Electron Paramagnetic Resonance (Electron Paramagnetic Resonance, 20)

Electron Paramagnetic Resonance Volume 20 A Specialist Periodical Report Electron Paramagnetic Resonance Volume 20 A ...

Author: Bruce C Gilbert | Michael J Davies | Damien M Murphy

44 downloads 996 Views 5MB 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

Electron Paramagnetic Resonance Volume 20

A Specialist Periodical Report

Electron Paramagnetic Resonance Volume 20 A Review of the Recent Literature Editors BC Gilbert, University of York, UK MJ Davies, Heart Research Institute, Sydney, Australia DM Murphy, Cardiff University, UK

Authors A Caragheorgheopol, Romanian Academy Institute of Physical Chemistry, Bucharest, Romania V Chechik, University of York, UK J-L Cle´ment, Universite´ d’Aix Marseille, Marseille, France D Collison, The University of Manchester, UK JB Feix, Medical College of Wisconsin, Milwaukee, US JT Hancock, University of the West of England, Bristol, UK SK Jackson, University of the West of England, Bristol, UK PE James, Cardiff University, UK CS Klug, Medical College of Wisconsin, Milwaukee, US EJL McInnes, The University of Manchester, UK MV Motyakin, Russian Academy of Science, Moscow, Russia F Neese, New Universita¨t Bonn, Germany ME Newton, University of Warwick, UK PC Riedi, University of St Andrews, UK S Schlick, University of Detroit Mercy, Michigan, US P. Tordo, Universite´ d’Aix Marseille, Marseille, France

If you buy this title on standing order, you will be given FREE access to the chapters online. Please contact [email protected] with proof of purchase to arrange access to be set up. Thank you.

ISBN-10: 0-85404-325-X ISBN-13: 978-0-85404-325-5 ISSN 1464-4622 A catalogue record for this book is available from the British Library r The Royal Society of Chemistry 2007 All rights reserved Apart from any fair dealing for the purpose of research or private study for non-commercial purposes, or criticism or review as permitted under the terms of the UK Copyright, Designs and Patents Act, 1988 and the Copyright and Related Rights Regulations 2003, this publication may not be reproduced, stored or transmitted, in any form or by any means, without the prior permission in writing of The Royal Society of Chemistry, or in the case of reprographic reproduction only in accordance with the terms of the licences issued by the Copyright Licensing Agency in the UK, or in accordance with the terms of the licences issued by the appropriate Reproduction Rights Organization outside the UK. Enquiries concerning reproduction outside the terms stated here should be sent to The Royal Society of Chemistry at the address printed on this page. Published by The Royal Society of Chemistry, Thomas Graham House, Science Park, Milton Road, Cambridge CB4 0WF, UK Registered Charity Number 207890 For further information see our web site at www.rsc.org Typeset by Macmillan India Ltd, Bangalore, India Printed by Henry Ling Ltd, Dorchester, Dorset, UK

Preface The topics described in Volume 20 of the EPR Specialist Periodical Report Series have been selected to reﬂect particularly exciting and timely examples of the renaissance in the development and application of EPR spectroscopy, including ENDOR, time-resolved and imaging techniques. We have again aimed to reﬂect the widespread application of EPR in chemistry and the cognate ﬁelds of physics, material science, biology, and medicine – and to provide updates for specialists as well as overviews for non-experts who may wish to learn of the scope of the technique. Authors with particular expertise have been invited to contribute authoritative and critical reviews which, whilst concentrating on major developments in 2004 and 2005, also provide descriptive and accessible accounts which are suitable for both existing and potential practitioners. We have intended to balance different types of article in this Volume. For example, in some cases we have continued our coverage of developments in relatively well-established ﬁelds for which detailed updates are appropriate – for example, in describing recent advances in spin-trapping methodology (especially in the design of new traps), in the study of exchange-coupled multi-spin oligomers, the application of high-ﬁeld EPR (especially to inorganic compounds and materials) and in biological and biomedical applications. In other areas we have invited practicing experts to illustrate the scope for application of EPR spectroscopy in new or rapidly-developing ﬁelds, and have accordingly chosen the use of EPR-imaging in studies of polymer degradation and the use of spin-labels and spin-probes for investigating a variety of nanomaterials; a review of site-directed spin-labelling (SDSL) in biological systems illustrates not only the substantial activity in this ﬁeld, but also, as with applications in nanomaterials and oligomers, shows how study and understanding of spin-spin interactions provides unique, subtle, and detailed information about molecular motion and distances. We have added two particularly novel chapters for this Volume – a timely review of quantum chemical approaches to aid the interpretation of spinHamiltonian parameters (reﬂecting the increasingly accessible and informative links which can be made between theoretical analysis and magnetic resonance observations) and an authoritative essay on defects in diamonds (which illustrates the beauty of the method and theoretical analysis, as well as potential materials-based applications).

v

vi

Electron Paramag. Reson., 2007, 20, v–vi

We hope that experts and potential beginners alike will value these reviews and be stimulated in their research. As always, we would value feedback on our choice of topics and styles, together with suggestions of possible topics for future coverage. Finally we thank all the authors for their care and promptness in contributing their articles, and acknowledge with gratitude the skilled cooperation of the RSC Editorial Staﬀ. Bruce Gilbert Michael Davies Damien Murphy

Contents Cover The cover depicts a singlecrystal EPR study of a dodecametallic Cr(III) cluster with an S=6 ground state.

ESR Imaging Beyond Phantoms: Application to Polymer Degradation Shulamith Schlick and Mikhail V. Motyakin 1

2

3

4 5

Introduction and Motivation 1.1 Polymer Degradation 1.2 Polymer Stabilization by Hindered Amines 1.3 Motivation for ESRI Studies of Polymer Degradation ESR Imaging and Applications to Polymer Systems 2.1 ESR Spectra in the Presence of Magnetic Field Gradients 2.2 ESR Imaging of Polymeric Systems 2.3 ESR Imaging of Polymers Stabilized by Hindered Amines Experimental Details 3.1 Hardware for ESRI 3.2 Intensity Proﬁling from 1D ESRI 3.3 Line-Shape Proﬁling from 2D Spectral-Spatial ESRI 3.4 The Polymers Studied ESR Spectra of Nitroxides Derived From Hindered Amine Stabilizers in Heterophasic Polymers 1D and 2D Spectral-Spatial ESRI of Heterophasic Polymers Stabilized by Hindered Amines vii

1

1 1 2 2 5 5 6 7 8 8 9 10 11 12 14

viii

Electron Paramag. Reson., 2007, 20, vii–xii

5.1 5.2 5.3

Thermal Treatment Photodegradation Advantages of ESRI Compared to FTIR and to Destructive Methods of Study 6 Conclusions Acknowledgments References

Advances in Spin Trapping J.-L. Cle´ment and P. Tordo 1 2 3 4

Introduction Basic Principles New Nitrones for Spin Trapping Spin Trapping Studies of Superoxide Radicals 4.1 Use of Cyclodextrins 4.2 Formation of Diastereoisomer Spin Adducts 4.3 Kinetic Studies 4.4 Chemical Exchange in Alkylperoxyl and Superoxide Radical Adducts of b-Phosphorylated Nitrones 5 Theoretical Calculations 6 Immuno Spin Trapping of Proteins 7 Separation and Identiﬁcation of Spin Adducts 8 Conclusion References

Site-Directed Spin-Labelling (SDSL) Applications in Biological Systems Jimmy B. Feix and Candice S. Klug 1 Introduction 2 Soluble Proteins 3 Integral Membrane Proteins 4 Membrane-Associated Proteins 5 Fibril-Forming Proteins 6 Nucleic Acids 7 Substrates 8 Distance Measurements 9 Conclusions Acknowledgements References

14 18 23 24 25 25

29

29 30 30 36 36 37 38 41 43 43 44 45 45

50

50 51 53 56 62 62 63 64 67 68 68

Electron Paramag. Reson., 2007, 20, vii–xii

Quantum Chemical Approaches to Spin-Hamiltonian Parameters Frank Neese 1 2

Introduction Theoretical Aspects 2.1 General Theory 2.2 Spin-Free Approaches 2.3 Two-Component Approaches 2.4 The Spin-Orbit Coupling Operator 3 Application to EPR Spin-Hamiltonian Parameters 3.1 g-Tensor 3.2 Hyperﬁne Couplings 3.3 Zero-Field Splittings 3.4 Quadrupole Couplings 3.5 Functional and Basis Set Eﬀects 3.6 Environment Eﬀects 4 Application Studies 4.1 Small Molecules 4.2 Organic and Biological Radicals 4.3 Transition Metals 5 Concluding Remarks Acknowledgements References Getting an Inside View of Nanomaterials with Spin Labels and Spin Probes Victor Chechik and Agneta Caragheorgheopol 1 2 3 4 5 6

Introduction Self-Assembled Supramolecular Structures Polymers Mesoporous Materials Adsorption in Mesopores Adsorption on Surfaces 6.1 Planar Surfaces 6.2 Rough Surfaces 7 Au Nanoparticles 8 Dendrimers 9 Conclusions Acknowledgements References

ix

73

73 74 74 75 77 78 80 80 82 83 84 84 85 86 86 86 87 87 88 88

96

96 98 102 107 112 115 115 117 119 123 127 128 128

x

Electron Paramag. Reson., 2007, 20, vii–xii

EPR, ENDOR and EPR Imaging of Defects in Diamond M.E. Newton 1 2

Introduction The Use of EPR and Related Techniques in the Study of Defects in Diamond 3 Intrinsic Defects in Diamond 4 Paramagnetic Nitrogen Defects in Diamond 5 Transition Metal Defects in Diamond 6 Paramagnetic Defects Incorporating Hydrogen in Diamond 7 EPR Imaging of Paramagnetic Defect Distributions in Diamond 8 Conclusions and Further Work References

EPR of Exchange Coupled Oligomers David Collison and Eric J.L. McInnes 1 2 3 4 5 6

Introduction p-Block d-Block Mixed p/d-Block Radicals Mixed d/f-Block Radicals Biological Systems 6.1 Methods 6.2 Nitrogenases 6.3 Copper 6.4 Manganese (Excluding Photosystems) 6.5 Diiron (Including 2Fe-2S) 6.6 Other Iron–Sulfur Centres 6.7 Photosystems 6.8 Nickel 6.9 Miscellaneous References

Biological Free Radicals and Biomedical Applications of EPR Spectroscopy Simon K. Jackson, John T. Hancock and Philip E. James 1 2

Introduction Reactive Oxygen Species

131

131 133 135 138 140 143 145 152 153

157

157 157 162 170 174 174 174 175 176 177 179 180 183 186 187 187

192

192 193

Electron Paramag. Reson., 2007, 20, vii–xii

2.1 Superoxide Radicals 2.2 Hydroxyl Radicals 3 Reactive Nitrogen Species 3.1 Nitric Oxide 3.2 Principles of dNO Measurement by EPR 3.3 Metal Chelate Complexes 3.4 Haemoglobin 3.5 Nitroxides as Spin Traps for dNO 3.6 Simultaneous Measurement of dNO and O2 3.7 Other Endogenous Paramagnetic Products of dNO 4 Enzyme-Mediated Free Radical Production 4.1 Cytochrome c 4.2 Cytochrome P450 4.3 NADPH Oxidase 5 Antioxidant Defences 5.1 Superoxide Dismutase 5.2 Catalase 5.3 Glutathione Peroxidase 5.4 Ascorbate (Vitamin C) 5.5 Vitamin E 5.6 Phenolic Antioxidants 6 Consequences of Free Radical Reactions with Biomolecules 6.1 Damage to Lipids 6.2 Damage to DNA 6.3 Damage to Proteins 6.4 Damage to Carbohydrates 7 Free Radicals and Disease 7.1 Cancer 7.2 Diabetes 7.3 Sepsis 7.4 Cardiovascular Disease 7.5 Oxidative Stress in other Disease Settings 7.6 Exercise and High Altitude Stress 7.7 Brain Injury 7.8 Other Systems 8 Apoptosis 9 Free Radicals in Plants 9.1 Reactive Oxygen Species Production by Plants 9.2 The Measurement of ROS in Plants by EPR 9.3 The Production of Nitric Oxide by Plant Cells 9.4 The Measurement of dNO in Plants by EPR 9.5 Other Processes in Plants Studied by EPR 10 Selected Biomedical Techniques

xi

193 195 196 196 196 197 198 198 199 199 201 201 202 203 204 204 205 206 206 207 208 209 209 210 212 215 216 216 217 218 219 221 221 221 222 222 223 223 224 226 226 228 229

xii

Electron Paramag. Reson., 2007, 20, vii–xii

10.1 In vivo EPR 10.2 EPR Oximetry 10.3 EPR Imaging References

Progress in High-Field EPR of Inorganic Materials Peter C. Riedi 1 2

Introduction Materials Research 2.1 Fullerenes 2.2 Semiconductors 2.3 Catalysis 2.4 Ferroelectrics 2.5 Transition Metal and Rare Earth Ions 2.6 Molecular Magnetic Clusters 2.7 Low-Dimensional Solids References

229 229 230 231

245

245 246 246 247 249 251 252 256 261 265

ESR Imaging Beyond Phantoms: Application to Polymer Degradation BY SHULAMITH SCHLICK1 AND MIKHAIL V. MOTYAKIN2 1 Department of Chemistry, University of Detroit Mercy, 4001 W. McNichols Rd, Detroit, Michigan 48221-3038, US 2 Institute of Chemical Physics, Russian Academy of Science, U1 Kosygina 4, Moscow 119991, Russia

1

Introduction and Motivation

1.1 Polymer Degradation. – Polymers undergo degradation when exposed to heat, mechanical stress, and ionizing or UV irradiation in the presence of oxygen, due to the formation of reactive intermediates such as free radicals Rd and ROOd, and hydroperoxides ROOH.1–4 Exposure to environmental factors leads to profound changes in polymer properties, on both molecular and macroscopic levels. The chemical structure is modiﬁed due, for example, to chain scission and cross-linking, resulting in changes of the elastic properties and of the degree of crystallinity.1 The chemistry of degradation is complicated because even small amounts of chromophores, free radicals, and metallic residues from polymerization reactions can introduce additional reaction pathways that usually enhance the rate of degradation. Polymer degradation is equivalent to corrosion in metals, a fundamental problem with important practical ramiﬁcations. Electron spin resonance (ESR)w methods have been used extensively for detecting and identifying the radicals formed, clarifying the degradation mechanism, and simulating the variation of the ESR spectra with temperature. Simulations of ESR line shapes for speciﬁc models of dynamics have been developed for the study of oxidative degradation of polymers due to ionizing radiation.5 Irradiation in vacuo has enabled the study of the type and mobility of alkyl radicals Rd derived from the polymer. These studies were initially performed on polytetraﬂuoroethylene (Teﬂon) and other perﬂuorinated polymers, because perﬂuoroalkyl radicals can be stabilized even at ambient temperature; in these polymers mid-chain and end-chain alkyl radicals have been w We will use ESR throughout this Chapter. The acronym EPR is reserved for ethylene-propylene rubber.

Electron Paramagnetic Resonance, Volume 20 r The Royal Society of Chemistry, 2007 1

2

Electron Paramag. Reson., 2007, 20, 1–28

detected. Admission of oxygen led to the formation of the corresponding peroxy radicals, ROOd. The motional mechanism of the peroxy radicals was deduced by simulation of the temperature dependence of the spectra; in this way a correlation between dynamics and reactivity has been established. This approach has also been extended to protiated polymers, for instance polyethylene and polypropylene.6 Direct ESR and spin-trapping experiments have identiﬁed oxygen radicals as well as membrane-derived fragments in Nafion, a perﬂuorinated ionomer used as a proton exchange membrane (PEM) in fuel cells, exposed to oxygen radicals produced in the Fenton reaction or by UV irradiation of hydrogen peroxide.7–9 Identiﬁcation of the radicals was possible by variation of sample preparation methods, temperature used for spectra acquisition, and annealing conditions. 1.2 Polymer Stabilization by Hindered Amines. – Recent research efforts on the effects of radiation and thermal treatment on polymeric materials have two main goals: understanding the degradation mechanism and predicting polymer lifetimes, and development of protective additives.10–14 Hindered amine stabilizers (HAS) rank among the most important recent developments for stabilization of polymeric materials. Nitroxides and amino ethers are major products of reactions involving HAS. The HAS-derived nitroxides (HASNO) are thermally stable, but can scavenge free radicals to yield diamagnetic species; the amino ethers can regenerate the original nitroxide, thus resulting in an efﬁcient protective effect. Some of these events are shown in Figure 1, where 4NH denotes the amine, 4NOd the nitroxide, and Rd, ROOd and ROOH the reactive intermediates derived from polymer chains exposed to oxygen and irradiation or heat. The intermediate radical N-peroxy radical 4NOOd has been detected by ESR.13 The most stable is 4NOd. Though stable, however, HAS-NO can react with alkyl radicals derived from polymeric precursors, and stabilize the polymer. In spite of numerous studies, important information on the degradation steps and kinetics is still incomplete or missing altogether.15 1.3 Motivation for ESRI Studies of Polymer Degradation. – The concept of diffusion-limited oxidation (DLO) has greatly contributed to the understanding of the mechanism for polymer degradation: if oxygen diffusion is slow R

ROO NH

• NO

•

•

ROOR

Figure 1 The chemistry of hindered amine stabilizers (HAS)

NOR

ROO

•

3

Electron Paramag. Reson., 2007, 20, 1–28

compared to the rate of degradation, as in accelerated degradation in the laboratory, only thin surface layers in contact with air are degraded, while the sample interior is little, if at all, affected; this is the DLO regime.16,17 The presence of oxygen is crucial for oxidative degradation of polymers. In the diffusion limited oxidation regime, the distribution of oxygen in the polymer can be described by equation (1),18,19 2 @½O2 @ ½O2 ¼D k½O2 @t @x2

ð1Þ

where the ﬁrst term on the right describes the rate of oxygen diffusion, and D is the oxygen diffusion coefﬁcient; the second term describes the rate of oxygen consumption due to degradation, assuming a ﬁrst order reaction, and k is the rate constant for oxygen consumption. For steady state conditions, the rate of oxygen consumption is equal to the oxygen supply by diffusion: D

2 @ ½O2 ¼ k½O2 @2x

ð2Þ

For the boundary conditions [O2] ¼ [O2]0 at x ¼ 0 and q[O2]/qx ¼ 0 at x ¼ l, the solution of equation (2) is (l is half sample thickness): ½O2 ¼

½O2 0 cosh½ðk=DÞ1=2 ðl xÞ cosh½ðk=DÞ1=2 l

ð3Þ

Assuming that all oxygen in the polymer reacts and the oxygen-containing products are not lost by diffusion, equation (3) describes the distribution of oxygen-containing products such as hydroxyl, carbonyl, and nitroxide. For large values of x equation (3) can be approximated as C/C0 ¼ exp [(k/D)1/2 x]

(4)

where C0 and C are the concentrations of oxidation products on the surface and at depth x, respectively. The parameter a ¼ (D/k)1/2, known in the literature as the degradation depth,4,18,19 shows the depth where most (90%) of oxygen-containing products are located. The degradation proﬁle for two values of a, and a sample depth of 4 mm are shown in Figure 2. For a degradation depth comparable with the sample depth, a ¼ 3.4 mm, the degradation proﬁle is essentially homogeneous; this is the case of low k, low oxygen consumption, and low degradation rates. For a degradation depth smaller than the sample depth, a ¼ 0.4 mm in Figure 2, the proﬁle is heterogeneous; this is the DLO regime, with high k, high oxygen consumption, and high degradation rates. The DLO concept implies that lifetimes of polymeric materials deduced from the study of average properties of samples involved in accelerated degradation cannot be used to estimate the durability of polymers in normal exposure;

4

c/c0

Electron Paramag. Reson., 2007, 20, 1–28

0.4 mm 3.4 mm

0

1

2

3

4

Depth / mm

Figure 2 Relative concentration of oxidation products during thermal degradation calculated from equation (3) by assuming [O2] ¼ C and [O2]0 ¼ C0 (see text), as a function of sample depth for a plaque with thickness 4 mm: Nearly homogeneous proﬁle when the degradation depth, a, is 3.4 mm, similar to sample thickness, and heterogeneous proﬁle when the degradation depth, a, is 0.4 mm, smaller than sample thickness

proﬁling methods, which determine the variation of the extent of degradation with sample depth, are needed. The primary motivation for ESR imaging (ESRI) studies was to develop methods for spatially-resolved degradation: to visualize degradation proﬁles and variation of degradation processes within sample depth. We have developed 1D and 2D spectra-spatial ESRI for the study of heterophasic systems such as poly(acrylonitrile-butadiene-styrene) (ABS)20–29 and heterophasic propylene-ethylene copolymers (HPEC) containing bis(2,2,6,6-tetramethyl-4piperidinyl) sebacate (Tinuvin 770) as the stabilizer, and exposed to thermal treatment and UV irradiation.30–32 The HAS-NO provided the contrast necessary in the imaging experiments. The major objectives were to examine polymer degradation under different conditions; to assess the effect of rubber phase, polybutadiene in ABS and ethylene-propylene rubber (EPR) in HPEC, on the extent of degradation; and to evaluate the extent of stabilization by HAS. The repeat units in ABS and the formula of Tinuvin 770 are shown in Figure 3. Recent advances in the ESRI ﬁeld, for instance the choice of pulsed vs continuous-wave (CW) experiments, combined ESR/NMR imaging, progress in resonators, and applications in various disciplines have been described in detail.33–37 The focus of this Chapter is on ESRI experiments and applications that have been used for, and are relevant to, polymeric systems. The experimental details reﬂect the imaging system in the Detroit laboratory.

5

Electron Paramag. Reson., 2007, 20, 1–28 (a) Tinuvin 770, bis(2,2,6,6-tetramethyl-4-piperidinyl) sebacate O H

N

O

OC(CH2)8CO

N

H

(b) Repeat units in ABS polymers -CH2-CHC

-CH2-CH=CHCH2-

-CH2-CH-

Acrylonitrile (AN)

-CH2-CH-

CH=CH2

N

1,4-trans- and cis-1,2 vinyl-Butadiene (B)

Styrene (S)

Figure 3 Tinuvin 770 and repeat units in ABS

This Chapter is organized as follows. In Section 2 we describe ESR spectra in the presence of magnetic ﬁeld gradients, review applications of ESRI to polymeric systems, and describe crucial experiments that led to the development of ESRI methods in our laboratory. Section 3 describes selected experimental details on sample preparation and treatment, and on the determination of the nitroxide intensity proﬁle (by 1D ESRI) and spectral proﬁle (by 2D spectral-spatial ESRI). The ESR spectra of HAS-NO in the heterophasic polymers are described in Section 4. 1D and 2D ESRI experiments are described in Section 5, which includes results for the ABS and HPEC systems, a comparison of ESRI and FTIR methods, and our experience with the effect of microtoming on crystalline polymers. Conclusions and prospects are presented in Section 6.

2

ESR Imaging and Applications to Polymer Systems

2.1 ESR Spectra in the Presence of Magnetic Field Gradients. – ESR spectroscopy can be transformed into an imaging method, ESRI, by measuring ESR spectra in the presence of magnetic-ﬁeld gradients. In ESR imaging experiments the microwave power is absorbed by the unpaired electrons located at point x when the resonance condition is fulﬁlled: n ¼ (g be/h) (Bres þ x Gx)

(5)

In equation (5), n is the microwave frequency, Bres is the resonant magnetic ﬁeld, Gx is the linear magnetic ﬁeld gradient (in Gauss cm1) at x, and the other parameters have their usual meaning. As in NMR imaging, the ﬁeld gradients produce a correspondence between spin location and Bres and allow the encoding of spatial information in the ESR spectra. If the sample consists of two point samples, for example, the distance between the samples along the gradient direction can be deduced if the ﬁeld gradient is known. In this way it is possible not only to verify the existence of the paramagnetic species in a sample, but also ‘‘to tell exactly where the signal came from’’.38 The ESRI methodology

6

Electron Paramag. Reson., 2007, 20, 1–28

is similar to that of NMR imaging (NMRI, or magnetic resonance imaging, MRI). The challenges in the application of the gradient approach to ESRI are numerous. First, higher gradients are needed compared to NMRI, usually 1001000 times larger. Second, the ESR spectra are often complex, with multiple lines due to hyperﬁne interactions and g-value anisotropy; these signals complicate the imaging experiments but in most cases do not add additional information. Third, most systems do not contain stable paramagnetic species on which imaging is based. ESRI experiments are usually performed on paramagnetic transition metal ions, radicals produced by irradiation, or stable nitroxide radicals as dopants; in some experiments triarylmethyl (‘‘trityl’’) radicals are used as probes, because their ESR spectra consists of a single narrow line, of width typically E50 mG in the absence of oxygen. The feasibility of ESRI was ﬁrst demonstrated in 1979 by Hoch and Day, who described the distribution of colour centres in natural diamonds.39 The instrumentation, software, and applications of ESRI have been described in a 1991 monograph40 and updated in recent reviews.33–37 The early efforts described the type and stability of the gradients necessary for speciﬁc ESRI experiments and the software necessary for image reconstruction in spatial and spectral dimensions. These studies also investigated the feasibility of ESRI experiments in a variety of ‘‘phantom’’ samples, and discussed and estimated the spatial resolution. Because of the short relaxation times of the electron spins, most ESRI experiments are performed in the CW mode, unlike MRI which is used in the pulsed mode. In vivo studies are usually performed at lower frequencies, typically below 2 GHz, which can accommodate large samples with high water content. Most experiments in materials science are performed at X band, E9 GHz. Gradients can be applied in the three spatial dimensions, and a spectral dimension can be added by the method of stepped gradients. The widening scope of ESRI studies was highlighted at the 2004 ESR Symposium in Denver,41 with focus on spatially-resolved degradation and software development for applications to polymers, and in vivo studies at 300 and 700 MHz for the detection of local oxygen proﬁles and the study of radical involvement in oxidative diseases. 2.2 ESR Imaging of Polymeric Systems. – While most ESRI efforts are directed to biological applications,42–46 a small number of studies on polymeric systems have appeared. Information on the spatial distribution of paramagnetic molecules deduced from ESRI experiments has been used for measuring macroscopic translational diffusion. Diffusion coefﬁcients, D, of paramagnetic diﬀusants can be deduced from an analysis of the time dependence of the concentration proﬁles along a selected axis of the sample. The determination of D for spin probes in liquid crystals and model membranes, and the effect of polymer polydispersity, have been described in a series of papers by Freed and coworkers.47 In our laboratory the diffusion coefﬁcients of paramagnetic guests in ion-containing polymers, polymer solutions, cross-linked polymers swollen by solvents and self-assembled polymeric surfactants have been determined by

Electron Paramag. Reson., 2007, 20, 1–28

7

1D ESRI.48–53 These papers represent an effort to extract quantitative information from ESRI experiments. Moreover, in some cases these experiments allow the comparison of macroscopic diffusion coefﬁcients in the presence of a concentration gradient (measured by ESRI) with the microscopic D values (measured by pulsed ﬁeld-gradient NMR). Some of these studies have resulted in measurements of diffusion coefﬁcients that can be deduced only by ESRI. An example is the measurement of D at 300 K for nitroxides probes that differ in their hydrophobicity, and were doped in the various phases (micellar, hexagonal, lamellar, and reverse micellar) of the triblock copolymers poly(ethylene oxide)-b-poly(propylene oxide)-b-poly (ethylene oxide), EOmPOnEOm, (commercial name Pluronics).51 The selfassembling is due to the different hydrophobicity of the two blocks, PEO and PPO, in water as solvent. Ionic, neutral and hydrophobic probes select speciﬁc sites in the self assembled system, and these sites are reﬂected in the rate of transport of the probes: in the value of D. The cationic probe 4-(N,N, N-trimethyl)ammonium-2,2,6,6-tetramethyl-piperidine-1-oxyl iodide (CAT1) and the hydrophobic probe 5DSE, the methyl ester of doxylstearic acid (5 indicates the carbon atom to which the doxyl group is attached) exhibited a contrasting transport behaviour in aqueous solutions of Pluronic L64, EO13PO30EO13. Although the molecular masses M are similar, 340 for CAT1 (213 for the cation) and 414 for 5DSE, the corresponding D values at each polymer content are very different. The ratio DCAT1/D5DSE is 35 in the micellar phase (polymer contents 20 and 35 % w/w), 11 in the hexagonal phase (polymer content 50 % w/w), and 3.1 in the mixed La (lamellar) and L2 (reverse micellar) phase (polymer content 80 % w/w). The range of D values measured in this study was 1.0 105 cm2 s1 – 1.0 107 cm2 s1. These results indicate that ESRI is the method of choice for the determination of diffusion coefﬁcients for guests present in low concentrations and located in various regions of selfassembled systems; this conclusion is relevant for drug delivery systems. 2.3 ESR Imaging of Polymers Stabilized by Hindered Amines. – Degradation processes can be studied by ESRI in polymers containing hindered amine stabilizers (HAS); this approach was originally suggested by Ohno, who presented 2D spectral-spatial ESRI images of radicals in polypropylene (PP) containing two different stabilizers, but no detailed analysis.54 The method is based on the formation of stable nitroxide radicals derived from HAS, HASNO, during UV- or thermal treatment. ESRI based on HAS-NO represents an important step in the development of ESRI beyond phantoms, because the nitroxides are part of the system, and participate in, and reﬂect, degradation processes. Detailed studies based on this approach have followed. Lucarini et al. have determined by 1D ESRI the distribution of the nitroxide radicals in UVirradiated PP containing a hindered amine stabilizer.55–57 The spatial variation of nitroxide intensity in the thicker samples irradiated for longer times was explained by the diffusion-limited oxidation (DLO) concept. The DLO regime and high oxidation rates lead to narrow penetration depth of oxygen. Recently,

8

Electron Paramag. Reson., 2007, 20, 1–28

Marek et al. have reported ESRI imaging experiments on polystyrene and polypropylene plaques exposed to thermal ageing and UV irradiation.58,59 As described below, the HAS-derived nitroxides in the heterophasic polymer systems studied in the laboratory perform a triple role. First, they provide the contrast needed in the imaging experiments. Second, they allow the visualization of polymer morphology, based on the detection of two dynamically different components detected in the ESR spectra of the nitroxides, as shown in Section 4; in ABS for instance, the two sites, fast (F) and slow (S), have been assigned to location of nitroxides in butadiene-rich and styrene-acrylonitrile (SAN)-rich domains, respectively. Third, the spatial variation of the ESR spectra of nitroxides (in terms of intensity and line shapes) with treatment time, t, provides detailed information on the extent of degradation in the different microdomains. These experiments make possible nondestructive determination of the nitroxide concentration proﬁles from 1D ESRI, and spectral proﬁles from 2D spectral-spatial ESRI. In these studies the nitroxides, which are the contrast agents, are part of the system; therefore these studies represent the evolution of ESRI techniques beyond phantoms.

3

Experimental Details

3.1 Hardware for ESRI. – An ESR imaging system can be built with small modiﬁcations of commercial spectrometers: gradient coils ﬁxed on the poles of the spectrometer magnet, regulated DC power supplies, and required computer connections. In most systems the software for image reconstruction in the spatial and spectral dimensions must be developed on site. ESR imagers built by Bruker Biospin Co have recently become available for selected applications, mostly in biological applications of ESRI.41 Because the ESR spectrum is measured by scanning the Zeeman magnetic ﬁeld, which is in the z direction, the gradients coils must supply gradients along x (qBz/qx), along y (dBz/dy), or along z (qBz/qz). The gradient coil arrangement that supplies a gradient along the y direction of the sample was the direction of choice in our 1D experiments, because it is along the longest dimension that can be visualized in ESRI experiments. The same ‘‘Figure 8’’ coils can be wired to produce a gradient in the z direction. The coils are cooled by water and protected by a temperature sensor that interrupts the current when the temperature is above a set limit. The coils in our laboratory supply a maximum linear ﬁeld gradient of E320 G cm1 in the direction parallel to the external magnetic ﬁeld (z axis), or E250 G cm1 in the vertical direction (along y), with a constant control current of 20 Amperes applied to each power supply. The coils are positioned so that the zero point of the gradient ﬁeld coincides with the center of the resonator. To calibrate the gradient and check its linearity, a sample consisting of two specks of DPPH at a distance of 10 mm is used; in the presence of gradients, two signals are detected. A straight line obtained by plotting the separation in

9

Electron Paramag. Reson., 2007, 20, 1–28

Gauss between the two signals as a function of the current indicated the linearity of the gradient, as required. This calibration is repeated regularly.

3.2 Intensity Proﬁling from 1D ESRI. – In the general case, the sample contains a distribution of paramagnetic centers and the ESR spectrum in the presence of the gradient is a superposition of signals from paramagnetic centers located at different positions. If the distribution is along a given direction, the intensity proﬁle can be obtained from 1D ESRI experiments, with the magnetic ﬁeld gradient along the direction of the distribution. The two ESR spectra needed to deduce the proﬁle are shown in Figure 4: the spectrum measured in the absence of magnetic ﬁeld gradient, F0(B), and the 1D image, F(B), measured in the presence of the gradient. Mathematically, F(B) is a convolution of F0(B) with the distribution function of the paramagnetic centers. It is essential to note that this convolution is correct only if the ESR line shape has no spatial dependence. This requirement has dictated the conditions for data acquisition in the 1D ESRI study of degradation processes, as described in Section 4 below. The two spectra were measured either at 240 K close to the rigid limit of both spectral component,20–25 or at 340 K in order to reach the motional narrowing regime of both spectral components;30–32 in this way the spatial dependence of the ESR signal was avoided, vide infra. In most cases the 1D images were obtained with a ﬁeld gradient of E200 G cm1. Gradient (G/cm)

0

207

3250

3300

3350

3400

Magnetic Field / G

Figure 4 1D ESRI at 240 K of a cylindrical ABS sample (height E 4 mm, diameter E 4 mm) containing a nitroxide derived from HAS; the polymer was treated at 393 K for 500 h. Top, X-band ESR spectrum recorded in the absence of gradient. Bottom, the 1D image, measured with a magnetic ﬁeld gradient of 207 G/cm. The inset shows the cylindrical sample and the direction of the gradient

10

Electron Paramag. Reson., 2007, 20, 1–28

The proﬁle can be obtained by deconvolution. Various optimization methods can then be applied. The process starts by assuming an initial distribution, which can be described by a set of parameters. Optimization methods use the convolution of this initial distribution function with the experimental spectrum in the absence of gradient in order to calculate the 1D image. The deviation between the calculated and the experimental spectra is then minimized by an optimization procedure. In our initial ESRI studies, the concentration proﬁles of the radicals were deduced by Fourier transform followed by optimization with the Monte-Carlo (MC) procedure.22–26 A disadvantage of this method is the high-frequency noise present in the optimized proﬁles. In more recent publications, the intensity proﬁle was ﬁtted by analytical functions and convoluted with the ESR spectrum measured in the absence of the ﬁeld gradient in order to simulate the 1D image. The best ﬁt was obtained by variation of the type and parameters of the analytical functions chosen (Gaussian or Boltzmann functions, for example) in order to obtain good agreement with the 1D image, and selected by visual inspection.24,25 Lately the genetic algorithm for minimization of the difference between simulated and experimental 1D images has been implemented; this procedure allowed the best ﬁt to be chosen automatically.31,32 A typical genetic algorithm (GA) is patterned after the Darwinian principle of reproduction and survival: creation of the initial population, calculation of the ﬁt to experimental data, selection of the couples, crossover (reproduction) and mutation. The approach and terminology are adopted from biology and resemble fundamental steps in evolution. 3.3 Line-Shape Proﬁling from 2D Spectral-Spatial ESRI. – These ESRI experiments provide the ESR spectrum as a function of a spatial coordinate (sample depth). The data collection consists of projections that examine an object viewed in the H (spectral) and L (spatial) coordinates.40 The angle a is between the L coordinate and the direction of a given projection: spatial information only is obtained when a ¼ 01, and spectral information only when a ¼ 901. The maximum attainable a value in a given experiment, amax, is given in equation (6), tan amax ¼ (L/DH) Gmax

(6)

where L is the sample length, DH is the spectral width, and Gmax is the maximum gradient. Each 2D image was reconstructed from a complete set of projections, collected as a function of the magnetic ﬁeld gradient, using a convoluted back-projection algorithm.21 The number of points for each projection (512– 1024) was kept constant. The maximum sweep width is SWmax ¼ O2DH/ cosamax. For a width DH E 65 G (which is typical for the slow-motional spectral component of HAS-NO present in stabilized polymers), a sample length of 4 mm, and a maximum ﬁeld gradient of 250 G cm1 along the vertical axis, we obtain amax E 601 and SWmax ¼ 169 G. A complete set of data for one image consists of 128–256 projections, taken for gradients corresponding to

Electron Paramag. Reson., 2007, 20, 1–28

11

equally spaced increments of a in the range 01–1801; of these projections, typically 86 (out of 128) are experimentally accessible, and the rest are projections at missing angles (for a in the intervals 601 to 1201). In the ﬁrst reconstruction stage, the projections at the missing angles were assumed to be identical to the projection measured at the largest available angle. In the second stage, the projections at the missing angles were obtained by the projection slice algorithm (PSA) with 2-10 iterations. 23–25,31–32 The spatial resolution is an important parameter in imaging, and can be deﬁned in various ways, as discussed recently;60 the resolution depends on the line width and line shape. The spatial resolution is most commonly expressed as the ratio of the line width to the ﬁeld gradient, dH/G; this definition implies that two signals separated by one line width due to the ﬁeld gradient can be resolved. Typical spatial resolution in our experiments was E100 mm.

3.4 The Polymers Studied. – ABS polymers are usually prepared in two steps. In the ﬁrst step polybutadiene (PB) with the required degree of cross-linking is prepared. In the second step, PB reacts with acrylonitrile and styrene monomers. During this step two processes take place: copolymerization of acrylonitrile-styrene monomers and grafting of the copolymer to polybutadiene. The resulting complex polymeric materials are phase-separated and consist of an acrylonitrile-styrene copolymer (SAN) continuous matrix phase in which PB particles are dispersed.61,62 The properties of ABS can be modiﬁed by variation of the preparation method, size and size distribution of the PB (rubber) particles, cross-link density, the amount of each repeat unit, and the molecular weight of the free SAN. The size of the rubber particles is in the range 0.1–1 mm for emulsion polymerization, and 0.5–5 mm (that can contain occluded SAN) in mass polymerization. Our experiments were performed on ABS containing 10% PB prepared by mass polymerization, and ABS containing 25% PB prepared by emulsion polymerization. Heterophasic propylene-ethylene copolymers (HPEC) consist of crystalline polypropylene (PP) modiﬁed by an elastomeric component, typically ethylenepropylene rubber (EPR),63 and are prepared by polymerization of propylene (P) in the presence of catalysts, and sequential polymerization of a propyleneethylene mixture with the same catalysts.64 The resulting polymeric materials are heterophasic, but the speciﬁc morphology depends on the preparation method and monomer ratio. In scanning electron microscopy (SEM) studies of HPEC, evidence for the presence of ‘‘two phases’’ has been reported: the continuous PP phase, and the dispersed EPR phase.65 The size of the dispersed elastomer particles was found to depend on the polymerization details; for HPEC prepared by sequential polymerization and containing 18 %wt EPR, the particle size measured by SEM was r1 mm. More recent papers have recognized the presence of four phases in HPEC: crystalline PP, amorphous PP, crystalline EPR (predominantly polyethylene, PE), and amorphous EPR.66 Our experiments were performed on HPEC with ethylene (E) content of 25% (notation HPEC1), and HPEC with E content of 10% (notation HPEC2).30–32

12

Electron Paramag. Reson., 2007, 20, 1–28

For the ESR and ESRI experiments, the polymers containing 1–2 % Tinuvin were prepared as 10 cm 10 cm 0.4 cm plaques, which were obtained by injection molding at 483 K. Thermal treatment of the plaques was performed in convection ovens, at 333, 353, and 393 K for ABS samples, and at 393 and 433 K for HPEC samples. UV-irradiation of the plaques carried out in weathering chambers at 338 K with Xe arc that mimics sunlight, or at 318 K with UVA (320–380 nm) or UVB (290–330 nm) sources. For the ESR and ESRI experiments, cylindrical samples with 4 mm in diameter were cut through the plaque thickness at selected time intervals, transferred to 5 mm in diameter ESR sample tube, and placed in the ESR resonator with the symmetry axis along the long (vertical) axis of the resonator and parallel to the direction of the magnetic ﬁeld gradient.

4

ESR Spectra of Nitroxides Derived From Hindered Amine Stabilizers in Heterophasic Polymers

Selected X-band ESR spectra at 300 K of HAS-NO in ABS, for the indicated irradiation times with the Xe source in the weathering chamber, are shown in Figure 5. All spectra, except that corresponding to the longest irradiation time, consist of a superposition of two components, from nitroxides differing in their mobility: a ‘‘fast’’ component (F) with a total width of E32 G, and a ‘‘slow’’ component (S) with a spectral width of E64 G. The corresponding rotational correlation times, tc, are 4 109 s rad1 and 5 108 s rad1, respectively, deduced by simulations of the spectra.20,22 The line shape of the fast component was simulated with a diffusion tilt angle, y, equal to 901.20 The angle y is between the direction of the N–O bond and the axis of the 2p orbital of the unpaired electron. The spectra indicate the presence of nitroxides in two different environments. The fast and slow components are assigned to nitroxides located respectively in low-Tg domains dominated by polybutadiene sequences (Tg E 200 K), and in high-Tg domains dominated by polystyrene (Tg E 370 K) or polyacrylonitrile sequences (Tg E 360 K). This assignment was veriﬁed by spin probe studies of the homopolymers polystyrene, polyacrylonitrile, and polybutadience and of ABS.67 As seen in Figure 5, the relative intensity of the F component as a function of irradiation time increases to a maximum (25%), decreases, and becomes negligible at the longest irradiation time. The decrease of the relative intensity of F with irradiation time is due to the consumption of the HAS-derived nitroxide radicals located in the butadiene-rich domains of the polymer, as butadiene is expected to be more vulnerable to degradation compared to the other repeat units in ABS. The ESR spectrum corresponding to the F component was isolated by subtracting the spectrum of the slow component presented at the top of Figure 5 from one of the composite spectra. After this step it was easy to superimpose the two components, and to reproduce all composite spectra; the relative concentration of each spectral component was then obtained by double integration. After examination of numerous samples we became convinced that the

13

Electron Paramag. Reson., 2007, 20, 1–28

Irradiation Time, h

% Fast

2425

0

Slow (S) 455

24

Fast (F)

141

3300

21

3325

3350

3375

3400

Magnetic Field / G

Figure 5

ESR spectra at 300 K of HAS-NO in ABS (2% HAS) for the indicated irradiation times with a Xe arc in the weathering chamber. Upward and downward arrows point to the signals from the ‘‘fast’’ (F) and ‘‘slow’’ (S) spectral components, respectively. The relative intensity of the F component, %F, at 300 K is indicated

same F and S components are detected in different samples, and the corresponding composite spectra differ only in the relative intensity of the two components. The percentage of the fast component, %F, calculated as described above, is given for all spectra shown in Figure 5. The F and S components were also detected in the ESR spectra at 300 K of thermally treated ABS; however, one component only was visible at 240 K, because both spectral components are close to the ‘‘rigid limit’’ of the nitroxide radical.24,25 A similar situation was encountered for thermally-treated and UV-irradiated HPEC samples.30–32 As nitroxide radicals are not expected to intercalate in crystalline domains, the spectra shown above reﬂect dynamics in the amorphous domains. The major conclusion deduced from the spectra shown in Figure 5 is that the two sites detected for the nitroxides in the heterophasic systems studied can serve as basis not only for describing the morphology of the system, but also to trace the evolution of the nitroxide signal as a function of treatment time, and the time-dependence of the degradation process, in the morphologically diﬀerent domains: in the butadiene-rich and SAN-rich domains in ABS, and in the amorphous PP-rich and EPR-rich domains in HPEC.

14

Electron Paramag. Reson., 2007, 20, 1–28

Preliminary experiments on cut samples have indicated that, as a result of UV and thermal treatments, the intensity ratio of the fast and slow components, [S]/[F], varies with sample depth. This information means that: (i) 1D ESRI experiments must be performed either in the rigid limit (240 K) or the motional narrowing limit (340 K) for the two spectral components; (ii) 2D spectral-spatial ESRI experiments are needed to determine the ESR spectrum as a function of sample depth and the [S]/[F] within sample depth, in a nondestructive way. As described below, 1D and 2D ESRI experiments, together with the determination of the nitroxide concentration and %F as a function of treatment time in whole samples, allowed the determination of the HAS-NO intensity in each morphologically distinct domain as a function of sample depth.

5

1D and 2D Spectral-Spatial ESRI of Heterophasic Polymers Stabilized by Hindered Amines

In this section we will describe selected ESRI experiments performed on thermally treated and UV-irradiated ABS and HPEC systems. The polymer plaques contained 1 or 2% HAS. As seen below, these experiments detected differences in the degradation mechanism depending on the type of treatment and wavelength of the irradiation source, and allowed the visualization of the degradation processes on two length scales: in distinct morphological domains with a resolution of 1–5 mm, and at different depth from the sample side in contact to oxygen with a resolution of E100 mm. The most important conclusions were based on thermal treatment of ABS at 353 and 393 K, and of HPEC at 393 and 433 K; and on ABS irradiated with Xe and UVB sources, and HPEC irradiated with UVA and UVB sources. 5.1 Thermal Treatment. – The concentration proﬁles of HAS-NO along the sample depth for ABS (2% HAS) for the indicated treatment times, t, at 393 K are presented in Figure 6; the proﬁles were deduced by deconvolution of 1D ESR images measured at 240 K. The four proﬁles on the right side are presented with the same maximum height; the proﬁles on the left were normalized by the nitroxide concentration measured in whole sample and are shown for one side of the samples (because of symmetry).24 The evolution from ﬂat proﬁles in the initial stages of thermal treatment (t ¼ 72 h) to spatially heterogeneous proﬁles due to DLO (for t Z 241 h) is clearly seen in Figure 6. The 1D proﬁles in the DLO regime show that the HAS-derived nitroxides are concentrated in the outer layers of the sample. The degradation depth, a, was described in Section 1.3. For the same oxygen diffusion constant, the degradation depth is narrower for more advanced degradation, and can be used to estimate degradation rates. The narrower oxidation depth detected in the case of the polymer containing 2% HAS suggested a higher degradation rate in these samples, compared with the polymer containing 1% HAS. This antiprotective effect of HAS in thermal

15

Electron Paramag. Reson., 2007, 20, 1–28

356 h

241 h

834 h

72 h

356 h

241 h

834 h 72 h 0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Depth / mm

0

25

50 75 100

Depth / %

0

25 50 75 100

Depth / %

Figure 6 Right: 1D concentration proﬁles of HAS-NO in ABS (2% HAS) for the indicated treatment times at 393 K. Left: 1D concentration proﬁles normalized to the corresponding nitroxide concentration in whole samples. Only one side of each (symmetrical) proﬁle is shown. Deconvolution method: optimization by ﬁtting with analytical functions

Spectral Slices

ABS, 834 h at 393 K

Fast / % 3.45 mm

3

2.45 mm 1.45 mm

0

0.45 mm

3 3380

1 De pt 2 h /m

m

3340

3 4

3320

c

eti

gn

Ma

3360 /G ld Fie

3320

3340

3360

3380

Magnetic Field / G

Figure 7 2D spectral-spatial perspective image at 300 K of HAS-NO in ABS (1% HAS) after 834 h of thermal treatment at 393 K, presented in absorption. The corresponding spectral slices in the derivative mode and %F at the indicated depths are shown. In the sample interior (depths 1.45 and 2.45 mm), the nitroxide signal was weak and %F could not be determined

16

Electron Paramag. Reson., 2007, 20, 1–28

treatment is due to the thermal instability of the amino ethers, 4NOR, formed in reaction of nitroxides with alkyl radicals: at 350 K and above the amino ethers decompose, with regeneration of initial products. Moreover, nitroxide radicals at these elevated temperatures are known to be powerful abstractors of hydrogen. ABS, 1% HAS 30

%F

20 241 h 834 h

10 0 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Depth / mm

Spectral proﬁling: % F as a function of sample depth for ABS containing 1% HAS (top) for the indicated treatment times at 393 K. The data were deduced from 200 mm thick virtual (nondestructive) slices in the corresponding 2D ESR images

Figure 8

HPEC1, 168 d at 393 K

Spectral Slices Fast / %

3.10 mm

38

3.00 mm

52 38

0 1 De p

th

2 /m m

3

3380 3360 /G 3340 ld e i 3320 cF eti gn 3300 a M

1.50 mm

48

0.15 mm

38

0.05 mm

3320

3340

3360

3380

Magnetic Field / G

Figure 9 2D spectral-spatial ESRI perspective image at 300 K for HAS-NO in HPEC1 (2% HAS) treated at 393 K for 168 d. The corresponding spectral slices and %F in each slice at the indicated depths are shown in the derivative mode

Electron Paramag. Reson., 2007, 20, 1–28

17

The spectral variation with sample depth was determined from 2D ESRI experiments on thermally-treated ABS containing HAS. The 2D spectralspatial perspective plot for ABS containing 1% HAS and t ¼ 834 h is shown in Figure 7. The ESR intensity is presented in absorption. The perspective plot clearly shows the distribution of the signal intensity, and the negligible signal intensity in the sample interior, as also seen in the concentration proﬁles deduced from 1D ESRI. The 2D images were ‘‘sliced’’ nondestructively to give the ESR spectra at various depths. These spectral slices indicate the spatial variation of the ESR line shapes, and the relative intensity of each spectral component (F and S) as a function of sample depth. The evolution of % F within the sample depth (‘‘spectral proﬁling’’) for ABS (1% HAS) is shown in Figure 8; corresponding proﬁles have been calculated for each treatment time. Since the F component is assigned to nitroxides located in the butadiene-rich domains, spectral proﬁling represents changes in elastomeric properties within the sample depth. Therefore, Figure 8 constitutes elastomer phase proﬁling. The line shape proﬁles clearly indicate the progressive disappearance of the fast component with treatment time. For all treatment times, the minimum and maximum values of % F were higher for ABS containing 1% HAS compared to ABS containing 2% HAS, suggesting a lower rate of loss of elastomeric properties. When the consumption of oxygen in chemical reactions is lower than the amount of oxygen provided by diffusion, oxidation occurs through the entire samples. This regime was observed in ABS samples treated at low temperature, typically 353 K,25 and/or for shorter treatment times at higher temperatures,24 and for HPEC treated at 393 K.31 Under these conditions the oxidative degradation is not limited by oxygen diffusion. Examination of the 1D proﬁles indicated the presence of a outside layer of thickness E100 mm that contained a lower concentration of nitroxide radicals. At 393 K the degradation process is very slow, as clearly seen by the absence of carbonyl peaks in the FTIR data (data not shown); therefore no serious complications due to consumption of HAS as stabilizer are expected. We suggested that the outer layer depleted in nitroxide radicals is due to the loss of HAS; the loss is initially more pronounced in HPEC1 because of the higher E content, which increases the mobility of the additive. The higher nitroxide content in the plaque center is also seen in the 2D spectral-spatial ESRI perspective plot at 300 K shown in Figure 9, for HPEC1 (2% HAS) treated at 393 K.31 We notice the high concentration of HAS-NO through the sample depth, indicating the low nitroxide consumption in the stabilization processes, and therefore a low degradation rate. The DLO regime was detected upon thermal treatment of HPEC1 during 10 d at 433 K, and of HPEC2 for 50 d at 433 K (data not shown). Both ESRI and FTIR experiments suggested a faster degradation rate in HPEC containing 25% E, compared to 10% E. As for the case of thermally-treated ABS, results for HPEC samples suggested that the effect of HAS is antiprotective: the presence and a larger of Tinuvin 770 content in the polymers led to less efﬁcient stabilization.

18

Electron Paramag. Reson., 2007, 20, 1–28

5.2 Photodegradation. – Exposure of polymeric materials to sunlight leads to degradation even at ambient temperatures. In the absence of speciﬁc absorptions, radiation with a wavelength l o 350 nm is expected to penetrate only to a thickness of r50 mm. In the laboratory we have exposed polymer plaques to Xe sources, which closely mimic the sunlight spectrum, and to UVA (maximum energy in the range 320-380 nm) and UVB (maximum energy in the range 290–330 nm) sources, as seen in Figure 10. UVA irradiation mimics the Xe region in the range 300 o l o 400 nm. The visible range of the Xe spectrum can penetrate through the entire samples and can catalyze degradation processes throughout the polymer depth. Because the repeat units in neat ABS and HPEC do not absorb in this range, we must implicate chromophores present as impurities or even HAS and the HAS-derived nitroxides as absorbers in the visible range. Experiments using UVA and UVB sources were performed in order to assess the effect of speciﬁc spectral regions on polymer degradation. In Figure 10 we present the concentration proﬁles of nitroxide radicals in ABS exposed to UVB and Xe irradiation in weathering chambers for the indicated times in hours. The proﬁles, deduced by deconvolution of 1D images

Irradiation time / h

Irradiation time / h

UVB

Xe

0

1

2

3

810

934

471

455

72

70

4

0

1

2

3

4

Depth / mm

Figure 10 1D concentration proﬁles of HAS-NO obtained by deconvolution of 1D ESRI images measured at 240 K, for the indicated UV irradiation times of ABS (2% HAS) in the weathering chamber. Left: UVB irradiation, Gy ¼ 207 G cm1. Right: Xe arc irradiation, Gy ¼ 207 G cm1. Horizontal arrows indicate the irradiated side of the plaque. Deconvolution method: Optimization by the Monte Carlo method

Electron Paramag. Reson., 2007, 20, 1–28

19

measured at 240 K, are spatially inhomogeneous in both cases. In contrast to the symmetric proﬁles deduced in thermally treated samples, the proﬁles shown in Figure 10 indicate a large nitroxide concentrations near the directly irradiated side of the samples, a clear indication the role of oxygen and irradiation in the formation of HAS-NO. The oxidation depths, a, at the directly irradiated side, determined from equation (4), are 77 mm for UVB-irradiation, and 90 mm for Xe irradiation; these values suggest that the degradation of the directly irradiated side is more advanced in the case of UVB irradiation. With increasing UVB irradiation time, nitroxide radicals appear throughout the sample depth. These results were taken as evidence for extensive damage on the irradiated side and slower thermal degradation in the entire sample at the exposure temperature of 318 K. The effect of UVB radiation is in marked contrast with results obtained by exposure of the same samples to a Xe source; after Xe source irradiation, strong signals from nitroxide radicals were detected at both the irradiated side and the opposite side, and their intensity was weak in the sample interior. The more homogeneous distribution of nitroxides during UVB exposure compared to Xe exposure led to conclusion that degradation at the non-irradiated side of ABS samples is more advanced in the case of Xe irradiation due to high intensity of light in the range l4340 nm, that can penetrate through the entire sample. In Figure 11 we present 2D spectral-spatial perspective images of HAS-NO in ABS (2% HAS) UV-irradiated in the weathering chamber by the UVB source for 72 h. The 2D image for samples irradiated with the Xe source is shown in Figure 12, for 70 h of treatment. The ESR intensity is presented in absorption mode. To the right of the 2D images in Figures 11 and 12 we also present ‘‘virtual’’ slices (in the derivative mode) obtained nondestructively at the indicated depths of the sample. The perspective plots in Figures 11 and 12 show very clearly the distribution of the signal intensity, the different effects of the two types of irradiation, and the weaker signal intensity in the sample interior in the Xe-irradiated ABS, as also seen in the concentration proﬁles deduced from 1D ESRI. The spectral slices indicate the line shape variation as well as the relative intensity of each spectral component as a function of depth. At short irradiation times, E70 h, %F is E42% for the Xe arc and E3% for UVB irradiation, indicating that the elastomeric properties at the directly irradiated polymer side are almost lost during UVB irradiation already at initial stage of irradiation. At and near the non-irradiated side, %F is significantly larger, 30% for UVB and 63% for Xe irradiation. For longer irradiation times, the spectra in the directly irradiated layer are similar, and show a negligible intensity of the fast component, a result that points to the loss of the elastomeric properties. Inspection of the ESR spectra and concentration proﬁles indicate major differences between the results obtained for the two UV irradiation sources: the degradation is more advanced on directly irradiated side in the case of UVB irradiation, but the non-irradiated side of polymer samples degrades faster under Xe irradiation. As for the thermally treated HPEC systems, the ESR spectra of the UVAand UVB-irradiated HPEC (whole samples) containing 1% HAS measured at

20

Electron Paramag. Reson., 2007, 20, 1–28

ABS, 2%HAS, 72 h

Spectral Slices Fast / % 30

3.4 mm

31

2.4 mm

29

1.4 mm

3

0.4 mm

UVB

0 De 1 pt h

2 /m m

3 4

Figure 11

3380 3360 G 3340 d/ iel F 3320 c eti 3300 gn Ma

3320 3340 3360 3380 Magnetic Field / G

2D spectral-spatial perspective image of HAS-NO in ABS (2% HAS) after 72 h of UVB irradiation in the weathering chamber, presented in absorption. The corresponding spectral slices and %F in each slice at the indicated depths are shown in the derivative mode. The 2D image was reconstructed from 83 real projections, Hamming ﬁlter width D ¼ 0.45, 2 iterations, L ¼ 4.5 mm, DH ¼ 69 G and were plotted on a 256 256 grid

300 K consist of two components, fast (F) and slow (S), which reﬂect the different dynamics of nitroxide radicals in the amorphous PP and in the rubber phase of the copolymers. The concentration proﬁles of HAS-NO in HPEC samples that were irradiated with UVA and UVB sources for the indicated treatment times are shown in Figure 13 for HPEC1. The proﬁles were normalized according to the total radical concentration in whole samples. For UVAtreated HPEC1, the concentration of HAS-NO increases with treatment time, and the radicals are almost homogenously distributed within the sample depth, with only a slightly higher concentration on the irradiated side (Figure 13A). The proﬁle for UVB-irradiated HPEC1 is different: spatially-heterogeneous degradation is seen both in the low nitroxide concentration for depth Z 1.5 mm from the irradiated side, and in the displacement of the maximum nitroxide concentration to the sample interior as treatment time increases. As for the ABS system, UVB irradiation led to more heterogeneous proﬁles, and therefore to more extensive degradation. In addition, the degradation was faster for HPEC2, which contains the larger amount of the PP component (spectra not given).

21

Electron Paramag. Reson., 2007, 20, 1–28 Spectral Slices ABS, 2%H, 70 h

Fast / % 63

3.4 mm

53

2.4 mm

39

1.4 mm

42

0.4 mm

Xe

0 3380

1

De

pt

h

2 /m

3360 m

3

3340 4

Figure 12

3320

ld

/G

3320 3340 3360 Magnetic Field / G

t

ne

g Ma

ie ic F

3380

2D spectral-spatial perspective image of HAS-NO in ABS (2% HAS) after 70 h of Xe irradiation in the weathering chamber, presented in absorption. The corresponding spectral slices and %F in each slice at the indicated depths are shown in the derivative mode. The 2D image was reconstructed from 83 real projections, Hamming ﬁlter width D ¼ 0.45, 2 iterations, L ¼ 4.5 mm, DH ¼ 70 G and were plotted on a 256 256 grid

UVA

(A) HPEC1

UVB

1440 h

960 h

480 h

480 h

240 h

360 h

96 h

144 h

24 h

24 h

0.0 0.5

1.0 1.5 2.0 Depth / mm

2.5 3.0

(B) HPEC1

0.0 0.5 1.0 1.5 2.0 2.5 3.0 Depth / mm

Figure 13 Normalized concentration proﬁles of HAS-NO in HPEC1 samples (1%HAS) deduced from 1D ESRI experiments at 340 K, for the indicated irradiation times. (A), UVA. (B), UVB. Horizontal arrows indicate the irradiated side of the plaque. Deconvolution method: Optimization by the genetic algorithm

22

Electron Paramag. Reson., 2007, 20, 1–28

The 2D spectral-spatial ESRI contour plot for HAS-NO in UVB-irradiated HPEC1 for t ¼ 480 hours is shown in Figure 14, together with spectral ‘‘virtual’’ slices and corresponding %F at several distances from the irradiated side. The intensity of the fast component increases from the irradiated side towards the sample interior, and is about twice higher on the non-irradiated side compared to the side exposed to the light. The high relative concentration of the F component on the non-irradiated side indicates the low consumption of the nitroxides in stabilization reactions. The ESRI experiments have made possible the visualization of the profound mechanistic differences between UV and thermal degradation. In thermallytreated HPEC systems, for example, the rate of degradation is higher in HPEC1, which contains 25 wt% E. This effect can be explained by the higher diffusion rate of oxygen and reactant mobility at the ageing temperatures (393 and 433 K) in copolymers containing more E. The higher degradation rate deduced in UV-irradiated HPEC2, which contained less E (10%), showed the dominant effect of PP sensitivity at the point of attack, the tertiary carbon, and suggests a different degradation mechanism in thermally-treated and UV-irradiated copolymers, as suggested previously. The effect of HAS content is also different in thermal and UV degradation: increased HAS content leads to a higher rate of degradation for thermally-treated samples; but HAS is effective as a light stabilizer, as seen by comparison of UV and thermally treated samples. UVB-Irradiated HPEC1 480 h

Spectral slices Fast / % 45

3.0 mm

35

1.6 mm

38

0.4 mm

22

0.1 mm

UVB

3320

4 5

3300

ld

3 mm

Fi e

th /

3340

tic

Dep

2

ne

1

/G

3380 3360

M ag

0

3300 3320 3340 3360 3380 Magnetic Field / G

Figure 14 2D spectral-spatial ESRI perspective image and corresponding virtual slices of HAS-NO in UVB-irradiated HPEC1 (1% HAS) during 480 h. %F as a function of depth from the irradiated side is also indicated

Electron Paramag. Reson., 2007, 20, 1–28

23

5.3 Advantages of ESRI Compared to FTIR and to Destructive Methods of Study. – Photodegradation of ABS68–70 and HPEC71,72 systems in the presence of oxygen has also been studied by FTIR spectroscopy. Jouan and Gardette have investigated exposure to radiation with wavelength l 4 300 nm of ABS ﬁlms with thicknesses in the range 42–211 mm, as well as ‘‘packed-multilayers’’, samples composed of several ﬁlms separated by a layer of cardboard that allowed oxygen penetration.68,69 The results have indicated that the extent of degradation depends on the ﬁlm thickness, and that directly exposed layers show more damage compared to the interior layers. As expected, the butadiene component was found to be the most susceptible to degradation.68 Photoacoustic (PA) FTIR spectroscopy has been used to follow non destructively chemical changes in ABS polymers at depths in the range 5–16 mm from the irradiated side. The spectra were measured after exposure of the samples in the interior of cars, Florida exposure, and Xe-arc irradiation in a weathering chamber. The results suggested that the butadiene and acrylonitrile components are degraded more extensively than the styrene component. Infrared microspectroscopy has been used for proﬁling the thermal oxidative products in nitrile rubber; the spatially heterogeneous degradation has been explained by diffusion-limited oxidation (DLO).73 The IR intensity of the acrylonitrile component was not affected by thermal treatment up to 413 K. We have evaluated the potential of ESRI studies by comparing results obtained by ESRI and FTIR in ABS24,74 and HPEC samples.31 The evolution of signals in the carbonyl and butadiene regions, 1650–1800 cm1 and 966 cm1 respectively, in ABS (0%, 1% and 2% HAS) as a function of thermal ageing time at 393 K was measured by attenuated total reﬂectance (ATR) FTIR in an outer layer of thickness 500 mm. Results from the carbonyl region were compared with the total nitroxide concentration in mol/g of sample; and results from the butadiene region were compared with %F in ABS (1 and 2% HAS). The presence and increase of the carbonyl peak and the decrease of the butadiene peak were detected by ATR-FTIR spectroscopy only in the advanced stages of ageing. ESR results indicated major changes in both nitroxide concentration and %F even in the early stages of degradation, for treatment times of 200 h or even less; under these conditions the intensity of the carbonyl peak and the decrease of the butadiene peak detected in the FTIR experiments were negligible. The major advantage of the ESRI approach is therefore the capability to provide details on ageing in the early stages of the process. For example, the evolution from the ﬂat proﬁles to the DLO regime was reﬂected in the 1D ESRI results, and in the depth variation of %F in the 2D ESRI results. Additional details were obtained by examination of the total nitroxide concentration and %F in whole samples. The ESRI and FTIR data are in agreement on the antiprotective effect of HAS in the advanced stage of thermal ageing for both ABS and HPEC systems. The conclusions from the spectroscopic data are in agreement with the visual appearance of the samples: for the same treatment time, considerably more discolouration and shape distortion were seen in samples containing more Tinuvin 770.31

24

Electron Paramag. Reson., 2007, 20, 1–28

The clear advantage of ESRI as a nondestructive method of study was revealed in the effect of microtoming on the relative intensities of the F and S components in HPEC: %F in the ESR spectra at 300 K of thermally-treated HPEC for 107 d at 393 K was compared in two samples. The ﬁrst sample was the typical cylinder of diameter E 4 mm and height also E 4 mm, cut from the plaque after treatment. The same sample was subsequently microtomed into 50 mm slices, and all slices were transferred to the ESR sample tube. In the whole sample %F ¼ 41 2, compared to %F ¼ 29 2 in the microtomed sample. The lowering of the relative intensity of the F component in the amorphous phase is related to the increased ordering and further crystallization in the PP domains. These domains restrain the mobility of nitroxide radicals located in the vicinal amorphous phases, and to the transformation from F-type to S-type nitroxides.30 The extreme separation measured immediately after microtoming was unchanged, but increased over a period of several days by E1 G at 300 K. The increase was detected in all ESR spectra measured in the temperature range 100–300 K. The lowering of %F and increase of the extreme separation were assigned to increased ordering and further crystallization upon microtoming. The results were interpreted in terms of an additional phase, the ‘‘rigid amorphous phase’’,75,76 whose extent and dynamics is reﬂected in the ESR spectra of the nitroxides. This phase is considered to be a part of the amorphous phase that is modiﬁed by the proximity to the crystalline phase. The ESR spectra have great sensitivity because the effect is observed mainly on the spectral component with narrow lines (the F component), and even small changes in the intensity lead to large differences in the heights of the corresponding signals. These conclusions demonstrate that nondestructive methods for polymer degradation are preferable in general, especially in crystalline systems.

6

Conclusions

Polymer degradation and stabilization are challenging topics of great fundamental and technological importance: fundamental, because this reﬂect changes in the properties of polymeric materials due to chemical phenomena that can vary as a function of a complex set of environmental conditions; and technological, because of the increasing sophistication needed in polymer properties. Historically, materials were used long before their properties were fully understood. In recent years, analytical tools such as microscopy, imaging, and computational techniques have made possible the determination of exquisite structural and functional details of materials. The ESRI experiments described in our publications and summarized in this Chapter lead to spatially-resolved information on the effect of treatment conditions, amount of stabilizer, and polymer composition on the degradation rate. In the heterophasic systems studied in our laboratory, ESRI has identiﬁed speciﬁc morphological domains where chemical processes are accelerated. The

Electron Paramag. Reson., 2007, 20, 1–28

25

combination of 1D and 2D spectral-spatial ESRI experiments lead to mapping of the stabilizer consumption on two length scales: within the sample depth on the scale of a few mm, and within morphological domains on the scale of a few mm. ESRI is a nondestructive method for the study of degradation, which is an important advantage, especially for crystalline polymers. The major advantage of ESRI compared with FTIR methods is its sensitivity to early events in the ageing process. Further developments of ESRI methods are expected to be of help in the ultimate goal: accurate predictions of lifetimes for polymeric materials and a better understanding of the environmental factors. The ESRI method requires the presence of a contrast agent, HAS-NO in our work. The implication is that ESRI is an exceptionally sensitive and speciﬁc method for observing degradation in HAS-stabilized polymers, but not in polymers in general; this advantage and this limitation is similar to ESR methods, which are applicable only when radicals are present.

Acknowledgments The development of ESRI methods in the Detroit laboratory was possible by the sustained and generous support of the Polymers Program of the National Science Foundation. Early efforts were also supported by a Founders Fellowship of the American Association of University Women (AAUW) to S. Schlick. We acknowledge with gratitude the collaboration with John L. Gerlock, which was supported by the University Research Program of Ford Motor Company. We are grateful to K. Ohno, P. Eagle, and S.-C. Kweon for getting us started in the ESRI ﬁeld; to J. Pilar and A. Marek from the Institute of Macromolecular Chemistry in Prague, Czech Republic for their major contributions to all aspects of ESRI in the Detroit laboratory; and to K Kruczala, T. Spalek and Z. Sojka of the Jagiellonian University, Cracow, Poland for implementation of the Genetic Algorithm in the ESRI software.

References 1. J.H. O’Donnell, in The Effects of Radiation on High-Technology Polymers, E. Reichmanis and J.H. O’Donnell, (ed), American Chemical Society, Washington, D.C., 1989, Ch. 1, p. 1–13. 2. ‘Handbook of Polymer Degradation’, ed. S.H. Hamid, M.B. Amin, and A.G. Maadhah, Marcel Dekker, New York, 1992. 3. ‘Irradiation of Polymeric Materials: Processes, Mechanisms, and Applications’, ed. E. Reichmanis, C.W. Frank, and J.H. O’Donnell, American Chemical Society, Washington, D.C., 1993. 4. ‘Polymer Durability: Degradation, Stabilization and Lifetime Prediction’, ed. R.L. Clough, N.C. Billingham, and K.T. Gillen, Adv. Chem. Series 249, American Chemical Society, Washington, D.C., 1996. 5. L. Kevan and S. Schlick, J. Phys. Chem., 1986, 90, 1998, and references therein.

26

Electron Paramag. Reson., 2007, 20, 1–28

6. (a) S. Schlick and L. Kevan, J. Am. Chem. Soc., 1980, 102, 4622; (b) M.G. AlonsoAmigo and S. Schlick, Macromolecules, 1987, 20, 795. 7. A. Bosnjakovic and S. Schlick, J. Phys. Chem. B, 2004, 108, 4332. 8. M.K. Kadirov, A. Bosnjakovic and S. Schlick, J. Phys. Chem. B., 2005, 109, 7664. 9. E. Roduner and S. Schlick, in ‘Advanced ESR Methods in Polymer Research’, ed. S. Schlick, Wiley Intercience, Hoboken, NJ, 2006, p. 197–228. 10. J.F. Rabek, Photostabilisation of Polymers Principles and Applications, Elsevier, London, 1990. 11. (a) J.L. Gerlock, D.R. Bauer and L.M. Briggs, Part I, Polym. Deg. Stab., 1986, 144, 53; (b) J.L. Gerlock and T. Riley and D.R. Bauer, ibid. Part II, 73; (c) J.L. Gerlock and D.R. Bauer, ibid. Part III, 97. 12. A. Faucitano, A. Buttafava, F. Martinotti and P. Bortolus, J. Phys. Chem., 1984, 88, 1190. 13. A. Faucitano, A. Buttafava, F. Martinotti and L. Greci, Polym. Deg. Stab., 1992, 35, 211 and references therein. 14. J. Pospisil, Adv. Polym. Sci., 1995, 124, 87. 15. O. Brede, D. Beckert, C. Windolph and H.A. Gottinger, J. Phys. Chem. A, 1998, 102, 1457. 16. (a) K.T. Gillen and R.L. Clough, Polymer, 1992, 33, 4359; (b) J. Wise, K.T. Gillen and R.L. Clough, Radiat. Phys. Chem., 1997, 49, 565. 17. K.T. Gillen and R.L. Clough, in Handbook of Polymer Science and Technology, N.P. Cheremisinoﬀ, M. Dekker, (ed), New York, 1989, Ch. 6, p. 167–202. 18. N.M. Emanuel and A.L. Buchachenko, ‘Chemical Physics of Molecular Destruction and Stabilization of Polymer’, VNU Science Press, Utrecht, Netherlands, 1987. 19. J.C.M. De Bruiijn, in ‘Polymer Durability: Degradation, Stabilization and Lifetime Prediction’, R.G. Clough, N.C. Billingham and K.T. Gillen (ed), Adv. Chem. Series 249, American Chemical Society, Washington, DC, 1996, ch. 36, p. 599–620. 20. M.V. Motyakin, J.L. Gerlock and S. Schlick, Macromolecules, 1999, 32, 5463. 21. K. Kruczala, M.V. Motyakin and S. Schlick, J. Phys. Chem. B., 2000, 104, 3387. 22. M.V. Motyakin and S. Schlick, Macromolecules, 2001, 34, 2854. 23. S. Schlick, K. Kruczala, M.V. Motyakin and J.L. Gerlock, Polym. Degrad. Stab., 2001, 73/3, 471. 24. M.V. Motyakin and S. Schlick, Polym. Degrad. Stab., 2002, 76(1), 25. 25. M.V. Motyakin and S. Schlick, Macromolecules, 2002, 35, 3984. 26. M. Lucarini, G.F. Pedulli, M.V. Motyakin and S. Schlick, Progress Polym. Sci., 2003, 28, 331. 27. S. Schlick and M.V. Motyakin, in Instrumental Methods in Electron Magnetic Resonance, Biological Magnetic Resonance. Vol. 21, C.J. Bender and L.J. Berliner, (ed), Kluwer Academic/Plenum Publishing Corporation, New York, 2004, p. 349–384. 28. M.V. Motyakin and S. Schlick, Polym. Degrad. Stab., 2006, 91, 1462. 29. S. Schlick and G. Jeschke, in Encyclopedia of Polymer Science and Engineering, J.I. Kroschwitz, (ed), Wiley-Interscience, New York NY, 2004, Ch. 9, p. 614–651. 30. K. Kruczala, B. Varghese, J.G. Bokria and S. Schlick, Macromolecules, 2003, 36, 1899. 31. K. Kruczala, J.G. Bokria and S. Schlick, Macromolecules, 2003, 36, 1909. 32. K. Kruczala, W. Aris and S. Schlick, Macromolecules, 2005, 38, 6979. 33. S.S. Eaton and G.R. Eaton, in Specialist Periodical Reports-Electron Paramagnetic Resonance, N.M. Atherton, B.C. Gilbert, and M.J. Davies, (ed), Royal Society of Chemistry, Cambridge, 1996, vol. 15, p. 169–185.

Electron Paramag. Reson., 2007, 20, 1–28

27

34. S.S. Eaton and G.R. Eaton, in Specialist Periodical Reports-Electron Paramagnetic Resonance, B.C. Gilbert, M.J. Davies, and K.A. McLauchlan, (ed), Royal Society of Chemistry, Cambridge, 2000, vol. 17, p. 109–129. 35. D.J. Lurie, in Specialist Periodical Reports-Electron Paramagnetic Resonance, B.C. Gilbert, M.J. Davies, and D.M. Murphy, (ed), Royal Society of Chemistry, Cambridge, 2002, vol. 18, p. 137–160. 36. S. Schlick and K. Kruczala, in ‘Advanced ESR Methods in Polymer Research’, S. Schlick, (ed), Wiley, Hoboken, NJ, 2006, p. 229–254. 37. S. Schlick, in Modern Magnetic Resonance, ed. G.A. Webb, Spinger Verlag, Dordrecht, The Netherlands, 2006, in press. 38. P. Lauterbur, Nature, 1973, 242, 190. 39. M.J.R. Hoch and A.R. Day, Solid State Commun., 1979, 30, 211. 40. ‘EPR Imaging and in vivo EPR’, G.R. Eaton, S.S. Eaton and K. Ohno, (ed), CRC Press, Boca Raton, FL, 1991. 41. ‘EPR Imaging Workshop’, 27th International EPR Symposium, Denver, CO, 1–5 August 2004. 42. H.J. Halpern, C. Yu, M. Peric, E. Barth, D.J. Grdina and B. Teicher, Proc. Natl. Acad. Sci. USA, 1994, 91, 13047. 43. G.M. Rosen, S. Pou and H.J. Halpern, Methods Mol. Biol., 1998, 108, 27. 44. J.L. Zweier and P. Kuppusamy, in Spatially Resolved Magnetic Resonance: Methods, Materials, Medicine, Biology, Rheology, Ecology, Hardware, P. Blu¨mler, B. Blu¨mich, R. Botto and E. Fukushima, (ed), Wiley–VCH, Weinheim, Germany, 1998, Ch. 34, p. 373–388. 45. H.J. Halpern, G.V.R. Chandramouli, E.D. Barth, B.B. Williams and V.E. Galtsev, Curr. Top. Biophys., 1999, 23, 5. 46. C. Mailer, B.H. Robinson and H.J. Halpern, Magnetic Reson. Med., 2003, 49, 1175. 47. D. Xu, E. Hall, C.K. Ober, J.K. Moscicki and J.H. Freed, J. Phys. Chem., 1996, 100, 15856 and references therein. 48. S. Schlick, J. Pilar, S.-C. Kweon, J. Vacik, Z. Gao and J. Labsky, Macromolecules, 1995, 28, 5780. 49. Z. Gao and S. Schlick, J. Chem. Soc. Faraday Trans., 1996, 92, 4239. 50. S. Schlick, P. Eagle, K. Kruczala and J. Pilar, in Spatially Resolved Magnetic Resonance: Methods, Materials, Medicine, Biology, Rheology, Ecology, Hardware, P. Blu¨mler, B. Blu¨mich, R. Botto and E. Fukushima, (ed), Wiley-VCH, Weinheim Germany, 1998, Ch. 17, p. 221–234. 51. E.N. Degtyarev and S. Schlick, Langmuir, 1999, 15, 5040. 52. J. Pilar, J. Labsky, A. Marek, C. Konak and S. Schlick, Macromolecules, 1999 32, 8230. 53. A. Marek, J. Labsky, C. Konak, J. Pilar and S. Schlick, Macromolecules, 2002, 35, 5517. 54. K. Ohno, in EPR Imaging and In vivo EPR, S.S. Eaton, G.R. Eaton and K. Ohno, (ed), CRC Press, Boca Raton FL, 1991, p. 181. 55. M. Lucarini, G.F. Pedulli, V. Borzatta and N. Lelli, Res. Chem. Intermed., 1996 22, 581. 56. M. Lucarini and G.F. Pedulli, Angew. Makromol. Chem., 1997, 252, 179. 57. P. Franchi, M. Lucarini, G.F. Pedulli, M. Bonora and M. Vitali, Macromol. Chem. Phys., 2001, 202, 1246. 58. J. Pospisil, J. Pilar, N.C. Billingham, A. Marek, Z. Horak and S. Nespurek, Polym. Degrad. Stab., 2006, 91, 417.

28

Electron Paramag. Reson., 2007, 20, 1–28

59. A. Marek, L. Kapralkova, P. Schmidt, J. Pﬂeger, J. Humlicek, J. Pospisil and J. Pilar, Polym. Degrad. Stab., 2006, 91, 444. 60. M. Van Kienlin and R. Pohmann, in Spatially Resolved Magnetic Resonance Methods Materials Medicine Biology Rheology Ecology Hardware, P. Blu¨mler, B. Blu¨mich, R. Botto and E. Fukushima, (ed), Wiley-VCH, Weinheim, 1998, Ch. 1, p. 3–20. 61. P.A.M. Steeman, R.J. Meier, A. Simon and J. Gast, Polymer, 1997, 38, 5455. 62. Y. Aoki, A. Hatano, T. Tanaka and H. Watanabe, Macromolecules, 2001, 34, 3100. 63. B. Pukansky, in Polymeric Materials Encyclopedia, J.C. Salamone, (ed), CRC Press, Boca Raton, 1996, p. 6615–6623. 64. E. Albizzati, U. Giannini, G. Collina, L. Noristi and L. Resconi, in Polypropylene Handbook, E.P. Moore, (ed), Hanser Publishers, Munich, 1996, Ch. 2, p. 92. 65. (a) F.M. Mirabella, Polymer, 1993, 34, 1729; (b) F.M. Mirabella, J Polym. Sci. B: Polym. Phys., 1994, 32, 1205. 66. F.M. Mirabella and D.C. McFaddin, Polymer, 1996, 37, 931. 67. B. Varghese and S. Schlick, J. Polym. Sci. B Polym. Phys., 2002, 40, 415. 68. X. Jouan and J.L. Gardette, J. Polym. Sci. A, 1991, 29, 685. 69. X. Jouan and J.L. Gardette, Polym. Degrad. Stab., 1992, 36, 91. 70. R.O. Carter III and J.B. McCallum, Polym. Degrad. Stab., 1994, 45, 1. 71. P. Delprat, X. Duteurtre and J.-L. Gardette, Polym. Degrad. Stab., 1995, 50, 1. 72. B.D. Penington, R.A. Ryntz and M.W. Urban, Polymer, 1999, 40, 4795. 73. M. Celina, J. Wise, D.K. Ottesen, K.T. Gillen and R.L. Clough, Polym. Degrad. Stab., 1998, 60, 493. 74. J.G. Bokria and S. Schlick, Polymer, 2002, 43, 3239. 75. S.Z.D. Cheng and B. Wunderlich, Macromolecules, 1988, 21, 789. 76. B. Wunderlich, Thermal Analysis, Academic Press, Boston, 1990.

Advances in Spin Trapping BY J.-L. CLE´MENT AND P. TORDO Universite´ de Provence, Faculte´ de St Jeroˆme Laboratoire SREP, Case 521, CNRS-UMR 6517 Av Esc. Normadie Niemen, 13397 Marseille, Cedex 20, France

1

Introduction

Various reactive free radicals and subsequent oxidant species are continuously formed in living systems.1,2 Generally, the concentration level of these free radicals and oxidant species is controlled by antioxidants such as superoxide dismutase (SOD), glutathione peroxidase (GPX), catalase, vitamins C, and E. However, under certain conditions such as ischemia-reperfusion injury or inﬂammation, the overproduction of radicals has been linked to cell injury and various physiopathologies. For example, during normal mitochondrial respiration, electrons may be released prior to oxidative phosphorylation and subsequently react with molecular oxygen to form the superoxide radical.1 This species is then converted to H2O2 by mitochondrial SOD and the level of H2O2 is kept relatively low through the action of mitochondrial GPX and cytosolic catalase. However, during ischemia-reperfusion injury, the mitochondria can produce an excess of H2O2 that can react with the metalloproteins in the electron transport chain, e.g. heme and iron-sulfur proteins, to form protein radicals.2 In some instances, living systems have learnt to handle free radicals, using them as mediators in regulatory physiological processes.3–5 Due to their implication in various physiological and pathophysiological processes, different approaches have been developed to detect and identify relevant biological radicals. Among these approaches, spin trapping coupled with EPR is the most popular6–9 and reviews have been published in the ﬁeld of radical damage to proteins,10–13 general oxidative stress and spin trapping,14–25 in vivo spin-trapping methods,26,27 spin-trapping of oxygen radicals in plants28 and the use of spectroscopic probes for detection of ROS (Reactive Oxygen Species).29 The aim of this review is to summarize recent developments in spin trapping. Owing to the huge amount of data published on this topic, our purpose will be mainly focused on spin trapping of superoxide radical. We will also report on recent methods that have been developed to push the limits of spin trapping coupled with EPR, and on the rationalization of spin trapping by theoretical calculations. Electron Paramagnetic Resonance, Volume 20 r The Royal Society of Chemistry, 2007 29

30

2

Electron Paramag. Reson., 2007, 20, 29–49

Basic Principles

The spin-trapping technique involves the addition of a reactive radical across the double bond of a diamagnetic spin trap to form a much more stable radical, a radical adduct, which will accumulate in the medium and can then be examined with EPR. The spin-trapping technique uses the unusual stability of nitroxide radicals; thus nitroso compounds and nitrones are the most widely used spin traps (Scheme 1). Trapping of alkyl radicals with nitroso compounds is efﬁcient. Persistent adducts with the radical moiety attached to the nitrogen atom are obtained and their EPR spectra contain important information on the identity of the trapped radical. However, due to poor water solubility, light and thermal sensitivity and very short life-time of spin adducts formed with oxygenand sulphur-centred radicals, the use of nitroso compounds is limited in biological media. The polar nitronyl function (4C ¼ N1–O) confers to nitrones a significant solubility in water, and this can be modulated by the presence of hydrophobic or hydrophilic substituents. Furthermore, persistent adducts usually result from the trapping of oxygen- or sulphur-centred radicals with nitrones. For these reasons, biological spin-trapping studies have been mostly developed with nitrones like the very popular DMPO (R2 ¼ R3 ¼ Me) and PBN (R4 ¼ Me) (Scheme 1). However, DMPO and PBN are poor traps of superoxide and alkylperoxyl radicals and during the last decade a considerable number of novel nitrones have been prepared and tested as spin traps.

R1

N

FR

R1

R3

FR

O

O Nitroso compounds R2

N

R2

FR N

R3

O Cyclic nitrones

FR N

H

O R4

R4

O

N C H

H N C

FR O

FR

Linear nitrones Scheme 1 Spin trapping by nitrones and nitroso compounds

3

New Nitrones for Spin Trapping

Tordo and co-workers ﬁrst showed30,38 that the replacement of a C-5 methyl group of DMPO with either a diethoxyphosphoryl- [(EtO)2P(O)] or an ethoxycarbonyl group [EtOC(O)] leads to nitrones (DEPMPO 2 and EMPO 17

Electron Paramag. Reson., 2007, 20, 29–49

31

respectively) that offer important advantages over DMPO, especially with regard to the persistence of the superoxide radical adduct. These authors observed the same effect when they replaced a methyl of the PBN tertiarybutyl group with a diethoxyphosphoryl group (PPN 71).51 After these pioneering studies, a large series of DEPMPO, EMPO and PPN analogues were prepared and tested for spin trapping (Scheme 2). Reductive cyclisation, with NH4Cl or CH3COOH in the presence of zinc, of the appropriate phosphorylated g-nitroaldehyde, provides a direct route to DEPMPO and its analogues. DEPMPO and analogues can also be obtained by oxidation of their amine precursor that is usually prepared either by aminophosphorylation of 5-chloropentan-2-one or phosphorylation of commerciallyavailable 2-methylpyrroline. Various phosphites are commercially available or may be easily prepared,63 thus making this last route particularly ﬂexible. DEPMPO analogues, 3-10,30–34 11,35 12,35 51,46 52,47 have been prepared according to either one or the other of these different approaches and all these nitrones, apart from 52, exhibited-spin trapping properties similar to those of DEPMPO. DIPPMPO 3,31 is a crystallisable analogue of DEPMPO with a partition coefﬁcient Kp of 2.1 and a slightly longer superoxide radical adduct half-life (21 min. vs 18 min. for DEPMPO-OOH in same conditions). The polyether arm bound to phosphorus should confer good solubility both in water and in lipid environment to nitrone 10.34 For nitrones 1135, 1235 and 1332, the electronegativity of the phosphorus group has been varied by combining alkoxyl and alkyl substituents onto the phosphorus atom. Compared with DEPMPO, the lifetime of the corresponding superoxide radical adducts decreases when the electronegativity of the phosphorus group decreases. Furthermore, for all kinds of spin adducts, the smaller is the phosphorus group electronegativity, the smaller is the phosphorus coupling. A DMPO analogue 14,34 in which one C-5 methyl has been replaced with a triﬂuoromethyl group, has been prepared by reduction of the corresponding g-nitroaldehyde. As for the DEPMPO series, introduction of an electronegative group at C-5 induces an increase in the superoxide radical adduct half-life (5 min. vs 50 sec. for DMPO-OOH under same conditions). For all adducts formed with various radicals and 14, the EPR lines were split into a quartet due to coupling with the three ﬂuorine atoms. Villamena et al. prepared 1536 (AMPO), an amide analogue of EMPO. The half-life time of AMPO-OOH in phosphate buﬀer at pH 7.2 was about 8 min. Stolze et al. have made a significant commitment to the synthesis and investigation of the spin trapping properties of the EMPO analogues 18-3539,40 and 37-4642,43 (Scheme 2). These nitrones are obtained by Michael addition on acrolein of a a-nitroester, R2(R1)C(H)NO2, (R1 ¼ CO2R; R and R2 represent various primary, secondary or tertiary alkyl groups), followed by reductive cyclisation. The nitrones synthesized cover a wide range of lipophilicity. Very persistent superoxide radical adducts (t1/2 4 18 0 ) were reported for 22-27, 29. EMPO analogues (53-62) substituted at C-3 or C-4 by a methyl group have been prepared and diastereoisomers separated. The superoxide

32

Electron Paramag. Reson., 2007, 20, 29–49

radical adducts obtained with some diastereoisomers of the nitrones 57-6248 have very impressive half-lifes (t1/2 4 45 min). TMINO, 63,49 was obtained by pyrolysis of the stable nitroxide TMIO. This trap gives very stable adducts with carbon-centred radicals, however it failed to trap superoxide radical. The amphiphilic PBN analogues, 64-70, with a polar head linked to the phenyl group and a hydrophobic tail attached to the t-Bu moiety, have been synthesized by Pucci et al.,50 and have been used in biological studies. Poly PPN 72-7552 were prepared in order to improve the trapping efﬁciency of radicals. A competitive kinetic study showed that the trinitrone 75 trapped the methyl radical 1.9 times more rapidly than either POBN or DEPMPO at pH 7.2. The superoxide radical adduct of the bisnitrone 74 has an interesting half-life time of 10.6 min. The PBN analogues 76-7953 were designed to be ‘‘anchored’’ at special depth within biological membranes. Bardelang et al. have studied the association of various EMPO analogues (47-49)44 and para-substituted PBNOH (80-84)54 with cyclodextrins. They also investigated the inﬂuence of this association on the trapping of the superoxide radical. The 5,5-diethoxycarbonyl-pyrrolin-N-oxide, 16, has been prepared by Karoui et al.37 and used to trap the superoxide radical. From the outset, a very intense signal of 16-OOH was observed; however, this signal was rapidly accompanied by the 16-OH signal. DFT calculations showed that one face of 16 is easily accessible to incoming radicals, and the unusually high concentration of superoxide radical adduct was ascribed to a high rate constant for radical trapping. This assumption was conﬁrmed by Allouch et al.59 who showed that the trapping rate of superoxide radical with 16, was 3 times higher than that of the parent EMPO. For the linear nitrones 9258 and 10659, the superoxide radical spin trapping rate was not improved by the presence of two ethoxycarbonyl groups. EPPN 8555, EPPyON 10558, and various analogues (86-9056, 9157, 9258, 93-10457, 10659) yield very persistent adducts with radicals. The half-life time of the superoxide radical adducts of 95 and 98 are over 10 min.; however, only carbon-centred spin adducts were obtained in all attempts to trap the hydroxyl radical with these nitrones. The deuterated DEPMPO derivatives, DEPMPO-d2,-d5 and DEPMPO-d7 (2-d2, 2-d5, 2-d7, Scheme 2)60 have been synthesized in order to simplify the EPR spectra of the superoxide and peroxyl radical adducts. Introduction of deuterium allowed the complete assignment of the g-couplings observed in these spectra. 13C-5 labelled DEPMPO has also been prepared and used to trap various radicals. The spin adducts resulting from the trapping of the radicals 13 CH3, 13CH2OH and 13COONa, yielded EPR signals with four resolved bhyperﬁne couplings; their preferred conformation was deduced from these coupling values and compared with the preferred conformation given by molecular mechanics calculations.61 The use of DMPO deuterated on C-3 (1-d2) and DFT calculations have permitted conﬁrmation that the small doublet (0.12 mT) observed in the EPR signal of DMPO-OOH results from a g-H coupling.62 The asymmetry of the

33

Electron Paramag. Reson., 2007, 20, 29–49 R1 R2

N O

34

Electron Paramag. Reson., 2007, 20, 29–49 R4

R3

R1 N

R2

O

N O

R1

Y H

OH OH OH HO

N

OH OH

O

O

OH

O

O HN

OH

R1

Y H N O

X

35

Electron Paramag. Reson., 2007, 20, 29–49 R1 R6

R2

R5

R3 R4

O R1

C H

OH

N

O R1 C H

N

R2

O R1 Z Y

X

C H

N

R2

36

Electron Paramag. Reson., 2007, 20, 29–49 OD D EtO P EtO D3C

N

O EtO EtO P H

O DEPMPO d5

H3C

D D

N

ODDD D EtO EtO P H

D3C

H

H313C

N

H

O

O

DEPMPO d7

DEPMPO 13C

O DEPMPO d2

N

O EtO P EtO

DD

N O

Scheme 2

Nitrones developed for spin-trapping applications

DMPO-OOH spectrum can be nicely reproduced assuming a chemical exchange between two conformers.

4

Spin Trapping Studies of Superoxide Radicals

The success of an in vivo spin trapping experiment is often limited by the poor resistance of spin adducts to biological reducing species. The nitrone-OOH species is a very fragile adduct susceptible to reduction into nitrone-OH or EPR-silent species. Some selected methodologies have been elaborated with the aim of solving this problem. In addition a discussion of the determination of trapping rate constants (kt) of superoxide radical with nitrones and the decay rate constant (kd) of the resulting adducts are discussed later. 4.1 Use of Cyclodextrins. – Cyclodextrins are versatile biocompatible macrocycles based on 1,4-glucopyranose units, with a truncated cone shape, which present external hydrophilic sites and an internal hydrophobic cavity allowing the complexation of a wide variety of compounds in aqueous solution.64 A number of authors have reported that nitroxides may be hosted into the cyclodextrin cavities65 and Rassat et al.66,67 have shown that reduction of 2,2,5,5-tetramethyl-pyrrolidin-1-oxyl (TEMPO) by ascorbate monoanion is prevented by its association with cyclodextrins. The half-life time of DEPMPO-OOH is close to 15 min in phosphate buﬀer, but decreases to 1-2 min in vivo. Karoui et al.68 have shown that the half-life of the spin adduct formed during the trapping of superoxide radical with either DEPMPO or DMPO, is increased 7-fold upon addition of RM-b-CD (randomly methylated

37

Electron Paramag. Reson., 2007, 20, 29–49 OH OH O2.-

O N

O

N

CH O2H

CH

R R OH O N

OH O N

CH

R Scheme 3

CH O2H

R Spin trapping with PBN analogues in the presence of CD

b-cyclodextrin). Moreover, under such conditions the rate of reduction of DEPMPO-OOH by ascorbate monoanion was also decreased significantly. The same trend was observed for the trapping of superoxide radicals with PBN.69 Further studies devoted to PBN and PBNOH analogues gave information on the geometry and the stoichiometry of cyclodextrin/nitrone and cyclodextrin/ adduct inclusion complexes (Scheme 3).54,70 4.2 Formation of Diastereoisomer Spin Adducts. – When the nitrone used to trap radicals contains a chiral centre, most radical additions result in the formation of two diastereoisomers (Scheme 4). Generally, these diastereoisomers have different EPR signatures allowing their characterization.31,38,40,42,43,60,71–73 After addition of methyl and hydroxymethyl radicals to DEPMPO, Khramtsov et al.74,75 reduced the resulting nitroxide spin adducts to hydroxylamines, and were able to characterize two diastereoisomers by 31P NMR. In some instances, the diastereoisomeric spin adducts show similar EPR coupling constants and are not easily characterized with EPR. Thus, the EPR signal resulting from the trapping of hydroxyl radical with DEPMPO has been assigned to the trans DEPMPO-OH diastereoisomer [HO trans to (EtO)2P(O)].76 However, using accurate spectral simulations, Culcasi et al.77 showed that hydroxyl radical addition on b-phosphorylated cyclic nitrones leads to both cis and trans DEPMPO-OH diastereoisomers. These authors established that the cis/trans ratio depends significantly on the mode of generation of the spin adduct (true HO. spin trapping or reduction of the corresponding hydroperoxyl nitroneOOH) and allows the determination of its origin.

38

Electron Paramag. Reson., 2007, 20, 29–49

R1

R1

*

N

R2 O

Hβ

.

FR

* R2

R1

H N

*

O trans adduct

FR

+

FR

* R2

N

* Hβ

O cis adduct

Scheme 4 Formation of cis/trans diastereoisomeric adducts from a chiral cyclic nitrone

4.3 Kinetic Studies. – 4.3.1 Spin Trapping of Superoxide Radical. The superoxide radical can undergo very fast reactions with various biochemicals or enzymes (with SOD the rate constant is about 109 M1s1) whilst the rate constant for the addition of this radical onto the nitrones routinely used for spin trapping is slow (typically kt o 100 M1s1). Therefore, in many instances, the spin trapping of superoxide radical is very slow, particularly during in vivo experiments. Knowledge of the rate constant for generation (kt) and decay (kd) of a spin adduct is therefore of crucial importance in the optimisation of spin trapping experiments, helping in the elucidation of the formation and decay mechanisms of spin adducts and facilitating the design of more efﬁcient traps. Determination of kt is usually achieved through competition kinetic experiments.78 Trapping of superoxide radical with the nitrone under study is conducted in the presence of a competitor, a known superoxide radical scavenger (noted SCV) with a well characterized rate constant, kSCV, (Scheme 5). When the nitrone concentration is high enough to trap out all the superoxide radicals produced, the spontaneous dismutation of superoxide radical (eqn. 3) can be neglected. In addition, it is assumed that the spin-adduct decay (eqn. 4) is negligible during the course of the reaction. With the above assumptions and assuming a steady state concentration of superoxide radical, equations 5 and 6, in which V and v represent rate of spin trapping in the absence and in the presence of competitor, respectively, can be written. A plot of V/v vs [SCV]/[N] gives a straight line whose slope is equal to kt/ kscv. Reported values of kt, for the formation of superoxide radical spin adduct with various nitrones are listed in Table 1. It is noteworthy that for DMPO, DEPMPO and BMPO, a wide range of kt values have been reported. Lauricella et al. have recently shown that for the competitive kinetic model, omission of important reactions such as superoxide radical dismutation (eqn. 3), scavenger consumption and spin adduct concentration decay (eqn. 4) results in significant overestimation of rate constant kt.78 These authors have developed a new kinetic approach avoiding the use of competing scavengers that allows the simultaneous evaluation of kt and kd (Table 2).79 Villamena et al. have calculated the positive charge on the carbon (C-2) of the nitronyl function of various cyclic nitrones.80 However, as shown in Table 3, no correlation exists between the charge on C-2 and kt. Tanigushi has shown that the trapping of radicals with DMPO is very sensitive to steric factors.81 Karoui et al.37 suggested that the increase of kt for 16 results from favourable steric factors.

