World Meteorological Organization Global Ozone Research and Monitoring Project (1989) Report No. 20; Scientiﬁc Assessment of Stratospheric Ozone, vol 1

160

T.J. Wallington · O.J. Nielsen

84. Wallington TJ, Schneider WF, Worsnop DR, Nielsen OJ, Sehested J, DeBruyn WJ, Shorter JA (1994) Environ Sci Tech 28:320A 85. Francisco JS, Good DA (2003) Chem Rev 103:4999 86. McCulloch A, Midgley PM, Ashford P (2003) Atmos Environ 37:889 87. Pinnock S, Shine KP, Smyth TJ, Hurley MD, Wallington TJ (1995) J Geophys Res 100:23227 88. Maricq MM, Szente JJ, Hurley MD, Wallington TJ (1994) J Phys Chem 98:8962 89. Wallington TJ, Hurley MD, Fracheboud JM, Orlando JJ, Tyndall GS, Sehested JJ, Møgelberg TE (1996) J Phys Chem 100:18116 90. Nölle A, Heydtmann H, Meller R, Schneider W, Moortgat GK (1992) Geophys Res Lett 19:281 91. Wallington TJ, Schneider WF, Szente JJ, Maricq MM, Sehested J, Nielsen OJ (1995) J Phys Chem 99:984 92. Schneider WF, Wallington TJ (1993) J Phys Chem 97:12783 93. Schneider WF, Wallington TJ, Minschwaner K, Stahlberg EA (1995) Environ Sci Tech 29:247 94. Molina LT, Molina MJ (1996) Geophys Res Lett 23:563 95. Wallington TJ, Schneider WF (1994) Environ Sci Tech 28:1198 96. Lovejoy ER, Huey LG, Hanson DR (1995) J Geophys Res 100:18775 97. DeBruyn W, Duan SX, Shi XQ, Davidovits P, Worsnop DR, Zahniser MS, Kolb CE (1992) Geophys Res Lett 19:1939 98. Wallington TJ, Hurley MD (1993) Environ Sci Tech 27:1448 99. Scollard DJ, Treacy JJ, Sidebottom HW, Balestra-Garcia C, Laverdet G, LeBras G, MacLeod H, Téton S (1993) J Phys Chem 97:4683 100. Rattigan OV, Wild O, Jones RL, Cox AR (1993) J Photochem Photobiol A: Chemistry 73:1 101. Zabel F, Kirchner F, Becker KH (1994) Int J Chem Kinet 26:827 102. Wallington TJ, Sehested J, Nielsen OJ (1994) Chem Phys Lett 226:563 103. Wine PH, Chameides WL (1989) World Meteorological Organization Global Ozone Research and Monitoring Project, Report No. 20; Scientiﬁc Assessment of Stratospheric Ozone, vol 1, p 271–295 104. Hayman GD, Derwent RG (1997) Environ Sci Tech 31:327 105. Schwartz J, Dockery DW, Neas LM (1996) J Air Waste Manage Assoc 46:927 106. Odum JR, Jungkamp TPW, Grifﬁn RJ, Flagan RC, Seinfeld JH (1997) Science 276:96 107. Finlayson-Pitts BJ, Pitts JN Jr (1986) Atmospheric Chemistry: Fundamentals and Experimental Techniques. Wiley, New York

Hdb Env Chem Vol. 2, Part M (2005): 161–192 DOI 10.1007/b138183 © Springer-Verlag Berlin Heidelberg 2005 Published online: 16 September 2005

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols and Anilines Gottfried Grabner1 (u) · Claire Richard2 1 Max

F. Perutz Laboratories, Department of Chemistry, University of Vienna, Campus-Vienna-Biocenter 5, 1030 Wien, Austria [email protected]

2 Laboratoire

de Photochimie Moléculaire et Macromoléculaire, UMR 6506 CNRS, Université Blaise Pascal, Ensemble Universitaire des Cézeaux, 63177 Aubière Cedex, France [email protected]

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

162

2 2.1 2.1.1 2.1.2 2.2

Halogenated Phenols . . . . . . . . . . . . . . . . . . Studies in Solution . . . . . . . . . . . . . . . . . . . Monohalogenated Phenols . . . . . . . . . . . . . . . Polyhalogenated Phenols . . . . . . . . . . . . . . . . Studies in Heterogeneous or Environmental Systems

. . . . .

163 163 163 168 170

3 3.1 3.2

Halogenated Anilines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monohalogenated Anilines . . . . . . . . . . . . . . . . . . . . . . . . . . Dihalogenated Anilines . . . . . . . . . . . . . . . . . . . . . . . . . . . .

172 172 178

4 4.1 4.2 4.3

Biocides . . . . . . . . . . . . . . . . . . . . . . . Phenoxyacetic Acid Derivatives . . . . . . . . . . Phenylurea Derivatives and Related Compounds Other Biocides . . . . . . . . . . . . . . . . . . .

. . . .

178 178 182 185

5

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

188

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

189

. . . .

. . . .

. . . . .

. . . .

. . . . .

. . . .

. . . . .

. . . .

. . . . .

. . . .

. . . . .

. . . .

. . . . .

. . . .

. . . . .

. . . .

. . . . .

. . . .

. . . . .

. . . .

. . . . .

. . . .

. . . . .

. . . .

Abstract The halogenated phenols and anilines are the parent chromophores of a number of widely used pesticides, among them phenoxyacetic acid and phenylurea derivatives, and of other biocides. The direct photolysis of these substances in the environment has become a subject of increasing interest. In the ﬁrst part of this review, investigations of the photochemistry of halogenated phenols and anilines carried out during the last decade are summarized. The second part is concerned with corresponding studies on biocides. The phototransformation mechanisms of these compounds have been found to share many common features. In particular, two kinds of heterolytic dehalogenation processes, aromatic carbon-halogen photohydrolysis on the one hand and formation of triplet carbenes or aryl cations on the other, have been described for many systems. The mechanistic knowledge gained in the investigations of model compounds is increasingly brought to bear on actual environmental situations.

162

G. Grabner · C. Richard

Keywords Environment · Halogenated aromatic compounds · Mechanism · Pesticides · Photochemistry

1 Introduction Halogenated aromatic compounds are among the most widespread ecosystem pollutants. Originating mainly from industrial production and from agrochemical use, they are ubiquitously present in soils and waters, with potential detrimental effects on the biosphere. The knowledge of their fate in the environment is therefore of paramount importance, and has been the subject of investigation for a long time [1]. Many of the substances involved are light absorbing in the near-UV range; they are therefore susceptible to photochemical transformation under the inﬂuence of solar irradiation. The study of the direct photochemistry of pollutants (as opposed to indirect, or sensitized, photochemical action) in the environment is a relatively recent subject, which is, however, increasingly attracting attention [2–4]. This is due to the accumulation of basic knowledge on photochemical mechanisms on the one hand, and to the progress of experimental techniques in environmental chemistry on the other. A related approach to the subject originates from inquiries into the possibility of pollutant abatement by photoirradiation techniques; although these studies frequently use radiation of higher energy (UV-C), their mechanistic ﬁndings may be of interest for the interpretation of lower-energy irradiation experiments. The present review is the update and extension of an earlier one dealing with the phototransformation of phenol and its derivatives [5]. We will now focus on the photochemistry of a particular class of halogenated aromatics, phenols or anilines containing one or more halogen substituents on the ring, and close derivatives thereof, such as phenoxy compounds or phenylureas. Belonging to this group of compounds are industrial pollutants, biocides, and their degradation products [6]. The study of the basic photochemistry of these substances is not a very old subject either, and much knowledge has been added during the last decade. At the same time, researchers have begun to investigate the photobehavior of biocides under simulated or realistic environmental conditions. Increasingly, these studies proceed from photodegradation to phototransformation, i.e. from merely measuring the disappearance of a given compound to identifying the products and the pathways leading to them, based on mechanistic photochemical knowledge gained under laboratory conditions. In the following, we will mainly refer to model studies carried out in aqueous solutions; however, it should be kept in mind that the behavior in other solvents or in heterogeneous model systems may be relevant for environmental conditions, which cannot always be likened to aqueous systems.

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

163

The halogenated phenols and anilines, as well as their derivatives, share some common chemical and spectroscopic properties. They undergo prototropic equilibria, which may profoundly alter their photochemical behavior. In the following, we will limit ourselves to the presentation of data obtained at pH values of environmental relevance; in particular, we will not discuss the photochemistry of the anionic forms of monohalogenated phenols, which require alkaline media. The absorption spectra of halogenated phenols and anilines are limited to the far-UV, in general not extending further than 320 nm; the overlap with the solar spectrum at ground level is therefore weak. However, the potential impact of direct photolysis in the environment basically depends on its efﬁciency relative to that of competing pathways, including indirect photoinduced transformation and biotransformation. For some of the pesticides considered in this review, biodegradation has been shown to be slow and incomplete [7]. As will be described below, the quantum yields of direct photolysis of the same compounds may be substantial. Consequently, direct photolysis may signiﬁcantly contribute to transformation of these biocides in the environment. Some of the topics addressed here have been reviewed by other authors. Studies of photoinduced processes (direct or indirect) of chlorophenols carried out before 1998 have been covered comprehensively [8]; this work will only cursorily be treated here. A detailed overview of the photochemical behavior of phenylurea herbicides was in press at the time of writing this article [9]. The related subject of the photodegradation of pharmaceuticals in the aquatic environment has been reviewed very recently [10].

2 Halogenated Phenols 2.1 Studies in Solution 2.1.1 Monohalogenated Phenols 2.1.1.1 ortho-halogenated Phenols Earlier studies [11, 12] have shown that photoexcitation of aqueous 2-chlorophenol or 2-bromophenol leads to contraction of the aromatic cycle to give cyclopentadienic acids (an example of a Wolff rearrangement [13]), and to substitution of the halogen by OH (photohydrolysis), with moderate quantum yields (Φ = 0.01–0.04). A carbene (2-oxocyclohexa-3,5-dienylidene) was suggested as a possible intermediate in the ring contraction pathway [11].

164

G. Grabner · C. Richard

The photoinduced degradation of 2-chlorophenol dissolved in surfactant solutions (Brij 35 or sodium dodecyl sulfate) was investigated by Shi et al. [14]. The formation of ring contraction products and of the photohydrolysis product catechol was conﬁrmed; additionally, unsubstituted phenol was found and attributed to a homolytic C – Cl dissociation producing an intermediate phenyl radical. Secondary photolysis processes were observed in the contraction pathway. The presence of surfactants increased the photodegradation yields. This was explained by the action of a micellar cage effect allowing efﬁcient sensitization of 2-chlorophenol photolysis by the photoproducts phenol and catechol, and by the H-donor function of the surfactants promoting phenol formation from phenyl radicals [14]. The intermediate species expected in the photocontraction pathway, fulvene 6-oxide (cyclopentadienylidenemethanone) and its hydration product, fulvene 6,6-diol, could be identiﬁed by nanosecond absorption spectroscopy of 2-chlorophenol [15]. The corresponding transients are characterized by absorption bands at 255 nm and 295 nm, respectively, and are in a precursorproduct relationship with k ≈ 106 s–1 ; they are insensitive to oxygen. These measurements gave no indication for the involvement of a carbene. In a subsequent study of aqueous 2-bromophenol, additional absorption bands were observed in the transient spectrum [16]. In oxygen-free solution, a previously unknown three-band species was detected (λmax = 388, 375 and 360 nm), which was converted into a product with λmax = 475 nm by reaction with O2 , and into the well-known phenoxyl radical (λmax = 400 and 380 nm) by reaction with aliphatic alcohols. The analogy of this reaction pattern with that of the 4-halophenol system (see below) allowed the unambiguous assignment of the unknown species to the carbene 2-oxocyclohexa-3,5-dienylidene in its triplet ground state. Since the Wolff rearrangement is expected to proceed on the excited singlet surface, a reaction scheme was proposed involving competitive photochemical pathways, both leading to heterolytic dehalogenation (Scheme 1) [16]. The absorption spectra of the carbenes produced from the 2-halophenols are shown in Fig. 1 along with those obtained from the 4-halophenols and 4haloanilines (see below). The carbene thus reacts with O2 to form an ortho-benzoquinone O-oxide, and with an aliphatic alcohol as H-donor to form a phenoxyl radical (plus an aliphatic radical not shown in Scheme 1). The ground state triplet electronic conﬁguration of this carbene accounts for its reaction behavior, in particular for the fact that it reacts very slowly with the solvent, H2 O. In agreement with the intrinsically faster intersystem crossing of 2-bromophenol compared to 2chlorophenol, the quantum yield of the carbene pathway was higher for the former (Φ = 0.04) than for the latter compound (Φ = 0.003). In contrast, the quantum yields of photocontraction were comparable (Φ = 0.04). The transient absorption data were conﬁrmed by photoproduct analysis, showing the formation of phenol from 4-bromophenol in the presence of H-donors [16].

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

165

Scheme 1 Primary steps of 2-halophenol photolysis in polar protic solvents

Fig. 1 Electronic absorption spectra of the carbenes produced by dehalogenation of 2-halophenols [16] (open circles), 4-halophenols [20] (full circles), and 4-haloanilines [55] (triangles)

A comparison between the photoreactions of 2-chlorophenol and 2-bromophenol in a low-temperature argon matrix was carried out by Akai et al. by means of IR spectroscopy [17, 18]. The formation of fulvene 6-oxide was evidenced in both systems. Homolytic C – Br cleavage was found as an additional pathway in the case of 2-bromophenol.

166

G. Grabner · C. Richard

The photochemistry of aqueous 2-iodophenol was the subject of a subsequent study [19]. Although ring contraction, carbene formation and photohydrolysis took place in this case too, they all were minor pathways (Φ < 0.01) in comparison with homolytic C – I cleavage forming phenyl radicals and I· atoms (Φ = 0.08), which could be evidenced via the reaction of the latter with added iodide to give I2 · – radical anions. It was concluded that the heterolytic dehalogenation mechanism observed for aqueous 2-chloro- and 2-bromophenol is overruled in the case of 2-iodophenol by a homolytic one promoted by the comparatively weaker carbon-halogen bond [19]. 2.1.1.2 para-halogenated Phenols The mechanism of the aqueous photochemistry of 4-chlorophenol has been reviewed earlier [5, 8]. Its basic features are the same as those of the carbene pathway described above for 2-bromophenol. The main differences are the fact that this is the only photolytic reaction of 4-chlorophenol and that its quantum yield is considerably higher than that of the 2-substituted analogues; Φ = 0.75 was obtained for the formation of the triplet carbene, 4-oxocyclohexa-2,5-dienylidene (λmax = 384 and 370 nm) from aqueous 4-chlorophenol (see Fig. 1) [20]. Photoproduct analysis yielded p-benzoquinone (in the presence of O2 ), phenol (in the presence of an alcohol), hydroquinone and isomeric chlorodihydroxybiphenyls, which could all be accounted for by carbene reactions [20]. The carbene mechanism of heterolytic dehalogenation of 4-chlorophenol was subsequently conﬁrmed by studies using ﬂash photolysis [21] and FTEPR [22]. A detailed account of the EPR measurements was published later [23], in which it was shown that the spin polarization of the phenoxylpropanoyl radical pair produced in the photolysis of 4-chlorophenol in 2-propanol is consistent with a triplet state precursor. The proposition that this precursor is the above-mentioned carbene was proved by generating the same radical pair, with identical spin polarization, by photolysis of p-benzoquinone diazide [22, 23]. The effect of the halogen substituent (ﬂuoro, chloro, bromo and iodo) on the yield and mechanism of 4-halophenol photolysis was investigated by Durand et al. [24]. Transient spectroscopy in aerated aqueous solutions indicated the formation of p-benzoquinone O-oxide from each derivative except 4-iodophenol for which no transients were detected; p-benzoquinone and hydroquinone were found as photoproducts for all four compounds. It was concluded that the carbene mechanism was valid for the whole series. Under continuous irradiation, the 4-halophenol degradation quantum yields were determined to be Φ = 0.31, 0.44, 0.08, and 0.022 in the series as stated above. The ﬂuorescence lifetimes decreased in the same order, from 2.1 ns for 4-ﬂuorophenol to 0.4 ns for 4-chlorophenol and < 0.1 ns for 4-bromophenol.

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

167

These results were interpreted with reference to an intramolecular heavy atom effect; the authors concluded the carbene is formed from the excited singlet state of the molecule, in competition to intersystem crossing. A direct proof of this statement could, however not be obtained [24]. It should be noted here that alternative explanations for the effect of halogen substitution on photolysis quantum yields have been put forward for the 4-haloanilines (see below). Further studies on the photodegradation of 4-chlorophenol were conducted, respectively, by combinations of ﬂash irradiation and product analysis [25], and by ﬂuorescence [26]. In the former study, a degradation quantum yield of Φ = 0.24 was reported. Vialaton et al. studied the phototransformation of aqueous 4-chloro-2methylphenol [27]. Analysis of the products formed upon direct irradiation yielded complete agreement with the mechanism of 4-chlorophenol photolysis. The dehalogenation quantum yield was found to be Φ = 0.66. The sensitization by humic substances was investigated in the same work. It was shown that it increased the degradation yield for λ > 300 nm by a mechanism involving oxidative attack by free radicals leading to ring opening products. Both carbene and radical pathways were found to operate in photolysis experiments in natural waters [27]. An attempt to clarify the ﬁrst stages of the mechanism of carbene formation was undertaken in a nanosecond transient absorption study of 5-chloro2-hydroxybenzonitrile (4-chloro-2-cyanophenol), based on the hypothesis

Scheme 2 Primary steps of 4-halophenol photolysis in polar protic solvents

168

G. Grabner · C. Richard

that introduction of an electron-attracting substituent on the ring might slow down some of the steps involved in dehalogenation [28]. Transients as well as photoproducts could be identiﬁed and conﬁrmed the validity of the carbene mechanism. The quantum yield of carbene formation in aqueous solution was determined to be Φ = 0.062. The yield of photodegradation was found to be inﬂuenced by added reactants: saturation with O2 reduced it to Φ = 0.038, while addition of bromide (2 M) or iodide (0.48 M) ions increased it to Φ = 0.20, indicating, respectively, quenching and enhancement of the formation of a molecular triplet state as a precursor of the carbene. In ethanolic solutions, the triplet lifetime of 5-chloro-2-hydroxybenzonitrile was long enough to allow the detection of the triplet-triplet absorption. Acrylamide quenching studies showed that the 2-cyanophenoxyl radicals arising from the reaction of the carbene with the solvent have the molecular triplet state as a precursor. It was concluded that intersystem crossing in the excited molecule precedes heterolytic dehalogenation [28]. Based on this evidence, the probable mechanism of the primary steps of 4-halophenol photolysis in polar protic solvents involves dehalogenation from the triplet state, as shown in Scheme 2. 2.1.2 Polyhalogenated Phenols Earlier work had shown that the quantum yield of dehalogenation upon direct excitation of polyhalophenols decreases when the number of halogen atoms increases [5]. This was conﬁrmed in a study comparing the degradation rates of 4-chlorophenol, 2,4-dichlorophenol and 2,4,6-trichlorophenol [29]. More recent work by Benitez et al. gave contradictory results in this respect. These authors applied a method allowing the estimation of photodegradation quantum yields under polychromatic UV irradiation produced by a high-pressure Hg lamp. For 2,4,6-trichlorophenol, the result was Φ = 0.0112 at pH 2, Φ = 0.0211 at pH 5, and Φ = 0.0307 at pH 7 and pH 9 [30]. In contrast, values of Φ = 0.20 at pH 3, Φ = 0.112 at pH 5, Φ = 0.0525 at pH 7 and Φ = 0.022 at pH 9 were obtained in the case of pentachlorophenol; the measured rate of degradation increased with increasing pH, however, but this was explained by an increase in solution absorbance with pH, overruling the decrease of the quantum yield [31]. It should be noted that the quantum yield values obtained in this study [31] are surprisingly high when compared with those of dihalogenated [5] or trihalogenated [30] phenols, and need to be conﬁrmed. Mechanistic studies of phototransformation of polyhalogenated phenols have been carried out by means of product analysis. A particular concern is the possibility of the formation of highly toxic polychlorinated dibenzodioxins (PCDDs) and dibenzofurans (PCDFs). In a detailed study of the photolysis of 2,4,5-trichlorophenol, already reviewed earlier [8], it was argued that ho-

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

169

molytic bond cleavage events producing a variety of chlorinated phenyl radicals are the initial steps which ultimately lead to PCDDs and PCDFs [32]. In contrast, oxidative dehalogenation initiated by singlet oxygen, 1 O2 , followed a different, probably heterolytic mechanism [32, 33]. An analogous behavior was obtained for pentachlorophenol [33, 34]. The phototransformation of pentachlorophenol was studied by Piccinini et al. by irradiating thin layers of the solid compound deposited in a photoreactor [35] with a high-pressure Hg lamp (λ > 290 nm). The photodegradation rate was considerably enhanced if a 50 v/v % methanol-water mixture was used as a deposition mixture instead of pure methanol. Tetrachlorophenols, hexachlorobenzene and pentachlorobenzene were found as major products along with chlorinated cyclopentadienes. PCDDs and PCDFs were also detected. The dehalogenation steps were attributed to homolytic cleavage processes [35]. Similar results were obtained by Hong et al. in aqueous solutions, who were able to identify a large number of products originating from reductive dechlorination on the one hand, and from oxidative processes on the other [36]. Production of PCDDs and PCDFs was conﬁrmed, although they appeared as minor products [36]. These were characterized in more detail by Zheng et al., who identiﬁed octachlorodibenzo-p-dioxin and octachlorodibenzofuran as the main components [37]. A correlative investigation of photodegradation and toxicity of aqueous pentachlorophenol was carried out by Ho and Bolton using a mediumpressure Hg lamp without ﬁlters [38]. A steady decrease in toxicity with progressing photolysis, corresponding to a continuous decrease in the average number of halogen ring substituents, was observed. The authors concluded that possible toxic products were absent, or were rapidly degraded during irradiation [38]. An entirely divergent result was obtained in a study of the photodegradation of aqueous 2,4,6-trichlorophenol using ﬁltered light of a xenon lamp (Xenotest 1200) [39]. These authors found that the toxicity of the solution increased during irradiation; using product analysis, 3,5-dichlorocatechol was found as a major product accumulating in the system in parallel to the increase in toxicity [39]. Comparing these two studies underlines the differences that may be encountered when different irradiation conditions are employed. Making use of the full UV lamp spectrum, as usually done in studies of pollutant abatement, may have the beneﬁcial effect of destroying potential toxic photoproducts much more efﬁciently than radiation limited to longer wavelengths, characteristic for studies of environmental photochemistry. It has to be stressed that the models proposed until now for the photodegradation of the polyhalophenols are based on indirect evidence only. As yet, real-time mechanistic studies may have been hindered by low aqueous solubilities and low quantum yields. In particular, this means that the mechanisms involving initial homolytic cleavage are not supported by direct evidence. The observation of preferential formation of 3,5-dichlorocatechol

170

G. Grabner · C. Richard

from 2,4,6-trichlorophenol [39], corresponding to photohydrolysis, indicates that heterolytic processes may also be involved. Further efforts to elucidate the photochemistry of the polyhalophenols are clearly needed. 2.2 Studies in Heterogeneous or Environmental Systems The last few years have seen the emergence of projects aiming at elucidating the photobehavior of monochlorophenols in heterogeneous systems. Two studies were concerned with the behavior of 4-chlorophenol in a surfaceadsorbed state, the substrates being silicalite and solid β-cyclodextrin [40], and cellulose and silica [41]. In both cases, nanosecond transient photolysis with diffuse reﬂectance detection and photoproduct analysis were the experimental techniques employed. The results of these investigations are instructive in demonstrating the inﬂuence of the solid support on the outcome of the photolytic reactions. Silicalite is a dealuminated analog of the zeolite ZSM-5, whereas βcyclodextrin is an oligomeric carbohydrate forming a toroidal cavity; both systems may be seen as prototypes of host-guest complexes. Phosphorescence emission was present in both cases, more intense in β-cyclodextrin, with lifetimes of a few 100 ns. The transient absorption spectra indicated the formation of 4-chlorophenoxyl radicals as the major intermediates in silicalite, whereas unsubstituted phenoxyl radicals were additionally detected in β-cyclodextrin. In this latter medium, phenol was detected as the main ﬁnal product [40]. Phenoxyl radical and phenol are the species expected from reaction of the putative carbene intermediate, 4-oxocyclohexa-2,5-dienylidene (see above) with the host serving as H-donor; however, the authors were reluctant to adopt this mechanism, since they were unable to detect the carbene itself or its reaction product with O2 , p-benzoquinone-O-oxide [40]. It should be noted, however, that this could well be due to the overwhelming efﬁciency of H abstraction from the host by the included carbene. In contrast, product analysis in silicalite surprisingly yielded ring opening species, such as 3-chloro-2-cyclopenten-1-one, as the major products; photodegradation yields were altogether lower in this medium compared with β-cyclodextrin. The authors speculated about electron photoejection being the source of 4chlorophenoxyl radicals [40]. In a subsequent study, Da Silva et al. investigated the photodegradation of 4-chlorophenol on the polysaccharide cellulose and on silica [41]. In both systems, transient spectra and products were consistent with the assumption of heterolytic dehalogenation to give 4-oxocyclohexa-2,5-dienylidene. In cellulose, unsubstituted phenoxyl radicals and phenol were the main products, indicating that the polymer serves as H-donor for the carbene. There was no effect of O2 on the reaction course, which was explained by the protective effect of the macromolecular structure. In silica, which contained

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

171

water and allowed O2 diffusion through its structure, the absorption of pbenzoquinone-O-oxide could be clearly identiﬁed, and p-benzoquinone and hydroquinone were detected as products, the latter being formed by reaction of the carbene with entrapped H2 O molecules [41]. This reaction is unable to compete with O2 addition in aqueous solution [20]; this again underlines the determining role of the solid support on the direction of the photodegradation process. The different pathways of heterogeneous photolysis of 4-chlorophenol are summarized in Scheme 3. The photochemistry of 2- and of 4-chlorophenol in water ice has been investigated by Klánová et al. by means of photoproduct analysis, motivated by the occurrence of halogenated pollutants in ice and snow media [42]. It was found that in solid ice solutions at – 10 ◦ C, chlorodihydroxybiphenyls were the major products from both chlorophenols. When the temperature was raised to – 5 ◦ C, catechol was the main product from 2-chlorophenol, whereas 5-chloro-2,4 -dihydroxybiphenyl remained the main product from 4-chlorophenol, with phenol and hydroquinone appearing in minor proportion. It was argued that the preponderance of dimeric products reﬂected the aggregation of the chlorophenols at the grain boundaries of polycrystalline ice. The drastic changes in product distribution upon raising the temperature were proposed to be due to formation of a quasi-liquid layer on the crystal surfaces, promoting photosolvolysis reactions. The quantum yields of photodegradation in solid ice, measured upon excitation at λ = 290 nm, were found to be Φ = 0.03 for 2-chlorophenol and Φ = 0.04 for 4-chlorophenol,

Scheme 3 Transients of 4-chlorophenol photolysis on solid supports

172

G. Grabner · C. Richard

i.e. of the same order of magnitude as those in liquid water for the former compound but one order of magnitude smaller for the latter one. In a detailed discussion of the reaction mechanism, homolytic cleavage processes were put forward as the most plausible primary mechanistic steps, since key products of aqueous photolysis such as cyclopentadienic acids (from 2chlorophenol) or p-benzoquinone (from 4-chlorophenol) were absent [42]. In a further study, the concentration of the chlorophenols in solid ice was lowered to 10–7 M, but dimeric products were still the major ones, underlining the decisive role of aggregation processes in this medium [43]. This intrinsic heterogeneity of solutions in solid ice was also shown to inﬂuence processes initiated by UV/H2 O2 and by β-irradiation in a determining way [43]. The role of photodegradation processes of pentachlorophenol in natural environments was the subject of several studies. Donaldson et al. studied the effect of water evaporation on the rates of photodegradation in sand and clay soils; it was shown that increased soil moisture inﬂuenced the amount of photolytic loss compared to loss by volatilization [44]. Puplampu and Dodoo investigated photodegradation in aqueous medium under tropical sunlight; they found that the yield of chloride was comparable to that obtained by irradiation at λ = 254 nm [45]. The formation of PCDDs by photolysis of pentachlorophenol adsorbed on different soils was measured by Liu et al.; octaand heptachlorodibenzo-p-dioxin were detected as products, the formation of which was inhibited by the presence of fulvic acids [46]. These studies concur in showing that photoinduced degradation may play a role for the environmental fate of pentachlorophenol, although they do not offer mechanistic insights.

3 Halogenated Anilines 3.1 Monohalogenated Anilines The photochemistry of the halogenated anilines had received little attention until about 10 years ago, when the elucidation of the mechanisms of the photolysis of the halophenols, as described above, raised the question of the more general applicability of the obtained results. As will be detailed below, there are indeed close similarities between the halogenated phenols and anilines, but characteristic differences also exist. The phototransformation of aqueous 2-chloroaniline was studied by Othmen and Boule by means of product analysis [47]; these authors later deepened and extended their work, including also 2-ﬂuoro- and 2-bromoaniline [48]. The range of photodegradation quantum yields was Φ = 0.02–0.04 for 2-chloro- and 2-bromoaniline, and Φ = 0.10–0.13 for ﬂuoroani-

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

173

line. Three main photoproducts were isolated from each of the three derivatives: 2-aminophenol as the major one, cyclopenta-1,3-diene-1-carbonitrile, and aniline. In addition, three isomeric diaminohalobiphenyls were identiﬁed from 2-bromoaniline, and, in minor proportion, from 2-chloroaniline too (see Scheme 4). The observation of mostly identical products from all congeners together with the fact that the ﬂuoro derivative showed the highest quantum yield was taken as a strong indication for the operation of a heterolytic dehalogenation mechanism. This conclusion was further strengthened by the failure to detect, by means of transient absorption spectroscopy, the formation of Br2 · – radicals in a solution of 2-bromoaniline containing excess Br– , which constitutes a sensitive test of the presence of Br atoms. This experiment set an upper limit of Φ < 0.005 for the homolytic cleavage of the C – Br bond [48]. In the discussion of the mechanism of product formation from the 2haloanilines, a strong similarity to the behavior of the 2-halophenols was noted. The two major monomeric products correspond to the processes of photohydrolysis and ring contraction which are also characteristic for this group of compounds (see above). The photohydrolysis reaction was proposed to be a concerted process of halide elimination and H2 O attack. Two possibilities were noted for the ring contraction process, the ﬁrst one similar to the mechanism observed for 2-chlorophenol, possibly via a singlet carbene, the other one involving the intermediate formation of phenylnitrene. No direct proof could be obtained for either mechanism, not least because it was impossible to observe any transients by nanosecond spectroscopy. The

Scheme 4 Products of 2-haloaniline photolysis in aqueous solution

174

G. Grabner · C. Richard

authors further proposed that the formation of aniline might reﬂect a reduction of a carbene, and that of the dimers a reaction of the 2-halophenol triplet state [48]. It may indeed be stated from the point of view of the present knowledge of 2-halophenol photochemistry that the formation of aniline and of the substituted biphenyls is compatible with the reactions of a putative triplet carbene as a photoreaction intermediate; there is, however, at present no immediate evidence to support this hypothesis. The photolysis of aqueous 3-chloroaniline was found to proceed via a clean photohydrolysis step to give 3-aminophenol with a quantum yield of Φ = 0.12 [49]. This behavior is analogous to that earlier observed for 3-chlorophenol [50] and chlorobenzene [51], and indicates either a concerted dehalogenation/hydrolysis process or a fast reaction of a primarily formed aminophenyl cation with H2 O (see Scheme 5). The mechanism of 3-haloaniline phototransformation was investigated in more detail in a nanosecond ﬂash photolysis and product analysis study in methanolic solution [52]. Photosolvolysis, as witnessed by the formation of anisidine, was again the major pathway, in particular for 3-ﬂuoroaniline. A photoreductive pathway yielding aniline was also observed, being most important for 3-bromoaniline; transient absorption indicated the formation of anilino radicals by singlet state dehalogenation as an intermediate on this pathway. Homolytic C – Br cleavage was additionally evidenced. The possibility to observe the triplet-triplet absorption spectra of 3-ﬂuoro- and 3-chloroaniline pointed to an unreactive triplet state [52]. The photochemistry of 4-chloroaniline had previously been investigated in organic solvents by Latowski, who suggested electron ejection as the primary step in polar media [53]. However, the observation of benzidine as a major photoproduct in methanol/H2 O mixtures indicated the intermediate formation of unsubstituted anilino radicals, and this led the authors to suggest that the actual mechanism might be related to that observed for 4-chlorophenol [54]. A detailed study of the photochemistry of 4-chloroaniline was subsequently undertaken by means of nanosecond transient spectroscopy and product analysis [55]. It indeed turned out that 4-chlorophenol and 4-chloroaniline basically share a common photolytic mechanism. The main photointermediate detected in polar solvents was again a carbene with a triplet ground state, 4-iminocyclohexa2,5-dienylidene, which could be characterized by its absorption spectrum (λmax = 407 and 390 nm, see Fig. 1) and by its characteristic reactivity, which

Scheme 5 Mechanism of sequential photohydrolysis of the 3-haloanilines

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

175

is very similar to that of its 4-oxo counterpart [55]. The quantum yields of carbene formation were found to be Φ = 0.53 in neutral solution and Φ = 0.40 in acidic solution (pH 1–2). The results of the transient measurements were fully conﬁrmed by the detected photoproducts, which are summarized in Scheme 6 (the reaction steps leading to the products are not shown in detail). Under steady-state conditions, the dehalogenation quantum yields were Φ = 0.50 from 4-ﬂuoroaniline, Φ = 0.43 from 4-chloroaniline and Φ = 0.27 from 4-bromoaniline [55]. There is one major difference between the phenol and aniline systems, however, related to the fact that the amino nitrogen is able to accommodate a positive charge: in the 4-haloanilines the primary photolytic dehalogenation step generates an aminophenyl cation which can be seen as a tautomer of the protonated carbene (this renders the assignment of this species

Scheme 6 Carbene reactions and products of 4-chloroaniline photolysis in aqueous solution

176

G. Grabner · C. Richard

somewhat ambiguous, as discussed below). This cation is the species that is detected by transient spectroscopy, whereas the cation generated from 4-chlorophenol must be deprotonated very rapidly to give the neutral carbene [20]. For this reason, heterolytic dehalogenation of 4-chlorophenol requires a protic solvent, whereas this is not necessary for 4-chloroaniline, as witnessed by the fact that the mechanism is basically the same in alcohols and in acetonitrile. For the same reason, 4-chloroaniline and 4-chloroN,N-dimethylaniline follow the same photolytic mechanism [55], whereas the pathways for 4-chlorophenol and 4-chloroanisole are different [20, 56]. Both species, the aminophenyl cation and the carbene 4-oxocyclohexa-2,5dienylidene, are isoelectronic, which accounts for the similarity of their absorption spectra [55]. The mechanism of 4-chloroaniline photochemistry was independently studied by Guizzardi et al. in organic solvents; they reached very similar conclusions [57]. These authors pointed out that the aminophenyl cation has a triplet-diradical character which fully explains its reactivity in organic solvents [57]. However, in aqueous solutions the cation reacted with hydroxyl ions with a rate constant of 3.1 × 1010 M–1 s–1 , which can only be interpreted as a deprotonation reaction [55]. The carbene 4-iminocyclohexa-2,5dienylidene thus must exist in aqueous solutions, even though its properties have not yet been characterized. This is partly due to an expected low extinction coefﬁcient, similar to the neutral anilino radical [55]. Following these arguments, the primary pathways of 4-chloroaniline photolysis in polar solvents may be pictured as shown in Scheme 7. The properties of both mentioned carbenes have been studied in detail by means of time-resolved resonance Raman spectroscopy and density functional calculations [58]. The measured spectra conﬁrmed the triplet character

Scheme 7 Primary steps of 4-chloroaniline photolysis in aqueous solution

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

177

of both species. Some characteristic differences were noted, in particular a lower cyclohexadienyl character of the imino species, which was found to be similar to the diphenylnitrenium ion. The authors concluded that the charge delocalization ability of the second hydrogen atom on the NH2 + moiety is the main cause of the observed differences [58]. The dependence of the photophysical and photodehalogenation properties of the 4-haloanilines and 4-halo-N,N-dimethylanilines on the nature of the halogen substituent in several organic solvents were investigated by Freccero et al. [59]. Heterolytic dehalogenation was monitored by trapping the phenyl cation by allyltrimethylsilane. It was found that photoreaction proceeded in all cases from the excited triplet state of the molecule; this is in agreement with the results of another recent computational study showing that intersystem crossing dominates the photophysics of 4-chloroaniline [60]. Freccero et al. found that 4-iodoaniline and 4-bromoaniline as well as their N,N-dimethylated analogues were poorly photoreactive. 4-Chloroaniline and 4-chloro-N,N-dimethylaniline had the highest reactivity in polar solvents (protic and non-protic), that of the methylated form being about twice higher. Interestingly, a protic polar solvent was required for efﬁcient dehalogenation of 4-ﬂuoroaniline. The results were rationalized by means of density functional theoretical calculations, which showed that C – X bond homolysis in the triplet state was endothermic in all cases, while C – X heterolysis in acetonitrile was exothermic for the C – Cl and C – Br bonds. Heterolysis of the C – F bond was prevented by the high bond strength unless a H – F bond could be concomitantly formed, which explained the necessity of a protic solvent. The poor reactivity of 4-bromoaniline was explained by the short lifetime of its triplet state [59]. It should, however, be noted that the difference in dehalogenation yield between 4-chloro- and 4-bromoaniline is far less dramatic in aqueous solution; these authors suggested an alternative explanation, involving an in-cage recombination reaction of the photogenerated carbene with bromide, based on the fact that this reaction is faster by two orders of magnitude than the corresponding reaction with chloride [55]. Investigations of the photochemistry of 4-haloanilines in heterogeneous and environmental systems are scarce. One study to be mentioned is that of adsorption and photodissociation processes on GaN, in which it was shown that 4-iodoaniline is homolytically dehalogenated while 4-chloroaniline is inert [61]. 4-Chloroaniline has been found to be poorly biodegradable in surface waters [62]; in an older study, the rate of its photoinduced removal from estuarine water was determined to be higher than its biodegradation rate throughout the year [63]. On account of the high quantum yield of direct dehalogenation, photolysis must be expected to contribute signiﬁcantly; the direct photolysis value of Φ = 0.50 mentioned above [55] may be compared with the degradation quantum yield of Φ = 0.015 recently measured for the phototransformation of 4-chloroaniline induced by aqueous Fe(III) at λ = 365 nm in acidic solution (pH 3–4) [64].

178

G. Grabner · C. Richard

3.2 Dihalogenated Anilines Following a preliminary study [65], the photochemistry of the six isomeric dichloroanilines in aqueous solution was investigated in a systematic way by Othmen and Boule [66]. Photoproduct analysis showed a strong preference for photohydrolysis from each of the isomers, forming an aminochlorophenol. Some speciﬁcity in the substitution pattern was observed, the relative probabilities for dechlorination being highest at the meta position. Accordingly, the highest degradation quantum yield upon monochromatic irradiation in the ﬁrst absorption band (around 293 nm) was obtained for 3,5dichloroaniline with Φ = 0.042; for the other isomers, the values ranged between 0.015 and 0.020. Interestingly, the products obtained from the orthoand para-substituted isomers showed no sign of the photochemistry of the respective monohaloanilines; in particular, ring contraction products were not formed. C – Cl bond homolysis was ruled out as a possible dehalogenation mechanism by the observing that the photoreaction was not inﬂuenced of O2 and that the quantum yields increased upon ﬂuorine substitution (Φ = 0.12 for 2,4-diﬂuoroaniline). Some aminochlorophenols were shown to undergo a secondary photoinduced cyclization, yielding aminochlorophenoxazones [66]. Dechlorination was also obtained as the main photoprocess in some related systems. Encinas et al. investigated the photochemistry of 2,6dichlorodiphenylamine, which is the parent chromophore of the nonsteroidal antiinﬂammatory drug diclofenac [67]. The ﬁrst stage of the reaction involved a rapid 6π-electrocyclization to form 1-chlorocarbazole; in a second step, the remaining chlorine atom was thought to be split homolytically [67]. The photodegradation of 3,3 -dichlorobenzidine, a suspected carcinogen, was studied by Nyman et al. using laser irradiation between 300 and 360 nm [68]. The authors observed two consecutive dechlorination steps ultimately leading to unsubstituted benzidine [68].

4 Biocides 4.1 Phenoxyacetic Acid Derivatives Phenoxyacetic acid is the parent chromophore of a number of commonly used pesticides. If ring-halogenated, they can be regarded as derivatives of haloalkoxybenzenes. The simplest model molecules of this series are the haloanisoles. The photochemistry of 4-chloro- and 4-ﬂuoroanisole has been studied previously, and it was shown that in aqueous solution photohydro-

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

179

lysis takes place with high quantum efﬁciency [56]. Although these authors proposed a pathway involving charge separation in a triplet excimer, it seems reasonable, based on current knowledge, to assume that photohydrolysis proceeds in an analogous way as in the cases of the halophenol and haloaniline molecules discussed above, i.e. either in a concerted mechanism or via a substituted phenyl cation, in this case a methoxyphenyl cation. A number of studies have been carried out on actual herbicides (see Table 1). The photochemistry of MCPA (4-chloro-2-methylphenoxyacetic acid) was investigated as a function of pH and irradiation wavelength [69]. The anionic form, which will form the majority under environmental conditions (pKa = 3.07), showed again a clean photohydrolysis reaction to give the corresponding phenol with a quantum yield of Φ = 0.59. In contrast, the neutral form mainly yielded the rearrangement product 5-chloro-2-hydroxy3-methylphenylacetic acid, indicating a previous homolytic ﬁssion of the Table 1 Halophenoxyacetic acid pesticides

180

G. Grabner · C. Richard

ether bond (photo-Claisen rearrangement). These pathways prevailed for excitation wavelengths between 254 and 310 nm (Scheme 8). Interestingly, photochemical reactions were also observed for longer irradiation wavelengths and even under sunlight. In this case, the main product was 4-chloro-2-methylphenol, i.e. dehalogenation did not take place. A hypothesis was formulated according to which MCPA is oxidized by p-benzoquinone derivatives present in the reaction mixture [69]. From the point of view of environmental photochemistry, these results are relevant since they document a photodegradation pathway leading to a 4chlorophenol derivative, which in turn is susceptible to be efﬁciently photolyzed [27]. In this connection, it may be noted that photoacceleration of the catalyzed oxidation of 2,4,6-trichlorophenol was observed by Lente and Espenson; this effect was attributed to the action of an intermediate pbenzoquinone derivative as well [70]. The photodegradation of chlorophenols and chlorophenoxyacetic acid derivatives at wavelengths λ > 300 nm was also studied by Klementová and Matousková, who attributed the photoreaction to the presence of traces of Fe(III) ions acting photocatalytically [71]. The mechanism of the phototransformation of mecoprop was found to be very similar to that of MCPA [72]. Again, photohydrolysis was the major pathway from the anionic form, whereas irradiation of the neutral form led to a radical rearrangement. 4-chloro-o-cresol was obtained at λ = 365 nm or under sunlight. The authors showed that the same product was obtained when Fe(III) ions or nitrite was used as a photosensitizer [72].

Scheme 8 Mechanisms of photolysis of MCPA

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

181

Triadimefon and triadimenol are bichromophoric fungicides containing a triazole group. Their aqueous photochemistry was investigated by Da Silva et al., who found 4-chlorophenol and 1,2,4-triazole as major products [73]. Transient absorption spectroscopy revealed the formation of 4chlorophenoxyl radicals. Triadimefon was efﬁciently degraded when irradiated at λ = 313 nm (Φ = 0.13), whereas triadimenol was stable under these conditions because of its blue-shifted absorption spectrum [73]. The phototransformation of triadimefon on glass and soil surfaces was studied by Nag and Dureja, who were able to identify a number of products mainly corresponding to side chain reactions [74]. Dichlorprop and 2,4-D are structurally related dichlorinated phenoxyacetic acid pesticides (see Table 1). The photochemistry of 2,4-D has been studied some time ago; 2,4-dichlorophenol was found as the the main product [75]. More recent photoproduct analysis studies are available for dichlorprop [76, 77]. Climent et al. identiﬁed several products corresponding to dechlorination on the one hand, and to decarboxylation on the other; homolytic bond cleavage steps were proposed to take place [76]. In a later study, Meunier et al. were able to demonstrate that 4-chloropyrocatechol is the major primary photoproduct [77]. A heterolytic cleavage of the ether bond and of the C – Cl bond was put forward as a possible mechanism (Scheme 9). A number of other, but minor primary photoproducts was also found, among them the products expected from a radical (photo-Claisen) rearrangement and from photohydrolysis of the ortho chlorine; 2- and 4-chlorophenol were detected too, but their formation remained unexplained. The photodegradation quantum yield of dichlorprop did not depend on pH and was 50 times smaller than that of the anionic form of the related monohalogenated compound mecoprop (see above) [77]. This is another example of the marked inﬂuence of the pattern of ring halogen substitution on the course and on the efﬁciency of photodegradation.

Scheme 9 Photolysis of dichlorprop

182

G. Grabner · C. Richard

4.2 Phenylurea Derivatives and Related Compounds Substituted phenylurea derivatives form an important group of pesticides [6]. As mentioned in the Introduction, their photochemistry has been described in detail in a recent review article, which also covers the older literature [9]. This allows us to be more brief here; readers are referred to this publication concerning detailed reaction mechanisms. In line with the topic of our review, we will concentrate on halogenated phenylureas, but also include some related compounds. The molecular structures of the halogenated phenylurea derivatives investigated in the recent literature are given in Table 2. They differ from each Table 2 Halophenylurea pesticides

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

183

other by the position and number of the halogen substituent(s), which deﬁne a relationship to the corresponding haloanilines discussed earlier, and by the presence or absence of a methoxy group in the side chain. The photodegradation quantum yields determined upon continuous excitation, mostly at λ = 254 nm, range between Φ = 0.02 and Φ = 0.13 [9]. It is of interest to note that the relatively high aqueous solubilities of the phenylurea herbicides render them mobile in soil; in addition, their biodegradation is slow [78]. They are therefore susceptible to migrate to aquifers. Their biotransformation initially involves side-chain reactions, leading to halogenated acetanilides or anilines which may be more toxic than the parent compounds [7, 79], and which may be persistent in surface waters, as noted above. This means that both the phenylureas and their metabolites are potential targets for transformation by photochemical processes in the environment. The ﬁrst two derivatives in the series, monuron and metobromuron, are related to the 4-haloanilines. Their primary photochemistry has been studied by Boulkamh and Richard by means of nanosecond absorption spectroscopy [80]. The transients detected from both compounds in aqueous solution could be assigned to the N-substituted 4-iminocarbene, imino-pbenzoquinone-O-oxide and anilino radical from a complete analogy of their spectral and reactive behavior with that of the species obtained from 4chloroaniline [55, 57]. The quantum yields of carbene formation were determined to be Φ = 0.051 for monuron and Φ = 0.060 for metobromuron [80]. It was concluded that the p-halogen-substituted phenylurea derivatives underwent the same heterolytic dehalogenation process as the 4-haloanilines, which could be understood with reference to the protonability of the amine nitrogen, as in the case of 4-chloro-N,N-dimethylaniline [55]. The ﬁnal products of the reactions of the carbene with H2 O and with the substrate molecule are expected to be, respectively, the p-hydroxylated derivative and dimers; both had indeed been found in earlier studies [9]. More recently, the products of carbene reactions were identiﬁed after irradiation of aqueous metobromuron along with those of side-chain demethoxylation [81]. It is interesting to note that products of a homolytic photo-Fries type rearrangement were not reported. This is in contrast to the related molecule 4-chloroacetanilide, which upon irradiation in the 280–300 nm range gave 4-hydroxyacetanilide, but also 2 -amino-5 -chloroacetophenone, the latter resulting from the photo-Fries reaction [82]. It thus seems that the presence of the urea nitrogen inhibits the homolytic cleavage of the side chain. The transformation of monuron photosensitized by fulvic acids (λ = 365 nm) was shown to yield the same products as direct photolysis or sensitization by hydroquinone [82]. It was concluded that in this case, sensitization was mediated by triplet energy transfer; in contrast, the presence of a humic acid induced an H atom transfer and the products were different [82]. The fulvic acid example shows that long-wavelength sensitization may in speciﬁc

184

G. Grabner · C. Richard

cases lead to a degradation mechanism similar to that obtained upon direct photoirradiation; this similarity will hold if direct photolysis proceeds via a triplet pathway as well, as is the case for the mono-p-substituted derivatives. It should, however be noted that in a different study, the photodegradation of pesticides, including chlorotoluron, metoxuron and diuron, in surface lake water was found to occur primarily by one-electron oxidation involving excited triplet states of dissolved organic matter, suggesting that triplet energy transfer may not be a frequent mechanism of sensitization of the halophenylureas [83]. Phenylurea derivatives mono-halogen-substituted in meta position were found to undergo nearly quantitative photohydrolysis, in the same way as 3-chloroaniline; this was evidenced for chlorotoluron [84, 85] and for metoxuron [86], but also for the related carbamate derivative chlorpropham (isopropyl-3-chlorocarbanilate) [49]. The mechanisms of the photodegradation of the dihalogenated derivatives diuron, linuron and chlorobromuron in aqueous solution were again found to be quite similar [87, 88]. In all cases, photohydrolysis was observed, exhibiting a remarkable dependence on the wavelength of excitation: substitution by OH in meta position was the major pathway at λ = 254 nm, whereas substitution in para position prevailed at λ = 365 nm [87, 88]. Demethoxylation of the side chain was additionally observed for linuron and chlorobromuron [88]. Jirkovský et al. also investigated the photolysis of diuron in dry aerobic conditions on solid supports (silica, clay), and found radically different degradation mechanisms, mainly involving oxidation and elimination of side chain methyl groups [87]. The primary steps of the photolysis of aqueous monuron and diuron were investigated by Canle et al. by means of transient absorption spectroscopy using an ArF laser (λ = 193 nm) for excitation [89]. Under these conditions, photoionization occurred with a quantum yield of about 10%. Radical cations were detected after the laser pulse and found to deprotonate to yield neutral radicals [89]. A further pesticide whose photodegradation was investigated is the bichromophoric derivative 1-(2-chlorobenzoyl)-3-(4-chlorophenyl)urea (CCU) [90]. An important number of products could be identiﬁed, the formation of which assumed to result from homolytic scissions of the urea chain at different positions [90]. The evolution of the toxicity of solutions of the herbicides metobromuron, metoxuron and diuron during photoirradiation was recently studied by Bonnemoy et al. [91]. In the case of metobromuron, initiation of the photoreaction was accompanied by a marked toxicity increase which subsided with further irradiation; intermediary quinone imines or quinones were proposed to be responsible. In contrast, photolysis of the other two compounds resulted in a lasting increase in toxicity, which was, however, not caused by photohydrolysis, but was due to unidentiﬁed minor photoreactions [91].

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

185

4.3 Other Biocides A few other biocides related to the halogenated phenols and anilines have been studied during the last years; this work is summarized in the following section. The compound structures are shown in Table 3. An important class of pesticides are the halobenzonitriles. Their hydroxylated derivatives, bromoxynil (see Table 3) and its chlorinated and iodinated congeners chloroxynil and ioxynil, are related to the 2,6-dihalophenols that had earlier been shown to give ring contraction photoproducts [92]. In contrast to this ﬁnding, irradiation of aqueous bromoxynil yielded the photohydrolysis product 3-bromo-4,5-dihydroxybenzonitrile, which accounted for 65% of the conversion [93]. In a later study, Clarke et al. found that the bromoxynil isomer 3,5-dibromo-2-hydroxybenzonitrile was photohydrolyzed much more efﬁciently than the former, even under sunlight, and attributed Table 3 Other biocides derived from halogenated phenols or anilines

186

G. Grabner · C. Richard

this to the fact that its absorption spectrum extends to longer wavelengths; they also observed that photohydrolysis was highly speciﬁc, substitution of Br by OH being observed only in para position with respect to the OH group [94]. Bromoxynil itself was found to be quite photostable in surface waters [95]. Aciﬂuorfen may be regarded as an ortho-halogen-substituted diphenyl ether. However, its spectroscopic properties are dominated by the nitro group on the second phenyl ring, conferring a strong absorption in the near-UV to the molecule. The compound was found to be highly photostable, the degradation quantum yields at λ = 313 nm being Φ = 6 × 10–5 at pH 4, and Φ = 1.2 × 10–4 at pH 12 [96]. The main photolytic process is decarboxylation in neutral or acidic solution, and scission of the ether bond, corresponding to a nucleophilic displacement of phenoxide by hydroxide, in alkaline medium. These two processes were found to be competitive and suggested to proceed from the triplet excited state of the molecule, based on the observation that high bromide concentrations were able to increase the degradation yield [96]. Similar photobehavior was found in a comparative study of 5-methoxy-2nitrobenzoic acid, showing that this is the photochemically relevant entity of aciﬂuorfen [96]. The chlorine-substituted phenyl ring is, however, not photochemically inert, as witnessed by the detection of the product of heterolytic chlorine photohydrolysis, but this pathway was found to be a minor one [96, 97]. The degradation of aciﬂuorfen was also studied in natural waters containing humic substances, exposed to solar light; direct photolysis was the main reaction pathway in this case, too [96]. A later comparative study of aciﬂuorfen in several solvents (water, acetonitrile, methanol and nhexane) found low photodegradation quantum yields in the polar solvents, while the quantum yield in n-hexane was higher by two orders of magnitude (Φ = 0.017) [98]. Experiments using excited state sensitizers and quenchers in acetonitrile indicated that the photoreaction originated from the singlet excited state of the molecule in this solvent [98]. Dicamba is a dichlorinated 2-methoxybenzoic acid derivative. It was found to be photodegraded in aqueous solution with a quantum yield of Φ = 0.0215 at λ = 275 nm [99]. The photoproducts were the dihydroxylated derivative and an unexpected dicyclic compound, only formed in the presence of O2 (Scheme 10). The proposed mechanism involved an initial photohydrolysis forming a 4-chlorophenol derivative, which was suggested to undergo a further het-

Scheme 10 Products of photolysis of dicamba

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

187

erolytic photodechlorination according to a carbene mechanism, with probably higher quantum yield. In the presence of O2 , the intermediate carbene would react to form a p-benzoquinone derivative, which was proposed to be the key intermediate in the cyclization step. When irradiated in the solid phase on a clay (laponite) support, dicamba was degraded in an unspeciﬁc way, giving products of methylation, decarboxylation, chlorination, dechlorination, and hydroxylation [99]. Dichlorophen (4,4 -dichloro-2,2 -methylenediphenol) is directly related to 4-chlorophenol. However, its ﬁrst acid dissociation constant is considerably lower (pKa = 7.5 and 11.5); the photochemistry of the monoanion is therefore of potential environmental signiﬁcance, since its absorption spectrum is signiﬁcantly red-shifted. Mansﬁeld and Richard studied the phototransformation of dichlorophen by means of product analysis [100]. The quantum yields of chloride formation were determined to be Φ = 0.50 and Φ = 0.34 for the neutral and for the anionic form, respectively. The neutral form yielded the products expected according to the carbene mechanism of 4chlorophenol photolysis (see above), i.e. a hydroquinone derivative in deoxygenated medium, a p-benzoquinone derivative in the presence of O2 , and reduction products in the presence of alcohols. All three products were found at pH 9, even in deoxygenated solution in the absence of alcohols. While the p-benzoquinone derivative was suggested to be formed by thermal oxidation in the course of the analytical procedure, a mechanism involving dismutation of semiquinone radicals originating from the carbene was proposed to explain the formation of the reduction product [100]. Triclosan (5-chloro-2-(2,4-dichlorophenoxy)phenol) is a widely used antimicrobial biocide who was recently found in appreciable concentrations in wastewater treatment plant efﬂuents [101]. Its pKa value is 8.1, which means that in this case too the photoactivity of the neutral as well as of the anionic form of the molecule had to be considered [102]. The quantum yields of phototransformation were determined to be Φ = 0.40 at λ = 279 nm and pH 5.9, Φ = 0.50 at λ = 292 nm and pH 11.5, and Φ = 0.31 at λ = 313 nm and pH 11.0. Because of these high quantum yields, photoremoval of triclosan from surface waters is a relevant process under actual environmental conditions: direct photodegradation was found to account for 80% of the total elimination of the compound from a Swiss lake during the summer season [102]. It is interesting to note that the photodegradation quantum yields of triclosan are signiﬁcantly higher than those of 3-chlorophenol [5], indicating that from the photochemical point of view, the molecule is more closely related to the p-chlorophenoxy derivatives MCPA and mecoprop discussed above. A comprehensive photoproduct analysis study of triclosan is not yet available. Previous studies have concentrated on the possibility of a cyclization yielding 2,8-dichlorodibenzo-p-dioxin (Scheme 11). In a recent study, it was established that this toxic product is indeed formed from the triclosan anion, both under laboratory conditions and in natural waters, with quan-

188

G. Grabner · C. Richard

Scheme 11 Photocyclization of triclosan

tum yields reaching Φ = 0.039 and chemical yields between 1 and 12% of total photodegradation [103].

5 Conclusions We have presented an overview of work done during the last decade on the photochemistry of the halogenated phenols and anilines, and of related biocides. The mechanisms of the phototransformation of the monohalogenated derivatives in aqueous solution are by now reasonably well understood. They exhibit two characteristic pathways that are relevant for many compounds studied in this review: ﬁrst, carbon-halogen photohydrolysis, i.e. substitution of the halogen by a hydroxyl group provided by the solvent; second, halide elimination to form a carbene or aryl cation. Both processes are heterolytic. Dehalogenation of p-monosubstituted derivatives occurs uniquely via the carbene pathway, with high quantum yields, from the triplet state of the molecules. In contrast, m-monosubstituted compounds often exhibit clean photohydrolysis. The picture is generally more complex for the o-monosubstituted derivatives, which can follow both mentioned pathways, and additionally a third one, contraction of the aromatic ring to form cyclopentadiene derivatives. Homolytic carbon-halogen bond cleavage is the exception; the only case were it has unambiguously been demonstrated is 2-iodoaniline. The pathways governing the phototransformation of polyhalogenated phenols and anilines are not as well documented. In particular, mechanistic studies for compounds carrying more than two halogen substituents are scarce. Homolytic cleavage mechanisms have often been assumed, but are not based on direct evidence. The photoreactions characteristic for the halogenated phenols and anilines have shown to occur in most members of the series of chlorophenoxy and halophenylurea derivative pesticides, and in some other related biocides as well. Photohydrolysis is a frequently encountered process, but carbene formation has been demonstrated in several cases too. The phototransformation mechanisms of these molecules, however, are frequently complex, owing to the greater complexity of the molecules in question, and homolytic cleavage steps have been found to contribute in several cases. Mechanistic studies of heterogeneous or realistic environmental systems are a recent subject. First results obtained on substances whose aqueous pho-

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols

189

tochemistry is well understood, in particular 4-chlorophenol, have demonstrated large variations in photoproduct distribution depending on the investigated system. A further question relevant to environmental conditions concerns the role of direct photolysis as compared to that of processes sensitized by other species, such as Fe(III) ions or dissolved organic matter. The relevance of direct photolysis has by now been shown for a number of systems, including several pesticides. The actual mechanisms may show combined features of several pathways encountered in the simpler model systems; for instance, a photosensitized step may lead to a product which is then degraded by direct photolysis. In the same way, combinations of bio- and photodegradation may be of relevance. Examples of such compound mechanisms have already been investigated. The next years will certainly see increased efforts to clarify these complex issues of environmental photochemistry.

References 1. Pignatello JJ, Martinson MM, Steiert JG, Carlson RE, Crawford RL (1983) Appl Environ Microbiol 46:1024 2. Vialaton D, Richard C (2002) Aquat Sci 64:207 3. Boule P, Meunier L, Bonnemoy F, Boulkamh A, Zertal A, Lavedrine B (2002) Int J Photoenergy 4:69 4. Blough NV, Sulzberger B (2003) Aquat Sci 65:317 5. Richard C, Grabner G (1999) Mechanism of phototransformation of phenol and derivatives in aqueous solution. In: Boule P (ed) Handbook of environmental chemistry, vol 2, part 1: Environmental photochemistry. Springer, Berlin Heidelberg, p 217 6. Tomlin C (ed) (2003) The pesticide manual, 13th edn. British Crop Protection Council Publications, Alton 7. Tixier C, Bogearts P, Sancelme M, Bonnemoy F, Twagilimana L, Cuer A, Bohatier J, Veschambre H (2000) Pest Manag Sci 56:455 8. Burrows HD, Ernestova LS, Kemp TJ, Skurlatov YI, Purmal AP, Yermakov AN (1998) Prog React Kinet 23:145 9. Amine-Khodja A, Boulkamh A, Boule P (2004) Photochem Photobiol Sci 3:145 10. Boreen AL, Arnold WA, McNeill K (2003) Aquat Sci 65:320 11. Guyon C, Boule P, Lemaire J (1982) Tetrahedron Lett 23:1581 12. Guyon C, Boule P, Lemaire J (1984) Nouv J Chim 8:685 13. Kirmse W (2002) Eur J Org Chem 2193 14. Shi Z, Sigman ME, Ghosh MM, Dabestani R (1997) Environ Sci Technol 31:3581 15. Boule P, Othmen K, Richard C, Szczepanik B, Grabner G (1999) Int J Photoenergy 1:49 16. Bonnichon F, Richard C, Grabner G (2001) Chem Commun 73 17. Akai N, Kudoh S, Takayanagi M, Nakata M (2001) J Photochem Photobiol A Chem 146:49 18. Akai N, Kudoh S, Takayanagi M, Nakata M (2002) Chem Phys Lett 363:591 19. Bonnichon F, Grabner G, Richard C, Lavédrine B (2003) New J Chem 27:591 20. Grabner G, Richard C, Köhler G (1994) J Am Chem Soc 116:11470 21. Durand APY, Brown RG, Worrall D, Wilkinson F (1996) J Photochem Photobiol A Chem 96:35

190

G. Grabner · C. Richard

22. Ouardaoui A, Steren CA, van Willigen H, Yang C (1995) J Am Chem Soc 117:6803 23. Ouardaoui A, Martino DM, Steren CA, van Willigen H (1997) Appl Magn Reson 13:275 24. Durand AP, Brown RG, Worrall D, Wilkinson F (1998) J Chem Soc, Perkin Trans 2:365 25. Ho TFL, Bolton JR, Lipczynska-Kochany E (1996) J Adv Oxid Technol 1:170 26. Svetlichnyi VA, Chaikovskaya ON, Bazyl’ OK, Kuznetsova RT, Sokolova IV, Kopylova TN, Meshalkin YP (2001) High Energy Chem 35:258 27. Vialaton D, Richard C, Baglio D, Paya-Perez AB (1998) J Photochem Photobiol A Chem 119:39 28. Bonnichon F, Grabner G, Guyot G, Richard C (1999) J Chem Soc, Perkin Trans 2:1203 29. Pandiyan T, Martinez Rivas O, Orozco Martinez J, Burillo Amezcua G, MartinezCarrillo MA (2002) J Photochem Photobiol A Chem 146:149 30. Benitez FJ, Beltran-Heredia J, Acero JL, Rubio FJ (1999) Ind Eng Chem Res 38:1341 31. Benitez FJ, Acero JL, Real FJ, García J (2003) Chemosphere 51:651 32. Skurlatov YI, Ernestova LS, Vichutinskaya EV, Samsonov DP, Semenova IV, Rod’ko IY, Shvidky VO, Pervunina RI, Kemp TJ (1997) J Photochem Photobiol A Chem 107:207 33. Skurlatov YI, Ernestova LS, Vichutinskaya EV, Samsonov DP, Pervunina RI, Semenova IV, Rod’ko IY, Shvydki VO (1997) Khim Fiz 16:16 34. Skurlatov YI, Ernestova LS, Vichutinskaya EV, Samsonov DP, Pervunina RI, Semenova IV, Shvydki VO (1998) Acta Hydrochim Hydrobiol 26:31 35. Piccinini P, Pichat P, Guillard C (1998) J Photochem Photobiol A Chem 119:137 36. Hong J, Kim DG, Cheong C, Jung SY, Yoo MR, Kim KJ, Kim TK, Park YC (2000) Anal Sci 16:621 37. Zheng MH, Liu PY, Xu XB (2002) Organohalogen Compounds 56:175 38. Ho TFL, Bolton JR (1998) Water Res 32:489 39. Svenson A, Hynning PÅ (1997) Chemosphere 34:1685 40. Da Silva JP, Vieira Ferreira LF, Da Silva AM, Oliveira AS (2002) J Photochem Photobiol A Chem 151:157 41. Da Silva JP, Vieira Ferreira LF, Da Silva AM, Oliveira AS (2003) Environ Sci Technol 37:4798 42. Klánová J, Klán P, Nosek J, Holoubek I (2003) Environ Sci Technol 37:1568 43. Klánová J, Klán P, Heger D, Holoubek I (2003) Photochem Photobiol Sci 2:1023 44. Donaldson SG, Miller GC (1997) J Environ Qual 26:402 45. Puplampu EL, Dodoo DK (2000) J Photochem Photobiol A Chem 135:81 46. Liu PY, Zheng MH, Xu XB (2002) Chemosphere 46:1191 47. Othmen K, Boule P (1997) Bull Environ Contam Toxicol 59:924 48. Othmen K, Boule P (2000) J Photochem Photobiol A Chem 136:79 49. David B, Lhote M, Faure F, Boule P (1998) Water Res 32:2451 50. Boule P, Guyon C, Lemaire J (1982) Chemosphere 11:1179 51. Boule P, Tissot A, Lemaire J (1985) Chemosphere 14:1789 52. Othmen K, Boule P, Richard C (1999) New J Chem 23:857 53. Latowski T (1968) Ann Soc Chim Pol 42:99 54. Szczepanik B, Latowski T (1997) Polish J Chem 71:807 55. Othmen K, Boule P, Szczepanik B, Rotkiewicz K, Grabner G (2000) J Phys Chem A 104:9525 56. Lemmetyinen H, Konijnenberg J, Cornelisse J, Varma CAGO (1985) J Photochem 30:315 57. Guizzardi B, Mella M, Fagnoni M, Freccero M, Albini A (2001) J Org Chem 66:6353

Mechanisms of Direct Photolysis of Biocides Based on Halogenated Phenols 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97.

191

Chan WS, Leung KH, Ong SY, Phillips DL (2002) J Phys Chem A 106:6254 Freccero M, Fagnoni M, Albini A (2003) J Am Chem Soc 105:13182 Artyukhov VY, Morev AV, Morozova YP (2003) Opt Spectrosc 95:361 Bermudez VM (2002) Surf Sci 519:173 Ahtinainen J, Aalto M, Pessala P (2003) Chemosphere 51:529 Hwang HM, Hodson RE, Lee RF (1987) Water Res 21:309 Mailhot G, Hykrdová L, Jirkovský J, Lemr K, Grabner G (2004) Appl Catal B Environ 50:25 Othmen K, Boule P (1999) J Photochem Photobiol A Chem 121:119 Othmen K, Boule P (1999) J Photochem Photobiol A Chem 121:161 Encinas S, Boscá F, Miranda MA (1998) Photochem Photobiol 68:640 Nyman MC, Haber KS, Kenttämaa HI, Blatchley III ER (2002) Environ Toxicol Chem 21:500 Zertal A, Sehili T, Boule P (2001) J Photochem Photobiol A Chem 146:37 Lente G, Espenson JH (2003) Chem Commun 1162 Klementová S, Matousková J (2000) Res J Chem Environ 4:25 Meunier L, Boule P (2000) Pest Manag Sci 56:1077 Da Silva JP, Vieira Ferreira LF, Da Silva AM (2003) J Photochem Photobiol A Chem 154:293 Nag SK, Dureja P (1996) Pestic Sci 48:247 Crosby DG, Tutass HO (1966) J Agric Food Chem 14:596 Climent MJ, Miranda MA (1997) J Agric Food Chem 45:1916 Meunier L, Gauvin E, Boule P (2002) Pest Manag Sci 58:845 Sørensen SR, Bending GD, Jacobsen CS, Walker A, Aamand J (2003) FEMS Microbiol Ecol 45:1 Tixier C, Sancelme M, Aït-Assa M, Widehem P, Bonnemoy F, Cuer A, Truffaut N, Veschambre H (2002) Chemosphere 46:519 Boulkamh A, Richard C (2000) New J Chem 24:849 Boulkamh A, Sehili T, Boule P (2001) J Photochem Photobiol A Chem 143:191 Richard C, Vialaton D, Aguer JP, Andreux F (1997) J Photochem Photobiol A Chem 111:265 Gerecke AC, Canonica S, Müller SR, Schärer M, Schwarzenbach RP (2001) Environ Sci Technol 35:3915 Millet M, Palm WU, Zetzsch C (1998) Environ Toxicol Chem 17:258 Tixier C, Meunier L, Bonnemoy F, Boule P (2000) Int J Photoenergy 2:1 Boulkamh A, Harakat D, Sehili T, Boule P (2001) Pest Manag Sci 57:1119 Jirkovský J, Faure V, Boule P (1997) Pestic Sci 50:42 Faure V, Boule P (1997) Pestic Sci 53:413 Canle LM, Rodríguez S, Rodríguez Vázquez LF, Santaballa JA, Steenken S (2001) J Mol Struct 565–566:133 Guoguang L, Xiangning J, Xiaobai X (2001) J Agric Food Chem 49:2359 Bonnemoy F, Lavédrine B, Boulkamh A (2004) Chemosphere 54:1183 Boule P, Guyon C, Tissot A, Lemaire J (1985) J Chim Phys 82:513 Machado F, Collin L, Boule P (1995) Pestic Sci 45:107 Clarke J, Hill RR, Roberts DR (1998) J Chem Res (S) 778 Nolte J, Heinlich F, Grass B, Zullei-Seibert N, Preuss G (1995) Fresenius J Anal Chem 351:88 Vialaton D, Baglio D, Paya-Perez A, Richard C (2001) Pest Manag Sci 57:372 Vulliet E, Emmelin C, Scrano L, Bufo SA, Chovelon JM, Méallier P, Grenier-Loustalot MF (2001) J Agric Food Chem 49:4795

192

G. Grabner · C. Richard

98. Scrano L, Bufo SA, D’Auria M, Méallier P, Behechti A, Shramm KW (2002) J Environ Qual 31:268 99. Aguer JP, Blachère F, Boule P, Garaudee S, Guillard C (2000) Int J Photoenergy 2:81 100. Mansﬁeld E, Richard C (1996) Pestic Sci 48:73 101. Singer H, Müller S, Tixier C, Pillonel L (2002) Environ Sci Technol 36:4998 102. Tixier C, Singer HP, Canonica S, Müller SR (2002) Environ Sci Technol 36:3482 103. Latch DE, Packer JL, Arnold WA, McNeill K (2003) J Photochem Photobiol A Chem 158:63

Hdb Env Chem Vol. 2, Part M (2005): 193–219 DOI 10.1007/b138184 © Springer-Verlag Berlin Heidelberg 2005 Published online: 6 September 2005

Recent Developments in the Environmental Photochemistry of PAHs and PCBs in Water and on Solids Richard M. Pagni1 (u) · Reza Dabestani2 1 Department

of Chemistry, The University of Tennessee, Knoxville, TN 37996-1600, USA [email protected] 2 Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA [email protected] 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

194

2

Photochemistry of PAHs in Water . . . . . . . . . . . . . . . . . . . . . . .

194

3

Photochemistry of PAHs in H2 O Ice . . . . . . . . . . . . . . . . . . . . . .

198

4

Photochemistry of PAHs in Heterogeneous Media . . . . . . . . . . . . . .

200

5 5.1 5.2 5.3 5.4 5.4.1 5.4.2 5.5 5.6 5.7

Photochemistry of PCBs . . . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photochemistry in Solution . . . . . . . . . . . . . . . . . . . . . . Photochemistry in Surfactant Solutions . . . . . . . . . . . . . . . Photochemistry at Interfaces . . . . . . . . . . . . . . . . . . . . . Silica and Siloxanes . . . . . . . . . . . . . . . . . . . . . . . . . . Titanium Dioxide . . . . . . . . . . . . . . . . . . . . . . . . . . . Reactions of Photogenerated Hydroxyl Radicals in Air and Water PCBs on Sediments and in Oils . . . . . . . . . . . . . . . . . . . . Ice Photochemistry . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .

205 205 206 208 210 210 210 213 214 215

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

216

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

Abstract This article updates the one on the same topic published in this series in 1999. The photochemistry of PAHs and PCBs in liquid water and on ice and other solids such as silica, soil and titanium dioxide continues to be actively studied. The photochemistry of PAHs in all phases continues to be dominated by oxidation by O2 , with superoxide (O2 –· ), excited singlet oxygen (1 O2 ), and hydroxyl radical (· OH) being the active agents. The recent photochemistry of PCBs has been dominated by practical considerations, i.e. how to use photochemistry to clean up environmental problems involving PCBs. The use of surfactants, the semiconductor TiO2 , and various sources of the powerful oxidant, the hydroxyl radical, in this regard has received considerable attention. Keywords Aqueous photochemistry · Hydroxyl radicals · Ice photochemistry · Photochemistry · Polycyclic aromatic hydrocarbons · Polychlorobiphenyls · Surfactants · Titanium dioxide

194

R.M. Pagni · R. Dabestani

1 Introduction The photochemistry of polycyclic aromatic hydrocarbons (PAHs) and polychlorobiphenyls (PCBs) in water and on solids was reviewed in this series in 1999 [1]. The large interest in this ﬁeld was due to the fact that many PAHs and mixtures of PCBs are on the United States Environmental Protection Agency’s list of priority pollutants (see [1] for the list of compounds). Photochemists wanted to know what the fate of these materials in the environment is when exposed to sunlight and if one could use photochemistry to remediate contaminated sites. This interest has continued and will be the focus of the present review. The review covers the recent literature through the middle of 2003.

2 Photochemistry of PAHs in Water The environmental impact of PAHs has rendered their photochemistry in aqueous media considerable attention in recent years. This is mainly due to the fact that PAHs constitute one of the most prevalent classes of aquatic environmental pollutants [2, 3]. Anthropogenic activity is believed to be mainly responsible for PAHs release into the environment. Combustion of fossil fuels, power generators, vehicle emission, refuse burning, industrial processes, petroleum spills, and coal liquefaction are among the major sources for PAH production. Low molecular weight PAHs are fairly water soluble and prone to microbial degradation and evaporation during their residence time in the environment. Higher molecular weight PAHs, on the other hand, are less water soluble making their biodegradation more difﬁcult. As a result, these PAHs are more likely to be affected by sedimentation and photooxidation in aquatic environments [4]. The following compilation details some of the most recent reports on photochemistry of PAHs in water. Photodegradation of acenaphthylene in water has been carried out by different treatments combining UV radiation, ozone, and hydrogen peroxide with conversions as high as 95–100% [5]. UV radiation alone showed slightly lower efﬁciencies compared to ozonation or ozonation in conjunction with hydrogen peroxide (10–3 M). Although detailed product studies were not carried out, the by-products identiﬁed by GC/MS were claimed to be mainly oxygenated species of the parent compound (ketones, aldehydes and carboxylic acids) formed by either direct reaction of the excited state of acenaphthylene with oxygen to form (a) peroxides that further degrade to low molecular weight products and/or (b) singlet oxygen that could add to acenaphthylene to form peroxides [5].

Recent Developments in the Environmental Photochemistry

195

Photodegradation kinetics of anthracene, pyrene, benz[a]anthracene, and benz[a,h]anthracene in aqueous solution at pH 5.6 and 7.6 have been studied by Lehto et al. using ﬂuorescence detection techniques to monitor the decomposition process without product identiﬁcation or quantiﬁcation [6]. Photodegradation was reported to be fast in all cases, with anthracene and benz[a]anthracene being the fastest to decompose. The apparent ﬁrst-order rate constant for degradation, which was essentially oxygen independent (except for pyrene), was reported to vary from 10–3 to 10–4 s–1 . Based on photodegradation kinetics and ﬂuorescence lifetime studies, the authors proposed a mechanism involving non-emitting excited state species in equilibrium with the excited state singlet state to be responsible for the observed decomposition. Vialaton et al. have reported that irradiation of naphthalene in air saturated aqueous solution produces mainly 7-hydroxy-1,4-naphthoquinone (major), 2-formylcinnamaldehyde, and 2-carboxycinnamaldehyde with a quantum yield of 0.0025 shown in Scheme 1 [7]. They proposed that phototransformation of naphthalene may proceed by a monophotonic ionization process to produce the naphthalene radical cation which can deprotonate and/or react with water to produce radicals that are oxidized by molecular oxygen to yield the observed 2-formylcinnamaldehyde and 2carboxycinnamaldehyde. It was further proposed that the formation of 7hydroxy-1,4-naphthoquinone (major) is likely to result from the photolysis of 1,4-naphthoquinone that can form as an unstable intermediate [8]. Natural sunlight induced photooxidation of naphthalene in aqueous solution has also been reported by McConkey et al. to produce six major products including 1-naphthol, coumarin, and two hydroxyquinone [9]. The authors proposed that the initially formed 2 + 2 and 2 + 4 photocycloaddition products undergo subsequent oxidation and/or rearrangement to form the observed products [9]. Grabner et al. have studied solvent effects on the photophysics of naphthalene and report that ﬂuorescence lifetime decreases by a factor of 2.5 in aqueous solution compared to organic solvents (e.g. ethanol, hexane, acetonitrile) [10]. Based on the observed differences in naphthalene excited triplet state properties in aqueous and organic media, the decrease

Scheme 1

196

R.M. Pagni · R. Dabestani

in ﬂuorescence lifetime was attributed to an increase in intersystem crossing rate induced by water. Simulated solar radiation of anthracene in aqueous media has been reported by Mallakin et al. to produce more than 20 photoproducts consisting of anthraquinones, benzoic acids, benzaldehydes, and phenols as determined by HPLC [11]. Photochemical oxidation of some of the anthraquinones formed initially (Scheme 2) has been attributed to the formation of secondary products observed in this process (not shown). However, no independent studies of anthraquinone photolysis under the same conditions were carried out to conﬁrm the proposed hypothesis. A pseudo ﬁrst-order reaction with a half-life of 2 h was deduced for the photooxidation kinetics of anthracene in aqueous media under simulated solar radiation conditions [11]. Bertilsson and Anneli report that photochemical degradation of anthracene and phenanthrene in water proceeds rapidly (half-lives 1 and 20.4 h, respectively) while that of naphthalene (half-life > 100 h) does not [12]. In the presence of humic substances, the anthracene degradation rate was unaffected while that of phenanthrene was slowed down considerably. Differences in the spectral absorbance of anthracene and

Scheme 2

Recent Developments in the Environmental Photochemistry

197

phenanthrene were invoked to partially explain the observed reaction kinetics of the two compounds [12]. The role of humic substances on photodegradation of 1-aminopyrene in aqueous media has been studied by Zeng et al. [13]. It was found that the photolysis rate of 1-aminopyrene in the presence of humic substances could be enhanced or inhibited depending upon the type and concentration of humic substance used. For example, the ﬁrst-order rate constant for the photolysis of 1-aminopyrene in phosphate buffer (pH 7.0, 1 mM) containing humic acid (20–80 ppm) was enhanced up to ﬁvefold while the same concentration of fulvic acid produced only a twofold increase in rate. When photolysis of 1aminopyrene was carried out in water containing soil humic acid, at 20 ppm concentration, a twofold enhancement was observed while at 80 ppm the rate decreased by up to fourfold [13]. Photodegradation of several PAHs in aerated pure water, Suwannee River fulvic acid, and natural waters has been studied by Fasnacht and Blough [14]. Quantum yields for degradation in pure water were reported to vary from 3.2 × 10–5 to 9.2 × 10–3 showing no clear correlation between the quantum yields and molecular properties. Rate constants for photodegradation of PAHs in solutions of Suwannee River fulvic acid or natural waters were practically unchanged compared to that of pure water. Based on these ﬁndings, the authors claim that photodegradation rate constant in natural waters can be estimated from the quantum yields in pure water, PAH absorption, and solar irradiance [14]. Photooxidation of benzo[a]pyrene, chrysene, and ﬂuorene in aqueous solution using the ozone assisted UV radiation method has also been reported [15, 16]. The highest rate constant for photodegradation was achieved in acidic media [15]. Quantum yields of photolytic decomposition were determined to be 0.014, 0.0031, and 0.0038 for benzo[a]pyrene, chrysene, and ﬂuorene, respectively [16]. Three pathways for PAH degradation were proposed that involve radical cation, singlet molecular oxygen, and hydroxyl radical formed in several reactions from superoxide anion [16]. In the presence of the radical scavenger tert-butanol, photodegradation of benzo[a]pyrene and chrysene was inhibited while that of ﬂuorene was slightly affected. The rate constants for hydroxyl radical reaction with benzo[a]pyrene, chrysene, and ﬂuorene were found to be 2.53 × 1010 , 9.82 × 109 , and 2.77 × 109 M–1 s–1 , respectively. Based on these ﬁndings, it was inferred that photooxidation of benzo[a]pyrene and chrysene proceeds by a radical mechanism while that of ﬂuorene was unclear [17]. Photodegradation of ﬂuorene in aqueous ethanolic solution has been reported to proceed more efﬁciently than its sensitized photolytic oxidation in the presence of TiO2 suspension [18]. The major oxidation products observed for ﬂuorene and its derivatives were 9-ﬂuorenone and its corresponding derivatives. The effect of solvents and substituent groups on photoxidation of ﬂuorene has been discussed by Moeini-Nombel and Matsuzawa [19]. The largest rate constant (3.4 × 10–4 s–1 ) and quantum yield (0.009) was obtained for dichloromethane and both values de-

198

R.M. Pagni · R. Dabestani

creased in the following order: dichloromethane > acetonitrile > methanol > acetonitrile/ water (70 : 30) > acetonitrile/water (50 : 50). The authors describe a relationship between the observed reactivities and solubility of oxygen in the solvents employed. Although 1-methylﬂuorene exhibited similar reactivity as ﬂuorene itself, 2-nitroﬂuorene appeared unreactive and showed good photostability [19]. Irradiation of anthracene and chrysene in natural water by UV and sunlight has been reported to be ﬁrst-order at low concentrations [20]. Photolysis rate and reaction efﬁciencies for anthracene and chrysene were highly dependent on oxygen concentration and increased as the temperature was raised [20]. As discussed above, a number of publications on the photochemistry of PAHs in water have appeared since our last overview of the subject. Although the general consensus is that PAHs undergo photodegradation and are photooxidized, the lack of detailed product and mechanistic studies in most of these reports makes it practically impossible to advance a general conclusion with regard to photodegradation/photooxidation paths. Thus, continued effort in this ﬁeld is required to better deﬁne the mechanism(s) which lead to photodegradation and photooxidation of PAHs in aqueous media.

3 Photochemistry of PAHs in H2 O Ice Photochemical transformation of organic compounds and in particular PAHs on ice, as a medium, has not received much attention from the photochemical community. As a result, information on such transformations is limited. Astrophysical research on water ice, on the other hand, has evolved at a rapid pace in recent years after its discovery on outer solar system bodies and in interstellar space [21–23]. A recent review article by Klan and Holoubek [24] on ice photochemistry provides the current knowledge on the distribution, accumulation, and chemical/photochemical transformation of persistent bioaccumulative and toxic compounds in water ice. Since PAHs constitute a substantial portion of the interstellar carbon inventory [25, 26], their photochemical behavior is of paramount importance in the radiative processing of interstellar ices. Bernstein et al. have used IR spectroscopy and mass spectrometry to study the products formed from photochemical transformation of naphthalene, anthracene, chrysene, phenanthrene, pyrene, tetracene, pentacene, perylene, benzo(e)pyrene, benzo(ghi)perylene, and coronene in water ices using ultraviolet radiation under astrophysical conditions [27]. The results of their investigation have revealed that peripheral carbon atoms can be oxidized to produce aromatic alcohols, ketones, ethers (when bay region is present,

Recent Developments in the Environmental Photochemistry

199

Scheme 3

e.g. benzo(ghi)perylene) and reduced to generate partially hydrogenated aromatic hydrocarbons. Scheme 3 shows typical products observed for coronene photochemistry on water ice. UV radiation of coronene in D2 O ice in a ratio of less than 1 part in 500 induces rapid exchange of its hydrogen atoms with deuterium by three different chemical processes which include D atom addition, D atom exchange at the aromatic edge site shown in Scheme 4 below and D atom exchange at the oxidized edge site shown in Scheme 5 [28]. Klan and his co-workers [29–31] have studied the photochemistry of halobenzenes in ice and reported that the observed photoproducts (biphenyl, terphenyl, and their halogenated isomers) were very different from those obtained in liquid water (mainly phenol derivatives). A more recent study by the same group on photolysis of 2- and 4-chlorophenol in water ice has revealed that the main products are chlorobiphenyldiols which are formed as a result of coupling reactions due to chlorophenol aggregation at the grain boundaries of the polycrystalline state [32]. Dubowski and Hoffmann have reported a quantum yield of 2.3 × 10–4 for photochemical degradation of 4-nitrophenol in ice pellets [33]. The photoproducts, hydroquinone, benzoquinone, 4-nitrosophenol, nitrate, and nitrite have been identiﬁed (Scheme 6) and the likely formation of organic polymers has been suggested. On the basis of the similarity of the products observed on ice and aqueous solution, a similar mechanism has been proposed to operate for the decomposition of 4-nitrosophenol in both media [33]. Although photooxidation seems to be the predominant path to photodegradation of PAHs in ice water, the available data do not provide a mechanistic path for the oxidation process. It is known that photooxidation of PAHs

Scheme 4

200

R.M. Pagni · R. Dabestani

Scheme 5

Scheme 6

in aqueous solution proceeds by either an electron transfer process (type I) or mediated by singlet oxygen (type II) depending upon the nature of PAH in question. The best conclusion that can be drawn from the available data on photochemical conversion of PAHs on water ice is that further research in this ﬁeld is required to better understand and deﬁne the mechanism(s) responsible for PAHs photooxidation.

