Nanoparticles in medicine and environment: Inhalation and health effects

Nanoparticles in Medicine and Environment

Jan C.M. Marijnissen

l

Leon Gradon´

Editors

Nanoparticles in Medicine and Environment Inhalation and Health Effects

Editors Jan C.M. Marijnissen Delft University of Technology Dept. Chemical Engineering Julianalaan 136 2628 BL Delft Netherlands

Leon Gradoń Faculty of Chemical and Process Engineering, Warsaw University of Technology, Warynskiego 1 00-645 Warsaw Poland

ISBN 978-90-481-2631-6 e-ISBN 978-90-481-2632-3 DOI 10.1007/978-90-481-2632-3 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2009941459 # Springer Science+Business Media B.V. 2010 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Cover design: Boekhorst Design Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

Introduction

Currently a huge effort is put into nanoparticle research and the production of nanoparticles. In many cases it is unavoidable that during the production processes or during the use of the particles or a product made from these particles, nanoparticles are released into the environment. It is also realized that combustion processes, including traffic and power plants release nanoparticles into the atmosphere. However it is not known how nanoparticles interact with the human body, especially upon inhalation. At the same time research activities are devoted to understand how nano-sized medicine particles can be used to administer medicines via inhalation. In any case it is absolutely necessary to know how the nanoparticles interfere with the inhalation system, how they deposit in and affect the human system. During May 30 and 31, 2008, a group of scientists met in the Jablonna Palace near Warsaw in Poland to discuss all aspects from the origin/production of the nanoparticles till the interaction with the lungs and the toxic/therapeutic effects. This workshop, the fourth in a series of very specialized workshops on aerosol particles and the human body, brought together top-experts from different disciplines but all in the field related to aerosol nanoparticle release/production and nanoparticles inhalation and the effects. The subjects were assembled in three main themes: Sources and Production, Inhalation and Deposition, Toxicological and Medical Consequences. The workshop was concluded with an overview and a roundtable discussion. The chapters in this book, including the last one, which reports the overview and discussions, make clear how complex the subject is and that it only can be attacked by an interdisciplinary approach. It also made evident that much is still unknown. Yet we are confident that this book presents the state of the art in the field and that it sets directions on how to proceed in each different part of this important health issue. It also made clear how crucial it is to work together with all the different subdisciplines. As mentioned, the Jablonna 4 workshop was held in the beautiful Renaissance Jablonna Palace, surrounded by a big park, which contributed greatly in making this very fruitful scientific meeting again a very pleasant get-together. The Polish v

vi

Introduction

hospitality was highly appreciated. Finally the fantastic music, from the Dutch composer/graphic artist Juriaan Andriessen, realized by the Dutch pianist Jetje van Wijk with contributions of the Polish puppeteer Andrzej Bocian and adorned with visionary paintings of the composer as made into a slide show by Allert Schallenberg, completed the wonderful workshop. The workshop was sponsored by the Warsaw University of Technology, Delft University of Technology, Alistore-ERI, the German Gesellschaft fu¨r Aerosol Forschung and some industries. Leon Gradon´ Jan C.M. Marijnissen

Warsaw University of Technology Delft University of Technology

Contents

1

The Origin and Production of Nanoparticles in Environment and Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Heinz Burtscher

2

Characterization of Combustion and Engine Exhaust Particles . . . . . . . . 19 M. Matti Maricq

3

Medicine Nanoparticle Production by EHDA . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Jan C.M. Marijnissen, Caner U. Yurteri, Jan van Erven, and Tomasz Ciach

4

Electrospray and Its Medical Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Da-Ren Chen and David Y.H. PUI

5

Generation of Nanoparticles from Vapours in Case of Exhaust Filtration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Markku Kulmala and Mikko Sipila¨

6

Measurement and Characterization of Aerosol Nanoparticles . . . . . . . . . . 91 Wladyslaw W. Szymanski and Gu¨nter Allmaier

7

Inhalation and Deposition of Nanoparticles. Fundamentals, Phenomenology and Practical Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Arkadiusz Moskal, Tomasz R. Sosnowski, and Leon Gradon´

8

Dosimetry of Inhaled Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 Wolfgang G. Kreyling and Marianne Geiser

9

Particles of Biomedical Relevance and Their Interactions: A Classical and Quantum Mechanistic Approach to a Theoretical Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Ewa Broclawik and Liudmila Uvarova vii

viii

Contents

10

Health Effects of Nanoparticles (Inhalation) from Medical Point of View/Type of Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Robert Baughman and Michal Pirozynski

11

Effects of Cigarette Smoke and Diesel Exhaust on the Innate Immune Function of the Airway Epithelium . . . . . . . . . . . . . . . . . . 203 P.S. Hiemstra

12

The Potential Harmful and Beneficial Effects of Nanoparticles in Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211 Karen G. Schu¨epp

13

Targeting Drugs to the Lungs – The Example of Insulin . . . . . . . . . . . . . 227 S. Ha¨ussermann, G. Scheuch, and R. Siekmeier

14

Protection of the Respiratory System Against Nanoparticles Inhalation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 Albert Podg´orski

15

Overview and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 Jan C.M. Marijnissen, Leon Gradon´, and Bob W.N.J. Ursem

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285

Contributors

Gu¨nter Allmaier Vienna University of Technology, Institute of Chemical Technologies and Analytics, Getreidemarkt 9/164-IAC, A-1060, Vienna, Austria Robert Baughman University of Cincinnati Medical Center, 1001 HH Eden Avenue and Albert Sabin Way, P.O. Box 670565, Cincinnati, OH 45267-0001, USA Ewa Broclawik Institute of Catalysis and Surface Chemistry, Polish Academy of Sciences, Krakow, Poland, [email protected] Heinz Burtscher Institute for Aerosol and Sensor Technology, University of Applied Sciences, Northwestern Switzerland, CH-5210 Windisch, Switzerland Tomasz Ciach Faculty of Chemical and Process Engineering, Warsaw University of Technology, Warsaw, Poland Da-Ren Chen Department of Energy, Environmental and Chemical Engineering, Washington University, St. Louis, MO USA Jan van Erven Nano Structured Materials, TU Delft, Juliananlaan 136, 2628 BL Delft, the Netherlands Marianne Geiser Institute of Anatomy, University of Bern, Baltzerstrasse 2, CH-3000 Bern 9, Switzerland ix

x

Contributors

Leon Gradon´ Faculty of Chemical and Process Engineering, Warsaw University of Technology, Warynskiego 1 00 – 645 Warsaw, Poland S. Ha¨ussermann Air Liquide Research Center CRCD, Paris, France P. S. Hiemstra Department of Pulmonology, Leiden University Medical Center, Leiden, the Netherlands, [email protected] Wolfgang G. Kreyling Institute for Inhalation Biology and Focus-Network Nanoparticles and Health, Helmholtz Center Munich, German Research Center for Environmental Health, Ingolstaedter Landstrasse 1, D-85764 Neuherberg/Munich, Germany Markku Kulmala Department of Physics, University of Helsinki, P.O. Box 64 (Gustaf Ha¨llstro¨minkatu 2), FI-00014, University of Helsinki, Finland Helsinki Institute of Physics, P.O. Box 64, FI-00014, University of Helsinki, Finland M. Matti Maricq Research and Advanced Engineering, Ford Motor Company, MD 3179, P.O. Box 2053, Dearborn, MI 48121, USA, [email protected] Jan C.M. Marijnissen Faculty of Applied Sciences, Delft University of Technology, Delft, the Netherlands Arkadiusz Moskal Faculty of Chemical and Process Engineering, Warsaw University of Technology, Warynskiego 1 00-645 Warsaw, Poland Michal Pirozynski Department of Anesthesiology and Intensive Therapy, CMKP, 241 Czerniakowska Street, 00-416 Warsaw, Poland Albert Podgo´rski Faculty of Chemical and Process Engineering, Warsaw University of Technology Waryn´skiego 1, 00-645 Warsaw, Poland David Y. H. PUI Director of Particle Technology Laboratory, Department of Mechanical Engineering, University of Minnesota, Minneapolis, MN USA

Contributors

xi

G. Scheuch Activaero GmbH, Gemu¨nden, Germany Karen G. Schu¨epp Department of Paediatric Respiratory Medicine, University Children’s Hospital, Bern, Switzerland and Swiss Paediatric Respiratory Research Group, Switzerland R. Siekmeier Federal Institute for Drugs and Medical Devices (BfArM), Bonn, Germany Mikko Sipila¨ Department of Physics, University of Helsinki, Finland, P.O. Box 64 (Gustaf Ha¨llstro¨minkatu 2), University of Helsinki, Finland and Helsinki Institute of Physics, P.O. Box 64, FI-00014 University of Helsinki, Finland. Tomasz R. Sosnowski Faculty of Chemical and Process Engineering, Warsaw University of Technology, Warynskiego 1 00 – 645 Warsaw, Poland Wladyslaw W. Szymanski Faculty of Physics, University of Vienna, Aerosol Physics, Biophysics and Environmental Physics Research Group, Boltzmanngasse 5, A-1090 Vienna, Austria Bob W.N.J. Ursem Faculty of Applied Sciences, Delft University of Technology Delft, Netherlands L. Uvarova Department of Applied Mathematics, Moscow State University of Technology “STANKIN”, Moscow, Russia Caner U. Yurteri Faculty of Applied Sciences, Delft University of Technology, Delft, the Netherlands

Chapter 1

The Origin and Production of Nanoparticles in Environment and Industry Heinz Burtscher

1.1

Introduction

Together with nitrogen oxides (NOX) particulate matter (PM) is considered one of the most important pollutants in ambient air. Many toxicological and epidemiological studies established adverse health effects by particulate matter. In most of these studies the particle mass in terms of PM10 or PM2.5 is used. There is increasing evidence that several health effects are associated with the ultra fine particles with diameters below 100 nm (Brown et al. 2001). Recent research shows that they can penetrate the cell membranes, enter into the blood and even reach the brain (Oberdo¨rster et al. 2004). Some investigations indicate that particles can induce heritable mutations (Somers et al. 2004). Usually approximately 90% of PM consists of fine and ultrafine particles (UBA 2005). Particulate matter in the ambient air is a mixture of directly emitted primary aerosol particles and secondary aerosol particles formed in the atmosphere. Coarse particles from primary aerosols originate mainly from mechanical processes (construction activities, road dust, re-suspension, wind, etc.) whereas fine particles are particularly produced through combustion (WHO 2006). So far the discussion is focused on vehicle emissions from diesel engines. Due to their adverse health effects and their abundance in the vicinity of roads, in particular in urban areas, they have become of great concern in the past years (Lighty et al. 2000, Wichmann and Peters 2000). However, other combustion systems as for example such for biomass combustion also have a significant contribution. Secondary aerosols are formed in the atmosphere through conversion of gaseous precursors such as sulphur oxides (SO2, SO3), nitrogen oxides (NO, NO2), ammonia

H. Burtscher 1 Institute for Aerosol and Sensor Technology, University of Applied Sciences, Northwestern Switzerland, CH-5210, Windisch, Switzerland e-mail: [email protected]

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_1, # Springer ScienceþBusiness Media B.V. 2010

1

2

H. Burtscher

(NH3) and Non-Methane Volatile Organic Compounds (NMVOC). Reaction products among many others are ammonium sulphates and ammonium nitrate, which often dominate PM mass in ambient air. According to (UBA 2005) important sources of precursors for secondary aerosols are agriculture (NH3), Diesel engines (NOX), other combustion processes (SO2, SO3 and NO, NO2) and the use of solvents, chemical industry and petro chemistry (NMVOC). In the next section some data on particulate matter in terms of mass concentrations will be given, and then emissions from some important contributors and measures taken to reduce these emissions will be discussed. This will by far not be complete, but I hope that the most important contributors are mentioned. The focus will lie on the submicron fraction.

1.2

Particulate Matter (PM10), Sources and Composition

Most available data for ambient air particulate matter concerns total mass concentrations (PM10, PM2.5). Only these are routinely measured and limited. Number concentrations, size distributions, and detailed chemical composition are determined only in research work. Main components of the ambient aerosol are l

l

l

l

l

l

Salts (most abundant ammonium sulphate and nitrate): This is secondary aerosol formed from gaseous precursors, mainly NH3, SO2 and NOx. The salts are found in the accumulation mode. Elemental carbon (EC): EC mainly arises from incomplete combustion, important sources being diesel engines. In ambient air EC usually is found as agglomerates in the accumulation mode with a number of volatile species adsorbed on the surface. Coarse mode EC particles are due to tire wear. Organic carbon (OC): In incomplete combustion (combustion engines, biomass combustion) a variety of organic species are produced, which are emitted in the gas phase and condense on solid cores or nucleate when the exhaust cools. More volatile material may form secondary aerosol upon oxidation in the atmosphere. Most OC is found in the fine fraction. Anthropogenic coarse OC may for example be due to tire wear or wood processing. Minerals (aluminium silicate, calcium carbonate, gypsum): From construction, road dust, agriculture, coarse particles. Sodium chloride: Near coasts from sea salt, during winter time road salt for de-icing. Metal particles: From traffic, engine wear, break wear and industrial processes.

Figure 1.1 shows examples for the composition of urban (Fig. 1.1a) and rural (Fig. 1.1b) PM10, measured at two locations in Switzerland. The urban site is in the city of Berne, the rural in Chaumont. Composition data from many other locations are similar.

1 The Origin and Production of Nanoparticles in Environment and Industry

a

17%

7%

b

10%

13%

ammonium 8%

4%

nitrate

12%

2% 5%

11%

3

sulfate EC

9%

OC 29%

17%

30%

21%

mineral dust trace elements unidentified

5%

Fig. 1.1 Typical (a) urban (city of Bern) and (b) rural (Chaumont) PM10 composition (data from EKL 2007)

domestic 11%

industry agriculture and forrestry traffic

29%

30%

30%

Fig. 1.2 Contribution of different sources to PM10 (from EKL 2007)

Important producers of anthropogenic particulate matter are l

l l

l

Road traffic (soot emissions, mainly from diesel engines, tire-, brake- and clutch wear, road dust). In urban areas traffic is the main source. Mobile off road diesel engines (forest work, agriculture, construction). Stationary sources: Combustion of wood, coal and other fuels, industrial emissions. In addition to primary aerosol these sources also emit gaseous precursors leading to the formation of secondary aerosol in the atmosphere (NH3, NOX, sulphur compounds, VOC).

The quantitative contribution of important sources to the primary aerosol mass is shown in Figs. 1.2 and 1.3. In a recent paper presenting data from India (Delhi) the contribution of traffic is much higher (Srivastava et al. 2008). Figure 1.4 shows emissions of the major precursor gases. NOX is mainly due to traffic, SOX and NMVOC are dominated by industrial emissions and NH3 stems almost completely from agriculture.

4

H. Burtscher wood combustion diesel

8%

gasoline wood combustion in forrest other combustion

17% 1% 57%

non combustion

7%

10%

Fig. 1.3 Contribution of combustion and non combustion to PM10 (from EKL 2007)

NMVOC

NOx

6%

13%

6%

22%

13% 22%

23%

domestic

23%

industry

6% 6%

57% 57%

agriculture and forrestry

14% 59%

14%

traffic

59% NH3

0% 2% 2%

SOx

1%

96%

9%

23%

67%

Fig. 1.4 Contribution of different emitters to the major precursors of secondary aerosol (NOX, non methane VOC (NMVOC), NH3 and SOX) (data from EKL 2007)

1 The Origin and Production of Nanoparticles in Environment and Industry

1.3

5

Particles from Diesel Engines

Diesel particles as well as particles from other combustion sources are a complex mixture of elemental carbon, a variety of hydrocarbons, sulphur compounds and other species. Particles differ in size, composition, solubility and therefore also in their toxic properties. Figure 1.5 shows the typical composition of particles from diesel engines (Kittelson 1998). The data shown in Fig. 1.5 are average values over a number of heavy-duty engines measured during a transient cycle. Operating conditions and engine type strongly influence the composition. For example the EC fraction may be greater than 80%, in particular in high load conditions. The exhaust contains a significant volatile fraction. Depending on temperature and other conditions the volatile fraction may l l l

Remain in the gas phase Condense on existing solid particles Nucleate and form new particles

Different amounts of the species mentioned above will consequently be measured as ‘particulate’ emissions, depending on how samples are taken (location of sampling, temperature, dilution, etc.). The sample contains not only particles, formed in the combustion process, but also secondary particles, formed during cooling in the exhaust and sampling lines. In addition to these ‘directly condensing’ volatile material secondary aerosol is formed later in the atmosphere. According to work by Robinson et al. (2007) the importance of aerosol is formed later in the atmosphere has been strongly underestimated up to now. On the other hand, much of the previously condensed material re-evaporates in the atmosphere according to the results of these authors. Diesel particles are agglomerates consisting of mainly spherical primary particles of about 15–40 nm in diameter. A study by Su et al. (2004) indicates that the primary particles emitted from modern engines fulfilling the EURO IV limits are smaller than those from older engines. These authors also find differences in

Fig. 1.5 Composition of diesel particles, average values for heavy duty engines during transient test cycle (from Kittelson 1998)

6

H. Burtscher 1.5E+13 2000RPM/100% 1400RPM/100% 2000RPM/50%

1E+13

part/kwh

1400RPM/50%

5E+12

0

10

100

d [nm]

1000

Fig. 1.6 Size distribution of diesel particles at different operating conditions (different speed and 50% or 100% load)

the particle microstructure. Whereas for conventional engines amorphous and graphitic structures are dominant, they observed a higher fraction of fullerenelike soot primary particles for the modern engine they tested. They also found that these particles can be oxidized more easily. The number size distribution of the agglomerated particles (accumulation mode) peaks almost always in the range of 60 to 120 nm. An example for size distributions of particles in the exhaust of a diesel engine used in machines for building or street construction, e.g. diggers, in different operating conditions is given in Fig. 1.6. As shown by this example the size is relatively insensitive to the operating conditions of the engine. Only few extreme conditions lead to significantly different size distributions. There is also no strong dependence of the particle size on the type of engine. The size distributions are lognormal with an almost constant geometric standard deviation of 1.8–1.9 (Harris and Maricq 2002). As will be shown later the accumulation mode can be accompanied by a ‘nucleation mode’, consisting of much smaller particles. Modern trap technology allows a very efficient removal of solid particles, also of the nanometer-sized fraction. Volatile material passes the trap in the gas phase. As the solid surface to condense on has been removed by the trap, nucleation becomes much more probable. The number concentration of particles downstream a particle trap is often dominated by volatile particles in the nucleation mode (see Fig. 1.7). Nucleation is further enhanced by oxidation due to catalytic active devices (catalytic converters or catalysts used for trap regeneration). For example the oxidation of SO2 to SO3, in combination with water may lead to the formation of sulphuric acid droplets. Whereas a strong correlation of the occurrence of nucleation and the fuel sulphur content is observed, chemical analysis of nucleation mode particles by thermal desorption particle beam mass spectrometry shows that the sulphur content of these particles is only a few percent or even less (Sakurai et al. 2003).

1 The Origin and Production of Nanoparticles in Environment and Industry

7

1000000

concentration [#/cm3]

without trap

100000

10000 with trap

1000

100

10

100 diameter [nm]

1000

Fig. 1.7 Size distribution of diesel particles with and without particle trap

They mainly consist of organic material. The finger print of the measured mass spectra indicates that this material stems mainly from lubricant oil and only to a small fraction from the fuel (Sakurai et al. 2003). This indicates that the first step in nucleation particle formation is the nucleation of sulphuric acid and water, followed by particle growth by condensation of organic species. A detailed theoretical investigation of this process has been done by Vouitsis et al. (2008). Downstream a trap the particle composition is dominated by volatile material (Burtscher 2005). Engine optimization led to a significant reduction in emitted mass; however, the number concentration is not significantly reduced. Comparing the size distribution of EURO3 to EURO5 engines shows that mainly the tail on the large particle side is reduced for modern engines (Mayer et al. 2007). The same study also shows that by open traps, introduced recently, only a small reduction in particle emissions is achieved. A significant reduction is only obtained by closed traps. They remove more than 99.9 %. Catalytic aftertreatment devices (oxidation catalysts to reduce hydrocarbon emissions, catalytically active filters to assist filter regeneration) also oxidize NO, which means that direct NO2 emissions are significantly increased (Du¨nnebeil et al. 2007), if such devices are applied. This leads to an increase in NO2 concentrations measured near highways. As NO is rapidly oxidized in the atmosphere anyway this effect is only observed in the direct vicinity of frequented roads. Emissions from very large diesel engines for marine applications or electrical power production differ significantly from those of smaller engines. This is on the one hand due to the different fuel (much higher sulphur and ash content) but also to

8

H. Burtscher

different operation conditions as for example the much lower speed. The fraction of volatile organic material is significantly higher, whereas the EC fraction is small. The average particle diameter is only about 50 nm (Kasper et al. 2007).

1.4 1.4.1

Partilces from Other Combustion Sources Spark Ignition Engines

Particles from gasoline engines are smaller than those from diesel engines (typically 40 nm average diameter) and they contain a large fraction of volatile material. Measurements from a small engine for electrical power production show a mean diameter of 40 nm and a fractal dimension df of three, when measured at ambient temperature. This indicates compact particles. If these particles are heated to a temperature of about 250 C, they become significantly smaller (mean diameter 20 nm) and df decreases from 3 to about 2.2. An explanation for this observation may be that at ambient temperature the solid, fractal part of the particles is encapsulated in volatile material. If this volatile material is removed by heating, the fractal core becomes ‘visible’ (Burtscher 2000). Emissions from well maintained port injection spark ignition engines are relatively low during stationary operation of the warm engine. However, they increase by orders of magnitude during acceleration (Kasper et al. 2005). Emissions from the cold engine are also much higher. The situation is very different for direct injection spark ignition engines (lean engines). These engines have been introduced because their fuel consumption in partial load conditions is lower. However, their particle emissions are somewhere in between those from conventional port injection gasoline engines and diesel engines (Aufdenblatten et al. 2002; Andersson et al. 2007). Two Stroke engines, for example in scooters, emit very high concentrations of particles, consisting mainly of organic carbon (Czerwinski et al. 2006).

1.4.2

Wood Combustion

Wood combustion in small domestic furnaces or stoves contributes significantly to ultrafine particle concentrations in ambient air. If the conditions for the combustion are good, most particles are in the submicron range. Emissions are high in the start up phase. Later the number concentration decreases and the particles become smaller. Typical size distributions in different phases of the burning cycle are shown in Fig. 1.8. The situation completely changes, if the furnace is not operated properly, in particular if the air supply is too low. In this case particle emissions may be dramatically higher and the size distribution is shifted to much larger particles (see for example Nussbaumer et al. 2008).

1 The Origin and Production of Nanoparticles in Environment and Industry

9

9.0E+07

start up

dn/dlog(d)(cm–3)

intermediate

6.0E+07

burn out

3.0E+07

0.0E+00 10

100 d [nm]

1000

Fig. 1.8 Size distribution of particles from wood combustion (log wood stove) at different phases of the combustion

Particles from wood combustion mainly consist of three fractions: (1) an inorganic fraction (minerals and salts, dominated by potassium and calcium compounds), (2) elemental carbon (EC) and (3) organic carbon (OC). For bad operating conditions in particular the organic fraction may become very high. The OC/EC ratio is much higher than for diesel engines. Pellet and chip furnaces allow a very good combustion. In this case the inorganic fraction is dominant. Typical mass emission factors are 30–60 mg/MJ. Emission factors for wood log stoves range from 10 mg/MJ for optimal operation up to several 1000 mg/MJ. In Fig. 1.9 total mass (TM) is plotted versus total organic carbon mass (OC) for different furnaces and operating conditions. The plot shows that TM and OC decrease both, when optimizing the combustion. After a base level is reached a further reduction of organic material by better combustion has no more significant influence of the total mass, which is now dominated by inorganic non-combustible material (Johansson et al. 2004). Toxicology tests (Klippel and Nussbaumer 2007) indicate that particles emitted from well operated furnaces (mainly minerals and salts) are less toxic than diesel particles, but those from high emitters (dominated by organic material) are much worse. Recently small electrostatic precipitators became available, which allow a reduction of emitted particles up to 90% (Schmatloch and Rauch 2005). The efficiency depends mainly on the flow rate of the exhaust gas.

10

H. Burtscher 10000 Wood pellets Wood logs

TM (mg/MJ)

1000

inorganic fraction organic fraction

100

10

1 1

10

100 OC (mg/MJ)

1000

10000

Fig. 1.9 Total mass (TM) emission factor versus emission factor for organic carbon (OC) (data from Johansson et al. 2004)

1.4.3

Waste Incineration

Waste incineration used to be a relevant source for emission of dust, heavy metals, acids, and many other species. Meanwhile the plants have been equipped with efficient flue gas cleaning devices, which led to a significant reduction of the above mentioned pollutants. Flue gas cleaning is done by the following devices: l

l l

Electrostatic precipitators (usually dry filters, sometimes wet filters are applied) or fabric filters for removal of particles. Wet scrubbers (removal of SO2, HCl, HF, Heavy metals, Aerosol particles). DeNOX system. Selective non catalytic reduction (SNCR) of NOX by NH3 injection (NOx is reduced at about 900 C) or more frequently Selective Catalytic Reduction (SCR) of NOx also by NH3 injection. For SCR significantly lower temperatures are sufficient. The catalyst usually is the last stage of the cleaning system.

The typical capacity of one combustion line of a waste incineration plant is in the order of 100000 Nm3/h. The plants have 2–3 lines. Figure 1.10 shows an example of size distributions, measured in the raw exhaust, after the electrostatic precipitator (ESP) and in the stack (after wet scrubber and SCR, Burtscher et al. 2002). The raw emissions are very high, but already the ESP removes about 99.9 % of the particles. The increase at the small particle side most probably has to be ascribed to nucleation of volatile material. The wet scrubber reduces the particle concentration by another order of magnitude. The resulting

1 The Origin and Production of Nanoparticles in Environment and Industry

11

1.E+09 solid: SMPS, dots: OPC

1.E+08

raw gas

dN/d log(d) [cm–3]

1.E+07 1.E+06

after ESP

1.E+05 1.E+04 1.E+03

stack

1.E+02 1.E+01 10

100

1000

10000

d [nm]

Fig. 1.10 Particles in the exhaust of a waste incineration plant. Concentrations in the raw gas, after the electrostatic precipitator, and in the stack are plotted

stack concentration is in the order of ambient air concentrations. Measurements at a number of plants corroborated these results. This demonstrates that the flue gas cleaning system of modern waste incineration plants is very efficient concerning fine particle removal. Properly maintained state of the art waste incineration plants are no relevant source for particulate matter.

1.4.4

Boilers and Furnaces

Particle emissions from domestic heating other than wood combustion are very small. Well maintained gas- and oil heatings allow an almost complete combustion. The same can be said about gas turbines. A comparison of the particle size distributions in the exhaust of a number of combustion systems can be found in Nussbaumer (2004). Table 1.1 shows a comparison of emissions from gas, oil and wood.

1.5

Noncombustion Particles

Traffic contributes to particulate pollution not only via tail pipe emissions, but also by road dust, tire wear and so on. In terms of PM10 emissions the contribution of this ‘non-combustion’ fraction is significant. Figure 1.11 shows the relative

12

H. Burtscher

Table 1.1 Emission factors for particle mass and number concentration for heating systems with different fuels

Mass (mg/MJ) Natural Gas <0.01 Oil (extra light) 1 Wood 70–170 Source: Data from Wieser et al. 2001.

Particle number (#/MJ) 2  1010 2  1013 5–15  1013

contribution of diesel and gasoline engines according to data of the Swiss Federal Office for the Environment (Keller and Zbinden 2004). The relative high contribution of gasoline reflects the percentage of gasoline cars, which is much higher in Switzerland, compared to EU countries. The mass emissions are dominated by the non-combustion fraction. However, these are mainly coarse particles and have almost no contribution to the submicron fraction. Beside road traffic railways also emit particles. When electric locomotives are used these are all ‘non-combustion’ particles. Railway emissions are dominated by metals (iron, copper, chromium and manganese). These emissions are due to abrasion of tracks, wheels, brakes and overhead traction lines. The latter are responsible for the copper. Iron is the dominant species. In the direct vicinity of highly frequented railway lines iron and copper concentration have been measured, which are about three times higher than at comparable background locations (Gehrig et al. 2007). However, already at distances of a few hundred meters the increase becomes insignificant. Very much higher concentrations are found in subway stations.

1.6

Occupational Exposition

So far emissions into ambient air have been discussed. Another topic of interest is occupational exposition. As combustion particle concentrations are particularly high on roads, the exposition is high for drivers. Car ventilations systems are meanwhile equipped with filters, however, most of this filters are inefficient for nanoparticles. The concentration in the car is almost the same as outside (Burtscher et al. 2008). Still higher concentrations occur at poorly ventilated construction sites, where diesel engines are used. An example is tunnel construction, where the needed fresh air supply is mainly determined by the requirement to meet the particle limits (in occupational health this is a limit for elemental carbon in many countries). The use of particle filters for construction machines used in tunnels already led to a significant reduction of particle concentrations. Another example for an occupational exposition is welding, where high concentrations of ultrafine particles are emitted (Lorenzo et al. 2008). Well designed ventilation systems are crucial to protect workers. Otherwise they have to wear masks, which is very inconvenient. A frequently discussed topic is emission of engineered nanoparticles, which are meanwhile used for many applications. This is discussed for example by Maynard and Kuempel (2005).

1 The Origin and Production of Nanoparticles in Environment and Industry

13

7%

a

27% 23%

Exhaust gasonline Exhaust diesel non exhaust gasoline

43%

non exhaust diesel

7%

b

27% 23%

Exhaust gasoline Exhaust diesel

43%

non exhaust gasoline non exhaust diesel

Fig. 1.11 Contribution of exhaust and non-exhaust emissions to traffic emissions (PM10) of diesel- and gasoline passenger cars

Measurements we recently performed at a plant where nanoparticle containing plastics are produced showed high mass concentrations up to 2 mg/m2, but no increased concentration of nanoparticles. Exposition of these particles may not only

14

H. Burtscher

occur during production, but also or perhaps mainly when applying them, for example when using sprays containing nanoparticles.

1.7

Nanoparticles in Ambient Air

In a recent study Imhof (2007) investigated the contribution of traffic to PM10 and ultrafine particles in the city of Zu¨rich. He compared data from a location close to a road and an urban background site. The results clearly show that number concentration and ultrafine particle mass are dominated by local traffic, whereas PM10 is due to more remote sources. Imhof also estimated the effect of particle exhaust filters on ultrafine particles and PM10. Figure 1.12 shows the result of this calculation. If all diesel cars (trucks and passenger cars) are equipped with filters, PM10 can be reduced by 30%, the fraction between 50 and 300 nm by 94%. The contribution by trucks is clearly higher than by passenger cars. A second important source of ambient particles is wood combustion. Glasius et al. (2006) performed a study in Denmark indicating that the contribution of wood combustion to PM2.5 is comparable to that from traffic. Investigations by optical absorption measurements, 14C analysis and analysis of organic material by Aerosol Mass Spectrometry show a dominant contribution of wood combustion in a Swiss alpine valley in wintertime and of traffic in Zu¨rich and near a highway (see Sandradewi et al. 2008 and references therein).

100

100

100

50–300nm 86

90

81

PM10

rel. concentration (%)

80 67

66

70 60 50 40 40 30 20

6

10 0

today

DPF for trucks

DPF for diesel passenger cars

DPF for all

Fig. 1.12 Estimated concentration of ultrafine particles and PM10, if vehicles (passenger cars, trucks) were equipped with diesel particle filters (DPF), normalized to the concentration without filters

1 The Origin and Production of Nanoparticles in Environment and Industry

1.8

15

Conclusions

Combustion is an important source of ultrafine particles in the atmosphere. Important sources are diesel engines and wood combustion in small units (stoves, domestic heating). Diesel engines emissions stem from traffic, but a significant part also is emitted by machines used for construction, agriculture and forestry. The contributions of larger plants for heating purposes, electricity production, or incineration are not significant if they are equipped with state of the art exhaust gas cleaning systems. The above mentioned emissions for cars are valid for well maintained vehicles. However, there is some evidence that the contribution to ambient air concentration is dominated by a few very high emitters (super polluters). This may be very old cars but the high emissions may also be due to a malfunction of the engine. Measurements by Kurniawan and Schmidt-Ott (2006) indicate that 5% of high emitters are responsible for more than 40% of the emissions. Not much is known up to now about the significance of engineered nanoparticles to ambient air particle concentrations. However, this is a current topic of concern. Most industrial emissions of particulate material do not belong to the ultrafine fraction.

References Andersson J, Giechaskiel B, Mun˜oz-Bueno R, Sandbach E, Dilara P (2007) Particle Measurement Programme (PMP) Light-duty Inter-laboratory Correlation Exercise (ILCE_LD) Final Report, document No. GRPE-54-08-Rev.1, European Commission, Joint Research Center Aufdenblatten S, Scha¨nzlin K, Bertola A, Mohr M, Przybilla K, Lutz T (2002) Charakterisierung der Partikelemission von modernen Verbrennungsmotoren. MTZ Motortechnische Zeitschrift 63(11):962–968 Brown DM, Wilson MR, MacNee W, Stone V, Donaldson K (2001) Size-dependent proinflammatory effects of ultrafine polystyrene particles: a role for surface area and oxidative stress in the enhanced activity of ultrafines. Toxicol Appl Pharmacol 175:191–199 Burtscher H (2000) Characterization of ultrafine particle emissions from combustion systems. SAE Technical Paper Series 2000-01-1997 Burtscher H, Zu¨rcher M, Kasper A, Brunner M (2002) Efficiency of flue gas cleaning in waste incineration for submicron particles. In: Mayer A (ed) Proceedings of the international ETH conference on nanoparticle measurement, 6–8 August 2001, BUWAL, 2002. Contrib Burtscher H (2005) Physical characterization of particulate emissions from diesel engines – a review. J Aerosol Sci 36:896–932 Burtscher H, Loretz S, Keller A, Mayer A, Kasper M, Artley R, Strasser R, Czerwinski J (2008) Nanoparticle filtration for vehicle cabins. SAE Technical Papers 2008-01-0827 Czerwinski J, Comte P, Reutimann F, Mayer A (2006) Influencing (nano)particle emissions of 2-stroke scooters. Int J Automotive Technol 7:237–244 Du¨nnebeil F, Lambrecht U, Kessler C(2007) Zuku¨nftige Entwicklung der NO2-Emissionen des Verkehrs und deren Auswirkung auf die NO2-Luftbelastung in Sta¨dten in Baden-Wu¨rttemberg. IFEU – Institut fu¨r Energie- und Umweltforschung Heidelberg GmbH, Commissioned by the Ministry of Environment Baden-Wu¨rttemberg, Stuttgart

16

H. Burtscher

EKL (2007) Eidgeno¨ssische Kommission fu¨r Luftreinhaltung (EKL), Feinstaub in der Schweit. Status Bericht der Eidg. Kommission fu¨r Luftreinhaltung, Bern, www.umwelt-schweiz.ch/div5013-d Gehrig R, Hill M, Lienemann P, Zwicky Ch, Bukowiecki N, Weingartner E, Baltensperger U, Buchmann B (2007) Contribution of railway traffic to local PM10 concentrations in Switzerland. Atmos Environ 41:923–933 Glasius M, Ketzel M, Wahlin P, Jensen B, Mønster J, Berkowicz R, Palmgren F (2006) Impact of wood combustion on particle levels in a residential area in Denmark. Atmos Environ 40:7115–7124 Harris S, Maricq M (2002) The role of fragmentation in defining the signature size distribution of diesel soot. J Aerosol Sci 33:935–942 Imhof D (2007) Nanopartikel am Strassenrand Das Potenzial des Partikelfilters am Beispiel einer dicht befahrenen Strasse. www.akpf.org/pub/2007_imhof.pdf Johansson LS, Leckner B, Gustavsson L, Cooper D, Tullin C, Potter A (2004) Emission characteristics of modern and old-type residential boilers fired with wood logs and wood pellets. Atmos Environ 38:4183–4195 Kasper A, Aufdenblatten S, Forss A, Mohr M, Burtscher H (2007) Particulate emissions from a low-speed marine diesel engine. Aerosol Sci Technol 41:24–32 Kasper A, Burtscher H, Johnson JP, Kittelson DB, Watts WF, Baltensperger U, Weingartner E (2005) Particle emissions from SI-engines during steady state and transient operation conditions. SAE Paper 2005-01-3136 Keller M, Zbinden R (2004) Luftschadstoff-Emissionen des Strassenverkehrs 1980-2030. Swiss Federal Office for the Environment, Report 355. www.bafu.admin.ch/php/modules/shop/files/ pdf/phpb75BgA.pdf Kittelson DB (1998) Engines and nanoparticles: a review. J Aerosol Sci 29:575–588 Klippel N, Nussbaumer Th (2007) Wirkung von Verbrennungspartikeln, Vergleich der Umweltrelevanz von Holzfeuerungeun und Dieslemotoren. www.energieforschung.ch, ISBN 3-908705-16-9 Kurniawan A, Schmidt-Ott A (2006) Monitoring the soot emissions of passing cars. Environ Sci Technol 40:1011–1915 Lighty JS, Veranth JM, Sarofim AF (2000) Combustion aerosols: factors governing their size and composition and implications to human health. J Air Waste Manage Assoc 50:1565–1618 Lorenzo R, Steinle P, Grobety B, Ka¨gi R (2008) Extending the assessment procedures for exposure situations at working environents to ultrafine particles. Subm to J Occup Environ Hyg Mayer A, Kasper M, Mosimann Th, Legerer F, Czerwinski J, Emmenegger L, Mohn J, Ulrich A, Kirchen P (2007) Nanoparticle-Emission of EURO 4 and EURO 5 HDV Compared to EURO 3 With and Without DPF. SAE 2007-01-1112 Maynard AD, Kuempel ED (2005) Airborne nanostructured particles and occupational health. J Nanoparticle Res 7:587–614 Nussbaumer Th (2004) Aerosols from biomass combustion, particle formation, relevance on air quality, and measures for particle reduction, Aerosol Symposium, Nordic Society for Aerosol Research, Stockholm 11.11.2004–12.11.2004 Nussbaumer Th, Czasch C, Klippel N, Johansson L, Tullin C (2008) Particulate emissions from biomass combustion in IEA countries survey on measurements and emission factors. International Energy Agency (IEA) Bioenergy Task 32 Swiss Federal Office of Energy (SFOE) Download www.ieabcc.nl, ISBN 3-908705-18-5 Oberdo¨rster G, Sharp Z, Atudorei V, Elder A, Gelein R, Kreyling W, Cox C (2004) Translocation of inhaled ultrafine particles to the brain. Inhal Toxicol 16:437–445 Robinson AL, Donahue NM, Shrivastava MK, Weitkamp EA, Sage AM, Grieshop A-P, Lane TE, Pierce JR, Pandis SN (2007) Rethinking organic aerosols: semivolatile emissions and photochemical aging. Sience 315:1259–1262 Sakurai H, Tobias HJ, Park K, Zarling D, Docherty KS, Kittelson DB, McMurry PH, Ziemann PJ (2003) On-line measurements of diesel nanoparticle composition and volatility. Atmos Environ 37:1199–1210

1 The Origin and Production of Nanoparticles in Environment and Industry

17

Sandradewi J, Prevot ASH, Weingartner E, Schmidhauser R, Gysel M, Baltensperger U (2008) A study of wood burning and traffic aerosols in an Alpine valley using a multi-wavelength Aethalometer. Atmos Environ 42:101–112 Schmatloch V, Rauch S (2005) Design and characterisation of an electrostatic precipitator for small heating appliances. J Electrostatics 63:85–100 Srivastava A, Gupta S, Jain VK (2008) Source apportionment of total suspended particulate matter in coarse and fine size ranges over Delhi. Aerosol Air Qual Res 8:188–200 Somers CM, McCarry BE, Malek F, Quinn JS (2004) Reduction of particulate air pollution lowers the rist of heritable mutations in mice. Science 304:1008–1010 Su DS, Mu¨ller JO, Jentoft RE, Rothe D, Jacob E, Schlo¨gl R (2004) Fullerene-like soot from EuroIV diesel engine: consequences for catalytic automotive pollution control. Top Catalysis 30(31):241–245 UBA (2005) Hintergrundpapier zum Thema Staub/Feinstaub (PM), Umweltbundesamt Berlin, p 3 Vouitsis E, Ntziachristos L, Samaras Z (2008) Theoretical investigation of the nucleation mode formation downstream of diesel aftertreatment devices. Aerosol Air Qual Res 8:37–53 WHO (2006) World Health Organization (WHO): Air quality guidelines for particulate matter, ozone, nitrogen, dioxide and sulphur dioxide, Global Update 2005, 2006 Wichmann HE, Peters A (2000) Epidemiological evidence of the effects of ultrafine particle exposure. Philos Trans R Soc Lond Ser A 358:2751–2768 Wieser U, Gaegauf Ch, Macquat Y (2001) Partikelemissionen aus Holzfeuerungen, Untersuchung der Partikelfrachten in Holzfeuerungen unter Praxisbedingungen. Report, Centre of Appropriate Technology, Langenbruck Switzerland, www.oekoze

Chapter 2

Characterization of Combustion and Engine Exhaust Particles M. Matti Maricq

2.1

Introduction

Combustion of carbon based fuels, whether fossil or renewable, represents the predominant source for society’s energy and heating needs. In some cases, for example stoichiometric premixed combustion, this process is rather clean, resulting in only parts per million quantities of emissions other than carbon dioxide and water vapor. In others it leads to a multitude of incomplete combustion byproducts ranging from small molecules to fractal-like soot structures hundreds of nanometers in size. The major combustion products are natural constituents of the atmosphere and this is where they have historically been emitted, with concerns over carbon dioxide’s role in climate change arising relatively recently (IPCC 2007). Emissions of the minor combustion byproducts, including hydrocarbons (HC), carbon monoxide, nitrogen oxides (NOx) and particulate matter (PM), have been regulated to an increasing extent over many decades owing to their potential adverse health impacts. Assessing CO and NO2’s effects on health is relatively straightforward. While some HCs are designated air toxics (United States Environmental Protection Agency 2007), hydrocarbons and NOx are regulated largely based on their role in enhancing tropospheric photochemical ozone formation. PM, though, is a widely diverse material over which concern arises primarily through epidemiological associations between ambient PM concentrations and incidence of adverse health endpoints (Health Effects Institute 2003). Specific mechanisms mediating possible health pathways are presently under investigation. Presumably these depend on the particle’s composition and morphology, but the precise relationship remains unclear. Combustion PM is also under scrutiny with respect to climate change, directly because it absorbs sunlight, and indirectly because it influences cloud formation and reduces the surface albedo of ice and snow (Ackerman et al. 2000). M. Matti Maricq Research and Advanced Engineering, Ford Motor Company, MD 3179, P.O. Box 2053, Dearborn, MI, 48121, USA e-mail: [email protected]

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_2, # Springer ScienceþBusiness Media B.V. 2010

19

20

M. Matti Maricq

To help guide and interpret health studies and develop a deeper understanding of climate change, it is important to characterize physically and chemically the particulate constituents of atmospheric aerosols. The present paper does this for a subset of PM sources, namely combustion aerosols and engine exhaust particles. Combustion PM can be classified broadly into three categories: particles generated (1) in the flame/engine, (2) by physical gas to particle conversion as the exhaust cools, and (3) from atmospheric chemical transformations of gaseous emissions. The first two categories are those commonly associated with combustion PM and that fall under PM emissions regulations. Recent research examines the interplay between freshly nucleated emissions and secondary organic aerosols, underscoring the difficulties in relating primary emissions from combustion sources to their eventual atmospheric PM burden (Robinson et al. 2007). Figure 2.1 provides a schematic pictorial of the life of a particle from its birth in the engine cylinder to the exhaust exiting the tailpipe (Kittelson et al. 2006). Engine and flame generated particles originate from fuel rich regions within the combustion mixture. Here, high temperature oxidation converts relatively stable fuel molecules into reactive intermediates, but there is insufficient oxygen to convert the fuel entirely into the CO2 and H2O end-products. At sufficiently high equivalence ratios these intermediates form polyaromatic hydrocarbons (PAH) that dimerize and grow by surface addition to primary particles, which then agglomerate into the fractal-like aggregates commonly associated with soot (Frenklach and Wang 1991; Appel et al. 2001; D’Anna et al. 2001). In many combustion processes, for example diesel engines, soot formation is followed by its oxidation. Particles

Particles formed by Diesel combustion carry a strong bipolar charge Carbon formation/oxidation t = 2 ms, p = 150 atm., T = 2500 K

There is potential to form solid nanoparticles here if the ratio of ash to carbon is high.

Formation

Ash Condensation t = 10 ms, p = 20 atm., T = 1500 K

Increasing Time

Exit Tailpipe t = 0.5 s, p = 1 atm., T = 600 K

There is where most of the volatile nanoparticles emitted by engines usually form.

Sulfate/SOF Nucleation and Growth t = 0.6 s, p = 1 atm., D = 10, T = 330 K

Atmospheric Aging Exposure Fresh Aerosol over Roadway-Inhalation/Aging t = 2 s, p = 1 atm., D = 1000, T = 300 K

Fig. 2.1 Particle formation – 2 s in the life of engine exhaust particles (reprinted from Kittelson et al. 2006 with permission from Elsevier)

2 Characterization of Combustion and Engine Exhaust Particles

21

formed in locally rich regions of the flame subsequently pass into lean regions, where the temperature remains sufficiently high to burn up large fractions (up to >90%) of the soot (Dec and Kelly-Zion 2000). The above scenario applies to the majority of carbon based fuels ranging from methane to coal. The main difference between pure molecular fuels and practical fuels, such as diesel fuel and coal, is the presence of impurities including sulfur and metals. The latter produce metal oxides, which when their concentrations are low condense onto soot surfaces (Miller et al. 2007). But under high metal to carbon ratios, as might occur with fuel-born additives, nucleation of metal nanoparticles can occur during combustion (Miller et al 2007; Kasper et al. 1999). As a result of extensive agglomeration and oxidation, soot/ash particles emerge from many combustion processes with a ubiquitous lognormal size distribution (Harris and Maricq 2001). Due to their high temperature origin these particles are nonvolatile and, therefore, relatively robust to sampling methodology, with the main concerns being losses and coagulation. Sampling plays a much more important role with respect to the semivolatile components of combustion aerosols. The precursors of this PM, including unburned fuel, lube oil, and partial combustion products, exit the combustion source as gases and in some cases are further transformed, for example the conversion of SO2 to SO3 over a diesel oxidation catalyst (DOC). As the exhaust cools some substances approach their saturation vapor pressures, whereupon two possibilities ensue: they condense onto existing soot particles or nucleate to form new particles. This can occur already in the exhaust system, or as the exhaust exits into the atmosphere (or dilution air of a sampling system). The degree to which each path is followed depends sensitively on a host of conditions such as temperature, dilution rate, humidity, soot concentration, fuel composition, and the presence of aftertreatment devices (Abdul-Khalek et al. 1998, 1999; Shi and Harrision 1999; Vaaraslahti et al. 2004). Condensation is the principal pathway in situations where soot dominates the PM. This is the case in conventional diesel engine exhaust, where condensed material often only modestly affects soot characteristics. As the ratio of condensable material to soot increases, nucleation predominates. At moderate levels this produces a bimodal particle size distribution, but at high levels it can alter the nature of the PM to one dominated by liquid heavy hydrocarbon/sulfuric acid droplets. Sampling conditions, too, can affect the propensity for nucleation and thereby alter the appearance of PM. In the following sections we will explore more deeply sampling methods and their influence on PM measurements, the methods used to characterize combustion particles, and the picture this provides of the soot and nuclei modes of particles from flames and internal combustion engines.

2.2

How Sampling Impacts PM Measurement

Ordinarily to study aerosol processes, for example how soot evolves in a flame, the goal is to probe non-invasively, or to use sampling methods that preserve the nascent aerosol characteristics. But how is this possible in the case of combustion

22

M. Matti Maricq

emissions, when some particles are produced in the flame or engine, whereas others may only form later? Approaches such as light scattering and laser induced incandescence (LII) (Dec et al. 1991; Snelling et al. 1999) probe particles as they exist in a flame, but they miss semivolatile PM that subsequently condenses or nucleates as the exhaust cools. Combustion PM characterization and sampling methods are, therefore, inexorably linked. But this is only one sampling issue. Here, at least, the emissions are combustion produced; the complexity lies in interpreting the gas–particle partitioning. A second issue concerns the possibility, depending on design, that the sampling system itself acts as a particle source, generating artifacts that cannot always be readily distinguished from combustion source PM.

2.2.1

Conventional Approach – Dilution Tunnels

The standard approach to motor vehicle PM emissions measurement collects test vehicle exhaust into a dilution tunnel (Hildemann et al. 1989; Code of Federal Regulations 2008). Figure 2.2 illustrates two such tunnels in a chassis dynamometer facility. Historically one is reserved for diesel and the other for gasoline vehicles based on their disparate emissions; however, the advent of diesel particulate filters (DPF) has blurred this distinction. Vehicle exhaust is combined in the tunnel with filtered, temperature and humidity controlled, dilution air (38 C and 9 C dew point) at a constant total volume. This implies a dilution ratio that varies during transient emissions tests, such as the US Federal Test Procedure (FTP) or New European Drive Cycle (NEDC). But it confers the advantage that emission rates per distance traveled are directly computed from the tunnel flow rate and PM concentration. The weakness of this approach lays not so much with the dilution tunnel as with the need to convey the exhaust from tailpipe to tunnel. This typically occurs

Fig. 2.2 Dilution tunnels used for PM measurement in a chassis dynamometer test cell

2 Characterization of Combustion and Engine Exhaust Particles

23

Fig. 2.3 Comparison of port fuel injection gasoline vehicle PM size distributions measured at the tailpipe versus from the dilution tunnel. The large nanoparticle peak originates from the sampling system transfer hose

through a heated corrugated stainless steel hose that can be meters in length and, therefore, introduces delays of a second or longer prior to dilution. Engine PM does not remain unchanged during this transit; particles are lost by diffusion and thermophoresis, and coagulation reduces their numbers while increasing their sizes. Moreover, hot vehicle exhaust can desorb materials deposited on the transfer hose walls, which subsequently nucleate in the dilution tunnel to generate an aerosol artifact difficult to distinguish from engine PM (Maricq et al 1999). This is illustrated in Fig. 2.3 by gasoline vehicle test data recorded after sustained driving at 100 km/h. The dilution tunnel measurement exhibits an intense nanoparticle peak, which is absent in the sample drawn directly via ejector diluter (at 200 C) from the tailpipe. One can argue, as discussed below, that the nanoparticles are real; ejector sampling simply suppresses their nucleation. However, this peak appears only after many minutes at the 100 km/h speed, and then dissipates over many minutes after the vehicle is returned to low speed. This lag correlates with transfer hose temperature and suggests that these nanoparticles arise from heat release of stored precursors, most likely heavy hydrocarbons.

2.2.2

Direct Sampling Methods

Redesigning the dilution tunnel to enable dilution at, or near, the tailpipe can alleviate artifact particle formation (Maricq et al. 2003). Alternatively, a number of options exist to sample combustion aerosols directly at the source including the rotating disc diluter (Matter Engineering) and ejector pump diluters (Dekati Ltd.). These methods permit control of sampling conditions to help avoid biases and artifacts, for example by the use of heated lines and heated first stage dilution to reduce thermophoretic losses and suppress storage–release mechanisms. Sampling is especially critical directly from a flame or engine cylinder where temperatures of ~2000 K and particle concentrations on the order of 1010 cm3 require immediate

24

M. Matti Maricq

dilution to quench combustion chemistry and particle agglomeration (Zhao et al. 2003). But sampling conditions can impact also the combustion aerosol itself, hence the argument with respect to Fig. 2.3 that one cannot distinguish without additional data whether direct sampling via the heated ejector pump prevents an artifact or suppresses nucleation of material from the engine exhaust. While it may complicate a precise definition of combustion PM, the ability to control sampling conditions can aid in aerosol characterization. The principal control parameters are temperature, residence time, and dilution factor (Abdul-Khalek et al. 1999). The EU Particulates Project, for example, analyzed motor vehicle exhaust via so called “wet” and “dry” branches to distinguish semivolatile and nonvolatile PM components (Ntziachristos et al. 2004). Sampling is done through a “perforated tube” diluter, which permits the use of cold dilution air while avoiding thermophoretic deposition. Cold dilution air enhances nucleation and condensation, hence the label “wet” branch. Passing this aerosol through a thermo denuder (Burtscher et al 2001), a heater followed by activated charcoal adsorbent, creates the “dry” branch, which yields the nonvolatile component of the combustion PM. Comparison to the “wet” branch then reveals the semivolatile component. In essence this accomplishes via sampling what thermal elemental carbon/organic carbon (EC/OC) analysis provides for filter collected PM (Chow et al. 2001). Furthermore, the “dry” branch forms the basis for the proposed European Particle Measurement Program (PMP) number based particle emissions standard (Kasper 2004; Giechaskiel et al. 2008). The sampling system design employs a heated residence chamber to evaporate semivolatile PM, and relies on dilution to prevent re-nucleation.

2.3

Particles Formed by Combustion

Bearing in mind sampling’s potential influences, we now turn to examine the nature of combustion PM. Recent reviews of physical and chemical characterization of diesel PM provide a detailed picture of current progress in this field (Burtscher 2005; Maricq 2007). The present discussion focuses on four qualities: morphology, density, volatility, and electrical charge, which are useful to distinguish the two modes prevalent in both flame and engine generated particles. These modes are illustrated in Fig. 2.4 in the case of a light duty diesel vehicle run over the transient FTP drive cycle. The peaks in time arise primarily from the increased exhaust flow during vehicle acceleration. As demonstrated below, the 20–200 nm mode can be associated with soot, that is, particles primarily formed in the engine cylinder or flame. The 2–20 nm mode remains more inscrutable; in some cases it is born from combustion chemistry and in others subsequently produced by nucleation. The present section examines the morphology and effective density of the so-called soot mode, whereas the next section investigates the nuclei mode via its volatility and electrical charge.

2 Characterization of Combustion and Engine Exhaust Particles

25

Fig. 2.4 Time resolved bimodal particle size distributions recorded from the exhaust of a light duty diesel vehicle (DOC and low sulfur fuel) via a dilution tunnel. Mobility equivalent diameter measured by Engine Exhaust Particle Sizer (TSI Inc.)

2.3.1

Size and Morphology

Transmission electron microscopy (TEM) has been much applied to the study of soot particle physical characteristics, usually via thermophoretic sampling directly from the flame (see for example (Dobbins 2007) and references therein). Here we examine soot that is first size selected with a differential mobility analyzer (DMA). Figure 2.5 displays TEM images of rich premixed ethylene flame soot particles at four distinct sizes ranging from 25 to 200 nm in mobility diameter, and produced with equivalence ratios from F ¼ 2.05–2.45. The latter two of this series represent particles that are larger than the vast majority actually present in these premixed flames, even at heights well past the flame front (see Fig. 2.9a for the flame soot size distribution at 20 mm above the burner). The 100 and 200 nm particles are grown by allowing soot sampled from the flame, quenched but not highly diluted (~5  109 particles/cm3), to coagulate for 3 s in a residence chamber at near room temperature. One consequence of electrical mobility sizing is that multiply charge particles appear smaller to the DMA than they actually are by a factor of approximately n½, where n represents the number of charges. Figure 2.5b displays one such outsized particle, presumably transmitted by the DMA due to multiple charges. Aside from this charge dependence, the chosen mobility diameter in each case matches closely, within ~10%, the projected area equivalent diameter of the corresponding TEM images. The major feature of Fig. 2.5 is the striking change in morphology that accompanies increasing soot size. At 25 nm the particles are at best disfigured spheres, with primary particle structure barely discernable due to filling in by surface growth. By 50 nm primary particle structure is clearly visible, but the aggregates

26

M. Matti Maricq

Fig. 2.5 TEM images of mobility size selected soot particles formed by a rich premixed ethylene flame. Soot from the flame was allowed to coagulate for up to 3 s in a residence chamber to grow the aggregates. The odd oversize particle in panel B is a doubly charged particle with the same electrical mobility as the 50 nm particles

remain compact, with growth dominated by aggregate–primary particle collisions. The 100 and 200 nm particles increasingly display the fractal-like structure commonly associated with soot, and arising from aggregate–aggregate collisions. Figure 2.6 displays nominally 60 nm mobility selected particles collected from the exhaust of an idling diesel vehicle; but clearly larger, multiply charged, particles are also evident. Compared to their flame counterparts these soot particles appear less fractal-like. They possibly contain condensed hydrocarbons, but morphology changes that might indicate evaporation were not observed under the TEM. Condensed material is not unexpected here as idle and low load are engine operating conditions often associated with a higher semivolatile to soot emissions ratio owing to lube oil or fugitive heavy end fuel components.

2.3.2

Effective Density

Another manifestation of particle morphology is revealed through measurement of its density. When particles assume complex shape, density takes on two interpretations: one is the material density and the other is the concept of effective density

2 Characterization of Combustion and Engine Exhaust Particles

27

Fig. 2.6 TEM images of 60 nm mobility size selected soot particles from a light duty diesel vehicle run at idle. Larger particles transmitted by the DMA on account of their multiple charge are also displayed

(Kelly and McMurry 1992). In the first case soot particle density remains constant with size and shape except if changes in composition take place, for example hydrocarbons condensed onto diesel soot particles were found to lower their density from 1.8 to 1.3 g/cm3 (Park et al. 2004). In contrast, effective density, defined here as particle mass divided by its mobility equivalent volume, decreases with increasing particle size owing to the rising fraction of voids caused by aggregation. As per definition, effective density can be determined by simultaneous measurement of a particle’s mobility diameter and mass. One procedure (Park et al. 2003) accomplishes this directly via a tandem DMA – aerosol particle mass analyzer (APM). Another approach is to select particles of known mobility diameter and sequentially, or in parallel, measure their aerodynamic diameter (Schleicher et al. 1995; Maricq and Xu 2004; Virtanen et al. 2004). The latter quantity is connected with the particle’s settling time and is therefore dependent on its mass. Effective density is related to the two particle equivalent diameters via re ðdm ÞCc ðdm Þdm2 ¼ ra Cc ðda Þda2

(2.1)

where ra is assigned unit density and Cc is the Cunningham slip correction (Hinds 1999). Figure 2.7a compares aerodynamic and mobility diameter measurements of mobility selected particles at dm ¼ 141 nm. The line marked “DMA” depicts the mobility diameter recorded using a second DMA. In decreasing order, the three peaks correspond to 141 nm particles having n ¼ 1, 2, and 3 charges as a result of passing through a 210Po neutralizer prior to the second DMA. In contrast, the aerodynamic diameter (histogram), as measured by electrical low pressure impactor (ELPI), is substantially smaller than the mobility diameter. By Eq. 2.1 this implies an “effective” density below 1 g/cm3, which is considerably lower than the commonly accepted 1.8 g/cm3 material density of soot (Park et al. 2004).

28

M. Matti Maricq

Fig. 2.7 Effective density measurement of flame and diesel soot particles. Panel A: Aerodynamic (ELPI) versus mobility (DMA) diameter of 141 nm particles. Panel B: Flame soot density from low and moderately sooting flames. Panel C: Diesel soot density at various vehicle speeds

Figures 2.7b and c display the mobility diameter dependence of soot effective density for flame and diesel soot, respectively. These data follow closely the power law expression re ¼ r0 ðdm =d0 ÞðDf 3Þ

(2.2)

expected for fractal-like particles, except at small diameter where the aggregate primary particle number approaches unity. Here d0 and r0 are the primary particle diameter and density, and Df is the fractal dimension. Particles from a moderately sooting flame, F ¼ 2.4, sampled from 20 mm above the burner exhibit a fractal dimension of Df ¼ 1.9, within the range found by light scattering and TEM (Dobbins 2007). Lowering the height to 10 mm and the equivalence ratio to F ¼ 2.0 produces a younger soot. The fractal dimension increases to Df ¼ 2.3, reflecting the reduced time for aggregation and consequently more spherically shaped particles. That these particles also exhibit a lower overall effective density suggests a decrease in their C/H ratio. Light duty diesel vehicle soot effective density also follows a Df ¼ 2.3 dependence, with only little variation between idle and 112 km/h operation, similar to what is found for heavy duty diesels by Park et al. 2003. Its higher fractal dimension relative to the more aged flame soot may originate from semivolatile hydrocarbon condensation filling voids in the fractal-like structure of dry soot (Ristima¨ki and Keskinen 2006), or from the partial soot oxidation that occurs in the latter stages of diesel combustion (Dec and Kelly-Zion 2000) leaving more compact structures. Condensation has little effect on the small particles, which already have a re ffi 1 g/cm3 close to that of hydrocarbons, but it would tend to increase the effective density of larger aggregates, where re falls well below 1 g/cm3. One practical application of effective density is that it enables calculation of PM mass from particle size distribution data, N(dm), via 1 ð p ddm re ðdm Þdm3 Nðdm Þ (2.3) M¼ 6 0

The soot (solid particle) mode of diesel PM exhibits a characteristic lognormal soot mode, with a width (sg ¼ 1.75) nearly independent of engine operation and fuel

2 Characterization of Combustion and Engine Exhaust Particles

29

Fig. 2.8 Particle number versus PM mass emission rates from various light duty engine technologies

choice (Harris and Maricq 2001) (in some cases high hydrocarbon/sulfate emissions can significantly broaden this or yield bimodal distributions). This leaves two free parameters to describe diesel soot, namely particle number and geometric mean diameter. Mean size varies with engine speed/load, level of exhaust gas recirculation, and fuel composition, but typically remains within the 50–100 nm range. Soot emissions from spark ignition engines, port as well as direct fuel injection, deviate somewhat from this characteristic size distribution, but not extensively. As a result soot mode mass and number emissions are correlated, as demonstrated in Fig. 2.8 (light duty vehicles). Here PM mass is recorded gravimetrically, so gaseous adsorption artifacts on the filter media are responsible for some of the data scatter at low emissions levels. Likewise, not applying strict PMP protocols to nuclei particle removal may contribute to some of the scatter in particle number. Nevertheless, a quite good correlation between PM mass and number emissions encompasses a variety of engine/aftertreatment technologies, the emission levels of which stretch over three orders of magnitude. One implication of this correlation is that the European EU5a number limit of 6x1011 particles/km for diesel vehicles is substantially more stringent than the 4.5 mg/km mass emissions standard.

2.4

Nuclei Mode Characterization

Although different in detail, soot modes from most combustion processes exhibit a qualitative similarity that derives from their high temperature origin. Differences originate principally from variations in ash content and from the extent of semivolatile material that condenses onto the soot. The situation with nucleation is different. As demonstrated below, a mode of 2–20 nm particles can appear either during combustion or as the exhaust subsequently cools. Size alone is insufficient to distinguish these two situations, but the volatility and electrical characteristics of the nuclei mode can help.

30

2.4.1

M. Matti Maricq

Particle Volatility

A variety of methods exists to examine the semivolatile versus solid components of aerosol particles, the common element of which involves heating the PM. The most widespread practice, EC/OC analysis, heats particulate matter collected onto quartz filters first in an inert atmosphere, and subsequently in an oxidizing atmosphere (Cadle et al. 1980; Huntzicker et al. 1982). The first stage evaporates semivolatile material, the carbon in which is then quantified by conversion to CO2 (or CH4). The second stage directly oxidizes the remaining nonvolatile carbon to CO2. Considerable efforts have been paid to ascertain the extent of pyrolysis during the first stage and to correct the resulting overestimate of EC (Chow et al. 2001). Unless separate provisions are made, non-carbonaceous constituents, such as sulfate and ash are not detected. Standard EC/OC analysis does not distinguish between different PM modes. A straightforward way to study the size dependence of particle volatility is via parallel “wet” and “dry” branch measurements (Ntziachristos et al. 2004). One limitation of this procedure is that it does not examine the EC/OC composition of individual particles; it reveals only the net contribution of semivolatile components to the overall size distribution. Another limitation is the small effect that the evaporation of condensed material has on a fractal-like particle’s mobility diameter, unless this represents a significant fraction of the particle (Ristima¨ki and Keskinen 2006). Alternatively, detailed examination of particle volatility is possible by a tandem DMA – evaporation tube – DMA approach (Sakurai et al. 2003; Kwon et al. 2003; Wehner et al. 2004). As in the measurement of effective density, the first DMA selects a narrow range of particle size for study. The selected particles are heated to a set temperature, and what remains of them after evaporation and desorption of the semivolatile components is recorded by a second DMA. Figure 2.9 illustrates the application of each method to the case of a sooting premixed flame. The parallel approach in Fig. 2.9a reveals that both the nucleation and accumulation modes appear to be nonvolatile, at least up to 330 C. The small reduction in particle concentration at 330 C, the “dry” branch, occurs because of thermophoretic and diffusion losses in the evaporation tube, which increase as temperature increases and particle size decreases. Figure 2.9b presents data taken with the tandem method that provide a closer look at the effect of heat on particles in the nuclei mode. As the temperature in the heat pipe increases from 25 C to 620 C, the mobility diameter decreases from 5.8 to 3.5 nm, and the originally monodisperse peak broadens. At this small size accounting for losses makes it difficult to determine the exact fraction of particles that survive to 620 C, but clearly a significant fraction does. This nuclei mode appears in relatively cooler flames as a result of new particle formation, which can continue well beyond the flame front (Maricq 2006a). These incipient soot particles are reported to have a C/H ratio lower than that of the soot aggregates that comprise the accumulation mode of larger particles (Sgro et al. 2007), an interpretation consistent with the decreased effective density observed in Fig. 2.7 for particles in the more lightly sooting flame. Thus, a

2 Characterization of Combustion and Engine Exhaust Particles

31

Fig. 2.9 Ethylene flame soot volatility. Panel A: Soot size distributions in “wet” branch (40 C) versus “dry” branch (330 C). Panel B: Tandem DMA volatility data for 5.8 nm particles from the nuclei mode

Fig. 2.10 Diesel exhaust PM volatility. Panel A: Semivolatile nuclei mode. Panel B: Nonvolatile nuclei mode. Measurements are made by passing non-size selected exhaust particles through heat pipe and then measuring their size distribution

plausible explanation of why the nuclei mode particles examined in Fig. 2.9b shrink, is that heating the particles increases their carbonization and leads towards a more graphitic structure (Dobbins 2002). Interestingly, nuclei mode particles in engine exhaust exhibit two volatility behaviors depending on engine operation, and perhaps other factors. The predominant case is the one of semivolatile particles illustrated in Fig. 2.10a. Here, heating the aerosol to above 200 C removes the nuclei mode particles, leaving behind the lognormal soot mode. The interpretation is that the former mode arises from semivolatile exhaust components as the exhaust cools, either from dilution into

32

M. Matti Maricq

a sampling system or emission into the atmosphere. This conclusion is supported by particle mass spectrometry, which reveals a composition primarily (>95%) of lube oil derived hydrocarbons in heavy duty diesel exhaust (Tobias et al. 2001), and of sulfate and heavy hydrocarbons for light duty diesel vehicles (Schneider et al. 2005). Under some conditions, for example at idle, diesel vehicles can also produce nonvolatile nuclei mode particles (De Filippo and Maricq 2008), as illustrated in Fig. 2.10b by their persistence to 450 C. This light duty vehicle exhaust is sampled upstream of any exhaust aftertreatment devices, and is diluted with hot (200 C) air, which tends to suppress condensation. Thus, here the mode appears almost entirely nonvolatile. Thermophoretic losses in the heat pipe are likely primarily responsible for the decrease the mode’s intensity with increasing temperature. The reduction in mean diameter from 9.5 to 7.5 nm could be due to removal of some condensed material or to restructuring from carbonization. Nonvolatile nuclei mode particles have also been observed as cores of much larger particles in the exhaust of a heavy duty diesel engine designed to meet EURO IV emissions standards without use of a diesel particulate filter. In that case condensed material consisted of semivolatile hydrocarbon emissions (Ro¨nkko¨ et al. 2007).

2.4.2

Particle Electrical Charge

Electrical charge represents another characteristic that is useful in tracing particle origins. Combustion chemistry includes chemiionization reactions, whereby at high temperatures two neutral molecules combine to produce positive and negative ions (Calcotte 1981). The ions rapidly collide with any particles that may be present, at rates enhanced by attractive image charge forces, and deposit their charge. Subsequent aggregation of the soot particles brings their charge into a size dependent Boltzmann distribution (Hinds 1999; Maricq 2006b)  fdm ðzÞ ¼

KE e2 pdm kT

1=2

  KE z2 e2 exp dm kT

(2.4)

where KE ¼ 9.0  109 Nm2/C2 in SI units, e is the charge of an electron, k is Boltzmann’s constant, and T represents temperature. As seen in Fig. 2.11, the soot mode (~20–200 nm) consistently exhibits a substantial fraction (40–60%) of charged particles, essentially evenly balanced between positive and negative charges. In the flame case, a temperature of ~1700 K describes the distribution of charges, which matches the flame temperature (Maricq 2006a). However, the characteristic temperature of the diesel soot electric charge is 800–1200 K, considerably lower than expected from diesel combustion temperatures (~2200 K) (Maricq 2006b). This is explained by the redistribution of charge to a lower temperature by the initially rapid particle aggregation that occurs

2 Characterization of Combustion and Engine Exhaust Particles

33

Fig. 2.11 Electrical charge of combustion particles. Panel A: rich premixed flame. Panel B: Light duty diesel vehicle, case 1 – semivolatile nuclei mode. Panel C: Light duty diesel vehicle, case 2 – nonvolatile nuclei mode. Panel D: DPF equipped diesel

as the soot exits the engine cylinder and cools in the exhaust system, but which cannot keep up with the exhaust cooling rate because aggregation drives down the particle collision rate. The nuclei mode again exhibits two contrasting behaviors. In Fig. 2.11a, the nuclei mode of incipient flame particles appears electrically neutral, even though it is formed within the flame. The hypothesis is that by 20 mm above the burner all the ions have already attached to the existing soot particles; hence, none are left to charge these newly born nuclei particles. The same does not appear to be the case for the nonvolatile diesel exhaust nuclei particles in Fig. 2.11c. Here, the fraction of charged particles ranges from ~1% for each polarity at 3 nm to ~10% each at 10 nm, which is consistent with a temperature of ~850 K. This value is similar to what is observed for the diesel soot mode, suggesting that these nuclei particles arise during combustion, and not subsequently in the exhaust system. Only the beginning of the accumulation mode is evident at ~50 nm in Fig. 2.11c, and it is likewise charged. The diesel exhaust nuclei particles in Fig. 2.11b and d are electrically neutral. In the former case, the volatility of this mode is the same as displayed in Fig. 2.10a; namely the particles are removed at temperatures above about 200 C. The conclusion is that this nucleation mode forms from organic compounds and sulfate that surpass their saturation vapor pressures as the exhaust temperature drops. This occurs at exhaust temperatures below ~200 C, at which point there is no mechanism to charge the resulting aerosol; hence, it remains electrically neutral. The situation shown in

34

M. Matti Maricq

Fig. 2.11d is identical, except that these particles are sampled downstream of a diesel particulate filter. The filter reduces the soot mode concentration, as evident from its much lower intensity. Without soot particles as a sink, semivolatile hydrocarbons and sulfate nucleate. At some point condensation onto the newly formed nuclei exceeds new particle nucleation and leads to particle growth, which explains why the mean nucleation mode diameter in panel D is the largest in Fig. 2.11.

2.5

Discussion and Conclusion

The size categories PM10, PM2.5, and ultrafine particles play a central role in the areas of ambient PM measurement, regulations, and health effects. However, with few exceptions, such as fly ash, combustion particles all fall well below the PM2.5 cutoff, and they are ineffectively distinguished by the conventional 100 nm ultrafine demarcation. Instead, they more naturally divide into nucleation and accumulation (soot) modes, the former generally extending from 2 to 20 nm and the latter from 10 to 300 nm. Because these two modes arise from distinct chemical and physical processes, they are not solely differentiated by size, but also by morphology, density, volatility, electrical charge, and chemical composition, although not always in a mutually exclusive manner. Posing questions within this more natural framework of particle modes may improve insight into issues of PM measurement, regulation, and health effects. The current method for engine PM emissions measurement is by dilution tunnel sampling onto filter substrates held at 47  5 C, followed by gravimetric analysis. Present day stack sampling from stationary sources follows EPA Method 5 to collect PM isokinetically onto glass fiber filters held at 120  14 C. As this discriminates against semivolatile particles, Method 202 is applied to collect condensable PM via impingers. Besides these mass based metrics, the European Union is introducing a particle number based emission standard for motor vehicles (GRPE 2007). It specifies hot dilution (150 C) followed by an evaporation tube held at 300 C to accelerate semivolatile particle removal, and stipulates a 23 nm lower cutpoint to communize an otherwise ill-defined limit for what is counted as a particle. Because the mass based methods are operationally defined, they do not provide consistent measures of PM. When soot dominates, this issue may be minor. These particles are solid, and so not much affected by sampling conditions. But these methods will register semivolatile components differently. The PMP number based approach is distinct, not only by the choice of metric, but also in the decision to specifically target solid exhaust particles, and thereby move away from an operational definition. In the absence of significant semivolatile PM mass, the solid particle number count is related to traditional mass measurement via soot size and effective density, a relationship supported by the correlation in Fig. 2.8 for light duty vehicle PM. As emissions regulations tighten, aftertreatment devices and ultralow sulfur fuel will change the exhaust PM character, primarily by a relative increase in semivolatile content. Currently, more thought is needed to identify

2 Characterization of Combustion and Engine Exhaust Particles

35

quantification methods specific to the characteristics of semivolatile combustion PM, and less dependent on operationally based definitions. Particulate matter remains too varied and complex to allow ready classification. But as illustrated here via morphology, effective density, volatility, and electrical charge, investigating their characteristics provides numerous benefits. It can help elucidate their origins, whether in the engine, as nucleation/condensation in the exhaust, or as a sampling artifact. It aids health effects studies, where the fate of respired particles depends on size, whether they are liquid or solid, and presumably on composition. And it provides the basis for designing new PM measurement instrumentation. These new methods may not always conform to the historical definition of PM, but they may suggest new ways to think about combustion aerosols. Acknowledgments The author would like to thank Mike Loos, Adolfo Mauti, Sandip Shah, and Joseph Szente (Ford Motor Co.) for their generous help with the motor vehicle emissions measurements, and Yi Liu (Wayne State University) for his gracious help producing the TEM images.