39

Electron Paramag. Reson., 2007, 20, 29–49 (1)

-. O2 +

(2)

-. O2

(3)

-. O2

SA

d[SA]/dt = k t [O2-.][N]

+

SCV

SCV'

d [SCV']/dt = kscv [O 2-.][SCV]

+

-. O2

+ 2H +

kd

SA

(4) (5)

N

H2O 2 + O2

Epr si lent

-d O2-./dt = k t [O2-.][N] + kscv [O2-.][SCV] -dO2-./dt

(6)

d[SA]/dt

=

V/v = 1 + k scv [SCV] / kt [N]

Scheme 5 Trapping of superoxide radical with nitrone N, in the presence of a superoxide radical scavenger SCV and kinetic analysis

Table 1

Reported rate constants for spin trapping of superoxide radical by various nitrones

Nitrone

kt (M1s1)

DMPO 1 DEPMPO 2 EMPO 17 MeMPO 18 nBuMPO 20 iPrMPO 22 BMPO 36 50 EPPN 85 EPPyON 105

17086 5080 20.187 1088 2.489 2.079 5880 1576 3.9579 0.5389 74.545 10.979 72.987 81.345 73.545 7745 59.587 780 o 386 0.2489 76.545 0.0279 0.3379

4.3.2 Decay of Superoxide Radical Adducts. Three superoxide radical generating systems: HX/XO, KO2/DMSO and light irradiation of riboﬂavin have been routinely used for spin-trapping experiments. In order to study the decay of the superoxide radical spin adduct, production of superoxide radical can be stopped either by adding a large amount of SOD, or by cessation of light exposure for the riboﬂavin system. Then, the kinetic decay curve is obtained by monitoring the intensity of a pure EPR line of the spin adduct. Replacement of a C-5 methyl of DMPO by an electron-withdrawing group such as P(O)(OEt)2,30 -CO2Et38, -CONH2,36 -CF3 34 resulted in a significant increase of the apparent half-life of the corresponding superoxide radical adduct. The same trend was observed on replacement of a methyl of the PBN tertiarybutyl group.55,82 No additive effect was observed with the introduction of a second electron-withdrawing group.37,59,83 However, as already mentioned, in the case of 16, the presence of two electron-withdrawing groups resulted in a 3 fold increase in the rate constant for addition of superoxide radical, compared with the parent EMPO.59

40

Table 2

Electron Paramag. Reson., 2007, 20, 29–49

Reported decay rate constants of various nitrone superoxide radical adducts

Nitrone

Apparent t1/2 (concentration of nitrone)a

DMPO 1 DEPMPO 2

77s79 (125 mM) 54s80 (100 mM) 1831 (20 mM) 15.536 (25 mM) 14.876 (100 mM) 14.380 (100 mM) 1172 (60 mM) 10.279 (100 mM) 2131 (20 mM) 4.834 (50 mM) 8.336 (25mM) 21.845 (50 mM) 1979 (10 mM) 1879 (30 mM) 9.936 (25 mM) 9.279 (200 mM) 8.639 (100 mM) 20.845 (50 mM) 5.639 (100 mM) 8.945(50 mM) 1.139 (130 mM) 18.840 (40 mM) 17.345 (50 mM) 26.339 (40 mM) 22.045 (50 mM) 15.739 (40 mM) 980 (100 mM) 16.143 (20 mM) 23.845 (50 mM) 4742 (40 mM) 25.143 (20 mM) 11.548 (20 mM) 4748 (20 mM) 5548 (20 mM) 5.956 (20 mM) 4.879 (50 mM) 5.7957 (20 mM) 10.6857 (20 mM) 7.2857 (20 mM) 2.479 (50 mM)

DIPMPO 3 CF3MPO 14 AMPO 15 EMPO 17 MeMPO 18 nBuMPO 20 iPrMPO 22 sBuMPO 23 BMPO 36 EPhPO 43 50 BEMPO-3 54 cis-4,5-EMPtPO 55 cis-3,5-EDPO 58 trans-3,5-EDPO 58 trans-3,3-DiPPO 62 EPPN 85 EPPyN-2 93 EPPyN-3 97 EPPyN-4 101 EPPyON 105 a

In minutes, except where indicated otherwise.

Table 3

Net atomic charges (Mulliken) on C-2 calculated at the B3LYP/631G(d) level37

Nitrone

Net charge

kst (M1s1)

DMPO 1 EMPO 17 DECPO 16 DEPMPO 2

þ0.052 þ0.061 þ0.069 þ0.074

2.079 10.979 31.159 3.9579

Nsanzumuhire et al.47 showed that for cyclic nitrones, the apparent half-life of superoxide radical adducts depends on the geometry of their preferred conformers. Characteristics in phosphate buﬀer (pH 7–7.3) of the superoxide adduct formed with DEPMPO and the cis diastereoisomers of 4-PhDEPMPO (51) and 3-PhDEPMPO (52) are shown in Scheme 6.46,47 The DEPMPO-OOH and cis-4-PhDEPMPO-OOH adducts have identical half-life times, and exhibit very similar coupling constants, indicating that for both adducts the geometry of the major conformer is almost the same. On the other hand, the coupling constants and half-life time value of cis-3-Ph

41

Electron Paramag. Reson., 2007, 20, 29–49 Ph

Ph (EtO)2(O)P H 3C

H N

(EtO)2(O)P

OOH

O

H3C

H N

(EtO)2(O)P

H N

H3C

OOH

t1/2 (min)

15.5

2

14.5

AHβ (mT)

<1.11>a

1.81

<1.17>a

AN (mT)

<1.31>a

1.36

<1.29>a

AP (mT)

<5.02>a

3.13

<5.32>a

a

OOH

O

O

: mean values

Scheme 6 Characteristics of superoxide radical adducts of DEPMPO, cis-3-PhDEPMPO and cis-4PhDEPMPO

Ph

OOH

Hβ R

Hβ +. N O -

OOH +. N O R = -P(O)(EtO)2

R

AHβ = 1.81 mT

AHβ = 1.11 mT

AP = 3.13 mT

AP = 5.02 mT

t1/2 = 2 min

t1/2 = 15 min

Scheme 7 Main structural features of DEPMPO-OOH and cis-4-PhDEPMPO-OOH adducts

DEPMPO-OOH are very different from those of DEPMPO-OOH and cis-4PhDEPMPO-OOH. For cis-3-PhDEPMPO-OOH, the large value of AHb and the small value of AP indicate pseudo-axial C-Hb and pseudo-equatorial C–P bonds. This geometry favours abstraction of the b-hydrogen and facilitates the decay of cis-3-PhDEPMPO-OOH via dismutation (Scheme 7). For the DEPMPO-OOH adduct the C-OOH bond eclipses the p orbitals of the aminoxyl group, giving rise to a stabilizing hyperconjugative interaction between the s*C-OOH and pN–O. orbitals. Stolze et al. have also observed a dramatic inﬂuence of substituents on the half-life time of 3,5-EDPO and its derivatives.48 In a series of acyclic nitrones (EPPN), they observed that increasing the bulkiness of the ester group does not improve the life time of the superoxide radical adducts and they suggested that due to freedom of rotation, the ester group exerts less protection. Using a spin trap for NOd, Villamena et al.84 showed, as predicted theoretically,85 that NOd is formed during the degradation of of DMPO, DEPMPO and EMPO superoxide radical adducts.

42

Electron Paramag. Reson., 2007, 20, 29–49 10%

(EtO)2(O)P * H3C

Hβ

N

ROO

Hγ (EtO)2(O)P

.

* γH3C

O

Hγ Hγ

N

Hγ Hβ * OOR

Hγ (EtO)2(O)P * γH3C

+

Hγ OOR

N

* Hβ

O

O

90%

Hγ Hγ

trans diastereoisomer

cis diastereoisomer

Scheme 8 Radical addition of peroxyl radicals to DEPMPO

a

⇓⇓

⇓⇓

⇓

⇓

⇓

⇓⇓

DEPMPO-OOt-Bu

⇓⇓

⇓

b narrow

narrow

narrow

broad

broad

broad

•

•

•

c

DEPMPO d7-OOt-Bu

•

•

•

d

Figure 1

(a): EPR spectrum of DEPMPO-OOt-Bu, (b): simulation, (c): EPR spectrum of DEPMPO d7-OOt-Bu, (d): simulation, : unidentiﬁed species

4.4 Chemical Exchange in Alkylperoxyl and Superoxide Radical Adducts of bPhosphorylated Nitrones. – As a result of the presence of the C-5 chiral carbon, addition of either a superoxide radical or an alkylperoxyl radical to DEPMPO can yield two diastereomeric adducts (Scheme 8). The spectra of the two diastereoisomers are superimposed to give the observed EPR signal. The signal obtained during the trapping of t-BuOOd is shown in Figure 1. The major signal (90%) in this spectrum has been attributed to the trans diastereoisomer ( + ). This diastereoisomer exhibits a dramatic alternating linewidth60,76,90 which was attributed to a chemical exchange between two conformers. The authors of these studies have suggested that this alternating line width results from a hindered rotation around the O–O peroxyl bond. This phenomenon was observed only for adducts of ROOd radicals (R ¼ H, alkyl). Rotation around the O–O bond induces ﬁve-membered ring conformational changes and generates exchanging conformers having significantly different b-hydrogen and phosphorus-atom couplings. A satisfactorily computed spectrum was obtained assuming a two-site chemical exchange (Figure 1b).91,92

Electron Paramag. Reson., 2007, 20, 29–49

43

In the case of DEPMPO-OOH signal, a satisfactorily simulation of the coupling pattern resulting from long-range couplings required the consideration of all the seven g-hydrogen splittings. This pattern could not be properly calculated considering the coupling with 6 g-hydrogens of an epoxy structure resulting from the degradation of DEPMPO-OOH adduct, as suggested by Dikalov et al.93 A less complex spectrum was obtained when t-BuOOd was trapped with DEPMPO-d7 (Figure 1c). Upon 2H substitution, EPR spectra of both diastereoisomers (k cis, + trans) were clearly distinguishable. The minor signal attributed to the cis diastereoisomer exhibits 12 equivalent EPR lines without further Hg splitting. Furthermore, for the trans diastereoisomer signal, the alternate linewidth effect was clearly observable. The chemical exchange observed for the superoxide radical adduct of DEPMPO was also considered to explain the asymmetrical pattern of DMPO-OOH signal.62

5

Theoretical Calculations

Quantum mechanical calculations have been performed to understand the spin trapping behaviour of various substituted cyclic nitrones.36,80,85,94–97 These studies have mainly employed the DFT approach, and concentrated on the evaluation of the trapping efﬁciency and adduct stability of the most popular nitrones (DMPO, DEPMPO, DIPPMPO, EMPO, BMPO) which are commonly used in in vivo as well as in vitro applications. The thermodynamics of spin trapping by these nitrones with biologically relevant radicals (superoxide, hydroperoxyl, hydroxyl, methyl, thiyl, and nitric oxide) were also investigated.36,94,95,97 In the gas phase, the order of increasing favourability of radical reaction with nitrones is NOd o O2d o HOOd o HSd o H3Cd o HOd based on DG calculations, with a similar trend predicted for the aqueous phase. The inertness of nitrones toward NOd can be ascribed to the endoergic reaction parameters, although transition states in both the aqueous and gas phases could be obtained for NOd addition to the investigated nitrones. The DFT approach has also been used to predict the isotropic hyperﬁne coupling constants arising from the N, Hb, 17O and Hg nuclei of DMPO-OOH including explicit interactions with surrounding water molecules.62,85 The calculated hyperﬁne coupling constants were in good agreement with experimental values, and predicted the existence of a long-range Hg coupling close to 0.15 mT, resulting from a W-arrangement of a hydrogen atom bound to C-3. These results show that the DMPO-OOH EPR signal cannot be attributed to the superimposition of individual EPR spectra associated with different conformers of DMPO-OOH.98 Cle´ment et al. have shown that the long range g-H coupling disappears upon deuterium substitution of the two C-3 hydrogens.92 Like for other superoxide radical adducts of pyrroline N-oxides, the asymmetry of the DMPO-OOH signal can be attributed to a chemical exchange between two conformers. This assumption has been conﬁrmed by an EPR temperature study of DMPO-OOH.93

44

6

Electron Paramag. Reson., 2007, 20, 29–49

Immuno Spin Trapping of Proteins

Mason and co-worker have developed a new technique, named ‘‘immuno-spin trapping’’, in which EPR spin trapping of protein-derived radicals is coupled with an immuno assay. After a protein radical has been trapped by DMPO, oneelectron oxidation can convert the spin adduct to a EPR silent DMPO-protein adduct. This DMPO-protein adduct can subsequently be detected using polyclonal antibodies (anti-DMPO). This immuno technique is speciﬁc and very sensitive; it was initially used to study the self-peroxidation of haem- and myoglobin.99–104 Subsequently, the technique has been used to investigate the formation of a protein radical on SOD induced by H2O2,105 nitrite-mediated oxidation of haemand myoglobin106,107 and Cu(II)/H2O2-induced radical damage to DNA.108

7

Separation and Identiﬁcation of Spin Adducts

When several radicals of the same type are generated, their trapping generates spin adducts exhibiting similar EPR spectra which are superimposed to form an EPR signal which is difﬁcult to interpret. The identiﬁcation of the trapped radicals therefore becomes complex and tedious and misinterpretations often occur. In order to separate the different components of complex spin-adduct mixtures, LC (HPLC) and EPR have been combined in an on-line-device. Furthermore, this system can be complemented with a combination of LC and MS or electrospray-ionisation tandem mass spectroscopy (ESI MS/MS). These combinations permit the separation, by LC, of radical adducts with very similar EPR spectra, followed by EPR detection, and ﬁnally accurate MS identiﬁcation. Frequently, hydroxylamines and nitrones resulting from disproportionation or oxido-reductive processes are detected in the MS experiments, rather than the initial parent paramagnetic adducts. A number of combinations between these different techniques are possible. In some cases, even with complex mixtures, preliminary LC separation has not been performed. Domingues et al. carried out a direct MS study of the reaction of HO. with tryptophan in either the absence, or presence, of a spin trap.109 They demonstrated the presence of various products including dimers and hydroxylated dimers. Hardy et al. reported the spin trapping of lipid peroxyl radicals, LOOd, with the phosphorylated nitrone 9.33 Tuccio et al.110 analyzed the fragmentation pathways of DEPMPO-OH, -CH3 and -CH2OH by means of ESI MS/MS. A LC/ELISA combination has been use by Guo et al.111 to identify an EPRsilent end-product of DMPO radical-adduct, obtained when microperoxidase11 was reacted with DTT. The ﬁnal analysis of this EPR-silent adduct was carried out with NMR. LC/EPR and LC/MS are particularly valuable for in vivo studies where high sensitivity is needed and where endogenous reductants often transform the EPR adducts to silent compounds. Qian et al. have identiﬁed and quantiﬁed in vitro adducts of POBN from DMSO/HOd oxidation, by examining the bile of

Electron Paramag. Reson., 2007, 20, 29–49

45

rat treated simultaneously with DMSO and POBN.112 Only a very weak signal was detected by LC/EPR, however, three other adducts reduced into hydroxylamines were detected by LC/MS. A typical application of LC/EPR and LC/MS is the study of the polyunsaturated fatty acid (PUFA) oxidation under aerobic conditions, using Fe21/H2O2 or soybean lipoxygenase, to mimic the complex radical-chain oxidation of cell membranes.113–118 Whilst examination of this system by EPR spin trapping affords a very simple EPR spectrum, more than thirteen different adducts were identiﬁed after LC separation. The spin trapping of the peroxyl PUFA (LOO.) was reported with 9 and with DMPO.33,115 However, DMPO has been reported to be unable to trap peroxyl radicals in water.119–122 An alternative approach related to these techniques is the use of TEMPO to scavenge transient alkyl radicals to form stable alkoxylamines that can be identiﬁed by MS analysis.123

8

Conclusion

About one hundred novel nitrone spin traps have been synthesized in the past few years and tested for spin trapping. 30–62,124 For some of these traps the superoxide radical spin-adduct exhibited very impressive in vitro life times. However efforts are still needed to design novel spin traps having a high rate constant for the trapping of the superoxide radical. New methods to study superoxide radical trapping kinetics are needed that should take in account the most important reactions involved both in the generation and decay of the spin adduct. Due to the high selectivity of LC, and the high sensitivity of MS, the on-line coupling of spin trapping with LC/EPR or LC/MS appears to be a very powerful tool to investigate biological free-radical processes. With the development of the highly sensitive immuno-spin trapping technique, the utility of the spin-trapping technique has been expanded to the detection of EPR-silent end products resulting from the trapping of protein and DNA radicals by DMPO, using a DMPO-antibody. Thus, the spin-trapping technique is no longer dependent solely on EPR as a detection method.

References 1. B. Halliwell and J.M. Gutteridge in Free Radicals in Biology and Medicine, Oxford University Press, Oxford, UK, 1999. 2. H. Sies, in Oxidative stress: Oxidants and antioxidants, Academic Press, New York, 1991. 3. Y.J. Suzuki, H.J. Forman and A. Sevanian, Free Radic. Biol. Med., 1997, 22, 269. 4. T. Finkel, Curr. Opin. Cell Biol., 1998, 10, 248. 5. W. Dro¨ge, Physiol. Rev., 2002, 82, 47. 6. P. Tordo, in Electron Paramagnetic Resonance, B.C. Gilbert, M.J. Davies and D.M. Murphy, (eds), Royal Society of Chemistry, Cambridge, UK, 1998, 16, 116.

46

Electron Paramag. Reson., 2007, 20, 29–49

7. M.J. Davies and G.S. Timmins, in Electron Paramagnetic Resonance, B.C. Gilbert, M.J. Davies and D.M. Murphy, (eds), Royal Society of Chemistry, Cambridge, UK, 2000, 17, 1. 8. M.J. Davies, in Electron Paramagnetic Resonance, B.C. Gilbert, M.J. Davies and D.M. Murphy, (eds), Royal Society of Chemistry, Cambridge, UK, 2002, 18, 47. 9. M.J. Burkitt, in Electron Paramagnetic Resonance, B.C. Gilbert, M.J. Davies and D. M. Murphy, (eds), Royal Society of Chemistry, Cambridge UK, 2004, 19, 33. 10. M.J. Davies and C. Hawkins, Biochim. Biophys. Acta., 2001, 1504, 196. 11. M.J. Davies and C. Hawkins, Free Radic. Biol. Med., 2004, 36, 1072. 12. M.R. Gunther, Free Radic. Biol. Med., 2004, 36, 1345. 13. O. Augusto, M.G. Bonini and D.F. Trindade, Free Radic. Biol. Med., 2004, 36, 1224. 14. R.A. Floyd and K. Hensley, Ann. N. Y. Acad. Sci., 2000, 899, 222. 15. L.J. Berliner, V. Khranstov, H. Fujii and T.L. Clanton, Free Radic. Biol. Med., 2001, 30, 489. 16. N. Khan and H. Swartz, Mol. Cell. Biochem., 2002, 234/235, 341. 17. Y.Z. Fang, S. Yang and G. Wu, Nutrition, 2002, 18, 872. 18. R.P. Mason and M.B. Kadiiska, in Biological Magnetic Resonance, L.J. Berliner, (ed), Kluwer Academic/Plenum Publisher, 2003, 18, 309. 19. B. Kalyanaraman, E.A. Konorev, J. Joseph and J.E. Baber, in New Comprehensive Biochemistry, C.A. Rice-Evans and R.H. Burdon, (eds), Elsevier, 1994, 28, 333. 20. B. Halliwell and M. Whiteman, Br. J. Pharmacol., 2004, 142, 231. 21. S. Venkataraman, F.Q. Schafer and G.R. Buettner, Antioxid. Redox Signal., 2004, 6, 631. 22. F. Villamena and J.L. Zweier, Antioxid. Redox. Signal., 2004, 6, 619. 23. C.J. Rhodes, Ann. Rep. Prog. Chem. Sect. C, 2004, 100, 149. 24. M. Kopa´ni, P. Celec, L. Danisˇ ovicˇ, P. Michalka and C. Biro´, Clin. Chim. Acta, 2006, 364, 61. 25. M. Valko, C.J. Rhodes, J. Moncol, M. Izakovic and M. Mazur, Chem. Biol. Interact., 2006, 160, 1. 26. G.R. Buettner and R. P. Mason, in Critical Reviews of Oxidative Stress and Aging, R.G. Cutler and H. Rodriguez (eds), World Scientiﬁc Publishing Co, 2003, 1, 27. 27. G.S. Timmins and K.J. Liu, in Biological Magnetic Resonance, L.J. Berliner (ed), Kluwer Academic/Plenum Publisher, 2003, 18, 285. 28. G. Bacˇic´ and M. Mojovic´, Ann. N. Y. Acad. Sci., 2005, 1048, 230. 29. G. Bartosz, Clin. Chim. Acta, 2006, in press. 30. C. Fre´javille, H. Karoui, B. Tuccio, F. Le Moigne, M. Culcasi, S. Pietri, R. Lauricella and P. Tordo, J. Chem. Soc. Chem. Commun., 1994, 15, 1793. 31. F. Chalier and P. Tordo, J. Chem. Soc. Perkin Trans. 2, 2002, 2110. 32. J.-L. Cle´ment, C. Fre´javille and P. Tordo, Res. Chem. Intermed., 2002, 28, 175. 33. M. Hardy, O. Ouari, L. Charles, J.-P. Finet, G. Iacazio, V. Monnier, A. Rockenbauer and P. Tordo, J. Org. Chem., 2005, 70, 10426. 34. P. Tordo et al., unpublished results. 35. K. Shioji, S. Tsukimoto, H. Tanaka and K. Okuma, Chem. Lett., 2003, 32, 604. 36. F.A. Villamena, A. Rockenbauer, J. Gallucci, M. Velayutham, C.M. Hadad and J.L. Zweier, J. Org. Chem., 2004, 96, 7994. 37. H. Karoui, J.-L. Cle´ment, A. Rockenbauer and P. Tordo, Tetrahedron Lett., 2004, 45, 149. 38. G. Olive, A. Mercier, F. Le Moigne, A. Rockenbauer and P. Tordo, Free. Radic. Biol. Chem., 2000, 28, 403.

Electron Paramag. Reson., 2007, 20, 29–49

47

39. K. Stolze, N. Udilova and H. Nohl, Biol. Chem., 2002, 383, 813. 40. K. Stolze, N. Udilova, T. Rosenau, A. Hoﬁnger and H. Nohl, Biol. Chem., 2003, 384, 493. 41. H. Zhao, J. Joseph, H. Zhang, H. Karoui and B. Kalyanaraman, Free Radic. Biol. Med., 2001, 31, 599. 42. K. Stolze, N. Udilova, T. Rosenau, A. Hoﬁnger and H. Nohl, Biochem. Pharmacol., 2005, 69, 297. 43. K. Stolze, N. Rohr-Udilova, T. Rosenau, A. Hoﬁnger, D. Kolarich and H. Nohl, Bioorg. Med. Chem., 2006, 14, 3368. 44. D. Bardelang et al., accepted for Org. Biomol. Chem. 45. P. Tsai, K. Ichikawa, C. Mailer, S. Pou, H.J. Halpern, B.H. Robison, R. Nielsen and G. Rosen, J. Org. Chem., 2003, 68, 7811. 46. M. Hardy, F. Chalier, J.-P. Finet, A. Rockenbauer and P. Tordo, J. Org. Chem., 2005, 70, 2135. 47. C. Nsanzumuhire, J.-L. Cle´ment, O. Ouari, H. Karoui, J.-P. Finet and P. Tordo, Tetrahedron Lett., 2004, 45, 6385. 48. K. Stolze, N. Rohr-Udilova, T. Rosenau, A. Hoﬁnger, R. Statmu¨ller and H. Nohl, Biochem. Pharmacol., 2005, 69, 1351. 49. S.E. Bottle, G.R. Hanson and A.S. Micallef, Org. Biomol. Chem., 2003, 1, 2581. 50. G. Durand, A. Polidori, O. Ouari, P. Tordo, V. Geromel, P. Rustin and B. Pucci, J. Med. Chem., 2003, 46, 5230. 51. A. Zeghdaoui, B. Tuccio, J.-P. Finet, V. Cerri and P. Tordo, J. Chem. Soc. Perkin Trans. 2, 1995, 2087. 52. V. Roubaud, H. Dozol, C. Rizzi, R. Lauricella, J.-C. Bouteiller and B. Tuccio, J. Chem. Soc. Perkin Trans. 2, 2002, 958. 53. A. Hay, M. J. Burkitt, C. M. Jones and R. C. Hartley, Arch. Biochem. Biophys., 2005, 435, 336. 54. D. Bardelang, J.-L. Cle´ment, J.-P. Finet, H. Karoui and P. Tordo, J. Phys. Chem., 2004, 108, 8054. 55. V. Roubaud, R. Lauricella, J.-C. Bouteiller and B. Tuccio, Arch. Biochem. Biophys., 2002, 397, 51. 56. K. Stolze, N. Udilova, T. Rosenau, A. Hoﬁnger and H. Nohl, Biochem. Pharmacol., 2003, 66, 1717. 57. K. Stolze, N. Udilova, T. Rosenau, A. Hoﬁnger and H. Nohl, Biochem. Pharmacol., 2004, 68, 185. 58. A. Allouch, V. Roubaud, R. Lauricella, J.-C. Bouteiller and B. Tuccio, Org. Biomol. Chem., 2003, 1, 593. 59. A. Allouch, V. Roubaud, R. Lauricella, J.-C. Bouteiller and B. Tuccio, Org. Biomol. Chem., 2005, 13, 2458. 60. J.-L. Cle´ment, J.-P. Finet, C. Fre´javille and P. Tordo, Org. Biomol. Chem., 2003, 1, 1591. 61. J.-L. Cle´ment, PhD thesis, Universite´ d’Aix-Marseille III, 1998. 62. J.-L. Cle´ment, N. Ferre´, D. Siri, H. Karoui, A. Rockenbauer and P. Tordo, J. Org. Chem., 2005, 70, 1198. 63. P. Pages, M.-R. Mazieres, J. Bellan, M. Sanchez and B. Chaudret, Phosphorus Sulfur Silicon Relat. Elem., 1992, 70, 205. 64. J. Szejtli, Chem. Rev., 1998, 98, 1743. 65. J.-J. Schneider, F. Hacket and V. Rudiger, Chem. Rev., 1998, 98, 1755. 66. A. Jeunet, B. Nickel and A. Rassat, New J. Chem., 1986, 10, 123. 67. J. Martinie, J. Michon and A. Rassat, J. Am. Chem. Soc., 1975, 97, 1818.

48

Electron Paramag. Reson., 2007, 20, 29–49

68. H. Karoui, A. Rockenbauer, S. Pietri and P. Tordo, Chem. Commun., 2002, 3030. 69. H. Karoui and P. Tordo, Tetrahedron Lett., 2004, 45, 1043. 70. D. Bardelang, A. Rockenbauer, H. Karoui, J.-P. Finet and P. Tordo, J. Phys. Chem., 2005, 109, 10521. 71. S. Dikalov, P. Tordo, A. Motten and R.P. Mason, Free Radic. Res., 2003, 37, 705. 72. K. Stolze, N. Udilova and H. Nohl, Free Radic. Biol. Med., 2000, 29, 1005. 73. J.-L. Cle´ment, S. Barbati, C. Fre´javille, A. Rockenbauer and P. Tordo, J. Chem. Soc. Perkin Trans. 2, 2001, 1471. 74. V. Khramtsov, L.J. Berliner and T.L. Clanton, Magn. Reson. Med., 1999, 42, 228. 75. L.J. Berliner, V. Khramtsov, T. L. Clanton and H. Fujii, Curr. Top. Biophys., 2002, 26, 21. 76. C. Fre´javille, H. Karoui, B. Tuccio, F. Le Moigne, M. Culcasi, S. Pietri, R. Lauricella and P. Tordo, J. Med. Chem., 1995, 38, 258. 77. M. Culcasi, A. Rockenbauer, A. Mercier, J.-L. Cle´ment and S. Pietri, Free Radic. Biol. Med., 2006, 40, 1524. 78. R.P. Lauricella, J.-C.H. Bouteiller and B.N. Tuccio, Phys. Chem. Chem. Phys., 2005, 7, 399. 79. R. Lauricella, A. Allouch, V. Roubaud, J.-C. Bouteiller and B. Tuccio, Org. Biomol. Chem., 2004, 2, 1304. 80. F.A. Villamena and J.L. Zweier, J. Chem. Soc., Perkin Trans., 2, 2002, 1340. 81. H. Tanigushi and K.P. Maiden, J. Am. Chem. Soc., 1999, 121, 11875. 82. V. Roubaud, R. Lauricella, B. Tuccio, J.-C. Bouteiller and P. Tordo, Res. Chem. Intermed., 1996, 22, 405. 83. G. Olive, F. Le Moigne, A. Mercier, A. Rockenbauer and P. Tordo, J. Org. Chem., 1998, 63, 9095. 84. E.L. Locigno, J.L. Zweier and F.A. Villamena, Org. Biomol. Chem., 2005, 3, 3220. 85. F.A. Villamena, J.K. Merle, C. Hadad and J.L. Zweier, J. Phys. Chem., 2005, 109, 6089. 86. S. Golstein, G.M. Rosen, A. Russo and A. Samuni, J. Phys. Chem. A, 2004, 108, 6679. 87. J. Weaver, P. Tsai, S. Pou and G.M. Rosen, J. Org. Chem., 2004, 69, 8423. 88. E. Finkelstein, G.M. Rosen and E.J. Rauckman, J. Am. Chem. Soc., 1980, 102, 4994. 89. A. Keszler, B. Kalyanaraman and N. Hogg, Free Radic. Biol. Med., 2003, 35, 1149. 90. S. Barbati, PhD thesis Universite´ d’Aix-Marseille I, 1997. 91. A. Rockenbauer and L. Korecz, App. Magn. Res., 1996, 10, 29. 92. A. Rockenbauer, Mol. Phys. Reports, 1999, 26, 117. 93. S. Dikalov, J. Jiang and R.P. Mason, Free Radic. Res., 2005, 39, 825. 94. F. Villamena, C.M. Hadad and J.L. Zweier, J. Phys. Chem. A, 2003, 107, 4407. 95. F. Villamena, C.M. Hadad and J.L. Zweier, J. Am. Chem. Soc., 2004, 126, 1816. 96. F. Villamena, J.K. Merle, C.M. Hadad and J.L. Zweier, J. Phys. Chem. A, 2005, 109, 6083. 97. F. Villamena, C.M. Hadad and J.L. Zweier, J. Phys. Chem. A, 2005, 109, 1662. 98. G.M. Rosen, A. Beselman, P. Tsai, S. Pou, C. Mailer, K. Ichikawa, B.H. Robinson, R. Nielsen, H.J. Halpern and A.D. MacKerell, J. Org. Chem., 2004, 69, 1321. 99. C.D. Detweiler, L.J. Deterding, K.B. Tomer, C.F. Chignell, D. Germolec and R.P. Mason, Free Radic. Biol. Med., 2002, 33, 364. 100. D.C. Ramirez, Y.R. Chen and R.P. Mason, Free Radic. Biol. Med., 2003, 34, 830. 101. Y.Y. He, D.C. Ramirez, C.D. Detweiler, R.P. Mason and C.F. Chignell, Photochem. Photobiol., 2003, 77, 585.

Electron Paramag. Reson., 2007, 20, 29–49

49

102. L.J. Deterding, D.C. Ramirez, J.R. Dubin, R.P. Mason and K.B. Tomer, J. Biol. Chem., 2004, 279, 11600. 103. R.P. Mason, Free Radic. Biol. Med., 2004, 36, 1214. 104. C.D. Detweiler, O.M. Lardinois, L.J. Deterding, P.R. Ortiz de Montellano, K.B. Tomer and R.P. Mason, Free Radic. Biol. Med., 2005, 38, 969. 105. D.C. Ramirez, S.E. Gomez Mejiba and R.P. Mason, Free Radic. Biol. Med., 2005, 38, 201. 106. K. Nakai and R.P. Mason, Free Radic. Biol. Med., 2005, 39, 1050. 107. A. Kesler, R.P. Mason and N. Hogg, Free Radic. Biol. Med., 2006, 40, 507. 108. D.C. Ramirez, S.E. Gomez Mejiba and R.P. Mason, Nat. Methods, 2006, 3, 123. 109. M.R.M. Domingues, P. Domingues, A. Reis, C. Fonseca, F.M.L. Amado and A.J.V. Ferre-Correia, J. Am. Soc. Mass Spectrom., 2003, 14, 406. 110. B. Tuccio, R. Lauricella and L. Charles, Int. J. Mass Spectrom., 2006, 252, 47. 111. Q. Guo, G. Gao, S.Y. Qian and R.P. Mason, Chem. Res. Toxicol., 2004, 17, 1481. 112. S.Y. Qian, M.B. Kadiiska, Q. Guo and R.P. Mason, Free Radic. Biol. Med., 2005, 38, 125. 113. S.Y. Qian, K.B. Tomer, G.-H. Yue, Q. Guo, M.B. Kadiiska and R.P. Mason, Free Radic. Biol. Med., 2002, 33, 998. 114. H. Iwahashi, Free Radic. Res., 2003, 37, 939. 115. A. Reis, M.R.M. Domingues, F.M.L. Amado, A.J.V. Ferre-Correia and P. Domingues, J. Am. Soc. Mass Spectrom., 2003, 14, 1250. 116. S.Y. Qian, Q. Guo and R.P. Mason, Free Radic. Biol. Med., 2003, 35, 33. 117. S.Y. Qian, G.-H. Yue, K.B. Tomer and R.P. Mason, Free Radic. Biol. Med., 2003, 34, 1017. 118. A. Reis, M.R.M. Domingues, F.M.L. Amado, A.J.V. Ferre-Correia and P. Domingues, Biomed. Chromatogr., 2006, 20, 109. 119. Q. Guo, S.Y. Qian and R.P. Mason, J. Am. Soc. Mass Spectrom., 2003, 14, 862. 120. S.I. Dikalov and R.P. Mason, Free Radic. Biol. Med., 1999, 27, 864. 121. S.I. Dikalov and R.P. Mason, Free Radic. Biol. Med., 2001, 30, 187. 122. C.M. Jones and M.J. Burkitt, J. Chem. Soc. Perkin Trans. 2, 2002, 2044. 123. P.J. Wright and A.M. English, J. Am. Chem. Soc., 2003, 125, 8655. 124. Y.-K. Xu, Z.-W. Chen, J. Sun, K.Liu, W. Chen, W. Shi, H.-M. Wang and Y. Liu, J. Org. Chem., 2002, 67, 7624.

Site-Directed Spin-Labelling (SDSL) Applications in Biological Systems BY JIMMY B. FEIX AND CANDICE S. KLUG Department of Biophysics, Medical College of Wisconsin, 8701 Watertown Plank Rd., Milwaukee, WI 53226, USA

1

Introduction

This chapter will review the application of site-directed spin labelling (SDSL) to proteins, peptides, substrates and RNA, focusing on literature for the period 2003–2005. Applications of SDSL in biological systems in the literature through 2002 were specifically reviewed in Volume 24 of Biological Magnetic Resonance1 and other recent reviews also exist.2–10 The application of SDSL in biological systems has experienced remarkable growth since its introduction in 1989.11,12 This has been based on significant improvements in the speed and reliability of molecular biology techniques for cloning and site-speciﬁc mutagenesis, advances in protein expression and puriﬁcation methods, and continuing developments in electron paramagnetic resonance (EPR) instrumentation and methodology. Nitroxide scanning, in which relatively large stretches of a protein sequence are systematically mutated to cysteine and spin labelled for analysis of secondary structure and local dynamics, has become a standard approach in SDSL. We deﬁne SDSL as the selective incorporation of a nitroxide spin label into a speciﬁc, experimentally-chosen site. Although it has been most commonly applied to proteins or peptides, this methodology has also been applied to nucleic acids and small substrates. We have not reviewed the extensive literature utilizing lipid-analogue spin labels. SDSL methodology typically focuses on a set of core techniques that includes changes in motion (to assess conformational changes and protein-protein interactions), accessibility to paramagnetic relaxation agents (to determine secondary structure and membrane insertion), and analysis of spin-spin interactions (to determine aggregation state and spatial proximity of labelled sites). The most significant recent advances in SDSL are in the area of distance measurements between selectively-introduced pairs of nitroxides, which in principle can be combined with a knowledge of secondary structure elements to generate a complete 3D structure. Electron Paramagnetic Resonance, Volume 20 r The Royal Society of Chemistry, 2007 50

Electron Paramag. Reson., 2007, 20, 50–72

2

51

Soluble Proteins

The motion of the spin-label side-chain is commonly used as a site-speciﬁc probe of side-chain interactions and thus allows the monitoring of both large and small conformational changes in protein structure. The correlation between the motion of the spin-label side-chain with local structure and dynamics has been previously analyzed for a-helical sites (e.g. ref. 13), and a detailed analysis of b-sheet sites is presented by Lietzow and Hubbell14 using the cellular retinol binding protein (CRBP) labelled with the methanethiosulfonate spin label (MTSL). The detailed analysis of numerous spin-labelled sites within CRBP demonstrates that the motion of interior strand sites are largely dependent on the degree of strand twisting, the identity of the nearest-neighbour residues, and whether the neighbouring site is located at the non-hydrogen-bonded or hydrogen-bonded position. These factors contribute directly to the variety of dynamic modes observed for spin-label side-chains located in interior b-sheets, whereas the residues located on edge strands have much more simple motional features and are more similar to those located on the outer surfaces of a-helices. During nitrate/nitrite signal transduction in E. coli, the presence of nitrate/nitrite modulates the transcription and expression of 40 genes involved in the respiratory and fermentation pathways. The cytoplasmic regulatory response protein NarL is phosphorylated at aspartate 59 in response to these chemicals, which triggers a conformational change within the protein and enables interaction with Narregulated promoters of downstream genes. The mechanism of this phosphorylation-induced conformational change in NarL was studied by SDSL methods and revealed that the phosphorylation itself triggers a hinge-like opening of the N- and C-terminal domains of NarL.15 This change was elucidated by the introduction of both single and double labels to monitor the changes in motion and distance between unphosphorylated and phosphorylated states of the regulator. The coiled-coil is an important structural motif for a number of regulatory and nucleic acid-binding proteins. SDSL was used to examine the structure of a model peptide from the E. coli osmoregulatory transporter, ProP.16 Analysis of spin-spin interactions indicated that this 30-residue peptide formed an antiparallel coiled-coil, while a 35-residue model peptide was shown to form a parallel coiled coil. Phosphoinositide (PI) 3-kinases regulate a wide variety of important physiological processes, and SDSL has been used to examine the structure of a PI 3-kinase regulatory domain.17 Analysis of mobility and inter-nitroxide distances for strategically selected mutants provided experimental evidence supporting the validity of a coiled-coil structure for the central region of the regulatory domain generated by homology modelling. Differences in backbone motion have been examined by SDSL methods18 in the DNA-binding transcription factor GCN4, in which the DNA-binding Nterminus comprises a gradient of backbone motion and the C-terminus forms a coiled coil leucine zipper motif with slower motion as determined by nuclear magnetic resonance (NMR) studies.19 Spin labels were introduced at more than 50 sites within GCN4, and mobility analysis revealed a motional gradient similar to that seen in the NMR studies in the resting state, indicating that the backbone

52

Electron Paramag. Reson., 2007, 20, 50–72

motion at the DNA-binding sites is reﬂected in the spectral lineshapes. Upon DNA binding the backbone motion was slowed but a gradient still existed, with the DNA-binding site having the slowest motion, as expected, thus showing that backbone dynamics can be reﬂected in the EPR spectra for sites with similar sidechain motion. DnaK is a molecular chaperone expressed in E. coli that uses the energy derived from ATP hydrolysis to assist protein folding. Seven-residue peptide substrates for DnaK were synthesized, each containing two cysteine residues, and the distances and mobilities of the introduced spin-label side-chains were compared when associated with DnaK in its ATP- and ADP-bound states.20 No structural changes were observed, suggesting that the ATPase activity of DnaK is not associated with structural changes in the peptide-binding pocket. The a-crystallins are members of the small heat-shock family of proteins. It is proposed that they bind and sequester non-native protein states, thus inhibiting aggregation. To characterize the ability of aA-crystallin to bind non-native protein states, a series of spin-labelled, destabilized mutants of T4 lysozyme (T4L) were developed.21 T4L mutants were selected with DG of unfolding values separated by B4 kJ mol1 and spin-labelled at a non-perturbing, surfaceexposed site. In solution, the EPR spectra of the T4L mutants indicated rapid tumbling, as reﬂected by a relatively narrow lineshape. However, upon binding to aA-crystallin a strongly-immobilized component was observed, allowing ready determination of the relative populations of bound and free T4L. This provided a method for detailed analysis of binding afﬁnity as a function of substrate (i.e., T4L) stability. These destabilized T4L mutants were also used to examine the chaperone activity of aB-crystallin.22 In this case, serine to aspartate mutants of aB-crystallin were constructed to mimic the effects of phosphorylation and the spin-labeled T4L mutants were used to determine binding afﬁnities and stoichiometries as a function of substrate stability. SDSL has also been used to study directly the stability of bB1-crystallin.23 In this case, a spin label introduced at its single wild-type cysteine gave a spectrum reﬂecting substantial motional restriction in the native state, and became significantly more mobile upon unfolding by urea or guanidine hydrochloride. Denaturation curves were used to examine the effects of deamidation, an important post-translational modiﬁcation, on the DG of unfolding. Force generation in muscle results from the cyclic interaction between myosin and actin. Current models propose that force is generated by reorientation of the myosin head, and that this is ampliﬁed by myosin regulatory light chain (RLC). This mechanism was investigated using a spin label rigidly attached to the RLC, which was then reconstituted into skeletal muscle ﬁbers.24 Outer hyperﬁne splittings approached the rigid limit, and orientation was examined by variation of the ﬁber orientation with respect to the magnetic ﬁeld. Global analysis revealed one population of spins with the label z-axis within 151 of the ﬁbre axis and a second, less-ordered population. The strong immobilization and orientation dependence allowed the spin label orientation to be interpreted in terms of orientation of the labelled domain. Smooth muscle contraction is activated by phosphorylation of the RLC. Spin labels were

Electron Paramag. Reson., 2007, 20, 50–72

53

introduced throughout the RLC and the effects of phosphorylation examined.25 Only sites in the N-terminal domain of the RLC showed significant changes in dynamics upon phosphorylation. Oxygen accessibility studies indicated that there was initially little discernible secondary structure, but that upon phosphorylation the ﬁrst 17 residues became much more a-helical. To investigate the structure of the actin-myosin interface, four single-cysteine mutants were prepared (two of actin and two of myosin).26 Changes in mobility and accessibility were not consistent with a simple model with discrete interfaces for weakly and strongly bound states. All four sites appeared to be involved in the strong actin-myosin interface. T4L continues to be a workhorse for advancing the SDSL technique. T4L is a highly a-helical protein that has been well-characterized by SDSL by the Hubbell lab. In recent studies, T4L helix mutants were used for measurement of accessibility and Heisenberg exchange rates by saturation recovery (SR) EPR.27 It was demonstrated that the spin-lattice relaxation and exchange rates can be used to determine secondary structure and to detect multiple states of the side chain and rapid exchange (on the spin-lattice relaxation EPR timescale) between these multiple conformations of the spin-label side-chain. In addition, accessibility measurements on fourteen T4L mutants were used to show that absolute Heisenberg exchange values can be determined from accessibility parameters such as P (derived from power saturation measurements). The authors reported a proportionality constant that directly relates the two values.28 The results of this study also show that the SDSL and crystallographic data from these sites correlate, which provides a basis for comparison of other solution and crystal structures. T4L was also used in a SDSL study to determine feasibility of an EPR technique to measure the structure and orientation of proteins adsorbed to planar lipid bilayers.29 Through the use of surface-tethered T4L mutants, several orientation angles of the planar surface with respect to the magnetic ﬁeld, and spectral simulations, a great deal of information was collected on the orientation and structure of the T4L helices with respect to the planar surface. This technique proved feasible and therefore can be applied to study important orientation and structural features of integral membrane proteins as well as membrane associated proteins.

3

Integral Membrane Proteins

SDSL continues to make significant contributions to our understanding of the structure and molecular dynamics of integral membrane proteins. Because such proteins typically do not readily crystallize and are too large, and tumble too slowly, for structure determination by NMR, SDSL is often the method of choice for the analysis of membrane protein structure. Even with advances in membrane protein crystallography, SDSL provides unique insights into functional dynamics and allows the study of membrane proteins in their native environment.

54

Electron Paramag. Reson., 2007, 20, 50–72

Rhodopsin is a well-studied G protein-coupled receptor involved in the visual signal transduction pathway. The intramolecular interactions that occur within rhodopsin upon disruption of the mechanistically important buried salt bridge between transmembrane helices 3 and 7 were studied by the clever introduction of spin labels on the cytoplasmic surface of rhodopsin that were known to report on significant conformational changes due to activation.30 Through the use of spectral motion as a conformational sensor and the introduction of point mutations to induce various activation states of rhodopsin, it was determined by SDSL that the salt bridge is responsible for maintaining the receptor in its resting state and that disruption of this salt bridge causes the conformational change upon activation. SDSL has also played a key role in characterizing transmembrane signalling events for a family of outer membrane transport proteins that are responsible for acquiring rare nutrients in Gram-negative bacteria. These proteins, which include the iron-transport proteins FepA, FecA, and FhuA, and the vitamin B12 transporter, BtuB, all interact with the inner membrane protein, TonB, and utilize the transmembrane potential gradient as an energy source to drive ligand translocation across the outer membrane. Coupling to TonB is believed to occur through a consensus sequence, the Ton box, that is conserved throughout this family of receptors. In the resting (i.e., unliganded) state, the Ton box is sequestered inside the transmembrane b-barrel domain, where it is unavailable for interaction with TonB. Caﬁso and co-workers placed spin labels at 17 consecutive sites at the N-terminus of BtuB, which encompasses the Ton box (residues 6-12).31 Addition of the vitamin B12 ligand to BtuB reconstituted into 1-palmitoyl-2-oleoyl phosphatidylcholine liposomes increased spin-label mobility along the entire length of the N-terminus and eliminated local secondary structure, indicating that the Ton box and surrounding residues had become ‘‘undocked’’ and extended into the aqueous phase – a conformational change that would allow interaction with TonB. This was consistent with previous SDSL studies on BtuB in intact E. coli outer membranes that demonstrated significant changes in mobility and accessibility along the Ton box upon ligand binding.32 Detergents and certain micelle compositions also induced unfolding of the Ton box of BtuB, while in large micelles the native conformation appeared to be largely conserved.33 In contrast to these studies, determination of the X-ray crystal structure of BtuB with and without bound ligand indicated only a minor change in the conformation of the Ton box.34 A subsequent SDSL study demonstrated that the high concentrations of osmolytes used in crystallization buﬀers prevented the vitamin B12 – induced conformational change.35 This provides a dramatic demonstration of the value of being able to study structure and dynamics under conditions similar to the native environment. SDSL studies of BtuB have now been extended to include a number of osmolytes, and have shown that the N-terminal region of this protein exists in a conformational equilibrium that is sensitive to solution osmolality.36 The SDSL methodology has also been used to examine ligand-induced conformational changes in the bacterial inner-membrane lipid A-transport protein, MsbA. MsbA is an ATP-binding cassette (ABC) transporter and therefore uses ATP as an energy source to ﬂip lipid A from the inner to the outer leaflet of the bacterial inner membranes of Gram-negative bacteria such as E. coli. Although

Electron Paramag. Reson., 2007, 20, 50–72

55

three crystal structures of MsbA from various bacteria have been solved,37–39 the mechanism of transport and the conformational dynamics of ATP and lipid A binding have yet to be elucidated. The resting state structure of MsbA and the structural changes upon ATP binding in the cytoplasmic nucleotide binding domains of E. coli MsbA have been studied by SDSL techniques. One group has shown that the resting state of this protein in solution is consistent with the open state of the crystal structure based on several sites within the transmembrane helices of the homodimer40 and the structure of the conserved Walker A motif of MsbA was determined to be consistent with that of other ATPases.41 In addition, the side-chain motion of each of these sites at each stage of ATP hydrolysis was monitored and revealed that sites 378-385 are directly involved in ATP binding and hydrolysis, while positions 386-390 remain unaffected by ATP turnover. Another group studied three pocket-facing transmembrane helices and reported on the changes observed in the solution state vs. the crystal structure, and the changes induced by the trapped transition state and by the presence of an ATP analogue.42 SDSL has been used to characterize the structure and dynamics of prokaryotic potassium channels.43–47 Voltage-dependent K1 channels are composed of a pore domain and a voltage-sensing domain. A crystal structure for the prokaryotic voltage-dependent channel KvAP has been quite controversial as it indicates that the charged S4 segment of the voltage sensor (containing four arginines in a 10-residue span) is exposed to membrane lipid. SDSL was carried out on a set of B120 single-cysteine mutants comprising the entire voltage-sensing domain in reconstituted, full-length KvAP.46 Under conditions designed to maintain the voltage sensor in a conformation corresponding to the open state, both spin label dynamics and accessibility to O2 and NiEDDA indicated clear boundaries for the individual transmembrane segments. The EPR data was far more compatible with the crystal structure of the isolated voltage-sensor domain than with that of the full-length channel. Oxygen accessibility and mobility data allowed development of a model for how the voltage-sensor domain is organized with respect to the pore domain. The prokaryotic K1 channel from Streptomyces lividans, KcsA, is activated by low pH. KcsA is a tetramer, with each subunit containing two transmembrane helices. Previous SDSL studies have shown that channel activation is accompanied by rigid-body movements of both helices, producing a large opening at the base of the channel.44 The strong spin-spin interactions observed for spinlabelled sites on the inner transmembrane helix (TM2) in the closed state of KcsA, and the loss of such interactions upon channel opening, were utilized to study channel gating mechanisms.47 These studies, in conjunction with a variety of physiological measurements and two new crystal structures, indicated a key role for the selectivity ﬁlter in governing channel gating. Phospholambin (Pln) is a 52-amino acid, single pass transmembrane protein that regulates the sarcoplasmic reticulum calcium ATPase (SERCA) in cardiac muscle. Spin labels attached to a cysteine (A11C) introduced into the basic N-terminal domain of Pln became more motionally restricted in the presence of SERCA.48 Accessibility studies carried out with O2, NiEDDA, and Ni21

56

Electron Paramag. Reson., 2007, 20, 50–72

complexed to the lipid headgroup (DOGS-NiNTA) indicated that this site interacted strongly with the membrane surface in the absence of SERCA and that association with SERCA greatly reduced this interaction. To further characterize the structure of Pln in the membrane, four Pln analogues were prepared by solidphase peptide synthesis incorporating the spin-labelled amino acid TOAC (2,2,6,6tetramethyl-piperidine-N-oxyl-4-amino-4-carboxylic acid).49 This probe, with the nitroxide moiety rigidly coupled to Ca, provides direct insight into the conformational dynamics of the peptide backbone. Although one site in the transmembrane domain of Pln gave a single, strongly-immobilized population, all three sites in the cytoplasmic domain had two clearly resolved spectral components suggesting an equilibrium between conformational states. For TOAC at position 11 (mentioned above), attachment of a N-terminal lipid anchor eliminated the more disordered spectral component. Accessibility studies using DOGS-NiNTA indicated that all three sites in the cytoplasmic domain were located near the membrane surface. Addition of SERCA to TOAC-labeled Pln decreased the mobility of the spin label at two different sites (11 and 46), and this change in mobility was used to determine the SERCA-Pln dissociation constant, Kd.50 SERCA increased the dynamically-disordered form of Pln both in micelles (under conditions used for NMR studies) and in lipid bilayers, and this effect was especially pronounced in the cytoplasmic domain. Based on complementary results from NMR and EPR, a model for the conformational equilibrium between Pln and SERCA was proposed.50 The ability to measure the depth of a spin-labelled site in the membrane bilayer is a unique aspect of SDSL. This methodology is based on the inverse concentration gradients of paramagnetic relaxation agents such as O2 and polar complexes of paramagnetic ions across the bilayer normal. A recent study examined EPR depth measurements using a series of single-cysteine mutants based on a well-studied a-helical transmembrane peptide (WALP) to position the label along the bilayer normal, and compared CW saturation measurements with those made by saturation recovery EPR.51 This model system may provide better localization of the spin label, since the peptide is less ﬂexible than the typicallyused lipid spin labels. Proﬁles for relaxation enhancement by O2 and NiEDDA across the bilayer have been obtained, and calculated depth parameters are in good agreement with the hyperbolic tangent function introduced by Frazier et al.52 A comparison of this data with depth measurements derived from lipidanalogue spin labels indicated that they were in excellent agreement.53

4

Membrane-Associated Proteins

SDSL has also been an invaluable tool in the study of proteins that associate with the membrane surface. Proteins that reversibly associate with the membrane play key physiological roles in mediating membrane fusion, bilayer organization, and signaling. The E. coli maltose transport complex, MalFGK2, is in the ABC transporter class of proteins described above and is one of the most well characterized

Electron Paramag. Reson., 2007, 20, 50–72

57

transporters. Maltose binding protein (MBP) is a soluble periplasmic protein that delivers maltose to the inner membrane transport complex MalFGK2, which stimulates ATP hydrolysis by the cytoplasmic MalK dimer. MBP binds maltose between two lobes of the protein as determined by x-ray crystallography studies.54,55 Davidson and co-workers introduced spin labels within each lobe of MBP to monitor binding of MBP to the transport complex proteins MalF and MalG.56 The EPR data followed the mobility of each introduced label in solution, in the presence of transporter alone, and in the presence of transporter and ATP, ADP, and the vanadate-induced transition state complex, in addition to measuring the distances between the two labels at each of these stages. Binding of maltose to MBP causes its lobes to close in on maltose and brings the spin labels from 420 A˚ apart to approximately 8 A˚ apart, which is a significant distance difference that enabled the investigators to identify the maltose-release step upon MBP binding to the transport complex. The resulting data suggested that MBP opened and released maltose into the transporter upon ATP hydrolysis and closure of the MalK dimer interface. Similarly, the general secretory system of E. coli includes a cytoplasmic chaperone protein (SecB), a peripheral ATPase (SecA), and a membrane translocon, which act together to bind unfolded protein and transfer the protein across the cytoplasmic membrane. SecB is a chaperone responsible for capturing the unfolded polypeptides and delivering them to SecA, which then undergoes ATP hydrolysis to transfer the protein across the membrane. Randall et al. introduced spin labels at 41 solvent-exposed positions on SecB to study the docking interface between polypeptide-bound SecB and SecA.57 Because SecB is a homotetramer and it was demonstrated that binding to SecA is not symmetric, the authors used spectral subtraction methods to reveal the actual lineshapes due to contact with the complex. A model of SecB binding to SecA was presented based on the changes in spectral motion, and is a productive step toward mapping out the interaction interface at the initial stage of protein export in the E. coli secretory system. Twenty sites within the F0 water-soluble subunits b2 of the F0F1-ATP synthase were modiﬁed with a spin label and monitored for changes in mobility upon binding to the cognate membrane-associated F1 subunits.58 Data indicated a tight binding interaction between b2 and F1 for the majority of sites studied, and differences in spectral motion of the sites were observed upon deletion of the d subunit of F1 indicating different binding interactions. The structure of the C-terminal domain of human apolipoprotein A-I (apoA-I) was studied by SDSL in both the lipid-free and lipid-associated states using 39 singly labelled and two doubly labelled proteins.59 The tertiary and quaternary structure in this region was mapped out using a combination of mobility, accessibility, and spin-spin measurements. Upon reconstitution into high density lipoprotein (HDL), of which apoA-I is the major component, a lipid-induced transition into a-helices occurred, similar to the conformational switches seen in viral fusion proteins. C2 domains are abundant motifs found in a wide variety of proteins that reversibly interact with the membrane surface. Synaptogamin I is a synaptic

58

Electron Paramag. Reson., 2007, 20, 50–72

vesicle-associated protein that is anchored to the bilayer through a single transmembrane helix, and contains two C2 domains that are proposed to form the Ca21-sensitive elements in exocytosis. SDSL of the isolated C2B domain of synaptogamin I was used to examine membrane binding and insertion.60 Significant decreases in spin-label motion at several sites in the Ca21-binding loops of C2B were observed upon membrane binding. The isolated C2B domain bound phosphatidylcholine (PC):phosphatidylserine (PS) membranes only in the presence of Ca21, but was able to bind membranes containing phosphatidylinositol-4,5-bisphosphate even in the absence of calcium. Power saturation studies showed that the Ca21-binding loops penetrated the membrane interface in a manner similar to that previously observed for the synaptogamin I C2A domain.61 SDSL of the C2 domain from protein kinase Ca indicated that its calciumbinding loops inserted into the membrane bilayer62 in a manner similar to that described above for synaptogamin I. Additional spin-labelling sites showed partial protection from collisions with NiEDDA, suggesting a model in which the C2 domain is oriented with its canonical b-sheet aligned parallel to the membrane surface. The C2 domain of cytosolic phospholipase A2 (cPLA2) triggers Ca21-dependent binding to membranes, allowing hydrolysis of arachidonic acid-containing phospholipids. In contrast to the majority of C2 domains, which are thought to utilize an electrostatic mechanism for membrane binding, binding of the cPLA2 C2 domain is primarily hydrophobic and thus able to bind to zwitterionic (e.g., PC) bilayers. Single cysteines were introduced at all 27 residues in the calcium-binding loops of the cPLA2 C2 domain and spin-labelled with MTSL.63 Of these, 24 retained Ca21-activated membrane binding. Membrane insertion was observed for several sites in the Ca21-binding loops. These data were used to model the structure of the C2 domain onto the membrane surface, and suggested an orientation of the C2 domain that was perpendicular to the membrane surface. The protein a-synuclein is a main component of Lewy body aggregates, a hallmark of Parkinson’s disease pathogenesis. Although its precise function is not yet known, a-synuclein has been proposed to play a role in a number of membrane-related processes. The N-terminal domain of a-synuclein contains seven repeats, each with 11 amino acids. Analysis of 47 singly-labelled asynuclein mutants demonstrated that membrane association caused a signiﬁcant conformational change in the N-terminal region of the protein, but had little effect on the C-terminal domain.64 Accessibility studies encompassing three of the N-terminal repeats indicated formation of a continuous helix associated with the membrane surface, each repeat having a well-deﬁned periodicity of 3.67 residues/turn. This slightly expanded helix allows for three complete turns by an 11-residue repeat, facilitating the membrane interaction of multiple consecutive units. SDSL has also been used to study ﬁbril formation by a-synuclein,65 and EPR studies using lipid-analogue spin labels have been used to examine the interaction of a-synuclein with membranes.66 Myelin basic protein (MBP) is another membrane-associated protein of considerable biomedical interest. Autoimmunity towards MBP may play an

Electron Paramag. Reson., 2007, 20, 50–72

59

important role in the inﬂammatory demyelinating disease, multiple sclerosis (MS). MBP is highly post-translationally modiﬁed, resulting in isoforms with varying degrees of positive charge, with the most basic form (C1) predominating in adult humans and the least basic form (C8) found at elevated levels in MS patients. Harauz and co-workers have used SDSL to characterize to membranebound structure of MBP.67–69 SDSL showed that the central and C-terminal regions of the less-basic C8 isoform were significantly less associated with myelinlike membranes than the more common C1 isoform.68 Nitroxide scanning of a region of MBP associated with eliciting a strong immune response indicated that, in the normal C1 isoform, this ‘‘immunodominant epitope’’ had a-helical secondary structure and was embedded in the membrane at a shallow, 91 tilt relative to the membrane surface.67 In contrast, the immunodominant epitope of the C8 isoform was much more highly exposed to the aqueous phase,69 which could facilitate induction of an autoimmune response. Annexins are a family of soluble cytosolic proteins that undergo reversible, Ca21-dependent binding to phospholipid bilayers. Although their physiological function has not been established, annexins have been implicated in a number of processes that rely on their ability to interact with membranes such as membrane fusion, vesicle trafﬁcking, and domain formation. High-resolution crystal structures exist for the soluble forms of several annexins; however SDSL has been instrumental in characterizing their interaction with membranes under a variety of conditions.70–75 Annexins have a conserved core domain and a variable N-terminal domain. The highly conserved core domain contains four repeats – each of which contains ﬁve a-helices connected by short interhelical loops. In solution, EPR mobility and accessibility parameters from nitroxide scanning experiments show periodic ﬂuctuations that readily identify helices and connecting loops, and are in excellent agreement with fractional accessible surface areas determined from the crystal structures.71 A continuous, 30residue nitroxide scan of helices D and E and the intervening loop in the second repeat of annexin B12 showed that upon Ca21-dependent membrane binding, although the overall backbone fold did not change, there was a very strong immobilization in the D-E loop and accessibility measurements showed that this loop became buried in the hydrophobic core of the bilayer.72 A subsequent SDSL study showed similar behaviour for the D-E loops in the other three repeat units as well, leading to the conclusion that these loops are involved in the formation of a highly immobile complex on the surface of the membrane.71 Langen and co-workers have also demonstrated that the Ca21dependent membrane-bound form of annexins A5 and B12 is a trimer.70,75 The observation of strong spin-spin interactions allowed identiﬁcation of the trimer interface, and spin-dilution studies, in which spin-labelled annexin was mixed with varying amounts of unlabelled annexin, established the trimeric nature of the oligomer. Time-resolved EPR was used to show that the trimer forms very rapidly (on the order of tens of milliseconds), but that subunit exchange of pre-formed trimers is relatively slow.70 Remarkably, the A1 and A2 isoforms do not form similar membrane-bound trimers despite the highly-conserved structure of their core domains.75 In addition to the Ca21-dependent membrane

60

Electron Paramag. Reson., 2007, 20, 50–72

binding described above, annexins also undergo membrane insertion at low pH.73,74 The exact pH required is strongly dependent on phospholipid composition.74 A continuous nitroxide scan of helix A-loop-helix B showed that this region of annexin B12 refolds into a transmembrane helix that lines an aqueous pore, possibly upon protonation of a key aspartate.73 SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) proteins are essential components of vesicle fusion in eukaryotic cells, mediating processes such as protein trafﬁcking and neurotransmitter release. It is believed that vesicle-associated (v-) SNAREs form selective complexes with target membrane (t-) SNAREs, bringing their respective membranes into close proximity to initiate fusion. Shin and coworkers have carried out an extensive set of studies on the membrane interactions of SNARE proteins and the mechanism of membrane fusion.76–82 For yeast SNARE proteins involved in post-Golgi vesicle trafﬁcking, spin labels attached to a soluble, coiled-coil domain from the v-SNARE Snc2p showed reduced mobility upon addition of the complementary soluble domains of cognate t-SNAREs, demonstrating spontaneous folding and complex formation, even in the absence of membranes.80 Similar changes were observed with a somewhat larger construct containing the membrane-interfacial domain and reconstituted into PC:PS liposomes. Such spontaneous assembly of the yeast SNARE complex is in contrast to what is observed for neuronal SNAREs. Power saturation analysis of reconstituted Snc2p indicated membrane insertion of a short, two-turn ahelix in the interfacial region of the bilayer at an angle of B401 relative to the membrane normal.80 The membrane-embedded structure of the transmembrane (TM) domain of Snc2p was also determined, showing a-helical secondary structure with a tilt of B241 relative to the membrane normal and evidence of helix-helix contact sites.82 The overall structure of Snc2p is similar to that previously determined for neuronal SNARE proteins,77,82 although with some important distinctions.80 The structure and dynamics of the yeast t-SNARE, Sso1p, was also examined.78 Accessibility to O2 and NiEDDA indicated that Sso1p contained a well-deﬁned TM a-helix, and a V-shaped proﬁle consistent with a TM orientation was also observed in the isotropic hyperﬁne coupling constant, demonstrating that this simple parameter can provide a useful measure of membrane depth. Sites involved in quaternary interactions (i.e., oligomer formation) were identiﬁed based on observation of a strongly-immobilized component in the room-temperature EPR spectra and was conﬁrmed by Fourier-deconvolution analysis of the low-temperature (130 K) EPR spectra that were indicative of spin-spin interaction. The fraction of non-interacting spins was determined based on a constant y-axis oﬀset in the Fourier deconvolution, consistent with a monomer-oligomer equilibrium.78 Spin-labelled SNAREs were also used to examine the proposed ‘‘zipper model’’ of complex formation.79 Time-resolved EPR was used to follow the kinetics of complex assembly for the yeast SNAREs, using a two-syringe stopped-ﬂow apparatus ﬁtted to the EPR cavity.81 Second order rate constants of B4000 – 5000 s1 M1 were observed for the soluble coiled-coil domains. Rate constants were B40-fold slower when SNAREs containing the transmembrane domains were

Electron Paramag. Reson., 2007, 20, 50–72

61

reconstituted before mixing the resulting vesicles. In both cases, rate constants were quite similar for all positions of the spin label, suggesting that complex formation may occur by a concerted, rather than a ‘‘zippered’’ mechanism. For peptides that partition between the membrane and aqueous phase, SDSL provides a highly effective method for determining the partition coefﬁcient. Membrane binding is typically accompanied by a substantial reduction in spinlabel mobility, so that at equilibrium one can observe distinct signals from both free and membrane-bound populations. The fraction of membrane-bound peptide can be quantitated either by spectral subtraction or from changes in line amplitudes.83,84 EPR studies can be done across a wide range of lipid concentrations, which can be problematic when using ﬂuorescence techniques, and do not require separation of bound and unbound species as must be done when using radioactive tracers. The interaction of a peptide derived from a secretory carrier membrane protein (SCAMP) with membranes containing phopsphatidylinositol-4,5-bisphosphate, PI(4,5)P2, was examined using a combination of SDSL and NMR.85 The 11-residue peptide, SCAMP-E, was derived from a loop sequence connecting two transmembrane a-helices and contained three basic and four aromatic residues. Binding afﬁnity for phosphatidylcholine liposomes was enhanced by two orders of magnitude upon incorporation of just 3 mole % PI(4,5)P2. A spin-labelled analogue of PI(4,5)P2 was used to directly verify binding to SCAMP-E. Depth measurements on spin-labelled, single cysteine mutants of SCAMP-E indicated that the nonpolar side chains of this peptide sit B7–9 A˚ below the lipid phosphates. Paramagnetic relaxation parameters for lipid protons in bicelles were consistent with the interfacial position of membrane-bound SCAMP-E determined by SDSL power saturation. Antimicrobial peptides (AMPs) are small (12–50 amino acid) cationic peptides that play an essential role in the innate immune response. A large number of these peptides exist in solution as random coils, but fold into amphipathic structures (either a-helices or b-strands) upon binding to cell membranes. Synthetic peptides composed of D-amino acids are just as potent as the natural L-enantiomers, suggesting that membranes are their primary targets of action. SDSL was used to characterize the interaction of CM15, a 15-residue chimeric peptide based on the insect AMPs cecropin A and mellitin, with bilayers mimicking the composition of the Gram-negative bacterial membrane.84 Membrane binding of CM15 produced a large reduction in mobility at spin-labelled sites along the entire length of the peptide. Binding afﬁnities (on the order of 104 M1) were found to be relatively independent of the position of the labelling site as long as the label was located in the hydrophobic, C-terminal domain. SDSL depth measurements on CM15 fully bound to liposomes at a high lipid:peptide ratio (200:1) indicated that the membrane-bound peptide was folded into an a-helix aligned parallel to the bilayer surface and inserted B5 A˚ below the membrane interface.84 This localization is essentially ideal for allowing the hydrophilic face of the amphipathic peptide to interact with polar lipid headgroups, while keeping the non-polar face of the helix buried in the hydrophobic phase of the membrane.

62

5

Electron Paramag. Reson., 2007, 20, 50–72

Fibril-Forming Proteins

Amyloid ﬁbrils formed from the aggregation of misfolded proteins play an important role in a number of devastating human diseases, including Alzheimer’s disease, Parkinson’s disease, prion-related encephalopathies, and non-insulin-dependent type II diabetes. The protein constituents of these ﬁbrils associate in a common cross-b structure, with b-strands arranged perpendicular to the ﬁbril axis. SDSL studies have provided insights into the structural organization of several types of amyloid ﬁbrils.65,86,87 Tau is a microtubuleassociated protein that forms ﬁlamentous inclusions in a number of neurodegenerative diseases, most notably Alzheimer’s disease. In solution, tau is thought to be a ‘‘natively unfolded’’ protein. Spin labels placed at widely spaced positions along tau exhibited fast rotational mobility in solution, although ﬂuorescence resonance energy transfer (FRET) experiments indicated some degree of tertiary structure.88 In a study of ﬁbril formation by tau, spin labels placed within one of tau’s microtubule-binding regions showed rapid motion in solution, however upon the formation of ﬁlaments almost all sites showed only a single line, indicating strong spin exchange.87 When spinlabelled protein was diluted with unlabelled tau (at a 1:3 ratio), spin exchange was alleviated and a characteristic strongly-immobilized nitroxide spectrum was observed. Spin exchange was observed at 20 consecutive sites within the microtubule binding domain, indicating that the peptides are aligned in a parallel fashion.87 This result is similar to that found for other amyloid ﬁbrils including islet amyloid polypeptide86 and a-synuclein.65 The ﬁlament assembly of the human type III intermediate ﬁlament (IF) protein vimentin was examined by SDSL methods under physiological conditions.89 Spin labels were introduced into 21 different positions along rod domain I in vimentin and the motional analysis provided direct evidence for the predicted coiled-coil motif of the vimentin dimer. In addition, it was discovered that the dimers were arranged in an anti-parallel and staggered orientation using a single residue at a critical point of overlap within rod domain I. Filament assembly was also followed and the sequence of interactions upon in vitro assembly was identiﬁed as starting with a loose coiled coil fold and progressing to tetramers and then to octamers or larger multimers. A later study, using these characterized reporter sites, investigated the impact of a known disease-causing mutation on vimentin structure and assembly, giving valuable insights into which speciﬁc stage of assembly is blocked by an inherited arginine to serine mutation.90

6

Nucleic Acids

A 23-nucleotide ribonucleic acid (RNA) molecule was spin-labelled and analyzed for motion and solution structure by Qin et al.91 demonstrating the general utility of the technique to the study of RNA structure and local interactions. Spin-labelling was accomplished by the introduction of a 4-thiouridine

Electron Paramag. Reson., 2007, 20, 50–72

63

derivative at the site of choice within the RNA and subsequently labelled with a spin-label reagent. Using this technique, conformational changes within a GAAA tetra-loop receptor upon docking of a GAAA tetra-loop were monitored at four positions within a molecule containing the receptor sequence.92 This study provided direct evidence that base unstacking is a key step in this complex formation in the presence of magnesium. SDSL of RNA was further used to identify the sites in the basic region of the Tat protein that are affected by complex formation with the transactivation responsive (TAR) RNA, which is the 5 0 leader sequence of the HIV-1 mRNA genome.93 This interaction occurs during transcription and is therefore a potential target for therapeutic intervention. The TAR RNA was spin-labelled with a nitroxide probe at four different uridine nucleotides and changes in spinlabel mobility were monitored in the presence of various mutated forms of the Tat protein. From these data, the arginine residues within Tat that are important for complex formation were identiﬁed, and some of the mutations allowed the complex to switch between rigid and ﬂexible states. Overall, this study identiﬁed the speciﬁc contacts required for rigid complex formation and was the ﬁrst example of using RNA spin labels to study the mutational effects of the RNA binding protein upon RNA binding. In an attempt to bypass the commonly used method of spin-labelling recombinant proteins following puriﬁcation from their native environments, Voss et al. evaluated the ability of Xenopus oocytes to incorporate injected nitroxide-containing tRNAs.94 The incorporation rate was low, but EPR signals were recorded and a tRNA carrying a nitroxide spin label was synthesized in the process. Future success of this approach will give the potential to examine labelled proteins in situ, and to introduce spin labels into protein with intact native cysteines.

7

Substrates

Vogel et al. used the spin-labelled photoafﬁnity ATP analogue, 2-N3-spin labeladenosine triphosphate, to show for the ﬁrst time that when ATP is bound to the isolated b-subunits of F1-ATPase it is in two distinctly different conformations.95 The spin-labelled ATP analogue was allowed to bind to the protein and then irradiated to covalently attach the probe to the protein. The ratios of the two different signals indicated that there is asymmetry within the catalytic sites, likely corresponding to the open and closed states of the subunits, with the closed state dominating. In addition, spin-labelled ATP (SL-ATP), where the spin label is attached to the ribose, was used as an ATP analogue to study ATP binding in the multidrug-resistance protein P-glycoprotein (P-gp).96 The SL-ATP analogue was hydrolyzed by the ABC transporter and allowed for stoichiometric binding. Interestingly, reduction of the nitroxide was observed when bound to native P-gp containing a cysteine residue in close enough proximity to the ribose-bound spin label. Repeating the experiments in a cysteine-free P-gp

64

Electron Paramag. Reson., 2007, 20, 50–72

background alleviated the reduction, and further studies indicated strong immobilization of the spin label upon trapping the transporter, indicative of a closed transition state conformation. Another spin labeled substrate used for binding studies is the neurokinin-1 receptor (Nk1r) agonist, substance P peptide. A spin label was attached to the third residue of substance P by amino group derivitization and was found to be an excellent substrate analogue for the G-protein coupled receptor.97 The spinlabelled substrate reported on the binding-pocket environment and served as a real-time endocytosis reporter through loss of the EPR signal due to the reducing environment of the endosome.

8

Distance Measurements

An increasingly useful aspect of SDSL is the ability to make distance measurements between two spin labels, either intramolecular distances between two labels in the same monomer or intermolecular distances between sites on different monomers in an oligomer (for reviews, see refs. 2–4,7,9,98). In the range from B8–20 A˚, distances can be determined from dipolar broadening of the CW EPR spectrum using spectral simulation approaches based on Fourier deconvolution,99,100 convolution of the spectra from non-interacting spins with a broadening function,101 or rigorously simulating spectra with consideration of both the distance between spin labels and their relative orientations.102 A recently developed ‘‘tether-in-a-cone’’ model takes into consideration the effects of the relatively restricted side-chain dynamics of MTSL on dipolarcoupled CW EPR spectra.103 In this approach, the unpaired electrons are placed at the end of a tether that is allowed to adopt all possible angles within a cone of variable half-width. When probes are highly ordered, this approach can provide additional information about their relative geometry. Experimentally, observation of dipolar coupling by CW methods requires that the individual labels be relatively immobilized on the EPR time scale to reduce motional averaging of the dipolar interaction. This may require samples to be frozen, but is also applicable in solution as long as the tumbling rate is slow relative to the dipolar coupling.100,104 For interspin distances in the range of 15–60 A˚, pulse electron-electron double resonance (DEER or pELDOR) and double-quantum coherence (DQC) methods can be used. At present, the need for sufﬁciently long T2 relaxation times require these measurements to be made at low temperature (typically 40–77 K). A number of recent SDSL studies have made use of these powerful pulsed methods.104–109 A comprehensive SDSL study has been carried out on the cytoplasmic domain of the erythrocyte anion exchange protein, AE1 (also referred to as band 3).104 AE1 consists of a transmembrane domain that spans the bilayer 12 to 14 times, and a 42.5 kDa cytoplasmic domain (cdb3) that mediates numerous important protein-protein interactions with both soluble proteins (e.g., hemoglobin) and the erythrocyte cytoskeleton. A crystal structure of cdb3 obtained at pH 4.8 showed an unexpectedly compact dimeric structure.110 Given the low

Electron Paramag. Reson., 2007, 20, 50–72

65

pH used in crystallization, along with extensive evidence of pH-dependent conformational changes in AE1, it was important to investigate the structure of cdb3 under more physiological conditions. Using SDSL, mobility and power saturation accessibility parameters veriﬁed that the solution structure of an ahelix in the peripheral domain of cdb3 was the same as that observed in the crystal structure.104 Similarly, a nitroxide scan of one of the helices forming the dimer interface (helix 10) also showed a-helical periodicity in side-chain mobility and accessibility to NiEDDA, which again was in excellent agreement with the structure and orientation of this helix in the crystal structure. One site in helix 10 (Q339R1) was particularly remarkable in that it displayed highly resolved dipolar coupling between the spin labels of each monomer. Analysis using the ‘‘tether-in-a-cone’’ model103 indicated an interspin distance of 14.7 0.4 A˚, in close agreement with the distance between these sites across the dimer interface from the crystal structure. To further characterize the orientation of the two monomers in the cdb3 dimer, DEER was used to measure distances extending further out along helix 10 (from 340–345). Distances in the range 24.9–37.0 A˚ were measured. The width of the distance distribution was very narrow (0.8–1.1 A˚) for sites proximal to the core of dimer, and increased (3.6– 6.6 A˚) towards the distal end of the helix in a manner that parallelled the general increase in side-chain mobility. Distance measurements were also made for an additional 11 sites, using either DEER or by ﬁtting CW EPR spectra with a Gaussian convolution model.101 Distances were reported in the range 6.2–47.7 A˚, and again veriﬁed that the solution structure of cdb3 at neutral pH was in close agreement with the pH 4.8 crystal structure. Taken together, the data in this extensive study demonstrate the power of the SDSL approach in deﬁning elements of local secondary structure and using distance measurements to elucidate how those secondary structure elements are arranged. The close agreement of the EPR data with that from the published crystal structure provides conﬁdence that these methods can be used to determine unknown structures using SDSL alone. Sensory rhodopsins are photoreceptor proteins that mediate phototrophic and photophobic responses in archebacteria. These integral membrane proteins have many structural similarities to bacteriorhodopsin and vertebrate rhodopsin, including a 7-transmembrane helix motif and a retinal chromophore. Constraints from ﬁve pairs of inter-residue distance measurements indicated that the photophobic sensory rhodopsin, NpSRII, exists with its cognate signal transducer (NpHtrII) in a 2:2 complex with 2-fold symmetry.111 Light-induced changes in dipolar couplings suggested a rotation of the second transmembrane helix, reminiscent of the light-induced conformational changes observed in vertebrate rhodopsin and bacteriorhodopsin. Significant structural differences were observed between membrane-reconstituted and detergent-solubilized complexes.112 To explore interspin distance measurements in RNA, model 10-mer RNA duplexes and HIV-1 TAR RNA motifs were spin-labelled at the 2 0 -amino groups of appropriately placed uridines and characterized by Fourier deconvolution of dipolar-broadened CW EPR spectra.113 Good agreement was

66

Electron Paramag. Reson., 2007, 20, 50–72

found between experimental distances and those from molecular dynamics calculations. The importance of explicitly incorporating the structure of the spin label into molecular models based on distance measurements has been examined.114 Experimental EPR distance measurements obtained from either CW line broadening, DEER, or relaxation enhancement in the range 8–50 A˚ gave a reasonably strong correlation with Cb-Cb distances from corresponding atomic models (R2 B 0.8). However, correlation was poor at distances less than 15 A˚ where the tether of the spin label makes a relatively larger contribution to the overall separation and the overall mean error was B6 A˚. In contrast, when the spin label was included in the structural modeling the correlation between experimental and simulated distances improved significantly and the mean error decreased to B3 A˚. SDSL and DEER time-domain experiments were used to begin characterization of the unknown structure of PutP, a Na1/proline symporter found in the inner membrane of E. coli.106 Previous studies have suggested that PutP has a somewhat unusual topology, with a predicted 13 transmembrane helices. To evaluate the proposed structure, three sets of double mutants were constructed from four mutation sites, all of which were in predicted interhelical loops. None of the double mutants showed evidence of dipolar broadening by CW EPR methods, but distances for all three pairs could be obtained from DEER experiments. One rather long distance of 48 A˚ indicated that two of the labelling sites were indeed located on opposite sides of the membrane, in agreement with the proposed structure. Much shorter distances (B18 A˚ and 22 A˚) between the remaining two pairs indicated that they were located on the same side of the bilayer. Relatively large distance distributions for all three pairs suggested that the structure of these loop regions is rather poorly deﬁned, consistent with the general expectation of ﬂexibility in interhelical loops. In a study of the chlorophyll a/b light-harvesting complex, four-pulse DEER was used to determine distances for six different spin-label pairs.107 Double cysteine mutants were constructed and labeled with an iodoacetamide spin label. Three of the pairs gave distances in the range 30–40 A˚ that were comparable to the corresponding distances in the light-harvesting complex crystal structure. The other three pairs each gave bimodal distance distributions. Comparison of data between monomeric and trimeric complexes indicated that the N-terminal domain exists in at least two conformational states in the monomer, only one of which is significantly populated in the trimer. In addition to the measurement of intramolecular distances, four-pulse DEER methodology has sufﬁcient range for the characterization of distances between subunits in oligomeric proteins. NhaA is a large protein that catalyzes pH-dependent Na1/H1 exchange across the inner membrane of E. coli. A reconstructed electron density map based on two-dimensional crystals diﬀracting at 4 A˚ indicated that NhaA crystallized as a dimer, with each monomer containing 12 transmembrane helices. Two sites thought to be involved in the pH-dependent regulation of NhaA were selected for SDSL.106 Conventional and low-temperature EPR showed evidence of a pH-dependent conformational

Electron Paramag. Reson., 2007, 20, 50–72

67

change at one of the sites (H225R1), and accessibility measurements indicated that both sites were buried either in the lipid or within the tertiary structure of the protein. To examine possible interaction between spin-labelled sites on different monomers, four-pulse DEER data were analyzed both in terms of a simple exponential decay, exp(kt), and in terms of a two-dimensional exponential decay, exp (kt2/3), as expected for a homogenous planar distribution. For H225R1 both types of analysis gave similar results, with an interspin distance of B44 A˚ and a distribution width of B3 A˚, suggesting formation of a well-deﬁned oligomer. For the other site (V254R1), a reliable distance (20 A˚, with a width of 2 A˚) could only be obtained for the assumption of a planar distribution. Analysis of the extent of oligomerization indicated that a significant fraction of the protein was present as dimers at pH 5.8, and that dimerization increased with increasing pH. Milov et al. have used CW and pulsed EPR methods to examine the secondary structure and aggregation state of a spin-labelled derivative of the fungal antibiotic peptide, tricogin GA IV.108 This 12-residue peptide (a peptaibiol) contains an N-terminal octanoyl, a C-terminal leucinol, and several residues of the unusual amino acid, a-aminoisobutyric acid (Aib). One or more Aib residues were replaced with the spin labelled amino acid, TOAC, and pulse ELDOR (pELDOR) studies indicated that the peptide formed tetrameric aggregates in an apolar chloroform-toluene glass, but remained monomeric in a polar chloroform-toluene-ethanol mixture. pELDOR studies of a doublylabelled peptide with TOAC residues at the 1 and 8 positions gave an intramolecular distance between labels consistent with a 310-helical structure.108 Excess unlabelled peptide was used to eliminate strong intermolecular spin-spin interactions between peptides in a given aggregate. Gramicidins are linear antimicrobial peptides composed of alternating Dand L-amino acids. Double-quantum coherence (DQC) EPR of gramicidin A, spin-labelled at its C-terminus, gave a well-deﬁned interspin distance of 30.9 A˚ in 1,2-dimyristoyl phosphatidylcholine bilayers, consistent with formation of head-to-head dimers.109

9

Conclusions

As can be seen from the studies reviewed in this chapter, SDSL has become widely used in recent years and continues to advance in its applications and techniques. It remains one of the leading techniques for the study of integral membrane and membrane-associated proteins and peptides, where the ability to study proteins in their native environment is a significant advantage, and has proven to be highly complementary to crystallography and NMR techniques. Spin-labelled substrates can give valuable information on substrate binding within a protein complex and many have been shown to be good binding substitutes. SDSL often can yield data on proteins and peptides not obtainable by other techniques and is especially amenable to studies of proteins and their binding partners since labels can be placed on either or both participants.

68

Electron Paramag. Reson., 2007, 20, 50–72

Improved methods for introducing spin labels into nucleic acids will yield many new opportunities. The rapidly expanding advances in distance measurement methodology and applications, in conjunction with improvements in molecular modeling and dynamics simulations, hold significant promise for further applications of site-directed spin labeling in biological systems. Acknowledgements We acknowledge support from NIH (GM068829 to JBF, and GM070642 and AI058024 to CSK). References 1. C.S. Klug and J.B. Feix, Biological Magnetic Resonance Volume 24, L.J. Berliner, S.S. Eaton, G.R. Eaton (Eds.), Kluwer Academic/Plenum Publishers, Hingham, MA, 2004, 269. 2. G.R. Eaton, S.S. Eaton and L.J. Berliner, Distance Measurements in Biological Systems by EPR, Volume 19 of Biological Magnetic Resonance, Kluwer, New York, 2000. 3. E.J. Hustedt and A.H. Beth, Biological Magnetic Resonance, L.J. Berliner, S.S. Eaton, G.R. Eaton (Eds.), Kluwer Academic/Plenum Publishers, New York, 2000, 155. 4. W. Xiao and Y.K. Shin, Biological Magnetic Resonance, Volume 19, L.J. Berliner, S.S. Eaton, G.R. Eaton (Eds.), Kluwer Academic/Plenum Publishers, New York, 2000, 249. 5. H.S. Mchaourab and E. Perozo, Biological Magnetic Resonance, Volume 19, L.J. Berliner, S.S. Eaton, G.R. Eaton (Eds.), Kluwer Academic/Plenum Publishers, New York, 2000, 185. 6. W.L. Hubbell, C. Altenbach, C.M. Hubbell and H.G. Khorana, Adv. Protein Chem., 2003, 63, 243. 7. H.J. Steinhoﬀ, Biol. Chem., 2004, 385, 913. 8. P.Z. Qin and T. Dieckmann, Curr. Opin. Struct. Biol., 2004, 14, 350. 9. S.S. Eaton and G.R. Eaton, Electron Paramagnetic Resonance, B.C. Gilbert, M.J. Davies, D.M. Murphy (Eds.), Royal Society of Chemistry, London, 2004, vol. 19, 318. 10. N.J. Malmberg and J.J. Falke, Annu. Rev. Biophys. Biomol. Struct., 2005, 34, 71. 11. C. Altenbach, S.L. Flitsch, H.G. Khorana and W.L. Hubbell, Biochemistry, 1989, 28, 7806. 12. A.P. Todd, J. Cong, F. Levinthal, C. Levinthal and W.L. Hubbell, Proteins, 1989, 6, 294. 13. L. Columbus, T. Kalai, J. Jeko, K. Hideg and W.L. Hubbell, Biochemistry, 2001, 40, 3828. 14. M.A. Lietzow and W.L. Hubbell, Biochemistry, 2004, 43, 3137. 15. J.H. Zhang, G. Xiao, R.P. Gunsalus and W.L. Hubbell, Biochemistry, 2003 42, 2552. 16. A. Hillar, B. Tripet, D. Zoetewey, J.M. Wood, R.S. Hodges and J.M. Boggs, Biochemistry, 2003, 42, 15170. 17. Z. Fu, E. Aronoﬀ-Spencer, J.M. Backer and G.J. Gerfen, Proc. Natl. Acad. Sci. U. S. A., 2003, 100, 3275.

Electron Paramag. Reson., 2007, 20, 50–72

69

18. L. Columbus and W.L. Hubbell, Biochemistry, 2004, 43, 7273. 19. C. Bracken, P.A. Carr, J. Cavanagh and A.G. Palmer III, J. Mol. Biol., 1999, 285, 2133. 20. S. Popp, L. Packschies, N. Radzwill, K.P. Vogel, H.J. Steinhoﬀ and J. Reinstein, J. Mol. Biol., 2005, 347, 1039. 21. H.S. Mchaourab, E.K. Dodson and H.A. Koteiche, J. Biol. Chem., 2002, 277, 40557. 22. H.A. Koteiche and H.S. Mchaourab, J. Biol. Chem., 2003, 278, 10361. 23. Y.H. Kim, D.M. Kapfer, J. Boekhorst, N.H. Lubsen, H.P. Bachinger, T.R. Shearer, L.L. David, J.B. Feix and K.J. Lampi, Biochemistry, 2002, 41, 14076. 24. B.A. Baumann, H. Liang, K. Sale, B.D. Hambly and P.G. Fajer, Biophys. J., 2004, 86, 3030. 25. W.D. Nelson, S.E. Blakely, Y.E. Nesmelov and D.D. Thomas, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 4000. 26. V.L. Korman, S.E. Anderson, E. Prochniewicz, M.A. Titus and D.D. Thomas, J. Mol. Biol., 2006, 356, 1107. 27. J. Pyka, J. Ilnicki, C. Altenbach, W.L. Hubbell and W. Froncisz, Biophys. J., 2005, 89, 2059. 28. C. Altenbach, W. Froncisz, R. Hemker, H. McHaourab and W.L. Hubbell, Biophys. J., 2005, 89, 2103. 29. K. Jacobsen, S. Oga, W.L. Hubbell and T. Risse, Biophys. J., 2005, 88, 4351. 30. J.M. Kim, C. Altenbach, M. Kono, D.D. Oprian, W.L. Hubbell and H.G. Khorana, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 12508. 31. G.E. Fanucci, K.A. Coggshall, N. Cadieux, M. Kim, R.J. Kadner and D.S. Caﬁso, Biochemistry, 2003, 42, 1391. 32. H.J. Merianos, N. Cadieux, C.H. Lin, R.J. Kadner and D.S. Caﬁso, Nat. Struct. Biol., 2000, 7, 205. 33. G.E. Fanucci, J.Y. Lee and D.S. Caﬁso, J. Am. Chem. Soc., 2003, 125, 13932. 34. D.P. Chimento, A.K. Mohanty, R.J. Kadner and M.C. Wiener, Nat. Struct. Biol., 2003, 10, 394. 35. G.E. Fanucci, J.Y. Lee and D.S. Caﬁso, Biochemistry, 2003, 42, 13106. 36. M. Kim, Q. Xu, G.E. Fanucci and D.S. Caﬁso, Biophys. J., 2006, 90, 2922. 37. G. Chang and C.B. Roth, Science, 2001, 293, 1793. 38. G. Chang, J. Mol. Biol., 2003, 330, 419. 39. C.L. Reyes and G. Chang, Science, 2005, 308, 1028. 40. A.H. Buchaklian, A.L. Funk and C.S. Klug, Biochemistry, 2004, 43, 8600. 41. A.H. Buchaklian and C.S. Klug, Biochemistry, 2005, 44, 5503. 42. J. Dong, G. Yang and H.S. Mchaourab, Science, 2005, 308, 1023. 43. E. Perozo, D.M. Cortes and L.G. Cuello, Nat. Struct. Biol., 1998, 5, 459. 44. E. Perozo, D.M. Cortes and L.G. Cuello, Science, 1999, 285, 73. 45. A. Gross, L. Columbus, K. Hideg, C. Altenbach and W.L. Hubbell, Biochemistry, 1999, 38, 10324. 46. L.G. Cuello, D.M. Cortes and E. Perozo, Science, 2004, 306, 491. 47. J.F. Cordero-Morales, L.G. Cuello, Y. Zhao, V. Jogini, D.M. Cortes, B. Roux and E. Perozo, Nat. Struct. Mol. Biol., 2006. 48. T.L. Kirby, C.B. Karim and D.D. Thomas, Biochemistry, 2004, 43, 5842. 49. C.B. Karim, T.L. Kirby, Z. Zhang, Y. Nesmelov and D.D. Thomas, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 14437. 50. J. Zamoon, F. Nitu, C. Karim, D.D. Thomas and G. Veglia, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 4747.

70

Electron Paramag. Reson., 2007, 20, 50–72

51. R.D. Nielsen, K. Che, M.H. Gelb and B.H. Robinson, J. Am. Chem. Soc., 2005, 127, 6430. 52. A.A. Frazier, M.A. Wisner, N.J. Malmberg, K.G. Victor, G.E. Fanucci, E.A. Nalefski, J.J. Falke and D.S. Caﬁso, Biochemistry, 2002, 41, 6282. 53. D. Marsh, B.G. Dzikovski and V.A. Livshits, Biophys. J., 2006, 90, L49. 54. A.J. Sharﬀ, L.E. Rodseth, J.C. Spurlino and F.A. Quiocho, Biochemistry, 1992, 31, 10657. 55. J.C. Spurlino, G.Y. Lu and F.A. Quiocho, J. Biol. Chem., 1991, 266, 5202. 56. M.I. Austermuhle, J.A. Hall, C.S. Klug and A.L. Davidson, J. Biol. Chem., 2004, 279, 28243. 57. J.M. Crane, C. Mao, A.A. Lilly, V.F. Smith, Y. Suo, W.L. Hubbell and L.L. Randall, J. Mol. Biol., 2005, 353, 295. 58. C. Motz, T. Hornung, M. Kersten, D.T. McLachlin, S.D. Dunn, J.G. Wise and P.D. Vogel, J. Biol. Chem., 2004, 279, 49074. 59. M.N. Oda, T.M. Forte, R.O. Ryan and J.C. Voss, Nat. Struct. Biol., 2003, 10, 455. 60. E. Rufener, A.A. Frazier, C.M. Wieser, A. Hinderliter and D.S. Caﬁso, Biochemistry, 2005, 44, 18. 61. A.A. Frazier, C.R. Roller, J.J. Havelka, A. Hinderliter and D.S. Caﬁso, Biochemistry, 2003, 42, 96. 62. S.C. Kohout, S. Corbalan-Garcia, J.C. Gomez-Fernandez and J.J. Falke, Biochemistry, 2003, 42, 1254. 63. N.J. Malmberg, S. Varma, E. Jakobsson and J.J. Falke, Biochemistry, 2004, 43, 16320. 64. C.C. Jao, A. Der-Sarkissian, J. Chen and R. Langen, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 8331. 65. A. Der-Sarkissian, C.C. Jao, J. Chen and R. Langen, J. Biol. Chem., 2003, 278, 37530. 66. M. Ramakrishnan, P.H. Jensen and D. Marsh, Biochemistry, 2006, 45, 3386. 67. I.R. Bates, J.B. Feix, J.M. Boggs and G. Harauz, J. Biol. Chem., 2004, 279, 5757. 68. I.R. Bates, J.M. Boggs, J.B. Feix and G. Harauz, J. Biol. Chem., 2003, 278, 29041. 69. A.A. Musse, J.M. Boggs and G. Harauz, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 4422. 70. R. Langen, J.M. Isas, H. Luecke, H.T. Haigler and W.L. Hubbell, J. Biol. Chem., 1998, 273, 22453. 71. J.M. Isas, Y.E. Kim, C.C. Jao, P.B. Hegde, H.T. Haigler and R. Langen, Biochemistry, 2005, 44, 16435. 72. J.M. Isas, R. Langen, W.L. Hubbell and H.T. Haigler, J. Biol. Chem., 2004, 279, 32492. 73. Y.E. Kim, J.M. Isas, H.T. Haigler and R. Langen, J. Biol. Chem., 2005, 280, 32398. 74. J.M. Isas, D.R. Patel, C. Jao, S. Jayasinghe, J.P. Cartailler, H.T. Haigler and R. Langen, J. Biol. Chem., 2003, 278, 30227. 75. D.R. Patel, J.M. Isas, A.S. Ladokhin, C.C. Jao, Y.E. Kim, T. Kirsch, R. Langen and H.T. Haigler, Biochemistry, 2005, 44, 2833. 76. D.H. Kweon, C.S. Kim and Y.K. Shin, Nat. Struct. Biol., 2003, 10, 440. 77. D.H. Kweon, C.S. Kim and Y.K. Shin, J. Biol. Chem., 2003, 278, 12367. 78. Y. Zhang and Y.K. Shin, Biochemistry, 2006, 45, 4173. 79. Y. Zhang, Z. Su, F. Zhang, Y. Chen and Y.K. Shin, J. Biol. Chem., 2005, 280, 15595. 80. Y. Chen, Y. Xu, F. Zhang and Y.K. Shin, EMBO J., 2004, 23, 681. 81. F. Zhang, Y. Chen, Z. Su and Y.K. Shin, J. Biol. Chem., 2004, 279, 38668.