4 Photochemistry of PAHs in Heterogeneous Media Photochemistry of PAHs in heterogeneous media has also received considerable attention in recent years. The major focus of most of these studies,

Recent Developments in the Environmental Photochemistry

201

however, has been on the fate and stability of these ubiquitous compounds to phototransformation when exposed to light in the environment. The emphasis of this review article will be on the kinetics and products of photodegradation of PAHs in heterogeneous media. Phototransformation of anthracene, pyrene, and phenanthrene on silica, alumina, and ﬂy ash has been reported by Teinemaa and Kirso [34]. Photodegradation was found to proceed much more readily on silica and alumina compared to ﬂy ash but no product(s) were reported. It was found that variation in the decay rates was signiﬁcantly different for the three PAHs studied with anthracene reacting much faster than pyrene followed by phenanthrene [34]. Phototransformation of perylene adsorbed on nonactivated silica gel and alumina surfaces has been reported by Sotero and Arce [35]. Using ﬂuorescence, diffuse reﬂectance laser ﬂash photolysis, and EPR techniques, the authors have examined the effect of oxygen, water, and PAH loading on the photodegradation rates. Results indicate that oxygen signiﬁcantly quenches the ﬂuorescence of perylene and accelerates the photodegradation rate suggesting the involvement of a reactive oxygen species in the observed photodecomposition process. Diffuse reﬂectance laser ﬂash photolysis of perylene samples on silica show transient spectral features consistent with the perylene radical cation (also conﬁrmed by EPR studies) and triplet state. Adsorbed water causes the displacement of perylene on the silica surface and leads to its aggregation as evidenced by absorption and emission spectroscopy. The authors claim that the observed photoproducts exhibit redshifted absorption bands consistent with incorporation of electron donating groups into the perylene’s ring system. However, because the photoproducts were not isolated and characterized, the exact mechanism of phototransformation could not be predicted [35]. A similar study has also been carried out on the photophysics and photochemistry of benzo[e]pyrene, a teratogenic and possibly mutagenic PAH, on non-activated silica and alumina as the model surface to examine the effect of pore size, coadsorbed gases (oxygen and argon), and water [36]. Based on the observed spectral changes upon irradiation of adsorbed benzo[e]pyrene, the authors propose that oxygen actively participates in the destruction pathway and that oxy radical species are involved [36]. A detailed product and mechanism study of pyrene on unactivated and activated silica surfaces has been published by Reyes et al. [37]. It is reported that pyrene photolysis proceeds very slowly to produce 1-hydroxypyrene, 1,6-pyrenedione, and 1,8pyrenedione as the main photoproducts (Scheme 7). Small amounts of minor products which show molecular weights and spectral properties consistent with oxygenated pyrene and 1,1 -bipyrene dimer were also formed at higher loadings of pyrene on the silica surface. On the basis of details of photoproducts and a mechanistic study, it was postulated that the precursor to photoproduct formation was the pyrene cation radical, formed by electron transfer from the pyrene excited state to oxygen (type I) and/or photoioniza-

202

R.M. Pagni · R. Dabestani

Scheme 7

tion of pyrene, which can react with physisorbed water on silica to form the oxygenated products [37]. The photochemistry and photophysics of 2-methyl, 2-ethyl, and 2-tertbutylanthracenes on silica gel was examined by Dabestani et al. [38] to assess the substituent effect on product distribution and mechanism. In contrast to anthracene, which forms ground-state pairs at very low surface coverages (1% of a monolayer) [39], substituted anthracenes show no evidence of pairing even at surface coverages as high as 53% of a monolayer. Further-

Recent Developments in the Environmental Photochemistry

203

more, photodecomposition of 2-substituted anthracenes on silica proceeds more efﬁciently than anthracene itself to produce mainly oxidation products and small amounts of isomeric dimers (Scheme 8). The enhanced reactivity of 2-substituted anthracenes was attributed to the absence of ground-state pairs which can act as a light sink without producing any net chemistry. The data indicated that oxidation products were formed by the addition of singlet molecular oxygen (sensitized by the interaction of the excited state of 2-substituted anthracenes with molecular oxygen) to the ground state of 2-substituted anthracenes while photodimers appeared to form from the crystalline forms of the compounds [38]. Photochemical reactions of substituted phenylethylenes, trans-stilbene and 1,1-diphenylethylene, on silica gel, has been reported by Sigman et al. [40]. Irradiation of trans-stilbene adsorbed on silica gel produced two dimers, along with cis-stilbene, phenanthrene, and a small amount of benzaldehyde which arose from a type II oxidation mechanism (Scheme 9). The formation of photodimers was claimed to be promoted by inhomogeneous surface loading and slow diffusion of trans-stilbene on silica [40]. Photochemical oxidation of 1,1-diphenylethylene on silica produced benzophenone as the major product as well as diphenylmethane and oxidized diphenylethanes as minor products as shown in Scheme 10 [40]. The oxidation of 1,1-diphenylethylene has been proposed to proceed by a type I mechanism [41].

Scheme 8

Scheme 9

204

R.M. Pagni · R. Dabestani

Scheme 10

Since the motion of PAH molecules on surfaces can impact their photochemical rate and behavior, it is important to gain information on the factors that inﬂuence such reaction rates. Fluorescence anisotropy has been employed as a powerful tool to examine the motion of physisorbed and chemically attached PAH molecules on silica surfaces [42, 43]. Molecular dynamics of chemically attached ﬂuorescent probes 1-pyrenebutanol and physically adsorbed 1-pyrenebutyric acid at the solid/air interface of cab-o-sil (fumed silica nanoparticles) have been investigated by optical polarization spectroscopy [42]. The dynamics of ﬂuorescence depolarization for both molecules (low and high surface loadings) were examined under steady-state and time-resolved conditions to learn about the motion of probes and excimer formation. Data indicated that pyrene excimer formation was mostly static in nature for physically adsorbed 1-pyrenebutyric acid, but showed a more dynamic character for the chemically attached 1-pyrenebutanol in the presence of the spacer molecule biphenyl co-attached on the surface. Fluorescence lifetimes were dependent on the concentration of the probe molecules and became shorter at higher surface coverages for both molecules. From the measured in-plane and out-of-plane rotational rate constants it was deduced that co-attached biphenyl spacer molecules provides a 2-D solvent-like environment for the pyrene to experience. Furthermore, in-plane rotation appeared to be sensitive to the nature of a probe’s attachment (physical or chemical) and relatively insensitive to a probe’s concentration. The out-ofplane rotational rate, on the other hand, was sensitive to both attachment and concentration [42]. In related work reported by the same group [43], the rate of molecular motion of pyrene and benzene moieties covalently attached to silica surfaces were determined by 13 C NMR for comparison with ﬂuorescence anisotropy data. A lower limit for the rate of phenyl group motion was estimated to be 1.5 × 105 s–1 from the NMR data consistent with the value of 2 × 106 s–1 determined by time-dependent ﬂuorescence anisotropy [43]. A nanosecond time-resolved diffuse reﬂectance laser ﬂash photolysis technique has also been employed to study the factors that impact energy and electron transfer reactions between co-adsorbed PAHs on silica gel [44]. Using anthracene and 9-anthracenecarboxylic acid as the donors and azulene as the acceptor, the bimolecular quenching rate constants for the triplet state and radical cation of both donor molecules by azulene have been deter-

Recent Developments in the Environmental Photochemistry

205

mined [44]. Based on these ﬁndings, the authors propose that reaction rates for anthracene and 9-anthracenecarboxylic acid on silica was mainly governed by the rate of quencher diffusion on the surface [44]. As outlined above, the photochemistry of PAHs in heterogeneous systems under aerated conditions leads mainly to oxidation products that can be formed by either a type I (electron transfer) and/or type II (energy transfer, singlet molecular oxygen formation) mechanism depending upon the nature of the PAH in question. Although the extent to which each mechanism contributes to the oxidation process depends on a number of factors that involve the PAH itself and the substrate in which it is adsorbed onto, the available data do not provide a clear clue as to how the surface properties can impact facilitating one mechanism over the other. Furthermore, the role of the surface in dictating product distribution and yields as well as the mechanistic pathway(s) that best operate on a particular surface are still not well understood. Thus, additional studies that clearly deﬁne the factors that can inﬂuence photodegradation of PAHs in heterogeneous media would greatly beneﬁt this ﬁeld of research.

5 Photochemistry of PCBs 5.1 Introduction Interest in the photochemistry of PCBs, both in the environment and the laboratory, continues to be quite strong [45–47]. Scientists want to learn more about the long-term fate of these compounds in the environment and discover more effective ways of destroying them in their various settings. Photochemistry plays an important role in both of these endeavors. Research carried out prior to circa 1997–1998 may be found in our earlier review in this series [1]. Other reviews which cover the topic, at least in part, are available, though mostly not in English [48–56]. The current chapter will cover the recent literature through the middle of 2003. Because many of the papers described below use a standardized numbering system for each of the 209 chlorobiphenyls, the interested reader is referred to papers where the system is described [57, 58]. The positions in the biphenyl are numbered as shown in Scheme 11.

Scheme 11

206

R.M. Pagni · R. Dabestani

5.2 Photochemistry in Solution The photochemistry of several PCB congeners in alkanes [59–63] and 2-propanol and alkaline 2-propanol [64–71] has been reported. Before describing this chemistry, it is worth spending a moment considering the factors which dictate whether a chlorobiphenyl will undergo heterolysis or homolysis in a given solvent based on bond dissociation energies of the solvent and its polarity (ET ). Alkanes of course are non-polar, water, the environmental liquid, described in our earlier review, is polar, and 2-propanol is moderately polar. See Table 1. Heterolysis is only observed in the very polar water, i.e. chlorine is replaced by hydroxyl. Of equal importance are bond strengths. An aryl C – H has a bond strength of about 110 kcal/mol. An aryl carbon-centered radical will readily abstract a hydrogen atom from an alkane or 2-propanol because the reaction is exothermic, but it will not do so from water because the reaction is endothermic. Reductive dechlorination is thus observed in hexane, other alkanes and 2-propanol, as described below (Scheme 12). No mechanistic studies have appeared on the photochemistry of PCB mixtures or their congeners in water, the environmental liquid, during the reporting period. Several studies on the photochemistry of chlorobiphenyls in Table 1 Bond dissociation energies and polarities of selected solvents Solvent

Bond dissociation energy (kcal/mol)a

Polarity (ET (30)) (kcal/mol)b

Water Alkane

119 (OH) 98 (1◦ CH) 95 (2◦ CH) 92 (3◦ CH) 91 (2◦ C – H)

63.1 31.0

2-Propanol a

48.4

CRC Handbook of Chemistry and Physics (1982) CRC Press, Boca Raton, FL Reichardt C. Solvents and Solvent Effects in Organic Chemistry, 2nd edn (1988) VCH, Weinheim, Germany b

Scheme 12

Recent Developments in the Environmental Photochemistry

207

alkanes have been reported, however [59–63]. These studies may have relevance to the behavior of the compounds in water. As shown below, the initial steps in the photohydrolysis of chloroarenes (ArCl) in water and the reductive dechlorination in alkanes are the same. The rates of decomposition in the polar and non-polar media may thus be correlated. In any event, photochemistry in an alkane is effective at dechlorinating chlorobiphenyl congeners without destroying the biphenyl ring. Xu and co-workers examined 3,3 ,4,4 ,5-pentachlorobiphenyl [59] and eight di- thru hexachlorobiphenyls [61] in hexane. Masazaki et al. studied all 209 PCB congeners in decane [60], a real tour de force, while Wang and co-workers examined 22 congeners in hexane [62, 63]. In all cases the photoreactions were pseudo ﬁrst order, the dominant reaction was reductive dechlorination in which chlorine is replaced with hydrogen (the fate of the chlorine is unknown); ortho chlorines (2,2 ,6,6 ) were removed faster than meta (3,3,5,5 ) and para (4,4 ) chlorines, probably because the homolysis of an ortho C – Cl bond is exothermic while that at the meta and para sites is endothermic, and in a few instances, migration of a chlorine atom from one site to another on the biphenyl was observed. As was the case where alkanes were used as solvents, a variety of PCB mixtures and congeners have been investigated in 2-propanol [69], 2-propanolH2 O (9 : 1) [64], and alkaline 2-propanol [65–68, 70, 71]. It may be assumed that base (NaOH) was added to the solvent to react with the HCl as it formed. In all cases reductive dechlorination, i.e. replacement of Cl with H, was the dominant photochemical reaction, presumably occurring via free radical chemistry. For the polychlorobiphenyl congeners, multiple losses of chlorine occurred. More chlorinated biphenyls were more reactive than less substituted ones [65]. When the biphenyl was unequally substituted with chlorine, the more substituted ring was more reactive [64, 65, 67]. Sterically hindered chlorines such as those at the ortho positions (2,2 ,6,6 ) and those adjacent to other chlorines were removed faster than non-hindered chlorine atoms [64]. Destruction of the biphenyl ring and polymer formation were also seen [66]. Rearrangement of the chlorine atoms on the biphenyl rings was also detected [68]. As was true for alkane solvents, it is not clear if these results in 2-propanol have any relevance to those in water. What is clear, however, is that photochemistry in 2-propanol is a very effective method of dechlorinating PCB mixtures and their congeners. Whether an alkane or 2-propanol is a better solvent for the dechlorination has not yet been addressed, however. Sinkkonen and Paasivirta have taken literature data to estimate the halflives of 11 tri- through heptachlorobiphenyls in air, water, and soil/sediment in the Baltic region of Europe at 7 ◦ C [72, 73]. They assumed that the degradation in air was totally photochemical in nature, but in water and soil/sediment occurred also by microbial degradation. For a given chlorobiphenyl, the half-life decreased in the order: soil/sediment > water > air. The slower rates of decomposition in water and soil/sediment may be due

208

R.M. Pagni · R. Dabestani

Scheme 13

to attenuation of light as it penetrates the medium. The presence of excited state quenchers on the soil/sediment may also account for the very slow rates of decomposition in this medium. The half-life of PBC 180, 2,2 ,3,4,4 ,5,5 -heptachlorobiphenyl, in soil/sediment, for example, is calculated to be 3.33 × 105 hours; PCB 28, 2,4,4 -trichlorobiphenyl, on the other hand, is expected to have a half-life of 72 hours in air. This contrast in rates also illustrates another trend in the data: as the chlorine content of the biphenyl increases, independent of the medium being examined, the rate of decomposition goes down signiﬁcantly. This suggests that the PCB congeners are undergoing oxidative degradation. On a related note Múˇcka and co-workers have studied the degradation of PCBs in alkaline 2-propanol by very energetic electrons (4.5 MeV) [74]. Although the PCBs were consumed, no organic products were identiﬁed. Chloride was, however, identiﬁed and quantiﬁed. It can be assumed that the dominant chemistry here is also reductive dechlorination (Scheme 13). 5.3 Photochemistry in Surfactant Solutions A number of studies have appeared on the photochemistry of chlorobiphenyls in surfactant/water solutions [75–87]. The surfactants used in these studies are shown in Table 2. Sodium dodecylsulfonate, SDS, is an ionic surfactant, while the others are non-ionic polyethers. Surfactants form micelles which solubilize the hydrophobic chlorobiphenyls. Surfactant/water solutions are also effective at extracting chlorobiphenyls from soils. Once inside the nonpolar interiors of the micelles, the chlorobiphenyls will undergo reductive dechlorination photochemically. Ghosh and co-workers have shown in several studies that Brij 35 and Pol 10 in water provide suitable environments for the photochemically induced reductive dechlorination of chlorobiphenyls [75, 77, 82, 83]. Ghosh and Sayler have demonstrated that photochemistry and microbes work symbiotically to degrade PCB mixtures [76, 80]. Photochemistry in a surfactant/water solution converts highly chlorinated biphenyls into much less chlorinated ones. The microbes, which do not oxidize the highly chlorinated biphenyls, then oxidize the less chlorinated biphenyls. For a related study on the combined degradation of a PCB mixture and microbes see [88]. Several quantitative studies (rates, quantum yields) for the photodegradation of chlorobiphenyls in surfactant solutions have been reported [79, 85–87]. Jafvert and co-workers have shown that Brij 58/water is a more effective medium for the photodecomposition of 2,3,4,5-tetrachlorobiphenyl

Recent Developments in the Environmental Photochemistry

209

Table 2 Names and structures of surfactants Surfactant Structure

Average value of n Reference

SDS Brij 35 Brij 58 Pol 10

CH3 (CH2 )10 SO3 Na CH3 (CH2 )11 (OCH2 CH2 )n OH CH3 (CH2 )15 (OCH2 CH2 )n OH CH3 (CH2 )11 (OCH2 CH2 )n OH

– 23 20 10

Triton X-100 Tween 80

(CH3 )2 CH(CH2 )5 (OCH2 CH2 )n OH

10

83, 85 75, 85 79 76, 77, 80, 82, 83 84

20 (w + x + y + z) 85–87

than water alone (higher quantum yield) [79]. Degradation of PCBs extracted from soil by Brij 58/water was counterbalanced by a loss of surfactant to the soil. Chu and Kwan have examined the effect of acetone (sensitizer) and triethylamine (sensitization through electron transfer) on the rate of photodecomposition of 4,4 -dichlorobiphenyl in SDS-, Brij 35- and Tween 80-water solutions [85]. Acetone accelerated the degradation at low concentration but retarded it at higher concentration, perhaps by light attenuation or self-quenching. Triethylamine, a good electron donor to chlorobiphenyl excited states, also had a good effect on the degradation. Interestingly, increases in the pH of the medium led to faster rates of decay of the substrate. Chu and Kwan have also studied the effect of added humic acid, a substance extractable from soil, on the photochemistry of 4,4 -dichlorobiphenyl in Tween 80/water [86]. At low concentrations (up to 1 mg/L) humic acid promoted degradation (source of hydrogen) while at higher concentrations it retarded it, probably by acting as an excited state quencher. Chu and Kwan lastly examined the effect that added acetone and squalane, an alkane and source of hydrogen, had on the photochemistry of 4,4 -dichlorobiphenyl in Tween 80/water micellar solution [87]. The effect of acetone was identical to what they reported earlier. Squalane enhanced the degradation at low concentration but retarded it at higher concentrations. The authors speculated that the retardation is due to an emulsifying effect, resulting in turbidity of the solution and thus to light scattering.

210

R.M. Pagni · R. Dabestani

5.4 Photochemistry at Interfaces 5.4.1 Silica and Siloxanes Just as the interiors of micelles in water provide a hydrophobic environment to solubilize chlorobiphenyls and then photodegrade them by reductive dechlorination, so do octadecyl-functionalized silica gel [89], normally used as reverse phase packing in HPLC and for solid-phase extraction, and the liquid-semisolid polydimethylsiloxane [– OSi(CH3 )2 –]n [90, 91], also used for solid-phase extraction. In both media, reductive dechlorination is the primary photochemical pathway. 5.4.2 Titanium Dioxide Numerous papers have appeared describing the photodegradation of chlorobiphenyls on the semiconductor, titanium dioxide (TiO2 ) [92–106]. This is not surprising because previous research, as described in our earlier review [1], has shown semiconductor photochemistry to be effective at destroying PCBs. Photochemistry in TiO2 slurred with water serves to generate the hydroxyl radical which can add to the biphenyl ring, both at C – H and C – Cl positions, to make phenols which are susceptible to further photochemical oxidation or to microbial oxidation. TiO2 -induced photochemistry of chlorobiphenyls ultimately mineralizes chlorobiphenyls, i.e. converts them into H2 O, CO2 and HCl. Studies have appeared on the use of TiO2 -induced photochemistry of chlorobiphenyls to purify water [94] and clean up ﬂue gases generated from waste incineration [100]; a batch reactor has also been developed [97]. Various schemes have been examined to improve or modify the destruction of chlorobiphenyls on TiO2 . These include the addition of oxidizers [92, 99], surfactants [102, 103], and metal ions [92, 104, 105]. The effect of added humic acid has also been examined [98]. Several quantitative studies of the photodegradation of 2-chlorobiphenyl on TiO2 from the group of Hong and co-workers have appeared [95, 96, 98, 99, 101, 102] and will be the focus of further discussion. In the ﬁrst of the six papers [95] Hong and his students examined the photochemistry of 2-chlorobiphenyl in TiO2 /H2 O slurries, in the presence and absence of O2 . They identiﬁed 12 aromatic products, 2-hydroxybiphenyl, the seven 2-chlorobiphenylols, 2-chlorobenzaldehyde, 2-chloroacetophenone, benzoic acid, and 2-chlorobenzoic acid, from the photochemistry (Scheme 14). The hydroxyl radical clearly played a dominant role in the degradation. The disappearance of 2-chlorobiphenyl and its degradation

Recent Developments in the Environmental Photochemistry

211

products was accelerated by dissolved gases in the order O2 > air > N2 . The O2 may affect the reaction by accepting an electron from the conduction band of TiO2 , thus facilitating the formation of · OH by reaction of the hole in the valence band with water. Other mechanisms are also possible. Hong, Wang and Bush showed that the adsorption of 2-chlorobiphenyl from water onto TiO2 ﬁts nicely onto a Langmuir isotherm [96]. The subsequent pseudo ﬁrst order decay of the substrate under illumination obeyed Langmuir–Hinshelwood kinetics with a quantum yield of 5 × 10–3 . The same organic products were observed here as in their earlier study. If the reaction was run long enough, mineralization occurred. In another study Hong, Wang and Fang examined how a series of naturally occurring addends affected the TiO2 photocatalyzed degradation of 2-chlorobiphenyl [98]. Lower pHs, for example, led to faster degradation. At lower pHs the surface of TiO2 is positively charged (isoelectric pH = 6.3) which facilitates · OH formation. At lower pHs the production of H2 O2 from the reaction of O2 – with H+ is also facilitated; the photochemistry of H2 O2 also generates · OH. Cl– , abundant in natural waters, did not affect degradation at pH = 10, but inhibited it at pHs of 7 and 4. Cl– adsorbs to the surface of TiO2 at neutral and acidic pHs. CO3 –2 /HCO3 – likewise inhibited the decomposition of the substrate. Mg+2 , also abundant in natural waters, inhibited the reaction in base (pH = 11) but had no effect at pH = 4. This may be due to the adsorption of Mg+2 at high pH to active sites on the surface of TiO2 . Humic acid, also common in natural waters, had an inhibitory effect on the degradation, possibly by acting as a light ﬁlter. Hong and Wang also examined the effect of the oxidizers, H2 O2 , S2 O8 –2 , and IO4 – , on the photochemistry of 2-chlorobiphenyl in TiO2 /water slurries [99]. The addition of the oxidizer to TiO2 resulted in a faster rate of decomposition than TiO2 alone, but a slower rate than the oxidizer alone, i.e. rate (oxidizer) > rate (oxidizer + TiO2 ) > rate (TiO2 ). This can be explained by the fact that photodegradation occurs both in solution (by oxidizer) and on

Scheme 14

212

R.M. Pagni · R. Dabestani

the surface of TiO2 , with the reaction on TiO2 being suppressed by adsorption of the oxidizer. In another study Hong and Wang examined the effect of another potential oxidizer, O2 , on the photodecomposition of 2-chlorobiphenyl in TiO2 /water slurries [101]. By seeing what effect partial pressures of O2 had on the decomposition of the substrate they deduced that the rate of decomposition ﬁrst increased rapidly and then levelled off as the O2 pressure increased, consistent with O2 adsorbing reversibly to the TiO2 and obeying Langmuir– Hinshelwood kinetics. The adsorbed O2 scavenges electrons from the conduction band, thus facilitating the formation of · OH. Oxygen also assisted in the photodecomposition of the primary photoproducts: 2-hydroxybiphenyl and the seven 2-chlorohydroxybiphenyls. In the last of the six papers Hong and Huang have examined how added surfactants such as FC-143, ammonium perﬂuorooctanoate [CF3 (CF2 )6 CO2 NH4 ], inﬂuenced the TiO2 -photocatalyzed decomposition of a large number of PCB congeners adsorbed on soils [102]. The authors concluded that the perﬂuorinated surfactants are photostable and do not get directly involved in the photochemistry of the chlorobiphenyls (unlike the surfactants mentioned earlier), and the surfactants extract the PCBs from the soil which are then expeditiously photoxidized on TiO2 . This represents a novel two-stage procedure to remove PCBs from soil and then destroy them. Krauss and Wilcke examined the TiO2 -photocatalyzed oxidation of 12 PCB congeners [and 20 polycyclic aromatic hydrocarbons (PAHs)] on various soil samples (four mineral topsoil horizons, six organic horizons, and four particle-site fractions in three different soils) [107]. When the TiO2 /soil mixture was irradiated in the absence of H2 O, no photooxidation of the chlorobiphenyls occurred. When slurried with water, however, chlorobiphenyl concentrations decreased by 40–50% after 48 hours of irradiation, while the PAH concentrations were unchanged. By way of contrast, PAHs and PCBs doped onto quartz sand diminished by 95–100% after 8 hours of photolysis. The pollutants are clearly more accessible to hydroxyl radicals on sand than on soil. It is also clear that the photooxidation occurred in the soil and not in solution. Thus, · OH is generated on one surface (TiO2 ), diffuses in the water to the other surface (soil), where the oxidation occurs. Duffy and co-workers have considered a two-stage scheme to clean up drinking water, ground water, and waste water [108]. The ﬁrst stage consists of a wet air oxidating (WAO) (oxidation with O2 under pressure and temperature; often catalytic) [109]. The second stage involved TiO2 -photocatalyzed oxidation, with acetic acid and 2-chlorobiphenyl being examined in this study. The substrates were chosen because they are recalcitrant to WAO. 2-Chlorobiphenyl was completely degraded photochemically at 85◦ in less than one minute. It was also possible to degrade selectively 2-chlorobiphenyl in the presence of excess acetic acid at neutral and basic pH. Crittenden and co-workers have developed a different water treatment strategy involving

Recent Developments in the Environmental Photochemistry

213

adsorption of the organic pollutants onto Ambersorb 563, a carbonaceous adsorbent with a very large surface area, desorption of the organics with steam (concentrates pollutants), and cleanup of the steam condensate on a ﬁxed bed photoreactor consisting of 1% Pt/TiO2 on silica gel [110]. Better than 98% of the 2-chlorobiphenyl in the initially polluted water could be destroyed in this manner. 5.5 Reactions of Photogenerated Hydroxyl Radicals in Air and Water It is well established that concentrations of PCBs in the atmosphere around the Great Lakes of North America have decreased markedly in the last few decades [111]. The mechanism of their removal from the troposphere involves their reaction with hydroxyl radicals. Anderson and Hites have developed a laboratory system (Scheme 15) to measure rate constants for the reaction of weakly volatile compounds with · OH in the gas phase [112]. · OH is generated by the photodecomposition of O3 in the presence of water vapor. At 298 K, for instance, the reaction of 4,4 -dichlorobiphenyl with · OH yielded k = (2.0 ± 0.5) × 10–12 cm3 s–1 . From this value the researchers estimated the lifetime of the substrate in the troposphere at 12 days. Brubaker and Hites then investigated the chemistry of biphenyl and eight chlorobiphenyls (2-,3and 4-chloro, 2,4-, 2,5-, 2,2 -, 2,3 ,-, and 2,4 -dichloro) with · OH in the gas phase [113]. In all cases benzoic acids were generated. It appears that · OH attacked one of the two benzene rings followed by attack of atmospheric O2 onto the resulting carbon-centered radical. It is likely that this type of chemistry takes place in the troposphere. Several papers have been published describing the use of ozone and UV radiation to clean up water [114–116]. Of note is the pilot plant operation

Scheme 15

214

R.M. Pagni · R. Dabestani

developed by Niessner and co-workers to purify waste water at high ionic strength [114]. In addition to O3 , they examined the effect of H2 O2 and O3 /H2 O2 plus light on the clean up. It is likely that · OH is the active oxidant in all cases. Besides chlorobiphenyls, PAHs and chlorinated phenols, dioxins, and furans were studied. Depending on the structure of the chlorobiphenyl, 23–96% of the pollutant could be removed from the water. Minami et al. also successfully degraded PCBs in water using ozone and light [116]. Kuo and Lo have reported their results on the use of a so-called photoFenton process to degrade 4-chloro- and 4,4 -dichlorobiphenyl in water [117]. Fenton’s reagent consisting of Fe+2 and H2 O2 generates · OH by the reaction: Fe+2 + H2 O2 6 Fe+3 + OH– + · OH. Normally, the photo-Fenton reagent is made up of Fe+3 + H2 O2 which yields · OH photochemically. In this instance, however, the researchers photolyzed the ordinary Fenton’s reagent (Fe+2 /H2 O2 ). Interestingly, the chlorobiphenyls degraded more quickly when light was present than in its absence. It is clear that the addition of light results in a greater concentration of · OH. 5.6 PCBs on Sediments and in Oils Four papers have appeared on the photodegradation of PCBs on sediments [118–121]. Tang and Myers, for example, have carried out a study to model the behavior of PCBs on sediments conﬁned in disposal facilities [118, 120]. In their glass aquariums the authors found PCB levels decreased by 40% over a 5-month span. Unfortunately, they were unable to pin point the source of the loss. Tang and Meyers believed the loss was due to a combination of volatilization, photodegradation and biodegradation. Poster et al. used a different approach to degrade PCBs on sediments [121]. PCBs in contaminated marine sediments, when mixed with water, were effectively dechlorinated by electrons from a linear accelerator. Adding 2-propanol enhanced chlorobiphenyl dechlorination. Photolysis of the contaminated sediment, on the other hand, led to little dechlorination. Adding triethy-

Scheme 16

Recent Developments in the Environmental Photochemistry

215

lamine, however, did lead to signiﬁcant dechlorination. Triethylamine opened up a new photochemical pathway involving photoinduced electron transfer (Scheme 16). Photochemistry may prove to be as effective as the electron beam in dechlorination of PCBs on sediments. Three papers have been published on the photodegradation of PCBs in transformer and insulation oils [122–124]. Photolyses successfully degraded PCBs in oils mixed with ground water in spite of the turbidity of the mixtures. 5.7 Ice Photochemistry Interest in the photochemistry of inorganic and organic compounds on water ice is quite high because of its relevance to astrochemistry, ozone depletion in the upper atmosphere, and degradation of pollutants by sunlight. Although no studies have appeared describing the photochemistry of PCBs on ice, several have appeared on the closely related chlorobenzenes [24, 29– 31, 125, 126]. Although liquid water provides a very polar environment for PCBs and chlorobenzenes to undergo photoheterolysis (Cl → OH), this is not the case for ice. It is likely that the surface of ice is also polar, perhaps even more so than the liquid [127]. Water molecules on the solid’s surface are not available for chlorine replacement, however. As a result, no phenols are produced on ice. Instead other free radical and ionic pathways are opened up. For example, photolysis of chlorobenzene on ice yielded 22 non-phenolic products (Scheme 17) [29].

Scheme 17

Acknowledgements This research was sponsored by the division of Chemical Sciences, Geosciences, and Biosciences, Ofﬁce of Basic Energy Sciences, U.S. Department of Energy under contract DE-AC05-960OR22464 with Oak Ridge National Laboratory, managed by UT-Battelle for the Department of Energy.

216

R.M. Pagni · R. Dabestani

References 1. Pagni RM, Sigman ME (1999) The Photochemistry of PAHs and PCBs in water and on solids. In: Boule P (ed) The Handbook of Environmental Chemistry, vol 2, part 2. Springer, Berlin Heidelberg New York, p 139 2. Neff JM (1979) Polycyclic aromatic hydrocarbons in the aquatic environment: sources, fates and biological effects. Applied Sciences, London 3. Cook RH, Pierce RC, Eaton PB, Lao RC, Onuska FI, Payne JF, Vavasour E (1983) Polycyclic aromatic hydrocarbons in the aquatic environment: sources, fate and effects on aquatic biota. NRCC 18981, National Research Council of Canada, Ottawa, Ontario, Canada 4. Lee RF, Gardner WS, Anderson JW, Blaylock JW, Barwell-Clarke J (1978) Environ Sci Technol 12:832 5. Rivas FJ, Beltran FJ, Acedo B (2000) J Hazardous Materials 75:89 6. Lehto K-M, Vuorimaa E, Lemmetyinen H (2000) J Photochem Photobiol A Chem 136:53 7. Vialaton D, Richard C, Baglio D, Paya-Perez A-B (1999) J Photochem Photobiol A Chem 123:15 8. Endicott JF, Feraudi G, Barber JR (1975) J Phys Chem 79:630 9. McConkey BJ, Hewitt LM, Dixon DG, Greenberg BM (2002) Water, Air, Soil Pollution 136:347 10. Grabner G, Rechthaler K, Mayer B, Kohler G, Rotkiewicz K (2000) J Phys Chem A 104:1365 11. Mallakin A, Dixon DG, Greenberg BM (2000) Chemosphere 40:1435 12. Bertilsson S, Anneli W (2002) Hydrobiologia 469:23 13. Zeng K, Hwang H-M, Yu H (2002) Inter J Mol Sci 3:1048 14. Fasnacht M, Blough NV (2002) Environ Sci Technol 36:4364 15. Ledakowicz S, Miller JS, Olejnik D (2001) Int J Photoenergy 3:95 16. Miller JS, Olejnik D (2000) Water Research 35:233 17. Ledakowicz S, Miller JS, Olejnik D (1999) Int J Photoenergy 1:55 18. Sabate J, Bayona JM, Solanas AM (2001) Chemosphere 44:119 19. Moeini-Nombel L, Matsuzawa S (1998) J Photochem Photobiol A Chem 119:15 20. Wang Y, Lin F, Lin Z, Xu Z, Tang Y (1999) Chemosphere 38:1273 21. Mayer E, Pletzer R (1986) Nature 319, 6051:298 22. Mukai T (1986) Astron Astrophys 164:397 23. Grim RJA, Greenberg JM (1987) Astron Astrophys 181:155 24. Klan P, Holoubek I (2002) Chemosphere 46:1201 25. Allamandola LJ, Tielens AGGM, Barker JR (1989) Astrophys J Suppl Ser 71:733 26. Puget JL, Leger A (1989) Annu Rev Astron Astrophys 27:161 27. Bernstein MP, Sandford SA, Allamandola LJ, Gillette JS, Clemett SJ, Zare RN (1999) Science 283:1135 28. Sandford SA, Bernstein MP, Allamandola LJ, Gillette JS, Zare RN (2000) Astrophys J 538:691 29. Klan P, Ansorgova A, Favero D, Holoubek I (2000) Tetrahedron Lett 41:7785 30. Holoubek I, Klan P, Ansorgova A, Favero D (2000) Organohalogen Compds 46:228 31. Klan P, Del Favero D, Ansorgova A, Klanova J Holoubek I (2001) Environ Sci Pollut R 8:195 32. Klanova J, Klan P, Nosek J, Holoubek I (2003) Environ Sci Technol 37:1568 33. Dubowski Y, Hoffmann MR (2000) Geophys Res Lett 20:3321 34. Teinemaa E, Kirso U (1999) Polycyclic Arom Compds 14, 15:275

Recent Developments in the Environmental Photochemistry

217

35. Sotero P, Arce R (1999) Polycyclic Arom Compds 14, 15:295 36. Fioressi S, Arce R (1999) Polycyclic Arom Compds 14, 15:285 37. Reyes CA, Medina M, Crespo-Hernandez C, Cedeno MZ, Arce A, Rosario O, Steffenson DM, Ivanov IN, Sigman ME, Dabestani R (2000) Environ Sci Technol 34:415 38. Dabestani R, Higgin J, Stephenson DM, Ivanov IN, Sigman ME (2000) J Phy Chem B 104:10 235 39. Dabestani R, Ellis KJ, Sigman ME (1995) Photochem Photobiol A Chem 86:231 40. Sigman ME, Barbas JT, Corbett S, Chen Y, Ivanov IN, Dabestani R (2001) J Photochem Photobiol A Chem 138:269 41. Aronovitch C, Mazur Y (1985) J Org Chem 50:149 42. Ivanov IN, Dabestani R, Buchanan AC, Sigman ME (2001) J Phys Chem B 105:10 308 43. Sigman ME, Read S, Barbas JT, Ivanov IN, Hagaman EW, Buchanan AC, Dabestani R, Kidder MK, Britt PF (2003) J Phys Chem A 107:3450 44. Worrell DR, Kirkpatrick I, Williams SL (2002) Photochem Photobiol Sci 1:896 45. Robertson LW, Hansen LG (eds) (2001) PCBs, Recent Advances in Environmental Toxicology, Health Effects. University of Kentucky Press, Lexington 46. Fieder H, Lau C (1998) In: Schürmann G, Markert B (eds) Ecotoxicology-Ecological Fundamentals, Chemical Exposure, Biological Effects. Wiley, New York, p 317 47. Muszkat L (1998) In: Kearney PC, Roberts T (eds) Pesticide Remediation in Soils, Water. Wiley, New York, p 307 48. Ulman JA (1997) Chimica Oggi 15:21 49. Lee BD, Kobayashi A, Hosomi M (1999) Yosui to Haisui 41:483 50. Fujita M (1999) Kagaka Keiza 46:26 51. Masuaki H (2000) Kagaka Kogaku 64:135 52. Gavrancic S, Skala D (2000) Hemijska Indutrija 54:53 53. Watanabe M (2000) Materials Integration 13:11 54. Matsunaga A, Yasuhara A (2002) Kankyo Kagaku 12:33 55. Taoda H (2002) Kogyo Zairyo 50:55 56. Kim YJ, Osako M (2002) Haikibutsu Gakkai Ronbunshi 13:341 57. Ballschmiter K, Zell M (1980) Fresenius Zeit Anal Chem 302:20 58. Schulte E, Malisch R (1983) Fresenius Zeit Anal Chem 314:545 59. Miao XS, Chu SG, Xu XB (1997) Toxicol Environ Chem 61:223 60. Masuzaki Y, Matsumura T, Tsubota H, Ikeda Y, Chisaki Y, Morita M, Ito H (1998) Organohalogen Compds 36:385 61. Miao XS, Chu SG, Xu SB (1999) Chemosphere 39:1639 62. Chang FC, Yen JH, Wang YS (2002) Organohalogen Compds 58:13 63. Chang FC, Chiu TC, Yen JH, Wang YS (2003) Chemosphere 51:775 64. Barr JR, Old T, Kimata K, McClure PC, Lapeza CR Jr, Muggio VL, Hosoya K, Tanaka N, Patterson DG Jr (1997) Organohalogen Compds 33:199 65. Yao Y, Kato Y, Hanada Y, Shinohara R (1997) Organohalogen Compds 33:246 66. Yao Y, Kakimoto H, Ogawa HI, Kato Y, Hanada Y, Shinohara R, Yoshino E (1997) Bull Environ Contam Toxicol 59:238 67. Yao Y, Kakimoto K, Ogawa HI, Kato Y, Hanada Y, Shinohara R, Yoshino E (1997) Chemosphere 35:2891 68. Yao Y, Kato Y, Kadokami K, Shinohara R (1998) Organohalogen Compds 36:381 69. Matsumoto K, Kawabata S, Yamuzaki S (1999) J Adv Oxid Technol 4:279 70. Yao Y, Kakimoto K, Ogawa HI, Kato Y, Kadokami K, Shinohara R (2000) Chemosphere 40:951 71. Yukio N, Muramuta T, Nishizawa K, Ohno M, Sakai SI (2002) Organohalogen Compds 56:413

218

R.M. Pagni · R. Dabestani

72. 73. 74. 75. 76.

Sinkkonen S, Paasivirta J (1998) Organohalogen Compds 36:509 Sinkkonen S, Paasivirta J (2000) Chemosphere 40:943 ˇ Múˇcka V, Silber R, Pospíil M, Camra M, Bartoníˇcek B (2001) Rad Phys Chem 57:489 Shi Z, Sigman ME, Ghosh MM (1997) Proc WEFTEC’97 1:197 Shi Z, LaTorre KA, Ghosh MM, Layton AC, Luna SH, Bowles L, Sayler GS (1998) Water Sci Technol 38:25 LaTorre KA, Shi Z, Ghosh MM, Layton AC (1998) In: Wiekramanayake GB, Hinchee RE (eds) Designing and Applying Treatment Technologies. Remediation of Chlorinated and Recalcitrant Compounds. Battelle Press, Columbus, p 225 Shukla SS, Maheshwan DK (1998) Book of Abstracts. 216th American Chemical Society Meeting Chu W, Jafvert CT, Diehlel CA, Marley K, Larson RA (1998) Environ Sci Technol 32:1989 Sampsel ER, Ghosh MM, Shi Z, Robinsin KG, Sanseverino J (1999) Hazard Ind Waste 31:129 Sanseverino J, Menn FM, Gregory B, Layton A, Sampsel E, Shi Z, Ghosh M, Schultz TW, Sayler G (1999) In: Allerman BC, Leeson A (eds) Int In Situ and On-Site Bioremediation Symp. Battelle, Columbus, 7:155 Sampsel E, Ghosh M, Shi Z, Sanseverino J, Menn FM, Wong MM, Gregory B, Sayler G (1999) In: Allerman BC, Leeson A (eds) Int In Situ and On-Site Bioremediation Symp. Battelle Press, Columbus, 7:161 Shi Z, Ghosh MM, Robinson KG (2000) Environ Sci Pollution Control Series 23:523 Vichutinskuya EV, Skurlatov YI, Ernestova LS, Sansonov DP, Semenova IV (2001) Khim Fiz 20:56 Chu W, Kwan CY (2002) Water Res 36:2187 Chu W, Kwan CY (2003) J Environ Eng 129:716 Chu W, Kwan CY (2003) Water Res 37:2442 Rhoﬁr C, Hawari J (2000) J Chomatogr A 873:53 Olda T, Barr JR, Kimata K, McClure C, Lapeza CR Jr, Hosoya K, Ikegami T, Smith CJ, Patterson DG Jr, Tanaka N (1999) Chemosphere 39:1795 Lores M, Llompart M, González-García R, González-Barreiro C, Cela R (2002) Chemosphere 47:607 Lores M, Llompart M, González-García R, Gonzales-Barreiro C, Cela R (2002) J Chromatogr A 963:37 Matsunaga K, Tamube S, Mori T (1997) Jpn J Toxicol Environ Health 43:174 Kuo CY, Lo SL (1997) J Colloid Interface Sci 196:199 Li T, Chen Z (1998) Haunjin Kexue Xueba 18:167 Wang Y, Hong CS, Bush B (1998) J Adv Oxid Technol 3:299 Hong CS, Wang Y, Bush B (1998) Chemosphere 36:1653 Funayama H, Kurosawa M, Yamakoshi Y, Sugawara K, Sugawa T (1999) Kagaku Kogaku Ronbushu 25:884 Wang Y, Hong CS, Fang F (1999) Environ Eng Sci 10:433 Wang Y, Hong CS (1999) Wat Res 33:2031 Taoda H, Okabe A, Kondo T, Tsuruda H, Shimauchi H, Aizawa K, Suzuki Y (2000) Organohalogen Compds 45:400 Wang Y, Hong CS (2000) Water Res 34:2791 Huang Q, Hong CS (2000) Chemosphere 41:871 Yuan Q, Ravikrishna R, Valsuraj KT (2001) Separ Purif Technol 24:309 Li C, Mai B, Pan H, Lin Z, Sheng G, Fu J (2002) Huanjin Kexue 23:120 Li CL, Mai BX, Pan HX, Lin Z, Jia KX (2002) Zhongguo Huanjing Kexue 22:123

77.

78. 79. 80. 81.

82.

83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105.

Recent Developments in the Environmental Photochemistry 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123.

124.

125.

126. 127.

219

Horikoshi S, Hidaka H (2003) Chemosphere 51:139 Krauss M, Wilcke W (2002) J Plant Nutr Soil Sci 165:173 Duffy JE, Anderson MA, Hill CG Jr, Zeltner WA (2000) Ind Eng Chem Res 39:3698 Luck F (1999) Catal Today 53:81 Suri RPS, Crittenden JC, Hund DW (1999) J Environ Eng 125:897 Buehler SS, Basu I, Hites RA (2002) Environ Sci Technol 36:5051 Anderson PN, Hites RA (1996) Environ Sci Technol 30:301 Brubaker WW Jr, Hites RA (1998) Environ Sci Technol 32:3913 Wenzel A, Gahr A, Niessner R (1999) Water Res 33:937 Hirose H (2000) Petrotech 23:758 Minami H, Horii Y, Nakao T, Hideaki M (2001) Organohalogen Compds 54:176 Kuo CY, Lo SL (1999) Chemosphere 38:2041 Tang NH, Myers TE (1998) Hazard Ind Waste 30:877 Wernersson AS, Dave G, Nilsson E (1999) Aquat Ecosys Health Manag 2:379 Tang NH, Myers TE (2002) Chemosphere 46:477 Poster DL, Chaychian M, Neta P, Huie RE, Silverman J, Al-Sheikhly M (2003) Environ Sci Technol 37:3808 Muto H, Funayama H, Kurosawa M, Sugawara K, Sugawara T (1997) Akita Daigaku Kozangakubu Kenkyui Hokoku 18:57 Deswael S, Vinnakota VL, Crine M, Bosmon B, Smitz F, Remacle J (1999) In: Hupka J, Miller JD (eds) Environ Technol for Oil Pollution, Proc of the US United Engineering Foundation Conference and 2nd Int Conference Analysis and Utilization of Oily Wastes AUZO. Tech Univ Gdansk, Gdansk, Pol 1:126 Urkiagu A, Marina V, De la Fuentes L, Susaeta I (2002) In: Garaskar AR, Chen ACS (eds) Remediation of Chlorinated and Recalcitrant Compounds–2002, Proc of the Int Conference on Remediation of Chlorinated and Recalcitrant Compounds. Battelle Press, Columbus, Ohio, p 2005 Klan P, Del Favero D, Anorgora A, Holoubek I (2001) In: Midgley JP, Reuther M, Williams M (eds) Transport and Chemical Transformation in the Troposphere, Proc of EUROTEAC Symposium. Springer, Berlin Heidelberg New York, Germany, p 841 Klan P, Klanova J, Holoubek I, Cupr P (2003) Geophys Res Lett 30:4611 Sigman M, Pagni RM, unpublished results

Hdb Env Chem Vol. 2, Part M (2005): 221–253 DOI 10.1007/b138185 © Springer-Verlag Berlin Heidelberg 2005 Published online: 23 September 2005

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions Davide Vione · Valter Maurino · Claudio Minero · Ezio Pelizzetti (u) Dipartimento di Chimica Analitica, Università di Torino, Via P. Giuria 5, 10125 Torino, Italy [email protected] 1 1.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Occurrence of Nitrate and Nitrite in Natural Waters. Their Observed or Expected Photochemical Incidence . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

224 224 226 226 227

Transformation Reactions in the Presence of Organic Compounds . Dark Processes in Acidic Solutions of Nitrate and Nitrous Acid . . . Transformation of Phenol upon Photolysis of Nitrite and Nitrate . . Nitrite Photolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nitrite Photooxidation . . . . . . . . . . . . . . . . . . . . . . . . . Nitrate Photolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanistic Implications . . . . . . . . . . . . . . . . . . . . . . . . Transformation of Substituted Phenols upon Photolysis of Nitrite and Nitrate . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 Nitrite Photolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.2 Nitrate Photolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Transformation of Other Organic Compounds upon Photolysis of Nitrite and Nitrate . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

228 228 230 230 232 234 237

. . . . . . . . . . . .

238 238 240

. . . .

244

Assessment of Oxidation, Nitration and Nitrosation under Environmental Conditions . . . . . . . .

247

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

249

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

250

3 3.1 3.2 3.2.1 3.2.2 3.2.3 3.2.4 3.3

4 5

Direct Photolysis of Nitrate and Nitrite Nitrate Photolysis . . . . . . . . . . . . Photolysis of Nitrite and Nitrous Acid . Nitrite . . . . . . . . . . . . . . . . . . Nitrous Acid . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

223

. . . . .

2 2.1 2.2 2.2.1 2.2.2

. . . . .

222

Abstract This review describes the reaction mechanisms of nitrate and nitrite ions, which play a role in the photochemical processes occurring in natural waters and in atmospheric aerosols. The UV excitation of nitrate ions leads to the formation of hydroxyl radicals, nitrogen dioxide, atomic oxygen and nitrite. Hydroxyl radicals induce the oxidation of organic substrates and account for the photodepollution effect of nitrate in natural waters. Nitrogen dioxide produces nitrate and nitrite, but it can also interact with organic compounds inducing nitration and (in the presence of nitric oxide) nitrosation processes. Further nitration and nitrosation are the consequence of the reactivity (dark or

222

D. Vione et al.

photoinduced) of photogenerated nitrite/nitrous acid. Nitro and nitroso derivatives are compounds of concern, since many of them are mutagenic. Nitrite behaves in a similar way to nitrate, but it also acts as a sink as well as a source of hydroxyl. Oxidation of nitrite by hydroxyl yields nitrogen dioxide, and the prevailing reactions induced by nitrite photolysis depend on its concentration values (oxidation at low concentration, nitration and nitrosation in more concentrated solution). Nitrite can also interact with other environmental factors (dissolved Fe(III), semiconductor oxides, Fenton reagent, nitrate) yielding nitrogen dioxide and therefore enhancing nitration processes. Due to the relevant transformation of organic molecules induced by these photochemical processes, the interaction between irradiated nitrate and nitrite and a number of organic compounds is reviewed and discussed. An attempt has also been made to include the majority of the results in a general framework of reactivity. Keywords Induced photooxidation · Nitrate ion · Nitrite ion · Photonitration · Photonitrosation Abbreviations HPLC-MS-MS High Performance Liquid Chromatography coupled with tandem Mass Spectrometry acidic dissociation constant (pKa = – log10 Ka ) Ka hν radiation (h: Planck’s constant; ν: frequency, s–1 ) λ wavelength (nm) M molar concentration (mol l–1 ); mM = 10–3 M, µM = 10–6 M UVA ultraviolet radiation (range: 315–400 nm) UVB ultraviolet radiation (range: 280–315 nm)

1 Introduction Phototransformation plays an important role in the degradation of organic substances present in natural waters, particularly in the case of poorly biodegradable pollutants. Direct photolysis takes place with substrates absorbing solar light, i.e. wavelengths longer than about 300 nm. In many instances, however, organic molecules do not absorb sunlight to an appreciable extent and direct photolysis reactions are negligible. In these cases, photodegradation can be induced by environmentally occurring compounds such as inorganic ions (e.g. nitrate and nitrite), transition metals (e.g. FeIII salts) and natural organic molecules containing carbonyl or phenolic groups (e.g. humic substances). These substances produce radical or oxidising species upon irradiation [1–4]. The present chapter focuses on the transformation reactions induced by the irradiation of nitrate and nitrite ions. It is the development of a chapter published in the 1999 edition (Boule P, Bolte M, Richard C (1999) Phototransformations induced in aquatic media by NO3 – /NO2 – , FeIII and humic substances).

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

223

1.1 Occurrence of Nitrate and Nitrite in Natural Waters. Their Observed or Expected Photochemical Incidence Nitrate ions are usually present in natural waters, including cloud water. Their concentration in sea water changes with latitude, most likely reﬂecting biological activity. Concentration levels of 30 µM have been reported near the Antarctic peninsula [5] and 10 µM in the Sargasso Sea [6]. In the case of freshwater, the values are usually higher, also due to the use of fertilisers in agriculture, and usually span in the range 0.1–1.0 mM [7, 8]. In the atmospheric aqueous phase, the typical concentration of the nitrate ion is again in the range 0.1–1.0 mM [9]. Ammonium and sodium nitrates are also important constituents of atmospheric particulate matter [10, 11]. Nitrate photolysis is a relevant source of hydroxyl in natural waters. A study carried out on the Greifensee Lake water indicates that nitrate photolysis is a much more important source of hydroxyl when compared with the photolysis of hydrogen peroxide or the Fenton reaction [12]. Nitrate photochemistry can thus lead to a steady-state hydroxyl concentration around 5 × 10–16 M [8]. Nitrite is usually present in the environment at a lower concentration than nitrate, but its higher molar absorptivity and photolysis quantum yield can make it a competitive photoreactant under environmental conditions [13]. Usual concentration values of nitrite are below 2 µM in seawater [13], below 0.1 mM in surface waters [7] and around 0.1–0.5 µM in the atmospheric aqueous phase in unpolluted areas [9]. Nitrite concentration was, however, found to reach up to 75 µM in fog water from California’s Central Valley, and nitrite photochemistry was shown to account for 50–100% of hydroxyl formation upon irradiation of the collected water samples [14]. One of the environmental sources of nitrite is represented by the irradiation of nitrate itself. The fact that nitrite, too, is a photoactive compound implies that the photochemical reactivity of nitrate and nitrite cannot often be dissociated, although the relative contributions to hydroxyl generation can be derived from the concentration values, photolysis quantum yields and radiation absorption calculations [6, 8, 12, 14]. Nitrite photochemistry can also be a relevant sink of nitrogen in surface waters [12]. For instance, the irradiation of aqueous NH4 NO2 was shown to yield N2 upon reaction between photoformed N2 O3 and NH3 , the reaction being a loss process for dissolved nitrogen [15]. Due to the generation of hydroxyl radicals, the photolysis of nitrate and nitrite can lead to the degradation of organic compounds [16]. This effect can result in decontamination induced by photolytic processes in natural waters, leading to the photoinduced degradation of organic pollutants. This depolluting role of nitrate and nitrite photochemistry is, however, reduced by the competition between natural organic matter, bicarbonate ions and organic

224

D. Vione et al.

pollutants for reaction with hydroxyl [13]. Furthermore, the overall antipollution role can be counterbalanced by the fact that the UV irradiation of solutions containing aromatic compounds and nitrate or nitrite can lead to the formation of mutagens [17–19]. Also the irradiation of aromatic amino acids (in particular tryptophan) was observed to lead to mutagen formation upon irradiation in the presence of nitrate and nitrite, with a larger effect observed for nitrite irradiation [20]. Detailed analytical and mechanistic studies are useful for a better understanding of all the reported processes. They have been the subject of several recent publications, and the main results are presented in the subsequent sections. Knowledge of the mechanisms is useful to assess photochemical transformation under environmental conditions.

2 Direct Photolysis of Nitrate and Nitrite 2.1 Nitrate Photolysis The UV spectrum of nitrate shows two absorption bands, a stronger one near 200 nm and a weaker one around 300 nm ([21, 22], see also Fig. 1). The weaker band, showing a maximum at 302 nm but extending towards the longer wavelengths up to at least 340 nm, accounts for the ability of nitrate to absorb sunlight in the UVB and UVA regions [23–25]. When considering both nitrate

Fig. 1 Left Y-axis scale: UV absorption spectra (molar absorptivity ε, M–1 cm–1 ) of NO3 – , NO2 – and HNO2 . Right scale: spectral solar photon irradiance in Central Europe (photons cm–2 s–1 nm–1 ) according to [25] (15th June, 15th December)

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

225

molar absorbivity and sunlight intensity, the latter increasing with increasing wavelength, it can be concluded that the sunlight absorption rate of nitrate under midday, midsummer, 50◦ N latitude irradiation conditions has a maximum around 320 nm [26]. UVA and UVB radiation absorption by nitrate results in photolysis, following two main pathways. The ﬁrst one yields · NO2 and · OH, the second NO2 – and O(3 P) [22, 23, 27, 28]: [Φ1 (305–313 nm) ≈ 0.01] (1) NO3 – + hν → · NO2 + O· – – + · · [pKa 2 = 11.9] O + H OH (2) NO3 – + hν → NO2 – + O(3 P)

[Φ3 (305 nm) ≈ 0.001]

(3)

Hydroxyl radicals can then oxidise other dissolved molecules or undergo dimerisation, oxygen atoms react with nitrate and nitrite [22, 27, 29], while nitrogen dioxide undergoes dimerisation and hydrolysis [30]: · OH + · OH → H2 O2 [k4 = 5.5 × 109 M–1 s–1 ] (4) O(3 P) + NO3 – → O2 + NO2 – 3

O( P) + NO2 → NO3 · NO2 + · NO2 N2 O4 –

(5)

–

(6) 8

–1 –1

[k7 = 4.5 × 10 M s ; k–7 = 6900 s–1 ] [k8 = 1000 s–1 ]

(7)

N2 O4 (+ H2 O) → NO3 – + NO2 – + 2H+ (8) – · · The species OH, NO2 , N2 O4 and NO2 /HNO2 are of particular interest for the further reactions that can take place in irradiated nitrate solutions in the presence of other molecules [23, 24]. It is interesting to observe that the irradiation of nitrate also results in the isomerisation to peroxynitrite/peroxynitrous acid. This reaction has been conclusively demonstrated to take place at λ < 280 nm. Its occurrence at higher wavelength is still uncertain [22, 27, 31, 32]. NO3 – + hν → ONOO– ONOO– + H+ HOONO

[Φ9 (254 nm) = 0.1] [pKa 10 ≈ 7]

(9) (10)

The upper wavelength range in which reaction 9 has been studied is relatively near the lower end of solar UVB radiation, but at present very little information is available on the real environmental role of such a process, if it has any. However, since the occurrence of reaction 9 under UVB irradiation cannot be excluded, Mack and Bolton discussed the need to revisit the reaction pathways proposed for the transformation of organic molecules in the presence of irradiated nitrate and nitrite in the light of the ﬁndings concerning the possible formation of HOONO/ONOO– [24]. Since new data concerning the transformation reactions of phenol in the presence of NO3 – + hν, NO2 – + hν and HOONO are now available, the possible role of peroxynitrous acid in that particular case will be discussed later (Sect. 3.2.3).

226

D. Vione et al.

Another pathway that can lead to the formation of peroxynitrous acid is the reaction between · OH and · NO2 , provided that no hydroxyl scavengers are present in solution. Another unstable species, peroxynitrate (O2 NOO– ), can also form upon reaction between O(3 P) and nitrate [22–24, 27, 29, 33, 34]: · OH + · NO2 → HOONO 3

[k11 = 1.3 × 109 M–1 s–1 ] 8

–1 –1

(11)

O( P) + NO3 → O2 O2 NOO– + H+ HOONO2 O2 NOO– → NO2 – + O2

[k12 = 3 × 10 M s ] [pKa 13 = 5.9] [k14 = 1.0 s–1 ]

(12) (13) (14)

HOONO2 → HNO2 + O2 HOONO2 → HO2 · + · NO2

[k15 = 7.0 × 10–4 s–1 ]

(15)

–

NOO–

–3 –1

[k16 = 8.6 × 10 s ]

(16)

2.2 Photolysis of Nitrite and Nitrous Acid 2.2.1 Nitrite Nitrite absorbs radiation in the UV region with a high absorption maximum at 210 nm, a shoulder maximum at 280–290 nm and a second maximum near 355 nm ([24, 35, 36], see also Fig. 1). The absorption band at 355 nm continues well into the visible and accounts for the environmental role of nitrite photochemistry. UV irradiation of nitrite yields · OH + · NO as the main pathway, while a minor pathway of uncertain importance might also yield · NO2 and aquated electrons [22, 28, 35, 37–42]: NO2 – + hν → · NO + O· – O· – + H+ · OH NO2 – + hν → · NO2 + e– aq

[Φ17 (355 nm) = 0.025] [pKa 18 = 11.9] [Φ19 ≤ 0.076Φ17 ]

(17) (18) (19)

The quantum yield of reaction 17 varies with the irradiation wavelength, from about 0.07 near 300 nm to 0.025 at 355 nm down to 0.015 at 371 nm [35, 40]. Differently from nitrate, nitrite is a source but also a sink for the hydroxyl radical, the latter reaction yielding nitrogen dioxide [35, 37–40, 43]: NO2 – + · OH → · NO2 + OH–

[k20 = 1.0 × 1010 M–1 s–1 ]

(20)

The presence in the system of · OH, · NO, · NO2 and possibly of e– aq gives rise to a series of further reactions, the main ones being reported below ([30, 35, 38, 44]; reactions 7 and 8 are to be added to the list, too): · NO + · NO2 → N2 O3 N2 O3 (+ H2 O) → 2NO2

[k21 = 1.1 × 109 M–1 s–1 ] –

+ 2H+

2 –1

[k22 = 5.3 × 10 s ]

(21) (22)

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

O2 + e– aq → O2 – · O2 – · + NO2 – + 2H+ → H2 O2 + · NO2

[k23 = 1.9 × 1010 M–1 s–1 ] 6

–1 –1

[k24 = 5 × 10 M s ]

227

(23) (24)

Obviously, reactions 23, 24 play a role only if the quantum yield of reaction 19 is relatively near its upper limit. The following reactions can also take place [35, 45–51], but kinetic analysis indicates that they are very likely to play a secondary role in the system [44]: · NO + O2 – · → ONOO–

HOONO HOONO → NO3 – + H+ HOONO → · NO2 + · OH · NO2 + · OH → HOONO ONOO–

+ H+

2· NO + O2 → 2· NO2 · NO2 + O2 – · → NO2 – + O2 NO3

–

+ · NO → · NO2 + NO2 –

[k25 = 6.7 × 109 M–1 s–1 ] [pKa 26 = 7] [k27 = 0.7 s–1 ] [k28 ≤ 0.3 s–1 ]

(25) (26) (27) (28)

[k29 = 1.3 × 109 M–1 s–1 ]

(29)

6

–2 –1

[2k30 = 4.2 × 10 M s ] 9

–1 –1

[k31 = 4.5 × 10 M s ] 4

–1 –1

[k32 ≤ 4 × 10 M s ]

(30) (31) (32)

Among the species generated upon nitrite photolysis, · OH, · NO2 , N2 O4 and N2 O3 are the most relevant ones when taking into account the transformation reactions of other molecules that can be present in solution [24]. 2.2.2 Nitrous Acid Nitrous acid also absorbs radiation in the UV range. In particular, its spectrum shows an absorption band around 360 nm, corresponding to the one of nitrite but with higher molar absorptivity. Quite interestingly, the UVA absorption spectrum of aqueous HNO2 shows the vibrational ﬁne structure, which is rather uncommon for dissolved molecules ([35, 36, 52]; see also Fig. 1). The UV irradiation of nitrous acid causes photolysis, generating · OH + · NO with an average quantum yield near 0.4 in the wavelength range 280–385 nm [12, 28, 35]: HNO2 + hν → · OH + · NO

[Φ33 = 0.35–0.45]

(33)

In a similar way to nitrite, although at a somewhat lower rate, nitrous acid can be oxidised by the hydroxyl radical [35, 43]. Furthermore, it also undergoes a thermal decomposition process that parallels the photoinduced reaction [53]. · OH + HNO2 → H2 O + · NO2 2HNO2 → · NO + · NO2 + H2 O

[k34 ≈ 2.6 × 109 M–1 s–1 ] [k35 = 1.3 M–1 s–1 (274 K), 28.6 M–1 s–1 (303 K)]

Reactions 33–35 are then followed by reactions 7, 8 and reactions 21, 22.

(34) (35)

228

D. Vione et al.

3 Transformation Reactions in the Presence of Organic Compounds Based on the direct photolysis data reported in Sect. 2, it is then interesting to consider how the photogenerated species can interact with organic compounds and transform them. Such reactions are important both in natural water chemistry and in the context of advanced oxidation processes for water and wastewater treatment, in the presence of dissolved nitrate and nitrite [24, 54]. This section will thus give an account of the studies carried out on the transformation of organic molecules in the presence of irradiated nitrate and nitrite. Particular attention will be paid to the transformation of phenol, the compound for which the relevant reactions have been most studied. For mechanistic discussion, only those studies will be considered where irradiation has been carried out at environmentally signiﬁcant wavelengths (above 290–300 nm). Before starting with the photoinduced processes, it is, however, necessary to take into account the thermal (dark) reactions that can take place in the presence of nitrate and nitrous acid. 3.1 Dark Processes in Acidic Solutions of Nitrate and Nitrous Acid Starting a description of photoinduced processes with a discussion of dark reactions is motivated by the fact that the acknowledgement of the relative role that photoinduced and dark processes have in different systems has allowed a development in the understanding of the relevant transformation pathways [55–57]. Furthermore, since irradiated systems might show a mixture of photoinduced and thermal processes, it is better to start isolating the thermal ones that occur alone in the dark. It is interesting to observe that thermal processes in the presence of nitrate and nitrite only occur in acidic solutions [55–58], while no such processes have been observed in neutral to basic systems [44, 59]. Furthermore, most dark reactions have been studied in the presence of phenol as substrate, although data on other molecules are also available. Phenol (H – Ph – OH) can be nitrated in the dark in acidic nitrate solutions. However, nitration of phenol by HNO3 occurs at an appreciable rate only for [HNO3 ] > 0.1 M and pH < 1 [58]. Such a process is initiated by the electrophilic agent NO2 + , formed upon protonation of HNO3 [60, 61]: HNO3 + H+ H2 NO3 + → NO2 + + H2 O H – Ph – OH + NO2 + → O2 N – (H)Ph+ – OH O2 N – (H)Ph+ – OH + H2 O O2 N – Ph – OH + H3 O+

(36) (37) (38)

The reaction yields the ortho and para isomers (2-nitrophenol, 4-nitrophenol), with a slight excess of the former.

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

229

In the presence of nitrous acid, phenol undergoes transformation under much milder conditions when compared with HNO3 . In fact, pH 5.5 is sufﬁcient to induce the reaction [55, 62]. The transformation of phenol in the presence of nitrous acid in the dark has been studied by different authors under a wide variety of conditions and focusing attention on different intermediates [62–67]. The main detected transformation intermediates of phenol in the presence of HNO2 are 2- and 4-nitrophenol and 2- and 4-nitrosophenol. The yield in each of these compounds varies, however, according to the adopted conditions. It is therefore not surprising that different authors have indicated different reaction pathways to account for the observed results. Proposed reactive species include NO+ , NO2 + , HNO2 , N2 O3 , · NO2 , and N2 O4 [63–67]. It is, however, interesting to observe that electrophilic reactive species (NO+ , NO2 + ) have been proposed by those authors working under rather extreme conditions (e.g. 10–20% H2 SO4 in water or CF3 COOH as solvent [63–65]). When considering milder conditions of higher environmental relevance, the pH trend of the initial formation rates of the detected compounds (2-nitrophenol, 4-nitrophenol, 4-nitrosophenol) suggests that an acid-base equilibrium is operating in the system. The curves have a ﬂexus at pH ≈ 3.3 [62], which is fully compatible with the pKa of nitrous acid [68]. In the presence of electrophilic reactive species, one would rather expect a steeper pH dependence of the kind {Rate ∝ [H+ ]} [57], considering that for instance NO+ has pKa = – 6.5 [69]. Accordingly, in the presence of phenol and HNO2 at pH around 3, the formation of 2- and 4-nitrophenol and of 4-nitrosophenol is likely to involve neutral species such as HNO2 , · NO2 , N2 O4 and N2 O3 [65–67], the last three formed upon HNO2 thermal decomposition (see reactions 35, 7, 21). The kinetic features of the process seem, however, to favour HNO2 [56, 62]. In the presence of 1.0 × 10–3 M phenol and 7.0 × 10–3 M NaNO2 , at pH 2.0 by addition of HClO4 , the yields are as follows: 2-nitrophenol 22%, 4-nitrophenol 11%, 4-nitrosophenol 40% [62]. Another thermal process that can take place in HNO2 solutions is the reaction between HNO2 and H2 O2 to yield peroxynitrous acid, HOONO [70]. The reaction rate between HNO2 and H2 O2 is proportional to the concentration product of the two reactants and shows a ﬂexus at pH 1.5, compatible with a rate-determining step involving H3 O2 + and HNO2 [57]. Peroxynitrous acid can also form upon reaction between nitric oxide and superoxide [71] and upon nitrate photolysis (see Sect. 2.1). Peroxynitrous acid is an isomer of nitric acid and the isomerisation to HNO3 is its main transformation pathway (see reaction 27). A secondary transformation process, in contrast, involves the generation of · OH + · NO2 (reaction 28). Furthermore, HOONO was postulated to undergo protonation in acidic solution with generation of electrophilic species [72]: HOONO + H+ H2 OONO+

(39)

230

H2 OONO+ NO2 + + H2 O H2 OONO+ NO+ + H2 O2

D. Vione et al.

(40) (41)

Differently from HNO3 , HOONO is a weak acid (see reaction 10). The anion, peroxynitrite, was shown to react with carbon dioxide to yield the nitrating agent ONOOCO2 – [73, 74]. In contrast, peroxynitrous acid is too short-lived to undergo such a reaction [75]. In the presence of organic compounds, peroxynitrous acid induces oxidation and nitration processes [72, 74, 76, 77]. Oxidation reactions are due to the generation of hydroxyl (reaction 28). In the case of phenol, nitration is most likely to be electrophilic as evidenced by the very steep pH trend of nitrophenol initial formation rates ({Rate ∝ [H+ ]} [57]). Dark processes have also been studied in the presence of benzene and naphthalene as substrates [62]. Nitration by HNO3 requires stronger acidic conditions when compared with phenol (HNO3 > 1 M, pH < 0). Furthermore, no transformation of benzene and naphthalene was observed in the presence of HNO2 . This ﬁnding is consistent with the hypothesis that phenol nitration in the presence of HNO2 is initiated by attack on the hydroxyl function [65]. Finally, both benzene and naphthalene can be hydroxylated and nitrated in the presence of HOONO [62]. 3.2 Transformation of Phenol upon Photolysis of Nitrite and Nitrate The case of nitrite will be dealt with before nitrate, since the phenol/nitrite/ UV system was the ﬁrst to be understood in sufﬁcient detail. 3.2.1 Nitrite Photolysis The UV irradiation of phenol in the presence of nitrite (pH around neutrality) yields 4-nitrosophenol, catechol, hydroquinone, 1,4-benzoquinone, 2-nitrophenol and 4-nitrophenol [44, 62, 78, 79]. It is interesting to observe that nitrophenols do not form in appreciable amount upon irradiation of 1.0 × 10–3 M phenol and nitrite, while relevant nitrophenol formation can be observed for [NO2 – ] > 1.0 × 10–3 M [44, 79]. The formation of catechol, hydroquinone and 1,4-benzoquinone is most likely due to the photoinduced generation of hydroxyl (reactions 17 and 18). The relevant processes are initiated by hydroxyl attack on phenol (H – Ph – OH) in the presence of oxygen ([78, 79]; HO – Ph – OH: catechol or hydroquinone, O = Ph = O: 1,4benzoquinone): H – Ph – OH + · OH → HO(H) – Ph· – OH HO(H) – Ph· – OH + O2 → HO – Ph – OH + HO2 ·

(42) (43)

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

HO – Ph – OH + · OH → HO – Ph – O· + H2 O 2HO – Ph – O· → HO – Ph – OH + O = Ph = O HO – Ph – O· + O2 → O = Ph = O + HO2 ·

231

(44) (45a) (45b)

Reaction 45a is very likely to play a more important role than reaction 45b. 1,4-Benzoquinone can also form upon direct photolysis of hydroquinone (λ < 320 nm) [79]. The formation of 4-nitrosophenol most likely takes place upon reaction between phenol and N2 O3 [44, 51, 79]. This is supported by evidence according to which · NO is not a nitrosating agent [61] and by the effect of hydroxyl scavengers. In fact, the formation of 4-nitrosophenol from phenol is strongly inhibited if, under conditions of nitrite irradiation, one adds a hydroxyl scavenger to the system (e.g. formate [79] or 2-propanol [44]). The generation of N2 O3 requires reaction between · NO and · NO2 (reaction 21), while the formation of · NO2 is linked with the oxidation of nitrite by the hydroxyl radical (reaction 20). The addition of hydroxyl scavengers to the system inhibits reaction 20 and, as a consequence, the production of N2 O3 in reaction 21. Further evidence in favour of N2 O3 as the reactive species for phenol photonitrosation comes from the pH trend of the formation rate of 4-nitrosophenol upon nitrite photolysis in basic solution. Such a trend shows the inﬂuence of the basic hydrolysis of N2 O3 [59]. The formation of nitrophenols most likely involves nitrogen dioxide (· NO2 /N2 O4 ), as evidenced by the fact that phenol nitration is inhibited by the addition of hydroxyl scavengers to the system (hydroxyl is necessary for the generation of nitrogen dioxide in reaction 20 [44]). Furthermore, the trend of nitrophenol initial formation rate as a function of nitrite concentration is compatible with nitration by nitrogen dioxide and seems for instance to exclude a nitration pathway involving · OH + · NO2 as in the gas phase [80, 81]. In fact, the · OH + · NO2 nitration pathway would be controlled by the steady-state hydroxyl concentration. In the presence of 1.0 × 10–3 M phenol, both [· OH] and the initial formation rate of nitrophenols would reach a maximum for 0.01 M nitrite [55, 82]. In contrast, the initial nitrophenol formation rate steadily increases in the nitrite concentration range 0.001–0.1 M, consistent with nitration by · NO2 /N2 O4 [44]. Furthermore, kinetic evidence seems to favour N2 O4 over · NO2 as the reactive species for nitration [62]. An additional pathway for nitrophenol formation upon nitrite photolysis is linked with the nitrosation of phenol to yield 4-nitrosophenol, followed by the oxidation of 4-nitrosophenol to 4-nitrophenol [44]. A similar pathway involving phenol, 2-nitrosophenol and 2-nitrophenol might tentatively be postulated [65, 66]. The nitroso derivative pathway is likely to be a secondary one in neutral to acidic solution [44, 55], but it is a major source of 4-nitrophenol upon nitrite photolysis under basic conditions [59]. An interesting aspect of phenol nitration and nitrosation in the presence of nitrite is that these processes are favoured with decreasing pH. The pH trend

232

D. Vione et al.

of the initial formation rates can be accounted for by an acid-base equilibrium with pKa ≈ 3.3 [55], most likely linked with the presence of HNO2 in the systems under study [68]. Nitrous acid is known to undergo photolysis upon UV irradiation, producing · NO, · NO2 , N2 O3 , and N2 O4 in a similar way as nitrite (reactions 7, 21, 33, 34), but with higher quantum yield than NO2 – . However, the generation of nitrophenols and of 4-nitrosophenol in the presence of phenol and nitrous acid is very likely to be a thermal process (see Sect. 3.1). Indeed, if irradiation and dark experiments are carried out under the same temperature and stirring conditions, the initial formation rates of 2nitrophenol, 4-nitrophenol and 4-nitrosophenol are the same [55]. In the case of irradiation, however, the transformation rate of phenol is a bit higher due to the formation of hydroxyl in reaction 33. In neutral solution, in contrast, phenol nitration and nitrosation are photoinduced processes since no thermal reaction has been observed between phenol and nitrite ion. The pH value where the thermal and photoinduced processes have similar importance is around 5.5. Thermal processes prevail at lower pH and photoinduced ones at higher pH [55, 62]. 3.2.2 Nitrite Photooxidation As discussed in Sect. 3.2.1, phenol nitration upon UV irradiation of nitrite in neutral solution is due to the photoinduced generation of · NO2 /N2 O4 . Nitrogen dioxide forms upon oxidation of nitrite by photogenerated hydroxyl (reaction 20). The fact that the photoproduction of nitrogen dioxide upon nitrite photolysis is not direct but involves nitrite both as photoactive species and substrate to be oxidised is an interesting phenomenon. In fact, it opens up the possibility for nitrite to interact with other hydroxyl sources, yielding nitrogen dioxide that can be involved in aromatic nitration [82]. It also suggests the possibility for different environmental factors to interact with one another, in addition to the already acknowledged interaction between environmental factors and organic compounds, resulting in oxidation processes [83, 84]. An enhancement of nitrophenol formation has, for instance, been observed upon irradiation of dissolved Fe(III) and nitrite/nitrous acid, when compared with NO2 – /HNO2 alone [85]. If the solution pH is sufﬁciently acidic to allow the existence of the photoactive Fe(III) species FeOH2+ , UV photolysis yields hydroxyl [86] that can oxidise nitrite to nitrogen dioxide (reaction 20): FeOH2+ + hν → Fe2+ + · OH [Φ46 (313 nm) = 0.14] (46) At higher pH (usually > 3–4) the mononuclear Fe(III) species are less stable and dimers and polymers tend to prevail. The irradiation of the latter yields hydroxyl at a lesser extent, but oxidation via charge-transfer processes

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

233

involving excited species can still take place [85, 87]. [FeIII ] + hν → [FeIII ]∗ [FeIII ]∗ + NO2 – → Fe(II) + · NO2

(47) (48)

Around neutral pH, by far the most common Fe(III) species are the oxides and hydroxides. Of these, α-Fe2O3 and β-FeOOH have been most studied from a photochemical point of view [88–92]. They are semiconductor oxides and irradiation in the visible promotes electrons from the valence band to the conduction one, leaving holes (h+ ) in the valence band [90, 91]. Electrons and holes can either thermally recombine or migrate to the oxide surface, where they can be trapped by surface species and react with dissolved molecules [93]. In aerated solution, trapped electrons are likely to reduce oxygen to superoxide, while trapped holes can oxidise various molecules. Quite interestingly, when irradiated in the visible, α-Fe2 O3 and β-FeOOH are not able to transform phenol. However, in the presence of phenol and nitrite, quantitative yield in nitrophenols is observed as a consequence of the oxidation of nitrite to nitrogen dioxide [85]. α-Fe2 O3 (β-FeOOH) + hν → e– CB + h+ VB e– CB + O2 → O2 – · h+ VB + NO2 – → · NO2

(49) (50) (51)

When the irradiation of α-Fe2 O3 and β-FeOOH is carried out at λ ≤ 360 nm, one observes the excitation of a charge-transfer band, with generation of Fe2+ and · OH from the surface FeIII – OH groups [88, 89]. Obviously, hydroxyl can then oxidise nitrite [57]. The photoproduction of Fe2+ is also interesting because this ion can react with hydrogen peroxide to yield further hydroxyl in the well-known Fenton reaction [94]. In the presence of Fe(III) (hydr)oxides and H2 O2 under UV irradiation (λ ≤ 360 nm) one obtains the so-called heterogeneous photo-Fenton system [95], and also in this case an enhancement of nitrophenol formation is observed [57]. Fe2+ + H2 O2 → Fe3+ + · OH + OH–

(52)

Generation of Fe2+ also takes place upon irradiation of dissolved Fe(III) in acidic solution (photo-Fenton reaction [96]). It is, however, difﬁcult to study phenol nitration in the Fe(III)/H2 O2 /HNO2 /UV system due to the very fast thermal reaction between H2 O2 and HNO2 to yield peroxynitrous acid, HOONO, that also nitrates phenol [57]. Enhancement of nitrophenol formation has also been observed upon irradiation of phenol, nitrite and two Mn(III,IV) (hydr)oxides, β-MnO2 and γ -MnOOH [97], which show some photocatalytic activity [98]. Finally, interaction can take place between nitrate and nitrite (reactions 1, 2, 20), which also results in enhanced nitrophenol formation [85].

234

D. Vione et al.

3.2.3 Nitrate Photolysis The UV irradiation of phenol in the presence of nitrate yields oxidation (hydroquinone, catechol, resorcinol, benzoquinone, hydroxybenzoquinone, trihydroxybenzenes), condensation (phenoxyphenols and dihydroxybiphenyls), nitration (2-nitrophenol, 4-nitrophenol, 4-nitrocatechol, nitrobenzoquinone) and nitrosation products (4-nitrosophenol) [54, 58, 79, 99, 100]. The oxidation pathways can be described in analogy with those already discussed in the case of nitrite photolysis (Sect. 3.2.1, reactions 42–45). The main difference with nitrite is that a far higher number of oxidation intermediates has been observed in the case of nitrate. A possible explanation is that the rate of hydroxyl photogeneration upon nitrate photolysis increases with increasing [NO3 – ]; thus, one can study the system at high nitrate concentration and obtain the primary oxidation intermediates in relatively high amounts, which also allows the identiﬁcation of secondary intermediates. With nitrite one is limited to low concentration values because at high nitrite levels nitration and nitrosation processes prevail over the oxidative ones, NO2 – being a sink for hydroxyl [44, 79, 82]. Reaction between phenol and hydroxyl yields the dihydroxybenzenes, which can then undergo further oxidation (hydroquinone to benzoquinone, further hydroxylated to hydroxybenzoquinone, catechol and resorcinol to trihydroxybenzenes [79, 100]). The condensation products, phenoxyphenols and dihydroxybiphenyls, most likely originate from the reaction between phenol and the phenoxyl radical [101]. Their presence indicates that some phenoxyl forms in the system, due to the reaction of phenol with · OH or · NO2 . The possibility for · NO2 to oxidise phenol to phenoxyl has been the object of a literature debate [102, 103] in the context of nitration processes. The problem can be tackled upon consideration of the reduction potentials of the various species. The reduction potential of phenoxyl to undissociated phenol is E = 1.34 V – 0.059 pH [104], while for the reduction of nitrogen dioxide to nitrite it is E = 0.90 V [105]. Accordingly oxidation of phenol to phenoxyl would be possible above pH 7.5, and of course in the presence of phenolate (pH > 10 [106]). Phenol can be nitrated upon nitrate irradiation, yielding 2- and 4-nitrophenol [54, 58, 79, 99, 100]. The generation of · OH + · NO2 upon nitrate photolysis (reactions 1 and 2) would suggest the possibility that phenol nitration might follow an · OH-mediated pathway as in the gas phase [80, 81]. Furthermore, hydroxyl-mediated nitration in aqueous solution has been described in the case of benzene [107]. However, the addition of hydroxyl scavengers to the system (formate [79], 2-propanol [58]) favours the formation of nitrophenols, while an · OH-mediated nitration would be inhibited by hydroxyl consumption. The positive effect of the scavengers can be accounted for in the hypothesis that phenol nitration takes place upon reaction with nitrogen

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

235

dioxide (· NO2 /N2 O4 ). Nitrate photolysis yields · OH + · NO2 (reactions 1 and 2), and the recombination of the two species to yield back nitrate decreases the steady-state concentration of nitrogen dioxide: · OH + · NO2 → NO3 – + H+ (53) The reaction of the scavengers with hydroxyl would inhibit the recombination and increase the steady-state concentration of nitrogen dioxide, thus enhancing phenol nitration [23, 58, 79]. A controversy has arisen as to the mechanism of phenol nitration by nitrogen dioxide, concerning in particular the possibility for phenoxyl to be an intermediate [102, 103]. The oxidation of phenol to phenoxyl by nitrogen dioxide can be expected to take place in basic solution, and is favoured with increasing pH [104, 105], while the formation of nitrophenols upon nitrate photolysis in basic solution is negligible [58]. Further evidence against nitration involving phenoxyl comes from the pH trend of nitrophenol formation upon nitrite photolysis in basic solution. When considering that nitrogen dioxide is able to rapidly oxidise the phenolate anion to phenoxyl [108], and that the nitration of phenolate is less efﬁcient than the nitration of undissociated phenol [59], one would conclude that existing evidence is hardly consistent with a relevant role played by phenoxyl as an intermediate of phenol nitration. A very interesting aspect of phenol nitration upon nitrate photolysis is the pH trend in the acidic range: many authors agree that the formation of nitrophenols is favoured by decreasing pH [54, 58, 79]. The pH trend of the initial formation rates of 2- and 4-nitrophenol is very similar to the one already observed in the case of nitrite, reﬂecting also in this case an acid-base equilibrium with pKa = 3.3–3.4 [56] (see Fig. 2). This result is compatible with a relevant role played by nitrous acid, and HNO2 actually forms upon irradiation of nitrate in acidic solution (reactions 1–3, 7, 8). The irradiation of nitrous acid in the presence of nitrate in excess is, however, unlikely to give a relevant contribution to the photogeneration of nitrogen dioxide. In contrast, the observed nitrous acid levels can account for nitrophenol formation in the hypothesis of a dark reaction. Obviously, irradiation is necessary for nitrous acid to be formed, but the further process leading to nitrophenols is probably not a photoinduced one [56]. This ﬁnding is in agreement with the results already discussed in the context of nitrite photolysis (Sect. 3.2.1). The pH trend of nitrophenol formation is interesting also because with this datum, it is possible to discuss the role of peroxynitrous acid in a transformation process induced by nitrate photolysis, thus giving a ﬁrst answer to the observation made by Mack and Bolton [24]. HOONO forms upon irradiation of nitrate (reactions 9 and 10), although its formation at λ > 280 nm is uncertain. In the presence of HOONO phenol undergoes various transformation reactions, among which nitration [57]. The pH trend of nitrophenol formation in the presence of HOONO is of the kind Rate ∝ [H+ ], as discussed in

236

D. Vione et al.

Fig. 2 Initial formation rates of 2-nitrophenol and 4-nitrophenol upon nitrate irradiation at 312 nm as a function of pH. Initial conditions: 1.0 × 10–3 M phenol, 0.10 M NaNO3 , photon ﬂux in solution 3.6 × 10–7 Ein s–1 . Ein = Einstein (1 Einstein = 1 mole of photons)

Sect. 3.1, and thus not compatible with the trend observed in acidic nitrate solutions under irradiation (Rate ∝ [H+ ]{Ka + [H+ ]}–1 , with Ka ≈ 10–3.3 [56]). Accordingly, the formation of nitrophenols upon UV irradiation of nitrate in acidic solution is due to the generation of HNO2 and not of HOONO. Apart from nitrophenols, 4-nitrocatechol and nitrobenzoquinone have also been detected as nitro derivatives [54, 79, 100]. They are secondary photoproducts and are thought to originate from the nitration of catechol and hydroquinone, in the latter case followed by the oxidation of nitrohydroquinone [54, 100]. 4-Nitrocatechol might in principle derive from catechol nitration or from 4-nitrophenol hydroxylation. However, the conversion of 4-nitrophenol into 4-nitrocatechol upon nitrate photolysis is rather limited [109] and cannot account for the observed time evolution starting from phenol [54]. Finally, 4-nitrosophenol has also been detected as a transformation intermediate of phenol upon nitrate photolysis [54, 58, 79, 99, 100]. The occurrence of 4-nitrosophenol might be accounted for by the irradiation of photoformed nitrite and by the thermal reactions induced by nitrous acid. Indeed, the formation of 4-nitrosophenol upon nitrate irradiation is favoured at acidic pH [58] and 4-nitrosophenol forms in relevant amount in the presence of both nitrite under irradiation [44, 59, 110] and nitrous acid in the dark [55, 65–67]. The formation of nitrite/nitrous acid upon nitrate UV irradiation has been the object of some attention also because the photochemical treatment of nitrate-containing water might result in nitrite concentration above the permitted levels. However, for a reasonable initial concentration of nitrate, the limit values of nitrite are extremely unlikely to be reached [32, 100, 111, 112].

