USF Libraries
USF Digital Collections

Characterization and formation of particulate nitrate in a coastal area

MISSING IMAGE

Material Information

Title:
Characterization and formation of particulate nitrate in a coastal area
Physical Description:
Book
Language:
English
Creator:
Evans, Melissa Cheryl Foster
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
mineral dust
modeling
nitric acid
sea salt
tampa bay
Dissertations, Academic -- Chemistry -- Doctoral -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Particulate nitrates play important roles in the atmosphere. They consist mainly of NH4NO3 and NaNO3, products from the reactions of gaseous HNO3 with gaseous NH3 and sea salt, respectively. The gas-to-particle phase conversion of nitrate changes its deposition characteristics and ultimately changes the transport and deposition rates of the locally produced species. Studies were conducted to develop background information on the particle concentrations and size distributions and the chemistry and kinetics behind the interactions. The predominant nitrate species in the Tampa Bay area was identified as coarse mode NaNO3. NH4NO3 was not detected as it has high volatility at ambient temperatures. Spatial distribution sampling determined a gradient of NaCl and NaNO3 with increased distance from the coastal shore and an increase in the gas-to-particle conversion of nitric acid along a predominant air mass trajectory. The EQUISOLV II thermodynamic equilibrium model was evaluated. It was determined that the model can be used to predict gas and size-distributed particulate matter concentrations. The model was also used to examine the gas-to-particle partitioning of nitric acid to nitrate by NaCl and CaCO3. Both sodium and calcium partitioned nitrate to the particle phase. The magnitude of the partitioning was directly dependent on the equilibrium coefficients. The fine mode percentage of the total nitrate was determined using two methods. The results were used to expand the current data set to account for the coarse mode nitrate, and they indicated that particle nitrate accounted for 9% of the total nitrogen deposition flux to Tampa Bay. The formation of particle nitrate was examined using a nitrate accumulation model. Results indicated that the equilibrium time for particles less than 10 um in diameter was significantly less than their atmospheric residence time, with fastest conversion occurring under the highest relative humidity conditions. This information is vital in the development of atmospheric nitrogen dry deposition estimates, which are used to assess water quality and nutrient loading. These data can be used to determine air-monitoring strategies and to develop models that account for the coarse particle nitrogen species.
Thesis:
Thesis (Ph.D.)--University of South Florida, 2003.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Melissa Cheryl Foster Evans.
General Note:
Includes vita.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 236 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001441469
oclc - 54018444
notis - AJM5909
usfldc doi - E14-SFE0000142
usfldc handle - e14.142
System ID:
SFS0024838:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Characterization and Formation of Particulate Nitrate in a Coastal Area by Melissa Cheryl Foster Evans A dissertation submitted in partial fulfillment of the requirement s for the degree of Doctor of Philosophy Department of Chemistry College of Arts and Sciences University of South Florida Co-Major Professor: Noreen D. Poor, Ph.D. Co-Major Professor: Julie P. Harmon, Ph.D. Scott W. Campbell, Ph.D. Milton D. Johnston, Jr., Ph.D. Date of Approval: November 5, 2003 Keywords: mineral dust, modeling, nitric acid, sea salt, Tampa Bay Copyright 2003, Melissa Cheryl Foster Evans

PAGE 2

Dedication To my loving husband.

PAGE 3

Acknowledgements Several people deserve special thanks for their contribution to this dissertation. I would like to extend my deepest gratitude to my advisor and co major professor, Noreen D. Poor, for her unending sup port, motivation and guidance throughout this project. I feel extremely fortunate to have had the opportunity to work with such a dedicated individual. I would like to thank my other co major professor Julie P. Harmon, for her willingness to support me as I studied outside the department. I am extremely grateful for my committee members. Many thanks to Scott W. Campbell for his valuable insight, gentle spirit and encouragement. I am especially appreciative for his gift in teaching and devotion to studen ts. Thanks to Milton D. Johnston, Jr., for his guidance and expertise. Thanks to Yehia Y. Hammad for serving as my defense chair and for his knowledge and instruction in aerosol science. Many other individuals deserve thanks for their contribution to t his project. I am grateful for the many graduate and undergraduate students I have had the pleasure of working with and for the assistance I have had from the Environmental Protection Commission of Hillsborough County. I extend my gratitude to my family and friends for their unconditional love, encouragement and support. This research was funded by the Florida Department of Environmental Protection and the Tampa Bay Estuary Program.

PAGE 4

i Table of Contents List of Tables iv List of Figures vii List of Symbols and Abbreviations xii Abstract xviii Introduction 1 Eutrophication 1 Aerosols 3 Mineral Dust 7 Sodium 8 Nitrate 10 Ammonium 12 Sulfate 14 Particle Diameter 16 Deposition Velocities 17 Sampling 19 Filtration Methods 19 Denuder Methods 21 Cascade Impactors 25 Virtual Impactors 29 Kinetics 34 Heterogeneous Reactions 34 Sea Salt 37 Mineral Dust 42 Statement of the Problem 45 Methods and Experimental 47 Instrumentation 47 Cascade Impactors 47 Annular Denuder System 48 Dichotomous Sampler 49 Total Suspended Particulate Collection 49

PAGE 5

ii Filters 50 Filter Extraction 51 Denuder Preparation and Extraction 52 Preparation 52 Extraction 53 Sample Analysis 53 Ion Chromatography Analysis 53 pH Analysis 54 Trajectory Analysis 55 Error Analysis 56 MOUDI 56 Annular Denuder System 68 Statistical Analysis 69 Grubbs’ Outlier Test 69 Paired t-Test 69 Wilcoxon’s Signed Rank Test 70 Experimental Studies 71 Preliminary Dichotomous Studies 71 Experimental 71 Results and Discussion 72 Size Distribution Determination 74 Experimental 74 Results 75 Discussion 78 Evidence of Macroparticles 80 Experimental 81 Results and Discussion 81 Retention of Nitric Acid by Nylon Filters 87 Experimental 88 Results and Discussion 90 Size Distributed Trajectory Study 96 Experimental 97 Results and Discussion 99 EQUISOLV II: A Thermodynamic Model 109 Inputs 111 Outputs 112 Limitations 113 Qualitative Analysis 114 Model Comparison 128 Case Studies 130 The Partitioning of Nitric Acid to Nitrate 130 EQUISOLV II Model 130 Results and Discussion 132

PAGE 6

iii Competitive Partitioning 137 Environmental Implications 139 The Prediction of Coarse Mode Nitra te from Fine Mode ADS Data 144 Environmental Implications 152 The Formation of Particulate Nitrate 153 Methods 154 Deposition Velocities 156 Residence Times and Dist ances Traveled 157 Reaction with Nitric Acid 158 Results and Discussions 168 Impact on Nitrogen Loading 171 Conclusions 174 References 177 Appendices 193 Appendix 1: Meteor ological Data 194 Appendix 2. Size Distributions and Ion Ratios from May 2002 197 Appendix 3. Density Calculation for Aqueous Aerosol 207 Appendix 4. Size Distributions and Data Inversion 209 About the Author End Page

PAGE 7

iv List of Tables Table 1. Average over water deposit ion velocities and their respective standard deviations for the MOUD I ten particle size ranges calculated using May 2002 meteorological data. 18 Table 2. Equilibrium models, species treated and numerical method used to solve equilibrium (Jacobson, 1999a). 33 Table 3. Addresses and coordina tes for each sampling site. 55 Table 4. Relative precision for t he MOUDI instrument during May 2002. 67 Table 5. Relative precisi on for the annular denuder system measurements. 68 Table 6. Total and averaged daily concentrations of dichotomous samples collected January 11-13, 2001. 75 Table 7. Concentrations for the A ndersen cascade impactor for January 11-13, 2001. 76 Table 8. Average chloride depletion, in percentage, for January 11-13, 2001. 79 Table 9. Dry deposition flux for parti culate nitrogen (nitrate + ammonium) for January 11-13, 2001. 80 Table 10. Experimental nitric ac id, denuded nitrate and undenuded nitrate concentrations for October November 2001. 91 Table 11. Ion ratios for May 4, 2002. 101 Table 12. Ion ratios for May 14, 2002. 103 Table 13. Ion ratios for May 6, 2002. 105 Table 14. Ion ratios for May 20, 2002. 106

PAGE 8

v Table 15. Comparison between measured HNO3 gas concentrations and those modeled by EQUISOLV II. 115 Table 16. Comparison between measured NH3 gas concentrations and those modeled by EQUISOLV II. 116 Table 17. Comparison between m easured HCl gas concentrations and those modeled by EQUISOLV II. 117 Table 18. Particle size fraction, instrumentation and data collection periods used for predicting coarse mode nitrate fractions. 146 Table 19. Percent fine mode nitrat e determined using actual coarse and fine mode nitrate samples. 149 Table 20. Computed percent fine m ode nitrate using lognormal analysis of cascade impactor data. 151 Table 21. The predicted nitrate conc entrations and resulting over water dry deposition fluxes. 152 Table 22. Averaged concentration fr om 37 MOUDI experimental samples and year 2000 ADS measurements. 155 Table 23. Annual averaged over water dry deposition velocities with their respective standard deviations. 157 Table 24. Traveling distances an d residence times for low (<2.4 m s-1), mid (2.4-6.0 m s-1) and high (>6.0 m s-1) wind speeds. 158 Table 25. Mass fractions of sodium and calcium in NaCl, sea salt and mineral dust. 162 Table 26. Calculated densitie s for NaCl, sea salt and CaCO3 at varying relative humidities. 163 Table 27. The percent of particulate nitrate formation based on an initial height of 100 m at the residenc e time and different ambient relative humidity values. 169 Table 28. The calculated nitrogen over water dry deposition flux. 172 Table 29. Temperature and relative humidity data for October November 2001. 194 Table 30. Temperature and relati ve humidity data for May 2002 Azalea Park sampling site. 195

PAGE 9

vi Table 31. Temperature and relati ve humidity data for May 2002 Gandy sampling site. 195 Table 32. Temperature and relati ve humidity for May 2002 Sydney sampling site. 196 Table 33. Ion ratios for May 10, 2002. 197 Table 34. Ion ratios for May 15, 2002. 198 Table 35. Ion ratios for May 16, 2002. 199 Table 36. Ion ratios for May 17, 2002. 200 Table 37. Ion ratios for May 19, 2002. 201 Table 38. Ion ratios for May 23, 2002. 202 Table 39. Ion ratios for May 24, 2002. 203 Table 40. Ion ratios for May 25, 2002. 204 Table 41. Ion ratios for May 31, 2002. 205 Table 42. Overall ion ratios for May 2002. 206

PAGE 10

vii List of Figures Figure 1. Cross-sectional view of an impactor stage (adapted from Hinds, 1999). 27 Figure 2. 24-hour backward trajectories for (a) terrestrial or land origin (October 17, 2001) and (b) mari ne origin (October 25, 2001) (HYSPLIT4, 1997). 55 Figure 3. Comparison of collocat ed measurements of instrument A ( ) and instrument B ( ) for May 4, 2002. 58 Figure 4. Comparison of collocated measurements of instrument A ( ) and instrument B ( ) for May 6, 2002. 59 Figure 5. Comparison of collocat ed measurements of instrument A ( ) and instrument B ( ) for May 10, 2002. 60 Figure 6. Comparison of collocat ed measurements of instrument A ( ) and instrument B ( ) for May 14, 2002. 61 Figure 7. Comparison of collocat ed measurements of instrument A ( ) and instrument B ( ) for May 15, 2002. 62 Figure 8. Comparison of collocat ed measurements of instrument A ( ) and instrument B ( ) for May 16, 2002. 63 Figure 9. Comparison of collocat ed measurements of instrument A ( ) and instrument B ( ) for May 17, 2002. 64 Figure 10. Comparison of collocat ed measurements of instrument A ( ) and instrument B ( ) for May 19, 2002. 65 Figure 11. Comparison of collocat ed measurements of instrument A ( ) and instrument B ( ) for May 20, 2002. 66 Figure 12. Coarse percentage for di chotomous samples collected during October 5-12, 2000. 73

PAGE 11

viii Figure 13. Normalized particle size di stributions of sodium, nitrate and chloride using Andersen instrum ent from January 11-13, 2001. 77 Figure 14. Normalized particle size di stributions of am monium and sulfate using the Andersen instrument from January 11-13, 2001. 77 Figure 15. Daily macroparticl e concentrations of (a) Na+ and Cland (b) Ca2+ and NO3 (October-November 2001). 83 Figure 16. Simple line ar regression for TSP NO3 versus dichotomous total NO3 using daily concentrations (October-November 2001). 84 Figure 17. Daily macroparticle nitrat e concentrations with correction for the nitric acid bias. 85 Figure 18. HNO3 + Denuded PM NO3 vs. Undenuded NO3 for Whatman nylon filters. 92 Figure 19. HNO3 (g) + Denuded PM NO3 vs. Undenuded NO3 for Nylasorb nylon filters. 93 Figure 20. Nitric acid from the AD S compared to that adsorbed by the Whatman nylon filters. 94 Figure 21. Nitric acid from the AD S compared to that adsorbed by the Nylasorb nylon filters. 95 Figure 22. Map of sampling sites dur ing the May 2002 intensive period. 99 Figure 23. Backward air mass trajectories for (a) May 4th, (b) May 14th, (c) May 6th and (d) May 20th, 2002. 100 Figure 24. Size distributions for ea ch sampling site on May 4, 2002. 101 Figure 25. Size distributions for ea ch sampling site on May 14, 2002. 103 Figure 26. Size distributions for ea ch sampling site on May 6, 2002. 105 Figure 27. Size distributions for ea ch sampling site on May 20, 2002. 106 Figure 28. Nitrate and ammonium particu late flux for select days in May 2002. 108 Figure 29. EQUISOLV II input file example. 112

PAGE 12

ix Figure 30. Comparison of exper imental, EQUISOLV II default mode and EQUISOLV II metastable mode data for ammonium at the Azalea site. 118 Figure 31. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for ammonium at the Gandy site. 119 Figure 32. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for ammonium at the Sydney site. 120 Figure 33. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode da ta for nitrate at the Azalea site. 121 Figure 34. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode da ta for nitrate at the Gandy site. 122 Figure 35. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for nitrate at the Sydney site. 123 Figure 36. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for chloride at the Azalea site. 124 Figure 37. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for chloride at the Gandy site. 125 Figure 38. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for chloride at the Sydney site. 126 Figure 39. The partitioning of HNO3 to nitrate by (a) NaCl and (b) CaCO3 by different ambient air concent rations and total nitrate at 78% RH. 133 Figure 40. The effect of relative humidity on the partitioning of HNO3 to nitrate by (a) NaCl and (b) CaCO3, where the total available nitrate was 3 g m-3. 134 Figure 41. The concentration and mass percent of water and the fraction of total nitrate within the particle for NaCl and CaCO3 at 78% RH. 136

PAGE 13

x Figure 42. The competitive nitrate partitioning effect between NaCl and CaCO3, in molar percentages at 78% RH. 138 Figure 43. The predicted HNO3 gas flux for the nitric acid partitioning by (a) NaCl and (b) CaCO3 at 78% RH. 140 Figure 44. The predicted nitrate particle flux for the nitrate partitioning by (a) NaCl and (b) CaCO3 at 78% RH. 141 Figure 45. The total (gas + particl e) predicted flux for the nitrate partitioning by (a) NaCl and (b) CaCO3 at 78% RH. 142 Figure 46. The predicted (a) nitric acid gas and (b) particulate nitrate flux from a molar percent mi xture of NaCl and CaCO3 3 248 ] [ ] [ m neq Ca Na at 78% RH. 143 Figure 47. Nitrate particle size distribution. 147 Figure 48. Linear regression for dichot omous fine mode and total nitrate. 147 Figure 49. Linear regression for dic hotomous fine and the inverted filter pack TSP nitrate. 148 Figure 50. Linear regression fo r annular denuder system fine and TTU particle nitrate. 148 Figure 51. Typical lognormal ni trate size distribution. 150 Figure 52. Time-resolved nitrate formation for NaCl. 165 Figure 53. Time-resolved nitrate formation for sea salt. 166 Figure 54. Time-resolved nitrate formation for CaCO3. 167 Figure 55. Typical size distribution of sodium, calcium and nitrate in the Tampa Bay area. 170 Figure 56. Size distributi ons for May 10, 2002. 197 Figure 57. Size distributi ons for May 15, 2002. 198 Figure 58. Size distributi ons for May 16, 2002. 199 Figure 59. Size distributi ons for May 17, 2002. 200 Figure 60. Size distributi ons for May 19, 2002. 201 Figure 61. Size distributi ons for May 23, 2002. 202

PAGE 14

xi Figure 62. Size distributi ons for May 24, 2002. 203 Figure 63. Size distributi ons for May 25, 2002. 204 Figure 64. Size distributi ons for May 31, 2002. 205 Figure 65. Average size distributions for May 2002. 206 Figure 66. Collection efficiencies as a function of particle diameter for the MOUDI sampler (adapted from Marple et al., 1991). 210 Figure 67. Deposition kernel functi ons for the MOUDI as functions of particle aerodynamic diameter. 212

PAGE 15

xii List of Symbols and Abbreviations mass accommodation coefficient (ads) adsorbed ADS annular denuder system (aq) aqueous phase avg average Bx fitting coefficient for electrolyte x C degree Centigrade c molecular speed C concentration cm centimeter(s) CRH crystallization relative humidity d day(s) d mean difference between data sets Da particle aerodynamic diameter den denuded Dg particle geometric diameter 3HNOD diffusion coefficient of nitric acid dichot dichotomous sampler

PAGE 16

xiii Dp particle diameter Dp50 particle diameter collected with 50% efficiency DRH deliquescent relative humidity F flux Faq mass fraction of the aqueous phase fi mass fraction of species i FP filter pack Fsolid mass fraction of the solid phase FTIR Fourier transform infrared spectroscopy (g) gas phase g gram(s) uptake coefficient measured measured uptake coefficient net net uptake probability 3 4NO NH mean mixed activity coefficient of NH4NO3 rxn surface reaction probability ha hectare(s) 3NOI nitrate accumulation rate i integer keq equilibrium coefficient kg kilogram(s) kg-N kilograms of nitrogen

PAGE 17

xiv kg-S kilograms of sulfur Ki kernel function for species i km kilometer(s) ks surface rate constant (l) liquid phase L liter(s) m meter(s) m molality Max maximum MBL marine boundary layer Mi molar mass of species i min minute(s) Min minimum mL milliliter(s) M megohm MOUDI Micro-Orifice Un iform Deposit Impactor mTorr milliTorr(s) air viscosity eq microequivalent(s) g microgram(s) m micrometer(s) mol micromole(s) MW molecular weight

PAGE 18

xv N nitrogen n number of samples or observations p gas phase partial pressure PM particulate matter PM10 particulate matter less than 10 m in diameter PM2.5 particulate matter less than 2.5 m in diameter PM10-2.5 particulate matter (2.5 < Dp < 10 m diameter) PTFE polytetrafluoroethylene q volumetric flow rate r correlation coefficient RB relative bias RH relative humidity aq density of aqueous phase o standard density (1 g cm-3) p particle density s density of solute solid density of the solid phase w density of water RP relative precision RT residence time s second(s) (s) solid phase

PAGE 19

xvi standard deviation SWilcoxon’s signed rank test negative value S+ Wilcoxon’s signed rank test positive value S sulfur SD standard deviation soln solution St Stokes number t paired t-test value T time tcrit 95% tcritical at the 95% confidence interval tcritical paired t-test critical value Tcritical Wilcoxon’s signed rank test critical value tobtained paired t-test obtained value Tobtained Wilcoxon’s signed rank test obtained value TTU Texas Tech University TD traveling distance of a particle Tg teragram(s) (1012 gram) TSP total suspended particulate matter TT traveling time of a particle UD undenuded Vd deposition velocity Vi molar volume of i Vo average air velocity

PAGE 20

xvii v/v volume/volume W nozzle diameter WS wind speed ws mass fraction of solute ww mass fraction of water w/v weight/volume w/w weight/weight {X} thermodynamic activity of species X Xi mole fraction of i yr year(s) Zcritical Grubbs’ outlier test critical value Zobtained Grubbs’ outlier test obtained value

PAGE 21

xviii Characterization and Formation of Part iculate Nitrate in a Coastal Area Melissa Cheryl Foster Evans ABSTRACT Particulate nitrates play important roles in the atmosphere. They consist mainly of NH4NO3 and NaNO3, products from the reactions of gaseous HNO3 with gaseous NH3 and sea salt, respectively. The gas-to-particle phase conversion of nitrate changes its depo sition characteristics and ultimately changes the transport and deposition rates of the locally produced species. Studies were conducted to develop bac kground information on the particle concentrations and size distributions and the chemistry and kinetics behind the interactions. The predominant nitrate species in the Tampa Bay area was identified as coarse mode NaNO3. NH4NO3 was not detected as it has high volatility at ambient temperatures. S patial distribution sampling determined a gradient of NaCl and NaNO3 with increased distance from the coastal shore and an increase in the gas-to-particle conversion of nitric acid along a predominant air mass trajectory. The EQUISOLV II thermodynamic equi librium model was evaluated. It was determined that the model can be used to predict gas and size-distributed

PAGE 22

xix particulate matter concentrations. The m odel was also used to examine the gasto-particle partitioning of nitric acid to nitrate by NaCl and CaCO3. Both sodium and calcium partitioned nitrate to the par ticle phase. The magnitude of the partitioning was directly dependent on the equilibrium coefficients. The fine mode percentage of the tota l nitrate was determined using two methods. The results were used to exp and the current data set to account for the coarse mode nitrate, and they indicat ed that particle nitrate accounted for 9% of the total nitrogen depositi on flux to Tampa Bay. The formation of particle nitrat e was examined using a nitrate accumulation model. Results indicated that the equilibrium time for particles less than 10 m in diameter was significant ly less than their atmospheric residence time, with fastest conversion occurring under the highest relative humidity conditions. This information is vital in the development of atmospheric nitrogen dry deposition estimates, which are used to assess water quality and nutrient loading. These data can be used to determine air-monitoring strategies and to develop models that account for the coarse particle nitrogen species.

PAGE 23

1 Introduction Eutrophication The atmospheric deposition of nitr ogen is a major concern for the Tampa Bay, Florida region (Tampa Bay National Estuary Program, 1996). Nitrogen is a primary limiting nutrient in marine waters (Conley, 2000; de Wit and Bendoricchio, 2001; Nixon et al., 1996) for phytoplankton growth (Pryor and Sorensen, 2000) and is a funda mental building block for plant cells (Pryor and Barthelmie, 2000a). Excess amounts of nitrogen can pollute the bay by accelerating phytoplankton growth (Bur ian et al., 2001; Tampa Bay National Estuary Program, 1996). These alga l blooms, or increased abundance of phytoplankton, have implications for wa ter quality, human health, and ecosystem health and productivity (Pryor and Barthel mie, 2000b). The algal blooms inhibit sufficient light penetration required for se agrass growth and result in seagrass death. This has a dramatic impact on t he local environment, as the seagrasses serve as a protective habitat and feeding gr ounds for fish and shellfish. Partial or complete oxygen depletion (hypoxia or anoxia) is o ften seen following the algal blooms as the algae begin to decay (Burian et al., 2001 ; de Leeuw et al., 2001). The condition associated with excess ni trogen, algal blooms, oxygen depletion

PAGE 24

2 and the degradation of water quality is known as eu trophication (Paerl, 2001; Tampa Bay National Estuary Program, 1996). Research has indicated that approximately 29% (o r 980 metric tons-N yr-1) of the bay’s total nitrogen loading co mes from atmospheric deposition of pollutants directly to the surface of t he bay (Burian et al ., 2001; Patwardhan and Donigian Jr., 1995; Tampa Bay National Estuary Progr am, 1996; Zarbock et al., 1996). These estimates are actually much higher if depositio n to the watershed is included, since much of this nitr ogen will eventually enter the bay in stormwater runoff (Tampa Bay National Estuary Program, 1996). EPA estimates that as much as 67% of Tampa Bay’ s total nitrogen could come from the atmosphere (Patwardhan and Donigian Jr., 1995). Other major sources of nitrogen loading include point sources (19%) and fertilizer application (14%) (Patwardhan and Donigian Jr., 19 95; Zarbock et al., 1996). Dry and wet deposition are the two ma in processes by which nitrogen is transferred to surface waters. Dry deposit ion is the transfer of gaseous species and particles in the absence of precipit ation and is determined by meteorological conditions, atmospheric conditions and particle and gas properties. In wet deposition, gaseous species and particles are transferred to the surface via precipitation (rain, fog, sleet and snow). Dry deposition fluxes are difficult to measure and are often calculated under the assumption that t he dry deposition flux is dire ctly proportional to the concentration of the species (Caffrey et al., 1998; Poor et al., 2001; Seinfeld and Pandis, 1998),

PAGE 25

3 15576 3 C V Fd (Equation 1) where F (kg ha-1 yr-1) is the dry particle or gas deposition flux, Vd (cm s-1) is the deposition velocity and C (g m-3) is the concentration of nitrogen. The Vd depends on particle size or Henry’s constant for gaseous species, meteorological conditions and the characterist ics of the depositing surfac e (Caffrey et al., 1998). Wet deposition fluxes are estimated using: P C F (Equation 2) where F (kg ha-1 yr-1) is the wet deposition flux, C (g m-3) is the concentration and P (m yr-1) is the precipitation rate (Luo et al., 2003). Aerosols Aerosols are defined as a suspension of liquid or solid particles in a gas (McMurry, 2000; Wayne, 2000). They are an extremely important component of our atmosphere; however, they are onl y partially understood. They play important roles in many biogeochemical cycles, acting as reaction sites and as carriers for many sorbed species (Dent ener et al., 1996). Particles also contribute to smog episodes and light and radiatio n scattering (Seinfeld and Pandis, 1998). Aerosols are produced by natural and anthropogenic activities. Natural sources include sea spray, mi neral dust, forest fires and volcanic emissions (Wayne, 2000). Anthropogenic activities predominantly include the burning of fossil fuels and biomasses and contribute mostly to the submicron (less than 1 m diameter) particle concent rations. These particles are primary

PAGE 26

4 pollutants, as they are emitted directly into the atmosphere (Finlayson-Pitts and Pitts Jr., 2000). Secondary pollutant or aerosol formation includes the gas-toparticle conversion of emitt ed primary pollutants, such as the oxidation of SO2 to sulfate compounds. There are two main size classifica tions of particles based on observed modes: (a) fine, those with a diameter less than 2.5 m and (b) coarse, those with a diameter greater than 2.5 m (Seinfeld and Pandis, 1998). This size distinction is important, as the two clas ses differ in their modes of production and removal, chemical composition, deposit ion and optical properti es (Seinfeld and Pandis, 1998). Within the fine particle ca tegory are two distinct modes: (a) the nuclei mode, particles with a diameter of 0.005 to 0.1 m and (b) the accumulation mode, diam eter of 0.1 to 2.5 m. Due to their size, nuclei mode particles compose a very small mass percent of all particulate matter. They are formed from combustion sources and nucleat ion of atmospheric species and are removed from the mode primarily through coagulation with la rger particles. Accumulation mode particles comprise a si gnificant amount of aerosol mass and surface area. These particles are form ed through nuclei particle coagulation and from the growing of existing particles due to the condensation of gases. As indicated by its name, the accumulati on mode has very inefficient removal mechanisms, resulting in long residence times (Seinf eld and Pandis, 1998). Mechanical processes generate the coarse mode particles. These particles have large gravitational settling velocities, allowing them to settle out of the atmosphere in a relatively short amount of time.

PAGE 27

5 Sea Salt Sea salt particles are considered to be the dominant particles originating from natural sources (Roth and Okada, 1998) They are the largest source of tropospheric particulate matter (Weis and Ewing, 1999) and ar e the principal constituent of the coarse particle mode (F itzgerald, 1991). The global emission rate of sea salt is estimated at 1000 to 2000 Tg yr-1 (Jaenicke, 1993). Sea salts, which are primarily composed of NaCl, play an important role in our atmosphere, as they provide a reactive surface for ma ny pollutants. Particulate sulfates can be formed from the gas-to-particle c onversion of biogenically emitted SO2 on sea salt particles. Heterogeneous chemical reactions with acidic species (i.e. HNO3) result in the emission of volatile HCl, creating a source for HCl in the atmosphere (Roth and Okada, 1998). Sea salts are produced at the ocean’s surface by the mechanical wave action and the bursting of entrained air bubbles (Allen et al., 1996; de Leeuw et al., 2001; O'Dowd et al., 1997). The mechanic al generation of ae rosols results in a wide size range of particle s produced from 0.1 to 100 m in diameter (Andreas et al., 1995; O'Dowd et al., 1997). The majori ty of sea salt particles are found in the coarse mode with a diameter greater than 1 m (Plate and Schulz, 1997). The coarse mode makes up over 95% of t he total mass but only contributes to 510% of the total particle number (Seinfeld and Pandis, 1998). Sea salt, as a primary component of ma rine aerosol, can affect the climate by scattering and absorbing radiation (Fit zgerald, 1991). Research has shown

PAGE 28

6 that sea salt particles can be responsible for up to 75% of light scattering by aerosols (Clegg and Toumi, 1998; Murphy et al., 1998). The concentration of sea salt aerosol is a function of wind sp eed, relative humidity and turbulence (de Leeuw, 1986; Lovett, 1978; O'Dowd et al., 1997). The relative humidity in the ma rine boundary layer is often above the deliquescent point of sea salt, the relative humidity point at which the substance absorbs atmospheric moisture to produc e a saturated aqueous solution (Pilinis et al., 1989). The sorbed waters are unbo und, and their quantity is governed by thermodynamic equilibrium. The deliquescent relative humidity for sodium chloride is 76% (Seinfeld and Pandis 1998). Consequently, most sea salt aerosols exist as concentrated aqueous dropl ets of salt. The available surface waters allow for the scav enging of atmospheric gases, chemical transformations and volatilization of products (Erickson III et al., 1999). Sea salts enriched in sulfur (as su lfate) are considered aged sea salts, with sulfur being greater than 20% (by weight) and sodi um plus chloride greater than 60% (Ebert et al., 2000) Freshly emitted sea sa lts are primarily composed of sodium and chloride. These particles are categorized as fresh sea salts, distinguished as having sodium and ch loride making up over 85% of the total mass. Over 90% of a fresh sea salt par ticle is 1:10 (by weight) magnesium and sodium chlorides (Laskin et al., 2002). Processing through reactions with other species results in the replacement of ch loride with nitrate in both salts.

PAGE 29

7 NaCl(aq,s) + HNO3(g) NaNO3(aq,s) + HCl(g) (Reaction 1) MgCl2(aq,s) + 2 HNO3(g) Mg(NO3)2(aq,s) + 2 HCl(g) (Reaction 2) Magnesium and sodium are untouched by these reactions and are therefore used as signatur e elements for particle apportionment. Sea salt particles are primarily processed dur ing the daytime hours, as gaseous NO2 is photochemically converted into reactive gaseous HNO3. Laskin et al. (2002) reported that daytime sea salt particles in Houst on, Texas were found to be almost completely processed, or converted NaNO3 particles. They noted by 8:00 a.m. chloride started to di sappear from the particle co mposition and four to five hours later was completely eliminated. However, nighttime sea salt particles were found to be almost completely unprocessed, and their composition was close to that of seawater. Mineral Dust Mineral dust aerosols are formed from wind-blown soils. It is estimated that 1000 to 3000 Tg of mineral aerosol are emitted annually into the atmosphere (Dentener et al., 1996; Grassi an, 2002). Mineral dusts pl ay an important role as a sink for nitric acid, converting it into a coarse mode species, Ca(NO3)2. CaCO3(s) + 2 HNO3(g) Ca(NO3)2(aq,s) + CO2(g) + H2O(g,aq) (Reaction 3) Dentener et al. (1996) stated, however, that the uptake of HNO3 can take place only in mineral dust aerosols with a high enough alkalinity to overcome the acidity associated with uptake. The alkalinity of the mineral dust is to a great

PAGE 30

8 extent determined by the concentration of CaCO3. Soils in arid regions of the United States can contain 3 to 8% by weight Ca2+ (Dentener et al., 1996). Sodium A significant amount of particulate sodi um exists in our atmosphere, as the “emission sources of sodium are widely spread on the Earth’s surface” (Ooki et al., 2002). Sodium along with chloride is one of the largest trace components of sea spray, which accounts for 36% of t he global sodium emissions (Ooki et al., 2002; Seinfeld and Pandis, 1998). Land-based mineral dust emissions contribute 42% of the total sodium emi ssions (Ooki et al., 2002; Seinfeld and Pandis, 1998). According to the 1987 California Air Resources Board emissions inventory, land-based sodium emissions were estimated at 5 metric tons day-1 (Jacobson, 1997). Land-based sources include paved road dust and diesel emissions. Sodium can be found in the coarse and fine particle modes. Refuse incineration is considered to be the most significant source of fine particle sodium in urban air (Ooki et al., 2002). In J apan, refuse incinerat ion accounted for 2443% of the total sodium emissions and for 79-91% of t he total anthropogenic sodium emissions (Ooki et al., 2002). In marine environments, sodium is primar ily seen in the coarse fraction. It is often used as an indicator for primar y aerosols, as it is a conservative nonreactive species (Ten Brink, 1998).

PAGE 31

9 Chloride The primary source for particulate chloride is sea salt, as NaCl. Landbased sources for particulate chloride include biomass burning, fossil fuel burning, chemical manufacturing and so il dust (Jacobson, 1997). Gaseous chloride is also emitted through c oal combustion (over 98%) and waste incineration (Saxena et al ., 1994). Chloride is mainly distributed in the coarse mode, as NaCl. However, combustion s ources contribute to fine mode chloride. HCl can also be produced through the reaction of nitric acid with sodium chloride (Reaction 1). The production of HCl gas is termed a chloride-depletion process with respect to particulate chloride because chloride changes phase from particle to gas (Cheng et al., 2000). Th is reaction with sea salt plays a role in the halogen release over the marine boundary layer (Aristarain and Delmas, 2002), which may be linked to ozone depl etion (Pryor and Sorensen, 2000). In freshly produced sea salt, Cl-:Na+ molar ratio is 1.17 (Zhuang et al., 1999a). Chloride depletion can be ca lculated using the estimated quantities of chloride compared to the experim ental quantities. % 100 ]) [ 174 1 ( ]) [ ] [ 174 1 ( % Na Cl Na dep Cl (Equation 3) Chloride depletion has been found to decrease wi th increasing particle size, suggesting a surface mechanism for the loss (Jordan et al., 2000). Two factors have been noted that affe ct the extent of chloride depletion. They include

PAGE 32

10 relative humidity and the competition of Ca2+ with Na+ for the acidic gases (Zhuang et al., 1999a). As nitric acid is taken up by the sea sa lt particle, it forc es HCl into the gas phase. This process is due to the Henr y’s constant and the binary activity coefficient of both HNO3 and HCl. The binary activi ty coefficient for dissolved HCl increases exponentially at high molalities, whereas the activity coefficient of dissolved nitric acid remains moderately low at high molalities (Jacobson, 1997; Jacobson et al., 1996). As HCl is re leased or degassed from the sea salt particle, the pH of the aerosol increases due to the loss of H+ (Brimblecombe and Clegg, 1988). Nitrate The primary oxidized reactive nitr ogen compounds of concern are nitric acid (HNO3) and particulate nitrate (NO3 -). These compounds are considered secondary pollutants, as t hey are not directly emi tted into the atmosphere (Blanchard, 1999). Nitric acid is pr oduced through the gas pha se oxidation of NOx (NOX = NO + NO2) (Zhuang et al., 1999b), which is emitted directly into the atmosphere from anthropogenic activities including the combustion of fossil fuels. Natural sources for nitrogen oxi des include soil, volcanic emissions and lightning; they make up less than 10% of the total emissions (Erisman et al., 1998).

