Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Reactive uptake of O₃ and N₂O₅ on organic mixtures and inorganic solutions coated with organic monolayers Cosman, Lori Marie 2008

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2008_fall_cosman_lori_marie.pdf [ 9.99MB ]
Metadata
JSON: 24-1.0061818.json
JSON-LD: 24-1.0061818-ld.json
RDF/XML (Pretty): 24-1.0061818-rdf.xml
RDF/JSON: 24-1.0061818-rdf.json
Turtle: 24-1.0061818-turtle.txt
N-Triples: 24-1.0061818-rdf-ntriples.txt
Original Record: 24-1.0061818-source.json
Full Text
24-1.0061818-fulltext.txt
Citation
24-1.0061818.ris

Full Text

REACTIVE UPTAKE OF 0 3 AND N 2 0 5 ON ORGANIC MIXTURES AND INORGANIC SOLUTIONS COATED WITH ORGANIC MONOLAYERS  by  LORI MARIE COSMAN  B.Sc. Honours (Chemistry), Acadia University, 2003  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  The Faculty of Graduate Studies (Chemistry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  MAY 2008  © Lori Marie Cosman, 2008  11  ABSTRACT  Atmospheric particles play a crucial role in climate, visibility, air pollution, and human health. Reactions between gas-phase molecules and particles (heterogeneous reactions) affect not only the particle composition and morphology, but also the composition of the atmosphere. This thesis investigates the heterogeneous chemistry of organic mixtures and inorganic solutions coated with organic monolayers as proxies for atmospheric particles. The first topic of interest was the reaction between N20 5 and aqueous inorganic solutions coated with organic monolayers. The goal of this work was to better understand how organic monolayers on aqueous particles affect the mass transport and kinetics of N20 5 uptake by aqueous aerosols, and consequently what effect the monolayer can have on predicted concentrations of N20 5 in the atmosphere. To investigate heterogeneous reactions of inorganic solutions coated with an organic monolayer a new rectangular channel flow reactor was developed. This newly developed flow reactor was described in detail and validated. Subsequently, the new flow reactor was used to study the reactive uptake of N205 on sulfuric acid solutions in the presence of a variety of 1- and 2-component monolayers with varying functional groups, solubilities, chain lengths, surface pressures, and molecular surface areas. Reactive uptake of N20 5 on aqueous sulfuric acid solutions was found to correlate most strongly with the molecular surface area or packing density of the monolayer. These results provide a good foundation for determining the influence of monolayers on heterogeneous reactions in the atmosphere, and highlight the need for characterization of monolayer surface properties of organic monolayers present on atmospheric particles. The second topic of interest was reactions between 0 3 and proxies for meat cooking aerosols with the goal to better understand the effect of the phase and microstructure of the mixtures on the lifetime of oleic acid (OA) in atmospheric particles. The reactive uptake of 03 was approximately 1 order of magnitude slower on binary solid-liquid mixtures and multicomponent mixtures that closely represent compositions of meat-cooking aerosols compared to the liquid solutions. Lifetimes up to 75 min were obtained for these mixtures.  iii  TABLE OF CONTENTS  ii  Abstract ^ Table of Contents ^ List of Tables ^  iii viii  List of Figures ^ Acknowledgements ^  xxiii  Dedication ^  xxiv  Co-Authorship Statement ^  xxv  1. Background and Motivation: Gas-Particle Interactions in the Troposphere ^ 1 1.1 Aerosols and Aerosol Particles ^  1  1.2 Importance of Aerosols ^  1  1.3 Sources, Atmospheric Transformations, and Sinks of Tropospheric Particles ^2 1.3.1 Sources of Tropospheric Particles ^  2  1.3.2 Atmospheric Transformations and Sinks of Tropospheric Particles ^2 1.4 Composition of Tropospheric Particles ^  4  1.4.1 Aqueous Inorganic Particles Coated With Organic Monolayers ^5 1.4.2 Organic Particles Produced From Charbroiling and Meat Cooking Operations ^ 1.4.3 Organic Compounds Used in This Thesis ^  7 8  1.5 Heterogeneous Chemistry ^  10  1.6 Reactive Uptake Coefficient ^  11  1.7 N205 Chemistry ^  11  1.7.1 N205 Chemistry in the Troposphere ^  11  1.8 Theory of Monolayers ^  15  1.8.1 Organic Surfactants ^  15  1.8.2 Insoluble Monolayers ^  17  1.8.3 Soluble Monolayers ^  20  iv  2.  1.9 Thesis Overview ^  21  1.10 References ^  23  A Novel Flow Reactor for Studying Reactions on Liquid Surfaces Coated by Organic Monolayers: Methods, Validation, and Initial Results ^29 2.1 Introduction ^  29  2.2 Experimental ^  30  2.2.1 New Flow Reactor ^  30  2.2.2 Details of the Reactive Uptake Experiments ^  33  2.2.3 Organic Monolayer Preparation ^  34  2.2.4 Computational Fluid Dynamics Simulations ^  34  2.2.5 Mathematical Framework to Derive the First-Order Wall Loss Rate Coefficient, kw ist , and the Reactive Uptake Coefficient, y ^40 2.2.6 Chemicals ^ 2.3 Results and Discussion ^  43 43  2.3.1 Validation of the Flow Cell and the Data Analysis Procedure ^43 2.3.2 N 2 0 5 Reactive Uptake Measurements on Sulfuric Acid Solution Coated With  3.  an Organic Monolayer ^  49  2.4 Atmospheric Implications ^  50  2.5 Conclusions and Summary ^  51  2.6 References ^  53  N205 Reactive Uptake on Aqueous Sulfuric Acid Solutions Coated With Branched and Straight-Chain Insoluble Organic Surfactants ^59 3.1 Introduction ^  59  3.2 Experimental ^  62  3.2.1 Flow Reactor and Experimental Conditions for the Reactive Uptake Measurements ^ 3.2.2 Preparation of the Organic Monolayers ^  62 66  V  3.2.3 Measurements of the Surface Pressure and the Surface Area Occupied by Each Surfactant Molecule in the Monolayer ^  67  3.2.4 Further Characterization of the Organic Monolayers in the Flow Reactor ^70 3.2.5 Chemicals ^ 70 3.3 Results and Discussion ^  70  3.3.1 Properties of the Monolayers ^  70  3.3.2 Reactive Uptake Coefficients on Aqueous Solutions Covered With Organic Monolayers ^  73  3.3.3 Correlation Between Reactive Uptake Coefficients and Carbon Chain Length, Surface Pressure, and Molecular Surface Area ^ 79 3.4 Atmospheric Implications ^  86  3.5 Conclusions and Summary ^  87  3.6 References ^  89  4. Reactive Uptake of N205 on Aqueous H2SO4 Solutions Coated With 1Component and 2-Component Monolayers ^  95  4.1 Introduction ^  95  4.2 Experimental ^  98  4.2.1 Rectangular Channel Flow Reactor and Determination of Reactive Uptake Coefficients ^  98  4.2.2 Preparation of 1-Component Monolayers for the Reactive Uptake Measurements ^  101  4.2.3 Measurements of Surface Pressure-Area (Ir-A) Isotherms of the 2-Component .  Monolayers ^  102  4.2.4 Preparation of the 2-Component Monolayers for the Reactive Uptake Measurements ^ 4.2.5 Chemicals ^ 4.3 Results and Discussion ^  103 104 104  4.3.1 N2O 5 Reactive Uptake in the Presence of a 1-Component Monolayer (1-Octadecanol) ^  104  4.3.2 ,r-A Isotherms and Miscibility of 2-Component Monolayers ^ 110  vi 4.3.3 N20 5 Uptake in the Presence of a 2-Component Monolayer ^ 117 4.4 Summary and Conclusions and Atmospheric Implications ^ 121 4.5 References ^  5.  123  N205 Reactive Uptake on Aqueous Sulfuric Acid Solutions Containing Dicarboxylic Acid Organic Surfactants ^  130  5.1 Introduction ^  130  5.2 Experimental ^  131  5.2.1 Measurement of Dicarboxylic Acid — H2O — H2SO4 Surface Tensions ^ 131 5.2.2 Flow Reactor and Experimental Conditions for the Reactive Uptake Measurements ^ 5.2.3 Chemicals ^ 5.3 Results and Discussion ^  132 134 134  5.3.1 Surface Tension and Surface Excess of Dicarboxylic Acids in Aqueous Sulfuric Acid at 260 K ^  134  5.3.2 Reactive Uptake of N205 With Aqueous Sulfuric Acid in the Presence of Dicarboxylic Acids at 260 K ^  141  5.3.3 Surface Tension and Surface Excess of Dicarboxylic Acids in Water at 293 K ^  145  5.3.4 Reactive Uptake of N20 5 With Dilute Aqueous Solutions in the Presence of Dicarboxylic Acids at 293 K ^  6.  148  5.4 Conclusions and Summary ^  148  5.5 References ^  150  Reactive Uptake of 03 by Multicomponent and Multiphase Mixtures Containing Oleic Acid ^  155  6.1 Introduction ^  155  6.2 Experimental ^  157  6.2.1 Chemicals ^  162  6.3 Results and Discussion ^  162  vii 6.3.1 Kinetics Experiments ^  162  6.3.2 Reactive Uptake Coefficients of 0 3 on OA ^  164  6.3.3 Reactive Uptake Coefficients of 03 on Liquid Binary Oleic Acid/Myristic Acid Mixtures ^  165  6.3.4 Reactive Uptake Coefficients of 03 on Solid-Liquid Mixtures ^ 169 6.3.5 Reactive Uptake Coefficients of 03 on Solid-Liquid Mixtures That Closely Represent the Composition of Meat-Cooking Aerosols ^ 171 6.3.6 Reactive Uptake Coefficients as a Function of Substrate Age ^ 172 6.4 Atmospheric Implications ^  175  6.4.1 General Atmospheric Implications ^  175  6.4.2 Atmospheric Lifetime Estimates of Oleic Acid ^  175  6.5 Conclusions and Summary ^  178  6.6 References ^  180  7. Concluding Remarks ^  187  7.1 Conclusions ^  187  7.2 Considerations for Future Work ^  190  7.3 References ^  192  A. Appendix ^  193  Appendix ^  196  B.  References ^  198  viii  LIST OF TABLES  Table 1.1: ^  8  Summary of Organic Compounds Used in This Thesis Table 2.1: ^  35  Experimental Conditions Under Which Flow Experiments and Corresponding CFD Simulations Were Conducted' Table 2.2: ^  42  Calculated Diffusion Coefficients of N20 5 in He, N20 5 in H2O (Vapor), and 03 in H e 146-151 Table 2.3: ^  48 Experimentally Obtained Reactive Uptake Coefficients for the Uptake of  N20 5 by Aqueous 60 and 80 wt % H 2 SO 4 Solutions and N2O 5 by Aqueous 80 wt % H2SO4 Solutions Coated With a Monolayer of 1-Octadecanol a Table 2.4: ^  50  Comparison of Measured Reactive Uptake Coefficients of N 2 0 5 on Aqueous H2SO4 and NaC1 Solutions Coated With Organic Monolayers of Different Chain Lengths a Table 3.1: ^  61  Structure and Melting Points of Organic Surfactants Used to Form Monolayers in This Study Table 3.2: ^  71 The Reactive Uptake of N 2 0 5 on Aqueous 60 wt % Sulfuric Acid Solutions at 273 K in the Presence and Absence of Organic Monolayers  Table 3.3: ^  75  A Comparison of Measured Reactive Uptake Coefficients of N2O 5 on  ix Aqueous H2SO 4 and NaC1 Solutions Coated With Straight-Chain Organic Monolayers of Different Chain Lengths Table 4.1:  ^  118  Measured Reactive Uptake Coefficients for N 2 0 5 on 60 wt % Aqueous Sulfuric Acid Solutions at 273 ± 1 K Coated with 1-Octadecanol-Phytanic Acid Monolayers at 21 ± 2 mN m -1 Table 5.1:  ^  142  The Reactive Uptake Coefficient (y) of N20 5 on 40 wt % Sulfuric Acid Solutions at 260 K in the Presence and Absence of Organic Monolayers Table 6.1 .  ^  165  Reactive Uptake Coefficients, 7, for 03 on Pure Liquid and Solid OA Table 6.2: ^  172  Obtained Reactive Uptake Coefficients and the Corresponding OA Lifetimes Given for Different Fast- and Slow-Cooled Solid-Liquid Mixtures Which Closely Represent Meat-Cooking Aerosols'  LIST OF FIGURES  Figure 1.1: ^  3  Atmospheric cycling of aerosol particles.' Figure 1.2: ^  4  Average urban and remote composition (by mass) for fine particulate aerosol particles based on a literature survey by Heintzenberg 24 of several field studies. Figure 1.3: ^  6  Core-shell model of organic coated aqueous atmospheric particles. 14 ' 41 Figure 1.4: ^  7  The composition (by compound class) of the identified organic mass fraction of fine particulate measured from charbroiling and meat cooking particles.  47-  49  Figure 1.5: ^  13  Daytime and night time reaction pathways that convert chemically active nitrogen species to less active HNO3. (adapted from Ravishankara 69 ) Figure 1.6: ^  15  Schematic representation of the dynamic equilibrium established for monolayers at the air-water interface. (adapted from Albrecht et al. 4 ) Figure 1.7: ^  16  A Wilhelmy plate partially submerged in an aqueous solution for measurement of surface pressure. (adapted from Gaines 73 ) Figure 1.8: ^  19  A schematic of a Langmuir trough 74 used to measure a 7r-A isotherm and a  xi typical 7r-A isotherm representative of a long, straight-chain organic surfactant. (adapted from Kaganer et al. 75 ) Figure  2.1:  31  New flow reactor. (A) Top view of the flow reactor without the cover. The main flow is from left to right. The liquid surface in a quartz trough, the movable T-shaped injector, and the corresponding coordinate system are indicated. (B) Side view of the flow reactor including the top cover. The carrier gas inlet, mixing barrier, length of the liquid surface, and corresponding coordinate system are indicated. (C) Cross section of the flow reactor. The width of the liquid surface is given, and the grooves that support the injector are represented. Figure  2.2.  37  Velocity flow profiles along the y-axis derived from CFD simulations for the conditions given in Table 2.1. (a), (b), and (c) represent the results obtained at x = 6, 13, and 20 cm, respectively. Solid diamonds, open circles, open squares, solid circles, solid triangles, and open diamonds represent the flow profiles obtained for conditions 1, 2, 3, 4, 5, and 6 given in Table 2.1, respectively. Solid lines indicate calculations of the flow profile for a rectangular channel given by Solbrig and Gidaspow. I38 Figure  2.3.  38  Velocity profiles across the width of the flow reactor (in the z-axis) derived from CFD simulations for the conditions given in Table 2.1. (a), (b), and (c) represent the results obtained at x = 6, 13, and 20 cm, respectively. Solid diamonds, open circles, open squares, solid circles, solid triangles, and open diamonds represent the flow profiles obtained for conditions 1, 2, 3, 4, 5, and 6 given in Table 2.1, respectively. Solid lines indicate calculations of the flow profile for a rectangular channel given by Solbrig and Gidaspow.138 Figure 2.4: ^  44  Experimentally derived natural logarithms of the gas-phase signals as a  xii function of reaction time (t). Solid circles and solid squares indicate the uptake of 03 by canola oil and the uptake of N205 by aqueous 80 wt % H2SO4 solution, respectively. Dashed lines indicate a linear fit to the data. Figure  2.5:  46  Uptake of 0 3 by canola oil as a function of pressure (a). Open circles indicate y values that have not been corrected for diffusion. Solid circles indicate 7  values that have been corrected for vertical diffusion and diffusion in the direction of bulk flow. The dashed line and gray shading represent the literature value with corresponding uncertainty, I53 respectively. (b) shows the amount of correction when taking diffusion into account. Figure  2.6:  47  Panel (a) shows uptake of 03 by canola oil as a function of flow velocity and Peclet number (Npe ). Open circles and solid circles indicate y values that are not corrected for diffusion and are corrected for vertical diffusion and diffusion in the direction of bulk flow, respectively. The dashed line and gray shading represent the literature value with corresponding uncertainty, I53 respectively. Panel (b) shows the amount of correction due to vertical diffusion and diffusion in direction of bulk flow as a function of velocity. Figure  3.1:  63  (a) A top view sketch of the rectangular channel flow reactor without the aluminum cover. The liquid solution is placed in a quartz trough located inside the flow reactor. (b) A side view of the rectangular channel flow reactor coupled to the chemical ionization mass spectrometer (CIMS). Figure  3.2:  69  Surface pressure — area isotherms for (a) 1-hexadecanol, (b) 1-octadecanol, (c) stearic acid, and (d) phytanic acid performed on aqueous 60 wt % sulfuric acid at 273 K. The horizontal dashed lines represent the surface pressure of each monolayer measured in the kinetic uptake experiments and the vertical dashed lines represent the corresponding molecular surface area at those  surface pressures. The collapse pressures are indicated by the horizontal dotted lines. Figure  3.3:  74  Natural logarithm of the observed N 2 0 5 signal as a function of reaction time. Experiments were performed on aqueous 60 wt % H2SO 4 solutions at 273 K. The lines represent the corresponding linear fits to the data. Open triangles: blank uptake, solid circles: 1-octadecanol, solid triangles: 1-hexadecanol, solid diamonds: stearic acid, solid inverted triangles: phytanic acid. Figure  3.4:  76  Reactive uptake coefficients of N205 on aqueous H 2 SO 4 and NaC1 solutions coated with straight-chain organic monolayers of different chain lengths. Solid circle: 1-octadecanol (this study), solid diamond: stearic acid (this study), solid triangle: 1-hexadecanol (this study), solid inverted triangle: 1butanol (Park et al.), 188 solid pentagon: hexanol (Park et al.), 188 solid star: 1octadecanol (Knopf et al.), 189 solid sideways triangle: hexanoic acid (Thornton and Abbatt), 187 solid square: SDS (McNeill et al.). 181 Panel (a) gives the ordinate as a function of y in the presence of the monolayer. In panel (b) the ordinate is normalized to y for the uncoated solution. Figure  3.5:  80  Reactive uptake coefficients for N205 on organic coated sulfuric acid solutions as a function of carbon chain length. Solid symbols represent straight chain molecules and the open symbols represent the branched molecule. Solid circle: 1-octadecanol (this study), solid diamond: stearic acid (this study), solid triangle: 1-hexadecanol (this study), open circle: phytanic acid (this study), solid inverted triangle: 1-butanol (Park et al.), 188 solid pentagon: hexanol (Park et al.), 188 solid star: 1-octadecanol (Knopf et al.). 189 Panel (a) gives the ordinate as a function of y in the presence of the monolayer. In panel (b) the ordinate is normalized to y for the uncoated solution.  xiv Figure  3.6:  82  Reactive uptake coefficients for N 2 0 5 uptake on organic coated aqueous sulfuric acid solutions as a function of monolayer surface pressure. Solid symbols represent straight chain molecules and the open symbols represent the branched molecule. Solid circle: 1-octadecanol (this study), solid diamond: stearic acid (this study), solid triangle: 1-hexadecanol (this study), open circle: phytanic acid (this study), solid inverted triangle: butanol (Park et al.), 188 solid pentagon: hexanol (Park et al.), 188 solid star: 1-octadecanol (Knopf et al.). 189 Panel (a) gives the ordinate as a function of 7 in the presence of the monolayer. In panel (b) the ordinate is normalized to yfor the uncoated solution. Figure  3.7:  83  Reactive uptake coefficients for N205 uptake on organic coated aqueous sulfuric acid solutions as a function of packing density. Solid symbols represent straight chain molecules and the open symbols represent the branched molecule. Solid circle: 1-octadecanol (this study), solid diamond: stearic acid (this study), solid triangle: 1-hexadecanol (this study), open circle: phytanic acid (this study), solid inverted triangle: butanol (Park et al.), 188 solid pentagon: hexanol (Park et al. ),188 solid star: 1-octadecanol (Knopf et al.). 189 The dashed line represents a sigmoidal fit to the straight chain molecules. This fit was chosen because it gave a reasonable fit to the data, but has no physical meaning. Panel (a) gives the ordinate as a function of yin the presence of the monolayer. In panel (b) the ordinate is normalized to y for the uncoated solution. Figure  3.8:  85  Reactive uptake coefficients for N205 uptake on aqueous solutions and aerosols as a function of packing density. Solid symbols represent data collected on aqueous sulfuric acid subphases and the open symbols represent data collected on other subphases. Solid circle: 1-octadecanol (this study), solid diamond: stearic acid (this study), solid triangle: 1-hexadecanol (this  XV study), solid square: phytanic acid (this study), solid inverted triangle: butanol (Park et al.), 188 solid pentagon: hexanol (Park et al.), 188 solid star: 1octadecanol (Knopf et al.), 189 open circle: hexanoic acid (Thornton and Abbatt), 187 open square: SDS (McNeill et al. ),181 open triangle: oleate (McNeill et al.). 195 Panel (a) gives the ordinate as a function of y in the presence of the monolayer. In panel (b) the ordinate is normalized to y for the uncoated solution. Figure  4.1:  100  Natural logarithm of the observed N205 signal as a function of reaction time. Symbols show representative data for a series of 1-octadecanol monolayers with varying molecular surface area. Solid squares: a molecular surface area of 20.7 A2 molec 4 , open squares: 21.2 A 2 molec 4 , open triangles: 22.6 A2 molec" 1 , and solid triangles: 24.2 A 2 molec 4 . The dashed lines represent the corresponding linear fits to the data. Figure  4.2:  105  Surface pressure — area isotherm for 1-octadecano1 276 (panel (a)) and reactive uptake coefficients (y) for N205 on organic coated aqueous 60 wt % sulfuric acid at 273 ± 1 K plotted as a function of molecular surface area of 1octadecanol (panel (b)). The solid line in panel (a) represents the variation of surface pressure (7r) with a change in molecular surface area. S', S, S*, and L2 * represent the different 2-D phases of the monolayer (see text for further details). 296 Also shown is the collapse pressure (n) of the 1-octadecanol monolayer. The open symbols in panel (a) represent the experimental conditions (surface pressures and molecular surface areas) at which reactive uptake experiments were performed for N205 on aqueous 60 wt % H2SO4 at 273 ± 1 K coated with 1-octadecanol. The solid square in panel (a) represents experimental conditions used previously by Cosman et al. octadecanol monolayers.  76  for 1-  xvi Figure  4.3:  107  Reactive uptake coefficients for N 2 0 5 on organic coated aqueous 60 wt % sulfuric acid at 273 + 1 K plotted as a function of the fractional monolayer surface coverage. In this figure the reactive uptake coefficient in the presence of the organic monolayer (7fil m ) is normalized to reactive uptake coefficient for the uncoated solution  uncoated) to  illustrate the change  in reactive uptake coefficient due to the presence of the monolayer. Open squares: 1-octadecanol (this study), solid square: 1-octandecanol (Cosman et al.). 276 Panel (a) gives the ordinate in a linear scale whereas panel (b) shows the ordinate in a log scale. Figure  4.4:  109  Reactive uptake coefficients for N205 on aqueous sulfuric acid in the presence of organic monolayers as a function of molecular surface area. Open squares: 1-octadecanol (this study), solid right facing triangles: phytanic acid (this study), solid squares: 1-octadecanol (Cosman et al.), 276 solid triangles: stearic acid (Cosman et al.), 276 solid circles: phytanic acid (Cosman et al.), 276 solid left facing triangles: 1-hexadecanol (Cosman et al.), 276 solid diamonds: butanol (Park et al. ),27° solid stars: hexanol (Park et al.), 27° solid inverted triangles: 1-octadecanol (Knopf et al.). 271 In Panel (a) the ordinate is the reactive uptake coefficient in the presence of the organic monolayer and in Panel (b) the ordinate is nu n, normalized to reactive uptake coefficient for the uncoated solution (vuncoated). Figure 4.5 .  111  Surface pressure — area isotherms for organic monolayers on aqueous 60 wt % sulfuric acid at 273 ± 1 K. Panel (a): the isotherm for 1-octadecano1  276  (xphytanic = 0) represented by the solid line and phytanic acid 276 (xp h ytan i c = 1) shown as the dashed line. 71-c, octadecanol and gc, phytanic represent the surface pressure at which monolayers of pure 1-octadecanol and pure phytanic acid collapse, respectively. Panel (b): the isotherm for 2-component monolayers of 1-octadecanol and phytanic acid. The solid line, dashed line, and bold-solid  xvii line represent the isotherms for compositions of  0.05, 0.2, and 0.7,  Xphytamc  respectively. Isotherms for x ph ytam , = 0.1 and 0.4 have been omitted for clarity. Figure  4.6:  113  The collapse pressures of 1-octadecanol-phytanic acid monolayers on aqueous 60 wt % sulfuric acid at 273 ± 1 K as a function of composition (Xphytanic)• Solid (4, mixture)  triangles and solid squares represent the 1 st collapse pressure  and  2 nd  collapse pressure (712 c, mixture) for the 2-component  monolayers, respectively. The open triangle and open square represent the collapse pressure of pure phytanic acid and pure 1-octadecanol monolayers, respectively. Figure 4.7 .  115  Panel (a) shows the measured molecular surface area (solid symbols) and the predicted molecular surface area (dashed line) for 2-component monolayers with R = 21 ± 2 mN ni l as a function of composition -  (Xphytanic). Panel  (b)  displays the excess area (A ex ) as a function of composition for 1-octadecanol phytanic acid monolayers. See text for further details. Figure  4.8:  117  Surface pressure — area isotherms for binary mixtures of 1-octadecanol and phytanic acid performed on aqueous 60 wt % sulfuric acid at 273 ± 1 K. Examples of experimentally measured isotherms for 2-component monolayers containing Xphytanic = 0.1 (bold solid line) and Xphytanic = 0.7 (solid line) are shown. The dashed lines are the predicted isotherms based on a linear combination of the isotherms for pure 1-octadecano1 276 and pure phytanic acid. 276 Figure  4.9:  119  The reactive uptake coefficient for N205 on aqueous sulfuric acid in the presence of 1-octadecanol - phytanic acid monolayers (2/mixed film) as a function of mole fraction of phytanic acid in the monolayer,  Xphytanic• Solid  squares  x viii represent the average 7 value for each corresponding monolayer. The error bars represent 26 . The bold dashed line represents the prediction based on eq 4.5 whereas the bold dash dot line represents the prediction based on eq 4.6. 282 The shaded regions reflect the uncertainty in the predictions based on the uncertainty in yi (Xphytanic = 0) and 72 (xphyta„,, = 1). In Panel (a) the ordinate is the reactive uptake coefficient in the presence of the organic monolayer (ymixed film), and in Panel (b) the ordinate is rmixed film normalized to reactive uptake coefficient for the uncoated solution (limcoated)• Figure 5.1. ^  136  Surface tension of glutaric acid — aqueous 40 wt % sulfuric acid solutions and surface excess ( ,F(H2 o)  ^  of glutaric acid as a function of glutaric acid  concentration at 260 K. Solid circles represent surface tensions (panel (a)) and surface excess of glutaric acid (panel (b)) in 40 wt % H2SO4 at 260 K. The solid line in panel (a) represents a third-order polynomial fit to the data. Figure 5.2: ^  140  Surface tension of azelaic acid — aqueous 40 wt % sulfuric acid solutions and (H 20) surface excess ( F chcarboxy/  )  of azelaic acid as a function of azelaic acid  concentration at 260 K. Solid circles represent surface tensions (panel (a)) and surface excess of azelaic acid (panel (b)) in 40 wt % H2SO4 at 260 K. The solid line in panel (a) represents a third-order polynomial fit to the data. Figure 5.3: ^  141  Experimentally derived natural logarithms of the gas-phase N205 signals as a function of reaction time. Open squares: blank uptake (no aqueous solution in the flow cell), open circles: uncoated aqueous 40 wt % H2SO4, solid triangles: glutaric acid, solid diamonds: azelaic acid. All data was collected at 260 K. Figure 5.4: ^  144  The reactive uptake coefficient (y) for N20 5 on aqueous solutions of sulfuric  xix acid coated with organic monolayers as a function of molecular surface area. Solid symbols represent the results from this study for y in the presence of dicarboxylic acids and the open symbols represent the results in the literature for y in the presence of alcohols and monocarboxylic acids. Solid square: glutaric acid (this study), solid circle: azelaic acid (this study), open square: 1-octadecanol (Cosman et al.), 335 open hexagon: phytanic acid (Cosman et al.), 335 open circle: 1-octadecanol (Cosman et al.), 333 open inverted triangle: stearic acid (Cosman et al.), 333 open star: phytanic acid (Cosman et al.), 333 open triangle: 1-hexadecanol (Cosman et al.), 333 open diamond: butanol (Park et al.), 331 open sideways triangle: hexanol (Park et al.), 331 x: 1-octadecanol (Knopf et al.). 334 Panel (a) gives the ordinate as a function of y in the presence of the monolayer (yfil.). In panel (b) the ordinate is normalized to y for the uncoated solution (I/ \ , uncoated). Figure 5.5: ^  146  Surface tension of glutaric acid — water solutions at 295 K and surface excess ( r (H 2 0) of glutaric acid as a function of glutaric acid concentration. Panel k I dicarboxy1  )  (a) - open circles represent surface tension data collected in this study, open squares are surface tension data from Shulman et a1. 370 The solid line represents a third-order polynomial fit to the data. Panel (b) - surface excess of glutaric acid in water at 295 K calculated by fitting raw data from both this study and Shulman et al. 37° Figure 5.6: ^  147  Surface tension of azelaic acid — water solutions at 295 K and surface excess (  F(dilicfbo) xyl) of azelaic acid as a function of azelaic acid concentration. Panel  (a) - open triangles are surface tension data from Tuckermann et al. 371 The solid line represents a third-order polynomial fit to the data. Panel (b) surface excess of azelaic acid in water at 295 K calculated by fitting raw data from Tuckermann et a1.371  XX  Figure 6.1: ^  157  Sketch of the rotating-wall flow-tube reactor coupled to the CIMS. Figure 6.2: ^  159  The phase diagram of MA/OA as a function of temperature and concentration. 416 S and L indicate the solid and liquid phases, respectively. xmA indicates the MA mole fraction. SoA(a) and SoA(7) represent two polymorphic forms of 0A. 416 The filled circles indicate the conditions of the conducted experiments. Figure 6.3: ^  161  Panels (a) and (b) show a substrate composed of 51 wt % LA/OA prepared by slow and fast cooling, respectively. Panels (c) and (d) show the same substrates as (a) and (b) but with a higher magnification. Panels (a) and (b) were obtained with a digital camera and (c) and (d) were obtained with an optical microscope. Figure 6.4: ^  163  Natural logarithm of the observed 0 3 signal as a function of reaction time. The filled circles and filled triangles correspond to uptake on pure OA and on a 16 wt % MA/OA solution, respectively. The filled and open diamonds represent the data obtained from solid-liquid 26 wt % MA/OA mixtures that have been prepared by slow and fast cooling, respectively. The lines represent the corresponding linear fits to the data. Figure 6.5: ^  166  Experimentally derived reactive uptake coefficient, y, as a function of MA/OA concentration conducted at 298 K. Figure 6.6: ^  167  Plot of inverse reactive uptake coefficient vs. inverse square root of OA concentrations for MA/OA solutions. Solid line indicates a linear fit to the  xxi data points disregarding the data point of pure OA at 0.563 (mol L- 1)-112. Dotted lines represent the 95% confidence interval of the linear fit. Figure  6.7:  168  Plot of inverse reactive uptake coefficient vs. inverse OA concentrations for MA/OA solutions. Solid line indicates a linear fit to the data points disregarding the data point of pure OA at 0.317 (mol L -1 ) -1 . Dotted lines represent the 95% confidence interval of the linear fit. Figure  6.8:  169  Experimentally derived reactive uptake coefficient, y, as a function of MA/OA concentration. The open circles correspond to data obtained from liquid MA/OA solutions shown in Figure 6.6. The filled circles indicate experiments in which the mixture was cooled slowly, and the filled triangles indicate experiments in which the mixture was cooled rapidly. The gray shaded area indicates the liquid-phase concentration range. Figure  6.9:  173  The 03 uptake coefficients as a function of substrate age are presented for a fast-cooled 35 wt MA/OA solution (triangles) and liquid OA (squares). The dotted lines are plotted to guide the eye and have no physical meaning. Figure  6.10:  177  OA lifetimes obtained at 298 K under typical polluted environments (100 ppb 0 3 ) for MA/OA particles, 0.2 pm in diameter. The open circles correspond to  particles in the liquid state. The solid triangles and solid circles indicate particles in the solid-liquid state which have been cooled rapidly and slowly, respectively. The gray shaded area indicates the liquid-phase concentration range. Figure  A.1:  193  Schematic of the physical and chemical processes which determine the overall uptake in gas-particle interactions.458  Figure  A.2:  194  Schematic of the Resistor Model for gas-particle interactions taking into consideration gas-phase diffusion, mass accommodation, solubility limited uptake, liquid-phase reaction, and surface reaction. Figure  B.1:  196  Schematic of the resistor model for gas-particle interactions taking into consideration gas-phase diffusion, mass accommodation, solubility limited uptake, liquid-phase reaction.458  ACKNOWLEDGEMENTS  Over the years there have been many people that have played a role in guiding me to where I am today. First, I'd like to thank Dr. J. Roscoe, who sparked my interest in research and encouraged me to pursue further studies. The past few years have been a journey beyond which I ever imagined I'd like to thank my colleagues at UBC — Daniel, Simone, Jackson, Emily, Sarah, Pedro, Matt, Ben, Magda, Jenna, Atul, Mike, Michael, and Ian. It has been a pleasure working with friends everyday, and was great sharing experiences in the lab with you. Simone, Jackson and Daniel— the trials and tribulations generating N20 5 will be imprinted on my brain forever. I'm glad I was not alone on the trip to the brink of insanity. Daniel, you were an excellent postdoc. Your endless patience and exceptional attention to detail will make you a great professor. I wish you all the best. I'd like to thank Brian, Des, Raz, Ken, Pritesh, and Ken in the machine shop, and Brian Ditchburn for all their effort and hard work in building and maintaining my flow system. They were always up for a challenge, and their skills and knowledge never cease to amaze me. Allan, you are a great mentor. You made me feel as though I was working with you, not for you. Your quiet leadership encouraged me to become independent, but I always knew that you were available if I had questions. For that I am extremely thankful. To my parents, I am eternally grateful for all your support. You've been there for me every step of the way, always letting me know how proud you are of me. You are the best role models. To Nick and my girls— everyday I appreciate how fortunate I am for our family. Without you I never would have come this far. It's been a long road, and I can't wait to continue to journey. The best is yet to come.  xxiv  DEDICATION  To my family  XXV  CO-AUTHORSHIP STATEMENT  Chapters 2 through 6 are co-authored published journal articles or manuscripts prepared for submission. The details of my contribution to each chapter are as follows: Chapter 2: (second author status on published article) •  Identified and designed research program in collaboration with coauthors  •  Helped build and develop experimental apparatus with co-authors  •  Performed all of the preliminary tests and calibration work  •  Obtained all of the experimental data presented in the publication  •  Shared the data analysis with first author  •  Shared manuscript text preparation with first author and my supervisor  Chapter 3: (first author status on published article) •  Identified and designed research program in collaboration with my supervisor  •  Obtained all of the experimental data presented in the publication  •  Performed all of the data analysis  •  Prepared all of the figures for the manuscript  •  Shared manuscript text preparation with my supervisor  Chapter 4: (first author status on accepted article) •  Identified and designed research program in collaboration with my supervisor  •  Obtained all of the experimental data presented in the publication  •  Performed all of the data analysis  •  Prepared all of the figures for the manuscript  •  Shared manuscript text preparation with my supervisor  Chapter 5: (first author status on prepared manuscript) •  Identified and designed research program in collaboration with my supervisor  xxv i  • Obtained half of the experimental data presented in this manuscript •  Shared the data analysis with second author  • Prepared all of the figures for the manuscript •  Shared manuscript text preparation with my supervisor  Chapter 6: (second author status on published article) • Identified and designed research program in collaboration with coauthors • Obtained half of the experimental data presented in the publication •  Shared the data analysis with first author  •  Shared manuscript text preparation with first author and my supervisor  1  Chapter 1^  1. BACKGROUND AND MOTIVATION:^GAS-PARTICLE INTERACTIONS IN THE TROPOSPHERE  1.1 Aerosols and Aerosol Particles Aerosols are solid or liquid particles suspended in a gas and are ubiquitous in the atmosphere.' Aerosol particles vary in mass, size, chemical composition, and number density, and each of these properties is linked to the source of the particles.' Aerosol particles are generally divided into two classes based on particle size: fine particles (diameters <2.5 ,um) and coarse particles (diameters >2.5 pm). 1 Particles with diameters in the range of 0.002 pm — 10 pm are the most important with respect to atmospheric chemistry and physics, and play a significant role in climate, visibility, air pollution, and human health.'  1.2 Importance of Aerosols Aerosol particles influence the global radiation budget by interacting directly with solar and terrestrial radiation. 2 Consequently, aerosol particles have a significant impact on global climate. 2 In addition, they can modify the radiative properties of clouds, the conditions required for cloud formation, and the lifetimes of clouds, by acting as cloud condensation nuclei (CCN) and ice nuclei  am  .2-10  Despite recent strides to understand  the effect of aerosols on climate, the intergovernmental panel on climate change (IPCC) shows that the largest uncertainties in the radiative budget of the atmosphere remain associated with the effects of aerosol particles. 2 The IPCC 2007 assessment declares that our understanding of the impact of aerosols on the radiative budget is still low compared to our understanding of the effect of greenhouse gases on radiative forcing. 2 As such, interest in the field of aerosol research has grown considerably recently, and great strides are being taken to increase our understanding of the chemistry and physics of aerosol particles. The sources, atmospheric transformations, sinks, and composition of tropospheric particles are discussed in detail below.  Chapter 1^  2  1.3 Sources, Atmospheric Transformations, and Sinks of Tropospheric Particles 1.3.1 Sources of Tropospheric Particles Tropospheric aerosols originate from a wide variety of both natural and anthropogenic, or man-made, sources. The troposphere is the region of the atmosphere extending from sea level to approximately 15 km. Aerosol particulates can either be emitted directly to the troposphere by primary sources or can be formed by secondary sources, which involves gas-to-particle conversion in the atmosphere. Primary emission sources include biomass burning, volcanic eruptions, sea salt, biological materials (pollen, spores, leaf abrasion, microorganisms, etc.), wind-driven suspension of mineral dust and soil, and incomplete combustion of fossil fuel." Examples of secondary sources include homogeneous nucleation and condensation of sulfuric acid and water to form new particles, or oxidation of precursor gas-phase organic molecules to form lowvolatility compounds followed by condensation (also known as secondary organic aerosol formation). 1,11,12 Aerosol particles that are emitted through mechanical processes such as wind action (wave-generated sea spray, dust storms, etc.) generally fall into the coarse particle size range.' Fine particulate aerosols usually arise from coagulation of ultra fine particles from primary emission sources or from vapor condensation.' The fine particle mode of the aerosol distribution contains the majority of the total number of aerosol particles (up to 10 8 cm 3 ), and nearly half the mass in urban environments.' 1.3.2 Atmospheric Transformations and Sinks of Tropospheric Particles Atmospheric particles undergo a variety of aging processes throughout their lifetime that consist of changes in particle size and structure." Gas uptake and chemical reaction of atmospheric particles can change the composition of an aerosol particle drastically." Figure 1.1 illustrates the various processes that aerosol particles can go through during their lifetime."  3  Chapter 1^ Primary Emission (Natural and Anthropogenic)  Secondary Formation (Condensation)  • II • is *••. -•  Physical and • •wo 4D Chemical^Cloud Processing Aging ^  N  ••• • • ■•• • • Aiiik  Dry  ^Wet Deposition (Sink)  Figure 1.1: Atmospheric cycling of aerosol particles."  Aerosol particles act as CCN and IN and take up water vapor, resulting in the formation of cloud droplets. These CCN or IN can then be removed by precipitation (i.e. wet deposition). Other aerosol particles can become entrained in the precipitation and removed from the atmosphere." Consequently, wet deposition is a major sink for atmospheric particles on a global scale." Dry deposition processes are a minor sink on a global scale, but are significant on a local scale with regards to air quality and human health." Inhalation of airborne particulate matter for example has a high impact on day to day life. Oxidation of organic material by gas-phase species such as 0 3 , OH, halogens, and NO3, plays a large role in the chemical aging of atmospheric particles. 13 ' 14 Oxidation of inert hydrophobic organic molecules at particle surfaces to hydrophilic molecules can significantly alter the physical and chemical characteristics of the particle. 13 ' 14 Consequently, the chemically-aged, oxygenated particle becomes a better CCN or IN than its precursor.15-18  4  Chapter 1^  Each of these transformations and deposition processes affect the lifetime of atmospheric particles. The lifetime of coarse particulate matter can range from 0.5-10 days in the troposphere prior to deposition. 16 Lifetimes of fine particulates in the troposphere can vary significantly, but are on the order of days to weeks.  16  Removal  usually occurs due to wet deposition.  1.4 Composition of Tropospheric Particles Tropospheric aerosol particles can be organic, inorganic, or composed of both organic and inorganic material, the composition of which can vary significantly depending on the emission source and chemical and physical aging.' The most common inorganic species present in aerosols are NH4 + , Na+ , SO4 2- , NO3 - , and C1 - . 19-23 In contrast, there are hundreds of organic species present in aerosol particles. Figure 1.2 shows the results of a literature survey by Heintzenberg 24 for the average composition of aerosols measured at several urban and rural sites. This figure highlights the variation in mass fraction of organic material, depending on location.  o  other C (elemental) I^I C (organic)  80 -  NH:  O  I 1 NO3  60 -  0. E 0 a) 40 cy) a)  MN SO 42-  > 20 co  Urban^Remote (continental) Remote (marine)  fine atmospheric particle source Figure 1.2: Average urban and remote composition (by mass) for fine particulate aerosol particles based on a literature survey by Heintzenberg 24 of several field studies.  5  Chapter 1^  The organic fractions shown in Figure 1.2 can be due to pure organic particles or organic species internally mixed with inorganic species. 25-32 This thesis focuses on two specific types of atmospheric particles. One type is an aqueous inorganic particle coated with an organic monolayer. The second is a pure organic particle produced from charbroiling and meat cooking operations. 1.4.1 Aqueous Inorganic Particles Coated With Organic Monolayers Insoluble organic material represents a large component of organic material in the atmosphere. A large fraction of this insoluble material contains fatty acids (n-alkanoic acids and n-alkenoic acids) with carbon chain lengths of 12-32 carbon atoms. Fatty acids are present in marine and continental aerosols originating from both biogenic and anthropogenic sources such as biomass burning, breakdown of biological material present on the ocean surface, combustion of fossil fuels, and food preparation methods.  334°  Other  insoluble organic compounds present in atmospheric particles include n-alkanes, nalkanols, aldehydes, polycyclic aromatic hydrocarbons, steroids, and nitrogen containing compounds. 