References Abdul-Khalek I, Kittelson D, Brear F (1999) The influence of dilution conditions on diesel exhaust particle size distribution measurements, SAE Technical Paper 1999-01-1142 Abdul-Khalek IS, Kittelson DB, Brear F (1998) Diesel trap performance: particle size measurements and trends, SAE Technical Paper 982599 Ackerman AS, Toon OB, Stevens DE, Heymsfield AJ, Ramanathan V, Welton EJ (2000) Reduction of tropical cloudiness by soot. Science 288:1042–1047 Appel J, Bockhorn H, Wulkow M (2001) A detailed numerical study of the evolution of soot particle size distributions in laminar premixed flames. Chemosphere 42:635–645 Burtscher H (2005) Physical characterization of particulate emissions from diesel engines: a review. J Aerosol Sci 36:896–932 Burtscher H, Baltensperger U, Bukowiecki N, Cohn P, Hu¨glin C, Mohr M, Matter U, Nyeki S, Schmatloch V, Streit N, Weingartner E (2001) Separation of volatile and non-volatile aerosol fractions by thermodesorption: instrumental development and applications. J Aerosol Sci 32:427–442 Cadle SH, Groblicki PJ, Stroup DP (1980) An automated carbon analyzer for particulate samples. Anal Chem 52:2201–2206 Calcotte HF (1981) Mechanisms of soot nucleation in flames – a critical review. Combust Flame 42:215–242 Chow JC, Watson JG, Crow D, Lowenthal DH, Merrifield T (2001) Comparison of IMPROVE and NIOSH carbon measurements. Aerosol Sci Technol 34:23–34 Code of Federal Regulations (2008) Title 40, Part 86.110-90, http://ecfr.gpoaccess.gov/cgi/t/text/ text-idx?c=ecfr&sid=7bc1209f844f49a99540cf1b44d6633e&rgn=div8&view=text&node=40: 18.0.1.1.2.2.1.13&idno=40 D’Anna A, Violi A, D’Alessio A, Sarofim AF (2001) A reaction pathway for nanoparticle formation in rich premixed flames. Combust Flame 127:1995 Dec JE, Kelly-Zion PL (2000) The effects of injection timing and diluent addition on latecombustion soot burnout in a DI Diesel engine based on simultaneous 2-D imaging of OH and soot, SAE Technical Paper 2000-01-0238

36

M. Matti Maricq

Dec JE, zur Loye AO, Siebers DL (1991), Soot Distribution in a D. I. Diesel Engine Using 2-D Laser-Induced Incandescence Imaging, SAE Technical Paper 910224 De Filippo A, Maricq MM (2008) Diesel nucleation mode particles: Semivolatile or solid? Environ Sci Technol 42:7957–7962 Dobbins RA (2002) Soot inception temperature and the carbonization rate of precursor particles. Combust Flame 130:204–214 Dobbins RA (2007) Hydrocarbon nanoparticles formed in flames and diesel engines. Aerosol Sci Technol 41:485–496 Frenklach M, Wang H (1991) Detailed modeling of soot particle nucleation and growth. Symp (Int) Combust 23:1559–1566 Giechaskiel B, Dilara P, Andersson J (2008) Particle measurement programme (PMP) light-duty inter-laboratory exercise: Repeatability and reproducibility of the particle number method. Aerosol Sci Technol 42:528–543 GRPE (2007) ECE/TRANS/WP.29/GRPE/2007/8 – (United Kingdom) Proposal for draft Supplement 7 to the 05 series of amendments to Regulations No. 83 (Emissions of M1 and N1 categories of vehicles). http://www.unece.org/trans/doc/2007/wp29grpe/ECE-TRANS-WP29GRPE-2007-08e.pdf Harris SJ, Maricq MM (2001) Signature size distributions for diesel and gasoline engine exhaust particulate matter. J Aerosol Sci 32:749–764 Health Effects Institute (2003) Revised analyses of time-series studies of air pollution and health. Special Report, Health Effects Institute, Boston, MA Hildemann LM, Cass GR, Markowski GR (1989) A dilution stack sampler for collection of organic aerosol emissions: design, characterization and field tests. Aerosol Sci Technol 10:193–204 Hinds WC (1999) Aerosol Technology. Wiley, New York Huntzicker JJ, Johnson RL, Shah JJ, Cary RA (1982) Analysis of organic and elemental carbon in ambient aerosol by a thermal-optical method. In Wolff GT, Klimisch RL (eds) Particulate carbon: atmospheric life cycle. Plenum, New York IPCC (2007) Climate change 2007: synthesis report. contribution of Working Groups I, II and III to the fourth assessment report of the intergovernmental panel on climate change [Core Writing Team, Pachauri, R.K and Reisinger, A. (eds.)]. IPCC, Geneva, Switzerland, 104 pp, http://www.ipcc.ch Kasper M (2004) The number concentration of non-volatile particles – Design study for an instrument according to the PMP recommendations, SAE Technical Paper 2004-01-0960 Kasper M, Sattler K, Siegmann K, Matter U, Siegmann HC (1999) The influence of fuel additives on the formation of carbon during combustion. J Aerosol Sci 30:217–225 Kelly WP, McMurry PH (1992) Measurement of particle density by inertial classification of differential mobility analyzer-generated monodisperse aerosol. Aerosol Sci Technol 17:199–212 Kittelson DB, Watts WF, Johnson JP (2006) On-road and laboratory evaluation of combustion aerosols Part 1: Summary of diesel engine results. J Aerosol Sci 37:913–930 Kwon S-B, Lee KW, Saito K, Shinozaki O, Seto T (2003) Size-dependent volatility of diesel nanoparticles: chassis dyamometer experiments. Environ Sci Technol 37:1794–1802 Maricq MM (2006a) A comparison of soot size and charge distributions for ethane, ethylene, acetylene, and benzene/ethylene premixed flames. Combust Flame 144:730–743 Maricq MM (2006b) On the electrical charge of motor vehicle exhaust particles. J Aerosol Sci 37:858–874 Maricq MM (2007) Chemical characterization of particulate emissions from diesel engines: a review. J Aerosol Sci 38:1079–1118 Maricq MM, Xu N (2004) The effective density and fractal dimension of soot particles from premixed flames and motor vehicle exhaust. J Aerosol Sci 35:1251–1274 Maricq MM, Chase RE, Podsiadlik DH, Vogt R (1999) Vehicle exhaust particle size distributions: a comparison of tailpipe and dilution tunnel measurements, SAE Technical Paper 1999-01-1461 Maricq MM, Chase RE, Xu N, Podsiadlik DH (2003) A constant volume rapid exhaust dilution system for motor vehicle PM number and mass measurements. J Air Waste Manage Assoc 53:1196–1203

2 Characterization of Combustion and Engine Exhaust Particles

37

Miller A, Ahlstrand G, Kittelson D, Zachariah M (2007) The fate of metal (Fe) during diesel combustion: morphology, chemistry, and formation pathways of nanoparticles. Combust Flame 149:129–143 Ntziachristos L, Giechaskiel B, Pistikopoulos P, Samaras Z, Mathis U, Mohr M, Ristima¨ki J, KeskinenJ, Mikkanen P, Casati R, Scheer V (2004) Overview of the European “Particulates” Project on the characterisation of exhaust particulate emissions from road vehicles: results for light-duty vehicles, Vogt R, SAE Technical Paper 2004-01-1439 Park K, Cao F, Kittelson DB, McMurry PH (2003) Relationship between particle mass and mobility for diesel exhaust particles. Environ Sci Technol 37:577–583 Park K, Kittelson DB, Zachariah MR, McMurry PH (2004) Measurement of inherent material density of nanoparticle agglomerates. J Nanoparticle Res 6:267–272 Ristima¨ki J, Keskinen J (2006) Mass measurement of non-spherical particles: TDMA-ELPI setup and performance tests. Aerosol Sci Technol 40:997–1001 Robinson AL, Donahue NM, Shrivastava MK, Weitkamp EA, Sage AM, Grieshop AP, Lane TE, Pierce JR, Pandis SN (2007) Rethinking organic aerosols: Semivolatile emissions and photochemical aging. Science 315:1259–1262 Ro¨nkko¨ T, Virtanen A, Kannosto J, Keskinen J, Lappi M, Pirjola L (2007) Nucleation mode particles with a nonvolatile core in the exhaust of a heavy duty diesel vehicle. Environ Sci Technol 41:6384–6389 Sakurai H, Park K, McMurry PH, Zarling DD, Kittelson DB (2003) Ziemann, PJ, Size-dependent mixing characteristics of volatile and nonvolatile components in diesel exhaust aerosols. Environ Sci Technol 37:5487–5495 Schleicher B, Ku¨nzel S, Burtscher H (1995) In situ measurement of size and density of submicron aerosol particles. J Appl Phys 78:4416–4422 Schneider J, Hock N, Weimer S, Borrmann S, Kirchner U, Vogt R, Scheer V (2005) Nucleation particles in diesel exhaust: composition inferred from in situ mass spectrometric analysis. Environ Sci Tech 39:6153–6161 Sgro LA, De Filippo A, Lanzuolo G, D’Alessio A (2007) Characterization of nanoparticles of organic carbon (NOC) produced in rich premixed flames by differential mobility analysis. Proc Combust Inst 31:631–638 Shi JP, Harrision RM (1999) Investigation of ultrafine particle formation during diesel exhaust dilution. Environ Sci Technol 33:3730–3736 Snelling DR, Smallwood GJ, Sawchuk RA, Neill WS, Gareau D, Clavel D, Chippior WL, Liu F, ¨ L Bachalo WD (1999) Particulate Matter Measurements in a Diesel Engine Exhaust Gu¨lder O by Laser-Induced Incandescence and the Standard Gravimetric Procedure, SAE Technical Paper 1999-01-3653 Tobias HJ, Beving DE, Ziemann PJ, Sakuri H, Zuk M, McMurry PH, Zarling D, Waytulonis R, Kittelson DB (2001) Chemical analysis of diesel engine nanoparticles using a nano-DMA/ thermal desorption particle beam mass spectrometer. Environ Sci Technol 35:2233–2243 United States Environmental Protection Agency (2007) The original list of hazardous air pollutants. http://www.epa.gov/ttn/atw/188polls.html Vaaraslahti K, Virtanen A, Ristima¨ki J, Keskinen J (2004) Nucleation mode formation in heavyduty diesel exhaust with and without a particulate filter. Environ Sci Technol 38:4884–4890 Virtanen A, Ristima¨ki J, Keskinen J (2004) Method for measuring effective density and fractal dimension of aerosol agglomerates. Aerosol Sci Technol 38:437–446 Wehner B, Philippin S, Wiedensohler A, Scheer V, Vogt R (2004) Variability of non-volatile fractions of atmospheric aerosol particles with traffic influence. Atmos Environ 38:6081–6090 Zhao B, Yang Z, Wang J, Johnston MV, Wang H (2003) Analysis of soot nanoparticles in a laminar premixed ethylene flame by scanning mobility particle sizer. Aerosol Sci Technol 37:611–620

Chapter 3

Medicine Nanoparticle Production by EHDA Jan C.M. Marijnissen, Caner U. Yurteri, Jan van Erven, and Tomasz Ciach

3.1

Introduction

Chemical products, in general, can be produced in the liquid or gas phase. However for most medicines, with their complex molecules, the liquid route will be the appropriate one. To make these medicines into nanoparticles can also be done via the wet route (colloids) or a dry route. To separate the nanoparticles from the liquid phase will be almost impossible without some contamination, so to avoid contamination a dry method might be favourable. So to be considered is the disintegration of bigger structures into (nano) fractions. Depending on the phase of the structures, liquid or solid, different disintegration techniques exist, such as grinding, liquid atomization, lithography and etching, and evaporation/condensation. Only attention will be given here to liquid atomization with the consequent droplet to particle conversion. From the several atomization methods we are only interested in methods which break up into rather uniform droplets, so we limit ourselves to jet breakup in the laminar flow region (Lefebvre 1989). Another limitation is the size of the initially generated droplets. To produce nanoparticles the initial droplet size should be already fairly small, because otherwise the begin concentration has to be unacceptably low. One should realize that the diameter of the final particle after drying equals the diameter of the initial droplet times the cube root of the volumetric concentration of the non-volatile material (van Erven et al. 2005). So in case of very low concentrations of the product material, the role of impurities might become very important. For methods, where atomization is brought about by J.C.M. Marijnissen and C.U. Yurteri Faculty of Applied Sciences, Delft University of Technology, Delft, the Netherlands e-mail: J.C.M. [email protected] J. van Erven Nano Structured Materials, TU Delft, Juliananlaan 136, 2628 BL Delft, the Netherlands e-mail: [email protected] T. Ciach Faculty of Chemical and Process Engineering, Warsaw University of Technology, Warsaw, Poland

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_3, # Springer ScienceþBusiness Media B.V. 2010

39

40

J.C.M. Marijnissen et al.

forcing liquid through a thin nozzle or orifice, such as for the Vibrating Orifice Aerosol Generator (TSI Model 3450) the generated droplet size will be about two times the orifice diameter, while the orifice size is restricted by clogging risk. So the best option is a method, which produces mono sized droplets with a diameter smaller than the inside nozzle diameter. Such a method is found in: ElectroHydrodynamic Atomization (EHDA) or Electrospraying. EHDA is a method to produce very fine droplets from a liquid (atomization) by using an electric field. By applying the right conditions, monodisperse droplets from nanometers to several micrometers can be produced. By means of an example, i.e., the production of nano platinum particles, a generic way to produce nanoparticles from a multitude of different precursors is given. After that, several examples of medicine particles made by EHDA will be given, in the nano- and micro-range, with different properties, such as controlled release, high porosity and elongated shape. Also a method, bipolar coagulation, where two sprays of opposite electrical potential are used will be discussed. In this method, each combination of a positive and a negative charged droplet can be seen as a nano- or micro-reactor, so being able to produce new chemical compounds in a liquid, aerosolized condition. Bipolar coagulation can also be used to apply nanoparticles on a carrier. Finally some attention will be given on EHDA instrumentation and out-scaling methods.

3.2

Electro Hydrodynamic Atomization and the Production of Nanoparticles

EHDA refers to a process where a liquid jet breaks up into droplets under influence of electrical forces. Depending on the strength of the electric stresses in the liquid surface relative to the surface tension stress, and on the kinetic energy of the liquid leaving the nozzle, different spraying modes will be obtained (Cloupeau and Prunet-Foch 1994; Grace and Marijnissen 1994). For the production of nanoparticles in our case the so called Cone-jet mode is the relevant one. In this mode a liquid is pumped through a nozzle at low flow rate (mL/h to mL/h). An electric field is applied between the nozzle and some counter electrode. This electric field induces a surface charge in the growing droplet at the nozzle. Due to this surface charge, and due to the electric field, an electric stress is created in the liquid surface. If the electric field and the liquid flow rate are in the appropriate range, then this electric stress will overcome the surface tension stress and transform the droplet into a conical shape, the Taylor cone (Taylor 1964). The tangential component of the electric field accelerates the charge carriers (mainly ions) at the liquid surface toward the cone apex. These ions collide with liquid molecules, so accelerating the surrounding liquid. As a result, a thin liquid jet emerges at the cone apex. Depending on the ratio of the normal electric stress over the surface tension stress in the jet surface, the jet will break up due to axisymmetric instabilities, also called varicose instabilities or due to varicose instabilities and also lateral instabilities, called kink instabilities (Hartman et al. 2000). At a low stress ratio in the varicose break-up mode the desired monodisperse droplets are produced. The droplets produced by EHDA carry a high electric charge close to the Rayleigh charge limit (Hartman et al. 2000). To avoid Rayleigh disintegration of

3 Medicine Nanoparticle Production by EHDA

41

the droplets (Davis and Bridges 1994; Smith et al. 2002), the droplets have to be completely or partially neutralized. Rayleigh disintegration happens when the mutual repulsion of electric charges exceeds the confining force of surface tension, a result here of the evaporation of the droplets. Neutralization is also desirable to make the droplets manageable, A possible method of discharging is with ions of opposite charge created by corona discharge. To estimate the right conditions and operational parameters to produce nanodroplets of a certain size, scaling laws can be used. de la Mora and Loscertales (1994) and Gan˜a´n-Calvo et al. (1997) developed scaling laws which estimate the produced droplet size (or jet diameter) and the electric current required for a liquid sprayed in the Cone-Jet mode as function of liquid flow rate and liquid properties. Hartman refined the scaling laws for EHDA in the Cone-Jet mode using his theoretically derived models for the cone, jet, and droplet size (Hartman et al. 1999, 2000). For the current scaling for liquids with a flat radial velocity profile in the jet, which is appropriate here because of the high conductivity of the solution, he derived the following relation 1

I ¼ bðgKQÞ2 ;

(3.1)

where Q is the flow rate (m3/s), I is the current through the liquid cone (A), g is the surface tension (N/m), K is conductivity (S/m), and b is a constant, which is approximately 2. The droplet diameter for the varicose break-up mode is given by Eq. 3.2:


dd;v

re0 Q4 ¼c I2

1 6

;

(3.2)

where dd,v is the droplet diameter for varicose break-up and c is a constant, which is approximately 2. Substituting Eq. 3.1 into Eq.3.2 yields:
dd;v ¼

16re0 Q3 gK

1 6

(3.3)

For a spherical particle, the diameter of the (final) particle (dp) is related to the droplet diameter (Eq. 3.3) by Eq. 3.4: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rdroplet 3 d dp ¼ 3 f rparticle droplet

(3.4)

where f is the mass fraction of the material in the solution (), rdroplet is the density of the solution and rparticle is the density of the final (product) particle (kg/m3).

42

J.C.M. Marijnissen et al.

This chapter describes the production of (medical) nanoparticles by EHDA. Other authors report already on the production of nanoparticles by EHDA (Rulison and Flagan 1994; Hull et al. 1997; Ciach et al. 2002; Lenggoro et al. 2000) but besides presenting two methods to produce specific nanoparticles by EHDA, our methods can according to us be seen as generic ways to produce well-defined nanoparticles of many different compositions on demand. Before we describe the production of medical nanoparticles, we start with an example, i.e., the production of Pt nanoparticles, to explain each step involved in detail, including chemical reactions. The two different EHDA configurations which have been used for the production of Pt nanoparticles, relate to the two different routes of the decomposition step of the platinum precursor into platinum. In the first one the precursor droplets are collected on a support and heat treated afterward. In the second route the produced precursor droplets are kept in airborne state, neutralized and heat treated before collection.

3.2.1

Experimental

Two production routes of platinum nanoparticles are used as described by van Erven et al. (2005). In both routes the droplets are produced from a solution of chloroplatinic acid (H2PtCl6.6H2O Alfa-Aesar 99.9%) in ethanol. When heated above 500 C the platinum precursor will decompose into platinum, gaseous hydrochloric acid, and chlorine (Hernandez and Choren 1983). In the first route the by EHDA produced chloroplatinic acid particles are deposited on a carrier support. After deposition the support is placed in a tubular furnace and the particles are decomposed forming platinum nanoparticles. In the second route the produced droplets are neutralized and ducted in an airborne state through a tubular furnace where they decompose. After the furnace the particles are deposited on a substrate, such as a TEM grid. The two different routes have different set ups. The first one, is referred to as the Capillary-plate set-up and the second one, as the Aerosol reactor set up.

3.2.2

Capillary-Plate Set-Up

The capillary-plate set-up is shown in Fig. 3.1. Droplets are produced by pumping (Harvard PHD2000) a 1 wt% solution of chloroplatinic acid in ethanol (K ¼ 0.04 S/ m, g ¼ 0.022 N/m) through a capillary (B). The flow rate of the solution was 13 mL/h. The required electrical field is created by applying a voltage between the capillary (B) (inner diameter 60 m, outer diameter 160 mm) and a grounded counter electrode (D) using a high voltage power supply (C) (FUG HCL 14-12500). For the experiments the potential difference between B and D was 1.26 kV and the distance between the tip of the capillary (B) and the carrier support (E) was 1 mm.

3 Medicine Nanoparticle Production by EHDA

43

Fig. 3.1 (a) Capillary-plate set-up. A – syringe; B – metal capillary; C – high voltage power supply; D – grounded plate; E – Si/SiO2 support. (b and c) SEM images of particles produced by capillary-plate set-up before and after 10 min decomposition at 700 C

The droplets are deposited on the carrier support (E) which in principle can be any material which is heat resistant at the decomposition temperature of chloroplatinic acid and is conductive to discharge the droplets. In this study thin plates of silicon, with an 0.4 mm oxidized top layer, of about 20 by 20 mm were used as carrier support. The set up was operated at room temperature. After evaporation of the solvent the support with the chloroplatinic acid nanoparticles was placed in a tubular furnace for 10 min at T ¼ 700 C to decompose the deposited chloroplatinic acid particles into platinum particles. The particles were examined before and after decomposition by a SEM (Hitachi Model S-4700).

3.2.3

Aerosol Reactor Set Up

The aerosol reactor set up is shown in Fig. 3.2a. The set up can be divided in two sections, A and B. Section A is the production part which is based on the Delft Aerosol Generator (Meesters et al. 1992). In section B the chloroplatinic acid particles are decomposed, in the airborne state, during their transport through the tubular furnace. A blow up of the production area, section A, is shown in the upper part of Fig. 3.2a. A 0.2 wt% solution of chloroplatinic acid in ethanol (K ¼ 0.01 S/ m, g ¼ 0.022 N/m) was pumped (Harvard PHD2000) through a metal capillary (I.D. 60 mm, O.D. 160 mm) with a flowrate of 8 mL/h. In this set up a ring is used as counter electrode. The ring is connected to a high voltage power supply (FUG HCL 14 12500), but at a lower voltage than the capillary, respectively 5.57 and 8.8 kV. The distance between the ring and capillary is approximately 15 mm. The potential difference between the nozzle and the ring creates the field to produce the droplets, which will pass through the ring. In this way the droplets are not deposited as in the capillary-plate set-up, but are kept airborne.

44

J.C.M. Marijnissen et al.

Fig. 3.2 (a) Aerosol reactor set up. In section A the particles are generated and dried. In section B the dried chloroplatinic particles are decomposed to form platinum particles, (b) single Pt particle of 7 nm, (c) Cluster of 3 Pt particles, (d) TEM-EDX spectrum of a platinum particle (Cu and Ni peaks are from TEM grid)

To discharge the highly charged droplets a grounded needle is used in this set up. The needle has a sharp tip and the high electric field strength there, creates a corona discharge, so supplying ions of opposite charge for the neutralization. The distance between the tip of the needle and the ring is 60 mm. The chloroplatinic acid particles are then ducted into a tubular furnace (T = 700 C) with filtered air (fv ¼ 1.5 L/min). The residence time is estimated to be 2 min. After the furnace the platinum nanoparticles are deposited on a TEM grid. The deposition takes place by two phenomena; thermophoresis and diffusion. In the beginning thermophoresis is important because the TEM grid is cold compared to the gas. When the grid has been heated up, diffusion will be the dominant process of deposition. After deposition the nanoparticles are examined by a HR-TEM (Philips CM30UT).

3 Medicine Nanoparticle Production by EHDA

3.2.4

45

Results of Pt Particles Production

A small area of a Si/SiO2 substrate with chloroplatinic acid particles, produced by the capillary-plate set-up, is shown in Fig. 3.1b. The surface concentration was obtained by spraying for 5 s. The spot sizes as seen in Fig. 3.1b vary between 80 and 120 nm. Substituting the values of the different variables as described in the experimental section in the scaling laws (Eq. 3.3) and using Eq. 3.4, yields a particle size of 63 nm (here in Eq. 3.4 f is the mass fraction of the chloroplatinic acid in ethanol, rdroplet is the density of ethanol and rparticle is the density of chloroplatinic acid). Realizing that some deformation might occur during deposition of still wet particles, the measured and calculated values are in reasonable agreement. Figure 3.1c shows the particles after the decomposition of the chloroplatinic acid in a tubular furnace for 10 min at 700 C. It can be seen that the original chloroplatinic acid particles are formed into clusters of supposedly platinum particles of 5–15 nm. This is caused by the fact that platinum does not evaporate at 700 C, while the other decomposition products are gaseous. Platinum particles produced by the aerosol reactor set up with the settings mentioned in Section 3.2.3 are shown in Fig. 3.2. In Fig. 3.2b, a TEM micrograph of a single particle of approximately 7 nm is shown. The produced particles are not charged and can therefore form agglomerates. An example of such an agglomerate is shown in Fig. 3.2c. Elemental analysis using EDX (Fig. 3.2d) showed that the particles only contain platinum. The TEM pictures also prove that the platinum particles are crystalline. Using the values of the variables as described in the experimental section the scaling laws (Eqs. 3.3 and 3.4) predict a particle size of 13 nm. By observing different areas of the TEM grid, we noticed that the particle size of non-agglomerated particles was very similar. To get an estimation of the size a limited number of particles was measured giving an average size in the order of 10 nm.

3.3

Medicine Nano- and Micro-Particles Produced with EHDA

In the preceding part a general introduction was given to produce nanoparticles with EHDA. From now we will focus on medicine particles. Drug particles with a narrow size distribution have the unique advantage of providing more regular and predictable drug release profiles from batch to batch compared to particles with the same mean size but wider distribution. Electrospraying is the ideal route for the production of such drug particles either in pure or polymer blended form. In this case a drug or a polymer/drug combination dissolved in a suitable solvent is electrosprayed. The aerosol reactor setup shown in Fig. 3.1 is used to demonstrate that nanosized medical particles or polymer blended combinations can be generated. Paclitaxel

46

J.C.M. Marijnissen et al.

Fig. 3.3 (a) Taxol; 1.0% in EtoH at 22 mL/h (21oC/38% RH), 60 s (spray time), 3 cm (spray to substrate distance), 2.1 kV (high voltage), (b) PVP; 0.3% in EtoH at 10 mL/h (22oC/47% RH), 60 s, 3 cm, 2.4 kV, (c) Taxol + PVP; 0.1% in EtoH at 20 mL/h (21oC/64% RH), 120 s, 3 cm, 2.0 kV

(Taxol) is selected to illustrate an example. Taxol is used to treat various forms of cancer such as breast, lung, ovarian, and it is also used as a way of prevention against restenosis. Taxol can be dispensed alone or blended with a biodegradable polymer such as PVP, PLA, or PLGA. Taxol (Sigma-Aldrich) is dissolved in Ethanol in a 1% mass ratio. With a flow rate of 22 mL/h and a potential difference of 2.1 kV, droplets are produced and targeted to an SEM stub placed 3 cm downstream of the nozzle. Based on the initial droplet diameter of 1.5 mm, a droplet evaporates to deposit as a dry Taxol particle in the size of 300 nm on the surface of the SEM stub, Fig. 3.3a. When the flow rate is lowered to 10 mL/h, particle sizes were in the order of 200 nm or less as confirmed by SEM (Philips XL20). In order to show the fabrication of polymer and mixed polymer/drug nanoparticles a PVP solution and a PVP/Taxol solution is utilized. In order to fabricate polymer sub¯ S-M.W. 1300000, K85-95) is utilized. Experiments are micron spheres PVP (ACRO carried out in ambient conditions. Figure 3.3b and c is an example of pure and Taxol blended polymer nanoparticles. Spray conditions for fabricating these particles are listed in the caption of Fig. 3.3. The size of the particles in Fig. 3.3 is between 200 and 300 nm.

3.3.1

Slow Release and Low Density Particles

Besides the production of medical nanoparticles as such, EHDA also offers the possibility to produce more complex particles such as slow release and low density particles, which e.g. can be used in inhalation treatment. The two following examples consider micrometer sized particles, but the methods can be equally used for the production of nanoparticles (Ciach et al. 2002) First we discuss the use of biodegradable polymer solutions for the production of slow release medicine particles. We selected poly-(lactic-co-glycolic acid) (PLGA) (50:50, Aldrich) as the primary polymer and Polyethylene glycol (PEG) as an additive to modify the decomposition rate. As an example of drug paclitaxel (taxol) was used again. As solvent a dichloromethane acetone mixture (4:1 weight) was employed. The solution was atomized with a set-up as in Fig. 3.2, but instead of a tube furnace a heating element around the outlet tube of the production part was used.

3 Medicine Nanoparticle Production by EHDA

47

Relative cumulative release [-]

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3

PLGA PLGA + PEG 4.6k

0.2 0.1 0.0 0

10

20

30

40

50

60

Time [days]

Fig. 3.4 (a) PLGA particles containing paclitaxel produced by EHDA. (b) Relative cumulative release of paclitaxel from polymer microparticles made of PLGA (poly(lactic-co-glycolic acid)) and PEG (polyethylene glycol).Conditions: room temperature, pH ¼ 7

Figure 3.4a shows PLGA particles containing paclitaxel, produced in the described way. As can be seen the size distribution is narrow but additional small particles are present. The contribution of these particles in the total mass of the system is negligible but some effort will be made in the future to avoid formation of these small particles. To investigate the drug release characteristics, particles together with the filter on which they were collected were immersed in a 200 mL buffer solution of pH ¼ 7 at room temperature with a small addition of sodium azide to prevent bacteria growth. To measure the paclitaxel release into the liquid as a function of time, samples of 1 mL of the solution were taken at certain time intervals and after passing them through a membrane filter analysed on the paclitaxel content with liquid chromatography. In the buffer solution slow release of the medicine takes place. The involved mechanisms are supposed to be hydrolytic decomposition of the polymer matrix followed by dissolution of medicine entrapped in the polymer. In addition, diffusion of active compound to the surface and dissolution probably also takes place. The results of the paclitaxel release as a cumulative release with time is presented in Fig. 3.4b. It is clear that the cumulative release of the medicine is rather linear with time, with some faster release in the first few days and a slowing down after about 35 days. This initial burst of the active substance could originate from decomposition of small particles and/or from the drug available on the particle surface. For a higher time span (some 30 days) the release rate is more or less constant. As can be seen choosing a proper polymer mixture (here PLGA + PEG) can serve as a tuning method for particle decomposition time. To produce low density particles with EHDA, we have used two different ways. Hollow or balloon like particles can be obtained with the right evaporation conditions and concentration. In reality also other factors play an important role, such as mechanical properties and porosity of the formed solid shell as well as the surface tension of the solution and the presence of surface-active compounds. If we do not

48

J.C.M. Marijnissen et al.

Fig. 3.5 (a) Budesonide1 particles produced by EHDA. The bar length in the picture is 5 mm. (b) Inflated PEG particle attached to a wire. (c) Surface of the inflated PEG particle

choose the composition of the droplets or the conditions of solvent evaporation properly, we can get the wrong particle structure such as small solid particles or remains of collapsed shells. There is still no scientific way to accurately predict if the solution we have will produce hollow spheres. The only way is still trial and error. An example of particles obtained from a 1% (wt) solution of Budesonide1 in a water-ethanol (1:10, wt) mixture, is shown in Fig. 3.5a. On the picture we can see shell-like particles. The calculated volumetric fraction of the walls in relation to the whole particle is about 1%. Some of the particles have broken walls. Among big shells we see small particles, which may have been formed from satellite droplets. Another way to obtain low-density particles is to inflate them by releasing a gas inside the polymer structure after particle formation. As a substance that can release gas we use NaHCO3 or (NH4)2CO3. These inflating agents decompose at elevated temperature (about 60) releasing carbon dioxide. At this temperature the polymer is already soft. The gases are formed inside the polymeric structure of the particle and the whole process can be compared with baking a cake where the biodegradable polymer is the dough. To verify this idea we used a solution of PEG (10 kDa M.W.) containing 0.5% (wt) of NaHCO3 and 0.1%(wt) of surfactant (related to the weight of the polymer). In a first test we created droplets on a 50 mm wire by immersing the wire in the solution. This resulted in tiny droplets hanging at the end of the wire. After evaporation of the solvent we put the wire with the particles for five minutes in an oven at 60. At this temperature the polymer became soft and the inflating agent decomposed, releasing CO2. The gas expanded the particle. An example particle is shown in Fig. 3.5b. We can see in Fig. 3.5c that the particle has a spongy structure with pores. The measured porosity of this particle is about 80%. The particle is slightly collapsed. We also try to accomplish the same process in the aerosol state by heating airborne particles. Particles are produced by EHDA and after solvent evaporation they pass a heated chamber where, we expect that gas is released inside the polymeric particles.

3 Medicine Nanoparticle Production by EHDA

49

Fig. 3.6 (a) Elongated polymer particles, (b) smooth PVP nanofibers, (c) TiO2 nanofibers, (d and e) electrospun nanotubes obtained by co-spinning olive oil/PVP-TiO2 precursor

3.3.2

Different Shapes

For certain medical applications it might be advantageous to use particle shapes different from spheres. EHDA is able to produce elongated shapes and fibres. Depending on the concentration, type of polymer and solvent and drying conditions it is possible to produce elongated particles see e.g. Fig. 3.6a. It is even possible to make fibres (electro spinning), see Fig. 3.6b and c. If fibres or particles are formed depends probably especially on the degree of chain entanglement of the polymer. By using a coaxial spinning system it is also possible to produce (nano) tubes of a certain material (polymer or ceramic) filled or not with another material (Fig. 3.6d and e).

3.4

Bipolar Coagulation and Carrier Particles

It is also possible to use two sprays of oppositely charged droplets. If they are directed towards each other coagulation between the droplets takes place through the electrical attraction between them. The two sprays are created using EHDA in the cone-jet mode. The coagulation can be used just to neutralize the droplets, but also a chemical reaction can take place in the newly formed droplet obtaining the desired product, see Fig. 3.7. If the right conditions are chosen, it is also possible to coat one material with another.

50

J.C.M. Marijnissen et al.

Fig. 3.7 Bipolar coagulation

More or less the same method can be used to load carrier particles with nano- or microsized medicine particles. This can be of high interest for the pharmaceutical industry. The electrospraying of nanoparticle laden liquids resolves an apparent problem of effectively dispersing nanoparticles. Electrospraying such suspension generates a spray of charged droplets that are seeded with nanoparticles. Thus electrospraying offers a solution for dispersing and depositing nanoparticles on a substrate. In order to enhance the efficiency of deposition, the charged nature of the nanoparticles can be exploited to coat host particles or to coat them with other droplets. EHDA leads to the formation of unipolarly charged suspensions of nanoparticles, while host particles can be charged with opposite polarity by means of tribocharging, corona or inductive charging, or get their charge due to the use of EHDA. When these particles are brought into contact in an appropriate way, the mutual electrostatic attraction force between the negative and positive charge will cause a coating to be deposited on the surface. Interaction can be realized in three ways: nanoparticles can be embedded in host particles, host particles can be encapsulated with a polymer and nanoparticles, and nanoparticles can be discretely deposited on the surface of host particle. We have studied several possibilities for mutually interacting oppositely charged particles in order to deposit nanoparticles on micro ones as an example of the latter case as depicted in Figs. 3.8a and b and 3.9a. These processes can be named as the grounded moving target (GMT) method (Dabkowski 2006; Dabkowski et al. 2007), falling curtain method (Coppens 2007; van Ommen et al. 2008), and vibrating dish method respectively. In the GMT method, in which 165 mm alumina host particles were coated with 65 nm PS nanoparticles, the host particles were charged by tribocharging on a particle feeder, while the suspension of nanoparticles were charged with opposite polarity by means of electrospraying. In the first experiment the charged host particles were fed onto the conveyor, which in this case was stationary, giving a deposition pattern as shown in Fig. 3.8c. We see good targeting of PS particles on the alumina due to the mutual attraction between the oppositely charged particles. Although in this stationary case a very high degree of deposition can be achieved, we prefer a continuous deposition method. The coating level then can be controlled by changing the residence time of the host particles in the

3 Medicine Nanoparticle Production by EHDA

51

Fig. 3.8 (a) A schematic representation of the Grounded Moving Target (GMT) set-up, (b) schematic representation of falling curtain setup, (c) Stationary coated 165 mm alumina with 65 nm PS, (d) GMT coated alumina; Three types of depositions are distinguished: single, in groups and agglomerates, (e) 200 mm glass beads coated with 500 nm PS in a falling curtain setup (conditions; mean counter air velocity of 1.26 m/s and two EHDA nozzles)

spraying zone via conveyor speed and changing the concentration of the suspension. When host particles are in motion, three types of nanoparticle deposits were identified: single, in groups and in agglomerates, Fig. 3.8d. The latter type is presumable explained by the deposition of droplets with a high concentration of nanoparticles. Tribo charging of host particles can be improved by constructing the particle feeder/charger out of a material, which is far away from the host particle in the tribo series, e.g. for glass host particle, the particle feeder is made out of Teflon. However, charging the host particles too high causes particles to stick on the feeder making it difficult to supply them to the conveyor for coating them on the conveyor. Too high charge also causes sticking of the particles to the conveyor. The falling curtain set up (Fig. 3.8b) omits the contact of particles with the conveyor surface. However the particle residence time is also reduced. Applying multiple electrosprays as well as applying a counter air flow increases the residence time of the particles (glass beads in this case) in the spray zone and thus enhances targeting of the nanoparticles on the glass beads. Figure 3.8e is an example of coating with counter air flow and two EHDA sprays. Besides these continuous processes, batch type processes are investigated to get more insight in the process involved using a vibrating dish setup as shown in Fig. 3.9a. One gram of 45 mm glass beads are tribo charged using a PTFE vibrating dish. These particles form an almost single layer of vibrating particles. Due to their confinement in the dish the electrostatic interaction with the spray is enhanced. Two cases are considered, in the first case the particles are charged by vibration and then the vibration is stopped. A small amount of these

52

J.C.M. Marijnissen et al.

Fig. 3.9 (a) Schematic of the vibrating dish coating unit, (b) 45 mm glass beads stationary coated with 100 nm PS spheres (c) 45 mm glass beads dynamically coated in a vibrating dish with 100 nm PS

particles are mounted on a SEM stub and exposed to electrosprayed nanoparticles, Fig. 3.9b. In the second case the vibration is continued and nanoparticles are directly sprayed on the vibrating host particles, Fig. 3.9c. In this setup a needle – ring configuration is used to avoid the influence of charged particles on the spray formation process.

3.5

Production Equipment

A typical EHDA setup for powder production is presented in Fig. 3.10. The set-up consists of a cylindrical glass tube of 10–20 cm diameter. One end acts as an inlet for filtered air and the other end directs the produced particles to a filter for collection. Sometimes a heating step before collection is necessary to evaporate the solvent. The EHDA spraying nozzle is positioned in a glass side tube in which also the counter electrode ring is placed close to the main glass cylinder. A corona discharge needle is inserted in the glass cylinder opposite to the spraying nozzle to neutralize the droplets generated. A more detailed description is already given in Section 3.2.3. Another set-up is proposed by Ciach (2007), which was designed to have better long term production stability, see Fig. 3.11. The reactor consists of a glass cylinder, 20 cm in diameter and 50 cm long. The EHDA nozzle is placed on top of the cylinder and is surrounded by a counter electrode in the shape of a tube with rounded edges. This tube also acts as an inlet for the air stream, which carries the particles away. Four or six corona discharge electrodes of opposite polarity as the nozzle and the ring, are placed symmetrically some distance from the bottom of the cylinder to neutralize the droplets. By carefully selecting the corona current, the particles will not be completely neutralized. They follow the air stream, which enters through the tube electrode and small holes in the upper cover near the cylinder wall (not shown). Due to their charge the particles are efficiently collected on a grounded collecting plate downstream of the reactor. The collecting plate is a disc of 18 cm diameter placed 2 cm below the cylinder outlet rim which rotates slowly. While the disc rotates slowly a Teflon

3 Medicine Nanoparticle Production by EHDA

53

Fig. 3.10 Setup for powder production by EHDA

Fig. 3.11 EHDA particle production setup by Ciach (not on scale)

scraper directs the particles to the powder container. The particles should be preferably dry before collection. The reactor produces half a gram of powder per hour and operates stably for at least 24 h. To avoid problems related to particle discharging, drying, and accumulation on the setup walls, particles can be collected in a liquid, see Fig. 3.12 (Ciach 2007). In this setup particles are atomized by EHDA 10–20 cm above the collecting solution surface. The grounded collecting liquids act as the counter electrode, in which the particles are immersed. Obviously particles should not be soluble in the collecting liquid. This method can also be used to produce porous particles and for encapsulation of poorly water soluble drugs like taxol. It is also possible to place the spraying nozzle in a liquid, with the nozzle on a high potential and a submerged

54

J.C.M. Marijnissen et al.

Fig. 3.12 Collection of EHDA particles in a liquid

Liquid feed HV

Collecting solution

grounded counter electrode. In this case emulsions, can be made. It is clear that the liquid in which is sprayed (the continuous phase of the emulsion) must have a low conductivity.

3.6

Future of EHDA – Out Scaling

As shown, Electrospraying enables controlled atomization. Therapeutic aerosols with a narrow size distribution can be generated of a desired size, chemical composition, charge and morphology, hence providing a safe and controlled way of respiratory drug delivery. Besides for the production of inhalation particles, EHDA can be used to coat particles or surfaces with medical nanoparticles in a very efficient way. This leads to cost savings in expensive pharmaceutical materials. However, industrial implementation still suffers from low production rates although much effort is put in up-scaling the production process. In order to generate small sized particles, low flow rates are required. For example, a flow rate of less than 0.1 mL/h for a single nozzle is needed to obtain droplets in the micrometer diameter range. To obtain a desired size is mainly determined by flow rate and conductivity of the liquid as dictated by the scaling laws (Eq. 3.3). For the same droplet size it is impossible to increase the production rate by increasing the flow rate. Thus an out-scaling rather than up-scaling is needed by means of using multiple sprays. There are many efforts reported on out-scaling methods including the use of an array of capillaries, an array of holes in combination of non-wetting material, serrations, grooves, multi jet mode operation as summarized by Deng and Gomez (2007). Increasing the number of capillary nozzles seems to be a simple and effective way of increasing the number of droplets. However, out-scaling mainly suffers from flow rate and field intensity variations and thus droplet size changes from nozzle to nozzle. The design may also be dependent on the nature of the liquid.

3 Medicine Nanoparticle Production by EHDA

55

There is therefore a need for systematic design tools. The challenge is a uniform delivery of the liquid and having an equal field intensity in each spraying point. Studies of Snarski and Dunn (1991) and Rulison and Flagan (1994) show that the voltage required for the steady cone-jet mode increases with a decrease in distance between the capillary nozzles. As the distance decreases in order to increase the nozzle packing density further issues will arise. Space charge, a dense charged droplet cloud, decreases the field strength at the nozzles, and so may cease the cone jet spraying of one or more nozzles. So for steady spraying a higher voltage setting is needed. So the electric field at the tip of a capillary nozzle is more often influenced by the nearby nozzles’ electric field. If the influence between the nozzles is large, also the radial component of the electric force acting on a cone is not negligible and the electric force deforms the cone at the tip of the capillary nozzle leading to no or interrupted droplet break up. As already discussed the droplets are highly charged and to avoid Rayleigh disintegration, they have to be discharged. The more jets result in the higher space charge in the gap between the cones and the counter electrodes. The higher the space charge is, the higher the required potential difference necessary for the formation of the cones. The space charge in the setup could also lead to differences in the electric field at the nozzles. The problem of the electric field can be solved by introducing a ring electrode close to the nozzle just as for a single nozzle. In that case, the electric field is determined by the field between the nozzle and the ring. Neighbouring nozzles have no longer an influence on the field at the nozzle. The problem of space charge can be solved in two ways; collecting the particles

Fig. 3.13 Schematic of multi nozzle system after Hartman

56

J.C.M. Marijnissen et al.

immediately after their generation on a conducting surface (counter electrode), or discharging and transporting the particles with a carrier gas flow. Figure 3.13 shows a schematic representation of a multiple nozzle system as suggested and realized by Hartman (1998), in where all the requirements have been fulfilled.

References Ciach T (2007) Application of Electrohydrodynamic atomization in drug delivery: a review. J Drug Deliv Sci Technol 17(6):367–375 Ciach T, Geerse KB, Marijnissen JCM (2002) EHDA in particle production. In: Kanuth P, Schoonman J (eds) Nanostructured materials. Kluwer Academic, Boston Cloupeau M, Prunet-Foch B (1994) Electrohydrodynamic spraying functioning modes: a critical review. J Aerosol Sci 25:1021–1036 Coppens PF (2007) Coating of tribocharged model particles with nanoparticles using EHDA, MS Thesis, Delft University of Technology, Faculty of Applied Sciences, Nanostructured Materials Research Group, Process and Product Engineering Dabkowski MF (2006) Coating of particles with nanoparticles by means of electrostatic forces, MS Thesis, Delft University of Technology, Faculty of Applied Sciences, Nanostructured Materials Research Group, Process and Product Engineering Research Group Dabkowski MF, van Ommen JR, Yurteri CU, Hochhaus G, Marijnissen JCM (2007) The coating of particles with nanoparticles by means of electrostatic forces. In: Schreglmann C, Peukert W (eds) Partec 2007 – CD proceedings, Nuernberg, Germany, paper S37_2 Davis EJ, Bridges MA (1994) The Rayleigh limit of charge revisited – light-scattering from exploding droplets. J Aerosol Sci 25(6):1179–1199 De la Mora JF, Loscertales IG (1994) The current emitted by highly conducting taylor cones. J Fluid Mech 260:155–184 Deng W, Gomez A (2007) Influence of space charge on the scale up of multiplexed electrosprays. J Aerosol Sci 38:1062–1078 Gan˜a´n-Calvo AM, Davila J, Barrero A (1997) Current and droplet size in the electrospraying of liquids. Scaling laws. J Aerosol Sci 28:249–275 Grace JM, Marijnissen JCM (1994) A review of liquid atomization by electrical means. J Aerosol Sci 25(6):1005–1019 Hartman RPA (1998) Electrohydrodynamic atomization in the cone-jet mode. From physical modeling to powder production. PhD thesis, Delft University of Technology Hartman RPA, Brunner DJ, Camelot DMA, Marijnissen JCM, Scarlett B (1999) Electrohydrodynamic atomization in the cone-jet mode physical modeling of the liquid cone and jet. J Aerosol Sci 30(7):823–849 Hartman RPA, Brunner DJ, Camelot DMA, Marijnissen JCM, Scarlett B (2000) Jet break-up in electrohydrodynamic atomization in the cone-jet mode. J Aerosol Sci 31(1):65–95 Hernandez JO, Choren EA (1983) Thermal stability of some platinum complexes. Thermochimica Acta 71(3):265–272 Hull P, Hutchison J, Salata O, Dobson P (1997) Synthesis of nanometerscale silver crystallites via a room-temperature electrostatic spraying process. Adv Mater 9(5):413–417 van Erven J, Moerman R, Marijnissen Jan CM (2005) Platinum nanoparticle production by EHDA. Aerosol Sci Technol 39(10):929–934 Lefebvre AH (1989) Atomization and sprays. Hemisphere Publishing, WA Lenggoro I, Okuyama K, de la Mora J, Tohge N (2000) Preparation of ZnS nanoparticles by electrospray pyrolysis. J Aerosol Sci 31(1):121–136 Meesters G, Vercoulen PHW, Marijnissen JCM, Scarlett B (1992) Generation of micron-sized droplets from the Taylor cone. J Aerosol Sci 23(1):37–49

3 Medicine Nanoparticle Production by EHDA

57

Rulison AJ, Flagan RC (1994) Synthesis of Yttria powders by electrospray pyrolysis. J Am Ceramic Soc 77:3244–3250 Smith JN, Flagan RC, Beauchamp JL (2002) Droplet evaporation and discharge dynamics in electrospray ionization. J Phys Chem A 106(42):9957–9967 Snarski SR, Dunn PF (1991) Experiments characterizing the interaction between two sprays of electrically charged liquid droplets. Exp Fluids 11(4):268–278 Taylor GI (1964) Disintegration of water drops in an electric field. Proc R Soc A280:383–397 van Ommen JR, Beetstra R, Nijenhuis J, Yurteri CU, Marijnissen JCM (2008) Coating of tribocharged host particles with nanoparticles using electrospraying, Particulate processes in the pharmaceutical industry II, San Juan, Puerto Rico, 3–7 February

Chapter 4

Electrospray and Its Medical Applications Da-Ren Chen and David Y. H. PUI

4.1

Introduction

Electrohydrodynamic atomization, commonly called “electrospray (ES)”, has recently attracted a great deal of interests in research communities. It is because of its enormous potential in practical applications. In traditional applications, the process has been applied to the surface coating (Hines 1966; Paul 1985; van Zomeren et al. 1994), agricultural treatments (Coffee 1964), emulsion (Nawab and Mason 1958) or supermicron aerosol production, fuel spraying (Jones and Thong 1971), micro-encapsulation (Langer and Yamate 1969), ink-jet printers (Tomita et al. 1986), and colloid micro-thrusters (Huberman et al. 1968). More recently new applications have been explored. Examples include (1) using the electrospray as ion sources for mass spectrometry (ES MS) for the macromolecular detection (Yamashita and Fenn 1984; Fenn et al. 1989; Thompson et al. 1985; Cole 1997; Smith et al. 1997; Dulcks and Juraschek 1999), (2) monodisperse nanoparticle generation (Chen et al. 1995), (3) biomolecule detection using gas-phase electrophoretic mobility molecular analyzer (GEMMA) (Kaufman et al. 1996; Kaufman 1998, 1999; Koropchak et al. 1999; Scalf et al. 1999; Bacher et al. 2001), (4) enhancement of droplet mixing by inter-electrospray (Dunn and Snarski 1991; Snarski and Dunn 1991; Dunn et al. 1994), (5) targeted drug delivery by inhalation (Tang and Gomez 1994), (6) micro-mixing for drug powder production (Borra et al. 1999), (7) inorganic nanoparticle preparation by electrospray pyrolysis (Lenggoro et al. 2000), (8) preparation of non-structured ceramic thin films (Chen et al. 1999), (9) electrospray gene transfection (Chen et al. 2000), (10) compound-jet D.-R. Chen (*) Department of Energy, Environmental and Chemical Engineering, Washington University, St. Louis, MO USA e-mail: [email protected] D.Y.H. PUI Director of Particle Technology Laboratory, Department of Mechanical Engineering, University of Minnesota, Minneapolis, MN USA

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_4, # Springer ScienceþBusiness Media B.V. 2010

59

60

D.-R. Chen and D.Y.H. PUI

electrospray for the nano-encapsulation and enhancement of targeted lung delivery of nano-medicines (Chen and Pui 2000). The general reviews of the electrostatic atomization and applications are covered in the books written by D. Michelson (1990) and by A. G. Bailey (1988). More advanced discussion on the functional modes in the process are presented in the works of Cloupeau and Prunet-Foch (1989, 1990, 1994), and Jaworek and Krupa (1999). Among all the electrospray operating modes, the most commonly used and studied one is the so-called “cone-jet mode”. It is because the cone-jet mode operation has the capability of producing monodisperse particles with the same electrical polarity. It means particles generated by the conejet electrospray are monodisperse and non-agglomerated. Many applications can be made possible or benefit from particles with such properties, especially for those making use of particles in the nanometer size range. The most common configuration of ES system is the point-to-plate (or orifice plate) arrangement. A single capillary with one end serving as the “point” is used to deliver the spray liquid into the spray chamber. The spray liquid is often delivered into the capillary by either a syringe pump or gravity force. High voltage is applied either at the capillary with the grounded plate, or on the plate with the grounded capillary. At a proper voltage the shape of liquid meniscus at the capillary end will form the conical shape with a tiny jet issued from the cone tip (so-called cone-jet mode). For some of ES systems an orifice plate is used in place of the solid plate, allowing generated particles to exit the spray chamber for further particle conditioning. The systems consisting of a single capillary is named as the single-capillary ES system in this article. Many applications have been explored with single-capillary ES systems. Several limitations on the ES operation exist for the single-capillary ES systems. First, the electrical conductivity of spray liquid or solution needs to be conditioned within a proper range in order to electrospray them in a stable operation. For some liquids, especially for non-polar solvents, conductivity conditioners (or ion additives) are difficult to be identified. For some cases only limited range of electrical conductivity can be varied even if the proper conditioners were found. Secondly, the particle size and the associated electrical charges cannot be controlled independently. Additional charge conditioner will then be needed if the charge level on ES-generated particles were critical for applications. Lastly, the coating of colloidal particles with a different material (e.g., polymers) by single capillary ES systems is limited for the cases that particles and coating material can stably co-exist in the same solvent. To further extend the applications for ES technique dual-capillary ES systems are thus proposed. In such a system a dual-capillary assembly (one served as the outer cylinder and the other as inner tube) is used. The outer and inner tubes are often coaxially aligned. The coaxial tubing arrangement creates two flow channels for introducing two liquids. One flow channel is in the annual spacing between the outer and inner tubes, and the other in the inner tube. The use of the dual-capillary assembly in ES systems expands its potential to overcome the limit of singlecapillary ES systems and opens for more applications based on the ES technique. The dual-capillary configuration of ES systems was first proposed to introduce biomaterial into cells for gene transfection (Chen and Pui 2000). It makes use of the

4 Electrospray and Its Medical Applications

61

space charge effect to propel biomaterial particles at high speed prior to reaching target cells. Here, the dual-capillary ES system was proposed to introduce more electrical charges on the particles by sheathing the biomaterial suspension (in the inner capillary) with highly conductive and volatile liquid (in the outer capillary) during the spray process. From the calculation it is found that the particle velocity in such an application is primarily attributed by the space charge effect, resulting from the production of highly charged particles at high concentration, upon the particle impact on the cell membrane. Using the coaxial electrospray, Loscertales et al. (2002) further demonstrated the production and control of monodisperse capsules in submicron sizes, varying from 0.15 to 10 mm. They also found that the diameter of capsules produced by coaxial electrospray is influenced not only by the operational parameters, such as liquid feed flowrates, but also by the physical properties of spray liquids as well as the interaction between inner and outer liquids during the spray process. Lopez-Herrera et al. (2003) investigated electrified coaxial jets of two immiscible liquids issuing from a structured Taylor cone. To interpret the experimental observation, they introduced the concept of a driving liquid and presented the linear scaling law for the compound jet diameters. Chen et al. (2005) studied compound jet electrospray modes using an ethanol–glycerol– tween80 (polysorbate detergents) mixture and cooking oil, two immiscible liquids. They found that the spray phenomena were mainly controlled by the property of outer liquid, which was very viscous and electrically conductive. When compared with other liquid atomization techniques the uniqueness of ES technique is its ability to produce highly-charged, non-agglomerated and monodisperse particles in a wide size range from the nanometer to supermicron range. The unique properties of ES-generated particles offer great control on the dispersion and deposition of particles. Such features on the technology opens up many modern medical and biological applications where the precise control of particle size, morphology, and electrical charges associated with, as well as the control on the particle dispersion and deposition are needed. In this chapter we will briefly introduce the electrospray technology and its variation. Examples of medical and biological applications using the electrospray technique will then be discussed. Note that the intention of this chapter is to make readers aware of the electrospray technology and its application in medical/ biological areas. A comprehensive review of the subject is not the intention of the authors for this chapter.

4.2 4.2.1

Basics of Single- and Dual-Capillary Electrospray Basics for Single-Capillary ES

All recent applications described herein are operating the electrospray in this mode, so-called “cone-jet electrospray”. Consequently the following review is primarily focused on the cone-jet electrospray. The fundamental characteristics of cone-jet

62

D.-R. Chen and D.Y.H. PUI

electrospray are based on the spray electrical current and droplet size produced. With the experimental observations and dimensional analysis of the parameters in a cone-jet electrospray, de la Mora and Loscertales (1994), and Rosell-Llompart and de la Mora (1994) proposed that the spray current, I, can be formulated as f(k) (GKQ/k)1/2 and the main droplet size, Dd, can be scaled with r* (the charge relaxation length) (where K is electrical conductivity of liquids, G liquid surface tension, m liquid absolute viscosity, and k relative dielectric constant of liquids, Q the feeding flow rate). The charge relaxation length, r* is defined as the characteristic traveling distance for electrical charges to make up its loss due to the jet breakup; it can be estimated as (Qt)1/3 where t is the electric charge relaxation time and is inversely proportional to the electrical conductivity of spray solutions). The function of f(k) and the proportionality for Dd were later experimentally determined by Chen and Pui (1997). Taking a different approach, Gan˜a´n-Calvo (1994) published a set of new scaling laws based on the asymptotic analysis of the 1-D governing equations for electrical and flow fields in the near region where the jet is emitted from a liquid cone. They are summarized as I ¼ CI*k1/4(gKQ/k)1/2 and Dd ¼ Cd*k1/6(Qt)1/3 for polar liquids; I ¼ KI*(QKg3/r)1/4 and Dd ¼ Kd*Q1/2(r/ gK)1/6 for non-polar liquids. However the proposed scaling laws were revised in later works of the same author (Gan˜a´n-Calvo 1997, 1999). On the other hand, the recent experimental and numerical work by Hartman et al. (2000) suggested that Dd ~ Q1/2 if varicose breakup involved and Dd ~ Q1/3 if whipping break-up process occurred. Unfortunately, the transition between two jet breakups cannot be confirmed or predicted by the proposed model. The experimental determination of this transition also presents a difficulty because of the poor quality of collected data. Table 4.1 summarizes the scaling laws on the droplet size and emitted current proposed in various publications. The other important spray characteristic is the mean charges of individual droplets produced by a cone-jet electrospray. The knowledge is of importance in many practical applications such as precision deposition of particles, droplet micromixing, detection of macromolecules, gene transfection, drug delivery and so on. However, disagreements exist among the literatures. According to the models Table 4.1 Summary of scaling laws for droplet size and emitted current in the electrospray operated at the cone-jet mode Emitted current, I References Droplet diameter, Dd f(k) (gKQ/k)1/2 Rosell-Llompart et al. (1994); Chen and g(k)(Qt)1/3 Pui (1997a) CI k1/4(gKQ/k)1/2 Cd k1/6 (Qt)1/3 Gan˜a´n-Calvo (1994) polar liquids KI (QKg3/r)1/4 Kd Q1/2(r/gK)1/6 non-polar liquids 3.78p2/30.6Q1/2(re0/ 4.25(gQK/ln(Q/Q0)1/2)1/2 Gan˜a´n-Calvo (1997) Kg)1/6 2.6 (QK/g)1/2 Gan˜a´n-Calvo (1999) 2.9 e0p2/3(rgQ3/K)1/6 1/2 I ~ (QK)1/2 Hartman et al. (2000) Dd ~ Q , varicose breakup Dd ~ Q1/3, whipping break-up

4 Electrospray and Its Medical Applications

63

proposed by Vonnegut and Neubauer (1952), Ryce and Patriarche (1965), Pfeifer and Hendricks (1967), the mean droplet charge was about 50% of the Rayleigh limit. Experimental data were provided for the confirmation. Although the principle applied in these models was questioned by Krohn (1973) the result was often cited in the literature. The analysis given by Jones and Thong (1971) predicted that the charge-to-mass ratio for electrosprayed droplets was a linear function of Dd1 in the small particle size range. No experimental data was given to verify the result until the works published by Cloupeau and Prunet-Foch (1989), Gomez and Tang (1993) for heptane droplets in 5–10 mm, Tang and Gomez (1994) for water droplets in 2–10 mm size range, and Chen et al. (1995) for water droplets of the sizes ranging from 0.1 to 1 mm. Gomez and Tang (1993) also found that for heptane droplets of the sizes larger than 10 mm, the charge-to-mass ratio could be characterized by the power function of Dd3/2. Meanwhile, based on the revised scaling laws, Gan˜a´nCalvo (1999) suggested that the maximal surface electric charge on electrosprayed droplets had a universal value independent of the jet size and the liquid flow rate, and could be given as 0.53*21/2p1/3(e0g2rK2)1/6. It should be noted that in all these previous studies the mean charge on individual droplets was derived from the droplet size and spray current given by scaling laws or obtained by measurements. The assumption underlying the derivation was that the charge distribution on electrosprayed particles was relatively narrow if relatively monodisperse droplets were produced. Thus the droplet production rate was estimated from the liquid mass flowrate with the droplet size given by the scaling law or measurements. The mean electrical charges were then calculated from the spray current and droplet production rate. However, the assumption was questioned in the experimental work of de la Mora (1997). The assumption was further challenged with the observation that the gas surrounding the cone-jet had an important effect on the spray current (Aguirre-de-Carcer and de la Mora 1995). Moreover, the charge-to-mass characterization is complicated by the presence of satellite droplets. They are often observed in the production of primary droplets due to the nonlinear dynamics and breakup of liquid jets. For neutral jets they came about through the mechanisms of pinching singularity (Eggers 1997a). Prior to the breakup a tiny neck is usually formed between the jet and drop. The singularity is initially localized and producing pinchoff at the location where the neck is attached to the drop. Since only a small amount of fluid is involved, it acts on time scales much shorter than the growth of disturbances on the jet, once it sets in. Due to the asymmetrical singularity the only way the liquid neck can be matched onto an outer solution is by pinching off again at the point where the neck is linked with the jet. Hence the liquid neck is pinched off from both sides and eventually contracts into a satellite droplet. The detail process was photographed and further simulated using the 1-D models (Eggers 1997b). For electrically charged jets similar process was numerically observed in the work of Setiawan and Heister (1997). However, electric charges carried by these satellites remain largely uncharacterized. Meanwhile, satellite droplets could also be produced by the droplet fission. Recent works on electrosprayed droplets of supermicron sizes have shown that under the influence of aerodynamic forces droplet fission occurred already at charge states of

64

D.-R. Chen and D.Y.H. PUI

about 70–80% of Rayleigh limit (Taflin et al. 1989; Gomez and Tang 1993; Shrimpton 2005). All experimental investigations have shown that there was no “explosion” of the droplet but the droplet surface formed a cone-like shape from the tip of which a number of smaller droplets was ejected. Measurements showed that this “uneven fission” the primary droplet lost about 15% of the charge and about 2% of its mass. However, whether this scenario can be taken for droplets in submicron and nanometer size range remains unknown. The mean charge characteristics are even more complex if evaporating droplets are produced. As the droplet size is continuously reduced through either solvent evaporation or droplet fission electrical charges may release from droplets (as ion emitters). It is because of the buildup of the electric field strength at the droplet surface (Iribarne and Thomson 1976). The scenario is often called “ion evaporation”.

4.2.2

Dual-Capillary ES

The operation of dual-capillary ES systems is more sophisticated than that of single-capillary ES systems. The main reason is due to the involvement of two liquids coaxially introduced into a highly non-uniform electrical field. The scientific knowledge on the liquid behavior under such a condition is generally insufficient. From the operational point of view the first question to be addressed for using the dual-capillary ES systems is under which condition a stable compound cone-jet mode can be established similar to that established in the single-capillary ES system to generate monodisperse particles. This question has been partially answered by Mei and Chen (2008). They found that a stable compound cone-jet mode can be easily established for miscible and partially miscible liquid pairs. For an immiscible liquid pair, the liquids should satisfy two sufficient conditions to form a stable compound cone-jet mode: (1) a liquid of high dielectric constant should be used as the inner one; and (2) the surface tensions of liquid pair should satisfy the spreading-coefficient criterion for the engulfing and partial engulfing cases in three-phase interaction. For the first condition, the density of electric flux emitted from the outer liquid cone base would be more than that from the inner liquid cone base, if the dielectric constant of the outer liquid was higher than that of the inner liquid. As a result, the normal electric force along the interface would be greatly reduced or negligible. The formation of the inner cone, based on the balance of normal electric stress and interfacial tension, thus became impossible. For the second condition, the dynamic condition of the liquid pair used in a dual-capillary ES system should not depart much from the static condition of three-phase interaction, because of the slow fluid motion in the liquid cones in the stable compound cone-jet mode. The 2nd question to be addressed is under which condition the encapsulation of droplets will occur in dual-capillary ES system. Based on the study of Mei and Chen (2007) it was found that two different types of droplet size distributions, e.g., uni-modal and bi-modal types were produced by dual-capillary ES systems. For the examples of ethanol–olive oil, TBP–olive oil, ethanol–mineral oil pairs (i.e., inner

4 Electrospray and Its Medical Applications

65

liquid–outer liquid), unimodal size distributions were generally observed. In these cases, the size of the produced capsules linearly increased with an increase of the inner liquid flowrate in the low flowrate regime. It implied that varicose or kink instability was the dominant mechanism in the low flowrate regime. After exceeding a certain value of inner liquid flowrate, the particle size remained constant, implying that the jet breakup was significantly influenced by the inertia effect of inner liquid at high flowrate. For other test cases in the study, bi-modal size distributions or polydisperse but uni-modal size distributions of produced particles were observed, which suggested uncontrollable or failed encapsulation for such liquid pairs. For the cases with bi-modal size distributions an increase of the inner liquid flowrate decreased only the concentration of particles in the large size peak but their sizes remained constant. Through the data analysis they identified the criteria to predict the formation of capsules by the dual-capillary ES process. Two domains (i.e., controllable and uncontrollable encapsulation) were indicated in the R*O/R*I vs r*O/r*I plot, where R* was the inertia length of liquid, r* the charge relaxation length of liquid, and the subscript o and I indicating outer and inner liquids used in the ES systems. The finding of two domains on the plot implied that the encapsulation using the dual-capillary ES mainly accounted for combined effects of the relative importance of the inertial and electrical force for the inner and outer liquids. Further work and study need to be performed on the investigation of spray current and droplet size produced by the dual-capillary ES systems. The authors are only aware of two studies on these subjects. Interested readers can refer to the publications to find out the details of both studies. In summary limited liquid pairs were tested in the study performed by Lopez-Herrera et al. (2003). Empirical models on spray current and droplet size generated from the dual-capillary ES system had been proposed to fit the data collected in the study. Unfortunately the proposed models cannot be applied to all the studied liquid pairs at the same category. A broader range of liquid pairs were studied in the dissertation performed by Mei (2008). This study also advances our understanding on the subjects but it is far from what we know for single-capillary systems.

4.3

4.3.1

Examples of Medical/Biological Applications Using Electrspray Drug Reformulation

Most newly synthesized medicines are poorly soluble in water. Because of their poor solubility, some have been abandoned for further development and others require the patients to take high dose of the medicines with potential negative side effects. One way to increase the solubility is to reduce the size of medicine particles and consequently increase the surface area of particles. The single-capillary

66

D.-R. Chen and D.Y.H. PUI 13.1

4.0

Impactor D50

3.5

3.28e+005

dN/dlogDp (#/cm³)[e5]

3.0 2.5 2.0 1.5 1.0 0.5 0.0

10

Diameter (nm)

100

1000

Fig. 4.1 Particle size distribution produced by single-capillary electrospray technique. TSI scanning mobility particle sizer (Model # 3936NL25) was used to measure the size distribution of particles

electrospray technique has been proposed to produce nanometer-sized pharmaceutical particles. An example of such application is shown in Fig. 4.1. The size distribution of steroid particles produced by single capillary ES technique was measured by the scanning mobility particle sizer (SMPS). It is evidenced that steroid particles of 13 nm diameter can be produced by the ES technique. In this case the solvent used was water and nitric acid was used to adjust the electrical conductivity of spray solutions to 520 mO/cm. The liquid feed flow rate was 0.1 mL/min. The issue of using the electrospray technique for drug reformulation is on the mass throughput of the ES systems. More development work needs to be done in this area to address the issue. One idea to increase the mass throughput of the singlecapillary ES system is to use multiple capillaries in parallel. However the tight packing of multiple capillaries presents a challenge on the successful implementation of the multi-capillary systems. It is primarily because of the space charge effect resulted from highly charged particles produced. At present all the successful works were accomplished for solutions with low surface tension and low electrical conductivity. For such solutions the electrical charges associated with ES-produced particles are much less than those produced by electrospraying highly conductive solutions of high surface tension.

4.3.2

Nanoparticle Dispersion for Toxicity Study

Nanoparticles are encountered in many industrial systems utilizing aerosol reactors. Such reactors are used in industries to make a wide variety of particulate

4 Electrospray and Its Medical Applications

67

commodities, such as carbon black, pigments, and materials for high technology applications such as optical waveguides and powders for advanced ceramics (Stamatakis and Natalie 1991). Kruis et al. (1998) have described newly proposed applications in nanotechnology. To realize these new applications, nanoparticles of different physical and chemical properties are synthesized through different routes with unknown toxicity. A similar scenario is encountered in many other systems, where a large quantity of the so-called “undesirable” aerosols is produced. Biswas and Wu (1998) cited municipal waste incinerators, hazardous waste incinerators, welding systems, exhausts, coke ovens, smelters, nuclear reactor accidents, utility boilers, and the exhausts from automobile, diesel engine, and jet aircraft as examples. In manufacturing, nanoparticles need to be collected to fabricate parts for applications. In the case of waste particle generation, the particles could be potentially very toxic, and their emission to the ambient atmosphere needs to be prevented. With recent increasing findings that nanoparticles may be associated with deleterious health effects (Wolfgang et al. 2006), it will be necessary to study the toxicity of nanoparticles. Most engineered nanoparticles are in the powder form or colloidal suspension in liquids. The dispersion of nanoparticles in gaseous phases is needed to investigate the toxicity of nanoparticles through the in-vivo and in vitro routes. Unfortunately the dispersion of nanoparticles from suspensions cannot be accomplished by conventional pressure atomization. As an example the suspension of PSL particles of 28 nm diameter was atomized using the Collison atomizer. Shown in Fig. 4.2a is the particle size distribution measured by SMPS when the diluted PSL suspension was dispersed. It is obvious that PSL nanoparticles cannot be isolated from the measured particle size spectrum. One of the reasons resulted in such observation is the impurity from the water and the original PSL suspension (e.g., the surfactant used to keep PSL nanoparticles apart). The other reason is due to the nature of pressure atomization. The pressure atomization technique often produced droplets with a broad size range. With the impurity presence in the colloidal suspension, atomizer-produced droplets in which only impurity is present will form residue particles in nanometer size range. Shown in Fig. 4.2b is the size distribution measured by electrospraying diluted PSL suspension. By tuning the droplet size slightly larger than 28 nm in diameter it is possible to isolate the PSL nanoparticles from the SMPS-measured size distribution. A further example to demonstrate the capability of ES technique to disperse nanoparticles suspension is given in Fig. 4.2c. In this case the slurry used in CMP (Chemo-mechanical polishing) process was used. The sprayed slurry was diluted from the original suspension by a factor of ten in concentration. A programmable syringe pump was used to linearly vary the liquid feed flow rate into the ES capillary. The size distribution shown in the figure was accomplished by continuously varying the feed flowrate (consequently varying the droplet size produced) to synchronize with SMPS scanning process. Nanoparticles of three different sizes were detected. Nanoparticles of three different sizes in the diluted slurry were also confirmed by the SEM imaging of the sprayed sample.

68

D.-R. Chen and D.Y.H. PUI

Fig. 4.2 (a) SMPS-measured particle size distribution by dispersing 28 nm PSL suspension with collison atomizer. (b) SMPS-measured Particle size distribution by dispersing 28 nm PSL suspension with single-capillary electrospray (c) SMPS-measured size distribution of particles airborne by electrospraying the diluted CMP slurry

4.3.3

Electrospray Inhaler for Asthma Patient

Inhalers are critical devices to deliver medicine into the target location of asthma patient’s lung for effective reduction of the asthma symptom. Research in developing a better inhaler has been progressing for a couple of decades. Different atomization techniques were proposed to accomplish the delivery task. Most of these techniques produce polydisperse particles. However, the particles size and

4 Electrospray and Its Medical Applications

69

electrical charges associated with them are two important factors to deposit medicine particles in the target lung area. With polydisperse particles generated by many atomization techniques some (or majority for some cases) of the particles would not be deposited on the target lung area. Because of the monodispersity of ES produced particles it has been proposed for inhaler applications. Battelle Memorial Institute has been a key player for the commercialization of MysticTM inhaler which is based on the electrospray technology. A spin-off company, Ventaira Pharmaceuticals Inc. was established to further develop and market the inhaler. The design of the MysticTM inhaler is based on the patents filed by Coffee et al. (2004). In addition to the electrospray component for atomizing medicines a corona discharge compartment is included in the inhaler to reduce the electrical charges on ES-produced droplets for the particle delivery. The product was initially scheduled to be introduced commercially in 2008 but was delayed due to the issues encountered in the Phase I clinical trial. The Ventaira Company is now ceased to exist and was purchased back by Battelle Memorial Institute.

4.3.4

Polymer Coating for the Control of Drug Release Rate

Coating of medicine particles for the controlled drug release is one of modern medical applications. By controlling the medicine release rate the patients can take the medicine on the daily, weekly or even monthly basis. The drug release control also allows the application of medicine when needed. The encapsulation of pharmaceutical particles can be accomplished by the dual-capillary ES technique (Mei 2008, 1991). The bio-degradable polymers, e.g., PEG and PLGA are often used as the coating material. For other medical applications proteins can be used as the coating material.

Fig. 4.3 (a) A typical stent for the human use. (b) Coated stent prepared by electrospray technique

70

D.-R. Chen and D.Y.H. PUI

Fig. 4.4 (a) The porous film morphology of the coated layer, prepared by electrospray technique, on the stents. (b) The continuous film morphology of the coated layer, prepared by electrospray technique, on the stents

4.3.5

Coating of Medical Devices

The coating of stents with medicines and/or polymers used the dual-capillary electrospray system is one of the examples. The bypass surgery was often performed to treat clogged blood vessel in the human hearts in the past. The surgery is not performed regularly with the invention of stents. Stents are often made by laser carving stainless steel tubes of tiny diameters to form a wire structure. An SEM image of a stent used in human heart is shown in Fig. 4.3a. The wire structure of stents allows it to be significantly expanded in volume. The heart surgeon can thus insert the un-expanded stent into the blood vessel through a tiny cut on the patient body and expand it after proper positioning it in the clogged vessel. In stent restenosis is an issue, expected in 6 months following its installation, leading to the clogging of the blood vessel again. One way to resolve this issue is to coat stents with the medicine to prevent the vessel tissue from growing around the opened stent wall. In the current practice the stent coating is done by the ultrasonic atomization of the medicine solution and depositing particles on the stent by inertial impact. Because of the size and porosity of stents majority of sprayed particles are not deposited on the stent and are wasted. Due to the nature of the electrospray technology it offers great improvement on the waste reduction and yield increase over the current practice. With the presence of electrical field in the ES system charged particles in sub-micrometer sized range will follow the existing electric field and deposited on the stent wall. Because of the electric field around the stent wire wall particles will be deposited around the wall wires, instead of merely on the outer side of the stent wall. The coating around the stent wall wires is possible when coating particles in the submicron sizes. Shown in Fig. 4.3b is the coated stent done by the electrospray process. The uniformity of coating layer around the stent wall wires is evidenced. Further, as demonstrated in Fig. 4.4a and b, the morphology of coated film on the stent can be controlled by using the dual-capillary ES technique.

4 Electrospray and Its Medical Applications

71

Fig. 4.5 (a) Nanogradient thin film of neuron growth factor, prepared by electrospray technique. (b) Guided neuron growth using the nanogradient thin film of neuron growth factor

4.3.6

Formation of Nano-Gradient Growth Factor for Guided Neuron Growth

Nano-scaled variation of the concentration of particles deposited on the substrate can be accomplished using the single-capillary ES technique. The task of creating the pattern with the deposited material in nano-gradient concentration on a substrate can be done by either direct electrospray writing or the use of patterned mask to filter electrosprayed particles. The ability of making the pattern with nano-scaled gradient of concentration opens its potential medical application. An example given herein is for the guided growth of neuron cells. Shown in Fig. 4.5a is the gradient line of the neuron growth factor prepared by electrospraying and masking growth factor with the careful control of the substrate moving. Fig. 4.5b evidences the guided growth of frog neuron cells after placing them on the prepared growth factor nano-gradient lines.

4.3.7

Gene Transfection Using the Electrospray Technique

Gene transfection at the cellular level offers much application and potential in plant improvement, cancer therapy and other applications in biology. Many techniques have been proposed to introduce genes into the cells. The electrospray technique offers an alternative way to accomplish the same task. An exploratory study was published in Chen et al. (2000). It has been demonstrated that the particles, produced by the dual-capillary electrospray with the introduction of additional electrical charges by outer ionized liquids, have sufficient velocities to enable them penetrating through the membrane of animal cells.

72

4.4

D.-R. Chen and D.Y.H. PUI

Summary

The uniqueness of the electrospray techniques relies on its ability to produce highly charged, monodisperse particles in the diameter range from nanometer to supermicrometer. Due to the electrical charges of the same polarity electrosprayed particles are non-agglomerated. The technology allows users to gain greater control on the generation, dispersion and deposition of particles over other atomization techniques. It allows users to tailor particles to the desired size, morphology and construction for improved particle function and transport properties, making the technique suitable for many particle applications, especially for medical and biological applications. In this chapter we have briefly reviewed the history and the evolution of the electrospray technique. Single- and dual capillary electrospray techniques were introduced. The up-to-date basics on the operation of single- and dual- capillary ES techniques were also summarized. The last part of this chapter has been devoted to provide example applications of the electrospray technique in medical and biological areas. Many studies have been reported on using the electrospray technique for medical applications. More work is needed to take this wonderful technology to the next level for practical applications. One issue is related to the mass throughput of the electrospray technique. Limited success has been reported using the multiple capillary systems. All the reported works with multiple capillary systems involved spraying solutions of low surface tension and electrical conductivity, resulting in less electrical charges on particles and thus lower space effect attributed by the charged particles. The space charge effect resulted from the charged particles in high concentration will eventually limit the number of capillaries that can be deployed in the multiple capillary ES systems. The mass throughput of multiple capillary systems cannot be scaled up indefinitely. A breakthrough design for implementing the multiple capillary systems will be needed in the future. Another concern of using electrospray for medical application is related to the viability of biomaterials after spraying even though the viability of some biomolecules has been established (Kwok et al. 2008; Clarke and Jayasinghe 2008). Potential damage to the bio-molecular structure exists, especially for fragile bio-molecules. To evaluate the bio-viability of sprayed biomaterial is always a necessary step in the technology development. The situation also calls for the development of a soft electrospray technique to ensure no damage to the sprayed bio-materials. Lastly, the control of electrical charges on the electrosprayed particles is often accomplished through the use of corona discharge devices, either DC or AC. It is because of the gradually tightened safety regulation for the use of radioactive materials for neutralizing the charged particles. Unfortunately, ozone is also produced in the corona discharge process. The strong oxidation ability of ozone presents the threat of bio-material damage; it also has adverse health effect on the patient if ES was used to deliver the medicine into the human lung. An alternative approach to controlling the charges on the electrosprayed particles will be much needed for the medical application of the electrospray technique.

4 Electrospray and Its Medical Applications

73

Acknowledgement This work was partially supported by a grant from the U.S. Department of Defense (AFOSR) MURI Grant (FA9550-04-1-0430) “Relationship between Physicochemical Characteristics and Toxicological Properties of Nanomaterials”.

