Characterization of Simple Saccharides and Other Organic Compounds in Atmospheric Particulate Matter and Source Apportionment Using Positive Matrix Factorization
Thesis by
Yuling Jia
In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy
APPROVED, THESIS COMMITTEE:
Matthew P. Fraser, Associate Professor, Director, Co-Chair Civil and Environmental Engineering
Robert Griffin, Associate Professor, Chair Civil and Environmental Engineering
Daniel Cohan, Assistant Professor Civil and Environmental Engineering
-4/ Carrie A. Masiello, Assistant Professor Earth Science
Pierre Herckes, Assistant Professor Chemistry & Biochemistry, Arizona State University
Rice University, Houston, TX May, 2010 UMI Number: 3421320
All rights reserved
INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.
Jn the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. UMT Dissertation Publishing
UMI 3421320 Copyright 2010 by ProQuest LLC. All rights reserved. This edition of the work is protected against unauthorized copying under Title 17, United States Code. uest
ProQuest LLC 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 ABSTRACT Characterization of Simple Saccharides and Other Organic Compounds in Atmospheric Particulate Matter and Source Apportionment Using Positive Matrix Factorization
by Yuling Jia
Ambient particulate matter samples were collected at various sites in Texas,
Arizona, and Austria from 2005 to 2009 to characterize the organic compositions and local PM sources. The primary biologically derived carbon sources, specifically the atmospheric entrainment of soil and associated biota and primary biological aerosol particles (PBAPs), are major sources contributing to ambient PM. This dissertation work proposes simple saccharides as well-suited tracers to characterize the contribution to ambient PM from these primary biologically derived carbon sources.
Saccharide concentrations in ambient PM were determined from various locations and various seasons. Aerosol saccharide compounds displayed seasonal variations, inter-correlations, and size fractionations (fine vs. coarse) that were consistent between samples and that can be used to determine sources. The difference in aerosol saccharide concentrations and relative species abundances was reflective of different climate patterns and ecosystems. Selected saccharide compounds including an established marker (levoglucosan) and novel markers (glucose, sucrose, trehalose, mannitol, and arabitol) were used along with other markers to model the major source contributions to ambient PM using a positive matrix factorization (PMF) model.
Major local PM sources were resolved at three Texas sites (San Augustine, Dallas, and Big Bend National Park) and one Arizona site (Higley), with two source factors enriched in the proposed novel saccharide markers that can be related to the primary
biologically derived carbon sources. The contribution to PM from the saccharide-rich
primary biological sources was estimated to range from 16% (remote area) to 36%
(rural and suburban area) at the four sampling sites studied. Other PM sources
identified by PMF included motor vehicles, secondary aerosol formation, meat
cooking, biogenic wax, sea salt, crustal material, and road dust.
To further characterize the primary biologically derived carbon sources,
different soil and source samples representing PBAPs (plants and spores) were
collected at Higley (AZ) to study their saccharide compositions in particle sizes
equivalent to PM2.5 and PM10. It was found that the total measured
non-levoglucosan saccharide content relative to PM mass in ambient aerosols (0.2%
on average in PM2.5 and 0.11% in PM10) was much higher than the soil samples
(<0.02% in both PM2.5 and PM10) but much lower than in the PBAP source samples
(2% on average in plant PBAP samples and 16% in spore PBAP samples). The
measured PBAP samples contained a concentration of sucrose and glucose that is
consistent with the saccharide-rich source profiles resolved from ambient aerosol data
analyzed by PMF while the measured soil samples did not. This can be interpreted as
confirmation that PBAPs are an important PM source in additional to soil and
associate biota at Higley, AZ. However, the saccharide levels in the measured
PBAP samples were several orders of magnitude higher than the PMF results,
suggesting that the ambient aerosol samples are a combination of high saccharide concentration PBAPs and lower saccharide concentration soils at Higley, AZ. Acknowledgements
I would like to express my deepest gratitude to my advisor, Dr. Matthew
Fraser, for his invaluable support, encouragement, supervision, and useful suggestions
throughout the work of this thesis. His far reaching knowledge and expertise has
helped improve the quality of this work and his constant help and support to my
research has made the past four years a rich learning experience for me.
I am sincerely grateful to the following people at Arizona State University for
their help and support with this thesis. First of all, Dr. Pierre Herckes and his research
group (James Hutchings, Nabin Upadhyay, and Hansina Hill), who provided much
support and help to the chemical analysis of various samples in this work. My
colleague Andrea Clements who assisted me with sample collection and analysis and
Dr. Tony Hartshorn who helped with soil taxonomy. Last but not least, Dr. Scott
Bates and Dr. Leslie Landrum who cordially helped me with the identification and
collection of plant and spore samples in this study. This thesis work would not have
been possible without their help.
I would like to thank Dr. Robert Griffin, Dr. Daniel Cohan, and Dr. Carrie
Masiello for serving on my committee, spending time on my thesis and providing
valuable advice and comments. I would also like to forward my gratitude to my
former colleagues at Rice University, Birnur Buzcu-Guven and Shagun Bhat, for their
help with my research.
A special thanks goes to Dr. Hans Puxbaum and his research group (Nicole
Jankowski, Magdalena Rzaca, and Heidi Bauer) at the Vienna University of V
Technology. Cooperation with them has been an inspiring experience to both me and my research.
Finally, I wish to thank those I love. To my husband Eddie, who has always been so supportive of me and my work. Whose love, help, and encouragement have pulled me through all the hardships throughout the years. My parents, who I am forever indebted to, have always trusted and supported me with their unconditional love. Without them none of my achievements, if there are any, would have been possible. Contents
Abstract ii Acknowledgement iv List of Tables xi List of Figures xiii
Chapter 1 Introduction 1 1.1 Motivation 1 1.2 Research Objectives 4 1.3 Hypotheses 5 1.4 Approach 5
Chapter 2 Background and literature review 9 2.1 PM 9 2.1.1 Health effect and regulation 9 2.1.2 Carbonaceous composition and sources 12 2.1.2.1 Elemental carbon 12 2.1.2.2. Organic carbon 13 2.2 Saccharides in aerosols 20 2.2.1 Importance 20 2.2.2 Saccharide species and associated sources 22 2.3 Analysis 26 2.3.1 Organic solvent extraction 26 2.3.2 GC-MS and derivatization 28 2.4 Statistical Tool - Correlation Analysis 30 2.5 Source apportionment 32 2.5.1 Sources and molecular markers 34 vii
2.5.2 PMF 37
Chapter 3 Characterization of saccharides in fine particles collected at rural and urban sites in Texas 42 3.1 Sampling 42 3.2 Analytical methods 43 3.2.1 Element carbon (EC) and organic carbon (OC) analysis 43 3.2.2 Gravimetric analysis 44 3.2.3 Organic speciation 44 3.2.3.1 Extraction 44 3.2.3.2 Analysis 45 3.3 Results 47 3.3.1 Detection limit 47 3.3.2 PM2.5, organic carbon (OC), elemental carbon (EC), and total measured saccharides 48 3.3.3 Aerosol saccharides and seasonal variation 50 3.3.4 Correlation Analysis 59 3.3.5 Rural and urban comparison 61 3.3.6 Source apportionment using PMF 62 3.4 Acknowledgement 69
Chapter 4 Characterization of Organic Fine Particulate Matter in Big Bend National Park and Source Apportionment using PMF 72 4.1 Introduction 72 4.2 Experimental Methods 73 4.2.1 Sample Collection 73 4.2.2 Carbon Analysis 74 4.2.3 Sample Extraction 74 viii
4.2.4 Sample Analysis 75 4.3 Results and Discussion 76 4.3.1 Organic speciation 76 4.3.2 Source Apportionment Using PMF 87 4.3.2.1 Data Analysis by PMF 87 4.3.2.2 Source Profiles 88 4.3.2.3 Relative Source Contributions 94 4.4 Conclusions 98 4.5 Acknowledgement 99
Chapter 5 Saccharide Composition in Fine and Coarse Particulate Matter and Soils in Central Arizona 100 5.1 Introduction 100 5.2 Experimental method 102 5.2.1 PM sampling 102 5.2.2 Soil sampling 103 5.2.3 Element carbon (EC) and organic carbon (OC) analysis 106 5.2.4 Gravimetric analysis 106 5.2.5 Organic speciation 106 5.3 Results and discussion 107 5.3.1 Ambient PM speciation 107 5.3.2 Ambient saccharides 113 5.3.2.1 Saccharide concentrations 113 5.3.2.2 PM10 vs. PM2.5 119 5.3.2.3 Correlation analysis 121 5.3.3 Saccharides in resuspended soil samples 126 5.3.3.1 Total measured saccharides and seasonal change 130 5.3.3.2 Agricultural soil vs. currently uncultivated soil vs. road dust 135 5.3.3.3 Fine vs. coarse soils 136 ix
5.3.4 Ambient saccharides vs. soil saccharides 137 5.3.5 Source Apportionment using PMF 143 5.3.5.1 PMF-resolved source factors 143 5.3.5.2 Relative source contribution to ambient PM2.5 and PM10 and comparison of saccharide rich factors with source profiles of soil samples 146
5.4 Acknowledgement 154
Chapter 6 Source Study of Primary Biological Aerosol Particles 155 6.1 Introduction 155 6.2 Sample collection 156 6.3 Analytical method 158 6.4 Results and discussion 158 6.4.1 Saccharide composition 158 6.4.2 Comparison with Ambient PM Samples and Soil Samples 163 6.4.3 Comparison with PMF Results 164 6.5 Acknowledgement 166
Chapter 7 Characterization of Saccharides in PM10 and Soil Samples from Austria 169 7.1 Introduction 169 7.2 Analytical method 169 7.3 Results and discussion 171 7.3.1 Soil saccharides and comparison with Higley soil samples 171 7.3.2 PM10 saccharides, GM/AX ratios, and method comparison 176 7.4 Acknowledgement 177
Chapter 8 Conclusions 180 8.1 Summary 180 X
8.2 Recommendations for future research 186
References 190 List of Tables
Table 2.1 Sources and markers of ambient aerosols 34
Table 3.1 Characteristics of saccharide standards analyzed by GC-MS as TMS derivatives 46
Table 3.2 GC/MS detection limits for representative saccharide standards 47
Table 3.3 PM2.5, OC, EC and total saccharide ambient concentration statistics 48
Table 3.4 Sugar Concentrations in aerosol samples collected at three sites in Texas
(San Augustine, Clarksville and Dallas) from November, 2005 to July, 2006 54
Table 3.5 Correlation coefficients among 10 saccharide compounds in aerosols collected at the three Texas sites 60
Table 4.1 Ambient concentrations of organic compounds in PM2.5 at BBNP 80
Table 5.1 Sampling date, sample type and site description of Higley soil samples...104
Table 5.2 Ambient concentrations of PM, EC, OC, and organic compounds in Higley
PM samples (Jan - Apr 2008) 110
Table 5.3 Ambient concentrations of saccharides in PM2.5 and PM10 samples at
Higley, AZ 114
Table 5.4 Comparison of total measured saccharide concentrations and the ratio of measured saccharides to OC reported in the literature 116
Table 5.5 Correlation matrix of the concentrations of major saccharide compounds in
Higley PM2.5 and PM10 samples 122 xii
Table 5.6 Correlation coefficients for measured saccharides in PM10 and PM2.5 during different sampling periods at Higley 125
Table 5.7 The range and average concentration of saccharide compounds in resuspended and size-selected soil samples at Higley, AZ 127
Table 5.8 The range and average carbon content of saccharide compounds relative to organic carbon in resuspended and size-selected soil samples at Higley, AZ 128
Table 5.9 Concentrations of major saccharides in soil and dust samples in the literature 131
Table 5.10 Total saccharide content (excluding levoglucosan) relative to PM mass and total carbon content of saccharide (excluding levoglucosan) relative to OC in Higley ambient aerosols and dust and soil samples 139
Table 5.11 Saccharide content relative to PM mass in Higley ambient and soil samples 142
Table 6.1 Plant and spore samples collected in the Phoenix desert area for PBAP source studies 157
Table 6.2 Saccharide composition in PBAP samples selected in current study 162
Table 7.1 Austria bare soil sample information 170
Table 8.1 Total measured saccharide concentrations, PM mass and organic carbon in
PM2.5 samples 189
Table 8.2 Estimated relative contribution of different sources to ambient PM mass for saccharide-related sources resolved by PMF 189 List of Figures
Figure 2.1 Chemical structures of selected organic compounds analyzed in this dissertation work 40
Figure 3.1 Map of Texas showing the three sampling sites at San Augustine,
Clarksville and Dallas, and counties impacted by the Texas-Oklahoma wildfires of
2005-06 43
Figure 3.2 Comparison of the monthly ratio of the carbon content of total measured saccharides to OC in PM2.5 at the three sampling sites in Texas 49
Figure 3.3 Monthly individual concentrations of major saccharides: (a) levoglucosan,
(b) glucose, (c) sucrose and (d) trehalose, in aerosol samples collected at three sites in
Texas from November 2005 to July 2006 56
Figure 3.4 Seasonal variation of sugar polyols at two rural sites - (a) San Augustine,
(b) Clarksville, and one urban site - (c) Dallas, from November 2005 to July
2006 58
Figure 3.5 Source profiles for Dallas samples (Jan - May 2006) 67
Figure 3.6 Relative source contributions to PM2.5 at (a) Dallas and (b) San Augustine site 68
Figure 3.7 Apportioned PM2.5 mass concentrations for Dallas and San Augustine plotted against the measured values 69 xiv
Figure 3.8 Back trajectories of PM2.5 samples with the two highest levoglucosan concentrations at the three Texas sampling sites - (a) San Augustine, (b) Clarksville, and (c) Dallas 70
Figure 4.1 Ambient OC and EC concentrations in PM2.5 samples at BBNP 77
Figure 4.2 Correlation of ambient OC and EC concentrations in PM2.5 samples at
BBNP 77
Figure 4.3 Saw-tooth alkane pattern of PM2.5 samples at BBNP 79
Figure 4.4 AlkaneCPI of BBNP PM2.5 samples 80
Figure 4.5 Ambient saccharide concentrations in BBNP PM2.5 samples (glucose, levoglucosan and sucrose) 84
Figure 4.6 Range of predicted contribution to biomass burning particulate matter at
BBNP 85
Figure 4.7 Source profiles of the 9 factor PMF solution of BBNP PM2.5 samples...92
Figure 4.8 Source contributions to PM2.5 at BBNP resolved by PMF as a function of time 95
Figure 4.9 Relative source contributions to BBNP PM2.5 estimated by PMF 97
Figure 4.10 Apportioned PM2.5 mass concentrations plotted against the measured values for BBNP PM2.5 samples 97
Figure 5.1 Map of Arizona State (USA) including the Phoenix metropolitan area and the ambient sampling location (Higley) of this study 103
Figure 5.2 Basic setup of the soil resuspension and sampling apparatus 105 XV
Figure 5.3 Correlation of ambient OC and EC concentrations in Higley PM2.5 and
PM10 samples 112
Figure 5.4 Distribution of n-alkane isomers showing odd-carbon enrichment as measured in Higley PM samples 112
Figure 5.5 Ambient concentrations of levoglucosan, glucose, sucrose and trehalose in
Higley PM10 and PM2.5 samples during the sampling period (January - April,
2008) 117
Figure 5.6 Average concentration of saccharide compounds measured in PM2.5 and
PM10 from samples collected in Arizona 121
Figure 5.7 PMF modeled factor profiles for Higley PM2.5 and PM10 samples (Jan -
Apr 2008) 149
Figure 5.8 Relative source contributions to (a) PM2.5 and (b) PM10 at Higley 151
Figure 5.9 Apportioned PM mass concentration plotted against the measured values for Higley PM2.5 and PM10 samples 152
Figure 5.10 Comparison of the two PMF source profiles (sucrose and trehalose rich bio-aerosol and glucose and sugar polyol rich bio-aerosol) with the saccharide composition in (a) PM2.5 and (b) inPMlO of the three soil types at Higley 153
Figure 6.1 Composition profiles of major saccharide compounds in PBAP samples 163
Figure 6.2 Comparison of the two PMF source profiles at Higley (sucrose and trehalose rich bio-aerosol and glucose and sugar polyol rich bio-aerosol) with the local PBAP sample source profiles: (a) Comparison with the sucrose and trehalose xvi
rich bio-aerosol factor (PM2.5 and PM10); (b) Comparison with the glucose and sugar polyol rich bio-aerosol factor (PM2.5 and PM10) 167
Figure 7.1 Map of Austria with the soil sampling locations circled in red 171
Figure 7.2 Saccharide compositions in Austria soil samples: (a) mass concentrations
(b) percentages of total measured saccharides 174
Figure 7.3 Comparison of saccharide composition between Austria and Arizona April soil samples: (a) mass concentrations (b) percentages of total measured saccharides 175
Figure 7.4 Saccharide compositions of Austria PM10 samples 178
Figure 7.5 GM/AX ratios and correlation with arabitol and mannitol in Austria PM10 samples 178
Figure 7.6 Comparison of saccharide concentrations in Austria PM10 samples based on two analytical methods (HPLC vs. GCMS) 179 1
Chapter 1
Introduction
1.1 Motivation
Ambient particulate matter (PM) is one of six criteria air pollutants regulated
by the National Ambient Air Quality Standards (NAAQS) established by the US EPA.
Ambient PM together with ozone are the most prevalent criteria pollutants in the
United States, and in particular, ambient PM has been identified as posing the greatest
threat to human health (Hackley et al. 2007). Particles that are 10 micrometers in
diameter or smaller are of most concern because they can be inhaled and penetrate
into the respiratory system, and directly impact health.
Sources of ambient PM are numerous. They can be directly emitted from
primary sources or be formed in the atmosphere from reactions between gases
released from primary sources. Important sources of ambient PM include forest fires,
motor vehicle emissions, fuel combustion, primary biogenic emissions, industrial
processes, cooking, soil and dust entrainment, and secondary aerosol formation.
Among these sources, emissions from biologically derived sources, including biomass
burning, the atmospheric entrainment of soil and dust, and primary biological aerosol particles (PBAPs) such as spores, pollens, fungi, algae, bacteria, virus, and fragments of plants and animals, are significant contributors to both the national and global
loadings of ambient PM. 2
The EPA's 1999 National Emission Inventory estimated that biomass burning and atmospheric entrainment of soil and fugitive dust contributes 20.3 millions tons per year to PM10 and 4.5 millions tons per year to PM2.5 in the United States. By comparison, the emission from fuel combustion for electric power generation plus all on-road and non-road engines and vehicles was estimated to be only 1.0 million tons for PM10 and 0.8 million tons for PM2.5 (NARSTO 2004). Chow et al. (1992) and
Chen et al. (2007) used multivariate receptor models based on ions and elemental markers to estimate that up to 40-60% of the ambient PM10 and 7-23% of the ambient PM2.5 in the San Joaquin Valley in California were derived from fugitive dust. Separate from fugitive dust sources, emissions of PBAPs are ubiquitous, and on a global scale estimates of the PBAP emissions range from 56 Tg (Penner et al. 2001) to 1,000 Tg (Jaenicke 2005) annually. Bauer et al. (2008) used spore counts and concentrations of mannitol and arabitol to estimate the fungal contribution to PM10 in
Vienna, Austria of between 3% and 7% at a suburban sampling location. While these studies are insightful, much less is known about the contribution of soil dust and
PBAPs to aerosol mass, particularly on a regional scale, than for combustion sources.
One approach to separate the contribution of these biologically derived sources and other important contributors to ambient PM levels is organic speciation.
By determining the molecular composition of organic PM in ambient and source samples, attribution studies have determined the relative contribution of different primary sources to ambient particles levels based on the development of specific markers (Schauer et al. 1996; Fraser et al. 2003; Chen et al. 2007; Jaeckels et al. 3
2007). These specific molecular markers include hopanes from lubricating oil in
motor vehicles (Schauer et al. 2002), elemental carbon (EC) for motor vehicle exhaust
(Miguel et al. 1998), levoglucosan from wood combustion (Simoneit et al. 1999),
n-alkanoic and n-alkenoic acids from meat charbroiling (Rogge et al. 1991), and high
molecular weight n-alkanes for resuspension of vegetative waxes and petroleum
hydrocarbons (Simoneit 1984). However, studies on characterizing the contribution of
primary biologically derived sources, specifically PBAPs and the resuspension of soil
and associated biota, have been limited by a lack of appropriate markers.
Recently, saccharides have been proposed as primary molecular tracers for
these biologically important organic materials in natural environments (Simoneit et al.
2004; Medeiros and Simoneit 2007). Saccharide compounds have been identified in
aerosols impacted by forest fires (Oros and Simoneit 1999; Simoneit et al. 1999; Fine
et al. 2004), atmospheric entrainment of soil and associated biota (Simoneit et al.
2004; Rogge et al. 2007), and emissions of PBAPs (Graham et al. 2002; Graham et al.
2003; Bauer et al. 2008). More importantly, unique associations have been found
between specific saccharide compounds and their sources. For example, anhydrous
saccharides such as levoglucosan have been used as tracers for biomass burning
(Simoneit 2002); monosaccharides and disaccharides including glucose, sucrose, and
trehalose have been proposed as marker compounds for the resuspension of soil and
associated biota into the air by wind erosion or agricultural practices (Rogge et al.
2007); and saccharide polyols, in particular mannitol and arabitol, have been reported to originate from PBAPs (Graham et al. 2003; Elbert et al. 2007; Bauer et al. 2008). Therefore, it is possible to elucidate the contribution of fugitive dusts and PBAPs to ambient PM based on the speciation of saccharide markers. By adding these markers for biologically derived organic aerosols, the ambient PM emission inventory will be verified and advanced on a regional basis, which in turn will aide in the development of new air quality improvement plans.
1.2 Research Objectives
The overall objectives of the proposed dissertation research are threefold. The first part is the identification and quantification of carbonaceous compounds with a focus on simple saccharides (i.e. sugars and sugar polyols) in ambient particulate matter. This part involves chemical analysis of multiple categories of organic compounds in aerosol samples (n-alkanes, PAHs, hopanes, acids, saccharides, etc).
The second part is the characterization of major sources contributing to ambient PM levels, with an emphasis on the biologically derived organic carbon sources, mainly atmospheric entrainment of soil and associated biota, and PBAPs. This part involves source studies of soil and PBAP samples, and the use of a multivariate receptor modeling technique to apportion the major PM-contributing sources on the basis of simple saccharides and other established molecular markers. The third part is the use of simple saccharides to compare different aerosol samples based on the location, particle size (fine vs. coarse), season of the year, chemical composition, and relative 5 source contribution. This part involves sampling and analysis of aerosol samples with different sizecuts, at different geographic locations, and during different seasons.
1.3 Hypotheses
The first hypothesis tested in this thesis is that saccharides can serve as useful indicators of carbohydrate production and utilization in the ecosystem, and that specific saccharide compounds can be used as markers in receptor modeling to characterize the contribution from primary biologically derived carbon sources to ambient particulate matter. The second hypothesis to be tested is that other major
PM sources contribute primary biological material to ambient PM2.5 and PM10 beyond soil and associated biota, as is commonly assumed (Simoneit et al. 2004).
Upon completion of this work, novel information about new markers for the primary biologically derived carbon sources will be added to the community of scientific knowledge and the relative contributing strengths of major ambient PM sources in regions with different climate patterns and ecosystems will be determined. This work will be the first of its kind to characterize the contribution of two PM sources - soil and primary biological aerosol material to both PM10 and PM2.5 levels using saccharides as tracers. This work will also answer the question of whether the atmospheric entrainment of soil and associate biota is the dominant contributor to the local ambient PM (as proposed by many studies) or if other routes of entrainment are important at a specific sampling location.
1.4 Approach 6
In order to achieve the goals stated above, several sampling campaigns have been conducted to collect environmental samples. The first sampling was carried out during the second Texas Air Quality Study (TexAQSII) between November 2006 and July 2007 at two rural sites, San Augustine and Clarksville, and one urban site,
Dallas, in eastern Texas. PM2.5 samples were collected for analysis of aerosol saccharide compositions. For this project, the ambient concentration, seasonal variation, and correlation of saccharide compounds have been determined, relative source contributions have been calculated based on the use of saccharides and other molecular marker compounds, and a quantitative comparison of urban versus rural site characteristics has been undertaken based on aerosols collected at the three different locations. Details of this campaign are discussed in Chapter 3 of this dissertation. The second sampling exercise was conducted in Big Bend National
Park (BBNP) in Texas. Ambient PM2.5 samples were collected during summer 2007 to characterize the carbonaceous composition of particles, including elemental carbon
(EC), organic carbon (OC), and organic species including n-alkanes, hopanes, poly cyclic aromatic hydrocarbons (PAHs), and saccharide compounds. These organic markers were also used to identify and quantify sources contributing to the PM levels at BBNP. Details of this campaign are presented in Chapter 4. The third sampling campaign was conducted in Higley, Arizona, with a series of ambient PM and soil samples collected between January and April 2008. Saccharides and other carbonaceous compounds were analyzed in size-selected soil and ambient PM samples. Comparisons were made between the ambient PM and three types of soil 7
dust samples based on the particle size (fine vs. coarse), season, and the relative composition of saccharide compounds. Based on the aerosol concentrations of saccharides and other molecular markers, a PMF model was performed to estimate the major sources and their relative contributions to ambient PM. Two saccharide-rich factor profiles resolved by PMF were also compared with the soil saccharide composition to see if they corroborate with each other. Details of these results are discussed in Chapter 5. Chapter 6 is a continuation of the PM source study, in which the saccharide profiles of another large source category - PBAPs, were characterized based on plant and spore samples that are typical for the desert environment.
Comparison was made between the saccharide compositions of the collected PBAP samples, soil samples, and two saccharide-rich factors resolved by PMF modeling to assess the relative importance of PBAPs and soils as ambient PM sources. Chapter 7 presents a sample exchange project with the Institute of Chemical Technologies and
Analytics at Vienna University of Technology, Austria, in which ambient PM10 samples and bare soil samples collected in 2005 at various locations in Austria were analyzed for saccharide composition. The saccharide profiles of Austria soil samples were compared with the saccharide profiles of Arizona soil samples. The saccharide concentrations of Austria PM10 samples were assessed and compared with results from an equivalent measurement conducted at the Vienna University of
Technology to check the consistency of two different analytical methods (GCMS vs.
HPLC). In the end, Chapter 8 summarizes the key findings in this thesis and provides recommendations for future research. 8
The analytical approaches used for these studies are based on organic solvent extraction and chemical analysis of aerosol and soil samples. Collected samples were extracted using a set of solvents, and target compounds were analyzed in either their original extracted forms or as molecular derivatives using gas chromatography (GC) coupled with mass spectrometry (MS). A review of currently available analytical protocols can be found in Chapter 2 (background and literature review), and a detailed description of each adopted approach is stated in the experiment section of each chapter. The modeling approach for source apportionment used in this work is positive matrix factorization (PMF), which is a multivariate receptor model used to analyze statistical trends in ambient data sets and to isolate the source profile and source contribution by using physically meaningful limitations. A detailed description of the PMF theory and application can be found in Chapter 2 as well. Another statistical approach adopted in this work is correlation analysis, which is used to resolve the relationship between different environmental parameters and help elucidate sources by distinguishing different groups of compounds that are highly correlated within each group. This approach is also briefly reviewed in Chapter 2. 9
Chapter 2
Background and literature review
2.1PM
Particulate matter (PM), also known as atmospheric particles or aerosols, refers to a complex mixture of solid material and liquid droplets suspended in the atmosphere. They are either emitted directly from a source as primary particles, or formed through in-situ reactions from the gas phase as secondary particles (Seinfeld and Pandis 1998). The size of particulate matter ranges from a few nanometers to tens of micrometers. Due to their potential for causing health problems, particles that are
10 micrometers in diameter or smaller are of primary concern, and are known as
PM10. PM10 can be further divided into two particle categories - PMf and PMC. PMf, also known as PM2.5 or fine particles, has an aerodynamic diameter less than or equal
to 2.5 urn, while PMC, also referred to as PM10-2.5 or inhalable coarse particles, has an aerodynamic diameter between 2.5 um and 10 um.
2.1.1 Health effect and regulation
PM10 including PM2.5 and PMC is of concern because of its threat to both our lungs and heart. PM2.5 has long been recognized as a great health concern because fine particles can be inhaled into and accumulate in the human respiratory system
(NARSTO 2004). Increased concentrations of fine particles have been associated in many studies with health problems including aggravated asthma, reduced lung 10
function and lung cancer, cardiovascular problems, and premature death (Wilson and
Spengler 1996). For example, Peters et al. (2000) reported that acute exposures of one
hour duration to elevated fine particle concentration were linked to irregular heart
rhythms in people with heart disease. Many long-term epidemiological studies have
also demonstrated that people living in areas with large fine particle concentrations have an increased risk of premature death compared to those in cities with lower fine
PM levels (Krewski et al. 2003). Pope et al. (2002) found the risk of premature death
from cardio-respiratory diseases and lung cancer is higher in more polluted areas, and
several studies have found the risk of mortality to increase almost linearly with PM
concentrations (Dockery and Pope 1994; Schwartz 1994). In a mortality study
conducted by Schwartz et al. (1996) in Amsterdam, an increase of 10 ug m"3 in the
48-hr mean PM2.5 concentrations resulted in an increase of daily mortality of 1.5%. It has also been reported that adverse health effects from fine particles can be
aggravated in sensitive groups such as the elderly, children, people with existing heart
and lung disease, and pregnant women (Bates 1995; Woodruff et al. 1997; Goldberg
et al. 2000; Pope 2000).
PMc also has the potential to affect lung function and aggravate symptoms,
particularly in asthmatics due to the deposition of coarse particles in the tracheobronchial region of the human lungs (U.S.EPA 2005). For example, recent studies have found cellular, immunological effects (Kleinman et al. 1995; Steerenberg et al. 2003) and pro-inflammatory effects (U.S.EPA 2004) associated with PMc.
Lipsett et al. (2006) observed a significant perturbation in heart rate variability 11 associated with PMc in seniors exposed to PMc in California's Coachella Valley; results peaked shortly after exposure and lasted a few hours. Under some conditions, crustal particles (main component of PMc) may undergo alteration to become more harmful (U.S.EPA 2005). For example, resuspended crustal particles may be contaminated with toxic trace elements and other components from previously deposited fine PM, e.g., metals from smelters or steel mills, PAHs from automobile exhaust, or pesticides from agricultural lands (U.S.EPA 2004).
Because of the impact on human health, both PM10 and PM2.5 have been regulated by the Clean Air Act as criteria air pollutants. In July 1997, U.S. EPA modified existing National Ambient Air Quality Standards for PM and set an annual average level of 50 ug m"3 and a 24-hr standard of 150 ug m"3 for PM10 and an annual average standard of 15 ug m"3 and a 24-hr standard of 65 ug m"3 for PM2.5. In
September 2006, EPA revised the national air quality standard, and lowered the annual average standard for PM2.5 to 35 ug m"3 as a result of mounting evidence of the short-term health effects of PM2.5, while the annual PM10 standard was revoked
"given that the available evidence does not suggest an association between long-term exposure to coarse particles at current ambient levels and health effects" (U.S.EPA
2006). The 24-hr PM10 standard was retained but was proposed to be replaced by a
PM10-2.5 (PMc) standard which will be revisited by the EPA in the near future
(U.S.EPA 2006). 2.1.2 Carbonaceous composition and sources
The carbonaceous fraction of ambient particulate matter consists of elemental carbon and a variety of organic compounds (Seinfeld and Pandis 1998). In the lower troposphere, carbonaceous particles account for up to -50% of fine air particulate matter (Poschl et al. 2005). This fraction of PM has been widely studied for a variety of reasons. From the health effect point of view, there are thousands of harmful and potentially harmful organic compounds associated with the carbonaceous portion of ambient PM (Seinfeld and Pandis 1998). From a climate modification perspective, the carbonaceous fraction of aerosols is able to act as cloud condensation nuclei thus play an important role by affecting cloud formation (Gelencser 2004). Finally, individual organic species contained in this fraction are useful in tracking emissions from sources (Schauer et al. 1996).
2.1.2.1 Elemental carbon
Elemental carbon (EC), also known as graphitic carbon, is predominantly emitted by combustion sources and is not formed by secondary reactions. Ambient concentrations of EC range from less than 1 (xg m"3 in remote areas to tens of ug m"3 in urban areas (Seinfeld and Pandis 1998). Sources of ambient EC mainly include biomass burning and fossil fuel combustion, especially by diesel engines (Gelencser
2004). The ratio of EC to total carbon is indicative of aerosol age, and a decrease of this ratio suggests the aging of an airmass (Burtscher et al. 1993). The ratio of EC to organic carbon (OC) can be used to distinguish primary OC from secondary OC, given EC and primary OC often have the same sources (Turpin and Huntzicker 1995; 13
Russell and Allen 2004). Ambient EC particles emitted by automobile are predominantly less than 0.2 micrometers (Venkataraman and Low 1994), and therefore are considered to be a significant human health concern. EC also has an impact on the global climate by absorbing sunlight and heating up the atmosphere
(Andreae and Gelencser 2006).
2.1.2.2. Organic carbon
Organic carbon (OC) makes up a large fraction of ambient PM, typically
20-50% of aerosols in mid-latitudes (Saxena and Hildemann 1996). In tropical forests, the percentage rises to about 90% (Kanakidou et al. 2005). The organic carbon fraction of ambient PM is made up of hundreds to thousands of individual organic species. Compounds that have been identified in the ambient fine particles include n-alkanes, n-alkanoic acids, n-alkanals, aliphatic dicarboxylic acids, diterpenoid acids, aromatic polycarboxylic acids, polycyclic aromatic hydrocarbons, steroids, triterpenoids, saccharides, N-containing compounds, steranes, hopanes, and others
(Mazurek et al. 1989; Hildemann et al. 1993; Rogge et al. 1993; 1993; 1993).