35 Much of the insoluble organic material in the atmosphere is known to be surface active. It has been suggested that this organic material will form organic monolayers at the interface of aqueous particles, and a core-shell structure for atmospheric aerosol particles consisting of an inorganic aqueous solution core with an insoluble organic film was proposed. 14 ' 41 A schematic of this core-shell model is shown in Figure 1.3.  6  Chapter 1^  Hydrophobic Group Hydrophilic Group  Aqueous Particle  vow.  jir-Jrjr  Figure 1.3: Core shell model of organic coated aqueous atmospheric particles. 14,41 -  Surface active organic molecules are often called surfactants, and usually consist of a hydrophilic head group and a hydrophobic tail group. The hydrophilic head group usually contains an oxygen atom that is available for hydrogen bonding with water molecules at the surface. 41 In general, the surface activity of the surfactant increases with length of the hydrophobic tail group. See Section 1.8 for a more in depth discussion of the characteristics of surfactant molecules and organic monolayers. The presence of organic films on aqueous particles can significantly alter the physical properties of aerosols, cloud droplets, and fog droplets by lowering surface tensions and affecting mass transfer across the air-liquid interface. 13 ' 14 ' 41-43 This, in turn, can alter both the CCN ability of a particle and the rate of gas-particle reactions in the atmosphere. 13,14,43,44 While the core-shell model in Figure 1.3 shows an inert hydrophobic surface, atmospheric processing of the surface film by atmospheric radicals can transform the inert hydrophobic layer into a reactive, hydrophilic layer 13 ' 14 as discussed in the atmospheric transformation section.  7  Chapter 1^  1.4.2 Organic Particles Produced From Charbroiling and Meat Cooking Operations A large fraction (— 20 %) of the fine particulate organic mass fraction in urban areas comes from these types of particles. 45 ' 46 Figure 1.4 shows the fractions of identified organic compounds in fine organic aerosols by urban meat cooking operations, organized by compound class. 4749  n-alkanes other^2% 18% lactones 2% n-alkenoic acids 30%  n-alkanoic acids 45%  amides 3%  Figure 1.4: The composition (by compound class) of the identified organic mass fraction of fine particulate measured from charbroiling and meat cooking particles. 47-49  While Figure 1.4 shows the initial compositions of aerosols near the source, after hours to days in the atmosphere, these organic compounds can undergo atmospheric processing by atmospheric oxidants (such as 03, OH, NO3, etc.). Particularly susceptible are those organic compounds with double bonds. Aging of organic species can lead to formation of new chemical compounds and increase the toxicity of organic aerosols.  8  Chapter 1^  1.4.3 Organic Compounds Used in This Thesis The organic compounds chosen for use in this thesis were selected to represent both surface active organic molecules present on atmospheric particles and organic molecules present in particles produced by meat cooking operations (See Figure 1.4). Table 1.1 lists the organic compounds used in this thesis and their atmospheric sources.  Table 1.1: Summary of Organic Compounds Used in This Thesis ^ compound^compound^chemical structure atmospheric class^  sources  n-alkanols^1 -hexadecanolbiomass burning OH  1 -octadecanol^  OH^  n-alkanoic^lauric acid  and leaf abrasion 50 ' 51  marine surface  acids  layer,"' 52 '" meat OH  cooking, 48 ' 49  myristic acid  biomass burning 5 " 1  palmitic acid  stearic acid  phytanic acid s  petroleum 54  n-alkenoic^palmitoleic  meat cooking48'49  acids^acid 0 1-1  oleic acid  aAlkane  derivative was found in tropospheric particles.  Chapter 1  ^  9  Table 1.1^(continued) atmospheric  compound^compound^chemical structure  sources  class  fuel burning, 46 '"  di-carboxylic succinic acid HO  acids  biomass burning, 30 ' 32 meat  OH  cooking 48,49 ,  glutaric acid  photochemical oxidation 56 adipic acid HO  azelaic acid HO  meat cookin g,48,49  n-alkanes^tetracosane  fuel burning46 ' 55  pentacosane n-alkanones^2-penta-  meat cookin g,48,49  C  leaf abrasion, 5°  decanone 2-hexa-  and biomass  decanone  burning 30 ' 32 ' 5°  2-octa-  0  decanone amides  hexadecanamide  meat cooking48'49  10  Chapter 1^  1.5 Heterogeneous Chemistry Aerosols in the atmosphere experience reactions with gas-phase species which potentially lead to the modification of the particle composition and morphology. These reactions between gas-phase species and solid or liquid particles are termed heterogeneous reactions, and are of importance for several reasons. To begin with, heterogeneous reactions play a crucial role in the composition of the atmosphere. Stratospheric and tropospheric ozone are created and destroyed in the atmosphere, and are directly influenced by heterogeneous chemistry. The destruction of stratospheric ozone is a classic example of how heterogeneous reactions can lead to drastic changes in the composition of the atmosphere. 57-59 In the troposphere, heterogeneous reactions of N20 5 on aqueous aerosols (which will be described in detail in the next chapter) play a crucial role in the atmospheric budget of 03 and other reactive species. 6 " 1 In addition to changing the composition of the atmosphere, heterogeneous reactions also play a large role in changing the hygroscopic properties and optical properties of organic aerosols in the atmosphere.  15,18,44,62-64  These reactions can influence  the ability of these particles to act as CCN and to scatter and absorb solar radiation. 44,64-68 In altering the hygroscopic properties of aerosols, heterogeneous reactions may affect the lifetime of the particle in the atmosphere. Furthermore, aerosols that contain organic material may form toxic or carcinogenic compounds; thus, they may be important for health-related issues. In short, these heterogeneous reactions may dictate the importance of organic particles in atmospheric chemistry, climate, and health-related issues. This thesis focuses on two specific heterogeneous reactions: 1) the heterogeneous reaction of N20 5 with inorganic aqueous solutions coated with organic monolayers and 2) the heterogeneous reaction between 03 and organic mixtures that are a good proxy for aerosols produced by meat cooking operations. The experimental goal in these projects is to determine the reactive uptake coefficient of the gas-phase species (N205 and 03) with the substrate or surface.  Chapter 1^  11  1.6 Reactive Uptake Coefficient The reactive uptake coefficient is defined as the fraction of collisions with a surface that leads to the irreversible loss of the gas-phase species due to a reaction. This is a convenient parameter for describing the heterogeneous reaction and for including this chemistry in atmospheric models. Two different methods can be used to determine the reactive uptake coefficient. The first involves using an aerosol flow tube, where the loss of the trace gas-phase species on the aerosol particle is determined. The second approach involves using a flow tube reactor that is coated with the material of interest and using this substrate to determine the reactive uptake coefficient. (See section 2.2.1 and section 6.2 for a discussion of the specific flow tube reactors used in this thesis.) This has an advantage in that the composition and structure of the substrate/surface can be controlled and characterized accurately. For our studies we use a flow tube reactor coated with the material of interest Under most cases the reactive uptake coefficient determined in these studies can be applied directly to aerosol particles.  1.7 N20 5 Chemistry 1.7.1 N20 5 Chemistry in the Troposphere The atmospheric lifetime of reactive species and aerosols in the atmosphere is a key parameter for determining the atmospheric abundance of these compounds and their overall effect on atmospheric chemistry and climate. 69 The atmospheric lifetime determines the magnitude of accumulation of a reactive species in the atmosphere for a given emission source, and how quickly it would be removed from the atmosphere if the emission ceases. When the atmospheric lifetime of a compound is known, it not only determines its abundance but also how this compound will interact and affect the composition of the atmosphere. This information is beneficial in evaluating the role of a species in not only climate issues but related issues such as air quality and health. 69 Processes that remove a chemically active species from the atmosphere, known as sinks, can aid in predicting atmospheric lifetimes. Some of these processes (sinks) include gas-phase free radical reactions, photolytic dissociation or conversion, dissolution  12  Chapter 1^ and deposition, rain out, and heterogeneous and multiphase reactions.  69  This last sink,  heterogeneous and multiphase reactions, is of key interest when discussing the link between N205 and the oxidative capacity of the troposphere. Photochemistry in the atmosphere plays a large role in the composition of the atmosphere, with photochemical reactions producing highly reactive free radicals such as 0 and NO. However, there are free radicals and reactive species that exist at nighttime as well. Since roughly half the earth is in darkness all the time, this creates a prime environment for chemistry to occur. For example, many ozonolysis reactions occur at night due to the presence of significant amounts of ozone. 1 ' 69 Also, the nitrate radical (NO 3 ), which is scarce during the day due to rapid photolysis, is present in significant amounts and available for chemical reactions during the night. 1 ' 69 To illustrate the different chemical pathways a reactive species can go through during the daytime versus at nighttime, let's consider the conversion of NO2. During the daytime, photolysis of NO2 leads to formation 03: R1.1  NO2 + hv --+ NO+ 0  R1.2  0+ 02 --1---> A >03  This is the primary source of tropospheric ozone, which contributes to the oxidative capacity of the troposphere. The major loss process for NO2 during the daytime is reaction with OH: R1.3  NO2 + OH --m-->, HNO3  This is a three-body reaction where M represents an inert molecule in the atmosphere such as N2 or 02. At nighttime, instead of acting as a source of tropospheric ozone, NO2 acts as a sink for 0 3 :  ^R1.4^  NO2 +03 --> NO3 + 02  Further reaction of the nitrate radical with NO2 leads to the reversible formation of N205:  ^R1.5  ^  NO2 + NO 3 <-11--> ' • N 2 05  13  Chapter 1^  Figure 1.5 shows a visual representation of these loss mechanisms for NO2 via the two different pathways: daytime and nighttime. 69  NO2 NO 4C . / N2 0 5 ^HNO 3 N2'5  Removal  (a)  (b)  < V NO  OH  HNO 3  NO NO2 N22O55 •••,,„„„„„,„,„„„„.„,„,,„„„„„,,„*HNO 3 Removal  Removal  Figure 1.5: Daytime and night time reaction pathways that convert chemically active nitrogen species to less active HNO3. (adapted from Ravishankara 69)  At nighttime, the abundance of OH in most regions of the atmosphere is essentially zero, shutting off pathway (a). 69 Figure 1.5 illustrates that the major nocturnal loss mechanism for NO2 is through pathway (b), conversion of N 2 O 5 to HNO 3 via hydrolysis on aqueous particles: R1.6  N2O5(g)  + H2 00) ---> 2HNO3(aq)  As mentioned above, the nitrate radical (NO 3 ) is photolabile, therefore the formation of N2O 5 through pathway (b) during the daytime is minimal. Night time loss of NO2 through N2O5 hydrolysis is an important removal mechanism for NO„ (= NO + NO2) in the troposphere because it prevents the cycling of N2O5 back to NO„ thereby reducing  Chapter 1^  14  formation of 03. Consequently, it is crucial to understand this loss process of N20 5 in order to accurately predict NO and 03 concentrations in the atmosphere. 6° ' 61 Laboratory measurements have shown that reaction R1.6 is efficient on aqueous particles with reactive uptake coefficients of 0.015-0.2. 7° The reactive uptake coefficient (y) is defined as the fraction of gas-particle interactions resulting in irreversible loss of  the gas. Recent field studies have measured nocturnal lifetimes of NO3 and N205 species in air parcels over the Northeast United States, 71 which they used to calculate reactive uptake coefficients. This study concluded that y for N20 5 was highly variable and appeared to be dependent on the aerosol composition and the sulfate-to-organic ratio in particular. Regions with high sulfate-to-organic mass loading yielded 7 = 0.02, which agrees laboratory measurements. However, regions with lower sulfate-to-organic ratios displayed uptake coefficients up to an order of magnitude lower than 0.02. A decrease in y of N205 by an order of magnitude can result in an increase in mass averaged tropospheric NO R , 03, and OH concentrations of 7 %, 4 %, and 8 %, respectively, according to global model simulations. 61 This demonstrates the need for laboratory investigations to gain more insight into N20 5 uptake on aqueous solutions using systems that contain both organic and aqueous inorganic material and more closely resemble atmospheric aerosols.  15  Chapter 1^  1.8 Theory of Monolayers 1.8.1 Organic Surfactants Organic molecules that contain both a hydrophobic carbon chain and a hydrophilic head group have an affinity for the air-aqueous interface due to the hydrophobic nature of their hydrocarbon chain. These types of molecules are often termed surfactants due to their tendency to segregate at surfaces. Organic surfactant solutions are in a dynamic equilibrium, with organic molecules constantly adsorbing/desorbing from the surface as shown in Figure 1.6: 4 crystallites 1•11  monolayer 1111111MMO1•1•1•1111101•1111401411401•1•1111101•1014111•/•14110141•.•1•1 ^  111111111111  wall^1 adsorption  air  I le)  0'  •• ,,,• • •• micelle  I 111  Tit'  subphase  dissolved free monomers  Figure 1.6: Schematic representation of the dynamic equilibrium established for monolayers at the air-water interface (adapted from Albrecht et al!).  The surface tension (o) of a solution is defined as the work required to change the surface area by a unit area. 72 Surfactants have the ability to significantly lower the surface tension of aqueous solutions, even at low surfactant concentrations. This reduction in the surface tension of a solution by the presence of a surfactant is labeled as the surface pressure (pr) of the surfactant monolayer, and is given by the following equation: 73 1.1  = (7 0  where o is the surface tension of the pure liquid or surfactant free solution, and crf i l rn is the surface tension of the monolayer covered solution.  Chapter 1^  16  One way to determine the surface activity of an organic species is to measure the surface tension lowering effect of the organic surfactant on an aqueous solution. There are several ways to measure the surface tension of a solution that include: capillary rise method, detachment or static Wilhelmy plate methods, ring method, drop-volume or drop-weight methods, pendant drop method, and oscillating jet method.  73  In this thesis  the detachment and static Wilhelmy plate methods are used to measure surface tensions and surface pressures of surfactant solutions. These methods are described in more detail below. The Wilhelmy method uses a thin plate, typically made of platinum, mica, glass or filter paper, 73 suspended in the liquid solution. Several forces act on the submerged plate: two downward forces due to gravity and surface tension effects, and an upward force due to buoyancy. The experimental set-up is illustrated in Figure 1.7: 73 connection to balance  air contact angle 0  aqueous solution 1  h  air aqueous solution  w front view^side view Figure 1.7: A Wilhelmy plate partially submerged in an aqueous solution for measurement of surface pressure. (adapted from Gaines73)  17  Chapter 1^  For a plate with dimensions of length (1), width (w), and thickness (t), with a density (pp), the force (F) on the plate is given by eq 1.2: 73  ^1.2^F = p p glwt + 2o- (t + w)cos — p gtwh ,  where g is the gravitational constant, a is the surface tension, B is the contact angle, p i is the density of the liquid, and h is the height of the liquid in which the plate is submerged. For the static Wilhelmy method the force is measured using a balance while the plate is maintained at a constant height in the solution." If the plate is completely wetted (ie., the contact angle is zero and cose = 1), then the surface pressure of the solution is given by eq 1.3: 73  ^1.3^  z= A=  ^OF  —  2(t + w)  The Wilhelmy plate method can also be used to measure the surface tension of a solution using the detachment method. The procedure is similar to the static method whereby the force on the plate is measured; however, instead of maintaining the plate at a constant height, the force on the plate is measured as a function of height in the liquid as the plate is immersed and detached from the liquid solution. The surface tension is then calculated from the maximum difference in the force on the plate between immersion and separation from the solution.  1.8.2 Insoluble Monolayers Organic surfactants that have both a hydrophilic head group and a hydrophobic tail group (usually one or more hydrocarbon chains with >14 carbon atoms) form insoluble organic monolayers at the air-water interface. Typically, these surfactants have very low solubility in aqueous media and in turn, form a relatively stable monolayer at the interface. Surfactant molecules in insoluble monolayers can exist in a variety of states, from low density, gaseous states, to more tightly packed, condensed or solid states. These states can be thought of as 2-dimensional (2-D) gases, liquids, and solids present on the surface of the aqueous subphase and are analogous to their three-dimensional  Chapter 1^  18  counterparts. 73 For a 2-D gas-phase monolayer, surfactant molecules are present at the interface, but are sufficiently separated from one another so as to exert little force on each other. Assuming 2-D kinetic analysis analogous to the ideal gas law, gas-phase surfactant molecules in the plane of the surface have a total kinetic energy of kT, which is described by the ideal gas law: 73 1.4  zA =kT  where r is the surface pressure of the monolayer, A is the molecular surface area of the surfactant, k is the Boltzmann constant, and T is the temperature. Since eq 1.4 is a limiting relation, it only holds at very low surface pressures and very large molecular surface areas. Equation 1.5 corrects the ideal gas law for the finite size of the polar head group and also the interaction of the hydrocarbon chains with the surface: 1.5  73  (z – z 0 )(A– A o ) =kT  where go corresponds to the spreading coefficient of the hydrocarbon part of the monolayer on water and A o represents the minimum area occupied by a surfactant molecule. Figure 1.8 illustrates a typical surface pressure-area (z-A) isotherm for a long, straight-chain molecule. -A isotherms are obtained using a Langmuir trough 74 by measuring the surface pressure of an insoluble monolayer as a function of surface area at constant temperature.  19  Chapter 1^ monolayer collapse condensed solid  Langmuir trough ^ barrier barrier  subphase  * • *  condensed liquid expanded liquid gaseous  area per molecule Figure 1.8: A schematic of a Langmuir trough 74 used to measure a , -A isotherm and a typical ' -A isotherm representative of a long, straight-chain organic surfactant. (adapted from Kaganer et al. 75 ) -  -  Figure 1.8 shows several "kinks" in the isotherm, which correspond to 2-D phase transitions. 73 ' 75 These phase transitions represent the different degrees of ordering of the organic molecules. 75 ' 76 At large molecular surface areas the monolayers exist as a 2-D gas on the aqueous acid surface as discussed above, with molecules on the surface exerting relatively little force on each other due to sufficient separation. 73 As the molecular surface area decreases, the monolayer undergoes several phase transitions until the monolayer collapses. The phases associated with these phase transitions are generally referred to as gaseous, expanded liquid, condensed liquid, and condensed solid phases. Monolayers that contain bent hydrocarbon chains (ie., unsaturated molecules that have cis-double bonds like oleic acid) or branched hydrocarbon chains do not exhibit the same phase behavior as illustrated in Figure 1.8. Bent and branched surfactants are not able to pack as efficiently at the interface as straight-chain surfactants. cis-double bonds or methyl groups have a much more pronounced effect on the packing ability of surfactant molecules than trans-double bonds. 73 As such, the 7r A isotherms of bent -  surfactants with cis-double bonds and branched surfactants tend to form expanded monolayers as opposed to condensed monolayers.73  Chapter 1^  20  1.8.3 Soluble Monolayers  Soluble surfactants are typically shorter-chain molecules that contain both a hydrophobic carbon chain and a hydrophilic head group. Molecular interactions between the hydrophilic head group of short-chain molecules and water make these organic species soluble in aqueous solutions. Despite their solubility, these molecules can exhibit high surface activity due to the hydrophobicity of their tail group. For soluble monolayers, it is typically the surface excess of the surfactant that is determined, rather than the surface concentration. The surface tension (a) of a solution is related to the chemical potentials of the components in a multi-component solution by the Gibbs adsorption equation: 72 1.6  do- = —SdT  —  En i r dA /  where S is the entropy, T is the temperature, T is the surface excess of component i, and A is the chemical potential of component i. For a two component system (solvent and  surfactant) at constant temperature, eq 1.6 simplifies to: 1.7  do- =  72  — F2 deu 2  If the dividing surface is chosen so as to make the surface excess of the solvent zero (Ti---0), then eq 1.7 further reduces to: 72 FT) = ( du  1.8  where FP is the surface excess of the surfactant with respect to the solvent. The Gibbs adsorption equation allows the extent of adsorption of the organic surfactant at the interface to be estimated from variation of surface tension with solution composition by substituting p 2 = + RT In a 2 into eq 1.8. 72 1.9  ^  F(1) —^1^a6 2^RT a2ria 2 )1.  Chapter 1^  21  The concentration dependence of FP at surfactant concentrations below phase separation is described by the Langmuir adsorption isotherm 72 (eq 1.10) and can be used to determine the surface excess at saturation (F," ): 1.10  F=  F [1] ^a (B + [i])  where [i] is the surfactant concentration and B represents a ratio of the rate constants for the adsorption and desorption from the surface. 77 In order for eq 1.10 to be applicable the Langmuir model assumes that the surface consists of a set number of adsorption sites and that all adsorbed species interact with only one site, and not each other. Consequently, adsorption is limited to a monolayer. 72  1.9 Thesis Overview This thesis focuses on the heterogeneous chemistry of organic mixtures and inorganic solutions coated with organic films as proxies for atmospheric particles. Specifically we focus on two topics. First, we focus on the reactive uptake coefficient of N205 on aqueous inorganic solutions coated with organic monolayers. The goal of this project is to better understand how organic monolayers on aqueous particles will affect the mass transport and kinetics of N205 uptake on aqueous aerosols, and consequently what effect the monolayer will have on predicted concentrations of N205 in the atmosphere. Secondly, we investigate the reactive uptake coefficient of 0 3 on proxies for meat cooking aerosols with the goal to better understand the discrepancy between the observed lifetime of oleic acid in tropospheric aerosols and the lifetime predicted by laboratory results to date. The main results from this research are presented in chapters 2-6. Chapter 2 focuses on validating a new instrument for accurately measuring the uptake of N205 on aqueous solutions coated with an organic monolayer. Chapter 3 examines the uptake of N205 on aqueous solutions through insoluble, 1-component (1-octadecanol, 1hexadecanol, stearic acid, or phytanic acid) monolayers at a single packing density, using the newly developed instrument. Chapter 4 focuses on uptake of N20 5 on aqueous solutions through 1-component (1-octadecanol) monolayers as a function of packing  Chapter 1^  22  density and through 2-component monolayers (1-octadecanol and phytanic acid) as a function of monolayer composition. Chapter 5 examines the uptake of N205 on aqueous solutions in the presence of soluble, dicarboxylic acid (glutaric acid, azelaic acid) surfactants. Finally, Chapter 6 focuses on the reactive uptake of 1-component (oleic acid) and multicomponent (oleic acid and myristic acid or meat cooking mixture) bulk organic substrates by ozone.  Chapter 1^  23  1.10 References 1. Finlayson-Pitts, B. J.; Pitts, J. N. Chemistry of the upper and lower atmosphere: Theory, experiments, and application; Academic Press: San Diego, CA, 2000. 2. Climate Change 2007: The Scientific Basis. Contribution of Working Group 1 to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change; Cambridge University Press: Cambridge, United Kingdom, 2007. 3. Twomey, S. Atmos. Environ. 1974, 8, 1251-1256. 4. Albrecht, B. A. Science 1989, 245, 1227-1230. 5. Twomey, S. Atmos. Environ. A-Gen. 1991, 25, 2435-2442. 6. Pincus, R.; Baker, M. B. Nature 1994, 372, 250-252. 7. Baker, M. B. Science 1997, 276, 1072-1078. 8. Pruppacher, H. R.; Klett, J. D. Microphysics of Clouds and Precipitation; Kluwer Academic Publishers: Dordrecht, The Netherlands, 1997. 9. Rosenfeld, D. Science 2000, 287, 1793-1796. 10. Ramanathan, V.; Crutzen, P. J.; Kiehl, J. T.; Rosenfeld, D. Science 2001, 294, 21192124. 11. Poschl, U. Angew. Chem. Int. Edit. 2005, 44, 7520-7540. 12. Kanakidou, M.; Seinfeld, J. H.; Pandis, S. N.; Barnes, I.; Dentener, F. J.; Facchini, M. C.; Van Dingenen, R.; Ervens, B.; Nenes, A.; Nielsen, C. J.; Swietlicki, E.; Putaud, J. P.; Balkanski, Y.; Fuzzi, S.; Horth, J.; Moortgat, G. K.; Winterhalter, R.; Myhre, C. E. L.; Tsigaridis, K.; Vignati, E.; Stephanou, E. G.; Wilson, J. Atmos. Chem. Phys. 2005, 5, 1053-1123. 13. Eliason, T. L.; Gilman, J. B.; Vaida, V. Atmos. Environ. 2004, 38, 1367-1378.  Chapter 1^  24  14. Ellison, G. B.; Tuck, A. F.; Vaida, V. I Geophys. Res.-Atmos. 1999, 104, 1163311641. 15. Cruz, C. N.; Pandis, S. N. Environ. Sci. Technol. 2000, 34, 4313-4319. 16. Pandis, S. N.; Wexler, A. S.; Seinfeld, J. H. I Phys. Chem. 1995, 99, 9646-9659. 17. Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. Atmos. Environ. A-Gen. 1993, 27, 1309-1330. 18. Saxena, P.; Hildemann, L. M.; Mcmurry, P. H.; Seinfeld, J. H. I Geophys. Res.Atmos. 1995, 100, 18755-18770. 19. Carslaw, K. S.; Clegg, S. L.; Brimblecombe, P. J. Phys. Chem. 1995, 99, 1155711574. 20. Clegg, S. L.; Brimblecombe, P.; Wexler, A. S.J. Phys. Chem. A 1998, 102, 21372154. 21. Clegg, S. L.; Brimblecombe, P.; Wexler, A. S. J. Phys. Chem. A 1998, 102, 21552171. 22. Massucci, M.; Clegg, S. L.; Brimblecombe, P. J. Phys. Chem. A 1999, 103, 42094226. 23. Wexler, A. S.; Clegg, S. L. I Geophys. Res. 2002, 107, 4207, doi:4210.1029/2001JD000451. 24. Heintzenberg, J. Tellus 1989, 41 B, 149-160. 25. Middlebrook, A. M.; Murphy, D. M.; Thomson, D. S. I Geophys. Res.-Atmos. 1998, 103, 16475-16483. 26. Murphy, D. M.; Thomson, D. S.; Mahoney, T. M. J. Science 1998, 282, 1664-1669.  Chapter 1^  25  27. Cavalli, F.; Facchini, M. C.; Decesari, S.; Mircea, M.; Emblico, L.; Fuzzi, S.; Ceburnis, D.; Yoon, Y. J.; O'Dowd, C. D.; Putaud, J. P.; Dell'Acqua, A. Geophys. Res.-Atmos. 2004, 109, D24215, doi:24210.21029/22004JD005137. 28. Decesari, S.; Facchini, M. C.; Fuzzi, S.; McFiggans, G. B.; Coe, H.; Bower, K. N. Atmos. Environ. 2005, 39, 211-222. 29. Decesari, S.; Facchini, M. C.; Fuzzi, S.; Tagliavini, E. J. Geophys. Res. -Atmos. 2000, 105, 1481-1489. 30. Graham, B.; Mayol-Bracero, 0. L.; Guyon, P.; Roberts, G. C.; Decesari, S.; Facchini, M. C.; Artaxo, P.; Maenhaut, W.; Koll, P.; Andreae, M. O.J. Geophys. Res. 2002, 107(D20), 8047, doi:8010.1029/2001JD000336. 31. Havers, N.; Burba, P.; Lambert, J.; Klockow, D. J. Atmos. Chem. 1998, 29, 45-54. 32. Mayol-Bracero, 0. L.; Guyon, P.; Graham, B.; Roberts, G.; Andreae, M. O.; Decesari, S.; Facchini, M. C.; Fuzzi, S.; Artaxo, P. J. Geophys. Res.-Atmos. 2002, 107(D20), 8091, doi:8010.1029/2001JD000522. 33. Barger, W. R.; Garrett, W. D. J. Geophys. Res. 1970, 75, 4561-4566. 34. Barger, W. R.; Garrett, W. D. J. Geophys. Res.-0c. Atm. 1976, 81, 3151-3157. 35. Saxena, P.; Hildemann, L. M. J Atmos. Chem. 1996, 24, 57-109. 36. Mochida, M.; Kitamori, Y.; Kawamura, K.; Nojiri, Y.; Suzuki, K. J. Geophys. Res. 2002, 107, 4325, doi:4310.1029/2001JD001278. 37. Peterson, R. E.; Tyler, B. J. Atmos. Environ. 2002, 36, 6041-6049. 38. Tervahattu, H.; Hartonen, K.; Kerminen, V. M.; Kupiainen, K.; Aarnio, P.; Koskentalo, T.; Tuck, A. F.; Vaida, V.J. Geophys. Res.-Atmos. 2002, 107, 4053, doi:4010.1029/2000JD000282.  Chapter 1^  26  39. Tervahattu, H.; Juhanoja, J.; Kupiainen, K. J Geophys. Res.-Atmos. 2002, 107, 4319, doi:4310.1029/2001JD001403. 40. Tervahattu, H.; Juhanoja, J.; Vaida, V.; Tuck, A. F.; Niemi, J. V.; Kupiainen, K.; Kulmala, M.; Vehkamaki, H. J Geophys. Res.-Atmos. 2005, 110, D06207, doi:06210.01029/02004JD005400. 41. Gill, P. S.; Graedel, T. E.; Weschler, C. J. Rev. Geophys. 1983, 21, 903-920. 42. Facchini, M. C.; Decesari, S.; Mircea, M.; Fuzzi, S.; Loglio, G. Atmos. Environ. 2000, 34, 4853-4857. 43. Donaldson, D. J.; Vaida, V. Chem. Rev. 2006, 106, 1445-1461. 44. Vaida, V.; Tuck, A. F.; Ellison, G. B. Phys. Chem. Earth (C) 2000, 25, 195-198. 45. Hildemann, L. M.; Markowski, G. R.; Cass, G. R. Environ. Sci. Technol. 1991, 25, 744-759. 46. Schauer, J. J.; Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R. Atmos. Environ. 1996, 30, 3837-3855. 47. Limbeck, A.; Puxbaum, H. Atmos. Environ. 1999, 33, 1847-1852. 48. Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simonelt, B. R. T. Environ. Sci. Technol. 1991, 25, 1112-1125. 49. Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T. Environ. Sci. Technol. 1999, 33, 1566-1577. 50. Standley, L. J.; Simoneit, B. R. T. Environ. Sci. Technol. 1987, 21, 163-169. 51. Fang, M.; Zheng, M.; Wang, F.; To, K. L.; Jaafar, A. B.; Tong, S. L. Atmos. Environ. 1999, 33, 783-795. 52. Hunter, K. A.; Liss, P. S. Mar. Chem. 1977, 5, 361-379.  Chapter 1^  27  53. Marty, J. C.; Saliot, A.; Buatmenard, P.; Chesselet, R.; Hunter, K. A. I Geophys. Res.-0c. Atm. 1979, 84, 5707-5716. 54. Simoneit, B. R. T.; Cox, R. E.; Standley, L. J. Atmos. Environ. 1988, 22, 983-1004. 55. Kawamura, K.; Kaplan, I. R. Environmental Science & Technology 1987, 21, 105110. 56. Kawamura, K.; Kasukabe, H.; Barrie, L. A. Atmos. Environ. 1996, 30, 1709-1722. 57. Molina, M. J.; Tso, T. L.; Molina, L. T.; Wang, F. C. Y. Science 1987, 238, 12531257. 58. Crutzen, P. J.; Arnold, F. Nature 1986, 324, 651-655. 59. Solomon, S.; Garcia, R. R.; Rowland, F. S.; Wuebbles, D. J. Nature 1986, 321, 755758. 60. Dentener, F. J.; Crutzen, P. J. I Geophys. Res.-Atmos. 1993, 98, 7149-7163. 61. Evans, M. J.; Jacob, D. J. Geophys. Res. Lett. 2005, 32, L09813, doi:09810.01029/02005GL022469. 62. Dick, W. D.; Saxena, P.; McMurry, P. H. I Geophys. Res.-Atmos. 2000, 105, 14711479. 63. Ming, Y.; Russell, L. M. AIChE J. 2002, 48, 1331-1348. 64. Shilling, J. E.; King, S. M.; Mochida, M.; Martin, S. T. I Phys. Chem. A 2007, 111, 3358-3368. 65. Asad, A.; Mmereki, B. T.; Donaldson, D. J. Atmos. Chem. Phys. 2004, 4, 2083-2089. 66. Broekhuizen, K.; Kumar, P. P.; Abbatt, J. P. D. Geophys. Res. Lett. 2004, 31, L01107. 67. Facchini, M. C.; Mircea, M.; Fuzzi, S.; Charlson, R. J. Nature 1999, 401, 257-259.  Chapter 1^  28  68. Novakov, T.; Penner, J. E. Nature 1993, 365, 823-826. 69. Ravishankara, A. R. Faraday Discuss. 2005, 130, 9-26. 70. Sander, S. P.; Golden, D. M.; Kurylo, M. J.; Moortgat, G. K.; Wine, P. H.; Ravishankara, A. R.; Kolb, C. E.; Molina, M. J.; Finlayson-Pitts, B. J.; Huie, R. E. "Chemical Kinetics and Photochemical Data for Use in Atmospheric Studies Evaluation Number 15," JPL Publication 06-2, 2006. 71. Brown, S. S.; Ryerson, T. B.; Wollny, A. G.; Brock, C. A.; Peltier, R.; Sullivan, A. P.; Weber, R. J.; Dube, W. P.; Trainer, M.; Meagher, J. F.; Fehsenfeld, F. C.; Ravishankara, A. R. Science 2006, 311, 67-70. 72. Adamson, A. W.; Gast, A. P. Physical chemistry of surfaces, 6th Ed. ed.; WileyInterscience: New York, 1997. 73. Gaines Jr., G. L. Insoluble monolayers at liquid-gas interfaces; Interscience Publishers: New York, 1966. 74. Langmuir, I. J. Am. Chem. Soc. 1917, 39, 1848-1906. 75. Kaganer, V. M.; Mohwald, H.; Dutta, P. Rev. Mod. Phys. 1999, 71, 779-819. 76. Lawrie, G. A.; Barnes, G. T. J. Colloid Interf ScL 1994, 162, 36-44. 77. Donaldson, D. J. J. Phys. Chem. A 1999, 103, 62-70.  Chapter 2^  29  2. A NOVEL FLOW REACTOR FOR STUDYING REACTIONS ON LIQUID SURFACES COATED BY ORGANIC MONOLAYERS: METHODS, VALIDATION, AND INITIAL RESULTS  2.1 Introduction Reactions between aerosol particles and gas-phase species, often termed heterogeneous reactions, have been identified to play a crucial role in the atmosphere. These reactions can lead to significant changes in atmospheric composition.  78-9°  For  example, the hydrolysis of N205 on and in aqueous H2SO4 aerosol particles represents an important heterogeneous reaction acting as a sink for NO x (NO and NO2) in the troposphere. 80 ' 89 See Chapter 1.7 for a further discussion. A majority of the previous work in the area of heterogeneous atmospheric chemistry has focused on reactions involving aqueous inorganic particles. However, tropospheric particles can consist of a large fraction of organic material (see, e.g., Kanakidou et al. 91 ) and some of these organic molecules can act as surfactants that may form organic monolayers on the surface of aqueous inorganic particles.  92-193  These  organic monolayers may limit the transfer of molecules across the air-aqueous interface and, hence, reduce reaction rates between gas-phase species and the aqueous particles. Recently, researchers have begun to investigate the effect of organic monolayers on heterogeneous chemistry of aqueous aerosol particles.  104-115  Nevertheless, the effect of  organic monolayers on atmospheric chemistry still remains unclear. A possible experimental technique for studying the effect of an organic monolayer on the heterogeneous chemistry of aqueous solutions involves aerosol flow tube reactors where the gas-phase loss to the aerosol particles is measured (see, for example, Hanson and Lovejoy, 116 Fried et al., I17 Hu and Abbatt, 118 Robinson et al., 119 Hallquist et al. ,120 Kane et al. ,121 Thornton and Abbatt, I°8 and McNeill et al. 1 "). A version of this chapter has been published. Knopf, D. A., Cosman, L. M., Mousavi, P., Mokamati, S., and Bertram, A. K., A Novel Flow Reactor for Studying Reactions on Liquid Surfaces Coated by Organic Monolayers: Methods, Validation, and Initial Results, J. Phys. Chem. A, 2007, 111 11021-11032. ,  Chapter 2^  30  However, in this case, determining important properties of the organic monolayers, such as surface tension and packing density, directly on submicron particles is difficult. Here we have developed a new flow reactor for studying heterogeneous reactions on aqueous solutions coated with organic monolayers. The advantage of this new flow reactor is that it allows us to study heterogeneous reactions using well characterized organic monolayers. For example, we can determine the surface tension and the packing density of the organic monolayer at the air-aqueous interface prior to and after studying the heterogeneous chemistry. This allows us to correlate properties of the organic monolayer with the heterogeneous reaction rates. In this paper we first describe the new flow reactor and discuss computational fluid dynamics simulations that are used to characterize the flow dynamics in the reactor. Then we present a mathematical procedure to derive the first-order wall loss rate constant, kw ist , and the reactive uptake coefficient, y, from our experimentally observed first-order wall loss rate constant, k o b s . (The reactive uptake coefficient is defined as the ratio of the molecules removed from the gas phase by reactions to the total gas-surface collisions.) Next, to validate our new apparatus and mathematical procedure for data analysis, we present measurements of the reactive uptake of N205 by aqueous H2SO4 solutions and measurements of the reactive uptake of 03 by liquid canola oil. We choose these reactions for validation purposes because they have been studied several times in the past using well developed experimental procedures. Finally, we carry out a preliminary study of the reactive uptake coefficient of N205 on aqueous H2SO4 solutions coated with an insoluble organic monolayer. We end by discussing the atmospheric implications of the latter results.  2.2 Experimental 2.2.1 New Flow Reactor Figure 2.1 shows a schematic of the newly developed flow reactor. Figure 2.1 A shows a top view with the cover removed, and Figure 2.1 B and Figure 2.1 C show a side view and a front view, respectively. The injector is not included in the front view for clarity.  31 movable injector  x  liquid surface  B carrier gas inlet  outlet  x  mixir  r` r  C  Figure 2.1: New flow reactor. (A) Top view of the flow reactor without the cover. The main flow is from left to right. The liquid surface in a quartz trough, the movable T-shaped injector, and the corresponding coordinate system are indicated. (B) Side view of the flow reactor including the top cover. The carrier gas inlet, mixing barrier, length of the liquid surface, and corresponding coordinate system are indicated. (C) Cross section of the flow reactor. The width of the liquid surface is given, and the grooves that support the injector are represented.  The main body of the reactor is made from aluminum, and it can be temperature controlled by circulating coolant through channels in the aluminum body. Located on the bottom surface of the reactor is a glass trough, which is filled with the aqueous solution, and this aqueous solution can be covered with an organic monolayer. The gas-phase reactant is introduced to the flow reactor by a movable T-shaped injector, which slides just above the liquid surface. The T-shaped injector is equipped with 6 exit holes 0.2 mm in diameter, which point toward the top of the reactor and which distribute the gas-phase reactant evenly across the width of the flow cell. The carrier gas enters the flow reactor  Chapter 2^  32  through inlets at the back of the flow cell. The gas stream entering the reactor first flows against a barrier to ensure mixing before reaching the liquid surface. Shown in Figure 2.1 C are the dimensions of the open channel above the liquid surface through which the gases flow. The open channel is close to a perfect rectangular channel geometry, except for the small grooves on the side, which support the movable injector. When designing the flow cell reactor the height of the open channel (distance above the liquid surface in the y-direction) was kept as small as possible to reduce the effect of diffusion to the aqueous solution on the overall loss process of the gas-phase reactants. This height was typically less than 10 mm. The width (distance in the zdirection) of the open channel was chosen in such a way that the ratio of the height to width, which is also called the aspect ratio (e) is as small as reasonably possible. This simplified the calculations necessary to extract kw, lst and y from the experimental data (see below). In addition, the length of the flow cell (distance in the x-direction) was chosen so that the carrier gas was fully developed over most of the length of the liquid surface (see below). A chemical ionization mass spectrometer (CIMS) is connected to the outlet of the flow reactor to measure the gas-phase reactant concentration.  122'123  A typical reactive  uptake experiment involves measuring the gas-phase reactant concentration as a function of position of the T-shaped injector. By varying the position of the injector, we varied the reaction time between the gas-phase reactant and the liquid surface, and from this data, we determined the observed first-order loss rate, ko b s . From ko b s we then determined the first-order wall loss rate, k,,,, ist , and the reactive uptake coefficient, y, of the trace gasphase species to the aqueous solution (see below for the procedure to determine k w lst and y from ko b s ). The total pressure in the flow reactor is measured through a 0.64 cm port in the center of the flow reactor cover using a capacitance pressure gauge. Three additional ports, 0.16 cm in diameter, allow the measurement of gas and solution temperature using K-type thermocouples. All aluminum surfaces inside the flow reactor are coated with Halocarbon wax to minimize loss of the gas-phase species to the walls.  33  Chapter 2^ 2.2.2 Details of the Reactive Uptake Experiments  Three different kinds of reactive uptake experiments have been conducted: First, to validate our apparatus and mathematical procedure for data analysis, we measured the reactive uptake coefficient of 0 3 on liquid canola oil and the reactive uptake coefficient of N205 on aqueous H2SO4 solutions (without organic monolayers). After the validation experiments, we measured the reactive uptake coefficient of N20 5 by an aqueous H2SO4 solution coated with a monolayer of 1-octadecanol (C18H370H). For the 0 3 uptake experiments, we generated 03 by passing a flow of 02 over an ultraviolet source. The generated 03 was collected and stored in a 5 L bulb. During the 0 3 uptake experiments the flow of 02/03 that passed through the movable injector varied between 0.9 and 4.5 cm 3 min -1 STP (standard temperature and pressure). The flow of the He carrier gas ranged between 0.14 and 2 L min -1 STP. This resulted in Reynolds numbers (Re) of 0.3-5, indicating laminar flow conditions. 03 was detected as 03 - in the mass spectrometer after its chemical ionization by SF6 - . 122,124 SF6 was generated by passing a trace amount of SF6 in about 2 L min -1 STP N2 through a 210Po source. 122 0 3 concentrations used in these reactive uptake measurements ranged from 0.53 x 10  11  to 1.4  x 10 11 molecules cm 3 . The total flow velocities used in these experiments ranged from 70 to 470 cm s d . Within experimental uncertainty, the reactive uptake coefficient was independent of flow rate. For the second set of experiments N205 was generated by reacting NO2 with an excess amount of 03 in a separate flow system. 123,125-127 The N205 resulting from this reaction was passed through a glass vessel containing P2O5 to convert any residual HNO 3 into N205 before N205 was collected in a glass trap held at 193 K. In the N205 uptake measurements a saturated flow of N20 5 of about 4.6-9 cm 3 mind STP mixed in a He flow of 36-80 cm 3 mind STP enters the flow reactor through the movable injector. The flow of the H20/He carrier gas varies between 50-780 cm 3 min d STP. This results in laminar flow conditions (Re = 0.1-1.6). The relative humidity of the H 2 0/He carrier gas is adjusted to the corresponding relative humidity of the aqueous H2SO4 solution, which was determined using the Aerosol Inorganics Model.  128-130 N2,-.5  was detected as NO3  after its chemical ionization by F. 124 r was generated by passing a trace amount of CH 3 I in about 2 L min -1 STP N2 through a 210 Po source. N205 concentrations in experiments  Chapter 2^  34  employing pure aqueous sulfuric acid solutions ranged between 2 x 10 10 to 1 x 10 11 molecules cm 3 . N205 concentrations in experiments employing aqueous sulfuric acid solutions coated by an organic monolayer ranged between 8 x 10 9 to 1 x 10 12 molecules cm 3 . Because we are using low concentrations of reactants in our experiments, the accumulation of reaction products during the course of our experiments is not a concern. For example, if all the HNO3 produced from the N20 5 hydrolysis remained in the solution, the maximum HNO3 concentration in the solution after 1 h would be at most (assuming y = 0.01) 0.02 wt % with the highest N205 concentrations. The fact that we do not see any dependence of the reactive uptake coefficient on time (see below), or the N205 concentrations used further confirms that accumulation of impurities is not an issue. The flow velocities used in these experiments ranged from 200 to 500 cm s -1 . Within experimental uncertainties, the reactive uptake coefficients were independent of this parameter. 