References Aguirre-de-Carcer I, de la Mora JF (1995) Effect of background gas on the current emitted from Taylor cones. J Colloid Interface Sci 171:512–517 Bacher G, Szymanski WW, Kaufman SL, Zollner P, Blass D, Allmaier G (2001) Charge-reduced nano electrospray ionization combined with differential mobility analysis of peptides, proteins, glycoproteins, noncovalent protein complexes and viruses. J Mass Spectrom 36:1038–1052 Bailey AG (1988) Electrostatic spraying of liquids. Wiley, NY Biswas P, Wu CY (1998) Control of toxic metal emissions from combustors using sorbents: a review. J Air Waste Manage Assoc 48(2):113–127 Borra JP, Camelot D, Chou K-L, Kooyman PJ, Marijnissen JCM, Scarlett B (1999) Bipolar coagulation for powder production: micro-mixing inside droplets. J Aerosol Sci 30(7):945–958 Chen D, Pui DYH (1997) Experimental investigation of scaling laws for electrospraying: dielectric constant effect. Aerosol Sci Technol 27(3):367–380 Chen D, Pui DYH (2000) Electrospraying apparatus and method for introducing material into cells. US Patent 6,093,557 Chen CH, Emond MHJ, Kelder EM, Meester B, Schoonman J (1999) Electrostatic sol-spray deposition of nanostructured ceramic thin films. J Aerosol Sci 30:959–965 Chen D, Wendt CH, Pui DYH (2000) A novel approach for introducing bio-material into cells. J Nanoparticle Res 2:133–139 Chen X, Jia L, Yin X, Cheng J (2005) Spraying modes in coaxial jet electrospray with outer driving liquid. Phys Fluids 17 Clarke JDW, Jayasinghe S (2008) Bio-elecrosprayed multicellular zebrafish embryos are viable and develop normally. Biomed Mater 3(2008) art: 011001 Cloupeau M, Prunet-Foch B (1989) Electrostatic spraying of liquids in cone-jet mode. J Electrostat 22:135 Cloupeau M, Prunet-Foch B (1990) Electrostatic spraying of liquids: main functioning modes. J Electrostat 25:165 Cloupeau M, Prunet-Foch B (1994) Electrohydrodynamic spraying functioning modes: a critical review. J Aerosol Sci 25:1021 Coffee RA (1964) Electrostatic spraying of liquids and powders. Electron Power Oct:349–352 Coffee RA, Pirrie AB, Davies DN (2004) Inhaler, US Patent 6,684,879 B1 (02/03/2004) Cole RB (1997) Electrospray ionisation mass spectrometry. Wiley, New York de la Mora JF (1997) Charge and size distributions of electrospray drops. J Colloid Interface Sci 186:280–293 de la Mora JF, Loscertales IG (1994) The current emitted by highly conducting Taylor cones. J Fluid Mech 260:155 Dulcks Th, Juraschek R (1999) Electrospray as an ionization method for mass spectrometry. J Aerosol Sci 30:927–943 Dunn PF, Snarski SR (1991) Velocity component and diameter distribution characteristics of droplets within two interacting electrohydrodynamic sprays. Phys Fluids 3:494 Dunn PF, Grace JM, Snarski SR (1994) The mixing of electrically charged droplets between and within electrohydrodynamic fine sprays. J Aerosol Sci 25:1213 Eggers J (1997a) Nonlinear dynamics and beakup of free-surface flows. Rev Mod Phys 69 (3):865–927 Eggers J (1997b) Singularities in droplet pinching without viscosity, Los Alamos preprint No. chao-dyn/9705006

74

D.-R. Chen and D.Y.H. PUI

Fenn JB, Mann M, Meng CK, Wong SF (1989) Electrospray ionization for mass-spectrometry of large biomolecules. Science 246(4926):64–71 Gan˜a´n-Calvo AM (1994) The size and charge of droplets in the electrospraying of polar liquids in cone-jet mode, and the minimum droplet size. J Aerosol Sci 25(S1):309 Gan˜a´n-Calvo AM (1997) Cone-jet analytical extension of Taylor’s electrostatic solution and the asymptotic universal scaling laws in electrospraying. Phys Rev Lett 79:217–220 Gan˜a´n-Calvo AM (1999) The surface charge in electrospraying: its nature and its universal scaling laws. J Aerosol Sci 30(7):863–872 Gomez A, Tang K (1993) Charge and fission of droplets in electrostatic spraying. Phys Fluids 6:404 Hartman RPA, Brunner DJ, Camelot DMA, Marijnissen JCM, Scarlett B (2000) Jet break-up in the electrohydrodynamic atomization in the cone-jet mode. J Aerosol Sci 31(1):65–95 Hines RL (1966) Electrostatic atomization and spraying painting. J Appl Phys 37:2730–2736 Huberman MN, Beynon JC, Cohen E, Goldin DS, Kidd PW, Zafran S (1968) Present status of colloid microthruster technology. J Spacecraft 5:1319–1324 Iribarne JV, Thomson BA (1976) On the evaporation of small ions from charged droplets. J Chem Phys 64(6):2287–2294 Jaworek A, Krupa A (1999) Classification of the modes of EHD spraying. J Aerosol Sci 30(7):873 Jones AR, Thong KC (1971) The production of charged monodisperse fuel droplets by electrical dispersion. J Phys D: Appl Phys 4:1159 Kaufman SL (1998) Analysis of biomecules using electrospray and nanoparticle methods: the gasphas electrophoretic mobility molecular analyzer (GEMMA). J Aerosol Sci 29:537–552 Kaufman SL (1999) Molecular cluster observed using high resolution DMA. J Aerosol Sci 30 (Suppl 1):S821–S822 Kaufman SL, Skogen JW, Dorman FD, Zarrin F, Lewis KC (1996) Macromolecule analysis based on electrophoretic mobility in air: globular proteins. Anal Chem 68:1895–1904 Koropchak JA, Sadain SY, Magnusson X, Heybroek L-E, Anisimov M, Kaufman SL (1999) Nanoparticle detection technology for chemical analysis. Anal Chem 71(11):386–394 Krohn VE (1973) Evidence that the minimum-energy state is not accessible to a system of droplets produced by electrohydrodynamic spraying. Appl Phys Lett 23:220 Kruis FE, Fissan H, Peled A (1998) Sythesis of nanoparticles in the gas phase for electronic, optical and magnetic application–a review. J Aerosol Sci 29(5–6):511–535 Kwok A, Arumuganathar S, Irvine S, McEwan JR, Jayasinghe SN (2008) A hybrid bio-jetting approach for directly engineering living cells. Biomed Mater 3(2008) 025008 Langer G, Yamate G (1969) Encapsulation of liquid and solid aerosol particles to form dry powders. J Colloid Interface Sci 29:450–455 Lenggoro IW, Okuyama K, de la Mora JF, Tohge N (2000) Preparation of ZnS nanoparticles by electrospray pyrolysis. J Aerosol Sci 31:121–136 Lopez-Herrera JM, Barrero A, Lopez A, Loscertales IG, Marquez M (2003) Coaxial jets generated from electrified Taylor cones. Scaling laws. J Aerosol Sci 34:535–552 Loscertales IG, Barrero A, Guerrero I, Cortijo R, Marquez M, Ganan-calvo AM (2002) Micro/ nano encapsulation via electrified coaxial liquid jets. Science 295(5560):1695–1698 Mei F (2008) Dual-capillary electrospray: fundamental and application, PhD Dissertation. Washington University in St. Louis, MO Mei F, Chen D (2007) Investigation of compound jet electrosray: particle encapsulation. Phys Fluids 19:103303. doi:10.1063/1.2775976 Mei F, Chen D (2008) Operational modes of dual-capillary electrospraying and the formation of stable compound cone-jet mode. Accepted by Air Qual Aerosol Res April 29 Michelson D (1990) Electrostatic atomization. Adam Hilger, Bristol and New York Nawab MA, Mason SG (1958) The preparation of uniform emulsions by electrical dispersion. J Colloid Sci 13:179–187 Paul S (1985) Surface coating: science and technology, 1st edn (1985) and 2nd edn (1996), Wiley Baffins Lane, Chichester, West Sussex, England

4 Electrospray and Its Medical Applications

75

Pfeifer RJ, Hendricks C (1967) Charge-to-mass relationships for electrohydrodynamically sprayed liquid droplets. Phys Fluids 10:2149 Rosell-Llompart J, de la Mora JF (1994) Generation of monodisperse droplets 0.3 to 4 mm in diameter from electrified cone-jets of highly conducting and viscous liquids. J Aerosol Sci 25:1093 Ryce SA, Patriarche DA (1965) Energy considerations in the electrostatic dispersion of liquids. Can J Phys 43:2192 Scalf M, Westphall MS, Krause J, Kaufman LS, Smith LM (1999) Controlling charge states of large ions. Science 283:194–197 Setiawan ER, Heister SD (1997) Nonlinear modeling of an infinite electrified jet. J Electrostat 42:243–257 Shrimpton JS (2005) Dielectric charged drop break-up at sub-Rayleigh limit conditions. IEEE Trans Dielect Electr Insul 12(3):573–578 Smith RD, Bruce JA, Wu Q, Lei P (1997) New mass spectrometric methods for the study of noncovalent associations of bipolymers. Chem Soc Rev 26:191 Snarski SR, Dunn PF (1991) Experiments characterizing the interaction between two sprays of electrically charged liquid droplets. Exp Fluids 11:268–278 Stamatakis P, Natalie CA (1991) Research needs in aerosol processing. Aerosol Sci Technol 14:316–321 Taflin DC, Ward TL, Davis EJ (1989) Electrified droplet fission and the Rayleigh limit. Langmuir 5:376 Tang K, Gomez A (1994) Generation by electrospray of monodisperse water droplets for targeted drug delivery by inhalation. J Aerosol Sci 25(6):1237–1250 Thompson JF, Warsi ZUA, Mastin CW (1985) Numerical grid generation: foundations and applications. North-Holland, New York Tomita Y, Ishibashi I, Yokoyama T (1986) Fundamental studies on an electrostatic ink jet printer. Bull Jpn Soc Mech Eng 29:3737–3743 van Zomeren AA, Kelder EM, Marijnissen JCM, Schoonman J (1994) The production of thin films of LiMn2O4 by electrospraying. J Aerosol Sci 25(6):1229–1236 Vonnegut B, Neubauer RL (1952) Production of monodisperse liquid particles by electrical atomization. J Colloid Sci 7:616–622 Wolfgang GK, Manuela S-B, Winfried M (2006) Health implications of nanoparticles. J Nanoparticle Res 5(8):543–562 Yamashita M, Fenn JB (1984) Electrospray ion source: another variation on the free-jet theme. J Phys Chem 88:4451–4459

Chapter 5

Generation of Nanoparticles from Vapours in Case of Exhaust Filtration Markku Kulmala and Mikko Sipila¨

5.1

Introduction

Generation, investigation and manipulation of nanostructured materials are of fundamental and practical importance for several disciplines including materials sciences and medicine. Recently, atmospheric new particle formation in the nanometer size range has been found to be a global phenomenon (Kulmala et al. 2004). The processes related to nanomaterials and atmospheric nanoparticles are at least similar and in some cases even identical. However, the detailed mechanisms for nucleation and nanoparticle formation are mostly unknown, largely depending on the incapability to generate and measure nanoparticles in a controlled way. In recent experiments an organic vapour (n-propanol) condenses on molecular ions as well as charged and uncharged inorganic nanoparticles via initial activation by heterogeneous nucleation (Winkler et al. 2008). In these experiments a smooth transition in activation behaviour as a function of size has been found, and activation did occur well before the onset of homogeneous nucleation. Furthermore, nucleation enhancement for charged particles and a significant negative sign preference were quantitatively detected. While fresh aerosol particle formation has been observed to take place almost everywhere in the atmosphere (Kulmala et al. 2004), several gaps in our knowledge regarding this phenomenon still exist. These gaps range from the basic processlevel understanding of atmospheric aerosol formation to its various impacts on atmospheric chemistry, climate, human health and environment. Until recent years nucleation pathways have been poorly understood even though several different mechanisms have been suggested (Kulmala 2003; Kulmala et al. 2006). Main

M. Kulmala (*) and M. Sipila¨ Department of Physics, University of Helsinki, P.O. Box 64 (Gustaf Ha¨llstro¨minkatu 2), FI-00014, University of Helsinki, Finland Helsinki Institute of Physics, P.O. Box 64, FI-00014, University of Helsinki, Finland e-mail: [email protected]

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_5, # Springer ScienceþBusiness Media B.V. 2010

77

M. Kulmala and M. Sipila¨

78

reason for that has been the instrumental inability to detect particles below 3 nm in diameter. Direct measurements of both the nucleation process itself and the initial growth of the clusters are crucial in order to resolve the detailed pathways of the particle formation. Only very recently observations of atmospheric neutral particles and clusters below 3 nm have shed light on the first steps of particle formation (Kulmala et al. 2005, 2007a). Those observations were made using newly developed instruments designed for maximal detection efficiency of small clusters, like UF02proto swirling flow condensation particle counter pair (Mordas et al. 2005, 2008) and an Air Ion Spectrometer (Mirme et al. 2007) equipped with an aerosol charger i.e. Neutral Cluster and Air Ion Spectrometer (NAIS) (Kulmala et al. 2007a). Here we first describe the recent instrumentation for physical detection of nanoparticles and nanoclusters (Section 5.2). In Section 5.3 we describe aerosol generation, and in Section 5.4 we present how filtration will affect on generation of nanoparticles both theoretically and experimentally. The concluding remarks are given in Section 5.5.

5.2

Detection of Nanoparticles

Studying nanoparticles e.g. during atmospheric aerosol formation or particle generation in laboratories requires the measurements of both physical and chemical properties of nucleation mode particles (3–20 nm) and clusters (<3 nm) (e.g. Kulmala and Kerminen 2008). Recently, the typical cut size of aerosol instruments has decreased from 3 to 1.5 nm. Here, we give a brief summary of the relevant physical methods, their characteristics and limitations. CONDENSATION NUCLEUS COUNTER (CNC), or CONDENSATION PARTICLE COUNTER (CPC), is widely used in atmospheric studies. The principle of the CPC consists of three processes: (1) creation of a supersaturated vapour (working fluid), (2) growth of aerosol particles by condensation of these supersaturated vapours, (3) optical detection of the particles after their growth. Different types of CPCs have different techniques of creating vapour supersaturation. The most common type is the laminar flow chamber CPC, which has no movable parts and therefore is suitable for long-term atmospheric observations. Recent developments of CPCs in general have aimed, for instance, at improving the detection efficiency (Stoltzenburg and McMurry 1991) and response time (Sgro and Ferna´ndez de la Mora 2004). For a comprehensive summary of the history and principles of CPCs see e.g. McMurry (2000). There are several widely-used commercial alcohol-based CPCs, in which the supersaturation of the alcohol vapour is created by conductive cooling (for performance details, see, e.g., Sem 2002; Ha¨meri et al. 2002). Different research groups have tested and calibrated these instruments. Recently a new generation of commercial CPCs was introduced, in which the water supersaturation is achieved due to differences in the molecular vapour diffusivity and thermal diffusivity of air (see Hering and Stolzenburg 2005). Currently, the state-of-art commercial CPCs have

5 Generation of Nanoparticles from Vapours in Case of Exhaust Filtration

79

lowest detection limit at 2.5–3 nm (in diameter). Using water and butanol CPCs with different cut sizes as CPC Battery, the composition of fresh aerosol particles can determined (Kulmala et al. 2007b). In laminar flow CPCs the detection efficiency can be modified by changing the temperature difference between the saturator and condenser (or growth tube). In practice, particle size distributions have been measured by changing the temperature difference (Brock 1998; Kulmala et al. 2005, 2007b; Mordas et al. 2008). Brock (1998) developed a nuclei-mode aerosol size spectrometer by utilizing Kelvin-effect sizing to obtain fast-time response (~1 s) airborne measurements. Kulmala et al. (2005) performed studies by using an increasing temperature difference between the saturator and condenser to have different cut-off diameters in order to be able to detect neutral clusters. Other versions of the CPC technique have been recently used to detect also sub-3 nm particles, namely a PH-CPC (pulse height CPC) and an expansion CPC (ECPC). Earlier the pulse height analysis method has been used in size distribution measurements between 3 and 10 nm (Wiedensohler et al. 1994; Weber et al. 1995; 1998), as well as in determining the composition of freshly-nucleated nanoparticles (O’Dowd et al. 2002, 2004). Recently, Sipila¨ et al. (2008) used a PH-CPC together with a modified version of the expansion condensation particle counter, an instrument described by Ku¨rten et al. (2005), in order to investigate neutral clusters. The pulse height condensation particle counter (PH-CPC) used to detect sub-3 nm clusters comprises a TSI-3025A ultrafine CPC with modified optics and a multichannel analyzer (MCA) (Dick et al. 2000). The PH-CPC exploits an axial gradient of butanol supersaturation inside the CPC condenser (Stoltzenburg and McMurry 1991). Particles entering the condenser are activated for growth at different axial positions depending on their initial size, which leads to a monotonic link between the original particle size and the final droplet size. Bigger particles activate earlier and since they have more time to grow, they reach larger droplet sizes, whereas smaller particles have to travel further inside the condenser before activation and therefore they yield smaller droplets in the end. By measuring the final droplet size by optical methods one can thus conclude the initial particle size. This method works for initial particle sizes smaller than ca. 15 nm (Saros et al. 1996). As discussed earlier, the detection efficiency of a CPC can be improved by increasing the supersaturation. This will finally induce homogeneous nucleation inside the CPC condenser. However, the pulse height analysis method allows one to distinguish between homogeneously nucleated droplets, and droplet nucleated heterogeneously on particles or clusters. Therefore very high supersaturations can be used. DIFFERENTIAL MOBILITY PARTICLE SIZER (DMPS), or SCANNING MOBILITY PARTICLE SIZER (SMPS), is the most widely-used instrument for investigating the evolution of a particle size distribution. In these instruments, operation of an electrical classifier upstream of a CPC enables the measurement of particle size distributions. Differential mobility analysers (DMA) segregate particles in an electrical field, and yield particles of a narrow monodisperse electrical mobility (Knutson and Whitby 1975). DMAs are available in various

80

M. Kulmala and M. Sipila¨

designs, with recent developments focusing on a more efficient transmission of the smallest sizes <10 nm (Winklmayr et al. 1991; Chen et al. 1998) and improved resolution (e.g. Rosser and Ferna´ndez de la Mora 2005). A frequently used instrumental set-up of a Differential Mobility Particle Sizer (DMPS) in ground-based or ship-based experiments involves two DMAs covering a wide size range, such as 3 to 1000 nm, and two separate CNCs to count particles (e.g., Birmili et al. 1999; Aalto et al. 2001). The time required for measuring an atmospheric aerosol size distribution depends primarily on the time required to obtain a statistically significant number of CNC counts at each classifying voltage. A measurement period of 10 min provides a viable compromise between the size and time resolution and particle counting statistics for most atmospheric applications. DMA-CNC systems may also be operated as Scanning Mobility Particle Sizers (SMPS; Wang and Flagan 1990), whereby particle concentrations are measured as the classifying voltage is increased at a continuous rate. SMPS scan times as short as 2 min are possible, albeit in a trade-off against sizing accuracy and particle counting statistics. AIR ION MOBILITY SPECTROMETERS are an alternative class of instruments based on electric mobility analysis (e.g., Misaki 1964; Ho˜rrak et al. 1998; and references therein). Ion mobilities are segregated very similarly as in a DMPS, but an array of electrometers is typically used to simultaneously measure the various mobility fractions. Unlike the DMPS and SMPS systems which utilize bipolar chargers in order to bring the aerosol particle population into Boltzmann equilibrium before classification by the DMA, ion mobility spectrometers measure naturally-occurring mobility distributions. Ion mobility spectrometers can detect charged particles of any size, extending down to the size of molecular ions (ca. 0.4 nm). A limitation is that the sensitivity of electrometers limits the lowest detectable particle concentration to ~50 cm3. The two recently-developed and already quite widely-used instruments are the Balanced Scanning Mobility Analyzer (BSMA) (Tammet 2004) and the Air Ion Spectrometer (AIS) (Mirme et al. 2007), both manufactured by AIREL Ltd., Estonia. The BSMA consists of two plain aspiration-type differential mobility analyzers, one for positive and the other for negative ions. The electric mobility range of the BSMA is 0.032–3.2 cm2 V1 s1. The mobility distribution is converted into the size distribution using a specific algorithm (Tammet 1995), and the size distribution range 0.4–6.3 nm. The Air Ion Spectrometer (AIS) measures the mobility distribution of air ions (naturally charged clusters and aerosol particles). The spectrometer consists of two identical cylindrical aspiration-type differential mobility analyzers, one for measuring positive ions and the second for measuring negative ions. The mobility range of the AIS is 0.000752.4 cm2 V1 s1 and the corresponding diameter ranges of single charged particles calculated according to the algorithm by Tammet (1995) are 0.4655 nm, respectively. Based on the AIS technique, a Neutral cluster and Air Ion Spectrometer (NAIS) was very recently developed (Kulmala et al. 2007a). The NAIS is able to measure neutral clusters down to 1.2–1.5 nm (mass diameter based on the Tammet’s algorithm).

5 Generation of Nanoparticles from Vapours in Case of Exhaust Filtration

81

Recently, an ion-Differential Mobility Particles Spectrometer (ion-DMPS) and related techniques have been constructed in order to measure the charging state of particles in the diameter range 3–15 nm (e.g. Iida et al. 2006; Laakso et al. 2007). The charging state is defined as the ratio between the actual charged fraction of ambient aerosol particles and their respective stationary-state charged fraction. In principle, charging state measurements at multiple particle sizes provide quantitative information about the contribution of ion-induced nucleation to total atmospheric nucleation (Kerminen et al. 2007).

5.3

Aerosol Generation Methods

Accurate calibrations are necessary for reliable use of any nanoparticle detector. Methods of generation and classification of particles with diameters exceeding ~10 nm are well established and calibration experiments at that size range are reasonably straight forward. Typical calibration setup consists of particle generator, aerosol charger, DMA for size separation, and an electrometer for concentration reference. Tube furnace is widely used device for generation of 2–50 nm particles (e.g. Scheibel and Porstendo¨rfer 1983). It bases on evaporation and subsequent nucleation of material, for example silver, gold, or NaCl. Output of the tube furnace at sub-3 nm regime is very low and thus other methods are needed. Commercially available Grimm tungsten oxide generator has reasonably high output down to ca. 1.5 nm. Particles, however are mainly neutral and aerosol needs to be charged before it can be classified in a DMA. Charging probabilities at nanometer range are very low and therefore only a small fraction of the particles can be used. This problem can be overcome by utilization of a hot wire generator (Peineke et al. 2006) that produces self-charged particles from metal wires down to atomic dimensions (~1nm in electrical mobility diameter). This method is also much cleaner since aerosol chargers can produce condensable impurities thus changing the composition of the calibration aerosol particle (Ferna´ndez de la Mora et al. 1998) For nanoparticle research the mobility standards are of great importance. These standards include a group of tetra-alkyl ammonium halide salts (Ude and Ferna´ndez de la Mora 2005) that can be electrosprayed from alcohol solution. In an electrospray source liquid in high electrical potential is atomized to sub-micron droplets by strong electric field. These highly charged droplets start to evaporate and finally explode due to repulsive Coulombic forces. Resulting aerosol contains a series of sharp peaks associated to clusters formed from anions and cations of the sprayed salt dominated by the bare cation. These peaks have well known mobilities corresponding to mobility equivalent diameters from 1.05 up to ca 3.5 nm and masses ranging from 74 amu up to thousands amu’s. High diffusivity of nanometric particles sets certain demands on classification methods. Ordinary Vienna-type DMA (Winklmayr et al. 1991) is capable to classify particles with reasonably good resolution down to particle diameters of ~10 nm. Resolution gets worse with decreasing particle size. Therefore effort has been put on

M. Kulmala and M. Sipila¨

82

developing DMAs with excellent resolution down to molecular sizes (see e.g. de Juan and Ferna´ndez de la Mora 1998; Ferna´ndez de la Mora et al. 1998; Herrmann et al. 2000; Rosser and Ferna´ndez de la Mora 2005). With a very high resolution Herrmann type DMA (Herrmann et al. 2000) the resolution of less than 2% (transfer function’s full width at half maximum in mobility space) at 1.5 nm can be achieved. Common feature in all high resolution DMA’s are that they can be used with Reynolds numbers from 5,000 up to even 35,000 before the flow becomes turbulent. Faraday cage electrometers (FCE) are commonly used as concentration reference for classified particles. Electrometer measures the electrical current carried by the aerosol. If the flow through the electrometer is high enough (usually ~10 Lpm) so that diffusion losses of nanometer sized particles can be neglected, and if particles are singly charged, the concentration can be immediately calculated. With the methods described above instruments like AIS, NAIS, ultrafine CPCs, pulse height CPC, and expansion-CPC can be calibrated accurately and reliably.

5.4

Aerosol Formation and the Effect of Filtration

We investigated experimentally the effect of existing particles on cluster concentrations in our laboratory in Helsinki in November 2007. The typical daily evaluation of size distribution and total particle concentration measured by DMPS is shown in Fig. 5.1. The corresponding cluster spectra measured by PH-CPC is given in Fig. 5.2. As seen in Fig. 5.2 the concentration of freshly nucleated clusters anticorrelate with the particle (5–1000 nm) concentration. As shown in Fig. 5.2 we are able to detect particles well below 3 nm. The detection of clusters down to 1.5 nm (mobility diameter) is possible both in laboratory and atmospheric conditions. The detection of size and concentration of clusters by PH-CPC is based on observation of homogeneous nucleation inside PH-CPC. However, in the case of huge nucleation bursts or polluted air the homogeneous nucleation will disappear. In Fig. 5.3 the suppression of particle formation inside CPC is shown as function of condensation sink. The filtration of existing particles as well as filtration of condensable gases will affect on particle formation. Actually, we are unable to measure the true nucleation rate, J*, that produces clusters of diameter d*, but rather the formation rate of particles of some larger diameter dp, termed J(dp). The diameter dp corresponds typically to the CPC (condensation particle counter) detection limit, which is presently 3 nm or greater. The relation between J* and J(dp) can be estimated based on the theory describing the competition between condensation growth and cluster scavenging (Kerminen and Kulmala 2002):     1 1 CS0  Jðdp Þ ¼ J exp g dp d  GR 

(5.1)

5 Generation of Nanoparticles from Vapours in Case of Exhaust Filtration

83

Fig. 5.1 Particle size distribution (above) and total number concentration measured by differential mobility particle sizer (DMPS)

Here CS0 (unit m2) is directly proportional to the condensation sink CS (CS ¼ 4pD  CS0 , where D is the gas-phase diffusion coefficient of the condensing vapour; see also Kulmala et al. 2001) caused by the pre-existing aerosol particle population and GR (nm h1) is the growth rate of nucleated particles. The parameter g depends on many factors but can usually be approximated by assuming it to be equal to 0.23 nm2 m2 h1. The relation given by Eq. 5.1 assumes implicitly that nucleated clusters grow in size at a constant rate, which may not be the case in the atmosphere. A more general form of Eq. 5.1, allowing GR to depend on the particle size, has been introduced by Kerminen et al. (2004). Lehtinen et al. (2007) modified Eq. 5.1 slightly further by starting from fundamental aerosol dynamical arguments. Also the vapour concentration (C) depends on source (Q) and existing condensation sink (CS) (Kulmala et al. 2001). dC ¼ Q  CS  C dt

(5.2)

In steady state this is C¼

Q CS

(5.3)

M. Kulmala and M. Sipila¨

84

Fig. 5.2 Size distribution of clusters (above) and total concentration of clusters (1.5–3 nm) and particles (>5 nm). Clusters measured by Pulse Height CPC

On the other hand growth rate is (Kulmala et al. 2001) GR ¼ A0  C

(5.4)

Therefore the production is well connected to filtration: with filtration we will decrease the CS and therefore both C and J(dp) are increasing. This can be seen from Fig. 5.4. If the sulphuric acid concentration without filtration is e.g. 107 cm3. After 90% of particles are filtrated out the vapour concentration will increase by a factor of 10, and formation rate of 3 nm particles ca 10 orders of magnitude and 100 nm particle more than 50 orders of magnitude. The corresponding results can also be obtained using PH-CPC with different CS values (filtration levels). The nucleation inside CPC is decreasing as function of CS (Fig. 5.3). If we make filtration more actively, we will obtain enhanced nucleation rates inside CPC.

5 Generation of Nanoparticles from Vapours in Case of Exhaust Filtration

85

Fig. 5.3 Detected particle concentration due to homogenous nucleation inside PH-CPC as a function of condensation sink

Fig. 5.4 The effect of filtration on nucleation rate ( j2) and on particle formation at 3 nm ( j3) and at 100 nm ( j100) as a function of sulphuric acid concentration. The particle formation mechanism is activation of clusters

M. Kulmala and M. Sipila¨

86

The filtration (decreased sinks) will also modify cluster concentrations. The evolution of ion (n) and neutral (N) concentrations can be given by following equation pair: dn ¼ Q  an2  nCoagSions dt dN ¼ J þ ean2  NCoagS dt Here Q = ion production rate, J = particle formation rate, a = recombination rate, CoagS = coagulation sink We made following assumptions: (1) CoagSions = CoagS, (2) e  1, (3) steady state. Then we obtain estimations of neutral particle concentrations as a function of ion concentrations: N¼

J þ an2 CoagS

As seen by decreasing the sink term we also increase the cluster concentration. Therefore we can state that if sink decreases due to filtration by a certain factor N is increasing by the same factor.

5.5

Conclusions

At small and moderate particle concentrations the more number the more mass we have. However, when the mass concentration is high enough the number concentration starts to decrease due to the suppressed new particle formation. Actually if the mass concentration starts to decrease e.g. due to filtration the number concentration starts to increase significantly as seen from Figs. 5.3 and 5.4. The main effects of filtration are that both the number and mass of existing particles will decrease. When the sink values are decreasing the vapour concentration increases and the new particle formation will be enhanced significantly. At the same time also coagulation scavenging decreases and the number concentration of newly formed particles increases rapidly. As a summary we can state that the particle generation in different exhaust plumes can be mastered using different filtration methods and levels of effectiveness. As shown in Section 5.4 the effect can be even more than 10 orders of magnitude. Recently the detection limit of particles has decreased from 3 to 1.2–1.8 nm, and actually atmospheric nucleation has been observed around 1.5–2 nm (Kulmala et al. 2007a). Also the atmospheric clusters around size of ~1.5–1.8 nm (700–1000 amu) has been detected (Sipila¨ et al. 2008). These new findings are based on (a) better and

5 Generation of Nanoparticles from Vapours in Case of Exhaust Filtration

87

more accurate particle generation methods used in detector calibrations, (b) development of new instruments like NAIS/AIS/BSMA and (c) improving CPC techniques with qualitative (Kulmala et al. 2007b) and quantitative (Winkler et al. 2008) understanding of heterogeneous nucleation. In practice it would be possible to apply these knowledge and experience also to other fields of aerosol science. Acknowledgements This work has been partially funded by European Commission 6th Framework programme project EUCAARI, contract no 036833-2 (EUCAARI). Maj and Tor Nessling foundation and the Academy of Finland are also acknowledged for financial support.

References Aalto P, Ha¨meri K, Becker E, Weber R, Salm J, Ma¨kela¨ JM, Hoell C, O’Dowd CD, Karlsson H, Hansson H-C, Va¨keva¨ M, Koponen IK, Buzorius G, Kulmala M (2001) Physical characterization of aerosol particles during nucleation events. Tellus 53B:344–358 Birmili W, Stratmann F, Wiedensohler A (1999) Design of a DMA-Based size spectrometer for a large particle size range and stable operation. J Aerosol Sci 30:549–553 Brock CA (1998) A fast-response nuclei mode spectrometer for determining particle size distribution in the 3–100 nm diameter range: technical description, Technical Report, University of Denver, Denver, CO Chen D-R, Pui DYH, Hummes D, Fissan H, Quandt FR, Sem GJ (1998) Design and evaluation of a nanometer aerosol differential mobility analyzer (nano-DMA). J Aerosol Sci 29:497–509 de Juan L, Ferna´ndez de la Mora J (1998) Size analysis of nanoparticles and ions: running a Vienna DMA of near optimal length at Reynolds numbers up to 5000. J Aerosol Sci 29:617–626 Ferna´ndez de la Mora JF, de Juan L, Eichler T, Rosell J (1998) Differential mobility analysis of molecular ions and nanometer particles. Trends Anal Chem 17:328–339 Dick WD, McMurry PH, Weber RJ, Quant R (2000) White-light detection for nanoparticle sizing with the TSI ultrafine condensation particle counter. J Nanoparticle Res 2:85–90 Ha¨meri K, Koponen IK, Aalto PP, Kulmala M (2002) The particle detection efficiency of the TSI3007 condensation particle counter. J Aerosol Sci 33:1463–1469 Hering SV, Stoltzenburg MR (2005) A method for particle size amplification by water condensation in a laminar, thermally diffusive flow. Aerosol Sci Technol 39:428–436 Herrmann W, Eichler T, Bernardo N, Ferna´ndez de la Mora JF, Turbulent transition arises at Reynolds number 35,000 in a short Vienna Type DMA with a large laminarization inlet. Abstract AAAR Conference, 15B5, 2000 Ho˜rrak U, Salm J, Tammet H (1998) Bursts of intermediate ions in atmospheric air. J Geophys Res 103:13909–13915 Iida K, Stolzenburg M, McMurry P, Dunn M, Smith J, Eisele F, Keady P (2006) Contribution of ion-induced nucleation to new particle formation: methodology and its application to atmospheric observations in boulder. Colorado J Geophys Res 111:D23201. doi:10.1029/ 2006JD007167 Kerminen V-M, Kulmala M (2002) Analytical formulae connecting the “real” and the “apparent” nucleation rate and the nuclei number concentration for atmospheric nucleation events. J Aerosol Sci 33:609–622 Kerminen V-M, Anttila T, Lehtinen KEJ, Kulmala M (2004) Parameterization for atmospheric new-particle formation: application to a system involving sulfuric acid and condensable watersoluble organic vapors. Aerosol Sci Technol 38:1001–1008 Kerminen V-M, Anttila T, Peta¨ja¨ T, Laakso L, Gagne´ S, Lehtinen KEJ, Kulmala M (2007) Charging state of the atmospheric nucleation mode: implications for separating neutral and ion-induced nucleation. J Geophys Res 112:D21205. doi:10.1029/2007JD008649

88

M. Kulmala and M. Sipila¨

Knutson EO, Whitby KT (1975) Aerosol classification by electric mobility: apparatus, theory, and applications. J Aerosol Sci 6:443–451 Kulmala M (2003) How particles nucleate and grow? Science 302:1000–1001 Kulmala M, Kerminen V-M (2008) On the formation and growth of atmospheric nanoparticles. Atmos. Res. 90:132–150 Kulmala M, Dal Maso M, Ma¨kela¨ JM, Pirjola L, Va¨keva¨ M, Aalto PP, Miikkulainen P, Ha¨meri K, O’Dowd CD (2001) On the formation, growth and composition of nucleation mode particles. Tellus 53B:479–490 Kulmala M, Vehkama¨ki H, Peta¨ja¨ T, Dal Maso M, Lauri A, Kerminen V-M, Birmili W, McMurry PH (2004) Formation and growth rates of ultrafine atmospheric particles: a review of observations. J Aerosol Sci 35:143–176 Kulmala M, Lehtinen KEJ, Laakso L, Mordas G, Ha¨meri K (2005) On the existence of neutral atmospheric clusters. Boreal Environ Res 10:79–87 Kulmala M, Lehtinen KEJ, Laaksonen A (2006) Cluster activation theory as an explanation of the linear dependence between formation rate of 3 nm particles and sulphuric acid concentration. Atmos Chem Phys 6:787–793 Kulmala M, Mordas G, Peta¨ja¨ T, Gro¨nholm T, Aalto PP, Vehkama¨ki H, Hienola AI, Herrmann E, Sipila¨ M, Riipinen I, Manninen H, Ha¨meri K, Stratmann F, Bilde M, Winkler PM, Birmili W, Wagner PE (2007a) The condensation particle counter battery (CPCB): a new tool to investigate the activation properties of nanoparticles. J Aerosol Sci 38:289–304 Kulmala M, Riipinen I, Sipila¨ M, Manninen HE, Peta¨ja¨ T, Junninen H, Dal Maso M, Mordas G, Mirme A, Vana M, Hirsikko A, Laakso L, Harrison RM, Hanson I, Leung C, Lehtinen KEJ, Kerminen V-M (2007) Toward direct measurement of atmospheric nucleation. Science 318:89–92. 10.1126/science.1144124 Ku¨rten A, Curtius J, Nillius B, Borrmann S (2005) Characterization of an automated, water-based expansion condensation nucleus counter for ultrafine particles. Aerosol Sci Technol 39: 1174–1183 Laakso L, Gagne S, Peta¨ja¨ T, Hirsikko A, Aalto PP, Kulmala M, Kerminen V-M (2007) Detecting charging state of ultrafine particles: instrumental development and ambient measurements. Atmos Chem Phys 7:1333–1345 Lehtinen KEJ, Dal Maso M, Kulmala M, Kerminen V-M (2007) Estimating nucleation rates from apparent particle formation rates and vice-versa: revised formulation of the Kerminen– Kulmala equation. J Aerosol Sci 38:988–994 McMurry PH (2000) A review of atmospheric aerosol measurements. Atmos. Environ. 34:1959–1999 Mirme A, Tamm E, Mordas G, Vana M, Uin J, Mirme S, Bernotas T, Laakso L, Hirsikko A, Kulmala M (2007) A wide-range multi-channel air ion spectrometer. Boreal Environ Res 12:247–264 Misaki M (1964) Mobility spectrums of large ions in the New Mexico semidesert. J Geophys Res 69:3309–3318 Mordas G, Kulmala M, Peta¨ja¨ T, Aalto PP, Matulevicius V, Grigoraitis V, Ulevicius V, Grauslys V, Ukkonen A, Ha¨meri K (2005) Design and performance characteristics of a condensation particle counter UF-02proto. Boreal Environ Res 10:543–552 Mordas G, Sipila¨ M, Kulmala M (2008) Nanometer particle detection by the condensation particle counter UF-02 proto. Aerosol Sci Technol 42:521–527 O’Dowd CD, Aalto PP, Ha¨meri K, Kulmala M, Hoffmann T (2002) Atmospheric particles from organic vapours. Nature 416:497–498 O’Dowd CD, Aalto PP, Yoon YJ, Ha¨meri K (2004) The use of the pulse height analyzer ultrafine condensation particle counter (PHA-UCPC) technique applied to sizing of nucleation mode particles of differing chemical composition. J Aerosol Sci 35:205–216 Peineke C, Attoui MB, Schmidt-Ott A (2006) Using a glowing wire generator for production of charged, uniformly sized nanoparticles at high concentrations. J Aerosol Sci 37:1652–1661 Rosser S, Ferna´ndez de la Mora J (2005) Vienna type DMA of high resolution and high flow rate. Aerosol Sci Technol 39:1191–1200

5 Generation of Nanoparticles from Vapours in Case of Exhaust Filtration

89

Saros M, Weber RJ, Marti J, McMurry PH (1996) Ultra fine aerosol measurement using a condensation nucleus counter with pulse height analysis. Aerosol Sci Technol 25:200–213 Scheibel HG, Porstendo¨rfer J (1983) Generation of monodisperse Ag- and NaCl-aerosols with particle diameters between 2 and 300 nm. J Aerosol Sci 14:113–126 Sem GJ (2002) Design and performance characteristics of three continuous-flow condensation particle counters: a summary. Atmos Res 62:267–294 Sgro LA, Ferna´ndez de la Mora J (2004) A simple turbulent mixing CNC for charged particle detection down to 1.2 nm. Aerosol Sci Technol 38:1–11 Sipila¨ M, Lehtipalo K, Kulmala M, Peta¨ja¨ T, Junninen H, Aalto PP, Manninen HE, Vartiainen E, Riipinen I, Kyro¨ E-M, Curtius J, Ku¨rten A, Borrmann S, O’Dowd CD (2008) Applicability of condensation particle counters to measure atmospheric clusters. Atmos Chem Phys 8: 4049–4060 Stoltzenburg MR, McMurry PH (1991) An ultrafine aerosol condensation nucleus counter. Aerosol Sci Technol 14:48–65 Tammet H (1995) Size and mobility of nanometer particles, clusters and ions. J Aerosol Sci 26:459–475 Tammet H (2004) Balance scanning mobility analyzer BSMA. In: Kasahara M, Kulmala M (eds) Nucleation and atmsopheric aerosols 2004, 16th Interantional Conference, Kyoto 2004, pp 294–297 Ude S, Ferna´ndez de la Mora J (2005) Molecular monodisperse mobility and mass standards from electrosprays of tetra-alkyl ammonium halides. J Aerosol Sci 36:1224–1237 Wang SC, Flagan RC (1990) Scanning electrical mobility spectrometer. Aerosol Sci Technol 13:230–240 Weber RJ, McMurry PH, Eisele FL, Tanner DJ (1995) Measurement of expected nucleation precursor species and 3–500-nm diameter partciles at Mauna Loa observatory Hawaii. J Atmos Sci 52:2242–2257 Weber RJ, Stolzenburg MR, Pandis SN, McMurry PH (1998) Inversion of ultrafine condensation nucleus counter pulse height distributions to obtain nanoparticle ( 3–10 nm) size distributions. J Aerosol Sci 29:601–615 Wiedensohler A, Aalto P, Covert D, Heintzenberg J, McMurry PH (1994) Intercomparison of four methods to determine size distributions of low-concentration ( 100 cm3), ultrafine aerosols (3 < Dp < 10 nm) with illustrative data from the arctic. Aerosol Sci Technol 21:95–109 Winkler PM, Steiner G, Vrtala A, Vehkama¨ki H, Noppel M, Lehtinen KEJ, Reischl GP, Wagner PE, Kulmala M (2008) Heterogeneous nucleation experiments bridging scale from molecular ion clusters to nanoparticles. Science 319:1374–1377 Winklmayr W, Reischl GP, Linde AO, Berner A (1991) A new electromobility spectrometer for the measurement of aerosol size distributions in the size range from 1 to 1000 nm. J Aerosol Sci 22:289–296

Chapter 6

Measurement and Characterization of Aerosol Nanoparticles Wladyslaw W. Szymanski and Gu¨nter Allmaier

6.1

Introduction

Aerosol nanoparticles are abundant in natural and man-made environment. Air quality studies and health impact of particles, in particular of nanoparticles, are being conducted with an increasing level of interest in the academic and also nonacademic community. The potential risks to the health and environment of engineered nanoparticles and consequently nanomaterials become increasingly the focus of general attention in recent years (Pui and Chen 1997; Royal Society and Royal Academy of Engineering 2004; Maynard and Pui 2007; Gazso et al. 2007). However, so far no standard procedures for measurement of aerosol nanoparticles and their properties were attempted to invoke. Moreover, there is even no standard definition regarding nanoparticles, not to mention aerosol nanoparticles. Definitions, as well as the upper limit in terms of particle size may vary depending on the source of information and application. The Publicly Available Specification (PAS) in the UK proposed that a nanoparticle is a particle “having one or more dimensions of the order of 100 nm or less”. This seems to be a useful starting point and this value of 100 nm will be used in this contribution as the upper size limit for nanoparticles; however this “static” definition does not touch on the difficulty of definition for airborne nanoparticles – the nanoaerosol. Description of a nanoaerosol should include besides static descriptors referring to geometry (diameter, surface area, fractal dimension, morphology, or mass), also its temporal changes, such as concentration, or degree of agglomeration; hence an observation time interval next to “geometric” values would be needed as a parameter.

W.W. Szymanski (*) and G. Allmaier University of Vienna, Faculty of Physics, Aerosol Physics, Biophysics and Environmental Physics Division, Boltzmanngasse 5, A-1090, Vienna, Austria e-mail: [email protected] Vienna University of Technology Institute of Chemical Technologies and Analytics, Getreidemarkt 9/164-IAC, A-1060, Vienna, Austria

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_6, # Springer ScienceþBusiness Media B.V. 2010

91

92

W.W. Szymanski and G. Allmaier

The metric for a nanoparticle in terms of size (diameter) only is somewhat weak. This has been already stated in a seminal publication by Preining (1998) devoted to the nature of “very, very small particles”. Bearing in mind that one nanometer corresponds to about 4 to 5 atomic diameters, particles with sizes of the order few nanometers contain several hundreds of atoms only. This brings with itself a fact of the increasingly larger surface to volume ratios, which in consequence influences the physical and/or chemical properties of nanoaerosols without changes of the chemical composition of the particle itself. Quantum size effects, material stability, conductivity, magnetic, or optical properties can be varied through the changes of dimension or structure and were intriguing scientists already more than a century ago, so in some sense the science of nanoparticles is not novel. Scientific investigations of colloids containing nanoparticles and their properties were probably first reported by Faraday (1857) about his experiments with gold suspensions using focussed sun light as a real-time monitoring tool, facts showing his utmost brilliance. He produced nanoparticles from gold leaf films and also using electrical discharge by means of a Layden battery - a method often used nowadays and called “exploding wire particle generator” (Kotov 2003; Aranchuk et al. 2004). Faraday postulated 1857 that many of those particles in the investigated suspensions are in metallic, yet “extreme divided state” and are too small to be observed by his microscopic possibilities. He actually estimated the particle size to be “very small as compared to the wave-length” of light. Based on his optical measurements he deduced that in “all these cases the ruby tint (of the colloid) is due simply to the presence of diffused finely-divided gold”, a fact which is now understood to comprise monodisperse particles in the size range below 50 nm. Somewhat later Zsigmondy (1905) describes his experiments with gold particles in liquids. With his newly developed “Ultramicroscope” he investigated the movement of colloidal gold particles, which he synthesised, and deduced that his gold hydrosols are extremely stable, and that in this “red gold sols” particles must be of same sizes ranging from 6 to 20 nm. Zsigmondy, who in 1925 received a Nobel Price for his research on colloids, said in his Nobel Lecture: “if I had known of Faraday’s results, it would have saved me much unnecessary work”. Despite the early interest in dispersed systems the real-time measurement capability of aerosol nanoparticles still needs further development. Especially high efficiency detection and characterisation for nanoaerosols containing nanoparticles below about 10 nm in size, along with invention of new techniques and/or improvement and standardization of existing methods for determination of particle size, concentration, morphology, structure and, if possible chemical and biological characterisation are posing a constant challenge. The posed questions are of interdisciplinary character leading to an increasing interdependence of particular research fields which are consequently merging into the nanospace determined by the particle size and time (Fig. 6.1). The drive towards the “nanospace” over the past decade, or so, promoted a rapid development in a field which can be regarded as truly interdisciplinary, and have resulted in collaboration between researchers in previously different areas using

6 Measurement and Characterization of Aerosol Nanoparticles

93

Fig. 6.1 Development of methods and techniques on their way into the “Nanospace”

different tools and techniques. An understanding of physics and chemistry of matter and processes at the nanoscale is seemingly relevant to many scientific disciplines, from chemistry and physics to engineering, biology and medicine. As always, the eyes and fingers of scientists are their instruments, which is also very true when it comes to nanoaerosols.

6.2

Measurement of Particle Size Distribution and Number Concentration

Among first steps towards understanding of aerosol systems are experimental procedures. The application of an appropriate methodology and measurement of relevant parameters to obtain information and insight into dynamic processes governing the behaviour of nanoparticles is the essential issue. Probably most frequently measured parameters of an aerosol are the size distribution of particles and their concentration, which is mirrored in a number of workshops devoted to that issue over past years (e.g. WUFA 1979; Liu et al. 1982; Ankilov et al. 2002; Gras et al. 2002; NARSTO 2007). When it comes to real-time measurement of nanoparticles electrostatic techniques utilizing electrical mobility of a charged particle in a defined electrostatic field seems to be among the best and convenient approaches currently available. This method was primarily developed to measure ions in gases and in the atmosphere (Erikson 1922; Zeleny 1929). Likely one of first who utilized this technique for observation of airborne particles was Rohmann (1923). In past years the electrical mobility approach to measure aerosols was continuously developed and refined (Knutson and Whitby 1975; Liu and Pui 1975; Winklmayr et al. 1991) allowing now precise measurement of particles even below 1 nm in size in terms of the electrophoretic mobility (Reischl 2008).

94

6.2.1

W.W. Szymanski and G. Allmaier

Electrostatic Particle Manipulation and Quantification of a Nanoparticle Ensamble

The working principle of a so-called differential mobility analyzer (DMA) is shown in Fig. 6.2. The usual arrangement is a co-axial placement of a cylindrical outer and inner electrode with radii R2 and R1, with an electrostatic field between them, which is defined by the voltage V applied to the inner electrode. Orthogonally to the electrostatic field there a constant, laminar flow (QS) of a particle-free air. Insertion of small air flow (QA) containing charged nanoparticles between the electrodes results in their movement, which is determined by both – the laminar air flow and the electric field. When the voltage V applied to the inner electrode is continuously changed, particles with different, well-defined electrophoretic mobilities Z (DP) corresponding to particles sizes DP will leave the DMA system at the particle outlet placed at the separation length L (Eq. 6.1) and can be consequently monitored and analyzed (Chen and Pui 1997).

ZðDP Þ ¼

Qtot  ln

  R2 R1

2p  V  L

(6.1)

Nanoparticles contained in the exiting flow (QEX) can not be observed and measured directly due to their very small dimensions. Basically two techniques allowing comfortable detection of nanoparticles are now commonly used: the condensation technique combined with optical detection and the electrical technique.

Fig. 6.2 Scheme showing the working principle of a differential mobility analyzer (DMA)

6 Measurement and Characterization of Aerosol Nanoparticles

a

95

b

Nanoaerosol input

Current measurement

Working vapor

Nanoaerosol input

Airflow out

Electrode

Optical detection of micrometer sized droplets

Charge transfer

Housing

Ceramic Insulator

Fig. 6.3 (a) Principle of the CPC function, (b) principle of the AE function

The first method is used in the Condensation Particle Counters (CPC), the second method, in various Aerosol Electrometers (AE) (Fig. 6.3a and b). In a CPC single nanoparticles are mixed with a working vapour (typically alcohol, water) which is then supersaturated due to e.g. rapid cooling of this mixture. The resulting droplet formed on the particle in question grows to a super-micrometer size and can be then easily optically monitored and counted. In an Aerosol Electrometer (AE) (sometimes called also the Faraday Cup) nanoparticles leaving a DMA will be detected resulting in an electrical current signal which can be straightforwardly converted into particle number concentration. The combination of single particle detector with a DMA operated in the scanning mode (voltage on the central electrode of the DMA (Fig. 6.2) is continuously varied) becomes increasingly popular device for nanoparticle size distribution measurement (Kinney et al. 1991; De la Mora et al. 1998; Collins et al. 2000; Bacher et al. 2001). However, although in principle straightforward the DMA technique seems to require a calibration with well-defined size or molecular mass standards for analysis of nanoaerosols including airborne macromolecules, proteins and polymers, particularly if a conversion of DMA data to molecular mass is attempted (Laschober et al. 2006, 2007). In recent years several designs of DMAs have been reported in the literature (Zhang and Flagan 1996; Reischl et al. 1997; Mesbah et al. 1997; Chen and Pui 1999). Also other electrical mobility spectrometers, e.g. Electrical Aerosol Spectrometer of Tartu University (Tammet 1995) and the Fast Mobility Particle Sizer (Biskos et al. 2005a) have been developed recently all of them with the focus on nanoparticle measurement.

96

W.W. Szymanski and G. Allmaier

6.2.1.1

Particle Generation and Charge Conditioning

The “condicio sine qua non” for utilization of electrostatic means for nanoparticle measurement is a known, well-defined and experimentally repeatable charge level on particles in question. Ideally, a highly efficient charging process resulting in a single charge per particle and independent of its size would be desirable. Unfortunately the currently known particle charging methods do not deliver such result. There are however few approaches to condition the charging status of aerosol particles. Likely the most widely spread method is the diffusion charging in a bipolar ion atmosphere, which is obtained by a weak radioactive source, typically Kr-85 (b radiation), Am-241, or Po-210 (both a radiation). The simplicity of use is counterbalanced by the difficulty of handling radioactive materials and by the extremely low charging probability for particles with sizes smaller than 20 nm (Fig. 6.4). As can be seen in Fig. 6.4 only about 10% of particles with 20 nm in size are charged. For 5 nm particles the charging probability drops to only about 1%, showing clearly how ineffectively this charging process works in this size range. However, particles with sizes below 100 nm, which were successfully charged, carry, from practical point of view, a single charge only due to the fact that double charging in this size range has very low probability. Another possibility to manipulate the charging level of aerosol particles is to generate ions due to a corona discharge, which can be for example obtained applying voltages of the order of few thousand volts to very thin wires (active electrode with high curvature), with the other electrode (low curvature) typically grounded. The corona discharge is a process, where the geometry of electrodes confines the gas ionizing processes to high electric field ionisation regions around 1.00000 Singly Charged

Charging Probability

0.10000

Doubly Charged

0.01000

0.00100

0.00010

0.00001 0.10

1.00 10.00 Particle Diameter [ nm ]

100.00

Fig. 6.4 Diffusion charging probability for nanoparticles based on the Fuchs–Boltzmann charge equilibrium approach (Reischl et al. 1997)

6 Measurement and Characterization of Aerosol Nanoparticles

97

the active electrode. The coronas can be positive, negative, bipolar, or high frequency alternating current (Biskos et al. 2005c; Goldman et al. 1985). The area of interest for charging purposes is the so-called ion drift region connecting the ionisation region with the passive electrode. Here the ions, or electrons drift and react with available particles, but with energy too low to ionize and too low density to react with other ionized particles. In unipolar coronas the drift region contains drifting ions with the corona polarity. Processes of particle charge manipulation are determined by the Nt-product (N – mean ion number concentration and t – mean aerosol residence time), which is a fundamental parameter to describe, judge and optimize the performance of any aerosol charger (Alonso et al. 2006). The Nt-values of about 107 are needed for stable charging conditions. The Nt-product combines geometrical dimension with the operational parameters and ion source production effectiveness in a charge conditioning device. An experimental arrangement utilizing the corona discharge to influence the charge levels of aerosols produced by means of an electrospray is shown in Fig. 6.5a. The device consists of an electrospray nanoparticle generator producing positively charged particles (chamber A )and a negative corona discharge (chamber B). Mixing of electrosprayed particles and their charge conditioning occurs in chamber C. These charging processes have been recently investigated (Laschober et al. 2006; Vivas et al. 2008). Results show that charging levels for nanoparticles well above the values achieved with radioactive sources are possible and they can be influenced by the operational parameter of the corona unit, such as flow rates, linked to particles’ interaction time with ions/electrons, or to the corona current resulting in the concentrations of generated number concentrations of charged species. The sketch (Fig. 6.5b) visualizes the process of charge conditioning in a setup shown in Fig. 6.5a. The presented arrangement makes use of the negative corona. Depending on the application various coronas can be considered. Simultaneous generation and interaction of corona ions with electrosprayed particles is a necessary process to provide the functioning of such a unit. The absence of corona electrons/ions results to a complete particle removal due their high initial charge and space charge effects. The utilization of corona process for particles below 100 nm seems to provide besides neutral particles essentially singly charged particles, which is plausible considering the outcome of a diffusion charging process (Fig. 6.4). The analysis of data presented in Fig. 6.6 indicates that corona discharge and a subsequent charge conditioning can be unquestionably more effective process, when compared with the bipolar diffusion charging occurring in a system driven by a radioactive source. The original high electrostatic charge of the ES-generated particles can be successfully controlled by the corona discharge process. In the presented arrangement a sort of “charge titration process” takes place. At low corona voltage values mainly positively charged particles will be exiting from the mixing unit. Increasing the corona current decreases continuously the amount of electrically negative particles – the dashed line in Fig. 6.6 corresponds the Boltzmann–Fuchs charge equilibrium. Simultaneously with the increasing applied voltage an increasing amount of particles with be eventually negatively

98

W.W. Szymanski and G. Allmaier

Fig. 6.5 (a) Schematics of an experimental Corona-Electrospray device, (b) visualisation of particle charge conditioning due to mixing of unipolarly, positively charged particles with electrons generated in a corona discharge

charged. As mentioned before the probability for doubly charged particles in the observed size range is practically negligible and was not observed (Laschober et al. 2006). Apparently it is possible to optimize a performance of such a device in terms of number concentration of charged particles, the drawback however which is linked with corona charging process is the fact that charging probability varies not only with operational parameters but seems also to be dependent on the kind/material of particles in question, a fact which is shown in Fig. 6.6. Although by and large increased charging efficiencies can be obtained (Fig. 6.6), much lesser correlation between particle size and the enhanced charging particle fraction of unipolar charging compared to bipolar charging was observed, to some degree in contrast to results reported in the literature (Chaolong et al. 2007). This may be caused by several factors. The observed process in the above discussed arrangement (Fig. 6.5) was not the charging of neutral or non-equilibrium charged aerosols, but rather the stepwise neutralisation (“titration process”) of highly

6 Measurement and Characterization of Aerosol Nanoparticles

99

a 4 Sucrose 6.3 nm

Normalized particle concentration

Silica 17.7 nm

3

Thyroglobulin 11.2 nm Streptavidin 6.3 nm

2

1

0 2

2.4

2.8 Corona voltage [ kV ]

3.2

3.6

3.2

3.6

b 4 Sucrose 6.3 nm

Normalized particle concentration

Silica 17.7 nm 3

Thyroglobulin 11.2 nm Streptavidin 6.3 nm

2

1

0 2

2.4

2.8 Corona voltage [ kV ]

Fig. 6.6 Corona charging of nanoparticles by means of corona discharge. Charged fractions (positive and negative) were measured with nano-DMAs having either positive, or negative voltage applied to the inner electrode

100

W.W. Szymanski and G. Allmaier

charged nanoparticles with only one polarity produced by an electrospray process. Another difference is that the investigated nanoaerosol particles in this work varied not only in size but consisted of different compound classes such as sucrose, various proteins, and silica particles. The charging process here and final charge state of nanoaerosols after the electrospray procedure is known to be dependent on the chemical surface properties of substances. This effect caused by the increasing importance of image forces for the ion capture by particles below about 50 nm. Image forces are a strong function of the dielectric constant of the particle material (Biskos et al. 2005b). This makes the corona charger, as it was used in this application, a very useful device for preparatory purposes or to enrich a certain particle kind, however the variability of the charging level linked with the particle material makes it rather difficult to use this kind of arrangement for charging of unknown particles and DMA data inversion in order to obtain information about nanoparticle aerosols in terms of their real size distribution and concentration.

6.3

Measurement of Particle Surface with Surface Area Monitoring Devices

It is well known fact that especially in developed countries mass concentrations of PM (Particulate Matter), especially in urban areas, are continuously decreasing, but their number concentrations and hence particle surface area per volume increases. The decreasing ambient particle mass is due to very efficient removal technologies for super-micrometer particles. In early industrial times ambient particles were rather coarse. Presence of fine particles was timely limited among other due to rapid coagulation effects and sedimentation of resulting aggregates (Fig. 6.7a) The postindustrial, present period is characterised by an increased presence of ultrafine, nanosized ambient particles, which are quite persistent with regard to coagulation effects and can stay airborne for days (Fig. 6.7b) resulting in large particle surface area per unit volume with substantial impact on humans and environment. Assuming that ambient conditions with pollution burden regarding PM2.5 are 10 mg/m3; this would result in number concentration of 1 particle/cm3 with a diameter of 2.5 mm, or 2,400,000 particles/cm3 with a diameter of 20 nm. The corresponding

a

b

Fig. 6.7 (a) Multimodal aerosols in earlier industrial period resulted in fast removal of ultrafine particles due to coagulation with coarse particles. (b) Nanoparticles in post-industrial environment remain size-stable and airborne for long periods of time

6 Measurement and Characterization of Aerosol Nanoparticles

101

particle surface area would be about 20 mm2, or 3,000 mm2 per cm3, respectively. This simple example shows evidently the rapid increase of particle number and surface with decreasing particle size. Currently all air pollution and exposure limits are linked with the particle mass. However, there is now an increasing scientific indication regarding nanoparticles that it is likely the particle surface area which corresponds well with the body reaction rather than other, conventional particle metric, such as e.g. particle mass (Porter et al. 1999; Brown et al. 2001; Maynard and Kumpel 2005; Wittmaack 2007). Figure 6.8 summarizes the same experimental findings (Oberdo¨rster 2000) showing evidently that particle surface area correlates in a unique way, independent of particle size, with an inflammatory reaction of an organism. The correlation is missing for particulate mass (Fig. 6.8a). Obviously same particulate mass causes various responses depending on the particle size with an increasing effect for decreasing particle size. These observations triggered recently an instrumental development towards the measurement of particle surface. The operating principle is based on a two-step technique: particle charging with ions generated by means of a corona discharge and a consecutive detection of the charged aerosol by measuring the current in an electrometer, which is converted to particle surface values. There are at present three commercial devices monitoring particle surface area: Nanoparticle Surface Area Monitor (NSAM) (TSI, Inc., USA), DC2000-CE (EcoChem Analytics, USA) and LQ1-DC (Matter Engineering AG, Switzerland). The latter two instruments are calibrated to measure the total detected particle surface area (Fig. 6.9), the first one however, NSAM, measures the total surface area of particles deposited in certain areas of human lung. The instrument utilizes an ion trap with variable voltage resulting in measured electrical currents representing the total deposited surface area of particles in either tracheobronchial (TB), or alveolar (AL) regions of a lung. We performed an evaluation of the NSAM with well-defined, monodisperse aerosols (DEHS – Di-Ethyl-Hexyl-Sebaccate) of known sizes and concentrations comparing the measured surface area of deposited particles by NSAM with the mathematical model provided by the International Commission on Radiation Protection (ICRP 1994; Vincent 2005). The data show that experimentally determined values relate very well the model for TB and AL total deposited nanoparticle surface area (Fig. 6.10).

6.4

Mass Measurement of Single Aerosol Particles

Traditional aerosol mass measurement using a filter, or inertial methods such as impactors delivers time-integrated information in terms of accumulated mass per unit volume. With the improving air quality this type of measurement becomes increasingly difficult using traditional metric such as PM10, or PM2.5. Moreover, the mass measurement of nanoaerosols (Dp < 100 nm, also called ultra fine particles) requires excessively long sampling periods resulting in virtually non

102

W.W. Szymanski and G. Allmaier

a 40%

Inflammatory Response

Dp = 25 nm 30%

20%

10%

0% 10

Dp = 250 nm

100 1000 Particulate Mass [ µg ]

10000

b 40%

Inflammatory Response

Dp = 25 nm 30%

20%

10%

0% 0.0001

Dp = 250 nm

0.001 0.01 Particulate Surface Area [ m²/m³ ]

0.1

Fig. 6.8 Inflammatory response to TiO2 particles of various sizes plotted as a function of (a) particulate mass, (b) particulate surface area (adapted from Oberdo¨rster 2000)

existing time resolution, hence making a source identification and apportionment of pollutants a very questionable task. In recent years the rise of aerosol mass spectrometry provided instruments delivering chemical information in real-time from single aerosol particles. The combination of aerosol measuring techniques and mass spectrometric techniques into a real-time measurement system offers unique possibility of physical and

6 Measurement and Characterization of Aerosol Nanoparticles

1:1 Relationship

8000

Measured surface area[ nm²]

103

SMPS TEM LQ1-DC

6000

DC 2000CE

4000

2000

0 0

1000

2000

3000 4000 5000 6000 Calculated surface area[ nm²]

7000

8000

9000

Fig. 6.9 The experimental evidence (adapted from Maynard 2006) indicates that electrically measured surface area of monodispersed test aerosols is in a very good agreement with calculated surface area obtained from scanning differential mobility analysis. The data has also been confirmed evaluating TEM images of particles

0.6 measured Al-Fraction

0.5

Deposition, fraction

calculated Al-Fraction measured TB-Fraction

0.4

calculated TB-Fraction

0.3

0.2

0.1

0 0

50

100

150

200

250

Particle Diameter [nm]

Fig. 6.10 Predicted regional deposition in the tracheobronchial (TB) and alveolar (AL) region for nose breathing and the response of the NSAM obtained with monodisperse nanoaerosols

chemical characterization of nanoparticles (Marijnissen et al. 1988; Noble and Prather 1998; Jayne et al. 2000; Harris et al. 2005; Mu¨ller et al. 2007). Utilization of mass spectrometric means for nanoparticle study will be given later.

104

W.W. Szymanski and G. Allmaier

Recently, in addition to these analytical chemical techniques an inventive mass measurement of single aerosol nanoparticles based on the aerodynamic movement of particles due to centrifugal and electric forces has been introduced.

6.4.1

Aerosol Particle Mass Analyzer (APMA, Mod. APM-10)

Many ambient aerosols are composed of non-spherical nanoparticles posing a specific health hazard. Their mass, or density cannot be easy derived from differential mobility analysis. However, a new approach published recently (Ehara et al. 1996, McMurry et al. 2002, Park et al. 2003) opens new avenues for more complete nanoaerosol characterisation as in the past, providing not only mass but also parameters such as fractal dimension, characteristic density and shape factors of particles in question. The principle of this measuring method is shown in Fig. 6.11. Balancing the centrifugal (FZ) and the electrostatic force (FEL) in a laminar flow field allows manoeuvring particles through the gap between the rotating inner and outer electrode with a certain specific mass-to-charge ratio represented by the specific mass parameter SC (Eq. 6.2). Prior to entering the instrument particles must be properly charged. Leaving the instrument they must be detected with methods used for particles leaving a DMA.

R1

HV

FEL = qE

Fig. 6.11 Schematic diagram and the principle of function of the APMA (Mod. AMP-10). The electrostatic field in the gap between the inner electrode (radius R1) and outer electrode (radius R2) is determined by the applied HV. Depending on the rotating speed o particles with given specific mass SC will be passing through the gap leaving the separating column

R2

FZ = mrω²

HV qE

mrω²

6 Measurement and Characterization of Aerosol Nanoparticles

sC ¼

V

  rC 2  o2  ln RR21

105

(6.2)

There is a continuous search for the proper metric of nanoparticles. Although, as discussed above, particle size measurement is now possible down to about 1 nm, surface area and nanoparticle number concentration are measurable and seem to be good indicators of their health hazard potential together with the chemical composition and also modern real-time mass measurement in the femtogram range is feasible, it is evident that each measuring method delivers only partial information. However, combinations of various measuring techniques open new insight and possibilities of enhanced description of nanoaerosol particle systems. 6.4.1.1

Grouping of Differential Mobility and Aerosol Particle Mass Analyzer

A measurement of each single particle over the entire size range of interest with size-resolved particle composition and physical properties would be the ultimate objective for a perfect nanoparticle characterisation. There is no single instrument capable of that task, but depending on posed questions combination of measuring techniques can deliver realistic answers. Particularly the grouping of a DMA and APMA provides real-time information regarding nanoparticle size, mass, density and fractal structure, properties of importance for particle transport and deposition in human respiratory system. Evidently for spherical particles simultaneous measurement of size (DMA) and mass (APMA) delivers instantaneously density of each detected particle (McMurry et al. 2002). The measurement of non-spherical particulate aggregates such as atmospheric soot particles poses much more challenging task. However, the combination of these two measuring systems allows approaching this problem as can be seen in Fig. 6.12. The assessment of effective density of measured soot agglomerates is possible by means of the Eq. 6.3, where r0=1 g/cm3 and DDMA and DAPMA are the electrophoretic mobility diameter determined by the DMA and diameter of the average mass determined by the APMA, respectively. reff ¼ r0 

6.4.1.2

DDMA 3 DAPMA 3

(6.3)

Grouping of Differential Mobility Analyzer with Mass Spectrometer

Continuously increasing need for more complete characterisation of ambient aerosols and better understanding of processes in aerosols on the nanosize level resulted in a fact that mass spectrometry (MS) obtained an outstanding position among

106

W.W. Szymanski and G. Allmaier

Effective Particle Density [ g/cm3 ]

1.4 Electrical mobility diameter increases

1.2 1.0 0.8 0.6 0.4 0.2 0.0

0

100 200 300 Electrical Mobility Diameter [ nm ]

400

Fig. 6.12 Effective density of soot agglomerates decreases with increasing complexity of the aggregate corresponding to the increasing electrical mobility diameter. Data show also that the load of an engine does not dominantly influence the effective particle density (courtesy McMurry 2008)

various aerosol measuring techniques (Dusek et al. 2006; Schneider et al. 2004) High sensitivity and diversity of applications of a MS techniques combined with aerosol methods, which in contrast to most MS systems allow measurement under ambient pressure conditions opens new ways of description of airborne nanoparticles including such species as dendrimeres, macromolecules and viruses (Bacher et al. 2001; Mu¨ller et al. 2007; Laschober et al. 2008; Allmaier et al. 2008). Particularly inviting and relatively straighforward is the combination of a DMA delivering size information with MS delivering mass information for globular nanoparticles such as many proteins, or dendrimeres. The latter one (also called dendritic molecules) represent a new class of molecular architecture have globular shape and various sizes (Tomalia and Frechet 2002). The examples below show a special class of dendrimeres – PAMAM dendrimeres, which resemble proteins in their size and chemical structure (Dentritech, Inc.). Their sizes correspond to the “Generation” covering mass range from a few kDa (Generation – G1) to MDa (Fig. 6.13). Evidently a combination of the information from two devices delivering different matrics of the same object – the size (DMA) and the mass (MS) of such nanoparticles – allows an accurate assessment of their density hence contributing to more complete characterisation of nanoaerosols and nanoparticles (Fig. 6.14). The measurement of mass spectra of nanoparticles with sizes corresponding to molecular masses of the order of several hundred kilodaltons, or more, becomes increasingly difficult with MS techniques. This is manifested also by the fact that the resolution of MS technique, which is indeed superior to aerosol instrumentation for the low kilodalton mass range becomes increasingly poorer for nanoparticles with sizes of the order of 15 nm, or larger. A comparison of largest “monodisperse” protein – immunoglobulin M – spectra obtained with MALDI-TOF-MS and GEMMATM (TSI, Inc.) illustrates the complementary strength of an aerosol

a

Relative Abundance

6 Measurement and Characterization of Aerosol Nanoparticles [M]+G6

107

[M]+G9

1.0

0.5

0 5

0

10

15

20

b

Relative Abundance

EM-Diameter [ nm ]

1.0

G6

0.5

0

Relative Abundance

20000

40000

60000 Mass / Charge

1.0

80000

G9

0.5

0 100000

200000

300000 Mass / Charge

400000

Fig. 6.13 (a) Size spectra of dendrimers (G6 and G9) measured by means of a differential mobility analysis and (b) mass spectrum of dendrimers of the G6 generation (adapted from Mu¨ller et al. 2007)

technique (DMA) vs MS characterisation of molecules and functional complexes in the MDa molecular mass range (Fig. 6.15).

6.5

Summary

The size of nanoparticle mainly governs their deposition in respiratory system as well and their fate within a human body. It is also quite certain now that nanoparticle surface is a critical indicator linked with the seditious response of organisms. Realtime measurement of aerosol nanoparticles is hence an indispensable approach in

108

W.W. Szymanski and G. Allmaier 1.4 experimental values PAMAM-ideal structure

1.2

Density [ kg/l ]

1 0.8 0.6 0.4 0.2 0 0

2

4

6

8

10

12

14

16

Dendrimer Diameter [ nm ] Fig. 6.14 Measured and theoretical values (PAMAM-ideal structure) of dendrimer densities for various dendrinmer generations obtained from data gathered by a DMA analysis and MALDITOF-MS (adapted from Mu¨ller et al. 2007)

Fig. 6.15 MALDI-TOF-MS spectrum of the immuno-globulin M ([M]+ at m/z = 982 kDa (monomeric protein) is indicated with the arrow) and GEMMA spectrum (insert) of the same compound with the peak at 17.4 nm corresponding to the monomer (adapted from Nelsen et al. 1995)

understanding adverse health effects associated with the environmental presence of nanoparticles and related exposures. Combinations of instruments measuring different particle properties allow further insight and understanding of dynamics of aerosols and their influence on humans. Table 6.1 summarizes selected instruments and methods indicating also their operating range and typical data sampling time.

6 Measurement and Characterization of Aerosol Nanoparticles

109

Table 6.1 Specification of discussed instrumental devices and measured aerosol parameters Metric of nanoaerosol Method instrument Measuring range Time resolution (s) Particle size, number Scanning DMAs + 3–1,000 nm 100 CPC Scanning DMAs + 0.5–1,000 nm 100 AE 1 Total particle surface Dc2000 (EcoChem) 0–2,000 mm2/cm3 LQ1-DC 0–2,000 mm2/cm3 1 (Matter Eng.) Lung deposited NSAM (TSI. Inc.) 0–2,500 mm2/ 1 cm3(TB) particle surface 1 NSAM (TSI. Inc.) 0–10,000 mm2/ cm3(A) Particle mass APMA–MOd. 10 0.01–100 fg 100 Molecular mass, chemical TOF–MS 0.1kDa–1MDa 100 composition

The benefits of measurement of aerosol nanoparticles using electrical techniques include real time determination of size distributions even below 1 nm, assessment of particle surface area and appraisal of particles’ effective density, or their fractal dimensions. The prerequisite for all electrical techniques is the well-defined charging of particles in question. So far, no genuine alternative to charging and/or neutralization process using radioactive sources has been established, although various approaches, including corona discharge are being continuously investigated. A successful replacement for the radioactive source would enormously contribute to broader exploitation of electrical aerosol measuring techniques. The development of nanotechnology, besides the presence of ambient nanoparticles, likely results in increased health risks linked with manufacturing and handling of nanoparticles and is a growing societal concern. Persons may be exposed to nanoparticles through of inhalation, but so far no ambient, or workplace standards exist to with regard to exposure to nanoparticles. Unremitting interdisciplinary efforts are needed for better, faster and more affordable methods for real-time characterisation of aerosol nanoparticles along with innovative metrics related to health risks, maybe to heath benefits and also to management of nanoparticle burden. Acknowlegements The support of the Austrian Science Foundation (FWF) grant P16185 (to W.W.S) and P15008 (to G.A.) is here gratefully acknowledged. Authors would like express their thanks to Nikolaus Fo¨lker for his help with the graphical layout.

References Allmaier G, Laschober C, Szymanski WW, Nano ES (2008) GEMMA and PDMA, new tools for the analysis of nanobioparticles – protein complexes, lipoparticles and viruses. J Am Soc Mass Spectrom 19(8):1062–1068

110

W.W. Szymanski and G. Allmaier

Alonso M, Martin MI, Alguacil FJ (2006) The measurement of charging efficiencies and losses of aerosol nanoparticles in a corona charger. J Electrostat 64:203–214 Ankilov A, Baklanov A, Colhoun M, Enderle KH, Filipovicova D, Julanov A, Lushnikov A, Vrtala A, Mavliev R, McGovern F, Mirme A, OConnor TC, Podzimek J, Preining O, Reischl G, Rudolf R, Sem G, Szymanski WW, Tamm E, Wagner P, Zagaynov A (2002) Particle size dependent response of aerosol counters. Atmos Res. 62:209–237 Aranchuk LE, Chuvatin AS, Larour J (2004) Compact submicrosecond, high current generator for wire explosion experiments. Rev Sci Instrum 75:69–75 Bacher G, Szymanski WW, Kaufman SL, Zo¨llner P, Blaas D, Allmaier G (2001) Charge reduced nano-electrospray combined with differential mobility analysis of peptides, proteins, glycoproteins, noncovalent protein complexes and viruses. J Mass Spectrom 36:1038–1052 Biskos G, Reavell K, Collings N (2005a) Electrostatic characterization of corona-wire aerosol charges. J Electrostat 63:69–82 Biskos G, Reavell K, Collings N (2005b) Description and theoretical analysis of a differential mobility spectrometer. Aerosol Sci Technol 9:527–541 Biskos G, Reavell K, Collings N (2005c) Unipolar diffusion charging of aerosol particles in the transition regime. J Aerosol Sci 36:247–265 Brown DM, Wilson MR, MacNee W, Stone V, Donaldson K (2001) Size-dependent proinflammatory effects of ultrafine polystyrene particles: a role for surface area and oxidative stress in the enhanced activity of ultrafines. Toxicol Appl Pharmacol 175:191–199 Chaolong Q, Chen DR, Pui DYH (2007) Experimental study of a new corona-based unipolar aerosol charger. J Aerosol Sci 38:775–792 Chen DR, Pui DYH (1997) Numerical modeling of the performance of differential mobility analyzers for nanometer aerosol measurement. J Aerosol Sci 28:985–1004 Chen DR, Pui DYH (1999) A high efficiency, high throughput unipolar aerosol charger for nanoparticles. J Nanoparticle Res 1:115–126 Collins DR, Nenes A, Flagan RC, Seinfeld JH (2000) The scanning flow DMA. J Aerosol Sci 31:1129–1144 De la Mora JF, De Juan L, Eichler T, Rosell J (1998) Differential mobility analysis of molecular ions and nanometer particles. TRAC 17:328–339 Dusek U, Frank GP, Hildebrandt L, Curtius J, Schneider J, Walter S, Chand D, Drewnick F, Hings S, Jung D, Borrmann S, Andreae MO (2006) Size matters more than chemistry for cloud nucleating ability of aerosol particles. Science 312:1375–1378 Ehara K, Hagwood C, Coakley KJ (1996) Novel method to classify aerosol particles according to their mass-to-charge ratio – aerosol particle mass analyser. J Aerosol Sci 27:217–234 Erikson HA (1922) On the nature of the negative and positive ions in air, oxygen and nitrogen. Phys Rev 20:117–126 Faraday M (1857) The Bakerian Lecture – experimental relations of gold (and other metals) to light. Philos Trans R Soc Lond 145:181 Gazso A, Gressler S, Schimer F (eds) (2007) Nano – Chancen und Risken aktueller Technologien, Springer, Vienna, New York Goldman M, Goldmann A, Sigmond RS (1985) The corona discharge, its properties and specific uses. Pure Appl Chem 57(9):1353–1362 Gras JL, Podzimek J, O’Connor TC, Enderle KH (2002) Nolan–Pollak type CN counters in the Vienna aerosol workshop. Atmos Res 62:239–254 Harris WA, Reilly PTA, Whitten WB (2005) MALDI of individual biomolecule-containing airborne particles in an ion trap mass spectrometer. Anal Chem 77(13):4042–4050 ICRP – International Commission on Radiological Protection (1994) Human Respiratory Tract Model for Radiological Protection, ICRP publication 66, Pergamon, Elmsford, NY Jayne JT, Leard DC, Zhang X, Davidovits P, Smith KA, Kolb CE, Worsnop DR (2000) Development of an aerosol mass spectrometer for size and composition analysis of submicron particles. Aerosol Sci Technol 33:49–70

6 Measurement and Characterization of Aerosol Nanoparticles

111

Kinney PD, Pui DYH, Mulholand GW, Bryer NP (1991) Use of the elecftrostatic classification method to 0.1 tm SRM particles – a feasibiiiy study. J Res Natl Inst Stand Technol 96:147 Knutson EO, Whitby KT (1975) Aerosol classification by electric mobility: apparatus, theory and applications. J Aerosol Sci 6:443–451 Kotov YA (2003) Electric explosion of wires as a method for preparation of nanopowders. J Nanoparticle Res 5:539–550 Laschober C, Kaufman SL, Reischl G, Allmaier G, Szymanski WW (2006) Comparison between an unipolar corona charger and a polonium-based bipolar neutralizer for the analysis of nanosized particles and biopolymers. J Nanosci Nanotechnol 6:1474–1481 Laschober C, Kaddis CS, Reischl GP, Loo JA, Allmaier G, Szymanski WW (2007) Comparison of various nano-differential mobility analysers (nDMAs) applying globular proteins. J Exp Nanoscience 2(04):291–301 Laschober C, Wruss J, Blaas D, Szymanski WW, Allmaier G (2008) Gas-phase electrophoretic molecular mobility analysis of size and stoichiometry of complexes of common cold virus with antibody and soluble receptor molecules. Anal Chem 80(6):2261–2264 Liu BYH, Pui DYH (1975) On the performance of the electrical aerosol analyzer. J Aerosol Sci 6:249–264 Liu BYH, McKenzie RL, Agarwal JK, Jaenicke R, Pohl FG, Preining O, Reischl G, Szymanski WW, Wagner PE (1982) Intercomparison of different “absolute” instruments for measurement of aerosol number concentration. J Aerosol Sci 13:429–450 Marijnissen J, Scarlett B, Verheijen P (1988) Proposed on-line aerosol analysis combining size determination, laser-induced fragmentation and time-of-flight mass spectroscopy. J Aerosol Sci 19(7):1307–1310 Maynard AD (2006) Nanotechnology: managing the risks. Nanotoday 1:22–33 Maynard AD, Kumpel ED (2005) Airborne nanostructured particles and occupational health. J Nanoparticle Res 7:587–614 Maynard AD, Pui DYH (eds) (2007) Nanotechnology and occupational health. Springer, Dordrecht, The Netherlands McMurry PH, Wang X, Park K, Ehara K (2002) The relationship between mass and mobility for atmospheric particles: a new technique for measuring particle density. Aerosol Sci Technol 36:227–238 McMurry PH, Private communication 2008 Mesbah B, Fitzgerald B, Hopke PK, Pourprix M (1997) A new technique to measure the mobility size of ultrafine radioactive particles. Aerosol Sci Technol 27:381–393 Mu¨ller R, Laschober C, Szymanski WW, Allmaier G (2007) Determination of molecular weight, particle size and density of high number generation PAMAM dendrimers using MALDITOF-MS and nES-GEMMA. Macromolecules 40:5599–5605 NARSTO (North American Research Strategy for Tropospheric Ozone) (2007) Aerosol Workshops, Crystal City, Virginia, USA, Webpage: http://www.narsto.org/ Nelsen RW, Dogruel D, Williams P (1995) Detection of human IgM at m/z .apprx. 1 MDa. Rapid Comm Mass Spectr 9:625 Noble CA, Prather KA (1998) Aerosol time-of-flight mass spectrometry: a new method for performing real-time characterization of aerosol particles. Appl Occup Environ Hygiene 13:439–443 Oberdo¨rster G (2000) Pulmonary effects of inhaled ultrafine particle. Int Arch Occup Environ Health 74:1–8 Park K, Feng C, Kittelson DB, McMurry PH (2003) Relationship between particle mass and mobility for diesel exhaust particles. Environ Sci Technol 37:577–583 Porter DW, Castranova V, Robinson VA (1999) Acute inflammatory reaction in rats after intratracheal instillation of material collected from a nylon flocking plant. J Toxicol Environ Health A 14:25–45

112

W.W. Szymanski and G. Allmaier

Preining O (1998) The physical nature of very, very small particles and its impact on their behaviour. J Aerosol Sci 29:481–495 Pui DYH, Chen DR (1997) Nanometer particles: a new frontier for multidisciplinary research. J Aerosol Sci 28:539–544 Reischl GP (2008) private communication Reischl GP, Makela JM, Necid J (1997) Performance of Vienna type differential mobility analyzer at 1.2–20 nanometer. Aerosol Sci Technol 27:651–672 Rohmann H (1923) Methode zur Messung der Gro¨ße von Schwebeteilchen. Zeitschrift fu¨r Physik 17:253–265 Royal Society and Royal Academy of Engineering (2004) Final Report on Nanotechnology and Nanoscience – Website http://www.nanotec.org.uk Schneider J, Borrmann S, Wollny AG, Bla¨sner M, Mihalopoulos N, Oikonomou K, Sciare J, Teller A, Levin Z, Worsnop DR (2004) Online mass spectrometric aerosol measurements during the MINOS campaign. Atmos Chem Phys 4:65–80 Tammet H (1995) Size and mobility of nanometer particles, clusters and ions. J Aerosol Sci 26:459–475 Tomalia DA, Frechet JMJ (2002) Discovery of dendrimers and dendritic polymers: a brief historical perspective. J Polymer Sci A 40:2719–2728 Vincent JH (2005) Health-related aerosol measurement: a review of existing sampling criteria and proposals for new ones. J Environ Monit 7:1037–1053 Vivas MM, Hontanon E, Schmidt-Ott A (2008) Reducing multiple charging of submicron aerosols in a corona diffusion charger. Aerosol Sci Technol 42:97–109 Winklmayr W, Reischl GP, Lindner AO, Berner A (1991) A new electromobility spectrometer for the measurement of aerosol size distribution in the size range from 1 to 1000 nm. J Aerosol Sci 22:289–296 Wittmaack K (2007) In search of the most relevant parameter for quantifying lung inflammatory response to nanoparticle exposure: particle number, surface area, or what? Environ Health Perspect 115:187–194 WUFA 1979 – Working Group on Ultrafine Aerosol, Vienna 1979, Results of the WUFAWorkshop, Vienna, May 20–June 15, 1979, Liu BYH, Pui DYH, McKenzie RL, Agarwal JK, Jaenicke R, Pohl FG, Preining O, Reischl G, Szymanski WW, Wagner PE. J Aerosol Sci 1980, 11:261-263 Zeleny J (1929) The distribution of mobilities of ions in moist air. Phy Rev 34:310–334 Zhang SH, Flagan RC (1996) Resolution of the radial differential mobility analyzer for ultrafine particles. J Aerosol Sci 27:1179–1200 Zsigmondy R (1905) Zur Erkenntnis der Kolloide – Ueber die irreversiblen Hydrosole und Ultramikroskopie, G. Fischer, Jena

Chapter 7

Inhalation and Deposition of Nanoparticles: Fundamentals, Phenomenology and Practical Aspects Arkadiusz Moskal, Tomasz R. Sosnowski, and Leon Gradon´

7.1

Introduction

The development of an improved theoretical model for predicting particle deposition during inhalation requires accurate knowledge of respiratory system geometry, air flow structure inside the airways related to the breathing patterns and physical properties of aerosol particles defined through their shape, morphology, geometric dimension and density. For most aerosols, the main particle deposition mechanisms are inertial impaction, sedimentation, interception and Brownian diffusion. The particles considered in this paper, both environmental and therapeutic, have a complex structure due to the process they are produced through. Generally such particles are nano-structured and consist of many spherical primary nanoparticles with diameters of the order of 10 nm. The main mechanisms of deposition of these particles within the respiratory system are Brownian diffusion and interception. In the present paper we will analyze some deposition models of diffusional particles at the characteristic part of human airways.