Sources of these compounds include biomass burning (Schauer et al. 2001; Simoneit
2002), fossil fuel combustion (Rogge et al. 1993; Kanakidou et al. 2005), tire and break wear abrasion (Rogge et al. 1993), cooking (Rogge et al. 1991), industrial boilers (Rogge et al. 1993; 1997), biogenic emission (Graham et al. 2003; Jaenicke
2005), and soil and dust resuspension (Simoneit et al. 2004). This portion of PM has been intensively studied during the past ten to twenty years (Simoneit and Mazurek
1982; Simoneit 1984; Fraser et al. 1998; Elias et al. 1999; Nolte et al. 1999; Schauer 14 et al. 1999; 1999; Nolte et al. 2001; Schauer et al. 2001; 2002; 2002). One reason for that is the possibility that these organic carbon compounds pose a great health hazard.
The number size distributions of fine PM are mainly in the submicrometer range
(McMurry and Zhang 1989), and this portion of particles is considered to be a significant human health concern because of the ability of these particles to penetrate deeply into the respiratory system (Oberdorster and Utell 2002). Also, certain fractions of particulate organic matter, especially those containing polycyclic aromatic hydrocarbons (PAHs), have been shown to be carcinogenic in animals and mutagenic in in vitro bioassays (Seinfeld and Pandis 1998).
n-Alkanes
Hydrocarbons which contain only single bonds are called alkanes. n-Alkanes
(n for normal) refer to the simplest isomers of alkanes in which the carbon atoms are arranged in a single chain without branches. n-Alkanes range in volatility and include species that are entirely in the gas phase, that partition between phases, and that exist entirely in the particle phase. Normally n-alkanes with carbon number greater than 26 are considered to be entirely in the particle phase (Fraser et al. 1997). n-Alkanes are primary pollutants that are commonly measured in urban areas, and their presence in the atmosphere originates from two major source categories: fossil fuel combustions and biogenic emissions (i.e., plant wax). Particulate n-alkanes in vehicular emissions are derived from incomplete combustion of fuel and from lubricant oil (Rogge et al.
1993), while n-alkanes of biological origin are produced by mechanical processes 15
(mainly in coarse particles) and biomass burning (mainly in fine particles) (Kadowaki
1994). Due to the decarboxylation process of even-numbered fatty acids (Harborne
1998), n-alkanes from leaf wax and other biogenic organic particle sources exhibit a strong odd carbon number predominance with the dominant wax n-alkanes being C29,
C31, and C33 (Rogge et al. 1993), contrasting n-alkanes from fossil sources that have roughly equal concentrations of odd and even homologues due to equilibrium of continuous decarboxylation within the sedimentary organic material (Cooper and
Bray 1963; Simoneit 1984). In this regard, a carbon preference index (CPI) can be used to provide information about the source origins of n-alkanes. CPI values are calculated as the ratio of the sum of odd-carbon isomers to the sum of even-carbon isomers for n-alkanes ranging from C26 to C33 (Kavouras and Stephanou 2002). A
CPI value close to 1 is regarded as characteristic of petroleum hydrocarbons, while a higher CPI value indicates impact from biogenic sources (Simoneit 1984; Mazurek et al. 1989).
PAH compounds
Polycyclic aromatic hydrocarbons (PAHs) are chemical compounds that contain fused aromatic rings and do not contain heteroatoms or carry substituents
(Fetzer 2007). PAHs are present in both gas and particle phase in the atmosphere, with lighter PAHs, such as phenanthrene, found almost totally in the gas phase, and the heavier PAHs, such as benzo[a]pyrene, found absorbed on to particles (Ravindra et al. 2008). PAHs in ambient PM are formed primarily in combustion processes of 16 either vegetative or fossil fuel materials (Simoneit 1984), with vehicular emissions and biomass burning as the major sources of PAHs (Ravindra et al. 2008). The anthropogenic PAH sources, mainly from fossil fuel combustion, are by far the major contributors of hydrocarbons with known hazards to human health (Simoneit 2002), and most of the probable human carcinogenic PAHs are commonly measured to be associated with particulate matter, especially in the fine mode (Ravindra et al. 2008).
Higher-molecular-weight PAHs largely originate from motor vehicle exhaust, and prior studies suggest gasoline vehicles emit more high-molecular-weight PAHs as compared to diesel vehicles (Miguel et al. 1998; Watson 1998; Zielinska et al. 2004).
The commonly analyzed PAH compounds include fluoranthene, pyrene, chrysene, benz[a]anthracene, benzofluoranthenes, benzo[a]pyrene, perylene, indeno[l,2,3-cd]pyrene, dibenz[a.h]anthracene and benzo[ghi]perylene. Structures of these PAH compounds can be found in Figure 2.1.
Hopanes
Hopanes are a group of hydrocarbons that have four C6 rings and one C5 ring as base structures. Hopanes are the degraded and saturated version of hopanoids, and have been stable over geological time scales (Hunt 1979). They have been extensively used in geochemistry as fossil petroleum markers (Peters and Moldowan
1993). In atmospheric chemistry, hopanes are also used as petroleum biomarkers to track vehicular emissions from gasoline and diesel engines (Rogge et al. 1993). The predominance of 17a(H), 21 P(H)-hopane is an indicator of fossil petroleum input or 17 thermal maturation (Standley and Simoneit 1987). The ratio between the ambient concentrations of hopanes and total OC can be used to estimate the contribution of mobile sources to primary aerosol OC mass based on vehicle emission source profiles which indicate that each gram of vehicle-emitted OC contains 0.002 grams of hopanes, including 18a(H)-22,29,30-trisnorneohopane, 17a(H), 2ip(H)-29-norhopane, 17a(H),
21p(H)-hopane, 22S-17a(H), 21p(H)-30-homohopane, 22R-17a(H),
2ip(H)-30-homohopane, 22S-17a(H), 2ip(H)-30-bishomohopane, and 22R-17a(H),
2ip(H)-30-bishomohopane (Rogge et al. 1993; Brown 2001). Other commonly analyzed hopanes also include 17a(H)-22,29,30-trisnorhopane and
18a(H)-29-norneohopane. The structure of the most common hopane compound,
17a(H), 21 P(H)-hopane, is listed in Figure 2.1. Other hopanes are isomers or derivatives of this species that share the similar parent structure (four C6 rings and one C5 rings).
Carboxylic acids
Carboxylic acids are organic hydrocarbons that contain a carboxylic acid group (-COOH). Important carboxylic acids of interest in tracking sources of particulate matter include n-fatty acids (n-alkanoic and n-alkenoic acids) with even number of carbons, and dicarboxylic acids with two carboxyl groups.
Sources of n-fatty acids can be both anthropogenic and biogenic.
Anthropogenic sources include the combustion of fossil fuels, wood and organic detritus (Rogge et al. 1993). In particular, meat cooking aerosols were found to be 18
dominated by n-alkanoic acids including hexadecanoic acid (CI6) and octadecanoic acid (CI8), and n-alkenoic acids including palmitoleic acid (C16:1A9) and oleic acid
(CI8:1) (Rogge et al. 1991). Biogenic emission sources of n-fatty acids include epicuticular plant waxes, fungi, bacteria, spores, pollen, algae, seeds and seed oils
(Simoneit 1989). An indicator of biogenic sources is the strong even-to-odd carbon preference (Rogge et al. 1993; Rogge et al. 1993). Simoneit and Mazurek (1981) have noted that fatty acids
C18:l to C18:0 ratio can be used as indicative of the degree of aging of particulates and also reflects the presence of photochemical oxidants (Simoneit and Mazurek 1982;
Zheng et al. 2000).
Dicarboxylic acids are predominantly of secondary origin, and mainly come from the photooxidation of unsaturated hydrocarbons and fatty acids, both biogenic and anthropogenic (Kawamura et al. 1996; Gelencser 2004). Among the various dicarboxylic acids that have been identified in ambient particulate matter, one that is worthy of special mention is pinic acid, which is formed by the photooxidation of a-pinene in presence of ozone (Pandis et al. 1992). Another important carboxylic acid similar to pinic acid is pinonic acid, also a photooxidation product of a-pinene with ozone and other oxidants. Both pinic and pinonic acid have low vapor pressures and are expected to contribute significantly to secondary organic aerosol (SOA) formation 19
(Yu et al. 1999). Analysis of the ambient concentrations of pinic and pinonic acids is of importance because it can be used to track a-pinene oxidation and SOA formation. a-Pinene is an important gas-phase monoterpene species that is highly reactive, and monoterpenes are one of the primary classes of biogenic volatile organic compounds that form SOA (Griffin et al. 1999). Dicarboxylic acids are also considered as important contributors to the water-soluble organic compounds (WSOC) due to their strong hydrophilic and hygroscopic properties, therefore they are potentially able to act as cloud condensation nuclei (CCN) altering cloud formation (Kawamura and
Sakaguchi 1999). The compound structures of pinic acid and pinonic acid are listed in Figure 2.1.
Saccharides
Saccharides, also called carbohydrates, are simple hydrocarbons that are aldehydes or ketones with many hydroxyl groups. Saccharides are the single most abundant class of organic molecules found in nature, and they represent the major form of photosynthetically assimilated carbon in nature (Cowie and Hedges 1984). Of relevance to particulate matter are mainly simple saccharides, i.e., sugars (referring to any monosaccharide or disaccharide) and sugar polyols (or sugar alcohols).
Significant amounts of sugars have been reported to be present in urban, rural, and marine aerosols worldwide (Simoneit and Elias 2000; Nolte et al. 2001; Graham et al.
2002; Pashynska et al. 2002; Zdrahal et al. 2002; Simoneit et al. 2004; Simoneit et al.
2004; Simoneit et al. 2004; Yttri et al. 2007). Sugar polyols (reduced sugars) were 20 also found in aerosol at various locales (Graham et al. 2002; Carvalho et al. 2003;
Graham et al. 2003; Yttri et al. 2007; Bauer et al. 2008). Compound structures of the saccharides analyzed in this work can be found in Figure 2.1. The next section will discuss in more details the characteristics and sources of saccharide species in ambient particulate matter.
2.2 Saccharides in aerosols
2.2.1 Importance
Saccharide compounds are important constituents of ambient aerosols.
Approximately 13-26% of the total compound mass identified in continental aerosols was saccharides and this number can reach as high as 63% in oceanic aerosols
(Simoneit et al. 2004; Simoneit et al. 2004). Saccharide compounds have been proposed as one of the primary biogenic markers observed in aerosol samples
(Simoneit 1999; Graham et al. 2002). They represent the contribution of carbohydrates in the ecosystem to the atmospheric compartment from biomass burning (Simoneit et al. 1999; Simoneit 2002), atmospheric entrainment of soil and associated biota (Simoneit et al. 2004; Rogge et al. 2007), and more recently primary biological aerosol particles (PBAPs) (Graham et al. 2003; Elbert et al. 2007; Bauer et al. 2008). The contributions of the three sources shown to include saccharides to ambient PM are significant. Forest fires and the atmospheric entrainment of soil and 21
dust are reported as major sources of airborne particulate matter in the United States
(NARSTO 2004). On a global scale, the contribution from primary biological aerosol particles (PBAPs) also constitutes a major portion of atmospheric aerosols (Jaenicke
2005).
Saccharides have been recognized as an important component of the
water-soluble organic compounds (WSOCs) within aerosols (Saxena and Hildemann
1996). The water-soluble organic matter documented for aerosol particles consists of
short-chain dicarboxylic acids, anhydrosaccharides, saccharides, and secondary
oxidation products such as methoxyphenolics and aromatic acids (Kawamura and
Usukura 1993; Kawamura and Sakaguchi 1999; Simoneit et al. 2004). Previous
studies have focused on the chemical characterization of the non-polar fraction of the
OC in fine particles (Mazurek et al. 1987; Zheng et al. 2002), which only accounted
for 10-30% of the total OC mass (Rogge et al. 1993; Schauer et al. 1996; Simoneit et
al. 2004). Recently, WSOCs have been widely studied because they are found to
contribute significantly to OC, especially in rural and remote atmospheres (Mader et
al. 2004). WSOCs generally account for 20-70% of total aerosol carbon (Saxena and
Hildemann 1996). WSOCs also play an important role in the formation of cloud
condensation nuclei (CCN) and thus the regulation of climate (Graham et al. 2002;
Kanakidou et al. 2005).
As a major form of photosynthesized carbon in the ecosystem, saccharides have been proposed to be potentially useful in understanding sources, processing and transport of biologically important organic materials in the natural environment 22
(Medeiros and Simoneit 2007). The analysis of sugar contents in ambient particulate
matter is important for aerosol source apportionment and for the understanding of
their related source categories on local and regional air quality.
2.2.2 Saccharide species and associated sources
Saccharide species that have been detected in aerosols and associated with
contributing sources are mainly simple saccharides, which include anhydrous
saccharides, monosaccharides, disaccharides, and saccharide polyols. The chemical
structures of saccharides analyzed in this dissertation work can be found in Figure 2.1.
Discussion of saccharide analysis in this thesis refers to simple saccharides only.
Andydrous saccharides
Anhydrous saccharides (i.e. levoglucosan, mannosan and galactosan) are
widely acknowledged as markers for biomass burning. Levoglucosan
(1,6-anhydro-p-D-glucopyranose) is a pyrolytic decomposition product of cellulose
and is the most important tracer for smoke particulate matter from burning of biomass
due to its atmospheric stability (Fraser and Lakshmanan 2000) and the fact it has been
shown to be generated only by combustive processes (Simoneit et al. 1999; Simoneit
and Elias 2000). Mannosan (1,6-anhydro-P-D-mannopyranose) and galactosan
(1,6-anhydro-P-D-galactopyranose) are two isomers of levoglucosan and are produced
from hemicellulose decomposition (Simoneit 2002). They are less often used as tracers for biomass burning than levoglucosan because the emitted amounts of 23 levoglucosan are substantially higher (Zdrahal et al. 2002). Levoglucosan is often the most prevalent sugar component in fine particles and can contribute to more than 90% of the total sugar mass in tropical areas impacted by wood smoke (Simoneit 2002).
Monosaccharides
Monosaccharides (also called simple sugars) are the simplest saccharides.
Monosaccharides consist typically of three to seven carbon atoms and are described either as aldoses or ketoses, depending on whether the molecule contains an aldehyde or a ketone functional group. The primary monosaccharides commonly found in PM are a- and 0- glucose. Glucose contains six carbon atoms and has an aldehyde group, therefore is referred to as an aldohexose. Glucose is the most common monosaccharide present in vascular plants (Cowie and Hedges 1984). It has been recognized as a product of cellulose pyrolysis (Shafizadeh et al. 1979) and serves as a major source of carbon for soil microbes (Paul and Clark 1996). Glucose has frequently been identified as a dominant sugar compound (apart from levoglucosan) in aerosols (Simoneit et al. 2004). It is worth mentioning that aerosol glucose may undergo degradation during sample transport and storage as it is an important carbon source for microorganisms. However, given the sample storage condition (-4°F) in this study, this impact should be small therefore was not taken into account.
Disaccharides 24
Disaccharides are the simplest oligosaccharides, which consist of two monosaccharide (the hexoses glucose, fructose, mannose, and galactose and the pentoses ribose and xylose) units linked by a glycosidic bond. Two commonly measured disaccharides in aerosols are sucrose and trehalose (or mycose). Sucrose is a disaccharide of glucose and fructose, and a major storage disaccharide for fixed CO2 in plants (Martin et al. 1988). Sucrose is the principal sugar in the phloem of plants and is particularly important in developing flower buds (Bieleski 1995). Pacini (2000;
2006) reported sucrose as a major component in pollen grains. Sucrose is also an important contributor to soil sugar content (Medeiros et al. 2006; Al-Mutlaq et al.
2007). Trehalose is a dissaccharide of two glucoses. It is widely present in bacteria, yeast, fungi, algae, a few plants, and invertebrates (Elbein 1974). Trehalose is known as a metabolite of mycorrhyzal fungi converted from glucose (Martin et al. 1988;
Niederer et al. 1989). It is generally regarded as a main reserve carbohydrate in yeast and is also highly efficient in protecting the structural integrity of the cytoplasm under environmental stress. Large quantities of trehalose (up to 35% of dry weight) are frequently found in anhydrobiotic organisms capable of surviving complete dehydration (Crowe et al. 1984). The primary role of trehalose is to protect components of the cytosol against adverse conditions such as desiccation, frost, and heat (Wiemken 1990). Trehalose was often the most abundant sugar found in soil samples (Simoneit et al. 2004; Medeiros et al. 2006; Rogge et al. 2006; 2007).
Saccharide polyols 25
Saccharide polyols (also known as saccharide alcohols, sugar polyols, or sugar alcohols) are the reduced form of saccharides when the aldo or keto group of a saccharide is reduced to the corresponding hydroxyl group. Saccharide polyols are widespread in fungi, algae, lichens, and higher plants (Lewis and Smith 1967). They are more reduced than other saccharide forms, therefore suitable to act as energy and carbohydrate reserves (Lewis and Smith 1967; Bieleski 1982). Other functions of saccharide polyols have also been addressed in literature, such as acting as agents of translocation and endogenous osmo-regulation against salt stress and water deficiency in plants and bacteria (Lewis and Smith 1967). Saccharide polyols might also be emitted by thermal stripping during wildfires, although only limited evidence was reported (Graham et al. 2002; Simoneit et al. 2004). Mannitol is by far the most abundant saccharide polyol in nature (Lewis and Smith 1967). It is widely present in the two major fungi groups Ascomycetes and Basidiomycetes (Pfyffer and Rast 1980).
Mannitol is often associated with another important polyol arabitol, as well as the disaccharide trehalose in the organisms of these two groups. These three compounds take the place of glucose and sucrose in metabolism acting as the main storage and respirable carbohydrates, with mannitol being the direct source for carbohydrate utilization (Lewis and Smith 1967). Mannitol and arabitol have been proposed as tracers for airborne fungal spores because their aerosol concentrations were found to be highly correlated with the measured fungal spore counts and because no other emission sources of these two compounds have been reported (Bauer et al. 2008).
Unlike mannitol and arabitol, sorbitol occurs generally in higher plants of the woody 26
Rosacease (Caseiro 2008). It can also be formed and accumulated in bacteria in order to overcome osmotic stress (Loos et al. 1994). Another major saccharide polyol is glycerol, which has been reported as a metabolic product in soil organisms (Simoneit et al. 2004) or could possibly derive from hydrolysis of fatty acid glycerides (Martin et al. 1988; Simoneit et al. 2004). Aerosol glycerol is also found to be highly correlated with black carbon (BC), organic carbon (OC) and potassium (K) at representative pasture and primary rainforest sites in Rondonia, Brazil, which suggests a combustion source (Graham et al. 2002). Other saccharide polyols such as ribitol, xylitol, galactitol, and erythritol are less prevalent. Sources of these polyols include fruits, hardwood, lichens, yeasts, fungi, and soil biota (Lewis and Smith 1967;
Cox and Wathes 1995; Simoneit et al. 2004). In aerosols collected from various environments, glycerol, mannitol, and arabitol are generally the most abundant of all the polyols reported (Graham et al. 2002; Pashynska et al. 2002; Carvalho et al. 2003;
Simoneit et al. 2004; Yttri et al. 2007).
2.3 Analysis
2.3.1 Organic solvent extraction
In the past decades, the identification and quantification of organic carbonaceous compounds in aerosols has mostly been achieved using GC-MS after organic-solvent extraction and silylation derivatization (Chizhov et al. 1967; DeJongh et al. 1969; Modzeleski et al. 1971; Martinez-Castro et al. 1989; Garcia-Raso et al. 27
1992; Fiehn et al. 2000). Existing analytical methods include the use of different suites of extracting solvents.
Analytical methods for the analysis of the non-polar organic carbonaceous portion of aerosols use n-hexane, benzene and isopropanol as extracting solvents
(Mazurek et al. 1987). This combination of extraction solvents has been studied extensively in previous work, especially for quantification of non-polar organics from petroleum products, fossil hydrocarbons, vascular plant waxes, and lipids (Simoneit
1989; Rogge et al. 1991; Rogge et al. 1993; Nolte et al. 2001; Fraser et al. 2002; Nolte et al. 2002; Zheng et al. 2002). However, in this method, a substantial portion of polar
(oxygenated) organic compounds, including the more water-soluble ones, remains unextracted (Decesari et al. 2000).
Recent studies targeting the more polar content of environmental samples involve the use of dichloromethane and methanol in different proportions (Pashynska et al. 2002; Simoneit et al. 2004; Simoneit et al. 2004; Simoneit et al. 2004; Jaffe et al.
2006; Medeiros et al. 2006; Medeiros et al. 2006; Rogge et al. 2006; Fuzzi et al. 2007;
Medeiros and Simoneit 2007; Rogge et al. 2007; Bauer et al. 2008). The advantage of using the dichloromethane/methanol mixture (1:1 to 4:1, v/v) is that this suite has been shown to extract both neutral (lipids) and polar (water-soluble) compounds, without hydrolysis of organic anhydrides. This is a well established and tested method in organic geochemistry (Simoneit et al. 2004) and has been shown to extract polar compounds in environmental samples with very high efficiency (Simoneit et al. 2004;
Rogge et al. 2006). Several studies also use solely methanol (Graham et al. 2003) or 28 dichloromethane and water (Alves et al. 2002; Carvalho et al. 2003) as the extraction solvent. However, these two methods are not widely practiced mainly because the first method is only efficient in extracting polar compounds, while the second one involves a two-step extraction and is thus more tedious.
In terms of extraction of saccharides and other water-soluble organic compounds, several experiments also used water as the extracting solvent and showed almost 100% recovery for these water-soluble compounds (Debosz et al. 2002;
Graham et al. 2002; Simoneit et al. 2004; Yttri et al. 2007). However, despite the fact that these target compounds are water-soluble, the extraction into water is limited because water also extracts polymeric humic-like substances that would not pass through the gas chromatographic column and cause rapid injector and column deterioration (Decesari et al. 2000).
2.3.2 GC-MS and derivatization
Gas chromatography coupled with mass spectrometry (GCMS) has been widely used in the analysis of carbonaceous compounds in ambient aerosols
(Simoneit and Mazurek 1982; Mazurek et al. 1987; Rogge et al. 1993; Medeiros and
Simoneit 2007; 2007) because of its capability of molecular identification of organic compounds of different polarities at a high level of sensitivity (Jacobson et al. 2000).
In terms of the analysis of saccharides, there are currently five categories of existing analytical methods: gas chromatography (GC)-based methods (Graham et al.
2002; Pashynska et al. 2002; Carvalho et al. 2003; Graham et al. 2003; Rogge et al. 29
2006; Medeiros and Simoneit 2007; Rogge et al. 2007; Bauer et al. 2008), high performance liquid chromatography (HPLC)-based methods (Cheng et al. 1992;
Engling et al. 2006; Caseiro et al. 2007; Pio et al. 2008; Iinuma et al. 2009), direct mass spectrometry (MS) methods (Black and Fox 1996), liquid chromatography-mass spectrometry (LC-MS) methods (Dye and Yttri 2005; Wan and Yu 2006), and capillary electrophoresis (CE)-based methods (Wang et al. 2001; Cheng et al. 2006).
Apart from GCMS, HPLC is the most used method for the analysis of saccharide compounds because it utilizes a chromatographic column suitable for carbohydrate analysis and a variety of detection methods, including electrochemical detection
(Clarke et al. 1991), refractive index (RI) detection, and evaporative light scattering detection (ELSD) (Wei and Ding 2000). However, due to the lack of chromophoric and fluorophoric moieties of saccharide compounds required for UV and fluorescence detection, less sensitive HPLC detection methods such s RI and ELSD are used.
Therefore, these methods are not practical for a trace level detection of aerosol saccharides (Cheng et al. 2006; Wan and Yu 2006).
Up to date, the GC-MS analytical procedure is still preferred and widely practiced because it is capable of trace-level detection and can simultaneously resolve compounds of different polarities. However, due to their high polarity, hydrophilicity, and low volatility, polar compounds including saccharides, acids, steroids, and terpenoids have to be converted into more volatile and stable derivatives before
GC-MS analysis. The most common derivatization technique is silylation, which converts the hydroxy 1 groups of polar compounds into their trimethylsilyl (TMS) 30
ether derivatives and makes them more volatile, less polar, and more thermally stable
(Sweeley et al. 1963). This results in improved separation and detection with GC techniques. The combination of BSTFA (N,0-bis(trimethylsilyl)trifluoroacetamide) and TMCS (trimethylchlorosilane) is a preferred reagent for trimethylsilylation of a series of polar compounds, and the resulting TMS derivatives have the advantages of easy formation, excellent chromatographic qualities, and long-term chemical stability
(Blau and Halket 1993). The use of pyridine as a catalyst accelerates the rate of reaction because it can act as an HC1 acceptor in silylation reactions (Langer et al.
1958).
2.4 Statistical Tool - Correlation Analysis
The use of a statistical tool such as correlation analysis can help resolve the relationship between different organic compounds in an environmental sample and help elucidate sources by distinguishing different groups of compounds that are highly correlated within each group. In the application of correlation analysis to saccharide compounds in environmental samples, previous studies have shown correlations between pairs of saccharide species in both soil and aerosol samples.
For instance, Lewis and Smith (1967) found interrelationships between the metabolism of trehalose and mannitol in fungi. Maenhaut et al.(2002) observed correlations between glucose and fructose, as well as between arabitol and mannitol, for aerosols collected in South Africa. Wan and Yu (2006) revealed moderate to 31 strong correlation (r=0.6-0.9) among the primary saccharide (e.g., glucose, sucrose) and saccharide polyols (e.g., erythritol, mannitol) in aerosols. Medeiros et al. (2006) observed a strong correlation between mannitol and trehalose (r=0.86), and a moderate correlation between glucose and most saccharides (r=0.2~0.8) in a ryegrass soil.
The fact that different saccharide compounds were well correlated with each other suggests that these compounds share common dominant sources. In an aerosol study conducted by Pashynska et al.(2002), levels of glucose and sucrose were reported to be highest in early summer, while arabitol and mannitol concentrations peaked together later in summer. They suggested that glucose and sucrose may have been associated with airborne detritus from developing leaves and the saccharide polyols with detritus from mature leaves. Also, Graham et al.(2003) found in
Amazonian aerosol samples that trehalose, arabitol, and mannitol appeared to be associated with yeasts and other fungal spores present in higher concentrations at night, while glucose and sucrose were more linked with the specific daytime release of pollen, fern spores, and other bioaerosols probably driven by lowered relative humidity and enhanced convection during the day.
The correlation analysis tool adopted in this study is the Pearson's correlation test using R statistical software (R 2.9.1, a free software environment originally developed in the Statistic Department of the University of Auckland). The Pearson's correlation (short for the Pearson Product Moment Correlation) is a parametric statistical test, which assumes the testing subject follows a normal distribution (Neter 32
et al. 1996). It calculates the correlation coefficients between two random variables measured on the same object, which reflects the degree of linearity between these two variables (Neter et al. 1996). Pearson's correlation coefficients range from +1 to -1, suggesting statistically the tendency of the variables to increase or decrease together.
A correlation of+1/-1 means that there is a perfect positive/negative linear relationship between variables. Cohen (1988) suggested an absolute value of correlation coefficient greater than 0.5 should be considered as an indication of statistically significant correlation.
2.5 Source apportionment
Quantitatively identifying the relative contributions of different source types to ambient particulate matter concentrations at a receptor location is referred to as source apportionment. Source apportionment techniques are valuable tools that aid in the development of effective emission control strategies. The approach of source apportionment is commonly "receptor"-oriented, which means the identification and quantification of the sources of ambient particulate matter at a receptor is primarily based on the ambient concentration measurements at that receptor (Bhave et al. 2001).
Receptor based models have been used frequently in the past to understand sources of elemental carbon and organic carbon, which comprise carbonaceous aerosols
(Schauer et al. 2002; Lough et al. 2005). 33
Two types of receptor-based model are commonly used in source apportionment of organic aerosols: chemical mass balance (CMB) models using molecular markers, ECOC data, and a few crustal elements (Schauer et al. 1996;
Fraser et al. 2003), and positive matrix factorization (PMF) using measurements of
ECOC, secondary inorganic ions, and trace elements (Ramadan et al. 2000; Song et al.
2001; Kim and Hopke 2007). CMB models provide a least squares solution to a set of equations where the concentrations of chemical species at each receptor are expressed as a sum of products of source profiles and source contributions (Coulter 2004). A disadvantage with CMB source apportionment is the requirement of a priori knowledge of the source profiles (Jaeckels et al. 2007). PMF uses statistical correlations in time series of ambient data to isolate common factors representing sources, and the contribution of each to ambient PM levels. Unlike CMB, PMF doesn't require knowledge of source profiles as model inputs. This reduces the uncertainty introduced by potential differences between representative and actual source profiles (Jaeckels et al. 2007). Therefore, PMF can be very useful in estimating the source contributions at a receptor where source profiles are not readily available or are subject to uncertainty. For this reason, a PMF model was chosen for the source apportionment of ambient PM in this study. Although PMF does not require source profiles as model inputs, it does require knowledge of sources and molecular markers to determine the relationship of factors derived from the model to ambient
PM sources. 34
2.5.1 Sources and molecular markers
Table 2.1 lists the sources that have previously been identified in the source apportionment of ambient aerosols and molecular markers that are associated with specific sources. Among the molecular markers, most of them have already been established, yet some (mostly saccharides) are proposed as potential markers that are helpful in resolving sources that have not been quantitatively identified but expect broader applications. The following is a brief review of these saccharide-related sources.
Table 2.1 Sources and markers of ambient aerosols
Source Key Markers Reference Biomass burning Anhydrous saccharides (Simoneit et al. 1999; Nolte et Methoxyphenols al. 2001; Oros and Simoneit Diterpenoids 2001; 2001; Simoneit 2002) Triterpenoids Phytosterols
Motor vehicles Hopanes (pentacyclic triterpanes) (Rogge et al. 1993; Schauer Steranes etal. 1996; Schauer etal. Tricyclic terpanes 2002) Isoprenoids
Secondary aerosol Pinic acid (Pandis et al. 1992; Buzcu et Pinonic acid al. 2003; Kim and Hopke Sulfate 2007) Nitrate
Meat cooking Cholesterol (Rogge etal. 1991; Nolte et cis-9-octadecenoic acid al. 1999; Schauer etal. 1999) Nonanal 2-Decanone
Vegetative detritus n-Alkanes (C27-C33, odd) (Simoneit and Mazurek 1982; n-Alkanoic acids (C20-32) Standley and Simoneit 1987; 35
Roggeetal. 1993) Primary biological Monosaccharides (Graham et al. 2003; aerosol particles Disaccharides Jaenicke 2005; Elbert et al. (PBAPs) Saccharide polyols 2007; Bauer et al. 2008; Cellulose Deguillaume et al. 2008)
Soil and Aluminum (Chow et al. 1992; Simoneit associated biota Silicon et al. 2004; Chen et al. 2007; Monosaccharides Rogge et al. 2007) Disaccharides Saccharide polyols
Road dust Silicon (Hildemann et al. 1991; Aluminum Roggeetal. 1993; 1993; Magnesium 1993) n-alkanes (C25-C33, odd)
Marine aerosols Chloride (Ramadan et al. 2000; Buzcu Sodium et al. 2003)
Sources of aerosol saccharides in documentation include mainly three categories - biomass burning, atmospheric entrainment of soil and associated biota, and primary biological aerosol particles (PBAPs). In particular, biomass burning is responsible for the great amount of anhydrous saccharides, such as levoglucosan from cellulose decomposition, and to a lesser extent mannosan and galactosan from hemicellulose decomposition (Simoneit et al. 1999; Simoneit 2002), that are present in aerosols at numerous locations worldwide. Speculatively, sugar polyols may also be emitted by thermal stripping during combustion processes in wildfires (Simoneit et al. 2004). Soil together with its associated biota contains a complex mixture of biogenic detritus, including plant detritus, airborne microbes such as bacteria, viruses, spores of lichens and fungi, small algae, and protozoan cysts (Simoneit and Mazurek 36
1981; Simoneit et al. 2004). This material can become airborne when winds are strong enough and soil dry enough to be entrained in the atmosphere. Soil particles, especially the organic fraction with sizes between 30-100 urn diameter, may be resuspended when wind speeds exceed ca 0.2 m s"1 (Cox and Wathes 1995). A wide range of saccharide compounds have been detected in different types of soil samples, including monosaccharides, disaccharides, and sugar polyols. Simoneit et al. (2004) proposed that saccharides could be source specific tracers for resuspension of soils and unpaved road dust from agricultural activities into the air. Following that, Rogge et al. (2006; 2007) did a thorough analysis of the organic compound compositions of surface soils from 16 fields growing different crops and vegetations and found similar concentration profiles of the major saccharides including glucose, sucrose and trehalose for all types of cultivated surface soils. As a result, Rogge et al. (2006; 2007) suggested these species to be marker compounds for fugitive dust from biologically active surface soils. A number of other studies have confirmed the presence of saccharides in different types of soils and the potential for using saccharides to evaluate the source contribution of soil resuspension into the atmosphere (Rushdi et al.
2005; Medeiros et al. 2006; Rushdi et al. 2006; Al-Mutlaq et al. 2007). However, source specific analysis of the contribution of soil dust to ambient PM has been limited to date, and the use of saccharides as tracers has not been quantitatively tested.