2.2.3 Organic Monolayer Preparation  In the experiments where we measured the reactive uptake of N205 on aqueous sulfuric acid solutions coated with organic monolayers, two types of methods were used to prepare the organic monolayers. The first consisted of depositing a few droplets of solution of 1-octadecanol dissolved in chloroform on the aqueous H2SO4 surface. 131 ' 132 The second method consisted of sprinkling 1-octadecanol crystals on the aqueous H2SO4 solution. 133-135 Both methods produced an organic monolayer in contact with solid 1octadecanol. Sulfuric acid solutions were prepared volumetrically using purified water (resistivity 18.2 MC2 cm). 2.2.4 Computational Fluid Dynamics Simulations  Computational fluid dynamics (CFD) simulations have been performed to show that the gas flow over the liquid surface has a well developed laminar flow profile for typical experimental conditions used in our experiments. Fully developed laminar flow conditions simplify the data analysis of our experimental results (see below). In addition we use these simulations to visualize the gas flow profiles in our experiments. Also we use the CFD simulations to show that the flow profile of the gas in our reactor is close to the flow profile that would be established in a perfect rectangular channel.  35  Chapter 2^  For the simulations, we chose conditions that were the same as some of the conditions used in the reactive uptake experiments. These conditions are listed in Table 2.1 and cover the typical range of conditions used in this study as well as typical conditions that we plan to use in future studies of reactive uptake measurements on aqueous solutions coated with organic monolayers.  Table 2.1: Experimental Conditions Under Which Flow Experiments and Corresponding CFD Simulations Were Conducted' experiment  a  mass flow He,  mass flow H 2 0,  pressure,  % relative  Vavg,  STP cm -3 min -1  STP cm -3 min-1  Torr  humidity  m S-1  1  176  10  2.6  0.5  1.4  2  875  17  3.9  0.25  4.5  3  1570  18  5.1  0.2  6.0  4  2708  410  9.1  4.1  6.5  5  2655  17  6.9  0.15  7.6  6  3946  17  8.8  0.12  8.7  All experiments and simulations were performed at 298 K.  The simulations were carried out with the software package Fluent. 136 Fluent is capable of modeling fluid flow velocity vectors and temperature and pressure contours. 136 The framework of Fluent is based on the conservation of mass, momentum, and energy. First a three-dimensional computational grid (or mesh) that corresponded to the actual experimental dimensions was constructed. For a detailed description of the computation grid see Knopf et al. 137 Each cross point of the grid represented a node at which the differential equations that described the conservation of mass, momentum, and energy were replaced by equivalent finite difference approximations. Those algebraic equations were solved numerically to yield the variables of the interest such as flow velocity, pressure, and temperature. Iteration of the overall equations using minimization of the corresponding residuals lead to convergence of the numerical solution. For these calculations, we focused on the flow dynamics of the gas phase and assumed the liquid surface was stationary. The three dimensional segregated solver for laminar conditions was applied for these simulations. Discretization was performed using  Chapter 2^  36  the second-order upwind scheme. Wall and gas temperature were set constant to 298 K. Fluid properties were obtained using ideal gas mixing laws and mass diffusivity was derived using a constant dilution approximation.  136  Figure 2.2 shows the flow profiles which develop along the y-axis evaluated at z = 0.0375 m (i.e., midpoint of the width). The different symbols correspond to the different conditions given in Table 2.1. Panel a, b, and c of Figure 2.2 correspond to calculations performed at x-axis positions of 6, 13, and 20 cm, respectively. The results indicate a Poiseuille flow along the y-axis (i.e., in the vertical direction) between the liquid surface and the top cover of the reactor. The flow profiles do not change when going from x = 6 cm to x = 20 cm, indicating a fully developed laminar flow has been established at x = 6 CM.  Figure 2.3 shows the flow profiles that developed across the width of the flow reactor (in the z-axis) evaluated at mid-height. Similar to above, the symbols represent the CFD calculations and the different symbols correspond to the different conditions given in Table 2.1. Panels a, b, and c of Figure 2.3 were evaluated at x-axis positions of 6, 13, and 20 cm. The profile at an x-position of 6 cm is the same as the profile at an xposition of 20 cm, indicating a fully developed laminar flow at an x-position of 6 cm. Figure 2.3 also indicates that the flow velocity is constant over a majority of the width (zdirection).  37  Chapter 2^  4 3 2 1  4 3  0 -1 -2 -3 -4  -3 -4  4 pj^ 7 4 3 7 ^ (0 7 3 t^ 2- 4^14 a. .^ - 2 n^ €4 • 1 "^ 19, * • ^- 1 A '''  .  0  r;  e.^ 4 4 '^ C .^ 4^ 3^ : "^ ,-,^ ^  • -1  -0  -1 2 -^, .e • . ,---)^7 -2 -3 - i^n 4 4,-,5^" - 3 "4 4 4 ,...:-..sr^ - -4 e  -  ;lc ■^i .  I . I^i . .  2 4 6 8 10 12 14 16 1  v im s j -  Figure 2.2. Velocity flow profiles along the y-axis derived from CFD simulations for the conditions given in Table 2.1. (a), (b), and (c) represent the results obtained at x = 6, 13, and 20 cm, respectively. Solid diamonds, open circles, open squares, solid circles, solid triangles, and open diamonds represent the flow profiles obtained for conditions 1, 2, 3, 4, 5, and 6 given in Table 2.1, respectively. Solid lines indicate calculations of the flow profile for a rectangular channel given by Solbrig and Gidaspow.138  38  Chapter 2^  0.0  ^  0.025  ^  0.05  ^  0.075  14  14  12  12  10  10 8  6  6  (a)  4 2  tM  2  0  0  14  14  12 000'0 00 O 00 i 0 0000 . 0,  12 10  —0) 10  E  4  8  8  6 (b  4  4  2  2  0  0 14  14 12  A . 6,  AAAAAA1a4A4AAAAAAA,  12  0 0 i• OO 0 ,00 OO 0 OO .0000,  10  10 8  6  I.  4 2^  ft  6 4  2  ^0  0.0^0,025^0.05^0.075 2: [nl]  Figure 2.3. Velocity profiles across the width of the flow reactor (in the z-axis) derived from CFD simulations for the conditions given in Table 2.1. (a), (b), and (c) represent the results obtained at x = 6, 13, and 20 cm, respectively. Solid diamonds, open circles, open squares, solid circles, solid triangles, and open diamonds represent the flow profiles obtained for conditions 1, 2, 3, 4, 5, and 6 given in Table 2.1, respectively. Solid lines indicate calculations of the flow profile for a rectangular channel given by Solbrig and Gidaspow.138  Chapter 2^  39  The flow profiles in our system are very close to the flow profiles one would predict for a rectangular channel with a width equal to 75 mm (which is the width of the liquid surface) and height equal to 9 mm (which is the height of the open channel above the liquid). This is also illustrated in Figures 2.2 and 2.3. The solid curves in these figures represent the predicted flow velocities for a rectangular channel with width = 75 mm and height = 9 mm, calculated using the equations presented in Solbrig and Gidaspow. 138 The equations by Solbrig and Gidaspow 138 correspond to fully developed laminar flow. The solid lines are in very good agreement with the predictions from our computational fluid dynamics simulations, which is not surprising because our geometry is very close to the geometry of a rectangular channel. In Figures 2.2 and 2.3 we show that the flows are fully developed after a short distance (<6 cm) in the reactor. Additional CFD analysis (not shown here) indicates that the time to reach a fully developed laminar laminar flow after the mixing barrier is less than 1 5 cm for all the different conditions given in Table 2.1. This is consistent with the approximate estimates of the time to reach a fully developed flow between two parallel plates with similar dimensions and flow conditions. According to Levich  139  the distance  required to reach fully developed flow between two parallel plates can be estimated by: 2.1  /,  0.1• a • Re  where a is half the height of the flow reactor and Re is the Reynolds number. If we use a Reynolds number and height consistent with our experimental conditions we obtain 1 e 0.2 cm. The effect of the T-shaped injector on the flow profiles has not been modeled, but it is assumed that the distance to reach a fully developed flow after the T-shaped injector will be similar to the time to reach a fully developed flow at the entrance of the flow reactor. This distance is much less than the length of the reactive surface. For example, for the conditions given in Table 2.1 the corresponding distances are maximum 5-15 mm. Also this distance will not effect our overall uptake measurements as long as it is relatively short and it remains constant during the uptake measurements. This is discussed in more detail below.  40  Chapter 2^  2.2.5 Mathematical Framework to Derive the First-Order Wall Loss Rate Coefficient, k w ist , and the Reactive Uptake Coefficient, y The overall goal of our experiments is to determine the reactive uptake coefficient, y, of the trace gas-phase species to the aqueous solution, where y is defined as the fraction of collisions with a surface that result in irreversible loss. As mentioned above, in our experiments we measure the concentration of the trace gas-phase species exiting the flow cell as a function of injector position. Assuming first-order kinetics, the concentration as a function of position can then be described by the following equation: 14° 7\  2.2^  C(l) = C o exp  11'  obs i  V avg  where 1 indicates the length of the reactive surface, C(l) is the concentration of the gasphase reactant at a position 1, Co is the concentration of the gas-phase reactant at 1 = 0, van is the average flow velocity. In our studies we plot the In [C] versus the reaction time given by t = 11 vavg to determine kbs. The position used for the 1 origin (i.e., 1 = 0) is arbitrary because eq 2.2 shows that k ob s can be determined from the relative concentration of C. 14° However, it must be fixed at some distance downstream from the T-shaped injector so that the measurements start only after the reactants are well mixed.  14°  Typically, we use a value of approximately 2 cm after the T-shaped injector to ensure the reactants are well mixed and the flow is fully developed. If the heterogeneous reaction at the aqueous surface is slow, then kob s equals the first-order wall loss rate constant (1c, 1St) , and then y can be calculated from kw ist using the following equation, which is also corrected for the non-Maxwellian velocity distribution 141,142  2.3  1 y  =  cA^1 + 41c1sT 2  where A = length x width is the reactive surface area and V is the gas flow volume above this area. c is the thermal molecular speed of the reactant gas species given by  Chapter 2^  41  c = -ART I 7zM , where R is the general gas constant, T is the temperature of the gas -  species, and M is the molecular weight of the gas species. For our analysis, we assume that the reactive uptake coefficient is independent of time, which is supported by experimental results. In a set of separate experiments we first pushed the injector past the liquid solution (no exposure to the liquid) and the signal of the reactive gas (either N205 or 0 3 ) is recorded. Then we pulled the injector back several centimeters and observed the signal over the period of several minutes. As soon as the injector was pulled back, the signal decreased and then rapidly stabilized. Over the observation time, the signal was constant. In addition, during the course of 0 3 and N205 uptake experiments we did not observe any changes in the reactive uptake coefficient with time. We performed up to eight uptake experiments on three freshly prepared liquid surfaces, which took 1-3 h. Within this time frame we did not observe any changes of the reactive uptake. The observation that the uptake coefficient was independent of time is consistent with the low reactant concentrations used in our experiments and only a small amount of reactant products accumulating during the course of our experiments as mentioned above. If the heterogeneous reaction at the aqueous surface is fast, then concentration gradients of the reactive species can develop in the open channel above the liquid surface. In this case, the observed loss rate of the reactive species is controlled by both diffusion to the liquid surface and heterogeneous reaction at the surface, and hence k o b s can be significantly less than kw ist . To determine kw ist from kobs, gas-phase transport to the reactive surface has to be considered. 140,143-145 Once kw l s t is determined, y can then be calculated using eq 2.3. In the case of a cylindrical flow reactor an analytical solution exists, 143 which has been implemented into a Fortran program code 140, to determine kw "' from ko b s . This solution takes into account concentration gradients of the reactive species that develop in both the axial and radial directions. 14° In other words, this solution takes into account diffusion in the axial and radial directions. For rectangular channel flow reactors with one catalytic wall, an analytical solution exits to determine kw 1st from kob s . 138,144 This analytical solution corrects for concentration gradients that develop in the vertical direction, but it does not correct for concentration gradients that develop in the direction  Chapter 2^  42  of the bulk flow. 138 ' 144 In most cases this is a very good approximation when fast flows are applied (see below). For completeness, however, a numerical solution has been developed to calculate Icw ist from ko b s . 137 This numerical solution corrects for concentration gradients that develop in both the vertical direction and the direction of the bulk flow for arbitrary laminar flow conditions. This numerical method basically decouples the effect of mass transport to and reaction at the liquid surface. See Knopf et a1. 137 for the derivation of this numerical solution. In our studies, we use this numerical method to calculate kw i s t from kobs . The reactive uptake coefficient, y, is then calculated from kwlst using eq  2  as mentioned above.  When calculating k w lst from kobs, the gas-phase diffusion coefficients are needed. The diffusion coefficients applied in this study are calculated using molecular parameters following the procedure outlined previously. 122,146-151 These calculated diffusion coefficients are given in Table 2.2 as a function of experiment temperature.  Table 2.2: Calculated Diffusion Coefficients of N205 in He, N2 0 5 in H2O (Vapor), and 03 in Ile 146-151 temp, K  DN 2 0  5  ^DN-112o A^  —He 5  '  Torr cm 2 s -1  Torr cm2 s -1  273  289  72  295  330  85  298  336  87  D03—He  Torr cm2 s -1  394  The binary N205 diffusion coefficient in the He/H 2 0 gas mixture for given water partial pressure, p H,o , and He partial pressure, p,„ is then calculated using: 152  ^( 2.4 ^D  ^■ -1 N 2 0,  =  PH20  , D N2 05 —He  \•  D N 2 05 —H 20  Chapter 2^  43  2.2.6 Chemicals Listed below are the chemicals, the corresponding purities, and manufacturer used in our studies: N2 (99.999%, Praxair), He (99.999%, Praxair), SF6 (99.995%, Praxair), 02 (99.993%, Praxair), H2SO4 (95-98%, Fisher), CH3I (99%, Aldrich), NO2 (99.5%, Matheson), P205 (97%, Aldrich), canola oil (not determined), 1-octadecanol (99%, Aldrich), chloroform (99.9%, Fisher).  2.3 Results and Discussion 2.3.1 Validation of the Flow Cell and the Data Analysis Procedure As mentioned above, to validate our flow cell and data analysis methodology, we measured the reactive uptake of 03 on canola oil and N20 5 on aqueous uncoated H2504 solutions. Shown in Figure 2.4 are examples of typical results. Plotted is the natural logarithm of the gas-phase reactant signal as a function of reaction time (determined from the average flow velocity). Each of the data points represents the gas-phase reactant concentration as a function of injector position. The data for each uptake experiment was fitted by a straight line and the observed first-order loss rate, kobs, was determined from the slope.  ^ 44  Chapter 2^  ^0.00^0.01^0.02^0.03^0.04^0.05^0.06^0.07 0.0^  0.0  -0.2^-0_ -(:).4^  i- --  -0.2  --o, -  ^  - 0.4  -0.6^ M --.. -0.8^ m c cr) -1.0^ . Cl) -E -1.2^  -1.2  -1.4^  -1.4  -1.6^  -1.6  -0.6 -0.8 -1.0  U  0.00^0.01^0.02^0.03^0.04^0.05^0.06^0.07 fime/s  Figure 2.4: Experimentally derived natural logarithms of the gas-phase signals as a function of reaction time (0. Solid circles and solid squares indicate the uptake of 03 by canola oil and the uptake of N205 by aqueous 80 wt % H2SO4 solution, respectively. Dashed lines indicate a linear fit to the data.  The example of the uptake of 0 3 by canola oil presented in Figure 2.4 results in ko b s = 7.8 s -1 . The calculated k is ' value derived from the experimental data using the correction procedure is 9.3 s -1 . Hence, kobs is corrected by 19 %. In the case of the reactive uptake of N205 by aqueous H2SO4 shown in Figure 2.4 the k o b s value is 116.5 s -1 and the corresponding k w lst value is 372 s -1 , which is about 318 % higher than k ob s . These results emphasize the importance of the correction for diffusion. For the reaction between 03 and canola oil we carried out two sets of experiments. In the first set, we measured y for 03 on canola oil as a function of total pressure while keeping the flow velocity relatively constant (the flow velocity was held  Chapter 2^  45  between 200 and 300 cm s -1 ). In the second set of experiments we measured y for 0 3 on canola oil as a function of flow velocity while keeping the total pressure constant (at 3 Torr). Figure 2.5a shows the results of uptake measurements of 03 by canola oil as a function of total pressure in the flow reactor. Each data point in the Figure 2.5 is the result of 3-16 individual uptake experiments. Uncorrected y values were obtained by using ko b s in eq 2.3 instead of kw lst . Corrected values were obtained by using kw ' s' in eq 2.3 as discussed above. The dashed line in Figure 2.5a represents the value reported in the literature determined with a cylindrical flow reactor.  153  The shaded region represents the  uncertainty in the number reported in the literature. 153 In all cases our measured y is consistent with the literature data. Figure 2.5a indicates that the correction for vertical diffusion and diffusion in flow direction is small in all cases. Figure 2.5b indicates that the correction for diffusion increases with the total pressure as expected. This is due to slower diffusion of the gas-phase reactants to the reactive surface at higher total pressures. Figure 2.6 shows y values obtained from the uptake of 03 by canola oil as a function of flow velocity. Also, on the secondary x-axis we have included the Peclet number, which can be interpreted as the ratio between system length and diffusion length. Figure 2.6a presents uncorrected y values derived from ko b s and corrected values derived from k  lst .  Figure 2.6 shows that for low flow velocities or Peclet numbers the  uncorrected value does not agree with the literature values. Correction for vertical diffusion and diffusion in the direction of the bulk flow is necessary to obtain agreement with the literature data. Figure 2.6b shows that the correction for diffusion can be large at low flow velocities, as expected. This is due to the increased importance of diffusion in the direction of the bulk flow as the flow velocity is decreased.  Chapter 2^  1.0x103  0  46  2^4^6^8^10^12^14^16  1.0x10-3  9.0x104  9.0x10 4  8.0x10 4  8.0x10`  7.0x104  7.0x10 4  6.0x104  6.0x10 4  45  45  40  40  35  35  30 _) 25  .1.  •  0  0 20  •  •  •  •  25 20  •  15 10  30  15 I^,^I  o  2^4^6^8^10^12^14^16  10  Pressure / torr  Figure 2.5: Uptake of 03 by canola oil as a function of pressure (a). Open circles indicate y values that have not been corrected for diffusion. Solid circles indicate y values that have been corrected for vertical diffusion and diffusion in the direction of bulk flow. The dashed line and gray shading represent the literature value with corresponding uncertainty, 153 respectively. (b) shows the amount of correction when taking diffusion into account.  47  Chapter 2^  N Pe 1.4^2.4^3.7^5.4  1.0x10 -3  ^  6.9 1.0x10 -3  (a) 9.0x10 -4  9.0x10 -4  8.0x10 -4  8.0x10  7.0x10 -4  7.0x10 -4  0  6.0x10 -4  6.0x10 -4 45  45  •  40  (b)  35 30  25  •  20 0  35 30  •  25  40  •  20  •  • 15  15  10 ^ ^ ^ 10 100^200^300^400 0 500 600 .  Velocity / cm S-1  Figure 2.6: Panel (a) shows uptake of 03 by canola oil as a function of flow velocity and Peclet number (Npd. Open circles and solid circles indicate y values that are not corrected for diffusion and are corrected for vertical diffusion and diffusion in the direction of bulk flow, respectively. The dashed line and gray shading represent the literature value with corresponding uncertainty, 153 respectively. Panel (b) shows the amount of correction due to vertical diffusion and diffusion in direction of bulk flow as a function of velocity.  Chapter 2^  48  As mentioned, for the validation experiments we also studied the uptake of N205 on aqueous H2SO4 solutions (not coated with organic monolayers). At 295 K we studied the uptake on 80 wt % solutions, and at 273 K we studied the uptake on 60 wt % solutions, and the results from these measurements are reported in Table 2.3.  Table 2.3: Experimentally Obtained Reactive Uptake Coefficients for the Uptake of N205 by Aqueous 60 and 80 wt % H2SO4 Solutions and N20 5 by Aqueous 80 wt % H2SO4 Solutions Coated With a Monolayer of 1-Octadecanor solution  uncorr y  con. y  % corr  literature b  60 ± 1 wt %  (1.6 ± 0.4) x10 -2  (4.9+_ 51 51 ) x10 -2  306  3.2-8.5 x10 -2  (1.9 ± 0.4) x10 -2  (5.0j'_01 95 ) x10 -2  263  4.2-10.8 x10 -2  (7.4 ± 2.6) x10 -4  (8.1 ± 3.2) x10 -4  21  H2SO4/1120 c  80 ± 1 wt % H2SO4/ -120 d C18H370H  80 ± 1 wt% H2S 04/H20 e Uncorrected y values do not consider diffusion of the gas-phase species. The error represents +1 o. Corrected y values are corrected for vertical diffusion and diffusion in direction of bulk flow. The corresponding error is due to an assumed 20% error in the diffusion coefficients. 154 The amount of correction due to consideration of vertical diffusion and diffusion in direction of bulk flow is given as percentage. In addition, the literature value for the uptake of N 2 0 5 by aqueous H 2 SO4 solutions is given for comparison. b Mozurkewich and Calvert, I55 Hanson and Ravishankara,' 56 Fried et al., 1 " Hu and Abbatt," 8 Robinson et al., 119 Hallquist et al., 12° Kane et al.' 21 'Measured at (273 ± 1) K. d Measured at (295 ± 0.5) K. e Measured at (298 ± 0.5) K. a  The uncorrected values in Table 2.3 (determined using /cob s in eq 2.3) do not agree with literature values. The error in the uncorrected values represents ±1a. However, after correcting for diffusion using the procedure outlined above, the corrected values are in agreement with the literature data. The uncertainties in the corrected values are mainly due to a 20% uncertainty in the diffusion coefficients.  154  The general conclusion from the 0 3 and N205 experiments discussed above is that the flow reactor and method of data analysis work well for both slow reactions where 7 is approximately 8 x 10 -4 and for fast reactions where 7 is approximately 0.1.  Chapter 2^  49  2.3.2 N205 Reactive Uptake Measurements on Sulfuric Acid Solution Coated With an Organic Monolayer  The reactive uptake of N205 by an aqueous 80 wt % H2SO4 surface coated with an organic monolayer of 1-octadecanol was studied. As mentioned above, these experiments were carried out with a monolayer in contact with a few crystals of 1octadecanol on the surface. In a separate set of experiments (using a commercial surface pressure sensor with a platinum plate and a commercial Langmuir film balance) we determined the surface pressure and the packing density of the organic monolayer. Using the surface pressure sensor with a platinum plate, we determined that the surface pressure is 36 ± 1 mN m -1 . Using the Langmuir film balance and a standard procedure, 131 we measured the pressure-area isotherm for the organic monolayer on the aqueous solution, and then from this information, we were able to conclude that the packing density of the organic monolayer is 22.5 A 2 molecule 1. After determining the properties of the monolayer, we then measured the reactive uptake coefficient. The obtained reactive uptake coefficient of N205 by aqueous 80 wt % H2SO4 coated by octadecanol is (8.1 ± 3.2) x 10 -4 . This number is based on eight uptake experiments for three individually prepared organic monolayers. We did not observe any time dependence of the reactive uptake coefficient during the course of these experiments. This value is about 2 orders of magnitude lower compared to the reactive uptake on pure aqueous 80 wt % H2SO4 solutions. In Table 2.4 we compare our measurements of N205 uptake with previous measurements that also studied the uptake of N2 05 in the presence of monolayers.  50  Chapter 2^  Table 2.4: Comparison of Measured Reactive Uptake Coefficients of N205 on Aqueous H2SO4 and NaC1 Solutions Coated With Organic Monolayers of Different Chain Lengths' literature  monolayer  subphase  temp  Y  (K)  Chain  factor  length  decrease in y  this study  octadecanol  H2SO4/ H2O  298  (8.1 + 3.2) x10 -4  18  62  McNeill et al. iii  SDS  NaC1/ H2O  295  (2 ± 1) x10 -3  12  10  Thornton  hexanoic  NaC1/ H2O  295  (8 ± 4) x10 -3  6  3.5  and  acid  Abbatt' °8 Park et al. 114  hexanol  H2SO4/ H2O  216  (6 ± 1) x10 -2  6  2.5  Park et al. 114  butanol  H2SO4/ H2O  216  (1 ± 0.2) x10 -1  4  1.5  a The factor decrease of the reactive uptake coefficient due to an organic monolayer coating compared to the corresponding bare aqueous solution is given.  The previous literature studies focused on organic monolayers of sodium dodecyl sulfate, hexanoic acid, hexanol, and butanol. 108,111,114 In these studies the decrease in y in the presence of the organic monolayer ranged from 1 to 10. For our results, we observed a larger decrease in reactive uptake, which is likely in part related to the chain length of the surfactant. In the future we will systematically study the effect of monolayer properties such as chain length, surface pressure, and packing density on N205 uptake.  2.4 Atmospheric Implications N205 heterogeneous reactions on aqueous particles are known to be an important  sink of NO in the atmosphere. 80 ' 89 Reactions of N2 0 5 on aqueous solutions have been studied extensively, and more recently researchers have begun to investigate the effect of organic monolayers on this chemistry (see Table 2.4 and the discussion above). This recent research has focused on mainly short chained and soluble monolayers, and in these studies a decrease in the reactive uptake coefficient of between 1 and 10 was observed. In  Chapter 2^  51  our studies with long chained and insoluble organic monolayers (C 18) we observed a decrease of a factor of approximately 62. Evans and Jacob 89 show that a decrease by a factor of 5 compared to the previous study of Dentener and Crutzen 8° can change predictions of NO N, 03, and OH concentrations by 7%, 4%, and 8%, respectively. A decrease in the reactive uptake by 2 orders of magnitude would drastically change the atmospheric composition. However, our results should be considered as a lower limit to y, because monolayers in the atmosphere will likely also contain organic molecules of shorter chain length and also branched organic molecules, which will most likely have a smaller effect on y. Further studies using the flow reactor presented here will address this point by using more atmospherically relevant monolayers.  2.5 Conclusions and Summary A new flow reactor has been developed that allows the study of heterogeneous kinetics occurring on a planar aqueous surface. Here, the effect of an organic monolayer of 1-octadecanol (C181-1370H) on the heterogeneous kinetics of N205 and aqueous H2SO4 surfaces has been studied. Computational fluid dynamics simulations have been applied in the development of the flow dynamics for various experimental conditions. These results were used to set up a novel mathematical framework to derive the true first-order wall loss rate coefficient, k w ist , from the experimentally observed wall loss rate, kobs, under consideration of vertical diffusion and diffusion in flow direction of the gas-phase reactant. The results indicate that neglecting diffusion can lead to measured reactive uptake coefficients that are erroneous by several 100 %. Validation of the new apparatus has been performed by measuring the uptake of 03 by canola oil as a function of pressure and flow velocity. Additional validation has been performed by measuring the reactive uptake coefficients of N205 by aqueous 60 wt % and 80 wt % H2SO4 solutions. The reactive uptake of N205 by aqueous H2SO4 surfaces coated with an organic monolayer of 1-octadecanol has been measured. The packing density of the 1octadecanol monolayer in the uptake experiments was 22.5 A 2 molecule -1 , indicating a monolayer in a condensed state. The measured reactive uptake coefficient was determined to be (8.1 ± 3.2) x 10 4 . This is almost 2 orders of magnitude lower than the uptake by the bare aqueous H2SO4 solution.  121,155,157  The data indicate that the uptake  Chapter 2^  52  may partly depend on the monolayer chain length based on a comparison of our results with previous literature values. 108,111,114 The reactive uptake coefficient on coated aqueous H2SO4 surfaces obtained here may serve as a lower limit for atmospheric aerosols because monolayers in the atmosphere, most likely, also contain organic molecules of different chain lengths, and branched structures which will have a smaller effect on the reactive uptake kinetics.  Chapter 2^  53  2.6 References 78. Solomon, S.; Garcia, R. R.; Rowland, F. S.; Wuebbles, D. J. Nature 1986, 321, 755758. 79. Molina, M. J.; Tso, T. L.; Molina, L. T.; Wang, F. C. Y. Science 1987, 238, 12531257. 80. Dentener, F. J.; Crutzen, P. J. J. Geophys. Res.-Atmos. 1993, 98, 7149-7163. 81. Dentener, F. J.; Carmichael, G. R.; Zhang, Y.; Lelieveld, J.; Crutzen, P. J. J. Geophys. Res. Atmos. 1996, 101, 22869-22889. 82. Ravishankara, A. R.; Rudich, Y.; Talukdar, R.; Barone, S. B. Philos. Trans. R. Soc. Lond. Ser. B-Biol. Sci. 1997, 352, 171-181. 83. Gard, E. E.; Kleeman, M. J.; Gross, D. S.; Hughes, L. S.; Allen, J. 0.; Monica!, B. D.; Fergenson, D. P.; Dienes, T.; Galli, M. E.; Johnson, R. J.; Cass, G. R.; Prather, K. A. Science 1998, 279, 1184-1187. 84. Zhang, Y.; Carmichael, G. R. J. AppL Meteorol. 1999, 38, 353-366. 85. Song, C. H.; Carmichael, G. R. J. Atmos. Chem. 2001, 40, 1-22. 86. Ammann, M.; Poschl, U.; Rudich, Y. Phys. Chem. Chem. Phys. 2003, 5, 351-356. 87. Rudich, Y. Chem. Rev. 2003, 103, 5097-5124. 88. Poschl, U. Angew. Chem. Mt. Edit. 2005, 44, 7520-7540. 89. Evans, M. J.; Jacob, D. J. Geophys. Res. Lett. 2005, 32, L09813, doi:09810.01029/02005GL022469. 90. Rudich, Y.; Donahue, N. M.; Mentel, T. F. Annu. Rev. Phys. Chem. 2007, 58, 321352.  Chapter 2^  54  91. Kanakidou, M.; Seinfeld, J. H.; Pandis, S. N.; Barnes, I.; Dentener, F. J.; Facchini, M. C.; Van Dingenen, R.; Ervens, B.; Nenes, A.; Nielsen, C. J.; Swietlicki, E.; Putaud, J. P.; Balkanski, Y.; Fuzzi, S.; Horth, J.; Moortgat, G. K.; Winterhalter, R.; Myhre, C. E. L.; Tsigaridis, K.; Vignati, E.; Stephanou, E. G.; Wilson, J. Atmos. Chem. Phys. 2005, 5, 1053-1123. 92. Barger, W. R.; Garrett, W. D. J. Geophys. Res.-0c. Atm. 1976, 81, 3151-3157. 93. Garrett, W. D. Pure Appl. Geophys. 1978, 116, 316-334. 94. Decesari, S.; Facchini, M. C.; Mircea, M.; Cavalli, F.; Fuzzi, S. J. Geophys. Res.Atmos. 2003, 108, 4685, doi:4610.1029/2003JD003566. 95. Duce, R. A.; Mohnen, V. A.; Zimmerman, P. R.; Grosjean, D.; Cautreels, W.; Chatfield, R.; Jaenicke, R.; Ogren, J. A.; Pellizzari, E. D.; Wallace, G. T. Rev. Geophys. 1983, 21, 921-952. 96. Facchini, M. C.; Decesari, S.; Mircea, M.; Fuzzi, S.; Loglio, G. Atmos. Environ. 2000, 34, 4853-4857. 97. Facchini, M. C.; Mircea, M.; Fuzzi, S.; Charlson, R. J. Nature 1999, 401, 257-259. 98. Gill, P. S.; Graedel, T. E.; Weschler, C. J. Rev. Geophys. 1983, 21, 903-920. 99. Mochida, M.; Kitamori, Y.; Kawamura, K.; Nojiri, Y.; Suzuki, K. J. Geophys. Res. 2002, 107, 4325, doi:4310.1029/2001JD001278. 100. Tervahattu, H.; Hartonen, K.; Kerminen, V. M.; Kupiainen, K.; Aarnio, P.; Koskentalo, T.; Tuck, A. F.; Vaida, V. J. Geophys. Res.-Atmos. 2002, 107, 4053, doi:4010.1029/2000JD000282. 101. Tervahattu, H.; Juhanoja, J.; Kupiainen, K. J. Geophys. Res.-Atmos. 2002, 107, 4319, doi:4310.1029/2001JD001403.  Chapter 2^  55  102. Tervahattu, H.; Juhanoja, J.; Vaida, V.; Tuck, A. F.; Niemi, J. V.; Kupiainen, K.; Kulmala, M.; Vehkamaki, H. J. Geophys. Res.-Atmos. 2005, 110, D06207, doi:06210.01029/02004JD005400. 103. Donaldson, D. J.; Vaida, V. Chem. Rev. 2006, 106, 1445-1461. 104. Rubel, G. 0.; Gentry, J. W. J Phys. Chem. 1984, 88, 3142-3148. 105. Rubel, G. 0.; Gentry, J. W. I Aerosol Sci. 1985, 16, 571-574. 106. Daumer, B.; Niessner, R.; Klockow, D. I Aerosol Sci. 1992, 23, 315-325. 107. Folkers, M.; Mentel, T. F.; Wahner, A. Geophys. Res. Lett. 2003, 30, 1644, doi:1610.1029/2003GL017168. 108. Thornton, J. A.; Abbatt, J. P. D. I Phys. Chem. A 2005, 109, 10004-10012. 109. Lawrence, J. R.; Glass, S. V.; Nathanson, G. M. I Phys. Chem. A 2005, 109, 74497457. 110. Lawrence, J. R.; Glass, S. V.; Park, S. C.; Nathanson, G. M. I Phys. Chem. A 2005, 109, 7458-7465. 111. McNeill, V. F.; Patterson, J.; Wolfe, G. M.; Thornton, J. A. Atmos. Chem. Phys.  2006, 6, 1635-1644. 112. Anttila, T.; Kiendler-Scharr, A.; Tillmann, R.; Mentel, T. F. J. Phys. Chem. A 2006, 110, 10435-10443. 113. Gilman, J. B.; Vaida, V. I Phys. Chem. A 2006, 110, 7581-7587. 114. Park, S. C.; Burden, D. K.; Nathanson, G. M. I Phys. Chem. A 2007, 111, 29212929. 115. Clifford, D.; Bartels-Rausch, T.; Donaldson, D. J. Phys. Chem. Chem. Phys. 2007, 9, 1362-1369.  Chapter 2^  56  116. Hanson, D. R.; Lovejoy, E. R. Science 1995, 267, 1326-1328. 117. Fried, A.; Henry, B. E.; Calvert, J. G.; Mozurkewich, M. I Geophys. Res 1994, 99, 3517-3532. 118. Hu, J. H.; Abbatt, J. P. D. J. Phys. Chem. A 1997, 101, 871-878. 119. Robinson, G. N.; Worsnop, D. R.; Jayne, J. T.; Kolb, C. E.; Davidovits, P. J. Geophys. Res.-Atmos. 1997, 102, 3583-3601. 120. Hallquist, M.; Stewart, D. J.; Baker, J.; Cox, R. A. I Phys. Chem. A 2000, 104, 3984-3990. 121. Kane, S. M.; Caloz, F.; Leu, M. T. J. Phys. Chem. A 2001, 105, 6465-6470. 122. Knopf, D. A.; Anthony, L. M.; Bertram, A. K. I Phys. Chem. A 2005, 109, 55795589. 123. Knopf, D. A.; Mak, J.; Gross, S.; Bertram, A. K. Geophys. Res. Lett. 2006, 33, L17816, doi:17810.11029/12006GL026884. 124. Huey, L. G.; Hanson, D. R.; Howard, C. J. I Phys. Chem. 1995, 99, 5001-5008. 125. Schott, G.; Davidson, N. J. Am. Chem. Soc. 1958, 80, 1841-1853. 126. Atkinson, R.; Carter, W. P. L.; Plum, C. N.; Winer, A. M.; Pitts, J. N. Int. I Chem. Kinet. 1984, 16, 887-898. 127. Wayne, R. P.; Barnes, I.; Biggs, P.; Burrows, J. P.; Canosamas, C. E.; Hjorth, J.; Lebras, G.; Moortgat, G. K.; Perner, D.; Poulet, G.; Restelli, G.; Sidebottom, H. Atmos. Environ. A-Gen. 1991, 25, 1-203. 128. Carslaw, K. S.; Clegg, S. L.; Brimblecombe, P. J. Phys. Chem. 1995, 99, 1155711574.  Chapter 2^ 129. Massucci, M.; Clegg, S. L.; Brimblecombe, P.1 Phys. Chem. A 1999, 103, 42094226. 130. Wexler, A. S.; Clegg, S. L. J Geophys. Res. 2002, 107, 4207, doi:4210.1029/2001JD000451. 131. Gaines Jr., G. L. Insoluble monolayers at liquid-gas interfaces; Interscience Publishers: New York, 1966. 132. Gericke, A.; Simonkutscher, J.; Huhnerfuss, H. Langmuir 1993, 9, 3115-3121. 133. Myrick, S. H.; Franses, E. I. Colloid. Surface. A 1998, 143, 503-515. 134. Park, S. Y.; Chang, C. H.; Ahn, D. J.; Franses, E. I. Langmuir 1993, 9, 3640-3648. 135. Park, S. Y.; Franses, E. Langmuir 1995, 11, 2187-2194. 136. Fluent Inc. 3d, segregated, laminar, version 6.2.16. 137. Knopf, D. A.; Cosman, L. M.; Mousavi, P.; Mokamati, S.; Bertram, A. K. J Phys. Chem. A 2007, 111, 11021-11032. 138. Solbrig, C. W.; Gidaspow, D. AIChE J. 1967, 13, 346-351. 139. Levich, V. G. Physicochemical hydrodynanmics; Prentice-Hall: Englewood Cliffs, New Jersey, 1962. 140. Brown, R. L. J. Res. Nat. Bur. Stand. 1978, 83, 1-8. 141. Hanson, D. R. I Phys. Chem. A 1998, 102, 4794-4807. 142. Motz, H.; Wise, H. J. Chem. Phys. 1960, 32, 1893-1894. 143. Walker, R. E. Phys. Fluids 1961, 4, 1211-1216. 144. Solbrig, C. W.; Gidaspow, D. Can. J. Chem. Eng. 1967, 45, 35-39. 145. Kulacki, F. A.; Gidaspow, D. Can. J. Chem. Eng. 1967, 45, 72-78. 146. Fuller, E. N.; Schettle.Pd; Giddings, J. C. Ind. Eng. Chem. 1966, 58, 19-27.  57  Chapter 2^  58  147. Mason, E. A.; Monchick, L. I Chem. Phys. 1962, 36, 2746-2757. 148. Monchick, L.; Mason, E. A. J. Chem. Phys. 1961, 35, 1676-1697. 149. Patrick, R.; Golden, D. M. Int. I Chem. Kinet. 1983, 15, 1189-1227. 150. Hanson, D. R.; Burkholder, J. B.; Howard, C. J.; Ravishankara, A. R. I Phys. Chem. 1992, 96, 4979-4985. 151. Moise, T.; Rudich, Y. I Geophys. Res.-Atmos. 2000, 105, 14667-14676. 152. Hanson, D. R.; Ravishankara, A. R. J. Geophys. Res. 1991, 96, 5081-5090. 153. de Gouw, J. A.; Lovejoy, E. R. Geophys. Res. Lett. 1998, 25, 931-934. 154. Vandoren, J. M.; Watson, L. R.; Davidovits, P.; Worsnop, D. R.; Zahniser, M. S.; Kolb, C. E. I Phys. Chem. 1991, 95, 1684-1689. 155. Mozurkewich, M.; Calvert, J. G. I Geophys. Res.-Atmos. 1988, 93, 15889-15896. 156. Hanson, D. R.; Lovejoy, E. R. Geophys. Res. Lett. 1994, 21, 2401-2404. 157. Hanson, D. R.; Ravishankara, A. R. I Phys. Chem. 1994, 98, 5728-5735.  59  Chapter 3^  3. N205 REACTIVE UPTAKE ON AQUEOUS SULFURIC ACID SOLUTIONS COATED WITH BRANCHED AND STRAIGHTCHAIN INSOLUBLE ORGANIC SURFACTANTS  3.1 Introduction Reactions between aerosol particles and gas phase species, termed heterogeneous reactions, can play a crucial role in the atmosphere.  158-161  Often the efficiency of these  heterogeneous reactions is described with the reactive uptake coefficient, y, which is defined as the fraction of collisions with a surface that leads to the irreversible loss of the gas-phase species due to a reaction. One heterogeneous reaction that has been studied extensively is the reaction between N205 and aqueous particles: R3.1  N205 (g) + 11 2 0(l)  aerosol  > 2 HNO3 (aq)  Modeling studies have demonstrated that this reaction can affect NO x , 03, and OH concentrations in the atmosphere. 162-164 For example, Dentener and Crutzen 162 demonstrated, using a global tropospheric model, that this heterogeneous reaction can decrease yearly averaged NO R , 03, and OH concentrations by 49%, 9%, and 9%, respectively, in the troposphere if the reactive uptake coefficient on aqueous particles is 0.1. Due to the importance of this heterogeneous reaction to the atmosphere, many research groups have investigated the reactive uptake coefficient of N2O5 on aqueous inorganic solutions and particles (see Sander et al.  165  and references therein). These  studies have shown that N2O 5 reactive uptake is efficient on aqueous inorganic solutions, with reactive uptake coefficients ranging from 0.015 to 0.2. 165 Most of the previous laboratory work on N2O5 reactive uptake has focused on aqueous inorganic solutions and particles free of organic surfactants. Nevertheless, field studies indicate that tropospheric inorganic aerosols can contain a significant amount of  A version of this chapter has been published. Cosman, L. M., Knopf, D. A., and Bertram, A. K., N 2 0 5 Reactive Uptake on Aqueous Sulfuric Acid Solutions Coated with Branched and Straight-Chain Insoluble Organic Surfactants, J Phys. Chem. A, 2008, 112, 2386-2396.  Chapter 3^  60  organic surfactants - both insoluble aqueous surfactants and/or soluble aqueous surfactants. See for example references  166-180  These surfactants can form organic  monolayers at the air-aqueous interface, 178-180 and depending on the composition and degree of compression of these organic monolayers, they may limit the transfer of molecules across the air-aqueous interface. 178-193 If this is the case, the N205 reactive uptake coefficients measured on uncoated aqueous inorganic solutions may not be applicable under all atmospheric conditions. Examples of atmospheric conditions where surfactants may be important include the marine boundary layer or continental regions influenced by forest fires, coal and straw burning. 166,170,173,175-177 Possibly related, recent field measurements over the northeast United States, by Brown et al. 194 showed that the reactive uptake coefficient of N205 can decrease significantly (by a factor of ?_ 10) when particles contain a large amount of organic material in addition to inorganic material. One possible explanation for these results is that the organic material formed a coating on the aqueous droplets, and this coating limited the transfer of N2O5 molecules across the air-aqueous interface. 194 More laboratory studies on the effects of organic monolayers on N205 heterogeneous chemistry would be useful to better understand the precise mechanism that led to the decrease in y observed by Brown et al. 194 A few laboratory studies have looked at the reactive uptake of N2O 5 on aqueous surfaces coated with monolayers consisting of straight-chain organic surfactants.  181,187-189  However, there has only been one study that has investigated the effect of monolayers consisting of a bent or branched surfactant on the N2O5 uptake coefficient, 195 even though a large fraction of atmospheric surfactants may have bent or branched structures. 179 ' 196 Furthermore, all of the previous measurements, except for the preliminary study carried out in our laboratory 189 focused on soluble organic surfactants. Additional studies with insoluble organic surfactants would be beneficial. In addition to the monolayer studies mentioned above, Folkers et al. 184 investigated the reactive uptake of N2O5 on aqueous inorganic aerosols coated with multilayers of organic material produced by the ozonolysis of a-pinene. Also, Badger et a1.  197  studied the reactive uptake of N2O5 on aerosol particles containing mixtures of humic acid (a water soluble surfactant) and ammonium sulfate.  Chapter 3^  61  The current manuscript focuses on the uptake of N20 5 on aqueous solutions coated with insoluble organic monolayers. Specifically, we focus on the reactive uptake coefficient of N205 on aqueous 60 wt % sulfuric acid solutions at 273 K coated with monolayers of 1-hexadecanol, 1-octadecanol, stearic acid, and phytanic acid (see Table 3.1 for chemical structures). 1-hexadecanol, 1-octadecanol and stearic acid are all straight-chain insoluble organic surfactants, while phytanic acid is a branched insoluble surfactant. The temperature of 273 K is relevant for the lower and middle troposphere. This temperature was chosen due to the atmospheric relevance and also due to experimental constraints. At warmer temperatures the vapor pressure of water over the aqueous solution is large, which significantly limits the range of reactive uptake coefficients we can measure with our experimental apparatus due to large gas-phase diffusion corrections (see below for further details). Table 3.1: Structure and Melting Points of Organic Surfactants Used to Form Monolayers in This Study  organic^melting^molecular structure compound^points' ('C) 1-hexadecanol^48-50  oo  N...,e4NN.e.eNNeee.NNNoe°4'44.,...eeN.N.F°N4Noee.N.NNon  d"..  1-octadecanol^56-59 stearic acid^67-72 (octadecanoic acid) phytanic acid^< 20 (3,7,11,15-  )CDH  tetramethylhexa -decanoic acid) a Melting points for 1-hexadecanol, 1-octadecanol, and stearic acid are obtained from Sigma-Aldrich, and the melting point for phytanic acid has not been reported but it is a liquid at room temperature.  Chapter 3^  62  For these studies we use a rectangular channel flow reactor to measure reactive uptake coefficients. One of the strengths of the experimental configuration is that we can prepare and study well characterized organic monolayers with this apparatus. In the following, we present measurements of the surface pressure of the prepared monolayers and the surface area occupied by each surfactant molecule in the monolayer (i.e. packing density of the monolayer). Then we present measurements of the reactive uptake coefficients. We then try to correlate trends in the measured reactive uptake coefficients with surfactant chain length, the surface pressure of the monolayer, and surface area occupied by each surfactant molecule to better understand the variables governing the reactive uptake coefficient. The atmospheric implications of these results are also discussed.  