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

237

3.2.4 Mechanistic Implications Phenol transforms in the presence of nitrate and nitrite under UV irradiation. In all cases of oxidation/hydroxylation, nitrosation and nitration products have been detected. The oxidation and hydroxylation of phenol is initiated by reaction with hydroxyl, while nitrosation is probably due to reaction with N2 O3 . The formation of 2- and 4-nitrophenol from phenol is a consequence of the reaction of the substrate with nitrogen dioxide (· NO2 /N2 O4 ), although the details of the mechanism are still uncertain. It is, however, very doubtful that phenol photonitration involves phenoxyl as an intermediate, since nitra-

Scheme 1 Phenol transformation pathways in the presence of nitrate and nitrite under irradiation. hν = direct photolysis

238

D. Vione et al.

tion under basic conditions (where phenoxyl is most likely to form) is very limited. A nitration pathway involving · OH + · NO2 can also be ruled out. The formation of nitrophenols from phenol is strongly favoured at acidic pH. This was initially attributed to the formation of an electrophilic species from nitrogen dioxide, but the pH trend of nitrophenol initial formation rate indicates that the pH effect is due to the presence of HNO2 , most likely involved in a dark reaction. In the case of nitrate photolysis, nitrous acid is generated in reactions 1–3, 7, 8. The cited pH trend also indicates that nitrophenol formation upon nitrate photolysis in acidic solution is due to the formation of HNO2 and not of HOONO. The transformation pathways of phenol upon photolysis of nitrate and nitrite are reported in Scheme 1. The further transformation of the primary intermediates is discussed in Sect. 3.3. 3.3 Transformation of Substituted Phenols upon Photolysis of Nitrite and Nitrate 3.3.1 Nitrite Photolysis The transformation of various substituted phenols has been studied in the presence of nitrite under irradiation: dihydroxybenzenes [78, 79, 113], nitrophenols [109], phenylphenols [114]. Obviously, the observed transformation intermediates vary according to the reaction rate of each substrate and intermediate with the various reactive species formed during irradiation and according to the absorption spectrum and direct photolysis quantum yield of each compound. The irradiation of nitrite in the presence of resorcinol mainly yields 4nitrosoresorcinol, with a certain amount of 2,4-dinitrosoresorcinol detected after prolonged irradiation and most likely arising from the nitrosation of the former intermediate [113]. The yield in 4-nitrosoresorcinol upon 366 nm irradiation of 5 × 10–4 M resorcinol in aerated solution varied from 0.44 at 5 × 10–4 M nitrite to 0.89 in the presence of nitrite 5 × 10–3 M [113]. At the same time, the quantum yield for resorcinol transformation passed from 1.21 × 10–2 to 1.05 × 10–2 . The overall effect is most likely due to the scavenging of hydroxyl at higher nitrite concentration (reaction 20), which inhibits the oxidative pathways and enhances the relative importance of nitrosation. The ﬁnding also constitutes indirect evidence of a possible involvement of nitrogen dioxide, produced in reaction 20, in the formation of 4-nitrosoresorcinol. The mechanistic studies (effect of hydroxyl scavengers and oxygen, absence of dark reaction with · NO) rule out · NO and · OH + · NO as possible reactive species for nitrosation. In contrast, N2 O3 or · NO2 + · NO, the latter involved in two consecutive steps, can possibly induce nitrosa-

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

239

tion [78, 79, 113]. It is obviously difﬁcult to differentiate the role of N2 O3 and of · NO2 + · NO; however, the fact that resorcinol nitrosation is inhibited in the presence of oxygen favours a two-step pathway. Quite interestingly, nitrosation of resorcinol also takes place in the presence of nitrous acid in the dark; indeed, the thermal decomposition of HNO2 (reaction 35) yields the species required for nitrosation to occur (· NO2 , · NO) [113]. In the presence of catechol the photolysis of nitrite yields hydroxybenzoquinone and 4-nitrocatechol [78, 79]. The yield of hydroxybenzoquinone decreases and that of 4-nitrocatechol increases with increasing [NO2 – ], comitantly with the fact that nitrite is a sink as well as a source of hydroxyl. In fact, with increasing nitrite concentration the generation of nitrogen dioxide prevails over that of hydroxyl, consumed in reaction 20 to yield · NO2 . The latter is probably involved in the formation of 4-nitrocatechol [55, 79, 82]. The involvement of nitrogen dioxide in the nitration of catechol is conﬁrmed by the fact that hydroxyl scavengers, inhibiting the formation of · NO2 in reaction 20, also inhibit the formation of 4-nitrocatechol [79]. A similar effect has also been observed in the case of phenol (Sect. 3.2.1). Hydroxybenzoquinone might derive from catechol hydroxylation followed by further oxidation initiated by hydrogen abstraction, in a similar way as shown in reactions 44 and 45. Differently from the case of resorcinol, no nitrosation derivatives were observed in the presence of catechol [78, 79]. Hydroquinone transforms in the presence of irradiated nitrite to yield benzoquinone and hydroxybenzoquinone [78, 79]. At the irradiation wavelength adopted in the cited works (365 nm), hydroquinone direct photolysis should be limited and benzoquinone most likely forms upon reaction between hydroquinone and hydroxyl (reactions 44 and 45; hydroquinone absorbs radiation at λ < 320 nm). Hydroxybenzoquinone is likely to be a product of benzoquinone photolysis. No nitration or nitrosation intermediates of hydroquinone were observed in the presence of nitrite under irradiation, differently from the cases of resorcinol and catechol [78, 79]. The reaction between hydroquinone and nitrogen dioxide is, however, quite rapid [106, 115], as conﬁrmed by the marked inhibition of phenol nitration upon nitrite photolysis by added hydroquinone [62]. The point is that the reaction between hydroquinone and · NO2 mainly yields benzoquinone [62]. Another interesting feature in the case of hydroquinone is the formation of the fairly stable semiquinone radical anion upon reaction between benzoquinone and deprotonated hydroquinone. The spectrum of the resulting solution shows the typical absorption bands of the semiquinone at 308, 315, 403, and 430 nm [79]. The irradiation of nitrophenols in the presence of nitrite and nitrous acid mainly results in hydroxylation processes upon reaction with · OH, leading to the corresponding dihydroxynitrobenzenes [109]. The absorption spectra of nitrophenols, nitrite and nitrous acid overlap, thus irradiation also results in the direct excitation of nitrophenols, with possible photolysis. Accordingly, it is very likely that many of the minor intermediates detected in the cited sys-

240

D. Vione et al.

tems (catechol and 2-nitrosophenol from 2-nitrophenol; hydroquinone, benzoquinone, hydroxybenzoquinone and 4-nitrosophenol from 4-nitrophenol) originate from direct photolysis rather than reaction with the species generated upon nitrite photoexcitation [109]. Indeed, these compounds have been observed as products of the direct photolysis of nitrophenols [116–118]. In the case of 3-nitrophenol, it was not possible to detect intermediates that might arise from direct photolysis [109]. Direct irradiation of 4-nitrophenol has also been carried out in ice pellets [119, 120] and practically yielded the same products as the direct photolysis in solution: hydroquinone, benzoquinone and 4-nitrosophenol. The transformation of 4-nitrophenol in ice is of some interest due to the detection of this compound in snow samples from Antarctica [121]. Interestingly, no nitration of nitrophenols could be observed upon nitrite photolysis: the electron-withdrawing effect of the nitro group evidently inhibits reaction between nitrophenols and nitrogen dioxide [109]. Finally, the irradiation under sunlight of nitrite and phenylphenols (ortho and para isomers) yielded the corresponding nitro derivatives (2-hydroxy3-nitrobiphenyl and 4-hydroxy-3-nitrobiphenyl) in detectable amounts. Furthermore, the irradiated solutions were extracted in ether and the extract proved to be highly mutagenic in the Salmonella typhimurium assay. The direct mutagenicity of the extract and the proportionality between nitro derivative concentration and mutagenic effects points to the nitrophenylphenols as the main mutagenic components [114]. 3.3.2 Nitrate Photolysis Various studies focused on the transformation of phenolic compounds in the presence of nitrate under irradiation. Studied substrates were dihydroxybenzenes [78, 79, 122], chlorophenols [123], nitrophenols [109] and phenylphenols [114, 124]. The irradiation of nitrate in the presence of resorcinol yielded trihydroxybenzenes, 4-nitroresorcinol, 4-nitrosoresorcinol and 2,4-dinitrosoresorcinol. Hydroxylation is due to the reaction of the substrate with hydroxyl, photogenerated in reactions 1 and 2. The formation of 4-nitroresorcinol is connected with the photogeneration of nitrogen dioxide (· NO2 /N2 O4 ) [122]. The formation of the nitroso and dinitroso derivative was also observed upon nitrite photolysis [113]. One could therefore postulate that, in the presence of nitrate under irradiation, the nitroso compounds form as a consequence of nitrite photogeneration, followed by nitrite photolysis. However, the role of photogenerated nitrite under conditions of nitrate photolysis is not yet completely clear [122]. Both nitration and nitrosation of resorcinol upon nitrate irradiation are favoured with decreasing pH. A similar pH effect has also been observed in the case of phenol (see Sect. 3.2.3), and it would be very interesting to plot the initial formation rates as a function of pH to check

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

241

a possible role of photoformed nitrous acid in the phenomenon. If the enhancement of resorcinol nitration and nitrosation was due to nitrous acid, one would observe a curve reﬂecting an acid-base equilibrium with pKa ≈ 3.3 (Rate ∝ [H+ ]{Ka + [H+ ]}–1 , [56]). In the presence of nitrate under irradiation, catechol yields 4-nitrocatechol and hydroxybenzoquinone. The effect of hydroxyl scavengers indicates that the formation of 4-nitrocatechol is due to nitrogen dioxide, whereas hydroxybenzoquinone forms upon reaction of catechol with hydroxyl [78, 79]. The irradiation of hydroquinone/nitrate mixtures yields benzoquinone and hydroxybenzoquinone plus a very small amount of nitrohydroquinone. The interpretation of the results is, however, complicated by the fact that the absorption spectra of hydroquinone and nitrate overlap. The oxidation intermediates mainly originate upon direct photolysis of hydroquinone [79], but the minor formation of nitrohydroquinone requires the contribution of nitrate photolysis [78]. The transformation of chlorophenols upon irradiation of nitrate mainly yields hydroxylated derivatives, while no nitration was observed. The absence of nitro derivatives can be attributed to the electron-withdrawing character of the chlorine atoms, inhibiting nitration. Dihydroxybenzenes and benzoquinone were also observed implying that a dechlorination process, induced by either hydroxyl or light, takes place in the system [123]. Absence of nitro derivatives was also observed upon irradiation of nitrophenols and nitrate. Also in this case, the electron-withdrawing character of the nitro group can account for the inhibition of nitration [109]. The difﬁculty to nitrate nitrophenols to dinitrophenols is widely recognised [125] and also constitutes a problem in environmental chemistry, since ﬁeld data seem, in contrast, to indicate that the nitration of 2-nitrophenol to 2,4-dinitrophenol in the atmospheric aqueous phase (e.g. cloud water) is an important process [126]. In fact, aqueous-phase nitration might be a relevant sink for 2-nitrophenol and possibly the main source of the dinitro compound, which is a powerful phytotoxic agent [127, 128]. In the presence of nitrate under irradiation the main transformation intermediates of nitrophenols are the hydroxyl derivatives, while other compounds may derive from the direct photolysis of the substrates (catechol and 2-nitrosophenol from 2-nitrophenol; hydroquinone, benzoquinone, hydroxybenzoquinone and 4-nitrosophenol from 4-nitrophenol) [109]. The irradiation of nitrate in the presence of 2-phenylphenol yields oxidation (phenylhydroquinone, phenylbenzoquinone, 2-hydroxydibenzofuran) and nitration intermediates (2-hydroxy-3-nitrobiphenyl and 2-hydroxy-5nitrobiphenyl) [124]. 2-Hydroxydibenzofuran is formed by photocyclisation of phenylbenzoquinone. Nitration might be due to reaction with nitrogen dioxide, but an interesting aspect is the effect of pH. The fact that nitration is favoured at low pH might indicate a role of photoformed nitrous acid, due to either photochemical or dark reactivity [124]. It is worth noting that the

242

D. Vione et al.

contribution of nitrous acid photolysis to the generation of nitrogen dioxide in the presence of 0.10 M nitrate (concentration adopted in [124]) is rather limited [56]; thus, nitrous acid might operate in a dark reaction. In fact, dark nitration by HNO2 has been observed in the case of a similar compound, 4-phenoxyphenol, and the species responsible for the primary attack is likely to be HNO2 [129]. In such a case, the nitration of 2-phenylphenol upon nitrate

Scheme 2 Transformation pathways of dihydroxybenzenes in the presence of nitrate and nitrite under irradiation. hν = direct photolysis

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

243

photolysis would show a close similarity to the one of phenol (Sect. 3.2.3). It would result from a process involving nitrogen dioxide in neutral solution combined with an enhancement of nitration at acidic pH, the latter due to photoinduced generation of nitrous acid followed by dark reaction of nitrous acid with the substrate. The irradiation of phenylphenols and nitrate results in the production of mutagenic compounds, most likely the nitro derivatives [114].

Scheme 3 Transformation pathways of other substituted phenols in the presence of nitrate and nitrite under irradiation. hν = direct photolysis, Ph = C6 H5 (phenyl)

244

D. Vione et al.

The transformation pathways of substituted phenols in the presence of nitrate and nitrite under irradiation are summarised in Schemes 2 and 3. 3.4 Transformation of Other Organic Compounds upon Photolysis of Nitrite and Nitrate The transformation of biphenyl upon irradiation of nitrate and nitrite has been the object of many studies due to the formation of mutagenic hydroxynitrobiphenyls in the reaction [130, 131]. In the presence of irradiated nitrate, the main transformation products of biphenyl are hydroxy- and hydroxynitrobiphenyls together with derivatives of further hydroxylation and nitration processes [130]. In the presence of nitrite under irradiation, nitrobiphenyls have also been detected [132]. The formation of mutagenic hydroxynitrobiphenyls is most likely due to the hydroxylation of biphenyl, followed by the nitration of the hydroxyl derivative [131, 133]. Biphenyl hydroxylation can take place upon reaction of the substrate with hydroxyl, but also upon direct photoexcitation [131]. In the latter case the key reaction is likely to take place between radiation-excited biphenyl and nitrate ion [133]. The light excitation of aromatic hydrocarbons to yield hydroxy, hydroxynitro and nitro derivatives in the presence of nitrogen dioxide has also been described in the cases of naphthalene and phenanthrene [134, 135]. Formation of mutagenic 1-nitropyrene has been observed upon irradiation of nitrite and pyrene in solution. Due to poor water solubility, pyrene was adsorbed on silica gel particles and the particle suspension irradiated. Mutagenicity was higher when carrying out irradiation at lower pH [136]. Quite interestingly, 1-nitropyrene also formed in the dark in a gas-solid system in the presence of nitrogen dioxide and pyrene adsorbed on silica [137]. Nitration of pyrene adsorbed on different metal oxides was also observed in the presence of nitrogen dioxide added by different means (directly, present in polluted indoor air or produced upon irradiation of solid nitrate and nitrite). All the systems were exposed to xenon lamp irradiation, simulating sunlight [138]. All the possible nitropyrene isomers were detected (1-, 2- and 4-nitropyrene), and the isomer distribution did not depend on the way nitrogen dioxide was introduced in the system. In contrast, it strongly depended on the metal oxide on which pyrene was adsorbed. In particular, with basic substrates such as MgO and Al2 O3 all the three isomers were detected, while in the case of SiO2 and TiO2 1-nitropyrene was the prevailing isomer or was exclusively present [138]. The prevalent formation of 1-nitropyrene on silica is consistent with the data presented before [136, 137]. In the case of aromatic hydrocarbons such as benzene and naphthalene nitration has been observed upon irradiation of nitrite (and to a lesser extent upon irradiation of nitrate). Differently from phenol and phenol derivatives, benzene and naphthalene do not react at an appreciable rate with HNO2

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

245

in the dark. Nevertheless, nitration upon irradiation of nitrite is favoured at acidic pH and the pH trend of the initial formation rate suggests an involvement of HNO2 [62]. This effect is possibly due to the higher molar absorbivity and photolysis quantum yield of nitrous acid when compared with nitrite [35, 36]. The transformation of azaarenes in the presence of nitrate and nitrite under irradiation yields a large variety of products, among which hydroxyl derivatives, quinones, ring-opening products (often dialdehydes and dicarboxylic acids), ring-rearrangement products and nitro derivatives. Apart from nitro derivatives, the formation of all the other compounds is likely to be initiated by · OH radical attack, followed by reactions involving radicals and oxygen or rearrangement of radical species (leading for instance to the forma-

Table 1 Main reactions induced on aromatic compounds by excitation of nitrate ions in aqueous solution Substrate

Concentration (mmol l–1 ) Oxidation Nitration NitrosationRef. Substrate

Phenol

0.5–1 0.13–1.6 1 1.1 Resorcinol 0.5 Catechol 0.5 Hydroquinone 0.5 Biphenyl 0.6 b 4 0.3 b 0.3 b Hydroxybiphenyls Formed in situ 0.3 b 0.025–0.25 Benzene 3 Naphthalene 0.1 Hydroxybenzoates 0.05 Nitrophenols 1–2 Chlorophenols 1 Monolinuron 0.0005–0.05

Nitrate 1.6–10 1–8 10–100 100 50 50 50 118 10–100 0.3–1.2 1.2 10–100 1.2 100 100 100 50–200 20–40 10–50 0.1–5

× × × × × × × × × × × × × × × × × ×

× × × × × × ×m × × × × × × × × × ×

× × ×

[79, 99] [100] [54] [58] [79, 113] [78, 79] [78, 79] a [130] [133] m [17] [131] [133] m [114] [124] [145] [62] [143] [109] [123] [144]

×m minor reaction a The oxidation reaction may result from the excitation of nitrate or hydroquinone, since both species absorb in the same range b Coated on suspended silica m In aqueous methanol

246

D. Vione et al.

Table 2 Main reactions induced on organic compounds by excitation of nitrite ions in aqueous solution Substrate

Concentration (mmol l–1 ) Oxidation Nitration Nitrosation Ref. Substrate

Phenol

Resorcinol Catechol Hydroquinone Biphenyl Hydroxybiphenyls Benzene Naphthalene 4-Hydroxybenzoate Pyrene Bromoxynil Nitrophenols Dimethylamine Monolinuron b

Nitrate

0.5 1 1.1 1–12 1.1 1–100 0.5 1 0.5 1–5 0.4 10 1.2 0.3 b 1.2 0.3 b 1–5 1–100 0.03–0.1 10–100 Not reported 0–1.2 0.25 b 0.0078 0–25 1–2 1–2 5 3–50 0.0005–0.05 0.01–1

× ×

×

× × ×

×

× × × × ×

× × × ×

× × × × × × × ×

× ×

[79] [110] [44, 55] [113] [78, 79] [78, 79] [131] [114] [145] [62] [143] [136] [140] [109] [141] [144]

Coated on suspended silica

tion of 5-atom rings in place of the initial 6-atom ones). Nitration of azaarenes was attributed to reaction with nitrogen dioxide (· NO2 or N2 O4 ) [139]. The transformation of the herbicide bromoxynil (3,5-dibromo-4-hydroxybenzonitrile) was studied upon direct photolysis in water and in the presence of added nitrite. The addition of nitrite reduces the transformation rate of the substrate, most likely due to competition for absorption of radiation, but induces the formation of nitro derivatives (3-bromo-4-hydroxy5-nitrobenzonitrile, 4-hydroxy-3-nitrobenzonitrile) [140]. The irradiation of dimethylamine and nitrite was shown to yield the wellknown carcinogenic species nitrosodimethylamine, indicating that carcinogenic compounds may form upon photolysis of nitrite in natural waters [141]. Probably nitrosating agents for amines in irradiated nitrite solution are · NO2 and N2 O3 [142]. Last but not least, the transformation of 4-hydroxybenzoate in the presence of nitrate under irradiation is quite interesting from a mechanistic point of view. The substrate undergoes nitration to 4-hydroxy-3-nitrobenzoate, and nitration is inhibited by addition of hydroxyl scavengers [143]. This marks a difference from all the other cases reported so far. Whenever the effect of · OH scavengers on nitration upon nitrate photolysis has been studied, the typical result was an enhancement effect consistent with nitration by

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

247

nitrogen dioxide (· NO2 /N2 O4 ). In the case of 4-hydroxybenzoate the observed inhibition is partially consistent with a nitration process involving · OH + · NO2 . Such a process would be reasonably depressed upon addition of compounds consuming hydroxyl. It is also noticeable that the nitration of 4hydroxybenzoate has been studied at pH 11 [143] and that under basic pH conditions the nitration of many substrates (among which phenol [54, 58]) did not proceed at a detectable rate. Comparison of 4-hydroxybenzoate with other compounds is thus not easy. An overview of the transformation pathways of various substrates in the presence of nitrate and nitrite under irradiation is reported in Tables 1 and 2, respectively. Quite interestingly, nitrosation intermediates have been detected only in the presence of phenolic compounds, dimethylamine and the herbicide monolinuron [144].

4 Assessment of Oxidation, Nitration and Nitrosation under Environmental Conditions It is well established that nitrate and nitrite ions are a source of hydroxyl radicals in natural waters and atmospheric aerosols and can induce the photochemical oxidation of most organic substrates [6, 8, 12, 14]. Hydroxyl generation upon photolysis of nitrate is proportional to the intensity of radiation absorption, while in the case of nitrite the hydroxyl-scavenging role is to be taken into account, too. Actually, oxidation reactions upon nitrite irradiation become disfavoured compared to nitration and nitrosation when increasing nitrite concentration. The role of nitrite as hydroxyl scavenger also implies that NO2 – can interact with other environmental factors (e.g. dissolved Fe(III), nitrate, semiconductor oxides, Fenton systems) with generation of nitrogen dioxide. This interaction can relevantly modify the resulting photoreactions when compared with the photolysis of nitrite alone. In the case of phenol it was observed that the interaction between nitrite and other environmental factors can lead to relevant formation of nitro derivatives at nitrite concentration levels where direct NO2 – photolysis only gives oxidation reactions [57, 85, 97]. Hydroxyl concentration in natural waters can be depressed by the scavenging effects of natural organic matter, bicarbonate ions and nitrite [8, 26, 83, 84]. Accordingly, both the photo-anti-polluting effect of nitrate and the oxidation of nitrite to nitrogen dioxide can be expected to be more important in waters with low concentration of hydroxyl scavengers. Furthermore, the concentration of nitrite ions as hydroxyl scavengers in natural aquifers can play an important role. In fact, relatively high NO2 – levels can interfere with the photo-oxidation processes induced by nitrate photolysis, at the same

248

D. Vione et al.

time producing relevant amounts of nitrogen dioxide that can be involved in nitration reactions ([83–85]; see Sect. 3.2.2). Nitration and nitrosation of aromatic compounds are likely to be rather complex processes involving different pathways. Involved reactive species are · NO2 , N2 O4 , N2 O3 , · NO2 + · NO and HNO2 . Aromatic nitration via such pathways is inhibited by electron-withdrawing substituents on the ring, such as Cl and the nitro group. As to the relevance of hydroxyl-mediated processes (nitration by · OH + · NO2 , nitrosation by · OH + · NO), it can be expected to be quite low for the majority of the compounds. The role of aromatic photonitration and photonitrosation processes in natural waters is still not clearly understood. It is, however, very interesting to report that nitration and nitrosation of monolinuron have been observed upon photolysis of nitrate and nitrite at environmental concentration levels [144]. Studies on the environmental importance of such reactions will indeed beneﬁt from the use of HPLC-MS-MS techniques for both process studies and ﬁeld analysis, as they allow much lower detection limits to be reached when compared with traditional HPLC techniques. In the case of process studies, it would be very interesting to determine the extent at which the pathways observed for the photonitration and photonitrosation of aromatic compounds at fairly high concentration levels can be extrapolated to environmental conditions. At the present state of knowledge, it is possible to postulate that nitration and nitrosation can be minor reactions in aerated aquatic media at low concentration values of substrate and NO3 – /NO2 – . In contrast, they might occur in waters containing relatively high concentration of nitrate (> 50 mg l–1 ) and nitrite (> 3 mg l–1 ). The presence of dissolved organic matter in water can be expected to have limited or even positive inﬂuence on photonitration upon nitrate photolysis, unless the organic matter can compete for nitration. In contrast, hydroxyl scavenging might inhibit nitration processes induced by nitrite photolysis or photooxidation. However, in a recent study on the photonitration of benzene in the presence of irradiated nitrite, no inhibition of nitration upon addition of hydroxyl scavengers has been observed, possibly because benzene nitration under the reported conditions is not initiated by · NO2 /N2 O4 [145]. This ﬁnding might indicate that some aromatic substrates behave differently from phenolic compounds under conditions of nitrite irradiation, and that in such a case, the possible effect of the scavengers needs reconsideration. In this context, it should also be mentioned that the possible impact of nitrate and nitrite photolysis on the transformation of natural organic matter in aqueous solution is a largely unexplored ﬁeld. Another interesting aspect to be taken into account and that will require further investigation is the adsorption of organic compounds at the air/water interface, coupled with the fact that the adsorption of organic compounds also favours the accumulation of inorganic species at the water surface [146].

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

249

When considering that the water surface layer is also intensely illuminated by sunlight, it can be postulated that such an interface can represent an efﬁcient photoreactor. Here, relatively concentrated compounds can take part to processes that can be rather different from the ones occurring at the concentration levels typical of the water bulk. Nitrate and nitrite photochemistry might also play a role in atmospheric hydrometeors. Nitrite photolysis has been shown to account for the majority of hydroxyl photoformation in irradiated fog water from a polluted site [14]. In addition, the generation of mutagenic and carcinogenic compounds from amino acids and amines dissolved in fog water [147] is a process that can be linked with nitrite photochemistry [20, 141]. Furthermore, the formation of atmospheric nitrophenols partially takes place in aqueous solution. Reactions in the aqueous phase can account for about 30% of the atmospheric sources of mononitrophenols and for the vast majority of the dinitrophenol ones [148], and irradiation of nitrate and nitrite can possibly play a role in the process (see Sect. 3.2). Mono- and dinitrophenols are toxic compounds, and their occurrence in rainwater is thought to be a contributory factor in forest decline [149–151].

5 Conclusion The photolysis of nitrate and nitrite in the presence of aromatic compounds yields oxidation, nitration and nitrosation intermediates. Oxidation is due to hydroxyl photogeneration, nitration is usually linked to the presence of nitrogen dioxide (· NO2 /N2 O4 ) and nitrosation with N2 O3 or · NO2 + · NO. Reaction of aromatic compounds with hydroxyl usually has an anti-polluting effect, while the formation of nitro- and nitroso derivatives is a cause of concern due to the mutagenicity of many of these compounds. Nitration and nitrosation processes are often enhanced at acidic pH, and the effect is often due to the presence or photochemical formation of HNO2 in acidic solution. Nitration and nitrosation by HNO2 may be due to dark or photolytic processes, their relative role depending on the substrate. The photolysis of nitrate in surface waters and of nitrite in atmospheric hydrometeors is a relevant source of hydroxyl, while the assessment of the environmental importance of nitration and nitrosation processes needs further studies. Recent results, however, indicate that photonitration and photonitrosation might play a more important role in the environment than previously suspected [144]. All such processes are inﬂuenced by substrate and ion concentration and in some cases by the amount of dissolved organic matter. Nitration processes are also of concern in the context of water treatment techniques. Indeed, relevant formation of nitro derivatives can take place under irradiation in the presence of high levels of nitrate and nitrite [44, 54,

250

D. Vione et al.

55, 58, 85]. Furthermore, dark processes leading to nitro derivatives can take place when nitrite-containing water is acidiﬁed and/or added with oxidants in general and hydrogen peroxide in particular (in the latter case HOONO would form in acidic solution) [55, 57, 97, 152].

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

Zepp RG, Schlotzhauer PF, Sink RM (1985) Environ Sci Technol 19:74 Zepp RG, Wolfe NL, Baughman GL, Hollis RC (1977) Nature 267:421 Haag WR, Hoigné J, Gassmann E, Braun AM (1984) Chemosphere 13:631 Zaﬁriou OC (1974) J Geophys Res 79:4491 Goeyens L, Sörensson F, Tréguer P, Morvan J, Panouse M, Dehairs F (1991) Mar Ecol Prog Ser 77:7 Zaﬁriou OC, True MB (1979) Marine Chem 8:33 Keeney D (1986) CRC Reviews 16:257 Russi H, Kotzias D, Korte F (1982) Chemosphere 11:1041 Warneck P (1999) Phys Chem Chem Phys 1:5471 Vogt R, Finlayson-Pitts BJ (1995) J Phys Chem 99:17269 Petriconi GL, Papee HM (1972) Water Air Soil Pollut 1:117 Zaﬁriou OC, True MB (1979) Marine Chem 8:9 Haag WR, Hoigné J (1985) Chemosphere 14:1659 Anastasio C, McGregor KG (2001) Atmos Environ 35:1079 Harrison CC, Malati MA, Smetham NB (1995) J Photochem Photobiol A Chem 89:215 Kotzias D, Parlar H, Korte F (1982) Naturwissenschaften 69:444 Suzuki J, Okazaki H, Sato T, Suzuki S (1982) Chemosphere 11:437 Suzuki J, Okazaki H, Nishi Y, Suzuki S (1982) Bull Environ Contam Toxicol 29:511 Suzuki J, Hagino T, Ueki T, Nishi Y, Suzuki S (1983) Bull Environ Contam Toxicol 31:79 Suzuki J, Ueki T, Shimizu S, Uesugi K, Suzuki S (1985) Chemosphere 14:493 Meyerstein D, Treinin A (1961) Trans Faraday Soc 57:2104 Wagner I, Strehlow H, Busse GZ (1980) Z Phys Chem N F (Frankfurt) 123:1 Warneck P, Wurzinger C (1988) J Phys Chem 92:6278 Mack J, Bolton JR (1999) J Photochem Photobiol A Chem 128:1 Frank R, Klöpffer, W (1988) Chemosphere 17:985 Zepp RG, Hoigné J, Bader H (1987) Environ Sci Technol 21:443 Daniels M, Meyers RV, Belardo EV (1968) J Phys Chem 72:389 Zellner R, Exner M, Herrmann H (1990) J Atmos Chem 10:411 Løgager T, Sehested K (1993) J Phys Chem 97:10047 von Grätzel M, Henglein A, Lilie J, Beck G (1969) Ber Bunsenges Physik Chem 73:646 Løgager T, Sehested K (1993) J Phys Chem 97:6664 Mark G, Korth H-G, Schuchmann H-P, von Sonntag C (1996) J Photochem Photobiol A Chem 101:89 Gonzales MC, Braun AM (1995) Res Chem Intermed 21:837 Gonzales MC, Braun AM (1996) J Photochem Photobiol A Chem 93:7 Fischer M, Warneck P (1996) J Phys Chem 100:18749 Arakaki T, Miyake T, Hirakawa T, Sakugawa H (1999) Environ Sci Technol 33:2561 Treinin A, Hayon E (1970) J Am Chem Soc 92:5821

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81.

251

von Grätzel M, Taniguchi S, Henglein A (1970) Ber Bunsenges Physik Chem 74:488 Strehlow H, Wagner I (1982) Z Phys Chem N F (Frankfurt) 132:151 Zaﬁriou OC, Bonneau R (1987) Photochem Photobiol 45:723 Bilski P, Szychlinski J, Oleksy E (1988) J Photochem Photobiol A Chem 44:269 Bilski P, Chignell CF, Szychlinski J, Borkowski A, Oleksy E, Reszka K (1992) J Am Chem Soc 114:549 Buxton GV, Greenstock CL, Helman WP, Ross AB (1988) J Phys Chem Ref Data 17:513 Vione D, Maurino V, Minero C, Pelizzetti E (2001) Chemosphere 45:893 Hughes MN, Nicklin HG (1968) J Chem Soc (A):450 Keith WG, Powell RE (1969) J Chem Soc (A):90 Goldstein S, Czapski G (1996) Inorg Chem 35:5935 Coddington JW, Hurst JK, Lymar SV (1999) J Am Chem Soc 121:2438 Goldstein S, Meyerstein D, Eldik van R, Czapski G (1999) J Phys Chem A 103:6587 Gerasimov OV, Lymar SV (1999) Inorg Chem 38:4317 Pires M, Rossi MJ, Ross DS (1994) Int J Chem Kinet 26:1207 Graedel TE, Weschler CJ (1981) Rev Geophys Space Phys 19:505 Park J-Y, Lee Y-N (1988) J Phys Chem 92:6294 Dzengel J, Theurich J, Bahnemann DW (1999) Environ Sci Technol 33:294 Vione D, Maurino V, Minero C, Pelizzetti E (2001) Chemosphere 45:903 Vione D, Maurino V, Minero C, Pelizzetti E (2002) Ann Chim (Rome) 92:919 Vione D, Maurino V, Minero C, Borghesi D, Lucchiari M, Pelizzetti E (2003) Environ Sci Technol 37:4635 Vione D, Maurino V, Minero C, Vincenti M, Pelizzetti E (2001) Chemosphere 44:237 Vione D, Maurino V, Pelizzetti E, Minero C (in press) Int J Environ Anal Chem March J (1992) Advanced Organic Chemistry. Wiley, New York Ridd JH (1998) Acta Chem Scand 52:11 Vione D (2001) PhD thesis, University of Torino, Italy Challis BC, Lawson AJ (1971) J Chem Soc 1971B:770 Uemura S, Toshimitsu A, Okano M (1978) J Chem Soc Perkin Trans I:1076 Al-Obaidi U, Moodie RB (1985) J Chem Soc Perkin Trans II:467 De la Bretèche M-L, Billion M-A, Ducrocq C (1996) Bull Soc Chim France 133:973 Lehnig M (1999) Tetrahedron Lett 40:2299 Martell AE, Smith RM (1974) Critical Stability Constants. Plenum Press, New York Bell KE, Kelly HC (1996) Inorg Chem 35:7225 Halfpenny E, Robinson PL (1952) J Chem Soc:928 Uppu RM, Lemercier JN, Squadrito GL, Zhang HW, Bolzan RM, Pryor WA (1998) Arch Biochem Biophys 358:1 Nonoyama N, Chiba K, Hisatome K, Suzuki H, Shintani F (1999) Tetrahedron Lett 40:6933 Lemercier J-N, Padmaja S, Cueto R, Squadrito GL, Uppu RM, Pryor WA (1997) Arch Biochem Biophys 345:160 Tibi S, Koppenol WH (2000) Helv Chim Acta 83:2412 Vayssie S, Elias H (1998) Angew Chem Int Ed 37:2088 Halfpenny E, Robinson PL (1952) J Chem Soc:939 Bach RD, Glukhovtsev MN, Canepa C (1998) J Am Chem Soc 120:775 Machado F, Boule P (1994) Trends Photochem Photobiol 3:29 Machado F, Boule P (1995) J Photochem Photobiol A Chem 86:73 Atkinson R, Aschmann SM, Arey J (1992) Environ Sci Technol 26:1397 Olariu RI, Klotz B, Barnes I, Becker KH, Mocanu R (2002) Atmos Environ 36:3685

252

D. Vione et al.

82. Vione D, Maurino V, Minero C, Vincenti M, Pelizzetti E (2003) Environ Sci Poll Res 10:321 83. Vaughan PP, Blough NV (1998) Environ Sci Technol 32:2947 84. Brezonik PL, Fulkerson-Brekken J (1998) Environ Sci Technol 32:3004 85. Vione D, Maurino V, Minero C, Pelizzetti E (2002) Environ Sci Technol 36:669 86. Benkelberg H-J, Warneck P (1995) J Phys Chem 99:5214 87. Mazellier P, Mailhot G, Bolte M (1997) New J Chem 21:389 88. Waite TD, Morel FMM (1984) Environ Sci Technol 18:860 89. Faust BC, Hoffmann MR (1986) Environ Sci Technol 20:943 90. Faust BC, Hoffmann MR, Bahnemann DW (1989) J Phys Chem 93:6371 91. Leland JK, Bard AJ (1987) J Phys Chem 91:5076 92. Siffert C, Sulzberger B (1991) Langmuir 7:1627 93. Serpone N, Pelizzetti E (1989) Photocatalysis. Fundamentals and Applications. Wiley, New York 94. Walling C (1975) Acc Chem Res 8:125 95. Mazellier P, Sulzberger B (2001) Environ Sci Technol 35:3314 96. Pignatello JJ, Liu D, Huston P (1999) Environ Sci Technol 33:1832 97. Vione D, Maurino V, Minero C, Pelizzetti E (2004) Chemosphere 55:941 98. Xyla AG, Sulzberger B, Luther III GW, Hering JG, Van Cappellen P, Stumm W (1992) Langmuir 8:95 99. Niessen R, Lenoir D, Boule P (1988) Chemosphere 17:1977 100. Guillaume D, Morvan J, Martin G (1989) Environ Technol Lett 10:491 101. Platz J, Nielsen OJ, Wallington TJ, Ball JC, Hurley MD, Straccia AM, Schneider WF, Sehested J (1998) J Phys Chem A 102:7964 102. Louw R, Santoro D (1999) Environ Sci Technol 33:3281 103. Dzengel J, Theurich J, Bahnemann DW (1999) Environ Sci Technol 33:3282 104. Surdhar PS, Armstrong DA (1987) J Phys Chem 91:6532 105. Bard AJ, Parsons R, Jordan J (1985) Standard Potentials in Aqueous Solution. Dekker, New York 106. Huie RE, Neta P (1986) J Phys Chem 90:1193 107. Eberhardt MK (1975) J Phys Chem 79:1067 108. Alfassi ZB, Huie RE, Neta P, Shoute LCT (1990) J Phys Chem 94:8800 109. Alif A, Boule P (1991) J Photochem Photobiol A Chem 59:357 110. Suzuki J, Yagi N, Suzuki S (1984) Chem Pharm Bull 32:2803 111. Bayliss NS, Bucat RB (1975) Aust J Chem 28:1865 112. Dubovitskii AV, Leksina LN, Manelis GB (1981) Khim Vys Energ 15:347 113. Machado F, Boule P (1994) Toxicol Environ Chem 42:155 114. Suzuki J, Sato T, Ito A, Suzuki S (1990) Bull Environ Contam Toxicol 45:516 115. Lind J, Shen X, Eriksen TE, Merényi G (1990) J Am Chem Soc 112:479 116. Alif A, Boule P, Lemaire J (1987) Chemosphere 16:2213 117. Alif A, Boule P, Lemaire J (1990) J Photochem Photobiol A Chem 50:331 118. Alif A, Pilichowski JF, Boule P (1991) J Photochem Photobiol A Chem 59:209 119. Dubowski Y, Hoffmann MR (2000) Geophys Res Lett 27:3321 120. Klán P, Holoubek I (2002) Chemosphere 46:1201 121. Vanni A, Pellegrino V, Gamberini R, Calabria A (2001) Int J Environ Anal Chem 79:349 122. Machado F, Boule P (1994) Toxicol Environ Chem 42:165 123. Schedel G, Lenoir D, Boule P (1991) Chemosphere 22:1063 124. Sarakha M, Boule P, Lenoir D (1993) J Photochem Photobiol A Chem 75:61 125. Harrison MAJ (2003) PhD thesis, University of Edinburgh, UK

Reactions Induced in Natural Waters by Irradiation of Nitrate and Nitrite Ions

253

126. Lüttke J, Scheer V, Levsen K, Wünsch G, Cape JN, Hargreaves KJ, Storeton-West RL, Acker K, Wieprecht W, Jones B (1997) Atmos Environ 31:2637 127. Hinkel M, Reischl A, Schramm K-W, Trautner F, Reissinger M, Hutzinger O (1989) Chemosphere 18:2433 128. Schüssler W, Nitschke L (2001) Chemosphere 42:277 129. Beake BD, Constantine J, Moodie RB (1992) J Chem Soc Perkin Trans II, 1653 130. Suzuki J, Sato T, Suzuki S (1985) Chem Pharm Bull 33:2507 131. Suzuki J, Sato T, Ito A, Suzuki S (1987) Chemosphere 16:1289 132. Kotzias D, Hustert K, Parlar H, Korte F (1983) Naturwissenschaften 70:413 133. Bunce NJ, Cater SR, Willson JM (1985) J Chem Soc Perkin Trans II, 2013 134. Barlas H, Parlar H, Kotzias D, Korte F (1982) Chemiker-Zeitung 106:293 135. Barlas H, Parlar H, Kotzias D, Korte F (1982) Z Naturforsch 37b:486 136. Suzuki J, Hagino T, Suzuki S (1987) Chemosphere 16:859 137. Ramdahl T, Bjørseth A (1984) Chemosphere 13:527 138. Sugiyama H, Watanabe T, Hirayama T (2001) J Health Sci 47:28 139. Beitz T, Bechmann W, Mitzner R (1999) Chemosphere 38:351 140. Kochany J, Choudhry GG (1990) Toxicol Environ Chem 27:225 141. Ohta T, Suzuki J, Iwano Y, Suzuki S (1982) Chemosphere 11:797 142. Goldstein S, Czapski G (1996) J Am Chem Soc 118:3419 143. Usui Y, Takebayashi S, Takeuchi M (1992) Bull Chem Soc Jpn 65:3183 144. Nélieu S, Kerhoas L, Sarakha M, Einhorn J (submitted) Environ Chem Lett 145. Vione D, Maurino V, Minero C, Lucchiari M, Pelizzetti E (submitted) Chemosphere 146. Sadiki M, Quentel F, Elléouet C, Huruguen J-P, Jestin J, Andrieux D, Olier R, Privat M (2003) Atmos Environ 37:3551 147. Anastasio C, McGregor KG (2001) Atmos Environ 35:1091 148. Barletta B, Bolzacchini E, Meinardi S, Orlandi M, Rindone B (2000) Environ Sci Technol 34:2224 149. Nojima K, Ohya T, Kanno S, Hirobe H (1982) Chem Pharm Bull 30:4566 150. Nojima K, Kawaguchi A, Ohya T, Kanno S, Hirobe M (1983) Chem Pharm Bull 31:1047 151. Rippen G, Zietz E, Frank R, Knacker T, Klöpffer W (1987) Environ Technol Lett 8:475 152. Vione D, Maurino V, Minero C, Pelizzetti E (2003) Ann Chim (Rome) 93:477

Hdb Env Chem Vol. 2, Part M (2005): 255–298 DOI 10.1007/b138186 © Springer-Verlag Berlin Heidelberg 2005 Published online: 10 September 2005

Role of Iron in Light-Induced Environmental Processes T. David Waite Centre for Water and Waste Technology, School of Civil and Environmental Engineering, The University of New South Wales, NSW 2052 Sydney, Australia [email protected] 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2

Mechanistic Aspects of Homogeneous Photochemical Processes Involving Iron Photolysis of Inorganic Iron Species . . . . . . . . . . . . Implications to Transformations of Other Species . . . . . Photolysis of Iron-Organic Complexes . . . . . . . . . . . Simple Ligands . . . . . . . . . . . . . . . . . . . . . . . . Well-Deﬁned Naturally Occurring Ligands . . . . . . . . . Iron-Related Photochemical Transformations Mediated by Natural Organic Matter . . . . . . . . . . . .

. . . . . .

258 259 260 266 266 279

. . . . . . . . .

280

3.1 3.2

Mechanistic Aspects of Heterogeneous Photochemical Processes Involving Iron . . . . . . . . Photoreductive Dissolution of Iron Oxides . . . . . . . . . . . . . . . . . . Photooxidation of Organic Compounds . . . . . . . . . . . . . . . . . . .

285 285 289

4

Field Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

291

5

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

294

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

295

2.1 2.1.1 2.2 2.2.1 2.2.2 2.3

3

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

255

Abstract In this chapter, photochemical processes in natural aquatic systems involving iron in one way or another are reviewed. Homogeneous and heterogeneous processes are examined with attention given to both simple model systems in which the species distribution is generally well understood and to complex systems more typical of natural or treatment environments. Insights into mechanistic aspects are provided where possible and implications to natural and treatment systems assessed. Keywords Environmental processes · Iron · Natural waters · Photochemistry

1 Introduction As noted in the previous edition of Environmental Photochemistry [1], iron is the most abundant transition metal in soils. Despite this, it has been proposed that phytoplankton production in major ocean regions is restricted by

256

T.D. Waite

low iron availability [2]. This paradox arises because of the unique transformations that iron undergoes in natural aqueous systems. Some of these transformations are light-induced or light-assisted and it is these processes to which we will give particular attention in this chapter. The major transformations of iron are shown in Fig. 1 and involve both homogeneous and heterogeneous processes. Iron is particularly soluble in the reduced ferrous state and quite high (up to millimolar) concentrations of iron in this form may accumulate in anoxic groundwaters, interstitial porewaters and hypolimnetic regions of lakes. On introduction to oxygenated waters, this ferrous iron will oxidise to ferric iron (Process 1 in Fig. 1) at a rate that increases with pH since the ferrous hydroxy complexes that dominate Fe(II) speciation at higher pH are particularly susceptible to oxygenation [3]. The ferric iron so produced is highly insoluble under the pH conditions of most natural waters and will quickly hydrolyse and form particulate iron oxyhydroxides such as the amorphous ferrihydrite (Process 2) [4]. Only under strongly acidic or alkaline conditions will particulates such as this undergo dissolution (Process 3). Indeed, amorphous oxides will age relatively quickly to more stable (less soluble) forms such as hematite and goethite. Natural organic matter (NOM) present in fresh and marine waters may inﬂuence the inorganic transformation pathways mentioned above by forming complexes with Fe(II) (Process 4). These complexes are typically relatively weak and will readily dissociate if inorganic Fe(II) is removed from solution (as may be the case if organisms speciﬁcally utilise ferrous iron) (Process 5). These complexes will also oxidise to Fe(III) in the presence of oxygen (Process 6) with the rate of oxygenation determined by the relative Fe(III) and Fe(II) binding strengths [5, 6]. Some of these ligands (particularly the so-called “siderophores” that are released by some organisms to render iron bioavailable) are strong enough to reduce the ferric iron activity below that necessary to sustain iron oxide precipitation (Process 7) [7]. Even when the ligands are not particularly strong and iron oxide precipitates are the thermodynamically stable state, Fe(III)-NOM complex dissociation (Process 8) may be slow with ﬁrst order constants reported to be in the range 10–6 s–1 to 10–3 s–1 . These dissociation half-times correspond to lifetimes of the ferric-organic complexes of up to a week in some instances. Organic compounds of both natural and anthropogenic origin can also adsorb to particulate iron oxide surfaces (Process 9) leading to the formation of either strongly bound ligands (“surface complexes”) or more loosely associated assemblages. The interaction of organic compounds with iron oxyhydroxide surface hydroxy groups is typically pH dependent and desorption would be expected to occur (Process 10) if other anions outcompete the organic ligands for surface sites or if the pH renders organic ligand adsorption unfavourable (as would be expected at very low pH due to proton competition for surface sites or high pH because of electrostatic effects).

Role of Iron in Light-Induced Environmental Processes

257

Fig. 1 Schematic showing major transformations of iron in aquatic systems with lightmediated processes indicated

258

T.D. Waite

While ferric species would be expected to dominate iron speciation under oxic conditions, light may enhance the transfer of electrons to the metal centre with a resultant reduction of Fe(III) to Fe(II). Both dissolved ferric hydroxy species as well as organically complexed Fe(III) are susceptible to light-induced reduction (Processes 11 and 12 respectively). Detailed consideration to these processes is given in Sects. 2.1 and 2.2 of this chapter. While natural organic matter does form strong complexes with iron which exhibit chromophoric behaviour related speciﬁcally to the nature of the complex formed, the NOM also absorbs light in its own right resulting in the possible production of a variety of active species (such as superoxide and hydroxyl radicals). These agents may, in turn, induce transformations in iron species and consideration to such processes is given in Sect. 2.3. Particulate iron oxides and oxyhydroxides may also be inﬂuenced by light, particularly as a result of adsorption of light either by surface iron hydroxy groups or by the bulk oxide, especially if it exhibits semiconducting properties as is the case for hematite (α-Fe2 O3 ) and magnetite (Fe3 O4 ). In both cases, light-induced reduction of Fe(III) to Fe(II) occurs with loss of Fe(II) to solution with concomitant dissolution of the particulate phase (Process 13). Organic surface complexes with ferric iron may be even better chromophores than surface Fe(III) hydroxy groups and lead to even more rapid dissolution of the particulate iron oxide/oxyhydroxide in the presence of light (Process 14). These heterogeneous processes are described in detail in Sect. 3. While a wide range of light-mediated transformations of iron are possible, it is important to place these transformations in perspective and assess the extent of their occurrence in natural aquatic systems. Within this context, evidence from ﬁeld studies for light-mediated transformations of iron is reviewed in Sect. 4 and a brief discussion provided of the implications of these transformations to both iron bioavailability and contaminant mobilisation.

2 Mechanistic Aspects of Homogeneous Photochemical Processes Involving Iron Mechanistic aspects of homogeneous photochemical processes involving iron are considered in this section with attention given to light-mediated transformations of both inorganic and organic complexes of iron. While the analysis is presented from a relatively fundamental level, consideration is only given to transformations that might be of signiﬁcance in natural aquatic environments or common aqueous treatment systems.

Role of Iron in Light-Induced Environmental Processes

259

2.1 Photolysis of Inorganic Iron Species That inorganic ferric iron species are active chromophores has been known for many years [8] with the FeOH2+ species (which dominates ferric speciation at 3 < pH < 5) recognised to be the most photoactive, particularly in the near UV region (Fig. 2a) [9, 10]. Photolysis of this species by ultraviolet (UV) or near-UV light results in the reduction of the metal centre to Fe(II) by ligand to metal charge transfer and generation of strongly oxidising hydroxyl radicals, HO· ; i.e. FeOH2+

hv

→ Fe2+ + HO·

As shown in Fig. 2b, HO· quantum yields decrease with increasing wavelength (an effect that Benkelberg and Warneck [10] ascribe to the energy required

Fig. 2 a Molar absorption coefﬁcients for ferric species of interest (Fe3+ and FeOH2+ ), and b HO· quantum yields obtained from photolysis of pH 3 ferric solutions in the wavelength range 280–370 nm in the presence and absence of oxygen (data from [10])

260

T.D. Waite

for hydroxyl radicals to escape the solvent cage around the metal ion) with similar quantum yields observed in the presence and absence of oxygen. 2.1.1 Implications to Transformations of Other Species The hydroxyl radicals generated on photolysis of FeOH2+ have been shown to be effective oxidants for a wide range of organic contaminants including 2-chlorophenol [11], 3-chlorophenol [12] and 4-chlorophenol [13], 2,6dimethylphenol (DMP) [14], the herbicides diuron (3-(3,4-dichlorophenyl)1,1-dimethylurea) [15], asulam (4-amino-benzosulfonyl-methylcarbamate) [16], and monouron [17], the plasticiser dibutyl phthalate (DBP) [18], an alkylphenol ethoxylate typical of those used as detergents and emulsiﬁers [19] and benzoic acid [20]. In all cases, the organic compounds do not form complexes with the non-excited ferric iron and FeOH2+ is the chromophore with the hydroxyl radicals generated by this route attacking the compounds at the most susceptible locations. Aliphatic chains are particularly susceptible to attack by HO· radicals and attack on these chains represents the major mode of contaminant degradation. This is certainly the case for degradation of dibutyl phthalate [18] with the most labile H atom being that on the methylene group in α-position to the ester function (Struct. 1).

Structure 1

The resulting carbon centred radical (R· ) is a highly reductive agent which can react either with oxygen or with Fe(III) to produce an hydroxylated compound (Scheme 1). The hydroxylated compound produced by this pathway is highly unstable and decomposes in water to the carboxylic derivative (Struct. 2).

Structure 2

A variety of other products are observed including aldehyde or ketone derivatives which presumably arise from hydroxyl radical attack on the last three carbons of the alkyl chain (Scheme 2).

Role of Iron in Light-Induced Environmental Processes

261

Scheme 1

Photolysis of a pH 3 × 300 µM solution of Fe(III) (most of which was in the form of FeOH2+ ) resulted in 85% conversion of DBP to reaction products within approximately 90 minutes. The concentration of Fe(II) produced was also monitored and found to reach a steady state concentration of around 100 µM after 4–5 hours. The plateauing in [Fe(II)] was attributed to the reoxidation of Fe(II) species by hydroxyl radicals; i.e. Fe2+ + HO· → Fe3+ + OH– This reaction, which exhibits rate constants in the range 1.8 to 4.0 × 108 mol–1 L s–1 [21], is believed to be an important factor in regenerating the ferric chromophore and thus in enabling the contaminant degradation process to proceed in a catalytic manner [18]. In addition to aliphatic chain attack, hydroxyl radicals may also directly attack the aromatic ring. This is believed to be the case for the iron-mediated photodegradation of 3-chlorophenol (3CP) [12] with hydroxyl radicals that are formed on photolysis of FeOH2+ rapidly reacting with 3CP to form rad-

262

Scheme 2

Scheme 3

T.D. Waite

Role of Iron in Light-Induced Environmental Processes

263

ical adducts such as the dihydroxycyclohexadienyl radical [22] (Scheme 3). In the presence of oxidants such as iron(III) or oxygen, these adducts can evolve to the major identiﬁed photoproducts in this instance, chlorohydroquinone (I) [which is rapidly oxidised to chlorobenzoquinone (II) by Fe(III)], 3-chlorocatechol (III) and 4-chlorocatechol (IV). In some cases, degradation products may be produced as a result of attack on both the aromatic ring and aliphatic components of the compound. This is the case for diuron where hydroxylated photoproducts 1 and 2 (Scheme 4) are produced by reaction of HO· with the aromatic ring (Process A) to form an adduct similar to that described above with subsequent loss of a chlorine atom [15]. The more important degradation pathway in this instance is via hydroxyl radical attack on the methyl of the dimethylurea group (Process B).

Scheme 4

264

T.D. Waite

The alkyl radical formed by hydrogen abstraction reacts with oxygen to form primary photoproducts 3 and 4. These intermediate products are, in turn, subject to hydroxyl radical attack with resultant formation of the secondary photoproducts 5 and 6 (Scheme 4). In addition to oxidation of organic compounds, organometallic compounds such as tributyl tin (TBT) have been shown to be readily degraded by ferric iron-mediated photochemical processes [23]. Photolysis of low ppm solutions of TBT resulted in almost complete conversion to inorganic tin within 24 hours with a 97% yield of CO2 obtained over long (> 40 h) irradiation times. GC-FID analysis indicated the formation of di- and monobutyl tin as

Scheme 5

Role of Iron in Light-Induced Environmental Processes

265

well as several organotin derivatives, ketones and alcohols. Not surprisingly, quantum yields of tin disappearance increase as irradiation wavelengths decrease (Table 1). A mechanism consistent with the observed degradation products is shown in Scheme 5 in which the attack of the hydroxyl radicals occurs primarily by abstraction of a hydrogen atom from the carbon of a butyl group in the α-position to tin. Interestingly, recent studies have suggested that photolysis of Fe(II) rather than Fe(III) solutions can also result in effective contaminant degradation [16]. Photolysis of ferrous iron solutions was only found to be effective in inducing contaminant degradation when oxygen was present and led to the conclusion that light was increasing the rate of oxygenation of ferrous iron to Fe(III) with production of hydroperoxy radicals. It was surmised that these radicals disproportionate to hydrogen peroxide which, in turn, react with Fe(II) via a Fenton process to produce additional hydroxyl radicals (Fig. 3). These observations require further analysis, particularly since the product yield of the primary photoprocess (i.e., the photolysis of FeOH2+ ) is oxygen independent. Table 1 Quantum yields of TBT disappearance as a function of wavelength of irradiation of FeOH2+ solutions (from [23]) λ (nm)

365

334

313

296

φTBT

1.6 × 10–3

7.0 × 10–3

1.0 × 10–2

1.85 × 10–2

Fig. 3 Schematic showing the proposed interplay between iron, oxygen and light in a system with light mediating both the reduction of Fe(III) and the oxidation of Fe(II) (after [22])

266

T.D. Waite

2.2 Photolysis of Iron-Organic Complexes Given the insolubility of ferric oxyhydroxides and ferric oxides under conditions typical of natural waters, the concentration of Fe(III) in equilibrium with these solids is very low. Thus, while the photochemistry of ferric hydroxy species is of some interest at low pH (i.e., 3–5), the concentrations of these species in the pH 6–8 range more typical of natural waters is very low and insufﬁcient to drive substantial change in iron speciation or induce contaminant degradation. The concentration of dissolved Fe(III) however can be increased substantially by the addition of strong ferric iron complexing agents. The complexes so formed may absorb light (often to a greater extent than FeOH2+ ) and undergo redox transformations which typically results in reduction of the metal centre and degradation of the ligand (so-called “ligand-tometal charge transfer”). Depending upon reaction conditions (e.g., presence of oxygen or hydrogen peroxide, pH), these light-mediated redox transformations may generate reactive species which are capable of degrading other (possibly contaminant) compounds that may be present. The ability of any ligand (L) to outcompete hydroxide for Fe(III) is simply a function of the strength of the FeIII L complex and the pH and may be deduced from straightforward thermodynamic analysis [24, 25]. As ﬂagged in the Introduction however, Fe(III) complexes may form under nonequilibrium conditions in natural systems and be present for periods of time sufﬁcient to inﬂuence the bioavailability of iron to aquatic organisms. This issue will be discussed further in Sect. 2.3. 2.2.1 Simple Ligands Iron complexes with both inorganic and organic ligands have perhaps undergone more intensive investigation of their photochemical and photophysical processes than any other group of compounds with comprehensive reviews providing detailed coverage of these studies [26, 27]. Of particular interest in this overview are compounds that either (i) occur in nature, (ii) provide good models of natural processes, or (iii) are used in waste treatment processes. We focus initially on the photochemical behaviour of complexes of Fe(III) with simple carboxylic acids and give particular attention to oxalic acid. This compound is prevalent in atmospheric aerosols [28], provides a simple example of environmentally important light-mediated ligand-to-metal charge transfer (LMCT) processes which result in ligand decarboxylation [27] and is used to initiate the degradation of contaminants both in the absence and presence of added hydrogen peroxide (via the so-called “modiﬁed photo-Fenton process” [29, 30]). In addition, the photochemistry of Fe(III)-oxalate complexes has been studied in detail, as it is the basis of

Role of Iron in Light-Induced Environmental Processes

267

the well-known ferrioxalate actinometer. The use of potassium ferrioxalate (K3 Fe(C2 O4 )3 ) as a chemical actinometer was ﬁrst proposed by Parker [31] and Hatchard and Parker [32]. The photolysis of Fe(III)-oxalate complexes such as FeIII (C2 O4 )3 3– was thought to occur as follows [33]: FeIII (C2 O4 )3 3– + hν ↔ FeII (C2 O4 )2 (C2 O4 · )3– FeII (C2 O4 )2 (C2 O4 · )3– → FeII (C2 O4 )2 2– + C2 O4 · – FeIII (C2 O4 )3 3– + C2 O4 · – → FeII (C2 O4 )2 2– + C2 O4 2– + 2CO2 C2 O4 · – + O2 → 2CO2 + O2 · – The ﬁrst step is the formation of an Fe(II) complex with an oxalate radical ligand. The oxalate radical can undergo an intramolecular reduction back to oxalate or be released into the bulk solution. C2 O4 · – is consumed by reacting with either Fe(III) or O2 . The details of the above process have been reﬁned somewhat by subsequent ﬂash photolysis studies [34–38]. However, in all of the mechanisms that have been discussed, a Fe(III)-oxalate complex is reduced by an intramolecular photochemical process, generating an unbound radical (either C2 O4 · – or CO2 · – ). If a C2 O4 · – radical is produced, it rapidly decomposes to form CO2 · – [39]: C2 O4 · – → CO2 · – + CO2 As shown above, the C2 O4 · – or CO2 · – radical reduces a second Fe(III)-oxalate complex, or can instead react with O2 [40, 41]: C2 O4 – + O2 → 2CO2 + O2 · – CO2 · – + O2 → CO2 + O2 · – The O2 · – (or HO2 , at pH < 4.8, the pKa of the HO2 · /O2 · – acid-base pair) radicals that are produced will participate in further reactions, including the generation of H2 O2 by disproportionation; i.e. HO2 · + HO2 · → H2 O2 + O2 Both C2 O4 · – and CO2 · – radicals can be protonated: C2 O4 · – + H+ ↔ HC2 O4 · CO2 · – + H+ ↔ HCO2 · The pKa of HCO2 · was reported to be 1.4 [42]. The pKa of HC2 O4 · is unknown but was estimated by Rowan et al. [43] to lie somewhere between the pKa of HC2 O4 · – (4.1) and the pKa of the HCO2 · radical (1.4). The protonated radicals may have reactivities different from the corresponding deprotonated radicals [28, 33]. These radicals could undergo radical-radical reactions but such reactions are only likely to be signiﬁcant at high light intensity when relatively high radical concentrations are generated [35].

268

T.D. Waite

Several authors have studied the effect of various parameters on the quantum yield of Fe(II) production (ΦFe(II) ) from Fe(III)-oxalate complexes in the absence of oxygen. Parker [33] and Hatchard and Parker [32] found that ferrioxalate concentration, pH, light intensity and temperature did not signiﬁcantly affect ΦFe(II) at 366 nm. Abrahamson et al. [44] found, however, that ΦFe(II) at 366 nm was approximately halved when the pH was increased from 2.7 to 4.0 or when a large excess of oxalate was added. At other wavelengths the solution pH and the presence of excess oxalate has a large effect on ΦFe(II) (Table 2). While Hatchard and Parker [32] concluded that complexes of Fe(III) with 1–3 oxalate ligands have equal quantum yields at 366 nm, Vincze and Papp [50] determined quantum yields at 254 nm of 0, 1.18 and 1.60 for Fe(C2 O4 )+ , Fe(C2 O4 )2 – and Fe(C2 O4 )3 3– , respectively. In view of these markedly different quantum yields, and the changes in Fe(III) oxalate speciation expected with change in solution composition (Fig. 4), it is to be expected that the quantum yields of Fe(II) production from Fe(III)-oxalate complexes will be signiﬁcantly affected by the solution composition and pH. While the quantum yields reported by Hatchard and Parker [32] for ferrioxalate actinometry have been conﬁrmed by several other authors (references cited in [26]), values of ΦFe(II) under other conditions have not been well studied. In acidic solutions oxygenation has little effect on the photoproduction of Fe(II) from ferrioxalate [32]. In neutral solutions, however, the presence of oxygen reduces ΦFe(II), especially in the presence of excess oxalate [31, 48]. Table 2 Effect of solution pH and excess oxalate on the quantum yield of Fe(II) production (ΦFe(II) ) for Fe(III)-oxalate complexes in the absence of oxygen (from [49])

λ (nm)

Strongly acidic solutions Hatchard and Baxendale and Parker [32] Bridge [45]

Neutral solutions Allmand and Livingston [48] Webb [46, 47]

254 313 365/366 405 436

1.25 1.24 1.21 1.14 1.11

1.59 1.16 0.91 0.89

1.22 1.20 1.04

1.14 0.88

Note: Strongly acidic solutions contained 0.006 M K3 Fe(C2 O4 )3 and 0.05 M H2 SO4 (Hatchard and Parker [32]) or 0.005 M K3 Fe(C2 O4 )3 and 0.05 M H2 SO4 (Baxendale and Bridge [45]); neutral solutions contained 0.02–0.06 M K3 Fe(C2 O4 )3 and 0.18 M K2 C2 O4 (Allmand and Webb [46, 47] or 0.02 M K3 Fe(C2 O4 )3 and 0.10 M K2 C2 O4 (Livingston [48]). Livingston [48] reported quantum yields for oxalate degradation, which have been converted to quantum yields of Fe(II) production.

Role of Iron in Light-Induced Environmental Processes

269

Fig. 4 Calculated Fe(III) speciation as a function of pH in systems containing 6 µM Fe(III) and 18 µM oxalate (a) or 180 µM oxalate (b)

The oxidation of Fe(II) by O2 is also relatively fast in neutral solutions and in the presence of excess oxalate [31]—under these conditions Fe(C2 O4 )2 2– will form, and O2 reacts much more quickly with Fe(C2 O4 )2 2– than with Fe(C2 O4 ) or Fe2+ [52]. The presence of O2 also increases the quantum yield of oxalate degradation, and therefore the reduction in ΦFe(II) cannot be caused only by the reoxidation of Fe(II) by O2 [48]. Oxygen may also reduce the photoproduction of Fe(II) by reacting with C2 O4 · – and CO2 · – radicals with the resulting generation of superoxide which disproportionates to hydrogen peroxide. The Fe(II) and H2 O2 so produced may react via the so-called Fenton reaction to produce hydroxyl radicals, i.e. FeII + H2 O2 → FeIII + · OH + OH– Key steps in this reaction are highlighted in Fig. 5 where reactions I involve the formation of an excited charge-transfer complex on photolysis of Fe(III)oxalate followed by lysis of this complex to Fe(II) and an organic radical, reaction of the organic radical with oxygen to give superoxide which disproportionates to H2 O2 (reactions II). Hydrogen peroxide then yields hydroxyl radicals by reaction with Fe(II) (reaction III). A number of other Fe(III)-carboxylate complexes are also photolysed, producing Fe(II), CO2 and ligand oxidation products. The mechanism is generally thought to be similar to that for Fe(III)-oxalate complexes [26, 54, 55]: Fe(RCO2 )2+ + hν → Fe2+ + R· + CO2 Fe(RCO2 )2+ + R· → Fe2+ + RCO2 – + R+

270

T.D. Waite

Fig. 5 Major reactions in a system involving photolysis of Fe(III) oxalate in the presence of oxygen (from [53])

A Fe(III)-carboxylate complex is reduced by an intramolecular photochemical process. This produces a radical that can reduce a second Fe(III)carboxylate complex. There have been only a small number of studies on the quantum yields of Fe(II) production by photolysis of other Fe(III)-carboxylate complexes. Abrahamson et al. [44] determined ΦFe(II) at 366 nm for several Fe(III)carboxylate complexes in the absence of oxygen. The quantum yield depended on the ligand, with oxalate > tartrate > citrate > isocitrate > malate at pH 2.7. When the pH was increased to pH 4.0, ΦFe(II) increased by about 50% for all Fe(III)-carboxylate complexes except for Fe(III)-oxalate, for which ΦFe(II) decreased by 50%. All of the quantum yields were reduced in the presence of a large excess of carboxylate. Faust and Zepp [51] measured ΦFe(II) for complexes of Fe(III) with oxalate, citrate and malonate in the absence of oxygen. The quantum yields at 436 nm for Fe(III)-oxalate complexes at pH 3–4 were greater than the equivalent quantum yields for Fe(III)-citrate complexes at pH 4–6. The quantum yields of Fe(III)-malonate complexes at 366 nm were very low compared with those of oxalate and citrate complexes at 436 nm. Kuma et al. [55] studied the photoreduction of Fe(III) in seawater (pH 8.1) in the presence of carboxylic acids. The rate of Fe(III) reduction by sunlight varied with the presence of different carboxylic acids in the order glucaric > tartaric citric > malic acid. The order was thought to depend upon both the complexation ability of each carboxylic acid with Fe(III) and the photoactivity of each Fe(III)-carboxylate complex.

Role of Iron in Light-Induced Environmental Processes

271

Another group of iron complexes for which the environmental photochemistry has been studied in some detail are the aminopolycarboxylic acids. These ligands are widely used as sequestering agents in industrial cleaning, household detergents, the photographic industry, pharmacy products and textile and paper manufacturing. Among them, ethylenediamino-tetraacetic acid (EDTA) and nitrilotriacetic acid (NTA) are the most commonly used [1]. EDTA is only slowly degraded in sewage treatment [56] thus substantial concentrations may accumulate in natural waterbodies where abiotic degradation pathways will determine its fate [57]. As a result, there have been a number of studies of the photochemical degradation of both EDTA and NTA.

Structure 3

The speciation of ferric iron in the presence of EDTA is strongly dependent upon solution conditions. As shown in Fig. 6, the FeEDTA– and Fe(OH)EDTA2– species dominate over the pH 4–8 range in solutions containing 1 µM of each of total iron and EDTA.

Fig. 6 Dissolved iron and EDTA species in a system with ionic strength of 0.1 M, [dissolved Fe(III)]TOTAL = 10–6 M and [EDTA]TOTAL = 10–6 M (after Kari et al. [57])

272

T.D. Waite

Fig. 7 Reaction quantum yield of Fe(III)EDTA in 1 mM phosphate buffer as a function of pH and photolysis wavelength (after Kari et al. [57])

Despite the strong pH dependence in relative importance of the various Fe(III) EDTA species, the reaction quantum yields of Fe(III) EDTA (as determined from Fe(III)EDTA degradation rates) are relatively pH independent (Fig. 7) suggesting that both major forms of Fe(III)EDTA complexes exhibit similar photochemical behaviour [57]. Other workers have reported strong pH dependence in the reaction quantum yield, however, these effects appear to be related to the formation of dimers at higher concentrations and high pH which are prone to photoexcitation and relaxation effects (rather than LMCT) which lead to a decrease in the reaction quantum yield in this pH range [58, 59]. As is typically the case for such LMCT processes, increasing irradiation wavelength results in a decrease in quantum yield. The major photoproducts formed are Fe(II) ethylenediaminotriacetic acid (ED3A), formaldehyde, CO2 and ferrous species that are most likely complexed to EDTA. A ratio of quantum yields ΦFe(II) : ΦCO2 : ΦHCHO of 2 : 1 : 1 has been reported and supports the mechanism shown in Scheme 6 [58, 60, 61]. A similar photodegradation mechanism is considered to hold for ferric nitrilotriacetate complexes at 365 nm in acidic medium [62]. The quantum yield as a function of irradiation wavelength and pH is shown in Table 3 and major photoproducts shown below. In near neutral solutions, photosolvation occurs with the release of NTA and an aquo complex of Fe(III). At 254 nm, photolysis of the aquocomplex results in production of hydroxyl radicals (as discussed earlier) which subse-

Role of Iron in Light-Induced Environmental Processes

273

Table 3 Quantum yields of FeNTA disappearance as a function of excitation wavelength at pH 4 and 7 (from [62]) λ (nm)

Φ (pH 4)

Φ (pH 7)

254 296 313 334 365

0.50 0.40 0.30 0.22 0.18

0.19 0.18 0.08 – 0.09

Scheme 6

Structure 4

quently oxidises the organic moiety but this process is relatively ineffective at 365 nm with the result that only a small portion of FeNTA is degraded. At lower pH, LMCT processes result in degradation of the complex. Under acidic conditions, the iminodiacetic acid (IDA) formed also binds to Fe(III) and is degraded to Fe(II) and formaldehyde [63]. 2.2.1.1 Implications to Transformation of Other Species As indicated above, the photolysis of Fe(III)-carboxylate and aminocarboxylate complexes may result in oxidative degradation of the ligand and reduction of the metal centre to Fe(II). In addition, these complexes may play a signiﬁcant role in the photochemical generation of reactive species such as O2 · – , H2 O2 and OH· [28, 51, 64]. The degradation of a variety of organic con-

274

T.D. Waite

taminants has been examined and attributed to attack by hydroxyl radicals generated on photolysis of Fe(III) carboxylate species [65]. A good example of such a process is the degradation of atrazine on photolysis of ferrioxalate solutions containing this triazine herbicide [53].

Structure 5

As can be seen in Fig. 8, atrazine degrades at an appreciable rate in the presence of Fe(III) alone, presumably as a result of hydroxyl radicals generated on photolysis of FeOH2+ . The rate is observed to increase on addition of 18 µM oxalate and to increase further on increase in oxalate concentration to 180 µM. Addition of methanol in this latter case prevents degradation supporting the contention that degradation is occurring as a result of hydroxyl radical attack. As can be seen from Fig. 4, increasing the oxalate concentration to 180 µM results in formation of the more photoactive di- and tri-oxalato species. More detailed studies of the pH dependence of atrazine degradation showed that in the presence of an initial 18 µM oxalate, the rate increased in the order of pH 7.5 < 5.6 < 3.2 < 4.3, and with 180 µM oxalate in the order of pH

Fig. 8 Effect of different initial oxalate concentrations (0, 18 and 180 µM) and of excess methanol on atrazine degradation in systems containing 6 µM total iron and 0.47 µM atrazine (initial concentrations) at pH ≈ 3 (from [53])

Role of Iron in Light-Induced Environmental Processes

275

7.9 < 3.2 < 4.6 ≈ 5.4. Balmer and Sulzberger [53] attributed the results to a number of competing processes. First, both pH and oxalate concentration control the Fe(III) speciation and thus the rate of photolysis of Fe(III) complexes. Second, both pH and oxalate concentration control Fe(II) speciation and thus the rate of oxidation of Fe(II) by H2 O2 . Thirdly, oxalate acts as a scavenger of hydroxyl radicals and hence competes with atrazine for these radicals. A very similar process to that described for atrazine degradation is believed to be the case for degradation of 4-chlorophenol (4-CP) mediated by photolysis of Fe(III)-NTA solutions [66]. The quantum yields for Fe-NTA disappearance, Fe(II) formation and 4-CP disappearance as a function of wavelength are shown in Table 4 and, as is the case for photolysis of Fe(III) aquocomplexes, decrease with increasing wavelength of irradiation. In the absence of oxygen, the quantum yields of Fe-NTA disappearance and Fe(II) formation were lower (by 25%) and the quantum yield of 4-CP disappearance was halved. Increasing the concentration of Fe-NTA resulted in an increase in extent of 4-CP removal though once Fe-NTA degradation was complete, the rate of 4-CP removal decreased substantially (Fig. 9). Some continuing Table 4 Quantum yields of 4-CP disappearance and Fe(II) formation in pH 4 solutions of Fe-NTA (from [66]) λ (nm)

ΦFeNTA disappearance

ΦFe(II) formation

Φ4–CP disappearance

313 313 (without O2 ) 365 365

0.50 0.36 0.27 without NTA

0.46 0.36 0.28 0.03

0.018 0.009 0.011 0.003

Fig. 9 Disappearance of 4-CP (initial concentration 0.2 mM) in the presence of 0.3 mM Fe-NTA (initial concentration). Both the disappearance of 4-CP (a) and Fe-NTA (b) are shown (from [66])

276

T.D. Waite

degradation however was observed, possibly as a result of hydroxyl radicals generated on photolysis of FeOH2+ or a result of the generation of iminodiacetic acid which, as discussed above, is itself a Fe(III) complexant and chromophore. This example highlights the drawback of such a process for waste treatment. While photolysis of oxic solutions of ferric complexes results in generation of hydroxyl radicals, the process induces the degradation of the chromophore. Additionally, the hydroxyl radicals so generated may react with the ligand in preference to any contaminant present. If such a process is to be sustainable, iron complexes which are photoactive but which enable lightmediated electron transfer without ligand destruction are required. It will also be important that the ferrous form of any such complex can react quickly with oxygen resulting in reformation of Fe(III) species which continue to be photoactive. It will also be important that any activated forms of oxygen that result do not attack the iron-binding ligand. It is likely that useful insights may come from analysis of the light capture and electron transfer processes in biological systems. Indeed, recent studies of the photochemistry of iron porphyrins suggest that systems may be engineered which mimic the catalytic activity of cytochrome P-450 oxygenases [67] though cost-effective assemblages of adequate efﬁciency are still some way off. In addition to transformation of organic contaminants by photolysis of iron carboxylate solutions, selected inorganic contaminants have also been shown to undergo light and iron-mediated changes. Thus, Hug et al. [68]

Fig. 10 Decrease of As(III) and increase of As(V) during illumination of (initially) 500 µg/L As(III) at pH 7.0–7.3 with 1 µM Fe(II, III), 0 and 50 µM citrate and 0 and 16.6 mM 2-propanol as indicated (from [68])

Role of Iron in Light-Induced Environmental Processes

277

have shown that As(III) is oxidised to As(V) on photolysis of solutions containing iron and citric acid (Fig. 10). These workers suggested that oxidation occurred as a result of oxidants generated on photolysis of ferric citrate complexes but, in view of the apparent lack of effect of isopropanol addition, deduced that hydroxyl radicals were not directly involved in the oxidation step. Other possible oxidants considered included superoxide radicals and high-valent iron species such as the ferryl complex Fe(IV) [69–71] or ferrate entities Fe(V, VI) [72] but deﬁnitive assignment was not possible. Kocar and Inskeep [73] have examined the photochemical oxidation of As(III) in ferrioxalate solutions and conclude that the hydroxyl radical is the major oxidising species. The difference between the ﬁndings of Kocar and Inskeep [73] and Hug et al. [68] may relate to the nature of the ligand and their ground-state one electron oxidation potentials which determine

Fig. 11 Cr(VI) concentrations versus illumination time (Xe light source ﬁltered to simulate solar spectrum). Filled symbols: aerated solutions; empty symbols: N2 - purged solutions. The lines are the results from a kinetic model developed to describe the data. The pH values and the initial concentrations in µM for Fe(III) and for oxalate are given (from [74])

278

T.D. Waite

whether hydroxyl radical addition or electron transfer (resulting in generation of ferryl species) are the dominant reaction pathways [71]. In contrast to oxidation of As(III) on photolysis of Fe(III) oxalate and Fe(III) citrate solutions, Cr(VI) is reported to be reduced on photolysis of solutions of these iron carboxylates [74]. As can be seen from Fig. 11, the rate and extent of Cr(VI) reduction was strongly dependent upon reactant concentrations, pH and the presence/absence of oxygen. By inspection of the results, Hug and coworkers [74] concluded that Cr(VI) was being reduced by both Fe(II) produced by light-mediated reduction of Fe(III) and superoxide/hydroperoxyl (O2 · – /HO2 · ) radicals produced by reduction of oxygen by CO2 · – radicals arising from the oxidation of oxalate. A schematic of the suggested pathways is shown in Scheme 7.

Scheme 7

Role of Iron in Light-Induced Environmental Processes

279

2.2.2 Well-Defined Naturally Occurring Ligands Dissolved iron (III) in the upper oceans occurs almost entirely in the form of complexes with strong organic ligands (siderophores) presumed to be of biological origin. Barbeau et al. [75] have shown strongly iron-binding ligands containing an α-hydroxy acid moiety are particularly prone to photochemical degradation and have used the well-characterised aquachelin series of siderophores to investigate this issue. The aquachelins are a suite of siderophores produced by the heterotrophic marine bacterium Halomonas aquamarina which was isolated from 40 m depth over the continental slope in the eastern equatorial Atlantic Ocean. The aquachelins each contain one of a series of fatty-acid tails and a peptidic headgroup that forms a 1 : 1Fe(III) : ligand complex via two hydroxamate groups and one βhydroxyaspartate residue [75] (Fig. 12). By analysis of the products formed, Barbeau et al. [75] have shown that photolysis of Fe(III)-aquachelin C results in oxidative cleavage of the ligand at the site of the β-hydroxyaspartate residue resulting in formation of a hydrophilic peptide component and a hydrophobic fatty-acid component (Fig. 12). The peptide photoproduct retains the two hydroxamate groups [75] and is capable of binding Fe(III) but exhibits a much weaker conditional stability constant than the original complex. Barbeau et al. [75] show that photolysis of the Fe(III) aquachelin substantially increases the bioavailability of iron and suggest that this process may be critical to the cycling of iron in the upper ocean. Barbeau et al. [76] have also characterised the siderophore Petrobactin, produced by the oil-degrading bacterium Marinobacter hydrocarbonaoclasti-

Fig. 12 Schematic representation of the photochemical reaction of Fe(III)-aquachelin complexes. The dashed line indicates the position of photolytic cleavage of the aquachelin ligand. Iron(III) is reduced to iron(II) via ligand-to-metal charge transfer (Orn = ornithine) (From [75])

280

T.D. Waite

Fig. 13 Schematic of the photochemical transformation of Fe(III)-Petrobactin (from [76])

cus. This is a bis-catechol α-hydroxy acid siderophore that readily undergoes a light-mediated decarboxylation reaction when bound to Fe(III) (Fig. 13). Barbeau et al. [76] suggest that the photodegradation of the complex may play a role in the biodegradation of petroleum hydrocarbons by facilitating microbial acquisition of iron, a limiting nutrient. Barbeau et al. [77] have compared the photoactivity of a range of marine, terrestrial and synthetic siderophores containing either hydroxamate, catecholate or α-hydroxy carboxylate moieties. They conclude that hydroxamate groups are photochemically resistant regardless of Fe(III) complexation. Catecholates, in contrast, are susceptible to photooxidation in the uncomplexed form but stabilised against photooxidation when ferrated. α-hydroxy carboxylate groups are stable as the uncomplexed acid, but when coordinated to Fe(III), these moieties undergo light-induced ligand oxidation and reduction of Fe(III) to Fe(II). It is concluded that these photochemical properties appear to determine the reactivity and fate of Fe(III)-binding siderophores in ocean surface waters which, in turn, might signiﬁcantly inﬂuence the biogeochemical cycling of iron [77]. 2.3 Iron-Related Photochemical Transformations Mediated by Natural Organic Matter While speciﬁc compounds such as siderophores have been found in natural waters, the bulk of the dissolved organic matter (DOM) is made up of relatively refractory compounds known as humic substances. These substances exhibit complex, ill-deﬁned structures with the actual structure depending markedly on the source of the organic material. DOM in the open ocean is almost entirely authochthonous and formed by condensation, polymerisation and partial oxidation of smaller molecules such as triglycerides, sugars and amino acids and exhibits very little aromatic character [78]. In contrast, the DOM in fresh and coastal waters is largely allochthonous and derived

Role of Iron in Light-Induced Environmental Processes

281

principally from the decaying material of higher plants. The resulting terrigenous humic compounds retain a relatively high degree of aromaticity from their precursors [79]. Both marine and freshwater humic substances possess carboxylic and hydroxy functional groups but phenolic groups are found predominantly in the DOM of waters near land with phenolic-to-carboxylic ratios in the range 1 : 2 to 1 : 4 [80]. While the metal-complexing behaviour of humic substances has been associated largely with chelation by hydroxy and carboxylate functional groups in close proximity to each other [24], redox behaviour has also been attributed to the presence of quinone moieties and the associated formation of semiquinones and hydroquinones [81]. Humic substances are widely recognised to strongly bind to iron and to inﬂuence the redox transformations of iron. These substances have been variously reported to accelerate, retard or have no effect on the rate of Fe(II) oxygenation under a variety of pH and dissolved oxygen conditions [82–87]. However, at typical environmental pH, recent studies have found that DOM in natural waters generally accelerates the oxidation reaction [5, 88], which has been attributed in part to the high density of carboxylate binding sites in many humic substances [89]. The mechanism via which organic substances affect the rate of oxygenation can be represented by a dual pathway approach (see Fig. 1), in which organic substances form complexes with Fe(II) that react with O2 at a different rate to inorganic Fe(II) [5, 82, 86, 88]. Binding of iron by organic substances can dramatically alter the redox potential of the Fe(II)/Fe(III) couple, which exhibits a remarkable range in E0 values from – 200 to + 300 mV when complexed with humic substances [90]. Rose and Waite [5] have developed a detailed kinetic model to describe the oxidation of Fe(II) in the presence of the well-characterised Suwannee River fulvic acid by extension of the Haber–Weiss model [91]. In this model, O2 reduces to OH– by oxidation of Fe(II) (either inorganic or complexed with an organic ligand L) in a four-step process: Fe(II)/FeII L + O2 → Fe(III)/FeIII L + O2 · – Fe(II)/FeII L + O2 · – + 2H+ → Fe(III)/FeIII L + H2 O2 Fe(II)/FeII L + H2 O2 → Fe(III)/FeIII L + OH· + OH– Fe(II)/FeII L + OH· → Fe(III)/FeIII L + OH– The superoxide produced on oxygenation of Fe(II) (and FeII L) may also reduce Fe(III) (and FeIII L) to Fe(II) (FeII L) [92]; i.e. Fe(III)/FeIII L + O2 · – → Fe(II)/FeII L + O2 though the concentrations of superoxide produced by the oxygenation reaction alone are unlikely to render this a signiﬁcant reduction pathway.

282

T.D. Waite

The transition between the inorganic and organically complexed states of ferrous iron is governed by the kinetics of the complexation reaction; i.e. Fe(II) + L → FeII L FeIII L complexes resulting from oxidation of FeII L may dissociate to inorganic Fe(III) which may hydrolyse and precipitate as Fe(OH)3 (s); i.e., for a wellbuffered system in which Fe(III) hydrolysis and precipitation is assumed to have a negligible effect on pH [4]: Fe(III) + Fe(III) → 2Fe(OH)3 (s) Fe(III) + Fe(OH)3 (s) → 2Fe(OH)3 (s) The tendency for Fe(III) to either remain in complexed (FeIII L) form or to precipitate will be a function of pH, the concentrations of iron and ligand L and the strength of the FeIII L complex. As noted by Rose and Waite [93], even when thermodynamically expected, the dissociation of FeIII L complexes can be slow with the result that iron may remain in complexed form for signiﬁcantly longer times (hours to days) than expected. The photolysis of natural waters containing DOM has long been recognised to induce the reduction of Fe(III) to Fe(II) species (coupled with a concomitant oxidation of the organic matter) [94–97]. The principal pathways by which these transformations have been suggested to occur [98–106] involve either (or both) (i) direct photolysis of the Fe(III)-DOM complexes (FeIII L) via ligand-to-metal charge transfer which results in formation of Fe(II) and organic free radicals; i.e. FeIII L + hν → Fe(II) + L· – with the possibility of secondary reactions between the organic free radicals and Fe(III) [producing Fe(II)] or oxygen (producing superoxide) as described earlier for ferric oxalate complexes; and (ii) absorption of light by chromophoric components of DOM (CDOM) resulting in production of organic free radicals (L· – ) which can reduce Fe(III) directly or, in the presence of oxygen, initiate a one electron reduction producing the superoxide radical anion which, as shown above, may induce ferric iron reduction; i.e. L + hυ → L· – FeIII /FeIII L + L· – → Fe(II)/FeII L + Lox L· – + O2 → O2 · – + Lox Fe(III)/FeIII L + O2 · – → Fe(II)/FeII L + O2 A number of investigators have attempted to construct kinetic models of the light-mediated transformations of iron in natural waters. For example, Miller et al. [98] used a relatively simple scheme incorporating LMCT, reduction of inorganic Fe(III) by superoxide, ferrous iron oxidation and superoxide disproportionation to model the generation and decay of ferrous iron and

Role of Iron in Light-Induced Environmental Processes

283

hydrogen peroxide in samples of coastal seawater illuminated with artiﬁcial sunlight. In contrast, Rose and Waite [105] developed a model of the timedependent generation and decay of Fe(II), superoxide and hydrogen peroxide under transient light/dark conditions based on the reduction of ferric iron by photo-generated superoxide and oxidation of ferrous by both oxygen and peroxyl radicals with no consideration of LMCT. Additional work is needed to clarify the dominant transformation pathways though the relative importance of the various mechanisms is likely to depend on the type of organic matter present, the pH and iron and DOM concentrations. While most studies of the effect of light and DOM on changes in iron speciation have been performed with single, well-characterised samples of natural organic matter, Kaiser and Sulzberger [106] demonstrate signiﬁcant differences between hydrophilic and hydrophobic fractions and between high and low molecular weight fractions of natural organic matter extracted from the same riverine source. Steady state Fe(II) concentrations, normalised to dissolved organic carbon concentrations, achieved on photolysis with simulated sunlight were higher by a factor of 4–5 in high molecular weight and hydrophilic fractions than in low molecular weight and hydrophobic fractions. As mentioned in the introduction, DOM-mediated photo-transformations of iron are likely to inﬂuence the bioavailability of iron since ferrous iron is less strongly bound by organic matter. In addition, the iron and lightmediated oxidation of DOM may result in a decrease in absorptivity (“photobleaching”) of humic-rich waters and alteration in DOM characteristics [107] with varying effects on bioavailability [108–110]. Brinkmann et al. [111] found that the wavelengths of maximum bleaching and of maximum absorbed energy to be coincident and presented evidence that hydrophilic entities of the DOM were preferentially photodegraded while hydrophobic components were relatively unaffected or even formed. Iron was shown to be an important factor in DOM photobleaching while copper, another redox active metal, was found to inhibit the process. The mechanisms described earlier (i.e., LMCT following absorption of light by FeIII L chromophores and direct absorption of light by CDOM followed by oxidative degradation of organic intermediates) may account for the photobleaching effects. Direct attack by hydroxyl radicals has also been proposed to account for DOM transformations [112] though it has been shown that such a pathway is only likely to be important in waters exhibiting very high OH· production rates [113]. These highly reactive species could be produced via a Fenton reaction in which both Fe(II) and H2 O2 are generated by photolysis of DOM [114] though signiﬁcant production of hydroxyl radicals under anaerobic conditions (Fig. 14) indicates the presence of an additional dioxygen-independent production pathway [106, 115] as suggested in earlier studies by Zhou and Mopper [116, 117]. While deﬁnitive assignment has yet to be made, hydrogen abstraction from water by photoexcited triplet states of

284

T.D. Waite

Fig. 14 Wavelength dependence of the quantum yield for 10 mg/L Suwannee River Fulvic Acid under aerobic () and anaerobic (·) conditions (from [115])

quinone moieties of DOM may represent a possible pathway for direct production of hydroxyl radicals [118, 119]. Secondary effects of iron-DOM photo-transformations on other water constituents have been reported though additional study is required before deﬁnitive mechanistic assignments can be made. Thus, Cr(VI) reduction to Cr(III) has been found to be dependent on both the amount of iron and DOM present and the nature of the DOM [120]. Similar effects were observed with simple organic acids [74] (see Fig. 11) and it was suggested that LMCTmediated Fe(II) production (which in turn induces Cr(VI) reduction) may account for the observed phenomenon with natural organic matter though the reduction of Fe(III) by superoxide arising from direct absorption of light by the DOM was also considered a possibility. Degradation of organic contaminants (which are otherwise non-photoreactive) on photolysis of iron-rich DOM solutions is normally associated with the production of hydroxyl radicals [121] though whether this is related to production via a Fenton process or direct production by photolysis of quinone moieties of the DOM as discussed above is unclear.