Another group of aerosol saccharide sources is PBAPs, which has received increasing attention as studies show their composition and quantify the large emission flux worldwide. PBAPs comprise a series of biologically derived material such as spores, 37 pollens, fungi, algae, protozoa, bacteria, viruses, and fragments of plants and animals
(Caseiro et al. 2007; Winiwarter et al. 2009). They are injected directly into the atmosphere and range in size from tens of nanometers to millimeters (Cox and
Wathes 1995; Jaenicke 2005). PBAPs have been found to be highly enriched in sugar and sugar polyols. For example, pollen grains are known to contain sucrose, glucose, and fructose (Pacini 2000; Pacini et al. 2006); trehalose and sugar polyols are well documented as products of fungal metabolism (Lewis and Smith 1967; Bieleski 1982;
Elbert et al. 2007); and mannitol and arabitol have been proposed as tracers for airborne fungal spores (Bauer et al. 2008). In relation to ambient observation, high concentrations of aerosol sugars and sugar polyols have been linked with PBAPs in several studies conducted in the Amazon rainforest (Graham et al. 2003) and different locales in Europe (Pashynska et al. 2002; Carvalho et al. 2003; Yttri et al. 2007).
However, current knowledge about the contribution of PBAPs to aerosol mass, specifically PM10 mass, is lacking (Jaenicke 2005; Winiwarter et al. 2009).
2.5.2 PMF
Positive Matrix Factorization (PMF) is a statistical analysis source apportionment model, which has been widely used with aerosol elemental data and, to a more limited extent, with organic molecular markers. Originally developed by
Paatero (Paatero and Tapper 1994; Paatero 1997), PMF is a multivariate receptor modeling technique that can be used to analyze sets of ambient aerosol data to estimate the number of sources of particles, the chemical composition of each source, 38
and the amount that each source contributes to each sample, based on the
compositional data of aerosols and key representative markers for difference sources.
Using a least square approach, PMF determines the best solution to:
p X = GF + E or xiJ^YdfitSkJ+ev(i^l.:,n;j = l,...,m;k = l,...,p) k=\
where X(n xw) is the environmental data matrix consisting of the concentration
measurements of n chemical species in m samples, G («xp) is the source profile matrix of the n chemical species from/? sources, F(p*m) is the source contribution
matrix for each p sources for the m samples, and E (n *m) is the residuals matrix of
random errors, which is defined as
p eu = xu -Hf'kSkji' = h-,n;j = \,...,m\k = \,...,p) .
The objective of the PMF approach is to determine the optimal solution to matrices G
and F, by minimizing Q with a constraint that sources cannot have negative species
concentration (g;*>0) and samples cannot have a negative source contribution (fkj>0),
e gCE) = £I u 1=1 7=1 S'J
where s/, is the uncertainty estimate in they chemical species concentration for the /
samples.
The PMF model developed by EPA (PMF 1.1) was used in this study for data
analysis and interpretation. This model utilizes a conjugated gradient method to
iteratively arrive at a solution that optimizes Q (Brown and Hafner 2005) by allowing
the user to decide the uncertainty in each xy- (i.e., sy). The model weighting of a
species can be downgraded from "strong" to "weak" or "bad" according to the data 39 statistics of initial model applications. A species identified as weak will have the uncertainty increased by a factor of 3, and the species identified as bad will be removed from the analysis. EPA PMF1.1 operates in a robust mode, which prohibits outlying data from overly influencing the fitting of the contributions and profiles
(Eberly 2005).
There are several important criteria to ensure a good PMF solution. First of all, the modeled Q's should be within 50% of the theoretical value (Brown and Haftier
2005). The theoretical Q is calculated based on the following equation:
Qtheoreticai ~ (# samples x # good species) + [(# samples x # weak species)/3] -
(# samples * # factors being estimated)
Secondly, the optimum number of factors is determined by the criterion that each factor has distinctively dominant grouping of compounds (Shrivastava et al. 2007);
Thirdly, the model uncertainty produced by bootstrapping should be small (Brown and Hafner 2005). A good PMF solution is also characterized by the fact that the majority of the measured PM or OC can be accounted for by the modeled results.
Details of other statistical bases can be found elsewhere (Paatero and Tapper 1994;
Paatero 1997; Paatero and Hopke 2003; Eberly 2005). 40
Figure 2.1 Chemical structures of selected organic compounds analyzed in this dissertation work
PAHs:
Fluoranthene Pyrene Chrysene Benzo[a]anthracene Benzo[b]fluoranthene
Benzo[k]fluoranthene Benzo[e]pyrene Benzo[a]pyrene Perylene Indeno[l,2,3-cd]pyrene
Dibenz[a,h]anthracen Benzo[ghi]perylene
Hopanes:
17a(H),21 pCH)-Hopanes
Carboxylic Acids:
OH \ Y / OH OH ds-Pinie add cis-Pinonic add 41
Saccharides: -OH
OH OH °H V-^ m
Glucose Levoglucosan Sucrose
OH
OH OH
OH Jo OH OK OH OH OH OH
Trehalose Sorbitol Mannitol
.OH OH
HO OH OH OH OH
Arabitol Iso-erythritol Ribitol
OH OH OH
OH OH
Glycerol Galactitol Xylitol
HQ D
^CH D
D-glucose-1,2,3,4,5,6,6-d7 42
Chapter 3
Characterization of saccharides in fine particles collected at rural and urban sites in Texas
3.1 Sampling
Samples of PM2.5 were collected on pre-baked quartz-fiber filters (Whatman,
8"xl0") by high-volume air samplers (Thermo Andersen) fitted with a PM2.5 inlet
(High Volume Virtual Impactor; MSP). Sampling was conducted at three sites in
Texas between November 2005 and July 2006. Of the sites, San Augustine (31°32'N,
94°10'W) and Clarksville (33°37'N, 95°3'W) were located in rural areas of eastern
Texas, and the Dallas sampling site (32°49'N, 96°5PW) was located in the urban core of that city (Figure 3.1). Samples were taken at a flow rate of 1.13 m3 min"1 for a total of 24 hrs on each sampling day and used for analysis of saccharide compounds.
In parallel, PM was also collected on pre-weighed Teflon-filters (Whatman, 47mm) using low-volume air samplers (Thermo, Partisol-Plus, Model 2025). These low volume Teflon-filter samples were taken at a flow rate of 16.7 LPM for 24 hrs on each sampling day and used for gravimetric mass determination. After collection, quartz-fiber filter samples were returned from the field sites in glass jars with
Teflon-lined caps and Teflon-filters were transported in dedicated Petri-dishes. All samples were stored in freezers at -4°F until analysis. A total of 174 samples were obtained for organic speciation. During this sampling period, a series of wildfires 43
broke out in northern Texas and southern Oklahoma from November 2005 to April
2006, with a reported tens of thousands of wildfires and millions of acres of land and
forest burned in this area (NCDC 2007).
Figure 3.1 Map of Texas showing the three sampling sites at San Augustine, Clarksville and Dallas, and counties impacted by the Texas-Oklahoma wildfires of 2005-06.
i . * • * Clarksville • ~ • Dallas •'-&-; \'; • .;.': .. i . •• ">\
: ' . . .••'."•:•.-. t . t _.
*fr Sampling Site lllllll Co«m»s tmpsietetl itf Wiklfires • ... . -
3.2 Analytical methods
3.2.1 Element carbon (EC) and organic carbon (OC) analysis
The elemental and organic carbon content was determined for the ambient samples at the Sunset Laboratory, Oregon, using the Thermo-Optical Carbon
Analyzer. A 1 cm x 1.5 cm punch was obtained from the quartz-fiber filter corresponding to each sample, and the analysis was performed following the thermal-optical method described by Birch and Cary (1996). 44
3.2.2 Gravimetric analysis
The total mass of each sample was determined using a microbalance.
Teflon-filters from the ambient samples were weighed under controlled temperature
(22-24°C) and humidity conditions (45-55%).
3.2.3 Organic speciation
3.2.3.1 Extraction
One fourth of each quartz-fiber filter sample was extracted and analyzed for sugar composition following the method described by (Simoneit et al. 2004). This method has proved highly efficient for extracting sugar compounds (Medeiros and
Simoneit 2007). Prior to extraction, an isotopically labeled internal standard
(D-Glucose-l,2,3,4,5,6,6-d7) was spiked onto one fourth of each quartz-fiber filter.
After spiking, samples were extracted using three 30 ml aliquots of l:l(v/v) dichloromethane methanol for 10 min each under mild ultrasonic agitation. After extraction, the solvent extracts were combined, filtered, and concentrated using a rotary evaporator to approximately 2 ml, followed by concentration with a gentle stream of high-purity nitrogen to further reduce the volume to approximately 200 ul.
Aliquots of the total extracts were blown down to complete dryness using pure nitrogen gas and then derivatized with N,0-bis(trimethylsilyl)trifluoroacetamide
(BSTFA) containing 1% trimethylchlorosilane (TMCS) and pyridine for 3 hrs at 70
°C. The converted trimethylsilyl derivatives were analyzed by GC/MS within 24 hrs. 45
3.2.3.2 Analysis
A 1 ul aliquot of the derivatized extracts were analyzed by a HP model 6890 gas chromatography (GC) system coupled to a HP model 5973 inert mass selective detector (MSD). An aliquot of 1-phenyldodecane (1-PD) was co-injected to ensure proper injection as well as for compound quantification. Separation was achieved using a 30-m HP-5MS capillary column. The GC operating conditions were as follows: temperature hold at 65°C for 10 min, increase from 65 to 300°C at a rate of
10°C/min with a final isothermal hold at 300°C for 5 min. The sample was injected splitless with the injector temperature at 250°C. The MS was operated in the electron impact mode at 70 eV and scanned from 50 to 500 Dalton.
Individual saccharide compounds were identified by their unique retention time and specific mass fragmentation pattern. Characteristics of saccharides analyzed in this study (including internal standard) are listed in Table 3.1. Generally, the key ions for the trimethylsilyl (TMS) monosaccharides are 204 (corresponding to the pyrano-ring (5 C)), 191, and 217 (corresponding to the furano-ring (4 C)) (Medeiros and Simoneit 2007); TMS saccharide polyols have additional key ions at m/z 205; and the internal standard D-glucose-1,2,3,4,5,6,6-d7 (isomer of glucose) is characterized by ions at m/z 206 and 192 as TMS derivatives. TMS disaccharides were mainly recognized by a key ion at m/z 361 (Chizhov et al. 1967; DeJongh et al. 1969;
Simoneit et al. 2004). TMS derivatives of glucose and D-glucose-l,2,3,4,5,6,6-d7 presented two GC peaks due to the la- and 1 (3-configurations of the OH on the pyrano- or furano-ring (Medeiros and Simoneit 2007). For quantification purposes, 46
the two peaks were summed to report one value. Saccharide compounds from the
same group or with similar molecular structures have very similar mass spectra
fragmentation patterns (e.g., trehalose and sucrose, or different saccharide polyol
species). As a result, GC retention time was also incorporated for the correct
identification of sugar compounds (Simoneit 1999). The retention times were checked
repeatedly for all sugar compounds during this work to ensure instrument stability.
Compounds were quantified using selected ion peak area and converted to compound
mass using relative response factors (RRF) determined by GC/MS injection of
authentic standards spanning the range of concentrations in ambient samples and
1-PD under the same instrumental operating conditions.
Table 3.1. Characteristics of saccharide standards analyzed by GC-MS as TMS derivatives
Molecular Molecular Retention time Compounds formula mass CAS# (min; a-,6-) m/z a- + B-Glucose CeHi20e 180 492-62-6 25.56, 26.42 204, 191 Levoglucosan C6H10O5 162 498-07-7 23.50 204,217 Sorbitol C6H14O6 182 50-70-4 26.06 205,217 Mannitol C6H14O6 182 69-65-8 25.98 205,217 Arabitol C5H12O5 152 488-82-4 23.73 205,217 Ribitol C5H12O5 152 488-81-3 23.79 205,217 Iso-Erythritol C4H10O4 122 149-32-6 21.17 205,217 Glycerol CaHsOs 92 56-81-5 17.78 205,218 Trehalose C12H22O11 342 99-20-7 33.39 361 Sucrose C12H22O11 342 57-50-1 32.49 361 D-Glucose-1,2,3,4,5,6,6-d7 CeHsDrOe 187 23403-54-5 25.53, 26.37 206, 192
The detection limit (DL) of saccharides by GC/MS was calculated as the
concentration that corresponds to three times the standard deviation of the peak areas 47
generated by injections (n=4) of approximately 2 ng ml"1 of the representative saccharide standards (Medeiros and Simoneit 2007). Glucose, levoglucosan, sucrose and mannitol were chosen as representatives of the monosaccharide, anhydrosaccharide, disaccharide and saccharide polyol groups, respectively, since these compounds were commonly observed in environmental samples.
3.3 Results
3.3.1 Detection limit
Detection limit (DL) for saccharides by GC/MS are given in Table 3.2. The
DLs varied from 120 to 550 ng ml"1 (or from 120 to 550 pg/injection), which are comparable to those reported by Medeiros and Simoneit (2007) (from 130 to 360 ng ml"1). Saccharide concentrations detected in a majority of the aerosol and soil samples in this study were well above these detection limits, assuring the confidence of the assembled data set.
Table 3.2 GC/MS detection limits for representative saccharide standards
Compound Group DL (ng ml"1) Glucose Monosaccharides 540 Levoglucosan Anhydrous Saccharides 550 Sucrose Disaccharides 120 Mannitol Saccharide Polyols 470 48
3.3.2 PM2.5, organic carbon (OC), elemental carbon (EC), and total measured saccharides
Table 3.3 summarizes the ambient concentrations of PM2.5, OC, EC, and total measured saccharides. The average PM2.5 concentrations at the two rural sites (San
Augustine and Clarksville) were very similar (10.1 vs. 10.2 ug m"3), and a slighter higher value (11.2 ug m"3) was measured at the urban site Dallas.
Table 3.3 PM2.5, OC, EC and total saccharide ambient concentration statistics San Augustine Clarksville Dallas PM2.5 (ug nf3) 10.1(2.5-23.2) 10.2(2.2-22.3) 11.2(3.5-34.0)
OC(ugCnT3) 2.7(1.2-9.7) 3.5(1.4-10.2) 3.4(2.0-6.9)
EC(ugCm"3) 0.18(0.06-0.65) 0.25(0.09-0.63) 0.58(0.18-1.46)
OC/TC(%) 93.8(91.1-95.7) 93.3(90.7-96.5) 85.7(75.6-92.4)
Total measured saccharides 70.4(15.4-355) 128(7.5-372) 52.4(15.8-196) (ng m"3) Carbon content of total 1.1(0.4-2.1) 1.8(0.1-4.2) 0.8(0.3-1.6) measured saccharides (ngC m"3)/OC (ngC m"3) (%) Total measured saccharides 1.0 (0.1 -4.1) 1.6 (0.1 -6.0) 0.6 (0.1 -1.6) (ng m 3)/PM2.5 (ng m3) (%)
The average OC and EC concentrations in PM2.5 were 2.7 u.gC m"3 and 0.2
(igC m"3 for San Augustine, 3.4 ugC m"3 and 0.2 ngC m"3 for Clarksville, and 3.4 ugC m"3 and 0.6 |igC m"3 for Dallas. On an average, the measured OC and EC in PM2.5 accounted for 34% and 2% for San Augustine, 39% and 3% for Clarksville, and 34% and 6% of PM2.5 mass for Dallas. The higher level of EC at Dallas was expected, as urban areas are more impacted by motor vehicle emissions, especially diesel 49 powered vehicles, which are the main source of elemental carbon (Hildemann et al.
1991).
The carbon content of total measured saccharides (ngC m"3) to OC (ugC m"3) ratio in PM2.5 was higher at the two rural sites (1.1% on average for San Augustine and 1.8% on average for Clarksville) than the urban site (0.8% on average for Dallas), which suggests biologically related sources are a greater contributor to the fine aerosol OC content in rural locations. This ratio was in general higher from February to April (Figure 3.2), which could be a result of mixing effects of both biomass burning and the decomposition and reproduction of the living biomass during this period (Simoneit et al. 2004). A detailed explanation of the seasonal variations of aerosol saccharide contents is provided in the next section
Figure 3.2 Comparison of the monthly ratio of the carbon content of total measured saccharides to OC in PM2.5 at the three sampling sites in Texas.
Jan Feb Mar Apri May Sampling Period (2006) 50
3.3.3 Aerosol saccharides and seasonal variation
Individual concentrations of measured saccharide species in aerosol samples taken during two sampling periods (November 2005 - December 2005; January 2006
- July 2006) are listed in Table 3.4. Each concentration was averaged over an approximately 2-week sampling period.
Concentrations of total measured aerosol saccharides ranged from 8 to 372 ng m"3 with an average of 128 ng m"3 at Clarksville, from 15 to 355 ng m"3 with an average of 70 ng m" at San Augustine, and from 16 to 196 ng m" with an average of
52 at Dallas. Higher concentrations were observed from January to April for the
Clarksville and Dallas sites, and to a lesser extent, the San Augustine site. This period parallels the carbohydrate production in the ecosystem, suggesting biologically derived carbon as a significant source of the measured saccharides in collected fine
PM. The concurrence of major wildfires in northern Texas during this period also indicates the impact of wood combustion on aerosol saccharide concentrations.
Levoglucosan, glucose, mannitol, arabitol and glycerol were dominant saccharide species measured at both rural and urban sites. The prevalence of those compounds has been reported in urban and rural aerosols elsewhere (Graham et al.
2002; Carvalho et al. 2003; Simoneit et al. 2004; Medeiros et al. 2006; Wan and Yu
2006; 2007; Yttri et al. 2007; Pio et al. 2008). Saccharide compositions in the aerosol samples also displayed a seasonal variation, reflecting the sugar production and utilization in the ecosystem (Medeiros et al. 2006). 51
Levoglucosan was the most abundant sugar at all three sites and represented
between 14% and 82% of the total saccharide mass measured. Concentrations of
levoglucosan were highest in winter and early spring (Figure 3.3 (a)), and it
contributed 47- 82% to the total saccharide mass during this period. This percentage
dropped to less than 30% in May and further decreased to less than 20% in July. The
highest recorded levoglucosan concentrations were found in February and March, 111
ng m"3 in Clarksville and 80 ng m"3 in San Augustine and Dallas. Levoglucosan is a
reliable indicator of biomass combustion (Simoneit et al. 1999), and concentrations of
levoglucosan in these samples reflect local biomass burning, as the elevated
levoglucosan levels were measured concurrently with the Texas-Oklahoma wildfires
and back trajectories (Figure 3.8) also suggested the aerosol parcels containing the
highest levoglucosan levels originated from the Texas-Oklahoma wildfire regions.
Glucose was the second most abundant sugar, contributing between 5% and
39% of measured sugar mass. As the growing season proceeded, concentrations of
glucose in Clarksville increased from 10 ng m"3 in January to a maximum of 43 ng m"3
in May. After that, a strong decrease followed, with a much lower concentration of 12 ng m"3 measured in July (Figure 3.3 (b)). The same trend was observed at the San
Augustine and the Dallas site, although not as strong. These changes seem to reflect the saccharide production in the ecosystem, with major synthesis of primary sugars early in the growing season (Medeiros et al. 2006).
Sucrose was significant only in the spring (March and April) (Figure 3.3 (c)).
It is the predominant sugar in the phloem of plants and is particularly important in 52
developing flower buds (Bieleski 1995). The strong decrease of sucrose amounts after
April was likely due to the translocation of saccharides to other parts of the plants for
supporting growth after spring (van Doom 2004).
Trehalose, as an important fungal metabolite (Martin et al. 1988; Sillje et al.
1999), had much higher concentrations in summer than in winter and spring (Figure
3.3 (d)), likely due to the increased contribution from fungal derived sources as a result of external stressors of heat and drought during summer months (Eleutherio et al. 1993). This trend was more pronounced for San Augustine and Clarksville compared to Dallas due to rural and urban site differences.
Similar to trehalose, sugar polyols also had enhanced concentrations in June and July (Figure 3.4). Sugar polyols, mainly mannitol and arabitol, are major fungal metabolites in many green-algal lichens (Dahlman et al., 2003) as well as important constituents of bacteria, fungi, and lower plants (Bieleski, 1982). The prevalence of sugar polyols in summer indicates an increase in contributions from microbially degraded materials during the period of leaf decay (Medeiros et al. 2006).
The only sugar polyol that didn't seem to follow this trend was glycerol. One potential explanation is that the presence of glycerol in aerosols originates from sources other than microbial input. Graham et al.(2002) reported a high correlation of aerosol glycerol with black carbon (BC), organic carbon (OC) and potassium (K) at representative pasture and primary rainforest sites in Rond6nia, Brazil, suggesting biomass combustion may influence atmospheric glycerol concentrations. 53
The seasonal variation of major aerosol saccharides (except levoglucosan) measured in our study was in general agreeable to values reported in Maine (USA)
(Medeiros et al. 2006), Vienna (Austria) (Caseiro et al. 2007), and Hong Kong (Wan and Yu 2007), reflecting a similar impact from major synthesis of primary saccharides during the growing season in very different ecosystems. 54
Table 3.4 Sugar Concentrations in aerosol samples collected at three sites in Texas (San Augustine, Clarksville and Dallas) from November, 2005 to July, 2006. Each concentration (ng/m3) was averaged over a half-a-month sampling period.
Sampling Period (2005) Sampling Period (2006)
1 Nov- 16 Nov- 1 Dec- 16 Dec- 1 Jan - 16 Jan - 1 Feb- 16 Feb- 1 Mar- 16 Mar 1 Apr- 16 Apr- 1 May- 16 May - 1 Jun- 16Jun - 1 Jul - 15 Jul-
Compounds 15 Nov 30 Nov 15 Dec 31 Dec 15 Jan 31 Jan 15 Feb 28 Feb 15 Mar -31 Mar 15 Apr 30 Apr 15 May 31 May 15 Jun 30 Jun 15 Jul 31 Jul
San Augustine a- + |3-Glucose 3.58 4.36 3.53 3.42 1.80 4.75 5.50 4.94 15.9 14.8 8.48 15.2 13.8 4.74 18.2 15.9 20.4 16.1
Levoglucosan 30.9 67.5 76.8 49.7 14.7 48.3 47.2 40.3 78.8 45.6 29.4 17.5 31.0 14.2 20.2 11.7 18.5 9.29
Sorbitol 0.24 1.15 0.44 0.42 0.12 0.25 0.59 0.36 0.89 0.26 0.22 0.21 0.38 0.17 0.97 0.85 1.53 0.88
Mannitol 2.83 3.08 1.25 1.26 1.07 1.00 1.15 0.75 2.22 1.12 1.14 3.53 5.71 1.51 4.60 9.79 15.9 10.5
Arabitol 1.95 6.43 0.92 2.19 0.80 0.90 1.37 1.02 2.42 1.28 0.92 2.91 4.05 1.36 3.34 5.76 9.51 7.10
Ribitol 0.15 3.14 0.21 0.77 0.07 0.08 0.24 0.17 0.89 0.25 0.11 0.05 0.14 0.27 0.25 0.28 0.42 0.27
Iso-Erythritol 0.27 0.53 0.24 0.44 0.16 0.22 0.31 0.16 0.69 0.36 0.18 0.39 0.63 0.56 0.99 0.81 1.17 1.58
Glycerol 7.35 5.93 4.53 6.50 3.22 3.66 4.44 5.91 7.64 5.71 3.83 6.30 8.68 7.26 7.01 3.66 4.50 3.39
Trehalose 0.61 1.00 0.70 0.53 0.37 0.46 0.35 0.47 0.69 0.34 0.32 0.49 1.12 0.33 1.34 2.41 3.24 1.44
Sucrose 0.42 1.81 4.28 2.75 0.53 2.40 3.03 1.28 54.2 19.8 7.34 0.82 0.60 0.36 0.50 0.75 0.53 0.34
Total 48.3 94.9 92.9 68.0 22.9 62.0 64.1 55.4 164 89.5 51.9 47.4 66.1 30.8 57.4 51.8 75.7 50.9
Clarksville a- + p-Glucose NA NA NA NA 10.6 19.8 12.1 20.8 25.8 26.6 45.2 51.5 19.1 24.4 24.1 12.0 13.2 NA
Levoglucosan NA NA NA NA 61.9 94.2 50.0 112 62.8 74.5 61.0 44.9 22.2 23.0 34.8 14.1 6.57 NA
Sorbitol NA NA NA NA 0.27 0.40 0.82 0.45 0.60 0.48 1.32 1.33 0.65 0.81 1.76 0.82 1.03 NA Mannitol NA NA NA NA 5.46 2.74 2.98 4.09 2.35 7.68 18.9 16.1 10.5 20.3 22.7 8.42 9.59 NA
Arabitol NA NA NA NA 4.20 2.51 6.60 3.35 3.38 4.07 12.2 9.40 5.59 7.69 11.9 9.01 6.15 NA Ribitol NA NA NA NA 0.17 0.52 3.46 0.75 1.18 0.28 0.47 0.36 0.28 0.34 1.15 3.35 0.60 NA o » 5 > 2 9 D + §= (D -< 9 a T 8! o S •S. 1 iH4 — o Sfi ^ CD Z2.