3.2 Experimental 3.2.1 Flow Reactor and Experimental Conditions for the Reactive Uptake Measurements Figure 3.1 is a detailed schematic of the flow reactor. Figure 3.1 a shows a top view with the cover removed, and Figure 3.1 b shows a side view. The apparatus has been described and characterized in detail in the previous chapter. Briefly, the flow cell is made entirely from aluminum, and it can be temperature controlled by circulating coolant through channels in the aluminum body. All interior aluminum walls are coated with halocarbon wax to minimize loss of N205 to the walls. Located on the bottom surface of the reactor is a glass trough, which is filled with the aqueous solution, and this aqueous solution can be covered with an organic monolayer. The surface of the liquid in the trough is 7.5 cm in width and 22 cm in length. The height of the head space (or open channel) above the liquid surface (Figure 3.1b) depends on the amount of liquid solution used in each experiment, and in most cases is less than 1 cm. N205 is introduced to the flow reactor by a movable T-shaped injector, which slides just above the liquid surface. The T-shaped injector is equipped with 6 exit holes 0.2 mm in diameter which point towards the top of the reactor and which distribute the gas phase reactant evenly across the width of the flow cell.  (a)  movable T-injector  N 2 0 5/He  L  carrier He/H 2 0  mixing barrier  ^  coated solution  (b)  Pcell thermocouples  N2/CH31  movable T-injector  N 2 0 5/He^T1^T2  ion^quadrupole pinhole optics^10-6Torr  II 1^  head space  210p ^ 1 1  I ^ carrier He/H 2 0  ,1 ion-molecule ^ reaction region 1 3-4 Torr^41 rotary mixing barrier^cooling coil^coated solution^quartz trough pump ,  Fl ow reactor  11 turbo  1 -10 I <,. 0  1  channeltron turbo multiplier pump  pump  CIMS  Figure 3.1: (a) A top view sketch of the rectangular channel flow reactor without the aluminum cover. The liquid solution is placed in a quartz trough located inside the flow reactor. (b) A side view of the rectangular channel flow reactor coupled to the chemical ionization mass spectrometer (CIMS).  Chapter 3^  64  The carrier gas (He) enters the flow reactor through inlets at the back of the flow cell. This gas stream within the reactor first flows against a barrier to ensure mixing before reaching the liquid surface. Thermocouples are used to determine the temperature of the liquid and the gas above the liquid. In all cases the liquid and the gas are within ± 0.5 K of each other. The pressure inside the flow cell is measured using an MKS baratron at the exit of the flow reactor (see Figure 3.1b). Prior to entering the flow cell, the carrier gas was first passed through a carbon filter (Supelco, Supelcarb HC) and a Drierite (W.A. Hammond Drierite Co. Ltd.) trap cooled with liquid nitrogen to remove any possible organic contamination. Then the carrier gas passed over a water reservoir held at a fixed temperature to adjust the relative humidity (RH) of the carrier gas. The RH was adjusted so that it matched the relative humidity over the specific aqueous sulfuric acid solution, calculated using the AIM model. 198-200 This ensured that there was no evaporation of water from the aqueous sulfuric acid solution over the course of the experiments. For these experiments the relative humidity above the aqueous solution was maintained at (14.5 ± 1.5)%. The RH of the carrier gas was verified with a dew point hygrometer. The open channel above the liquid surface has a rectangle geometry. The flow dynamics of the gas above the liquid surface has been characterized in our previous publication using computation fluid dynamics simulations. 189 These calculations show that the carrier gas reaches a fully developed laminar flow in less than 1.5 cm, which is much shorter than the length of the reactive surface (i.e. liquid surface). See our previous publication for further discussion on the gas flow dynamics in the system.  189  N205 was produced by reacting NO2 with an excess of 0 3 . 03 was generated by passing a flow of 02 over an ultraviolet source (Jelight, model #600). To remove water vapor from the 0 2 carrier gas, a Drierite trap was placed immediately before the UV lamp. The NO2 flow was passed through a P2O5 trap to remove trace amounts of water, prior to reaction with 03. N205 produced by this reaction flowed through an additional P2O5 trap to reduce the concentration of nitric acid, and was then collected and stored in a glass trap immersed in an ethanol bath cooled to 193 K. N205 condensed as white crystals inside the glass trap.  Chapter 3^  65  During uptake experiments, a saturated flow of N2 0 5 between 6 to 10 cm 3 min d at STP was mixed with 20-100 cm 3 min d at STP of dry He prior to entering the flow reactor through the T-shaped injector. Total mass flow rates inside the flow reactor ranged from 200 to 700 cm 3 mind at STP and total pressures ranged from 2.4 to 3.1 Torr. Under these conditions the Reynolds number varied from 0.4 to 1.4, indicating laminar flow conditions. The exit of the flow cell is connected to a chemical ionization mass spectrometer (CIMS), which is used to measure the change in the gas-phase reactant concentration as a result of reactive uptake at the liquid surface. 201,202 N2 0 5 was detected as NO3 - after its chemical ionization by I. 203-205 F was formed by flowing trace amounts of CH3I diluted in 1000 to 2000 cm 3 min d at STP of N2 through a polonium-210 source (NRD, model Po2031) for ionization. For N20 5 detection in the presence of H2O the chemical ionization region was biased to -122 V in order to fragment weakly bound ion-H20 clusters. 204,205 N20 5 concentrations of 2 x 10 10 to 1 x 10 11 molec cm -3 were used for the uptake measurements. N205 concentrations were based on the I - + N205 chemical ionization reaction rate that has been reported in the literature. 203 In a typical uptake experiment we measured the N2 0 5 signal as a function of injector position. The natural log of this signal was then plotted as a function of reaction time (time was calculated from the reaction length and the flow velocity), in order to determine the observed first order loss rate constant, k obs . Then from ka s we determined the first order wall loss rate constant, kw , using the procedure developed in our previous work. 189 This procedure corrects for any concentration gradients that can develop in the flow reactor due to a fast heterogeneous loss at the liquid surface. In other words, this procedure decouples reaction and diffusion to the aqueous surface in order to determine the true first order wall loss rate constant. In order to calculate kw from ko b s diffusion coefficients were needed. The diffusion coefficients applied in this study were taken from Knopf et al.'" and are based on calculations using molecular parameters. 206-208 For the diffusion coefficient of N2 0 5 in He (DN205-He) at 273 K, we used a value of 289 Ton cm2 coefficient of N2 0 5 in H2O  (DH205-H20)  s-1  ^for the diffusion  at 273 K, we used a value of 72 Ton cm 2 s d . To  Chapter 3^  66  calculate the diffusion coefficient, D, of N2 05 in a mixture of helium and water we used the following equation: 2°9  1 .^ PHe^PHp + D DN 2 0 -He DArp,-H p  3.1  5  where PHe and PH2,9 are the partial pressures of helium and water vapor in the flow reactor, respectively. The reactive uptake coefficient, 7, was determined from Ic ,„ using the following • 210,211 equation:  3.2  1^c S 1 = y 4k., V 2  where c is the mean molecular velocity of N205, S is the reactive surface area inside the flow reactor, and V is the volume of the open channel above the liquid surface (i.e. the volume of the head space illustrated in Figure 3.1b). 3.2.2 Preparation of the Organic Monolayers Prior to preparing an organic monolayer the surface of the aqueous solution (60 wt % sulfuric acid) was thoroughly cleaned with an aspirator to remove any organic contamination on the surface. Solutions of 1-octadecanol, 1-hexdecanol, stearic acid, and phytanic acid dissolved in chloroform (-1 mg cm 3 ) were prepared. Several droplets of the organic solution were deposited on the clean aqueous sulfuric acid surface. The chloroform evaporated, leaving behind an organic monolayer. An excess amount of organic material was used so that a few micro-crystals or lenses in the case of a solid or liquid organic monolayer, respectively, were left on the surface after a monolayer was established. Typically 25 % more organic material was added than required to attain a tightly packed monolayer. The presence of micro-crystals or lenses on the surface (which were verified visually) ensured that the aqueous solutions were completely coated with a surfactant monolayer. This method was chosen due to the reproducibility and ease of preparing this type of monolayer. Monolayers in contact with micro-crystals or lenses were used in all reported reactive uptake coefficient measurements. The presence of these crystals or lenses were not expected to affect the overall uptake coefficient of N205  Chapter 3^  67  since the surface area covered by the crystals or lenses was very small compared to the overall surface area exposed to N20 5 . To confirm this point, we carried out experiments where we varied the amount of excess organic material used in the uptake experiments. For example, we carried out some experiments where only enough organic material was added to the aqueous surface to form a tightly packed monolayer (i.e. no micro-crystals or lenses were formed). In these experiments the reactive uptake coefficients were within the experimental uncertainties of the uptake coefficients determined in the presence of the micro-crystals or lenses. This confirms that the presence of the micro-crystals or lenses had little effect on our uptake measurements. Also, if small islands consisting of multilayers of surfactant did form in our experiments when excess organic was used, they did not influence our results. 3.2.3 Measurements of the Surface Pressure and the Surface Area Occupied by Each Surfactant Molecule in the Monolayer  Prior to measuring the reactive uptake coefficients we first determined the surface pressure of the prepared monolayers (monolayers in contact with micro-crystals or lenses) and surface area occupied by each surfactant molecule in these monolayers, in order to better understand the physical properties of the monolayers under investigation. The surface pressure (g) of an organic monolayer at 273 K in the rectangular flow reactor was determined by first measuring the surface tension of the solution coated by an organic monolayer (o-fil m ) and the surface tension of the uncoated solution  (a0 ).  Surface  pressure was then calculated using the following equation: 212 33  7r = Cr 0 - Cr film  The surface tensions of coated and uncoated solutions were determined with the Whilhelmy plate method. In short, we measured the force (using a surface pressure sensor; NIMA Technology, model PS4) on a platinum plate while the plate was immersed and detached from the liquid solution. The surface tension was calculated from the maximum difference in the force on the plate between immersion and separation from the solution. 213,214 For the solutions coated with the organic monolayers, we measured the surface tension just before and after the reactive uptake experiments. In all cases the  Chapter 3^  68  results were the same within experimental error. The following procedure was employed to measure surface tensions within the rectangular channel flow reactor. First we degassed the aqueous solution. Second we prepared the monolayers and assembled the flow cell and established experimental conditions such as RH, temperature, and mass flow in order to condition the flow cell. Then after about 1 hour of conditioning, we removed the cover of the flow cell, measured the surface tension of the coated solution, reassembled the flow cell and performed the uptake experiments. In order to determine the surface area occupied by each surfactant molecule in the monolayers, we measured in a separate set of experiments the surface pressure-area isotherm for each organic surfactant on an aqueous 60 wt % H2504 solution at 273 K. The pressure-area isotherm illustrates the variation of the surface pressure with the area occupied by each surfactant molecule. Pressure-area isotherms were carried out using a commercial temperature controlled Langmuir film balance (NIMA Technology, model 611D). The Langmuir film balance consisted of a PTFE trough (with dimension of 20 cm by 30 cm), two movable barriers, and a surface pressure sensor (NIMA Technology, model PS4) with a platinum plate. The experimental procedure that was followed is described in detail by Myrick and Franses. 215 Briefly, the aqueous sulfuric acid solution was placed in the trough. The surface of the acid solution was cleaned thoroughly using an aspirator. A known volume of an organic solution (containing the surfactant and chloroform) was added to the clean H2SO4-H2O surface. The chloroform was allowed to evaporate leaving behind a known number of molecules on the surface. The surface pressure was then recorded as the moveable barriers reduced the available surface area, resulting in a surface pressure-area isotherm. 212 Examples of typical pressure-area isotherms determined in our studies are illustrated in Figure 3.2. Once the pressure-area isotherms and the surface pressure of the prepared monolayers were known, the corresponding molecular surface area occupied by each surfactant molecule within the monolayer can be easily read off the surface pressure-area isotherms.  ^  Chapter 3^  69  20  40  60  80  IIIIIIIIIIIIIII  60  (a)  7  E 40 Z E 20 0  ^I  60  E Z  E  60 40 20  1 -^1  'l  I  I  ^I^i ^I^I^I^ ^I^f^I^ ^I^I^I^  I  0 60  (e) 40  40  20  20  0 60  1^1^tit  ^I^1^1^1^I^1^1  0 I  60 40  E 40  z  E 20  20  0 60 7  0 (d) -..  60  E 40  40  20  20  Z  0  0 20^40^60^80  molecular surface area / A 2 molee Figure 3.2: Surface pressure - area isotherms for (a) 1-hexadecanol, (b) 1octadecanol, (c) stearic acid, and (d) phytanic acid performed on aqueous 60 wt % sulfuric acid at 273 K. The horizontal dashed lines represent the surface pressure of each monolayer measured in the kinetic uptake experiments and the vertical dashed lines represent the corresponding molecular surface area at those surface pressures. The collapse pressures are indicated by the horizontal dotted lines.  Chapter 3^  70  3.2.4 Further Characterization of the Organic Monolayers in the Flow Reactor During the reactive uptake coefficient measurements there is a steady gas flow above the organic monolayers. A flow above a monolayer could produce an additional horizontal force on the monolayers (i.e. surface shear stress), which could cause a pressure gradient in the monolayer along the length of the flow reactor. 216 However, since we are using low flow rates and low pressures, this force is minor, and at most can cause the surface pressure of the monolayer to increase by approximately 0.03 mN/m, 216 which is small compared to the surface pressures used in our experiments. Also note that in our experiments we used a range of flow velocities. For example, for a phytanic acid monolayer we used flow velocities ranging from 165 to 425 cm s -1 . We did not see any dependence of the reactive uptake coefficient on the flow velocity. If the force from the steady gas flow on the monolayer was significant in our experiments, we would expect to see a dependence of the uptake coefficient on the flow velocity since the surface stress is proportional to the square of the velocity. 3.2.5 Chemicals Listed below are the chemicals, the manufacturer, and the corresponding purities of the chemicals used in our studies: He (Praxair, 99.999% Purity), N2 (Praxair, 99.999%), NO2 (Matheson, 99.5%), 02 (Praxair, 99.5%), P2O5 (Aldrich, 97%), 1hexadecanol (Sigma-Aldrich, 99+%), 1-octadecanol (Sigma-Aldrich, 99%), stearic acid (Sigma-Aldrich, 98+%), phytanic acid (Sigma-Aldrich, 96%), chloroform (Fisher, 99+%), sulfuric acid (Fisher, 95-98%).  3.3 Results and Discussion 3.3.1 Properties of the Monolayers Table 3.2 lists the surface pressures of the prepared monolayers (monolayers in contact with a few micro-crystals or liquid lenses) on aqueous 60 wt % H2SO4 at 273 K measured in these experiments. The surface pressures for 1-hexadecanol, 1-octadecanol, stearic acid, and phytanic acid on 60 wt % sulfuric acid-water solutions at 273 K are 27.4 + 0.6 mN m -1 , 30.8 + 1.6 mN m -1 , 4.4 + 0.5 mN m -1 , and 23.9 ± 0.7 mN m -1 , respectively.  Chapter 3^  71  The stearic acid monolayer has a much lower surface pressure compared to the other surfactants.  Table 3.2: The Reactive Uptake of N20 5 on Aqueous 60 wt % Sulfuric Acid Solutions at 273 K in the Presence and Absence of Organic Monolayers monolayer^r of^surface monolayer^area  y^lower^upper^  limit^limit^  (mN m -1 ) (A 2/molec  none a  Nam/  Yuncoated b  4.9 x 10 -2  3.8 x 10 -2  1.0 x 10 -1  1  hexadecanol  27.4 ± 0.6  18.8 ± 0.5  8.9 x 10 -4  8.0 x 10 -4  9.8 x 10 -4  0.018  octadecanol  30.8 ± 1.6  19.8 ± 0.5  8.0 x 10 -4  6.3 x 10 -4  9.7 x 10 -4  0.016  stearic acid  4.4 ± 0.5  24.1 ± 0.5  3.0 x 10 -3  2.1 x 10 -3  3.8 x 10 -3  0.060  phytanic  23.9 ± 0.7  44.8 ± 0.5  5.4 x 10 -2  3.9 x 10 -2  1.5 x 10 -1  1.100  acid a Knopf et al. 189 b the uncertainty for nli n /yuncoated can be calculated using the lower and upper limits for n um and , uncoated and are shown in Figure 5.4.  Shown in Figure 3.2 are surface pressure-area (,r A) isotherms for 1-hexadecanol, 1-octadecanol, stearic acid, and phytanic acid on aqueous 60 wt % sulfuric acid solutions at 273 K. if-A isotherms were also obtained for 1-octadecanol on water at 295 K (not shown) and compared with literature data for validation of our experimental procedure. The results for 1-octadecanol on water agree well with those in the literature. 217 Also, the isotherms for 1-hexadecanol, 1-octadecanol, stearic acid, and phytanic acid on aqueous 60 wt % show similar trends (i.e. the positions of the "kinks" in the n-A isotherm - see description below) to the isotherms for the same compounds on water and other aqueous solutions reported in the literature. 212,217,218 -  For the straight-chain surfactants, at a high value for the molecular surface area per molecule the surface pressure is nearly zero. For a molecular surface area close to  72  Chapter 3^  approximately 25 A 2 molec -I the surface pressure increases rapidly until the monolayer collapses. The collapse pressure is indicated by the horizontal dotted lines in Figure 3.2. The collapse pressure of a monolayer depends not only on the nature of the surfactant molecules and the temperature and composition of the subphase, but also on experimental conditions such as rate of compression, previous history of the monolayer, and presence of impurities on the surface. 212 In our experiments we observed the collapse pressure to vary between 36 - 60 mN m -I for the straight-chain monolayers and 25.7 — 26.5 mN tn -1 for the branched monolayer. Besides the collapse pressures, the isotherms did not change significantly when the rate of compression was varied from 20 to 50 cm 2 min -I . In n-A isotherms for the straight-chain surfactants, there are several "kinks" in the isotherms, due to 2-D phase transitions. 212,218 These phase transitions correspond to different degrees of ordering of the organic molecules. 217 ' 218 At large molecular surface areas the monolayers exist as a 2-D gas on the aqueous acid surface, with molecules on the surface exerting relatively little force on each other due to sufficient separation.  212  For decreasing  molecular surface areas the monolayers undergo several phase transitions until they reach their collapse pressure. The phases associated with these phase transitions are generally referred to as gaseous, expanded liquid, condensed liquid, and condensed solid phases. 212 The behavior for the branched-chain monolayer is significantly different than that for the straight-chain surfactants. Between 45 and 80 A 2 molec -I , the branched-chain monolayer is in a liquid expanded state. 212 At a surface pressure of —26 mN m -I the monolayer collapses, and further compression of the monolayer results in the formation of liquid lenses in equilibrium with a monolayer at a molecular surface area of —44.5 A 2 molec -I . The 7r-A area isotherm obtained here is in good agreement with the 7r-A area isotherms of other branched surfactants reported in the literature. 212 Branched surfactants typically do not form condensed solid or condensed liquid states, because the side chains hinder a closed-packed molecular arrangement of the surfactant molecules. 212 Also indicated in Figure 3.2 as horizontal dashed lines are the measured surface pressures of the monolayers prepared for the reactive uptake experiments (i.e. monolayers in contact with micro-crystals or liquid lenses). The corresponding molecular surface areas occupied by the organic surfactants in these monolayers are given by vertical dashed lines and thus can be readily read off the Figure. The molecular  Chapter 3^  73  surface area occupied by each surfactant molecule is also included in Table 3.2. 1hexadecanol and 1-octadecanol are the most tightly packed monolayers with a molecular surface area per molecule of 18.8 ± 0.5 A 2 molec -1 and 19.7 ± 0.5 A 2 molec -1 , respectively. Stearic acid is intermediate with a surface area of 24.1 ± 0.5 A 2 molec -1 . Phytanic acid has a much larger surface area per molecule, 44.8 ± 0.5 A 2 molec -1 , compared to the straight-chain monolayers, indicating that the phytanic acid is less densely packed on the aqueous acidic surface. Phytanic acid is less efficient at packing due to the branched nature of the molecule, as mentioned above. The four methyl side chains on the long hydrocarbon tail of the phytanic acid prevent it from attaining a tightly packed molecular arrangement. The effect of the chain length, surface pressure, and molecular surface area on the reactive uptake coefficients is explored below. 3.3.2 Reactive Uptake Coefficients on Aqueous Solutions Covered With Organic Monolayers The reactive uptake coefficients are determined from the irreversible removal of N20 5 as a function of injector position as mentioned above. Shown in Figure 3.3 are plots of the natural logarithm of the N 2 0 5 signal as a function of reaction time for the loss of N205 on coated aqueous 60 wt % sulfuric acid solutions at 273 K. The data for each uptake experiment was fit to a straight line, and from the slope of this line the first-order rate constant, ko b„ was determined. From ko b„ we calculate kw and the reactive uptake coefficient.  Chapter 3^  0.00 0.0 t,^  74  0.02^0.04^0.06^0.08^0.10 1 1^I^1 0.0 IL  A  4,11  -0.3  -0.3  -0.6  -0.9  -0.9  -1.2  -1.2  0.00^0.02^0.04^0.06^0.08^0.10  fime/s  Figure 3.3: Natural logarithm of the observed N205 signal as a function of reaction time. Experiments were performed on aqueous 60 wt % H2SO4 solutions at 273 K. The lines represent the corresponding linear fits to the data. Open triangles: blank uptake, solid circles: 1-octadecanol, solid triangles: 1hexadecanol, solid diamonds: stearic acid, solid inverted triangles: phytanic acid.  The reactive uptake coefficients for N205 on aqueous 60 wt % H2SO4 at 273 K in the presence of monolayers prepared with branched and straight-chain surfactants are given in Table 3.2. 7-values reported in this study were based on at least 6 different uptake experiments performed on 2-3 freshly prepared monolayers. The upper and lower limit for ytake into account 20 % error in the diffusion coefficients. For organic coatings prepared with straight-chain surfactants, the reactive uptake coefficient was significantly less than for the uncoated solutions. The reactive uptake coefficient for N205 on aqueous 60 wt % sulfuric acid coated with a monolayer of 1-  75  Chapter 3^  octadecanol was (8.0 ± 1.7) x 10 4 , which is approximately a factor of 61 less than the observed uptake on aqueous 60 wt % H2SO4. Surfaces coated with a monolayer of a stearic acid and 1-hexadecanol monolayer showed a decrease in y by a factor of approximately 17 and 55, respectively, compared to the uncoated solution. In Table 3.3 and Figure 3.4 we compare our results for the straight-chain surfactants with previous measurements reported in the literature that also used straightchain organic surfactants. Previous studies have investigated monolayers of butanol, hexanol, hexanoic acid, sodium dodecyl sulfate (SDS), and 1-octadecano1.181,187189 With the exception of 1-octadecanol, these molecules are all soluble species in the aqueous subphase and partition to the surface to form a surface excess.  Table 3.3: A Comparison of Measured Reactive Uptake Coefficients of N205 on Aqueous H2SO4 and NaCI Solutions Coated With Straight-Chain Organic Monolayers of Different Chain Lengths literature  monolayer  subphase  T (K)  chain  yfii,/  length )uncoated  this study  1 -o ctadecanol  H2SO4/H20  273  18  0.016  this study  stearic acid  H2SO4/H20  273  18  0.060  Knopf et al: 89 93  1 -octadecanol  H2SO4/H20  298  18  0.018  this study  1-hexadecanol  H2SO4/H20  273  16  0.018  McNeill et al. 181 85  SDS  NaCl/H20  295  12  0.133  Thornton and  hexanoic acid  NaCI/H20  300  6  0.333  Park et al. 188 92  hexanol  H2 SO4/H20  216  6  0.400  Park et al. 188 92  butanol  H2SO4/H20  216  4  0.667  Abbatt 187 91  Chapter 3^  76  0 2 4 6 8 10 12 14 16 18 20 1•1  .  1•1•1•1•1•1'1  .  1'1  (a) 0.1  0.1  I  0.01  0.01  1 E-3  1 E-3 11111111111111111111i  .  (b):  1  1 a) 0  = E  0.1  0.1  0.01  0.01 0 2 4 6 8 10 12 14 16 18 20 Carbon chain length  Figure 3.4: Reactive uptake coefficients of N205 on aqueous H2SO4 and NaCI solutions coated with straight-chain organic monolayers of different chain lengths. Solid circle: 1-octadecanol (this study), solid diamond: stearic acid (this study), solid triangle: 1-hexadecanol (this study), solid inverted triangle: 1butanol (Park et al.), 188 solid pentagon: hexanol (Park et al.), 188 solid star: 1octadecanol (Knopf et al.), 189 solid sideways triangle: hexanoic acid (Thornton and Abbatt), 187 solid square: SDS (McNeill et al.). 1511 Panel (a) gives the ordinate as a function of y in the presence of the monolayer. In panel (b) the ordinate is normalized to y for the uncoated solution.  77  Chapter 3^  In Table 3.3 the reactive uptake coefficient in the presence of the organic monolayer, ;u m , was normalized to the reactive uptake coefficient for the uncoated solutions, Yuncoated, s imilar to the study by Park et al. 188 In Figure 3.4 we plot both the reactive uptake coefficient,  yfilm,  and the normalized reactive uptake  coefficient, yfilin/2/uncoated. The data in both Table 3.3 and Figure 3.4 illustrates that the reactive uptake coefficient decreases as the chain length increases. For reactive uptake of N205 in the presence of a branched monolayer of phytanic acid we measured a y value of 0.054+ror, . This value is not statistically different from the reactive uptake coefficient corresponding to the uncoated 60 wt % H2SO4-H2O solution (see Table 3.2). Thus, the presence of the branched phytanic acid monolayer does not appear to significantly affect the reactive uptake of N20 5 . Our work is the first to investigate the reactive uptake of N205 by aqueous H2SO4 solutions in the presence of branched monolayers. McNeill et al. 195 studied the loss of N20 5 on sub-micron aqueous NaCI particles coated with mixtures of oleate and oleic acid (bent aliphatic surfactant). These authors found that when the aqueous particles were covered with a full monolayer the N205 reactive uptake coefficient decreased by a factor of approximately 20. The difference between our results for a branched surfactant and the results by McNeill et al. 195 for a bent surfactant may be due to the structure of the monolayers or the difference in the aqueous subphases used in the two experiments. The latter point is discussed in more detail below. The effect of branched surfactants or bent surfactants on the transfer rate of other gases across the air-aqueous interface has been studied in the past. 183,185,219 Daumer et a1. 183 showed that a coating of 1-(hydroxymethyl)-adamantane (a branched hydrocarbon) had no significant effect on the neutralization rate of H2SO4 with NH3. Gilman and Vaida 185 demonstrated that the transport of acetic acid across the interface is impeded by long chain organic molecules such as 1-octadecanol, but unaffected by bent molecules such as cis-9-octadecen-1 -ol. Xiong et al. 219 did not observe any retardation of the hygroscopic growth of acidic droplets by a single monolayer of oleic acid. However, care should be exercised when extrapolating these results to the N205 system, since a different mechanism may be important in each case, and also each different study will  • Chapter 3^  78  have a different sensitivity to a possible change in the reactive uptake coefficients or mass accommodation coefficients. The mechanism responsible for the reaction between N205 and aqueous sulfuric acid solutions is believed to be an acid catalyzed mechanism: 220,221  ^R3.2^ N205 + H+ HNO3 + NO; ^R3.3^ NO + H2 O -+ HNO3 + H + Also, the overall uptake of N205 into the H2SO4 aqueous solution can be expressed using the resistor model: 222-224 ^1 3.4  _  1^1 S ^1 1^1 r  ▪ b  + Fs:of  s ^k sol  kdesorb where S is the sticking coefficient (fraction of collisions at the surface that result in accommodation on the surface), Icso , is the rate coefficient for the transfer from the surface into the liquid, kdesorb is the rate coefficient for the transfer of molecules from the surface into the gas phase, rb is the rate of reaction in the bulk of the solution, normalized to the gas-phase collision frequency, and T urf represents the surface reaction. See appendix A for a further discussion of this equation. The presence of the organic monolayer can influence y a number of different ways. First, the monolayer can influence the sticking coefficient, S, of N205 onto the surface and/or the transfer of N205 molecules from the surface into the liquid (k 01), by acting as a barrier to mass transfer. In addition, the organic monolayer could influence possible surface reactions by modifying T ug. We assume the presence of the monolayer does not influence the bulk reaction rate. Also, if the hydrolysis reaction occurs close to the surface, the carboxylic and/or alcohol functional groups on the surfactant molecules could potentially play a role. However, from our results the reactive uptake coefficient correlates best with the packing density of the organic surfactants, not the functional groups on the surfactants (see below for more details).  79  Chapter 3^  3.3.3 Correlation Between Reactive Uptake Coefficients and Carbon Chain Length, Surface Pressure, and Molecular Surface Area  The reactive uptake coefficient is expected to be a function of several parameters including the molecular surface area of the surfactant, the carbon chain length, the structure of the surfactant, the surface pressure, and the aqueous subphase. In the following analysis we compare measured y values with several of these parameters (carbon chain length, surface pressure, and molecular surface area) in order to determine if there is one parameter that dominates the reactive uptake coefficient. For the first part of this analysis we only use our data and the data from Park et al.  188  and Knopf et al.' 89  since these data sets were all carried out with the same subphase (aqueous sulfuric acid solutions). The data from Park et a1. 188 were carried out on 72 wt % H2SO4 solution at 216 K. We assume here that our data carried out at 273 K is directly comparable to the data obtained by Park et al.  ;  188  however, further work is needed to verify this assumption.  After this analysis, we will also discuss the results obtained with other subphases. In Figure 3.5 we have plotted the reactive uptake coefficient measured on aqueous sulfuric acid solutions as a function of carbon chain length. Figure 3.5a shows nu m and Figure 3.5b shows rfilm/iuncoated, similar to Park et al. 188 The solid symbols are results for straight-chain organic surfactants including data obtained in this study and those reported in the literature.' 88 " 89 The open symbols represent the results obtained in this study for the branched monolayer. For the straight-chain surfactants there appears to be a correlation between y and the length of the hydrocarbon chain, although there is some scatter in the data for carbon chain lengths between C 16 and C 18. This scatter may be due to differences in the properties used in the different experiments, such as molecular surface area (see below). Nevertheless, a trend is apparent. In contrast, the branched result deviates drastically from the straight-chain trend.  Chapter 3^  80  0 2 4 6 8 10 12 14 16 18 20 1^•^1^•^1^1^1^1^1^1  0.1  0.1  0.01  0.01  1E-3  1 E-3 II  IiIlIlIltjIlIlIlIl  (b): 1  1  is  0 0  E  0.1  0.1  L.=  0.01  0.01 1^.^1^.1^1.^1.^1^.1^1.^1^.  0 2 4 6 8 10 12 14 16 18 20  Carbon chain length Figure 3.5: Reactive uptake coefficients for N205 on organic coated sulfuric acid solutions as a function of carbon chain length. Solid symbols represent straight chain molecules and the open symbols represent the branched molecule. Solid circle: 1-octadecanol (this study), solid diamond: stearic acid (this study), solid triangle: 1-hexadecanol (this study), open circle: phytanic acid (this study), solid inverted triangle: 1-butanol (Park et al.), 188 solid pentagon: hexanol (Park et al.), 188 solid star: 1-octadecanol (Knopf et al.). 189 Panel (a) gives the ordinate as a function of y in the presence of the monolayer. In panel (b) the ordinate is normalized to y for the uncoated solution.  Chapter 3^  81  The reactive uptake coefficient was plotted as function of surface pressure in Figure 3.6. Figure 3.6a plots 'film and Figure 3.6b shows 7mm/2/uncoated. Similar to Figure 3.5, the solid symbols correspond to the results for straight-chain surfactants and the open symbols correspond to the branched surfactant. Clearly, ymm and 7mm/2/uncoated do not correlate with the surface pressure of the monolayers. In Figure 3.7, we have plotted the reactive uptake coefficient as a function of the surface area occupied by each surfactant molecule. The solid symbols correspond to the results for straight-chain surfactants and the open symbols correspond to the branched surfactant. The dashed line represents a sigmoidal fit (Figure 3.7) to the straight-chain data. This fit was chosen because it resulted in the best fit to the data. However, it serves solely to guide the eye. The data corresponding to the surfactants follow a trend: as the molecular surface area decreases the reactive uptake coefficient decreases. This is to be expected, because, as the molecular surface area decreases, the monolayer becomes more densely packed, and should limit the transfer of N205 across the air-aqueous interface. Within the experimental uncertainty the reactive uptake coefficient obtained for the branched surfactant follows the trend observed for the reactive uptake coefficients obtained for the straight-chain surfactants. Therefore, we speculate that the reason the branched monolayers do not decrease y significantly is because these monolayers are not densely packed. Based on this very limited set of data, we suggest that the molecular surface area is the best parameter for predicting the influence of an organic monolayer on the reactive uptake coefficient (at least for an aqueous sulfuric acid subphase) since it can explain, reasonably well, the trends for both the straight-chain surfactants and the branched surfactants.  Chapter 3^  82  0 5 10 15 20 25 30 35 40 45 50  (a) 0.1 F^4-, 4-,  0.1  E 0.01  0.01  4L  1 E-3  1 E-3  ,  (b) 1  ++  T  1  to 0  0 C  E  0 1 .  0.1  + 0.01  0.01 1^.^1^.^1^.^1^.^1^.^1  1^•^1  0 5 10 15 20 25 30 35 40 45 50  it / mN m -1 Figure 3.6: Reactive uptake coefficients for N205 uptake on organic coated aqueous sulfuric acid solutions as a function of monolayer surface pressure. Solid symbols represent straight chain molecules and the open symbols represent the branched molecule. Solid circle: 1-octadecanol (this study), solid diamond: stearic acid (this study), solid triangle: 1-hexadecanol (this study), open circle: phytanic acid (this study), solid inverted triangle: butanol (Park et al.), 188 solid pentagon: hexanol (Park et al.), 188 solid star: 1-octadecanol (Knopf et al.). 189 Panel (a) gives the ordinate as a function of yin the presence of the monolayer. In panel (b) the ordinate is normalized to yfor the uncoated solution.  Chapter 3^  83  20 30 40 50 60 70 80 90 I^•^I  •^I^•^I^I^•  (a) 0.1  0.1  E 0.01  0.01  1 E -3  1 E-3  (b) 1  0.1  0.01  0.01 I^.^I  20 30 40 50 60 70 80 90  molecular surface area / A 2 molec -i Figure 3.7: Reactive uptake coefficients for N205 uptake on organic coated aqueous sulfuric acid solutions as a function of packing density. Solid symbols represent straight chain molecules and the open symbols represent the branched molecule. Solid circle: 1-octadecanol (this study), solid diamond: stearic acid (this study), solid triangle: 1-hexadecanol (this study), open circle: phytanic acid (this study), solid inverted triangle: butanol (Park et al.), 188 solid pentagon: hexanol (Park et al. ),188 solid star: 1-octadecanol (Knopf et al.). 189 The dashed line represents a sigmoidal fit to the straight chain molecules. This fit was chosen because it gave a reasonable fit to the data, but has no physical meaning. Panel (a) gives the ordinate as a function of yin the presence of the monolayer. In panel (b) the ordinate is normalized to y for the uncoated solution.  Chapter 3^  84  In Figure 3.5-Figure 3.7 we only included data obtained with an aqueous sulfuric acid subphase. In Figure 3.8 we also included data corresponding to aqueous sea salt aerosols and aqueous NaC1 particles. The solid symbols correspond to data obtained with an aqueous sulfuric acid subphase (the same data that was shown in Figure 3.7), and the open symbols correspond to data obtained with other subphases. Thornton and Abbatt 187 studied the loss of N205 on aqueous sea salt aerosols coated with hexanoic acid (straightchain C6 surfactant). McNeill et a1. 181 studied the reactive uptake coefficient on aqueous NaCl aerosols coated with sodium dodecyl sulfate (straight-chain C12 surfactant) and McNeill et al. 195 studied the loss of N20 5 on sub-micron aqueous NaCI particles coated with mixtures of oleate and oleic acid (bent C 18 surfactant). In the studies by Thornton and Abbatt 187 the molecular surface areas were estimated from bulk surface tension measurements. McNeill et al. 181 and McNeill et al. 195 estimated the molecular surface areas of their monolayers indirectly based on an observed plateau in the kinetics N 2 0 5 uptake data.  85  Chapter 3^  20 30 40 50 60 70 80 90 I^I^I  0. 1  0.1 E 0.01  0.01  1 E-3  1 E-3  (b)  1  1  0. 1  0.01  0.01 I^.1^1^I^.^I^.  20 30 40 50 60 70 80 90 molecular surface area / A 2 molec -1  Figure 3.8: Reactive uptake coefficients for N205 uptake on aqueous solutions and aerosols as a function of packing density. Solid symbols represent data collected on aqueous sulfuric acid subphases and the open symbols represent data collected on other subphases. Solid circle: 1-octadecanol (this study), solid diamond: stearic acid (this study), solid triangle: 1-hexadecanol (this study), solid square: phytanic acid (this study), solid inverted triangle: butanol (Park et al.), 188 solid pentagon: hexanol (Park et al.), 188 solid star: 1-octadecanol (Knopf et al.), 189 open circle: hexanoic acid (Thornton and Abbatt), 187 open square: SDS (McNeill et al.), 181 open triangle: oleate (McNeill et al.). 198 Panel (a) gives the ordinate as a function of y in the presence of the monolayer. In panel (b) the ordinate is normalized to yfor the uncoated solution.  Chapter 3^  86  It is apparent that the data obtained with an aqueous sulfuric acid subphase is not in agreement with other subphases when the reactive uptake coefficient is plotted versus the molecular surface area. The reason for the difference is not clear, but perhaps it suggests that a different mechanism is important for the different subphases. Alternatively, molecular surface area is only one of the important parameters, and other variables such as chain length and solution temperature need to be considered when assessing the overall reactivity and explaining all the experimental data. Yet another alternative may be related to the experimental techniques. All the aqueous sulfuric acid experiments were carried out with bulk solutions coated with surfactants, whereas the other experiments were all carried out with aerosols. As pointed out by McNeill et a1. 181 ' 195 and Thornton and Abbatt, 187 their aerosol generation method leads to uncertainty in the actual mixing state of the surface active organics. Nonetheless, their interpretation of the N205 kinetics yields predictions of an area per molecule that are strikingly similar to equilibrium values determined on macroscopic systems. Thus, while perhaps the aerosol measurements with NaCI or seawater aerosols are not directly comparable with the sulfuric acid data, the differences illustrated in Figure 3.8 suggest that the nature of the subphase may play an important role in the effect of surface-active organics on net reactive uptake. Also interesting, in contrast to Figure 3.8, the data from McNeill et al 181'195 and Thornton and Abbatt 187 are in good agreement with the data obtained with aqueous sulfuric acid subphases (except for our phytanic acid data) when the reactive uptake coefficient is plotted against the carbon chain length (See Table 3.3 and Figure 3.4). The issues mentioned above should be addressed by future experiments.  3.4 Atmospheric Implications Our results suggest that insoluble straight-chain organic surfactants can decrease the reactive uptake coefficient by a factor of 17-61. This decrease in the reactive uptake coefficient is the same order of magnitude as the decrease in reactive uptake coefficient observed by Brown et al. 194 during a recent field measurement over the northeast United States. These authors showed that the reactive uptake coefficient of N205 can decrease significantly (by a factor of 10) when particles contain a large amount of organic material in addition to inorganic material. However, keep in mind that when Brown et  Chapter 3^  87  al. 194 observed a drastic decrease in the reactive uptake coefficient, the inorganic material consisted of ammonium sulfate. As a result, our results may not be directly comparable to the field measurements by Brown et al. 194 since we used a different subphase. Experiments with an aqueous ammonium sulfate subphase would be beneficial. In contrast to straight-chain monolayers, phytanic acid (an insoluble branched monolayer) showed no significant effect on the uptake of N205 (the decrease in the uptake coefficient was less than the uncertainty in our measurements). This result highlights the need for studies that focus on the physical and chemical properties of organic surfactants that reside on the surface of aqueous particles in the atmosphere. Researchers have begun to consider the detailed structure of organic monolayers on aerosol particles (see for example Seid1 225 ), but more studies in this direction would be beneficial. Also, studies that investigate the effect of other types of branched and bent surfactants on the reactive uptake of N205 employing atmospherically relevant subphases would be helpful. In the atmosphere, aerosol particles will most likely consist of a mixture of different organic surfactants. Studies on the effect of mixed monolayers (i.e. monolayers containing both straight-chain and branched surfactants) would also be informative.  