PAGE 33

11 During daylight hours, HNO3 is produced by NO2 reaction with a hydroxyl radical (Blanchard et al., 2000; Seinfeld and Pandis, 1998): •OH + NO2 HNO3 (Reaction 4) At night, HNO3 is produced through a multi-step process (Blanchard et al., 2000; Seinfeld and Pandis, 1998): O3 + NO2 •NO3 + O2 (Reaction 5) NO2 + •NO3 N2O5 (Reaction 6) N2O5 + H2O 2 HNO3 (Reaction 7) During atmospheric chemical proce sses, these oxides are transformed into more water-soluble and thermally stable pollutants (nitrates), which are subject to wet or dry atmospheric deposition (Ali-Mohamed and Jaffar, 2000; Mamane and Mehler, 1987; Pilinis and Seinfeld, 1987 ; Pio and Harrison, 1987; Russell and Cass, 1984; Watson et al., 1994). Because there is no primary source for coarse particle nitrate, it is assumed to originate from atmospheric reactions of gaseous nitrogen species with coarse particles (Evans and Poor 2001; Pakkanen et al., 1996b). Sea salt and mineral dust particles play an important role in the incorporation of nitrate aerosols in the coarse m ode (greater than 2.5 m). The oxidized nitrogen species (nitrous and nitric acids) can reac t with NaCl on the surface of sea salt aerosols or CaCO3 on mineral dust to form NaNO3 or Ca(NO3)2 and HCl (Reactions 1 and 3) (Clarke et al., 1999; de Leeuw et al., 2001; Dentener et al., 1996; Evans and Poor, 2001; Goodman et al., 2000; Pakkanen et al., 1996a; Tabazadeh et al., 1998; Ten Brink and Spoelstra, 1998; Zhuang et al., 1999b).

PAGE 34

12 Because NaNO3 is formed on the surface of an existing coarse mode sea salt or mineral dust particle, NaNO3 and Ca(NO3)2 will be collected in the coarse mode (de Leeuw et al., 2001). The produc tion of coarse mode nitrogen from gaseous HNO3 greatly affects its deposition rate Larger particles have a higher “efficiency of precipitational scavenging vi a inertial impaction” (de Leeuw et al., 2001) and a change in gravitational settling velocities, Vd, as compared to gas phase species. This change in deposit ion rate may change current nitrogen loading estimates when accounting for coar se nitrate particles (Torseth et al., 2000). Ammonium Ammonia, NH3, is the predominant alkaline atmospheric gas (Seinfeld and Pandis, 1998) and is a prim ary pollutant. Over 90% is emitted from agricultural practices as animal waste and fertilizer lo ss (Erisman et al., 1998; Seinfeld and Pandis, 1998). Other ammonia sources include landfil ls, wastewater treatment plants, industry and combustion. Industrial sources release ammonia during the manufacturing of ammonia-based products, such as fertilizer. It is also used to make nitric acid and to remove NOX from industrial coal boiler flue gases (Chang et al., 2003). Ammonia is readily absorbed by water and soil surfaces, resulting in a relatively low residence time of appr oximately 10 days (Seinfeld and Pandis,

PAGE 35

13 1998). In 1994, the total global ammonia emissions were estimated at 45 Tg-N yr-1 (Dentener and Crutzen, 1994; Seinfeld and Pandis, 1998). Annual national (U.S.A.) ammonia emissions were es timated in 1997 to be 4.47 Tg-N yr-1 (Gilliland et al.). Both ammonia and the ammonium ion, NH4 +, contribute to nutrient loadings as sources for nitrogen. Proce sses in which this nitrogen is converted into secondary aerosols include the reacti ons of ammonia with nitric and sulfuric acids. Fine particulate nitrate and ammonium are produced from the heterogeneous chemical reactions of nitr ic and sulfuric acids and ammonia, producing ammonium nitrate (NH4NO3) (Reaction 8) and ammonium bisulfate (NH4HSO4) (Reaction 9) or a mmonium sulfate ((NH4)2SO4) (Reaction 10) (Paerl, 2001; Seinfeld and Pandis, 1998; Wall et al., 1988). NH3(g) + HNO3(g, aq) NH4NO3(aq, s) (Reaction 8) NH3(g) + H2SO4(aq) NH4HSO4(aq, s) (Reaction 9) NH3(g) + NH4HSO4(aq, s) (NH4)2SO4(aq, s) (Reaction 10) The ammonium nitrate reaction exists in equilibrium and is reversible. This reaction usually occurs in fogs and clouds. Ammonium nitrate is quite unstable (Watson et al., 1994) and is easily forced back into its reactive gaseous components. Production of NH4NO3 can range from 1 to 90 percent per hour, depending on the time of day and meteorological conditions, mainly temperature and relative humidity (Calvert and Sto ckwell, 1983; Watson et al., 1994). It is important to note that amm onium nitrate is a major secondary component of suspended particles in urba n areas (Watson et al., 1994). With

PAGE 36

14 increasing relative humidity, ammonium nitrate adsorbs water, resulting in particle growth and increased light scattering A wintertime study in Denver, CO, found ammonium nitrate concentrations as high as 28 g m-3, contributing to 180 Mm-1 light extinction (Watson et al., 1994). Ammonium may also be seen in the coarse mode aerosol. Ammonia may neutralize with acidic species on sea sa lt particles, or physical processes may transfer the NH4NO3, NH4HSO4 and (NH4)2SO4 to the coarse mode (Cheng et al., 2000; Yeatman et al., 2001) Sulfate Sulfur dioxide, SO2, is a primary pollutant and is the “predominant anthropogenic sulfur-containing air pollut ant” (Seinfeld and Pandis, 1998). In 1995, global sulfur emissions we re estimated at 98-120 Tg-S yr-1, with approximately 80 Tg-S yr-1 emitted as SO2 (Berresheim et al., 1995; Seinfeld and Pandis, 1998). Sulfuric acid, H2SO4, is produced when SO2 is oxidized in the presence of hydroxyl radicals and water vapor, forming sulfuric ac id droplets (Watson et al., 1994). These droplets are then readily neutralized by ammonia (Reactions 910), forming fine particulate species (A llen and Miguel, 1995). The reaction of ammonia with sulfuric acid occurs in two stages (H arrison, 1993; Mehlmann and Warneck, 1995). In an environment with lo w concentrations of ammonia, sulfate and nitrate compete for available ammoni a. However, sulf ate preferentially

PAGE 37

15 scavenges ammonia and tends to drive th e unstable ammonium nitrate to the gas phase, resulting in very low particulate ammonium nitrate levels. Significant amounts of ammonium nitrate are form ed only when the concentrations of ammonia exceed that of sulfate by a mola r ratio of two or more (Watson et al., 1994). Below this ratio, the aerosol phase will be acidic and sulfate may exist as ammonium bisulfate (Seinf eld and Pandis, 1998). Transformation rates range from 0.01 to 5 percent per hour of the SO2 concentration (Calvert and Stockwell, 1983), being most active during da ylight hours (Watson et al., 1994). Coarse mode sulfate arises from the reaction of sulf uric acid with sea salt or mineral dust particles (Wa ll et al., 1988; Zhuang et al., 1999a). H2SO4(aq) + 2 NaCl(aq,s) Na2SO4(aq,s) + 2 HCl(g) (Reaction 11) H2SO4(aq) + CaCO3(aq,s) CaSO4(aq,s) + H2O(g,aq) + CO2(g) (Reaction 12) The sulfate formed from the reaction with sea salt is termed non-sea salt sulfate, as it was not part of the initial particle origi nating from the sea (Zhuang et al., 1999b). This reaction is also important as it leads to par ticulate chloride depletion. It has also been found that coarse mode sulfate is formed through the heterogeneous oxidation of SO2 by O3 in freshly formed sea salt particles. Because these particles contain water and are alkaline, this process proceeds rapidly. It has been noted that the conversion of SO2 to sulfate is strongly dependent on the amount of av ailable surface waters and water volume of the particle (Chameides and Stel son, 1992; Sievering et al., 1995; Zhuang et al., 1999a).

PAGE 38

16 Particle Diameter Atmospheric particles are usually cat egorized by their diameters, implying the particles are spherical. Many atmo spheric particles, however, have irregular shapes and are not easily ca tegorized by their geometric radii or diameters. Equivalent or effective diameters ar e often used as they depend on physical, rather than geometric, properties (Finlayson-Pitts and Pitts Jr., 2000). The most common effective diameter used is the aerodynamic diameter, Da, which is the “diameter of a spherical particle with a standard density of 1000 kg m-3 (density of a water droplet) that has the same settling velocity as the particle” (Hinds, 1999). The aerodynamic di ameter is used to correct for particle morphology as it standardizes for shape (a sphere), density and settling velocities. It is the primary particle pr operty for characterizing filtration and respiratory deposition (Fin layson-Pitts and Pitts Jr., 2000; Hinds, 1999) and can be calculated using: o p g aD D (Equation 4) where Da is the aerodynamic diameter, Dg is the geometric diameter, p is the particle density, and o is the standard density. Another type of diameter us ed is the Stokes diameter, Ds, which is the “diameter of a sphere having the same density and settling velocity as the particle” (Hinds, 1999). The Stokes diam eter standardizes for settling velocity,

PAGE 39

17 whereas the aerodynamic diameter standar dizes for both settling velocity and density. The term “diameter” in this work r epresents the aerodynam ic diameter of atmospheric particles unless otherwise stat ed. As will be discussed later, the particles in the area of interest were determined to be predominantly in the metastable or deliquescent st ate. Their densities were near that of the standard density, allowing the assumption that the aerodynamic diam eter approximated the geometric diameter. Deposition Velocities The significance in the gas-toparticle conversion of gaseous HNO3 to particle NO3 relates to the deposition velociti es of the substances. The dry deposition velocity of HNO3 is relatively high wih re spect to fine particulate matter. However, since most of the NO3 particles formed in the Tampa Bay area are in the coarse size r ange (diameter greater than 2.5 m), the conversion may have an increased significance. To det ermine if the phase change results in increased or decreased ni trogen loadings to the Ta mpa Bay watershed, the deposition velocities of the fo rmed species were reviewed. Typical dry deposition velocities for gaseous HNO3 are reported as 4 and 1 cm s-1 for over land and over open water, respectively (Seinfeld and Pandis, 1998). Particle dry deposition velocities over water can be calculated using a combination of the NOAA Buoy and Will iams models (Bhethanabotla, 2002; Poor

PAGE 40

18 et al., 2001; Valigura, 2001; Williams, 1982). T he NOAA Buoy model is an iterative bulk exchange model for momentum, heat and moisture; and the Williams model is a two-layer multiple -path model for the dry deposition of particles to surface waters. The comb ined or integrated NOAA Buoy-Williams model includes the effects of wave break ing, particle hygroscopic growth and turbulent heat flux. The model calculates the over water dry deposition velocities on the basis of turbulent heat transfer a nd gravitational settling of particles. Table 1 lists the average over wate r deposition velocity and one standard deviation for the MOUDI ten particle size ranges using May 2002 meteorology. Particle Range ( m) Geometric Mean Dp50 ( m) Average Deposition Velocity (cm s-1) 18-30 23 1.7 0.17 3.2-18 7.6 0.21 0.12 1.8-3.2 2.4 0.02 0.01 1.0-1.8 1.3 0.01 0.01 0.56-1.0 0.75 0.01 0.01 0.32-0.56 0.42 0.01 0.01 0.18-0.32 0.24 0.01 0.01 0.10-0.18 0.13 0.02 0.02 0.056-0.10 0.075 0.03 0.03 0.01-0.056 0.024 0.07 0.07 Table 1. Average over water deposition ve locities and their respective standard deviations for the MOUDI ten particle size ranges calculated using May 2002 meteorological data. The geometric mean of each si ze bin was used as the Dp50 for the model input. The geometric mean can be calculated by:

PAGE 41

19 n nx x x Mean Geometric ) )...( )( (2 1 (Equation 5) Sampling It is essential to collect accurate measurements of atmospheric gases and particulate species in order to under stand the processes occurring in our environment. These species are difficult to co llect, as they exist at trace levels in multiple phases. Many species are partitioned between the gaseous and particulate phases. Numerous methods have been developed to facilitate collection of atmospheric s pecies. They include filtration (Allen et al., 1989; Appel et al., 1980; Okita et al., 1976; Spicer et al ., 1982), diffusion denuders (Forrest et al., 1982; Harrison and Ki tto, 1990; Possanzini et al., 1983) and spectroscopy (Platt et al., 1980). Filtration Methods Filtration sampling is the most widely used technique for sampling atmospheric gaseous and particulate specie s due to its low cost and simplicity (Kitto and Colbeck, 1999). Many types of filters have been used, the most popular being glass fiber, quartz fiber, PTFE (polytetrafluoroethylene) Teflon membrane and nylon filters. PTFE Teflon is a choice of many scientists, as it is inert towards chemical species (Kitto and Colbeck, 1999) and contains low blank analyte levels. Quartz fiber filters are ty pically used for organic sampling, as they

PAGE 42

20 can be heated to remove blank analytes. Ny lon filters have the ability to adsorb acidic gaseous species, making them id eal for nitric and hydrochloric acid collection (Dasch et al., 1989; Grosjean, 1982; Hering et al., 1988; Spicer et al., 1982). In addition to these filters, impregnated filters may also be used. Nitric acid has been collected using sodium chlori de(Appel et al., 1980; Forrest et al., 1980; Karakas and Tuncel, 1997), potassi um carbonate(K im and Allen, 1997) and tetrabutyl ammonium hydroxideimpregnated filters (Huebert and Lazrus, 1980). SO2 has been collected using potassium hydroxide(Nodop and Hanssen, 1986) and various carbonate(Karakas a nd Tuncel, 1997; Kim and Allen, 1997) impregnated filt ers. Ammonia has been co llected using a variety of acid-impregnated filters, including citric oxalic and phosphoric acids (Harrison and Kitto, 1990; Karakas and Tuncel, 1997; Masia et al., 1994). Filters are deployed in the field fo r atmospheric gaseous and particulate species collection in a filter pack. Filt er packs can be set up using a single filter or multiple filters, allowing for the colle ction of multiple s pecies in both the gas and particle phases. A multiple filter setup is used to collect and determine the volatilization of particulates, gas-particl e reactions and particle-particle reactions (Kitto and Colbeck, 1999). Under changing temperature, relati ve humidity and acidity, unstable species, such as ammonium nitrate and am monium chloride, ma y volatilize into their parent gases: NH4NO3(s) NH3(g) + HNO3(g) (Reaction 13) NH4Cl(s) NH3(g) + HCl(g) (Reaction 14)

PAGE 43

21 Gas-particle artifacts can form from the interaction of incoming reactive gases with pre-collected aerosols. Exam ples of these types of interferences include: 2 NH3(g) + H2SO4(aq) (NH4)2SO4(aq,s) (Reaction 15) 2 HNO3(g) + CaCO3(aq,s) Ca(NO3)2(aq,s) + H2O(g,aq) + CO2(g) (Reaction 16) HNO3(g) + NaCl(aq,s) NaNO3(aq,s) + HCl(g) (Reaction 17) Particle-particle interactions may result in the loss of a species of interest. Examples of this type of reaction include: NH4NO3(s) + H2SO4(aq) NH4HSO4(aq,s) + HNO3(g) (Reaction 18) 2 NaCl(aq,s) + H2SO4(aq) 2 HCl(g) + Na2SO4(aq,s) (Reaction 19) All of these processes can result in a sampling bias, overor underestimation, of the measurements. To mi nimize these effects, filter packs are often coupled with denuder samplers, which remove the reactive gas phase species prior to the filter pack. Like the previous filter pack se tup, multiple filters can be used in conjunction with denuders. Denuder Methods In denuders, atmospheric gaseous components are removed from the airstream by diffusion to the walls of the instrument, l eaving the particulate matter untouched. There are several types of denuders including cylindrical, annular, honeycomb and others. In this work, annular denuders we re used. These consisted of concentric etched glass tu bes contained within a Teflon-coated

PAGE 44

22 stainless steel cylindrical housing. Th e walls of the denuders are coated with a solution designed to react with the species desired to be co llected. In order to be efficient, a denuder must maintain stable and laminar gas flow, ha ve an infinitely large collection capacity wi th a large sink for the sp ecies, and not generate or destroy the species of interest in side the denuder (Kitto and Colbeck, 1999). A typical configuratio n of an annular denuder system (ADS) includes an inlet, a series of coated denuders and a filt er pack. There ar e many types of inlets, including cyclones and elutriators, and they are generally Teflon-coated. When using a cyclone inlet in marine environm ents, the interior wall of the ADS cyclone may become coated by s ea salt particles. It ha s been found that sea salt aerosols react with acids in the atmosphere, such as H2SO4 and HNO3, forming sulfates and nitrates (Li-Jones et al., 2001). The reaction between HNO3 and sea salts within the cyclone may resu lt in a substant ial loss of HNO3 (Appel et al., 1988; Vossler et al., 1988). This reacti on leads to the mis apportionment of both nitrate and chloride ions, as HCl is rele ased into the vapor phase. Under marine conditions, it has been sugges ted that the ADS be used with a different type of inlet device such as a small impactor (Li-Jones et al., 2001). In an impactor, large sea salt particles will be deposited within a relatively small area on the impaction surface, making the area of sea sa lt exposed to the sample air stream smaller. Because they we re readily available, 2.5 m Teflon-coated cyclones were chosen for the ADS assembly inlet device during the course of these experiments. These cyclones were cleane d on a regular basis to prevent the buildup of sea salt.

PAGE 45

23 As mentioned, the wa lls of the denuders should be coated with solutions appropriate for the target species to be collected. Typical species collected include HCl, HNO2, HNO3, SO2 and NH3. The first four species are acidic and are easily collected using basic coating solutions. Ammonia sampling uses acidic coating solutions. The criteria for choosing the proper denuder coating are: (1) the coating must be very select ive, (2) it must be a good sink for the species to be determined, (3) it must hav e high collection efficiency, and (4) it must be unreactive towards products resu lting from the prim ary reaction between species and the coating layer (Perrino et al., 1990). In previous work, a coating solution of NaCl or NaF was used for the collection of HNO2 and HNO3 (Allegrini et al., 1987; Allegr ini et al., 1994), with a collection efficiency for HNO3 greater than 97%. Ther e are several problems associated with using NaCl. First, excess chloride from the coating solution in ion chromatography (IC) analysis can over whelm the analytical column, creating a masking peak that inva lidates the nitrite and nitrate analysis (from HNO2 and HNO3, respectively). Second, when HNO3 reacts with NaCl, gaseous HCl is formed. As a result, atmospheric HCl will be mixed with the coating solution reaction byproduct HCl and, therefore, cannot be quantifi ed in a subsequent denuder in series. Third, the collection efficiency for HNO2 on NaCland NaFcoated denuders is poor. It has been shown that HNO2 was almost entirely released from those denuders; only to be captured by the subsequent Na2CO3coated denuder (Perrino et al., 1990).

PAGE 46

24 Na2CO3and glycerol-coated (1% + 1% w/w in 50:50 methanol-water solution) denuders can be used for the collection of SO2, HNO2, HNO3 and HCl (Allegrini et al., 1987; Allegrini et al., 1994; Vossler et al., 1988). Perrino et al. (2001) determined the collection efficiency for SO2 collection to be greater than 99.9% of incoming SO2 with deposition of particulate SO4 2at 0.5-2%. Breakthrough was tested using a second backup Na2CO3-coated denuder. The backup denuder showed collecti on of less than 0.1% of SO2. Na2CO3 has been proven to be a good sink for HNO2, with a removal constant near infinity. Difficulty arrives in the quantification of NO2 (from HNO2) when collecting HNO3 on the same denuder. HNO3 is collected and analyzed as NO3 -, whereas HNO2 is partially oxidized during sa mpling and is analyzed as both NO2 and NO3 -. Exposure to gaseous NO can cause an increase in NO2 formation on the Na2CO3-coated denuders. This interferent am ount can be taken into account by using the differential technique (F ebo et al., 1989; Perrino et al., 1990). For these experiments, Na2CO3 was used as the bas ic denuder coating. Collection efficiencies have been determi ned to be greater than 99.5% for HCl, greater than 98.5% for HNO2, greater than 97% for HNO3 and greater than 99% for SO2 on Na2CO3-coated denuders (Allegrini et al ., 1987; Perrino et al., 2001). Acid-coated denuders are used to determine ambient ammonia concentrations. The following acids were considered: citric, oxalic, phosphoric and phosphorus. The collection efficienci es of oxalic acid, citric acid and phosphorus acid (H3PO3) have been compared (Perri no and Gherardi, 1999). The data indicated the collection efficiencie s of oxalic and phos phorus acid were

PAGE 47

25 equivalent (>99%). However, the efficiency of citric acid had great variation from the values that were expected for a per fect sink, with a negative bias of approximately 20% (McCulloch and Shendrikar, 2000). Phosphorus and oxalic acids yielded very good r eproducibility in determini ng ammonia. The drawbacks of oxalic acid are its to xicity and desorbing potentia l. Perrino and Gherardi (1999) found significant amounts of oxalat e releasing from the denuder into the incoming air stream and displacing nitr ate ion on the Teflon filter, causing an excess of nitrate ions on the backup filt er. Their conclusion indicated that both oxalic and phosphorus acid pr oved to be suitable for ammonia determination, but only the latter is able to a ssure a correct determination of ammonium salts on the backup filter packs. After direct discussion with Dr. Perrino, it has concluded that the performances of phosphoric and phosphorus acid are very similar. In these experiments, both citric and phosphoric acid coating solutions were used. Cascade Impactors Many instruments have been devel oped for the co llection of PM10, particulate matter with an aer odynamic diameter less than 10 m. Cascade (or inertial) and virtual impactors are the met hods of choice for collecting particles 10 m and smaller (Spurny, 1999). Both devices size segregate particles based on their aerodynamic diameter. The microorifice uniform deposit impactor (MOUDI) and Andersen Mark II sampler are exampl es of cascade impactors. Virtual

PAGE 48

26 impactors are used in instruments like t he dichotomous or trichotomous sampler, where only two or three size frac tions, respectively, are collected. Impaction methods have been used for pa rticle collection for quite some time. In the early twentieth century, dust particles were collected through impaction to evaluate occupational env ironments (Hinds, 1999). Since the 1960s, cascade impactors have been develope d to measure the particle size distributions. The collected particle mass distributions were determined by weighing the impaction plates before and after sampling. All cascade (or inertial) impactors are composed of two basic parts: an impaction nozzle and an impaction plate. Together, they make up a collection stage. Cascade impactors c an have multiple stages, allowing for particles to be collected in multiple size bins. The airstream enters the impactor through the impaction nozzle, or jet, which directs t he airstream towards the flat impaction plate. The plate is placed perpendicular to the airstream, deflecting the airflow to a near 90 bend (Figure 1). Due to the inertia and size, some particles are unable to follow the bend in the airflow and collide or impact on the impaction plate, whereas the smaller particles fo llow the bend in the airstream and are not collected on the impaction pl ate. As a result, the impactor separates particles by size. The larger particles are remov ed from the airstream and collected on the impaction plates, leaving the smaller particl es in the remainder of the airflow to be collected on latter stages.

PAGE 49

27 Jet Streamline Impaction Plate Nozzle Jet Streamline Impaction Plate Nozzle Jet Streamline Impaction Plate Nozzle Jet Streamline Impaction Plate Nozzle Jet Streamline Impaction Plate Nozzle Figure 1. Cross-sectional view of an impactor stage (adapted from Hinds, 1999). Impactors have been used in parallel to collect particles in multiple size bins. However, this is uncommon due to the complexity of maintaining multiple flow rates required. The use of impacto rs in series has been developed and is commonly practiced to collect multiple si ze bin particles. Stages are arranged in order of decreasing cutoff di ameters, collecting the lar gest particles first. The device is referred to as a cascade impactor, with each separate unit called an impactor or collection stage. The smalle r particles are collected by continually decreasing the nozzle diameters, keeping th e volumetric flow rate the same for all stages. Each stage in a cascade impa ctor is fitted with a plate, and a final backup collection filter is pl aced at the exit of the in strument to collect those remaining particles. When in operation, each collection plate may be used as is, coated, or have a filter placed on top. Uncoated plates are used for the collection of liquid

PAGE 50

28 particles as they stick when impacted ont o the collection plate. Hard, solid particles have a tendency to bounce on ce impacted and end up collected in a lower stage. To reduce particle bounce, impaction stages may be coated with a thin film of oil or grease. Filters are used for the collection of both liquid and solid particles and are placed on top of the co llection plate. T hey are preferred because they minimize sample contamin ation by being easily removed and replaced without the requirem ent of thorough cleaning of each impaction plate. The Stokes number parameter, St, predicts whether a particle will impact or follow the airstream out of the impaction region (Hi nds, 1999; Marple et al., 1991). The Stokes number can be defined as: W D V C Stp o p 92 (Equation 6) where p is the particle density, C is the slip correction factor (also known as the Cunningham correction factor, Cc), Vo is the average air velocity at the nozzle, Dp is the particle diameter, is the air viscosity, and W is the nozzle diameter. Vo can be determined using: 24 W q Vo (Equation 7) where q is the volumetric flow rate through the nozzle. The Stokes number is the primary parameter that governs the size of particl es collected in an inertial impactor. Each collection stage in a cascade impactor has a cut point diameter, in which 50% of the particles greater than a certain si ze are collected and those smaller continue with the ai rstream. The Stokes number is used to characterize

PAGE 51

29 the collection efficiency of each impacti on stage (Hinds, 1999). The cut points for each impaction plate are defined by the particle diameter that is collected with 50% efficiency, Dp50. 50 509 St V C W Do p p (Equation 8) where 50St is the dimensionless particl e size defining the value of St corresponding to Dp50 (Hinds, 1999). Particle concentrations collected us ing a cascade impactor are calculated by dividing the mass collected by the volume of air sampled. total i Stg i StgVolume Mass PM) ( ) ( (Equation 9) where PMStg(i) (g m-3) is the particulate concent ration from stage (i) of the impactor, MassStg(i) (g) is the mass collected and Volumetotal (m-3) is the volume of air through the impactor. Virtual Impactors Virtual impactors are used in t he dichotomous sampler to separate particles into two size fractions. Unlik e a traditional impac tor, particles are collected on filters the air is drawn th rough, instead of im pacting on a plate. Ambient air is drawn through the virtual impactor and is split into major and minor airstreams. During typical operation of a dichotomous sampler, the total flow rate

PAGE 52

30 is 16.7 L min-1, with respective 15.0 and 1.67 L min-1 directed onto the PM2.5 and PM10-2.5 filters. Particles are separated in a similar fash ion to that of a traditional impactor. Larger particles, those with enough inertia, are drawn into a collection probe and are carried by the minor flow. Smaller par ticles are carried away from the nozzle and continue with the major flow stream. All particles are collected onto a filter through which the air flows. Due to the complexity of controlling multiple more airflows, virtual impactors are used for the collection of one or two stages. A trichotomous sampler has been developed us ing virtual impactors, allowing the collection of particles in three size bins. The traditional cutpoint for a dic hotomous sampler is 2.5 m, naming the collective PM2.5 bin “fine” and the PM10-2.5 bin “coarse”. PM2.5 concentrations can be calculated using 5 2 5 2 5 2Volume Mass PM (Equation 10) where PM2.5 ( g m-3) is the concentration of particulate matter, Mass2.5 ( g) is the collected mass and Volume2.5 (m-3) is the volume of air (Poor et al., 2002). The PM10-2.5 concentrations are calculated using total totalVolume Volume PM Volume Mass PM5 2 10 5 2 5 2 10 5 2 10 (Equation 11) where PM10-2.5 ( g m-3) is the concentration of particulate matter, Mass10-2.5 ( g) is the collected mass and Volumetotal and Volume10-2.5 (m-3) are the volumes of air (Poor et al., 2002). This equation serves to correct for the PM2.5 collected on the PM10-2.5 filter.

PAGE 53

31 Models Models are tools used to “provide t he necessary framework for integration of our understanding of indi vidual atmospheric proce sses and study of their interactions” (Seinfeld and Pandis, 1998). Models are need ed because of the complexity of the atmosphere, with numerous processes occurring simultaneously. Atmospheric sampling provides a “snapshot of atmospheric conditions at a particular time and lo cation” (Seinfeld and Pandis, 1998) and helps identify the state of the atmos phere. Models are us ed to extend our snapshot to understand processes on a local or regional scale. Thermodynamic models have been devel oped over the past twenty years using thermodynamic equilibrium principl es to predict the composition and physical state of atmospheric aerosols The basis for these models is the assumption that gas and aer osol volatile species are in equilibrium (Ansari and Pandis, 1999; Bassett and Seinfeld, 1983; Pi linis and Seinfeld, 1987; Stelson and Seinfeld, 1982). Current thermodynam ic equilibrium models include EQUIL (Bassett and Seinfeld, 1983), KEQUIL (B assett and Seinfeld, 1984), MARS (Saxena et al., 1986), SEQULIB (Pilinis and Seinfeld, 1987), AIM (Wexler and Seinfeld, 1991), SCAPE (Kim et al., 1993a; Kim et al., 1993b), SCAPE2 (Kim and Seinfeld, 1995; Meng et al., 1995), MA RS-A (Binkowski and Shankar, 1995), EQUISOLV (Jacobson et al., 1996), AI M2 (Clegg et al., 1998), ISORROPRIA (Nenes et al., 1998), GFEMN (Ansari and Pandis, 1999) and EQUISOLV II (Jacobson, 1999a). During the developm ent of many of these models,

PAGE 54

32 maintaining computer efficiency wa s a primary endeavor. As a result, assumptions were made to simplify pr oblems and allow the use of these equilibrium models in atmospheric chem ical transport models (Ansari and Pandis, 1999). Most of these equilibrium models solv e for equilibrium using the iterative Gibbs free energy minimization method (J acobson, 1999a). The following table taken from Jacobson 1999 summarizes the equilibrium models, the system solved and their solution method.