zzzzzzzzzzz > > > > > >>>>>>>>>>> z z z z z ZZZZ2ZZ2ZZ > > > > > >>>>>>>>>>
ZZZZZZZZZZZ 2 Z z z z >>>>>>>>>>> > > > > >
z z z Z Z Z Z Z Z Z > > > z z z z z >>>>>>> > > > > >
-» o -» p e -"• CO Cn 00 Cn 00 00
-* en ro co
-* o Cn o o -A -I o -» M -» _o ^ en bo KJ 00 en '.£*
O) O cn ^
o CO
00 CO (Jl 00
M _i o en ro oo en -* oo en ro cn -b. -* oo o -» bo '-g 1> !fc cn cn ro en ro en lo 0 Cn 0> CO b en en en to -» cn cn §s 2 Z z z z z z 2 Z > > Z Z Z Z > > > > > > > > > > > 56 Figure 3.3. Monthly individual concentrations of major saccharides: (a) levoglucosan, (b) glucose, (c) sucrose and (d) trehalose, in aerosol samples collected at three sites in Texas from November 2005 to July 2006. (a) Levoglucosan • San Augustine • Clarksville • Dallas Nov Dec Jan Feb Mar Apri May Jun Jul Sampling period (Nov 2005 - Jul 2006) (b) Glucose 50 • San Augustine • Clarksville 40 • Dallas Nov Dec Jan Feb Mar Apri May Jun Jul Sampling period (Nov 2005 - Jul 2006) 57 (c) Sucrose 45 40 • San Augustine • Clarksville 35 • Dallas 30 125 O) 20 Nov Dec Jan Feb Mar Apri May Jun Jul Sampling period (Nov 2005 - Jul 2006) (d) Trehalose 7 H San Augustine 6 ~ M Clarksville 5 _ • Dallas g> 3 2 1 0 1 1 rv- Nov Dec Jan Feb Mar Apri May Jun Jul Sampling period (Nov 2005 - Jul 2006) 58 Figure 3.4 Seasonal variation of sugar polyols at two rural sites - (a) San Augustine, (b) Clarksville, and one urban site - (c) Dallas, from November 2005 to July 2006. (a) —•— -Sorbitol —m- - Mannitol -Arabitol --X- - - Ribitol -- •- • Glycerol Nov Dec Jan Feb Mar Apri May Jun Jul Sampling period (Nov 2005 - Jul 2006) —•— -Sorbitol —m- -Mannitol -Arabitol ---X- - Ribitol Iso-Erythritol -•*- - Glycerol Sampling period (Jan - Jul 2006) 59 —•— Sorbitol —•— Mannitol —a—Arabitol - - X- - • Ribitol —©— Iso-Erythritol --#••• Glycerol Jan Feb Mar Apri May Jun Jul Sampling period (Jan - Jul 2006) 3.3.4 Correlation Analysis Pearson's correlation test was performed for the dataset containing ambient concentrations of 10 saccharide compounds in aerosol samples collected at the three sampling sites, and results are summarized in Table 3.5. Concentrations of trehalose and sugar polyols, particularly mannitol and arabitol, were well correlated (r= 0.4-0.8) in PM samples at the three sites, serving as statistical support for a common source for these compounds (Lewis and Smith 1967; Bieleski 1982). Glucose was correlated with most sugars and polyols in varying degrees. Since glucose is a fundamental simple sugar in biology, all other sugars and polyols can be considered derivatives in the sense that glucose serves as a primary carbon source for plants and microorganisms (Pigman and Horton 1980). These correlations, between trehalose and the sugar polyols and glucose and most other saccharides has been reported for 60 both aerosol and soil samples (Medeiros et al. 2006; Wan and Yu 2006). No consistent correlation was observed between levoglucosan and the other measured saccharides except for fine particle glycerol, which is consistent with our earlier interpretation that aerosol glycerol may be linked to biomass combustion sources. Furthermore, all correlations shown in this study were stronger for aerosol samples from Clarksville and San Augustine (rural sites) and weaker for samples from the Dallas site (urban), meaning local biogenic sources have less influence on saccharides in aerosols at the urban site than at the rural site. Table 3.5 Correlation coefficients among 10 saccharide compounds in aerosols collected at the three Texas sites, level of significance a=0.05 (two-tailed test) Levoglu- Iso- Trehalose Glucose cosan Sucrose Sorbitol Mannitol Arabitol Ribitol Erythritol Glycerol San Augustine Trehalose 1 Glucose 0.43 1 Levoglucosan -0.08 -0.02 Sucrose -0.01 0.35 0.51 1 Sorbitol 0.63 0.43 0.03 0.06 1 Mannitol 0.82 0.59 -0.24 -0.12 0.67 1 Arabitol 0.71 0.53 0.07 0.09 0.75 0.88 1 Ribitol 0.06 -0.10 0.35 0.11 0.34 -0.01 0.38 1 Iso-Erythritol 0.35 0.46 -0.02 0.03 0.39 0.57 0.66 0.08 1 Glycerol -0.06 0.12 0.29 0.11 0.08 -0.04 0.04 0.08 0.17 Clarksville Trehalose 1 Glucose 0.51 1 Levoglucosan -0.06 0.24 Sucrose -0.08 0.38 0.38 1 Sorbitol 0.64 0.57 0.04 0.02 1 Mannitol 0.78 0.63 -0.03 0.06 0.78 1 Arabitol 0.64 0.59 0.03 0.05 0.90 0.85 1 Ribitol -0.01 -0.28 -0.06 -0.11 0.21 -0.18 0.24 1 Iso-Erythritol 0.60 0.51 0.11 -0.07 0.63 0.71 0.71 -0.14 1 Glycerol -0.02 0.28 0.26 0.00 0.14 0.06 0.23 0.05 0.19 61 Dallas Trehalose 1 Glucose 0.60 1 Levoglucosan 0.07 0.20 1 Sucrose 0.22 0.50 0.48 1 Sorbitol 0.58 0.63 0.00 0.24 1 Mannitol 0.35 0.35 -0.31 -0.03 0.50 1 Arabitol 0.49 0.57 -0.21 -0.07 0.66 0.64 1 Ribitol 0.00 0.08 0.33 0.12 0.02 -0.20 -0.09 1 Iso-Erythritol 0.08 0.21 0.34 0.15 0.20 0.00 0.26 0.27 1 Glycerol 0.04 0.12 0.53 0.51 0.15 -0.23 -0.11 0.31 0.25 3.3.5 Rural and urban comparison Total measured saccharide concentrations in aerosol samples were, in general, higher at the two rural sites than at the urban sampling location. The total measured aerosol saccharide concentration averaged 68 ng m" at the San Augustine site and 117 ng m"3 at the Clarksville site, compared to 55 ng m"3 at the Dallas site. The total measured saccharides also constituted higher percentages of PM2.5 mass and OC at the two rural sites than at the urban site (Table 3.3). The relative composition of aerosol saccharides collected at the three sites was similar, and ambient saccharide concentrations followed a similar seasonal trend at all three sites, although this trend was much weaker at the urban site (Figure 3.3). As an urban site, Dallas can be expected to have a smaller contribution of primary biogenic carbon sources, and this is corroborated by the weaker seasonal trend of aerosol saccharide compositions of glucose, sucrose, trehalose and sugar polyols, all of which are potential tracers for PBAPs and other biologically derived carbohydrate sources such as resuspended soil dust and associated biota (Graham et al. 2003; Simoneit et al. 2004; Caseiro et al. 62 2007) . The only exception to this dampened seasonal trend at Dallas was levoglucosan, which, as a marker for wood smoke, indicates an impact from wildfires and residential wood smoke in both urban and rural locations. The higher levels of total saccharides measured at the Clarksville site compared to those measured at the San Augustine site may reflect a different degree of impact from primary biogenic sources between the two areas, and the influence of the proximity of Clarksville to the regional wildfires in North Texas. 3.3.6 Source apportionment using PMF Positive Matrix Factorization (PMF) is a statistical source apportionment model which has been widely used with elemental data and to a more limited extent, on organic molecular markers. A recent application of PMF modeling using organic molecular markers was based on apportioning ambient PM to motor vehicle emission, secondary organic aerosol, wood combustion, and two point sources (Jaeckels et al. 2007). However, no report is currently available characterizing the contribution to ambient PM from the primary biogenic carbon sources based on carbohydrate molecular markers. Given the documented roles of saccharides as tracers for biomass burning, PBAPs, and other biologically derived carbohydrate sources, we propose here to use saccharide compounds as markers in PMF modeling to trace their related sources. By adding saccharides to the existing molecular marker pool, the local emission inventory of PM2.5 can be further verified. 63 The PMF model developed by EPA (PMF1.1) was used for the modeling of PM2.5 samples collected at Dallas and San Augustine (PMF was not applied to Clarksville samples due to lack of data for other molecular markers), and a detailed description of the model can be found elsewhere (Brown and Hafner 2005).. In this application, 24 particulate molecular markers were selected, including EC and OC, 7 normal alkanes, 2 hopanes, 4 acids, one anion (sulfate), and 8 saccharide compounds (Figure 3.5). Concentrations of saccharides were obtained from this study. Concentration of sulfate was obtained by parallel sampling conducted by the Texas Commission on Environmental Quality, and other organic markers in the same samples were obtained from Bhat (2007). Data from Dallas included 25 sample days and San Augustine 37 sample days. Similar source profiles were resolved by PMF for both the Dallas and San Augustine datasets, and the isolated profiles for Dallas are shown in Figure 3.5. The first factor isolated was dominated by two petroleum biomarkers (17a(H), 2ip(H)-hopane and 17a(H), 2ip(H)-29-norhopane), which are characteristic of motor vehicle emissions (Schauer et al. 2002). Other sources such as diesel trains, ships, and lawn mowers could also contribute to this factor, thus this factor is labeled as petroleum-powered engine sources. The second factor was enriched in hexadecanoic acid (C16:0), octadecanoic acid (C18:0) and octadecenoic acid (C18:l), which has a chemical composition consistent with meat cooking (Rogge et al. 1991). A third factor was attributed to a-pinene derived secondary organic aerosol, as the dominant species in this source profile (pinic acid) is an oxidation product of a-pinene (Pandis 64 et al. 1992). The fourth factor, secondary sulfate, is identified by high sulfate concentration, and the fifth factor includes a series of alkanes (C25 to C31). Sources of these alkanes include both vegetative waxes and petroleum hydrocarbons (Simoneit 1984; Rogge et al. 1993). This source profile does not display enrichment in odd carbon number alkanes, indicating the alkanes were predominately from fossil sources (Standley and Simoneit 1987) as opposed to biogenic sources. In addition to the above five factors that have commonly been identified in PMF or other receptor modeling, three additional source factors were also successfully isolated based on using saccharides as molecular markers. The sixth factor was dominated by levoglucosan therefore labeled as a wood smoke factor. The sub-dominance of glycerol in this factor is consistent with the correlation analysis, suggesting glycerol may originate from biomass combustion. The seventh and eighth factor are both enriched in sugar and sugar polyols, with one factor dominated by sucrose and to a lesser extent trehalose and glucose, and another factor dominated by mannitol and arabitol. While it is clear these two factors are of primary biologically derived aerosol origin, they could be from mixed effects. Sucrose and glucose are known major constituents of pollen (Pacini et al. 2006) and are also found in some fungi and spores (Feofilova et al. 2000; Tereshina et al. 2000). Trehalose, mannitol and arabitol are widely present in bacteria, fungi, lower plants, and invertebrates (Bieleski 1982), serving as storage or transport carbohydrates and cell protectants against external stresses (Eleutherio et al. 1993; Eleutherio et al. 1993; Elbert et al. 2007). In particular, mannitol and arabitol are documented as tracers for fungal spores 65 » (Bauer et al. 2008). Given the dominance of different saccharides in the source profile, these two factors are labeled as sucrose rich bio-aerosol and mannitol and arabitol rich bio-aerosol. However, it is worth noticing that they could be mixed effects of PBAPs and other biologically derived carbohydrate sources such as soil, dust, and associated biota. The successful separation of the wood smoke, sucrose rich bio-aerosol, and mannitol and arabitol rich bio-aerosol factors from other source factors by PMF using saccharide compounds suggests receptor modeling can be a useful tool for the source apportionment of PM2.5 from the primary biologically derived carbon sources, and sugar and sugar polyols can serve as molecular markers for these source categories. PMF results can be further utilized to calculate the relative source contributions for the total PM2.5 mass using a multiple linear regression between the isolated factor strengths and measured PM2.5 mass as shown in Figure 3.6 (Song et al. 2001). Contributions from the first and fifth factors were combined as they both characterize petroleum based emissions and are labeled as mobile sources. Similarly, contributions from the seventh and eighth factors were combined together into a single "primary biologically derived carbon sources" factor. Mobile sources and secondary sulfate were the largest contributors to PM2.5 at both Dallas and San Augustine sites (around 31% and 25%). The fact that mobile sources contributed similar fractions to PM2.5 mass at both sites was not expected. However, the primary difference between the two sites lies in the contributions from other source categories, including wood smoke (22% at San Augustine and 16% at Dallas) and primary 66 biologically derived carbon sources (14% at San Augustine and 5% at Dallas). This is consistent with the fact that San Augustine is a rural site which was more impacted by primary biogenic sources. Similarly, the meat cooking factor contributed a greater fraction to PM2.5 at Dallas (13%, compared with 3% at San Augustine) due to its urban location. Biogenic secondary organic aerosol formation had a slightly larger contribution at Dallas (9%) than at San Augustine (4%). It should be noted that this factor likely did not include the contribution of secondary organic aerosol formation from anthropogenic precursors. Also note that this relative source contribution analysis is based solely on factors identified by PMF, which is dependent on the markers used in the analysis. There may be other unidentified sources, but the comparison of measured to apportioned mass for the two sites (Figure 3.7) suggests that the unidentified sources would be small and that most of the sources having significant contributions to PM2.5 mass have been isolated by the model. This analysis is the first attempt to use saccharides as molecular markers in PMF modeling for the contribution to PM2.5 from PBAPs and other biologically derived carbohydrate sources such as soil, dust and associated biota. Due to the limited size of input data sets, the modeling results of this analysis may be subject to bias or variability, but the model has successfully resolved eight sources that provide meaningful physical interpretation. This approach can serve as a baseline for the further expansion of saccharides as tracers for primary biologically derived carbon sources in receptor modeling using organic molecular markers. 67 Figure 3.5 Source profiles for Dallas samples (Jan - May 2006). (OC: organic carbon; EC: elemental carbon; C25-C31: normal alkane series; HOP1: 17a(H), 2ip(H)-hopane; HOP2: 17a(H), 2ip(H)-29-norhopane; C160: hexadecanoic acid; C180: octadecanoic acid; C181: octadecenoic acid; PINIC: pinic acid; TRE: trehalose; GLUC: glucose; LEV: levoglucosan; SUC: sucrose; MANN: mannitol; ARAB: arabitol; ERY: erythritol; GLY: glycerol; SUL: sulfate) Dallas - Petroleum-powered Engine Sources Dallas - Meat Cooking ».,l,i,i, a, •,-,-, inss o o u aI. e Dallas - Biogenic Secondary Organic Aerosol Dallas - Secondary Sulfate O 50 -I t O 40 » lil Dallas - Petroleum Waxes Dallas - Wood Combustion u i — Jl O Ui U > 5O z < a: -i 3 J w < a ui o v> u o o o o o o i 2X 2i 7o7 o o 5IT p Dallas -Sucrose Rich Bio-aerosol Dallas - Mannitol and Arabitol Rich Bio-aerosol fact o one d t o | , T j ! T appor t ! o f specie s 1 .iiiil iil I 1 l < n i n o > y z o; =I ui 3 z < u o o o u u o o = s o o o o (Sources of model input: Concentrations of saccharides were obtained from this study. Concentration of sulfate was obtained by parallel sampling conducted by the Texas 68 Commission on Environmental Quality, and other organic markers were obtained from Bhat (2007).) Figure 3.6 Relative source contributions to PM2.5 at (a) Dallas and (b) San Augustine site ("motor vehicles" and "petroleum waxes" were combined into "mobile sources"; "sucrose rich bio-aerosol" and "mannitol and arabitol rich bio-aerosol" were combined into "primary biologically derived carbon sources"). (a) Wood Smoke Bioge™ S0A 16% °% Secondary - Sulfate ^/^JpV- 25% Primary Mobile Sources / \C-»i ^IKC Biologically 32% ^<2_^^ ^\^ Derived Carbon Sources Meat Cooking 5% 13% (b) Biogenic SOA 4% Wood Smoke Secondary Sulfate 22% 26% Primary Biologically Derived Carbon Sources Mobile Sources 13% 31% Meat Cooking 4% 69 Figure 3.7 Apportioned PM2.5 mass concentrations for Dallas and San Augustine plotted against the measured values II o o 10 20 30 Apportioned Mass Concentration (pg m3) 10 20 30 Apportioned Mass Concentration (pg m"3) 3.4 Acknowledgement This project was funded by the EPA STAR Program. I would like to thank the Sunset Lab in Oregon for analyzing EC and OC, the Texas Commission on Environmental Quality (TCEQ) for providing the aerosol sulfate concentrations that were used in the PMF model. I also would like to thank my previous colleague Shagun Bhat for shipping and handling of all quartz filter samples in this project, and for giving me the permission to include the non-saccharide organic marker data in the PMF model. 70 Figure 3.8 Back trajectories of PM2.5 samples with the two highest levoglucosan concentrations at the three Texas sampling sites - (a) San Augustine, (b) Clarksville, and (c) Dallas. (a) San Augustine NOAAHYSPLIT MODEL NOAA HYSPLIT MODEL Backward trajectories ending at 1800 UTC 13 Mar 06 Backward trajectories ending at 1800 UTC 22 Nov 05 EDAS Meteorological Data EDAS Meteorological Data i - -4 ; i> " ,, / ' i "- - i ! i '. X "--••J \ * ' > . V 100 r ____—r~*^*~ ° 500 it ~ ^~^J^^**~^ • •• 500 12 06 oo is n;22 Job ID: 352312 Job Start: Wed Jul 15 02:05:32 GMT 2009 Source 1 lat; 31.53 Ion.:-94.02 heights: 100, 500 m AGL Trajectory Directs: Backward Duration 24 hrs Meteo Data: EDAS40 Traiectcry Direction: Backward Duration: 24 hrs Mateo Data' EDAS40 Vertical Motion Catenation Method: Mode) Vertical Velocity Vertical Motion Calculation Method: Model Vertical Velocity Produced with HYSPLIT from the NOAA ARL Website fhtrp:.?wiww.arl noaa qcwready/j Produced with HYSPLIT from the NOAA ARL Website fhttptfwww arl.noaa^w'ready,'! (b) Clarksville NOAA HYSPLIT MODEL NOAA HYSPLIT MODEL Backward trajectories ending at 1800 UTC 13 Mar 06 Backward trajectories ending at 1800 UTC 08 Feb 06 EDAS Meteorological Data EDAS Meteorological Data Job ID.352627 Source 1 lat: 33.62 Ion.: -95.05 heights. 100, 500 m AGL Trajectory Direction: Backward Duration' 24 hrs Meteo Data EDAS40 i 49 hrs Meteo Data. EDAS40 Vertical Motion Calculation Method: Model Vertical Velocity al Vertical Velccily Produced with HYSPLIT from the NOAA ARL Website (http:ywww art, noaa^ov/ready/j Produced with HYSPLIT from the NOAA ARL Website {http^wwVK.arl.naaaqwJ'eady/) 71 (c) Dallas NOAA HYSPLIT MODEL NOAA HYSPLIT MODEL Backward trajectories ending at 1800 UTC 13 Mar 06 Backward trajectories ending at 1800 UTC 25 Mar 06 EDAS Meteorological Data EDAS Meteorological Data 2000 1500 1000 500 J Produced with HYSPLIT from the NOAA ARL Website jhttp'/wwarl roaa.gpv/read^ 72 Chapter 4 Characterization of Organic Fine Particulate Matter in Big Bend National Park and Source Apportionment using PMF 4.1 Introduction Big Bend National Park (BBNP) in West Texas, located on the border of the United States and Mexico, experiences degraded visibility despite its remote location. Particulate matter (PM) and PM precursors from emission sources on both sides of the border contribute to this problem (Pitchford 2004). Several studies have been launched by the U.S. EPA and the National Park Service to investigate the particulate problem in Big Bend, particularly the cause of haze and degraded visibility (Schichtel et al. 2006). The sources contributing to PM levels in Big Bend were found to be both local and regional (Brown et al. 2002). A set of source profile was constructed later by Chow et al. (2004) to characterize the industrial, mobile and area sources in the Big Bend region. For the most part, elements and ions were included in these source profiles, and specific organic compounds (e.g., PAHs, hopanes, steranes, organic acids, alcohols, saccharides, etc.) were shown to be promising in differentiating among source types and subtypes (Chow et al. 2004). These previous studies motivate a more detailed investigation to further determine the origin of fine 73 particles in the Big Bend region and suggest the use of source apportionment using ' organic molecular markers to better isolate important contributions. To better understand the various particulate sources in this region, a high volume particulate matter air sampler was deployed in BBNP to collect samples of carbonaceous particulate matter during the summer of 2007. The goals of this study were to characterize the molecular composition of organic fine particulate matter present in BBNP and to use organic markers to identify sources contributing to PM levels at this location. To accomplish these goals, the ambient concentrations of various organic species in aerosol were determined through chemical analysis, and different contributing sources were isolated using molecular markers through positive matrix factorization (PMF) modeling. 4.2 Experimental Methods 4.2.1 Sample Collection PM2.5 samples were collected on pre-baked quartz-fiber filters (Whatman, 20.3 x 25.4 cm) using a high-volume particulate matter sampler (Thermo Andersen) fitted with a PM2.5 inlet (High Volume Virtual Impactor, MSP). Continuous sampling was conducted every sixth day from Jan 24 to Feb 23, 2007, and every second day from May 15 to July 2, 2007 in BBNP. High-volume samplers were operated at a flow rate of 1.13 m3 min"1 for a total of 24 h on each sampling day. After 74 collection, samples were returned from the field sites in glass jars with Teflop-lined caps and stored in a freezer at -4°F until extraction. 4.2.2 Carbon Analysis Total fine PM levels of organic and elemental carbon were measured at Sunset Laboratory by the thermal-optical method (Birch and Cary 1996). 4.2.3 Sample Extraction Filter samples were extracted and analyzed using the method described by Mazurek et al. (1987). Prior to extraction, a recovery standard was first added to each sample to monitor the extraction efficiency and loss on target analytes. This recovery standard included a group of isotopically labeled compounds of varying polarity and volatility to span the range of species to be used as markers, including n-dodecane-d26, n-hexadecane-d34, n-eicosane-d42, n-octacosane-d58, acenaphthene-dlO, chrysene-dl2, dibenz[a,h]anthracene-dl4 and D-glucose-l,2,3,4,5,6,6-d7. After spiking, samples were extracted with 2 30-mI aliquots of n-hexane and 3 30-ml aliquots of 2:1 benzene: isopropanol under mild ultrasonic agitation for a period of 1 Omin for each aliquot. The sample extract was concentrated and derivatized following the same procedure as described in Section 3.2.3.1 in Chapter 3. 75 4.2.4 Sample Analysis Aliquots of 1 ul from the original and derivatized extracts were analyzed by a Hewlett-Packard model 6890 gas chromatograph (GC) coupled to a HP model 5973 mass selective detector (MSD). The GC/MS conditions and operating procedures for the analysis of saccharide compounds in the derivatized extracts were the same as described in Section 3.2.3.2 in Chapter 3. The GC operating conditions for the analysis of non-polar organic compounds in the original extracts were as follows: temperature hold at 65°C for 10 min, increase from 65 to 285°C at a rate of 10°C min"1 and hold for lOmin, then increase from 285 to 310°C at a rate of 40°C min"1 with final isothermal hold at 310°C for 1.5 min. Marker compounds used to track primary sources were quantified, including alkanes, hopanes, polycyclic aromatic hydrocarbons (PAHs), and saccharides. Species quantified include the n-alkane series (C20-C34), 18a(H)-22,29,30-trisnorneohopane, 17a(H)-22,29,30-trisnorhopane, 17a(H),21 P(H)-29-norhopane, 18a(H), 29-norneohopane, 17 22R-17a(H),21 p(H)-30-homohopane, 22S-17a(H),21 P(H)-3O-bishomohopane, 22R-17a(H),2ip(H)-30-bishomohopane, fluoranthene, pyrene, chrysene/triphenylene, benzo[a]anthracene, benzofb] fluoranthene, benzofk] fluoranthene, benzo[e]pyrene, benzo[a]pyrene, perylene, indeno[l,2,3-cd]pyrene, dibenz[a,h]anthracene, benzo[ghi]perylene, levoglucosan, glucose, sucrose and trehalose. These species include a variety of potential markers for primary sources of fine particulate matter 76 including wood combustion, fossil fuel combustion, soil resuspension, and primary biological aerosol particles. 4.3 Results and Discussion 4.3.1 Organic speciation OC and EC The ambient concentrations of organic carbon (OC) and elemental carbon (EC) during the sampling program in summer 2007 are plotted in Figure 4.1. The average concentration of OC in PM2.5 measured in this study was higher compared to the measurements by Brown et al. (2002), July through October 1999 (5.63 ugC m"3 vs. 1.08 ugC m'3), while the average concentration of EC was more comparable in two studies (0.18 ugC m"3 vs. 0.17 ugC m"3). The samples of fine PM collected in Big Bend National Park are characterized by larger concentrations of OC compared to EC, which is consistent with the remote location where biogenic sources and secondary organic aerosol formation dominate over EC- rich sources such as diesel combustion. The ratio of OC to EC ranged from 16.2 to 297 with a mean of 49.5 and a median of 33.4 in the summer of 2007. This OC/EC ratio is large compared to measurements in more urban locations and consistent with the assumption that air quality at Big Bend is likely influenced by secondary organic aerosol formation from transport of emissions in other areas. A loose correlation between the measured OC and EC concentrations is shown in Figure 4.2. The weak correlation between EC and OC 77 and the relatively large intercept are likely due to the importance of regional transport and secondary organic aerosol formation compared to an urban location where primary emissions from EC and OC rich sources dominate. Figure 4.1 Ambient OC and EC concentrations in PM2.5 samples at BBNP E o C O 4) O C o o o o c o !5 E < Figure 4.2 Correlation of ambient OC and EC concentrations in PM2.5 samples at BBNP 10 9 8 • • 7 E 6 y = 7.641 x + 4.0459 O _ 4 o) 5 R2 = 0.3117 2 1 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 EC (ugC/m3) 78 n-Alkanes Normal alkanes ranging from C20 to C33 were quantified in all fine PM samples (Table 4.1). An example of the distribution of n-alkane homologues is given in Figure 4.3, showing the average concentrations of C25 to C33 n-alkanes measured in all sample collected in BBNP. This example, typical of the Big Bend samples in the summer of 2007, shows a strong enrichment in odd carbon number alkanes (C27, C29, C31 and C33), indicating biogenic origin for the measured alkanes. This so-called saw-tooth alkane pattern, where odd carbon alkanes are enriched over even isomers, is typical of leaf waxes and other biogenic organic particle sources (Rogge et al. 1993). This is very different from the smooth alkane pattern typical of fossil sources of n-alkanes that have roughly equal concentrations of odd and even homologues (Simoneit 1984). The highest average concentrations were found for C29 and C31, which is also characteristic of leaf wax alkanes (Rogge et al. 1993). In order to estimate the relative contributions from biogenic and anthropogenic sources, the carbon preference index (CPI) was calculated for each sample, which was the ratio of the sum of odd-carbon isomers to the sum of even-carbon isomers for n-alkanes ranging from C26 to C33 (Kavouras and Stephanou 2002). A CPI value close to 1 is regarded as characteristic of petroleum hydrocarbons, while a higher CPI value indicates impact from biogenic sources (Simoneit 1984). The CPI values in the 2007 Big Bend samples ranged from 1.06 to 3.29 with a mean of 1.95 and a median value of 1.98 (Table 4.1), indicating a strong influence of biogenic sources to alkane concentrations in this region. The CPI 79 values for samples for the summer sampling campaign are plotted in Figure 4.4. As shown clearly, generally higher CPI values were measured in the late spring and early summer samples, which may have resulted from the possible biogenic emissions from plant growth during this period. Alkane CPIs were lower in the late summer samples, which is consistent with BRAVO study observations (Brown 2001), suggesting anthropogenic emissions were more dominant in this period. Another parameter defining the contribution of biogenic organic carbon (Cwax) was calculated to estimate the contribution to alkanes from plant wax sources C +Cn+1 [Cwax = £(Cn - ""' )] (Gogou et al. 1996) and the values of Cwax n=25 ^ accounted for between 11% and 55% of the total measured alkanes (mean of 35%, median of 37%) (Table 4.1). This value is relatively high compared to values calculated for urban areas (Yue and Fraser 2004), but still indicated that more than 50% of the alkanes species originated from petroleum sources during most of the sample period. Figure 4.3 Saw-tooth alkane pattern of PM2.5 samples at BBNP 2 0 < C25 C26 C27 C28 C29 C30 C31 C32 C33 80 Figure 4.4 Alkane CPI of BBNP PM2.5 samples a=.• 30 2» • • a> • -. • •=o 2.0 a> u | 1.5 • • • • a£. 1.0 | 0.5 ni. O 0.0 -t 1 1 1 1 1 1 1 1 1 r 5/8 5/13 5/18 5/23 5/28 6/2 6/7 6/12 6/17 6/22 6/27 7/2 7/7 Table 4.1. Ambient concentrations of organic compounds in PM2.5 at BBNP (ng m") mean median minimum maximum n-Alkanes C20 (Eicosane) 0.39 ± 0.06 0.36 0.10 0.87 C21 (Heneicosane) 0.53 ± 0.07 0.51 0.16 0.96 C22 (Docosane) 0.71 ±0.10 0.68 0.23 1.31 C23 (Tricosane) 0.92 ±0.12 0.90 0.33 1.51 C24 (Tetracosane) 1.61 ±0.25 1.65 0.46 3.09 C25 (Pentacosane) 1.37±0.19 1.39 0.51 2.47 C26 (Hexacosane) 2.65 ± 0.35 2.79 0.98 4.15 C27 (Heptacosane) 3.02 ± 0.41 3.06 1.08 5.05 C28 (Octacosane) 2.15 ±0.33 2.07 0.66 3.66 C29 (Nonacosane) 3.82 ± 0.52 3.84 0.90 6.60 C30 (Triacontane) 1.88 ±0.35 1.64 0.58 4.15 C31 (Hentriacontane) 5.21 ±0.76 4.95 1.09 9.51 C32 (Dotriacontane) 1.15 ±0.25 1.03 0.08 2.56 C33 (Tritriacontane) 2.32 ± 0.38 2.31 0.23 4.18 C34 (Tetratriacontane) 0.20 ± 0.09 0.15 BDL 0.99 CPI 1.95 ±0.21 1.98 1.06 3.29 Cwax 8.21 ±1.44 8.13 1.16 17.73 ZCn(25-33) 23.6 ±3.0 24.7 7.24 36.8 Biogenic contribution (%) 35.2 ±4.5 36.9 10.8 55.0 PAHs Fluoranthene 0.04 ±0.01 0.03 BDL 0.11 Pyrene 0.03 ±0.01 0.03 BDL 0.11 Chrysene 0.04 ± 0.02 0.01 BDL 0.20 Benzo[a]anthracene 0.01 ± 0.003 0.004 BDL 0.03 Benzo[b]fluoranthene 0.01 ± 0.002 0.006 BDL 0.02 Benzo[k]fluoranthene 0.01 ± 0.003 0.01 BDL 0.04 Benzo[e]pyrene 0.01 ± 0.004 0.01 BDL 0.04 Benzo[a]pyrene 0.01 ± 0.003 0.01 BDL 0.03 Perylene 0.02 ± 0.005 0.02 BDL 0.05 lndeno[1,2,3-cd]pyrene BDL BDL BDL BDL Dibenz[a,h]anthracene BDL BDL BDL BDL Benzo[ghi]perylene BDL BDL BDL BDL Hopanes 18a(H)-22,29,30-trisnorneohopane 0.23 ± 0.05 0.26 BDL 0.45 17a(H)-22,29,30-trisnorhopane 0.21 ± 0.04 0.22 BDL 0.43 17a(H), 21p(H)-29-norhopane 0.42 ± 0.09 0.44 BDL 1.04 18a(H)-29-norneohopane 0.22 ± 0.05 0.23 BDL 0.57 17a(H), 2ip(H)-hopane 0.60 ±0.116 0.62 0.14 1.34 22S-17a(H), 0.13 ±0.04 0.12 BDL 0.35 21 (3(H)-30-homohopane 22R-17a(H), 0.15 ±0.04 0.15 BDL 0.42 213(H)-30-homohopane 22S-17a(H), 0.12 ±0.03 0.10 BDL 0.39 21 (B(H)-30-bishomohopane 22R-17a(H), 0.09 ± 0.03 0.03 BDL 0.27 21 B(H)-30-bishomohopane Saccharides Levoglucosan 3.70 ±1.46 2.44 0.08 11.9 Glucose 10.9 ±2.8 9.32 1.23 32.2 Sucrose 1.33 ±0.47 0.94 0.10 4.73 Trehalose 0.23 ±0.11 0.12 BDL 1.46 Saccharides Saccharides in aerosols have been proposed as one of the primary markers for biogenic sources such as biomass burning, primary biological aerosol particles, atmospheric entrainment of soil and associated biota (Simoneit and Mazurek 1981; 82 Bieleski 1982; Oros and Simoneit 2001; 2001; Graham et al. 2003; Simoneit et al. 2004; Caseiro et al. 2007). Several important saccharide compounds including levoglucosan, glucose, sucrose and trehalose were quantified in each sample (Table 4.1) to characterize the contribution from these primary biogenic sources in Big Bend. Glucose was found to be the most abundant saccharide compound, with concentrations ranging from 1.23 to 32.21 ng m"3, mean of 10.94 ng m"3, and median of 9.32 ng m"3. Glucose is the simple sugar present in vascular plants (Cowie and Hedges 1984) and serves as a primary carbon reservoir for plants and microorganisms (Pigman and Horton 1980). Limited samples collected in BBNP in the winter and spring indicate a possible trend of increasing glucose concentrations through the growing season starting in the late spring and summer (Figure 4.5), as a result of plant development, which is consistent with studies of other regions (Pashynska et al. 2002; Medeiros et al. 2006; Wan and Yu 2007). Levoglucosan was the second most abundant saccharide, ranging in concentration from 0.08 to 11.89 ng m"3 with a mean of 3.70 ng m"3 and a median of 2.44 ng m"3. As a key tracer for biomass burning (Simoneit et al. 1999), the elevated levoglucosan concentration in the first half of the sampling period paralleled the fire season when forest and agriculture burning in southern Mexico is generally strongest (Pitchford 2004). Higher levels of sucrose were measured in May (Figure 4.5), which likely resulted from the blooming of flowers during this period, as sucrose is important in development of flower buds and is a major constituent in pollen (Bieleski 1995; Pacini et al. 2006). Trehalose was detected in small amounts in each sample during the summer sampling period. Trehalose is a fungal metabolite and also serves as a stress protectant in bacteria, yeast, insects, and a few higher plants (Eleutherio et al. 1993; Eleutherio et al. 1993; Caseiro et al. 2007). The low concentrations of aerosol trehalose likely indicate a lack of environment stress during this period. Typically May and June are hot and dry at BBNP; however, 2007 was an exception when plentiful rains (average 0.07 inch in May and June) kept the desert green and the temperatures pleasant (average 76°F in May and June) (Weather Underground 2007). Concentrations of levoglucosan can be used to estimate the fraction of OC contributed from wood smoke. Source tests of biomass combustion aerosols indicate that between 4.5% and 10% of aerosol OC (fxgC m"3) from biomass combustion is the single compound levoglucosan (ng m"3) (Schauer et al. 2001). Using this range of source characterization, the estimated OC from wood combustion would vary from about 1 ng m"3 to 264 ng m"3, and the percentage of the estimated OC from wood smoke to the measured OC ranged from less than 0.1% to 6.2%. This indicates that biomass combustion is never the dominant source of carbonaceous particles in Big Bend. Figure 4.6 shows the range of the estimated OC from wood combustion to the measured OC during the sampling period and from this data, the contribution to the total OC appears greater during winter and spring than during summer periods. 84 Figure 4.5 Ambient saccharide concentrations in BBNP PM2.5 samples (glucose, levoglucosan and sucrose) Glucose Levoglucosan 11 • •. a I „ 11 Sucrose .,, ,1, 111 1 •U JU III. ill 85 Figure 4.6 Range of predicted contribution to biomass burning particulate matter at BBNP. Top of range indicates maximum predicted impact, diamond mean predicted impact, and bottom of range is minimum predicted impact. c •Js 12% in 3 XI E o 10% o TJ O O 8% 5 E p 6% o o 4% J » £ 2% j in l 0% JP J^ X& W& J^ <& J^ «.