7.2 7.2.1

Geometry of the Human Respiratory System Upper Airways

The upper airways in human beings serve an important role in humidification and warming of the air that we breathe in. They also protect the lower airways from pathogens and toxic substances present in the environment. Therapeutic drugs delivered into the human body via inhalation of aerosol particles are partially removed from the aerosol stream in the upper airways. A. Moskal (*), T.R. Sosnowski, L. Gradon´ Faculty of Chemical and Process Engineering, Warsaw University of Technology, Warynskiego 100 – 645, Warsaw, Poland

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_7, # Springer ScienceþBusiness Media B.V. 2010

113

114

7.2.1.1

A. Moskal et al.

Nose

The interior of the nose is divided by the nasal septum into two separate cavities. These begin at the nostrils and are divided by the septum until the posterior nares, where the cavities unite and join the nasopharynx. The nasal vestibules narrow to a slit of three to four mm, commonly called the nasal valve, which form the junction between the vestibule and the main nasal cavity. The nasal valve is the narrowest point of the nasal cavity. The nasal chamber is a high and narrow channel of complex geometry owing to the presence of the turbinates – three large mucosal folds. The height of the chamber decreases along the passageway while the width increases to maintain a constant cross-sectional area. Behind the nasal chamber, the channel again becomes constricted and makes a severe bent into the oropharynx (Langes, 1989). Figure 7.1. presents the longitudinal and crossectional (at the same position) views of the nasal cavity.

7.2.1.2

Oral Cavity Air Passages

The oral cavity is bounded anteriorly by the lips, posteriorly by the anterior tonsillar pillar, inferiorly by the floor of the mouth, and superiorly by the hard and soft palates. The pharynx is a 12–13 cm long muscular tube in the adult divided into three parts: the oropharynx, nasopharynx and hypopharynx. Its function is to transport with air any particulate matter in the form of aerosol particles from the external environment into the lung. Taking into account the characteristic part of oral passages with respect to transport properties, the following parts can be distinguished: (a) the buccal region from the back of the teeth to the soft palate, (b) the nasopharynx region

Fig. 7.1 The upper airways structure

7 Inhalation and Deposition of Nanoparticles

115

incorporating the nasal airways to the tip of epiglottis, (c) the laryngeal-pharynx from the tip of epiglottis to just above the vocal chords and (d) the laryngeal cavity defined from just above the vocal chords to the trachea. Such a complicated structure is difficult to describe using regular geometrical approaches. The nasal or oral cavity geometry is transformed into a numerical description of the computational domains or for formation of the replicas, through advanced imaging techniques using in vivo studies. It can be determined in vivo by computerized tomography (CT) or magnetic resonance imaging (MRI). The oropharyngeal models have been developed also through the application of X-ray CT scans. Over the last two decades MRI has provided much more information on the detailed anatomy of the upper airways.

7.2.2

Airway System of Bronchial Region

The function of the bronchial tree is to conduct air to the alveolar surface where gas transfer takes place between respired air and gas dissolved in the blood of pulmonary capillaries. The inspired air should be distributed evenly to the alveolar capillary bed with minimal resistance to flow, Fig. 7.2. The conducting zone of the lung includes the trachea, bronchi and nonalveolated bronchioles in which air cannot diffuse through the well developed wall. The bronchial tree exhibits a complicated branching pattern expanded from the trachea to the membranous bronchioles. Data describing the branching structure of the lung are derived from (Horsfield and Cumming 1968), and (Weibel 1963). The basic branching pattern of the lung is dichotomous. The branching may be symmetric – two branches equal in all respects, or asymmetric with variation in the diameter or length of branches in a given generation, the number of divisions to the final branches or combination of

Fig. 7.2 Geometry of the tracheobronchial and pulmonary airways

116

A. Moskal et al.

Fig. 7.3 Geometry of single bifurcation

both. The experimental or mathematical models of deposition require the definition of geometric parameters of each branching system, i.e. dimensions and shapes of the cross-section of the ducts, their lengths, branching angles and curvatures (Gradon´ and Orlicki 1990) (Fig. 7.3). Taking into account the computational abilities of the system used, a single bifurcation or a sequence of a few consecutive ones is considered in modeling the deposition.

7.2.3

Pulmonary Region

The third part of the airways forms the respiratory zone which consists of all structures that participate in gas exchange. It is the alveolated region of the lung. This portion of the lung distal to one terminal bronchioles forms an anatomical unit, called an acinus. A typical acinus contains about 14 respiratory bronchioli, each of which has shallow alveoli in its wall. Beyond these are alveolar ducts (approximately 100 per acinus). Alveolar ducts are completely lined with alveoli. The modeling of this part of the respiratory system is reduced to the description of gas flow and particle motion in the spherical region attached to the cylinder. Sometimes a deep lung morphology depicts spherical alveoli grouped in a grape-like cluster. The models of gas flow and particle motion incorporate the constitutive parameters of a lung tissue for the description of time-dependent alveolar wall displacement during breathing.

7.3

Gas Flow in the Respiratory Tracts

Temporary and local patterns of the air flow in the human respiratory system are complicated because of the geometrical complexity of a particular part of the system and the pulsative character of the driving force for breathing, i.e. time

7 Inhalation and Deposition of Nanoparticles

117

dependent pressure gradient during the respiratory cycle. We can observe turbulent and laminar flows occurring simultaneously at different parts of the respiratory tract, and similar effects can be observed at the same place, but for different moments of the respiratory cycle. This complexity requires careful analysis of the stated problem and responsibility in the selection of methods for its solution. The Navier–Stokes equations describe the variation of pressure and velocity in the fluid. Using Cartesian index notation, Xi, j, k with the summation convention, we can write the first equation, which is an expression of the mass conservation, as: @Ui ¼0 @xi

(7.1)

And the second, which is an expression of the conservation of momentum, as: @Ui @Ui Uj 1 @p @ 2 Ui þ ¼  þ nm  @t @xj @x2 r @xi

(7.2)

I II III IV where: the first term I means the accumulation, II – advection, i.e. momentum carried by flow, III – driving force, i.e. pressure gradient and IV – diffusion in which momentum is dispersed by the viscous effect. In the above equation U means fluid velocity, t – time, p – pressure, vm – molecular kinematic viscosity and r fluid density, (Daley and Harlow 1970; Pope 2000). Navier–Stokes equations (7.1, 7.2) can be used to derive the relationships which are convenient for establishing dynamic similitude. Let us begin with ordered – laminar flow, and focus our attention on two terms (II and IV) in Eq. 2, i.e. advection and diffusion. Advection is related to kinematic effects, which is the transport of fluid properties by the motion of the fluid, and thus accounts for momentum transport along streamlines. The diffusion terms represent viscous effects that cause momentum to diffuse between streamlines, thereby leading to diminish and sharp velocity gradients. Introducing the length scale L and the fluid velocity in the free stream U0 into the reduced terms II and IV, Eq. 7.2, we can define the dimensional number which indicates the ratio of the advection and diffusion of momentum:

Re ¼

Uo  L nm

(7.3)

called the Reynolds number, which is a comparative measure of inertial and viscous effects within the flow field. When the geometry of an object is defined for the fluid flow, and the initial and boundary conditions for the Navier–Stokes equations are properly stated, the problem can be reasonably easily solved. The fluid velocity distribution, U(x, t), is determined.

118

A. Moskal et al.

In reality, the real laminar flow condition doesn’t exist in the human respiratory system during breathing. Oscillatory excitation of the flow through a complicated geometry causes the flow to lose its stable structure. The flow is disordered. Flows that are simple in Eulerian terms (see the next section of the paper) may nevertheless produce very complicated trajectories of a Lagrangian tracer (e.g. a nanosized, inertialess particle). This should be noted when the deposition of particles in the respiratory system is considered. The situation is even more complicated when real disorder is introduced into the flow caused by increasing the Reynolds number above its critical value. The inertia contribution to mean-flow momentum that cannot be dissipated by viscous stress must be absorbed by the formation of new structures – turbulent eddies. Turbulence energy is described in terms of an effective turbulence viscosity nt defined as the ration of turbulence-shear to the mean-flow strain rate. With this an effective turbulence Reynolds number is defined as: Re ¼

Uo  L nm þ nt

(7.4)

Turbulence modeling is based on two simplifying assumptions with respect to the full Navier-Stokes equations: (a) fluid is incompressible (r  U ¼ 0) and (b) density r is constant everywhere (which is the case of the respiratory flow condition). Even with these assumptions a direct solution of the Navier–Stokes equation is very difficult. Theoretical approaches to turbulence modeling use a type of averaging – either temporally, spatial or ensemble. Ensemble averaging is the most general type of averaging with the fewest restrictions and is most conveniently formulated in terms of moments of an appropriate distribution function. The Reynolds stress transport equation supplied with additional constraints or empirical information (the classic concept) is a common approach for turbulence modeling. Simpler transport models of turbulence related to the description of turbulent viscosity use the concept of the mixing length but for more complex flow the k – e model is useful. To effectively use turbulence modeling, one must decide which length scales will be considered mean flow and which will be considered turbulence. With this clear vision a wide spectrum of codes exist for a numerical solution of the turbulence flow problem. Recently, the Large Eddy Simulation (LES) method became very common in turbulence simulations of the airflow in human upper airways. The objective of LES is to compute three-dimensional, time-dependent details of the largest scales of motion, those responsible for the primary transport. LES is intended to be useful in the study of turbulence physics at high Re numbers, in the development of turbulence models and for flows in a complex situation where simpler approaches (for example, the Reynolds stress transport) are inadequate. Some groups of models describe turbulence by reducing the phenomenon to the near-wall. The distance yt from the wall within which the phenomenon is considered, is scaled through the shear velocity of the fluid flow region, using the viscous sublayer

7 Inhalation and Deposition of Nanoparticles

119

concept. The stochastic nature of fluid velocity components is expressed by the Ornstein–Uhlenbeck process, as the consequence of the Langevin equation (see the next section). At this point one has to remember that the viscous sublayer concept shadows the occurrence of the near-wall coherent structures formation. The “bursting”, i.e. production of turbulence within the boundary layer by violent ejection of a near-wall fluid, is the source of a strong effect of the lifting force acting on the particle. The streak structures eject from the viscous sublayer (y+ < 5) and then move downstream with the flow direction x. Two concepts are currently used when considering the violent, temporally intermitted ejection of the fluid from the wall or localized ejection of the fluid from the wall by quasi-streamwise vortices which are persistent. Ignoring such an effect, we lose important information about the nature of turbulence. A simple stochastic approach causes velocity structure to be treated as computable output rather than assumed input in the process description. At the end of this section we have to mention that a very promising method of modeling the fluid flow, dynamically developed in recent years, is the cellular automata concept. This method can be also useful for modeling the particle transport in the respiratory system.

7.4

Particles

The size range of common airborne particulate pollutants presented in the atmospheric air extends from about 1 nm in particle diameter, for example combustion nuclei, fumes, diesel exhaust, to more than five orders of magnitude. Observing characteristic ambient particle size distributions, frequently more than 50% of the total ambient mass can be related to fine aerosols with the particle diameter smaller than about 1 mm. Furthermore, over 80% of the total number of atmospheric particles are typically smaller than about 100 nm. Freshly generated nanosized particles when inhaled as singlets or small aggregates, at very low mass concentrations, can be highly toxic. It is caused by their high deposition efficiency in the lower respiratory tract, their large number per unit mass, and their increased surface areas available for interaction with cells. Similar properties of particles are also used constructively for the formation of therapeutic particles delivered into the human lung via inhalation. Therapeutical aerosol particles are delivered for lung treatment (asthma, COPD) or for systemic drug delivery in aerosoltherapy. The common root for both types of inhaled particles, ambient or therapeutic, is their size and morphology. These two factors determine the mechanical properties of particles. The main mechanism of deposition of such particles in the human respiratory system is diffusion. The diffusion coefficient characterizes particle properties with respect to its movement and deposition in the respiratory system. If a Brownian particle size is very large by comparison with the molecules of the fluid in which it is immersed, it “sees” the fluid as a continuous medium, and we may use a

120

A. Moskal et al.

hydrodynamic definition of the friction coefficient on the basis of Stokes’ law. The viscous retarding force on a spherical particle of radius R moving with velocity V in a fluid of viscosity m is equal 6pmRV, so the friction coefficient ’ is equal: ’ ¼ 6pmR

(7.5)

The assumption of motion with a steady velocity V is equivalent to the customary neglect of accelerations in theories of diffusion. This interpretation of ’ is not generally applicable to the diffusion of particles of arbitrary size. When the particle radius is comparable to the mean free path of the gas molecule l in which the particle is immersed, the friction coefficient ’ derived for continuum should be modified through the introduction of the Cunningham correction factor Cc into the Eq. 7.5. This factor is related to striking the surface of the particles by molecules (slip effect) and is a function of the Knudsen number Kn ¼ l/R: 
 d Cc ¼ 1 þ Kn a þ b  exp  Kn

(7.6)

where a, b and d are constants obtained from experiment. The diffusion coefficient D of the particles is related to the friction coefficient ’ and the Cunningham correction factor Cc by the equation: D¼

kT  Cc ’

(7.7)

where k is the Boltzman constant and T is the absolute temperature (Friedlander 2000). It should be mentioned at this point that Cc coefficient for nanosized particles immersed in the air at standard conditions has significant values. For particles of 10 nm in diameter, Cc is of the order of 10 and for particles of 100 nm, it equals 3. The values of D for spherical particles are available in literature on the subject, or can be calculated from the Stokes-Einstein formula: D¼

kT  Cc 6pmR

(7.8)

Environmental and therapeutic aerosol particles frequently have a complex shape determined by the process condition in which they are produced. Typical therapeutic particles produced in precipitation or in spray drying processes, for example, have a compact spherical shape. Particles morphology is frequently modified during their production and porous (foam-like) structures are created. Particles can have a spherical or toroidal (for the development of an external surface) shapes and be empty inside. The particle material forms a surface shell. Sometimes, for special

7 Inhalation and Deposition of Nanoparticles

121

Fig. 7.4 Nanostructured therapeutic particles: (a) spherical shell and foamy particles, (b) toroidal particles, (c) mesoporous spherical particle Fig. 7.5 Diesel soot particle

purposes, spherical shell particles can have mezopores of controlled size distributed on their surface. Examples of the above mentioned structures are shown in Fig. 7.4. The diffusion coefficient of compact particles can be calculated using a classical Stokes–Einstein formula. There is an important group of particles with the shape of a significantly developed structure. Their shape could be characterized by fractal geometry (Hausdorff dimension), Fig. 7.5. Colloidal aggregates of small nanosized primary particles are the examples of an object for which a new approach for topological description and hydrodynamics should be used in comparison with compact (spherical) particles. The description of the structure of colloidal aggregates can be facilitated using the fractal model, Fig. 7.6. According to this, for relation between mass m of the cluster of primary particles of radius Rp and characteristic radius of the cluster, R is rendered by a power-law formula:
m ¼ kf 

R Rp

Df (7.9)

122

A. Moskal et al.

Fig. 7.6 Fractal-like diesel particle model

where kf is a prefactor that depends on the selected characteristic dimension R and Df is a fractal dimension of the cluster. Two characteristic cluster sizes are particularly important: the radius of gyration, Rg, and hydrodynamic radius Rh. Both, Rg and Rh, depend on the geometrical arrangement of particles in the cluster. The Rh is used for calculation of the diffusion coefficient of the cluster through the general expression: D¼

kT  TrðX1 Þ 3

(7.10)

where X is a friction tensor of the cluster. The relation between Rg, Rh and Rs – the maximum radius of the cluster, are elaborated by (Lattuada et al. 2003). Another approach for calculation of diffusion coefficient of clusters is given by (Moskal and Payatakes 2006). Diffusional deposition of the fractal-like aggregates using the Brownian dynamics method is shown by (Balazy and Podgorski 2007).

7.5

Classification of Diffusional Particle Deposition Models

As mentioned above, in our paper we consider the deposition of nanoparticles in the human respiratory system. For such particles the predominant mechanism of deposition is diffusion. At this moment, it is worth presenting a more general review of deposition modeling. There are considered two major descriptions of particle balance within the compartment they enter, namely Lagrangian and Eulerian. Continuum density rðX; tÞ and velocity U(X, t) are Eulerian fields and they are indexed by the position X in an inertial frame. Lagrangian fields are indexed by the position at the reference time t0. The common root for both descriptions in the case of analysis of the behavior of a diffusional aerosol particle within a defined region is the Langevin equation. It was originally proposed as

7 Inhalation and Deposition of Nanoparticles

123

a stochastic model for the velocity of a microscopic particle undergoing the Brownian motion. The stochastic process b(t) generated by the Langevin equation is called the Ornstein–Uhlenbeck process (Uhlenbeck and Ornstein 1930) and its probability density function evolves by the Fokker–Planck equation. In the terminology of stochastic processes b(t) is a diffusion process, and the Langevin equation is a stochastic differential equation. Let us consider the relationship between Lagrangian and Eulerian models, using the nanoparticles deposition process as an example.

7.5.1

Single Particle Trajectory Concept (Lagrangian Approach)

The object of our observation is the aerosol particle of diameter dp and density r, entering a compartment of arbitrary shape at position Xo (Fig. 7.7). The flow field within the compartment is given by vector U, and the particle has velocity V. According to Newton’s law, particle motion is described by: Stk

dV ¼ U  V þ FðX; tÞ þ AðtÞ dt

(7.11)

where velocities U and V are related to average gas velocity U, and time t is related to the characteristic time of the process. The first term on the right side of Eq. 7.11 represents interaction between the particle and surrounding fluid, which to be governed by Stokes’ law.  is2 assumed  rd U where d is a characteristic dimension of The Stokes number is defined as Stk dtp the compartment, U – an average gas velocity within the compartment, and F(X, t) is an external force acting on the particle (gravity, electrostatic, etc.). Another force A(t) acting on the particle is the result of collisions with molecules of the surrounding gas. F(X, t) and A(t) are both in a dimensionless form. These collisions make momentary changes of particle acceleration and A(t) has a random pattern with respect to its quantity and direction. The following principal assumptions are made for the fluctuating part A(t):·

Fig. 7.7 Limiting trajectory of aerosol particle in the compartment

124 l l l

A. Moskal et al.

A(t) is statistically independent of V(t) A(t) varies extremely rapidly in comparison with the variation of V(t) average in time of V(t) is zero

The second assumption implies that time intervals of duration Dt are such that during Dt the variations of V expected are very small indeed, while during the same interval A(t) may undergo several fluctuations. No correlation between A(t) and Dðt þ DtÞ exists. Equation (7.l1) in its general form is stochastic and describes the Brownian motion of an aerosol particle. It is called the Langevin equation in its general form (Schuss 1980). When the external force field F(X, t) does not exist or can be neglected in comparison to other effects, and gas molecules are in the state of disordered motion because of disappearing forced convective diffusion, particles interact with the surrounding fluid being in a chaotic state on molecular or turbulent levels. For the first case, the Brownian displacement of a particle is the predominant mechanism of particle motion. When the fluid is in turbulent disorder, the displacement of a particle also has a random form difficult to describe in terms of particle trajectory. The interaction of particles with the complex structure of an eddy (fluid particles) of different morphology could be rather described in the form of a population balance. When the fluctuation of aerosol particles caused by thermal interaction with surrounding molecules has significant influence on the displacement, the process of particle motion is a stochastic one. Langevin equation, which describes the motion of Brownian particles, has two terms: deterministic, representing an interaction of particle with surrounding fluid, aV, and f1uctuating A(t): dV ¼ aV þ AðtÞ dt

(7.12)

But “solving” a stochastic differential equation like Eq. 7.12 differs from the ordinary procedure of solving a deterministic differential equation. It has to be understood rather in the sense of specifying probability distribution pðV; t; V0 Þ which describes the probability that velocity has value V at time t, assuming V ¼ V0 at t ¼ 0. The function p has the following properties: pðV; t; V0 Þ ! dðV  V0 Þ when t ! 0

(7.13)

where d is Dirac’s d function, and pðV; t; V0 Þ tends to Maxwellian distribution for temperature T of the surrounding fluid, independently of V0 as t ! 1: !  m 32 mjVj2 (7.14) exp pðV; t; V0 Þ ! 2kT 2pkT where m is particle mass and k is the Boltzmann constant. The position of a Brownian particle is given in the form: Z t XðtÞ ¼ X0 eat þ ebðt;sÞ A ðsÞds (7.15) 0

7 Inhalation and Deposition of Nanoparticles

125

The integral in expression Eq. 15 can be calculated as the stochastic integral in the sense of Ito, (Schuss 1980). The construction of such an integral is a complex procedure so more useful practically as the integration of Eq. 15, using the Monte Carlo simulation. This procedure is simple and well described in the literature, (Zielin´ski 1970). More interesting and important is another property of Brownian particle motion described by the Langevin equation. Displacements of a particle obtained from the Langevin equation form Markov’s process, which is analogous to the diffusion process, (Gradon´ and Podgorski 1996). For that process the probability distribution function p(V, t, Vo) satisfies the Fokker–Planck equation. The evolution of density, s, of system particles in the phase space is defined in classical mechanics by the Liouville equation: ds @H @s @H @s ¼    dt @p @q @q @p

(7.16)

Where H is the Hamiltonian of the system (entire energy of the system) and p and q are generalized coordinates of the position and momentum of the system. If random effects are introduced into the system, then density s is described by the generalized Liouville equation, called the Fokker–Planck or prospective Kolmogoroff equation written in the form of evolution of the transition probability p(s, x, t, y), which is the probability of the event that the system being at time s in x will be in time t in y: dp @½aðy;tÞp 1 @ z ½bðy;tÞp  þ ¼ 0 p ! @ðx  yÞ @y dt 2 @yz

if t ! s

(7.17)

where a( y, t) is the deterministic drift of particles in the system and b(y, t)/2 is the diffusion coefficient (in the probabilistic sense). If we are interested in the distribution of Brownian particles for time intervals Dt, r dp 2 Cc Dt  t, very large in comparison with the time of particle relaxation t ¼ p18m then using the analogy of derivation of the Fokker-Planck equation, we receive the diffusion equation (Eq. 7.18).

7.5.2

Eulerian Approach

With this path of transformations we came to the deterministic equation describing diffusional particle population balance i.e. the Eulerian approach of a model. The balance equation within a controlled differential space has the following form in the dimensionless coordinate system: 1 @C þ Pe  Vr2 c þ B ¼ 0 Fo @t

(7.18)

126

A. Moskal et al.

where t is reduced time with respect to the characteristic time of the process (for example, breathing cycle time T) and the space coordinates are reduced with respect to the characteristic cross-sectional dimension of compartment d, Fourier number Vd Fo ¼ DT d2 , Pe ¼ D , and D is the diffusional coefficient of a Brownian particle. C means particle concentration resulting from the integral of particles probability distribution over the particle velocity space. Thus C means the number of particles in the control space volume at moment t. The first term in Eq. 7.18 means the accumulation of particles of a given size; the second term means the net transport of particles caused by ordering effects of forced convection (or other acceleration field); the third term means the dispersion of particles due to Brownian and/or turbulent diffusion and the fourth term means the sink of particles due to their deposition or due to changes in their size caused by coagulation, evaporation or chemical reaction with surrounding gas. The last term for the deposition process can be expressed as: B¼


 1 @C Pe  Vn  C  rp @n

(7.19)

where rp – particle diameter, Vn – velocity component of particle normal to the surface deposition, n – direction normal to the surface of deposition. The integration of Eq. 7.18 with appropriate boundary and initial conditions gives us the possibility of balancing particles within considered compartments and then calculating the efficiency of deposition.

7.6

The Results of Recent Modeling of Nanoparticles Deposition

Recent advances in computational software have made it possible to solve threedimensional transport equations numerically to provide information on local and regional deposition of nanoparticles. The geometric complexity of the mouth-to-trachea region causes a large variety of transition and turbulent effects observed for any flow rates of inhaled aerosol. Simple and more realistic models of the upper airways are used for flow modeling and submicron particle deposition, (Cheng et al. 1999; Renotte et al. 2000; Stapleton et al. 2000; Katz and Martonen 1999; Zhang et al. 2002; Zhang et al. 2004). The k-e, k-o and LES turbulence models were used for the description of a gas flow pattern in the oral airways models. The deposition of nanosized particles in a sequence of bifurcations of the tracheobronchial tree are reported by Longest and Xi (2007), Balashazy et al. (1999), Zhang and Kleinstreurer (2002). The above authors use Eulerian– Lagrangian or Eulerian–Eulerian models for particles transport.

7 Inhalation and Deposition of Nanoparticles

127

There are only few publications reported in the literature for modeling submicron particles in the alveolar region of the respiratory system, where the alveolar model incorporates the constitutive equation of a lung tissue (Gradon´ et al. 2000 Dailey and Ghadiali 2007). The authors of the above mentioned papers describe the details of used geometries of objects, more or less realistic flow patterns, focusing their attention on the model used for the fluid flow. For upper airways deposition the above authors indicate the advantages and disadvantages of turbulent models k-e, k-o or LES, depending on which model they use, without any special reflection on what essence of the process they work on is. The last step of modeling is a simple introduction of the equations of particle motion. There is no controversy between the aforementioned authors. The experimental data of particle deposition in the respiratory system indicates the influence of the breathing pattern – the shape of a real inspiratory-expiratory curve. Time dependent flow evolution during breathing shows how complicated patterns of temporary and local effects could be achieved, especially when the transition between inspired and expired streams of air causes complex circulation influencing the deposition of diffusional particles at any part of the respiratory system. Let us consider an object of simple geometry, i.e. a segment of the bent tube of the diameter D ¼ 20 mm and radius of bending R ¼ 70 mm. The segment is ended with two cylinders of the lengths equal to the tube diameter (Fig. 7.8). Aerosol particles of diameter 100 nm are introduced into the system with two patterns of the flow rate: the first, frequently used for deposition analysis, uniform in time with constant volumetric flow rate Q ¼ 1 dm3/s and the second one corresponding to the natural pattern of breathing defined by the time dependent volumetric flow rate QðtÞ ¼ p2 sin ðp2 tÞ; (Fig. 7.9). The exposure time for both flowrates corresponds to the inspiratory time of breathing and is equal to 2 s. The inhaled volume of aerosol for both patterns is the same and equals 2 dm3. The question arises whether the temporary and local deposition of aerosol particles depends on the type of a flow rate. The k-o model of turbulence was used to calculate the air flow structure within the tube and the Lagrangian approach – to calculate particle deposition. As a result of the above analysis, temporary and spatial deposition efficiencies of aerosol particles in the object

Fig. 7.8 Geometry of the bent tube

128

A. Moskal et al.

Fig. 7.9 Two flow patterns through the bent tube

Fig. 7.10 Temporal deposition of the 100 nm aerosol particles in the bent tube for steady and unsteady flows

were calculated. The temporal deposition TD(t) is defined as a ratio of mass deposited at the Dti, i.e. the flux m(ti) over the surface area of the object, to the total mass deposited in the object: TDðtÞ ¼ PN

miðDti Þ

i¼o

½½mi ðDti Þ

(7.20)

The results of calculations for both types of the flow rates are shown in Fig. 7.10. For the steady flow, temporal deposition, except for the first time steps, is almost uniform and is significantly lower than that for the unsteady flow, where the linear velocity of air increases in time and has its maximum in the middle of inspiration. There are only few papers published recently, in which the authors consider the deposition of particles for an unsteady flow pattern. In one of them, (Zhang and

7 Inhalation and Deposition of Nanoparticles

129

Fig. 7.11 Instantaneous deposition efficiency in the bent tube for steady and unsteady flows

Kleinstreuer 2004), the authors calculate the deposition of diffusional particles in the oral airways during the inspiratory cycle. They use the following definition of instantaneous deposition efficiency: mi ðDti Þ IDEðDtÞ ¼ R t ti1 QðtÞdt

Dt ¼ ti  ti1

(7.21)

which is a ratio of mass deposited during the time step Dt to the mass introduced into the system during Dti. Taking this controversial definition, we calculated the IDE for steady and unsteady flow patterns. The results shown in Fig. 7.11 confirm the differences in deposition for steady and unsteady flows. Additionally, we have calculated the local deposition of particles in this simple geometry for both types of flow rates. The results show significant difference in the distribution of the ‘hot spots’ of particle deposition. For the steady flow, for which the linear velocity of the gas is constant and reasonably high from the beginning of the exposure, the ‘hot spots’ of deposition are located in the middle of the bent tube. For an unsteady flow, when the velocity of air increases from zero to maximum in the middle of inspiration, the ‘hot spots’ are shifted towards the exit of the compartment. Summarizing this part of our consideration, we conclude that for the objective information about the temporary and local deposition patterns of nanoparticles in any part of the respiratory system a natural breathing pattern (or at least its simplified analytical form) should be used. It is discussed in more detail in the next section of this chapter.

7.6.1

Practical Issues of Ex-vivo Studies of Particle Deposition in Models of Human Airways

In practice, a quantitative analysis of aerosol deposition in the respiratory system can be accomplished with three types of approach: (i) in vivo, i.e. by experiments on

130

A. Moskal et al.

living subjects, (ii) in silico, i.e. by computer modeling, and (iii) in vitro, i.e. doing experiments with adequate artificial models of physiological objects. Data from in vivo studies are apparently the most reliable. However, due to a difficult control of experimental constraints (fixed airways geometry, airflow, etc.) they demonstrate a large scatter and irreproducibility. The amount of useful data from such studies is limited owing to ethical issues and relatively high costs, so their statistics is often unsatisfactory. Moreover, the spatial and temporal resolution of in vivo studies is usually too low to provide precise information on how the distribution of particle deposition depends on particle size, breathing pattern, aerosol properties, etc. Last but not least, in vivo data are typically obtained from healthy volunteers, so their validity in the discussion of deposition patterns in diseased lungs can be questionable. In this part of the paper we will focus only on practical aspects of in silico and in vitro studies done for selected parts of the respiratory system. These methods have an advantage in terms of precise control over simulation/experimental conditions, a large number of tests can be done in a relatively short time and at reasonable cost, and a wide range of parameters available for studies (in contrast to in vivo tests). As already discussed in the previous section, the factors to be considered in a methodologically correct analysis of aerosol particles behavior following their inhalation are as follows: (a) a representative geometry of the airways structure, (b) a realistic airflow pattern (as the function of, e.g., health status, activity level, etc.), (c) for in silico models: the proper numerical simulator of an airflow and of aerosol particles dynamics, both combined with appropriate boundary conditions. This is of special importance for nano-sized and nano-structured particles which have to be analyzed in a special way taking into account both their size (Brownian diffusion must not be neglected) and their unique geometry (dendrite-like structure). By considering the above factors during in silico/in vitro studies, we can better understand the physical mechanisms responsible for health effects resulting from aerosol nanoparticles inhalation, which is equally important both from a toxicological and a therapeutic point of view.

7.6.2

In Silico and In Vitro Studies of the Extrathoracic Region

To study the influence of inhalation patterns on aerosol flow and deposition in the mouth and throat we will focus on a simplified but standard (therefore: reproducible) geometry of the USP inlet, sometimes called the ‘USP throat’ (US Pharmacopeia 1996). As shown in Fig. 7.12, it holds the important feature of the human oropharynx, i.e. a strong bend of the air passage. This model can be used for the parallel computational fluid mechanics (CFD) analysis of aerosol dynamics and

7 Inhalation and Deposition of Nanoparticles

131

Fig. 7.12 USP inlet (USP “throat”) geometry used for in silico (a) and in vitro (b) studies Fig. 7.13 The instantaneous airflow field in the simplified model of the human mouth and throat

in vitro deposition measurements. The idealized, but still a realistic breathing pattern can be described by: h  p i VðtÞ ¼ A 1  cos t B

(7.22)

where: A , B are constants, V(t) is the volume of air that flows through the mouth. Equation 7.21 after differentiation results in the instantaneous airflow rate in the mouth, which can be used as a boundary condition in numerical modeling. The commercial CFD package (Fluent ver. 6.2.16 with Gambit pre–processor ver. 2.0 – Fluent Inc., Lebanon, NH) has been used to calculate the fluid flow in the model. The structured numerical mesh contained 4·105 hexagonal elements. During reallike inhalation the airflow is changing from laminar to turbulent, therefore the Large Eddy Simulation (LES) model was used to handle the peculiarities of the unsteady flow. For good and fast convergence the time step of 1 ms was used in our calculations. Figure 7.13 shows the contours of air velocity magnitude calculated at three time-instants of the inhalation curve described by Eq. 7.21 with the parameters: A ¼ 1.368 dm3 and B ¼ 2.75s. The velocity distributions calculated

132

A. Moskal et al.

Fig. 7.14 Example of calculated temporal deposition of 300 nm particles in the first two elements of the model throat (Fig. 7.12a) during sinusoidal airflow

for the average steady flow are also shown to demonstrate that even for those instants when the value of the flow rate is equal to the average, the velocity fields are completely different. This proves that the results of the steady-state calculations done for the average value of the flow rate falsify the complex behavior of air flowing through the airways. Due to the lack of physical background a simplified steady-flow approach can be a source of erroneous predictions of aerosol deposition in the human breathing system. The influence of airflow unsteadiness on the 300 nm particles deposition was computed employing the Lagrangian trajectory approach described in the previous section. As a result, temporal deposition in different parts of the geometry was calculated (Fig. 7.14). The deposition intensity varies among different locations with time of inspiration. Interestingly, the peak of deposition rises at the beginning of expiration as a result of the deposition of particles which have not been deposited during inspiration but still remain in the computational domain. These results demonstrate that a priori averaging of a dynamic and spatially inhomogeneous problem (e.g., by assuming an average flow or using an average dimension to characterize complex geometry) oversimplifies the process and may lead to overlooking some phenomena which can be essential for health effects (e.g., the localized hot-spots of deposition). The importance of flow variations for spatial distribution of deposition was also confirmed during the in vitro experiments done with medicinal, micrometer-sized particles. The aluminum ‘USP throat’ was manufactured as a three-compartmental assembly shown in Fig. 7.12b, consisted of the inlet zone (i.e., mouth), the bend (throat) and outlet zone (pharynx). Using a programmable air pump incorporated in the Artificial Lung Apparatus (ALA) a real-like inspiratory curve was produced, as shown in Fig. 7.15. Constant flow conditions (45, 60 and 90 LPM) were also studied using a vacuum pump. With all flows, DSCG (disodium cromoglycate) aerosol generated from a DPI (dry powder inhaler) was drawn through the airway model, the interior of which was coated with a thin layer of a viscous liquid to avoid powder rebound and re-entrainment. The mass of particles deposited in each compartment was determined with a chemical assay of aqueous washings. The results in Fig. 7.16 show that the regional distribution of particle deposition in the USP throat is different for a real-like inhalation pattern when compared to the

7 Inhalation and Deposition of Nanoparticles

133

Fig. 7.15 (a) Experimental set-up with the Artificial Lung Apparatus (ALA), and (b) one of the real-like non-steady inhalation patterns produced with ALA Fig. 7.16 Local depososition of aerosol particles in different parts of the USP throat vs. airflow pattern. 45, 60, 90 LPM – constant flows, INH – real-like flow depicted in Fig. 7.15b

depositions obtained at constant flows (60 and 90 LPM are approximately equal to the average and peak airflow rates during real-like inhalation, respectively). In particular, the deposition in the inlet (part A) is overestimated in constant flow experiments comparing to the real-like flow. Analogous results are found for the overall deposition efficiency in the model. The conclusions from this simplified geometry are valid also for the more realistic mouth-and-throat geometry as demonstrated by Sosnowski et al. (2007a, b). In accurate predictions of deposition, real particles morphology must be accounted for as well. For nano-structured particles the typical geometry is nonspherical, since they usually have the form of aggregates composed of many primary nanoparticles (typically with the size of 5–50 nm, Friedlander and Pui 2003 attached together by the Van der Waals forces. Such aggregates can be of the size from about 100 nm to several micrometers (Friedlander 2000). The aggregates create composite geometrical structures of fractal-like complexity. As discussed in the previous sections of the paper, many physical quantities of fractal-like aggregates are related to their size by a power-law. The relation used to describe such aggregates is that between the total number of primary particles in a given aggregate and its radius (Eq. 7.9). The fractal dimension characterizes the aggregate structure and gives information about the structure compactness. The aggregates with Df close to 3 (~2.5) have a spherical compact structure but when Df goes to

134

A. Moskal et al.

Fig. 7.17 Examples of the aerosol aggregates. Both structures contain 100 primary particles with diameter equal to 100 nm

Fig. 7.18 The diffusion coefficient for aggregates as a function of the number of primary particles composed in the structure for three different models

1 (~1.6), the structure becomes more open and dendrite-like (Fig. 7.17). Three main approaches to model the dynamics of aerosol aggregates are generally used: (i) the aggregate is treated as a porous sphere, (ii) the behavior of the aggregate is mimicked by the equivalent sphere with its radius equal to the hydrodynamic radius of the aggregate, (iii) the aggregate is treated as a rigid body assembled from spherical primary particles. The first two approaches are widely used e.g., (Bałazy and Podgo´rski 2007; Gmachowski 2008; Lattuada et al. 2003) but they are restricted only to aggregates containing several thousand primary particles to fulfill the assumption of spherical symmetry in a cluster. The model proposed by Moskal and Payatakes (2006) based on classical rigid body mechanics can be used even for the simplest aggregates containing only a few primary particles. Based on the assumption that only three main forces (Brownian, drag and gravity) are acting on the primary particle assembled in the cluster structure, the present authors proposed a set of differential equations which allow to determine the trajectory of a given aggregate in a given environment, taking into account translational and rotational displacement. Using the proposed approach, the diffusion coefficient of aggregates can be evaluated. Figure 7.18 shows the comparison of the predicted aggregate diffusion coefficient as a function of the number of primary particles in the aggregate, obtained from different models. The influence of aggregate morphology on deposition in the human airways was investigated in a simplified model of the nasal cavity using the equivalent sphere radius concept Moskal et al. (2006). Four

7 Inhalation and Deposition of Nanoparticles

135

Fig. 7.19 Deposition of aerosol aggregates in the model of human nose as a function of aggregates fractal dimension

different populations of aggregates with the size range from 250 to 312 nm and the fractal dimension from 1.7 to 2.1 were created using the diffusion limited aggregation (DLA) algorithm. Their trajectory and deposition in the nasal cavity was calculated using the Lagrangian approach incorporated into Fluent ver. 6.2.16 CFD package. Figure 7.19 illustrates how the fractal dimension of aggregates is related to particle deposition in the model nasal airways. The lack of strong tendency in the relation between deposition and fractal dimension can be explained by the imperfection of equivalent sphere estimation which does not take into account the real complexity of the aggregates structure approximated by the perfect spherical shape.

7.6.3

In Silico and In Vitro Studies of Bronchial Bifurcation

The results of in silico analysis of aerosol flow and deposition during realistic nonsteady flow conditions in the idealized asymmetrical double bifurcation representing the geometrical structure repetitively met in the bronchial airways has been presented in details by Moskal and Gradon´ (2002). The geometry of the bifurcating tubes implemented into the CFD model is shown in Fig. 7.20. The authors demonstrated that for nano-range particles deposition in the bifurcation is elevated during the flow-reversal (inhalation–exhalation), which is a new information on temporal distribution of bronchial deposition. Experimental validation of the in silico results done by Gradon´ et al. (2003) using the plastic hollow cast of the bifurcation confirmed the formation of deposition hot-spots in the locations predicted by the computations. Additional data have been obtained from in vitro studies using three different inhalation patterns depicted in Fig. 7.21, which were simulated with the ALA system. The distribution of particle deposition in the bifurcating

136

A. Moskal et al.

Fig. 7.20 The representative meshed geometry of the first two bifurcation of bronchial tree

Fig. 7.21 Visualization of in vitro particle deposition in the model bronchial bifurcations as a function of breathing pattern

tube is noticeably different depending on the inhalation profile and the peak inspiratory flow. These results confirm that different dynamics of breathing (which can be the result, e.g., of variable activity or lung pathology) has an influence on the local efficiency of particle deposition also in the bronchial airways.

7 Inhalation and Deposition of Nanoparticles

7.6.4

137

In Silico and In Vitro Studies of Pulmonary Region

In comparison to other regions of the respiratory tract, the information on aerosol particle behavior in the pulmonary zone is limited due to the fact that the geometrical structure of this region is extremely complex. There are millions of fine tubes and small alveoli, which continuously change their size during breathing. It is not possible to reconstruct a reliable geometrical model of the whole region, so we have to focus on simpler elements. Proper modeling of the local airflow first requires information on the pattern of alveolar wall motion. For that the mechanical properties of the alveolar tissue should be known. There are only a few publications which take into account the mechanical aspects of alveolus oscillations (e.g., Orlicki and Gradon´ 1996; Dailey and Ghadiali 2007). On the other hand, because of the lack of suitable mechanical data on the human lung, a different approach can be proposed. In a simple geometrical model of an individual alveolus (Fig. 7.22), an arbitrary function describing the time-variations of the radius of an alveolus can be proposed as: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 h  p i 9B 3 A 1  cos (7.23) t RðtÞ ¼ R0 3 þ 4 p3 B where A and B are constants, and R0 is the minimum radius of the alveolus in the cycle, equal to the radius of the attached bronchiole. Equation 7.23 can be incorporated into the CFD code with the Moving Mesh model (e.g., Fluent) to describe the moving boundaries in the simulations of the alveolar airflow during the breathing cycle. Using this techniques the fluid – structure interaction can be modeled without employing the stress balances relationship and viscoelestic constitutive relationship for tissue deformation. Equation 7.23 can be also modified in a way to mimic pathological cases of radius variation. Figure 7.23 presents an example of the velocity magnitude in the model of an oscillating alveolus. Knowing the time-dependent airflow field, the deposition of particles in the alveolus can be calculated in the common way using the Lagrangian approach.

7.7

Interaction of Deposited Particles with the Lung Surfactant

The problem of deep lung deposition is specific also because of the extraordinary properties of the fluids covering the tissue in that region. They influence the fate of nano-sized and nano-structured particles deposited there. In contrast to bronchial

Fig. 7.22 The spherical model of the alveolus

138

A. Moskal et al.

Fig. 7.23 The air velocity field in a model of alveolus

airways, where the tissue is covered with the mucus and the movement of cilia creates the mechanism of airways clearance from deposited particles (‘mucociliary escalator’), the alveoli are lined with a very thin layer of liquid (hypophase) containing the lung surfactant (LS), which is partly responsible for particle clearance. The presence of surface active compounds in the alveoli, which undergo dynamic variations of the shape and area during breathing, has biophysical consequences. By lowering the surface tension during expiration, the surfactant facilitates homogeneous lung expansion at air inspiration and eases up the work of breathing. It was proposed (Schu¨rch et al. 1990) that by promoting the particles wetting, the surfactant facilitates their dislocation towards lung epithelium. However, for particle-lung interactions it is also important that under dynamic conditions caused by area oscillations, there is propagation of surface tension gradients, which can lead to the tangential flows of liquid (Marangoni effects) capable of displacing the immersed particles and cells (e.g., alveolar macrophages – AMs) from alveoli to bronchioles. The schematic representation of such action is depicted in Fig. 7.24. This mechanism was quantitatively analyzed in silico by Gradon´ and Podgo´rski (1989), Podgo´rski and Gradon´ (1993), and later by Espinosa and Kamm (1997). Such transport was also verified in several in vitro studies, e.g. Gradon´ et al. (1996), Sosnowski et al. (2003), Sosnowski (2004), which indirectly supported the role of LS in hydrodynamic mechanisms of alveolar clearance from particulate deposits. It was also postulated that LS may play role in stimulating alveolar macrophages by means of hydrodynamic signaling (Gradon´ and Podgo´rski 1995; Sosnowski 2001). All discussed defense functions of LS may be fulfilled only when the interfacial activity of the surfactant is maintained. It can be evaluated by means of the surface tension hysteresis found during the breathing-induced area oscillations of alveolar hypophase. It is documented that a decrease of hysteresis is connected with surfactant inactivation in vivo and with respiratory failure (Notter

7 Inhalation and Deposition of Nanoparticles

139

Fig. 7.24 Schematic of the hydrodynamic clearance of the alveolus from deposits due to the lung surfactant activity (Marangoni effects)

et al. 1982; Miller et al. 2005). Hysteresis can be measured in vitro under simulated breathing-like area oscillations, e.g., by the oscillating bubble method or with the Langmuir–Wilhelmy film balance (Sosnowski and Podgo´rski 1999), and then correlated with the physiological function of the lung surfactant, e.g., (Sosnowski et al. 2000; Sosnowski 2003). A convenient quantitative parameter for hysteresis characterization is the area enclosed by the loop, normalized against the value measured for the LS with the standard quality (STD): h

i hR i R Amax Amax Amin sdA expansion  Amin sdA compression

(7.24) c ¼ h i h i R Amax R Amax sdA  sdA Amin Amin expansion

compression

STD

Discussing the potential health effects caused by inhaled nanoparticles at the alveolar level, one can speculate about their direct physicochemical interactions with the lung surfactant. Since such health effects are known to correlate with particles surface area rather than their mass e.g., McClellan (1996), it seems plausible that interfacial phenomena may be of key significance. We have studied in vitro the possibility of LS inactivation by aerosol nanoparticles (20–100 nm) emitted from the graphite aerosol generator (model GFG-1000, Palas GmbH, Germany) by exposing the samples of bovine surfactant (Survanta – Abbott Laboratories). The aerosol concentration was 5 mg/m3 and the LS samples (0.5 ml each) were exposed to the aerosol up to 2 hrs. The surface tensions hysteresis was measured with Pulsating Bubble Surfactometer (Electronetics Corp., USA), where a tiny air bubble (0.8 mm in diameter) was pulsated with a frequency of

140

A. Moskal et al.

Fig. 7.25 Reduction of surface tension hysteresis area (indicated by c – Eq. 7.23) during lung surfactant exposure to graphite nanopatrticles

0.25 Hz in the LS sample kept at 37 C. The experiments allowed to evaluate the parameter c (Eq. 7.23) for each sample. The cumulative results shown in Fig. 7.25 indicate that the surfactant biophysical activity is partially reduced by nanoparticles deposited in the liquid, and that this response is time-dependent. The most probable mechanism here is the adsorption of LS molecules on large surface of aggregatetype particles. A local decrease of surfactant concentration in the liquid and on the air–liquid interface results in the alteration of surface-tension variations during bubble oscillations. Similar results have been obtained by Sosnowski et al. (2000) during surface activity studies of a model LS (Infasurf) in mixtures with soot particles. Recently, Bakshi et al. (2008) demonstrated that the surfactant phospholipids are inactivated in vitro by metallic nanoparticles. In any case, the reduction of biophysical activity of the LS, if it happens in vivo, may induce certain physiological effects (e.g., a lower clearance rate, and – in high doses – respiratory symptoms). The relationship between the surface area of deposited particles and the loss of surfactant activity has been also demonstrated for nanostructured silica particles (Sosnowski et al. 2003; Sosnowski 2004), the examples of which are shown in Fig. 7.4b, c. It was demonstrated that with the same mass load, particles with different morphology induce the response proportional to their interfacial area. The value of c dropped from 1 (for the control) to 0.53 (0.15) for nonporous toroidal particles, and to 0.37 (0.03) for mesoporous particles. This implies that mesoporous insoluble particles can adsorb higher amounts of LS from the liquid in their surrounding than nonporous particles, so in this way, by reducing the rate of surfactant-mediated clearance, they can stay longer in the region of deposition. These observations open up the possibility of applying engineered particles in specialized drug delivery to the lungs. Targeting alveoli with adjusted amounts of open-structured (porous) particles will cause their prolonged residence time due to

7 Inhalation and Deposition of Nanoparticles

141

a controllable decrease in the local LS concentration. If the active pharmaceutical ingredient (e.g., in the form of nanoparticles) is loaded into such micrometer-sized porous carriers, it will be slowly washed-out by the surfactant while the carrier remains in the alveolar region. This allows the sustained-release of a drug delivered to the deep lungs, giving an opportunity of using nano-structured and nano-sized particles in the modern systems of inhalation drug delivery. On the other hand, overloading the lungs with large amounts of nanoparticles characterized by a high surface area can reduce LS concentration so much that the clearance functions are noticeably impaired. This mechanism explains the hazards posed by inhaling high doses of nanoparticles, e.g., in occupational environments.

7.8

Conclusions

Advanced analysis of a nanoparticle aerosol flow and deposition in different parts of the respiratory system requires a physically-correct description of the phenomena governing the airflow dynamics and particle motion. Starting from the basic equations which describe these processes under various circumstances, we intended to stress the importance of the proper definition of the physical problem to be solved either numerically (in silico) or experimentally (in vitro). We focused only on selected issues (correct flow representation, advanced analysis of aggregate-type particles motion, flow dynamics in the geometries with moving walls) to demonstrate currently available tools for more thorough modeling. It was pointed out that averaging time-dependent and spatially inhomogeneous processes of an aerosol flow and deposition can be a source of oversimplification which does not enable a realistic assessment of health effects resulting from aerosol inhalation. The comprehensive model of any process or phenomena should emphasize the set of operational condition and parameters the process is sensitive for. Deposition of nanosized particles occurs at the diffusional boundary layer of the air attached to the respiratory wall. The thickness of this boundary depends on the flow patterns. Our investigations show that for real breathing pattern, i.e. cyclic flow, flow structure is more complex than for steady-state flows used regularly in recently presented models. Complex turbulent structures in upper airways and at the bifurcations of tracheobronchial tubes, significantly changes deposition pattern of diffusional particles, which are strongly involved in such flow structures. The same is in the alveolar region when inspired and expired air, due to sequence of stretching and folding patterns, reduces thickness of the boundary layer. It facilitates the diffusional particle deposition. Concluding that part of our remarks we noticed that the proper and deep analysis of the gas flow during breathing is a key point for the successful modeling of nanoparticles deposition. Ignoring this effect we lose important information about the process. Lagrangian and Eulerian models of nanoparticle deposition are already well defined when the flow structure is properly described. We have also showed that it can be important to extend the analysis of the deposition process to particle-lung interaction problems which are also

142

A. Moskal et al.

determined by physicochemical mechanisms (e.g., in alveolar region). Including the discussed issues into ex vivo modeling, it is possible to improve the in vitroin vivo correlation (IVIVC), which is not satisfactory by now. In the case of nanoparticles our knowledge is even less complete. We hope that with the presented techniques of modeling and experimental investigation the understanding of nanoparticles deposition and resulting in health effects can be improved in the near future. Acknoledgement This work was supported by the grant PBZ-MEIN-3/2/2006: “Process engineering for the abatement of harmful and greenhouse gases emissions and their utilization”.

References Bakhsi MS, Zhao L, Smith R, Possmayer F, Petersen NO (2008) Metal nanoparticle pollutants interfere with pulmonary surfactant function in vitro. Biophys. J. 94:855–868 Balashazy I, Hoffman W, Heistraicher T (1999) Computation of local enhancement factors for the quantification of particle deposition pattern in airway bifurcation. J. Aerosol Sci. 30:185–203 Balazy A, Podgorski A (2007) Deposition efficiency of fractal-like aggregates in fibrous filters calculated using Brownian dynamics method. J. Colloid Interface Sci. 311:323–337 Cheng YS, Zhou Y, Chen BT (1999) Particle deposition in a cast of human oral airways. Aerosol Sci. Tech 31:286–300 Dailey HL, Ghadiali SN (2007) Fluid–structure analysis of microparticle transport in deformable pulmonary alveoli. J. Aerosol Sci. 38:269–288 Daley BJ, Harlow FH (1970) Transport equations in turbulence. Physics of Fluids 13:2634–2649 Espinosa FF, Kamm RD (1997) Thin layer flows due to surface tension gradients over a membrane undergoing nonuniform, periodic strain. Ann. Biomed. Eng. 25:913–9125 Friedlander SK (2000) Smoke, Dust, and Haze: fundamentals of aerosol dynamics. Oxford University Press, New York Friedlander SK, Pui D (2003) NSF Workshop Report on Emerging Issues. In: Nanoparticle Aerosol Science and Technology (NAST). University of California, Los Angeles. [accessed on 22 Jan, 2006] http://nnco5.nano.gov/html/res/NSFAerosolParteport.pdf Gmachowski L (2008) Free settling of aggregates with mixed statistics. Coloids Surf. A: Physicochem. Eng. Aspects 315:57–60 Gradon´ L, Orlicki D (1990) Deposition of inhaled aerosol particles in a generation of the tracheobronchial tree. J. Aerosol Sci. 25:3–19 Gradon´ L, Podgorski A (1996) Deposition of inhaled particles. Discussion of present modeling techniques. J. Aerosol Medicine, 9:343–355 Gradon´ L, Podgo´rski A (1989) Hydrodynamical model of pulmonary clearance. Chem. Eng. Sci. 44:741–749 Gradon´ L, Podgo´rski A (1995) Displacement of alveolar macrophages in the air space of human lung. Med. Biol. Eng. Computing 33:575–581 Gradon´ L, Podgo´rski A, Sosnowski TR (1996) Experimental and theoretical investigations of transport properties of DPPC monolayer. J. Aerosol Med. 9:357–367 Gradon´ L, Orlicki D, Podgorski A (2000) Deposition and retention of ultrafine aerosol particles in the human respiratory system. Normal and pathological cases, JOSE 6:189–207 Gradon´ L, Sosnowski TR, Moskal A (2003) Resuspension of powders and deposition of aerosol particles in the upper human airways. In: Gradon´ L, Marijnissen J (eds) Optimization of aerosol drug delivery. Kluwer, Dordrecht, pp 123–137

7 Inhalation and Deposition of Nanoparticles

143

Horsfield K, Cumming G (1968) Morphology of the bronchial tree in man. J. Appl. Physiol. 24:373–383 Katz IM, Martonen TB (1999) A numerical study of particle motion within the human larynx and trachea. J. Aerosol Sci. 30:173–183 Langes J (1989) Clinical anatomy of the nose, nasal cavity and paranasal sinuses. Blackwell Scientific Publ, Cambridge Lattuada M, Wu H, Morbidelli M (2003) Hydrodynamic radius of fractal clusters. J. Colloid Interface Sci. 286:96–105 Longest PW, Xi J (2007) Computational investigation of particle inertia effects on submicron aerosol deposition in the respiratory tract. J. Aerosol Sci. 39:111–130 McClellan RO (1996) A mechanistic approach to assessing the lung cancer risk of diesel exhaust and carbon black. In: Marijnissen JCM, Gradon´ L (eds) Aerosol inhalation: recent research frontiers. Kluwer, Dordrecht, pp 27–78 Miller E, Piroz˙yn´ski M, Jastrze˛bski J, Pawelec M, Sosnowski TR (2005) Pulmonary surfactant function in patients with acute lung injury/acute respiratory distress syndrome treated by mechanical ventilation. Critical Care, 9(S1):39–40 Moskal A, Gradon´ L (2002) Temporary and spatial deposition of aerosol particles in the upper human airways during breathing cycle. J Aerosol Sci. 33:1525–1539 Moskal A, Makowski Ł, Sosnowski TR, Gradon´ L (2006) Deposition of fractal-like aerosol aggregates in a model of human nasal cavity. Inhalation Toxicology 18:725–731 Moskal A, Payatakes AC (2006) Estimation of the diffusion coefficient of aerosol particle aggregates using Brownian simulation in the continuum regime. J. Aerosol Sci. 37:1081–1101 Notter RH, Taubold RM, Mavis RD (1982) Hysteresis in saturated phospholipid films and its potential relevance for lung surfactant functions in vivo. Exp. Lung Res. 3:109–127 Orlicki D, Gradon´ L (1996) The influence of normal and pathological inhomogeneities of the lung tissues on particle transport and deposition. In: Marijnissen JCM, Gradon´ L (eds) Aerosol inhalation: recent research frontiers. Kluwer, Dordrecht, pp 195–204 Podgo´rski A, Gradon´ L (1993) An improved mathematical model of hydrodynamical self-cleansing of pulmonary alveoli. Ann. Occup. Hyg. 37:347–365 Pope SP (2000) Turbulent flows. Cambridge University Press. Cambridge Renotte C, Bouffioux U, Wilquem F (2000) Numerical 3D analysis of oscillatory flow in the time varying laryngeal channel, J. Biomechanics 33:1637–1644 Schuss Z (1980) Theory and Application of Stochastic Differential Equations. J. Wiley, New York Schu¨rch S, Gehr P, Im Hof V, Geiser M, Green F (1990) Surfactant displaces particles toward the epithelium in airways and alveoli. Respir Physiol. 80:17–32 Sosnowski TR (2001) Sorption-induced Marangoni microflows in the pulmonary surfactant system. Inz˙. Chem. Proces. 22:251–267 Sosnowski TR (2003) Dynamic surface tension as an indicator of lung function in health and disease. Biocyb. Biomed. Eng. 23:89–98 Sosnowski TR (2004) Interactions of deposited particles and lung liquids (surfactants). In: Sheuch G (ed) Aerosole in der Inhalationstherapie. Dustri-Verlag Dr. Karl Feistle, Mu¨nchen, pp 10–20 Sosnowski TR, Podgo´rski A (1999) Assessment of the pulmonary toxicity of inhaled gases and particles with physicochemical methods. Int. J. Occupational Safety and Ergonomics 5:433–449 Sosnowski TR, Gradon´ L, Podgo´rski A (2000) Influence of insoluble aerosol deposits on the surface activity of the pulmonary surfactant: a possible mechanism of alveolar clearance retardation? Aerosol Sci. Technol. 32:52–60 Sosnowski TR, Gradon´ L, Iskandar F, Okuyama K (2003) Interaction of deposited aerosol particles with the alveolar liquid layer. In: Gradon´ L, Marijnissen J (eds) Optimization of aerosol drug delivery. Kluwer, Dordrecht, pp 205–216 Sosnowski TR, Moskal A, Gradon´ L (2007a) Mechanims of aerosol particle deposition in the oropharynx under non-steady airflow. Ann. Occup. Hyg. 51:19–25

144

A. Moskal et al.

Sosnowski TR, Gradon´ L Moskal A (2007a) Regional deposition pattern of aerosol particles in the replica of upper airways. Proceedings of EAC2007 (CD), Salzburg, 9–14.09.2007, Abstract T08A006 Stapleton KW, Guentsch E, Hoskinson MK, Finlay WH (2000) On the suitability of k-e turbulence modeling for aerosol deposition in the mouth and throat. J. Aerosol Sci. 31:739–749 Uhlenbeck GH, Ornstein LS (1930) On the theory of Brownian motion, Phys. Review 36:824–841 United States Pharmacopeial Convention (1996) General tests and assays: preparations for inhalations. Pharmacopeial Forum 22:3048–3098 Weibel R (1963) Morphometry of the human lung. Springer-Verlag, Heidelberg Zhang Z, Kleinstreuer C (2002) Transient airflow structures and particle transport in a sequentially branching lung airway model. Physics and Fluids 14:862–880 Zhang Z, Kleinstreuer C (2004) Airlfow structures and nano-particle deposition in human upper airway model. J. of Computational Physics 198:178–210 Zhang Z, Kleinstreuer C, Kim CS (2002) Micro-particle transport and deposition in a human oral airway model. J. Aerosol Sci. 33:1635–1652 Zielin´ski R (1970) Monte Carlo Methods. WNT, Warszawa

Chapter 8

Dosimetry of Inhaled Nanoparticles Wolfgang G. Kreyling and Marianne Geiser

8.1

Introduction

The ample and consistently found evidence for the association between adverse health effects and increased concentrations of ambient fine and ultrafine particles (e.g. Ibald-Mulli et al. 2002; Laden et al. 2000, 2006; Pope 2004; Schulz et al. 2005), as well as the exponentially growing production of engineered nanoparticles (NP) urgently require risk assessment of the potential for adverse health effects by such NP. The risk assessment paradigm comprises of exposure assessment, hazard identification and characterization and risk characterization; this chapter will focus on hazard identification and characterization. Thereby, data for inhaled NP are particularly pressing because this is the major route for unwanted exposure. On the other hand, medicinal use of NP as diagnostic tools or as therapeutics may offer advanced treatment (Rytting et al. 2008; Duncan 2006). Also in this case, the full evaluation of the entire risk assessment is mandatory. Moreover, since NP have altered properties compared to microparticles (MP) of the same material (Ferin et al. 1992), health risk identification and characterization is also necessary for already tested materials. Dosimetry comprises the deposition behavior of inhaled particles and their subsequent biokinetics fate in the respiratory tract and in the entire organism. It is the first step in hazard identification and characterization. Besides exposure assessment dosimetry is a prerequisite for toxicology and epidemiology. Within dosimetry ultrastructural analyses at the individual particle level provide detailed W.G. Kreyling ð*Þ Institute of Lung Biology and Disease and Focus-Network Nanoparticles and Health, Helmholtz Center Munich – German Research Center for Environmental Health, Ingolstaedter Landstrasse 1, D-85764, Neuherberg/Munich, Germany e-mail: [email protected] M. Geiser Institute of Anatomy, University of Bern, Baltzerstrasse 2, CH-3000, Bern 9, Switzerland e-mail: [email protected]

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_8, # Springer ScienceþBusiness Media B.V. 2010

145

146

W.G. Kreyling and M. Geiser

information about which cells and intracellular components are the targets of inhaled MP and NP. In this article we reflect the results of our basic research to investigate the deposition, retention and clearance of NP, as well as their translocation to secondary organs, in order to contribute to further hazard identification and characterization by inhaled NP. We give a macroscopic as well as a microscopic view. In addition, we present the methodology for such studies in greater detail as has been done before, so that the protocols allow their application by the interested reader.

8.2

Methods

8.2.1

Macroscopic Methods of Particle Dosimetry

8.2.1.1

Quantitative Nanoparticle Biokinetics

Quantitative biokinetics can be performed after NP administration by any route; the most prominent routes are inhalation or intratracheal instillation delivering NP to the respiratory tract (RT), intraoesophageal instillation (gavage) delivering NP to the gastro-intestinal-tract (GIT), intravenous or intra-arterial injection delivering NP directly to blood circulation and finally dermal applications to the skin. The concept of quantitative biodistribution is simple, since it aims to quantify the total amount of NP distributed in the entire body at a certain time point and in the total excretion until this time point. Hence, it does not only determine NP fractions in the organs and tissues of interest but the entire remaining carcass and the total excretion are also analyzed. Therefore, a 100% balance of the administered NP is achieved providing precise fractions for all samples analyzed (Kreyling et al. 2002; Semmler et al. 2004). Quantitative biokinetics is achieved by measurements of the biodistribution of NP at several time points, as shown in Fig. 8.1. It is generally accepted that quantitative biokinetics is limited to small laboratory animals and hence, it is mainly performed in rodents.

Fig. 8.1 Concept of quantitative NP biokinetics

8 Dosimetry of Inhaled Nanoparticles

147

Analysis of NP can be performed by chemical analytical technologies like ICPMS or AA-MS etc. However, care is required at the sample preparation step, namely when NP need to be dissolved and the extremely low concentration of ions of the NP may be lost at the walls of the instrumentation prior to the final analytical step. In limited cases, the radio-analytical approach provides an elegant alternative, since the native organs and tissues are directly radio-spectroscopically analyzed without any pre-treatment. However, the radio-label in use needs to be firmly integrated into the NP matrix without any leaching. While this requirement is difficult to fulfill for a radio-isotope blended into the NP during production, the use of a radio-isotope which belongs to the same chemical element as the matrix of the NP allows the stable labeling of the NP. So, the radio-label needs to be integrated during NP production, which is a real hurdle in many cases, since nanotechnology labs are usually not equipped with radio-chemistry instrumentation and the necessary permission. However, nuclear reaction within the previously generated non-radio-labeled NP upon neutron, proton or any other ion bombardment allows for the radio-activation of one to a few atoms of the NP. In fact, at most times the amount of radioactivity is sufficiently high for the subsequent radioanalysis when only one atom per NP was converted by the nuclear reaction. A well established example is gold NP which can be neutron-activated in a nuclear research reactor such that the gold NP are radio-labeled with the 198Au radio-isotope.

8.2.1.2

Animals and Organ Preparation

For our studies with radio-labeled NP, the deeply anaesthetized (isoflurane 5%) rats were sacrificed at each time point after NP administration by exsanguinations via the prepared abdominal aorta. Like this, about 70% of the total blood volume, estimated from the body weight, was collected. As described earlier (Semmler et al. 2004; Semmler-Behnke et al. 2007a), all organs and tissues, the entire remaining carcass as well as the total excretions were sampled for radio-analysis: – Organs: lungs, liver, spleen, kidneys, reproductive organs, brain, heart, gastrointestinal tract, total exsanguinated blood, total skin – Tissues: samples of muscle and of bone (femur) – Remainder: total remaining carcass beyond the listed tissues and organs – Excretions: total urine and feces, collected separately To avoid any cross contamination, the organs were removed in toto and all body fluids were immediately cannulated, when vessels or excretory ducts had to be cut. Organ and tissue samples were taken without any tissue waste and weighed in wet state. It is essential to use clean dissection techniques in order to avoid cross-contamination of organs, in particular in inhalation studies, where fur contamination is present like in whole body or nose-only exposures; these techniques may include stratified dissection of exposed versus unexposed organs and tissues, and systematic

148

W.G. Kreyling and M. Geiser

change of fresh and clean dissection tools and equipment. In addition, whole body vascular perfusion to get rid of blood is recommended as a means to determine the retention in the organ tissue itself, free of blood contamination.

8.2.2

Ultrastructural Analyses

Ultrastructural analysis of NP, i.e. at the individual NP level, requires (1) adequate organ preservation, (2) representative tissue sampling for analysis, and (3) unambiguous identification of the NP in ultrathin tissue sections. The methods for tissue preservation using chemical fixative solutions are well established and can be found in the respective text books or methodological papers. The following methods are generally used for whole lung fixation.

8.2.2.1

Lung Fixation

Airway Instillation. The fixative is introduced into the airways of a collapsed lung, in deeply anaesthetized animals or in cadaver lungs (Weibel 1984). Thereby, the air spaces evenly expand and the interalveolar septa remain unfolded. Briefly, a cannula is inserted into the trachea, close to the larynx, and securely fastened with surgical thread. Following laparatomy, the lungs are collapsed by cutting the diaphragm. Immediately thereafter, the tracheal cannula is connected to a reservoir containing buffered 2.5% glutaraldehyde solution (30 mM potassium phosphate, 350 mOsm, pH 7.4) and the flow of fixative initiated, maintaining a pressure of about 20–25 cm of H2O above the lung. The volume of fixative is limited by the capacity of the thorax, since the lungs are fixed in the closed chest. Thereafter, the trachea is ligated to avoid leakage of the instilled fixative. The chest organs are removed in toto and completely immersed in glutaraldehyde solution for at least 24 h before any further processing. Cave: Airway instillation of aqueous fixatives does not preserve the lung-lining layer and dislocates cells, i.e. macrophages residing on the inner surface of the lungs, from their native positions (Brain et al. 1984). Vascular Perfusion. When fixatives are delivered to the lungs via the blood vessels, the airways and alveoli remain in their natural air-filled state (Im Hof et al. 1989; Weibel 1984). Briefly, at first a cannula is inserted into a tracheostoma of heparinized and anaesthetized animals, which are then artificially ventilated with oxygen. Following thoracotomy, the pulmonary artery is cannulated via the rightheart ventricle, the left-heart auricle is perforated, and the lung circulation is flushed with a plasma substitute used in clinical medicine. After three inflation-deflation cycles to total capacity (TLC, at 25 cm of H2O), the lungs are maintained at about 60% TLC by reducing the inflation pressure to 5 cm of H2O (on the deflation limb). Sequential intravascular perfusion of buffered 2.5% glutaraldehyde (30 mM potassium phosphate, 510 mOsm), 1% osmium tetroxide (100 mM sodium cacodylateHCl, 350 mOsm) and 0.5% uranyl acetate (50 mM sodium hydride maleate-NaOH,

8 Dosimetry of Inhaled Nanoparticles

149

Fig. 8.2 Airway wall of mouse lung fixed by vascular triple-perfusion (a and b) and of horse trachea fixed by immersion in non-polar fixative (c and d). Note that the lung-lining layer (asterix) including the surfactant film (arrow) at the air-liquid interface has been preserved with both fixation techniques. EP ¼ epithelium, CI ¼ Cilia. Bars: (a) and (c) ¼ 2 mm, (b) and (d) ¼ 0.2 mm

100 mOsm) then follows. The appropriate oncotic pressure is obtained by addition of 3% dextran T-70 to the fixatives. This triple-perfusion system leads to good preservation of the lungs, especially of small airways and lung parenchyma, including the surface lining layer with its phagocytic cells and the surfactant film at the air–liquid interface (Fig. 8.2a and b) (Geiser et al. 1997; Gil and Weibel 1971). This method is less effective in larger airways and especially in large animals, where the distance between vasculature and airway surface is greater. Cave: Fixation solely with glutaraldehyde or paraformaldehyde, or mixtures thereof, does not adequately preserve the surface lining layer and its associated free cells. Osmium tetroxide and uranyl acetate are necessary for cross-linking and stabilizing this layer. Other Methods. Vapor fixation of lungs at a given pressure is another means to preserve whole lungs (e.g. Hulbert et al. 1982; Yoneda 1976). The method is fast and technically not difficult. However, shrinkage and organ distortion are present and the ultrastructural preservation of cells is inferior to the fixation methods mentioned above. Non-polar fixatives, i.e. 1% osmium tetroxide dissolved in inert fluorocarbon (FC, FluorinertTM Liquid, 3M, Belgium) are suitable to preserve large airways, either by immersion of excised specimens (Fig. 8.2c and d) or by airway

150

W.G. Kreyling and M. Geiser

instillation (Geiser et al. 1997; Sims et al. 1991; Thurston et al. 1976). A variety of other fixation techniques have been developed with a view to specifically preserving the inner surface of airways and alveoli, but most of these have a narrow range of application and cannot be used to fix the whole lungs or lung lobes (for review, see e.g. Geiser et al. 1997).

8.2.2.2

Recovery of Phagocytic Cells by Bronchoalveolar Lavage

There are essentially two ways of studying the uptake of inhaled particles by phagocytes in the lungs (Geiser 2002; Geiser et al. 1994, 2008). We either isolate the cells from the organ for further processing for microscopic analysis, or we fix whole lungs or parts thereof and study the cells of interest in situ, within their functional environment. Lung-surface macrophages, i.e. airway and alveolar macrophages, are readily accessible to bronchoalveolar lavage (BAL). Conversely, these cells’ location on surfaces makes them difficult to study in situ and special fixation protocols are required (see above). In small laboratory animals, lung surface cells are usually recovered by lavage of the whole organ. After animals are sacrificed by exsanguinations and following induction of a pneumothorax by cutting the diaphragm, a cannula is inserted into the trachea and connected to a syringe. The lungs are then lavaged 5–10 times with physiological saline (0.85% NaCl) or divalent-cation-free balanced salt solution; the volume of each wash cycle being five times greater than that of the tissue. The recovered BAL fluid is centrifuged; the number of pelleted cells estimated within a haemocytometer chamber and differential cells counts are performed on stained smear or cytocentrifuge slide preparations. Normally, more than 90% of BAL-cells are lung-surface macrophages. For morphological investigations, the cell pellet may be fixed in either Ito-Karnovsky’s solution (2% formaldehyde and 2% glutaraldehyde in 0.1 M sodium cacodylate, 0.5% calcium chloride, 0.01% trinitrocresol, pH 7.2–7.3) or 2.5% glutaraldehyde (in 30 mM potassium phosphate, 350 mOsm, pH 7.4) and then processed for light or electron microscopic analysis. In humans, BAL permits the recovery of cells (and solutes) from the lower respiratory tract (bronchi, bronchioli and alveoli). The standard site of sampling is generally the middle lobe or lingula. The procedure has to be in line with the official guidelines and recommendations for BAL in humans.

8.2.2.3

Representative Tissue Sampling and Quantitative Morphologic Analysis

The analysis of particles in lungs at the microscopic level requires high magnification and therefore, the analyzed tissue volume is extremely small. Furthermore, particle distribution in lungs may not be uniform random and the lung structures have specific orientations in space (e.g. the airway tree). Therefore, appropriate tissue sampling is crucial to obtain representative samples for analysis. Today, the

8 Dosimetry of Inhaled Nanoparticles

151

methods of unbiased stereology allow direct measurements of nearly all parameters (Howard and Reed 2005). To estimate the number of particles, cells, cell organelles etc. we would need a systematic sample of disectors. This can be easily applied on thick slices at the light microscopic level (the optical disector, e.g. Cruz-Orive et al. 2004; West et al. 1996). The alternative is the use of the physical disector (Sterio 1984) consisting of the reference section plus an adjacent look-up section, where a particle is counted if it is observed in the former but not in the latter one (see e.g. Cruz-Orive and Geiser 2004). We have implemented this technique to the lungs for the analysis of MP and of airway macrophages at the light microscopic level (e.g. Geiser et al. 1990, 1994). However, to properly identify NP by energy filtering transmission electron microscopy (EFTEM) we have to use ultrathin sections in the range of 40–60 nm thickness (see below) that precludes the use of thick

Fig. 8.3 Multistep sampling design for nanoparticle analysis in lungs. Stage 1 ¼ systematic sampling and epon embedding of slices from agar-embedded lungs that were exhaustively cut perpendicular to their longitudinal axis. Stage 2 ¼ systematic sub-sampling of tissue blocks from epon embedded lung slices, cutting of ultrathin sections and mounting on hexagonal cupper TEM grids. Stage 3 ¼ systematic sub-sampling of quadrats (marked in grey) from virtual field completely contained in the ultrathin section and subdivided into 35 quadrats of equal size, at a magnification of 80. Stage 4 ¼ systematic sub-sampling of hexagonal fields for particle analysis from sampled quadrats, at a magnification of 6300. For details see also text in the Section 8.2

152

W.G. Kreyling and M. Geiser

sections. In addition, the physical disector cannot be put into practice, because the sections are very thin, hence difficult to manipulate, yet very extensive at the required magnification; their matching is therefore difficult. In addition, the grids consist of hexagonal windows (Fig. 8.3), such that 40% of the section area is covered by the copper network. Consequently, we found section matching - and thereby the disector – utterly impracticable in this case. Hence, for the moment we have to resort to the model based approach to obtain number estimates for NP (Geiser et al. 2008). We have developed the following multi-step systematic sampling designs for the analysis of NP in lungs at the individual particle level (Fig. 8.3) Stage 1 – Systematic Sampling of Lung Slices. Individual lung lobes are embedded in agar and exhaustively cut (with random start) perpendicular to the longitudinal axis into slices of equal thickness using a tissue slicer (Geiser et al. 1990). Thereof a systematic sample of slices is withdrawn, e.g. every second slice (with random start), dehydrated in a graded series of ethanol and embedded in Epon. Stage 2 – Systematic Sub-sampling of Tissue Blocks and Ultrathin Sections. From the embedded lung slices, tissue blocks are systematically sub-sampled using a point counting test system. From these tissue blocks ultrathin (50 nm) sections are cut with an ultra-microtome, placed on 600-mesh hexagonal cupper grids, and stained with lead citrate and uranyl acetate for EFTEM analysis. Stage 3 – Systematic Sub-sampling of Quadrats on Ultrathin Sections. In the EFTEM, the image of the complete ultrathin section is displayed on the screen in the observation chamber at a magnification of 80. A virtual field, which is completely contained within the ultrathin section, is generated with the EFTEM software and then subdivided into a predetermined number of quadrats of equal size i.e. into 35 quadrats (seven rows, five quadrats per row). Every third quadrat is then sampled, resulting in a total of 12 sub-sampled quadrats. Stage 4 – Systematic Sub-sampling of Hexagonal Fields and NP Counting. In each sub-sampled quadrat, a main field for analysis confined by the hexagonal cupper grid is sampled using a point counting test system at a magnification of 6300. For NP sampling, the six adjacent hexagonal fields are included; hence, the total area of analysis per sub-sampled quadrat equals the total area of seven hexagonal fields. The volume of the analyzed tissue can be obtained by multiplication of the total area analyzed with the section thickness, t. Within the subsampled hexagons, particles with TiO2-NP morphology are sampled according to the unbiased counting frame rule (Gunderson 1978). Sampled NP are counted only if they are confirmed to consist of TiO2 by elemental microanalysis as described earlier (Kapp et al. 2004; Geiser et al. 2005, 2008) and below. In addition, the localization of each NP with respect to the lung tissue and cellular compartments is recorded and the equivalent diameter of the NP measured. In case of NP agglomerates, the total number of NP transects therein are recorded. The volume densities of the respective tissue and cellular compartments can be obtained with a point counting test system that is superimposed over the subsampled hexagonal fields. Stages 3 and 4 can also be applied on ultrathin sections from cell pellets.