3.5 Conclusions and Summary A rectangular channel flow reactor coupled to a chemical ionization mass spectrometer was used to study the reactive uptake coefficients of N20 5 on aqueous 60 wt % sulfuric acid solutions at 273 K coated with insoluble organic monolayers. Both straight-chain (1-hexadecanol, 1-octadecanol, and stearic acid) monolayers and branched (phytanic acid) monolayers were studied. The reactive uptake coefficients obtained were (8.9 ± 0.9) x 10 4 , (8.0 ± 1.7) x 10 4 , (3.0 ± 0.8) x 10 -3 , and (5.4+_:) x 10 -2 for 1hexadecanol, 1-octadecanol, stearic acid, and phytanic acid respectively. The reactive uptake coefficient decreased drastically for straight-chain surfactants. The decrease ranged from a factor of 17 to a factor of 61 depending on the type of straight-chain surfactant. In contrast to the straight-chain data, N205 uptake in the presence of phytanic acid, which has a branched structure, did not have a significant effect on the N205  Chapter 3^  88  reactive uptake coefficient (the decrease was less than the uncertainty in the data) compared to the uncoated solution. In addition to measuring the reactive uptake coefficients, we also tried to correlate properties of the monolayers with the reactive uptake coefficients. Based on a limited set of data, the reactive uptake coefficients measured on aqueous sulfuric acid subphases coated with organic monolayers show a relationship to the surface area occupied by each surfactant molecule. The apparent correlation between y and the molecular surface area of the surfactant may be tentatively explained by the fact that mass transport is hindered by tight-packed surfactants compared to less densely packed ones (branched surfactants). This leads to the possible conclusion that the overall uptake process is governed by mass transport rather than by reaction. On the other hand the aqueous subphase seems to influence y significantly (see Figure 3.8) which suggests a dependence on the reaction mechanism. This apparent complexity should be investigated with future studies. Our results also highlight the need for further studies that focus on the physical and chemical properties of the organic surfactants that reside on the surface of aqueous particles in the atmosphere.  Chapter 3^  89  3.6 References 158. Poschl, U. Angew. Chem. Int. Edit. 2005, 44, 7520-7540. 159. Zhang, Y.; Carmichael, G. R. J. Appi. Meteorol. 1999, 38, 353-366. 160. Molina, M. J.; Molina, L. T.; Kolb, C. E. Annu. Rev. Phys. Chem. 1996, 47, 327367. 161. Ravishankara, A. R. Science 1997, 276, 1058-1065. 162. Dentener, F. J.; Crutzen, P. J. J. Geophys. Res.-Atmos. 1993, 98, 7149-7163. 163. Evans, M. J.; Jacob, D. J. Geophys. Res. Lett. 2005, 32, L09813, doi:09810.01029/02005GL022469. 164. Tie, X.; Brasseur, G.; Emmons, L.; Horowitz, L.; Kinnison, D. I Geophys. Res.Atmos. 2001, 106, 22931-22964. 165. Sander, S. P.; Golden, D. M.; Kurylo, M. J.; Moortgat, G. K.; Wine, P. H.; Ravishankara, A. R.; Kolb, C. E.; Molina, M. J.; Finlayson-Pitts, B. J.; Huie, R. E. "Chemical Kinetics and Photochemical Data for Use in Atmospheric Studies Evaluation Number 15," JPL Publication 06-2, 2006. 166. Barger, W. R.; Garrett, W. D. J. Geophys. Res. 1970, 75, 4561-4566. 167. Barger, W. R.; Garrett, W. D. J. Geophys. Res.-0c. Atm. 1976, 81, 3151-3157. 168. Cavalli, F.; Facchini, M. C.; Decesari, S.; Mircea, M.; Emblico, L.; Fuzzi, S.; Ceburnis, D.; Yoon, Y. J.; O'Dowd, C. D.; Putaud, J. P.; Dell'Acqua, A. I Geophys. Res.-Atmos. 2004, 109, D24215, doi:24210.21029/22004JD005137. 169. Decesari, S.; Facchini, M. C.; Fuzzi, S.; Tagliavini, E. I Geophys. Res.-Atmos. 2000, 105, 1481-1489.  Chapter 3^  90  170. Mochida, M.; Kitamori, Y.; Kawamura, K.; Nojiri, Y.; Suzuki, K. I Geophys. Res.  2002, 107, 4325, doi:4310.1029/2001JD001278. 171. O'Dowd, C. D.; Facchini, M. C.; Cavalli, F.; Ceburnis, D.; Mircea, M.; Decesari, S.; Fuzzi, S.; Yoon, Y. J.; Putaud, J. P. Nature 2004, 431, 676-680. 172. Peterson, R. E.; Tyler, B. J. Appl. Surf Sci. 2003, 203, 751-756. 173. Peterson, R. E.; Tyler, B. J. Atmos. Environ. 2002, 36, 6041-6049. 174. Russell, L. M.; Maria, S. F.; Myneni, S. C. B. Geophys. Res. Lett. 2002, 29, doi:10.1029/2202GL014874. 175. Tervahattu, H.; Hartonen, K.; Kerminen, V. M.; Kupiainen, K.; Aarnio, P.; Koskentalo, T.; Tuck, A. F.; Vaida, V. J. Geophys. Res.-Atmos. 2002, 107, 4053, doi:4010.1029/2000JD000282. 176. Tervahattu, H.; Juhanoja, J.; Kupiainen, K. I Geophys. Res.-Atmos. 2002, 107, 4319, doi:4310.1029/2001JD001403. 177. Tervahattu, H.; Juhanoja, J.; Vaida, V.; Tuck, A. F.; Niemi, J. V.; Kupiainen, K.; Kulmala, M.; Vehkamaki, H. I Geophys. Res. Atmos. 2005, 110, D06207, doi:06210.01029/02004JD005400. 178. Ellison, G. B.; Tuck, A. F.; Vaida, V. I Geophys. Res.-Atmos. 1999, 104, 1163311641. 179. Gill, P. S.; Graedel, T. E.; Weschler, C. J. Rev. Geophys. 1983, 21, 903-920. 180. Donaldson, D. J.; Vaida, V. Chem. Rev. 2006, 106, 1445-1461. 181. McNeill, V. F.; Patterson, J.; Wolfe, G. M.; Thornton, J. A. Atmos. Chem. Phys.  2006, 6, 1635-1644.  Chapter 3^  91  182. Anttila, T.; Kiendler-Scharr, A.; Tillmann, R.; Mentel, T. F. J. Phys. Chem. A 2006, 110, 10435-10443. 183. Daumer, B.; Niessner, R.; Klockow, D. I Aerosol Sci. 1992, 23, 315-325. 184. Folkers, M.; Mentel, T. F.; Wahner, A. Geophys. Res. Lett. 2003, 30, 1644, doi:1610.1029/2003GL017168. 185. Gilman, J. B.; Vaida, V. I Phys. Chem. A 2006, 110, 7581-7587. 186. Lawrence, J. R.; Glass, S. V.; Park, S. C.; Nathanson, G. M. I Phys. Chem. A 2005, 109, 7458-7465. 187. Thornton, J. A.; Abbatt, J. P. D. I Phys. Chem. A 2005, 109, 10004-10012. 188. Park, S. C.; Burden, D. K.; Nathanson, G. M. I Phys. Chem. A 2007, 111, 29212929. 189. Knopf, D. A.; Cosman, L. M.; Mousavi, P.; Mokamati, S.; Bertram, A. K. J. Phys. Chem. A 2007, 111, 11021-11032. 190. Clifford, D.; Bartels-Rausch, T.; Donaldson, D. J. Phys. Chem. Chem. Phys. 2007, 9, 1362-1369. 191. Barnes, G. T. Colloid. Surface. A 1997, 126, 149-158. 192. Jefferson, A.; Eisele, F. L.; Ziemann, P. J.; Weber, R. J.; Marti, J. J.; McMurry, P. H. I Geophys. Res.-Atmos. 1997, /02, 19021-19028. 193. Rubel, G. 0.; Gentry, J. W.J. Aerosol Sci. 1985, 16, 571-574. 194. Brown, S. S.; Ryerson, T. B.; Wollny, A. G.; Brock, C. A.; Peltier, R.; Sullivan, A. P.; Weber, R. J.; Dube, W. P.; Trainer, M.; Meagher, J. F.; Fehsenfeld, F. C.; Ravishankara, A. R. Science 2006, 311, 67-70.  Chapter 3^  92  195. McNeill, V. F.; Wolfe, G. M.; Thornton, J. A. J. Phys. Chem. A 2007, 111, 10731083. 196. Kanakidou, M.; Seinfeld, J. H.; Pandis, S. N.; Barnes, I.; Dentener, F. J.; Facchini, M. C.; Van Dingenen, R.; Ervens, B.; Nenes, A.; Nielsen, C. J.; Swietlicki, E.; Putaud, J. P.; Balkanski, Y.; Fuzzi, S.; Horth, J.; Moortgat, G. K.; Winterhalter, R.; Myhre, C. E. L.; Tsigaridis, K.; Vignati, E.; Stephanou, E. G.; Wilson, J. Atmos. Chem. Phys. 2005, 5, 1053-1123. 197. Badger, C. L.; Griffiths, P. T.; George, I.; Abbatt, J. P. D.; Cox, R. A. J. Phys. Chem. A 2006, 110, 6986-6994. 198. Massucci, M.; Clegg, S. L.; Brimblecombe, P. 1 Phys. Chem. A 1999, 103, 42094226. 199. Wexler, A. S.; Clegg, S. L. J. Geophys. Res. 2002, 107, 4207, doi:4210.1029/2001JD000451. 200. Carslaw, K. S.; Clegg, S. L.; Brimblecombe, P. J. Phys. Chem. 1995, 99, 1155711574.^  •  201. Knopf, D. A.; Anthony, L. M.; Bertram, A. K. I Phys. Chem. A 2005, 109, 55795589. 202. Knopf, D. A.; Mak, J.; Gross, S.; Bertram, A. K. Geophys. Res. Lett. 2006, 33, L17816, doi:17810.11029/12006GL026884. 203. Huey, L. G.; Hanson, D. R.; Howard, C. J. I Phys. Chem. 1995, 99, 5001-5008. 204. Hu, J. H.; Abbatt, J. P. D. I Phys. Chem. A 1997, 101, 871-878. 205. Thornton, J. A.; Braban, C. F.; Abbatt, J. P. D. Phys. Chem. Chem. Phys. 2003, 5, 4593-4603.  Chapter 3^ 206. Mason, E. A.; Monchick, L. J. Chem. Phys. 1962, 36, 2746-2757. 207. Monchick, L.; Mason, E. A. J Chem. Phys. 1961, 35, 1676-1697. 208. Patrick, R.; Golden, D. M. Int. I Chem. Kinet. 1983, 15, 1189-1227. 209. Hanson, D. R.; Ravishankara, A. R. I Geophys. Res. 1991, 96, 5081-5090. 210. Hanson, D. R. J. Phys. Chem. A 1998, 102, 4794-4807. 211. Motz, H.; Wise, H. I Chem. Phys. 1960, 32, 1893-1894. 212. Gaines Jr., G. L. Insoluble monolayers at liquid-gas interfaces; Interscience Publishers: New York, 1966. 213. Rolo, L. I.; Caco, A. I.; Queimada, A. J.; Marrucho, I. M.; Coutinho, J. A. P. J. Chem. Eng. Data 2002, 47, 1442-1445. 214. Adamson, A. W.; Gast, A. P. Physical chemistry of surfaces, 6th Ed. ed.; WileyInterscience: New York, 1997. 215. Myrick, S. H.; Franses, E. I. Colloid. Surface. A 1998, 143, 503-515. 216. Vines, R. G. Q. J. Roy. Meteor. Soc. 1959, 85, 159-162. 217. Lawrie, G. A.; Barnes, G. T. J. Colloid Interf Sci. 1994, 162, 36-44. 218. Kaganer, V. M.; Mohwald, H.; Dutta, P. Rev. Mod. Phys. 1999, 71, 779-819. 219. Xiong, J. Q.; Zhong, M. H.; Fang, C. P.; Chen, L. C.; Lippmann, M. Environ. Sci. Technol. 1998, 32, 3536-3541. 220. Hallquist, M.; Stewart, D. J.; Baker, J.; Cox, R. A. I Phys. Chem. A 2000, 104, 3984-3990. 221. Robinson, G. N.; Worsnop, D. R.; Jayne, J. T.; Kolb, C. E.; Davidovits, P. J. Geophys. Res.-Atmos. 1997, 102, 3583-3601.  93  Chapter 3^  94  222. Schwartz, S. E. "Chemistry of multiphase atmospheric systems"; NATO ASI Series, 1986, Springer- Verlag, Berlin. 223. Jayne, J. T.; Duan, S. X.; Davidovits, P.; Worsnop, D. R.; Zahniser, M. S.; Kolb, C. E. I Phys. Chem. 1992, 96, 5452-5460. 224. Hanson, D. R. I Phys. Chem. B 1997, 101, 4998-5001. 225. Seidl, W. Atmos. Environ. 2000, 34, 4917-4932.  95  Chapter 4^  4. REACTIVE UPTAKE OF N205 ON AQUEOUS H2SO4 SOLUTIONS COATED WITH 1-COMPONENT AND 2-COMPONENT MONOLAYERS  4.1 Introduction Heterogeneous reactions between aerosol particles and gas phase species can play a key role in the atmosphere. 226-236 One heterogeneous reaction that has been studied extensively is the reaction between N205 and aqueous particles: R4.1  N 2 05 (g)  + H 2 0(l)  aeroso l >  2 HNO3 (aq)  Modeling studies have demonstrated that the rate for this reaction needs to be known accurately in order to predict tropospheric concentrations of NO,,,  03,  and OH. 227,231,237  Often the efficiency of N 2 0 5 heterogeneous reactions is described by the reactive uptake coefficient (y), which is defined as the fraction of collisions with a surface that leads to the irreversible loss of the gas-phase species due to a reaction. Since the initial modeling studies that demonstrated the atmospheric importance of the N205 heterogeneous reaction, there has been extensive laboratory investigations of the reactive uptake coefficient of N205 on aqueous inorganic solutions and particles. 238-249 These studies have shown that N205 reactive uptake is large (y is between 0.02 and 0.15) on aqueous inorganic particles. In addition to inorganic material, tropospheric particles can also contain a significant fraction of organic surfactants. 25° These surfactants can form organic monolayers at the air-aqueous interface, 251-261 and depending on the composition and degree of compression of these organic monolayers, they may limit the transfer of molecules across the air-aqueous interface. 251-253,262-276 Possibly related, recent field measurements over the northeast United States, by Brown et al. 277 showed that the reactive uptake coefficient of N205 can decrease A version of this chapter has been accepted for publication. Cosman, L. M. and Bertram, A. K., Reactive uptake of N 2 0 5 on aqueous H2 SO 4 solutions coated with 1component and 2-component monolayers, J. Phys. Chem. A, 2008, 112.  96  Chapter 4^  significantly (by a factor of 10) when particles contain a large amount of organic material in addition to inorganic material. One possible explanation for these results is that the organic material formed a coating on the aqueous droplets, and this coating limited the transfer of N2O5 molecules across the air-aqueous interface. 277 Recently, several research groups have investigated the effect of organic monolayers on reactive uptake of N2O5 on aqueous solutions or particles. 262,263,269-271,276 Park et al. 27° studied N205 uptake on H2SO4 solutions and showed that 7 decreased by a factor of 2.5 in the presence of hexanol monolayers, and decreased by a factor of 1.5 in the presence of butanol monolayers. Thornton and Abbatt 269 investigated N20 5 uptake on sea salt aerosols and determined that the uptake coefficient decreased by a factor of 3-4 in the presence of hexanoic acid monolayers. McNeill et  al.  262,263  studied uptake of N205 on  aqueous salt aerosols, and they observed a decrease in 7 by a factor of 7.5 in the presence of monolayers formed from sodium dodecyl sulfate, 262 and a factor of 20 in the presence of monolayers formed from sodium oleate. 263 More recently, we measured the reactive uptake of N205 on H2SO4 solutions coated with 1- octadecanol, 1-hexadecanol, and stearic acid and observed a decrease in the reactive uptake coefficient by a factor of 61, 55, and 17, respectively. 276 In contrast, we found that an insoluble, branched monolayer containing 16 carbon atoms (phytanic acid) did not have an effect on y within our experimental error. 276 Related to the above, Folkers et al. 266 investigated the uptake of N205 on aqueous NH4HSO4 particles coated with organics produced by the ozonolysis of a-pinene. In these studies it was observed that an organic film approximately 15 nm thick reduced the reaction probability of N205 by approximately a factor of 5. Also, Badger et al. 278 investigated N2O5 uptake on aqueous ammonium sulfate aerosols containing humic acid (a surfactant). For aerosols containing only 6 % humic acid by dry mass, a decrease in reactivity of more than a factor of 2 was observed compared with the case for singlecomponent ammonium sulfate. Analysis of the uptake coefficients using the water concentration data showed that the change in reactivity could not be explained by the change in water content alone. The authors attributed the larger than expected change in reactivity to the formation of organic films by the humic acid surfactants.278  Chapter 4^  97  Despite the recent progress on the effect of organic monolayers on N205 reactive uptake, much remains to be done. For example, in most of the previous studies only single-component monolayers were investigated. More work on multi-component monolayers (monolayers containing more than one type of organic surfactant) is needed. Also, much remains to be learned in order to understand the mechanism by which monolayers and films affect N205 uptake. Uptake studies as a function of molecular surface area of the surfactant, the 2-D phase of the organic monolayer, and temperature would be useful to better understand the physical chemistry. This type of information is also necessary in order to accurately extrapolate laboratory data to atmospheric conditions. As mentioned above, we recently studied N20 5 reactive uptake on aqueous sulfuric acid solutions coated with 1-octadecanol, 1-hexadecanol, stearic acid, and phytanic acid. 276 In these studies we only investigated 1-component monolayers and for each organic monolayer we only measured the reactive uptake at one surface pressure, which corresponds to one molecular surface area of the surfactant (i.e. packing density of the surfactant). Experiments as a function of molecular surface area with the same surfactant were not carried out. The current paper is an extension of our previous work, and consists of two series of measurements. In the first series of measurements we investigate the uptake of N205 on aqueous sulfuric acid solutions coated with a 1octadecanol monolayer (a 1-component monolayer) as a function of the molecular surface area of the surfactant. This allows us to isolate the effect of molecular surface area on the reactive uptake coefficient. In addition, the results from this study allow us to investigate whether or not the 2-D phase of the surfactant monolayer has a large impact on uptake. Monolayers can go through a series of 2-D phase transitions (see below), and in the 1-octadecanol experiments as a function of molecular surface area mentioned above the monolayer undergoes a phase transition. This allows us to observe how the 2-D phase alters the reactive uptake coefficient. Below, we present these results and also discuss the results in terms of the Accessible Area Theory, 279 which has been used in the past to predict the effect of organic monolayers on water evaporation. 273,280,281 In the second series of measurements we focus on a 2-component monolayer consisting of 1-octadecanol (a straight-chain surfactant) and phytanic acid (a branched  Chapter 4^  98  surfactant). This mixed monolayer is of atmospheric relevance since monolayers in the atmosphere may consist of more than one type of surfactant. It is also reasonable to expect that some atmospheric monolayers will consist of mixtures of straight-chain and branched surfactants. Prior to our uptake measurements, we first determine if 1-octadecanol-phytanic acid monolayers are miscible or immiscible, which is necessary for understanding our uptake data. 282 In order to determine the miscibility, we measure the surface pressurearea isotherms of the 2-component monolayer. The analysis of this data allows us to show that 1-octadecanol-phytanic acid monolayers are immiscible. After proving this 2component monolayer is immiscible, we study the uptake coefficient of N 2 05 on aqueous H2SO4 solutions in the presence of these monolayers as a function of the composition of the monolayer (i.e. mole fraction of phytanic acid in the 2-component monolayer). The results from these studies are presented below and compared to predictions based on two different models used to describe mass transfer across the air-aqueous interface in the presence of a 2-component monolayer. The atmospheric implications of these results are also discussed.  4.2 Experimental 4.2.1 Rectangular Channel Flow Reactor and Determination of Reactive Uptake Coefficients  A rectangular channel flow reactor 271 '276 coupled to a chemical ionization mass spectrometer276 is used to measure heterogeneous reaction rates for the uptake of N205 on aqueous solutions coated with organic monolayers. The strength of the rectangular channel flow reactor configuration is that we can determine the surface pressure and molecular surface area (i.e. packing density) of the organic monolayer before and after each kinetic measurement. The experimental apparatus has been described in detail in the previous two chapters (sections 2.2 and 3.2), and therefore will not be described again. N20 5 was produced by reacting NO2 with an excess of 03. 03 was generated by passing a flow of 02 over an ultraviolet source (Jelight, model #600). The details of this procedure have been outlined in depth in the previous chapter (section 3.2).  Chapter 4^  99  A saturated flow of N205 between 7 to 8 cm 3 min -1 at STP was mixed with 70-80 cm 3 min -1 at STP of dry He prior to entering the flow reactor through the T-shaped injector. Total mass flow rates inside the flow reactor ranged from 260 to 360 cm 3 min -1 at STP and total pressures were between 2.6 to 3 Ton. Laminar flow conditions were achieved under these conditions as Reynold's numbers ranged from 0.4 to 0.8. A chemical ionization mass spectrometer (CIMS) was connected to the exit of the flow cell, which was used to measure the change in the gas phase reactant concentration as a result of reactive uptake at the liquid surface. 271,276,283,284 N2 0 5 was detected as NO3 after its chemical ionization by F.240,246,285 Trace amounts of CH3I diluted in 1250 to 1500 cm 3 min -1 at STP of N2 were passed through a polonium-210 source (NRD, model Po2031) in order to form I - . For N205 detection in the presence of H2 O the chemical ionization region was biased to -122 V in order to decluster weakly bound ion-H20 clusters. 240,246 N2 0 5 concentrations ranged from 2.1 x 10 10 to 1.1 x 10 11 molec cm -3 in these experiments. N205 concentrations were based on the F + N205 chemical ionization reaction rate that has been reported in the literature. 285 The reactive uptake coefficients were determined from the irreversible removal of N205 as a function of injector position in the rectangular channel flow reactor. Shown in Figure 4.1 are plots of the natural logarithm of the N205 signal as a function of reaction time for the loss of N205 on coated and uncoated aqueous 60 wt % sulfuric acid solutions at 273 ± 1 K. The data for each uptake experiment is represented with a linear fit, which yields the observed first-order rate constant (k o b s ) from the slope of each data set. The first order wall loss rate (kw ) was calculated from ko b s using a procedure developed by Knopf et a1. 271 This procedure corrects for any concentration gradients that can develop in the flow reactor as a result of a fast heterogeneous reaction at the surface.  100  Chapter 4^  0.01  0.02  0.03  0.04  0.01  0.02^0.03  0.04  fime/s Figure 4.1: Natural logarithm of the observed N2 0 5 signal as a function of reaction time. Symbols show representative data for a series of 1-octadecanol monolayers with varying molecular surface area. Solid squares: a molecular 2 • 1 , open triangles: surface area of 20.7 A 2 molec -1 , open squares: 21.2 A moiec 22.6 A 2 molec -1 , and solid triangles: 24.2 A 2 molec -1 . The dashed lines represent the corresponding linear fits to the data.  The determination of k w from ko b s requires diffusion coefficients for N20 5 in He ( D N205—He) and N205 in H2O (DN205-H20 )• The values used for DN 2 05 -He and D1,1205-H20 at 273 K were 289 Torr cm 2 s -1 and 72 Torr cm2 s -1 , respectively. These values were taken from Knopf et al., 271 and were calculated based on molecular parameters. 286-288 The diffusion coefficient (D) for N20 5 in a binary mixture of helium and water is then given by:289  101  Chapter 4^ ^ 4.1  1^PHe^PH20 D D N205—He D N 0 -11,0 2  5  where PHe and PH20 are the partial pressures of helium and water vapor in the flow cell, respectively. The reactive uptake coefficient () was determined from k v, using: 290 291 1^c S 1 + 4k,, V 2  4.2  where c is the mean molecular velocity of N205, S is the reactive surface area inside the flow cell, and V is the volume above this area. 4.2.2 Preparation of 1-Component Monolayers for the Reactive Uptake Measurements The first set of experiments involved N205 uptake experiments on 1-component monolayers (specifically 1-octadecanol) as a function of molecular surface area of the surfactant (i.e. packing density of the surfactant). We prepared monolayers with various molecular surface areas by varying the amount of organic surfactant added to the surface. Since the total area of the surface was known accurately and the total amount of organic surfactant added to the surface was known accurately, the molecular surface area of the surfactant molecules could be accurately calculated. The molecular surface area of the surfactant was also verified by measuring the surface tension of the monolayers before and after the reactive uptake measurements (see below). More details on preparation of the monolayers are as follows: first, the surface of the aqueous acid was thoroughly cleaned using a PTFE nozzle aspirator to remove any organic contamination on the surface prior to preparing an organic monolayer. Second, solutions of 1-octadecanol were prepared by dissolving the surfactant in chloroform ( 1 mg cm -3 ) to make organic -  solutions. Finally, monolayers were prepared by depositing a known volume of the organic solution on a clean aqueous sulfuric acid surface. The chloroform was allowed time to evaporate, leaving behind an organic monolayer, with a known molecular surface area of the surfactant.  Chapter 4^  102  To verify the molecular surface area of the surfactants, we measured the surface tension of the prepared monolayers using the Wilhelmy plate method prior to and after the uptake measurements. From the surface tension we calculated the surface pressure, and then from the surface pressure we determined the molecular surface area of the surfactant molecules from the pressure-area isotherm of octadecanol on aqueous sulfuric acid solutions. The pressure-area isotherm for 1-octadecanol was determined in previous experiments 276 using a commercial Langmuir film balance, and is shown in Figure 4.2a (see below). The isotherm shows the relationship between the surface pressure of the monolayer and the molecular surface area of the surfactant, so once the surface pressure is known, the molecular surface area can be determined. The molecular surface areas determined from the surface tension measurements were always within a few percent of the molecular surface areas calculated from knowledge of the mass of organic surfactant added to the surface. More details on the surface tension measurements are as follows: 276 surface tension measurements were carried out with a platinum Wilhelmy plate (23.32 mm perimeter) connected to a tensiometer (NIMA Technology, model PS4). The platinum plate was cleaned using a 1:1 solution of H2SO4 and HNO3 prior to each measurement, followed by thorough rinsing with purified water. The Wilhelmy plate was hung from the tensiometer and was immersed in the aqueous acid solution contained in the temperature controlled flow cell. The force on the Wilhelmy plate was measured while the plate was immersed and detached from the liquid solution contained in the quartz trough. The surface tension was calculated from the maximum difference in the force on the plate between immersion and separation from the solution. 276,292,293 The surface tension in the presence of the monolayer was measured before and after each kinetic experiment and used to calculate the surface pressure. 4.2.3 Measurements of Surface Pressure-Area (71-A) Isotherms of the 2-  Component Monolayers  The second set of experiments involved N205 uptake experiments on a 2component monolayer (consisting of 1-octadecanol and phytanic acid) as a function of mole fraction of phytanic acid in the monolayer  (xphytanic).  First, in order to characterize  Chapter 4^  103  these monolayers, which was necessary for interpreting the uptake results, we measured surface pressure-area isotherms 282 as a function of mole fraction of phytanic acid in the monolayer, using a commercial Langmuir film balance (NIMA Technology, model #611). The main point of these experiments was to determine if the 2-component monolayer was immiscible or miscible. The Langmuir film balance consisted of a PTFE trough (with dimension of 20 cm by 30 cm), two movable barriers, and a surface pressure sensor (NIMA Technology, model PS4) with a platinum plate. The experimental procedure that was followed is described in detail by Myrick and Franses. 294 Briefly, the trough was filled with an aqueous sulfuric acid solution, and the surface of the acid solution was cleaned thoroughly using an aspirator. A known volume of an organic solution (containing 1octadecanol and phytanic acid dissolved in chloroform) was added to the clean H2SO4H2O surface. The chloroform was allowed to evaporate leaving behind a known number of surfactant molecules on the surface. The surface pressure was then recorded as the moveable barriers reduced the available surface area, resulting in a surface pressure-area isotherm. 282 Surface pressure-area isotherms for 2-component monolayers were measured at a constant compression rate of 20 cm2 min -1 . 4.2.4 Preparation of the 2-Component Monolayers for the Reactive Uptake Measurements  After characterizing the 1-octadecanol-phytanic acid monolayers, we then investigated the reactive uptake of N205 on aqueous 60 wt % H2SO4 solutions in the presence of these monolayers. Measurements were carried out as a function of the mole fraction of phytanic acid in the monolayer. For these experiments, the surface pressure of the monolayers was kept at 21 ± 2 mN m -1 , which is less than the first collapse pressure of the 2-component monolayer (see below for more details). Monolayers with a surface pressure of 21 ± 2 mN m -1 were prepared by dissolving a known amount of 1-octadecanol and phytanic acid (with a fixed ratio) in chloroform. Drops of this solution were than added to a clean aqueous sulfuric acid surface until the desired surface pressure was reached, which was verified by measuring the surface tension of the solution. The surface tension was measured with the procedure discussed above.  Chapter 4^  104  4.2.5 Chemicals Listed below are the chemicals, the manufacturer, and the corresponding purities of the chemicals used in our studies: He (Praxair, 99.999% Purity), N2 (Praxair, 99.999%), NO 2 (Matheson, 99.5%), 02 (Praxair, 99.5%), P20 5 (Aldrich, 97%), 1-octadecanol (Sigma-Aldrich, 99%), phytanic acid (Sigma-Aldrich, 96%), chloroform (Fisher, 99+%), sulfuric acid (Fisher, 95-98%).  4.3 Results and Discussion 4.3.1 N20 5 Reactive Uptake in the Presence of a 1-Component Monolayer (1Octadecanol) The reactive uptake of N20 5 on aqueous H2SO4 was measured in the presence of a monolayer of 1-octadecanol at varying molecular surface areas of the surfactant. The surface pressure-area isotherm for 1-octadecanol is shown in Figure 4.2a.  276  The solid  line in Figure 4.2a illustrates the isotherm for 1-octadecanol on 60 wt % H2SO4 at 273 K. 276 The molecular surface areas at which reactive uptake coefficients were determined  in this study are shown in Figure 4.2a (open symbols). The solid symbol represents the conditions used in our previous study for 1-octadecanol monolayers. 276 Figure 4.2a illustrates that as the molecular surface area decreases, the monolayer undergoes a series of phase transitions shown by a change in the slope of the isotherm. 282 ' 295 These phase transitions correspond to different degrees of ordering of the organic surfactants on the surface. 295 ' 296 At large molecular surface areas (greater than approximately 40 A 2 molec -1 ) the monolayers exist as a 2-D gas on the aqueous acid surface, with molecules on the surface exerting relatively little force on each other due to sufficient separation.  282  For  decreasing molecular surface areas the monolayers undergo several phase transitions until they reach their collapse pressure, ge (which occurs at approximately 19 A 2 molec -1 for 1octadecanol). 282 In Figure 4.2 the different phase regions (defined by discontinuities in the slope of the isotherm) are labeled as S', S, S*, and L2 * , based on previous studies of 1-octadecanol monolayers on water. 296 The different phases correspond to different orientations of the organic chains and also different unit cells. Traditionally, S' and S would be classified as a condensed solid phase and S* and L2 * would be classified as a  Chapter 4^  105  liquid condensed phase. 282 In our experiments we studied N205 uptake in the presence of S* and L2 * phases. The main difference between these two phases is a change in the tilt angle of the hydrocarbon chains in the monolayer. 296  18^22^24^26  ^  28  Area / A 2 molec -1  Figure 4.2: Surface pressure — area isotherm for 1-octadecano1 276 (panel (a)) and reactive uptake coefficients (y) for N205 on organic coated aqueous 60 wt % sulfuric acid at 273 ± 1 K plotted as a function of molecular surface area of 1octadecanol (panel (b)). The solid line in panel (a) represents the variation of surface pressure (n) with a change in molecular surface area. S', S, S*, and L2 * represent the different 2-D phases of the monolayer (see text for further details). 296 Also shown is the collapse pressure (xi) of the 1-octadecanol monolayer. The open symbols in panel (a) represent the experimental conditions (surface pressures and molecular surface areas) at which reactive uptake experiments were performed for N205 on aqueous 60 wt % H2SO4 at 273 ± 1 K coated with 1-octadecanol. The solid square in panel (a) represents experimental conditions used previously by Cosman et al. 276 for 1-octadecanol monolayers.  Chapter 4^  106  Figure 4.2b shows the reactive uptake coefficients (y) in the presence of 1octadecanol monolayers determined in our studies using different molecular surface areas. The error bars for y take into account reproducibility in the kinetic experiments and a 20% uncertainty in the diffusion coefficients. Figure 4.2 shows that the reactive uptake coefficient depends strongly on the molecular surface area of the organic monolayer, with the reactive uptake coefficient decreasing as the molecular surface area was decreased. This observation is consistent with previous studies of the rate of water evaporation through organic monolayers. 281,282,297 In these previous studies, it was generally observed that the resistance to evaporation increased as the molecular surface area of the surfactant decreased. Figure 4.2 also suggests that the reactive uptake coefficient does not change drastically when the phase of the monolayer converts from S* to L2 * . The y values shown in Figure 4.2b increase smoothly with increasing molecular surface area of the surfactant, with no apparent step function change observed upon transition between monolayer phases S* and L2 * . This is also generally consistent with studies of the rate of water evaporation through organic monolayers. For example, La Mer et al. 297 showed that the resistance to water evaporation increased smoothly (i.e. no apparent step function change in the evaporation resistance was observed) as monolayers underwent phase transitions from condensed liquid to solid phases. Our results are also not surprising as the surface pressure changes relatively smoothly between S* to L 2 *, and the structural differences between S* to L2 * are thought to be relatively minor Due to experimental constraints, we were not able to compress the monolayer to molecular surface areas beyond that at equilibrium. Hence we were not able to monitor the change in reactive uptake coefficient when going from the phase S* to S, where the surface pressure changes most steeply. In Figure 4.2 we have plotted y versus the molecular surface area of the organic monolayer. Another way to present the reactivity data is versus the fractional surface coverage. Here we define the fractional surface coverage as the molecular surface area of the surfactant normalized to the molecular surface area of the surfactant at the collapse pressure, 7rc . Shown in Figure 4.3 is the reactive uptake coefficient in the presence of the  107  Chapter 4^  0.80  0.75  0.85  0.90  0.95  1.00  (a) 0.12  0.12  0.08  0.08  co  o  0 0  T  47-  0.04  0.04  0  ^  I  0.00  0.00 I  I^I  ^  I^I  (b)  0.1  0.1  o a)  -  0  = E  0.01  0.01  0.75^0.80^0.85^0.90^0.95  1.00  fractional surface coverage  Figure 4.3: Reactive uptake coefficients for N20 5 on organic coated aqueous 60 wt % sulfuric acid at 273 ± 1 K plotted as a function of the fractional monolayer surface coverage. In this figure the reactive uptake coefficient in the presence of the organic monolayer (7film ) is normalized to reactive uptake coefficient for the uncoated solution (vuncoated) to illustrate the change in reactive uptake coefficient due to the presence of the monolayer. Open squares: 1-octadecanol (this study), solid square: 1-octandecanol (Cosman et al.). 27 Panel (a) gives the ordinate in a linear scale whereas panel (b) shows the ordinate in a log scale.  Chapter 4^  108  organic monolayer (nu m ) normalized by the uptake coefficient for the uncoated case (/uncoated) as  a function of this fractional surface coverage.  Figure 4.3 shows that the fractional surface coverage does not have to be 1 in order to significantly decrease the reactive uptake coefficient. Even at 0.75 of the maximum surface coverage (i.e. fractional surface coverage = 0.75), the monolayer still decreases the reactive uptake coefficient by a factor of 10. Park et al. 27° and McNeill et ai.,262,263 also performed a similar analysis to the above (although they defined the fractional surface coverage slightly differently). Our conclusion mentioned above is similar to the conclusions reached by these authors previously. For example, McNeill et al. 262  observed a significant decrease in y (approximately a factor of 3 decrease) even  when the fractional surface coverage was 0.08 for a NaCI aerosol coated with sodium dodecyl sulfate monolayers. In Figure 4.4, we compare our 1-octadecanol results with previous measurements of N205 reactive uptake measured on aqueous sulfuric acid solutions coated with 1component monolayers. For this comparison we use the molecular surface area on the xaxis, since in our previous work we have shown that the uptake results for coated aqueous sulfuric acid solutions correlate better with the molecular surface area of the surfactant compared to other parameters such as surface pressure. 276 Panel (a) displays  /film  and  panel (b) displays the ratio of 2/fil m to the reactive uptake coefficient of the uncoated solution (juncoated). Our current data for 1-octadecanol is represented with open symbols and all the other data (obtained with soluble and insoluble monolayers) is represented by solid symbols. First, our current data generally fits well the trend observed with all the previous data. 270,271,276 Second, considering all the data, there appears to be a strong correlation between the reactive uptake coefficient and the molecular surface area. This trend was observed in our previous publication, 276 and our current data adds more support to this conclusion.  Chapter 4^  109  20 30 40 50 60 70 80 90 100 I^'^I  I^I  (a) 0.1  0.1  0.01  0.01  1 E-3  1 E-3  E  (b) ID  1  1  a) ca  C  z E  0.1  0.01  0.1  I^.^I  I^I  0.01  20 30 40 50 60 70 80 90 100  Area / A2 molec -1  Figure 4.4: Reactive uptake coefficients for N205 on aqueous sulfuric acid in the presence of organic monolayers as a function of molecular surface area. Open squares: 1-octadecanol (this study), solid right facing triangles: phytanic acid (this study), solid squares: 1-octadecanol (Cosman et al.), 76 solid triangles: stearic acid (Cosman et al.), 276 solid circles: phytanic acid (Cosman et al.), 276 solid left facing triangles: 1-hexadecanol (Cosman et al.), 276 solid diamonds: butanol (Park et al.), 27° solid stars: hexanol (Park et al.), 27° solid inverted i triangles: 1octadecanol (Knopf et al.). 271 In Panel (a) the ordinate is the reactive uptake coefficient in the presence of the organic monolayer (yrd.), and in Panel (b) the ordinate is yrd m normalized to reactive uptake coefficient for the uncoated solution (Vuncoated)•  Chapter 4^  110  The fact that the data in Figure 4.4 correlates well with the surface area is broadly consistent with the Accessible Area Theory. 279 The Accessible Area Theory has previously been used to predict the effect of organic monolayers on the evaporation rate of water. This theory captures well the general trends in evaporation of water in the presence of several long-chain alcohols. However, the agreement between theory and experiments is not quantitative. 273 ' 28 ' The Accessible Area Theory suggests that transport occurs only through open sections of the surface. Such sections may be formed through random fluctuations or by incomplete packing of the monolayer. One limitation of this theory is that it does not consider the effect of chain length on the mass transport of gases. 279 Note that Figure 4.4 should not be considered as a rigorous test of the Accessible Area Theory since the figure largely separates the short chain molecules (right hand side in Figure 4.4) from the long chain molecules (left hand side in Figure 4.4, with the exception of phytanic acid). More work is needed to fully test the Accessible Area Theory. Studies of the reactive uptake of N 2 0 5 in the presence of long-chain molecules with large molecular surface areas would be useful. From Figure 4.4, we conclude that the trend in reactive uptake coefficient of N205 with molecular surface area is broadly consistent with the Accessible Area Theory (i.e. a correlation between the reactive uptake coefficient and the molecular surface area is observed). More work is needed, however, to determine if the theory is quantitative. Also, note that in Figure 4.4 we have only plotted data measured with aqueous sulfuric acid solutions. As discussed in the previous chapter, data measured with a sea salt or NaCI subphase are not consistent with the trend shown in Figure 4.4. More work is needed to understand this apparent conflict. 4.3.2 7r-A Isotherms and Miscibility of 2-Component Monolayers  A 2-component monolayer may be miscible or immiscible. An immiscible monolayer can be thought of as made up of two separate monolayers, whereby one component forms patches of a monolayer distributed within a monolayer of the second component. 282 The main point of this section is to determine if the 2-component system we studied (1-octadecanol-phytanic acid monolayer) is immiscible.  Chapter 4^  111  Shown in Figure 4.5 are the pressure-area isotherms for the 1-octadecanolphytanic acid monolayers (with various  Xphytanic  values) as well as the isotherms for the  pure components (i.e. pure 1-octadecanol and phytanic acid). The z-A isotherm for 1octadecanol (see Figure 4.5a) shows several "kinks" in the isotherm due to 2-D phase transitions, as discussed above. 282,295,296 As the molecular surface area decreases the monolayer undergoes several phase transitions until it reaches its collapse pressure (labeled in Figure 4.5 as 2rc,octadecanoi), 276 and a further compression of the monolayer results in the formation of a new bulk phase on the surface. 282  70 60 50 40 30 20 10 0 20^40^60^80  Area / A 2 =lee  0^20^40^60  80  Area / A 2 molee  Figure 4.5: Surface pressure — area isotherms for organic monolayers on aqueous 60 wt % sulfuric acid at 273 ± 1 K. Panel (a): the isotherm for 1-octadecano1 276 (xphytanic = 0) represented by the solid line and phytanic acid 276 (xphytanic 1) shown as the dashed line. Irc, octadecanol and 7rc, phytanic represent the surface pressure at which monolayers of pure 1-octadecanol and pure phytanic acid collapse, respectively. Panel (b): the isotherm for 2-component monolayers of 1octadecanol and phytanic acid. The solid line, dashed line, and bold-solid line represent the isotherms for compositions of Xphytanic 0.05, 0.2, and 0.7, respectively. Isotherms for Xphytanic 0.1 and 0.4 have been omitted for clarity.  The behavior for the branched surfactant, phytanic acid, is significantly different than for the straight-chain surfactant, 1-octadecanol (see Figure 4.5a). Between 45 A2  112  Chapter 4^  molec -1 and approximately 80 A2 molec -1 , the branched-chain monolayer is in a liquid expanded state. 276,282 At a surface pressure of 26 mN m - 1 the monolayer collapses (labeled 7rc,phytanic in Figure 4.5a), and further compression of the monolayer results in the formation of liquid lenses in equilibrium with a monolayer at a molecular surface area of 44.5 A 2 molec -1 . 276 Panel (b) of Figure 4.5 illustrates the 7r-A isotherms for the 2-component monolayers. The isotherms for the mixtures appear to be intermediate between the pure component isotherms and also exhibit two discontinuities which we assign as collapse pressures of the monolayers, labeled c,mixture and R 2 c,mixture• The ,r-A isotherms for Xphytanic =  0.1 and 0.4 have been omitted for clarity.  One useful tool for establishing the miscibility of a 2-component monolayer is the interfacial phase rule. 282,298,299 According to this phase rule, when a monolayer is immiscible it will exhibit two separate collapse events that correspond to the collapse pressures of the individual pure components, independent of monolayer composition.  282  If, however, the two components of the monolayer at the interface are miscible, the surface pressure at which the monolayer collapses is dependent on the monolayer composition. 282 Figure 4.6 shows the collapse pressure for the 2-component monolayers of 1-octadecanol and phytanic acid as a function of monolayer composition (i.e.  Xphytanic),  as well as the collapse pressures for the single component monolayers. The solid symbols represent the collapse pressures for the 2-component monolayer and the open symbols represent the collapse pressures for the 1-component systems.  Chapter 4^  90  0.0  0.2  113  0.4  0.8  0.6  1.0  80  80  ■  70  2 7C^.  c, mixture  ■ ■^■  z 60 E  70 60  ■  50  40  90  50  - c, mixture  30  40  A • A  •  30  •  20  20 0.0  0.2  0.4 X  0.6  0.8  1.0  phytanic  Figure 4.6: The collapse pressures of 1-octadecanol-phytanic acid monolayers on aqueous 60 wt % sulfuric acid at 273 f 1 K as a function of composition (Xphytanid• Solid triangles and solid squares represent the 1 st collapse pressure (r i c , mixture) and 2" collapse pressure (117 c, mixture) for the 2-component monolayers, respectively. The open triangle and open square represent the collapse pressure of pure phytanic acid and pure 1-octadecanol monolayers, respectively.  