^ /l^ y.^ A^ Petroleum Biomarkers Motor vehicle emissions can be tracked using fossil petroleum biomarkers such as hopanes and steranes (Rogge et al. 1993), which are organic compounds present in crude petroleum deposits as remnants of biological material (Simoneit 1984). Nine hopane compounds were quantified in the fine PM samples from BBNP (Table 4.1). The most abundant species include 17a(H),2ip(H)-29-norhopane and 17a(H),21 P(H)-hopane, reaching an average ambient concentration of 0.42 and 0.60 ng m"3, respectively. These two compounds were commonly detected in urban atmosphere with relatively high concentrations (Schauer et al. 1996; Fraser et al. 1999; Schauer et al. 2002). The total concentration of the nine measured hopane species ranged from 0.14 to 5.18 ng m"3 with a mean of 2.14 ng m" , representing only a small 86 fraction of the total measured OC. However, this small fraction of up to 0.01% is comparable to even some of the urban areas (Yue and Fraser 2004; Li et al. 2006), indicating that mobile sources contributed to the ambient fine organic particles at Big Bend (Zielinska et al. 2004). In general, hopane concentrations measured in this study were relatively larger than those observed in the BRAVO study (Brown et al. 2002), for example, the most abundant hopane species 17a(H),2ip(H)-hopane measured in the BRAVO study was less than 0.01 ng m"3, compared with 0.60 ng m"3 in current study. This was possibly due to the differences between seasons or meteorological conditions for each sample period. The contribution of mobile sources to primary fine aerosol OC mass can be estimated using the ratio between the concentration of petroleum markers and total OC emitted from vehicle exhaust based on the emission source profiles (Rogge et al. 1993; Brown 2001; Brown et al. 2002). Based on source tests of fine PM OC from mobile sources (Hildemann et al. 1991; Rogge et al. 1993), concentrations of petroleum biomarkers measured in this study were used to estimate that OC from mobile sources contributed from 1.8% to 32% of the measured fine PM OC with a mean contribution of 16%, and a median of 17%. Poly cyclic Aromatic Hydrocarbons (PAH) Polycyclic aromatic hydrocarbons (PAHs) are formed primarily from combustion processes of either vegetative or fossil fuel materials (Simoneit 1984). The anthropogenic PAH sources, mainly from fossil fuel combustion processes, are 87 by far the major contributors of hydrocarbons with known hazards to human health (Simoneit 2002). A range of PAH compounds were analyzed, and concentrations of individual PAH are listed in Table 4.1. Generally, very low levels of PAHs were detected, consistent with the BRAVO study results (Brown 2001; Brown et al. 2002). Total concentrations of the measured PAH compounds averaged 0.175 ng m"3, ranging from 0.025 to 0.502 ng m"3. Fluoranthene, pyrene, chrysene, and perylene were found to be the more abundant PAHs, with mean concentrations of 0.039, 0.033, 0.036 and 0.019 ng m"3, respectively. Higher-molecular-weight PAHs including indeno[l,2,3-cd]pyrene, dibenz[a,h]anthracene, and benzo[ghi]perylene were not detected in any of the fine particle samples. 4.3.2 Source Apportionment Using PMF 4.3.2.1 Data Analysis by PMF The EPA PMF 1.1 was applied to isolate the potential source contributions to the Big Bend PM2.5 samples. A total of 23 marker species were used in the modeling, which included OC, EC, n-alkanes ranging from C25 to C33, hopanes including 17a(H)-21 P(H)-hopane, 17 22S,17a(H),21p(H)-30-homohopane, and 22R,17a(H),2ip(H)-30-homohopane, PAHs including benzo[k]fluoranthene, benzo[e]pyrene, benzo[a]pyrene and perylene, saccharides including levoglucosan, glucose, sucrose, and trehalose, and other components including sulfate, aluminum, silicon, and sodium. Other organic species 88 identified in this study were not included in the PMF data set because of their small concentrations. Concentrations of OC, EC, alkanes, hopanes, PAHs and saccharides were determined by the chemical analysis in this study (concentrations below detection limits were substituted with DL concentrations), while the concentrations of sulfate, aluminum, silicon and sodium were obtained from parallel sampling by the Texas Commission on Environmental Quality. In addition to isolating factor composition, PMF also can be used to determine the optimal number of sources to fit the ambient data set based on agreement between the theoretical and calculated Q values, where Q value is the sum of the squares of the residuals weighted inversely by the variation of the data points (Paatero 1997). A PMF application with 9 resolved sources was chosen as the best solution to the PMF model for this application. The specific source profiles resolved by the nine-factor model are shown in Figure 4.7. The fact that each isolated source is responsible for the majority of the contribution from at least one marker indicates good model application and separation of source contributions. 4.3.2.2 Source Profiles The first factor was dominated by the well-established marker for biomass burning - levoglucosan (Simoneit et al. 1999). Four PAH compounds (benzo[k]fluoranthene, benzo[e]pyrene, benzo[a]pyrene and perylene) were also relatively abundant in this factor as they are byproducts of the pyrolysis processes in wood fires (Simoneit 2002). A time series plot of this factor (Figure 4.8) shows a 89 much higher contribution in winter and spring, paralleling the fire season when forest fires and agricultural burns are more prevalent (Pitchford 2004; Simoneit et al. 2004). The factor isolated with the high concentrations of two elements - aluminum and silicon, combined with lesser contributions from organic species including alkanes, hopanes and PAHs has been labeled as road dust based on source knowledge (Schauer et al. 1996). Aluminum and silicon represent common elements associated with mineral material, while the organic species are consistent with vehicular emission deposited, then resuspended, from road surfaces. The temporal trend in the source contribution of this factor (Figure 4.8) shows a peak in summer time when warm and dry weather conditions facilitate the resuspension of road dust by vehicle traffic. The isolated factor dominated by sulfate is labeled as secondary sulfate formation. The relatively high concentration of EC in this factor suggested that this source could be related to coal combustion processes (Chen et al. 2002; Ogulei et al. 2005), and possible contributors to this factor include power plants on both sides of the US-Mexico border. The seasonal variation (Figure 4.8) shows higher concentrations in summer time when higher concentrations of atmospheric oxidants result in more rapid conversion of emitted S02 from coal combustion into particulate sulfate (Song etal. 2001). A factor was isolated with Na as the major species and is labeled as sea salt. The contribution of sea salt to PM2.5 concentrations did not display a seasonal pattern 90 (Figure 4.8), but elevated contributions at certain times suggest incursions of marine air masses carried by the winds most likely from the Gulf of Mexico (Pitchford 2004). A factor was isolated with high concentrations of a series of alkanes (C25-C33), which could originate either from vegetative waxes or petroleum hydrocarbons (Simoneit 1984). This factor was also rich in OC and some of the hopane and PAH species; the source profile does not display enrichment in odd carbon number alkanes, as is expected for biogenic wax emissions (Standley and Simoneit 1987). This implies the factor originated from petroleum emissions as opposed to biogenic sources. Two factors with enrichment of saccharide compounds were resolved separately. The first factor was dominated by glucose and trehalose, while the second factor was dominated by sucrose. Glucose is as a major source of carbon for microbes (Paul and Clark 1996), and trehalose is a well-known constituent of fungal spores (Bieleski 1982), serving as a reserve carbohydrate and/or cell protectant against stressful conditions (Eleutherio et al. 1993; Eleutherio et al. 1993; Chaturvedi et al. 1997); therefore, the linking of glucose and trehalose may indicate some fungal spore derived bio-aerosols. The fact that this source profile has very low levels of silicon and aluminum suggests this factor has little contribution from soil and associated biota. On the other hand, sucrose, as an important sugar for developing flower buds (Bieleski 1995), is particularly enriched in pollen grains (Pacini 2000). The separation of this sucrose-rich factor may suggest an origin from pollen. Graham et al. (2003) observed a nighttime increase in yeasts and other small fungal spores and 91 higher pollen, fern spore, and insects counts during the day. Again, small contributions of silicon and aluminum to this factor suggest it is mainly derived from PBAPs, not soil resuspension. The separation of the PBAP source into two different factors by PMF may indicate a time of day dependence on the two factors between the glucose and trehalose rich bio-aerosol (fungal related) and the sucrose rich bioaerosol (pollens related). The last two factors were characterized by high contributions of PAH, EC, OC and hopanes. Separation of these two factors is based on larger concentrations of PAHs and OC in gasoline vehicle emissions, while the diesel vehicle emission factor was distinguished by a peak concentration of EC and an elevated contribution from the hopane species (Figure 4.7). Gasoline powered vehicles tend to emit larger amount of OC, and large concentrations of EC provide a fingerprint for diesel engine exhaust (Schauer et al. 1996; Seinfeld and Pandis 1998; Ogulei et al. 2005). Gasoline vehicle exhaust also contains more high-molecular-weight PAHs as compared with diesel exhaust (Miguel et al. 1998; Watson 1998; Zielinska et al. 2004), as is observed in the gasoline powered vehicle factor in Figure 4.7. Hopanes are compounds present in crude oil as a result of the decomposition of hopanoids and other biomass (Simoneit 1985) and have been used as molecular markers of vehicle emissions (Simoneit 1985; Rogge et al. 1993; Fraser et al. 1999). Large emissions of hopanes were observed both from gasoline and diesel fueled vehicles (Zielinska et al. 2004; Lough 2007; Lough 2007), and the enrichment of hopanes in the sucrose rich bio-aerosol factor resolved by PMF cannot be readily explained. 92 Figure 4.7 Source profiles of the 9 factor PMF solution of BBNP PM2.5 samples (Qrobust=290.4). (OC: organic carbon; EC: elemental carbon; C25-C33: alkane series; HOP1: 17a(H)-2ip(H)-hopane; HOP2: 17a(H)-21p(H)-29-norhopane; HOP3: 22S, 17 0 - ,1,1,ii ,1, . Ji, • , w. 140 C 120 Glucose and Trehalose Rich Bio-aerosol 8 * 100 &; ao 60 (Sources of model input: Concentrations of OC, EC, alkanes, hopanes, PAHs and saccharides were determined by the chemical analysis in this study, while the concentrations of sulfate, aluminum, silicon and sodium were obtained from parallel sampling by the Texas Commission on Environmental Quality.) 4.3.2.3 Relative Source Contributions The relative contribution of each isolated source to the total PM2.5 mass was calculated based on the PMF modeling result and a multiple linear regression between factor strengths and measured PM2.5 mass (Song et al. 2001), as shown in Figure 4.9. To resolve the contribution of different wax sources, the biogenic contribution to that factor was estimated using the 35.2% biogenic contribution of the total wax (listed in Table 4.1). Gasoline and diesel exhausts were lumped with the petroleum portion of the isolated wax factor into a combined mobile source emissions. The glucose- and trehalose-rich factor and the sucrose-rich factor were combined to characterize the contribution from PBAPs. The largest predicted contributor to the fine particulate matter level at Big Bend was the secondary sulfate factor, which accounted for about 43% of the total measured PM2.5 mass. Sulfate is the major cause of the Big Bend haze any time of year, and this contribution is consistent with the previous BRAVO findings (Pitchford 2004). Particulate sulfate at Big Bend originates from numerous SO2 sources across various geographic regions (Pitchford 2004). Mobile sources were the second largest contributor, which accounted for 21% of the PM2.5 mass. This percentage is relatively small compared to similar rural settings in Texas (Jia 2008), likely a result of the remote location of BBNP. Apart from the two major contributors discussed 95 above, there were four source factors that contributed almost equally to the Big Bend PM2.5 mass, including sea salt (11%), wood smoke (9%), PBAPs (7%), and road dust (8%). The smallest contributor to PM2.5 is plant wax, which accounts only about 2% compared to other major sources. Finally, the good agreement between the apportioned PM2.5 mass and the measured values (Figure 4.10) demonstrates that the model has successfully isolated the important sources contributing to the BBNP PM2.5 levels. In summary, PMF resolved nine source factors that contributed to the PM2.5 mass at Big Bend. Particularly, saccharides were proven once again as useful indicators of primary biologically derived sources resuspended into the atmosphere, and the separation of gasoline and diesel vehicle emissions was possible based on the abundances of different petroleum and combustion generated species. Due to the limited size of input data sets, the modeling results of this analysis may be subject to variability, as a larger sample size would yield more robust modeling results. In a modeling domain where source profiles are not available, PMF can be a good preliminary approach of source apportionment. The modeling results can also expand the PM2.5 source inventory by adding previously unknown sources or refining existing source profiles. Figure 4.8 Source contributions to PM2.5 at BBNP resolved by PMF as a function of time. Source contribution is determined by a multiple linear regression of source factor scores to the measured PM2.5 mass (The Y axis is unitless, which represents the relative contribution of a specific factor at each sampling day spanning the whole sampling period, and the average of all Y values equals to 1). 96 s: s; i** 97 Figure 4.9 Relative source contributions to BBNP PM2.5 estimated by PMF. Source contribution is determined by a multiple linear regression of source factor scores to the measured PM2.5 mass. Plant wax 2% Primary biological Wood smoke aerosol particles 9% 7% Road dust Sea salt 8% 10% Mobile sources 21% Secondary sulfate 43% Figure 4.10 Apportioned PM2.5 mass concentrations plotted against the measured values for BBNP PM2.5 samples 10 15 25 Measured PM2.5 Mass (ug/m3) 98 4.4 Conclusions Fine particulate matter samples collected in the Big Bend National Park were characterized with a high OC/EC ratio (49.5 on average), indicating the importance of regional transport and secondary aerosol formation. Odd carbon alkanes were enriched over even isomers, creating a CPI value between 1.06 and 3.29. This suggested plant wax particles were an important source of alkanes for this region, and their contribution to the alkane wax concentrations ranged from 11% to 55%. Saccharide compounds were also identified in fine particulate samples, with glucose (10.94 ng m"3 on average) and levoglucosan (3.70 ng m"3 on average) being the two most abundant species. Calculation of the estimated OC from wood smoke was low compared to the measured OC, showing that biomass combustion is never the dominant source of carbonaceous particles at BBNP. For petroleum markers, high concentrations of hopanes were measured (total average 2.1 ng m"3), implying that mobile sources were a major contributor to the ambient fine particles at BBNP. Analysis of PAHs further suggests this motor vehicle source was more influenced by diesel-powered vehicles, as only lower-molecular-weight PAH species were detected. A Positive Matrix Factorization (PMF) model was later applied using organic compounds measured in this study as well as inorganic species. Nine sources were isolated, including sources identified as secondary sulfate, gasoline vehicle, diesel vehicle, wax, wood smoke, sea salt, road dust, sucrose rich bio-aerosol, and glucose and trehalose rich bio-aerosol. The PMF results were similar to the preliminary interpretation of the ambient data and suggested the secondary sulfate and mobile 99 sources contributed a great portion to the ambient PM level in Big Bend (43% and 21%, respectively). These results indicate the broader ability of the PMF approach to investigate source apportionment using organic molecular markers. 4.5 Acknowledgement This project was funded by the Texas Commission on Environmental Quality (TCEQ). All the organic speciation analysis was conducted in the Chemistry Lab at Arizona State University with the help of Dr. Pierre Herckes and his research group. I would like to thank John Forsythe for sample collection, my former colleague Shagun Bhat for sample shipping and handling, Sunset Lab for analyzing EC and OC, and TCEQ for providing the Big Bend ion speciation data that was used in the PMF model. 100 Chapter 5 Saccharide Composition in Fine and Coarse Particulate Matter and Soils in Central Arizona 5.1 Introduction The desert southwestern United States routinely exceeds the federal 24-hr NAAQS (National Ambient Air Quality Standards) for coarse particulate matter (AirData 2008). PM10 concentrations are large in both urban and rural areas and are believed to originate from fugitive dust sources including agricultural fields, roads, and soil erosion from the surrounding desert locations. Other potential sources to PM10 may include PBAPs given the broad mixture of flower, grass, and fungal species that thrive in the Sonoran desert area and actively release pollens and spores throughout the year (O'Rourke 1990). Notably, epidemiological studies have shown that pollen, fungi spores, and other allergens are responsible for the prevalence of allergic rhinitis (also known as hay fever) and asthma in the southwest which occurs in approximately 35% and 7% of the population, respectively (Schumacher 2008). However, because soil is also believed to contain a large number of these biological particles and is considered a secondary host of PBAPs (Cox and Wathes 1995; Simoneit et al. 2004), separating the ambient entrainment route for biological aerosol components in the desert southwest has not been clearly stated or investigated. In an effort to identify and assess the relative contribution of these and other 101 major PM sources in the southwestern US region, and particularly to assess the contribution of soil entrainment, a series of ambient PM and soil samples was collected in Higley, AZ, a suburb of Phoenix. This suburban region has seen rapid urban sprawl onto agricultural lands, and the role of agricultural sources compared to more traditional urban sources is an important consideration for local environmental policy. Because of their suggested ability to track biologically important organic materials from soils (Simoneit et al. 2004; Medeiros and Simoneit 2007; Rogge et al. 2007), saccharide compounds were analyzed in size-segregated soil and ambient PM samples at Higley; intra- and inter- comparisons were made between the ambient PM and three types of soil dust samples (agricultural soil, currently uncultivated soil, and road dust) based on the particle size (fine vs. coarse), season of the year, and relative composition of 12 saccharide compounds. Based on the ambient concentrations of major saccharides and a number of other specific compounds (including elemental and organic carbon, ions, metals, n-alkanes, n-alkanoic acids, steroid, and hopanes) that are simultaneously resolved in Higley PM samples, a Positive Matrix Factorization (PMF) model was performed to determine the key contributors to PM10 and PM2.5 levels. The goals were to test the suitability of saccharides as molecular markers to trace the entrainment of soil dusts from the ground into the atmosphere and to elucidate other major sources that contribute to the PM levels in this location in the arid southwestern US. To our knowledge, this study is the first of its kind to compare the saccharide between the fine and coarse fraction of different soil types in two seasons and to relate the contribution from soil dust to ambient PM using saccharide 102 compounds. 5.2 Experimental method 5.2.1 PM sampling The sampling of ambient PM was conducted from January to April 2008 near Higley, Arizona (33°18'N, 111°43'W), a site of rapid urbanization in a traditionally agricultural location (Figure 5.1). Both PM2.5 and PM10 samples were collected over a 24-hr period every other day. Two sets of PM2.5 and PM10 samples were collected on pre-baked quartz-fiber filters (Whatman, 8"xl0") by high-volume air samplers (Thermo) fitted with PM2.5 and PM10 inlets respectively (MSP Corp). These quartz-fiber filter samples were taken at a flow rate of 1.13 m3min_1 to be used for bulk chemical compositional analysis and analysis of organic markers. In parallel, two additional PM2.5 and PM10 samples were collected on pre-weighed Teflon-filters (Whatman, 47mm) by low-volume air samplers (Thermo, Partisol-Plus Model 2025). Teflon-filter samples were taken at a flow rate of 16.7 LPM for gravimetric mass determination and metal speciation. After collection, quartz-fiber filter samples were returned from the field sites in glass jars with Teflon-lined caps and Teflon-filters in dedicated Petri-dishes. All samples were stored in a freezer at -4°F until analysis. A total of 45 sets of PM2.5 and 46 sets of PM10 samples were obtained, each set including both quartz-fiber filters and Teflon-filters. 103 Figure 5.1 Map of Arizona State (USA) including the Phoenix metropolitan area and the ambient sampling location (Higley) of this study r^" •"&*~^• , m igfeyi 5.2.2 Soil sampling During the PM sampling period, soil samples were taken in the vicinity of the Higley sampling site in January and April to characterize a likely source of ambient PM. Samples were taken from the surface earth (0-5 cm depth) in currently uncultivated soil fields (no agricultural practice), agricultural fields, and alongside major roads. Note that the term of "currently uncultivated soil" used in this thesis refers to soils without cultivation for the past at least ten years. It is commonly assumed that it takes ~50 years for the recovery of soil carbon after deforestation or abandonment of cultivation, although this rate may vary in different life zones (moist vs. dry vs. wet) (Lugo and Sanchez 1986). A total of 28 soil samples were obtained (half in January and half in April), which included 14 agricultural soil samples, 6 currently uncultivated soil samples, and 8 road dust samples. Table 5.1 lists the date, the sample type, a brief description, and taxonomy of the samples taken in this study. 104 Each sample was collected in pre-baked glass jars (250 ml) sealed with Teflon-lined lids and Teflon tape and stored in a freezer at -4°F after collection. Prior to analysis, each sample was dried at 110°C for 24 hrs and then resuspended. Size-segregated samples representing fractions equivalent to PM10 and PM2.5 were collected using two different cyclone separators (URG Model, 2000-3 OEA and 2000-3 0EC). Three 47mm filters were collected in parallel for each resuspension, one Teflon (Whatman) and two quartz (Whatman). Mass determination, elemental and organic carbon analysis, and organic speciation were achieved from these collected samples. Figure 5.2 shows the basic setup of the soil resuspension and sampling apparatus. Table 5.1 Sampling date, sample type and site description of Higley soil samples Site Sample Sampling ID Type Date Site Description Soil Taxonomy 2 Agricultural 1/15/2008 Soil from fairly bare agricultural field East of the sampling site Mohall loam1 2B Agricultural 4/12/2008 Soil from fairly bare agricultural field East of the sampling site Mohall loam1 3 Agricultural 1/17/2008 Taken about 50 feet into the field covered with dry, brown plant stems Mohall loam1 3B Agricultural 4/14/2008 Taken about 50 feet into the field covered which is dry and brown with Mohall loam1 some plant stems and a couple of green plants here and there 4 Agricultural 1/17/2008 Taken from along the base of a soil ridge dividing the field of green plants Vecont clay2 about 1ft tall at approximately 100 ft into the field. 2 4B Agricultural 4/8/2008 Taken from along the base of a soil ridge dividing the field approximately Vecont clay 100 ft into the field. The plants have been cut down and the field is now covered in dense stubble. Soil is somewhat wet. 5 Agricultural 1/21/2008 Taken along the base of a soil ridge within the dense green plant (~1ft Mohall loam1 high) growth about 100 ft into the field 5B Agricultural 4/2/2008 Taken along the base of a soil ridge within the dense green plant (~1ft Mohall loam1 high) growth about 100 ft into the field 6 Agricultural 1/21/2008 Taken about 100 feet within the bare field Mohall loam1 6B Agricultural 4/2/2008 Taken about 50-100 feet within a field that used to be bare but now has Mohall loam1 little plant sprouts 7 Agricultural 1/29/2008 Taken about 150 ft into an field which seems to have some grass-like Mohall loam1 cover which is dead, unwatered, and mowed. 7B Agricultural 4/12/2008 Taken about 150 ft into an open field with patches of ground cover and Mohall loam1 exposed dirt near an exposed dirt track. 105 8 Agricultural 1/29/2008 Taken within the plants about 150 ft into a field with very green grass-like Mohall loam plants 8B Agricultural 4/8/2008 Taken at the edge of the dense green grass-like plants which are now Mohall loam1 approximately 3 ft tall and very dense 1 Currently 1/15/2008 Local soil about 50 yards from the sampling site Contine clay loam3 uncultivated 1B Currently 4/2/2008 Local soil about 50 yards from the sampling site Contine clay loam3 uncultivated 11 Currently 2/20/2008 Taken from no growth area under the sampling platform at the site Contine clay loam3 uncultivated 11B Currently 4/14/2008 Taken from no growth area under the sampling platform at the site Contine clay loam3 uncultivated 13 Currently 2/20/2008 Taken from center of the open area with no plant growth East of the fire Mohall loam1 uncultivated station 13B Currently 4/12/2008 Taken from center of the open area with no plant growth East of the fire Mohall loam1 uncultivated station 12 Road Dust 2/20/2008 Taken from dirt shoulder on the West side of Higley Southwest of the site Contine clay loam3 12B Road Dust 4/4/2008 Taken from dirt shoulder on the West side of Higley Southwest of the site Contine clay loam3 14 Road Dust 2/20/2008 Taken from dirt shoulder on the South side of Williams Field Road Mohall loam1 14B Road Dust 4/4/2008 Taken from dirt shoulder on the South side of Williams Field Road Mohall loam1 15 Road Dust 2/26/2008 Taken from paved bike lane about 150 ft from the intersection Contine clay loam3 15B Road Dust 4/4/2008 Taken from paved bike lane about 150 ft from the intersection Contine clay loam3 16 Road Dust 2/26/2008 Taken from around the center divider at the entrance of the Farm Mohall loam1 restaurant 16B Road Dust 4/4/2008 Taken from around the center divider at the entrance of the Farm Mohall loam1 restaurant Mohall loam1: fine-loamy, mixed, hyperthermic, Typic Haplargids; Vecont clay2: fine, mixed, hyperthermic, Typic Haplargids; Contine clay loam3: fine, mixed, hyperthermic, Typic Haplargids Figure 5.2 Basic setup of the soil resuspension and sampling apparatus. HERA Filter To Waste Suction Pump 106 5.2.3 Element carbon (EC) and organic carbon (OC) analysis The elemental and organic carbon content of collected samples was determined using a Sunset Laboratory Thermo-Optical Carbon Analyzer. A 1 cm x 1.5 cm punch was obtained from the quartz-fiber filter corresponding to each sample, and the analysis was performed following the thermal-optical method described by Birch and Cary (1996). This part of analysis was conducted by my colleague Andrea Clements at Arizona State University. 5.2.4 Gravimetric analysis The total mass of each sample was determined using a microbalance (Mettler Toledo). Teflon-filters were pre- and post-weighed after equilibration under controlled temperature (22-24°C) and humidity conditions (45-55%). 5.2.5 Organic speciation A portion of the quartz-fiber filters from both the ambient PM and soil resuspension samples was extracted and analyzed for saccharide composition following the method described by Medeiros and Simoneit (2007). This method has proved highly efficient for extracting saccharide compounds (Medeiros and Simoneit 2007). Prior to extraction, an isotopically labeled internal standard (D-Glucose-l,2,3,4,5,6,6-d7) was spiked onto one fourth of each quartz-fiber filter. After spiking, samples were extracted, concentrated, and derivatized following the same procedure as described in Section 3.2.3.1 in Chapter 3. The analysis procedures 107 for the original and derivatized extracts were the same as described in Section 4.2.4 in Chapter 4. 5.3 Results and discussion 5.3.1 Ambient PM speciation The total PM mass, OC, EC, n-alkanes (C20-C34), one steroid compound (cholesterol), hopanes, even-carbon-number n-alkanoic acids (C8-C24), and saccharides were analyzed in both the PM10 and PM2.5 samples at Higley, and the average concentrations, ratios between compounds, and carbon preference index (CPI) are summarized in Table 5.2. PAHs and triterpenoids were also analyzed but not detected, and therefore are not listed here. During the sampling period, both PM2.5 and PM10 were below the NAAQS 24-hr standards, with an average concentration of 5.8 ug m"3 for PM2.5 and 29.9 ug m"3 for PM10. In general, PM2.5 (ug m"3) contained a higher percentage of OC (ugC m"3) (27.6% on average) than was measured in PM10 (12.5% on average). The ratio of OC to EC ranged from 1.1 to 12.4 with a mean value of 5.4 in PM2.5, and ranged from 3.8 to 31.1 with a mean value of 7.8 in PM10. This OC to EC ratio was lower than that measured in the Big Bend PM2.5 samples (Chapter 4), but higher than the primary OC/EC ratio in the Los Angeles Basin (Turpin et al. 1991), likely indicating an impact from both primary emissions and secondary organic aerosol formation (Turpin and Huntzicker 1995). The relationship between the measured OC and EC 108 showed greater variability in PM10 than in PM2.5 (Figure 5.3), which suggests a greater diversity of sources for PM10 relative to PM2.5. n-Alkanes measured in both PM2.5 and PM10 samples showed an enrichment in odd carbon number isomers (Figure 5.4), with the highest concentrations found for C29 and C31, typical of leaf waxes and other biogenic sources (Simoneit 1984; Rogge et al. 1993). A CPI value higher than 1 (2.3 on average for PM2.5 and 1.9 on average for PM10) also confirms the impact from biogenic sources (Simoneit 1984), which was estimated to contribute 38% and 31% on average (Table 5.2) to the quantified alkanes in PM2.5 and PM10, respectively (Gogou et al. 1996). This estimated biogenic contribution to alkanes was comparable to that calculated in the Big Bend PM2.5 samples (Chapter 4), but lower compared to the value calculated for the Houston urban area (Yue and Fraser 2004). Eight hopane compounds were detected in most of the PM samples, with 17a(H), 2ip(H)-hopane and 17a(H), 21|3(H)-29-norhopane as the most abundant two species (Table 5.2), consistent with other measurements in urban areas (Schauer et al. 1996; Fraser et al. 1999; Schauer et al. 2002). However, the fact that higher concentrations of hopanes were measured in PM10 than PM2.5 samples was not expected (total average 6.3 ng m"3 in PM10 vs. 0.6 ng m"3 in PM2.5), and can not be readily explained. A hypothesis might be the effect of heterogeneous oxidation of condensed phase hopanes due to a larger surface area of PM2.5 especially at low RH levels (Weitkamp et al. 2008) and coagulation of fine particles containing hopane 109 compounds and incorporation into road dust during the aerosol aging process (Wang et al. 2009). n-Alkanoic acids (C8 to C24) with even carbon numbers were also quantified, with hexadecanoic acid (CI6) and octadecanoic acid (CI8) being the two dominant species (Table 5.2). These two n-alkanoic acids were often identified as the two predominant n-alkanoic species in urban and background aerosols (Simoneit 1989; Limbeck and Puxbaum 1999). Only trace levels of alkanoic acids with 20 or more carbon atoms (>C20) were detected, suggesting the contribution from vegetative plants was small (Simoneit and Mazurek 1981; Simoneit 1989). The dominance of n-alkanoic acids (Rogge et al. 1991; Schauer et al. 2002), fuel oil combustion (Rogge et al. 1997), or motor vehicle emissions (Rogge et al. 1993; Schauer et al. 1999; Schauer et al. 2002). However, it is worth noting that the meat cooking tracer cholesterol was not detected, which implies the contribution from the meat cooking sources is likely small. The measured n-alkanoic acids were more enriched in PMC (PM10 - PM2.5) than PMf (or PM2.5), which is consistent with the observations in Amazonian aerosols by Graham et al. (2003). Twelve saccharide compounds were measured, including one anhydrous saccharide, one monosaccharide, two disaccharides, and eight saccharide polyols (Table 5.2). The next section (5.3.2) will discuss in details the ambient concentrations, 110 seasonal variation during the sampling period, size distribution, and correlation analysis of the measured saccharides. Table 5.2 Ambient concentrations of PM, EC, OC, and organic compounds in Higley PM samples (Jan - Apr 2008) (ng m'3) PM2.5 (n=45) PM10(n=46) PM mass (ug m3) 5.8 + 2.2 29..9±10.6 OC mass (pgC m-3) 1.6 ±0.5 3.7 + 0.9 EC mass (pgC m-3) 0.3 ±0.2 0.6 + 0.3 OC/EC 5.4 ±1.7 7.8 ±4.5 OC/PM (%) 27.6 ± 20.7 12.5 ±8.8 n-Alkanes C20 (Eicosane) 0.50 ± 0.45 0.70 ±0.41 C21 (Heneicosane) 0.89 ± 0.68 1.33 ±0.68 C22 (Docosane) 0.87 ± 0.60 1.77 ±0.89 C23 (Tricosane) 0.99 ± 0.73 2.69 ±1.37 C24 (Tetracosane) 0.92 ± 0.74 3.59 ±2.18 C25 (Pentacosane) 1.42 ±2.68 4.31 ±5.18 C26 (Hexacosane) 0.55 ± 0.54 1.78 ±1.35 C27 (Heptacosane) 0.80 ± 0.70 1.77 ±1.35 C28 (Octacosane) 0.71 ± 0.65 1.20 ±0.74 C29 (Nonacosane) 1.68 ±1.26 3.11 ±2.82 C30 (Triacontane) 0.88 ± 0.74 1.32 ±0.83 C31 (Hentriacontane) 2.29 ±1.75 4.04 ± 3.56 C32 (Dotriacontane) 0.82 ± 0.84 1.56 ±1.24 C33 (Tritriacontane) 1.20 ±1.34 2.59 ± 2.47 C34 (Tetratriacontane) 0.36 ± 0.67 1.37 ±1.53 CPI 2.32 ±1.19 1.94 ± 0.84 *Cwax 3.91 ±4.05 8.34 ± 9.06 ICn(25-33) 10.3 ±9.12 23.4 ±16.5 tBiogenic contribution (%) 37.9 ±15.4 31.1 ±14.4 .