Electron Paramag. Reson., 2007, 20, 50–72

71

82. Y. Xu, F. Zhang, Z. Su, J.A. McNew and Y.K. Shin, Nat. Struct. Mol. Biol., 2005, 12, 417. 83. K.G. Victor and D.S. Caﬁso, Biophys. J., 2001, 81, 2241. 84. K. Bhargava and J.B. Feix, Biophys. J., 2004, 86, 329. 85. J.F. Ellena, J. Moulthrop, J. Wu, M. Rauch, S. Jaysinghne, J.D. Castle and D.S. Caﬁso, Biophys. J., 2004, 87, 3221. 86. S.A. Jayasinghe and R. Langen, J. Biol. Chem., 2004, 279, 48420. 87. M. Margittai and R. Langen, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 10278. 88. S. Jeganathan, M. von Bergen, H. Brutlach, H.J. Steinhoﬀ and E. Mandelkow, Biochemistry, 2006, 45, 2283. 89. J.F. Hess, M.S. Budamagunta, J.C. Voss and P.G. FitzGerald, J. Biol. Chem., 2004, 279, 44841. 90. J.F. Hess, M.S. Budamagunta, P.G. FitzGerald and J.C. Voss, J. Biol. Chem., 2005, 280, 2141. 91. P.Z. Qin, K. Hideg, J. Feigon and W.L. Hubbell, Biochemistry, 2003, 42, 6772. 92. P.Z. Qin, J. Feigon and W.L. Hubbell, J. Mol. Biol., 2005, 351, 1. 93. T.E. Edwards, B.H. Robinson and S.T. Sigurdsson, Chem. Biol., 2005, 12, 329. 94. A.M. Shafer, T. Kalai, S.Q. Bin Liu, K. Hideg and J.C. Voss, Biochemistry, 2004, 43, 8470. 95. S. Burgard, M. Harada, Y. Kagawa, W.E. Trommer and P.D. Vogel, Cell Biochem. Biophys., 2003, 39, 175. 96. S. Delannoy, I.L. Urbatsch, G. Tombline, A.E. Senior and P.D. Vogel, Biochemistry, 2005, 44, 14010. 97. A.M. Shafer, V.J. Bennett, P. Kim and J.C. Voss, J. Biol. Chem., 2003, 278, 34203. 98. P.P. Borbat and J.H. Freed, Biological Magnetic Resonance, L.J. Berliner, S.S. Eaton, G.R. Eaton (Eds.), Kluwer Academic/Plenum Publishers, New York, 2000, 383. 99. M.D. Rabenstein and Y.K. Shin, Proc. Natl. Acad. Sci. U. S. A., 1995, 92, 8239. 100. C. Altenbach, K.J. Oh, R.J. Trabanino, K. Hideg and W.L. Hubbell, Biochemistry, 2001, 40, 15471. 101. H.J. Steinhoﬀ, N. Radzwill, W. Thevis, V. Lenz, D. Brandenburg, A. Antson, G. Dodson and A. Wollmer, Biophys. J., 1997, 73, 3287. 102. E.J. Hustedt, A.I. Smirnov, C.F. Laub, C.E. Cobb and A.H. Beth, Biophys. J., 1997, 72, 1861. 103. E.J. Hustedt, R.A. Stein, L. Sethaphong, S. Brandon, Z. Zhou and S.C. Desensi, Biophys. J., 2006, 90, 340. 104. Z. Zhou, S.C. Desensi, R.A. Stein, S. Brandon, M. Dixit, E.J. McArdle, E.M. Warren, H.K. Kroh, L. Song, C.E. Cobb, E.J. Hustedt and A.H. Beth, Biochemistry, 2005, 44, 15115. 105. D. Hilger, H. Jung, E. Padan, C. Wegener, K.P. Vogel, H.J. Steinhoﬀ and G. Jeschke, Biophys. J., 2005, 89, 1328. 106. G. Jeschke, C. Wegener, M. Nietschke, H. Jung and H.J. Steinhoﬀ, Biophys. J., 2004, 86, 2551. 107. G. Jeschke, A. Bender, T. Schweikardt, G. Panek, H. Decker and H. Paulsen, J. Biol. Chem., 2005, 280, 18623. 108. A.D. Milov, Y.D. Tsvetkov, F. Formaggio, M. Crisma, C. Toniolo and J. Raap, J. Am. Chem. Soc., 2001, 123, 3784. 109. B.G. Dzikovski, P.P. Borbat and J.H. Freed, Biophys. J., 2004, 87, 3504. 110. D. Zhang, A. Kiyatkin, J.T. Bolin and P.S. Low, Blood, 2000, 96, 2925.

72

Electron Paramag. Reson., 2007, 20, 50–72

111. A.A. Wegener, J.P. Klare, M. Engelhard and H.J. Steinhoﬀ, EMBO J., 2001, 20, 5312. 112. J.P. Klare, E. Bordignon, M. Doebber, J. Fitter, J. Kriegsmann, I. Chizhov, H.J. Steinhoﬀ and M. Engelhard, J. Mol. Biol., 2006, 356, 1207. 113. N.K. Kim, A. Murali and V.J. DeRose, Chem. Biol., 2004, 11, 939. 114. K. Sale, L. Song, Y.S. Liu, E. Perozo and P. Fajer, J. Am. Chem. Soc., 2005, 127, 9334.

Quantum Chemical Approaches to Spin-Hamiltonian Parameters BY FRANK NEESE Lehrstuhl fu¨r Theoretische Chemie, Institut fu¨r Physikalische und Theoretische Chemie, Universita¨t Bonn, D-53115 Bonn, Germany

1

Introduction

EPR spectroscopists have reﬁned the art of interpreting experimental results by means of an effective spin-Hamiltonian (SH) to a high degree of sophistication.1 The SH parameters are the principal outcome of an EPR experiment and represent a concise summary of the information content of the experiments. However, the power of the SH approach extends far beyond summarizing experimental results. The SH describes the physics of spin systems so well that it can be used creatively to design new experiments. The behaviour of the spin system can be simulated in advance through exact solutions of the quantummechanical equations of motions in the SH formalism. In this respect it is of major utility that the SH is so simple – it usually works in a low-dimensional Hilbert space which is only spanned by the (effective) spin degrees of freedom of the system under investigation. Due to this simplicity exact solutions are relatively easy to generate with ordinary computational hardware, or, in many cases, just with paper and pencil. The price to pay for this invaluable convenience is that the SH contains adjustable parameters, the values of which are determined from ﬁtting them to experimental measurements. Clearly, if the SH formalism is well designed, as in fact it is, there must be a direct connection between the SH parameters and the microscopic physics which governs the behaviour of molecules under all conceivable circumstances. Unfortunately, since exact solutions to those microscopic (relativistic) wave equations are, with very few exceptions, out of the question, a great many different pathways exist which match the SH parameters to the outcome of quantum chemical calculations at various levels of sophistication. The purpose of this Chapter is to describe some of the recent developments in this ﬁeld and to point the interested reader to the literature on the subjects covered. The reader is also referred to a recently published book which contains many valuable chapters dealing with various aspects of the material covered here in more detail.2 Plane-wave approaches will not be reviewed in this Chapter. Electron Paramagnetic Resonance, Volume 20 r The Royal Society of Chemistry, 2007 73

74

2

Electron Paramag. Reson., 2007, 20, 73–95

Theoretical Aspects

2.1 General Theory. – The problem at hand involves the calculation of the behaviour of a N-electron system under the inﬂuence of a homogenous, timeindependent magnetic ﬁeld. The time dependence introduced by the oscillating B1 ﬁeld and the associate dynamics of the spin-system are best dealt with at the level of the SH and need not be covered in the microscopic treatment. Reﬁned experiments such as the linear-electric ﬁeld effect3 require the introduction of further terms in the Hamiltonian, which will not be treated in this chapter. The rigorous starting point should then be a fully relativistic N-particle Hamiltonian which contains the external magnetic ﬁeld (or the associated vector potential, usually represented in the Coulomb gauge). While there are still adjustments in the fundamental relativistic N-particle theory, in EPR (and NMR) spectroscopy one does not need to go further than the level of the DiracCoulomb-Breit (DCB) Hamiltonian.4 In this approach, the fundamental relativistic one-particle Dirac equation is summed over all N-electrons and the electron-electron interaction is treated by the (nonrelativistic) Coulomb-repulsion together with the Breit correction. The relativistic one-particle wavefunctions are so-called four-component spinors which are composed of a so-called ‘large’ (electron like) and a ‘small’ (positron like) component which both contain two contributions that represent the two possible spin-cases. Such four component objects and the associated N-particle wavefunctions are hard to deal with and one looks for simpliﬁcations. There are various procedures to decouple the four-component equations into two components. Historically, the Breit-Pauli approach (an expansion in c2) is the most well-known procedure. However, it leads to singularities and can not be used in variational approaches. Presently, the two most popular approaches to the reduction of the four-component equations into two-components are the Douglas-Kroll-Hess (DKH)5–9 method and the zero’th order regular approximation (ZORA).10–14 At this level, the relativistic effects show up in two ways: ﬁrst, there are the kinematic effects. They are usually called ‘scalar relativistic effects’ and are spin-independent. Secondly, there are spin-dependent effects which are referred to as ‘spin-orbit coupling’ (SOC). While this division is extremely convenient, it has been shown that there is no unique way of separating the spin-dependent and spin-independent effects and therefore a residual ambiguity has to be tolerated.15 From this point on, two possible routes can be taken. Either the scalarrelativistic and SOC effects are treated together which leads to the so-called ‘two-component’ approaches. In this case, the one-particle wavefunctions used to construct N-electron state functions are mixtures of spin-up and spin-down components.16 Alternatively, the SOC can be introduced at a later stage after solving spin-free wave equations which may or may not contain the scalar relativistic effects (they are negligible for lighter elements, perhaps up to Argon and small for most EPR and NMR properties, perhaps up to Kr). The latter approach is then referred to as one-component or (if scalar relativistic effects

Electron Paramag. Reson., 2007, 20, 73–95

75

are neglected) as ‘non-relativistic’. The latter term might be somewhat misleading since the SOC term is relativistic in origin. In any case, the expansion of a two-component approximation to the DCB Hamiltonian gives rise to all terms that are necessary to predict all SH parameters. That is, it contains all necessary terms to describe the Zeeman effect (g-tensor), the hyperﬁne couplings (HFC), the zero-ﬁeld splitting (ZFS) and the nuclear quadrupole coupling (NQC). However, full four-component calculations of EPR properties are certainly feasible and have recently been pursued by Quiney and Belanzoni17 (see also recent results for NQC’s at the four component level18,19). Given the appropriate terms in the Hamiltonian, one now has to face the Nelectron problem and the challenge of electron correlation. This problem is usually either tackled by approximations based on Hartree-Fock (HF) theory, such as many-body perturbation theory (MBPT), coupled-cluster (CC) theory or conﬁguration interaction (CI). Alternatively, the large success of density functional theory (DFT) during the past 15 years has led to much activity in the ﬁeld of DFT approaches to EPR and NMR parameters and it is these approaches that currently dominate the ﬁeld of theoretical EPR spectroscopy. 2.2 Spin-Free Approaches. – In the spin-free approaches, the wave equations which are originally solved do not contain the electron spin. Consequently, the total spin of the system is a good quantum number and the approximations to N-particle wavefunctions should be spin-eigenfunctions. The starting point of such approaches is almost invariably the Born-Oppenheimer (BO) Hamiltonian, which just contain the leading electrostatic interactions between the electrons and nuclei together with the kinetic energy of the electrons. Addition of scalar-relativistic effects to the BO Hamiltonian is straightforward and in practical applications usually amount to a modiﬁcation of the one-electron part of the BO Hamiltonian (however see ref. 20 for a treatment of two-electron contributions to the DKH procedure). All other ﬁeld- and spin-dependent terms are introduced by some form of perturbation theory. Historically, the most popular approaches to EPR and NMR parameters expanded the perturbed N-electron wavefunctions in a set of a complete and orthonormal N-electron eigenstates |aSMi (S is the total spin, M-the projection and a is a compound label which contains all information to properly identify a given state) of the BO Hamiltonian (see refs 21–23 for details). Despite the fact that such an expansion must always be possible, it is not self-evident that the resulting perturbed energies or the associated effective Hamiltonian have the same form as the terms in the SH. However, at least to second order in perturbation, this has been accomplished through McWeeny’s partitioning theory22,23 for all terms in the SH; the most complicated case was represented by the ZFS which has been expressed in standard SH form for the ﬁrst time in 1998.21 The SOS approaches lead to explicit expressions of EPR and NMR parameters as long as the state under consideration is orbitally nondegenerate or only nearly-degenerate. A convenient definition of what is meant by nondegenerate is, that the only electronic levels which are populated to any

76

Electron Paramag. Reson., 2007, 20, 73–95

significant extent under the conditions of the experiment are the 2S þ 1 levels of the multiplet with total spin S which is under study. If another multiplet is nearby, the SH formalism becomes invalid (or at least the meaning of the parameters become obscure and such cases are best dealt with extensions of the SH formalism in order to keep the physics that one wishes to model transparent). These sum-over-states (SOS) approaches have formed the cornerstone of EPR and NMR theory for a long time in the form of Ramsey’s theory of NMR parameters (reviewed e.g. by Pyykko¨24) and the Abragam and Price theory25 of the SH in EPR. The Abragam and Pryce theory was formulated in terms of ligand-ﬁeld theory and concise general treatments have been given by McWeeny22,23 and Harriman.26 A famous example of SOS expressions is also the g-tensor theory of Stone,27 which is formulated in terms of single-determinantal representations of the N-electron state functions. In the SOS approaches, ﬁrst- and second-order properties are readily identiﬁed. First order properties are determined by the reference wavefunction alone while second-order properties involve an (inﬁnite) summation over excited states as described above. First-order EPR properties are the NQC, the Fermi contact and spin-dipolar contributions to the HFC, the reduced mass and gauge corrections to the g-tensor and the direct spin-spin contribution to the ZFS. Second order properties are the SOC contribution to the HFC and the ZFS and the orbital-Zeeman/SOC contribution to the g-tensor. Convenient as these SOS approaches are, they are of limited utility in practical calculations of second-order properties since it is rarely possible to calculate more than a few dozen low-lying states explicitly and truncation of the inﬁnite sum after these few dozen terms renders the convergence of the perturbation sum uncertain. In the case of transition metal complexes some arguments have recently been brought forward to suggest that the convergence of the SOS expansion is probably favourable.28 SOS approaches to the g-tensor have frequently been pursued. The most rigorous work was initiated by Lushington and Grein29–32 at the multireference conﬁguration interaction (MRCI) level and serves well for the interpretation of the EPR properties for small molecules.29–33 Similar work has recently been pursued by Brownridge et al.34 and by this author35 at the levels of MRCI and MR-MBPT. The ZFS and HFC second-order contributions have also been treated by this methodology.28,36,37 However, approaches which are not plagued by the uncertainties of SOS expansions are clearly attractive. Recently, it became evident that linear response theory (LRT38–42) provides a very convenient framework for reformulating the theory of SH parameters. LRT is just one realization of a family of methods which are all formulated in a similar spirit. If time does not explicitly occur in the equations (which is not necessary for EPR and NMR parameters), these methods can also be called ‘analytic derivative approaches’43 or in the framework of HF and DFT methods they are known as ‘coupledperturbed self-consistent ﬁeld’ (CP-SCF44–47) methods or ‘double-perturbation theory’ (DPT48) respectively. All of these acronyms stand for computational

Electron Paramag. Reson., 2007, 20, 73–95

77

methods which provide identical results and it is a matter of taste which framework one prefers. All these methods start from the realization that the perturbing ﬁeld- and spin-dependent operators lead to shifts of the total energies which are typically orders of magnitude smaller than the spacing of the eigenvalues of the BO Hamiltonian. Since the leading SH parameters always occur in bilinear form it is the second partial derivatives with respect to two perturbations that have to be studied. Thus, as one ‘turns on’ these perturbations, say lh^l and kh^k one can expand the total energy in a Taylor series which can be safely truncated after @2 E the bilinear term, i.e. @l@k . Through comparison with the equivalent l¼k¼0

expansion of the SH, the relevant terms of the SH can be identiﬁed. For example, perturbation lh^l may be the interaction of the x-component of the static magnetic ﬁeld with the orbital moment of the electrons and perturbation lh^l may be the z-component of the SOC operator. Although, this has at most a few times been spelled out in the EPR literature, it became evident that all these second derivatives which involve one-electron or effective one-electron perturbing operators can be written in a uniﬁed form which embraces a wide variety of seemingly very different electronic structure methods: D E X @P D E X @ 2 E mn mh^k n ¼ P mh^lk n þ ð1Þ @l@kl¼k¼0 mn mn @l mn Here a set of basis functions {jm, jn, . . . } has been introduced in which the orbitals {cp, cq, . . . } are expanded. The quantities P mn represent the electron (‘þ’) or the spin density (‘’) of the state under investigation and the oneelectron operators h^lk and h^k represent the second- and ﬁrst derivative of the total Hamiltonian with respect to the perturbation order parameters l and k. The two terms in eq (1) replace the ﬁrst- and second-order contributions in the SOS expansions. In fact, a clean connection between the two-formalisms can be made rather straightforwardly at the CI level.49 The only place where the speciﬁc electronic structure method enters is the calculation of the density matrices P. For variational electronic structure methods (HF, DFT, MCSCF), with the N-electron wavefunction |Ci approximating the state of þ interest, these may be deﬁned by the expectation values P pq ¼ hC|apaaqa þ þ apbaqb |Ci where aps and aqs are the creation and annihilation operators for an electron with spin s(¼a,b) in molecular orbital p. For non (MBPT, CC) or partially-(CI) variational methods, the density matrices are the so-called ‘response-densities’ which are well-known from the gradient theory of correlated ab initio methods.50 The most difﬁcult terms are represented by the derivatives of the electron- or spin-density matrices with respect to the external perturbation as this almost invariably involves the non-trivial solution of a large set of linear equations (vide infra). 2.3 Two-Component Approaches. – A conceptionally different approach to magnetic resonance parameters is offered by so-called two-component methods. In these methods the SOC is not introduced as a perturbation but

78

Electron Paramag. Reson., 2007, 20, 73–95

is included in the orbital optimization step. Consequently, the MOs become complex valued and the SOC is treated to ‘inﬁnite’ order already at the SCF level. The scalar relativistic effects can be included at the same time, of course, without any additional computational cost. The calculation of the g-tensor then only involves ﬁrst-order perturbation theory with the Zeeman operator.51,52 The SOC contribution to the HFC can be similarly treated.53 The ﬁrst such method which was applied to magnetic resonance calculations was the ZORA approach by van Lenthe and co-workers as implemented in the ADF code.53 This approach applies to a single unpaired electron and optimizes the twocomponent spinor of this electron in a spin-restricted approach which allows the program to maintain the correct double-group symmetry. However, the sometimes important effects of spin-polarization are lost. More recently, alternative approaches have been formulated on the basis of the DKH Hamiltonian.54–56 A problem common to the two-component DFT methods is that in the presence of SOC, the spin-quantization axes are no longer obvious. Therefore, the spin-densities that enter the expressions of the exchange-correlation energy are also no longer well deﬁned. Van Wu¨llen has discussed the relative merits of ‘collinear’ versus ‘non-collinear’ treatments of the XC terms and has concluded that the non-collinear approach is more justiﬁed.16 Since in two-component methods the matrices to be diagonalized are a factor of twolarger and complex valued it can be estimated that ultimately they will be about an order of magnitude more expensive than spin-free SCF approaches. Nevertheless, once the formal problems of loss of spin-polarization versus symmetry breaking are satisfactorily solved, it can be expected that two-component methods become attractive for problems which involve very large SOC interactions, such as present in compounds with heavy elements, or in the presence of near degeneracies where a ﬁrst-order treatment of the SOC is no longer sufﬁcient. 2.4 The Spin-Orbit Coupling Operator. – The question of how to best approximate the SOC in molecular calculations has been found to be surprisingly controversial (reviewed by Patchkovskii and Schreckenbach57). Here, the discussion in ref 58 will be followed. From the Dirac equation, it is deduced that: 2 Veff a ! h^SOC ¼ ^s rV ^p ð2Þ 2 where a is the ﬁne structure constant (B1/137 in atomic units), sˆ is the vector operator for the electron spin and p is the momentum operator. In this form the SOC operator has the appearance of a single-particle operator and can for molecular applications be written: X Veff Veff ^si ^zi h^SOC ¼ ð3Þ i

The sum is over electrons. In a DFT framework the potential V may be interpreted as the Kohn-Sham potential. In this case, one obtains a

79

Electron Paramag. Reson., 2007, 20, 73–95

computationally attractive formulation of DFT that has been implemented in most DFT programs. Another way to arrive at a single particle SOC operator is the semi-empirical approach used by many authors and systematized by Koseki.59–62 In this case, one uses semiempirical effective nuclear charges and writes the SOC as: Zeff h^SOC ¼

X a2 X ðAÞ Z eff X Zeff A ^l ^si ^si ^zi i 3 2 A jri RA j i i

ð4Þ

Here, ri denotes the position of the i’th electron, RA the position of the A’th (A) nucleus with semiempirical effective nuclear charge Z eﬀ the angular A and ıˆ i momentum of the i’th electron relative to nucleus A. The Zeﬀ operator shows relatively good performance for 2p and 3d elements. Its drawback is obviously that the effective nuclear charges are somewhat system and basis-set dependent. A third way, not discussed here, is to employ spin-orbit pseudopotentials in the SOC calculation.63 Finally, in the authors’ opinion the most accurate and attractive method is the spin-orbit mean-ﬁeld (SOMF) approach invented by Hess et al.7,64 Here, one treats the two-particle Breit-Pauli SOC operator (or its DKH or ZORA analogue) by a mean-ﬁeld method which resembles HF theory. As HF theory gives the total energy accurate to B98–99%, the SOMF operator is expected (and conﬁrmed) to give SOC energies accurate to B98–99%, which means that SOC effects are usually predicted to within a few wavenumbers of their Breit-Pauli values. The SOMF operator also takes the form of an effective one-electron operator: X SOMF SOMF ^si ^zi h^SOC ¼ ð5Þ i

with matrix elements:

58,65

" E a2 X D ðAÞ E D E D SOMF ^ 3 m^z ZA m^l r3 r mn l n ¼ n 12 iA 12 r 2 A þ32

X kt

Pþ kt

nD

# E D Eo ^ 3 ^ 3 mkl12 r12 tn þ kn l12 r12 mt

ð6Þ

Here ^l12 is the angular momentum of electron ‘1’ relative to electron ‘2’. It is evident that the evaluation of the matrix elements of this operator requires a higher computational effort than in the Veﬀ and Zeﬀ cases, since a large number of complicated two-electron SOC integrals must be calculated. An efﬁcient implementation which largely avoids computational bottle-necks has recently been discussed.58 Malkina et al. have pointed out that the Veﬀ SOC operator does not include the important spin-other-orbit (SOO) interaction and have attributed this omission to the fact that the Veﬀ operator tends to overestimate SOC effects.63 The subject has recently been reinvestigated.58 It was pointed out that the SOO contribution to the Coulomb term (the second term in equation 6) vanishes exactly while it gives an exchange contribution (the last two-terms in eq 6) of twice the spin-same-orbit (SSO) contribution which is the origin of the

80

Electron Paramag. Reson., 2007, 20, 73–95

factor 3/2. Moreover, it was numerically demonstrated that the Veﬀ operator introduces an exchange contribution of the wrong sign, which, together with the effect of the neglect of the SOO interaction, accounts for overestimation of SOC matrix elements of up to 30%. Subsequently, an attempt to introduce the SOO interaction into DFT treatments was reported, but has not been found systematically to improve g-tensor predictions.66 3

Application to EPR Spin-Hamiltonian Parameters

3.1 g-Tensor. – In order to put the spin-free formalism outlined above on more solid ground, the calculation of g-tensors using HF or DFT theory which can be treated together will be brieﬂy discussed and will be subsumed under the term ‘SCF’ theory. Following the treatment in ref. 67 the spin-unrestricted HF or Kohn-Sham (KS) SCF equations for the orbitals and their energies read: ( Z 1 ! 2 X ZA rðr0 Þ r þ dr0 2 jr R A j j r r0 j A ) ð7Þ X dEXC ½r ^ cHF þcDF K js cis ðrÞ ¼ eis cis ðrÞ drðrÞ js with: jcis i ¼

X

csmi jm s

ð8Þ

nsi csmi csni

ð9Þ

m

Psmn ¼

X i

rðrÞ ¼

X Pamn þ Pbmn jm ðrÞjn ðrÞ

ð10Þ

mn

Here, the terms represent the kinetic energy of the electron, the electron-nuclear attraction, the Coulomb repulsion, the DFT exchange correlation potential and the Hartree-Fock exchange respectively. (cDF ¼ 1, cHF ¼ 0 corresponds to pure DFT, cDF P ¼ 0, cHF ¼ 1 to pure HF theory and 0 o cHF o 1 to hybrid DFT). r(r) ¼ mn Pþ mnjm(r)jn(r) is the total electron density, EXC is the exchangecorrelation functional, K^ js is the non-local Hartree-Fock exchange operator for the spin-orbital js and cis is the i’th spin-orbital with orbital energy eis. In the case of the g-tensor, one has three ﬁrst-order contributions which are all small.67 The more challenging term, is the second-order contribution, which also dominates the g-tensor. In order to calculate the density matrix derivative one ﬁrst has to solve the CP-SCF equation with respect to an imaginary perturbation represented by the orbital Zeeman term: X Aias; jbs0 Ujbs0 ¼ Vias ð11Þ ias;jbs0

81

Electron Paramag. Reson., 2007, 20, 73–95

D E Vais ¼ csa lðOÞ csi þ ðGIAO termsÞ

ð12Þ

Aias; jbs0 ¼ esa esi dias; jbs0 þ cHF dss0 ½ðibjjaÞ ðij jabÞ

ð13Þ

ðpqjrsÞ ¼

Z Z

cp ðr1 Þcq ðr1 Þcr ðr2 Þcs ðr2 Þr1 12 dr1 dr2

ð14Þ

(real orbitals assumed). Here i,j represent occupied orbitals in the reference determinant, and a,b empty orbitals. The coefﬁcients U are the ﬁrst-order orbital coefﬁcients to be determined, l(O) is the angular momentum relative to the gauge origin and the Vais right-hand sides may contain additional terms, if the orbitals or basis functions have been chosen to be magnetic ﬁeld-dependent as in the individual gauge for localized orbitals (IGLO) and gauge including atomic orbitals (GIAO) approaches to the gauge problem (vide infra). It is evident that the magnetic response matrix A is diagonal in the absence of Hartree-Fock exchange (in general: in the absence of nonlocal potentials) and that the linear equation system is then trivially solved by Uias ¼ Vais/(esa esi ) which makes ‘pure’ DFT calculations of magnetic properties particularly efﬁcient. Once the coefﬁcients U are known, the electron- and spin-density derivatives are readily calculated by:

@P 0 Ua a mn 0 Ub b bþ c ð15Þ c c ¼ caþ Ua 0 Ub 0 @B B¼0

The perturbed electron density may be used to calculate NMR chemical shifts while the perturbed spin-density enters the g-tensor calculation. In addition, the perturbed electron density is contracted with integrals over the nucleus-orbit interaction operator while the g-tensor is contracted with integrals over the spin-orbit operator. Thus, a close analogy between the ﬁelds of theoretical EPR and NMR spectroscopy becomes evident. The formalism outlined above is particularly compact and makes it straightforward to efﬁciently compute EPR properties at various levels of electronic structure theory with minimal modiﬁcations of existing code. This approach has been implemented in the EPRNMR module of the ORCA program68 which computes various EPR and NMR spectral parameters at the spin-unrestricted HF and DFT levels.67,69,70 A spin-restricted DFT treatment has been presented by Rinkevicius et al. and gives similar results for most molecules.71 At the MCSCF level, the ﬁrst implementation of an equivalent theory was reported by Vahtras and coworkers38,40,41 in the Dalton code. MBPT and CC methods have recently been implemented into the ACESII code by Gauss and co-workers72 and these methods for the ﬁrst time give access to highly accurate correlated ab initio calculations of electronic g-tensors. At the MRCI level, the formalism was ﬁrst implemented by the present author into the ORCA code. Although, the orbital relaxation has not yet been included in this variant, the CC calculations have shown it to be small.49

82

Electron Paramag. Reson., 2007, 20, 73–95

A second subject to be addressed is the treatment of the gauge dependence of the g-tensor. Although it is presumably small if a ‘reasonable’ global origin is chosen (e.g. the centre of electronic charge73), it is certainly the best to eliminate any dependence of the computational results on the choice of origin. The IGLO method which is popular in the ﬁeld of NMR calculations does not appear to be very successful in this respect. Malkin and co-workers have argued that the separate localization of spin-up and spin-down orbitals introduces artifacts and symmetry breaking into the calculation74 and these ﬁndings have been conﬁrmed in the author’s experience. Alternatively, the gauge dependence can be eliminated by GIAO’s. Indeed, the early implementations by Schreckenbach and Ziegler75 and the two-component method of van Lenthe et al.,51 both in the framework of the ADF program, used GIAO’s. The Respect program has also been reported to feature GIAO’s but details of the implementation have not been provided. Unless special techniques are used, the calculation of the extra terms arising from the ﬁeld-dependent basis functions comes at a significant computational expense. Directly programmable equations have recently been given in the NMR framework by Friesner and co-workers.76 As other details of g-tensor calculations approach maturity, the introduction of GIAO’s in an efﬁcient program will probably turn out to be worthwhile. 3.2 Hyperﬁne Couplings. – The theory of the hyperﬁne couplings is a rather well studied ﬁeld in quantum chemistry as far as the isotropic Fermi contact interaction is concerned. A recent review by Improta and Barone provides the state of the art of the ﬁeld as applied to organic radicals.77 In transition metal complexes, the situation appears to be relatively well understood too. The work of Munzarova and Kaupp78,79 as well as ref. 70 have shown that DFT methods tend to underestimate the core-level spin-polarization systematically. The amount of underestimation depends apparently on the metal and the number of unpaired electrons and amount to 10–80%.80–95 Scalar relativistic effects also start to become significant for transition metals and can now be treated in the ZORA53 and DKH96 frameworks. Filatov and Cremer97 have recently discussed the appropriate modiﬁcations for the IORA98,99 method. The dipolar contribution to the HFC is available through measurements of frozen solutions or single crystals. The performance of DFT appears to be mixed here. The dominant spin centres in a molecule appear to be well treated by DFT. For small couplings of remote nuclei (superhyperﬁne coupling) and centres without ‘ﬁrst-order’ spin population the results are usually somewhat worse and deviations of up to a factor of two have been observed in a number of applications.82,87–90,92 However, the orientation of the HFC tensors appears to be rather well predicted by DFT methods which makes them an invaluable aid in the interpretation of high-resolution EPR experiments which give access to a large number of weakly coupled nuclei.82 The calculation of the SOC contribution to the HFC, although well known in ligand ﬁeld theory25 and the earlier EPR literature100 is a more recent development in modern DFT programs.70,101–103 The theory closely follows the methods discussed above for the g-tensor and has been fully described in refs 101 and 70 and is available in

Electron Paramag. Reson., 2007, 20, 73–95

83

the ORCA code. Additional information and similar implementations are available in the papers by Kaupp and co-workers and the Respect program.102,103 The two-component framework was discussed by van Lenthe et al. within the ZORA scheme as implemented in the ADF program.53 The ORCA implementation also gives access to the ZORA method but instead of using a two-component scheme, the SOC is introduced by perturbation/response theory.58,70,84,87,88 Since transition-metal HFCs may contain contributions of similar magnitude but opposing signs from all three sources of HFC, quantitative accuracy is difﬁcult to achieve. As with transition metal complex gtensors, the SOC contributions are likely to be significantly underestimated by DFT methods, dipolar HFCs are probably reasonable if hybrid functionals are used and the Fermi contact term suffers from the defect described above. A detailed case study in relation to ab initio methods is available.28 More recent developments in the ﬁeld of HFCs are the use of MR-MBPT methods104,105 as well as the development of alternative spin-density operators which are designed to side-step the difﬁculties associated with the Fermi contact term which probes just a single point in space.106,107 3.3 Zero-Field Splittings. – The ZFS is the most recent property to receive significant attention from the quantum chemical community. In general, the ZFS tensor contains contributions from the direct dipolar electron-electron spin-spin coupling (to ﬁrst-order) and from the SOC (to second-order).26 Traditionally, the former is expected to dominate for organic triplets while the latter has usually been held solely responsible for the ZFSs of transition metal complexes. Vahtras and co-workers have implemented a MCSCF method to deal with the spin-spin coupling and have also discussed the SOC contribution.108–110 A number of applications have demonstrated rather good accuracy of the CASSCF method and have conﬁrmed the smallness of the SOC contribution for p-triplets.110–113 The spin-spin term has been implemented in a DFT framework ﬁrst by Petrenko et al. and applied to CH2114 and more recently by the author.36 Our initial experiences with DFT calculations of the spin-spin term are mostly positive. The SOC term is somewhat less straightforward to treat since the general theory21 shows that contributions from excited states with DS ¼ 0, 1 contribute and it is not evident how to best treat them in a DFT framework. The ﬁrst DFT method has been proposed by Pederson and Khanna115 in the spin-polarized framework and has been applied to large single molecule magnets with surprisingly good success.116 Upon application to mononuclear transition metal complexes it became evident that the DFT methods tend to underestimate the ZFS by up to a factor of two.36,81,86,117,118 An alternative DFT method which works with spin-eigenfunctions has been proposed more recently and shown to give results which are very similar to the Pederson-Khanna method.36,81,86,118 Most recently, the SOC contributions have been incorporated into a MRCI program and applied to atoms and small molecules with very encouraging results.37 In the framework of the SORCI method119 these calculations can also be applied to somewhat larger molecules. A recent study on [Mn(acac)3], a d4 system with ground state

84

Electron Paramag. Reson., 2007, 20, 73–95

spin S ¼ 2, has challenged the traditional view of the smallness of the direct spin-spin contribution. At the DFT as well as the CASSCF level (in a meanﬁeld approximation), the spin-spin contributions to the ZFS exceeded 30% of the ﬁnal values and also improved the agreement between DFT calculations and experiments.36 In addition, it was shown, that the spin-ﬂip contributions will in many cases dominate over the same-spin contributions in transition metal complexes and must be properly taken into account if quantitative accuracy is requested.36 The DFT methods were also applied to a series of Mn(II) complexes with mostly encouraging results.120 In this case it also became clear, that the spin-spin contributions can certainly not be neglected in transition-metal complexes which invalidates many ligand-ﬁeld based treatments which take neither account of spin-ﬂip nor of spin-spin contributions. Two component DFT methods have also been developed55,121 and appear to give similar results to the perturbation theoretical treatments under the same conditions under which the two-component g-tensor calculations agree with their perturbation theoretical counterparts. 3.4 Quadrupole Couplings. – The nuclear quadrupole coupling arises from the interaction of the nuclear quadrupole moment with the electric ﬁeld gradient. It is a relatively simple ﬁrst-order property which can be evaluated from the knowledge of the ground state electron density.122 Results with DFT methods appear generally to be reasonable for 2H123 and 14N80 and possibly also 17 88,124 O as well as 57Fe81,84,86,91,93,94,125–131. Since the ﬁeld-gradient operator features a steep r3 dependence it is important to treat the region around the core somewhat more carefully by using basis sets that are sufﬁciently ﬂexible to describe small aspherical core-density distortions which may have large effects on the calculated NQC constants. For transition metals and heavier elements, scalar relativistic effects also start to become significant and should be considered. A method to do this has been proposed by van Lenthe and Baerends in the ZORA framework.132 In the DKH framework, the so-called picture change consistent calculation of EFGs has been discussed by a number of authors.18– 19,133–136 A recent implementation has been provided by Malkina et al.136 and in another recent study.19 The latter study demonstrated for the ﬁrst time, that with the DKH method, even if only taken to second-order in the external potential, accurate four-component results can be almost quantitatively recovered. The ZORA method was shown to not reach the same level of agreement with reference four component results. At least in our study, the scaling proposed by van Lenthe and Baerends did not improve the results relative to the four component reference values.19 The results of the two implementations by Malkina et al.136 and Neese et al.19 are only in moderate agreement which may be related to basis-set effects. 3.5 Functional and Basis Set Effects. – The dependence of DFT results on the functional and basis set used was studied by several authors.63,67,75,137 The general result was that g-tensor calculations are only moderately basis set dependent if the basis set is at least of double-z plus polarization quality. In

Electron Paramag. Reson., 2007, 20, 73–95

85

terms of Gaussian basis sets, the EPR-II basis set138 appears to perform well for ﬁrst row atoms. The performance of most GGA functionals is generally very similar (however, see ref. 139). No consensus has apparently yet been reached whether hybrid functionals represent a deﬁnitive improvement of GGA’s. At least in the experience gained in this laboratory, the question can be positively answered-hybrid functionals, with B3LYP being the standard functional of this type, do give somewhat better g-tensor predictions for organic radicals as well as transition metal complexes. The PBE0 functional is of the same quality. A number of more modern functionals with dependence on the Laplacian of the density and the kinetic energy were tested by Arbuznikov et al.140 but no dramatic improvement over the performance obtainable with standard GGA or hybrid functionals was obtained. Slightly improved results were reported by the same authors for so-called ‘local hybrid functionals’ after ﬁtting of the fraction of Hartree-Fock exchange.141 The subject certainly deserves further future investigations. Consistent improvements of the quality of the results, in particular for transition-metal complexes, are presumably only achievable with density functionals which have an improved physical content and perhaps also a dependence on the current density. 3.6 Environment Effects. – Realistic modelling of environment effects has found increasing attention recently. The simplest approach is to model the environment of a radical or transition metal complex through dielectric continuum approaches. The studies of Cioﬁni et al.142–145 and Rinkevicius et al.146 used the polarisable continuum (PCM) model. The former study did not introduce solvent speciﬁc terms in the response equations, while the latter did. In a later study,147 the conductor like screening model (COSMO) was evaluated in some detail. It was pointed out, that the solvent operator is local and therefore, no solvent terms enter the magnetic response equations unless the basis functions depend on the external perturbation.147 Thus, in this case, all solvent effects are already included in the ‘solvated orbitals’ and their energies. All studies agreed that the major part of the unspeciﬁc solvent effect is well captured by the continuum approaches but that speciﬁc interactions, in particular hydrogen bonding, require a more detailed treatment. Most studies so far have used a supermolecule approach.123 In many cases, in particular in conjunction with a polariazable continuum, results could be obtained which are quite close to the experimental data. A recent study reported errors of less than 100 ppm for a series of nitroxide radicals if large basis sets (IGLO-III plus diffuse functions), the B3LYP functional, the supermolecule approach and the COSMO model are used.147 In this paper, the self-consistent generalization of the COSMO-RS model (RS ¼ realistic solvent) was reported. It was shown that a substantial fraction of the hydrogen bonding effects can, in fact, be included in this approach without the need to resort to a supermolecule model.147 A more detailed modelling of the environment is offered by QM/MM approaches which is particularly attractive for studies of proteins, DNA and other macromolecules. Magnetic resonance calculations within a QM/MM framework have recently been reported for compound I in cytochrome P45086 and the

86

Electron Paramag. Reson., 2007, 20, 73–95

active site of copper-plastocyanin.80,148,149 At least in the latter case, where experimental data abundantly exists, the results were improved relative to gas phase or dielectric continuum calculations. A promising approach to accurately model solvent effects is the use of Carr-Parinello molecular dynamics calculations. In this case, one takes snapshots of different molecular conformations during the simulation run and computes the properties at these geometries. If a sufﬁciently large number of points is used such that meaningful statistics can be obtained, one can obtain detailed insights into the behaviour of the system under investigation. The cost of such calculations is high, but they become increasingly feasible as exempliﬁed in the work of Asher et al. on solvated quinones.150

4

Application Studies

There have been numerous applications of EPR parameter calculations being carried out in recent years and space only allows the presentation of a crude overview of the ﬁeld. 4.1 Small Molecules. – In the ﬁeld of small molecules the numerous studies and pioneering work of Grein and co-workers on the g-values and HFCs of small radicals at the MRCI level stands out.29–34,151–158 Theoretical isotropic HFCs of 29Si, 31P, 33S were recently studied in detail by Hermodilla et al. with DFT methods.159,160 Kaupp et al. studied the spin-density distribution of hydrogen atoms trapped in silasequiooxanes.161 Knight et al. studied diatomic transition metal oxides experimentally and with ab initio methods.162 Prassad studied the ﬁne- and hyperﬁne-structures of NCO, CNO, N31, N3 and Nd using multireference ab initio techniques.163,164 The cases of CCCD and HCCS were considered in detail with ab initio methods by Peric et al. and Mladenovic et al.165,166 Fþ S centres on the surface of MgO were recently studied with two-component DFT methods.167 The HFCs of XeF were studied with relativistic DFT methods and high-resolution experiments.168 Numerous studies used information available on small molecules for calibration purposes.19,28,34,35,37,49,56,58,66,70–72,96,97,102,103,110,111,113,134,139,141,146,159,160 4.2 Organic and Biological Radicals. – Explosive activity in the ﬁeld of organic and biological radicals has been noted but it is difﬁcult to claim a comprehensive overview of the ﬁeld. Nevertheless, a number of studies of amino acid derived radicals74,145,169–176, DNA bases and derivatives,177–180, nitroxide radicals146–147,181–184, semiquinones,123,150,185–190 ﬂavin derived radicals191 and others,192 as well as some triplets and diradicals109–110,112,193–195 have been quoted. The vast majority of the cited works use DFT methods with emphasis on hybrid density functionals of which B3LYP is the most popular. Many studies include solvent effects through dielectric continuum or more sophisticated approaches as eluded to above. The computation of isotropic HFCs is still the most intensively pursued subject but g-tensor computations become

Electron Paramag. Reson., 2007, 20, 73–95

87

more and more common and are now at a stage where quantitative accuracy (B100 ppm deviation from experiments) is observed more frequently. The vast majority of studies are focused on S ¼ 1/2 species but given the latest developments for high-spin systems and their properties together with the recent progress in instrumentation, it is readily foreseen that at least triplet states and diradicals will be studied more frequently in the future. 4.3 Transition Metals. – The ﬁeld of transition-metal EPR spectroscopy is still the most challenging for theoreticians as the accuracy of DFT methods is not as high as for organic radicals and there are many complicated bonding situations and electronic structure problems to be considered. Nevertheless, there have been numerous studies that demonstrate that essential insights can be achieved. However, this requires precaution in the interpretation of the results which are best viewed in conjunction with experimental data. Again, the ﬁeld has grown to considerable size and a comprehensive survey cannot be claimed. Nevertheless, studies on complexes of vanadium,90,124,196–201 chromium,202 manganese116,36,85,89,120,124 iron,69,81,84,86,91,203–208,93–95,118,125,130,209–213 124,131,214–218 83,118,219–230 cobalt, nickel, copper,80,82,92,101,231–238 molybdenum87,88 118,239–243 as well as some heavier metals and transition-metal atoms and ions37,244 have been reported. More intricate situations, like magnetically coupled dimers82,89,93 and metal-radical assemblies86,91,94,95,118,131,212,215,227,237,239,242–245 have also been considered. These studies originate from very different areas of chemistry ranging from solid-state chemistry over zeolites to coordination and bioinorganic chemistry which demonstrates the usefulness of the computational methods already in this ﬁeld. Based on the recent progress in the computational technologies an even higher popularity of theoretical methods in the analysis of experimental data is anticipated.

5

Concluding Remarks

From the preceding sections it should become evident that the quantum theory of EPR parameters is now a very active ﬁeld with many important recent contributions. DFT methods have almost matured from a technical point of view and their accuracy is relatively well understood. The consistent treatment of scalar relativistic effects is under intense investigation and more progress is to be expected in the near future. Highly correlated ab initio methods, at least in their traditional formulations, generally offer higher accuracy but still suffer from high computational cost. Simpliﬁed approaches are being developed but have not yet reached the same level of standardization as DFT methods. Significant progress in this ﬁeld is expected in the coming years. Already in the present form, quantum chemistry has become a strong partner of experimental investigations and there are many recent studies which report high-resolution EPR spectra in conjunction with DFT calculations. The theoretical results have already often been proven invaluable in making correct assignments. In particular, the many small couplings that are now measurable in larger molecules

88

Electron Paramag. Reson., 2007, 20, 73–95

with modern EPR techniques almost require theoretical calculations for their interpretation since full tensors together with their orientations are difﬁcult to deduce from the experiments without additional information. It is therefore very likely that the interaction between theory and experiment will strongly intensify in the years to come.

Acknowledgements I am grateful to my students, postdocs, co-workers and collaborators listed in the bibliography for their invaluable scientiﬁc input, their creativity and cooperation. I also gratefully acknowledge funding from the German-Israeli Foundation (grant I * 746-137.9/2002), the Deutsche Forschungsgemeinschaft (SPP 1137 ‘Molecular Magnetism’) and the SFB 663 (‘Molecular Response to Electronic Excitation’; University of Du¨sseldorf). References 1. A. Schweiger and G. Jeschke, Principles of Pulse Electron Paramagnetic Resonance, Oxford University Press, Oxford, 2001. 2. M. Kaupp, V. Malkin and M. Bu¨hl, The Quantum Chemical Calculation of NMR and EPR Properties, Wiley-VCH, Heidelberg, 2004. 3. W.B. Mims, The Linear Electric Field Effect in Paramagnetic Resonance, Clarendon Press, Oxford, 1976. 4. P. Strange, Relativistic Quantum Mechanics, Cambridge University Press, Cambridge, 1998. 5. B.A. Hess, Phys. Rev. A, 1985, 32, 756. 6. B.A. Hess, Phys. Rev. A, 1986, 333, 3742. 7. B.A. Hess and C.M. Marian, in Computational Molecular Spectroscopy, ed. P. Jensen and P.R. Bunker, New York, 2000. 8. G. Jansen and B.A. Hess, Phys. Rev. A, 1989, 39, 6016. 9. A. Wolf, M. Reiher and B.A. Hess, J. Chem. Phys., 2002, 117, 9215. 10. S. Faas, J. van Lenthe, A.C. Hennum and J.G. Snijders, J. Chem. Phys., 2000, 113, 4052. 11. W. Klopper, J. van Lenthe and A.C. Hennum, J. Chem. Phys., 2000, 113, 9957. 12. E. van Lenthe, J.G. Snijders and E.J. Baerends, J. Chem. Phys., 1996, 105, 6505. 13. J.H. van Lenthe, S. Faas and J.G. Snijders, Chem. Phys. Lett., 2000, 328, 107. 14. M. Filatov, Chem. Phys. Lett., 2002, 365, 222. 15. L. Visscher and E. van Lenthe, Chem. Phys. Lett., 1999, 306, 357. 16. C. van Wu¨llen, J. Comp. Chem., 2002, 23, 779. 17. H.M. Quiney and P. Belanzoni, Chem. Phys. Lett., 2002, 353, 253. 18. L. Visscher, T. Enevoldsen, T. Saue and J. Oddershede, J. Chem. Phys., 1998, 109, 9677. 19. F. Neese, M. Reiher, T. Fleig and B.A. Hess, J. Chem. Phys., 2005, 122, 204107/1. 20. A. Mateev and N. Ro¨sch, J. Chem. Phys., 2003, 118, 3997. 21. F. Neese and E.I. Solomon, Inorg. Chem., 1998, 37, 6568. 22. R. McWeeny, Methods of Molecular Quantum Mechanics, Academic press, London, 1992. 23. R. McWeeny, Spins in Chemistry, Academic Press, New York, 1970.

Electron Paramag. Reson., 2007, 20, 73–95

89

24. P. Pyykko¨, in The Quantum Chemical Calculation of NMR and EPR Properties, ed. M. Kaupp, M. Bu¨hl, and V. Malkin, Heidelberg, 2004. 25. A. Abragam and B. Bleaney, Electron Paramagnetic Resonance of Transition Ions, Clarendon Press, Oxford, 1970. 26. J.E. Harriman, Theoretical Foundations of Electron Spin Resonance, Academic press, New York, 1978. 27. A.J. Stone, Proc. Roy. Soc. London (A), 1963, 271, 424. 28. F. Neese, Mag. Res. Chem., 2004, 42, S187. 29. G.H. Lushington, P. Bu¨ndgen and F. Grein, Int. J. Quant. Chem., 1995, 55, 377. 30. G.H. Lushington and F. Grein, Theor. Chim. Acta, 1996, 93, 259. 31. G.H. Lushington and F. Grein, Int. J. Quant. Chem.: Quant. Chem. Symp., 1996, 30, 467. 32. G.H. Lushington and F. Grein, J. Chem. Phys., 1997, 106, 3292. 33. P.J. Bruna and F. Grein, Int. J. Quant. Chem, 2000, 77, 324. 34. S. Brownridge, F. Grein, J. Tatchen, M. Kleinschmidt and C.M. Marian, J. Chem. Phys., 2003, 118, 9552. 35. F. Neese, Chem. Phys. Lett., 2003, 380, 721. 36. F. Neese, J. Am. Chem. Soc., 2006, 128, 10213. 37. D. Ganyushin and F. Neese, J. Chem. Phys., 2006, 125, 024103. 38. O. Vahtras, M. Engstro¨m and B. Schimmelpfennig, Chem. Phys. Lett., 2002, 351, 424. 39. O. Vahtras, B. Minaev and H. A˚gren, Chem. Phys. Lett., 1997, 281, 186. 40. M. Engstro¨m, B. Minaev, O. Vahtras and H. A˚gren, Chem. Phys. Lett., 1998, 237, 149. 41. M. Engstro¨m, O. Vahtras and H. A˚gren, Chem. Phys. Lett., 1999, 243, 263. 42. Y. Luo, D. Jonsson, P. Norman, K. Ruud, O. Vahtras, B. Minaev, H. A˚gren, A. Rizzo and K.V. Mikkelsen, Int. J. Quant. Chem., 1998, 70, 219. 43. Y. Yamaguchi, J.D. Goddard, Y. Osamura and H. Schaefer, Analytic Derivative Approaches - A New Dimension to Quantum Chemistry, Oxford University Press, Oxford, 1994. 44. J.A. Pople, R. Krishnan, H.B. Schlegel and J.S. Binkley, Int. J. Quant. Chem.: Quant. Chem. Symp., 1979, 13, 225. 45. S.M. Colwell and N.C. Handy, Chem. Phys. Lett., 1994, 217, 271. 46. A.M. Lee, S.M. Colwell and N.C. Handy, Chem. Phys. Lett., 1994, 229, 225. 47. J. Gerrat and I.M. Mills, J. Chem. Phys, 1968, 49, 1719. 48. J. Autschbach and T. Ziegler, Coord. Chem. Rev., 2003, 238/239, 83. 49. F. Neese, 2006, in preparation. 50. J. Gauss and D. Cremer, Adv. Quant. Chem., 1992, 23, 205. 51. E. van Lenthe, P.E.S. Wormer and A. van der Avoird, J. Chem. Phys., 1997, 107, 2488. 52. G. Schreckenbach and T. Ziegler, Theor. Chem. Acc., 1998, 99, 71. 53. E. van Lenthe, A. van der Avoird and E.S. Wormer, J. Chem. Phys., 1998, 108, 4783. 54. K.M. Neyman, D.I. Ganyushin, A.V. Matveev and V.A. Nasluzov, J. Phys. Chem. A., 2002, 106, 5022. 55. M. Mayer, S. Kru¨ger and N. Ro¨sch, J. Chem. Phys., 2001, 115, 4411. 56. I. Malkin, O. Malkina, V. Malkin and M. Kaupp, J. Chem. Phys., 2005, 123, 244103. 57. S. Patchkovskii, and G. Schreckenbach, in Calculation of NMR and EPR Parameters, ed. M. Kaupp, V. Malkin, and M. Bu¨hl, Weinheim, 2004.

90

Electron Paramag. Reson., 2007, 20, 73–95

58. F. Neese, J. Chem. Phys., 2005, 122, 034107/1. 59. S. Koseki, M.W. Schmidt and M.S. Gordon, J. Chem. Phys., 1992, 96, 10768. 60. S. Koseki, M.S. Gordon, M.W. Schmidt and N. Matsunaga, J. Phys. Chem., 1995, 99, 12764. 61. S. Koseki, M.W. Schmidt and M.S. Gordon, J. Phys. Chem. A., 1998, 102, 10430. 62. S. Koseki, D.G. Fedorov, M.W. Schmidt and M.S. Gordon, J. Phys. Chem. A, 2001, 105, 8262. 63. O.L. Malkina, J. Vaara, B. Schimmelpfenning, M. Munzarova, V. Malkin and M. Kaupp, J. Am. Chem. Soc., 2000, 122, 9206. 64. B.A. Hess, C.M. Marian, U. Wahlgren and O. Gropen, Chem. Phys. Lett., 1996, 251, 365. 65. A. Berning, M. Schweizer, H.J. Werner, P.J. Knowles and P. Palmieri, Molec. Phys., 2000, 98, 1823. 66. S. Patchkovskii, R.T. Strong, C.J. Pickard and S. Un, J. Chem. Phys., 2005, 122, 214101. 67. F. Neese, J. Chem. Phys., 2001, 115, 11080. 68. F. Neese, in ORCA - an ab initio: density functional and semiempirical program package, Version 2.5.0, University of Bonn, Bonn, Germany, 2006, http:// www.thch.uni-bonn.de/tc/orca. 69. F. Neese, Inorg. Chim. Acta, 2002, 337C, 181. 70. F. Neese, J. Chem. Phys., 2003, 3939. 71. Z. Rinkevicius, L. Telyatnyk, P. Salek, O. Vahtras and H. Agren, J. Chem. Phys., 2003, 119, 10489. 72. J. Gauss, M. Kallay and F. Neese, J. Chem. Phys. 2006, to be submitted. 73. A.V. Luzanov, E.N. Babich and V.V. Ivanov, J. Mol. Struc. (Theochem), 1994, 311, 211. 74. M. Kaupp, T. Gress, R. Reviakine, O.L. Malkina and V.G. Malkin, J. Phys. Chem. B, 2003, 107, 331. 75. G. Schreckenbach and T. Ziegler, J. Phys. Chem. A., 1997, 101, 3388. 76. Y. Cao, M.D. Beachy, B. D.A.L. Morrill, M.N. Ringnalda and R.A. Friesner, J. Chem. Phys., 2005, 122, 224116. 77. R. Improta and V. Barone, Chem. Rev., 2004, 104, 1231. 78. M.L. Munzarova and M. Kaupp, J. Phys. Chem. A, 1999, 103, 9966. 79. M.L. Munzarova, P. Kubacek and M. Kaupp, J. Am. Chem. Soc., 2000, 122, 11900. 80. S. Sinnecker and F. Neese, J. Comp. Chem., 2006, 27, 1463. 81. F. Neese, J. Inorg. Biochem. 2006, 716. 82. S. Kababya, N.J.C. Calle, F. Neese and D. Goldfarb, J. Am. Chem. Soc., 2006, 128, 2017. 83. M. van Gastel, C. Fichtner, F. Neese and W. Lubitz, Biochem. Soc. Trans., 2005, 33, 7. 84. S. Sinnecker, L. Slep, E. Bill and F. Neese, Inorg. Chem., 2005, 44, 2245. 85. S. Sinnecker, F. Neese and W. Lubitz, J. Biol. Inorg. Chem., 2005, 10, 231. 86. J. Scho¨neboom, F. Neese and W. Thiel, J. Am. Chem. Soc., 2005, 127, 5840. 87. M. Mader-Cosper, F. Neese, A.V. Astashkin, M.A. Carducci, A.M. Raitsimring and J.H. Enemark, Inorg. Chem., 2005, 44, 1290. 88. A.V. Astashkin, F. Neese, A.M. Raitsimaring, J.J.A. Cooney, E. Bultman and J.H. Enemark, J. Am. Chem. Soc., 2005, 127, 16713. 89. S. Sinnecker, L. Noodleman, F. Neese and W. Lubitz, J. Am. Chem. Soc., 2004, 126, 2613.

Electron Paramag. Reson., 2007, 20, 73–95

91

90. T. Paine, E. Bothe, E. Bill, T. Weyhermu¨ller, L. Slep, F. Neese and P. Chaudhuri, Inorg. Chem., 2004, 43, 7324. 91. R. Garcia Serres, C.A. Grapperhaus, E. Bothe, E. Bill, T. Weyhermu¨ller, F. Neese and K. Wieghardt, J. Am. Chem. Soc., 2004, 126, 5138. 92. D. Baute, D. Arieli, H. Zimmermann, F. Neese, B. Weckhuysen and D. Goldfarb, J. Am. Chem. Soc., 2004, 126, 11733. 93. L.D. Slep, A. Mijovilovich, W. Meyer-Klaucke, T. Weyhermu¨ller, E. Bill, E. Bothe, F. Neese and K. Wieghardt, J. Am. Chem. Soc., 2003, 125, 15554. 94. P. Ghosh, E. Bill, T. Weyhermu¨ller, F. Neese and K. Wieghardt, J. Am. Chem. Soc., 2003, 125, 1293. 95. M. Li, D. Bonnet, E. Bill, F. Neese, T. Weyhermu¨ller, N. Blum, D. Sellmann and K. Wieghardt, Inorg. Chem., 2002, 41, 3444. 96. I. Malkin, O. Malkina, V.G. Malkin and M. Kaupp, Chem. Phys. Lett., 2004, 396, 268. 97. M. Filatov and D. Cremer, J. Chem. Phys., 2004, 121, 5618. 98. K.G. Dyall and E. van Lenthe, J. Chem. Phys., 1999, 111, 1366. 99. M. Filatov and D. Cremer, J. Chem. Phys., 2003, 119, 11526. 100. P.J.M. Geurts, P.C.P. Bouten and A. van der Avoird, J. Chem. Phys., 1980, 73, 1306. 101. F. Neese, J. Phys. Chem. A., 2001, 105, 4290. 102. C. Remenyi, A. Arbuznikov, R. Reviakine, J. Vaara and M. Kaupp, J. Phys. Chem. A, 2004, 108, 5026. 103. A. Arbuznikov, J. Vaara and M. Kaupp, J. Chem. Phys., 2004, 120, 2127. 104. M. Pak and M.S. Gordon, J. Chem. Phys., 2003, 118, 40. 105. S. Chawla and R.P. Messmer, J. Chem. Phys., 1995, 103, 7442. 106. B. Wang, J. Baker and P. Pulay, Phys. Chem. Chem. Phys., 2000, 2, 2131. 107. V.A. Rassolov and D.M. Chipman, in The Quantum Chemical Calculation of NMR and EPR Properties, ed. M. Kaupp, M. Bu¨hl, and V. Malkin, Heidelberg, 2004. 108. O. Vahtras, O. Loboda, B. Minaev, H. Agren and K. Ruud, Chem. Phys., 2002, 279, 133. 109. O. Loboda, B. Minaev, O. Vahtras, B. Schimmelpfennig, H. Agren, K. Ruud and D. Jonsson, Chem. Phys., 2003, 286, 127. 110. O. Rubio-Pons, O. Loboda, B. Minaev, B. Schimmelpfennig, O. Vahtras and H. A˚gren, Mol. Phys., 2003, 101, 2103. 111. B.F. Minaev, E.M. Khomenko, E.A. Bilan and L.B. Yashchuk, Optics and Spectroscopy, 2005, 98, 209. 112. O. Loboda, I. Tunell, B. Minaev and H. Agren, Chem. Phys., 2005, 312, 299. 113. B. Minaev, L. Yaschuk and V. Kukueva, Spectrochimica Acta Part a-Molecular and Biomolecular Spectroscopy, 2005, 61, 1105. 114. T.T. Petrenko, T.L. Petrenko and V.Y. Bratus, J. Phys: Condens. Matter, 2002, 14, 12433. 115. M.R. Pederson and S.N. Khanna, Phys. Rev. B., 1999, 60, 9566. 116. J. Kortus, Phys. Rev. B., 2002, 66, paper 092403. 117. J. Fernandez, J. Kortus, F. Neese, and E. Ruiz, 2006, manuscript in preparation. 118. K. Ray, T. Weyhermu¨ller, F. Neese and K. Wieghardt, Inorg. Chem., 2005, 44, 5345. 119. F. Neese, J. Chem. Phys., 2003, 119, 9428. 120. S. Zein, J.-M. Mouesca, C. Duboc and F. Neese, 2006, manuscript in preparation. 121. R. Reviakine, A.V. Arbuznikov, J.C. Tremblay, C. Remenyi, O.L. Malkina, V.G. Malkin and M. Kaupp, J. Chem. Phys., 2006, 125, 054110.

92

Electron Paramag. Reson., 2007, 20, 73–95

122. P. Schwerdtfeger, M. Pernpointer and W. Nazarewicz, in Calculation of NMR and EPR Parameters, ed. M. Kaupp, M. Bu¨hl, and V. Malkin, Weinheim, 2004. 123. S. Sinnecker, F. Neese and W. Lubitz, J. Am. Chem. Soc., 2004, 126, 3280. 124. D. Baute and D. Goldfarb, J. Phys. Chem. A, 2005, 109, 7865. 125. Y. Zhang, J. Mao and E. Oldﬁeld, J. Am. Chem. Soc., 2002, 124, 7829. 126. N. Godbout, R.H. Havlin, R. Salzmann, P.G. Debrunner and E. Oldﬁeld, J. Phys. Chem. A, 1998, 102, 2342. 127. N. Godbout, L.K. Sanders, R. Salzmann, R.H. Havlin, M. Wojdelski and E. Oldﬁeld, J. Am. Chem. Soc., 1999, 121, 3829. 128. R.H. Havlin, N. Godbout, R. Salzmann, M. Wojdelski, W. Arnold, C.E. Schulz and E. Oldﬁeld, J. Am. Chem. Soc., 1998, 120, 3114. 129. M.T. McMahon, A.C. deDios, N. Godbout, R. Salzmann, D.D. Laws, H. Le, R.H. Havlin and E. Oldﬁeld, J. Am. Chem. Soc., 1998, 120, 4784. 130. T. Lovell, W.G. Han, T. Liu and L. Noodleman, J. Am. Chem. Soc., 2002, 124, 5890. 131. K. Ray, A. Begum, T. Weyhermu¨ller, S. Piligkos, J. van Slageren, F. Neese and K. Wieghardt, J. Am. Chem. Soc., 2005, 127, 4403. 132. E. van Lenthe and E.J. Baerends, J. Chem. Phys., 2000, 112, 8279. 133. V. Kello¨ and A.J. Sadlej, Int. J. Quant. Chem., 1998, 68, 159. 134. V. Kello¨ and A.J. Sadlej, J. Chem. Phys., 2004, 120, 9424. 135. V. Kello¨, A.J. Sadlej and B.A. Hess, J. Chem. Phys., 1996, 105, 1995. 136. I. Malkina, O.L. Malkina and V. Malkin, Chem. Phys. Lett., 2002, 361, 231. 137. S. Patchkovskii and T. Ziegler, J. Chem. Phys., 1999, 111, 5730. 138. V. Barone, in Recent Advances in Density Functional Methods, Part I, ed. D.P. Chong, Singapore, 1996. 139. M. Filatov and D. Cremer, J. Chem. Phys., 2005, 123, 124101. 140. A. Arbuznikov, M. Kaupp, V. Malkin, R. Reviakine and O.L. Malkina, Phys. Chem. Chem. Phys., 2002, 4, 5467. 141. A. Arbuznikov and M. Kaupp, Chem. Phys. Lett., 2004, 391, 16. 142. I. Cioﬁni, in Calculation of NMR and EPR Parameters, ed. M. Kaupp, V. Malkin, and M. Bu¨hl, Weinheim, 2004. 143. I. Cioﬁni, R. Reviakine, A. Arbuznikov and M. Kaupp, Theor. Chem. Acc., 2004, 111, 132. 144. I. Cioﬁni, Magn. Reson. Chem., 2004, 42, S48. 145. I. Cioﬁni, C. Adamo and V. Barone, J. Chem. Phys., 2004, 121, 6710. 146. Z. Rinkevicius, L. Telyatnyk, O. Vahtras and K. Ruud, J. Chem. Phys. 2004, 2004, 121, 5051. 147. S. Sinnecker, A. Rajendran, A. Klamt, M. Diedenhofen and F. Neese, J. Phys. Chem. A, 2006, 110, 2235. 148. S. Moon, S. Patchkovskii and D.R. Salahub, J Mol Struct THEOCHEM, 2003, 632, 287. 149. M. Swart, Ph.D. Thesis, Rijksuniversiteit Groningen, 2002. 150. J.R. Asher, N.L. Doltsinis and M. Kaupp, J. Am. Chem. Soc., 2004, 126, 9854. 151. P. Bu¨ndgen, G.H. Lushington and F. Grein, Int. J. Quant. Chem., 1995, 29, 283. 152. P.J. Bruna, G.H. Lushington and F. Grein, Chem. Phys. Lett., 1996, 258, 427. 153. P.J. Bruna and F. Grein, Chem. Phys. Lett., 2000, 318, 263. 154. F. Grein, Chem. Phys. Lett., 2005, 418, 100. 155. P.J. Bruna and F. Grein, J. Chem. Phys., 1998, 109, 9439. 156. P.J. Bruna and F. Grein, Int. J. Quant. Chem., 2000, 77, 324. 157. P.J. Bruna and F. Grein, Chem. Phys., 2001, 276, 1.

Electron Paramag. Reson., 2007, 20, 73–95

93

158. P.J. Bruna and F. Grein, Int. J. Quant. Chem., 2002, 90, 472. 159. L. Hermodilla, P. Calle, J.M.G. de la Vega and C. Sieiro, J. Phys. Chem. A, 2005, 109, 7626. 160. L. Hermodilla, P. Calle, J.M.G. de la Vega and C. Sieiro, J. Phys. Chem. A, 2005, 109, 1114. 161. M. Kaupp, J.R. Asher, A. Arbuznikov and A. Patrakov, Phys. Chem. Chem. Phys., 2002, 4, 5458. 162. L.B. Knight, J.G. Kaup, B. Petzold, R. Ayyad, T.K. Ghanty and E.R. Davidson, J. Chem. Phys., 1999, 110, 5658. 163. R. Prassad, J. Chem. Phys., 2004, 120, 10089. 164. R. Prassad, J. Chem. Phys., 2003, 119, 9549. 165. M. Mladenovic, M. Peric, M. Jerosimic and B. Engels, Molec. Phys., 2004, 102, 2623. 166. M. Peric, M. Mladenovic and B. Engels, Chem. Phys. Lett., 2004, 393, 552. 167. C. Di Valentin, K.M. Neyman, T. Risse, M. Sterrer, E. Fischbach, H.-J. Freund, V.A. Nasluzov, G. Pacchioni and N. Ro¨sch, J. Chem. Phys., 2006, 124, 044708. 168. E.Y. Misochko, A.V. Akimov, U. Goldschleger, D.A. Tyurin and D.N. Laikov, J. Chem. Phys., 2005, 122, 034503. 169. M. Van Gastel, G. Lassman, W. Lubitz and F. Neese, J. Am. Chem. Soc., 2004, 126, 2237. 170. J. Fritscher, E. Artin, S. Wnuk, G. Bar, J.H. Robblee, S. Kapraczak, M. Kaupp, R.G. Grifﬁn, M. Bennati and J. Stubee, J. Am. Chem. Soc., 2005, 127, 7729. 171. N. Rega, M. Cossi and V. Barone, J. Am. Chem. Soc., 1997, 119, 12962. 172. F. Himo, G.T. Babcock and E.A. Eriksson, Chem. Phys. Lett., 1999, 313, 374. 173. L. Benisvy, R. Bittl, E. Bothe, C.D. Garner, J. McMaster, S. Ross, C. Teutloﬀ and F. Neese, Angew. Chem. Int. Ed., 2005, 44, 5314. 174. F. Himo, E.A. Eriksson, M.R.A. Blomberg and P.E. Siegbahn, Int. J. Quant. Chem., 2000, 76, 714. 175. G. Lassman, E.A. Eriksson, F. Lendzian and W. Lubitz, J. Phys. Chem. A, 2000, 104, 9144. 176. D. Biglino, P.P. Schmidt, E.J. Reijerse and W. Lubitz, Phys. Chem. Chem. Phys., 2006, 8, 58. 177. C. Adamo, M. Heitzmann, F. Meilleur, N. Rega, G. Scalmani, A. Grand, J. Cadet and V. Barone, J. Am. Chem. Soc., 2001, 123, 7113. 178. F. Jolibois, J. Cadet, A. Grand, R. Subra, N. Rega and V. Barone, J. Am. Chem. Soc., 1998, 120, 1864. 179. S.M. Mattar, Chem. Phys. Lett., 2005, 405, 382. 180. S.M. Mattar, J. Sanford and A.D. Goodfellow, Chem. Phys. Lett., 2006, 418, 30. 181. M. Beyer, J. Fritscher, E. Feresin and O. Schiemann, J. Org. Chem., 2003, 68, 2209. 182. J. Cirujeda, J. Vidal-Gancedo, O. Ju¨rgens, F. Mota, J.J. Novoa, C. Rovira and J. Veciana, J. Am. Chem. Soc., 2000, 122, 11393. 183. M. Pavone, C. Benzi, F. De Angelis and V. Barone, Chem. Phys. Lett., 2004, 395, 120. 184. P. Cimino, M. Pavone and V. Barone, Chem. Phys. Lett., 2006, 419, 106. 185. J.R. Asher, N.L. Doltsinis and M. Kaupp, Mag. Res. Chem., 2005, 43, S237. 186. M. Kaupp, Biochemistry, 2002, 41, 2895. 187. M. Kaupp, C. Remenyi, J. Vaara, O.L. Malkina and I. Malkin, J. Am. Chem. Soc., 2002, 124, 2709. 188. S. Kapraczak, M. Kaupp and F. McMillan, J. Am. Chem. Soc., 2006, 128, 5659. 189. S.M. Mattar, A.H. Emwas and A.D. Stephens, Chem. Phys. Lett., 2002, 363, 152.

94

Electron Paramag. Reson., 2007, 20, 73–95

190. S.M. Mattar, J. Phys. Chem. B, 2004, 108, 9449. 191. M.R. Fuchs, E. Schleicher, A. Schnegg, C.W.M. Kay, J. To¨rring, R. Bittl, A. Bacher, G. Richter, K. Mo¨bius and S. Weber, J. Phys. Chem. B, 2002, 106, 8885. 192. M. Straka, M. Kaupp and E. Roduner, Theor. Chem. Acc., 2005, 114, 318. 193. R. Takeda, S. Mitsuo, S. Yamanaka and K. Yamaguchi, Polyhedron, 2005, 24, 2238. 194. S. Sinnecker and F. Neese, J. Phys. Chem. A, 2006, in press. 195. M. Shoji, K. Koizumi, T. Hamamoto, T. Taniguchi, R. Takeda, Y. Kitagawa, T. Kawakami, M. Okumura, S. Yamanaka and K. Yamaguchi, Polyhedron, 2005, 24, 2708. 196. A.C. Saladino and S.C. Larsen, J. Phys. Chem. A, 2002, 106, 10444. 197. P.J. Carl, S.L. Isley and S.C. Larsen, J. Phys. Chem. A, 2001, 105, 4563. 198. S.C. Larsen, J. Phys. Chem. A, 2001, 105, 8333. 199. M.L. Munzarova´ and M. Kaupp, J. Phys. Chem. B, 2001, 105, 12644. 200. C. Remenyi, M.L. Munzarova and M. Kaupp, J. Phys. Chem. B, 2005, 109, 4227. 201. E. Bernhardt, H. Willner, A. Kornath, J. Breidung, M. Bu¨hl, V. Jonas and W. Thiel, J. Phys. Chem. A, 2003, 107, 859. 202. R. Kapre, K. Ray, I. Sylvestre, T. Weyhermu¨ller, S. DeBeer-George, F. Neese and K. Wieghardt, Inorg. Chem., 2006, 45, 3499. 203. E. van Lenthe, A. van der Avoird, W.R. Hagen and E.J. Reijerse, J. Phys. Chem. A, 2000, 104, 2070. 204. V.V. Vramasu, E.L. Bominaar, J. Meyer and E. Mu¨nck, Inorg. Chem., 2002, 41, 6358. 205. T.A. Jackson, J. Xie, E. Yikilmaz, A.-F. Miller and T.C. Brunold, J. Am. Chem. Soc., 2002, 124, 10833. 206. M.I. Davis, A.M. Orville, F. Neese, J.M. Zaleski, J.D. Lipscomb and E.I. Solomon, J. Am. Chem. Soc., 2002, 124, 602. 207. H. Andres, E.L. Bominaar, J.M. Smith, N.A. Eckert, P.L. Holland and E. Munck, J. Am. Chem. Soc., 2002, 124, 3012. 208. T.A. Jackson, E. Yikilmaz, A. Francis-Miller and T.C. Brunold, J. Am. Chem. Soc., 2003, 125, 8348. 209. S. Y. Wu, W. H. Wei and H.N. Dong, Zeitschrift Fur Naturforschung Section a-a Journal of Physical Sciences, 2004, 59, 769. 210. A. Fiedler and T.C. Brunold, Inorg. Chem., 2005, 44, 9322. 211. J.F. Berry, E. Bill, F. Neese, R. Garcia-Serres, T. Weyhermu¨ller and K. Wieghardt, Inorg. Chem., 2006, 45, 2027. 212. V.K.K. Praneeth, F. Neese and N. Lehnert, Inorg. Chem., 2005, 44, 2570. 213. N. Aliaga-Alcade, S. DeBeer George, E. Bill, K. Wieghardt and F. Neese, Angew. Chem. Int. Ed., 2005, 44, 2908. 214. M. Atanasov, E.J. Baerends, P. Baettig, R. Bruyndonckx, C. Daul, C. Rauzy and M. Zbiri, Chem. Phys. Lett., 2004, 399, 433. 215. E. Bill, E. Bothe, P. Chaudhuri, K. Chlopek, D. Herebian, S. Kokatam, K. Ray, T. Weyhermu¨ller, F. Neese and K. Wieghardt, Chem. Eur. J., 2005, 11, 204. 216. T.A. Stich, N.R. Buan, J.C. Escalante-Semerena and T.C. Brunold, J. Am. Chem. Soc., 2005, 127, 8710. 217. T.A. Stich, M. Yamanishi, R. Banerjee and T.C. Brunold, J. Am. Chem. Soc., 2005, 127, 7660. 218. T.A. Stich, J. Seravalli, S. Venkateshrao, T.G. Spiro, S.W. Ragsdale and T.C. Brunold, J. Am. Chem. Soc., 2006, 128, 5010.

Electron Paramag. Reson., 2007, 20, 73–95

95

219. R.G. Hayes, Inorg. Chem., 2000, 39, 156. 220. M. Stein and W. Lubitz, Curr. Op. Chem. Biol., 2002, 6, 243. 221. M. Stein, E. van Lenthe, E.J. Baerends and W. Lubitz, J. Phys. Chem. A, 2001, 105, 416. 222. M. Stein, E. van Lenthe, E.J. Baerends and W. Lubitz, J. Am. Chem. Soc., 2001, 123, 5839. 223. M. Stein and W. Lubitz, J. Inorg. Biochem., 2001, 86, 442. 224. W. Lubitz, M. Stein, M. Brecht, O. Trofanchuk, S. Foerster, Y. Higuchi, E. van Lenthe and F. Lendzian, Biophys. J., 2000, 78, 1660. 225. J.L. Craft, B.S. Mandinutsira, K. Fujita, C.G. Riordan and T.C. Brunold, Inorg. Chem., 2003, 42, 859. 226. Z.Y. Yang, Y. Hao, C. Rudowicz and Y.Y. Yeung, J. Phys. Cond. Mat., 2004, 16, 3481. 227. S. Blanchard, F. Neese, E. Bothe, E. Bill, T. Weyhermu¨ller and K. Wieghardt, Inorg. Chem., 2005, 44, 3636. 228. R. Schenker, M.T. Mock, M.T. Kieber-Emmons, C.G. Riordan and T.C. Brunold, Inorg. Chem., 2005, 44, 3605. 229. A.T. Fiedler, P.A. Bryngelson, M.J. Maroney and T.C. Brunold, J. Am. Chem. Soc., 2005, 127, 5449. 230. A. Goenka, M. van Gastel, W. Ga¨rtner and W. Lubitz, J. Phys. Chem. B, 2006, 110, 8142. 231. P.J. Carl, S.L. Baccam and S.C. Larsen, J. Phys. Chem. B, 2000, 104, 8848. 232. A.R. Jaszewski and J. Jezierska, Chem. Phys. Lett., 2001, 343, 571. 233. K. Pierloot, A. Delabie, M.H. Groothaert and R.A. Schoonheydt, Phys. Chem. Chem. Phys., 2001, 3, 2174. 234. R.J. Deeth, J. Chem. Soc. Dalton Trans., 2001, 664. 235. A. Delabie, K. Pierloot, M.H. Groothaert, B.M. Weckhuysen and R.A. Schoonheydt, Phys. Chem. Chem. Phys., 2002, 4, 134. 236. A.C. Saladino and S.C. Larsen, J. Phys. Chem. A, 2003, 107, 5583. 237. M. Davidova, D. Nachtigallova, P. Nachtigall and J. Sauer, J. Phys. Chem. B, 2004, 108, 13674. 238. R. Arratia-Perez, L. Alvarez-Thon and P. Fuentealba, Chem. Phys. Lett., 2004, 397, 408. 239. X. Sun, H. Chun, K. Hildenbrand, E. Bothe, T. Weyhermu¨ller, F. Neese and K. Wieghardt, Inorg. Chem., 2002, 41, 4295. 240. R. Arratia-Perez and L. Hernandez-Acevedo, J. Chem. Phys., 2003, 118, 7425. 241. L. Alvarez-Thon, L. Hernandez-Acevedo and R. Arratia-Perez, J. Chem. Phys., 2001, 115, 726. 242. S. Frantz, H. Hartmann, H. Doslik, M. Wanner, W. Kaim, H.-J. Ku¨mmerer, G. Denninger, A.-L. Barra, C. Duboc-Toia, J. Fiedler, C. Urba and M. Kaupp, J. Am. Chem. Soc., 2002, 124, 10563. 243. M. Wanner, T. Scheiring, W. Kaim, L.D. Slep, L.M. Baraldo, J.A. Olabe, S. Zalis and E.J. Baerends, Inorg. Chem., 2001, 40, 5704. 244. F. Neese, T.T. Petrenko, D. Ganyushin and G. Olbrich, Coord. Chem. Rev. 2006, in press. 245. C. Remenyi and M. Kaupp, J. Am. Chem. Soc., 2005, 127, 11399.

Getting an Inside View of Nanomaterials with Spin Labels and Spin Probes BY VICTOR CHECHIK1 AND AGNETA CARAGHEORGHEOPOL2 1 Department of Chemistry, University of York, Heslington, York, Y010 5DD, UK 2 ‘Ilie Murgulescu’ Institute of Physical Chemistry, Romanian Academy, Spl. Independentei 202, 060021 Bucharest, Romania

1

Introduction

The world as we know it depends heavily on materials chemistry. From the dawn of civilisation, synthetic materials have been crucial for all industrial development. Until recently, materials research was mainly focussed on the bulk properties of materials; this proved very successful and many new materials with superior properties were discovered. Nowadays, however, the demands and expectations in this area have grown significantly, and modern industry requires ﬁne-tuning of multifunctional materials properties to a strict speciﬁcation. Fortunately, our ability to obtain information about material properties on a molecular scale has significantly improved over the years. The new analytical tools include modern transmission and scanning electron microscopy, surface probe microscopy (e.g., AFM and STM), readily available spectroscopic tools such as XPS, mass-spectrometry, solid state NMR spectroscopy. The development of analytical methods thus led to the explosion of interest in designing and manipulating molecular structure of materials, which is often called nanoscience and/or nanotechnology. More and more elaborate, complex and multifunctional materials are being designed and studied. It became possible to rationally develop new materials, in which bulk properties result from subtle but controlled changes on nanometre scale. The examples of the new generation of smart materials include selfassembling surfactant aggregates (e.g., amphiphilic block-copolymers); surfactant-templated mesoporous materials with narrow distribution of pore sizes and well-organised pore distribution; organically-functionalised porous materials for applications in catalysis; integrated organic-inorganic hybrid compounds; supramolecular assemblies built from pre-formed complementary building blocks, such as polymer-clay nanocomposites; inorganic nanoparticles; Electron Paramagnetic Resonance, Volume 20 r The Royal Society of Chemistry, 2007 96

Electron Paramag. Reson., 2007, 20, 96–130

97

functionalised surfaces etc. In most of these cases the structural units have nanometre-scale dimensions. Despite the development of new analytical tools, it still remains challenging to obtain information about the molecular environment in these complex materials, and development of new methods is essential for future development of nanoscience. We believe that modern EPR spectroscopy can contribute significantly to this area. After the pioneering work on organic free radicals, paramagnetic complexes and spin-labelling, EPR spectroscopy was experiencing a period of stagnation in the 80’s. However, in the last decade, dramatic improvements in EPR instrumentation (in particularly, development of very high ﬁeld spectroscopy and pulsed methods) mean that completely new experiments can be carried out and new types of information can be obtained. Coincidentally, the parallel development in molecular biology (site-directed mutagenesis) made it possible to specifically label proteins with paramagnetic groups in any given position. This has led to an explosion of interest in using advanced EPR methods for structural studies in molecular biology. The information which can be obtained using modern spin-labelling methods includes: Measurements of distance between several spin labels. Short distances (e.g., continuous wave (CW) up to 2–2.5 nm) can be probed using dipolar broadening in spectra; however much longer distances (up to 7–8 nm) are now accessible through pulsed double electron-electron resonance (usually abbreviated as DEER or PELDOR). Spin-label dynamics can be deciphered from the line shape of CW spectra, as the magnetic tensors of the nitroxide do not completely average out for slowly tumbling species. The development of very high ﬁeld methods makes it possible to analyse dynamics in multifrequency experiments which enables deconvolution of complex motions (e.g., motion of the spin-labelled species as a whole vs. local motion). Excellent g resolution at very high ﬁeld makes it possible to obtain separate EPR spectra for co-existing species which cannot be distinguished at commonly used lower ﬁelds (e.g., X-band). Accessibility of the immobilised spin labels to solution-based reagents can be probed using paramagnetic quenching reagents, such as dioxygen or paramagnetic metal complexes. Polarity and acidity of the immediate microenvironment around the spin label can be probed, particularly using high ﬁeld EPR and pH-sensitive nitroxides. The microenvironment can also be characterised through short range electron-nuclear interactions using pulsed techniques such as ENDOR and ESEEM. As mentioned above, most of these methods were developed to study biological substrates, and this proved really successful. There is however certain similarity

98

Electron Paramag. Reson., 2007, 20, 96–130

between problems in molecular biology and material science. Therefore, the approaches (both theoretical and experimental) used for probing biological systems can be translated into the materials area. In fact, all types of EPRaddressable information mentioned above are much needed by materials chemists and are very difﬁcult to obtain using other techniques. It is hence surprising that, until recently, EPR spectroscopists concentrated their efforts so heavily on the biological systems, and somewhat neglected other areas. In the last few years, scientists have started applying advanced spin-labelling methods to materials chemistry, but there is still much scope for more research. This review surveys recent applications of EPR and spin labels/probes to different areas of material characterisation, in particular those which aimed to understand material properties on a molecular level. The review is not intended to be comprehensive; instead we have tried to select representative studies which illustrate the potential of EPR in materials science, the range of different approaches which can be applied to this area, and how these methods can be used to obtain important information about the materials structure and properties. In several areas (e.g., polymer chemistry, colloidal chemistry), EPR has been successfully used to probe molecular environment and structure for decades, e.g., long before the word ‘‘nano’’ has become ubiquitous. In these cases, we have attempted to emphasise how the new methods of advanced EPR complement the old techniques and thus significantly expand the scope of EPR applications. In the next sections, we give speciﬁc examples categorised by the type of material studied. The basic principles of the EPR techniques used in these studies are introduced brieﬂy as they have been reviewed many times elsewhere.

2

Self-Assembled Supramolecular Structures

The application of spin labelling to colloidal systems (e.g., micelles, vesicles, liquid crystals) has been exploited by a number of research groups in the 80’s and 90’s and the results of this work have been recently reviewed.1 Colloidal systems can rightfully claim their place in the nanoscience domain; however they do not have the same control over the composition and functional group localisation which has become essential in modern nanochemistry. More precise and rational design is possible with supramolecular assemblies (aggregates of speciﬁc molecular units held together by non-covalent bonds) which represent one of Nature’s most important building and functioning principles. Noncovalent interactions are involved in recognition processes responsible for the basic processes in the functioning of living organisms. On the basis of these principles, synthetic supramolecular systems are being developed which can potentially lead to molecular scale photonic and opto-electronic devices. Host-guest compounds of cyclodextrins (CDs) were ﬁrst studied by EPR using cyclic nitroxide spin probes. However, the changes of EPR parameters upon complexation are small and the distinction between the spectra of unbound and bound species is hard to quantify. Lucarini and co-workers

99

Electron Paramag. Reson., 2007, 20, 96–130

developed a series of benzyl nitroxides which show much more significant spectral changes following interaction with supramolecular hosts such as cyclodextrins and calixarenes (Figure 1).2 The enhanced sensitivity of these molecules to supramolecular interactions is due to conformational changes upon complexation which significantly modify the hyperﬁne splittings from the b-protons (as the hyperﬁne interaction depends on the dihedral angle between the C–H bond and the SOMO). Simulation of the spectra of the unbound radical and its complex with the CD made it possible to determine binding constants. Unfortunately, benzyl nitroxides are significantly less stable than the commonly used TEMPO, PROXYL or DOXYL derivatives which precludes their wide-spread use. The authors also recognised the potential of EPR to study the kinetics of association/dissociation of such host-guest complexes, due to the frequency range (105–109 Hz) of the processes involved, which is speciﬁc for EPR spectroscopy. Thus, at higher temperatures some lines broadened due to more rapid exchange of the radicals between the complexed and the free state. In these cases, the exchange rates were determined from the spectral simulations including the two-site jump model for each temperature point. This made it possible to determine the activation parameters for the binding. The authors rationalised the results in terms of CD type (size, substituents) and radical structure. The association/dissociation rates with CDs were compared with those for different types of micelles. The reverse approach – spin labelling the host rather than the guest molecules – has also been explored. A series of spin-labelled cyclodextrin derivatives has been prepared by Ionita and Chechik.3 EPR data led the authors to suggest formation of self-inclusion complexes by these compounds. Although EPR parameters were not sensitive to binding of small guest molecules, formation of supramolecular complexes with large structures (e.g., dendrimer derivatives) can be readily studied by EPR. Supramolecular interactions can also be probed using the sensitivity of EPR to interspin distances. Tsvetkov et al.4,5 used this approach to characterise secondary structure of doubly labelled peptides (analogues of the antibiotic trichogin) and to gain information about their aggregation in solvents of different polarity. Two methods were used to measure the interspin distances: analysis of dipolar broadening in CW EPR spectra and pulsed electron-electron double resonance (PELDOR).

OH

OH N

HO

O

OH

Figure 1 Binding of a benzyl nitroxide by b-cyclodextrin

100

Electron Paramag. Reson., 2007, 20, 96–130

At room temperature, the authors succeeded in separating inter- and intramolecular spin-spin interactions by comparing line-widths of CW spectra in samples of mono- and doubly-labelled peptides at different concentrations in organic solvents as well as their mixtures with unlabelled peptides. The exchange interaction can be neglected if the spin-labelled species is rigid and no collisions between adjacent spin labels can take place due to steric factors. The contribution of direct exchange at the interspin distances used in this work can also be neglected. Therefore, the observed line broadening was due to the decrease in the relaxation time T2 caused by dipole-dipole interaction. The cross-relaxation time T2 for a pair of spin-labels at a ﬁxed distance r is related to the rotational correlation time (eqn 1). 1/T2 ¼ 9g4i2t/20r6

(1)

This made it possible to evaluate the interspin distances from room temperature EPR spectra of peptide solutions. The distances in doubly-labelled peptides estimated by this method, were consistent with the a-helical conformation of the peptides. This study illustrates how a relatively simple CW approach can be used to provide important information about the conformations of the spinlabelled supramolecular structures. Analysis of dynamics of the peptide solutions revealed formation of aggregates at higher concentration. The aggregate dimensions could be roughly evaluated from the analysis of the tc values. However, detailed information about the aggregate structure required measurement of distances between the monomer units. The interspin distances in peptide aggregates cannot be accurately determined by the CW methods. In this case, application of pulsed EPR methods becomes essential. The PELDOR (or DEER) technique (which is commonly used to determine interspin distances following pioneering studies of Tsvetkov et al. in 1980’s) is based on the electron spin echo experiment. The echo pulse sequence is modiﬁed with an additional pumping pulse that modulates the dipole-dipole coupling of the spins. The intensity of the echo signal is then measured as a function of time delay between the ﬁrst pulse and the additional pulse. In this way, the dipolar coupling can be readily separated from inhomogeneous broadening of the EPR spectra. It is this inhomogeneous broadening which hinders the measurement of relatively weak dipolar couplings (e.g., for large interspin distances) in CW spectra. PELDOR data provide direct distance information for systems in which all interspin distances within the EPR range (at least 1.5–8 nm) have the same value. Very often, however, spin-labelled supramolecular systems show a distribution of interspin distances. These include the distances between spin labels in the same assembly, as well as distances between spin labels in adjacent structures. In such cases accurate information about average distances and distance distributions can be obtained. For instance, for a system with a nonhomogeneous spin distribution, i.e. consisting of aggregates, with the inter-aggregate distance larger than the average distance within aggregates it is convenient to represent the decay of the PELDOR signal as a product of

Electron Paramag. Reson., 2007, 20, 96–130

101

inter- and intra-molecular contributions (eqn 2). V(t) ¼ V(t)inter V(t)intra

(2)

The analytical shape of the intermolecular contribution is well known (e.g., for homogeneous 3D solutions V(t)inter ¼ eat). In many practical situations, the generic shape of intramolecular distance distribution is known from models or other experimental data. In this case, the DEER decay can be ﬁtted with the theoretical function derived for the model distribution. The deconvolution of the overall decay curves becomes trivial and can provide information not only about the interspin distances, but also about the number of interacting spins (from V(t)intra) and the concentration of supramolecular aggregates in solution (from V(t)inter), see Figure 2. This makes PELDOR (or DEER) a very attractive method to study spinlabelled supramolecular aggregates. The information about the number of interacting spins and the distances between them is often sufﬁcient to accurately re-create the structure of the aggregate. For instance, in a very elegant recent study, Goldfarb and co-workers applied the DEER technique to determine the size of micelles and the aggregation numbers by exploring the distance distributions in micellar solution.6 The method relies on the selection of an appropriate hydrophobic probe which

Figure 2 Decay of PELDOR signal (on logarithmic scale) for different concentrations of doubly labelled peptide (dotted lines). The solid line shows the background signal V(t)inter due to spin-spin interactions between spin labels in different peptide molecules (Reprinted with permission from Raap et al.4 Copyright (2001) American Chemical Society)

102

Electron Paramag. Reson., 2007, 20, 96–130

would only dissolve in the micelle core. The Goldfarb group used 4-hydroxyTEMPO benzoate and a series of polyethylene/polypropylene block-copolymer surfactants (Pluronics). The experimental DEER decays were then ﬁtted with the theoretical function developed for a system with a non-homogeneous spin distribution, i.e. consisting of aggregates (eqn 2). This procedure yielded the average number of radicals per micelle and rmax which represents the maximum distance between spins in the same micelle, i.e. the diameter of a spherical micelle, or the length of a cylinder. From these data, the aggregation number can be deduced.

3

Polymers

Classical CW EPR studies of spin probes and spin labels have long been successfully used to describe the dynamics of polymer chains. For instance, Pilar et al.7,8 in a series of publications reported data on local segmental dynamics of polymers and co-polymers which were spin labelled with nitroxidefunctionalised methacrylic acid. Rotational diffusion of the nitroxide attached to chain segments of the polymer in dilute solutions, was approximated as a superposition of the isotropic rotational diffusion of the polymer chain segment (characterised by rotational diffusion coefﬁcient RS), and the internal rotation of the spin label about the tether RI. For short tethers, the axis of internal rotation should be identical to the bond axis. This approximate model gives rise to an axial rotational diffusion tensor with two components Rprp ¼ RS and Rpll ¼ RS þ RI. The experimental spectra for different concentrations of the polymer in methanol at different temperatures were ﬁtted using the MOMD model. This analysis yielded the values of diffusion tensors. The shape of the orienting potential described the distribution of effective orientations of the nitroxides’ rotational tensor. The construction of the Arrhenius plots of the rotational parameters made it possible in some cases to estimate the barrier for local conformational transitions and comment on the conformational changes. Dynamics studies were also applied to Nafion membranes by Schlick et al. In order to make spin probes compatible with the polymers, they were functionalised with perﬂuorinated chains.9 The exact location of the spin probes varied as a function of the length of ﬂuorinated side chain. The authors found evidence for the presence of multiple sites with different dynamics. This was tentatively explained by the existence of a range of amorphous phases which are affected by the proximity of crystalline domains. Other contributions describe dynamic heterogeneity in the interfacial region of microphase-separated block-copolymers and analysis of high mobility chainends in polymers.10 Recently, it became possible to obtain direct information about the organisation of the polymer molecules using pulsed techniques (e.g., DEER) or advanced mathematical treatment of CW data.

Electron Paramag. Reson., 2007, 20, 96–130

103

Jeschke et al.11,12 used spin probes to study the organisation of counter-ions in dilute polyelectrolyte solutions. The electrostatic interactions in these systems can bring about changes in polymer conformation as a function of solvent polarity and relative concentrations of the polymer and the counter-ions. Introduction of spin probes was achieved simply and elegantly by mixing polymer solutions with multiply charged paramagnetic species. These probes were found to bind electrostatically to the polymer. The microscopic structure of the system was analysed using CW and FT EPR methods. The rotational correlation times obtained from analysis of EPR data showed that the rotational mobility of the probes is significantly hindered in the presence of the polyelectrolyte, but not to the extent of tight binding. This proved that the interaction represents a dynamic attached-detached equilibrium which is fast on the EPR time scale. Furthermore, the spectral component corresponding to the free counter-ion spin probe was not observed in the presence of polyelectrolyte. The rotational diffusion in the mixtures of doubly charged probe Fremy’s salt (FS) with poly(diallylmethylammonium) chloride (PDADMAC) could only be described by a highly anisotropic axial tensor with the unique axis along one N–S bond of FS assigned to a speciﬁc ion-pair. This led the authors to suggest that at the short time scale of the radical pair existence, the binding of the counter-ion to the polymer preferentially proceeds via only one sulfonate moiety (Figure 3). The spectral changes observed in the presence of polyelectolyte could be separated into the contributions of the slowing rotational dynamics and enhanced spin exchange contribution. However, the spectra could not be ﬁtted with a single value for the Heisenberg exchange frequency.