Role of Iron in Light-Induced Environmental Processes

285

3 Mechanistic Aspects of Heterogeneous Photochemical Processes Involving Iron As indicated in Fig. 1, light may induce the reduction of Fe(III) present as particulate iron oxides and oxyhydroxides (Processes 13 and 14). The resulting Fe(II) species could induce the formation of a mixed valence oxide at the particle surface or, more likely, be released to solution resulting in dissolution of the solid phase. Whether the ferrous iron remains in solution or oxidises and reprecipitates as a ferric oxyhydroxide will be dependent particularly upon solution pH. If other species are adsorbed to the oxide surface, they may also undergo redox transformation as a result of reduction of the metal centre. Such photo-transformations are described in this section. 3.1 Photoreductive Dissolution of Iron Oxides While iron oxides such as hematite (α-Fe2 O3 ) possess semiconducting properties and may generate charge carriers (holes and electrons) at the oxide surfaces as a result of absorption of visible light corresponding to the d-d transition induced 2.3 eV band gap of the bulk oxide [122, 123], such a process appears to contribute little to redox transformations at the particle surface at least compared to those arising from LMCT transitions involving a speciﬁcally adsorbed ligand and a surface metal ion [124]. The light enhanced dissolution of iron oxides in the presence of a variety of adsorbing ligands has been attributed to LMCT processes occurring at the surface of iron oxides [125–127]. Faust and Hoffmann [128], Litter and Blesa [129] and Siffert and Sulzberger [130] investigated the wavelength dependence of the rate of light-induced reductive dissolution of Fe(III) oxides using hematitebisulﬁte, maghemite-EDTA and hematite-oxalate systems, respectively. The wavelength dependence of the rate of light-induced reductive dissolution of hematite in the presence of oxalate from the latter study is shown in Fig. 15. Siffert and Sulzberger [130] concluded that either an LMCT transition of the surface complex and/or a FeIII ← O-II charge-transfer transition of the hematite surface lattice are the oscillators that drive the redox reaction leading to reductive dissolution of the solid phase. A schematic of the various elementary steps involved in the surface photoredox reaction leading to hematite dissolution are shown below in Scheme 8. Rapid formation of a (hypothetical) bidentate, mononuclear surface complex is the important ﬁrst step followed by photoexcitation of this surface complex and subsequent electron transfer to Fe(III) with formation of an oxalate radical and Fe(II). The oxalate radical undergoes a rapid decarboxyla-

286

T.D. Waite

Fig. 15 Rate of the photochemical reductive dissolution of hematite (d[Fe(II)]/dt) in the presence of oxalate as a function of the wavelength at constant incident light intensity (Io = 1000 µE L–1 h–1 ). The hematite suspensions were deaerated; [Oxalate]o = 3.3 mM; pH = 3. In order to maintain the rate of thermal dissolution constant, [Fe2+ ] was maintained high and constant by adding Fe2+ at commencement of the experiment ([Fe2+ ] = 0.15 mM) (from [130])

tion reaction yielding CO2 and the CO2 · – radical, which is a strong reductant that can reduce a second surface Fe(III) in a thermal reaction [124]. Thus, two surface Fe(II) and two CO2 may theoretically be formed per absorbed photon. However, the quantum yield of this surface redox reaction is less than two presumably because of loss reactions such as thermal deactivation from the excited state. Ferrous iron released to solution will be complexed by oxalate and, according to Siffert and Sulzberger [130], may form a ternary surface complex which subsequently undergoes internal charge transfer as shown below.

Structure 6

The resulting Fe(III) oxalate complex is an excellent chromophore and will undergo homogeneous photolysis to produce additional Fe(II) species which, in turn, enhance dissolution of the iron oxide.

Role of Iron in Light-Induced Environmental Processes

287

Scheme 8

Scheme 9

An autocatalytic process is thus established in which the rate of dissolution is found to increase with time. Litter et al. [131] examined this autocatalytic process in more detail using maghemite (γ -Fe2 O3 )-EDTA systems and developed a kinetic model of the process which satisfactorily described the results obtained. Consistent with the proposed process, addition of soluble ferric iron increased dissolution rates.

288

T.D. Waite

Interestingly, Siffert and Sulzberger [130] found that while photodissolution of hematite occurred in anoxic systems, very little iron was released to solution in the presence of oxygen. Despite this, oxalate was oxidised 3–4 times more quickly in oxic compared to anoxic systems. The limited rate of hematite dissolution in the presence of oxygen was attributed to competition for surface-located photoproduced electrons by adsorbed oxygen— a process that was considered feasible because of the stability of the hematite lattice. The possibility that reprecipitation of iron oxide was occurring was discounted because of the low pH (pH = 3) of the study though the possibility of generation of more powerful oxidants (such as O2 · – and H2 O2 ) was not considered. The fact that photoreductive dissolution of other oxides such as lepidocrocite (γ -FeOOH) [125] and amorphous iron oxyhydroxide [132] has been observed in the presence of oxygen has been considered due to the lower crystallinity (higher solubility) of these more disordered phases. Indeed, the rate limiting step in the photoreductive dissolution of iron oxides is considered to be the rate of detachment of ferrous ions with this rate strongly dependent upon the degree of crystallinity of the iron oxide [124]. While the thermodynamic stability of the underlying oxide may control the rate of detachment of ferrous ions from the particle surface, the rate of dissolution on any particular oxide appears to be dependent upon a variety of factors including the concentration of the photoactive surface complex [133], the photoactivity of the Fe(III)-ligand complex, the backoxidation rate of photoproduced Fe(II) and the hydrolytic precipitation rate of reoxidised Fe(III). That this suite of processes were important factors in the light-induced dissolution of iron oxides was supported by Kuma et al. [132] in studies of the photoreduction of both dissolved and particulate Fe(III) in seawater (pH 8.1) over a range of natural sunlight intensities in the presence of a range of hydroxycarboxylic acids (HCAs). Under these conditions, the Fe(III) photoreduction abilities of the HCAs were in the order: glucaric-1,4lactone > tartaric > gluconic citric > glyceric = malic > glucuronic acids. That natural organic matter can induce the photoreductive dissolution of iron oxides has been known for some time [134] and, indeed, has been suggested to be one of the means by which iron is rendered available to organisms, particularly in marine systems [135–137]. Laboratory studies [100, 134] conﬁrm that humic and fulvic acids do enhance the photoreductive dissolution of iron oxides such as lepidocrocite under acidic conditions via lightassisted LMCT transitions in surface complexes in much the same manner as described for hydroxycarboxylic acids above. A decrease in dissolution rate is observed on increase in pH and has been attributed to both a decrease in the concentration of surface Fe(III)-fulvate complexes and reoxidation of surface Fe(II) by O2 · – and H2 O2 before detachment can take place. These factors are likely to limit the signiﬁcance of photoreductive dissolution of iron oxides in marine systems suggesting that light-induced formation of Fe(II) in surface seawaters occurs principally as a result of homogeneous processes.

Role of Iron in Light-Induced Environmental Processes

289

3.2 Photooxidation of Organic Compounds As is clear from the above discussion, reduction of surface-located Fe(III) (which may or may not lead to oxide dissolution) is associated in most instances with oxidation of the electron donor at the particle surface and many of the same factors that inﬂuence the rate of reductive dissolution will also affect the rate of donor oxidation. Leland and Bard [138] found that the rate constants of photooxidation of oxalate and sulﬁte varied by about two orders of magnitude with different Fe(III) oxides and concluded that “this appears to be due to differences in crystal and surface structure rather than to differences in surface area, hydrodynamic diameter or band gap”. A comprehensive set of studies into the photooxidation of halogenated acetic acids (HAA) by a variety of iron oxyhydroxides under conditions similar to those found in atmospheric water droplets has been undertaken by Pehkonen et al. [139]. Halogenated acetic acids are predicted to be one of the major degradation products of hydrochloroﬂuorocarbons and hydroﬂuorocarbons in the troposphere and their oxidation in Fe(III)-containing aerosols is of considerable interest. The order of reactivity toward monohaloacetate oxidation was am-FeOOH > γ -Fe2 O3 > γ -FeOOH ≥ α-Fe2O3 > α-FeOOH and was roughly in accord with the surface area of these oxides. The kinetics of photodissolution of the oxide and photooxidation of the HAA were consistent with photolysis of surface located O2– → Fe(III) charge transfer transitions with the rate limiting step (rls) being hydrogen-atom abstraction from the HAA by surface-bound hydroxyl radicals to produce haloacetate radicals which in turn yield the corresponding halide and glycolic acid (Fig. 16). Light-mediated redox processes other than reduction of surface-located Fe(III) and oxidation of the electron donor can occur in heterogeneous systems. Thus, Mazellier and Sulzberger [140] observed the degradation of diuron in irradiated goethite (α-FeOOH)/oxalate systems and attributed the degradation to production of hydroxyl radicals arising from the reaction of photoproduced Fe(II) and H2 O2 (an end product of the reduction of oxygen by oxalate radicals). Conﬁrmation that Fe(II)(aq) production through photoreductive dissolution of goethite is the rate-limiting step was provided by the similarity in dependence of the initial rate of Fe(II) formation and the rate of diuron disappearance on the concentration of adsorbed oxalate (Fig. 17). The change in rates of Fe(II) formation and diuron degradation at a threshold value of the oxalate surface concentration was hypothesised to be due to a transition from a bidentate, binuclear Fe(III) oxalate surface complex at low surface coverage to a monodentate, binuclear Fe(III) oxalate surface complex at high surface coverage [140] though characterisation of the nature of the surface complex is required to verify this hypothesis. The possibility of using particulate iron oxides to induce contaminant degradation has resulted in a variety of studies into the possibility of at-

290

T.D. Waite

Fig. 16 Mechanism for the photoreduction of iron oxides and the concomitant photooxidation of haloacetic acids (such as bromoacetate) proposed by Pehkonen et al. [139]

taching the iron oxides to inert support materials in order to (i) aid the separation of the iron oxide from the solution, and (ii) improve the efﬁciency by maintaining the particles in dispersed (albeit attached) form. Various inert substrates have been examined including Naﬁon ion exchange membranes [141–143], polyethylene block-copolymer ﬁlms containing negative anhydride groups [144], alumino-borosilicate ﬁbres [145] and laponite RD clays [146]. Early trials of these technologies are promising and enable oper-

Role of Iron in Light-Induced Environmental Processes

291

Fig. 17 Initial rates of Fe(II)(aq) formation and rate of diuron disappearance as a function of the surface concentration of oxalate in irradiated, aerated suspensions containing initially 80 mg L–1 goethite and 10 µM diuron at pH 4 (after [140])

ation at neutral to basic pH without the need for addition of a complexant to maintain the iron in solution.

4 Field Observations The issues addressed above have been investigated in laboratory studies, often under conditions simulating natural or treatment systems. In this section, observations of light-mediated transformations involving iron in marine and freshwaters are reviewed. While the observations of increased oxygen consumption in iron and humic rich waters in Florida during the day compared to the night were attributed to light-induced production of Fe(II) by Miles and Brezonik [94], the ﬁrst distinct evidence of a diurnal cycle in [Fe(II)] was provided by McKnight et al. [147] in studies of the changes in Fe(II) concentration in a small mountain stream in Colorado that received acid mine drainage. As can be seen from Fig. 18, maximum Fe(II) concentrations coincide with times of peak light intensity with more detailed analysis revealing that daytime production of ferrous iron by photoreduction was almost four times as great as nighttime oxidation of ferrous iron. Diurnal cycling of Fe(II) was even more distinct in the same waters for which the pH was increased from 4.0 to 6.5 as a result of the increased rate of Fe(II) oxidation [148] and was dramatic in the

292

T.D. Waite

Fig. 18 Changes in ferrous iron concentration, light intensity and pH in St Kevin Gulch on 19 and 20 August 1986 (from [147])

these same waters to which two organic acids (phthalic and aspartic acids) had been added [149]. Hydrogen peroxide was also found to follow a diurnal trend though at the concentrations generated was not considered a major oxidant of Fe(II) [149]. Results of investigations of iron speciation in Lake Cristallina, an acidic lake (pH 5.2) in the southern Alps of Switzerland, are strongly indicative of a diurnal cycle of Fe(II) production and have been interpreted in terms of photochemical reductive dissolution of Fe(III) oxyhydroxides and subsequent thermal oxidation of Fe(II) [124]. Distinct diurnal trends in [Fe(II)] have also been reported in Lakes Greifensee and Melchsee in Switzerland [150]. These lakes are of higher pH (8.0–8.5) than Lake Cristallina and Fe(II) concentrations are much lower but the diurnal cycling is clear with Fe(II) concentrations of 0.1–0.2 nM at night and up to 0.9 nM near the surface during the day. While a number of deckboard incubations of seawater samples show an increase in Fe(II) on irradiation [98, 136, 151], only a few studies have examined the change in iron speciation in situ as a function of time of day. Waite and Szymczak [152] measured the concentration of iron in waters overlying a coral reef on One Tree Island on Australia’s Great Barrier Reef and observed

Role of Iron in Light-Induced Environmental Processes

293

an increase in the concentration of iron that passed through a 0.22 µm membrane ﬁlter from an overnight low of around 50 nM to around 150 nM by noon followed by a subsequent decrease to the night-time level. Hydrogen peroxide concentrations were observed to follow an almost identical temporal trend rising from a night-time value of 30 nM to a noon value of 130 nM [152]. Much lower concentrations of ﬁlterable iron were observed in waters of the

Fig. 19 Results of a die1 study of surface-water ferrozine-reactive iron concentrations while following a body of water in the Gulf of Carpentaria, north of Australia. Plot A shows the light intensity measured over the day with the ﬁrst point being at sunrise and the last at sunset. Plot B depicts ﬁlterable (< 0.4 µm) ferrozine-reactive iron concentrations for the same local time period. Plot C shows particulate () and total (•) ferrozine-reactive iron concentrations. Error bars indicate 98% conﬁdence limits (from [153])

294

T.D. Waite

Timor Sea between Australia and Indonesia but results of a Lagrangian study in which the same parcel of water was tracked and samples taken over one day revealed a distinct diurnal cycle in < 0.4 µm ferrozine-reactive iron concentrations (Fig. 19) [153]. While a large diel variation in the concentration of ferrozine-reactive iron is apparent, this represented only a small fraction (< 1%) of the total iron pool in these waters.

5 Conclusions In this chapter we have overviewed the key photochemical processes in which iron plays a role in both natural and treatment systems. The discussion has been grouped around commonality of mechanisms with attention given to both homogeneous and heterogeneous systems. Insights from simple model systems have been critical to understanding both the complex interrelationships between light-mediated transformations of iron and oxygen species and the nature of processes involving complex natural organic matter. While the volume of work in this area is large and considerable advance in understanding has occurred, gaps in knowledge remain. Selected areas requiring further investigation are described (in no particular order) below. • While the key photochemical transformations involving iron carboxylate complexes have been extensively studied, scope remains for further insight into the impact of speciation change during the photolysis process, particularly with regard to the change in nature and concentration of the primary chromophore as reactant concentrations change. • Photolysis of iron carboxylates has been used to initiate the degradation of contaminant species (often via photo-Fenton processes) but the process results in degradation of the carboxylate ligand. Alternate ligands which are more resistant to oxidation (either via LMCT processes or hydroxyl radical attack) would appear necessary if such approaches are to be adopted in practice. • The key processes operating on photolysis of natural organic matter remain unclear with an apparent interplay between quinone and hydroxycarboxylate properties. The implications of these functionalities to iron transformation in natural systems require additional attention. • The relative importance of homogeneous and heterogeneous photoprocesses involving iron in natural waters is far from clear and requires additional ﬁeld investigation. • The possible use of inert supports for iron oxides as a means of maintaining photoactivity, even at higher pH is intriguing and represents a fruitful area for further development.

Role of Iron in Light-Induced Environmental Processes

295

References 1. Boule P, Bolte M, Richard C (1999) Phototransformations induced in aquatic media by NO3 – /NO2 – , FeIII and humic substances. In: Boule P (ed) The Handbook of Environmental Chemistry, vol 2, Part L Environmental Photochemistry. Springer, Berlin Heidelberg New York, p 181 2. Turner DR, Hunter KA (2001) The biogeochemistry of iron in seawater. In: IUPAC Series on Analytical and Physical Chemistry of Environmental Systems, vol 7. Wiley, Chichester 3. Wehrli B (1990) In: Stumm W (ed) Aquatic Chemical Kinetics. Wiley-Interscience, New York, p 311 4. Rose AL, Waite TD (2003) Environ Sci Technol 37:3897 5. Rose AL, Waite TD (2002) Environ Sci Technol 36:433 6. Rose AL, Waite TD (2003) Environ Sci Technol 37:4877 7. Albrecht-Gary A-M, Crumbliss AL (1998) In: Sigel A, Sigel H (eds) Metal Ions in Biological Systems, vol 35: Iron Transport and Storage in Microorganisms, Plants and Animals. Marcel Dekker, New York, p 239 8. David F, David PG (1976) J Phys Chem 80:579 9. Faust BC, Hoigné J (1990) J Atmos Environ 24A:79 10. Benkelberg HJ, Warneck P (1995) J Phys Chem 99:5214 11. Kawaguchi H, Inagaki A (1994) Chemosphere 28:57 12. Mazellier P, Bolte M (2001) Chemosphere 42:361 13. Mazellier P, Sarakha M, Bolte M (1999) New J Chem 23:133 14. Mazellier P, Mailhot G, Bolte M (1997) New J Chem 21:389 15. Mazellier P, Jirkovský J, Bolte M (1997) Pestic Sci 49:259 16. Catastini C, Sarakha M, Mailhot G, Bolte M (2002) Sci Tot Env 298:219 17. Meˇsánková H, Mailhot G, Krysa J, Jirkovský J, Bolte M (2004) Wat Sci Tech 49:165 18. Bajt O, Mailhot G, Bolte M (2001) Appl Cat B Environ 33:239 19. Brand N, Mailhot G, Bolte M (1998) Environ Sci Technol 32:2715 20. Andreozzi R, Marotta R (2004) Wat Res 38:1225 21. Anbar M, Neta P (1967) Int J Appl Radiat Isot 18:493 22. Stafford U, Gray KA, Kamat PV (1994) J Phys Chem 98:6343 23. Mailhot G, Astruc M, Bolte M (1999) Appl Organometal Chem 13:53 24. Morel FMM, Hering JG (1993) Principles and applications of aquatic chemistry. Wiley-Interscience, New York 25. Waite TD (2001) Thermodynamics of the iron system in seawater. In: Turner DR, Hunter KA (eds) The Biogeochemistry of Iron in Seawater, chap 7. Wiley, Chichester, p 291 26. Balzani V, Carassiti V (1970) Photochemistry of coordination compounds. Academic Press, New York 27. Sima J, Makanova J (1997) Coord Chem Rev 160:161 28. Zuo Y, Hoigné J (1994) Atmos Environ 28:1231 29. Safarzadeh-Amiri A, Bolton JR, Cater SR (1996) Solar Energy 56:439 30. Safarzadeh-Amiri A, Bolton JR, Cater SR (1997) Water Res 31:787 31. Parker CA (1953) Proc R Soc Lond 220A:104 32. Hatchard CG, Parker CA (1956) Proc R Soc Lond 235A:518 33. Parker CA (1954) Trans Faraday Soc 50:1213 34. Parker CA, Hatchard CG (1959) J Phys Chem 63:22 35. Cooper GD, DeGraff BA (1971) J Phys Chem 75:2897 36. Cooper GD, DeGraff BA (1972) J Phys Chem 76:2618

296

T.D. Waite

37. Jamieson RA, Perone SP (1972) J Phys Chem 76:830 38. Patterson JIH, Perone SP (1973) J Phys Chem 77:2437 39. Mulazzani QG, D’Angelantonio M, Venturi M, Hoffman MZ, Rodgers MAJ (1986) J Phys Chem 90:5347 40. Neta P, Huie RE, Ross AB (1988) J Phys Chem Ref Data 17:1027 41. Huston PL, Pignatello JJ (1996) Environ Sci Technol 30:3457 42. Buxton GV, Sellers RM (1973) J Chem Soc Faraday Trans 1 69:555 43. Rowan NS, Hoffman MZ, Milburn RM (1974) J Am Chem Soc 96:6060 44. Abrahamson HB, Rezvani AB, Brushmiller JG (1994) Inorg Chim Acta 226:117 45. Baxendale JH, Bridge NK (1955) J Phys Chem 59:783 46. Allmand AJ, Webb WW (1929) J Chem Soc 1929:1518 47. Allmand AJ, Webb WW (1929) J Chem Soc 1929:1531 48. Livingston R (1940) J Phys Chem 44:601 49. Aplin R (2002) PhD Thesis, The University of New South Wales, Sydney, Australia 50. Vincze L, Papp S (1987) J Photochem 36:289 51. Faust BC, Zepp RG (1993) Environ Sci Technol 27:2517 52. Park JSB, Wood PM, Davies MJ, Gilbert BC, Whitwood AC (1997) Free Radical Res 27:447 53. Balmer ME, Sulzberger B (1999) Environ Sci Technol 33:2418 54. Glikman TS, Kalibabchuk VA, Sosnovskaya VP (1965) J Gen Chem USSR 35:1533 55. Kuma K, Nakabayashi S, Matsunaga K (1995) Water Res 29:1559 56. Alder AC, Siegrist H, Guser W, Giger W (1990) Water Res 24:733 57. Kari FG, Hilger S, Canonica S (1995) Environ Sci Technol 29:1008 58. Carey JH, Langford CH (1973) Can J Chem 51:3665 59. Schugar HJ, Hubbard AT, Anson FC, Gray HB (1969) J Am Chem Soc 90:71 60. Natarajan P, Endicott JF (1973) J Phys Chem 77:2049 61. Carey JH, Langford CH (1975) Can J Chem 53:2436 62. Andrianirinaharivelo SL, Mailhot G, Bolte M (1995) Transition Metal Chem 18:37 63. Mailhot G, Bordes AL, Bolte M (1995) Chemosphere 30:1729 64. Sedlak DL, Hoigné J (1993) Atmos Environ 27A:2173 65. Zepp RG, Faust BC, Hoigné J (1992) Environ Sci Technol 26:313 66. Abida O, Emilio C, Quici N, Gettar R, Litter M, Mailhot G, Bolte M (2004) Wat Sci Technol 49:123 67. Maldotti A, Andreotti L, Molinari A, Carassiti V (1999) J Biol Inorg Chem 4:154 68. Hug SJ, Canonica L, Wegelin M, Gechter D, Von Gunten U (2001) Environ Sci Technol 35:2114 69. Bossmann SH, Oliveros E, Gob S, Siegwart S, Dahlen EP, Payawan L, Straub M, Worner M, Braun AM (1998) J Phys Chem A 102:5542 70. Kremer ML (1999) Phys Chem Chem Phys 1:3595 71. Bossmann SH, Oliveros E, Kantor M, Niebler S, Bonﬁll A, Shahln N, Worner M, Braun AM (2004) Wat Sci Technol 49:75 72. Pignatello JJ, Liu D, Huston P (1999) Environ Sci Technol 33:1832 73. Kocar BD, Inskeep WP (2003) Environ Sci Technol 37:1581 74. Hug SJ, Laubscher HU, James BR (1997) Environ Sci Technol 31:160 75. Barbeau K, Rue EL, Bruland KW, Butler A (2001) Nature 413:409 76. Barbeau K, Zhang G, Live DH, Butler A (2002) J Am Chem Soc 124:378 77. Barbeau K, Rue EL, Trick CG, Bruland KW, Butler A (2003) Limnol Oceanogr 48:1069 78. Stuermer DH, Harvey GR (1978) Mar Chem 6:55 79. Wilson MA, Barron PF, Gillam AH (1981) Geochim Cosmochim Acta 45:1743 80. Ritchie JD, Perdue EM (2003) Geochim Cosmochim Acta 67:85

Role of Iron in Light-Induced Environmental Processes

297

81. Scott DT, McKnight DM, Blunt-Harris EL, Kolesar SE, Lovley DR (1998) Environ Sci Technol 32:2984 82. Kurimura Y, Ochiai R, Matsuura N (1968) Bull Chem Soc Japan 41:223 83. Jobin R, Ghosh MM (1972) J AWWA 64:590 84. Harris DC, Aisen P (1973) Biochimica et Biophysica Acta 329:156 85. Theis TL, Singer PC (1974) Environ Sci Technol 8:569 86. Liang L, McNabb JA, Paulk JM, Gu B, McCarthy JF (1993) Environ Sci Technol 27:1864 87. Santana-Casiano JM, González-Dávila M, Rodríguez MJ, Millero FJ (2000) Mar Chem 70:211 88. Emmenegger L, King DW, Sigg L, Sulzberger B (1998) Environ Sci Technol 32:2990 89. Voelker BM, Sulzberger B (1996) Environ Sci Technol 30:1106 90. Straub KL, Benz M, Schink B (2001) FEMS Microbiology Ecology 34:181 91. King DW, Lounsbury HA, Millero FJ (1995) Environ Sci Technol 29:818 92. Rush JD, Bielski BHJ (1985) J Phys Chem 89:5062 93. Rose AL, Waite TD (2003) Mar Chem 84:85 94. Miles CJ, Brezonik PL (1981) Environ Sci Technol 15:1089 95. Francko DA, Heath RT (1982) Limnol Oceanogr 27:564 96. Collienne RH (1983) Limnol Oceanogr 28:83 97. Waite TD, Morel FMM (1984) Anal Chim Acta 162:263 98. Miller WL, King DW, Lin J, Kester DR (1995) Mar Chem 50:63 99. Voelker BM, Sedlak DL (1995) Mar Chem 50:93 100. Voelker BM, Morel FMM, Sulzberger B (1997) Environ Sci Technol 31:1004 101. Gao H, Zepp RG (1998) Environ Sci Technol 32:2940 102. Fukushima M, Tatsumi K (1999) Colloids and Surfaces A 155:249 103. Fukushima M, Tatsumi K, Nagao S (2001) Environ Sci Technol 35:3683 104. Powell RT, Wilson-Finelli A (2003) Aquat Sci 65:367 105. Rose AL, Waite TD (2003) Aquat Sci 65:375 106. Kaiser E, Sulzberger B (2004) Limnol Oceanogr 49:540 107. White EM, Vaughan PP, Zepp RG (2003) Aquat Sci 65:402 108. Wetzel RG, Hatcher PG, Bianchi TS (1995) Limnol Oceanogr 40:1369 109. Benner R, Biddanda B (1998) Limnol Oceanogr 43:1373 110. Obernosterer I, Reitner B, Herndl GJ (1999) Limnol Oceanogr 44:1645 111. Brinkmann T, Sartorius D, Frimmel FH (2003) Aquat Sci 65:415 112. Westerhoff P, Aiken G, Amy G, Debroux J (1999) Environ Sci Technol 33:2265 113. Goldstone JV, Pullin MJ, Bertilsson S, Voelker BM (2002) Environ Sci Technol 36:364 114. Southworth B, Voelker BM (2003) Environ Sci Technol 37:1130 115. Vaughan PP, Blough NV (1998) Environ Sci Technol 32:2947 116. Mopper K, Zhou X (1990) Science 250:661 117. Zhou X, Mopper K (1990) Mar Chem 30:71 118. Ononye AI, McIntosh AR, Bolton JR (1986) J Phys Chem 90:6266 119. Alegria AE, Ferrer A, Santiago G, Sepulveda E, Flores W (1999) J Photochem Photobiol A:Chemistry 127:57 120. Gaberell M, Chin YP, Hug SJ, Sulzberger B (2003) Environ Sci Technol 37:4403 121. Andreozzi R, Rafaele M, Paxeus N (2003) Chemosphere 50:1319 122. Marusak LA, Messier R, White WB (1980) J Phys Chem Solids 41:981 123. Sherman DM, Waite TD (1985) Am Mineral 70:1262 124. Sulzberger B (1990) In: Stumm W (ed) Photoredox reactions at hydrous metal oxide surfaces: a surface coordination chemistry approach. Wiley, New York, p 401 125. Waite TD, Morel FMM (1984) J Colloid Interface Sci 102:121

298

T.D. Waite

126. 127. 128. 129. 130. 131.

Waite TD, Torikov A, Smith JD (1986) J Colloid Interface Sci 112:412 Waite TD, Torikov A (1987) J Colloid Interface Sci 119:228 Faust BC, Hoffmann MR (1986) Environ Sci Technol 20:943 Litter MI, Blesa MA (1988) J Colloid Interface Sci 125:679 Siffert C, Sulzberger B (1991) Langmuir 7:1627 Litter MI, Baumgartner EC, Urrutia GA, Blesa MA (1991) Environ Sci Technol 25:1907 Kuma K, Nakabayashi S, Matsunaga K (1995) Wat Res 29:1559 Stumm W, Sulzberger B (1992) Geochim Cosmochim Acta 56:3233 Waite TD, Morel FMM (1984) Environ Sci Technol 18:860 Finden DAS, Tipping E, Jaworski GHM, Reynolds CS (1984) Nature 309:783 Wells ML, Mayer LM (1991) Deep-Sea Res 38:1379 Wells ML, Mayer LM, Donard OFX, De Souza Sierra MM, Ackelson SG (1991) Nature 353:248 Leland JK, Bard AJ (1987) J Phys Chem 91:5076 Pehkonen SO, Siefert RL, Hoffmann MR (1995) Environ Sci Technol 29:1215 Mazellier P, Sulzberger B (2001) Environ Sci Technol 35:3314 Fernandez J, Bandara J, Lopez A, Albers P, Kiwi J (1998) Chem Commun 1998:1493 Fernandez J, Bandara J, Lopez A, Buffar Ph, Kiwi J (1999) Langmuir 15:185 Fernandez J, Djananjeyan MR, Kiwi J, Senuma Y, Hilborn J (2000) J Phys Chem B 104:5298 Dhananjeyan MR, Mielczarski E, Thampi KR, Buffat Ph, Bensimon M, Kulik A, Mielczarski J, Kiwi J (2001) J Phys Chem B 105:12 046 Bozzi A, Yuranova T, Mielczarski J, Lopez A, Kiwi J (2002) Chem Commun 2002:2202 Yue PL, Feng JY, Hu X (2004) Wat Sci Technol 49:85 McKnight DM, Kimball BA, Bencala KE (1988) Science 240:637 McKnight DM, Kimball BA, Runkel RL (2001) Hydrol Process 15:1979 Scott DT, Runkel RL, McKnight DM, Voelker BM, Kimball BA, Carraway ER (2003) Water Resour Res 39:1308 Emmenegger L, Schonenberger R, Sigg L, Sulzberger B (2001) Limnol Oceanogr 46:49 Johnson KS, Coale KH, Elrod VA, Tindale NW (1994) Mar Chem 46:319 Waite TD, Szymczak R (1994) Photoredox transformations of iron and manganese in marine systems: review of recent ﬁeld investigations. In: Helz GR, Zepp RG, Crosby DG (eds) Aquatic and surface photochemistry. Lewis, Boca Raton, p 39 Waite TD, Szymczak R, Espey QI, Furnas MJ (1995) Mar Chem 50:79

132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152.

153.

Hdb Env Chem Vol. 2, Part M (2005): 299–323 DOI 10.1007/b138187 © Springer-Verlag Berlin Heidelberg 2005 Published online: 16 September 2005

Aquatic Phototransformation of Organic Contaminants Induced by Coloured Dissolved Natural Organic Matter Claire Richard1 (u) · Silvio Canonica2 1 Laboratoire

de Photochimie Moléculaire et Macromoléculaire, UMR 6505 CNRS-Université Blaise Pascal, 63177 Aubière Cedex, France [email protected]

2 Swiss

Federal Institute for Environmental Science and Technology (EAWAG), Ueberlandstrasse 133, 8600 Dübendorf, Switzerland [email protected]

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

300

2

Natural Organic Matter: Deﬁnitions and Characterization . . . . . . . . .

300

3 3.1 3.2 3.3 3.4 3.5 3.6 3.7

Reactive Species Produced upon Irradiation of CDOM Hydrated Electron . . . . . . . . . . . . . . . . . . . . . Hydroperoxyl Radical and Superoxide Radical Anion . Excited Triplet States of the CDOM . . . . . . . . . . . Singlet Oxygen . . . . . . . . . . . . . . . . . . . . . . . Hydroxyl Radical . . . . . . . . . . . . . . . . . . . . . Carbonate Radical . . . . . . . . . . . . . . . . . . . . . DOM-Derived Radicals . . . . . . . . . . . . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

302 304 304 305 306 306 307 307

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

308 308 310 311 315

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

320

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

321

4

Phototransformation of Organic Contaminants in the Presence of CDOM or Humic Substances 4.1 Inhibition or Absence of Effect . . . . . . . . . . 4.2 Enhancing Effect . . . . . . . . . . . . . . . . . 4.2.1 Sulphur-Containing Compounds . . . . . . . . . 4.2.2 Other Compounds . . . . . . . . . . . . . . . . . 5

. . . . .

. . . . .

. . . . .

. . . . .

Abstract This review will ﬁrst give a brief description of coloured dissolved organic matter (CDOM) from the aquatic environment or extracted from soil. Data on chemical structure and spectroscopic properties are presented and the major photochemical processes involving CDOM as well as the main reactive species produced upon their irradiation are described. The fourth part of this chapter is an overview of studies related to the phototransformation of organic contaminants in the presence of CDOM, either under ﬁeld or laboratory conditions. CDOM was found to be able to photo-induce the transformation of numerous organic pollutants. Hydroxyl radicals play a minor role, probably because they are efﬁciently scavenged by the CDOM itself. Examples of the oxidation of pollutants by singlet oxygen and by reactive triplet states are given, as are examples of photoreduction.

300

C. Richard · S. Canonica

Keywords Dissolved organic matter · Humic substances · Organic pollutants · Photo-induced transformation · Reactive species

1 Introduction Photochemical reactions provide degradation pathways of organic contaminants present in the aquatic environment. Two basic types of reactions may occur. Contaminants can undergo photolysis via absorption of solar radiation, which is referred to as direct photolysis. The importance of this phenomenon depends on the capacity of the contaminant to absorb photons and to undergo chemical changes after light absorption. Alternatively, the degradation may be mediated by chromophoric components of environmental waters, and these photo-induced or photosensitized processes are often called indirect phototransformations. The efﬁciency of such reactions depends on the number of reactive species produced by excitation of the chromophores as well as on the ability of these species to react with the contaminant. Indirect phototransformations may be the dominant degradation processes in the case of contaminants that absorb terrestrial sunlight poorly or that are stable against direct photolysis. Various review articles [1–14] and textbook chapters [15, 16] have dealt with the description and assessment of both direct and indirect phototransformations in the aquatic environment in the last 25 years. A major aspect of these studies was identiﬁcation of the various reactive transients that are formed in sunlit natural waters, and to estimate or calculate their concentrations and their second-order reaction rate constants with contaminants in order to assess indirect phototransformation rates under environmental conditions. In this study, we maintain this reactive transient approach, outlining the main pathways of formation after absorption of light by natural organic matter. We also present a survey of studies addressing organic-matterinduced phototransformations of organic contaminants conducted during the last decade. This survey shows the importance of indirect phototransformations in the aquatic environment and illustrates how the transients responsible for such reactions may be identiﬁed.

2 Natural Organic Matter: Definitions and Characterization Natural organic matter comprises any organic chemical substance that is not part of living organisms, but is derived from them either as a product of their metabolism or of their decomposition. Since we are particularly inter-

Phototransformation of Organic Contaminants by CDOM

301

ested in organic matter that is contained in natural surface waters, this section is focused on dissolved organic matter (DOM), which is much more abundant than the other fraction, the particulate organic matter, not treated here. Despite the chemical complexity of DOM (see [17] for a recent review), it has been possible to identify various components according to their size and polarity, acid/base properties and compound classes. DOM contains both autochthonous material, which is produced in the natural water, as well as allochthonous material, which is derived from soil. Two major fractions of the DOM, the humic and fulvic acids, are classiﬁed as humic substances and are typical of the organic matter present in soils [18]. Structural units in humic substances can be identiﬁed using various methods [19, 20] and comprise the most common organic functional groups, such as carboxylic, phenolic, ketone, ether and ester moieties. Aromatic moieties may be assessed by 13 C nuclear magnetic resonance (NMR) spectroscopy, and the obtained aromaticity index has been shown to be positively correlated with the speciﬁc absorption coefﬁcient at the wavelength of 280 nm [21]. Of special importance in redox chemistry are quinone moieties, which can be detected using electron paramagnetic resonance (EPR) spectroscopy [22] and are thought to play a key role as electron-transfer mediators in the reduction of organic contaminants under anoxic conditions [23]. DOM, and in particular dissolved humic substances give, depending on their concentration, a yellowish or brown colour to the natural waters. To underline the fact that the light-absorbing components, and, thus, those that are photochemically active, are also responsible for the colour, the term coloured (or chromophoric) dissolved organic matter (CDOM) [24, 25] has been proposed for use in environmental photochemistry literature, and will be used here. UV-visible absorption spectra of CDOM, which are crucial in connection with its photochemistry since they enable the determination of the rate of light absorption by CDOM and the attenuation of light in a water body [26], exhibit a nearly exponential shape and a shoulder in the 250–300 nm region. Above about 290 nm, the spectra can be ﬁt reasonably to a single exponential function of the form: a(λ) = a(λ0 ) e–S(λ–λ0 )

(1)

where a is the absorption coefﬁcient (usually decadic for chemists but Napierian for limnologists or physicists!), λ the wavelength, λ0 a reference wavelength and S is termed spectral slope [24]. It was observed that CDOM of terrestrial origin, rich in humic acids, is more aromatic than oceanic CDOM and presents a smaller spectral slope [25]. Little is known of the relationship between ﬂuorescence spectra (and quantum yields) of CDOM and its photochemical activity, but such spectra, especially in the form of excitation– emission matrices (EEM), can be used to assess the presence of some groups of ﬂuorophores [27]. Characteristic peaks in EEM spectra may be used to as-

302

C. Richard · S. Canonica

sess the presence of humic-like [17] or protein-like [28] compounds. Until now, it has not been possible to relate the occurrence of photo-induced transformation of organic contaminants to the presence of speciﬁc chromophores or structural units in the CDOM.

3 Reactive Species Produced upon Irradiation of CDOM Some major photochemical processes involving CDOM as the light-absorbing component (photosensitizer) are summarized in Scheme 1. Chromophores of the DOM reach their excited singlet manifold upon absorption of a photon. Subsequently, either charge separation with formation of a hydrated electron and a CDOM radical cation or vibronic relaxation to the lowest excited singlet state (vibrational excitation according to a Maxwell-Boltzmann distribution) takes place. Molecular oxygen, usually present at an appreciable concentration in the photic zone of surface waters, scavenges most of the hydrated electron producing the superoxide anion radical. The ﬁrst excited singlet state is short-lived enough (lifetime in the order of nanoseconds) to be insigniﬁcant for any photochemical reaction with dissolved species. Two deactivation pathways, ﬂuorescence and internal con-

Scheme 1 Major photochemical processes of coloured dissolved natural organic matter (CDOM) that may lead to the transformation of organic contaminants. 1 X∗ Excited singlet state of the species X, 3 X∗ excited triplet state of the species X, P an organic contaminant , POX oxidation product(s) formed upon reaction of P with singlet oxygen, full arrow heads chemical reactions, open arrow heads energy transfer/loss processes, continuous arrows radiative processes, dashed arrows radiationless processes

Phototransformation of Organic Contaminants by CDOM

303

version, lead to the ground state of the chromophore, while intersystem crossing generates the lowest excited triplet state of the chromophore, which may undergo several types of reactions with dissolved species owing to its longer lifetime (in the order of microseconds). Two intramolecular processes, phosphorescence and intersystem crossing (which may be enhanced by the presence of certain molecular moieties in the DOM molecule containing the chromophore) contribute to the deactivation of the excited triplet state to the singlet ground state of the chromophore. In aerated water under common environmental conditions, quenching by molecular oxygen is a very efﬁcient deactivation process, which limits the lifetime of the excited triplet state to about 2 µs and leads either to both deactivated reaction partners or to the formation of singlet (molecular) oxygen, 1 O . The latter is a very well studied oxidant that, despite its rapid deacti2 vation in water solution (mostly due to intersystem crossing to the triplet ground state) is able to oxidize a variety of organic compounds. There are two other modes of interaction of excited triplet states of the CDOM with organic contaminants. One is triplet–triplet energy transfer, a radiationless process that leads simultaneously to deactivation of the CDOM triplet and formation of the excited triplet state of the organic contaminant, which may then undergo deactivation (by the same processes as described above for 3 CDOM∗ ) or photochemical transformation. The other is a photoredox reaction, leading to reduced CDOM (with one excess electron), CDOM–· , and oxidized organic contaminant (lacking one electron), P+· . The representation as a one-electron transfer reaction is purely formal; hydrogen atom transfer reactions from the contaminant to the CDOM are also included. Although the opposite redox reaction, i.e. reduction of the organic contaminant and oxidation of 3 CDOM∗ , is also possible, it is not considered in the scheme because it is unlikely to occur in aerated solution (see the comments given in the Hydrated Electron subsection). The CDOM is regenerated by reaction of CDOM–· with oxygen under the formation of a superoxide anion. P+· can react in various ways, which should eventually lead to even-electron species as the oxidation products, but can also undergo reduction (e.g., by DOM), in which case the parent compound is formed again. Other reactive species are or may be formed upon irradiation of CDOM, but their production rates were determined (or are considered) to be much lower than for the species represented in Scheme 1. Nevertheless, these species can be important for the transformation of organic contaminants that are not susceptible to reaction with the main reactive species. They comprise the hydroxyl radical, the carbonate radical and all the radicals derived from the DOM (carbon-centred, oxyl, peroxyl radicals). The following subsections describe in more detail the role that each reactive species plays in the transformation of aquatic organic contaminants. Hydrogen peroxide, although considered a reactive oxygen species (ROS) and an important player in aquatic photochemistry [7], is neglected because there is no evidence that

304

C. Richard · S. Canonica

its direct reaction with organic contaminants is of relevance for their aquatic transformation. 3.1 Hydrated Electron This species could be observed in oxygen-free aqueous solutions of humic substances after laser excitation and transient absorption spectroscopy in the visible range [29–31]. The hydrated electron, exhibiting a broad absorption band with maximum at about 700 nm, appears to be formed immediately after a picosecond pulse excitation [29], which means that it originates, most probably, from excited singlet states of the CDOM before vibrational relaxation has taken place. As far as the formation quantum yield is concerned, there is a large discrepancy (not yet resolved) between values determined by laser excitation using intense picosecond [29, 32] or nanosecond pulses [31, 33] and the values determined under low intensity photostationary irradiation (sunlit water conditions) using reactive probe compounds [33]. Possible explanations are non-linear effects that may occur using pulsed high intensity radiation or fast (re)combination of DOM and hydrated electron, which would then be unavailable for reaction with dissolved compounds. The importance of the solvated electron for the transformation of organic contaminants in aerated water has been estimated to be very minor [7, 8, 33]. Only rare cases of photo-induced reduction in surface waters are known (see Sect. 4). 3.2 Hydroperoxyl Radical and Superoxide Radical Anion These radicals constitute an acid–base pair (pKa = 4.8 [34]) and, although the superoxide radical anion (O–2· ) is the dominating species in most surface waters, its conjugated acid cannot be neglected because of its much higher standard reduction potential: HO2 · + e– → HO2 – E◦ = 0.75 V [35] (2) – E◦ < 0.21 V (3) O2 –· + e → O2 2– Superoxide can also react as a reductant: O2 + e– → O2 –· E◦ = – 0.16 V[35]

(4)

As a consequence, HO2 · /O2 –· are involved as intermediate species in aquatic redox cycles of metals such as copper or iron [36]. There are many different possible pathways of formation of superoxide [7, 37], only two of which 1 Calculated from the reduction potential and pK of the hydroperoxyl radical and assuming a pKa (HO2 – ) > 14.

Phototransformation of Organic Contaminants by CDOM

305

are listed in Scheme 1. It has been shown [37, 38] that disproportionation of HO2 · /O2 –· according to the equation: HO2 · + HO2 · /O2 –·

(catalyst)

→ H2 O2 /HO2 – + O2

(5)

is almost quantitative and leads to the formation of hydrogen peroxide. Therefore, the formation rate of hydrogen peroxide (which can be analysed more easily than HO2 · /O2 –· ) can be assumed to be half of the formation rate of HO2 · /O2 –· . The disproportionation rate, recently determined by Voelker and co-workers [38, 39], is controlled by catalytic moieties of the DOM (quinones, copper complexes) and exceeds the rate for the uncatalysed reaction [34] by two to three orders of magnitude in seawater. The concentration of HO2 · /O2 –· in sunlit surface waters may be considered to be too low to induce appreciable transformations of organic contaminants. 3.3 Excited Triplet States of the CDOM Broad-band transient absorption signals with lifetimes in the microsecond to millisecond range were observed after nanosecond ﬂash photolysis of aqueous humic substances and tentatively assigned to excited triplet states of the CDOM [7, 29]. There are several additional pieces of evidence that these excited states are formed in sunlit surface water and may play a major role in the transformation of organic contaminants. The formation of singlet oxygen (see Singlet Oxygen section) is one such indication. Zepp and coworkers showed that, using 1,3-cyclopentadiene (E or (Z isomer) as a probe compound, CDOM-photosensitized isomerization took place [40]. In the absence of direct photo-irradiation of the probe compound (as was the case in that study) such a reaction indicates the occurrence of triplet–triplet energy transfer. Excited triplet states of the CDOM were proposed to be the main reactive transients responsible for the transformation of electron-rich phenols in aqueous solutions of humic and fulvic acids and in natural waters irradiated with UVA and visible light [41]. Using three aromatic ketones as surrogates of CDOM chromophores and comparing phenol transformation rates under photostationary conditions with triplet-quenching rate constants obtained by laser ﬂash photolysis [42], the effective one-electron reduction potential of CDOM excited triplet states was estimated to be at least 1.36 V versus the normal hydrogen electrode (NHE). Such triplet states were also postulated to be the transient oxidants that induced the transformation of various phenylurea herbicides [43, 44].

306

C. Richard · S. Canonica

3.4 Singlet Oxygen The lowest excited singlet state of molecular oxygen is a well characterized chemical species and is called singlet oxygen for simplicity. The signiﬁcance of singlet oxygen in aquatic photochemistry was already recognized in the 1980s [45, 46]. We refer the reader to the various reviews on indirect photochemical transformations for more detailed information [7–9, 12, 13]. The concentration of singlet oxygen in photo-irradiated waters can be most easily determined by following the depletion kinetics of selective probe compounds, in particular furane derivatives, such as 2,5-dimethylfurane [40] and furfuryl alcohol [47, 48]. The second-order rate constants for the reaction of singlet oxygen with various chemical compounds are available [49] and the determination of such constants for further compounds (including contaminants) is straightforward (see e.g. [50]). Alternatively, rate constants can be calculated using quantitative structure–activity relationships (QSARs) [50, 51]. Combining these two parameters, it is possible to calculate depletion rates of a large number of organic contaminants in any natural water. Photostationary state concentrations of singlet oxygen are of the same order of magnitude as those of their immediate precursors, the CDOM excited triplet states [40]. Since only a fraction of the latter are capable of inducing the transformation of dissolved compounds (by energy transfer [40] or oxidation [41]), singlet oxygen is potentially capable of inducing higher depletion rates than reactive triplet states. However, the second-order reaction rate constants of many organic contaminants with singlet oxygen are relatively low (several orders of magnitude below the diffusion limit), and the relevance of singlet oxygen for the depletion of organic contaminants in natural waters could be shown only in a relatively small number of cases [12, 52]. 3.5 Hydroxyl Radical Despite its prominent role in atmospheric chemistry [53, 54], the hydroxyl radical is considered a minor oxidant in the great majority of surface waters owing to the low formation rates [9] and efﬁcient scavenging by DOM and the bicarbonate/carbonate anions. DOM is at the same time both sensitizer and scavenger of the hydroxyl radical, but CDOM is responsible only for minor formation pathways, which can be summarized as: (6) CDOM + H2 O + hv → CDOM–· + · OH + H+ Below neutral pH, hydrogen peroxide formed by CDOM photoreactions participates in Fenton’s reaction to produce hydroxyl radical [55]: (7) H O + Fe(II)+n → Fe(III)OH+n + · OH 2 2

Phototransformation of Organic Contaminants by CDOM

307

Fenton type reactions are particularly important in atmospheric waters [10], which are acidic. An additional source of hydroxyl radical that is signiﬁcant only when using UV light of wavelength shorter than about 300 nm (i.e. in the case of water treatment using UV lamps, but not for sunlight) is represented by photolysis of hydrogen peroxide: (8) H2 O2 + hv → 2· OH Since hydroxyl radical reacts at very high rates with most organic contaminants (with the exception of some polyhalogenated compounds), it may end in being the main reactive species for the transformation of contaminants that do not appreciably react with the main transient species, i.e. triplet states and singlet oxygen. 3.6 Carbonate Radical One of the scavenging pathways of the hydroxyl radical leads directly to the formation of the carbonate radical: · OH + HCO3 – /CO3 2– → H2 O/– OH + CO3 –· (9) Other photochemical formation pathways of the carbonate radical may exist, but have not been investigated yet. Assuming that only the above reaction generates the carbonate radical, its formation rate cannot exceed the formation rate of the hydroxyl radical. DOM is thought to be its main scavenger present in natural waters, but it reacts at much lower rates than it does with the hydroxyl radical [56]. The result is that photostationary carbonate radical concentrations are typically about two orders of magnitude higher than hydroxyl radical concentrations [57]. Therefore, for organic contaminants that react fast with the carbonate radical, this may be much more effective than the hydroxyl radical. 3.7 DOM-Derived Radicals The great variety of possible radicals that might be formed upon DOM photoirradiation makes their identiﬁcation and the quantiﬁcation of their contribution to the transformation of contaminants a very complex task. Carboncentred radicals as a whole class of radicals have been detected by EPR spectroscopy using scavenging by nitroxides [58]. Only a few of these radicals, consisting of small molecular fragments, have been tentatively identiﬁed by trapping with nitroxide followed by derivatization and liquid chromatography with ﬂuorescence detection [59]. Although such radicals have never been shown to react appreciably with organic contaminants, they may play an important role as the precursors of DOM-derived oxyl and peroxyl radicals,

308

C. Richard · S. Canonica

which result upon addition of dissolved oxygen and subsequent reactions according to the following mechanism: R· + O2 → ROO· ROO· + ROO· → ROOOOR ROOOOR → 2RO· + O2 ROO· + PH → ROOH + P· RO· + PH → ROH + P·

(10) (11) (12) (13) (14)

where R· denotes a carbon-centred radical and PH an organic compound that undergoes oxidation. Peroxyl radicals play an important role in the autoxidation of organic compounds [60], in particular of lipids, and the chemistry of simple peroxyl radicals in aqueous solution has been studied extensively by pulse radiolysis [61]. In sunlit surface waters, DOM-derived oxyl and peroxyl radicals have been postulated to be the reactive species that effect the oxidation of various contaminants [62]. As discussed above, such an oxidation could also be brought about by excited triplet states of the CDOM [41] and a possible way to distinguish between (short-lived) triplet states and longer lived radicals has been suggested [63]. Radical cations, formed, for example, by photo-ionization of CDOM (see Scheme 1), may also be considered reactive oxidant species. They possibly contribute to the pool of DOM-derived photo-oxidants that cannot be assigned to excited triplet states.

4 Phototransformation of Organic Contaminants in the Presence of CDOM or Humic Substances The reported inﬂuence of humic substances or DOM on organic pollutants phototransformation is one of contrasts. Depending on the pollutant structure, inhibiting or enhancing effects were found. 4.1 Inhibition or Absence of Effect The photodegradation of 12 polycyclic aromatic hydrocarbons (PAHs) in pure water, in solutions of Suwannee River fulvic acid (5 mg L–1 ) and in natural waters under polychromatic light (λ > 290 nm) was investigated [64]. In general, PAH loss followed an apparent ﬁrst-order kinetics. The ratios of PAH photodegradation rate constants in natural waters or in the presence of fulvic acid compared to pure water were generally close to 1, indicating no effect, or were slightly lower, indicating a weak inhibition by DOM. In the case of anthracene (I) only was the ratio higher than 1, in accordance with

Phototransformation of Organic Contaminants by CDOM

309

Scheme 2

an enhancement effect of DOM on the photodegradation. However, Bertilsson [65] reported a non-effect of water-extracted fulvic acids added at a level of 25 mg L–1 on anthracene photodegradation while reporting a signiﬁcant inhibiting effect on phenanthrene (II). The inhibiting effect could not be satisfactorily rationalized. It could not have been due to light attenuation at the chosen experimental conditions. As an alternative, binding of PAHs to DOM might reduce photolysis by decreasing the lifetime of excited states. However, ﬂuorescence of PAHs was found to be only slightly reduced indicating little interaction. Moreover, a strong increase in fulvic acid concentration did not increase the inhibition. The peculiar behaviour of anthracene also remained unclear and the authors concluded that the photolysis of PAHs is not affected by DOM. The photodecomposition of carbofuran (2,3-dihydro-2,2-dimethyl-7-benzofuranyl-N-methylcarbamate) (III) was investigated in pure water and in the presence of various samples of organic matter (soil-extracted humic and fulvic acids) [66]. Irradiation of carbofuran (1.3 × 10–5 M) in pure water at 254 nm yielded several photoproducts; however, only IIIa and IIIb could be ﬁrmly identiﬁed. A step reaction was proposed: ﬁrst, formation of IIIa by cleavage of the O – CO bond, and second, formation of IIIb by photohydrolysis of the furan ring (see Scheme 3). The presence of the DOM samples was found to inhibit the carbofuran photolysis. The inhibition ranged from 19 to 70% for a concentration of DOM of 5–30 ppm. This decrease cannot be explained by a screen effect alone; indeed, at the highest DOM level (30 ppm), the decrease in light absorption by carbofuran was only of 27%, while the rate was decreased by 70%. Using the fact that carbofuran is ﬂuorescent, the authors determined its binding to

Scheme 3 Photolysis of carbofuran in pure water

310

C. Richard · S. Canonica

DOM. They monitored the ﬂuorescence emission intensity decay of carbofuran in solutions with various DOM concentrations, as described by Chen et al. [67]. Using the relationship: F0 /F = 1 + kb [DOM], where F0 and F are the ﬂuorescence intensity of carbofuran and carbofuran/DOM solution respectively, they evaluated kb , the binding constant. Values lay in the range 22–43 L g–1 depending on the DOM studied. In comparison, values of 91.5, 23.2 and 20.1 L (g of carbon)–1 were reported for ﬂuoranthene, napropamide and 1-naphthol, respectively [67]. This ability of carbofuran to bind to DOM was given as an explanation for the inhibiting effect of DOM on the photolysis. Since the UV spectrum of carbofuran was found to be unaffected by the binding, it means that carbofuran still absorbs energy. The inhibiting effect of DOM can be explained in different ways. An energy transfer to surrounding DOM molecules may be suggested; however, it seems unlikely, because it is generally admitted that the cleavage of the O – CO bond is a very fast reaction involving the singlet excited state, and the energy transfer should not be competitive. Alternatively, the binding may favour radiationless deactivation processes. It may also favour the back-recombination of radicals produced in the cleavage reaction, rendering the photochemical process inefﬁcient. Aciﬂuorfen (5-(2-chloro-α, α, α,-triﬂuoro-p-tolyl)-2-nitrobenzoic acid) (IV) mainly undergoes photodecarboxylation upon irradiation at 254 or 313 nm. This compound was exposed to solar light, ﬁrst in the presence of humic acids (10 mg L–1 ), and then dissolved in a natural water sampled in an artiﬁcial lake [68]. No effect of humic acids was observed. The lack of a photo-inductive effect of humic substances can be explained by the fact that aciﬂuorfen bears electron-withdrawing substituents rendering it poorly oxidizable by reactive triplet states. In contrast, the half-life of IV was found to decrease by a third in the natural water. This photo-inductive effect might be due to the photogeneration of hydroxyl radicals by chromophores other than humic components of DOM; nitrate and nitrite ions are probable candidates.

Scheme 4

4.2 Enhancing Effect DOM was also found to have an enhancing effect on the degradation of organic substances, which could be due to various processes. Several examples are presented in this section.

Phototransformation of Organic Contaminants by CDOM

311

4.2.1 Sulphur-Containing Compounds The fungicides carboxin (5,6-dihydro-2-methyl-1,4-oxathi-ine-3-carboxanilide) (V) and oxycarboxin (5,6-dihydro-2-methyl-1,4-oxathi-ine-3-carboxanilide-4,4-dioxide) (VI) were photolysed in artiﬁcial solar light (λ > 290 nm) in pure water and in the presence of soil-extracted humic and fulvic acids (Scheyern, Germany) (1 mg L–1 ) [69].

Scheme 5

Carboxin was much more readily photodegraded than oxycarboxin in pure water. The oxidation of the sulphur atom yielded the sulphoxide Va and the subsequent loss of the heterocyclic ring produced the oxanilic acid Vb (see Scheme 6). Humic substances enhanced the phototransformation of both fungicides. For carboxin, this enhancement is possibly related to the presence of the sulphur atom that might be oxidized by singlet oxygen or other photo-oxidants. Radiolabelled derivatives of the herbicide ﬂorasulam (N-(2,6-diﬂuorophenyl)-5-methoxy-8-ﬂuoro(1,2,4)-triazolo-[1,5-c]-pyrimidine-2-sulphonamide) (VII) were exposed to natural sunlight in a sterile pH 5 buffer water and in a natural lake water collected from 20 to 30 cm below the surface [70]. The photodegradation was much faster in the natural water system, with a half-life of 3.3 days against 73 days in the buffered aqueous medium. Moreover, the photoproducts produced in the distilled and natural waters were found to be different. Direct photolysis led to the cleavage of the N – S bond with formation of the sulphonic acid derivative (VIIa) after 10% of conversion (see Scheme 7).

Scheme 6 Photolysis of carboxin in pure water

312

C. Richard · S. Canonica

Scheme 7 Direct photolysis of ﬂorasulam in pure water

In the natural water, many photoproducts were found (see Scheme 8). Florasulam underwent photosubstitution of the methoxy by the hydroxy group to form VIIb, loss of the diﬂuorophenyl moiety yielding VIIc with subsequent replacement of the methoxy by the hydroxy to produce VIId. The opening of the pyrimidine ring to form the carboxylic acid derivative led to VIIe and VIIf.

Scheme 8 Phototransformation of ﬂorasulam in natural water

Phototransformation of Organic Contaminants by CDOM

313

If indirect photolytic processes are clearly important contributors to the photolytic degradation of ﬂorasulam in the aquatic environment, the role of DOM in the reaction cannot be evaluated without additional experiments using DOM or humic substances as the only photosensitizers. Irgarol 1051 (2-methylthio-4-tert-butylamino-6-cyclopropylamino-s-triazine) (VIII) is a newly developed herbicide used as an alternative to tributyltin in antifouling paints. The photochemical degradation was investigated in distilled water and sea, river and lake waters as well as in aqueous solutions of Fluka humic acid or International Humic Substances Society (IHSS) standard fulvic acids isolated from Suwannee River. Natural waters that were used without ﬁltration or sterilization contained 2.5–10 mg L–1 in organic carbon. Their pHs lay in the range 7.67–7.9. Humic and fulvic acids were used at a level ranging from 4 to 24 mg L–1 . All the solutions were irradiated in artiﬁcial solar light (λ > 290 nm) and in solar light (May–June in Ioannina, Greece) [71]. The rate of Irgarol 1051 (1 mg L–1 ) photolysis was found to be slightly lower in environmental waters than in distilled water, while humic and fulvic acids signiﬁcantly increased (4.5- to 14-fold) the rate of substrate consumption. This discrepancy may be due to the presence of particulate matter scattering incident light in the environmental waters. Gas chromatography-mass spectrometry (GC-MS) analyses of solid-phase extracted irradiated samples revealed the presence of photoproducts VIIIa to VIIIe produced via dealkylation and/or oxidation of the sulphur atom (see Scheme 9). This study does not indicate differences in photoproducts between DOM-free and DOM-containing solutions. The sulphur atom is oxidized and the alkyl groups released through direct absorption of light; reactive species photogenerated by DOM are also expected to attack these substituents. The photochemical fate of histamine H2 -receptor antagonists cimetidine (IX) and ranitidine (X) in Mississipi River water (16 mg L–1 of dissolved organic carbon, pH = 8) was investigated [52]. Ranitidine was found to undergo direct photolysis under midsummer sunlight (Minneapolis, 45 latitude) with a quantum yield of 5.5 × 10–3 . In contrast, cimetidine was shown to react negligibly under the same irradiation conditions due to lack of appreciable absorption in the wavelength region supplied by the solar spectrum. In the natural water, both compounds were photodegraded. The involvement of singlet oxygen in these transformations was evidenced using the quenching technique. Indeed, the addition of 1 O2 quenchers 1,4-diazabicyclo [2,2,2] octane (DABCO) and sodium azide diminished the rate of IX and X depletion to a value comparable to that in distilled water, while spiking with 2-propanol (1%) did not alter the photodegradation rates. Moreover, replacing H2 O by D2 O yielded a solvent isotope effect that conﬁrmed the reaction with singlet oxygen. To assess the sites of reaction of IX and X with singlet oxygen, the authors measured the bimolecular rate constants of singlet oxygen quenching by parent and model compounds using laser ﬂash photolysis. These data showed that the reactive moiety is the dialkylimidazole ring in IX

314

C. Richard · S. Canonica

Scheme 9 Photodegradation pathways of irgarol in natural waters

and the furfuryl group in X. The sulphide functionality, which is a potential site for oxygenation, does not seem to contribute signiﬁcantly to the reaction as no sulphoxide was found among the photoproducts. From this study, it was concluded that the fate of X in the aquatic environment is mainly governed by direct photolysis, while that of IX is governed by reaction with singlet oxygen.

Scheme 10

Phototransformation of Organic Contaminants by CDOM

315

4.2.2 Other Compounds The photodegradation of carbaryl (1-naphthyl-N-methyl carbamate) (XI), an insecticide, was also found to be enhanced in natural compared to distilled water [72].

Scheme 11

The study was focused on both nitrate and DOM effects. For this purpose, the natural water was sampled at different times in Old Woman Creek Estuarine Reserve, Ohio, USA from June to September, and each sample was characterized for nitrate and DOM contents. The nitrate concentration varied within the range 0.0185–1.01 mM, and that of organic carbon within the range 2–7 mg L–1 . In the waters with relatively high concentration of nitrate, the rate of carbaryl phototransformation drastically increased, as expected since photolysis of nitrate is known to produce hydroxyl radical. In waters with low nitrate content, the photodegradation of carbaryl was promoted too and this effect was attributed to DOM. To test the role of iron or other trace metals in these reactions, metals were removed by passing the samples through a cation-exchange resin. This operation did not affect the results of the photochemical experiments, indicating that the photo-Fenton-type reaction played a minor role. Propiconazole (1-[2-(2,4-dichlorophenyl)-4-propyl-1,3-dioxolan-2-ylmethyl]-1H-1,2,4-triazole) (XII), which is used as a fungicide, was found to undergo different types of reaction in pure water depending on the wavelength of irradiation [73]. Photocyclization after HCl elimination to give XIIa (see Scheme 12) and subsequent photohydrolysis of XIIa prevailed over oxidation at 254 nm while the opposite result was observed in natural solar light (June, Clermont-Ferrand, France, 46◦ N, 3◦ E) or simulated solar light. A group of peaks showing a mass of M + 14 or M + 16 and corresponding to oxidation products (M + O – 2H or M + O) appeared in high performance liquid chromatography (HPLC). Their exact structures could not be determined; they are likely the result of oxidation of the dioxolane ring or of the aliphatic chain. The photodegradation of XII was enhanced by humic substances (10 mg L–1 ) or in a natural water sampled in a dam (Villerest, France) (7.2 mg C L–1 ). Formation of the photoproducts was measured by HPLCatmospheric pressure chemical ionization-MS (HPLC-APCI-MS) using se-

316

C. Richard · S. Canonica

Scheme 12 Photocyclization of propiconazole in pure water

lected ion monitoring (SIM) signals corresponding to [M + H]+ at the detected compounds. It was thus possible to evaluate the distribution pattern of photoproducts in pure and DOM-containing water. The presence of humic substances was found to slightly reduce the formation rate of XIIa, and to increase twofold that of oxidation products. It can be concluded that photocyclization is characteristic of direct photolysis, and therefore slightly reduced by a screen effect, while oxidation results from both direct photolysis and photo-induced reactions. Dehydroabietic acid (XIII) is the most abundant of the resin acids that occur naturally in wood and tree bark. These compounds are toxic and persistent in the environment. The effect of DOM on the photolysis of XIII was investigated at 254 nm and at λ > 300 nm [74]. In pure water, XI (3 mg L–1 ) was degraded both by 254 nm radiation and simulated solar light. The degradation of XIII at 254 nm was slower in the humic surface water (Lake Sarojärvi, Finland) than in pure water. The opposite effect was observed at λ > 300 nm. These ﬁndings show that the direct photolysis of dehydroabietic acid is the dominant process at short wavelength, DOM acting as an inner ﬁlter. At λ > 300 nm, XIII absorbs poorly and photo-inductive processes become more important than direct photolysis. The analytical study revealed the presence of nine photoproducts corresponding to decarboxylation or oxidation reactions. The number of photoproducts was found to be highest in simulated

Scheme 13

Phototransformation of Organic Contaminants by CDOM

317

sunlight or in the natural water. The lack of readily oxidizable moieties in XIII and the variety of photoproducts formed led us to conclude that hydroxyl radical is the probable reactive species. The pesticide Mirex (dodecachloropentacyclo[5.3.0.02,6 .04,8 ]decane) (XIV) constitutes a very special case, since the DOM-mediated photodechlorination into 8-monohydromirex (XIVa) was demonstrated as occurring wholly within humic acid molecules. This result was shown by modelling the effects of humic acids and scavengers concentrations on the dechlorination rates [75]. A kinetic model based on a homogeneous phase reaction could not ﬁt the change in the apparent Mirex reaction rate with humic acid concentration. In contrast, a good ﬁt was achieved by postulating that the reaction of Mirex occurs in separate bound and dissolved environments. In addition, the effect of additives such as nitrate, 2-chloroethanol or lindane on the apparent reaction rate was studied. The results of the decrease effect modelling indicated that the photoreactivity of Mirex is restricted to the bound phase. Another work was devoted to the mechanism of this intrahumic dechlorination [76]. Using N,N-dimethylaniline solution, which is a model system for generating hydrated electrons it was shown that Mirex reacts with hydrated electrons yielding XIVa. Then, by comparing the reduction rate of lindane, Mirex and dichlorobenzene in humic acids solutions, in N,N-dimethylaniline solution, and in humic acids solutions containing N,N-dimethylaniline, they could conclude that only bound Mirex molecules react with solvated electrons. The quantum yield of Mirex photoreduction at 300 nm (0.03) fell within the range of values previously measured for hydrated electron generation using ﬂash photolysis at 355 nm (0.22–0.0017) for natural waters and humic acids solutions. These latter values were found to be much higher than those obtained under steady-state conditions using 2-chloroethanol as a hydrated electron quencher [33] showing the poor ability of 2-chloroethanol to insert into hydrophobic zones and to have access to these transient species. The photodegradation of the fungicide chlorothalonil (2,4,5,6-tetrachloroisophtalonitrile) (XV), studied under the same conditions as VIII, was found

Scheme 14

318

C. Richard · S. Canonica

Scheme 15

to be faster in river or lake waters or in water containing humic substances than in distilled water [77]. The half-life was reduced sixfold in the two former cases and twofold in the latter. Benzamide and reduction photoproducts (mono- and dichlorinated derivatives) could be characterized by GC-MS. Chlorothalonil, which bears several electron-withdrawing substituents, could be only oxidized by a strong oxidant and the involvement of hydroxyl radicals was very likely. Photoreduction also occurred. Since chlorothalonil is very poorly soluble in water, it might sorb to DOM and react with solvated electrons produced within the hydrophobic zones of DOM in a similar way to Mirex. The high production volume chemical, 4-chloro-2-methylphenol, XVI was photolysed in distilled and natural waters and aqueous solutions of soilextracted humic substances (Ranker, Spain) under various irradiation conditions: with monochromatic light at 280 nm, using polychromatic lamps emitting between 300 and 450 nm, and in natural sunlight (June, ClermontFerrand, France) [78, 79]. With excitation at 280 nm, XVI was shown to be very efﬁciently dechlorinated (φ = 0.66), mainly forming methylbenzoquinone in suroxygenated solution that in turn is photolysed into methylhydroquinone and methylhydroxybenzoquinone. In the wavelength range where XVI does not absorb (λ > 300 nm), humic substances were found to photo-induce the transformation, the rate of photodegradation being about 40 times higher in the presence than in the absence of humic acids. The ring-opening photoproducts XVIa and XVIb were detected. Their formation was proposed to involve the

Scheme 16

Phototransformation of Organic Contaminants by CDOM

319

reactive triplet states of humic substances since the addition of radical scavengers such as 2-propanol and azide anions did not affect the reaction. It was suggested to result from an initial abstraction of the phenolic hydrogen atom by the reactive triplets to yield the phenoxyl radical, followed by the addition of the superoxide anion radical at the carbon atom bearing the methyl group. The cleavage of the CO – C(OOH)CH3 bond is then expected to take place producing the trans–trans (XVIa) and cis–trans(XVIb) muconic acid derivatives. Irradiation of XVI in natural waters with solar light yielded a mixture of methylhydroquinone, methyl hydroxybenzoquinone, XVIa and XVIb. It shows that under these conditions, XVI is phototransformed via both direct photolysis and DOM-mediated photoreactions. The phenylurea herbicides are an interesting series of compounds to study, owing to the fact that substituents on the ring can inﬂuence the global reactivity. Unsubstituted fenuron (XVII) was proved to undergo a photo-induced transformation in the presence of soil-extracted humic acids on irradiation at 365 nm [43, 80]. The addition of a Cl atom in the para position (monuron) was found to decrease the rate of reaction by a factor of 6 while the addition of two Cl in para and meta positions (diuron) did so by a factor of 14. The strong inﬂuence of ring substituents on the reactivity was conﬁrmed in another study [44] where the phototransformations of a series of 11 substituted phenylurea herbicides were compared. The apparent ﬁrst order rate constant of photosensitized transformation in an aqueous solution of Suwannee River fulvic acid was found to increase with electron-donating substituents and to decrease with electron-withdrawing substituents. Such a substituent effect is a good indication that the ﬁrst step in the phototransformation is a oneelectron oxidation (or a H atom transfer) from the phenylurea to the reactive triplets, as proposed for phenolic compounds [41]. Deoxygenation of the aqueous solution was found to increase the phototransformation of monuron (XVIII) by a factor of 2 in the presence of soilextracted fulvic acids at 365 nm, but to drastically inhibit that of XVII [80]. These results were tentatively rationalized by postulating a triplet–triplet energy transfer in the former case. Fenuron (XVII) was used to compare the photo-inductive properties of several soil-extracted humic and fulvic acids [81]. Solutions containing humic

Scheme 17

320

C. Richard · S. Canonica

Scheme 18

or fulvic acids (100 mg L–1 ) and XVII (2 × 10–4 M) were irradiated at 365 nm. In all cases, XVII was found to disappear. As a general rule, the rate of XVII consumption was higher in the presence of the fulvic rather than the humic acids extracted from the same soil. The rate decreased as the absorbance of the humic substances at 365 nm increased, showing that photo-inductive chromophores contribute weakly to the absorbance at 365 nm. Most of the light is absorbed by DOM moieties that exhibit poor or no photo-inductive activity. Microcystins are cyanobacterial toxins released in the aquatic environment through cell lysis. The removal of these persistent and highly toxic compounds is of particular concern in drinking water production. MicrocystinLR (XIX) was found to be photostable in sunlight. However, the photodegradation was found to be sensitized by fulvic acids and CDOM [82, 83]. A detailed study on the CDOM-sensitized phototransformation of XIX was undertaken using the sorbate ion to mimic the photoreactive moiety of XIX [83]. Photo-isomerization of the sorbate ion was observed upon irradiation at 365 nm in the presence of CDOM extracted using a XAD-8 resin. This reaction can only be explained by a triplet–triplet energy transfer reaction between CDOM triplet excited states and ground state sorbate ion. Photooxidation involving singlet oxygen was also shown to occur using the scavenging technique.

5 Conclusion Different types of reactive species are produced upon irradiation of CDOM. Their contribution to the fate of organic pollutants in the aquatic environment can be assessed on the basis of the reported literature. It generally

Phototransformation of Organic Contaminants by CDOM

321

competes with photochemical processes due to direct light absorption of the pollutants. PAHs are recalcitrant to photo-induced transformation. Aromatic compounds bearing electron-withdrawing substituents are either recalcitrant or undergo photoreduction in the bound phase. Sulphur-containing compounds or aromatics bearing hydroxyl or NH functionalities undergo photo-induced oxidation. An original triplet–triplet energy transfer reaction between CDOM triplet excited states and ground state compounds bearing a sorbate moiety was observed. In general, photo-induced reactions are slow compared with the direct photolysis rate of photoreactive compounds.

References 1. Zepp RG, Baughman GL (1978) In: Hutzinger O, van Lelyveld IH, Zoeteman BCJ (eds) Aquatic pollutants: transformation and biological effects. Pergamon, Oxford, UK, p 237 2. Larson RA (1978) Critical reviews in environmental control, vol 8. CRC, Boca Raton, FL, USA, p 197 3. Zepp RG (1980) In: Haque R (ed) Dynamics, exposure and hazard assessment of toxic chemicals. Ann Arbor Science, Ann Arbor, Michigan, p 69 4. Mill T (1980) In: Hutzinger O (ed) The handbook of environmental chemistry, vol 2, part A. Springer, Berlin Heidelberg New York, p 77 5. Zaﬁriou OC, Joussot-Dubien J, Zepp RG, Zika RG (1984) Environ Sci Technol 18:A358 6. Mill T (1989) Environ Toxicol Chem 8:31 7. Cooper WJ, Zika RG, Petasne RG, Fischer AM (1989) Adv Chem Ser 219:333 8. Hoigné J, Faust BC, Haag WR, Scully FE, Zepp RG (1989) Adv Chem Ser 219:363 9. Hoigné J (1990) In: Stumm W (ed) Aquatic chemical kinetics: reaction rates of processes in natural waters. Wiley-Interscience, New York, NY, p 43 10. Faust BC (1994) Environ Sci Technol 28:A217 11. Mill T (1999) Chemosphere 38:1379 12. Larson RA, Marley KA (1999) In: Boule P (ed) The handbook of environmental chemistry, vol 2, part L. Springer, Berlin Heidelberg New York, p 123 13. Faust BC (1999) In: Boule P (ed) The handbook of environmental chemistry, vol 2, part L. Springer, Berlin Heidelberg New York, p 101 14. Boule P, Bolte M, Richard C (1999) In: Boule P (ed) the handbook of environmental chemistry, vol 2, part L. Springer, Berlin Heidelberg New York, p 181 15. Brezonik PL (1994) Chemical kinetics and process dynamics in aquatic systems. Lewis, Boca Raton, FL 16. Schwarzenbach RP, Gschwend PM, Imboden DM (2003) Environmental organic chemistry, 2nd edn. Wiley, Hoboken, NJ 17. Leenheer JA, Croue JP (2003) Environ Sci Technol 37:18A 18. Steelink C (2002) Anal Chem 74:326a 19. Averett RC, Leenheer JA, McKnight DM, Thorn KA (eds) (1989) Humic substances in the Suwannee River, Georgia: interactions, properties, and proposed structures. Open File Report 87–557. US Geological Survey, Denver, CO 20. Gaffney JS, Marley NA, Clark SB, (eds) (1996) Humic and fulvic acids – isolation, structure, and environmental role. ACS Symp Ser, vol. 651 21. Chin YP, Aiken G, Oloughlin E (1994) Environ Sci Technol 28:1853

322

C. Richard · S. Canonica

22. 23. 24. 25.

Choudhry GG (1981) Toxicol Environ Chem 4:261 Dunnivant FM, Schwarzenbach RP, Macalady DL (1992) Environ Sci Technol 26:2133 Green SA, Blough NV (1994) Limnol Oceanogr 39:1903 Blough NV, Green SA (1995) In: Zepp RG, Sonntag C (eds) Role of nonliving organic matter in the earth’s carbon cycle. Wiley, Chichester, UK, p 23 Zepp RG, Cline DM (1977) Environ Sci Technol 11:359 Coble PG (1996) Mar Chem 51:325 Yamashita Y, Tanoue E (2003) Mar Chem 82:255 Power JF, Sharma DK, Langford CH, Bonneau R, Joussot-Dubien J (1987) ACS Symp Ser 327:157 Fischer AM, Kliger DS, Winterle JS, Mill T (1985) Chemosphere 14:1299 Fischer AM, Winterle JS, Mill T (1987) ACS Symp Ser 327:141 Bruccoleri A, Pant BC, Sharma DK, Langford CH (1993) Environ Sci Technol 27:889 Zepp RG, Braun AM, Hoigné J, Leenheer JA (1987) Environ Sci Technol 21:485 Bielski BHJ, Cabelli DE, Arudi RL, Ross AB (1985) J Phys Chem Ref Data 14:1041 Stanbury DM (1989) Adv Inorg Chem 33:69 Sedlak DL, Hoigné J (1993) Atmos Enviro A – Gen Top 27:2173 Sturzenegger VT (1989) PhD thesis. Swiss Federal Institute of Technology Zurich Goldstone JV, Voelker BM (2000) Environ Sci Technol 34:1043 Voelker BM, Sedlak DL, Zaﬁriou OC (2000) Environ Sci Technol 34:1036 Zepp RG, Schlotzhauer PF, Sink RM (1985) Environ Sci Technol 19:74 Canonica S, Jans U, Stemmler K, Hoigné J (1995) Environ Sci Technol 29:1822 Canonica S, Hellrung B, Wirz J (2000) J Phys Chem A 104:1226 Aguer JP, Richard C (1996) Pestic Sci 46:151 Gerecke AC, Canonica S, Muller SR, Scharer M, Schwarzenbach RP (2001) Environ Sci Technol 35:3915 Zepp RG, Baughman GL, Schlotzhauer PF (1981) Chemosphere 10:119 Haag WR, Hoigné J (1986) Environ Sci Technol 20:341 Haag WR, Hoigné J, Gassman E, Braun AM (1984) Chemosphere 13:631 Haag WR, Hoigné J, Gassman E, Braun AM (1984) Chemosphere 13:641 Wilkinson F, Helman WP, Ross AB (1995) J Phys Chem Ref Data 24:663 Tratnyek PG, Hoigné J (1991) Environ Sci Technol 25:1596 Canonica S, Tratnyek PG (2003) Environ Toxicol Chem 22:1743 Latch DE, Stender BL, Packer JL, Arnold WA, McNeill K (2003) Environ Sci Technol 37:3342 Atkinson R (1986) Chem Rev 86:69 Seinfeld JH (1986) Atmospheric chemistry and physics of air pollution. Wiley, New York, NY Southworth BA, Voelker BM (2003) Environ Sci Technol 37:1130 Larson RA, Zepp RG (1988) Environ Toxicol Chem 7:265 Sulzberger B, Canonica S, Egli T, Giger W, Klausen J, von Gunten U (1997) Chimia 51:900 Blough NV (1988) Environ Sci Technol 22:77 Kieber DJ, Blough NV (1990) Anal Chem 62:2275 Ingold KU (1969) Acc Chem Res 2:1 von Sonntag C, Schuchmann H-P (1997) In: Alfassi ZB (ed) Peroxyl radicals. Wiley, New York, NY p 173 Mill T, Hendry DG, Richardson H (1980) Science 207:886 Canonica S, Freiburghaus M (2001) Environ Sci Technol 35:690 Fasnacht MP, Blough NV (2002) Environ Sci Technol 36:4364

26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64.

Phototransformation of Organic Contaminants by CDOM 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83.

323

Bertilsson S, Widenfalk A (2002) Hydrobiologia 469:23 Bachman J, Patterson HH (1999) Environ Sci Technol 33:874 Chen SJ, Inskeep WP, Williams SA, Callis PR (1994) Environ Sci Technol 28:1582 Vialaton D, Baglio D, Paya-Perez A, Richard C (2001) Pest Manag Sci 57:372 Hustert K, Moza PN, Kettrup A (1999) Chemosphere 38:3423 Krieger MS, Yoder RN, Gibson A (2000) J Agric Food Chem 48:3710 Sakkas VA, Lambropoulou DA, Albanis TA (2002) J Photochem Photobiol A Chem 147:135 Miller PL, Chin YP (2002) J Agric Food Chem 50:6758 Vialaton D, Pilichowski JF, Baglio D, Paya-Perez A, Larsen B, Richard C (2001) J Agric Food Chem 49:5377 Corin NS, Backlund PH, Kulovaara MAM (2000) Environ Sci Technol 34:2231 Burns SE, Hassett JP, Rossi MV (1996) Environ Sci Technol 30:2934 Burns SE, Hassett JP, Rossi MV (1997) Environ Sci Technol 31:1365 Sakkas VA, Lambropoulou DA, Albanis TA (2002) Chemosphere 48:939 Vialaton D, Richard C, Baglio D, Paya-Perez AB (1998) J Photochem Photobiol A Chem 119:39 Vialaton D, Richard C (2002) Aquat Sci 64:207 Aguer JP, Richard C (1999) Chemosphere 38:2293 Aguer JP, Richard C, Trubetskaya O, Trubetskoj O, Leveque J, Andreux F (2002) Chemosphere 49:259 Welker M, Steinberg C (2000) Environ Sci Technol 34:3415 Chasson G (2003) PhD thesis, University of Poitiers

Hdb Env Chem Vol. 2, Part M (2005): 325–366 DOI 10.1007/b138188 © Springer-Verlag Berlin Heidelberg 2005 Published online: 16 September 2005

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment Marta I. Litter Unidad de Actividad Química, Centro Atómico Constituyentes, Comisión Nacional de Energía Atómica, Av. Gral. Paz 1499, Gral. San Martín, 1650 Buenos Aires, Argentina [email protected] 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

327

2 2.1 2.2 2.3 2.4 2.5 2.5.1 2.5.2 2.5.3 2.5.4 2.6 2.6.1 2.6.2 2.6.3 2.6.4 2.6.5 2.6.6 2.6.7 2.6.8 2.6.9 2.7 2.8 2.9 2.10

Photochemical Advanced Oxidation Technologies . . . . . . . . . Direct Photolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . Sensitization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photolysis of Water in the Vacuum Ultraviolet (VUV) . . . . . . . UV/H2 O2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UV/O3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thermal Ozonation . . . . . . . . . . . . . . . . . . . . . . . . . . O3 /H2 O2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photoinduced Ozonation . . . . . . . . . . . . . . . . . . . . . . . UV/O3 /H2 O2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photo-Fenton and Related Reactions . . . . . . . . . . . . . . . . . Fenton Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Photo-Fenton Reaction and Other Iron-Based Photoprocesses Photo-Ferrioxalate and Other Fe(III) Complexes . . . . . . . . . . SORAS Technology . . . . . . . . . . . . . . . . . . . . . . . . . . Zero-Valent Iron . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photo-Fenton and Ozonation . . . . . . . . . . . . . . . . . . . . . Photoelectro-Fenton . . . . . . . . . . . . . . . . . . . . . . . . . . Immobilized Photo-Fenton Systems . . . . . . . . . . . . . . . . . Active Species in Fenton and Photo-Fenton Systems . . . . . . . . UV/Periodate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Heterogeneous Photocatalysis . . . . . . . . . . . . . . . . . . . . Comparative Practical Examples . . . . . . . . . . . . . . . . . . . Combination of PAOTs with Biological Treatments . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . .

330 331 332 333 335 339 339 341 342 343 344 344 346 349 352 352 353 353 354 354 355 356 359 359

3

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

361

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

363

. . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . .

Abstract In this chapter, an overview of Photochemical Advanced Oxidation Technologies (PAOTs) is given, together with recent relevant literature examples and references. ShortUV and VUV photolysis, UV/H2 O2 , UV/O3 , UV/O3 /H2 O2 , photo-Fenton and iron-based technologies, photo-ferrioxalate and UV/periodate, are exposed, together with a brief introduction of heterogeneous photocatalysis. Fundamental grounds with mechanistic pathways are described in each case. Combination of PAOTs with other treatments (espe-

326

M.I. Litter

cially biological methods) is also illustrated. Limitations, advantages and drawbacks are pointed out, together with different examples of real cases. Keywords Photochemical Advanced Oxidation Technologies · Vacuum ultraviolet · Ozonation · Direct photolysis · Photo-Fenton · Photo-ferrioxalate · Photocatalysis Abbreviations 1O singlet oxygen 2 2,4,5-T 2,4,5-trichlorophenoxyacetic 2,4-D 2,4-dichlorophenoxyacetic acid 2,4-DNT 2,4-dinitrotoluene 2,6-DNT 2,6-dinitrotoluene 4-CP 4-chlorophenol AMBI 5-amino-6-methyl-2-benzimidazolone AOPs Advanced Oxidation Processes AOTs Advanced Oxidation Technologies biological oxygen demand during an incubation period of 5 days at 37 ◦ C BOD5 C2 O4 •– oxalyl radical COD chemical oxygen demand CPC compound parabolic solar collector DOC dissolved organic carbon EDTA ethylenediaminetetraacetic acid FBR ﬁxed bed reactor FeOx ferrioxalate GAC granular activated carbon hydroxyl radical HO• hydroperoxyl radical HO2 • LMCT ligand to metal charge transfer MTBE methyl tert-butyl ether NB nitrobenzene NOM natural organic matter NTA nitrilotriacetic acid superoxide radical O2 •– OM organic matter PAOPs photochemical Advanced Oxidation Processes PAOTs photochemical Advanced Oxidation Technologies PCBs polychlorinated biphenyls PET polyethyleneterephthalate RB rose bengal Sens sensitizers TCE trichloroethylene THM trihalomethanes TNT trinitrotoluene TOC total organic carbon VUV vacuum ultraviolet

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

327

1 Introduction The growing demand from society for disinfection and detoxiﬁcation of polluted waters from different sources, materialized in very strict governmental regulations, has led, in the last few decades, to the development of new and more effective water puriﬁcation technologies. In most cases, anthropogenically polluted water can be efﬁciently treated by biological methods, activated carbon adsorption or other adsorbents, or by conventional physical and chemical treatments (ﬂocculation, ﬁltration, thermal oxidation, chlorination, ozonation, potassium permanganate, etc.). Nevertheless, in some cases, these procedures are not adequate to reach the degree of purity required by law or by the ﬁnal use. In those cases, Advanced Oxidation Technologies or Processes (AOTs, AOPs) are efﬁcient novel methods for water treatment, which have afforded very good results in industrialized countries and are beginning to be employed in developing regions [1–5]. AOTs are based on physicochemical processes that produce profound changes in the structure of chemical species. The concept was initially established by Glaze et al. [2, 6, 7], who deﬁned AOPs as processes involving generation and use of powerful transitory species, principally the hydroxyl radical (HO• ). This species can be generated by photochemical means (including solar light) or by other forms of energy, and has a high efﬁciency for organic matter (OM) oxidation. Some AOTs, such as heterogeneous photocatalysis, radiolysis and others, can also produce reducing agents, allowing the transformation of pollutants that are difﬁcult to oxidize, such as some metal ions or halogenated compounds. AOTs are usually divided into nonphotochemical and photochemical processes, as listed in Table 1. In this article, only photochemical technologies (PAOTs, PAOPs) will be reviewed, with some references to the non-photochemical process in the cases of ozonation and the Fenton reagent. For other technologies, the references indicated in Table 1 can be consulted, as well as references [8] and [9]. Concerning heterogeneous photocatalysis, this subject will be treated in detail in another article of this book, and we will only make a brief mention in this chapter. Examples and references are principally those covering the last ﬁve years, with the exception of the most relevant papers on the subject. Older references may be consulted in the referenced papers. The high efﬁciency of AOPs is supported on thermodynamic and kinetic grounds, due to the participation of radicals. The hydroxyl radical can attack virtually all organic compounds and it reacts 106 –1012 times more rapidly than alternative oxidants such as O3 . In Table 2, it can be observed that, after ﬂuorine, HO• is the most energetic oxidant. Table 3 shows that the reaction constant rates of different compounds with HO• are several orders of magnitude higher than those with O3 . However, we must emphasize that

328

M.I. Litter

Table 1 Advanced Oxidation Technologies and other related processes Non-photochemical processes

Photochemical processes

Alkaline ozonation (O3 /OH– ) [2, 7, 35, 41, 45] Ozonation with hydrogen peroxide (O3 /H2 O2 ) [42, 45] Fenton and related processes (Fe2+ /H2 O2 ) Electrochemical oxidation [124, 125] γ -Radiolysis and electron-beam treatment [126–128] Non-thermal plasma [129] Electrohydraulic discharge—ultrasound [2, 130–133] Oxidation in sub/and supercritical water [134–137] Zero-valent iron [94, 138, 139] Ferrate (K2 FeO4 , Fe(VI)) [140]

Water photolysis in vacuum ultraviolet (VUV) UV/ hydrogen peroxide UV/O3 Photo-Fenton and related processes UV/periodate Heterogeneous photocatalysis

Table 2 Redox potentials of some oxidants [1] Species

E0 (V, 25 ◦ C)1

Fluorine Hydroxyl radical Atomic oxygen Ozone Hydrogen peroxide Perhydroxyl radical Permanganate Chlorine dioxide Hypochlorous acid Chlorine Bromine Iodine

3.03 2.80 2.42 2.07 1.78 1.70 1.68 1.57 1.49 1.36 1.09 0.54

1

Redox potentials referred to normal hydrogen electrode (NHE)

the efﬁciency of AOTs resides in the generation of high concentrations of hydroxyl radicals in the steady state. Another active oxygen species is the superoxide radical, O2 •– , and its conjugate acid form, the hydroperoxyl radical, HO2 • , and these are also produced in many AOTs, but they are far less active than HO• .

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

329

Table 3 Rate constants (k in M–1 s–1 ) for some organic compounds with hydroxyl radical and ozone [4] Compound

HO•

O3

Chlorinated alkenes Phenols Aromatics Ketones Alcohols Alkanes

109 –1011 109 –1010 108 –1010 109 –1010 108 –109 106 –109

10–1 –103 103 1–102 1 10–2 –1 10–2

When a target pollutant compound is attacked by HO• , three main mechanisms may be involved in the degradation of organics: hydrogen abstraction, OH addition or substitution, and electron transfer. Hydrogen abstraction is generally the ﬁrst step in many acid compounds [1, 4]: RH + HO• → H2 O + R• → further oxidation reactions R• + O2 → ROO• → further oxidation reactions

(1) (2)

If the target is an aromatic compound, the ﬁrst stage is ring hydroxylation, but further HO• attack leads to the opening of the ring and the formation of open conjugated structures:

Scheme 1

The majority of AOTs can be applied to the remediation and detoxiﬁcation of low or medium volumes of waters. Ground, surface, and wastewater can be treated, giving rise to the destruction or transformation of hazardous or refractory pollutants. Point sources of toxic pollutants such as pesticides, heavy metals and others can be treated in small-scale mobile treatment units, easy to install in industrial plants. The methods can be used alone or combined with other AOTs or with conventional methods. The use of modular units allows the selection of the best technology or combination of technologies to treat a speciﬁc wastewater. AOTs can also be applied to pollutants in the air and soil, and they may even allow disinfection or sterilization of bacteria, viruses, and other microorganisms. AOTs offer several advantages over conventional methods of treatment. One of the most important characteristics is that pollutants are not merely

330

M.I. Litter

transferred from one phase to another (as in air stripping or activated carbon treatment), but they are chemically transformed, leading, in many cases, to complete mineralization (destruction) of the pollutant. In consequence, Advanced Oxidation Processes are very useful for treating refractory pollutants resistant to other treatments such as, for example, biological technologies. AOTs can treat contaminants at very low levels (ppb), and reaction by-products are generally not formed. The technologies are also useful for improving the organoleptic properties of water, or can just be used to discolor dark industrial wastes. In most cases, they consume much less energy than some conventional methods such as, for example, incineration. Nevertheless, it must be taken into account that wastes with relatively high chemical oxygen demand (COD) contents (> 5.0 gL–1 ) cannot be suitably treated by AOTs because they would require large amounts of expensive reagents or electrical energy for irradiation [10]. As the total destruction of the pollutant is not always required, AOTs are especially useful in two cases: (a) as a pre-treatment to transform recalcitrant pollutants in more biodegradable compounds; or (b) as a post-treatment, to polish waters before their discharge to the receptor bodies [11]. The main idea of the combination is the use of a more expensive technology only in the ﬁrst or ﬁnal step of the treatment, to reduce costs. PAOTs are developed and commercialized to a variable degree and are undergoing constant change as technological advances take place. At present, UV/H2 O2 , UV/O3 , UV/H2 O2 /O3 , UV/Fenton and UV/TiO2 are totally or partially commercialized.

2 Photochemical Advanced Oxidation Technologies To produce photochemical changes in a molecule, irradiation of light in the UV–visible range must occur within the system. The visible spectrum covers wavelengths between 400 and 800 nm. The UV range is usually divided into four regions, UV-A (also called near-UV light, long-wave light or black-light), UV-B, UV-C (short-UV light) and VUV (vacuum ultraviolet light), as shown in Table 4. Sunlight irradiation may be used in some applications, but it must be taken into account that only 3–5% of UV light is present in the solar spectrum. Usually, light appreciably increases the reaction rate of AOTs in comparison with the same technology in the absence of illumination. As a source of light, high-pressure mercury or xenon arc lamps, with good emission in the near-UV range, can be used. Some applications require short-UV irradiation, as we will see later and, in this case, cheap germicide lamps are easily available. Operative costs are reduced due to a lower power consumption to generate HO• compared to other rather more expensive AOPs such as radiolysis or supercritical technologies. If solar light can be used, a consequent

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

331

Table 4 Regions of the UV-Vis electromagnetic spectrum and their application in Photochemical Advanced Oxidation Technologies for water treatment Type

λ (nm)

Energy (kJ mol–1 ) Uses

UV-A∗

315–400 (365)∗∗ 280–315 190–280 (254, 185) < 190 nm (172)∗∗

380–299 327 427–380 629–427 (471, 646) > 629 695

UV-B UV-C∗ VUV∗ ∗

Almost all photochemical AOTs Some AOTs Disinfection and sterilization, H2 O2 Some applications

Used in environmental applications ∗∗ The most used wavelength

saving of electrical power will be produced, with safer industrial installations. As the light is totally directed to the system, the photochemical industrial equipment used is more compact, and smaller tanks can be employed. As we will see later, the use of light increases the ﬂexibility of the system, allowing the use of a variety of oxidants and operational conditions. Another advantage of the photochemical technologies is that pH changes in the efﬂuents need not be as drastic as for example with alkaline ozonation. It is worthwhile to point out, however, that light-mediated AOPs, especially the homogeneous processes, are not adequate for treating mixtures of substances of high absorbance, or containing high amounts of solids in suspension, because the quantum efﬁciency decreases through loss of light, dispersion and/or by competitive light absorption. 2.1 Direct Photolysis It is possible to use a direct photolytic process for the treatment of waters and efﬂuents, without the addition of chemical reagents. It is worthwhile to bear in mind that, for example, a 254-nm photon is equivalent to 4.89 eV, enough energy to produce homolytic or heterolytic breakages in the molecules. Direct irradiation leads to the promotion of a molecule from the fundamental state to an excited singlet state, which may then intersystem cross to produce triplets. Such excited states can undergo homolysis, heterolysis or photoionization, among other processes. In most cases, homolytic rupture produces radicals: (3) R – R + hν → R – R∗ → 2R• These radicals initiate chain reactions to produce the ﬁnal low-weight products. In the presence of oxygen, additional reactions generating the superoxide radical are possible: (4) R – R∗ + O2 → R – R•+ + O2 •–

332

M.I. Litter

Although its oxidizing power is not very high, the superoxide radical is able to degrade substituted aromatic compounds with high absorption in the UV range. Direct photolysis is important for compounds that react very slowly with HO• or do not react at all, for example nitrophenols, NO2 – , and halogenated compounds. Some pesticides can be degraded by direct short-UV photolysis with good yield [12]. Degradation of trihalomethanes (THM), chloromethanes, chloroethanes, chlorinated aromatics and chlorinated phenols by the use of 254 nm irradiation is well documented in the literature [1, 4]. For irradiation at this wavelength, low-pressure mercury germicidal lamps are easy to employ. Irradiation with KrCl excimer lamps (222 nm) is used for chlorinated aliphatics such as CCl4 or 1,1,1-trichloroethane, because the rupture of the C – Cl bond takes place at 210–230 nm. Generally, the technology is combined with other conventional methods. Limitations of the process are: (i) low efﬁciency; (ii) application only to compounds absorbing at 200–300 nm; (iii) only one target compound can be treated with reasonably good results. The mechanism and products of UV radiation decomposition have been described for important pollutants such as DDT, lindane, PCP, TNT and atrazine ([13] and references therein). Direct 254-nm UV photolysis is effective for discoloring textile dyes at low concentrations, as seen in the recently described case of Solophenyl Green BLE 155% [14]. When direct photolysis was compared with other processes as 254-nm UV/TiO2 and combined TiO2 photocatalysis/activated carbon, it was demonstrated that, at low dye concentrations (5–10 mgL–1 ), the photolytic treatment is 2–3 times faster than the other processes for color removal. 2.2 Sensitization In many cases, direct photolysis may be favored in the presence of oxygen and substances which can act as photosensitizers. Sensitizers (Sens) are compounds that absorb visible light and are excited to a higher energy state from which an energy transfer occurs, the excess energy then being transferred to other molecules present in the system [15]. In this sense, some dyes like Rose Bengal (RB), phthalocyanines or methylene blue promote singlet oxygen (1 O2 ) formation in excellent quantum yield [16]; singlet oxygen is an oxidant powerful enough to attack OM and microorganisms [3]: Sens + hν → 1 Sens∗ → 3 Sens 3 Sens + 3 O → Sens + 1 O 2 2 1 O + A → AO 2 2

(5) (6) (7)

For water puriﬁcation, the efﬁciency is strongly dependent on the production rate of singlet oxygen in the aqueous solution.

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

333

Signiﬁcant degradation of compounds can also be obtained in the presence of electron-acceptor sensitizers. In a very recent example, triadimenol, a systemic pesticide widely applied in horticulture and viticulture that is very difﬁcult to degrade by direct UV photolysis, could be signiﬁcantly decomposed in the presence of electron acceptors such as 9,10-dicyanoanthracene or 2,4,6-triphenylpyrylium tetraﬂuoroborate. Decomposition was accelerated by the presence of oxygen [17]. This process has not been commercialized yet; one of the main problems is the necessity of removing the dye from the water after the treatment. For this reason, attempts at immobilization to different supports have been reported recently, but this process leads to a decrease in the efﬁciency of 1 O2 production. For example, when RB is immobilized on a polymer, its efﬁciency is reduced one hundred-fold compared with the sensitizer in a homogeneous water solution [18]. More research is needed to improve this technology, taking into account that the system demonstrates an effective disinfection ability for drinking water. 2.3 Photolysis of Water in the Vacuum Ultraviolet (VUV) This process uses light irradiation of wavelengths lower than the UV-C, i.e., lower than 190 nm. Generally, Xe excimer lamps (λexc = 172 nm) are used. The excitation leads, in the majority of the cases, to the homolytic breakage of chemical bonds, degrading OM in condensed and gaseous phases (for example, ﬂuorinated and chlorinated hydrocarbons) [1, 3]. However, its application is limited, and the most important use of VUV radiation is in water photolysis (Eq. 8): H2 O + hν → HO• + H•

(8)

This process generates hydroxyl radicals and hydrogen atoms in situ, without the addition of external agents1 . Due to the high absorption cross-section of water, the total incident radiation is absorbed within a very narrow layer around the lamp shaft [19]. The quantum yield of reaction 8 depends on the irradiation wavelength, varying between 0.33 at 185 nm and 0.72 at 147 nm [20]. Aqueous electrons (strong reductants) are also produced, but with a lower quantum yield (0.05), almost independent of the irradiation wavelength in the range 160–190 nm [19]. H2 O + hν → HO• + H+ + e– aq

(9)

1 A similar in situ HO• generation can be obtained using high power ultrasound sources or by processes using subcritical or supercritical water (at very high temperatures or pressures). [9] and other references therein and in Table 3 can be consulted.

334

M.I. Litter

Fig. 1 Degradation of atrazine by photolysis of water under VUV irradiation in argon (a), air (b), and oxygen (c); [atrazine]0 = 0.1 mM [23]

In aerated solutions, HO2 • and O2 •– are rapidly generated from the primary active species: O2 + H• → HO2 • O2

+ e–

aq

→ O2

•–

kHO2 • = 1 × 1010 M–1 s–1 kO2

•–

= 2 × 10

10

M–1 s–1

(10) (11)

The generated oxidants (HO• , HO2 • , O2 •– ) and reductants (H• , e– aq , HO2 • , O2 •– ) make possible simultaneous reductions and oxidations in the chemical system. The technology can be used for the degradation of pollutants in water and in a current of air with a high humidity content, for ultrapure water production and for treating oxidizable compounds that are difﬁcult to treat, such as chlorinated and ﬂuorinated hydrocarbons (for instance, ClCH3 ). The process is highly efﬁcient because VUV lamps generally have a high radiant power of illumination and water has a high cross-section of absorption in the wavelength range. This technology does not require the addition of chemical agents, and is simple and competitive. However, it requires an oxygen supply, the use of quartz and high power provisions. The technology has not yet been commercialized, and is presently in the development stage. González and Braun have thoroughly studied various systems submitted to this process, such as nitrate and nitrite photolysis [21, 22] and mineralization of the very resistant pesticide atrazine [23]. The results of this work are shown in Fig. 1.

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

335

2.4 UV/H2 O2 H2 O2 is a weak acid, a powerful oxidant and an unstable compound that disproportionates with a maximal rate at the pH of its pKa : H2 O2 ⇔ HO2 – + H+ pKa = 11.6 – + H2 O2 + 2e + 2H → 2H2 O Eo =+ 1.78 V, pH0 1 H2 O2 → H2 O + O2 2 H2 O2 + HO2 – → H2 O + O2 + HO–

(12) (13) (14) (15)

Hydrogen peroxide has been widely used in the removal of low levels of pollutants from wastewaters (chlorine, nitrites, sulﬁtes, hypochlorites, etc.) and as a disinfectant [24]. However, low reaction rates make its use—at reasonable concentrations—in the treatment of high levels of refractory pollutants, such as highly chlorinated aromatic compounds and some inorganic compounds (e.g. cyanides), ineffective. The oxidizing power of hydrogen peroxide can be sensibly improved by HO• generation through cleavage of the O – O union with photons of enough energy (higher than 213 kJ mol–1 , the energy bond, which corresponds to wavelengths lower than 280 nm). The reaction has a low quantum yield (φHO• = 0.5) due to strong recombination of the radicals in solution [19, 25], and produces almost quantitatively one HO• per quantum of radiation absorbed in the 200–300 nm range: H2 O2 + hν → 2HO•

(16)

H2 O2 photolysis is usually performed with low- or medium-pressure mercury vapor lamps. Almost 50% of the energetic consumption is lost in the form of heat or emissions less than 185 nm, which are absorbed by the quartz jacket. Generally, cheap germicidal lamps are used; however, as H2 O2 absorption is maximal at 220 nm, it is more convenient to use Xe/Hg lamps that—although more expensive—emit in the 210–240 nm range. In addition to H2 O2 (ε = 18.6 M–1 cm–1 at 254 nm), other species can absorb photons at these short wavelengths, and can act as light ﬁlters. However, if the contaminants can be directly photolyzed, this may improve the efﬁciency of the oxidative destruction process. As the intensity of UV radiation decays exponentially towards the bulk of the solution, it is necessary to establish conditions of turbulent ﬂow to continuously renew the solution surrounding the luminous source. In the presence of oxygen, multiple pathways are operative in the UV/H2 O2 system, as shown in Fig. 2 [1]. The photochemical process is more efﬁcient in alkaline media because the concentration of the conjugate anion of hydrogen peroxide increases with pH (reaction 12), and this species has a higher absorption coefﬁcient

336

M.I. Litter

Fig. 2 Sequence of reactions occurring in the UV/H2 O2 system [1]

(ε254 = 240 M–1 cm–1 ) than H2 O2 , favoring light absorption and increasing HO• production [1, 26]. However, a high pH should be avoided because bicarbonate and carbonate ions (coming from the mineralization or present in the waters) are competitive HO• trapping species: HO• + HCO3 – → CO3 •– + H2 O HO• + CO3 2– → CO3 •– + HO–

(17) (18)

Of course, this will occur in every AOT involving HO• in carbonated solutions. As a general rule, pH changes due to mineralization processes should be taken into account in almost all AOTs because they may affect reaction rates [8]. In most degradations performed by UV/H2 O2 , it has been found that the rate is very dependent on the concentration of H2 O2 , increasing to an optimum value, beyond which an inhibitory effect takes place [19, 27]. At high HO• concentrations, competitive reactions occur because these radicals are prone to recombination, regenerating H2 O2 (reverse of reaction 16), or react in accordance with the following scheme [28]: HO• + H2 O2 → HO2 • + H2 O HO2 • + H2 O2 → HO• + H2 O + O2 2HO2 • → H2 O2 + O2 HO2 • + HO• → H2 O + O2

(19) (20) (21) (22)

Reactions 19 and 22 consume HO• and decrease the probability of oxidation. HO2 • radicals are produced through reaction 19, but one must remember that they are much less reactive than HO• . In all cases, it is necessary to determine the optimal H2 O2 concentration, to avoid an excess that could retard the degradation, and this depends on the concentration and chemical nature

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

337

of the pollutants in the efﬂuent stream. Consequently, treatability tests are needed to determine the right amount of H2 O2 and to validate the technology. López et al. [19] were able to determine the optimal concentration of hydrogen peroxide that led to the fastest degradation of 4-chloro-3,5-dinitrobenzoic acid as a function of the initial concentration of the organic compound. If low-pressure mercury vapor lamps are used, a high [H2 O2 ] is needed to generate enough HO• , making the process less effective. To overcome this limitation, high-intensity UV lamps can be employed. The use of UV/peroxide offers some advantages: the oxidant is commercially accessible, thermally stable, and can be stored in the site of use (with the required precautions). As H2 O2 has an inﬁnite solubility in water, it is an effective source of HO• , producing 2HO• per each H2 O2 . There are no mass transfer problems associated with gases, as we will see in the case of ozone. The capital investment is minimal and the operation is simple. In contrast, due to the low H2 O2 cross-section absorption at 254 nm, high concentrations of the oxidant are required, and depletion of the reagent must be controlled throughout the reaction span. The method has a low efﬁciency for treating waters of high absorbance at λ < 300 nm, or containing substances that compete with HO• generation. In these cases, a large amount of H2 O2 is again needed. The UV/H2 O2 technology is one of the oldest AOPs and has been successfully used in the removal of contaminants from industrial efﬂuents, including organochlorinated aliphatics, aromatics, phenols (chlorinated and substituted) and pesticides [1, 8]. It has been considered a very good treatment for the reuse of wastewater from the dye industry [29]. A recent example is the case of Hispamin Black CA, a dye widely used in the Peruvian textile industry [27]. Using UV/H2 O2 , it was possible not only to decolorize but also to mineralize the dye in reasonable reaction times (Fig. 3). A strongly absorbing solution was completely decolorized after 35 min, and an 82.1% reduction of the total organic carbon (TOC) was obtained after 60 min. Care must be taken to control the formation of toxic compounds during the process, as has been observed during the degradation of Remazol BlackB. However, an absence of toxicity was reported as occurring at the end of the process [29]. At present, UV/H2 O2 technology is totally commercialized. The method can be sensibly improved by combination with ultrasound [30] or by pretreatment with ozone [31]. The combination UV/H2 O2 /O3 has also been proposed, as we will see later. Recently, the degradation kinetics of two pharmaceutical intermediates [5-methyl-1,3,4-thiadiazole-2-methylthio (MMTD-Me) and 5-methyl-1,3,4thiadiazole-2-thiol (MMTD)] has been studied in order to assess the effectiveness and the feasibility of UV processes. For both substrates, the results showed that no degradation occurred when H2 O2 was used alone and that UV and UV/H2 O2 processes were both effective for degrading the substrates, but

338

M.I. Litter

Fig. 3 Treatment of Hispamin Black CA by UV(366 nm)/peroxide: (a) Variation of the normalized absorption at 471 nm with irradiation time under UV irradiation and under UV irradiation in the presence of hydrogen peroxide; (b) variation of TOC during reaction with UV/peroxide. Conditions: [Hispamin Black CA] = 40 mg L–1 , pH 7.5, [H2 O2 ] = 565.8 mg L–1 [27]

photo-oxidation was always faster than direct photolysis. The results showed that to remove 99% of some µg L–1 of the pharmaceutical intermediates with a H2 O2 dose of 1 mg L–1 , 55 min for MMTD-Me and 2.6 min for MMTD are necessary, showing the feasibility of the decontamination process suggested in this study [32].

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

339

2.5 UV/O3 2.5.1 Thermal Ozonation Ozone is a powerful oxidant (see Table 2) and an efﬁcient bactericide. Lately, ozone has been increasingly used for the treatment of drinking water, because the method does not produce THM or other chlorinated compounds that can be generated through disinfection with chlorine or chlorine oxide. The use of ozone allowed a remarkable improvement of organoleptic properties, ﬁltration characteristics and biodegradability of drinking water. Additionally, the use of ozone decomposition by different initiators for the decontamination of water has triggered a study of the different mechanisms taking place in the chemical processes. Ozone is industrially applied for water treatment either alone or in combination with hydrogen peroxide and/or activated carbon. Recent reviews describe improvements of the ozone technology, including combinations with catalysts and AOTs [33, 34]. In the absence of light, ozone can react directly with an organic substrate, through a slow and selective reaction 23, or through a fast and non-selective radical reaction that produces HO• , Eq. 24 [2, 35–37]: O3 + S → Sox

k ≈ 1 – 100 M–1 s–1

2O3 + 2H2 O → 2HO• + O2 + 2HO2 •

(23) 8

10

–1 –1

k ≈ 10 – 10 M s

(24)

As stated earlier, the rate constants of ozone with organic compounds differ greatly for both types of processes (Table 3). The ﬁrst reaction is important in acid media and for solutes that react very fast with ozone such as, for example, unsaturated compounds and compounds containing amine or acid groups. The results support the electrophilic nature of the reaction, either by electrophilic substitution or by dipolar cycloaddition [37]. This route leads to a very limited mineralization of the organic compounds, and its use for the removal of pollutants must be reinforced by modiﬁcation of the method. It has been demonstrated that ozone decomposition in aqueous solution forms HO• , especially when initiated by OH– [10]: O3 + HO– → O2 + HO2 – HO2 – + O3 → O3 •– + HO2 • HO2 • ⇔ O2 •– + H+ O2 •– + O3 → O3 •– + O2 O3 •– + H+ → HO3 • HO3 • → HO• + O2 O3 + HO• ⇔ O2 + HO2 •

(25) (26) (27) (28) (29) (30) (31)

340

M.I. Litter

In some cases, singlet oxygen is formed when ozone reacts by O-atom transfer, for example with sulﬁdes, disulﬁdes, methanesulﬁnic acid, nitrite, etc. A detailed description of these results is beyond the scope of this paper; for more information see [38]. Ozonation follows a rapid zero-order initial step, limited by the mass transfer of the gas to water. A second step takes place when the aqueous medium is saturated by ozone, and the rate of this step is limited by slower reaction pathways [39]. The increase of the ozone dose plays a relevant role in enhancing the reaction efﬁciency. Typical ozone doses are 3–15 mg L–1 , depending on the initial concentration of the target compound. The indirect pathway is less selective, because the species formed in the process have a higher oxidant ability than the ozone itself, especially HO• . The route can be initiated in different ways, by HO2 – , HCOO– , Fe2+ , humic substances or principally by HO– . This is why, in principle, ozonation is more efﬁcient in alkaline media, presenting an optimum around pH 9. Figure 4 shows a scheme of the main species of ozone decomposition in pure water initiated by hydroxide ions [7]. The addition of Fe(II), Mn(II), Ni(II), Co(II) or Ag(I) salts as well as solid oxides such as Fe(III)/Al2 O3 , goethite, MnO2 , TiO2 , Cu/Al2 O3 or Cu/TiO2 (Catazone process) improve the technology [37] (see Sect. 2.6.6). The combination of both direct and indirect routes enhances sensibly OM degradation. This obviously depends on the composition and pH of the solution, and on the ozone dose. The pH should be carefully controlled due to the already mentioned HO• scavenging action of bicarbonate and carbonate ions produced as mineralization takes place (reactions 17 and 18). Intermedi-

Fig. 4 Scheme of the main species of ozone decomposition in pure water initiated by hydroxide ions [7]

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

341

ate oxidation products, like acetic and oxalic acids, are refractory compounds that often resist mineralization. In a relatively recent example, total depletion of 5 × 10–6 M nitrobenzene (NB) and 2,6-dinitrotoluene (2,6-DNT), model compounds of nitroaromatic hydrocarbons, could be accomplished with ozone in 10 and 40 min respectively at neutral or weakly basic pH. The rate constants of the direct reaction between ozone and NB or DNT are very low, indicating that the process develops in these cases more through hydroxyl radicals than through the direct reaction [40]. Ozonation is a very well-known commercialized technology for water treatment. It has been successfully used in the discoloration of kaolin and cellulose pulp and, in general, in the treatment of extremely polluted aqueous efﬂuents. It must be highlighted that ozone is transformed merely into O2 and H2 O, making the method less toxic when compared with other conventional treatments that use Cl2 or chromic acid. Ozonation is a good pre-treatment to a biological treatment, because complex organics are transformed into aldehydes, ketones or carboxylic acids, all easily biodegradable compounds. Ozonation is versatile and can be combined with other conventional or Advanced Oxidation Technologies. Ozone can be simply produced in situ by electric discharge in a current of oxygen or air, leaving neither odors nor residual tastes. In contrast, from the operational point of view, the use of ozone is not as trivial as the use of a totally water miscible oxidant such as hydrogen peroxide, and there are mass transfer limitations due to the difﬁcult access of the gaseous molecule to the aqueous phase [41]. Consequently, the process requires efﬁcient stirring, the use of line mixers, venturis, contact towers, etc. To improve the process, another possibility is to increase the retention time in the reactor by large bubble columns or to increase the solubility of ozone by increasing the pressure to several atmospheres. However, any additional modiﬁcation adds high investment costs. Furthermore, a rather high O3 /pollutant molar ratio (more than 5 : 1) is generally needed for the complete destruction of a compound, which makes the treatment even more expensive. As an additional drawback, in some cases, the method does not lead to complete mineralization. Care must be taken to control the temperature, because of the risk of volatilization of initial or intermediate compounds. Final degassing devices in the circuit are necessary to completely deplete ozone, which will be deleterious in a possible biological post-treatment; this also increases the costs. 2.5.2 O3 /H2 O2 The addition of hydrogen peroxide to the ozonation system provides a better result [42]. The process, called Perozone, combines the direct and indirect ozone oxidation of organic compounds. H2 O2 initiates O3 decomposition by

342

M.I. Litter

electron transfer [2]; alternatively, the reaction can be envisaged as the activation of H2 O2 by ozone. The set of reactions already seen (27 to 31) is initiated by the HO• producing reaction 32 [43]: O3 + H2 O2 → HO• + O2 + HO2 •

(32)

The process is expensive but fast, and can treat organic pollutants at very low concentrations (ppb), at a pH between 7 and 8; the optimal O3 /H2 O2 molar ratio is ∼ = 2 : 1. It has been suggested that the acceleration of ozonation is due to the fact that H2 O2 increases ozone transfer within water [44]. The treatment is effective for decomposing organochlorinated compounds such as trichloroethylene (TCE), tetrachloroethylene, etc. It is excellent for the post-treatment of water submitted to disinfection treatments with chlorine or chlorine dioxide because it can decompose THM or related compounds. One of the principal ﬁelds of application is in the degradation of pesticides [45]. 2.5.3 Photoinduced Ozonation The UV irradiation of ozone in water produces H2 O2 quantitatively: O3 + hν + H2 O → H2 O2 + O2

(33)

The generated hydrogen peroxide is photolyzed (see Eq. 16), generating HO• radicals, and also reacts with the excess of ozone, according to Eq. 32. This method might be considered in principle as just an expensive way of generating H2 O2 and then HO• . Indeed, it is a combination of UV/H2 O2 and O3 /H2 O2 , but the advantage is that the ozone has a higher absorption coefﬁcient than H2 O2 (ε254 = 3300 M–1 cm–1 ), and can be used to treat water with a high UV absorption background. The efﬁciency is higher than that of O3 or direct UV, and the reactor does not need to be in quartz because UV-B light (280–315 nm) can be used. The method has been applied to potable water, to treat highly contaminated wastewater, in disinfection, in discoloration of waters from the paper industry, in the degradation of chlorinated aliphatic hydrocarbons (saturated and unsaturated), etc. In [43], the ﬁrst applications of the technology are mentioned. If wavelengths lower than 300 nm are used, photolysis of O3 takes place, generating additional HO• and other oxidants, with a subsequent increase in the efﬁciency [46]: O3 + hν → O2 (1 ∆g) + O(1 D) O(1 D) + H2 O → 2HO•

(34) (35)

Gurol and Akata [43] studied the kinetics of ozone photolysis following a conceptual model based on possible reaction pathways. They obtained experimentally the primary quantum yield of ozone photolysis at 254 nm (0.48).

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

343

The rate of ozone photolysis increased with increasing light intensity, ozone concentration and pH, and decreased with increasing inorganic carbon concentration. As the formation of HO• is tied to ozone decomposition, this model can be extended to predict the oxidation rates of water contaminants by HO• generated in the process. Generally, an increase of the ozone concentration increases the degradation rate of the pollutant, as demonstrated for the case of atrazine [47]. Although direct ozonation can contribute, 87% of the oxidation process proceeds, in the atrazine case, through the radical pathway. In contrast with the results in the absence of light, alkaline pH reduces the reaction rate, as has been observed in the case of 2,6-DNT degradation. The decrease of the rate is due to the dissociation of the hydroxyl radical in the less active oxygen anion radical (Eq. 36) and to the lower solubility of ozone at high pH [40]. HO• → O•– + H+

(36)

Although ozonation is improved under UV light, it was found that the use of high initial concentrations of ozone (1000 mg L–1 ) (without irradiation) was more effective than the combination UV/O3 to treat formulated pesticides like atrazine, alachlor, carbofuran, etc., because of the presence of large amounts of hydroxyl radical scavengers in the formulations [39]. It was recently demonstrated that solar light is also valuable for enhancing ozonation, as proved in the degradation of two model organic compounds, phenol and malic acid. This process has been called Heliozon. The rates of OM removal were also higher and faster, and complete mineralization was achieved even at high initial TOC values (as high as 49 000 ppm). This provides a possible way of increasing ozone reactivity at low cost. The simultaneous presence of sunlight and Fe(II) in solution also produced a beneﬁcial effect in the mineralization; this was, however, less effective with other metal ions like Cu(II), Ni(II), Mn(II) and Co(II) [49]. Ozonation is greatly improved when UV irradiation is combined with a heterogeneous photocatalyst such as TiO2 (see Sect. 2.8). 2.5.4 UV/O3 /H2 O2 The addition of light to the H2 O2 /O3 process produces a net increase in the efﬁciency. The thermal process is accelerated, especially the very slow reaction (32). The three separate processes, UV/H2 O2 , UV/O3 and UV/H2 O2 /O3 , have been shown to be very effective for the decontamination of groundwater and for soil remediation [2, 3]. In contrast to UV/O3 and UV/H2 O2 technologies, which are commercially available [3], UV/H2 O2 /O3 application studies are at present only at the pilot plant scale.

344

M.I. Litter

2.6 Photo-Fenton and Related Reactions 2.6.1 Fenton Reaction Fenton’s well-known experiments at the end of the 19th century demonstrated that solutions of hydrogen peroxide and ferrous salts were able to oxidize tartaric and malic acids as well as other organic compounds [50]. Haber and Weiss suggested later that HO• was formed through reaction (37) [2, 51]: Fe2+ aq + H2 O2 → Fe3+ aq + HO– + HO•

(37)

The attack of HO• radicals on OM was proposed, in principle, as the oxidizing pathway according to Eqs. 1–3 (see however Sect. 2.6.9). Radicals produced by these processes can be additionally oxidized by Fe3+ , reduced by Fe2+ or dimerized, according to the following sequence [24, 52]: R• + Fe3+ → R+ + Fe2+ R• + Fe2+ → R– + Fe3+ 2R• → R – R

(38) (39) (40)

HO• can also oxidize Fe2+ , leading to the following unproductive reaction: Fe2+ aq + HO• → Fe3+ aq + HO–

(41)

At pH < 3, the reaction system is autocatalytic, because Fe3+ decomposes H2 O2 in O2 and H2 O through a chain mechanism [51, 53–57]: Fe3+ + H2 O2 ⇔ Fe – OOH2+ + H+ Fe – OOH2+ → HO2 • + Fe2+ HO2 • + Fe2+ → Fe3+ + HO2 – HO2 • + Fe3+ → Fe2+ + O2 + H+ HO• + H2 O2 → HO2 • + H2 O

(42) (43) (44) (45) (46)

As can be seen, the process can be initiated by Fe3+ , and it is then known as Fenton-like or as a Fenton-type process. This reaction is, however, slow and, as stated, HO2 • is much less reactive than HO• . The Cu(II)/Cu(I) couple can play the same role as the Fe(III)/Fe(II) couple. The Fenton process is very effective for HO• generation, but an excess of 2+ Fe , H2 O2 , hydroperoxyl radicals or halogens (if present) can act as HO• scavengers. In the presence of an excess of peroxide, the Fe2+ concentration is small compared with that of Fe3+ , because Fe2+ is quickly oxidized to Fe3+ (in seconds or minutes)[53]. It is believed that the destruction of wastes by the

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

345

Fenton reagent is simply due to the catalytic cycle of H2 O2 decomposition, a process that generates HO• radicals. Generally, the degree and rate of total mineralization are independent on the initial oxidation state of iron. Conversely, the efﬁciency and the initial mineralization rate are higher when starting from Fe2+ ; as a counterpart, Fe3+ salts produce a stationary Fe2+ concentration. The application of the Fenton process to the destruction of toxic organic material began in 1960 [2]. In [58], several applications have been reported, including a large pilot plant for wastewater treatment. The method is effective for the treatment of water from the manufacturing or processing of chemicals, pharmaceuticals, insecticides, from petroleum reﬁneries and fuel terminals, for color removal in efﬂuents from the dye industry [59], for explosives such as trinitrotoluene (TNT), etc. The Fenton process degrades chlorinated aliphatic and aromatic compounds, polychlorinated biphenyls (PCBs), nitroaromatics, azo dyes, chlorobenzene, pentachlorophenol, phenols, chlorinated phenols, octachloro-p-dioxine, formaldehyde and many others. Compounds that cannot be attacked by this reaction are few but include acetone, acetic acid, oxalic acid, parafﬁns and organochlorinated compounds [60]. It has been successfully applied in the COD reduction of municipal and groundwaters and in the treatment of lixiviates. It is useful as a pre-treatment for non-biodegradable compounds [55]. Recently, it has begun to be effectively applied to the treatment of soils as a good oxidant of herbicides and other contaminants such as hexadecane or Dieldrin (see for example [61, 62]). The advantages of the method are various: Fe2+ is abundant and nontoxic, hydrogen peroxide can be easily handled and it is an environmentally friendly compound. No chlorinated compounds are formed as in other oxidative techniques, and there are no mass transfer limitations because all of the reagents are in solution. The design of reactors for technological application is rather simple [60]. At variance, it requires a high iron concentration and the continuous or intermittent addition of H2 O2 and Fe2+ . However, one must remember that an excess of both reagents, Fe2+ and H2 O2 , cause HO• trapping. Although the degradation rate increases with Fe2+ concentration, no effect is observed above a certain value; oppositely, a large amount should be avoided because it contributes to an increase in the content of total dissolved salts in the efﬂuent stream [8]. Generally, the reaction rate is high until full H2 O2 depletion. Theoretically, the H2 O2 /substrate molar ratio needed for destruction of soluble compounds oscillates between 2 and 10. However, in practice, this ratio may be sometimes as high as 1000, because in environmental samples there are usually other HO• competing species. Obviously, hydrogen peroxide must be completely eliminated before passing the efﬂuent on for biological treatment [8]. The maximum catalytic activity of the Fe(II)/Fe(III) – H2 O2 system is at a pH of about 2.8–3.0. At pH > 5, particulate Fe(III) is generated and at a lower pH, the complexation of Fe(III) with H2 O2 (reaction 42) is inhib-

346

M.I. Litter

Fig. 5 Oxidation of chlorobenzene by Fenton reagent (excess H2 O2 ). Initial conditions: [chlorobenzene] = 1.6 mM; [Fe2+ ] = 5.0 mM; pH 3.0 [64]

ited [55]; therefore, the pH must be kept constant. The type of buffer used also has an effect on the degradation efﬁciency. Phosphate and sulfate buffers are the worst, probably due to formation of stable Fe(III) complexes, which decreases the concentration of free iron species in solution and inhibits the formation of free radicals [53]. At the end of the process, even though this means further management of the generated sludge, it is common to alkalinize the waters, with the simultaneous addition of a ﬂocculant to eliminate the remaining iron. In the laboratory, the metal is traditionally added as pure ferrous salts, but at a larger scale, the use of these salts becomes prohibitively expensive, and normally Fe(NH4 )2 (SO4 )2 , which contains 20% of active iron, is used. Other iron compounds have been employed, including solids such as goethite, which has been used, for example, for TCE destruction [63]. In Fenton reactions, complete mineralization cannot generally be achieved; resistant intermediates such as carboxylic acids, which react very slowly with HO• , are formed, with the unproductive reaction (41) predominating. Sometimes, as Fig. 5 shows, products more toxic than the initial ones—a quinone in this example—can be formed, whose presence must be carefully monitored until total depletion [64]. 2.6.2 The Photo-Fenton Reaction and Other Iron-Based Photoprocesses As mentioned in the previous section, Fenton processes do not generally lead to mineralization, the recycling of Fe2+ is slow, and a scavenging of HO• or

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

347

other competitive reactions take place. Complexation of Fe(III) with organic compounds in the system, mainly carboxylic acids, leads to the formation of very stable iron (III) compounds, whose further oxidation and mineralization is difﬁcult. One example is oxalate, a common intermediate in many oxidative degradations. The photochemistry of the Fe(III) species in solution is a very common process in natural waters, and can also be of interest for use in oxidation processes for water treatment. The chemistry implicated in photoinduced processes of Fe(III)-complexes has been recently reviewed, together with the degradation of organic compounds in aqueous solutions initiated by them [65]. Fe(III)-hydroxocomplexes undergo photochemical reduction to Fe(II) under UV irradiation, and Fe(II) is reoxidized by oxidants like dissolved oxygen, giving rise to the basic Fe(III)/Fe(II) redox cycle. Fe(III)(OH)2+ is the dominant complex from pH 2.5 to 5 and it absorbs light in the UV range with higher absorption coefﬁcients than that of aqueous Fe3+ . Its photolysis leads to Fe(II) and HO• , as shown in Eq. 47. The quantum yield of this reaction is low and depends on the irradiation wavelength, but it is higher than that of other Fe(III)-aquo or -hydroxo species in solution [65]: Fe(III)(OH)2+ + hν → Fe(II) + HO•

(47)

A set of recent results [66–74] shows that the iron(III)-photoinduced degradation by itself is a homogenous photocatalytic process, efﬁcient under solar light and useful to be employed in decontamination systems. It can be used as a physicochemical pre-treatment to transform biorecalcitrant pollutants or as a complete treatment leading to mineralization. The advantage of this process is that it only needs the addition of iron at low concentrations, compatible with the environment (ca. 5 ppm). The process must be rationalized as follows: 1. If there is no interaction between Fe(III) and the pollutant, Fe(III)hydroxocomplexes are the source of HO• , according to reaction (47). The interest of this process resides in its catalytic aspect. HO• radicals react with iron (II) at a high rate, according to reaction (41), which allows the regeneration of the absorbing species. The aqueous FeOH2+ complex plays a fundamental role in this process. The efﬁciency of this system in degrading benzene, phenols, chloro-organic carboxylates and triazines was tested under either UV or solar light (for a list of references, see [65]). 2. If the pollutant is a carboxylic acid such as oxalic acid or others used in the formulation of detergents (ethylenediaminetetraacetic acid (EDTA), nitrilotriacetic acid (NTA), etc.), Fe(III) forms stable complexes or associated ionic pairs that exhibit ligand-to-metal charge transfer (LMCT) bands in the UV-Vis spectrum; these complexes are, in general, photochemically active and, under irradiation, they generate Fe(II): Fe(III)(O2 CR)2+ + hν → Fe(II) + CO2 + R•

(48)

348

M.I. Litter

Light irradiation of these complexes can be used to promote an important enhancement of the degradation of organic compounds [53–56, 75]. We will return to ferrioxalate in Sect. 2.6.3. Recently, Park and Choi [76] reported the visible light (λ ≥ 420 nm) and Fe(III)- mediated discoloration of Acid Orange 7 (AO7) in the absence of H2 O2 . Ferric ions form complexes with AO7 mainly through the azo chromophoric group of the dye and, under irradiation, production of ferrous ions accompanies AO7 photodegradation. The reaction was not inhibited in the presence of an excess of an HO• scavenger (2-propanol), which indicated that HO• radicals were not responsible for the dye degradation. From the evidence that addition of an excess of sulﬁtes and sulfates, which inhibit complex formation, decreased the photodegradation efﬁciency, it was suggested that the actual active species was the Fe(III)-AO7 complex. Although the process does not reduce TOC concentration, it does not require hydrogen peroxide addition and it can be proposed as an economically viable method to pre-treat or decolorize azo dye wastewaters using sunlight. The above-mentioned processes in the absence of H2 O2 also take place in the presence of the oxidant, making both Fenton and Fenton-like reactions more efﬁcient due to radical generation through Eqs. 47 and 48 and iron recycling. In these photo-Fenton processes, wavelengths from 300 nm up to the visible can be used, in contrast to UV/H2 O2 , which needs short-UV light. As expected, irradiation under 360 nm produces H2 O2 photolysis (Eq. 16), yielding also HO• . However, as in the case of thermal Fenton systems, H2 O2 must be continuously added and acid conditions are needed. Iron concentrations can be orders of magnitude lower than in the conventional Fenton reaction; either Fe3+ or Fe2+ can be used, in the 5–15 mg L–1 range, supplied as FeSO4 , Fe(ClO4 )3 or FeCl3 . Iron salts must be eliminated after the treatment by neutralization and precipitation of Fe(OH)3 , as in classic Fenton processes. The most frequent use of the photo-Fenton technology has been the treatment of industrial waters and lixiviates. Nitroaromatics, polychlorinated phenols, herbicides (2,4,5-trichlorophenoxyacetic (2,4,5-T), 2,4dichlorophenoxyacetic acid (2,4-D)) and pesticides have been successfully degraded [39]. When comparing different technologies, photo-Fenton is generally the most efﬁcient. An interesting example is the comparative efﬁciency of three different AOP systems, direct photolysis, UV/H2 O2 and UV/Fenton reagent, for the degradation of 2,4-dinitrotoluene (2,4-DNT, 100 ppm) [77]. While direct photolysis resulted in incomplete and slow 2,4-DNT decomposition, UV/H2 O2 was faster (98% degradation in 60 min, 88 mM optimal H2 O2 concentration). However, 94% TOC reduction after 2 h and complete mineralization after 60 min occurred with the Fenton reagent (3 : 1 H2 O2 to FeSO4 .7H2 O molar ratio), while 96% TOC reduction after 2 h was observed with UV photoFenton oxidation using a 125W UV lamp and the same ratio of reagents.

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

349

One practical use of Fenton and photo-Fenton processes is the removal of natural organic matter (NOM) from organic rich waters before the chlorine disinfection of drinking water. It was observed that, under optimal conditions, both processes achieved more than 90% TOC removal, leading to the potential formation of trihalomethanes at concentrations below 10 µg L–1 , well under UK and US standards [78]. 2.6.3 Photo-Ferrioxalate and Other Fe(III) Complexes Oxalic acid forms complexes with Fe(III) that absorb strongly from 254 to 442 nm. The absorption corresponds to a LMCT band, with εmax values around 103 –104 M–1 cm–1 . Photolysis of trisoxalatoferrate(III) (ferrioxalate, FeOx) constitutes the most used chemical actinometer; the quantum yield of Fe2+ formation is high (φ = 1.0 – 1.2) and almost independent of the wavelength [79]. If H2 O2 is added, the photochemical reduction of the Fe(III)-complex will be coupled to a Fenton reaction (Eq. 37) [56, 80]. Thus, the use of illuminated mixtures of H2 O2 and FeOx is very efﬁcient for the photodegradation of organic contaminants: the energy required to treat the same volume of a selected wastewater is ca. 20% of the energy required by the common photoFenton system [56, 81, 82]. The main reactions in the photo/FeOx/H2 O2 system are described by the following sequence of reactions [83]. After light absorption, oxalyl radical (C2 O4 •– ) is produced through a LMCT: Fe(C2 O4 )3 3– + hν → Fe2+ + 2C2 O4 2– + C2 O4 •–

(49)

Then, a rapid decarboxylation takes place from the oxalyl radical: C2 O4 •– → CO2 •– + 2CO2

(50)

The fate of CO2 •– depends on the competitive reactions between dissolved oxygen and ferrioxalate: CO2 •– + O2 → O2 •– + CO2 CO2 •– + Fe(C2 O4 )3 3– → Fe2+ + 3C2 O4 2– + CO2

(51) (52)

The superoxide radical (or its conjugate acid) has three reaction pathways, depending on the oxidation state of iron or the H2 O2 concentration and the pH: HO2 • (or O2 •– ) + Fe2+ + H+ (2H+ ) → Fe3+ + H2 O2 HO2 • (or O2 •– ) + Fe(C2 O4 )3 3– → Fe2+ + 3C2 O4 2– + O2 + H+ HO2 • (or O2 •– ) + Fe(OH)2+ → Fe2+ + O2 + H2 O

(53) (54) (55)

350

M.I. Litter

Reaction (54) is the predominant one at high [H2 O2 ] (in the mM range) and acid pH, while at low [H2 O2 ] (µM range), reaction (53) is the preferred path. HO2 • can produce H2 O2 and O2 by disproportionation: HO2 • + HO2 • (or O2 •– + H+ ) → H2 O2 + O2

(56)

Then, the Fenton reaction (37) takes place. The method is useful to treat waters with high absorbance at λ < 300 nm, because of the high ferrioxalate absorption cross-section in the 200 to 400 nm range. Solar light can be used, and that makes the technology very attractive from the economical point of view. As said, the energy required to treat the same volume of a wastewater is about 20% of the energy required by the photo-Fenton system [56], and this high efﬁciency is attributed to the broad range of absorbance of the reagent, and the high quantum yield of Fe2+ formation. The reagents are totally water soluble, and there are no mass transfer limitations. The process is cheap and the oxidant is accessible. Ferrioxalate technology has been used for the treatment of aromatic and chloroaromatic hydrocarbons, chlorinated ethylenes, ethers, alcohols, ketones and other compounds. Figure 6 compares the destruction of 2-butanone by three different AOTs, and shows the high efﬁciency of ferrioxalate. Figure 7 compares solar photo-Fenton and solar FeOx for toluene treatment. Nevertheless, it must be stated that total mineralization is seldom attained and that the contaminants are only transformed into other organic compounds. Aromatic pollutants producing hydroxyderivatives as intermediates that strongly absorb in the same UV range as H2 O2 and Fe3+ , present a low rate of destruction [10].

Fig. 6 Destruction of 2-butanone in a contaminated groundwater with different UV treatment processes [55]

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

351

Fig. 7 Comparison of toluene destruction in a polluted groundwater by solar irradiation in the presence of ferrioxalate/H2O2 and Fe(III)/H2 O2 [55]

A recent paper [84] presents a very complete study of the inﬂuence of different operational parameters on the FeOx process, such as light intensity, concentration of the reagents, and the presence of anions and HO• scavengers. The case study was the herbicide 2,4-D. It was demonstrated that the system presented a higher efﬁciency than the photo-Fenton process, that the removal rate increased with light intensity and that ferrioxalate concentration determined the light absorption fraction, then controlling the removal rate. Iron carboxylates other than oxalate were tested [81, 85]. For example, an enhancement of the TiO2 -photocatalytic degradation of 4-chlorophenol (4CP) was found when Fe(III)-NTA was added, this effect being larger than that in the presence of non-complexed Fe(III) [86]. However, when the carboxylate is itself the target pollutant, the addition of oxalate only causes a competition for Fe(III). This has been observed when EDTA degradation (in the mM range) at pH 3 was treated with the FeOx/H2 O2 process under solar irradiation. A rapid TOC removal was attained in all cases, reaching almost 100% after 1 h solar exposure under the best conditions. However, the extent of degradation was found to decrease at high ferrioxalate concentrations, probably because of the competition of oxalate with EDTA or its degradation products. In the absence of oxalate, EDTA could also be degraded to a reasonably good extent, with a TOC removal only slightly lower than that obtained when using ferrioxalate; this constitutes a good advantage from the economical point of view [87].

352

M.I. Litter

2.6.4 SORAS Technology The Solar Oxidation and Removal of Arsenic (SORAS) method is a very simple process in which As can be removed in the presence of iron and citric acid; the technology has been applied with relatively good success in the poorer regions of the planet such as in Bangladesh, India and other countries [88–90]. Water contained in transparent polyethyleneterephthalate (PET) bottles, to which some drops of lemon or lime juice have been added, are irradiated with sunlight for a few hours. Generally, natural water contains an amount of iron salts or they are intentionally added in proper quantities to the water. Although As(III) is partly oxidized in the dark by the addition of Fe(II) to aerated water, presumably by reactive intermediates formed in the reduction of oxygen by Fe(II), over 90% of As(III) can be oxidized photochemically after 2–3 h solar illumination. In the SORAS process, where Fe(III) citrate complexes participate, Fenton-like reactions strongly accelerate As(III) oxidation. The resulting As(V) is adsorbed or incorporated into the precipitating solid in a better way than As(III); clear water is then obtained by decantation or ﬁltration. Topics that have been recently explored include the way in which the nature of the solids formed under solar irradiation differ from those obtained by the normal hydrolysis of Fe(III) salts, and how the presence of complexing agents such as citrate inﬂuence the nature of solids formed by oxidative hydrolysis. It was concluded that the role of solar energy is to direct the pathway of the formation of solids towards structures that are adequate for As(V) uptake, and to achieve these reactions in time spans that permit coupling with the photocatalyzed oxidation of As(III) [91]. 2.6.5 Zero-Valent Iron The use of zero-valent iron (Fe(0)) as a reducing agent to treat compounds recalcitrant to oxidative treatments (e.g., halogenated oleﬁns such as TCE to ethylene) is an emerging technology, which can also convert metal ions (for example Cr(VI) to Cr(III)). For details of this promising new technology, see [92] and references therein. The combined action of UV and Fe(0) or H2 O2 and Fe(0) has been assessed in these systems, and this actually transforms the technology into a Fentonbased process. The role of UV light is to affect Fe(0) dissolution. Recent examples are the enhancement of atrazine degradation [93] and the improvement of discoloration of three reactive dyes, C.I. reactive red 2, C.I. reactive blue 4 and C.I. reactive black 8, using Fe(0) and 254-nm UV irradiation [94]. However, information is still rare regarding the effects of ultraviolet light on the zero-valent iron system. In the case of nitrate reduction by Fe(0), a detrimental effect of 254-nm irradiation on ferrous ion dissolution and ni-

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

353

trate removal was reported. It seems that the role of UV light is strongly dependent on the solution composition [95]. These processes deserve profound further research. 2.6.6 Photo-Fenton and Ozonation The combination of photo-Fenton and ozonation results in an important enhancement of the destruction efﬁciency of organic compounds like phenol [96], 2,4-D [97], aniline or 2,4-chlorophenol ([33] and references therein). As mentioned in Sect. 2.5.1, metal ions catalyze ozone decomposition. In the dark, Fe(II) catalyzes O3 degradation giving the ferryl intermediate (FeO2+ , see Sect. 2.6.9), which can directly oxidize the organic pollutant or evolve to a hydroxyl radical: Fe2+ + O3 → FeO2+ + O2 FeO2+ + H2 O → Fe3+ + HO• + HO–

(57) (58)

The combination of ozone with UV light and iron as the catalyst improves the oxidative capability of the system due to regeneration of Fe(III). In the presence of UV light, Fe(III) ions can be reduced to Fe(II) by a photo-Fenton process, closing a loop mechanism where Fe species act as catalysts while generating additional HO• and ferryl radicals. Irradiation with UV light also causes HO• generation by the direct UV/O3 pathway and photo-Fenton reactions. The interaction of Fe(III) and ligand species in solution, which ends in photochemical active complexes, can also take place in these complex systems. 2.6.7 Photoelectro-Fenton The photoelectro-Fenton method [98] complements the photo-Fenton and electro-Fenton reactions. In the latter, a potential is applied between two electrodes immersed in a solution containing Fenton reagent and the target compound. The recent study of the herbicide 2,4,5-T, performed in an undivided cell with a Pt anode and an O2 -diffusion cathode, showed that the photoelectrochemical process was more powerful than the electro-Fenton process, which can yield only about 60–65% of decontamination. The electro-Fenton method provides complete destruction of all reaction intermediates, except oxalic acid, which, as already mentioned, forms stable complexes with Fe3+ that remain in the solution. The fast photodecarboxylation of such Fe(III)oxalate complexes by UV light explains the highest oxidative ability of the photoelectro-Fenton treatment, which allows a fast and total mineralization of highly concentrated acidic aqueous solutions of 2,4,5-T at low current and temperature. A similar behavior was found for the herbicide 3,6-dichloro-2methoxybenzoic acid [99].

354

M.I. Litter

2.6.8 Immobilized Photo-Fenton Systems As has been already mentioned, homogeneous Fe3+ /H2 O2 reactions need up to 50–80 ppm of iron ions in solution, a value far above the established regulations in industrial countries (around 2 ppm). Moreover, Fenton systems require working at acid pH to avoid iron precipitation, but additional alkalinization and redissolution steps, required to eliminate and recover iron, elevate the costs of the process. To avoid these drawbacks, supported Fenton catalysts, with iron containing membranes or beads, have been developed in recent years. The supporting material needs to be a good complexing agent for Fe2+ and Fe3+ , stable in aqueous solution, resistant to oxidative conditions and transparent to UV/Vis radiation. In this sense, Fecontaining Naﬁon® [100] and perﬂuorinated Naﬁon® membranes [101, 102] were reported useful in degrading Orange II, 2,4-dichlorophenol and other chlorophenols at a pH between 2.8 and 11 with rates similar and even faster than those of homogeneous photo-Fenton reactions [102]. Naﬁon-silica composites [103], C-Naﬁon structured fabrics [104], polyethylene copolymers [105], alginate gel beads [106], structured silica fabrics [107], brick grain [108], MgO [109], SiO2 [110] and zeolites [111] have also been successfully tested as supports. Another advantage of these systems is that it is possible to work at a pH at which it is not necessary to make a ﬁnal adjustment before a biological post-treatment. Interestingly, when industrial wastewaters were treated with Fe-containing silica fabrics, the ﬁnal BOD5 /TOC ratio was higher than that obtained with a homogeneous photo-Fenton process, indicating a higher biodegradability extent [107]. 2.6.9 Active Species in Fenton and Photo-Fenton Systems Although several studies indicate that HO• is formed in Fenton systems according to Eq. 37 and it is responsible for the efﬁciency of degradative reactions, it is presently believed that other Fe(IV) or Fe(V) species like FeO3+ and ferryl complexes, are also active agents in the processes [53– 55, 58, 112]. For example, Kremer [112] identiﬁed a mixed valence binuclear species, {FeOFe}5+ , and proposed a new mechanism for the Fenton reaction, in which FeO2+ acts as the key intermediate. Bossmann et al. [58] proposed the initial formation of a hydrated Fe(II)– H2 O2 complex, leading to a steady-state concentration of iron(II) bound to H2 O2 , according to: [Fe(OH)(H2 O)5 ]+ + H2 O2 ⇔ [Fe(OH)(H2 O2 )(H2 O)4 ]+ + H2 O

(59)

The authors based their argument on the fact that an outer-sphere electrontransfer reaction between Fe2+ aq and H2 O2 , as indicated in the classical re-

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

355

action (37), is thermodynamically not possible, because the formation of the H2 O2 – species is not favorable. Subsequently, an inner-sphere two-electrontransfer reaction takes place, with the formation of a Fe(IV) complex: [Fe(OH)(H2 O2 )(H2 O)4 ]+ → [Fe(OH)3 (H2 O)4 ]+

(60)

This complex may give rise to HO• and Fe(III): [Fe(OH)3 (H2 O)4 ]+ + H2 O → [Fe(OH)(H2 O)5 ]2+ + HO• + HO–

(61)

Pignatello et al. [54] performed nanosecond laser ﬂash photolysis experiments with a 355-nm laser pulse in the Fe(III)/H2 O2 system in the absence of organics. They observed a broad positive signal in the visible region, indicative of the formation of a light-induced transient. They proposed different possible species in agreement with the observed signal, such as H3 Fe(V)O4 , FeO3+ , FeO2+ or a triplet excited state of the [Fe(III) – OOH]2+ peroxo complex. The decay of this species produces HO2 • radicals and new high valent oxoiron, ferryl-like species, which can be precursors of the Fenton reaction, although their identity remained undetermined: [Fe(III) – OOH]2+ + hν → [Fe(III) – OOH]2+∗ [Fe(III) – OOH]2+∗ → HO2 • + Fe(II) [Fe(III) – OOH]2+∗ → {Fe(III) – O• ↔ Fe(IV) = O} + HO• [Fe(III) – OOH]2+∗ → Fe(V) = O + OH–

(62) (63) (64) (65)

In summary, it could be emphasized that both HO• as well as ferryl species coexist in the Fenton systems; depending on the experimental conditions (type of substrate, iron-H2O2 ratio, presence or addition of scavengers, etc.), one of them will predominate. 2.7 UV/Periodate Periodic acid, H5 IO6 , and periodate, IO4 – , are strong oxidants: H5 IO6 + H+ + 2e– → IO3 – + 3H2 O

E0 =+ 1.60 V

(66)

Irradiation of periodate solutions under short-UV light generates radicals (IO3 • , HO• , IO4 • ) and other oxidative species (IO3 – , HOI, I2 , H2 O2 , O3 ). The oxidation of a system containing this reagent under UV light is less selective but more efﬁcient than other AOTs. The proposed mechanism may be very complex, as illustrated in Fig. 8 [113]. With this technology, a wide variety of compounds at low concentrations can be destroyed. It can be used for discoloration of dye-containing waters and for the treatment of other wastewaters. For improved effectiveness, waters should have a low absorbance. So far, there are no legislated discharge requirements for iodine compounds, from which I2 and I– are the

356

M.I. Litter

Fig. 8 Possible reduction pathway of periodate to iodide based on radiolysis studies and UV irradiation of iodine species [113]

more toxic (but still of low toxicity). Iodine can be recovered by ionic exchange, and periodate can be electrochemically regenerated. For example, the treatment of a real wastewater of high COD containing triethanolamine with UV/periodate reduces COD to acceptable values in relatively short times. The technology is faster than other photochemical AOTs and seems very promising, although there are no more recent references in the literature concerning its use. 2.8 Heterogeneous Photocatalysis This AOT will be discussed later in this book; therefore, only a brief introduction is included here. Heterogeneous photocatalysis is a process based on the direct or indirect absorption of visible or UV radiant energy by a solid, normally a wide-band semiconductor. In the interfacial region between the excited solid and the solution, destruction or removal of contaminants takes place, with no chemical change in the catalyst. Figure 9 shows a scheme of processes occurring in a particle of semiconductor when it is excited by light of energy higher than that of the band gap. Under these conditions, electron-hole pairs are created, whose lifetime is in the nanosecond range; during this time interval, electrons and holes migrate to the surface and react with adsorbed species, acceptors (A) or donors (D) [114]. Electron-hole pairs that cannot separate and react with surface species, recombine with energy dissipation. The net process is the catalysis of the reaction between the oxidant A and the reductant D (for example, between O2 and OM).

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

357

Fig. 9 Processes occurring in the semiconductor-electrolyte interface under irradiation with light of E > Eg

Various materials are candidates to act as photocatalysts such as, for example, TiO2 , ZnO, CdS, iron oxides, WO3 , ZnS, etc. These materials are economically available, and many of them participate in chemical processes in nature. Besides, most of these materials can be excited with light of a wavelength in the range of the solar spectrum (λ > 310 nm); this increases the interest in the possible use of sunlight. So far, the most investigated photocatalysts are metallic oxides, particularly TiO2 ; this semiconductor presents a high chemical stability and can be used in a wide pH range, being able to produce electronic transitions by light absorption in the near ultraviolet range (UV-A). The driving force of the electron transfer process in the interface is the difference of energy between the levels of the semiconductor and the redox potential of the species close to the particle surface. The thermodynamically possible processes occurring in the interface are represented in Fig. 9: the photogenerated holes give rise to the D → D•+ oxidative reaction while the electrons of the conduction band lead to the A → A•– reductive process. The most common semiconductors present oxidative valence bands (redox potentials from +1 to + 3.5 V) and moderately reductive conduction bands (+ 0.5 to – 1.5 V) [115]. Thus, in the presence of redox species close or adsorbed to the semiconductor particle and under illumination, simultaneous oxidation and reduction reactions can take place in the semiconductor-solution interface. Holes react with adsorbed substances, in particular with adsorbed water or OH– ions, generating HO• radicals and/or other radicals, as in other AOTs. Normally, in environmental applications, the photocatalytic processes take place in aerobic environments, and adsorbed oxygen is the principal electron acceptor species: O2 + e–cb → O2 •–

(67)

358

M.I. Litter

If noble or heavy metal ions are present in solution, they can be reduced by conduction band electrons to a lower oxidation state: Mz+ + n e–cb → M(z–n)+

(68)

Reduction to the zero-valent state or formation of other metal solid phases like oxides, causes the element to deposit onto the semiconductor surface. The efﬁciency of the photocatalytic reaction depends on different factors. One of the most critical aspects is the high probability of electron-hole recombination, which competes with the separation of the photogenerated charges. On the other hand, as there is no physical separation between the anodic reaction site (oxidation by holes) and the cathodic one (reduction by electrons), back reactions can be of importance. The low efﬁciency is one of the most severe limitations of heterogeneous photocatalysis. Heterogeneous photocatalysis over TiO2 can be also combined with other AOTs. For example, addition of Fe(III) and H2 O2 combines UV/TiO2 with photo-Fenton; in this way, the destruction of some resistant pollutants can be improved. For example, EDTA, NTA and other oligocarboxylic acids are more rapidly mineralized in the presence of Fe(III)/H2 O2 than when using TiO2 alone [116–118]. Similarly, the presence of photochemical Fe(III) complexes such as Fe(III)-NTA helps the photocatalytic degradation of 4-CP [86]. In these cases, an important effect of the Fe(III)-complexes formed with the initial compound or with possible degradation intermediates takes place: these complexes can be photolyzed and even photocatalyzed in the reaction medium, generating Fe(II) and other active radical species. Combination of UV/TiO2 and ozone is also possible. Ozone acts as a powerful oxidant in place of oxygen, which has a slow electron transfer from TiO2 (reaction 67) [33, 119]. In the presence of TiO2 , ozone generates HO• through the formation of an ozonide radical in the adsorption layer: O3 + e–cb → O3 •–

(69)

Then, direct and indirect ozonation reactions take place, with HO• generation: O3 •– + H+ → HO3 • HO3 • → HO• + O2

(70) (71)

HO• generation from O3 is pH dependent and increases with decreasing pH. This avoids the use of high alkaline pH to induce HO• formation from O3 . Photocatalytic ozonation of organic compounds such as 2,4-d, glyoxal, pyrrole-2-carboxylic acid, p-toluenesulfonic acid, monochloroacetic acid, phenol, aniline and others was found to be much faster than UV/TiO2 , UV/O3 or ozonation alone ([33] and references therein). In many cases, higher extents of mineralization than with single AOTs are reached.

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

359

Another interesting combination is heterogeneous photocatalysis with ultrasonic irradiation, because this process hinders the inactivation of the catalyst by reaction intermediates, which usually block the catalyst. Ultrasound also reduces mass transfer limitations occurring in the case of immobilized catalysts (see [8] for a detailed description of this combined process). 2.9 Comparative Practical Examples Two recent interesting examples will be brieﬂy commented upon to evaluate AOTs in real application cases. However, it is worthwhile to point out that it is not possible to generalize the results due to variable experimental conditions. In a recent work, different AOPs (O3 , O3 /H2 O2 , UV, UV/O3 , UV/H2 O2 , UV/O3 /H2 O2 , Fe2+ /H2 O2 and UV/TiO2 ) have been compared for the degradation of the model pollutant phenol [120]. Different variables (pH, oxidant, catalyst and reagent concentration) were studied to select the best conditions for each process, and pseudo-ﬁrst order constants were calculated and compared among the cases. None of the ozone combinations improved the degradation rate of the single ozone process and even inhibited it. The UV/H2 O2 process was almost ﬁve-times faster than photocatalysis and UV alone. Fenton reagent showed the fastest degradation, 40-times faster than UV and photocatalysis and 5-times faster than ozonation. Nevertheless, the relatively high degradation rate combined with lower costs made ozonation the most suitable choice for phenol degradation under the studied conditions. Another interesting case is the study of the treatability of methyl tertbutyl ether (MTBE) in ﬁve groundwaters with highly variable water quality characteristics. Air stripping, granular activated carbon (GAC) adsorption, O3 /H2 O2 and UV/H2 O2 were compared in a mobile water treatment pilot plant under a variety of conditions. For high-ﬂow rates, air stripping showed the lowest treatment costs, although relatively tall towers were required. However, at low ﬂow rates and low COD, AOTs were the least expensive treatments [121]. 2.10 Combination of PAOTs with Biological Treatments Reference [122] offers an overview of recent works (1998–2002) where photoassisted AOPs and biological processes were coupled for wastewater treatment. This overview conﬁrms the beneﬁcial effects of such two-step treatments at the laboratory scale and the lack of studies carried at a ﬁeld scale with the same approach. A general strategy to develop combined photochemical and biological systems for biorecalcitrant wastewater treatment was proposed, taking into ac-

360

M.I. Litter

count the following points: the biodegradability of the initial solutions, the operation mode of the coupled reactor, the chemical and biological characteristics of the phototreated solutions, the evaluation of different photoassisted AOPs, the optimal conditions for both the photochemical and biological processes, and the efﬁciency of the coupled reactor. The strategy to couple photochemical and biological processes is illustrated by case studies of four different biorecalcitrant pollutants: p-nitrotoluene-ortho-sulfonic acid, a pollutant derived from the manufacture of dyes, surfactants and brighteners, metobromuron and isoproturon, two of the most commonly used herbicides in Europe and 5-amino-6-methyl-2-benzimidazolone (AMBI), a model biorecalcitrant compound of the dye industry. Three kinds of combined systems were developed using either photo-Fenton, Fe3+ /UV, or TiO2 supported on glass rings for the photocatalytic pre-treatment and, in all cases, immobilized biomass for the biological step. However, the authors indicate that this strategy is not a universal solution. Chemical, biological, and kinetic studies must always be carried out to ensure that the photochemical pre-treatment increases the biocompatibility of the treated wastewater. Some ﬁeld experiments using a solar reactor indicated that a coupled photochemical-biological treatment system at the pilot scale is a possible way to achieve the complete mineralization of the biorecalcitrant pollutant compounds, but it can only be justiﬁed if the resulting intermediates are easily degradable in a further biological treatment [123]. An innovative coupled solar-biological system at ﬁeld pilot scale for the treatment of biorecalcitrant pollutants has been described in [122]. The strategy to develop this system implicates the choice of the most appropriate solar collector and the most efﬁcient AOP, the optimization of this AOP, the choice of the biological oxidation system, the monitoring of the chemical and biological characteristics of photo-treated solutions and the evaluation of the performance of the coupled solar-biological ﬂow system. The coupled system is conformed by a Compound Parabolic Solar Collector (CPC) and a Fixed Bed Reactor (FBR). AMBI was selected for tests. The results showed that CPC was the most appropriate photoreactor to be coupled with a biological reactor and that the photo-Fenton system was the most appropriate AOT for the degradation of the model pollutant, generating a biocompatible efﬂuent. The coupled reactor operated in semicontinuous mode, and a mineralization performance between 80 and 90% was reached in the range of initial dissolved organic carbon (DOC) concentration of 300–500 mg C L–1 . With this coupled system, wastewaters coming from textile, pulp and paper, surfactants, explosive military industries, and from olive washing, as well as efﬂuents contaminated with pesticides, were tested. For the 16 cases studied, two of them were previously biologically pre-treated to remove the easily biodegradable fraction before leading to the classical AOT-biological treatment schema, in which the main aim of the AOT is to produce biodegradable intermediates or partial mineralization. This result indicates the plausibility

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

361

Fig. 10 Schematic representation of the coupled solar-biological ﬂow reactor, according to [122]

of using the coupled approach at the pilot scale to treat real industrial wastewaters. Figure 10 shows a scheme of the proposed coupled system.

3 Conclusions The relevant parameter that determines if a photochemical AOT can result in an effective alternative to traditional processes (e.g., chlorination, biological treatment) is mostly the concentration of the pollutants. In general, AOTs are more adequate for the treatment of small ﬂows (or volumes) and not too high concentrations. Small COD contents, not higher than 5 g L–1 , can be suitably treated. Higher concentrations would require high concentrations of expensive reagents and/or high electrical power consumption [10]. The great utility of the technologies resides in the fact that they can process wastewaters resistant to conventional treatments and are complementary to them. However, the selection of the technology to be used must be based on its effectiveness and cost. The effectiveness depends on the nature of the contaminants to be destroyed, and the cost is strongly determined by the required equipment, the amount of energy required and the necessity for further treatment. Among the chemical reagents, the advantages of using O2 or H2 O2 as oxidants are clear, they are cheap, easy to handle and do not generate substances that must be removed later. Ozone shares the last advantage, but its manipulation is not as simple. A generalization on the application of an AOT can never be made. Each efﬂuent must be previously characterized, and treatability tests at the laboratory scale must be performed to choose the most appropriate method in economical and efﬁciency terms. It is important to evaluate the existing options to choose the most adequate. A knowledge of the kinetics, with establishment of the limiting step and limiting reagent(s), and a comparison with other conventional treatments should be available before applying

362

M.I. Litter

the technology. With a study of the kinetics, reliable information about substrate decay can be obtained, using analytical techniques such as HPLC or spectrophotometric measurements. Continuous TOC measurements should be performed to follow the degree of mineralization during the process. Obviously, a complex chemical composition always has a higher difﬁculty than simple mixtures, and HO• scavengers are usually the main source of efﬁciency reduction. From the technical point of view, suitable UV sources and appropriate photochemical reactors must be chosen and designed. A more extended exploitation of solar radiation would ensure a reduction of the costs of photochemical AOTs. In the case of photoreactors, a proper design should warrant the highest possible absorption of light by the reaction system. If an ozonebased technology is used, a rather expensive ozone generator is needed, with a cooling system, air-dryer and abatement of residual ozone at the end of the treatment. Furthermore, gas-liquid contactors, bubbling devices and good stirring must be provided to reduce mass transfer limitation problems. The use of ozone also requires resistant materials that cannot be attacked by the reagent, such as stainless steel. It should be remembered that each AOT has an optimum working pH value and that in addition to adjusting the initial pH, the variation of pH during the reaction must be continuously controlled. Of course, this is dependent on the composition of the mixture: some pollutants are transformed to acid intermediates and give rise to a pH decrease, while others such as amino compounds produce amines or ammonia that increases pH. As constantly repeated in this article, carbonate or bicarbonate, either initially present in the treated water or formed during the reaction, are strong HO• scavengers. At the end of the process, another pH adjustment will be needed in many cases before a biological treatment or to comply with local regulations before discharge of the efﬂuent to receiving bodies. The use of toxicological tests (Microtox, Amphitox, etc.), to control the formation of noxious by-products along the process path is mandatory. The purpose is to use the technology until toxicity is reduced to a certain level, beyond which a conventional, less expensive method, can bring about the mineralization process with the obvious reduction of costs. Although much research has been done to understand the mechanistic and kinetic aspects of AOTs, which can be improved in the future by new investigations, some requirements are still needed for wide commercialization. These requirements refer mainly to reactor optimization and modeling from the point of view of chemical engineering. In the case of solar light, ﬂuctuation in solar irradiation through the year or because of varying weather conditions on different days, makes reactor design difﬁcult. Another important point is the control of variables that can affect the reactivity. This last task needs the support from expertise coming from varied scientiﬁc areas. Research in solid state physics can lead to an improved semiconductor activity;

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment

363

development of analytical techniques would allow the discovery of methods for evaluating very low concentrations of the target and intermediates, and there are many other examples. As possible future actions in the area, it is necessary to widely disseminate photochemical AOTs, especially in the less industrially developed countries, as alternative technologies of treatment that can be successfully used instead of other more expensive or less productive ones. These underdeveloped countries are fortunate in that they possess the highest sunlight irradiation powers on the planet, in contrast to the richest countries. From the point of view of scientiﬁc research, knowledge of the mechanisms taking place in PAOTs is extremely important as a way of improving existing drawbacks that hinder the use of the technology. Acknowledgements This work is part of Comisión Nacional de Energía Atómica CNEA P5-PID-36-4 Program and Consejo Nacional de Investigaciones Cientíﬁcas y Técnicas, PIP662/98 CONICET project. M.I.L. is a member of CONICET.

References 1. Legrini OR, Oliveros E, Braun AM (1993) Chem Rev 93:671 2. Huang CP, Dong Ch, Tang Z (1993) Waste Manag 13:361 3. US/EPA Handbook of Advanced Photochemical Oxidation Processes (1998) EPA/625/R-98/004 4. The AOT Handbook, Calgon Carbon Oxidation Technologies (1996) Ontario 5. Bolton JR, Cater SR (1994) Helz GR, Zepp RG, Crosby DG (eds) Aquatic and Surface Photochemistry. Lewis, Boca Raton, FL, p 467 6. Glaze WH (1987) Environ Sci Technol 21:224 7. Glaze WH, Kang JW, Chapin DH (1987) Ozone Sci Technol 9:335 8. Gogate PR, Pandit AB (2004) Adv Environ Res 8:553 9. Domènech X, Jardim W, Litter M (2001) Tecnologías avanzadas de oxidación para la eliminación de contaminantes. In: Blesa MA (ed) Eliminación de contaminantes por fotocatálisis heterogénea. Texto colectivo elaborado por la Red CYTED VIII-G.1 Digital Graﬁc, La Plata, p 3. Available at: http://www.cnea.gov.ar/xxi/ambiental/cyted.asp (last accessed 7th September 2005) 10. Andreozzi R, Caprio V, Insola A, Marotta R (1999) Catal Today 53:51 11. Scott JP, Ollis DF (1995) Environ Progress 14:88 12. Burrows HD, Canle M, Santaballa JA, Steenken S (2002) J Photochem Photobiol B: Biol 67:71 13. Golimowski J, Golimowska K (1996) Anal Chim Acta 325:111 14. da Silva CG, Faria JL (2003) J Photochem Photobiol A: Chem 155:133 15. Faust D, Funken K-H, Horneck G, Milow B, Ortner J, Sattlegger M, Schäfer M, Schmitz C (1999) Solar Energy 65:71 16. Wilkinson F, Helman WP, Ross AB (1993) J Phys Chem Ref Data 22:113 17. Iesce MR, Graziano ML, Cermola F, Montella S, di Gioia L, Stasio C (2003) Chemosphere 51:163 18. Schaap AP, Thayer AL, Blossey EC, Neckers DC (1975) J Am Chem Soc 97:3741

364

M.I. Litter

19. López JL, García Einschlag FS, González MC, Capparelli AL, Oliveros E, Hashem TM, Braun AM (2000) J Photochem Photobiol A: Chem 137:177 20. Heit G, Neuner A, Saugy P-Y, Braun AM (1998) J Phys Chem A 102:5551 21. González MC, Braun AM (1995) Res Chem Intermed 21:837 22. González MC, Braun AM (1996) J Photochem Photobiol A: Chem 95:67 23. González MC, Braun AM (1994) Chemosphere 28:2121 24. Neyens E, Baeyens J (2003) J Hazardous Mater B98:33 25. Bircher KG, Lem W, Simms KM, Dussert BW (1997) J Adv Oxid Technol 2:435 26. Beltrán FJ, Ovejero G, Rivas J (1996) Ind Eng Chem Res 35:883 27. López Cisneros R, Gutarra Espinoza A, Litter MI (2002) Chemosphere 48:393 28. Baxendale JH, Wilson JA (1957) Trans Faraday Soc 53:344 29. Ince NH, Stefan MI, Bolton JR (1997) J Adv Oxid Technol 2:442 30. Fung PC, Huang Q, Tsui SM, Poon CS (1999) Wat Sci Tech 40:153 31. Arslan I, Blacioglu IA (2001) J Chem Technol Biotechnol 76:53 32. López A, Bozzi A, Mascolo G, Kiwi J (2003) J Photochem Photobiol A: Chem 156:121 33. Kasprzyk-Hordern B, Ziólek M, Nawrocki J (2003) Appl Catal B: Environ 46:639 34. von Gunten U (2003) Wat Res 37:1443, 1469 35. Hoigné J, Bader H (1976) Wat Res 10:377 36. Hoigné J, Bader H (1983) Wat Res 17:173 37. Legube B, Karpel Vel Leitner N (1999) Catal Today 53:61 38. Muñoz F, Mvula E, Braslavsky SE, von Sonntag C (2001) J Chem Soc Perkin Trans 2:1109 39. Chiron S, Fernández-Alba A, Rodríguez A, García-Calvo E (2000) Wat Res 34:366 40. Beltrán FJ, Encinar JM, Alonso MA (1998) Ind Eng Chem Res 37:25 41. Roche P, Volk C, Carbonnier F, Paillard H (1994) Ozone Sci Eng 16:135 42. Glaze WH, Beltrán FJ, Tuhkanen T, Kang JW (1992) Water Poll Res J Canada 27:23 43. Gurol MD, Akata A (1996) AIChE J 42:3283 44. Roche P, Prados M (1995) Ozone Sci Eng 17:657 45. Beltrán FJ, González M, Rivas J, Marín M (1994) Ind Eng Chem Res 33:125 46. Peyton GR, Glaze WH (1988) Environ Sci Technol 22:761 47. Beltrán FJ, García Araya JF, Acedo B (1994) Wat Res 28:2165 48. Beltran FJ, Encinar JM, Alonso MA (1998) Ind Eng Chem Res 37:32 49. Domènech X, Casado J, Peral J (2003) Chemosphere 50:1085 50. Fenton HJJ (1894) J Chem Soc 65:899 51. Walling Ch (1975) Acc Chem Res 8:125 52. Tang WZ, Tassos S (1997) Wat Res 31:1117 53. Pignatello JJ (1992) Environ Sci Technol 26:944 54. Pignatello JJ, Liu D, Huston P (1999) Environ Sci Technol 33:1832 55. Safarzadeh-Amiri A, Bolton JR, Cater SR (1996) J Adv Oxid Technol 1:18 56. Safarzadeh-Amiri A, Bolton JR, Cater SR (1997) Wat Res 31:787 57. De Laat J, Gallard H (1999) Environ Sci Technol 33:2726 58. Bossmann SH, Oliveros E, Göb S, Siegwart S, Dahlen EP, Payawan L Jr, Straub M, Wörner M, Braun AM (1998) J Phys Chem A 102:5542 59. Lin SH, Lo CC (1997) Wat Res 31:2050 60. Bigda RJ (1995) Chem Eng Progress 63 61. Hue N, Quan AL, Teel R, Watts J (2003) J Hazard Mater B102:277 62. Watts RJ, Stanton PC, Howsawkeng J, Teel AL (2002) Wat Res 36:4283 63. Teel AL, Warberg CR, Atkinson DA, Watts RJ (2001) Wat Res 35:977 64. Sedlak DL, Andren AW (1991) Environ Sci Technol 25:777 65. Feng W, Nansheng D (2000) Chemosphere 41:1137

Introduction to Photochemical Advanced Oxidation Processes for Water Treatment 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105.