* Hopanes 18a(H)-22,29,30-trisnorneohopane 0.05 ±0.15 0.58 ± 0.49 17ct(H)-22,29,30-trisnorhopane 0.01 ± 0.05 0.59 ± 0.57 17a(H), 21(3(H)-29-norhopane 0.19 ±0.34 1.53 + 1.16 18a(H)-29-norneohopane 0.05 ±0.13 0.49 + 0.46 17a(H), 21B(H)-hopane 0.25 ± 0.39 2.13 ±1.63 22S-17a(H), 21B(H)-30-homohopane BDL 0.22 + 0.28 22R-17a(H), 21B(H)-30-homohopane 0.02 ± 0.09 0.30 + 0.37 Ill 22S-17a(H), 2ip(H)-30-bishomohopane 0.01 ±0.03 0.21 ±0.28 22R-17a(H), 21p(H)-30-bishomohopane BDL 0.21 ±0.31 n-Alkanoic acids (even-carbon-number) C8 (Octanoic acid) 0.4 ±0.6 1.2 ±1.4 C10 (Decanoic acid) 0.2 ±0.5 1.0 ± 1.1 C12 (Dodecanoic acid) 0.8 ± 0.6 2.4 ±1.4 C14 (Tetradecanoic acid) 1.1 ±0.9 4.8 ± 3.4 C16 (Hexadecanoic acid) 7.7 ±6.6 23.0 ±14.1 C18 (Octadecanoic acid) 7.7 ± 7.5 23.1 ±17.5 C20 (Eicosanoic acid) 0.2 ±0.5 0.4 + 0.8 C22 (Docosanoic acid) 0.3 ±0.9 0.6 ± 3.4 C24 (Tetracosanoic acid) BDL BDL Steroid Cholesterol BDL BDL Saccharides Levoglucosan 14.3 ±17.6 17.4 ±21.4 Glucose 2.07 ± 1.40 5.84 ± 3.30 Sucrose 5.97 ± 8.23 21.0 ± 25.8 Trehalose 0.69 ± 0.76 2.59 ± 2.02 Glycerol 0.52 ± 0.41 0.77 ± 0.47 Mannitol 0.21 ±0.18 0.87 ± 0.36 Arabitol 0.20 ± 0.22 0.63 ± 0.46 Erythritol 0.03 ± 0.07 0.04 ± 0.07 Xylitol 0.12 ±0.15 0.10 ±0.17 Ribitol BDL BDL Sorbitol BDL BDL Galactitol BIX BDL Note: *Cwax = f](Cn --^±±£=±1) (Gogou et al. 1996) n=25 2 tBiogenic contribution = Cwax/XCn(25-33) BDL: below detection limit 112 Figure 5.3 Correlation of ambient OC and EC concentrations in Higley PM2.5 and PM10 samples 7.0 • PM10 y = 2.32 x +2.35 6.0 2 • PM2.5 R = 0.52 — 5.0 to | 4.0 OJ •5- 3.0 O ° 2.0 y = 2.58 x + 0.73 1.0 R2 = 0.73 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 EC (ugC/m3) Figure 5.4 Distribution of n-alkane isomers showing odd-carbon enrichment as measured in Higley PM samples. CO £ 5.0 ^> c 4.5 c o 4.0 '& 2 3.5 *-» c 3.0 cu0) o 2.5 u c 2.0 !oo 1.5 E 1.0 A 0) D) m 0.5 0.0 ver ; < C25 C26 C27 C28 C29 C30 C31 C32 113 5.3.2 Ambient saccharides 5.3.2.1 Saccharide concentrations Ambient concentrations of individual saccharide compounds in both PM2.5 and PM10 samples were determined and summarized in Table 5.3. Concentrations of total measured saccharides in PM2.5 samples ranged from 1 to 83 ng m"3 with a mean concentration of 24 ng m"3, and in PM10 samples the range was from 6 to 121 ng m"3 with a mean concentration of 49 ng m"3. Consistent with the low levels of PM2.5 (5.9 ug m"3 on average) and PM10 (29.6 ug m"3 on average), the total measured saccharide concentrations in Higley aerosol samples were less than those reported in many other sampling locations worldwide (Table 5.4). Based on the measured mass of organic carbon, the carbon content of the total measured saccharides in Higley aerosols constituted 0.7% and 0.6% of the measured OC in PM2.5 and PM10, respectively (Table 5.10). Again, the carbon content of saccharide to OC ratios were lower than those found for aerosols collected in Hong Kong (Wan and Yu 2007), in the Amazon rainforest (Graham et al. 2003), and in the urban center of Ghent, Belgium (Pashynska et al. 2002), but were comparable to the numbers observed in aerosols collected at an urban site located in Dallas, TX (Jia 2008), and at four sites in Norway (Yttri et al. 2007) (Table 5.4). 114 Table 5.3 Ambient concentrations of saccharides in PM2.5 and PM10 samples at Higley, AZ (ng m"3) 1/16-1/23 1/24-1/31 2/1-2/8 2/9-2/16 2/17-2/24 2/25-3/3 3/4-3/11 3/12-3/19 3/20-3/27 3/28-4/4 4/5-4/12 4/13-4/19 PM2.5 a- + P-Glucose 2.07 1.78 1.01 2.71 2.55 3.99 2.01 2.18 1.26 1.99 2.12 0.88 Levoglucosan 18.3 25.1 40.5 29.7 30.2 16.0 10.2 1.58 0.66 2.53 2.62 2.27 Trehalose 0.91 0.17 0.21 0.48 0.26 2.01 0.96 0.71 0.81 0.70 0.79 0.27 Sucrose 2.32 0.66 0.43 0.83 2.18 18.8 5.99 11.3 14.3 5.21 4.74 1.75 Mannitol 0.53 0.25 0.15 0.26 0.09 0.49 0.25 0.19 0.07 0.13 0.18 0.07 Arabitol 0.68 0.19 0.11 0.25 0.20 0.50 0.27 0.27 0.07 0.05 0.05 0.01 Sorbitol 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Ribitol 0.00 0.00 0.00 0.00 0.00 0.00 0.02 0.07 0.00 0.00 0.00 0.00 Xylitol 0.00 0.10 0.05 0.08 0.24 0.17 0.15 0.32 0.15 0.02 0.05 0.00 Iso-Erythritol 0.14 0.04 0.01 0.02 0.01 0.14 0.01 0.02 0.00 0.00 0.02 0.00 Glycerol 0.72 1.06 0.41 0.86 0.78 0.65 0.58 0.25 0.09 0.35 0.44 0.11 Total 25.6 29.3 42.9 35.2 36.5 42.7 20.5 16.9 17.4 11.0 11.03 5.37 PM10 a- + B-Glucose 3.07 3.97 3.16 7.60 9.30 5.24 3.98 10.6 5.89 6.20 5.37 4.85 Levoglucosan 37.1 26.9 39.1 27.1 29.8 8.67 10.6 8.49 7.15 4.21 11.1 1.93 Trehalose 1.39 0.64 0.56 1.36 1.00 3.40 3.32 6.20 3.99 3.46 2.65 2.66 Sucrose 1.72 0.85 0.80 2.36 2.02 25.4 17.6 65.2 65.7 31.4 15.8 16.7 Mannitol 1.14 0.84 0.74 0.80 0.75 1.08 0.71 1.30 0.85 0.77 0.69 0.74 Arabitol 1.11 0.68 0.57 0.56 0.38 0.64 0.52 0.88 0.40 0.78 0.79 0.32 Sorbitol 0.02 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.01 0.01 0.01 Ribitol 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0.00 115 Xylitol 0.00 0.01 0.00 0.02 0.06 0.12 0.45 0.11 0.05 0.27 0.09 0.05 Iso-Erythritol 0.04 0.03 0.03 0.04 0.03 0.00 0.00 0.00 0.00 0.08 0.21 0.04 Glycerol 0.79 1.08 0.86 0.65 0.43 0.51 0.43 1.03 0.68 0.82 1.61 0.54 Total 46.4 35.0 45.8 40.5 43.8 45.0 37.6 93.8 84.7 48.0 38.3 27.9 116 Table 5.4 Comparison of total measured saccharide concentrations and the ratio of measured saccharides to OC reported in the literature. Sampling Concentration PM Size Location Period range (ng m"3) % of OC* Source PM2.5 Hong Kong Univ Campus Aug - next April 38-1316 1.3% (Wan and Yu 2007) Howland Experimental PM2.5 Forest, Maine May - Oct 10-180 - (Medeiros et al. 2006) PM2.5 Four sites in Norway Nov - Dec - 0.2 - 0.7% (Yttri et al. 2007) PM10 Four sites in Norway Nov - Dec - 0.5 - 2.7% (Yttri et al. 2007) PM<2.5 Amazonia, Brazil July 61 2.1% (Graham et al. 2003) PM>2.5 Amazonia, Brazil July 237 4.2% (Graham et al. 2003) PM10 Ghent, Belgium Winter 190-2800 3.6% (Pashynska et al. 2002) PM10 Ghent, Belgium Summer 390-1320 5.7% (Pashynska et al. 2002) PM10 Melpitz, Germany April - May > 56, <525 - (Carvalho et al. 2003) PM10 Hyytiala, Finland Aug > 3.8, <359 - (Carvalho et al. 2003) PM2.5 San Joaquin Valley, CA Dec - next Jan 144 - 3644 - (Nolteetal. 2001) PM2.5 San Augustine, TX Feb - June 19-355 1.1% (Jia 2008) PM2.5 Clarksville, TX Jan - April 7-372 1.8% (Jia 2008) PM2.5 Dallas, TX Jan - May 20-196 0.8% (Jia 2008) PM2.5 Higley, AZ Jan - April 1 -83 0.7% Current study PM10 Higley, AZ Jan - April 6-121 0.6% Current study * the carbon content of total measured saccharide concentration (ngC m") as a ratio of OC (ngC m") Of the measured saccharide compounds, levoglucosan was the most abundant species, constituting on average 46% and 35% of the total measured saccharide concentrations in PM2.5 and PM10, respectively. Levoglucosan levels were elevated in the first half of the sampling period, indicating the prevalence of biomass burning sources during the winter and early spring (Figure 5.5). Sucrose was the second most abundant saccharide compound, comprising on average 32% of the PM2.5 saccharide concentrations and 36% of the PM10. Sucrose is an important sugar in developing flower buds (Bieleski 1995) and a known component of pollen grains (Pacini 2000). The high aerosol sucrose measured between the end of February and the beginning of April (Figure 5.5) suggests this compound may originate from pollen 117 or pollen fragment entrainment into the atmosphere. Another sugar that had higher concentrations during this same period was trehalose, a primary saccharide serving as a reserve carbohydrate or stress protectant in bacteria, yeast, insects and a few higher plants (Caseiro et al. 2007). Previous studies have reported the enrichment of aerosol trehalose mainly in the fall (Medeiros et al. 2006; Wan and Yu 2007; Jia et al. 2010). However, samples in this study showed an increase of trehalose as early as March (Figure 5.5), indicating a unique influence from the desert heat and restricted water availability on the metabolic activities of micro-organisms and plants in Arizona. Another important sugar found in Higley PM samples was glucose, which contributed 12% and 16% to the measured total saccharide concentrations in PM2.5 and PM10. Concentrations of glucose did not display strong seasonable variation (Figure 5.5), indicating a uniform source of this most common simple sugar present in the atmosphere during the sampling period. Figure 5.5 Ambient concentrations of levoglucosan, glucose, sucrose and trehalose in Higley PM10 and PM2.5 samples during the sampling period (January - April, 2008). 118 Levoglucosan - Higley Glucose - Higley 20 n 18 —•—PM10 16 • PM2.5 14 - 0 « 1 10 c 8 J 6 > » 4 ' * a * a V«s* « ® 2 j»— 0 -eo to co eo » » co eo oo eo oooooooo o © © © o © o © g O © O © 1 8 sI a s " s B s Sucrose - Higley Trehalose - Higley Several sugar polyols were also detected in Higley PM2.5 and PM10 samples, including glycerol, mannitol, arabitol, sorbitol, ribitol, xylitol, and iso-erythritol. However, the concentrations of sugar polyols were low compared to other saccharide compounds (Table 5.3), and the total measured sugar polyols contributed only 3% and 5% to the total measured saccharide concentrations in PM2.5 and PM10 respectively. Similar to glucose, little seasonal variation was observed for sugar polyols, suggesting a relatively constant contribution from the fungi, lichens, and bacteria during this period (Eleutherio et al. 1993; Dahlman et al. 2003; Graham et al. 2003). The seasonal variation of saccharide compounds in PM2.5 has been reported in previous studies (Medeiros et al. 2006; Wan and Yu 2007; Yttri et al. 2007), and, in general, good agreement was found between the results of the current study and literature reports with a few exceptions. For example, in Hong Kong fine aerosol 119 samples, levoglucosan concentrations were enhanced in both fall and winter, and sucrose was measured at the highest concentrations in autumn as opposed to spring (Wan and Yu 2007). In PM2.5 samples at one suburban site in Norway, higher concentrations of glucose were measured in winter compared to summer (Yttri et al. 2007). These results can be interpreted as indicators of ecosystem function, climate and local sources. For example, the enhancement of levoglucosan measured in Hong Kong during fall and winter is a result of crop residue burning during the harvest season, and more incidents of hill fires during the drier weather of fall (Wan and Yu 2007). The maximum of sucrose in Hong Kong aerosols in fall instead of spring represents the flowering season of the Bauhinia flower which is locally widespread and starts to bloom in September. The glucose peak in winter aerosols in Norway is possibly related to fungal spores (Yttri et al. 2007) and other microbially degraded material. For these reasons, saccharides in aerosols can be used as tracers for the change of carbohydrate production and utilization in a local ecosystem across specific seasons. 5.3.2.2 PM10 vs. PM2.5 Several studies have shown that different saccharide compounds are enriched in different size ranges of ambient aerosols (Pashynska et al. 2002; Carvalho et al. 2003; Graham et al. 2003; Elbert et al. 2007; Fuzzi et al. 2007; Yttri et al. 2007). At the Higley sampling site, the analysis of saccharide concentrations in PM2.5 and PM10 samples collected simultaneously can track the enrichment of individual 120 saccharide compounds in fine and coarse particulate matter to elucidate local contributing sources. While saccharide compounds were generally measured at higher concentrations in PM10 samples compared to parallel PM2.5 samples (Figure 5.6), two distinct patterns in this relationship were observed. Levoglucosan and glycerol had nearly equivalent PM10 and PM2.5 concentrations while the remainder of the measured saccharide compounds (glucose, sucrose, trehalose, mannitol, arabitol, sorbitol, erythritol, ribitol and xylitol) had elevated PM10 relative to PM2.5 concentrations. These two different size fractionation patterns indicate that levoglucosan and glycerol are predominately associated with the fine fraction of PM while other saccharide compounds exist in both PM10 and PM2.5. This is consistent with levoglucosan being a tracer for wood smoke (Simoneit et al. 1999), with primary particle emissions in the submicron accumulation aerosol mode (Kleeman et al. 1999; Fuzzi et al. 2007; Pio et al. 2008). While the size distribution of aerosol glycerol primary emissions has not been documented before, Graham et al.(2002) reported a high correlation of aerosol glycerol with black carbon (BC), organic carbon (OC) and potassium (K) in particles collected at pasture and primary rainforest sites in Rondonia, Brazil, linking glycerol with combustion sources. Simoneit et al. (2004) also link some sugar polyols to the emission of particles by thermal stripping during wildfires. The enrichment of glucose, sucrose, trehalose, mannitol and arabitol in larger particles is consistent with the main contribution of these saccharides from PBAPs (Caseiro et al. 2007; Elbert et al. 2007; Bauer et al. 2008; Bauer et al. 2008; Deguillaume et al. 2008; Winiwarter et al. 2009), which have been reported to 121 contribute mainly to coarse mode aerosols (Cox and Wathes 1995; Taylor et al. 2002; Bauer et al. 2008; Deguillaume et al. 2008). This observation is consistent with interpretation of ambient saccharide measurements reported in the literature (Carvalho et al. 2003; Graham et al. 2003; Fuzzi et al. 2007; Yttri et al. 2007; Pio et al. 2008). Figure 5.6 Average concentration of saccharide compounds measured in PM2.5 and PM10 from samples collected in Arizona (Total number of samples analyzed n=45 for PM2.5 and n=46 for PM10). «S co 30 o L c E D PM2.5 O O) 25 ^ c E3PM10 n — 20 go- £•£ 15 O (0 U 0 a) 2 10 B> b to " C £ 0) u 5 I > o < n ^ r-^ w 0 1 • • ft® <* %** ^ 5.3.2.3 Correlation analysis To further investigate the correlation between saccharide compounds in the observational dataset and the implication of their different enrichment patterns in PM2.5 and PM10, Pearson's correlation coefficients were calculated for aerosol PM2.5 and PM10 samples with results shown in Table 5.5. 122 Table 5.5 Correlation matrix of the concentrations of major saccharide compounds in Higley PM2.5 and PM10 samples, level of significance a=0.05 (two-tailed test) Levoglucosan Glucose Sucrose Trehalose Glycerol Erythritol Arabitol Mannitol PM2.5 Levoglucosan 1 Glucose 0.17 1 Sucrose -0.23 0.63 1 Trehalose -0.15 0.72 0.84 1 Glycerol 0.54 0.46 -0.06 0.12 1 Erythritol 0.12 0.62 0.60 0.77 0.32 1 Arabitol 0.19 0.64 0.44 0.62 0.45 0.75 1 Mannitol 0.18 0.73 0.51 0.73 0.49 0.83 0.84 1 PM10 Levoglucosan 1 Glucose 0.10 1 Sucrose -0.33 0.42 1 Trehalose -0.33 0.38 0.82 1 Glycerol 0.20 0.20 0.14 0.15 1 Erythritol 0.09 0.08 -0.14 -0.07 0.53 1 Arabitol 0.14 0.21 0.16 0.36 0.53 0.39 1 Mannitol 0.01 0.25 0.30 0.44 0.25 0.03 0.77 1 Statistically significant correlations were found between mannitol and arabitol in both PM2.5 and PM10 (r=0.84 in PM2.5 and r=0.77 in PM10), and between trehalose and erythritol (r=0.77) as well as mannitol (r=0.73) and arabitol (r=0.62) in PM2.5. This correlation serves as a statistical support for trehalose and these polyols sharing a common source (Lewis and Smith 1967; Bieleski 1982). Trehalose was also highly correlated with sucrose in both PM2.5 and PM10 (r=0.84 in PM2.5 and r=0.82 in PM10). Although trehalose resembles sucrose both structurally and functionally as a carrier of energy and carbon (Wiemken 1990), these two sugars originate from different sources. Trehalose is widely present in bacteria, fungi, lower plants, and invertebrates (Bieleski 1982) serving as a cell protectant against stressful conditions 123 (Eleutherio et al. 1993; Wong et al. 1993; Chaturvedi et al. 1997). Sucrose, on the other hand, is more prevalent in the phloem of plants and has been found to be enriched in pollen grains (Bieleski 1995; Pacini 2000). The statistically significant correlation found between trehalose and sucrose in Higley PM samples may result from the onset of pollen production concurring with increased microbial metabolism in the spring under desert heat and restricted water availability conditions in Arizona. The dominant pollen types found in Arizona originate from ragweed, mulberry trees, cypress, and mormon tea, all of which have their peak pollen releases in March and April (O'Rourke 1990). March and April also typically represents a local dry season in Arizona, between winter frontal and summer monsoonal rains, and this seasonal dry period may correspond to trehalose being produced as a cell protectant against desiccation for microorganisms. Glucose was correlated with most sugars and polyols to varying degrees. Because glucose is a fundamental simple sugar in biology, all other sugars and polyols can be considered derivatives as glucose serves as the primary carbon source for plants and microorganisms (Pigman and Horton 1980). These correlations, between trehalose and the sugar polyols and glucose and most other saccharides have been reported for both aerosol and soil samples (Medeiros et al. 2006; Wan and Yu 2006). No consistent correlation was observed between levoglucosan and the other measured saccharides except for fine particle glycerol. This is consistent with the earlier interpretation that aerosol glycerol may be linked to biomass combustion sources. 124 In general, the correlation between saccharide compounds was consistent between the analysis of PM2.5 and PM10 samples collected at the Higley site. However, all correlations between saccharides were stronger in interpretation of PM2.5 data compared to PM10 data. While no direct comparison of this statistical observation has been reported in the literature, results from several studies on the size distribution of aerosol sugars and sugar polyols can be used to corroborate this observation. Yttri et al. (2007) found that the majority of the sugars and the sugar polyols measured at four sites in Norway were predominately in aerosols in the size range 0.25-1.0 um during winter, and in aerosols larger than 2 um in summer. This was attributed to snow on the ground which effectively reduced resuspension of decaying biogenic material (coarse particles) from the local soils in winter. In general, this is in agreement with data of Carvalho et al. (2003), which reported a varying maximum of the aerosol saccharide size distribution between locations. In recognition of these earlier results, the stronger correlations found between saccharides in PM2.5 compared to PM10 may have originated from lower temperatures and atmospheric stagnation during the Arizona sampling campaign that prevented resuspension of PBAP into the atmosphere. To corroborate this, correlations on the first and second halves of the ambient data did show a greater degree of correlation for PM10 saccharides in the second half of the data (March - April) and greater degree of correlation for PM2.5 saccharides in the first half of the data (January - February) (see Table 5.6). This is consistent with biomass combustion 125 dominating the first half of the sampling period (fine aerosol source) and soil resuspension dominating the second half (coarse aerosol source). Table 5.6 Correlation coefficients for measured saccharides in PM10 and PM2.5 during different sampling periods at Higley, level of significance a=0.05 (two-tailed test) Levoglucosan Glucose Sucrose Trehalose Glycerol Erythritol Arabitol Mannitol PM2.5 (Jan - Feb) Levoglucosan 1 Glucose -0.01 1 Sucrose -0.22 0.72 1 Trehalose -0.25 0.79 0.94 1 Glycerol 0.27 0.32 0.00 0.08 1 Erythritol -0.10 0.64 0.81 0.88 0.16 1 Arabitol -0.16 0.62 0.58 0.71 0.25 0.79 1 Mannitol -0.15 0.64 0.69 0.81 0.26 0.85 0.83 1 PM2.5 (Mar - Apr) Levoglucosan 1 Glucose 0.23 1 Sucrose 0.01 0.66 1 Trehalose 0.37 0.69 0.70 1 Glycerol 0.65 0.60 0.12 0.39 1 Erythritol 0.16 0.77 0.51 0.50 0.50 1 Arabitol 0.52 0.59 0.53 0.58 0.46 0.68 1 Mannitol 0.46 0.88 0.52 0.74 0.71 0.82 0.78 1 PM10 (Jan - Feb) Levoglucosan 1 Glucose 0.20 1 Sucrose -0.22 0.21 1 Trehalose -0.13 0.37 0.79 1 Glycerol 0.28 -0.09 -0.08 -0.01 1 Erythritol 0.56 0.41 -0.14 0.18 0.40 1 Arabitol 0.07 0.03 0.15 0.53 0.42 0.65 1 Mannitol -0.07 -0.01 0.23 0.52 0.13 0.38 0.80 1 126 PMIO(Mar-Apr) Levoglucosan 1 Glucose 0.15 1 Sucrose 0.20 0.69 1 Trehalose 0.29 0.51 0.71 1 Glycerol 0.52 0.42 0.16 0.19 1 Erythritol 0.02 -0.05 -0.32 -0.28 0.59 1 Arabitol 0.50 0.43 0.33 0.56 0.65 0.38 1 Mannitol 0.30 0.60 0.61 0.72 0.37 -0.10 0.73 5.3.3 Saccharides in resuspended soil samples In light of recent suggestions that saccharides are potential organic tracers for the entrainment of soil into the atmosphere (Simoneit et al. 2004; Rogge et al. 2007), 28 sets of soil samples (including currently uncultivated soil, agricultural soil, and road dust samples) were collected from locations near the ambient aerosol sampling site for source characterization. Each soil sample was resuspended and size-segregated to represent material entrained as PM10 and PM2.5. Table 5.7 summarizes the range and average concentrations of saccharide compounds in fine and coarse fractions for each of the three types of soil samples collected in January and April. Please note that the coarse fraction here and in discussions to follow refers to soil particles equivalent to ambient PM10. 127 Table 5.7 The range and average concentration of saccharide compounds in resuspended and size-selected soil samples at Higley, AZ (reported as ng mg"1) PM10 PM10 PM10 PM2.5 PM2.5 PM2.5 Compound Currently Currently Agricultural soil (7) uncultivated soil (3) Road dust (4) Agricultural soil (7) uncultivated soil (3) Road dust (4) January Glucose 1.4-215(35.0) 0.9-11.2(4.5) 0.6-12.5(5.0) nd - 5.9 (2.6) nd-1.1 (0.4) 0.6-5.5 (2.5) Sucrose nd - 8.6 (2.8) 0.9-1.6(1.2) nd - 7.9 (3.3) nd-1.3 (0.8) nd-2.8 (1.0) nd-1.1 (0.7) Trehalose 3.5-29.4(14.3) 1.6-41.5(21.7) nd-1815 (56.6) 0.9-36.1 (9.5) nd - 54.4 (20.3) nd - 14.0 (7.0) Mannitol nd-1.7 (0.4) nd - 0.5 (0.2) nd - 3.8 (0.9) nd-2.8 (1.3) nd - 2.4 (0.8) nd-1.2 (0.3) Arabitol nd-11.8 (2.1) nd-2.7 (1.2) nd-7.2 (1.8) nd-6.5 (1.6) Sorbitol nd-11.4 (3.4) nd-1.1 (0.4) nd - 6.7 (2.2) nd-0.4 (0.1) Erythritol Xylitol nd-2.1 (0.7) nd-9.1 (2.3) Ribitol nd-1.0 (0.3) Galactitol nd-1.0 (0.3) Glycerol nd-124 (34.2) nd - 65.6 (36.6) nd-44.1 (17.0) nd-23.0 (10.4) nd - 65.9 (22.0) nd - 27.6 (6.9) Levoglucosan nd - 0.9 (0.2) Total measured sugar polyol nd-18.9 (6.0) nd-4.8 (2.4) nd-22.1 (5.5) nd-15.0 (5.1) nd-2.4 (0.8) nd-1.6 (0.4) (excluding glycerol) Total measured saccharides 7.2-268(92.3) 3.5-112(66.5) 8.6 - 200 (87.7) 6.3-79.0(28.4) nd-126 (44.4) 6.5-30.3(17.4) April Glucose nd - 23.4 (6.8) 0.90-3.68(2.19) 12.3-212(73.9) nd-8.0 (3.1) nd - 9.2 (3.3) 18.4-75.8(43.6) Sucrose nd-12.3 (3.7) 0.76-2.80(1.44) 6.7 - 82.9 (29.9) nd-5.0 (1.6) nd - 7.9 (2.7) 6.4-119(41.5) Trehalose 3.8 - 223 (48.2) 2.7 - 288 (99.6) 16.3-358(141) 0.5-237(42.6) 0.8 - 229 (76.9) 13.8-415(125) Mannitol nd-9.1 (1.7) nd - 2.4 (0.8) 3.5-13.4(6.3) nd - 2.0 (0.6) nd-4.1 (1.4) 0.8-3.9(1.8) Arabitol nd- 6.2 (1.9) nd-4.4 (1.5) nd - 5.7 (2.4) nd-3.8 (1.5) nd - 3.9 (2.0) Sorbitol nd-9.8 (1.8) nd-3.2 (1.1) nd - 0.4 (0.2) nd-7.1 (1.1) nd-1.6 (0.5) nd-1.1 (0.4) Erythritol nd - 2.5 (0.6) nd-0.8 (0.2) Xylitol nd - 2.6 (0.4) nd - 0.9 (0.2) nd - 6.4 (0.9) nd-1.8 (0.4) Ribitol nd-1.3 (0.8) nd - 0.7 (0.3) Galactitol 0.65 ±1.73 nd-1.3 (0.3) nd-0.4 (0.1) Glycerol nd - 59.9 (24.7) nd-39.2 (15.4) 19.6-38.5(28.3) nd - 26.8 (9.9) nd-95.2 (31.7) nd-24.2 (13.4) Levoglucosan nd-0.6 (0.1) Total measured sugar polyol nd - 32.3 (6.5) nd-10.1 (3.4) 4.7-25.4(10.9) nd-9.2 (4.1) nd-5.7 (1.9) 0.8 - 9.3 (5.3) (excluding glycerol) Total measured saccharides 10.7-317(89.9) 10.4-343(122) 62.2 - 522 (284) 1.3-283(61.4) 0.8-347(117) 51.1 -623(228) Table 5.8 The range and average carbon content of saccharide compounds relative to organic carbon in resuspended and size-selected soil samples at Higley, AZ (reported as %) PM10 PM10 PM10 PM2.5 PM2.5 PM2.5 Compound Currently uncultivated Currently Agricultural soil (7) soil (3) Road dust (4) Agricultural soil (7) uncultivated soil (3) Road dust (4) January Glucose 0.004-0.80(0.15) nd - 0.008 (0.004) nd - 0.01 (0.004) nd - 0.04 (0.03) nd - 0.02 (0.008) nd - 0.01 (0.008) Sucrose nd-0.03 (0.01) nd - 0.004 (0.004) nd - 0.02 (0.008) nd - 0.03 (0.008) Trehalose 0.02-0.17(0.06) 0.004-0.20(0.10) nd - 0.07 (0.03) 0.02-0.31 (0.10) nd - 0.63 (0.25) nd - 0.04 (0.02) Glycerol nd-0.67 (0.16) nd-0.29 (0.14) nd-0.03 (0.01) nd-0.18 (0.08) nd - 0.71 (0.24) nd-0.14 (0.04) Total measured sugar polyol (excluding glycerol) nd - 0.06 (0.02) nd - 0.01 (0.006) nd - 0.02 (0.004) nd-0.12 (0.03) nd-0.03 (0.01) Total measured sugar and sugar polyol 0.03 - 0.99 (0.40) 0.005 - 0.51 (0.26) 0.02 - 0.09 (0.05) 0.10-0.66(0.24) 0.004-1.40(0.51) 0.01 -0.15(0.06) April Glucose 0.004 - 0.26 (0.06) 0.004 - 0.01 (0.008) 0.01 - 0.08 (0.03) 0.004-0.12(0.03) nd - 0.11 (0.04) 0.01 -0.24(0.14) Sucrose nd-0.14 (0.03) 0.004-0.008(0.004) 0.004-0.02(0.01) nd-0.003 (0.01) nd-0.10 (0.03) 0.008-0.52(0.16) Trehalose 0.03 - 0.27 (0.13) 0.01 -1.01 (0.36) 0.02-0.11 (0.06) 0.04 - 0.80 (0.24) 0.008 - 2.88 (0.97) 0.008-1.83(0.52) Glycerol nd - 0.49 (0.10) nd - 0.13 (0.06) 0.008 - 0.02 (0.02) 0.04-0.16(0.08) nd- 1.11 (0.37) nd-0.10 (0.05) Total measured sugar polyol (excluding glycerol) nd - 0.09 (0.02) nd - 0.03 (0.01) nd - 0.009 (0.006) 0.004-0.10(0.04) nd - 0.07 (0.02) nd - 0.03 (0.02) Total measured sugar and sugar polyol 0.06 -1.25 (0.34) 0.05 -1.19 (0.43) 0.08-0.17(0.13) 0.11 -0.94(0.41) 0.01 -4.27(1.43) 0.02 - 2.73 (0.89) 130 5.3.3.1 Total measured saccharides and seasonal change The average total measured saccharides ranged from 17 ug g"1 (fine road dust samples in January) to 284 ug g"1 (coarse road dust samples in April) for all soil samples (Table 5.7). For all samples, the carbon content of measured saccharides constituted from 0.05% (fine and coarse road dust samples in January) to 1.4% (fine currently uncultivated soil samples in April) of the measured organic carbon (Table 5.8). There have been several prior literature reports on the saccharide content of soils at different locations, and results from these studies are summarized in Table 5.9. Note that all of these results reported in the literature were based primarily on analysis of bulk soil samples with particle diameter less than 0.6 mm, and not for size selected soil samples. For comparison, additional samples of the current soil sampled with particle diameter less than 0.6 mm were analyzed following the procedure described by Medeiros et al. (2006). This characterization showed less enrichment of saccharides in bulk soil when compared to aerosol size selective soil resuspension samples by a factor of 2-8. If this increased enrichment is incorporated into the data interpretation, the total saccharide levels measured in Higley soil samples were within the range of those reported in different soil types at various locations elsewhere (Simoneit et al. 2004; Medeiros et al. 2006; Rogge et al. 2007; Jia 2008). 131 Table 5.9 Concentrations of major saccharides in soil and dust samples in the literature. Note that concentrations are given as \ig g" (dry weight of bulk soil with particle diameter <0.6mm). Ryegrass Safflower Almond Grape Unpaved Paved road soil8, Native soil", Native soilb, Native soil", Amazonia Cotton field" Tomato field0 field' orchard" orchard0 road dust0 dust0 Willamette San Augustine, Clarksville, Hempstead, soil0 San Joaquin San Joaquin San Joaquin San Joaquin San Joaquin San Joaquin San Joaquin Valley, TX TX TX Brazil Valley, CA Valley, CA Valley, CA Valley, CA Valley, CA Valley, CA Valley, CA Oregon (temperate, (temperate, (temperate, (temperate, (temperate, (temperate, (temperate, (temperate, (temperate) (tropical) (temperate) (temperate) summer) summer) summer) summer) summer) summer) summer) summer) 0.14-0.69 0.16-7.58 0.45-1.94 0.28-1.71 0.02-0.27 0.47-0.72 0.05-4.8 2.9-93.3 Glucose 0.1-2.2 (0.43) (1.63) (0.91) 43.4 (1.10) 0.3-4.2 (2.3) 0.5-1.3(0.6) (0.15) (0.6) (0.9) (48.1) 0.07-0.21 0.08-0.39 0.27-10.7 0.84-1.78 0.01-1.14 0.19-0.77 0.34-10.7 Sucrose 0.14-2.9 (0.16) (0.20) (3.64) 3.9 (1.31) 0.9-2.8(1.9) (0.4) (0.48) 0.3-0.6 (0.4) (2.8) 7.6-7.7 (7.6) 11.8-43.4 1.44-56.5 43.3-59.9 5.80-6.01 21.7-38.7 0.37-4.9 1.03-7.52 6.2-19.0 Trehalose 3.5-19.8 (24.8) (29.4) (50.2) 48.2 (5.91) (30.2) (2.8) (4.3) 3.4-6.8(5.1) 3.3-7.2 (5.3) (12.6) 0.22-0.47 nd-0.94 nd-0.18 Mannitol 0.005-0.13 (0.32) (0.17) (0.08) 0.17-1.10 0.01-0.13 Arabitol (0.55) (0.09) 0.01-0.03 0.01-0.10 Sorbitol (0.03) (0.04) 21.7 1.4-5.3(3.4) 0.01-1.18 nd-0.03 Erythritol (0.32) (0.01) nd-0.43 Xylitol (0.22) 132 0.10-0.65 0.01-0.13 Ribitol (0.30) (0.05) Galactitol 0.26-1.08 0.28-2.23 0.79-3.98 0.07-0.28 0.02-0.28 0.02-0.21 0.23-0.30 0.06-0.19 0.73-1.25 Glycerol (0.58) (1.08) (2.44) 16.9 (0.18) 3.2-3.7 (3.4) (0.11) (0.12) (0.26) (0.1) (0.99) 0.02-0.06 0.02-0.36 0.01-0.12 nd-0.21 Levoglucosan (0.04) (0.07) (0.07) (0.11) 13.21-48.0 2.32-60.8 46.1-76.2 Total mean 3.8-23.7 (27.5) (32.7) (57.3) 167 8.9 41.2 3.91 5.05 6.36 9.1 85.3 Note:a Medeiros et al., 2006b. b Jia 2008 c Simoneit et al., 2004 133 In general, higher soil saccharide concentrations were measured in April than in January reflecting the increased plant growth and microbial activities during this period (Medeiros et al. 2006). Measured saccharide concentrations increased consistently in both the coarse and fine fraction of the resuspended soil samples from January to April. The only exception to this trend was the coarse agricultural soil in which a slight decrease of total saccharides was observed. Only slight increases were observed in the total measured saccharide to organic carbon ratio from January to April, except in the fine fraction of currently uncultivated soil and road dust samples, where the increases were significant (Table 5.8). This observation indicates the dominant role of other organics in determining the organic content of soil, particularly in the coarse fraction, and candidates include fatty acids, biopolymers like proteins, cellulose, and other unidentified organic components from fragmented primary biological structures such as fungal spores, pollen, algae, bacteria, leaf and insect parts (Graham et al. 2003). Most of the observed increases in soil saccharide content between the January soil sampling to the April soil sampling in both the total measured saccharide concentration and the saccharide to OC content can be attributed to the single disaccharide trehalose. Trehalose has previously been reported in many studies as the most abundant sugar in soil (Simoneit et al. 2004; Medeiros et al. 2006; Rogge et al. 2006; 2007), and trehalose comprised on average about 40% to 70% of the total measured saccharides in Higley soil samples, with the greatest enrichment found in currently uncultivated soil. A slightly higher percentage of trehalose was measured in 134 the fine soil samples compared to the coarse portion of the soils, which may imply the re-entrainment of this fungal metabolite released by primary biological particles into the soil or production of this stress protectant by the soil microbial community, occurs preferentially in the fine size range. The second most abundant saccharide in soils was glycerol, which contributed on average between 15% and 32% of the total measured saccharides. Glycerol in soils was found to be more abundant in January than in April and comprised a higher percentage of the total measured saccharides in the fine soils than in the coarse. These two characteristics of glycerol are consistent with the association of glycerol with biomass burning and locally may represent the burning of crops and vegetation residue in the fields (Graham et al. 2002; Jia 2008). Glucose was the third most dominant saccharide, followed by sucrose, and each contributed on average from 9% to 23% and from 4% to 12% to the total measured saccharides, respectively. No uniform enrichment pattern of glucose or sucrose was found in either the fine or coarse soils, but glucose and sucrose did appear to be enriched in road dust samples as opposed to agricultural soil and currently uncultivated soil samples. Given glucose and sucrose are known as the main components of pollen grains used to maintain water content and pollen viability (Pacini et al. 2006), this may be attributable to pollen grains and fragment enrichment in road dust particles. Sugar polyols (excluding glycerol) were minor saccharide components in Higley soils, contributing from less than 1% up to 16% of the total measured saccharides, with sorbitol, arabitol, and mannitol the dominant polyol species. Sorbitol, a sugar polyol occurring in higher woody plants but not commonly found in lower plants (Wallaart 135 1980), was mostly measured in agricultural soils. By comparison, mannitol and arabitol, two sugar polyols that originate from fungal spores (Bauer et al. 2008), were found in higher concentrations in road dust samples. Finally, the wood smoke tracer levoglucosan, unlike its prevalence in aerosols, was detected in only trace levels in all soil and road dust samples from Higley. 5.3.3.2 Agricultural soil vs. currently uncultivated soil vs. road dust Saccharide concentrations varied between different soil types and this variation could potentially be used to separate entrainment of different types of soil to the atmosphere. Agricultural and currently uncultivated soils had similar saccharide compositions, except that glucose was more enriched in agricultural soils and trehalose was more abundant in currently uncultivated soils. This may originate from different saccharide production and utilization mechanisms in these two different types of soils. In agricultural soils, glucose is likely released from the decomposition of plant debris or the addition of manures or other organic material to increase the agricultural productivity of the soil. In currently uncultivated desert soils, where fewer plants grow, soil microbes are more inclined to suffer from heat and desiccation, and as a result, enhanced trehalose levels may indicate a mechanism to maintain the structural integrity of the cytoplasm in stressed microbes (Wiemken 1990). Road dust samples contained the highest levels of saccharides, especially in April, with major saccharides including glucose, sucrose, trehalose, mannitol and arabitol all measured at higher concentrations in road dust than soil. Similarly, higher levels of saccharides 136 have been measured in local paved road dust samples compared to other local soil samples in a study in the San Joaquin Valley, CA by Rogge et al. (2007). This may indicate that wind-blown primary biological particles, including pollen grains, fungal spores, yeasts, bacteria, along with their fragments, are the original sources of saccharides measured in road dust, and that primary biological particles are blown off their host bodies (i.e., plants, fungi, yeasts, etc) by wind and this biological material deposited on the road surface. 