Figure 3 Binding of Fremy’s salt to the cationic groups of polyelectrolyte PDADMAC (Reprinted with permission from Jeschke et al.12 Copyright (2004) American Chemical Society)

104

Electron Paramag. Reson., 2007, 20, 96–130

The unusual line shapes of the CW spectra were analysed taking into consideration a broad distribution of line widths. Initial approximate ﬁts (obtained with Freed’s NLSL programme13) for the relative intensities of the three lines of the nitroxide’s spectrum (assuming the intrinsic line width observed in the absence of polyelectrolyte) were Fourier-transformed to give the time domain data Vsim(t), and convolution with an empirical (stretched exponential) function was performed to account for the concentration broadening (eqn 3). V(t) ¼ Vsim(t) exp(kt)x

(3)

Here, k is a characteristic decay time constant related to the average concentration and x is the stretch factor characterising the width of the distribution of line widths. The additional broadening stems from inhomogeneous distribution of counter-ions (e.g., decrease of local concentrations with the increased distance from the polyelectrolyte chain). This approach made it possible to ﬁt the experimental spectra with good precision, and hence extract reliable information about the dynamics in the system. Additional information about the polymer structure was obtained from DEER and ESEEM experiments. DEER measurements provided the information about distribution of distances between adjacent counter-ions and hence made it possible to explore the conformation of the polymer chain. Interestingly, the precise location of the counter-ion can be probed using electron spin echo envelope modulation (ESEEM). This pulsed EPR technique is basically complementary to ENDOR and is sensitive to weak, short range interactions between the unpaired electron and surrounding nuclei. In ESEEM of FS-PDADMAC mixtures, the modulation observed was assigned to the quaternary 14N of the PDADMAC repeat unit. The electron nuclear distance of closest approach (e.g., the distance between the unpaired electron of the counter-ion and the nitrogen atom of polyelectrolyte) was 0.43 nm, as expected for an ion-pair. The fraction of such contact pairs was found to be close to 20% irrespective of polyelectrolyte concentration. The other spin-probe ions bound territorially by non-speciﬁc electrostatic interaction. Four-pulse DEER data showed that for high polyelectrolyte concentration, the counter-ions are distributed homogeneously in 3D space, while for low polyelectrolyte concentrations they appear to be linearly distributed along the locally extended chain. Another contribution from the Jeschke group14,15 explored the ionic clusters in ionomers (macro-zwitterions, e.g., diblock-copolymers functionalised with oppositely charged groups, Figure 4). The ion cluster sizes and inter-cluster distances were determined as described in section 2 (eqn 2) using a four-pulse DEER experiment. It was found that the time-domain DEER signal consists of three components that can be modelled by two Gaussian distance distributions with mean values of about 1.5–2 and 6–7 nm and a uniformly distributed background. The Gaussian component at about 2 nm was assigned as a measure of the cluster size while the Gaussian component at 6–7 nm corresponded to inter-cluster distances. These numbers were in good agreement with

Electron Paramag. Reson., 2007, 20, 96–130

105

Figure 4 Schematic illustration of ionomers and spin probes used to study clustering (Reprinted with permission from Jeschke et al.14 with minor changes. Copyright (2000) American Chemical Society)

molecular modelling and results of independent X-ray scattering (SAXS). However, DEER can be applied to study ionic clustering in any ionomers including ionically end-capped polymer for which SAXS experiment failed. Apart from the structural studies of bulk polymers and polymer solution, EPR can be successfully applied to probe nanostructured polymeric materials. For instance, spin-labelling recently provided much information about the structure and dynamics of the surfactant layer in polymer-clay nanocomposites.16 Here, special cationic spin-probes were required which had structure similar to the cationic surfactants (Figure 5). Room and high temperature CW EPR data reported on the binding of the surfactant heads to the inorganic phase through their dynamic properties. This was studied in media of different polarities for polymer-clay nanocomposites prepared by different approaches: in situ polymerisation, intercalation from melt, intercalation from solution. High frequency ENDOR revealed hyperﬁne interaction with 19F nuclei from the clay (expandable ﬂuoromica) thus providing evidence of the proximity of the spin labels to the clay surface. For dispersions of the organoclay in deuteriated toluene, the ESEEM data made it possible to assess the penetration of solvent molecules into the surfactant layer down to the region of anchor groups. The spin probe technique was also applied to colloidal polymer dispersions (latices).17 The authors succeeded in characterising the different components of this micro-heterogeneous system. Small hydrophobic spin probes were found to reach the particle interior using the polymer-free volume, whereas amphiphilic radicals were included in the ordered surfactant layers adsorbed on the surface of latex particles. Small polar nitroxides remained in the serum. In the dried dispersions, the mobility of small spherical spin probes correlated with the Tg of the polymer. While TEMPO probed the polymer free volume, without any speciﬁc interaction, TEMPOL had a more hindered rotation. Surfactant-type

106

Electron Paramag. Reson., 2007, 20, 96–130

Figure 5 Schematic illustration of a surfactant layer on the clay particle (Reprinted with permission from Jeschke et al.16)

spin probes provided evidence for surfactant aggregation in the dried dispersions. These results represent a basis for studies of ﬁlm formation, ﬁlm annealing and rewetting. Size and shape selectivity of molecular aggregates is important for selfassembly of materials by supramolecular interactions. Polymeric materials with well-deﬁned porous structures can serve this purpose and are therefore useful in applications such as chemical catalysis, selective binding of drugs, proteins and other biological molecules. Jeschke et al.18 applied W-band EPR spectroscopy of various spin probes to study porous, highly cross-linked resins based on styrene-divinylbenzene polymers. Pores with speciﬁc sizes were created by template imprinting with AOT reverse micelles, while pore functionalisation was achieved by introducing polymerisable co-surfactants. Various spin probes were used to obtain information about the accessibility of pores to molecules with different sizes, charges, polar characteristics as well as on the preferred locations of these probes. W-band EPR (95 GHz) was chosen, as it increases spectral resolution and advantageously modiﬁes the time window for dynamics. Thus, tumbling which is in the motional narrowing regime in X-band appear in W-band in the slow motional regime, a region where line-shapes are more sensitive to motional parameters, and their extraction by spectra simulations is much more reliable. The dynamics of an amphiphilic probe (7-doxyl stearic acid) changed significantly in the presence of functionalised porous polymers. It was found that when this probe entered the

Electron Paramag. Reson., 2007, 20, 96–130

107

pores, it occupied the place of the washed-out template surfactant, an effect named ‘‘structural memory’’ by the authors. It might be interesting to mention that Caldararu and co-workers observed a similar behaviour by adsorbing spin probes from a surfactant solution in the empty pores of dried mesoporous silica. The surfactants and the spin probes replicated the template aggregates in the pores.29

4

Mesoporous Materials

Mesoporous materials are usually prepared by hydrolysis and condensation of soluble precursors (e.g., tetraalkoxysilanes) around a colloidal (e.g., surfactant) template. This route (introduced by Mobil Oil researchers in 1992) leads to materials with ordered cavities or channels with tuneable diameter in the 2–10 nm range.19 This range of pore diameters can be further extended (up to 30 nm) by using block-copolymers as templates. For instance, mesoporous silica MCM-41 is formed hydrothermally in the presence of a cationic surfactant (cetyltrimethylammonium bromide (CTAB) or hydroxide) with tetraethoxysilane (TEOS) as the silica source. While it is certain that the surfactant plays a vital role in the formation of MCM-41, the precise mechanism of the process is hard to unravel. Two mechanistic pathways have been suggested: (A) the liquid-crystal phase may form prior to the addition of the silica source, or (B) the silicate species generated in the reaction mixture may trigger formation of the liquid-crystal phase. Literature data report formation of hexagonal, cubic and lamellar structures by varying the silica concentration at constant surfactant concentration, thus implying the important role of the silicate anions. It is clear that each pathway may be active, depending on the speciﬁc reaction conditions. EPR methodology based on line-shape analysis has been extensively used in previous studies of surfactant aggregation for the detection of unimer-micelle transitions as well as for evaluation of polarity proﬁles, dynamics and order of surfactant chains in micelles.1 It is therefore not surprising that the synthesis of mesoporous silica, alumina and other oxides using surfactant aggregates as templates was a topic of interest for quite a few EPR groups. Importantly, combination of EPR with other methods enabled chemists to unravel the complex mechanism of formation of mesoporous silica. Early papers by Galarneau and Ottaviani20–22 followed the synthesis of MCM-41 on CTAB template using cationic (4-(N,N 0 -dimethyl-Nhexadecyl)ammonium-2,2 0 ,6,6 0 -tetramethyl-piperidine-1-oxyl iodide, CAT 16) and anionic (5-doxyl stearic acid, 5-DSA) spin probes. The spin probes reported on the changes of diffusion properties in the organic template during silica synthesis. The authors selected an inorganic source of silica (sodium silicate) to slow down the reaction in order to observe intermediate steps of mesophase formation. Important changes in the micellar structure of CTAB were observed a few minutes after addition of the silica solution: increase of rotational correlation time for CAT 16 and appearance of order, reported by

108

Electron Paramag. Reson., 2007, 20, 96–130

5-DSA, which was attributed to the effect of silica layer on the micellar surface. Both these results suggest very fast formation of a more ordered and rigid structure of the micelles. After 3 h, a series of complex changes were reported. Importantly, the CAT-16 spectra revealed the appearance of an increasing fraction of molecules immobilised by strong interactions with the silica layer. The authors seem to have found evidence for the sphere-to-rod transition of the micelles in the line shape of 12-DSA probe, as simulations required the use of a tilt angle between the molecular axis and the director. It is however surprising that a similar change was not reported for 5-DSA spectra, where the diffusion around the y-axis should produce more visible changes, especially at temperatures used in this study (e.g., above room temperature). The effect of additives and surfactant chain length in the CnTAB series (n ¼ 8–18) on silica formation was also investigated by the same authors. The synthesis of MCM-41 silica on the same template but with TEOS as the silica source was studied by Goldfarb et al.23 by in situ CW EPR measurements of several spin probes dissolved in the reaction mixture. The line shape analysis revealed the existence of different stages during the evolution of precipitate. The spectral line shape was simulated using the MOMD model (Brownian diffusion) using the NLSL programme.13 Thus, the spectrum of CAT-16 prior to TEOS addition represents a superposition of micellar and solution spectra. After TEOS addition, the micellar spectrum changed gradually into a rigid limit spectrum (Figure 6). This observation provided direct evidence that the micellar structures present in the initial reaction mixture serve as precursors for the ﬁnal mesoporous product and there are no intermediate colloidal structures. The temporal evolution of the spectrum is characteristic of an isotropic system undergoing a gradual increase in microviscosity. The evolution of the EPR spectrum was compared to that of the X-ray diffraction pattern which indicated that the hexagonal long-range order is already formed 5–8 minutes after mixing the reagents. Unfortunately, this could not be conﬁrmed unambiguously by EPR since the lateral diffusion rate of the probe around the cylindrical axis is too slow at the synthesis temperature to produce the lineshape changes speciﬁc for the sphere-to-rod transition. Formation of the inorganic phase, which could be followed by the slow-down of the spin-probe motion, is considerably slower (41.5 h). The use of the 5-DSA spectra recorded in situ during formation of MCM41,24 showed that as the reaction progresses, the probe experiences a growing ordering potential while the rotational diffusion rates decrease. The time evolution of these parameters indicated the presence of two stages. During the ﬁrst 12 minutes the changes were rapid, but during the next hour the diffusion constants remained constant, while the order potential increased slightly (Figure 7). The kinetic data can be complemented by the information about the precise location of various components of the system at different stages of silica synthesis. The location of the spin probes relative to the template and to the inorganic phase have been established by electron spin echo envelope modulation (ESEEM), by following k(2H), the deuterium modulation depth due to

Electron Paramag. Reson., 2007, 20, 96–130

109

Figure 6 EPR spectra of the CAT16 spin probe during the synthesis of MCM-41 after subtraction of the aqueous component (Reprinted with permission from Goldfarb et al.23 Copyright (1997) American Chemical Society)

the deuteriated surfactant derivatives: CTAB-d9, CTAB-a-d2, and D2O.25 The measurements showed that 5-DSA is located in the organic phase, near the polar heads, so the spectra are relevant for the processes occurring at the interface. The possibility to describe the hyperﬁne interaction with surrounding magnetic nuclei using ESEEM was previously demonstrated for colloidal systems by Kevan and co-workers in the 80’s, in an impressive range of papers (see ref. 160–205 in ref. 1). Recently, Golfarb et al. synthesised a specifically designed spin-labelled silica precursor (SL-TEOS) probe, which produced direct evidence about the silica layer formation.26 Once again, ESEEM measurements established the probe position. It was found that before precipitation, the probe was located near the CTAB polar heads, subsequently shifting towards silica layer during precipitation. The results also showed that the as-synthesised silica layer contained a large quantity of water which can be removed by simple ﬁltration, when a contraction of the silica layer around micelles takes place. The process is reversible unless a hydrothermal stage is applied, which leads to a higher degree of silica cross-linking. The synthesis of another hexagonal mesoporous silica, SBA-15, where the template is a Pluronic block-copolymer [poly(ethylene oxide)-b-poly(propylene oxide)-b-poly(ethylene oxide)] was also examined by Goldfarb and

110

Electron Paramag. Reson., 2007, 20, 96–130

Figure 7 Temporal evolution of rotational diffusion during synthesis of MCM-41 as determined by EPR line shape simulation (Reprinted with permission from Goldfarb et al.24 Copyright (2000) American Chemical Society)

co-workers.27 In this case, a series of three Pluronics with different lengths of poly(ethylene oxide) chain (L-62, P123, F127) were spin-labelled at the terminal hydroxyl group and these molecules were used as spin probes with the radical moiety placed at different radial positions in the corona (conﬁrmed by ESEEM). The longest spin probe, extending outside the micelle, was included in the silica layer. This fact is a direct evidence that poly(oxyethylene) chains included in the silica layer are at the origin of the micropores observed in SBA-15. It could also be concluded that TEOS is initially dissolved in the micelle core and that silica formation propagates outward from the core-corona interface. Somewhat unexpectedly, the as-synthesised materials have a rather ﬂuid structure unless a (hydro-) thermal treatment is applied. This behaviour seems to be quite general; for instance it was also observed in the synthesis of mesoporous alumina. In this reaction, Pluronic P123 was used as a template in organic solvents, with Al(sec-BuO)3 as the alumina source. The authors28 found that significant structural changes occur in the as-synthesised alumina when treated with different solvents after precipitation: the micelle aggregates serving as templates modify their packing (while in the precipitate) during treatment with different solvents and – more importantly – this induces corresponding changes in the pore dimensions of the ﬁnal material. The process can be monitored by the line shape changes of 5-DSA probe (Figure 8). The

Electron Paramag. Reson., 2007, 20, 96–130

111

Figure 8 EPR spectra of spin probe 5-DSA in as-synthesised alumina washed with (a) acetonitrile, (b) water, and (c) sample b washed with acetonitrile again (Reprinted with permission from Caragheorgheopol et al.28 Copyright (2005) American Chemical Society)

modiﬁcations induced are reversible; however heating to ca. 1001C stabilises the actual structure. The neutral synthesis pathway which relies on self-assembly of dodecylamine (DDA) in EtOH/water mixtures and TEOS as a silica source, was examined by Caldararu et al.29 using a series of neutral, cationic and doxylstearic acid spin probes. DDA aggregates in initial solutions were characterised for two EtOH/ water mixture compositions as (i) micelles, when the solvent was a homogeneous, ethanol-rich mixture and (ii) emulsion droplets in a biphasic system when the solvent was water-rich. These characteristics of the starting solutions could be related to the pore structures of the ﬁnal silica. For instance, the structural mesopores most likely result from the micelle templates, while textural pores are formed around emulsion droplets. The authors also observed the structural ﬂuidity of silica layer in the as-synthesised material, in agreement with 29Si NMR data which conﬁrm the low level of network cross-linking. Reversible changes were observed in the drying-hydrating cycle, unless the sample was heated to ca. 1001C for at least 2 h. DDA was also used in a different solvent system, i.e. aqueous acetonitrile (AN).30 Caragheorgheopol and co-workers used three different solvent compositions (v/v): (a) AN/water ¼ 2:8, (b) AN/water ¼ 1:1 and (c) AN/water ¼ 8:2. The EPR study of starting solutions (using spin probes) showed that they are very different: (a) is a heterogeneous system containing lamellar liquidcrystalline structures, (b) is a micellar solution and (c) is a molecular solution which did not show formation of any aggregates. Surprisingly, all three systems produce very similar mesoporous silicas.

112

Electron Paramag. Reson., 2007, 20, 96–130

In another synthesis of mesoporous alumina, anionic sodium dodecyl sulphate (SDS) was used as a surfactant, Al(NO3)3 as alumina source and urea as a precipitating agent (by heating).31 The formation of mesostructured hexagonal alumina was monitored using spin probes incorporated in the SDS aggregates. No structural changes were observed during urea hydrolysis up to pH 6.1, before the onset of the precipitation. In the precipitate, the presence of the surfactant template was demonstrated as well as its interaction with the aluminium species. The EPR spectra showed structural changes at the template/silica interface during precipitation, i.e. tilting of CAT 16 head group after about 6 h and tilting of the C12NO head group (positioned deeper in the micelle) after ca. 12 h. While spin probes cannot unfortunately provide direct information about the geometry of template organisation, they proved to be very sensitive to subtle local changes during synthesis of mesoporous materials, and some changes can only be probed by EPR. In particular, spin probes can help identify the aggregation state of surfactants in the starting solutions, observe aggregation, follow the order, viscosity, dynamics, polarity at different sites in the system (core, interface, silica layer), follow the evolution of template or inorganic phase during precipitation, provide kinetic data about changes during precipitation, evidence effects of post-precipitation treatment (washing with different solvents, drying/re-hydration). Thus, in each speciﬁc case EPR methods significantly helped to uncover important mechanistic details.

5

Adsorption in Mesopores

The physico-chemical behaviour of liquids in conﬁned systems play an important role in many ﬁelds of science and has been extensively studied in zeolites, silica gel, micelles and other nano-heterogeneous systems. The phenomena involved are capillary condensation, freezing/melting, liquid-liquid equilibriums, host-guest interactions, chemical reactions. MCM-41 silica and its analogues have regular, well-deﬁned nanometer-scale pores and therefore represent ideal materials for the type of studies mentioned. For instance, Okazaki et al.32–34 used EPR of spin probes to model the diffusion and ﬂow at the level of molecular interactions of solute, solvents and silica nanochannel walls. Starting with the observation of a photoreduction reaction in MCM-41 nanotubes, the authors were led to the conclusion that molecular diffusion is inhibited in the longitudinal direction of the nanotubes. This was attributed to the lack of excess volume for exchanging the position of solutes or solvent molecules, and the molecular interactions between molecules. As a result, molecules ﬂow collectively and the pressure necessary to ﬂow the solution is far below that estimated by classical equations (Hagen-Poiseuille law). The research in this area was concentrated on the dynamics of spin probes in MCM-41 nano-channels in different solvents. The di-tert-butyl nitroxide (DTBN) spin probe dissolved in 2-propanol was found to have a preferential

Electron Paramag. Reson., 2007, 20, 96–130

113

rotation around the axis connecting the two t-Bu groups, supposedly in the axial region of the channel. When cyclohexane was used as a solvent, DTBN interacted strongly with the silica and an immobilised spectrum was observed. Addition of a small quantity of 2-propanol stripped the DTBN molecule from the surface, thus suggesting a possible inhomogeneous distribution of liquid molecules in the nanochannel. Upon cooling a DTBN/2-propanol mixture in the nano-channels, an immobilised spectrum appeared 601C higher than the bulk melting point of 2-propanol, along with a population of rapidly rotating radicals. This may occur because in the small volume of the nanochannel (diameter and length are about 3.4 nm and 5 mm, respectively), the molecules in solution do not diffuse as freely as in the bulk. The two-component spectra observed several degrees above the melting point were explained by the phase separation model: an ordered phase in the central portion of the channel and a random phase near the channel wall. Since the guest molecules can get a larger free volume in the random phase, the spin probe gives a sharp signal because of rapid rotation. On the other hand, spin probe molecules in the ordered phase may be immobilised due to the reasons presented above. This tentative assignment is however somewhat counter-intuitive. Host-guest interactions of dendrimers with porous materials are of interest in relation to their possible applications and for possible modiﬁcations of solid surface properties. Dendritic polymers are composed of layers of branched monomer units radiating from a polyfunctional central core, with each layer forming the next dendrimer generation. The controlled structure of dendrimers is reﬂected in the size monodispersity and uniform structural characteristics which can be tuned by using selected generations. The globular structure and the controlled branching of dendritic structures provide a three dimensional architecture that allows functional groups to be incorporated in a geometrically well-deﬁned fashion. The nanoscopic dimensions of well-deﬁned functionalised dendrimers may give rise to new properties in, for example, molecular recognition, catalysis, or molecule-based electronics and optics. A systematic study of the interaction of dendrimers with porous surfaces of different chemical nature and porosity was undertaken by Ottaviani et al.35 by using spin-labelled poly(amidoamine) dendrimers (PAMAM) of two generations: G2 and G6, which differ largely by their size and ﬂexibility. The solids involved were silicas of different pore sizes (40–500 A˚), activated alumina (acid, neutral and basic) and MCM-41 zeolite. The quantity of spin-labelled dendrimer adsorbed on the above porous materials was determined from the concentration change of the supernatant solution as a function of the starting dendrimer concentration (as assessed by quantitative EPR measurements). The solid was separated by ﬁltration and gently dried on ﬁlter paper. The corresponding EPR spectra were found to consist of an ‘‘interacting’’ (immobilised) and an ‘‘essentially free’’ component. Spectra simulations made it possible to separate the two contributions. The free component can be washed out from the silica surface, while the interacting part cannot. The tc values of the ‘‘essentially free’’ component increased

114

Electron Paramag. Reson., 2007, 20, 96–130

substantially when conﬁned in small volumes. The authors also followed the effect of pH on the described interactions. The authors concluded that the adsorption and interaction of dendrimers with the porous surfaces is determined by the pore size of the host (e.g., porous material) relative to the guest (dendrimer). If the pore size is smaller than the dendrimer diameter, the adsorption occurred exclusively on the external surface and small amount of dendrimer is adsorbed. Conversely, if the pore size is bigger than the dendrimer size, large amounts are adsorbed on the inner surface. Another factor strongly affecting the dendrimer-porous materials interactions is the nature of the surface groups on both host and guest, as it determines the strength of intermolecular forces (e.g., electrostatic, hydrophobic or dispersion forces). The effect of intermolecular interactions on the behaviour of pore-adsorbed molecules was nicely demonstrated by a spin-labelling study. Ottaviani et al. attached a relatively polar nitroxide spin probe to the surface of mesoporous silica.36 Importantly, the mobility of the probes showed very strong solvent dependence. In non-polar solvents or in the dried state, the pore-conﬁned probe showed a nearly immobilised spectrum. In water, however, the spectrum was typical of a fast motion regime. This behaviour was interpreted by the competition of spin-probe interactions with the solvent and silica surface – the phenomenon which lies at the heart of silica chromatography. Smirnov et al.37 reported an ingenious approach to explore the behaviour of spin labels incorporated into an organised porous material. The substrate used in this study was prepared from anodic aluminium oxide (AAO) disks, which have an aligned through-ﬁlm porous structure, macroscopically homogeneous and hexagonally packed with pore diameter tuneable from ca. 4 to 200 nm. Such inorganic substrates can be functionalised by self-assembly of lipid bilayers on their surfaces, a convenient model of cellular membranes. Biofunctionalisation of substrates is one of the most attractive ways for building hybrid nanoscale devices for studying membrane-protein interactions and mechanisms of intracellular processes. Lipid bilayers assembled in nanotubes can more easily be kept hydrated, have larger surface area and are much more robust compared to the planar ones, which expose the entire surface to the environment and can be easily damaged or contaminated. Bilayers of self-assembling cylindrical phospholipid structures were supported on the inner surfaces of AAO. Spin-labelled dimyristoylphosphatidylcholine (5-PC) was introduced with the parent unlabelled lipid in the nanotubes. The fully hydrated samples were characterised by X- and W-band EPR at different temperatures. A simple change in the AAO substrate orientation changes the alignment of the oriented lipid molecules with respect to the magnetic ﬁeld. Indeed, 150 K spectra in W-band showed preferentially x or z features upon changing the chip orientation (Figure 9). This conﬁrms the good alignment of the phospholipids in the pores of the support. Room temperature spectra are less orientation-dependent, with the order parameters being close to those of aqueous dimyristoylphosphatidylcholine liposomes. This indicates the mobility of the lipids inside AAO nanopores.

Electron Paramag. Reson., 2007, 20, 96–130

Figure 9

6

115

Rigid limit W-band EPR spectra of AAO substrate with lipid bilayer incorporating a 5-PC spin probe at two orientations with respect to the magnetic ﬁeld (Reprinted with permission from Smirnov et al.37 Copyright (2003) American Chemical Society)

Adsorption on Surfaces

6.1 Planar Surfaces. – Self-assembled monolayers on ﬂat substrates are the most basic organised assemblies and hence attract much academic interest. Besides, monolayer ﬁlms have many important technological applications (e.g. as highly speciﬁc sensors, in optical and electronic devices). Many analytical tools do not have high enough sensitivity for studying monolayers. Here, the advantages of superior sensitivity of EPR can be fully exploited. Conventional X-band spectrometer can easily detect an incomplete monolayer on a ﬂat substrate with surface area of just a few cm2. As spin probes are powerful tools to investigate interfacial phenomena,38 these methods are highly applicable to monolayer studies. Risse et al.39–41 investigated fatty acid monolayer ﬁlms grown on a ﬂat Al2O3 surface. Such studies can elucidate the molecular dynamics of surface-bound molecules (which is distinct from LB multilayer ﬁlms in which the interactions with the surface are negligible). A number of doxyl stearic acid (DSA) spin probes with the nitroxide moiety positioned at different distances from the carboxylic acid head group were introduced into the stearic acid ﬁlm and their EPR spectra at different concentrations and temperatures were studied. The experimental spectra could only be ﬁtted by considering 2-4 different populations of spin labels having different dynamic parameters. Increasing the temperature over a range of more than 1001C changed the relative contributions of these populations rather than their dynamic parameters. Careful simulation of

116

Electron Paramag. Reson., 2007, 20, 96–130

the variable temperature spectra for 16-, 12- and 5-DSA showed that: (i) the onset temperature for rotational motion on the EPR time scale decreases with increased distance of the spin label from the surface; (ii) the line shape of the spectra could only be described by a combination of spectra with different rotational diffusion tensors; (iii) the fraction of the more mobile components increases with increasing temperature; (iv) for a given temperature the rotational motion of the molecules increases drastically with increasing distance to the surface; (v) the rotational motion of the molecules is not isotropic. In the case of 12- and 16-DSA probes, a strong change in the adsorption behaviour with the temperature of the deposition solution was observed (at variance with 5-DSA), which can be explained by the formation of loop structures with both the carboxylic acid and the nitroxide groups bound to the surface. Self assembled monolayers of stearic acid spin probes were also investigated on planar GaAs substrates by Goldfarb et al.42 Broad-line spectra were obtained for 5- and 16-DSA probes, due to spin-exchange and dipolar interactions. For as-prepared samples, the EPR spectra of 5-DSA, with the spinlabel group located close to the substrate, were poorly resolved. However, as the samples were allowed to age (up to 1 week), the resolution of the 14N hyperﬁne coupling increased revealing a better organised monolayer with the molecules more homogenously spaced. Moreover, the spectrum of this organised ﬁlm after reaching equilibrium, showed that there is no motional freedom near the GaAs surface. Orientation-dependence measurements on the equilibrated sample showed the presence of a preferred orientation of the molecules although with a wide distribution. The spectrum of the 16-DSA monolayer, where the nitroxide spin label is located at the end of the chain and far from the substrate surface, also showed a poorly resolved spectrum for as-prepared monolayers. However, unlike 5-DSA, ageing did not lead to any spectral changes. Through EPR line-shape simulations and by comparison with FT-IR results, the differences between 5-DSA and 16-DSA layers were attributed to difference in coverage caused by the bulky spin label near the surface in the case of 5-DSA. One of the most important applications of self-assembled monolayers is in biotechnology, where they are used to control protein adsorption and cell adhesion. Moreover, such protein-functionalised surface interactions can be used to mimic biological processes. However, little is known about the structure and dynamics of the surface-adsorbed proteins. Most traditional methods, such as IR, CD or ﬂuorescence, can provide only global information on the secondary structure of absorbed proteins, or report on the properties of a speciﬁc chromophore. Risse, Hubbell et al. used site directed spin-labelling (SDSL) to specifically label protein annexin XII with nitroxide moieties and study the secondary and tertiary structure of this protein adsorbed on a single planar phospholipid bilayer.43 Nitroxide side-chains were introduced at several topological sites and their EPR spectra were recorded with the protein in adsorbed state or in solution.

Electron Paramag. Reson., 2007, 20, 96–130

117

Figure 10 EPR spectra of bilayer-adsorbed annexin XII (spin labelled at residue 213) at different angles between the magnetic ﬁeld and the bilayer normal (Reprinted with permission from Risse et al.43 with minor changes. Copyright (2003) American Physical Society)

The EPR spectra of the labelled protein reﬂect the mobility of the side chain which is uniquely determined by the local structure of the protein and the backbone dynamics. Thus, the EPR spectra for a set of side chains at speciﬁc sites throughout the protein provide a ‘‘ﬁngerprint’’ for the global protein structure and can be used to determine sequence-speciﬁc secondary structure, reveal tertiary organisation, monitor structural changes related to protein function, and map the backbone dynamics. The authors found that the spectra of the label at each site in the adsorbed state are very similar to those in solution. Since the location of labelled sites was distributed throughout the molecule, this result strongly suggests that the basic fold of the protein is preserved upon adsorption to the lipid bilayer. For one speciﬁc site, the spectrum was orientation-dependent (Figure 10). The average orientation of the nitroxide relative to the surface could be extracted from these data. The fact that the side-chains with hindered internal motions have a net orientation can thus be used to reveal the orientation of helical segments in surface-adsorbed proteins. The authors conclude that most methods developed for structural studies of spin-labelled biological molecules in solution (e.g., side-chain dynamics, interresidue distance determination using dipolar interactions, analysis of backbone dynamics) can be directly employed to analyse behaviour of surface-adsorbed proteins. 6.2 Rough Surfaces. – Somasundaran et al.38 studied adsorption of surfactants and polymers on alumina commonly used to modify particle surface properties. Spin probes 7- and 12-DSA were covalently bound to alumina by esteriﬁcation of surface hydroxy groups. Subsequently, the structure of sodium dodecyl sulphate (SDS) and poly(acrylic acid) (PAA) layers adsorbed on these labelled surfaces was investigated by observing the changes in the EPR spectra

118

Electron Paramag. Reson., 2007, 20, 96–130

of the bound probes. The results were compared to the probes electrostatically adsorbed on the surfaces. Surfactant (e.g., SDS) adsorption caused an increase of the rotational correlation time of both covalently and electrostatically bound probes, due to the formation of hemimicelles. The effect of the polymer (PAA) on the bound probes is much weaker compared to the surfactant and was attributed to the adsorption of the polymer as a network at the interface, with a much less compact structure than the hemi-micelles. In the case of covalently bound probes, a competition between the probes and the polymer for adsorption sites was observed, and probe aggregation was revealed by the appearance of a spinexchange-narrowed spectral line. Turro, Ottaviani et al. used adsorption of spin probes to explore the external surface of the MFI family of zeolites.44–46 These studies were developed in connection with the possibility to ‘‘measure’’ the external surface of these materials, an important parameter for the description of catalytic reactions. Materials studied were the ZSM-5 zeolite and poly- and mono-disperse silicalites [the same structure as ZSM-5 but containing only SiO4 tetrahedra (e.g., no AlO4)]. The external surface of ZSM-5 zeolite was studied by adsorption of various probe molecules exclusively on the external surface and subsequently measuring the rotational mobility of the probes at different surface coverage. Subsequent adsorption of ortho-methyl dibenzyl ketone (oMeDBK) – an EPR-silent compound – gradually displaced the nitroxides, starting from the strongest binding sites. When all strong-binding sites were occupied by oMeDBK, the nitroxides were displaced into solution or remained bound to the weak sites. This was clearly signalled by the appearance of sharp lines in the EPR spectra. The critical loading of oMeDBK required to bring about the onset of the fast motion of the nitroxide probes was therefore expected to be a parameter directly proportional to the number of strong binding sites. These were postulated to be pore openings; their number hence depended on the total external surface area of the silicalite sample. This assumption was checked by comparison with external surface area measured by Hg porosimetry. This study was then extended for a series of monodisperse silicalites with different dimensions by using nitroxide spin probes of different chemical structure, some of them labelled with 15N (TEMPONE, 15N-TEMPONE, TEMPOL, 4-diphenylacetyl-TEMPO (DPA-TEMPO) and 15N-DPA-TEMPO). A series of elegant experiments exploited the possibility to distinguish between 14N- and 15N-labelled nitroxides. Displacement of a 1:1 mixture of 14 N- and 15N-TEMPONE from the silicalite surface with oMeDBK led to the appearance of sharp lines which represented the two radicals in the original 1:1 ratio. However, when oMeDBK was used to displace a 1:1 mixture of 14 N-TEMPONE and 15N-DPA-TEMPO, the EPR spectra showed the two displaced nixtroxides in a different proportion. This was interpreted by stronger binding of TEMPONE compared to DPA-TEMPO. Finally, simultaneous and successive adsorption of TEMPONE and 15N-DPA-TEMPO in

Electron Paramag. Reson., 2007, 20, 96–130

119

1:1 proportion produced identical results proving the existence of efﬁcient exchange between the probes adsorbed on strong binding sites and those on weak binding sites or in solution. This methodology can be extended to the characterisation of surface binding sites of different active surfaces.

7

Au Nanoparticles

Ligand-protected inorganic nanoparticles are truly supramolecular species possessing properties typical of organic molecules and bulk materials. The most known and stable are thiol-coated Au particles 2-5 nm in diameter. These materials behave like ordinary compounds; they can be isolated in pure form, precipitated, dried and redissolved in solvents, or stored as solids. At the same time, the inorganic cores of these particles contain several hundred or thousand atoms. Nanoparticles are usually stabilised by a well-packed layer of organic ligands. The ligands are often attached to the nanoparticle surface with an anchor group (e.g., thiol) which has strong afﬁnity for the core material.47 The size and hence many properties of ligand-protected nanoparticles can be tuned by changing the experimental conditions of their synthesis. For instance, optical properties of Au nanoparticles and ﬂuorescent properties of semiconductor nanoparticles strongly depend on particle size. Nanoparticle stability, simplicity of their preparation, and useful size-dependent properties made them promising candidates for a variety of applications, including sensing, catalysis, biological labelling, electronics. Au nanoparticles are a highly suitable system for studying by EPR and spin probes/labels. High sensitivity of EPR is very beneﬁcial as it makes it possible to probe rather low concentrations of spin labels/probes. The protective organic layer is well packed, and the local microenvironment (e.g., viscosity, polarity) can be anisotropic and different from the bulk solution. The tumbling of the particle as a whole could also fall in the EPR time scale, depending on the particle size. Hence, EPR can provide direct information about the complex dynamics of the organic layer in the nanoparticles. On the other hand, the distances between several spin labels incorporated into the same nanoparticle can be in the range from very short to 7–8 nm. This range is accessible to EPR through CW measurements (for shorter distances) or pulsed methods (for longer distances). Consequently, spin labels/probes have been used to characterise several aspects of organisation and reactivity of the organic layer in Au nanoparticles. Spin labels can be attached to the nanoparticle shell using ordinary organic reactions. However, control over stoichiometry and spin-label location in this case is rather limited. A very appealing alternative way of introducing the spin labels into Au nanoparticles is through a ligand exchange reaction. In this simple method, the appropriate ligand is added to the pre-synthesised nanoparticles in order to replace the weakly-bound ligands on the nanoparticle surface. The ligand exchange reaction is usually carried out with thiols; however, nitroxide-labelled thiols are unstable and do not survive ligand exchange.

120

Electron Paramag. Reson., 2007, 20, 96–130 O N

NH O

S S S

S

S S

Au S

S S

S S S S

S

+

S

S S

S

S S

Au S

S S

S S S

NH O

N

O

O NH

N O

Figure 11 Schematic illustration of a ligand exchange reaction of Au nanoparticles with a spin-labelled ligand

Hence the easiest way to introduce a spin label into the nanoparticle is through a reaction with a spin-labelled disulﬁde (Figure 11). This seemingly simple reaction, however, showed a remarkably complex mechanism. The two branches of the spin-labelled disulﬁde ligand did not adsorb next to each other on the nanoparticle surface. This was unambiguously shown by the absence of exchange (at room temperature)48 and dipole-dipole (in frozen solutions) interactions between the spin labels after exchange reaction.49 Moreover, another EPR-active species was formed during the exchange reaction but disappeared at the end of the reaction. This species was therefore an intermediate; EPR analysis showed that it is a highly mobile mono-nitroxide. The disappearance of the spin-spin interaction in the spin-labelled disulﬁde molecule during exchange made it possible to monitor the reaction kinetics by EPR. The reaction was shown to be zeroth order with respect to the concentration of spin-labelled disulﬁde. This suggested the dissociative mechanism for the ligand exchange, e.g., the rate determining step of the reaction involves dissociation of the ligand from the nanoparticle surface. Not all binding sites on the Au nanoparticle surface were reactive: only 4–5 ligands (out of ca. 100) can be replaced with spin-labelled disulfides. Moreover, the reactivity of the Au nanoparticles in disulﬁde exchange was shown to depend strongly on their age. Simply keeping the nanoparticles in solution at room temperature over the period of hours to days significantly reduced their reactivity.50 EPR studies therefore made it possible to create a clear picture of the general mechanism of disulﬁde exchange in thiol-protected Au nanoparticles. The reaction is likely to occur at defect sites at the nanoparticle surface; the number of these sites depends on batch preparation and history. The reaction starts by dissociation of the thiolate ligands from the nanoparticle creating a vacant site. This active species then reacts with the spin-labelled disulﬁde. One branch of the disulﬁde is attached to the nanoparticle surface, while the other forms a mixed disulﬁde with the outgoing ligand. At longer reaction times, this mixed disulﬁde can also react with the nanoparticle surface with the same mechanism (Figure 12).

121

Electron Paramag. Reson., 2007, 20, 96–130

R'S-SR', k2

SR

RS

RS-SR', k2/2

R'S- Au R'S- Au

+ RS-SR

RS- Au

+ RS-SR'

RS- Au

+ RS-SR

rate-determining Au Thiolprotected particles

step, k1

Au Reactive species

RS-SR', k2/2

RS-SR, k2

+ RS-SR'

New thiolDisulfides protected particles Figure 12 Proposed mechanism for ligand exchange of Au nanoparticles with disulfides

As the ligand exchange does not lead to two branches of nitroxide-labelled disulﬁde adsorbing next to each other on the nanoparticle surface, it could be used to prepare nanoparticles with a controlled coverage of spin labels randomly distributed on the surface. At low coverage (e.g., mostly singlylabelled), such nanoparticles can provide much information about the dynamics of the organic layer. The mobility of the nanoparticle-adsorbed spin label was shown to depend on both the length of the linker connecting the spin label to the nanoparticle surface, and the nature of the surrounding ligand (Figure 13). The mobility is the highest when the spin label on a long linker protrudes above the layer of the surrounding ligand. Conversely, the spin label is most immobilised if it is attached to the nanoparticle surface with a short linker and is surrounded by a well-packed layer of long chain ligands.51 Interestingly, the nanoparticle size also affects the mobility of the surfaceattached spin label, with smaller nanoparticles showing higher mobility. Two factors probably contribute to this effect: (i) the tumbling of the whole nanoparticles, and (ii) looser packing on the surface of smaller nanoparticles (e.g., with higher surface curvature).52 High coverage spin-labelled nanoparticles can be prepared by reaction of spin-labelled disulﬁde with the particles coated with a very weakly bound ligand, e.g., triphenylphosphine. In this way, nearly complete coverage of spin label on the nanoparticle surface can be achieved. EPR spectra of high-coverage spin-labelled nanoparticles show broad lines in both ﬂuid and frozen solution, due to exchange and dipole-dipole interactions. At intermediate coverage, analysis of these interactions can provide important information about the distribution of the spin labels on the nanoparticle surface and their conformation. For instance, dipolar broadening was the subject of a CW EPR study. The broadening was correlated with the distribution of interspin distances on the nanoparticles, calculated using simple geometrical models and statistical analysis. The results showed that the spin labels

122

Electron Paramag. Reson., 2007, 20, 96–130

Figure 13 EPR spectra of Au nanoparticles protected by butane- (a, d), octane- (b, e) and octadecanethiol (c, f). The linker between the spin label and the Au surface contained 2 (a–c) and 10 (d–f) methylene groups, respectively (Reprinted with permission from Chechik et al.51)

protruding above the layer of the surrounding ligands do not extend into solution, but instead wrap around the nanoparticle surface.53 Similar results were obtained from a complementary DEER study. The location of the spin labels on the nanoparticle surface was best described by a random distribution function. This further conﬁrms the lack of cooperative adsorption of spin labels on the nanoparticle surface.54 The properties of the organic layer in the nanoparticles can also be assessed using spin probes. In a very elegant study, Pasquato, Lucarini and co-workers used dialkyl nitroxides to investigate the partitioning of molecules between the bulk solution and nanoparticle shell.55 The Au nanoparticles were protected by an amphiphilic thiol. The outer, hydrophilic part of the organic shell on the nanoparticle surface ensured solubility in water. The inner, hydrophobic part of the shell served as a pool to solubilise hydrophobic molecules. These nanoparticles can thus be thought of as covalently linked micelles assembled around the metal core (Figure 14). Partitioning of the dialkyl nitroxides between the hydrophilic bulk solution and the hydrophobic interior of the organic shell can be easily monitored by EPR thanks to the sensitivity of the nitrogen and b-proton hyperﬁne splitting to the polarity changes (Figure 14). The rate of nitroxide exchange between the bulk solution and the nanoparticle shell happened to be slow or fast on the EPR timescale depending on the temperature of the experiment. The simulation of the exchange-derived line broadening hence made it possible to assess not

Electron Paramag. Reson., 2007, 20, 96–130

123

Figure 14 Partitioning of a nitroxide spin probe between bulk solution and the shell of Au nanoparticles coated with an amphiphilic ligand (Reprinted with permission from Lucarini et al.56 Copyright (2004) American Chemical Society)

only the equilibrium, but also the rates of nitroxide incorporation into the nanoparticle shell. The same authors also explored the effect of particle size on the partitioning of the dialkyl nitroxide probe.56 They found that the binding becomes significantly weaker with increased nanoparticle size. This is likely to be due to the effect of nanoparticle curvature on the packing. In small particles, the packing of the organic shell is expected to be loose, thus facilitating easy incorporation of the guest molecules. In large particles, the packing is likely to be much tighter, which makes guest inclusion more difﬁcult both thermodynamically and kinetically. It would be interesting to compare these results with micelles of similar composition in order to see how the packing is inﬂuenced by the different assembling principles. 8

Dendrimers

The poly(amidoamine) (PAMAM) family of dendrimers possesses branches consisting of amidoamine units. The branches at the external surface terminate with amino groups (full generations, integer G’s) or carboxylate groups (half generations, half-integer G’s). Molecular simulations of the PAMAM structure showed that a change in the dendrimer morphology occurs for full generations at around G4. The earlier generations are characterised by an open and solvent accessible external structure, whereas the later generations are characterised by a densely packed external surface. All these properties make dendrimers very important synthetic supramolecular hosts. Several applications of dendrimers in biomedicine have been described. The internal structure and size of PAMAM dendrimers mimic biological macromolecules. They are biocompatible and can be used in the pharmaceutical and biochemical ﬁelds as drugs and vehicles of biological materials. PAMAM dendrimers have been used as viral vectors and gene carriers to deliver DNA sequences in cells. Ottaviani, Turro and Tomalia addressed dendrimer interactions with components of biological systems. This research follows earlier work of the same

124

Electron Paramag. Reson., 2007, 20, 96–130

authors on the interaction of PAMAM dendrimers with surfactants and paramagnetic ions.57,58 Multicomponent spectra were observed. The approach used by the authors consisted of subtraction-addition procedures of experimental spectra obtained in different conditions in order to get the individual spectra. These were then computer simulated and added in the proper relative amounts to reproduce the experimental line shape. In one application, the interactions of PAMAM dendrimers (G2 and G6) with calf thymus DNA were studied by EPR spectroscopy and a number of other techniques (CD, UV-Vis).59 The dendrimers were labelled with TEMPO and their adducts with DNA were observed at various ratios of dendrimer surface groups to the DNA base pairs (abbreviated as r). The self-complementary results led to the following model. At small dendrimer/DNA ratios (r o 1) only a small effect of dendrimer on DNA is observed. For intermediate values of r (10–200), precipitation of the components occurs due to charge neutralisation. However, at higher r values, (e.g., r 4 200), re-solubilisation was observed due to a ‘‘salt in’’ effect, when stable dendrimer-DNA supramolecular structures were formed. This effect was accompanied by a saturation behaviour of EPR parameters; the equilibrium constant for the adduct of the G6 PAMAM dendrimer with DNA could be roughly estimated as 1.5 102 M1. This result demonstrated that PAMAM dendrimers of different generations are able to bind DNA by electrostatic interactions between the positively charged protonated amino groups on the dendrimer surface and negatively charged phosphate groups of DNA. Other papers from the same group studied the interaction of G2 and G6 PAMAM dendrimers with vesicles formed from pure dimyristoylphosphatidylcholine (DMPC)60 and its mixtures with the corresponding phosphatidylcholate sodium salt (DMPA-Na).61 In order to follow the dendrimer/vesicle interactions, two spin probes were selected, 5-DSA and CAT 16, with nitroxide groups located near the water interface. Spin-labelled dendrimers were also used in this study. The authors followed the effects of concentration and pH on the dendrimer-vesicle interactions by monitoring the changes in the following parameters (obtained from spectra simulations): rotational correlation times of spin probes in vesicles, the order parameters, the diffusion tilt angles of the rotation axis vs magnetic molecular direction, intrinsic line-widths. The results were compatible with a model in which vesicles wrap around large dendrimers (G6 PAMAM) and simply interact with the small ones (G2 PAMAM). Protein-carbohydrate interactions play a critical role in biological recognition events. Saccharide-modiﬁed dendrimers are an attractive model system displaying a multivalent carbohydrate surface. Cloninger et al. have used dendrimers as scaffolds for carbohydrate attachment by functionalising G4 PAMAM with mannose.62,63 Heterogeneous functionalisation of dendrimer terminal groups with a different number of saccharide residues along with other functionalities such as hydroxyl groups, is attractive because it makes it possible to ﬁne-tune the protein afﬁnity and protein clustering properties of the macromolecules. Not only can the size of the dendrimer framework be

Electron Paramag. Reson., 2007, 20, 96–130

125

varied easily by changing the generation of the dendrimer used, but the density and environment of the sugar units on the dendrimer surface can also be readily manipulated through heterogeneous functionalisation. In order to design heterogeneous multivalent recognition systems effectively, it is necessary to characterise the distribution of functional groups on the dendrimer surface. For this purpose, the authors functionalised the dendrimer with mannose and TEMPO units in various ratios and measured the linebroadening produced by dipole-dipole interactions between the dendrimerbound radicals (approximated as the ratio of ﬁrst/second line in the rigid limit spectrum) as a function of the density of TEMPO units on the dendrimer surface. The result of this study conﬁrmed the random character of functionalisation. Afﬁnity chromatography followed by EPR examination of eluted fractions established that binding of the mannose-TEMPO dendrimers to Concanavalin A (a mannose-binding lectin) does not depend on the speciﬁc distribution of the mannose groups on the dendrimer surface; the observed binding is due to the dendrimer molecules randomly functionalised with mannose functionalities. Dendrimer functionalisation is thus a very attractive way of creating multifunctional supramolecular systems. The interaction between pendant functional groups at the dendritic surface and their dynamic behaviour is of fundamental importance for understanding the properties of dendrimers. In analogy to their linear macromolecular counterparts, the dynamic behaviour of pendant end groups and interactions between pendant groups on the surface of the same dendrimer molecule can be probed using EPR spectroscopy and spin labelling. Bosman, Janssen and Meijer prepared a series of ﬁve generations of fully functionalised poly(propylene imine) dendrimers (DAB-dendr-(NH2)n; n ¼ 2, 4, 8, 16, 32, 64) with carboxy-proxyl end groups.64,65 The EPR spectra of the ﬁrst two generations of spin-labelled dendrimers consist of 5 and 9 lines and can be easily assigned to exchange-coupled 2 and 4 radicals, respectively. For n ¼ 8 one can still distinguish partly resolved features separated by aN/8. For higher generations (e.g., n ¼ 16, 32 and 64), the decreasing spacing between hyperﬁne lines leads to the loss of resolution and appearance of a single broad line with decreasing peak-to-peak line width (Figure 15). These spectra are well simulated assuming interactions of n paramagnetic units with hyperﬁne couplings of aN/n with the same line width as observed for the parent monoradical under the same conditions. The authors discuss the theoretical basis of the observed spin-spin exchange for the case of biradical. The room temperature line shape is consistent with the presence of at least two classes of conformers, abbreviated as A and B (Figure 16), which are characterised by significantly different exchange interaction: in an extended structure (e.g., A), the exchange integral J is small compared to the 14N hyperﬁne coupling (JA { aN), whereas in a conformation with two nitroxyl radicals in close proximity (such as B), a strong exchange interaction is present (JB c aN). In solution, both classes of conformers will

126

Electron Paramag. Reson., 2007, 20, 96–130

Figure 15 EPR spectra of proxyl-functionalised DAB dendrimers of different generations. The number n shows the number of spin labels on the dendrimer surface (Reprinted with permission from Janssen et al.65 Copyright (1997) American Chemical Society)

Electron Paramag. Reson., 2007, 20, 96–130

127

Figure 16 Schematic of the proposed diradical conformers (Reprinted with permission from Janssen et al.65 Copyright (1997) American Chemical Society)

co-exist and the actual EPR spectrum depends on the values of JA and JB and the rate of interconversion between A and B. The observed line shape can be explained by two models: (i) slow exchange, when the two spectra of conformers A and B are overlapped or (ii) rapid exchange, in the intermediate rate region, which leads to alternating linebroadening. Spectra simulations as well as the evolution of spectral shapes at higher temperature conﬁrmed the second hypothesis. Moreover, the ratio of the doubly integrated intensities in the experimental spectra is very close to 1:2:3:2:1, further supporting the fast modulation of exchange interaction. Caragheorgheopol et al. introduced a spin probe at the focal point of an L-lysine based dendritic system.66 The basic spin probe (4-amino-TEMPO) was bound to the carboxylic acid group at the dendrimer core through hydrogen bonding interactions. Different generations of these dendrimers formed organogels with diamines. The spin probe can thus be used to prove the carboxylate-amine binding, which is vital for the gel formation. The authors demonstrated that the mobility of 4-amino-TEMPO is diminished on binding to the dendritic branch. Notably, this effect is generation dependent, with larger dendritic branches having a more important effect on the radical tumbling. Control experiments clearly proved that the binding of amines is speciﬁc to the carboxylic acid groups and demonstrated that the effective binding only occurs in non-polar solvents.

9

Conclusions

We have attempted to illustrate how classical EPR methods of spin labels/spin probes can be complemented by new techniques, including high ﬁeld EPR, and pulsed techniques such as ESEEM, ENDOR and PELDOR/DEER. We believe that application of these new methods to nanoscience will help obtain molecular information about many complex phenomena taking place in nanostructured materials. Thanks to the efforts of several research groups, much has already been done (as we have tried to outline in this review); however we

128

Electron Paramag. Reson., 2007, 20, 96–130

believe that these studies only scratch the surface, and ingenious use of new methodology, smart spin labels, elaborate data processing will uncover much more information about the molecular structure and processes in the world of nanomaterials. We hope that highlighting the scope of problems in nanoscience which can be addressed by EPR will help further stimulate the interest in this area.

Acknowledgements The authors thank the Romanian Ministry of Education and Research for funding AC travel to York (project CEEX III No. 2/2005) which greatly facilitated preparation of this review. References 1. A. Caragheorgheopol and H. Caldararu, in Specialist Periodical Reports: Electron Paramagnetic Resonance, Vol. 17, Royal Society of Chemistry, 2000, p. 205. 2. P. Franchi, M. Lucarini and G.F. Pedulli, Curr. Org. Chem., 2004, 8, 1831. 3. G. Ionita and V. Chechik, Org. Biol. Chem., 2005, 3, 3096. 4. A.D. Milov, Yu.D. Tsvetkov, F. Formaggio, M. Crisma, C. Toniolo and J. Raap, J. Am. Chem. Soc., 2001, 123, 3784. 5. A.D. Milov, Yu.D. Tsvetkov, F. Formaggio, M. Crisma, C. Toniolo, G.L. Millhauser and J. Raap, J. Phys. Chem., 2001, 105, 11206. 6. S. Ruthstein, A. Potapov, A.M. Raitsimring and D. Goldfarb, J. Phys. Chem. B, 2005, 109, 22843. 7. A. Marek, J. Czernek, M. Steinhart, J. Labsky, P. Stepanek and J. Pilar, J. Phys. Chem. B, 2004, 108, 9482. 8. J. Pilar and J. Labsky, Macromolecules, 2003, 36, 913. 9. I. Dragutan, J.G. Bokria, B. Varghese, E. Szajdzinska-Pietek and S. Schlick, J. Phys. Chem. B, 2003, 107, 11397. 10. (a) K. Yamamoto, K. Kato, Y. Sugino, S. Hara, Y. Miwa, M. Sakaguchi and S. Shimada, Macromolecules, 2005, 38, 4737; (b) Y. Miwa, K. Yamamoto, M. Sakaguchi, M. Sakai, S. Makita and S. Shimada, Macromolecules, 2005, 38, 832; (c) Y. Miwa, Y. Sugino, K. Yamamoto, T. Tanabe, M. Sakaguchi, M. Sakai and S. Shimada, Macromolecules, 2004, 37, 6061. 11. D. Hinderberger, G. Jeschke and H.W. Spiess, Macromolecules, 2002, 35, 9798. 12. D. Hinderberger, H.W. Spiess and G. Jeschke, J. Phys. Chem. B, 2004, 108, 3698. 13. D.E. Budil, S. Lee, S. Saxena and J.H. Freed, J. Magn. Res. A, 1996, 120, 155. 14. M. Pannier, V. Schadler, M. Schops, U. Wiesner, G. Jeschke and H.W. Spiess, Macromolecules, 2000, 33, 7812. 15. M. Pannier, M. Schops, V. Schadler, U. Wiesner, G. Jeschke and H.W. Spiess, Macromolecules, 2001, 34, 5555. 16. G. Jeschke, G. Panek, S. Schleidt and U. Jonas, Polym. Eng. Sci., 2004, 44, 1112. 17. S.E. Cramer, G. Jeschke and H.W. Spiess, Macromol. Chem. Phys., 2002, 203, 182. 18. D. Leporini, X.X. Zhu, M. Krause, G. Jeschke and H.W. Spiess, Macromolecules, 2002, 35, 3977. 19. C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli and J.S. Beck, Nature, 1992, 359, 710.

Electron Paramag. Reson., 2007, 20, 96–130

129

20. A. Galarneau, F. Di Renzo, F. Fajula, L. Mollo, B. Fubini and M.F. Ottaviani, J. Coll. Interface Sci., 1998, 201, 105. 21. M.F. Ottaviani, A. Galarneau, D. Desplantier-Giscard, F. Di Renzo and F. Fajula, Microporous Mesoporous Mater., 2001, 44–45, 1. 22. M.F. Ottaviani, A. Moscatelli, D. Desplantier-Giscard, F. Di Renzo, P.J. Kooyman, B. Alonso and A. Galarneau, J. Phys. Chem. B, 2004, 108, 12123. 23. J. Zhang, Z. Luz and D. Goldfarb, J. Phys. Chem. B, 1997, 101, 7087. 24. J. Zhang, Z. Luz, H. Zimmermann and D. Goldfarb, J. Phys. Chem. B, 2000, 104, 279. 25. J. Zhang, P.J. Carl, H. Zimmermann and D. Goldfarb, J. Phys. Chem. B, 2002, 106, 5382. 26. D. Baute, V. Frydman, H. Zimmermann, S. Kababya and D. Goldfarb, J. Phys. Chem. B, 2005, 109, 7807. 27. S. Ruthstein, V. Frydman and D. Goldfarb, J. Phys. Chem. B, 2004, 108, 9016. 28. A. Caragheorgheopol, H. Caldararu, G. Ionita and F. Savonea, Langmuir, 2005, 21, 2591. 29. H. Caldararu, A. Caragheorgheopol, F. Savonea, D.J. Macquarrie and B.C. Gilbert, J. Phys. Chem. B, 2003, 107, 6032. 30. H. Caldararu, A. Caragheorgheopol, D.J. Macquarrie, B.C. Gilbert, R. Luque and F. Savonea, manuscript in preparation. 31. A. Caragheorgheopol, H. Caldararu, M. Vasilescu, A. Khan, D. Angelescu, N. Zilkova and J. Cejka, J. Phys. Chem., 2004, 108, 7735. 32. M. Okazaki and K. Toriyama, J. Phys. Chem. B, 2003, 107, 7654. 33. M. Okazaki and K. Toriyama, J. Phys. Chem. B, 2005, 109, 13180. 34. M. Okazaki and K. Toriyama, J. Phys. Chem. B, 2005, 109, 20068. 35. M.F. Ottaviani, N.J. Turro, S. Jockusch and D.A. Tomalia, J. Phys. Chem. B., 2003, 107, 2046. 36. M.F. Ottaviani, L. Mollo and B. Fubini, J. Coll. Interface Sci., 1997, 191, 154. 37. A.I. Smirnov and O.G. Poluektov, J. Am. Chem. Soc., 2003, 125, 8434. 38. X. Yu and P. Somasundaran, Langmuir, 2000, 16, 3506. 39. T. Risse, T. Hill, M. Beckendorf, U.J. Katter, H. Schlienz, H. Hamann and H.-J. Freund, Langmuir, 1996, 12, 5512. 40. T. Risse, T. Hill, J. Schmidt, G. Abend, H. Hamann and H.-J. Freund, J. Phys. Chem. B, 1998, 102, 2668. 41. T. Risse, T. Hill, J. Schmidt, G. Abend, H. Hamann and H.-J. Freund, J. Chem. Phys., 1998, 108, 8615. 42. S. Rubinstein, R. Artzi, D. Goldfarb and R. Naaman, Phys. Chem. Chem. Phys., 2005, 7, 524. 43. T. Risse, W.L. Hubbell, J.M. Isas and H.T. Haigler, Phys. Rev. Lett., 2003, 91, 188101-1. 44. Z. Liu, M. F. Ottaviani, L. Abrams, Z. Liu and N.J. Turro, J. Phys. Chem.A, 2004, 108, 8040. 45. N.J. Turro, X.-G. Lei, W. Li, Z. Liu, A. McDermott, M.F. Ottaviani and L. Abrams, J. Am. Chem. Soc., 2000, 122, 11649. 46. M.F. Ottaviani, X.-G. Lei, Z. Liu and N.J. Turro, J.Phys.Chem.B., 2001, 105, 7954. 47. M.-C. Daniel and D. Astruc, Chem. Rev., 2004, 104, 293. 48. P. Ionita, A. Caragheorgheopol, B.C. Gilbert and V. Chechik, J. Am. Chem. Soc., 2002, 124, 9048. 49. P. Ionita, A. Caragheorgheopol, B.C. Gilbert and V. Chechik, Langmuir, 2004, 20, 11536.

130

Electron Paramag. Reson., 2007, 20, 96–130

50. V. Chechik, J. Am. Chem. Soc., 2004, 126, 7780. 51. V. Chechik, H.J. Wellsted, A. Korte, B.C. Gilbert, H. Caldararu, P. Ionita and A. Caragheorgheopol, Faraday Disc., 2004, 279. 52. H. Wellsted, E. Sitsen, A. Caragheorgheopol and V. Chechik, Anal. Chem., 2004, 76, 2010. 53. P. Ionita, A. Caragheorgheopol, B.C. Gilbert and V. Chechik, J. Phys. Chem. B, 2005, 109, 3734. 54. G. Jeschke, V. Chechik, P. Ionita, A. Godt, H. Zimmermann, J. Banham, C.R. Timmel, D. Hilger and H. Jung, Appl. Magn. Res., Submitted. 55. M. Lucarini, P. Franchi, G.F. Pedulli, C. Gentilini, S. Polizzi, P. Pengo, P. Scrimin and L. Pasquato, J. Am. Chem. Soc., 2005, 127, 16384. 56. M. Lucarini, P. Franchi, G.F. Pedulli, P. Pengo, P. Scrimin and L. Pasquato, J. Am. Chem. Soc., 2004, 126, 9326. 57. M.F. Ottaviani, E. Cossu, N.J. Turro and D.A. Tomalia, J. Am. Chem. Soc., 1995, 117, 4387. 58. M.F. Ottaviani, P. Andechaga, N.J. Turro and D.A. Tomalia, J. Phys. Chem. B., 1997, 101, 6057. 59. M.F. Ottaviani, F. Furini, A. Casini, N.J. Turro, S. Jockusch, D.A. Tomalia and L. Messori, Macromolecules, 2000, 33, 7842. 60. M.F. Ottaviani, R. Daddi, M. Brustolon, N.J. Turro and D.A. Tomalia, Langmuir, 1999, 15, 1973. 61. M.F. Ottaviani, P. Favuzza, B. Sacchi, N.J. Turro, S. Jockusch and D.A. Tomalia, Langmuir, 2002, 18, 2347. 62. E.D. Walter, K.B. Sebby, R.J. Usselman, D.J. Singel and M.J. Cloninger, J. Phys. Chem. B, 2005, 109, 21532. 63. L.E. Samuelson, K.B. Sebby, E.D. Walter, D.J. Singel and M.J. Cloninger, Org. Biomol. Chem., 2004, 2, 3075. 64. A.J. Maliakal, N.J. Turro, A.W. Bosman, J. Cornel and E.W. Meijer, J. Phys. Chem. A, 2003, 107, 8467. 65. A.W. Bosman, R.A.J. Janssen and E.W. Meijer, Macromolecules, 1997, 30, 3606. 66. G.M. Dykes, D.K. Smith and A. Caragheorgheopol, Org. Biomol. Chem., 2004, 2, 922.

EPR, ENDOR and EPR Imaging of Defects in Diamond BY M.E. NEWTON Department of Physics, University of Warwick, Coventry CV4 7AL, UK

1

Introduction

Defects in diamond are no longer just an interesting curiosity in a material used either for its hardness in mechanical applications or unsurpassed allure as a gem. Recent and continuing developments in diamond synthesis have focused attention towards the possible exploitation of the amazing combinations of extreme properties which diamond offers. There is considerable technological interest in understanding defects in diamond, because of their importance in inﬂuencing the useful properties of diamond. Defects, even in low concentrations, inﬂuence virtually all the exploitable properties, including thermal conductivity (pure diamond conducts heat ﬁve times better than copper), optical transparency (diamond is transparent to radiation ranging from the extreme ultraviolet, through the visible and infrared and into the microwave region), carrier mobility (the highest of any wide band-gap bulk semiconductor) and the more mundane but equally important mechanical applications (e.g. the advantages offered by diamond tooling in the oil exploration industry bring resources into reach). There is insufﬁcient room in a review of this type to even list all the remarkable properties of diamond; these are covered in a variety of reference works.1–4 Defects in diamond are also inherently interesting, providing a variety of challenging problems for experimentalists and theoreticians, who endeavour to explain how so many diverse defect structures can arise; and also from a study of their formation and evolution, to understand the mechanisms which occur in diamond during and after synthesis under the inﬂuence of temperature, pressure, optical illumination, and particle irradiation. The tetrahedral symmetry about each lattice site and the simplicity of the purely covalent nature of the directed bonds resulting from the atomic s and p orbital of the carbon atoms makes the study of diamond very appealing. The importance of determining the nature, concentrations and distributions of defects in diamond is not in doubt. Electron Paramagnetic Resonance (EPR) and related techniques are amongst the most powerful weapons in our armoury, especially when used in combination with optical spectroscopy. Electron Paramagnetic Resonance, Volume 20 r The Royal Society of Chemistry, 2007 131

132

Electron Paramag. Reson., 2007, 20, 131–156

However, for many years only natural diamonds were available for study. Researchers often had little knowledge of the nature and concentration of impurities in them, or the environments in which the diamonds had resided for perhaps hundreds of millions of years (e.g. annealing history) and naturally this led to considerable speculation about the spectra observed in natural diamonds. It appears that a wider variety of paramagnetic defects are observed in covalently bonded materials like silicon,5 quartz6,7 and diamond,8 than in ionic crystal like alkali halides9 or alkaline earth halides.10 There is a critically important distinction between recognising and cataloging a centre and assigning a molecular structure. The former is a matter of measuring as many of the distinguishing parameters in the spin Hamiltonian, from which the symmetry and constituents of a defect may be apparent. Tables of characteristic parameters are undoubtedly useful;5,8 however the real challenge is the identiﬁcation of the detailed atomic structure which requires careful analysis of the magnitude and symmetry of the spin Hamiltonian matrices, particularly of any hyperﬁne structure. Recent advances in synthesis technology have significantly aided these spectroscopic studies. It was the synthesis of high purity diamond, with controlled isotopic composition, which enabled the investigation of intrinsic defects in diamond. The advances made by EPR and related techniques in our understanding of vacancies and self-interstitials in diamond are discussed in Section 3. Traditional high pressure and high temperature (HPHT) diamond synthesis, in which carbon is dissolved in a metal solvent/catalyst at high temperatures and high pressures, and re-crystallised under conditions where diamond is the stable carbon allotrope, has advanced so much that even though the conditions are extreme (e.g. 15001C and 6 GPa) in a handful of laboratories considerable control over the process is possible. Large single crystal diamonds can be produced with low impurity concentrations and high structural perfection. However, unless precautions are taken nitrogen is readily incorporated in HPHT diamond in high concentrations. Nitrogen is also the major impurity in approximately 98% of natural diamond. The progress made with EPR in investigating the many nitrogen related defects in diamond is discussed in Section 4. Some of the metal solvent/catalysts can also be incorporated in diamond and the data obtained on these is discussed in Section 5. Other dopants, such as boron, hydrogen and silicon can be introduced in HPHT diamond, although the technology does not lend itself to controlled doping, especially if one needs to control the distribution as well as the concentration of dopants. Diamond synthesis via chemical vapour deposition (CVD) has developed dramatically over the last 20 years, and now provides a route for the controlled synthesis of doped as well as high-purity synthetic single-crystal diamond. The process is essentially a simple one in which a carbon containing gas (usually methane) is mixed with an excess of hydrogen, energy is supplied, typically in the form of microwaves, to produce atomic hydrogen and a variety of reactive carbon gaseous species are transported to a substrate. By controlling the

Electron Paramag. Reson., 2007, 20, 131–156

133

growth conditions (e.g. source gas concentrations, pressure, input power, substrate temperature etc) it is possible to deposit diamond. If this is done onto a foreign substrate (heteroepitaxy) the result is polycrystalline or nanocrystalline diamond, with both forms having a wide variety of proven and potential applications.3,4 However, if the substrate is diamond then singlecrystal diamond can be grown by CVD. Several groups are researching into the heteroepitaxial growth of single crystal diamond, but to date the only route for the CVD growth of large single crystals of diamond is via homoepitaxy. Growth from the gas phase provides a route for controlled doping, even of impurities which do not have a high equilibrium solubility, as well as the synthesis of extremely pure, low-defect-density material.11 The nature of the process suggests that under appropriate synthesis conditions hydrogen may be a significant impurity. Paramagnetic hydrogen-related defects in diamond are discussed in Section 6. Nitrogen is readily incorporated in CVD diamond, and the defects resulting are discussed in Section 4 and 6. Boron is also readily incorporated as a substitutional impurity, introducing a relatively shallow acceptor level. A variety of unipolar diamond semiconductor devices, exploiting boron-doped p-type diamond and the high electrical breakdown strength and carrier mobilities of intrinsic diamond, are being considered which should offer unsurpassed performance in high-power or high-frequency applications.12 At high boron-doping levels diamond becomes metallic and one of the most exciting applications of this material is in electrochemistry. Boron-doped diamond electrodes are resistant to chemical attack, provide a wide potential window with low background currents and are therefore ﬁnding a variety of applications.13 Little is known about boron-related paramagnetic defects in diamond, and although this situation is likely to change in the next few years they will only be discussed brieﬂy in this article. It is not only the presence of defects and impurities which can inﬂuence the properties of materials, but also their distribution. EPR imaging (EPRI) in several different forms has successfully been used to investigate defect distributions in as grown and treated synthetic diamond, and the ﬁeld is reviewed in Section 7. Before embarking on a description of the highlighted areas where EPR has made a significant contribution to our understanding of defects and impurities in diamond, in Section 2 an attempt is made to assess the potential of data from EPR, electron nuclear double resonance (ENDOR), and pulsed EPR experiments to help in unravelling the mysteries of defects in diamond.

2

The Use of EPR and Related Techniques in the Study of Defects in Diamond

The principal information from EPR and related techniques which can be interpreted in terms of structural models of paramagnetic defects in solids is: (i) the anisotropy in the electronic g matrix, which is related to the nature of the wave function of the unpaired electrons, the crystal ﬁeld which acts upon them, and the symmetry of the site; (ii) the zero ﬁeld coupling in defects with S 4 1/2 also gives information about the unpaired electron wave function and defect

134

Electron Paramag. Reson., 2007, 20, 131–156

symmetry; (iii) the hyperﬁne interaction with the principal nucleus, if there is one, can facilitate chemical identiﬁcation, and together with hyperﬁne interaction of neighbouring nuclei indicates the delocalisation of the probability density of the unpaired electron and symmetry; (iv) the nuclear gN value is determined in ENDOR spectroscopy, unambiguously identifying the nucleus giving rise to the hyperﬁne splitting; (v) when I 4 1/2 the quadrupole interaction which relates to the symmetry of the site and the electric ﬁeld gradient at the nucleus set up by the distribution of all electrons in the defect; (vi) for defects with high symmetry, the value of the electron spin S cannot be unambiguously determined from EPR spectra alone, but the value can be resolved from ENDOR, or pulsed experiments; (vii) motional effects in EPR spectra arising from bond switching or atomic reorganisation provide insight into how the defect may be formed/decay. For example, for silicon the anisotropy of the electronic g-matrix is typically fairly large, so that symmetry and principal values are easily measured and may be interpreted in terms of the interaction with the surroundings. The 4.7% abundant isotope 29Si gives observable hyperﬁne lines to allow the mapping of the unpaired electron probability density over silicon atoms near the defect. The availability of samples tailor-made for spectroscopy (e.g. controlled dopant concentrations and isotopic composition) enabled unambiguous experimental studies. Furthermore, in most cases in silicon the Jahn-Teller splitting dominates over electron-electron coupling and a simple one-electron picture proves very powerful and reliable in the determination of defect ground states and hence the analysis of EPR spectra.14 For diamond the differentiation of defects has been much less easy. The much smaller spin-orbit coupling and much wider band gap for diamond compared with silicon makes the anisotropy in the electronic g-matrix about 25 times smaller in diamond than silicon. Even though EPR lines in diamond can be narrow (at low paramagnetic defect concentrations, in the absence of broadening from ferromagnetic inclusions and random strains) the spin-spin interaction between the unpaired electron and distant 1.1% abundant 13C nuclei produces an inhomogeneous broadening of the EPR linewidth of about 8 mT. This makes it much more difﬁcult to resolve even the symmetry-related copies of a single defect centre, let alone resolve different defect centres. Many centres have lines near the free electron g value, so this part of the spectrum can be very rich and confusing. Whether such lines can be resolved depends upon their width (typically much greater than 8 mT) and the microwave frequency at which the experiment is performed. It appears that a solution to the problem is to go to higher frequencies, however most defects in diamond have very long relaxation times and saturation problems typically become worse at higher frequencies, often precluding useful experimental investigation. However, the recent and continuing improvements in the performance of high frequency EPR spectrometers suggest that a significant contribution could yet be made in the exploration of the properties of defects in diamond. The g-matrix is rarely sufﬁciently different from the free electron value to interpret it in terms of interactions with the surroundings. The 1.1% abundance

Electron Paramag. Reson., 2007, 20, 131–156

135

of 13C (compared with the 4.7% of 29Si) often makes the hyperﬁne structure from natural abundance 13C difﬁcult to observe by EPR and ENDOR. Given that until the last decade of the 20th century most experiments were performed on natural diamond it is not surprising that our understanding of paramagnetic defects in diamond lagged a long way behind that of silicon. Little information was gleaned from 13C hyperﬁne studies, since the lines could rarely be unambiguously identiﬁed, whereas in silicon 29Si hyperﬁne structure was often crucial in defect structural determination. Hence until the 1990’s EPR contributed relatively little to the understanding of defects in diamond apart from producing catalogues of zero ﬁeld parameters from unknown defects with S Z 1. However, in the 1990’s great advances were made in diamond synthesis and in improving the sensitivity of EPR spectrometers. The latter made what was impossible 15 years ago now routine, and the former meant that samples became available which were tailor-made to enable the solution of some of the most taxing problems. These advances have opened up the ﬁeld, and in some areas our understanding of defects in diamond now exceeds that in silicon. It is important to note that the theoretical approaches developed to interpret the properties of paramagnetic defects in silicon cannot usually be used for their counterparts in diamond, although many people try! However, the recent experimental breakthroughs in studying defects in diamond have been accompanied by tremendous advances in the modelling of defects in diamond. The synergy between experiment and advanced theoretical models which can make robust predictions of spin Hamiltonian parameters for defects in diamond,15 is one of the reasons the ﬁeld is moving so quickly.

3

Intrinsic Defects in Diamond

Diamond is accepted to be a radiation-hard material. The strong covalent bonds mean that defects such as interstitials and vacancies are only produced by irradiation with relatively high energy particles. The lattice vacancy in diamond is one of the primary products of radiation damage. The early theoretical defect molecule calculations of Coulson and Kearsley for the vacancy in diamond have proved remarkably successful in predicting the properties of the neutral (V0) and negatively (V) charged vacancies in diamond.16 This is because they recognized that there are several possible ground states for the vacancy, and as a consequence electron-electron correlation effects are important. In the defect molecule calculation they assumed that the electronic properties are primarily determined by the electrons in the dangling orbitals surrounding the vacancy. The Td-point group requires that these orbitals transform as a1 and t2. For the neutral vacancy these orbitals are ﬁlled with four electrons in the conﬁgurations a12t22, a11t23, or t24 with the ﬁrst conﬁguration giving rise to the many-electron states 1A1, 1E, 3T1 and 1T2. Optical absorption measurements conﬁrmed that the ground state of V0 is 1E.17 EPR cannot be used to probe the diamagnetic ground state; however the a11t23 conﬁguration gives rise to (amongst others) the 5A2 many-electron state, which

136

Electron Paramag. Reson., 2007, 20, 131–156

following the predictions of Hund’s rules would be the ground state. Although the optical measurements prove that it is not, it should still be a low-lying state that could be populated via optical excitation. For the negatively charged vacancy the a1 and t2 states are ﬁlled with ﬁve electrons. The conﬁguration a12t23 gives rise to the many electron states 4A2, 2E, 4T1, 2T1 and 2T2, of which 4 A2 is predicted to be the ground state. The ﬁrst problem for the EPR spectroscopist is that the value of the unpaired electron spin S, which is related to the number of unpaired electrons in the defect, is not measurable from EPR alone for a site with Td symmetry. This is because D ¼ 0 (from the zero ﬁeld splitting term S D S in the spinHamiltonian) and hence for example for the 4A2 state of V0 with S ¼ 3/2 the MS ¼ 3/2 to 1/2, 1/2 to 1/2 and 1/2 to 3/2 transitions are coincident. However, the 13C ENDOR transitions for MS ¼ 3/2 and 3/2, occur at different frequencies to the transitions for MS ¼ 1/2 and 1/2, so that their presence indicates that S ¼ 3/2 rather than 1/2. A synthetic diamond enriched with 10% 13C, doped with the nitrogen donor (see Section 4) and irradiated with 2 MeV electrons to produce the vacancy in the negative charge state, was studied using ENDOR.18 V was identiﬁed; ENDOR conﬁrmed that the ground state had S ¼ 3/2 and analysis of the 13C hyperﬁne satellites conﬁrmed that the unpaired electron probability density is predominately localized in the carbon dangling orbitals and that the symmetry remains as Td.18 Furthermore, in later work it was proved that the 5A2 excited state of V0 could be populated by optical excitation. ENDOR conﬁrmed the identity of this state (since it is the only possible state of V0 with S ¼ 2), that again the unpaired electron probability density was shown to be predominately localized in the carbon dangling orbitals and that the symmetry remains as Td.19 The situation is somewhat different in silicon where the electronic structure of the different charge states of the isolated vacancy are successfully modelled using a oneelectron defect molecule (concentrating on the dangling bonds), where it is assumed that the electrons pair oﬀ whenever possible and the Jahn-Teller effect dominates over any corrections to the one electron model.20 V1 has not yet been conclusively identiﬁed in diamond, even though it should have a paramagnetic ground state. Theoretical predictions suggest that the ground state is susceptible to a Jahn Teller distortion lowering the symmetry from Td to D2d, and if this is dynamic the lines could be broadened beyond observation. This defect is likely to be the target of future work because of its possible importance in boron doped semiconducting diamond. The activation energy for neutral vacancy migration has been determined to be 2.3(1) eV.17 The high migration energy implies that high-temperature annealing (4900 K) is required for di-vacancy formation by aggregation. Mobile vacancies are readily trapped by nitrogen centres (see Section 4) and at even moderate nitrogen concentrations the probability of di-vacancy formation is dramatically reduced. EPR studies of high purity radiation damaged and annealed diamond enriched to 5% 13C resulted in the identiﬁcation of the nearest-neighbour di-vacancy centre. It was the isotopic enrichment, and consequent observation of 13C hyperﬁne lines, that permitted the structure to

Electron Paramag. Reson., 2007, 20, 131–156

137

be determined more than 30 years after the discovery of the centre, known as R4 or W6.21 There are several surprising features about this defect (e.g. temperature-dependent EPR linewidth, temperature-dependent zero ﬁeld splitting, etc) and this along with an investigation of larger vacancy cluster deserves further study. Diamond is the only elemental semiconductor in which the isolated selfinterstitial has been identiﬁed. EPR studies show that the electrically neutral self-interstitial, I0, occupies a split [001] conﬁguration with D2d symmetry,22 in agreement with early23,24 and recent25 theoretical predictions. The same local atomic conﬁguration is predicted, but not yet conclusively identiﬁed, for the interstitial carbon atom trapped at a carbon anti-site in silicon carbide.26,27 In the limit of low irradiation temperature, the production rate of I0 is closely equal to the production rate of the neutral vacancy, when pure diamonds are irradiated with MeV electrons.28 At temperatures greater than about 100 K, a radiation-enhanced migration of I0 occurs when self interstitials move with migration energy of 0.3 eV.28 Thermally activated migration of I0 occurs near 600–700 K, with an activation energy of 1.7 eV,29 so I0 is stable at room temperature. The EPR-active state of I0 has S ¼ 1, and lies about 50 meV above the ground state of this centre. The EPR spectrum is very interesting; the linewidth of the DMS ¼ 1 transitions is far too great to be due to the electronelectron spin-spin interaction, and the narrow DMS ¼ 2 lines observed for B close to [100] show that the broadening is not due to short relaxation times. The angular variations of the linewidths were found to be consistent with a speciﬁc distribution of values of the components of D about the mean value.22 Measurements of 13C hyperﬁne structure of samples grown with enriched isotopic abundance of 13C, and the EPR of samples annealed under uniaxial stress helped to conﬁrm the structure of I0.22 Studies of the structure of I0, through the use of uniaxial stress measurements and isotope-substitution effects on the optical absorption lines near 1685 and 1859 meV (which correlate with the EPR signal) have shown that this defect has many interesting optical properties.30 The data show that different ground and excited states involved in the optical transitions tunnel rapidly between equivalent symmetry conﬁgurations but that the motion is sufﬁciently rapid for I0 to have the observed effective D2d point group. It appears that pulsed EPR should be used to study the dynamics of the paramagnetic state of this defect. Upon migration, self interstitials are either annihilated at vacancies, trapped by impurities or form aggregates. EPR has been particularly effective at identifying the [001] di-interstitial31 and tri-interstitial complexes,32 which have S ¼ 1 ground states. First-principles theoretical methods have been used to investigate the self-interstitial and its aggregates in diamond. The calculated spin-spin interaction tensors are in good agreement with the experimental D tensors supporting the assignments of the microscopic models.25 Taking heart from the excellent agreement these authors went on to suggest a low-energy structure for a four-interstitial cluster and a generic model for an extended defect called the platelet, which has been widely accepted as the explanation to a problem which has vexed researchers for 30 years.25

138

Electron Paramag. Reson., 2007, 20, 131–156

So far the discussion has been restricted to simple isolated vacancies and interstitials and small aggregates of either. The literature is packed with the results of EPR (and optical) studies on diamond, but it is often difﬁcult to assess data when the samples are not well characterised, and in many cases it is not clear whether some defects are intrinsic or impurity related. Comprehensive catalogues of the paramagnetic defects observed in diamond are available.2,5 It is likely that work in this area will continue, with research focusing on the properties of larger aggregates of vacancies and interstitials, as well as into how intrinsic radiation damage defects inﬂuence the optical and electronic properties of diamond and interact with impurities.

4

Paramagnetic Nitrogen Defects in Diamond

Nitrogen is the most prevalent impurity in natural, HPHT and CVD diamond. In the latter two cases great care has to be taken in synthesis to exclude nitrogen from the diamond. The neutral single substitutional nitrogen centre (NS0) in diamond has been extensively studied by EPR33–35 and ENDOR36,37 and optical techniques. This defect is common to CVD and HPHT diamond and often observed in natural diamond, although in this case the bulk of the nitrogen is usually in the form of aggregates. In NS0 the nitrogen replaces a carbon atom on a substitutional site, forming four bonding orbitals with neighbouring carbon atoms, whilst the remaining unpaired electron goes into a highly localised anti-bonding orbital directed along one of the N–C directions. The spin Hamiltonian parameters are very accurately known, and simple theory indicates that the unpaired electron probability density is almost entirely localised in the unique N–C antibonding orbital.39 The extension of the unique N–C bond is large, around 25–30% longer than the normal C–C bond38 and the relaxation of the neighbouring carbon atoms is significant. Substitutional phosphorus in silicon remains on the Td site forming a shallow hydrogenic donor; unfortunately the oﬀ-site distortion (Td - C3v) of nitrogen in diamond lowers the energy of the donor level significantly (to approximately 1.7 eV below the conduction band) and hence NS0 is not electrically useful. However, it can act as a deep donor providing an electron to other defects (e.g. NS0 þ V0 - NS1 þ V). The distortion cannot be described as manifestation of the Jahn-Teller effect;40 it is in fact the bonding-antibonding occupancy which is responsible for the lowering of the symmetry. Above 600 K motional averaging starts to occur for the NS0 centre as the unpaired electron hops between the four possible N–C antibonding orbitals. At 1200 K the motional averaging is so fast that the EPR spectrum is isotropic. Measurements of the temperature range 600–1200 K show that the barrier to reorientation is 0.7(1) eV.41 At low temperature, reorientation of the centres occurs by tunnelling between thermally populated excited vibrational states,42 and in the absence of other defects it has been suggested that tunnelling processes in which MS is not conserved provides the dominant spin-lattice relaxation mechanism.41

Electron Paramag. Reson., 2007, 20, 131–156

139

In diamonds containing high concentrations of NS0 EPR spectra have been found in three well-separated regions.42 A family of lines around the EPR spectrum of the NS0 centre was attributed to three defect centres, named NOC1, NOC2, and NOC3 (NOC stands for the Novosibirsk-Oxford collaboration); each one of the centres corresponds to a coupled pair of NS0 atoms with different separations. The closest distance at which the pairs behave like S ¼ 1 is for the NS0 separated by 0.357 nm in NOC2. Another spectrum consisting of a family of weak lines at about one-half of the NS0 magnetic ﬁeld was also attributed to separated pairs of NS0 centres, but is a superposition of spectra from all pairs with separation greater than about 0.7 nm. The third spectrum consists of a family of weak lines at about twice the P1 magnetic ﬁeld and was attributed to isolated NS0 centres which were resonated by a two microwave photon transition.42 Although the NS0 centre has been studied by EPR for nearly 50 years, its spin physics deserves further research. As 14N (I ¼ 1) is nearly 100% abundant, its hyperﬁne structure is much easier to observe in EPR than that of 13C. However, in many centres involving nitrogen the unpaired electron probability density on the nitrogen is small, so the resulting hyperﬁne splitting is small and difﬁcult to resolve. In addition for many centres the 14N quadrupole interaction is large, and hence when the hyperﬁne interaction is also small the distinction between allowed and forbidden transitions is meaningless and the EPR spectrum may become too complicated to recognize without the aid of ENDOR or Electron Spin Echo Envelope Modulation. 14N ENDOR studies of a variety of nitrogen containing defects in natural diamond have been reported in the literature,43–49 and will not be discussed here. Transition metal nitrogen complexes will be discussed in Section 5. Nitrogen defects are effective at trapping migrating vacancies, and one of the most studied paramagnetic defects is the negatively charged nitrogen vacancy centre (NV) which consists of a substitutional nitrogen adjacent to a vacancy.50 This defect is formed in irradiated (to produce vacancies) and annealed (to allow the vacancies to migrate) NS0 doped diamond as follows: NS0 þ V0 NV0 and NV0 þ NS0 - NV þ NS1.17,51 Both NV and NV0 are found in as grown CVD diamond when the source gases are contaminated with nitrogen. It is uncertain if the NV defect is grown into the diamond lattice as a unit or formed via a vacancy diffusing to NS0 during the growth process. NV has been observed using a variety of experimental techniques, including EPR, optical absorption and photoluminescence (PL). NV0 has been observed using techniques such as optical absorption, cathodoluminescence and photoluminescence, but not EPR. The lack of observation of NV0 by EPR is an unresolved problem. This defect must be paramagnetic, but perhaps its signal is obscured by other EPR active defects. The NV defect attracts attention for use in quantum information processing, either as a qubit in a quantum computer or as a single photon source.52–55 The NV optical centre was ﬁrst discovered in 1965 by du Preez,56 where it was named as the 1.945 eV absorption line (637 nm). In 1976, Davies and Hamer57 performed uniaxial stress measurements on the 1.945 eV absorption band and showed that it originated from a transition between an A ground state and an E excited state of a trigonal centre. Loubser

140

Electron Paramag. Reson., 2007, 20, 131–156

and van Wyk went on to observe EPR from the NV centre.58 Two-laser hole burning measurements by Reddy, Manson and Krausz,59 conﬁrmed that the NV centre has a 3A ground state. Various measurements using spin coherence,60 cross-relaxation,61 Raman heterodyne,62–66 and four-wave mixing techniques67 followed. It has been determined from the nitrogen hyperﬁne parameters that there is negligible unpaired electron probability density on the nitrogen atom, and the unpaired electrons are instead localised on the three neighbouring carbon dangling orbitals. The NV is an excellent example of a system where the large quadrupole interaction makes any characteristic 14N hyperﬁne interaction difﬁcult to interpret. Figure 1 shows half-ﬁeld forbidden EPR transitions from NV produced in two different diamonds, one doped with 14N and the other with 15N. The two-line pattern from 15N is clear, but the 14 N spectra are very complicated because of the large quadrupole interaction. The ﬁts shown in Figure 1, were determined using gJ ¼ 2.0029(2), g> ¼ 2.0027(2), D ¼ 2873(8) MHz, 14N:AJ ¼ 2.17(5) MHz, 14N:A> ¼ 2.81(5) MHz, PJ ¼ 4.79(5) MHz, 15N:AJ ¼ þ3.15(5) MHz and 15N:A> ¼ þ3.85(5) MHz, which are discussed elsewhere.68 The electronic structure of the NV defect can be described by the procedure developed by Coulson and Kearsley.16 The single-electron molecular orbitals can be constructed from the dangling orbital on the nitrogen and the three carbon dangling orbitals, see Figure 2. Ignoring overlap of the orbitals, there are two levels of a1 symmetry (u ¼ fN and v ¼ [fa þ fb þ fc]/3) and a doubly degenerate e level (e ¼ [2 fa–fb–fc]/O6 and [fb–fc]/O2). It is assumed that the energies of the one-electron orbitals follow the sequence u o vo e, so that for NV the conﬁguration u2v2e2 gives rise to the states 1A1, 1E and 3A2, the latter being the ground state. For NV0 the conﬁguration u2v2e1 gives rise to an 2E ground state, which is susceptible to a Jahn-Teller distortion. For NV the small nitrogen hyperﬁne interaction and relatively large quadrupole interaction are predicted and observed for the 3A2 ground state, however the consequences of the dynamical behaviour of this centre (e.g. tunnelling of nitrogen into the vacancy) are poorly understood. Could it be that the Jahn-Teller splitting of the NV0 ground state is dynamic and this is broadening the EPR lines beyond observation? The hunt for the EPR signal from the NV0 centre is on!

5

Transition Metal Defects in Diamond

Point defects involving transition metals attract attention both because of their likely presence in HPHT diamond grown from transition metal solvent catalysts, and the possible exploitation of the optical properties of these centres. When investigating transition metal defects in diamond one needs to establish the element involved, the symmetry of the site, the electronic conﬁguration, the nature of the ligand nuclei and the unpaired spin probability density on them and the overall charge of the defect. This is not always straightforward. Common solvent catalysts include iron, nickel and cobalt, but we know of no evidence for the incorporation of iron so we will not consider this element

141

Electron Paramag. Reson., 2007, 20, 131–156

(a)

EXPT.

SIMULATION

164.8

165.0

165.2

165.4

165.6

Magnetic Field (mT) (b)

EXPT.

SIMULATION

164.6

164.8

165.0

165.2

165.4

165.6

165.8

Magnetic Field (mT) Figure 1 Half ﬁeld EPR spectra from the NV defect in (a) 14N doped diamond, and (b) 15 N doped diamond. Spectra recorded with magnetic ﬁeld parallel to the h111i crystallographic axis and a microwave frequency of approximately 9.75 GHz. Below each experimental spectrum is a simulated spectrum produced using the parameters given in the text. The relatively large 14N quadrupole interaction and the small hyperﬁne splitting results in the allowed and forbidden hyperﬁne transitions having comparable intensity giving rise to the complex 14NV EPR spectrum observed. On the other hand the 15NV EPR spectrum shows the expected characteristic simple two-line hyperﬁne pattern

further. Nickel and cobalt could be identiﬁed by characteristic hyperﬁne structure in EPR. In the case of nickel isotopic enrichment may be necessary to enable observation of the satellites, 61Ni (natural abundance 1.14%) is the only non-zero-spin stable isotope (I ¼ 3/2) and enrichment is not usually

142

Electron Paramag. Reson., 2007, 20, 131–156

C3v

c

b

a

N

Figure 2

Structure of the nitrogen-vacancy centre in diamond, showing the nearest neighbour atoms surrounding the vacancy. The bottom right atom is nitrogen, the remaining atoms are carbon. The trigonal (C3v) axis of the defect is indicated. The labelling of the nitrogen (fN) and carbon (fa, fb, fc) dangling orbitals is discussed in the text

practicable (we know of only one example69). Cobalt on the other hand is readily identiﬁable since 59Co (I ¼ 7/2) is 100% abundant. The EPR centre W8, has been shown beyond doubt to be substitutional nickel in the negative charge state (NiS: 3d7), the extra electron being donated by NS0. 61Ni hyperﬁne structure conﬁrms the identity69; 13C hyperﬁne structure from four nearest neighbours and twelve next-nearest neighbours shows that it is a substitutional site with Td symmetry70; the effective spin state of S ¼ 3/2 was determined from Fourier-transform nutational EPR spectroscopy, since this could not be determined from EPR for an ion in a Td site.72 In the substitutional site the tetrahedral crystal ﬁeld splits the 3d orbitals into the subsets e and t2, with e lower in energy, hence from 3d7 we get an orbitally nondegenerate 4A2 (from e4t23). To conﬁrm the existence of interstitial nickel using 13C hyperﬁne structure we would expect evidence of a site with four nearest neighbours and six nextnearest neighbours but no such pattern has so far been observed for any defect assigned to interstitial nickel. In diamonds grown from a nickel-containing solvent and boron- rather than nitrogen-doped, the negatively charged nickel defect is not expected. Three defects have been observed, but there is still a debate over the defect structures. The original interpretations of NIRIM-1 as interstitial nickel Nii1: 3d9 at a Td interstitial site71 and NIRIM-2 Nii1: 3d9 perturbed by a nearby defect such as a vacancy,71 has been questioned in subsequent analysis72 leading to the suggestion that as NIRIM-1 is a substitutional nickel defect (NiS1: 3d5); and that NIRIM-2 is probably Nii1: 3d9 with the nearest C–C bond directed away from the Ni atom replaced by C–B, with B further away from the nickel.72 It was suggested that the third defect labelled NOL1 (probably the same as NIRIM-573) is axially distorted Nii21: 3d8 with a B along h111i at an unspeciﬁed distance.74 The boron hyperﬁne

Electron Paramag. Reson., 2007, 20, 131–156

143

structure is very small suggesting little unpaired spin density on the boron. Reanalysis of this trigonal, boron-related defect has led to the suggestion that a more likely structure is NiS1: BS0 nearest neighbour conﬁguration with no connecting covalent bond.72 Calculations of the formation energy and mobility of the interstitial species suggest that nickel will be predominantly of the substitutional form, supporting the re-assignment of NIRIM-2 and NOL1 to substitutional nickel defects.75 Without new data it is unlikely that the correctness of the models can be proved. When diamonds containing substitutional nickel and high concentrations of nitrogen are annealed at high temperatures and pressures a series of nickel– nitrogen complexes NE1–NE3, NE5 and NE8 (NE centres) are produced.76–79 The g-matrix anisotropy suggests that these defects contain a transition metal. It has been proposed that these centres are formed upon annealing through the capture of mobile nitrogen atoms by the NE4 centre. This defect is considered as a template of the nickel–nitrogen complexes NE1–NE3, NE5, and NE8, which differ only in the number and in the positions of the nitrogen atoms in the coordination shell of nickel. NE4 is formed on HPHT treatment through the ejection of a self-interstitial by a substitutional NiS, resulting in the nickel semi-vacancy structure (e.g. nickel ion at the centre of a di-vacancy), with D3d symmetry. Recently it has been shown that the NE4 and AB1 are in fact the same defect.80 The series of nickel-related paramagnetic centres labelled AB1– AB7 have been found in as grown and annealed HPHT diamonds.81–83 These ABx centres are probably formed by a process different from the nitrogen aggregation to nickel, because no hyperﬁne splitting due to involved nitrogen has been resolved. AB1, AB3 and AB6 centres exhibit a transitory behaviour upon HPHT annealing and anneal out by 19001C.84 It is difﬁcult to speculate on the models for these defects as only the g-matrix is known. The solubility of cobalt in diamond appears to be much less than that of nickel, nevertheless cobalt-related defects have been identiﬁed. A defect labelled O4 was observed at low temperatures (o30 K) in HPHT synthetic diamonds grown with a cobalt containing metal-solvent catalyst after a high-temperature anneal at 18001C.85 The centre has S ¼ 1/2, monoclinic I symmetry and a hyperﬁne splitting from a 100% abundant nucleus with I ¼ 7/2. Analysis of the measured hyperﬁne interaction in terms of 3d5 gives the unpaired electron wave function on the cobalt atom, and correctly predicts the measured g matrix. Even though no nitrogen hyperﬁne splitting was resolved, analysis of the EPR linewidth gave support to the argument that O4 is a cobalt-nitrogen complex, with the cobalt in a semi-vacancy position (e.g. like nickel in the NEx centres). Other cobalt centres have subsequently been observed, and it appears that the discovery of a series of cobalt defects similar to the nickel NEx has begun.80

6

Paramagnetic Defects Incorporating Hydrogen in Diamond

Given the high concentration of hydrogen in the CVD growth environment, it is possible that hydrogen is readily incorporated into CVD diamond. It is worth

144

Electron Paramag. Reson., 2007, 20, 131–156

starting the discussion with reference to the H1 defect,86–88 observed in polycrystalline diamond ﬁlms grown by a variety of different CVD techniques. The H1 centre is sometimes called the ‘‘g ¼ 2.0028’’ defect, since this is its g-value, but it has been shown that other S ¼ 1/2 defects in polycrystalline CVD diamond also have g ¼ 2.0028.87,88 The incorporation of hydrogen in the H1 centre is deduced from satellites that can be observed when EPR measurements are made at microwave frequencies above about 8 GHz (at lower microwave frequencies the EPR lineshape is very different88). The position and intensity of the satellites as a function of microwave frequency suggests that they arise from the forbidden nuclear-spin ﬂip transitions of a hydrogen atom weakly coupled to the unpaired electron spin.89 Analysis of the EPR lineshape between 9.8 and 35 GHz led Zhou et al.86,87 to conclude that H1 was a single well-deﬁned defect consisting of an unpaired electron in a carbon dangling orbital coupled to a hydrogen atom about 0.2 nm away. The hydrogen hyperﬁne coupling parameters are AJ ¼ 27.5(2.5) MHz and A> ¼ 5.5(2.5) MHz.86,87 Zhou et al.87 proposed that H1 was formed by a hydrogen atom entering a stretched C–C bond at a grain boundary, allowing the carbons to relax back, one bonding to the hydrogen and the other with an unpaired electron primarily localised in its dangling bond. Pulsed and continuous-wave EPR measurements show that the H1 EPR centre has a much higher spin-lattice relaxation rate (4100 times) than the single substitutional nitrogen centre known to be incorporated into the diamond lattice. The concentration of H1 centres can be reduced by annealing above 1700 K.90 Annealing above 1700 K also produces changes in the CH infrared absorption bands seen in polycrystalline CVD diamond.90 1H ENDOR measurements on the H1 centre indicate that within 0.2–1.0 nm of the defect there is a significant concentration of hydrogen atoms. Although this strong 1H matrix ENDOR was observed, no ENDOR signal was detected from the near neighbour unique hydrogen identiﬁed by EPR.88 All the experimental evidence is consistent with H1 being located on hydrogen decorated grain boundaries or in inter-granular material rather than bulk diamond. H1 has been reported in what are referred to as single crystal CVD diamond ﬁlms; however it appears that this material always contains non-epitaxial crystallites and a defective nature. Hence H1 defects are likely to be located in the grain boundaries, voids etc. in the defective regions and not in the bulk single crystal. If such material is removed the H1 signal disappears. The nitrogen-vacancy-hydrogen complex (NVH) was identiﬁed in a free-standing nitrogen-doped isotopically engineered single crystal diamond synthesized by CVD.91 The hydrogen atom is located in the vacancy of a nearest-neighbour nitrogen-vacancy defect and at room temperature is on the C3v axis (at least this is the time average position) of the defect maintaining the trigonal symmetry of the centre. The defect is observed by EPR in the negative charge state in samples containing a suitable electron donor (e.g., substitutional nitrogen, NS0). The NVH complex is a very common defect in nitrogen-doped CVD diamond (often present at concentrations Z 1017 cm3), and does not start to anneal out until E16001C. Theoretical treatments92–94 suggest that the

Electron Paramag. Reson., 2007, 20, 131–156

145

hydrogen is tunnelling between the three carbon dangling orbitals pointing into the vacancy. More work needs to be done at low temperatures and with deuterium-doped samples to investigate the possible dynamical nature of this defect. NVH0 is an electron trap, and in some samples there is evidence for substantial concentrations of NVH0, hence it is important to investigate if this defect inﬂuences the electronic properties. In single-crystal diamond synthesized by CVD another defect was identiﬁed by EPR95 which has been attributed to both the negatively charged vacancy hydrogen defect (VH)95 and the negatively charged di-vacancy decorated with a hydrogen (V2H).94 This defect has only ever been observed at the sub-ppm level.96 The assignment to the former model is based on the fact that for both the nitrogen vacancy defect (Section 4) and the VH the electronic structure should be very similar (we can picture the C–H unit as analogous to the N-lonepair). This means that in both cases the electronic properties are determined by the electrons localised in the three carbon dangling orbitals. Hence we would expect NV and VH to have almost identical zero ﬁeld splitting tensors, as is found experimentally.95 However, it is argued that this model cannot correctly explain the very small hyperﬁne coupling, whereas the V2H model with dynamical tunnelling of the hydrogen gives reasonable agreement.94 However, an argument against the V2H model is that the limited supply of vacancies appears to restrict the formation of vacancy related defects. The concentration of NV and NVH is less than NS0/1 and although it would be stable at growth temperatures the undecorated di-vacancy has not been observed. The VH or V2H defect anneals out at around 12001C, with an activation energy appropriate for breaking a C–H bond.97 Given the importance of hydrogen in silicon, where it can both compensate useful donors and passivate unwanted traps, and the propensity for hydrogen to decorate vacancies, it seems reasonable to assume that there will be much more work on hydrogen related defects in diamond.