365

Mazellier P, Sarakha M, Bolte M (1999) New J Chem 133 Brand N, Mailhot G, Bolte M (2000) Chemosphere 40:395 Mailhot G, Asif A, Bolte M (2000) Chemosphere 41:363 Brand N, Mailhot G, Sarakha M, Bolte M (2000) J Photochem Photobiol A: Chem 135:221 Mazellier P, Bolte M (2000) J Photochem Photobiol A: Chem 132:129 Brand N, Mailhot G, Bolte M (2000) J Information Recording 25:439 Mazellier P, Bolte M (2001) Chemosphere 42:361 Mazellier P, Brand N, Mailhot G, Bolte M (2000) Entropie 228:44 Bajt O, Mailhot G, Bolte M (2001) Appl Catal B: Environ 33:239 Ruppert G, Bauer R, Heisler G (1993) J Photochem Photobiol A: Chem 73:75 Park H, Choi W (2003) J Photochem Photobiol A: Chem 159:241 Celin SM, Pandit M, Kapoor JC, Sharma RK (2003) Chemosphere 53:63 Murray CA, Parsons SA (2003) Chemosphere 54:1017 Hatchard CG, Parker CA (1956) Proc Roy Soc (London) A 235:518 Zuo Y, Hoigné J (1992) Environ Sci Technol 26:1014 Safarzadeh-Amiri A, Bolton JR, Cater SR (1996) Solar Energy 56:439 Nogueira RFP, Jardim WF (1999) J Adv Oxid Technol 4:1 Lee Y, Jeong J, Lee C, Kim S, Yoon J (2003) Chemosphere 51:901 Lee Y, Lee C, Yoon J (2003) Chemosphere 51:963 Nogueira RFP, Alberici RM, Mendes MA, Jardim WF, Eberlin MN (1999) Ind Eng Chem Res 38:1754 Abida O, Emilio C, Quici N, Gettar R, Litter M, Mailhot G, Bolte M (2004) Wat Sci Technol 49:123 Emilio CA, Jardim WF, Litter MI, Mansilla HD (2002) J Photochem Photobiol A: Chem 151:121 Wegelin M, Gechter D, Hug S, Mahmud A, Motaleb A (2000) In: Water, sanitation, hygiene: challenges of the millennium. 26th WEDC Conference, Dhaka, p 379 Hug SJ, Canonica L, Wegelin M, Gechter D, Von Gunten U (2001) Environ Sci Technol 35:2114 Hug SJ, Leupin O (2003) Environ Sci Technol 37:2734 García MG, d’Hiriart J, Giulitti J, Hidalgo MV, Lin H, Custo G, Litter MI, Blesa MA (2004) Solar Energy 77:601 Su C, Puls RW (2001) Environ Sci Technol 35:1487 Pulgarin CO, Schwitzguebel J-P, Peringer PA (1996) J Adv Oxid Technol 1:94 Deng N, Luo F, Wu F, Xiao M, Wu X (2000) Wat Res 34:2408 Liao C-H, Kang S-F, Hsub Y-W (2003) Wat Res 37:4109 Canton C, Esplugas S, Casado J (2003) Appl Catal B: Environ 43:139 Brillas E, Calpe JC, Cabot P-L (2003) Appl Catal B: Environ 46:381 Boye B, Dieng MM, Brillas E (2003) J Electroanal Chem 557:135 Brillas E, Baños MÁ, Garrido JA (2003) Electrochimica Acta 48:1697 Maletzky P, Bauer R (1999) Chemosphere 38:2315 Fernández J, Bandara J, López A, Buffat Ph, Kiwi J (1999) Langmuir 15:185 Sabhi S, Kiwi J (2001) Wat Res 35:1994 Dhananjeyan M, Kiwi J, Albers P, Enea O (2001) Helv Chim Acta 84:3433 Parra S, Guasaquillo I, Enea O, Mielczarski E, Mielczarski J, Albers P, Kiwi-Minsker L, Kiwi J (2003) J Phys Chem B 107:7026 Dhananjeyan M, Mielczarski E, Thampi R, Buffat P, Bensimon M, Kulik A, Mielczarski J, Kiwi J (2000) J Phys Chem 105:12046

366

M.I. Litter

106. Fernández J, Dhananjeyan MR, Kiwi J, Senuma Y, Hilborn J (2000) J Phys Chem B 104:5298 107. Bozzi A, Yuranova T, Mielczarski E, Mielczarski J, Buffat PA, Lais P, Kiwi (2003) J Appl Catal B: Environ 42:289 108. Chou S, Huang C, Huang Y-H (2001) Environ Sci Technol 35:1247 109. Pak DW, Chang WS (1999) Wat Sci Technol 40:115 110. Huling SG, Arnold RG, Jones PK, Sierka RA (2000) J Environ Eng 126:348 111. Centi G, Perathoner S, Torre T, Verduna MG (2000) Catal Today 55:611 112. Kremer ML (1999) Phys Chem Chem Phys 1:3595 113. Weavers LK, Hua I, Hoffmann MR (1997) Wat Environ Res 69:1112 114. Mills A, Le Hunte S (1997) J Photochem Photobiol A: 108:1–35 115. Morrison SR (1980) Electrochemistry at Semiconductor and Oxidized Metal Electrodes. Plenum Press, New York, p 186 116. Babay PA, Emilio CA, Ferreyra RE, Gautier EA, Gettar RT, Litter MI (2001) In: Vogelpohl A, Geissen SU, Kragert B, Sievers M (eds) Oxidation technologies for water and wastewater treatment (II), Water Sci Technol 44:79 117. Babay PA, Emilio CA, Ferreyra RE, Gautier EA, Gettar RT, Litter MI (2001) Int J Photoenergy 3:193 118. Blesa MA, Chocrón M, Litter MI, Gettar R, Babay P, Paolella M, Repetto P, Quici N, Piperata G (2003) Revista de la Comisión Nacional de Energía Atómica, 9/10:11 119. Gilbert E (2002) Ozone Sci Eng 24:75 120. Esplugas S, Giménez J, Contreras S, Pascual E, Rodríguez M (2002) Wat Res 36:1034 121. Sutherland J, Adams C, Kekobad J (2003) Wat Res 38:193 122. Sarria V, Kenfack S, Guillod O, Pulgarin C (2003) J Photochem Photobiol A: Chem 159:89 123. Sarria V, Parra S, Adler N, Péringer P, Benitez N, Pulgarin C (2002) Catal Today 76:301 124. Brillas E, Mur E, Sauleda R, Sánchez L, Peral J, Domènech X, Casado J (1998) Appl Catal B: Environ 16:31 125. Kraft A, Stadelmann M, Blaschke M (2003) J Hazard Mater B103:247 126. Makogon O, Fliount R, Asmus K-D (1998) J Adv Oxid Technol 3:11 127. Chaychian M, Silverman J, Sheikhly MA (1999) Environ Sci Technol 33:2461 128. Innovative Technology Evaluation Report EPA/540/R-96/504 (1997), National Risk Management Research Laboratory, Cincinnati 129. Rosocha LA, Korzekwa RA (1999) J Adv Oxid Technol 4:247 130. Joseph JM, Destaillats H, Hung HM, Hoffmann MR (2000) J Phys Chem: A 104:301 131. Destaillats H, Hung HM, Hoffmann MR (2000) Environ Sci Technol 34:311 132. Destaillats H, Colussi AJ, Joseph JM, Hoffmann MR (2000) J Phys Chem: A 104:8930 133. Olson TM, Barbier PF (1994) Wat Res 28:1383 134. Kronholm J, Riekola ML (1999) Environ Sci Technol 33:2095 135. Zhang Q, Chuang KT (1999) Environ Sci Technol 33:3641 136. Martino CJ, Savage PE (1999) Environ Sci Technol 33:1911 137. Dinsdale RM, Almemark M, Hawkes FR, Hawkes DL (1999) Environ Sci Technol 33:4092 138. Deng B, Burris DR, Campbell TJ (1999) Environ Sci Technol 33:2651 139. Tang WZ, Chen RZ (1996) Chemosphere 32:947 140. Sharma VK, Rivera W, Joshi VN, Millero FJ, O’Connor D (1999) Environ Sci Technol 33:2645

Hdb Env Chem Vol. 2, Part M (2005): 367–423 DOI 10.1007/b138189 © Springer-Verlag Berlin Heidelberg 2005 Published online: 15 September 2005

Photocatalytic Detoxification of Water and Air Peter K. J. Robertson1 (u) · Detlef W. Bahnemann2 · Jeanette M. C. Robertson1 · Fiona Wood1 1 Centre

for Research in Energy and the Environment, School of Engineering, The Robert Gordon University, Schoolhill, Aberdeen AB10 1FR, UK [email protected] 2 Institut für Technische Chemie, Universität Hannover, Callinstrasse 3, 30167 Hannover, Germany [email protected] 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

368

2

Primary Processes upon Bandgap Irradiation of Semiconductor Particles

369

3

Chemical Nature of Trapped Charge Carriers . . . . . . . . . . . . . . . .

370

4 4.1 4.2 4.2.1 4.2.2 4.2.3

Fate of Trapped Charge Carriers . . . . . . . . . . . . . . . . . . . . . . . Recombination Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . Charge Transfer Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . Interfacial Electron Transfer . . . . . . . . . . . . . . . . . . . . . . . . . Direct Interfacial Hole Transfer . . . . . . . . . . . . . . . . . . . . . . . Hole Transfer through the Intermediate Formation of Hydroxyl Radicals

. . . . . .

372 372 374 374 374 376

5 5.1 5.2 5.3

Novel Mechanistic Concepts in Photocatalysis The Antenna Mechanism . . . . . . . . . . . . Metal Nanocontacts . . . . . . . . . . . . . . . The De-Aggregation Concept . . . . . . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

380 380 381 384

6 6.1 6.2 6.3 6.4 6.5 6.6 6.7

Photocatalytic Treatment of Waste Water . . Haloaromatic and Aliphatic Compounds . . Phenols and Polyaromatic Hydrocarbons . . Agrochemicals . . . . . . . . . . . . . . . . . Dye Compounds . . . . . . . . . . . . . . . . Explosive Residues and Oilﬁeld Chemicals . Potable Water Applications . . . . . . . . . . Disinfection of Pathogenic Micro-Organisms

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

385 385 389 390 392 394 395 398

7

Air Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

404

8 8.1 8.1.1 8.1.2 8.1.3 8.2

Solar Reactors and Pilot Plants . . . . . . . . . . . Reactor Types . . . . . . . . . . . . . . . . . . . . . Thin-Film-Fixed-Bed Reactor (TFFBR) . . . . . . . Compound Parabolic Collecting Reactor (CPCR) . . Double Skin Sheet Reactor (DSSR) . . . . . . . . . . Solar Photocatalytic Treatment of Real Waste Water

. . . . . .

410 410 411 411 412 412

9

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

416

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

416

. . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

368

P.K.J. Robertson et al.

Abstract The use of semiconductor photocatalysis for treatment of water and air has been a topic of intense research activity over the past 20 years. This chapter provides a review of this highly effective technology. The fundamental processes involved in the technique are initially detailed with a discussion of some recent novel concepts in photocatalysis. A range of applications of water and air treatment are subsequently described with examples of mechanistic description of the major breakdown pathways of some key compounds. Examples of large-scale water treatment applications are also discussed. Keywords Photocatalysis · Reaction mechanisms · Novel photocatalysis · Water treatment · Disinfection · Photocatalytic reactor

1 Introduction The generation of hazardous industrial efﬂuents is a serious problem experienced by nations throughout both the developed and developing world. In recent years there has been a huge proliferation of legislation within the European Union aimed at protecting the environment. Directives such as the Water Framework Directive, Urban Waste Water Directive, the Integrated Pollution Prevention and Control Directive, the Dangerous Substances Directive, Directive on Air Pollution from Industrial Plants and Directive on the Control of Volatile Organic Compound (VOC) Emissions have all emphasised the importance of developing new effective and environmentally acceptable water and air treatment technologies. Sectors such as the petrochemical, pharmaceutical, textile, agricultural, food, and chemical industries all produce waste efﬂuent contaminated with organic compounds such as aromatics, haloaromatics, aliphatics, dyes, dioxins and a wide range of other polluting materials. Many of these materials are extremely toxic and the waste streams must be treated prior to discharge. Traditional waste treatment systems have involved the use of techniques such as coagulation, chlorination or ozonation which utilise potentially hazardous or polluting materials. Chlorination presents a particular problem since it will often generate trihalomethanes as by-products when used to treat water contaminated with organic compounds. An effective treatment system is required which can degrade the polluting materials prior to discharge of the efﬂuent. The TiO2 photocatalysis process is a technique which could address both these requirements as has been indicated by the massive range of compounds that have been successfully mineralised using this technique [1–6]. The technology has also proven to be effective for removal of hazardous organic compounds from air [7–9]. TiO2 photocatalysis has developed into a massive research topic particularly over the past decade and the commercialisation of the technology has been applied in a wide range of ﬁelds. This is indicated by the market for photocatalytic products which is already considerable. For example the turnover in Japan in 2003 was 450 Million and it is expected to grow up to 1 Billion

Photocatalytic Detoxiﬁcation of Water and Air

369

by 2006. This chapter reviews the applications of photocatalysis for air and water treatment with a consideration of the basic principles of the photocatalytic process through to examples of application to treatment of air, waste and potable waters.

2 Primary Processes upon Bandgap Irradiation of Semiconductor Particles Absorption of a photon with an energy hν greater or equal to the bandgap energy Eg of a semiconductor (i.e., 3.2 eV for titanium dioxide in its anatase modiﬁcation) generally leads to the formation of an electron/hole pair in the semiconductor particle (reaction 1 and Fig. 1). TiO2 + hν → TiO2 (e–cb + h+vb )

(1)

where e– cb represents a conduction band (CB) electron and h+vb a positive hole in the valance band (VB) of the semiconductor. After their generation according to reaction 1 both conduction band electrons and valence band holes migrate to the surface of the semiconductor particle. The transit time τ needed by these charge carriers to reach the surface of the particle is given by Eq. 2 τ=

R2 πD

(2)

Fig. 1 Schematic presentation of the processes occurring in photocatalysis upon bandgap irradiation of a semiconductor particle

370

P.K.J. Robertson et al.

where R is the radius of the particle and D the diffusion coefﬁcient of the excited charge carriers [10]. Taking a value of D = 5 × 10–3 cm2 s–1 [11] and a particle radius of 2.5 nm (radius of particles typically used in experiments with colloidal titanium dioxide but also typical particle sizes of several commercially available photocatalyst powders, i.e., Millennium PC 500 or Sachtleben UV 100) the average transit time is only 1.3 ps. Even with bigger particles such as those used in photocatalytic systems for water treatment (e.g., 21 nm for Degussa P25 titanium dioxide [12]) the transit time is only some ten picoseconds. Reaching the surface these charge carriers are trapped in the subsurface and surface states of the particle (reactions 3 and 4) e–cb → e–tr h+vb → h+tr

(3) (4)

where e–tr and h+tr represent the trapped electron and trapped hole, respectively. These trapped charge carriers exhibit strong optical transient absorptions. The position of the absorption maximum is deﬁnitely affected by the presence of suitable electron acceptors and donors in the surrounding aqueous phase. Exploiting this effect it has been shown in early laser ﬂash photolysis studies that the trapped electron exhibits a strong optical absorption around 650 nm while the trapped hole absorbs predominantly at shorter wavelengths, i.e., around 430 nm or even shorter [13, 14]. As has been shown by time-resolved ﬂash photolysis measurements in colloidal titanium dioxide suspensions trapping is a very fast process. Rothenberger et al. performed picosecond and nanosecond transient absorption experiments on titanium dioxide and observed that the electron trapping time was faster than 30 ps, the time resolution of their laser system [15]. The trapping time for holes was estimated to be < 250 ns. In a picosecond study by Serpone et al. on titanium dioxide colloid solutions of varying diameters it was observed that the spectra of trapped electrons as well as of trapped holes are fully developed after a laser pulse of 30 ps [16]. Another picosecond study by Bowmann et al. had shown that the time needed for the full development of the spectrum of the electron was approximately 200 fs [17].

3 Chemical Nature of Trapped Charge Carriers Generally it is assumed, that TiIV cations at the surface of the titanium dioxide particle are reduced by the light-induced electrons forming TiIII cations [18, 19] which can be considered to be intrinsic surface states localised about 0.1 eV below the conduction band edge, i.e., within the bandgap [20, 21]. Equilibrium between these trapped electrons and free electrons is assumed, but in an acidic medium nearly all electrons are trapped in surface states [22].

Photocatalytic Detoxiﬁcation of Water and Air

371

On the basis of their laser ﬂash photolysis measurements Hoffmann et al. have extended this mechanistic picture [23, 24]. These authors assume that the CB electrons are trapped in two different TiIII sites (reactions 5 and 6) e–cb + TiIV OH ⇔ TiIII OH e–cb + TiIV → TiIII

(5) (6)

where TiIII OH is a surface-trapped electron while TiIII denotes a bulk trapped electron. The dynamic equilibrium of reaction 5 represents a reversible trapping of a CB electron in a shallow trap. Hoffmann et al. have estimated that these trapped electrons lie in the range of 25 to 50 meV below the conduction band edge of titanium dioxide Degussa P25 [23, 24]. Reaction 6 represents the irreversible trapping in a deep trap. From the analysis of their experimental results of the investigation of the charge carrier recombination kinetics in titanium dioxide colloidal solutions and in dispersions Serpone et al. and Bowman et al have also assumed the existence of two different traps [16, 17]. Concerning the energetic nature of the electron-trapping centres, several problems arise. Assuming that the absorptions of the trapped electrons around 650 nm (1.7 eV) correspond to a transition between a classical surface state and the conduction band, this surface state is located near the middle of the band gap. As has been already pointed out by Bahnemann and co-workers this assignment cannot be correct because it has been found experimentally that the reduction of molecular oxygen occurs via transfer of a trapped electron [25]. This process would not be possible thermodynamically if the electron originated from an energy state being 1.7 eV below the conduction band of titanium dioxide. Accordingly, Bahnemann et al. assumed that the absorption is due to an excitation of a trapped electron within a surface molecule [26]. The chemical nature of the trapped holes has not been clearly clariﬁed yet. Older reports assume that the holes are trapped at the titanium dioxide surface in adsorbed hydroxy groups yielding weakly adsorbed hydroxyl radicals (reaction 7) [27–31]. (7) h+ + TiIV – O2– – TiIV – OH → TiIV – O2– – TiIV ∼ · OH+ vb

Howe and Grätzel deduced from esr investigations that the hole is trapped in a subsurface oxygen anion (reaction 8) [22]. h+ + TiIV – O2– – TiIV – OH → TiIV – O· – – TiIV – OH (8) vb

Other groups assume that the trapped hole is an oxygen radical centred at the surface of the titanium dioxide particle, having an energy state lower than the valence band edge of the semiconductor (reaction 9) [32]. h+ + TiIV – O2– – TiIV – OH → TiIV – O2– – TiIV – O· + H+ (9) vb

On the other hand, an extended model with two different types of traps appears to be much better suited to explaining the experimental observations

372

P.K.J. Robertson et al.

obtained by Bahnemann and co-workers [26]. They have therefore proposed the following modiﬁcation of reaction 4. It is envisaged that at least two different trap sites for holes exist on the surface of the TiO2 particle. While holes which are trapped in energetically deep traps, h+tr,d , can be characterised by their transient absorption around 430 nm, those initially residing in shallow traps, h+tr,s , do not possess such spectral features. Following generation, all holes are rapidly trapped in either of these energy states (reactions 10 and 11). h+ → h+tr,d h

+

⇔ h+tr,s

(10) (11)

The holes trapped in shallow traps are probably excited thermally into the valence band, so that equilibrium with free holes is indicated in reaction 11. Shallowly trapped holes, h+tr,s , will therefore have a comparable reactivity to detrapped holes, h+ . While both types of trapped holes will recombine with the trapped electrons within the ﬁrst 200 ns after their generation following reaction 1, only holes excited thermally from the shallow traps have the chance to migrate to the energetically more favoured h+tr,d site (cf. reaction 12). h+tr,s → h+tr,d

(12)

It is assumed that deeply trapped holes, h+tr , are chemically equivalent to surface-bound hydroxyl radicals. Weakly trapped holes, on the other hand, that are readily detrapped apparently possess an electrochemical potential close to that of free holes and can therefore be considered to be chemically similar to the latter. Their shallow traps are probably created by surface imperfections of the semiconductor nanocrystal. From these traps the charge carriers recombine or they are transferred by interfacial charge transfer to suitable electron acceptors or donors adsorbed at the surface of the semiconductor.

4 Fate of Trapped Charge Carriers 4.1 Recombination Kinetics The charge carriers formed upon absorption of light (reaction 1) can recombine in a radiative or nonradiative way according to reactions 13 to 16. This is clearly seen from the rather rapid depletion of the transient absorption

Photocatalytic Detoxiﬁcation of Water and Air

373

spectra recorded during laser ﬂash photolysis studies. e–cb + h+vb → TiO2 + energy e–tr + h+vb → TiO2 + energy e–cb + h+tr → TiO2 + energy e–tr + h+tr → TiO2 + energy

(13) (14) (15) (16)

The recombination kinetics of the charge carriers have been studied in detail by the groups of Grätzel, Serpone and Colombo [15–17]. Since recombination of electrons and holes is monitored by transient absorption techniques most of the observed decay is due to reaction 16. The ﬁrst picosecond laser spectroscopic study to examine charge carrier trapping and recombination dynamics was reported by Grätzel, Serpone and co-workers [15]. The mean lifetime of a single electron/hole pair was determined to be 30 ± 15 ns at low occupancy of electron/hole pairs in the titanium dioxide particles. At high occupancies, where recombination followed second-order kinetics, the bulk rate coefﬁcient for recombination was (3.2 ± 1.4) × 10–11 cm3 s–1 . Serpone et al. have examined colloidal titanium dioxide sols (prepared by hydrolysis of TiCl4 ) with mean particle diameters of 2.1, 13.3, and 26.7 nm by picosecond transient absorption and emission spectroscopy [16]. Absorption decay for the 2.1 nm sols was found to be a simple ﬁrst-order process, and electron/hole recombination was 100% complete by 10 ns. For the 13.3 and 26.7 nm sols absorption decay follows distinct second-order biphasic kinetics; the decay times of the fast components decrease with increase in particle size. 10 ns after the excitation pulse, about 90% or more of the photogenerated electron/hole pairs have recombined such that the quantum yield of photooxidations must be 10% or less. The faster components are due to the recombination of shallow-trapped charge carriers, whereas the slower components (τ > 20 ns) reﬂect recombination of deep-trapped electrons and holes. Bowman and co-workers characterised the subpicosecond dynamics of titanium dioxide sols employing particle sizes of about 2 nm prepared by hydrolysis of titanium tetraisopropoxide [17]. From their spectral results the authors inferred that the average lifetime of an electron/hole pair is 23 ± 5 ps, and substantial electron/hole recombination occurs within the ﬁrst 30 ps. A second-order recombination rate constant of (1.8 ± 0.7) × 10–10 cm3 s–1 for trapped electrons with holes has been obtained [33].

374

P.K.J. Robertson et al.

4.2 Charge Transfer Kinetics 4.2.1 Interfacial Electron Transfer In most experiments and applications with titanium dioxide photocatalysts, molecular oxygen is present to act as the primary electron acceptor. Usually the electrons trapped as Ti(III) are transferred to molecular oxygen adsorbed at the semiconductor surface yielding peroxyl radical anions (reaction 16) [34–37]. e–tr + O2,ads → O2 · – O2 · – + H+ → HO2 ·

(17) (18)

Depending on the pH of the suspension these superoxide radical anions can also exist in the protonated form (reaction 18) [38]. Besides the electron transfer from the semiconductor to adsorbed molecular oxygen also the direct transfer to an organic molecule is possible. This type of photocatalytic reaction, yielding an organic radical anion, has been found to occur with 1,4-benzoquinone [39], tetrachloromethane [40, 42], and several nitroaromatic compounds [43, 44]. But electrons can also be transferred very efﬁciently to (adsorbed) metal cations [45–47]. In the investigations of Bahnemann and co-workers different decay kinetics and evolution of the transient absorption spectra of titanium dioxide colloidal solutions upon bandgap irradiation have been observed depending upon the presence of molecular oxygen, air, or molecular nitrogen, respectively [25]. In every case, a biphasic decay of the transient absorption signal was observed. Following a fast initial decay, the remaining 20–40% of the original signal height decayed much more slowly. While in the presence of molecular nitrogen this portion of the signal appeared to be stable even over a period of 200 ms, its decay rate increased with increasing O2 concentration. Considering the limited number of data points a rate constant k = 7.6 × 107 L mol–1 s–1 has been determined by Bahnemann and co-workers for the reaction of a trapped electron with molecular oxygen evidencing that this is a really slow process that in many cases could be the rate limiting step in photocatalytic processes [25]. 4.2.2 Direct Interfacial Hole Transfer A signiﬁcant body of literature proposes that the photocatalytic oxidation of organic or inorganic solutes may occur by either indirect oxidation via a surface-bound hydroxyl radical (i.e., a trapped hole at the particle surface)

Photocatalytic Detoxiﬁcation of Water and Air

375

or directly via the valence-band hole before it is trapped either within the particle or at the particle surface. Interfacial hole transfer from titanium dioxide to organic and inorganic solutes has been studied by Grabner et al. [48], Colombo and Bowman [49], and Bahnemann and co-workers [26]. Grabner et al. have shown that in titanium dioxide sols containing chloride (which is either introduced into the solution as HCl to adjust the pH or is present on the particle surface when TiCl4 is used as the starting compound to prepare TiO2 ) Cl2 · – radical anions are formed. Their formation was postulated to occur by direct valence-band hole oxidation of surface adsorbed Cl– (reactions 19 and 20) [48]. h+vb + Cl– → Cl· Cl· + Cl– → Cl2 · –

(19) (20)

It has been observed that these Cl2 · – radical anions oxidise phenol yielding phenoxyl radicals (reaction 21) [48]. PhOH + Cl2 · – → PhOH· + + 2Cl–

(21)

Interfacial hole transfer dynamics from titanium dioxide (Degussa P 25) to SCN– has been investigated by Colombo and Bowman using femtosecond time-resolved diffuse reﬂectance spectroscopy [49]. A dramatic increase in the population of trapped electrons was observed within the ﬁrst few picoseconds, demonstrating that interfacial charge transfer of an electron from the SCN– to a hole on the photoexcited titanium dioxide effectively competes with electron-hole recombination (reactions 13–16) on an ultrafast time scale [49]. Direct hole transfer was also observed when dichloroacetate (DCA) was used as the electron donor [26]. While it was obvious that deeply trapped holes h+tr,d do not react with dichloroacetate, it was observed that the h+tr,d concentration is reduced considerably in the presence of DCA– , either the free holes, h+ , can be directly transferred to adsorbed DCA– molecules (reaction 19) or shallowly trapped holes, h+tr,s , are detrapped (reaction 11) to react with DCA– in the nanosecond time scale via reaction 22. h+ + DCA– → DCA·

(22)

A similar reactivity of trapped holes has previously reported by Bahnemann et al. [13, 14] when they studied reactions in colloidal TiO2 /Pt suspensions with an average particle diameter of approximately 12 nm. While the addition of ethanol as a hole scavenger resulted in a considerable increase of the rate of disappearance of the h+tr absorption, the addition of citrate and acetate mainly led to a decrease of its initial absorption height. It was concluded that strongly adsorbed ionic species would primarily react with free holes while weakly adsorbed molecules will mainly react with long-lived h+tr in a diffusion-controlled process [13, 14].

376

P.K.J. Robertson et al.

The direct charge transfer to dichloroacetate proposed in reaction 22 requires that the scavenging molecules are adsorbed on the TiO2 surface prior to the adsorption of the photon. Otherwise, this reaction could not compete with the normal hole-trapping reactions 10 and 11. So the adsorption of the model compound DCA– on the titanium dioxide surface prior to the bandgap excitation appears to be a prerequisite for an efﬁcient hole scavenging. A detailed kinetic analysis of the time-resolved spectroscopic data revealed an extremely good correlation with independent adsorption measurements [50]. It has been calculated that 20% of all TiO2 particles carry on average one adsorbed DCA– anion. The direct one electron oxidation of dichloroacetate immediately follows the hole transfer from the bulk to the TiO2 surface and, in principle, a maximum photonic efﬁciency of 0.2 would be possible under the experimental conditions. However, much lower efﬁciencies have been observed during the steady-state photocatalytic oxidation of dichloroacetate in the presence of TiO2 colloids [50], suggesting that a considerable number of holes either recombine with the electrons or are trapped at the surface hydroxyl groups yielding the transient absorption around 430 nm. These surface-bound hydroxyl-radicals are apparently unreactive toward dichloroacetate. Thus, the model incorporating the direct hole trapping by adsorbed dichloroacetate molecules, which has been proposed by Bahnemann and coworkers, appears to be probable [26]. Moreover, calculations using the Marcus electron transfer theory for adiabatic processes which result in a reorientation energy of 0.64 eV suggest that also in the case of SCN– the hole transfer occurs in the adsorbed state [26]. 4.2.3 Hole Transfer through the Intermediate Formation of Hydroxyl Radicals In photocatalytic degradation experiments with acetate in dioxygen-containing suspensions of TiO2 evidence had been obtained that holes as well as hydroxyl radicals are acting as oxidising species [11, 51]. Acetate is readily degraded when aqueous suspensions of TiO2 and acetate are irradiated in the presence of molecular oxygen [11, 51] with the degradation rates depending strongly on the pH of the suspension. In acidic suspensions (pH 3.0) formate and formaldehyde have been detected as the only products of the photocatalytic oxidation of acetate, while in alkaline suspension (pH 10.6) the main products are glycolate and formate accompanied by smaller amounts of glyoxylate and formaldehyde. Comparing this product distribution with the product distribution obtained in homogeneous solutions upon oxidation of acetate with hydroxyl radicals or by direct one-electron oxidation, e.g., on a Pt electrode, shows that both oxidising species contribute to the photocatalytic oxidation of acetate [52]. It has been established in detailed radiation chemical investigations that hydroxyl radicals attack acetate ions mainly at

Photocatalytic Detoxiﬁcation of Water and Air

the methyl group according to [53]: CH3 COO– + · OH → · CH2 COO– + H2 O

377

(23)

In the presence of air the radicals thus formed react quickly with molecular oxygen leading to the products given in reaction (24) [54]. · CH2 COO– + O2 → · O2 CH2 COO– →→→ (OCH2 COO– )2 , CHOCOO– , CH2 OHCOO– , CH2 O

(24)

Direct oxidation of acetate results in the well-known Kolbe decarboxylation with the formation of methyl radicals (reaction 25) [55–60]. CH3 COO– + h+ → CH3 COO· → CH3 · + CO2 (25) A considerably different product distribution results when these methyl radicals react with oxygen (reaction 26 [55–60]. CH3 · + O2 → CH3 OO· →→→ CH3 OOH, CH3 OOCH3 , CH2 O, CH3 OH, HCOO–

(26)

Figure 2 summarises both described pathways. The formation of glycolate and glyoxylate during its photocatalytic oxidation has been taken as evidence for the photocatalytic oxidation of acetate via hydroxyl radicals. The relative importance of this reaction path seems to be higher with increasing pH. In alkaline suspensions the surface of the TiO2 particles is negatively charged (pHZPC = 6.0–6.4) and the resulting electrostatic repulsion should hinder the adsorption of the negatively charged car-

Fig. 2 Proposed reaction mechanism for the oxidation of acetate by h+VB or • OHS , respectively (adopted from [51])

378

P.K.J. Robertson et al.

boxyl group of the acetate anion thus favouring an attack of surface bound hydroxyl radicals onto the methyl group. On the other hand, negatively charged carboxyl groups are directed towards positively charged surface groups of the semiconductor particles at pH values below the pHZPC and an attack leading to the subsequent decarboxylation of the acetate molecule is favoured. It should be noted that the formation of formate does not unambiguously indicate that the oxidation of acetate occurs also via a direct electron transfer from the carboxylate group. Formate itself is the main oxidation product of glycolate and glyoxylate and thus a secondary reaction product of the photocatalytic oxidation of acetate. Furthermore, it is evident that in acidic suspensions of TiO2 only formaldehyde and formate are formed during the photocatalytic oxidation of acetate. Here a different mechanism appears to be operative, probably a direct oxidation of the acetate molecule via holes. It can be concluded that the formation of glycolate and glyoxylate during the photocatalytic oxidation of acetate strongly suggests that hydroxyl radicals are formed on TiO2 surfaces upon band-gap illumination [11, 51]. An additional support of hydroxyl radicals as reactive oxidants is the observation that the intermediates detected during the photocatalytic degradation of aromatic compounds in the presence of titanium dioxide are typically hydroxylated structures [61]. These intermediates are consistent with those found when similar aromatics are reacted with a known source of hydroxyl radicals. In addition, esr studies have veriﬁed the existence of hydroxyl radicals in aqueous solutions of irradiated titanium dioxide [62–65]. Mao et al. have found that the rate of the oxidation of chlorinated ethanes correlates with the C – H bond strengths of the ethanes under investigation which indicates that the abstraction of hydrogen by a hydroxyl radical is an important factor in the rate-determining step of the photocatalytic oxidation of this class of organics [66]. On the other hand, these authors have observed that trichloroacetic acid and oxalic acid (compounds which have no hydrogen atom available for abstraction by a hydroxyl radical) are oxidised primarily by valence-band holes via a photo-Kolbe reaction [66]. Kinetic isotope work by Cunningham and Srijaranai [67] and Robertson et al. [68] also provides evidence for hydroxyl radical attack. Cunningham and Srijaranai [67] observed a primary kinetic isotope effect of 3.3 for the destruction of isopropanol using TiO2 . A similar effect of 3 was reported by Robertson [68] for the photocatalytic destruction of the cyanotoxin, microcystin-LR. The results of both studies suggest that the formation of the hydroxyl species may be a rate limiting process in the photocatalytic process. It was proposed that the reduced rate of photocatalytic decomposition in D2 O was due to the lower quantum efﬁciency for the formation of · OD radicals on the TiO2 surface [67]. This would therefore result in a relatively lower surface concentration of · OD radicals on the TiO2 surface for subsequent attack on the target molecules. The lower rate of oxidation may, however, be due to the · OD radical having a lower oxidation potential compared to the

Photocatalytic Detoxiﬁcation of Water and Air

379

· OH radical and therefore having a reduced oxidising power. Whatever the reason for the inﬂuence of the kinetic isotope effect on the photocatalytic process, Cunningham proposed that such effects strengthened the supposition that the photogeneration of hydroxyl radicals was the rate determining process for the photocatalytic process. It is interesting that the magnitude of kinetic isotope effects observed by Cunningham and Robertson were so similar. Robertson [68] proposed that an additional possibility was that the destruction of the substrates may be mediated by hydroxyl radicals generated via the superoxide radical anion produced at the conduction band. This is subsequently hydrated or deuterated by the solvent. This may be rate determining since the O2 has to be generated at the conduction band prior to interaction with the solvent and subsequent formation of OH· or OD· species. Therefore the kinetic isotope effect could be due to the interaction of the solvent with the superoxide species rather than the attack on the toxin. If this is the case it was suggested that a similar kinetic isotope effect would be observed no matter what substrate was being destroyed. Further kinetic isotope studies will help elucidate the potential of this proposed mechanism. Interestingly other workers have also suggested the possibility that species (O2 · – , HO2 · and H2 O2 ) generated following conductance band electron transfer to oxygen were involved in photocatalytic oxidation processes [69, 70]. Lu et al. used 18 O2 to establish the involvement of such species in the destruction of chloromethane on TiO2 [71]. Richard found evidence that both holes and hydroxy radicals are involved in the photocatalytic oxidation of 4-hydroxybenzyl alcohol [72]. These results suggest holes and hydroxyl radicals have different regioselectivities in the photocatalytic transformation of this compound: hydroquinone is thought to result from the direct oxidation by a valence-band hole, dihydroxybenzyl alcohol from the reaction with a hydroxyl radical, while 4-hydroxybenzaldehyde is produced by both pathways. In the presence of a hydroxyl radical quencher, the formation of dihydroxybenzyl alcohol is completely inhibited while the formation of 4hydroxybenzaldehyde is only partly reduced. The strongest evidence for direct hole oxidation as the principal step in the photooxidation step comes from a recent study performed by Draper and Fox that failed to detect any of the expected intermediate hydroxyl radical adducts following diffuse reﬂectance ﬂash photolysis of several titanium dioxide/substrate combinations [73]. In each case where the product of hydroxyl radical-mediated oxidation was known to be different from that of direct electron transfer oxidation, the authors observed only the products of the direct electron-transfer oxidation.

380

P.K.J. Robertson et al.

5 Novel Mechanistic Concepts in Photocatalysis 5.1 The Antenna Mechanism When they studied a new kind of preparation method for Fe(III)-doped colloidal TiO2 photocatalysts, Wang et al. observed a clear superiority of their newly synthesised material in the photocatalytic oxidation of methanol [74, 75]. For a better understanding of this effect, the photocatalysts have been characterised by cryo-TEM and high resolution TEM (HRTEM), evidencing that the following three features are unique for these Fe(III)-doped TiO2 particles: (1) the particle sizes fall in the range of 2–4 nm; (2) the particles possess the crystal modiﬁcation of anatase; and (3) the particles tend to form three-dimensional networks along a given crystallographic orientation. Cryo-TEM is a combination of cryo techniques with TEM, which allows observation of the nanoparticles in their native environment [74, 75]. Thus, the observation by cryo-TEM indicates that three-dimensional TiO2 networks indeed exist in aqueous suspensions. Based on this observation, the authors propose a novel mechanism that is shown in Fig. 3, i.e., the so-called antenna mechanism. Even if the target molecule is adsorbed on a photocatalyst particle at a certain distance from the light-absorbing particle, the latter can transfer the energy from particle to particle provided these particles are aggregated and possess the same crystallographic orientation. Once the energy has reached the particle with the adsorbed target molecule like methanol, the latter will act as a hole trap thus inducing the separation of the original exciton. Thus, a long chain of TiO2 -particles aligned as shown in Fig. 3 will act as an antenna system transferring the photon energy from the location of light absorption to the location of reaction.

Fig. 3 The Antenna Mechanism (reproduced from [74] with permission from the Royal Society of Chemistry)

Photocatalytic Detoxiﬁcation of Water and Air

381

It is envisioned that this antenna effect plays a signiﬁcant role in all types of photocatalytic systems. Most currently used and commercially available photocatalyst powders, for example, consist of nanocrystalline primary particles that are aggregated to form secondary structures with dimensions in the micrometer range. A material with a strong electronic coupling between the primary nanoparticles should hence exhibit a more pronounced antenna effect and thus a higher photocatalytic activity. Similarly, such an effect will also be operative in nanocrystalline transparent photocatalytic coatings where the target pollutant molecule or micro organism will not necessarily need to be located at the very location of the photon absorption. Consequently, it should be possible to apply smart molecular engineering approaches to design photocatalytic systems exhibiting considerably higher activity than existing systems through an improvement of the electronic coupling between the semiconducting nanoparticles. This could lead to an important breakthrough in photocatalytic research since the photonic efﬁciency still represents the greatest bottleneck on the way from interesting laboratory results to a successful commercialisation of most photocatalytic systems. 5.2 Metal Nanocontacts The presence of platinum as an electron sink (Fig. 4) has long been known to enhance the separation of photogenerated electron-hole pairs [76] and, hence, to improve the photocatalytic efﬁciency in general [77–79]. However, most of these studies were focused on the inﬂuence of the extent of platinisation on the photocatalytic efﬁciency. Photodeposition of metals on TiO2 has been studied extensively. Wang et al. investigated the deposition of Au from AuCl4 – on illuminated TiO2 by means of TEM, HRTEM and XPS [80, 81]. Well-deﬁned Au clusters formed

Fig. 4 The effect of metal nanocontacts on photocatalytic processes (reproduced from [82] with permission from the American Chemical Society)

382

P.K.J. Robertson et al.

Fig. 5 Models of PtTi-S1 formed by photodeposition of Pt a and of PtTi-S2 formed by mixing of colloidal Pt, TiO2 b (reproduced from [82] with permission from the American Chemical Society)

on the particle surface by the growth of Au nuclei were identiﬁed. Similarly, the analogous preparation of platinum islands by photodeposition of Pt from PtCl6 2– should yield Pt clusters on the TiO2 surface (PtTi-S1, cf. Fig. 5a). Alternatively, when Platinum deposits are prepared by the mixing of colloidal Pt with excess colloidal TiO2 , the Pt particles should rather be surrounded by TiO2 particles (PtTi-S2, cf.Fig. 5b). A corresponding proposal for the structure of the Pt promotion centres in the two photocatalysts is given in Fig. 5. The in-situ growth of Pt clusters during photodeposition leaves little doubt that the electronic interaction between Pt and TiO2 in photodeposited Ptislands is stronger than in physically mixed ones. Thus, the former is expected to exhibit a stronger promotion effect than the latter under the same speciﬁc conditions. This has been veriﬁed by the CW photolysis in the presence of O2 and repetitive laser-pulse photolysis in the absence of O2 [82]. Wang et al. have investigated the photocatalytic performance of three photocatalysts including neat TiO2 , PtTi-S1 (i.e., photodeposited Pt-islands) and PtTi-S2 (i.e., physically mixed Pt- and TiO2 -particles), by measuring the quantum yield, φHCHO , of HCHO formed from aqueous methanol at pH 3.5 under different conditions [82]. The improvement of φHCHO by platinisation was found both in continuous (CW) UV-photolysis and after repetitive 351-nm laser-pulse photolysis. Upon exposure of O2 -saturated aqueous suspensions, containing one of the photocatalysts and methanol, to 300–400 nm CW light, HCHO was produced and identiﬁed quantitatively by HPLC analysis. Other products were not detected by HPLC. HCHO was not formed in the dark or during photolysis in the absence of the TiO2 -based samples, evidencing that the process is truly photocatalytic [82]. It was found that platinisation has a strong promoting effect on the formation of HCHO in TiO2 -based photocatalysis. φHCHO increases by up to 50–70% when compared with that value on neat TiO2 particles. Moreover, it was also observed that photochemical platinisation of the TiO2 particles yields a more efﬁcient photocatalyst (PtTi-S1) than mixing of the colloidal components of Pt and TiO2 (PtTi-S2). This was attributed to the structural dif-

Photocatalytic Detoxiﬁcation of Water and Air

383

ference between PtTi-S1 and PtTi-S2 including the dispersity of Pt on TiO2 , and the electronic interaction between Pt and TiO2 (cf. Fig. 5). Photocatalytic oxidation of aqueous methanol was also studied by applying repetitive 351-nm laser-pulse photolysis both in the presence and absence of oxygen as an electron acceptor [82]. In the oxygenated sample solutions, the quantum yields of HCHO formation were measured after 0.5-Hz illumination by 200 laser pulses (10–20 ns pulse width). In the presence of neat colloidal TiO2 , φHCHO was 0.032 ± 0.002, ca. 50% larger than for CW photolysis. Likewise, φHCHO was increased by a factor of ca. 1.5 when the laser-pulse photolysis was applied on suspensions containing the platinised photocatalysts (1 wt % Pt), regardless of their preparations. Most importantly, it was found that repetitive laser-pulse photolysis improved φHCHO substantially and by the same factor (ca. 1.5) over CW photocatalysis for all the photocatalysts studied although in both modes of photolysis the timeaveraged molar photon absorption rate was approximately the same (ca. 8 × 10–7 Einstein L–1 s–1 ) [82]. In general, the presence of O2 is considered as a prerequisite for efﬁcient TiO2 -photocatalysed oxidation of organic pollutants in wet systems [83]. Similarly, the gas-phase photocatalytic oxidation of 2-propanol to produce acetone and water on single-crystalline TiO2 proceeds only in the presence of co-adsorbed O2 [84]. Contrary to expectation, removal of O2 by purging the suspensions with N2 resulted in only a small decrease of φHCHO (ca. 15%) for neat TiO2 and no change for PtTi-S1. A somewhat larger decrease of φHCHO (ca. 30%) was observed for PtTi-S2 [82]. Although O2 may not have been removed completely by purging with N2 , the result suggests that in (nearly)

Fig. 6 Mechanism of TiO2 -photocatalyzed HCHO in the presence of O2 saturation a and via current-doubling b in the presence of e– -acceptor A (H+ and/or residual O2 ) in N2 purged suspension. (1) Photogeneration of the charge carriers e– and h+ ; (2) and (3) surface trapping of the charge carriers; (4) and (5) recombination channels; (6) e– -transfer to the acceptor: formation of O2 –· /HO2 · (a) or A–· ; (7) ﬁrst oxidation step of CH3 OH; (8) formation of HCHO by e– -transfer from · CH2 OH to O2 ; (9) formation of HCHO by e– injection from · CH2 OH into the conduction band of TiO2 (current-doubling) (reproduced from [82] with permission from the American Chemical Society)

384

P.K.J. Robertson et al.

oxygen-free suspensions a species A other than O2 acts as the photoelectron acceptor, and the · CH2 OH intermediate is converted to HCHO via a different route. This is illustrated in Fig. 6. A potential e– -acceptor in the absence of molecular oxygen is the hydrogen ion (A in Fig. 6b) present in the suspensions at pH 3.5. Reduction of H+ to give H2 requires the transfer, on the same TiO2 particle, of two electrons from the conduction band or state which is ca. 30 mV [23, 24] below the conduction band. The two-electron requirement is met even after the absorption of just a single photon in case of current-doubling (Fig. 6b). The reduction of H+ to give H2 is thermodynamically allowed even at unity partial pressure of H2 , since the potential of conduction-band electrons in TiO2 particles is – 0.12 ± 0.02 V vs. the hydrogen electrode at the same pH [85]. In fact, H2 evolution has been observed upon UV-illumination of a deoxygenated TiO2 sol at pH 3 [85]. Dispersed Pt formed as a cluster on the TiO2 particles in PtTi-S1 (Fig. 5a) should strongly catalyse the H2 formation, while the compact Pt particles in PtTi-S2 (Fig. 5b) should be the poorer electrocatalysts for H2 evolution. From these considerations and the assumption that the reduction of H+ is ratecontrolling for the formation of HCHO in an O2 -free suspension, it can be understood why φHCHO on PtTi-S1 is found to be larger than on PtTi-S2 [82]. 5.3 The De-Aggregation Concept Regardless of the photocatalyst employed, repetitive laser pulse illumination increased φHCHO in the oxygenated suspensions by a factor of ca. 1.5 in comparison with CW illumination, although in both modes of photolysis the time-averaged photon absorption rate was approximately the same [82]. Previous studies on intermittent illumination in TiO2 photocatalysis offer no satisfying explanation for this observation [86–91]. Cornu et al. investigated the quantum yield, φF , of photocatalytic formate oxidation in oxygenated acidic TiO2 suspension under continuous and periodic pulse illumination [91]. For very short illumination times in periodic pulse photolysis, Cornu et al. demonstrated that φF was identical to the value obtained under continuous illumination at the same time-averaged photon absorption rate. This is in sharp contrast to the results of Wang et al. for photocatalytic methanol oxidation [82]. As a tentative explanation of the increase of φHCHO obtained by repetitive laser pulse illumination the latter authors proposed laser-pulse stimulated deaggregation of the TiO2 particle agglomerates [82]. Such a process is expected to increase the surface area available for reactant adsorption at a given photocatalyst loading and would thus lead to an increase of the photocatalytic reaction rate, and of φHCHO , upon applying a large number of laser pulses at a given average photon absorption rate.

Photocatalytic Detoxiﬁcation of Water and Air

385

Fig. 7 Laser-pulse induced deaggregation of a close packed TiO2 particle tetramer resulting in single particles with additional sites for reactant adsorption (reproduced from [82] with permission from the American Chemical Society)

The energetic requirement for deaggregation is met under the laser-pulsed conditions used by Wang et al. [82]. The average number of photons absorbed per TiO2 particle per laser pulse was calculated to be 0.3. The average photon energy deposited in the particle (Fig. 7) is therefore at least 341 kJ mol–1 per laser pulse at 351 nm. The inter-particle bond energy is only ca. 30 kJ mol–1 [92]. The photocatalytic formation of HCHO from CH3 OH + O2 is exothermal. The energy supplied by the ﬁrst few of the 200 pulses applied for photolysis would therefore be sufﬁcient for the supposedly stepwise deaggregation of even larger agglomerates than the tetramer. Reaggregation of the particles is a slow process. This is concluded from the temporal evolution of small UV-spectral changes observed after suspending the neat TiO2 colloid and from a kinetic study [93] of reversible agglomeration of TiO2 nanocolloids. Diffusion of reactants is not rate controlling under the conditions of this and related studies of photocatalytic processes [91]. In CW photolysis the absorption of a photon is a rare event compared with the high-intensity illumination by the laser pulse. The aggregation remains intact under CW illumination because there is enough time for redistribution of the photon energy among the many vibrational modes before the same aggregate absorbs a second photon.

6 Photocatalytic Treatment of Waste Water 6.1 Haloaromatic and Aliphatic Compounds The ﬁrst reported application of semiconductor photocatalysis for the destruction of an organic compound in water was performed by Carey et al. [94] in 1976. The successful degradation of biphenyl and chlorobiphenyls in the presence of TiO2 was investigated although complete mineralisation of the

386

P.K.J. Robertson et al.

compounds was not reported. Fox and Chen [95] subsequently reported the oxidative cleavage of several aromatic hydrocarbons in irradiated TiO2 suspensions. They developed this investigation further to include a number of other compounds such as alkanes, alkenes, arenes, acids, esters, amides, amines, ethers, thioethers, organohalides, aldehydes, ketones, organosilanes and phosphonates [96–99]. Since these early reports the destruction of a vast range of compounds in water using TiO2 photocatalysts have been investigated including alkanes, haloalkanes, aromatics, haloaromatics, surfactants, agrochemicals, dyes, explosives and oil ﬁeld chemicals. The ﬁrst complete mineralisation of a compound in water was reported by Ollis et al. [100, 101] in 1983. Using a TiO2 photocatalyst, the complete destruction of several halogenated hydrocarbons including trichloroethane, methylene chloride, chloroform and carbon tetrachloride was achieved. A detailed kinetic analysis of the destruction was carried out as part of this work and it was found that a simple Langmuirian rate equation represented the destruction of these materials. The use of the Langmuir model for photocatalytic destruction of organic compounds has been subsequently applied by a number of researchers with varying degrees of success. Halogenated aromatic and aliphatic compounds probably represent the most comprehensively studied class of compound for which photocatalytic water treatment has been applied. Mills et al. performed extensive investigations into the photocatalytic degradation of 4-chlorophenol. These included studies on the effects of different titania samples [102], effects of annealing temperature on the photocatalytic efﬁciency of titania [103] and a mechanistic study of the decomposition process. The rate of chlorophenol destruction was found to drop when using titania photocatalysts that had been heated above 600 ◦ C. This was believed to be due to a build up of the rutile phase and a reduction of surface area following heat treatment above these temperatures. A number of intermediates were reported including 4-chlorocatechol, hydroquinone, benzoquinone and 4-chlororesorcinol [104]. The applicability of the Langmuir model originally developed by Turchi and Ollis [56] was also investigated for this process. This model provided a good ﬁt with the experimental data. In a comparative study of commercial and laboratory prepared specimens it was suggested that no intermediates were generated on the lab prepared samples. Theurich et al. [59] in a study of the degradation of 4-chlorophenol also found the kinetics of the photocatalytic destruction process could also be described by the L-H model. Although the rate of 4-CP was uninﬂuenced by pH it was found the level of mineralisation was particularly sensitive to changes in this parameter. The mechanism of destruction also appeared to be affected by the nature of titania used. With Degussa P25 seven intermediates were detected while with Hombicat UV100 only three by-products were generated (Scheme 1).

Photocatalytic Detoxiﬁcation of Water and Air

387

Scheme 1 Mechanism for the photocatalytic destruction of 4-chlorophenol (reproduced from [59] with permission from the American Chemical Society)

The complete mineralisation of pentachlorophenol has been achieved within 3 h [58]. As with chlorophenol both quinoidal and hydroquinone byproducts were detected which again suggested hydroxyl radical attack as the primary oxidation route. Subsequent ring cleavage was reported to be slow with cleavage products including acetate and formate being detected. Wang and Hong investigated the effect of additives such as hydrogen peroxide, periodate and persulphate on the photocatalytic destruction of 2-chlorobiphenyl in water [105]. No rate enhancement for the destruction process was observed when any of these compounds were used as additives. In this study greater destruction rates of the chlorobiphenyl were in fact

388

P.K.J. Robertson et al.

achieved in the absence of titania. Polychlorinated biphenyls in sediment and soil slurries have also been degraded using TiO2 irradiated with simulated solar light [106]. In a 24 h period up to 81% of PCB contaminants in sediment from the St. Lawrence River were successfully degraded. The photocatalytic destruction of haloaliphatic compounds has been studied in detail by Hoffmann and co-workers [107, 108]. A detailed study of the mechanism of the photocatalytic destruction of chloroform using TiO2 was investigated by Kormann et al. [107]. The process was believed to be initiated by hydrogen abstraction by the hydroxyl radicals generated on the surface of the TiO2 (Eqs. 27–32). The chloroform radical subsequently reacts with oxygen forming trichloromethanol oxidation to trichloroformaldehyde and ultimately complete mineralisation. · OH + CHCl3 → · CCl3 + H2 O · CCl3 + O2 → · CCl3 O2 2· CCl3 O2 → 2CCl3 O· + O2 CCl3 O· + HO2 · → CCl3 OH + O2 CCl3 OH → Cl2 C = O + H+ + Cl– Cl2 C = O + H2 O → CO2 + 2H+ + 2Cl–

(27) (28) (29) (30) (31) (32)

Choi and Hoffmann demonstrated that hydrogen abstraction by hydroxyl radicals was a major hole scavenging process in the photocatalytic destruction of chloroform by TiO2 . The efﬁciency of this process was also found to be greatly enhanced through the addition of alcohols or organic acids which acted as electron donors. Hydrogen abstraction appears to be an important step in the degradation of carbon tetrachloride [108]. The effect of the addition of 0.1 M solutions of seven different alcohols as electron donors was investigated. Signiﬁcant rate enhancements were observed in each case particularly for isopropanol and 2-butanol. The dechlorination rate was also found to be pH dependent maximising at pH 6. Under the acidic conditions a number of by-products were detected which were not observed in the basic solutions. Crittenden et al. [109] reported that modiﬁcation of Aldrich TiO2 with platinum and silver signiﬁcantly improved the efﬁciency of the destruction of chlorinated hydrocarbons in water. No such enhancement was observed for Degussa P-25. The rates for trichloroethylene destruction with TiO2 samples containing 1% Ag and Pt, were three and two times that observed for unmodiﬁed titania. It was proposed that this enhancement in photocatalytic activity was probably due to an inhibition of the recombination of the electron/hole pair. A similar lack of enhancement in the degradation rate of trichloroethylene and chloroform was observed when Degussa P-25 TiO2 was modiﬁed with Pt and Pd by Chen et al. [110]. A reasonable acceleration in the destruction of

Photocatalytic Detoxiﬁcation of Water and Air

389

methanol and ethanol was, however, observed. The presence of both Pt and Pd was found to enhance the oxygen reduction reaction on the titania surface. 6.2 Phenols and Polyaromatic Hydrocarbons Phenolic compounds are common pollutants in a range of industrial waste waters and have contaminated ground waters in regions with signiﬁcant levels of industrialisation. The photocatalytic destruction of phenol has been the subject of extensive investigations [69, 111–115]. The mineralisation of a 10 ppm phenol solution using TiO2 suspensions activated by sunlight in shallow pools was demonstrated to occur within 110 min. Not surprisingly the rate of phenol destruction depended on the season with rates in winter roughly half that observed for the summer and autumn. Interestingly this destruction rate did not appear to be related to the irradiance as the summer irradiance was more than twice that observed in winter and 1.6 times that of autumn levels. Pelizetti et al. detailed the photocatalytic mineralisation of ortho-, paraand meta-cresols [116]. As with chlorophenol, methyl catechol and hydroquinone were detected as major by-products. The o- and p-cresols were observed to degrade with ﬁrst order kinetics while the m-cresol degradation followed zero order kinetics at pH 3. Under 3 h photocatalysis time in alkaline conditions m-cresol kinetics become ﬁrst order which was proposed to be due to lower surface coverage of the cresol at this pH due to electrostatic repulsion with the titania surface. In air saturated solutions the time for mineralisation was around 8 h. The mineralisation time was, however, reduced to 2.5 h in suspensions sparged with oxygen. Catechol and hydroquinone by-products were also detected in the photocatalytic destruction of nitrophenols [117]. Ninety-nine percent destruction of 4-nitrophenol was achieved within 3 h, although for complete mineralisation up to 9 h irradiation was necessary. The presence of oxygen not only affected the level of mineralisation of the nitrophenol but also the extent of denitration, believed to be the major initial step of the photocatalytic process. Hydrogen peroxide was reported to signiﬁcantly improve the rate of photocatalytic process increasing the destruction rate by up to 3 fold. Addition of Cu ions, however, had a detrimental effect [118]. Pichat et al. [119] previously reported an increase in the apparent ﬁrst order rate constant with increasing hydrophobicity of a series of chlorophenols. In a follow up study of meta and para substituted methoxybenzenes no simple correlation was observed between the hydrophobicity (Kow ) of the compound and the ﬁrst order rate constant for the photocatalytic decomposition process [120]. They proposed this discrepancy was possibly due to an interaction between adsorbate molecules resulting from van der Waals forces.

390

P.K.J. Robertson et al.

Polyaromatic hydrocarbons are generated by a range of industrial activities from town gas generation to activities of the oil and gas industry. These compounds can be very persistent and are often resistant to biodegradation. Semiconductor photocatalysis has been successfully applied to the destruction of a wide range of these compounds, although not surprisingly the decomposition rates are signiﬁcantly lower than that observed for simpler compounds [121]. For example in a study of the destruction of 16 polyaromatic hydrocarbons in aqueous suspensions of TiO2 the t1/2 varied from 2.7 h for anthracene to 380 h for ﬂuorene [121]. Dass et al [122] studied the photocatalytic destruction of a series of polyaromatic hydrocarbons including anthracene, naphthalene, ﬂuorene and acenaphthene using both sunlight and artiﬁcial sources. From isolated by-products it was proposed that hydroxyl radical attack was the primary mechanism of the degradation process. A more detailed mechanistic study of the photocatalytic degradation of naphthalene has been reported by Theurich et al. [60]. This process appeared to be initiated by hydroxyl radical attack on the structure generating a naphthol product. Subsequent hydroxyl radical attack resulted in the generation of diols and ortho and para-naphthoquinones. Further photocatalysis subsequently induced ring cleavage with coumarin and endoperoxides being detected as by-products. 6.3 Agrochemicals TiO2 photocatalysis has demonstrated efﬁcacy against a range of agrochemicals such as herbicides and pesticides. One of the ﬁrst reports detailed the mineralisation of the pesticide permethrin in TiO2 hexane/water slurry. Using solar irradiation 90% photodegradation of a 17 000 ppm solution was achieved in 8 h [123]. The use of solar powered systems for destruction of a mixture of pesticides in well water including dichloroaniline, benzopyran, atrazine, propazine, alachlor, prometryn, bromacil and cyanobenzoate has been achieved [124]. Variable efﬁciencies for the decomposition of these compounds was observed with over 90% of dichloroaniline destroyed in 10 h compared to 50% of propazine in the same period. In general for most of the pesticides examined the half life time for the process in the 100 ppb concentration range was less than 1 h. The contact insecticide acrinathrin has been completely mineralised by TiO2 under sunlight within 8 h. The addition of peroxydisulphate to the system resulted in a signiﬁcant enhancement of the process efﬁciency with complete mineralisation achieved within 2 h [125]. The solar photocatalytic destruction of Aldrin has been reported by Bandala et al. using both concentrated and nonconcentrated solar processes [126]. As with other processes the addition of peroxide enhanced

Photocatalytic Detoxiﬁcation of Water and Air

391

the process although in this case only marginally at levels of 6%. Degradation products identiﬁed for this process included dieldrin, chlordene and 12-hydroxydieldrin. A mechanistic study of the photocatalytic degradation of cyromazine has been performed using LCMS to identify a range of by-products [127]. The initial step in the process involved ring cleavage and subsequent fragmentation into a range of short chain aliphatic compounds. This investigation, however, demonstrated that total mineralisation of the pesticide did not occur after 22 h photocatalysis. The ﬁnal product of the process cyanuric acid was resistant to further degradation and it was proposed that the removal of this compound could be achieved using microbiological processes. Hequet et al. reported that the ultimate product for the photocatalytic decomposition of atrazine using TiO2 was also cyanuric acid [128]. In this study the initial decomposition of atrazine appeared relatively efﬁcient with a half life of 20 min. Muneer et al. [129] examined the photocatalytic oxidation of three pesticide derivatives, propham, propachlor and tebuthiuron in aqueous TiO2 suspensions. The rates of degradation of each compound were found to be strongly affected by the type of TiO2 used, pH, catalyst and substrate concentration. For each compound several intermediate products were identiﬁed using GCMS. This study indicated that the photocatalytic oxidation process proceeded by reactions involving electron transfer, hydroxyl radical and superoxide radical anions. Scheme 2 displays the proposed mechanism for the photocatalytic decomposition of propham. Ninety-ﬁve percent mineralisation of the herbicides propane and pollinate has been achieved in 240 min using simulated solar irradiation of TiO2 suspensions [130]. Using GCMS a mechanism for this process was proposed which involved hydroxyl radical attack, dechlorination, dealkylation and subsequent oxidation. Zertal et al. [131] investigated the effectiveness of ﬁve commercially available TiO2 photocatalysts on the destruction of herbicide 4-chloro2-methylphenoxyacetic acid (MCPA) including Millennium PC50, PC100, PC105, PC500 and Degussa P25. The order of activity for the herbicide decomposition on these materials was found to be P25 > PC50 > PC500 > PC100 > PC105. No relationship between surface area and rate of photocatalytic destruction of MCPA was observed. The main intermediate of the photocatalytic destruction was found to be 4-chloro-2-methylphenol and the authors proposed two mechanistic pathways for the generation of this byproduct through hydroxyl radical attack and direct hole oxidation. The photocatalytic destruction of the fungicide 2-phenylphenol generated a range of by-products including hydroquinones, benzoquinones and dihydroxybiphenyls [132]. This indicated that the major routes of degradation included hydroxylation and scission of the phenyl-phenol ring. The rate of

392

P.K.J. Robertson et al.

Scheme 2 Proposed mechanism for the photocatalytic destruction of propham (reproduced from [129] with permission from Elsevier BV)

photocatalytic oxidation was found to be enhanced with increasing pH and addition of hydrogen peroxide. 6.4 Dye Compounds The photocatalytic destruction of the thiazine dye, methylene blue, is probably the most extensively studied application for dye destruction [133–136]. This is commonly used as a standard molecule in demonstration experiments of photocatalysis as it provides an excellent visual representation of the process. This molecule may, however, be a rather “ambiguous system” to quote Mills [135]. The reason for this is that it appears to undergo simultaneous reduction to a colourless leuco dye while also undergoing photocatalytic oxidation. Herrmann et al. investigated the degradation pathway of methylene blue and reported that the initial attack appeared to involve cleavage of the nitrogen-carbon double bond with subsequent hydroxyl radical attack on the aromatic rings of the by-products [136] (Scheme 3). The hydroxyl radical attack on aromatic structures appeared also to be an important step in the photocatalytic destruction of a series of other dye molecules including Crocein orange G, methyl red, Congo red and Alizarin

Photocatalytic Detoxiﬁcation of Water and Air

393

Scheme 3 Proposed mechanism for the photocatalytic destruction of methylene blue (reproduced from [136] with permission from Elsevier Science B.V.)

S [137]. TiO2 photocatalysis appeared to be effective for the decolourisation and detoxiﬁcation of water samples containing each of these dyes. The application of photocatalysis to textile chemicals such as the azo dyes acid red 14 and acid orange 8 and acid orange 7 has been reported [138–140]. Complete colour removal was achieved for each of these materials within 3 h while use of additives such as H2 O2 in millimolar quantities enhanced the destruction rates. The addition of ethanol, a hydroxyl radical quencher, was

394

P.K.J. Robertson et al.

found to inhibit the photocatalytic oxidation of acid red, suggesting again the involvement of hydroxyl radicals as the major route to photocatalytic destruction [138]. Stylidi et al. [140] suggested that the azo bond was the initial point of attack in the photocatalytic destruction of acid orange 7 under irradiation with visible light. A number of naphthaquinoidal compounds were detected as initial degradation by-products, which rapidly degraded to phthalic derivatives, aromatic and subsequently aliphatic acids on extended photocatalysis. The COD of the solution was, however, only reduced to 45% of the starting level after 120 h irradiation suggesting the system was only effective when coloured compounds were present to sensitise the titania. Hydrogen peroxide also enhanced the photocatalytic destruction of the triphenyl methane dye, gentian violet [141]. The degradation products were analysed using GCMS and it appeared that hydroxyl radical attack was again the primary oxidation pathway for this molecule. In the case of the indigoid dye acid blue 74 (indigo carmine) it was observed that the photocatalytic destruction on a series of anatase photocatalysts was in fact slower than where the solutions were treated with UV/H2 O2 [142]. Relatively poor kinetics were observed when rutile titania samples were used as photocatalysts. Virtually complete mineralisation of indigo and of indigo carmine using TiO2 irradiated with UV light was reported by Vautier et al. [143]. Interestingly, although irradiation of titania with visible light did result in decolourisation mineralisation was not achieved. The effect of metallic species on the photocatalytic degradation of acid orange-7, tartrazine and 3-nitrobenzenesulphonic acid was reported by Rao et al. [144]. Only silver ions demonstrated any degree of enhancement with little inﬂuence being observed with Cu2+ . Vanate, however, resulted in a reduction in photocatalytic efﬁciency. 6.5 Explosive Residues and Oilfield Chemicals Trinitrotoluene (TNT) contamination is a problem experienced in many former military installations in the United States and Europe [145, 146]. The destruction of TNT has been achieved in a TiO2 slurry reactor with 90% mineralisation occurring within 120 min [145]. Wang and Kutal also observed 95% mineralisation of TNT within 5 h photocatalysis time [147]. Under anaerobic conditions, although rapid photodegradation of the compound was achieved mineralisation was not observed. Dillert et al. [146] investigated the photocatalytic destruction of a variety of nitroaromatic compounds in TiO2 suspensions. The effectiveness of the process was variable depending on the target molecule with the nitro groups inﬂuencing the attack of the electrophilic reagents on the aromatic group. In a subsequent study it was reported that the rate of photocatalytic destruction of TNT was signiﬁcantly faster than direct photolysis or UV/H2 O2 [148]. The rate of the UV/TiO2 sys-

Photocatalytic Detoxiﬁcation of Water and Air

395

tem was relatively unaffected by pH, however, the addition of H2 O2 inhibited the process. Schmelling et al. also observed that altering pH and the presence of inorganic anions and humic compounds did not signiﬁcantly alter the rate of TNT destruction but did affect the level of mineralisation [149]. These results indicate that TiO2 photocatalysis is a very effective technique for the destruction of these highly toxic and explosive materials. The decomposition of crude oil slicks on the surface of water has been achieved using TiO2 coated to plastic spheres or other ﬂoating substrates. Berry and Mueller achieved complete mineralisation of hydrocarbons on the surface of water using TiO2 coated on pine wood chips and irradiated with sunlight [150]. Using a titania modiﬁed with KOH complete oil decomposition was achieved within 2 h irradiation [151]. Ziolli and Jardim [152] also found that the use of TiO2 signiﬁcantly increased the rate of the water soluble fraction of crude oil in sea water on photolysis, reducing the time for decomposition from six days irradiation to between one and two days. This technology may be of potential use to the oil and gas industry for destruction of hydrocarbons in produced waters. This has been applied to oil ﬁeld produced waters from the Campos Basin in Brazil and ﬁelds in the North Sea [153, 154]. A 90% reduction in hydrocarbon content was observed following a 60 min treatment period for the Brazilian sample, while for the North Sea sample this was achieved in 75 min. A problem with such applications, however, is the volume of water generated in many oil ﬁelds. For example, typically ﬁelds in the North Sea may produce between 103 and 105 m3 of contaminated water per day. This would require large reactor volumes which would be prohibitive in off-shore applications. 6.6 Potable Water Applications In recent years scarcity of safe drinking water has become a signiﬁcant challenge to the existence of many communities throughout the world with the Middle East and Southern Europe experiencing particular problems. Disease is a common consequence of water shortage with diarrhoea caused by ingestion of contaminated water being a signiﬁcant problem and killer particularly for children under ﬁve years old. In addition new challenges in the form of cyanotoxins in drinking water are presenting serious problems to the water industry as existing technologies are not always effective in the removal of such compounds. Titania photocatalysis has, however, proven to be rather effective in removal of a wide range of chemical compounds and microbiological pathogens from potable water samples. Humic compounds are naturally occurring compounds which are present in many water supplies from which drinking water is abstracted. These compounds must be removed prior to distribution of the potable water as not only do they colour the water, but they may solubilise pesticides. Eggins

396

P.K.J. Robertson et al.

et al. [155] demonstrated the rapid decomposition of humic substances using a TiO2 photocatalyst. The effectiveness of the technique for the removal of humic substances has been subsequently conﬁrmed by other workers indicating the overall effectiveness of the process for the removal of such compounds [156, 157]. Cyanobacterial toxins produced and released by cyanobacteria in freshwater around the world are well documented [158, 159]. Microcystins are the most common of the cyanobacterial toxins found in water, as well as being the ones most often responsible for poisoning animals and humans who come into contact with toxic blooms and contaminated water [160]. Acute exposure results in hepatic injury, which can in extreme cases prove fatal. One such incident occurred that resulted in the death of over 50 dialysis patients due to the use of microcystin-contaminated water in the treatment [161]. Chronic exposure due to the presence of microcystin in drinking water is thought to be a contributing factor in primary liver cancer (PLC) through the known tumour-promoting activities of these compounds [162]. The use of a titanium dioxide photocatalyst for the removal of microcystinLR in water has been demonstrated by Robertson et al. [163]. They reported a rapid photocatalytic degradation of this toxin using a Degussa P25 photocatalyst. In a subsequent study [68] a primary kinetic isotope effect of approximately 3 was observed when the destruction was performed in a heavy water solvent. Hydroxylated compounds were observed as products of the destruction process, while no destruction was observed when the process was investigated under a nitrogen atmosphere. A more detailed mechanistic study of the photocatalytic destruction of microcystin showed that the toxin disappearance was accompanied by the appearance of seven UV-detectable compounds [164]. Six of the observed reaction products did not appear to undergo further degradation during prolonged photocatalysis (100 min). The degree of mineralisation of the toxin was found to be less than 10%. Since more than 90% of the toxin was not mineralised the toxicity of the degradation products were subsequently assessed using brine shrimp and protein phosphatase bioassays [165]. This study established that the photocatalytic process removed any residual toxicity from the water together with potential tumour promoting activity. The mechanism of the photocatalytic destruction of microcystin-LR was studied in more detail using liquid chromatography-mass spectrometry analysis [166]. This investigation demonstrated that the major destruction pathway of the toxin appeared to be initiated via three mechanisms: UV irradiation, hydroxyl radical attack, and direct oxidation (Scheme 4). Shepard et al. [167] also reported the use of TiO2 photocatalysis for the degradation of three microcystin variants, -LR, -YR and -YA. Each of these toxins rapidly decomposed on photocatalysis and a half life of less than 5 min was reported for each microcystin. The effect of pH was found to strongly inﬂuence the destruction of a series of microcystins including LR, RR, LW and

Photocatalytic Detoxiﬁcation of Water and Air

397

Scheme 4 Proposed major pathway for the photocatalytic destruction of microcystin-LR (reproduced from [166] with permission from The American Chemical Society)

398

P.K.J. Robertson et al.

LF [168]. This pH dependence can be associated with both changes to surface charge of the photocatalyst and altered hydrophobicity and net charge on the toxin. These ﬁndings show that TiO2 photocatalysis is a very effective method for the removal of these potentially hazardous compounds from drinking water. When considering large-scale application of this process for removal of microcystins from potable waters, optimisation of reaction conditions must be performed with a range of representative microcystin variants. The presence of oestrogenic compounds in potable water supplies has also raised concerns in recent years. The destruction of 17-β-oestradiaol on immobilised TiO2 ﬁlms has been achieved with 98% decomposition observed in 3.5 h photocatalysis [169]. 6.7 Disinfection of Pathogenic Micro-Organisms To date, numerous studies have been carried out demonstrating the germicidal effects of titanium dioxide photocatalysis [170–176]. The vast majority of theses focus on the destruction of bacterial pathogens in particular Escherichia coli, most likely due to the fact that coliforms are traditionally used as indicators of faecal contamination in water supplies. Coliform bacteria may not necessarily cause disease, but can be indicators of pathogenic organisms that do. The latter could cause intestinal infections, dysentery, hepatitis, typhoid fever, cholera and other illnesses. Intestinal infections and dysentery are generally considered minor health problems. They can, however, prove fatal to infants, the elderly, and those who are ill. However, the effect of TiO2 photocatalysis on many other bacterial species [177–181], several viruses [170, 182], algae [183], fungi [170, 184] and protozoa [185] as well as on the destruction of bacterial spores [186] and toxins [187] has also been investigated. Both the hydroxyl radical and superoxide radical anion are believed to be involved in the photocatalytic disinfection reaction (Fig. 8) [171]. Matsunaga et al. [170] were the ﬁrst to report on the “novel concept of photochemical sterilisation”. They looked at the destruction of Lactobacillus acidophilus, Saccharomyces cerevisiae and Escherichia coli (103 cells mL–1 ) using platinum-loaded titanium dioxide and showed that cell death occurred following incubation with TiO2 /Pt particles under metal halide lamp irradiation for 60–120 min. This, along with other early work highlighted the importance of several parameters namely TiO2 concentration, UV light intensity, microbial starting concentration, temperature, pH and aeration on experimental outcome. The photocatalyst may be used in the form of an aqueous suspension or on a solid support. A suspension must be constantly stirred to ensure mixing of catalyst and target species and to prevent settling of the catalyst. Reaction mixtures are often aerated to scavenge electrons and prevent electron/hole

Photocatalytic Detoxiﬁcation of Water and Air

399

Fig. 8 a–c Schematic illustration of the process of E. coli photokilling on TiO2 ﬁlm. In the lower row, the part of the cell envelope is magniﬁed (reproduced from [171] with permission from Elsevier BV)

recombination from occurring. Furthermore the use of sonicated TiO2 has been shown to be very effective in increasing the rate of bacterial destruction [173, 177, 182] as well as causing greater intracellular damage in damaged bacterial cells. This effect is thought to be due to the increased surface area of the catalyst which makes it more available to react with target species and/or the ability of smaller TiO2 particles to enter bacterial cells with already damaged cell membranes more readily and cause additional damage to intracellular components. The disadvantages of this technique are that at the end of the treatment period the catalyst needs to be separated from the liquid (water). Furthermore, the resulting opacity of the slurry may interfere with UV light penetration through the solution. The use of TiO2 ﬁlms has been investigated by several authors and found to be as effective as the suspended form [171, 174, 175, 188]. This method obviously overcomes the disadvantage of having to separate catalyst from liquid at the end of the experimental period. Unfortunately, UV intensity has not been standardised and therefore great variation exists among the UV intensity used among reported works with some studies making use of natural sunlight [179] while others employ a UV lamp light source [173, 177, 184, 189]. It is therefore difﬁcult to directly compare the effects of UV intensity among different studies since other experimental parameters such as, for example, starting inoculum, pH or aeration, may vary. However, experiments in which all other parameters have remained

400

P.K.J. Robertson et al.

constant have shown that the rate of microbial destruction increases with increasing UV light intensity [172–174, 190, 191]. In terms of the effects of microbial starting concentration, the general trend appears to be that the higher the initial concentration of micro organisms the longer the period of time required for inactivation [172, 192]. However, Dunlop et al. [191] who used an applied electrical potential to enhance the photocatalytic destruction of E. coli K12, showed that at a high bacterial starting concentration the effect of the applied potential was greatest and suggested that higher bacterial numbers increased the likelihood of an interaction occurring between the catalyst and the bacterium. As well as the importance of the starting inoculum concentration, a recent study has suggested that the generation of the bacteria is also important. Rincon et al. [192] have shown that E. coli harvested at the third generation of culture were less sensitive to irradiation than those harvested at the seventh generation. It is well known that bacteria undergo mutations with each successive sub-culture, these authors suggest that such changes make the bacteria more susceptible to photocatalytic attack. An important parameter when working with micro organisms is pH. There is a pH optimum for each micro organism at which growth is maximal. Moving away from the pH optimum in either direction slows microbial growth. Thus the pH of the reaction mixture during photocatalytic treatment is of utmost importance. Most micro organisms are neutrophiles i.e. they prefer a neutral pH, the vast majority of human pathogens fall into this category. Most of the reported photocatalytic reactions have been performed in the pH range 5–8 which does not seem to overtly affect experimental outcome [173, 174, 182]. Herrera Melián et al. [179] examined the photocatalytic destruction of coliform bacteria and Streptococcus faecalis in urban waste water using both direct solar and UV lamp light. They found no signiﬁcant differences in bacterial destruction between TiO2 photocatalysis and UV light only (either solar or UV lamp light source) at natural pH (7.8). If, however, the pH was lowered to 5, the inactivation rate was increased in the presence of the catalyst. Watts et al. [182] on the other hand found that the photocatalytic destruction of coliform bacteria and poliovirus in secondary waste water efﬂuent was unaffected by pH changes in the range 5–8. The role of oxygen in the photocatalytic destruction process is also unclear. For some studies oxygen was found to enhance, or in some cases be a prerequisite for, photocatalytic-induced cell death, although the optimum concentration required is equivocal [173, 190]. However, Herrera Melián et al. [179] showed that it was not necessary to bubble oxygen into a waste water sample undergoing photocatalytic treatment since comparison of an airbubbled sample and a stirred sample resulted in similar rates of destruction. Studies in which additional parameters have been included as a means of enhancing the photocatalytic efﬁciency include the use of an applied electrical potential or the addition of hydrogen peroxide to the reaction mix-

Photocatalytic Detoxiﬁcation of Water and Air

401

ture [183, 186, 191]. Interestingly, Butterﬁeld et al. showed that exposure of E. coli and Clostridium perfringens spores (103 CFU mL–1 ) to UV light and TiO2 catalyst was ineffective but application of an electrical ﬁeld resulted in complete destruction of E. coli and a signiﬁcant reduction in Cl. perfringens spores [186]. Dunlop et al. [191] also examined the photocatalytic destruction of E. coli in Ringers solution with the addition of an applied potential and reported a 99.996% reduction in bacterial numbers 120 min after treatment. Given the variability that exists among experimental set up and conditions it is not surprising that great variation exists among reported rates of destruction for many of these studies. For example Watts et al. [182] reported a 2 log order reduction in coliform bacterial numbers in a secondary waste water efﬂuent sample after 150 min photocatalytic treatment. Whereas Herrera Melián et al. [179] observed a 2 log order reduction in coliform number from waste water samples after 60 min photocatalytic treatment at pH 7.8 in an aerated system. Another important factor that may affect experimental outcome is the choice of culture media for enumerating bacterial recovery following photocatalytic treatment [192]. Compared to a nonselective agar, the use of selective agars has been shown to reduce the growth of even the target organism and therefore give misleading results. Thus, if it is deemed necessary to use a selective agar for example when treating a natural water sample, then it is also important to include a nonselective agar to conﬁrm results. To complicate matters further, however, recent works have shown that experimental outcome is also affected by organism type [184, 192]. It is well known that different organisms respond differently to the effects of UV light [193, 194] however it has recently been shown that the susceptibility of an organism to photocatalytic treatment is directly related to its cell wall structure [184, 192]. Hence, Gram negative bacteria like E. coli and Pseudomonas aeruginosa were shown to be more sensitive to photocatalytic attack than Gram positive organisms like Staphylococcus aureus and Enterococcus faecium whereas the fungus Candida albicans was shown to be the most resistant to photocatalytic attack [184]. Much of the current work in this ﬁeld is now focused on conﬁrming the mode of microbial destruction. While several methods for bacterial destruction have been proposed the common factor appears to be initial cell membrane damage followed by destruction of intracellular components. In the ﬁrst study of its kind, Matsunaga et al. [170] showed that Coenzyme A, a coenzyme involved in a variety of biochemical reactions, was photoelectrochemically oxidised causing an inhibition in cellular respiratory activity and hence cell death. Saito et al. [180] reported a rapid leakage of potassium ions from Streptococcus sobrinus following treatment with TiO2 and UV light. This leakage occurred in parallel to cell death. Furthermore, they demonstrated the slow release of protein and RNA from cells during treatment time. Sunada et al. [187] examined the photocatalytic destruction of E. coli endotoxin, which is an integral component of the cell membrane,

402

P.K.J. Robertson et al.

using a thin TiO2 ﬁlm. They showed that death of E. coli cells was accompanied by degradation of endotoxin and suggested that their results showed that TiO2 photocatalysis destroys the outer membrane of the bacterial cell. Maness et al. [195] proposed that E. coli were killed by a mechanism known as “lipid peroxidation”. They suggested that the reactive oxygen species (OH· , O2 – and H2 O2 ) generated on the irradiated TiO2 surface act together to attack the polyunsaturated phospholipids in the bacterial cell membrane. This primary attack causes the cell membrane to break down, TiO2 particles may then gain access to the damaged cell via phagocytosis and further photocatalytic damage may occur inside the cell. Huang et al. [196], using E. coli as a model organism, suggested that cell death occurred as a result of rapid cell wall damage which was followed by cytoplasmic membrane damage and subsequent attack of intracellular components. This damage occurred within 20 minutes of treatment. It has become apparent from all of these studies that the primary target for photocatalytic attack is the cell wall. This fact goes some way towards explaining the observed differences in susceptibility of particular microbes to photocatalytic attack, i.e., the Gram positive cell wall while structurally less complex than that of the Gram negative cell wall, is surrounded a thick peptidoglycan layer and fungi, which were shown to be much less susceptible to photocatalytic attack than bacteria [184] have a thick eukaryotic cell wall. Once the integrity of the cell wall has been breached leakage of intracellular contents takes place which ultimately results in cell death. In order for TiO2 photocatalysis to be accepted as a reliable water disinfection method, a residual effect of treatment is necessary after initial treatment has taken place. Bacterial reappearance studies have shown that this may not be the case. Herrera Melián et al. [179] examined bacterial re-growth in natural waste water samples previously treated at pH 5 and pH 7.8, 24 to 48 h after photocatalytic treatment, a signiﬁcant rise in bacterial numbers was observed, although re-growth was slower in samples treated at pH 5. Wist et al. [189] looked at bactericidal activity of TiO2 after stopping illumination. They examined the destruction of E. coli in crude water samples compared to distilled water both during and after the irradiation period. They showed an initial decrease in bacterial numbers in both samples following a 1.5 h irradiation period. Twenty four hours after stopping irradiation there was no increase in bacterial cell number in the distilled water sample. However, there was a signiﬁcant increase in bacterial cell number in the natural water sample. The most likely reason that bacterial re-growth occurs is that not all of the bacteria have been killed during the photocatalytic treatment period thus continue to grow and reproduce after the treatment time. Conversely, Rincón et al. [192] looked at the residual effect of TiO2 photocatalysis of E. coli K12 in deionised water, and showed that even up to 60 h after photocatalytic treatment, there was no bacterial re-growth. Although they did point out that this was dependent on the light intensity used. Huang

Photocatalytic Detoxiﬁcation of Water and Air

403

et al. [174] also demonstrated a residual bactericidal effect although they only looked at samples which had been kept in the dark for 30 min after a 30 min irradiation period. On the other hand, Dunlop et al. [191] did not detect any bacterial re-growth 48 h after photocatalytic disinfection with an applied potential. They suggested that such treatment had caused irreversible damage to E. coli cells although did point out the possibility that cells may have entered a viable but nonculturable state and therefore would not be detectable by traditional methods. What these studies have also highlighted is that signiﬁcant differences exist between the photocatalytic response of microbes in natural water conditions and those under simulated laboratory conditions [179, 189, 192]. Dillert et al. [197] examined the photocatalytic treatment of raw and pre-treated municipal waste water, they showed that the total CFU decreased within 6 h from 106 mL–1 to 104 mL–1 and from 104 mL–1 to 102 mL–1 , respectively. However, in the same paper, they also examined the photocatalytic treatment of E. coli in a physiological sodium chloride solution and reported a reduction of 107 CFU mL–1 to 101 mL–1 after 1 h treatment time. It has recently been shown also by Rincon et al. [192] that the stage of growth at which bacteria are examined has a signiﬁcant effect on their response to photocatalytic treatment [192]. Stationary phase cultures were shown to be less readily inactivated than log phase cultures and since the bacteria present in natural water will all be at different growth stages at any one time, differences may exist in their susceptibility to photocatalytic treatment. Furthermore, this group looked at two waste water samples and found that considerable differences existed in photoreactivity of both samples taken at the same place but on different dates. Thus, it is clear that extrapolation of results obtained from the photocatalytic treatment of pure cultures of micro organisms, grown under uniform laboratory conditions, in distilled water samples to the natural situation is almost impossible. Furthermore, consideration should be given as to how results are actually reported. For example many authors report their results as % survival rates or % inactivation rates, however, caution must be taken when interpretating such results. For example, a reduction of 108 CFU mL–1 to 103 CFU mL–1 equates to a 99.99% inactivation yield. If such results are reported as % inactivation only this throws a very different light on the experimental outcome. Rarely do authors quote actual bacterial numbers when reporting results. This is a practice that, if adopted, would beneﬁt all researchers in this ﬁeld. While many authors claim to have achieved complete destruction of pathogens in water others have not. It should be remembered that in terms of acceptable rates of microbial destruction there is no tolerable limit for pathogens in water intended for consumption, for preparing food or drink or for personal hygiene. Bacterial contamination falls under the category of pathogens. The EPA Maximum Contaminant Level (MCL) for coliform bacteria in drinking water is zero (or no) total coliform per 100 mL of water.

404

P.K.J. Robertson et al.

The application of this technique to water treatment systems is hence restricted by several factors, most notably the length of time required to achieve bacterial destruction (best results appear to be achieved in those studies which have a low bacterial starting concentration i.e. 105 CFU mL–1 or less) and the possibility that re-growth may occur following photocatalytic treatment. Thus, unless a complete removal of all microbes from a treated water sample is obtained, followed by a period of no re-growth, the photocatalytic technique cannot be used as a reliable disinfection process.