5.3.3.3 Fine vs. coarse soils In January, higher concentrations of saccharides were measured in the coarse soil fraction compared to the fine fraction for all soil and dust types. In April, this same pattern was found for most samples, with the exception of road dust samples where higher saccharide content for some major saccharide species (sucrose, trehalose) was found in the fine fraction (Table 5.7). When considering the increase in saccharide concentration from January to April, this might indicate the enhanced entrainment of saccharides in the fine mode possibly contributed from the fragmented pollen grains and fungal spores in road dust samples (Pacini et al. 2006; Bauer et al. 2008). Taylor et al. (2002; 2004) reported that ruptured grass pollens caused by a cycle of wetting, drying and wind shedding were measured in the size range of 30 nm - 4.7 um. Gorny et al. (2002; 2003) also reported that large quantities of sub-micronic fungal and actinomycete fragments are released together with spores from moldy building surfaces, and the number of released fragments was greater than the number 137 of intact spores released from moldy surfaces. Considering this, it is likely that the primary biological particles not only contribute to the global loading of coarse particulate matter as has been currently acknowledged (Graham et al. 2003; Bauer et al. 2008; Bauer et al. 2008; Winiwarter et al. 2009), but also contribute significantly to the fine particulate matter as well (Puxbaum and Tenze-Kunit 2003; Jaenicke 2005; Elbert et al. 2007). During stagnant atmospheric conditions and the weather unfavorable to the resuspension of coarse particles, the role of fine primary biological particles may be more clearly observed. 5.3.4 Ambient saccharides vs. soil saccharides Saccharide content in ambient PM samples and soil samples can be compared on a mass to mass basis by dividing the ambient aerosol saccharide concentration by the corresponding measured PM mass. Saccharide concentrations in soils have already been presented as a ratio of saccharide mass to the corresponding resuspended soil mass. Because soil is not regarded as a potential source for levoglucosan in aerosols and very little levoglucosan was quantified in soil samples, the comparison in the following discussion will exclude this compound. The total saccharide content (excluding levoglucosan) relative to PM mass in ambient aerosol samples is 0.20% for PM2.5 and 0.11% for PM10. When compared to a ratio of saccharide to aerosol mass of 0.01% and 0.02% for the fine and coarse fraction of road dust (which was the highest saccharide content level found in three types of soil and dust samples as reported in Table 5.10), it is clear that the overall 138 saccharide content was much higher in ambient aerosols than in the resuspended and size-segregated soil samples. This was unexpected because prior literature reports propose that soils and associated microbiota are a major source contributing to sugars and polyols in ambient PM (Simoneit et al. 2004). If the same comparison is based on saccharide to OC content, the saccharide to OC ratio found for soil samples (Table 5.10) is still not much higher than the corresponding value for ambient aerosol samples. This indicates there are likely other dominant sources contributing to sugars and polyols in aerosols at Higley in addition to soils and dust during the current study period. Primary sources as proposed in the literature could include PBAPs originating from pollen grains, fungal spores, lichens, yeast, bacteria, to leaf fragmentation and other biogenic sources (Bauer et al. 2002; Graham et al. 2002; Pashynska et al. 2002; Yttri et al. 2007; Bauer et al. 2008; Deguillaume et al. 2008). These primary sources become incorporated into the soil system and re-entrained through fugitive dust. The content of PBAPs have been measured in many studies, with the concentration of trehalose in stationary vegetative yeast as high as 5-10% of the dry weight and up to 20% of the dry weight in spores of the fungus Myrothecium verrucaria (Mandels et al. 1965; Dey and Harborne 1997). Mature pollen is generally endowed with a high content of sucrose, ranging from 7 to 23% of the pollen dry weight (Black and Pritchard 2002). Conidia of A. niger, a very common fugal species found throughout much of the world, contains mannitol representing 10.9% of dry weight. Based on a range of 5% to 20% dry weight for trehalose, sucrose, and mannitol in PBAPs such as yeasts, fungal spores, or pollens, and an assumed 50% of 139 water inside these primary biological aerosol (PBA) (Bidochka et al. 1990; Dijksterhuis and Samson 2007), one could estimate 2.5% to 10% of primary biological aerosols are saccharides on a wet weight basis. This estimated range is well above the total saccharide content (excluding levoglucosan) relative to PM mass or organic carbon in the ambient aerosols collected at Higley, and confirms the likelihood of PBAP as a primary source to aerosol saccharides. To further investigate this hypothesis, the characterization of PBAPs based on saccharide composition is presented in Chapter 6. Table 5.10 Total saccharide content (excluding levoglucosan) relative to PM mass and total carbon content of saccharide (excluding levoglucosan) relative to OC in Higley ambient aerosols and dust and soil samples. Total saccharide content (excluding Total carbon content of saccharide levoglucosan) relative to PM mass (excluding levoglucosan) relative to OC PM2.5 PM10 PM2.5 PM10 Ambient PM 0.20 ± 0.25% 0.11 ±0.09% 0.26 ± 0.27% 0.37 ± 0.37% Agricultural Soil 0.004 ±0.007% 0.009 ±0.010% 0.32 ±0.27% 0.37 ± 0.43% Currently uncultivated Soil 0.008 ± 0.014% 0.009 ±0.013% 0.97 ±1.70% 0.36 ±0.51% Road Dust 0.012 ±0.020% 0.019 ±0.018% 0.47 ±0.93% 0.09 ± 0.05% Further investigating into the fine and coarse fraction of aerosol and soil samples, the data show a higher saccharide percentage relative to PM mass (i.e., 140 saccharide content) in ambient PM2.5 than PM10 for all compounds without exception (Table 5.11). However, in most soil samples, the corresponding percentage of saccharides to PM mass was higher in the coarse fractions (Table 5.11). If soils and dust are one of the sources that contribute to sugars and polyols in aerosols, and PBAP is taken as a separate source of coarse particles (Madelin and Madelin 1995; Graham et al. 2003; Elbert et al. 2007; Winiwarter et al. 2009), then the data here may suggest these sources undergo a series of physical processes (i.e. resuspension, wetting, drying) which separate coarse particles into numerous fine components to become airborne (Taylor et al. 2002; Taylor et al. 2004). Separately there may be a significant contribution from fine material in soil or PBA particles that are mostly of microbiological origin that have previously been underestimated (Gorny et al. 2002; Gorny et al. 2003; Jaenicke 2005). In a relevant comparison, Puxbaum and Tenze-Kunit (2003) unexpectedly measured a large fraction of atmospheric cellulose in fine aerosol fractions (< 1.6 urn AD) at a downtown site in Vienna. It was proposed that the large amount of leaf litter generated fine particles during the biological decomposition process involving microorganisms and other decomposers. Although atmospheric cellulose is not equivalent to atmospheric saccharides, the same speculation could explain the enrichment of saccharides in fine aerosols. Alternatively, unfavorable weather conditions (i.e., lower temperatures, air stagnation, low relative humidity) between January and April may have prevented the fragmentation of coarse soil and PBA particles and subsequent resuspension into the atmosphere. In short, a larger fraction of non-levoglucosan saccharides in the fine Higley aerosols (compared 141 to coarse aerosols) was not expected, and could not be explained by a comparison with the local soil samples and the current knowledge of PBAP in the coarse fraction. This suggests the contributions in the fine size fraction from biogenic sources may be greater than previously acknowledged (Jaenicke 2005). Table 5.11 Saccharide content relative to PM mass in Higley ambient and soil samples (reported as 10"3%) Ambient samples Agricultural soil Currently uncultivated soil Road dust PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 Levoglucosan 1-1208 (259) 2-418(75) Glucose 2-178(42) 5 - 84 (22) nd - 0.8 (0.3) nd - 22 (2) nd - 0.9 (0.2) 0.1 - 1 (0.3) 0.1 -8(2) 0.1-21 (4) Sucrose nd-1123 (122) 1 - 346 (68) nd-0.5 (0.1) nd -1 (0. 3) nd - 0.8 (0.2) 0.1-0.3(0.1) nd -12 (2) nd - 8 (2) Trehalose nd-122(14) nd - 29 (9) 0.1-24(3) 0.3 - 22 (3) nd - 23 (5) 0.2 - 29 (6) nd - 42 (7) nd-36(10) Mannitol nd - 24 (4) 1 -15(3) nd-0.3 (0.1) nd-0.9 (0.1) nd-0.4 (0.1) nd-0.2 (0.1) nd-0.4 (0.1) nd -1 (0.4) Arabitol nd - 29 (4) nd - 8 (2) nd - 0.6 (0.2) nd -1 (0.2) nd-0.4 (0.1) nd-0.4 (0.1) nd - 0.7 (0.2) Glycerol nd-69(11) nd - 8 (3) nd-3(1) nd-12(3) nd - 1 (0.3) nd - 7 (3) nd-3(1) nd - 4 (2) Total sugar polyols (excluding glycerol) nd-30(12) nd-20(4) nd - 2 (0.5) nd-3(1) nd-0.6 (0.1) nd - 1 (0.3) nd-0.9 (0.3) nd - 3 (0.8) Total saccharides 20-2040(460) 30-550(180) Total saccharides (excluding levoglucosan) 10-1500(200) 20-440(110) 1 - 28 (5) 0.7-32(1) nd - 35 (8) 0.3 - 34 (9) 0.6-62(12) 0.8-52(19) 143 5.3.5 Source Apportionment using PMF An investigation into the sources of measured aerosols at Higley was conducted using the EPA PMF1.1 model. Details of the PMF theory and the EPA model have been described in other chapters. The marker compounds used in this model application included OC and EC, n-alkanes (C25-C32), hopanes (17a(H)-21 p(H)-hopane, 17a(H)-21 p(H)-29-norhopane, 22S, 17a(H),21 P(H)-30-homohopane, and 22R, 17a(H),21 p(H)-3O-homohopane), n-alkanoic acids (hexadecanoic acid and octadecanoic acid), saccharides (levoglucosan, glucose, sucrose, trehalose, glycerol, arabitol, and mannitol), and ions (sodium, magnesium, sulfate, nitrate, and chloride). A total of 39 sample days were available for the source apportionment of PM2.5 and 43 sample days for PM10. 5.3.5.1 PMF-resolved source factors Based on the ambient concentrations of the marker compounds, PMF was able to resolve eight common factors that contribute to both PM2.5 and PM10 at Higley, with an additional factor isolated only from the PM10 data. The first common factor is dominated by hopane species (Figure 5.7), with additional enrichment of n-alkanes ( The second factor is dominated by the n-alkane series ranging from C25 to C32 (Figure 5.7). The profile resolved for PM2.5 has an equal distribution of the 144 even- and odd- carbon numbered alkanes, while the profile for PM10 has a slightly enrichment of the even-carbon-number alkanes. The source of this factor is consistent with petroleum wax rather than biogenic (leaf) wax (Simoneit 1984; Standley and Simoneit 1987; Rogge et al. 1993), and is labeled n-alkane wax. The third factor was resolved with the dominance of palmitic acid (or n-hexadecanoic acid) and stearic acid (or n-octadecanoic acid) (Figure 5.7). In numerous prior studies, these two n-alkanoic acids have been associated with meat cooking (Rogge et al. 1991; Schauer et al. 2002). Factors four and five are characterized by the enrichment of ion species (Figure 5.7). One factor contains ions commonly associated with sea salt (Ramadan et al. 2000). However, given the distance of the Higley site from the ocean, it was likely that the chloride in this factor was also formed by the photochemical reaction of anthropogenically emitted molecular chlorine with hydrocarbons and excessive ammonia present in the atmosphere (Seinfeld and Pandis 1998; Chang and Allen 2006). This factor is labeled as a salt factor. The other factor is dominated by sulfate and nitrate, with enrichment of Mg, Na, and OC. This factor is consistent with, and labeled as, secondary sulfate and nitrate. Factors sixth, seven, and eight which are common to both PM10 and PM2.5 are separated based on saccharide compounds (Figure 5.7). The factor dominated by levoglucosan is labeled as a wood smoke factor (Simoneit et al. 1999). The factor characterized by the enrichment of sucrose and trehalose is labeled as a sucrose and trehalose rich bio-aerosol factor. Sources contributing to this factor are likely pollens 145 (Pacini et al. 2006), bacteria, fungi, lichens, and lower plants (Lewis and Smith 1967; Bieleski 1982; Feofilova et al. 2000; Tereshina et al. 2000). The association of sucrose and trehalose in the same factor is a reflection of the high correlation between these two compounds as discussed earlier in Section 5.3.2.3. The final factor common to PM2.5 and PM10 is resolved with the enrichment of mannitol, arabitol, glycerol, and glucose to different extents, therefore is labeled as glucose and sugar polyol rich bio-aerosol. Since mannitol and arabitol have been used as markers for fungal spores (Bauer et al. 2008) and glycerol and glucose are also known constituents of spores (Lewis and Smith 1967), this factor could be of fungal spore derived origin. Again, it is worth noting that these last two factors could be mixed effects of PBAPs and other biologically derived carbohydrate sources from soil and associated biota. Finally, a factor that was unique to PM10 is identified with high percentages of Na, Mg, and odd-carbon numbered alkanes. Na and Mg are common elements of the lithosphere (Ramadan et al. 2000), and odd-carbon numbered alkanes are characteristic of vegetative detritus (Rogge et al. 1993). Therefore, this factor is likely contributed from weathered crustal materials and incorporated plant abrasion particles. 146 5.3.5.2 Relative source contribution to ambient PM2.5 and PM10 and comparison of saccharide rich factors with source profiles of soil samples The relative source contributions to PM2.5 and PM10 were calculated as described in earlier chapters. The apportionment of the biogenic contribution of the wax n-alkanes (37.9% for PM2.5 and 31.1% for PM10) was based on the calculation presented in Table 5.2. Using the estimated biogenic and petroleum fraction based on n-alkane data, the petroleum fraction of the wax factor was lumped with the petroleum-powered engine into a combined mobile source factor. The sucrose and trehalose rich bio-aerosol and the glucose and sugar polyol rich bio-aerosol factors were also combined to characterize the contribution from primary biologically derived carbon sources. The apportionment to aerosol sources for PM2.5 shows six major source categories that dominate PM2.5 mass (Figure 5.8). These dominant sources include wood smoke (19%), mobile sources (18%), secondary sulfate and nitrate (18%), primary biologically derived carbon sources (16%), .salt particles (15%), and cooking related sources (13%). In a comparison, these six source categories contributed to PM10 at different magnitudes, with secondary sulfate and nitrate (27%) and primary biologically derived carbon sources (22%) being the largest two contributors, followed by mobile sources (16%), the cooking related factor (11%), salt particles (11%), and wood smoke (9%). Contributions of biogenic wax and crustal material were minor compared with the six major source categories. 147 These results can be compared with earlier work done by Ramadan et al. (2000) which identified seven sources of ambient aerosols in Phoenix through the application of a PMF model based solely on OC, EC, and elements. Ramadan et al. (2000) estimated mobile sources, wood smoke, and emissions from coal-fired power plants (similar to secondary sulfate and nitrate) to be the major sources of fine PM with minor contributions from sea salt, soil, and industrial smelter emissions, while soil and construction dusts were the two predominant factors responsible for over 90% of the coarse PM mass. The major difference between the results from this current study and the work by Ramadan et al. (2000) is based on the estimation of contribution from soil and construction. To determine if there were any major unidentified sources of aerosol mass in the current study, the mass apportioned by PMF was plotted against the measured mass value (Figure 5.9). The result suggests that most of the sources having significant contributions to PM2.5 mass have been isolated by the model and that any unidentified sources would be minor contributors. The discrepancy between the current work (where soil is not resolved as a PM10 source) and prior work identifying soil as a major component of PM10, likely arises from the selection of marker compounds used to track sources. The current focus on organic markers likely results in the incorporation of soil resuspension into the factors dominated by saccharides and labeled as primary biologically derived carbon sources. Comparison shows moderate consistency between the source profiles of these saccharide rich source factors and the saccharide composition of the Higley size-segregated soil samples (Figure 5.10). However, a major discrepancy was 148 found in the relative abundance of sucrose and glucose to a lesser extent. This implies soil was not solely responsible for these two PMF source factors resolved for Higley aerosol samples. In a summary, the major factors contributing to the Higley aerosol levels based on the PMF modeling included mobile sources, wood smoke, secondary sulfate and nitrate, primary biologically derived carbon sources, cooking operations, and sea salt particles. Comparison of the modeled source profiles with the local size-segregated soil samples based on saccharides showed moderate consistency with a major discrepancy between the relative abundance of sucrose and glucose. This is consistent with the earlier conclusion that soil may not be the only significant PM source at Higley introducing saccharides to ambient aerosols, as suggested by other studies, and there may be other major sources such as PBAPs that are injected directly into the atmosphere instead of residing in the surface soil and being re-entrained through soil resuspension. However, caution should be taken to extrapolate this conclusion to other locations, as local PM and soil samples will need to be studied to decide whether soil is or is not the only important source of aerosol saccharides at a specific location. 149 Figure 5.7 PMF modeled factor profiles for Higley PM2.5 and PM10 samples (Jan - Apr 2008). (Unit on Y-axis: % of species apportioned to factor.) (OC: organic carbon; EC: elemental carbon; Na: sodium; Mg: magnesium; C25-C32: alkane series; HOP1: 17a(H)-21p(H)-hopane; HOP2: 17 1)PM2.5 Petroleum-powered engine ,-B-, , i lljl n-Alkane wax iiiiiiii.U,.,.,,, ,TI Cooking related factor Salt particle i i ^ ^ ^ Secondary sulfate and nitrate liaiil -I ii,. a. . I Wood smoke Glucose and sugar polyol rich bio-aerosol ,*l«i,i,i. ouuouooo 111« I«*j i * 2)PM10 Petroleum-powered engine n-Alkane wax •r.Illii.1. . x . , ^ r « 120 Cooking related factor 100 i.l. ,i, i Salt particle Sucrose and trehalose rich bio-aerosol 140 Glucose and sugar polyol rich bio-aerosol 120 100 JUL Crustal materials ii.ii.i. §s ng s § OS § S3 II g S i fl I 151 Figure 5.8 Relative source contributions to (a) PM2.5 and (b) PM10 at Higley. These source contributions are determined by a multiple linear regression of the modeled source factor scores to the measured PM mass. (a) Biogenic n-alkane Cooking related Secondary sulfate factor and nitrate 13% 18% Primary Wood smoke biologically 19% derived carbon source 16% Salt particle Mobile source 15% 18% (b) Biogenic n-alkane Cooking related Secondary sulfate factor and nitrate 11% 27% Primary biologically derived carbon Crustal material source 1% 22% Salt particle Mobile source 11% 16% Wood smoke 9% 152 Figure 5.9 Apportioned PM mass concentration plotted against the measured values for Higley PM2.5 and PM10 samples 0 2 4 6 8 10 12 14 16 Measured PM2.5 Mass (ug/m3) 0 10 20 30 40 50 60 Measured PM10 Mass (ug/m3) 153 Figure 5.10 Comparison of the two PMF source profiles (sucrose and trehalose rich bio-aerosol and glucose and sugar polyol rich bio-aerosol) with the saccharide composition in (a) PM2.5 and (b) PM10 of the three soil types at Higley (a) 600 D Sucrose and trehalose rich bio-aerosol (PM2.5) • Glucose and sugar polyol rich bio-aerosol (PM2.5) • Agricultural soil (PM2.5) 500 • Native soil (PM2.5) m Road dust (PM2.5) 400 E |> 300 200 100 0 Glucose Sucrose Trehalose Glycerol Arabitol Mannitol (b) • Sucrose and trehalose rich bio-aerosol (PM10) • Glucose and sugar polyol rich bio-aerosol (PM10) H Agricultural soil (PM10) • Native soil (PM10) • Road dust (PM10) ! JB_ Glucose Sucrose Trehalose Glycerol Arabitol Mannitol 154 5.4 Acknowledgement This project was funded by a grant from the EPA STAR program. I would like to thank the Maricopa County Air Quality Department for access and power at the field site, my colleague Andrea Clements who collaborated with me closely on collecting the ambient and soil samples and analyzed the elemental carbon, organic carbon, and ions used in analysis in this chapter, and Dr. Tony Hartshorn at the School of Earth and Space Exploration at Arizona State University who helped me with soil taxonomy. My gratitude also extends to Dr. Pierre Herckes and his research group at the Arizona State University for their support of me conducting all the chemical analysis work in their Environmental Chemistry Lab. 155 Chapter 6 Source Study of Primary Biological Aerosol Particles 6.1 Introduction In Chapter 5, soil samples collected in the vicinity of the aerosol sampling site in Higley, AZ were compared to ambient PM samples of equivalent size. The total nonlevoglucosan-saccharide content relative to PM mass was shown to be much higher in ambient aerosol samples when compared to equivalent soil samples (Table 5.10). PMF modeling of ambient aerosol data resolved two factors representing sources enriched in nonlevoglucosan saccharides. The source profiles of these resolved factors, however, were not consistent with the saccharide composition of the local soils analyzed (Section 5.3.5.2). This suggests that the atmospheric entrainment of soil and associated biota was not the dominant contributor to the local ambient PM saccharides, and instead, that there may be additional major sources of ambient PM that were not considered at the Higley sampling site. The most likely candidate source is PBAPs, which have recently been shown to be responsible for a significant fraction of ambient PM saccharides in both urban and rural atmospheres (Pashynska et al. 2002; Carvalho et al. 2003; Graham et al. 2003; Elbert et al. 2007; Yttri et al. 2007; Bauer et al. 2008). To characterize this potential source category, several local plant and spore samples representative of the Arizona desert environment were collected 156 and analyzed for their saccharide composition. Size equivalent samples were collected to identify the chemical composition of PBAPs to determine the importance of this source to ambient PM2.5 and PM10 levels. 6.2 Sample collection A total of 9 plant and spore samples were obtained representing vegetation and fungi common to the southwest US desert (Table 6.1). The plant samples were collected with the assistance of Dr. Leslie Landrum of the Herbarium at Arizona State University and two mushroom spore samples were obtained from the Arizona Mycota Project with the help of Dr. Scott Bates at Arizona State University. The mold spore sample was collected from the open atmosphere in Phoenix, separated and cultured for two weeks by Dr. Scott Bates of the Arizona State University, School of Life Sciences. 157 Table 6.1 Plant and spore samples collected in the Phoenix desert area for PBAP source studies. Species Division or Phylum* Family Description Abbreviation Sample Note Prosopis juliflora Magnoliophyta Fabaceae a shrub or small tree native to Mexico, South America PJ green leaves with presence of a small amount of and the Caribbean flowers Prosopis juliflora Magnoliophyta Fabaceae a shrub or small tree native to Mexico, South America PJ_dead dry and dead branches without leaves and the Caribbean Simmondsia chinensis Magnoliophyta Simmondsiaceae a shrub native to the Sonoran and Mojave deserts of SC green leaves (Jojoba) Arizona, California, and Mexico Atriplex lentiformis Magnoliophyta Amaranthaceae a saltbush native to the southwest United States and AL green leaves northern Mexico Baccharis sarothroides Angiosperms Asteraceae a flowering shrub native to northern Mexico and the BS branches without leaves (desert broom) southwestern United States Justicia califomica Angiosperms Acanthaceae a floweringshru b native to the deserts of southern JC green leaves (hummingbird bush) California, Arizona, and northern Mexico Podaxis pistillaris Basidiomycota Agaricaceae a mushroom with a wood stem (~15cm) and a large PP spores cap which houses a large number of spores and thrives in deserts and semi-desert areas Battarrea phalloides Basidiomycota Tulostomataceae a mushroom with a woody stem (~20cm) and a spore BP spores sac atop, common in dry, sandy locations throughout North America, primarily in western regions. Aspergillus niger Ascomycota Trichocomaceae a fungus that is ubiquitous in soil and common in the AN spores air, growing as molds on the surface of a substrate ' Division is used in botany and phylum in zoology 158 6.3 Analytical method After collection, plant samples were air-dried for a week and oven-dried at 110°C for 24hrs prior to grinding into powders to represent the grinding and milling process of plant vegetation that occurs as material is deposited on roads and other surfaces and resuspended as dusts. The spore samples were gently shaken upside down for the spore particles (2-10 microns in diameter) to be released and collected onto pre-cleaned aluminum foil. Particles from these pretreated samples were then placed in a 1L heavy wall filtering flask for resuspension as described in Section 5.2.2. Collected filter samples were analyzed for organic and elemental carbon, total mass, and saccharide composition following the procedures described in Section 5.2.3, 5.2.4, and 5.2.5, respectively. 6.4 Results and discussion 6.4.1 Saccharide composition The saccharide contents of the locally representative size-selected PBAP samples are reported in Table 6.2. In general, there was no statistically significant difference in saccharide profiles between the PM2.5 and PM10 fraction of the collected samples (P-values ranged from 0.7 to 0.9, Welch two sample t-test, two-tailed, 95% confidence interval). However, the plant and spore samples displayed different patterns in total measured saccharide concentration and relative abundance of saccharides between the two sample categories. 159 The six plant samples contained a total measured saccharide concentration ranging from 6 to 71 ug mg", comprising from 1% to 7% of the measured sample mass, and the carbon content of measured saccharides also comprising from 1% to 7% of the measured OC. The total measured saccharide contents (% wet mass) are within the ranges reported in woody plant samples (Chow and Landhausser 2004). The most abundant saccharide compounds in the collected plant samples were sucrose and glucose, with one plant species (Prosopis juliflora) particularly enriched in sucrose and another (Atriplex lentiformis) in glucose. This is consistent with prior knowledge because sucrose and glucose are the most abundant storage and photosynthetic carbohydrates in plants (Cowie and Hedges 1984). The elevated level of sucrose measured in Prosopis juliflora was likely due to the presence of flowers in the sample, as sucrose is an important sugar in developing flower buds (Bieleski 1995) and a major component in pollen grains (Pacini et al. 2006). The one exception to the enrichment of sucrose and glucose in plant samples was the dead branches of Prosopis juliflora, which contained a slightly higher concentration of trehalose than sucrose and glucose. This particular sample also contained approximately an order of magnitude greater concentration of mannitol compared to the other plant samples. Trehalose and mannitol have been labeled as "fungus-specific" saccharides as they are rarely found in plants (Lewis and Smith 1967; Almeida et al. 2007). The presence of trehalose and mannitol in plant tissues is often associated with desiccation conditions, under which they are produced as protectants against osmotic stress and to stabilize dry membranes (Schobert 1977; Crowe et al. 1984; Stoop et al. 1996). The 160 detection of both mannitol and trehalose in plants has also been attributed to the presence of fungi (Mellor 1992; Murelli et al. 1996). Apart from mannitol, other saccharide polyols were also detected, including glycerol, arabitol, and sorbitol. However, the concentrations of these species were small compared to the four major saccharide species discussed above. The size-separated spore samples contained higher saccharide contents than the size-separated plant samples, with total measured saccharide concentration ranging from 93 to 268 ug mg'1. This measured saccharide content represents between 9% and 27% of collected mass. In the spore samples, trehalose and mannitol were the most dominant saccharide compounds, comprising more than 97% of the total measured saccharides. Trehalose and mannitol are well-known fungal metabolites (Martin et al. 1988; Sillje et al. 1999), functioning mainly as storage carbohydrates (Blumenthal 1976) and as membrane stabilizers (Lewis and Smith 1967; Chandrasekhar and Gaber 1988). Mandels et al. (1965) reported that trehalose and mannitol account for 19% and 2% of the dry weight in spores of the fungus Myrothecium verrucaria, respectively. Other literature reports suggest mannitol represents between 10% and 15% of the dry weight in spores of the filamentous fungus {Aspergillus niger) (Ruijter et al. 2003). If the samples analyzed in the present work are assumed to contain -50% water by weight (Bidochka et al. 1990; Dijksterhuis and Samson 2007), then the present analysis is consistent with these earlier literature reports. Other saccharide compounds (such as sucrose and glucose) were also measured in the spore samples, although at much lower concentrations, a 161 characteristic consistent with higher fungi (Ascomycota and Basidiomycota) (Pfyffer and Rast 1980). The composition profiles of major saccharide compounds in the PBAP samples are shown in Figure 6.1. 162 Table 6.2 Saccharide composition in PBAP samples selected in current study (ng nig"1). Abbreviations are given in Table 6.1. Plant Spore PJ PJ dead SC AL BS JC PP BP AN PM10 PM2.5 PM10 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM10 PM2.5 PM10 PM2.5 PM10 Glucose 1706 1795 1910 6738 12635 20262 24464 868 834 699 640 888 324 274 32 Sucrose 63155 69089 2337 5029 5614 5990 6609 3901 5170 3728 129 88 55 5 10 Trehalose 285 238 2908 897 217 962 324 965 312 637 87107 86715 126753 123878 183316 Glycerol 193 174 349 305 191 166 269 432 237 338 112 114 126 62 202 Iso-erythritol 24 30 15 18 8 101 129 44 52 10 2 3 15 13 4224 Xylitol BDL BDL BDL 31 29 14 9 BDL BDL BDL 2 3 15 10 0 Arabitol 13 18 169 17 22 71 93 BDL 24 23 16 27 15 15 1836 Ribitol 12 9 BDL 33 20 23 25 BDL 16 8 2 2 18 19 BDL Mannitol 81 79 1173 224 128 146 94 198 137 77 5658 9069 11826 16999 77799 Sorbitol 18 21 122 33 28 57 76 96 48 30 11 17 123 181 75 Galactitol 14 12 26 22 14 28 22 69 44 25 172 102 73 52 771 Total Measured Saccharides 65501 71466 9011 13348 18907 27819 32113 6573 6874 5574 93851 97029 139344 141507 268265 % of total measured saccharide to mass 7% 7% 1% 1% 2% 3% 3% 1% 1% 1% 9% 10% 14% 14% 27% OC/Mass 45% 42% 31% 30% 25% 28% 34% 37% 28% 38% 23% 25% 30% 28% 31% % of carbon content of total measured saccharide to OC 6% 7% 1% 2% 3% 4% 4% 1% 1% 1% 17% 16% 19% 21% 36% 163 Figure 6.1 Composition profiles of major saccharide compounds in PBAP samples (values were averaged from concentrations in particles equivalent to PM2.5 and PM10). Abbreviations are given in Table 6.1. 7.E+04 £..^^V\J 6.E+04 ^mZ* v> .. o U D.-> 5.E+04 Q. E to n m to ** •(P0 "-" 1 4.E+04 £ CL o CQ DL o" 1.E+05 tfl 0. — Q. 3.E+04 o> "E5 Qr>. E D) "& C 2.E+04 c 1.E+04 O.E+00 n cj.nn e>° c? # <0* 6.4.2 Comparison with Ambient PM Samples and Soil Samples In Chapter 5, the total measured saccharide content (excluding levoglucosan) relative to PM mass (and PM OC based on carbon content) in Higley ambient aerosol samples was 0.20% (0.26%) on average in PM2.5 and 0.11% (0.37%) in PM10 (Table 5.10). The equivalent comparison of saccharide content in size-segregated samples of local soils showed lower total saccharide levels relative to sample mass and OC excluding soil as the dominant source of measured ambient aerosol saccharides at Higely (Section 5.3.4). Separately, the PBAP size-selected samples analyzed in the current study had total measured non-levoglucosan saccharide content ranging from 1% to 27% relative to PM mass and from 1% to 36% relative to measured OC (based 164 on carbon content) in the PM2.5 and PM10 fraction of plant and spore samples. These values are one to two orders of magnitude greater than equivalent ratios measured in ambient PM samples, which supports the likelihood of PBAPs as an important source of ambient aerosol saccharides in the Phoenix desert area. The major non-levoglucosan saccharide species measured in aerosols, namely sucrose, glucose, and trehalose, were also found in significant amount in PBAP samples, with glucose and sucrose enriched in plant samples and trehalose enriched in fungal samples. 6.4.3 Comparison with PMF Results In Chapter 5, PMF modeling of ambient aerosol data from Higley, AZ resolved two factors with enrichment of non-levoglucosan saccharides (Section 5.3.5.1). The saccharide profiles of these two factors were initially compared with the measured saccharide composition of the local soil samples. This comparison presented a major discrepancy in the relative abundance of sucrose and glucose between soil and ambient data (Section 5.3.5.2). An equivalent comparison is made here between the PBAP saccharide profiles and PMF factors resolved from ambient data (Figure 6.2). Although the measured saccharide composition of PBAPs is several orders of magnitude higher when compared to the modeled source factors, the relative abundance of each saccharide species in the plant samples is fairly consistent with the resolved "sucrose and trehalose rich bio-aerosol" source factor. In particular, Prosopis juliflora has a similar distribution of non-levoglucosan saccharides as the 165 resolved source factor. The saccharide profiles of the collected spore samples are not consistent with the resolved "glucose and sugar polyol rich bio-aerosol" factor when comparing the relative abundance of major saccharide compounds. However, this may be due to the small number of spore samples analyzed, and mature spores of other common fungal species may significantly differ from those analyzed in terms of their saccharide composition. For example, S. rugosaannulata and L. edodes are enriched in arabitol (23% and 28% of total saccharides) and P. ostreatus spores are rich in sucrose (10% of total saccharides) than other fungal spore species (Tereshina et al. 2000). Nevertheless, all saccharide species in the modeled "glucose and sugar polyol rich bio-aerosol" factor are present at significantly higher concentrations in the measured spore samples, supporting the likelihood that spores are one potential source influencing this factor. As the saccharide levels in all the measured PBAP samples are significantly higher than the modeled results, PBAPs may have undergone some physical transformation (i.e. resuspension, wetting, drying) that incorporate other material into the resuspension process that dilutes the saccharide concentration prior to release to the atmosphere. Alternatively, the modeled factor could represent a combination of both PBAPs and the less saccharide-abundant sources of soil and associated biota. Nevertheless, the measured PBAP samples contained saccharides at a high enough level to match the modeled source profiles while the measured samples of soil did not. This confirms the importance of PBAPs as a source of PM source in additional to soil and associate biota at the Higley sampling site. 166 6.5 Acknowledgement The assistance of Dr. Leslie Landrum and Dr. Scott Bates at the School of Life Science at Arizona State University in identification and collection of the plant and spore samples is gratefully acknowledged. The work presented in this chapter wouldn't be possible without their help. I would also like to thank my colleague Andrea Clements for her assistance in analyzing the elemental and organic carbon content of the PBAP samples, and for Dr. Pierre Herckes and his research group at Arizona State University for their support. 