7

EPR Imaging of Paramagnetic Defect Distributions in Diamond

It is not only the presence of defects and impurities in diamond which are important but also their distribution. It is well known that the concentration of nitrogen, boron and transition metal impurities incorporated into different growth sectors of the same HPHT diamond depends on the nature of the growth sector, the growth temperature and the growth medium. Even within a growth sector the incorporation is not always homogeneous and can be dramatically affected by ﬂuctuations in the growth conditions. The morphology of HPHT synthetic diamond crystals depends on the growth conditions. At low growth temperatures {100} surfaces are dominant and at high temperatures {111} faces dominate, and a variety of other minor surfaces are often observed. Nitrogen in the HPHT growth cell is partitioned between the growing diamond and metal solvent/catalyst, the relative afﬁnity of the metal for nitrogen controls the incorporation in the diamond, and typically the nitrogen concentration is highest

146

Electron Paramag. Reson., 2007, 20, 131–156 (001)

-

(113)

B

(113)

(011)

(101)

A

-

(111)

(111)

(311)

(311)

(131)

-

(110) (100)

--

(111)

--

(110)

-

(311)

(311)

(010)

-

(111)

(131)

-

(101)

Figure 3

Schematic cubo-octahedral HPHT synthetic diamond. The crystallographic faces are identiﬁed using the standard notation. The dotted lines refer to the slices through the sample shown in Figure 4

in the {111} sectors, with progressively decreasing amounts found in {100}, {113} and {110} sectors.100 This is emphasised in Figure 3, which shows a stylistic cubooctahedral HPHT synthetic diamond, from which slices are cut, Figure 4, showing the inhomogeneous incorporation of nitrogen. There have been several examples where EPR imaging (EPRI) has been used to determine the distribution of nitrogen in HPHT diamond.99–101 Imaging of the nitrogen distribution in diamond is possible using infrared (IR) absorption, where the detection limit is of the order of a few ppm. Is EPRI competitive with IR microscopy for studying the distribution of NS0 in diamond? We estimate that the conventional detection sensitivity in an EPRI probe for NS0 is B7 1013 spins/mT. On ﬁrst inspection this may seem poor, but one has to remember that microwave power saturation is a problem (since even at room temperature T1 is long for NS0 in diamond) and spectra must be recorded with low microwave powers. Figure 5 shows a photograph of light transmission through a thin slice of HPHT synthetic diamond. The slice was taken approximately perpendicular to a h110i direction (see Figure 4a) and was 240 mm thick. Dark regions correspond to high nitrogen (strong absorption) and the clear regions to relatively pure diamond. The growth sectors can be identiﬁed by reference to Figure 4a. Figure 6 shows a 2D EPR image of the NS0 distribution in the thin plate of diamond shown in Figure 5. The image was obtained by taking 45 projections with a ﬁeld gradient of 1.3 mT/mm. For an EPR linewidth of 0.1 mT the spatial resolution is gradient-limited to approximately 80 mm. The image produced using deconvolution and back projection has a voxel size of 120 120 240 mm3, hence we estimate that the minimum detectable NS0 concentration/voxel is B12 ppm. In order to acquire the data for the image in Figure 6 in less than 24 hours, signal to

147

Electron Paramag. Reson., 2007, 20, 131–156

(a)

-

(b)

-

(113)

(113)

--

(011)

(111) (101)

--

(113)

(001)

-

(111)

-

(111)

(001)

(111)

(111)

(113)

(113)

-

-

(110)

(110)

-(111)

- (111)

(113) (011)

(101) (111)

Seed Figure 4

Thin slices through the cubo-octahedral HPHT synthetic diamond shown in Figure 3. In (a) the slice has been taken perpendicular to the [110] direction and in (b) through perpendicular to the [001] direction, at the positions shown in Figure 3. The shading of the growth sectors indicates the relative nitrogen incorporation, the darker the colour the greater the nitrogen incorporation (see text for details). In (a) the position of the seed crystal is shown. The larger HPHT crystal would be grown on this seed

noise was compromised somewhat and hence we estimated that the true detectable NS0 concentration/voxel is B25 ppm. Often there is a temptation to select a higher cut oﬀ than justiﬁable when presenting EPRI data to artiﬁcially enhance the image. Comparison of the image shown in Figure 6 with data obtained using infra-red imaging showed that the local concentration determined by EPR was accurate to 10%, but conﬁrmed that EPRI was less sensitive than IR imaging for studying the distribution of NS0. There is an underlying problem in EPRI of NS0 in diamond which has not been addressed by other workers. The EPR linewidth is a function of paramagnetic defect concentration for concentrations in excess of B5 ppm. This problem is particularly important when the sample contains regions with substantially different concentrations, for example a local concentration of 200 ppm will give rise to a linewidth of B0.2 mT, whereas that of 2 ppm corresponds to only B0.02 mT.102 Thus the EPR spectrum recorded with ﬁeld gradient oﬀ contains mixtures of EPR linewidths and cannot be used for deconvolution of the gradient on the spectrum. However, by using a sufﬁciently high ﬁeld modulation to broaden the narrow EPR lines, reasonable success can be obtained over a limited range of concentrations, as seen in Figure 6. If the paramagnetic defect concentration is less than B2 ppm the EPR linewidth is independent of concentration, and this problem does not arise. However the EPR signals are weak and signalto-noise becomes the limiting factor in obtaining high spatial resolution in a reasonable amount of time. Hence we have to search for ways of increasing the EPR signal-to-noise ratio, without resorting to excessive averaging. For defects which do not show prohibitive microwave power saturation, a solution to the sensitivity issue can be to go to higher microwave frequency. This was demonstrated by a study of the (R1) di-interstitial defect

148

Electron Paramag. Reson., 2007, 20, 131–156

Figure 5 Photograph of light transmission through a thin slice of HPHT synthetic diamond (before the sample was irradiated). The slice was taken approximately perpendicular to a h110i direction (see Figure 4a) and was 240 mm thick. The darker the growth sector the higher the concentration of nitrogen. The growth sectors can be identiﬁed by reference to Figure 4a

(see Section 3).103 It was believed that this defect was not produced in regions containing high concentrations of NS0, because the nitrogen trapped the migrating interstitials-preventing formation of the interstitial pairs. This was conﬁrmed by 35 GHz EPRI of the sample shown in Figure 6 post-production of the di-interstitial defect by high-energy electron irradiation at room temperature. The EPR image of the di-interstitial distribution is shown in Figure 6. This image was obtained by taking 45 projections with a ﬁeld gradient of 0.3 mT/mm. For an EPR linewidth of 0.03 mT the spatial resolution is gradient-limited to approximately 100 mm. The image produced using deconvolution and back projection has a voxel size of 150 150 240 mm3, hence we estimate that the minimum detectable di-interstitial defect concentration/voxel is B0.04 ppm given a probe sensitivity for this defect of approximately 4 1011 spins/mT. In this case the EPR spectra of the di-interstitial and NS0 defects do not overlap, so spectral-spatial imaging is straight forward. The image leaves no doubt that the di-interstitial is not produced in the high nitrogen regions. Furthermore, the concentration of di-interstitial produced is consistent with production only in the low nitrogen material.103 Another method for high sensitivity EPRI of the distribution of low concentrations of NS0 involves the utilization of adiabatic rapid passage (ARP), to

149

Electron Paramag. Reson., 2007, 20, 131–156

1.5

1.5

1.5

1.5

1.0

1.0

1.0

1.0

0.5

0.5

0.5

0.5

0.0

0.0

0.0

0.0

-0.5

-0.5

-0.5

-0.5

-1.0

-1.0

-1.0

-1.0

-1.5

-1.5

-1.5

-1.5

-2.0 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 Position (mm)

25

40

55

70

85

Concentration NS0 (ppm)

100

Position (mm)

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 2.0 2.0

Position (mm)

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 2.0 2.0

-2.0 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 Position (mm)

0.045 0.050 0.055 0.060 0.065 0.070

Concentration R1 (ppm)

Figure 6 EPR images of the NS0 (left) and di-interstitial (right) defect distributions in the thin slice of HPHT synthetic diamond shown in Figure 5. See the text for details of the data acquisition parameters. EPR imaging conclusively shows that the di-interstitial defect is only produced in the low nitrogen-containing regions of the sample

rapidly accumulate EPR spectra with a good signal-to-noise ratio.103 A comprehensive review of passage effects has been undertaken by Weger.104 ARP detection requires long spin-lattice relaxation times (T1) so that the EPR signal can be easily saturated. For the NS0 defect this is always the case so long as the paramagnetic defect density is low. ARP detection with ﬁeld modulation produces EPR signals with an absorption shape rather than the conventional ﬁrst derivative lineshape. The ARP signal is out-of-phase with that of an unsaturated EPR signal, the phase-shift depending on T1 and the modulation frequency. The critical experimental parameters are the microwave magnetic ﬁeld, the ﬁeld modulation frequency and amplitude, the EPR line width, the rate of ﬁeld sweep and of course T1, which is temperature dependent. With saturation, low modulation amplitude, and a rapid ﬁeld sweep, ARP EPR signals can routinely be observed from NS0 defects when the concentration is less than a few ppm. ARP of the NS0 EPR under conditions of strong saturation produces a much greater signal-to-noise ratio than conventional slow passage (SP) EPR. It is well known that small changes of nitrogen concentration in the gas phase can cause considerable change in the growth rate, morphology, and texture of CVD-deposited diamond ﬁlms. The relative growth rates of different crystal facets, depends markedly on the precise growth conditions. As a polycrystalline ﬁlm grows the faster growing crystals overgrow the slower ones and ﬁlm texture evolves. Since the NS0 incorporation can vary significantly between different growth sectors in CVD (e.g. nitrogen incorporation into {111} growth sectors

150

Electron Paramag. Reson., 2007, 20, 131–156

was shown to be a factor of 3–4 higher than in the {100} growth sectors in a speciﬁc CVD study where no other parameters were changed105) as well as HPHT synthesis, it is hardly surprising that it was demonstrated by conventional EPRI that the NS0 incorporation often varies with thickness of a polycrystalline ﬁlm of CVD diamond.106 Figure 7 shows two 1D EPR images (obtained with a ﬁeld gradient of 1.2 mT/mm enabling a 1D spatial resolution of 20 mm) through a polycrystalline ﬁlm parallel to the growth direction (e.g. nucleation surface on left), one recorded using conventional SP EPR and the other using ARP. The distributions obtained in the two different ways are identical to within experimental error showing that ARP is a viable technique for measuring NS0 distributions in lightly doped CVD diamond samples. Note that the data acquisition time for the gradient-on and gradient-oﬀ ARP spectra was B20 minutes, while the equivalent SP spectra (with the same signal to noise ratio) took almost B34 hours to acquire: a significant time saving of a factor of B100. To acquire 1D images through samples requires that the sample be precisely oriented with respect to the gradient qBz/qz and static ﬁeld Bz, since misalignment of the sample will broaden the image. The gradual increase in NS0 concentration at the beginning of the image in Figure 7 suggests that the sample is slightly misaligned. The time saving in ARP imaging allows for the orientation to be changed iteratively until the narrowest image and hence true distribution is found. In the example shown in Figure 7, where the NS0 concentration decreases smoothly the true distribution can be extracted from the misaligned image by ﬁtting the data to a simple functional form and including misalignment. This method produced (when a misalignment of 2.21 was included) the NS0 distribution shown in inset to Figure 7, which (within errors) was identical to that obtained from a correctly aligned sample. EPRI provides a detection sensitivity and spatial resolution for the variation of NS0 through a CVD diamond ﬁlm (1D image) which is unsurpassed by any other non destructive technique. A combination of ARP and SP imaging has been used for spectral spatial imaging.103 The H1 defect (see Section 6) often observed in polycrystalline CVD diamond ﬁlms does not saturate easily90 and can only be observed in SP. At high incident microwave powers, the SP NS0 EPR signal is so strongly saturated that it is often not observable. Under these conditions H1 EPR dominates the spectra observed from polycrystalline CVD diamond ﬁlms. However, ARP conditions suppress the H1 EPR signal very effectively so that only the NS0 EPR signal is observed. Hence, by recording spectra and images under two sets of experimental conditions the NS0 and H1 distributions can be individually recorded, even though the EPR spectra overlap. An example image is shown in Figure 8, where the distributions of H1 and NS0 through a polycrystalline CVD diamond ﬁlm have been determined by SP and ARP imaging, respectively. The nucleation face is on the left of the image. This ﬁlm was initially grown at a high growth rate (estimated 30 mm/hour) from a mixture of CH4 and H2, but part way through the run a small quantity of oxygen was added to the source gases and the growth-rate dropped (estimated

151

NS0 Concentration (ppm)

Electron Paramag. Reson., 2007, 20, 131–156 1.2

NS0 Concentration (ppm)

1.0

0.8

1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 0

200

400

600

800

1000

Distance from nucleation face (µm)

0.6

0.4

0.2

0.0 0

200

400

600

800

1000

Distance from nucleation face (µm) Figure 7 EPR images of the NS0 concentration though a plate of polycrystalline CVD diamond recorded using conventional slow passage EPRI (black curve) and ARP EPRI (lighter curve). The 1D image was taken with the ﬁeld gradient approximately parallel to the growth direction. The uncertainties in the concentration determined at any distance from the nucleation face are indicated by the error bars. Inset is the calculated NS0 distribution with the ﬁeld gradient exactly parallel to the growth direction (the misalignment in the original image was approximately 2 degrees – see text for further details)

2 mm/hour). It can be seen that the addition of oxygen had a dramatic effect on the H1 defect concentration which dropped to near zero, while at the same time the ﬁlm colour improved dramatically (black to colourless). This is consistent with H1 being incorporated in non diamond inter-granular material which is not present at lower growth rates when oxygen is added to the source gases. The NS0 concentration is constant after addition of the oxygen, and the same as the maximum value during the early growth. The reason for the low initial incorporation of NS0 and the reduction in incorporation during the transition period (from growth without to with oxygen) is unclear, but may relate to morphology and structural changes. EPRI is a useful tool for studying defect incorporation in single crystal and polycrystalline diamond. Spatial resolution is often limited by spin sensitivity, not the available ﬁeld gradients because the EPR lines in diamond are typically very narrow. The variation of EPR linewidths with defect concentration is a problem. The technique of EPRI would be used more widely if it were possible to acquire the images more quickly. To this end approaches and instrumentation developing for EPR microscopy utilising modulated magnetic ﬁeld

152

Electron Paramag. Reson., 2007, 20, 131–156

P1 or NS0 H1

Defect Concentration (ppm)

30

25

20

15

10

5

0 -400

-200

0

200

400

600

800

1000

1200

1400

Distance from nucleation face (µm) Figure 8 EPR images of the NS0 and H1 defect concentrations though a plate of polycrystalline CVD diamond recorded using conventional slow passage EPRI (H1 black curve) and ARP EPRI (NS0 grey curve). The 1D images were taken with the ﬁeld gradient parallel to the growth direction. At around 300 mm into the ﬁlm the growth conditions were changed, see text for further details

gradients appear to provide the sensitivity and speed required to make EPRI practicable for routine studies of defect distributions in diamond.107

8

Conclusions and Further Work

In the last decade EPR and related techniques have made a significant contribution to our understanding of the incorporation and nature of defects and impurities in diamond. It is unsurprising that this progress has occurred at the same time that there have been great advances in the techniques of diamond synthesis (especially CVD). EPR has helped researchers start to build a sound understanding of defect physics in diamond, and more is known about speciﬁc intrinsic defects in diamond (e.g. self-interstitial) than the counterparts in much more extensively studied materials (e.g. silicon). The advances in EPR technology, which have made many of the recent studies possible, provide a platform from which progress in answering many out-standing issues could be made. There is currently considerable interest in the exploitation of the outstanding properties of synthetic diamond in a variety of technological applications. It appears to the author that those that are most likely to be

Electron Paramag. Reson., 2007, 20, 131–156

153

successful exploit several, not just one, of the exceptional properties. This naturally suggests that research efforts which combine different techniques to study combinations of properties (e.g. optical detected magnetic resonance, electrically detected magnetic resonance, etc) are likely to be the most useful. Future research should not be restricted to bulk defects. Given modern EPR technology it seems that paramagnetic surface species on diamond could soon be studied. References 1. J.E. Field, Properties of Natural and Synthetic Diamond, Academic Press, UK, 1992. 2. M.H. Nazare and A.J. Neves, Properties, Growth and Applications of Diamond (EMIS Datareviews Series 26), INSPEC, UK, 2000. 3. C. Nebel and J. Ristein, Thin-Film Diamond I, Volume 76, First Edition: (part of the Semiconductors and Semimetals Series), Academic Press, UK, 2004. 4. C. Nebel, Thin-Film Diamond II, Volume 76, First Edition: (part of the Semiconductors and Semimetals Series), Academic Press, UK, 2004. 5. C.A.J. Ammerlaan, in Landolt-Bornstein Numerical Data and Functional Relationships in Science and Technology, New Series 111, O. Madelung and M. Schulz, (ed.), Spinger, Berlin, 1990, vol. 22b, p. 365–381. 6. J.A. Weil, Phys. Chem. Minerals, 1984, 10, 149. 7. J.A. Weil, in Physics and Chemistry of SiO2 and the Si-SiO2 Interface, C.R. Helms, D.E. Deal, (ed), Plenum, New York, 1993, vol. 2, p. 131–144. 8. C.A.J. Ammerlaan, in Landolt-Bornstein Numerical Data and Functional Relationships in Science and Technology New Series 111, O. Madelung and M. Schulz, (ed), Spinger, Berlin, 1990, vol. 22b, p. 117–206. 9. J.-M. Spaeth, J.R. Niklas and R.H. Bartram, Structural Analysis of Point Defects in Solids - An Introduction to Multiple Magnetic Resonance Spectroscopy, Springer, Berlin, 1994. 10. W. Hayes, Crystals with the Fluorite Structure, Oxford University Press, UK, 1974. 11. J. Isberg J. Hammersberg, E. Johansson, T. Wikstro¨m, D.J. Twitchen, A.J. Whitehead, S.E. Coe and G.A. Scarsbrook, Science, 2004, 297, 1670. 12. D.J. Twitchen, A.J. Whitehead, S.E. Coe, J. Isberg, J. Hammersberg, T. Wikstro¨m and E. Johansson, IEEE Transactions on Electron Devices, 2004, 51, 826. 13. A. Fujishima, Y. Einaga, T. N. Rao and D. A. Tryk, Diamond Electrochemistry, Elsevier Science, Amsterdam, 2005. 14. G.D. Watkins, Physics of the Solid State, 1999, 41, 746. 15. M.J. Rayson, J.P. Goss and P.R. Briddon, Physica B - Condensed Matter, 2003, 340, 673. 16. C.A. Coulson and M.J. Keresley, Proc. R. Soc. London, Ser. A., 1957, 241, 433. 17. G. Davies, S.C. Lawson, A.T. Collins, A. Mainwood and S.J. Sharp, Phys. Rev. B, 1992, 46, 13157. 18. J. Isoya, H. Kanda, Y. Uchida, S.C. Lawson, S. Yamasaki, H. Itoh and Y. Morita, Phys. Rev. B., 1992, 45, 1436. 19. J.A. van Wyk, O.D. Tucker, M.E. Newton, J.M. Baker, G.S. Woods and P. Spear, Phys. Rev. B, 1995, 52, 12657. 20. G.D. Watkins, in Defects and Their Structure in Non-Metallic Solids, B. Henderson and A.E. Hughes, (ed), Plenum, New York, 1976.

154

Electron Paramag. Reson., 2007, 20, 131–156

21. D.J. Twitchen, M.E. Newton, J.M. Baker, T.R. Anthony and W.F. Banholzer, Phys. Rev. B., 1999, 59, 12900. 22. D.C. Hunt, D.J. Twitchen, M.E. Newton, J.M. Baker, T.R. Anthony, W.F. Banholzer and S.S. Vagarali, Phys. Rev. B, 2000, 62, 3863. 23. C. Weigel, D. Peak, J.W. Corbett, G.D. Watkins and R.P. Messmer, Phys. Rev. B, 1973, 8, 2906. 24. A. Mainwood, F.P. Larkins and A.M. Stoneham, Solid-State Electron., 1978, 21, 1431. 25. J.P. Goss, B.J. Coomer, R. Jones, T.D. Shaw, P.R. Briddon, M. Rayson and S. O¨berg, Phys. Rev. B, 2001, 63, 195208. 26. A. Mattausch, M. Bockstedte and O. Pankratov, Mater. Sci. Forum, 2002, 389–393, 481. 27. A. Gali, P. Deak, N.T. Son and E. Janzen, Mater. Sci. Forum, 2002, 389–393, 477. 28. M.E. Newton, B.A. Campbell, D.J. Twitchen, J.M. Baker and T.R. Anthony, Diamond Relat. Mater., 2002, 11, 618. 29. L. Allers, A.T. Collins and J. Hiscock, Diamond Relat. Mater., 1998, 7, 228. 30. H.E. Smith, G. Davies, M.E. Newton and H. Kanda, Phys. Rev. B, 2004, 69, 045203. 31. D.J. Twitchen, M.E. Newton, J.M. Baker, O.D. Tucker, T.R. Anthony and W. F. Banholzer, Phys. Rev. B, 1996, 54, 6988. 32. D.C. Hunt, D.J. Twitchen, M.E. Newton, J.M. Baker, J. K. Kirui, J.A. van Wyk, T.R. Anthony and W.F. Banholzer, Phys. Rev. B, 2000, 62, 6587. 33. W.V. Smith, P.P. Sorokin, I.L. Gelles and G.I. Lasher, Phys. Rev., 1959, 115, 1546. 34. R.C. Barklie and J. Guven, J. Phys. C: Solid State Phys., 1981, 14, 3621. 35. J.H.N. Loubser and L. du Preez, Br. J. Appl. Phys., 1965, 16, 457. 36. R.J. Cook and D.H. Whiﬀen, Proc. Roy. Soc. A, 1966, 295, 99. 37. A. Cox, M.E. Newton and J.M. Baker, J. Phys.: Condens. Matter, 1994, 6, 551. 38. S.A. Kajihara, A. Antonelli, J. Bernholc and R. Car, Phys. Rev. Lett., 1991, 66, 2010; P.R. Briddon, M.I. Heggie, R. Jones, in Proc. 2nd Int. Conf. on New Diamond Science and Technology, ed. R. Messier, T.J. Glass, J.E. Butler and R. Roy, 1991, Pittsburg: Materials Research Society, p. 63; P.R. Briddon and R. Jones, Physica B, 1993, 185, 179. 39. L.A. Shul’man, I.M. Zaritskii and G.A. Podzyarei, Sov. Phys. Solid State, 1967, 8, 1029; J.H.N. Loubser, W.P. Van Ryneveld, Br. J. Appl. Phys., 1967, 18, 1029. 40. C.A. Ammerlaan and E.A. Burgemeister, Phys. Rev. Lett., 1981, 47, 954. 41. J.A. van Wyk, E.C. Reynhardt, G.L. High and I. Kiﬂawi, J. Phys. D: Appl. Phys., 1997, 30, 1790; E.C. Reynhardt, G.L. High, J.A. van Wyk, J. Chem. Phys., 1998, 109, 8471. 42. V.A. Nadolinny, A.P. Yelisseyev, J.M. Baker, D.J. Twitchen, M.E. Newton, A. Hofstaetter and B. Feigelson, Phys. Rev. B, 1999, 60, 5392. 43. O.D. Tucker, M.E. Newton and J.M. Baker, Phys. Rev. B, 1994, 50, 15586. 44. J.A. van Wyk and J.H.N. Loubser, J. Phys.: Condens. Matter, 1993, 5, 3019. 45. J.A. van Wyk, J.H.N. Loubser, M.E. Newton and J.M. Baker, J. Phys.: Condens. Matter, 1992, 4, 2651. 46. A. Cox, M.E. Newton and J.M. Baker, J. Phys.: Condens. Matter, 1992 4, 8119. 47. A. Cox, M.E. Newton and J.M. Baker, J. Phys.: Condens. Matter, 1991, 3, 3591. 48. M.E. Newton and J.M. Baker, J. Phys.: Condens. Matter, 1989, 1, 9801. 49. M.E. Newton and J.M. Baker, J. Phys.: Condens. Matter, 1989, 1, 10549.

Electron Paramag. Reson., 2007, 20, 131–156

155

50. J.P. Goss, R. Jones, P.R. Briddon, G. Davies, A.T. Collins, A. Mainwood, J.A. van Wyk, J.M. Baker, M.E. Newton, A.M. Stoneham and S.C. Lawson, Phys. Rev. B, 1997, 56, 16031. 51. F. Jelezko, C. Tietz, A. Gruber, I. Popa, A. Nizovtsev, S. Kilin and J. Wrachtrup, Single Mol., 2001, 2–4, 255. 52. A. Beveratos, R. Brouri, T. Gacoin, A. Villing, J.P. Poizat and P. Grangier, Phys. Rev. Lett., 2002, 89, 187901. 53. A. Drabenstedt, L. Fleury, C. Tietz, F. Jelezko, S. Kilin, A. Nizovtzev and J. Wratchup, Phys. Rev. B, 1999, 60, 11503. 54. S. Kuhn, C. Hettich, C. Schmitt, J.-Ph. Poizat and V. Sandoghdar, Jnl. Micros., 2000, 202, 2. 55. F.T. Charnock and T.A. Kennedy, Phys. Rev. B, 2001, 64, 041201. 56. L. du Preez, PhD Thesis, University of Witwatersrand, South Africa, 1965. 57. G. Davies and M.F. Hamer, Proc. R. Soc. A., 1976, 348, 285. 58. J.H.N. Loubser and J.A. van Wyk, Diam. Res. Suppl. to Int. Diamond Rev., 1977, 11, 11. 59. N.R.S. Reddy, N.B. Manson and E.R. Krausz, J. Lumin., 1987, 38, 46. 60. E. Van Oort, N.B. Manson and M. Glasbeek, J. Phys. C: Solid State Phys., 1988, 21, 4385. 61. E. Van Oort and M. Glasbeek, Phys. Rev. B, 1989, 40, 6509. 62. K. Holliday, X. F. He, P.T.H. Fisk and N.B. Manson, Opt. Lett., 1990, 15, 983. 63. N.B. Manson, X.F. He and P.T.H. Fisk, Opt. Lett., 1990, 15, 1094. 64. N.B. Manson, X.F. He and P.T.H. Fisk, J. Lumin., 1992, 53, 49. 65. X.F. He, P.T.H. Fisk and N.B. Manson, J. Appl. Phys., 1992, 72, 211. 66. X.F. He, N.B. Manson and P.T.H. Fisk, Phys. Rev. B, 1993, 47, 8809. 67. D.A. Redman, S. Brown, R.H. Sands and S.C. Rand, Phys. Rev. Lett., 1991, 67, 3420. 68. C. Glover and M.E. Newton, manuscript submitted to Phys. Rev. B. 69. M.I. Samoilovich, G.N. Bezrukov and V.P. Butuzov, JETP Lett., 1971 14, 379. 70. J. Isoya, H. Kanda, J.R. Norris, J. Tang and M.K. Bowman, Phys. Rev. B, 1990, 41, 3905. 71. J. Isoya, H. Kanda and Y. Uchida, Phys. Rev. B, 1990, 42, 9843. 72. J.M. Baker, J. Phys.: Condens. Matter, 2003, 15, S2929. 73. J. Isoya J, H. Kanda and Y. Morita, in Advances in New Diamond Science and Technology, Tokyob MYU 1994, p. 351. 74. V.A. Nadolinny, J.M. Baker, M.E. Newton and H. Kanda, Diamond Relat. Mater., 2002, 11, 627. 75. J.P. Goss, P.R. Briddon, R. Jones and S. O¨berg, J. Phys.: Condens. Matter, 2004, 16, 4567. 76. V.A. Nadolinny and A.P. Yelisseyev, Diamond Relat. Mater., 1994, 3, 17. 77. V.A. Nadolinny and A.P. Yelisseyev, Diamond Relat. Mater., 1994, 3, 1196. 78. V.A. Nadolinny, A.P. Yelisseyev, J.M. Baker, M.E. Newton, D.J. Twitchen, S.C. Lawson, O.P. Yuryeva and B.N. Feigelson, J. Phys.: Condens. Matter, 1999, 11, 7357. 79. V.A. Nadolinny, J.M. Baker, O.P. Yuryeva, M.E. Newton, D.J. Twitchen and Y.N. Palyanov, Appl. Magn. Reson., 2005, 28, 365. 80. W. Gehlhoﬀ and R.N. Pereira, J. Phys.: Condens. Matter, 2002, 14, 13751. 81. A.J. Neves, R. Pereira, N.A. Sobolev, M.H. Nazare, W. Gehlhoﬀ, A. Naser and H. Kanda, Diamond Relat. Mater., 2000, 9, 1057.

156

Electron Paramag. Reson., 2007, 20, 131–156

82. A.J. Neves, R. Pereira, N.A. Sobolev, M.H. Nazare, W. Gehlhoﬀ, A. Naser and H. Kanda, Physica B, 1999, 273–274, 651. 83. R.N. Pereira, A.J. Neves, W. Gehlhoﬀ, N.A. Sobolev, L. Rino and H. Kanda, Diamond Relat. Mater., 2002, 11, 623. 84. R.N. Pereira, W. Gehlhoﬀ, A.J. Neves, N.A. Sobolev, L. Rino and H. Kanda, J. Phys.: Condens. Matter, 2003, 15, S2941. 85. D.J. Twitchen, J.M. Baker, M.E. Newton and K. Johnston, Phys. Rev. B, 2000, 61, 9. 86. X. Zhou, G. Watkins and K.M. McNamara-Rutledge, Mater. Sci. Forum, 1996, 196–201, 825. 87. X. Zhou, G. Watkins, K. M. McNamara-Rutledge, R. P. Messmer and S. Chawla, Phys. Rev. B, 1996, 54, 7881. 88. D.F. Talbot-Ponsonby, M.E. Newton, J.M. Baker, G.A. Scarsbrook, R.S. Sussmann, A.J. Whitehead and S. Pfenninger, Phys. Rev. B, 1998, 57, 2264. 89. S.L. Holder, L.G. Rowan and J.J. Krebs, Appl. Phys. Lett., 1994, 64, 1094. 90. D.F. Talbot-Ponsonby, M.E. Newton, J.M. Baker, G.A. Scarsbrook, R.S. Sussmann and A.J. Whitehead, Phys. Rev. B, 1998, 57, 2302. 91. C. Glover, M.E. Newton, P.M. Martineau, D.J. Twitchen and J.M. Baker, Phys. Rev. Lett., 2003, 90, 185507. 92. A. Kerridge, A.H. Harker and A.M. Stoneham, J. Phys.: Condens. Matter, 2004, 16, 87431. 93. J.P. Goss, P.R. Briddon, R. Jones and S. Sque, J. Phys.: Condens. Matter, 2003, 15, S2903. 94. M.J. Shaw, P.R. Briddon, J.P. Goss, M.J. Rayson, A. Kerridge, A.H. Harker and A.M. Stoneham, Phys. Rev. Lett., 2005, 95, 105502. 95. C. Glover, M.E. Newton, P.M. Martineau, S. Quinn and D.J. Twitchen, Phys. Rev. Lett., 2004, 92, 135502. 96. ppm: parts per million carbon atoms. 97. R.J. Cruddace, M.E. Newton, P.M. Martineau and D.J. Twitchen, submitted Phys. Rev. B. 98. R.C. Burns, V. Cvetkovic, C.N. Dodge, D.J.F. Evans, M.T. Rooney, P.M. Spear and C.M. Welbourn, J. Crystal Growth, 1990, 104, 257. 99. M. Furusawa and M. Ikeya, J. Phys. Soc. Japan, 1990, 59, 2340. 100. M. Ikeya, Annual Review of Materials Science, 1991, 21, 45. 101. G.A. Watt, M.E. Newton and J.M. Baker, Diamond Relat. Mater., 2001, 10, 1681. 102. J.A. van Wyk, E.C. Reynhardt, G.L. High and I. Kiﬂawi, J. Phys. D: Appl. Phys., 1997, 30, 1790. 103. G.A. Watt, PhD Thesis, Univeristy of Oxford, 2002. 104. M. Weger, Bell Syst. Tech. J., 1960, 1013. 105. R. Samlenski, C. Haug, R. Brenn, C. Wild, R. Locher and D. Koidl, Appl. Phys. Lett., 1995, 67, 2798. 106. D.F. Talbot-Ponsonby, M.E. Newton and J.M. Baker, J. Appl. Phys., 1997, 82, 1201. 107. A. Blank, C.R. Dunnam, P.P. Borbat and J.H. Freed, J. Magn. Reson., 2003, 165, 116.

EPR of Exchange-Coupled Oligomers BY DAVID COLLISON AND ERIC J.L. McINNES School of Chemistry, The University of Manchester, Manchester M13 9PL, UK

1

Introduction

This review is an update of our previous SPR reviews on magnetically exchange-coupled oligomers,1,2 and covers the literature in the two calendar years 2004 and 2005. As before we review the publications involving EPR spectroscopy of discrete molecular compounds containing more than one radical centre – polymeric materials are not covered. The discussion is largely restricted to compounds where 3 or more radical centres are present, but also include selected dimeric systems where we feel that these are of particular interest. The discussion is organised into coupled (i) p-block radicals, (ii) d-block radicals, (iii) mixed p/d-block radicals, (iv) mixed d/f-block radicals and (v) biological systems.

2

p-Block

This area is dominated by work on nitroxide species, driven by efforts towards fundamental understanding of magnetic exchange interactions in simple coupled S ¼ 12 systems and ultimately towards use as components in magnetic materials. We should note here that a major area and application of the study of interactions between nitroxide species is in distance measurements by PELDOR/DEER methods – in particular as applied to labelled biological systems – this is not covered in this review article. Baumgarten and coworkers reported the pyrazolylbipyridine-bridged bis (nitronyl)nitroxide 1 and bis(iminonitroxide) 2, and compare them to their terpyridine (terpy)-bridged analogues.3 They describe these as ‘‘multifunctional’’ biradicals because of the potential metal chelating sites. For 1 and 2 the ﬂuid toluene solution EPR spectra are characteristic of the strong exchange limit (J c aN; J ¼ isotropic exchange coupling constant, aN ¼ isotropic hyperﬁne interaction to 14N) giving equally spaced 9- and 13-line spectra, respectively. Spin-triplet spectra are observed from frozen solutions (120 K), Electron Paramagnetic Resonance, Volume 20 r The Royal Society of Chemistry, 2007 157

158

Electron Paramag. Reson., 2007, 20, 157–191

and simulation gives the axial zero-ﬁeld splitting (ZFS) parameter |D| as ca. 38 103 cm1. Fitting the temperature dependence of the intensity of the ‘‘halfﬁeld’’ DMS ¼ 2 transitions to the Bleaney-Bowers equation gives the singlettriplet energy gaps (2J) as ca. 13 cm1 (ferromagnetic) for 1 and 0.06 r 2|J| r 0.7 cm1 for 2. These J-values are lower than those in the terpy analogues whilst the D-values are larger. This is consistent with the longer inter-radical distance in the terpy-bridged systems. The increase in J is concluded to be due to the two different spin polarisation pathways available in 1 and 2 due to the 5membered heterocyclic rings in the bridges. Komaguchi et al. have investigated the efﬁciency of –(SiMe2)- based bridging for communication between two TEMPO radicals.4 They studied the ﬂuid solution EPR spectra of TEMPO(SiMe2)n-TEMPO (3) with n ¼ 1 to 4. With n ¼ 1 the strong exchange limit is observed, in contrast to single atom –C(¼O)– or –S- links, from which the authors conclude that the Si atom enhances the through-bond exchange interaction. For n ¼ 4 a monomer-like spectrum is observed (weak exchange limit). The spectra for n ¼ 3 are temperature dependent, changing between (crudely) slow and rapid exchange on increasing temperature. They ascribe this to enhanced through-space intramolecular interactions due to the conformational freedom of the longer-chain links, and estimate activation barriers for this process (rotation about the Si–Si bonds) from variable-temperature line-shape analysis.

N

N

N N

R O

R

N N R=

1

O O N

2 N

A very strong antiferromagnetic exchange interaction has been observed in the ethylene-bridged biradical 4.5 Spin triplet spectra are observed from 4 isolated in a polystyrene matrix, with |D|/gmB ¼ 168 G (mB ¼ the electronic Bohr magneton), and the intensity as a function of temperature gives a singlettriplet gap of 479 K. The authors note that analysis of D in a point-dipole approximation is not valid due to the strong exchange. Stroh et al. have shown that the exchange interaction between nitronyl nitroxides can be mediated by a diamagnetic Pt(II) fragment (complex 5 and its isomer with meta-substituted aryl linkers).6 Strong exchange limit spectra are observed in ﬂuid solution, although hyperﬁne coupling is not observed to 195Pt or 31P. There is a significant difference between the isotropic g-values for the ortho- and meta-substituted species (2.0094 and 2.0066, respectively) which the authors tentatively ascribe to different orbital overlap between the p-radical and metal d-orbitals.

159

Electron Paramag. Reson., 2007, 20, 157–191

The upper limit of |J| is estimated as 1 cm1. O O

N

N N N

O

O

4 O

O PPh3

N

N

Pt N

N

PPh3

O

O

5

Takui et al. have continued their elegant studies of covalently bonded biradical-monoradical composites with the triradical 6.7 Here they synthesised strongly ferromagnetically coupled dimeric nitronyl-nitroxide fragments (via a meta-aryl linker, with exchange interaction J) which are then appended to a further nitronyl nitroxide via a s-bound link which is expected to give rise to only weak exchange between the dimer and monomer (J 0 , J00 ), thereby isolating S ¼ 1 and 12 species in the same molecule. By this route they hope to be able to produce ‘‘single component ferrimagnets’’ in the solid state. Solution EPR spectra in toluene reveal a 13-line spectrum due to hyperﬁne coupling to six 14N nuclei, aN (2.5 G) being about one-third of the splitting typically observed for monomeric nitronyl nitroxides indicating that both J and J 0 are significantly greater than aN. However, on every third 14N hyperﬁne line of this spectrum there is further coupling to 1H, with aH similar to that observed for a msubstituted benzoic acid derivative of the monoradical. The spectra can be analysed successfully by a perturbative approach conﬁrming the result that JcJ 0 , J00 caN. They conclude that the molecule can be described, in terms of magnetic behaviour, as independent S ¼ 1 and 12 moieties. Despite this, solid state samples of 6 do not show ferrimagnetic behaviour, due to only weak intermolecular coupling.

O N

N O

O O

O N

N N

N O

O

6

160

Electron Paramag. Reson., 2007, 20, 157–191

Turek and co-workers have extended their systematic study of phenylacetylene-, biphenyl- and meta-phenylene-bridged polyradicals to the tri- and tetraradicals 7–9.8 The symmetrically substituted tri-radicals 7 and 8 were all found to have spin-quartet ground states from variable temperature EPR studies of frozen solutions, consistent with ferromagnetic coupling via the m-aryl and m-phenylacetylene or biphenyl pathways. The exchange interaction within the biradical moiety is found to be strongest for R ¼ nitronyl nitroxide. The measured D-values for the quartet states (DQ) were compared to those of the triplet states (DT) for the corresponding m-aryl linked dimers (i.e. the diradical building blocks of the triradicals) through the relationship DQ ¼ DT/3, valid if the dipolar interaction between the diradical and monoradical parts is insignificant. For the nitronyl nitroxide species there is good agreement, indicating that the coupling within the diradical part is not significantly perturbed by appending the monoradical moiety. However, for the imino nitroxide (and mixed nitronyl nitroxide-imino nitroxide) species there is significant disagreement which the authors interpret as being due to significant changes in the torsion angles of the radicals between the bi- and tri-radicals. In contrast to the tri-radicals, the tetra-radicals 9 are found to have singlet ground states, in deﬁance of the ‘‘topological rules’’ based on spin-polarisation ideas. R1 R3 R2

7

R R R

8

O

O

N R

R

N or

R= N

N

O R

9

R

Veciana and co-workers have published two papers exploring the exchange between nitroxides in the substituted triphenyl phosphine oxides 10 and 11.9,10 In the di- and tri-nitronyl nitroxide substituted species 10 (n ¼ 2, 3) the isotropic hyperﬁne interaction to 14N is approximately one half and one third, respectively, of that seen for the monoradical n ¼ 1. Coupling is seen to all nitrogens (strong exchange limit), and 1H and 31P coupling is also resolved for n ¼ 1.9 The spin density at P is found to increase with increasing n. They ﬁnd a change in the sign of the spin density at P between the phosphine oxide and

161

Electron Paramag. Reson., 2007, 20, 157–191

analogous phosphine species (i.e. depending on whether the P has a lone pair of electrons or not) and speculate that this should have a significant effect on the sign of the exchange interaction. Unfortunately these proved to be too weak to measure by susceptibility methods. In order to address this they reported, in a second paper, the aminoxyl species 11 where the NO radical is conjugated directly into the aryl rings.10 Coupling to three equivalent 14N, six 1H and one 31 P is observed in the ﬂuid EPR spectrum (strong-exchange limit). Although they expected this species to be ferromagnetically coupled based on spinpolarisation arguments, it turns out to be antiferromagnetically coupled with J ¼ 7.55 cm1 and the authors conclude there must be another mechanism. The frozen solution EPR spectrum at 5 K is that of an isolated S ¼ 12 species. An antiferromagnetically coupled equilateral triangle would be expected to lead to a degenerate pair of S ¼ 12 states (spin frustration), thus the EPR behaviour suggests distortion to an isosceles triangle. O N PPh3-n N O

O n

10 Cl Cl

Cl But

N

O

Cl Cl

Cl Cl

Cl Cl

Cl Cl

Cl

Cl Cl

Cl Cl O

O P

P

N N O P P O N O O

O t

t

Bu

N O

N

11

O

Bu

12

Carriedo et al. have also studied exchange interactions, mediated by phosphorus, in the disubstituted triphosphazine derivative 12.11 Fluid-solution EPR of 12 shows hyperﬁne coupling to a single 31P (consistent with exchange via the –O–P–O– fragment) and to twelve 1H’s. The aH coupling is half that observed for the monoradical analogue hence JcaH. The resolution in further tri- and tetra-substituted analogues is much poorer, consistent with much smaller hyperﬁne interactions, as expected in the strong exchange limit, but that can be extracted by simulation. These show 31P coupling to two nuclei, hence there is electronic communication via the P-N-P backbone of the phosphazine ring. Finally, Rajca et al. have prepared the diradical 13, an analogue of trimethylenemethane (TMM), prepared by oxidation of its tetra-anion in THF.12 EPR studies show that there is a strong ferromagnetic coupling

162

Electron Paramag. Reson., 2007, 20, 157–191

between the radical centres, with a triplet-singlet gap of at least 100 K and a value in the triplet state of D ¼ 0.004 cm1, smaller than that observed in TMM. Ar

Ar

Ar

Ar Ar

Ar But

Ar =

13

3

d-Block

Elsenbroich and co-workers have published a beautiful series of papers on electronic communication in the trovocenyl-based polyradicals {trovocene, TVC ¼ [(Z7-C7H7)V(Z5-C5H5)]}.13–16 In a series of dimers where the TVCs are linked by acetylides via the cyclopentadienyl (Cp) moieties [(C7H7)V(C5H4R-C5H4)V(C7H7)],13 biradical-like spectra are observed in ﬂuid solution with 15-line patterns due to the two equivalent 51V nuclei, but with intensities that deviate from the binomial pattern expected in the strong exchange limit. This allows the J-values to be determined by spectrum simulation, giving (antiferromagnetic) J ¼ 0.92 [R ¼ –(CRC)-bridge], 0.56 [R ¼ –(CRC)2–] and 0.005 cm1 [R ¼ p-(CRC)2–C6H4], compared to the ‘‘parent’’, directly linked, ditrovocenyl which has J ¼ 2.8 cm1. They conclude from the systematic attenuation of J with increasing n (0–2) that it should be possible to detect exchange up to n ¼ 15 by EPR. The J-values are weak because the V 3dz2 magnetic orbitals are orthogonal to the molecular orbitals of the Cp ligands with little spin density on the bridging ligands. In a second paper they study further ditrovocenyls where R ¼ –CH2–, –CH2CH2–, –C(¼CH2)–, and cis and trans –C(Me)¼C(Me)–.14 The strong exchange limit is only observed with R ¼ –CH2–. They argue that in this species the single sp3 carbon forces tilting of the two Cp rings towards each other, facilitating exchange via overlap of the psystems. The next strongest interaction is observed for the sp2 bridged R ¼ –C(¼CH2)–. There is an order of magnitude difference in J between the two isomeric ethylene bridged species, in favour of the trans isomer. The authors argue that this is because the Cp rings can lie co-planar in the trans isomer. Surprisingly a significant J-value is observed for R ¼ –CH2CH2– which must either be due to a s-spin polarisation mechanism or through hyperconjugation of the C–H bonds in the Cp rings. The same group have extended this work to a tetra-trovocenyl radical, [1,2,4,5-(C7H7)V(C5H4)]4(C6H2).15 29 lines are

163

Electron Paramag. Reson., 2007, 20, 157–191

observed in the ﬂuid solution EPR spectrum with aV ca. one quarter of that for trovocenyl itself. Hence, this approaches the strong exchange limit, although the three different exchange pathways precluded determination of the J-values. Finally, these authors have probed the effect of a diamagnetic metal fragment as a linker between derivatised vanadocenes, in the complexes [{(Me2P-C6H5)2V}2Ni] and [{(Me2P-C6H5)2V}2CoH].16 The V. . .V exchange is found to be similar between the two complexes, ca. 0.3 cm1, and there is no resolution of 59Co hyperﬁne in the latter, so there is little effect on the exchange. A further example of biradical-like spectra from transition metal ions is given by Wong et al.17 for the di-copper(II) macrocycle 14. Fluid-solution EPR gives an intermediate exchange-type spectrum (J E aCu). However, on formation of a [2]catenane with the di-gold(III) analogue (a dication), the exchange is ‘‘switched oﬀ’’ or, alternatively, the two Cu(II) ions are insulated from one another by the diamagnetic, non-covalently bound, Au(III) ring. The EPR spectrum of the catenane species is indistinguishable from that of a monomeric Cu(II) bis-dithiocarbamate complex.

Bu

N

S

N S

S

Cu

Bu

S Cu

S

S

Bu

S

N

S N

Bu

14

Bond et al. have published EPR studies of the Cu(II) dimers [Cu2(CN)L]31, where L is an aza-cryptand (for example, ligand 15).18 The Cu(II) ions have trigonal bipyramidal coordination geometry with the unique axis along the Cu. . .Cu vector bridged by CN. The Cu(II) ions are antiferromagnetically coupled but, unusually for Cu dimers, have dz2 ground states as conﬁrmed by the EPR spectra of the S ¼ 1 excited states which have gx,y 4 gz E ge with gz collinear with the principal axis of the ZFS tensor.

HN

NH N

N HN

NH

HN

NH

15

164

Electron Paramag. Reson., 2007, 20, 157–191

Yano et al. have reported a low-temperature single-crystal X- and Q-band study of the mixed-valence Mn(III)Mn(IV) dimer [Mn2O2(phen)4](PF6)3,19 showing unusually well-resolved spectra for an undiluted crystal, with good resolution of the 55Mn hyperﬁne splittings. At 9 K only the S ¼ 12 ground state of this antiferromagnetically coupled dimer is populated. The data are analysed with an effective S ¼ 12 spin Hamiltonian. Near axial g-values are found with the ‘‘unique’’ lowest g-value (gz) perpendicular to the Mn2O2 core. The hyperﬁne couplings to Mn(III) and Mn(IV) are determined and, as expected, those for Mn(III) are significantly anisotropic (with the ‘‘unique’’ Az axis co-parallel with gz) while those for Mn(IV) are nearer isotropic. These data are rationalised in terms of a vector-coupling model based on data for Mn(III) and Mn(IV) ions doped in rutile. The authors discuss the relevance of their results for biological systems such as photosystem II, highlighting that the oft-quoted ‘‘2AMn(IV) ¼ AMn(III)’’ relationship used to interpret multi-line mixed-valence manganese spectra is not valid when the anisotropy of the hyperﬁne interaction is taken into account. A series of papers from several groups deals with antiferromagnetically coupled triangles of S ¼ 12 paramagnets.20–23 Such species are of interest because for an equilateral triangle an orbitally degenerate ground state should result (a pair of degenerate S ¼ 12). This is the strict definition of ‘‘spin frustration’’24 although many authors use this term simply to mean competing antiferromagnetic interactions. Such systems can relieve the orbital degeneracy via spontaneous distortion to isosceles or lower symmetry (energy gap d, a magnetic Jahn-Teller distortion) or via antisymmetric exchange interactions (G), giving rise to a total zero-ﬁeld splitting (D) between the two S ¼ 12 states of D2 ¼ d2 þ 3G2. Liu et al. report such interactions in [Cu3X(Hpz)2(m2-pz)3(m3-OMe)]X [X ¼ Cl, Br; Hpz¼3{5}-(mesityl)pyrazole] where the three Cu(II) ions are in a (near) isosceles triangle bridged by a central m3-OMe.20 Powder Q-band EPR at 4 K reveal two strong transitions at effective g-values of geﬀ ¼ 2.21 and 1.47 (X ¼ Cl) and 2.19 and 1.52 (X ¼ Br). In an axial system subject to antisymmetric exchange EPR transitions are expected at the true gJ and at g> 0 which is related to the true g> by g> 0 ¼ g>[{d2-(hn)2}/{D2-(hn)2}]1/2 which simpliﬁes to g> 0 /g> E d/D if hn { D,d and this can be used to estimate the contribution of the lower than trigonal symmetry to the total ZFS D. The authors note that in order to see g> 0 { ge (and hence the true g>) antisymmetric exchange effects must be in action (i.e. d a D), and these are estimated from powder magnetic susceptibility data in conjunction with the EPR spectra. In this work the molecule has lower than trigonal symmetry anyway. Ideally such effects would be studied in single crystals of a crystallographically trigonally-symmetric trimer, and this is the subject of a much more detailed study by Solomon and co-workers on the D3 (crystal structure determined at 1301C) symmetry [Cu3(DBED)3(m3-OH)](ClO4)3 (DBED ¼ N,N 0 -di-tBu-ethylenediamine).21 Here the (antiferromagnetic) isotropic J is 105 cm1. Powder EPR spectra at 5 K show a single broad resonance at geﬀ ¼ 2.32 with a partially resolved four-line hyperﬁne pattern with AJ ¼ 74 104 cm1. X-band single crystal studies at 3.7 K show a remarkable orientation dependence: with the

165

Electron Paramag. Reson., 2007, 20, 157–191

applied ﬁeld (H) parallel to the 3-fold axis of the molecule (z, y ¼ 01) a spectrum is observed similar to that from the powder; when H is rotated away from z this broadens and shifts up-ﬁeld rapidly, reaching geﬀ E 1.2 before being unobservable beyond y ¼ 601. The total ZFS D is determined from magnetic circular dichroism (MCD) studies as 67 cm1. These data are analysed via the spin Hamiltonian: ^ ¼ 2JðS^1 :S^2 þ S^2 :S^3 þ S^1 :S^3 Þ þ dðS1 :S2 Þ H þ dðS1 :S3 Þ þ G S^1 S^2 þ S^2 S^3 þ S^3 S^1 þ ðgz cos y þ gxy sin yÞbHðS^1 þ S^2 þ S^3 Þ where G is the antisymmetric exchange vector. When y a 01, the two S ¼ 12 states mix giving a non-linear dependence on H and low geﬀ. Transitions within the lower doublet should only be observed when d a 0 and hence the EPR provides direct evidence for a spontaneous symmetry lowering from D3, and ﬁtting the angular dependence of the EPR gives d ¼ 17.5 cm1 which in turn gives G ¼ 36 cm1. The authors point out that the origin of the antisymmetric exchange is in efﬁcient ground state-excited state exchange interactions between the Cu(II) ions. The same group followed this work with a study of a ferromagnetically coupled, trigonally (C3) symmetric Cu(II) trimer, [Cu3(16) (m3-O)](ClO4)4.22 There is a rather large ZFS in the spin-quartet ground state (D ¼ 2.5 cm1), determined from the temperature dependence of the intensity of the solid state EPR spectrum (geﬀ ¼ 3.64 and 2.06 at X-band).

N

NH

N H

N H

N

N N

N

HN HN

NH N

16

The authors discuss the origin of this large splitting in terms of the anisotropic exchange which they point out has the same physical basis as the antisymmetric exchange – coupling between the ground state of one metal ion [in this case dz2 because of the trigonal bipyramidal geometry at each Cu(II) ion] and the excited state of another. Glaser et al. report another ferromagnetically coupled Cu(II) trimer, [Cu3(talen)], where talen is a tris-salen derived ligand.25 Wellresolved perpendicular and parallel mode X-band spectra are observed from frozen solutions at 3 K. Although D is too small to resolve ﬁne structure, modelling the relative intensities of the allowed and forbidden (half-ﬁeld) transitions gave |D| ¼ 74 104 cm1. The spectra show 10-line hyperﬁne patterns due to the three equivalent Cu(II) ions, which they claim is the ﬁrst time this has been observed. Nellutla et al. used EPR to prove the localisation of the spin ground state on a single copper ion in the penta-copper

166

Electron Paramag. Reson., 2007, 20, 157–191

[Cu5(OH)4(H2O)2(A-a-SiW9O33)2]10 where the Cu(II) ions essentially deﬁne a rectangular based pyramid with the apical Cu(II) equally bound to the other four.26 This cluster is sandwiched between the diamagnetic polyoxotungstates. Low-temperature EPR gives a rhombic S ¼ 12 spectrum with hyperﬁne to a single Cu(II) ion which they interpret, in conjunction with magnetic susceptibility data, as a well isolated S ¼ 12 ground state located on the apical Cu(II) ion. Halcrow and co-workers report a heptametallic Cu(II) cluster [{Cu3(Hpz)6 (m3-Cl)(m3-OH)3}2Cu] based on two vertex-sharing {Cu4O4} heterocubanes.27 From modelling magnetic susceptibility data the S ¼ 12 ground state arises largely from the linear Cu3 fragment in this topology. Q-band EPR of a powder at 4 K reveals an axial S ¼ 12 spectrum with a well-resolved eight-line hyperﬁne pattern on the parallel feature. This can be simulated by assuming hyperﬁne coupling to two equivalent and one inequivalent Cu nuclei with a 2.5:1 ratio of magnitudes. Significantly above or below this ratio gives simpler ﬁve or sevenline patterns, respectively, towards the expected limits for coupling to two or three equivalent Cu. The three copper ions have similar coordination environments, and the difference in the coupling to the central Cu ion is rationalised from a simple vector coupling analysis assuming the |S13,Stotal 4 ¼ |1,1/24 ground state of a linear trimer. Three papers have appeared detailing EPR studies on ‘‘molecular wheel’’ complexes. Pilawa et al. studied the antiferromagnetically coupled [Fe6(tea)6] (tea ¼ trianion of triethanolamine) where the six Fe(III) ions are bridged by the alkoxide arms of the ligand.28 The antiferromagnetic coupling leads to a diamagnetic ground state but EPR transitions are observed from the ﬁrst S ¼ 1, 2, 3 and 4 excited states in an X-band single crystal study at 30 K. The spectra had previously been analysed by a full interaction Hamiltonian to obtain the single-ion and exchange parameters: in this work the authors probed relaxation dynamics via the linewidths. Below ca. 30 K the linewidths are dominated by inter-cluster dipolar interactions and in order to model this it is necessary to consider states up to S ¼ 4, while above this temperature spin-lattice relaxation dominates. The odd-membered ring [(C6H11)2NH2][Cr8NiF9(O2CCMe3)18] has been studied by Sessoli and co-workers.29 Here the antiferromagnetic coupling gives rise to an S ¼ 0 ground state, but the odd number of metal ions means that a simple ‘‘up-down’’ arrangement of spins cannot be satisﬁed around the ring – a form of frustration – and the authors make the analogy to a Mo¨bius strip. The ﬁrst excited state is S ¼ 1 but the nature of this state is different depending on the relative magnitude of the Cr. . .Cr and Cr. . .Ni exchange interactions. The g-value of the S ¼ 1 state is expected to be approximately equal to gNi [singleion g-value of the Ni(II) ion] if JCrCr c JCrNi from consideration of the form of the wavefunction by a simple vector coupling model. If JCrCr { JCrNi then a similar analysis predicts g ¼ 3/2gCr-1/2gNi. Because the gCr and gNi are expected to be substantially different this allows determination of the relative amplitude of the exchange interaction from the EPR spectrum of the S ¼ 1 excited state. Low temperature powder spectra at 285 GHz suggest that JCrCr { JCrNi. A ferromagnetically coupled decametallic Cr(III) ring, [Cr10(OMe)20(O2CCMe3)10], has been studied by Sharmin et al.30 Although magnetic susceptibility measurements

Electron Paramag. Reson., 2007, 20, 157–191

167

show that the average J-value is ferromagnetic, the low-temperature EPR spectra measured between 50 and 70 GHz on powder and single crystal samples are more consistent with an S ¼ 9 ground state (D/kB ¼ 0.045 K, where kB is the Boltzmann constant). This can only result if some of the Cr. . .Cr interactions are antiferromagnetic, which is consistent with the molecular symmetry – there are in principle ﬁve different J-values given the two-fold symmetry of the molecule. Clusters with such high-spin ground states are of interest because this can lead to unusual low-temperature physics: for example, when combined with a significant negative ZFS this can lead to a barrier to relaxation of magnetisation and hence a molecular magnetic memory effect. Such materials have become known as ‘‘single molecule magnets’’ (SMMs). One of the smallest SMMs prepared to date is the heterodimetallic [MnIIICuII(5-Br-sap)2(MeOH)] (5-Br-sap ¼ 5-bromo-2-salicylideneamino-1-propanol) reported by Oshio et al.31 The two metal ions are strongly coupled ferromagnetically via two alkoxide arms of the ligand to give a well isolated S ¼ 5/2 ground state. Wellresolved 342 GHz EPR spectra are consistent with this and give D ¼ 1.81 cm1. This cluster D is then simply related to the single ion Mn(III) ZFS via D5/2 ¼ (16/25)DMn, although this analysis neglects dipolar contributions. One of the important magnetic phenomena observed in SMMs is quantum tunnelling of magnetisation (QTM) – this can occur between pairs of MS states on either side of the energy barrier to relaxation, e.g. in a spin S with negative D it can be possible for the system to tunnel between MS ¼ S to þS in zero ﬁeld where these states are degenerate. This can often be observed as a step in magnetisation vs applied magnetic ﬁeld loops. An unusually large step is observed in such a measurement for the S ¼ 4 ground state SMM [Ni(hmp)(tBuEtOH)Cl]4 (hmp ¼ anion of hydroxymethylpyridine; tBuEtOH ¼ 3,3dimethyl-1-butanol) implying very fast QTM at zero ﬁeld.32,33 This molecule has axial (S4) symmetry, hence rhombic E terms in the spin Hamiltonian are forbidden. Single crystal measurements with the applied ﬁeld varied in the hard plane of magnetisation (xy) reveal a four-fold periodicity. These can be reproduced using a Hamiltonian including a fourth order term B44(Sˆ4þ Sˆ4) which induces mixing between states differing between MS 4. Simulation gives B44 ¼ 4 104 cm1 which equates to a large 10 MHz tunnel splitting between the MS 4 states in zero ﬁeld. Cornia and co-workers have observed important symmetry effects on the barrier height in a family of tetrametallic SMMs based on centred triangles of Fe(III).34,35 Antiferromagnetic coupling leads to an S ¼ 5 ground state. The parent molecule [Fe4(OMe)6(dpm)6] (Hdpm ¼ dipivaloylmethane) has only C2 point symmetry and EPR studies reveal a ground state ZFS of D/kB ¼ 0.29 K. The six methoxides can be replaced with the tripodal tris-alkoxide ligands RC(CH2O)3 (R ¼ Me or Ph). When R ¼ Ph C2 is maintained, but with R ¼ Me higher symmetry (D3) is imposed. For both derivatives the ground states have much higher D/kB (0.64 and 0.61 K, respectively, determined by 230 GHz EPR measured on powders between 30 and 10 K) leading to much larger magnetisation blocking temperatures. The authors speculate that this may be due to different extents of trigonal distortion

168

Electron Paramag. Reson., 2007, 20, 157–191

of the apical iron sites in the different molecules and hence differing single-ion ZFS. Christou’s group report a further example of the importance of molecular symmetry on the magnetic behaviour of clusters.36 The tetrametallic Mn(III)3Mn(IV) cluster [Mn4O3(O2CPh)4(dbm)3] has an S ¼ 9/2 ground state, and is part of a family of {Mn4O3X} heterocubane structures studied by this group. When X ¼ Cl the cluster has pseudo-C3 symmetry. Here X is an Z2,m3-carboxylate which opens up one edge of the heterocubane reducing the symmetry to pseudo-CS. This leads to a better alignment of the Mn(III) JahnTeller axes with each other than in the parent X ¼ Cl cluster (where they intersect at X). 249 GHz EPR studies on a single crystal gives D ¼ 0.646, E ¼ 0.14 cm1, both significantly larger than those observed for X ¼ Cl (D ¼ 0.53 cm1, E ¼ 0). Despite the larger |D| value the lower symmetry molecule shows faster magnetisation relaxation rates, and this is ascribed to the non-zero E which leads to rapid QTM in zero applied ﬁeld. van Slageren et al. have reviewed the application of frequency-swept EPR to SMMs and related species,37 pointing out that it has the advantage for measuring ZFS parameters of being measured in zero applied ﬁeld! A nice example of the application of this technique is on Brechin and co-workers’ S ¼ 17/2 ground state [Mn9O7(O2CMe)11(thme)(py)3(H2O)2] {H3thme ¼ tris(hydroxymethyl)ethane}, the core of which contains a trigonal {Mn(IV)3 (m3-O){Mn(III)4Mn(II)2O6} ring.38 Frequency-swept EPR spectra in zero applied magnetic ﬁeld, measured on a powder between 10 and 2 K and in the frequency range 2.8–5.0 cm1, show three resonances at 3.79, 3.41 and 2.95 cm1. The highest frequency transition is the most intense, and increases in intensity with decreasing temperature, showing that D is negative – this is the MS ¼ 17/2 to 15/2 transition. Fitting to these data gives D ¼ 0.247 cm1, E ¼ 0, B40 ¼ þ4.6 104 cm1 where the fourth-order operator is: ^0 ¼ 35S^2 ½30SðS þ 1Þ 25S^2 6SðS þ 1Þ þ 3S2 ðS þ 1Þ2 O 4 z z Kirchner et al. studied the lineshapes of the transitions in this complex, and in the anionic (PPh4)[Mn12O12(O2CEt)16(H2O)4] (the one-electron reduction product of the parent class of SMMs, see later) which has an S ¼ 19/2 ground state.39 In the former the lineshapes are Gaussian and temperature independent, i.e. the resonances are inhomogeneously broadened, and this is likely due to D-strain effects. In contrast, those of (PPh4)[Mn12O12(O2CEt)16(H2O)4] are Lorentzian and temperature dependent – they are homogeneously broadened. The authors relate this to the lifetimes of the excited-state MS levels, estimating lifetimes of 50–58 ps from the MS ¼ 19/2 to 17/2 transition depending on temperature. This observation of homogeneous broadening is unusual for a SMM, and stands in contrast to work by the same group on the archetypal SMM, [Mn12O12 (O2CMe)16(H2O)4] 2(MeCO2H) 4H2O (‘‘Mn12Ac’’; S ¼ 10 ground state), where inhomogeneously broadened Gaussian lines are observed. The cause of this difference between Mn12Ac and (PPh4)[Mn12O12(O2CEt)16(H2O)4] is

Electron Paramag. Reson., 2007, 20, 157–191

169

unclear. Vongtragool et al. have also used the frequency-swept EPR experiment to measure the relaxation of Mn12Ac as follows:40,41 the molecule is cooled in an applied ﬁeld, thus magnetising the system, and only the MS ¼ 10 to 9 transition is observed at a higher frequency cf. zero-ﬁeld measurements. Now the applied ﬁeld is reversed and a new peak (MS ¼ þ10 to þ9) is observed to grow in down-frequency (cf. the zero-ﬁeld experiment) as the molecule slowly depopulates from the meta-stable MS ¼ 10 to the new ground state of MS ¼ þ10. Monitoring evolution of the intensities with time gives a relaxation time of ca. 150 minutes for an applied ﬁeld of 0.45 T. The literature on EPR studies of SMMs is still dominated by Mn12Ac and its ever-increasing number of derivatives. In addition to the above studies, several reports have appeared from Hill and co-workers highlighting the importance of crystal disorder in determining observed magnetic properties. Although Mn12Ac crystallises in a tetragonal space group and has global 4-fold symmetry, Cornia et al. had previously shown that disorder of the solvents of crystallisation leads to a symmetry-breaking effect leading to six different isomers in the lattice, only two of which retain four-fold symmetry.42 Hill provided evidence for this from detailed single-crystal studies with the applied ﬁeld oriented in the hard plane of magnetisation (and narrow ranges of angles deviating from this plane).43 They demonstrate that the observed four-fold periodicity in this plane and the presence of anomalous peaks (peaks not expected from a simple model assuming a single magnetically unique molecule) can be modelled by assuming (i) local distortions from four-fold symmetry (introducing non-zero E) in subsets of molecules as described by Cornia, and (ii) tilting of the easy axis of magnetisation (molecular z axes) of subsets of molecules away from the unique crystal axis. This latter effect introduces subsets of molecules with transverse ﬁeld components when a magnetic ﬁeld is applied along the crystal’s easy axis of magnetisation, and the authors speculate that this could be the mechanism for QTM transitions between MS states differing by odd integer values. These have been observed in magnetisation vs. applied ﬁeld studies, but cannot be accounted for by non-zero E (which mixes states differing by MS 2) or B44 (MS 4). Christou and co-workers have prepared new variants of Mn12 in order to try to remove the solvent disorder problem.44,45 The extra structure observed in the high frequency EPR spectra of Mn12Ac (see above) is absent in those of [Mn12O12(O2CCH2Br)16(H2O)4] 4CH2Cl2 and [Mn12O12(O2CCH2But)16(H2O)4] MeOH which both have crystallographic S4 symmetry (as does Mn12Ac, notwithstanding solvate disorder), thus implying that these may have true fourfold molecular symmetry without disorder and that there is no tilting of the molecules’ easy axis of magnetisation with respect to each other. In contrast to Mn12Ac, the solvate molecules in these two materials do not hydrogen-bond to the cluster. Hill also reports evidence for a low-lying S ¼ 9 excited state in [Mn12O12(O2CCH2Br)16(H2O)4] 4CH2Cl2 from variable temperature single crystal EPR studies at 60–80 GHz.46 With the applied magnetic ﬁeld perpendicular to the easy axis of magnetisation, weak extra peaks are observed between those assignable to the S ¼ 10 ground state. From the variable temperature behaviour – these peaks reach a maximum in intensity at ca. 20 K – they

170

Electron Paramag. Reson., 2007, 20, 157–191

conclude these are due to an S ¼ 9 excited state with otherwise similar spin Hamiltonian parameters to the S ¼ 10 ground state. They place the MS ¼ 9 states of the S ¼ 9 at ca. 40 K above the MS ¼ 10 states of the S ¼ 10, which means that the two S manifolds overlap considerably. EPR studies of two other Mn12 variants have been reported. Hendrickson gives another example of a ‘‘Jahn-Teller isomer’’, in [Mn12O12(O2CC6H4-p-Me)16(H2O)4] CH2Cl2.47 The Mn12 family contain eight Mn(III) ions and four Mn(IV) ions and in Mn12Ac itself the Jahn-Teller distortion axes of the high spin d4 Mn(III) ions are coparallel with each other and with the molecular z axis (easy axis of magnetisation). It is this that gives rise to the large, negative cluster ZFS. However, in some Mn12 variants one or more of the Jahn-Teller axes are tilted with respect to the rest, and [Mn12O12(O2CC6H4-p-Me)16(H2O)4] CH2Cl2 is an example of this. Curiously 324 GHz EPR gives D ¼ 0.47 cm1 which is similar to that reported for Mn12Ac itself. Hachisuka et al. report a Mn11Cr cluster where one Mn(III) ion in Mn12Ac has been replaced by Cr(III) giving rise to a S ¼ 19/2 ground state.48 Surprisingly they ﬁnd the other ground state spin Hamiltonian parameters to be identical to those of Mn12Ac – replacement of an anisotropic Mn(III) ion with a pseudo-isotropic Cr(III) ion had little inﬂuence on the cluster anisotropy.48 Finally, in addition to the frequency-swept EPR studies reviewed above, there are two interesting ‘‘new’’ experimental procedures reported for high spin ground state clusters. Piligkos et al. have performed a single crystal parallel mode X-band EPR study of the S ¼ 6 ground state [Cr12O9(OH)3(O2CCMe3)15].49 Although widely exploited in the study of high spin FeS clusters in biological systems, this is the ﬁrst application of parallel mode modulation to very high spin molecular clusters. The authors discuss the consequences of the different selection rules (DMS ¼ 0 cf. 1) for selective connection of pairs of eigenstates in high spin clusters. Two groups report the use of magnetisation-detected EPR as illustrated on the S ¼ 10 ground state SMM [Fe8O2(OH)2(OH)12(tacn)6]Br8 (‘‘Fe8’’; tacn ¼ 1,3,5-triazacyclononane).50–53 Petukhov et al. monitor the magnetisation response of a single crystal of Fe8 under microwave radiation (pulsed or c.w.) as the applied magnetic ﬁeld is swept, using a Hall-probe magnetometer (an array of 10 10 mm2 Hall bars).50 The sensitivity of the Hall-probe technique allows measurements on micron-sized single crystals. Data are measured at 118 GHz and between 2 to 20 K. The authors also illustrate that the spectra can be transformed to determine the spin temperature of the sample by mapping onto variable temperature magnetisation vs. ﬁeld curves (in the absence of microwave radiation). Cage et al. report a conceptually similar experiment but using a commercial SQUID magnetometer where the Fe8 single crystal (ca. 3 mm3) is under irradiation at 95 or 141 GHz.51,52

4

Mixed p/d-Block Radicals

Wieghardt and co-workers have continued their systematic studies of coordination complexes of redox non-innocent ligands. The o-diiminobenzosemiquinonate

171

Electron Paramag. Reson., 2007, 20, 157–191

radical anion ligand (17) complexes to Fe(III) with a co-ligand X to form ﬁvecoordinate species [Fe(17)2X]n1 where X ¼ I (n ¼ 0) or PBu3 (n ¼ 1).53 Xband EPR spectra at 10 K in frozen solutions reveal S ¼ 12 spectra, consistent with this ground state as shown by magnetic susceptibility data. Both species have rhombic g-values with all g-values greater than ge and this is inconsistent with a simple assignment as low spin Fe(III) with two diamagnetic ligands as had been previously suggested. However, this is consistent with two radical ligands coordinated to an intermediate spin S ¼ 3/2 Fe(III) ion and this is supported by Mo¨ssbauer spectra. The S ¼ 12 ground state arises from strong antiferromagnetic coupling between the metal ion and the radical ligands, where the inter-ligand exchange is negligible. Interestingly, when X ¼ I a wellresolved hyperﬁne pattern due to the 127I nucleus is observed that can be simulated with inclusion of the 127I hyperﬁne (A) and electric quadrupole (P) interactions. They further note that the spectra can only be reproduced by including non-coincidence between the A and P principal axes. Ph N

NH

17

The relatively rare intermediate spin state for iron(III) is also invoked in the bis-(benzene-1,2-dithiolato) complexes [Fe2(18)3(18 . )] and [Fe(18 . )2 (PMe3)]1.54 The X-band spectrum of the dimetallic species at 10 K shows a near-axial spectrum with low g-anisotropy (g1–g3 ¼ 0.05), only consistent with an S-centred radical. This is consistent with the magnetic model where the antiferromagnetic exchange between the two Fe(III) ions is much stronger than that between the metal ions and the radical ligand. In the monometallic species a rhombic S ¼ 12 is observed at 10 K in frozen solution, with much larger ganisotropy (g1–g3 ¼ 0.15). In contrast to the cationic [Fe(18 . )2(PMe3)]1, the reduced, anionic [Fe(18)2(PMe3)] exhibits effective g-values typical of an S ¼ 3/2 state with large D, consistent with the ligands being in their reduced, dianionic and diamagnetic form. While EPR is obviously useful in characterisation of the redox states of ligand and metal, the authors conclude that the best ‘‘ﬁngerprints’’ for oxidation state of both metal and benzene-1,2-ditholate ligands in such complexes come from infrared (IR), visible/near-IR and Mo¨ssbauer spectroscopies. t

Bu

tBu

S

S

-e+e-

tBu

S

18

tBu

S

18

Thomas et al. have probed the electronic structure of copper(II) complexes of the bis-salen type ligand 19, [Cu(19)], as a possible model for galactose

172

Electron Paramag. Reson., 2007, 20, 157–191

oxidase.55 The complex undergoes two one-electron oxidations which are assigned as oxidations of the ligand to phenoxyl radicals by comparison with the redox behaviour of the free ligands. [Cu(19)] has a normal tetragonal Cu(II) EPR spectrum, while the monocation [Cu(19)]1 is found to be EPR silent at Xband and 4–100 K. The dication [Cu(19)]21 has resonances at geﬀ ¼ 8.0, 4.7, 2.9 and 1.5 which the authors state belong to the MS ¼ 1/2 (4.7, 2.9) and 3/2 (8.0, 1.5) doublets of a zero-ﬁeld split S ¼ 3/2 state resulting from ferromagnetic coupling of the p-radical ligands with the Cu(II) ion. The interesting hetero-three-spin system 20 is reported by Kirk et al.56 The X-band EPR spectrum of 20 in frozen solution at 5 K gives effective g-values typical of an S ¼ 3/2 state with large |D|, implying ferromagnetic coupling of the three components; simulation gives |D| ¼ 0.5 cm1. The authors attempt to rationalise the strong ferromagnetic coupling between the two organic radical centres in a valence-bond conﬁguration-interaction model. O Cu(Tp) tBu

N tBu

O

N

OH

tBu

HO

N

O N

tBu

tBu

O

19 20 N

N N

Mn N

O

N

Mn O

N

O

N

N

O

O

N

Mn N

O

N

Mn O

O

N

N

N O

N

Mn N

O

N

Mn O

O

N

N O

21

N O

22 O N

N O

23

Finally, Bill, Wieghardt and co-workers have published a detailed multifrequency (S, X, Q-band) EPR study of heterovalent Mn(III)Mn(IV) dimers coupled to nitroxide free radicals (21–23),57 with relevance to biological systems such as the tyrosine radical – manganese cluster in photosystem II. The Mn(III)Mn(IV) dimers are strongly antiferromagnetically coupled (J E 130 cm1) to give well isolated S ¼ 12 ground states and these couple weakly (J 0 ) with the appended free radical. Studies on the isolated Mn(III)Mn(IV) dimers (no free radical) and Mn(IV)Mn(IV)-appended radical systems deﬁne the magnetic

Electron Paramag. Reson., 2007, 20, 157–191

173

parameters of the isolated fragments. The large J and weak J 0 mean that the low temperature EPR spectra are dominated by the singlet-triplet system arising from the coupling of the two ‘‘isolated’’ S ¼ 12 moieties. Parallel and perpendicular mode X-band spectra of frozen solutions were simulated using the full interaction spin Hamiltonian [considering only the ground state of the Mn dimer (Mn2) and the free radical (rad) spins]: ^ pair ¼S^rad ðJd J 0 1Þ S^0 H Mn2 X X 0 ^ þ mB Sj gj B þ S^Mn2 Ai I^i with j ¼ rad, Mn2; i ¼ Mn(III), Mn(IV). Jd is the anisotropic (dipolar) coupling matrix, J 0 is the isotropic exchange, 1 is the unitary matrix, and the Mn hyperﬁne interaction is included as this is a potential mechanism for singlettriplet mixing. The effect of J 0 on EPR spectra is investigated by systematic simulation of parallel and perpendicular mode spectra and monitoring the behaviour of singlet-triplet transitions. The authors note that parallel mode is more sensitive to changes in J 0 because the otherwise intense DMS ¼ 1 transitions are attenuated. Although such transitions are not observed in 21–23, this can still provide upper or lower limits for J 0 . Jd – magnitude and orientation – can in principle be determined from the half-ﬁeld transitions. For complexes 21 and 23, J 0 is found to be significantly larger than the microwave quantum and hence these can be analysed in the strong exchange limit, consistent with the J 0 ¼ 1.1 and 2.8 cm1 values, respectively, found from magnetic susceptibility studies. There is a significant difference in the relative intensities of the half-ﬁeld (g ¼ 4) and g ¼ 2 regions of the spectra between these two complexes, implying a significantly smaller anisotropic coupling for 21 despite the shorter Mn2. . .rad distance, cf. 23. This is found to be because Jd is unexpectedly large for 23 given its inter-spin separation (see below). No significant half-ﬁeld signal is found for 22 and no singlet-triplet transitions are observed, and a much smaller |J 0 | of ca. 0.02 cm1 is concluded from simulation; this species also exhibits the smallest anisotropic coupling of the series. The dipolar couplings are calculated based on the known inter-spin distances (from X-ray crystallography) using a three-spin point model: ^ dipole ¼ðm0 m2 =4pÞ H B

X

ðg0Mn2 grad =r3 ÞKi

i¼1;2

0 0 SMn2 Srad 3 SMn2 ri;rad Srad ri;rad P where g0Mn2 ¼ iKigi is the g value of the manganese dimer, and Ki are the projection coefﬁcients for Mn(III) and Mn(IV). The calculated dipolar interactions for 21 and 23 (and their relative magnitudes) are consistent with the X-ray structures, and also give some information on the orientations of the nitroxyls, while the calculated value for 22 is much too small. The authors

174

Electron Paramag. Reson., 2007, 20, 157–191

conclude that the assumption that the radical spin is localized on the N–O moieties may not be valid and that spin density leaks onto the aryl spacer. 5

Mixed d/f-Block Radicals

There is one report of an EPR study of a mixed d/f-block system, for the pair of complexes (NMe4)[LnNi6(pro)12](ClO4)4 (Hpro ¼ proline) where Ln ¼ Gd or La.58 These structures consist of an octahedron of Ni(II) ions arranged about the trivalent lanthanide ion. Using a Kambe´ approach to analyse magnetic susceptibility data for the complex containing the f 7 Gd(III) ion, assuming only Gd. . .Ni and nearest-neighbour Ni. . .Ni interactions, the authors conclude that there is a ﬁve-fold orbitally degenerate S ¼ 13/2 ground state – a strictly frustrated system. When Ln is the diamagnetic f 0 La(III) ion, weak antiferromagnetic exchange is observed between the Ni(II) ions. The Gd complex only gives an EPR transition at g E 2, and the upper limit of a ZFS is estimated. In contrast, the La complex is EPR silent at X-band but gives only two features at high frequency (ca. 150–500 GHz). The authors assign the low-ﬁeld transition as the DMS ¼ 2 of a spin triplet and the high-ﬁeld transition as a doublequantum transition, although they do not justify this assignment. From a frequency-resonance ﬁeld plot they determine |D| ¼ 5.5, E ¼ 0.8 cm1 for an S ¼ 1 state. They conclude that the magnetic exchange is much less in magnitude than the single-ion anisotropy in this complex. 6

Biological Systems

6.1 Methods. – Pulsed electron-electron double resonance (ELDOR) or double electron-electron resonance (DEER) has become a widely employed technique used to examine distances between paramagnetic centres and in biological systems these centres are often nitroxide-based and introduced via chemical modiﬁcation. Application of pulsed ELDOR to pairs of intrinsic metal-centred paramagnets has been undertaken by Bittl’s group, and their prototypical examples were an as-isolated form of a [NiFe] hydrogenase and a [3Fe-4S]1 cluster.59 These oligomers introduced the additional factors (compared to nitroxides) of significant g-value anisotropy and a resultant spin, in the case of the latter, derived from magnetic exchange, wherein spin projection needs to be considered. The ELDOR spectra for the [3Fe-4S]1 and [NiFe] centres in the hydrogenase from D. vulgaris Miyazaki F gave an experimentally observed dipole frequency of 8.5 MHz, which was larger than anticipated for the known inter-cluster distance from X-ray diffraction of 21.4 A˚, assuming two point dipoles. Considering spin projection, the authors showed that the Fe ion of the [3Fe-4S]1 cluster closest to the [NiFe] cluster had a large positive spin projection factor, which itself led to the calculation of a spin mixing parameter, a, that can also be determined independently from analysis by Mo¨ssbauer spectroscopy. Hence a further conclusion via values of a was that the [3Fe-4S] ferredoxin from Azotobacter vinelandii and that from the hydrogenase of the present study were very similar.

Electron Paramag. Reson., 2007, 20, 157–191

175

The Frankfurt group led by Prisner have introduced a method based on pulsed EPR and T1 relaxation times to extract individual components from overlapping spectra. The method, termed relaxation ﬁltered hyperﬁne (REFINE) spectroscopy60 was applied to mitochondrial complex I, which is a multi-subunit, membrane-bound protein from the respiratory chain and contains several Fe-S centres. In particular the iron-sulfur cluster designated N2 has EPR signals that overlap significantly with other signals from this type of cluster. The method utilises an inversion-recovery ﬁlter, and a ﬁlter time TF is deﬁned such that this is the time at which the inversion recovery amplitude is zero, and no Hahn echo is observed in the pulsed detection sequence. Thus, for samples with components having different TF values, individual spectra can be extracted from the mixture. It is suggested that a 2-dimensional time versus magnetic ﬁeld sequence can be employed, without necessarily determining accurate TF values for the individual components, to deconvolute spectra. A powerful extension of the method was demonstrated by applying it to ESEEM spectra with the results that it was shown that cluster N1 contains a ferredoxin [2Fe-2S] type cluster interacting with a peptide backbone nitrogen atom and that cluster N2 shows no direct bonding to nitrogen atoms. 6.2 Nitrogenases. – Hoﬀman and Seefeldt have embarked on a series of studies of intermediates in nitrogenase reactivity and have reviewed substrate interactions with the active site of nitrogenase.61 They used ENDOR and ESEEM at Q-band on the S ¼ 32 states of protein-bound compared with extracted FeMo-cofactor to assess whether the unassigned electron density located by X-ray crystallography in the centre of the [MoFe7S9:homocitrate] cluster of nitrogenase might belong to a nitrogen atom.62 They deduce from a detailed comparative study that observed 14/15N responses all arise from protein-bound nitrogen nuclei and not from the cofactor. Hence they assert that unless the proposed nitrogen atom in the centre of the cluster is magnetically uncoupled (i.e. hyperﬁne splitting B0 MHz, where the authors claim an experimental detection limit for A(14N) on the S ¼ 32 system of B0.1 MHz) from the electron spin system, then in their standard preparative conditions there is no nitrogen atom at the centre of the cofactor. Rapid freeze experiments on reduction of propargyl alcohol (HCRC– CH2OH) by nitrogenase trapped an intermediate S ¼ 12 state, which was investigated by c.w. and pulsed 13C ENDOR, and using Mims and stochastic-ﬁeld modulated methods for 1,2H ENDOR at Q-band frequency.63 Using the derived hyperﬁne interaction matrices an organometallic, doubly reduced, side-on bonded allyl alcohol product is proposed.64 Wild-type and mutants of nitrogenase, designed for early (e), middle (m) and late (l) stage studies of turnover using nitrogen (N2), methyldiazene, (CH3N¼NH) and hydrazine (N2H4), respectively were investigated by 15N Mims-pulsed and 1H c.w. ENDOR at Q-band frequency to provide evidence for the structure of trapped intermediates. These were assigned from single- and double-labelling with 15N, and were consistent with a cluster-substrate intermediate. Work on hydrazine trapping was extended65 and turnover of a double mutant nitrogenase, which

176

Electron Paramag. Reson., 2007, 20, 157–191

also involved EPR signals from a S ¼ 32 spin system, produced a freeze-trapped S ¼ 12 state with a rhombic EPR signal (g-values: 2.09, 2.01, 1.93) optimised at pH ¼ 7.4, showing the same Curie-Law behaviour to that of the resting state EPR signal. 15N Mims pulsed ENDOR at Q-band with samples of isotopically enriched hydrazine showed that a hydrazine-derived species was incorporated at the FeMo-cofactor and the low ﬁeld g-value carried a 15N hyperﬁne interaction of 1.5 MHz. These workers also studied dihydrogen evolution using 1,2H ENDOR on another variant protein that allowed a S ¼ 12 state to be trapped during reduction of protons to H2.66 Hyperﬁne matrices were determined from 2-dimensional ﬁeld-frequency plots that gave a large isotropic hyperﬁne coupling Aiso B 23 MHz, and hence deduced to be bound to the cofactor, but in addition the two hydrogen atoms have the same principal values of their hyperﬁne matrices, and so are possibly two chemically equivalent bound hydride ions. Newton, Fisher and co-workers have puriﬁed nitrogenase proteins from Gluconacetobacter diazotrophicus, a sugar-cane colonizing bacterium and obtained characteristic c.w. EPR spectra at X-band frequency reporting results from MoFe and Fe proteins.67 The EPR signals from component 2 of G. diazotrophicus (Gd2) were at g-values ¼ 4.16 (4.27), 3.69 (3.64) and 2.01 (2.01), where the values in parentheses are those from dithionite-reduced protein from A. vinelandii, and additional features at g ¼ 5.09 and 3.96, which were also present in whole cells, crude extracts and throughout the puriﬁcation procedure. The S ¼ 12 signal of the Fe protein was reported to be identical to that from the protein isolated from A. vinelandii. In a separate study68 using nitrogenase from A. vinelandii, this group investigated turnover under carbon monoxide, which has the effect of removing the characteristic S ¼ 32 signal and producing several new ones. These signals have been investigated for a series of mutant MoFe proteins, but no new parameters were observed, which implied that no new CO binding sites had been generated in the variants. Ribbe et al. identiﬁed two [8Fe-7S] P-cluster variants of nitrogenase, which they characterised by analysis, activity and a range of spectroscopic techniques including c.w. EPR at X-band and showed S ¼ 12 signals.69 Petersen and co-workers investigated Mn21-adenosine nucleotide complexes in the presence of the iron protein of nitrogenase from Klebsiella pneumoniae by c.w. EPR and 2D-ESEEM at Xband frequency.70 By changing the redox state of the Fe-protein and detecting conformational rearrangements directly at the nucleotide binding site it was suggested that the Fe-S cluster communicates with that site. 6.3 Copper. – Copper incorporation (reconstitution) in the multi-copper oxidase known as CueO (formerly yacK) followed by c.w. X-band spectroscopy did not show any resonances in addition to those attributed to the type 1 and type 2 centres.71 The plasma-membrane, multi-copper oxidase protein Fet3p in Saccharomyces cerevisiae also contains one each of type 1, type 2 and bimetallic type 3 sites and is essential for iron uptake in yeast. A study of mutants, by Solomon et al. using EPR, CD and MCD spectroscopies, reported no unusual EPR signatures of the resultant tricopper (type 2 plus type 3) cluster.72 In Rhus

177

Electron Paramag. Reson., 2007, 20, 157–191

vernicifera laccase a similar set of copper centres is found made up of the three types of copper sites. Multi-frequency c.w. EPR at X- and Q- bands and at 285 GHz were used to study the trimetallic active site and DFT calculations of models were used to estimate isotropic magnetic exchange constants (J), but the EPR data were not able to provide experimental support for the theoretical results.73 The exchange coupling between the high spin Fe(III) haem o3 and CuB(II) in quinol oxidase (cytochrome bo3) from Escherichia coli was studied by c.w. X-band EPR and variable temperature variable ﬁeld (VTVF) MCD by Thomson et al.74 Both parallel and perpendicular mode detection were used in the EPR spectra and spectrum simulations were used to show that the EPR responses could only arise from a weakly coupled (by anisotropic exchange) interaction with |J| B 1 cm1, which led to a ‘‘four-level’’ energy state model. Four aniontreated forms of cytochrome bo3 were studied: ﬂuoride, azide, chloride and formate as well as the oxidised form and a spin-Hamiltonian of the form: 2 2 2 2 ^ ¼½~ H g b H S^ þ DðS^z S^ =3Þ þ EðS^x S^y ÞFe

^ Cu þ S^Fe J~ S^Cu þ ½~ g b H S was employed. The EPR spectra in each case were very similar to each other, and it was suggested that overlap of magnetic orbitals between iron and copper was hindered and hence was the cause of the structural similarity in these forms. The authors went on to consider published magnetic and spectroscopic data for the coupled Fe-Cu pair in cytochrome c oxidase and concluded that the interpretation of the data requiring a strongly coupled pair with a resultant S 0 ¼ 2 ground state is not necessarily correct. In addition, it was suggested that the ‘‘four-level’’ model might have wider applicability for instance in the iron-nitrosyl (S ¼ 32) and semiquinone QAd radical in Photosystem II. Strong magnetic exchange leading to EPR-silent states as measured at X-band frequency has been reported for the coupled cytochrome b3 with CuB coupling in a NO reductase homologue from Roseobacter denitriﬁcans.75 6.4 Manganese (Excluding Photosystems). – Multifrequency (X-, Q- and W-band) measurements on the antiferromagnetically coupled mixed valence MnIII-MnIV and MnII-MnIII states of di-manganese catalase from Thermus thermophilus together with some related low molecular weight analogues have been reported as part of a special issue of Magnetic Resonance in Chemistry on High-ﬁeld EPR in Biology, Chemistry and Physics.76 Data analysis used a simpliﬁed spin-Hamiltonian to exchange couple the dimetallic: ^ ¼ 2J S^1 S^2 þ g~1 b H S^1 þ g~2 b H S^2 H _

_

_

_

þ S^1 a~1 I þ S^2 a~2 I 2 þ S^1 d~1 S 1 þ S^2 d~2 S 2 where the terms have their usual meaning, with lower case letters deﬁning single centre interactions. Also, a spin-coupled Hamiltonian was employed: _

_

^ ¼ G~1 b H S^ þ S^1 A~1 I þ S^2 A~2 I 2 H

178

Electron Paramag. Reson., 2007, 20, 157–191

where upper case letters represent resultant interactions derived from single centre interactions by spin projection factors deﬁned for a two-centre spin system as: S1 ðS1 þ 1Þ S2 ðS2 þ 1Þ þ SðS þ 1Þ and 2SðS þ 1Þ S2 ðS2 þ 1Þ S1 ðS1 þ 1Þ þ SðS þ 1Þ c2 ¼ 2SðS þ 1Þ

c1 ¼

and where: 1 c2 G ¼ c1 g1 þ c2 g2 þ c5J ðg1 g2 Þ½ð3c1 þ 1Þd1 ð3c2 þ 1Þd2 a1 A1 ¼ c1 a1 5J c1 c2 ½ð3c1 þ 1Þd1 ð3c2 þ 1Þd2 a2 A2 ¼ c2 a2 5J c1 c2 ½ð3c1 þ 1Þd1 ð3c2 þ 1Þd2

Iterative ﬁtting of simulated spectra across all three frequencies was used to obtain a ﬁnal set of spin-Hamiltonian parameters. The spectra for MnIIIMnIV (c1 ¼ 2 for S1 ¼ 2, and c2 ¼ 1 for S2 ¼ 32) all showed overall 16 line patterns and hence |J| c |d|, whereas for MnIIMnIII (c1 ¼ 7/3 for S1 ¼ 5/2, and c2 ¼ 4/3 for S2 ¼ 2) complex, broad spectra were reported. New data (tabulated below) and those from the literature were analysed and compared with the results of DFT calculations.

Enzyme

Anisotropy

g

MnIIIMnIV catalase

X Y Z

2.0048 2.0040 1.9876

MnIIMnIII catalase

X Y Z

1.935 1.995 2.016

A (MnIII)/ mT

A (MnIV)/ mT

15.2 14.7 10.7 A (MnII)/ mT 19.3 18.4 25.6

8.0 8.3 9.1 A (MnIII)/ mT 8.0 7.8 7.7

giso

Aiso (MnIII)/ mT

Aiso (MnIV)/ mT

1.9988

13.5

8.5

Aiso (MnII)/ mT

Aiso (MnIII)/ mT

21.2

7.8

1.982

The exchange coupling switched from the strong-exchange limit in MnIIIMnIV to one in which zero-ﬁeld splitting could not be ignored because there was only small exchange coupling for MnII-MnIII. Solvent-bridging in di-manganese(II) phophoesterase has been investigated by c.w. X-band EPR spectroscopy allied to spectrum simulation by assigning spectra to the resultant S 0 ¼ 2 state of an exchange coupled pair using Boltzmann weighting of transitions from a full matrix diagonalisation treatment of the spin-Hamiltonian: _

_

_

_

H ¼ 2JS1 S 2 þ S 1 d~12 S2 þ

2 X

_

_

~ Si ½S i D

i¼1 _

_

þ g~ b H S^i þ bn g~n H I i þ S^i A~i I i

Electron Paramag. Reson., 2007, 20, 157–191

179

An exchange coupling constant of J ¼ 2.7 0.2 cm1 was determined, consistent with a hydroxide and carboxylate bridge, and a rhombic zero-ﬁeld splitting (D ¼ 0.055 and E ¼ 0.0067 cm1) are reported. The sulfate-oxidising enzyme system in Paracoccus pantotrophus is comprised of four periplasmic proteins (SoxXA, SoxYZ, SoxB and SoxCD), mediating cytochrome c reduction dependent on hydrogen sulﬁde, sulfur, thiosulfate and sulﬁte, although separately each protein is inactive. A variable temperature X- and Q-band EPR study78 of the di-manganese SoxB protein by Lubitz et al. also used a spin-Hamiltonian of the form above relying on simulations with XSophe and Easyspin software as well as some in-house routines, and overall parameters were obtained from single ion values (for g, D and A interactions) by spin projection. The rich hyperﬁne structure (11 lines) was ﬁltered out from broad background resonances. The analysis was simpliﬁed by assuming that the g-value and zero-ﬁeld splitting contributions of the two manganese ions were identical, that the g-value was isotropic (reﬁned to 1.9 0.01), that the dipole tensor (d~12 ) was axial with its largest component co-parallel with the Mn–Mn direction, and that the hyperﬁne coupling matrix was isotropic (8.12 0.1 mT). Features were identiﬁed for each of the resultant S 0 ¼ 1, S 0 ¼ 2 and S 0 ¼ 3 spin states and gave J ¼ 7 1 cm1. The single ion zero-ﬁeld splitting tensor was set to have its largest principal component at an angle y to the Mn–Mn direction and the inter-metal distance was also reﬁned via the dipolar tensor, where: 2

b d~12 ¼ 3 r

3ðg1 rÞðrg2 Þ g~1 g~2 r2

whereupon the remaining parameters were determined as rMnMn ¼ 3.4 0.1 A˚, D ¼ 0.09 cm1, E/D ¼ 0 and y ¼ 151. 6.5 Diiron (Including 2Fe-2S). – The aldehyde oxidoreductase from Desulfovibrio gigas contains two [2Fe-2S]21/11 clusters and a molybdenum cofactor, and when all three sites are at paramagnetic oxidation levels there are splittings in the X- and Q-band EPR spectra resulting from inter-centre spinspin couplings.79 This study was facilitated by the observably different EPR linewidths of the three centres and their structural arrangement within the enzyme. The theory for this analysis followed from a previously published method by the same group in which local spins were considered to interact by a dipolar mechanism in which the magnetic moment on each individual metal site was used. Variable temperature measurements were needed to obtain regimes of T1 in which the additional dipolar splittings could be observed. Thus spin-spin interactions between the Mo centre and the proximal [2Fe-2S]11 cluster was only seen in the splitting of Mo(V) EPR signals. In combination with the reported X-ray crystal structure these EPR results were used to propose an electron-transfer pathway within the enzyme. The study by EPR spectroscopy of iron-sulfur clusters in the NADH:Ubiquinone oxidoreductase (complex I) of E. coli has continued (see also Section 6.6).