PAGE 55

33 Model Name Reference System Solved Solution Method AIM Wexler and Seinfeld (1991) NH4 +-Na+-NO3 --SO4 2--ClIterative Gibbs free energy minimization method AIM2 Clegg et al. (1998) NH4 +-Na+-NO3 --SO4 2--ClIterative Gibbs free energy minimization method EQUIL Bassett and Seinfeld (1983) NH4 +-NO3 --SO4 2Iterative Gibbs free energy minimization method EQUISOLV Jacobson et al. (1996)NH4 +-Na+-NO3 --SO4 2--ClMass-flux iteration method EQUISOLV II Jacobson (1999a) NH4 +-Na+-NO3 --SO4 2--Cl-Ca2+-Mg2+-K+-CO3 2Analytical equilibrium iteration + mass-flux iteration GFEMN Ansari and Pandis (1999) NH4 +-Na+-NO3 --SO4 2--ClIterative Gibbs free energy minimization method ISORROPRIA Nenes et al. (1999) NH4 +-Na+-NO3 --SO4 2--ClIterative bisection + bisectionNewton for H+ KEQUIL Bassett and Seinfeld (1984) NH4 +-NO3 --SO4 2Iterative Gibbs free energy minimization method MARS Saxena et al. (1986) NH4 +-NO3 --SO4 2Iterative Newton-Raphson method MARS-A Binkowski and Shankar (1995) NH4 +-NO3 --SO4 2Iterative Newton-Raphson method SCAPE Kim et al. (1993) NH4 +-Na+-NO3 --SO4 2--ClIterative bisection + bisectionNewton for H+ SCAPE2 Kim and Seinfeld (1995), Meng et al. (1995) NH4 +-Na+-NO3 --SO4 2--Cl-Ca2+-Mg2+-K+-CO3 2Iterative bisection method SEQUILIB Pilinis and Seinfeld (1987) NH4 +-Na+-NO3 --SO4 2--ClIterative bisection method Table 2. Equilibrium models, specie s treated and numerical method used to solve equilibrium (Jacobson, 1999a). The primary thermodynamic equilibri um model used throughout this work was the EQUISOLV II (Jacobson, 1999a), which solves for thermodynamic equilibrium through analytical equilibrium it erations and mass flux iterations. The model can be used over a range of tem peratures as it corrects equilibrium constants, deliquescent relative humid ities and activity coefficients for

PAGE 56

34 temperature dependence. This model is very versatile as it can be used for NH4 +, Na+, NO3 -, SO4 2-, Cl-, Ca2+, Mg2+, K+ and CO3 2systems. The Aerosol Inorganics Model (AIM2 Model III) (Clegg et al., 1998) was used for model comparison. AIM2 solv es for thermodynamic equilibrium by minimizing the Gibbs free energy of t he system through the use of sequential quadratic programming algorithms. This model is limited to a si ngle particle bin, NH4 +, Na+, NO3 -, SO4 2and Clsystems and 25C environments. Kinetics Heterogeneous Reactions Heterogeneous chemical reactions are those occurring between gases and either solids or liquids in the atmosphere (Finlayso n-Pitts and Pitts Jr., 2000). These reactions occur on many different ty pes of surfaces including ice crystals, liquid aerosols, sea salts, soot, metal oxides, clouds and surface waters as well as a number of other surfac e types (Kolb et al., 1995). There are a few common term s used to describe heterogeneous reactions. The surface reaction probability rxn is the fraction of collisions between the gas and condensed phases that l eads to the irreversible uptake of the gaseous species due to chemical reaction. rxn is also known as the reaction probability. The mass accommodation coefficient is the fraction of collisions between the gas and condensed phases that re sult in the uptake of the gaseous

PAGE 57

35 species by the condensed phase. It is t he fundamental parameter that measures the rate at which molecules cross the interface between the gas and condensed phases (Kolb et al., 1995). The mass acco mmodation coefficient is not the net uptake because it does not include the reve rsible effect of evaporation of the gaseous species from the condensed phas e; however, it determines the maximum rate of mass transport. is also referred to as the sticking coefficient for the uptake on solid surfaces. T he overall or net uptake probability net is the net rate of uptake normalized to the rate of collisions. When determining uptake coefficients experimentally measured net measured (Finlayson-Pitts and Pitts Jr., 2000). Several types of instrumentati on have been used to experimentally determine heterogeneous chemis try kinetics. The most common instrumentation is the Knudsen cell (Caloz et al., 1997; Fenter et al., 1997; Finlayson-Pitts and Pitts Jr., 2000). It has been used over t he last 30 years for kinetic measurements of both homogeneous and heterogeneous systems. According to Caloz et al. (1997), the technique is well adapted fo r heterogeneous reactions because the reactant interacts with the substrate without boundary layer effects and because the collision frequency between the gas eous reactant and substrate can be accurately determined. The Knudsen cell is comprised of a chamber containing a reactive surface through which the reac tant gas is allowed to pass. It is typically coupled with a mass spectrometer, which detects the gas phase concentration. The condensed phase is l oaded on a sample plate, which is either placed in a separat e chamber or covered with a lid, depending if the

PAGE 58

36 Knudsen cell is dual or single chambered The reactant gas enters the cell through a valve where it evenly disperses. The gas is often allowed to flow through the system prior to exposure to the condensed phase for some time to minimize loss of the reactant gas to t he cell walls. When the reactant gas is exposed to the condensed phase, the su rface takes up the gas reducing the concentration of the gas in the cell. Th is change in gas concentration is detected by the mass spectrometer, allowing the net uptake of the gas by the surface to be determined. The major limitation for Knudsen cells is that they are operated at low pressures (<10 mTorr) to increase the mean free path of the gas molecules (Davies and Cox, 1998; Finlayson-Pitts and Pitts Jr., 2000). Reactions can be studied only under dry and very low relative humidity conditions. This limits environmental applications, as the humidity in coastal ar eas is often greater than 60%. Other types of instrumentation hav e been developed whic h are not limited to low relative humidity conditions. The aerosol flow tube has been used to determine rate constants for gas-phase r eactions (Finlayson-Pitts and Pitts Jr., 2000), and it has been recently adapted to study heterogeneous reactions (Abbatt and Washewsky, 1998; Hu and Abbatt, 1997). The aerosol flow tube can be used in one of two ways. The walls of the flow tube can be coated with the condensed phase of interest. However, it is often difficult to coat the walls of t he flow tube. In the second method, the condensed phase is generated as an aerosol that can be aqueous or solid and

PAGE 59

37 passed through the flow tube along with the reactant gas phase. The exit orifice of the flow tube is connected to a detec tion device, which is typically a mass spectrometer. The exiting gas is analyze d in a similar fashion to that in a Knudsen cell apparatus. The aerosol flow reactor is not constrained to low pressure, and it is very versatile for a wi de range of relative humidity conditions. Sea Salt Understanding the kinetics behind the uptake of nitric acid by sea salt is essential in modeling the gas phase nitric acid to particulate nitrate conversion and its effects on nitrogen deposition. Past studies have focused on determining the uptake of nitric acid by NaCl, a ma jor component of s ea salt (77% w/w) (Beichert and Finlayson-Pitts, 1996; De Haan and Finlayson-Pitts, 1997; Ghosal and Hemminger, 1999; Zangmei ster and Pemberton, 1998; Zangmeister and Pemberton, 2001). However, other res earch has suggested that the reaction probabilities, the frac tional loss of a species from the gas phase due to reaction with a surface (Jacobson, 1999b), for sea salt are at least an order of magnitude faster than those for sodium chloride. Other measurements have suggested that NaCl may not be the most reactive component of sea salt (De Haan and Finlayson-Pitts, 1997; Langer et al., 1997) Under most marine conditions, sea salt is present in the form of deliquescent aerosol particl es (Guimbaud et al., 2002a). However, the majority of laboratory studies investi gating the kinetics behind the HNO3 and sea salt

PAGE 60

38 interaction were performed on solid crystals or thin films at low relative humidity conditions. A model describ ing the step-by-step process of the uptake of nitric acid and conversion to nitrate has been proposed (Laskin et al., 2002). HNO3 and other nitrogen oxide gases are first taken up into the surface adsorbed water of a sea salt particle. Even at conditi ons well below the deliquescent point at ambient temperature, NaCl (and sea salt) has sign ificant amounts of adsorbed water on its surface, which is associat ed with its steps, edges and defect sites (Beichert and Finlayson-Pitts, 1996; Davi es and Cox, 1998; Finlayson-Pitts and Hemminger, 2000). Nitric acid dissolves and remains inactive until the particle becomes significantly acidifi ed. This happens by the further uptake of acids or the drying of the particle. As it dries, the particle bec omes more concentrated in all components, including acids. When th e acidity reaches ~pH 1.7, HCl begins to degas, NaNO3 precipitates and additional HNO3 is taken up (Beichert and Finlayson-Pitts, 1996; Finlayson-Pitts and Hemminger, 2000). A thermodynamic model (Clegg et al., 1998) predicts that HCl evaporates faster than HNO3, leaving the dried particles deficient in chloride rather than nitrate. The reaction between HNO3 and NaCl has been wi dely studied (Beichert and Finlayson-Pitts, 1996; FinlaysonPitts and Hemminger, 2000; Langer et al., 1997; Ten Brink, 1998; Vogt and Finl ayson-Pitts, 1994; Zangmeister and Pemberton, 1998). NaCl is a hygroscopic salt; its deli quescent relative humidity (DRH) is 75% (Seinfeld and Pandis, 1998). Below the deliquescent point, the aerosol exists as a metastable supe rsaturated liquid droplet until the efflorescence point, or crystallization rela tive humidity, is reached (~37% RH), at

PAGE 61

39 which point the salt begins to crysta llize (Davies and Cox, 1998). Studies have been done at different relative humidities to determine the effect of water on the HNO3 absorption. At 30% relative humidit y, dry aerosol NaCl was reacted with gaseous HNO3. Reaction analysis revealed the absence of measurable substitution of chloride by ni trate, providing evidence that HNO3 does not react with dry NaCl (Ten Brink, 1998). However, this absence of nitrate detection does not provide enough evidence to exclude the possibility of a small surface interaction of HNO3 with NaCl (Ten Brink, 1998; Vo gt et al., 1996). It is possible that the detection limits were too high at the time of experimentation. The experiment was repeated at 80% relative hu midity, and it result ed in a substantial amount of nitrate format ion (Ten Brink, 1998). The experiment was repeated again, this time at 50% RH (below the 75% DRH of NaCl). The formation of nitrate was observed, contr adicting the results reported at 30% RH. After further analysis, it was concluded that these parti cles were formed fr om NaCl droplets that were “dried”. The aerosols, in ac tuality, were not thoroughly “dry” aerosols but supersaturated droplets due to the h ysteresis effect of aerosol water (Seinfeld and Pandis, 1998; Tang, 1980). Evidence suggests that the reaction pr obability significantly increases with the presence of adsorbed surface water (Beichert and Finlayson-Pitts, 1996; Langer et al., 1997). The water layers crea te a quasi-liquid la yer allowing the surface molecules to deliquesce and incr ease ionic mobility. This greatly facilitates the uptake and reaction of HNO3 with the NaCl solution (Zangmeister and Pemberton, 1998). NaNO3 is formed on the surface of the salt particle.

PAGE 62

40 Without the further uptake of additional water molecules, the reaction rate of additional NaNO3 formation is greatly decreased or stops due to the formation of a protective layer or film on the particle’s surface (A llen et al., 1996). Upon the exposure to increased relative humidit y, the particle adsorbs additional water molecules, increasing the ionic (nitrate ) mobility (Finlayson-Pitts and Hemminger, 2000). The components of the particl e are then allowed to rearrange, regenerating the NaCl reactive surface. A two-step model has been developed representing this process: HNO3(g) + NaClsite NO3 (ads) + HCl(g) (Reaction 20) NO3 (ads) + H2O(l) NaClsite + NaNO3(s) (Reaction 21) where NaClsite is the unreacted surface site and NO3 (ads) is the newly formed immobile NaNO3 that is blocking the NaCl reac tive site (Finlayson-Pitts and Hemminger, 2000; Ghosal and Hemminger, 1999). The availability of surface waters on NaCl appears to govern the uptake of HNO3 (Beichert and Finlayson-Pitts, 1996; De Haan and Finlayson-Pitts, 1997; Ten Brink, 1998). Beichert (1996) did not observe the uptake of HNO3 on single crystals. He proposed the ex planation that a single crystal is relatively free of defects and does not readily hold adso rbed water. The measured dry HNO3 uptake by a NaCl single cr ystal was approxim ately two orders of magnitude less than that on finely ground NaCl powders. The uptake coefficient for a given reaction is the probability that a molecule is removed from the gas phas e upon colliding with a particle surface (Guimbaud et al., 2002a). It is “simply a measure of how likely the molecule will

PAGE 63

41 be taken up by the surface, through eit her adsorption or reaction, on a per collision basis” (Grassian, 2002). The uptake coefficients for gaseous HNO3 on NaCl and sea salt have been determined experimentally. For Na Cl, the uptake coefficient was characterized by the value of 2 0 (Abbatt and Washewsky, 1998). This value was determined using the aerosol kine tics flow tube technique at room temperature with deliquescent NaCl at 75% relative humidity. The rate of uptake was determined to be independent of particle size for deliquescent particles (Guimbaud et al., 2002b; Ten Brink, 1998). Abbatt and Waschewsky (1998) report that this elevated uptak e coefficient is driven by the very high solubility of HNO3 in aqueous salt solution but limited by the gas-diffusion rate. Other uptake coefficients for HNO3 on NaCl have been reported, ranging from 10-4 to 10-2, depending on the amount of water on the salt surface (Abbatt and Washewsky, 1998; Beichert and Finlays on-Pitts, 1996; Fenter et al., 1994; Laux et al., 1994; Vogt and Finlayson-Pitts, 1994). Lo w-pressure Knudsen cell flow reactors are other tools used to determine heterogeneous kinetics of these types of interactions. However, the Knuds en cell flow reactor is limited to very low relative humidity conditions (Davies and Cox, 1998). Beichert and FinlaysonPitts (1996) determined a constant steady-state uptake of 210 6 0 4 1 for particles 0.5 and 4 m at room temperature. F enter et al. (1994) determined a value of 210 3 0 8 2

PAGE 64

42 The value determined by Abbatt and Waschewsky (1998) was used in the modeling and is considered a lowe r limit for the uptake of HNO3 by NaCl as the average relative humidity for the Tampa Bay coastal area is approximately 80%. For sea salt, the uptake coefficient was estimated to be 20 0 50 0 for deliquescent sea salt at 55% relative humidity (Guimbaud et al., 2002a). This value was also determined using an aerosol flow tube technique. The difference between the coefficients for sea salt (which is primarily NaCl) and NaCl lie in the composition of sea salt. Sea salt c ontains hygroscopic hydrates, such as MgCl2•6H2O, which provide additional surface waters for the uptake and reaction of HNO3 (De Haan and Finlayson-Pitts, 1997). Other uptake coefficient values for HNO3 on sea salt have been reported. Knudsen cell flow reactor studies resulted in a steady-state HNO3 uptake rate of 2 0 (De Haan and Finlayson-Pitts, 1997). Mineral Dust The irreversible reactions betw een nitrogen oxides and mineral dust surfaces have been investigated (Dent ener et al., 1996; Grassian, 2002). The research indicated that mineral dust aer osols provide an important sink for HNO3 (Goodman et al., 2000), and these particles may have a significant impact on the chemistry of the tropos phere (Dentener et al., 1996). Mineral aerosols are composed of metallic and nonmetallic oxides, silicates and carbonates.

PAGE 65

43 Carbonates are of special interest as they may be especially effective in removing gaseous HNO3 from the atmosphere (Reaction 3). Grassian (2002) studied the uptake of gaseous HNO3 by CaCO3 using in situ FTIR spectroscopic methods. Her re sults indicate the dependence of water vapor on the HNO3 uptake. Under dry co nditions, exposure of CaCO3 to HNO3 resulted in very little changes to the so lid particle. However, when the CaCO3 particle was exposed to HNO3 in the presence of 20% relative humidity, several changes were noted. Spectral absorpti on bands indicated the presence of the formation of Ca(NO3)2, the formation of gas-phase CO2 and the presence of OH stretching vibrations from the adsorbed water. T he quantity of adsorbed water increased with the increased Ca(NO3)2 formation. This is related to the solubility and hygroscopicity of the particle; calc ium nitrate is more hygroscopic and approximately 100 times more soluble t han calcium carbonate (Lide, 1991). The formation of Ca(NO3)2 allows more water to adsor b onto the particle surface. Morphological studies of the surfac e of the calcium carbonate particles have been conducted using transmission electron microscopy. Unreacted calcium carbonate particles have a smoot h shape. However, those particles exposed to nitric acid at 20% RH had irregular shape with jagged edges. Smooth CaCO3 particles were seen to become more irregular with increased exposure to nitric acid. The jagged edges also resulted in an increase in the surface area of the particle. Similar studi es were conducted using NaCl crystals. Upon exposure to HNO3, the micrographs indicated a physical change in

PAGE 66

44 morphology indicating the formation of NaNO3 (Allen et al., 1996; Allen et al., 1998; Grassian, 2002). Due to the presence of water vapor and the particle’s ability to adsorb water, the reaction between nitric acid and carbonaceous mineral dust particles is not limited to the surface but can c ontinue within the bul k of the particle (Goodman et al., 2000; Grassian, 2002). The adsorbed water is thought to be involved with the subsequent reac tivity of the particle. Th is uptake of nitric acid and formation of Ca(NO3)2 has been found to be irreversible. For CaCO3, the uptake coefficients have been determined under a few different conditions (Goodman et al ., 2000; Grassian, 2002; Hanisch and Crowley, 2001). At 0% and 20% relative humidities, they were estimated to be 410 4 2 and 310 5 2 respectively (Goodman et al., 2000; Grassian, 2002). Hanisch and Crowle y determined the uptake coefficients for “dry” heated and “damp” unheated CaCO3 to be 210 5 2 10 and 210 5 4 18 respectively. The value estimated under “damp” conditions by Hanisch and Crowley (2001) is thought to be more relevant under atmos pheric conditions. Their reported value is considered a lower limit for the uptake of HNO3 by CaCO3.

PAGE 67

45 Statement of the Problem The atmospheric d eposition of nitrogen has had a detrimental environmental impact on the water quality and biodiversity in Tampa Bay as well as other regions around the world. Nitrogen is deposited through wet and dry deposition processes in the form of both gaseous and particulate species. The dominant nitrogen-containing gaseous forms include nitric acid and ammonia; nitrate and ammonium are the do minant particulate species. In the Tampa Bay area, gaseous ammonia and nitric acid and fine particle ammonium and nitrate (less than 2.5 m in diameter) have been monitored since 1996. Research in coastal regions, however, has revealed the dominance of nitrate in the coarse mode (greater than 2.5 m). By virtue of their increased mass, the coarse particles may have a greater local envir onmental impact than the fine particles as they have greater deposition velocities and shorter residence times. The purpose of this study was to in vestigate the format ion of particulate nitrate species in a coastal urban env ironment through the use of ambient monitoring and modeling. The goals of this study were: To characterize ambient air nitrat e concentrations and particle size distributions through a network of sampling campaigns

PAGE 68

46 To determine the retention of ni tric acid by nylon filters To determine the spatial di stribution of particles al ong an air mass trajectory To evaluate the use of a thermodynam ic equilibrium model for predicting ambient aerosol phases and concentrations To expand the current data set (fro m 1996 to the present) to account for coarse mode nitrate form ation and its contribution to the local nitrogen deposition estimates To determine if macroparticles, thos e with a diameter greater than 10 m, contain nitrogen To determine the partitioning of nitric acid gas to par ticles nitrate by NaCl and CaCO3 (mineral dust) To model the formation of particulate nitrate on NaCl, sea salt and mineral dust particles, with a focus on macropar ticle formation, and to determine the environmental implications This knowledge will be us eful in developing more accurate estimates of the atmospheric contribution of ni trogen to Tampa Bay. The need and importance of coarse particl e nitrate monitoring will be addressed as well as the chemistry behind the formation of these particles.

PAGE 69

47 Methods and Experimental Instrumentation Cascade Impactors Two sampling systems were used thr oughout these experiments to collect size-segregated particulate matter. T he Andersen impactor (Mark-II Cascade Impactor, Thermo Andersen) had eight fractionated stages with nominal cutpoints of 10, 9.0, 5.8, 4.7, 3.3, 2.1, 1.1, 0.7, 0.4 m and a backup filter. The flow rate was factory set at 28.3 L min-1 and was verified using a dry gas meter. Custom-cut 81-mm quartz filters (Pa ll Gelman Sciences) were used as the collection media. Quartz was chosen for its low SO2 absorption (Batterman et al., 1997) and low blank analyte concentrations. Four different non-rotating MicroOrifice Uniform Deposit Impactors (MOUDI™, MSP Corporation) (Marple et al., 1991) were used during May 2002. Two MOUDIs (MDI-242 and MDI-245) had ten fractionated stages with nominal cut-points of 18, 10, 5.6, 3.2, 1.8, 1.0, 0.56, 0.32, 0.18, 0.10, 0.056 m and a backup filter. One MOUDI (MDI-020) had eight fractionated stages with nominal cut-points of 18, 3.2, 1.8, 1. 0, 0.56, 0.32, 0.18, 0.10, 0.056 m and a backup

PAGE 70

48 filter. Nominal cut-points fo r the MDI-079 MOUDI were 18, 10, 5.6, 3.2, 1.8, 1.0, 0.56, 0.32, 0.18 m and a backup filter. Each MOUDI was equipped with Pall Gelman Sciences Teflo™ PTFE membr ane filters. To prevent an excess pressure drop, 2 m pore size filters were used for all backup filters. The flow rate for the MOUDIs was set at 30 L min-1 and was verified using a dry gas meter. Teflo filters were chosen for their low blank analyte concentrations and their inertness towards the species of interest. Annular Denuder System The annular denuder system (ADS) consisted of (a) a Teflon-coated cyclone inlet to remove particles 2.5 m or greater in diameter, (b) annular denuders to quantitate acidic and basic gase s, and (c) a filter pack for particle collection. The de nuders are a series of concent ric glass tubes that have been etched and coated with chemicals to adsor b gaseous species of interest. During operation, ambient air is drawn in through the cyclone, passed through the denuders, and then filtered. The filter pack typically held a single Teflon PTFE or Nylasorb nylon membrane filter. A series of filters or impregnat ed filters may be used to collect species that may have volatilized off the denuders of preceding filters. A typical deployment setup for th e ADS included (a) a denuder housing, to protect the ADS from the har sh elements, (b) rigid air tubing, to prevent the line from collapsing, and (c) a pump, typically encased in a weatherproof housing.

PAGE 71

49 Flow rates can vary from system to system depending on the type of equipment purchased. In these experiments, the flow rate was set at 10 L min-1. The flow rate was checked using a dry gas meter. Dichotomous Sampler The Rupprecht and Patashnick Partisol-Plus Dichotomous Model 2025 Sequential Air Sampler was an automated monitoring device allowing for the collection of PM2.5 and PM10-2.5 (Poor et al., 2002). The PM2.5 and PM10-2.5 fractions are termed the “fine” and “coarse” particulate matter fractions, respectively. The combined flow rate for the instrument was 16.7 L min-1, which was split, directing 15.0 L min-1 and 1.7 L min-1 of ambient air onto the fine and coarse filters, respectively. The inle t apparatus was approx imately 2.5 meters above ground level. Samples we re collected using Whatman PTFE 46.2-mm filters, integrat ing over 24 hours. Total Suspended Particulate Collection An inverted filter pack from URG Corporation was used to collect total suspended particulate matter (TSP) to determine the concentration of macroparticles (greater than 10 m in diam eter). The filter pack was attached to a burette stand and placed on a platform, with the inlet approximately 2.5 m above the ground. The ambien t airflow was 28.3 L min-1, sufficient to collect

PAGE 72

50 particles up to approximately 80 m in diameter. Samples were collected daily on Whatman 46.2-mm P TFE filters, integr ating over 24 hours. Filters Several different types of filters were used during the course of the experiments. The instrum entation limited the size of filters required for sample collection, and the types of experiment governed the type of filter used. The dichotomous, macroparticle and MOUDI samplers used 47-mm filter media, where the filter media included: Pall Gelman Sciences Nylasorb membrane, Whatman PTFE and Teflo f ilters. The PTFE filter s were inert towards the species of interest and had the advantage of low HNO3 adsorption, hence the minimization of nitrate bias. Nylasorb membrane filters were used to prevent volatilization and loss of particles. The Andersen cascade impactor used custom cut 81-mm quartz fiber filters, purc hased from Pall Gelman Sciences. Due to the presence of background analytes, lab blanks were used to correct for background concentration. Fi eld and trip blanks were also collected for all media types. Because the nature of env ironmental work is similar to trace analysis, filters and equipment were treated with s pecial care to avoid contamination. While working with the instrumentation and filters, powderless gloves were worn and clean tweezers were used. All equi pment and glassware were washed, double rinsed in deionized water and allowed to air dry.

PAGE 73

51 Filter Extraction Filters were removed from the sa mpling apparatus using tweezers while wearing powderless gloves. Filters were placed into 15-mL centrifuge tubes, and 5-15 mL of >18 M -cm deionized water was added The volume of water depended on the type and size of filter to be extracted. For 47-mm Nylasorb and Teflon PTFE filters prior to May 2002, 10.0 mL of deionized water was used. For 81-mm quartz filters, 15.0 mL of deioni zed water was used. For all MOUDI samples collected during May 2002, 5.0 mL of deionized water was used. Smaller aliquots of water were used to lower detection limits during specified sampling events. The deionized water was added to t he centrifuge tubes using a calibrated pipettor. The filters were then sonicat ed for 45 min. The extract was decanted into vials for ion chromatography analysis, leaving ~0.5 mL for pH analysis. If samples were not analyzed that day, t hey were stored at 4C until analysis.

PAGE 74

52 Denuder Preparati on and Extraction Preparation The denuder coatings chosen fo r these experiments were sodium carbonate (Na2CO3) and either citric (C6H8O7) or phosphoric (H3PO4) acid. A Na2CO3 coating removes acidic atmos pheric gases, such as HCl, HNO2, HNO3 and SO2 (Allegrini et al., 1994). Citric an d phosphoric acids capture ammonia (Allegrini et al., 1994). Alkaline denuders were coated with Na2CO3 and glycerol (1% + 1% w/w) in a 50:50 v/v water-methanol solution (Allegr ini et al., 1987; Allegrini et al., 1994; Vossler et al., 1988). Collection effici encies were determined to be >99.5% for HCl, >98.5% for HNO2, >97% for HNO3 and >99% for SO2 (Allegrini et al., 1987; Perrino et al., 2001). Acidic denuders were coated with either citric or ph osphoric acid (1% w/v) in an 80% v/v methanol solu tion (Allegrini et al., 1994). The collection efficiency for ammonia was determined to be >99% wi th negligible deposition of particulate NH4 + (Allegrini et al., 1987; Perrino et al., 2001; Vossler et al., 1988). Prior to coating, each denuder was ri nsed with deionized water for one minute followed by approximately 5 mL of the coating solution. The coating solution was then decanted and the denuder filled with ~10 mL of the fresh coating solution. The denuder was shaken or placed on a rotating table for 10 min. The solution was then decanted, and the denuders were dried using either

PAGE 75

53 filtered compressed air or zero air (Vo ssler et al., 1988). The compressed air was filtered through silica gel, activat ed carbon and acid-coated glass beads. The acid coating was used on the beads to capture any ammonia. Extraction Denuders were extracted using 10.0 mL of >18 M -cm deionized water. Denuders were capped and eit her shaken or placed on a rotating table for 10 min. The extract was decanted into vi als for analysis by ion chromatography. The filters from the ADS filter pack were removed using tweezers and placed into 15 mL centrifuge tubes. 10.0 mL of >18 M -cm deionized water was added to each tube, which were then capped and sonicated for 30 min (Vossler et al., 1988). An aliquot of the filter extract was taken for pH analysis. The remainder was filtered using a Pall Gelman Sciences 0.45 m syringe tip filter and placed into vials for analysis by ion chromatography. Lab blanks were used to correct for background concentrations. Sample Analysis Ion Chromatography Analysis Prior to June 2001, samples were analyzed using a Dionex 2000i ion chromatograph. After June 2001, sample s were analyzed using a Dionex DX-

PAGE 76

54 600 ion chromatograph. In both systems, cations (Na+, NH4 +, K+, Mg2+ and Ca2+) were analyzed using CS12G guard and CS12A analytical columns. Prior to June 2001, anions (F-, Cl-, NO2 -, NO3 -, PO4 3and SO4 2-) were analyzed using AS4G guard and AS4A analytical columns. Post June 2001, AS14G and AS14A columns separated the anions. For both ca tions and anions, isocratic elution and self-regenerating suppressors were used. Calibration curves were used to determine the concentration of each analyte. Curves were created us ing external standards from SPEX CERTIPREP. Three check standards were run for every ten samples for verification. These standards were pr epared from a separat e batch of standards from which the curves were run. pH Analysis pH of samples was taken using an Accumet AR50 pH meter fitted with a Thermo Orion pH probe. The pH meter was calibrated using four points: 4.00, 5.00, 6.00 and 7.00. Prior to analysis samples were brought to room temperature. Measurements were taken wi th ~0.5 mL of sample. The pH probe was rinsed with deionized wa ter, blotted using Kimwipes and place into the sample. The pH was allowed to stabilize. The sample was stirred, and the pH was allowed to re-stabilize. The pH was then recorded. The probe filling solution was replaced every 30 days, and calibration was checked daily.

PAGE 77

55 Trajectory Analysis Incoming air masses to the Tampa Bay area were classified using backward trajectories obtained from the National Oceanographic and Atmospheric Administration’s Air Res ources Laboratory (NO AA ARL) website (Draxler and Hess, 1998; HYSPLIT4, 1997) The Hybrid Single-Particle Lagrangian Integrated Trajecto ry (HYSPLIT) model plots the trajectory of an air mass on a three-dimensional grid usin g archived two-hour meteorological data and Lagrangian and Eulerian ca lculations. Trajectories were plotted for each 24hour sampling period at each sampling site (Table 3). Site Street Address City Latitude Longitude Azalea Park 7200 22nd Ave. N. St. Petersburg 27N 47' 03" 82W 44' 24" Gandy 5121 Gandy Blvd. Tampa 27N 53' 33" 82W 32' 15" Sydney Dover & Sydney Rds.Dover 27N 57' 56" 82W 13' 56" Table 3. Addresses and coordi nates for each sampling site. Figure 2. 24-hour backw ard trajectories for (a) terrest rial or land or igin (October 17, 2001) and (b) marine origin (Oct ober 25, 2001) (HYSPLIT4, 1997).

PAGE 78

56 Terrestrial or land air masses consis ted of masses that originated from the north, swept across the state of Florida, or swirled around Tampa (as moderately stagnant air). Marine air masses were those that spent the majority of their path over the Gulf of Mexico or the Atlantic Ocean. Error Analysis MOUDI Error was estimated for the MOUDI sa mplers using collocated instrument data from the May 2002 Sydney site in tensive monitoring period. Two instruments were deployed side-by-side. The flow rates of each instrument were adjusted so they were identical at 30 L min-1. Instrument A was comprised of twelve collection bins and instrument B on ly ten. The three smallest bins of instrument A collected the same diamet er particles as the smallest bin of instrument B. Bins ni ne through twelve of instru ment A were summed and then compared to bin ten of inst rument B. This resulted in ten comparable bins for May 2-20, 2002. On May 21, 2002, the instruments were exchanged and rotated at the collection sites. In strument C was collocated with instrument B. However, instrument C did not contain the same cut-points as inst rument B. The only way to compare these instruments would have been to sum all of the bins. Because

PAGE 79

57 of this, samples collected after May 20, 2002, were not used in the error estimation. Figures 3-11 display the size distributions for the collocated measurements. Size dist ributions for fluoride, ni trite and phosphate were not created because these analytes we re near the detection limits.

PAGE 80

58 Na +D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 NH4+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 K +D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 0000 0001 0002 0003 0004 0005 Mg 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 Ca 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 Cl -D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) -0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06 NO3 -D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 000 002 004 006 008 0.10 SO4 2-D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 Figure 3. Comparison of collocated measurement s of instrument A ( ) and instrument B ( ) for May 4, 2002.

PAGE 81

59 Na +D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 NH4+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 K +D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.001 0.002 0.003 0.004 0.005 0.006 0.007 Mg 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 Ca 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 Cl -D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 NO3 -D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 000 002 004 006 008 0.10 SO4 2-D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 Figure 4. Comparison of collocat ed measurements of instrument A ( ) and instrument B ( ) for May 6, 2002.

PAGE 82

60 Na +D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 NH4+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 K +D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.001 0.002 0.003 0.004 0.005 Mg 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 Ca 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 Cl -D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 NO3 -D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 000 002 004 006 008 SO4 2-D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 Figure 5. Comparison of collocated measurement s of instrument A ( ) and instrument B ( ) for May 10, 2002.

PAGE 83

61 Na +D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 NH4+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 K +D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 0000 0001 0002 0003 0004 0005 Mg 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.002 0.004 0.006 0.008 0.010 0.012 Ca 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 Cl -D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) -0.01 0.00 0.01 0.02 0.03 0.04 0.05 NO3 -D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) -001 000 001 002 003 004 005 006 SO4 2-D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 Figure 6. Comparison of collocated measurement s of instrument A ( ) and instrument B ( ) for May 14, 2002.

PAGE 84

62 Na +D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 NH4+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 K +D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 0000 0001 0002 0003 0004 0005 0006 0007 Mg 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 0.018 Ca 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 Cl -D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 NO3 -D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 000 002 004 006 008 SO4 2-D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 Figure 7. Comparison of collocated measurement s of instrument A ( ) and instrument B ( ) for May 15, 2002.

PAGE 85

63 Na +D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 NH4+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 K +D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 0000 0001 0002 0003 0004 0005 0006 Mg 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 Ca 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 Cl -D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 NO3 -D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) -001 000 001 002 003 004 005 006 SO4 2-D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 Figure 8. Comparison of collocated measurement s of instrument A ( ) and instrument B ( ) for May 16, 2002.

PAGE 86

64 Na +D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 NH4+D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 000 002 004 006 008 0.10 K +D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.0000 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030 0.0035 Mg 2+D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 0000 0002 0004 0006 0008 0010 0012 0014 Ca 2+D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.002 0.004 0.006 0.008 0.010 0.012 Cl -D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 000 002 004 006 008 NO3 -D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 000 002 004 006 008 SO4 2-D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 000 001 002 003 004 005 Figure 9. Comparison of collocated measurement s of instrument A ( ) and instrument B ( ) for May 17, 2002.

PAGE 87

65 Na +D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) -0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06 NH4+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 K +D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 0000 0001 0002 0003 0004 0005 0006 Mg 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.002 0.004 0.006 0.008 Ca 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) -0.001 0.000 0.001 0.002 0.003 0.004 0.005 Cl -D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) -0.01 0.00 0.01 0.02 0.03 0.04 0.05 0.06 NO3 -D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) -001 000 001 002 003 004 SO4 2-D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 Figure 10. Comparison of collocat ed measurements of instrument A ( ) and instrument B ( ) for May 19, 2002.

PAGE 88

66 Na +D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 NH4+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 000 005 0.10 0.15 020 025 030 K +D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0.0000 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030 0.0035 Mg 2+D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 0000 0002 0004 0006 0008 0010 0012 Ca 2+D p ( m) 001 0.1 1 10 dC/dLogD p (umol m -3 ) 0.000 0.002 0.004 0.006 0.008 Cl -D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 000 002 004 006 008 NO3 -D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) -001 0.00 0.01 0.02 0.03 0.04 0.05 0.06 SO4 2-D p ( m) 0.01 0.1 1 10 dC/dLogD p (umol m -3 ) 000 002 004 006 008 0.10 0.12 0.14 Figure 11. Comparison of collocat ed measurements of instrument A ( ) and instrument B ( ) for May 20, 2002.

PAGE 89

67 Relative bias and relative precision were calculated fo r the error analysis period using Equations 12-14 (Poor et al ., 2002). The relative precision was calculated for each analyte in each size bin (Table 4). ) x y ( ) x y ( RBi i i i i 2 (Equation 12) i iRB n RB 1 (Equation 13) i iRB n RP22 1 (Equation 14) Geo Mean Bin Max ( m) ( m) Na+ NH4 +K+ Mg2+Ca2+FClNO2 -NO3 PO4 3SO4 2-AVG 23 30 19% 77%50%19%27%73%19%92%28% 19% 20%40% 13 18 26% 100%57%28%35%103%26%116%31% 48% 24%54% 7.5 10 9% 82%28%12%17%70%10%90%16% 50% 9% 36% 4.2 5.6 19% 68%14%18%15%65%23%98%15% 56% 50%40% 2.4 3.2 25% 64%27%22%28%83%25%98%16% 13% 14%38% 1.3 1.8 39% 53%23%53%43%82%48%75%34% 77% 58%53% 0.75 1.0 28% 20%40%50%85%47%112%97%90% 1% 17%53% 0.42 0.56 71% 48%35%106%103%47%85%96%82% 82% 44%73% 0.24 0.32 73% 53%39%86%110%47%99%109%83% 67% 40%73% 0.042 0.18 98% 45%80%124%87%47%96%100%109% 67% 42%81% Average 41% 61%39%52%55%66%54%97%50% 48% 32%54% Table 4. Relative precision for the MOUDI instrum ent during May 2002. Relative precision impr oves at higher analyte concentration. Chloride is predominantly a coarse mode species. W hen analyzing the relative precision for chloride in the coarse parti cle bins, the relative precision is, on average, 20%. When analyzing the chloride in the fine particle bins, however, the relative precision is near 100%. Analyte concentrations of F-, NO2 and PO4 3were near

PAGE 90

68 the detection limits during the sampling per iod, resulting in higher relative precision values. The relative precision values listed in Table 4 were used to place error bars on the MOUDI measurements. Annular Denuder System Error estimation for the annular denuder system (ADS) was done using collocated measurement dat a from August 1996 through November 2002. The Environmental Protection Commission of Hillsborough County collected data on a one-in-six day measurement cycle at t he Gandy monitoring site. The relative bias and precision were calculated usi ng Equations 12-14. The values are reported in Table 5. Number of Relative Samples (n) Precision (%) HCl 19 37% HNO3 304 24% NH3 305 15% SO2 303 14% NO3 280 23% NH4 + 278 28% SO4 2281 26% Table 5. Relative precision for the annular denuder syst em measurements.

PAGE 91

69 Statistical Analysis Grubbs’ Outlier Test The Grubbs’ outlier test (GraphP ad Software, 2000) is also called the extreme studentized deviate (ESD) me thod. The test computes a Zobtained value, SD value mean Z | | (Equation 15) where the mean was the arithmetic mean of the data set, value was the numerical value in question and SD was the standard deviation of the data set. Both the mean and the standard deviation were calculated using all of the values, including the outlier valu e in question. The Zobtained value was compared to a Zcritical. The null hypothesis was rejected if |Zobtained| > |Zcritical|, and the value in question was identified as an outlier. Paired t-Test The paired t-test was used to compar e the means of two groups of data. The test assumes the sampled data se t was taken from a Gaussian bell-shaped distributed population. A null hypothesis was developed stating the two data sets were not statistically different. The difference between each pair of measurements was calc ulated. The mean, d, and the standard deviation, SD,

PAGE 92

70 of the differences were then calculated and used to obtain a t-value (Glover and Mitchell, 2002; Pagano and Gauvreau, 2000), n SD d t (Equation 16) where n was the number of observations. The null hypothesis was accepted if |tobtained| < |tcritical|, indicating the data sets were not statistically different. Wilcoxons Signed Rank Test The Wilcoxons signed rank test is a non-parametric test used to test the differences between two paired data sets taken from a non-Gaussian distributed population. Values were ranked accordi ng to the absolute value of their size from the smallest to the largest (O tt, 1993; Pagano and Gauvreau, 2000). The ranks were summed as: 2 1 )n(n S (Equation 17) where n was the number of observations. Ranks were then summed according to their association with positive or negative values. )S,S(min Tobtained (Equation 18) The smaller of the two values (S+ or S-) was used as the test statistic. Tcritical can be obtained from a table. The null hypothesis was accepted if |Tobtained| < |Tcritical| indicating the paired data sets were not statistically different.