7 Air Treatment TiO2 photocatalysis has also proven to be a most effective treatment process for air pollution control/abatement. This process has several advantages over other oxidation processes such as high destruction rates being achieved at ambient temperature and pressure with low-intensity UV, use of oxygen in air as an oxidant, is effective in destroying a broad range of common pollutants (Table 1), demonstrates potential for lower operating costs, elimination of reagents or electron acceptors such as H2 O2 and the possible recovery, regeneration and reuse of the photocatalyst. Over the past decade research into the use of semiconductor photocatalysis as a method of treating trace level organic contaminants in air has advanced signiﬁcantly [198–236]. Table 1 presents a representative sample of gaseous pollutants that have undergone effective photocatalytic treatment. Current research has shown that one of the main areas of industrial gaseous photocatalysis is indoor air treatment and this will be discussed below. The efﬁciency of the photocatalytic reaction may be inﬂuenced by a number of factors such as:

Table 1 Compounds which have undergone gas-phase photocatalytic destruction Compound

Reference

Trichloroethylene Toluene Formaldehyde Acetone Benzene Acetaldehyde 1,3-Butadiene Tetrachloroethylene Ethanol

201–207 208–213 214–216 216–219 220–222 216, 223, 225 215 225, 226 218, 223

Photocatalytic Detoxiﬁcation of Water and Air

405

• Effect of water vapour; • Light intensity—research has shown a variety of artiﬁcial UV sources have been used but to date there is little or no research done using solar power for gas-phase photocatalysis; • Initial inlet concentration of pollutant; • Temperature; • Radiant energy ﬂux—has been adopted by IUPAC as equivalent to radiant power; • Quantum yield—is deﬁned as the number of molecules of a given product formed per photon of light of a given wavelength that is absorbed by the photocatalyst; • Catalyst modiﬁcations—use of supports or doped catalysts e.g.; • Effect of oxygen. Gas phase photocatalytic oxidation using TiO2 was ﬁrst investigated by Dibble and Raupp [200]. They reported high levels of destruction of trichloroethylene (TCE) when TiO2 was irradiated with ultraviolet light. TCE is a priority substance under EU Regulation 793/93 and consequently the photodegradation of TCE in the gas phase has been extensively investigated [200–207]. TCE has also been used in some areas of research as a model pollutant to test a variety of gas phase reactor designs. High conversion levels of TCE were achieved in a bench scale ﬂat plate ﬂuidised bed photoreactor with a silica supported titania catalyst [204]. A supported catalyst was used as the titania ﬂuidisation characteristics were considered to be poor. It was found that the stoichiometric reaction 34 required the simultaneous presence of oxygen, water vapour and TCE. In order to maximise the titania threshold of 350 to 400 nm a single 4-W ﬂuorescent UV source was used. At low concentrations of TCE the oxidation rate was independent of the water concentration whereas the rate of oxidation of the TCE was inhibited by water vapour when the concentrations of the pollutant had increased. Without any water vapour the photooxidation activity of the catalyst rapidly declines. In the presence of water, however, at high TCE concentrations there was a marked deactivation with the photocatalyst. Cl2 C = CHCl + 3/2O2 + H2 O → 2CO2 + 3HCl

(33)

Dibble and Raupp [200] found that when the ﬂow rate of the TCE was reduced but the concentration was held constant the TCE conversion rose to 49%. The TCE concentration was then reduced and the conversion rate rose rapidly to ca. 96%. The decrease mirrored the fact that the observed reaction rate was controlled or limited by the feed rate of the TCE rather than by the kinetic considerations under nearly complete conversion conditions. Hager et al. [207] conﬁrmed the results of Dibble and Raupp [200] in that degradation of TCE was enhanced at lower inlet concentrations. Conversion decreased from 60.1% to 19.9% when the initial contaminant concentration

406

P.K.J. Robertson et al.

was raised to 78 g/m3 from 10 g/m3 [207]. The addition of water to the reacting mixture ﬁlled the hydroxyl ion reservoir on the photocatalyst surface and resulted in an enhancement of photocatalytic activity by the additional hydroxyl radical formation. Hager et al. [207], however, also concluded that the presence of moisture was not necessarily essential to keep up the photocatalytic process which was in contrast to the conclusions of Dibble and Raupp [200]. Alberici et al. [226] used two forms of reactor to analyse the breakdown of TCE—ﬁxed bed and ﬂuidised bed. In the presence of 20% O2 the main breakdown products identiﬁed using GC-MS were phosgene, carbon tetrachloride, dichloroacetyl chloride, dichloroacetic acid and pentachloroethane. In the absence of oxygen the products detected were pentachloroethane, 1pentachloropropene and 1,1,3,4-tetrachloro-1,3-butadiene. Kim et al. [230] concluded that the degradation rate of TCE decreased with increasing water vapour. They concluded that there was an optimum water concentration of 0.383 mol m–3 (vol%). It was also reported that molecular oxygen was an essential component because it trapped photogenerated electrons on the semiconductor surfaces and decreased the recombination of electrons and holes [230]. Wang et al. [231] proposed a pathway for the heterogeneous photocatalysis of TCE in line with their investigation (Scheme 5). Kaneko and Okura [232] found the formation of TCE decay products decreased considerably with increases in the relative humidity. It would appear that the retardation seemed to be related to consumption of Cl or Clads by water and was also due to competitive adsorption between the water vapour and TCE to the surface of TiO2 . They also concluded that in the presence of water vapour there were other active species apart from O2 – and Cl involved in the photochemical reaction, however, OH radicals may be the main contributor to TCE degradation under humid conditions. Kaneko and Okura suggested that humidity retarded decomposition of TCE and as such it was better to remove water vapour from the incoming reaction gas before it reached the photocatalytic reaction area [232]. Wang and Ku [221] proposed a possible degradation pathway for benzene by using GC-MS and FTIR spectroscopy. Using an optical ﬁbre array reactor the major intermediate detected in the initial stages of the process was phenol, suggesting hydroxyl radical attack. Total mineralisation was achieved on extended photocatalysis. The breakdown products for benzene have been quoted as phenol, hydroquinone, 1,4-benzoquinone and propanedioic acid [221]. Kaneko and Okura [232] found that the only products detected were carbon dioxide and carbon monoxide, while operating at a relative humidity level of 65%. They found that the benzene conversion decreased with decreasing water concentrations. They concluded that the presence of water retarded the formation of carbon deposits on the catalyst surface and therefore reduced the risk of catalyst deactivation due to accumulation

Photocatalytic Detoxiﬁcation of Water and Air

407

Scheme 5 Proposed pathway for TCE degradation (reproduced from [231] with permission from Elsevier Science B.V.)

of such deposits. Kaneko and Okura [232] also reported that as the benzene concentration increased the conversion rates decreased and the catalyst underwent a colour change from white to brown, which became more intense. Other researchers also observed catalyst discolouration during benzene photo-oxidation and formation of polymeric species was proposed. Fu et al. [222] concluded that the conversion of benzene was not affected by the addition of water vapour to the reactant feed stream. They concluded that once benzene conversion began to occur, sufﬁcient water would be produced to allow the reaction to continue as one of the end products of the process was water. Initiation of the reaction at the beginning of the experiment was most likely due to the presence of adsorbed water on the surface of the photocatalyst. It was also concluded that when there was oxygen in the feed stream the superoxide radical played a role in the degradation of benzene [222]. Hager et al. [207] also found that long operating times were required to approach steady state conditions when pure TiO2 was employed but when a platinised TiO2 was used the operating time was reduced to two–three hours.

408

P.K.J. Robertson et al.

As research into gaseous photocatalysis progressed a potential major disadvantage was the possibility of catalyst deactivation. Einaga et al. [208] concluded that the key factors which inﬂuenced catalyst deactivation were the formation of carbon deposits on the photocatalyst and their decomposition to COx . The photooxidation rate of benzene decreased with decreasing humidity due to the increasing amount of carbon deposits on the catalyst, however, photo irradiation in humidiﬁed air decomposed the deposits and regenerated the catalyst [208]. D’Hennezel et al. [203] proposed a possible degradation pathway for the photooxidation of gaseous toluene (Scheme 6). Their work was carried out over water and used TiO2 which had been pre-treated with HCl as it has been shown that this pre-treatment enhanced the removal rate of the toluene. Using GC-FIC, GC-MS and HPLC-UV they found that the intermediate products formed were benzoic acid, benzaldehyde and benzyl alcohol with trace derivatives of these products and of toluene that were nonhydroxylated on the ring. Marci et al. [233] concluded that the presence of O2 , a catalyst and irradiation was required for toluene degradation. They found that the intermediate breakdown products were benzyl alcohol, benzoic acid and benzaldehyde. However, the benzaldehyde was only generated in small quantities and large amounts of benzoic acid were found adsorbed onto the catalyst surface. With the two forms of TiO2 used in this study, Degussa P25 and Merck, it was observed that the P25 material deactivated even in the presence of water vapour. Luo and Ollis [204] compared the inﬂuence of water vapour on toluene with other compounds and found that the inﬂuence of water depended upon the characteristics of the contaminants. The research indicated that in a toluene-air mixture there was no toluene degradation in the absence of water, the toluene oxidation rate began to decrease when the water concentration started to reach saturation levels. Martra et al. [209] found that water vapour was needed to keep steady state toluene conversion to benzaldehyde and that in a dry toluene/air mixture an irreversible deactivation of the catalyst occurred. Their results further indicated that the produced benzaldehyde could undergo further oxidation but only in the presence of water. Kaneko and Okura [232] reported that the concentration of CO2 increased linearly with increases in the relative humidity and that the yield at 60% relative humidity was greater by one order than under dry conditions. The yield of benzaldehyde, however, decreased sharply with increased relative humidity and it was proposed that an increase in hydroxyl radicals may compete and/or hinder adsorption of toluene on the surface of TiO2 hence resulting in retardation of toluene oxidation. Kaneko and Okura [232], however, concluded that the effect of water vapour on the photocatalytic oxidation of toluene may depend on the initial pollutants concentration and its adsorptivity. Pengyi et al. [234] observed that the effect of water vapour was two sided in that a little humidity can improve the decomposition of toluene whilst too much can suppress the decomposition. This was explained by the fact that hu-

Photocatalytic Detoxiﬁcation of Water and Air

409

Scheme 6 a Primary toluene photocatalytic oxidation pathways b secondary toluene photocatalytic oxidation pathways (reproduced from [203] with permission from Elsevier Science B.V.)

410

P.K.J. Robertson et al.

Table 2 Examples of commercially available semiconductor photocatalytic systems for puriﬁcation of air Company

Country

Patents

Ref.

Toyoda Gosei Co. Ltd. Nippon Muki Co. Ltd. Trojan Technologies Incorporated Puriﬁcs Environmental Technologies Inc. Kawasaki Heavy Industries Ltd.

Japan Japan Canada Canada

236 237 238 239

Mitsubishi Materials Co. Certech

Japan Belgium

JP9038190 JP9057112 US6179972, US6179971 US6136203, US5589078, US5462674 JP2918112, JP3122082, JP3055684 JP2988376 Patent pending

Japan

240 241 242

midity could enhance the generation of hydroxyl radicals to decompose more of the contaminant while simultaneously impeding the contact of the toluene with the photocatalyst. Many have found that during the photocatalytic destruction of toluene the catalyst was susceptible to deactivation. This may be due to the accumulation of benzoic acid on the catalyst surface. Studies have also shown that the addition of TCE or ozone can either have a positive or negative effect on toluene degradation [235]. Larson and Falconer [235] observed that toluene reacts quickly on TiO2 to form strongly bound intermediates which react more slowly to form CO2 and water. As has been shown there are a number of pollutants that have undergone gaseous photocatalytic oxidation and the research within this ﬁeld is continually progressing. Unlike water photocatalytic oxidation which has found application on both a scientiﬁc and industrial basis gaseous PCO development has currently been limited to terrestrial indoor air puriﬁcation (Table 2). The vast majority of industrial development and application is conﬁned to Japan and the USA.

8 Solar Reactors and Pilot Plants 8.1 Reactor Types The artiﬁcial generation of photons required for the detoxiﬁcation of polluted water is the most important source of costs during the operating of photocatalytic waste water treatment plants. This would suggest that use of sunlight

Photocatalytic Detoxiﬁcation of Water and Air

411

would be an economically and ecologically sensible light source. With a typical UV-ﬂux near the surface of the earth of 20 to 30 Wm–2 the sun puts 0.2 to 0.3 mol photons m–2 h–1 in the 300 to 400 nm range at the disposal of the process [243]. Principally these photons are suitable for destroying water pollutants in photocatalytic reactors. Over the last 15 years several reactors for solar photocatalytic water treatment have been developed and tested. In the following, the currently most frequently used reactor concepts will be presented. 8.1.1 Thin-Film-Fixed-Bed Reactor (TFFBR) One of the ﬁrst solar reactors not applying a light-concentrating system and thus being able to utilise the diffuse as well as the direct portion of the solar UV-A irradiation for the photocatalytic process was the thin-ﬁlm-ﬁxedbed reactor (TFFBR) [244–246]. It should be noted that under AM (air mass) 1.5 conditions the diffuse (Edif (300–400 nm) = 24.3 Wm–2 ) and direct (Edir (300–400 nm) = 25.0 Wm–2 ) portion of the solar radiation reaching the surface of the earth are almost equal [247]. This means that a light concentrating system, e.g., a parabolic trough reactor can in principal only employ half of the solar radiation available in this particular spectral region. 8.1.2 Compound Parabolic Collecting Reactor (CPCR) A compound parabolic collecting reactor (CPCR) is a trough reactor without light concentrating properties. It differs from a conventional parabolic trough reactor by the shape of its reﬂecting mirrors. A reﬂector of a parabolic trough reactor has a parabolic proﬁle with the reaction pipe in its focal line. Consequently only parallel light entering the parabolic trough can be focused into the reaction pipe and a sun-tracking system is required. The shape of a CPCR’s reﬂector usually consists of two half circular proﬁles side by side, and a parabolic continuation at both outer sides of the circles. The focal line is located closely above the connection of the two circles. This geometry enables light entering from almost any direction to be reﬂected into the focal line of the CPCR, i.e., most of the diffuse light entering the module can also be employed for the photocatalytic reaction. Due to this geometry a CPCR exhibits only a small concentration factor (< 1.2). The CPCRs at the Plataforma Solar de Almeria in Spain (PSA) manufactured by Industrial Solar Technology Corporation, Denver, Colorado (USA), have a concentration factor of 1.15, i.e., this type of reactor has practically no light concentrating properties [248, 249]. Moreover, a CPCR must not necessarily track the sun due to its geometry. The azimuth should be adjusted to the complementary angle a of

412

P.K.J. Robertson et al.

the geographical altitude and the pipes should be aligned south and from top to bottom. 8.1.3 Double Skin Sheet Reactor (DSSR) Another nonconcentrating reactor is the double skin sheet reactor (DSSR). This consists of a ﬂat and transparent structured box made of PLEXIGLAS [250, 251]. The inner structure of the reactor is schematically drawn in Fig. 9. The suspension containing the waste water and the photocatalyst is pumped through these channels. The comparison of the spectral irradiance of the sun (AM 1.5) with the transmission spectrum of the PLEXIGLAS used to manufacture the double skin sheets evidences that the UV-A portion of the solar spectrum below 400 nm nicely matches with the onset of the PLEXIGLAS-transmission [251]. This type of reactor can utilise both the direct and the diffuse portion of the solar radiation in analogy to the CPCR. After the degradation process the photocatalyst has to be removed from the suspension either by ﬁltering or by sedimentation for both reactors. 8.2 Solar Photocatalytic Treatment of Real Waste Water Despite its obvious potential for the detoxiﬁcation of polluted water, there has been very little commercial or industrial use of photocatalysis as a technology so far. According to a review by Goswami [252] the published literature shows only two engineering scale demonstrations for groundwater treatment in the

Fig. 9 Schematic view of a DSSR reactor showing the inner structure of the transparent structured box made of PLEXIGLAS (reproduced from [267] with permission from Elsevier BV)

Photocatalytic Detoxiﬁcation of Water and Air

413

U.S. and one industrial waste water treatment in Spain. Engineering scale ﬁeld experiments have been conducted by the National Renewable Energy Laboratory (NREL) at the Lawrence Livermore National Laboratory (LLNL) treating ground water contaminated with trichloroethylene (TCE) [253]. This ﬁeld system consisted of 158 m2 of parabolic trough reactors and used Degussa P25 particles (0.1%) as the photocatalyst in a slurry ﬂow conﬁguration. With this relatively low titanium dioxide content the TCE concentration was reduced from 200 ppb to less than 5 ppb. Engineering scale demonstrations of the nonconcentrating solar reactor technology were conducted at Tyndall Air Force Base, Florida (USA), in 1992 [254–257] treating groundwater contaminated with fuel, oil, and lubricants which was leaking from underground storage tanks. Crittenden and co-workers used a tubular photoreactor equipped with a reﬂector and packed with platinised titanium dioxide supported on silica gel. This system was used in a single-pass mode for the photocatalytic degradation of benzene, toluene, ethylbenzene, and xylene (BTEX). Even on a rainy day, BTEX compounds (> 2 mg L–1 inﬂuent concentration) were destroyed within 6.5 min of operation [255]. Goswami and co-workers conducted ﬁeld tests with a slurry tubular reactor without a reﬂector (SOLTOX-reactor). Five modules connected in series with a total aperture area of 10 m2 were used. The photoreactor array was set up facing south and tilted at an angle of 15◦ from the horizontal. The experiments were performed in a batch mode by recycling the contaminated groundwater between a tank and the reactor array [256, 257]. While these ﬁeld tests successfully demonstrated the feasibility of destroying benzene, toluene, ethylbenzene, and xylenes (BTEX) in the groundwater, the observed reaction rates were rather low. However, in a laboratory test with city water spiked with the same amounts of BTEX as those present in the groundwater and using the same reactors under similar sunlight conditions, Goswami et al. observed reaction rates which were an order of magnitude higher [257]. These results suggested that a careful site treatability study and the establishment of an appropriate pre-treatment will be very important prerequisites for the successful ﬁeld deployment of solar photocatalytic processes [258]. A very similar pilot plant has been used by Goswami and co-workers for the treatment of a BTEX contaminated groundwater at Gainesville, Florida (USA) [259]. The ﬂat plate (DSSR) type photoreactor used in this study was developed from a double skinned acrylic panel. Experimental results showed that 1.9 m3 of water contaminated with 1000 ppb of BTEX could be cleaned reaching a ﬁnal concentration of 10 ppb in about 3 h provided that an average solar UV energy ﬂux density of 28 W m–2 was employed [259]. Dillert et al. have treated biologically pre-treated industrial waste waters from factories in Wolfsburg (Germany) and Taubaté (Brazil) of the Volkswagen AG in laboratory and bench-scale experiments [260, 261]. The results of the experiments, which have been performed using the DSSR, were so

414

P.K.J. Robertson et al.

promising, that a pilot plant was installed in the Wolfsburg factory during the summer of 1998 [262]. Figure 10 shows a photograph of this plant. As a consequence of this oxidation process as well as due to the increase of the reaction temperature, the concentration of molecular oxygen in the reﬂux pipe between the photoreactors and the tank as well as inside the tank decreased. More than 50% of the organic pollutants initially present in the mixed water inside the pilot plant could thus be degraded within 8 to 11 h of illumination. Naturally, the total mass of the degraded contaminants was found to depend on the initial pollutant concentration, the time of illumination, and, in particular, on the solar UV energy ﬂux density. In 1997 Freudenhammer et al. [263] reported their results from a pilot study using TFFBR reactors which was performed in various Mediterranean countries and showed that biologically pretreated textile waste water can be cleaned by solar photocatalysis with a maximum degradation rate of 3 g COD h–1 m–2 . Moreover, a comparison of reaction rates measured under artiﬁcial and under solar illumination, respectively, which was also performed during this study proved the necessity of outdoor experiments. It was concluded that photocatalysis should be a suitable technology as the ﬁnal stage of puriﬁcation of biologically or physically pre-treated waste water in particular in sun-rich areas. However, it was also pointed out that industrial applications of solar waste water treatment on a larger scale is in strong demand for cheap

Fig. 10 Photograph of the DSSR pilot plant installed in the Wolfsburg factory of the Volkswagen AG (Photo: J. Lohmann, ISFH) (reproduced from [267] with permission from Elsevier BV)

Photocatalytic Detoxiﬁcation of Water and Air

415

Fig. 11 Photograph of the TFFBR pilot plant installed at the site of a textile factory in Tunisia (Photo: D. Hufschmidt, University of Hannover) (reproduced from [267] with permission from Elsevier BV)

photocatalysts of higher activity to be competitive with treatment methods already established on the market [263]. Based on these results, a pilot plant (shown in Fig. 11) has been built at the site of a textile factory in Tunisia (Menzel Temime). The TFFBR was chosen because previous studies showed sufﬁcient degradation rates with the selected textile waste water in combination with its simple, low cost construction and the low energy consumption [263]. However, to integrate results obtained with suspended catalysts showing in some cases a higher efﬁciency than the ﬁxed system, the possibility of using suspended catalysts has also been considered. The plant was designed to operate the two reactors independently. This allowed a comparison of different catalysts, ﬁxation techniques, hydrodynamics and efﬂuents/model compounds under identical conditions. The plant could be operated with suspended and ﬁxed catalysts. A sedimentation tank was connected to the reactors (via a storage tank) to separate and recycle the catalyst. The experiments could be carried out in continuous or recycling mode [264].

416

P.K.J. Robertson et al.

9 Conclusion As can be seen from the above discussion semiconductor photocatalysis is a highly versatile technique which has proven to be very effective for the treatment of a wide range of chemical and biological species in water and air. The process has been applied on an industrial scale and promising results on large scale water and air treatment have been achieved [236–267]. One of the main short comings of the technique is the fact that the most effective photocatalyst, TiO2 , is a UV absorber. Although the use of sunlight for activation of this material has been achieved in large scale applications [246] it would be highly desirable if effective and affordable visible absorbing materials were developed. Although there have been signiﬁcant advances in research in such visible absorbing titania, this is still only very much on a lab development scale and no large scale manufacture has been undertaken [268–271]. In some cases, however, TiO2 photocatalysis may offer the only viable process for effective removal of hazardous compounds, for example in the case of cyanotoxins. Photocatalytic oxidation has started to make an important impact in environmental remediation of both contaminated water and air and commercial applications of this versatile environmental treatment technology will multiply over the next decades. Acknowledgements Sections two to ﬁve have been adopted with permission from the chapter by Detlef W. Bahnemann, Ralf Dillert, and Peter K.J. Robertson entitled Photocatalysis: Electron Transfer Serving the Environment , produced in Chemical Physics of Nanostructured Semiconductors edited by A.I. Kokorin and D.W. Bahnemann, Brill Academic Publishers, Inc., Herndon, VA, USA, 2003. This was reproduced with kind permission from Brill Academic Publishers, Inc., Herndon, VA, USA and the authors gratefully acknowledge the publishers for this consent. The authors wish to thank Dr. Cathy McCullagh and Mr. Morgan Adams of The Robert Gordon University Aberdeen for the preparation of diagrams and proof reading of this manuscript.

References 1. Kamat PV (1993) Chem Rev 93:671 2. Mills A, Davies RH, Worsley D (1993) Chem Soc Rev 417 3. Ollis DF, Al-Ekabi H (1993) Photocatalytic puriﬁcation and treatment of water and air, vol. 3. Elsevier, Amsterdam 4. Pichat P (1994) Catal Today 19:313 5. Hoffmann MR, Martin ST, Choi W, Bahnemann DW (1995) Chem Rev 95:69 6. Mills A, Le Hunte S (1997) J Photochem Photobiol A Chem 108:1 7. Sauer ML, Ollis DF (1996) J Catal 158:570 8. Alberici RM, Jardim WF (1997) App Catal B Environ 14:55 9. Pichat P, Disdier J, Hoang-Van C, Mas D, Goutailler G, Gaysse C (2000) Catal Today 63:636

Photocatalytic Detoxiﬁcation of Water and Air

417

10. Grätzel M, Frank AJ (1982) J Phys Chem 86:2964 11. Wolff K (1993) Mechanistische Untersuchungen zum Oxidationsprozess an der belichteten Titandioxid/Wasser Grenzﬂäche. PhD Thesis, University of Hannover, Hannover, Germany 12. Degussa AG (1989) Hochdisperse Metalloxide nach dem Aerosil-Verfahren [Schriftenreihe Pigmente], 4th ed, Frankfurt 13. Bahnemann D, Henglein A, Lilie J, Spanhel L (1984) J Phys Chem 88:4278 14. Bahnemann D, Henglein A, Spanhel L (1984) Faraday Discuss Chem Soc 78:151 15. Rothenberger G, Moser J, Grätzel M, Serpone N, Sharma DK (1985) J Am Chem Soc 107:8054 16. Serpone N, Lawless D, Khairutdinov R, Pelizzetti E (1995) J Phys Chem 99:16655 17. Colombo DP, Bowman RM (1995) J Phys Chem 99:11752 18. Howe RF, Grätzel M (1987) J Phys Chem 91:3906 19. Moser J, Punchihewa S, Infelta PP, Grätzel M (1991) Langmuir 7:3012 20. Schindler K-M, Kunst M (1990) J Phys Chem 94:8222 21. Warman JM, de Haas MP, Pichat P, Serpone N (1991) J Phys Chem 95:8858 22. Howe RF, Grätzel M (1985) J Phys Chem 89:4495 23. Martin ST, Herrmann H, Choi W, Hoffmann MR (1994) Trans Faraday Soc 90:3315 24. Martin ST, Herrmann H, Hoffmann MR (1994) Trans Faraday Soc 90:3323 25. Hilgendorff M (1996) Untersuchungen zur Bedeutung der Adsorption in der Photokatalyse. PhD Thesis, University of Hannover, Hannover, Germany 26. Bahnemann DW, Hilgendorff M, Memming R (1997) J Phys Chem B 101:4265 27. Gonzalez-Elipe AR, Munuera G, Soria J (1979) J Chem Soc Faraday Trans I 75:748 28. Jaeger CD, Bard AJ (1979) J Phys Chem 83:3146 29. Anpo M, Shima T, Kubokawa Y (1985) Chem Lett 1799 30. Kasinski JJ, Gomez-Jahn LA, Faran KJ, Gracewski SM, Miller RDJ (1989) J Phys Chem 90:1253 31. Lawless D, Serpone N, Meisel D (1991) J Phys Chem 95:5166; Serpone N, Lawless D, Terzian R, Meisel D (1992) In: RA Mackay, J Texter (eds) Electrochemistry in Colloids and Dispersions. VCH, New York, Weinheim, 399 32. Micic OI, Zhang Y, Cromack KR, Trifunac AD, Thurnauer MC (1993) J Phys Chem 97:7277 33. Colombo DP, Roussel KA, Saeh J, Skinner DE, Cavaleri JJ, Bowman RM (1995) Chem Phys Lett 232:207 34. Bickley RI, Jayanty RKM (1974) Faraday Discuss Chem Soc 58:194 35. Tafalla D, Salvador P (1987) Ber Bunsenges Phys Chem 91:475 36. Brezová V, Stasko A, Lapcik L (1991) J Photochem Photobiol A Chem 59:115 37. Peterson MW, Turner JA, Nozik AJ (1991) J Phys Chem 95:221 38. Buxton GV, Greenstock CL, Helman WP, Ross AB (1988) J Phys Chem Ref Data 17:513 39. Richard C (1994) New J Chem 18:443 40. Choi W, Hoffmann MR (1995) Environ Sci Technol 29:1646 41. Hilgendorff M, Hilgendorff M, Bahnemann DW (1993) In: M Tomkiewicz, R Haynes, H Yoneyama, Y Hori (eds) Environmental Aspects of Electrochemistry and Photoelectrochemistry. The Electrochemical Society, Pennington, 112 42. Hilgendorff M, Hilgendorff M, Bahnemann DW (1996) J Adv Oxid Technol 1:35 43. Peyton GR, Bell OJ, Girin E, Lefaivre MH (1995) Environ Sci Technol 29:1710 44. Nahen M, Bahnemann DW, Dillert R, Fels G (1997) J Photochem Photobiol A Chem 110:191 45. Ward MD, Bard AJ (1982) J Phys Chem 86:3599

418

P.K.J. Robertson et al.

46. 47. 48. 49. 50. 51.

Sclafani A, Palmisano L, Davi E (1991) J Photochem Photobiol A Chem 56:113 Prairie MR, Evans LR, Stange BM, Martinez SL (1993) Environ Sci Technol 27:1776 Grabner G, Li G, Quint RM, Getoff N (1991) J Chem Soc, Faraday Trans 87:1097 Colombo DP, Bowman RM (1996) J Phys Chem 100:18445 Bahnemann DW (1993) Isr J Chem 33:115 Wolff K, Bockelmann D, Bahnemann DW (1991) Proc IS&T Symp on Electronic and Ionic Properties of Silver Halides 44th IS&T Annual Conf, St Paul, Minnesota, May 12–17, 1991 (ed: Levy B), Springﬁeld, Virginia, IS&T, 259 Neta P, Simic M, Hayon E (1969) J Phys Chem 73:4207 Schuchmann MN, Zegota H, von Sonntag C (1985) Z Naturforsch B 40:215 Schuchmann H-P, von Sonntag C (1984) Z Naturforsch B 39:217 Augugliaro V, Palmisano L, Sclafani A, Minero C, Pelizzetti E (1988) Toxicol Environ Chem 16:89 Turchi CS, Ollis DF (1990) J Catal 122:178 Terzian R, Serpone N, Draper RB, Fox MA, Pelizzetti E (1991) Langmuir 7:3081 Mills G, Hoffmann MR (1993) Environ Sci Technol 27:1681 Theurich J, Lindner M, Bahnemann DW (1996) Langmuir 12:6368 Theurich J, Bahnemann DW, Vogel R, Ehamed FE, Alhakimi G, Rajab I (1997) Res Chem Intermed 23:247 Theurich J (1999) Kinetische und mechanistische Untersuchungen zum photochemischen Abbau organischer Schadstoffe in wässriger Phase. PhD Thesis, University of Hannover, Hannover, Germany Jaeger CD, Bard AJ (1979) J Phys Chem 83:3146 Anpo M, Shima T, Kubokawa Y (1985) Chem Lett 1799 Kasinski JJ, Gomez-Jahn LA, Faran KJ, Gracewski SM, Miller RDJ (1989) J Phys Chem 90:1253 Gonzalez-Elipe AR, Munuera G, Soria J (1979) J Chem Soc Faraday Trans I 75:748 Mao Y, Schoneich C, Asmus KD (1991) J Phys Chem 95:80 Cunningham J, Srijaranai S (1988) J Photochem, Photobio A Chem 43:329 Robertson PKJ, Lawton LA, Cornish BJPA, Jaspars M (1998) J Photochem Photobiol A Chem 116:215 Okamoto K, Yamamoto Y, Tanaka H, Itaya A (1985) Bull Chem Soc Jpn 58:2023 Anpo M, Chiba K, Tominari M, Coluccia S, Chen M, Fox MA (1991) Bull Chem Soc Jpn, 64:543 Lu G, Linsebigler A, Yates YT (1995) J Phys Chem 99:7626 Richard C (1993) J Photochem Photobiol A Chem 72:179 Draper RB, Fox MA (1990) Langmuir 6:1396 Wang C-Y, Böttcher C, Bahnemann DW, Dohrmann JK (2003) J Mater Chem 13:2322 Wang C-Y, Böttcher C, Bahnemann DW, Dohrmann JK (2004) J Nanoparticle Res 6:119 Bahnemann DW, Henglein A, Lilie J, Spanhel L (1984) J Phys Chem 88:709 Nosaka Y, Kishimoto M, Nishino J (1998) J Phys Chem B 102:10279 Vorontsov AV, Stoyanova IV, Kozlov DV, Simagina VI, Savinov EN (2000) J Catal 189:360 Kim S, Choi W (2002) J Phys Chem B 106:13311 Wang C-Y, Liu CP, Zheng X, Chen J, Shen T (1998) Colloid Surf A Physicochem Eng Asp 131:271 Wang C-Y, Liu CP, Chen J, Shen T (1997) J Colloid Interface Sci 191:464 Wang C-Y, Pagel R, Bahnemann DW, Dohrmann JK (2004) J Phys Chem B 108:14082 Wang C-Y, Liu CY, Shen T (1997) J Photochem Photobiol A Chem 109:65

52. 53. 54. 55. 56. 57. 58. 59. 60. 61.

62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83.

Photocatalytic Detoxiﬁcation of Water and Air 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119.

120. 121. 122. 123. 124. 125.

419

Brinkley D, Engel T (1998) J Phys Chem B 102:7596 Duonghong D, Ramsden J, Grätzel M (1982) J Am Chem Soc 104:2977 Sczechowski JG, Koval CA, Noble RD (1993) J Photochem Photobiol A Chem 74:273 Upadhya S, Ollis DF (1997) J Phys Chem B 101:2625 Buechler KJ, Nam CH, Zawistowski TM, Noble RD, Koval CA (1999) Ind Eng Chem Res 38:1258 Buechler KJ, Noble RD, Koval CA, Jacoby WA (1999) Ind Eng Chem Res 38:892 Buechler KJ, Zawistowski TM, Noble RD, Koval CA (2001) Ind Eng Chem Res 40:1097 Cornu CJG, Colussi AJ, Hoffmann MR (2001) J Phys Chem B 105:1351 Pagel R (2004) PhD Thesis, Institut fuer Physikalische Chemie, Freie Universität Berlin, Germany, in preparation Vorkapic D, Matsoukas T (1999) J Colloid Interface Sci 214:283 Carey JH, Lawrence J, Tosine HM (1976) Bull Environ Contam Toxicol 16:697 Fox MA, Chen CC (1981) J Am Chem Soc 103:6757 Fox MA, Chen MJ (1983) J Am Chem Soc 105:4497 Fox MA, Younathan JN, Sackett D (1987) Tetrahedron 43:1643 Fox MA, Kim Y-S, Dulay MT (1990) Catal Lett 5:369 Dulay MT, Washington-Dedeaux D, Fox MA (1991) J Photochem Photobiol 61:153 Hsiao C-Y, Lee C-L, Ollis DF (1983) J Catal 82:418 Turchi, CS, Ollis DF (1989) J Catal 119:483 Mills A, Sawunyama P (1994) J Photochem Photobiol A Chem 84:305 Mills A, Morris S (1993) J Photochem Photobiol A Chem 71:285 Mills A, Morris S, Davies R (1993) J Photochem Photobiol A Chem 70:183 Wang Y, Hong C-S (1999) Wat Res 33:2031 Chiarenzelli J, Scrudato R, Wunderlich M, Rafferty D, Jensen K, Oenga K, Roberts R, Pagano J (1995) Chemosphere 31:3259 Kormann C, Bahnemann DW, Hoffmann MR (1991) Environ Sci Tech 25:494 Choi W, Hoffmann MR (1995) Environ Sci Tech 29:1646 Crittenden JC, Liu J, Hand DW, Perram DL (1997) Wat Res 31:429 Chen J, Ollis DF, Rulkens WH, Bruning H (1999) Wat Res 33:661 Kawaguchi H (1984) Environ Tech Lett 5:471 Augugliaro V, Palmisano L, Sclafani A (1988) Toxicol Environ Chem 16:89 Augugliaro V, Davi E, Palmisano L, Schiavello M, Sclafani A (1990) J Appl Catal 65:101 Matthews RW, McEvoy SR (1992) J Photochem Photobiol A Chem 64:231 Matthews RW, McEvoy SR (1992) Solar Energy 49:507 Terzian R, Serpone N, Minero C, Pelizzetti E (1991) J Catal 128:352 Dieckmann MS, Gray KA (1996) Water Res 30(5):1169–1163 San N, Hatipoglu A, Koçtürk G, Çinar Z (2002) J Photochem Photobiol A Chem 146:189 Pichat P, Guillard C, Maillard C, Amalric L, D’Oliveira J-C (1993) In: Ollis DF, AlEkabi H (eds) Puriﬁcation and Treatment of Water and Air. Elsevier, Amsterdam, pp 207–223 Amalric L, Guillard C, Blanc-Brude E, Pichat P (1996) Wat Res 30:1137 Ireland JC, Dávila B, Moreno H (1995) Chemosphere 30:965 Dass S, Muneer M, Gopidas KR (1994) J Photochem Photobiol A Chem 77:83 Hidaka H, Nohara K, Zhao J, Serpone N (1992) J Photochem Photobiol A Chem 64:247 Muszkat L, Bir L, Feigelso L (1995) J Photochem Photobiol A Chem 87:85 Malato S, Blanco J, Fernandez-Alba AR, Aguera A (2000) Chemosphere 40:403

420

P.K.J. Robertson et al.

126. Bandala ER, Gelover S, Leal MT, Arancibia-Bulnes C, Jiminez A, Estrada CA (2002) Catal Today 76:189 127. Goutailler G, Valette J-C, Guillard C, Pisse O, Faure R (2001) J Photochem Photobiol A Chem 141:79 128. Hequet V, Gonzalez C, Le Cloirec P (2001) Wat Res 35:4253 129. Muneer M, Qamar M, Saquib M, Bahnemann DW (2004) Res Chem Intermed 30:663 130. Konstantinou IK, Sakkas VA, Albanis TA (2001) Appl Catal B Environ 34:227 131. Zertal A, Molnár-Gábor D, Malouki MA, Sehili T, Boule P (2004) Appl Catl B Environ 49:83 132. Khodla AA, Sehili T, Pilichowski J-F, Boule P (2001) J Photochem Photobiol A Chem 141:231 133. Mathews RW (1989) J Chem Soc Faraday Trans 85:1291 134. Lakshimi S, Renganathan R, Fujita S (1995) J Photochem Photobiol A Chem 88:163 135. Mills A, Wang J (1999) J Photochem Photobiol A Chem 127:123 136. Houas A, Lachheb H, Ksibi M, Elaloui E, Guillard C, Herrmann JM (2001) App Catal B Environ 31:145 137. Lachheb H, Puzenat E, Houas A, Ksibi M, Elaloui E, Guillard C, Herrmann JM (2002) App Catal B Environ 39:75 138. Daneshvar N, Salari D, Khataee AR (2003) J Photochem Photobiol A Chem 157:111 139. Sapquib M, Muneer M (2003) Desalination 155:255 140. Stylidi M, Konarides DI, Verykios XE (2004) App Catal B Environ 47:189 141. Sapquib M, Muneer M (2003) Dyes Pigments 56:37 142. Galdindo C, Jacques P, Kalt A (2001) J Photochem Photobiol A Chem 141:47 143. Vautier M, Guillard C, Herrmann JM (2001) J Catal 201:46 144. Rao KVS, Lavédrine B, Boule P (2003) J Photochem Photobiol A Chem 154:189 145. Schmelling DC, Gray KA (1995) Wat Res 29:2651 146. Dillert R, Brandt M, Fornefett I, Siebers U, Bahnemann D (1995) Chemosphere 12:2333 147. Wang Z, Kutal C (1995) Chemosphere 30:1125 148. Dillert R, Fornefett I, Siebers U, Bahnemann D (1995) J Photochem Photobiol A Chem 94:231 149. Schmelling DC, Gray KA, Kamat PV (1997) Wat Res 31:1439 150. Berry RJ, Mueller MR (1994) Microchem J 50:28 151. Grzechulska J, Hamerski M, Morawski AW (2000) Wat Res 34:1638 152. Ziolli RL, Jardim WF (2003) J Photochem Photobiol A Chem 155:243 153. Bessa E, Sant’Anna GL, Dezotti M (2001) App Catal B Environ 29:125 154. Campbell I, Robertson PKJ (2005) Chem Eng Sci, in preparation 155. Eggins BR, Palmer FL, Byrne JA (1997) Wat Res 31:1223 156. Bekbolet M, Suphanndag A, Uyguner CS (2002) J Photochem Photobiol A Chem 148:121 157. Wiszniowski J, Robert D, Surmacz-Gorska J, Milksch K, Weber J-V (2003) Int J Photoenergy 5:69 158. Carmichael WW (1995) In: Hallegraef GM, Anderson GM, Cembella AD (eds) Manual on Harmful Marine Microalgar. United Nations Educational, Scientiﬁc and Cultural Organization, Paris 159. Sivonen K (1996) Phycologia 35:12 160. Codd GA, Bell SG, Broks WP (1989) Water Sci Tech 21:1 161. Dunn J (1996) Brit Med J 312(7040):1183–1184 162. Yu SZ (1994) Toxic Cyanobacteria: Current Status of Research and Management. Australia Centre for Water Quality Research, Salisbury, Australia

Photocatalytic Detoxiﬁcation of Water and Air

421

163. Robertson PKJ, Lawton LA, Münch B, Rouzade J (1997) Chem Comm 4:393 164. Lawton LA, Robertson PKJ, Cornish BJPA, Jaspars M (1999) Environ Sci Tech 33:771 165. Liu I, Lawton LA, Cornish BJPA, Robertson PKJ (2002) J Photochem Photobiol A Chem 148:349 166. Liu I, Lawton LA, Cornish BJPA, Robertson PKJ (2002) Environ Sci Tech 37:3214 167. Shepard GS, Stockenstrom S, De Villiers D, Engelbrecht WJ, Sydenham EW, Wessels GFS (1998) Toxicon 36:1895 168. Lawton LA, Robertson PKJ, Cornish BJPA, Marr IL, Jaspars M (2003) J Catal 213:109 169. Coleman HM, Eggins BR, Byrne JA, Plamer FL, King E (2000) Appl Catal B Environ 24:L1 170. Matsunaga T, Tomoda R, Nakajima T, Wake H (1985) FEMS Microbiol Lett 29:211 171. Sunada K, Watanabe T, Hashimoto K (2003) J Photochem Photobiol A Chem 156:227 172. Bekbölet M (1997) Water Sci Tech 35:95 173. Li XZ, Zhang M, Chua H (1996) Water Sci Tech 33:111 174. Huang N, Xiao Z, Huang D, Yuan C (1998) Supramolec Sci 5:559 175. Rincón GG, Pulgarin C (2003) Appl Catal B Environ 44:263 176. Ireland JC, Klostermann P, Rice E, Clark R (1993) Appl Environ Microbiol 1668 177. Ibáñez JA, Litter MI, Pizarro RA (2003) J Photochem Photobio A Chem 157:81 178. Amézaga-Madrid P, Nevárez-Moorillón GV, Orrantia-Borunda E, Miki-Yoshida M (2002) FEMS Microbiol Lett 211:183 179. Herrera Melián JA, Doña Rodríguez JM, Viera Suárez A, Tello Rendón E, Valdés do Campo C, Arana J, Pérez Peña J (2000) Chemosphere 41:323 180. Saito T, Iwase T, Horie J, Morioka T (1992) J Photochem Photobiol B Biol 14:369 181. Kim B, Kim D, Cho D, Cho S (2003) Chemosphere 52:277 182. Watts RJ, Kong S, Orr MP, Miller GC, Henry BE (1995) Water Res 29:95 183. Cornish BJPA, Lawton LA, Robertson PKJ (2000) Appl Catal B Environ 25:59 184. Kühn KP, Chaberny IF, Massholder K, Stickler M, Benz VW, Sonntag H-G, Erdinger L (2003) Chemosphere:53:71 185. Curtis TP, Walker G, Dowling BM, Christensen PA (2002) Water Res 36(9):2410 186. Butterﬁeld IM, Christensen PA, Curtis TP, Gunlazuardi J (1997) Water Res 31:675 187. Sunada K, Kikuchi Y, Hashimoto K, Fujishima A (1998) Env Sci Tech 32:726 188. Kikuchi Y, Sunada K, Iyoda T, Hashimoto K, Fujishima A (1997) J Photochem Photobiol A Chem 106:51 189. Wist J, Sanabria J, Dierolf C, Torres W, Pulgarin C (2002) J Photochem Photobiol A Chem 147:241 190. Wei C, Lin W, Zainal Z, Williams NE, Zhu K, Kruzic AP, Smith RL, Rajeshwar K (1994) Env Sci Tech 28:934 191. Dunlop PSM, Byrne JA, Manga N, Eggins BR (2002) J Photochem Photobiol A Chem 148:355 192. Rincón A-G, Pulgarin C (2004) Appl Catal B Environ 49:99 193. Fernandez RO, Pizarro RA (1996) Photochem Photobiol 64:334 194. Oppezzo OJ, Pizzarro RA (2001) Photochem Photobiol 62:158 195. Maness PC, Smolinski S, Blake DM, Huang Z, Wolfrum EJ, Jacoby WA (1999) Appl Environ Microbiol 65(9):4094 196. Huang Z, Maness P-C, Blake DM, Wolfrum EJ, Smolinski SL, Jacoby WA (2000) J Photochem Photobiol A Chem 130:163 197. Dillert R, Siemon U, Bahnemann D (1999) J Adv Oxid Technol 4:55 198. Anpo M (1997) Catal Surv Jpn 1:169 199. Mills A, Hunte SL (1997) J Photochem Photobiol A Chem 108:1 200. Dibble LA, Raupp GB (1992) Environ Sci Tech 26:492

422

P.K.J. Robertson et al.

201. Peral J, Domenech X, Ollis DF (1997) J Chem Technol Biotechnol 70:117 202. Jacoby WA, Blake DM, Fennell JA, Boulter JE, Vargo LM, George MC, Dolbergy SK (1996) J Air Waste Manage Assoc 46:891 203. D’Hennezel O, Pichat P, Ollis DF (1998) J Photochem Photobiol A Chem 118:197 204. Luo Y, Ollis DF (1996) J Catal 163:1 205. Beuchler K, Nobel RD, Koval CA, Jacoby WA (1999) Ind Eng Chem Res 38:1258 206. Kim JS, Itoh Murabayashi KM (1998) Chemosphere 36:483 207. Hager S, Bauer R (1999) Chemosphere 38:1549 208. Einaga H, Futamura S, Ibusuki T (2002) Appl Catal B Environ 38:215 209. Martra G, Coluccia S, Marchese L, Augugliaro V, Loddo V, Palmisano L, Schiavello M (1999) Catal Today 53:695 210. Blount MC, Falconer JL (2002) Appl Catal B Environ 39:39 211. Blanco J, Avila P, Bahamonde A, Alvarez E, Sanchez B, Romero M (1996) Catal Today 29:437 212. Augugliaro V, Coluccia S, Loddo V, Marchese L, Martra G, Palmisano L, Schiavello M (1999) Appl Catal B Environ 20:15 213. Mendez-Roman R, Cardona Marinez N (1998) Catal Today 40:353 214. Obee T (1996) Environ Sci Tech 30:3578 215. Obee TN, Brown RT (1995) Environ Sci Tech 29:1223 216. Stevens L, Lanning JA, Anderson LG, Jacoby WA, Chornet N (1998) J Air Waste Manage Assoc 48:979 217. Coronado JM, Zorn ME, Tejedor-Tejedor I, Anderson MA (2003) Appl Catal B Environ 43:329 218. Vorontsov AV, Savinov ENN, Barannik GB, Troitsky VN, Parmon VN (1997) Catal Today 39:207 219. Zorn ME, Tompkins DT, Zeltner WA, Anderson MA (1999) Appl Catal B Environ 23:1 220. D’Hennezel O, Ollis DF (1997) J Catal 167:118 221. Wang W, Ku Y (2003) J Photochem Photobiol A Chem 159:47 222. Fu X, Zeltner WA, Anderson MA (1995) Appl Catal B Environ 6:209 223. Muggli DS, Ding L (2001) Appl Catal B Environ 32:181 224. Ibrahim H, De Lasa H (2002) Appl Catal B Environ 38:201 225. Hegedus M, Dombi Kiriesi AI (2001) Reaction Kinetic Catal Lett 74(2):209 226. Alberici R, Jardim WF (1997) Appl Catal 14:55 227. Demeestere K, De Visscher A, Dewulf J, Van Leeuwen M, Van Langenhove H (2004) Appl Catal B 54:216 228. Placidus B Amama, Itoh K, Murabayashi M (2004) J Mol Catal A 217:109 229. Lim TH, Kim SD (2004) Chemosphere 54:305 230. Kim JS, Joo HK, Lee TK, Itoh K, Murabayoshi M (2000) J Catal 194:484 231. Wang K-H, Jehng J-M, Hsieh Y-H, Chang C-Y (2002) J Hazard Mater B90:63 232. Kaneko M, Okura I (eds) (2003) Photocatalysis: Science and Technology. Springer, Berlin Heidelberg New York 233. Marci G, Addamo M, Augugliaro V, Coluccia S, Garcia-Lopez E, Loddo V, Martra G, Palmisano L, Schiavello M (2003) J Photochem Photobiol A Chem 160:105 234. Pengyi Z, Fuyan L, Gang Y, Qing C, Wanpeng Z (2003) J Photochem Photobiol A Chem 156:189 235. Larson SA, Falconer JL (1997) Catal Lett 44:57 236. http://www.toyoda-gosei.co.jp/english/topics/001108.html 237. http://www.nipponmuki.co.jp/e/index.html 238. http://www.oceta.on.ca/proﬁles/trojan/air/trojan2_5tech.html 239. http://www.puriﬁcs.com

Photocatalytic Detoxiﬁcation of Water and Air 240. 241. 242. 243. 244.

245. 246. 247. 248. 249. 250.

251. 252. 253. 254. 255. 256. 257. 258. 259. 260. 261. 262. 263. 264. 265. 266. 267. 268. 269. 270. 271.

423

http://www.khi.co.jp/ http://www.mmc.co.jp http://www.certech.be/dossier/ang/1.pdf Bahnemann DW (1994) Nachr Chem Tech Lab 42:378 Hilgendorff M, Bockelmann D, Nogueira RFP, Weichgrebe D, Jardim WF, Bahnemann DW, Goslich R (1992) Proc 6th Int Symp Solar Thermal Concentrating Technologies 2:1167 Bockelmann D, Weichgrebe D, Goslich R, Bahnemann DW (1995) Solar Energy Mater Solar Cells 38:441 Goslich R, Dillert R, Bahnemann DW (1997) Wat Sci Tech 35:137 Bird RE, Hulstrom RL, Lewis LJ (1983) Solar Energy 30:563 Blanco J, Malato S (1993) personal communication Goß A (1995) Diploma Thesis, University of Clausthal-Zellerfeld, Germany Benz V, Müller M, Bahnemann DW, Weichgrebe D, Brehm M (1996) Reaktoren für die photokatalytische Abwasserreinigung mit Stegmehrfachplatten als Solarelemente, Deutsche Offenlegungsschrift DE 195 14 372 A1 van Well M, Dillert RHG, Bahnemann DW, Benz VW, Müller MA (1997) J Solar Energy Eng 119:114 Goswami DY (1997) J Solar Energy Eng 119:101 Mehos MS, Turchi CS (1993) Environ Prog 12:194 Crittenden JC, Zhang Y, Hand DW, Perram DL, Marchand EG (1996) Water Environ Res 68:270 Zhang Y, Crittenden JC, Hand DW, Perram DL (1996) J Solar Energy Eng 118:123 Turchi CS, Klausner JF, Goswami DY, Marchand E (1994) Vol. Date 1993. Chem Oxid 3:216 Goswami DY, Klausner J, Mathur GD, Martin A, Schanze K, Wyness P, Turchi C, Marchand E (1993) Proc Am Solar Energy Soc Ann Conf Solar p 235 Goswami DY (1997) J Solar Energy Eng 119:101 Srinivasan M, Klausner JF, Jotshi CK, Goswami DY (1997) ASME Int Solar Energy Conf, Orlando, FL Dillert R, Vollmer S, Schober M, Theurich J, Bahnemann D, Arntz HJ, Pahlmann K, Wienefeld J, Schmedding T, Sager G (1999) gwf Wasser Abwasser 140:293 Dillert R, Vollmer S, Schober M, Theurich J, Bahnemann D, Arntz HJ, Pahlmann K, Wienefeld J, Schmedding T, Sager G (1999) Chem Ing Techn 71:396 Dillert R, Vollmer S, Gross E, Schober M, Bahnemann D, Wienefeld J, Pahlmann K, Schmedding T, Arntz HJ, Sager G (1999) Z Phys Chem 213:141 Freudenhammer H, Bahnemann D, Bousselmi L, Geissen SU, Ghrabi A, Saleh F, SiSalah A, Siemon U, Vogelpohl A (1997) Wat Sci Tech 35:149 Bousselmi L, Geissen SU, Schröder H (2004) Wat Sci Tech 49:331 Funken KH (1992) Nachr Chem Tech Lab 40:378 Alpert DJ, Sprung JL, Pacheco JE, Prairie MR, Reilly HE, Milne TA, Nimlos MR (1991) Solar Energy Mater 25:594 Bahnemann DW (2004) Solar Energy 77:445 Sakthivel S, Kisch H (2003) Angew Chem Int Ed 42:4908 Sakthivel S, Kisch H (2003) Chem Phys Chem 4:487 Yamashita H, Anpo M (2004) Catal Surv Asia 8:35 Zang L, Macyk W, Lange C (2000) Chem Eur J 6:379

Hdb Env Chem Vol. 2, Part M (2005): 425–450 DOI 10.1007/b138190 © Springer-Verlag Berlin Heidelberg 2005 Published online: 20 September 2005

Photocatalytic Active Surfaces and Photo-Induced High Hydrophilicity/High Hydrophobicity Hiroshi Irie · Kazuhito Hashimoto (u) 4-6-1 Komaba, Meguro-ku, 153-8904 Tokyo, Japan [email protected], [email protected] 1

Introduction and Historical Background . . . . . . . . . . . . . . . . . .

426

2

Photocatalytic Oxidation Reaction . . . . . . . . . . . . . . . . . . . . . .

427

3 3.1 3.2 3.3 3.4 3.4.1 3.4.2 3.5 3.5.1 3.5.2 3.5.3 3.6

Photoinduced High Hydrophilic Conversion . . . . . . . . . . Wettability on the Surface . . . . . . . . . . . . . . . . . . . . What is High Hydrophilicity? . . . . . . . . . . . . . . . . . . Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanism for Highly Hydrophilic Conversion . . . . . . . . Negative Experimental Results for Conversion by Oxidative Decomposition . . . . . . . . . . . . . . . . . . . Proposed Mechanism for the Highly Hydrophilic Conversion High Sensitization Under Weak UV-Light Irradiation . . . . . Heterogeneous System, TiO2 /WO3 . . . . . . . . . . . . . . . Control of Surface Nano-Structure . . . . . . . . . . . . . . . Introduction of Residual Tensile Stress . . . . . . . . . . . . . Visible Light Sensitivity . . . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

429 429 430 431 434

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

434 437 440 440 442 443 444

4 4.1 4.2

High Hydrophobicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Highly Hydrophobic Films and Surfaces . . . . . . . . . . . . . . . . . . . Applications of Highly Hydrophobic Thin Films . . . . . . . . . . . . . . .

445 445 447

5

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

448

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

448

Abstract There are two types of photo-effects on a TiO2 surface. One is the well-known photocatalytic reaction and another is the highly hydrophilic conversion under ultraviolet light irradiation. Using these functions, TiO2 has already been put to practical use in water puriﬁcation, air puriﬁcation, anti-bacteria, and so on. In addition, various materials coated with this type of TiO2 transparent ﬁlm can also show self-cleaning and antifogging effects. These industrial applications are herein described. In particular, high hydrophilicity is focused on and is attributed to the increase in the amount of OH groups. The mechanism is explained in detail in this article. The enhancement of high hydrophilicity under weak UV-light irradiation and sensitization to Vis light are also described. Moreover, the highly hydrophobic surface, the exact opposite of the highly hydrophilic surface, is also described. Keywords Titanium dioxide · Oxidation · Wettability · High hydrophilicity · High hydrophobicity

426

H. Irie · K. Hashimoto

Abbreviations UV ultra-violet Vis visible

1 Introduction and Historical Background It is well-known that the photocatalytic oxidation-reduction reaction of TiO2 irradiated with UV light can decompose organic compounds adsorbed on the surface. This photocatalytic oxidation-reduction reaction originated from the discovery of water-splitting on TiO2 semiconductor electrodes irradiated with UV light, which is known as the Honda–Fujishima effect and was reported in Nature in 1972 [1]. This effect attracted considerable attention after the ﬁrst oil crisis. Initial studies based on this effect were on systems that could utilize solar energy to produce hydrogen. In addition, following the analogy of a plants’ production of organic compounds through the reduction of carbon dioxide, studies were begun to utilize the effect to reduce CO2 to produce organic compounds [2–6]. Studies in this area are being actively pursued even today. However, scientiﬁc studies involving the Honda–Fujishima effect have been redirected towards a new line of research that has been aimed at environmental cleaning [7–14]. Although the Honda–Fujishima effect was discovered as an electrode reaction, its application to environmental cleaning was conceived as a system based on the use of TiO2 powder suspended in water. This concept may be understood if we consider the ﬁne particles of TiO2 as micro-scale electrochemical cells. The use of TiO2 powder suspended in water, however, requires an extra treatment step after water-cleaning, to remove the powder. The process would not be practical and would be difﬁcult to apply unless we had a simple reaction system. There was another difﬁculty with this research. Because the original research started with the decomposition of water by solar light, we tended to work on technologies that would process large amounts of substances, using intense light. In fact, the light that TiO2 can utilize is limited to UV. The fraction of UV light in the whole of the solar spectrum is small. In addition, the production of man-made UV light was found to be unexpectedly costly. The problem was then approached from the opposite direction. If naturally existing UV light alone were sufﬁcient to obtain photocatalytic effects, we would be relieved from the cost of supplying additional light energy. As long as the object we are treating stays within a limit treatable by naturally existing UV light (a maximum of about 1 mW/cm2 ) or the UV light present in indoor ﬂuorescent lamps (a maximum of about 1 µW/cm2 ), the technology should be very easy to put into practical use. There are several other details we need to consider in coming up with the most efﬁcient way to use TiO2 .

Photocatalytic Hydrophilicity/Hydrophobicity

427

As the photocatalytic reactions occur almost exclusively on the surface of the material, light needs to be able to get to the surface efﬁciently. Also, only substances adsorbed on the surface can react with the highly reactive radicals that are produced, such as ·OH and O2 .– [15–22]. Working within these parameters, the most effective way to use TiO2 was found to be in the form of thin ﬁlms that could be coated onto the surfaces of various materials. Additional beneﬁts are that the amount of material actually needed is very small, and since the ﬁlms are essentially transparent, thus unobtrusive, the presence of TiO2 is visually undetectable. In 1995 Watanabe et al. found by chance the remarkable change of surface wettability before and after UV-light irradiation through the process of investigating the photocatalytic properties of the TiO2 thin ﬁlm [23–29]. Because of this phenomenon materials coated by TiO2 showed various interesting properties including self-cleaning, antifogging, cooling effects and so on, which will be discussed later [30]. Already, a number of studies that have been carried out have demonstrated the effectiveness of these functions, oxidative decomposition and high hydrophilicity, in applications involving building materials. This approach promises to make both indoor and outdoor environmental cleaning possible, with antibacterial, stain-resistant, self-cleaning, and deodorizing functions.

2 Photocatalytic Oxidation Reaction We will now discuss the photocatalytic reactions occurring on the surface of TiO2 . In the use of the photocatalyst for environmental cleaning, reactions always occur in the presence of oxygen from air. The reaction starts with the exposure of TiO2 to UV light. After the light is absorbed by TiO2 , two types of carriers, electrons (e– ) and holes (h+ ), are generated. In most materials that are electrically conductive (i.e. metals) these are immediately recombined. With semiconductors such as TiO2 , however, they survive for longer periods of time. One of the notable characteristics of TiO2 is that the oxidizing power of the holes is greater than the reducing power of the excited electrons. On the surface of the photocatalyst, there is a single layer of tightly adhering (adsorbed) water molecules. When these adsorbed water molecules are oxidized by the holes, hydroxyl radicals (·OH), which are strongly oxidizing, are formed (Fig. 1). The hydroxyl radicals can then react with organic compounds, initially producing free radicals (unstable molecules that have one unpaired electron). Molecular oxygen also has unpaired electrons and when it is present it is expected to react with these free radicals, producing organic peroxyl radicals, which, in addition to containing an unpaired electron, also now contain two oxy-

428

H. Irie · K. Hashimoto

Fig. 1 Reaction mechanism of TiO2 photocatalysis

gens. These radicals can then take part in chain reactions. Within a short time the organic compounds are completely degraded, that is, converted into CO2 and H2 O [15–22]. Simultaneously, photoexcited holes, without oxidizing water molecules into hydroxyl radicals, directly oxidize organic compounds. Meanwhile, the electrons that are produced in the electron-hole pairs are also put to work. These electrons are used to reduce atmospheric oxygen. Because oxygen is easier to reduce than water, it will tend to be reduced, producing the superoxide radical anion (O2 .– ) (Fig. 1) [15–22]. The superoxide anion attaches itself to the peroxyl radicals. The resulting unstable product now contains at least four oxygens and can decompose organic compounds to produce CO2 . In addition to this mechanism, another interpretation proposed is that the formation in the air of so-called atomic oxygen (·O), which is extremely reactive, directly acts on the C – C bonds in organic materials. In general, organic compounds are more likely to be oxidized than water. Therefore, when the concentration of the organic compounds is high, the probability increases that the photogenerated holes will react directly with these compounds, rather than ﬁrst reacting with water to produce hydroxyl radicals. The holes are effectively “trapped” before they recombine with electrons. Under these conditions, the rate-determining step is the transfer of electrons to molecular oxygen. As a consequence, the overall efﬁciency of the photocatalytic process can increase by just speeding up the transfer of electrons to oxygen molecules. This can easily be accomplished by depositing small particles of metals (diameter of several nanometers) such as Pd and Pt on TiO2 particles.

Photocatalytic Hydrophilicity/Hydrophobicity

429

3 Photoinduced High Hydrophilic Conversion 3.1 Wettability on the Surface Wettability is one of the most important properties on solid surfaces. The surface wettability is generally evaluated by the contact angle. The contact angle (θ) is deﬁned as the angle between the solid surface and the tangent line of the liquid phase at the interface of three phases (Fig. 2). For the simplest case, the wettability of the solid surface is commonly evaluated by the contact angle given by Young’s equation [31]: γSV = γSL + γLV cos θ where γSV , γSL and γLV are the interfacial free energies per unit area of the solid-gas, solid-liquid and liquid-gas interfaces, respectively. The surface energy is determined by the chemical composition of the surface. Dropping the water droplet onto a solid surface with a higher surface energy, the droplet spreads more, that is, the water contact angle is lower. In general, a metal surface has a very high surface energy, hence the water droplet spreads completely over the surface. However, it is often observed that the metal oxide surface has a ﬁnite value of water contact angle. A surface with a high surface energy is thermodynamically unstable, so the adsorption of some chemical compound proceeds, causing a decrease in the surface energy. For example, metal surfaces such as iron and aluminum cause oxygen adsorption on their surfaces and are covered with oxide layers. Young’s equation is applicable only to an ideal surface, that is, homogeneous, rigid, insoluble and ﬂat. However, real surfaces have surface roughness and surface heterogeneity. Wentzel modiﬁed Young’s equation considering the surface roughness to obtain the following equation [32]. cos θ = γ cos θ where θ is the apparent contact angle, γ is the surface roughness ratio between the actual surface area and the apparent surface area. In the case of the

Fig. 2 Water contact angle deﬁned by Young’s equation

430

H. Irie · K. Hashimoto

hydrophilic surface, the rougher the surface the smaller the contact angle and vise versa in the case of hydrophobic surfaces. Another modiﬁcation was performed by Cassie considering the surface heterogeneity [33]. If two phases (phases 1 and 2) with different surface energies exists on the identical surface, the apparent contact angle is described as follows: cos θ = f1 cos θ1 + f2 cos θ2 which gives the contact angle θ of a liquid on a heterogeneous surface composed of fractions f1 and f2 , where θ1 and θ2 are the contact angles on the homogeneous surfaces of phases 1 and 2, respectively. For a surface composed of a solid and air, the water contact angle can be expressed as follows. When a unit area of the surface has a wetted solid surface area fraction f with a water contact angle θ, the contact angle on the surface is: cos θ = f cos θ + (1 – f ) cos 180◦ = f cos θ + 1 – f assuming the water contact angle for air is 180◦ . 3.2 What is High Hydrophilicity? In our daily environment, the surface of a material will repel water to some degree. The degree of the water repellency of a substrate can be expressed in terms of the contact angle of a water drop on its surface. On glass or other inorganic materials, water has a contact angle ranging from 20 to 30◦ . With water-repellent plastics, such as silicone resins and ﬂuoro-resins, the angle could be higher than 90◦ . Almost no substrates have been known to show angles lower than 10◦ , with the exception of some water-absorbing substances and surfaces that have been activated with soap or similar agents. These surfaces, however, do not have long-lasting effects. A TiO2 thin ﬁlm shows an initial water contact angle of several tens of degrees. On this surface, when it is exposed to UV light, water starts to show a decreasing contact angle, that is, it tends to spread out (ﬂat) instead of beading up. Finally, the contact angle reaches almost 0◦ . At this stage, the surface becomes completely non-water repellant and is called “highly hydrophilic”. The surface retains a contact angle of a few degrees for water for a day or two without being exposed to UV light. Then the contact angle slowly increases, and the surface becomes hydrophobic again. At this point, the high hydrophilicity can be recovered simply by exposing the surface again to UV light. In short, this type of photocatalyst is the only known practical highly hydrophilic material that shows stable, semipermanent properties. Figure 3 shows the results of an experiment designed to demonstrate this effect. Before the surface was exposed to UV light, the water contact angle of

Photocatalytic Hydrophilicity/Hydrophobicity

431

Fig. 3 Change in the shape of water droplets after irradiation: (left) before irradiation, (right) after irradiation

the surface was greater than 40◦ , and the water beaded up, forming droplets on the surface. After being exposed to UV light, the contact angle reached 0◦ , and the water took the form of a highly uniform thin ﬁlm. How can we explain this phenomenon? We considered various pointsof-view in our analysis and ﬁnally came up with the answer which will be discussed later in this article. With the TiO2 photocatalyst alone, this high hydrophilicity disappears once the exposure of the surface to UV light is stopped. When the photocatalyst is combined with a water-storing substance, such as SiO2 , the effect of high hydrophilicity continues even in the dark. Photocatalytic technology inherently requires some quantity of light, and highly hydrophilic technology is no exception. When it is to be used indoors, the quantity of light provided by indoor lighting may not be sufﬁcient. Our research and development efforts have focused on this point, and we will discuss this later. 3.3 Applications Fogging of the surfaces of mirrors and glasses occurs when steam cools down on these surfaces to form water droplets. On a highly hydrophilic surface, no water droplets are formed. Instead, a uniform ﬁlm of water is formed on the surface. This uniform water ﬁlm prevents fogging. We expect that various glass products, mirrors and eyeglasses, for example, can be imparted with antifogging functions using this technology, with simple processing and at low cost. In fact, Japanese-made cars are being equipped with antifogging highly hydrophilic side-view mirrors, as shown in Fig. 4. Strain-prooﬁng and self-cleaning effects can be enhanced through the highly hydrophilic function. As an example, a plastic surface smeared with oil cannot be cleaned unless one uses detergent. A highly hydrophilic surface,

432

H. Irie · K. Hashimoto

Fig. 4 Automobile side-view mirror: (left) conventional mirror, (right) TiO2 -coated mirror

however, has a higher afﬁnity for water than for oil. An oil smear on a plastic utensil is released from the plastic surface when the utensil is simply soaked in water. Based on this characteristic, a kitchen exhaust fan, which is likely to be covered with oil, could be easily cleaned by water if the fan blades were coated with a highly hydrophilic photocatalyst. Outdoor applications of this technique are also possible. Most of the exterior walls of buildings become soiled from automotive exhaust fumes, which contain oily components. If the original building materials are coated with a highly hydrophilic photocatalyst, the dirt on the walls will wash away with rainfall, keeping the building exterior clean at all times. Fig. 5 shows the smear-resistant effect of a highly hydrophilic photocatalyst coated on exterior concrete panels in a checkerboard fashion, the contrast was easily seen between the regular exterior panels, on which soiling was very conspicuous, and

Fig. 5 Field test of strain-resistant exterior tiles in polluted urban air: a highly hydrophilic treated tiles, b normal tiles

Photocatalytic Hydrophilicity/Hydrophobicity

433

Fig. 6 Comparison of self-cleaning effects: (left) conventional photocatalytic oxidation process, (right) washing away by rainwater due to photo-induced highly hydrophilic surface

the photocatalyst exterior material, which was not soiled at all. This photocatalytic building material is projected to have a life of 10 years or longer. Photocatalytic technology has been applied to the decomposition of noxious gases, odor-causing gases, dirt, and so on. In other words, the technology involves materials that have been carried to the surface from the surrounding environment. A high hydrophilic effect is based on a concept of altering the properties of the surface itself through the photocatalytic effect. Although both technologies (photocatalytic oxidative decomposition and high hydrophilicity) are applied to the prevention of soiling, their basic mechanisms are quite different as shown in Fig. 6. The basic idea is that the same TiO2 material can have both types of properties, photocatalytic oxidation and high hydrophilicity, in varying proportions, depending on the conditions where the TiO2 material is used. For example, at a roadside with heavy trafﬁc and no rainfall the self-cleaning effect (of surface strains) is based just on the property of photocatalytic oxidative decomposition, whereas in locations where there is rainfall both the photocatalytic oxidative decomposition and the high hydrophilicity effects are active and involved in the self-cleaning. The applications of this highly hydrophilic technology are not limited to antifogging and self-cleaning. The highly hydrophilic surface has many additional beneﬁts, so the scope of this technology has been extended to other ﬁelds. When water vapor pressure is lower than saturated vapor pressure, water evaporates, generating latent heat ﬂux, accompanied by a cooling down of the surrounding atmosphere. Sprinkling a very small amount of water continuously onto the highly hydrophilic TiO2 surface, effective cooling can be achieved. A very thin water layer with a thickness of approximately 0.1 mm can cover all of the highly hydrophilic TiO2 material, even though it stands vertically, if only a small amount of water is supplied continuously. This tech-

434

H. Irie · K. Hashimoto

nology is applicable to the building of walls. The TiO2 -coated walls become highly hydrophilic through exposure to solar light and thus the sprinkled water will form a thin layer and be evaporated efﬁciently from the surface. The evaporation of the water generates latent heat ﬂux, which cools the building surfaces and the surrounding atmosphere. The cooling effect reduces the usage of air conditioning, saves on energy consumption, and reduces the artiﬁcial heat emission. This technology would be useful for the prevention of the heat island phenomenon and for saving energy, for which the photoinduced highly hydrophilic characteristic of the TiO2 surface plays an important role in minimizing the amount of sprinkled water needed for the formation of the water ﬁlm. 3.4 Mechanism for Highly Hydrophilic Conversion 3.4.1 Negative Experimental Results for Conversion by Oxidative Decomposition Generally, metal oxides have large surface energies, so a clean TiO2 surface is essentially hydrophilic. Earlier in our work, the present authors considered that this highly hydrophilic conversion originated from the clean surface produced by the decomposition of the strains adsorbed on the surface through conventional photocatalytic oxidation. However, after a large number of experimental results which did not suggest the conversion by photocatalytic decomposition was taking place, the authors came to the conclusion that the surface had in fact undergone some structural change. Some examples are listed below. 3.4.1.1 The Relationship Between Adsorbed Surface Strains and Water Contact Angles Figure 7 shows the change in water contact angles under UV-light irradiation after applying oleic acid to the surfaces of TiO2 and SrTiO3 . Just after applying the oleic acid, both the TiO2 and SrTiO3 surfaces were converted to the hydrophobic state with contact angles of around 70◦ due to the hydrophobic property of oleic acid adsorbed on the surfaces. When the SrTiO3 surface was irradiated with UV light, the water contact angle decreased to 20◦ , which was the same value as the water contact angle before applying oleic acid. As a result of the conventional photocatalytic oxidation processes, the surface adsorbed oleic acid was decomposed and removed, reproducing the initial clean surface. For the TiO2 surface, however, the water contact angle decreased to 0◦ , which was lower than the initial value, showing that the highly hydrophilic surface of TiO2 is not simply attributed to the effect of the power of photocatalytic oxidation [29].

Photocatalytic Hydrophilicity/Hydrophobicity

435

Fig. 7 Changes in water contact angle after applying oleic acid to the surfaces of TiO2 and SrTiO3

In fact, though the surface strains were completely removed by the treatment of the TiO2 surface in a warm concentrated NaOH solution, this leads to a water contact angle of around 10–20◦ . At this point, a contaminant C1s peak (approximately at 285 eV) was not detected in the X-ray photoemission spectroscopy (XPS) measurement, i.e. no strains remained at the surface. On the contrary, it was conﬁrmed that even if the surface strains remained, the highly hydrophilic conversion still occurred under UV-light irradiation [34]. 3.4.1.2 Hydrophobic Conversion in the Dark by External Perturbation As mentioned above, the highly hydrophilic state generated by UV light gradually returns to the initial hydrophobic state—of the surface— in the dark. This reaction can be accelerated by external perturbation. Figure 8 shows the ambient temperature dependence of hydrophobic changes in water contact angle on a TiO2 surface in the dark. Because this measurement was carried out in a vessel ﬁlled with synthetic air, where the only condition varied was temperature, even if strains were adsorbed on the surface any inﬂuence from the strains could be ignored. It is obvious that the hydrophobic reaction was enhanced by the increasing temperature. When the TiO2 thin ﬁlm was sonicated in pure water after highly hydrophilic treatment by UV light, the contact angles were found to increase to around 10◦ , as shown in Fig. 9. When a similar highly hydrophilic TiO2 was simply stored in the dark, it required approximately three weeks for the contact angle to increase to 10◦ . Subjected to cycles of alternating ultrasonic treatment and UV-light irradiation, the water contact angles switched between 0◦ and 10◦ reversibly. As the highly hydrophilic TiO2 was sonicated in pure water, the surface strains were removed and the hydrophilicity was enhanced. However, contrary to our expectations, the ultrasonic treatment leads

436

H. Irie · K. Hashimoto

Fig. 8 Ambient temperature dependence of the change in water contact angle on a highly hydrophilic TiO2 surface in the dark

Fig. 9 Changes in water contact angle subjected to alternate ultrasonic treatment in pure water (•) and UV-light irradiation ()

to a less hydrophilic state [26]. Figure 10 shows the changes in water contact angles on a highly hydrophilic TiO2 surface in a vacuum and in ambient air. Both are under dark conditions. As shown in Fig. 10, the hydrophobic reaction was accelerated by simply storing under vacuum conditions. Other external perturbations, except for heat and vacuum treatments, could accelerate the hydrophobic reaction in the dark. For example, Kamei et al. reported that “wet-rubbing” with a piece of cloth realized a switch of the hydrophilic surface (water contact angle of 3◦ ) to the hydrophobic surface (80◦ ) [35]). Taking all these experimental results into consideration, the removal of strains by oxidative decomposition can make the surface “moderately” hydrophilic with a water contact angle of 10–20◦ . Therefore, it can be concluded that the “highly” hydrophilic state with a water contact angle of 0◦ was achieved by the generation of a metastable state with a large surface energy irradiated with UV light.

Photocatalytic Hydrophilicity/Hydrophobicity

437

Fig. 10 Changes in water contact angle on a highly hydrophilic TiO2 surface in the dark, in a vacuum and in ambient air

Recently, some groups reported that the phenomenon of this UV-lightinduced highly hydrophilic conversion is caused by the conventional photocatalytic reaction, that is, the oxidative decomposition of surface strains. For example, Henderson et al. reported hydrophobic and hydrophilic switching by adsorbing an organic compound (trimethyl acetate) and removing it by irradiation with UV light under ultra-high vacuum conditions (10–6 Torr) [36]. They showed that the clean surface which had shown hydrophilicity became hydrophobic after applying trimethyl acetate and that this hydrophobic surface regained hydrophilicity under UV-light irradiation. Schultz et al. concluded that the oxidative decomposition of the hydrocarbon ﬁlm on the TiO2 surface was responsible for the hydrophilic conversion [37]. However, Henderson et al. and Schultz et al. only showed that the surface organic compounds were decomposed by the conventional photocatalytic reaction, accompanied by a decrease in the water contact angle to a “moderately” hydrophilic state. Neither discussed the “highly” hydrophilic state. 3.4.2 Proposed Mechanism for the Highly Hydrophilic Conversion Some experimental evidences obtained showed that the photoinduced high hydrophilicity peculiar to TiO2 is attributed to some physical structural changes at the surface, which are different from the photocatalytic oxidation process. To gain information about surface wettability at a microscopic level, friction force microscopy (FFM) was utilized. A rutile TiO2 (110) single crystal was used. Before UV-light irradiation, no difference in contrast was observed (Fig. 11), indicating microscopically homogeneous wettability. After irradiation, hydrophilic (bright) and hydrophobic (dark) areas of size

438

H. Irie · K. Hashimoto

30–80 nm were clearly observed (Fig. 11) [23, 25]. A gradual reversion to smaller contrast was observed during the storage of the crystal in the dark. The conclusion stated that the nanoscale separation between the hydrophilic and the hydrophobic phases accounts for the highly hydrophilic character on the TiO2 surface. Hardness changes were also observed on the surface of the rutile TiO2 (110) single crystal before and after being irradiated with UV light. Figure 12 shows the dependence of the surface (within 50 nm from the surface) and inner (300 nm from the surface) hardness and the water contact angles on the switching on and off of the UV light. When irradiated with UV light, water contact angles decreased to 0◦ , accompanied by an increase in the surface hardness. In contrast, the inner hardness remained constant. These ﬁndings suggested that the highly hydrophilic surface had a compressive stress.

Fig. 11 FFM images of a rutile TiO2 (110) single crystal surface, (left) before UV-light irradiation, (right) after UV-light irradiation

Fig. 12 Relationship between changes in water contact angle and hardness at the surface (within 50 nm) and inside (300 nm)

Photocatalytic Hydrophilicity/Hydrophobicity

439

Some evidences for the chemical change at the TiO2 surface were also obtained. This chemical change was due to the increase in the amount of hydroxyl (OH) groups at the highly hydrophilic TiO2 surface, detected by X-ray photoemission spectroscopy (XPS) [27]. The O1s spectrum for the highly hydrophilic TiO2 surface exhibited a broad shoulder to the higher binding energy side of the main O1s peak. The shoulder was ﬁtted with two bands, which are associated with dissociatively adsorbed water on the TiO2 surface as well as physically adsorbed molecular water on the dissociatively adsorbed OH groups [25]. The shoulder gradually decreased during a period of storage in the dark. Therefore, these photogenerated surface OH groups were considered to be thermodynamically metastable. Fourier transformation infrared (FTIR) spectroscopy provides another way to investigate the relationship between hydrophilicity and the amount of OH groups. TiO2 shows IR bands positioned at 3695 cm–1 , assigned to the stretching of OH groups chemisorbed on the surface, 3300 cm–1 , assigned to hydroxyls for both dissociated water and molecularly adsorbed water, and 1623 cm–1 , pertaining to H – O – H bending for molecular water. This observation denotes the coexistence of dissociated water and molecular water on the TiO2 surface. Storage in the dark for one week resulted in the decrease of all the bands, correlated with both dissociated and molecular water desorptions. The signiﬁcant increase in the water contact angle after storage in the dark is consistent with this surface conversion. The surface conversion is ascribed to the removal of the chemisorbed hydroxyl groups. With UV light irradiation, however, the 1623 cm–1 band decreased whereas the 3695 cm–1 band increased, demonstrating the increase in the amount of adsorbed dissociated water and the decrease of adsorbed molecular water. Nosaka et al. also observed the increase in the amount of OH groups under UV-light irradiation with nuclear magnetic resonance (NMR) measurements [38]. The photogenerated holes, not the electrons, were found to be responsible for the highly hydrophilic conversion, as they are enhanced under anodic polarization and depressed in the presence of sodium sulﬁte, which is known as a hole scavenger [24]. On the basis of the previous studies mentioned above, the mechanism for the highly hydrophilic conversion under UV-light irradiation is proposed as shown in Fig. 13. When the photogenerated holes are diffused to the surface and trapped at lattice oxygen sites of the TiO2 lattice, producing OH radicals or oxidizing the adsorbed organics directly, the binding length between the titanium cation and the lattice oxygen expands. Actually, it has been reported recently that the lattice spacing for the TiO2 surface increases after UV-light irradiation; this was shown using energy dispersive total-reﬂection X-ray diffraction. Then, as the expanded Ti – O bond induces molecular water adsorption the molecular water simultaneously releases a proton for charge compensation, and then the new OH groups form, leading to an increase in the amount of OH groups at the

440

H. Irie · K. Hashimoto

Fig. 13 Schematic illustrations of reversible changes in the amount of hydroxyl group under UV-light irradiation and in the dark

surface. Thus, one OH group doubly coordinated to Ti atoms at an oxygen defect site converts to two OH groups singly coordinated to each Ti atom. It is considered that the new singly coordinated OH groups produced by UV-light irradiation are thermodynamically less stable compared to the initial doubly coordinated OH groups. Therefore, the surface energy of the TiO2 surface covered with the thermodynamically less stable OH groups is higher than that of the TiO2 surface covered with the OH groups initially. Because a water droplet is substantially larger than the hydrophilic (or hydrophobic) domain as shown in Fig. 11, it instantaneously spreads completely over such a surface, thereby resembling the two-dimensional capillary phenomenon. 3.5 High Sensitization Under Weak UV-Light Irradiation Particularly when utilizing photoinduced hydrophilic properties, TiO2 has not been applied to indoor use so far, as this highly hydrophilic conversion only occurs under UV-light irradiation with the approximate intensity of natural sunlight. Therefore, a very sensitive TiO2 surface is required for highly hydrophilic conversion. An example would be TiO2 irradiated with extremely weak UV light with an intensity of 1 µW/cm2 , which can be obtained from regular indoor ﬂuorescent lighting. Studies of this are shown below. 3.5.1 Heterogeneous System, TiO2 /WO3 Besides single-component photocatalysts, heterogeneous metal-semiconductor systems, such as Pt/TiO2 , Pd/TiO2 , Ag/TiO2 , and Cu/TiO2 [39–41],

Photocatalytic Hydrophilicity/Hydrophobicity

441

and heterogeneous semiconductor systems, such as CdS/TiO2 , TiO2 /SnO2 , WO3 /WS2 , and TiO2 /WO3 [42–44] have been studied. It has been reported that heterogeneous semiconductor systems suppress the recombination of photogenerated charge carriers and improve the efﬁciency of photocatalytic reactions, because of vectorial transfers of photogenerated carriers from one semiconductor to another. The modiﬁcation of TiO2 by WO3 increased the sensitivity of the photoinduced hydrophilic reaction and achieved high hydrophilicity under 1 µW/cm2 UV-light irradiation (ﬂuorescent lamp), as shown in Fig. 14 [45]. In the TiO2 /WO3 system, as the band-gaps of TiO2 (anatase) and WO3 are 3.2 eV and 2.8 eV, respectively, and the ﬂat band potentials are – 0.1 V and + 0.5 V (vs. hydrogen electrode at pH = 0), respectively, the valence and conduction bands of WO3 are considered to be at a lower level when compared to those of TiO2 . As is shown in Fig. 15, upon optical excitation, photogenerated electrons accumulated at the lower-lying conduction band of WO3 , whereas holes accumulated at the valence band of TiO2 . Accumulated electrons at the conduction band of WO3 can be transferred to oxygen adsorbed on the surface. Also, the accumulation of holes at the va-

Fig. 14 Changes in water contact angle, under irradiation with a ﬂuorescent light bulb, of TiO2 and TiO2 /WO3

Fig. 15 Energy diagrams for the heterogeneous TiO2 /WO3 thin ﬁlm

442

H. Irie · K. Hashimoto

lence band of TiO2 is favorable for increasing OH groups, as holes play an important role for hydrophilicizing. For the heterogeneous TiO2 /WO3 system, as photogenerated holes were effectively accumulated to TiO2 without recombination with electrons, and as WO3 could absorb Vis light contained in the ﬂuorescent light up to a wavelength of around 440 nm due to its band-gap of 2.8 eV, the hydrophilicizing reaction was greatly enhanced [45]. 3.5.2 Control of Surface Nano-Structure As the reconstruction of surface OH groups is responsible for the hydrophilic conversion, the different hydrophilic conversion behaviors among the various single crystal faces should be observable by comparing their surface atom alignments. In fact, the (100) and (110) faces of a single crystal of rutile exhibited higher hydrophilic properties than the (001) face, as the former faces have two-fold oxygens, which are higher in position and energetically more reactive than their surrounding atoms and are so-called “bridging site oxygens” whereas the latter has three-fold oxygens, which are lower in position and energetically less active. Therefore, the (100) and (110) faces were favorable for increasing the number of OH groups. This ﬁnding could serve as a guideline for the selection of faces of a TiO2 crystal to expose as the surface. The photoetching technique was applied to the (001) face of a single crystal of rutile. As shown in Fig. 16, a large number of rectangular porous holes which look like “wells” with a size of 50–100 nm emerged with a fairly regular arrangement. According to Sugiura et al. [46] and Nakato et al. [47], the rectangular holes proceed toward the <001> direction (c-axis direction) and the (100) face or equivalents were exposed selectively at the walls. This rectangular porous surface increased the sensitivity of the photoin-

Fig. 16 SEM image of rectangular holes emerging after photoetching treatment on the (001) surface of a single crystal of rutile

Photocatalytic Hydrophilicity/Hydrophobicity

443

Fig. 17 Changes in water contact angles on the (001) surface with and without the photoetching treatment

duced hydrophilic reaction and achieved high hydrophilicity under UVlight irradiation of 1 µW/cm2 , whereas the surface without the photoetching treatment did not, as shown in Fig. 17 [48]. Because the surface roughness enhances the hydrophilicity as expressed by Wentzel’s equation, the high sensitization for hydrophilic conversion provided by the photoetching treatment was caused by the exposure of the face with bridging site oxygens and the increase in the surface roughness. This high sensitization was also observed on the rutile polycrystal after the photoetching treatment, which suggested that this technique was also applicable to an anatase thin ﬁlm. 3.5.3 Introduction of Residual Tensile Stress As the hydrophilic conversion is caused by the increase in the amount of surface OH groups, the Ti – O bond length should be greater in the highly hydrophilic surface than in the less hydrophilic surface. Actually, it was reported that the lattice spacing for the TiO2 surface increases after UV-light irradiation. Therefore, it can be expected that the TiO2 surface in the presence of tensile stress is favorable for hydrophilic conversion. In fact, the residual tensile stress in the anatase thin ﬁlm fabricated by the sputtering technique greatly enhanced the hydrophilic behavior, compared to the thin ﬁlm without residual stress, as shown in Fig. 18 [49]. In contrast, the photocatalytic oxidation activities were identical for the two ﬁlms. The surface morphologies of the two ﬁlms were both the same, therefore residual stress is an important factor inﬂuencing highly sensitive photoinduced hydrophilicity of a ﬁlm.

444

H. Irie · K. Hashimoto

Fig. 18 Changes in water contact angles when irradiated with UV light, a a thin ﬁlm without residual stress and b one with residual tensile stress

3.6 Visible Light Sensitivity TiO2 is activated only by irradiation with UV light. A current area of research in this ﬁeld has been to modify TiO2 so that it is sensitive to Vis light. One approach was to substitute Cr, Fe or Ni, etc., for a Ti site [50– 52]. Anpo et al. substituted Cr3+ or V3+ (V4+ ) at the lattice positions of Ti4+ in TiO2 using a metal ion-implantation method [51, 52]. They showed that the absorption band of the Cr3+ -doped TiO2 shifted to the Vis light region and that irradiating with Vis light (wavelength > 450 nm) photocatalytically decomposed NO into N2 , O2 and N2 O. Another approach is to form Ti3+ sites by introducing an oxygen vacancy in TiO2 [53, 54]. However, there seemed to be some problems in terms of reproducibility and chemical stability. Recently, nitrogen-doped TiO2 has been reported to be a Vis light corresponding photocatalyst [55–61]. The nitrogen-doped TiO2 , both yellowish and transparent (Fig. 19), can be fabricated by sputtering a Ti or TiO2 target in a gas ﬂow containing N2 . The dopant nitrogen was found to be located at the oxygen site considering the N1s XPS peak at 396 eV. The hydrophilic conversion behaviors of nitrogen-doped TiO2 thin ﬁlm with a dopant concentration of 0.6 mol % irradiated with Vis light (400–530 nm) are shown in Fig. 20. The high hydrophilicity could be achieved only by Vis-light irradiation. The incorporated nitrogen forming the isolated (or localized) N2p narrow band above O2p, constituting the valence band, or a narrowing band gap by mixing N2p and O2p was responsible for the Vis-light sensitivity [56–58]. In addition to nitrogen, sulfur-doped or carbon-doped TiO2 were conﬁrmed to be Vis-light sensitive photocatalysts [62–65].