180

Electron Paramag. Reson., 2007, 20, 157–191

Uhlmann and Friedrich80 used site-directed mutants to show that the previously reported and spectroscopically unusual ‘‘N1c’’ with g-values ¼ 1.92, 1.94 and 2.00 does not derive from a new [2Fe-2S] cluster, but from cluster N1a. Verkhovskaya et al.81 propose that ‘‘N1c’’ is a degradation product of other [2Fe-2S] clusters in complex I. Daldal et al. used angular variation of X-band spectra of layered membrane samples of cytochrome bc1 from Rhodobacter capsulatus to investigate the [2Fe-2S] cluster, in particular to ascertain the orientation of the cluster in mutations that led to changes in mid-point potential for the redox active cluster.82 A model of the quinone oxidation site was produced from these EPR data. The S ¼ 12 signal from the mixed valence di-iron cluster in recombinant ferritin from Pyrococcus furiosus expressed in E. coli was used in a redox titration to determine mid-point potential.83 The state had all of its g-values below 2.0, did not saturate up to 200 mW and down to 9 K and broadened out to become undetectable above 20–25 K. However, a broad signal with unusual temperature dependence in the range 6.8 to 44 K also appeared at intermediate potentials, but was not seen using parallel mode detection. The signal was tentatively ascribed to a modiﬁed mixed valence di-iron centre resulting from ferritin biosynthesis in the heterologous host. The narrower signals were stated to deﬁne the stable prosthetic group of this protein and to be part of a three redox state cycle. Recombinant E. coli biotin synthase (BioB), which converts dethiobiotin into biotin by inserting a sulfur atom in a S-adenosylmethionine (SAM)-dependent reaction and contains a [2Fe-2S] cluster, is inactive in vitro without the addition of exogenous iron. Anaerobic reconstitution of BioB with Fe21 and S2 produces a form that contains both a [2Fe-2S] and a [4Fe-4S] cluster and a combination of EPR (X-band) and Mo¨ssbauer spectroscopies have been used to monitor turnover.84 The EPR data conﬁrmed that S ¼ 12 states of the Fe-S clusters are generated during turnover, and that radical production takes place. In addition cluster degradation of the [2Fe-2S] cluster was seen to accompany turnover. 6.6 Other Iron–Sulfur Centres. – The ‘‘radical-SAM’’ super-family of enzymes is thought to have more than 600 members and these contain iron-sulfur clusters and S-adeosylmethionine (SAM or AdoMet). Pyruvate formate-lyase (PFL), an enzyme involved in anaerobic glucose metabolism and involving a long lived (t12 Z 24 h under anaerobic conditions) glycyl radical, was chosen by Broderick et al.85 as the basis to make a generic study of this family by EPR and ENDOR spectroscopy, and using an activating system that included an activating enzyme (AE), SAM, Fe(II) and reduced ﬂavodoxin. No crystal structure had yet been reported for PFL-AE. c.w. X-band measurements were used to assess activity and to follow turnover under a variety of conditions, and Q-band pulsed ENDOR spectra were obtained as ﬁeld-frequency plots using isotopically enriched versions of SAM with 13 C, 17O and 15N. A proposed structural model and mechanism were built using a point dipole approximation to analyse the results in conjunction with published Mo¨ssbauer data, and in comparison with reported crystal structures

Electron Paramag. Reson., 2007, 20, 157–191

181

on three other members of the radical-SAM super-family. Thus a site-diﬀerentiated [4Fe-4S] cluster is proposed to be chelated at one iron by a carboxyl oxygen and amino nitrogen of SAM to give a ﬁve-membered ring, and the sulfur atom of the ligand interacts with a bridging sulﬁde bonded to the iron in the cluster. Also within this super-family, c.w. X-band EPR spectra at 13 K showed that lipoyl synthase from E. coli binds two [4Fe-4S] clusters, distinguished by their g-values, but Mo¨ssbauer data were not able to achieve this distinction.86 The proton-pumping NADH-quinone oxidoreductase from E. coli contains nine iron-sulfur clusters of which eight are found in its mitochondrial counterpart, complex I. These clusters can be individually deactivated by mutation of the binding motif using replacement of four cysteine residues with alanine. Cicchillo et al.87 used reducing conditions to produce EPR signals in the ‘‘g ¼ 2’’ region, and determined that a set of rhombic signals at g-values 1.91, 1.94 and 2.05, visible between 6 and 20 K, were assigned to cluster N1c and hence conﬁrmed that it is a [4Fe-4S] cluster that ligates to the N1c motif, rather than a [2Fe-2S] cluster as had been reported elsewhere. Work on complex I by Verkhovskaya et al.88 followed activation by detergent and phospholipids by c.w. X-band EPR spectroscopy and suggested conformational rearrangements were involved and associated with two [4Fe-4S] clusters. The characteristic S ¼ 12 EPR signal for [3Fe-4S] clusters was used to identify an unexpected variation within three minor Fe-S proteins, designated V-I, V-II and V-III, found whilst purifying recombinant Bacillus thermoproteolyticus ferredoxin (BtFd) from E. coli.89 A crystal structure at 1.6 A˚ resolution of the BtFd-CoA complex (V-II) showed that each complex in the asymmetric unit comprised a [3Fe-4S] cluster coordinated by 3 cysteines, and that the polypeptide chain was superimposable on the original [4Fe-4S] cluster of BtFd. However, the fourth cysteine of this binding domain was rotated away towards the surface of the protein and formed a disulﬁde bond with the terminal sulfhydryl group of CoA. Hagen et al.90 have performed a detailed multi-frequency (X-, Q- and D-bands) EPR study on the [4Fe-4S]31 oxidation level (S ¼ 12) of a series of seven high potential iron-sulfur proteins (HiPIPs) from six microbial sources. The g-values were found to be frequency independent, but the form of the spectra was subject to significant g-strain broadening, and rapid-passage effects dominated the Q- and D-band spectra. Multi-component g-strain simulations at X-band identiﬁed three to four discernible species, which were assigned to valence isomers, i.e. the mixed valence pair [FeII–FeIII] can be in any one of six positions in the cluster. The results were compared to a model in the literature proposed from paramagnetic NMR studies. Mutants of the [4Fe-4S]-containing Ech hydrogenase from Methanosarcina barkeri were probed by c.w. X-band EPR spectroscopy at 10 K in a study91 of the S ¼ 12 signals of the reduced clusters. The spectra were used to identify the cluster giving a g ¼ 1.89 signal as the proximal cluster located in EchC and the g ¼ 1.92 signal arose from one of the clusters of subunit EchF. Barton et al.92 used c.w. X-band EPR at 10 K in conjunction with the results of

182

Electron Paramag. Reson., 2007, 20, 157–191

electrochemistry on DNA-modiﬁed electrodes to study DNA repair glycosylases containing [4Fe-4S] clusters. A functional role for the iron-sulfur clusters was established and characteristic EPR signals identiﬁed that both [4Fe-4S]31 and [3Fe-4S]11 were present. A multi-spectroscopic approach showed that L-serine deaminase from E. coli uses a [4Fe-4S] cluster, and the dithionitereduced reconstituted protein has typical parameters of: gJ ¼ 2.03 and g> ¼ 1.93, but treatment of the reduced protein with L-serine resulted in broadening of the former signal, indicative of a direct interactions of the amino acid with the cluster.93 Weiner et al. used EPR-monitored (X-band, liquid helium temperatures, S ¼ 12 signals) redox titrations in an investigation of the environment surrounding iron-sulfur cluster 4 of E. coli dimethyl sulfoxide reductase.94 The group showed that this cluster is the ﬁrst electron acceptor in the relay deﬁned by four such clusters and that the [3Fe-4S] form of this cluster can act as a barrier to electron transfer by virtue of its high redox potential. Weiner and co-workers have also used EPR spectroscopy, redox potentiometry and protein crystallography to characterise the novel [4Fe-4S] cluster (designated FS0) of nitrate reductase A from E. coli,95 which contains one histidine and three cysteine ligands. At temperatures of less than 15 K at X-band frequency EPR signals at effective g-values 5.023 and 5.556 were detected, which varied as a function of redox potential, and the midpoint (standard) potential of both signals was 55 mV at pH ¼ 8.0. The signals were assigned to a S ¼ 32 ground state of a [4Fe-4S]11 cluster. The characterisation of a new formate dehydrogenase (FDH) from the sulfate-reducing organism Desulfovibrio alaskensis NCIMB 13491 by EPR spectroscopy following dithionite reduction showed two ferredoxins with sets of g-values of: 1.882, 1.946 and 2.046 (cluster I) and 1.868, 1.931 and 2.066 (cluster II), with evidence for additional iron-sulfur clusters.96 57 Fe-pulsed Davies-ENDOR at X- and W-bands have been used to probe the [4Fe-4S] cluster in the heterodisulﬁde reductase (Hdr) from Methanothermobacter marburgensis, which catalyzes the reversible two electron reduction of the mixed disulﬁde between coenzyme M and coenzyme B.97 Measurements on Hdr-CoM indicated that the cluster was oxidised, i.e. [4Fe-4S]31, and that there was direct interaction of the cluster with the substrate. The results were contrasted with those reported elsewhere for a ferredoxin:thioredoxin reductase (FTR) that also catalyses a disulﬁde cleavage, where the latter enzyme has significant anisotropy of 57Fe hyperﬁne coupling at the unique iron site, but also gav 4 2.0 is reported for FTR (typical of [4Fe-4S]31 clusters) whereas gav o 2.0 for Hdr-CoM. In addition, it was noted that the cluster-binding cysteines were not conserved between the two proteins and possible analogies to the multiple binding sites found in radical-SAM enzymes were also suggested. Walters and Johnson have reviewed the structural and spectroscopic studies on ferredoxin:thioredoxin reductase,98 collecting together data from electronic absorption, EPR, ENDOR, VT-MCD, resonance Raman and Mo¨ssbauer spectroscopies. Mechanisms for the catalytic cycles of FTR and

Electron Paramag. Reson., 2007, 20, 157–191

183

Hdr were suggested and open questions were posed and future prospects for research were identiﬁed. Dual mode c.w. X-band spectra at liquid helium temperatures were used as part of the characterisation of MOCS1A, an oxygen-sensitive iron-sulfur protein involved in human molybdenum cofactor biosynthesis.99 In addition to signals in the ‘‘g ¼ 2’’ region, signals at effective g-values 9.4 in both modes of detection were assigned as being characteristic of a S ¼ 2 [3Fe-4S]0 cluster. A [4Fe-4S] cluster with g-values 2.067, 1.933 and 1.89 measured at X-band at 20 K was identiﬁed from isolation of the photosynthetic reaction centre Heliobacterium modesticaldum (HbRC) and also in whole cells and isolated membranes.100 Illumination produced a complex spectrum consistent with exchange interaction between two iron-sulfur clusters. 6.7 Photosystems. – Britt et al. have reviewed101 recent pulsed EPR studies of the Photosystem II oxygen-evolving complex (PSII OEC) from a mechanistic viewpoint and begin their review acknowledging the debt owed to Jerry Babcock by workers in the ﬁeld of EPR of PSII, noting his guidance, enthusiasm and significant contributions to the area. Substrate and cofactor interactions were highlighted in the context of proposed mechanisms of action, with a suggestion that substrate water molecules bind early in the S-state cycle, for which pulsed EPR data are limited to the lower S-states, S0–S2. Blondin et al. report on c.w. X-band spectra of IR sensitive, untreated PSII and MeOH- and NH3-treated PSII from spinach in the S2 state.102 Spectrum simulation used co-parallel g- and Mn-hyperﬁne matrices to obtain principal values that indicated a 1 MnIII-3 MnIV cluster and these were in good agreement with previously published values. A magnetic exchange coupling scheme involving four J-values was explored against derived spin densities to yield antiferromagnetic interactions with values between 290 and 130 cm1, giving a S ¼ 12 ground state and ﬁrst excited state of S ¼ 52 at þ30 cm1 responsible for a geﬀ ¼ 4.1 signal. c.w. EPR, electron-spin-echo (ESE)-detected ﬁeld swept and electron spin echo envelope modulation (ESEEM) spectra at X-band frequency were reported on the effect of adding isotopically labelled azide (15N-14N-N) to PSII membrane samples (BBY preparation) with or without chloride.103 Analysis of the results showed that a binding site close to the manganese cluster was competitive for azide and chloride, implying that chloride is bound proximally to the cluster, and it was also shown that inhibition induced by azide could be partially recovered by addition of bicarbonate. Britt and co-workers have also demonstrated, by c.w. EPR and ESEEM spectroscopies, that the presence or absence of the 33 KDa manganese-stabilising protein, which limits access of small molecules to the metal site, bound to the lumenal side of PSII close to the Mn4Ca cluster, has no effect in the number or distance of deuterons magnetically coupled to manganese after equilibration with D2O.104 The samples of PSII were trapped in the dark stable S1 state prior to illumination at 200 K for

184

Electron Paramag. Reson., 2007, 20, 157–191

ESEEM measurements. Analysis of modulation depths led to estimation of distance between deuteron and manganese of 2.5–2.6 A˚. Glycerol-treated samples were also investigated, which led to changes in the ‘‘g ¼ 4.1’’ signal, but it was shown that unlike treatment with smaller alcohols such as MeOH, there was no direct binding of glycerol to the manganese site. It was suggested that the 33 KDa protein might control access to higher S-state transitions in order to manage possible side reactions. This group has also investigated the calcium-binding site of the OEC of PSII using 87Sr ESEEM spectroscopy in conjunction with spectrum simulation and DFT calculations.105 A stimulated echo modulation function for the I ¼ 9/2 nucleus in the absence of nuclear quadrupole effects was derived following the method of Dikanov and coworkers. However, the analysis was extended using a numerical diagonalisation procedure to include a non-zero quadrupolar interaction, the results of which were compared directly with quadrupole couplings calculated by DFT methods, for both 43Ca and 87Sr. It was concluded that the DFT method overestimated these couplings by up to a factor of two. Use of a simple point dipole approximation on the modulation depth gave a Mn-Sr(Ca) distance of 4.5 A˚, which was converted to a range of 3.8 to 5.0 A˚, if a tetrametallic magnetic model for the cluster was considered. However, this range deﬁnes an upper bound because the effect of nuclear quadrupole coupling was not included in those calculations. Q-band pulsed ESE-detected ﬁeld swept and inversion-recovery spectra on PSII membranes have been used to measure the electron spin-lattice relaxation times, T1, for the S0 state of the OEC, in comparison with two di-manganese complexes.106 The relaxation times in the S0 state were measured between 4.3 to 6.5 K and indicated that an Orbach mechanism exists with a ﬁrst excited state at 21.7 0.4 cm1 above the ground state. 55Mn pulsed ENDOR spectra at Q-band frequency have been reported for both S2 and S0 states,107 and these are taken to indicate that all four manganese ions are magnetically coupled. Principal, absolute values for the hyperﬁne interaction matrices in the two states are given below: S2 state MnA MnB MnC MnD

S0 state

A>/MHz

AJ/MHz

A>/MHz

AJ/MHz

235 185 310 175

265 245 265 230

270 190 320 170

200 280 400 240

The similarity in magnitude between the two sets of values led to the suggested assignment of Mn4(III, III, III, IV) and Mn4(III, IV, IV, IV) for the S0 and S2 states, respectively. A series of inversion-recovery and microwave power saturation measurements on the tyrosine YD radical of PSII between 4 and 25 K have been

185

Electron Paramag. Reson., 2007, 20, 157–191

reported by a different group.108 The dark stable S1 and ﬁrst turnover S2 states were studied in sucrose and in ethyleneglycol-glycerol as cryo-protectants. A through-space dipolar relaxation model was shown to dominate and was only slightly temperature-dependent and was independent of S-state. A background exponential decay was an order of magnitude weaker and was temperature and S-state dependent, consistent with an interaction of YDd and the manganese cluster over a range of ca. 30 A˚. X- and W-band c.w. EPR spectra at ca. 10 K have been used to probe the tyrosine Z radical and manganese cluster in the S1 state.109 Characteristic resonances with g-values of 2.019, B2.00 and 1.987 were deﬁned at W-band and were simulated by assuming that the manganese cluster had two low lying states, S ¼ 1 and S ¼ 0, with the former interacting with the S ¼ 12 state of the radical. Simultaneous ﬁtting at both frequencies gave the following parameters (estimated errors on the last digit in parentheses) for the manganese cluster and for the ferromagnetic exchange with the radical: gx ¼ 2.0055(3), gy ¼ 2.0043(2), gz ¼ 1.992(3), D ¼ 0.048(2) cm1, l ¼ 0, Jx ¼ 0.087(2) cm1, Jy ¼ Jz ¼ 0 cm1. The extracted parameters were of similar magnitude and so the authors stated that further data are required in order to deﬁne the various interactions with more certainty. A series of pulsed X-band experiments, including 55Mn-ELDOR on the S2state of the OEC of PSII have been reported,110 and the broad peak found in the double frequency experiment is consistent, by spectrum simulation, with manganese hyperﬁne interactions determined from 55Mn-ENDOR, and these are tabulated below, and show good agreement with the results given in reference 107.

MnA MnB MnC MnD

A>/MHz

AJ/MHz

PJ/MHz

232 200 311 180

270 250 270 240

3 3 8 1

The hyperﬁne and nuclear quadrupole interactions were strongly correlated such that moving within a range of 10 MHz about the tabulated A-values could be compensated by changes in the P parameters and maintain the same degree of correspondence between simulated and experimental spectra. A random acquisition technique was used to measure the ENDOR response, wherein the radio-frequency ﬁeld was not swept linearly, but was varied randomly within the required range in order to decrease local heating. The heterogeneous nature of the multi-line EPR spectrum of the S2-state of the OEC of PSII was shown to be manifest in two forms of EPR signal, and significantly it was demonstrated that MeOH favoured the narrow form on ﬁrst turnover of the enzyme, as determined by c.w. X- band and Q-band EPR spectroscopy.111 The effect was removed on subsequent turnovers and led to a

186

Electron Paramag. Reson., 2007, 20, 157–191

predominant broad form of the multi-line spectrum. ESEEM characterisation of the narrow form of the spectrum was also reported, and a structural model was proposed for the MeOH binding to the OEC. X-band ESEEM spectroscopy of the S2-state in the presence of H217O (ﬁnal enrichment of sample 20-30%) showed modulations around the 17O Larmor frequency at ca. 2 MHz and further features up to ca. 18 MHz assigned to 17O, but only observed when the applied magnetic ﬁeld was at the centre of the multi-line spectrum between effective g-values of 1.98 and 1.87.112 Analysis of the data gave Aiso B 5 MHz and a quadrupole interaction 46 MHz, and it was suggested that this meant that water, rather than OH, was the bound species. c.w. and ELDOR studies have been reported on a mutant in which a glutamate proposed to act as a direct ligand to the manganese cluster was replaced by a glutamine.113 The S2-state spectra and the S2-state to YDd distance determined by ELDOR were unaffected by the modiﬁcation, although redox potential and FTIR spectral data were changed relative to those of the unmodiﬁed OEC.

6.8 Nickel. – HYSCORE and ENDOR spectra at X-band frequency at liquid helium temperatures on two states, Ni-A (‘‘unready’’) and Ni-B (‘‘ready’’) of the [Ni-Fe] hydrogenase from Desulfovibrio vulgaris Miyazaki F, isolated and puriﬁed under aerobic conditions, showed that both have an exchangeable proton, but with different hyperﬁne coupling.114 DFT calculations were used in the analysis and the proton was assigned to a bridging hydroxide ligand between Ni and Fe in Ni-B and either a differently bonded hydroxide or a hydroperoxide in Ni-A. A similar and very detailed experimental study of the reduced Ni-C (‘‘active’’) state has been reported by the same group,115 and the problem of ambiguity of signs (in the choice of direction cosines) for the location of the non-exchangeable protons was discussed. Two strongly coupled protons were assigned to b-CH2 of a ligating cysteine residue and a third, slightly more weakly coupled proton, was assigned to a different cysteinyl b-CH2. D2O exchange experiments showed the presence of hydride in the bridge between the iron and nickel centres. 14N interactions were also detected and these were assigned to a histidine residue hydrogen-bonded to the cysteine with the strongly coupled (to the unpaired electron) protons. S-, X-, Q- and D-band c.w. EPR spectra were used to examine the COdehydrogenase/acetyl-CoA synthase enzyme from Moorella thermoacetica, for which the identity of essential metal ions was not clear.116 Ni, Cu and Zn had previously been implicated as having active roles as part of a heterodimetallic site, with proximal and distal positions relative to bridging to a [4Fe-4S] cluster. Using enzyme preparations with a wide range of Ni and Cu stoichiometries per dimetallic unit, followed by removal of the metal from the proximal site using selective chelators and then monitoring by EPR and EXAFS spectroscopy led to a consensus view of the active site. It was reported that the ‘‘NiFeC’’ EPR signal derived from a [4Fe-4S]21 (S ¼ 0) cluster bridged to Ni11 (S ¼ 12) that was itself bridged to a planar Ni21 (S ¼ 0).

Electron Paramag. Reson., 2007, 20, 157–191

187

6.9 Miscellaneous. – Ethanolamine deaminase from Salmonella typhimurium is coenzyme B12-dependent and catalyses elimination of ammonia from aminoethanol to yield ethanal via production of a low spin CoII-cobalamin and a carbon-centred deoxyadenosyl radical pair.117 c.w. EPR and ESEEM spectroscopy at X-band frequency on the radical pair have been used to give a partial structural model for the active site and for the hydrogen atom abstraction event in the catalysis. References 1. D. Collison and E.J.L. McInnes, Specialist Periodical Reports Electron Spin Resonance Series, B.C. Gilbert, M.J. Davies and D.M. Murphy (eds), Royal Society of Chemistry, Cambridge, 2002, 18, 161. 2. D. Collison and E.J.L. McInnes, Specialist Periodical Reports Electron Spin Resonance Series, B.C. Gilbert, M.J. Davies and D.M. Murphy (eds), Royal Society of Chemistry, Cambridge, 2004, 19, 374. 3. G. Zoppellaro, V. Enkelmann, A. Geies and M. Baumgarten, Org. Lett., 2004, 6, 4929. 4. K. Komaguchi, T. Iida, Y. Goh, J. Oshita, A. Kanui and M. Shiotani, Chem. Phys. Lett., 2004, 387, 327. 5. R. Ziessel, C. Stroh, H. Heise, F.H. Ko¨hler, P. Yurek, N. Claiser, M. Souhassou and C. Lecomte, J. Am. Chem. Soc., 2004, 126, 12604. 6. C. Stroh, M. Mayor and C. von Hanisch, P Turek, Chem. Commun., 2004, 2050. 7. K. Maekawa, D. Shiomi, T. Ise, K. Sato and T. Takui, J. Phys. Chem. B, 2005, 109, 3303. 8. L. Catala, J. Le Moigne, N. Gruber, J.J. Novoa, P. Rabu, E. Belorizky and P. Turek, Chem. Eur. J., 2005, 11, 2440. 9. C. Rancurel, H. Heise, F.H. Ko¨hler, U. Schatzschneider, E. Rentschler, J. VidalGancedo, J. Veciana and J.-P. Sutter, J. Phys. Chem. A, 2004, 108, 5903. 10. O.B. Borobia, P. Gionneau, H. Heise, F.H. Ko¨hler, L. Ducasse, J. Vidal-Gancedo, J. Veciana, S. Golhen, L. Ouahab and J.-P. Sutter, Chem. Eur. J., 2005, 11, 128. 11. G.A. Carriedo, F.J. Garcia Alonso, P.G. Elipe, E. Brillas, A. Labarta and L. Julia´, J. Org. Chem., 2004, 69, 99. 12. A. Rajca, K. Shiraishi, M. Vale, H. Han and S. Rajca, J. Am. Chem. Soc., 2005, 127, 9014. 13. C. Elschenbroich, J. Plackmeyer, M. Nowotny, K. Harms, J. Pebler and O. Burghaus, Inorg. Chem., 2005, 44, 955. 14. C. Elschenbroich, J. Plackmeyer, M. Nowotny, A. Behrendt, K. Harms, J. Pebler and O. Burghaus, Chem. Eur. J., 2005, 11, 7427. 15. C. Elschenbroich, O. Schiemann, O. Burghaus and K. Harms, Chem. Commun., 2005, 2149. 16. C. Elschenbroich, M. Wu¨nsch, A. Behrend, B. Metz, B. Neumu¨ller and K. Harms, Organometallics, 2005, 24, 5509. 17. W.W.H. Wong, J. Cookson, E.A.L. Evans, E.J.L. McInnes, J. Wolowska, J.P. Maher, P. Bishop and P.D. Beer, Chem. Commun., 2005, 2214. 18. A.D. Bond, S. Derossi, C.J. Harding, E.J.L. McInnes, V. McKee, C.J. McKenzie, J. Nelson and J. Wolowska, Dalton Trans., 2005, 2403. 19. J. Yano, K. Sauer, J.-J. Girerd and V.K. Yachandra, J. Am. Chem. Soc., 2004, 126, 7486.

188

Electron Paramag. Reson., 2007, 20, 157–191

20. X. Liu, M.P. de Miranda, E.J.L. McInnes, C.A. Kilner and M.A. Halcrow, Dalton Trans., 2004, 59. 21. J. Yoon, L.M. Mirica, T.D.P. Stack and E.I. Solomon, J. Am. Chem. Soc., 2004, 126, 12586. 22. J. Yoon and E.I. Solomon, Inorg. Chem., 2005, 44, 8076. 23. I. Gautier-Luneau, D. Phanon, C. Duboc, D. Luneau and J.-P. Pierre, Dalton Trans., 2005, 3795. 24. O. Kahn, Chem. Phys. Lett., 1997, 265, 109. 25. T. Glaser, M. Heidmeier, S. Grimme and E. Bill, Inorg. Chem., 2004, 43, 5192. 26. S. Nellutla, J. Van Tol, N.S. Dalal, L.-H. Bi, U. Kortz, B. Keita, L. Nadjo, G.A. Khitrov and A.G. Marshall, Inorg. Chem., 2005, 44, 9795. 27. X. Liu, J.A. McAllister, M.P. de Miranda, E.J.L. McInnes, C.A. Kilner and M.A. Halcrow, Chem. Eur. J., 2004, 10, 1827. 28. B. Pilawa, R. Bofﬁnger, I. Keilhauer, R. Leppin, I. Odenwald, W. Wendl, C. Berthier and M. Horvatic´, Phys. Rev. B, 2005, 71, 184419. 29. O. Cador, D. Gatteschi, R. Sessoli, A.-L. Barra, G.A. Timco and R.E.P. Winpenny, J. Magn. Magn. Mat., 2005, 290–291, 55. 30. S. Sharmin, A. Ardavan, S.J. Blundell, A.I. Coledea, E.J.L. McInnes and D. Low, Appl. Phys. Lett., 2005, 86, 032507. 31. H. Oshio, M. Nihei, A. Yoshida, H. Nohiri, M. Nakano, A. Yamaguchi, Y. Karaki and H. Ishimoto, Chem. Eur. J., 2005, 11, 843. 32. C. Kirman, J. Lawrence, S. Hill, E.-C. Yang and D.N. Hendrickson, J. Appl. Phys., 2005, 97, 10M501. 33. D.N. Hendrickson, E.-C. Yang, R.M. Isidro, C. Kirman, J. Lawrence, R.S. Edwards, S. Hill, A. Yamaguchi, H. Ishimoto, W. Wernsdorfer, C. Ramsey, N. Dalal and M.M. Olmstead, Polyhedron, 2005, 24, 2280. 34. A. Cornia, A.C. Fabretti, P. Garrisi, C. Mortalo´, D. Bonacchi, D. Gatteschi, R. Sessoli, L. Sorace, W. Wernsdorfer and A.-L. Barra, Angew. Chem. Int. Ed., 2004, 43, 1136. 35. A. Cornia, A.C. Fabretti, P. Garrisi, C. Mortalo´, D. Bonacchi, R. Sessoli, L. Sorace, A.-L. Barra and W. Wernsdorfer, J. Magn. Magn. Mat., 2004, 272–276, e749. 36. N. Aliaga-Alcalde, R.S. Edwards, S.O. Hill, W. Wernsdorfer, K. Folting and G. Christou, J. Am. Chem. Soc., 2004, 126, 12503. 37. J. van Slageren, S. Vontragool, B. Gorshunov, A.A. Mukhin, N. Karl, J. Krzystek, J. Telser, A. Mu¨ller, C. Sangregorio, D. Gateschi and M. Dressel, Phys. Chem. Chem. Phys., 2003, 5, 3837. 38. S. Piligkos, G. Rajaraman, M. Soler, N. Kirchner, J. Van Slageren, R. Bircher, S. Parsons, H.-U. Gu¨del, J. Kortus, W. Wernsdorfer, G. Chrsitou and E.K. Brechin, J. Am. Chem. Soc., 2005, 127, 5572. 39. N. Kirchner, J. van Slageren, E.K. Brechin and M. Dressel, Polyhedron, 2005, 24, 2400. 40. S. Vongtragool, B. Gorshunov, J. van Slageren, M. Dressel and A.A. Mukhin, J. Magn. Magn. Mat., 2004, 272–276, e769. 41. S. Vongtragool, A. Mukhin, B. Gorshunov and M. Dressel, Phys. Rev. B, 2004, 69, 104410. 42. A. Cornia, R. Sessoli, L. Sorace, D. Gatteschi, A.-L. Barra and C. Daiguebonne, Phys. Rev. Lett., 2002, 89, 257201. 43. S. Takahashi, R.S. Edwards, J.M. North, S. Hill and N.S. Dalal, Phys. Rev. B, 2004, 70, 094429.

Electron Paramag. Reson., 2007, 20, 157–191

189

44. S. Hill, N. Anderson, A. Wilson, S. Takahashi, K. Petukhov, N.E. Chakov, M. Murugesu, J.M. North, E. del Barco, A.D. Kent, N.S. Dalal and G. Christou, Polyhedron, 2005, 24, 2284. 45. S. Hill, N. Anderson, A. Wilson, S. Takahashi, N.E. Chakov, M. Murugesu, J.M. North, N.S. Dalal and G. Christou, J. Appl. Phys., 2005, 97, 10M510. 46. K. Petukhov, S. Hill, N.E. Chakov, K.A. Abboud and G. Christou, Phys. Rev. B, 2004, 70, 054426. 47. E.M. Rumberger, E. Del Barco, J. Lawrence, S. Hill, A.D. Kent, L.N. Zakharov, A.L. Rheingold and D.N. Hendrickson, Polyhedron, 2005, 2557. 48. H. Hachisuka, K. Awaga, T. Yokoyama, T. Kubo, T. Goto and H. Nojiri, Phys. Rev. B, 2004, 70, 104427. 49. S. Piligkos, D. Collison, V.S. Oganesyan, G. Rajaraman, G.A. Timco, A.J. Thomson, R.E.P. Winpenny and E.J.L. McInnes, Phys. Rev. B, 2004, 69, 134424. 50. K. Petukhov, W. Wernsdorfer, A.-L. Barra and V. Mosser, Phys. Rev. B, 2005, 72, 052401. 51. B. Cage, S.E. Russek, D. Zipse and N.S. Dalal, J. Appl. Phys., 2005, 97, 10M507. 52. B. Cage, S.E. Russek, D. Zipse, J.M. North and N.S. Dalal, Appl. Phys. Lett., 2005, 87, 082501. 53. K. Chlopek, E. Bill, T. Weyhermu¨ller and K. Wieghardt, Inorg. Chem., 2005, 44, 7087. 54. K. Ray, E. Bill, T. Weyhermu¨ller and K. Wieghardt, J. Am. Chem. Soc., 2005, 127, 5641. 55. F. Thomas, O. Jarjayes, C. Duboc, C. Philouze, E. Saint-Aman and J.-L. Pierre, Dalton Trans., 2004, 2662. 56. M.L. Kirk, D.A. Shultz and E.C. Depperman, Polyhedron, 2005, 24, 2880. 57. D.S. Marlin, E. Bill, T. Weyhermu¨ller, E. Bothe and K. Wieghardt, J. Am. Chem. Soc., 2005, 127, 6095. 58. Y. Yukawa, G. Aromi, S. Igarashi, J. Ribas, S.A. Zvyagin and J. Krzystek, Angew. Chem. Int. Ed., 2005, 44, 1997. 59. C. Elsaesser, M. Brecht and R. Bittl, Biochem. Soc., 2004, 33, 15. 60. T. Maly, F. MacMillan, K. Zwicker, N. Kashani-Poor, U. Brandt and T.F. Prisner, Biochemistry, 2004, 43, 3969. 61. P.C. Dos Santos, R.Y. Igarashi, H.-I. Lee, B.M. Hoﬀman, L.C. Seefeldt and D.R. Dean, Acc. Chem. Res., 2005, 38, 208. 62. T.-C. Yang, N.K. Maeser, M. Laryukhin, H.-I. Lee, D.R. Dean, L.C. Seefeldt and B.M. Hoﬀman, J. Amer. Chem. Soc., 2005, 127, 12804. 63. H.-I. Lee, R.Y. Igarashi, M. Laryukhin, P.E. Doan, P.C. Dos Santos, D.R. Dean, L.C. Seefeldt and B.M. Hoﬀman, J. Amer. Chem. Soc., 2004, 126, 9563. 64. B.M. Barney, T.-C. Yang, R.Y. Igarashi, P.C. Dos Santos, M. Laryukhin, H.-I. Lee, B.M. Hoﬀman, D.R. Dean and L.C. Seefeldt, J. Amer. Chem. Soc., 2005, 127, 14960. 65. B.M. Barney, M. Laryukhin, R.Y. Igarashi, H.-I. Lee, P.C. Dos Santos, T.-C. Yang, B.M. Hoﬀman, D.R. Dean and L.C. Seefeldt, Biochemistry, 2005, 44, 8030. 66. R.Y. Igarashi, M. Laryukhin, P.C. Dos Santos, H.-I. Lee, D.R. Dean and L.C. Seefeldt, and B.M. Hoﬀman, J. Amer. Chem. Soc., 2005, 127, 6231. 67. K. Fisher and W.E. Newton, Biochim. Biophys. Acta, 2005, 1750, 154. 68. Z. Maskos, K. Fisher, M. Sørlie, W.E. Newton and B.J. Hales, J. Biol. Inorg. Chem., 2005, 10, 394. 69. Y. Hu, M.C. Corbett, A.W. Fay, J.A. Webber, B. Hedman and K.O. Hodgson, Proc. Nat. Acad. Sci., 2005, 102, 13825.

190

Electron Paramag. Reson., 2007, 20, 157–191

70. J. Petersen, C. Gessner, K. Fisher, C.J. Mitchell, D.J. Lowe and W. Lubitz, Biochem. J., 2005, 391, 527. 71. I. Galli, G. Musci and M.C. Bonaccorsi di Patti, J. Biol. Inorg. Chem., 2004 9, 90. 72. L. Quintanar, C. Stoj, T.-P. Wang, D.J. Kosman and E.I. Solomon, Biochemistry, 2005, 44, 6081. 73. L. Quintanar, J. Yoon, C.P. Aznar, A.E. Palmer, K.K. Andersson, R.D. Britt and E.I. Solomon, J. Amer. Chem. Soc., 2005, 127, 13832. 74. M.R. Cheesman, V.S. Oganesyan, N.J. Watmough, C.S. Butler and A.J. Thomson, J. Amer. Chem. Soc., 2004, 126, 4157. 75. Y. Matsuda, T. Uchida, H. Hori, T. Kitagawa and H. Arata, Biochim. Biophys. Acta, 2004, 1656, 37. 76. C. Teutloﬀ, K.-O. Scha¨fer, S. Sinnecker, V. Barynin, R. Bittl, K. Weighardt, F. Lendzian and W. Lubitz, Magn. Reson. Chem., 2005, 43, S51. 77. C.R. Samples, T. Howard, F.M. Raushel and V.J. DeRose, Biochemistry, 2005, 44, 11005. 78. B. Epel, K.-O. Scha¨fer, A. Quentmeier, C. Friedrich and W. Lubitz, J. Biol. Inorg. Chem., 2005, 10, 636. 79. C. More, M. Asso, G. Roger, B. Guigliarelli, J. Caldeira, J. Moura and P. Betrand, Biochemistry, 2005, 44, 11628. 80. M. Uhlmann and T. Friedrich, Biochemistry, 2005, 44, 1653. 81. M. Verkhovskaya, L. Sinegina, M. Wikstro¨m and M. Verkhovsky, Biochim. Biophys. Acta, 2004, 1658, 141. 82. J.W. Cooley, A.G. Roberts, M.K. Bowman, D.M. Kramer and F. Daldal, Biochemistry, 2004, 43, 2217. 83. J. Tatur and W.R. Hagen, FEBS Lett., 2005, 579, 4729. 84. G.N.L. Jameson, M.M. Cosper, H.L. Herna´ndez, M.K. Johnson and B.H. Huynh, Biochemistry, 2004, 43, 2022. 85. C.J. Walsby, D. Ortillo, J. Yang, M.R. Nnyepi, W.E. Broderick, B.M. Hoﬀman and J.B. Broderick, Inorg. Chem., 2005, 44, 727. 86. E. Nakamura-Ogiso, T. Yano, T. Yagi and T. Ohnishi, J. Biol. Chem., 2005, 280, 301. 87. R.M. Cicchillo, K.-H. Lee, C. Baleanu-Gogonea, N.M. Nesbitt, C. Krebs and S.J. Booker, Biochemistry, 2004, 43, 11770. 88. L. Sinegina, M. Wikstro¨m, M.I. Verkhovsky and M.L. Verkhovskaya, Biochemistry, 2005, 44, 8500. 89. T. Shirakawa, Y. Takahashi, K. Wada, J. Hirota, T. Takao, D. Ohmori and K. Fukuyama, Biochemistry, 2005, 44, 12402. 90. A.H. Priem, A.A.K. Klaassen, E.J. Reijerse, T.E. Meyer, C. Luchinat, F. Capozzi, W.R. Dunham and W.R. Hagen, J. Biol. Inorg. Chem., 2005, 10, 417. 91. L. Forzi, J. Koch, A.M. Guss, C.G. Radosevich, W.W. Metcalf and R. Hedderich, FEBS J., 2005, 272, 4741. 92. A.K. Boal, E. Yavin, O.A. Lukianova, V.L. O’Shea, S.S. David and J.K. Barton, Biochemistry, 2005, 44, 8397. 93. R.M. Cicchillo, M.A. Baker, E.J. Schnitzer, E.B. Newman, C. Krebs and S.J. Booker, J. Biol. Chem., 2004, 279, 32418. 94. V.W.T. Cheng, R.A. Rothery, M.G. Bertero, N.C.J. Strynadka and J.H. Weiner, Biochemistry, 2005, 44, 8068. 95. R.A. Rothery, M.G. Bertero, R. Cammack, M. Palak, F. Blasco, N.C.J. Strynadka and J.H. Weiner, Biochemistry, 2004, 43, 5324.

Electron Paramag. Reson., 2007, 20, 157–191

191

96. C.D. Brondino, M.C.G. Passeggi, J. Caldeira, M.J. Almendra, M.J. Feio, J.J.G. Moura and I. Moura, J. Biol. Inorg. Chem., 2004, 9, 145. 97. M. Bennati, N. Weiden, K.-P. Dinse and R. Hedderich, J. Amer. Chem. Soc., 2004, 126, 8378. 98. E.M. Walters and M.K. Johnson, Photosynth. Res., 2004, 79, 249. 99. P. Ha¨nzelmann, H.L. Herna´ndez, C. Menzel, R. Garcı´ a-Serres, B.H. Huynh, M.K. Johnson, R.R. Mendel and H. Schindelin, J. Biol. Chem., 2004, 279, 34721. 100. M. Heinnickel, G. Shen, R. Agalarov and J.H. Golbeck, Biochemistry, 2005, 44, 9950. 101. R.D. Britt, K.A. Campbell, J.M. Peloquin, M.L. Gilchrist, C.P. Aznar, M.M. Dicus, J. Robblee and J. Messinger, Biochim. Biophys. Acta, 2004, 1655, 158. 102. M.-F. Charlot, A. Boussac and G. Blondin, Biochim. Biophys. Acta, 2005, 1708, 120. 103. H. Yu, P. Aznar, X. Xu and R.D. Britt, Biochemistry, 2005, 44, 12022. 104. W. Gregor, R.M. Cinco, H. Yu, V.K. Yachandra and R.D. Britt, Biochemistry, 2005, 44, 8817. 105. S.H. Kim, W. Gregor, J.M. Peloquin, M. Brynda and R.D. Britt, J. Amer. Chem. Soc., 2004, 126, 7228. 106. L.V. Kulik, W. Lubitz and J. Messinger, Biochemistry, 2005, 44, 9368. 107. L.V. Kulik, B. Epel, W. Lubitz and J. Messinger, J. Amer. Chem. Soc., 2005, 127, 2392. 108. F. Mamedov, P.J. Smith, S. Styring and R.J. Pace, Phys. Chem. Chem. Phys., 2004, 6, 4890. 109. D. Koulougliotis, C. Teutloﬀ, Y. Sanakis, W. Lubitz and V. Petrouleas, Phys. Chem. Chem. Phys., 2004, 6, 4859. 110. L. Kulik, B. Epel, J. Messinger and W. Lubitz, Photosynth. Res., 2005, 84, 347. 111. K.A. A˚rhling, M.C.W. Evans, J.H.A. Nugent and R.J. Pace, Biochim. Biophys. Acta, 2004, 1656, 66. 112. M.C.W. Evans, J.H.A. Nugent, R.J. Ball, I. Muhiuddin and R.J. Pace, Biochemistry, 2004, 43, 989. 113. Y. Kimura, N. Mizusawa, A. Ishii, S. Nakazawa and T.-a. Ono, J. Biol. Chem., 2005, 280, 37895. 114. M. van Gastel, C. Fichtner, F. Neese and W. Lubitz, Biochem. Soc. Trans., 2005, 33, 7. 115. S. Foerster, M. van Gastel, M. Brecht and W. Lubitz, J. Biol. Inorg. Chem., 2005, 10, 51. 116. J. Seravalli, Y. Xiao, W. Gu, S.P. Cramer, W.E. Antholine, V. Krymov, G.J. Gerfen and S.W. Ragsdale, Biochemistry, 2004, 43, 3944. 117. J.M. Canﬁeld and K. Warncke, J. Phys. Chem. B, 2005, 109, 3053.

Biological Free Radicals and Biomedical Applications of EPR Spectroscopy BY SIMON K. JACKSON,1 JOHN T. HANCOCK1 AND PHILIP E. JAMES2 1 Centre for Research in Biomedicine, Faculty of Applied Sciences, University of the West of England, Frenchay Campus, Coldharbour Lane, Bristol, BS16 1QY 2 Department of Cardiology, School of Medicine, Cardiﬀ University, Heath Park, Cardiﬀ CF14 4XN

1

Introduction

EPR spectroscopy retains its position as the most important and powerful technique for the detection and identiﬁcation of biological radicals. Recent developments in the technology have seen an increasing use of its application in the biomedical and healthcare setting. Thus, the scope of EPR spectroscopy in biomedicine is immense and methodological advances have seen exciting developments in new and challenging areas, particularly in vivo EPR and EPR imaging. Many of these advances have allowed the technique to reach a point where it is ready for clinical evaluation. Combinations of reaction rates and relaxation times have conspired to prevent the direct EPR visualization of most biomedically relevant radicals under normal physiological conditions. Instead, their involvement in a process is usually inferred or conﬁrmed through analysis of products or, commonly, identiﬁed by spin trapping the reactive intermediate radicals to form stable radical adducts analysed by EPR. Excellent reviews and updates on the spin trapping technique and its application in biology have recently been written (see the chapter by Clement and Tordo in this series1 and a recent review by Burkitt2). Use is made of stable radicals designed to report on the molecular motions of their environs in the technique of spin labelling. This has facilitated a wealth of information on processes centered on lipid membranes and ligandreceptor interactions. Again, excellent recent reviews of this technique and its applications are available (see, e.g. Feix and Klug in this series,3 Becker et al.4). This review will not attempt therefore, to cover comprehensively the techniques of spin trapping or spin labelling. Neither will it present an exhaustive list of biological radicals in their entirety. Rather, radical generation and processes in biomedically relevant settings will be focussed on and the Electron Paramagnetic Resonance, Volume 20 r The Royal Society of Chemistry, 2007 192

Electron Paramag. Reson., 2007, 20, 192–244

193

application of techniques that have been developed to study speciﬁc radical species that are key to many biomedical processes and conditions, especially in vivo, will be outlined. The chapter will ﬁrst provide an overview of the biologically relevant radical species and the processes that generate them and defences against them. This includes a more detailed analysis of the production and measurement of nitric oxide (NO), a radical species that is gaining significance in a growing number of biomedical settings. Some of the physiological consequences of the altered expression of these radicals and selected disease states associated with them, is then explored. A section has been included on the study of radical generation in plants, many of which have biomedical relevance to higher organisms. Finally, a brief overview of the techniques that have precipitated the increase in biomedical applications of EPR spectroscopy and their potential clinical use will be presented.

2

Reactive Oxygen Species

2.1 Superoxide Radicals. – Superoxide anion radicals (O2d) may be generated enzymatically or from the oxidation of a variety of endogenous and exogenous molecules. Among the enzymes with known O2d-producing capacity are those of the mitochondrial electron-transport chain, leukocyte NADPH oxidase and xanthine oxidase. Enzymatically generated O2d is an important source of cellular oxidant-mediated injury and may also contribute cell signalling functions. Within respiring cells, NADH dehydrogenase (NDH) in mitochondria is a critical site of O2d production although its mechanism of generation is not known. Chen and colleagues5 investigated the catalytic function of NDH in the mediation of O2d generation by EPR spin-trapping. In the presence of NADH, O2d generation from NDH was observed and was inhibited by diphenyleneiodinium chloride (DPI), indicating involvement of the FMNbinding site of NDH. Generation of O2d could also damage NDH itself and using a novel immunospin-trapping technique with anti-5,5-dimethyl-1-pyrroline-N-oxide (DMPO) antibody and mass spectrometry, the sites of oxidative damage of NDH were found on the 51-kDa subunit and shown to involve protein thiyl radicals. The reaction between xanthine and xanthine oxidase is also well-known to produce O2d. A study6 on the potential for reactive oxygen species (ROS) to regulate matrix metalloproteinases (MMPs) utilised the xanthine/xanthine oxidase (X/XO) reaction in an investigation of the activation of proMMP-2 by X/XO in cultured vascular smooth muscle cells (SMCs). EPR analysis showed that X/XO produced O2d, which was completely scavenged by SOD, and activated proMMP-2. However, XO alone also induced activation of proMMP-2 that could not be inhibited by combination of SOD and catalase. These results suggest that that proteolytic activity contained in XO, rather than the ROS derived from X/XO, is responsible for proMMP-2 activation in cultured SMCs. This report highlights the need for care when assessing the results of enzyme-generated radicals.

194

Electron Paramag. Reson., 2007, 20, 192–244

Perhaps the most celebrated enzyme system for the generation of O2d in biological systems is the NADPH oxidase found in phagocytic and other cells. However, in addition to generating O2d (and hence H2O2) activated phagocytes also release the haem enzyme myeloperoxidase (MPO). Reaction of MPO with H2O2 in the presence of chloride ions generates HOCl and exposure of glycosaminoglycans to this system generates long-lived chloramides (R-NClC(O)-R 0 ). Transition-metal ions can lead to decomposition of these species producing further nitrogen- and carbon-centred radicals. A study of the exposure of glycosaminoglycan chloramides to O2d showed that this promotes chloramide decomposition and glycosaminoglycan fragmentation.7 EPR spintrapping experiments using DMPO and 2-methyl-2-nitrosopropane (MNP) have provided evidence for both O2d and polymer-derived carbon-centred radicals as intermediates. This synergistic damage to glycosaminoglycans induced by HOCl and O2d may be of significance at sites of inﬂammation where both oxidants are generated concurrently. NADPH oxidase activity is present in many cells, and low levels of O2d produced in these cells may act as a second messenger. A recent study8 demonstrated that cell-free extracts of rabbit corneal epithelial and stromal cells produced O2d in an NADPH-dependent manner as determined by SODinhibitable spin adduct formation and EPR spectroscopy. Western blot analysis showed these cells to constitutively express ﬁve proteins known to comprise a classic neutrophil-like NADPH oxidase complex. Thus oxidation of NADPH via the NADPH oxidase complex is a potential contributor to signal transduction pathways as well as a potential participant in processes that occur during inﬂammation. The enzymatic oxidation of 3-(3,4-dihydroxyphenyl)-DL-alanine (DOPA)9 was found to generate O2d and further reaction of O2d with DOPA produced an EPR spectrum of DOPA-semiquinone and H2O2. Studies on the non-enzymatic generation of O2d include the photo-irradiation of sunscreen products containing TiO2,10 and the cancer-chemopreventive agent oltipraz, a 1,2-dithiolethione.11 EPR spin trapping demonstrated that oltipraz slowly reacts in the presence of oxygen to generate O2d that the authors suggest could be involved in the mechanism by which it exerts chemoprotection. In another study, oxidative damage produced by phthalocyanines used in photodynamic therapy of tumours was investigated.12 Plasma low-density lipoproteins (LDL) are important carriers of phthalocyanines in the blood, but on exposure to visible light, phthalocyanine-loaded LDL undergo an oxidation process that propagates to erythrocytes. These authors attempted to identify the reactive species involved in LDL and erythrocyte oxidation using EPR spectroscopy with 2,2,6,6-tetramethyl-4-piperidone (TEMPO) and the spin trap 5,5 0 -dimethyl-1-pyrroline-N-oxide (DMPO). The spectra obtained during the irradiation of phthalocyanineloaded LDL in the presence of these traps suggested the simultaneous production of O2d and singlet oxygen (1O2) implying that Type-I and Type-II mechanisms of photochemistry are simultaneously operative in phthalocyanine-loaded LDL.

Electron Paramag. Reson., 2007, 20, 192–244

195

2.2 Hydroxyl Radicals. – Hydroxyl radicals are the most reactive and damaging species in biological systems. They are generated by multiple reactions, many involving H2O2 or O2d and transition metal ions, including the Fenton reaction: Fe21 þ H2O2 - Fe3+ þ OHd þ OH and the metal-catalysed Haber-Weiss reaction: Fe3+ þ O2d - Fe21 þ O2 Fe21 þ H2O2 - OHd þ OH þ Fe3+ Thus, the presence of transition metals, such as iron or copper, often in trace amounts, can catalyse the production of dOH. This can have serious consequences for biological systems and is an important mechanism of oxidant damage to cells and tissues. An illustration of this is the use of carbon nanotubes, nano-cylinders with extremely small diameters (1-2 nm) that have many potential applications in medicine. Manufactured nanotubes can contain significant amounts of iron that can catalyse the production of dOH in macrophages, the primary cell type that respond to particulates. In a recent study of the potential effects of nanotubes on a murine macrophage cell line,13 non-puriﬁed nanotubes were more effective in generating dOH (as shown by EPR spin-trapping with DMPO), than puriﬁed (iron-depleted) nanotubes. In an interesting study on heat-shock treatment, Ilangovan and colleagues14 proposed that dOH is formed from O2d leakage from the electron-transport chain that oxidatively damages mitochondrial aconitase, releasing free Fe21. The released Fe21 combines with H2O2 to generate dOH via a Fenton reaction and the oxidized Fe31 recombines with the inactivated aconitase, is reduced, and reactivates the enzyme. Particulate matter (PM), including diesel exhaust particles have been shown to be linked with adverse health effects. Experimental evidence has showed that PM contains redox-active transition metals, redox cycling quinoids and polycyclic aromatic hydrocarbons (PAHs) which act synergistically to produce reactive oxygen species (ROS). An investigation on the synergistic effect of transition metals and persistent quinoid and semiquinone radicals associated with PM on the generation of ROS15 revealed that the cytotoxic potential of PM can be partly the result of redox cycling of persistent quinoid radicals, which generate large amounts of ROS. Furthermore, the water-soluble fraction of extracted PM was found to elicit DNA damage via reactive transition metaldependent formation of dOH.16 High-energy dissociation of water molecules can generate dOH and a study into the wound-healing effect of the Er-YAG laser in dentistry showed that when water was irradiated with an Er-YAG laser at an energy of 100–130 mJ cm2 and 10-30-Hz pulse repetition rate was dOH generated as assessed by spin trapping.17 UV irradiation of compounds containing a hydroperoxy moiety have been shown to generate dOH. However, UV irradiation (4300 nm) of several naphthalene diimides, which are devoid of the hydroperoxyl moiety and

196

Electron Paramag. Reson., 2007, 20, 192–244

therefore unable to generate dOH, produced the corresponding anion-radicals, positively identiﬁed by EPR, concomitant with the formation of radicals from the solvent. In the presence of oxygen O2d radicals were detected as spin adducts of DMPO, supporting the notion that hydroperoxynaphthalene diimides can induce oxidative damage via photo processes that are independent of d OH generation.18 Hydroxyl radicals have also been generated from herbicides in yeast cells as a eukaryotic model organism19 and from particulate wear debris from orthopaedic composites.20 3

Reactive Nitrogen Species

3.1 Nitric Oxide. – Perhaps the most productive and exciting ﬁeld of mammalian biochemistry and physiology since its discovery some 20 years ago is that of the ubiquitous and simple gas, dNO. Like oxygen, this diatom is itself paramagnetic with the free electron being shared between nitrogen and oxygen. d NO is known to be critical to normal biological homeostasis; at the cellular level it can modulate the respiratory chain, inﬂuence protein synthesis, combine with other radicals to defend against invading organisms, and modulate ion channels and membrane surface receptors. The paracrine function of dNO is focussed around its role as a potent dilator of blood vessels. Produced by vascular endothelium in response to endogenous agonists and stimuli such as shear stress and ﬂow, it relaxes underlying smooth muscle and so causes vasodilatation. A more recent discovery is that of an endocrine role, where d NO is ‘‘stored’’ as a metabolite and re-cycled at vascular and cellular sites distant from its site of production. Whereas EPR techniques can report on levels of dNO in vitro and in vivo, it is perhaps best regarded not as the simplest and most practical means of measuring dNO (given that other techniques can be used for multiple sampling and often without need for specialised equipment) but it is certainly the only method by which dNO and its paramagnetic derivatives can be unambiguously identiﬁed. We provide here an overview of techniques, illustrating examples of their recent use and potential advantages and limitations in their application to biological systems. 3.2 Principles of dNO Measurement by EPR. – An accurate technique for the direct detection of dNO is critical for a complete understanding of the physiological and pathophysiological processes in which dNO is implicated. This especially needs to be measured where it has been produced and where it plays an important role. Measurement of dNO concentrations, until recently, has been limited to ex vivo determinations of complexes in blood or other body ﬂuids,21 with the exception of an electrochemical catheter method that has been used to detect dNO released into superﬁcial veins.22 Ozone-based chemiluminescence has recently emerged as the technique of choice and beneﬁts from exquisite sensitivity (o5 nM). While these techniques may provide important

Electron Paramag. Reson., 2007, 20, 192–244

197

and complimentary data, they have some potential disadvantages. For example, measurement of nitrite and nitrate (NOx, believed to be the major metabolites of dNO in the circulation) by the Griess method is subject to false positives and also cannot distinguish between the various sources of these species (for example, nitrite derived from peroxynitrite). It also only provides limited time resolution because the half life of some metabolites is long and there is accumulation in the system being studied. Measurement of dNO directly via electrodes certainly has advantages, such as real-time measurement and exquisite sensitivity (B0.2 mM), and has been used with good effect in vitro. Its use in vivo is limited primarily by the fact it is invasive, electrodes are extremely brittle (and consequently expensive), and the baseline is extremely temperature sensitive. Although dNO is itself paramagnetic, like oxygen its EPR spectrum is extremely broad (resulting from an extremely fast relaxation time) and is not detectable at practical temperatures. EPR techniques for measurement of dNO therefore largely depend on its entrapment to form a more stable NO-adduct that has readily measured EPR spectra. The principal approach has been to use metal complexes that trap NO and form an EPR detectable product. It also may be feasible to use organic molecules as spin traps for NO. Typically, the amount of adduct present is proportional to the intensity of the signal, or more accurately, the area under the peaks of its EPR spectrum. Each dNO species or adduct has a characteristic EPR spectrum, resulting from delocalisation of the electron over the nitrogen and oxygen atoms, with the spectra usually consisting of three hyperﬁne lines as a result of the nitrogen nuclear moment. 3.3 Metal Chelate Complexes. – Diethyldithiocarbamate (DETC) and methylglucamine-diothiocarbamate (MGD) have both been utilized. The technique depends on formation of a trapping complex between the chelate and ferrous iron (for example, Fe-(DETC)2) that then forms a complex with dNO. The NOFe21-(DETC)2 formed is stable and exhibits a three-line EPR signal typical of a hyperﬁne splitting from nitrogen. The precise spectral characteristics vary according to which species is used. Our group and others developed several systems utilizing this technique some years back.23–26 Together with developments in EPR instrumentation, the improved sensitivity has enabled measurements of dNO to be made in other tissues and proﬁles of tissue NO levels with time. One interesting difference between DETC and MGD is the solubility of the resulting complex with iron: DETC is hydrophobic and consequently readily enters into cells and tissues, while MGD is more hydrophilic and therefore tends to remain in the intravascular compartment. One area where this technique has offered new insights recently is in the spin trapping of dNO from brain tissue. For example, the interaction between endothelial, neuronal and inducible dNO synthases (eNOS, nNOS and iNOS, respectively) is complex and EPR spin trapping of dNO has been utilised to study nNOS-derived dNO27 and vascular derived dNO production (eNOS) in genetically modiﬁed mouse models.28 The technique has also been utilised to great effect in studying oxidative stress in vitro and in vivo. The important

198

Electron Paramag. Reson., 2007, 20, 192–244

neuroprotective role of iNOS derived NO against oxidative stress following brain surgery has been demonstrated29 and there are many examples of biomedical and potential diagnostic application. Particularly important work on the mechanism of S-nitrosothiol decomposition30 and vasodilatory activity of nitroglycerin31 are good examples where the technique has contributed to pharmacological advantage. A further consideration is the oxidation/reduction state of the iron in the DETC-Fe complex being utilised. NO will bind both Fe31-(DETC)2 and Fe21(DETC)2 yet only the latter forms a characteristic three-line spectrum of NOFe-(DETC)2 that is specifically detected by EPR. The redox state is particularly important for in vivo measurements. Some investigators have recently shown a significant contribution arises from NOFe31-(DETC)2 and ex-vivo reduction of this can improve yield and detectability of NO from tissue dramatically.32 3.4 Haemoglobin. – Interactions of dNO with haemoglobin (Hb) are among the most extensively studied. Hb meets the criteria for an effective spin trap for d NO: it is present in high concentration in vivo, displays high afﬁnity for dNO, and the HbNO produced following reaction with deoxyHb has a characteristic EPR spectrum.33 This is particularly attractive for detection of dNO in vivo because of the high concentration of Hb. Oxidation of HbNO may occur in oxygen-rich environments, so immediate freezing of red blood cell samples in liquid nitrogen is desirable. Other small molecular ligands (CO and oxygen) exhibit lesser afﬁnity than dNO and form EPR-silent complexes. An important consideration is the extent of saturation of Hb with oxygen, since interaction of d NO with oxy-Hb occurs at almost the same rate as with deoxy-Hb, but the products are MetHb and nitrate (rather than paramagnetic HbNO). Although this limits its application to some extent, the fact that differential metabolism of d NO by Hb and its various haem groups has proven invaluable in elucidating the transfer and metabolism of dNO in blood. EPR analysis of HbNO was instrumental in suggesting the potential role for nitrite as an alternative source of dNO in vivo34 and can be protective during ischemia-reperfusion injury. More recently, it has been used to study the potential that NOS itself may be consuming nitrite under low oxygen conditions to produce dNO even in the absence of other substrate.35 Biomedical applications of dNO inhalation also depend on HbNO,36 and beneﬁcial effects of gaseous dNO on the healing of skin wounds has been studied by analysing HbNO, NO-Fe-(DETC)2 and myoglobin-bound dNO.37 EPR analysis of HbNO is still regarded as the gold standard by which all other ‘‘HbNO’’ measuring methodologies is compared (such as ozone-based chemiluminescence). 3.5 Nitroxides as Spin Traps for dNO. – EPR-detectable nitronyl nitroxides have been used to trap dNO in vitro in solution for some time.38 They can be used in cell systems with relative ease, since they are well tolerated at relatively high concentrations. Carboxy-PTIO [carboxy (2-phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl 3-oxide] exhibits an EPR spectrum with a ﬁve-line hyperﬁne

Electron Paramag. Reson., 2007, 20, 192–244

199

splitting. Its NO-adduct is characterized by a shift to a nine-line species (however, this is observed as 7 lines in the EPR spectrum because of overlap of some lines). Interestingly, both the spin trap (i.e. the nitroxide) and the spin-adduct can be monitored independently under certain circumstances. We utilized this technique in several studies to detect dNO production by endothelial cells.39 Others have studied the involvement of dNO in the interaction between human neutrophils and platelets. More recently, studies in pancreatic islet cells have demonstrated important inﬂuences of dNO on insulin and/or glucagon secretion40 and the inﬂuence of glucose on NOS expression.41 As with most nitroxides, these spin traps are subject to reduction and cellular metabolism with consequent loss of the EPR signal as the nitroxide group (and possibly the spin trap-radical adduct) is reduced to its corresponding hydroxylamine. This instability limits their use in vivo, although direct detection in the blood circulation by EPR spin trapping of dNO was demonstrated using nitronyl nitroxides in the tails of septic mice42,43 where the concentration of d NO is high. Importantly, cPTIO and its non-cell permeable derivative trimethylammonio-PTIO have been used with great success to scavenge dNO in chemical and biological systems. In particular, the involvement of dNO rather than its metabolites (such as nitrite) has been conﬁrmed in cell44 and tissue preparations45 by the ability of these spin traps to scavenge dNO. 3.6 Simultaneous Measurement of dNO and O2. – The technique relies on the fact that the EPR signal(s) arising from the paramagnetic species used for detection of dNO and the oxygen reporting species do not unacceptably overlap in the resulting EPR spectrum.46 Within the last several years there has been a very significant amount of progress in EPR oximetry, based on developments in several different laboratories.47,48 EPR oximetry is reviewed below and will not be detailed here. However, there are several possible combinations from the approaches of EPR oximetry and dNO detection that allow simultaneous measurement of these important parameters. For example, particulate materials reporting on pO2 can generally be used in conjunction with nitroxides or chelates, since the g-value of the former lies between the middle and high-ﬁeld lines of the latter. Since both materials are relatively inert, this technique can be applied to a broad range of studies involving dNO and oxygen. 3.7 Other Endogenous Paramagnetic Products of dNO. – In addition to Hb (discussed above) many other endogenous iron-based and non-haem traps exist for dNO that are capable of forming paramagnetic complexes and indeed make up a biologically important group of active molecules. Their detectability depends largely on achieving large enough concentrations in biologically relevant samples, such as cells and tissues. Clearly, dNO interacts strongly with most metal containing enzymes, for example the mitochondrial electrontransport chain is effectively inhibited, halting cellular respiration, yet the

200

Electron Paramag. Reson., 2007, 20, 192–244

concentration of metal-NO adducts is invariably too low to be studied by routine EPR techniques. Similar to detection of HbNO, detection typically relies on carrying out measurement at cryogenic temperatures (such as liquid nitrogen or helium). In vivo, non-haem Fe-NO complexes have been observed in regard to growth of primary and secondary solid tumours.49 At the cellular level, NO has been shown to inhibit the peroxidase activity of cytochrome c/cardiolipin complex, blocking cardiolipin oxidation that is an important event during execution of the apoptotic cell death program.50 These authors clearly demonstrate how low temperature EPR spectroscopy can be used to discern cardiolipin-facilitated interactions of ferro- and ferri-states of cytochrome c with dNO and NO, respectively, to yield a mixture of penta- and hexa-coordinate nitrosylated cytochrome c. As already alluded to above, the study of NOS by EPR continues to yield important information not obtainable by other methodologies. The role of tetrahydrobiopterin (BH4) as a cofactor of NOS has been the object of much intense research, yet only recently it was found that in addition to its established effects on the NOS haem spin state, substrate afﬁnity, and enzyme dimerization, BH4 is required as a one electron donor to oxyferrous [Fe(II).O2] haem that is formed as an intermediate in the catalytic cycle.51 Furthermore, properties of the NOS oxygenase/FMN domain were investigated using calmodulin-induced shifts in high spin ferrihaem EPR spectra, and through mutual broadening of haem and FMNH radical signals in iNOS (the inducible form of the enzyme).52 Hemopexin (HPX) serves as a trap for toxic plasma haem, ensuring its complete clearance by transportation to the liver. HPX has also been postulated to play a key role in homeostasis of dNO. The dNO binding characteristics to produce HPX-haem-NO was studied and the pH transition from six- to ﬁvecoordinate product (exhibiting a three-line EPR spectrum) has been monitored by EPR.53 Although not strictly the detection of paramagnetic dNO species, the study of nitrated lipids by a combination of chemiluminescence and EPR has highlighted important insights into these species as endogenous sources of dNO. Homolytic cleavage of nitrated lipid (nitrolinoleate or cholestorol linoleate) yields a lipid-derived radical (which has been trapped using 3,5-dibromo-4nitrobenzenesulfonic acid) and diffusible dNO54 that can actively dilate blood vessels. The study of nitrated lipids is a very recent development but it is already evident this represents a very large pool of vasoactive dNO. This ﬁeld has emerged from the pioneering work by Stamler55,56 and by other groups who have eloquently demonstrated the presence and physiological relevance of dNO ‘‘stores’’ in blood and in tissue.57 EPR methods have played an important role in these studies, both as a standard for studying HbNO interaction but also in order to elucidate the potential for oxygen to modulate binding of dNO in an allosteric, and possibly cooperative manner. Perhaps more importantly, they have a particular role in pathophysiology, especially in regard to pulmonary hypotension and respiratory biology.

Electron Paramag. Reson., 2007, 20, 192–244

4

201

Enzyme-Mediated Free Radical Production

4.1 Cytochrome c. – Mitochondrial O2d production is an important mediator of oxidative cellular injury. Zweier and co-workers58 investigated the succinate-cytochrome c reductase (SCR) of the electron-transport chain as an important site of O2d generation, and an alternative target of dNO in t he regulation of mitochondrial respiration. O2d generation by SCR was measured with the EPR spin-trapping technique using DEPMPO (5-diethoxylphosphoryl-5-methyl-1-pyrroline N-oxide) and found to be dependent on the haem group and oxidation of ferrocytochrome b. In the presence of NO, DEPMPO/OOH adduct production was progressively diminished, implying that NO interacted with SCR or trapped the O2d. The mechanism of electron coupled proton transfer in cytochrome c oxidase (CcO) is still poorly understood. The P(M)-intermediate of the catalytic cycle is an oxoferryl state whose generation requires one additional electron, which cannot be provided by the two metal centres. Utilising conserved variants it has been shown59 that Trp-272 is the electron donor which could be replenished from Tyr-167 or from the Tyr-280 – His-276 cross link in the natural cycle. Cytochrome c peroxidase (CcP) and ascorbate peroxidase (APX) are haem peroxidases which have very similar active site structures, yet differ substantially in the properties of compound I, the intermediate formed upon reaction with peroxides. These results suggest that the electrostatics of the proximal pocket and the shielding of propionate groups by salt bridges are critical factors controlling the location of a stable compound I radical in haem peroxidases. The radicals formed in the reactions between hydroperoxides and the haem group of proteins and enzymes, have been implicated in a number of pathological conditions where oxygen-binding proteins interact with H2O2 or other peroxides; this area has been recently reviewed.60 The radical, generated in the interaction of the oxidized Fe(III) state with hydroperoxides and formation of the Fe(IV)QO (oxoferryl) haem state, is transferred to an amino acid residue of the protein and undergoes further transfer and transformation. By analysis of previously published EPR spectra, the identity of the amino acid radicals has been ascertained, and the sequence of events leading to radical formation, transformation and transfer, both intra-and inter-molecularly, unravelled. The identity of the amino acid radical formed following H2O2 treatment of cytochrome c oxidase is the subject of debate.61 The identity and functional importance of this species is reviewed for cytochrome c oxidase from Paracoccus denitriﬁcans and bovine heart.62 Several other reports have appeared on the identity, and interactions, of the radicals generated in cytochrome oxidase reactions.63–65 The release of cytochrome c from mitochondria during apoptosis results in the enhanced production of O2d, which is converted to H2O2 by Mn-superoxide dismutase. The mechanism of this process has been explored, with the data consistent with the formation of a peroxidase compound I species, with one oxidizing equivalent present as an oxo-ferryl haem intermediate and the

202

Electron Paramag. Reson., 2007, 20, 192–244

other as the tyrosyl radical.66 These ﬁndings provide a physico-chemical basis for the redox changes that occur during apoptosis. Excessive changes (possibly catalysed by cytochrome c) may have implications for the redox regulation of cell death, including the sensitivity of tumour cells to chemotherapeutic agents. The effects of both dNO and peroxynitrite on complexes I (NADH dehydrogenase) and III (cytochrome c reductase) isolated from bovine heart have been examined.67 It was found that peroxynitrite can react with the protein moieties of the enzymes to derivatize a select few tyrosine residues in both complexes I and III forming 3-nitrotyrosine, as detected by immunoblots.

4.2 Cytochrome P450. – Human cytochrome P450 is a superfamily of related enzymes that catalyse the oxygen-dependent metabolism of a wide range of substrates including many drugs and other xenobiotics. During the metabolic conversions, the enzyme can generate reactive oxygen species and other radical intermediates and may be a source of oxidative stress. Mammalian cytochrome P450 (CYP) and cytochrome P450 reductase play important roles in organic nitrate bioactivation although the mechanism of NO formation is not understood. Recent work by Li and colleagues has shown that CYP reductase catalyzes organic nitrate reduction, producing nitrite, whereas CYP can mediate further nitrite reduction to dNO.68 Further reaction of organic nitrite with free or microsome-associated thiols led to dNO or nitrosothiol generation and thus stimulated the activation of soluble guanylyl cyclase. In a study to identify new anti-mycobacterial drugs, the cytochrome P450 enzyme CYP51, from Mycobacterium avium complex that catalyses an early step in sterol metabolism, was cloned and characterized.69 Inactivation of CYP51 by azoles led to accumulation of methylated sterol precursors and compromise of cell membrane integrity, resulting in growth inhibition or cell death. Classes of azole antibiotics were tested in this system to identify potentially more rapid antimycobacterial activity. To gain insights into the molecular basis of the design for the selective azole anti-fungals, the binding properties of azole-based inhibitors for cytochrome P450 sterol 14-alphademethylase (CYP51) from human (HuCYP51) and Mycobacterium tuberculosis (MtCYP51) were compared.70 Density functional calculations on ferric peroxo, ferric hydroperoxo, Compound I and protonated Compound I haem active-site models have been made for the interaction of haemoproteins with peroxide.71 The theoretical results, including calculated isotropic Fermi contact couplings and anisotropic spindipole couplings, have been compared with experimental EPR/ENDOR data.72 Emerging from these data is the possibility that a protonated Compound I has already been detected in ENDOR experiments on cytochrome P450. The general feasibility of a protonated Compound I in P450 monooxygenases is probed in light of these ﬁndings. Several studies have investigated the characterization of iron-oxo intermediates73,74 and the binding of dNO to haem thiolate complexes75,76 relevant to the study of CYP450.

Electron Paramag. Reson., 2007, 20, 192–244

203

4.3 NADPH Oxidase. – Reactive oxygen species produced by phagocytic cells during inﬂammation are generated by the NADPH oxidase. Many cell types are now known to express an NADPH oxidase, although its role in these cells is not well understood. While reactive oxygen and nitrogen species have antimicrobial and tissue-damaging effects in inﬂammation, they are also able to serve as second messengers or regulators of signal transduction. The characterization of enzymes such as NADPH oxidase and NOS are therefore a ﬁrst step in our greater understanding of these functions in different cells under different conditions. A study of a constitutively expressed NADPH oxidase complex that includes the component Nox4, showed it to be a source of O2d produced by rabbit corneal epithelial and stromal cells. O2d produced by the oxidation of NADPH via the NADPH oxidase complex is a potential contributor to signal transduction pathways as well as a potential participant in processes that occur during inﬂammation.77 Duox2 (and probably Duox1) is a glycoﬂavoprotein involved in thyroid hormone biosynthesis, as the thyroid H2O2 generator functionally associated with thyroperoxidase. El-Hassani and co-workers78 have managed to overcome difﬁculties with expression of Duox into nonthyroid cell lines and investigated the H2O2-generating activity in the particulate fractions from Duox2- and Duox1-transfected HEK293 and Chinese hamster ovary cells. They conﬁrmed that mature thyroid NADPH oxidase does not release O2d but H2O2 and also demonstrated that the partially glycosylated form of Duox2, located in the endoplasmic reticulum, generates O2d in a calcium-dependent manner. In an analysis of dichlorodihydroﬂuorescein (H2-DCF) and dihydrocalcein (H2-calcein) as probes for the detection of intracellular reactive oxygen species in vascular smooth muscle,79 intracellular ROS, as generated by the angiotensin II (Ang II)-activated NADPH oxidase, did not increase the oxidation of H2-calcein but increased the oxidation of H2-DCF by approximately 50%. Inhibition of the NADPH oxidase using gp91 silencing prevented the Ang II-induced increase in DCF ﬂuorescence, without affecting cells loaded with H2-calcein. Calcein accumulated in the mitochondria, whereas DCF was localized in the cytoplasm. In submitochondrial particles, H2-calcein, but not H2-DCF, inhibited complex I activity. These observations indicate that H2DCF is an indicator for intracellular ROS, whereas the oxidation of H2-calcein most likely occurs as a consequence of direct electron transfer to mitochondrial complex I. Another study on vascular NADPH oxidases utilized vascular smooth muscle cells (VSMC) to show that in contrast to the neutrophil enzyme, VSMCs can use Nox1 rather than gp91phox as a catalytic centre in the p22phox-based oxidase and that these two proteins are preassembled at or near the plasma membrane and submembrane vesicular structures in unstimulated cells.80 Vascular production of ROS may be modulated by protein levels and polymorphisms of the NADPH oxidase subunit p22(phox). By targetting expression of p22phox to smooth muscle in transgenic mice, Laude and colleagues81 analysed ROS generation and physiological parameters. Surprisingly, aortas from the transgenic p22phox-overexpressing mice produced much

204

Electron Paramag. Reson., 2007, 20, 192–244

more NO and this was found to regulate endothelium-dependent relaxation and blood pressure. These data indicate that chronic oxidative stress caused by excessive ROS production evokes a compensatory response involving increased eNOS expression and NO production that increases extracellular SOD protein expression and counterbalances increased ROS production leading to the maintenance of normal vascular function and hemodynamics. Glucocorticoids were found to have a differential effect on the expression of p22phox subunit of NADPH oxidase in Chlamydia-primed monocytes that may induce unexpected effects on the inﬂammatory response.82 Effects of the C242T CYBA polymorphism in p22phox in neutrophils showed that it is associated with reduced respiratory burst and that vascular oxidative stress is likely to be reduced in these individuals.83 In non-phagocytic cells, O2d formation has been implicated in physiological and pathological functions. Studies on keratinocytes reveal that they express a NOX that is distinct from the phox isoform in phagocytes, and that this can provide a source of O2d that may regulate cell proliferation and host defence in skin.84

5

Antioxidant Defences

5.1 Superoxide Dismutase. – Superoxide dismutases (SOD) are ubiquitous in nature and are speciﬁc for the catalytic removal of O2d: O2d þ O2d þ 2H1 - H2O2 þ O2 The CuZn containing enzymes (CuZn SOD) are highly stable and found in virtually all eukaryotic cells. In animals, CuZn SOD is found in the cytosol but some is also located in organelles including the nucleus and mitochondria. MnSOD is located almost exclusively in mitochondria. An iron-containing enzyme, FeSOD, is found in many bacteria and plants. Although SOD is an intracellular enzyme, an extracellular form of CuZnSOD, ecSOD, that is a large tetrameric glycoprotein, is found in plasma and associated with cell surfaces, where it might prevent formation of peroxynitrite from O2d and NO. Increasing SOD is an aim of antioxidant therapy and the development of ‘SOD mimics’ feature largely in current SOD research. Functional models for SOD include Cu(II) peptide complexes that are able to mimic copper-containing oxidase enzymes.85 A copper complex of curcumin, a phytochemical from turmeric, was synthesized and examined for its superoxide dismutase (SOD) activity. The results conﬁrm that the new Cu(II)curcumin complex possesses SOD activity, radical-neutralizing ability, and antioxidant potential.86 A Mn complex of curcumin was also found to have SOD activity and be stable in different aqueous environments including plasma.87 However, Co(II) complexes synthesized as a model for MnSOD, were found to have a redox potential outside the range required to convert O2d to H2O2.88 Using X-band EPR and electron nuclear double resonance (ENDOR) spectroscopy at liquid helium temperatures, the Cu(II) coordination geometry at

Electron Paramag. Reson., 2007, 20, 192–244

205

the active site of human (hSOD1) and bovine copper,zinc-superoxide dismutase (bSOD1) treated with H2O2 and bicarbonate (HCO3) was examined.89 It was found that HCO3 does not alter significantly the Cu(II) active-site geometry and histidine coordination to Cu(II) in SOD1, as does H2O2 alone. However, the oxidant derived from HCO3 (i.e. carbonate anion radical) reacts with surface-associated Trp-32 in hSOD1 to form the corresponding radical. Dinuclear Cu(II) complexes with N-substituted sulfonamide ligands90 and two Cu(II) complexes of 1,4,7-triazacyclononane with benzimidazole groups91 were found to exhibit high SOD activity. A tris-malonic acid derivative of the fullerene C60 molecule (C3) was found capable of removing O2d with a rate constant k(C3) of 2 106 mol1 s1, approximately 100-fold slower than SOD.92 Its targetting to mitochondria suggests it might have useful MnSOD mimetic activity. 5.2 Catalase. – Generation of H2O2 by SOD enzymes and several other oxidases may itself add to oxidative stress via transition metal conversion to d OH. Catalases catalyse the decomposition of H2O2 to O2 while peroxidases consume H2O2 to oxidize another substrate. An Mn-containing enzyme with catalase activity is found in some bacteria. Catalases may be essential for bacterial resistance to oxidative stress and catalase inhibitors are being investigated as novel antimicrobials. In the gastric pathogen Helicobacter pylori, catalase (KatA) and alkyl hydroperoxide reductase (AhpC) are two highly abundant enzymes that are crucial for oxidative stress resistance and survival of the bacterium in the host. A recent study observed that the catalase in AhpC mutant cells is partially inactivated, and the decrease of catalase activity correlates with the perturbation of the haem environment, as detected by EPR spectroscopy.93 To understand the reason for this catalase inactivation, they examined the inhibitory effects of hydroperoxides on H. pylori catalase by monitoring the enzyme activity and the EPR signal of catalase. The results showed that the total amount of extractable lipid hydroperoxides in the AhpC mutant cells is approximately three times that in the wild type cells revealing a novel role of the organic hydroperoxide detoxiﬁcation system in preventing catalase inactivation. The isoniazid susceptibility of Mycobacterium tuberculosis is mediated by the product of the KatG gene, which encodes a haem-containing enzyme catalaseperoxidase. The native KatG and KatG(S315T) puriﬁed from M. tuberculosis, and KatG(R463L) puriﬁed from Mycobacterium bovis were characterized by EPR and the properties of the native enzymes were compared and contrasted with those reported for recombinant KatG, KatG(S315T), and KatG(R463L) in order to assess the ability of the recombinant enzymes to act as good models for the native enzymes.94 Hydroxyurea has recently been approved for the treatment of sickle cell disease. A number of in vitro studies show that the oxidation of hydroxyurea results in the formation of NO, which also has attracted interest as a potential therapy. EPR studies have provided a proposed mechanism for this conversion that includes initial H2O2-dependent oxidation of hydroxyurea by catalase to

206

Electron Paramag. Reson., 2007, 20, 192–244

form the nitroso species, hydrolysis of this nitroso species to produce nitroxyl, and reductive nitrosylation of the Fe(III) haem of catalase by nitroxyl to yield the Fe(II)-NO catalase complex. These studies have demonstrated NO production from the ferric catalase oxidation of nitroxyl, and have identiﬁed a catalase-mediated pathway as a potential source of NO production from hydroxyurea.95 5.3 Glutathione Peroxidase. – Four analogues of Ebselen were synthesized and their glutathione peroxidase activity and antioxidant property evaluated and compared to Ebselen. Among the studied compounds, only diselenide exhibited both glutathione peroxidase activity and radical-scavenging capability.96 The radical-reducing activity and the membrane ﬂuidity of liver microsomes from selenium-deﬁcient (SeD) rats were examined by EPR spin labelling using nitroxyl-labeled stearic acids.97 SeD caused the induction of liver microsomal cytochrome P-450 activity, and the reduction rate of nitroxyl radical near the membrane surface was increased. Selenium-deﬁcient rats experienced an increase in H2O2 due to a pronounced loss of glutathione peroxidase activity that masked the overall reduction rate of the nitroxyl spin probe by reoxidation of the hydroxylamine form. An increased level of glutathione in SeD liver was also evident, likely due to the absence of GSH-Px activity. Peroxiredoxin I (Prx I) is a key cytoplasmic peroxidase that reduces intracellular hydroperoxides in concert with thioredoxin. To study the role of tissue Prx I in protection from oxidative stress, Prx I-mice were generated and the effect of acute-phase tissue damage caused by ferric-nitrilotriacetate (Fe-NTA) was observed.98 Using real-time EPR imaging, the reduction of the stable paramagnetic nitroxyl radical 3-carbamoyl-2,2,5,5-tetramethylpyrrolidine-1oxyl in vivo was followed and the half-life of this spin probe in the liver and kidney was significantly prolonged in the Prx I-mice. These results stress the importance of glutathione/thioredoxin and demonstrate that Prx I-mice have less reducing activity and are more susceptible to the damage mediated by ROS in vivo than wild-type mice. 5.4 Ascorbate (Vitamin C). – The water-soluble vitamin ascorbic acid is an essential co-factor for several enzymes, including those involved in the biosynthesis of collagen, and is considered an important antioxidant in humans. The antioxidant properties of ascorbic acid (present as the monoanion, ascorbate, at physiological pH) come from its reduction potential that allows reduction of most oxidizing radicals. In this process, ascorbate is oxidized to produce the ascorbyl radical (semidehydroascorbate) which is sometimes used as an indicator of oxidative stress. Ascorbyl is quite unreactive and can be further oxidized to dehydroascorbate. Interestingly, transition metals will oxidize ascorbate and generate H2O2 and dOH and can thus act as a pro-oxidant. The physiological or pathological relevance of this is being investigated. The effects of vitamin C prophylaxis during vascular surgery were investigated and increased ascorbyl radical and lipid-derived PBN spin adducts were measured in the peripheral blood of supplemented patients during reperfusion.

Electron Paramag. Reson., 2007, 20, 192–244

207

The study suggested that ascorbate prophylaxis could promote iron-induced oxidative lipid damage via Fenton-type reactions in the ischemic phase of surgery.99 The pro- and anti-oxidative properties of ascorbate on Cr(VI)induced damage were examined using the yeast Saccharomyces cerevisiae as a model organism.100 In vitro measurements employing EPR and the results of supercoiled-DNA cleavage revealed that the pro-oxidative action of ascorbic acid during Cr(VI) reduction was concentration-dependent and that dOH and Cr(V) had formed following Cr(VI) reduction. However, the in vivo results highlighted the important role of increased cytosol-reduction capacity related to modiﬁcation of Cr(V) formation, increased chromium accumulation, better scavenging ability of O2d and H2O2, and consequently decreased cytotoxicity and genotoxicity in ascorbic acid pretreated cells. Thus, the authors conclude that ascorbic acid inﬂuenced Cr(VI) toxicity both as a reducing agent, by decreasing Cr(V) persistence, and as an antioxidant, by decreasing intracellular O2d and H2O2 formation and by quenching radicals formed during Cr(VI) to Cr(III) reduction. They suggest that increased DNA damage and a fall in reduced glutathione in ascorbic acid-treated cells might induce an endogenous antioxidant defense system and thus increase cell tolerance against subsequent Cr-induced stress. Stable radicals, such as many nitroxides, have been proposed to offer therapeutic potential as antioxidants. However a study by May and co-workers101 in endothelial cells found that the stable nitroxide Tempol (4-hydroxy2,2,6,6-tetramethylpiperidine-N-oxyl) led to increased oxidative stress, as measured by disappearance of the Tempol EPR signal, increased oxidation of dihydroﬂuorescein and glutathione, that could be reversed by ascorbate. Oxidation of ascorbate was investigated in sucrose aqueous model systems by measuring the reduction of Tempol by ascorbate102 and in a study comparing ﬂuorescence and spin-trapping EPR spectroscopy methods.103

5.5 Vitamin E. – Vitamin E comprises eight naturally occurring tocopherols and tocotrienols that form an important chain-breaking antioxidant activity to prevent lipid peroxidation. Research has focused on synthesizing a-tocopherol analogues that have increased water solubility or additional antioxidant effects. In an unusual oxidative role for a vitamin E analogue, the use of alphatocopheryl succinate (alpha-TOS), with pro-apoptotic and anti-cancer activity was studied in human malignant mesothelioma (MM) cells.104 Alpha-TOS was found to down-regulate ﬁbroblast growth factor-2 (FGF-2) in MM cells, via oxidative stress induced by alpha-TOS, whereas nonmalignant cells did not show this response. This activity suppressed MM cell proliferation by generating oxidative stress and point to the agent as a selective drug against fatal mesotheliomas. Rezk and colleagues105 have described antioxidant activities of vitamin E phosphate (VEP), and EPR experiments using the spin label 16-doxylstearic acid have shown that VEP reduces membrane ﬂuidity, in contrast to vitamin E. This suggests that VEP acts as a detergent and forms a barrier that might

208

Electron Paramag. Reson., 2007, 20, 192–244

inhibit the transfer of radicals from one polyunsaturated fatty acid to another, suggesting the basis for a new class of antioxidants. EPR has been used to evaluate radical scavenging properties of hepatopancreas extracts of the shrimp, Pleoticus muelleri, fed different antioxidants, against the stable 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical. Vitamin E showed the strongest effect on DPPH radicals, indicating that vitamin E in the diet can provide immediate protection against radicals.106 Hypoxia is known to limit wound healing and both normobaric (1 atm) and hyperbaric oxygen (HBO) approaches have been used clinically to oxygenate wound tissue. The increases in oxidant stress associated with HBO treatment were studied in a rat model and found to be completely neutralized by excess vitamin E in the diet.107 The kinetics of the reaction of alpha-tocopheroxyl radical with green tea polyphenols was examined by stopped-ﬂow EPR108 and the inhibition of linoleic acid peroxidation by these antioxidants, and the decay of alphatocopherol during the peroxidation, was measured. It was found that the green tea polyphenols could increase the potency of vitamin E by reducing the atocopheroxyl radical to regenerate a-tocopherol. Peroxynitrite can produce damaging ROS and RNS and react directly with many biomolecules. The kinetics and mechanisms of interaction between Trolox C, a water-soluble analogue of vitamin E, and peroxynitrite and its modulation by CO2, were evaluated by direct EPR. The results suggest that H1 or CO2-catalyzed homolysis of peroxynitrite is required to cause oxidation109 of Trolox C, and thus, a-tocopherol. 5.6 Phenolic Antioxidants. – There is much current interest in the potential antioxidant activity of plant phenolics and they are being cited as possible components in the protective effect of the ‘‘Mediterranean diet’’. The plant polyphenol tannic acid was observed to limit dOH formation by preventing Fe(III)-induced ascorbate oxidation and Fe(II) autoxidation110 and by forming Cu complexes that possess dOH scavenging activity.111 The safety and whole-body antioxidant potential of a novel formulation of edible berries rich in anthocyanins was tested in experimental animal models and found to be effective.112 The related C-phycocyanin ameliorated doxorubicin-induced oxidative stress and apoptosis in cardiomyocytes.113 The use of doxorubicin, a potent antineoplastic agent, is dose-limited due to the associated risk of developing cardiomyopathy and congestive heart failure. Epidemiological studies have shown that moderate intake of red wine reduces the risk of coronary heart disease possibly due to the scavenging of reactive oxygen species by polyphenols and ethanol or an effect on endothelial NO production. Several groups have investigated the antioxidant potential of polyphenolic compounds from red wine and other plant origin and reported disparate results. Studies of polyphenols in olive oil showed an antioxidant effect114 while those in red wine were suggested to act synergistically with another phenolic, resveratrol.115,116 Resveratrol synthesis requires activation of the gene for stilbene synthase and introduction of this gene into tomato plants was performed to modify its antioxidant activity.117 The dOH scavenging activity of

Electron Paramag. Reson., 2007, 20, 192–244

209

the products, into which the gene had been introduced, was almost double than that of natural products and resveratrol concentrations in modiﬁed tomatoes were much higher. Moreover, in isolated rat hearts subjected to ischemia/ reperfusion, the rats fed with modiﬁed tomatoes exhibited better cardiac performance, reduced myocardial infarct size and decreased number of apoptotic cardiomyocytes, and reduced oxidative stress compared to unmodiﬁed tomatoes, or resveratrol alone, indicating superior cardioprotective abilities of modiﬁed tomatoes. Huisman and co-workers118 determined the reaction rates of O2d with four different polyphenols and ethanol and found that at concentrations found in vivo (low nM range), the scavenging of O2d by polyphenols and ethanol is negligible. They suggest that the observed antiatherogenic effects must be caused by a mechanism other than direct scavenging of O2d or inﬂuence on maximal endothelial NO production. Other studies also showed that the antioxidant activity of polyphenols occurs at concentrations at least 1 order of magnitude higher than their bioavailability.119

6

Consequences of Free Radical Reactions with Biomolecules

The outcome of the interactions between radical generation and antioxidant defences determine the consequences of radical production on health. If the radical production is high and/or the antioxidant defences depleted, then radical damage to biomolecules will accumulate (oxidative stress) with subsequent tissue damage or loss of function manifest as symptoms of disease. Radical damage to biomolecules is therefore outlined and the consequent disease outcomes reviewed in the next sections. 6.1 Damage to Lipids. – Radical damage to lipids and membranes is usually a secondary event after generation of reactive oxygen- and nitrogen-centred radicals. Thus, radical damage to lipid membranes is usually reliant on the detection of non-radical lipid peroxidation products that may accumulate at sites distant from their generation. This has made the exact temporal and spatial identity of primary lipid damage by reactive species difﬁcult to determine and thus assumptions about the precise mechanisms of membrane damage remain unproven. Several recent methodological developments have made the identiﬁcation of initial lipid radical species more achievable. A recent review of the detection of lipid radicals by EPR provides a useful guide to experimental approaches that have been used to detect lipid radicals.120 Development of new EPR spin-trapping agents designed to trap radicals at a predetermined depth within biological membranes were tested in a model system consisting of large unilamellar vesicles exposed to a copper-dependent (t-BuO ) generating system.121 To assist in the identiﬁcation of the radicals detected, preliminary studies were performed in methanolic solution, where the major radical trapped was shown to be CH2OH, resulting from H-atom abstraction from the alcohol by t-BuO . This supports the proposal that the primary species trapped

210

Electron Paramag. Reson., 2007, 20, 192–244

in the lipid vesicles were radicals derived from membrane fatty acids. The ability to trap lipid-derived radicals ex vivo has been used122,123 and with conﬁrmation that storage of the adducts at low temperatures [196 1C (liquid nitrogen) and 80 1C] does not significantly diminish them with time124 will allow the study of the lipid radical products in laboratories at a distance from where the radicals may be generated. Conventional EPR techniques were used to show that a novel nitroxidebased sunscreen, 2-ethylhexyl-4-methoxycinnamate, combined with the piperidine nitroxide TEMPOL was efﬁcient in preventing photo-oxidative damage to lipids induced by UVA, natural sunlight and 4-tert-butyl-4-methoxydibenzoylmethane, a photo-unstable sunscreen which generates radicals upon UV radiation.125 Metal-ion induced generation of ROS is associated with membrane lipid damage and lipid peroxidation. Several studies have used EPR techniques to determine the level of free iron and iron chelation strategies and correlate this with membrane damage markers such as lipid peroxidation, in experimental models of iron overload, iron supplementation and neurodegenerative conditions in which iron misregulation has been reported.126–128 Recent data suggest that the primary mechanism of amyloid beta-protein neurotoxicity may be mediated by radicals. A study to evaluate this hypothesis used the spin traps DMPO and a-(4-pyridyl-1-oxide)-N-tert-butylnitrone (POBN), to reveal that dOH, and a lipid radical were produced in the process of cell damage by amyloid beta-protein.129 This study provides EPR evidence that amyloid beta-protein neurotoxicity is derived from hydrogen abstraction from polyunsaturated lipid acid by dOH as a cause of lipid peroxidation. Radical-mediated lipid peroxidation has been strongly suggested to be the main cause of neuronal toxicity in the rat brain, including neonatal brain damage. In an in vivo study of experimental neuronal toxicity, the spin trap POBN was perfused through a probe in the hippocampus before and after hypoxia and the EPR spectra analysed in the dialysates.130 Comparison with adducts from a lipid radical generating system (lipoxygenase/linoleic acid) conﬁrmed the generation of lipid radicals that are suggested to participate in the cascade of reactions leading to neuronal damage in the hippocampus following ischemic-hypoxic insult in neonatal rats. 6.2 Damage to DNA. – Protein iron binding has been suggested to have both protective and degradative effects on the integrity of nuclear DNA. EPR experiments have been carried out using human H-chain ferritin to discover if the DNA protective effects are due to the proteins ability to prevent Fenton chemistry.131 EPR spin-trapping experiments demonstrated that a ferritin-like DNA binding protein from Listeria innocua attenuates the production of dOH by Fenton chemistry. DNA cleavage assays showed that the protein, while not binding to DNA itself, protects it against the deleterious combination of Fe2+ and H2O2.132 In novel studies on oxidative DNA damage from a distance, ﬂash/quench experiments monitored by EPR spectroscopy proposed a model for the redox activation of DNA repair proteins through DNA charge transport, with guanine radicals, the ﬁrst product under oxidative stress, in oxidizing

Electron Paramag. Reson., 2007, 20, 192–244

211

the DNA-bound repair proteins, providing the signal to stimulate DNA repair.133 Despite a low copy number within the cell, base excision repair (BER) enzymes readily detect DNA base lesions and mismatches. These enzymes also contain [Fe4S4] clusters, yet a redox role for these iron cofactors had been unclear. An important EPR study from the same group, using chemically modiﬁed bases, shows electron trapping on DNA in solution with bound BER enzymes with electron transfer from two BER proteins, Endonuclease III and MutY, to modiﬁed bases in DNA containing oxidized nitroxyl radical EPR probes.134 These results are consistent with DNA binding leading to the activation of the repair proteins toward oxidation and support a mechanism for DNA repair that involves DNA-mediated charge transport. To study the spatial properties of trapped radicals produced in heavy-ionirradiated solid DNA at 77 K, pulsed electron paramagnetic double resonance (PELDOR or DEER) techniques were employed using a refocused echo detection sequence that allows dipolar interaction between trapped radicals to be observed.135 The EPR spectrum obtained was attributed to electron loss/ gain DNA base radicals and neutral carbon-centred radicals that probably arise from sugar damage. Measurements revealed a radical concentration of 13.5 1018 cm3 in the tracks and a track radius of 6.79 nm. The cross section of these tracks was 144 nm2, yielding a lineal radical density of 2.6 radicals/nm. These measurements of radical density and spatial extent provide the ﬁrst direct experimental determination of track characteristics in irradiated DNA. A review of time-resolved EPR techniques applied to study radical pairs particularly in DNA repair mediated by DNA photolyase provides a useful resource for mechanistic approaches to DNA repair.136 Other studies have investigated possible oxidative DNA damage in benzo(a)pyrene-induced genotoxicity and, eventually, carcinogenicity.137 ROS formation by rat lung and liver microsomes following a single oral dose of benzo(a)pyrene was studied in vitro by EPR spectrometry and subsequent bulky DNA adduct formation and 8-oxo-dG levels were examined in vivo in rat lung and liver. The results indicated that ROS are generated during the CYP450-dependent metabolism of benzo(a)pyrene can lead to bulky DNA adducts that were more persistent in rat lung than in liver. However, this may not necessarily result in increased levels of oxidative DNA damage in vivo, due to induction of DNA repair mechanisms. Protection from radiation damage to DNA has been investigated with N(2-mercaptopropionyl) glycine (PSH).138 Electron loss from PSH gave an EPR detectable radical anion, PSSPd. When the PSH derivative was frozen in aqueous DNA solutions to 77 K and exposed to ionizing radiation, normal damage to the DNA was detected by EPR spectroscopy. However, on annealing above 77 K, EPR features for the DNA base radical cations and anions gave central features assigned to PSSPd sigma*-radical anions, together with outer features of 5-6-dihydro-5-thymyl radicals. The author concluded that the primary effect of PSH is to capture the G1. centres, and thereby either prevent or repair radiation damage to DNA. Characterisation of the quantitative and qualitative production of thymine radicals produced by monochromatic

212

Electron Paramag. Reson., 2007, 20, 192–244

ultrasoft X-(USX) or 60Co gamma-rays was undertaken using direct irradiation of thymine powder pellets in an X-band microwave cavity in a vacuum chamber.139 Differences observed in the EPR dose-response relationship reﬂect the difference in the K-absorption cross-sections of carbon, nitrogen and oxygen in the thymine molecule which govern the photo-/Auger electron energy spectrum. EPR studies showed that a semiquinone-radical was produced by some benzoquinone mustard bifunctional alkylating agents, and that they are likely to exert their action through DNA cross-linking and/or by inducing oxidative stress.140 Although DNA topoisomerase II is not a direct target of these agents, the study suggests that this enzyme may play a role in processing the consequences of direct DNA adduction and/or oxidative DNA damage. Epidemiological evidence linking exposure to ambient particulate matter (PM) and increased morbidity and mortality has prompted studies into the potential mechanisms. These have mostly explored ROS, especially dOH, generation by transition metals associated with PM. A number of studies have utilised EPR and spin trapping to examine dOH generation and DNA radicals induced by PM.141–143 Selective DNA cleavage has a wide range of applications from molecular cloning to anti-cancer agents. Selective photo-induced cleavage using DNA binding Cu(II) complexes has been investigated by several groups that suggest major groove binding and DNA cleavage via ROS radical generation.144–147 Application of naphthalene diimides as selective photonucleases has also been explored.148 A development from the detection of DNA (and protein) radicals by EPR spin trapping has been the detection of these radicals by immuno–spin trapping. This is based on the detection of the nitrone adducts of spin traps, such as DMPO, with an anti-DMPO antibody, the complex being ‘‘visualized’’ by a variety of techniques including ELISA and Western blotting. Further analysis with mass spectroscopy affords more structural information. The technique has been developed to allow the detection of reactive DNA and protein radicals in laboratories that do not have access to EPR spectroscopy and complements existing methods for the measurement of oxidatively generated DNA or protein damage. It has been pioneered by Mason and collaborators149 who have used it for the detection of DNA radicals and protein radicals (see below). 6.3 Damage to Proteins. – Protein radicals can be generated through oxidative stress and via enzymatic intermediates. Protein radical intermediates generated in reactions of peroxidases, oxidases, reductases and photosynthetically are dealt with in separate sections. ROS can induce a multitude of protein alterations including oxidation of thiol groups, hydroxylation of tyrosine residues, formation of protein peroxides and protein carbonyls. Reactive nitrogen species may attack tyrosine residues to produce 3-nitrotyrosine. The generation and reactions of dNO and its metabolites have been reviewed in Section 4 and this section concentrates on protein damage initiated by ROS.