PAGE 93

71 Experimental Studies Preliminary Dichotomous Studies Local air quality and particulate ma tter size distributions depend on the local environment and meteorological cond itions. Particulate ammonium and nitrate can be found in the fine fraction as NH4NO3, as seen in California’s urban environment (Grosjean, 1982); and nitrat e can be almost exclusively in the coarse fraction as NaNO3, as seen in Hong Kong’s maritime environment (Zhuang et al., 1999a; Zhuang et al., 1999b). Previous measurements of coarse particulate matter in urban coasta l Tampa have only included mass determination; they have not included inorganic specie s determination through chemical analysis. The main focus of this experiment was to develop a background understandi ng of the inorganic aerosol distribution for the Tampa area. Experimental Samples were collected at the Gandy Bridge sampling site, adjacent to Tampa Bay, with the he lp of the Environmental Protection Commission of

PAGE 94

72 Hillsborough County. T he samples were collected during a seven-day period from October 5-12, 2000, us ing a dichotomous sample r. The sampler was equipped with Nylasorb membrane filters, integrating over 24 hours. After sampling, the filters were brought back to the laboratory where they were extracted and analyzed. Results and Discussion Samples were analyzed to determine which particulate species dominated the coarse and fine size fractions. Re sults indicated the following: sodium, calcium, chloride and nitrate were the species dominating the coarse size fraction (Figure 12). The remainder, ammoni um, potassium, fluoride, sulfate and hydronium, were primarily f ound in the fine fraction. The coarse percent of a species was calculated by: % ] Na [ ] Na [ ] Na [ ] Na [ % Coarsefine coarse coarse100 (Equation 19) The high concentrations of sodium and chloride in the c oarse fraction can be attributed to the presence of sea salt at the bayside Gandy site. Calcium was also found predominantly in the coarse mode. The Ca2+:Na+ molar ratios for this sampling period averaged to be 0.41, where t he ratio to that in seawater is 0.044 (Zhuang et al., 1999a). The excess calcium in the coarse fraction, which was not accounted for by sea salt, can be attributed to mineral dust (CaCO3) particles (Zhuang et al., 1999a).

PAGE 95

73 Species Coarse Percentage 0 25 50 75 100 Na + NH 4 + K + Ca 2+ F Cl NO 3 SO 4 2H + Figure 12. Coarse percentage for di chotomous samples collected during October 5-12, 2000. About two-thirds, or 67%, of nitrat e was found in the c oarse fraction. Coarse mode nitrate is assu med to be predominantly NaNO3, a product from the reaction with sea salt. The presence of nitrate in the fine fraction was not thought to be NH4NO3, as this species is very volatile and unlikely to form in warm, humid environments (Allen et al., 1989; Grosjean, 1982). Instead, the presence of fine mode nitrate can be attributed to two r easons: (1) the filter media used for collection was Nylasorb memb rane filters. These nylon filters have a tendency to adsorb nitric acid from the air st ream, potentially biasing the nitrate concentration. In the dichotomous sampler, the air stream was split by a virtual

PAGE 96

74 impactor, 1.67 L min-1 of the incoming air was directed onto the coarse filter, with the remaining 15 L min-1 directed onto the fine filter (Poor et al., 2002). This increase in airflow on the fi ne filter greatly increased t he potential for bias during the sampling period. (2) T he distribution of coarse mode nitrate may extend into the fine (<2.5 m) fraction. This theory will be examined using size distribution studies. Size Distribution Determination Cascade impactors are instrum ents used to obtain a clearer understanding of particle size distributions These instruments separate particles by their aerodynamic diameter into multip le size bins instead of simply coarse and fine particle bins. The purpose of th is experiment was to determine the size distribution of nitrate and ammonium particles. Experimental An Andersen cascade impactor was deployed during a three-day period, January 11-13, 2001. The instrument was equipped wi th custom-cut quartz fiber filters, integrating over 72 hours. For instrument comparison, a dichotomous sampler was collocated with the Anders en impactor. The dichotomous sampler was equipped with Nylasorb membrane f ilters, integrati ng over 24 hours.

PAGE 97

75 Results Table 6 shows the range and threeday averaged concentrations for the dichotomous fine and coarse size fr actions. Table 7 shows the total concentrations for the Andersen cascade impactor samples for the three-day integrated period based on size fraction. Dichotomous Particle Concentrations ( g m-3) Fine (Dp<2.5 m) Coarse (2.5
PAGE 98

76 Cascade Impactor Particle Concentrations ( g m-3) Size ranges, m 0.0-0.4 0.4-0.70.7-1.11.1-2.12 .1-3.33.3-4.74.7-5.85.8-9.0 9.0-10 H+ 0.00 0.0020.0020.0010. 00 0.0010.0050.00 0.002 Cl0.00 0.00 0.00 0.00 0. 00 0.00 0.00 0.068 0.00 NO3 0.02 0.51 0.15 0.23 0. 39 0.68 0.40 0.64 0.43 SO4 20.69 2.9 3.4 1.1 0. 36 0.34 0.22 0.24 0.30 Na+ 0.068 0.0920.15 0.23 0. 27 0.55 0.33 0.53 0.33 NH4 + 0.11 0.47 0.59 0.16 0. 00 0.00 0.00 0.00 0.00 K+ 0.011 0.0200.0260.0590. 00 0.0200.0090.090 0.015 Mg2+ 0.028 0.0290.0020.0280. 0290.0500.0350.050 0.035 Ca2+ 0.029 0.00 0.0720.24 0. 0630.15 0.18 0.12 0.12 Table 7. Concentrations for the A ndersen cascade impactor for January 11-13, 2001. Size distributions were calculated from the Andersen cascade impactor data. The data was fit to a normalized di stribution by dividi ng the experimental analyte concentration, Cexperimental, by the difference in the Log10 diameter for the bin maximum and minimum. min max10 10 al experiment p p pD Log D Log C dLogD dC (Equation 20) Sodium and nitrate had bimodal dist ributions (Figure 13), both peaking between 3-4 and 9-10 m. Ammonium and sulfate we re both of a single mode (Figure 14), peaking around 0.65 m.

PAGE 99

77 Particle Diameter, Dp ( m) 0.1110 dC/dLogDp (umol m-3) 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 Na+ NO3 ClFigure 13. Normalized particle size dist ributions of sodium, nitrate and chloride using Andersen instrument from January 11-13, 2001. Particle Diameter, Dp ( m) 0.1110 dC/dLogDp (umol m-3) 0.00 0.05 0.10 0.15 0.20 NH4 + SO4 2Figure 14. Normalized particle size dist ributions of ammonium and sulfate using the Andersen instrument from January 11-13, 2001.

PAGE 100

78 Discussion Coarse mode nitrate can be attribut ed to the adsorption and reaction of gaseous HNO3 with calcium and magnesium in mineral dust and sodium in sea salt particles (Jordan et al., 2000). Reac tions with the calcium or magnesium do not appear to be the major pathway under the sodium-rich marine conditions. The fine mode nitrate may be due to t he gas-to-particle conversion and neutralization of HNO3 and NH3, forming NH4NO3. During the sampling times, the environmental conditi ons did not favor the formation of volatile NH4NO3. As a result, very little nitrate was assum ed to be associated with ammonium. The formation of fine mode sulfat e results from the condensation and neutralization of sulfuric acid with a mmonia. During the sampling period, the majority of the sulfate and ammonium was collected in the fine size fractions. Figure 14 shows a strong 1:1 molar ratio of ammonium to sulf ate. It appears that these constituents are pres ent in the form of NH4HSO4, leaving the fine aerosol slightly acidic. The dichotomous and size-fractioned samples revealed the presence of chloride in the coarse fraction. The major source of chloride is sea salt, as NaCl. Seawater has a molar ratio of Cl-:Na+ of 1.16 (Aherne and Farrell, 2002), but the molar ratios of Cl-:Na+ for the coarse and fine fr actions were 0.10 and 0.05, respectively. This indicated a deficienc y of chloride relative to the sodium concentration. The loss of chloride was a ttributed to the reacti ons of NaCl with HNO3 and H2SO4, producing gaseous HCl.

PAGE 101

79 Chloride depletion was calculated us ing Equation 3. The percentage of chloride depletion for this sampling per iod ranged from 86 to 100%, results are reported in Table 8. % Cl-depletion Andersen Fine (Dp > 2.1 m) 100% Dichot Fine (Dp > 2.5 m) 94% Andersen Coarse (2.1 < Dp < 10.0 m) 99% Dichot Coarse (2.5 < Dp < 10.0 m) 86% Table 8. Average chloride depletion, in percentage, for January 11-13, 2001. One factor for determining the extent of chloride depletion is the relative humidity. The water content of the hygr oscopic salts increases with relative humidity. The additional su rface waters play a role in the uptake of nitric acid and the gas-particle nitrate equilibriu m (Guimbaud et al., 2002b). During this sampling period, the relati ve humidity was greater than 90%, and the air mass was of marine origin. High ch loride depletion was expected. In Table 9, the dry deposition flux (Equation 1) was ca lculated for three particle diameters. The nine different si ze bins of the cascade impactor were assigned to one of the three particle diam eter categories. Concentrations were the nitrate particle concentrati on sums for all stages in that size range. As the square of the particle diameter increases its gravitational settling velocity increases. Results in Table 9 indicate that 5.8 to 10.0 m particles accounted for 60% of the total nitrogen flux for particles less than 10.0 m.

PAGE 102

80 Particle Particle Particle Settling Flux Percent Diameter Dp RangeConc. Velocity (kg-N of ( m) ( m) ( g-N m-3)(cm s-1) ha-1 yr-1) Flux 1.0 0.0-3.3 1.3 0.017 0.071 23% 4.7 3.3-5.8 0.24 0.069 0.053 17% 9.0 5.8-10.0 0.24 0.25 0.19 60% Sum of Flux 0.31 100% Table 9. Dry deposition flux for particu late nitrogen (nitrate + ammonium) for January 11-13, 2001. Previous research has modeled the dry deposition settling velocities for different nitrogen containing species (Poor et al., 2001). Over a three-year averaged study, gaseous ammonia and nitric acid accounted for 53% and 40% of the 7.3 kg-N ha-1 yr-1 of total nitrogen deposition, resp ectively. Particulate nitrate only accounted for 3.5%, but this esti mate only accounted for fine mode particulate nitrogen compounds. As seen from Table 9, coarse mode nitrate would account for 60% of particle nitr ogen deposition for particles less than 10 m. New estimates need to be developed to account for coarse particle nitrate. Evidence of Macroparticles In recent years, research has conf irmed the presence of coarse particle nitrogen. Coarse particles of NaNO3 can form, for example, when a marine air mass laden with sodium chloride mixe s with an anthropogenic air mass rich in nitric acid. These particles are typically less than 10 m in diameter. The

PAGE 103

81 purpose of this investigation was to det ermine if rapidly depositing atmospheric particles or particle aggregates with diameters 10 to 80 m contain nitrogen. Experimental Macroparticles are those with a diameter greater than 10 m. They were investigated during an intensive sixweek period during Oc tober and November 2001. Particles up to 10 m were collected using the dichotomous sampler, collecting coarse and fine species. Total suspende d particulates (TSP) were collected using the TSP sampler, or simp ly an inverted filter pack. In both instruments, Whatman PTFE Teflon membr ane filters were used as the collecting media and were integrated over 24 hours. An annular denuder system was collocated with the samplers to collect ambient gaseous nitric acid and fine particulate matter. Results and Discussion Macroparticle concentrations were fi rst determined by simply subtracting the dichotomous (coarse + fine) conc entration from the TSP concentration (Equation 21). macro s dichotomou TSPNO NO NO] [ ] [ ] [3 3 3 (Equation 21) Figure 15 represents t he daily macroparticle co ncentrations of Na+, Cland NO3 -. All values greater t han zero indicate the pres ence of that species in

PAGE 104

82 the greater than 10 m size fraction. Na+ and Clconcentrations were significantly correlated (r=0.75) but not Na+ and NO3 (r=0.26). NO3 concentrations were plotted to visualize the correlations with Ca2+ (r=0.69). These correlations pointed to macroparticle NO3 as possibly Ca(NO3)2 rather than NaNO3.

PAGE 105

83 TSP Dichot Concentration ( mol m-3) 0.00 0.02 0.04 0.06 0.08 0.10 Na+ ClDay 10/22/01 10/29/01 11/5/01 11/12/01 11/19/01 0.00 0.02 0.04 0.06 0.08 0.10 Ca2+ NO3 Figure 15. Daily macroparticl e concentrations of (a) Na+ and Cland (b) Ca2+ and NO3 (October-November 2001).

PAGE 106

84 Due to the nature of coarse part iculate matter sampling, ambient gas species may play a role in the final result of the species collected. In this study, gaseous HNO3 was not denuded from the airflow of neither the TSP nor the dichotomous samplers, and the potential existed for HNO3 to react with coarse particles that had accumulated on the fi lter (Perrino et al., 1988). In the dichotomous sampler, however, the airflo w through the coarse filter was low at 1.67 L min-1 (as compared to the 28.3 L min-1 of the TSP sampler) thus reducing the nitric acid bias associated with coarse particle interactions. n = 29 slope = 1.462 + 0.059 standard error of regression = 0.011 R2 = 0.846 Dichot NO3 Concentration (umol m-3) 0.000.020.040.060.08 TSP NO3 Concentration (umol m-3) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Figure 16. Simple linear regression for TSP NO3 versus dichotomous total NO3 using daily concentrations (October-November 2001).

PAGE 107

85 To estimate the HNO3 bias between the TSP and dichotomous samplers, TSP NO3 was plotted against the dichotomous NO3 (Figure 16). From the linear regression, it was found that the dichotomous total NO3 concentrations explained ~85% of the va riability in the TSP NO3 concentrations. The daily TSP NO3 concentrations were ~46% higher than the dichotomous total NO3 concentrations. The slope of the regre ssion curve was defined as the possible nitric acid bias between the two methods. For each sampling day, the macroparticle concentrations were corre cted for this bi as (Equation 22). 3 3 3] [ ] [ 46 1 ] [macro s dichotomou TSPNO NO NO (Equation 22) Day 10/22/01 10/29/01 11/5/01 11/12/01 11/19/01 NO3 TSP (1.46NO3 dichotomous) Concentration (umol m-3) -0.02 -0.01 0.00 0.01 0.02 0.03 0.04 Nitrate Concentration Median + 2 StErr = 0.022 Median = 0.001 = 0.00 Median 2 StErr = -0.020 Figure 17. Daily macroparticle nitrate conc entrations with correction for the nitric acid bias.

PAGE 108

86 The mean and standard error of the ni trate macroparticle concentrations were plotted in Figure 17 after correction for the nitric acid bias. The mean and standard error excluded the two statistica l outliers. All but two concentrations were within two times the standard error of the regression about zero. These two concentrations offer tentative evidenc e of ambient nitrate macroparticles. The formation of NaNO3 and Ca(NO3)2 on the filter was likely dependent on the air velocity through the filter, t he particle number and size, the relative humidity and the ambient nitric acid concentration; and it may not be well modeled with a linear regre ssion. It was postulated that relative humidity played a role, providing an aqueous layer for the chemical conversion (Ten Brink, 1998). Trends were addressed to find an affilia tion for the macroparticle nitrate. Ammonium-to-sulfate ratios have been used as an indicator of an air mass age. NH4 +:SO4 2molar ratios near 2.0 indicate an aged air mass, whereas those less than or equal to 1.0 represent a re latively fresh air mass, as NH3 has not been completely neutralized with sulfate. The two days of possible macroparticle formation were consistent with NH4 +:SO4 2molar ratios of near 2.0, suggesting an aged air mass. The air mass origin on these days was pr edominantly from terrestrial sources, with higher than average Ca2+ concentrations and HNO3 concentrations. This suggest ed that the ma croparticle NO3 was affiliated with mineral dust particles, possibly as Ca(NO3)2. The nitrate affiliated with the mineral dust Ca2+ particles may also be a product of depositing HNO3 reacting on previously settled soil dust. There was a minimum amount of rain during the samp ling period, leaving the potential for

PAGE 109

87 water-soluble species to exist on the ground. The bayside sampling site was located <100 m from a major roadway. Re-suspended soil or road dust was a possible source of macroparticle Ca2+ and NO3 -. A 20-m particle re-suspended to a height of 100 m could travel more than 30 km before re-d epositing to the surface. Other possible sources for ma croparticle nitrate include dust from local agricultural areas. These particles may be more enriched with nitrate than those from other sources. Retention of Nitric Acid by Nylon Filters Several techniques have been used to measure gaseous HNO3 and particulate NO3 concentrations in the atmosphere. Some of these techniques include filter pack (FP) methods (Anlauf et al., 1986; Perrino et al., 1988; Spicer, 1986; Torseth et al., 2000) and annular denuder systems (ADS) (Torseth et al., 2000; Tsai et al., 2000; Vossler et al., 1988) Filter packs have been the method of choice for a number of government agencies, such as U.S. Environmental Protection Agency National Dry Depositi on Network (NDDN) and U.S. Clean Air Status and Trends Network (CASTNet) (K im and Allen, 1997; Sickles II and Hodson, 1999) because they have low maintenance problems, light weight, low cost and the same collection efficiency as other methods (Karakas and Tuncel, 1997). The ADS was designed to collect HNO3 and particulate NO3 -, differentiating between the vapor phase HNO3 and HNO3 produced from the dissociation of NH4NO3 during sampling (Vossler et al., 1988).

PAGE 110

88 Several studies have indicated the potential for nylon filters to adsorb gaseous nitric acid (Karakas and Tuncel, 1997; Perrino et al., 1988), but most of these studies were perform ed in environments where c onditions were suitable for the formation of NH4NO3. This experimental study was done in a coastal environment where the average relative humidity and temperature were above the deliquescent relative humidity (S einfeld and Pandis, 1998) and dissociation temperatures (Stelson et al., 1979) for NH4NO3, thus avoiding the issues of formation and dissociation. PM2.5 cyclone inlets were chosen to av oid the collection of sea salt particles, which have bimoda l diameters greater than 3 m (Evans and Poor, 2001). HNO3 can be liberated from sea salt parti cles by the reaction of nitrate salts (NaNO3) with H2SO4 and HCl (Appel et al., 1984). Experimental Ambient gas and particulate concentrations were collected using two separate channels on an annul ar denuder system (ADS) fr om URG Corporation. Channel one consisted of a Teflon-coated PM2.5 cyclone inlet, two 242-mm denuders and a filter pa ck, in series. The denuders we re prepared as stated in the methods section using citric acid as the acidic denuder coat ing solution. The filter pack was equipped with a Whatman ny lon or Nylasorb nylon filter. The second channel consisted of a Teflon-coated PM2.5 cyclone inlet and a filter pack,

PAGE 111

89 in series. The filter pa ck was equipped with a Whatm an nylon or Nylasorb nylon filter. Denuders and filters were extracted and analyzed. The samples were collected during a si x-week period in the fall of 2001. The sampling site was located on the prop erty of the State of Florida Fish and Wildlife Conservation Commission, which is on the eastern side of the Gandy Bridge on Tampa Bay. The samples we re integrated over 24 hours with an ambient airflow of 10 L min-1, which was checked daily using a dry gas meter. During the first three weeks, 47-mm, 1-m pore size Whatman nylon filters were used. Pall Gelman Nylasorb 47-mm, 1-m pore size were used during the second three-week period. When using a cyclone inlet in marine env ironments, the interior wall of the ADS cyclone may be become coat ed by sea salt particles. It has been found that sea salt aerosols react with acids, such as H2SO4 and HNO3, in the atmosphere forming sulfates and nitrates (Li-Jones et al., 2001). The reaction between HNO3 and sea salts within the cyclone may resu lt in a substant ial amount of HNO3 loss. This reaction leads to the misapportionment of both NO3 and Cl-, as HCl is released into the vapor phase. To prevent nitric acid loss within the cyclone, the cyclones were thoroughly cleaned with deionized water and dried between deployments.

PAGE 112

90 Results and Discussion Two channels on the ADS we re used to collect (a) HNO3 and “denuded” particulate matter (PM) nitrate. T he term denuded is used to describe the PM2.5 fraction of nitrate of which the HNO3 and other reactive gases have been removed, or denuded, prior to collecti on of the particulate matter; and (b) “undenuded” PM NO3 -. Undenuded refers to the se cond channel of collection in which both HNO3 and PM2.5 NO3 are collected on the filter. In theory if nylon substrates prove to be 100% efficient in HNO3 collection, the sum of analytes collected on both the denuder and denuded filter will equal the sum of analytes on the undenuded filter. The reason for using two different types of nylon filters was the availability of the media types at the time of samp ling. Pall Gelman Sciences had halted production of the Nylasorb ny lon filter, and they were u navailable for purchase. As an alternative, nylon filters were purchased from Whatman. One major drawback for the use of t he Whatman filters was t he presence of existing analytes, especially nitrates. Both types of filters were treat ed identically from deployment to analysis. Data correction by blank subtraction is common for this type of analysis, but was a major factor in the Whatman filters. Pall Gelman Sciences certifies their filters with respect to nitrate concentrations. The temperature and relative humidity (RH) values for the sampling period are listed in Table 29 (A ppendix 1). Experimental HNO3, denuded PM nitrate and undenuded (UD) nitrate concentrati ons are given in Table 10.

PAGE 113

91 Whatman ( mol m-3) Nylasorb ( mol m-3) HNO3 gas den PM NO3 UD HNO3 + NO3 HNO3 gas den PM NO3 UD HNO3 + NO3 10/10/01 0.009 0.009 0.017 11/1/01 0.007 0.017 0.020 10/11/01 0.007 0.007 0.010 11/2/01 0.006 0.014 0.016 10/12/01 0.007 0.011 0.015 11/3/01 0.006 0.012 0.013 10/13/01 0.008 0.010 0.014 11/4/01 0.019 0.007 0.008 10/14/01 0.007 0.008 0.014 11/5/01 0.008 0.015 0.023 10/15/01 0.023 0.013 0.022 11/6/01 0.010 0.002 0.028 10/16/01 0.012 0.011 0.020 11/7/01 0.033 0.017 0.031 10/17/01 0.013 0.007 0.014 11/8/01 0.023 0.029 0.052 10/18/01 0.012 0.010 0.018 11/9/01 0.059 0.037 0.047 10/19/01 0.010 0.015 0.019 11/11/010.090 0.033 0.085 10/20/01 0.007 0.011 0.015 11/12/010.032 0.011 0.043 10/21/01 0.014 0.008 0.010 11/13/010.015 0.009 0.012 10/24/01 0.012 0.008 0.008 11/14/010.026 0.007 0.009 10/26/01 0.043 0.004 0.023 11/15/010.009 0.010 0.018 10/27/01 0.011 0.003 0.008 11/16/010.038 0.011 0.026 10/28/01 0.009 0.010 0.015 11/17/010.038 0.016 0.019 10/30/01 0.026 0.010 0.015 11/18/010.011 0.019 0.022 10/31/01 0.006 0.012 0.016 11/19/010.024 0.041 0.049 Table 10. Experimental nitric acid denuded nitrate and undenuded nitrate concentrations for October November 2001. The Grubbs’ outlier test (GraphPad Software, 2000) was performed for the Whatman nylon filter data set, and no outliers were det ected. All data points were included in the linear regression analysis (Figure 18) (GraphPad Software, 2000). From the linear regression (Figure 18), the sum of HNO3 plus the denuded PM NO3 can be estimated to be 37% great er than that collected on the undenuded NO3 fraction. These filters do not appear to be completely efficient in adsorbing gaseous nitric acid.

PAGE 114

92 n = 18 slope = 1.374 + 0.380 intercept = 0.002 + 0.006 standard error of regression = 0.007 R2 = 0.450 Undenuded NO3 ( mol m-3) 0.0060.0080.0100.0120.0140.0160.0180.0200.0220.024 HNO3 (g) + Denuded PM NO3 (umol m-3) 0.01 0.02 0.03 0.04 0.05 Figure 18. HNO3 + Denuded PM NO3 vs. Undenuded NO3 for Whatman nylon filters. The paired t-test (GraphPad Softw are, 2000) was performed to determine the statistical significance of the data set. The test was based on a two-tailed test at the 95% confidence interval. The two data sets were not statistically different if |tobtained|<|tcritical|. For the Whatman nyl on filters data set, the tobtained = 4.6 (tcrit 95% = 2.1). The test resulted in a rejection of the null hypothesis, indicating the data sets for the (HNO3 + denuded NO3 -) compared to the undenuded NO3 (gas plus particulate) we re statistically different. A linear regression was created for t he Nylasorb nylon filters. The Grubbs’ outlier test was performed, not detecting any outliers (GraphPad

PAGE 115

93 Software, 2000). All data points were in cluded in the linear regression analysis (Figure 19). n = 18 slope = 1.244 + 0.193 intercept = 0.006 + 0.007 standard error of regression = 0.016 R2 = 0.721 Undenuded NO3 ( mol m-3) 0.000.020.040.060.080.10 HNO3 (g) + Denuded PM NO3 (umol m-3) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Figure 19. HNO3 (g) + Denuded PM NO3 vs. Undenuded NO3 for Nylasorb nylon filters. The paired t-test (GraphPad Softw are, 2000) was performed to determine the statistical significance of the dat a set. The test was based on the 95% confidence interval. For the Nylasorb nylon filter data set, tobtained = 3.5 (tcrit 95% = 2.1). The data sets (HNO3 + denuded PM NO3 -) compared to the undenuded NO3 (gas + particulate) were statistically different. To determine the collection efficien cy of both the Whatman and Nylasorb nylon filters, the particulate matter fraction was subtracted from the undenuded nitrate (Equation 23), leavi ng only the adsorbed nitric acid fraction to compare. PM undenuded adsorbedNO NO HNO] [ ] [ ] [3 3 3 (Equation 23)

PAGE 116

94 The adsorbed nitric acid, adsorbedHNO] [3, from both filt er types was compared to the nitric acid collect ed on the denuders through a linear regression (Figures 20-21). n = 18 slope = 0.423 + 0.040 standard error of regression = 0.003 R2 = 0.537 HNO3 (g) from the ADS (umol m-3) 0.000.010.020.030.040.05 HNO3 (g) adsorbed from the Whatman filter (umol m-3) 0.000 0.005 0.010 0.015 0.020 0.025 Figure 20. Nitric acid fr om the ADS compared to that adsorbed by the Whatman nylon filters.

PAGE 117

95 n = 18 slope = 0.454 + 0.073 standard error of regression = 0.010 R2 = 0.449 HNO3 (g) from the ADS (umol m-3) 0.000.020.040.060.080.10 HNO3 (g) adsorbed from the Nylasorb filter (umol m-3) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 Figure 21. Nitric acid from the ADS co mpared to that adsorbed by the Nylasorb nylon filters. The Whatman and Nylasorb filters appear to have collected only 42% and 45% of the total nitric acid, respectively. To verify this was correct, the t-test was performed comparing the HNO3 adsorbed by each filter type to the linear regression slope times the HNO3 collected by the ADS. For the Whatman filter type, wher e the regression sl ope was 0.42, tobtained = 0.55 (tcrit 95% = 2.1). The t-test resulted in a non-statistical difference between the two sets, further indicating the co llection efficiency of the Whatman nylon filters to be approximately 42%. For the Nylasorb nylon filters, wh ere the regression slope was 0.45, tobtained = 0.20 (tcrit 95% = 2.1). The t-test resulted in a non-statistical difference between

PAGE 118

96 the two data sets. This indicated the co llection efficiency of the Nylasorb nylon filters to be approximately 45%. A similar study has been done looking at the uptake of nitrous acid (HNO2) and NOX by nylon surfaces (Perrino et al., 1988). The collection efficiencies of HNO2 on nylon filters was determined experimen tally at different flow rates. Efficiencies ranged from 25% at 12 L min-1 to 80% at 2 L min-1. At 10 L min-1, the collection efficien cy ranged from 30-40%. The results from Perrino et al. ( 1988) are comparable to those reported here with nitric acid, HNO3. Both the Whatman and Nyla sorb nylon filter media resulted in a collection effici ency of approximately 40%. Size Distributed Trajectory Study As freshly emitted sea salt particl es mix with the urban plume, sodium nitrate particles begin to fo rm. With transport, this gas-to-particle conversion of nitric acid on the sea salt particles is expected to reach equi librium and contribute to particulate nitrate deposition within the Tampa Bay area watershed. A trajectory study was completed looking at the transformation and deposition of particles as an air mass moves throughout the Tampa Bay area.

PAGE 119

97 Experimental Three monitoring sites were es tablished during May 2002, taking advantage of monitoring already in progress. At each site, a MOUDI sampler, an ADS and various meteorological instrum entation were deployed. Collected data included 23-hour integrated, size-segr egated inorganic particulate species concentrations, 12-hour integrated acidic and alkaline gas concentrations, wind speed, wind direction, ambient temper ature, relative humidity, and water temperature. The Azalea Park site is located in the southwester corner of Pinellas County, with close proximity to the Gulf of Mexico (Figure 22). The Gandy Bridge site is located on the eastern side of t he Gandy Bridge, adjacent to Tampa Bay. The third site, Sydney, is located in a semi-rural area, approximately 20 km from Tampa Bay. The three sites span over a 55 km distance, providing sufficient travel distance for particle conversion and some deposition. There are many sources of error and uncertainties in this type of experimental fieldwork. They need to be addressed in order to determine if there was a real difference in particle conc entrations and flux between the sampling sites. The greatest uncertainty lies in t he pump flow control. Each pump was set at a 30 L min-1 flow rate and was checked week ly using a dry gas meter. The pumps were not installed in a climate-c ontrolled shelter but were deployed in the field open to the elements. With cont inual changes to ambi ent temperature and pressure, the flow rate was assumed to change. During May 2002, the daily

PAGE 120

98 temperature ranged from 18 to 33C, with an 8 to 10C difference between the daily minimum and maximum temperatures Applying a 10C temperature change to the ideal gas law resulted in a 3% change to the 30 L min-1 volumetric flow rate, which was within the range of the quality assuranc e protocol flow restriction guidelines of 5%. The overall error of the MOUDI samp lers was estimated during this time period and was averaged at 54%, ranging from 1 to ov er 100% (Table 4). The average sizeand species-dependent erro r was applied to each data set and was seen in the following particle size di stributions. In order to determine a real or significant change in the particle conc entrations, the entire data range with its error needs to be compared. As a resul t, minor changes in concentration cannot be considered significant as the data point s may, in actualit y, be of identical value. In this study, trends were used to determine significant changes in the particle concentration and flux. Ten percent was assigned as the minimum difference values needed before they were considered signi ficantly different from one another.

PAGE 121

99 A ZALEA PAR K Tampa Point sources of NH3 Point sources of NOx S YDNEY GANDY 32 km 23 km Figure 22. Map of sampling sites during the May 2002 intensive period. Results and Discussion Four specific days during the May 2002 intensive were investigated as they represented the predomi nant air mass pattern duri ng the sampling period. May 4, 2002, represents a southwestern air origin day (Fi gure 23a), originating from the Gulf of Mexico and the southweste rn Florida region. The air came into the area from the west, crossing the si tes eastward in order from Azalea to Sydney. May 14, 2002, r epresents a northwestern air day (Figure 23b), originating from the northern states, coming across the Gulf of Mexico for only a short period of time. The air on May 6, 2002, originat ed from the Atlantic Ocean (Figure 23c). The air mass traveled ac ross the state before reaching the Tampa Bay area. The air traveled in towards the west, reaching the Sydney site first and

PAGE 122

100 then traveling onwards past Gandy toward s Azalea. On May 20, 2002, the air originated in the Atlant ic Ocean area near North and South Carolina (Figure 23d). The air traveled over the ocean for some time and then proceeded to cross the state, reaching the Sydney site first, followed by Gandy and Azalea. (a) May 4th South Western Winds (b) May 14th North Western Winds (c) May 6th Eastern Winds (d) May 20th North Eastern Winds Figure 23. Backward air mass trajectories for (a) May 4th, (b) May 14th, (c) May 6th and (d) May 20th, 2002. Many interesting things can be look ed at for each episode. Ammonium to sulfate molar ratios can be used to determine the approxim ate age of an air mass. Ratios near 1.0 indicate a fresh marine air mass, however those near 2.0 indicate an aged or pollution-laden air mass. Chloride to sodium and nitrate to sodium molar ratios and chloride depletio n can be used to determine the extent of the reaction between sea salt and nitric acid. The trends of the magnitudes of concentrations can be used to indicate particulate deposition and re-suspension.

PAGE 123

101 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 24. Size distributions for each sampling site on May 4, 2002. Azalea Gandy Sydney Cl-:Na+ 0.77 0.75 0.55 NO3 -:Na+ 0.48 0.50 0.75 Cl--dep % 35% 36% 54% NH4 +:SO4 21.6 1.8 1.7 Table 11. Ion ratios for May 4, 2002.

PAGE 124

102 On May 4, 2002, the air mass originat ed from the Gulf of Mexico and southwestern Florida region (F igure 23a). Upon arrival, the NH4 +:SO4 2ratio was 1.57, indicating a moderately fresh or slightly aged air parcel (Figure 24d-f and Table 11). As the air mass travel ed eastward, towards Gandy, the ratio increased. However, as the air parce l traveled towards Sydney, the ratio decreased slightly. This may be attributed to the increased SO2 and SO4 2concentrations over urban Tampa. As the air parcel moved through the downtown area, it is possible that additional sulfate was picked up. Upon the arrival at the Azalea site, the chloride to sodium ratio was 0.77 (Figure 24a-c and Table 11). As the air mass traveled throughout the area, this ratio steadily decreased, reac hing 0.55 at the Sydney si te. Nitrate to sodium ratios increased from 0.48 to 0. 75 as the air picked up urban NOX or nitric acid, converting it to particulate nitrate. In agreement with the above trends, chloride depletion increases from 35% to 54% from Azalea to Sydney. Combined, these three indicators represent the adsorption an d conversion of gaseous nitric acid to particulate nitrate. As predicted by Reaction 1, HCl is released during the process. This is seen through a steady increase in the percentage of chloride depletion. On May 14, 2002, winds originated fr om the northern st ates, approaching the Tampa area from the nor thwestern direction (Figure 23b). When this air mass arrived, the ammonium to sulfate ratio was 1.90 (Figure 25d-f and Table 12). This is indicative of an aged air ma ss. As the air mass crossed Tampa Bay, the ratio remained unchanged (ratio values of 1.9 and 1.8 are not significantly

PAGE 125

103 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 25. Size distributions for each sampling site on May 14, 2002. Azalea Gandy Sydney Cl-:Na+ 0.81 0.66 0.56 NO3 -:Na+ 0.50 0.50 0.73 Cl--dep % 31% 44% 52% NH4 +:SO4 21.9 1.8 2.3 Table 12. Ion ratios for May 14, 2002.

PAGE 126

104 different) and then continued to increase, reaching over 2.3 at the Sydney site. There was a significant in crease in the ammonium co ncentration from the Gandy to Sydney sites. There are anhydr ous ammonia loading docks and a large wastewater treatment plant located at the Port of Tampa, which lies between the Gandy and Sydney sites. Th is increase in ammonium ma y be due to activities at these facilities. The concentration of su lfate was seen to only increase slightly. When the air first reaches the Azalea site on May 14, 2002, the nitrate to sodium ratio is 0.50 (Figure 25a-c Tabl e 12). As the air travels through urban Tampa and picks up nitric acid, the nitrate to sodium ratio continues to increase, reaching 0.73 at the Sydney site. Coupled with the chloride to sodium ratios and chloride depletion indicators, an uptake and transformation of nitric acid to particulate nitrate was seen. All three indicators reveal an increase in the uptake and conversion to nitrate. It appears that some sea salt deposition is occurring between the Gandy and Sydney sites, as the concentra tion of sodium decreases between these monitoring locations. On May 6, 2002, the air mass originat ed in the Atlantic Ocean, traveling west across the state befor e reaching the Tampa Bay area. Upon arrival to the Sydney site, the NH4 +:SO4 2ratio was over 2.0 (Figure 26d-f Table 13). This is indicative of the air aging as it traveled across the state. The ratio continued to remain over 2.0 as the air trav eled past the Gandy and Azalea sites.