8 Dosimetry of Inhaled Nanoparticles

153

Alternative Stages 3 and 4. Systematic random samples of lung tissue or BALmacrophage profiles for particle analysis can also be obtained by picking a random hexagon on the ultrathin section as starting point at a magnification of 80. From there on the automated goniometer allows the unbiased screening the tissue and/or cells in vertical and horizontal direction. We usually analyze at least 80 hexagons on 4–5 ultrathin sections per animal. The tissue located within these hexagons is then investigated for the presence and localization of TiO2 NP by their morphology as well as by elemental microanalyses as described below.

8.2.2.4

Morphological Characterization and Elemental Microanalysis of Nanoparticles

The identification of NP in ultrathin tissue sections requires morphological classification as well as elemental microanalysis. The morphological characterization is established on aerosol samples collected on 600-mesh formvar coated hexagonal TEM grids, as well as on ultrathin sections of aerosol samples collected onto Teflon membranes that are embedded into Epon and cut perpendicularly to the membrane (Fig. 8.4). Like this the morphology of the NP, especially also that of their section profile, is established for particle recognition in the ultrathin tissue section. However, particle morphology is – with a few exceptions – not sufficient for unambiguous identification of NP in ultrathin sections, since there are other structures morphologically resembling the NP (“false positives”). Additional elemental microanalysis of the NP is required and may be performed in a transmission electron microscope equipped with an in-column energy filter (e.g. LEO 912, Oberkochen, Germany). We have adapted three methods, which provide high resolution and sensitivity, for elemental analysis of NP in ultrathin sections (Kapp et al. 2004): (1) Spectra recorded of a small region (parallel electron energy-loss spectroscopy, parallel-EELS), (2) images taken at a defined energy loss (electron spectroscopic

Fig. 8.4 Nanoparticle morphology. Profile of ultrathin section of epon embedded TiO2 aerosol sampled on filters. Bar: 50 nm

154

W.G. Kreyling and M. Geiser

Fig. 8.5 Elemental microanalysis of a particle in lung tissue by electron spectroscopic imaging (ESI, three window method), demonstrating that the NP (arrow) consists of titanium. Three images are taken: two below the element-specific edge in order to extrapolate a background image (390eV and 440eV) and one image within the maximum (464eV) of the element specific signal. The net titanium signal is calculated by subtraction of the extrapolated background image from the titanium specific signal. The obtained image reflects the titanium distribution in white pixels. Bar: 500 nm

imaging, ESI), and (3) a mixture of the two methods called image-EELS, the simultaneous recording of multiple electron energy-loss spectra from a series of electron spectroscopic images. In our studies with TiO2 aerosols, we routinely identify elemental titanium by ESI, whereby elemental mapping is achieved with the three-windows method (Fig. 8.5). Two background windows are set below the specific edge of titanium at DE ¼ 390 eV and DE ¼ 440 eV, respectively. The titanium-specific window is set within the maximum of the titanium-signal at an energy-loss of DE ¼ 464 eV (L2,3 edge of titanium). The acceleration voltage is 120 kV and the energy selection slit width is 10 eV. Bright-field and structure-sensitive images (recorded at 250 eV) as well as element-specific contrast for titanium are obtained by digital acquisition (iTEM, Olympus Soft Imaging Solutions GmbH, Mu¨nster, Germany). ESI uses solely inelastically scattered electrons with an energy loss producing a dark field

8 Dosimetry of Inhaled Nanoparticles

155

image. Therefore, these images have reversed contrast compared to usual brightfield micrographs.

8.3 8.3.1

Nanoparticle Dosimetry Particle Deposition

The mechanisms, the pattern and the deposition efficiency of MP and NP in the respiratory tract largely depend on the aerodynamic or thermodynamic diameter of the inhaled particles (Fig. 8.6). NP deposit with high efficiency in the entire respiratory tract, due to diffusion; therefore only their thermodynamic diameter is relevant during inhalation while the aerodynamic diameter remains irrelevant because of the absence of sufficiently strong drag forces. Hence NP deposit in the extra- and intrathoracic regions of the respiratory tract: in head airways, in the air conducting trachea, bronchi, and bronchioles, in zones consisting of air conducting and gas exchange regions, i.e. the respiratory bronchioles, and in the gas exchanging alveoli, which amount to about 300  106 and comprise a surface area of roughly 140 m2 in the adult human lung. The Human Respiratory Tract Model (HRTM) of the International Commission of Radiological Protection (ICRP) provides deposition data of inhaled particles from 1 nm to 10 m for different breathing patterns of adult female or male healthy human subjects at different breathing patterns and physiological activities (ICRP 66). Data are provided for the extrathoracic region, bronchiolar region (i.e. sum of large bronchi and small bronchioles) and the alveolar region, which we have recently discussed (Kreyling et al. 2006a, b). These data represent a meta-analysis of the knowledge at the time of publication; they are widely accepted and used in many applications. Particle deposition data for children provided therein are solely based on numerical scaling factors applied to the data from adults. In fact, there is considerable lack of knowledge on the deposition in infants and children.

Fig. 8.6 The respiratory tract (a), adapted from Scha¨ffler A. and Menche N., Mensch Ko¨rper Krankheit, Urban and Fischer, Munich (1999), and particle deposition in a normal adult male mouth breathing human subject at rest as a function of particle size (b), adapted from ICRP Publication 66 (1994). Data of bronchi are the sum of the deposition in bronchi and bronchioles

156

W.G. Kreyling and M. Geiser

We conducted an inhalation study in rats of 7, 14, 21, 35 and 90 days (adult) of age and found maximal deposition of NP of 20 and 80 nm in size in the developing lungs at the age of 21 days, when the alveolar structures are largely developed but the lung size is still rather small, resulting in high diffusional deposition of NP (Semmler-Behnke et al. 2007c). Extrapolation of these data to human infants would result in highest NP deposition and hence an increased risk from inhaled NP at the age between 1 and 3 years. Furthermore, there is a body of evidence that NP deposition in patients with diseased lungs significantly deviates from those obtained in healthy subjects because of altered breathing patterns and serious changes of the fine pulmonary structures (for review see Kreyling et al. 2006a, b; Mo¨ller et al. 2008a). The size and cellular composition of airways and alveoli vary considerably; in addition, there are substantial interspecies differences. Though, airway and alveolar walls are built of the same basic structural elements (Fig. 8.7a): (1) the liquid lining layer consisting of the aqueous phase with the lubricant pericellular and mucus layers, and the surfactant film at the air–liquid interface that is continuous from the alveoli to the trachea (Geiser et al. 1997; Im Hof et al. 1997); (2) the mobile cells, i.e. mainly resident airway and alveolar macrophages submersed in the aqueous phase; (3) the highly differentiated epithelium with its basement membrane and (4)

Fig. 8.7 Components of the inner surface of the lungs (a); wetting (b) and complete displacement (c) of the deposited particle (white sphere) into the lung-lining layer by surfactant

8 Dosimetry of Inhaled Nanoparticles

157

the subepithelial connective tissue containing the blood vessels and further cells of the immune system. The inner surface of the lungs functions as a physical, biochemical and immunological barrier to separate outside from inside. It is also these structures deposited particles first interact with. Microscopic studies of rodent lungs and biophysical investigations with particles of 1–6 mm in diameter and with fibers clearly demonstrated the wetting and displacement of all particle types from the air into the aqueous phase by surfactant (Fig. 8.7b and c); regardless of their shape, surface topography and surface free energy (Geiser et al. 2003a, b). The same can be expected for NP, since this process becomes even more efficient with decreasing particle size.

8.3.2

Particle Retention and Relocation Pathways within the Lungs

In general, MP remain on the epithelial surface in airways and alveoli of rodents and are accessible to BAL (Lehnert et al. 1989; Ellender et al. 1992). The retention of inhaled particles on the epithelium of airways is presumed to be short because of their efficient clearance by the mucociliary escalator and coughing. Mucus velocity decreases with airway generations and consequently the retention time of particles in lungs increases with increasing airway generation. However, this cannot explain the prolonged retention of both NP and MP in the small human airways as has been demonstrated in a series of studies, which have been reviewed recently (Kreyling and Scheuch 2000; Mo¨ller et al. 2008b). Different mechanisms can be considered for slow particle clearance in the airways, resulting in a loss of particles from mucociliary transport, such as (i) penetration of particles through the mucus deep into the periciliary phase (Schu¨rch et al. 1990; Gehr et al. 1990), or (ii) deposition of particles in areas with reduced lung-lining layer. Smaller particles may have a higher probability of penetrating into the liquid lining layer in between cilia, where they may be readily taken up by lung surface macrophages and/or dendritic cells (Blank et al. 2007), or penetrate and/or cross the lung epithelium. The fraction of long term retained particles in the airways decreases with increasing particle size. It was just recently shown that only 25% of 100-nm carbon particles targeted to the human airways by shallow aerosol bolus inhalation were cleared within 24 h by mucociliary action, while 75% of the NP were retained for more than 48 h (Mo¨ller et al. 2008a). There are only few in vivo data on the retention of NP in lung-epithelial cells or deeper in the lungs. We found 20-nm TiO2 NP, though a small fraction, to penetrate into epithelial cells and deeper into the lung tissue within only 1 h after aerosol inhalation in rats (Geiser et al. 2005). Very few of these NP even penetrated into the blood vessels, favoring systemic distribution and translocation of such NP into secondary organs. In this electron microscopy study, we did not observe any NP accumulation in the epithelium or in the tissue compartments beyond, at 24 h after

158

W.G. Kreyling and M. Geiser

Table 8.1 Recovery of inhaled iridium NP by BAL and NP association with BAL macrophages

Time of BAL after aerosol inhalation (h) 0 6 24 72 168

NP recovered by BAL (% lung retention) 46 19 11 10 9

NP associated with BAL macrophages (% lavaged NP) 22 62 70 92 97

NP inhalation. In a 6-months study in rats, however, after a single 1-h inhalation most of the inhaled 20-nm iridium NP had disappeared from the lung surface into the epithelium and into the interstitial spaces within a few days, since the iridium NP were no longer accessible for recovery by BAL (Semmler et al. 2004, 2007a). As shown in Table 8.1, 46% of the NP were accessible to exhaustive lung lavage immediately after the 1-h inhalation and most of these NP (78%) were not associated with BAL macrophages. At 24 h and beyond, about 10% of the NP were accessible to BAL, whereby from 72 h on more than 90% of the lavaged NP were associated with BAL macrophages. To clarify whether these studies stand for differences in the materials to penetrate cells and epithelial barriers we need to confirm the localization of the iridium NP at the individual particle level, i.e. by TEM. When iridium NP were retained in the epithelium and interstitial spaces, they were accessible to lymphatic drainage and they very close to blood vessels. Yet, there was no noticeable NP accumulation in lymph nodes surrounding the trachea and there was little NP translocation into the blood, as we found in total about 5% of the deposited NP retained in all secondary target organs and in the skeleton and soft tissue (Kreyling et al. 2008 submitted). Instead, these long-term retained NP apparently re-appeared on the epithelium for subsequent clearance by lung-surface macrophages. Although not yet clear, the most likely mechanism for NP re-appearance on the epithelial surface is by macrophages. Subsequent lung-surface macrophagemediated clearance of these NP to the larynx is likely, since more than 90% of the iridium NP were found to be associated with BAL macrophages in any lavage performed beyond 72 h after aerosol inhalation. Despite of the different retention pathway of these NP, the clearance kinetic was identical to that of MP retained on the epithelium (Semmler et al. 2004).

8.3.3

Particle Clearance Towards the Larynx and to Regional Lymph Nodes

8.3.3.1

Airways

Particle clearance from airways by mucociliary transport is fast but may not remove all particles from the lung surface as summarized in Section 3.2, particularly

8 Dosimetry of Inhaled Nanoparticles

159

in humans. The transport rate depends on both the beating cilia and the aqueous lung-lining layer as reviewed by Kreyling and Scheuch (2000); it is fastest in the central airways and gets slower with increasing airway generation. From studies with MP there are conflicting results whether the aerodynamic or the geometric particle diameter is the determining parameter for prolonged retention and hence for a reduced clearance rate of MP from airways (Mo¨ller et al. 2008b). In the former case, the deposition location primarily determines the clearance time. In the latter one the geometric size of the deposited particle is likely to affect the interaction with proteins and cells of the airway epithelium, which then in turn has an impact on the clearance time, may be even by a change in the clearance mechanism. Thereby, uptake by lung-surface macrophages and/or epithelial cells is possible. Resident airway macrophages substantially contribute to the clearance of MP from the lung surface. Quantitative data for 3–6 mm particles of different materials showed an average uptake of 27.7% (SD 16.0%) of all deposited particles by airway macrophages in hamsters within less than 1 h after the beginning of the inhalation, whereby 12–15% of the macrophages contained particles (Geiser 2002). From this study there was also evidence for fast clearance of airway macrophages with high MP loads. Overall, there are no data for prolonged retention (beyond 24–48 h) of MP in rodent airways. Kinetic studies on inhaled iridium NP in rats and mice also suggest that the fraction of NP that are slowly cleared from airways may be rather small in rodents, because clearance fractions of about 30% of the NP deposit within the first 24 h after inhalation were found (rat: Kreyling et al. 2002; mouse: Alessandrini et al. 2008). The rat data correspond rather well with theoretical NP deposition estimates on rat airways (Asgharian et al. 2001a); free public software “Multipath Model of Particle Deposition, MPPD” is available for particle deposition estimates in humans and rats (http://www.thehamner.org/mppd/helpfiles/index.htm). In contrast, there is evidence from a large body of inhalation studies for prolonged retention of MP and NP in human airways (Mo¨ller et al. 2008b). Prolonged retention of NP increases with decreasing particle size; with a maximal retention of 80% of the airway deposit of NP 100 nm in diameter. This apparent difference between rodents and man in the retention time of particles in airways and hence their clearance from this lung compartment has to be utterly considered when extrapolating rodent data to man. The prolonged retention of MP was also observed in the airways of dogs (Kreyling et al. 1999). As mentioned already in the previous paragraph, the mechanisms for prolonged retention and, hence, delayed clearance of particles from airways as observed in humans and dogs remain unclear. Yet, prolonged retention will result in a drastic increase of the dose delivered to airway tissue from toxic particles and may well be one of the factors for the development of small airway cancer. In fact, the Human Respiratory Tract Model (HRTM model) of the International Commission on Radiological Protection (ICRP) (ICRP Publication 66 1994) has already included a term of prolonged airway retention.

160

8.3.3.2

W.G. Kreyling and M. Geiser

Alveoli

Lung-Surface Macrophage-Mediated Particle Transport Towards the Larynx. The uptake of NP and MP by resident alveolar macrophages and further transport to the larynx is the predominant mechanism for particle clearance from the peripheral lungs. As in the airways, the engulfment of MP by alveolar macrophages is rapid and the process is essentially completed within 24 h (Geiser et al. 1994; Lehnert and Morrow 1985; Sorokin and Brain 1975). The transport rate of MP varies between species; it is one order of magnitude lower in man, monkey, dog and guinea pigs than in other rodents and sheep (Kreyling 1990; Kreyling and Scheuch 2000). This species-specific dissimilarity may result from differences in lung anatomy, i.e. the number of generations of respiratory bronchioles: there are 1–3 generations in rat, mouse, hamster and sheep, while there are 5–10 generations in man, monkey and dog. This results in different . mean pathway lengths from the alveoli to the conducting airways, and the mucociliary escalator. The guinea pig, however has a slow particle clearance rate despite of having only 1–3 generations of respiratory bronchioles. Another explanation may be that these particles are retained in the interstitium of the human or canine lungs, whereas they remain on the epithelium of the rodent lungs. While the latter was shown for MP in hamsters (Ellender et al. 1992) and rats (Lehnert et al. 1990), MP relocation to interstitial sites and subsequent retention was estimated from biokinetics studies in dogs (Kreyling and Scheuch 2000) and later on morphologically confirmed in dogs (Kreyling et al. 2001). Surprisingly, these MP were cleared only to a rather small fraction to the hilar lymph nodes, and there was no detectable particle translocation into the blood circulation despite their close proximity to the lymphatic drainage and to blood capillaries during the entire long-term retention period. Instead, the most important clearance mechanism of insoluble MP from peripheral lungs remains lung-surface macrophage-mediated transport to the larynx. Therefore, particles retained in the interstitium need to re-appear on the epithelium. Although not yet studied, macrophage-mediated re-appearance is very likely, since BAL analyses showed that >90% of the MP are associated with lung-surface macrophages at any long-term retention time after inhalation. Unfortunately, there are no human long-term lymphatic clearance data. However, according to the striking similarity of the predominant mechanisms of human and canine particle clearance kinetics, namely the progression of lung retention and lung-surface macrophage-mediated long-term particle clearance towards the mucociliary escalator and larynx, a similar interstitial retention of MP in human lungs as that shown in dogs is expected. Still unclear is the role of bronchus associated lymphatic tissue (BALT) by which particles would find their way back to the airway epithelium through the lymphatic drainage, as it was hypothesized very early on (Macklin 1955; Adamson and Bowden 1981, Bowden and Adamson, 1984). In summary, the major differences in the clearance kinetics of MP between rodents and man may relate to morphological (length of respiratory bronchioles) as well as to functional differences (on top or below the epithelium).

8 Dosimetry of Inhaled Nanoparticles

161

Since the effective transport rates of lung-surface macrophage-mediated clearance is the same for MP and NP in rodent lungs (Semmler-Behnke et al. 2007a), one could extrapolate from both lines of evidence that NP also penetrate the human epithelium for long term retention in the interstitial spaces and that the clearance kinetics of NP in humans is not faster but as slow as that of MP. However, the underlying mechanisms are not yet fully understood and presumably more complex. Note, however, that due to the very slow particle clearance kinetics in humans, with declining particle clearance rates over increasing retention time, there is an estimated fraction of 10–20% of insoluble particles which will never be cleared out of the human lungs under physiological conditions (Kreyling and Scheuch 2000). In cases of very high particle exposure, like smoking or in some occupational settings (mining, milling, etc.), the fraction of never-cleared particles may be substantially enhanced and associated with fibrotic pathogenesis. We have recently assessed the clearance of inhaled 20-nm TiO2 NP by lungsurface macrophages at the individual NP level by EFTEM (Geiser et al. 2008). The data from this study in rats showed that lung-surface macrophages do not efficiently phagocytose these NP but take them up rather sporadically and unspecifically within the first 24 h after NP inhalation: 1. There was only 0.06–0.12% of the TiO2 NP taken up by lung-surface macrophages within 24 h after aerosol inhalation, compared to >10% of MP that were phagocytosed already within the first hour (Geiser 2002) and >80% within 24 h (Geiser et al. 1990, 2000a, b; Lehnert and Morrow 1985; Sorokin and Brain 1975). 2. As little as 0.2% and 1.7% of the BAL macrophage populations, respectively, contained NP at 1 h and at 24 h after the aerosol inhalation, which are about two orders of magnitudes less than what was shown for 3 to 6-mm particles of different materials (Geiser 2002; Geiser et al. 1994). 3. The TiO2 NP in BAL macrophages were not tightly enclosed by the vesicular membrane, as it is known from phagocytic uptake of MP. Instead, they were located in large vesicles compared to NP size and the vesicles contained other material (Fig. 8.8). This also points to a rather sporadic uptake of TiO2 NP by lung-surface macrophages, maybe during the process of phagocytic uptake of other material. Hence, there is evidence that, at least within the first 24 h after aerosol inhalation, NP bypass the most important clearance mechanisms for particles deposited in the alveoli, namely phagocytic uptake by lung-surface macrophages. Consequently, the probability of NP uptake by lung epithelial cells and/or the translocation of NP through the thin epithelial barrier increases. Particle Transport to Regional Lymph Nodes. In a previous review (Kreyling and Scheuch 2000) we pointed out species differences between rodents versus humans, dogs and monkeys, i.e. there is less MP accumulation found in lymphatic nodules adjacent to airways in rodents than in large animals and humans. This may be physiologically plausible in the case where most long-term retained MP are

162

W.G. Kreyling and M. Geiser

Fig. 8.8 TEM micrograph of BAL macrophage with TiO2 NP located in a large phagolysosome containing other phagocytosed material. Bar ¼ 500 nm

phagocytosed by macrophages residing on the (rodent) lung epithelium. But it is not conceivable for NP being long-term retained in interstitial spaces with minimal lymphatic clearance. In large species, there are only MP clearance data for dogs available. These studies demonstrated that under physiological conditions, only a small, inter-subject variable fraction of 1–5% of deposited MP was drained to the regional lymph nodes (Kreyling et al. 1986; Kreyling and Scheuch 2000), although these MP were retained in interstitial spaces right next to the lymphatic drainage system. There are no newer particle kinetic data available for larger species and particularly not for human lungs. Particle Transport from the Interstitium. There is evidence for re-appearance of MP on the epithelium of man and dog, because the most prominent clearance mechanism was shown to be lung-surface macrophage-mediated and directed towards the larynx (Kreyling et al. 2001). Note that there is negligible interstitial retention of MP in rodents (Ellender et al. 1992; Lehnert et al. 1990). While there are no human or large species data available on the long-term retention and clearance of insoluble NP from the alveolar region, in rodents NP retention in the interstitium is prominent and NP re-appear on the lung epithelium for lung-surface macrophage-mediated clearance towards the larynx (Semmler-Behnke et al. 2007a). It remains to be resolved whether NP re-appear on the alveolar epithelium at large from the interstitial connective tissue or from BALT, as discussed much earlier on (Macklin 1955; Adamson and Bowden 1981; Bowden and Adamson 1984). The re-appearance of NP on the alveolar epithelium from the interstitium is likely to be macrophage-mediated as discussed above.

8.3.4

Translocation of Nanoparticles to the Blood and Their Subsequent Accumulation in Secondary Target Organs

Since about a decade, epidemiological studies continue to indicate associations between exposure to increased concentrations of ambient fine and ultrafine particles

8 Dosimetry of Inhaled Nanoparticles

163

and adverse health effects in susceptible individuals (Ibald-Mulli et al. 2002; Peters and Pope 2002; Pope 2004; Schulz et al. 2005). Cardio-vascular effects observed in these studies triggered the discussion on enhanced translocation of ultrafine particles from the respiratory epithelium towards the blood circulation and subsequently to target organs, like the heart, liver and brain, eventually causing adverse effects on cardiac function and blood coagulation, as well as on functions of the central nervous system (Oberdo¨rster et al. 2005). From the latter review it appears that NP are subjected to transport across membrane boundaries, which has not been reported for MP. A possible mechanism provides the formation of complexes of NP with proteins of the lung-lining layer such that this complex acts like a ferry boat for NP. In contrast, MP would not be able to make use of those ferry boats because of their much larger size (Kreyling et al. 2007; Cedervall et al. 2007; Lynch and Dawson 2008). In addition, since MP are rapidly phagocytosed by lung-surface macrophages, they are only shortly available for protein-mediated transport. In fact, modifications of the surface of NP are currently intensively investigated in the discipline of Nanomedicine aiming to design, test, and optimize specific biokinetic behaviors of medicinal NP as diagnostic and therapeutic tools to reach high target organ specificity (ESF 2005); for example, drug delivery to the central nervous system via circulating NP requires surface modifications facilitating receptor-mediated translocation across the tight blood–brain barrier (e.g., apolipoprotein-E coating for LDL-receptor–mediated endocytosis in brain capillaries) (Kreuter 2001, 2004; Kreuter et al. 2002). Such highly desirable properties of NP must be carefully weighed against potential adverse cellular responses to targeted NP drug delivery; a rigorous risk assessment is mandatory. Besides the large surface area that can be modified, the large number of NP allows to disperse them into many more cells and intracellular compartments than MP of the same mass. For instance, a particle mass of 100 ng corresponds to only 2.4  104 particles (spheres of unit density) of 2 m in diameter, but to 2.4  108 particles of 20 nm in diameter or 2.4  1011 particles of 2 nm in diameter. Note that 20-nm particles comprise a major fraction of the number concentration of ambient aerosol particles (Kreyling et al. 2003), and 2 to10-nm particles are the primary particles originating from many combustion processes of which aggregated ambient ultrafine particles are made of. When we assume a NP mass of 100 ng to have accumulated in a secondary target organ like the heart or the brain, this mass would usually not be considered to be of any toxicological relevance for low toxicity particle; particularly, if one considers their low number in the entire organ, when they are 2-m in diameter. However, if they are 2-nm NP, they are to exceed the number of cells in the organ easily by a factor of ten (cell estimate is based on a 100 g organ and an average cellular volume of 6.5  1011 cm corresponding to a cell of 5 m in diameter), and the triggering of adverse effects is more likely. Of course, the induction of such effects depends on other factors as well, including NP localization within subcellular structures, NP chemistry and surface characteristics. Although there is a huge set of literature on in-vitro NP-cell interaction usually the number of NP per cell were not estimated but only the total NP mass added to the cellular system is provided. Rough

164

W.G. Kreyling and M. Geiser

estimates indicate that most of these in vitro studies used NP-to-cell ratios far beyond 1000:1. It would be most relevant indeed to provide such an estimate of these ratios to better understand the relevance of those in vitro studies to the NP doses per cell under real exposure conditions.

8.3.4.1

Experimental Translocation Studies

Human Studies. The comprehensive analysis of NP accumulation and retention in organs and tissues is practically impossible in humans; because of ethical reasons, but also because of the limiting resolution of existing detection systems. Currently, there are no reliable human experimental data reporting a translocated NP mass fraction of more than 1% of the dose delivered to the lungs (Brown et al. 2002; Wiebert et al. 2006a, b; Mills et al. 2006; Mo¨ller et al. 2008a). However, there is indirect evidence of NP translocation in humans from recent exposure studies in healthy subjects using diluted Diesel exhaust (Mills et al. 2005). Inhalation of dilute diesel exhaust was found to impair two major functions of the vasculature, i.e. the regulation of the vascular tone and fibrinolysis. In addition, in a similar study in healthy volunteers, spontaneous alterations of EEG signals in the frontal cortex were observed during and until 1 h after exposure to Diesel exhaust (Cruts et al. 2008). From these studies it is not clear, whether the observed effects were initiated by the translocated NP or by mediators released from the lungs upon their interaction with deposited NP. However, neither of the initiation pathways can be excluded. Until now, classical pathology has reported substantial MP loads in secondary target organs only for long term and massive particle exposure conditions; e.g. the accumulation of tar in the lungs of smokers leading to increasing blackening of the lungs, or the accumulation of particles or fibres in the liver and other organs of the reticuloendothelial system in coal miners and asbestos workers (Auerbach et al. 1980; LeFevre et al. 1982). Thereby, particle translocation via the lymphatic system into the blood circulation was assumed. Similarly, in overload conditions particle translocation and accumulation particularly in the reticulo-endothelial organs were observed in experimental animals (dogs: Bianco et al. 1974; rodents: reviewed by Miller 2000). Animal Studies. There is evidence for the translocation of NP such as gold, silver, TiO2, polystyrene and carbon, in the size range of 5–100 nm, across the air– blood barrier from animal experiments. These studies demonstrated NP either in the blood circulation (Berry et al. 1977; Kapp et al. 2004; Geiser et al. 2005) and in secondary target organs (Oberdo¨rster et al. 2002; Kreyling et al. 2002, Takenaka et al. 2001, 2006; Semmler et al. 2004, 2007a, b), or they revealed thrombogenic effects (Nemmar et al. 2002; Silva et al. 2005; Khandoga et al. 2004). However, it still remains unclear, whether the translocated NP fractions exceeded 5% of the delivered lung dose (see also Kreyling et al. 2004, 2006b). Recently, Chen et al. (2006) reported an estimated translocated fraction of 1–2% of 50 and 200 nm polystyrene particles.

8 Dosimetry of Inhaled Nanoparticles

165

Quantitative Assessment of NP Translocation. Quantitative NP biokinetics as described in the Methods section of this chapter allows for a rather precise determination of total and organ-specific translocated NP fractions. Such data currently exist for the following NP: – Inhaled iridium NP, 20 and 80 nm, in rats and mice (Kreyling et al. 2002; Semmler et al. 2004; Semmler-Behnke et al. 2007a) – Inhaled carbon NP, 25 nm, spiked with radio-labeled primary iridium NP in rats (Kreyling et al. 2008 submitted) – Instilled gold NP, 1.4 and 18 nm, in rats (Semmler-Behnke et al. 2007b, 2008) From the inhalation studies with iridium NP it became clear that NP accumulate not only in secondary target organs but also in soft (connective) tissue and skeletal bone including bone marrow. In these studies, 20-nm NP accumulation in all secondary target organs (liver, spleen, kidneys, heart, brain, reproductive organs) was in the range of 1–2% of the deposited dose at 24 h after administration. A similar fraction was found in the skeleton and up to 5% in the soft tissue. Hence, the total translocated NP fraction reached just about 10% (Kreyling et al. 2008 submitted). Furthermore, 20-nm iridium NP were poorly cleared from secondary target organs such that six months after a single one-hour NP inhalation exposure, the total NP fraction in the secondary target organs was still close to 1% of the initial NP deposit in the lungs, and all organs studied still contained NP (Semmler et al. 2004; Semmler-Behnke et al. 2007a). Unfortunately, there are no further data on longterm translocation of NP. There were no significant differences in NP translocation and accumulation in secondary target organs at 24 h after inhalation observed between rats and mice (Semmler-Behnke and Kreyling, personal communication 2008). Even in the fetuses of pregnant rats in their third trimester, small but detectable translocated NP fractions were registered (Semmler-Behnke et al. 2007b). NP Size Dependence. There is evidence from experimental studies that the translocation and accumulation of NP in secondary target organs depend on their size; i.e. inhaled 80-nm iridium NP were shown to translocate about one order of magnitude less than the 20-nm iridium NP, including accumulation in the skeleton and soft tissue (Kreyling et al. 2002; Kreyling et al. 2008 submitted). Additionally, significant differences in the translocation and accumulation between 1.4-nm and 18-nm gold NP have been observed, with total translocated fractions of 8% and 0.2%, respectively, at 24 h after intratracheal NP instillation (Semmler-Behnke et al. 2008). Both NP sizes were found in all secondary target organs investigated. NP Material Dependence. Consequences of different materials on NP translocation and 24-h accumulation in secondary organs can be derived from inhalation studies with 20-nm iridium and 25-nm carbon NP. There were significant, 5–10 times lower carbon than iridium NP fractions found in any of the secondary target organs (except in the liver) studied, as well as in the skeleton or soft tissue, clearly indicating a material dependence. Note that both NP types are chain agglomerates made up of either 2 to 5-nm iridium or 5 to 10-nm carbon primary particles. Caution is required when these data are compared with those of 18-nm gold NP of similar

166

W.G. Kreyling and M. Geiser

size: there is not only the material difference, but gold NP are spherical and have a smooth surface. In addition, the methods of NP administration to the lungs was different (inhalation versus instillation). Yet, the total difference of almost 10% of translocated iridium and about 2% carbon chain agglomerates compared to 0.2% of the spherical gold NP within 24 h is striking, when considering NP material dependence only. There are more quantitative data required to better understand biokinetics associated with different NP materials. Summary Remarks on NP Translocation Across the Air–Blood-Barrier. Besides the discussed importance of size and material, other NP characteristics such as the surface charge (zeta potential) and surface structures are very likely to influence the NP biokinetics. They determine the interactions of NP with proteins and cellular components and thereby the transport mechanisms responsible for NP translocation and accumulation in extra-pulmonary organs. However, it needs to be emphasized that according to the current knowledge, NP translocation and accumulation in extra-pulmonary organs is a minor clearance pathway for NP from the lungs compared to lung-surface macrophages mediated NP clearance towards the larynx. Yet, while the latter pathway leads to NP excretion via the gastro-intestinal tract, NP translocation into the blood circulation distributes NP in the body and permits access to e.g. the cardio-vascular system, the central-nervous system and the reticulo-endothelial, i.e. the immune system. Despite the potential toxicological consequences for the organism when NP interact with these organ systems, it is still unknown whether it is the translocated NP that cause the epidemiologically established adverse effects. Particularly, it remains to be shown whether chronic exposure leads to sufficiently high NP doses to trigger or mediate responses leading to initiation and/or progression of disease. In addition, the release of mediators into the blood circulation needs thorough investigations: these mediators may be triggered or modulated by the well-known oxidative stress and pro-inflammatory responses to NP. Yet, even the importance of the dose metric in the lungs or in extra-pulmonary organs is still debated: if NP mass is the effect-determining metric, it appears very unlikely to reach sufficiently high doses in extra-pulmonary organs by inhalation. However, if NP number and (reactive) surface are the cause-effect-determining metrics, then chronic NP exposure may well be a health hazard; particularly, in susceptible individuals such as infants, the elderly and individuals with pre-existing cardiovascular and lung diseases. Furthermore, the interaction of NP with the organism has to be studied at cellular and molecular levels, in lungs as well as in those secondary target organs which receive a sufficiently high NP dose. Microscopic analyses of organs from animal inhalation experiments may give more detailed information about possible pathways for (adverse) effects by inhaled NP. It will be important to know what cell and tissue types or what intracellular compartments NP interact with and what NP properties are responsible for these interactions. Unrestricted crossing of the cellular membranes by NP facilitates not only NP translocation into basically any organ but also provides access to any subcellular compartment. While unexpected NP access to secondary target organs at non negligible doses on a macroscopic scale as well as unexpected NP access to parenchymal and

8 Dosimetry of Inhaled Nanoparticles

167

immune cells and to their subcellular structures like mitochondria and nuclei may result in adverse health effects, these interactions and pathways provide unforeseeable opportunities in the design of NP for diagnostic or therapeutic medical use in the new field of Nanomedicine. Acknowledgements This work was supported in part by EU FP6 PARTICLE_RISK 012912 (NEST), U.S. National Institutes of Health grant HL070542, the German Research Foundation FOR 627, the Swiss National Science Foundation grant 3200B0-105419 and the Swiss Society for Cystic Fibrosis. We acknowledge B. Kupferschmid, M. Casaulta and C. Wigge and B. Krieger for their microscopic and technical contributions to the graphics.

References Adamson IY, Bowden DH (1981) Dose response of the pulmonary macrophagic system to various particulates and its relationship to transepithelial passage of free particles. Exp Lung Res 2:165–175 Alessandrini F, Semmler-Behnke M, Jakob T, Schulz H, Behrendt H, Kreyling W (2008) Total and regional deposition of ultrafine particles in a mouse model of allergic inflammation of the lung. Inhal Toxicol 20:585–593 Asgharian B, Hofmann W, Bergmann, R (2001a) Particle deposition in a multiple-path model of the human lung. Aerosol Sci Technol 34:332–339. Free public shareware is available at http:// www.thehamner.org/mppd/helpfiles/index.htm Auerbach O, Conston A S, Garfinkel L, Parks VR, Kaslow HD, Hammond EC (1980) Presence of asbestos bodies in organs other than the lung. Chest 77:133–137 Berry JP, Arnoux B, Stanislas G (1977) A microanalytic study of particles transport across the alveoli: role of blood platelets. Biomedicine 27:354–357 Bianco A, Gibb FR, Kilpper RW, Landman S, Morrow PE (1974) Studies of tantalum dust in the lungs. Radiology 112:549–556 Blank F, Rothen-Rutishauser B, Gehr P (2007) Dendritic cells and macrophages form a transepithelial network against foreign particulate antigens. Am J Respir Cell Mol Biol 36:669–677 Bowden DH, Adamson IY (1984) Pathways of cellular efflux and particulate clearance after carbon instillation to the lung. J Pathol 143:117–125 Brain JD, Gehr P, Kavet RI (1984) Airway macrophages. The importance of the fixation method. Am Rev Respir Dis 129:823–826 Brown JS, Zeman KL, Bennett WD (2002) Ultrafine particle deposition and clearance in the healthy and obstructed lung. Am J Respir Crit Care Med 166:1240–1247 Cedervall T, Lynch I, Lindman S, Berggard T, Thulin E, Nilsson H, Dawson KA, Linse S (2007) Understanding the nanoparticle-protein corona using methods to quantify exchange rates and affinities of proteins for nanoparticles. Proc Natl Acad Sci USA 104:2050–2055 Chen J, Tan M, Nemmar A, Song W, Dong M, Zhang G, Li Y (2006) Quantification of extrapulmonary translocation of intratracheal-instilled particles in vivo in rats: effect of lipopolysaccharide. Toxicology 222:195–201 Cruts B, van Etten L, Tornqvist H, Blomberg A, Sandstrom T, Mills NL, Borm PJA (2008) Exposure to diesel exhaust induces changes in EEG in human volunteers. Part Fibre Toxicol 5:5 Cruz-Orive LM, Geiser M (2004) Estimation of particle number by stereology: an update. J Aerosol Med 17:197–212

168

W.G. Kreyling and M. Geiser

Cruz-Orive LM, Insausti A, Insausti R, Crespo D (2004) A case study from neuroscience involving stereology and multivariate analysis. In: Evans SM, Janson AM, Nyengaard JR (eds) Quantitative methods in neuroscience, Chapter 2, Practical Neuroscience Series, Oxford University Press, Oxford, UK, pp 16–64 Duncan R (2006) Polymer conjugates as anticancer nanomedicines. Nat Rev Cancer 6:688–701 Ellender M, Hodgson A, Wood KL, Moody JC (1992) Effect of bronchopulmonary lavage on lung retention and clearance of particulate material in hamsters. Environ Health Perspect 97:209–213 ESF (2005) ESF forward look on Nanomedicine. European science foundation policy briefings, http://www.esf.org/newsrelease/83/SPB23Nanomedicine.pdf Ferin J, Oberdo¨rster G, Penney DP (1992) Pulmonary retention of ultrafine and fine particles in rats. Am J Respir Cell Mol Biol 6:535–542 Gehr P, Schu¨rch S, Berthiaume Y, Im Hof V, Geiser M (1990) Particle retention in airways by surfactant. J Aerosol Med 3:27–43 Geiser M (2002) Morphological aspects of particle uptake by lung phagocytes. Micr Res Tech 57:512–522 Geiser M, Cruz-Orive LM, Im Hof V, Gehr P (1990) Assessment of particle retention and clearance in the intrapulmonary airways of hamster lungs with the fractionator. J Microsc 160:75–88 Geiser M, Baumann M, Cruz-Orive LM, Im Hof V, Waber U, Gehr P (1994) The effect of particle inhalation on macrophage number and phagocytic activity in the intrapulmonary conducting airways of hamsters. Am J Respir Cell Mol Biol 10:594–603 Geiser M, Im Hof V, Siegenthaler W, Grunder R, Gehr P (1997) Ultrastructure of the extracellular lining layer in hamster airways: is there a two-phase system? Micr Res Techn 36:428–437 Geiser M, Gerber P, Maye I, Im Hof V, Gehr P (2000a) Retention of teflon particles in hamster lungs: a stereologic study. J Aerosol Med 13:43–55 Geiser M, Leupin N, Maye I, Im Hof V, Gehr P (2000b) Interaction of fungal spores with the lungs: distribution and retention of inhaled Calvatia excipuliformis spores. J Allergy Clin Immunol 106:92–100 Geiser M, Matter M, Maye I, Im Hof V, Gehr P, Schu¨rch S (2003a) Influence of airspace geometry and surfactant on the retention of man-made vitreous fibers (MMVF 10a). Environ Health Perspect 111:895–901 Geiser M, Schu¨rch S, Gehr P (2003b) Influence of surface chemistry and topography of particles on their immersion into the lung’s surface-lining layer. J Appl Physiol 94:1793–1801 Geiser M, Rothen-Rutishauser B, Kapp N, Schu¨rch S, Kreyling W, Schulz H, Semmler M, Im Hof V, Heyder J, Gehr P (2005) Ultrafine particles cross cellular membranes by non-phagocytic mechanisms in lungs and in cultured cells. Environ Health Perspect 113:1555–1560 Geiser M, Casaulta M, Kupferschmid B, Schulz H, Semmler-Behnke M, Kreyling W (2008) The role of macrophages in the clearance of inhaled ultrafine titanium dioxide particles. Am J Respir Cell Mol Biol 38:371–376 Gil J, Weibel ER (1971) Extracellular lining of bronchioles after perfusion-fixation of rat lungs for electron microscopy. Anat Rec 169:185–200 Gunderson HJG (1978) Estimators of the number of objects per area unbiased by edge effects. Microsc Acta 81:107–117 Howard CV, Reed M (2005) Unbiased stereology: three-dimensional measurement in microscopy, 2nd edn. Taylor & Francis Inc., Abingdon, UK Hulbert WC, Forster BB, Laird W, Pihl CE, Walker DC (1982) An improved method for fixation of the respiratory epithelial surface with the mucous and surfactant layers. Lab Invest 47:354–363 Ibald-Mulli A, Wichmann HE, Kreyling W, Peters A (2002) Epidemiological evidence on health effects of ultrafine particles. J Aerosol Med 15:189–201

8 Dosimetry of Inhaled Nanoparticles

169

ICRP Publication 66 (1994) Human respiratory tract model for radiological protection. A report of a Task Group of the International Commission on Radiological Protection. Ann ICRP 24 (1–3):1–482 Im Hof V, Scheuch G, Geiser M, Gebhard J, Gehr P, Heyder J (1989) Techniques for the determination of particle deposition in lungs of hamsters. J Aerosol Med 2:247–259 Im Hof V, Gehr P, Gerber V, Lee MM, Schu¨rch S (1997) In vivo determination of surface tension in the horse and in vitro model studies. Respir Physiol 109:81–93 Kapp N, Kreyling W, Schulz H, Im Hof V, Semmler M, Gehr P, Geiser M (2004) Identification of inhaled ultrafine titanium oxide particles by analytical electron microscopy in rat lungs. Microsc Res Techn 63:298–305 Khandoga A, Stampfl A, Takenaka S, Schulz H, Radykewicz R, Kreyling W, Krombach F (2004) Ultrafine particles exert prothrombotic but not inflammatory effects on the hepatic microcirculation in healthy mice in vivo. Circulation 109:1320–1325 Kreuter J (2001) Nanoparticulate systems for brain delivery of drugs. Adv Drug Deliv Rev 47:65–81 Kreuter J (2004) Influence of the surface properties on nanoparticle-mediated transport of drugs to the brain. J Nanosci Nanotech 4:484–488 Kreuter J, Shamenkov D, Petrov V, Ramge P, Cychutek K, Koch-Brandt C, Alyautdin R (2002) Apolipoprotein-mediated transport of nanoparticle-bound drugs across the blood–brain barrier. J Drug Target 10:317–325 Kreyling WG (1990) Interspecies comparison of lung clearance of “insoluble” particles. J Aerosol Med 3:S93–S110 Kreyling W, Scheuch G (2000) Clearance of particles deposited in the lungs. In: Heyder J, Gehr P (eds) Particle lung interactions. Marcel Dekker, New York, USA, pp 323–376 Kreyling WG, Ferron GA, Haider B (1986) Metabolic fate of inhaled Co aerosols in beagle dogs. Health Phys 51:773–795 Kreyling WG, Blanchard JD, Godleski JJ, Haeussermann S, Heyder J, Hutzler P, Schulz H, Sweeney TD, Takenaka S, Ziesenis A (1999) Anatomic localization of 24- and 96-h particle retention in canine airways. J Appl Physiol 87:269–284 Kreyling WG, Takenaka S, Schumann G, Ziesenis A (2001) Particles are predominantly transported from the canine epithelium towards the interstitial spaces and not to larynx! Analogy to human lungs? Am J Respir Crit Care Med 163:A166 Kreyling WG, Semmler M, Erbe F, Mayer P, Takenaka S, Schulz H (2002) Translocation of ultrafine insoluble iridium particles from lung epithelium to extrapulmonary organs is size dependent but very low. J Toxicol Environ Health 65:1513–1530 Kreyling WG, Tuch T, Peters A, Pitz M, Heinrich J, Sto¨lzel M, Cyrys J, Heyder J, Wichmann HE (2003) Diverging long-term trends in ambient urban particle mass and number concentrations associated with emission changes caused by the German unification. Atmos Environ 37:3841–3848 Kreyling WG, Semmler M, Moeller W (2004) Dosimetry and toxicology of ultrafine particles. J Aerosol Med 17:140–152 Kreyling WG, Semmler-Behnke M, Moller W (2006a) Ultrafine particle-lung interactions: does size matter? J Aerosol Med 19:74–83 Kreyling WG, Semmler-Behnke M, Moeller W (2006b) Health implications of nanoparticles. J Nanopart Res 8:543–562 Kreyling WG, Mo¨ller W, Semmler-Behnke M, Oberdo¨rster G (2007) Particle dosimetry: Deposition and clearance from the respiratory tract and translocation towards extra-pulmonary sites. In: Donaldson K, Borm P (eds) Particle toxicology. CRC, Taylor & Francis, Boca Raton, pp 47–74 Kreyling WG, Semmler-Behnke M, Seitz J, Scymczak W, Wittmaack K, Wenk A, Mayer P, Takenaka S, Oberdo¨rster G (2009) Size and material dependency of Translocation of inhaled iridium or carbon Nanoparticles from the lungs of rats to blood. Inhal Toxicol 21(S1):55–60

170

W.G. Kreyling and M. Geiser

Laden F, Neas LM, Dockery DW, Schwartz J (2000) Association of fine particulate matter from different sources with daily mortality in six U.S. cities. Environ Health Perspect 108:941–947 Laden F, Schwartz J, Speizer FE, Dockery DW (2006) Reduction in fine particulate air pollution and mortality: extended follow-up of the Harvard Six Cities study. Am J Respir Crit Care Med 173:667–672 LeFevre ME, Green FH, Joel DD, Laqueur W (1982) Frequency of black pigment in livers and spleens of coal workers: correlation with pulmonary pathology and occupational information. Hum Pathol 13:1121–1126 Lehnert BE, Morrow PE (1985) Association of 59Iron oxide with alveolar macrophages during alveolar clearance. Exp Lung Res 9:1–16 Lehnert BE, Valdez YE, Tietjen GL (1989) Alveolar macrophage-particle relationships during lung clearance. Am J Respir Cell Mol Biol 1:145–154 Lehnert BE, Ortiz JB, London JE, Valdez YE, Cline AF, Sebring RJ, Tietjen GL (1990) Migratory behaviors of alveolar macrophages during the alveolar clearance of light to heavy burdens of particles. Exp Lung Res 16:451–479 Lynch I, Dawson KA (2008) Protein-nanoparticle interactions. Nano Today 3:40–47 Macklin CC (1955) Pulmonary sumps, dust accumulations, alveolar fluid and lymph vessels. Acta Anat 23:1–33 Miller FJ (2000) Dosimetry of particles in laboratory animals and humans in relationship to issues surrounding lung overload and human health risk assessment: a critical review. Inhal Toxicol 12:19–57 Mills NL, Tornqvist H, Robinson SD, Gonzalez M, Darnley K, MacNee W, Boon NA, Donaldson K, Blomberg A, Sandstrom T, Newby DE (2005) Diesel exhaust inhalation causes vascular dysfunction and impaired endogenous fibrinolysis. Circulation 112:3930–3936 Mills NL, Amin N, Robinson SD, Anand A, Davies J, Patel D, de la Fuente JM, Cassee FR, Boon NA, MacNee W, Millar AM, Donaldson K, Newby DE (2006) Do inhaled carbon nanoparticles translocate directly into the circulation in humans? Am J Respir Crit Care Med 173:426–431 Mo¨ller W, Felten K, Sommerer K, Scheuch G, Meyer G, Meyer P, Haussinger K, Kreyling WG (2008a) Deposition, retention, and translocation of ultrafine particles from the central airways and lung periphery. Am J Resp Crit Care Med 177:426–432 Mo¨ller W, KreylingWG, Schmid O, Semmler-Behnke M, Schulz H (2008b) Deposition, retention and clearance, and translocation of inhaled fine and nano-particles in the respiratory tract. In Gehr P, Blank F, Mu¨hlfeld C, Rothen-Rutishauser B (eds) Particle-Lung Interactions Second Edition, Series: Lung Biology in Health and Disease Volume 241 Nemmar A, Hoylaerts MF, Hoet PH, Dinsdale D, Smith T, Xu, H, Vermylen J, Nemery B (2002) Ultrafine particles affect experimental thrombosis in an in vivo hamster model. Am J Respir Crit Care Med 166:998–1004 Oberdo¨rster G, Sharp Z, Atudorei V, Elder A, Gelein R, Lunts A, Kreyling WG, Cox C (2002) Extrapulmonary translocation of ultrafine carbon particles following whole-body inhalation exposure of rats. J Toxicol Environ Health 65:1531–1543 Oberdo¨rster G, Oberdo¨rster E, Oberdo¨rster J (2005) Nanotoxicology: an emerging discipline evolving from studies of ultrafine particles. Environ Health Perspect 113:823–839 Peters A, Pope CA III (2002) Cardiopulmonary mortality and air pollution. Lancet 360:1184–1185 Pope CA III (2004) Air pollution and health – good news and bad. N Engl J Med 351:1132–1134 Rytting E, Nguyen J, Wang X, Kissel T (2008) Biodegradable polymeric nanocarriers for pulmonary drug delivery. Expert Opin Drug Deliv 5:629–639 Scha¨ffler A, Menche N (1999) Mensch Ko¨rper Krankheit. Urban and Fischer, Munich, Germany Schulz H, Harder V, Ibald-Mulli A, Khandoga A, Koenig W, Krombach F, Radykewicz R, Stampfl A, Thorand B, Peters A (2005) Cardiovascular effects of fine and ultrafine particles. J Aerosol Med 18:1–24 Schu¨rch S, Gehr P, Im Hof V, Geiser M, Green F (1990) Surfactant displaces particles toward the epithelium in airways and alveoli. Respir Physiol 80:17–32

8 Dosimetry of Inhaled Nanoparticles

171

Semmler M, Seitz J, Erbe F, Mayer P, Heyder J, Oberdo¨rster G, Kreyling WG (2004) Long-term clearance kinetics of inhaled ultrafine insoluble iridium particles from the rat lung, including transient translocation into secondary organs. Inhal Toxicol Jun 16:453–459 Semmler-Behnke M, Takenaka S, Fertsch S, Wenk A, Seitz J, Mayer P, Oberdo¨rster G, Kreyling WG (2007a) Efficient elimination of inhaled nanoparticles from the alveolar region: Evidence for interstitial uptake and subsequent reentrainment onto airways epithelia. Environ Health Perspect 115:728–733 Semmler-Behnke M, Fertsch S, Schmid O, Wenk A, Kreyling WG (2007) Uptake of 1.4 mm versus 18mm Gold particles by secondary target organs is size dependent in control and pregnants rats after intertracheal or intravenoiz application. In: Euro nanoforum: nanotechnology in industrial applications, pp 102–104, http://www.euronanoforum2007.de/download/Proceedings%20ENF2007.pdf. Semmler-Behnke M, Bolle I, Moeller W,.Schulz H, Takenaka S, Tsuda A, Kreyling WG (2007) Ultrafine particle deposition differs consistently between the developing and adult rat lung. European Aerosol Conference 2007, Salzburg, Abstract T08A013 Semmler-Behnke M, Kreyling WG, Lipka J, Fertsch S, Wenk A, Takenaka S, Schmid G, Brandau W (2008) Biodistribution of 1.4 nm and 18 nm Gold particles in rats. Small 4(12):2108–2111 Silva VM, Corson N, Elder A, Oberdo¨rster G (2005) The rat ear vein model for investigating in vivo thrombogenicity of ultrafine articles (UFP). Toxicol Sci 85:983–989 Sims DE, Westfall JA, Kiorpes AL, Horne MM (1991) Preservation of tracheal mucus by nonaqueous fixative. Biotech Histochem 66:173–180 Sorokin SP, Brain JD (1975) Pathways of clearance in mouse lungs exposed to iron oxide aerosols. Anat Rec 181:581–626 Sterio DC (1984) The unbiased estimation of number and sizes of arbitrary particles using the disector. J Microsc 134:127–136 Takenaka S, Karg E, Roth C, Schulz H, Ziesenis A, Heinzmann U, Schramel P, Heyder J (2001) Pulmonary and systemic distribution of inhaled ultrafine silver particles in rats. Environ Health Perspect 109 (Suppl 4):547–551 Takenaka S, Karg E, Kreyling WG, Lentner B, Moller W, Behnke-Semmler M, Jennen L, Walch A, Michalke B, Schramel P, Heyder J, Schulz H (2006) Distribution pattern of inhaled ultrafine gold particles in the rat lung. Inhal Toxicol 18:733–740 Thurston RJ, Hess RA, Kilburn KH, McKenzie WN (1976) Ultrastructure of lungs fixed in inflation using a new osmium-fluorocarbon technique. J Ultrastruct Res 56:39–47 Weibel ER (1984) Morphometric and stereological methods in respiratory physiology including fixation techniques. In: Otis AB (ed) Techniques in the life sciences. Elsevier Scientific, Ireland, pp 1–35 West MJ, Østergaard K, Andreassen OA, Finsen B (1996) Estimation of the number of somatostatin neurons in the striatum: an in situ hybridization study using the optical fractionator method. J Comp Neurol 370:11–22 Wiebert P, Sanchez-Crespo A, Falk R, Philipson K, Lundin A, Larsson S, Mo¨ller W, Kreyling WG, Svartengren M (2006a) No significant translocation of inhaled 35-nm carbon particles to the circulation in humans. Inhal Toxicol 18:741–747 Wiebert P, Sanchez-Crespo A, Seitz J, Falk R, Philipson K, Kreyling WG, Mo¨ller W, Sommerer K, Larsson S, Svartengren M (2006b) Negligible clearance of ultrafine particles retained in healthy and affected human lungs. Eur Respir J 28:286–290 Yoneda K (1976) Mucous blanket of rat bronchus. Am Rev Respir Dis 114:837–842

Chapter 9

Particles of Biomedical Relevance and Their Interactions: A Classical and Quantum Mechanistic Approach to a Theoretical Description Ewa Broclawik and Liudmila Uvarova

9.1

Introduction

The need for theoretical modeling in biological and medicinal chemistry stems from many reasons. Experimental results are often difficult to interpret and the assignment of the observed signals to certain species is arbitrary in many cases. Theoretical models give direct insight into the electronic, energetic and geometric properties of the chemical systems (molecules and their aggregates), and processes they undergo, These then allow relating the obtained quantities with the measured ones. Direct investigation of materials of biological importance by modern experimental techniques is usually hardly possible, at most time- and funds consuming processes. Computational methods allow investigating a variety of hypothetical structures, where the most promising ones may be selected for further study. In addition, unstable or extremely short-living species, crucial for overall processes but being hardly proven by experiments, may be still investigated by theoretical tools. However, it should be noted that theoretical studies are performed not on actual biological systems, but rather on relevant model systems. Thus the reliability of theoretical results may strongly depend on the quality of the model, that mimics the biological system. When it comes to implementations of the theoretical and computational models, one must, for practical reasons, employ certain level of uncertainty. Therefore, modeling results must always be verified by experiments. As a rule of thumb, computational methods can be done either very accurate but feasible only for small molecules, or fast but much less accurate. The latter one is often the only possibility for the computational treatment of large systems, e.g. for tissues, enzymes or heterogeneous catalysts. It should be noted, however, that in E. Broclawik (*) Institute of Catalysis and Surface Chemistry, Polish Academy of Sciences, Krakow, Poland e-mail: [email protected] L. Uvarova Department of Applied Mathematics, Moscow State University of Technology “STANKIN”, Moscow, Russia

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_9, # Springer ScienceþBusiness Media B.V. 2010

173

174

E. Broclawik and L. Uvarova

many cases the current theoretical methods are able to predict chemical and/or biological properties with experimental accuracy, even for quite large and complex systems. In conclusion, experimental and theoretical methods should be treated as complementary approaches to gain a feasible overall view of the studied problems. Proper links between the two may be provided by mathematical models of mesoscopic phenomena, based either on atomistic scale modeling or phenomenological approaches for these physical processes. Interesting perspectives of this emerging field of quantum medicines is given in a monograph edited by Carloni and Alber 2003.

9.2 9.2.1

Theoretical Approaches Basics of Quantum Chemistry

Experiments from the beginning of the twentieth century showed that atoms and subatomic particles cannot be described by classical physics, but should obey the laws of quantum mechanics (QM) (Atkins and Friedman 2005; Szabo and Ostlund 1996). Quantum chemistry (QC) in that regard, evolved as the application of the quantum mechanics for chemical systems, namely atoms, molecules or their aggregates. The basic concept of the quantum theory is a wavefunction (WF), being in general the complex function of the coordinates, spin and time. For systems without the time dependent interactions WFs may be defined in a 4N-dimensional spinconfiguration space (with accuracy to the time dependent phase shift) C ¼ Cðq1 ; q2 ; . . . qN Þ;

q ¼ x; y; z; s

(9.1)

A WF itself has no physical meaning, but the square of the WF determines linearly the probability density of finding the species under study, i. e. giving the probability of finding the system with coordinates in the range between rN and rNþdrN and with spin value sN. It is further stressed that a WF must obey the Schro¨dinger equation (SE), which for the time independent and nonrelativistic case has the form: H^ C ¼ EC

(9.2)

where H^ is the total energy operator (Hamiltonian) and E is total energy of the system. Despite the apparent simplicity, SE can only be solved exactly for simple model problems. One of the approximations commonly used in QC to make solving SE possible, comes from the fact that nuclei are much heavier than electrons, thus move much slower and can be separated from electrons. In this case, the Hamiltonian as defined in Eq. 9.2 can be simplified to the electronic Hamiltonian H^e , which depends only on the electronic coordinates, with the nuclear coordinates becoming fixed parameters. In the BO approximation, the set of electron energies for the various nuclei positions constitutes the potential energy surface (PES), hence, providing the relevant data for molecular geometry, vibrational properties and reaction pathways.

9 Particles of Biomedical Relevance and Their Interactions

175

The strict solutions of SE are known for hydrogen atoms and hydrogen-like ions, containing one nucleus and one electron. One-electron functions, being the solutions, are called spinorbitals. The simplest approximation of the WF for the N-electron systems could be assumed as the product of the N one-electron WF (spinorbitals) keeping in mind that electrons are the fermions (they have spin equal to one half). The Pauli principle for fermions says that WFs of N fermions must be antisymmetrical with respect to the odd permutation of fermions. The simplest function fulfilling this requirement is the determinant function (determinant is the ‘antisymmetrised product’) The variational method of searching for the best spinorbitals to construct such a wave function, i.e. by minimizing the energy, is known as the Hartree–Fock (HF) method (Szabo and Ostlund 1996). HF is the mean field approximation, because within this approach electrons do not interact among each other directly, instead, each electron interacts with the averaged electric field originating from the remaining (N  1) electrons and from the Coulomb field of the nuclei. The HF procedure of energy minimization leads to the set of one-electron equations: ^ i ðri Þ ¼ Ei ’i ðri Þ h’

(9.3)

where h^ is one-electron energy operator and ’i(ri) are the orbitals, i.e. spatial parts of the spinorbitals. Equation 9.3 is still too complex to be solved even for systems of very small molecules and certainly for systems containing species such as particles. For more complex molecular systems, molecular orbitals are expanded as the linear combinations of atomic orbitals (Linear Combination of Atomic Orbitals – Molecular Orbitals): ’i ðrÞ ¼ Sk Cik wk ðrÞ

(9.4)

and the whole HF procedure is then simplified so as to find an optimal set of coefficients of the expansion. Here, HF is used as a mean field approach, which means that electrons do not interact instantaneously: they experience the averaged electric field of the other electrons. As a result, HF neglects the correlation, coming from additional repulsion between the charges of equal sign and various spins. The correlation energy is defined as the difference between the exact nonrelativistic energy and the limit of the HF energy. Although being the small fraction of the total energy, the correlation energy may bring large contribution to the relative energies, which are the object of interest in chemistry. The idea to represent the total energy as the functional relation of the particle density (observable defined in physical space), instead of abstract WF, was proposed to overcome the typical drawbacks of traditional QC and so as to avoid problems with electron correlation (Parr 1994). Such an approach has appeared to be quite successful in solid state physics, enabling a qualitatively acceptable description of many properties. However, the errors in the predicted relative energies were far from sufficient in comparison to the accuracy required in chemistry. Moreover, this density based approach had no formal justification for a long time.

176

E. Broclawik and L. Uvarova

In 1964 Hohenberg and Kohn (1964), and later, in more general form, Levy (Levy 1982), showed that energy (or any other observable parameter) can be equivalently described in the terms of WF formalism and as the functional relations of the particle density (the electron density for chemical systems in the BO approximation). Hohenberg–Kohn theorems form the background of the density functional theory (DFT). The next important step in the development of DFT was the Kohn–Sham (KS) method (Kohn and Sham 1965), being now the most commonly applied variant of DFT in computational chemistry. A detailed description of the KS method may be found in many handbooks (Parr 1994; Koch and Holthausen 2001) and review articles (Baerends and Gritsenko 1997; Jones and Gunarsson 1987). In formulae, the result is a set of KS one-electron equations: KS KS KS h^ ’KS i ðri Þ ¼ ei ’i ðri Þ

where

(9.5)

rðrÞ ¼ Sk nk ’KS kðrÞ =2

Kohn–Sham equations (Eq. 9.5) have similar form as the Hartree–Fock equations, nevertheless, differences between the HF and KS approaches are essential. It must be noted that, despite the use of orbitals in Eq. 9.5, KS is the density functional method where the electron density is the direct outcome. The total energy is calculated without any assumption about the form of the exact WF of a real system (while in HF it is approximated by the one-electron approximation). The great popularity of DFT in KS formulation comes from the fact that this is the only correlated computational method with low computational cost (time of calculations). As for the accuracy, the DFT results are typically much better than HF, comparable to the very advanced time-consuming methods. DFT is usually able to give at least qualitatively correct results for strongly correlated systems, for which HF completely fails (for example in bulk metals). In the field of modeling of large systems, like in periodic solids or biological macromolecules, DFT is the only way to obtain reasonable results.

9.2.2

Molecular Mechanics and Molecular Dynamics

Facing the fact, that Schrodinger equation is solvable only for the simple model systems and finding even approximate solutions for the complex ones is difficult, the simplified theoretical method for macromolecules and extended systems has been proposed and developed parallel to QM, namely molecular mechanics (MM) (Kettering et al. 1930; Jensen 2006). Within this approach an electronic structure does not appear explicitly. Instead, the total energy of the investigated system is presented as dependent on a certain force field, being the function of nuclear positions (treated as the mass centre of the atoms). The detailed form of this scalar field is expressed by assumed classical formulas and includes a number of arbitrary

9 Particles of Biomedical Relevance and Their Interactions

177

parameters, characterizing the interatomic interactions. These parameters are fitted either to the experimental data or to the theoretical results, obtained by more accurate computational methods for model systems. The functions appearing in the described above formula for energy are called force fields (FF). FF with properly fitted parameters is supposed to predict selected chemical properties, such as the geometries of local minima on PES for defined set of investigated systems. The expectation of limited transferability of parameters is justified by certain experimental results. For example, it is known from infrared spectroscopy that vibrational modes of some group of atoms are very similar for different molecules containing given groups. For example the stretching mode of C¼C bond is very similar for different hydrocarbons. From this observation one can expect that properties of this bond can be approximately described by a single empirically fitted parameterization for a wide range of molecules containing C¼C bonds. The main advantage of MM is a very short time of calculations in comparison to the QM methods. Therefore MM may be easily applied for large systems, even those not tractable at QM level. MM formalism is not only applicable for the simple search of the local minima of energy, but it can be easily extended to the investigation of time evolution. Such an approach, called the classical molecular dynamics (MD), considers solving classical equations of motion (Newton or Lagrange) for the system of species interacting via FF. Classical MD calculations may be performed on large systems and relatively long time scales. MM calculations are often used for initial optimizations of the structures, which are to be calculated then at QM level. Very attractive are combined QM/MM methods. The limitations of classical MM/MD are, however, numerous. First of all, the transferability of MM parameters is limited, thus caution must be paid to the choice of a proper set of parameters. Due to the simplified nature of MM approach, often only qualitative agreement with experiment can be obtained. There is no simple prescription which FF should be applied for the studied system to give reasonable values. The quality of parameters strongly depends on the fitting procedure and the database of quantities used for fitting. For example FF parameters fitted for some set of experimental structures may produce accurate geometrical properties of studied systems, but give poor relative energies or vibrational frequencies. Traditional MM/ MD is not able to describe bond breaking or formation; therefore it cannot be applied for the modeling of chemical reactions, like search for the transition states. The direct combination of DFT with classical molecular dynamics, called Car– Parinello molecular dynamics (CP-MD, Car and Parrinello 1985), widely extended the scope of applications towards medicinal chemistry (Carloni et al. 2002).

9.2.3

Phenomenological Approaches

In spite of the rapid development of methods focused on investigating the system of interest on an atomistic scale, vast variety aspects of reality, especially in life and environmental sciences, remain out of reach at that level and require phenomenological

178

E. Broclawik and L. Uvarova

treatment. The rapid development of the modern applied science and the technical and technological components of the civilization lead in particular to the increase of atmosphere pollution. Therefore a lot of scientific studies were devoted to the problems of an interaction of nanoparticles, particles with magnetic properties, and drug drops with a human organism, or aerosol particles with the lung (Korn and Krause 2007; Pauluhn 2005; Frampton et al. 2006). Nanoparticles effectively spread in all parts of the lung tract at an inhalation that can lead to complex consequences. For the description of e.g. the transport of the clusters in the lung systems it is necessary to investigate the interaction of the clusters with the lung wall. This may be accomplished by the use of phenomenological models in addition to atomistic investigation by QM or MD methods. One example is the Hydrodynamics Transport Model (Kleinstreuer and Zhang 2007; Shi and Kleinstreuer 2007). The hydrodynamics transport model relates derivatives of the velocity components, ui on the density r, the pressure p, the coefficient of viscosity n, and the roughness-viscosity nR in a manner similar to the Eddy-viscosity concept in turbulent flow models. By introducing nR, the effect of the wall roughness on the laminar flow profile was included in the airflow simulations. For the correct calculations it is also necessary to add the heat transport equation which must be considered for agglomerates. One of possible methods is based on the set theory (Uvarova 2007). It is interesting to find the particle dimensions for which the macroscopic models may be used. As an example, the calculation experiments carried out for nicotine drops may be invoked. The internal energy was determined for different number of molecules constituting a drop, beginning with two molecules. The potential energy calculation was carried out using binary interactions of a molecule with any other molecule. The potential of the interaction was assumed in the form of the Lenard–Jones modification taking chemical interactions into account. For the most effective calculations, the Metropolis’s scheme was used. It was found that by adding subsequently molecules, the energy in one drop stabilizes at approximately 110 molecules. Similar calculations for a particle of other substances (water, metals and other materials) were carried out as well (Vnukova 2008). The experiments here showed that the energy stabilizes at 40–130 molecules by adding one molecule to a drop. It allowed determining the particle dimensions for which using of the continuum models then should be acceptable. The transport processes of small particles with sizes close to the molecule dimensions in alveolus, may be studied in this way, also by the molecular dynamics (MD) method. The force was determined with the help of the surface potential given in the form of a modified Lennard–Jones potential. The characteristics for membranes were taken from the lipids parameters (Zieder 2008). The calculation experiments for water clusters were carried out by Vnukova 2008 (Vnukova, 2008; Babarin and Uvarova 1997). The water clusters moved and precipitated on the surface of the alveolus. The geometry for the alveolus was simulated by the cone with two different diameters, i.e. 12 and a 10 nm, respectively (Patton 1996). The initial quantities for the co-ordinates and particle velocity came from a random numbers generator. In a given case (for water clusters) the density of the clusters at n ¼ 3, 4, . . . , 10 exceeded the density of water in its normal state. As a result,

9 Particles of Biomedical Relevance and Their Interactions

179

clusters with increased density precipitated on the surface more quickly then clusters of the appropriate diameter with the water density. The cluster structure and orientation differed too. Calculation experiment results agreed with the experimental data on the precipitating particles (Nel et al. 1996) where a significant percentage of the precipitation of the nanoparticles was shown.

9.3

Selected Theoretical Case-Studies for Systems of Biological Importance

Theoretical enzyme chemistry took a major step forward about 10 years ago when it became possible to attack mechanistic problems involving enzyme active sites containing transition metal centers (Siegbahn and Borowski 2006). This is a very important part of biochemistry containing enzymes such as photosystem II of photosynthesis, cytochrome c oxidase in the respiratory chain, ribonucleotide reductase in DNA synthesis, and methane monooxygenase for converting alkanes to alcohols. Quantum chemical methods have now been used to examine these enzymes and many others, and appropriate methods and models have eventually emerged. In the present section, examples will be given where theory has provided significant new contributions, with the emphasis on problems where theory has been shown to be particularly useful. In this context, comparison of different systems yielding trends is a major point. Another advantage with theory is that the proposed mechanisms sometimes can be shown relatively easy to be unlikely and that new suggestions need to be made. To prove that a certain mechanism is the preferred one is quite a task, both experimentally and theoretically, and will usually require a long combined effort. In this context, theory has risen in importance during the past decade to become a part that needs to be considered before a mechanism can be fully accepted.

9.3.1

Iron-Containing Enzymes

Since the advent of a robust computational techniques based on density functional theory (DFT) quantum mechanical modeling has provided extensive information on catalytic cycle of cytochromes. This included transformation of the initial, inactive form of the enzyme into the active oxyferryl form, model ligand binding and subsequent metabolism. Recently a few comprehensive reviews have been published on the subject (Sono et al. 2006, Denisov et al. 2005; Shaik et al. 2005). The cytochrome P450 (CYP) enzymes are membrane bound proteins that catalyze primary oxidations of endobiotics and xenobiotics. CYP3A4 is a major CYP450 isoform and contributes extensively to human drug metabolism due to its high level of expression in the liver and broad capacity to oxidize structurally diverse

180

E. Broclawik and L. Uvarova

substrates. It metabolizes 50% of the drugs used in human being. Hydroxylation of the C–H bond of the drug is one of the most important metabolism steps that can influence their bioavailability by transforming them either to active form or to toxic compounds. A detailed understanding of this metabolism step and prediction of metabolites is thus a major challenge being crucial to screen drugs in an early stage of a leading development. Since experimental investigation of the catalytically active species in the metabolism requires already the presence of a substrate to initiate the reaction cycle, computational methods are very important to accomplish this task. Such techniques involve docking in the active site, pharmacophore modeling, quantitative structure activity relationship (QSAR) and/or quantum chemical/molecular mechanical (QM/MM) studies. Apart from practical significance, the recognition of the mechanism making enzymatic oxidation a highly potential tool capable of activating the inert C–H bond in hydrocarbons, and is in itself of high scientific interest, and, thus, deserves strong attention. One example discussed here (Shaikh et al. 2006, 2007) is focused on the compound (S)-N-[1-(3-morpholin-4-yl phenyl)ethyl]-3-phenylacrylamide, the novel KCNQ2 potassium channel opener that was found to have significant oral activity in a cortical spreading depression model of migraine. The substrate has excellent oral bioavailability in dogs and rats; however, CYP3A4 MDI studies indicated that it forms reactive intermediate after metabolism. On the other hand, its difluoro analogue, (S)-N-[1-(4-fluoro-3-morpholin-4-yl phenyl)ethyl]3-(4-fluorophenyl)acrylamide was found to be orally bioavailable KCNQ2 opener free of CYP3A4 MDI. The existence of a pair of closely related compounds with different MDI properties provides promising material and good guidance for examining details of selected steps in their metabolism; it also confirms the position of primary oxidation of the phenyl ring. Another example is clavaminic acid synthase (CAS) (Borowski et al. 2007), a remarkable nonheme iron dioxygenase that catalyzes three separate oxidative reactions in the biosynthesis of clavulanic acid, a clinically used inhibitor of serine b-lactamases. Notably, all three oxidative reactions, i.e. hydroxylation, cyclization, and desaturation, take place at the single active site of CAS, which in the native state hosts a single high-spin ferrous ion coordinated by the 2-histidine-1-carboxylate binding motif and three water molecules. CAS belongs to a large superfamily of 2-oxoglutarate (2-OG) dependent oxygenases, a group of mononuclear non-heme iron enzymes that couples the oxidative decarboxylation of the 2-oxoglutarate cosubstrate to the 2-electron oxidation of a primary organic substrate. The results of the classical MD simulations suggest that the active site region of the CAS-Fe (IV)¼O-succinate-PCA has a well-defined structure consistent with the concept of “negative catalysis” which proposes that enzymes with highly reactive intermediates achieve product selectivity via ablation of unproductive/undesirable pathways. Interestingly, in this structure the hydroxyl group of PCA lies substantially closer to the oxoferryl group than the C40 -bound hydrogen of PCA. Thus, this structure, together with a repulsive PMF calculated for an approach of 40 (S) hydrogen of PCA toward the oxo group, suggests that the alcohol group of PCA is oxidized first by the

9 Particles of Biomedical Relevance and Their Interactions

181

reactive oxoferryl species. The DFT investigations indeed show that oxidation of an alcohol group is markedly easier than activation of the C40 –H bond. Moreover, based on the DFT results, a novel mechanism is proposed for the cyclization reaction by CAS. This new mechanistic hypothesis involves O-radical fragmentation, ylide formation, and 1,3-dipolar cycloaddition with aldehyde as dipolarophile. Importantly, this new mechanism is consistent with the isotope kinetics data and predicts formation of a relatively long-lived intermediate, whose accumulation might be experimentally verifiable.

9.3.2

Quantitative Structure–Activity Relationship

In the next example, Quantitative Structure–Activity Relationship (QSAR) of a series of novel phenanthrene-based tylophorine derivatives with anticancer activity has been studied (Liao et al. 2008) by using the Density Functional Theory (DFT), Molecular Mechanics (MM), and statistical methods. The established model shows not only significant statistical quality, but also predictive ability. It was found that the anticancer activity expressed as pIC50, which is defined as the negative value of the logarithm of necessary molar concentration of this series of compounds to cause 50% growth inhibition against the human A549 lung cancer cell line, closely relates with the energy of the Highest Occupied Molecular Orbital, the net charge of the terminal H atom of substituent R2 (QHR2), the hydrophobic coefficient of substituent R2 (log PR2 ), and the net charges of the first atom of substituent R1 (QFR1 ). The same model was further applied to predict the pIC50 for six recently reported congeneric compounds as external test set, and the predicted pIC50 values are close to the experimental ones, and thus it further confirms that this QSAR model has high predictive ability. The theoretical results can offer some useful references for understanding the action mechanism and designing new compounds with anticancer activity. Based on this QSAR equation, ten new compounds with higher anticancer activity have been theoretically designed and are expecting experimental confirmation, hopefully soon. Antagonists of the 5-HT2A receptor studied by Borowski et al. (2000), are used to treat many psychiatric disorders. The work focuses on a group of 27 antagonists possessing varying affinities toward the receptor. These are 26 title compounds and clozapine as a reference antagonist. The active conformers of the conformationally flexible ligands were proposed by using the active rigid analogue approach and performing similarity calculations. The calculations involved genetic neural network (GNN) computations deriving QSAR from similarity matrices with crossvalidated correlation coefficients exceeding 0.92. The performance of neural networks with variety of architectures was studied. As the computations were performed for cations and neutral molecules separately, the relevance of the ligand charging is discussed.

182

9.3.3

E. Broclawik and L. Uvarova

Miscellaneous Examples of Atomistic Simulation for Biosystems

The problem of designing new potent and selective muscarinic agonists, which could be utilized in treatment of Alzheimer disease, is among the top problems in medicinal chemistry nowadays. The most promising way towards potential and selective drugs is driven by understanding the receptor–ligand interactions. The work by Broclawik and Borowski (2000), focuses on electronic and conformational structure of the bicyclic analogues of arecoline and sulfoarecoline – muscarinic receptor agonists structurally related to 4,5,6,7-tetrahydroisoxazolo[4,5-c]pyridin-3-ol (THPO) and its S-methylsulfonium derivative (DHTO). Conformational freedom of six-member rings containing sulphur and nitrogen has been investigated by means of the semiempirical AM1 method. Interaction between ‘cationic heads’ of two representative compounds and a carboxyl group of Asp in the muscarinic receptor has been modeled using the DFT method. The electrostatic potential (ESP) around the studied complexes and ligands with an extra electron (simulation of complex formation) was analyzed. The position and depth of the ESP minima in a series of studied ligands correlated well with their activity as muscarinic agonists. On the basis of these results the mechanism of the ligand–binding site interaction was hypothesized. The calculations allowed also for the comparison of bicyclic analogues of arecoline with an already existing model for muscarinic pharmacophore and to rationalize the model parameters. In the next example (Watanabe et al. 2007), the conformational changes of p47phox–p22phox complexes of wild-type and three mutants, which have been detected in CGD patients, have been analyzed using molecular dynamics (MD) simulations. The phagocyte NADPH oxidase complex plays a crucial role in the host defense against microbial infection through the production of superoxides. The chronic granulomatous disease (CGD) is an inherited immune deficiency caused by the absence of certain components of the NADPH oxidase. Key to the activation of the NADPH oxidase is the cytoplasmic subunit p47phox, which includes the tandem SH3 domains (N-SH3 and C-SH3). In active phagocytes, p47phox forms a stable complex with the cytoplasmic region of the membrane subunit p22phox that forms a left-handed polyproline type-II (PPII) helix conformation. It was found that in the wild-type, two basal planes of PPII prism in cytoplasmic region of p22phox interacted with N-SH3 and C-SH3. In contrast, in the modeled mutants, the residue at the ape of PPII helix, which interacts simultaneously with both of the tandem SH3 domains in the wild-type, moved toward C-SH3. Furthermore, interaction energies of the cytoplasmic region of p22phox with C-SH3 tended to decrease in these mutants. All these findings allowed concluding that interactions between N-SH3 of p47phox and PPII helix, which is formed by cytoplasmic region of p22phox, may play a significant role in the activation of the NADPH oxidase. The last example (Snyder and Madura 2008) deals with silica dust particles known to promote pulmonary diseases. The workers with exposures to certain respirable dusts are at high risk of developing diseases such as silicosis or coal

9 Particles of Biomedical Relevance and Their Interactions

183

workers pneumoconiosis. Fine respirable-sized crystalline silicon dioxide mineral dusts (quartz or other polymorphs) are well documented etiological agents for pulmonary fibrosis by epidemiological studies of human occupational exposures and by animal model inhalation or installation studies. Moreover, there is sufficient evidence that crystalline silica exposures can increase the risk of lung cancer. The precise mechanism of silica cytotoxicity at the molecular level is not completely known. Upon deposition in the lung, the respirable dust surfaces may be conditioned by interaction with biological fluids and materials such as surfactant components of the pulmonary bronchiole-alveolar surface. Hence, adsorption processes on silica dusts play an important role in understanding the pathogenic origin of pneumoconiosis. Crystalline silica exposures cause silicosis, while silica in the form of aluminum silicates (clays, kaolin) does not. It was proposed that quartz and kaolin have a comparable membranolytic potential on a specific surface area basis and a comparable cytotoxic potential for lavaged pulmonary macrophages. To develop some insight into this phenomenon, the interaction between a phospholipid and silica particles was examined by performing ab initio DFT calculations on clusters constructed with representative parts of the silicate surface and the phospholipid head group. Fully optimized geometries of the complexes were used to determine binding energies, –OH vibrational frequency shifts, and NMR chemical shieldings. Results indicated that the interaction of an unprotonated aluminol group (Al–OH) with the phospholipid head group is stronger compared to that with a silanol group (Si–OH). The presence of the choline moiety increased the magnitude of the –OH vibrational frequency shifts, and the shifts were significantly larger in complexes with protonated aluminol groups relative to silanol complexes. Calculated 31 P NMR chemical shieldings were increased slightly by the presence of the choline unit and were also larger in aluminol complexes relative to those in silanol complexes. These results shed some light on differences in toxicity observed for silica versus aluminosilicate surfaces.