Electron Paramag. Reson., 2007, 20, 192–244

213

An excellent review by Davies and Hawkins150 summarizes recent developments in EPR spin trapping applied to the detection of protein radicals and provides a useful overview of radical formation on proteins. EPR spin trapping coupled with immuno-spin trapping (see above) has been developed to identify protein radicals.151 Using a combination of spectrophotometry, immuno-spin trapping, and EPR, the formation of radicals during the oxidation of oxyhemoglobin (oxyHb) and oxymyoglobin (oxyMb) by nitrite and H2O2 was investigated.152 Using EPR/spin trapping, a protein radical was observed in the case of oxyMb, but not oxyHb, and was inhibited by catalase. When DMPO spin trapping was combined with Western blot analysis using an anti-DMPO-nitrone antibody, globin/DMPO adducts of both oxyHb and oxyMb were detected, and their formation was inhibited by catalase. Catalase effects conﬁrm the intermediacy of H2O2 as a haem oxidant in this system. Other studies have used immuno-spin trapping and EPR to demonstrate that NO and NO2 both react with the tyrosyl radical formed in sperm whale myoglobin by reaction with H2O2153 and the radical is located at Tyr-103 as detected by MS/MS and site-speciﬁc mutagenesis.154 In a reinvestigation of the biochemistry of H2O2-induced Cu,Zn-superoxide dismutase (SOD1)-centred radicals, immuno-spin trapping was used to show that these radicals are involved in H2O2-induced structural and functional damage to SOD1, and their mechanism of generation depends on copper and/or (bi)carbonate (i.e., CO2, CO32, or HCO3).155 NADH dehydrogenase (NDH) is a critical site of mitochondrial O2d production and an important mediator of oxidative cellular injury. Its mechanism of O2d generation is poorly understood and this was investigated in an EPR spin-trapping study by Chen and colleagues.5 O2d production by NDH was seen to induce self-inactivation and immuno-spin trapping with antiDMPO antibody and subsequent mass spectrometry was used to deﬁne the sites of oxidative damage of NDH. A DMPO adduct was detected on the 51-kDa subunit and was O2d dependent. Alkylation of the cysteine residues of NDH significantly inhibited NDH-DMPO spin adduct formation, indicating involvement of protein thiyl radicals and LC/MS/MS analysis revealed that Cys-206 and Tyr-177 were speciﬁc sites of NDH-derived protein-radical formation. Proteins are major targets for leukocyte-mediated oxidative injury although limited information is available on the mechanisms of this damage and the intermediates formed. Hawkins and Davies156 analysed the reactions between proteins and hypobromous acid (HOBr), produced by activated eosinophils and found the formation of bromamines and bromamides, from side-chain and backbone amines and amides, and 3-bromo- and 3,5-dibromo-Tyr, from Tyr residues accounted for approximately 70% of the oxidant consumed. The bromamines/bromamides were unstable and induced further oxidation and radical formation detected by EPR spin trapping. Evidence was obtained for the generation of nitrogen-centred radicals on side-chain and backbone amide groups of amino acids, peptides, and proteins. These radicals readily undergo rearrangement reactions to give carbon-centred radicals.

214

Electron Paramag. Reson., 2007, 20, 192–244

A study comparing the effect of photogenerated singlet oxygen (1O2) and radicals on cytochrome c structure and reactivity utilised photoexcitation of methylene blue to generate the reactive species.157 Matrix-assisted laser desorption ionization time-of-ﬂight mass spectrometry of the haem group and the polypeptide chain of cytochrome c with Soret band at 405 nm (cytc405) revealed no alterations in the mass of the cytc405 haem group but oxidative modiﬁcations on Met-65 and Met-80 and Tyr-74 residues. Damage of the cytc405 tyrosine residue impaired its reduction by diphenylacetaldehyde, but not by beta-mercaptoethanol, which was able to reduce cytc405, generating cytochrome c Fe(II) in the high-spin state (spin 2). The mechanism of UV-B radiation induced damage to the light harvesting apparatus of the cyanobacterium Synechocystis 6803 was investigated.158 Investigations performed on phycobilisomes or isolated biliproteins irradiated with moderate UV-B intensity revealed rapid destruction of beta-phycocyanin and a slower damage of the other biliproteins. EPR spin-trapping measurements revealed generation of carbon-centred radicals suggesting that radicals produced from bilins probably attack the polypeptide chain of protein inducing its degradation. Diphenylacetaldehyde (DPAA) is able to promote damage in mitochondrial DNA, lipids, and proteins and previous studies suggest that DPAA is a redox cytochrome c modiﬁer. Recent analysis suggests involvement of two tyrosine residues, probably Tyr-67 and Tyr-74, related to DPAApromoted haem iron reduction.159 The cytochrome c-promoted DPAA electron abstraction quickly produced the expected enol-derived radical, as indicated by 3,5-dibromo-4-nitrosobenzenesulfonate (DBNBS) spin-trapping EPR measurements, that could react with O2, producing a peroxyl radical intermediate that promotes formation of benzophenone as the major product of this reaction. EPR spectroscopy has been extensively used to identify and characterize protein-based redox-active amino acid radicals based on their g-values and hyperﬁne couplings. The theoretical g-values and proton hyperﬁne tensors of three models corresponding to the tyrosyl, tryptophanyl and glycyl radicals were calculated and found not to differ significantly from previous calculations.160 In the case of the tyrosyl radical, it was shown that the para-position substituent that is opposite to the C-O group can break the symmetry of the phenyl ring, leading to different hyperﬁne tensors for the two large orthoproton couplings. For the tyrosyl and tryptophanyl models, the calculated hyperﬁne couplings to hydrogen-bonding protons were in very good agreement with measured values for both the tyrosyl and tryptophanyl models. Probing the reaction centres in Rhodobacter sphaeroides was investigated with mutants containing a tyrosine residue near a highly oxidizing bacteriochlorophyll dimer.161,162 The mutants all showed evidence of tyrosyl radical formation at high pH, and the extent of radical formation at Tyr L135 with pH differed depending on the identity of L144 and L164. The results show that tyrosine residues within approximately 10 A˚ of the dimer can become oxidized when provided with a suitable protein environment. Human recombinant copper-zinc superoxide dismutase (CuZnSOD) was inactivated by peroxynitrite and the radical species formed led to the proposal

Electron Paramag. Reson., 2007, 20, 192–244

215

that peroxynitrite reacts with CuZnSOD leading to nitrogen dioxide plus a copper-bound dOH that reacts with histidine residues, forming histidinyl radicals.163 Identiﬁcation of the site of protein-bound tyrosyl radicals is important as they catalyze many important enzymatic reactions and can initiate oxidative damage to cells. Svistunenko and Cooper164 recently reported a new method of computer simulation of tyrosyl radical EPR spectra that allows the determination of the rotational conformation of the phenoxyl ring in a radical with great accuracy (ca. 2 degrees). When coupled with a new online database, all tyrosine residues in a protein can be screened for that particular conformation. 6.4 Damage to Carbohydrates. – Activated phagocytes release the haem enzyme MPO (myeloperoxidase) and also generate O2d and hence H2O2, via an oxidative burst. Reaction of MPO with H2O2 in the presence of chloride ions generates HOCl, producing a potent oxidatitive mixture at inﬂammatory sites. Studies by Davies and co-workers165 have shown that HOCl and O2d can act synergistically to induce fragmentation of hyaluronan and chondroitin sulphates. Furthermore, heparan sulphate proteoglycans are key components of the extracellular matrix and cell surfaces and are known to bind myeloperoxidase (MPO) via their negatively charged heparan sulphate chains. Reaction of heparan sulfate with HOCl generates polymer-derived N-chloro derivatives that are decomposed in the presence of redox-active transitionmetal ions and O2d. Radical intermediates in these processes have been identiﬁed by EPR spectroscopy and spin trapping.166 Evidence has been obtained that the N-chloro derivatives undergo reductive homolysis to nitrogen-centred (aminyl, N-chloroaminyl, sulfonamidyl, and amidyl) radicals that generate carbon-centred radicals via rapid, intramolecular hydrogen atom abstraction reactions The degradation of heparan sulfate via reductive homolysis of its N-chloro derivatives may be of significance at sites of inﬂammation, where MPO-derived HOCl is produced in high concentration and transition-metal ions and O2d are known to be present or generated. Investigation of the oxidant peroxynitrite on hyaluronan has been studied using an integrated spectroscopic approach, using EPR, NMR, and MS.167 Although NMR and MS experiments did not reveal peroxynitrite-mediated modiﬁcation of hyaluronan, spin-trapping EPR experiments indicated that peroxynitrite induces the formation of carbon radicals, most probably by its d OH-like reactivity. These EPR data support a role for oxidation in the degradation of hyaluronan, a probable event in the development and progression of rheumatoid arthritis. The one electron interactions of carbohydrates with various radicals has been used to examine their potential use as radical scavengers in a range of settings. Carotenoid/cyclodextrin complexes have been examined as antioxidants. The results obtained show that cyclodextrin protects carotenoids from reactive oxygen species but results in a considerable decrease in antioxidant ability of the carotenoid.168 The interaction between the sunscreen agents, phenylbenzimidazole sulphonic acid169, octyl-dimethylaminobenzoate, oxybenzone

216

Electron Paramag. Reson., 2007, 20, 192–244

and octyl-methoxycinnamate170 and hydrophilic a, b, g and-cyclodextrin derivatives was investigated. Irradiation-induced decomposition of sunscreen agent in the emulsion vehicle was markedly reduced by complexation with hydroxypropyl-b-cyclodextrin (HP-b-CD). Spin-trapping studies showed that the inclusion of the sunscreen agent into the HP-b-CD cavity completely inhibited the formation of radicals generated on exposure to simulated sunlight, thereby suppressing its photosensitising potential. The effect of stabilization produced by b-cyclodextrin on carbon-centered radicals afforded by some anti-cancer mercaptopyridine congeners has been studied.171 In the presence of b-cyclodextrin, photochemically produced alkyl radicals are involved in an adduct with b-cyclodextrin, which works as a cage preventing the interaction of such radicals with the medium. Furthermore, the presence of b-cyclodextrin has been shown to delay the generation of radicals in water due to a possible interaction between the cyclodextrin cavity and the mercaptopyridine. Several groups have investigated the effect of ionizing radiation on chitosan microparticles,172 sucrose173 and D-fructose.174

7

Free Radicals and Disease

7.1 Cancer. – 7.1.1 Chemotherapy Metabolites. There is continuing and considerable interest in the role of radicals in the mechanism of actions of drugs used in cancer chemotherapy and chemoprevention. This includes radicals implicated in the direct cytotoxic activity of the drug and radicals generated in metabolites of these agents. A novel approach to cancer chemotherapy is the use of radical producing substances as pro-drugs, able to produce irreversible damage to the tumour cell, so stimulating cellular apoptosis.175 EPR spin trapping studies using DMPO showed that in the presence of glutathione (GSH), O2d is formed from 7-methyl-6,8-bis-methyldisulfanylpyrrolo[1,2-a]pyrazine, a precursor to the metabolite of the cancer chemopreventive drug oltipraz.176 It was hypothesised that GSH reduces the metabolite to a radical-anion, that in turn donates an electron to oxygen resulting in O2d formation; this suggests a possible mechanism by which the parent compound, oltipraz, might effect the cancer chemopreventive increase in the transcription of phase two enzymes that is mediated by transcription factor Nrf2. EPR spectroscopy has been used to ascertain the oxidation status of a V(III)-Lcysteine compound which exhibits antimetastatic, antioxidant and inhibition of neutral endopeptidase activities,177 to demonstrate that the iron centre and tyrosyl radical are involved in ribonucleotide reductase (RR) activity, and show that the susceptibility differences to RR inhibitors may lead to a new direction in drug design for human cancer treatment.178 In a recent thought-provoking study,179 glucoraphanin, the bioprecursor of the widely extolled chemopreventive agent sulforaphane found in broccoli, was shown to induce Phase-I xenobiotic metabolizing enzymes, and increase (unidentiﬁed) radical generation in rat liver, suggesting that long-term administration of glucoraphanin may be

Electron Paramag. Reson., 2007, 20, 192–244

217

a potential health hazard. Alpha-tocopheryl succinate was shown to inhibit malignant mesothelioma by disrupting the ﬁbroblast growth factor autocrine loop by a mechanism that involves oxidative stress.180 EPR spin-labelling has been used to show that the microviscosity of the interior of erythrocytes increases during the early stages of treatment with doxorubicin but not idarubicin, anthracycline antibiotics widely used in human cancer treatment.181 This suggests that idarubicin is less toxic than doxorubicin to erythrocytes from acute myeloid leukaemia patients. 7.1.2 Carcinogenic Agents and Free Radicals in Tumours. Several studies have investigated the production of reactive oxygen and nitrogen species in tumour cells and how these might relate to cell hypoxia,182 estrogen exposure183 or as tumour prognostic markers.184 Mechanisms of carcinogenesis by particulate matter (PM) pollutants linked to the production of reactive oxygen species and nucleotide radicals that result from PM exposure in vitro, were investigated by spin trapping studies with 2-methyl-2-nitrosopropane (MNP).16 The presence of Fe-NO complexes in the interaction between primary and secondary tumours was detected by EPR spectroscopy at liquid nitrogen temperatures in a convenient model for research on ‘‘concomitant immunity’’ against in vivo growing tumours.185 7.1.3 Radiotherapy, Ultrasonic and Photodynamic Therapy. Sensitization of tumours to phototoxicity is an important area of research. The photodynamic effects of phthalocyaninatosilicon covalently linked to 2,2,6,6-tetramethyl-1piperidinyloxyl radicals toward HeLa cells in vitro were found to be due to the large singlet oxygen yield and the inhibition of aggregation due to the bulky TEMPO radicals.186 Photoirradiated water solutions of C(60) fullerenes have been shown to be potentially useful for photodynamic therapy of tumours as a ROS generating system.187 Ultrasonic therapy has also been investigated by EPR as a potential additional cancer treatment utilising radical generation.188,189 In a novel study, a polypeptide-nitroxide conjugate was used for non-invasive thermometry in hyperthermia as an adjuvant with radiation and chemotherapy, a protocol that has shown promise in the treatment of cancer.190 7.1.4 Tumour Oxygenation. Establishing the pO2 of tumours is an important factor in predicting their response to cytotoxic therapy. Several groups have developed novel EPR techniques for the measurement of tumour oxygenation both in vitro and in vivo;191–193 these have been reviewed by Gallez et al.194 7.2 Diabetes. – Endothelial dysfunction is a feature of both Type 1 and Type 2 diabetes and is characterised by an oxidative stress due to increased O2d production and decreased NO bioactivity. The role of tetrahydrobiopterin (BH4), an essential cofactor of endothelial NO synthase (eNOS), in protection against oxidative endothelial dysfunction was examined in hyperglycemic human aortic endothelial cells (HAEC) by adenovirus-mediated gene transfer of GTP cyclohydrolase I (GTPCH, the rate-limiting enzyme for the de novo

218

Electron Paramag. Reson., 2007, 20, 192–244

BH4 synthesis).195 High glucose was seen to decrease HAEC NO and increase O2d production, in association with reductions in both total biopterin and BH4 levels, whereas GTPCH gene transfer increased cellular biopterin levels and NO production, but decreased O2d production. This study indicates that GTPCH is a rational target to augment endothelial BH4 and recover eNOS activity in cases of hyperglycaemic endothelial dysfunction. The deterioration of endothelial dysfunction in response to post-prandial lipaemia has been shown to be attenuated by vitamin C in patients with Type 2 diabetes.196,197 The production of the potent oxidant peroxynitrite can also contribute to the oxidative stress associated with diabetes. The oxidation initiated by peroxynitrite of the ethyl esters of acetoacetate (EAA) metabolites that accumulate in diabetes was investigated by EPR spin trapping using the spin traps 3,5dibromo-4-nitrosobenzenesulfonic acid and 2-methyl-2-nitrosopropane.198 This showed the intermediacy of methyl and carbon-centered (dCH2COR) radicals in the oxidation of EAA by peroxynitrite, and has provided further insight into the molecular basis of diabetes. The structure of vanadyl complexes of different ligands have been characterised by EPR spectroscopy and found to have insulin-enhancing or mimetic effects.199,200 7.3 Sepsis. – Sepsis is a severe host inﬂammatory response to infection that includes an oxidant stress, endothelial damage and organ dysfunction. No speciﬁc therapies for sepsis exist and it remains a significant cause of morbidity and mortality. In many experimental models of sepsis, the bacterial outer membrane component endotoxin (lipopolysaccharide; LPS) is often used as a relevant trigger of disease. dNO is necessary to maintain vascular tone, but sepsis is associated with excessive dNO production from iNOS that induces a profound hypotension and may mediate direct toxic effects. Thus, many studies in sepsis have utilised EPR to detect and measure dNO production in tissues and blood. A separate section in this chapter has been devoted to this biomedically important radical and details its measurement at various sites under different conditions (see Section 3). Paradoxically, excess dNO production in the macrocirculation in sepsis may be accompanied by a collapse of the microcirculation possibly due to insufﬁcient eNOS activity to maintain microvessel tone. Efforts to alleviate this situation have been made through addition of dNO release agents in experimental models of sepsis201 and the use of anti-ischemic drugs.202 The role of iron-dextran on inﬂammatory cell dNO production has been assessed in LPS-treated rats by EPR spin trapping.203 The role of dNO in the maldistribution of O2 in tissues observed in sepsis has been investigated in vivo by EPR oximetry (see Section 10.2)204 allowing new insights into this condition to be gained. A recent novel study has used low temperature EPR spectroscopy and a deconvolution method based on spectra subtraction with variable coefﬁcients, to quantify individual paramagnetic components of human muscle biopsies taken from critically ill patients with severe sepsis.205 Analysis of the paramagnetic centres in the muscle showed that a mitochondrial radical assigned to a spin-coupled pair of semiquinones (SQd-SQd) negatively correlates with the illness severity of the patient, and a decreased concentration of mitochondrial

Electron Paramag. Reson., 2007, 20, 192–244

219

Complex I iron-sulfur redox centres is linked to mortality. A study by Nakai et al. on the inﬂammatory reactions induced by LPS in skin suggest that both iNOS and xanthine oxidase, but neither NADPH oxidase nor Fe(III), have synergistic effects to form lipid radicals and 3-nitrotyrosine early in the skin inﬂammation caused by LPS.206

7.4 Cardiovascular Disease. – Oxidative stress is an important cause of cardiovascular disease and contributes to the progression of chronic heart failure as well as mediating direct cardiac injury following infarct. EPR spectroscopy has been applied to the study of many aspects of cardiovascular and heart disease and these are summarized in the following section. Reperfusion injury, mediated at least partly by a burst of reactive oxygen species (ROS), occurs with the reintroduction of O2 to ischemic tissue and is a source of tissue damage in the heart following myocardial infarct. Several studies have investigated the role of ROS inhibitors on cardiac function in the ischemia-reperfusion setting by EPR spin-trapping,207–209 and the oxidation of 1-hydroxy-3-carboxy-pyrrolidine (CP-H) to paramagnetic 3-carboxy-proxyl (CP).210 EPR spectroscopy has been used to detect ROS generation in isolated rat hearts that had been subjected to global ischemia followed by controlled reperfusion with solutions of different concentrations of O2 and with different ROS inhibitors.211 It was found that recovery of left ventricular function during reperfusion was inversely related to the oxygen radical burst, highlighting the importance of myocardial O2 tension during initial reperfusion. The protective effects of inhibitors of the Na1-H1 exchange pump were investigated in ischemia-reperfusion injury in guinea pig hearts, and found to directly quench ROS, improve cardiac performance and play an important role in apoptosis.212 In addition to ROS, recent EPR studies, that measured nitrosylhaem, have shown that nitrite-mediated NO-haem formation occurs progressively during myocardial ischemia with these complexes serving as a store of dNO with concordant activation of NO-signalling pathways.213 Further studies from the same laboratory examined the activation of soluble guanylyl cyclase from organic nitrate reduction by xanthine oxidase under anaerobic conditions.214 Organic nitrates have been used clinically in the treatment of ischemic heart disease for more than a century. This study reveals that organic nitrite is the initial product in the process of xanthine oxidase-mediated organic nitrate biotransformation, and is the precursor of dNO and nitrosothiols, serving as the link between organic nitrate and soluble guanylyl cyclase activation. Investigation of the inﬂuence of deep hypothermia (4 1C) during ischemiareperfusion in the isolated rat heart was studied with EPR spectroscopy to detect ascorbyl radicals, as markers of radical production.215 EPR analysis showed that the post-ischemic release of ascorbyl radicals decreased in the hypothermia-treated group, and suggest that the protective effect of hypothermia against functional injury caused by ischemia-reperfusion may decrease the radical burst at reperfusion. EPR detection of ascorbyl radical has also been used as a marker of oxidative stress to monitor potential adverse effects of fatty

220

Electron Paramag. Reson., 2007, 20, 192–244

acids on myocardial ischemic-reperfusion injury although direct measurement of ROS was not made.216 The nature of any oxidative stress in the failing human heart and the activity of antioxidant enzyme systems are incompletely understood. Ventricular myocardium from failing, explanted human hearts in patients with nonischemic dilated cardiomyopathy at the time of heart transplant was used to examine whether ROS production and antioxidant enzyme activity or expression were altered in end-stage human heart failure.217 EPR with the spin trap 5diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide demonstrated that formation of O2d was increased more than 2-fold in the failing myocardium with concomitant increases in MnSOD mRNA, consistent with the thesis that there is increased oxidative stress in failing myocardium that leads to increase transcription of antioxidant enzymes. EPR spin trapping was also applied to demonstrate that oxidative stress contributes to the progression of hypertrophy and chronic heart failure218,219 and are more prevalent in the aging heart.220 EPR spectroscopy using the spin labels 5-doxylstearic acid and 16-doxylstearic acid in erythrocytes has been employed to show that the ﬂuidity of erythrocyte membranes from patients with heart failure was decreased near the membrane surface and may be related to the reduced tissue supply of O2 seen in patients with heart failure.221 Atrial ﬁbrillation (AF) is associated with an increased risk of stroke. EPR and superoxide dismutase-inhibitable cytochrome c reduction assays have been utilised to demonstrate that AF increases O2d production in both the left atrium and left atrial appendage (LAA).222 Increased NAD(P)H oxidase and xanthine oxidase activities contributed to the observed increase in LAA O2d production. This increase in O2d and its reactive metabolites may contribute to the pathological consequences of AF such as thrombosis, inﬂammation, and tissue remodelling. A study in dogs used EPR to monitor ascorbyl radical concentrations as a measure of oxidative stress and dNO production following epicardial direct current shocks for deﬁbrillation.223 They found that dNO contributes to radical generation and nitrosative injury after epicardial shocks, and that NOS inhibitors decrease radical generation by inhibiting the production of peroxynitrite. The spin trap a-phenyl-N-tert-butylnitrone (PBN) has been studied as a novel therapeutic in a rat model of acute cardiac allograft transplantation.224 EPR spectroscopy revealed nitrosylation of myocardial haem proteins in untreated allografts, which was decreased by treatment with PBN. PBN also decreased iNOS protein and iNOS mRNA, whilst RT-PCR analysis revealed enhanced cytokine gene expression for interferon-gamma, interleukin-6, and interleukin-10 in untreated allografts. Expression of these genes was potently inhibited or abolished in recipients treated with PBN. Further studies have investigated ﬁbroblast remodelling of the heart,225 and the rate of dNO release from glyceryl trinitrate, in the treatment of angina pectoris.226 The techniques of EPR oximetry and EPR imaging are given separate sections in this chapter to reﬂect their importance in the biomedical applications of EPR. However, they have found particular application to the

Electron Paramag. Reson., 2007, 20, 192–244

221

study of the heart and the group of Zweier has pioneered the use of these techniques in this setting. They have provided an excellent review of the cardiac applications of EPR imaging227 and studies utilising EPR oximetry in the beating heart.228,229 7.5 Oxidative Stress in other Disease Settings. – Oxidative damage mediated by reactive oxygen- and nitrogen species have been described in many different organisms, and in different organs and tissues. This oxidative stress has largely been measured by EPR spin trapping. A good summary of this area, which also provides experimental considerations, has appeared.230 Due to the interest of the role of oxidative stress in diseases and the possible role of red blood cells in oxidative stress, redox reactions of hemoglobin have become important. A review focused on the important contributions of EPR to our understanding of hemoglobin redox reactions has shown how EPR not only identiﬁes the paramagnetic species formed, but can also be used to provide insights into the mechanism involved in these redox reactions.231 7.6 Exercise and High Altitude Stress. – Exhaustive exercise has been shown to increase oxidant stress and several studies have reported increased radical ﬂux in human muscle232 and rat brain233 during exercise. A study of cerebral hypoperfusion as a precursor to oxidant stress and dementia used EPR spectroscopy with PBN and DMPO spin traps to detect dOH generation (detected using DMSO via the trapping of dCH3). Direct detection of the ascorbyl radical has been reported to be useful in monitoring chronic cerebral hypoperfusion.234 Combined molecular and neuroimaging techniques have been used to determine if radical-mediated damage to barrier function in hypoxia results in extracellular edema, raised intracranial pressure and accounts for the neurological symptoms typical of high-altitude headache (also known as acute mountain sickness).235 EPR spectroscopy identiﬁed a clear increase in the blood and CSF concentration of oxygen and carbon-centred radicals, subsequently identiﬁed as lipid-derived alkoxyl and alkyl species. Magnetic resonance imaging (MRI) demonstrated a mild increase in brain volume that resolved within 6 h of normoxic recovery. However, there was no detectable evidence for gross barrier dysfunction, elevated lumbar pressures, T2 prolongation or associated neuronal and astroglial damage. Thus, the study concludes that radical mediated vasogenic edema is not an important pathophysiological event that contributes to the mild brain swelling observed in acute mountain sickness. An earlier study236 also concluded that although antioxidant prophylaxis decreased the concentration of carbon-centred radicals at high altitude this did not inﬂuence markers of inﬂammation, appetite-related peptides, ad libitum nutrient intake or body composition. Thus radicals do not appear to be involved in the inﬂammatory response and subsequent control of eating behaviour at high altitude. 7.7 Brain Injury. – Studies of experimental traumatic brain injury (TBI) suggest both deleterious and protective effects of inducible NO synthase (iNOS). In

222

Electron Paramag. Reson., 2007, 20, 192–244

iNOS-deﬁcient mice, enhanced oxidative stress has been detected after traumatic brain injury, supporting a neuroprotective role of iNOS.237 Evaluation of ascorbate prophylaxis for ischemia-reperfusion injury during vascular surgery found it may have promoted iron-induced oxidative lipid damage via a Fentontype reaction initiated during the ischemic phase of surgery.238 7.8 Other Systems. – EPR spectroscopy has been applied to the study of tissue damage in kidney,239,240 liver,241 airway obstruction242 and inborn errors in metabolism.243 An interesting study on the mechanism of antimalarial activity of clotrimazole through its role in inhibiting haemoperoxidases, and inducing oxidative stress in Plasmodium falciparum, identiﬁed the one-electron oxidation product of clotrimazole by spin trapping with DMPO. The results indicate that the antimalarial activity of clotrimazole might be due to the inhibition of haemoperoxidases and subsequent development of oxidative stress in P. falciparum.244 Oxidative stress was measured in peripheral blood of thalassaemic patients245 using an innovative hydroxylamine ‘radical probe’ to detect oxygen-radicals. The study suggests the usefulness of the approach to study new strategies of chelation, new chelators, or the efﬁcacy of antioxidant formula in patients affected by thalassaemia. The impact of environmental pollutants and their bioaccumulation has been investigated in ﬁsh and other marine creatures with EPR techniques.246–250

8

Apoptosis

Programmed cell death or apoptosis is an important biological regulator, controlling responses such as embryonic development, inﬂammation, and cell growth. Not surprisingly, inappropriate or inhibited apoptosis has a plethora of pathological consequences. Apoptotic cells display certain characteristics, such as increased phosphatidylserine (PS) on their surface that can be monitored by annexin binding. A more quantitative measure of PS externalization has been described utilizing paramagnetic iron as the reporter molecule for annexin-PS binding.251 ROS are suspected to play a key role in Fas (CD95)induced cell death although the identity of speciﬁc species involved in this process are largely unknown. Medan and co-workers252 have established a role for H2O2 and dOH as key participants in Fas-induced cell death. They suggest that as the Fas death pathway is implicated in various inﬂammatory and immunologic disorders, utilization of antioxidants and apoptosis inhibitors as potential therapeutic agents may be advantageous. Induction of apoptosis could be a useful therapeutic strategy in cancer, and a number of groups have assessed the induction of apoptosis induced by ultrasound253,254 and high energy proton beams255 by monitoring oxidant production and DNA double strand breaks. Caspases are enzymes that mediate apoptosis pathways. The effect of different apoptosis-inducing stimuli on caspases has been assessed in human leukaemia HL60 cells.256 Hydroxyl radicals were generated in the culture medium after exposure to radiation or H2O2, and activity assay of

Electron Paramag. Reson., 2007, 20, 192–244

223

caspases revealed that caspase-3, -8 and -9 were activated 2 h after exposure to H2O2, whereas in irradiated cells caspase-3 and -9 activation occurred 4 h after exposure; increased caspase-8 activation was not observed. The results suggest that, unlike the caspase cascade of H2O2-induced apoptosis, cytochrome c and caspase-9 are important for the intrinsic pathway of radiation-induced apoptosis, independent of caspase-8.

9

Free Radicals in Plants

9.1 Reactive Oxygen Species Production by Plants. – As with animal cells, reactive oxygen species (ROS) can be produced in plants by the leakage of electrons from electron transport chain systems. However, as well as mitochondria leakage, there is also ROS production from chloroplasts. In fact the effects of H2O2 on chloroplasts was being studied thirty years ago. Once excited, direct energy transfer from chlorophyll can produce singlet oxygen, or alternatively there can be a single electron reduction of oxygen at Photo System I (PSI), in the Mehler reaction. After the O2d is formed, H2O2 and dOH can result. Chloroplast ROS generation is especially high under high light conditions.257 Furthermore, there are dedicated systems in plant cells for the production of ROS,258 and the search for these has largely centred on the search for homologues of the mammalian NADPH oxidase system. However, unlike mammalian systems where a complex of several proteins is involved, the plant enzyme appears to have one large polypeptide that shows homology to the gp91-phox protein. In Arabidopsis there are 10 such isoforms,259 with multiple copies also found in other plant species.260 Other enzymes found to be able to contribute to the overall ROS generation in plants include peroxidases, which are enzymes that might be involved in ROS removal as well as it generation.258 The role of ROS in plants is very wide ranging. ROS may simply have an anti-microbial effect, relying on their inherent destructive nature, although this does not seem to be the case in all infections. However, it is now clear that ROS can partake in cell signalling, and such events may be modulated by the complement of antioxidants in, or around, cells. These antioxidants include ﬂavonoids, polyamines, glutathione and ascorbic acid, amongst others, and certainly the presence of key antioxidants in plants give the antioxidant effects so highly debated for the human diet. Some of the key events in plants which are controlled, perhaps in part, by ROS, include stomatal closure, root growth, and programmed cell death, including the hypersensitive response. Certainly profound effects on gene expression have been shown.261 With the generation of O2d by oxidases, the presence of H2O2 is likely to be inevitable, either from the dismutation of the O2d, catalysed by superoxide dismutase or more spontaneously at low pH, or from dedicated enzymatic sources such as peroxidases. Further reactions would lead to the production of d OH. Although the signalling role of these radicals in plants has not had the

224

Electron Paramag. Reson., 2007, 20, 192–244

attention of other ROS, they have been shown to be involved. For example, d OH generation is promoted by auxin, and the formation of this species has been shown to mediate both root gravitropism and elongation growth in maize (reviewed in 258).

9.2 The Measurement of ROS in Plants by EPR. – There are many ways to access the production and presence of ROS in plant cells and tissues, most of which are based on the systems developed for animal studies, but one of the methods of choice is the use of ﬂuorescent dyes262 in conjunction with confocal microscopy. However, nearly all these methods are beset with problems, especially the speciﬁcity of the dyes and probes, giving doubt as to the robust nature of the measurements. However, not all ROS are open to being measured by EPR, and EPR therefore has been restricted mainly to the measurement of d OH, and occasionally O2d. The use of EPR in such measurements has recently been reviewed.263,264 EPR has major advantages over other methods. Using the spin trap DEPMPO the detection of O2d is reported to be 40 times more sensitive than, for example, the cytochrome c assay, which is commonly used.264 However, care needs to be exercised when using biological materials, as the consumption of O2 will reduce the potential signal obtained, but this can be overcome by the use of gas-permeable (e.g. Teﬂon) tubes or ﬂat cells.264,265 The generation and detection of dOH has often employed a system using the spin trap a-(4-pyridyl-1-oxide)-N-tert-butylnitrone (4-POBN) and ethanol. This trap was ﬁrst used by Janzen266 and although previously used in biological systems,267 has not been widely adopted for in vivo studies. The reaction of dOH with ethanol produces a carbon-centred a-hydroxyethyl radical, which leads to the formation of a POBN adduct, which being relatively long-lived, means that a characteristic EPR spectrum is detected; this allows an estimation of dOH captured by the trap from measuring the signal height. POBN will react with both dOH and O2d, but the adduct formed from its reaction with O2d is so unstable that it would not normally appear in the spectrum. However, if the concentration of POBN is too high, then new adducts may form which confuses the speciﬁcity of the trap.264 To ensure speciﬁcity of the reaction, the experiments can be repeated in the presence of dOH scavengers, such as thiourea, adenine or Na-salicylate. As an example, in one study 850 mM ethanol with 50 mM POBN was used, and the measurements made at room temperature, in a ﬂat cell at 63 mW microwave power, 100 KHz modulation frequency, 0.2 mT modulation amplitude, 9.687 GHz microwave frequency using a Bruker ESP 300 X-band spectrophotometer.268 Hydroxyl radicals could be shown to be produced from coleoptiles and roots.268,269 It was proposed that dOH could be generated in the cell wall, and that such radicals would partake in the cleavage of wall polymers. This would subsequently mediate cell wall loosening and extension growth. To investigate this hypothesis, maize (Zea mays L.) coleoptiles and sunﬂower (Helianthus

Electron Paramag. Reson., 2007, 20, 192–244

225

annuus L.) hypocotyls were used, and the generation of dOH investigated using EPR and ethanol/POBN spin trapping,268 with the generation of dOH driven by the addition of externally applied NADH which would potentially produce ROS in a peroxidase-mediated manner. A similar mechanism has been proposed for roots, and using primary roots of maize (Zea mays) seedlings and the ethanol/POBN system as a trapping method, the production of dOH could be demonstrated.269 The EPR signal was reduced if dead roots were used, or if the experiment was repeated in the presence of dOH scavengers. Furthermore, the signal was reduced by the presence of SOD or catalase, indicating that the dOH was emanating from a chemical route that involved O2d and H2O2. However, using DEPMPO spin traps with isolated plasma membranes from maize roots, which enabled the detection of both O2d and dOH, it was shown that dOH could be produced without the initial generation of O2d.270 It was suggested that O2d was produced in a manner that was sensitive to both the NADPH oxidase inhibitor diphenylene iodonium and O2, and was dependent on NADH as a reductant. On the other hand, dOH could be generated in a manner that was insensitive to O2, and under conditions when the O2d synthase system was inhibited or not active.270 Further work on the isolated cell walls of pea, also using DEPMPO as a spin-trap, showed that dOH generation as also not dependent on NADH as a reductant, and it was proposed that neither plasma membranes or a source of O2d was required. Clearly the application of EPR, which allows the measurement of more than one radical under the same conditions, should enable the mechanisms of production to be unravelled. Recently, a reﬁnement of the method which allows localisation of dOH in plant tissues has been reported,271 using the ethanol/POBN system. Studying the roots of cucumber and Arabidopsis, Renew et al.271 were able to vary the position of the plant tissue inside the spectrophotometer, and therefore only record a signal from speciﬁc sections of the tissue. Of particular interest is that the roots were still intact and undamaged, and therefore this method could be adopted for further in vivo measurements. Furthermore, by the use of mutants of the NADPH oxidase-homologues in Arabidopsis, in particular the root hair defective 2 mutant272, they could demonstrate the involvement of this enzyme in d OH formation. This would ﬁt with the previous observations that the dOH was emanating from O2d, as the O2d would be a likely product of NADPH oxidase activity in these tissues. Other studies using EPR to investigate the involvement of ROS in plant growth and development include the work of Kateria et al.273 This group investigated the relationship between the hormone cytokinin and UV-B light. UV-B enhanced the levels of ROS, and these could be measured by EPR, and therefore the role of ROS in the UV-B induced inhibition of cytokininpromoted expansion growth could be studied, using cucumber cotelydons as a model system. Exposure of previously hypoxically grown roots to air caused a rapid two-fold increase in radical concentration, as shown using EPR by Garnczarska et al.274 In a related study by the same group, using lupine

226

Electron Paramag. Reson., 2007, 20, 192–244

(Lupinus luteus L., cv. Juno) seedlings, hypoxic pre-treatment for 12 and 24 h and then exposure to air caused an increase in the level of radicals, again measured using EPR.275 Naglreiter et al.276 used EPR to measure ROS generation from seeds and concluded that increased ROS production did not correlate with increased germination rates in their system. Interestingly, ROS were shown to be involved in the digestion of prey by pitcher plants, where DMPO was used as a spin trap in an EPR study.277 Finally, instead of using spin traps that form a stable adduct that can be measured, spin probes can be used, where the signal is reduced on reaction with the target radical. This approach allows localisation studies to be carried out, as probes can be chosen that are able to localise in membranes, or in the soluble phase. Such an approach has been used in maize, where the spin probes used were Tempone and 7-DS.264 The use of scavengers and inhibitors further allows the investigation of the likely radicals involved in the reactions with the probes. 9.3 The Production of Nitric Oxide by Plant Cells. – In a manner similar to the search for ROS producing enzymes in plants, which was based on the assumption that they would be similar to those characterised in mammals, the early work on the elucidation of the enzyme responsible for NO production in plants was based on the search for a dNO synthase enzyme. However, while the search was still taking place, the whole genome of the model plant Arabidopsis was being completed and there was no evidence of a gene with enough homology to mammalian NOS to make one in plants readily identiﬁable. It was proposed that the protein must contain small domains that conferred on the protein the function of NOS activity, without the need for a similar overall structure. Using a snail gene as a basis for a search, a plant NOS-like enzyme was ﬁnally identiﬁed,278 but it lacks the obvious domains that would seem to be required for oxidation and reduction of arginine to citrulline and dNO. In other work, the enzyme nitrate reductase was shown to be able to produce NO, both in vitro and in vivo.279 dNO emanating from this enzyme was subsequently shown to be important in dNO mediated events in plants, such as stomatal closure. The role of dNO in plant parallels that found for ROS. There are a myriad functions in which NO has been found to be involved, and certainly there is a wide range of signalling in which dNO appears to play a part.280 Downstream signalling from dNO appears to be of a similar type to that seen in animals, and includes cGMP and S-nitrosylation, often leading to gene expression.281,282 9.4 The Measurement of dNO in Plants by EPR. – In a study of transgenic alfalfa root cultures with modiﬁed haemoglobin levels (both increased and decreased) by expressing sense and antisense barley haemoglobin transcripts, Dordas et al.283 used EPR measurements to assess the presence of dNO. The dNO spin trap Fe21-(MGD)2 was used, which was prepared by mixing N-(dithiocarbamoyl)-N-methyl-D-glucamine with fresh FeSO4. On reaction with NO, the stable Fe(II)–(MGD)2NO spin adduct is formed, which gives a

Electron Paramag. Reson., 2007, 20, 192–244

227

characteristic EPR spectrum. In these experiments Dordas et al.283 used a Bruker EMX spectrometer (Billerica, MA, USA) set for microwave power 20 mW, modulation amplitude 0.4 mT, modulation frequency 100 kHz, microwave frequency 9.243 GHz, receiver gain of 2 x 104 and magnetic ﬁeld centred at 325.5 mT with a 10 mT scan range. They estimated that the detection limit was approximately 0.5 mM dNO. This study allowed the authors to postulate that a major role of haemoglobin in plants may be to modulate dNO levels. Vanin et al.284 measured dNO in the leaves of Arabidopsis thaliana and Vicia faba. By using different probes they were able to quantify dNO in membranes and in extracellular solution: diethyldithiocarbamate complexes was used for the former, while N-methyl-D-glucamine dithiocarbamate was used for the latter. Basal rates of dNO agreed with that measured by others, at approximately 1 nmol g1 h1, but they suggested that much of the dNO produced by the cells is removed by the presence of O2d, and correcting for this they suggested that the true basal rate of dNO production was more likely to be approximately 18 nmol g1 h1. Nitrite increased this rate, while nitrate decreased it, and the authors suggested that dNO was being generated by the enzyme nitrate reductase. Furthermore, dNO can form dinitrosyl iron complexes (DNIC) which are paramagnetic with endogenous free iron and thiolcontaining complexes. The EPR signal seen has a g value of g ¼ 2.037 and gJ ¼ 2.012 (gav B 2.03). When Chinese rose leaves were treated with nitrite a signal was indeed detected, which was decreased with co-treatment with nitrate. Similar EPR signals were also seen when such leaves were treated with dNO gas, suggesting that the EPR spectrum seen was indeed from DNIC caused by the presence of dNO. EPR measurements of dNO has also been carried out to investigate its subcellular localization in leaves from young and senescent pea plants (Pisum sativum).285 The spin trap used was Fe(MGD)2 and a characteristic three-line spectrum (g ¼ 2.05 and aN ¼ 12.8 G), indicating the formation of the stable NO-Fe(MGD)2 complex was shown using peroxisomes. As a positive control isolated neuronal dNO synthase (nNOS) was used, whereas pre-treatment of the peroxisome samples with L-NAME, a NOS inhibitor, or leaving out the reductant NADPH, resulted in lower signal intensities. The formation of the adduct MGD-Fe-NO was used to study the role of dNO during the germination of sorghum seeds.286 Under conditions where the substrates of NR and NOS were not limiting, the authors claimed to be the ﬁrst to use EPR to assess the activities of both these enzyme activities simultaneously in a tissue that was actively growing. Removal of dNO in plants may involve haemoglobins, as discussed above, and these were the subject of an EPR study by Ioanitescu et al.287 With a combination of continuous wave EPR along with UV/Vis and resonance Raman spectroscopy, they showed that the deoxy form was a mixture of pentaand hexacoordinate haem, but the ferric form was predominantly bis-histidine ligated. They further revealed some detail of the planar arrangement within the protein, and compared this tomato haemoglobin with others, giving clues to the mechanisms involved in protein action.

228

Electron Paramag. Reson., 2007, 20, 192–244

9.5 Other Processes in Plants Studied by EPR. – As well as studying the production of signalling molecules such as ROS and dNO in plants, other systems can be investigated using EPR too. Photosynthesis involves the excitation of electrons by photosystems, and EPR has often used to investigate this activity. Haddy et al.288 used EPR to investigate the effects of azide. More recent studies on the same topic, using both continuous wave and pulsed EPR, showed that there was an azide binding site in the immediate vicinity of a Mn cluster. As there appeared to be competitive binding between azide and chloride, it was suggested that the chloride-binding site was proximal to the Mn site.289 Electron transfer in the photosystems has been investigated using EPR.290 Razeghifard et al.291 used time-resolved EPR spectroscopy at room temperature to study the photosystem of the cyanobacterium Acaryochloris marina (A. marina), as this photosytem contains chlorophyll d. A sequence of electron transfer events was also proposed for the spinach photosytem II, again with the use of time-resolved EPR.292 They focused on the triplet state within the photosystem, and in particular with reference to the reduction of the ironquinone complex, while Semin et al.293 focused on hydrogen bonding. Other work on photosystems include studies involving comparison with X-ray diffraction structural data to give an insight into the mechanisms of water oxidation in photosynthetic O2 evolution,294 the mechanisms involved in O2,295 the loss of O2 evolution on heating,296 structural studies,297 including the possible structure of the Mn(4) cluster,298 and using EPR during calcium depletion.299 As ammonia is a structural analogue for water binding to photosystem II, and an inhibitor of water oxidation, Fang et al.300 investigated the effects on this of ethylene glycol and methanol, to give a better understanding of the mechanisms involved in the action of this photosystem. The effects of other compounds have also been examined. Jajoo et al.301 investigated the relationship and interaction of the effects of chloride and formate on photosystem II, while others have examined at iodide action302 and cadmium inhibition.303 Using pulsed EPR techniques Gregor et al.304 investigated the role of a protein that limits access of small molecules to the metal site of photosystem II. The so called 33 kDa manganese-stabilizing extrinsic protein binds to the lumen side of the photosystem II, close to the Mn(4)Ca cluster of the O2 evolving complex. Other extrinsic proteins to the photosystem II have also been investigated using EPR. For example, a 23kDa protein was shown to contain Mn at room temperature305 and Jeschke et al.,306 looked at the structural details of the major light-harvesting chlorophyll a/b protein (LHCIIb) by use of spin labels placed in various domains of the protein. Other enzymes have also come under the scrutiny of EPR, including those that are receptive to light, such as the blue-light sensitive photoreceptor, phototropin, which is a ﬂavoprotein that is involved in the regulation of the phototropism response of higher plants.307 EPR has also been used for studies in plants that involve proteins not involved in light perception. For example, along with Raman spectroscopy, EPR was used to investigate the enzyme sulﬁte oxidase from Arabidopsis thaliana, and was shown to have characteristics

Electron Paramag. Reson., 2007, 20, 192–244

229

similar to that already reported for the vertebrate enzyme.308 The same group also studied the Mo(V) centre of this enzyme using EPR.309 They found three different Mo(V) EPR signals, which depended on pH of the solution and the way the Mo(V) state was generated. Such studies give an insight to the conﬁrmation of the enzyme, in this case centred around Arg-374, and it was proposed that the enzyme had at least two conformations, an open or closed state, allowing the mechanism of the catalysis to be further unravelled. The cytosolic form of the metalloenzyme glyoxalase 2 (GLX2-2) from Arabidopsis thaliana, has been the subject of EPR studies. A mixture of metal-loaded and metal-unloaded enzymes could be identiﬁed, and metal loading-showed positive co-operativity. A comprehensive study of the enzyme glyoxalase 2 was carried out by Marasinghe et al.,310 including EPR, 1H NMR spectroscopy and X-ray crystallographic studies. Glyoxalase 2 is an enzyme containing a beta-lactamase fold. They cloned, over-expressed and puriﬁed a mitochondrial isozyme (GLX2-5) from Arabidopsis thaliana, and using EPR found that recombinant GLX2-5 contains several metal centres, including a predominant Fe(III)Zn(II) centre and an anti-ferromagnetically coupled Fe(III)Fe(II) centre. However, unlike the cytosolic isoform GLX2-5 did not appear to bind manganese. 10

Selected Biomedical Techniques

10.1 In vivo EPR. – Developments in low-frequency EPR instrumentation have allowed larger biological samples to be studied and imaged. Most of the instrumental development has been at L-Band ( B 1 GHz) or radio frequencies (B300 MHz). Various detection techniques can be used for low frequency EPR. In addition to the conventional continuous-wave experiment, lowfrequency pulsed EPR, longitudinal detection and indirect detection via the Overhauser enhancement in NMR have been developed for biomedical samples. The EPR signals are detected in the microwave resonator and several designs including ‘loop-gap’ and ‘birdcage’ resonators have been developed. These techniques have been applied successfully across a range of experimental applications that may now be translated into potential clinical applications. These include EPR oximetry, imaging and dosimetry that are brieﬂy reviewed below. Several excellent reviews have recently been written on instrumental aspects,311 potential clinical applications,312,313 monitoring drug delivery,314 and pH measurements.315 In vivo EPR techniques have been applied to the study of oxidative and nitrosative stress in the brain,316 skin317 and during sepsis.318 10.2 EPR Oximetry. – The critical fact that allows the EPR oximetry experiment is that the ground state of molecular oxygen has two unpaired electrons. These two unpaired electrons will interact via Heisenberg exchange with other unpaired electrons in the system in a manner that is a function of the amount of O2 present. The resulting EPR spectra of the unpaired-spin system will be thus be broadened and the degree of broadening is directly proportional to the concentration or partial pressure of O2. While this effect occurs with all paramagnetic

230

Electron Paramag. Reson., 2007, 20, 192–244

materials, it is much greater in some and these have been selected as O2 probes for EPR oximetry. The EPR signal arising from these probes is a sensitive reporter of pO2 or [O2] at that location. Typically, the EPR spectral linewidth is measured and converted to pO2 or [O2] using an appropriate calibration curve. The choice of material to use as probes for EPR oximetry depends on the system and parameter to be measured. Both soluble and particulate O2-sensitive EPR probes have been developed and rely on a physical interaction with O2 that does not affect the local O2 concentration. This is a distinct advantage over many other methods which consume O2 in the process used to measure its concentration. EPR oximetry has been developed by several groups, with the laboratory of Swartz particularly active in its application.319 This group has utilised EPR oximetry to measure myocardial O2 tension in isolated perfused rat hearts and compared the experimental data to the steady-state model of O2 supply to living tissue developed by Krogh and Erlang.320 The same group have used EPR oximetry in concert with near-infrared spectroscopy,321 ﬂuorescence ﬁbre-optic sensors,322 and in comparison with Eppendorf polarographic electrodes323 to study oxygenation in the brain during normoxia and hyperoxia. In their study pO2 and regional blood ﬂow in a rabbit model of limb ischemia, they demonstrate, for the ﬁrst time, the applicability of EPR oximetry in animals larger than rodents.324 Measurement of pO2 in the mammary gland pad and femoral muscle of female mice was achieved using EPR oximetry and implanted crystals of lithium phthalocyanine (lipc) as O2 probes.325 This study further validates the use of lipc crystals and EPR oximetry for long-term non-invasive assessment of pO2 in tissues, underscores the importance of maintaining normal body core temperature during the measurements, and demonstrates that mammary tissue functions at a lower pO2 level than muscle in female mice. Ilangovan and colleagues used EPR oximetry in the beating heart and found that the myocardial O2 consumption rate during ischemia-reperfusion is a reliable index of postischemic recovery.326 Nitroxides have been used as probes of O2 supersaturated solutions327 but a recent study, using nuclear spin-relaxation spectroscopy, found the proton spin-lattice relaxation rate induced by O2 is dependent on solvent and that nitroxide based EPR oximetry may suffer a reference concentration shift of an order of two in aqueous media.328 To aid the biocompatibility of oxygensensitive probes in EPR oximetry, ﬁlms of polytetraﬂuoroethylene polymers holding lithium phthalocyanine crystals have been developed and used to report on pO2 in mice and rabbits over time.These have been shown to be promising tools for future pre-clinical and clinical developments of EPR oximetry.329 Development of EPR oximetry for use in the clinical setting is a goal of current research and the applicability of India ink as a suitable probe for pO2 measurements in such clinical studies has been shown.330 10.3 EPR Imaging. – By applying magnetic ﬁeld gradients across the sample, spatial information in the EPR spectra can be obtained and spectroscopic imaging achieved. EPR imaging technology has progressed rapidly over recent years allowing many biomedical applications to be introduced. Several approaches including continuous wave and time-domain modes as well as

Electron Paramag. Reson., 2007, 20, 192–244

231

Overhauser-enhanced imaging and spin echo have been described. 331–334 However, the imaging time taken to acquire a given number of projections may limit the in vivo applications of the technique where relatively rapid kinetics occur. Therefore, a major effort in recent years has been to develop approaches to increase data acquisition times in EPR imaging.335–339 Nonetheless, the potential for EPR imaging to obtain physiological information in a non-invasive manner has seen a rapid growth in applications of the method including applications in cardiology,340 cancer,341 renal injury,342,343 brain injury344 and pharmacology.345 References 1. J.-L. Clement and P. Tordo, Chapter in this Volume, p. 29. 2. M.J. Burkitt, Electron Paramagnetic Resonance, vol. 19, Royal Society of Chemistry, Cambridge, Chapter 9, 2004, pp. 33–81. 3. J.B. Feix and C.S. Klug, Chapter in this Volume, p. 50. 4. C.F. Becker, K. Lausecker, M. Balog, T. Kalai, K. Hideg, H.J. Steinhoﬀ and M. Engelhard, Magn. Reson. Chem., 2005, 43, S34. 5. Y.R. Chen, C.L. Chen, L. Zhang, K.B. Green-Church and J.L. Zweier, J. Biol. Chem., 2005, 280, 37339. 6. W. Liu, G.A. Rosenberg, H. Shi, T. Furuichi, G.S. Timmins, L.A. Cunningham and K.J. Liu, Arch. Biochem. Biophys., 2004, 426, 11. 7. M.D. Rees, C.L. Hawkins and M.J. Davies, Biochem. J., 2004, 381, 175. 8. W.J. O’Brien, C. Krema, T. Heimann and H. Zhao, Invest. Ophthalmol. Vis. Sci., 2006, 47, 853. 9. D.A. Komarov, I.A. Slepneva, V.V. Glupov and V.V. Khramtsov, Free Rad. Res., 2005, 39, 853. 10. V. Brezova, S. Gabcova, D. Dvoranova and A. Stasko, J. Photochem. Photobiol. B., 2005, 79, 121. 11. M. Velayutham, F.A. Villamena, J.C. Fishbein and J.L. Zweier, Arch. Biochem. Biophys., 2005, 435, 83. 12. J. Martins, L. Almeida and J. Laranjinha, Photochem. Photobiol., 2004, 80, 267. 13. V.E. Kagan, Y.Y. Tyurina, V.A. Tyurin, N.V. Konduru, A.I. Potapovich, A.N. Osipov, E.R. Kisin, D. Schwegler-Berry, R. Mercer, V. Castranova and A.A. Shvedova, Toxicol. Lett. 2006, 165, 88. 14. G. Ilangovan, C.D. Venkatakrishnan, A. Bratasz, S. Osinbowale, A.J. Cardounel, J.L. Zweier and P. Kuppusamy, Am. J. Physiol. Cell Physiol., 2006, 290, C313. 15. A. Valavanidis, K. Fiotakis, E. Bakeas and T. Vlahogianni, Redox Rep., 2005, 10, 37. 16. A. Valavanidis, T. Vlahoyianni and K. Fiotakis, Free Rad. Res., 2005, 39, 1071. 17. R. Lubart, G. Kesler, R. Lavie and H. Friedmann, Photomed. Laser Surg., 2005, 23, 369. 18. K.J. Reszka, M. Takayama, R.H. Sik, C.F. Chignell and I. Saito, Photochem. Photobiol., 2005, 81, 573. 19. M.C. Teixeira, J.P. Telo, N.F. Duarte and I. Sa-Correia, Biochem. Biophys. Res. Commun., 2004, 19, 1101. 20. M. DiCicco, R. Compton and S.A. Jansen-Varnum, J. Biomed. Mater. Res. B. Appl. Biomater., 2005, 15, 146. 21. P.H.P. Groeneveld, K.M.C. Kwappenberg, J.A.M. Langermans, P.H. Nibbering and L. Curtis, Cytokine., 1997, 9, 138.

232

Electron Paramag. Reson., 2007, 20, 192–244

22. P. Vallance, S. Patton, K. Bhagat, R. MacAllister, M. Radomski, S. Moncada and T. Malinski, Lancet, 1995, 346, 153. 23. P.E. James, K.J. Liu and H.M. Swartz, Adv. Exp. Med., 1997, 454, L181. 24. H. Fujii, J. Koscielniak and L.J. Berliner, Mag. Res. Medicine, 1997, 38, 565. 25. P.E. James, M. Miyake and H.M. Swartz, Nitric oxide; Chemistry and Biology, 1999, 3, 292. 26. A.M. Komarov and C-S. Lai, Biochim. Biophys. Acta, 1995, 1272, 29. 27. J. Weaver, S. Porasuphatana, P. Tsai, T. Budzichowski and G.M. Rosen, Free Rad. Res., 2005, 39, 1027. 28. J.P. Khoo, N.J. Alp, J.K. Bendall, S. Kawashima, M. Yokoyama, Y.H. Zhang, B. Casadei and K.M. Channon, Nitric Oxide, 2004, 10, 156. 29. H. Bayir, V.E. Kagan, G.G. Borisenko, Y.Y. Tyurina, K.L. Janesko, V.A. Vagni, T.R. Billiar, D.L. Williams and P.M. Kochanek, J. Cereb. Blood Flow Metab., 2005, 25, 673. 30. A.F. Vanin, A.A. Papina, V.A. Serezhenkov and W.H. Koppenol, Nitric Oxide, 2004, 10, 60. 31. A.L. Kleschyov, M. Oelze, A. Daiber, Y. Huang, H. Mollnau, E. Schulz, K. Sydow, B. Fichtlscherer, A. Mulsch and T. Munzel, Circ. Res., 2003, 93, 104. 32. A.F. Vanin, A.P. Poltorakov, V.D. Mikoyan, L.N. Kubrina and E. van Faassen, Nitric Oxide. 2006, 15, 295. 33. H. Kosaka and T. Shiga, In Methods in Nitric Oxide Research 1996 (M. Freelisch and J.S. Stamler eds.), John Wiley and Sons Ltd., Chapter 25, p. 373–381. 34. K. Tsuchiya, Y. Kanematsu, M. Yoshizumi, H. Ohnishi, K. Kirima, Y. Izawa, M. Shikishima, T. Ishida, S. Kondo, S. Kagami, Y. Takiguchi and T. Tamaki, Am. J. Physiol. Heart Circ. Physiol., 2005, 288, H2163. 35. C. Gautier, E. Van Faassen, I. Mikula, P. Martasek and A. Slama-Schwok, Biochem. Biophys. Res. Commun., 2006, 341, 816. 36. B. Piknova, M.T. Gladwin, A.N. Schechter and N. Hogg, J. Biol. Chem., 2005, 280, 40583. 37. A.B. Shekhter, V.A. Serezhenkov, T.G. Rudenko, A.V. Pekshev and A.F. Vanin, Nitric Oxide, 2005, 12, 210. 38. J. Joseph, B. Kalyanaraman and J.S. Hyde, Biochem. Biophys. Res. Comms., 1993, 192, 926. 39. A. Barchowsky, L.R. Klei, E.J. Dudek, F.A. Gesek, H.M. Swartz and P.E. James, Free Rad. Biol. Med., 1999, 27, 1405. 40. S.S. Qader, I. Lundquist, M. Ekelund, R. Hakanson and A. Salehi, Regul. Pept., 2005, 128, 51. 41. J. Jimenez-Feltstrom, I. Lundquist and A. Salehi, Cell Tissue Res., 2005, 319, 221. 42. C-S. Lai and A.M. Komarov, FEBS Letters, 1994, 345, 120. 43. A.M. Komarov, D.A. Wink, M. Feelish and H.H. Schmidt, Free Rad. Biol. Med., 2000, 28, 739. 44. K. Turpaev, C. Bouton and J.C. Drapier, Biochemistry, 2004, 43, 10844. 45. M. Soler, M. Camacho and L. Vila, J. Infect. Dis., 2003, 188, 927. 46. P.E. James and H.M. Swartz, Methods in Enzymology Volume 359 (L. Packer and E. Cadenas, eds) in: Nitric Oxide (Parts D and E), 52-66 (2002). 47. H.M. Swartz and R.B. Clarkson, Phys. Med. Biol., 1998, 43, 1957. 48. H.M. Swartz in: Biological Magnetic Resonance-Volume 20: In vivo EPR (ESR): Theory and Applications (Book). L.J. Berliner (ed.), Plenum Publishing Co., NY, 2002.

Electron Paramag. Reson., 2007, 20, 192–244

233

49. D. Szczygiel, J. Pawlus, P.M. Plonka, M. Elas, M. Szczygiel, B.K. Plonka and S.J. Lukiewicz, Nitric Oxide, 2004, 11, 279. 50. II. Vlasova, V.A. Tyurin, A.A. Kapralov, I.V. Kurnikov, A.N. Osipov, M.V. Potapovich, D.A. Stoyanovsky and V.E.Kagan. J. Biol. Chem. 2006, 281, 14554. 51. A.C. Gorren, M. Sorlie, K.K. Andersson, S. Marchal, R. Lange and B. Mayer, Methods Enzymol., 2005, 396, 456. 52. D.K. Ghosh, M.A. Holliday, C. Thomas, J.B. Weinberg, S.M. Smith and J.C. Salerno, J. Biol. Chem., 2006, 281, 14173. 53. M. Fasano, A. Bocedi, M. Mattu, M. Coletta and P. Ascenzi, J. Biol. Inorg. Chem., 2004, 9, 800. 54. E.S. Lima, M.G. Bonini, O. Augusto, H.V. Barbeiro, H.P. Souza and D.S. Abdalla, Free Rad. Biol. Med., 2005, 39, 532. 55. B. Gaston, D. Singel, A. Doctor and J.S. Stamler Am. J. Respir. Crit. Care Med. 2006, 173, 1186. 56. T.J. McMahon, G.S. Ahearn, M.P. Moya, A.J. Gow, Y.C. Huang, B.P. Luchsinger, R. Nudelman, Y. Yan, A.D. Krichman, T.M. Bashore, R.M. Caliﬀ, D.J. Singel, C.A. Piantadosi, V.F. Tapson and J.S. Stamler, Proc. Natl. Acad. Sci U S A, 2005, 102, 14801. 57. B.P. Luchsinger, E.N. Rich, Y. Yan, E.M. Williams, J.S. Stamler and D.J. Singel, J. Inorg. Biochem., 2005, 99, 912. 58. Y.R. Chen, C.L. Chen, A. Yeh, X. Liu and J.L. Zweier, J. Biol. Chem. 2006, 281, 13159. 59. F. MacMillan, K. Budiman, H. Angerer and H. Michel, FEBS Lett., 2006, 580, 1345. 60. D.A. Svistunenko, Biochim. Biophys. Acta., 2005, 1707, 127. 61. M. Fabian, L. Skultety, D. Jancura and G. Palmer, Biochim. Biophys. Acta., 2004, 1655, 298. 62. D.A. Svistunenko, M.T. Wilson and C.E. Cooper, Biochim. Biophys. Acta., 2004, 1655, 372. 63. T.P. Barrows, B. Bhaskar and T.L. Poulos, Biochemistry, 2004, 43, 8826. 64. K. Budiman, A. Kannt, S. Lyubenova, O.M. Richter, B. Ludwig, H. Michel and F. MacMillan, Biochemistry, 2004, 43, 11709. 65. B. Burlat, J.D. Gwyer, S. Poock, T. Clarke, J.A. Cole, A.M. Hemmings, M.R. Cheesman, J.N. Butt and D.J. Richardson, Biochem. Soc. Trans., 2005, 33, 137. 66. M. Burkitt, C. Jones, A. Lawrence and P. Wardman, Biochem. Soc. Symp., 2004, 71, 97. 67. L.L. Pearce, A.J. Kanai, M.W. Epperly and J. Peterson, Nitric Oxide, 2005, 13, 254. 68. H. Li, X. Liu, H. Cui, Y.R. Chen, A.J. Cardounel and J.L. Zweier, J. Biol. Chem., 2006, 281, 12546. 69. M.P. Pietila, P.K. Vohra, B. Sanyal, N.L. Wengenack, S. Raghavakaimal and C.F. Thomas, Am. J. Respir. Cell Mol. Biol. 2006, 35, 236. 70. K. Matsuura, S. Yoshioka, T. Tosha, H. Hori, K. Ishimori, T. Kitagawa, I. Morishima, N. Kagawa and M.R. Waterman, J. Biol. Chem., 2005, 280, 9088. 71. R. Silaghi-Dumitrescu and C.E. Cooper, Dalton Trans., 2005, 3477. 72. R. Davydov, R. Perera, S. Jin, T.C. Yang, T.A. Bryson, M. Sono, J.H. Dawson and B.M. Hoﬀman, J. Am. Chem. Soc., 2005, 127, 1403. 73. C. Jung, V. Schunemann, F. Lendzian, A.X. Trautwein, J. Contzen, M. Galander, L.H. Bottger, M. Richter and A.L. Barra, Biol. Chem., 2005, 386, 1043.

234

Electron Paramag. Reson., 2007, 20, 192–244

74. C. Jung, V. Schunemann and F. Lendzian, Biochem. Biophys. Res. Commun., 2005, 338, 355. 75. A. Franke, G. Stochel, N. Suzuki, T. Higuchi, K. Okuzono and R. Eldik Van, J. Am. Chem. Soc., 2005, 127, 5360. 76. V.K. Praneeth, E. Haupt and N. Lehnert, J. Inorg. Biochem., 2005, 99, 940. 77. W.J. O’Brien, C. Krema, T. Heimann and H. Zhao, Invest. Ophthalmol. Vis. Sci., 2006, 47, 853. 78. R. Ameziane-El-Hassani, S. Morand, J.L. Boucher, Y.M. Frapart, D. Apostolou, D. Agnandji, S. Gnidehou, R. Ohayon, M.S. Noel-Hudson, J. Francon, K. Lalaoui, A. Virion and C. Dupuy, J. Biol. Chem., 2005, 280, 30046. 79. A. Keller, A. Mohamed, S. Drose, U. Brandt, I. Fleming and R.P. Brandes, Free Rad. Res., 2004, 38, 1257. 80. I.R. Hanna, L.L. Hilenski, A. Dikalova, Y. Taniyama, S. Dikalov, A. Lyle, M.T. Quinn, B. Lassegue and K.K. Griendling, Free Rad. Biol. Med., 2004, 37, 1542. 81. K. Laude, H. Cai, B. Fink, N. Hoch, D.S. Weber, L. McCann, G. Kojda, T. Fukai, H.H. Schmidt, S. Dikalov, S. Ramasamy, G. Gamez, K.K. Griendling and G. Harrison, Am. J. Physiol. Heart Circ. Physiol., 2005, 288, H7. 82. A. Mouithys-Mickalad, G. Deby-Dupont, M. Mathy-Hartert, Y. Habraken, M. Nys, Y. Henrotin, M. Lamy and C. Deby, Biochem. Biophys. Res. Commun., 2004, 318, 941. 83. K.E. Wyche, S.S. Wang, K.K. Griendling, S.I. Dikalov, H. Austin, S. Rao, B. Fink, D.G. Harrison and A.M. Zafari, Hypertension, 2004, 43, 1246. 84. W. Chamulitrat, W. Stremmel, T. Kawahara, K. Rokutan, H. Fujii, K. Wingler, H.H. Schmidt and R. Schmidt, J. Invest. Dermatol., 2004, 122, 1000. 85. A. Jancso, Z. Paksi, N. Jakab, B. Gyurcsik, A. Rockenbauer and T.M Gajda, Dalton Trans., 2005, 3187. 86. A. Barik, B. Mishra, L. Shen, H. Mohan, R.M. Kadam, S. Dutta, H.Y. Zhang and K.I. Priyadarsini, Free Rad. Biol. Med., 2005, 39, 811. 87. O. Vajragupta, P. Boonchoong and L.J. Berliner, Free Rad. Res., 2004, 38, 303. 88. M. Scarpellini, A.J. Wu, J.W. Kampf and V.L. Pecoraro, Inorg. Chem., 2005, 44, 5001. 89. C. Karunakaran, H. Zhang, J.P. Crow, W.E. Antholine and B. Kalyanaraman, J. Biol. Chem., 2004, 279, 32534. 90. R. Cejudo-Marin, G. Alzuet, S. Ferrer, J. Borras, A. Castineiras, E. Monzani and L. Casella, Inorg. Chem., 2004, 43, 6805. 91. Q.X. Li, Q.H. Luo, Y.Z. Li and M.C. Shen, Dalton Trans., 2004, 2329. 92. S.S. Ali, J.I. Hardt, K.L. Quick, J.S. Kim-Han, B.F. Erlanger, T.T. Huang, C.J. Epstein and L.L. Dugan, Free Rad. Biol. Med., 2004, 37, 1191. 93. G. Wang, R.C. Conover, S. Benoit, A.A. Olczak, J.W. Olson, M.K. Johnson and R.J. Maier, J. Biol. Chem., 2004, 279, 51908. 94. N.L. Wengenack, B.D. Lane, P.J. Hill, J.R. Uhl, G.S. Lukat-Rodgers, L. Hall, G.D. Roberts, F.R. Cockerill 3rd, P.J. Brennan, K.R. Rodgers, J.T. Belisle and F. Rusnak. Protein Expr. Purif. 2004, 36, 232. 95. J. Huang, D.B. Kim-Shapiro and S.B. King, J. Med. Chem., 2004, 47, 3495. 96. A. Mouithys-Mickalad Mareque, J.M. Faez, L. Chistiaens, S. Kohnen, C. Deby, M. Hoebeke, M. Lamy and G. Deby-Dupont, Redox Rep., 2004, 9, 81. 97. K.I. Matsumoto, A. Suzuki, H. Washimi, A. Hisamatsu, A. Okajo, I. Ui and K. Endo. J Nutr Biochem. 2006, 17, 677.

Electron Paramag. Reson., 2007, 20, 192–244

235

98. J. Uwayama, A. Hirayama, T. Yanagawa, E. Warabi, R. Sugimoto, K. Itoh, M. Yamamoto, H. Yoshida, A. Koyama and T. Ishii, Biochem. Biophys. Res. Commun., 2006, 339, 226. 99. D.M. Bailey, S. Raman, J. McEneny, I.S. Young, K.L. Parham, D.A. Hullin, B. Davies, G. McKeeman, J.M. McCord and M.H. Lewis, Free Rad. Biol. Med., 2006, 40, 591. 100. B. Poljsak, Z. Gazdag, S. Jenko-Brinovec, S. Fujs, M. Pesti, J. Belagyi, S. Plesnicar and P. Raspor, J. Appl. Toxicol., 2005, 25, 535. 101. J.M. May, Z.C. Qu, S. Juliao and C.E. Cobb, Free Rad. Res., 2005, 39, 195. 102. D. Champion, D. Simatos, E.P. Kalogianni, P. Cayot and M. Le Meste, J. Agric. Food Chem., 2004, 52, 3399. 103. S. Veljovic-Jovanovic, B. Kukavica, T. Cvetic, M. Mojovic and Z. Vucinic, Ann. N Y Acad. Sci., 2005, 1048, 500. 104. M. Stapelberg, N. Gellert, E. Swettenham, M. Tomasetti, P.K. Witting, A. Procopio and J. Neuzil, J. Biol. Chem., 2005, 280, 25369. 105. B.M. Rezk, G.R. Haenen, W.J. Van Der Vijgh and A. Bast, Biochim. Biophys. Acta., 2004, 1683, 16. 106. A.C. Diaz, A.V. Fernandez Gimenez, S.N. Mendiara and J.L. Fenucci, J. Agric. Food Chem., 2004, 52, 3189. 107. V. Patel, I.V. Chivukula, S. Roy, S. Khanna, G. He, N. Ojha, A. Mehrotra, L.M. Dias, T.K. Hunt and C.K. Sen, Antioxid Redox Signal., 2005, 7, 1377. 108. B. Zhou, L.M. Wu, L. Yang and Z.L. Liu, Free Rad. Biol. Med., 2005, 38, 78. 109. H. Botti, M. Trujillo, C. Batthyany, H. Rubbo, G. Ferrer-Sueta and R. Radi, Chem. Res. Toxicol., 2004, 17, 1377. 110. R.G. Andrade Jr, J.S. Ginani, G.K. Lopes, F. Dutra, A. Alonso and M. HermesLima. Biochimie. 2006, 88, 1287. 111. R.G. Andrade Jr, L.T. Dalvi, J.M. Silva Jr, G.K. Lopes, A. Alonso and M. Hermes-Lima, Arch. Biochem. Biophys., 2005, 437, 1. 112. D. Bagchi, S. Roy, V. Patel, G. He, S. Khanna, N. Ojha, C. Phillips, S. Ghosh, M. Bagchi and C.K. Sen, Mol. Cell Biochem., 2006, 281, 197. 113. M. Khan, S. Varadharaj, J.C. Shobha, M.U. Naidu, N.L. Parinandi, V.K. Kutala and P. Kuppusamy, J. Cardiovasc. Pharmacol., 2006, 47, 9. 114. A. Valavanidis, C. Nisiotou, Y. Papageorgiou, I. Kremli, N. Satravelas, N. Zinieris and H. Zygalaki, J. Agric. Food Chem., 2004, 52, 2358. 115. A. Bertelli, M. Falchi, R. Lo Scalzo and R. Morelli, Drugs Exp. Clin. Res., 2004, 30, 111. 116. A. Bertelli, R. Morelli and M. Falchi, Drugs Exp. Clin. Res., 2004, 30, 177. 117. R. Morelli, S. Das, A. Bertelli, R. Bollini, R. Lo Scalzo, D.K. Das and M. Falchi, Mol. Cell Biochem., 2006, 282, 65. 118. A. Huisman, A. Van De Wiel, T.J. Rabelink and E.E. Van Faassen, J. Nutr. Biochem., 2004, 15, 426. 119. F. Ferroni, A. Maccaglia, D. Pietraforte, L. Turco and M. Minetti, J. Agric. Food Chem., 2004, 52, 2866. 120. S. Venkataraman, F.Q. Schafer, G.R. Buettner, Antioxid Redox Signal. 2004, 6 (3), 631. 121. A. Hay, M.J. Burkitt, C.M. Jones and R.C. Hartley, Arch. Biochem. Biophys., 2005, 435, 336. 122. R.A. Anderson, G.R. Ellis, Y.Y. Chirkov, A.S. Holmes, N. Payne, D.J. Blackman, S.K. Jackson, M.J. Lewis, J.D. Horowitz and M.P. Frenneaux, Eur. J. Heart Fail., 2004, 6, 47.

236

Electron Paramag. Reson., 2007, 20, 192–244

123. A.K. Nightingale, P.J. James, J. Morris-Thurgood, H. Fraser, R. Tong, S.K. Jackson, J.R. Cockcroft and M.P. Frenneaux, Am. J. Physiol. Heart Circ. Physiol., 2001, 280, H1334. 124. N. Khan, O. Grinberg, C. Wilmot, H. Kiefer and H.M. Swartz, J. Biochem. Biophys. Methods., 2005, 62, 125. 125. E. Damiani, P. Astolﬁ, L. Cionna, F. Ippoliti and L. Greci, Free Rad. Res., 2006, 40, 485. 126. O. Sergent, A. Tomasi, D. Ceccarelli, A. Masini, H. Nohl, P. Cillard, J. Cillard, Y.A. Vladimirov and A.V. Kozlov, Biometals, 2005, 18, 567. 127. B. Sreedhar, R. Subramaniyan and K.M. Nair, Biochem. Biophys. Res. Commun., 2004, 318, 992. 128. H. Zheng, M.B. Youdim, L.M. Weiner and M. Fridkin, J. Pept. Res., 2005, 66, 190. 129. Y. Hayashi, Y. Ueda, A. Nakajima and Y. Mitsuyama, Brain Res., 2004, 1025, 29. 130. Y. Ueda, J.I. Noor, K. Nagatomo, T. Doi, T. Ikeda, A. Nakajima and T. Ikenoue, Exp. Brain Res., 2006, 169, 117. 131. G. Zhao, P. Arosio and N.D. Chasteen, Biochemistry, 2006, 45, 3429. 132. M. Su, S. Cavallo, S. Stefanini, E. Chiancone and N.D. Chasteen, Biochemistry, 2005, 44, 5572. 133. E. Yavin, A.K. Boal, E.D. Stemp, E.M. Boon, A.L. Livingston, V.L. O’Shea, S.S. David and J.K. Barton, Proc. Natl. Acad. Sci. U S A, 2005, 102, 3546. 134. E. Yavin, E.D. Stemp, V.L. O’shea, S.S. David and J.K. Barton, Proc. Natl. Acad. Sci. U S A, 2006, 103, 3610. 135. M.K. Bowman, D. Becker, M.D. Sevilla and J.D. Zimbrick, Radiat. Res., 2005, 163, 447. 136. R. Bittl and S. Weber, Biochim. Biophys. Acta., 2005, 1707, 117. 137. J.J. Briede, R.W. Godschalk, M.T. Emans, T.M. De Kok, E. Van Agen, J. Van Maanen, F.J. Van Schooten and J.C. Kleinjans, Free Rad. Res., 2004, 38, 995. 138. A. Pezeshk, Life Sci., 2004, 74, 2423. 139. K. Akamatsu, K. Fujii and A. Yokoya, Int. J. Radiat. Biol., 2004, 80, 849. 140. B.B. Hasinoﬀ, X. Wu, A. Begleiter, L.J. Guziec, F. Guziec Jr, A. Giorgianni, S. Yang, Y. Jiang and J.C. Yalowich, Cancer Chemother. Pharmacol., 2006, 57, 221. 141. A. Valavanidis, T. Vlahoyianni and K. Fiotakis, Free Rad. Res., 2005, 39, 1071. 142. T. Shi, R. Duﬃn, P.J. Borm, H. Li, C. Weishaupt and R.P. Schins, Environ. Res., 2006, 101, 18. 143. A. Valavanidis, K. Fiotakis, E. Bakeas and T. Vlahogianni, Redox Rep., 2005, 10, 37. 144. S.S. Tonde, A.S. Kumbhar, S.B. Padhye and R.J. Butcher, J. Inorg. Biochem., 2006, 100, 51. 145. R. Cejudo, G. Alzuet, M. Gonzalez-Alvarez, J.L. Garcia-Gimenez, J. Borras and M. Liu-Gonzalez, J. Inorg. Biochem., 2006, 100, 70. 146. T. Gupta, S. Dhar, M. Nethaji and A.R. Chakravarty, Dalton Trans., 2004, 1896. 147. P.A. Reddy, B.K. Santra, M. Nethaji and A.R. Chakravarty, J. Inorg. Biochem., 2004, 98, 377. 148. K.J. Reszka, M. Takayama, R.H. Sik, C.F. Chignell and I. Saito, Photochem. Photobiol., 2005, 81, 573. 149. D.C. Ramirez, S.E. Mejiba and R.P. Mason, Nat. Methods., 2006, 3, 123. 150. M.J. Davies and C.L. Hawkins, Free Rad. Biol. Med., 2004, 36, 1072. 151. R.P. Mason, Free Rad. Biol. Med., 2004, 36, 1214. 152. A. Keszler, R.P. Mason and N. Hogg, Free Rad. Biol. Med., 2006, 40, 507.

Electron Paramag. Reson., 2007, 20, 192–244

237

153. K. Nakai and R.P. Mason, Free Rad. Biol. Med., 2005, 39, 1050. 154. C.D. Detweiler, O.M. Lardinois, L.J. Deterding, P.R. de Montellano, K.B. Tomer and R.P. Mason, Free Rad. Biol. Med., 2005, 38, 969. 155. D.C. Ramirez, S.E. Gomez Mejiba and R.P. Mason, Free Rad. Biol. Med., 2005, 38, 201. 156. C.L. Hawkins and M.J. Davies, Free Rad. Biol. Med., 2005, 39, 900. 157. M.L. Estevam, O.R. Nascimento, M.S. Baptista, P. Di Mascio, F.M. Prado, A. Faljoni-Alario, R. Zucchi Mdo and I.L. Nantes, J. Biol. Chem., 2004 279, 39214. 158. S. Rinalducci, E. Hideg, I. Vass and L. Zolla, Biochem. Biophys. Res. Commun., 2006, 341, 1105. 159. T.A. Rinaldi, I.L. Tersariol, F.H. Dyszy, F.M. Prado, O.R. Nascimento, P. Di Mascio and I.L. Nantes, Free Rad. Biol. Med., 2004, 36, 802. 160. S. Un, Magn. Reson Chem., 2005, 43, S229. 161. L. Kalman, A.J. Narvaez, R. LoBrutto, J.C. Williams and J.P. Allen, Biochemistry, 2004, 43, 12905. 162. A.J. Narvaez, R. LoBrutto, J.P. Allen and J.C. Williams, Biochemistry, 2004, 43, 14379. 163. B. Alvarez, V. Demicheli, R. Duran, M. Trujillo, C. Cervenansky, B.A. Freeman and R. Radi, Free Rad. Biol. Med., 2004, 37, 813. 164. D.A. Svistunenko and C.E. Cooper, Biophys. J., 2004, 87, 582. 165. M.D. Rees, C.L. Hawkins and M.J. Davies, Biochem. J., 2004, 381, 175. 166. M.D. Rees and M.J. Davies, J. Am. Chem. Soc., 2006, 128, 3085. 167. M.M. Corsaro, D. Pietraforte, A.S. Di Lorenzo, M. Minetti and G. Marino, Free Rad. Res., 2004, 38, 343. 168. N.E. Polyakov, T.V. Leshina, T.A. Konovalova, E.O. Hand and L.D. Kispert, Free Rad. Biol. Med., 2004, 36, 872. 169. S. Scalia, A. Molinari, A. Casolari and A. Maldotti, Eur. J. Pharm. Sci., 2004, 22, 241. 170. A. Molinari, R. Tursilli, A. Casolari, A. Maldotti and S. Scalia, Pharmazie., 2006, 61, 41. 171. E. Castagnino, M. Cangiotti, S. Tongiani and M.F. Ottaviani, J. Control Release., 2005, 108, 215. 172. K.G. Desai and H.J. Park, Drug Deliv., 2006, 13, 39. 173. K. Nakagawa and Y. Sato, Radiat Res., 2005, 164, 336. 174. F. De Proft, E. Pauwels, P. Lahorte, V. Van Speybroeck and M. Waroquier, P. Geerlings, Magn. Reson Chem., 2004, 42, S3. 175. E. Castagnino, M. Cangiotti, S. Tongiani and M.F. Ottaviani, J. Control Release., 2005, 108, 215. 176. M. Velayutham, F.A. Villamena, M. Navamal, J.C. Fishbein and J.L. Zweier, Chem. Res. Toxicol., 2005, 18, 970. 177. A. Papaioannou, M. Manos, S. Karkabounas, R. Liasko, A.M. Evangelou, I. Correia, V. Kalfakakou, J.C. Pessoa and T. Kabanos, J. Inorg. Biochem., 2004, 98, 959. 178. J. Shao, B. Zhou, L. Zhu, W. Qiu, Y.C. Yuan, B. Xi and Y. Yen, Cancer Res., 2004, 64, 1. 179. P. Perocco, G. Bronzetti, D. Canistro, L. Valgimigli, A. Sapone, A. Aﬀatato, G.F. Pedulli, L. Pozzetti, M. Broccoli, R. Iori, J. Barillari, V. Sblendorio, M.S. Legator, M. Paolini and S.Z. Abdel-Rahman. Mutat Res. 2006, 595, 125.

238

Electron Paramag. Reson., 2007, 20, 192–244

180. M. Stapelberg, N. Gellert, E. Swettenham, M. Tomasetti, P.K. Witting, A. Procopio and J. Neuzil, J. Biol. Chem., 2005, 280, 25369. 181. A. Marczak, A. Kowalczyk, A. Wrzesien-Kus, T. Robak and Z. Jozwiak, Cell Biol. Int., 2006, 30, 127. 182. A.P. Burlaka, E.P. Sidorik, II. Ganusevich and S.P. Osinsky, Exp. Oncol., 2006, 28, 49. 183. S. Loibl, J. Bratengeier, V. Farines, G. von Minckwitz, B. Spankuch, V. SchiniKerth, F. Nepveu, K. Strebhardt and M. Kaufmann, Pathol. Res. Pract., 2006, 202, 1. 184. A. Bratasz, I. Kuter, R. Konior, I. Goscinski and S. Lukiewicz, Antioxid. Redox Signal., 2004, 6, 613. 185. D. Szczygiel, J. Pawlus, P.M. Plonka, M. Elas, M. Szczygiel, B.K. Plonka and S.J. Lukiewicz, Nitric Oxide, 2004, 11, 279. 186. K. Ishii, A. Takayanagi, S. Shimizu, H. Abe, K. Sogawa and N. Kobayashi, Free Rad. Biol. Med., 2005, 38, 920. 187. A.P. Burlaka, Y.P. Sidorik, S.V. Prylutska, O.P. Matyshevska, O.A. Golub, Y.I. Prylutskyy and P. Scharﬀ, Exp. Oncol., 2004, 26, 326. 188. L.B. Feril Jr, Y. Tsuda, T. Kondo, Q.L. Zhao, R. Ogawa, Z.G. Cui, K. Tsukada and P. Riesz, Cancer Sci., 2004, 95, 181. 189. R. Satchi-Fainaro, M. Puder, J.W. Davies, H.T. Tran, D.A. Sampson, A.K. Greene, G. Corfas and J. Folkman, Nat. Med., 2004, 10, 255. 190. M.R. Dreher, M. Elas, K. Ichikawa, E.D. Barth, A. Chilkoti, G.M. Rosen, H.J. Halpern and M. Dewhirst, Med. Phys., 2004, 31, 2755. 191. G. Ilangovan, A. Bratasz, H. Li, P. Schmalbrock, J.L. Zweier and P. Kuppusamy, Magn. Reson. Med., 2004, 52, 650. 192. H. Hou, N. Khan, J.A. O’Hara, O.Y. Grinberg, J.F. Dunn, M.A. Abajian, C.M. Wilmot, M. Makki, E. Demidenko, S. Lu, R.P. Steffen and H.M. Swartz, Int. J. Radiat. Oncol. Biol. Phys., 2004, 59, 834. 193. M. Sentjurc, M. Cemazar and G. Sersa, Spectrochim Acta A Mol. Biomol. Spectrosc., 2004, 60, 1379. 194. B. Gallez, C. Baudelet and B.F. Jordan, NMR Biomed., 2004, 17, 240. 195. S. Cai, J. Khoo and K.M. Channon, Cardiovasc Res., 2005, 65, 823. 196. R.A. Anderson, L.M. Evans, G.R. Ellis, N. Khan, K. Morris, S.K. Jackson, A. Rees, M.J. Lewis and M.P. Frenneaux, Diabet. Med., 2006, 23, 258. 197. R.A. Anderson, G.R. Ellis, L.M. Evans, K. Morris, Y.Y. Chirkov, J.D. Horowitz, S.K. Jackson, A. Rees, M.J. Lewis and M.P. Frenneaux, Diab. Vasc. Dis. Res., 2005, 2, 88. 198. L.O. Royer, F.S. Knudsen, M.A. de Oliveira, M.F. Tavares and E.J. Bechara, Chem. Res. Toxicol., 2004, 17, 1725. 199. V. Monga, K.H. Thompson, V.G. Yuen, V. Sharma, B.O. Patrick, J.H. McNeill and C. Orvig, Inorg. Chem., 2005, 44, 2678. 200. M. Yamaguchi, K. Wakasugi, R. Saito, Y. Adachi, Y. Yoshikawa, H. Sakurai and A. Katoh, J. Inorg. Biochem., 2006, 100, 260. 201. A.V. Kozlov, M. Albrecht, E.M. Donnelly, M. Jafarmadar, Z. Szelenyi, H. Nohl, H. Redl and S. Bahrami, Vascul. Pharmacol., 2005, 43, 411. 202. N. Sjakste, L. Baumane, J.L. Boucher, M. Dzintare, D. Meirena, J. Sjakste, L. Lauberte and I. Kalvinsh, Basic Clin. Pharmacol. Toxicol., 2004, 94, 46. 203. M. Galleano, M. Simontacchi and S. Puntarulo, Mol. Aspects Med., 2004 25, 141. 204. P.E. James, M.P. Thomas and S.K. Jackson, NMR Biomed., 2004, 17, 319.

Electron Paramag. Reson., 2007, 20, 192–244

239

205. D.A. Svistunenko, N. Davies, D. Brealey, M. Singer, C.E. Cooper. Biochim. Biophys. Acta. 2006, 1757, 262. 206. K. Nakai, M.B. Kadiiska, J.J. Jiang, K. Stadler and R.P. Mason, Proc. Natl. Acad. Sci. U S A, 2006, 103, 4616. 207. M. Dambrova, E. Liepinsh, O. Kirjanova, R. Petrovska, O. Pugovich, L. Baumane, S. Uhlen, I. Kalvinsh, D. Oliver and J.E. Wikberg, J. Cardiovasc. Pharmacol., 2004, 44, 178. 208. A.L. Bulteau, K.C. Lundberg, M. Ikeda-Saito, G. Isaya and L.I. Szweda, Proc. Natl. Acad. Sci. U S A., 2005, 102, 5987. 209. N. Engberding, S. Spiekermann, A. Schaefer, A. Heineke, A. Wiencke, M. Muller, M. Fuchs, D. Hilﬁker-Kleiner, B. Hornig, H. Drexler and U. Landmesser, Circulation, 2004, 110, 2175. 210. S. Lecour, L. Rochette and L. Opie, Cardiovasc. Res., 2005, 65, 239. 211. M.G. Angelos, V.K. Kutala, C.A. Torres, G. He, J.D. Stoner, M. Mohammad and P. Kuppusamy, Am. J. Physiol. Heart Circ. Physiol., 2006, 290, H341. 212. Y. Hotta, H. Nishimaki, T. Takeo, G. Itoh, M. Yajima, H. Otsuka-Murakami, N. Ishikawa, N. Kawai, L. Huang, K. Yamada, S. Yamamoto, K. Matsui and N. Ohashi, Eur. J. Pharmacol., 2004, 503, 109. 213. E. Tiravanti, A. Samouilov and J.L. Zweier, J. Biol. Chem., 2004, 279, 11065. 214. H. Li, H. Cui, X. Liu and J.L. Zweier, J. Biol. Chem., 2005, 280, 16594. 215. S. Gambert, S. Bes-Houtmann, D. Vandroux, C. Tissier, C. Vergely-Vandriesse, L. Rochette and P. Athias, J. Heart Lung Transplant., 2004, 23, 487. 216. S. Gambert, C. Vergely, R. Filomenko, D. Moreau, A. Bettaieb, L.H. Opie and L. Rochette, Mol. Cell Biochem., 2006, 283, 147. 217. F. Sam, D.L. Kerstetter, D.R. Pimental, S. Mulukutla, A. Tabaee, M.R. Bristow, W.S. Colucci and D.B. Sawyer, J. Card Fail., 2005, 11, 473. 218. V. Mellin, M. Isabelle, A. Oudot, C. Vergely-Vandriesse, C. Monteil, B. Di Meglio, J.P. Henry, B. Dautreaux, L. Rochette, C. Thuillez and P. Mulder, Eur. Heart J., 2005, 26, 1544. 219. M. Kakishita, K. Nakamura, M. Asanuma, H. Morita, H. Saito, K. Kusano, Y. Nakamura, T. Emori, H. Matsubara, T. Sugaya, N. Ogawa and T.J. Ohe, J Cardiovasc. Pharmacol. 2003, 42, Suppl 1, S67-S70. 220. A.V. Kozlov, L. Szalay, F. Umar, K. Kropik, K. Staniek, H. Niedermuller, S. Bahrami and H. Nohl, Biochim. Biophys. Acta., 2005, 1740, 382. 221. O.V. Rodnenkov, O.G. Luneva, N.A. Ulyanova, G.V. Maksimov, A.B. Rubin, S.N. Orlov and E.I. Chazov, Pathophysiology, 2005, 11, 209. 222. S.C. Dudley Jr, N.E. Hoch, L.A. McCann, C. Honeycutt, L. Diamandopoulos, T. Fukai, D.G. Harrison, S.I. Dikalov and J. Langberg, Circulation, 2005, 112, 1266. 223. C.B. Clark, Y. Zhang, S.M. Martin, L.R. Davies, L. Xu, K.C. Kregel, F.J. Miller, G.R. Buettner and R.E. Kerber, Resuscitation, 2004, 60, 351. 224. G.M. Pieper, V. Nilakantan, X. Zhou, A.K. Khanna, C.P. Johnson, A.M. Roza, M.B. Adams, G. Hilton and C.C. Felix, J. Pharmacol. Exp. Ther., 2005, 312, 774. 225. I. Papparella, G. Ceolotto, L. Lenzini, M. Mazzoni, L. Franco, M. Sartori, L. Ciccariello and A. Semplicini, J. Hypertens., 2005, 23, 793. 226. G. Clermont, S. Lecour, C. Vergely, M. Zeller, C. Perrin, V. Maupoil, O. Bouchot and L. Rochette, Fundam. Clin. Pharmacol., 2003, 17, 709. 227. P. Kuppusamy and J.L. Zweier, NMR Biomed., 2004, 17, 226. 228. G. Ilangovan, T. Liebgott, V.K. Kutala, S. Petryakov, J.L. Zweier and P. Kuppusamy, Magn. Reson. Med., 2004, 51, 835. 229. X. Zhao, G. He, Y.R. Chen, R.P. Pandian, P. Kuppusamy and J.L. Zweier, Circulation, 2005, 111, 2966.