PAGE 127

105 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 0.30 Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 0.30 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 0.30 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 26. Size distributions for each sampling site on May 6, 2002. Azalea Gandy Sydney Cl-:Na+ 0.53 0.56 0.64 NO3 -:Na+ 0.63 0.53 0.46 Cl--dep % 55% 52% 46% NH4 +:SO4 22.1 2.4 2.1 Table 13. Ion ratios for May 6, 2002.

PAGE 128

106 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 27. Size distributions for each sampling site on May 20, 2002. Azalea Gandy Sydney Cl-:Na+ 0.72 0.71 0.74 NO3 -:Na+ 0.53 0.46 0.45 Cl--dep % 39% 39% 37% NH4 +:SO4 22.1 2.0 2.2 Table 14. Ion ratios for May 20, 2002.

PAGE 129

107 Upon arrival at the Sydney site, the chloride depletion had already reached 46%, steadily increasing to 55% along the prevailing wind direction (Figure 26a-c and Table 13). Coupled with increasing chlo ride depletion, the Cl-:Na+ and NO3 -:Na+ ratios are indicative of the nitric acid gas to particulate nitrate conversion. On May 20, 2002, the air originated in the Atlantic Oc ean area near North and South Carolina (Figure 23). NH4 +:SO4 2ratios at all three sites are over 2.0 (Figure 27 and Table 14), indicating the a rrival of an aged air mass. Chloride depletion and Cl-:Na+ and NO3 -:Na+ ratios remain nearly constant, with only a very small insignificant fraction of change between the monitoring sites. The ammonium and nitrate particulate nitrogen flux was compared for these four sampling days (Figure 28). For each day, regardless of the air mass origin, the particulate nitrate dominat ed the particulate nitrogen flux. These particles were significantly larger than those of ammonium, giving them a significantly larger deposition velocity. For May 4th and 14th, when the air originated out of the wes t, there does not appear to be any significant trend in deposition. For May 6th and 20th, when the air originated out of the east, there does appear to be a slight trend only for nitr ate. Particulate nitrate flux increased along the prevailing wind di rection. In general, the air masses originating from the east were aged more than those from t he west. Aging allows for the gas-toparticulate conversion of nitr ic acid to nitrate. In an older air mass, higher nitrate deposition fluxes would be expected. Fo r the majority of the sampling days during May 2002, the concentration of sodi um was greatest at the Azalea site.

PAGE 130

108 With its proximity to the Gulf of Mexico there appeared to be a salt gradient over land regardless of the wind direction. May 4, 2002 AzaleaGandySydney Particle Flux (kg-N ha-1 yr-1) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 NH 4 + NO 3 May 14, 2002 AzaleaGandySydney Particle Flux (kg-N ha-1 yr-1) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 NH 4 + NO 3 May 6, 2002 AzaleaGandySydney Particle Flux (kg-N ha-1 yr-1) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 NH 4 + NO 3 May 20, 2002 AzaleaGandySydney Particle Flux (kg-N ha-1 yr-1) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 NH 4 + NO 3 Figure 28. Nitrate and ammonium particula te flux for select days in May 2002. Size distributions and ion ratios for the remainder of the May 2002 samples are given in Appendix 2. Met eorological data for the entire sampling period is listed in Tabl es 26-28 in Appendix 1.

PAGE 131

109 EQUISOLV II: A Thermodynamic Model EQUISOLV II is an aerosol thermody namic equilibrium model, which is an updated version of EQUISOLV (Jacobson, 1999a; Jacobson, 1999b; Jacobson et al., 1996) written by Mark Jacobson from Stanford University. The original EQUISOLV model included sodium, ammoni um, chloride, nitrate, and sulfate species. Advances from EQUISOLV, EQ UISOLV II includes potassium, calcium, magnesium and carbonate specie s. The model works by solving sets of equilibrium equations, similar to aA + bB cC + dD, utilizing the temperaturedependent equilibrium coefficient, keq. b a d c eqB A D C T k ) ( (Equation 24) where {X} is the thermodynamic activity of species X. The model iteratively solves sets of equilibrium equations in a “positivedefinite, mass-conserving, and charge-conserving” pr ocess using analytical equilibrium iterations and mass flux iterat ions (Jacobson, 1999a). After sufficient iterations and a positive solution exis ts to a set of equilibrium equations, EQUISOLV II converges. Species can be gases, dissolved liquids, dissolved ions or solids. Many of the equilibrium reactions and corresponding constants are listed in Jacobson (1999a).

PAGE 132

110 The above equilibrium expression (E quation 24) requires the mean mixed activity coefficients. T he equilibrium expression for NH3(g) + HNO3(g) NH4 + + NO3 is: 3 3 3 4 3 42 3 3 3 4} { } { } { } {HNO NH NO NH NO NH eqp p m m HNO NH NO NH k (Equation 25) keq (mol2 kg-2 atm-2) is the equilibrium constant, m (mol kg-1) is the molality, p (atm) is the gas phase partial pressure and 3 4, NO NH is the mean mixed activity coefficient of NH4NO3. In EQUISOLV II, the mean mixed activity coefficients are calculated using the Bromle y’s method for the empirical mixing rule (Bromley, 1973; Jacobson, 1999b) using temperat ure-dependent coefficients given in Jacobson et al. (1996). The expression for the activity coefficients is as follows: ... m B m B m B B ln/ / b 2 3 12 3 12 2 2 1 12 1 0 0 1 12 (Equation 26) where B0, B1, … are the fitting coefficients for each electrolyte and m12 the molalities of electrolytes 1 and 2. T hese values can be found in Table 2 of Jacobson (1999b). The model can be used in two differ ent modes. It can be used to solve for internal equilibrium within a single aerosol bin. Or, it can be used to determine equilibrium for species between the gas p hase and multiple, internally mixed, aerosol size bins. EQUISOLV II can be set to run under various conditions. In the default mode, aerosols are solids at relative hum idity conditions below the particle’s deliquescent relative humidity (DRH) and are aqueous aerosols above the DRH point. Metastable conditions are obtaine d by deactivating the solid formation

PAGE 133

111 reactions in the program, limiting aer osols to the aqueous phase. Metastable conditions exist when the ambient relati ve humidity falls below the particle’s DRH; and the aerosol exhibits a h ysteresis effect and remains as a supersaturated droplet until the relati ve humidity reaches the particle’s crystallization relative humidit y (CRH). This is the relative humidity at which the aerosol crystallizes and becomes a solid. Inputs The model is initialized by creating a text file with the following data: temperature, relative humid ity, pressure, number of collection stages, cut-point diameters for each collection stage, aerosol concentrations and gas concentrations. The model was altered from the original version to accept twelve input bins. Aerosol concentrations of Na+, NH4 +, K+, Mg2+, Ca2+, Cl-, NO3 and SO4 2are entered for each bi n in units of ng m-3. Gas concentrations of HNO3, NH3, HCl and SO2 are also entered in units of ng m-3. Ion imbalances were automatically corrected for using hydrogen and carbonate ions. The following is an example of an input file.

PAGE 134

112 Figure 29. EQUISOLV II input file example. Upon execution, the model redistributes the volatile species between the multiple size bins and the gas phase unti l equilibrium is reache d. The volatile species include nitrate, chloride and ammonium. Outputs The output for the model includes the predicted equilibrium gas, aqueous and solid particle concentrations. T he list of predictable solids includes: NH4NO3, AEROSOL-COMPONENT CONCENTRATION VERSUS SIZE DATE: 5/04 GANDY T = TEMPERATURE (C), EITHER SAMPLING OR AMBIENT RH = RELATIVE HUMIDITY (%), EITHER SAMPLING OR AMBIENT P = AIR PRESSURE (MB), EITHER SAMPLING OR AMBIENT D50 = 50% CUTOFF DIAMETERS BEGIN SAMPLING T = 26.2 C RH = 76.6 %; AMBIENT T = 26.2 C; RH = 76.6 % P = 1019.13 MB P = 1019.13 MB 12 STAGES D50(UM) 18.00 10.0 5.60 3.20 1.80 1.00 0.56 0.32 0.180 0.10 0.056 0.01 DLO, DHI (UM) 0.001 30.0 FORMAT(A1,1X,A6,12(0PF6.2)) AEROSOL CONCENTRATIONS (NG M-3) BEGIN A NA+ 188.7 127.3 552.7 784.0 309.0 151.0 12.88 4.300 2.960 0.000 1.650 0.400 A NH4+ 0.000 0.000 0.000 0.000 1.240 24.46 258.8 363.3 270.3 110.9 72.70 17.65 A K+ 7.760 5.730 22.00 43.89 22.73 10.55 11.33 14.22 11.91 6.990 4.470 4.520 A MG2+ 29.32 19.07 75.96 107.3 44.84 23.98 3.300 1.570 1.070 0.000 0.030 0.110 A CA2+ 152.8 105.5 206.1 265.7 101.6 44.62 7.280 7.050 3.370 0.760 3.550 5.520 A CL319.6 208.8 806.3 848.5 218.1 52.06 0.000 0.000 0.000 0.000 0.000 0.000 A NO3115.0 89.46 550.0 1228. 604.9 238.1 0.000 0.000 0.000 2.000 2.090 20.09 A SO4287.74 57.89 243.6 403.0 234.2 283.8 800.0 1035. 798.0 322.2 211.6 38.16 END GAS CONCENTRATIONS (NG M-3) BEGIN A HNO3 395.1 A NH3 782.0 A HCL 1152. A SO2 7421. END

PAGE 135

113 NH4Cl, NH4HSO4, (NH4)2SO4, (NH4)3H(SO4)2, NH4HCO3, NaNO3, NaCl, NaHSO4, Na2SO4, NaHCO3, Na2CO3, KNO3, KCl, KHSO4, K2SO4, KHCO3, K2CO3, Ca(NO3)2, CaCl2, CaSO4•2H2O, CaCO3, MgCl2, Mg(NO3)2, MgSO4 and MgCO3. The predictable gases and liquids include: HCl, H2O, H2SO4, SO2, HNO3, NH3 and CO2. The remaining analytes are constrained to the aqueous phase. Limitations The EQUISOLV II model is a mass and charge conserving thermodynamic model. The model cannot be used simply to predict a gas phase concentration of a species given a fixed amount in the particle phase. The model only redistributes a given amount of a substance. Multiple, iterativ e model runs must be completed to derive this relationship. One benefit from using the model is t hat it is capable of differentiating between the solid and aqueous phases, allowing it to predict the solid phase formation. This information cannot be directly determi ned from aerosol measurements, which indicates the presenc e and size of the species but not its phase (Campbell et al., 2002).

PAGE 136

114 Qualitative Analysis Before the EQUISOLV II therm odynamic model can be used for theoretical work, it must be investigated whether it can predict size-resolved concentrations of ammonium, chloride and nitrate. Measurements taken during the May 2002 intensive samp ling period were used for the model evaluation. Sample s were taken using the MOUDI sampler with 23-hour integration. Measurem ents were collocat ed with an annular denuder system for the collection of ac idic and basic gas species. Meteorological data was averaged for each 23-hour period fo r the model input. All ions analyzed except fluoride, nitrite and phosphate were included in the model runs. These excluded ions were only present in minute quantities, near the detection limit of the ion chromatograph, and are coincidentally not treated by the model. The model was run in two modes, default and metastable, for analysis to see which mode gave results closest to the experimental data. In the metastable mode, analytes were constrained to the gaseous and aqueous phases.

PAGE 137

115 HNO3 (g m-3) Measured Modeled (Default) Modeled (Metastable) 5/4/02 Azalea 0.55 1.09 0.24 Gandy 0.40 0.40 0.37 Sydney 0.37 1.72 0.35 5/6/02 Azalea 0.89 0.39 0.13 Gandy 1.09 1.51 0.97 Sydney 0.63 1.46 0.10 5/10/02 Azalea 0.37 0.00 0.00 Gandy 1.24 3.52 0.98 Sydney 0.58 1.24 1.05 5/14/02 Azalea 0.42 0.18 0.13 Gandy 0.88 0.88 0.84 Sydney 0.65 0.24 0.21 5/15/02 Azalea 0.68 2.73 0.00 Gandy 0.77 0.48 0.47 Sydney 0.84 2.49 0.69 5/16/02 Azalea 0.69 1.09 0.37 Gandy 0.96 1.37 0.69 Sydney 0.28 0.70 0.34 5/17/02 Azalea 0.21 0.15 0.13 Gandy 0.39 0.64 0.42 Sydney 0.44 0.49 0.44 5/19/02 Azalea 0.56 0.23 0.20 Gandy 0.23 0.20 0.18 Sydney 0.12 0.37 0.36 5/20/02 Azalea 0.60 0.39 0.40 Gandy 0.68 1.45 1.21 Sydney 0.28 0.73 0.40 5/23/02 Azalea 0.60 0.57 0.37 Gandy 0.58 0.64 0.44 Sydney 0.39 0.01 0.01 5/24/02 Azalea 0.27 0.00 0.00 Gandy 0.56 0.94 0.71 Sydney 0.37 0.51 0.38 5/25/02 Azalea 0.36 0.08 0.07 Gandy 0.57 0.97 0.76 Sydney 0.51 0.31 0.22 Table 15. Comparison between measured HNO3 gas concentrations and those modeled by EQUISOLV II.

PAGE 138

116 NH3 (g m-3) Measured Modeled (Default) Modeled (Metastable) 5/4/02 Azalea 2.41 1.92 2.34 Gandy 0.78 0.75 0.82 Sydney 1.63 1.55 1.61 5/6/02 Azalea 2.45 2.66 2.73 Gandy 2.56 3.05 3.05 Sydney 1.09 1.30 1.40 5/10/02 Azalea 3.54 3.78 3.80 Gandy 1.23 1.64 1.64 Sydney 1.73 1.97 1.94 5/14/02 Azalea 1.28 1.28 1.32 Gandy 0.64 0.56 0.62 Sydney 1.19 1.50 1.53 5/15/02 Azalea 2.35 2.38 2.52 Gandy 0.84 0.75 0.84 Sydney 1.62 1.57 1.66 5/16/02 Azalea 1.76 1.36 1.65 Gandy 3.42 2.47 3.25 Sydney 1.53 1.80 1.83 5/17/02 Azalea 4.51 4.38 4.43 Gandy 1.79 1.73 1.75 Sydney 3.38 3.40 3.39 5/19/02 Azalea 1.10 1.23 1.25 Gandy 0.59 0.62 0.62 Sydney 1.59 1.75 1.75 5/20/02 Azalea 2.00 1.90 1.96 Gandy 0.62 0.58 0.55 Sydney 1.75 1.97 1.90 5/23/02 Azalea 2.28 2.22 2.19 Gandy 1.92 1.82 1.80 Sydney 1.64 1.63 1.65 5/24/02 Azalea 2.43 2.52 2.53 Gandy 2.18 2.23 2.19 Sydney 0.91 0.79 0.76 5/25/02 Azalea 2.44 2.55 2.57 Gandy 5.63 5.53 5.48 Sydney 1.17 1.19 1.18 Table 16. Comparison between measured NH3 gas concentrations and those modeled by EQUISOLV II.

PAGE 139

117 HCl (g m-3) Measured Modeled (Default) Modeled (Metastable) 5/4/02 Azalea 2.70 0.90 0.14 Gandy 1.15 0.33 0.33 Sydney 1.06 0.22 0.01 5/6/02 Azalea 0.58 0.09 0.04 Gandy 1.62 0.12 0.30 Sydney 1.01 0.66 0.05 5/10/02 Azalea 0.25 0.00 0.00 Gandy 1.32 0.05 0.03 Sydney 1.51 0.07 0.05 5/14/02 Azalea 0.42 0.03 0.02 Gandy 0.82 0.11 0.05 Sydney 0.51 0.01 0.00 5/15/02 Azalea 0.51 0.03 0.00 Gandy 1.01 0.08 0.03 Sydney 1.38 0.16 0.05 5/16/02 Azalea 2.17 0.89 0.16 Gandy 4.76 2.47 0.46 Sydney 0.94 0.90 0.38 5/17/02 Azalea 0.56 0.03 0.13 Gandy 0.69 0.06 0.19 Sydney 0.35 0.03 0.01 5/19/02 Azalea 0.11 0.11 0.10 Gandy 0.07 0.11 0.11 Sydney 0.66 0.56 0.57 5/20/02 Azalea 0.87 0.04 0.07 Gandy 1.01 0.17 0.06 Sydney 0.46 0.04 0.02 5/23/02 Azalea 0.72 0.04 0.03 Gandy 0.72 0.03 0.04 Sydney 0.43 0.01 0.01 5/24/02 Azalea 0.31 0.00 0.00 Gandy 0.78 0.04 0.02 Sydney 0.60 0.09 0.06 5/25/02 Azalea 0.28 0.00 0.00 Gandy 1.00 0.01 0.02 Sydney 0.43 0.06 0.03 Table 17. Comparison between measured HCl gas concentrations and those modeled by EQUISOLV II.

PAGE 140

118 May 14, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 May 16, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 May 15, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 May 17, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 May 20, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 May 19, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 May 23, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 Experimental Modeled Default Modeled Metastable May 25, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 May 24, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 May 4, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 10, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 May 6, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 Figure 30. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for ammonium at the Azalea site.

PAGE 141

119 May 14, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 16, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 15, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 May 17, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 20, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 May 19, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 May 23, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 Experimental Modeled Default Modeled Metastable May 25, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 May 24, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 May 4, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 10, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 6, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 Figure 31. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for ammonium at the Gandy site.

PAGE 142

120 May 14, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 May 16, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 May 15, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 May 17, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 May 20, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 May 19, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 May 23, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Experimental Modeled Default Modeled Metastable May 25, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 May 24, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 May 4, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 10, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 6, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 Figure 32. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for ammonium at the Sydney site.

PAGE 143

121 May 14, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 May 16, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 May 15, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 May 17, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 May 20, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 19, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 May 23, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Experimental Modeled Default Modeled Metastable May 25, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 May 24, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 May 4, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 May 10, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 May 6, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 Figure 33. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for nitrate at the Azalea site.

PAGE 144

122 May 14, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 May 16, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 May 15, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 May 17, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 20, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 May 19, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 May 23, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Experimental Modeled Default Modeled Metastable May 25, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 0.018 May 24, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 May 4, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 10, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 6, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 Figure 34. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for nitrate at the Gandy site.

PAGE 145

123 May 14, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 May 16, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 May 15, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 May 17, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 May 20, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 May 19, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 May 23, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 Experimental Modeled Default Modeled Metastable May 25, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 May 24, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 0.018 May 4, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 10, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 May 6, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 Figure 35. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for nitrate at the Sydney site.

PAGE 146

124 May 14, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 16, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 15, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 17, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 20, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 May 19, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 23, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Experimental Modeled Default Modeled Metastable May 25, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 May 24, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 May 4, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 10, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 6, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 Figure 36. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for chloride at the Azalea site.

PAGE 147

125 May 14, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 16, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 May 15, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 May 17, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 20, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 May 19, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 May 23, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 Experimental Modeled Default Modeled Metastable May 25, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 24, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 May 4, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 10, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 6, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 Figure 37. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for chloride at the Gandy site.

PAGE 148

126 May 14, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 May 16, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 May 15, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 May 17, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 20, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 19, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.000 0.005 0.010 0.015 0.020 0.025 May 23, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 Experimental Modeled Default Modeled Metastable May 25, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 May 24, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 May 4, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 May 10, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 May 6, 2002 D p (um) 0.010.1110 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Figure 38. Comparison of experim ental, EQUISOLV II default mode and EQUISOLV II metastable mode data for chloride at the Sydney site.

PAGE 149

127 The comparison between the meas ured gaseous species and those modeled by EQUISOLV II are displayed in Tables15-17. Both the defaultand metastable-modeled results agreed best for the NH3 gas scenario. The model continually under pr edicted HCl and HNO3 concentrations. The comparisons for the particulate species are categorized by analyte and site in Figures 30-38. Of the thr ee volatile species, ammonium had the greatest agreement between t he modeled and measured va lues. The agreement seen for both gaseous NH3 and particulate NH4 + infers that the system is in or is near an equilibrium state. These particles are primarily fine mode species, which have the potential to reach equilibrium in a relative short period of time (Campbell et al., 2002; Moya et al., 2002). For both particulate chloride and nitr ate, their situati ons are nearly the same. The modeled and experimental particle data showed some agreement, with the metastable mode giving better re sults. The model continually over predicted particle concentrati ons and, for nitrate, o ften resulted in a size distribution mode shift. From the gas results, the modeled metastable mode under predicted the gas concentration for the majority of the m odel runs. Since the modeled gas phase is und er predicted and the m odeled particle phase is over predicted, it appears that the nitrate/nitric acid and chloride/hydrochloric acid systems during May 2002 were not at ther modynamic equilibrium. This is a different scenario from the ammonium/amm onia system as chlo ride and nitrate are predominantly coarse mode species.

PAGE 150

128 Model Comparison The EQUISOLV II model (Jacobson, 1999a) was compared against the AIM2 – Model III thermodynamic equilibr ium model (Clegg et al., 1998). The main difference between the two models lies in their method of solution. EQUISOLV II used analytical equilibrium iterations (e quilibrium constants), and AIM2 uses Gibbs free energy minimization it erations. These models were run in parallel, both in default and metastable modes, to co mpare their predicted solid and aqueous phase concentrations. The focu s of this study was to determine whether or not EQUISOLV II should be r un in the metastable mode, in which the aerosols are constrained to the aqueous phase, or the defau lt mode in which both solid and aqueous aerosols are allowed. A simple system was developed, comprising only Na+, Cland NO3 -. The first scenario tested was a 1:1 NaCl:HNO3 molar ratio system and a 1:2 NaCl:HNO3 system. The models agreed extremel y well at 60% (low) and at 90% (high) relative humidities. The systems were expanded to account for NH4 + and SO4 2-. The models no longer agreed as they did with the simple Na+, Cland NO3 system in the default mode. However, t he models were in closer agreement when run in the metastable mode. The compounds treat ed by each of the models were investigated. AIM2 was found to handle more solid compounds than EQUISOLV II. This is most likely due to the temperat ure limitations of the models. Data for equilibrium reactions are typically obtai ned under 25C conditions. The wide

PAGE 151

129 range of temperatures us ed by EQUISOLV II limit s the number of compounds treated by the model because data for many of the equations are not available at temperatures other than 25C. Actual ambient concentrations of Na+, NH4 +, K+, Mg2+, Ca2+, Cl-, NO3 and SO4 2were used to compare the model s under simulated environmental conditions. The AIM2 model does not treat K+, Mg2+ and Ca2+, so EQUISOLV II was run with and without thos e species. Simulations were run at the ambient relative humidity and at both the ambi ent temperature reco rded during sampling and at 25C. Results indi cate the EQUISOLV II model is sensitive to both temperature and t he presence of K+, Mg2+ and Ca2+. The models were most comparable at 25C and condit ions without those three species. The EQUISOLV II model predicted more aeros ol nitrate and chloride th an AIM2, which predicted higher gas phase concentration s of these species. Ag reement was greatest in the metastable mode. Through all of these simulati ons, the models best agreed in the metastable mode. Modeling in the me tastable mode improved the agreement between actual and modeled results, as seen in Fi gures 30-38. As noted in literature, metastable aeros ols are ubiquitous and are mo st likely the prevalent form in the Tampa Bay area (Rood et al., 1989).

PAGE 152

130 Case Studies The Partitioning of Nitric Acid to Nitrate Atmospheric particulate nitrate is pr imarily formed from the transfer and/or reaction of gaseous nitric acid, HNO3, onto existing particles’ surfaces (Pakkanen et al., 1996a). The most important mech anism for coarse particulate nitrate formation is the reaction between nitric acid and sea salt (NaCl) and mineral dust (CaCO3) particles. The partitioning of gas phase nitric acid and particle phase nitrate is an important process with lo cal environmental implicat ions. The change from one species to another changes the residence times and removal mechanisms and rates. The partitioning from the gas ph ase to small particle s (diameter less than 10 m) often decreases the dry deposition locally but can increase the deposition over open waters (Pryor and Sorensen, 2000). EQUISOLV II Model The EQUISOLV II model was used to ca lculate the partiti oning of nitrate between the particle and gas phases. T he specified model inputs included

PAGE 153

131 temperature, relative humidity and vary ing concentrations of sodium, chloride and calcium. The carbonate affiliated with the calcium was not directly entered into the model; instead, it was automat ically accounted for by ion balance corrections. The concent ration of NaCl and CaCO3 used in the model were similar to ambient air concentrations s een in the Tampa Bay area. The model gave predicted gas and parti cle phase concentrations in its output. The EQUISOLV II model was also used to compute the hygroscopic growth and water mass fraction of the particles. The ambient air concentrations (f rom annular denuder measurements) and meteorological data (NO AA, 2003a) were based on t he year 2000 averages. During this time period, t he averaged ambient air concentrations of nitric acid and particulate nitrate we re 1.2 and 1.7 g m-3, respectively. The average sodium concentration was 1.3 g m-3 (or 3.3 g m-3 as NaCl), and the average calcium concentration was 0.5 g m-3 (or 1.2 g m-3 as CaCO3). The dry gas and particle fluxes were calculated using Equation 1 (dV C F ). The deposition velocities for each species were calculated using the integrated NOAA Buoy – Williams m odel (Bhethanabotla, 2002) The particle diameter was set at 4 m, which is the approximat e modal diameter for the particles of interest. A weighted average of the val ues for chloride and nitrate was used based on the compos ition of the particle, computing the densitydependent particle deposition velocity. T he deposition velocity for the gas phase was computed specifically for HNO3.

PAGE 154

132 Results and Discussion The model results revealed that the ni tric acid to nitrate partitioning was dependent on the ambient air sodium a nd calcium concentrations, the total nitrate in the system (Figure 39) and re lative humidity (Figure 40). Increased concentrations of both NaCl and CaCO3 increased the fraction of nitrate in the particulate phase (Equation 27). total particleNO NO particle the in NO of Fraction ] [ ] [3 3 3 (Equation 27) However, increased concentration of the to tal available nitrate (gas plus particle phase) reduced the particulate nitrate fraction. In the NaCl example (Figure 39a), the nitrate is always divided in equilibrium between both the gas and particle phases. In the CaCO3 example (Figure 39b), the fraction of particulate nitrate significantly increas ed with a linear slope until the nitrat e was completely in the particle phase. Ga s phase nitric acid was only seen when the particulate phase was completely saturated with nitrat e, where the remainder was forced to the gas phase. At this saturation point, nitrate and calcium were present in equiequivalent amounts, which is the nano equivalents of calcium equaled the nanoequivalents of nitrate. It was obser ved that the particulate nitrate was preferentially formed in a calcium-rich environment.

PAGE 155

133 (a) Sodium[Na+] (ug m-3) 0.00.51.01.52.02.53.03.5 Fraction of NO3 in the Particle 0.0 0.2 0.4 0.6 0.8 1.0 NO3 Total = 1 ug m-3 NO3 Total = 2 ug m-3 NO3 Total = 3 ug m-3 NO3 Total = 4 ug m-3 NO3 Total = 5 ug m-3 (b) Calcium[Ca2+] (ug m-3) 0.00.51.01.52.02.5 Fraction of NO3 in the Particle 0.0 0.2 0.4 0.6 0.8 1.0 NO3 Total = 1 ug m-3 NO3 Total = 2 ug m-3 NO3 Total = 3 ug m-3 NO3 Total = 4 ug m-3 NO3 Total = 5 ug m-3 Figure 39. The partitioning of HNO3 to nitrate by (a) NaCl and (b) CaCO3 by different ambient air concentrations and total nitrate at 78% RH.

PAGE 156

134 (a) Sodium[NO3 Total] / [Na+] 110 Fraction of NO3 in the Particle 0.0 0.2 0.4 0.6 0.8 1.0 65% RH 78% RH 90% RH (b) Calcium[NO3 Total] / [Ca2+] 110 Fraction of NO3 in the Particle 0.0 0.2 0.4 0.6 0.8 1.0 65% RH 78% RH 90% RH Figure 40. The effect of relative humidity on the partitioning of HNO3 to nitrate by (a) NaCl and (b) CaCO3, where the total available nitrate was 3 g m-3.

PAGE 157

135 For NaCl, increased relative humidity increased the fraction of nitrate in the particle phase (Figure 40a). The hygr oscopic nature of the NaCl and NaNO3 particle encouraged the uptake and reaction of nitric acid with increased relative humidity. The hygroscopicity and water solubility for CaCO3 are substantially lower than those of NaCl. As a result, no (or very little) change in the nitrate particle formation with changing relative humidity was seen for CaCO3 (Figure 40b). In the figure, all three relati ve humidity scenarios are superimposed onto each other, giving one repr esentative line for CaCO3. The concentrations and mass percent of water and the fraction of nitrate within the particle as a f unction of sodium and calciu m concentrations for NaCl and CaCO3 at 78% relative humidity are disp layed in Figure 41. For NaCl and CaCO3, the amount of absorbed water increased as the fraction of particulate nitrate and sodium or calcium concent rations increased (Figure 41a-d). Therefore, the formation of NaNO3 and Ca(NO3)2 increased the hygroscopicity and, hence, the amount of absorbed water in the particle (Grassian, 2002). The water mass percent (Figure 41e-h) for Na Cl increased as particulate nitrate formation increased. However, the water mass percent for CaCO3 began to decrease as the fraction of particula te nitrate reached 1.0. As CaCO3 (MW = 100 g mol-1) is converted to Ca(NO3)2 (MW = 164 g mol-1), the amount of adsorbed water increases or remains the same; however, the reflecting water mass percent may actually decrease due to the change in molecular weight of the salt.

PAGE 158

136 (a) Sodium NO3 Total = 1 ug m-3[Na + ] (ug m -3 ) 0.00.5101.52.02.53.035 Fraction of NO 3 in the Particle 0.0 0.2 0.4 0.6 0.8 1.0 [H 2 O]Particle (ug m -3 ) 0 5 10 15 20 25 (b) Sodium NO3 Total = 5 ug m-3[Na + ] (ug m -3 ) 0.0051.01.52.02.5303.5 Fraction of NO 3 in the Particle 0.0 0.2 0.4 0.6 0.8 [H 2 O]Particle (ug m -3 ) 0 5 10 15 20 (c) Calcium NO3 Total = 1 ug m-3[Ca 2+ ] (ug m -3 ) 0.00.51.01.52.0 Fraction of NO 3 in the Particle 0.0 0.2 0.4 0.6 0.8 1.0 [H 2 O]Particle (ug m -3 ) 0.0 0.5 1.0 1.5 2.0 2.5 (d) Calcium NO3 Total = 5 ug m-3[Ca 2+ ] (ug m -3 ) 0.00.51.01.52.0 Fraction of NO 3 in the Particle 0.0 0.2 0.4 0.6 0.8 1.0 [H 2 O]Particle (ug m -3 ) 0 2 4 6 8 10 (e) Sodium NO3 Total = 1 ug m-3[Na + ] (ug m -3 ) 0.00.5101.52.02.53.035 Fraction of NO 3 in the Particle 0.0 0.2 0.4 0.6 0.8 1.0 [H 2 O]Particle (Mass %) 67 68 69 70 71 72 73 74 (f) Sodium NO3 Total = 5 ug m-3[Na + ] (ug m -3 ) 0.0051.01.52.02.5303.5 Fraction of NO 3 in the Particle 0.0 0.2 0.4 0.6 0.8 [H 2 O]Particle (Mass %) 56 58 60 62 64 66 68 70 (g) Calcium NO3 Total = 1 ug m-3[Ca 2+ ] (ug m -3 ) 0.00.51.01.52.0 Fraction of NO 3 in the Particle 0.0 0.2 0.4 0.6 0.8 1.0 [H 2 O]Particle (Mass %) 20 30 40 50 60 70 (h) Calcium NO3 Total = 5 ug m-3[Ca 2+ ] (ug m -3 ) 0.00.51.01.52.0 Fraction of NO 3 in the Particle 0.0 0.2 0.4 0.6 0.8 1.0 [H 2 O]Particle (Mass %) 56 57 58 59 60 61 Fraction of NO3 in the Particle [H2O] Figure 41. The concentration and mass perc ent of water and the fraction of total nitrate within the particle for NaCl and CaCO3 at 78% RH.

PAGE 159

137 Competitive Partitioning The competitive partitioni ng between NaCl and CaCO3 for nitrate was modeled at 78% relative humid ity (Figure 42). When t he total available nitrate (gas plus particle) was set at 1 g m-3 for a 0% Na+ and 100% Ca2+ system, the fraction of particulate nitrat e was 1.0. As the percent of sodium in the particle increased, the fraction of particulate nitrate decreased. The modeled results indicated that sodium partitioned nitrat e to the gas phase be tter than calcium, which partitioned nitrate to the particle phase. From the previous figures, it c an be concluded that both calcium and sodium play a role in the partitioning of ni tric acid to nitrate. For NaCl, there was no dominant formation of one species or the other. Calcium, on the other hand, primarily partitioned nitrate to the particle phase, forming gaseous nitric acid only when the calcium was completely satu rated with particulate nitrate. The partitioning ability of sodium and calcium is directly related to the equilibrium constants for these reactions From the EQUI SOLV II code, the equilibrium constants fo r the NaCl and CaCO3 reactions with HNO3 were determined to be: ) g ( ) s ( k ) g ( ) s (HCl NaNO HNO NaCleq 3 3 keq = 4.0 ) g ( ) aq ( k ) g ( ) s (HCl NaNO HNO NaCleq 3 3 keq = 48 ) g ( ) aq ( k ) g ( ) aq (HCl NaNO HNO NaCleq 3 3 keq = 1.3 ) ( 2 ) ( 2 ) ( 2 3 ) ( 3 ) ( 3) ( 2l g s k g sO H CO NO Ca HNO CaCOeq 1610 9 6 eqk

PAGE 160

138 ) ( 2 ) ( 2 ) ( 2 3 ) ( 3 ) ( 3) ( 2l g aq k g sO H CO NO Ca HNO CaCOeq 2210 6 4 eqk ) ( 2 ) ( 2 ) ( 2 3 ) ( 3 ) ( 3) ( 2l g aq k g aqO H CO NO Ca HNO CaCOeq 3010 2 9 eqk Despite the different keq values between the aqueous and solid phase reactions, the keq values for the calcium reactions are several orders of magnitude different than those for t he sodium reactions. The fundamental equilibrium constant parameter explai ns why calcium preferentially forms particulate nitrate over gas phase ni tric acid and why calcium carbonate partitions nitrate to the particulate phase better than sodium chloride. 0% Na6% Na12% Na25% Na50% Na100% Na Fraction of NO3 in the Particle 0.2 0.4 0.6 0.8 1.0 100% Ca94% Ca88% Ca75% Ca50% Ca0% Ca NO3 Total = 1 ug m-3 NO3 Total = 2 ug m-3 NO3 Total = 3 ug m-3 NO3 Total = 4 ug m-3 NO3 Total = 5 ug m-3 Figure 42. The competitive nitrate partitioning effect between NaCl and CaCO3, in molar percentages at 78% RH.