9.4

Conclusions

Quantum chemical calculations emerged nowadays as a key element in biological research and computational medicinal chemistry. They can aid the formulation of hypotheses that provide connecting links between experimentally determined structures and biological functions. The calculations can be used to understand e.g. enzyme mechanisms, hydrogen bonding, ligand binding and other fundamental processes both in normal and aberrant biological contexts. The fundamental assumption of the rational developing new therapeutics is that beneficial effects of drugs come from molecular recognition and binding of ligands to the active sites of specific targets, such as enzymes, receptors, and nucleic acids. The effect of binding can be either promotion or inhibition of signal transduction, enzymatic activity, or

184

E. Broclawik and L. Uvarova

molecular transport. The design of small molecules able to affect the biological function of the latter is one of the major aims in the future of medicinal chemistry. Obviously, the choice of the computational strategy depends on the ability of the method (i.e. the types of atoms and/or molecules, and the type of property that can be treated satisfactorily) and the size of the system to be investigated. In biochemical applications the method of choice – if we are interested in the dynamics and effects of temperature on an entire protein with, say, 10,000 atoms – is classical molecular dynamics (MD) simulation. The key problem then is to choose a relevant force field. On the other hand, if we are interested in electronic and/or spectroscopic properties or explicit reaction with bond breaking and formation in an enzymatic active site, we must introduce quantum chemical methods where electrons are treated explicitly. Among the various quantum chemical approaches available, the density functional theory (DFT) has become a key method over the past decade, with applications ranging from interstellar space, to the atmosphere, the biosphere and the solid state. In this chapter we presented an introduction to the theory and exemplified the wide range of problems that can be addressed, with some illustrative results taken from our work and other recent works in the field of drug design (quantum chemistry) and particle transport in physiological tracts.

References Atkins PW, Friedman RS (2005) Molecular Quantum Mechanics, 4th edn. Oxford University Press, Oxford Babarin SS, Uvarova LA (2007) Mathematical modeling of non-equilibrium processes in nano volume the happening as the Casimir’s force acting. Dynamical heterogeneous sys, Moscow 29 (1), 11:60–69 (in Russian). Borowski T, de Marothy S, Broclawik E, Schofield C, Siegbahn PEM (2007) Mechanism for cyclization reaction by clavaminic acid synthase. Insights from modeling studies. Biochem-US 46:3682–3691 Borowski T, Krol M, Brocławik E, Baranowski TC, Strekowski L, Mokrosz MJ (2000) Application of similarity matrices and genetic neural networks in quantitative structure–activity relationships of 2- or 4-(4-Methylpiperazino)pyrimidines: 5-HT2A receptor antagonists. J Med Chem 43:1901–1909 Baerends EJ, Gritsenko OV (1997) A quantum chemical view of density functional theory. J Phys Chem A 101:5383–5403 Broclawik E, Borowski T (2000) Characteristics of the ligand–binding site interaction for a series of arecoline-derived muscarinic agonists: a quantum chemical study. Comp Chem 24:411–420 Car R, Parrinello M (1985) Unified approach for molecular dynamics and density-functional theory. Phys Rev Lett 55:2471–2474 Carloni P, Rothlisberger U, Parrinello M (2002) The role and perspective of ab initio molecular dynamics in the study of biological systems. Acc Chem Res 35:455–464 Carloni P, Alber F (eds) (2003) Quantum medicinal chemistry: methods and principles in medicinal chemistry, vol 17. Wiley-VCH, Weinheim

9 Particles of Biomedical Relevance and Their Interactions

185

Denisov IG, Makris TM, Sligar SG, Schlichting I (2005) Structure and chemistry of cytochrome P450. Chem Rev 105:2253–2278 Frampton MW, Stewart JC, Oberdo¨rster G, Morrow PG, Chalupa D, Pietropaoli AP, Frasier LM, Speers DM, Cox C, Huang L, Utell MJ (2006) Inhalation of ultrafine particles alters blood leukocyte expression of adhesion molecules in humans. Environ Health Perspect 114(1):51–58 Hohenberg P, Kohn W (1964) Inhomogeneous electron gas. Phys Rev 136:B864 Jensen F (2006) Introduction to computational chemistry. Wiley-VCH, Weinheim Jones RO, Gunarsson O (1989) The density functional formalism, its applications and prospects. Rev Modern Phys 61:681–746 Kettering CF, Shutts LW, Andrews DH (1930) A representation of the dynamic properties of molecules by mechanical models. Phys Rev 36:531–543 Koch W, Holthausen MC (2001) A chemist guide to density functional theory. Wiley-VCH, Weinheim Kohn W, Sham LJ (1965) Self-consistent equations including exchange and correlation effects. Phys Rev 140:A1133 Korn K, Krause E (2007) Cell-based high content screening of small-molecule libraries. Curr Opin Chem Biol 11:503–511 Levy M (1982) Electron densities in search of Hamiltonians. Phys Rev A 26:1200–1208 Li Z, Kleinstreuer C, Zhang Z (2007) Particle deposition in the human tracheobronchial airways due to transient inspiratory flow patterns. J Aerosol Sci 38:625–644 Si-Yan L, Jin-Can C, Qian L, Shen Y, Kang-Cheng Z (2008) QSAR studies and molecular design of phenanthrene-based tylophorine derivatives with anticancer activity. QSAR Comb Sci 27:280–288 Nel A, Xia T, Madler L, Li N (2006) Toxic potential of materials at the nanolevel. Science 311:622–627 Parr RG, Yang W (1994) Density-functional theory of atoms and molecules. Oxford University Press, Oxford Patton JS (1996) Mechanisms of macromolecule absorption by the lungs. Adv Drug Deliv Rev 19:3–36 Pauluhn J (2005) Overview of inhalation exposure techniques: strengths and weaknesses. Exp Toxicol Pathol 57:111–128 Shi H, Kleinstreuer C, Zhang Z (2007) Modeling of inertial particle transport and deposition in human nasal cavities with wall roughness. J Aerosol Sci 38:398–419 Siegbahn Per EM, Borowski T (2006) Modeling enzymatic reactions involving transition metals. Acc Chem Res 39:729–738 Shaik S, Kumar D, de Visser SP, Altun A, Thiel W (2005) Theoretical perspective on the structure and mechanism of cytochrome P450 enzymes. Chem Rev 105:2279–2328 Shaikh AR, Broclawik E, Tsuboi H, Koyama M, Endou A, Takaba H, Kubo M, Del Carpio CA, Miyamoto A (2007) Oxidation mechanism for the metabolism of (S)-N-[1-(3-morpholin-4ylphenyl)ethyl]-3-phenylacrylamide on oxyferryl active site in CYP3A4 enzyme: DFT modeling. J Mol Model 13:851–860 Shaikh AR, Broclawik E, Ismael M, Tsuboi H, Koyama M, Endou A, Takaba H, Kubo M, Del Carpio CA, Miyamoto A (2006) Hyperconjugation with lone pair of morpholine nitrogen stabilizes transition state for phenyl hydroxylation in CYP3A4 metabolism of (S)-N-[1-(3morpholin-4-yl phenyl) ethyl]-3-phenylacrylamide. Chem Phys Lett 419:523–527 Snyder JA, Madura JD (2008) Interaction of the phospholipid head group with representative quartz and aluminosilicate structures: an ab initio study. J Phys Chem B 112:7095–7103 Sono M, Roach MP, Coulter ED, Dawson JH (1996) Heme-containing oxygenases. Chem Rev 96:2841–2888 Szabo A, Ostlund NS (1996) Modern quantum chemistry: introduction to advanced electronic structure theory, Dover Publications Uvarova l A (2007) Mathematical modeling of heat transport in aerosol systems with particles of complex geometry. In: European Aerosol Conference. Salzburg, Abstract T12A015

186

E. Broclawik and L. Uvarova

Vnukova KV (2008) Mathematical model of physical clusters with using density functional theory. In: Fundamental physical and mathematical problems and modeling for technical and technological systems. Moscow, Yanus –K 11, 51–54 (in Russian) Watanabe Y, Tsuboi H, Koyama M, Kubo M, Del Carpio CA, Broclawik E, Ichiishi E, Kohno M, Miyamoto A (2006) Molecular dynamics study on the ligand recognition by tandem SH3 domains of p47phox, regulating NADPH oxidase activity. Comput Biol Chem 30:303–312 Zieder A (2008) Thermodynamic studies and binding mechanisms of cell-penetrating peptides with lipids and glycosaminoglycans. Adv Drug Deliv Rev 60:580–597

Chapter 10

Health Effects of Nanoparticles (Inhalation) from Medical Point of View/Type of Diseases Robert Baughman and Michal Pirozynski

10.1

Introduction

Our understanding of the functioning of the human body at the molecular and nanometer scale has improved tremendously, our diagnostic and therapeutic options for the effective treatment of severe and chronic diseases have increased only slowly over the past. Diseases like cancer, interstitial lung diseases, airway disorders, diabetes, lung and cardiovascular problems, inflammatory and infectious diseases, and neurological disorders are serious challenges to be dealt with. Applied nanotechnology to medical problems – nanomedicine – can offer new concepts. Understanding the role of nanoparticles in pathogenesis of diseases, but also their use in therapy may be the future of medicine. The respiratory system, skin and intestinal tract are always in direct contact with the environment. These organs are the first entry ports for all particles. Inhalation is the most significant route of exposure for airborne particles including the smallest – the nanoparticles. The lung consists of two functional compartments – the conducting zone made up of airways (trachea, bronchi, and bronchioles) and respiratory zone – alveoli (gas exchange units). The conducting zone consists of the first 16 generations of airways (trachea, main, lobar, segmental, subsegmental bronchi, subdividing into smaller bronchi and finally bronchioles). The respiratory zone consists of all structures that participate in gas exchange beginning with respiratory bronchioles and ending with the alveoli. The human lung is made of approximately 2300 km of airways and 500 million alveoli (Stone et al. 1992). R. Baughman (*) University of Cincinnati Medical Center, 1001 HH Eden Avenue and Albert Sabin Way, P.O. Box 670565, Cincinnati, OH 45267-0001, USA e-mail: [email protected] M. Pirozynski Department of Anesthesiology and Intensive Therapy, CMKP, 241 Czerniakowska Street, 00-416, Warsaw, Poland

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_10, # Springer ScienceþBusiness Media B.V. 2010

187

188

R. Baughman and M. Pirozynski

The effect of small particles can be beneficial to humans but more often it is harmful. Air pollution is a cocktail of different components, gaseous and particulate. The particulate component of air pollutant is measured as particle mass (PM) including PM10 and PM2,5. Particles deposit in different regions of the respiratory system depending on their aerodynamic diameter (Fig. 10.1). Levels of PM in any area (urban, rural) vary temporally (a fluctuating level around a mean) and spatially (depending on level of traffic or local industrial sources). Most health effects of PM occur at the level seen in modern cities. The most common adverse events are shown in (Table 10.1). Toxicological studies have helped to understand the mechanism of the adverse effect on the respiratory and cardiovascular systems. The complexicity and variability of ultrafine particles (UFPs) present in the air have a heterogeneous effect on the respiratory system. Particles such as sea salt, ammonium nitrates and sulfates, road dust, and crustal dust exhibit low potency in causing inflammation in the lungs. The most harmful are primary combustionderived nanoparticles (PCDNs), derived predominantly from automobiles, especially diesel, known to cause pulmonary inflammation in humans and animals.

Fig. 10.1 Deposition curves. The relationship between particle size and deposition in different anatomical structures of the respiratory system Table 10.1 Human adverse health effects of increased ultrafine particle levels

Increased mortality from cardiovascular and respiratory causes Increased admission to hospitals due to respiratory and cardiovascular causes Exacerbation of asthma and COPD Increased asthma symptoms Increased asthma medication usage Decrease of lung function Increase incidence of lung cancer

10 Health Effects of Nanoparticles (Inhalation)

189

Nanoparticles are extremely small particles with a diameter less than 100 nm, and those found in the general environment are principally derived from traffic. They have enhanced ability to generate highly reactive “free radical” molecules that damage and activate lung cells to produce proinflammatory mediators. In long-term, high-dose inhalation studies, UFPs of various types have been shown to cause chronic pulmonary inflammation, increased chemokine expression, epithelial cell hyperplasia, pulmonary fibrosis, and lung tumor (Dasenbrock et al. 1996; Driscoll et al. 1996; Nikula et al. 1995; Oberdorster et al. 1994). Short-term, low-dose exposure studies on ultrafine carbon black (CB) led to more inflammation than the larger fine CB. This was associated with increased oxidative stress, and modulation of one of the coagulation systems in normal rats (Ferin and Oberdorster 1992; Peters et al. 1997a, b). Particle size and surface area of the nanomaterials are important characteristics from a toxicological viewpoint. Carbon black nanoparticles of similar mass and composition but with different specific surface areas (300 versus 37 m2/g) were compared. It is recognized that the biologically available area was a critical parameter of the effects of nanomaterials. Also particle surface chemistry, biodegradability, number, shape, and solubility were all found to be significant factors in determining harmful biological effects (King et al. 2001). Biological effects such as inflammation, genotoxicity, and histology were related to surface area and not particle mass. Similar findings have been reported regarding tumorogenic effects of inhaled particles (Driscoll et al. 1997). Nanoparticle deposition in the respiratory system is determined by diffusional displacement due to thermal motion of air molecules interacting with the inhaled and exhaled air stream. Deposition occurs in all regions of the respiratory system and depends on particle size, shape and ventilation parameters. With decreasing particle diameter below 500 nm the deposition increases in all regions of the lung. Nanofibers with a small diameter will penetrate deeper into the lung while long fibers (>20 mm) deposit mainly in the upper airways. The fate of inhaled nanoparticles depends on regional distribution in the lung. Depending on where they deposit, the nanoparticles interact with the mucous lining fluid of the airways or the surfactant layer within the alveoli. Airway mucous (5 mm in depth) is a complex aqueous layer comprised of cell debris, electrolytes, proteins, and glycoproteins. The mucous varies depending on environmental and disease states. The surfactant layer (10–20 nm in thickness) covers the alveolar surface. Both airway and alveolar surface liquids are coated with a monolayer of highly surface active lung surfactant comprised of water insoluble long chain phospholipids. After deposition the nanoparticles are submerged in the lining fluid regardless of their nature. The smaller the particles in size, the more they can be incorporated into the surfactant layer. This does not affect the surface pressure of the surfactant, thus nanoparticles do not destabilize the lung surfactant film. Once deposited on to the lining fluid the particles act differently depending on their solubility in the fluid. Particles that are either lipid soluble or soluble in intracellular or extracellular fluids undergo chemical dissolution in situ. The kinetics of the diffusion in the alveoli is much faster than in the small airways. Only a small fraction of the nanoparticles are

190

R. Baughman and M. Pirozynski

absorbed from the tracheobronchial tree. Inhaled nanoparticles that are insoluble are not able to be rapidly absorbed and undergo physical translocation depending on the lung region in which they are deposited (Oberdorster et al. 2005). Immersion of the slowly dissolving or insoluble nanomaterials in the airway lining fluid leads to close association with the epithelial cells and cells of the host defense system. This phenomenon may induce the inflammatory cascade in the lungs (Geiser et al. 2003) Inflammation plays a dominant role of all small particles’ effect on the respiratory system. Small particles exert their proinflammatory effects in the airways but also in the interstitium of the lung. PM has shown to activate nucleus factor kappa B and other proinflammatory signaling pathways in the lungs to increase the levels of proinflammatory mediators. These effects help to explain the increased exacerbation of asthma and COPD seen during increased levels of pollution(Arenz et al. 2006; Nel et al. 1998; Riedl and Nel 2008; Takano et al. 1997). Inflammation and systemic effects of ultrafine particles are the driving force of cardiovascular effects. There is evidence of systemic inflammation following

Fig. 10.2 Pathways from inflammation to pulmonary and cardiovascular injury

10 Health Effects of Nanoparticles (Inhalation)

191

increases in fine particle levels as shown by elevated levels of C-reactive protein, blood leukocytes, platelets, fibrinogen and increased plasma viscosity (Fig. 10.2).

10.2

Nanoparticle Toxicity

Nanoparticles may also cause pulmonary toxicity. As noted, particles such as silica are known to cause pulmonary fibrosis. Interstitial lung disease and emphysema have been described in animals exposed to nanoparticles (Donaldson et al. 2006; Sayes et al. 2007; Warheit et al. 2007). To date, neither has been described in humans. Most of the information regarding pulmonary toxicity from nanoparticles is based on animal testing. In vitro testing allows for screening for potential toxicity of a wide range of materials against various cell lines and primary cells from the airways and the alveoli (Sayes et al. 2007). These studies allow one to test for cell cytokine release and cytotoxicity. However, these studies must be verified by whole host response. The most common testing for pulmonary toxicity with nanoparticles is by intratracheal administration, although there are inhaled studies as well. Assessment of response is usually performed by bronchoalveolar lavage (BAL) at fixed time points. Figure 10.3 summarizes the percentage of neutrophils of the nucleated cells of rats exposed to either phosphate buffered saline (saline) or carbonyl iron (CI), two relatively bland substances. These are

Fig. 10.3 The percentage of neutrophils in the nucleated cells retrieved by BAL in rats exposed to either saline, carbonyl iron, nano zinc oxide (5 mg/kg), mined silica (5 mg/kg) or nano quartz (5 mg/kg dose). Lavages were performed at the specified time points. Adapted from studies by Sayes et al. (2007) and Warheit et al. (2007)

192

R. Baughman and M. Pirozynski

compared to other particles, including mined silica, nano sized zinc oxide and nano quartz, all at 5mg/kg dose (Sayes et al. 2007; Warheit et al. 2007). One day after instillation, all groups have increased neutrophils in the BAL. By day 7, animals exposed to saline or carbonyl iron no longer have increased percentage of neutrophils in the BAL. For those exposed to nano zinc oxide, the BAL by day 30 is now normal. For those animals exposed to silica particles, increased neutrophils were found in the BAL through day 90. Those animals exposed to mined silica have a higher percentage of neutrophils than those exposed to the manufactured nano quartz. This figure demonstrates that larger particles (mines silica) may lead to larger reactions than nano particles. However, one can still demonstrate inflammatory reaction if a large enough dose is instilled. A variable reaction has been reported with other nanoparticles (Warheit et al. 2006; Zhu et al. 2008). Pathologic studies have confirmed that the inflammatory reaction found in the BAL is associated with inflammatory changes in the lung parenchyma (Warheit et al. 2007). Foamy macrophages are also seen in the alveolar space. These suggest that alveolar macrophage activation may be occurring. This has been associated with increased production of proinflammatory cytokines such as interleukin 6 and tumor necrosis factor-alpha (Sayes et al. 2007). This inflammatory reaction can be quite intense. Chen et al. found that titanium oxide nanoparticles led to not only alveolar inflammation, but also emphysematous changes (Chen et al. 2006). The authors demonstrated that the emphysematous changes were associated with apoptosis of the epithelial cells. In humans, progressive massive silicosis can be associated with emphysematous changes (Gamble et al. 2004). To date, there have been no reports of bronchiolitis from nanoparticles. This represents an area of the lung which leads to fairly nonspecific symptoms such as cough and dyspnea. Routine chest roentgenogram and pulmonary function are of limited value for mild to moderate disease. Since nanoparticles will deposit in the bronchioles, the potential for bronchiolitis does exist. The most extensively studied disorders resulting from nanoparticles interaction with the respiratory system are of allergic diseases. The specific effects of diesel exhaust particles (DEPs) and its nanoparticles on allergic respiratory disease have been explored in a number of animal, in vitro, and human clinical studies. The most striking finding is the profound adjuvant effects of DEPs on the development and intensity of allergic inflammation. Animal studies (instilment of diesel particles in the upper airways) have demonstrated an increase in total and antigen-specific IgE levels, as well as increases in IL-4, IL-5, and GM-CSF levels in response to DEP exposure. In addition, DEPs reproducibly induce increased airway eosinophilic inflammation, goblet cell hyperplasia, and airway hyperreactivity (AHR) in murine models of asthma (Nel et al. 1998; Takano et al. 1997). All these pathophysiological findings are consistent with clinical signs of asthma and chronic bronchitis. Asthma is a disease characterized by periodic airflow limitation, airway inflammation, and airway hyperresponsiveness (AHR) (Bochner and Busse 2005).

10 Health Effects of Nanoparticles (Inhalation)

193

Evidence on the effects of particulate air pollution on asthma exacerbation and hospital admissions is increasing. A panel study on subjects with asthma found that UFP number concentration correlated closely with alterations in lung function; furthermore, variations in the concentration of fine particles (PM 2.5) and UFP correlated with use of asthma medications (Peters et al. 1997b).

10.3

Screening for Toxicity

Since nanoparticles may cause interstitial and airway disease, screening for toxicity seems reasonable. The highest risk population would be workers involved in manufacturing the particles. These workers may be exposed during routine operation as well as exposures after accidents such as spills, which could lead to high levels of exposure. A lower risk would be individuals exposed to single doses for medical or other uses. However, patients treated with nanoparticles for underlying lung disease may be at risk because of underlying damage to the lung. An example of this would be acute exacerbation in the IPF patient. Table 10.2 lists the potential methods one could use to detect toxicity. Various factors need to be considered when evaluating these tests. For example, a respiratory questionnaire has low cost and low risk. However, the questionnaires will only be positive when a patient becomes symptomatic. They are of limited value for trying to detect subclinical disease. Pulmonary function studies are perhaps the most useful screening method for detecting early airway and lung disease. Serial testing allows for detecting trends, such as a drop of FEV-1. Decreasing FEV-1 may be occurring in response to an exposure, even when the individual test results remain within the predicted normal range (Kreiss et al. 2002). For interstitial lung disease, the earliest changes have

Table 10.2 Potential tests for detecting interstitial and airway lung disease Test Interstitial Bronchiolar Respiratory questionnaire Good sensitivity Fair sensitivity Not specific Not specific Serial Good sensitivity Fair sensitivity Spirometry Fair specifity Fair specifity Serial DLCO Good sensitivity Fair sensitivity Good specifity Fair specifity HRCT Very good sensitivity Very good sensitivity Good specifity Good specifity Bronchoscopy with BAL Fair sensitivity Fair sensitivity Fair specifity Not specific Open lung biopsy Excellent sensitivity Excellent sensitivity Excellent specifity Excellent specifity Exhaled gas condensate measurement Good sensitivity Good sensitivity Fair specifity Fair specifity

Cost Very low Low Moderate High High Very high Moderate

194

R. Baughman and M. Pirozynski

been noted with the DLCO (Murphy et al. 1971). This test is more expensive and less reproducible than spirometric measurements of the FEV-1 and FVC. The DLCO is also not useful in detecting bronchiolitis. Radiologic evaluation is another method for screening for lung disease. Chest roentgenographic reading has proved useful in detecting various dust related lung diseases, especially coal workers pneumoconiosis. With specific training, the films can be interpreted with good reproducibility (Lawson et al. 2001). However, the chest roentgenogram has less sensitivity and specificity than HRCT. This technique has proved useful in detecting interstitial (American Thoracic Society 2000; Baughman et al. 1991) and bronchiolar (de Jong et al. 2006) disease. In addition, the findings of HRCT are relatively specific for individual lung diseases (Wells and Hansell 2004). The major differences between chest roentgenogram and HRCT are cost and radiation risk (Chodick et al. 2007). However, newer protocols for HRCT have limited the amount of radiation exposure (Sone et al. 2007). As noted, bronchoalveolar lavage (BAL) has been useful in detecting disease in animal models. The technique has been widely used for diagnosis of infectious and noninfectious interstitial lung diseases (Meyer 2007). However, for occupational lung diseases, BAL lacks specificity. For example, in farmer’s lung, a form of hypersensitivity pneumonitis, increased lymphocytes in the BAL are routinely found (Leatherman et al. 1984). However, increased lymphocytes in the BAL are also found in asymptomatic farmers. Pulmonary evaluation of these asymptomatic farmers more than 5 years after the initial evaluation found no evidence for interstitial lung disease despite abnormal BAL findings (Lalancette et al. 1993). Surgical lung biopsy provides the most sensitive and specific information regarding interstitial and bronchiolar lung disease. However, the technique is associated with toxicity, including death (Utz et al. 2001). Therefore, the technique is usually reserved for those in whom a specific diagnosis would have impact on treatment. Bronchiolitis can have a subtle presentation. The diagnosis of occupational bronchiolitis in workers at microwave popcorn and other flavor manufacturers required a fair amount of detective work. The studies in this area demonstrated that there was at least two fold increase in respiratory symptoms for those workers exposed to diacetyl(Kreiss et al. 2002). These symptoms include cough, shortness of breath, wheezing, and asthma symptoms. These are fairly nonspecific complaints. The authors did find that the highest level of exposure to diacetyl was associated with the lowest FEV1 by quartile. However, the FEV1 for those with the highest exposure were still over 80% of predicted, therefore fitting into the normal range. Thus screening by symptoms and pulmonary function studies alone may fail to detect early disease. On the other hand, HRCT may be useful in providing a more specific diagnosis in these patients. Figure 10.4a demonstrates the mosaic pattern suggesting bronchiolitis in a worker at a flavor manufacturer. She had developed increasing dyspnea. As the Fig. 10.4b shows, she has gone on to develop fibrosis as well. An open lung biopsy demonstrated both fibrosis and bronchiolitis. The fibrosis was as an end-stage product of the bronchiolitis.

10 Health Effects of Nanoparticles (Inhalation)

195

Fig. 10.4 HRCT inspiratory and expiratory views of worker in flavour manufacturing plant who developed increasing dyspnea and cough. Patient was a non-smoker. Her inspiratory film demonstrates areas of ground glass density (circled) surrounded by areas of decreased density (mosaic pattern). This was accentuated on the expiratory view (circled)

10.4

Nanoparticles for Therapy

The use of nanoparticles to treat interstitial and bronchial lung disease seems quite reasonable (Bergamaschi et al. 2006). One of the advantages of nanoparticles is their area of deposition. There are several factors that affect the area of deposition. One is the size of the particles; another is the charge of the particles, and finally the properties of particles. For example, hydroscopic particles, which absorb water rapidly, will deposit a larger part of the airways. The size of the nanoparticle should deposit for the most part in the bronchioles and in the alveoli with minimal deposition in the nose and larynx. This, therefore, would make them an ideal treatment for interstitial and bronchial lung disease. There are several diseases of the small airways and interstitium which could be potential targets for therapy with nanoparticles. At the same time, the nanoparticles could cause disease in these same areas. We will discuss diseases of the small airways and interstitium and point out potential toxicity and therapy for these areas. Bronchiolitis is a reflection of the small airways of the lung. This was often referred to as the silent part of the lung. It can lead to both obstruction and restriction of disease. Unfortunately, bronchiolitis can be very difficult to diagnose and treat. There are several conditions, which have led to bronchiolitis that have been relatively well studied: chronic lung rejection to transplant and the bronchiolitis after exposure to diacetyl by workers in flavor manufacturing plants (de Jong et al. 2006; Kreiss et al. 2002) (Fig. 10.4). Lung transplant patients who develop chronic rejection of the new lung have bronchiolitis obliterans. Advanced bronchiolitis obliterans is a significant cause of death in the lung transplant population and several methods have been proposed to detect early disease. Monitoring with serial pulmonary function tests has been

196

R. Baughman and M. Pirozynski

useful in indicating possible disease. A drop in the forced expiratory volume in one second (FEV-1) of less than 80% of post transplant values usually leads to more evaluation (de Jong et al. 2006). The use of high resolution CT scan (HRCT) has been useful in detecting bronchiolitis obliterans. The usual findings for bronchiolitis obliterans are areas of ground glass mixed with areas of increased attenuation (mosaic pattern) (de Jong et al. 2006). The localized air trapping is often more obvious on an expiratory HRCT (Fig. 10.4b). A lung transplant patient with dropping FEV-1 or increasing dyspnea will often undergo bronchoscopy with transbronchial biopsies and bronchoalveolar lavage. Unfortunately, the bronchoscopy techniques have not been reliable for detecting bronchiolitis. However, they are quite useful in detecting opportunistic infections. A recent report has suggested analysis of exhaled gases for markers of increased inflammation may be useful for detection of early bronchiolitis (Van et al. 2007). Since this technique is noninvasive, it may prove a useful screening method in the appropriate population. Interstitial lung disease can be caused by several processes. One is an acute inflammation, which is a pneumonia-like pattern with increased neutrophils. This is usually self-limited. However, it can occur on top of other interstitial diseases and cause much shortness of breath. An example is an acute decompensation of idiopathic pulmonary fibrosis (Collard et al. 2007). Patients with acute hypersensitivity pneumonitis will have cough, wheezing, and squeaks. Symptoms can be more chronic, with just unexplained cough and gradually worsening of dyspnea. For both acute and chronic pulmonary function studies demonstrate both obstruction and restriction, with a reduced DLCO being an early feature. HRCT findings include nodularity and areas of ground glass. Hypersensitivity pneumonitis tends to be an upper lobe disease. Fibrosis is more prominent in chronic disease (Adler et al. 1992). One can also have chronic inflammation such as granulomatous lung disease, including sarcoidosis and hypersensitivity pneumonitis (Baughman et al. 2003). Fire fighters at “ground zero” after the 9/11 disaster in United States were noted to have a marked increase in the rate of a sarcoidosis-like disease (Izbicki et al. 2007). Inhalation of particle dust appears to be related to granulomatous disease. Animal studies have found that carbon nanoparticles can cause granulomatous reactions (Donaldson et al. 2006). Sarcoidosis is a granulomatous disease of unknown etiology affecting multiple organs. Lung involvement is identified in more than 90% of patients at time of diagnosis (Baughman et al. 2001). While the HRCT of sarcoidosis patients can be similar to hypersensitivity pneumonitis (Fig. 10.5). While many patients with sarcoidosis will have remission within 2 years of presentation, a quarter of patients require long term treatment (Baughman et al. 2006b). Novel treatments for chronic sarcoidosis include cytotoxic agents such as methotrexate and monoclonal antibodies which block tumor necrosis factor (Baughman et al. 2006a). Nanoparticles represent a potential approach to deliver immune modifiers to the lung of patients with chronic refractory sarcoidosis. Interstitial lung disease can be predominantly fibrotic. Among the chronic fibrotic lung diseases are silicosis and asbestosis. Inhalation of silica and other dusts are well

10 Health Effects of Nanoparticles (Inhalation)

197

known to cause lung disease in humans (Chong et al. 2006). The profile for clinically significant silicosis is in a young individual with heavy exposure (Castranova and Vallyathan 2000). Changes in lung function include both restrictive and obstructive disease (Gamble et al. 2004). Cigarette smoking appears to be an important cofactor leading to more extension disease (Gamble et al. 2004). Over the past few years, a group of diseases causing progressive pulmonary fibrosis have been studied. Among the idiopathic interstitial lung diseases is idiopathic pulmonary fibrosis (IPF) (2000). This disease is associated with digital clubbing and crackles on auscultation of the chest. Pulmonary function studies tend to show just restriction, with DLCO decrease out of proportion to the loss in lung volume. High-resolution CT (HRCT) shows diffuse interstitial disease changes, with a basilar prominence for many of the conditions. The presence of subpleural honeycombing is a characteristic feature seen in patients with idiopathic pulmonary fibrosis (Fig. 10.6) (2000). IPF is a progressive disease. There is no effective treatment for all patients. Some patients appear to have a response to antiinflammatory agents such as corticosteroids and/or cytotoxic drugs (Demedts et al.

Fig. 10.5 Areas of patchy infiltrate in patient with biopsy confirmed sarcoidosis

Fig. 10.6 Patient with idiopathic pulmonary fibrosis. Subpleural honeycombing is identified. Disease is more extensive at the base of the lung (b) than in mid lung (a)

198

R. Baughman and M. Pirozynski

2005; Raghu et al. 1991). However, it is not clear that any of the currently available agents change the natural course of the disease for most patients (Walter et al. 2006). Progressive pulmonary fibrosis can lead to respiratory failure in patients with scleroderma and other collagen vascular diseases. These patients have a difference in their clinical presentation with more ground glass and traction bronchiectasis, and less subpleural honeycombing than seen with IPF (Chan et al. 1997). It is a progressive disease and if untreated, the patients go on to die, although not as rapidly as in IPF. Recent studies have demonstrated a modest benefit from the use of the cyclophosphamide, a cytotoxic agent associated with significant morbidity (Tashkin et al. 2006). Pulmonary fibrosis is an increasing problem for the last 20 years (Gribbin et al. 2006). Part of this has been increased recognition on the basis of high-resolution CT scan. However, a true rise in cases appears due to our aging population. Idiopathic pulmonary fibrosis is the disease of the elderly with the average age greater than 65. The disease is associated with significant morbidity and a mortality of greater than 50% within 5 years of diagnosis (Walter et al. 2006). Agents that have been used as anti-inflammatory drugs have limited effectiveness for IPF. Nanotechnology has the advantage of being able to provide drugs that will be delivered directly to the lung. This would minimize the toxicity of other drugs given by intravenous dose. One of the limitations of nanotechnology for pulmonary fibrosis is aerosol deposition. The traction bronchiectasis in honeycomb leads to areas in which there is limited ventilation. While drug delivery may be problematic if immediate treatment by the aerosol was planned. However, the potential for prolonged deposition in the

Fig. 10.7 Patient with idiopathic pulmonary fibrosis who was clinically stable over a 6 month period (Baseline). She developed acute decompensation and respiratory failure (Acute Decompensation). Her repeat HRCT now shows diffuse ground glass opacification. She underwent multiple studies, including bronchoscopy, but no infectious agent could be identified. She subsequently died of respiratory failure

10 Health Effects of Nanoparticles (Inhalation)

199

lung and movement into areas not well ventilated may enhance effectiveness of agents delivered by nanoparticles. Patients with pulmonary fibrosis can also develop an acute exacerbation (Collard et al. 2007). Figure 10.7 demonstrates a patient with idiopathic pulmonary fibrosis with mild subpleural honeycombing and some mild ground glass on the left. She subsequently developed respiratory failure. Evaluation for infection, including bronchoscopy, was negative. HRCT demonstrates new diffuse ground glass with air bronchograms. This patient went on to die of respiratory failure. The cause of acute exacerbation is often unknown. Any aerosol treatment would have to allow for the possibility of enhancing pulmonary toxicity. While there are several treatment options for bronchiolitis, granulomatous and other interstitial lung disease available, all of these have their limitations. Nanoparticles have the potential of offering new treatment for all of these diseases. However, nanoparticles themselves have the potential of causing toxicity in these areas of the lung. The respiratory system is an attractive route for non-invasive drug delivery with advantages for both systemic and local applications . Incorporating therapeutics with polymeric nanoparticles offers additional degrees of manipulation for delivery systems, providing sustained release and the ability to target specific cells and organs. However, nanoparticle delivery to the lungs has many challenges including formulation instability due to particle–particle interactions and poor delivery efficiency due to exhalation of low-inertia nanoparticles. Also their potential toxic effects should be recognized. Further exploration into the effect of particle physicochemical properties (e.g. nanoparticle size and material) on extending particle persistence in the lungs and their influence on particle fate is necessary to aid the design of improved systems. The development of inhalable nanoparticle-based delivery systems should draw from the extensive nanoparticle research for injectable applications, including surface modification to target-specific sites. Further coordination with environmental health research is the key to understanding the implications that particle fate and toxicology might have on drug delivery.

References Adler BD, Padley SP, Muller NL, Remy-Jardin M, Remy J (1992) Chronic hypersensitivity pneumonitis: high-resolution CT and radiographic features in 16 patients. Radiology 185 (1):91–95 American Thoracic Society. Idiopathic pulmonary fibrosis: diagnosis and treatment. International consensus statement. American Thoracic Society (ATS), and the European Respiratory Society (ERS) (2000) Am J Respir Crit Care Med 161(2 Pt 1), 646–664 Arenz A, Hellweg CE, Stojicic N, Baumstark-Khan C, Grotheer HH (2006) Gene expression modulation in A549 human lung cells in response to combustion-generated nano-sized particles. Ann N Y Acad Sci 1091:170–183 Baughman RP, Shipley RT, Loudon RG, Lower EE (1991) Crackles in interstitial lung disease. Comparison of sarcoidosis and fibrosing alveolitis. Chest 100(1):96–101

200

R. Baughman and M. Pirozynski

Baughman RP, Teirstein AS, Judson MA, Rossman MD, Yeager H Jr, Bresnitz EA, DePalo L, Hunninghake G, Iannuzzi MC, Johns CJ, McLennan G, Moller DR, Newman LS, Rabin DL, Rose C, Rybicki B, Weinberger SE, Terrin ML, Knatterud GL, Cherniak R (2001) Clinical characteristics of patients in a case control study of sarcoidosis. Am J Respir Crit Care Med 164 (10 Pt 1):1885–1889 Baughman RP, Lower EE, du Bois RM (2003) Sarcoidosis. Lancet 361(9363):1111–1118 Baughman RP, Judson MA, Teirstein A, Yeager H, Rossman M, Knatterud GL, Thompson B (2006a) Presenting characteristics as predictors of duration of treatment in sarcoidosis. QJM 99 (5):307–315 Baughman RP, Drent M, Kavuru M, Judson MA, Costabel U, du BR, Albera C, Brutsche M, Davis G, Donohue JF, Muller-Quernheim J, Schlenker-Herceg R, Flavin S, Lo KH, Oemar B, Barnathan ES (2006a) Infliximab therapy in patients with chronic sarcoidosis and pulmonary involvement. Am J Respir Crit Care Med 174(7):795–802 Bergamaschi E, Bussolati O, Magrini A, Bottini M, Migliore L, Bellucci S, Iavicoli I, Bergamaschi A (2006) Nanomaterials and lung toxicity: interactions with airways cells and relevance for occupational health risk assessment. Int J Immunopathol Pharmacol 19(4 Suppl):3–10 Bochner BS, Busse WW (2005) Allergy and asthma. J Allergy Clin Immunol 115(5):953–959 Castranova V, Vallyathan V (2000) Silicosis and coal workers’ pneumoconiosis. Environ Health Perspect 108(Suppl 4):675–684 Chan TY, Hansell DM, Rubens MB, du Bois RM, Wells AU (1997) Cryptogenic fibrosing alveolitis and the fibrosing alveolitis of systemic sclerosis: morphological differences on computed tomographic scans. Thorax 52(3):265–270 Chen HW, Su SF, Chien CT, Lin WH, Yu SL, Chou CC, Chen JJ, Yang PC (2006) Titanium dioxide nanoparticles induce emphysema-like lung injury in mice. FASEB J 20(13):2393–2395 Chodick G, Ronckers CM, Shalev V, Ron E (2007) Excess lifetime cancer mortality risk attributable to radiation exposure from computed tomography examinations in children. Isr Med Assoc J 9(8):584–587 Chong S, Lee KS, Chung MJ, Han J, Kwon OJ, Kim TS (2006) Pneumoconiosis: comparison of imaging and pathologic findings. Radiographics 26(1):59–77 Collard HR, Moore BB, Flaherty KR, Brown KK, Kaner RJ, King TE Jr, Lasky JA, Loyd JE, Noth I, Olman MA, Raghu G, Roman J, Ryu JH, Zisman DA, Hunninghake GW, Colby TV, Egan JJ, Hansell DM, Johkoh T, Kaminski N, Kim DS, Kondoh Y, Lynch DA, MullerQuernheim J, Myers JL, Nicholson AG, Selman M, Toews GB, Wells AU, Martinez FJ (2007) Acute exacerbations of idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 176(7):636–643 Dasenbrock C, Peters L, Creutzenberg O, Heinrich U (1996) The carcinogenic potency of carbon particles with and without PAH after repeated intratracheal administration in the rat. Toxicol Lett 88(1–3):15–21 de Jong PA, Dodd JD, Coxson HO, Storness-Bliss C, Pare PD, Mayo JR, Levy RD (2006) Bronchiolitis obliterans following lung transplantation: early detection using computed tomographic scanning. Thorax 61(9):799–804 Demedts M, Behr J, Buhl R, Costabel U, Dekhuijzen R, Jansen HM, Macnee W, Thomeer M, Wallaert B, Laurent F, Nicholson AG, Verbeken EK, Verschakelen J, Flower CD, Capron F, Petruzzelli S, De VP, van den Bosch JM, Rodriguez-Becerra E, Corvasce G, Lankhorst I, Sardina M, Montanari M (2005) High-dose acetylcysteine in idiopathic pulmonary fibrosis. N Engl J Med 353(21):2229–2242 Donaldson K, Aitken R, Tran L, Stone V, Duffin R, Forrest G, Alexander A (2006) Carbon nanotubes: a review of their properties in relation to pulmonary toxicology and workplace safety. Toxicol Sci 92(1):5–22 Driscoll KE, Carter JM, Howard BW, Hassenbein DG, Pepelko W, Baggs RB, Oberdorster G (1996) Pulmonary inflammatory, chemokine, and mutagenic responses in rats after subchronic inhalation of carbon black. Toxicol Appl Pharmacol 136(2):372–380

10 Health Effects of Nanoparticles (Inhalation)

201

Driscoll KE, Deyo LC, Carter JM, Howard BW, Bertram TA (1997) Effects of particle exposure and particle elicited inflammatory cells on mutation in rat epithelial cells. Carcinogenesis 18:423–430 Ferin J, Oberdorster G (1992) Polymer degradation and ultrafine particles: potential inhalation hazards for astronauts. Acta Astronaut 27:257–259 Gamble JF, Hessel PA, Nicolich M (2004) Relationship between silicosis and lung function. Scand J Work Environ Health 30(1):5–20 Geiser M, Schurch S, Gehr P (2003) Influence of surface chemistry and topography of particles on their immersion into the lung’s surface-lining layer. J Appl Physiol 94(5):1793–1801 Gribbin J, Hubbard RB, Le JI, Smith CJ, West J, Tata LJ (2006) Incidence and mortality of idiopathic pulmonary fibrosis and sarcoidosis in the UK. Thorax 61(11):980–985 Izbicki G, Chavko R, Banauch GI, Weiden MD, Berger KI, Aldrich TK, Hall C, Kelly KJ, Prezant DJ (2007) World Trade Center “sarcoid-like” granulomatous pulmonary disease in New York City Fire Department rescue workers. Chest 131(5):1414–1423 King TE Jr, Schwarz MI, Brown K, Tooze JA, Colby TV, Waldron JA Jr, Flint A, Thurlbeck W, Cherniack RM (2001) Idiopathic pulmonary fibrosis: relationship between histopathologic features and mortality. Am J Respir Crit Care Med 164(6):1025–1032 Kreiss K, Gomaa A, Kullman G, Fedan K, Simoes EJ, Enright PL (2002) Clinical bronchiolitis obliterans in workers at a microwave-popcorn plant. N Engl J Med 347(5):330–338 Lalancette M, Carrier G, Laviolette M, Ferland S, Rodrique J, Begin R, Cantin A, Cormier Y (1993) Farmer’s lung Long-term outcome and lack of predictive value of bronchoalveolar lavage fibrosing factors. Am Rev Respir Dis 148(1):216–221 Lawson CC, LeMasters MK, Kawas LG, Simpson RS, Rice CH, Lockey JE (2001) Reliability and validity of chest radiograph surveillance programs. Chest 120(1):64–68 Leatherman JW, Michael AF, Schwartz BA, Hoidal JR (1984) Lung T cells in hypersensitivity pneumonitis. Ann Intern Med 100(3):390–392 Meyer KC (2007) Bronchoalveolar lavage as a diagnostic tool. Semin Respir Crit Care Med 28 (5):546–560 Murphy RL, Jr, Ferris BG, Jr, Burgess WA, Worcester J, Gaensler EA (1971) Effects of low concentrations of asbestos. Clinical, environmental, radiologic and epidemiologic observations in shipyard pipe coverers and controls. N Engl J Med 285(23):1271–1278 Nel AE, az-Sanchez D, Ng D, Hiura T, Saxon A (1998) Enhancement of allergic inflammation by the interaction between diesel exhaust particles and the immune system. J Allergy Clin Immunol 102(4 Pt 1) 539–554 Nikula KJ, Snipes MB, Barr EB, Griffith WC, Henderson RF, Mauderly JL (1995) Comparative pulmonary toxicities and carcinogenicities of chronically inhaled diesel exhaust and carbon black in F344 rats. Fundam Appl Toxicol 25(1):80–94 Oberdorster G, Cherian MG, Baggs RB (1994) Correlation between cadmium-induced pulmonary carcinogenicity, metallothionein expression, and inflammatory processes: a species comparison. Environ Health Perspect 102(Suppl 3):257–263 Oberdorster G, Oberdorster E, Oberdorster J (2005) Nanotoxicology: an emerging discipline evolving from studies of ultrafine particles. Environ Health Perspect 113(7):823–839 Peters A, Dockery DW, Heinrich J, Wichmann HE (1997a) Short-term effects of particulate air pollution on respiratory morbidity in asthmatic children. Eur Respir J 10(4):872–879 Peters A, Wichmann HE, Tuch T, Heinrich J, Heyder J (1997b) Respiratory effects are associated with the number of ultrafine particles. Am J Respir Crit Care Med 155(4):1376–1383 Raghu G, Depaso WJ, Cain K, Hammar SP, Wetzel CE, Dreis DF, Hutchinson J, Pardee NE, Winterbauer RH (1991) Azathioprine combined with prednisone in the treatment of idiopathic pulmonary fibrosis: a prospective double-blind, randomized, placebo-controlled clinical trial. Am Rev Respir Dis 144(2):291–296 Riedl MA, Nel AE (2008) Importance of oxidative stress in the pathogenesis and treatment of asthma. Curr Opin Allergy Clin Immunol 8(1):49–56

202

R. Baughman and M. Pirozynski

Sayes CM, Reed KL, Warheit DB (2007) Assessing toxicity of fine and nanoparticles: comparing in vitro measurements to in vivo pulmonary toxicity profiles. Toxicol Sci 97 (1):163–180 Sone S, Nakayama T, Honda T, Tsushima K, Li F, Haniuda M, Takahashi Y, Suzuki T, Yamanda T, Kondo R, Hanaoka T, Takayama F, Kubo K, Fushimi H (2007) Long-term follow-up study of a population-based 1996–1998 mass screening programme for lung cancer using mobile low-dose spiral computed tomography. Lung Cancer 58(3):329–341 Stone KC, Mercer RR, Gehr P, Stockstill B, Crapo JD (1992) Allometric relationships of cell numbers and size in the mammalian lung. Am J Respir Cell Mol Biol 6(2):235–243 Takano H, Yoshikawa T, Ichinose T, Miyabara Y, Imaoka K, Sagai M (1997) Diesel exhaust particles enhance antigen-induced airway inflammation and local cytokine expression in mice. Am J Respir Crit Care Med 156(1):36–42 Tashkin DP, Elashoff R, Clements PJ, Goldin J, Roth MD, Furst DE, Arriola E, Silver R, Strange C, Bolster M, Seibold JR, Riley DJ, Hsu VM, Varga J, Schraufnagel DE, Theodore A, Simms R, Wise R, Wigley F, White B, Steen V, Read C, Mayes M, Parsley E, Mubarak K, Connolly MK, Golden J, Olman M, Fessler B, Rothfield N, Metersky M (2006) Cyclophosphamide versus placebo in scleroderma lung disease. N Engl J Med 354(25):2655–2666 Utz JP, Ryu JH, Douglas WW, Hartman TE, Tazelaar HD, Myers JL, Allen MS, Schroeder DR (2001) High short-term mortality following lung biopsy for usual interstitial pneumonia. Eur Respir J 17(2):175–179 Van MA, Knoop C, Estenne M (2007) Early detection of chronic pulmonary allograft dysfunction by exhaled biomarkers. Am J Respir Crit Care Med 175(7):731–736 Walter N, Collard HR, King TE Jr (2006) Current perspectives on the treatment of idiopathic pulmonary fibrosis. Proc Am Thorac Soc 3(4):330–338 Warheit DB, Webb TR, Sayes CM, Colvin VL, Reed KL (2006) Pulmonary instillation studies with nanoscale TiO2 rods and dots in rats: toxicity is not dependent upon particle size and surface area. Toxicol Sci 91(1):227–236 Warheit DB, Webb TR, Colvin VL, Reed KL, Sayes CM (2007) Pulmonary bioassay studies with nanoscale and fine-quartz particles in rats: toxicity is not dependent upon particle size but on surface characteristics. Toxicol Sci 95(1):270–280 Wells AU, Hansell DM (2004) In: Baughman RP, du Bois RM, Lynch JP (eds) Radiologic evaluation. In diffuse lung disease. Arnold, London, pp 15–45 Zhu MT, Feng WY, Wang B, Wang TC, Gu YQ, Wang M, Wang Y, Ouyang H, Zhao YL, Chai ZF (2008) Comparative study of pulmonary responses to nano- and submicron-sized ferric oxide in rats. Toxicology 247(2–3):102–111

Chapter 11

Effects of Cigarette Smoke and Diesel Exhaust on the Innate Immune Function of the Airway Epithelium P. S. Hiemstra

11.1

Introduction

The lung is exposed to a large volume of inhaled air that contains numerous inhaled particles and gases. This way potential pathogenic micro-organisms reach the epithelial surface and need to be dealt with by host defense mechanisms to prevent severe lung infections (Bals and Hiemstra 2004; Diamond et al. 2000). The same inspired air that contains these respiratory pathogens may also contain cigarette smoke and air pollutants. These toxic compounds have been shown to cause lung inflammation and affect the ability of the lung to mount an efficient host defense response against these pathogens. In addition to impairing host defense, lung injury caused by inhalation of toxic substances may also pave the way for respiratory infections, because of the increased ability of pathogens to adhere to the injured lung mucosa. This way, smoke and air pollutants contribute to the development of airways inflammation and respiratory infections. These respiratory infections are a major cause of morbidity and mortality worldwide. In the past decades intensive research has led to a major increase in our understanding of the mechanisms that are involved in host defense against respiratory infections. An increasing number of clinical and basic science studies are addressing the way that inhaled toxic substances present in cigarette smoke and air pollutants affect this defense system and contribute to both respiratory infections and chronic lung disease. This review provides a selected overview of this area with a focus on the airway epithelium and addresses some of the remaining gaps in our knowledge.

P.S. Hiemstra Department of Pulmonology, Leiden University Medical Center, Leiden, the Netherlands e-mail: [email protected]

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_11, # Springer ScienceþBusiness Media B.V. 2010

203

204

11.2

P.S. Hiemstra

Host Defense Against Infection in the Lung

Despite the fact that we daily inhale large numbers of pathogenic micro-organisms, respiratory infections are relatively rare. Host defense mechanisms that prevent respiratory infections are operative in our lungs, and the observation that respiratory infections are rare indicates that these mechanisms are very effective. Various cell types and mechanisms are involved in protecting the lung from respiratory pathogens. These include not only phagocytic macrophages and neutrophils, but also a range of other inflammatory and immune cells. Although the focus of this paragraph and chapter is on the airway epithelium, it should be noted that cigarette smoke and diesel exhaust may also directly or indirectly affect the host defense function of these other cell types in the lung. The airway epithelium is crucial to an effective host defense system because of the large epithelial surface of the lung, and airway epithelial cells are the first cell type to interact with airborne pathogens. It provides both a passive barrier against entry of inhaled pathogens, and is increasingly recognized as an essential element of immunity in the lung (Diamond et al. 2000; Strieter et al. 2003; Bals and Hiemstra 2004) (Table 11.1). The various types of epithelial cells that constitute the pseudostratified epithelium that lines the conducting airways provide mucociliary clearance. Mucus produced by specialized airway epithelial cells and submucosal glands forms a layer on top of the airway epithelium that traps inhaled particles including micro-organisms (Bals and Hiemstra 2004; Thornton et al. 2008). This layer is positioned on top of a thin fluid layer that covers the airway epithelium, and that allows the ciliated cells of the airway epithelium to move the mucus layer that contains the trapped substances. This clearance mechanism is further supported by coughing. Airway epithelial cells also produce a range of molecules that kill micro-organisms that may have passed the mucus layer or reach the epithelial surface where little or no mucus is present (e.g. in the alveoli) (Bals and Hiemstra 2004; Rogan

Table 11.1 Mechanisms that contribute to host defense against infection by airway epithelium and submucosal glands Mechanism Function Barrier formation Prevention of penetration of micro-organisms Mucociliary clearance Trapping and removal Production of antimicrobial substances Killing and growth inhibition (such as AMPs, ROI and RNI) Cytokine and chemokine production Recruitment and activation of inflammatory and immune cells; antiviral activity of interferons Transport of antibodies (submucosal Prevention of adherence, complement-mediated glands) killing and opsonization AMPs, antimicrobial peptides and proteins; ROI, reactive oxygen intermediates; RNI, reactive nitrogen intermediates.

11 Effects of Cigarette Smoke and Diesel Exhaust on the Innate Immune Function

205

et al. 2006). These molecules include peptides and proteins such as lysozyme and lactoferrin, and are collectively referred to as antimicrobial peptides and proteins (AMPs). Whereas most of these AMPs were originally discovered based on their direct antimicrobial activity, it is now well established that these compounds also display a range of other activities and are involved in inflammation, immunity and repair processes. The major classes of antimicrobial peptides present in human airway secretions are the defensins and the cathelicidins. Based on their structure, the human defensin family can be divided into the a- and the b-defensin subfamilies. Human airway secretions contain a-defensins that are produced and secreted by neutrophils, whereas b-defensins present in these secretions are mainly derived from epithelial cells. In contrast to other species that contain a range of cathelicidin antimicrobial peptides, in humans only one member of the cathelicidin family is expressed: hCAP18/LL-37. Airway secretions contain this hCAP18/ LL-37 that is derived from neutrophils, but also e.g. epithelial cells and mast cells may contribute to its production. Antimicrobial proteins detected in airway secretions include lysozyme, lactoferrin, cationic serine proteinase inhibitors (secretory leukocyte proteinase inhibitor (SLPI) and elafin), and surfactant proteins SP-A and SP-D. In addition to these AMPs, reactive oxygen and nitrogen intermediates (ROI and RNI, respectively) that are formed by specialized enzyme systems present in the epithelium and in recruited inflammatory cells contribute to killing of micro-organisms. The expression of some of the components of this defense system show constitutive expression, thus providing a constant level of protection. The expression of other components is inducible following microbial exposure, inflammation and/or tissue repair processes. Airway epithelial cells may signal the presence of microbial exposure using so-called pattern recognition receptors that detect conserved molecular patterns present on pathogens (Akira et al. 2006; Bals and Hiemstra 2004). The cells respond to this exposure by e.g. increasing the production of certain AMPs, and by producing cytokines and chemokines. Cytokines signal to a vast range of cell types, including endothelial cells that line the vessels. This signaling results in increased expression of adhesion molecules on these endothelial cells that facilitate the recruitment of inflammatory and immune cells, such as the phagocytic neutrophils and monocytes, antigen-presenting dendritic cells and a range of other cell types. Chemokines produced by the airway epithelium form a chemotactic gradient along which the inflammatory and immune cells can migrate to the site of infection. In addition, viral exposure leads to the epithelial production of interferons with direct and indirect antiviral activities. Another mechanism that contributes to host defense against infection is the transport of antibodies that are produced by the adaptive immune system to mucosal secretions, which primarily occurs in the submucosal glands. These secreted antibodies (mainly of the IgA and IgM classes) prevent adherence of micro-organisms, activate the complement system and cause opsonization that facilitates recognition and removal by phagocytic cells.

206

11.3

P.S. Hiemstra

Chronic Obstructive Pulmonary Disease – Role of Cigarette Smoking and Air Pollution in Chronic Inflammatory Lung Disease

Chronic obstructive pulmonary disease (COPD) is a major health problem on a global scale (Pauwels and Rabe 2004; Mannino and Buist 2007; Barnes 2008). COPD is a very common disease characterized by progressive and irreversible airways obstruction, and its incidence is increasing world-wide. The major cause is cigarette smoking, but also indoor air pollution that is caused by cooking and heating in poorly ventilated houses may contribute to the development. COPD is characterized by chronic airway inflammation caused by inhalation of the toxic compounds present in cigarette smoke and air pollution, leading to the development of lung injury and irreversible airflow obstruction. Airway inflammation is markedly increased during acute episodes; these exacerbations are frequently accompanied by respiratory infections. Bacterial and viral infections are associated with COPD both in clinically stable patients as well as during exacerbations, and may contribute to the decline in lung function. At present no effective pharmaceutical intervention is available that slows the progressive loss of lung function, and smoking cessation is by far the most effective treatment. A large number of epidemiological studies have shown an association between levels of various air pollutants and hospital admissions for a variety of reasons, including pulmonary disorders (Sydbom et al. 2001; Riedl and az-Sanchez 2005). Traffic emissions that result from the combustion of fossil fuels constitute a major source of this air pollution. Despite advances in technology aimed to reduce the emission of hazardous compounds, diesel engines contribute more to traffic-related air pollution than gasoline engines. Combustion of diesel fuel results in the generation of diesel exhaust particles (DEPs), as well as various gaseous compounds that have adverse health effects. A large number of animal and human studies have shown that diesel causes pulmonary inflammation (Sydbom et al. 2001).

11.4

Modulation of Host Defense

Cigarette smoke has marked effects on epithelial gene expression, and these effects are thought to be involved in the development of lung cancer as well as COPD. The effect of cigarette smoke on airway epithelial gene expression has been explored in animal models and using in vitro cultures of human airway epithelial cells. These studies have shown that induction of goblet cell hyperplasia is an important consequence of smoke exposure that leads to increased mucus production in the airways. This effect has been shown to be mediated by smoke-induced activation of the epidermal growth factor receptor (EGFR) (Takeyama et al. 2001). Other studies have used epithelial cells derived from never smokers, current and ex-smokers to identify genes that are associated with the acute effects of cigarette smoke and those

11 Effects of Cigarette Smoke and Diesel Exhaust on the Innate Immune Function Table 11.2 Mechanisms involved in impaired host defense resulting from exposure to cigarette smoke or diesel exhaust

207

Direct Decreased mucociliary clearance l Increased microbial adherence to epithelial surfaces l Decreased production of antimicrobial molecules l Decreased activity of phagocytes Indirect l Increased inflammation l

associated with long-term effects of smoke exposure, and studied whether or not changes in gene expression are reversible after smoking cessation (Spira et al. 2004; Chari et al. 2007; Beane et al. 2007). These smoke-induced changes in gene expression were identified using gene expression profiling techniques such as microarray and serial analyis of gene expression (SAGE). The studies showed that in vivo smoke exposure not only affects mucin expression, but also changes expression of a range of genes involved in xenobiotic metabolism and antioxidant responses, as well as several oncogenes. Cigarette smoke not only causes airways inflammation, but also markedly affects host defense against respiratory infections. The increased risk of smokers and those exposed to environmental tobacco smoke (ETS; also referred to as “passive smoke”) to respiratory infections is well-documented (Arcavi and Benowitz 2004; Sopori 2002). A variety of respiratory infections are associated with smoking, including community-acquired pneumonia (CAP) which is an important cause of morbidity and mortality (Almirall et al. 2008; Almirall et al. 1999; Baik et al. 2000). Animal models of respiratory infections have confirmed the effect of smoke exposure on sensitivity to infection. Cigarette smoke has a suppressive effect on many elements of the innate and adaptive immune system (Sopori 2002) (Table 11.2). Smoke-induced inflammation itself may impair an efficient defense against respiratory pathogens by causing tissue injury and impairing effective clearance by phagocytes. Cigarette smoke decreases mucociliary clearance by inhibiting ciliary activity and increasing mucus production. It also increases bacterial adherence to epithelial surfaces (Ozlu et al. 2008) and thus promotes infection. The remodeling process of the epithelium that occurs during smoke- and/or inflammation-induced injury and subsequent repair has been shown to favor bacterial infections (Puchelle et al. 2006). Finally, cigarette smoking may decrease local defense mechanisms as shown by studies on alveolar macrophage function in smokers (Hodge et al. 2007). This study demonstrated a decreased phagocytic ability of macrophages from smokers, that was partially reversed by smoking cessation in COPD patients. In another study in COPD patients, a decreased ability to ingest non-typeable Haemophilus influenzae was reported. Both active smoking and COPD itself (by investigating ex-smokers with COPD) were identified as determinants in the decreased ability of macrophages from smokers with and without COPD to ingest Haemophilus influenzae (Berenson et al. 2006). Diesel exhaust particles (DEPs) constitute another important environmental risk factor for respiratory infections. The majority of DEPs are found in the size range classified as fine (0.1–2.5 mm) and ultrafine (<0.1 mm) and are thus present in the

208

P.S. Hiemstra

respirable fraction of inspired air that deposits along the respiratory tract and reaches the lower airways and alveoli (Maier et al. 2008; Riedl and az-Sanchez 2005; Dobson 2007). Deposition of DEPs and presence of the gaseous phase of diesel exhaust is especially problematic in the alveoli because of the extremely thin air-blood barrier in this part of the lung. A large number of studies have documented an increase in susceptibility to respiratory infections resulting from diesel exposure (Hahon et al. 1985; Castranova et al. 2001; Ciencewicki and Jaspers 2007). Several mechanisms have been shown to contribute to impaired host defense that increases the risk of respiratory infections. These mechanisms are reminiscent of the mechanisms involved in cigarette smoke induced impairment of host defense, and include decreased mucociliary clearance, increased microbial adherence, and decreased production of antimicrobial molecules (Table 11.2). In addition, diesel induced inflammation may impair host defense analogous to smoke induced inflammation. These findings are illustrated by a recent study from Gowdy et al. (2008). They used exposure to moderate and high levels of diesel exhaust to study the effect of diesel on inflammation and innate immunity in mice, using daily exposures for 1 or 5 days. DEP caused a marked increase in tissue injury, and markers of inflammation in the lung, including neutrophils, as well as increased expression of proinflammatory cytokines such as TNF-a and IL-6, and the adhesion molecule ICAM-1. ICAM-1 not only serves as an adhesion molecule for a variety of inflammatory cells, but also as an adhesion molecule for the major group of rhinovirus, the cause of the common cold. In addition to causing inflammation, DEP exposure resulted in a decrease in host defense molecules such as surfactant protein A and D (SP-A and SP-D), as well as Clara cell secretory protein (CCSP). SP-A and SP-D are members of the collectin family of innate immune proteins that bind a range of respiratory pathogens and contribute to their removal, whereas CCSP (also known as CC16 or CC10) is a anti-inflammatory protein that also contributes to host defense. Another recent in vitro study using epithelial cell culture showed a synergy between the proinflammatory cytokine IL-1b and DEPs in the induction of both proinflammatory gene expression, as well as an increase in b-defensin-2 (hBD-2) expression (Nam et al. 2006). These data show that under certain circumstances DEPs not only cause inflammation, but may also increase expression of selected AMPs. Findings such as these may help to explain the increased susceptibility to respiratory infections following diesel exposure.

11.5

Concluding Remarks and Directions for Further Research

It is evident that active smoking and environmental exposure to cigarette smoke and diesel exhaust increases the risk of infections of the upper and lower airways. The mechanisms underlying this increased risk for infections have been partially unraveled by the various studies that show a suppressive effect of smoke and diesel exhaust on inflammation, mucociliary clearance, microbial adherence to epithelial

11 Effects of Cigarette Smoke and Diesel Exhaust on the Innate Immune Function

209

surfaces, and innate and adaptive immunity. Further research is needed to better understand the mechanisms underlying this increased susceptibility. It is incompletely understood how cigarette smoke and diesel exhaust interact with other environmental factors that affect host defenses, such as inhaled allergens and allergic inflammation resulting from allergen exposure in susceptible individuals. This is highly relevant for the large group of patients with an allergy to inhaled allergens, including patients with asthma. In addition, the mechanisms that regulate passage of inhaled particles through mucus, and how a different mucus composition in inflammatory lung disease affects this passage need to be clarified. Other individual risk factors for the hazardous effects of smoke and diesel on host defense need to be studied, including genetic polymorphisms relevant to immunity. Since both cigarette smoke and diesel exhaust are complex mixtures containing particles and gaseous compounds, the relative contribution of the various components to the impairment of host defense needs to be unraveled. Finally, the interaction of smoke and diesel exhaust with specific components of the immune system requires further attention: do they affect the action of antimicrobial peptides, and how do they affect recruitment and maturation of pulmonary dendritic cells? These are just a few of the many questions that remain to be solved in this area of research that is highly relevant to human health.

References Akira S, Uematsu S, Takeuchi O (2006) Pathogen recognition and innate immunity. Cell 124:783–801 Almirall J, Bolibar I, Serra-Prat M, Roig J, Hospital I, Carandell E, Agusti M, Ayuso P, Estela A, Torres A (2008) New evidence of risk factors for community-acquired pneumonia: a population-based study. Eur Respir J 31:1274–1284 Almirall J, Gonzalez CA, Balanzo X, Bolibar I (1999) Proportion of community-acquired pneumonia cases attributable to tobacco smoking. Chest 116:375–379 Arcavi L, Benowitz NL (2004) Cigarette smoking and infection. Arch Intern Med 164:2206–2216 Baik I, Curhan GC, Rimm EB, Bendich A, Willett WC, Fawzi WW (2000) A prospective study of age and lifestyle factors in relation to community-acquired pneumonia in US men and women. Arch Intern Med 160:3082–3088 Bals R, Hiemstra PS (2004) Innate immunity in the lung: how epithelial cells fight against respiratory pathogens. Eur Respir J 23:327–333 Barnes PJ (2008) Immunology of asthma and chronic obstructive pulmonary disease. Nat Rev Immunol 8:183–192 Beane J, Sebastiani P, Liu G, Brody JS, Lenburg ME, Spira A (2007) Reversible and permanent effects of tobacco smoke exposure on airway epithelial gene expression. Genome Biol 8:R201 Berenson CS, Garlipp MA, Grove LJ, Maloney J, Sethi S (2006) Impaired phagocytosis of nontypeable Haemophilus influenzae by human alveolar macrophages in chronic obstructive pulmonary disease. J Infect Dis 194:1375–1384 Castranova V, Ma JY, Yang HM, Antonini JM, Butterworth L, Barger MW, Roberts J, Ma JK (2001) Effect of exposure to diesel exhaust particles on the susceptibility of the lung to infection. Environ Health Perspect 109(Suppl 4):609–612 Chari R, Lonergan KM, Ng RT, MacAulay C, Lam WL, Lam S (2007) Effect of active smoking on the human bronchial epithelium transcriptome. BMC Genomics 8:297

210

P.S. Hiemstra

Ciencewicki J, Jaspers I (2007) Air pollution and respiratory viral infection. Inhal Toxicol 19:1135–1146 Diamond G, Legarda D, Ryan LK (2000) The innate immune response of the respiratory epithelium. Immunol Rev 173:27–38 Dobson J (2007) Toxicological aspects and applications of nanoparticles in paediatric respiratory disease. Paediatr Respir Rev 8:62–66 Gowdy K, Krantz QT, Daniels M, Linak WP, Jaspers I, Gilmour MI (2008) Modulation of pulmonary inflammatory responses and antimicrobial defenses in mice exposed to diesel exhaust. Toxicol Appl Pharmacol 229:310–319 Hahon N, Booth JA, Green F, Lewis TR (1985) Influenza virus infection in mice after exposure to coal dust and diesel engine emissions. Environ Res 37:44–60 Hodge S, Hodge G, Ahern J, Jersmann H, Holmes M, Reynolds PN (2007) Smoking alters alveolar macrophage recognition and phagocytic ability: implications in chronic obstructive pulmonary disease. Am J Resp Cell Mol Biol 37:748–755 Maier KL, Alessandrini F, Beck-Speier I, Hofer TP, Diabate S, Bitterle E, Stoger T, Jakob T, Behrendt H, Horsch M, Beckers J, Ziesenis A, Hultner L, Frankenberger M, KraussEtschmann S, Schulz H (2008) Health effects of ambient particulate matter–biological mechanisms and inflammatory responses to in vitro and in vivo particle exposures. Inhal Toxicol 20:319–337 Mannino DM, Buist AS (2007) Global burden of COPD: risk factors, prevalence, and future trends. Lancet 370:765–773 Nam H, Ahn EK, Kim H, Lim Y, Lee C, Lee K, Vallyathan V (2006) Diesel exhaust particles increase IL-1beta-induced human beta-defensin expression via NF-kappaB-mediated pathway in human lung epithelial cells. Part Fibre Toxicol 3:9 Ozlu T, Celik I, Oztuna F, Bulbul Y, Ozsu S (2008) Streptococcus pneumoniae adherence in rats under different degrees and durations of cigarette smoke. Respiration 75:339–344 Pauwels RA, Rabe KF (2004) Burden and clinical features of chronic obstructive pulmonary disease (COPD). Lancet 364:613–620 Puchelle E, Zahm JM, Tournier JM, Coraux C (2006) Airway epithelial repair, regeneration, and remodeling after injury in chronic obstructive pulmonary disease. Proc Am Thorac Soc 3:726–733 Riedl M, az-Sanchez D (2005) Biology of diesel exhaust effects on respiratory function. J Allergy Clin Immunol 115:221–228 Rogan MP, Geraghty P, Greene CM, O’Neill SJ, Taggart CC, McElvaney NG (2006) Antimicrobial proteins and polypeptides in pulmonary innate defence. Respir Res 7:29 Sopori M (2002) Effects of cigarette smoke on the immune system. Nat Rev Immunol 2:372–377 Spira A, Beane J, Shah V, Liu G, Schembri F, Yang X, Palma J, Brody JS (2004) Effects of cigarette smoke on the human airway epithelial cell transcriptome. Proc Natl Acad Sci USA 101:10143–10148 Strieter RM, Belperio JA, Keane MP (2003) Host innate defenses in the lung: the role of cytokines. Curr Opin Infect Dis 16:193–198 Sydbom A, Blomberg A, Parnia S, Stenfors N, Sandstrom T, DahlenSE(2001) Health effects of diesel exhaust emissions. Eur Respir J 17:733–746 Takeyama K, Jung B, Shim JJ, Burgel PR, Dao-Pick T, Ueki IF, Protin U, Kroschel P, Nadel JA (2001) Activation of epidermal growth factor receptors is responsible for mucin synthesis induced by cigarette smoke. Am J Physiol Lung Cell Mol Physiol 280:L165–L172 Thornton DJ, Rousseau K, McGuckin MA (2008) Structure and function of the polymeric mucins in airways mucus. Ann Rev Physiol 70:459–486

Chapter 12

The Potential Harmful and Beneficial Effects of Nanoparticles in Children Karen G. Schu¨epp

12.1

Introduction

The inhaled route is one which potentially allows safe and effective delivery of drugs to children, not only to treat local airway diseases but also in the context of systemic diseases. On the other hand, inhalation of small particles such as nanoparticles has recently gained interest due to their potentially negative effects on children. Tobacco smoke exposure has garnered world-wide attention due to its harmful effects, not only on smokers themselves but also on people passively exposed to smoking. Other airborne particles such as end-products of ambient indoor as well as outdoor pollution from traffic, power-plants and other combustion sources have likewise the potential to harm children. Children interact with their environment in different and unique ways. Due to their developing organs, immature enzyme systems and behavioural differences vis-a`-vis adults, children seem to be the most vulnerable group with regard to harmful effects of air pollution. In medical science, it is specifically the class of particles with size <100nm that have gained interest: With nanotechnology emerging, significant advances in diagnosis and treatment of diseases are currently expected from nanoengineered materials. Applications in drug/gene delivery, diagnostics, imaging, and production of biocompatible materials are promising research fields. Unique features of inorganic nanoparticles, such as their increased surface to mass ratio, their modified quantum properties and their ability to absorb and carry probes and proteins open new opportunities and challenges for the development of improved drug delivery systems. Epidemiological and experimental data on combustion-derived nanoparticles,

K.G. Schu¨epp Department of Paediatric Respiratory Medicine, University Children’s Hospital, Bern, Switzerland Swiss Paediatric Respiratory Research Group, Switzerland e-mail: [email protected]

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_12, # Springer ScienceþBusiness Media B.V. 2010

211

K.G. Schu¨epp

212

such as e.g. diesel exhaust particles, demonstrates that any exposure can lead to a wide variety of effects. All in all, there is an enormous and constantly increasing body of research conducted for the controlled use of engineered nanoparticles in medicine. As alluded to earlier, a potential application of nanotechnology in medical science is drug delivery in young infants. As current drug solutions rely on large particle sizes, the available devices do not sufficiently allow the region of the lungs to be reached in infants. In the future, smaller particles such as nanoparticles may represent a promising therapeutic option to treat paediatric diseases. It is the objective of this chapter to discuss both the possible harmful and beneficial effects of small particles on children.

12.2 12.2.1

Why Are Children Not Small Adults Exposure Differences

Exposure to fine particles differs between children and adults in several ways. Compared to adults, children tend to spend more time outdoors and their physical outdoor activity additionally increases ventilation rates. Additionally, there is another group of children spending most time indoors before television with little activity. For these young children indoor pollution such as tobacco smoke seems to be more important than outdoor pollution. Due to their smaller airway geometry but less reduced minute ventilation compared to adults children have higher flow rates throughout their complete respiratory system. At increased ventilation rates, children start to breath through the mouth thus leading to a decreased filtering function of the nose for coarse particles; as a consequence, more particles are deposited in the lungs. On the other hand during physical activity air velocity is high which diminishes the amount of particles reaching the lower regions of the lungs thus particles are mainly deposited in the upper airways. Nanoparticles deposit mainly due to molecular diffusion while larger particles mainly deposit by inertial impaction. Therefore breathing pattern will mostly determine nanoparticle deposition by determining residential time within the lung system. Children spend most time outdoors during the second half of the day when outdoor pollution levels increases. The effect of ambient pollutants derived from motor-vehicle traffic on children are even more accentuated as these particles are emitted close to the ground and are dispersed afterwards: The breathing zone of small children, e.g. while seated in a pram, is closer to the ground than that of adults. It is therefore very likely that small children inhale more particles compared to adults (Heinrich et al. 2007). Last but not least, young children typically put their hands and anything that they pick up into their mouths, giving rise to a potential increase in exposure via non-nutritive ingestion (Sly and Pronczuk 2007).

12 The Potential Harmful and Beneficial Effects of Nanoparticles in Children

12.2.2

213

Differences in Developmental Biology

The respiratory tract and especially the respiratory epithelial-cell surfaces present a large fragile interface with the external environment, which during respiration is continuously exposed to a broad range of particles and antigens. Children are especially at risk from this exposure due to their developing organ systems. The harmful effects of ambient air pollution start even before a child’s birth. Contrary to the popular belief that the placental barrier protects the unborn child from all harm, it is known that the mother transmits drugs and toxicants to the foetus. The same applies later through transmission via breast milk. Many organs, such as the lungs and the brain, are not fully developed at birth and have long postnatal maturation periods. Cell number, cell type and cell function depend on the developmental stage. Therefore the timing of exposure is critical, i.e. the exposure of many toxicants is likely to have an age-specific effect. Infants differ from adults in that they have a lower body weight, a higher relative weight of the liver, a higher ratio between body surface and body weight, a smaller lung calibre and a higher particle deposition in the respiratory tract (Heinrich et al. 2007). Corrected for body weight, children have a minute ventilation of up to 50% of adults (Zeltner et al. 1987; Schwartz 2004). This is even more evident when corrected for surface area. Furthermore, the lung is generally immature and the epithelium is not fully developed thus leading to a decreased barrier function of the epithelium with a greater ease for toxicants to penetrate (Heinrich et al. 2007). Finally, young children have a greater surface area to volume ratio, which increases the potential for dermal exposure to pollutants (Sly and Pronczuk 2007).

Table 12.1 Why are children not small adults?

Exposure differences Playing activity Outdoor activity Time when playing outside Breathing zone Higher flow rates Non-nutritive ingestions Developmental biology Prenatal exposure Exposure by breast milk Long postnatal maturation periods Age specific toxicity Smaller airway geometry Lower body weight Higher ratio of body surface and body weight Increased deposition in the respiratory tract Higher minute ventilation per body weight and surface area Decreased barrier function of the lung epithelium Higher ratio of surface area to volume ratio

K.G. Schu¨epp

214

12.2.3

Normal Lung Development

The formation of the human respiratory system with its pulmonary and bronchial circulation is the net result of a complex interaction. Subsequent branching leads to the development of the conduction airways by about 16 weeks. The terminal respiratory units – the gas exchanging portion of the lung – develop from the latter third trimester of gestation through the first few months after birth, when alveoli with adult morphology are formed. By 24 weeks of gestation, the airways have the same wall structure as they have in the adult. The consequent thinning of the epithelium during gestation is led by capillaries which develop under the epithelium. This leads to the formation of a blood-gas barrier as thin as that of the adult (~0.6 mm) (Merkus and Hislop 2006). The number of alveoli increases with gestational age, and by term, between a third and a half of the adult number is present. The increase in lung volume seen during late gestation is mainly caused by the increase in alveolar number; alveolar surface area likewise shows a linear relationship between age and body weight. After birth, alveoli continue to rapidly increase in terms of number, size and complexity through a process of septation of the primary saccules. They reach adult values at the age of ~2 years. Once this process is completed, the number of alveoli remains constant and the lung continues to grow through increasing dimensions of all lung structures (De Jong et al. 2003). Still, compared to adults, airway diameters of children remain smaller than in adults. With increasing age, the breathing pattern changes as tidal volume gradually increases and the respiratory rate decreases. As the development of the lung is not complete until around school age, its integrity must remain uninterrupted. When the developing lung is exposed to toxicants during pregnancy or during the postnatal phase, this may significantly impact the overall growth and function of the respiratory system in children.