Photocatalytic Hydrophilicity/Hydrophobicity

445

Fig. 19 Optical absorbance spectra of TiO2 and nitrogen-doped TiO2 thin ﬁlms

Fig. 20 Changes in water contact angles on TiO2 and nitrogen-doped TiO2 thin ﬁlms when irradiated with Vis light

4 High Hydrophobicity 4.1 Highly Hydrophobic Films and Surfaces Various industrial products require not only hydrophilicity but also hydrophobicity. Currently, a surface with a water contact angle above 150◦ (i.e. a highly hydrophobic surface, Fig. 21) is attracting great attention [66]. Given the limited contact area between the solid surface and water, chemical reactions or bond formation through water are limited on a highly hydrophobic surface. Accordingly, various phenomena such as the adherence of snow, oxidation, and current conduction are expected to be inhibited on such a surface. So far, various methods have been developed for the processing of highly hydrophobic surfaces or ﬁlms. Highly hydrophobic surfaces or ﬁlms require a large surface roughness and a low surface energy, considering Cassie’s equation. From the viewpoint of surface roughness, hydrophobicity and transparency are competitive properties [67–73]. Providing

446

H. Irie · K. Hashimoto

Fig. 21 Water droplet on the highly hydrophobic surface, with water contact angle of more than 150◦

surface roughness means introducing sources of light scattering. When the roughness increases, the hydrophobicity also increases, whereas the transparency decreases. Therefore, precise roughness control, less than 100 nm, is required to satisfy both properties. Only a few groups have succeeded in fabricating transparent highly hydrophobic thin ﬁlms. For example, Nakajima et al. [73] imparted surface roughness to boehmite or silica ﬁlms by the sublimation of aluminum acetylacetonate (Al(C5 H7 O2 )3 ) during calcination, and subsequently they were able to prepare transparent highly hydrophobic thin ﬁlms from these materials by coating them with heptadecaﬂuorodecyl trimethoxysilane. Over long periods of outdoor exposure, the highly hydrophobic surface, artiﬁcially constructed, gradually degrades due to the accumulation of strains that adhere to the surface. However, natural highly hydrophobic surfaces such as lotus leaves avoid this problem by continuous metabolism of their surface wax layer, which makes it possible to maintain hydrophobicity throughout their lifetime [74–76]. Since the exact metabolic mechanism of a lotus leaf is impossible to mimic, practical applications of highly hydrophobic surfaces have not been successful. Recently, Nakajima et al. [77], and Yamauchi et al. [78], discovered that the addition of a few percent of TiO2 photocatalyst effectively provided a self-cleaning property to highly hydrophobic ﬁlms and maintained high contact angles during long periods of outdoor exposure (Fig. 22). Surfaces with a high roughness commonly show fewer mechanical properties than ﬂat surfaces, and this is a crucial problem for the application of highly hydrophobic surfaces. A needle-like structure is known as an ideal surface for high hydrophobicity [79]. However, this structure is generally insufﬁcient for practical use. To increase the hardness of highly hydrophobic surfaces, a crater-like structure is thought to be preferable to a needle-like one. In the case of a crater-like structure, however, it is difﬁ-

Photocatalytic Hydrophilicity/Hydrophobicity

447

Fig. 22 Effect of TiO2 addition on the hydrophobicity of a highly hydrophobic thin ﬁlm during outdoor exposure; the numbers refer to TiO2 concentration (wt%)

cult to obtain a highly hydrophobic state on the surface due to the increase in the contact area between the solid and liquid. Recently, a hard highly hydrophobic silica ﬁlm with Vis light transmission was successfully prepared by combining two different roughness dimensions, combining a craterlike roughness (about 800 nm) prepared by a phase separation achieved by the sol gel method and a ﬁne roughness (about 20 nm) provided by colloidal silica [80, 81]. The concept of roughness combination might be important for the design of highly hydrophobic ﬁlms with long durability. 4.2 Applications of Highly Hydrophobic Thin Films By applying a highly hydrophobic ﬁlm to the surface, snow was reported to be less likely to adhere to the surface due to the large amount of air between the surface and snow. A satellite antenna covered with a highly hydrophobic ﬁlm, reduced the disruption of communication by the adherence of snow to the antenna [78, 82]. Besides, these results portended well for techniques to apply highly hydrophobic ﬁlm coatings to trafﬁc lights, billboards, and roofs in areas with heavy snowfall. It was reported that the friction resistance between the highly hydrophobic surface and water was reduced to 20–45%, by supplying a small amount of air to the highly hydrophobic surface, as a ﬁlm of air ﬂow formed along the surface in water [83, 84]. This report suggests that the highly hydrophobic ﬁlms are effective for use on the hulls of ships and for tubes or pipes. In addition to these properties, highly hydrophobic coatings are expected to offer beneﬁcial properties for anti-oxidation or anti-current conduction.

448

H. Irie · K. Hashimoto

5 Conclusion This paper provides an overview of the TiO2 photocatalyst, including photocatalytic oxidative decomposition and photoinduced high hydrophilicity and high hydrophobicity (the opposite to high hydrophilicity). TiO2 is a basic and simple material, and more than thirty years have passed since the discovery of the TiO2 photocatalytic reaction. The wettability of solid surfaces is also a basic and familiar property. It is amazing that the excellent characteristics (high hydrophilicity among other properties) of the already well understood TiO2 were discovered only recently. In this paper, we introduced the surface reconstruction model of hydroxyl groups as the mechanism for the highly hydrophilic conversion. However, areas for further research still remain. The wettability of a solid surface, in terms of both high hydrophilicity and hydrophobicity, is complex and affects various other surface properties. This is a key technology for industries, and further investigations on the control of the wettability of solid materials are expected.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.

Fujishima A, Honda K (1972) Nature 238:37 Fujihira M, Satoh Y, Osa T (1981) Nature 293:206 Kraeutler B, Bard AJ (1977) J Am Chem Soc 99:7729 Ollis DF (1985) Environ Sci Technol 19:480 Hashimoto K, Kawai T, Sakata T (1984) J Phys Chem 88:4083 Sakata T, Kawai T, Hashimoto K (1984) J Phys Chem 88:2344 Nosaka Y, Koenuma K, Ushida K, Kira A (1996) Langmuir 12:736 Fu X, Zeltner WA, Anderson MA (1995) Appl Catal B Environ 6:209 Heller A (1995) Acc Chem Res 28:503 Paz Y, Luo Z, Rabenberg L, Heller A (1995) J Mater Res 10:2842 Sopyan I, Watanabe M, Murasawa S, Hashimoto K, Fujishima A (1996) J Electroanal Chem 415:183 Dillert R, Siemon U, Bahnemann D (1999) J Adv Oxidation Tech 4:55 Benz V, Mueller M, Bahnemann D, Weichgrebe D, Brehm M (1996) Eur Pat Appl 7 Schwitzgebel J, Ekerdt JG, Gerischer H, Heller A (1995) J Phys Chem 99:5633 Fox MA, Chen CC (1981) J Am Chem Soc 103:6757 Cunningham J, Srijaranai S (1988) J Photochem Photobiol A: Chem 43:329 Brezová V, Stasko A, Lapcik L Jr (1991) J Photochem Photobiol A: Chem 59:115 Kamat PV (1993) Chem Rev 93:267 Courbon H, Formenti M, Pichat P (1977) J Phys Chem 81:550 Anpo M, Chiba K, Tomonari M, Coluccia S, Che M, Fox MA (1991) Bull Chem Soc Jpn 64:543 Sun L, Bolton JR (1996) J Phys Chem 100:4127 Grela MA, Coronel MEJ, Colussi AJ (1996) J Phys Chem 100:16940

Photocatalytic Hydrophilicity/Hydrophobicity

449

23. Wang R, Hashimoto K, Fujishima A, Chikuni M, Kojima E, Kitamura A, Shimohigoshi M, Watanabe T (1997) Nature 388:431 24. Sakai N, Fujishima A, Watanabe T, Hashimoto K (2001) J Phys Chem B 105:3023 25. Wang R, Hashimoto K, Fujishima A, Chikuni M, Kojima E, Kitamura A, Shimohigoshi M, Watanabe T (1998) Adv Mater 10:135 26. Sakai N, Wang R, Fujishima A, Watanabe T, Hashimoto K (1998) Langmuir 14:5918 27. Wang R, Sakai N, Fujishima A, Watanabe T, Hashimoto K (1999) J Phys Chem B 103:2188 28. Watanabe T, Nakajima A, Wang R, Minabe, Koizumi S, Fujishima A, Hashimoto K (1999) Thin Solid Films 351:260 29. Miyauchi M, Nakajima A, Fujishima A, Hashimoto K, Watanabe T (2000) Chem Mater 12:3 30. Hata S, Kai Y, Yamanaka I, Oosaki H, Hirota K, Yamazaki S (2000) JSAE Rev 21:97 31. Hiemenz PC (ed) (1986) Principles of Colloid and Surface Chemistry. Marcel Dekker, New York, p 307 32. Wentzel RN (1949) J Phys Colloid Chem 53:1466 33. Cassie ABD (1948) Discuss Faraday Soc 3:11 34. Sun RD, Nakajima A, Fujishima A, Watanabe T, Hashimoto K (2001) J Phys Chem B 105:1984 35. Kamei M, Mitsuhashi T (2000) Surf Sci 463:L609 36. White JM, Szanyi J, Henderson MA (2003) J Phys Chem B 107:9029 37. Wang C, Groenzin H, Schultz MJ (2003) Langmuir 19:7330 38. Nosaka AY, Kojima E, Fujiwara T, Yagi H, Akutsu H, Nosaka Y (2003) J Phys Chem B 107:12042 39. Avalle L, Santos E, Leiva E, Macagno VA (1992) Thin Solid Films 219:7 40. Papp J, Shen HS, Kershaw R, Dwight K, Wold A (1993) Chem Mater 5:284 41. Lee W, Shen HS, Dwight K, Wold A (1993) J Solid State Chem 106:288 42. Serpone N, Maruthamuthu P, Pichat P, Pelizzetti E (1995) J Photochem Photobiol A 85:247 43. Serpone N, Borgarello E, Pelizzetti E (1988) J Electrochem Soc 135:2760 44. Bedja I, Kamat PV (1995) J Phys Chem 99:9182 45. Miyauchi M, Nakajima A, Hashimoto K, Watanabe T (2000) Adv Mater 12:1923 46. Sugiura T, Yoshida T, Minoura H (1999) Electrochem 67:1234 47. Tsujiko A, Kisumi T, Magari Y, Murakoshi K, Nakato Y (2000) J Phys Chem B 104:4873 48. Shibata T, Watanabe T, Hashimoto K, Nakajima A (2001) Photocatalysis 4:45 (in Japanese) 49. Shibata T, Irie H, Hashimoto K (2003) J Phys Chem B 107:10696 50. Borgarello E, Kiwi J, Grätzel M, Pelizzetti E, Visca M (1982) J Am Chem Soc 104:2996 51. Yamashita H, Ichihashi Y, Takeuchi M, Kishiguchi S, Anpo M (1999) J Synchrotron Rad 6:451 52. Anpo M, Takeuchi M (2001) Int J Photoenergy 3:1 53. Nakamura I, Negishi N, Kutsuna S, Ihara T, Sugihara S, Takeuchi K (2000) J Molecular Catalysis A: Chem 161:205 54. Ando K, Ando M, Fujii M (2001) Photocatalysis 6:20 (in Japanese) 55. Sakatani Y, Koike H (2001) Japan Patent P2001-72419A (in Japanese) 56. Asahi R, Ohwaki T, Aoki K, Taga Y (2001) Science 293:269 57. Irie H, Watanabe Y, Hashimoto K (2003) J Phys Chem B 107:5483 58. Irie H, Washizuka S, Yoshino N, Hashimoto K (2003) Chem Comm 11:1298 59. Sakthivel S, Kisch H (2003) Chem Phys Chem 4:487 60. Sato S (1986) Chem Phys Lett 123:126

450

H. Irie · K. Hashimoto

61. Noda H, Oikawa K, Ogata T, Matsuki K, Kamata H (1986) The Chem Soc Jap 8:1084 (in Japanese) 62. Umebayashi T, Yamaki T, Itoh H, Asai K (2002) Appl Phys Lett 81:454 63. Umebayashi T, Yamaki T, Tanaka S, Asai K (2003) Chem Lett 32:330 64. Ohno T, Mitsui T, Matsumura M (2003) Chem Lett 32:364 65. Irie H, Watanabe Y, Hashimoto K (2003) Chem Lett 32:772 66. Nakajima A, Hashimoto K, Watanabe T (2001) Monatshefte fur Chemie 132:31 67. Ogawa N, Soga M, Takada Y, Nakayama I (1993) Jpn J Appl Phys 32:L614 68. Tadanaga K, Katata N, Minami T (1997) J Am Ceram Soc 80:1040 69. Hozumi A, Takai O (1997) Thin Solid Films 303:222 70. Takai O, Hozumi A, Inoue Y, Komori T (1997) Bull Mater Sci 20:817 71. Tadanaga K, Katata N, Minami T (1997) J Am Ceram Soc 80:3213 72. Hozumi A, Takai O (1997) Thin Solid Films 334:54 73. Nakajima A, Fujishima A, Hashimoto K, Watanabe T (1999) Adv Mater 11:1365 74. Neinhuis C, Barthlott W (1997) Ann-Bot 79:667 75. Wirthensohn MG, Sedgley M (1996) Australian J Botany 44:691 76. Bitterlich I, Upadhyaya MK (1990) Canadian J Botany 68:1911 77. Nakajima A, Hashimoto K, Watanabe T, Takai K, Yamauchi G, Fujishima A (2000) Langmuir 16:7044 78. Takai K, Saito H, Yamauchi G (1997) Proceedings of the Composites: Design for Performance. Lake Louise, Canada, p 220 79. Onda K (1997) Tsukuba Research Consortium 4:56 (in Japanese) 80. Nakajima A, Abe K, Hashimoto K, Watanabe T (2000) Thin Solid Films 376:140 81. Nakajima A, Yoshimitsu Z, Saiki C, Hashimoto K, Watanabe T (2001) Proc 7th Int Conf on Ceramic Processing Science. Inuyama, Japan 82. Yamauchi G, Saito H, Takai K (1998) Proc of the Surface Characterization of Adsorption and Interfacial Reactions II. Keauhou-Kona, Hawaii, USA, p 121 83. Tokunaga J, Nobunaga T, Nakatani T, Iwasaki T, Fukuda K, Kunitake Y (1999) J Soc Naval Architects Jap 183:453 (in Japanese) 84. Fukuda K, Tokunaga J, Nobunaga T, Iwasaki T, Kunitake Y (1999) J Soc Naval Architects Jap 186:73

Hdb Env Chem Vol. 2, Part M (2005): 451–481 DOI 10.1007/b138191 © Springer-Verlag Berlin Heidelberg 2005 Published online: 6 September 2005

The Applications of Photocatalytic Waste Minimisation in Nuclear Fuel Processing Colin Boxall1 (u) · Gwénaëlle Le Gurun2 · Robin J. Taylor2 · Shaorong Xiao3 1 Centre

for Materials Science, University of Central Lancashire, Preston PR1 2HE, UK [email protected] 2 Nexia Solutions Ltd., BNFL Sellaﬁeld, Seascale CA20 1PG, UK 3 Department

of Physical Sciences, University of New Brunswick, St John, NB E2L 4L5,

Canada 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

452

2

Nuclear Fuel Processing: the PUREX Process . . . . . . . . . . . . . . . . .

454

3

Actinide Photochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . .

457

Photocatalytic Valence Control of Actinide Metal Ions I: the Twentieth Century . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Photocatalytic Valence Control of Neptunium . . . . . . . . . . . . . . . . 4.1.1 Neptunium Valence Control: Photochemistry and Heterogeneous Catalysis 4.1.2 Neptunium Valence Control: Heterogeneous Photocatalysis . . . . . . . . . 4.2 Photocatalytic Valence Control of Uranium . . . . . . . . . . . . . . . . . . 4.3 Photocatalytic Valence Control of Technetium . . . . . . . . . . . . . . . .

461 461 461 464 466 467

4

5

Photocatalytic Valence Control of Actinide Metal Ions II: Recent Developments . . . . . . . . . . . . . . . . . . . . . . . . Choice of Semiconductor Material for Actinide Valence Control Choice of Hole Scavenger for Actinide Valence Control . . . . . Photocatalytic Valence Control of a Pu Simulant . . . . . . . . . Photocatalytic Valence Control of Uranium . . . . . . . . . . . . Photocatalytic Valence Control in Mixed Solvent and Mixed Actinide Systems . . . . . . . . . .

. . . . .

467 467 470 470 472

. . . . . .

475

Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

477

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

478

5.1 5.2 5.3 5.4 5.5 6

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

Abstract Nuclear fuel processing has two main waste management requirements: (1) the disposal of waste organic solvent (secondary waste) generated by solvent extraction processes during the separation and puriﬁcation of uranium and plutonium in nuclear fuel and materials processing; and (2) the management of the small fractions of U and Pu that are inseparable during reprocessing (primary waste). Environmental impact associated with fuel use and reprocessing can be minimised by addressing either of these requirements. Semiconductor particles and ﬁlms may act as efﬁcient photocatalysts for a range of environmentally and industrially useful reactions including heavy metal recovery from efﬂuent streams by manipulation of the metal valence state. The manipulation of actinide

452

C. Boxall et al.

metal ion oxidation states plays an important role in nuclear fuel and materials processing. Thus, this review explores the potential use of heterogeneous photocatalysis in actinide valence state control in the context of actinide photochemistry and minimised primary and secondary waste management requirements in the plutonium–uranium reduction extraction (PUREX) nuclear fuel processing route. Criteria are deﬁned for the selection of heterogeneous semiconductor catalysts and sacriﬁcial charge scavengers for use within reprocessing scenarios and two main applications discussed: (1) the photocatalytic control of the neptunium ion oxidation state and consequent separation of Np from Pu and U; and (2) the photocatalytic control of U and Pu ion oxidation states and their consequent separation from each other. A quantum efﬁciency, φ, of 0.27 is reported for the photocatalytic reduction of the Pu(IV) simulant, Ce4+ to Ce3+ at pH 0. The high value of φ is attributed to both the forward and reverse charge transfer processes occurring via a dynamic quenching mechanism. Yields of 100% 4+ and Ce4+ to Ce3+ . are reported for the reductions of UO2+ 2 to U Keywords Actinides · Colloidal semiconductor · Nuclear fuel processing · Photocatalysis · Valence control Abbreviations e–CB Conduction band electron HA Highly active HAN Hydroxylamine nitrate HAW Highly active waste Valence band hole h+VB i.e.p. Isoelectric point MAW Medium active waste OK Odourless kerosene, n-decane or decane Ox Oxidised form of the standard system: Ox + ne– → Red Reduced form of Ox; equivalent to Red Ox– PUREX Plutonium–uranium reduction extraction Red Reduced form of the standard system: Ox + ne– → Red Red+ Oxidised form of Red; equivalent to Ox SCE Saturated calomel electrode TBP Tributyl phosphate

1 Introduction Semiconductor particles and ﬁlms have been found to act as heterogeneous photocatalysts in a number of environmentally important reactions. Materials such as TiO2 and CdS have been found to be efﬁcient in laboratoryscale pollution abatement systems, reducing both organic and inorganic pollutants/impurities to harmless species (see [1–35] and references therein). Photocatalysts have been shown to be useful for decomposition of O3 [36], destruction of bacteria [37–39] and viruses [40], puriﬁcation of air [41], photosplitting of water [42, 43] and clean-up of oil spills [44, 45]. Photocat-

The Applications of Photocatalytic Waste Minimisation

453

alytically driven oxidative and reductive manipulation of metal ion valence states (Cr(VI) [46], Cu(II) [47], Ni(II) [48]) has been studied extensively, most especially in the reductive deposition of heavy metals from a variety of industrial process and efﬂuent streams (e.g. Hg(II) [49, 50], Rh(III) [51], Pd(II) [52], Ag(I) [53], Pt(IV) [54], Cd(II) [55], Au(III) [56],and Cr(VI) [57]). The control of the actinide metal ion valence state plays a pivotal role in the separation and puriﬁcation of uranium and plutonium during the processing of spent nuclear fuel. Most commercial plants use the plutonium– uranium reduction extraction process (PUREX) [58], wherein spent fuel rods are initially dissolved in nitric acid. The dissolved U and Pu are subsequently extracted from the nitric solution into a non-aqueous phase of tributyl phosphate (TBP) dissolved in an inert hydrocarbon diluent such as dodecane or odourless kerosene (OK). The organic phase is then subjected to solvent extraction techniques to partition the U from the Pu, the extractability of the ions into the TBP/OK phase being strongly dependent upon the valence state of the actinide in question. Given the importance of the oxidation states of the actinides in reprocessing and the capability of heterogeneous semiconductor photocatalysts to manipulate the valence states of a wide range of metal ions, attention has been given to the potential applications of photocatalysis in nuclear reprocessing scenarios [59–65]. The attractions of this are twofold, and lead to a further minimisation of the environmental impact associated with fuel use and processing: 1. More effective control of the actinide valence state. This results in enhanced separation of U from Pu and consequently (a) more efﬁcient recycling and therefore use of fuel; and (b) a lower primary waste management requirement by decreasing the output of U and Pu to vitriﬁcation through the medium active waste (MAW) stream. The latter in turn leads to reduced requirements for special handling and long-term storage as well as their associated problems of accident/release and consequent effects on human health. 2. It is a salt-free process, thus offering simplicity and a lower requirement for use of chemical reductants with the cost beneﬁt of diminished reagent requirements and the environmental beneﬁt of less secondary waste. Many of the solvents used are organophosphate-based and so can give rise to the same generic environmental concerns (especially toxicity to humans and mammals) associated with the analogous pesticides. It is therefore the purpose of this chapter to review applications of photocatalysis in actinide ion valence control and nuclear fuel and actinide materials processing. The review will proceed as follows. As the primary processes of semiconductor photocatalysts and the applications of photocatalysis in the manipulation of the valence states of (predominantly transition) metal ions are

454

C. Boxall et al.

reviewed elsewhere in this volume ([34, 35] respectively), the chapter begins with a description of the radioactive waste reprocessing and spent fuel recycling that occurs during the PUREX process. This will be followed by a brief discussion of relevant actinide solution photochemistry. Applications of heterogeneous photocatalysis in actinide photochemistry will then be reviewed and the chapter will conclude with a discussion of results obtained in our laboratories from experiments on actinide valence control through heterogeneous semiconductor photocatalysis.

2 Nuclear Fuel Processing: the PUREX Process The principal objectives of reprocessing spent nuclear fuel from a nuclear reactor are [66] (1) the recovery of U, Pu and thorium, if present, for reuse as nuclear fuels, (2) the removal of radioactive, neutron-absorbing ﬁssion products and minor actinides from fuels, neptunium being one of the most prevalent and difﬁcult to control, and ﬁnally (3) the reconversion of the radioactive elements included in the spent fuel into safe forms suitable for long-term storage. The PUREX process is an effective method of reprocessing spent fuel and is the basis for all currently operating industrial-scale plants. It uses a series of solvent extractions and reductions, shown schematically in Fig. 1. However, signiﬁcant improvements are still under investigation, particularly in the adaptation of PUREX technology to meet the needs of advanced fuel cycles. As a typical example, British Nuclear Fuels plc Thermal Oxide Reprocessing (BNFL THORP) [67], based in Sellaﬁeld, UK, uses PUREX. After mechanical handling (e.g. dismantling and shearing), the irradiated nuclear fuel is dissolved in strong reﬂuxing HNO3 in the “head end” plant. After dissolution, the liquor is stored in a holding tank before being treated by the PUREX process. This storage period may affect the speciation of the actinides and ﬁssion products through radiolysis and other reactions, e.g. disproportionation of Np(V) to Np(IV) and NP(VI) which may then be undesirably extracted into the TBP/OK solvent used in the PUREX process (vide infra). TBP is used in the separation process because of its good radiolytic and chemical stabilities, low aqueous solubility and its chelating properties. As the density of TBP (0.98 kg dm–3 ) is so close to that of water, dilution with a non-polar, low density diluent, OK, is necessary to achieve phase separation [58]. The salting agent used is 3 M nitric acid, chosen because it facilitates the formation of weak, neutral complexes between the actinides and the organic solvent, so giving the greatest extraction efﬁciency. After dissolution of the chopped irradiated fuel in boiling 3 M HNO3 , the resultant mixture contains mainly UO2 2+ , Pu4+ and a small amount of PuO2 2+ that are extractable into the organic phase, ﬁssion products including inex-

The Applications of Photocatalytic Waste Minimisation

455

Fig. 1 Schematic diagram of the PUREX process. TBP Tributyl phosphate, OK odourless kerosene, HA highly active

tractable Np(V) and extractable Np(VI), and minor actinides. Extractable Np(IV) may also be present after storage due to the disproportionation reaction of Np(V) according to: 2NpO2 + + 4H+ ↔ Np4+ + NpO2 2+ + 2H2 O

(1)

After storage, the solution is ﬁltered and subjected to further puriﬁcation before being fed to the highly active (HA) cycle of the PUREX process, shown in Fig. 1. In the ﬁrst of a series of contactors in the HA cycle, the uranyl and plutonium nitrates are extracted into the organic solvent, 30% TBP/diluent, leaving the ﬁssion products in the aqueous phase which is directed to highly active waste (HAW) for further treatment. The Np that exists as NpO2 + will also remain in the aqueous phase. Conversely, if it exists as Np4+ or NpO2 2+ ions, then it is extractable into the organic phase and it will be co-extracted together with the Pu and U. The disproportionation reaction represented by Eq. 1 therefore presents a problem: the greater the rate of disproportionation of Np(V) to Np(IV) and Np(VI), the more Np will be extracted with the U and Pu, so requiring an additional puriﬁcation stage later in the cycle. In addition to the disproportionation of Np(V), a greater problem for the control of the valence of Np ion species arises from the dispersion of nitrous acid throughout the system. This is constantly formed through the radiolysis of HNO3 depending on radiation level, acidity and temperature [68]. The presence of nitrous acid is a problem as it catalyses the oxidation of Np(V) by

456

C. Boxall et al.

nitric acid [69, 70]: HNO2

2NpO2 + + NO3 – + 3H+ ––––––→ 2NpO2 2+ + HNO2 + H2 O

(2)

Nitrous acid also oxidises Pu and U. Thus, in order to obviate these difﬁculties, hydrazine is added as a scavenger for nitrous acid: N2 H5 + + HNO2 → HN3 + 2H2 O + H+ N2 H5 + + 2HNO2 → N2 + N2 O + 3H2 O + H+

(3) (4)

In excess hydrazine, the reaction in Eq. 3 occurs, producing hydrazoic acid, HN3 . In excess HNO2 , the reaction in Eq. 4 occurs. Hydrazoic acid is known to react vigorously with HNO2 and has been suggested as an intermediate in Eq. 5 [71]: HN3 + HNO2 → N2 + N2 O + H2 O

(5)

The formation of HN3 is undesirable as it is explosive and toxic. Goldacker reported [66] that a part of the small amount of HN3 produced is extracted in the aqueous phase, the remainder being removed by ventilation. However, these difﬁculties do not present intractable problems and so hydrazine is added to scavenge HNO2 and prevent, inter alia, the nitrous-acid-driven re-oxidation of Pu(III) to Pu(IV) [72]. Upon emerging from the ﬁrst contactor, the organic solvent containing the uranyl, Pu(IV) and neptunyl nitrates is injected into a second contactor, where U–Pu partitioning occurs: the “U/Pu split”. At THORP, for instance, Pu(IV) is reduced to Pu(III) using an aqueous solution of U(IV), which can be catalytically generated from U(VI). This method does not add any extra salts to the process, so does not increase waste [72]. U(IV) itself is oxidised to U(VI), which is re-extracted into the organic phase, leaving the Pu(III) in the aqueous phase [73]: U4+ + 2Pu4+ + 2H2 O → UO2 2+ + 2Pu3+ + 4H+

(6)

Some Pu(III) may be extracted into the organic phase which can act as a catalyst for the oxidation of U(IV), consuming the reducing agent [74]: 2Pu(NO3 )3 + 3HNO3 ↔ 2Pu(NO3 )4 + HNO2 + H2 O 2Pu4+ + U4+ + 2H2 O → 2Pu3+ + UO2 2+ + 4H+

(7) (8)

Thus, a large excess of U(IV) is required to reduce Pu(IV), diluting the enrichment of the recovered U. Further, the reactions described by Eqs. 7 and 8 will reduce the efﬁciency of the separation process and may ultimately lead to incomplete separation of U and Pu due to lack of reducing agent, so increasing the cost of the procedure. Alternative methods to reduce Pu include the use of hydroxylamine nitrate with hydrazine as stabiliser or the electroreduction of Pu(III) [66] in situ. Nonetheless, the current use of U(IV) as a reductant for

The Applications of Photocatalytic Waste Minimisation

457

Pu(III) and the dispersion of Np oxidation states throughout the process remain areas where potential improvements could be made to the HA stage of PUREX.

3 Actinide Photochemistry Photochemical manipulation of actinide valence states has been studied as a means towards a salt-free method for actinide separation during fuel processing. This section summarises the points of relevance to photocatalytic valence manipulation and places them in the context of the dark redox reactions of actinide ions in aqueous solution. The PUREX process exploits two features of U chemistry: (1) the UO2 2+ ion is the thermodynamically most stable form of U in aqueous solution; both PuO2 2+ and NpO2 2+ are easily reduced to Pu4+ and NpO2 + under similar conditions (vide infra); and (2) in general, the actinide MO2 2+ ions can be extracted from nitrate solutions into non-polar organic solvents [75] such as the phosphate esters, e.g. TBP. Since most other metal ions are not extracted under similar conditions, solvent extraction provides a convenient route for the puriﬁcation of U and Pu from practically all other metals. Np can also be rendered extractable by manipulation of its oxidation state. Similarly, U can be separated from Pu by the selective reduction of Pu(IV) to Pu(III), rendering it inextractable into TBP/OK. The dark redox chemistries and the associated thermodynamics of valence state interconversion of the actinides are most easily understood through the use of potential–pH diagrams. Figs. 2–4 show such diagrams for the U-H2 O, Np-H2O and Pu-H2 O systems respectively. Toth et al. [76] and Yusov & Shilov [77] have reviewed the photochemically induced redox reactions of U, Np and Pu. The UV–visible absorption spectrum of UO2 2+ is characterised by a peak with a band maximum at 225–245 nm and a less intense, inverted-rubber-glove-shaped feature at 400–480 nm. Photo-excitation is most commonly achieved through the second of these adsorption bands, the lowest excited state of ∗ UO2 2+ emitting photons in the wavelength range 460–600 nm [78]. UO2 + has several bands in the near IR at 770, 970 and 1420 nm [79], and evidence exists to suggest the presence of an absorption band in the range 366–388 nm in the near UV [80]. U4+ exhibits a number of sharp bands in the UV–visible at 426, 492, 548 and 646 nm, the last being the most intense. The main bands associated with PuO2 2+ occur at 350 and 830 nm. Pu4+ exhibits a number of bands in the UV–visible at ∼ 400, 425, 480 (sharpest and most intense), 545 and 640–660 nm. The primary Pu3+ band occurs at 563 nm [81]. NpO2 2+ has two bands in the UV at ∼ 350 and 400 nm and one in the near IR at 1225 nm. NpO2 + has a strong band in the near IR at 980 nm

458

C. Boxall et al.

Fig. 2 Potential–pH diagram of the uranium–water system at 298 K with SnO2 conduction band edge overlaid. Dissolved uranium activity= 0.01SCE Saturated calomel electrode

and weaker bands at 620 and 1100 nm, while the main Np4+ bands occur at 500, 722 and 960 nm [81]. Excited uranyl ion, ∗ UO2 2+ , is a strongly oxidising species (Eθ = + 2.35 ± 0.1 V vs. SCE [82]) and has been shown to be capable of oxidising a variety of substrates [78]. In the absence of any deliberately added solution phase organic compound, it has been suggested that UV photo-excitation of UO2 2+ in aqueous sulphuric acid solution initiates the following sequence of reactions to produce the observed primary products of UO2 + and OH· [83, 84]: hν

UO2 2+ –→ ∗ UO2 2+ ∗ UO 2+ + H O → UO + + H+ + OH· 2 2 2 · 2OH +→ H2 O2 2UO2 + + 4H+ → UO2 2+ + U4+ + 2H2 O

(9) (10) (11) (12)

Eq. 12 being the disproportionation of U(V). Alternatives to Eq. 10, involving the transient formation of UO2 (OH)3 – have been suggested by Bouby et al. [85]. In contrast to UO2 2+ , PuO2 2+ is inert under UV illumination in aqueous perchloric (> 0.5 mol dm–3 ), hydrochloric and sulphuric acids [77], despite ∗ PuO 2+ having an extrapolated Eθ of ∼ + 4.32 V vs. SCE [81]. However, at 2 [HClO4 ] < 10–2 mol dm–3 , Pu(VI) is converted to Pu(V) [77]. In nitric acid,

The Applications of Photocatalytic Waste Minimisation

459

Fig. 3 Potential–pH diagram of the neptunium–water system at 298 K with SnO2 conduction band edge overlaid. Dissolved neptunium activity= 0.01

Fig. 4 Potential–pH diagram of the plutonium–water system at 298 K with SnO2 conduction band edge overlaid. Dissolved plutonium activity= 0.01

UV-illuminated NpO2 2+ is converted to Np(V), although this is primarily as a result of the photolysis of the nitrate anion (see Sect. 4.1.1). In perchloric acid, UV-illuminated neptunyl is reduced to Np(V); in this case, it has been assumed that the products of water photolysis act as a reductant [86].

460

C. Boxall et al.

In UV-illuminated aqueous H2 SO4 solutions, both Np(IV) and Np(V) are quantitatively oxidised to Np(VI) [77], implying that, like PuO2 2+ , NpO2 2+ is intrinsically inert in perchloric and sulphuric acids, despite ∗ NpO2 2+ having an extrapolated Eθ of ∼ + 4.01 V vs. SCE [81]. Both U(VI) and Pu(VI) react with added reductant according to: hv

MO2 2+ + Red –→ MO2 + + Red+

(13)

where Red is an alcohol, an aldehyde, hydroxylamine or hydrazine. The MO2 + species disproportionates to M4+ and MO2 2+ in accordance with Eq. 12 with a rate dependant upon H+ and M(V) concentration. Oxygen is found to decrease the ultimate rate of U(IV) generation in the UO2 2+ /ethanol system, although this is thought to be a result of the reaction of O2 with the initial oxidation product of ethanol, the α-hydroxylethyl radical, which is itself a powerful reducing agent for U(VI) ions [87]. In the presence of ethanol, neptunyl is reduced to NpO2 + in accordance with Eq. 13, although this is only the ﬁrst reaction in a series of steps by which Np(VI) is photochemically reduced to Np(III). The quantum efﬁciency for the reaction described by Eq. 13 in 1 M acid with ethanol present is 50 times greater for UO2 2+ (irradiated in the wavelength range 250–600 nm) than for PuO2 2+ (irradiated at wavelengths < 350 nm). Direct photochemical reduction of Pu(IV) also occurs in the same system: hν

Pu4+ + Red –→ Pu3+ + Red+

(14)

In the absence of any deliberately added reductant, Pu(IV) remains unchanged under UV irradiation in aerated or de-aerated aqueous carbonate solutions. However, in nitric acid media, both Pu(III) and Pu(IV) are transformed to Pu(VI), although both oxidations are almost certainly mediated by disproportionation of Pu(IV) to Pu(III) and Pu(VI): 3Pu4+ + 2H2 O → 2Pu3+ + PuO2 2+ + 4H+

(15)

Similar results are observed in perchloric acid in the absence of any deliberately added reductant, the rate of disproportionation being increased considerably by UV irradiation [88]. Pu(III) is also converted to Pu(IV) by UV irradiation in 0.5–5.5 mol dm–3 HCl, the reaction occurring in both air- and CO2 -purged solutions, the process being somewhat slower in the latter [89]. Behaviour of Pu in sulphuric solutions is the same as that in hydrochloric. The III, IV and V states of Np can all be oxidised in sequential oneelectron transfer reactions to form Np(VI) when illuminated with UV light in HClO4 , HCl and H2 SO4 solutions. At [HCl] > 3 mol dm–3 and [H2 SO4 ] = 1–3 mol dm–3 , the oxidations of Np(IV) and Np(V) respectively are quantitative [90, 91]. In the presence of a deliberately added reductant such as ethanol, aldehyde etc., photochemical reduction of Np(VI), Np(V) and Np(IV) ions

The Applications of Photocatalytic Waste Minimisation

461

can occur in perchloric, sulphuric and hydrochloric acid solutions, long-term illumination resulting in the conversion of all Np into the IV state [92]. As mentioned above, in HNO3 solutions, the photochemistry of Np is complicated by the photolysis of the nitrate anion. As a result, Np(VI) and Np(IV) solutions can be completely converted to Np(V). This, and its utility in the separation of Np from Pu and U in acid solution, will be discussed in more detail in the next section.

4 Photocatalytic Valence Control of Actinide Metal Ions I: the Twentieth Century In summary, potential improvements could be made to the PUREX process in the following areas: (1) separation of Np from U and Pu prior to the U/Pu split; and (2) in the requirement to use a large excess of U(IV) reductant to reduce Pu(IV) to Pu(III). The majority of published work on the applications of photocatalysis in actinide redox chemistry has concentrated on solving the ﬁrst of these difﬁculties through Np valence control. A smaller volume of literature exists on the applications of photocatalysis in valence state control of U and the radioactive d block metal, technetium. This section will review both of these aspects. 4.1 Photocatalytic Valence Control of Neptunium 4.1.1 Neptunium Valence Control: Photochemistry and Heterogeneous Catalysis Ideally, during the HA cycle of PUREX, all Np should be directed into one of three streams: (1) the aqueous HAW stream with the ﬁssion products, pre U/Pu split; (2) the aqueous Pu stream at the U/Pu split; or (3) the nonaqueous U stream at the U/Pu split. The ﬁrst two require all Np to be in the V oxidation state, and the last that all Np be in the IV or VI state at those points in PUREX. Two reactions mitigate against such control of Np: disproportionation of Np(V) via Eq. 1 to form Np(IV) and Np(VI); and the nitrous acid (derived from the radiolysis of HNO3 )-promoted equilibrium between Np(VI) and Np(V): 2NpO2 2+ + HNO2 + H2 O ⇔ NO3 – + 2NpO2 + + 3H+

(16)

Between them, these two reactions give rise to a dispersion of Np oxidation states (IV, V & VI) throughout the PUREX process. In HNO3 solutions, the redox photochemistry of Np is further complicated by the photolysis of the nitrate anion. At [HNO3 ] < 1 mol dm–3 , and in the ab-

462

C. Boxall et al.

sence of any deliberately added reductant, irradiation of Np(IV) or Np(VI) solutions with UV light results in the accumulation of Np(V) in solution, which can be considered to proceed according to [68, 93]: hv HNO3 –→ OH· + NO2 2NO2 + H2 O → HNO3 + HNO2 Np4+ + OH· + H2 O → NpO2 + + 3H+ NpO2 + + OH· + H+ → NpO2 2+ + H2 O

(17) (18) (19) (20)

followed by the reaction in Eq. 16. An alternative mechanism suggests that the products of the photolysis of HNO3 are NO2 – and · O· and it is these species that participate in the Np reduction and oxidation processes of Eqs. 16 and 19 [93]. At [HNO3 ] = 2 or 4 mol dm–3 , again in the absence of any reductant, irradiation results in an increase in the Np(V) concentration in the ﬁrst 10 min, after which it decreases in favour of Np(VI) formation. This has been explained in terms of the reversibility of Eq. 16 [94] i.e. Eqs. 17–19 are now followed by the acid-promoted, (photolytically generated) nitrousacid-catalysed reaction (Eq. 2). This explanation is in keeping with results obtained in the dark under conditions typical of the PUREX process, i.e. at 3–6 mol dm–3 HNO3 , wherein radiolytically generated HNO2 catalyses the oxidation of Np(V) by nitric acid (Eq. 16) [70]. Under similar conditions to those under which Np(V) is photogenerated from Np(IV) and/or Np(VI) (Eqs. 16–20), Pu(III), (IV) & (VI) are photochemically converted to Pu(IV) and (VI) while all U species are ultimately converted to UO2 2+ . Pu(IV), Pu(VI) and U(VI) are extractable into TBP while Np(V) is not; thus, this suite of reactions may form the basis of a process for Np separation. Within an advanced PUREX ﬂowsheet, this would ﬁnd utility at the point of ﬁssion product removal, the Np being routed with the HAW stream. An alternative strategy for Np separation would be to co-extract it as Np(IV) and/or Np(VI) with UO2 2+ at the U/Pu split and perform a U/Np split further downstream. Photochemical valence adjustment of Np into its IV and VI oxidation states in nitric acid has been demonstrated, both processes requiring the introduction of additives to the reaction mixture. The addition of urea to UV-irradiated Np/HNO3 solutions promotes conversion to Np(VI) [68]. In this instance, Np(VI) photogeneration is commonly thought to occur by Eqs. 17–20 with the reaction of urea with HNO2 , shown in Eq. 21, preventing the reduction of Np(VI) via Eq. 16. (NH2 )2 CO + 2HNO2 → 2N2 + CO2 + 3H2 O

(21)

Interestingly, when the acid concentration is decreased in the range 1–0.05 mol dm–3 (in the presence of 1 mol NO3 – dm–3 ), the ﬁnal concentration of Np(VI) decreases in favour of Np(V). Shilov and Yusov [95] explain this by noting that NpO2 2+ has two absorption bands in the UV at 350 and 400 nm.

The Applications of Photocatalytic Waste Minimisation

463

They suggest that NpO2 2+ , photogenerated in accordance with Eqs. 17–20, is itself photoexcited and reacts, via excimer formation, according to a net reaction given by Eq. 22. This is promoted at low [H+ ], hence the decrease in Np(VI) yield at higher pH. hν

2NpO2 2+ + 2H2 O –→ 2NpO2 + + H2 O2 + 2H+

(22)

At low pH, Pu(III) and Pu(IV) are photo-oxidised to Pu(VI) in UV-irradiated urea/nitric acid solutions while Np(VI) is generated from Np species under identical conditions. As both Pu(VI) and Np(VI) are extractable into TBP/diluent, UV treatment in the presence of urea could potentially be used for the co-extraction of Np and Pu in a revised PUREX ﬂowsheet [96]. Hydrazine and hydroxylamine nitrate (HAN) are two other common additives for revised PUREX ﬂowsheets. In the presence of hydrazine/HAN, Pu(III) is converted to Pu(IV) in UV-irradiated nitric acid solutions. In the dark, Np(V) is reduced to Np(IV) by both HAN and hydrazine, but the rates are slow [97, 98]: 2NpO2 + + 4H+ + 2NH3 OH+ → 2Np4+ + N2 + 6H2 O 2NpO2 + + 8H+ + 2N2 H4 → 2Np4+ + N2 + 4H2 O + 2NH4 +

(23) (24)

Under UV irradiation, Np(IV) is oxidised to Np(V) (e.g. Eqs. 17–19) and subsequently to Np(VI) (e.g. Eqs. 2 and 20). Hydrazine is also a scavenger for nitrous acid (Eqs. 3 and 4), suggesting that reduction of Np(VI) back to Np(V) via Eq. 16 is unlikely to occur and that Np(VI) would be expected as the ultimate product of the UV irradiation of a HAN/hydrazine-containing solution of Np in low pH nitric acid. However, in a study of the photo-induced valence adjustments that occur in mixtures of Pu(III) and Np(V), Wada et al. found that, in HAN/hydrazine-containing nitric acid solutions, Pu(III) was oxidised to Pu(IV) while the Np(V) was unaffected [81, 96, 99]. Whilst this is useful from the point of view of developing a means to separate Np and Pu, it is counter to the expectation that Np(VI) should be the ultimate product of UV-induced valence adjustment in HAN/hydrazinecontaining nitric acid solutions. The results obtained could be explained by involvement of the reaction described by Eq. 22 but for the fact that the experiments of Wada et al. were conducted in 3 M HNO3 , thus rendering the reaction in Eq. 22 unlikely to occur. It is more likely that any Np(VI) generated is reduced back to Np(V) by Pu(III), either directly, so producing Pu(IV), or by the products of the photo-oxidation of Pu(III) to Pu(IV) by nitric acid. Having reviewed photochemical valence adjustment of Np in nitric acid solution, we will brieﬂy consider the effect of heterogeneous Pt particle catalysts on Np speciation reactions in the dark before discussing applications of heterogeneous photocatalysis in Np redox chemistry.

464

C. Boxall et al.

Catalytic reduction of Np(V) to Np(IV) has been reported by Nakamura et al. [100]. The reaction was conducted in 3 mol dm–3 HNO3 using platinum black as a catalyst and HAN as reducing agent. Without a catalyst, the reduction described by Eq. 23 occurs slowly, even at elevated temperatures between 333 and 363 K. However, in the presence of a Pt catalyst, 0.625 mol dm–3 HAN converted 6 × 10–3 mol dm–3 Np(V) to Np(IV) in < 30 min. Pt and other Pt group elements are known to act as catalysts for H atom formation. Nakamura et al. observed that Np(V) reduction did not occur in the absence of HAN. Thus, they have suggested that the reduction of Np occurs as a result of the generation of H atoms via the oxidation of HAN on the Pt black surface: NH3 OH+ –→ NHOH + 2H+ + e– H+ + e– –→ H (H atom adsorbed at Pt surface) 2NHOH –→ (NHOH)2 –→ N2 + 2H2 O NpO2 + + H –→ NpO(OH)+ NpO(OH)+ + 3H+ –→ Np4+ + 2H2 O

(25) (26) (27) (28) (29)

the overall reaction being written as: Pt

NpO2 + + 2NH3 OH+ + 4H+ –→ 2Np4+ + N2 + 6H2 O

(30)

Interestingly, while the rate of the Pt-catalysed Eq. 30 was found to be signiﬁcantly higher than the uncatalysed dark reaction of Eq. 23, the rate was found to be higher still in simulated high-level waste solutions than in pure nitric acid. Nakamura et al. note that under conditions where Np is reduced to TBP-extractable Np(IV), thermodynamic calculations indicate that Pu is reduced to inextractable Pu(III) while U exists as either U(IV) or U(VI), both of which are extractable by TBP. Thus, the reaction described by Eq. 30 offers a means by which all Np can be directed with the U stream at the U/Pu split (Fig. 1). Homogeneous photochemical and heterogeneous catalytic processes are attractive because they do not increase the volume of secondary waste. Heterogeneous photocatalysis offers the further advantage that, post valence control, any additives such as HAN can be photocatalytically degraded to readily disposable gases such as N2 and CO2 . This will be explored further in the next section. 4.1.2 Neptunium Valence Control: Heterogeneous Photocatalysis In common with other applications of photocatalysis (degradation of gas and solution phase pollutants, super hydrophilic coatings, metal recovery etc.), the material that has received the most attention in studies of heterogeneous pho-

The Applications of Photocatalytic Waste Minimisation

465

tocatalytic control of actinide oxidation states is titanium dioxide. In a study by Bondietti and Trabalka, particulate TiO2 was used as the principal component of a process to isolate PuO2 + [101]. However, its mode of action in this study was as an adsorbent. The actinide metal to which the photocatalytic, as well as surface chemical, properties of TiO2 have most frequently been applied is Np, and speciﬁcally in the oxidation of Np(V) to Np(IV). Photocatalysis has also been applied to the manipulation of the oxidation state of the radioactive d-block metal technetium and this will be discussed in Sect. 4.3. Fukasawa et al. have examined the effect of TiO2 photocatalysis on the oxidation of Np(V) to Np(VI) [60, 63–65]. Studies were conducted in 3 mol dm–3 nitric acid in both the absence and presence of urea as a scavenger for nitrous acid (Eq. 21), HNO2 being generated as a result of the photolysis of HNO3 (Eqs. 17 and 18). As discussed in Sect. 4.1.1, UV-irradiation of Np/HNO3 solutions in the absence of a scavenger for HNO2 can result in the quantitative conversion of all solution phase Np to Np(V). Fukasawa et al. report that photocatalytic TiO2 has no effect on either the rate or yield of this process. However, the same is not true of the process that occurs in the presence of a scavenger for nitrous acid. As discussed in Sect. 4.1.1, UV-irradiation of Np/HNO3 solutions in the presence of urea can result in the quantitative conversion of solution phase Np to Np(VI). Fukasawa et al. report that both the rate and the yield of the reaction are substantially increased when the system is irradiated in the presence of platinised TiO2 (10% Pt, 90% TiO2 ). They suggest the following mechanism operates in parallel with the reactions given by Eqs. 16–21: hν

Photocatalyst –→ h+VB + e–CB 2e–CB + NO3 – + 3H+ –→ HNO2 + H2 O h+VB + NpO2 + –→ NpO2 2+

(31) (32) (33)

hν From Eqs. 17 & 18: HNO3 –→ HNO2 + OH· From Eq. 20: OH· /Ox + NpO2 + –→ NpO2 2+

(34) (35)

where HNO2 is scavenged by urea through Eq. 21. OH· is generated as a result of Eq. 34 or the action of photogenerated holes on surface adsorbed water/hydroxyl groups, while Ox represents the product of the action of OH· on (adventitious) reductants in solution. Np(VI) is only generated by direct valence band hole transfer (Eq. 33) if the reorganisation energy of NpO2 + permits. Fukasawa et al. also explored the effect of semiconductor photocatalysis on the reduction of Np(V) to Np(IV) by hydrazine in analogy to Eq. 24 [64]. In this case, they employed platinised SiC as a photocatalyst (10% Pt, 90% SiC) due to the greater reducing power of the conduction band electrons of SiC than TiO2 . They found that the rate and yield of the reaction were substantially enhanced and suggested the following mechanism: The reaction in

466

C. Boxall et al.

Eq. 31 is followed by 2h+VB + N2 H4 –→ N2 H2 + 2H+ e–CB + H+ –→ H

(36) (37)

where H again represents a H atom adsorbed at the Pt surface. Np(IV) is then produced through Eqs. 28 and 29. 4.2 Photocatalytic Valence Control of Uranium Photodeposition of metals onto semiconductor colloidal suspensions has been explored in the recovery of precious metals from industrial process or efﬂuent liquors [7]. Using illuminated particulate TiO2 in the presence of hole scavengers (propan-2-ol, ethanoate, methanoate) Amadelli et al. [61] and Chen et al. [102] have applied photocatalytic deposition to the separation of U species from aqueous media for, inter alia, environmental, analytical and wastewater remediation applications. From a study of adsorption phenomena in the dark, both groups found that the amount of UO2 2+ , and the hydrolysis products thereof, adsorbed onto the colloid surface depends on solution pH: as pH increases toward the isoelectric point (i.e.p.) of TiO2 , the positively charged U species are more strongly adsorbed onto the particle surface. Complexation of uranyl ions with different hole scavengers was also found to play an important role in the photoredox process. For instance, surface adsorption is enhanced by the presence of carboxylic acid salts at pH
The Applications of Photocatalytic Waste Minimisation

467

tially present [102]) can be photodeposited as the oxide, so allowing for the recovery of this potentially harmful metal from waste solutions. 4.3 Photocatalytic Valence Control of Technetium All isotopes of Tc are unstable toward β decay or electron capture, and traces exist in nature only as fragments from the spontaneous ﬁssion of U. Thus, while it is not a member of the actinide series, it is radioactive and, therefore, the role of photocatalysis in control of its valence state will be brieﬂy considered here. The use of photocatalysis for the reduction of pertechnetate (TcO4 – ) to Tc metal has been discussed by Fukasawa et al. [59]. They suggest that when Tc-containing acidic liquid radioactive waste is irradiated by UV light in the presence of platinised TiO2 and a suitable hole scavenger, the photogenerated conduction band electrons directly/indirectly reduce the pertechnetate ions to Tc metal at the platinum surface. Suggested hole scavengers include ethanol, ammonium carbonate and combinations thereof. Upon completion of the reaction, the Tc-coated TiO2 may be removed from the liquid waste by ﬁltration, so converting a liquid waste management problem to a solid one. Whilst the primary source of illumination discussed is a xenon lamp, the authors also suggest that photocatalyst excitation could be achieved using other sources of electromagnetic waves such as radiation, although no results are given to support this assertion.

5 Photocatalytic Valence Control of Actinide Metal Ions II: Recent Developments The use of photocatalysis as a means of controlling the valence state of Np, and so facilitate its separation from U and Pu immediately before or after the U/Pu split has been the focus of a number of publications in the open and patent literature [60, 63–65]. The use of photocatalysis to either generate or replace the U(IV) reductant used at the U/Pu split has received little, if any, attention. It is our work on the U and Pu systems, and the non-radioactive analogues thereof, which forms the focus of this section. 5.1 Choice of Semiconductor Material for Actinide Valence Control In choosing suitable particulate semiconductor photocatalysts for use in nuclear fuel processing scenarios, it is necessary that certain performance criteria should be fulﬁlled. Suitable materials should have the following properties:

468

C. Boxall et al.

1. Chemical resistance to all chemical and radiolytic environments encountered in the reprocessing of nuclear fuels, most especially aqueous environments of high acidity (pH< 2); 2. Conduction band thermodynamics that have been either identiﬁed or tailored through deliberate synthesis as being appropriate to perform the target reduction; and 3. It should be photostable (i.e. not liable to photoanodic or photocathodic corrosion), environmentally benign and, preferably, cheap. The redox potential(s) of the photogenerated valence/conduction band holes/electrons must be sufﬁciently positive/negative to perform the desired oxidation/reduction. Different semiconductors exhibit different band edge energetics; thus, matching the semiconductor energetics to the thermodynamics of the desired reaction—e.g. U(VI) to U(IV) or Pu(IV) to Pu(III)—is crucial for overall process efﬁciency. Control of the particle valence/conduction band oxidation/reduction potential is not only achieved through a judicious choice of particle component material; band edge redox thermodynamics of a single material are also affected by solution pH, semiconductor doping level and particle size. The relevant properties of the actinide metal are its range of available valence states and, for aqueous systems, the pH dependence of the thermodynamics of inter-valence conversion. Consequently, any study of semiconductor-particleinduced valence control has to be conducted in close consultation with the thermodynamic potential–pH speciation diagrams of both the targeted actinide metal ion system and the semiconductor material. By following such a procedure, the semiconductor that is most commonly used for photocatalytic applications, TiO2 , must be excluded. TiO2 has a band gap of 390 nm or 3.2 eV [103]. Thermodynamic calculations indicate that TiO2 should dissolve to form TiO2+ in pH< 0 acid solutions in the dark [103]. Further, calculations [104] indicate that when TiO2 is subjected to ultraband gap illumination, the photogenerated conduction band electrons are energetic enough to reduce TiO2 to Ti3+ at pH< 0, while the concomitantly produced valence band holes may oxidise TiO2 to TiO2 2+ at pH< 1. Consequently, illuminated TiO2 may be expected to undergo photodissolution at pH< 1, rendering it unsuitable for use as a valence control photocatalyst in the majority of solution environments typical of those found in most reprocessing streams. Similar arguments can be used to discount ZrO2 , HfO2 , Si, SiC, Ge, ZnO, PbO2 and CdS. However, inspection of the Sn – H2 O potential–pH diagram indicates that SnO2 may ﬁnd wide utility as a photocatalyst in actinide valence control. SnO2 is used in white pigmented paint coatings [105]. Stucki and coworkers have noted that it is an efﬁcient photocatalyst for the oxidation of a wide range of organics [106]. It has a band gap of approximately

The Applications of Photocatalytic Waste Minimisation

469

350–360 nm or 3.5 eV and crystallises with the rutile structure [107, 108]. Inspection of the potential–pH diagram for the tin–water system at 298 K [109] indicates that tin oxide is thermodynamically stable towards acid dissolution under oxidising conditions at low pH. Further, band edge calculations indicate that both the valence and conduction band edges of tin oxide fall comfortably within the area of SnO2 stability at pH 2 to – 2, indicating that the material will not photocorrode upon illumination with ultra-band gap energy photons at pH < 2. These theoretical predictions have since been conﬁrmed by experiments conducted within these laboratories [110]. The identiﬁcation of SnO2 as a candidate for valence control applications within a modiﬁed PUREX process prompted an investigation of relative positions of the tin oxide band edges with respect to areas of stability on the U-H2 O, Np-H2 O and Pu-H2 O system potential–pH diagrams (see Figs. 2–4). This allowed identiﬁcation of a pH range where illuminated colloidal SnO2 could facilitate the aqueous phase separation of Np from Pu species before the primary TBP/diluent solvent extraction step of the HA cycle during a modiﬁed PUREX process. As in the work of Fukasawa et al. [60, 63–65], described in Sect. 4.1.2, upon illumination of the particles with light of ultra-band gap energy, the primary process is the generation of conduction band electrons and valence band holes, (Eq. 31). For a SnO2 suspension containing Np and HNO3 and no scavenger for nitrous acid, it is expected that this will be followed by the reactions described by Eqs. 16 and 32–35, i.e. hole-driven oxidation to Np(VI) followed by nitrous-acid-driven reduction back to Np(V). Therefore, the overall process as described by Fukasawa et al. for TiO2 in the absence of a scavenger for HNO2 represents a short circuit if the initial oxidation state of Np was Np(V). However, inspection of Fig. 3 indicates that the photogenerated conduction band electrons on SnO2 may be utilised to produce a useful net reaction, speciﬁcally, the reduction of NpO2 + at pH > – 1 to produce insoluble NpO2 : e–CB

NpO2 + –––→ NpO2

(38)

Examination of the potential–pH diagram for the Pu–water system (Fig. 4) indicates that the products of Pu photolysis reactions (analogous to those described above for Np) are the solution species PuO2 2+ and Pu3+ in the pH range – 1 to 2. Examination of the U–water potential–pH diagram (Fig. 2) indicates that the products of analogous U photolysis under the same solution conditions are solution phase U4+ and UO2 2+ , the latter produced through hole-driven processes (in analogy to Eqs. 33 and 35), the former through direct reduction by conduction band electrons in accordance with: e–CB

UO2 2+ + 4H+ –––→ U4+ + 2H2 O

(39)

Thus, at pH – 1 to 0 (typical of those values found in reprocessing liquors), thermodynamic arguments indicate that semiconductor-particle-driven pho-

470

C. Boxall et al.

tolysis of U, Pu and Np species results in the generation of solution phase U and Pu species, and the precipitation of Np species. It is expected that this strategy can be facilitated by the presence of an appropriate hole scavenger (e.g. ethanol, methanol). Through a competition reaction, the scavenger will disable the hole-driven oxidation processes to Np(VI), U(VI) and Pu(VI), so further limiting the products of the photocatalysis to Pu(III), U(IV) and insoluble Np(IV). Choice of hole scavenger will be the subject of the next section. 5.2 Choice of Hole Scavenger for Actinide Valence Control The electron donor/hole scavenger to be used in actinide valence control must have two properties. The ﬁrst is that it is capable of undergoing an irreversible oxidation, either by valence holes formed by the photo-excitation of the semiconductor photocatalyst or by means of the hydroxide radicals generated by the oxidation of water by the same valence band holes. Suitable materials include organic acids, alcohols, aldehydes, amino acids, and hydrazine and its oxidation products, such as hydroxylamine and dihydroxylamine. The second property is that it must be capable of acting as a stabiliser for actinide metal ion reduction in aqueous nitric acid, predominantly through the scavenging of the nitrous acid generated both photolytically and photocatalytically upon UV-irradiation of HNO3 (Eqs. 17, 18, 32 and 34). The nitrous acid is capable of reducing Np(VI) to Np(V) (Eq. 16), oxidising U(IV) to U(VI): U4+ + 2HNO2 –→ UO2 2+ + 2NO + 2H+

(40)

and, similarly, capable of oxidising Pu(III) and Pu(IV) to Pu(VI). Such redox reactions prevent the efﬁcient separation of Pu and U and, in the cases of Np and U, essentially short circuit the valence manipulation of those metals. However, these nitrous-acid-driven oxidation reactions can be obviated by the addition of a stabiliser such as hydrazine to the reprocessing liquor, the stabiliser destroying the nitrous acid in accordance with Eqs. 3 and 4. Thus, the addition of hydrazine to the reprocessing liquor as a hole scavenger also results in photogenerated U(IV) and Pu(III) being stabilised against reoxidation, so rendering the U and Pu separable by solvent extraction. 5.3 Photocatalytic Valence Control of a Pu Simulant With the ﬁndings of Sects. 5.1 and 5.2 in mind, we began a series of experiments to investigate the control of actinide oxidation states by photocatalysis. In order to avoid complications associated with the photolysis of

The Applications of Photocatalytic Waste Minimisation

471

nitric acid (vide supra), preliminary studies were carried out in sulphuric acid media using ethanol as a hole scavenger. The ﬁrst system investigated was the reduction of Pu4+ to Pu3+ ; it is a simple one-electron transfer and does not involve the breaking of metal–oxygen bonds, so should be kinetically facile. To minimise waste management concerns and other issues associated with the manipulation of radioactive materials, experiments were conducted on a thermodynamic simulant for Pu. Cerium has previously been used as a simulant for Pu as the two systems exhibit similar thermodynamics [75, 111]. A series of experiments were carried out to observe the effect of illuminating Ce(IV) with light of wavelength 312 nm in the presence of colloidal SnO2 , sulphuric acid (pH 0) and ethanol as an electron scavenger. From Fig. 5, it can be seen that, upon illumination, the concentration of Ce(III), which was initially zero, increases with illumination time as a result of photocatalysed reduction of Ce4+ . The concentration of Ce3+ continues to increase until it is equal to the original concentration of Ce(IV). Upon the removal of the illumination, the cerium remains in the Ce3+ state. The quantum efﬁciency, φ for the generation of useful charge carrier pairs (i.e. those that can potentially be used in a valence control process) per photon absorbed, was found to be ∼ 0.27. This is high compared to values reported for similar processes in the literature and is a direct consequence of the charge transfer mechanism in operation. Under the conditions employed in the experiment of Fig. 5, both the particle and the Ce4+ cation are

Fig. 5 Concentration of Ce(III) as a function of illumination time, as determined by photopotentiometry. The Ce3+ is generated by photocatalysed reduction of Ce4+ achieved through illuminating a pH 0 H2 SO4 solution containing 0.12 mol Ce4+ m–3 , 55 mol ethanol m–3 and 100 g SnO2 m–3 with light of wavelength 312 nm. Illumination is removed at 920 s

472

C. Boxall et al.

positively charged, effectively eliminating the possibility of electron transfer occurring by a static quenching mechanism. Photocatalytic reduction of the ceric ion therefore occurs by a dynamic quenching mechanism. Importantly, the product of that reduction, the cerous ion, is also positively charged. Thus, the adsorption of the reduction product, and the re-injection of the transferred charge back into the particles (either into the conduction band or ﬁlling photogenerated valence band holes in an extrinsic recombination mechanism) are both unlikely. The importance of the back reaction was highlighted in a recent study from these laboratories of the CdS/Fe(CN)6 3– semiconductor/electron acceptor system [112]. We reported high yields for the photocatalytic reduction of Fe(CN)6 4– by a dynamic quenching mechanism and attributed this to inhibition of the reverse electron transfer by coulombic repulsion between the negatively charged particles and the anionic ferrocyanide ion. It is this same effect that is responsible for the large value of φ obtained for the SnO2 /Ce4+ system in Fig. 5, an observation that has farreaching implications for the design of photocatalytic systems for metal ion valence control. 5.4 Photocatalytic Valence Control of Uranium Having demonstrated the feasibility of using photocatalysis for controlling the valence state of the Pu simulant, Ce, we next investigated the possibility of controlling the valence state of a non-simulated actinide system. In this case, we decided to study the reduction of UO2 2+ . As stated in Sect. 2, there are difﬁculties associated with the generation and storage of the U(IV) reductant in PUREX reprocessing. Post-generation and pre-use, the U(IV) reductant may exhibit some instability in its speciation and be at least partially oxidised to U(VI) by nitric and nitrous acids. One possible mechanism for this oxidation is given by the reaction between U(IV) and nitric acid to give U(V), in the form of UO2 + , and nitrous acid: 2U4+ + HNO3 + 3H2 O –→ 2UO2 + + HNO2 + 6H+ +

U4+

2+

(41)

The UO2 disproportionates to form and UO2 while the nitrous acid oxidises further U(IV), although with a higher rate than nitric acid. It would therefore be desirable to provide an alternative process for controlling the oxidation states of U in the reprocessing of nuclear fuel, preferably by the generation of U(IV). Figure 6 shows the effect of illuminating UO2 2+ with light of wavelength 312 nm in the presence of colloidal SnO2 , sulphuric acid (pH 0) and ethanol as an electron scavenger. Neither ethanol, sulphuric acid, UO2 2+ or its anticipated reduction products, UO2 + or U4+ possess a signiﬁcant absorption band at this wavelength; thus, any photo-induced reduction reactions observed can be attributed to the action of the particles rather than the intrinsic photo-

The Applications of Photocatalytic Waste Minimisation

473

Fig. 6 Concentration of U(IV) as a function of illumination time. The U4+ is generated by photocatalysed reduction of UO2 2+ achieved by illuminating a solution containing 10 mol UO2 2+ m–3 , 55 mol ethanol m–3 and 100 g SnO2 m–3 with light of wavelength 312 nm

electrochemistries of the U system, electron scavenger or supporting acid. UV–visible spectrophotometry indicated that the primary reduction product formed during these experiments was U4+ . As can be seen from Fig. 6, upon illumination, the concentration of U(IV), which was initially zero, increases with illumination time as a result of photocatalysed reduction of UO2 2+ . The concentration of U4+ continues to increase until it is equal to the original concentration of U(VI). However, it remains to be determined whether the U(IV) is generated by direct reduction of UO2 2+ by photogenerated conduction band electrons, or whether the UO2 2+ is ﬁrst photocatalytically reduced to UO2 + , which then disproportionates to form U(IV) and U(VI). Having demonstrated that SnO2 -photocatalysed reduction of U(VI) to U(IV) is possible in sulphuric acid media, we elected to study the same phenomenon in HNO3 , the acid most commonly used in reprocessing cycles. Figure 7, series 1, shows the effect of illuminating U(VI) with light of wavelength 350 nm in the presence of colloidal SnO2 , nitric acid (pH 0) and ethanol as an electron scavenger for the semiconductor photocatalyst. As in the experiment of Fig. 6, neither ethanol, HNO3 , UO2 2+ or its anticipated reduction products, UO2 + or U4+ possess a signiﬁcant absorption band at this wavelength; thus, any photo-induced reduction reactions observed can be attributed solely to the action of the photocatalytic particles. Further, irradiation at 350 nm avoids the photogeneration of nitrous acid, HNO2 being capable of reoxidising any photocatalytically generated U(IV) back to U(VI) (Eq. 37, vide supra).

474

C. Boxall et al.

Fig. 7 [U(IV)] as a function of illumination time. U4+ is generated by photocatalysed reduction of UO2 2+ achieved through illuminating a solution of 10 mol UO2 2+ m–3 , 100 g SnO2 m–3 and 550 mol ethanol m–3 (Series 1) or 550 mol hydrazine m–3 (Series 2) with light of wavelength 350 nm

U(IV) was again found to be the principal product. Figure 7, series 1, shows that, upon illumination, the concentration of U(IV), which was initially zero, increases with illumination time as a result of photocatalysed reduction of UO2 2+ . [U4+ ] continues to increase until it is equal to the original concentration of U(VI). Upon removal of illumination, the concentration of photogenerated U(IV) decreases with time in both the presence and absence of oxygen. The former is expected due to the oxidising action of O2 on U4+ ; the latter is indicative of the presence of a second oxidising agent in solution. Nitric acid is capable of undergoing photolytic reduction to nitrous acid (Eqs. 17 and 18), which in turn is capable of oxidising U(IV) to U(VI). However, generation of HNO2 by this means is unlikely at the irradiation wavelength employed (350 nm) due to reasons discussed above. Thus, any HNO2 present must have been photocatalytically generated in accordance with Eq. 32. In order to test this hypothesis, the experiment of Fig. 7, series 1 was repeated in the presence of hydrazine which, as was discussed in Sect. 5.2, is capable of acting as both an electron donor for the semiconductor photocatalyst and as a scavenger for HNO2 . Figure 7, series 2, shows the effect of illuminating U(VI) with light of wavelength 350 nm in the presence of colloidal SnO2 , nitric acid (pH 0) and hydrazine. Hydrazine exhibits no signiﬁcant absorption bands at 350 nm. The results are broadly similar to those recorded in the presence of ethanol, series 1, the sole difference being that the reduction of U(VI) to U(IV) is slightly slower in the presence of hydrazine. This difference may be attributed to the respective kinetics/efﬁciencies of the ethanol and hydrazine systems as electron donors for SnO2 .

The Applications of Photocatalytic Waste Minimisation

475

Upon removal of illumination, the concentration of photogenerated U(IV) was found to be invariant with time, indicating that hydrazine is fulﬁlling its role as a stabiliser for actinide metal reduction by the scavenging of HNO2 and that, as in the case of photocatalysed reduction Ce(IV) to Ce(III) (Sect. 5.3), the photocatalysed reduction of U(VI) to U(IV) can be rendered permanent on timescales appropriate for nuclear reprocessing. An even more important observation is that, in the presence of hydrazine, U(IV) is also stable in the presence of oxygen, meaning that measures will not have to be taken to exclude oxygen in any ultimate plant-based deployment of the process. In analogy to the PUREX process, photocatalysed reduction of actinide metal ions can be employed in tandem with solvent extraction in order to achieve actinide separation. A process can be envisaged wherein photocatalysed reduction occurs in one of two solvent phases in contact—one aqueous and one non-aqueous—and wherein, as a result, the reduced metal ion is selectively retained by one of the solvent phases, either the phase it originated in, or as a consequence of a phase transfer reaction. Thus, experiments were conducted to assess both the efﬁcacy of photocatalysed actinide metal ion reduction in a two-solvent phase system, and the efﬁcacy of simultaneous selective product retention in one of those two phases. 5.5 Photocatalytic Valence Control in Mixed Solvent and Mixed Actinide Systems Figure 8, series 2, shows the effect of illuminating a two solvent system with light of wavelength 350 nm, the two phases being: (1) an aqueous phase containing U(VI) in the presence of colloidal SnO2 , nitric acid (pH 0) and hydrazine; and (2) a non aqueous phase of TBP. TBP exhibits no signiﬁcant absorption bands at 350 nm. Again, U4+ was found to be the principal reduction product. As U(IV) is extractable into the non-aqueous organic phase, the time dependence of the concentration of U(IV) in TBP is shown in the diagram. Figure 8 shows that the concentration of U(IV) in the TBP phase increases with illumination time as a result of photocatalysed reduction of UO2 2+ originating from the aqueous phase. The concentration of U4+ in TBP continues to increase until it is equal to the original concentration of U(VI) in the aqueous phase. At the end of the period of illumination, the concentrations of both U(IV) and U(VI) in the aqueous phase are virtually zero, indicating that ∼ 100% of the U(VI) has been reduced to U(IV) and transferred from the aqueous phase to the non-aqueous phase. Figure 8 also compares this two-phase data with the comparable one-phase data of Fig. 7 and it can be seen that the rates of U(IV) evolution are virtually identical, suggesting that the semiconductor-photocatalysed reduction

476

C. Boxall et al.

Fig. 8 Series 1: [U(IV)] as a function of illumination time. U4+ is generated by photocatalysed reduction of UO2 2+ achieved through illuminating a one-solvent phase, aqueous solution containing 10 mol UO2 2+ m–3 , 100 g SnO2 m–3 and 550 mol hydrazine m–3 with light of wavelength 350 nm. Series 2: [U(IV)] in TBP as a function of illumination time. U4+ is generated by photocatalysed reduction of UO2 2+ achieved through illuminating a two-solvent phase system (one solvent phase being an aqueous solution initially containing 10 mol UO2 2+ m–3 , 100 g SnO2 m–3 and 550 mol hydrazine m–3 , the other solvent phase initially containing only TBP) with light of wavelength 350 nm

of U(VI) to U(IV) occurs almost exclusively in the aqueous phase, and that it is then followed by a fast phase transfer of U(IV) to the non-aqueous phase. Upon removal of illumination, the concentration of photogenerated U(IV) in the TBP phase was found to be invariant with time, indicating that, as in the aqueous phase photocatalysed reductions of Ce(IV) to Ce(III) and U(VI) to U(IV), the photocatalysed reduction of U(VI) to U(IV) in the non-aqueous TBP phase can be rendered permanent on timescales appropriate for nuclear reprocessing. When deployed on-line, the semiconductor photocatalyst may be required to photoreduce more than one type of actinide metal ion simultaneously. Figure 9 shows the effect of illuminating U(VI) with light of wavelength 350 nm in the presence of colloidal SnO2 , nitric acid (pH 0) and ethanol as an electron scavenger for the semiconductor photocatalyst and Ce(IV) as a nonradioactive, thermodynamic analogue for Pu(IV). Comparison of the data in Fig. 9 with the data recorded under similar conditions as shown in Fig. 7 indicates that the presence of Ce(IV) has no effect on the rate of photocatalysed reduction of U(VI) to U(IV). Furthermore, spectroscopic analysis indicates that virtually all of the Ce(IV) has been reduced to Ce(III) over the same timescale, suggesting that the simultaneous photocatalysed reduction of two or more different types of (actinide) metal ion can be accomplished with no loss of yield for either reaction.

The Applications of Photocatalytic Waste Minimisation

477

Fig. 9 [U(IV)] as a function of illumination time. U4+ is generated by photocatalysed reduction of UO2 2+ achieved through illuminating an aqueous solution containing 10 mol UO2 2+ m–3 , 0.05 mol Ce4+ m–3 , 100 g SnO2 m–3 and 550 mol hydrazine m–3 with light of wavelength 350 nm

6 Concluding Remarks The control of the valence state of actinide metal ions by heterogeneous photocatalysis is possible in both aqueous and non-aqueous solution environments with high yields and quantum efﬁciencies. Semiconductor catalysts that are stable under conditions typical of nuclear reprocessing streams have been identiﬁed, suggesting that photocatalysis may ﬁnd utility in a number of reprocessing scenarios. In the highly active cycle of the PUREX process these include: (1) the control of Np ion oxidation state and consequent separation of Np from Pu and U; and (2) the control of U and Pu ion oxidation states and their consequent separation from each other. It is foreseen that these adjustments to PUREX would lead to further improvements in the efﬁciency of separation of U from Pu in nuclear fuel and materials processing with the following environmental consequences: (i) an even smaller fraction of U and Pu will remain inseparable, so decreasing the need for primary waste management by vitriﬁcation; and (ii) a diminished reagent requirement all round and especially in the use of organophosphate-based solvents during the solvent extraction steps of U/Pu separation, thus lessening the environmental impact of the disposal of those solvents. Factors found to be contributory to efﬁcient actinide valence control include: 1. The thermodynamic stability of the particles. 2. Semiconductor band edge energetics: manipulation of the band edges by solution pH or the addition of dopants to the particle matrix allows for the tailoring of the photogenerated electron energetics to speciﬁc actinide reductions.

478

C. Boxall et al.

3. Particle surface charge: ideally this should be the same as the charge of the target actinide ion, i.e. cationic, so encouraging electron transfer via a dynamic quenching mechanism with its attendant high quantum efﬁciencies. 4. Choice of sacriﬁcial charge scavenger: in photoreductively induced actinide valence changes (e.g. U(VI) to U(IV)), it is advantageous to use reductive additives such as hydrazine that act as both hole scavengers and stabilising agents for the valence state generated. It has been suggested that the ambient ionising radiation ﬁeld of a reprocessing stream may induce electron–hole generation in semiconductor particles/ﬁlms in situ, so obviating the need for an external UV light source. If such were found to be the case, heterogeneous photocatalytic systems could be deployed more easily. Acknowledgements The authors wish to thank the University of Central Lancashire for a period of sabbatical leave for CB during which the manuscript was prepared and BNFL for their ﬁnancial support of the work described in Sect. 5.

References 1. Barni B, Cavicchioli A, Riva E, Zanoni L, Bignoli F, Bellobono IR, Gianturco F, De Giorgi A, Muntau H, Montanarella L, Facchetti S, Castellano L (1995) Chemosphere 30:1861 2. Bellobono IR, Carrara A, Barni B,Gazzotti A (1994) J Photochem Photobiol A: Chem 84:83 3. Legrini O, Oliveros E, Braun AM (1993) Chem Rev 93:671 4. Hagfeldt A, Grätzel M (1995) Chem Rev 95:49 5. Ollis DF, Al-Ekabi H (1993) In: Ollis DF, Al-Ekabi H (eds) Photocatalytic puriﬁcation and treatment of water and air. Elsevier, Amsterdam 6. Hoffman MR, Martin ST, Choi W, Bahnemann D (1995) Chem Rev 95:69 7. Litter MI (1999) Appl Catal B: Environ 23:89 8. Mills A, Davies RH, Worsley D (1993) Chem Soc Rev 22:417 9. Boxall C (1994) Chem Soc Rev 23:137 10. Boxall C (1999) In: Compton RG, Hancock G (eds) comprehensive chemical kinetics vol 37: applications of kinetic modeling. Elsevier, Amsterdam, p 281–370 11. Grätzel M (1992) In: Mackay RA, Texter J (eds) Electrochemistry in colloids and dispersions. VCH, New York, p 357–374 12. Serpone N, Lawless D, Terzian R, Meisel D (1992) In: Mackay RA, Texter J (eds) Electrochemistry in colloids and dispersions. VCH, New York, p 399–416 13. Kamat PV (1993) Chem Rev 93:267 14. Kamat PV (1994) Prog React Kinet 19:277 15. Henglein A (1995) Ber Bunsenges Phys Chem 99:903 16. Henglein A (1989) Chem Rev 89:1861 17. Fox MA, Dulay MT (1993) Chem Rev 93:341 18. Stafford U, Gray KA, Kamat PV (1996) Heterogen Chem Rev 3:77 19. Bahnemann DW (1993) Isr J Chem 33:115

The Applications of Photocatalytic Waste Minimisation

479

20. Bahnemann D, Cunningham J, Fox MA, Pelizzetti E, Serpone N (1994) In: Helz GR, Zepp RG, Crosby DG (eds) Aquatic and Surface Photochemistry. Chap 21. Lewis, Boca Raton, FL, p 261–316 21. Linsebigler AL, Lu G, Yates JT (1995) Chem Rev 95:735 22. Halmann MM (1996) Photodegradation of water pollutants. CRC, Boca Raton, FL 23. Nageswara Rao N, Natarajan P (1994) Curr Sci 66:742 24. Wang Y (1991) Acc Chem Res 24:133 25. Pichat P (1994) Catal Today 19:313 26. Pelizzetti E, Serpone N (eds) (1986) Homogeneous and heterogeneous Photocatalysis. Reidel, Dordrecht 27. Lewis NL (1993) Chem Rev 93:2693 28. Grätzel M (1989) Heterogeneous photochemical electron transfer. CRC, Boca Raton, FL 29. Serpone N, Pelizzetti E (eds) (1989) Photocatalysis: fundamentals and Applications. Wiley, New York 30. Mills A, Le Hunte S (1997) J Photochem Photobiol A: Chem 108:1 31. Herrmann JM (1999) Catal Today 53:115 32. Fujishima A, Rao TN, Tryk DA (2000) J Photochem Photobiol C: Photochem Rev 1:1 33. Tryk DA, Fujishima A, Honda K (2000) Electrochim Acta 45:2363 34. Robertson PKJ, Bahnemann D (2006) In: Boule P, Bahnemann D, Robertson PKJ (eds) The handbook of environmental chemistry, vol. 2, part M: environmental photochemistry part II. Springer, Berlin Heidelberg New York (in this volume) 35. Litter MI (2006) In: Boule P, Bahnemann D, Robertson PKJ (eds) The handbook of environmental chemistry, vol. 2, part M: environmental photochemistry part II. Springer, Berlin Heidelberg New York (in this volume) 36. Ohtani B, Zhang SW, Nishimoto S, Kagiya T (1992) J Chem Soc Faraday Trans 88:1049 37. Matsunaga T, Okochi M (1995) Environ Sci Technol 29:501 38. Zhang P, Scrudato RJ, Germano G (1994) Chemosphere 28:607 39. Dunlop PSM, Byrne JA, Manga N, Eggins BR (2002) J Photochem Photobiol A: Chem 148:355 40. Lee S, Nishida K, Otaki M, Ohgaki S (1997) Water Sci Technol 35:101 41. Bellobono IR, Gianturco F, Chiodaroli CM (1997) Fresenius Environ Bull 6:469 42. Pleskov YV, Krotova MD (1993) Electrochim Acta 38:107 43. Grätzel M (1981) Acc Chem Res 14:376 44. Gerisher H, Heller A (1992) J Electrochem Soc 139:113 45. Berry RJ, Mueller MR (1994) Microchem J 50:28 46. Prairie MR, Evans LR, Martinez SL (1992) Chem Oxid 2:428 47. Domenech J, Prieto A (1986) Electrochim Acta 31:1317 48. Morishita S, Suzuki K (1994) Bull Chem Soc Jpn 67:843 49. Serpone N, Ah-You YK, Tran TP, Harris R, Pelizzetti E, Hidaka H (1987) Solar Energy 39:491 50. Skubal LR, Meshkov NK (2002) J Photochem Photobiol A: Chem 148:211 51. Kriek RJ, Engelbrecht WJ, Cruywagen JJ (1995) J S Afr Inst Min Metall 75 52. Jacobs JWM (1986) J Phys Chem 90:6507 53. Vamathevan V, Amal A, Beydoun D, Low G, McEvoy S (2002) J Photochem Photobiol A: Chem 148:233 54. Hufschmidt D, Bahnemann D, Testa JJ, Emilio CA, Litter MI (2002) J Photochem Photobiol A: Chem 148:223

480

C. Boxall et al.

55. Skubal LR, Meshkov NK, Rajh T, Thurnauer M (2002) J Photochem Photobiol A: Chem 148:393 56. Borgarello E, Serpone N, Emo G, Harris R, Pelizzetti E, Minero C (1986) Inorg Chem, 25:4499 57. Garcia-Gonzalez ML, Martinez Chaparro A, Salvador P (1993) J Photochem Photobiol A: Chem 73:221 58. Malvyn J, McKibben (1984) Radiochim Acta 36:3 59. Fukasawa T, Kamoshida M, Kawamura F, Miura N (1996) European Patent 96:300785.1 60. Kitamori T, Nishi T, Fukasawa T, Fujimori H, Sasahira A, Ozawa Y, Suzuki K, Yusa H (1987) European Patent 87-1000807.4 61. Amadelli R, Maidotti A, Sostero S, Carassiti V (1991) J Chem Soc Faraday Trans 87:3267 62. Boxall C, Taylor RJ, Xiao S, LeGurun G (2003) International Patent Application PCT/GB02/004875 63. Fukasawa T, Kawamura F (1990) Japanese Patent 02-208210A 64. Fukasawa T, Takahashi M, Ikeda T, Kawamura F (1991) Proc 3rd Int Conf Nucl Fuel Reprocessing Waste Management, RECOD ’91, Sendai, Japan, vol. II, pp 803 65. Fukasawa T, Kawamura F (1991) J Nucl Sci Technol 28:27 66. Goldacker H (1993) Design of an interstage electrochemical reactor for the separation of uranium and plutonium by solvent extraction. PhD thesis, University of Newcastle upon Tyne 67. Phillips C (1993) Chem Eng Res Des 71(A2):134 68. Fukasawa T, Ikeda T, Kawamura F (1991) Eur J Solid State Inorg Chem 28:73 69. Wehrey F, Guillaume B (1989) Radiochim Acta 46:95 70. Tochiyama O, Nakamura Y, Katayama Y, Inoue Y (1995) J Nucl Sci Technol 31:50 71. Goherty AMM, Howes KR, Stedman G, Naji MQ (1995) J Chem Soc Dalton Trans 19:3103 72. Baumgartner F, Schmeider H (1978) Radiochim Acta 25:191 73. Bennet WE (1985) Chem Eng 415:35 74. Eccles H, Naylor S (1987) Chem Ind (London) 6:174 75. Cotton FA, Wilkinson G (1988) Advanced Inorganic Chemistry, 5th edn. Wiley, Chichester 76. Toth TM, Bell JT, Friedman HA (1980) ACS Symp Ser 117:253 77. Yusov AB, Shilov VP (1995) Usp Khim 64:888 78. Burrows HD, Kemp TJ (1974) Chem Soc Rev 3:139 79. Miyake C, Nakase T, Sano Y (1993) J Nucl Sci Technol 30:1256 80. Mizuguchi K, Park YY, Tomiyasu H (1993) J Nucl Sci Technol 30:542 81. Wada Y, Wada K, Goibuchi T, Tomiyasu H (1994) J Nucl Sci Technol 31:700 82. Sarakha M, Bolte M, Burrows HD (2000) J Phys Chem 104:3142 83. Khamidullina LA, Lotnik SV, Kazakov VP (1994) High Energy Chem 28:122 84. Arvis M, Keller N, Folcher G, Hickel B (1983) J Photochem 21:313 85. Bouby M, Billard I, Bonnenfant A, Klein G (1999) Chem Phys 240:353 86. Zielen AJ, Sullivan JC, Cohen D (1958) J Inorg Nucl Chem 7:378 87. Nagaishi R, Katsumura Y, Ishigure K, Aoyagi H, Yoshida Z, Kimura T (1996) J Photochem Photobiol A Chem 96:45 88. Shilov VP, Astafurova LN, Yusov AB (1995) Radiokhimiya 37:507 89. Palei PN, Nemodruk AA, Bezrogova EV (1969) Radiokhimiya 11:300 90. Nemodruk AA, Bezrogova EV, Ivanova SA, Novikov YP (1972) Zh Anal Khim 27:1270 91. Nemodruk AA, Bezrogova EV, Ivanova SA, Novikov YP (1973) Zh Anal Khim 28:379 92. Nemodruk AA, Bezrogova EV, Ivanova SA, Novikov YP (1972) Zh Anal Khim 27:73

The Applications of Photocatalytic Waste Minimisation

481

93. Nemodruk AA, Bezrogova EV, Ivanova SA, Novikov YP (1972) Zh Anal Khim 27:2414 94. Galkin VY, Krutnitskij CV, Lazarev LN, Mishin KY, Romanovskij VN (1985) Radiokhimiya 27:494 95. Shilov VP, Yusov AB (1996) Radiokhimiya 38:424 96. Wada Y, Morimoto K (1996) Radiochim Acta 72:195 97. Koltunov VS, Tikhonov MF (1977) Radiokhimiya 19:620 98. Koltunov VS, Marchenko VI, Zhuravleva GI, Tikhonov MF, Shapavalov MP (1976) Radiokhimiya 18:65 99. Wada Y, Morimoto K, Goibuchi T, Tomiyasu (1995) Radiochim Acta 68:233 100. Nakamura T, Takahashi M, Fukasawa T, Utamura M (1992) J Nucl Sci Technol 29:393 101. Bondietti EA, Trabalka JR (1980) Radiochim Radioanal Lett 42:169 102. Chen J, Ollis DF, Rulkens WH, Bruning H (1999) Colloids & Surfaces A: Physicochem Eng Aspects 151:339 103. Boxall C, Kelsall GH (1991) J Chem Soc Faraday Trans 87:3537 104. Pourbaix M (1974) Atlas of electrochemical equilibria in aqueous solutions, 2nd edn. NACE, Houston, Texas 105. Yoshizumi M, Wakabayashi K (1987) Plast Eng 87:61 106. Stucki S, Kötz R, Carcer B, Suter W (1991) J Appl Electrochem 21:99 107. Armstrong NR (1988) In: Finklea H (ed) Semiconductor electrodes. Elsevier, Amsterdam 108. Kung HH, Jarret HS, Sleight AW, Ferretti A (1977) J Appl Phys 48:2463 109. Boxall C, Kelsall GH (1992) J Electroanal Chem 328:75 110. LeGurun G, Holden AJ, Boxall C (1997) Personal communication 111. Choppin GR (1995) J Alloys & Compounds 223:174 112. Boxall C (2002) J Photochem Photobiol A: Chem 148:375