PAGE 161

139 Environmental Implications Deposition rates were used to assess the effects of the partitioning on the local environment. Gas and particle fluxes were calculated us ing the respective model-predicted gas and particle phase concentrations and deposition velocities. Gas phase nitric acid fluxes are displayed for NaCl and CaCO3 in Figure 43. Particulate nitrate fluxes are disp layed in Figure 44. For both NaCl and CaCO3 scenarios, the gas phase ni tric acid contributed to the majority of the local nitrogen deposition, with the nitrogen flux increasing as the total available nitrate increased. However, as the sodium or calcium concentration increased, the total nitrogen flux decreased (Figure 45). As a result of the pres ence of sodium and calcium particles, there was a decrease in the local nitrogen fl ux at conditions representative to the Tam pa Bay area. Instead of bei ng directly deposited as nitric acid, these 4 m particles can travel out of the area (~250 km) when suspended to a height of 100 m. This creates a local flux divergence as the particles are subject to horizontal transport. The gas phase nitric acid and partic ulate nitrate fluxes for the NaCl and CaCO3 mixture are reported in Figure 46. The gas phase nitric acid flux increased with increasing perce nt sodium concentrations as calcium preferred to have nitrate in the particle phase. As a result, calcium played a bigger role in creating a local nitrogen fl ux divergence than sodium.

PAGE 162

140 (a) Sodium[Na+] (ug m-3) 0.00.51.01.52.02.53.0 HNO3 Gas Flux (kg-N ha-1 yr-1) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 NO3 Total = 1 ug m-3 NO3 Total = 2 ug m-3 NO3 Total = 3 ug m-3 NO3 Total = 4 ug m-3 NO3 Total = 5 ug m-3 (b) Calcium[Ca2+] (ug m-3) 0.00.51.01.52.0 HNO3 Gas Flux (kg-N ha-1 yr-1) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 NO3 Total = 1 ug m-3 NO3 Total = 2 ug m-3 NO3 Total = 3 ug m-3 NO3 Total = 4 ug m-3 NO3 Total = 5 ug m-3 Figure 43. The predicted HNO3 gas flux for the nitric ac id partitioning by (a) NaCl and (b) CaCO3 at 78% RH.

PAGE 163

141 (a) Sodium[Na+] (ug m-3) 0.00.51.01.52.02.53.0 NO3 Particle Flux (kg-N ha-1 yr-1) 0.0 0.1 0.2 0.3 0.4 NO3 Total = 1 ug m-3 NO3 Total = 2 ug m-3 NO3 Total = 3 ug m-3 NO3 Total = 4 ug m-3 NO3 Total = 5 ug m3 (b) Calcium[Ca2+] (ug m-3) 0 00 51 01 52 0 NO3 Particle Flux (kg-N ha-1 yr-1) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 NO3 Total = 1 ug m-3 NO3 Total = 2 ug m-3 NO3 Total = 3 ug m-3 NO3 Total = 4 ug m-3 NO3 Total = 5 ug m-3 Figure 44. The predicted nitrate particle flux for the nitrate partitioning by (a) NaCl and (b) CaCO3 at 78% RH.

PAGE 164

142 (a) Sodium[Na+] (ug m-3) 0.00.51.01.52.02.53.0 Total Flux (kg-N ha-1 yr-1) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 NO3 Total = 1 ug m-3 NO3 Total = 2 ug m-3 NO3 Total = 3 ug m-3 NO3 Total = 4 ug m-3 NO3 Total = 5 ug m-3 (b) Calcium[Ca2+] (ug m-3) 0.00.51.01.52.0 Total Flux (kg-N ha-1 yr-1) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 NO3 Total = 1 ug m-3 NO3 Total = 2 ug m-3 NO3 Total = 3 ug m-3 NO3 Total = 4 ug m-3 NO3 Total = 5 ug m-3 Figure 45. The total (gas + particle) predicted flux for the nitrate partitioning by (a) NaCl and (b) CaCO3 at 78% RH.

PAGE 165

143 (a) 0% Na6% Na12% Na25% Na50% Na100% Na HNO3 Gas Flux (kg-N ha-1 yr-1) 0.0 0.5 1.0 1.5 2.0 100% Ca94% Ca88% Ca75% Ca50% Ca0% Ca NO3 Total = 1 ug m-3 NO3 Total = 2 ug m-3 NO3 Total = 3 ug m-3 NO3 Total = 4 ug m-3 NO3 Total = 5 ug m-3 (b) 0% Na6% Na12% Na25% Na50% Na100% Na NO3 Particle Flux (kg-N ha-1 yr-1) 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 100% Ca94% Ca88% Ca75% Ca50% Ca0% Ca NO3 Total = 1 ug m-3 NO3 Total = 2 ug m-3 NO3 Total = 3 ug m-3 NO3 Total = 4 ug m-3 NO3 Total = 5 ug m-3 Figure 46. The predicted (a) ni tric acid gas and (b) partic ulate nitrate flux from a molar percent mixture of NaCl and CaCO3 3 248 ] [ ] [ m neq Ca Na at 78% RH.

PAGE 166

144 The Prediction of Coarse Mode Ni trate from Fine Mode ADS Data Until recently, the monitoring of particulate matter around the Tampa Bay area only included fine particulate matte r, having a diameter based on a cyclone inlet with a cutpoint of 2.5 m (PM2.5) (Poor et al., 2001). These fine particles are responsible for visibility reductions as they scatter and absorb light more efficiently than larger particles (Sei nfeld and Pandis, 1998). These smaller particles are also responsible for seri ous health effects and mortality in humans, as well as environmental implications (Clarke et al., 1999). Coarse mode particles, however, have been found to co ntain nitrogen specie s, and they have the potential to have greater environmental implications than fine mode particles (Evans and Poor, 2001; Pryor and Bart helmie, 2000b; Pryor and Sorensen, 2000). Direct atmospheric deposition (w et and dry) of inorganic nitrogen to Tampa Bay has been estimated from 1996-1999 to be 7.3 1.3 kg-N ha-1 yr-1 or 760 140 metric tons yr-1 (Poor et al., 2001). T he dry deposition of nitrogen directly to Tampa Bay accounted for 44%, or 3.2 kg-N ha-1 yr-1. The data used to develop these estimates included one-in-s ix day ambient moni toring of gaseous ammonia, nitric acid and sulfur dioxide and fine mo de particulate ammonium, nitrate and sulfate. These species were collected using an annular denuder system (ADS) at the Gandy Bridge m onitoring station (Poor et al., 2001). The goal of this project was to ex pand the current data set by predicting the coarse mode (PM10-2.5) nitrate concentrations from the available PM2.5 data

PAGE 167

145 set. The predicted coarse mode and tota l concentrations can then be used to develop new nitrogen deposition estimates. Results and Discussion Two approaches were used to es timate the coarse mode nitrate concentrations. The first approach esti mated the coarse mode nitrate fraction from actual coarse and fine samples colle cted in the field. The second approach used the lognormal particle nitrate size di stributions to predict the fine and coarse mode fractions. For the first approach, three sets of fine and coarse mode data were used. Dichotomous coarse (PM10-2.5) and fine (PM2.5) particle samples were collected during October and November 2001 using an R&P dichotomous sampler. Total suspended particles (TSP) were also coll ected during this campaign using the inverted filter pack sampler. A third set of data was colle cted during May 2002 using an annular denuder system (PM2.5) and an open inlet automated particle sampler (>PM2.5) run by Texas Tech University ( TTU). Both sets of data were collected at the Sydney site during the in tensive May 2002 monitoring campaign.

PAGE 168

146 Dp Instrumentation Data Collection Period PM2.5 Dichotomous Sampler October 2001 PM2.5 Annular Denuder SystemMay 2002 >PM2.5* TTU Automated SamplerMay 2002 PM10 Dichotomous Sampler October 2001 >PM10* Inverted Filter Pack October 2001 True cut point is unknown Table 18. Particle size fraction, instru mentation and data co llection periods used for predicting coarse mode nitrate fractions. The samples were used to develop a relationship between the coarse (>PM2.5) and fine size fractions (Figure 47). The samples were compared through direct analysis and linear regressi on. Figures 48-50 display the linear regressions for the dichotomous fine mode and total nitrate (Figure 48), dichotomous fine mode and inverted filter pack TSP nitrate (Figure 49), and ADS fine mode and TTU coarse mode nitrate (Figure 50).

PAGE 169

147 2.5 um Particle Diameter (um) 0.1110 dC/dLogD (umol m-3) 0.00 0.01 0.02 0.03 0.04 0.05 PM2.5(Collected) >PM2.5(Missing) Figure 47. Nitrate particle size distribution. Dichotomous Total NO3 (umol m-3) 0.000.020.040.060.08 Dichotomous Fine NO3 (umol m-3) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 n = 31 slope = 0.30 + 0.02 standard error = 0.004 R2 = 0.66 October 2001 Figure 48. Linear regression for dic hotomous fine mode and total nitrate.

PAGE 170

148 Total Suspended Particle NO3 (umol m-3) 0.000.020.040.060.080.10 Dichotomous Fine NO3 (umol m-3) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 n = 29 slope = 0.22 + 0.02 standard error = 0.005 R2 = 0.54 October 2001 Figure 49. Linear regression for dichot omous fine and the inverted filter pack TSP nitrate. TTU Total Suspended Particle NO3 (umol m-3) 0.000.010.020.030.040.050.06 ADS Fine NO3 (umol m-3) 0.000 0.005 0.010 0.015 0.020 0.025 n = 45 slope = 0.40 + 0.02 standard error = 0.003 R2 = 0.33 May 2002 Figure 50. Linear regression for annular denuder system fine and TTU particle nitrate.

PAGE 171

149 Direct comparison on a per sample basis was done using the dichotomous samples and was determined by: % ] NO [ ] NO [ ] NO [ Nitrate Mode Fine %. PM Coarse PM1005 2 3 3 5 2 3 (Equation 28) This direct daily comparison analysis resulted in a percent fine mode nitrates summarized in Table 19. Fine Mode Nitrate Instrumentation Coarse Mode Nitrate Instrumentation Linear Regression Results Direct Comparison Results Dichot PM2.5 Dichot PM10-2.5 30 2% 31 12% Dichot PM2.5 Inverted FP (TSP)22 2% 21 10% ADS PM2.5 TTU sampler 40 2% 48 16% Table 19. Percent fine mode nitrate determined using actual coarse and fine mode nitrate samples. Three different coarse mode parti cle samplers were used in these analyses. Of the instrum ents, only the dichotomous sampler had a characterized PM10 inlet. The others were simply open-in let instruments with no characterized cut point. The Texas Tech University data give the greates t percentage of fine mode nitrate, whereas the in verted filter pack gives the smallest. It is apparent that the TTU instrument may have a cut diameter less than 10 m and the inverted filter pack a cut greater than 10 m. Due to the uncertainty of the cut diameters, only the result s from the dichotomous sampler were used in the prediction of coar se mode nitrate.

PAGE 172

150 The second approach was performed using lognormal analysis of sizedistributed particulate nitrate, a method used by aerosol scientists for representing quantitie s that cannot have negative va lues. The width of the distribution is used to characterize the standard deviation of t he distribution, and the height for the magnitude of the concentration. Figure 51 displays a typical lognormal distribution for parti culate nitrate in the Tampa Bay area. The sample was collected at the Gandy site on Ma y 4, 2002 using a MOUDI sampler. Particle Size (um) 0.010.1110 dC/dLogD (umol m-3) 0.00 0.01 0.02 0.03 0.04 0.05 Figure 51. Typical lognormal ni trate size distribution. The lognormal distributions were devel oped based on the three-parameter lognormal equation using SigmaPlot 2002 version 8.0 by SPSS Inc.:

PAGE 173

151 2ln 5 0 b x xoe a y (Equation 29) where 2 1 a and b, the standard deviation of the distribution. The value for ox, the particle diameter at which t he distribution peaked, was set at an averaged value of 3.15 m. The set value of ox and the resulting standard deviation, b, were used in conjunction with the lognormal function in Microsoft Excel to compute the frac tion of nitrate with 2.5 m and smaller diameter. Results are summarized in Table 20. Method # Samples Percent Fine Mode Nitrate Andersen Cascade Impactor14 40 2% MOUDI Cascade Impactor 38 36 7% Table 20. Computed perc ent fine mode nitrate using lognormal analysis of cascade impactor data. The Andersen cascade impactor is known to have poor resolution between its stages. As a result, this inst rument was not used in the prediction of coarse mode nitrate. The dichotomous sample comparis on and lognormal analysis method results were averaged, giving an approxim ate value for the percent fine mode nitrate of 33%. This estimate was then applied to the 19962002 ADS data to determine the coarse mode nitrate concentrations.

PAGE 174

152 The gas and particle over-water dry deposition velocities were calculated using the integrated NOAA Buoy – Williams model (Bhethanabotla, 2002) to determine the effects of coar se mode nitrate on the local nitrogen flux estimates. The averaged over-water deposition velocities for the gas and 3.15 m diameter particulate species were 0. 84 0.52 and 0.034 0.013 cm s-1, respectively. The particulate nitrate concentrati ons and dry deposition fluxes (dV C F ) as well as the gaseous nitric acid fluxes were ca lculated and are reported in Table 21. Concentration ( g m-3) Flux (kg-N ha-1 yr-1) HNO3 1.3 0.9 HNO3 0.76 0.51 Fine Mode NO3 0.80 0.59 CM p-NO3 0.04 0.03 Coarse Mode NO3 1.6 1.2 Total p-NO3 0.06 0.04 Total p-NO3 2.4 1.8 HNO3 + p-NO3 0.82 0.52 % CM p-NO3 6.0 4.4% % Total p-NO3 8.9 6.5% Table 21. The predicted ni trate concentrations and resulting over water dry deposition fluxes. Environmental Implications From the 1999 dry deposition estimate the fine mode particulate nitrate and nitric acid accounted for 1.9% (or 0.06 kg-N ha-1 yr-1) and 19% (or 0.61 kg-N ha-1 yr-1) of the estimated 3.2 kg-N ha-1 yr-1 nitrogen loading. The remainder resulted from gaseous ammonia (72% or 2.3 kg-N ha-1 yr-1) and fine mode particulate ammonium (6.6% or 0.21 kg-N ha-1 yr-1) (Poor et al., 2001).

PAGE 175

153 Using the predicted PM10 nitrate concentrations from this study, dry nitrogen deposition from 1999 can be estimated at 3.2 kg-N ha-1 yr-1. Total particulate nitrate accounted for 1.9% (or 0.06 kg-N ha-1 yr-1). This value is unchanged from the original estimate, as the original estimate computed the nitrogen deposition flux using a dry deposition velocity of 0.1 cm s-1, possibly overestimating the particle size. T he new size-dependent particle deposition velocity (Dp = 3.15 m) was recalculated at 0.034 cm s-1. Despite the change in the particulate nitrate concentration from the previous estimate until now, the discrepancy in the particle deposition ve locity results in an unchanged annual flux estimate. The Formation of Particulate Nitrate Experimental studies have shown that supermicron particles, those with a diameter greater than 1 m, containing nitrogen exist in our environment (Evans et al., 2002; Pakkanen et al., 1996a; Pr yor and Barthelmie, 2000b; Pryor and Sorensen, 2000). The origin of t hese compounds is unknown, but analysis reveals the possible affiliation with miner al dust or sea salt particles from the reactions with gaseous nitric acid (R eactions 1 and 3) (Clarke et al., 1999; de Leeuw et al., 2001; Dentener et al., 1996; Evans and Poor, 2001; Goodman et al., 2000; Pakkanen, 1996; Tabazadeh et al., 1998; Ten Brink, 1998; Zhuang et al., 1999a).

PAGE 176

154 Because of their size, macroparticles have a larger deposition velocity and shorter residence time than smaller particles (Pakkanen et al., 1996a). The formation of nitrate macroparticles ma y cause increased local nitrogen deposition and possibly increased eutrophication probl ems if deposited to surface waters. Using published uptake coefficients (Abbatt and Washewsky, 1998; Guimbaud et al., 2002a; Hanisch and Crowley, 2001) and a nitrate accumulation model (Kerminen and Wexler, 1995) with our rec ent measurements of inorganic aerosol distribution made near Tampa Ba y, the formation of NaNO3 and Ca(NO3)2 is estimated from reactions of NaCl sea salt and mineral dust (CaCO3) particles with nitric acid. This was a theoretical exercise to determine the potential for nitrogen-containing coarse particles to c ontribute significantly to the atmospheric nitrogen deposition to t he Tampa Bay estuary. Methods To study the theoretical formation of particulate nitrate, a closed system approach was adopted and assumed externally mixed particles of multiple size categories. Sodium and calcium were us ed as the non-volatile species. The system was initialized with a fixed amount of ni trogen (as gas phase nitric acid), which was allowed to dist ribute between the gas phase and across all particle bins. The model (described below) allowed t he particulate nitrate concentrations to be calculated in a time-step process. At each time interval, the sodium, calcium and nitric acid concentrations we re adjusted to account for the formation

PAGE 177

155 of particulate nitrate. The time st eps were allowed to continue until the particulate nitrate concentrations reached thermodynamic equilibrium with nitric acid. Experimental values were used to initialize the system The average annual nitric acid concentration was take n from one-in-six day annular denuder system (ADS) measurements made at t he Gandy site in 2000. Average particulate sodium, calcium and nitrate si ze distributions and concentrations were taken from 37 micro-orifice uniform deposit impactor (MO UDI) measurements made at three (Azalea Park, Gandy and Sydney) sites in 2 002. All of the nitrogen in the system (particula te nitrate plus nitric acid) was initialized as nitric acid. The geometric mean of seven MO UDI collection bins was used for the particle classification. The particle di ameters compared were: 0.2, 0.49, 0.77, 1.4, 2.5, 4.3 and 7.6 m. The species concentrations used in this modeling exercise are listed below in Table 22. GeoMean Bin Max] Na [ ] Ca [2 ] NO [3 ] HNO [3 ( m) ( m) ( mol m-3)( mol m-3)( mol m-3)( mol m-3) 0.20 0.40 5.0E-05 3.0E-04 3.0E-02 2.0E-02 0.49 0.60 8.0E-05 1.1E-04 0.77 1.0 5.5E-04 1.5E-04 1.4 2.0 4.0E-03 8.2E-04 Total nitrogen: 2.5 3.2 9.4E-03 1.8E-03 0.05 mol m-3 4.3 5.7 2.3E-02 4.0E-03 7.6 10 1.4E-02 2.6E-03 Table 22. Averaged concent ration from 37 MOUDI experimental samples and year 2000 ADS measurements.

PAGE 178

156 Deposition Velocities The over water deposition velocities of each particle size groups and gaseous nitric acid were calculated us ing the integrated NO AA Buoy-Williams model (Bhethanabotla, 2002) and the year 2000 hourly meteorological data. The integrated NOAA Buoy-Williams model is a combination of a two-layer multiplepath model for dry deposition of particles to surface waters (Williams) and an iterative bulk exchange model for mom entum, heat and moisture (NOAA Buoy). The model includes the effects of wave breaking, particle hygroscopic growth and turbulent heat flux. The dry depositi on velocities were calculated on the basis of the turbulent heat transfer and gr avitational settling of particles. The meteorological data used in the model included su rface weather and water observations, which were obtained from t he NOAA National Climatic Data Center (NOAA, 2003a) and the NOAA Na tional Ocean Service’s C enter for Operational Oceanographic Products and Services (NO AA, 2003b), respectively. Surface weather observations were collected at the Tampa Inte rnational Airport, located approximately 11 km from the Gandy sa mpling site. Water measurements were taken at NOAA’s Clearwater Beach station, located on the Gulf of Mexico coastline approximately 30 km from t he Gandy site. The meteorological data were divided into three categories based on wind speed. The low (<2.4 m s-1) and high (>6.0 m s-1) wind speeds were classified as the respective 25th and 75th percentile winds during 20 00. The midrange wind speeds were those between

PAGE 179

157 2.4 and 6.0 m s-1. The resulting dry deposition va lues were averaged, giving an annual deposition velocity (Table 23). Dp Deposition Velocity, Vd (cm s-1) ( m) Low WS Mid WS High WS 0.20 0.002 0.0010.006 0.0030.03 0.02 0.49 0.003 0.0020.005 0.0020.02 0.02 0.77 0.006 0.0040.007 0.0040.02 0.02 1.4 0.006 0.0030.007 0.0030.02 0.02 2.5 0.020 0.0000.02 0.00 0.03 0.02 4.3 0.06 0.00 0.06 0.00 0.07 0.06 7.6 0.17 0.00 0.17 0.00 0.25 0.17 HNO3 0.28 0.17 0.67 0.23 1.48 0.41 Table 23. Annual averaged over water dry deposition velocities with their respective standard deviations. Residence Times and Distances Traveled The distance a particle will travel is dependent on the particle’s deposition velocity, starting height and wind speed. For all scenarios tested, the starting height was set at 100 m. The traveling distance, TD, was calculated using: Velocity Deposition Height Starting Speed Wind TD (Equation 30) The traveling distance was calculat ed for three different wind scenarios during 2000. Table 24 displays the dist ance traveled by a particle at the three wind scenarios. The residence time, RT, of a particle (Tabl e 24) is the time required by a particle to deposit through dr y deposition to the su rface of a body of

PAGE 180

158 water. It is only dependent on the starti ng height and the depo sition velocity of the particle, and it wa s calculated using: Velocity Deposition Height Starting RT (Equation 31) Dp Distance Traveled (km) Residence Time (hr) ( m) Low WSMid WSHigh WSLow WSMid WSHigh WS 0.20 730 7100 3700 1300 470 110 0.49 4800 8600 5200 840 570 160 0.77 2800 6200 5400 480 410 170 1.41 2800 6000 5200 490 400 160 2.53 790 2000 3000 140 130 91 4.27 290 750 1200 50 49 38 7.55 94 250 350 16 16 11 Table 24. Traveling distances an d residence times for low (<2.4 m s-1), mid (2.46.0 m s-1) and high (>6.0 m s-1) wind speeds. Reaction with Nitric Acid The rate at which particulate nitrat e is formed from nitric acid depends on several interacting processes (Pakkanen et al., 1996a). Under the assumption of spherical particles and su rface-limited chemical reactions, the rate at which nitrate accumulates into a single particle, 3NOI (mol s-1) can be obtained from the relation (Kerminen and Wexler, 1995; Pakkanen et al., 1996a):

PAGE 181

159 1 ] [ 2 ) () ( 33 3g p HNO p NOHNO d D d I (Equation 32) where p s HNOd k D 32, pd is the particle diameter, DHNO3 (m2 s-1) is the diffusion coefficient of nitric acid, ] [3HNO (mol m-3) is the gas phase concentration, and ks (m s-1) is the surface rate const ant for the conversion of HNO3 to particulate nitrate ( 4 c ks, where c is the molecular speed of HNO3 and is the reaction uptake coefficient). The total concentration of particulate ni trate as a result of uptake of nitric acid by sea salt over a certain size range, j, can be obtained by integrating 3NOI over all particles in j and time, Tj (s) (Pakkanen et al., 1996a): ] [ 1000 ] [ 6 ] [) ( 3 ) ( ) ( ) ( ) ( ) ( 3 g Na ss p ss j p ss j ss j s Na j ss jHNO f d T k M Na NO (Equation 33) The nitrate concentration as a result of th e uptake of nitric acid by calcium in mineral dust can be obtained by (Pakkanen et al., 1996a): ] [ 1000 ] [ 6 ] [) ( 3 ) ( ) ( ) ( ) ( 2 ) ( 3 g Ca md p md j p md j md j s Ca j md jHNO f d T k M Ca nss NO (Equation 34) where ] [Na (mol m-3) is the concentration of sodium, ] [2 Ca nss (mol m-3) is the concentration of non-sea salt calciu m (that affiliated with mineral dust), Mi (g mol-1) is the molar mass of species i, dp,j ( m) is the representative particle diameter in the size range, p (kg m-3) is the particle density and fi is the mass fraction of sodium in sea salt and calcium in mineral dust particles.

PAGE 182

160 The underlying assumptions for this model include (Pakkanen, 1996): (1) These particles consist of two externally mixed particle types on which nitrate is formed: sea salt and mineral dust. (2) T he water-soluble sodium originates only from sea salt and calcium from mineral du st. (3) Particles of the same type and size have roughly the same co mposition. (4) The size distribution function is temporally constant over the size range. (5) The heter ogeneous conversion of NO2 and other nitrate pr ecursor gases is negligible co mpared to nitric acid. (6) The nitrate accumulation in both sea salt an d mineral dust particles is limited by the rate ks. This model considers only sodium and calcium as reactive species for nitric acid and is limited by not consideri ng the reactions between nitric acid and other particulate matter types. However, since little or no data are available for the reactions between nitric acid and othe r species, they were excluded. The particle density and mass fraction were held constant for each time step. The hygroscopicity for NaNO3 closely resembles that of NaCl, allowing for the assumption that the mass fracti on and density remained constant. The uptake coefficients for HNO3 on NaCl, sea salt and CaCO3 have been determined experimentally. For NaCl, t he uptake coefficient was characterized by the value 2 0. (Abbatt and Washewsky, 1998). This value was determined using an aerosol kinetics flow tube te chnique at room temperature with deliquescent NaCl at 75% re lative humidity. Other upt ake coefficients for NaCl have been determined using a low-pressure Knudsen cell flow reactor: ] 10 ) 6 0 4 1 ( [2 (Beichert and Finlayson-Pitts, 1996) and

PAGE 183

161 ] 10 ) 3 0 8 2 ( [2 (Fenter et al., 1994). The Knudsen cell flow reactor is limited to very low relative humidity c onditions. The val ue determined by Abbatt and Waschewsky (1998) was used in this modeling and is considered a lower limit for the uptake of HNO3 by NaCl as the average relative humidity for the Tampa Bay area is approximately 80%. The uptake coefficient for deliquesc ent sea salt was estimated to be 20 0 50 0 at 55% relative humidity (Guimbaud et al., 2002a). The difference between the coefficients for sea salt and NaCl lie in the composition of sea salt. Sea salt contains hy groscopic hydrates, such as MgCl2•6H2O, which provide additional surface waters for the uptake and reaction of HNO3. Other uptake coefficient values have been determ ined for sea salt by De Haan and Finlayson-Pitts (1997) ] 2 0 [ using a Knudsen cell at low relative humidity. Uptake coefficients for CaCO3 have been determined under a few conditions. At 0% and 20% relative hu midities, they were estimated to be 410 4 2 and 310 5 2 respectively (Goodman et al., 2000; Grassian, 2002). Hanisch and Crowle y (2001) determined the uptak e coefficients for “dry” heated and “damp” unheated CaCO3 to be 210 ) 5 2 10 ( and 210 ) 5 4 18 ( respectively. The value estimated under “damp” conditions by Hanisch and Crowley (2001) is thought to be more relevant under atmospheric conditions. Their reported va lue is a considered a lower limit for the uptake of HNO3 by CaCO3.

PAGE 184

162 The mass fractions of sodium or calciu m in the original NaCl, sea salt or mineral dust particles were estimated us ing the EQUISOLV II model, an aerosol thermodynamic equilibrium model (Campb ell et al., 2002; Jacobson, 1999a). Input files were created using the com position of dry particles and varying temperature and relative hum idity. EQUISOLV II ca lculated the equilibrium concentrations of the particl es, including the amount of absorbed water. From a series of modeling runs, te mperature was not found to have a significant effect on the mass fraction of sodium or calcium. However, the relative humidity had the greatest effect on the mass fraction. Due to the hygr oscopicity of the salts, the quantity of adsorbed water depends on t he relative humidity. The mass fraction was calculated for four different relative humidity scenarios. At 60% relative humidity, the parti cles were treated as dry so lids, while those above 75% relative humidity were deliquescent aeros ols. Table 25 gives the mass fraction of sodium and calcium within NaCl, sea sa lt and mineral dust particles at varying relative humidities. Mass Fraction RH Na+ Ca2+ (%) in NaCl in sea salt in dust 60 0.39 0.19 0.15 78 0.10 0.07 0.13 90 0.06 0.05 0.11 95 0.03 0.03 0.09 Table 25. Mass fractions of sodium and calcium in NaCl, sea salt and mineral dust.

PAGE 185

163 As with the mass fraction, the densit y of the particles changes with the uptake of water. The densities were ca lculated as follows. The derivation of these equations is given in Appendix 3. Th e density of the aqueous phase of the particle was calculated using: w w s s aqw w 1 (Equation 35) where s and w are the respective densities of the solute (salt) and water and ws and ww are the respective mass fractions of the solute and water. The density of the entire aerosol partic le was calculated using: solid solid aq aq partF F 1 (Equation 36) where Faq and Fsolid are the respective mass fr actions of the aqueous and solid phases within the entire particle and aq and solid are the respective densities of the aqueous and solid phases. The resulti ng density values ar e listed in Table 26. RH Density (g cm-3) (%) NaCl Sea salt CaCO3 60 1.3 2.2 1.3 78 1.2 1.2 1.2 90 1.1 1.1 1.1 95 1.0 1.1 1.1 Table 26. Calculated densities for NaCl, sea salt and CaCO3 at varying relative humidities.

PAGE 186

164 The Equisolv II model was also used to compute the thermodynamic equilibrium concentrations of particle ni trate. The modeled results were dependent on the relative humidity and the composition of the particle. The reactions of NaCl and sea salt with nitric acid we re greatly affected by the equilibrium, as the predicted keq is several orders of m agnitude smaller than that of mineral dust. The large keq value for mineral dust di d not limit the reaction between CaCO3 and nitric acid, which was allow ed to react until the particle was completely saturated with nitrate.

PAGE 187

165 NaCl 0.8 umResidence Time (hr) 0.0020.40.6081.0 [NO3 -] (umol m-3) 0 1e-4 2e-4 3e-4 [NO3 -]:[Na+] 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 NaCl 0.5 umResidence Time (hr) 000.10.2030.40.5 [NO3 -] (umol m-3) 0 1e-5 2e-5 3e-5 4e-5 5e-5 [NO3 -]:[Na+] 00 0.1 02 03 0.4 05 06 0.7 NaCl 0.2 umResidence Time (hr) 0.00.10.20.30.40.5 [NO3 -] (umol m-3) 0 1e-5 2e-5 3e-5 [NO3 -]:[Na+] 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 HNO3Residence Time (hr) 0246810 [NO3 -] (umol m-3) 001 002 003 004 005 % HNO3 in gas phase 20 40 60 80 100 NaCl 7.6 umResidence Time (hr) 0246810 [NO3 -] (umol m-3) 0.000 0.002 0.004 0.006 0.008 [NO3 -]:[Na+] 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.760% RH 78% RH 90% RH 95% RH NaCl 4.3 umResidence Time (hr) 012345 [NO3 -] (umol m-3) 0.000 0.004 0.008 0.012 [NO3 -][Na+] 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 NaCl 2.5 umResidence Time (hr) 0.0051.0152.02.53.0 [NO3 -] (umol m-3) 0000 0001 0002 0003 0004 0005 0006 [NO3 -][Na+] 00 0.1 02 03 0.4 05 06 0.7 NaCl 1.4 umResidence Time (hr) 0.00.20.406081.01.21.4 [NO3 -] (umol m-3) 0.0 5.0e-4 1.0e-3 1.5e-3 2.0e-3 2.5e-3 [NO3 -]:[Na+] 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Figure 52. Time-resolved nitrate formation for NaCl.

PAGE 188

166 Sea Salt 0.8 umResidence Time (hr) 0.0020.40.6081.0 [NO3 -] (umol m-3) 0 1e-4 2e-4 3e-4 4e-4 5e-4 [NO3 -]:[Na+] 0.0 0.2 0.4 0.6 0.8 1.0 Sea Salt 0.5 umResidence Time (hr) 000.10.20.30.4050.6 [NO3 -] (umol m-3) 0 2e-5 4e-5 6e-5 8e-5 [NO3 -]:[Na+] 00 02 0.4 06 08 10 Sea Salt 0.2 umResidence Time (hr) 0.0 0.1 0.2 0.3 0.4 [NO3 -] (umol m-3) 0 1e-5 2e-5 3e-5 4e-5 5e-5 [NO3 -]:[Na+] 0.0 0.2 0.4 0.6 0.8 1.0 HNO3Residence Time (hr) 012345 [NO3 -] (umol m-3) 000 001 002 003 004 005 % HNO3 in gas phase 0 10 20 30 40 50 60 70 80 90 100 Sea Salt 7.6 umResidence Time (hr) 0246810 [NO3 -] (umol m-3) 0.000 0.002 0.004 0.006 0.008 0.010 0.012 [NO3 -]:[Na+] 0.0 0.2 0.4 0.6 0.8 1.060% RH 78% RH 90% RH 95% RH Sea Salt 4.3 umResidence Time (hr) 01234 [NO3 -] (umol m-3) 0.000 0.004 0.008 0.012 0.016 0.020 [NO3 -][Na+] 0.0 0.2 0.4 0.6 0.8 1.0 Sea Salt 2.5 umResidence Time (hr) 0.0 0.5 10 1.5 2.0 [NO3 -] (umol m-3) 0 2e-3 4e-3 6e-3 8e-3 [NO3 -][Na+] 00 02 0.4 06 08 10 Sea Salt 1.4 umResidence Time (hr) 0.0020.40.60.81.012 [NO3 -] (umol m-3) 0.0 1.0e-3 2.0e-3 3.0e-3 [NO3 -]:[Na+] 0.0 0.2 0.4 0.6 0.8 1.0 Figure 53. Time-resolved nitr ate formation for sea salt.

PAGE 189

167 CaCO3 0.8 umResidence Time (hr) 0 0051 01 52 0 [NO3 -] (umol m-3) 0.0 1.0e-4 2.0e-4 3.0e-4 [NO3 -]:[Ca2+] 0.0 0.5 1.0 1.5 2.0 CaCO3 0.5 umResidence Time (hr) 00 0.5 10 1.5 2.0 [NO3 -] (umol m-3) 0.0 5.0e-5 1.0e-4 1.5e-4 2.0e-4 2.5e-4 [NO3 -][Ca2+] 00 05 10 15 20 CaCO3 0.2 umResidence Time (hr) 0.00.20.40.60.810 [NO3 -] (umol m-3) 0 2e-4 4e-4 6e-4 [NO3 -]:[Ca2+] 0.0 0.5 1.0 1.5 2.0 HNO3Residence Time (hr) 0 5 10 15 20 [NO3 -] (umol m-3) 2.5e-2 3.0e-2 3.5e-2 4.0e-2 4.5e-2 5.0e-2 % HNO3 in gas phase 50 60 70 80 90 100 CaCO3 7.6 umResidence Time (hr) 0510152025 [NO3 -] (umol m-3) 0 1e-3 2e-3 3e-3 4e-3 5e-3 [NO3 -]:[Ca2+] 0.0 0.5 1.0 1.5 2.060% RH 78% RH 90% RH 95% RH CaCO3 4.3 umResidence Time (hr) 02468101214 [NO3 -] (umol m-3) 0 2e-3 4e-3 6e-3 8e-3 [NO3 -]:[Ca2+] 0.0 0.5 1.0 1.5 2.0 CaCO3 2.5 umResidence Time (hr) 02468 [NO3 -] (umol m-3) 00 10e-3 20e-3 30e-3 40e-3 [NO3 -]:[Ca2+] 0.0 0.5 1.0 1.5 2.0 CaCO3 1.4 umResidence Time (hr) 01234 [NO3 -] (umol m-3) 00 50e-4 10e-3 15e-3 [NO3 -]:[Ca2+] 0.0 0.5 1.0 1.5 2.0 Figure 54. Time-resolved nitrate formation for CaCO3.