12.3

The Role of the Placenta

During human pregnancy, the foetal and maternal circulation are separated by the placental barrier which consists of five layers in the first trimester. This barrier undergoes drastic changes throughout pregnancy; the layers decrease in thickness and finally dissolve (Myren et al. 2007). Several cytochrome P450 proteins from the placental tissue have been identified to be largely responsible for the detoxication mechanisms of drugs and toxins. The transfer of compounds through the placenta can occur by different mechanisms: passive diffusion, carrier-mediated down a concentration gradient, active transport against a gradient or pinocytosis (Myren et al. 2007). Possible pathways of cellular uptake of nanoparticles are (macro) pinocytosis, clathrin-mediated endocytosis, non-mediated endocytosis or diffusion (Unfried et al. 2007). For example, Ziaee et al. (2007) compared maternal and foetal blood levels of heavy metals released from an artificial hip replacement in pregnant women.

12 The Potential Harmful and Beneficial Effects of Nanoparticles in Children Table 12.2 The role of the placenta

215

Transfer varies for each compound Placental activity adapts to exposure level Nanoparticles are expected to cross the placental barrier

The findings indicate that the amount of compounds transferred from the mother to the foetus vary substantially. This suggests that (a) nanoparticles may have the capacity to cross the placental barrier and be transferred from the mother to the foetus as well as (b) the respective amount of nanoparticle transfer likely depends on the chemical and physical characteristics of the particles. Kaiglova et al. compared (Kaiglova` et al. 2001) the placenta activity of smoking and non-smoking mothers living from areas with different levels of air pollution and environmental heavy metal pollution. The enzyme activity of lactate dehydrogenase in the placenta was significantly lower with non-smoking mothers living in a less polluted area than either smoking mothers and/or females living in more polluted areas. This is supported by the studies of Dussias et al. (1997) who found that placental activity can be amplified with higher exposure levels to lead. There the data show that the placenta has the ability to step up its protective function by increasing both metabolism and enzyme activity in order to adapt to increasing exposure levels.

12.4

Enzyme Systems

Many air pollutants such as tobacco smoke (Gamieldien and Maritz 2004) and diesel exhaust (Fanucchi et al. 1997a) are metabolized by the human cytochrome P 450 system. Driven by the age-specific metabolism activity of organs, prenatal detoxification mainly takes place in the liver where enzyme levels increase with gestational age. However, the enzyme levels in the neonatal lung are still significantly lower than in adults (Omiecinski et al. 1994). Enzyme levels of total P450 content in human liver microsomes remain stable at about one third of the adult value and do not reach adult levels before the end of the first year of age (Treluyer et al. 1996). Irrespective of enzyme levels, the activity of enzymes in neonates still remains remarkably low in comparison to adult enzyme systems (Day et al. 2006; Fanucchi et al. 2000). It is hence to be expected that the developing organism has a reduced ability to detoxify any possible toxic effects of nanoparticles.

12.5

Repair After Injury

Lung epithelial injury is known to be an initiating factor in several lung diseases such as asthma, chronic obstructive lung disease and emphysema, and pulmonary fibrosis. The proportion between injury and repair ultimately determines to what extent the structure of the lung can be maintained, whether remodelling occurs, or

K.G. Schu¨epp

216 Table 12.3 Repair functions of the immature lung system after exposure to particulate matter

Increased susceptibility to injury by volatile compounds Susceptibility at doses below the no-effect limit of adults Developing epithelial cells may fail to repair after injury Timing of exposure determines pattern of injury and repair Critical periods or points of increased susceptibility

whether emphysema or fibrosis may result. Infants are more susceptible to injury by volatile compounds, even at doses below the no-effect limit of adults (Fanucchi et al. 1997a, b; Plopper et al. 1994). It has furthermore been shown that developing epithelial cells in tracheobronchial airways may fail to repair following acute injury (Smiley-Jewell et al. 1998). Timing of exposure during the development seems to play an important role in the pattern of injury and repair. There appears to be critical time periods or points in time when tissue susceptibility is more pronounced (Fanucchi et al. 1997a); these time periods include both foetal and early postnatal periods of life (Pinkterton and Joad 2006).

12.6

Neural Development and Oxidative Stress

Normal lung development requires precise nerve connections. Restructuring of neural processes after an airway injury may provide the basis for hyperactivity resulting in decreased airway function and/or asthma. In infant monkeys, repeated exposure to allergen or oxidant air pollutants induces significant changes in the neural component within the epithelial compartment of pulmonary airways (Larson et al. 2004): Epithelial innervation re-establishes after cessation of exposure, but subsequent recovery leads to an exaggerated increase in airway nerve density or atypical nerve distribution. This implies that with immature lungs, allergen- or oxidant-induced airway injuries permanently disturb processes that are crucial in the establishment of normal airway innervation. Thus they may trigger persistent airway diseases in children (Kajekar et al. 2007). Before delivery, the foetus lives in a hypoxic environment. With the sudden onset of lung ventilation at birth, partial oxygen pressure increases fivefold (from ~20 to 100 mmHg) which produces many free oxygen radicals and imposes a significant oxidative stress on the foetus. Levels of glutathione, an antioxidant enzyme, as well as other antioxidant enzymes such as superoxidedismutase and catalase, remain low during early gestation. In late gestational age, enzyme levels then increase which is considered as preparation for birth. Antioxidant enzyme activity is not stable, i.e. substantial pre-birth activity increases might often be followed by activity declines just following birth (Rickett and Kelly 1990). In the past, this led to the hypothesis that the antioxidant enzyme system is virtually complete before delivery (Tanswell and Freeman 1984). However meanwhile it has been shown that adult enzyme activity levels are not reached before the age of around 6 months; some enzyme systems may take even longer to develop (WHO). This is particularly important since nanoparticles can induce oxidative stress as

12 The Potential Harmful and Beneficial Effects of Nanoparticles in Children

217

mentioned elsewhere in this book. The role of the antioxidant system in preterm babies has so far not been investigated. However, as the prenatal antioxidant system may not yet be fully developed, preterm babies are likely to be even more at risk of oxidative stress and injury.

12.7

The Immune System

Respiratory epithelial-cell surfaces present a large fragile interface to the external environment: during respiration they are continuously exposed to a broad range of antigens. Maintenance of local immunological homeostasis and hence the integrity of these gas-exchange surfaces must remain uninterrupted. Nanoparticles may either attenuate or amplify the immune response depending on their size and physical–chemical properties (Goppert and Muller 2005; Gref et al. 1994; Leu et al. 1984). As the placenta contains many cytokines and other immune mediators, nanoparticles have an enormous potential to influence the foetal immune development in a complex and unpredictable manner. Furthermore, placenta cells are likely to be sensitive to adverse environmental exposures that may affect pathways responsible for the alarming increase in immune disease such as asthma in very early life (Prescott 2008). For example, placentas exposed to diesel exhaust show inflammation, a congestion of immune cells as well an increase in different markers of inflammation (Fujimoto et al. 2005). Immune cells in cord blood of children with prenatal tobacco smoke exposure show at birth a decrease in all leukocytes including dendritic cells (Pachlopnik et al. 2007). Dendritic cells are recognised as a key antigen-presenting cell with regard to immunity and induction of T cell tolerance (Chen et al. 2006). Reduced dendritic cell levels thus increase susceptibility to infections. Type 1 and type 2 immunity is distinguished by CD4 T helper cell subsets, with type 1 (Th1) cells producing IFN-g and IL-2 and type 2 (Th2) cells secreting IL-4 and IL-10. At birth responses to common environmental antigens show a skew towards Th2 cell responses and are virtually present in all newborn infants (Prescott et al. 1998). In healthy infants, cell responses are later directed towards the Th1 cell response and thus immune tolerance. Infants exposed to relatively high levels of inhalant allergens during the first few months of life have an increased risk of allergic respiratory disease when exposed to the same allergens in their adult life (Holt et al. 1990). The immune system of newborns is still partially locked into the Th-2 skewed cytokine phenotype. It has been hypothesised that this may increase the risk of the infants developing a potentially pathogenic Th-2skewed allergen specific Th-cell memory. This risk may be even more pronounced in the atopic genotype (Holt 1998) or in the presence of prenatal allergen exposure. Nanoparticles acting as an allergen might therefore contribute to the risk of the development of allergic respiratory disease after prenatal or exposure early in life. During normal development major changes also occur in other immunocompetent and inflammatory cells. Alveolar macrophages are important suppressors of immune responses, but they further have the vital role of removing and killing

K.G. Schu¨epp

218 Table 12.4 The developmental immune system and particulate matter

Nanoparticles may attenuate or amplify an immune response Prenatal influence on immune system (immune cells and markers of inflammation) by placental cytokines and other immune mediators Placenta cells are likely to be sensitive to adverse environmental exposure Skew towards Th2-response at birth Early life exposure to allergens increases the risk for allergic respiratory disease Low number of alveolar macrophages and functional impairment in infants

infections material arising at the air-tissue interface. The number of alveolar macrophages is increased below 2 years of age. In addition, this is accompanied by immature functional impairment which may add to the vulnerability of children to environmental insults and increase the susceptibility to respiratory infections or antigens (Grigg et al. 1999). Carbonaceous particles are shown to deposit in the lower airway of healthy children where they can be phagocytised by alveolar macrophages. The percentage of particle containing alveolar macrophages does not increase with age. It is however increased in children living near a main road compared to those living on a quiet residential road (Bunn et al. 2001). Carbon content of airway macrophages in induced sputum of healthy children is shown to correlate positively with modelled exposure to PM10. Additionally there is evidence of an inverse relationship between the carbon content in the macrophages and overall lung function (Kulkarni et al. 2006). This strengthens the hypothesis of a causal association between inhaling carbonaceous particles and impaired lung function. In the same study, carbon levels in airway macrophages were lower for children with asthma vis-a`-vis healthy children, despite the higher levels of modelled PM10. If this finding supports the hypothesis that the phagocytosis of carbon particles by airway macrophages may be impaired in severe asthma or that the activity of the macrophages is decreased due to high local corticosteroids deposition from daily inhaled medication remains unclear.

12.8

Pre- and Postnatal Exposure

First exposure and hence potential harmful effects to air pollutants begin even before a child is born. The placenta insufficiently protects the foetus from harmful pollutants that enter the maternal blood circulation. Pollutants entering the foetal circulation may affect foetal growth and development. It is therefore not surprising that a range of papers report an association between prenatal ambient air exposure and early foetal loss, preterm delivery, lower birth weight and intrauterine growth retardation as reviewed extensively (Lacasan˜a et al. 2005; Schwarz 2004; Salvi 2007). Recently Latzin et al. additionally showed that prenatal exposure to PM10 is related to impaired lung function shortly after birth (Latzin et al. 2009). Significant

12 The Potential Harmful and Beneficial Effects of Nanoparticles in Children

219

relationships between respiration-related causes of death and PM10 (Woodruff et al. 2006) as well as PM2.5 (Woodruff et al. 2008) have been identified. In contrast to SO2 and NO2 exposure (Dales et al. 2004), so far no clear evidence has been found of any relationship between particulate exposure and sudden infant death syndrome. Infant mortality does not seem to be correlated with fine particle exposure on the same day and/or the two previous days, but has been correlated with levels of exposure to particulate matter on day 3 to 5 (Loomis et al. 1999). These findings are in accordance with clinical studies showing a pronounced effect of particulate matter with a 3 to 4 days delay both in children (Andersen et al. 2008) as well as in adults (Desqueyroux et al. 2002). The adverse effects of prenatal exposure to air pollutants such as tobacco smoke are not limited to the last weeks of pregnancy and are associated with decreased respiratory function in infants (Hoo et al. 1998; Tager et al. 1995; Brown et al. 1995). Lung function in childhood is a strong predictor for lung function in early adulthood, which, in turn, is an important health predictor throughout life. There is growing evidence of a foetal origin of later diseases. Children of smoking parents have more respiratory symptoms and poorer lung function in adulthood. Thus, parental tobacco smoking appears to influence the development of the child’s airways and to negatively impact the long-term respiratory health (Svanes et al. 2004). Chronic obstructive pulmonary disease is conventionally thought as a disease of adult smokers, related to airway inflammation and structural airway changes. However, there is important epidemiological evidence from several studies including subjects from birth to late middle age, that early life events such as antenatal interference with lung growth may program the child to be at increased risk for future chronic obstructive pulmonary disease (Bush 2008). Exposure to fine particulate matter may also lead to respiratory effects in children. It has been suggested that the combustion-generated component of particulate matter strongly influences lung function in exposed children (Allen et al. 2008; Horak et al. 2002). Asthmatic children seem especially sensitive to particulate matter. In addition, an increased exposure to PM2.5 is likely to contribute to the exacerbation of asthmatic conditions (Tang et al. 2007). Reports of cough without infection and dry cough at night as well as wheezing and bronchitis were found to be associated with exposure to fine particulate matter (Gehring et al. 2002; Hertz-Picciotto et al. 2007). In a large cohort study comparing the lung function at 10 and 18 years of age, local exposure to freeway traffic was shown to have an adverse effect on lung development, which was independent of regional air quality, and which could result in important deficits in attained lung function in later life (Gauderman et al. 2007). Respiratory infections are very common in the first months in life, and a positive correlation with prenatal smoke exposure has been identified. However, the influence of prenatal smoking declines in the first year of life (Latzin et al. 2007). Similar findings come from studies researching the children’s respiratory effects upon relocation to areas with altered air pollution levels. For subjects who moved during childhood, PM10 level changes (in either a positive or a negative direction) were reflected in statistically significant changes in rates of annual growth in mean

K.G. Schu¨epp

220 Table 12.5 Clinical effects of prenatal and postnatal exposure to particulate matter

Respiratory effects Lung function Lung growth Exacerbation of asthmatic conditions Cough, wheeze Bronchitis and other respiratory infections Other clinical effects Early fetal loss Preterm delivery Low birth weight Intrauterine growth retardation Infant mortality Influence declines after exposure cessation Fetal origin of later disease

expiratory flow (Avol et al. 2001). Accordingly, a decline in ambient air pollution has a positive effect on respiratory symptoms such as bronchitis and cold in children (Heinrich et al. 2000).

12.9

Beneficial Aspects of Nanoparticles in Children

Despite the harmful effects of environmental exposure to nanoparticles, there is beneficial potential in nanoparticles as a drug delivery tool. This is particularly significant due to existing problems with drug delivery in children. Total lung deposition is a key determinant of clinical efficacy and systemic side effects of any aerosolized drug. An orally deposited and swallowed drug is reabsorbed in the gastrointestinal tract and mostly eliminated by the first pass metabolism in the liver. Patient-related factors such as airway anatomy, breathing patterns and co-operation and compliance complicate drug deposition in childhood compared to adults. Airway diameters in children are smaller than in adults, thus leading to increased central lung deposition due to increased impaction. Young children are preferential nose breathers. As the nose is a very efficient filter, most of the aerosol inhaled through the nose does not reach the lung. In adults, Everard et al. (1993) show that lung deposition by nasal breathing is only half as much as by mouth breathing. Children have characteristic age-dependent breathing patterns which greatly influence the main deposition mechanics (impaction, sedimentation and diffusion) and hence lung deposition. Tidal volume increases and respiratory rate decreases with increasing age. Respiratory rate and peak inspiratory flow are higher in children who are awake than when asleep. When crying, the inspiratory flow of children increases six- to sevenfold. This is clinically relevant as compliance with aerosol therapy is often relatively low with small children. Small particles could at least partially overcome these problems. It has been shown both in vitro (Schu¨epp et al. 2005) and in vivo (Schu¨epp et al. 2004) that using smaller particles facilitate higher amount of lung deposition. Patients treated

12 The Potential Harmful and Beneficial Effects of Nanoparticles in Children

221

with smaller aerosol droplet sizes tend to have greater improvement in pulmonary function than those receiving larger droplets of aerosol (Geller et al. 1998). However for commercially available inhalation devices, the ratio of lung deposition to drug delivered remains generally low. It has been shown that submicron or nanosize particles are able to reach the lungs of young children without excessive deposition in the upper airways (Schu¨epp et al. 2007; Kropp et al. 1993). With a higher deposition efficacy, inhalation procedure could be shortened and compliance with drug administration amended. Nanoparticles deposit mainly due to molecular diffusion thus deposition will mainly be determined by residential time within the lung system, but a considerable part will be exhaled. This should be considered when developing devices for inhalation of nanoparticles, rebreathing devices with sufficient oxygen supply might be an option. The delivery of therapeutic genes and drugs directly to the lung via nanoparticles carriers is therefore a promising area of research for the treatment of paediatric respiratory disorders. Delivery of therapeutic genes with nanoparticles for cystic fibrosis, the most common lethal genetic disease in children, has gained recent attention (reviewed in Rosenecker et al. 2006). The disease progressively destroys the lungs over a period of a few decades, and has no cure at present. It mainly affects the lungs, but other tissues can also be affected. Due to a genetic defect the lungs build up a large amount of mucus and the airway epithelium is coated by a thicker and more viscous mucus layer than in healthy subjects. Researchers are therefore focusing on using magnetic nanoparticles to penetrate the mucus layer in the lungs and to reach the underlying epithelium for gene transfer to the affected cells. This technique involves attaching the therapeutic gene to magnetic nanoparticles. By positioning strong, high gradient magnets over the target sites on the surface of the body (such that the target tissue is between the particles and the magnet), the inhaled nanoparticles may then be targeted to the airway epithelial cells (Dobson 2006 Feb, 2007; Xenariou et al. 2006). Another potential use for nanoparticles lies in the way of vaccines. Infectious diseases represent a substantial threat to emerging countries. Suffering, disability and death resulting from infectious diseases have become a significant challenge for low-income countries and international health agencies. Vaccines are considered as a most effective and economic method of globally preventing infectious diseases. Most vaccines require multiple injections which are a substantial limitation, especially in childhood during which most vaccines have to be administered. Moreover the immune response to vaccination is variable, and a small proportion of children receiving certain antigens simply do not mount a response that is capable of conferring protection. For this reason, every vaccine has a failure rate (Va´zquez et al. 2001; Galil et al. 2002), which is a constraint in various parts of the world. This is even more pronounced for paediatric populations in emerging countries, where the risk of infectious diseases is greatest and any follow-up often proves to be impractical. Thanks to recent advances, it is now possible to trigger an immune response with nanoparticles. The huge area of the lungs favours a delivery via inhalation (Gehr et al. 1978). The high density of dendritic cells in the entire respiratory tract that

222 Table 12.6 Beneficial aspects of nanoparticles in children

K.G. Schu¨epp Less dependent on breathing pattern and compliance Reach all regions of the lung independent on calibre size ! lung deposition Higher deposition efficacy ! shorter inhalation procedure, improved compliance Targeting of lung regions ! local side effects Cell targeting ! efficiency

sample and scrutinize inhaled antigen (Holt et al. 2008) make the lungs furthermore a delicate organ to induce any immune response. If further research achieves an optimal immune response with nanoengineered inhaled vaccines without requiring a needle and thereby overcoming the limitations of current vaccines, an incredible increase of vaccination compliance rate and safety may be within reach. For this reason, nanovaccines attract a great deal of financial and research interest and are likely to yield some promising results in the years to come. In summary, due to their developing organs, immature enzyme systems and behavioural differences vis-a`-vis adults, children seem to be the most vulnerable group with regard to harmful effects of air pollution. It is unknown to date if the body has mechanisms to protect itself from harms from nanoparticles. In contrast, smaller particles such as nanoparticles may represent a promising therapeutic option to treat paediatric diseases and overcome the limitations of current inhalation devices and solutions. Most of the current knowledge is based on information on aerosols in the micron range delivered to children. However aerosol particles in the nanorange do have different properties than bigger particles and it is likely that the minor part of nanoparticles reaches and deposit in the peripheral regions of the lung. But the impact on the long run of nanoparticles in the lung can be more important due to its continuing effects on growth and delicate balances between injury and repair, between maturating and protecting processes. Future studies are therefore needed to outline harmful and beneficial effects of nanoparticles in children.

References Allen RW, Mar T, Koenig J, Liu LJ, Gould T, Simpson C, Larson T (2008 Feb) Changes in lung function and airway inflammation among asthmatic children residing in a woodsmoke-impacted urban area. Inhal Toxicol 20(4):423–433 Andersen ZJ, Loft S, Ketzel M, Stage M, Scheike T, Mette MN, Bisgaard H (2008 Aug) Ambient air pollution triggers wheezing symptoms in infants. Thorax 63(8):710–716 Avol EL, Gauderman WJ, Tan SM, London SJ, Peters JM (2001 Dec 1) Respiratory effects of relocating to areas of differing air pollution levels. Am J Respir Crit Care Med 164(11): 2067–2072 Brown RW, Hanrahan JP, Castile RG, Tager IB (1995 Jan) Effect of maternal smoking during pregnancy on passive respiratory mechanics in early infancy. Pediatr Pulmonol 19 (1):23–28 Bunn HJ, Dinsdale D, Smith T, Grigg J (2001 Dec) Ultrafine particles in alveolar macrophages from normal children. Thorax 56(12):932–934

12 The Potential Harmful and Beneficial Effects of Nanoparticles in Children

223

Bush A (2008 Feb) COPD: a pediatric disease. COPD 5(1):53–67 Chen M, Wang YH, Wang Y, Huang L, Sandoval H, Liu YJ, Wang J (2006 Feb 24) Dendritic cell apoptosis in the maintenance of immune tolerance. Science 311(5764):1160–1164 Dales R, Burnett RT, Smith-Doiron M, Stieb DM, Brook JR (2004 Jun) Air pollution and sudden infant death syndrome. Pediatrics 113(6):e628–e631 Day KC, Plopper CG, Fanucchi MV (2006 Jul) Age-specific pulmonary cytochrome P-450 3A1 expression in postnatal and adult rats. Am J Physiol Lung Cell Mol Physiol 291(1):L75–L83 De Jong PA, Nakano Y, Lequin MH, Merkus PJ, Tiddens HA, Hogg JC, Coxson HO (2003 Aug) Estimation of lung growth using computed tomography. Eur Respir J 22(2):235–238 Desqueyroux H, Pujet JC, Prosper M, Squinazi F, Momas I (2002 May) Short-term effects of lowlevel air pollution on respiratory health of adults suffering from moderate to severe asthma. Environ Res 89(1):29–37 Dobson J (2006 Feb) Gene therapy progress and prospects: magnetic nanoparticle-based gene delivery. Gene Ther 13(4):283–287 Dobson J (2007 Mar) Toxicological aspects and applications of nanoparticles in paediatric respiratory disease. Paediatr Respir Rev 8(1):62–66. Epub 2007 Mar 21 Dussias V, Stefos T, Stefanidis K, Paraskevaidis E, Karabini F, Lolis D (1997) Lead concentrations in maternal and umbilical cord blood in areas with high and low air pollution. Clin Exp Obstet Gynecol 24(4):187–189 Everard ML, Hardy JG, Milner AD (1993 Oct) Comparison of nebulised aerosol deposition in the lungs of healthy adults following oral and nasal inhalation. Thorax 48(10):1045–1046 Fanucchi MV, Buckpitt AR, Murphy ME, Plopper CG (1997a) Naphthalene cytotoxicity of differentiating Clara cells in neonatal mice. Toxicol Appl Pharmacol 144(1):96–104 Fanucchi MV, Murphy ME, Buckpitt AR, Philpot RM, Plopper CG (1997b) Pulmonary cytochrome P450 monooxygenase and Clara cell differentiation in mice. Am J Respir Cell Mol Biol 17(3):302–14 Fanucchi MV, Buckpitt AR, Murphy ME, Storms DH, Hammock BD, Plopper CG (2000 Nov 1) Development of phase II xenobiotic metabolizing enzymes in differentiating murine clara cells. Toxicol Appl Pharmacol 168(3):253–267 Fujimoto A, Tsukue N, Watanabe M, Sugawara I, Yanagisawa R, Takano H, Yoshida S, Takeda K (2005 Aug) Diesel exhaust affects immunological action in the placentas of mice. Environ Toxicol 20(4):431–440 Galil K, Lee B, Strine T, Carraher C, Baughman AL, Eaton M, Montero J, Seward J (2002 Dec 12) Outbreak of varicella at a day-care center despite vaccination. N Engl J Med 347(24):1909–1915 Gamieldien K, Maritz GS (2004 Mar) Postnatal expression of cytochrome P450 1A1, 2A3, and 2B1 mRNA in neonatal rat lung: influence of maternal nicotine exposure. Exp Lung Res 30 (2):121–133 Gauderman WJ, Vora H, McConnell R, Berhane K, Gilliland F, Thomas D, Lurmann F, Avol E, Kunzli N, Jerrett M, Peters J (2007 Feb 17) Effect of exposure to traffic on lung development from 10 to 18 years of age: a cohort study. Lancet 369(9561):9571–7 Gehr P, Bachofen M, Weibel ER (1978 Feb) The normal human lung: ultrastructure and morphometric estimation of diffusion capacity. Respir Physiol 32(2):121–140 Gehring U, Cyrys J, Sedlmeir G, Brunekreef B, Bellander T, Fischer P, Bauer CP, Reinhardt D, Wichmann HE, Heinrich J (2002 Apr) Traffic-related air pollution and respiratory health during the first 2 yrs of life. Eur Respir J 19(4):690–698 Geller DE, Eigen H, Fiel SB, Clark A, Lamarre AP, Johnson CA, Konstan MW (1998) Effect of smaller droplet size of dornase alfa on lung function in mild cystic fibrosis. Dornase Alfa Nebulizer Group. Pediatr Pulmonol Feb;25(2):83–87 Goppert TM, Muller RH (2005) Polysorbate-stabilized solid lipid nanoparticles as colloidal carriers for intravenous targeting of drugs to the brain: comparison of plasma protein adsorption patterns. J Drug Target 13:179–187 Gref R, Minamitake Y, Peracchia MT, Trubetskoy V, Torchilin V, Langer R (1994) Biodegradable long-circulating polymeric nanospheres. Science 263:1600–1603

224

K.G. Schu¨epp

Grigg J, Riedler J, Robertson CF, Boyle W, Uren S (1999 Nov) Alveolar macrophage immaturity in infants and young children. Eur Respir J 14(5):1198–1205 Heinrich J, Slama R (2007 Oct) Fine particles, a major threat to children. Int J Hyg Environ Health 210(5):617–622 Heinrich J, Hoelscher B, Wichmann HE (2000 Jun) Decline of ambient air pollution and respiratory symptoms in children. Am J Respir Crit Care Med 161(6):1930–1936 Hertz-Picciotto I, Baker RJ, Yap PS, Dosta´l M, Joad JP, Lipsett M, Greenfield T, Herr CE, Benes I, Shumway RH, Pinkerton KE, Sra´m R (2007 Oct) Early childhood lower respiratory illness and air pollution. Environ Health Perspect 115(10):1510–1518 Holt PG, McMenamin C, Nelson D (1990) Primary sensitisation to inhalant allergens during infancy. Ped Allergy Immunol 1:3–13 Holt PG (1998 Jun) Programming for responsiveness to environmental antigens that trigger allergic respiratory disease in adulthood is initiated during the perinatal period. Environ Health Perspect 106(Suppl 3):795–800 Holt PG, Strickland DH, Wikstro¨m ME, Jahnsen FL (2008 Feb) Regulation of immunological homeostasis in the respiratory tract. Nat Rev Immunol 8(2):142–152 Hoo AF, Henschen M, Dezateux C, Costeloe K, Stocks J (1998 Sep) Respiratory function among preterm infants whose mothers smoked during pregnancy. Am J Respir Crit Care Med 158 (3):700–705 Horak F Jr, Studnicka M, Gartner C, Spengler JD, Tauber E, Urbanek R, Veiter A, Frischer T (2002 May) Particulate matter and lung function growth in children: a 3-yr follow-up study in Austrian schoolchildren. Eur Respir J 19(5):838–845 Kaiglova´ A, Reichrtova´ E, Adamca´kova´ A, Wso´lova´ L (2001) Lactate dehydrogenase activity in human placenta following exposure to environmental pollutants. Physiol Res 50(5):525–528 Kajekar R, Pieczarka EM, Smiley-Jewell SM, Schelegle ES, Fanucchi MV, Plopper CG (2007 Jan 15) Early postnatal exposure to allergen and ozone leads to hyperinnervation of the pulmonary epithelium. Respir Physiol Neurobiol 155(1):55–63 Kropp J, Overbeck B, Klumpp J, Hotze A, Biersack HJ (1993 Mar) Inhalation scintigraphy with an ultrafine aerosol in infants with functional bronchial stenoses. Clin Nucl Med 18(3):223–226 Kulkarni N, Pierse N, Rushton L, Grigg J (2006 Jul 6) Carbon in airway macrophages and lung function in children. N Engl J Med 355(1):21–30 Lacasan˜a M, Esplugues A, Ballester F (2005) Exposure to ambient air pollution and prenatal and early childhood health effects. Eur J Epidemiol 20(2):183–199 Larson SD, Schelegle ES, Walby WF, Gershwin LJ, Fanuccihi MV, Evans MJ, Joad JP, Tarkington BK, Hyde DM, Plopper CG (2004 Feb 1) Postnatal remodeling of the neural components of the epithelial-mesenchymal trophic unit in the proximal airways of infant rhesus monkeys exposed to ozone and allergen. Toxicol Appl Pharmacol 194(3):211–220 Latzin P, Frey U, Roiha HL, Baldwin DN, Regamey N, Strippoli MP, Zwahlen M, Kuehni CE (2007) Swiss Paediatric Respiratory Research Group. Prospectively assessed incidence, severity, and determinants of respiratory symptoms in the first year of life. Pediatr Pulmonol Jan;42 (1):41–50 Latzin P, Ro¨o¨sli M, Huss A, Kuehni CE, Frey U (2009 Mar) Air pollution during pregnancy and lung function in newborns: a birth cohort study. Eur Respir J 33(3):594–603 Leu D, Manthey B, Kreuter J, Speiser P, DeLuca PP (1984) Distribution and elimination of coated polymethyl [2–14C]methacrylate nanoparticles after intravenous injection in rats. J Pharm Sci 73:1433–1437 Loomis D, Castillejos M, Gold DR, McDonnell W, Borja-Aburto VH (1999 Mar) Air pollution and infant mortality in Mexico City. Epidemiology 10(2):118–123 Merkus P and Hislop A (2006) Respiratory diseases in infants and children. In: Frey U, Gerritsen J (eds) European Society Journals. Eur Respir Mon 37:8–21 Myren M, Mose T, Mathiesen L, Knudsen LE (2007 Oct) The human placenta–an alternative for studying foetal exposure. Toxicol In Vitro 21(7):1332–1340

12 The Potential Harmful and Beneficial Effects of Nanoparticles in Children

225

Omiecinski CJ, Aicher L, Swenson L (1994 Apr) Developmental expression of human microsomal epoxide hydrolase. J Pharmacol Exp Ther 269(1):417–423 Pachlopnik Schmid JM, Kuehni CE, Strippoli MP, Roiha HL, Pavlovic R, Latzin P, Gallati S, Kraemer R, Dahinden C, Frey U (2007) Swiss Pediatric Respiratory Research Group. Maternal tobacco smoking and decreased leukocytes, including dendritic cells, in neonates. Pediatr Res Apr;61(4):462–466 Pinkerton KE, Joad JP (2006 Mar) Influence of air pollution on respiratory health during perinatal development. Clin Exp Pharmacol Physiol 33(3):269–272 Plopper CG, Weir AJ, Nishio SJ, Chang A, Voit M, Philpot RM, Buckpitt AR (1994 May) Elevated susceptibility to 4-ipomeanol cytotoxicity in immature Clara cells of neonatal rabbits. J Pharmacol Exp Ther 269(2):867–880 Prescott SL (2008 Mar) Effects of early cigarette smoke exposure on early immune development and respiratory disease. Paediatr Respir Rev 9(1):3–9; quiz 10 Prescott SL, Macaubas C, Holt BJ, Smallacombe TB, Loh R, Sly PD, Holt PG (1998 May 15) Transplacental priming of the human immune system to environmental allergens: universal skewing of initial T cell responses toward the Th2 cytokine profile. J Immunol 160(10):4730–4737 Rickett GM, Kelly FJ (1990 Feb) Developmental expression of antioxidant enzymes in guinea pig lung and liver. Development 108(2):331–336 Rosenecker J, Huth S, Rudolph C (2006 Oct) Gene therapy for cystic fibrosis lung disease: current status and future perspectives. Curr Opin Mol Ther 8(5):439–445 Salvi S (2007 Dec) Health effects of ambient air pollution in children. Paediatr Respir Rev 8 (4):275–280 Schuepp K, S Vlasenko, S Minocchieri, R Kraemer, M Ammann, J Wildhaber (2007) Aerosol particles less than 1.3m reach the lungs of young children P096, ISAM Tours, France Schu¨epp KG, Devadson S, Roller C, Wildhaber JH (2004 Apr 3) A complementary combination of delivery device and drug formulation for inhalation therapy in preschool children. Swiss Med Wkly 134(13–14):198–200 Schu¨epp KG, Jauernig J, Janssens HM, Tiddens HA, Straub DA, Stangl R, Keller M, Wildhaber JH (2005) In vitro determination of the optimal particle size for nebulized aerosol delivery to infants. J Aerosol Med Summer 8(2):225–235 Schwartz J (2004 Apr) Air pollution and children’s health. Pediatrics 113(4 Suppl):1037–1043 Sly PD, Pronczuk J (2007 Dec) Guest editorial: susceptibility of children to pollutants. Paediatr Respir Rev 8(4):273–274 Smiley-Jewell SM, Nishio SJ, Weir AJ, Plopper CG (1998 Apr) Neonatal Clara cell toxicity by 4-ipomeanol alters bronchiolar organization in adult rabbits. Am J Physiol 274(4 Pt 1): L485–L498 Svanes C, Omenaas E, Jarvis D, Chinn S, Gulsvik A, Burney P (2004 Apr) Parental smoking in childhood and adult obstructive lung disease: results from the European Community Respiratory Health Survey. Thorax 59(4):295–302 Tager IB, Ngo L, Hanrahan JP (1995 Sep) Maternal smoking during pregnancy. Effects on lung function during the first 18 months of life. Am J Respir Crit Care Med 152(3):977–983 Tang CS, Chang LT, Lee HC, Chan CC (2007 Aug 15) Effects of personal particulate matter on peak expiratory flow rate of asthmatic children. Sci Total Environ 382(1):43–51 Tanswell AK, Freeman BA (1984 Jul) Pulmonary antioxidant enzyme maturation in the fetal and neonatal rat. I. Developmental profiles. Pediatr Res 18(7):584–587 Treluyer JM, Cheron G, Sonnier M, Cresteil T (1996 Aug 9) Cytochrome P-450 expression in sudden infant death syndrome. Biochem Pharmacol 52(3):497–504 Unfried K, Albrecht C, Klotz L, Mikecz A, Grether-Beck S, Schins R (2007) Cellular responses to nanoparticles: Target structures and mechanism. Nanotoxicology 1(1):52–71 Va´zquez M, LaRussa PS, Gershon AA, Steinberg SP, Freudigman K, Shapiro ED (2001 Mar 29) The effectiveness of the varicella vaccine in clinical practice. N Engl J Med 344(13):955–960 Woodruff TJ, Parker JD, Schoendorf KC (2006 May) Fine particulate matter (PM2.5) air pollution and selected causes of postneonatal infant mortality in California. Environ Health Perspect 114(5):786–790

226

K.G. Schu¨epp

Woodruff TJ, Darrow LA, Parker JD (2008 Jan) Air pollution and postneonatal infant mortality in the United States, 1999–2002. Environ Health Perspect 116(1):110–115 World health organisation WHO (2006) Environmental Health Criteria 237. Principles for evaluation health risks in children associated with exposure to chemicals. Xenariou S, Griesenbach U, Ferrari S, Dean P, Schedule RK, Cheng SH, Geddes DM, Plank C, Alton EW (2006 Nov) Using magnetic forces to enhance non-viral gene transfer to airway epithelium in vivo. Gene Ther 13(21):1545–1552 Zeltzer TB, Burri PH (1987) The postnatal development and growth of the human lung. II. Morphology. Respir Physiol Mar;67(3):269–382 Ziaee H, Daniel J, Datta AK, Blunt S, McMinn DJ (2007 Mar) Transplacental transfer of cobalt and chromium in patients with metal-on-metal hip arthroplasty: a controlled study. J Bone Joint Surg Br 89(3):301–305

Chapter 13

Targeting Drugs to the Lungs – The Example of Insulin S. Ha¨ussermann, G. Scheuch, and R. Siekmeier

13.1

Introduction

In the last three decades methods for recombinant synthesis of peptides and proteins were developed allowing the production of large amounts of these substances for clinical treatment (e.g. growth factors, hormones, monoclonal antibodies and cytokines) (Agu et al. 2001; Patton and Byron 2007). Because of their biochemical properties (high molecular weight, hydrophilia, sensitivity against chemicals and proteolytic enzymes) these compounds cannot be administered orally, but require parenteral administration resulting in negative effects on convenience and compliance of the patients in cases of chronic diseases (e.g. diabetes mellitus). Inhaled application (via nose or mouth) of high molecular weight compounds seems to be a method of choice. However, much better conditions for absorption are found in the lung periphery (i.e. the alveolar region) making the lung to an important target for inhalative administration of drugs with systemic mode of action. Firstly, the size of the alveolar surface is about the half of a tennis court depending on the distension of the lung and much larger than that of the nose (about 180 cm2) (Niven 1995; Wolff 1998). Another advantage is the thin alveolar epithelium. Its thickness in most regions is between 0.1 and 0.2 mm (Patton and Platz 1992) resulting in a total distance between epithelial surface and blood between 0.5 and 1.0 mm (Wolff 1998) which is much less than in the bronchial tract where the deposited substances have to pass a distance of 30–40 mm and more between mucus surface and blood (Wolff 1998, Patton and Byron 2007). Several preconditions must be fulfilled to allow

S. Ha¨ussermann (*) Air Liquide Research Center, CRCD, Paris, France e-mail: [email protected] R. Siekmeier Activaero GmbH, Gemu¨nden, Germany G. Scheuch Federal Institute for Drugs and Medical Devices (BfArM), Bonn, Germany

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_13, # Springer ScienceþBusiness Media B.V. 2010

227

S. Ha¨ussermann et al.

228

administration of adequate and reproducible drug doses for treatment of systemic diseases by inhaled aerosols. These are biophysical and physiological factors (e.g. aerosol particle size, and breathing manoeuvre (inspired volume, inspiratory flow, endinspiratory breathhold time)) which are subject of other reviews (Patton and Byron 2007; Scheuch et al. 2006; Scheuch and Siekmeier 2007) as well as physical and biochemical stability of the pharmaceutical compounds designed for aerosolisation (aqueous solution, dry powder, suspension or solution in propellants (Niven 1995; Yu and Chien 1997).

13.2

Physical Methods for Aerosol Administration

Requirements for inhalative drug administration are high efficiency of drug delivery, reproducible dosing, targeted delivery of the inhaled drug to the site of action, ease of device operation, short duration of treatment, minimised risk to the patient and the medical personnel, environmental protection and cost-effectiveness (Dhand 2004). For this purpose a number of products have been developed differing strongly in respect to their suitability for nebulisation and administration of the various compounds. However, in the past, low rates of pulmonary drug absorption were observed because the nebulisers used were not qualified for production of an adequate aerosol particle spectrum (Wolff 1998; Yu and Chien 1997) and did not take the breathing patterns of the patients into account.

13.2.1

Nebulisers

The appropriateness of nebulisers for administration of macromolecular compounds depends on the performance of the nebulisers (e.g. aerosol output, distribution width and variability of the aerosol particle spectrum) as well as the stability of the biochemical compounds used for nebulisation. Within the nebulisation process in air-jet nebulisers protein structure and function can be compromised independently from the molecular weight of the protein by surface denaturation, shearstress induced denaturation as well as desiccation of the aerosol droplets (Niven 1995). In principle, protein stability can be increased by several additives like lipids, surfactant, amino acids, albumin, polyols as well as packing into liposomes (Niven 1995; Ko¨hler and Fleischer 2000). For example, liposomes, in form of nanoparticles, are used for stabilisation of low molecular weight compounds (e.g. beclomethasone and b-agonists) as well as peptides and proteins (e.g. antioxidative enzymes, cyclosporine A, interleukin-2 (IL-2)) (Scheuch et al. 2006; Niven 1995; Ko¨hler and Fleischer 2000). Ultrasonic nebulisers act by a disruption of liquid surfaces by means of ultrasound and allow a production of high concentration aerosols (Ko¨hler and Fleischer 2000). Usually, aerosol particles are produced with this type of nebulisers which are not appropriate for deep lung delivery.

13 Targeting Drugs to the Lungs – The Example of Insulin

229

Vibrating mesh nebulisers produce a liquid aerosol by means of a vibrating mesh or plate with multiple apertures. Devices of this type allow the generation of aerosols with a high fine-particle fraction. The aerosols are generated as a fine mist without requirement of an internal baffling system (Dhand 2004; Ko¨hler and Fleischer 2000). Compared to conventional jet nebulisers and ultrasonic nebulisers they have a higher efficiency for the delivery of drugs to the respiratory tract. Some other advantages are that these devices effectively aerosolise solutions, have only a minimal residual volume of medication left in the device (cost economic effect) and might be breath-actuated, thereby limiting the release of aerosolised drug into the environment (Dhand 2004). However, they sometimes fail when liposomal formulations should be aerosolised and usually it is difficult to aerosolise suspensions (exception: nano-suspensions).

13.2.2

Dry Powder Inhalers (DPI)

In DPI aerosols are produced by desaggregation of preformed (e.g. milled or spray-dried) micronised particles. The energy required for desaggregation is supplied by the inhalation manoeuvre or alternatively by means of an external energy (Niven 1995; Telko and Hickey 2005). Advantages of DPI are their environmental sustainability (due to a propellant-free design) and the ease to use (no much patient coordination needed). Typical disadvantages are the dependency of the deposition efficiency on the patient’s inspiratory airflow, their potential for dose uniformity problems and their relative high complexity and costs for development and manufacture. Aerosols produced by DPI are established for treatment of asthma and chronic obstructive pulmonary disease (COPD), e.g. by means of b-mimetics, anticholinergics or steroids whereas there is up to now only little experience on inhalative administration of biomolecules except insulin (Exubera1) for systemic treatment (Agu et al. 2001; Siekmeier and Scheuch 2005; Telko and Hickey 2005). This is caused by specific problems for the use of proteins or peptides occurring in the processes of lyophilisation or spray drying, micronisation, completeness of dispersion and desaggregation as well as the surveillance of the latter. The inspiratory air flow of the patient is an essential parameter in passive systems. If it is not sufficient for complete desaggregation large aggregates are inhaled which cannot reach the alveolar region. On the other hand, a high air flow rate increases oropharyngeal deposition also followed by a reduction of pulmonary aerosol deposition. Humidity can also be a large problem because it impairs the stability of proteins and peptides and also affects desaggregation and dispersion (Ko¨hler and Fleischer 2000; Irngartinger et al. 2004; Niven 1995; Telko and Hickey 2005). However, if the underlying problems especially in particle engineering are solved by novel techniques (Shoyele and Cawthorne 2006) dry powders, which dissolve fast might also be considered as carriers for nanoparticles.

S. Ha¨ussermann et al.

230

13.2.3

Metered Dose Inhalers (MDI)

In MDI compounds are dissolved or suspended in a pressurised propellant which requires to be nontoxic, noninflammable, compatible with drugs formulated as suspensions or solutions and to have appropriate boiling points and densities. For consistent dosing the vapour pressure must remain constant throughout the product’s life. These requirements are typically fulfilled by chlorofluorocarbons (e.g. dichlorodifluoromethane, dichlorotetrafluoroethane and trichlorofluoromethane, which are phased out and replaced by hydrofluoralcans (HFAs) because of environmental concerns (ozone hole)) but not by pressurised carbon dioxide. After its release with high velocity the mixture rapidly expands forming an aerosol. Because of the high velocity of the aerosol directly after its release different types of spacers are often required for optimisation of the aerosol deposition (Newman 2005; Niven 1995). Aerosols from metered dose inhalers are established in clinical treatment of patients with asthma or COPD for about 50 years and many different types of metered dose inhalers have been developed (Ko¨hler and Fleischer 2000; Newman 2005; Niven 1995). Unfortunately, these devices up to now cannot be used for treatment with macromolecules (e.g. peptides and proteins) because a number of prerequisites (stability of the compound within storage in the inhaler, no denaturation of the compound within the nebulisation process, production of an aerosol with appropriate particle distribution pattern) are not sufficiently fulfilled.

13.3

History of Insulin Inhalation

In 1924 and 1925 – i.e. only 2 years after the start of the therapeutic insulin era – the first studies on insulin inhalation were published. Laqueur and Grevenstuk published their investigation on intratracheal administration of insulin in 1924 and reported a more rapid onset of action after intratracheal administration compared to subcutaneous administration (Laqueur and Grevenstuk 1924). A first study on inhalation of insulin in patients was performed by Heubner et al. also in 1924. These investigators reported a dose-dependent effect of insulin inhalation on blood glucose. However, Heubner et al. required a 30-times higher dose for inhalation than for subcutaneous administration and assumed the problem in the requirement of the high amounts of insulin even though they also emphasised the advantage of this type of administration for the patients (Heubner et al. 1924). At the same time and independently from the investigations of Heubner et al. Ga¨nsslen performed also investigations in patients. Ga¨nsslen reported that the inhalation of insulin was well tolerated and caused a significant decrease of the blood glucose concentration and that the amounts of insulin required for inhalation in relation to subcutaneous application were not as high as described by Heubner et al. (Ga¨nsslen 1925; Heubner et al. 1924). However, because of the large number of unsolved problems it required 46 more years until Wigley et al. published their pivotal study of insulin

13 Targeting Drugs to the Lungs – The Example of Insulin

231

inhalation offering the proof of principle of this therapy (Cefalu 2004; Wigley et al. 1971). Wigley et al. investigated three subjects without diabetes mellitus and four patients with diabetes and they were able to demonstrate that pork-beef insulin administered by a nebuliser caused a prompt increase in plasma immunoreactive insulin and that hypoglycemia showed a temporal relationship with the increase in plasma immunoreactive insulin (Wigley et al. 1971). However, even after the investigation of Wigley et al. inhalative insulin therapy was far away from its introduction into clinical therapy and in the next two decades several studies established the basics of insulin inhalation (Elliott et al. 1987, Ko¨hler et al. 1987; Ko¨hler and Fleischer 2000; Laube et al. 1993). In these years it was observed that the bioavailability of inhaled insulin in case of improved application procedures was only about 20% to 25% of that after subcutaneous administration, but also that inhalation might be an important alternative administration route (Cefalu 2004; Ko¨hler and Fleischer 2000; Niven 1995). However, the methods under investigation were not able to administer sufficient drug doses in a reproducible way, because their particle spectrum was optimised for aerosol deposition in the bronchial system and not in the alveoli (Cefalu 2004; Gonda 2000). In the following years several companies developed devices for inhalative administration of insulin, which are very different in respect to the technical and pharmacological principles (e.g. manual or semi-automated systems for inhalation, powder aerosol or liquid aerosol) (Cefalu 2004). The most advanced method was Exubera1 from Pfizer/Nektar, which received the approval from the American and European Drug Agencies (FDA and EMEA, respectively) in early 2006 for patients with diabetes mellitus types 1 and 2 and was marketed since September 2006. Some others in later phase of development were AERx1 iDMS (Aradigm/Novo Nordisk), HIIP1 (Alkermes/Eli Lilly) and Technosphere1 (Pharmaceutical Discovery Corporation/Mannkind Pharmaceuticals). Exubera1 was based on a recombinant human insulin which was spray-dried and supplemented with the excipients mannitol, glycine and sodium citrate. The insulin content of the final product, a large low-density particle, packed into small blisters was 60% (Guntur and Dhand 2007). However, in October 2007 Pfizer announced it would be dropping Exubera1, citing that the drug had failed to gain market acceptance. After stop of marketing for Exubera1 and a recent press release of the American Drug Agency reporting a potentially increased risk for bronchial carcinoma in ex-smokers treated with inhalative insulin (FDA MedWatch Alert 2008) all manufacturers except Mannkind Pharmaceuticals stopped their development in this field.

13.3.1

Devices for Insulin Inhalation

In the last decade a number of very different systems for insulin inhalation have been developed often by cooperating pharmaceutical companies. The devices differ strongly in respect to the underlying inhalation technology and allow the inhalation of insulin powder and liquid insulin. In 2006 a number of methods was in the late

S. Ha¨ussermann et al.

232

phase of development. However, the only system introduced into the market was Exubera1 (see Table 13.1).

13.3.2

Pharmacokinetics of Inhaled Insulin

Many studies investigated pharmacokinetics of insulin in healthy subjects as well as patients with diabetes mellitus types 1 and 2. Unfortunately, the comparison of the obtained data is hampered by differences of the used inhalers, administered formulations and doses of insulin, small numbers of included individuals (healthy individuals or patients), inappropriately used pharmacological models and distinct determined parameters (Patton et al. 2004; Sakagami 2004). However, it was observed that inhaled regular insulin is absorbed at least as fast as subcutaneously administered insulin (time to peak concentration in plasma (tmax): 7–90 min vs 42–274 min (see Table 13.2) (Cefalu 2004; Ko¨hler and Fleischer 2000; Niven 1995; Patton et al. 2004; Sakagami 2004) with a biphasic pharmacokinetics (first peak rapidly after inhalation followed by a slow release comparable to that after subcutaneous injection) which suggests that inhaled insulin might have some therapeutic benefit in the treatment of prandial or postprandial hyperglycemia when compared to conventionally administered insulin (Farr et al. 2000; Patton et al. 2004). In the first 60 min after drug administration the area under the concentration vs time curve (AUC) is larger for inhaled insulin than for subcutaneously administered insulin whereas the opposite takes place if an observation period of 6 h is considered (Patton et al. 2004). In most studies the bioavailability was calculated by comparison of the AUC after inhalation to that after subcutaneous administration of insulin. In contrast, the bioeffectivity describes the hypoglycemic effect of inhaled insulin compared to a defined insulin dose administered by subcutaneous injection (Patton et al. 2004). Commercially available systems for pulmonary administration of insulin are characterised by bioavailabilities and bioeffectivities of 9–22% and 8–16%, respectively (see Table 13.2). In consequence, the insulin dose which is required to achieve the same therapeutic effect after inhalation is up to 11-times higher compared to subcutaneous administration (Patton et al. 2004). Between 50% and 80% of the insulin filled in the inhalation system does not reach the lung, but is remaining in the nebuliser, is deposited in the mouth or the oropharynx or is expired. Taking this into account, the bioavailability from the lung deposited fraction is about two to five times of the subcutaneously given insulin dose. However, from this dose more than 50% are deposited in the airways (bronchial system) and are removed from the lung by the mucociliary transport and/or degradation. Only about 40% are rapidly absorbed into the circulatory system. If this is also considered, it is obvious that the “pulmonary extradose” for insulin inhalation is two to three times of the dose required for injection (Patton et al. 2004) which strongly affects the price of the inhalative therapy. Even though the price is quite high, prescribing doctors see the need for inhaled insulin, since they expect a higher compliance of treated patients.

AERx1 iDMS (Aradigm/Novo Nordisk)

Phase III

Dosing via the number of blisters Pneumatic release of the aerosol out of the blister in an inhalation chamber Particle diameter of <5 mm Liquid insulin packed into single strips and dosed in single units Regulation of the breathing manoeuvre by means of microprocessors and electronic optimisation of insulin release within the inspiratory flow Particle diameter of 1–3 mm

Table 13.1 Devices for inhalative administration of insulin Trade name (developer/ Status of development Principle or pharmaceutical form partner) in 2006 Market approval 2006 Dry powder insulin packed into blisters of Exubera1 (Nektar Therapeutics/Pfizer) 1 mg (3 U insulin) or 3 mg

More rapid absorption and onset of action than subcutaneously administered insulin Variability of pharmacodynamic parameters in patients with diabetes mellitus type 1 similar to those after subcutaneous administration of insulin Similar quality of metabolic adjustment in patients with diabetes mellitus type 2 and subcutaneous administration of insulin (continued)

Normal short acting insulin More rapid onset of action than subcutaneously administered insulin or lispro Proven reproducibility for all pharmacokinetic and pharmacodynamic parameters Applicability for monotherapy or combined therapy in patients with diabetes mellitus type 2 insufficiently treated with oral antidiabetics Similar quality of metabolic adjustment (HbA1c) in patients with diabetes mellitus types 1 and 2 and subcutaneous administration of insulin Bioavailability: 10–16% Most data regarding long time efficiency and safety available

Selected clinical data

13 Targeting Drugs to the Lungs – The Example of Insulin 233

Microdose DPI1 (Microdose Technologies/Elan Corporation)

Technosphere1 (Pharmaceutical Discovery Corporation/ Mannkind Pharmaceuticals)

HIIP1 (Alkermes/Eli Lilly)

Table 13.1 (continued) Trade name (developer/ partner)

Phase II

Phase III

Phase III

Status of development in 2006

Dry powder recombinant insulin combined with a derivative of diketopiperazine for absorption enhancement Self-assembly into an ordered lattice array at low pH-value; mass median aerodynamic diameter 2–4 mm; dissolution of particles and insulin release at neutral pH-value on alveolar surface Formulation developed for administration by means of a dry powder inhaler and passive desagglomeration (MedTone1) Dry powder insulin packed into blisters Desaggregation of drug powder by means of a piezo vibrator Mass median aerodynamic diameter approximately 1.5 mm, 84% of the particles <4.7 mm

Mechanical system with breath activated release of particles Porous particles of low density with a geometric diameter of 5–30 mm (aerodynamic diameter of <5 mm)

Dry powder insulin packed into blisters

Principle or pharmaceutical form

Few clinical data; well tolerated in clinical studies, bioavailability about 18% compared to subcutaneous administration; more rapid absorption than subcutaneously administered insulin

Bioavailability: 10–16% Studies in patients with diabetes mellitus types 1 and 2 Development of a formulation with rapid release (pharmacokinetic similar to Humulin R) and a formulation with sustained release (pharmacokinetic similar to Humulin L) Bioavailability: 10–16% Studies in healthy individuals and patients with diabetes mellitus types 1 and 2 Low interindividual variability of the therapeutic effect of insulin Rapid absorption and onset of action as well as the short duration of the therapeutic effect indicate a potential use for treatment of postprandial hyperglycemia Bioavailability: 16–46%

Selected clinical data

234 S. Ha¨ussermann et al.

Unknown (Epic Therapeutics/-)

Alveair1 (CoreMed/ Fosun and Xuzhou)

Bio-Air1 (BioSante Pharmaceuticals/-)

Aerodose1 (Aerogen/-)

Unknown (Abbott (formerly Kos Pharmaceuticals)/-)

Phase I

Phase I

Phase I

Phase II (stopped in 2003)

Phase II

Coated dry particles based on calcium phosphate nanoparticle carriers Administration by means of a calcium phosphate nanoparticulate delivery system Liquid insulin Administration by means of a generic handheld device delivering inhaled insulin with the same units as injected insulin Microspheres of recombinant human insulin (PROtein MAtriX microspheres; ProMaxx1) Administration by means of a dry powder inhalation device (Cyclohaler) >90% insulin

Dry crystals of a recombinant insulin formulation Administration by means of a handheld breath actuated inhaler (BAI) driven by a propellant Liquid insulin Administration by means of a breath activated multiple dose inhaler Mean particle diameter of 3.2 mm, 87% of the particles between 1 and 6 mm

(continued)

Few clinical data; well tolerated in a phase I trial, bioavailability >12% compared to subcutaneous administration; more rapid absorption than subcutaneously administered insulin

Few clinical data; the manufacturer states a very high level of bioavailability of the compound

More rapid onset of action than subcutaneously administered insulin Reproducibility of pharmacokinetic parameters similar to that after subcutaneous administration Linear dose-response-relationship of pharmacodynamic parameters Important effect of the nebulisation time on the biological properties Bioavailability: 10–22% Few data available; preclinical studies demonstrated an extension of the hypoglycemic effect after insulin inhalation

Comparison to Lantus1 (insulin glargine) revealed a comparable effectiveness in controlling blood glucose concentrations

13 Targeting Drugs to the Lungs – The Example of Insulin 235

Phase I (stopped in 2004)

Status of development in 2006 95% of the microspheres with mass median aerodynamic diameters 0.95– 2.1 mm (mean: 1.5 mm), 95% of the particles <4.7 mm Dry powder insulin packed into blisters

Principle or pharmaceutical form

Small number of studies only in healthy individuals Administered doses consistent over a wide range of inspiratory flow rates

Selected clinical data

Release by means of a hand held battery driven multiple dose inhalator Development of a novel powder dispersion system (Spiros-S2) without the electromechanical components of the Spiros for administration at low inspiratory flow rates (15–30 L/min) Source: Modified according to Cefalu (2004), Charlish (2006), de Galan et al. (2006), Laube (2005), Mastrandrea and Quattrin (2006), Owens (2002), Pearson (2006), Patton et al. (2004).

Spiros1 (Elan Pharmaceuticals (formerly Dura Pharmaceuticals)/-)

Table 13.1 (continued) Trade name (developer/ partner)

236 S. Ha¨ussermann et al.

6

12

Spiros1

Spiros1

5

12

Aerodose1

Inhalator M1 (Dry powder inhaler)

17

n

Healthy and non-diabetic individuals AERx1 iDMS

Subjects and device (type of insulin)

s.c.

s.c. s.c. s.c.

s.c.

s.c.

Route of comparator

Regular 10 U

Regular 2 blisters Regular 3 blisters Regular 4 blisters Regular 5 blisters Regular 8 U Regular 14 U Regular 20 U Technosphere 100 U

Regular 2.31 mg (emitted)

Regular 0.15 U/kg

Regular 1.5 U/kg

Regular 2 U AERx1 h Regular 4 U AERx1 Regular 6 U AERx1 Regular 8 U AERx1 Regular 6 U

Type of insulin and dose

121

30 20 45 20 120 105 120 13

30

85

50

Serum insulin tmaxa(min)

204 pmol/L

24 mU/L 31 mU/L 38 mU/L 46 mU/L 31 mU/L 34 mU/L 59 mU/L 2225 pmol/L

1.13 mg/L

364 pmol/L

360 pmol/L

33.3 mU/L

47.1 mU/L

40.3 mU/L

33.1 mU/L

29.6 mU/L

Serum insulin Cmaxb

96.4 nmol min/L (0–3 h) 104.2 nmol min/L (0–6 h) 41.6 nmol min/L (0–3 h) 71.5 nmol min/L (0–6 h)

168 mU h/L (0–10 h) 162 mU h/L (0–10 h) 210 mU h/L (0–10 h) 236 mU h/L (0-10 h) 204 mU h/L (0–10 h) 48900 pmol min/L (0-6 h) 56400 pmol min/L (0–6 h)

Serum insulin AUCc

163

187 129 161 162 227 241 241 39

10.6

2.0 3.3 3.4 4.2 4.1 5.1 6.1 16.7

235 mg/ min

2590

7.75

112

3130

2600

2310

2170

GIRmaxe (mg/kg/ f min)

8.7

8.12

6.9

6.35

GIRd tmax(min)

Table 13.2 Selected studies on pharmacokinetics and pharmacodynamics of inhaled insulin

1150 (0–3 h) 2170 (0–6 h)

1940 (0–3 h) 2940 (0–6 h)

1520 (0–6 h) 1750 (0–6 h)

GIR AUCg (mg/kg)

Bioavailability 26% (0–3 h), 16% (0–6 h) (relative); 15% (0–3 h), 16% (0–6 h) (absolute); bioefficacy 19% (0–3 h), 14% (0–6 h) (relative);

Not dependent on the inhalation technique

Bioavailability 9.3%, bioefficacy 10.3%

Bioavailability and bioefficacy of inhaled insulin

(continued)

Steiner et al. (2000), Steiner et al. (2002)

Rave et al. (2001)

Chien et al. (2001)

Fishman, et al. (2000)

Heise et al. (2001)

References

13 Targeting Drugs to the Lungs – The Example of Insulin 237

11

18

15

24

Type 1 diabetes mellitus AERx1 iDMS

Type 2 diabetes mellitus Aerodose1

Aerodose1

n

Turbuhaler1 DPI

Subjects and device (type of insulin)

Table 13.2 (continued)

Regular 5 U

i.v.

s.c.

s.c.

65

Regular 1.8 U/kg

60 97 73

Regular 160 U

Regular 240 U

193

76

Regular 80 U

Regular 24 U

Regular 240 U

119

62

Regular 1.2 U/kg

Actrapid1 0.12 U/ kg

48

Regular 0.6 U/kg

1

49

Regular 5 U

i.v.

106

24

5

Serum insulin tmaxa(min)

Regular 0.3 U/kg

Regular 10 U

s.c.

Regular 99 U

Type of insulin and dose

Route of comparator

47 mU/L

96 mU/L

35.1 mU/L

77.4 mU/L

53.8 mU/L

32.9 mU/L

23.4 mU/L

5692 pmol/L

254 pmol/L

305 pmol/L

3692 pmol/L

Serum insulin Cmaxb

12 mU min/mL (0–3 h) 22 mU min/mL (0–8 h) 4.9 mU min/mL (0–3 h) 14 mU min/mL (0–8 h) 7.6 mU min/mL (0–8 h) 17.1 mU min/mL (0–8 h) 26.2 mU min/mL (0–8 h)

79 mU h/L (0–10 h) 122 mU h/L (0–10 h) 200 mU h/L (0–10 h) 315 mU h/L (0–10 h) 184 mU h/L (0–10 h)

35.2 nmol min/L (0–3 h) 36.0 nmol min/L (0–6 h) 27.0 nmol min/L (0–6 h) 37.1 nmol min/L (0–6 h) 25.8 nmol min/L (0–3 h)

Serum insulin AUCc

237

173

167

244

170

189

218

157

136

94

14

147

108

14

GIRd tmax(min)

4.3

5.5

3.2

6.5

4.7

2.5

1.6

17.6

9.1

6.2

18.9

GIRmaxe (mg/kg/ f min)

2400 (0–8 h)

1700 (0–8 h)

1100 (0–8 h)

421 (0–3 h) 1333 (0–8 h)

687 (0–3 h) 1684 (0–8 h)

765 (0–10 h)

1695 (0–10 h)

1029 (0–10 h)

452 (0–10 h)

165 (0–10 h)

830 (0–6)

1900 (0–6 h)

1440 (0–6 h)

1330 (0–3 h) 1700 (0–6 h)

GIR AUCg (mg/kg)

Bioavailability 22%, bioefficacy 16%

Bioavailability 18%, bioefficacy 13%

Bioavailability 16% (0– 8 h), bioefficacy 13% (0–8 h)

Bioavailability 12.9% (0–6 h), bioefficacy 12.7%

Bioavailability 7.8% (relative), 5.6% (absolute); bioefficacy 7.6% (relative), 9.5% (absolute)

8% (0–3 h), 10% (0–6 h) (absolute)

Bioavailability and bioefficacy of inhaled insulin

Kim et al. (2002)

Perera et al. (2002)

Brunner et al. (2001)

Heinemann et al. (1997)

References

238 S. Ha¨ussermann et al.

12

s.c.

Regular 24 U

s.c.

Technosphere1 50 U Technosphere1 100 U Regular 10 U

Technosphere1 25 U

Regular 16 U

s.c.

153

192

56

21

47

349

347

366

52

4.5 mU min/mL (0–8 h) 8.1 mU min/mL (0–8 h) 12.6 mU min/mL (0–8 h)

18

12

274

223

225

Source: According to Patton et al. (2004). a tmax: Time to reach Cmax. b Cmax: Maximum concentration of insulin in serum. c AUC: Area under the serum insulin concentration-time curve (between specified limits). d GIR: Glucose infusion rate. e GIRmax: Maximum GIR (peak). f If not stated otherwise. g GIR AUC: Area under the GIR-time curve (between specified time limits). h 1 U of AERx1  10 regular units.

DPI

Regular 8 U

s.c.

1500 (0–8 h)

1100 (0–8 h)

900 (0–8 h)

Bioavailability 28% (0–3 h) Bioavailability 42% (0–3 h) Bioavailability 46% (0–3 h)

Bioavailability 22%, bioefficacy 16%

Rave et al. (2000)

13 Targeting Drugs to the Lungs – The Example of Insulin 239

S. Ha¨ussermann et al.

240

The target group are elderly type 2 diabetes patients, who often start too late with insulin therapy with the consequence that the follow up costs outweigh the treatment cost.

13.3.3

Factors Affecting the Pharmacokinetics of Inhaled Insulin

The pulmonary deposition is mainly influenced by biological and physical parameters of the substance, the nebuliser, the breathing manoeuvre and the oropharyngeal filter efficiency of the patient (Owens 2002; Patton et al. 2004; Sakagami 2004; Scheuch et al. 2006; Scheuch and Siekmeier 2007). An optimal deposition of the inhaled insulin is achieved if the aerosol is released at the beginning of a slow and deep inhalation manoeuvre. This enables the particles to penetrate deeply into the lung and they can be deposited in the alveolar region (Farr et al. 2000; Owens 2002; Scheuch et al. 2006; Scheuch and Siekmeier 2007). Farr et al. observed in their study a later and weaker effect of insulin administered by a shallow inhalation manoeuvre (40% of inspiratory vital capacity (IVC)) than after a deep inspiration manoeuvre (80% IVC) (Farr et al. 2000; Owens 2002). The reproducibility of insulin pharmacokinetics which is influenced by a number of factors (e.g. breathing manoeuvre, smoking, physical stress and lung perfusion) was investigated in several studies. In summary, these demonstrated a similar or even better reproducibility of insulin administrated by inhalation when compared to subcutaneous injection (Patton et al. 2004) which can be explained by the missing of some parameters (e.g. physical exercise, smoking, temperature, body position and injection) affecting absorption after subcutaneous injection. In detail, the observed coefficients of variation were between 13.7% and 23.0% in studies performed with different types of inhaled insulin in patients with diabetes mellitus types 1 and 2 (Patton et al. 2004). The absorption of inhaled insulin is up to three to five times higher in smokers than in nonsmokers (see Table 13.3) (Cefalu 2004; Himmelmann et al. 2003; Ko¨hler 1990; Ko¨hler and Fleischer 2000; Ko¨hler et al. 1987; Patton et al. 2004, Sakagami 2004). For example, Ko¨hler et al. reported a higher absorption (Cmax) and bioavailability (65% vs 25%) of inhaled insulin which was accompanied by a more pronounced decrease of the glucose concentration in smokers compared to nonsmokers (Ko¨hler et al. 1987; Sakagami 2004). In another study Himmelmann et al. reported a higher absorption of inhaled insulin (AUC; 63.2 mU h/l vs 40.0 mU h/l, p < 0.01), a higher peak concentration (Cmax; 42.0 mU/L vs 13.9 mU/L, p < 0.001) and a shorter time to peak (tmax; 31.5 min vs 53.9 min, p < 0.001) in smokers compared to nonsmokers. In addition, the mean residence time (MRT) in smokers was less than half of that in the nonsmoker group (p < 0.0001) and accordingly the apparent elimination rate constant of exogenous insulin was almost twice as high in smokers compared to nonsmokers (p = 0.0019). However, the intraindividual variability was similar in both groups (see Table 13.3) (Himmelmann et al.

13 (NS) 23 (S)

Non-diabetic nonsmokers (NS) vs smokers (S) AERx1 iDMS Nonsmokers (NS) vs smokers (S) Exubera1

Non-diabetic individuals with upper respiratory tract infection (URTI) vs postupper respiratory tract infection (PU) AERx1 iDMS Non-asthmatic (NA) vs asthmatic (A) nondiabetic individuals AERx1 iDMS

30 (NS) 38 (S)

Nonsmokers (NS) vs smokers (S) Exubera1

Regular 45 U

Regular 45 U

Regular 45 U

20 (no URTI)

28 (NA)

16 (A)

30.0

Regular 1 mg

45

50

80

59

37.5

Regular 1 mg

Regular 45 U

20.0

Regular 1 mg

8310 pmol/L/ kg

9872 pmol/L/ kg

17.3 mU/L

15.4 mU/L

29.2 mU/mL

15.8 mU/mL

26.8 mU/mL

9.7 mU/mL

42.0 mU/l

31.5 52.5

13.9 mU/L

72 mU/L

16 mU/L

Serum insulin Cmaxb

53.9

31

Regular 2 mg

Regular 33.8 U Regular 33.8 U Regular 1 mg

53

Serum insulin tmaxa (min)

Type of insulin and dose Regular 2 mg

20 (URTI)

10 (NS) 20 (S)d) 20 (S)e) 20 (S)f)

n

Subjects/System

1.45  106 pmol min/L kg (0–6 h) 1.07  106 pmol min/L kg (0–6 h)

44.0 mU h/L (0–6 h) 47.7 mU h/L (0–6 h)

Partially reversible on smoking cessation

1410 mU min/mL (0–6 h) 4847 mU min/mL (0–6h) 40.0 mU h/L (0–6 h) 63.2 mU h/l (0–6 h) 1645 mU min/mL (0–6 h) 2583 mU min/mL (0–6 h) 1887 mU min/mL (0–6 h) 3156 mU min/mL (0–6 h)

Significant difference in AUC

Significant differences in tmax, Cmax and AUC between NS and S, values of S approaching to those of NS after smoking cessation, smoking resumption reversed the effect of smoking cessation Significant difference of tmax, but no effect of URTI on the other parameters of pharmacokinetics

Significant differences in tmax, Cmax and AUC

Additional comments

Serum insulin AUCc

(continued)

Henry et al. (2003a)

McElduff et al. (2005)

Becker et al. (2006)

Himmelmann et al. (2003)

Sha et al. (2002)

References

Table 13.3 Studies investigating factors influencing the pharmacokinetics of regular human insulin inhaled by devices developed for insulin inhalation

13 Targeting Drugs to the Lungs – The Example of Insulin 241

Regular 45 U

27 (Y)

28 (O)

Younger (Y, 18-45 years) vs older (O, >65 years) patients with type 2 diabetes AERx1 iDMS 30

40

Serum insulin tmaxa (min)

221 pmol/l

219 pmol/L

Serum insulin Cmaxb

Source: According to Patton et al. (2004). a tmax: Time to reach Cmax. b Cmax: Maximum concentration of insulin in serum. c AUC: Area under the serum insulin concentration-time curve (between specified limits). d Before smoking cessation. e 7 days after smoking cessation. f 9–10 days after smoking resumption.

Type of insulin and dose Regular 45 U

n

Subjects/System

Table 13.3 (continued)

38055 pmol min/ L (0–6 h) 37892 pmol min/ L (0–6 h)

Serum insulin AUCc Comparable pharmacokinetic profiles in both groups, but lower glucose lowering effect in elderly individuals

Additional comments

Henry et al. (2003b)

References

242 S. Ha¨ussermann et al.

13 Targeting Drugs to the Lungs – The Example of Insulin

243

2003). In another study Becker et al. investigated the effect of smoking cessation and subsequent resumption on the absorption of inhaled insulin. It was found that AUC and Cmax were higher in smokers than in nonsmokers whereas tmax was shorter. Smoking cessation resulted in a rapid change of the values obtained in smokers towards those of nonsmokers. In contrast, smoking resumption completely reversed the effect of smoking cessation (see Table 13.3) (Becker et al. 2006). However, the effects of acute cigarette smoke inhalation on the absorption of inhaled insulin are largely different from that of chronic cigarette consumption as cigarette consumption just before insulin inhalation significantly blunts the enhanced insulin absorption in smokers. (Himmelmann et al. 2003). Last not least it should be noted that even acute passive cigarette smoke exposure may affect the pharmacokinetics of inhaled insulin resulting in a modest decrease of the bioavailability (Fountaine et al. 2008). Because of the effects of smoking, inhalative insulin therapy was not approved in current smokers and individuals who quitted smoking less than 6 months before therapy until marketing was stopped by the manufacturers. Only a minor number of studies addressed the effect of pulmonary diseases on the pharmacokinetics of inhaled insulin. For example, upper respiratory tract infections had no effect (McElduff et al. 2005; Patton et al. 2004), whereas asthma bronchiale resulted in a mild decrease of Cmax, and a distinct decrease of AUC (bioavailability) and plasma glucose concentration (bioeffectivity) after insulin inhalation in asthma patients compared to healthy individuals. Furthermore, patients with asthma showed a higher variability of Cmax and AUC (see Table 13.3) (Henry et al. 2003a; Patton et al. 2004). Presumably, these undesirable effects can be improved by administration of bronchodilators in these patients (Mudaliar and Henry 2007). However, there are two primary concerns regarding the inhalation of insulin in asthma patients. Firstly, drug inhalation especially by means of DPI can induce bronchospasm. Secondly, in asthma exacerbation respiratory effort and bronchospasm limit the deposition of inhaled insulin in the lung alveoli due to a variation of pulmonary convective gas transport, a smaller airway diameter and in consequence the higher rate of particle deposition in the central airways of these patients. Data regarding the pharmacokinetics of inhaled insulin in patients with COPD are limited and conflicting as COPD patients demonstrated a variable (higher or lower) absorption of insulin compared to subjects without COPD and it is not clear whether this variability is secondary to differences in inhalation devices or different study populations (Mudaliar and Henry 2007; Rave et al. 2007). Elder individuals show a decrease of the alveolar surface, a variation of lung elasticity, a decrease of the alveolar capillary volume combined with a decline of the ventilation/perfusion ratio, a decrease of the pulmonary diffusion capacity for carbon monoxide (DLCO) as well as an increase of the pulmonary residual volume (RV) (Belmin and Valensi 2003). Therefore, the age is another important parameter influencing the pharmacokinetics of inhaled insulin. Henry et al. reported similar values of Cmax and AUC in patients with diabetes mellitus type 2 aged >65 years and young individuals of an age between 18 and 45 years. The variability of these parameters showed also no differences between both study groups. However, the observed decrease of plasma glucose concentrations was more pronounced in

S. Ha¨ussermann et al.

244

younger individuals than in elder patients indicating a requirement of higher doses in aged patients (Henry et al. 2003b; Patton et al. 2004).

13.3.4

Safety of Inhaled Insulin

Several studies investigated the safety of inhaled insulin focussing on lung function, local inflammation, allergic reaction, formation of antibodies against insulin, pulmonary fibrosis as well as lipodystrophy (Cefalu 2004; Patton et al. 2004). Most data regarding the long-term tolerability are published for the Exubera1 system and the AERx iDMS1 system for study periods of up to 2 years and more in patients with diabetes mellitus types 1 and 2 (Barnett et al. 2006a, b, 2007; Ceglia et al. 2006; DeFronzo et al. 2005; Laube 2005; Owens 2002; Patton et al. 2004; Rosenstock et al. 2005; Valente et al. 2003). Diabetes and consecutive insulin treatment is associated with morphological changes of lung structure (e.g. thickening of the alveolar membrane and the capillary basal lamina, vascular hyalinosis, granulomas, intraseptal nodular fibrosis and emphysema-like septal obliteration) depending on duration and severeness of the disease as well as additional factors like smoking and often not noticed by the patients (Black et al. 2007; Davis and Davis 2007). In most studies inhalation of insulin caused no changes of spirometric parameters of lung function (e.g. forced expiratory volume in 1 s (FEV1), forced vital capacity (FVC)) as well as parameters of diffusion capacity for carbon monoxide (DLCO) and blood gas analysis (Barnett et al. 2006a, b, 2007; Cefalu 2004; Ceglia et al. 2006; DeFronzo et al. 2005; Hollander et al. 2004; Laube 2005; Owens 2002; Patton et al. 2004; Quattrin et al. 2004; Rosenstock et al. 2005; Skyler et al. 2001, 2005; Valente et al. 2003). However, based on the results of the lung function tests and the observed variations in individual patients the manufacturer recommended spirometric measurement of lung function before treatment, after 6 months and thereafter at least annually in the product information for Exubera1 (NDA 21-868/EXUBERA: U.S. package insert 2006). Beside its strong metabolic effect insulin also acts as a weak growth factor (effectivity only 1/100 of insulin-like growth factor-1(IGF-1)) after binding to the receptor for IGF-1. Up to now there is no evidence for a relevant competitive effect of inhaled insulin at the IGF-1 receptors in the lung (Patton et al. 2004). However, about 6 months after end of the marketing of Exubera1 the American Food and Drug Administration (FDA) published a press release reporting a potentially increased risk for bronchial carcinoma in ex-smokers treated with inhalative insulin (FDA MedWatch Alert 2008). Many studies observed increased serum titres of non-neutralising IgG antibodies against insulin in patients treated with inhaled insulin. However, the development of these antibodies had no therapeutic relevance, i.e. there were no correlations to the metabolic control, lung function and adverse events (Barnett et al. 2006a, b; Belmin and Valensi 2003; Cefalu 2004; Ceglia et al. 2006; DeFronzo et al. 2005;

13 Targeting Drugs to the Lungs – The Example of Insulin

245

Hollander et al. 2004; Laube 2005; Owens 2002; Patton et al. 2004; Quattrin et al. 2004; Rosenstock et al. 2005; Skyler et al. 2005; Valente et al. 2003). Lipodystrophy is a phenomenon observed in up to 30% of patients treated with subcutaneous injections of insulin developing at the site of injection and includes lipohypertrophy (caused by the anabolic effect of insulin promoting the synthesis of protein and fat) and lipoatrophy (caused by an inflammatory process) (Radermecker et al. 2007). However, even though adipocytes are also located in the lung up to now it is not known if and how inhaled insulin affects pulmonary adipocytes (Ghosh and Collier 2007). Cough is a typical symptom in clinical treatment with inhalation of dry powder aerosols which might affect patient convenience and compliance and therefore was addressed in a number of studies on inhaled insulin. Mild to moderate cough was reported to occur rapidly after inhalation (seconds to minutes) in up to 20% to 30% of patients. However, the reported symptoms were transient, settled with continuation of the therapy and seldom resulted in treatment withdrawal (Barnett et al. 2006a, b; DeFronzo et al. 2005; Hollander et al. 2004; Quattrin et al. 2004; Rosenstock et al. 2005; Skyler et al. 2001, 2005). Hypoglycemia is a common problem in patients treated with antidiabetics, especially insulin. Even though study data in this field are conflicting there seems to be no or only little difference regarding the risk for th e occurrence of hypoglycemia between inhaled and subcutaneous insulin (Royle et al. 2004), whereas the risk is expectedly higher for patients treated with inhaled insulin when compared to treatment with oral antidiabetics (Barnett et al. 2006a, b; DeFronzo et al. 2005; Rosenstock et al. 2005).