Figure 4.6 shows that the 1-octadecanol-phytanic acid monolayers exhibit two collapse pressures that are consistent with the collapse pressures of the pure component systems and are roughly independent of composition, as one would expect for an immiscible monolayer. The scatter in A—2 c,mixture as a function of 712 e,mixture corresponds  xphytanic  is not surprising.  to the collapse pressure of 1-octadecanol in the 2-component  monolayer. In this case the 1-octadecanol is in a 2-D solid phase just before collapse, and it is well known that the collapse pressure of a 2-D solid phase can vary significantly  Chapter 4^  114  from experiment to experiment and is sensitive to trace levels of impurities in the monolayer, since the collapse of the of a 2-D solid phase is a nucleation event with a large kinetic barrier. 282 Also the scatter observed in Figure 4.6 is typical for immiscible monolayers. 30° Another useful tool for investigating the miscibility of a monolayer is the average molecular surface area in the 2-component monolayer and the excess area in the 2component monolayer. If the 2-component monolayer is completely immiscible, then the average molecular surface area of the surfactant molecules in this monolayer should be related to the molecular surface area in the 1-component monolayers at the same surface pressure, according to the following equation: 282 4.3  A 12 ,predided = A l x 1 + A 2 x 2  where A 12,predicted is the average molecular surface area predicted for an immiscible 2component monolayer. A1 and  A2  are the molecular surface areas in the 1-component  monolayers 1 and 2, respectively, and x 1 and x2 are their respective mole fractions. For eq 4.3 to apply, Al2,predicted, A1, and A2, need to be evaluated at the same surface pressure. In Figure 4.7a we plot A 12,predicted as well as the measured average molecular surface area in the 2-component monolayer, A 12,meas urea, evaluated at a surface pressure of 21 mN m-1.  Chapter 4^  115  0.2  0.0  0.4  0.6  0.8  52  c.) a) 0 E (No<  (a)  48  cu cu  a) cs) Cu L.  48 44  40  40  36  36  32  32  28  28 24  U a  20 1  °  52  44  24  Q  1.0  20 1  4  (b) - 4  2  2  0  0 -  -2  -2  -4  -4 0.0  02  0.4  0.6  0.8  1.0  xphytanic  Figure 4.7: Panel (a) shows the measured molecular surface area (solid symbols) and the predicted molecular surface area (dashed line) for 2-component monolayers with r = 21 ± 2 mN ni l as a function of composition (xphytanid• Panel (b) displays the excess area (A,) as a function of composition for 1-octadecanol phytanic acid monolayers. See text for further details.  A 12,measured was determined from the surface pressure-area isotherms shown in Figure 4.5b. In Figure 4.7a, symbols represent A 12,measured and the line represents A 12,predicted• The good agreement between the two gives more support for an immiscible monolayer. To further explore this point, we also calculated the excess area, A ex , as a function of monolayer  Chapter 4^  116  composition (evaluated at 21 mN m -I ) for the 2-component monolayer. The excess area is a measure of non-ideality of the monolayer, and is given by: 300,301 4.4  Aex =  A l2,measured A l2,predicted  where A izmeasured and A 12,predicted are as defined above. 2-component monolayers that are completely immiscible have A,=0. 30° Positive or negative values for the excess area indicate non-ideal mixing in the 2-component systems. 30I Figure 4.7b shows the excess areas as calculated using eq 4.4. The values of A, do not appear to vary in any systematic way from zero. In fact the error bars for all data points overlap zero except for one. A final way we explored the miscibility of the monolayer was by comparing our measured pressure-area isotherms with predictions. If the monolayers are completely immiscible, then the isotherms of the 2-component monolayer should be a linear combination of the 1-component monolayers, scaled to the mole fractions of the single components. In Figure 4.8 we have shown two of our measured isotherms, along with predictions (shown as dashed lines) assuming the monolayers were immiscible. The good agreement gives further support that our monolayers are immiscible. The above evidence gives strong support that the 1-octadecanol-phytanic acid monolayers are completely immiscible and hence one can think of the monolayers as forming patches of phytanic acid distributed within a monolayer of 1-octadecanol (or vice versa). At the first collapse pressure (gic,mixture) phytanic acid is squeezed out resulting in a monolayer that contains only 1-octadecanol until the second collapse pressure (712c,mixture),  at which the 1-octadecanol collapses as well. Since at the first collapse  pressure all the phytanic acid is squeezed out of the monolayer, we performed all our uptake measurements using a surface pressure of 21 ± 2 mN m -1 . This ensures that we still had two components in the monolayer.  Chapter 4^  0  20  117  40  60  80  70  70  60  60  50  50  z 40  40  30  30  20  20  10  10  E  0  0 0  ^  20^40^60  ^  80  Area / A2 rnolec -1 Figure 4.8: Surface pressure — area isotherms for binary mixtures of 1octadecanol and phytanic acid performed on aqueous 60 wt % sulfuric acid at 273 ± 1 K. Examples of experimentally measured isotherms for 2-component monolayers containing xphytanic 0.1 (bold solid line) and xphytanic = 0.7 (solid line) are shown. The dashed lines are the predicted isotherms based on a linear combination of the isotherms for pure 1-octadecano1 276 and pure phytanic acid. 276  4.3.3 N205 Uptake in the Presence of a 2-Component Monolayer The reactive uptake of N205 on aqueous 60 wt % H2SO4 solutions at 273 K was measured in the presence of 1-octadecanol-phytanic acid monolayers. 2-component monolayers with a mole fraction of phytanic acid  (Xphytanic) equal  to 0, 0.05, 0.1, 0.2, 0.4,  0.7, and 1 were studied. Reactive uptake coefficients and average molecular surface areas for each binary mixture studied are reported in Table 4.1. The upper and lower limits for yin Table 4.1 take into account 20% error in the diffusion coefficients.  118  Chapter 4^  Table 4.1: Measured Reactive Uptake Coefficients for N205 on 60 wt % Aqueous Sulfuric Acid Solutions at 273 ± 1 K Coated with 1-Octadecanol-Phytanic Acid Monolayers at 21 ± 2 mN m -1 Xphytanic  7F  (mN m 1 )  average  lower  upper  surface area  limit  limit  (A 2 /molec ) 0  21.1  21.3  0.00116  0.00091  0.00141  0.05  19.7  22.1  0.00579  0.00405  0.00752  0.1  21.7  23.5  0.0109  0.00761  0.0141  0.2  20.1  26.3  0.0217  0.0174  0.0261  0.4  21.4  32.2  0.0352  0.0246  0.0457  0.7  21.9  38.1  0.0625  0.0410  0.0841  1  21.5  47.2  0.0663  0.0439  0.0894  The reactive uptake coefficients determined in these studies are also shown in Figure 4.9. This figure shows that y increases as  Xphytame increases.  This was expected as  the uptake coefficient on a H2SO4-H20 solution coated with a phytanic acid monolayer (at a surface pressure of 21 ± 2 mN m -1 ) is much larger than for the same solution coated with a 1-octadecanol monolayer (at a surface pressure of 21 ± 2 mN m -1 ). As discussed in the previous chapter, the reason phytanic acid has a smaller effect on the reactive uptake coefficient is likely due to the branched structure of the molecule, which prevents the surfactant from packing densely on the surface (or achieving a small molecular surface area).  Chapter 4^  0.0  119  0.2  0.4  0.6  0.8  1.0 0.1  0.1  0.01  1 E-3  1 E-3  0.1  0.1  0.01  0.0  0.2  0.4 X  0.6  0.8  0.01 1.0  phytanic  Figure 4.9: The reactive uptake coefficient for N20 5 on aqueous sulfuric acid in the presence of 1-octadecanol - phytanic acid monolayers (‘ rmixed film) as a function of mole fraction of phytanic acid in the monolayer, xphytanic• Solid squares represent the average y value for each corresponding monolayer. The error bars represent 20 . The bold dashed line represents the prediction based on eq 4.5 whereas the bold dash dot line represents the prediction based on eq 4.6. 282 The shaded regions reflect the uncertainty in the predictions based on the uncertainty in Ti (xph y tanic = 0) and y (Xphytanic 1). In Panel (a) the ordinate is the reactive uptake coefficient in the presence of the organic monolayer (jmixed film), and in Panel (b) the ordinate is ymixe d rim normalized to reactive uptake coefficient for the uncoated solution (vuncoated)• .  Chapter 4^  120  Two different models have been suggested to explain the mass transfer of species across a 2-component organic monolayer. Gaines 282 suggests that for a 2-component immiscible monolayer the resistance to mass transfer of patches of unmixed monolayers might be expected to combine as resistances in parallel. This model can be expressed with the following equation: 282 4.5  712 = X1 . 71 + X2 . 72  where xi and x2 are the mole fractions of component one and component two in the 2component monolayer and 712, 71, and 72 represent the reactive uptake coefficients for the 2-component and 1-component monolayers, respectively (all evaluated at the same surface pressure). In contrast, Barnes and LaMer 302 used the Energy Barrier Theory 3°2-3°4 to predict the resistance to mass transfer through an ideal, miscible monolayer. In terms of reactive uptake coefficients, this model can be expressed with the following equation: 3°2 1 ^r 1 j ln^= x1 .1n — + x 2 .In 1 ^712^\ 71^72 ^t 4.6  The dashed lines in Figure 4.9 are the result of plotting eq 4.5 whereas the dash dot lines in Figure 4.9 are the results of plotting eq 4.6 for our experimental conditions. The shaded regions in Figure 4.9 reflect the uncertainty in the model predictions based on the uncertainty in 71 and n Our data is consistent with the model for reactive uptake in .  the presence of an immiscible monolayer presented by Gaines 282 (eq 4.5). However, our data does not agree well with the model prediction for a miscible monolayer (eq 4.6). Our data is the first case where the results for an immiscible monolayer are consistent with eq 4.5. More on this is included below. Related to this discussion, several researchers have investigated the evaporation of water coated with 2-component organic monolayers. 280,297,304-310 Researchers have generally found that the resistance to water evaporation due to an ideal, miscible monolayer can be explained with eq 4.6. 297 '31° For immiscible monolayers, however, none of the water evaporation studies could be explained by eq 4.5 (or eq 4.6).273,281,305,307,309  For example, several studies have investigated the evaporation  resistance of 1-octadecanol-cholesterol monolayers (which are expected to be  Chapter 4^  121  immiscible) and have shown resistances that are appreciably lower than those predicted by eq 4.5. 3°5 '3°9 For example, deviations from model predictions by up to a factor of 2.2 were observed by Barnes et a1. 305 Possible explanations for why evaporation resistance measurements through immiscible monolayers are not consistent with eq 4.5 include partial miscibility or mass transfer occurring mainly at grain boundaries. 273 Note that the reactive uptake coefficient of N20 5 on aqueous sulfuric acid solutions coated with 1octadecanol-phytanic acid monolayers may also deviate slightly from the predictions based on eq 4.5, but our work shows that this deviation is less than the uncertainty in our measurements as illustrated by Figure 4.9. In addition to water evaporation studies, Gilman and Vaida267 studied the uptake of acetic acid in the presence of mixed monolayers consisting of 1-triacontanol and cis-9octadecen-1 -ol, which are expected to form immiscible monolayers. These authors studied equimolar mixtures and observed that the mixed monolayers had permabilities that were between that of the single-component monolayers that comprise the mixture, which is consistent with our findings. However, the applicability of eq 4.5 and eq 4.6 to their results was not studied. Since the transfer of molecules across 2-component monolayers is an important step toward a complete understanding of the mechanism of mass transfer across the airaqueous interface in the presence of monolayers, more work in this direction is required. Further experiments that would be beneficial could include studies of N20 5 uptake on the same 2-component immiscible monolayers studied in the water evaporation experiments where the results deviated significantly from eq 4.5.  4.4 Summary and Conclusions and Atmospheric Implications The first part of this chapter focused on the reactive uptake of N205 on aqueous sulfuric acid solutions coated with a 1-component monolayer (1-octadecanol). Our results showed that the reactive uptake coefficient depends strongly on the molecular surface area of the surfactant in the monolayer. We also observed no step function increase in y when transitioning between 2-D phases. We also demonstrated that when the fractional surface coverage was less than 1, the monolayer still showed significant resistance to mass transfer, consistent with previous studies of N205 reactivity on  Chapter 4^  122  aqueous particles in the presence of surfactants. 262,263,270 For example, even at 0.75 of the maximum surface coverage, the monolayer still decreased the reactive uptake coefficient by a factor of 10. This observation may be of atmospheric relevance, since monolayers in the atmosphere are not expected to always have a fractional surface coverage of 1. When we compared all previous measurements of reactive uptake of N20 5 on aqueous sulfuric acid solutions coated with a 1-component monolayer, we observed a strong correlation between the reactive uptake coefficient and the molecular surface area of the surfactant. This observation is broadly consistent with the Accessible Area Theory for permeation through monolayers. The second part of the chapter focused on 2-component monolayers consisting of a mixture of a straight-chain surfactant (1-octadecanol) and a branched surfactant (phytanic acid). These studies may be of more atmospheric relevance since monolayers in the atmosphere probably consist of more than one component and a combination of straight-chain and branched surfactants. The 2-component monolayers were first shown to be immiscible through a series of pressure-area isotherm measurements. Then we focused on measurements of the N205 reactive uptake. These measurements showed that when the monolayer contained 100 % straight-chain molecules, the decrease in the reactive uptake coefficient was approximately a factor of 42 due to the presence of the monolayer. However, our results showed that when the mole fraction of the branched surfactant was only 0.20 the decrease in the reactive uptake coefficient was only a factor of 2 (down from 42). Hence, a small amount of branched surfactant drastically changes the overall resistance to reactive uptake on aqueous sulfuric acid solutions. This highlights the importance of understanding the composition of mixed organic monolayers in acidic atmospheric particles. Also, our results showed that the overall resistance to reactive uptake of immiscible monolayers can be predicted reasonably accurately using eq 4.5, which assumes the resistances to mass transfer can be combined in parallel. This equation may be useful for making predictions of reactive uptake of aqueous particles coated with multi-component monolayers in the atmosphere.  Chapter 4^  123  4.5 References 226. Solomon, S.; Garcia, R. R.; Rowland, F. S.; Wuebbles, D. J. Nature 1986, 321,755758. 227. Dentener, F. J.; Crutzen, P. J. I Geophys. Res. Atmos. 1993, 98, 7149-7163. 228. Rudich, Y. Chem. Rev. 2003, 103, 5097-5124. 229. Gard, E. E.; Kleeman, M. J.; Gross, D. S.; Hughes, L. S.; Allen, J. 0.; Monica!, B. D.; Fergenson, D. P.; Dienes, T.; Galli, M. E.; Johnson, R. J.; Cass, G. R.; Prather, K. A. Science 1998, 279, 1184-1187. 230. Dentener, F. J.; Carmichael, G. R.; Zhang, Y.; Lelieveld, J.; Crutzen, P. J. I Geophys. Res.-Atmos. 1996, 101, 22869-22889. 231. Evans, M. J.; Jacob, D. J. Geophys. Res. Lett. 2005, 32, L09813, doi:09810.01029/02005GL022469. 232. Poschl, U. Angew. Chem. Int. Edit. 2005, 44, 7520-7540. 233. Ravishankara, A. R.; Rudich, Y.; Talukdar, R.; Barone, S. B. Philos. Trans. R. Soc. Lond. Ser. B-Biol. Sci. 1997, 352,171-181. 234. Song, C. H.; Carmichael, G. R. J. Atmos. Chem. 2001, 40, 1-22. 235. Zhang, Y.; Carmichael, G. R. I Appl. Meteorol. 1999, 38, 353-366. 236. Molina, M. J.; Tso, T. L.; Molina, L. T.; Wang, F. C. Y. Science 1987, 238, 12531257. 237. Tie, X.; Brasseur, G.; Emmons, L.; Horowitz, L.; Kinnison, D. I Geophys. Res. Atmos. 2001, 106, 22931-22964. 238. Hallquist, M.; Stewart, D. J.; Baker, J.; Cox, R. A. I Phys. Chem. A 2000, 104, 3984-3990.  Chapter 4^  124  239. Hanson, D. R.; Lovejoy, E. R. Geophys. Res. Lett. 1994, 21, 2401-2404. 240. Hu, J. H.; Abbatt, J. P. D. J Phys. Chem. A 1997, 101, 871-878. 241. Lovejoy, E. R.; Hanson, D. R. I Phys. Chem. 1995, 99, 2080-2087. 242. Robinson, G. N.; Worsnop, D. R.; Jayne, J. T.; Kolb, C. E.; Davidovits, P. J. Geophys. Res.-Atmos. 1997, 102, 3583-3601. 243. Hallquist, M.; Stewart, D. J.; Stephenson, S. K.; Cox, R. A. Phys. Chem. Chem. Phys. 2003, 5, 3453-3463. 244. Hanson, D. R. Geophys. Res. Lett. 1997, 24, 1087-1090. 245. Mentel, T. F.; Sohn, M.; Wahner, A. Phys. Chem. Chem. Phys. 1999, 1, 5451-5457. 246. Thornton, J. A.; Braban, C. F.; Abbatt, J. P. D. Phys. Chem. Chem. Phys. 2003, 5, 4593-4603. 247. Fried, A.; Henry, B. E.; Calvert, J. G.; Mozurkewich, M. I Geophys. Res 1994, 99, 3517-3532. 248. Kane, S. M.; Caloz, F.; Leu, M. T. J. Phys. Chem. A 2001, 105, 6465-6470. 249. Stewart, D. J.; Griffiths, P. T.; Cox, R. A. Atmos. Chem. Phys. 2004, 4, 1381-1388. 250. Kanakidou, M.; Seinfeld, J. H.; Pandis, S. N.; Barnes, I.; Dentener, F. J.; Facchini, M. C.; Van Dingenen, R.; Ervens, B.; Nenes, A.; Nielsen, C. J.; Swietlicki, E.; Putaud, J. P.; Balkanski, Y.; Fuzzi, S.; Horth, J.; Moortgat, G. K.; Winterhalter, R.; Myhre, C. E. L.; Tsigaridis, K.; Vignati, E.; Stephanou, E. G.; Wilson, J. Atmos. Chem. Phys. 2005, 5, 1053-1123. 251. Donaldson, D. J.; Vaida, V. Chem. Rev. 2006, 106, 1445-1461. 252. Gill, P. S.; Graedel, T. E.; Weschler, C. J. Rev. Geophys. 1983, 21, 903-920.  Chapter 4^  125  253. Ellison, G. B.; Tuck, A. F.; Vaida, V. J. Geophys. Res.-Atmos. 1999, 104, 1163311641. 254. Mochida, M.; Kitamori, Y.; Kawamura, K.; Nojiri, Y.; Suzuki, K. J. Geophys. Res. 2002, 107, 4325, doi:4310.1029/2001JD001278. 255. Barger, W. R.; Garrett, W. D. J. Geophys. Res.-0c. Atm. 1976, 81, 3151-3157. 256. Decesari, S.; Facchini, M. C.; Mircea, M.; Cavalli, F.; Fuzzi, S. J. Geophys. Res.Atmos. 2003, 108, 4685, doi:4610.1029/2003JD003566. 257. Tervahattu, H.; Hartonen, K.; Kerminen, V. M.; Kupiainen, K.; Aarnio, P.; Koskentalo, T.; Tuck, A. F.; Vaida, V. J. Geophys. Res.-Atmos. 2002, 107, 4053, doi:4010.1029/2000JD000282. 258. Tervahattu, H.; Juhanoja, J.; Kupiainen, K. I Geophys. Res. Atmos. 2002, 107, 4319, doi:4310.1029/2001JD001403. 259. Duce, R. A.; Mohnen, V. A.; Zimmerman, P. R.; Grosjean, D.; Cautreels, W.; Chatfield, R.; Jaenicke, R.; Ogren, J. A.; Pellizzari, E. D.; Wallace, G. T. Rev. Geophys. 1983, 21, 921-952. 260. Garrett, W. D. Pure Appl. Geophys. 1978, 116, 316-334. 261. Tervahattu, H.; Juhanoja, J.; Vaida, V.; Tuck, A. F.; Niemi, J. V.; Kupiainen, K.; Kulmala, M.; Vehkamaki, H. I Geophys. Res.-Atmos. 2005, 110, D06207, doi:06210.01029/02004JD005400. 262. McNeill, V. F.; Patterson, J.; Wolfe, G. M.; Thornton, J. A. Atmos. Chem. Phys. 2006, 6, 1635-1644. 263. McNeill, V. F.; Wolfe, G. M.; Thornton, J. A. J. Phys. Chem. A 2007, 111, 10731083.  Chapter 4^  126  264. Anttila, T.; Kiendler-Scharr, A.; Tillmann, R.; Mentel, T. F. I Phys. Chem. A 2006, 110, 10435-10443. 265. Daumer, B.; Niessner, R.; Klockow, D. I Aerosol Sci. 1992, 23, 315-325. 266. Folkers, M.; Mentel, T. F.; Wahner, A. Geophys. Res. Lett. 2003, 30, 1644, doi:1610.1029/2003GL017168. 267. Gilman, J. B.; Vaida, V. I Phys. Chem. A 2006, 110, 7581-7587. 268. Lawrence, J. R.; Glass, S. V.; Park, S. C.; Nathanson, G. M. I Phys. Chem. A 2005, 109, 7458-7465. 269. Thornton, J. A.; Abbatt, J. P. D. J. Phys. Chem. A 2005, 109, 10004-10012. 270. Park, S. C.; Burden, D. K.; Nathanson, G. M. I Phys. Chem. A 2007, 111, 29212929. 271. Knopf, D. A.; Cosman, L. M.; Mousavi, P.; Mokamati, S.; Bertram, A. K. J. Phys. Chem. A 2007, 111, 11021-11032. 272. Clifford, D.; Bartels-Rausch, T.; Donaldson, D. J. Phys. Chem. Chem. Phys. 2007, 9, 1362-1369. 273. Barnes, G. T. Colloid. Surface. A 1997, 126, 149-158. 274. Jefferson, A.; Eisele, F. L.; Ziemann, P. J.; Weber, R. J.; Marti, J. J.; McMurry, P. H. J. Geophys. Res.-Atmos. 1997, 102, 19021-19028. 275. Rubel, G. 0.; Gentry, J. W. J. Aerosol Sci. 1985, 16, 571-574. 276. Cosman, L. M.; Knopf, D. A.; Bertram, A. K. I Phys. Chem. A 2008, 112, 23862396.  Chapter 4^  127  277. Brown, S. S.; Ryerson, T. B.; Wollny, A. G.; Brock, C. A.; Peltier, R.; Sullivan, A. P.; Weber, R. J.; Dube, W. P.; Trainer, M.; Meagher, J. F.; Fehsenfeld, F. C.; Ravishankara, A. R. Science 2006, 311, 67-70. 278. Badger, C. L.; Griffiths, P. T.; George, I.; Abbatt, J. P. D.; Cox, R. A. J Phys. Chem. A 2006, 110, 6986-6994. 279. Barnes, G. T.; Quickenden, T. I.; Saylor, J. E. J. Colloid Interf. Sci. 1970, 33, 236243. 280. Costin, I. S.; Barnes, G. T. J. Colloid Interf. Sci. 1975, 51, 122-132. 281. Barnes, G. T. Adv. Colloid Interfac. 1986, 25, 89-200. 282. Gaines Jr., G. L. Insoluble monolayers at liquid-gas interfaces; Interscience Publishers: New York, 1966. 283. Knopf, D. A.; Anthony, L. M.; Bertram, A. K. J. Phys. Chem. A 2005, 109, 55795589. 284. Knopf, D. A.; Mak, J.; Gross, S.; Bertram, A. K. Geophys. Res. Lett. 2006, 33, L17816, doi:17810.11029/12006GL026884. 285. Huey, L. G.; Hanson, D. R.; Howard, C. J. J. Phys. Chem. 1995, 99, 5001-5008. 286. Mason, E. A.; Monchick, L. J. Chem. Phys. 1962, 36, 2746-2757. 287. Monchick, L.; Mason, E. A. J. Chem. Phys. 1961, 35, 1676-1697. 288. Patrick, R.; Golden, D. M. Int. J. Chem. Kinet. 1983, 15, 1189-1227. 289. Hanson, D. R.; Ravishankara, A. R. J. Geophys. Res. 1991, 96, 5081-5090. 290. Hanson, D. R. J. Phys. Chem. A 1998, 102, 4794-4807. 291. Motz, H.; Wise, H. J. Chem. Phys. 1960, 32, 1893-1894.  Chapter 4^  128  292. Rolo, L. I.; Caco, A. I.; Queimada, A. J.; Marrucho, I. M.; Coutinho, J. A. P. J Chem. Eng. Data 2002, 47, 1442-1445. 293. Adamson, A. W.; Gast, A. P. Physical chemistry of surfaces, 6th Ed. ed.; WileyInterscience: New York, 1997. 294. Myrick, S. H.; Franses, E. I. Colloid. Surface. A 1998, 143, 503-515. 295. Kaganer, V. M.; Mohwald, H.; Dutta, P. Rev. Mod. Phys. 1999, 71, 779-819. 296. Lawrie, G. A.; Barnes, G. T. J. Colloid Interf. Sci. 1994, 162, 36-44. 297. La Mer, V. K.; Aylmore, L. A. G.; Healy, T. W. J. Phys. Chem. 1963, 67, 27932795. 298. Chou, T. H.; Chu, I. M. Colloid. Surface. B 2003, 27, 333-344. 299. Motomura, K. J. Colloid Interf. Sci. 1974, 48, 307-318. 300. Seoane, R.; Minones, J.; Conde, 0.; Minones, J.; Casas, M.; Iribarnegaray, E. J. Phys. Chem. B 2000, 104, 7735-7744. 301. Barnes, G. T. J. Colloid Interf. Sci. 1991, /44, 229-300. 302. Barnes, G. T.; La Mer, V. K. Retardation of Evaporation by Monolayers: Transport Processes; Academic Press: New York, 1962. 303. Archer, R. J.; La Mer, V. K. J. Phys. Chem. 1955, 59, 200-208. 304. Langmuir, I.; Schaefer, V. J. J. Franklin Inst. 1943, 235, 119-162. 305. Barnes, G. T.; Bacon, K. J.; Ash, J. M. J. Colloid Interf. Sci. 1980, 76, 263-264. 306. Garrett, W. D. J. Atmos. Sci. 1971, 28, 816-819. 307. Rao, Y. K.; Shah, D. 0. 1 Colloid Interf. Sci. 1989, 137, 25-29. 308. Fukuda, K.; Kato, T.; Machida, S.; Shimizu, Y. J. Colloid Interf. Sci. 1979, 68, 8295.  Chapter 4^  129  309. McNamee, C. E.; Barnes, G. T.; Gentle, I. R.; Peng, J. B.; Steitz, R.; Probert, R. J. Colloid Interf. Sci. 1998, 207, 258-263. 310. Rosano, H. L.; La Mer, V. K. J. Phys. Chem. 1956, 60, 348-353.  Chapter 5^  130  5. N205 REACTIVE UPTAKE ON AQUEOUS SULFURIC ACID SOLUTIONS CONTAINING DICARBOXYLIC ACID ORGANIC SURFACTANTS  5.1 Introduction Research has shown that aqueous particles in the atmosphere can be coated with organic monolayers. 311-315 These monolayers can consist of insoluble and soluble surfactants. 311 ' 313-328 Organic monolayers on atmospheric aerosols may decrease reaction rates. Recently, researchers have focused on the effect of organic monolayers on the hydrolysis of N205. 329-336 In these studies both soluble and insoluble monolayers were used. For insoluble monolayers, long chain alcohols and mono-carboxylic acids were studied. 333-335 Soluble surfactants with similar functional groups have also been investigated. 33°-332 '336 One class of molecules that has been largely overlooked are dicarboxylic acids. Thornton et al. 337 have shown that N20 5 uptake on solid dicarboxylic acid aerosols containing malonic acid or azelaic acid is slow. Their studies of N205 uptake on aqueous malonic acid aerosols showed no decrease in reactive uptake due to the presence of the dicarboxylic acid. 337 However, more work in this area is needed to determine if dicarboxylic acids with longer chain lengths (ie. more carbon atoms) can potentially shut off N205 uptake at the aqueous interface. Dicarboxylic acids have been shown to be an important component of atmospheric aerosols. 338 ' 339 Concentrations of dicarboxylic acids in the atmosphere range from 175 ng m -3 to 2044 ng m -3 during biomass burning events in rural areas 324 and 199 ng m -3 to 312 ng m 3 for annual average concentrations in urban environments. 339 Sources include fossil fuel burning, 340,341 biomass burning, 324,326 meat cooking, 339 ' 342 and photochemical oxidation. 343 These dicarboxylic acids may impact the transfer of gasphase species across the air-aqueous interface. In the following, we focus on glutaric acid (pentanedioic acid) and azelaic acid A version of this chapter will be submitted for publication. Cosman, L. M., Duffy, I., and Bertram, A. K., N 2 05 Reactive Uptake on Aqueous Sulfuric Acid Solutions Containing Dicarboxylic Acid Organic Surfactants, J. Phys. Chem. A.  Chapter 5^  131  (nonanedioic acid), two types of dicarboxylic acids. Concentrations of these species range from 1.8 ng m -3 to 47.1 ng m -3 for glutaric acid, 324 ' 339 and 1.3 ng m -3 to 44.7 ng m -3 for azelaic acid. 324 ' 339 Here, we investigate if glutaric acid and azelaic acid partition to the interface in aqueous 40 wt % H2SO4 solutions at 260 K and whether these organics influence N20 5 uptake. We also investigate the partitioning of glutaric acid and azelaic acid to the interface of water at 293 K, and speculate on whether or not these monolayers on water at 293 K can impact N205 reactivity. Glutaric acid was selected for N20 5 uptake studies due to its large surface excess and high solubility in the aqueous media. Azelaic acid was chosen for the uptake experiments due to its chain length, which could be important for shutting off the reaction. 40 wt % H2SO4 solutions at 260 K were chosen for atmospheric relevance and due to experimental constraints. At warmer temperatures the vapor pressure of water over the aqueous solution is large, which significantly limits the range of reactive uptake coefficients we can measure with our experimental apparatus due to large gas-phase diffusion corrections (see below for further details).  5.2 Experimental To produce dicarboxylic acid monolayers we dissolve the organic compounds in aqueous acid solutions. The organics then partition to the air-aqueous interface to form a surface excess. To characterize the monolayers we measure the surface tension of the organic-aqueous acid solution versus each dicarboxylic acid concentration and calculate the surface excess of dicarboxylic acid as a function of concentration using Gibbs' adsorption isotherm. This allows us to characterize the properties of the monolayers. Reactive uptake coefficients are measured using a rectangular channel flow reactor 333-335 coupled to a chemical ionization mass spectrometer. 333-335 ' 344 '345 Each of these processes is described in detail below. 5.2.1 Measurement of Dicarboxylic Acid — H2O — H2SO4 Surface Tensions The surface tensions (o) of dicarboxylic acids in aqueous 40 wt % H 2 SO 4 at 260 K and dicarboxylic acids in water at 295 K were measured using the Wilhelmy plate method. 346 We measured the downward force (using a surface pressure sensor; NIMA  Chapter 5^  132  Technology, model PS4) on a 23.32 mm platinum plate while the plate was immersed and detached from the liquid solution. The surface tension was calculated from the maximum difference in the force on the plate between immersion and separation from the solution. The surface of the platinum plate was roughened to ensure a contact angle of nearly zero. 347 Measurements at 260 K were made in a closed environment purged with dry N2 to minimize water uptake. The nitrogen passed through a drierite trap cooled with an ethanol-dry ice bath to minimize organic impurities. Sulfuric acid solutions were prepared using sulfuric acid and Millipore water. Dicarboxylic acid solutions were made by weighing the desired mass of solid organic with a balance that has an accuracy of ± 1 mg, followed by dissolving the solid in the aqueous solution. Each concentration was prepared individually in 50 mL or 100 mL Pyrex volumetric flasks. Surface tension measurements were performed in a 5 cm diameter by 7 cm tall Pyrex beaker. For measurements at 260 K the beaker was then submerged in a coolant bath. The platinum plate was immersed in the dicarboxylic acid solution for approximately 5 minutes to stabilize prior to measuring the surface tension. 6-12 consecutive measurements were recorded over a 15-30 minute timeframe, with a complete set of experiments requiring 68 hours. The standard deviation of the 6-12 measurements was ± 0.3 mN m -1 . 5.2.2 Flow Reactor and Experimental Conditions for the Reactive Uptake Measurements The experimental apparatus has been described in detail in sections 2.2 and 3.2, and therefore will not be described again. One modification in the experimental set-up is that a glass trough with a liquid surface of 7.5 cm in width and 6 cm in length was used for the uptake measurements at 260 K. During uptake experiments, a saturated flow of N205 between 8.5 to 10 cm 3 min -1 at STP was mixed with 20-40 cm 3 mid i at STP of dry He prior to entering the flow reactor through the T-shaped injector. Total mass flow rates inside the flow reactor ranged from 550 to 685 cm 3 min-1 at STP and total pressures ranged from 3.1 to 3.4 Toff. Under these conditions the Reynolds number varied from 1.2 to 1.5, indicating laminar flow conditions.  133  Chapter 5^  The exit of the flow cell is connected to a chemical ionization mass spectrometer (CIMS), which is used to measure the change in the gas phase reactant concentration as a result of reactive uptake at the liquid surface. 344 ' 345 N20 5 was detected as NO3 - after its chemical ionization by F. 3" ' 348 '349 I - was formed by flowing trace amounts of CH3I diluted in 1250 to 1500 cm 3 min d at STP of N2 through a polonium-210 source (NRD, model Po-2031) for ionization. For N 2 0 5 detection in the presence of H2O the chemical ionization region was biased to -122 V in order to fragment weakly bound ion-H20 clusters. 337 ' 349 N205 concentrations of 2.2 x 10 10 to 1.2 x 10 11 molec cm -3 were used for the uptake measurements. N205 concentrations were based on the I - + N205 chemical ionization reaction rate that has been reported in the literature. 348 In a typical uptake experiment the N205 signal was measured as a function of injector position. The natural logarithm of this signal was then plotted as a function of reaction time, which was calculated from the reaction length and the flow velocity, in order to determine the observed first order loss rate constant (ko b s ). From /cob s we then determined the first order wall loss rate constant (k w ) using the procedure developed in our previous work. 334 This procedure corrects for any concentration gradients that can develop in the flow reactor due to a fast heterogeneous loss at the liquid surface. This procedure decouples reaction and diffusion to the aqueous surface in order to determine the true first order wall loss rate constant. Diffusion coefficients were needed in order to calculate kw from ko b s . The diffusion coefficients used in this study were calculated using molecular parameters. 35°352 For the diffusion coefficient of N205 in He  (DN205 _ He ) at 260 K, we used a value of  266 Torr cm 2 s d and for the diffusion coefficient of N205 in H2O ( D N205—H20) at 260 K, we used a value of 64.5 Torr cm 2 . To calculate the diffusion coefficient of N20 5 in a mixture of helium and water, D, we used the following equation: 353 ^ 5.1  1^PHe^PH20 D D N205—He DN 0 -11 0 2  5  2  where PH, and PH20 are the partial pressures of helium and water vapor in the flow reactor, respectively.  Chapter 5^  134  The reactive uptake coefficient (y) was determined from lf,,, using the following equation. 354 '355 5.2^  1^c S 1 = y 41c,, V 2  where c is the mean molecular velocity of N205, S is the reactive surface area inside the flow reactor, and V is the volume of the open channel above the liquid surface. 5.2.3 Chemicals Listed below are the chemicals, the manufacturer, and the corresponding purities of the chemicals used in our studies: He (Praxair, 99.999% Purity), N2 (Praxair, 99.999%), NO 2 (Matheson, 99.5%), 0 2 (Praxair, 99.5%), P 2 O 5 (Aldrich, 97%), glutaric acid (Aldrich, 99%), azelaic acid (Fluka, 99+%), and sulfuric acid (Fisher, 95-98%).  5.3 Results and Discussion 5.3.1 Surface Tension and Surface Excess of Dicarboxylic Acids in Aqueous Sulfuric Acid at 260 K Prior to the N205 uptake we first studied the surface partitioning of dicarboxylic acids to the interface. These studies were necessary for understanding and interpreting the N205 uptake data on aqueous sulfuric acid at 260 K. We started by investigating the surface tension and surface excess of glutaric acid and azelaic acid in aqueous 40 wt % H2SO4 solutions at 260 K. The extent of dicarboxylic acid partitioning to the interface will depend, in part, on the bulk and surface properties of the sulfuric acid solution. 356 For aqueous 40 wt % H2SO4 at 260 K the bulk composition consists primarily of H2O, H 3 0 ± , and HSO4, with less than 4 % SO4 2- and negligible molecular H2SO4. 357-359 Gibbs adsorption analysis of H2SO4 solutions at 294 K suggest that sulfuric acid species are depleted at the surface by 6 x 10 12 cm -2 for an aqueous 40 wt % (11 mol %) sulfuric acid solution.36°  135  Chapter 5^  Dicarboxylic acids are readily protonated in sulfuric acid and undergo the following equilibria:  ^R5.1 ^R5.2 ^R5.3 ^R5.4  ^ ^ ^ ^  H2SO4 + H2O 4-3' HSO4 - +  HSO4 + H2O <---> 504 2- +  R - (CO2 H )2 + H2 S O4 HSO4- + H O 2C -R - CO2H2 +  HO2C - R - CO2H2 + + H2SO4 4+). HSO4 + R - (CO2H2)2 2+  where R—(CH2),, and n = 3 and 7 for glutaric acid and azelaic acid, respectively. Both glutaric acid and azelaic acid are partially ionized as diacid bases.  361  As such, surface  species are expected to be a mixture of R-(CO2H)2, HO2C-R-CO2H2 + , and R-(CO2H2)2 2+ according to the above equilibrium equations. Decarboxylation can occur in concentrated sulfuric acid (concentrations > 90 wt % H2SO4) but is expected to be negligible under our experimental conditions using aqueous 40 wt % H2SO4. 362 The surface tension measurements do not distinguish between the surface active dicarboxyl species but investigate the overall effect of these species on the surface tension. Figure 5.1a shows the surface tension of glutaric acid in aqueous 40 wt % H2SO4 at 260 K (solid circles) as a function of glutaric acid concentration over the concentration range of 0 — 0.35 M glutaric acid. The surface tension values gradually decrease as the glutaric acid concentration increases in the aqueous acid solutions. The minimum surface tension for glutaric acid solutions near its saturation with respect to the bulk crystal (0.35 M glutaric acid) is 69.8 ± 0.3 mN in 1 in 40 wt % H2SO4 at 260 K.  136  Chapter 5^  0.0  0.1  0.2  0.3  0.4  80  (a)  80  78  78  76  76  74  74  72  72  70  70  68  68  16  16  14 12 10 8 6 4 2 0 —2  • •  •  •  •  ••  •  •  •  (b)  12  Nc  10 8  O  6 Xn  4 2  •  0.0  14  0 0.1  0.2  0.3  0.4  —2  glutaric acid concentration / M  Figure 5.1: Surface tension of glutaric acid — aqueous 40 wt % sulfuric acid solutions and surface excess (FCaill:j °0) xyr ) of glutaric acid as a function of glutaric acid concentration at 260 K. Solid circles represent surface tensions (panel (a)) and surface excess of glutaric acid (panel (b)) in 40 wt % H2SO4 at 260 K. The solid line in panel (a) represents a third-order polynomial fit to the data.  In order to understand the surface properties of dicarboxylic acid monolayers in aqueous 40 wt % H2SO4 at 260 K and how they may affect N20 5 uptake, the surface  Chapter 5  137  excess of dicarboxyl species in each solution was calculated from the Gibbs' adsorption isotherm equation: 346 — do. = vn  5.3^  r(H 2 0)dpi  where ,u, is the chemical potential of component  i  and F, (H2°) is the surface excess of  component i relative to H2O. The ternary dicarboxylic acid — H 2 O — H2SO 4 system potentially contains eight species: H2O, H2SO4, HSO4 - , SO4 2- , H30 + , R-(CO2H) 2 , HO2C-R-CO2H2 + , and R(CO2H2)2 2+ . The Gibbs equation for this system is: +F (H 2 o) d^r (H 2 o )d 5.4^ do- =F (H2°) d,u^ +F(H2°)d,u^ H 2 SO4^H2SO4^ H 3 o-^H 3 o.^ -  HSOY  P HSOY^801- 11 SO: -  F (H 2 o)^d ,^±r(H,o) (H 0) R_(092 1- )2 fr.R-(cog ),^HO2 c-R-co 2 HZd^+ F 2 il Ho2 c-R-0O2 H ;^R_(co2 H ) 3+ dp R_(co2H)3+  The surface excess of a species  i  can be determined from F, =  --(au/ 1 fli ( -^au --I/T,  j *i,H 2 0)  if  the chemical potential of the other species are fixed. The temperature and the initial starting concentration of sulfuric acid are controlled, but may vary as the dicarboxylic acid is added. Equation 5.4 can be simplified by relating the chemical potentials through reactions R5.1 - R5.4 and also by imposing electrical neutrality using the following equation: 011 20 ) r(H 2 0)^+2r(H20)^= r(H 2 0) + 2-r(H 2 0)  5.5  H 3 0 +^HO2C-R-0O2112'^R-(CO211);+^HSOy^  These conditions allow eq 5.4 to be rearranged in terms of  I SQV  dp R „,-,, „ ,^ , fi t, Cyl k,  and  12  4i/ 2 o  to give:  = r^ (H20)^ 5.6^— (ao I au^ r(H20)^ 2F(H2o) 1 (H2°) + rui,o) ,92 11),T )^&carboxyl ± [2FSol— HSOy^ — HO2C-R-0O21-12'^ R-(co2HT ' -  L  -  (aPH 2 0 / aPR-(co2H)2)T +  Fr (H20) + F(H20) + F  2 0)  L 2 ,504^HSOy^SQV K ali H 2 SO4  I 4311 R-(CO2 H) 2  )T  uf 2 o)^+ F uf2 o)^+ (H °) where F &carboxyl (H26) =— -1-rR(CO2102  ^(HO2C)R(CO2H2)+ ^R(CO2112)i+ •  Equation 5.6 demonstrates that the change in surface tension depends not only on the surface excess of the dicarboxyl species, rrca2rObo) xyi , but also on the changes in chemical potentials of water and sulfuric acid with added dicarboxylic acid.  138  Chapter 5^  Torn and Nathanson 356 have investigated surface segregation of 1-butanol in 0-72 wt % H2SO4. The change in  H 2 Op has  been measured for the addition of butanol to pure  water and 2-butanol to 0.1 M Na 2 SO4, and (a .-PH 2 0 /0PR0H)T was shown to be small in both cases (< 0.005) at low butanol concentrations. 363 While the change in chemical potentials with the addition of butanol is more difficult to assess in concentrated H2SO4, Torn and Nathanson 356 estimate that the change in (  II,S0 4  I (311ROH)T contributes less  than 10% to /butyl. The change in chemical potential of H 2 O and H2SO4 with added glutaric acid or azelaic acid has not been measured. Here we assume that the change in the chemical potentials of H2O and H2SO4 is small upon addition of small quantities of glutaric acid or azelaic acid (similar to the assumptions made by Torn and Nathanson 356 ) but further work is needed to verify this assumption. If the chemical potential of H2SO4 or H2O does change upon the addition of organic species the observed decrease in the surface tension could arise from this change in chemical potentials. We can not definitively identify the cause for the change in the surface tension of the solutions, but we assume it is due to the organic species partitioning to the interface. However, once again this assumption needs to be verified. Using this assumption, the relative surface excess of dicarboxyl species can be approximated using: r (H,o)^ a dicarboxylic acid dicarboxyl RT  •  —  5.7  \  aa dicarboxylic acid  p) where adicarboxylic acid is the activity of the dicarboxylic acid, and Fdi( Hcarbo ^iss the relative  surface excess of all dicarboxyl-containing species as defined above. Assuming the activity of the dicarboxylic acid varies linearly with concentration 364 ' 365 (Cdicarboxylic acid) the relative surface excess may be determined using the surface tension data shown in Figure 5.1 a according to eq 5.8: ■ 5.8^  H1'20)^Cdicarboxylic acid  dicarboxyl  RT^,s ac dicarboxylic acid j  139  Chapter 5^  Figure 5.1 b shows the variation of surface excess as a function of glutaric acid concentration in aqueous 40 wt % H2SO 4 solutions at 260 K. Surface excess values were calculated by fitting the surface tension data in Figure 5.1 a to a third-order polynomial fit.  366  (H2O) The derivative of that fit was then used in eq 5.8 to calculate -' &carboxyl • 364 ' 366 Other  types of functions may be used (second-order polynomial, exponentialpolynomial) 3 64,365,367,368 and give comparable fits to the data. We estimate the uncertainty in the surface excess at saturation with respect to the bulk crystal to be of the order of 35%, based upon the different values obtained using different functions to fit the data. Glutaric acid solutions reached a surface excess of 13 ± 4 x 10  13  cm-2 near the crystalline  saturation point of glutaric acid. Presented in Figure 5.2 are the surface tension and surface excess results for azelaic acid in aqueous sulfuric acid solutions as a function of azelaic acid concentration over the concentration range of 0 — 5.3 x 10 4 M azelaic acid. Panel (a) shows the decrease in surface tension as a function of increased azelaic acid concentration in aqueous 40 wt % H2SO4 at 260 K (solid circles), for concentrations up to the saturation of crystalline azelaic acid in the solution. The solid line represents a third-order polynomial fit to the data. Note that the solubility of azelaic acid is much lower than that for glutaric acid in aqueous 40 wt % H2SO4 at 260 K, as evidenced by the smaller concentration range studied. Panel (b) shows the surface excess values for azelaic acid — aqueous sulfuric acid solutions at 260 K. Surface excess values were calculated using the same procedure discussed above. Surface excess values for azelaic acid in aqueous sulfuric acid at 260 K are 6 ± 2 x 10 13 cm -2 for the most concentrated solutions. This is lower than that for glutaric acid, despite the longer chain length of azelaic acid.  140  Chapter 5^  0.0000  0.0002  0.0004  0.0006  78  78  76  76  74  74  8  (b) -  • ••  8  •• • • • • _ 6  •  4  0 D  2 0 0  •  0  0.0000^0.0002^0.0004^0.0006  azelaic acid concentration / M  Figure 5.2: Surface tension of azelaic acid — aqueous 40 wt % sulfuric acid solutions and surface excess ( F.:.1 2r°b0) xyr ) of azelaic acid as a function of azelaic acid concentration at 260 K. Solid circles represent surface tensions (panel (a)) and surface excess of azelaic acid (panel (b)) in 40 wt % H2SO4 at 260 K. The solid line in panel (a) represents a third-order polynomial fit to the data.  Chapter 5^  141  5.3.2 Reactive Uptake of N 2 0 5 With Aqueous Sulfuric Acid in the Presence of Dicarboxylic Acids at 260 K The reactive uptake coefficients N205  (y)  are determined from the irreversible loss of  as a function of injector position as mentioned above. Shown in Figure 5.3 are  plots of the natural logarithm of the removal of  N205  N205  signal as a function of reaction time for the  on aqueous 40 wt % sulfuric acid solutions at 260 K, and aqueous  sulfuric acid solutions containing dicarboxylic acids. The data for each uptake experiment was fit to a straight line, and from the slope of this line the first-order rate constant, kobs, was determined. kW is then calculated from kob s , and used to calculate the reactive uptake coefficient after accounting for changes in the concentration profiles due to uptake and diffusion in vertical direction and in direction of the bulk flow. 334  0.000^0.002^0.004^0.006^0.008^0.010 0.0 -0.2  ^ „\.  - 0.0 - -0.2  -0.4  - -0.4  -0.6  - -0.6  -0.8  - -0.8  -1.0 Z" c -1.2  - -1.0  (7)  - -1.2  -1.4 -^  - -1.4  -1.6 '^'^'^'^'^1.6 0.000^0.002^0.004^0.006^0.008^0.010 fime/s  Figure 5.3: Experimentally derived natural logarithms of the gas-phase N205 signals as a function of reaction time. Open squares: blank uptake (no aqueous solution in the flow cell), open circles: uncoated aqueous 40 wt % H2SO4, solid triangles: glutaric acid, solid diamonds: azelaic acid. All data was collected at 260 K.  142  Chapter 5^  The reactive uptake coefficient for the uptake of N205 on uncoated aqueous 40 wt % H2SO4 at 260 K is given in Table 5.1, in addition to y for aqueous 40 wt % H2SO4 solutions containing glutaric acid or azelaic acid. Solution concentrations of 280 mM and 0.48 mM were used for glutaric acid and azelaic acid, respectively, which are close to their solubility limits. Also shown in Table 5.1 are the surface excess r (o H2 ) ) values for these solution concentrations, which were used to calculate the & carboxyl  molecular surface area for glutaric acid and azelaic acid in aqueous 40 wt % H2SO4. values reported in this study are based on 3-12 uptake experiments performed. The upper and lower limits for y take into account 20 % error in the pressure independent diffusion coefficients.  Table 5.1: The Reactive Uptake Coefficient (y) of N20 5 on 40 wt % Sulfuric Acid Solutions at 260 K in the Presence and Absence of Organic Monolayers  organic^[organic] (mM) uncoated  77-  (mNin -1 )  r(H 20) dicarboxyl  x o" cm -2 )  surface area (A2 molec -1 )  0^0^0^0  +0 087 0.053 — . 0.014  glutaric acid^280^7.2^12 ± 4^83 ± 29  0.040 +0.021 —0.007  azelaic acid^0.48^3.8^6.0 ± 2^167 ± 58  0.051 +0.042 —0.012  The reactive uptake coefficient for N205 on aqueous 40 wt % H2SO4 was not found to decrease significantly within the uncertainty of our experiment when comparing the 7 for the uncoated aqueous solution to that for solutions containing glutaric acid or azelaic acid. This is most likely because the packing density of the organic molecule is small. This is discussed in more detail below. In our previous manuscript we showed that the uptake coefficient for N205 on aqueous sulfuric acid solutions correlates with the packing density of the monolayer. 335 In Figure 5.4 we include the data measured in this study, and the data presented in our  Chapter 5^  143  previous study, to determine if dicarboxylic acids follow the same trend of y decreasing as the packing density of the surfactant increases. The reactive uptake coefficients for N205 uptake on aqueous 40 wt % H2SO4 solutions at 260 K in the presence of glutaric acid and azelaic acid are shown as solid symbols in Figure 5.4. The open symbols represent literature results of yfor N205 uptake on sulfuric acid solution performed previously. 331 ' 333-335 Our data for glutaric acid and azelaic acid fits the strong correlation of y decreasing with increasing surface density presented by the literature data. This is consistent with the Accessible Area Theory. 369 This data interpretation assumes that the change in the surface tension results from partitioning of the organic to the interface and not a change in the chemical potential of H2SO4 or H 2 O, which needs to be verified in further experiments (see discussion in section 7.3.1 for more details). Although we only show data for N2O5 uptake in the presence of monolayers on sulfuric acid solutions, other studies have been done using different subphases, which show a decrease in yin the presence of partial monolayers. McNeill et a1. 339 showed that y was decreased for N205 uptake on NaCI and natural seawater in the presence of partial monolayers of sodium dodecyl sulfate. McNeill et a1. 339 found that they saw a decrease in y for monolayers with surface densities of — 14 x 10 13 cm -2 , which is comparable to surface densities for glutaric acid investigated in this study (12 ± 4 x 10 13 cm 2 ). While our results do not agree with McNeill et al. based solely on packing density of the organics on the surface, yfor N2O5 on aqueous solutions has also been shown to correlate with chain length, 333 which may explain this discrepancy since sodium dodecyl sulfate has a hydrocarbon chain containing 12 carbon atoms and glutaric acid has only 5 carbon atoms. It is also possible that the different subphases in these two studies may play a role in the N2O5 uptake kinetics and warrant further investigation.333'335  Chapter 5^  144  20 40 60 80 100 120 140 160 180 I^'^I  0. 1  0.1  (7_ =^0.01  0.01  i 1 E-3  1 E-3  1  1  a) co  0  C.)  0.1  0.1  E  0.01  0.01 I^I^I  20 40 60 80 100 120 140 160 180  molecular surface area / A 2 mole&'  Figure 5.4: The reactive uptake coefficient 01 for N20 5 on aqueous solutions of sulfuric acid coated with organic monolayers as a function of molecular surface area. Solid symbols represent the results from this study for yin the presence of dicarboxylic acids and the open symbols represent the results in the literature for y in the presence of alcohols and monocarboxylic acids. Solid square: glutaric acid (this study), solid circle: azelaic acid (this study), open square: 1-octadecanol (Cosman et al.), 335 open hexarn: phytanic acid (Cosman et al.), 335 open circle: 1octadecanol (Cosman et al.), 33 open inverted triangle: stearic acid (Cosman et al.), 333 open star: phytanic acid (Cosman et al.), 333 open triangle: 1-hexadecanol (Cosman et al.), 333 open diamond: butanol (Park et al.), 331 open sideways triangle: hexanol (Park et al.), 331 x: 1-octadecanol (Knopf et al.). 334 Panel (a) gives the ordinate as a function of yin the presence of the monolayer (yfil m ). In panel (b) the ordinate is normalized to yfor the uncoated solution (vuncoated)•  145  Chapter 5^  5.3.3 Surface Tension and Surface Excess of Dicarboxylic Acids in Water at 293 K For acidic particles in the atmosphere, we have shown that the packing density of dicarboxylic acids is not sufficient to have a significant effect on the hydrolysis of N20 5 . This is most likely because the partitioning of the dicarboxylic acids to the surface is low. However, what about in pure water droplets or dilute aqueous aerosols? To investigate this, we have determined the surface excess of these same organics in pure water at 293 K, using surface tensions collected in this study in addition to data previously reported in the literature. Figure 5.5a shows the surface tension of glutaric acid — water solutions at 295 K for data collected in this study (open circles) and those of Shulman et al. squares) as a function of glutaric acid concentration. The surface excess  37°  (open  ( 1—,1( ,H cfbo) xy  ) of  glutaric acid as a function of concentration is shown in panel (b), calculated from the derivative of the fit of the surface tension data in panel (a). Surface segregation of glutaric acid shows that, in pure water at 295 K, glutaric acid can attain a surface excess up to 20 ± 7 x 10 13 cm 2 . This corresponds to a molecular surface area of 50 ± 18 A 2 molec -1 . Figure 5.6a shows the surface tension of azelaic acid — water solutions at 295 K for data collected by Tuckermann and Cammenga 371 as a function of azelaic acid (H20 ) concentration. The surface excess (F &caoxyl ) of azelaic acid as a function of concentration rb  is shown in panel (b), calculated from the derivative of the fit of the surface tension data in panel (a). Surface segregation of azelaic acid shows that in pure water at 295 K the surface excess at saturation with respect to crystalline azelaic acid is 23 ± 8 x 10 13 cm 2 . This corresponds to a molecular surface area of 43 ± 15 A 2 molec -1 . Here the packing density is much higher than in acidic solutions at 260 K. This increased packing density suggests azelaic acid may be important for influencing N205 uptake.  Chapter 5^  146  0.0^0.5^1.0^1.5^2.0^2.5 72  72  68  68  64  64  60  60  56  56  52  52  24  24  •  20  E 0  (b) • - 20  16  16  12  • • •  8 4 0  •  ••  •  -•  -•  12 8 4 0  I^A^I  0.0^0.5^1.0^1.5^2.0^2.5  glutaric acid concentration / M  Figure 5.5: Surface tension of glutaric acid — water solutions at 295 K and surface r uf 2 0 ) excess (idicarbo of glutaric acid as a function of glutaric acid concentration. xyl Panel (a) - open circles represent surface tension data collected in this study, open squares are surface tension data from Shulman et al 3 70 The solid line represents a third-order polynomial fit to the data. Panel (b) - surface excess of glutaric acid in water at 295 K calculated by fitting raw data from both this study and Shulman et al.3" )  Chapter 5^  147  0.000 0.005 0.010 0.015 0.020 0.025 0.030 75^ 75 70  70  65  65  60  60  55  55  50  50  45  45 I  28 24  E  0  20 16  • 8 - • 4 - •  •  •  •  ^I^I^I  (b) •  12 5,  0  O  t  0  28 24 20 16 12 8 4 0  0.000 0.005 0.010 0.015 0.020 0.025 0.030  azelaic acid concentration / M  Figure 5.6: Surface tension of azelaic acid — water solutions at 295 K and surface excess (. Fu C Hca2r°b0)xyr ) of azelaic acid as a function of azelaic acid concentration. Panel (a) - open triangles are surface tension data from Tuckermann et 0. 371 The solid line represents a third-order polynomial fit to the data. Panel (b) - surface excess of azelaic acid in water at 295 K calculated by fitting raw data from Tuckermann et al. 371  Chapter 5^  148  5.3.4 Reactive Uptake of N205 With Dilute Aqueous Solutions in the Presence of Dicarboxylic Acids at 293 K Above we discussed the surface concentrations of glutaric acid and azelaic acid on water solutions at 295 K, and we show the surface concentrations are 20 ± 7 x 10 13 cm -2 for glutaric acid and 23 ± 8 x 10 13 cm -2 for azelaic acid. Due to diffusion limitations, it was not possible to measure the reactive uptake coefficient of N205 on water surfaces at 293 K coated with these organic monolayers. However, as a starting point we speculate on the effect of the monolayer on N205 uptake using the surface segregation data and the data shown in Figure 5.4. Based on a comparison of the packing density of dicarboxylic acids on water to that on acidic solutions, we observe that the packing density of dicarboxylic acids increases as the particles become more dilute. From this observation we speculate, based on the packing density studies in this paper combined with the N205 uptake trends in the literature, that very dilute particles containing dicarboxylic acids may have sufficient packing densities to inhibit heterogeneous chemistry by up to a factor of 2-3.  5.4 Conclusions and Summary The surface activities of two atmospherically relevant dicarboxylic acids (glutaric acid and azelaic acid) were characterized in aqueous 40 wt % sulfuric acid solutions at 260 K. Glutaric acid showed the largest decrease in the surface tension of the aqueous acid solutions; a decrease in surface tension up to 10%. Glutaric acid exhibited the largest surface excess, 13 ± 4 x 10 13 cm 2 , near its solubility limit in aqueous 40 wt % H2SO4 at 260 K. The solubility of azelaic acid was very low in aqueous 40 wt H2SO4 at 260 K, and yielded a surface excess of 6 ± 2 x 10 13 cm -2 at the solubility limit with respect to crystalline azelaic acid. Keep in mind that for this analysis we assumed the change in chemical potentials of H2O and H2SO 4 are small upon addition of small quantities of glutaric acid or azelaic acid. Further work is needed to verify this. The reactive uptake coefficient (y) for N205 on aqueous sulfuric acid solutions at 260 K was studied in the presence of glutaric acid and azelaic acid monolayers. y for N205 on aqueous 40 wt % H2SO4 at 260 K was determined to be 0.0531 °08, 74 . In the  Chapter 5^  149  presence of glutaric acid and azelaic acid monolayers y for N205 on aqueous 40 wt % 1 and 0.0511 0122 respectively. These 7 values H2SO4 at 260 K was found to be 0.040+ 00 0020 .7 are not significantly different from y for the uncoated solution within experimental uncertainty. Reactive uptake coefficients were correlated with the molecular surface area of glutaric acid and azelaic acid on the aqueous acid surface and are in agreement with results observed by Cosman et a1. 335 (and references within). This correlation now includes organic components consisting of mono- and di-functional organic molecules that form both soluble and insoluble monolayers. Based on the correlation of N 2 05 uptake with packing density discussed above, our analysis of surface tension data 370 '371 to yield packing densities for dicarboxylic acids in water at 295 K allows us to predict the effect of dicarboxylic acid monolayers on N205 uptake on water droplets and dilute aqueous particles in the atmosphere. The surface excess values for glutaric acid and azelaic acid on water at 295 K were determined to be 20 ± 7 x 10 13 cm -2 and 23 ± 8 x 10 13 cm 2 , respectively. This is much higher than observed for the same dicarboxylic acids on aqueous 40 wt % acid aerosols at 260 K. Thus, the surface of water droplets and dilute aqueous particles in the atmosphere will likely be more densely packed with dicarboxylic acids than particles that have higher acidity. We predict that N20 5 uptake on atmospheric particles coated with a monolayer of water soluble dicarboxylic acids will decrease by a factor of 2-3.  Chapter 5^  150  5.5 References 311. Peterson, R. E.; Tyler, B. J. Atmos. Environ. 2002, 36, 6041-6049. 312. Russell, L. M.; Maria, S. F.; Myneni, S. C. B. Geophys. Res. Lett. 2002, 29, doi:10.1029/2202GL014874. 313. Tervahattu, H.; Hartonen, K.; Kerminen, V. M.; Kupiainen, K.; Aarnio, P.; Koskentalo, T.; Tuck, A. F.; Vaida, V. J Geophys. Res.-Atmos. 2002, 107, 4053, doi:4010.1029/2000JD000282. 314. Tervahattu, H.; Juhanoja, J.; Kupiainen, K. J Geophys. Res.-Atmos. 2002, 1 0 7, 4319, doi:4310.1029/2001JD001403. 315. Tervahattu, H.; Juhanoja, J.; Vaida, V.; Tuck, A. F.; Niemi, J. V.; Kupiainen, K.; Kulmala, M.; Vehkamaki, H. I Geophys. Res. Atmos. 2005, 110, D06207, doi:06210.01029/02004JD005400. 316. Barger, W. R.; Garrett, W. D. I Geophys. Res. 1970, 75, 4561-4566. 317. Barger, W. R.; Garrett, W. D. J. Geophys. Res.-0c. Atm. 1976, 81, 3151-3157. 318. Cavalli, F.; Facchini, M. C.; Decesari, S.; Mircea, M.; Emblico, L.; Fuzzi, S.; Ceburnis, D.; Yoon, Y. J.; O'Dowd, C. D.; Putaud, J. P.; Dell'Acqua, A. I Geophys. Res.-Atmos. 2004, 109, D24215, doi:24210.21029/22004JD005137. 319. Decesari, S.; Facchini, M. C.; Fuzzi, S.; McFiggans, G. B.; Coe, H.; Bower, K. N. Atmos. Environ. 2005, 39, 211-222. 320. Decesari, S.; Facchini, M. C.; Fuzzi, S.; Tagliavini, E. I Geophys. Res. Atmos. 2000, 105, 1481-1489. 321. Facchini, M. C.; Decesari, S.; Mircea, M.; Fuzzi, S.; Loglio, G. Atmos. Environ. 2000, 34, 4853-4857.  Chapter 5^  151  322. Facchini, M. C.; Fuzzi, S.; Zappoli, S.; Andracchio, A.; Gelencser, A.; Kiss, G.; Krivacsy, Z.; Meszaros, E.; Hansson, H. C.; Alsberg, T.; Zebuhr, Y. J. Geophys. Res.-Atmos. 1999, 104, 26821-26832. 323. Facchini, M. C.; Mircea, M.; Fuzzi, S.; Charlson, R. J. Nature 1999, 401, 257-259. 324. Graham, B.; Mayol-Bracero, 0. L.; Guyon, P.; Roberts, G. C.; Decesari, S.; Facchini, M. C.; Artaxo, P.; Maenhaut, W.; Koll, P.; Andreae, M. 0. J. Geophys. Res. 2002, 107(D20), 8047, doi:8010.1029/2001JD000336. 325. Havers, N.; Burba, P.; Lambert, J.; Klockow, D. J. Atmos. Chem. 1998, 29, 45-54. 326. Mayol-Bracero, 0. L.; Guyon, P.; Graham, B.; Roberts, G.; Andreae, M. 0.; Decesari, S.; Facchini, M. C.; Fuzzi, S.; Artaxo, P. J. Geophys. Res.-Atmos. 2002, 107(D20), 8091, doi:8010.1029/2001JD000522. 327. Mochida, M.; Kitamori, Y.; Kawamura, K.; Nojiri, Y.; Suzuki, K. J. Geophys. Res.  2002, 107, 4325, doi:4310.1029/2001JD001278. 328. O'Dowd, C. D.; Facchini, M. C.; Cavalli, F.; Ceburnis, D.; Mircea, M.; Decesari, S.; Fuzzi, S.; Yoon, Y. J.; Putaud, J. P. Nature 2004, 431, 676-680. 329. Badger, C. L.; Griffiths, P. T.; George, I.; Abbatt, J. P. D.; Cox, R. A. J. Phys. Chem. A 2006, 110, 6986-6994. 330. McNeill, V. F.; Patterson, J.; Wolfe, G. M.; Thornton, J. A. Atmos. Chem. Phys.  2006, 6, 1635-1644. 331. Park, S. C.; Burden, D. K.; Nathanson, G. M. J. Phys. Chem. A 2007, 111, 29212929. 332. Thornton, J. A.; Abbatt, J. P. D. J. Phys. Chem. A 2005, 109, 10004-10012. 333. Cosman, L. M.; Knopf, D. A.; Bertram, A. K. 2007.  Chapter 5^  152  334. Knopf, D. A.; Cosman, L. M.; Mousavi, P.; Mokamati, S.; Bertram, A. K. J Phys. Chem. A 2007, 111, 11021-11032. 335. Cosman, L. M.; Bertram, A. K. 2007. 336. McNeill, V. F.; Wolfe, G. M.; Thornton, J. A. J. Phys. Chem. A 2007, 111, 10731083. 337. Thornton, J. A.; Braban, C. F.; Abbatt, J. P. D. Phys. Chem. Chem. Phys. 2003, 5, 4593-4603. 338. Saxena, P.; Hildemann, L. M. I Atmos. Chem. 1996, 24, 57-109. 339. Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. Atmos. Environ. A-Gen. 1993, 27, 1309-1330. 340. Schauer, J. J.; Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R. Atmos. Environ. 1996, 30, 3837-3855. 341. Kawamura, K.; Kaplan, I. R. Environmental Science & Technology 1987, 21, 105110. 342. Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simonelt, B. R. T. Environ. Sci. Technol. 1991, 25, 1112-1125. 343. Kawamura, K.; Kasukabe, H.; Barrie, L. A. Atmos. Environ. 1996, 30, 1709-1722. 344. Knopf, D. A.; Anthony, L. M.; Bertram, A. K. I Phys. Chem. A 2005, 109, 55795589. 345. Knopf, D. A.; Mak, J.; Gross, S.; Bertram, A. K. Geophys. Res. Lett. 2006, 33, L17816, doi:17810.11029/12006GL026884. 346. Adamson, A. W.; Gast, A. P. Physical chemistry of surfaces, 6th Ed. ed.; WileyInterscience: New York, 1997.  Chapter 5^ 347. NIMA Technology Ltd. 348. Huey, L. G.; Hanson, D. R.; Howard, C. J. J. Phys. Chem. 1995, 99, 5001-5008. 349. Hu, J. H.; Abbatt, J. P. D. J. Phys. Chem. A 1997, 101, 871-878. 350. Mason, E. A.; Monchick, L. J. Chem. Phys. 1962, 36, 2746-2757. 351. Monchick, L.; Mason, E. A. J. Chem. Phys. 1961, 35, 1676-1697. 352. Patrick, R.; Golden, D. M. Int. J. Chem. Kinet. 1983, 15, 1189-1227. 353. Hanson, D. R.; Ravishankara, A. R. J. Geophys. Res. 1991, 96, 5081-5090. 354. Hanson, D. R. J. Phys. Chem. A 1998, 102, 4794-4807. 355. Motz, H.; Wise, H. J. Chem. Phys. 1960, 32, 1893-1894. 356. Torn, R. D.; Nathanson, G. M. J. Phys. Chem. B 2002, 106, 8064-8069. 357. Calculations from the aerosol inorganics model at http://www.aim.env.uea.ac.uk/aim.html. 358. Carslaw, K. S.; Clegg, S. L.; Brimblecombe, P. J. Phys. Chem. 1995, 99, 1155711574. 359. Massucci, M.; Clegg, S. L.; Brimblecombe, P. J. Phys. Chem. A 1999, 103, 42094226. 360. Raduge, C.; Pflumio, V.; Shen, Y. R. Chem. Phys. Lett. 1997, 274, 140-144. 361. Wiles, L. A. J. Chem. Soc. 1953, 996-997. 362. Liler, M. Reaction Mechanisms in Sulphuric acid; Academic Press: New York, 1971. 363. Harkins, W. D.; Wampler, R. W.J. Am. Chem. Soc. 1937, 53, 850-859. 364. Donaldson, D. J.; Vaida, V. Chem. Rev. 2006, 106, 1445-1461. 365. Donaldson, D. J.; Anderson, D.J. Phys. Chem. A 1999, 103, 871-876.  153  Chapter 5^  154  366. Gurkov, T. D.; Dimitrova, D. T.; Marinova, K. G.; Bilke-Crause, C.; Gerber, C.; Ivanov, I. B. Colloid. Surface. A 2005, 261, 29-38. 367. Donaldson, D. J.; Guest, J. A.; Goh, M. C. J Phys. Chem. 1995, 99, 9313-9315. 368. Donaldson, D. J. J Phys. Chem. A 1999, 103, 62-70. 369. Barnes, G. T.; Quickenden, T. I.; Saylor, J. E. J. Colloid Interf Sci. 1970, 33, 236243. 370. Shulman, M. L.; Jacobson, M. C.; Carlson, R. J.; Synovec, R. E.; Young, T. E. Geophys. Res. Lett. 1996, 23, 277-280. 371. Tuckermann, R.; Cammenga, H. K. Atmos. Environ. 2004, 38, 6135-6138.  Chapter 6^  155  6. REACTIVE UPTAKE OF 03 BY MULTICOMPONENT AND MULTIPHASE MIXTURES CONTAINING OLEIC ACID  6.1 Introduction Aerosol particles have a significant impact on global climate by interacting directly with solar and terrestrial radiation, thereby influencing the global radiation budget. 372 Furthermore, they can modify the radiative properties of clouds, the conditions required for cloud formation, and the lifetimes of clouds, by acting as cloud condensation nuclei (CCN) and ice nuclei (IN). 372-380 Condensed-phase organic material is abundant throughout the atmosphere. 38 "86 In urban areas of the U. S., for example, organic material typically accounts for 10-40% of the fine particulate mass, and in rural and remote areas of the U. S., organic material typically accounts for 30-50% of the fine particulate mass. 387 Organic aerosols, while in the atmosphere, experience reactions with gas-phase species which potentially lead to the modification of the particle composition and morphology. These reactions, which are often referred to as heterogeneous reactions, are of importance for several reasons. They may form toxic or carcinogenic compounds; thus, they may be important for healthrelated issues. Heterogeneous reactions may also change the hygroscopic properties 388-391 and optical properties of organic particles, and therefore, influence the ability of these particles to act as CCN 392-395 and to scatter and absorb solar radiation. Additionally, heterogeneous reactions may change the ability of organic particles to act as IN.  396  In  short, these heterogeneous reactions may dictate the importance of organic particles in atmospheric chemistry, climate, and health related issues. Over the past few years, researchers have begun to focus on the heterogeneous chemistry of organic particles. 397-4°3 As a first step toward understanding this chemistry, researchers have investigated single-component organic particles and substrates as surrogates or proxies for atmospheric organic particles. Organic particles in the  A version of this chapter has been published. Knopf, D. A., Anthony, L. M., and Bertram, A. K., Reactive Uptake of 0 3 by Multicomponent and Multiphase Mixtures Containing Oleic Acid, J. Phys. Chem. A., 2005, 109, 5579-5589.  156  Chapter 6^  atmosphere, however, are most likely multicomponent and in some cases possibly multiphase mixtures. 404,405 At present, the heterogeneous chemistry of multicomponent and multiphase organic particles is basically unexplored with few exceptions.  406  Research  is needed on this topic in order to obtain a complete understanding of the heterogeneous chemistry of organic particles in the atmosphere. One specific organic heterogeneous reaction that has been extensively studied is the reaction of 03 and pure cis-9-octadecenoic acid (oleic acid). 400,407-413 Experiments using coated wall flow-tube reactors obtained reactive uptake coefficients (7) of 03 on pure oleic acid (OA) of around 8 x 10- 4 . 4 00, 407,410 In contrast, reactive uptake measurements employing pure OA aerosol particles derived values ranging from 7.5 x 10 4 to 9.8 x  10-3.406,408,409,412,413  reaction have also been investigated.  Gas-phase and condensed-phase products for this 406,407,409-413  In the atmosphere, OA is most likely found in mixtures, 404,405 and the other components of the mixtures may influence the reaction rates, reaction mechanisms, and atmospheric lifetime of OA. Field measurements together with estimated source fluxes suggest that OA has a lifetime in the atmosphere on the order of days.  404,408  In contrast,  the laboratory studies on pure OA imply the lifetime is a few minutes for typical particle diameters under polluted conditions. Studies of multicomponent organic aerosols are needed to better understand the lifetime and fate of OA in the atmosphere. 406 One major source of OA-containing particles is from meat cooking.  404,405,414  These aerosol particles may comprise up to 20% of the primary fine organic carbon particle emissions in the Los Angeles area. 404 Approximately 10-20 wt % of the mass of meat-cooking particles has been identified. The composition of the identified mass is 45% n-alkanoic acid, 30% n-alkenoic acids (which is mainly OA), 3% lactones, 2% nalkanes, and 3% amides. 404,405,414 A minor source of OA in the atmosphere is leaf abrasion. 415 In the following, we present a comprehensive study of the reaction between 0 3 and oleic acid/alkanoic acid mixtures. The purpose of this study is to gain a better understanding of reactions on multicomponent and multiphase organic particles, as well as to improve the understanding of the reactivity and lifetime of OA in the atmosphere. For these studies, we focused on myristic acid/oleic acid (MA/0A) mixtures. MA is a  Chapter 6^  157  saturated fatty acid, with 14 carbon atoms (C14H2802). MA is a significant component of aerosols from meat cooking. 404,405 In addition, MA/OA systems are well-suited for the proposed studies, as this system has a wide composition range over which the mixtures are completely liquid and also a wide composition range over which the mixtures form a solid in equilibrium with a liquid. 416 This allows us to probe systematically the effect of the phase, morphology, and OA composition on the reactivity. Additionally, we investigated the uptake on multicomponent mixtures that closely represent compositions of the identified mass of meat-cooking aerosols. 404,405 For the remainder of the document, we refer to these multicomponent mixtures as meat cooking mixtures.  6.2 Experimental Experiments were conducted in a rotating-wall flow-tube reactor coupled to a chemical ionization mass spectrometer (CIMS). 417421 The apparatus is illustrated in Figure 6.1. rotating glass tube substrate 03  ion quadrupole optics^10 Torr  -6  He  movable injector  cooling jacket  It  2  ion-molecule N /S F6 reaction region 2-3 Torr rotary turbo pump pump  1^i ll  II  channeltron turbo multiplier pump  Figure 6.1: Sketch of the rotating-wall flow-tube reactor coupled to the CIMS.  The rotating glass tube is 1.77 cm in diameter and fits snugly inside the flow tube, which is surrounded by a cooling jacket for temperature control. 03 enters the flow-tube reactor through a movable injector. It is important to note that in these experiments less than 3% of the OA in the substrates is oxidized, even for long exposure experiments.  Chapter 6^  158  (This was determined by comparing the total moles of OA in the substrates with the total moles of 0 3 taken up by the substrates.) This is in contrast to aerosol experiments where the particles are typically completely oxidized. This should be kept in mind when comparing our experiments with aerosol experiments and also when extrapolating our results to the atmosphere. The setup and procedure of the experiments are similar to Moise and Rudich 407 and Thornberry and Abbatt. 41° Ozone was generated by passing a flow of 02 over an ultraviolet source (Jelight, model #600) and then stored in a 5 L bulb. In the uptake experiments, the flow of the 02/03 mixture varied between 1 and 30 cm 3 min d at STP (standard temperature and pressure). The carrier gas in the experiments was He, and a flow rate of 30-100 cm 3 mind at STP was used. Under all conditions, the flow inside the reactor was laminar, based on the Reynold's number. All uptake experiments were conducted at a constant temperature of 298 K. 03 concentrations in the flow-tube reactor of approximately 2 x 10 12 to 4 x 10 12 molecules cm  -3  were used for the uptake  experiments if not indicated otherwise. An 03 concentration of 2 x 10 12 molecules cm -3 corresponds roughly to polluted atmospheric conditions. Typical pressures in the flow tube were between 2 and 3 Torr. 03 was detected as 03 in the mass spectrometer after its chemical ionization by -  SF6 - . SF6 was generated by passing a trace amount of SF6 in about 1000 STP cm 3 mind N2 through a 210 Po source (NRD, model Po-2031).  Shown in Figure 6.2 is the phase diagram for the MA/OA binary system. 416 According to the phase diagram, the fatty acids are immiscible in the solid phase and exist as separate solid phases below the eutectic temperatures. 416 The solid circles represent the concentrations we investigated for these systems. When the MA weight percent was less than 17 wt % for the MA/OA system the mixtures were completely liquid at room temperature. When the concentration of the saturated fatty acid was higher than this value, the binary solutions formed solid-liquid mixtures, where solid MA was in equilibrium with a liquid at room temperature, and the ratio of liquid to solid could be determined by the lever rule. These solid-liquid mixtures are also referred to as semisolid mixtures. 422-424  159  Chapter 6^ XMA 0 .1 .2 .3 .4 .5 .6 .7^.8^.9^1 ^ I  ^  I  ^ ^ ^ I I I  330  320 310  ED-  300  E  290 280 SOA(a) SMA SOA(i) SMA^•  270  -1-a.a.I.E.P.I...1.  260 0 10 20 30 40 50 60 70 80 90 100  Myristic Acid /wt% Figure 6.2: The phase diagram of MA/OA as a function of temperature and concentration. 416 S and L indicate the solid and liquid phases, respectively. xmA indicates the MA mole fraction. SoA(a) and SoA(y) represent two polymorphic forms of 0A. 416 The filled circles indicate the conditions of the conducted experiments.  Liquid substrates were prepared by depositing approximately 0.6 mL of the solution on the bottom surface of the inner glass tube. This tube was then rotated at 5-10 rotations per minute and held at 298 K, resulting in a smooth liquid substrate on the inside surface. Two different methods were used to prepare the solid-liquid mixtures in order to investigate the effect of substrate preparation technique on the observed uptake coefficients. In the first method, the solid-liquid mixtures were melted and dispersed on the inside of the flow tube, which was fixed at a temperature above the complete melting temperature of the mixture. Then, while the flow tube was rotated, the temperature was decreased to room temperature at a rate of approximately 1 K min -I by changing the temperature of the coolant that was circulated in the cooling jacket of the flow cell. While the temperature decreased, the substrate crystallized. For the remainder of the document, this will be referred to as slow cooling. In the second method, the solid liquid mixture  Chapter 6^  160  was completely melted and dispersed on the inside of a glass tube, which was preheated to above the complete melting temperature of the mixture. The glass tube was rotated to form a smooth liquid substrate on the inside of the glass. Then, the glass tube was rapidly immersed into liquid N2. A similar method has been used previously to prepare solid substrates for uptake experiments. 425 For the remainder of the document, this will be referred to as fast cooling. Subsequently, the tube was taken out of the liquid N2 and located inside the flow-tube reactor and held at 298 K for the uptake measurements. To investigate the physical microstructure of the substrates we used optical microscopy. The substrates were prepared on microscope slides in order to view the substrates with an optical microscope. The methods of preparing these substrates were the same as discussed above. For all the solid-liquid substrates that we prepared, the substrates had a very high viscosity and appeared as a solid or a wax, even when the mixture contained a significant amount of liquid based on the lever rule. We did not record high quality digital images of the MA/OA substrates, but in previous experiments we did record high quality digital images of lauric acid/oleic acid (LA/OA) solid-liquid substrates. The physical properties of these substrates are very similar to the MA/OA substrates so we show them here to help illustrate the physical properties of the MA/OA substrates. Figure 6.3 shows images recorded with a digital camera of a 51 wt % LA/OA solid-liquid substrate prepared by slow (panel (a)) and fast (panel (b)) cooling. As mentioned, for all the solid-liquid substrates that we prepared, the substrates had a very high viscosity and appeared as a solid or a wax, even when the mixture contained a significant amount of liquid based on the lever rule. The high viscosity of the solid-liquid mixtures is probably due to the microstructure of the solid in the solid-liquid system, since the rheology of solid-liquid mixtures not only depends on the solid fraction but also is highly correlated with the degree of aggregation and the interconnected network of the solid crystals in the solid-liquid mixture. 422-424,426-428 For example, previous results obtained on semisolid alloys exhibit large viscosities even at relatively low solid fractions because of the dendritic structure of the solid. 423,424,429 In all cases, the slow-cooled substrates were rougher and appeared to have larger crystals (panel (c)) than the fast cooled substrates (panel (d)), as illustrated in Figure 6.3.  Chapter 6  ^  161  Figure 6.3: Panels (a) and (b) show a substrate composed of 51 wt % LA/OA prepared by slow and fast cooling, respectively. Panels (c) and (d) show the same substrates as (a) and (b) but with a higher magnification. Panels (a) and (b) were obtained with a digital camera and (c) and (d) were obtained with an optical microscope.  Formation of a solid-liquid mixture from a liquid solution involves nucleation and then crystal growth. The larger crystals produced by slow cooling (panel (c)) compared to fast cooling (panel (d)) can be explained by different rates of nucleation in the two methods. For slow cooling, only a few nucleation events probably occur followed by crystal growth. For rapid cooling, many nucleation events may occur, followed by crystal growth, resulting in smaller but more numerous crystals. As mentioned above, in addition to studying the MA/OA systems, we studied mixtures that closely represent the composition of meat-cooking mixtures. These mixtures consisted of up to 15 different substances that correspond to the most abundant identified components of particles produced from meat-cooking processes (see below for more details). We also prepared these substrates using both the slow and fast cooling processes described above, and the substrates also appeared as solids or waxes. We estimated the topography of the solid-liquid substrates using optical microscopy. First, we calibrated the z scale of our optical microscope using a gauge of  Chapter 6^  162  known thickness. Then, we measured the heights of the peaks and valleys of the solidliquid substrates with our microscope, while we scanned across the surfaces of the substrates. From this information, we estimate that the surface area of the slow- and fastcooled substrates deviated from the surface area of the glass substrate by 3-8%. In all cases, we have calculated y using the geometric surface area. Therefore, we overestimate y by at most 8%, which is a small uncertainty compared to the overall uncertainty in the  uptake measurements for solid-liquid mixtures. 6.2.1 Chemicals All gases were purchased from Praxair, and solid and liquid chemicals were purchased from Aldrich. Listed below are the chemicals and the corresponding purities used in our studies: N2 (99.999%), He (99.999%), SF6 (99.995%), 02 (99.993%), OA (99+%), MA (99.5+%), lauric acid (99%), palmitic acid (99%), stearic acid (98+%), succinic acid (99.8%), glutaric acid (99%), adipic acid (99+%), palmitoleic acid (98%), tetracosane (99%), pentacosane (99%), 2-pentadecanone (98+%), 2-hexadecanone (98%), 2-octadecanone (99%), hexadecanamide (purity not verified).  6.3 Results and Discussion 6.3.1 Kinetics Experiments The reactive uptake coefficients were determined from the irreversible removal of 0 3 as a function of injector position. Figure 6.4 shows the natural logarithm of this 0 3 signal as a function of reaction time (determined from the average flow velocity) for various mixtures. The data for each uptake experiment was fit to a straight line, and the observed first-order loss rate, kobs, was determined from the slope.  Chapter 6^  0.00 0.0  0.02  163  0.04  0.06  0.08 0.0  —0.2  —0.2  —0.4  —0.4  oc° -0 .6  —0.6  —0.8  —0.8  cj  • C7)  0.00  ^  0.02  ^  0.04^0.06  ^  0.08  Time / s Figure 6.4: Natural logarithm of the observed 0 3 signal as a function of reaction time. The filled circles and filled triangles correspond to uptake on pure OA and on a 16 wt % MA/OA solution, respectively. The filled and open diamonds represent the data obtained from solid-liquid 26 wt % MA/OA mixtures that have been prepared by slow and fast cooling, respectively. The lines represent the corresponding linear fits to the data.  The first-order wall loss rate,  kwall, was  calculated from ko b s using the iterative  numerical method of Brown, 43° which corrects for concentration gradients that form close to the flow-tube wall because of uptake at the organic surface. This results in a correction of less than 7% in all cases. The gas-phase diffusion constant of 03 in He was taken as 394 Torr cm2 s -1 at 298 K. 431 '432 The reactive uptake coefficient, 7, was obtained using the equation:433'434  Chapter 6 6.1  164 7=  2 rk wall  c o, + rk wall  where r is the flow-tube radius and co, is the mean molecular velocity of 03. The y values reported in this study are derived from at least three reactive uptake measurements. The reactive uptake coefficients represent the mean values of the performed uptake measurements, and the error represents two times the standard error of the mean. 6.3.2 Reactive Uptake Coefficients of 03 on OA  The reactive uptake coefficients derived in this study for liquid and crystallized OA are given in Table 6.1. The uptake of 03 on liquid OA at room temperature is in agreement with the previous flow-tube reactor studies by Moise and Rudich 407 and Thornberry and Abbatt. 41° Reactive uptake measurements employing pure OA aerosol particles and aerosol mass spectrometry obtained reactive uptake coefficients ranging from 7.5 x 10 -4 to 9.8 x 10 -3 , and in general, the y values measured with these techniques are higher than those measured with the coated-wall flow-tube technique. A recent study by Hearn et a1:113 suggests that most of these apparent discrepancies can be explained by secondary reactions between OA and the Criegee intermediate, 435 which results from the 03 reaction. The coated-wall technique measures the loss of 03, whereas the aerosol measurements determine the rate of loss of OA, and therefore includes the reaction between Criegee intermediates and OA. Hearn et al. 413 estimated that approximately 36% of OA loss is due to the reaction with the Criegee intermediate, and when this value was taken into account, a reactive uptake coefficient of 8.8 x 10 4 was obtained, which is in good agreement with the flow-tube measurements. More details on the discrepancies between the aerosol studies and the flow-tube studies are given in Hearn et al.413  Chapter 6  165  Table 6.1: Reactive Uptake Coefficients, y, for 03 on Pure Liquid and Solid OA  7 (liquid) a  y (solid)"  x10 4  x104  this study  7.9 ± 0.3  Moise and Rudich407  0.64 ± 0.05  8.3 ± 0.2  0.52 ± 0.01  Thornberry and Abbatt 4I°  8±1  Morris et a1.408  16 ± 2 c  Smith et al.409 Hearn and Smith412  58-98 C 7.5 ± 1.2 c  Hearn et al.413  13.8 ± 0.6 c  Ziemann406  13 ± 2  a  Values for y (liquid) were obtained at temperatures between 286 and 298 K. Values for 7 (solid) were obtained at temperatures between 278 and 285 K. c For these studies, y values were calculated by assuming that the loss of 0 3 to OA is 1:1. The recent study by Hearn et a1. 413 shows that in pure OA particles the ratio is actually 1:1.36 because of the reaction between OA and the Criegee intermediate. 435 b  Table 6.1 also shows the reactive uptake coefficient obtained from crystallized OA substrates. The 2/value is approximately 1 order of magnitude lower than for the case of liquid OA. This clearly shows that the physical state of OA has a significant impact on the reactive uptake as shown previously by Moise and Rudich. 407 The small difference between the 7 value derived by Moise and Rudich 407 and this study may be attributed to different surface areas or structures of the substrates. 6.3.3 Reactive Uptake Coefficients of 03 on Liquid Binary Oleic Acid/Myristic Acid Mixtures  Figure 6.5 shows the resulting reactive uptake coefficients for MA/OA solutions, as well as the uptake coefficient for pure OA. As the concentration of MA is increased, the reactive uptake coefficient decreases as expected. However, the change in y does not seem to be a continuous function. When a small amount of MA is added to OA, there appears to be a step change in the uptake coefficient, after which, the uptake follows a continuous function. This step change is not large, but it appears to be bigger than the uncertainty in the measurements. (Note that each data point for MA/OA concentrations  Chapter 6^  166  between 0 and 4 wt % were derived from six separate measurements in order to reduce the uncertainty in the reported values.)  1 E-3  0^2^4^6^8^10 12 14 16 18 '1^l'I^I^I^1'1'  9E-4 8E-4  7E-4  9E-4  4  8E-4  + . • TI ,T •^•  6E-4  5E-4  1 E-3  7E-4  6E-4  I.I.I  I^I^IgI^I^I  5E-4 0^2^4^6^8^10 12 14 16 18  Mrystic Acid / wt %  Figure 6.5: Experimentally derived reactive uptake coefficient, y, as a function of MA/OA concentration conducted at 298 K.  Assuming that the uptake of 03 is dominated by reaction in the bulk phase,  y can  be represented by the following equation: 436 '437 ^— 6.2  =  1^1^co,  +  y a 411RTVD03 k 2 [0A]  where a is the mass accommodation coefficient, H is the Henry's law solubility constant of 0 3 in the mixture, R is the gas constant, T is the temperature, Do, is the diffusion constant of 03 in the mixture, k2 is the second-order rate coefficient for reaction in the condensed phase, and [OA] is the concentration of OA in the liquid. See appendix B for a further discussion of this equation.  ▪ 167  Chapter 6^  According to eq 6.2, a plot of 1/ y versus 1/-‘1[0,4] should be linear if a, H, D03 , and k2 do not change significantly with the addition of MA. Figure 6.6 shows a linear fit to all points except the data point of pure OA. Also shown are the 95% confidence intervals for the fit. Pure OA falls outside this 95% confidence interval. A possible explanation for the trend in the data is that a, H, D03 , or k2 decrease significantly with the addition of a small amount of MA.  1800  0.56^0.57^0.58^0.59^0.60^0.61^0.62 ^^1800  1600  - 1600  1400  - 1400  1200  - 1200  1000 ^'^I^I^I^I^I^ 1000 0.56^0.57^0.58^0.59^0.60^0.61^0.62 -1/2 ) Oleic Acid / (mol L -1 -1/2  Figure 6.6: Plot of inverse reactive uptake coefficient vs. inverse square root of OA concentrations for MA/OA solutions. Solid line indicates a linear fit to the data points disregarding the data point of pure OA at 0.563 (mol L ) I/2 Dotted lines represent the 95% confidence interval of the linear fit. .  Equation 6.2 assumes that the reaction occurs in the bulk and that the rate of reaction is limited by diffusion of 03 in the liquid mixtures. Experimental data from several studies suggest that the reaction between 03 and pure OA can be described by this modeL4o6-410,412 Recently, however, Hearn et al. 413 suggested (on the basis of experimental data) that the reaction occurs solely at the interface. Assuming that the  Chapter 6^  168  uptake is governed solely by a surface reaction, 7 can be represented by the following equation: 438 ' 439 ^co3  1. 1 + — 7 S 4k; I/ s RTK s [OA]  6.3  where S is the adsorption coefficient of 0 3 on the surface, Hs is the Henry's law solubility constant of 03 in the mixture at the surface, k2 s is the second-order rate coefficient for reaction at the surface, and Ks is an equilibrium constant linking the surface concentration to the bulk concentration [OA]. According to eq 6.3, a plot of 1/ 7 versus 1/[0A] should be linear if S, Hs, Ks, and k2 s do not change significantly with the addition of MA. Figure 6.7 shows a plot of 1/ 7 versus 1/[0A]. All data points are included in the linear least-squares analysis except the data point of pure OA.  0.31 0.32 0.33 0.34 0.35 0.36 0.37 0.38 0.39 1800 1800  7  1600  - 1600  1400  1400  1200  1200  --  1000 ^.^1 ^I^•^I •^ 1000 0.31 0.32 0.33 0.34 0.35 0.36 0.37 0.38 0.39 Oleic Acid ' / (mol L -1 -  )  -1  Figure 6.7: Plot of inverse reactive uptake coefficient vs. inverse OA concentrations for MA/OA solutions. Solid line indicates a linear fit to the data points disregarding the data point of pure OA at 0.317 (mol L -1 )-1 . Dotted lines represent the 95% confidence interval of the linear fit.  Chapter 6^  169  Similar to Figure 6.6, the data point corresponding to pure OA falls outside the 95% confidence interval for liquid MA/OA mixtures. We conclude that our complete data set cannot be explained by a surface reaction, assuming S, Hs, Ks, and k2 5 do not change with composition. Also, note that the quality of the linear fits shown in Figure 6.6 and Figure 6.7 are similar. Hence, we cannot determine from our data if eq 6.2 or eq 6.3 is more appropriate for the description of the reactive uptake of 03 by liquid MA/OA mixtures. Further studies are needed to understand this behavior on a molecular level. 6.3.4 Reactive Uptake Coefficients of 03 on Solid-Liquid Mixtures Figure 6.8 shows the 7 values for the solid-liquid MA/OA mixtures. Additionally, the y values obtained from the binary liquid solutions are plotted for comparison. The gray shading indicates the concentration over which this mixture is liquid.  