PAGE 190

168 Results and Discussions The model was run in time steps until the particulate phase reached the thermodynamic equilibrium concentration of nitrate that was predicted by the EQUISOLV II model. The results are giv en in Figures 51-53 for NaCl, sea salt and CaCO3, respectively. The concentration of particulate ni trate was extrapolated at the particle residence time from the time-resolved curves given in Figures 51-53. For NaCl and sea salt, the particles reached the equ ilibrium nitrate concentration at speeds much faster than their residence ti mes. For calcium, the extrapolated concentrations were compared to the saturation equation: % 100 ]) [ ] ([ Reacted %) ( 3 (mod) 3 equilNO NO (Equation 37) where ] [(mod) 3NO (mol m-3) is the model-predicted ni trate concentration at a given relative humidity, and ] [) ( 3equilNO (mol m-3) is the equilibrium particle nitrate concentration as predicted by the EQUISOLV II model. This was done to determine the extent of the reaction at the residence time Those particles with a nitrate concentration less than calcium (i n microequivalents) were assumed to react only partially. The results are given in Table 27.

PAGE 191

169 Percent Reacted Nitrate at Reside nce Time (Time of Deposition) Ca2+ in mineral dust Low Wind Speed Mid Wind Speed High Wind Speed Dp 60% 78% 90% 95% 60%78%90%95% 60% 78% 90%95% 0.2 0.5 100% Reacted 0.8 1.4 2.5 4.3 7.6 94% 97% 99% 99% 94%96%99%99% 85% 90% 95%97% Table 27. The percent of particulate ni trate formation based on an initial height of 100 m at the residence time and differ ent ambient relative humidity values. Under the given conditions, the majority of the particles reached the nitrate equilibrium concentration at or before thei r residence time. The uptake of nitrate was greatest at 95% relative humidit y and low wind speed conditions. The largest particle size bin coll ected particles up to 10 m in diameter. According to Table 27, these particles contain a significant amount of nitrate. Figure 55 displays the typical size distribution for sodium, calcium and nitrate in the Tampa Bay area. All thr ee species are primarily in the coarse fraction, with the distri bution beginning at ~0.5 m. Upon reaction with sodium or calcium, particulate nitrate exhibits a simi lar size distribution as its parent NaCl or CaCO3 particle.

PAGE 192

170 Particle Size (um) 0.010.1110 dC/dLogD (umol m-3) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Na+ Ca2+ NO3 Figure 55. Typical size di stribution of sodium, calciu m and nitrate in the Tampa Bay area. The aerosol sampling devices used at our bayside sampling site have an upper collection limit of 10 m particles. Several studies have been conducted measuring the size distribution of marine aerosol (de Leeuw, 1986; Haaf and Jaenicke, 1980; Hoppel et al., 1989; Me szaros and Vissy, 1974; Seinfeld and Pandis, 1998). These studies reveal f our major modes of marine aerosol: (a) 0.05-0.06, (b) 0.2-0.3, (c) 6-7 and (d) 11-15 m. Continental aerosol size distribution has also been characterized (Jaenicke, 1993), revealing three major modes: (a) 0.01-0.02, (b ) 0.15-0.25 and (c) 6-7 m. In Figure 55, the particle size distribution at 10 m begins to show what may be the beginning of the largest mode (11-15 m) of marine aerosol.

PAGE 193

171 Particles with a geometric mean diameter of 15 m have an approximate deposition velocity and resi dence time of 0.7 cm s-1 and 4 hr, respectively. From Figures 51-52, all particles in the larges t size bin reached equilibrium in less than 8 hours, with the equilibrium time significa ntly less for deliquescent aerosols. If the modeling exercises were to continue for the largest mode of marine aerosols, results may indicate that the particle s would contain a significant amount of nitrate even though they ma y not reach complete equilibrium before their residence time. Impact on Nitrogen Loading To determine if these particles c ould have an environm ental impact, their effect on the total nitrogen loading th rough dry deposition was examined. Particle and gas deposition velocities we re calculated using the integrated NOAA Buoy-Williams model (B hethanabotla, 2002). The dry deposition flux was calculated using Equation 1, where dV C F During this modeling exercise, 0.05 mol m-3 of nitrate was partitioned between the gas and particle phases. If all of this nitrate were to remain in the gas phase as nitric acid, the resulting fl uxes for the low, mid and high wind speed conditions would be 0.61, 1.5 and 3.3 kg-N ha-1 yr-1, respectively. However, particles have lower deposit ion velocities and longer residence times than nitric acid, giving them the ability to create a flux divergence as nitric acid is transferred

PAGE 194

172 Nitric Acid and Particulate Nitrate Flux (kg-N ha-1 yr-1) Na+ in NaCl Low Wind Speed Mid Wind Speed High Wind Speed Dp Relative Humidity Relative Humidity Relative Humidity ( m) 60% 78% 90%95% 60%78%90%95% 60% 78% 90%95% 0.2 2 E-6 3 E-6 3 E-63 E-6 6 E-67 E-68 E-67 E-6 2 E-5 3 E-5 3 E-53 E-5 0.5 5 E-6 6 E-6 7 E-67 E-6 7 E-69 E-61 E-51 E-5 3 E-5 3 E-5 4 E-53 E-5 0.8 6 E-5 7 E-5 8 E-58 E-5 7 E-58 E-51 E-49 E-5 2 E-4 2 E-4 2 E-42 E-4 1.4 4 E-4 5 E-4 6 E-46 E-4 5 E-46 E-47 E-47 E-4 1 E-3 2 E-3 2 E-32 E-3 2.5 4 E-3 4 E-3 5 E-35 E-3 4 E-35 E-35 E-35 E-3 5 E-3 7 E-3 8 E-37 E-3 4.3 2 E-2 3 E-2 3 E-23 E-2 2 E-23 E-23 E-23 E-2 3 E-2 4 E-2 4 E-24 E-2 7.6 5 E-2 5 E-2 6 E-26 E-2 5 E-25 E-26 E-26 E-2 7 E-2 8 E-2 9 E-29 E-2 Total Particle 0.07 0.09 0.1 0.1 0.070.090.1 0.1 0.11 0.13 0.150.14 Gas + Particle 0.42 0.38 0.350.36 0.910.8 0.7 0.741.95 1.69 1.461.55 % Reduction 69% 62% 57%59%61%53%47%49%59% 51% 44%47% Na+ in sea salt Low Wind Speed Mid Wind Speed High Wind Speed Dp 60% 78% 90%95% 60%78%90%95% 60% 78% 90%95% 0.2 3 E-6 3 E-6 4 E-63 E-6 8 E-69 E-69 E-69 E-6 3 E-5 4 E-5 4 E-54 E-5 0.5 7 E-6 8 E-6 8 E-68 E-6 1 E-51 E-51 E-51 E-5 4 E-5 4 E-5 4 E-54 E-5 0.8 9 E-5 9 E-5 1 E-41 E-4 1 E-41 E-41 E-41 E-4 2 E-4 3 E-4 3 E-43 E-4 1.4 6 E-4 7 E-4 7 E-47 E-4 7 E-48 E-49 E-48 E-4 2 E-3 2 E-3 2 E-32 E-3 2.5 5 E-3 6 E-3 6 E-36 E-3 5 E-36 E-36 E-36 E-3 8 E-3 8 E-3 9 E-39 E-3 4.3 3 E-2 4 E-2 4 E-24 E-2 3 E-24 E-24 E-24 E-2 4 E-2 5 E-2 5 E-25 E-2 7.6 6 E-2 7 E-2 8 E-27 E-2 6 E-27 E-28 E-27 E-2 9 E-2 1 E-1 1 E-11 E-1 Total Particle 0.1 0.11 0.120.12 0.1 0.110.120.12 0.15 0.16 0.180.17 Gas + Particle 0.34 0.32 0.290.3 0.680.610.530.561.41 1.26 1.071.14 % Reduction 56% 52% 48%49%45%41%35%37%43% 38% 32%35% Ca2+ in dust Low Wind Speed Mid Wind Speed High Wind Speed Dp 60% 78% 90%95% 60%78%90%95% 60% 78% 90%95% 0.2 6 E-5 6 E-5 6 E-56 E-5 2 E-42 E-42 E-42 E-4 7 E-4 7 E-4 7 E-47 E-4 0.5 3 E-5 3 E-5 3 E-53 E-5 5 E-55 E-55 E-55 E-5 2 E-4 2 E-4 2 E-42 E-4 0.8 8 E-5 8 E-5 8 E-58 E-5 9 E-59 E-59 E-59 E-5 2 E-4 2 E-4 2 E-42 E-4 1.4 4 E-4 4 E-4 4 E-44 E-4 5 E-45 E-45 E-45 E-4 1 E-3 1 E-3 1 E-31 E-3 2.5 3 E-3 3 E-3 3 E-33 E-3 3 E-33 E-33 E-33 E-3 5 E-3 5 E-3 5 E-35 E-3 4.3 2 E-2 2 E-2 2 E-22 E-2 2 E-22 E-22 E-22 E-2 3 E-2 3 E-2 3 E-23 E-2 7.6 4 E-2 4 E-2 4 E-24 E-2 4 E-24 E-24 E-24 E-2 5 E-2 5 E-2 6 E-26 E-2 Total Particle 0.06 0.06 0.060.06 0.060.060.060.06 0.09 0.09 0.090.09 Gas + Particle 0.44 0.44 0.440.44 0.970.970.970.972.12 2.11 2.1 2.1 % Reduction 72% 72% 72%72% 65%65%65%65%64% 64% 64%64% Table 28. The calculated nitrogen over water dry deposition flux.

PAGE 195

173 to the particle phase. The resulting parti cle and total flux values are reported in Table 28. Due to their relatively higher sodium or calcium concentration and greater deposition velocity, particles 4.3 m and greater contributed to the majori ty of the particle dry deposition flux, especially when the relative humidity was 90%. The gas phase nitric acid dominated the total flux and was gr eatest at high wind speed conditions. Low ambient concentra tions of calcium resulted in lower nitrogen flux contributions fr om mineral dust with respec t to sea salt or NaCl. Flux divergence of HNO3 in the marine boundary layer has been modeled by Pryor and Sorensen (2000) to determi ne the importance of nitric acid reactions on sea salt particles. Their results indicated that under near-neutral stability and wind speeds between 3.5 and 10 m s-1, the transfer of nitric acid to the particle phase decreased t he deposition velocity of nitrogen by over 50%. This transfer to the particle phase led to greater horizontal transport prior to deposition of the nitrogen part icles. For low and high wind speeds (<3.5 and >10 m s-1), the transfer of nitric acid to the particle phase increased the deposition rate and decreased the horizontal transport.

PAGE 196

174 Conclusions A better understanding of the characte ristics and formation of particulate nitrate is essential for the development of nitrogen loading estimates and control. Nitrogen particle size distributions hav e been identifi ed and investigated during the course of this study. Through a network of ambient air sa mpling, a snapshot of our local environment was created giving us estima tes of particle concentrations and size distributions. The data collected durin g a series of intensive monitoring campaigns allowed the formation and charac terization of the nitrate particles to be deduced. This work researched the in teraction between nitric acid and sea salt, focusing on the chemistry and its effect on the local nitrogen deposition estimates. The partitioning of nitric acid to pa rticulate nitrate was examined for NaCl and CaCO3. Both salts exhibited an increase in particulate nitrate formation with an increase in sodium or calcium and total available nitrate concentrations in the system. The extent of parti culate nitrate formation wa s directly related to each equation’s equilibrium constant. Calcium has an equilibrium constant several orders of magnitude gr eater than sodium, pa rtitioning nitrate to the particle phase until its saturation concentration. The water content of the particles was

PAGE 197

175 examined. Both nitrate salts were more hygroscopic than their parent salt. As nitrate was partitioned to the particle, th e amount of absorbed water increased. Nitrogen dry deposition flux estimates we re calculated. For both NaCl and CaCO3 scenarios, the gas phase ni tric acid contributed to the majority of the local nitrogen deposition with the nitrogen flux in creasing as the total available nitrate increased. However, as the sodium or calcium concentration increased, the total nitrogen flux decreased. The particles cr eated a local nitrogen flux divergence as they have longer residence time s and greater horizontal transport. The coarse mode particulate nitrate was predicted using fine mode nitrate data and new dry deposition estimates we re calculated. Analysis from dichotomous samples and lognormal di stributions obtained with a cascade impactor resulted in an approximate value for percent fine mode (PM2.5) nitrate of 33%. The coarse mode (PM10-2.5) nitrate accounted for 0.04 of the 0.06 kg-N ha-1 yr-1 of the dry deposition flux. The gas plus particle dry deposition flux was estimated at 3.2 kg-N ha-1 yr-1. This value is unchanged from the previous estimate, but it was calculated using coarse mode particulate concentrations and size-dependent particle deposition velocities. The theoretical formation of particles ni trate from the reacti on of nitric acid with NaCl, sea salt and mineral dust (CaCO3) was estimated using a nitrate accumulation model and a t hermodynamic equilibrium model The extent of the reaction for NaCl and sea sa lt was limited to the pr edicted equilibrium nitrate concentration, as the keq values for the NaCl and sea salt reactions are relatively small when compared to that of mineral dust. The equilibrium concentration was

PAGE 198

176 determined to be dependent on the relative humidity. Size distributions for marine and mineral dust aerosols were used to determine the major modes for these species. Marine aerosols were found to have a macroparticle mode greater than the cutoff of our instru ments. Ambient sodium and calcium concentrations are unavailable for particles greater than 10 m. However, the use of this modeling has in ferred that these particles, when mixed with urban air, may contain a significant amount of nitrate despite their relatively short residence time. The particle contribution to the ni trogen dry deposition fl ux was examined. The gas phase nitric acid contributed to the majority of the nitrogen dry deposition flux, with particles only cont ributing a small perc entage. The highest particle contribution o ccurred under high humidity and high wind speed conditions. However, for mineral dust, the low ambient concentrations of calcium resulted in lower particulate nitrate concen trations with respect to sodium. Our work was found to be consistent with that of others, where t he transfer of nitric acid to particle phase nitrate decr eased the deposition of locally produced species, as the particles have greater horizontal transport and residence times.

PAGE 199

177 References Abbatt, J.P.D., Washewsky, G.C.G., 1998. Heterogeneous interactions of HOBr, HNO3, O3, and NO2 with deliquesce nt NaCl aerosols at room temperature. Journal of Ph ysical Chemistry A 102, 3719-3725. Aherne, J., Farrell, E.P., 2002. Deposition of sulphu r, nitrogen and acidity in precipitation over Ireland: chemistr y, spatial distri bution and long-term trends. Atmospheric Environment 36, 1379-1389. Ali-Mohamed, A.Y., Jaffar, A.H., 2000. Estimation of atmospheric inorganic water-soluble aerosols in the we stern region of Bahrain by ion chromatography. Chemosphere Global Change Science 2, 85-94. Allegrini, I., De Sanits, F., Di Palo, V ., Febo, A., Perrino, C., Possanzini, M., Liberti, A., 1987. Annular denuder met hod for sampling reactive gases and aerosols in the atmospher e. The Science of the Total Environment 67, 116. Allegrini, I., Febo, A., Perrino, C., Masi a, P., 1994. Measurem ent of atmospheric nitric acid in gas phas e and nitrate in particul ate matter by means of annular denuders. Internat ional Journal of Envi ronmental Analytical Chemistry 54, 183-201. Allen, A.G., Harrison, R. M., Erisman, J.W., 1989. Fi eld measurements of the dissociation of ammonium nitrat e and ammonium chloride aerosols. Atmospheric Environment 23, 1591-1599. Allen, A.G., Miguel, A.H., 1995. Aerosol chemistry in the urban atmosphere of Sao Paulo, Brazil. Journal of Aerosol Science 26, S375-S376. Allen, H.C., Laux, J.M., V ogt, R., Finlayson-Pitts, B.J., Hemminger, J.C., 1996. Water-induced reorganization of ultr athin nitrate films on NaCl: Implications for the tropospheric chemistr y of sea salt particles. Journal of Physical Chemistry 100, 6371-6375. Allen, H.C., Mecartney, M.L., Hemminger, J.C., 1998. Minimizing transmission electron microscopy beam damage during the study of surface reactions on sodium chloride. Microsc opy and Microanalysis 4, 23-33.

PAGE 200

178 Andreas, E.L., Edson, J.B., Monahan, E. C., Rouault, M.P., Sm ith, S.D., 1995. The spray contribution to net evaporation from the sea: A review of recent progress. Boundary-Layer Meteorology 72, 3-52. Anlauf, K.G., Wiebe, H.A., Fellin, P., 1986. Characterization of several integrative sampling methods for nitric acid sulphur dioxide and atmospheric particles. Journal of the Air Pollu tion Control Association 36, 715-723. Ansari, A.S., Pandis, S.N., 1999. Pr ediction of multic omponent inorganic atmospheric aerosol behavior. Atmo spheric Environment 33, 745-757. Appel, B.R., Povard, B., Kothny, E.L., 1988. Loss of nitric acid within inlet devices intended to exclude c oarse particles during atmospheric sampling. Atmospheric Environment 22, 2535-2540. Appel, B.R., Tokiwa, Y., Haik, M., Kothny E.L., 1984. Artifact particulate sulfate and nitrate formation on f ilter media. Atmospher ic Environment 18, 409416. Appel, B.R., Wall, S.M., Tokiwa, Y., Ha ik, M., 1980. Simultaneous nitric acid, particulate nitrate and ac idity measurements in ambient air. Atmospheric Environment 14, 549-554. Aristarain, A.J., Delmas, R.J., 2002. Snow chemistry measurements on James Ross Island (Antarctic Peninsula) s howing sea-salt aerosol modifications. Atmospheric Environment 36, 765-772. Bassett, M.E., Seinfeld, J.H., 1983. Atmos pheric equilibrium model of sulfate and nitrate aerosols. Atmospher ic Environment 17, 2237-2252. Bassett, M.E., Seinfeld, J.H., 1984. Atmos pheric equilibrium model of sulfate and nitrate aerosols II. Particle size ana lysis. Atmospheric Environment 18, 1163-1170. Batterman, S., Osak, I., Gelman, C., 1997. SO2 sorption characteristics of air sampling filter media using a new laboratory test. Atmospheric Environment 31, 1041-1047. Beichert, P., Finlayson-Pitts, B.J., 1996. Knudsen cell studies of the uptake of gaseous HNO3 and other oxides of nitrogen on solid NaCl: The role of surface-adsorbed water. Journal of Physical Chemis try 100, 15218-15228. Berresheim, H., Wine, P.H ., Davis, D.D., 1995. Sulfur in the atmosphere. In: Singh, H.B. (Ed.), Compositi on, Chemistry and Climate of the Atmosphere. Van Nostrand Reinhol d, New York, NY, pp. 251-307. Bhethanabotla, V.R., 2002. NOAA Buoy Williams Model, Web address: http://www.eng.usf.edu/~bhethana.

PAGE 201

179 Binkowski, F.S., Shankar, U., 1995. The regional particulate matter model 1. Model description and preliminary re sults. Journal of Geophysical Research 100, 26191-26209. Blanchard, C.L., 1999. Methods for attribut ing ambient air pollutants to emission sources. Annual Review of Ener gy and the Environment 24, 329-365. Blanchard, C.L., Roth, P.M., Tanenbaum, S.J ., Ziman, S.D., Seinfeld, J.H., 2000. The use of ambient measurements to identify which precursor species limit aerosol nitrate formation. J ournal of the Air & Waste Management Association 50, 2073-2084. Brimblecombe, P., Clegg, S. L., 1988. The solubility and behaviour of acid gases in the marine aerosol. Journal of Atmospheric Chemistry 7, 1-18. Bromley, L.A., 1973. Thermody namic properties of strong electrolytes in aqueous solutions. American Institute of Chem ical Engineers Journal 19, 313-320. Burian, S.J., Streit, G.E., McPherson, T.N., Brown, M.J., Turin, H.J., 2001. Modeling the atmospheric deposition an d stormwater washoff of nitrogen compounds. Environmental M odelling and Software 16, 467-479. Caffrey, P.F., Ondov, J.M., Zufall, M.J., Davidson, C.I., 1998. Determination of size-dependent dry particle deposition velocities with multiple intrinsic elemental tracers. Environmental Science & Tec hnology 32, 1615-1622. Caloz, F., Fenter, F.F., Tabor, K.D ., Rossi, M.J., 1997. Paper I: Design and construction of a Knudsen-cell r eactor for the study of heterogeneous reactions over the temperature range 130-750 K: Performances and limitations. Review of Scient ific Instrum ents 68, 3172-3179. Calvert, J.G., Stockwell, W.R., 1983. Ac id generation in the troposphere by gasphase chemistry. Environmental Sc ience & Technolog y 17, A428-A443. Campbell, S.W., Evans, M. C., Poor, N.D., 2002. Predi ctions of size-resolved aerosol concentrations of ammonium, ch loride and nitrate at a bayside site using EQUISOLV II. Atmospheric Environment 36, 4299-4307. Chameides, W.L., Stelson, A.W., 1992. Aqueous-phase chemic al processes in deliquescent sea-salt aerosols: A mechanism that couples the atmospheric cycles of S and sea sal t. Journal of Geophysical Research 97, 20565-20580. Chang, J.-S., Urashima, K., Tong, Y.X., Li u, W.P., Wei, H.Y ., Yang, F.M., Liu, X.J., 2003. Simultaneous removal of NOx and SO2 from coal boiler flue gases by DC corona discharge ammoni a radical shower systems: pilot plant tests. Journal of Electrostatics 57, 313-323.

PAGE 202

180 Cheng, Z.L., Lam, K.S., Chan, L.Y., Wa ng, T., Cheng, K.K. 2000. Chemical characteristics of aerosols at coasta l station in Hong Kong. I. Seasonal variation of major ions, halogens and mineral dusts between 1995 and 1996. Atmospheric Envi ronment 34, 2771-2783. Clarke, A.G., Azadi-Boogar, G.A., Andr ews, G.E., 1999. Pa rticle size and chemical composition of urban aerosols. The Science of the Total Environment 235, 15-24. Clegg, N.A., Toumi, R., 1998. Non-seasalt-sulphate formation in sea-salt aerosols. Journal of Geophysi cal Research 103, 31095-31102. Clegg, S.L., Brimblecombe, P., Wexler, A.S., 1998. T hermodynamic model of the system H+-NH4 +-Na+-SO4 2--NO3 --Cl--H20 at 298.15 K. Journal of Physical Chemistry A 102, 2155-2171. Conley, D.J., 2000. Biogeochemical nutrient cycles and nutrient management strategies. Hydrobiologia 410, 87-96. Cooper, D.W., Spielman, L.A., 1976. Data inversion using nonlinear programming with physical constrai nts: Aerosol size distribution measurement by impactors. Atmo spheric Environment 10, 723-730. Dasch, J.M., Cadle, S.H., Kennedy, K.G. Mulawa, P.A., 1989. Comparison of annular denuders and filter packs for at mospheric sampling. Atmospheric Environment 23, 2775-2782. Davies, J.A., Cox, R.A., 1998. Kinet ics of the heterogeneous reaction of HNO3 with NaCl: Effect of water vapor. Journal of Physical Chemistry A 102, 7631-7642. De Haan, D.O., Finlayson-Pitts, B.J., 1997. Knudsen cell studies of the reaction of gaseous nitric acid with synthetic sea salt at 298 K. Journal of Physical Chemistry A 101, 9993-9999. de Leeuw, G., 1986. Vertical profiles of giant particles close above the sea surface. Tellus 38B, 51-61. de Leeuw, G., Cohen, L., Frohn, L.M., Geernaert, G., Hertel, O., Jensen, B., Jickells, T., Klein, L., Kunz, G.J ., Lund, S., 2001. Atmospheric input of nitrogen into the North Sea: ANICE pr oject overview. Continental Shelf Research 21, 2073-2094. de Wit, M., Bendoricchio, G., 2001. Nutrient fluxes in the Po basin. The Science of the Total Environment 273, 147-161.

PAGE 203

181 Dentener, F.J., Carmichael, G.R., Zhang, Y., Crutzen, P.J., 1996. Role of mineral aerosol as a reactive surface in the global tropospher e. Journal of Geophysical Research 101, 22869-22889. Dentener, F.J., Crutzen, P.J., 1994. A three-dimensional m odel of the global ammonia cycle. Atmospheric Environment 19, 331-369. Draxler, R.R., Hess, G. D., 1998. An overview of the HYSPLIT_4 modelling system for trajectories, disper sion and deposition. Australian Meteorological M agazine 47, 295-308. Ebert, M., Weinbruch, S., Hoffmann, P., Ortner, H.M., 2000. Chemical characterization of North Sea aerosol particles. Journal of Aerosol Science 31, 613-632. Erickson III, D.J., Seuzaret, C., Keene, W.C., Gong, S.L., 1999. A general circulation model based calc ulation of HCl and ClNO2 production from sea salt dechlorination: Reactive chlori ne emissions inventory. Journal of Geophysical Research 104, 8347-8372. Erisman, J.W., Brydges, T., Bull, K., Co wling, E., Grennfelt, P., Nordberg, L., Satake, K., Schneider, T., Smeul ders, S., Van der Hoek, K.W., Wisniewski, J.R., Wisniewski, J., 1998. Nitrogen, the Confer-N-s, First International Nitrogen Conferenc e 1998: Summary Statement. Evans, M.C., Hendrix-Holmes, H., Poor, N.D., 2002. Do macroparticles contain nitrogen? Proceedings of the 95th Annual Air and Waste Management Conference and Exhibition, June 23-27, Baltimore, MD. A&WMA, Pittsburgh, PA. Evans, M.C., Poor, N.D., 2001. Scrubbing of atmospheric nitric acid and sulfuric acid by marine air. Proceedings of the 94th Annual Air and Waste Management Conference and Exhibiti on, June 24-28, Orlando, FL. A&WMA, Pittsburgh, PA. Febo, A., De Sanits, F., Perrino, C., Gi usto, M., 1989. Evaluation of laboratory and field performance of denuder tubes: a theoretical approach. Atmospheric Environment 23, 1517-1530. Fenter, F.F., Caloz, F., Rossi, M.J., 1994. Ki netics of nitric acid uptake by salt. Journal of Physical Chemistry 98, 9801-9810. Fenter, F.F., Caloz, F., Rossi, M.J., 1997. Paper II: Simulation of flow conditions in low-pressure flow reactors (Knudsen cells) using a Monte Carlo technique. Review of Scient ific Instrum ents 68, 3180-3186.

PAGE 204

182 Finlayson-Pitts, B.J., Hemminger, J.C., 200 0. Physical chemistry of airborne sea salt particles and their components. J ournal of Physical Chemistry A 104, 11463-11477. Finlayson-Pitts, B.J., Pitts Jr., J.N ., 2000. Chemistry of the Upper and Lower Atmosphere: Theory, Experiments, and Applications. Academic Press, New York, NY, pp. 969. Fitzgerald, J.W., 1991. Marine aerosols: A review. Atmospheric Environment 25A, 533-545. Forrest, J., Spandau, D.J., Tanner, R.L ., Newman, L., 1982. Determination of atmospheric nitrate and nitric acid employing a diffusion denuder with a filter pack. Atmospheric Environment 16, 1473-1584. Forrest, J., Tanner, R.L., Spandau, D.J. D'Ottavio, T., Newman, L., 1980. Determination of total inor ganic nitrate utilizing collec tion of nitric acid on NaCl-impregnated filters. Atmos pheric Environment 14, 137-144. Ghosal, S., Hemminger, J.C., 1999. Effect of water on the HNO3 pressure dependence of the reaction between gas-phase HNO3 and NaCl surfaces. Journal of Physical C hemistry A 103, 4777-4781. Gilliland, A.B., Pierce, T.E., Dennis, R.L., Roselle, S., Inverse Modeling to Estimate Seasonal Ammonia Emissions. USEPA Atmospheric Modeling Division, Research Triangle Park, NC, pp. Glover, T., Mitchell, K., 2002. An Introduc tion to Biostatistics. McGraw Hill, Boston, MA, pp. 416. Goodman, A.L., Underwood, G.M., Grassian, V.H., 2000. A laboratory study of the heterogeneous reaction of nitric ac id on calcium carbonate particles. Journal of Geophysical Research 105, 29053-29064. GraphPad Software, 2000. GraphP ad Software, Web address: http://www.graphpad.com. Grassian, V.H., 2002. Chemic al reactions of nitrogen ox ides on the surface of oxide, carbonate, soot, and mineral dus t particles: Implications for the chemical balance of the troposphere. Journal of Physical Chemistry A 106, 860-877. Grosjean, D., 1982. The stabi lity of particulate nitrate in the Los Angeles atmosphere. The Science of t he Total Environment 25, 263-275. Guimbaud, C., Arens, F., Gutzwiller, L ., Gaggeler, H.W., A mmann, M., 2002a. Uptake of HNO3 to deliquescent sea-salt par ticles. Atmospheric Chemistry and Physics Discussions 2, 739-763.

PAGE 205

183 Guimbaud, C., Arens, F., Gutzwiller, L ., Gaggeler, H.W., A mmann, M., 2002b. Uptake of HNO3 to deliquescent sea-salt particles: a study using the shortlived radioactive isotope tracer 13N. Atmospheric Chemistry and Physics Discussions 2, 249-257. Haaf, W., Jaenicke, R., 1980. Result s of improved size distribution measurements in the Aitken range of atmospheric aerosols. Journal of Aerosol Science 11, 321-330. Hanisch, F., Crowley, J.N., 2001. Heterogen eous reactivity of gaseous nitric acid on Al2O3, CaCO3, and atmospheric dust samples: A Knudsen cell study. Journal of Physical C hemistry A 105, 3096-3106. Harrison, R.M., 1993. The Chemistry and Deposition of Particulate Nitrogencontaining Species. In: Cocks, A.T. (Ed.), The Chemistry and Deposition of Nitrogen Species in the Troposphere. The Royal Society of Chemistry, Cambridge, pp. 95-105. Harrison, R.M., Kitto, A.-M.N., 1990. Fi eld inter-comparison of filter pack and denuder sampling methods for reac tive gaseous and particulate pollutants. Atmospheric Environment 24A, 2633-2640. Hering, S.V., Lawson, D.R., Allegrini, I., Febo, A., Pe rrino, C., Possanzini, M., Sickles II, J.E., Anlauf, K.G., Wiebe, A ., Appel, B.R., John, W., Ondo, J.L., Wall, S.M., Braman, R.S., Sutton, R., Cass, G.R., Solomon, P.A., Eatough, D.J., Eatough, N.L., Ellis, E.C., Grosjean, D., Hicks, B.B., Womack, J.D., Horrocks, J., Knapp, K.T., Ellestad, T.G., Paur, R.J., Mitchell, W.J., Pleasant, M., Peake, E., MacLean, A., Pierson, W.R., Brachaczek, W., Schiff, H.I., Mackay, G.I., Spicer, C.W ., Stedman, D.H., Winer, A.M., Biermann, H.W., Tuaz on, E.C., 1988. The nitric acid shootout: Field comparison of m easurement methods. Atmospheric Environment 22, 1519-1539. Hinds, W.C., 1999. Aerosol Technol ogy: Properties, Behavior, and Measurements of Airborne Particles. John Wiley & Sons, Inc., New York, NY, pp. 483. Hoppel, W.A., Fitzgerald, J.W., Frick, G.M., Larson, R.E., Mack, E.J., 1989. Atmospheric Aerosol Size Distributi ons and Optical Properties in the Marine Boundary Layer over the At lantic Ocean. NRL Report 9188, Washington, DC. Hu, J.H., Abbatt, J.P.D., 1997. Reaction probabilities for N2O5 hydrolysis on sulfuric acid and ammonium sulfate aer osols at room tem perature. Journal of Physical Chemistry A 101, 871-878.

PAGE 206

184 Huebert, B.J., Lazrus, A.L., 1980. Troposphe ric gas-phase and particulate nitrate measurements. Journal of G eophysical Research 85, 7322-7328. HYSPLIT4, 1997. HYSPLIT4 (HYbrid Si ngle-Particle Lagrangian Integrated Trajectory) Model, NOAA Air Resource s Laboratory, Silver Spring, MD. Web address: http://www.arl. noaa.gov/ready/hysplit4.html. Jacobson, M., 1999a. Studying the effect s of calcium and magnesium on sizedistributed nitrate and ammonium with EQUISOLV II. Atmospheric Environment 33, 3635-3649. Jacobson, M.Z., 1997. Development an d application of a new air pollution modeling system Part III. Aeroso l-phase simulations. Atmospheric Environment 31, 587-608. Jacobson, M.Z., 1999b. Fundamentals of Atmospheric Modeling. Cambridge University Press, New York, NY, pp. 656. Jacobson, M.Z., Tabazadeh, A., Turco, R. P., 1996. Simulating equilibrium within aerosols and nonequilibr ium between gases and aerosols. Journal of Geophysical Research 101, 9079-9091. Jaenicke, R., 1993. Tropospheric aerosols. In : Hobbs, P.V. (Ed.), Aerosol-CloudClimate Interactions. Academic Pr ess, San Diego, CA, pp. 1-31. Jordan, C.E., Talbot, R.W., Keim, B.D., 2000. Water-soluble nitrogen at the New Hampshire sea coast: HNO3, aerosols, precipitati on, and fog. Journal of Geophysical Research 105, 26403-26431. Karakas, D., Tuncel, S.G., 1997. Optimiza tion and field applicati on of a filter pack system for the simultaneous sa mpling of atmospheric HNO3, NH3 and SO2. Atmospheric Envi ronment 31, 1657-1666. Kerminen, V.-M., Wexler, A.S., 1995. Gr owth laws for atmospheric aerosol particles: An examination of the bi modality of the accumulation mode. Atmospheric Environment 29, 3263-3275. Kim, J.C., Allen, E.R., 1997. Effects of filter pack sampling conditions on observed ambient concentrations of dry acid deposition species. Chemosphere 34, 587-610. Kim, Y.P., Seinfeld, J.H., 1995. Atmo spheric gas-aerosol equilibrium: III. Thermodynamics of crustal elements Ca2+, K+, and Mg2+. Aerosol Science and Technology 22, 93-110. Kim, Y.P., Seinfeld, J.H., Saxena P., 1993a. Atmospheric gas-aerosol equilibrium I. Thermodynamic model Aerosol Science and Technology 19, 157-181.