13.4

Conclusions

In the last years there were large advances in pharmacology and inhalation technique regarding the inhalation of macromolecules for systemic treatment. Feasibility and safety of pulmonary administration have been shown for a number of drugs and biomolecules (Agu et al. 2001; Byron and Patton 1994; Ko¨hler, 1994; Ko¨hler and Fleischer 2000; Niven 1995; Scheuch et al. 2007; Siekmeier and Scheuch 2005; Valente et al. 2003; Wolff 1998). However, there are only little data regarding the long time effects of inhaled macromolecules except insulin and heparin (Agu et al. 2001; Cefalu 2004; Ceglia et al. 2006; Guntur and Dhand 2007; Ko¨hler 1994; Ko¨hler and Fleischer 2000; Laube 2005; Mastrandea and Quattrin 2006; Patton et al. 2004; Valente et al. 2003; Wolff 1998) and a lot of open questions should be subject of future investigations. For example, inhaled pharmaceuticals as well as additives for absorption enhancement may induce an incompatibility due to an acute or chronic toxicity. Peptides and proteins can cause an immunisation (Patton and Platz 1992; Patton et al. 2004; Valente et al. 2003; Wolff 1998) but also can have potential specific effects on the target organ lung (e.g. growth stimulating effect of insulin, risk for lung cancer) (Cefalu 2004; Patton et al. 2004).

246

S. Ha¨ussermann et al.

Furthermore pulmonary diseases (e.g. COPD and asthma) as well as individual lifestyle (e.g. smoking) may strongly affect pharmacokinetics, efficiency and safety of inhaled pharmaceuticals. Finally, the costs of this type of therapy should be considered and minimised. However, if the open questions are answered, inhalation therapies for treatment of systemic diseases can be an important clinical tool in the future, because it seems to be a safe and reliable technique for drug application improving patient compliance due to its noninvasive character.

References Agu RU, Ugwoke MI, Armand M, Kinget R, Verbeke N (2001) The lung as a route for systemic delivery of therapeutic proteins and peptides. Respir Res 2:198–209 Barnett AH, Dreyer M, Lange P, Serdarevic-Pehar M (2006a) On behalf of the Exubera Phase III Study Group. An open, randomized, parallel-group study to compare the efficacy and safety profile of inhaled human insulin (Exubera) with metformin as adjunctive therapy in patients with type 2 diabetes poorly controlled on a sulfonylurea. Diabetes Care 29:1282–1287 Barnett AH, Dreyer M, Lange P, Serdarevic-Pehar M (2006b) On behalf of the Exubera Phase III Study Group. An open, randomized, parallel-group study to compare the efficacy and safety profile of inhaled human insulin (Exubera) with glibenclamide as adjunctive therapy in patients with type 2 diabetes poorly controlled on metformin. Diabetes Care 29:1818–1825 Barnett AH, Lange P, Dreyer M, Serdarevic-Pehar M (2007) On behalf of the Exubera1 Phase 3 Study Group. Long-term tolerability of inhaled human insulin (Exubera1) in patients with poorly controlled type 2 diabetes. Int J Clin Pract 61:1614–1625 Becker RHA, Sha S, Frick AD, Fountaine RJ (2006) The effect of smoking cessation and subsequent resumption on absorption of inhaled insulin. Diabetes Care 29:277–282 Belmin J, Valensi P (2003) Novel drug delivery systems for insulin. Clinical potential for use in the elderly. Drugs Aging 20:303–312 Black C, Cummins E, Royle P, Philip S, Waugh N (2007) The clinical effectiveness and costeffectiveness of inhaled insulin in diabetes mellitus: a systematic review and economic evaluation. Health Technol Assess 11(33) Brunner GA, Balent B, Ellmerer M, Schaupp L, Siebenhofer A, Jendle JH, Okikawa J, Pieber TR (2001) Dose-response relation of liquid aerosol inhaled insulin in type I diabetic patients. Diabetologia 44:305–308 Byron PR, Patton JS (1994) Drug delivery via the respiratory tract. J Aerosol Med 7:49–75 Cefalu WT (2004) Concept, strategies, and feasibility of non-invasive insulin delivery. Diabetes Care 27:239–246 Ceglia L, Lau J, Pittas AG (2006) Meta-analysis: efficacy and safety of inhaled insulin therapy in adults with diabetes mellitus. Ann Intern Med 145:665–675 Charlish P (2006) Therapeutic focus – How would you like your insulin. BioPartnering Today Autumn:6–9 Chien JY, Wise SD, Sathirakul K (2001) Time action profile of inhaled insulin via Spiros dry powder inhaler is consistent among user inhalation techniques. Diabetologia 44(Suppl 2):A211 (abstract) Davis TME, Davis WA (2007) An assessment of eligibility for inhaled insulin (Exubera). Diabetes Care 30:360–361 de Galan BE, Simsek S, Tack CJ, Heine RJ (2006) Efficacy and safety of inhaled insulin in the treatment of diabetes mellitus. Neth J Med 64:319–325 DeFronzo RA, Bergenstal RM, Cefalu WT, Pullman J, Lerman S, Bode BW, Phillips LS (2005) For the Exubera Phase III Study group. Efficacy of inhaled insulin in patients with type

13 Targeting Drugs to the Lungs – The Example of Insulin

247

2 diabetes not controlled with diet and exercise: A 12-week, randomized, comparative trial. Diabetes Care 28:1922–1928 Dhand R (2004) New frontiers in aerosol delivery during mechanical ventilation. Respir Care 49:666–677 Elliott RB, Edgar BW, Pilcher CC, Quested C, McMaster J (1987) Parenteral absorption of insulin from the lung in diabetic children. Aust Paediatr J 23:293–297 Farr SJ, McElduff A, Mather LE, Okikawa J, Ward ME, Gonda I, Licko V, Rubsamen RM (2000) Pulmonary insulin administration using the AERx system: Physiological and physicochemical factors influencing insulin effectiveness in healthy fasting subjects. Diabetes Technol Ther 2:185–197 Fishman RS, Guinta D, Chambers F (2000) Insulin administration via the AerodoseTM inhaler: comparison to subcutaneously injected insulin. Diabetes 49 (Suppl 1):A9 (abstract) Fountaine R, Milton A, Checchio T, Wei G, Stolar M, Teeter J, Jaeger R, Fryburg D (2008) Acute passive cigarette smoke exposure and inhaled human insulin (Exubera) pharmacokinetics. Br J Clin Pharmacol 65:864–870 ¨ ber Inhalation von Insulin. Klin Wochenschr 4:71 Ga¨nsslen M (1925) U Ghosh S, Collier A (2007) Inhaled insulins. Postgrad Med J 83:178–181 Gonda I (2000) The ascent of pulmonary drug delivery. J Pharm Sci 89:940–945 Guntur VP, Dhand R (2007) Inhaled insulin: extending the horizons of inhalation therapy. Respir Care 52:911–922 Heinemann L, Traut T, Heise T (1997) Time-action profile of inhaled insulin. Diabet Med 14:63–72 Heise T, Scharling B, Bellaire S (2001) Dose-response of pulmonary insulin with the AERx insulin diabetes management system in healthy subjects. Diabetologia 44 (Suppl 2):A212 (abstract) Henry RR, Mudaliar SRD, Howland WC, Chu N, Kim D, An B, Reinhardt RR (2003a) Inhaled insulin using the AERx insulin diabetes management system in healthy and asthmatic subjects. Diabetes Care 26:764–769 Henry RR, Mudaliar S, Chu N, Kim D, Armstrong D, Davis TT, An B, Reinhardt RR (2003b) Young and elderly type 2 diabetic patients inhaling insulin with the AERx1 insulin diabetes management system a pharmacokinetic and pharmacodynamic comparison. J Clin Pharmacol 43:1228–1234 ¨ ber Inhalation von Insulin. Klin Wochenschr Heubner W, de Jongh SE, Laquer E (1924) U 3:2342–2343 Himmelmann A, Jendle J, Mellen A, Petersen AH, Dahl UL, Wollmer P (2003) The impact of smoking on inhaled insulin. Diabetes Care 26:677–682 Hollander PA, Blonde L, Rowe R, Mehta AE, Milburn JL, Hershon KS, Chiasson JL, Levin SR (2004) For the Exubera Phase III Study Group. Efficacy and safety of inhaled insulin (Exubera) compared with subcutaneous insulin therapy in patients with type 2 diabetes: results of a 6-month, randomized, comparative trial. Diabetes Care 27:2356–2362 Irngartinger M, Camuglia V, Damm M, Goede J, Frijlink HW (2004) Pulmonary delivery of therapeutic peptides via dry powder inhalation: effects of micronisation and manufacturing. Eur J Pharm Biopharm 58:7–14 Kim D, Mudaliar S, Plodkowski R (2002) Dose-response relationships of inhaled and subcutaneous insulin in type 2 diabetic patients. Diabetes 51 (Suppl 2):A47 (abstract) Ko¨hler D (1990) Aerosols for systemic treatment. Lung 168 (Suppl):677–684 Ko¨hler D (1994) Aerosolized heparin. J Aerosol Med 7:307–314 Ko¨hler D, Fleischer W (2000) Medikamente. In: Ko¨hler D, Fleischer W (eds) Theorie und Praxis der Inhalationstherapie. Arcis Verlag, Mu¨nchen, pp 71–99 Ko¨hler D, Schlu¨ter KJ, Kerp L, Matthys H (1987) Nicht radioaktives Verfahren zur Messung der Lungenpermeabilita¨t. Inhalation von Insulin. Atemw-Lungenkrkh 13:230–232 ¨ ber die Wirkung intratrachealer Zufu¨hrung von Insulin. Klin Laqueur E, Grevenstuk A (1924) U Wochenschr 3:1273–1274

248

S. Ha¨ussermann et al.

Laube BL (2005) The expanding role of aerosols in systemic drug delivery, gene therapy, and vaccination. Respir Care 50:1161–1176 Laube BL, Georgopoulos A, Adams GK (1993) Preliminary study of the efficacy of insulin aerosol delivered by oral inhalation in diabetic subjects. JAMA 269:2106–2109 Mastrandrea LD, Quattrin T (2006) Clinical evaluation of inhaled insulin. Adv Drug Deliv Rev 58:1061–1075 McElduff A, Mather LE, Kam PC, Clauson P (2005) Influence of acute upper respiratory tract infection on the absorption of inhaled insulin using the AERx1 insulin diabetes management system. Br J Clin Pharmacol 59:546–551 FDA MedWatch Alert (2008) Exubera (insulin human rDNA origin) Inhalation Powder. Available from http://www.drugs.com/fda/exubera-insulin-human-rdna-origin-inhalation-powder12372.html Mudaliar S, Henry RR (2007) Inhaled insulin in patients with asthma and chronic obstructive pulmonary disease. Diabetes Technol Ther 9(Suppl 1):S83–S92 NDA 21-868/EXUBERA: U.S. package insert (2006) New York, NY, Pfizer Labs. Available from http://www.pfizer.com/pfizer/download/uspi_exubera.pdf Newman SP (2005) Principles of metered-dose inhaler design. Respir Care 50:1177–1190 Niven RW (1995) Delivery of biotherapeutics by inhalation aerosol. Crit Rev Ther Drug Carrier Syst 12:151–231 Owens DR (2002) New horizons – alternative routes for insulin delivery. Nat Rev Drug Discov 1:529–540 Patton JS, Byron PR (2007) Inhaling medicines: delivering drugs to the body through the lungs. Nat Rev Drug Discov 6:67–74 Patton JS, Platz RM (1992) Routes of delivery: case studies. (2) Pulmonary delivery of peptides and proteins for systemic action. Adv Drug Deliv Rev 8: 179–196 Patton JS, Bukar JG, Eldon MA (2004) Clinical pharmacokinetics and pharmacodynamics of inhaled insulin. Clin Pharmacokinet 43:781–801 Pearson J (2006) Inhalation technologies – a breath of fresh air. Drug Deliv Rep Spring/Summer:19–21 Perera AD, Kapitza C, Nosek L, Fishman RS, Shapiro DA, Heise T, Heinemann L (2002) Absorption and metabolic effect of inhaled insulin. Intrapatient variability after inhalation via the Aerodose insulin inhaler in patients with type 2 diabetes. Diabetes Care 25:2276–2281 Quattrin T, Belanger A, Bohannon NJV, Schwartz SL (2004) For the Exubera Phase III Study Group. Efficacy and safety of inhaled insulin (Exubera) compared with subcutaneous insulin therapy in patients with type 1 diabetes. Diabetes Care 27:2622–2627 Radermecker RP, Pierard GE, Scheen AJ (2007) Lipodystrophy reactions to insulin Effect of continuous insulin infusion and new insulin analogs. Am J Clin Dermatol 8:21–28 Rave KM, Heise T, Pfu¨tzner A (2000) Results of a dose-response study with a new pulmonary insulin formulation and inhaler. Diabetes 49(Suppl 1):A75 (abstract) Rave K, Muchmore D, Gonzales C (2001) Inhaled insulin with an improved Spiros1 dry powder inhaler: dose response and time-action profiles. Diabetologia 44(Suppl 2):A211 (abstract) Rave K, de la Pena A, Tibaldi FS, Zhang L, Silverman B, Hausmann M, Heinemann L, Muchmore DB (2007) AIR inhaled insulin in subjects with chronic obstructive pulmonary disease: Pharmacokinetics, glucodynamics, safety, and tolerability. Diabetes Care 30:1777–1782 Rosenstock J, Zinman B, Murphy LJ, Clement SC, Moore P, Bowering CK, Hendler R, Lan SP, Cefalu WT (2005) Inhaled insulin improves glycemic control when substituted for or added to oral combination therapy in type 2 diabetes: a randomized, controlled trial. Ann Intern Med 143:549–558 Royle P, Waugh N, McAuley L, McIntyre L, Thomas S (2004) Inhaled insulin in diabetes mellitus (Review). Cochrane Database Syst Rev CD003890 Sakagami M (2004) Insulin disposition in the lung following oral inhalation in humans. A metaanalysis of its pharacokinetics. Clin Pharmacokinet 43:539–552

13 Targeting Drugs to the Lungs – The Example of Insulin

249

Scheuch G, Siekmeier R (2007) Novel approaches to enhance pulmonary delivery of proteins and peptides. J Physiol Pharmacol 58(Suppl 5, Pt 2):615–625 Scheuch G, Kohlhaeufl MJ, Brand P, Siekmeier R (2006) Clinical perspectives on pulmonary systemic and macromolecular delivery. Adv Drug Deliv Rev 58:996–1008 Scheuch G, Brand P, Meyer T, Herpich C, Mu¨llinger B, Brom J, Weidinger G, Kohlha¨ufl M, Ha¨ussinger K, Spannagl M, Schramm W, Siekmeier R (2007) Anticoagulative effects of the inhaled low molecular weight heparin certoparin in healthy subjects. J Physiol Pharmacol 58 (Suppl 5, Pt 2):603–614 Sha S, Becker RHA, Willavise SA (2002) The effect of smoking cessation on the absorption of inhaled insulin (Exubera1). Diabetes 51(Suppl 2):A133 (abstract) Shoyele SA, Cawthorne S (2006) Particle engineering techniques for inhaled biopharmaceuticals. Adv Drug Deliv Rev 58:1009–1029 Siekmeier R, Scheuch G (2005) Systemische Therapie mit Aerosolen. Beispiele zur pulmonalen Verabreichung von Makromoleku¨len zur systemischen Therapie. Atemw-Lungenkrkh 31:391–410 Skyler JS, Cefalu WT, Kourides IA, Landschulz WH, Balagtas CC, Cheng SL, Gelfand RA (2001) Efficacy of inhaled human insulin in type 1 diabetes mellitus: a randomised proof-of-concept study. Lancet 357:331–335 Skyler JS, Weinstock RS, Raskin P, Yale JF, Barrett E, Gerich JE, Gerstein HC (2005) The Inhaled Insulin Phase III Type 1 Diabetes Study Group. Use of inhaled insulin in a basal/bolus insulin regimen in type 1 diabetic subjects: A 6-month, randomized, comparative trial. Diabetes Care 28:1630–1635 Steiner S, Rave KM, Heise T (2000) Bioavailability and pharmacokinetic properties of inhaled dry powder Technosphere/Insulin. Diabetes 49(Suppl 1):A126 (abstract) Steiner S, Pfu¨tzner A, Wilson BR, Harzer O, Heinemann L, Rave K (2002) Technosphere/insulin– proof of concept study with a new insulin formulation for pulmonary delivery. Exp Clin Endocrinol Diabetes 110:17–21 Telko MJ, Hickey AJ (2005) Dry powder inhaler formulation. Respir Care 50:1209–1227 Valente AXCN, Langer R, Stone HA, Edwards DA (2003) Recent advances in the development of an inhaled insulin product. Biodrugs 17:9–17 Wigley FW, Londono JH, Wood SH, Ship JC, Waldman RH (1971) Insulin across respiratory mucosae by aerosol delivery. Diabetes 20:552–556 Wolff RK (1998) Safety of inhaled proteins for therapeutic use. J Aerosol Med 11:197–219 Yu J, Chien YW (1997) Pulmonary drug delivery: physiologic and mechanistic aspects. Crit Rev Ther Drug Carrier Syst 14:395–453

Chapter 14

Protection of the Respiratory System Against Nanoparticles Inhalation Albert Podg´orski

14.1

Introduction

Aerosol filtration in fibrous filters is one of the most widely used methods for high efficiency removal of fine and ultrafine aerosol particles. The technique is utilized in both traditional (e.g., steel, cement and mining) and high-tech industries, modern applications (microelectronics, opt electronics, biotechnology, pharmacy, nuclear plants, military); in the latter case, extremely high standards of filter efficiency are usually imposed. The same is true of the fibrous filters used in the respiratory protective devices (worn by surgeons, soldiers, firemen, miners, construction and farm workers), and of the laminar boxes in modern research laboratories, clean rooms, and operating rooms in hospitals. A lot of fibrous filters are also used in many daily applications, including HVAC systems in public and private buildings, filtering systems in cars, airplanes, and vacuum cleaners. A brief review of the classical single fiber theory and its possible extensions are presented first in this chapter and then specific issues related to nano-scale aspects of aerosol filtration are discussed (filtration of nanoparticles in electret filters, filtration of aerosol particles in nanofibrous media, filtration of fractal-like nanoaggregates, filtration of nanoparticles in polydisperse fibrous filters). The main objective of this work is to collect the information necessary to design efficient fibrous filters for protecting people against inhalation of harmful nanoparticles.

A. Podg´orski Faculty of Chemical and Process Engineering, Warsaw University of Technology, Waryn´skiego 1, 00-645 Warsaw, Poland e-mail: [email protected]

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_14, # Springer ScienceþBusiness Media B.V. 2010

251

A. Podgo´rski

252

14.2

Classical Single Fiber Theory

Aerosol particles, depending on their diameter and process conditions, can be deposited onto the fibers in fibrous filters due to several mechanisms, such as diffusional mechanism (Brownian motion), interception, inertial impaction, gravitational settling and electrostatic attraction, Fig. 14.1. According to the classical theory of depth filtration, the total single fiber efficiency, E, can be determined on the basis of the single fiber efficiencies obtained separately for particular mechanisms. The most commonly used approach assumes that E can be calculated as a sum of the efficiencies due to different mechanisms: E ¼ E D þ ER þ E I þ EQ þ E G

(14.1)

where ED, ER, EI, EQ and EG mean the single fiber efficiency due to diffusional, direct interception, inertial impaction, electrostatic and gravitational settling mechanisms, respectively. Another simple method utilized frequently for determination of the total single fiber efficiency is based on the assumption that all individual mechanisms act independently, thus: E ¼ 1  ð1  ED Þð1  ER Þð1  EI Þð1  EQ Þð1  EG Þ

(14.2)

More detailed analysis based on the Brownian dynamics method indicates, however, that various mechanisms of deposition are neither exactly additive nor completely independent. A simple formula to account approximately for coupling between stochastic and deterministic mechanisms of deposition was quoted by Hinds (1999). The single fiber efficiencies due to particular mechanisms can be obtained using the correlations that may be found in the literature. In the case of a mechanical filter and neutralized nano- and submicrometer particles, diffusion and interception are the only mechanisms that influence particles’ deposition. Particles larger than about 1 mm may deposit also due to inertial impaction mechanism and the trajectories of very large particles might be influenced by gravitational settling. Electrostatic mechanism should be considered only when particles or fibers carry some charges

Fig. 14.1 Mechanisms of aerosol particle deposition onto a fiber

14 Protection of the Respiratory System Against Nanoparticles Inhalation

253

or a filter is placed in an external electric field. Below we present a list of the most commonly used correlations for the single fiber efficiency due to action of various deposition mechanisms. To calculate the single fiber efficiency for diffusional mechanism, the following equation is frequently used:


1a ED ¼ 2:9 Ku

1=3

Pe2=3

(14.3)

In the above equation a means a filter packing density, Ku is the Kuwabara hydrodynamic factor: Ku ¼ 0:5 ln a  0:75 þ a  0:25a2 , and Pe is the Peclet number: Pe ¼ dF U0 =D

(14.4)

where dF is the fiber diameter, U0 denotes the air velocity, and D is a particle diffusion coefficient that is defined as follows: D ¼ kB T=f

(14.5)

In the Eq. 14.5 kB means the Boltzmann constant, T is the absolute temperature, and f is the particle friction coefficient: f ¼ 3 p m dP =CC , where dP denotes the particle diameter, m is the gas viscosity and CC is the Cunningham slip correction factor that is calculated as: CC ¼ 1 þ KnP ½aCc þ bCc expðdCc =KnP Þ

(14.6)

Numerical coefficients in Eq. 14.6 have the following values: aCc=1.142, bCc ¼ 0.558, dCc ¼ 0.999. The single fiber efficiency due to direct interception mechanism can be estimated as: ER ¼

ð1  aÞ NR2 Ku ð1 þ NR Þ

(14.7)

and NR ¼ dP/dF is the interception parameter. Larger particles can be deposited onto the fibers due to inertial impaction mechanism and EI can be approximated using the following formula: EI ¼

Stk J 2 Ku2

(14.8)

where J ¼ ð29:6  28a0:62 ÞNR 2  27:5NR 2:8 and Stk is the Stokes number, which is defined as: Stk ¼

dP 2 U0 rP CC 2 dP U0 rP CC 18m dF

(14.9)

A. Podgo´rski

254

Fig. 14.2 Filtration efficiencies for various deposition mechanisms and the total filter efficiency calculated on the basis of the classical theory assuming independence of particular deposition mechanisms

In the Eq. 14.9 rP denotes the particle density. According to the classical single fiber theory, the total filter efficiency, , is calculated on the basis of the total single fiber efficiency using the following formula:


4a E L  ¼ 1  exp p dF ð1  aÞ

 (14.10)

where L denotes the filter thickness. Figure 14.2 shows an example of the total filter efficiency as well as the efficiencies due to particular mechanisms calculated for a filter made of fibers with 10 mm in diameter, having packing density 0.1 and thickness 2 mm. These results were obtained at the gas face velocity 0.2 m/s, assuming the independence rule, Eq. 14.2. We can observe that the deposition efficiency for Brownian diffusion decreases with the increase of particle diameter, whilst the deposition efficiency for all other mechanisms increases then. As a consequence, a minimum of fractional efficiency is observed for particles with around 300 nm in diameter (the particle diameter, for which filtration efficiency reaches a minimum, is called the most penetrating particle size, MPPS). And finally, we can conclude that in the case of fibrous filters made of micrometer-sized fibers and when the electrostatic forces between particles and fibers are absent, the total efficiency is practically the same as that one for Brownian diffusion for particles smaller than around 100 nm. Classical single fiber theory outlined briefly above is very popular because of its simplicity. However, it is based on several very strong simplifications (e.g., mutual independence of various deposition mechanisms, the

14 Protection of the Respiratory System Against Nanoparticles Inhalation

255

same efficiency of all fibers, etc.), thus its accuracy may be in some instances unsatisfactory.

14.3

Brownian Dynamics Method

Instead of assuming that various deposition mechanisms act independently, it would be much better to calculate the total single fiber deposition efficiency directly, taking simultaneously all the mechanisms into account. It may be achieved using the Brownian dynamics approach. In this method different forces acting on a particle during its motion are considered simultaneously. The particle motion can be influenced by the drag and resistance force, F(DR), the external force (gravitational force, van der Waals force, electrostatic force, etc.), F(ext), and the fluctuating Brownian force, F(B); thus, a stochastic equation of particle motion has the following form (Langevin’s equation): mP

dv ¼ FðDRÞ þ FðextÞ þ FðBÞ dt

(14.11)

Because of the stochastic nature of the above equation, the trajectories of identical particles starting from the same position can differ. Thus, only a probability distribution ’i(Dvi, DLi) that during the time interval Dt the particle will change its ith component of the velocity by Dvi and it will be displaced by the distance DLi in ith direction can be derived, which may be expressed by the bivariate normal distributions as follows: (

"
 #) 1 1 Dvi  hDvi i 2 ’i ðDvi ;DLi Þ ¼ pffiffiffiffiffiffi exp  2 svi 2psvi 8 2 !2 39 = < DLi  hDLi i  rsi sviLi ðDvi  hDvi iÞ 1 1 5 ffi exp4 pffiffiffiffiffiffiffiffiffiffiffiffi  pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; : 2pð1  r2i ÞsLi 2 sLi 1  r2i (14.12) In the Eq. 14.12 hDvii is the expected value of the particle velocity change and hDLii means the expected value of the linear displacement of the particle over the time-step Dt, and they can be expressed as: h i ðextÞ hDvi i ¼ ui  vi þ Fi =ðmP bÞ ½1  expðbDtÞ # " # ðextÞ ðextÞ Fi 1 Fi Dt  ½1  expðbDtÞ ui  vi þ hDLi i ¼ ui þ mP b mP b b

(14.13)

"

(14.14)

A. Podgo´rski

256

where b ¼ f =mP . svi and sLi in Eq. 14.12 are the standard deviations defined as: svi ¼ sLi ¼

qffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi hDv2i i ¼ ð1  e2bDt Þ kB T=mP

qffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi hDL2i i ¼ ð2bDt  3 þ 4ebDt  e2bDt Þ kB T=ðmP b2 Þ

(14.15) (14.16)

and the coefficient of correlation, ri, is given by: pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi hDLi Dvi i  2    1=2 1  e2bDt 2bDt  3 þ 4ebDt  e2bDt ¼ 1  ebDt

ri ¼

(14.17)

The change of the particle velocity, Dvi, and the particle linear displacement, DLi, during the time-step Dt are calculated from the expressions accounting for deterministic and stochastic motion: Dvi ¼ hDvi i þ Gvi svi DLi ¼ hDLi i þ

ri sLi ðDvi  hDvi iÞ þ svi

(14.18) qffiffiffiffiffiffiffiffiffiffiffiffiffi 1  r2i GLi sLi

(14.19)

where GLi and Gvi are non-correlated random numbers that have Gaussian distribution with a zero mean and unit variance. The new particle velocity at a moment t þ Dt is obtained as vi(t þ Dt) ¼ vi(t) þ Dvi, and the new particle position as: Li(t þ Dt) ¼ Li(t) + DLi. The simulations are continued until the particle deposits on the fiber or leaves a considered area. The procedure described above will be called the standard Brownian dynamics algorithm based on bivariate distribution sampling. More detailed description of this and other Brownian dynamics methods can be found in the work by Podgo´rski (2002) and by Ermak and Buckholz (1980). Two examples of aerosol particles’ trajectories around a fiber, calculated with the use of the Brownian dynamics method, are shown in Fig. 14.3 for particle diameter 0.5 and 0.1 mm. Drastically increasing degree of randomness of particle motion can be observed with decrease in particle diameter. It may be interesting to compare results of the Brownian dynamics simulations with the classical theory that assumes independence of the stochastic and deterministic mechanisms of deposition. Such a comparison is shown in Fig. 14.4. The deposition efficiency for deterministic mechanisms only (in this case – for inertial impaction and interception), Edet, was obtained by the method of the limiting trajectories solving the deterministic equation of motion, while the single fiber efficiency for Brownian diffusion was calculated from Eq. 14.3. Good agreement between these two approaches is observed for two limiting cases: (i) for very fine particles, dP < 0.2 mm (diffusional regime), and (ii) for larger particles, dP >1 mm (inertial limit). For intermediate particle diameters, the Brownian dynamics method predicts

14 Protection of the Respiratory System Against Nanoparticles Inhalation

257

Fig. 14.3 Examples of aerosol particles’ trajectories around a fiber calculated using the Brownian dynamics method for particle diameter 0.5 mm (left) and 0.1 mm (right)

Fig. 14.4 Comparison of the independence rule of standard single fiber theory with Brownian dynamics results

a higher deposition efficiency than the combination of the deterministic efficiency from the trajectory analysis and diffusional efficiency from the solution of the convective-diffusion equation. Thus, the common method of the total deposition efficiency calculation based on the Eq. 14.2 seems to be inappropriate as it neglects the coupling effects between particle inertia and Brownian motion. Figure 14.5 shows comparison of aerosol particles penetration calculated on the basis of the standard single fiber theory and using the Brownian dynamics method with experimental data. Much better agreement with experiment is observed in the case of

258

A. Podgo´rski

Fig. 14.5 Comparison of the standard single fiber theory (dashed line) with Brownian dynamics results (solid line) and experimental data (points)

Brownian dynamics simulations. Thus, the Brownian dynamics approach, which enables one to calculate the total deposition rate due to all mechanisms acting simultaneously, is a very advantageous approach as no other method for combining the fractional efficiencies has a solid physical background. We may therefore conclude that coupling between Brownian motion and inertia is an important phenomenon and it may be accounted properly for only with the Brownian dynamics method. The results discussed so far concern standard implementation of the Brownian dynamics method, i.e., for the constant friction and diffusion coefficients. Podgo´rski (2001, 2002) generalized the Brownian dynamics method taking into account short-range interactions particle-fiber, namely, hydrodynamics interactions resulting in variation of the friction and diffusion coefficients for an aerosol particle moving near a solid surface and van der Waals forces between a particle and a fiber. Sample result of calculations done for various strengths of van der Waals forces (measured by the value of the Hamaker constant, A) is shown in Fig. 14.6. We can observe that deposition efficiency calculated using the LevichSmoluchowski approximation (i.e., assuming that hydrodynamic interactions and van der Waals forces counterbalance) may be either higher or lower than the efficiency predicted by the generalised Brownian dynamics algorithm, depending on the value of the Hamaker constant. Since Hamaker constants rarely exceed 2  1020J, the results obtained here indicate that effect of hydrodynamic interactions is more important and therefore it should be expected that in most cases the Levich-Smoluchowski approximation will lead to a too optimistic estimation of the deposition efficiency.

14 Protection of the Respiratory System Against Nanoparticles Inhalation

259

Fig. 14.6 Effect of the van der Waals interactions strength on the single fiber efficiency

14.4

Filtration of Nanoparticles in Electret Filters

One of the most effective ways to enhance aerosol particles’ separation in fibrous filters, especially these with diameters close to MPPS, is to make use of electrostatic interactions between the particles and the fibers. The electrostatic forces that can be useful in the intensification of fibrous filtration of aerosol particles can be classified as follow: (i) Coulombic force, when both particles and fibers are charged; (ii) polarization force, when a non-uniform electric field around a permanently charged fiber induces a dipole in a neutral aerosol particle; (iii) image force, when charged particles induce dipoles in a non-charged fiber; (iv) dielectrophoretic force, when both fibers and particles are non-charged dielectrics and the filter is placed in an external electric field. This work deals with the second case, that is with utilization of electret fibers (having “frozen” electric charges) to enhance collection efficiency of electrically neutral dielectric aerosol particles. Depending on the charge state, the unipolar (that have charges of one sign) and the bipolar (that have charges in the form of a line-dipole) electrets can be distinguished; the latter ones, which are more common, are considered in this work. Podgo´rski and Bałazy (2008) used the Brownian dynamics method to calculate the single fiber deposition efficiencies of electrically neutral aerosol particles in bipolarly charged fibrous filters for several values of the fiber charge density, qF, and for various particle diameters. Simulations were performed for the following set of parameters: fiber diameter, dF ¼ 7.84 mm, filter packing density a ¼ 0.069, and air velocity U0 ¼ 0.129 m/s. As can be seen in Fig. 14.7, a noticeable increase in E is observed with the rise in qF. Results of direct Brownian dynamics simulations were validated using the experimental data of Bałazy et al. (2006) of aerosol penetration through a commercial, triple-layer respirator, containing an electret middle layer. As seen

260

A. Podgo´rski

Fig. 14.7 Single fiber efficiencies calculated for various fibers’ charge densities

Fig. 14.8 Experimental validation of Brownian dynamics simulations using data of Bałazy et al. (2006)

in Fig. 14.8, the agreement is satisfactory if one assumes the fiber charge density of the order 1.2–1.3 nC/m. It should be also noticed that both theory and experiment predict a significant shift of the MPPS towards smaller particles in electret filters (30–40 nm for the respirator analyzed in Fig. 14.8) compared to the case of standard, mechanical filters (around 300 nm). Thus, standard certification procedure based on testing a filter against 300 nm aerosol particles is improper for electret filters as it does not represent the worst scenario.

14 Protection of the Respiratory System Against Nanoparticles Inhalation

261

A gain in the single fiber efficiency attributed to the action of the polarization force was calculated as: DEel ¼ Eel  Emech, where Eel and Emech refer to the electret and structurally identical mechanical filter, respectively. It is used to correlate DEel with the dimensionless polarization parameter, which is defined as: Ns0 ¼

2ðeP  1ÞdP2 q2F CC 3e0 ðeP þ 2Þð1 þ eF Þ2 mU0 dF3

(14.20)

wherein eP and eF denote the particle and the fiber dielectric constant, respectively, and e0 is the vacuum permittivity. We found that for each considered charge density of the fiber, the numerical data of DEel(Ns0) can be very precisely interpolated using the following formula: DEel ¼ aNs0b/(1 þ cNs0d), which contains 4 fitting parameters (a,b,c,d). The results of extensive numerical simulations allowed us to derive simple formulae relating an increase in the deposition efficiency due to action of the polarization force as a function of the dimensionless polarization parameter. We proposed two equations, depending on the range of the Ns0 variability: DEel ¼

DEel ¼

0:771Ns0

for 103 < Ns0 < 103

(14.21)

for 5  103 < Ns0 < 103

(14.22)

1 þ 0:973ðKuNs0 Þ0:785 0:74 0:528Ku0:26 Ns0

1 þ 0:366ðKuNs0 Þ

0:63

Equation 14.21 was derived using a complete set of discrete numerical data, while Eq. 14.22 was obtained neglecting results for the three smallest particles in each considered case of the fiber charge density, when numerical data deviate from a common curve DEel(Ns0). Figure 14.9 illustrates comparison of these two interpolating functions with all discrete numerical data obtained from the Brownian dynamics simulations. Applicability of our new correlations were verified using the experimental data of Kim et al. (2005) and of Lee et al. (2002), see Figs. 14.10 and 14.11, and the results were also compared to the literature correlations proposed by Brown (1981): DEel ¼ 0:54 Ku0:6 Ns0 0:4 for1 < Ns0 < 100

(14.23)

and by Otani et al. (1993): DEel ¼ 1:48Ns0 0:93 for104 < Ns0 < 102

(14.24)

DEel ¼ 0:51 Ku0:35 Ns0 0:73 for102 < Ns0 < 1

(14.25)

It may be concluded that both newly derived correlations for the single fiber efficiency due to the polarization force describe better available experimental

262

A. Podgo´rski

Fig. 14.9 Comparison of all numerical data with the proposed correlations

Fig. 14.10 Verification of proposed correlations using experimental data of Kim et al. (2005)

data than other formulae that can be found in the literature. Moreover, our new theoretical correlations cover in a continuous way a very broad range of values of the polarization force parameter. Comparing predictions of Eqs. 14.21 and 14.22 with data of Kim et al. (2005) it seems that Eq. 14.21 works better for very fine

14 Protection of the Respiratory System Against Nanoparticles Inhalation

263

Fig. 14.11 Verification of proposed correlations using experimental data of Lee et al. (2002)

particles. All discussed formulae overestimate slightly experimental data in the case of strong electrostatic interactions (i.e., for high values of Ns0). It may be due to the fact that all of them are based on the standard Kuwabara model that neglects presence of neighboring fibers, particularly – effect of other fibers on the electric field intensity around the fiber in question.

14.5

Filtration of Fractal-like Nanoaggregates

Whilst the filtration theory of spherical aerosol particles was being developed for many decades, much less effort has been put so far into theoretical analysis of nonspherical particles removal in fibrous filters, even though agglomerated soot particles emitted by diesel engines are particularly adverse for environment and hazardous for humans. Results of extensive numerical studies of the deposition efficiency in fibrous filters of fractal-like aggregates composed of nanosized primary particles were presented by Bałazy and Podgo´rski (2007). The simulations were performed using the modified Brownian dynamics method that allows one all the mechanisms of deposition to be taken into account straightforwardly. Effect of various parameters, like the aggregate fractal dimension, primary particle size, fiber radius and gas velocity, on the single fiber deposition efficiency was investigated theoretically. Results obtained for the aggregates were compared with these ones got for spherical particles of the same masses and with particles having the same

264

A. Podgo´rski

Fig. 14.12 The ratios of the single fiber efficiencies for spherical particles to these ones for the aggregates of various fractal dimensions. RF ¼ 10 mm, Rpp ¼ 0.01 mm, U0 ¼ 0.2 m/s, a ¼ 0.01

mobility or outer radii. Those data were also interpreted on the basis of the dimensionless numbers (Peclet, Stokes, interception) commonly used in the classical theory of depth filtration. The results obtained indicate that the collection efficiency of the aggregates can significantly differ from that observed for spherical particles and this discrepancy may have a various nature, depending on the predominating mechanism of deposition. For example, effect of the aggregate fractal dimension, Df, on the single fiber deposition efficiency computed using the modified BD algorithm is shown in Fig. 14.12 for five different fractal dimensions: 1.65, 1.85, 2.05, 2.5, and 2.8. The results are presented as the ratio of the single fiber efficiency for an aggregate to that one for a mass-equivalent spherical particle as a function of the mass-equivalent spherical particle radius, Rms. This ratio may significantly exceed one, especially for the particles with Rms in the range 0.03– 0.9 mm. The deposition efficiency of the fractal-like aggregate of a fixed mass strongly depends on its fractal dimension. For the diffusion-controlled deposition (i.e., for smaller particles), the efficiency of deposition for fractal-like aggregates is smaller than for spherical particles of the same masses and among the aggregates the lowest efficiency is observed for clusters with the smallest fractal dimension. This is due to the fact that the coefficient of Brownian diffusion for aggregates decreases with the decrease of Df. In contrast to that, for the interception-dominated range, the deposition efficiency is higher for smaller values of Df and aggregates are captured more effectively than spheres of the same masses. This is caused by the fact that for a fixed particle mass, the outer radius of the aggregate increases with the decrease of Df. The most penetrating particle size (the particle size for which the efficiency achieves minimum) shifts towards smaller clusters with the fractal

14 Protection of the Respiratory System Against Nanoparticles Inhalation

265

dimension decrease. The lowest value of the efficiency for the spherical particle (Df = 3) is observed for the particles with mass-equivalent radius about 300 nm, while for Df = 1.65 the minimum of the efficiency is for the particles with Rms only about 55 nm. Plotting the data of the single fiber efficiencies versus the mobility radius a unique curve is obtained for all aggregates with various fractal dimensions and for spherical particles for the diffusion-controlled deposition, but the same is not true of the case for larger particles. In this case a unique curve for all aggregates may be obtained if the results are related to the particle maximum radius; however, they do not coincide with the data for spherical particles, for which the deposition efficiency is higher than for clusters with the same outer radii. This observation leads to two conclusions. First, the deposition efficiency of small aggregates can be predicted using their mobility radii. Secondly, for larger aggregates, the deposition is controlled by direct interception, thus, it can be described by the outer radius of the aggregate, and the inertial effects for porous clusters are much weaker than they are for impermeable spherical particles. Next we examined the effect of the radius of a primary particle, Rpp, on the deposition efficiency for a constant gas velocity (U0 ¼ 0.2 m/s), fiber radius (RF ¼ 10 mm) and cluster fractal dimension (Df ¼ 1.85). For a fixed particle mass, the deposition efficiency due to Brownian diffusion is the lowest for the aggregates composed of the smallest primary particles, whilst for the deposition due to interception such a trend is reversed. It should be noticed that the most penetrating particle size (expressed by the mass-equivalent radius of particles corresponding to the maximum of penetration; Fig. 14.13) shifts towards smaller particles with decrease of a primary particle radius. It was also found that the unique relationship between the single fiber efficiency and the Peclet number is obtained for small

Fig. 14.13 Effect of the primary particle size on penetration through a filter with thickness of 50 mm. Results for Df ¼ 1.85

A. Podgo´rski

266

clusters and another unique dependence of the single fiber efficiency on the interception parameter exists for large clusters in all examined cases of aggregates composed of primary particles with various radii.

14.6

Aerosol Filtration in Nanofibrous Filters

Theoretical analysis suggests that substantial improvement of a filter performance, especially for MPPS, can be achieved using filtering media composed of nanofibers, Podgo´rski et al. (2006). To verify these predictions experimentally we produced six polypropylene fibrous filters using the melt-blown technology. The complete structural characteristics of these filters are shown in Tables 14.1 and 14.2. We intended to manufacture one conventional filter containing fibers with a few dozens of micrometers in diameter (it will be called the backing layer, BL) and five filters consisting predominantly of much finer, preferably submicrometer-sized fibers (these five filters will be consecutive labeled as nanofibrous layer #1, NL1, . . ., nanofibrous layer #5, NL5). SEM photographs of these filters are shown in Fig. 14.14. Fractional penetrations of DEHS particles (in the size range from 10 to 500 nm) passing through the sets of the backing layer (support) and a facial nanofibrous layer were measured using the Wide-Range Particle Spectrometer (MSP Corp., USA) The results obtained are shown in Fig. 14.15. In all experiments the same filter was used as a backing layer; however, with each nanolayer a new piece of the backing filter was taken. These results indicate that using an additional layer of the filter Table 14.1 Main parameters of the fibers’ size distributions Filter Mean fiber diameter Coefficient of variation dFCSD (mm) CV= SD/dFC () BL 18.03  6.84 0.38 NL1 0.74  0.30 0.41 NL2 0.87  0.63 0.72 NL3 1.30  0.77 0.59 NL4 1.18  0.59 0.50 NL5 1.41  0.84 0.60

Range of the fiber diameters (mm) 10.24–37.48 0.38–1.67 0.21–3.16 0.42–3.68 0.31–2.37 0.35–4.03

Table 14.2 Macroscopic structural characteristics of the manufactured filters Filter Thickness Porosity Basis weight L (mm) e () rSF (g/m2) BL 2.1 0.851 284.9 NL1 1.4 0.965 44.4 NL2 2.5 0.967 75.1 NL3 3.1 0.971 79.4 NL4 5.5 0.980 100.5 NL5 4.3 0.986 53.0

14 Protection of the Respiratory System Against Nanoparticles Inhalation

267

Fig. 14.14 SEM photographs of the investigated filters

Fig. 14.15 Fractional penetration for bilayer systems with the BL at the rear

composed of nanofibers the filter efficiency considerably increases, especially in the range of the most penetrating particle size, where the efficiency achieves the lowest values. Only for one nanolayer (NL1) the efficiency of the set of this nanolayer and the backing layer was lower than that observed for two layers of the backing filter.

A. Podgo´rski

268

Fig. 14.16 Pressure drop per unit filter thickness as a function of air face velocity

However, the NL1 was much thinner than the backing layer and, as all other nanolayers, it was very porous. We can also note a shift of MPPS towards smaller particles for bilayer composites containing a nanolayer compared to the case of two BLs. For all filters the pressure drop was measured at several gas velocities and these results are presented in Fig. 14.16 in the form of the pressure drop per unit filter thickness. It can be observed that the resistance of all nanofibrous filters is lower than that one for the BL filter. On the basis of the experimental data of the filters’ efficiencies and the pressure drops across the filters, the quality factor (QF) was calculated, which is defined as the ratio of the minus logarithm of the penetration, P, to the pressure drop: QF ¼

 ln P Dp

(14.26)

The quality factor defined in such a way should be independent of the filter thickness, thus, allowing one various filtering media to be quantitatively compared. Figure 14.17 shows the comparison of the QF values of the tested filters, which indicates that using the filter with the micrometer-sized fibers is the least profitable, since the QF values obtained for backing filter were generally the lowest. Only using one nanofiber layer (NL2) the quality factor for the particles smaller than about 30 nm had lower values than that for the BL; however, the pressure drop across this nanolayer was relatively high compared to others nanolayers. The QF values indicate that the best solutions were the sets with nanolayers NL4 and NL5, which resistances to the air flow (per unit filter thickness) were the lowest and their efficiencies were the highest in the entire measured size range. The count mean

14 Protection of the Respiratory System Against Nanoparticles Inhalation

269

Fig. 14.17 Quality factor for bilayer systems with the BL at the rear

fiber diameters of the filters NL4 and NL5 were not the smallest among the tested nanofibrous filters; however, their equivalent fiber diameters, determined on the basis of the measured pressure drop, were the lowest. Since the filter efficiency depends on the filter thickness, the increase of the number of nanolayers added to the backing layer should increase the filter efficiency. Experiments with one, two and three layers of NL1 (which is the thinnest filter among the tested ones) placed on the backing layer were carried out. The results of the measured penetrations are presented in Fig. 14.18, and a considerable increase of the efficiency with addition of the consecutive layers may be observed, particularly for the size of the most penetrating particles. These nanofibrous media have been also evaluated in terms of nonstationary filtration of polydisperse solid particles with diameters between 0.2 and 10 mm. Measurements were performed for bi- and triple-layered sets of filters, where a microfibrous support constituted the rear layer, which was covered by a nanofibrous layer (bi-layered sets), and that one was additionally covered by a microfibrous, facial layer (the same as the support) in the case of triple-layered sets. The momentary number and mass filtration efficiency as well as the momentary pressure drop were registered during continuous loading of all examined bi- and triplelayered sets of filters, and on this basis the values of the momentary quality factors were calculated. Figure 14.19 shows examples of the photographs for the NL3 layer in both configurations taken at the end of the experiments. One can observe formation of a thick external cake in the case of the bi-layered set, which is absent for the triple-layered system, when the biggest particles are collected mostly within the external microfibrous layer.

270

A. Podgo´rski

Fig. 14.18 Fractional penetrations for multilayer systems with the BL at the rear and a few NL1 layers at the front

Fig. 14.19 Photos of loaded layers at the end of experiments for the sets containing the NL3 filter: (a) NL3 layer in the bi-layered set; (b) facial microfibrous layer in the triple-layered set; (c) middle NL3 layer in the triple-layered set

The two following plots, Figs. 14.20 and 14.21, present how the momentary mass collection efficiency varies in time during the continuous loading of all examined bi-layered and triple-layered systems. As can be observed, the efficiency is comparable for both configurations. Information about the filter collection efficiency is insufficient to compare various filtering systems and to judge which solution is the best one. For this purpose, knowledge on the rise in the momentary pressure drop in time during the loading is also needed, and such experimental results obtained for all examined sets are presented in Fig. 14.22 for the bi-layered systems and in Fig. 14.23 for the triple-layered systems. We can observe that the lowest initial pressure drops are detected for systems that consisted of the thinnest and made of the finest fibers NL1 and NL2 nanofibrous layers. The initial pressure drops measured for the triplelayered systems are obviously higher than these ones for the bi-layered sets.

14 Protection of the Respiratory System Against Nanoparticles Inhalation

271

Fig. 14.20 Comparison of the momentary values of the mass efficiency obtained for bi-layered sets of the filters containing various nanofibrous media

Fig. 14.21 Comparison of the momentary values of the mass efficiency obtained for triple-layered sets of the filters containing various nanofibrous media

However, rapid clogging of the frontal nanofibrous filter in the bi-layered configuration results in much faster increase of the pressure drop in time compared to that one for the triple-layered system. As a consequence of a quick transition from depth to surface filtration, the momentary quality factors (based here on the instantaneous total mass efficiency and the momentary pressure drop) for the bi-layered sets rapidly drop after some time, while the QF for the triple-layered systems, although

272

A. Podgo´rski

Fig. 14.22 Comparison of the momentary values of the pressure drop obtained for bi-layered sets of the filters containing various nanofibrous media

Fig. 14.23 Comparison of the momentary values of the pressure drop obtained for triple-layered sets of the filters containing various nanofibrous media

initially being lower, decreases much slower during the prolonged loading, c.f., Figs. 14.24 and 14.25. This clearly demonstrates that nanofibrous media should be screened from larger particles by a thicker, microfibrous facial filter if

14 Protection of the Respiratory System Against Nanoparticles Inhalation

273

Fig. 14.24 Comparison of the quality factors obtained for bi-layered sets of the filters containing various nanofibrous media

Fig. 14.25 Comparison of the quality factors obtained for triple-layered sets of the filters containing various nanofibrous media

a polydisperse aerosol containing micrometer-sized particles is to be filtered out. Thus, we recommend strongly a triple-layered solution (micro–nano–micro), as shown schematically in Fig. 14.26. Comparing quality factors for various

274

A. Podgo´rski

triple-layered sets containing different nanofibrous media one can conclude that from the QF standpoint the best ones are these that were composed of the finest fibers, i.e., NL1 and NL2. Thus, we can finally conclude that application of nanofibrous filtering structures is very profitable also in the case of removal of polydisperse aerosols containing large, micrometer-sized particles. However, the results acquired during continuous loading of bi- and triple-layered systems containing nanofibers with solid polydisperse aerosol particles clearly indicate the necessity of the nanofibrous media protection with an external microfibrous cover in order to avoid too fast clogging of the filter and formation of the cake on the nanofibrous layer surface, which would be difficult to remove without a potential damage of a delicate nanofibrous structure.

Fig. 14.26 Concept of a triple-layer structure: External microfibrous layer (1) with large pores and a high porosity, relatively thick – to collect big particles and to protect the next layer against clogging; Middle nanofibrous layer (2) to remove precisely most penetrating submicrometer and nanometer-sized particles; Rear backing layer (3) made of micrometer-sized fibers, thinner but more densely packed than layer (1) acting as a support to assure a proper mechanical strength of the entire filter

14 Protection of the Respiratory System Against Nanoparticles Inhalation

14.7

275

Filtration of Nanoparticles in Polydisperse Filters

Research into filtration of nanoparticles is relatively new and experimental data are scarce and sometimes contradictory. While Heim et al. (2005) claimed that there is no measurable deviation from the classical theory of depth filtration even for particle sizes as small as 2.5 nm, a significant discrepancy between this theory and experiments was observed by Wang et al. (2007) for particles smaller than 20 nm, and this discrepancy increased with a particle diameter decrease. Similar phenomenon of a lower collection efficiency of nanoparticles measured experimentally in comparison to the classical theory predictions was reported by Podgo´rski et al. (2007). One possible reason of such discrepancy might be a hypothetical phenomenon of thermal rebound and thermal re-entrainment of nanoparticles. Another possible explanation of a higher penetration of nanoparticles measured experimentally than this one predicted theoretically based on the classical theory may be segregation of flow in polydisperse filters and improper use of a mean fiber diameter for filters made of fibers with various sizes. The fully segregated flow model (FSFM) to describe penetration of nanoparticles through a polydisperse fibrous filter was proposed by Podgo´rski (2009) and it was solved numerically in the case of log-normal distribution of fiber diameter. This solution is illustrated in Fig. 14.27, where penetration, PFSFM, through a polydisperse filter calculated

Fig. 14.27 Relationship between the penetration, PFSFM, through a polydisperse filter calculated using the FSFM and the penetration, Pg, calculated for the geometric mean fiber diameter. Numerical results for several values of the geometric standard deviation of the fiber size distribution

A. Podgo´rski

276

according to the FSFM is plotted vs the penetration, Pg, calculated as for a monodisperse filter consisting of fibers with the diameter equal to the geometric mean fiber diameter of the polydisperse filter in question. This solution is presented for several values of the geometric standard deviation of the fiber size distribution, sgdF, between 1.1 and 2.0. As can be seen in Fig. 14.27, the model formulated predicts that the penetration of nanoparticles for a polydisperse filter is always higher than the penetration Pg calculated on the basis of the geometric mean fiber diameter, dFg. It can also be shown that the same is true if another mean fiber diameter (e.g., the arithmetic mean diameter or the pressure drop equivalent diameter) is used instead of dFg. The relative difference between PFSFM and Pg increases with decrease of Pg, which means that the effect of filter polydispersity is more and more pronounced when the nanoparticles diameter decreases (since Pg decreases then). Obviously, for a fixed particle diameter and for a given mean fiber diameter, the penetration through a polydisperse filter will be higher for a more polydisperse filter (i.e., for greater values of sgdF).

14.8

Summary

Fibrous filters play a key role in protecting people against inhalation of harmful particulates, including nanoparticles. Because of a complicated filter structure and of the complexity of the phenomena of transport and deposition of aerosol particles in a filter, the fibrous media must be designed and tested with a special care. First, one should bear in mind that the classical single fiber theory was developed for an idealized case of spherical aerosol particles and a homogeneous filter, composed of fibers with the same diameters distributed uniformly in a space. Additionally, it assumes that all deposition mechanisms can be accounted for separately and the short range interactions counterbalance. All these assumptions may be a source of serious errors. Generalized Brownian dynamics approach described in this chapter allows one to avoid many of doubtful simplifications. It was shown that a considerable change of the most penetrating particle size is expected in the case of using the most promising filtering media, either nanofibrous or electret ones, thus limitation of testing the filters against 300 nm particles, which are believed to be the most penetrating ones in standard mechanical filters, is insufficient and it may lead to erroneous conclusions. Also, special attention must be paid to filtration of nonspherical particles, e.g., fractal-like nanoaggregates, which behave in a different way than the solid spheres. Finally, one should put special emphasis on filtration of nanoparticles in polydisperse filters, since the classical theory applied to a mean fiber diameter may underpredict nanoparticles penetration even a few orders of magnitude. It should be also borne in mind that in most instances nanoparticles constitute only a fraction of a polydisperse aerosol to be filtered out. Under such circumstances multi-layer fibrous structures are recommended to avoid rapid clogging and an unfavorable fast increase of the filter resistance.

14 Protection of the Respiratory System Against Nanoparticles Inhalation

277

Acknowledgements This work was supported by Polish Ministry of Science (project PBZMEiN-3/2/2006). The author expresses gratitude to Prof. Leon Gradon´, Dr. Anna Bałazy and Anna Jackiewicz, M.Sc., for fruitful cooperation.

References Bałazy A, Podgo´rski A (2007) Deposition efficiency of fractal-like aggregates in fibrous filters calculated using brownian dynamics method. J Colloid Inter Sci 311:323–337 Bałazy A, Toivola M, Reponen T, Podgo´rski A, Zimmer A, Grinshpun SA (2006) manikin-based performance evaluation of N95 filtering-facepiece respirators challenged with nanoparticles. Ann Occup Hyg 50:259–269 Brown RC (1981) Capture of dust particles in filters by line-dipole charged fibers. J Aerosol Sci 12:349–356 Ermak DL, Buckholz H (1980) Numerical integration of the langevin equation: Monte Carlo simulation. J Comput Phy 35:169–182 Heim M, Mullins BJ, Wild M, Meyer J, Kasper G (2005) Filtration efficiency of aerosol particles below 20 nanometers. Aerosol Sci Technol 29:782–789 Hinds WC (1999) Aerosol technology. Properties, behavior, and measurement of airborne particles. Wiley, New York Kim JC, Otani Y, Namiki N, Kimura K (2005) Initial collection performance of resin wool filters and estimation of charge density. Aerosol Sci Technol 39:501–508 Lee M, Otani Y, Namiki N, Emi H (2002) Prediction of collection efficiency of high-performance electret filters. J Chem Eng Japan 35:57–62 Otani Y, Emi H, Mori J (1993) Initial collection efficiency of electret filter and its durability for solid and liquid particles. J Chem Eng Japan 11:207–247 Podgo´rski A (2001) Brownian dynamics – II. Algorithms for stochastic simulations of a solid spherical aerosol particle motion near a solid wall. J Aerosol Sci 32(Suppl. 1):S713–S714 Podgo´rski A (2002) On the transport, deposition and filtration of aerosol particles in fibrous filters: selected problems. The Publishing House of the Warsaw University of Technology, Warsaw Podgo´rski A (2009) Estimation of the upper limit of aerosol nanoparticles penetration through inhomogeneous fibrous filters, J Nanoparticle Res 11:197–207 Podgo´rski A, Bałazy A, Gradon´ L (2006) Application of nanofibers to improve the filtration efficiency of the most penetrating aerosol particles in fibrous filters. Chem Eng Sci 61:6804–6815 Podgo´rski A, Balazy A (2008) Novel formulae for deposition efficiency of electrically neutral, submicron aerosol particles in bipolarly charged fibrous filters derived using Brownian dynamics approach. Aerosol Sci Technol 42:123–133 Podgo´rski A, Bałazy A, Gradon´ L (2007) Nano-scale aspects of aerosol filtration in fibrous filters. Chem Process Eng 28:615–627 Wang J, Chen DR, Pui BYH (2007) Modeling of filtration efficiency of nanoparticles in standard filter media. J Nanoparticle Res 9:109–115

Chapter 15

Overview and Discussion Jan C.M. Marijnissen, Leon Gradon´, and Bob W.N.J. Ursem

This book is a compilation of chapters, written by the participants of a workshop on all the aspects from the origin/production of nanoparticles till their interaction with the lungs and their toxic/therapeutic effects. The workshop took place May 30 and 31, 2008 in the Jablonna Palace near Warsaw in Poland. The isolation from the rest of the world resulted in a very intensive scientific interaction between the participants. This, combined with the beautiful ambiance of the Jablonna Palace, the fantastic Polish hospitality and the fine music/theatre evening event, resulted in a very constructive workshop First of all the origin of nanoparticles in ambient air is discussed. Important sources of these particles are combustion processes, in particular diesel engines and biomass combustion (Burtscher). Many other sources have a significant contribution to the particle mass (PM10), but do not emit high concentrations of ultrafine particles. Optimization of the combustion led to a strong reduction in mass emissions, but the number concentration is only slightly reduced. Filters reduce the mass and initially the number greatly. However when the mass concentration is reduced due to filtration, the number concentration increases significantly by nucleation after the filter (Kulmala). Secondary particle formation plays here a role. Of course there are many more occasions that secondary particles are formed in the atmosphere from gaseous precursors. Panelists underlined that these observations give an indication for the design of filter structures for the construction of particulate matter separators and personal respirators. Another remark was that all the

Jan C.M. Marijnissen (*) Faculty of Applied Sciences, Delft University of Technology Delft, Netherlands e-mail: [email protected] L. Gradon´ Faculty of Chemical and Process Engineering, Warsaw University of Technology, Warynskiego 1, 00–645, Warsaw, Poland Bob W.N.J. Ursem Faculty of Applied Sciences, Delft University of Technology Delft, Netherlands

J.C.M. Marijnissen and L. Gradon´ (eds.), Nanoparticles in Medicine and Environment, DOI 10.1007/978-90-481-2632-3_15, # Springer ScienceþBusiness Media B.V. 2010

279

280

J.C.M. Marijnissen et al.

knowledge, built up for especially health aspects, on particle formation and particle characteristics can also be and has to be used in models on climate change. An in depth study on combustion related aerosol particles accordingly demonstrates that combustion particles, such as from engine exhaust, generally fall into two modes (Maricq). There is a mode of carbonaceous particles (soot) ranging from 10 to 300 nm in diameter (accumulation mode) often accompanied by a mode of 2– 20 nm spherical particles (nucleation mode). Soot mode TEM images reveal a progressively more fractal-like structure with increasing size, resulting from primary soot particles. Soot particles exhibit extensive electrical charge, evenly balanced between positive and negative, that provides a signature of their high temperature origin. The nucleation mode particles in engine exhaust predominantly arise from liquid sulfuric acid/hydrocarbon droplets formed as the exhaust dilutes and cools in the atmosphere or sampling system. However some engine conditions produce a solid nucleation mode, the particles of which are electrically charged, suggesting they are formed already during combustion. During the roundtable discussion it was argued that the (electrical) mobility of electrons is much higher than for positive ions, giving rise to the question why the electrical charge on soot particles is evenly balanced between positive and negative. This point still needs clarification. There is also an increasing activity in the production of nanoparticles e.g. for catalytic and medical purposes. And although these nanoparticles are aimed to be used in a very positive way, care should be taken during production, handling and use to prevent dangerous health situations. (Marijnissen, e.a.). For the production of medicine nanoparticles by aerosol processes, Electro HydroDynamic Atomization (EHDA), also called Electrospraying, might be the most adequate method (Marijnissen e.a., Chen/Pui). EHDA is able to produce monodisperse droplets, and after drying particles, from nanometers till several micrometers, from many different precursor solutions. EHDA is an atomization method, using an electric field. Particles with diameters less than 10 nm can be produced in this way. Examples of medical nanoparticles produced by EHDA are taxol particles, controlled release particles of taxol in PLGA, high porosity particles, elongated particles and nanotubes. Other medical applications of EHDA are medical device coating, guided neuron growth, gene transfection and research in nanoparticle toxicity. Two concentric nozzles can be used for the encapsulation of pharmaceutical particles. By using two sprays of opposite polarity, nano- or micro-reactors can be created or encapsulation can take place (bipolar coagulation). It is also possible in this way to apply medicine nanoparticles on a carrier. During the roundtable discussion it was questioned how the EHDA technique can be scaled up to industrial quantities. It proves that the only way to increase production is out-scaling, by using parallel nozzles. Much research on this topic is in progress. A first requirement considering all aspects of nanoparticles is the ability to measure their size. At this moment no standard measurement procedures are available, moreover, there is even no standard definition regarding nanoparticles. The value of 100 nm is suggested as the upper size limit for airborne nanoparticles. Real-time measurements, utilizing electrical mobility of charged particles seem to

15 Overview and Discussion

281

be among the best and convenient techniques currently available. Since the surface area of nanoparticles proves to be a critical indicator linked with the seditious response of organisms, real time instruments to measure the surface of nanoparticles have been developed. A new electrical aerosol mass analyzer provides a tool to determine the mass of nanoparticles and in combination with a differential mobility analyzer it allows the assessment of the effective density of measured nanosized agglomerates. High sensitivity mass spectrometry techniques, combined with aerosol methods open new ways of describing airborne nanoparticles, including such species as dendrimeres, macromolecules and viruses. During the roundtable discussion it became clear that more measuring techniques are under consideration and that e.g. Photon Correlation Spectroscopy, an in-line measuring technique, has great promises for measuring airborne nanoparticles. To be able to predict and understand the effect of nanoparticles upon inhalation, first of all the deposition behavior of the particles must be known (Moskal. e.a.). The main mechanism of nanoparticle deposition in the respiratory system is the Brownian diffusion. The flux of particles towards the wall of the respiratory system strongly depends on the local and momentary flow structure during the breathing cycle. For any part of the respiratory tract the transition between inspiratory and expiratory parts of the breathing strongly affects the flow structures. The thickness of the momentum boundary layer significantly decreases, causing the enhancement of diffusional particle deposition. Besides this effect the morphology of nanostructured particles (diesel and medical) influences the mobility of particles. The approximation of the real particle shape, through its equivalent sphere diameter can also cause significant discrepancies in deposition for some particles of real shapes. Both effects cause the deposition pattern of diffusional particles and the position of “hot spots” of deposits to be significantly different than for simplified models, assuming a uniform, steady-state flow during breathing and spherical particles. In addition, the role of lung surfactant in the hydrodynamics of pulmonary fluids in relation to the clearance of deposited particles is important. This essential function of the surfactant is assured by its innate surface activity, i.e. the ability to optimally change the surface tension during breathing cycles. Using specialized experimental in vitro techniques, the influence of nano-sized and nano-structured particles on dynamic surface activity of the surfactant has been demonstated. Due to a hypothetical preferential adsorption of surfactant molecules on the surface of nanoparticles, the surfactant function was reduced, which suggests a decrease of pulmonary clearance rate and a prolonged retention of deposited particles in the lung. Panelists of the round table discussion underlined the relation of the results of this presentation with those presented by Kreyling/Geiser and Pirozynski/Baughman. Also the importance to take the hygroscopic properties of the particles into account, was discussed. With good deposition models, dosimetry of inhaled nanoparticles becomes more realistic (Kreyling, Geiser). Risk assessment includes hazard characterization, identification and risk evaluation. Dosimetry, a prerequisite for toxicology, epidemiology and risk assessment is a first step in risk identification. Emphasis is put on the results of the research for investigation the deposition, retention and clearance

282

J.C.M. Marijnissen et al.

of nanoparticles as well as their translocation to secondary organs in order to contribute to further risk assessment by inhaled nanoparticles. The concept of quantitative biodistribution of inhaled matter with explanation of used techniques of the dosimetry and ultrastructured analyses was presented. There is evidence that the translocation and accumulation of nanoparticles in secondary organs depends on their size. For example, inhaled 80 nm iridium particles were shown to translocate about one order of magnitude less than 20 nm iridium particles, including accumulation in the skeleton and soft tissue. It was also shown that 20 nm particles of carbon translocate significantly slower than iridium particles of the same size. Beside the size and material other particle characteristics like surface charge and surface structure significantly influence the biokinetics. Results of this presentation were discussed during the round table meeting in relation to the health effects and mechanisms on deposition and interaction of nanoparticles with the lung surfactant. There was also some discussion on the effect on deposition and on health effects of electrical charge on particles. The effect depends among others on the amount of charge, and so also on the electrical mobility of the particles. However more research is needed on this. This brings medical quantum chemistry in scope (Uvarova, Broclawik). New, but becoming very important and invaluable in medicine nanotechnology and biology are the use of quantum chemical calculations. It can be used to understand e.g. enzyme mechanisms, hydrogen bonding, ligand binding and other fundamental processes. Benefical effects of drugs are assumed to come from molecular recognition and binding of ligands to the active sites of specific targets, such as enzymes, receptors and nucleic acids. Medical quantum chemistry can be used for molecular modeling of particle properties and interaction and of mechanisms for biomedical processes at an atomistic level, including e.g. toxicity of silica dust particles. Another new approach in the field is phenomenological modeling of particle transport in human tissues. Here molecular mechanics and dynamics are used. It was emphasized during the round table discussion that the specific interaction of the nanoparticles with the organism through their interaction with cells is the result of the activity of electrons in the configuration of the molecules in their nanosized clusters, which is significantly different than it is in the case of coarse particles. It was also discussed that the effect of a very small, non spherical particle, deposited e.g. in the lung, depends on the orientation of the particle. Although the understanding of the functioning of the human body at the molecular and nano- level has improved a lot, the effective treatment of severe and chronic diseases follows only slowly (Baugham, Pirozynski). Understanding the role of nanoparticles is still a big challenge. This understanding can also be used in therapy, using medicine nanoparticles. Toxicological studies are used to reveal the mechanisms of the adverse effects on the respiratory and cardiovascular systems. Primary combustion nanoparticles, as from diesel engines, cause pulmonary inflammation in humans and animals. They have an enhanced ability to generate highly reactive free radical molecules, that damage and activate lung cells to produce pro-inflammatory mediators. Long term studies on inhalation of ultrafine particles show chronic pulmonary inflammation, increased chemokine expression,

15 Overview and Discussion

283

epithelial cell hyperplasia, pulmonary fibrosis and lung tumors, while short term studies on inhalation ultrafine carbon black show more inflammation, than for larger fine carbon black. Biological effects are not related to particle mass but to size and surface area (biologically available area), particle surface chemistry, biodegradability, number, shape and solubility. After deposition in the respiratory system, the nanoparticles are submerged in the lining fluid. Soluble particles are dissolved, insoluble particles undergo translocation, leading to close association with the epithelial cells and cells of the host defense system, possibly inducing inflammation. Inflammation and systemic effects are the driving force of cardiovascular effects. The results of the health studies of nanoparticles, can be used in the research on the inhalation of medicine nanoparticles for both systemic and local applications, including sustained release and targeting specific cells and organs. However, even for medical nanoparticles care should be taken for possible specific nano-effects, as fate or toxicity. As specific example of the effect of nanoparticle inhalation on health, cigarette smoke and diesel exhaust are discussed (Hiemstra). Both cigarette smoke and diesel exhaust cause lung inflammation and increased susceptibility to respiratory infection. The epithelium that lines the airways is considered as an important element in the innate defense against these respiratory infections. It is both a target for inhaled respiratory pathogens, as well as for cigarette smoke and diesel exhaust. The deposition of diesel exhaust particles in the alveoli is especially problematic because of the extremely thin air–blood barrier in this part of the lung. The effect of the exposure to (nano) particles by children is different than for adults (Schuepp). This starts already before a child is born. It is expected that children have a reduced ability to detoxify possible effects of nanoparticles, so injury can occur even at doses below the no-effect limit for adults, and also repair functions are restricted in early life. The development of different systems can be impaired. There prove to be many clinical and respiratory effects after prenatal or postnatal particle exposure. Children seem to be the most vulnerable group with regard to harmful effects of air pollution. It should also be understood that children interact with their environment differently from adults. Children e.g. tend to spend more time outdoors, often with heavy physical activity, so increasing ventilation rates. However nanoparticles might also have a beneficial potential. Nanoparticles represent a promising therapeutic option to treat pediatric diseases, as nanoparticles reach all regions of the lung. By targeting specific lung regions local side effects could be diminished and cell targeting by nanoparticles could lead to a higher efficiency in inhalation therapy. Of course all effects of administration of nanoparticles have to be considered. Medicine aerosol inhalation might, besides for the treatment of pulmonary diseases, be advantageous for the treatment of systemic diseases (Haussermann). Especially for high molecular weight compounds the method seems to have a great potential. The alveolar region seems to be the most appropriate target because of the enormous alveolar surface area, and the thin alveolar epithelium. Recently new aerosol delivery systems have been developed, enabling to generate aerosols with a defined and optimized aerosol particle size and improved way of administration, so

284

J.C.M. Marijnissen et al.

overcoming some of the former objections of this type of delivery. Several developments are under way such as packing pharmaceutical nanoparticles into carriers, in this way strongly affecting physical and pharmacological properties of the inhaled compounds. However there are still a lot of open questions, especially on potential bad side effects. It is clear that in certain environmental and production situations, protection against inhalation of nanoparticles is necessary. Filters seem to be the proper approach. However, the filtration behavior of filters for nanoparticles is not correctly described by the conventional theories, where e.g. the different filtration mechanisms are treated as being additive, and filter fibers are supposed to have equal diameters and are uniformly distributed in space (Podgorski). A much better approach is the coupling of the stochastic and deterministic deposition mechanisms. This approach is used to simulate the filtration for both spherical nanoparticles and for fractal-like nanoaggregates. The calculations, verified by experimental results, show a considerable change of the most penetrating particle size, in case of using the most promising filtering media, either nanofibrous or electret. Also fractal-like nanoaggregates behave differently from solid spheres. It also proves that in case of filtration of nanoparticles by fibrous filters of non-homogeneous fiber diameter, the particle penetration is much higher than calculated with the classical theory, using a mean fiber diameter. The overview above includes the presentations and roundtable discussions. It should be emphasized that the fruitful discussions during the workshop sessions significantly enriched the scientific atmosphere and introduced new aspects in all particular problems arisen. Very detailed questions, as posed during the discussions, which do not really contribute to the overall picture, are not included.

Index

A

D

Aerosol, 1–5, 10, 14, 20, 21, 23, 24, 27, 30, 31, 33, 35, 40, 42–45, 48, 54, 59, 66, 67, 77–87, 91–109, 113, 114, 119, 120, 122–124, 126–130, 132–135, 137, 139, 141, 153, 154, 157, 158, 161, 163, 178, 198, 199, 220–222, 228–231, 240, 245, 251, 252, 256–260, 263, 266–274, 276, 280, 281, 283 Aerosol therapy, 119, 220 Airways inflammation, 192, 203, 206, 207, 219

Deposition, 24, 42, 44, 50, 51, 61, 62, 72, 103, 105, 107, 113–142, 145, 146, 155–157, 159, 183, 188, 189, 195, 198, 208, 212, 213, 220, 221, 229–231, 240, 243, 252–259, 261, 263–265, 276, 281–284 Detection efficiency, 78, 79, 92 Developmental biology, 213 Diesel emissions, 1, 5, 7–9, 12, 13, 15, 22, 29, 32, 206 Diesel exhaust, 6, 13, 25, 26, 31–33, 119, 164, 192, 203–209, 212, 215, 217, 283 Diesel particles, 5–9, 14, 15, 25–28, 32, 33, 119, 121, 122, 192, 207, 209, 212, 263, 283 Dosimetry, 145–167, 281, 282 Droplet, 6, 21, 39–46, 48–52, 54, 55, 59, 62–65, 67, 69, 79, 81, 95, 221, 228, 280 Drug delivery, 54, 59, 62, 119, 140, 141, 163, 198, 199, 211, 212, 220, 221, 228 Drug reformulation, 65–66

B Biokinetics, 145–147, 163, 165, 166, 282 Bipolar coagulation, 40, 49–52, 280 Brownian dynamics, 122, 252, 255–261, 263, 276

C Cigarette smoke, 197, 203–209, 243, 283 Clearance, 138–141, 146, 157–162, 166, 204, 207, 208, 281 Clusters, 44, 45, 78–86, 116, 121, 122, 134, 178, 179, 183, 264–266, 282 Coating, 50–52, 59, 60, 69–71, 163, 280 Combustion particles, 9, 12, 21, 33, 34, 280 Controlled drug release, 69–70 Corona discharge, 41, 52, 69, 72, 96–99, 101, 109

E Effective particle density, 27, 106 Electret filter, 251, 259–263, 276, 284 Electrohydrodynamic atomization (EHDA), 39–56, 280 Electrospray, 59–72, 97, 98, 100 Enzymatic reaction mechanisms, 180 Epithelial cells, 157, 159, 161, 189, 190, 192, 204–206, 208, 213, 216, 217, 283 Exposure differences to adults, 212

285

286

F Fetal origin of later disease, 220 Fibrous filter, 251, 252, 254, 259, 263, 266, 275, 276, 284 Filtration, 77–87, 251, 252, 254, 259–276, 279, 284 Formation and growth of nanoparticles, 7, 71, 78 Fractal-like aggregates, 20, 26, 28, 121, 122, 133, 135, 263–265

G Gene transfection, 59, 60, 62, 71, 221, 280 Guided neuron growth, 71, 280

H Health effects, 1, 19, 34, 35, 67, 72, 108, 130, 132, 139, 141, 142, 145, 163, 167, 187–199, 206, 282

I Improved inhalation, 113, 231 Inhalation, 46, 54, 59, 109, 113–142, 146, 147, 155–161, 164–166, 178, 183, 187–199, 203, 206, 211, 221, 222, 229–246, 251–276, 281–284 Innate immunity, 203–209 Insulin, 227–246

L Liposomes, 228, 229 Lung surfactant, 138–140, 189, 281, 282 Lung tract, 116–119, 137, 145–147, 150, 155, 178, 187, 208, 213, 221, 229, 243, 281 Lymph nodes, 158–162

Index

Medicinal quantum chemistry, 174–176, 184, 282 Medicine, 39–56, 60, 65, 68–70, 72, 77, 93, 148, 174, 187, 212, 280, 282, 283 Microspheres, 235, 236 Modeling, 116, 118, 119, 122, 126–137, 141, 142, 173, 174, 176, 177, 179, 180, 282 Molecular mass, 95, 106, 107 Molecular modeling, 176, 181, 282

N Nano-medicine, 60, 163, 167, 187 Nanoaggregate, 251, 263–266, 276, 284 Nanofibers, 49, 189, 266–268, 274 Nanomass measurement, 101–107 Nanoparticle charging, 96–98, 101 Nanoparticle toxicity, 66–68, 163, 191–193, 195, 198, 199, 280, 282 Nanoparticles, 1–15, 21, 23, 39–56, 59, 66–68, 77–87, 91–109, 113–142, 145–167, 178, 179, 187–199, 211–222, 228, 229, 251–276, 279–284 Nanosize measurement, 281 Nanostructured particles, 113, 130, 133, 137, 141, 281 New therapeutic options, 183, 187, 212, 222, 283 Non-steady flow, 128, 129, 133, 135 Nucleation, 6, 7, 10, 21, 23, 24, 29, 30, 33–35, 77–79, 81, 82, 84–87, 279, 280 Nuclei mode, 21, 24, 29–34, 79

O Out scaling, 40, 54–56, 280

P M Mass spectrometry, 6, 14, 32, 59, 102, 103, 105–107, 281 Medical application, 49, 59–72, 280 Medical device coating, 70–71, 280 Medical nanoparticles, 42, 46, 54, 280, 283

Particulate matter, 1–4, 11, 15, 19, 30, 35, 100, 114, 218–220 Phagocytosis, 218 Polydispersity, 276 Production, 1–15, 39–56, 59, 61, 63, 84, 86, 97, 119, 120, 145, 147, 182, 192,

Index

205–208, 211, 227, 228, 230, 279, 280, 284

287

Surface monitoring, 100–101 Surface tension, 40, 41, 64, 66, 72, 138–140, 281

R Relocation, 157–158, 160, 219 Respiratory diseases, 192, 198, 199, 203, 206, 217, 219, 221, 222, 243 Retention, 146, 148, 157–162, 164, 281 Rich flames, 20, 21, 25, 26, 33

S Scaling, 41, 45, 54, 61–63, 155 Single- and dual- capillary electrospray, 60–66, 68–72 Small particles, 10, 28, 47, 48, 63, 92, 178, 188–190, 211, 212, 220 Soot, 3, 6, 19–21, 24–34, 105, 106, 121, 140, 263, 280

T Therapy, 71, 187, 195–199, 231, 232, 240, 243, 245, 246, 282, 283 Toxicity, 66–68, 163, 183, 191, 193–195, 198, 199, 245, 280, 282, 283 Translocation, 146, 157, 158, 160–167, 190, 282, 283 Tribo, 51

W Welding emissions, 12, 67 Wood combustion, 8–11, 14, 15