240

Electron Paramag. Reson., 2007, 20, 192–244

230. F.A. Villamena and J.L. Zweier, Antioxid. Redox Signal., 2004, 6, 619. 231. J.M. Rifkind, S. Ramasamy, P.T. Manoharan, E. Nagababu and J.G. Mohanty, Antioxid. Redox Signal., 2004, 6, 657. 232. D.M. Bailey, I.S. Young, J. McEneny, L. Lawrenson, J. Kim, J. Barden and R.S. Richardson, Am. J. Physiol. Heart Circ. Physiol., 2004, 287, H1689. 233. A. Toldy, K. Stadler, M. Sasvari, J. Jakus, K.J. Jung, H.Y. Chung, I. Berkes and C. Nyakas, Z. Radak. Brain Res. Bull., 2005, 65, 487. 234. S. Kasparova, V. Brezova, M. Valko, J. Horecky, V. Mlynarik, T. Liptaj, O. Vancova, O. Ulicna and D. Dobrota, Neurochem. Int., 2005, 46, 601. 235. D.M. Bailey, R. Roukens, M. Knauth, K. Kallenberg, S. Christ, A. Mohr, J. Genius, B. Storch-Hagenlocher, F. Meisel, J. McEneny, I.S. Young, T. Steiner, K. Hess and P. Bartsch, J. Cereb. Blood Flow Metab., 2006, 26, 99. 236. D.M. Bailey, P.N. Ainslie, S.K. Jackson, R.S. Richardson and M. Ghatei, Clin. Sci. (Lond)., 2004, 107, 589. 237. H. Bayir, V.E. Kagan, G.G. Borisenko, Y.Y. Tyurina, K.L. Janesko, V.A. Vagni, T.R. Billiar, D.L. Williams and P.M. Kochanek, J. Cereb. Blood Flow Metab., 2005, 25, 673. 238. D.M. Bailey, S. Raman, J. McEneny, I.S. Young, K.L. Parham, D.A. Hullin, B. Davies, G. McKeeman, J.M. McCord and M.H. Lewis, Free Rad. Biol. Med., 2006, 239, 591. 239. T. Oteki, S. Nagase, H. Yokoyama, H. Ohya, T. Akatsuka, M. Tada, A. Ueda, A. Hirayama and A. Koyama, Biochem. Biophys. Res. Commun., 2005, 332, 326. 240. T. Oteki, S. Nagase, H. Yokoyama, H. Ohya, T. Akatsuka, M. Tada, A. Ueda, A. Hirayama and A. Koyama, Biochem. Biophys. Res. Commun., 2005, 337, 48. 241. F.M. Megli and K. Sabatini, FEBS Lett., 2004, 573, 68. 242. P. Kramaric, Z. Pavlica, T. Koklic, A. Nemec, N.K. Erzen and M. Sentjurc, J. Chem. Inf. Model., 2005, 45, 1708. 243. L.O. Royer, F.S. Knudsen, M.A. de Oliveira, M.F. Tavares and E.J. Bechara, Chem. Res. Toxicol., 2004, 17, 598. 244. V. Trivedi, P. Chand, K. Srivastava, S.K. Puri, P.R. Maulik and U. Bandyopadhyay, J. Biol. Chem., 2005, 280, 41129. 245. A. Filosa, L. Valgimigli, G.F. Pedulli, A. Sapone, A. Maggio, D. Renda, C. Scazzone, R. Malizia, L. Pitrolo, C. Lo Pinto, Z. Borsellino, L. Cuccia, M. Capra, D. Canistro, M. Broccoli, A. Soleti and M. Paolini, Br. J. Haematol., 2005, 131, 135. 246. H. Shi, Y. Sui, X. Wang, Y. Luo and L. Ji, Comp. Biochem. Physiol. C Toxicol. Pharmacol., 2005, 140, 115. 247. H. Shi, X. Wang, Y. Luo and Y. Su, Aquat. Toxicol., 2005, 74, 365. 248. Y. Sun, H. Yu, J. Zhang, Y. Yin, H. Shi and X. Wang. Chemosphere. 2006, 63, 1319. 249. Y. Luo, Y. Su, R.Z. Lin, H.H. Shi and X.R. Wang. Chemosphere. 2006, 65, 1064 . 250. G. Malanga, M.S. Estevez, J. Calvo and S. Puntarulo, Aquat. Toxicol., 2004, 69, 299. 251. J.P. Fabisiak, G.G. Borisenko and V.E. Kagan, Methods Mol. Biol., 2005, 291, 457. 252. D. Medan, L. Wang, D. Toledo, B. Lu, C. Stehlik, B.H. Jiang, X. Shi and Y. Rojanasakul, J. Cell Physiol., 2005, 203, 78. 253. L.B. Feril Jr, T. Kondo, K. Takaya and P. Riesz, Int. J. Radiat. Biol., 2004 80, 165.

Electron Paramag. Reson., 2007, 20, 192–244

241

254. L.B. Feril Jr, T. Kondo, Z.G. Cui, Y. Tabuchi, Q.L. Zhao, H. Ando, T. Misaki, H. Yoshikawa and S. Umemura, Cancer Lett., 2005, 221, 145. 255. C. Di Pietro, S. Piro, G. Tabbi, M. Ragusa, V. Di Pietro, V. Zimmitti, F. Cuda, M. Anello, U. Consoli, E.T. Salinaro, M. Caruso, C. Vancheri, N. Crimi, M.G. Sabini, G.A. Cirrone, L. Raffaele, G. Privitera, A. Pulvirenti, R. Giugno, A. Ferro, G. Cuttone, S.L. Nigro, R. Purrello, F. Purrello and M. Purrello, Apoptosis, 2006, 11, 57. 256. Y. Hosokawa, Y. Sakakura, L. Tanaka, K. Okumura, T. Yajima and M. Kaneko, J. Radiat. Res. (Tokyo)., 2005, 46, 293. 257. P. Mullineaux and S. Karpinski, Curr. Opin. Plant Biol., 2002, 5, 42. 258. S.J. Neill, R. Desikan and J.T. Hancock, Curr. Opin. Plant Biol., 2002, 5, 388. 259. M.A. Torres, J.D. Jones and J.L. Dangl, Nat. Genet., 2005, 37, 1130. 260. H. Yoshioka, N. Numata, K. Nakajima, S. Katou, K. Kawakita, O. Rowland and J.D. Jones, N. Doke, Plant Cell, 2003, 15, 706. 261. S. Vanderauwera, P. Zimmermann, S. Rombauts, S. Vandenbeele, C. Langebartels, W. Gruissem, D. Inze´ and F. Van Breusegem, Plant Physiol., 2005, 139, 806. 262. A. Keller, A. Mohamed, S. Drose, U. Brandt, I. Fleming and R.P. Brandes, Free Rad. Res., 2004, 38, 1257. 263. E. Hideg, Methods Mol. Biol., 2004, 274, 249. 264. G. Bacic and M. Mojovic, Ann. NY Acad Sci., 2005, 1048, 230. 265. M. Mojovic, M. Vuletic and G.G. Bacic, Ann. N.Y. Acad. Sci., 2005, 1048, 471. 266. E.G. Janzen, Y.Y. Wang and R.V. J. Shetty, Am. Chem. Soc., 1978, 100, 2923. 267. C.L. Ramos, S. Pou, B.E. Britigan, M.S. Cohen and G.M. J. Rosen, Biol. Chem., 1992, 267, 8307. 268. A. Liszkay, B. Kenk and P. Schopfer, Planta, 2003, 217, 658. 269. A. Liszkay, E. van der Zalm and P. Schopfer, Plant Physiol., 2004, 136, 3114. 270. M. Mojovic, M. Vuletic, G.G. Bacic and Z.J. Vucinic, Exp. Bot., 2004, 55, 2523. 271. S. Renew, E. Heyno, P. Schopfer and A. Liskay, The Plant Journal, 2005, 44, 342. 272. J.W. Schiefelbein and C. Somerville, Plant Cell, 1990, 2, 235. 273. S. Kateria, K. Jain and K.N. Guruprasad, Indian J. Exp. Biol., 2005, 43, 910. 274. M. Garnczarska, W. Bednarski and Morkunas. J. Plant Physiol. 2004, 161, 415. 275. M. Garnczarska and W. Bednarski, Plant Physiol. Biochem., 2004, 42, 233. 276. C. Naglreiter, T.G. Reichenauer, B.A. Goodman and H.R. Bolhar-Nordenkampf, Plant Physiol. Biochem., 2005, 43, 117. 277. T.F. Chia, H.H. Aung, A.N. Osipov, N.K. Goh and L.S. Chia, Redox Rep., 2004, 9, 255. 278. F.Q. Guo, M. Okamoto and N.M. Crawford, Science, 2003, 302, 100. 279. P. Rockel, F. Strube, A. Rockel, J. Wildt and W.M. Kasier, Journal of Experimental Botany, 2002, 53, 103. 280. S.J. Neill, D. Desikan and J.T. Hancock, New Phytol., 2003, 159, 11. 281. C. Lindermayr, G. Saalbach and J. Durner, Plant Physiol., 2005, 137, 921. 282. C. Lindermayr, G. Saalbach, G. Bahnweg and J. Durner, J. Biol. Chem., 2006, 281, 4285. 283. C. Dordas, B.B. Hasinoﬀ, A.U. Igamberdiev, N. Manac’h, J. Rivoal and R.D. Hill, The Plant Journal, 2003, 35, 763. 284. A.F. Vanin, D.A. Svistunenko, V.D. Mikoyan, V.A. Serezhenkov, M.J. Fryer, N.R. Baker and C.E. Cooper, J. Biol. Chem., 2004, 279, 24100. 285. F.J. Corpas, J.B. Barroso, A. Carreras, M. Quiro´s, A.M. Leo´n, M.C. RomeroPuertas, F.J. Esteban, R. Valderrama, J.M. Palma, L.M. Sandalio, M. Go´mez and L.A. del Rı´ o, Plant Physiol., 2004, 136, 2722.

242

Electron Paramag. Reson., 2007, 20, 192–244

286. M. Simontacchi, S. Jasid and S. Puntarulo, Plant Science, 2004, 167, 839. 287. A.I. Ioanitescu, S. Dewilde, L. Kiger, M.C. Marden, L. Moens and S. Van Doorlaer, Biophys. J., 2005, 89, 2628. 288. A. Haddy, R. Allen Kimel and R. Thomas, Photosynth. Res., 2000, 63, 35. 289. H. Yu, C.P. Aznar, X. Xu and R.D. Britt, Biochemistry, 2005, 44, 12022. 290. C. Zhang, A. Boussac and A.W. Rutherford, Biochemistry, 2004, 43, 13787. 291. M.R. Razeghifard, M. Chen, J.L. Hughes, J. Freeman, E. Krausz and T. Wydrzynski, Biochemistry, 2005, 44, 11178. 292. W.O. Feikema, P. Gast, I.B. Klenina and I.I. Proskuryakov, Biochim. Biophys. Acta., 2005, 1709, 105. 293. B.K. Semin, E.R. Lovyagina, K.N. Timofeev, I.I. Ivanov, A.B. Rubin and M. Seibert, Biochemistry, 2005, 44, 9746. 294. H. Mino and S. Itoh, Photosynth. Res., 2005, 84, 333. 295. M.R. Raeghifard, D. Kuzek and R.J. Pace, Biochim. Biophys. Acta., 2005, 1708, 35. 296. M. Barra, M. Haumann and H. Dau, Photosynth. Res., 2005, 84, 231. 297. A. Kawamori, T.A. Ono, A. Ishii, S. Nakazawa, H. Hara, T. Tomo, J. Minagawa. R. Bittl and S.A. Dzuba, Photosynth. Res., 2005, 84, 187. 298. M.F. Charlot, A. Boussac and G. Blondin, Biochim. Biophys. Acta., 2005, 1708, 120. 299. H. Mino and S. Itoh, Biochim. Biophys. Acta., 2005, 1708, 42. 300. C.H. Fang, K.A. Chiang, C.H. Hung, K. Chang, S.C. Ke and H.A. Chu, Biochemistry, 2005, 44, 9758. 301. A. Jajoo, S. Bharti and A. Kawamori, J. Bioenerg. Biomembr., 2005, 37, 49. 302. D.I. Bryson, N. Doctor, R. Johnson, S. Baranov and A. Haddy, Biochemistry, 2005, 44, 7354. 303. K.G. Sigfridsson, G. Bernat, F. Mamedov and S. Styring, Biochim. Biophys. Acta., 2004, 1659, 19. 304. W. Gregor, R.M. Cinco, H. Yu, V.K. Yachandra and R.D. Britt, Biochemistry, 2005, 44, 8817. 305. N. Bondarava, P. Beyer and A. Krieger-Liszkay, Biochim. Biophys. Acta., 2005, 1708, 63. 306. G. Jeschke, A. Bender, T. Schweikardt, G. Panek, H. Decker and H. Paulsen, J. Biol. Chem., 2005, 280, 18623. 307. E. Schleicher, R.M. Kowalczyk, C.W. Kay, P. Hegemann, A. Bacher, M. Fischer, R. Bittl, G. Richter and S. Weber, J. Am. Chem. Soc., 2004, 126, 11067. 308. C. Hemann, B.L. Hood, M. Fulton, R. Hansch, G. Schwarz, R.R. Mendel, M.L. Kirk and R. J. Hille, Am. Chem. Soc., 2005, 127, 16567. 309. A.V. Astashkin, B.L. Hood, C. Feng, R. Hille, R.R. Mendel, A.M. Raitsimring and J.H. Enemark, Biochemistry, 2005, 44, 13274. 310. G.P. Marasinghe, I.M. Sandeer, G. Bennet, Periyannan, K.W. Yang, C.A. Makaroﬀ. M.W. Crowder, J. Biol. Chem. 2005, 280, 40668. 311. H. Fujii and L.J. Berliner, NMR Biomed., 2004, 17, 311. 312. H.M. Swartz, N. Khan, J. Buckey, R. Comi, L. Gould, O. Grinberg, A. Hartford, H. Hopf, H. Hou, E. Hug, A. Iwasaki, P. Lesniewski, I. Salikhov and T. Walczak, NMR Biomed., 2004, 17, 335. 313. S.K. Jackson, M.P. Thomas, S. Smith, M. Madhani, S.C. Rogers and P.E. James, Faraday Discuss. 2004, 126, 103-117; discussion 169-183. 314. D.J. Lurie and K. Mader, Adv. Drug. Deliv. Rev., 2005, 57, 1171. 315. V.V. Khramtsov, I.A. Grigor’ev, M.A. Foster and D.J. Lurie, Antioxid Redox Signal., 2004, 6, 667.

Electron Paramag. Reson., 2007, 20, 192–244

243

316. S. Liu, G.S. Timmins, H. Shi, C.M. Gasparovic and K.J. Liu, NMR Biomed., 2004, 17, 327. 317. K. Takeshita, C. Chi, H. Hirata, M. Ono and T. Ozawa, Free Rad. Biol. Med., 2006, 40, 876. 318. P.E. James, M.P. Thomas and S.K. Jackson, NMR Biomed., 2004 17, 319. 319. H.M. Swartz, Antioxid Redox Signal., 2004, 6, 677. 320. O. Grinberg, B. Novozhilov, S. Grinberg, B. Friedman and H.M. Swartz, Adv. Exp. Med. Biol., 2005, 566, 127. 321. Y.S. Sakata, O.Y. Grinberg, S. Grinberg, R. Springett and H.M. Swartz, Adv. Exp. Med. Biol., 2005, 566, 357. 322. J.A. O’Hara, H. Hou, E. Demidenko, R.J. Springett, N. Khan and H.M. Swartz, Physiol. Meas., 2005, 26, 203. 323. J.A. O’Hara, N. Khan, H. Hou, C.M. Wilmo, E. Demidenko, J.F. Dunn and H.M. Swartz, Physiol. Meas., 2004, 25, 1413. 324. O.Y. Grinberg, H. Hou, S.A. Grinberg, K.L. Moodie, E. Demidenko, B.J. Friedman, M.J. Post and H.M. Swartz, Physiol. Meas., 2004, 25, 659. 325. A. Matsumoto, S. Matsumoto, A.L. Sowers, J.W. Koscielniak, N.J. Trigg, P. Kuppusamy, J.B. Mitchell, S. Subramanian, M.C. Krishna and K. Matsumoto, Magn. Reson. Med., 2005, 54, 1530. 326. G. Ilangovan, T. Liebgott, V.K. Kutala, S. Petryakov, J.L. Zweier and P. Kuppusamy, Magn. Reson. Med., 2004, 51, 835. 327. A. Moscatelli, T.K. Chen, S. Jockusch, M.D. Forbes, N.J. Turro and M.F. Ottaviani, J. Phys. Chem. B. Condens. Matter Mater Surf. Interfaces Biophys., 2006, 110, 7574. 328. G. Diakova and R.G. Bryant, J. Magn. Reson., 2006, 178, 329. 329. M. Dinguizli, S. Jeumont, N. Beghein, J. He, T. Walczak, P.N. Lesniewski, H. Hou, O.Y. Grinberg, A. Sucheta, H.M. Swartz and B. Gallez, Biosens. Bioelectron., 2006, 21, 1015. 330. N. Khan, H. Hou, P. Hein, R.J. Comi, J.C. Buckey, O. Grinberg, I. Salikhov, S.Y. Lu, H. Wallach and H.M. Swartz, Adv. Exp. Med. Biol., 2005, 566, 119. 331. S. Subramanian, K. Matsumoto, J.B. Mitchell and M.C. Krishna, NMR Biomed., 2004, 17, 263. 332. C. Mailer, S.V. Sundramoorthy, C.A. Pelizzari and H.J. Halpern, Magn. Reson Med., 2006, 55, 904. 333. H. Hirata, T. Itoh, K. Hosokawa, Y. Deng and H. Susaki, J. Magn. Reson., 2005, 175, 177. 334. B.B. Williams, X. Pan and H.J. Halpern, J. Magn. Reson., 2005, 174, 88. 335. Y. Deng, P. Kuppusamy and J.L. Zweier, J. Magn. Reson., 2005, 174, 177. 336. R.H. Pursley, G. Salem, N. Devasahayam, S. Subramanian, J. Koscielniak, M.C. Krishna and T.J. Pohida, J. Magn. Reson., 2006, 178, 220. 337. R. Ahmad, B. Clymer, Y. Deng, G. He, D. Vikram, P. Kuppusamy and J.L. Zweier, J. Magn. Reson., 2006, 179, 263. 338. Y. Deng, G. He, S. Petryakov, P. Kuppusamy and J.L. Zweier, J. Magn. Reson., 2004, 168, 220. 339. N. Devasahayam, R. Murugesan, K. Matsumoto, J.B. Mitchell, J.A. Cook, S. Subramanian and M.C. Krishna, J. Magn. Reson., 2004, 168, 110. 340. P. Kuppusamy and J.L. Zweier, NMR Biomed., 2004, 17, 226. 341. T. Mikuni, G. He, S. Petryakov, M.M. Fallouh, Y. Deng, R. Ishihara, P. Kuppusamy, M. Tatsuta and J.L. Zweier, Cancer Res., 2004, 64, 6495.

244

Electron Paramag. Reson., 2007, 20, 192–244

342. A. Hirayama, S. Nagase, A. Ueda, T. Oteki, K. Takada, M. Obara, M. Inoue, K. Yoh, K. Hirayama and A. Koyama, Am. J. Physiol. Renal Physiol., 2005, 288, F597. 343. A. Hirayama and S. Nagase, Nephron. Clin. Pract., 2006, 103, c71. 344. H. Yokoyama, Y. Ueda, O. Itoh, T. Ikeda, J.I. Noor and T. Ikenoue, Magn. Reson. Imaging., 2004, 22, 1305. 345. K. Matsumoto, T. Yahiro, K. Yamada and H. Utsumi, Magn. Reson. Med., 2005, 53, 1158.

Progress in High-Field EPR of Inorganic Materials BY PETER C. RIEDI School of Physics and Astronomy, University of St Andrews, Fife KY16 9SS, UK

1

Introduction

This review continues earlier surveys1–5 of work using High Field EPR (HFEPR) and covers the years 2004 and 2005. Over the last 10 years there has been a rapid expansion of research using HFEPR, conventionally deﬁned as requiring frequencies at and above 90 GHz. In particular the introduction of commercial spectrometers, including pulsed spectrometers, at 90 GHz has allowed laboratories without mm-wave expertise to enter the ﬁeld. Meanwhile the specialist high-ﬁeld laboratories have continued to extend the range and ﬂexibility of their facilities, in some cases providing almost continuous coverage of a wide frequency range. The National High Magnetic Field Laboratory (NHMFL), Tallahassee, for example, now has tunable-frequency sources up to 700 GHz with a high stability resistive magnet providing a maximum ﬁeld of 25 T6–9 and Kobe10,11 and Osaka12 provide frequencies into the THz range combined with pulsed ﬁelds of more than 35 T. A series of articles on these facilities and many other aspects of HFEPR appear in a special issue of the journal Magnetic Resonance in Chemistry.13,14 Considering this review as a whole, we note an increased use of multifrequency spectra, combined with a ﬁt to the whole data set, in order to provide more accurate parameters for complicated materials with large Zero Field Splitting (ZFS). It is surprising, however, that the powerful combination of Inelastic Neutron Scattering (INS) and HFEPR is still not being used more extensively. INS is often used to study materials with large ZFS but combined with HFEPR it also provides the possibility to study dynamic effects since the time scale of INS (1012 s) is much shorter than that of HFEPR.15 [Notation: we shall frequently need to refer to the standard spin-Hamiltonian that, to second order reads, H ¼ mB.S.g.B þ DSz2 þ E(Sx2 Sy2).

Electron Paramagnetic Resonance, Volume 20 r The Royal Society of Chemistry, 2007 245

246

Electron Paramag. Reson., 2007, 20, 245–270

In some cases the values of D and E given in a paper were derived from a spectrum ﬁtting analysis including higher order terms. We shall indicate this by D* and E*. The abbreviations used are: AFM (antiferromagnetic), CW (continuous wave), DFT (Density Functional Theory), DM (Dzyaloshinskii–Moriya interaction), FM (ferromagnetic), LRO (Long Range Order), QMT (quantum mechanical tunnelling), SMM (single molecule magnet), ZFS (zero-ﬁeld splitting), 1D (one-dimensional chain).]

2

Materials Research

2.1 Fullerenes. – A review of the Jahn-Teller effect in fullerene ferromagnets has been given by Arcon and Blinc.16 Strong evidence was provided from resonance measurements that the orbital and spin ordering in TDAE-C60 is correlated and allows the determination of the Jahn-Teller dynamics. The degree of localization of the La ion in La @ C82 has been the subject of some controversy. CW and echo-detected HFEPR at 94 GHz from 10 K to room temperature showed17 that there were only small changes to the spectra below 80 K. The spectrum at 10 K could be simulated using a non-axial g-matrix showing that the La ion is positioned on the C2 axis of the fullerene cage. The multi-line pulsed ENDOR spectra, measured at 10 K, were only obtained after long signal-averaging times. The La hyperﬁne and quadrupole parameters were determined from a simulation of the ENDOR spectra. The La ion therefore has a unique binding site in C82 at low temperature and the largescale motion of the ion at room temperature has been frozen out. Sc3@C82 has been investigated18 by X- and W-band EPR. The isotropic and anisotropic hyperﬁne interaction between the electron spin and the 45Sc nuclei of the Sc3-triangle was evaluated. A statistical model of the 13C hyperﬁne coupling of the C82 cage was applied and its consequences for the charge transfer discussed. The observed EPR line-broadening with decreasing temperature was modelled with a simple exchange model that considers the molecular motion of the Sc3-triangle inside the C82 cage. This allowed the estimation of the linewidth contributions of the hyperﬁne- and g-anisotropies as well as the temperature dependence of the exchange rate, which increased rapidly with temperature and followed an exponential law. The excited-triplet states of a series of fullerene C60 mono- and bis-adducts were studied19 using HFEPR at 240 GHz. Fullerene derivatives are easier to process than the parent molecule and are therefore useful building blocks for the synthesis of C60-based materials. HFEPR was shown to have greatly improved resolution compared with X-band and provided accurate ZFS parameters and g-values. The excited state was populated using a pulsed Nd-YAG laser and TR-EPR used to record the spectra over the temperature range 50-270 K. The spectra showed both absorption and emission lines and at 240 GHz revealed the g-factor anisotropy that was not resolved at X-band. Ten adducts were studied and the values of D, E and g listed. The effect of the

Electron Paramag. Reson., 2007, 20, 245–270

247

number, position and nature of the substituents on the triplet parameters was discussed in detail. The monoanion of bis-trans2-[N-methyl-3,4-fulleropyrrolidine] was investigated20 by CW and pulsed EPR. X-band spectra showed signals due to 14N and 13 C hyperﬁne couplings. DFT calculations for different conformations of the pyrrolidinic rings agreed well with the experimental results when averaged over the different conformations. Accurate g-tensor principal values were obtained by HFEPR. The dynamical processes affecting the electron spin relaxation were also discussed. Porphyrin-fullerene complexes are attractive as electron donor-acceptor (D-A) systems for conversion of solar to chemical energy, due to the electron-donor properties of porphyrins (D) and the unique electron-accepting properties of fullerenes (A). An important requirement for an effective electrontransfer (ET) process is to minimize undesirable energy-wasting reactions such as charge recombination and energy transfer (EnT) from the donor to the acceptor. An axial coordination Zn-porphyrin-pyridylfullerene (ZnP-PyrF) complex has been studied21 by TR-EPR at X- and W-band. It was found that photoexcitation of the ZnP donor in frozen matrices at low temperature resulted mainly in singlet-singlet EnT to the pyridine-appended fullerene acceptor, but that in ﬂuid phases ET is the dominant process. 2.2 Semiconductors. – Zvanut22 has provided an extensive review of the use of HFEPR to investigate the defects found in electronic grade SiC substrates. The intrinsic defects observed in 4H-SiC were tabulated with their respective EPR parameters. While most work has been carried out on irradiated material, to increase the concentration of defects, three defects have also been observed in as-grown material. It was pointed out that many features of the spectra have not yet been identiﬁed with a particular defect structure. The advantage of HFEPR over X-band in resolving different defect structures has been shown by work at frequencies up to 240 GHz. An extensive review of EPR and related phenomena in III-V compounds has been given by Meisels.23 Three dimensional material is brieﬂy discussed but the emphasis of the review is on low-dimensional structures and quantum dots. The g-factors in these materials are usually small so this is high-ﬁeld but so far not high-frequency EPR, e.g. some 53 GHz at 10.2 T for AlGaAs. The signals are weak and EPR is often only observable by using the sample as a detector. In this case a double modulation technique is used to measure the change of the conductivity of the two-dimensional sample as a function of magnetic ﬁeld. Full details of the technique and the special problems of EPR in low-dimensional structures are given in the review. A study24 of semi-insulating GaAs in a new W-band spectrometer for optically detected HFEPR and ENDOR, using magnetic circular dichroism of the optical absorption (MCDA) as well as photo-luminescence-detected EPR, found the ﬁrst MCDA-detected paramagnetic EL21 defect. The higherorder effects, which prevented the unambiguous analysis of previous MCDAdetected spectra, were suppressed at W-band. The paramagnetic EL21 defect

248

Electron Paramag. Reson., 2007, 20, 245–270

was found to be an As antisite defect, which has four almost equivalent nearest 75 As neighbours differing less than 1.5% in the superhyperﬁne interactions, suggestive of an isolated As antisite, while the third As shell (ﬁfth neighbour shell) is clearly of lower symmetry than expected for an isolated As antisite. It was proposed that EL21 is an isolated antisite at room temperature, but at low temperature, where all magnetic resonance experiments are performed, it associates itself with shallow acceptors such as Zn-Ga more than two nearest neighbour distances away. W-band HFEPR was also measured using the photo-luminescence for detection, to investigate P dopants in 6H-SiC. Produced under reducing conditions, b-Ga2O3 becomes an n-type semiconductor. The conduction electrons produce a very strong EPR signal and a strong HFI with the Ga nuclei. The advantages of working at high frequency were clearly shown25 in a W-band study of a single crystal of b-Ga2O3. The Overhauser-shift technique at W-band resolved all the spectral lines and allowed them to be assigned to their corresponding electronic and nuclear states. The probability densities of the electrons at the nuclei in the two nonequivalent crystallographic positions - the lattice sites with octahedral and tetrahedral coordination - could also be determined directly. The enhanced resolution revealed an otherwise hidden substructure in the nuclear resonance signals. This structure could be used to probe the environment of the oxygen vacancy and to determine the extension of the electronic wave function of the donor electrons. Quantum conﬁnement in the smaller nanoparticles of CdSe in polybutadiene-based polymer matrices leads to an increase in the band-gap energy with associated band-edge luminescence emission. A new broad luminescence feature has been detected26 for particles smaller than 3 nm and an associated new paramagnetic defect centre observed by W-band HFEPR. The new defect was found to be located at a trigonally-distorted tetrahedral site, with spin S ¼ 1/2, gJ ¼ 2.321(6) and g> ¼ 2.053(6). A poorly resolved HFI with a nucleus of spin I ¼ 3/2 suggested that the defect is due to copper in its d 9 state. It was proposed that the defect normally resides within the valence band for bulk and large particles and only becomes apparent as a defect within the band-gap as the energy increases above 2.6 eV in the smaller particles. HFEPR measurements at 95 GHz at temperatures down to 5 K have been made on three cubic boron nitride single crystals doped with beryllium.27 The samples exhibited different spectra at low temperature that were sensitive to low temperature in situ illumination using a series of Kr and Ar laser lines. The analysis of the spectra resulted in the identiﬁcation of several native paramagnetic centres. Silver halides are of both fundamental and technological interest due to their position intermediate between covalent semiconductors and ionic insulators and their importance as photosensitive materials. The effect of reduced dimensionality on the properties of AgCl crystals is of particular interest since it is an indirect-gap material. The properties of shallow electron centres (SEC), selftrapped holes (STH), and self-trapped excitons (STE) have been investigated28 in nanocrystals of AgCl embedded in a crystalline KCl matrix. Time-resolved

Electron Paramag. Reson., 2007, 20, 245–270

249

photoluminescence (PL) measurements revealed behaviour different from that of bulk AgCl. HFEPR optically detected magnetic resonance (ODMR) measurements at 95 GHz revealed that the transitions of SEC and STH centres were only observed in the extreme low-energy tail of the PL emission. They exhibited higher g-values than the respective bulk values with a broad and asymmetric g-value distribution of width some 0.07. A complex behaviour of the ODMR spectrum in the STE region was revealed. Analysis of the measurements at 4.6 K yielded a ZFS in the STE triplet state in good agreement with the bulk value, while the g-values were slightly increased relative to bulk, which was ascribed to the nanocrystal environment. HFEPR and ENDOR experiments at 95 GHz on ZnO nanoparticles revealed29 the presence of shallow donors related to interstitial Li and Na atoms. The effect of conﬁnement on the shape of the electronic wave function could be determined for the ﬁrst time. In addition, it was observed that the 67Zn nuclear spins become polarized upon saturation of the EPR transition. This Overhauser-effect is induced in the nanoparticles by the zero-point vibrations of the phonon system. 2.3 Catalysis. – Cu zeolites are important as catalysts for the catalytic decomposition of the nitrogen oxides produced by engines and industrial boilers. The Cu(I)-NO adsorption complexes have been formed30 over copper exchanged and autoreduced high siliceous Cu-ZSM-5 and Cu-MCM-22 zeolites and studied by EPR at X-, Q-, and W-band. ZSM-5 has a three-dimensional structure with a large variety of potential sites for the Cu(I) cations but quantum chemical calculations suggest that the most probable locations for the cations are the M7 and I2 sites. MCM-22 is a medium-pore high silica zeolite. There are known to be two different Cu(II) sites in this zeolite but the exact location of the Cu(I) formed after reduction of Cu(II) is not known. The increased spectral resolution and separation of the Cu(I)-NO signals from the Cu(II) signal obtained from HFEPR allowed a reliable determination of the spin Hamiltonian parameters from the powder spectra. Two signals were identiﬁed for the Cu(I)-NO complexes in both zeolites. All four complexes had nearly identical g-values and hyperﬁne parameters so the zeolite matrix has little inﬂuence on the spin Hamiltonian parameters. The measured parameters for ZSM-5 agreed well with the calculated values for the M7 and I2 sites and it was also concluded that similar sites must exist in Cu-MCM-22. Another paper from the same group31 provided a critical assessment of EPR studies of Cu(I)-NO complexes in Cu-ZSM-5 zeolites. Two different Cu(I)-NO species, A and B, were identiﬁed with the ratio of A to B a function of the method of preparation of the complexes. EPR from X- to W-band revealed the rigid structure of the adsorption complexes at 6 K but at 77 K a complex dynamic structure had formed. Pulsed EPR showed that in fact a motional process of the adsorbed NO complexes began at temperatures above 10 K. The results of this paper support the conclusion drawn above that the NO complexes are formed at two different Cu(I) cationic sites in the ZSM-5 framework. It was stressed that multi-frequency EPR is essential for reliable determination

250

Electron Paramag. Reson., 2007, 20, 245–270

of the spin Hamiltonian parameters of the adsorption complexes. A review of the work of this group on zeolites has been given recently.32 The structural features of Mn(II) incorporated into two large cage zeotypes, Mn-UCSB-10Mg and Mn-UCSB-6Mg, have been explored33 by combining multifrequency CW-EPR with W-band ENDOR spectroscopy. HFEPR experiments at 95 and 180 GHz resolved two main types of Mn(II) framework sites with significantly different 55Mn hyperﬁne couplings and slightly different g-values. These results were supported by 55Mn ENDOR spectra, which also revealed the presence of a third species. The sign of Aiso(55Mn) was shown to be negative. The various species were assigned to framework sites with different degrees of water coordination. While one species is similar to the distorted (pseudo) tetrahedral sites found in the reference samples, the other two have an interaction with weakly bound water ligands. Hydration and dehydration result in a reversible transformation between the three Mn(II) types. It was noted that single crystal spectra did not show improved resolution over powder spectra. A broadline 31P NMR study34 of the incorporation of moderate amounts of Ni(II), Co(II), Fe(II/III), and Mn(II/III) into aluminophosphate zeotype AlPO4-34 and Fe(II/III) into aluminophosphate zeotype AlPO4-36 was shown to be complementary to HFEPR. NMR provided direct evidence of isomorphous substitution of framework aluminum by transition metals and allowed the determination of the extent of the substitution. Unlike ENDOR studies, where the optimum Me/Al ratio is below 1%, 31P NMR worked best with a moderate or large fraction of transition ions incorporated in the lattice. The macrocyclic ligand py2(NMe)2, where py2(NMe)2 is N,N 0 -dimethyl-2,11diaza[3,3](2,6)pyridinophane, is of interest because an Fe/py2(NMe)2 complex has been shown to catalyze the degradative oxidation of catechols by O2 in a dioxygenase-like reaction. Two manganese complexes, (py2(NMe)2)MnIICl2 (1) and [(py2(NMe)2)MnIIIF2]1 (2), were investigated35 by EPR. The powder spectrum of complex (1).H2O could be observed at 9 GHz and showed a large isotropic resonance with g ¼ 2. Complex (2)(PF6) required HFEPR because MnIII is a non-Kramers ion. The powder spectrum at 285 GHz was observed for temperatures at and below 15 K. Resonances were observed in the ﬁeld ranges 0.5–1.0 T, 2.0–5.0 T and 7.5–8.5 T. A preliminary ﬁt to the data showed that D is about 4 cm1 and the temperature dependence of the intensity of the spectra showed that the | 2> state is the lowest energy manifold. This conclusion was supported by DFT calculations that revealed a low-lying spin triplet state and the elongated Jahn-Teller distortion shown by X-ray diﬀraction experiments. The best technique for the preparation of blue aluminosilicate sodalite pigments is still not well established. Three blue pigments36 were prepared by a new method and characterized by CW and pulsed X-band EPR and HFEPR at 94 GHz. Pulsed ENDOR at 94 GHz was also used. Calcination generated the S3 and S2 radicals, and the effects of the Al/Si ratio and the calcinations temperature on the nature and amounts of the radicals were examined. The HFEPR measurements clearly resolved the g-anisotropy of S3 and ENDOR measurements detected strong coupling with extra-framework 23Na cations and

Electron Paramag. Reson., 2007, 20, 245–270

251

weak coupling with framework 27Al. The EPR signals were attributed to three different S3 radicals located within the sodalite cages. Two of these radicals were well isolated but the third one was associated with an exchange-narrowed signal originating from S3 radicals in nearby sodalite cages.

2.4 Ferroelectrics. – Lead titanate, PbTiO3, is a ferrroelectric material widely used as a dielectric in capacitors and in electro-optical components. Transition-metal ions have been added to PbTiO3 to improve its performance in devices and HFEPR has proved useful in characterizing the effect of the defect structures around the impurity. The intrinsic Fe31 impurity centre in polycrystalline PbTiO3 has been investigated37 by HFEPR at frequencies up to 190 GHz. The spectrum was attributed to Fe31 ions substituted for Ti41 at the B site of the perovskite lattice. The value of D ¼ þ35.28 GHz, with a variance of about 1 GHz, could be understood as arising from a directly coordinated Fe – oxygen vacancy defect associate. The distribution of D values was attributed to interactions with more distant defects and vacancies. The large value of D means that even at 93 GHz the resonance only satisﬁes the ‘‘medium ﬁeld’’ condition but the spectrum could be satisfactorily analysed at 190 GHz. An X-, Q-, and W-band study of Cr-doped PbTiO3 micro- and nanopowder samples revealed38 three Cr31 centres in tetragonal phase samples with different axial ZFS parameters, C1, C2, and C3. The C1 centre was similar to that observed in previous X-band crystal and ceramic samples. The ZFS data was used to deduce the local displacements inside the distorted oxygen octahedra of the PbTiO3 lattice. It was also found that there is a size-driven tetragonal-tocubic phase transition. Below a critical particle size only the cubic phase formed at all temperatures and a new Cr31 centre spectrum, C4, consisting of a single line with an isotropic g-factor was observed. Copper(II)- and iron(III)-modiﬁed Pb[ZrxTi1x]O3 ferroelectrics have also been investigated39 by HFEPR. It was found that Cu21 and Fe31 both substitute as acceptor centres for [ZrTi]41. In the iron-doped system the charge-compensating oxygen vacancies lead to the formation of charged defect associates, but no such associates were observed for the copper-modiﬁed system. The model of a mesoscopic mixing of the pure-member phases was reﬁned to a picture in which there is a nanoscale distribution of composition. Multiferroic materials have two order parameters, e.g. spontaneous polarization (ferroelectric or antiferroelectric) and spontaneous magnetization (FM or AFM). One of the ﬁrst multiferroic materials to be found was BiFeO3, bismuth ferrite, which has a perovskite-like structure. Ferroelectric ordering occurs below 1103 K and spin AFM ordering below 643 K. HFEPR at 4.2 K in the frequency range 115-360 GHz and in ﬁelds of up to 25 T was used40,41 to study the magnetic order of bismuth ferrite single crystals. Significant changes were observed in the spectra as the ﬁeld was increased and it was concluded that there was an induced phase transition from an incommensurately cycloidal modulated state to one with homogeneous spin order.

252

Electron Paramag. Reson., 2007, 20, 245–270

2.5 Transition Metal and Rare Earth Ions. – In this section we collect together HFEPR experiments on materials that do not ﬁt into other sections. An extensive review of the theory and methods of measuring ZFS in metal complexes has been given by Boca.42

2.5.1 Fe and Co Ions. The advantages of combining swept-ﬁeld HFEPR in ﬁelds of up to 25 T with swept frequency measurements (FDMRS) in zero or ﬁxed ﬁelds for the investigation of high-spin systems has been illustrated9 by measurements on [Fe(H2O)6](ClO4)2 (3) and (NH4)2[Fe(H2O)6](SO4)2 (4). Both compounds contain hexaaqua Fe(II). The resulting spectra were analyzed with S ¼ 2 to yield highly accurate spin Hamiltonian parameters. The complexes were also studied by other magnetic techniques to corroborate the resonance results. For (3), all the techniques were in excellent agreement and gave as consensus values: D ¼ 11.2(2) cm1, E ¼ 0.70(1) cm1. For (4) FDMRS and HFEPR gave D ¼ 14.94(2) cm1, E ¼ 3.778(2) cm1. It was concluded that the spin Hamiltonian parameters for the perchlorate best represent those for the isolated hexaaqua Fe(II) complex. The electronic parameters for the [Fe(H2O)6]21 ion will be important for future studies of biologically relevant systems containing high-spin Fe21. The Co(II) ion is of great interest, particularly for its ability to replace zinc in many enzymes, but has not been widely studied since its large ZFS usually makes HFEPR essential. The Co(II) ion in its high spin (S ¼ 3/2) state in the complex Co(PPh3)2Cl2, where Ph ¼ phenyl, was investigated7 by variable-ﬁeld magnetic circular dichroism (VTVH-MCD) and by HFEPR over the frequency range 150–700 GHz. The loose powder spectra were complicated, due to the reorientation of the particles in the magnetic ﬁeld, and were supplemented by measurements on a pellet sample. The spin Hamiltonian parameters found from the whole HFEPR data set were in good agreement with the values found by VTVH-MCD. This proved that the two techniques are equally valid routes to the determination of the ZFS parameters and settled a long-standing controversy. HFEPR is more accurate but VTVH-MCD is more sensitive. The HFEPR values were: D ¼ 14.76(2) cm1, E ¼ 1.141(8) cm1, gx ¼ 2.166(4), gy ¼ 2.170(4), gz ¼ 2.240(5). 2.5.2 Cr. The strong inﬂuence of dynamical effects on the electronic and molecular structure of octahedrally coordinated high-spin d 4 complexes was demonstrated15 by experiments on Cr Tutton’s salt. The Cr(II) Tutton’s salts, (MI)2Cr(X2O)6(SO4)2, where MI ¼ ND41, Rb1, or Cs1, and X ¼ H or D, were studied by inelastic neutron scattering (INS) and HFEPR over the frequency range 95–285 GHz. HFEPR spectra were taken using well-ground samples immobilized in wax. A new theoretical approach for the calculation of the variable-temperature EPR spectra of high-spin d 4 complexes allowed a rigorous definition of the ground-state electronic structure from 1.5 up to 296 K, which is unprecedented for a high-spin d 4 complex. Modelling of the INS data using a conventional S ¼ 2 spin Hamiltonian revealed a dramatic variation in

Electron Paramag. Reson., 2007, 20, 245–270

253

the axial and rhombic ZFS as a function of temperature. For the ammonium salt, D and E are 2.454(3) and 0.087(3) cm1 at 10 K and 2.29(2) and 0.16(3) cm1 at 250 K, respectively. A temperature variation in the stereochemistry of the [Cr(D2O)6]21 complex was also identiﬁed, with an apparent coalescence of the long and medium Cr-O bond lengths at temperatures above 150 K. Smaller changes were found for the rubidium and cesium salts. The experimental quantities were interpreted using a Jahn-Teller Hamiltonian perturbed by a temperature-dependent anisotropic strain. The differences between the spin Hamiltonian parameters determined by INS and HFEPR were shown to be due to the time scale of INS being much shorter (1012 s) than that of HFEPR. 2.5.3 Mn Ions. High–spin mononuclear Mn(II) complexes are of importance in catalysis and biochemistry.The isolation, structural characterization, and electronic properties of a series of high-spin mononuclear ﬁve-coordinated Mn(II) complexes, [Mn(terpy)(X)2] (terpy¼2, 2 0 :6 0 ,200 -terpyridine; X ¼ I (5), Br (6), Cl (7), or SCN (8)) has been reported.43 The X-ray analysis of the complexes showed that the manganese ion lies in the centre of a distorted trigonal bipyramid for complexes 5, 6, and 8, while complex 7 is better described as a distorted square pyramid. They were studied by HFEPR over the frequency range 95-285 GHz and at temperatures from 5 to 30 K. All the powder spectra were found to be in the high-ﬁeld limit. The magnitude of D varied between 0.26 and 1.00 cm1 with the nature of the anionic ligand. D was found to be positive for the iodo and bromo complexes and negative for the chloro and thiocyano complexes. The sign of D was correlated with the structure of the complexes. There was a significant rhombic term for all the complexes, with E/D values between 0.17 and 0.29. The spin Hamiltonian parameters of these ﬁve-coordinated complexes were compared with those of previously reported dihalo four- and six-coordinated complexes. The isolation, structural characterization and electronic properties of another new mononuclear, pentacoordinate MnII complex [Mn(tButerpy)(N3)2] (9; tButerpy ¼ 4,4 0 ,400 -tri-tert-butyl- 2,2 0 :6 0 ,200 -terpyridine) was reported by the same group.44 The X-ray structure of 9 showed that the manganese ion lies in the centre of a distorted trigonal bipyramid. HFEPR spectra were recorded between 190 and 285 GHz on a neat powder of 9 and also on 9 magnetically diluted in the corresponding zinc complex. The spin-Hamiltonian values deduced from the neat powder HFEPR spectra at 95 and 250 GHz were D ¼ 0.250(5) cm1, E ¼ 0.044(5) cm1 and gx ¼ gy ¼ gz ¼ 2.000(5). The magnetically diluted sample allowed the determination of the hyperﬁne coupling constant to be 77.5(5) G and conﬁrmed the values of the parameters determined with the neat powder. A survey of published values of D showed that there is good agreement between the magnitude of D and the coordination number of the MnII ion: |D| is larger than 0.2 cm1 for pentacoordinate, but smaller than 0.1 cm1 for hexacoordinate, complexes. A third paper45 discussed the six-coordinated chloromanganese(III) complexes [Mn(terpy)Cl3] 10 and [Mn(Phterpy)Cl3] 11, (Phterpy ¼ 4 0 -phenyl-2,20 :6 0 ,200 terpyridine). Both complexes exhibited a Jahn-Teller distortion of the octahedron,

254

Electron Paramag. Reson., 2007, 20, 245–270

characteristic of a high-spin Mn(III) complex (S ¼ 2). The manganese ion lies in the centre of a distorted octahedron characterized by an elongation along the tetragonal axis. HFEPR in the frequency range 190-475 GHz at temperatures between 5 and 15 K showed that the sign of D was correlated with the nature of the tetragonal distortion, but the magnitude of D was not sensitive to the nature of the anions in this series of rhombic complexes. The |E/D| values (0.124 for 10 and 0.085 for 11) were smaller than those found earlier for the [Mn(L)(X)3] complexes (in the range of 0.146 to 0.234). The E term was also found to increase as the ligand ﬁeld strength of the equatorial ligands decreased. The mononuclear high-spin Mn(III) complex [Mn(dbm)2(py)2](ClO4) (dbm ¼ anion of 1,3-diphenyl-1,3-propanedione (dibenzoylmethane), py ¼ pyridine) was found46 by X-ray crystallography to have a tetragonally distorted geometry with the axial positions occupied by the py ligands and the equatorial positions by the dbm ligands. HFEPR spectra were recorded for both a solid powder and in a frozen dichloromethane solution. High quality HFEPR spectra were recorded at frequencies between 95 and 440 GHz in ﬁelds up to 15 T. The complete data set of resonant magnetic ﬁelds versus transition energies was analyzed using an automated ﬁtting procedure to give the spin-Hamiltonian parameters, D* ¼ 4.504(2) cm1, E* ¼ 0.425(1) cm1, gx ¼ 1.993(1), gy ¼ 1.994(1), and gz ¼ 1.983(1). It was noted that the fourth-order ZFS terms are usually difﬁcult to obtain even from single crystal spectra but were obtained here from powder samples. The complex [MnIII(taa)], where H3(taa) is tris(1-(2-azolyl)-2-azabuten-4-yl) amine, is the ﬁrst example of a material showing a spin-crossover transition that can be driven by a magnetic ﬁeld. (There is a thermally induced spincrossover transition around 46 K from the low spin S ¼ 1 to the high spin S ¼ 2 state.) The total entropy gain associated with the transition is much greater than that expected just from the change of the value of S. It has been proposed that structural disorder in the high spin state due to the dynamic Jahn-Teller effect is important for driving the spin-crossover. This was conﬁrmed by the powder HFEPR spectra12 of [MnIII(taa)] obtained at 77 K over the ﬁeld range of 10 to 50 T at 1017.6 GHz. The spin-Hamiltonian values D ¼ 8.49 K, E ¼ 0.72 K and g ¼ 2.0 provide a good ﬁt to the spectrum. The ﬁnite value of E conﬁrmed the existence of a rhombic anisotropy term in the high-spin state of [MnIII(taa)]. It was proposed that the trigonal symmetry shown by X-ray diffraction is the time-averaged structure while HFEPR detects the Jahn-Teller distortion because the characteristic frequency of the molecular reorientation at 77 K is lower than the HFEPR frequency. The 17O isotope is an important probe in many systems in biology, catalysis and material science but its large quadrupole interaction often makes it difﬁcult to interpret X-band ENDOR spectra. The pulsed ENDOR spectrum at 95 GHz of the 17O hyperﬁne couplings of a frozen solution of Mn(H217O)621 was found47 to be in good agreement with single-crystal measurements at Q-band.This was a preliminary experiment to demonstrate the extra resolution possible at high frequency before investigating the V(H2O)521 complex discussed in the next section.

Electron Paramag. Reson., 2007, 20, 245–270

255

2.5.4 V Ions. The vanadyl ion VO21 is of particular interest because it may be substituted in biological systems for EPR-silent ions such as Mg21, Ca21, or Zn21. Pulsed ENDOR at 95 GHz at a temperature of 7.5 K has been used to measure the 17O hyperﬁne interaction of the water ligands and the V¼O oxygen of the vanadyl V(H2O)521 aquo complex.47 Orientation-selective ENDOR spectra of the vanadyl complex exhibited two distinct signals assigned to the vanadyl oxygen and the water ligands. DFT calculations showed good agreement with the experimental values thus providing clear evidence that the vanadyl oxygen is exchangeable. The interaction of the vanadyl oxygen, especially its anisotropic part, is significantly larger than that of the water oxygens due to a relatively large negative spin density on the oxygen p orbitals. Vanadium(III) complexes are of great theoretical interest because they are the simplest systems in which to study the combined effects of a ligand ﬁeld and inter-electronic repulsion. There is however a lack of experimental data to compare with calculations such as the angular overlap model (AOM) because X-band EPR is usually impossible due to the large ZFS of the ground state of V(III). In two very detailed linked papers48,49 the HFEPR of a number of salts containing the [(V(OH2)6]31 cation was compared with AOM calculations. HFEPR experiments were carried out over the frequency range 95–460 GHz at temperatures between 5 and 20 K on powder and single crystal samples. The spectra of V(III) as an impurity in guanidinium gallium sulfate (the guanidinium ion is [C(NH2)3]1) was found to be particularly instructive,48 with ﬁnestructure observed attributable to crystallographically distinct [V(OH2)6]31 cations, hyperﬁne coupling, and ferroelectric domains. The electronic structure of the complex was shown by single-crystal neutron- and X-ray diffraction measurements to depend strongly on the mode of coordination of the water molecules to the V(III) cation, and was also sensitive to the isotopic abundance. The AOM gave a good account of the change in the electronic structure as a function of the geometric coordinates of the [V(OH2)6]31 cation. However, the ligand-ﬁeld analysis was inconsistent with the proﬁles of electronic transitions between ligand ﬁeld terms. In the second paper49 the effect of dynamic Jahn-Teller coupling on the energies and magnetic properties of the low-lying vibronic states of the sixcoordinate trigonally distorted complex was calculated. The ZFS of the ground state of the [V(OH2)6]31 cation was shown to be significantly reduced when the Jahn-Teller and spin-orbit coupling is of comparable magnitude. The properties of the V(III) ion in two types of pseudooctahedral complexes: V(acac)3, where acac ¼ anion of 2,4-pentanedione, and VX3(thf)3, where thf ¼ tetrahydrofuran and X ¼ Cl and Br, have been investigated6 using a wide range of techniques. HFEPR was carried out over the frequency range 95 to 700 GHz in steady ﬁelds of up to 25 T. The spin Hamiltonian parameters for V(acac)3 were found to be D ¼ þ7.470(1) cm1, E ¼ þ1.916(1) cm1, gx ¼ 1.833(4), gy ¼ 1.72(2), gz ¼ 2.03(2). For VCl3(thf)3, HFEPR detected a single zero-ﬁeld transition at 15.8 cm1 (474 GHz), which was insufﬁcient to determine the complete set of spin Hamiltonian parameters. For VBr3(thf)3, however, a

256

Electron Paramag. Reson., 2007, 20, 245–270

particularly rich data set was obtained: D ¼ 16.162(6) cm1, E ¼ 3.694(4) cm1, gx ¼ 1.86(1), gy ¼ 1.90(1), gz ¼ 1.710(4). (This may be the largest magnitude of D ever obtained by EPR.) The inconsistencies in previous discussions of the bonding of V(acac)3 were removed by combining the HFEPR data with that from MCD experiments. The fact that the VX3(thf)3 complexes have a large negative value of D may make them useful building blocks for SMM. 2.5.5 Rare Earth Ions. Three types of spectra were observed50 for KPb2Cl5:Tb31 by HFEPR in the frequency range 74-200 GHz. The most intensive spectrum with resolved hyperﬁne structure corresponded to transitions between sublevels of the 159Tb31 ground quasi-doublet with ZFS close to 48 GHz. The terbium ions are in low-symmetry Pb21 positions with chlorine seven-fold coordination and a charge compensating vacancy in the nearest potassium site. The calculated values of g-factors and ZFS are in agreement with the experimental data. The nature of a broad EPR line with ZFS of about 180 GHz and of additional weak EPR lines observed as satellites of the main Tb31 lines was also discussed. The D-band pulsed HFEPR spectra of several Gd31 complexes in glassy water-methanol solutions were analyzed51 by a specially developed stochastic superposition model that predicted the essential features of the distribution of the crystal ﬁeld interaction (CFI) parameters in glassy systems. The HFI parameters for water ligand protons could be determined accurately since the D-band Mims 1H electron-nuclear double resonance spectra were free from CFI-related distortions. It was also shown that the HFI distribution is solely related to the distribution of the Gd-H distances. 2.6 Molecular Magnetic Clusters. – 2.6.1 Introduction. An extensive review of the theory and various methods of measuring ZFS in metal complexes has been given by Boca.42 The analysis of the properties of a cluster is often aided if spectra are also recorded for a diluted sample where the magnetic ions do not interact. The parameters are then shown as Di and Ei. A particular interest of magnetic clusters (MMC) is the possibility of forming a single molecule magnet (SMM). The potential energy barrier for reversal of the direction of magnetization is given by U ¼ |D|S2. (Note that a useful SMM must have uniaxial anisotropy so the value of D must be negative.) It is found that the ac susceptibility of an SMM follows an Arrhenius law, but that the activation energy, Ueﬀ, is usually lower than U, because the effective barrier height is lowered by quantum-mechanical tunnelling (QMT). SMM behaviour will be observed when the temperature is well below Ueﬀ. Quantum computing may also be possible using the QMT of SMM. 2.6.2 Mn Clusters. Although there have been extensive HFEPR measurements on [Mn12O12(CH3COO)16(H2O)4].2CH3COOH.4H2O, (Mn12-Ac), there remains significant disagreement between the details of the analysis based on data from different spectrometers. The Grenoble work has been reviewed

Electron Paramag. Reson., 2007, 20, 245–270

257

recently.52 It has been claimed for example that true line-shapes are not obtainable with simple transmission spectrometers and that a multi-parameter ﬁt to HFEPR spectra is less reliable than performing experiments over a wide frequency range and extrapolating to zero ﬁeld.53 It also remains controversial whether or not reliable values of the spin parameters can be obtained from polycrystalline pellet samples or if single crystals are essential. A further complication is that some weak sample dependent resonances may be of intrinsic origin or due to the loss of solvent material with time. The magnetic core of Mn12-Ac consists of an external ring of eight MnIII ions, each with S ¼ 2, and an internal tetrahedron of four MnIV ions with S ¼ 3/2. The cluster has an S ¼ 10 ground state with a negative value of the ground state splitting parameter D. The barrier to magnetization reversal is about 60 K. The ground state S multiplet is split to leave the Ms ¼ 10 sublevels with lowest energy. Below 2 K there is evidence for QMT from measurements of the superparamagnetic relaxation time and a staircase proﬁle in the hysteresis loop. Disorder appears to be inherent in Mn12-Ac and generates non-exponential MQT relaxation and tunnelling that would not be allowed in a perfect crystal. A review of QMT in Mn12-Ac has appeared recently.54 A review53 of swept frequency measurements of SMM over the range 170600 GHz has been illustrated by measurements on Mn12-Ac. Analysis of the spectra collected over a wide frequency range above 300 GHz gave the values; D* ¼ 0.454 cm1, and a locally varying value of E* with a maximum value of 0.01 cm1. It was pointed out that several variations of the basic Mn12-Ac structure are known to exist. One that exhibited faster relaxation than usual at low temperature showed clearer resonances at higher ﬁelds than have been observed before. The faster relaxation is believed to be due to a tilting of the Jahn-Teller elongation axes of one or more of the Mn(III) ions. A combination of magnetic measurements at 0.6 K, which is in the pure quantum regime, and HFEPR over the range 40-200 GHz has been used to investigate the MQT of Mn12-Ac single crystals in great detail.54 It was concluded that MQT is due to a disorder-induced locally-varying quadratic transverse anisotropy associated with rhombic distortions in the molecular environment. This is superimposed on a fourth-order transverse-magnetic anisotropy consistent with the global (average) molecule site symmetry. The angular dependence and ﬁne structure of the EPR absorption peaks and the detailed behaviour of the MQT was shown to be due to the resulting interplay between local and global symmetries. HFEPR conﬁrmed that there are discrete tilts of the local molecular magnetic easy-axis away from the average easy-axis of the crystal. The HFEPR spectrum of [Mn12O12(O2CCH2Br)16(H2O)4 4CH2Cl2], (Mn12-BrAc), has recently been shown55 to be significantly simpler than that of Mn12-Ac and may therefore be a more suitable candidate for measurements of quantum effects in high symmetry SMM with S ¼ 10. The Mn12-BrAc molecule consists of four Mn41 ions, each with spin S ¼ 3/2, surrounded by eight Mn31 ions with spin S ¼ 2. The orbital moment is quenched and a JahnTeller distortion associated with the Mn31 ions is largely responsible for the

258

Electron Paramag. Reson., 2007, 20, 245–270

magnetic anisotropy. HFEPR was carried out on sub-mm sized single crystals of Mn12-BrAc. The spectra observed with the magnetic ﬁeld in the hard (x-y) plane were similar to those of Mn12-Ac. It was noted that even a few degrees misalignment of the crystal was sufﬁcient to produce extra lines in the spectra but that these could be understood without assuming the solvent-induced disorder found in the Mn12-Ac lattice. Measurements at temperatures above 15 K revealed a new line in the spectra that was attributed to the ﬁrst excited state with S ¼ 9 some 40 K above the ground state. The S ¼ 9 state was found to have D ¼ 0.430 cm1 (0.62 K), some 5% smaller than for S ¼ 10, and a barrier height of 50 K, 23% smaller than for S ¼ 10. The pure material [Mn11CrO12(O2CCH3)16(H2O)4] 2CH3COOH 4H2O, (Mn11Cr), could not be prepared but was obtained as a mixed crystal with Mn12-Ac.56 The Cr31 ion was found from X-ray analysis to occupy a speciﬁc Mn31 site in the Mn12-Ac lattice. The coercive ﬁeld of Mn11Cr was found to be 0.95 T, nearly half that of Mn12-Ac. HFEPR at 381.5 GHz in pulsed ﬁelds of up to 30 T showed that Mn11Cr has a ground state spin S ¼ 19/2. The value of D was almost the same as for Mn12-Ac but the barrier height was reduced to 56.8 K because of the lower value of S compared to Mn12-Ac. Measurements of the QTM resonance ﬁeld of Mn11Cr suggested the importance of dipolar-biased tunnelling in this material. Single crystals of the new monoclinic material [MnIIICuIICl(5-Br-sap)2 (MeOH)] (12) were studied57 by magnetization measurements and HFEPR at frequencies up to 342 GHz. The ground state was found to be S ¼ 5/2 with an exchange-coupling constant of þ78 cm1. The HFEPR spectrum at 4.2 K consisted of four peaks and two further peaks appeared at higher ﬁelds as the temperature was increased. Values of D ¼ 1.86 cm1 and 1.81 cm1 were obtained from magnetization and HFEPR measurements respectively. The gap value is 10.5 K. It was noted that it should be possible to obtain compounds with higher spin ground states than (12) by similar preparation techniques. A theoretical analysis58 of the multi-frequency HFEPR spectra of the supermolecular dimer [Mn4]2 of the SMM [Mn4O3Cl4(O2CEt)3(py)3]2.2C6H14 (where EtCO2 is propionate, py is pyridine, and C6H14 is hexane)59 was made using a perturbation method in which the high-order corrections to the level splittings of degenerate states are included. It was shown that the corresponding eigenvectors are composed of entangled states of two molecules. The anisotropy constant and exchange coupling were obtained from the best ﬁt of the theoretical level splittings with the measured values. The prediction59 that the two Mn4 units within the dimer are coupled quantum-mechanically by the AFM exchange interaction was conﬁrmed. The supermolecular dimer behaves in a similar fashion to artiﬁcially fabricated quantum dots. 2.6.3 Ni Clusters. The compound (NMe4)[GdNi6(pro)12](ClO4)4 (13) was prepared60 to investigate the coupling between rare earth and Ni ions, that has been less extensively studied than other combinations in molecular clusters. The properties of (13) were compared with the compound where GdIII is replaced by diamagnetic LaIII (14). Both molecules consisted of a LnIII centred

Electron Paramag. Reson., 2007, 20, 245–270

259

cage of six symmetry-equivalent NiII ions, that form a perfect octahedron. Susceptibility measurements on (13) were analysed to give a ﬁve-fold degenerate spin-frustrated ground state with a total spin of 13/2 and a three-fold degenerate ﬁrst excited state with spin 15/2 only 0.01 cm1 higher in energy. The Ni–Ni exchange interaction was AFM but the Gd–Ni interaction was found to be weakly FM. X-band and HFEPR over the range 150–600 GHz showed a single broad resonance for complex (13). The g-value decreased from 2.20(2) at 186 GHz to 2.11(2) at 368 GHz showing the presence of a small ZFS. Complex (14) has integer spin and large ZFS so no X-band resonance was observed but HFEPR at low temperature revealed two broad lines. The best ﬁt to the spectra of complex (14) was obtained by assuming weak coupling between the six NiII ions and a large single-ion ZFS. The spin-Hamiltonian values were: S ¼ 1, D ¼ 5.5(1) cm1, E ¼ 0.8(3) cm1, gx ¼ 2.06(3), gy ¼ 2.21(4), and gz ¼ 2.30(3). The NiII4 complexes have a ground state with spin S¼4 and a negative value of D but their potential usefulness as SMM is limited by rapid QMT. Single crystals of [Zn3.91Ni0.09(hmp)4(dmb)4Cl4] (15), where dmb is 3,3-dimethyl-1butanol and hmp is the monoanion of 2-hydroxymethylpyridine, were prepared61 in order to investigate the origin of the rapid tunnelling observed in [Ni4(hmp)4(dmb)4Cl4] (16). HFEPR over the frequency range 40-350 GHz was performed on a single crystal of (15). The use of rather a large crystal, and the consequent non-uniformity of the mm-wave ﬁeld, complicated the observed spectra but single-ion parameters were determined for each NiII site. Angledependent measurements gave optimum single-ion values of Di ¼ 5.30(5) cm1, Ei ¼ 1.20(2) cm1, gz ¼ 2.30(5), and gx ¼ gy ¼ 2.20(5). The NiII ions easy axes were found to be tilted 151 away from the c-axis that is the easy axis for the Ni4 SMM. Using these values the parameters for the S ¼ 4 ground state of (16) were calculated to be: E* ¼ 0 and D* ¼ 0.69 cm1, taking the positive value of Ei or D* ¼ 0.66 cm1 for the negative value, in good agreement with the experimental value D ¼ 0.600 cm1. The observed rapid QMT in (16) was attributed to a large fourth order term in the ZFS due to the tilt of the NiII ions from the c-axis and the significant value of Ei. A second pair of NiII4 complexes, [NixZn1x(hmp)(t-BuEtOH)Cl]4, where t-BuEtOH is 3,3-dimethyl-1-butanol, were studied by the same group62 for x ¼ 1 and 0.02. The x ¼ 1 complex is a SMM, again exhibiting fast magnetization tunnelling. The single-ion NiII parameters were determined from measurements on the x ¼ 0.02 compound to be Di ¼ 5.27 cm1 and Ei ¼ 1.2 cm1 and shown to give good agreement with the values for the x ¼ 1 compound. The explanation given for the fast tunnelling was the same as that provided in the previous paragraph. The mononuclear NiII octahedral complex [Ni-(HIM2-py)2NO3]NO3 with the bidentate ligand HIM2-py was prepared63 to investigate the single-ion anisotropy and possible usefulness of this complex as a building block for SMM. Magnetization measurements lead to the parameters D ¼ 11.2 cm1, E ¼ 0 and giso ¼ 2.16. HFEPR at 475 GHz and frequency-domain magnetic resonance gave the more accurate results, D ¼ 10.1(1) cm1, E/|D| ¼ 0.02(1).

260

Electron Paramag. Reson., 2007, 20, 245–270

Such a large Ising-type anisotropy has not previously been reported for an isolated NiII complex and suggests that this complex may be useful as the basis of new SMM. 2.6.4 Copper Clusters. The new complex [Cu4(NH3)4(HL)4][CdBr4]Br2 3dmf H2O was shown64 to have a S ¼ 2 ground state by magnetization measurements and HFER over the range 95-380 GHz. The HFEPR spectra were simulated with gx ¼ 2.138(1), gy ¼ 2.142(1), gz ¼ 2.067(1), D* ¼ 0.3529(3) cm1, and E* ¼ 0.0469(8) cm1. The intermolecular interactions in this material were shown to be negligible from measurements over the temperature range 1.8–300 K. The X-band EPR powder spectra were not understandable but HFEPR conﬁrmed the S ¼ 2 ground state. The excited spin triplet state was not observed up to 300 K, possibly due to the fast relaxation that has been proposed to explain the spectra of other copper complexes. The pentameric copper core [Cu5(OH)4(H2O)2]61 of the dimeric, pentacopper(II)-substituted, tungstosilicate polyanion [Cu5(OH)4(H2O)2(A-aSiW9O33)2]10 exhibits strong AFM interactions. The ground state has spin 1/2. HFEPR experiments were carried out65 on polycrystalline powder samples over the frequency range 34–190 GHz. The spectra were only observable below 60 K. The parameters, gzz ¼ 2.4073, gyy ¼ 2.0672, gxx ¼ 2.0240 and Azz ¼ 340 MHz that provided a ﬁt to the spectra are typical values for a Cu21 ion in an axially–elongated octahedral coordination with oxygen. The unpaired electron spin density is therefore localized on the spin-frustrated apical Cu21 ion. The trinuclear complex [Cu3(mal)3 (H2O)]3 of Na3[Cu3(mal) (H2O)].8H2O is trapped in a three-dimensional network with sodium cations. The three copper atoms are connected by alkoxo bridges and form an almost isosceles triangle. Two of the copper ions are also bridged by an extra aquo ligand. Magnetic measurements and HFEPR showed66 that the ground state is an AFM coupled triangular system. The 285 GHz HFEPR spectrum is characteristic of a spin state S ¼ 1/2 with rhombic anisotropy. The rhombic pattern shows that the electronic spin density is delocalised on the three copper ions. 2.6.5 Other Clusters. HFEPR showed67 that Cr10(OMe)20(O2CCMe3)10 is an SMM with one of the smallest ZFS, D ¼ 0.045(4) cm1, yet reported. The ten Cr(III) ions are predicted by DFT to have FM interactions leading to a total spin S ¼ 15 for the ground state of the cluster. However, the HFEPR spectra are consistent with an S ¼ 9 ground state with the higher spin multiplet some 10–20 K higher in energy. The polyoxometalates, K11H[(VO)3(SbW9O33)2] 27H2O (17) and K12[(VO)3 (BiW9O33)2] 29H2O (18) are frustrated AFM that exhibit magnetization jumps with distinct hysteresis for the S ¼ 1/2 to S ¼ 3/2 level crossing under pulsed magnetic ﬁelds. X-band spectra were obtained68 for (17) and (18) as polycrystalline material and in aqueous solution. HFEPR at 110 and 135 GHz was used to observe the spectra around the S ¼ 1/2 to S ¼ 3/2 level-crossing ﬁeld. Two broad lines were observed in the HFEPR at g ¼ 1.96 and 1.94 that may be due

Electron Paramag. Reson., 2007, 20, 245–270

261

to a quasi-axial symmetry. The relative intensity of the two lines as a function of temperature has not yet been explained in detail. A review of the properties of SMM in their magnetically ordered state discussed the theory of canted AFM resonance in some detail.69 When the external ﬁeld along the easy axis of an AFM is increased beyond a critical value, the spin-ﬂop ﬁeld, the sub-lattices rotate into the plane normal to the easy axis. There are therefore two pairs of equations for AFM resonance in this case and HFEPR may be used to ﬁnd the spin-ﬂop ﬁeld. The material p-NCC6F4CNSSN is a canted AFM with an ordering temperature of 36 K. The crystals grow as needles parallel to the c-axis but the easy axis is along b. HFEPR at 94 GHz with the applied ﬁeld perpendicular to the c-axis clearly show the AFM transition at 36 K. Multi-frequency measurements could be used to display the full magnetic phase diagram. 2.7 Low-Dimensional Solids. – 2.7.1 Introduction. Quantum effects are more pronounced in lower dimensions so the magnetic properties of 1D and 2D arrays of localized spins are of great theoretical interest. These materials often exhibit huge ZFS and HFEPR at the very highest frequencies is then essential. The Kobe laboratory has been particularly prominent in this area.10,11 A new theory70 of the EPR line shift has been applied to the 1D XXZ and transverse Ising models in the high ﬁeld limit. A key concept in the theory of 1D magnetism is the difference between the low temperature properties of chains with half-integer and integer spin. The ground state of an S ¼ 1/2 AFM chain is a singlet with no energy gap to the excited states; the elementary excitation is called a spinon. However for S ¼ 1 there is an energy (Haldane) gap of some 0.41 J, where J is the exchange interaction between the spins. A second feature of 1D systems is that they are potentially unstable towards dimerization; in the magnetic context this is called a spin-Peierls transition. 2.7.2 Chains. The interchain interactions of the compound BaCu2 (Si1xGex)2O7 cancel at x ¼ 0.65. (The interactions are FM for x ¼ 0 and AFM for x ¼ 1.) Ohta et al.10,71 carried out HFEPR in the frequency range 40–500 GHz down to 0.5 K to try to observe the breather excitation predicted theoretically by Oshikawa and Afﬂeck72 (OA). Earlier work between 1.8 and 256 K had conﬁrmed that the temperature dependence of the linewidth and of the ﬁeld for resonance for the external ﬁeld parallel to the chain direction (c-axis) agreed well with OA theory in the spinon regime but below 15 K the linewidth at 80 GHz remained constant down to 0.5 K. The ﬁeld for resonance with the ﬁeld parallel to the chain direction was measured at 0.5 and 1.8 K between 40 and 500 GHz and found to have the functional form of the breather mode predicted by OA theory. It was concluded that the breather mode may well exist in the x ¼ 0.65 system, as in the well-known material Cu-benzoate, and that the random interchain interactions in the mixed crystal are averaged out within the coherence length of the material.

262

Electron Paramag. Reson., 2007, 20, 245–270

The compound CsCuCl3 is an S ¼ 1/2 hexagonal AFM with FM chains and AFM coupling between the chains. It becomes an AFM below 10.7 K with the spins in the c-plane forming a 1201 structure on a triangular lattice. Pulsed magnetic ﬁelds up to 16 T have been applied along the c-axis of a single crystal73 to investigate resonance between 120 and 360 GHz at 6 K at atmospheric pressure and 3.1 kbar. There were some differences between the spectra obtained in the pressure cell and earlier work that were attributed to eddycurrent heating of the Be-Cu cell by the pulsed magnetic ﬁeld but the differences found between atmospheric pressure and 3.1 kbar were believed to be intrinsic. The quantum-ﬂuctuation induced phase transition ﬁeld and the AFM gap were both found to increase under pressure. It was concluded that both the interchain interaction and the easy-plane anisotropy increased under pressure in agreement with theoretical predictions. CsFeBr3 has a similar spin structure to CsCuCl3. In the low energy region the magnetic properties may be described by a pseudo-spin S ¼ 1 and a large easyplane anisotropy. When a ﬁeld is applied along the c-axis the gap between the non-magnetic S ¼ 0 ground state and the doublet S ¼ 1 states closes at Hg ¼ 2.89 T. The properties of the high ﬁeld AFM state were investigated11 at 2.0 K by HFEPR in the frequency range 160–496 GHz in pulsed ﬁelds up to 30 T. Two very broad resonance lines with g-values of about 4.8 were observed and there was some evidence of critical phenomena around 17 T. The zero ﬁeld gap was estimated to be 1.3 THz but measurements at higher frequency are needed to conﬁrm this. Below 1.9 K Cu3(OH)2(CO3)2 (azurite) is a model substance of the S ¼ 1/2 distorted diamond chain AFM whose magnetic properties arise from competition between quantum ﬂuctuations and spin frustration. The chain direction is along the b-axis. A magnetization plateau is observed at 1.5 K between 16-26 T when the ﬁeld is along the b-axis and 11-30 T perpendicular to the b-axis. The direct transition from the singlet ground state to the ﬁrst excited state is forbidden but was observed10,74,75 at 1.8 K for the ﬁeld parallel to the chain direction and along the a-axis due to the presence of the anisotropic DM interaction. Frequencies up to 900 GHz and ﬁelds of 32 T were required to obtain a full spectrum. The direct transition was found to be anisotropic and related to the anisotropy of the magnetization plateau. Measurements below the ordering temperature at 0.5 K showed unconventional AFM resonance spectra.74 HFEPR at 1.8 K with the ﬁeld parallel to the c-axis also showed a direct transition.11 The value of the zero ﬁeld gap was found to be 985 GHz which is close to that for the ﬁeld along the a-axis. Magnetization measurements on the frustrated spin-chain system KCu5V3O13 showed AFM ordering below 7 K but with only one-ﬁfth of the Cu spins in the ordered state. The remaining Cu spins form a local singlet state with a spin gap of 44.3 K. HFEPR over the frequency range 40 to 500 GHz at temperatures between 1.7 and 86 K was used76 to investigate this unusual ground state. In the paramagnetic state at 86 K the full anisotropic powder pattern was not observable even at 483 GHz. Speciﬁc heat measurements showed the expected peak at the 7 K AFM ordering temperature and a second

Electron Paramag. Reson., 2007, 20, 245–270

263

sharp peak at 3.9 K. The broad HFEPR spectra were nearly independent of temperature down to 10 K but several sharp peaks began to appear as the temperature was decreased. The ﬁeld for resonance at ﬁxed frequency shifted near 7 K and 4 K. There was also some evidence for another transition at about 2 K. The frequency-ﬁeld diagram at 1.9 K was typical of an easy axis AFM but could not be fully explained within the conventional two sublattice model. A brief account was given in the same paper76 of an experiment on the related compound KMgCu4V3O13 in which one of the Cu21 ions of KCu5V3O13 is substituted by the non-magnetic Mg21 ion. A peculiar ground state is also expected for KMgCu4V3O13 from the magnetization measurements but the HFEPR from the local singlet state could not be observed on a powder sample in this case. Another frustrated 1D S ¼ 1/2 compound is Bi4Cu3V2O14. HFEPR from 105 to 540 GHz in pulsed ﬁelds up to 35 T revealed77 very broad lines at temperatures below 20 K. At 1.9 K AFM resonance modes were observed with g-values of 1.43 and 1.40. The ground state was found to be more complicated than a simple two sublattice AFM model. The quasi-1D Heisenberg AFM Ni(C5H14N2)2N3(PF6), (NDMAP), is an S ¼ 1 compound that forms chains of Ni21 and N3 ions along the c-axis. There is a Haldane gap with a critical ﬁeld (Hc) of 3.4 T along the c-axis and 5.8 T along the a-axis. (A recent theory78 of the magnetization of NDMAP, and other S ¼ 1 systems, in a ﬁeld sufﬁcient to close the Haldane gap has been applied to INS but not yet to HFEPR.) HFEPR at frequencies up to 700 GHz at 1.5 K in ﬁelds of up to 16 T was carried out79 along the a- and c-axes of a single crystal of NDMAP. Broad resonance lines were observed with the ﬁeld along the c-axis with a maximum width near 10 T. Sharp symmetric lines were observed above Hc with the ﬁeld along the a-axis. There is no LRO at 1.5 K for ﬁelds along the c-axis. The frequency-ﬁeld diagram was shown to be in quite good agreement with the Haldane model of a singlet ground-state and triplet excited-state at low ﬁeld and one of the triplet states as the ground state at high ﬁeld. Three branches were observed in the high-ﬁeld region with the ﬁeld along a-axis that corresponded to the transitions from the ground observed by INS. The ground state of an S ¼ 1 Haldane chain can be described by a model of two S ¼ 1/2 states. In the ground state of the perfect chain each S ¼ 1/2 spin forms a singlet with its nearest neighbour, but if the chain is broken two effective S ¼ 1/2 spins are created at the end of the chain. The orthorhombic crystal Y2BaNiO5 is a model Haldane system with the Ni21 chains along the a-axis. Doping breaks the chains to produce a compound such as Y2BaNi0.96Mg0.04O5. The edge-states of this compound were investigated80 by HFEPR at frequencies between 90 and 372 GHz in pulsed magnetic ﬁelds of up to 15 T. The main peak observed in the EPR spectra was attributed to long chains with subsidiary peaks due to chains with length 1, 3, 5 and 7. (It was also shown that it is difﬁcult to detect chains with an even number of spins.) This is the ﬁrst time that a dependence on chain length has been observed. The correlation length in the real system was somewhat larger than that calculated

264

Electron Paramag. Reson., 2007, 20, 245–270

by a simple Heisenberg model. It was emphasized that multi-frequency results are essential for an understanding of these complicated materials. The spin-gap systems SrCu2(PO4)2 and PbCu2(PO4)2 have four kinds of dimers. The ground state is a spin singlet. HFEPR was carried out81 on magnetically-aligned powder samples over the frequency range 80–315 GHz in pulsed magnetic ﬁelds of up to 35T. The observed spectra of SrCu2(PO4)2 showed a single peak in the temperature range from 4.2 to 80 K. The linewidth increased and the intensity of the signal decreased as the temperature was lowered. The g-anisotropy of SrCu2(PO4)2 was found to be very small compared to the usual anisotropic powder spectra of copper compounds but a frequency-dependent divergence of the linewidth was observed below 40 K. The spectrum of PbCu2(PO4)2 at 70 K was more anisotropic than that of SrCu2 (PO4)2 although the only difference between the two structures is a small change in the angle between Cu-O polyhedra. The low value of the g-anisotropy of the two compounds was attributed to the unique dimer structure of MCu2(PO4)2. HFEPR measurements of the 1D spin-gap system Pb2V3O9 in the frequency range 120 to 364 GHz at low temperature showed82 signals in both the paramagnetic phase and the high ﬁeld phase above the critical ﬁeld. The signals showed the characteristic features in each phase, indicating that Pb2V3O9 undergoes ﬁeld-induced magnetic ordering above the critical ﬁeld. The spin-gap system TlCuCl3 undergoes a change to a ﬁeld-induced ordered state at low temperature. An anomaly in the magnetic-ﬁeld dependence of the g-value was found83 to correspond to the phase-transition ﬁeld. The critical ﬁeld was found to decrease under pressure. HFEPR measurements of powder and aligned samples of MCuP2O7 (M ¼ Sr, Pb) have been performed84 in the frequency range from 40 GHz to 370 GHz from 1.7 K to 265 K. A typical powder absorption line of Cu21 was observed in the powder sample. In the higher temperature region, gJ and g> are estimated to be 2.40 and 2.10 for SrCuP2O7 and 2.41 and 2.09 for PbCuP2O7, respectively. Both gJ and g> increased with decreasing temperature below 7 K. The gshifts were found to be frequency dependence but the direction of the shifts was not consistent with theoretical predictions for the 1D quantum magnet. Magneto-optical techniques were used85 to investigate the HFEPR spectra of single crystal and powder samples of the organic superconductor b00 -(BEDTTTF)4[(H3O)Fe (C2O4)3]C6H5CN in the frequency range 55–190 GHz. The Fe31 ions were found to occupy a site with distorted octahedral symmetry. The temperature- and frequency-dependences of the magneto-optical spectra were used to identify the nature of the interaction between localized Fe spins and between the local moments and the delocalized carriers. 2.7.3 Spin Peierls. The spin-Peierls transition in the organic compound CuGeO3 has been intensively studied. A lattice dimerization is found to take place below TSP E 14 K. In the dimerized phase the ground state is a spin singlet, separated from the ﬁrst excited triplet by an energy gap. The application of an external magnetic ﬁeld tends to suppress quantum ﬂuctuations and

Electron Paramag. Reson., 2007, 20, 245–270

265

eventually collapses the energy gap at 12.5 T. Above the threshold ﬁeld, CuGeO3 undergoes a transition from the commensurate dimerized spin-liquid phase to the incommensurate phase, where the periodicity of the spin-polarization and lattice deformation is incommensurate with crystallographic lattice parameters. The low-ﬁeld incommensurate region can be described by a formation of a regular array of domain walls (solitons). The low-ﬁeld region has been investigated8 by studying the EPR linewidth of a pure single crystal of CuGeO3 and one containing 0.8% Si. HFEPR was carried out at many frequencies between175 and 510 GHz in ﬁelds of up to 17 T. The pronounced increase of the magnon spin resonance linewidth with increasing ﬁeld, with a sharp maximum at 13.8 T, indicated the development of a soliton-like incommensurate superstructure. The anomaly was explained in terms of magnon–soliton scattering and a soliton-like phase close to the boundary of the dimerized-incommensurate phase transition. 2.7.4 2D Compounds. The properties of lithium–ion batteries are of great interest at present. Prospective cathode materials of the form (1-a) LiNi1yAlyO2.aLi[Li1/3Ni2/3]O2, where a is in the range 0 to 0.4, were formed by solid-state reaction under high pressure and investigated86 by HFEPR of Ni31 at up to 115 GHz. The oxidation state of nickel ions and cationic distribution of the various oxides was studied by measuring the EPR spectra as a function of temperature and frequency. The composite oxides display different electrochemical properties to Al-substituted LiNiO2. The layered-solid solutions of LiCoO2 with a-LiAlO2 are also of interest as cathode materials for lithium ion batteries since they are able to deintercalate/ intercalate lithium reversibly at a potential higher than 4 V. The reversible electrochemical extraction of Li takes place along with the reversible oxidation of Co31 to Co41. The local coordination of Ni31 trigonal LiAlyCo1yO2 was studied87 by HFEPR at frequencies up to 115 GHz. Ni31 ions were found to occupy sites with trigonal and tetragonal crystal ﬁelds. In a uniform Co environment the orbitally degenerate state for Ni31 is preserved. A local tetragonal distortion was observed for Ni31 ions located in a Co6yAly environment. References 1. Y.S. Lebedev, in Electron Spin Resonance, Vol. 14, ed. N.M. Atherton, M.J. Davies and B.C. Gilbert, The Royal Society of Chemistry, Cambridge, 1994, p. 63. 2. A.A. Doubinski, in Electron Spin Resonance, Vol. 16, ed. N.M. Atherton, M.J. Davies and B.C. Gilbert, The Royal Society of Chemistry, Cambridge, 1998, p. 211. 3. G.M. Smith and P.C. Riedi, in Electron Paramagnetic Resonance, Vol. 17, ed. N.M. Atherton, M.J. Davies and B.C. Gilbert, The Royal Society of Chemistry, Cambridge, 2000, p. 164. 4. G.M. Smith and P.C. Riedi, in Electron Paramagnetic Resonance, Vol. 18, ed. B.C. Gilbert, M.J. Davies and D.M. Murphy, The Royal Society of Chemistry, Cambridge, 2002, p. 254.

266

Electron Paramag. Reson., 2007, 20, 245–270

5. P.C. Riedi and G.M. Smith, in Electron Paramagnetic Resonance, Vol. 19, ed. B.C. Gilbert, M.J. Davies and D.M. Murphy, The Royal Society of Chemistry, Cambridge, 2004, p. 338. 6. J. Krzystek, A.T. Fiedler, J.J. Sokol, A. Ozarowski, S.A. Zvyagin, T.C. Brunold, J.R. Long, L.C. Brunel and J. Telser, Inorganic Chemistry, 2004, 43, 5645. 7. J. Krzystek, S.A. Zvyagin, A. Ozarowski, A.T. Fiedler, T.C. Brunold and J. Telser, Journal of the American Chemical Society, 2004, 126, 2148. 8. S.A. Zvyagin, J. Krzystek, P.H.M. van Loosdrecht, G. Dhalenne and A. Revcolevschi, Physica B-Condensed Matter, 2004, 346, 1. 9. J. Telser, J. van Slageren, S. Vongtragool, M. Dressel, W.M. Reiﬀ, S.A. Zvyagin, A. Ozarowski and J. Krzystek, Magnetic Resonance in Chemistry, 2005 43, S130. 10. H. Ohta, S. Okubo, Y. Inagaki, Z. Hiroi and H. Kikuchi, Physica B-Condensed Matter, 2004, 346, 38. 11. S. Okubo, H. Ohta, Y. Inagaki and T. Sakurai, Physica B-Condensed Matter, 2004, 346, 627. 12. S. Kimura, T. Otani, Y. Narumi, K. Kindo, M. Nakano and G. Matsubayashi, Journal of Magnetism and Magnetic Materials, 2004, 272–76, 1102. 13. H. Gunther, Magnetic Resonance in Chemistry, 2005, 43, S1. 14. W. Lubitz, K. Mobius and K.P. Dinse, Magnetic Resonance in Chemistry, 2005, 43, S2. 15. C. Dobe, C. Noble, G. Carver, P.L.W. Tregenna-Piggott, G.J. McIntyre, A.L. Barra, A. Neels, S. Janssen and F. Juranyi, Journal of the American Chemical Society, 2004, 126, 16639. 16. D. Arcon and R. Blinc, Structure and Bonding, 2004, 109, 231. 17. N. Weiden, T. Kato and K.P. Dinse, Journal of Physical Chemistry B, 2004, 108, 9469. 18. J. Rahmer, L. Dunsch, H. Dorn, J. Mende and M. Mehring, Magnetic Resonance in Chemistry, 2005, 43, S192. 19. M. Bortolus, M. Prato, J. van Tol and A.L. Maniero, Chemical Physics Letters, 2004, 398, 228. 20. A. Zoleo, M. Bellinazzi, M. Prato, M. Brustolon and A.L. Maniero, Chemical Physics Letters, 2005, 412, 470. 21. T. Galili, A. Regev, A. Berg, H. Levanon, D.I. Schuster, K. Mobius and A. Savitsky, Journal of Physical Chemistry A, 2005, 109, 8451. 22. M.E. Zvanut, Journal of Physics-Condensed Matter, 2004, 16, R1341. 23. R. Meisels, Semiconductor Science and Technology, 2005, 20, R1. 24. J.M. Spaeth, I. Tkach, S. Greulich-Weber and H. Overhof, Magnetic Resonance in Chemistry, 2005, 43, S153. 25. H.J. Kummerer and G. Denninger, Magnetic Resonance in Chemistry, 2005, 43, S145. 26. N.R.J. Poolton, G.M. Smith, P.C. Riedi, J.W. Allen, A.V. Firth, D.J. Cole-Hamilton and E.J.L. McInnes, Physica Status Solidi B-Basic Solid State Physics, 2005, 242, 829. 27. E. Goovaerts, S.V. Nistor, D. Ghica and T. Taniguchi, Physica Status Solidi a-Applied Research, 2004, 201, 2591. 28. G. Janssen, A. Bouwen and E. Goovaerts, Physical Review B, 2005, 71, art. no. 035415. 29. S.B. Orlinskii, H. Blok, E.J.J. Groenen, J. Schmidt, P.G. Baranov, C.D. Donegg and A. Meijerink, Magnetic Resonance in Chemistry, 2005, 43, S140.

Electron Paramag. Reson., 2007, 20, 245–270

267

30. V. Umamaheswari, M. Hartmann and A. Poppl, Journal of Physical Chemistry B, 2005, 109, 1537. 31. V. Umamaheswari, M. Hartmann and A. Poppl, Journal of Physical Chemistry B, 2005, 109, 19723. 32. V. Umamaheswari, M. Hartmann and A. Poppl, Magnetic Resonance in Chemistry, 2005, 43, S205. 33. D. Arieli, T.F. Prisner, M. Hertel and D. Goldfarb, Physical Chemistry Chemical Physics, 2004, 6, 172. 34. G. Mali, A. Ristic and V. Kaucic, Journal of Physical Chemistry B, 2005, 109, 10711. 35. B. Albela, R. Carina, C. Policar, S. Poussereau, J. Cano, J. Guilhem, L. Tchertanov, G. Blondin, M. Delroisse and J.J. Girerd, Inorganic Chemistry, 2005, 44, 6959. 36. D. Arieli, D.E.W. Vaughan and D. Goldfarb, Journal of the American Chemical Society, 2004, 126, 5776. 37. H. Mestric, R.A. Eichel, K.P. Dinse, A. Ozarowski, J. van Tol and L.C. Brunel, Journal of Applied Physics, 2004, 96, 7440. 38. E. Erdem, R. Bottcher, H.J. Glasel and E. Hartmann, Magnetic Resonance in Chemistry, 2005, 43, S174. 39. R.A. Eichel, H. Mestric, K.P. Dinse, A. Ozarowski, J. van Tol, L.C. Brunel, H. Kungl and M.J. Hoﬀmann, Magnetic Resonance in Chemistry, 2005, 43, S166. 40. B. Ruette, S. Zvyagin, A.P. Pyatakov, A. Bush, J.F. Li, V.I. Belotelov, A.K. Zvezdin and D. Viehland, Physical Review B, 2004, 69, art. no. 06414. 41. D. Viehland, J.F. Li, S. Zvyagin, A.P. Pyatakov, A. Bush, B. Ruette, V.I. Belotelov and A.K. Zvezdin, Ferroelectrics, 2004, 301, 229. 42. R. Boca, Coordination Chemistry Reviews, 2004, 248, 757. 43. C. Mantel, C. Baﬀert, I. Romero, A. Deronzier, J. Pecaut, M.N. Collomb and C. Duboc, Inorganic Chemistry, 2004, 43, 6455. 44. C. Mantel, C. Philouze, M.N. Collomb and C. Duboc, European Journal of Inorganic Chemistry, 2004, 3880. 45. C. Mantell, H.Y. Chen, R.H. Crabtree, G.W. Brudvig, J. Pecaut, M.N. Collomb and C. Duboc, Chemphyschem, 2005, 6, 541. 46. G. Aromi, J. Telser, A. Ozarowski, L.C. Brunel, H.M. Stoeckli-Evans and J. Krzystek, Inorganic Chemistry, 2005, 44, 187. 47. D. Baute and D. Goldfarb, Journal of Physical Chemistry A, 2005, 109, 7865. 48. P.L.W. Tregenna-Piggott, D. Spichiger, G. Carver, B. Frey, R. Meier, H. Weihe, J.A. Cowan, G.J. McIntyre, G. Zahn and A.L. Barra, Inorganic Chemistry, 2004, 43, 8049. 49. P.L.W. Tregenna-Piggott and G. Carver, Inorganic Chemistry, 2004, 43, 8061. 50. G.S. Shakurov, B.Z. Malkin, A.R. Zakirov, A.G. Okhrimchuk, L.N. Butvina, N.V. Lichkova and V.N. Zavgorodnev, Applied Magnetic Resonance, 2004, 26, 579. 51. A.M. Raitsimring, A.V. Astashkin, O.G. Poluektov and P. Caravan, Applied Magnetic Resonance, 2005, 28, 281. 52. A.L. Barra, D. Gatteschi, R. Sessoli and L. Sorace, Magnetic Resonance in Chemistry, 2005, 43, S183. 53. R.S. Edwards, S. Hill, P. Goy, R. Wylde and S. Takahashi, Physica B-Condensed Matter, 2004, 346, 211. 54. E. del Barco, A.D. Kent, S. Hill, J.M. North, N.S. Dalal, E.M. Rumberger, D.N. Hendrickson, N. Chakov and G. Christou, Journal of Low Temperature Physics, 2005, 140, 119. 55. K. Petukhov, S. Hill, N.E. Chakov, K.A. Abboud and G. Christou, Physical Review B, 2004, 70, art. no. 054426.

268

Electron Paramag. Reson., 2007, 20, 245–270

56. H. Hachisuka, K. Awaga, T. Yokoyama, T. Kubo, T. Goto and H. Nojiri, Physical Review B, 2004, 70, art. no. 10442. 57. H. Oshio, M. Nihei, A. Yoshida, H. Nojiri, M. Nakano, A. Yamaguchi, Y. Karaki and H. Ishimoto, Chemistry-a European Journal, 2005, 11, 843. 58. C.T. Xu and J.Q. Liang, European Physical Journal B, 2005, 44, 469. 59. S. Hill, R.S. Edwards, N. Aliaga-Alcalde and G. Christou, Science, 2003, 302, 1015. 60. Y. Yukawa, G.A. Aromi, S. Igarashi, J. Ribas, S.A. Zvyagin and J. Krzystek, Angewandte Chemie-International Edition, 2005, 44, 1997. 61. E.C. Yang, C. Kirman, J. Lawrence, L.N. Zakharov, A.L. Rheingoid, S. Hill and D.N. Hendrickson, Inorganic Chemistry, 2005, 44, 3827. 62. C. Kirman, J. Lawrence, S. Hill, E.C. Yang and D.N. Hendrickson, Journal of Applied Physics, 2005, 97, art. no. 10M501. 63. G. Rogez, J.N. Rebilly, A.L. Barra, L. Sorace, G. Blondin, N. Kirchner, M. Duran, J. van Slageren, S. Parsons, L. Ricard, A. Marvilliers and T. Mallah, Angewandte Chemie-International Edition, 2005, 44, 1876. 64. E.A. Buvaylo, V.N. Kokozay, O.Y. Vassilyeva, B.W. Skelton, J. Jezierska, L.C. Brunel and A. Ozarowski, Inorganic Chemistry, 2005, 44, 206. 65. S. Nellutla, J. van Tol, N.S. Dalal, L.H. Bi, U. Kortz, B. Keita, L. Nadjo, G.A. Khitrov and A.G. Marshall, Inorganic Chemistry, 2005, 44, 9795. 66. I. Gautier-Luneau, D. Phanon, C. Duboc, D. Luneau and J. L. Pierre, Dalton Transactions, 2005, 3795. 67. S. Sharmin, A. Ardavan, S.J. Blundell, A.I. Coldea, E.J.L. McInnes and D. Low, Journal de Physique IV, 2004, 114, 645. 68. T. Yamase, E. Ishikawa, K. Fukaya, H. Nojiri, T. Taniguchi and T. Atake, Inorganic Chemistry, 2004, 43, 8150. 69. J.M. Rawson, A. Alberola, H. El-Mkami and G.M. Smith, Journal of Physics and Chemistry of Solids, 2004, 65, 727. 70. Y. Maeda and M. Oshikawa, Journal of the Physical Society of Japan, 2005, 74, 283. 71. B. Ohta, S. Okubo, D. Fukuoka, Y. Inagaki, T. Kunimoto, M. Kimata, K. Koyama, M. Motokawa and Z. Hiroi, Journal of Magnetism and Magnetic Materials, 2004, 272–76, 929. 72. M. Oshikawa and I. Afﬂeck, Physical Review B, 2002, 65, art. no. 134410. 73. T. Sakurai, M. Saruhashi, Y. Inagaki, S. Okubo, H. Ohta, H. Tanaka and Y. Uwatoko, Physica B-Condensed Matter, 2004, 346, 221. 74. S. Okubo, T. Kamikawa, T. Kunimoto, Y. Inagaki, H. Ohta, H. Nojiri and H. Kikuchi, Journal of Magnetism and Magnetic Materials, 2004, 272–76 912. 75. S. Okubo, A. Taketani, H. Ohta, T. Kunimoto, Y. Inagaki, T. Saito, M. Azuma, M. Takano and H. Kikuchi, Progress of Theoretical Physics Supplement, 2005, 11. 76. Y. Inagaki, D. Fukuoka, T. Sakurai, S. Okubo, H. Ohta, H. Sakurai, K. Yoshimura, K. Kosuge, H. Michor and G. Hilscher, Physica B-Condensed Matter, 2004, 346, 60. 77. S. Okubo, T. Hirano, Y. Inagaki, H. Ohta, H. Sakurai, H. Yoshimura and K. Kosuge, Physica B-Condensed Matter, 2004, 346, 65. 78. A.K. Kolezhuk and H.J. Mikeska, Journal of Magnetism and Magnetic Materials, 2004, 272–76, 902. 79. Z. Honda, M. Hagiwara, K. Katsumata and K. Yamada, Journal of Applied Physics, 2004, 95, 7543. 80. M. Yoshida, K. Shiraki, S. Okubo, H. Ohta, T. Ito, H. Takagi, M. Kaburagi and Y. Ajiro, Physical Review Letters, 2005, 95, art. no. 117202.

Electron Paramag. Reson., 2007, 20, 245–270

269

81. T. Kaji, S. Okubo, H. Ohta, Y. Inagaki, A.A. Belik, M. Azuma and M. Takano, Journal of Physics and Chemistry of Solids, 2005, 66, 2068. 82. M. Kodama, M. Yoshida, S. Okubo, H. Ohta, T. Waki, Y. Morimoto, C. Michioka, M. Kato and K. Yoshimura, Progress of Theoretical Physics Supplement, 2005, 114. 83. A. Taketani, T. Sakurai, M. Kodama, S. Okubo, H. Ohta, H. Tanaka and Y. Uwatoko, Progress of Theoretical Physics Supplement, 2005, 407. 84. K. Shiraki, M. Yoshida, S. Okubo, H. Ohta, A.A. Belik, M. Azuma and M. Takano, Progress of Theoretical Physics Supplement, 2005, 168. 85. A. Narduzzo, A. Ardavan, J. Singleton, L. Pardi, V. Bercu, A. Akutsu-Sato, H. Akutsu, S.S. Turner and R. Day, Journal De Physique IV, 2004, 114, 347. 86. E. Shinova, E. Zhecheva, R. Stoyanova, G.D. Bromiley, R. Alcantara and J.L. Tirado, Journal of Solid State Chemistry, 2005, 178, 2692. 87. R. Stoyanova, E. Zhecheva, R. Alcantara and J.L. Tirado, Journal of Physical Chemistry B, 2004, 108, 4053.