PAGE 207

185 Kim, Y.P., Seinfeld, J.H., Saxena P., 1993b. Atmospheric gas-aerosol equilibrium II. Analysis of common approxim ations and activity coefficient calculation methods. Aerosol Sc ience and Technology 19, 182-198. Kitto, A.-M.N., Colbeck, I., 1999. Filt ration and Denuder Sa mpling Techniques. In: Spurny, K.R. (Ed.), Analytical C hemistry of Aerosols. Lewis Publishers, Boca Raton, FL, pp. 103-132. Kolb, C.E., Worsnop, D.R., Zahniser, M. S., Davidovits, P., Keyser, L.F., Leu, M.T., Molina, L.T., Hanson, D.R., Ravishankara, A.R. 1995. Laboratory studies of atmospheric heterogeneous chemistry. In : Barker, J.R. (Ed.), Progress and Problems in Atmospheric Chemistry. World Scientific, New Jersey, pp. 771-875. Langer, S., Pemberton, R.S., FinlaysonPitts, B.J., 1997. Diffuse reflectance infrared studies on the reaction of syn thetic sea salt mixtures with NO2: A key role for hydrates in the kinet ics and mechanism. Journal of Physical Chemistry A 101, 1277-1286. Laskin, A., Iedema, M.J., Cowin, J. P., 2002. Quantitative time-resolved monitoring of nitrate formation in s ea salt particles using a CCSEM/EDX single particle analysis. Environm ental Science & Technology 36, 49484955. Laux, J.M., Hemminger, J.C., FinlaysonPitts, B.J., 1994. X-ray photoelectron spectroscopic studies of the heterogeneous reaction of gaseous nitric acid with sodium chloride: Kinetics and c ontribution to the chemistry of the marine troposphere. Geophysical Research Letters 21, 1623-1626. Lide, D.R., 1991. Handbook of Chemistry and Physics. CRC Press, Boca Raton, FL, pp. Li-Jones, X., Savoie, D.L., Prospero, J.M., 2001. HNO3 losses within the cyclone inlet of a diffusion-denuder system under simulated marine environments. Atmospheric Environment 35, 985-993. Lovett, R.F., 1978. Quantitativ e measurement of airborne sea-salt in the North Atlantic. Tellus 30, 358-364. Luo, Y., Yang, X., Carley, R.J., Perk ins, C., 2003. Effects of geographical location and land use on atmospheric depo sition of nitrogen in the State of Connecticut. Environmental Pollution 124, 437-448. Mamane, Y., Mehler, M., 1987. On the natur e of nitrate particles in a coastal urban area. Atmospheric En vironment 21, 1989-1994.

PAGE 208

186 Markowski, G.R., 1987. Improving Twomey's algorithm for inversion of aerosol measurement data. Aerosol Sc ience and Technology 7, 127-141. Marple, V.A., Rubow, K.L., Behm, S.M. 1991. A microorifice uniform deposit impactor (MOUDI): description, calibra tion and use. Aerosol Science and Technology 14, 434-446. Masia, P., Di Palo, V., Possanzini, M., 1994. Uptake of ammoni a by nylon filters in filter pack systems. Atmos pheric Environment 28, 365-366. McCulloch, R.B., Shendrikar, A.D., 2000 Concurrent atmospheric ammonia measurements using citric-aci d-coated diffusion denuders and a chemiluminescence analyzer. Atmo spheric Environment 34, 4957-4958. McMurry, P.H., 2000. A review of atmospheric aerosol measurements. Atmospheric Environment 34, 1959-1999. Mehlmann, A., Warneck, P., 1995. Atmospheric gaseous HNO3, particulate nitrate, and aerosol size distributions of major ionic species at a rural site in western Germany. Atmospher ic Environment 29, 2359-2373. Meng, Z., Seinfeld, J.H., Saxena, P., Ki m, Y.P., 1995. Atmospheric gas-aerosol equilibrium: IV. Thermodynamics of carbonates. Aerosol Science and Technology 23, 131-154. Meszaros, A., Vissy, K., 1974. Concentra tion, size distribution and chemical nature of atmospheric aeros ol particles in remote ocean areas. Journal of Aerosol Science 5, 101-109. Moya, M., Pandis, S.N., Jacobson, M.Z., 2 002. Is the size di stribution of urban aerosols determined by thermodynamic equilibrium? An application to Southern California. Atmospher ic Environment 36, 2349-2365. Murphy, D.M., Anderson, J.R., Quinn, P.K., McInnes, L.M. Brechtel, F.J., Kreidenweis, S.M., Middlebrook, A. M., Posfai, M., Thompson, D.S., Buseck, P.R., 1998. Influence of sea-salt on aerosol radiative properties in the Southern Ocean marine boundary layer. Nature 392, 62-65. Nenes, A., Pilinis, C., Pandis, S.N., 1998. ISORROPIA: a new thermodynamic model for inorganic multicomponent atmospheric aerosols. Aquatic Geochemistry 4, 123-152. Nixon, S.W., Ammerman, J.W., Atkinson, L.P., Berounsky, V.M., Billen, G., Boicourt, W.C., Boynton, W.R., Church T.M., Ditoro, D.M., Elmgren, R., Garber, J.H., Giblin, A.E., Jahnke, R.A., Owens, N.J.P., Pilson, M.E.Q., Seitzinger, S.P., 1996. The fate of nitrogen and phosphorus at the land sea margin of the North Atlant ic Ocean. Biogeochemistry 35, 141-180.

PAGE 209

187 NOAA, 2003a. NOAA National Climat ic Data Center, Web address: http://ncdc.noaa.gov. NOAA, 2003b. NOAA National Ocean Service, Center for Operational Oceanographic Products and Services Web address: http://www.coops.nos.noaa.gov. Nodop, K., Hanssen, J.E., 1986. Field in tercomparison of samp ling and analytical methods for SO2 and SO4 in ambient air. In: A ngeletti, G., Restelli, G. (Ed.), Physico-chemical Behaviour of Atmospheric Pollut ants. D. Reidel Publishing Company, Dordrecht, Holland, pp. O'Dowd, C.D., Smith, M.H., Consterdi ne, I.E., Lowe, J.A., 1997. Marine aerosol, sea-salt, and the marine sulphur cycl e: A short review. Atmospheric Environment 31, 73-80. Okita, T., Morimoto, S., Izawa, S., Konno, W., 1976. Measurement of gaseous and particulate nitrates in the atmo sphere. Atmospheric Environment 10, 1085-1089. Ooki, A., Uematsu, M., Mi ura, K., Nakae, S., 2002. Sources of sodium in atmospheric fine particles. Atmo spheric Environment 36, 4367-4374. O'Shaughnessy, P.T., Raabe, O.G., 2003. A comparison of cascade impactor data reduction methods. Aerosol Sc ience and Technology 37, 187-200. Ott, R.L., 1993. An Introduction to Stat istical Methods and Data Analysis. Duxbury Press, Belmont, CA, pp. 1051. Paerl, H.W., 2001. Atmos pheric Deposition of Nitrogen in Coastal Waters: Biogeochemical and Ecologic al Implications. In: Va ligura, R.A., Alexander, R.B., Castro, M.S., Meyers, T.P., Paer l, H.W., Stacey, P.E., Turner, R.E. (Ed.), Coastal and Est uarine Studies. Americ an Geophysical Union, Washington, D.C., pp. 11-52. Pagano, M., Gauvreau, K., 2000. Principles of Biostatistics. Duxbury, Pacific Grove, CA, pp. Pakkanen, T.A., 1996. Study of formati on of coarse particle nitrate aerosol. Atmospheric Environment 30, 2475-2482. Pakkanen, T.A., Kerminen, V.-M., Hilla mo, R.E., Makinen, M., Makela, T., Virkkula, A., 1996a. Distribution of nitr ate over sea-salt and soil derived particles Implications from a fi eld study. Journal of Atmospheric Chemistry 24, 189-205. Pakkanen, T.A., Kerminen, V.-M., Hilla mo, R.E., Makinen, M., Makela, T., Virkkula, A., 1996b. Erratum: Distributi on of nitrate over sea-salt and soil

PAGE 210

188 derived particles implications from a field study. Journal of Atmospheric Chemistry 25, 227. Patwardhan, A.S., Donigian Jr., A.S., 1995. Assessment of Nitrogen Loads to Aquatic Systems. USEP A, Prepared by AQUA TERRA Consultants, Mountain View, CA. Perrino, C., De Sanits, F., Febo, A., 1990. Criteria for the choice of a denuder sampling technique devoted to the m easurement of atmospheric nitrous and nitric acids. Atmospher ic Environment 24A, 617-626. Perrino, C., De Sanits, F., Febo, A., 1988. Uptake of nitrous acid and nitrogen oxides by nylon surfaces: Implicat ions for nitric acid measurement. Atmospheric Environment 22, 1925-1930. Perrino, C., Gherardi, M., 1999. Optimi zation of the coating layer for the measurement of ammonia by di ffusion denuders. Atmospheric Environment 33, 4579-4587. Perrino, C., Ramirez, D., Allegrini, I., 2001. Monitoring acidic air pollutants near Rome by means of diffusion lines: development of a specific quality control procedure. Atmospher ic Environment 35, 331-341. Pilinis, C., Seinfeld, J. H., 1987. Continued developm ent of a general equilibrium model for inorganic multicomponent at mospheric aerosols. Atmospheric Environment 21, 2453-2466. Pilinis, C., Seinfeld, J.H., Grosjean, D., 1989. Water content of atmospheric aerosols. Atmospheric Environment 23, 1601-1606. Pio, C.A., Harrison, R.M., 1987. The equi librium of ammonium chloride aerosol with gaseous hydrochloric ac id and ammonia under tropospheric conditions. Atmospheric Environment 21, 1243-1246. Plate, E., Schulz, M., 1997. Coarse nitr ate formation in a coastal area of the North Sea. Journal of Aero sol Science 28, S333-S334. Platt, U., Perner, D., Harris, G.W., Winer, A.M., Pitts, J.N., 1980. Observations of nitrous acid in an urban atmosphere by differential optical absorption. Nature 285, 312-314. Poor, N., Pribble, R., Greening, H., 2001. Direct wet and dry deposition of ammonia, nitric acid, ammonium and nitrate to the Tampa Bay Estuary, FL, USA. Atmospheric Environment 35, 3947-3955. Poor, N.D., Clark, T., Nye, L., Tamanini, T., Tate, K., Stevens, R.K., Atkeson, T., 2002. Field performance of dichotom ous sequential PM air samplers. Atmospheric Environment 36, 3289-3298.

PAGE 211

189 Possanzini, M., Febo, A., Liberti, A ., 1983. New design of high-performance denuder for sampling of atmospheric po llutants. Atmospher ic Environment 17, 2605-2610. Pryor, S.C., Barthelmie, R.J., 2000a. Part icle dry deposition to water surfaces: Processes and consequences. Marine Pollution Bulletin 41, 220-231. Pryor, S.C., Barthelmie, R.J., 2000b. Part icle dry deposition to water surfaces: processes and consequences. In: Sheppar d, C. (Ed.), Seas at the Millennium: An Environmental Evaluat ion. Elsevier Science Ltd., New York, NY, pp. 197-209. Pryor, S.C., Sorensen, L.L., 2000. Nitric acid-sea salt reactions: Implications for nitrogen deposition to water surfaces. Journal of Applied Meteorology 39, 725-731. Puttock, J.S., 1981. Data inversion for ca scade impactors: Fitting sums of lognormal distributions. Atmospher ic Environment 15, 1709-1716. Ramachandran, G., Johnson, E.W., Vincen t, J.H., 1996. Inversion techniques for personal cascade impactor data. Journal of Aerosol Science 27, 10831097. Ramachandran, G., Kandlikar, M., 1996. Bayesian analysis for inversion of aerosol size distribution data. Jour nal of Aerosol Science 27, 1099-1112. Rood, M.J., Shaw, M.A., Larson, R.V., Co vert, D.S., 1989. Ubiquitous nature of ambient metastable aerosol. Nature 337, 537-539. Roth, B., Okada, K., 1998. On the modificati on of sea-salt particles in the coastal atmosphere. Atmospheric Environment 32, 1555-1569. Russell, A.G., Cass, G.R., 1984. Acquis ition of regional air quality model validation data for nitrate, sulfate, ammonium i on and their precursors. Atmospheric Environment 18, 1815-1827. Saxena, P., Hudischewskyj A.B., Seigneur, C., Seinfeld, J.H., 1986. A comparative study of equilibrium approaches to the chemical characterization of secondary aerosol s. Atmospheric Environment 20, 1471-1483. Saxena, P., Mueller, P.K ., Hildemann, L.M., 1994. S ources and chemistry of chloride in the troposphere: a review. In: Chow, W., Connor, K.K. (Ed.), Managing Hazardous Air Poll utants: State of the Art. CRC Press, Boca Raton, FL, pp. 173-190.

PAGE 212

190 Seinfeld, J.H., Pandis, S.N., 1998. Atmo spheric Chemistry and Physics: From Air Pollution to Climate Change. John W iley & Sons, Inc., New York, NY, pp. 1326. Sickles II, J.E., Hodson, L.L., 1999. S hort communication retention of sulfur dioxide by nylon filters. Atmo spheric Environment 33, 2423-2426. Sievering, H., Gorman, E., Ley, T., Pszenny, A.A.P., Sp ring-Young, M., Boatman, J.F., Kim, Y.P., Nagamo to, C., Wellman, D.L., 1995. Ozone oxidation of sulfur in sea-salt aerosol particles during the Azores marine aerosol and gas exchange experiment. Journal of Geophysical Research 100, 2307523081. Spicer, C.W., 1986. Patterns of atmos pheric nitrates, sulfate, and hydrogen chloride in the central Ohio Rive r valley over a one-year period. Environmental International 12, 513-518. Spicer, C.W., Howes Jr., J.E., Bishop, T.A., Arnold, L.H., Stevens, R.K., 1982. Nitric acid measurement methods: An intercomparison. Atmospheric Environment 16, 1487-1500. Spurny, K.R., 1999. New Concepts for Samp ling, Measurement, and Analysis of Atmospheric Anthropogenic Aerosols. In: Spurny, K.R. (Ed.), Analytical Chemistry of Aerosols. Lewis Publishers, Boca Raton, FL, pp. 49-99. Stelson, A.W., Friedlander, S.K., Seinfeld J.H., 1979. A note on the equilibrium relationship between ammonia and nitric acid and particulate ammonium nitrate. Atmospheric Environment 13, 369-371. Stelson, A.W., Seinfeld, J.H., 1982. Relative humidity and temperature dependence of the ammonium nitrate dissociation constant. Atmospheric Environment 16, 983-992. Tabazadeh, A., Jacobson, M. Z., Singh, H.B., Toon, O.B. Lin, J.S., Chatfield, R.B., Thakur, A.N., Talbot, R.W., Di bb, J.E., 1998. Nitric acid scavenging by mineral and biomass burning aerosols. Geophysical Research Letters 25, 4185-4188. Tampa Bay National Est uary Program, 1996. Charting the Course for Tampa Bay. St. Petersburg, FL, pp. 263. Tang, I.N., 1980. Deliquescence prope rties and particle size change of hygroscopic aerosols. In: Willeke, K. (Ed.), Generation of Aerosols. Ann Arbor Publications, Ann Arbor, MI, pp. 153-170. Ten Brink, H.M., 1998. Reactive uptake of HNO3 and H2SO4 in sea-salt (NaCl) particles. Journal of Aerosol Science 29, 57-64.

PAGE 213

191 Ten Brink, H.M., Spoelstra, H., 1998. T he dark decay of HONO in environmental (SMOG) chambers. Atmospher ic Environment 32, 247-251. Torseth, K., Semb, A., Schaug, J., Hanss en, J.E., Aamlid, D., 2000. Processes affecting deposition of oxidised nitr ogen and associated species in the coastal areas of Norway. Atmo spheric Environment 34, 207-217. Tsai, C.-J., Perng, S.-B., Chiou, S.-F., 2000. Use of two different acidic aerosol samplers to measure acidic aerosols in Hsinchu, Taiwan. Journal of the Air & Waste Management Association 50, 2120-2128. Valigura, R.A., 2001. An Introduction to the First Assessment of Nitrogen Loads to US Estuaries with an Atmospheric Perspective In: Valigura, R.A., Alexander, R.B., Castro, M.S., Meyers T.P., Paerl, H.W., Stacey, P.E., Turner, R.E. (Ed.), C oastal and Estuarine Studi es. American Geophysical Union, Washington, D.C., pp. 1-10. Vogt, R., Elliott, C ., Allen, H.C., Laux, J.M., He mminger, J.C., Finlayson-Pitts, B.J., 1996. Some new laboratory appr oaches to studying tropospheric heterogeneous reactions. Atmospher ic Environment 30, 1729-1737. Vogt, R., Finlayson-Pitts, B.J., 1994. A diffuse reflectance infrared Fourier transform spectroscopic (DRIFTS) study of the surface reaction of NaCl with gaseous NO2 and HNO3. Journal of Physica l Chemistry 98, 37473755. Vossler, T.L., Stevens, R.K., Paur, R.J ., Baumgardner, R.E., Bell, J.P., 1988. Evaluation of improved inlets an d annular denuder systems to measure inorganic air pollutants. Atmos pheric Environment 22, 1729-1736. Wall, S.M., John, W., Ondo, J.L., 19 88. Measurement of aerosol size distributions for nitrate and major ionic species. Atmospheric Environment 22, 1649-1656. Watson, J.G., Chow, J.C., Lurmann, F. W., Musarra, S.P., 1994. Ammonium nitrate, nitric acid, and ammonia eq uilibrium in wintertime Phoenix, Arizona. Journal of the Air & Wast e Management Associ ation 44, 405-412. Wayne, R.P., 2000. Chemistry of Atmos pheres. Oxford University Press, New York, NY, pp. Weis, D.D., Ewing, G.E., 1999. The reacti on of nitrogen dioxide with sea salt aerosol. Journal of Physical Chemistry A 103, 4865-4873. Wexler, A.S., Seinfeld, J.H., 1991. Seco nd-generation inorganic aerosol model. Atmospheric Environment 25A, 2731-2748.

PAGE 214

192 Williams, R.M., 1982. A model for the dry deposition of particles to natural water surfaces. Atmospheric Environment 16, 1933-1938. Yeatman, S.G., Spokes, L.J., Jickells, T.D., 2001. Comparis ons of coarse-mode aerosol nitrate and ammonium at two polluted coastal sites. Atmospheric Environment 35, 1321-1335. Zangmeister, C.D., Pemberton, J.E., 1998. In situ moni toring of the NaCl + HNO3 surface reaction: The observation of mobile surface strings. Journal of Physical Chemistr y B 102, 8950-8953. Zangmeister, C.D., Pembert on, J.E., 2001. Raman spectr oscopy of the reaction of sodium chloride with nitric acid: Sodium nitrate growth and effect of water exposure. Journal of Ph ysical Chemistry A 105, 3788-3795. Zarbock, H.W., Janicki, A.J., Janicki, S.S., 1996. Estimates of Total Nitrogen, Total Phosphorus, and Total Suspended Solids to Tampa Bay, Florida. Technical Appendix: 19921994 total nitrogen loadings to Tampa Bay, Florida. Tampa Bay National Es tuary Program, Prepared by Coastal Environmental, Inc., St. Petersburg, FL. Zhuang, H., Chan, C.K., Fang, M., Wexler A.S., 1999a. Formation of nitrate and non-sea-salt sulfate on coarse parti cles. Atmospheric Environment 33, 4223-4233. Zhuang, H., Chan, C.K., Fang, M., Wexler, A.S., 1999b. Size distributions of particulate sulfate, nitr ate, and ammonium at a c oastal site in Hong Kong. Atmospheric Environment 33, 843-853.

PAGE 215

193 Appendices

PAGE 216

194 Appendix 1: Met eorological Data Temp (C) RH (%) Temp (C) RH (%) Avg Min Max AvgMinMax AvgMinMax Avg MinMax 10/10/01 26 21 30 69 46 84 11/1/0125 21 29 73 51 90 10/11/01 27 22 31 68 42 87 11/2/0125 21 29 84 65 96 10/12/01 26 21 31 68 50 87 11/3/0126 21 31 76 51 94 10/13/01 27 22 31 68 51 84 11/4/0124 21 27 78 62 91 10/14/01 27 23 30 87 72 97 11/5/0120 17 23 73 59 87 10/15/01 26 21 30 71 51 94 11/6/0120 14 26 60 43 78 10/16/01 25 21 29 74 50 87 11/7/0119 13 24 67 50 87 10/17/01 19 14 24 57 44 75 11/8/0119 13 25 73 54 90 10/18/01 22 14 29 65 46 81 11/9/0119 14 24 78 55 100 10/19/01 25 19 31 75 57 91 11/11/0119 13 25 74 45 97 10/20/01 27 23 31 80 57 94 11/12/0121 16 27 65 45 84 10/21/01 27 23 30 88 72 97 11/13/0121 16 27 82 50 96 10/24/01 27 24 30 89 72 97 11/14/0120 18 22 83 68 91 10/26/01 20 15 25 47 32 91 11/15/0120 18 22 79 73 90 10/27/01 16 12 20 42 28 62 11/16/0121 16 27 65 49 81 10/28/01 16 10 21 56 42 77 11/17/0122 15 28 68 53 87 10/30/01 21 16 26 72 58 81 11/18/0123 18 27 76 50 93 10/31/01 22 17 27 72 54 87 11/19/0123 18 27 81 52 93 Table 29. Temperature and relative hum idity data for October November 2001.

PAGE 217

195 Appendix 1: (Continued) Temp (C) RH (%) Avg Min Max Avg Min Max 5/2/02 27 24 31 76 59 97 5/4/02 28 23 32 70 52 91 5/6/02 28 24 33 68 41 90 5/10/02 28 24 31 61 44 84 5/14/02 26 22 30 60 48 71 5/15/02 26 22 32 53 39 76 5/16/02 27 24 33 71 40 94 5/17/02 28 25 32 69 46 89 5/19/02 22 18 26 79 63 100 5/20/02 23 18 28 55 45 71 5/23/02 26 20 31 54 35 80 5/24/02 26 20 31 50 32 78 5/25/02 26 21 30 53 37 78 5/31/02 27 23 32 70 47 91 Table 30. Temperature and relative humidity data for May 2002 Azalea Park sampling site. Temp (C) RH (%) Avg Min Max Avg Min Max 5/4/02 26 24 30 76 56 86 5/6/02 28 24 33 67 44 88 5/10/02 28 24 32 61 42 82 5/14/02 25 21 29 63 55 70 5/15/02 27 22 31 50 35 76 5/16/02 26 23 32 74 43 91 5/17/02 27 24 31 71 48 87 5/19/02 21 17 25 81 74 87 5/20/02 22 18 28 58 45 72 5/23/02 25 21 30 54 33 74 5/24/02 25 20 31 52 30 78 5/25/02 26 22 29 50 37 68 5/31/02 26 22 31 73 54 89 Table 31. Temperature and relative humidity data for May 2002 Gandy sampling site.

PAGE 218

196 Appendix 1: (Continued) Temperature (C) Relative Humidity (%) Avg Min Max Avg Min Max 5/6/02 27 22 35 77 44 99 5/10/02 29 21 35 5/14/02 25 19 36 66 53 77 5/15/02 26 20 32 62 39 89 5/16/02 27 23 33 79 51 99 5/17/02 28 22 35 75 38 99 5/19/02 22 16 28 88 70 97 5/20/02 21 16 28 69 53 81 5/23/02 25 19 31 64 39 93 5/24/02 25 17 31 63 35 98 5/25/02 25 18 30 65 41 95 Table 32. Temperature and relative humidity for May 2002 Sydney sampling site.

PAGE 219

197 Appendix 2. Size Distributions and Ion Ratios from May 2002 D p ( m) 001 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 D p ( m) 001 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 56. Size distribu tions for May 10, 2002. Azalea Gandy Sydney Cl-:Na+ 0.79 0.69 0.73 NO3 -:Na+ 0.41 0.55 0.46 Cl--dep % 32% 41% 38% NH4 +:SO4 22.3 2.8 2.7 Table 33. Ion ratios for May 10, 2002.

PAGE 220

198 Appendix 2: (Continued) D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 57. Size distribu tions for May 15, 2002. Azalea Gandy Sydney Cl-:Na+ 0.83 0.81 0.83 NO3 -:Na+ 0.39 0.38 0.40 Cl--dep % 30% 31% 29% NH4 +:SO4 22.0 1.8 2.1 Table 34. Ion ratios for May 15, 2002.

PAGE 221

199 Appendix 2: (Continued) D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 58. Size distribu tions for May 16, 2002. Azalea Gandy Sydney Cl-:Na+ 0.80 0.93 0.94 NO3 -:Na+ 0.49 0.37 0.32 Cl--dep % 32% 21% 20% NH4 +:SO4 21.7 1.9 2.3 Table 35. Ion ratios for May 16, 2002.

PAGE 222

200 Appendix 2: (Continued) D p ( m) 001 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 Dp ( m) 001 0.1 1 10 dC/dLogD (umol m -3 ) 000 002 004 006 008 0.10 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 59. Size distribu tions for May 17, 2002. Azalea Gandy Sydney Cl-:Na+ 0.90 0.89 0.67 NO3 -:Na+ 0.29 0.36 0.50 Cl--dep % 23% 24% 43% NH4 +:SO4 21.7 1.9 2.0 Table 36. Ion ratios for May 17, 2002.

PAGE 223

201 Appendix 2: (Continued) D p ( m) 001 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 Dp ( m) 001 0.1 1 10 dC/dLogD (umol m -3 ) 000 005 0.10 0.15 020 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 60. Size distribu tions for May 19, 2002. Azalea Gandy Sydney Cl-:Na+ 0.77 0.70 0.95 NO3 -:Na+ 0.45 0.43 0.65 Cl--dep % 34% 41% 20% NH4 +:SO4 22.0 1.9 2.3 Table 37. Ion ratios for May 19, 2002.

PAGE 224

202 Appendix 2: (Continued) D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 61. Size distribu tions for May 23, 2002. Azalea Gandy Sydney Cl-:Na+ 0.93 0.97 0.89 NO3 -:Na+ 0.26 0.26 0.23 Cl--dep % 21% 17% 24% NH4 +:SO4 22.0 1.9 1.6 Table 38. Ion ratios for May 23, 2002.

PAGE 225

203 Appendix 2: (Continued) D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 62. Size distribu tions for May 24, 2002. Azalea Gandy Sydney Cl-:Na+ 0.88 0.84 0.84 NO3 -:Na+ 0.28 0.33 0.34 Cl--dep % 25% 28% 29% NH4 +:SO4 22.1 2.3 1.7 Table 39. Ion ratios for May 24, 2002.

PAGE 226

204 Appendix 2: (Continued) D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 63. Size distribu tions for May 25, 2002. Azalea Gandy Sydney Cl-:Na+ 0.77 0.73 0.71 NO3 -:Na+ 0.37 0.37 0.45 Cl--dep % 34% 38% 39% NH4 +:SO4 22.1 2.3 2.1 Table 40. Ion ratios for May 25, 2002.

PAGE 227

205 Appendix 2: (Continued) D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) -0.005 0.000 0.005 0.010 0.015 0.020 0.025 D p ( m) 001 0.1 1 10 dC/dLogD (umol m -3 ) -0.005 0.000 0.005 0.010 0.015 0.020 0.025 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) -0.005 0.000 0.005 0.010 0.015 0.020 0.025 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.05 0.10 0.15 0.20 0.25 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 64. Size distribu tions for May 31, 2002. Azalea Gandy Sydney Cl-:Na+ 0.58 0.58 0.36 NO3 -:Na+ 0.81 0.82 1.32 Cl--dep % 50% 50% 70% NH4 +:SO4 22.0 2.2 2.2 Table 41. Ion ratios for May 31, 2002.

PAGE 228

206 Appendix 2: (Continued) D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 (a) Azalea (b) Gandy (c) Sydney Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 Dp ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 D p ( m) 0.01 0.1 1 10 dC/dLogD (umol m -3 ) 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 NH 4 + SO4 2(d) Azalea (e) Gandy (f) SydneyNa + Cl NO 3 Na + Cl NO 3 Na + Cl NO 3 NH 4 + SO4 2NH 4 + SO4 2Figure 65. Average size distributions for May 2002. Azalea Gandy Sydney Cl-:Na+ 0.78 0.76 0.72 NO3 -:Na+ 0.45 0.45 0.54 Cl--dep % 34% 36% 39% NH4 +:SO4 22.0 2.1 2.1 Table 42. Overall ion ratios for May 2002.

PAGE 229

207 Appendix 3. Density Calc ulation for Aqueous Aerosol Variables defined: Vi molar volume of i Xi mole fraction of i Mi molar mass of i i density of i wi mass fraction of i in aqueous phase Fi mass fraction of i in total particle soln solution s solute (salt) in aqueous phase of the particle w water solid solid or crystalline phase of the particle Aqueous phase: w w s s aqV X V X V w w w s s s aq aqM X M X M w w s s aqM X M X M w w w s s w w s w w s s s s aqX M X M X M X M X M X M 1 1 1

PAGE 230

208 Appendix 3: (Continued) w w s s aqw w 1 w w s s aqw w 1 Overall particle phase: solid aq partV V V solid aq partMass Mass Mass solid solid aq aq solid aq solid aq solid aq partMass Mass Mass Mass V V Mass Mass solid aq aq aqMass Mass Mass F solid solid aq aq partF F 1

PAGE 231

209 Appendix 4. Size Distributions and Data Inversion Cascade impactors are the standard in struments for measuring particle size distributions (Puttock, 1981). They separate particles by size according to their inertial properties (O'Sha ughnessy and Raabe, 2003 ). The size distributions result from either dire ctly assigning the collected mass or concentration to the given size on each st age or indirectly through the reduction of the impactor data (O 'Shaughnessy and Raabe, 2003). The direct application of the data assumes that t he instrument response is ideal, that the deposition step functions of each stage are perfectly sharp (Cooper and Spielman, 1976; Ramachandran et al., 1996). This is the most widely used approach for treating ca scade impactor data (Cooper and Spielman, 1976). The ideal or perfect step functi on efficiency curve assumption says that all particles greater than a certain size are collected on a particular stage and all particles smaller than that size pass through (Hinds, 1999; Puttock, 1981). This allows the mass or concentration on a st age to be directly assigned to a given particle size. Most well defined impactors can be assumed to be ideal, in that they have a sharp cut-off curve that approaches the ideal curve (Cooper and Spielman, 1976; Hinds, 1999). The following displays the collection efficiencies as a function of particle diameter for the MOUDI sampler (Figure 66).

PAGE 232

210 Appendix 4: (Continued) Particle Diameter (um) 0.1110 Collection Efficiency (%) 0 20 40 60 80 100 Stage 2 Stage 1 Stage 3 Stage 4 Stage 5 Stage 6 Stage 7 Stage 8 Stage 9 Stage 11 Stage 10 Figure 66. Collection efficiencies as a function of particle diameter for the MOUDI sampler (adapted from Marple et al., 1991). The cutpoint for each collection st age of a cascade impactor is determined by the 50% collection efficiency point (H inds, 1999). Each collection stage, or particle bin, has a range of particle sizes collected. The r ange is determined by the cutpoint characteristi cs of the adjacent stages. For cascade impactors with non-i deal, broad responses, the measured data must be reduced through a process known as inversion (Markowski, 1987). Data inversion is the “inference of the par ticle size distributi on function from the

PAGE 233

211 Appendix 4: (Continued) measured stage loadings” (Puttock, 1981). Data inversion is used when the collection characteristics of cascade impac tors deviate from ideal and when the impactors provide too few size separatio ns to accurately resolve the complex size distributions occurring in the at mosphere (Puttock, 1981; Ramachandran and Kandlikar, 1996). The mass on each stage loading is related to the kernel function, which is the fraction of particles that enters the instrument and is collected on a particular stage (Ramachandran et al., 1996; Rama chandran and Kandlikar, 1996). Kernel functions are determined for each stage of the instrument and are a function of the collection efficiency curves. For non-ideal response impactors, kernel functions overlap as particles of a gi ven size are collected on more than one stage (Ramachandran and Kandlikar, 1996) The kernel function can be calculated using: a a i a i a i a iD E D E D E D E D K1 2 11 1 1 (Equation 38) where Ei(Da) is the collection efficiency of the ith stage as a function of the particle aerodynamic diameter. The following show s the kernel functions for the eight stages of the MOUDI sampler used in these studies (Figure 67).

PAGE 234

212 Appendix 4: (Continued) Aerodynamic Diameter (um) 0.1110 Kernel Function, Ki 0.0 0.2 0.4 0.6 0.8 1.0 K1K2K3K4K5K6K7K8K9Kfilter Figure 67. Deposition kernel functions for the MOUDI as functions of particle aerodynamic diameter. The kernel functions are not strict ly independent; there is overlap between them (Ramachandran and Kandlikar, 1996) For stage one of the MOUDI sampler, the kernel function shows poor efficiency for particle collection. The inlet of this instrument was designed for high efficiency for particles less than 10 m. The top two stages result in less than perfect collection and need to be taken under consideration.

PAGE 235

213 The data inversion process takes the kernel functions and uses them in algorithms. The results are often of poorer quality than those obtained from a direct application of the data, as the inve rsion problem is ill defined and results in multiple iterations and solutions (Ramachandran and Kandlikar, 1996).

PAGE 236

About the Author Melissa Cheryl Foster Evans was born October 11, 1978 in Michigan. She graduated from Indian Rocks Christ ian High School as co -valedictorian in 1995. In 1999, she earned a Bachelor of Arts degree in Chemistry from the University of South Flor ida, graduating cum laude fr om the University Honors Program. She contin ued her education at that inst itution, where she completed her requirements for a Doct or of Philosophy degree in Physical Chemistry in December 2003.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001441469
003 fts
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 031203s2003 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000142
035
(OCoLC)54018444
9
AJM5909
b SE
SFE0000142
040
FHM
c FHM
090
QD31.2
1 100
Evans, Melissa Cheryl Foster.
0 245
Characterization and formation of particulate nitrate in a coastal area
h [electronic resource] /
by Melissa Cheryl Foster Evans.
260
[Tampa, Fla.] :
University of South Florida,
2003.
502
Thesis (Ph.D.)--University of South Florida, 2003.
504
Includes bibliographical references.
500
Includes vita.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 236 pages.
520
ABSTRACT: Particulate nitrates play important roles in the atmosphere. They consist mainly of NH4NO3 and NaNO3, products from the reactions of gaseous HNO3 with gaseous NH3 and sea salt, respectively. The gas-to-particle phase conversion of nitrate changes its deposition characteristics and ultimately changes the transport and deposition rates of the locally produced species. Studies were conducted to develop background information on the particle concentrations and size distributions and the chemistry and kinetics behind the interactions. The predominant nitrate species in the Tampa Bay area was identified as coarse mode NaNO3. NH4NO3 was not detected as it has high volatility at ambient temperatures. Spatial distribution sampling determined a gradient of NaCl and NaNO3 with increased distance from the coastal shore and an increase in the gas-to-particle conversion of nitric acid along a predominant air mass trajectory. The EQUISOLV II thermodynamic equilibrium model was evaluated. It was determined that the model can be used to predict gas and size-distributed particulate matter concentrations. The model was also used to examine the gas-to-particle partitioning of nitric acid to nitrate by NaCl and CaCO3. Both sodium and calcium partitioned nitrate to the particle phase. The magnitude of the partitioning was directly dependent on the equilibrium coefficients. The fine mode percentage of the total nitrate was determined using two methods. The results were used to expand the current data set to account for the coarse mode nitrate, and they indicated that particle nitrate accounted for 9% of the total nitrogen deposition flux to Tampa Bay. The formation of particle nitrate was examined using a nitrate accumulation model. Results indicated that the equilibrium time for particles less than 10 um in diameter was significantly less than their atmospheric residence time, with fastest conversion occurring under the highest relative humidity conditions. This information is vital in the development of atmospheric nitrogen dry deposition estimates, which are used to assess water quality and nutrient loading. These data can be used to determine air-monitoring strategies and to develop models that account for the coarse particle nitrogen species.
590
Co-adviser: Harmon, Julie P.
Co-adviser: Poor, Noreen D.
653
mineral dust.
modeling.
nitric acid.
sea salt.
tampa bay.
690
Dissertations, Academic
z USF
x Chemistry
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.142