40  60  80  100 1 E-3  5E-4  1 E-4  20^40^60^80  ^  2E-5 100  Mrystic Acid / wt %  Figure 6.8: Experimentally derived reactive uptake coefficient, y, as a function of MA/OA concentration. The open circles correspond to data obtained from liquid MA/OA solutions shown in Figure 6.6. The filled circles indicate experiments in which the mixture was cooled slowly, and the filled triangles indicate experiments in which the mixture was cooled rapidly. The gray shaded area indicates the liquid-phase concentration range.  Chapter 6^  170  This figure clearly shows that a small amount of solid saturated fatty acid can decrease the reactivity by an order of magnitude compared to the liquid mixtures. For example, when the slow-cooled MA/OA mixture contains only about 7% solid by mass, 7 decreases by 1 order of magnitude compared to the liquid solution. The magnitude of the decrease in 7 is comparable to the magnitude of the decrease observed by Moise and Rudich407 when going from pure liquid oleic acid to pure solid oleic acid. This behavior is most likely due to the microstructure of the solid-liquid mixtures as discussed above. The solid can form an interconnected network by dendritic growth, aggregation of crystals, or eutectic solidification, which can result in lamellar or rod microstructures. 424,426,427,440  422-  An interconnected network may efficiently reduce the effective diffusion of  03 and OA in the mixture, leading to a decrease in the uptake coefficient. Furthermore, the interconnected network may "trap" some of the OA, making it inaccessible for reaction. "Trapping" of a liquid or solvent during crystallization is a frequent problem in industrial crystallization and is associated with rapid crystallization and dendritic growth:44i In general, the slow-cooled solid-liquid mixtures exhibit smaller reactive uptake coefficients than the fast-cooled ones. This is also most likely due to the microstructure of the solid. As shown in Figure 6.3, the different methods of preparing the substrates produce different microstructures of the solids. The microstructure produced by slow cooling appears to reduce diffusion in the mixtures or "trap" OA more efficiently than the microstructure produced by fast cooling. Reactive uptake experiments were also performed on solid MA substrates. Within the experimental uncertainty, no uptake of 03 occurred. As mentioned above, the flow tube was rotated at 5-10 rpm in most experiments. We performed additional reactive uptake experiments on solid-liquid mixtures while the flow tube was not rotated, obtaining the same results as when the flow tube was rotated. Therefore, we conclude that the rotation has no significant effect on our uptake measurements. We have also carried out reactive uptake measurements on solid-liquid mixtures with smaller 03 concentrations of about 7 x 10 10 molecules cm 3 . Within our experimental uncertainty, we obtained the same results as determined with higher concentrations of ozone (2 x 1 0 x 10 12 molecules cm-3).  12  -4  Chapter 6^  171  6.3.5 Reactive Uptake Coefficients of 03 on Solid-Liquid Mixtures That Closely Represent the Composition of Meat-Cooking Aerosols For these studies, the experimental procedure and the data analysis are similar to that of OA and OA/MA mixtures. As mentioned above, these mixtures are composed of up to 15 different substances that correspond to the most abundant identified components of particles produced from meat cooking. 404 The phase behavior for these multicomponent systems is not well-understood. On the basis of the enthalpies of fusion of the different components and the van't Hoff equation, 441 we speculate that the substrates are solidliquid mixtures and the mixtures contain solid palmitic acid and solid stearic acid, whereas OA remains in the liquid state. This assumes that the mixtures behave as ideal solutions and that no mixed phases precipitate, and OA is not soluble in any solids that precipitate in the mixtures. Further work is needed to verify these assumptions. As determined by optical microscopy, the substrates appear as a solid or a wax similar to the MA/OA substrates. Table 6.2 gives the experimentally obtained y values from various mixtures (see notes  442-444  for the specific compositions). The obtained reactive uptake coefficients from  these solid-liquid meat-cooking mixtures range from 1.6 x 10 -5 to 6.9 x 10 -5 . Mix 3 444 consists of 15 different substances and, therefore, may represent the most realistic organic mixture found in aerosol particles with respect to the identified mass.  Chapter 6^  172  Table 6.2: Obtained Reactive Uptake Coefficients and the Corresponding OA Lifetimes Given for Different Fast- and Slow-Cooled Solid-Liquid Mixtures Which Closely Represent Meat-Cooking Aerosols' mix 1 42  mix 2 443^mix 3 444  n-alkanoic acids/wt%  40.0%  38.1%  37.0%  n-alkenoic acids/wt%  60.0%  49.8%  50.4%  dicarboxylic acids/wt %  2.8%  2.4%  n-alkanes/wt%  8.0%  2.6%  n-alkanones/wt%  6.0%  amides/wt%  1.3%  7 (fast cooling)  (6.9±1.2) x 10 -5  7 (slow cooling)  (4.3±0.7) x 1 0-5  1.6% (3.3±0.5) x 10 -5  (5.2±1.1) x 10 -5  r (fast cooling)/min^21.2^  (1.6±0.3) x 10 -5 36.1  r (slow cooling)/min^38.3^28.6^74.5 aThe  lifetime, T, is derived for 100 ppb 0 3 and particles diameters of 0.2pm.  6.3.6 Reactive Uptake Coefficients as a Function of Substrate Age During the course of our experiments, we also discovered that the reactive uptake on semisolid mixtures increased after the substrates were first generated and eventually reached a maximum in y after approximately 10 h. Shown in Figure 6.9 (solid triangles) are results from measurements of y on a fast-cooled 35 wt % MA/OA solution as a function of substrate age.  173  Chapter 6^  1 E-3  0  100  200  300  400  500  5E-4  1 E-4  2E-5  0^100^200^300^400  ^  500  Film Age / min Figure 6.9: The 03 uptake coefficients as a function of substrate age are presented for a fast-cooled 35 wt % MA/OA solution (triangles) and liquid OA (squares). The dotted lines are plotted to guide the eye and have no physical meaning.  In this experiment, the substrate was prepared and then an uptake measurement was performed approximately every 1 h. Between measurements the substrates were not exposed to ozone, but rather, they were only exposed to He gas. A similar trend in 7 was observed for slow-cooled binary mixtures as well as for slow and fast-cooled meatcooking mixtures. We have also performed experiments where the substrates were exposed to 03 continuously, and similar results were obtained. We conclude from this that the change of y did not depend on 0 3 exposure, but rather on the age of the substrate from which it was first prepared. For comparison, we have also included in Figure 6.9 the change in 7 on OA as a function of the liquid substrate age. (In this experiment, the  Chapter 6^  174  substrate was continuously exposed to 03.) As expected, the reactive uptake on liquids did not change with substrate age. The increase in uptake with age was only observed for semisolid mixtures. Also, note that in our long-term experiments less than 3% of the OA was consumed, in contrast to most aerosol experiments where the particle is completely oxidized. The increase in the uptake coefficient with time for the solid-liquid mixtures can be explained either by Ostwald's ripening of the crystal microstructure428,445-448 or by the formation of a non-equilibrium phase that relaxes to the stable phase. 416 The first process, Ostwald's ripening, is also referred to as particle coarsening and can be an important aging process for a solid that remains in equilibrium with the liquid. 448 This process refers to the change in the solid microstructure, driven by the tendency to minimize the total surface free energy of the solid in equilibrium with a liquid. 445-448 For example, after the solidification of a dendritic structure, the solid microstructure becomes gradually coarser as a result of the remelting of dendrite arms of smaller radius.  424,448  As mentioned  above, crystallization can lead to an interconnected network of solid crystals in the solidliquid mixture by several processes including dendritic growth, aggregation, and eutectic 422424,426428,440,441 solidification. In general, the ripening process may decrease the  connections in the solid network, which would lead to an increase of the diffusion of 03 and OA in the solid-liquid mixture and would result in a faster uptake coefficient. The second process, the formation of a non-equilibrium phase followed by its relaxation to the stable phase, is based on the experimental findings of Inoue et al. 416 These authors suggest that when a binary MA/OA solution crystallizes it first forms a non-equilibrium solid in which OA is partially soluble. Following crystallization, the solid phase undergoes phase separation or so-called demixing as it relaxes toward its equilibrium state where OA is not miscible in MA. This process would make more OA accessible for reaction; hence, it should lead to an increase in y with time.  Chapter 6^  175  6.4 Atmospheric Implications 6.4.1 General Atmospheric Implications We have shown that the uptake of 0 3 on solid-liquid mixtures can be significantly smaller than the uptake on the liquid mixtures even when the solid-liquid mixture contains only a small amount of solid. We have also shown that the uptake depends strongly on the method of generating the solid-liquid mixtures. Additionally, the heterogeneous uptake coefficient increases with the age of the semisolid mixture (when the amount of the substrate that is oxidized is small). These findings may be important for meat-cooking aerosols and also may be important for other aerosols in the atmosphere. For example, at low relative humidities, inorganics, such as (NH4)2SO4, may crystallize in aqueous inorganic-organic solution droplets to form solid-liquid particles. 449-451 The resulting particles may consist of an interconnected network of solid crystals in equilibrium with a liquid. This microstructure will likely influence the reactivity of these particles similar to the organic mixtures discussed above. The microstructure clearly plays an important role for the reactivity of the particles. As mentioned above, we have shown that the uptake coefficient increases with the age of the semisolid mixture when the amount of the substrate that is oxidized is small. On the basis of our measurements, y increases by approximately a factor of 2 after 10 h for semisolid substrates. The y values reported in Figure 6.8 and Table 6.2, which were obtained from substrates less than 1.5 h in age, can be used to estimate the uptake of ozone on unoxidized particles less than approximately 2 h in age. For particles that are greater than approximately 2 h in age, the increase in y due to Ostwald's ripening should be considered. 6.4.2 Atmospheric Lifetime Estimates of Oleic Acid Reactive uptake coefficients for 03 on OA suggest an atmospheric lifetime of OA on the order of seconds to minutes assuming 03 concentrations of about 100 ppb 452 '453 and submicron particles. 454 However, field measurements together with source fluxes suggest lifetimes on the order of days. 404,408 Morris et al:408 and Worsnop et al. 439 speculated that the long lifetime of OA in the atmosphere is because particles in the  Chapter 6^  176  atmosphere that contain OA are likely highly viscous and semisolid. Here, we estimate the atmospheric lifetime (which is the time it takes for the concentration of OA to decrease to 1/e of its initial value) using the y values we derived for binary and meatcooking mixtures. In all cases, we applied y values which have been obtained from substrates less than 1.5 h in age. The increase in y with substrate age as shown above has not been taken into account for these lifetime estimates. As mentioned, y increases by approximately a factor of 2 after 10 h for solid-liquid substrates. The lifetimes we calculated from our measurements (see below) are at most approximately 2 h. Hence, it is more appropriate to use y values determined from substrates less than 2 h in age rather than uptake coefficients determined from substrates 10 h in age for these calculations. To calculate the lifetime, r, of OA, we first determine HVD03 k2 from our uptake measurements using eq 6.2 (assuming that a »y). Then, we use HVD0, 1c2 in the following equation to predict the lifetime of OA in submicron particles  400,408,409  ^k 2 ^ 3P03 6.4^1/[49A] = ii[Oi] o^" t Zr  where P0, is the 0 3 partial pressure, t is the time the particle was exposed to 03, r is the radius of the particle, and [0A]3 is the initial OA concentration. Note that eqs 6.2 and 6.4 assume that diffusion of OA in the mixture is fast. This is most likely valid for the liquid substrates but may not be appropriate for solid-liquid mixtures. Nevertheless, we use these equations to estimate r for solid-liquid mixtures in the absence of a model to describe the uptake in these mixtures. Also, note that this calculation assumes a 1:1 loss of 03 to OA. The recent study by Hearn et al. 413 shows that in pure OA particles the ratio is actually 1:1.36 because of the reaction between OA and the Criegee intermediate. 435 However, the ratio is expected to be closer to 1:1 in MA/OA mixtures because of the lower concentration of OA in these solutions. From eqs 6.2 and 6.4, it follows that our assumption of 1:1 will lead at most to an overestimation of the lifetime by 36%. Figure 6.10 shows the derived lifetimes of aerosol particles 0.2 pm in diameter exposed to 100 ppb 03 consisting of different MA/OA compositions. The lifetime, z, of a pure OA particle is also plotted. The z values range between 4.7 and 5 min for the liquid  177  Chapter 6^  MA/OA solutions. This shows that the lifetime of OA is not significantly modified by the addition of a second component as long as the solution remains liquid. For the solidliquid MA/OA mixtures, the lifetime can change by more than an order of magnitude because of the microstructure of the solid-liquid mixtures. Lifetimes of up to 71 min are predicted for these binary solid-liquid particles.  0  20  40  60  80  100  80  80  70  70  60  60  -. .^50  50  - —. a) 40  40  E  t:T5 30  30  20  20  10  10  .J  0  '^  0  20^40^60^80  0 100  Myristic Acid / wt % Figure 6.10: OA lifetimes obtained at 298 K under typical polluted environments (100 ppb 03) for MA/OA particles, 0.2 ,um in diameter. The open circles correspond to particles in the liquid state. The solid triangles and solid circles indicate particles in the solid-liquid state which have been cooled rapidly and slowly, respectively. The gray shaded area indicates the liquid-phase concentration range.  Table 6.2 shows derived lifetimes which correspond to more complex meatcooking mixtures. Lifetimes of up to 75 min are derived for mixtures containing up to 15  Chapter 6^  178  substances. The influence of the different temperature history on the lifetime is clearly shown. The lifetime of OA in these mixtures increased by up to a factor of 16 compared to mixtures composed of liquid 0A/MA solutions. Recently, Ziemann 4°6 also showed that lifetimes of OA in a solid-liquid mixture can be significantly longer than for pure OA. The lifetimes above were determined with eqs 6.2 and 6.4, which assume that the reaction occurs in the bulk and that the reaction rate is limited by diffusion of 03 in the mixtures. We have also calculated the lifetimes assuming that the reaction occurs at the surface of the particles, using eq 6.3 shown above and eq 16 from Worsnop et al. 439 In this case, all the calculated lifetimes were approximately 25% longer than the lifetimes calculated with the assumption that the reaction occurs in the bulk and is limited by diffusion of 03 in the mixtures. Our study shows how the physical state and morphology of the organic mixture can influence the reactivity and, as a consequence, alter the lifetime of OA. However, the lifetimes determined in our studies for OA may be lower limits to the lifetime of OA in the atmosphere for the following reasons: First, we have only studied uptake on the identified composition of meat-cooking aerosols, which comprise 10-20% of the total aerosol mass. 404,405 The remaining 80-90% of the mass may decrease the uptake coefficient and increase the lifetime of OA. Second, Ziemann 4°6 recently showed that the reactivity of solid-liquid aerosol particles consisting of palmitic acid/oleic acid and margaric acid/oleic acid decrease more than expected after approximately 50% of the OA was oxidized. In other words, two kinetic regimes were observed. As mentioned previously, less than 3% of the OA was oxidized in our experiments. If the reaction in the meat-cooking mixtures does slow down more than expected after a large fraction of the OA is oxidized, then our analysis will underestimate the lifetime of OA in the atmosphere.  6.5 Conclusions and Summary The obtained reactive uptake coefficients for the binary liquid MA/OA solutions range between 6.5 x 10 -4 and 7.2 x 10 -4 . This entire data set cannot be explained by assuming that the reaction occurs exclusively in the bulk and that a, H, D03 , and k2 do  Chapter 6^  179  not change with composition. Furthermore, the complete data set cannot be explained by assuming that the reaction occurs at the surface and that S, Hs, Ks, and k2 5 do not change with composition. The uptake coefficients derived from binary solid-liquid mixtures range between 3.4 x 10 -5 and 1.7 x 10 -4 , which is approximately 1 order of magnitude lower compared to the liquid solutions. The obtained reactive uptake coefficients for multicomponent mixtures that closely represent compositions of meat-cooking aerosols range from 1.6 x 10 -5 to 6.9 x 10 -5 . We have shown that the reactive uptake coefficient is strongly dependent on the method of preparing the solid-liquid mixtures. This is most likely because the different preparation methods lead to different microstructures of the solid in the mixtures. This needs to be considered when modeling the lifetime and chemical processing of atmospheric particles. Also, this should be kept in mind when comparing laboratory results that use different preparation methods of substrates and particles. In some cases, a direct comparison between laboratory results may not be possible. We observed an unexpected increase in the reactivity with time for solid-liquid OA-containing mixtures. This has been attributed to Ostwald's ripening or the relaxation of a non-equilibrium phase to the stable phase. This also needs to be considered when comparing laboratory results and when extrapolating to the atmosphere. The OA lifetimes obtained for semisolid OA mixtures are up to a factor of 16 higher than for liquid mixtures. Lifetimes up to 75 min were obtained for these mixtures.  Chapter 6^  180  6.6 References 372. Climate Change 2001: The Scientific Basis. Contribution of Working Group 1 to the Third Assessment Report of the Intergovernmental Panel on Climate Change; Cambridge University Press: Cambridge, United Kingdom, 2001. 373. Albrecht, B. A. Science 1989, 245, 1227-1230. 374. Baker, M. B. Science 1997, 276, 1072-1078. 375. Pincus, R.; Baker, M. B. Nature 1994, 372, 250-252. 376. Twomey, S. Atmos. Environ. 1974, 8, 1251-1256. 377. Twomey, S. Atmos. Environ. A-Gen. 1991, 25, 2435-2442. 378. Pruppacher, H. R.; Klett, J. D. Microphysics of Clouds and Precipitation; Kluwer Academic Publishers: Dordrecht, The Netherlands, 1997. 379. Ramanathan, V.; Crutzen, P. J.; Kiehl, J. T.; Rosenfeld, D. Science 2001, 294, 21192124. 380. Rosenfeld, D. Science 2000, 287, 1793-1796. 381. Middlebrook, A. M.; Murphy, D. M.; Thomson, D. S. J. Geophys. Res. Atmos. 1998, 103, 16475-16483. 382. Murphy, D. M.; Thomson, D. S.; Mahoney, T. M. J. Science 1998, 282, 1664-1669. 383. Simoneit, B. R. T.; Cardoso, J. N.; Robinson, N. Chemosphere 1991, 23, 447-465. 384. Simoneit, B. R. T.; Kobayashi, A.; Kawamura, K.; Mochida, M. Geochim. Cosmochim. Acta 2003, 67, A437-A437. 385. Simoneit, B. R. T.; Sheng, G. Y.; Chen, X. J.; Fu, J. M.; Zhang, J.; Xu, Y. P. Atmos. Environ. A-Gen. 1991, 25, 2111-2129.  Chapter 6^  181  386. Tervahattu, H.; Juhanoja, J.; Vaida, V.; Tuck, A. F.; Niemi, J. V.; Kupiainen, K.; Kulmala, M.; Vehkamaki, H. J Geophys. Res.-Atmos. 2005, 110, D06207, doi:06210.01029/02004JD005400. 387. Air Quality Criteria for Particulate Matter; United States Environmental Protection Agency: Washington, D. C., 1996. 388. Cruz, C. N.; Pandis, S. N. Environ. Sci. Technol. 2000, 34, 4313-4319. 389. Dick, W. D.; Saxena, P.; McMurry, P. H. I Geophys. Res.-Atmos. 2000, 105, 14711479. 390. Ming, Y.; Russell, L. M. AIChE J. 2002, 48, 1331-1348. 391. Saxena, P.; Hildemann, L. M.; Mcmurry, P. H.; Seinfeld, J. H. I Geophys. Res.Atmos. 1995, 100, 18755-18770. 392. Asad, A.; Mmereki, B. T.; Donaldson, D. J. Atmos. Chem. Phys. 2004, 4, 20832089. 393. Broekhuizen, K.; Kumar, P. P.; Abbatt, J. P. D. Geophys. Res. Lett. 2004, 31, L01107. 394. Facchini, M. C.; Mircea, M.; Fuzzi, S.; Charlson, R. J. Nature 1999, 401, 257-259. 395. Novakov, T.; Penner, J. E. Nature 1993, 365, 823-826. 396. Gorbunov, B.; Baklanov, A.; Kakutkina, N.; Windsor, H. L.; Toumi, R. I Aerosol Sci. 2001, 32, 199-215. 397. Ellison, G. B.; Tuck, A. F.; Vaida, V. I Geophys. Res.-Atmos. 1999, 104, 1163311641. 398. Wadia, Y.; Tobias, D. J.; Stafford, R.; Finlayson-Pitts, B. J. Langmuir 2000, 16, 9321-9330.  Chapter 6^  182  399. Moise, T.; Talukdar, R. K.; Frost, G. J.; Fox, R. W.; Rudich, Y. I Geophys. Res.Atmos. 2002, 107, 223-242. 400. Rudich, Y. Chem. Rev. 2003, 103, 5097-5124. 401. Finlayson-Pitts, B. J.; Wingen, L. M.; Sumner, A. L.; Syomin, D.; Ramazan, K. A. Phys. Chem. Chem. Phys. 2003, 5, 223-242. 402. Eliason, T. L.; Gilman, J. B.; Vaida, V. Atmos. Environ. 2004, 38, 1367-1378. 403. Dubowski, Y.; Vieceli, J.; Tobias, D. J.; Gomez, A.; Lin, A.; Nizkorodov, S. A.; McIntire, T. M.; Finlayson-Pitts, B. J. J Phys. Chem. A 2004, 108, 10473-10485. 404. Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simonelt, B. R. T. Environ. Sci. Technol. 1991, 25, 1112-1125. 405. Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T. Environ. Sci. Technol.  1999, 33, 1566-1577. 406. Ziemann, P. J. Faraday Discuss. 2005, 130, 469-490. 407. Moise, T.; Rudich, Y. J. Phys. Chem. A 2002, 106, 6469-6476. 408. Morris, J. W.; Davidovits, P.; Jayne, J. T.; Jimenez, J. L.; Shi, Q.; Kolb, C. E.; Worsnop, D. R.; Barney, W. S.; Cass, G. Geophys. Res. Lett. 2002, 29, 1357, doi:1310.1029/2002GL014692. 409. Smith, G. D.; Woods, E.; DeForest, C. L.; Baer, T.; Miller, R. E. I Phys. Chem. A  2002, 106, 8085-8095. 410. Thornberry, T.; Abbatt, J. P. D. Phys. Chem. Chem. Phys. 2004, 6, 84-93. 411. Katrib, Y.; Martin, S. T.; Hung, H. M.; Rudich, Y.; Zhang, H. Z.; Slowik, J. G.; Davidovits, P.; Jayne, J. T.; Worsnop, D. R. I Phys. Chem. A 2004, 108, 66866695.  Chapter 6^  183  412. Hearn, J. D.; Smith, G. D..1 Phys. Chem. A 2004, 108, 10019-10029. 413. Hearn, J. D.; Lovett, A. J.; Smith, G. D. Phys. Chem. Chem. Phys. 2005, 7, 501-511. 414. Limbeck, A.; Puxbaum, H. Atmos. Environ. 1999, 33, 1847-1852. 415. Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. Atmos. Environ. A-Gen. 1993, 27, 1309-1330. 416. Inoue, T.; Hisatsugu, Y.; Ishikawa, R.; Suzuki, M. Chem. Phys. Lipids 2004, 127, 161-173. 417. de Gouw, J. A.; Lovejoy, E. R. Geophys. Res. Lett. 1998, 25, 931-934. 418. Hanson, D. R.; Lovejoy, E. R. J Phys. Chem. 1996, 100, 6397-6405. 419. Huey, L. G.; Hanson, D. R.; Howard, C. J. I Phys. Chem. 1995, 99, 5001-5008. 420. Longfellow, C. A.; Imamura, T.; Ravishankara, A. R.; Hanson, D. R. I Phys. Chem. A 1998, 102, 3323-3332. 421. Lovejoy, E. R.; Huey, L. G.; Hanson, D. R. J. Geophys. Res. 1995, 100, 1877518780. 422. Koke, J.; Modigell, M. J. Non-Newton Fluid Mech. 2003, 112, 141-160. 423. Barbe, J. C.; Perez, M.; Papoular, M. I Phys.-Condens. Matter 2000, 12, 25672577. 424. Flemings, M. C. Metall. Trans. B 1991, 22, 269-293. 425. Bertram, A. K.; Ivanov, A. V.; Hunter, M.; Molina, L. T.; Molina, M. J. I Phys. Chem. A 2001, 105, 9415-9421. 426. Kurz, W.; Giovanola, B.; Trivedi, R. Acta Metall. Mater. 1986, 34, 823-830. 427. Barthelmes, G.; Pratsinis, S. E.; Buggisch, H. Chem. Eng. Sci. 2003, 58, 2893-2902. 428. Chou, C. M.; Hong, P. D. Macromolecules 2004, 37, 5596-5606.  Chapter 6^  184  429. Wu, S. S.; Wu, X. P.; Xiao, Z. H. Acta Mater. 2004, 52, 3519-3524. 430. Brown, R. L. J Res. Nat. Bur. Stand. 1978, 83, 1-8. 431. Fuller, E. N.; Schettle.Pd; Giddings, J. C. Ind. Eng. Chem. 1966, 58, 19-27. 432. Moise, T.; Rudich, Y. I Geophys. Res.-Atmos. 2000, 105, 14667-14676. 433. Davison, B. Proc. Phys. Soc. A 1951, 64, 881-902. 434. Motz, H.; Wise, H. I Chem. Phys. 1960, 32, 1893-1894. 435. Criegee, R. Angew. Chem. Mt. Edit. 1975, 14, 745-752. 436. Danckwerts, P. V. Trans. Faraday Soc. 1951, 47, 1014-1023. 437. Kolb, C. E.; Worsnop, D. R.; Zahniser, M. S.; Davidovits, P.; Keyser, L. F.; Leu, M. T.; Molina, M. J.; Hanson, D. R.; Ravishankara, A. R. Laboratory Studies of Atmospheric Heterogeneous Chemistry. In Progress and Problems in Atmospheric Chemistry; Barker, J. R., Ed.; World Scientific: Singapore, 1995; Vol. 3. 438. Hanson, D. R. J. Phys. Chem. B 1997, 101, 4998-5001. 439. Worsnop, D. R.; Morris, J. W.; Shi, Q.; Davidovits, P.; Kolb, C. E. Geophys. Res. Lett. 2002, 29, GL015542. 440. Elliott, R. Eutectic Solidification Processing: Crystalline and Glassy Alloys; Butterworth & Co., Ltd.: London, 1983. 441. Mullin, J. W. Crystallization, 4th ed.; Elsevier, Ltd.: Amsterdam, 2001. 442. Mix 1. The composition of this multicomponent solution is as follows: OA = 40.11 wt %, LA = 2.33 wt %, MA = 6.14 wt %, palmiticacid = 33.98 wt %, stearic acid = 17.44 wt %.  Chapter 6^  185  443. Mix 2. The composition of this multicomponent solution is as follows: OA = 38.11 wt %, LA = 7.11 wt %, palmitic acid = 32.16 wt %, stearic acid = 16.5 wt %, succinic acid = 0.84 wt %, glutaric acid = 0.76 wt %, adipic acid = 1.18 wt %, pentacosane = 8.08 wt %,hexadecanamide = 1.26 wt %. 444. Mix 3. The composition of this multicomponent solution is as follows: OA = 33.86 wt %, LA = 1.97 wt %, MA = 5.15 wt %, palmitic acid = 28.69 wt %, stearic acid = 14.72 wt %, succinic acid ---- 0.714 wt %, glutaric acid = 0.62 wt %, adipic acid = 1.03 wt %, palmitoleic acid = 3.1 wt %, tetracosane = 1.23 wt %, pentacosane = 1.33 wt %, 2-pentadecanone = 1.38 wt %, 2-hexadecanone = 1.88 wt %, 2octadecanone = 2.77 wt %, hexadecanamide = 1.57 wt %. 445. Ostwald, W. Z. Phys. Chem. 1900, 34, 495-503. 446. Lifshitz, I. M. J. Phys. Chem. Solids. 1961, 19, 31-50. 447. Wagner, C. Z. Elektrokhemiya 1961, 65, 581-591. 448. Ratke, L. Growth and Coarsening: Ostwald's Ripening in Materials Processing; Springer: Berlin, 2002. 449. Brooks, S. D.; DeMott, P. J.; Kreidenweis, S. M. Atmos. Environ. 2004, 38, 18591868. 450. Braban, C. F.; Abbatt, J. P. D. Atmos. Chem. Phys. 2004, 4, 1451-1459. 451. Parsons, M. T.; Knopf, D. A.; Bertram, A. K. I Phys. Chem. A 2004, 108, 1160011608. 452. Abelson, P. H. Science 1988, 241, 1569-1569. 453. Meng, Z.; Dabdub, D.; Seinfeld, J. H. Science 1997, 277, 116-119.  Chapter 6^ 454. Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics; John Wiley & Sons: New York, 1998.  186  Chapter 7^  187  7. CONCLUDING REMARKS  7.1 Conclusions This thesis has described the heterogeneous chemistry of organic mixtures and inorganic solutions coated with organic monolayers as proxies for atmospheric particles. The objectives outlined in Chapter 1 were achieved using two separate flow reactors coupled to a chemical ionization mass spectrometer. Specifically, a new rectangular channel flow reactor was developed and validated for examining heterogeneous reactions of N205 and aqueous inorganic solutions coated with organic monolayers. The results of these studies were compared to atmospheric observations and the atmospheric implications of our findings were discussed. Additionally, a rotating-wall flow-tube reactor was used to study the reactions between 03 and multicomponent and multiphase organic mixtures as proxies for meat cooking aerosols. In Chapter 2, a new rectangular channel flow reactor was developed for studying reactive uptake coefficients of coated aqueous solutions. A novel mathematical framework was used to derive the true first-order wall loss rate coefficient, kw ist , from the experimentally observed wall loss rate, k o b s , under consideration of vertical diffusion and diffusion in flow direction of the gas-phase reactant. Validation of the new apparatus was performed by measuring the uptake of 0 3 by canola oil as a function of pressure and flow velocity. Additional validation was performed by measuring the reactive uptake coefficients of N205 by aqueous 60 wt % and 80 wt % H2504 solutions. In both cases, the kinetic results were in good agreement with literature values. The first measurement of N205 reactive uptake coefficients on aqueous H2504 in the presence of an insoluble monolayer (1-octadecanol) showed a decrease in N20 5 uptake by nearly 2 orders of magnitude compared to uptake by the bare aqueous H2SO4 solution. In Chapter 3, the reactive uptake coefficient for N205 by aqueous H2SO4 was measured in the presence of four different insoluble organic monolayers (1-octadecanol, 1-hexadecanol, stearic acid, and phytanic acid) using the validated rectangular channel flow reactor. Each monolayer was investigated at a single surface pressure. The reactive  Chapter 7^  188  uptake coefficient decreased drastically in the presence of straight-chain surfactants, but no significant decrease was observed in the presence of the branched-chain surfactant. Additionally, reactive uptake coefficients were correlated with monolayer properties such as surfactant chain length, monolayer surface pressure, and surface area occupied by each surfactant molecule. Based on a limited set of data, the reactive uptake coefficients measured on aqueous sulfuric acid subphases coated with organic monolayers show a relationship to the surface area occupied by each surfactant molecule. In Chapter 4, we extended the investigation in Chapter 3 by measuring uptake of N20 5 by aqueous solutions through 1-component (1-octadecanol) monolayers as a function of surface density and through 2-component monolayers (1-octadecanol and phytanic acid) as a function of monolayer composition. For 1-component monolayers, we demonstrated that when the fractional surface coverage was less than 1, the monolayer still showed significant resistance to mass transfer, which is consistent with previous studies of N20 5 reactivity on aqueous particles in the presence of surfactants. 455457 This observation may be of atmospheric relevance, since monolayers in the atmosphere are not expected to always have a fractional surface coverage of 1. For 2-component monolayers, we showed that 1-octadecanol (straight-chain) and phytanic acid (branched) form immiscible monolayers, and that N205 uptake in the presence of 2-component monolayers is strongly dependent on the fraction of branched molecules in the monolayer. When the mole fraction of branched surfactant was only 0.20 the decrease in the reactive uptake coefficient, compared to the uncoated solution, was only a factor of 2 (down from 42). Our results for 2-component monolayers showed that the overall resistance to reactive uptake of immiscible monolayers can be predicted reasonably accurately by a model that assumes the resistances to mass transfer can be combined in parallel. This observation may be useful for making predictions of reactive uptake of aqueous particles coated with multi-component monolayers in the atmosphere, and highlights the importance of understanding the composition of mixed organic monolayers in atmospheric particles. In Chapter 5, we investigated the reactive uptake of N20 5 in the presence of soluble surfactant monolayers. The surface activities of two atmospherically relevant dicarboxylic acids (glutaric acid and azelaic acid) were characterized in aqueous 40 wt %  Chapter 7^  189  sulfuric acid solutions. Glutaric acid and azelaic acid were found to have surface excesses of 13 ± 4 x 10 13 cm -2 and 6 ± 2 x 10 13 cm 2 , respectively, in aqueous 40 wt % H2SO4 at 260 K. Reactive uptake coefficients for N205 in the presence of these solutions were not found to be significantly different than that for the uncoated aqueous solutions. These reactive uptake coefficients were correlated with molecular surface area and compared to results in previous chapters. Good agreement was observed and this correlation of N20 5 uptake on aqueous sulfuric acid solutions as a function of molecular surface area now includes organic components consisting of mono- and di-functional organic molecules that form both soluble and insoluble monolayers. Using surface tension data for glutaric acid and azelaic acid in water to calculate surface excesses, combined with the correlation of y with molecular surface area, we predict that N205 uptake on atmospheric particles coated with a monolayer of water soluble dicarboxylic acids will decrease by a factor of 2-3. Finally, in Chapter 6, the oxidation of oleic acid by ozone was measured for mixtures of 0A/MA and meat cooking mixtures. The reactive uptake coefficients obtained for the binary liquid MA/OA solutions were slightly lower than that for pure oleic acid. In contrast, the reactive uptake of 03 was approximately 1 order of magnitude slower on binary solid-liquid mixtures and multicomponent mixtures that closely represent compositions of meat-cooking aerosols compared to the liquid solutions. We showed that different substrate preparation methods resulted in changes in reactivity resulting from formation of distinct microstructures of the solid-liquid substrates. Also, we observed an unexpected increase in the reactivity with time for solid-liquid OAcontaining mixtures, attributed to Ostwald's ripening or the relaxation of a nonequilibrium phase to the stable phase. Both of these observations should be considered when modeling the lifetime and chemical processing of atmospheric particles. The OA lifetimes obtained for semisolid OA mixtures are up to a factor of 16 higher than for liquid mixtures. Lifetimes up to 75 min were obtained for these mixtures. The overall conclusion from Chapters 3-5 is that the uptake process for N205 on aqueous H2SO4 appears to be governed by mass transport rather than by reaction, as evidenced by the strong correlation of the reactive uptake coefficient with the surface area of organic surfactants. However, the aqueous subphase seems to influence N205  Chapter 7^  190  uptake significantly when comparing uptake results for different aqueous subphases. This apparent complexity should be investigated with future studies and will be discussed below.  7.2 Considerations for Future Work A number of experiments would be beneficial in explaining some of the results obtained in this thesis. The following are suggestions for further studies that would strengthen the results in previous chapters: 1. In Chapter 3, it was first noted that experimental data obtained with aqueous sulfuric acid subphases is not in agreement with other subphases when the reactive uptake coefficient is plotted versus the molecular surface area of the monolayer. There are several possible explanations for this discrepancy: it is possible that a different mechanism is important for the different subphases, another explanation may be differences in results related to the experimental techniques (bulk solution studies versus aerosol measurements). Alternatively, it is possible that not one single variable (molecular surface area) can explain the overall reactivity of N205. To rule out differences in experimental techniques, comparison of N20 5 uptake studies investigating the same organic subphase and the same organic surfactant with both aerosol and bulk solution techniques would be an asset. Likewise, measuring reactive uptake of N20 5 using the same organic monolayer on both H2SO4 and NaCl would shed light onto the issue of different mechanisms of reaction. Finally, it is quite possible that more than one variable (molecular surface area, chain length, temperature, etc.) needs to be considered and incorporated into a model to explain all the experimental data. 2. In Chapter 4, in order to provide a more rigorous evaluation of the applicability of the Accesible Area Theory for both short- and long-chain organic surfactants more reactive uptake coefficients for long-chain surfactants at high molecular surface areas would be beneficial. This would enable direct comparison of reactive uptake coefficients in the presence of short- and long-chain surfactants with the same molecular surface area.  Chapter 7^  191  3. In Chapter 5, in order to calculate the surface excess of glutaric acid and azelaic acid in aqueous sulfuric acid solutions it was assumed that the change in the chemical potentials of H 2 O and H2SO4 is small upon addition of small quantities of glutaric acid or azelaic acid. Further work measuring the change in chemical potential of H2O and H 2 SO 4 with the addition of glutaric acid or azelaic acid is needed to verify this assumption and determine the concentration range under which this relationship is valid. 4. In general, a better understanding of the composition and phase of organic monolayers on aqueous inorganic aerosols would be beneficial in simulating atmospheric monolayers in laboratory scenarios.  A natural extension of this research would be to investigate heterogeneous uptake of N205 and other atmospherically relevant gases in the presence of monolayers composed of organic material collected from atmospheric samples. Although the physical properties of these atmospheric films would be difficult to characterize, the kinetic results of these studies would be directly applicable for use in atmospheric models. Also of atmospheric interest would be the investigation of N205 reactive uptake in the presence of organic monolayers exposed to atmospheric oxidants, as a function of oxidant exposure time, to simulate the atmospheric processing of organic monolayers by atmospheric oxidants. These studies would lead to a better understanding of how the uptake of N205 changes through the evolution of an aging organic monolayer.  Chapter 7^  7.3 References 455. McNeill, V. F.; Patterson, J.; Wolfe, G. M.; Thornton, J. A. Atmos. Chem. Phys.  2006, 6, 1635-1644. 456. Park, S. C.; Burden, D. K.; Nathanson, G. M. J. Phys. Chem. A 2007, 111, 29212929. 457. McNeill, V. F.; Wolfe, G. M.; Thornton, J. A. I Phys. Chem. A 2007, 111, 10731083.  192  A. Appendix^  193  A. APPENDIX  The uptake of gases by surfaces is a complex interaction that is governed by gasand condensed-phase parameters. Figure 0.1 shows a schematic of the various processes that may influence uptake: 458  Liquid Phase Diffusion Gas Phase^Adsorptio Diffusion Solvation Desorption  Liquid Phase Reaction  Surface Reaction Figure 0.1: Schematic of the physical and chemical processes which determine the overall uptake in gas-particle interactions ass Heterogeneous uptake of a gas into a liquid particle, followed by a reaction involves a variety of physical and chemical processes as described in Figure 0.1. First, the gas-phase molecule must diffuse through the atmosphere to the liquid surface. It can then undergo adsorption and desorption at the surface, or undergo salvation into the bulk liquid. Once in the liquid, it can diffuse through bulk liquid where it may undergo a chemical reaction in the bulk liquid. A gas-phase species may undergo one or more of these steps during a gas-particle interaction.  194  A. Appendix^  The uptake of a gas into a liquid particle (and hence the reactive uptake coefficient, y) can be expressed by a series of coupled differential equations, each describing the steps discussed above. These equations can not be solved analytically for all cases, and often the steps in the reaction are treated individually. 459 One approximation that is used frequently when discussing gas-particle interactions is the Resistor Mode1, 459-462 an approximation that decouples each of these physical and chemical processes and treats each step as a resistance in an electrical circuit. This model simplifies heterogeneous interactions by allowing each step to be examined individually, and allows one to obtain a relationship between y and the physical processes involved in the uptake process. Here we will discuss the Resistor Model, considering both surface reactions and reactions in the bulk liquid. A schematic for the Resistor Model for this situation is shown Figure 0.2, 458  1 1 Fg^S  1 S ksol kdesorb  1 Frxri + Fsat  1 "  surf  Figure 0.2: Schematic of the Resistor Model for gas-particle interactions taking into consideration gas-phase diffusion, mass accommodation, solubility limited uptake, liquid-phase reaction, and surface reaction.  where S is the sticking coefficient (fraction of collisions at the surface that result in accommodation on the surface), ks .1 is the rate coefficient for the transfer from the surface into the liquid, kdesorb is rate coefficient for the transfer of molecules from the surface into the gas phase, Frx„ is rate of reaction in the bulk of the solution, normalized to the gas-  A. Appendix^  195  phase collision frequency, Tat is related to the solubility limitation (liquid-phase saturation), and T urf represents the surface reaction. The net uptake for gas-particle interactions, is then described by combining the resistors in series and in parallel to yield eq 0.1. 462 Equation 0.1 considers reaction in both the bulk solution and at the surface. 462,463 0. 1  1^1^1 —= + + y Fd S  1  1 1 ^1 + Frxa + Fsa^ks 1  S ^°  1  If we assume that diffusion in the gas-phase is fast, we obtain eq 3.4 shown in Chapter 3: 1^1^1  — = ±^  7 S^1  1 ^1 + ks 01 F^ b S k5 kdesorb  where Fb = Fnm + Fat.  + Fswf  A. Appendix^  196  B. APPENDIX  Appendix A introduces the Resistor Model and considers the case where both reaction at the surface and in the bulk can occur. If reaction only in the bulk is considered, then Figure 0.1 can be used to represent the Resistor Mode1: 458  1  Fs at Figure 0.1: Schematic of the resistor model for gas-particle interactions taking into consideration gas-phase diffusion, mass accommodation, solubility limited uptake, liquid-phase reaction ass  where a is the mass accommodation coefficient, Fg , Frxn, and Fsat are the rates of gasphase diffusion, reaction in the bulk of the solution, and the solubility limitation (liquidphase saturation), respectively, normalized to the gas-phase collision frequency. The two resistors 1/S and 1 /S(ksoadesorb) in Figure 0.2 are represented by a single resistor (1/a) in Figure 0.1. The net uptake for gas-particle interactions, is then described by combining the resistors in series and in parallel to yield eq 0.1. 460 ' 461 0.1  1^1^1^1 = + +^ 7 Fd a F. + Fsat  197  A. Appendix^  When the chemical reaction is fast, and the gas-phase solubility is low  (r \-- sat< < Frxn),  which we assume is the case for the reaction of 0 3 and oleic acid, eq 0.1 reduces to eq 0.2: 464 0.2  ^  1^1^1 y a Frx,,  —=+  where the rate of reaction in the bulk of the solution, normalized to the gas-phase collision frequency, Tom,,, is given by eq 0.3. 459  0.3  ^  4HRT .142, k, F rxn  C 0,  Substituting Frxt, in eq 0.2, where k1=k2[0A], eq 0.2 becomes eq 6.2 in Chapter 6. 459 '460  + ^1 = 1^cn —3 6.2 ^±  ^7---  y a 4HRT.003k2[0,4]  A. Appendix^  198  References 458. Davidovits, P.; Hu, J. H.; Worsnop, D. R.; Zahniser, M. S.; Kolb, C. E. Faraday Discuss. 1995, 65-81. 459. Danckwerts, P. V. Trans. Faraday Soc. 1951, 4 7, 1014-1023. 460. Kolb, C. E.; Worsnop, D. R.; Zahniser, M. S.; Davidovits, P.; Keyser, L. F.; Leu, M. T.; Molina, M. J.; Hanson, D. R.; Ravishankara, A. R. Laboratory Studies of Atmospheric Heterogeneous Chemistry. In Progress and Problems in Atmospheric Chemistry; Barker, J. R., Ed.; World Scientific: Singapore, 1995; Vol. 3. 461. Schwartz, S. E. "Chemistry of multiphase atmospheric systems"; NATO ASI Series, 1986, Springer- Verlag, Berlin. 462. Hanson, D. R. J. Phys. Chem. B 1997, 101, 4998-5001. 463. Shi, Q.; Davidovits, P.; Jayne, J. T.; Worsnop, D. R.; Kolb, C. E. J. Phys. Chem. A 1999, 103, 8812-8823. 464. Finlayson-Pitts, B. J.; Pitts, J. N. Chemistry of the upper and lower atmosphere: Theory, experiments, and application; Academic Press: San Diego, CA, 2000.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0061818/manifest

Comment

Related Items