167 Figure 6.2 Comparison of the two PMF source profiles at Higley (sucrose and trehalose rich bio-aerosol and glucose and sugar polyol rich bio-aerosol) with the local PBAP sample source profiles (values were averaged from concentrations in particles equivalent to PM2.5 and PM10): (a) Comparison with the sucrose and trehalose rich bio-aerosol factor (PM2.5 and PM10); (b) Comparison with the glucose and sugar polyol rich bio-aerosol factor (PM2.5 and PM10). Abbreviations of the PBAP samples are given in Table 6.1. (a) 700 70000 IPJ-plant ISC- plant SAL- plant 60000 H BS - plant • JC-plant 50000 X Sucrose and trehalose rich bio-aerosol (PM2.5) u 40000 3 O in a. 30000 < CO 0. 20000 E c 10000 X -x- -X- -X- Glucose Sucrose Trehalose Glycerol Arabitol Mannitol 700 70000 0 PJ - plant ESC-plant -. 600 SAL - plant 60000 • BS-plant DJC - plant X Sucrose and trehalose rich bio-aerosol (PM10) 50000 0) u 40000 3 O (A a. 30000 < CD a.o> 20000 E o> c 10000 X JEfc -x- -X- -X- Glucose Sucrose Trehalose Glycerol Arabitol Mannitol 150 200000 X S PP - spore O BP - spore 150000 g o Q. O 100 BAN -spore ^O 3 O V) X Glucose and sugar polyol •o 100000 _o> Qi. rich bio-aerosol (PM2.5) a> < •o m £ 50 •) E £ 50000 "3> c c Glucose Sucrose Trehalose Glycerol Arabitol Mannitol 50 200000 H PP - spore X £ 40 B BP - spore 150000 £ o BAN -spore 3 30 o (0 X X Glucose and sugar polyol •a .2 rich bio-aerosol (PM10) o> ? 20 E. E o) 10 X c Glucose Sucrose Trehalose Glycerol Arabitol Mannitol 169 Chapter 7 Characterization of Saccharides in PM10 and Soil Samples from Austria 7.1 Introduction In cooperation with the Institute of Chemical Technologies and Analytics at Vienna University of Technology, Austria, 8 ambient PM10 samples and 7 bare soil samples collected in 2005 at different locations in Austria were analyzed for saccharide composition. The goal of this collaboration was to compare the saccharide composition between soil samples collected in Austria and Arizona, and to compare the measurement of aerosol saccharides by two different analytical methods (HPLC-based and GCMS-based). Table 7.1 and Figure 7.1 summarize the soil sampling dates, sampling locations, and sample description. The 8 PM10 samples were taken at Voelkermarkterstrasse, Koschatstrasse, Gurtschitschach, and Lavamuend, representing a subset of the soil sampling sites, between May and August 2005. 7.2 Analytical method Bare soil samples were resuspended and a size-selected sample of particles with atmospheric diameter equivalent to PM2.5 and PM10 were collected following 170 the same procedure as described in Section 5.2.2. Collected soil samples and ambient PM10 samples were analyzed for EC and OC, total mass (for soil filter samples only), and saccharide composition as described in Section 5.2.3, 5.2.4, and 5.2.5, respectively. For comparison of analytical methods, an equivalent PM10 sample was analyzed at the Institute of Chemical Technologies and Analytics at Vienna University of Technology using an anion-exchange HPLC and pulsed-amperometric detection based method. Details of this method have been presented in detail elsewhere (Caseiro et al. 2007). Table 7.1 Austria bare soil sample information Sample Sampling Site Sampling Date Material Description of Sampling site ID P79 Stixneusiedl 2005-04-25 Currently Knoll in lowland area; meadow, forest, agricultural uncultivated/Agric activities ultural Soil P80 Schwechat 2005-04-25 Street Dust Lowland; agricultural activities, airport, regional road P81 St. Poelten 2005-04-25 Street Dust Expanded valley in downs; major road, urban residential area, light industrial area P83 Lavamuend 2005-03-27 Agricultural Soil Hills; agricultural activities P84 Gurtschitschach 2005-03-27 Agricultural Soil Hills; agricultural activities P86 Koschatstrasse 2005-03-26 Street Dust Lowland; basin, surrounded by mountains, moderately frequented road, urban residential area P87 Voelkermarkterstra 2005-03-26 Street Dust Lowland; basin, surrounded by mountains, major sse road, urban residential area 171 Figure 7.1 Map of Austria with the soil sampling locations circled in red Note: the sampling sites "Koschatstrasse" and "Voelkermarkterstrasse" refer to specific sampling sites near major roadways in the city of Klagenfurt. 7.3 Results and discussion 7.3.1 Soil saccharides and comparison with Higley soil samples The total measured saccharide (TMS) concentrations varied between the two types of soil samples collected in Austria as shown in Figure 7.2 (a). The TMS concentration in Austria agricultural soil samples ranged from 76 to 217 ng mg"1 of PM, and this comprised between 0.2% and 0.43% of the measured OC. In street dust samples, the TMS value ranged from 598 to 3134 ng m"1, representing 0.47% to 1.86% of the measured OC. This difference may originate from the carbon consumption by crops and vegetation in agricultural soils. Although higher saccharide content was found in the street dust compared to the agricultural soil, the relative species composition as a percentage of the TMS was consistent between both types of 172 soils as shown in Figure 7.2 (b). Trehalose was the dominant compound, contributing between 40% and 85% of the TMS. Mannitol was the second most significant saccharide, with greater prevalence in street dust samples where it contributed between 7% and 37% of the TMS concentrations. Glucose and sucrose were present in all samples, although at lower concentrations. The most striking difference in the relative saccharide composition between the two types of soil samples is found in the saccharide polyols that were elevated in street dusts relative to levels found in agricultural soils. This may have resulted from a preferential consumption of saccharide polyols (reduced sugars) by plant growth in agricultural fields, or from the contribution of different fungal sources which are typically enriched in polyols (Pfyffer and Rast 1980). No consistent statistical difference was observed between the chemical composition of the PM2.5 and PM10 fraction of the Austria soil samples (P-values ranged from 0.5 to 0.9, Welch two sample t-test, two-tailed, 95% confidence interval). Comparisons between the Austria soil samples (collected in March and April) and Higley soil samples (collected in April) show comparable saccharide contents in agricultural soils, but elevated saccharide content from the street dust collected in Austria as shown in Figure 7.3 (a). The comparison based on the relative compound abundance to TMS (Figure 7.3 (b)), shows both agricultural soils displayed similar relative saccharide abundances. Differences are observed between the road dust samples, with the Higley road dust enriched in glucose and sucrose while the Austria street dust samples are enriched in mannitol, arabitol, and other sugar polyols. This 173 may indicate plants and pollens were more important as the original source of material deposited to roads in Higley (Pacini et al. 2006) while fungal spores were more important in deposition to Austria street dusts (Bauer et al. 2008). While both sets of samples from Austria and Higley were collected at the same time of the year, the climatic and ecological conditions at the two sites are very different. Therefore, no definite conclusion could be reached from this saccharide composition difference. However, pollen calendars for Arizona (O'Rourke 1990) and Austria (HON Foundation 2004) indicate that local pollen levels peak in March and April in Arizona and in April and May in Austria. On the other hand, the emission of fungal spores measured in three Austrian cities (Vienna, Graz, and Salzburg) increased starting in April and lasts through November (Bauer 2008). This indicates the saccharide composition difference in road dust samples between Arizona and Austria was likely a reflection of the different contribution strengths from two different PBAP sources (pollen vs. spores) under two dramatically different climatic conditions. •n 'oi Saccharides in Soil (rq T^—l ' (% of Total Measured Saccharides) Saccharides in Soil (ng/mg PM) c •1 m -kNio^ *•••'•• *•*••*' {•* M% •»..«f..JM!.«i..^. M^V*. »r^,. «-— .„ ^^-.«.w;^ J ^ o o" aft - OJ o F5^^ %&«*«*•»«#«.*#« M i ma IHI :XM>*# o o os e ito l ito l - o 4 o Arythrito l rT 5 D 0) alos e It- lito l ctito l os e os e 3- IT ito l B) ito l o o Arythrito l o ^ o_ CD o_ o_ o_ 3 I 175 Figure 7.3 Comparison of saccharide composition between Austria and Arizona April soil samples: (a) mass concentrations (b) percentages of total measured saccharides (Abbreviation note: H - Higley, AZ, A - Austria, Ag - Agricultural soil, Native - Currently uncultivated soil, Road - road dust) (a) 1600 [DGalactitol 1400 - • Sorbitol 1 1 BMannitol I 1200 B Q Ribitol 1000 • Arabitol f"3 (A BXylitol O 800 - TO Q. E3lso-erythritol E o - s o 600 • Glycerol • Trehalose •a 400 " • 1 * • 0 Sucrose 200 - :v •11 u H ED Glucose TO i flrH „„ ! B I V) 0 B M B < a , H , B £ 5! ^ •^ /> r / S//A o*> /// *w^ **• v* > (b) • Galactitol • Sorbitol BMannitol • Ribitol • Arabitol BXylitol • Iso-erythritol • Glycerol El Trehalose E3 Sucrose B Glucose N N ^ .# «* & «* /// r JCV" r f jr s //s **v ^ *dr f/^^/ / 176 7.3.2 PM10 saccharides, GM/AX ratios, and method comparison PM10 samples were collected in May and again in August 2005 at four sites in Austria. While the measured PM10 mass did not show much variation between May (23.2 ± 4.9 ug m"3) and August (23.5 ± 5.9 ug m"3), the average of total measured saccharides increased from 58 ng m" to 122 ng m" during this period (Figure 7.4). Most of the increase was due to an increase in the concentration of mannitol, arabitol, and trehalose, consistent with elevated fungal spore emissions in Austria during this period (Bauer 2008). The GM/AX ratio (defined as the ratio between the sum of galactose and mannose and the sum of arabinose and xylose (Oades 1984)) can quantify the relative importance of fungal spore derived aerosol sources versus plant derived sources. This ratio has been used in soil science to differentiate plant derived saccharides and microbial saccharides (Kiem and Kogel-Knabner 2003; Spielvogel et al. 2008). Plant material with significant levels of arabinose and xylose contrasts microorganisms that mainly synthesize galactose, glucose, and mannose, but little arabinose and xylose (Cheshire 1979). Therefore plant material typically has GM/AX ratios less than 0.5 while microbial material has a GM/AX ratio greater than 2 (Oades 1984). Figure 7.5 shows the calculated GM/AX ratio at the four sampling sites in Austria. Slightly higher GM/AX ratios were measured PM10 samples in August compared to May, and one site (Koschatstrasse) had higher GM/AX ratios compared to the other sites. The fungal spore tracer arabitol and mannitol is correlated with the GM/AX ratio in August PM10 samples as shown in Figure 7.5. However, this 177 correlation was not observed in May PM10 samples, possibly due to contributions from other microbial sources. In general, the GM/AX ratio was mostly greater than 1 in Austria PM10 samples, indicating a strong influence from microbially derived sources relative to plant sources. Finally, equivalent measurements conducted separately using two different analytical methods were compared, as described earlier in Section 7.2. As shown in Figure 7.6, slightly higher concentrations of saccharides were measured by the HPLC-based method relative to the GCMS-based method. Overall, a good agreement was observed within error ranges between the two data sets, providing independent verification of the analytical methods used in this thesis. 7.4 Acknowledgement This project was initiated and coordinated by Prof. Hans Puxbaum at the Institute of Chemical Technologies and Analytics at Vienna University of Technology. The sample collection, shipping, and analysis were performed by researchers including Nicole Jankowski, Magdalena Rzaca, and Heidi Bauer at Vienna University of Technology. The organic speciation analysis at Arizona State University was conducted in the Environmental Chemistry Lab with the assistance of Dr. Pierre Herckes and his research group. The EC and OC analysis at ASU was conducted by my colleague Andrea Clements. My appreciation goes to all of them. 178 Figure 7.4 Saccharide compositions of Austria PM10 samples 160 140 8/2/2005 • Sorbitol 2 — • Mannitol 120 g| • Arabitol 100 5/30/200 oI Io Slso-erythritol 80 T3 S • Glycerol "E 0. 60 0 Trehalose 5.E u 40 13 Sucrose o S Glucose wra 20 • Levoglucosan // / / //// /// " ^ A° Figure 7.5 GM/AX ratios and correlation with arabitol and mannitol in Austria PM10 samples 35 2.0 May August 30 1.8 I 1.6 25 1.4 ra 1.2 nc 20 X 1.0 < 15 s •ac 0.8 o 10 Arabitol 0.6 Mannitol 0.4 .Q 5 TO GM/AX Ratio 0.2 0 0.0 St> c? «? 8> & '// «s? y //// s // ov • Sample Site 179 Figure 7.6 Comparison of saccharide concentrations in Austria PM10 samples based on two analytical methods (HPLC vs. GCMS) CO E "3> c c a> E a n •a a> TO .a 0. GCMS-based Measurements (ng/m3) 180 Chapter 8 Conclusions 8.1 Summary The ultimate goal of this dissertation was to characterize ambient PM samples and their local sources through quantification of simple saccharide compounds. From 2005 to 2009, series of ambient aerosol and representative source samples were collected at various sites in Texas, Arizona, and Austria, with sampling locations ranging from arid desert to temperate and humid forests, from rural to suburban and urban areas. Carbonaceous compounds were analyzed to detail the chemical composition of different PM samples, and molecular markers were used in a source apportionment model (PMF) to estimate the relative contributions of various local PM sources. In particular, this thesis advanced the use of simple saccharide compounds as indicators of biologically important organic carbon in nature, which reflect the impact from the biologically derived sources of organic matter in different ecosystems. For example, the total measured saccharide concentrations in PM2.5 samples collected at the Arizona site (24 ng m"3 on average for Higley) was much lower compared to the three Texas sites (52 ng m" on average for Dallas, 70 ng m" on average for San Augustine, and 128 ng m"3 on average for Clarksville). This is partly due to lower PM2.5 mass concentrations as well as a lower fraction of PM2.5 that was organic carbon (Table 8.1), and is possibly a result of the arid desert location. 181 However, the average fraction of PM2.5 mass and OC that was quantified as saccharides at Higley (0.5% and 0.7%) was comparable to that measured at Dallas (0.6% and 0.8%) (Table 8.1). Both sites are located in large urban areas but have different climatological and ecological conditions. Among the three sampling sites in Texas, total measured saccharides were higher at the two rural sites (70 ng m"3 on average for San Augustine and 128 ng m~3 on average for Clarksville) than the urban site (52 ng m"3 on average for Dallas), which is consistent with lower expected contributions from biogenic sources at the urban location. This corroborates the suggested role of saccharides as tracers for biologically derived aerosol sources. On the other hand, soil samples collected in Arizona and in Austria at the same time of the year showed different saccharide compositions, with Arizona soil enriched in pollen-related saccharides and glucose and Austrian soil enriched in fungal spore derived saccharides (Section 7.3.1). Taking into account the annual emission cycle of pollen and spores in Arizona and Austria, this could indicate different contribution strengths from two different PBAP sources under two different climatic patterns. Despite the varying sampling locations, aerosol saccharide concentrations displayed consistent seasonal variations and correlations between species that may indicate their utility as markers. The elevation of aerosol levoglucosan was observed almost exclusively in winter and early spring samples, reflecting local and regional biomass burning. Sucrose was found to be significant only during spring sampling, paralleling the flowering period and consistent with sucrose as an important component in pollen grains. Glucose was ubiquitous as it is the primary simple sugar 182 in photosynthetic cycles. Trehalose, mannitol, arabitol and other minor saccharide polyols were more indicative of microbially derived carbon sources. The concentrations of polyols were enhanced when environmental stressors such as heat and drought were present. Through this study, the ambient concentration of saccharide species with common sources were well correlated (r=0.4~0.9), and the seasonal variations of aerosol saccharides were compared with other studies (Section 5.3.2.1). Often, data from this work showed good agreement with literature reports, and the few exceptions could be interpreted as indicators of ecosystem function, climate, and local sources. Again, this bolsters the argument that saccharides in aerosols can be used as tracers for the seasonal change of carbohydrate production and utilization in a local ecosystem or may represent regional source trends. Molecular markers can be used in source apportionment techniques to estimate the contributions to ambient PM from primary sources. While much work has focused on markers for anthropogenic emissions, the markers for tracking the primary biologically derived carbon sources have not been as well developed. One hypothesis tested in this thesis is whether specific saccharide compounds can be used as markers in receptor modeling to characterize the contribution from the broad source category of PBAPs to ambient particulate matter. The saccharide markers selected for this application were levoglucosan (an established marker for biomass combustion) and novel markers including glucose, sucrose, trehalose, mannitol, and arabitol. By using these novel markers in conjunction with established organic and inorganic tracer compounds, matrix separation techniques have been successfully 183 applied to the ambient PM datasets collected at three Texas sites and one Arizona site and resolved relevant source factors with meaningful interpretation. Among these factors, two were dominated by the proposed novel saccharide markers, and these factors may represent either PBAPs or other biologically derived carbohydrate sources such as soil and associated biota. By combining these two factors, the contribution from primary biologically derived carbohydrate sources to ambient PM was estimated (Table 8.2). The sources for which saccharides are useful as tracers, including both biomass burning and primary biologically derived carbohydrate sources, contributed to over one third of the measured PM mass at both a rural site (San Augustine, TX) and a suburban site (Higley, AZ), implying the benefit of quantifying saccharides in PM. The comparison between different sites also demonstrated that saccharide-rich sources contributed a higher percentage to the ambient PM2.5 for a rural site (36% for San Augustine) when compared to an urban site (21% for Dallas) or a remote site (16% for BBNP) in Texas, that biomass burning contributed a greater fraction to PM2.5 (20%) compared to PM10 (13%), and that the primary biologically derived carbohydrate sources contributed more to PM10 (21%) than PM2.5 (16%) at Higley. The greater contribution of saccharide-rich sources to PM2.5 at the suburban site Higley (36%) in Arizona compared to the urban site Dallas (21%) in Texas may indicate the influence from diverse flora and fungal species present in the Sonoran desert. In addition to the saccharide-related sources, other major sources were also identified in source attribution modeling, including motor vehicles, secondary aerosol formation, cooking operations, biogenic wax, sea salt, 184 crustal material, and road dust, each with varying contributions to aerosol levels at different sites. The modeled source profiles and relative PM source contributions of these major source factors can be useful in improving local emission inventories and optimizing air quality improvement strategies. In this work, the abundance of organic compounds in different size particles was determined based on the Arizona field sampling (Chapter 5). The size distributions of aerosol saccharides were consistent with their related sources with many saccharides derived from PBAPs elevated in PM10 with the exception of levoglucosan and glycerol derived from biomass combustion. Interestingly, the fraction of non-levoglucosan saccharides to aerosol mass was greater in fine ambient aerosols compared to coarse ambient aerosols (Section 5.3.4). This was not expected as aerosol saccharides were thought to be largely associated to coarse particles, but may indicate a greater contribution to the fine size fraction of ambient PM from the primary biologically derived sources than currently acknowledged. An important consideration of this work is whether the atmospheric entrainment of soil and associated biota dominates local PM saccharide concentrations as proposed by previous studies (Simoneit et al. 2004; Rogge et al. 2007). To this end, soil samples were collected in the vicinity of the ambient sampling site in Higley, AZ, including currently uncultivated soil, agricultural soil, and road dust. The saccharide content was measured in the resuspended and size-fractionated dust particles. Comparison between the fine (size cut equivalent to PM2.5) and coarse (size cut equivalent to PM10) dust samples and between agricultural soil, 185 currently uncultivated soil, and road dust showed some important differences (Section 5.3.3). For example, higher saccharide levels were measured in the coarse fraction relative to the fine fraction of almost all soil samples with the exception of some road dusts. Road dust samples also contained higher concentrations of the major saccharides than agricultural and currently uncultivated soils. Possibly, this could originate from wind-blown PBAPs depositing to road surfaces. Agricultural and currently uncultivated soils had similar saccharide compositions, except that glucose was more enriched in agricultural soils while trehalose was more abundant in currently uncultivated soils. This may indicate the saccharide production and utilization mechanisms in agricultural and currently uncultivated soils are different. When comparing soil samples to aerosol samples, major differences were seen (Section 5.3.4). The total measured saccharide content (excluding levoglucosan) relative to PM mass was much higher in ambient aerosols (0.20% on average in PM2.5 and 0.11% in PM10) than in the resuspended soil particles of the same size range (<0.02%). The soil saccharide composition also differed from the profiles of two saccharide-enriched source factors resolved by PMF modeling. While there was some consistency in the major saccharides present, the relative abundance of sucrose and glucose was much higher in ambient particles (5.3.5.2). This suggests that other sources may also contribute to ambient aerosol saccharide levels at the Higley sampling site. For this reason, local plant and spore samples were collected to characterize another potentially large source of aerosols - PBAPs (Chapter 6). These samples 186 were resuspended and particles with size ranges equivalent to ambient PM2.5 and PM10 were collected. Contrary to the observation in soil samples, the total measured saccharide content (excluding levoglucosan) relative to PM and OC mass was one or two orders of magnitude higher in these PBAP source samples than in ambient aerosols. The dominant saccharide species in PBAP samples were also consistent with aerosol saccharide data. The measured PBAP saccharide profiles were compared with the two PMF-resolved saccharide-rich factors, with one plant (Prosopis juliflora) source profile showing good agreement with the sucrose rich bio-aerosol factor. The measured PBAP samples contained a concentration of sucrose and glucose that is consistent with the modeled source profile while the measured soil samples did not. This serves as confirmation that PBAPs are an important PM source in additional to soil and associated biota. However, the saccharide levels in the measured PBAP samples were significantly higher than the PMF results, suggesting that the ambient samples at Higley are a combination of high saccharide concentration PBAPs and lower saccharide concentration soils. 8.2 Recommendations for future research The first recommendation is to conduct an extended field sampling campaign. The longest sampling period in this work lasted from November to July the next year. Seasonal variations of the aerosol organic composition from winter to summer were captured with the exception of fall. However, an extended sampling period covering 187 four seasons would be needed in order to fully characterize the annual cycle of aerosol composition variations, particularly for compounds such as saccharides that do not have a consistent source contribution throughout the year. An enhanced ambient dataset with more sample days will also improve the overall PMF modeling performance by reducing uncertainties (Jaeckels et al. 2007). The second recommendation is to further evaluate the chemical composition of potential saccharide sources. This would include sampling of more soil and representative plants, pollen, bacteria, spores, fern, and insects in different seasons and at different locations. The ultimate goal of improved source characterization is to improve the understanding of different sources by adding the profile of these primary biologically derived carbohydrate sources to existing emission inventories. As PBAPs are considered to be significant PM sources, their inclusion in emission inventories is vital. This recommendation would also ease comparison of PBAP sources at different locations and test if the conclusions made in this work (that aerosol saccharides derive from both soil and PBAPs) can be extrapolated to other locations. The third recommendation is to perform alternative receptor modeling. Chemical Mass Balance (CMB) calculations are based on known source profiles which have been developed in this work. Similar to PMF, the CMB approach is based on: n /t=i 188 where c(ij) is the concentration of chemical species i in the fine particles at receptor site/, n is the total number of source types, a(jk) is the contribution to total fine particle mass concentrations at receptor sitey' from source k, and s(ik) is the contribution of species / from source k. The CMB determines the optimum relative emission contribution by minimizing the error term e(ij) which represents the difference between the ambient concentration of each component (i) in each sample (j) and the reconstructed ambient concentration of molecular markers determined through source attribution. The quantitative apportionment of sources by CMB calculations can be compared with the relative source contributions estimated by PMF to see if these disparate approaches corroborate one another. The ultimate goal is to refine the knowledge of major PM sources and their contributions to ambient PM. The CMB approach based on improved source profiles is also important to differentiate the contribution from different primary biologically derived sources and further determine the contribution of soil and associated biota versus PBAPs. Other recommendations to improve the research field include: using isotopically labeled saccharide standards that are representative of anhydrous saccharides, monosaccharides, and disaccharides, respectively, for a better evaluation of recovery efficiency; increasing the variety of other carbonaceous compounds to be analyzed for assessment of other important PM sources, i.e., n-alkanoic acids with both even- and odd- carbon numbers, dicarboxylic acids, additional saccharide compounds such as fructose and inositols; incorporating other biological parameters 189 such as phospholipid and neutral lipid fatty acids to assess the saccharide metabolism difference in the fungal and bacterial communities in soils (Medeiros et al. 2006). Table 8.1 Total measured saccharide concentrations, PM mass and organic carbon in PM2.5 samples (n= total number of samples analyzed). Higley, AZ San Augustine, TX Clarksville, TX Dallas, TX (Jan - Apr 2008) (Nov 2005 - Jul 2006) (Jan - Jul 2006) (Jan - Jul 2006) Total Measured Saccharide Concentration (ng m"3) 1.1-83.0(24.1) 15.4-355(70.4) 7.5-372(128) 15.8-196(52.4) PM2.5 Mass (ug m"3) 1.1-12.4(5.9) 2.5-23.2(9.9) 2.2-22.3(10.2) 3.5-34.0(11.2) OC(ugm'3) 0.7-2.9(1.6) 1.2-9.7(2.7) 1.4-10.2(3.5) 2.0-6.9(3.4) Total Saccharides/PM2.5 (%) 0.02-2.0(0.5) 0.1-4.1(1.0) 0.1-6.0(1.6) 0.1-1.6(0.6) Carbon Content of Total 0.04-2.3(0.7) 0.4-2.1(1.1) 0.1-4.2(1.8) 0.3-1.6(0.8) Saccharides/OC (%) Table 8.2 Estimated relative contribution of different sources to ambient PM mass for saccharide-related sources resolved by PMF Source factor characterized by the Relative source contribution enrichment of saccharide compounds Higley, AZ San Augustine, TX Dallas, TX . BBNP, TX PM2.5 Biomass burning 20% 22% 16% 9% Primary biologically derived carbohydrate 16% 14% 5% 7% sources PM10 Biomass burning 13% - - Primary biologically derived carbohydrate 21% - - sources 190 References AirData (2008). "AirData - US EPA: Access to Air Pollution Data." 2008, from http://www.epa.gov/air/data/index.html. Al-Mutlaq, K., A. I. Rushdi and B. R. T. Simoneit (2007). "Organic compound tracers of fine soil and sand particles during summer in the metropolitan area of Riyadh, Saudi Arabia." Environmental Geology 52(3): 559-571. Almeida, A. M., L. A. Cardoso, D. M. Santos, J. M. Torne and P. S. Fevereiro (2007). "Trehalose and its applications in plant biotechnology." In Vitro Cellular & Developmental Biology-Plant 43(3): 167-177. Alves, C, A. Carvalho and C. Pio (2002). "Mass balance of organic carbon fractions in atmospheric aerosols." Journal of Geophysical Research-Atmospheres 107(D21). Andreae, M. O. and A. Gelencser (2006). "Black carbon or brown carbon? The nature of light-absorbing carbonaceous aerosols." Atmospheric Chemistry and Physics 6(10): 3131-3148. Bates, D. V. (1995). "The Effects of Air-Pollution on Children." Environmental Health Perspectives 103: 49-53. Bauer, H. (2008). Three Cities' Study of Biogenic Emissions: Annual Trends and Contributions of Wood Smoke and Fungal Spores to Organic Carbon in PM10 Aerosols in Austria. 9th International Conference onCarbonaceous Particles in the Atmosphere, Berkeley, California. 191 Bauer, H., M. Claeys, R. Vermeylen, E. Schueller, G. Weinke, A. Berger and H. Puxbaum (2008). "Arabitol and mannitol as tracers for the quantification of airborne fungal spores." Atmospheric Environment 42(3): 588-593. Bauer, H., M. Fuerhacker, F. Zibuschka, H. Schmid and H. Puxbaum (2002). "Bacteria and fungi in aerosols generated by two different types of wastewater treatment plants." Water Research 36(16): 3965-3970. Bauer, H., E. Schueller, G. Weinke, A. Berger, R. Hitzenberger, I. L. Marr and H. Puxbaum (2008). "Significant contributions of fungal spores to the organic carbon and to the aerosol mass balance of the urban atmospheric aerosol." Atmospheric Environment 42(22): 5542-5549. Bhat, S. (2007). Characterization, seasonality and source apportionment of fine particulate organic matter at urban and rural sites. Civil and Environmental Engineering. Houston, TX, Rice University. Doctor of Philosophy. Bhave, P. V., D. P. Fergenson, K. A. Prather and G. R. Cass (2001). "Source Apportionment of Fine Particulate Matter by Clustering Single-Particle Data: Tests of Receptor Model Accuracy." Environmental Science & Technology 35(10): 2060-2072. Bidochka, M. J., N. H. Low and G. G. Khachatourians (1990). "Carbohydrate Storage in the Entomopathogenic Fungus Beauveria-Bassiana." Applied and Environmental Microbiology 56(10): 3186-3190. 192 Bieleski, R. L. (1982). "Sugar alcohols [in plants]." Encycl. Plant Physiol, New Ser. FIELD Full Journal Title .Encyclopedia of Plant Physiology, New Series 13A(PlantCarbohydr. 1: Intracell. Carbohydr.): 158-92. Bieleski, R. L. (1995). "Onset of Phloem Export from Senescent Petals of Daylily." Plant Physiology 109(2): 557-565. Birch, M. E. and R. A. Cary (1996). "Elemental Carbon-Based Method for Monitoring Occupational Exposures to Particulate Diesel Exhaust." Aerosol Science and Technology 25(3): 221 - 241. Black, G. E. and A. Fox (1996). "Recent progress in the analysis of sugar monomers from complex matrices using chromatography in conjunction with mass spectrometry or stand-alone tandem mass spectrometry." Journal of Chromatography A 720(1-2): 51-60. Black, M. and H. W. Pritchard (2002). Desiccation and Survival in Plants: Drying Without Dying, CABI Publishing. Blau, K. and J. Halket (1993). Handbook of Derivatives for Chromatography. New York, Wiley. Blumenthal, H. J. (1976). "RESERVE CARBOHYDRATES IN FUNGI." Smith, John E. And David R. Berry (Ed.). The Filamentous Fungi, Vol. 2. Biosynthesis and Metabolism. Xiv+520p. Illus. HalstedPress/Wiley: New York, NY., U.S.A. Isbn 0-470-15505-X: 292-307. Brown, S. G. and H. R. Hafner (2005). Multivariate receptor modeling workbook (prepared for: United States Environmental Protection Agency). Sonoma Technolgy, Inc Brown, S. G., P. Herckes, L. Ashbaugh, M. P. Hannigan, S. M. Kreidenweis and J. L. Collett Jr (2002). "Characterization of organic aerosol in Big Bend National Park, Texas." Atmospheric Environment 36(38): 5807-5818. Brown, S. G., Herckes, P., Kreidenweis, S.M., Collett, Jr., J.L. (2001). Characterization of Carbonaceous Aerosol during the Big Bend Regional Aerosol and Visibility Observational Study. Depart of Atmospheric Science, Colorado State University Burtscher, H., A. Leonardi, D. Steiner, U. Baltensperger and A. Weber (1993). "Aging of Combustion Particles in the Atmosphere - Results from a Field-Study in Zurich." Water Air and Soil Pollution 68(1-2): 137-147. Buzcu, B., M. P. Fraser, P. Kulkarni and S. Chellam (2003). "Source identification and apportionment of fine particulate matter in Houston, TX, using positive matrix factorization." Environmental Engineering Science 20(6): 533-545. Carvalho, A., C. Pio and C. Santos (2003). "Water-soluble hydroxylated organic compounds in German and Finnish aerosols." Atmospheric Environment 37(13): 1775-1783. Caseiro, A. (2008). European Atmospheric Aerosol Chemical Composition. Environment and Planning. Vienna, Austria, Vienna University of Technology: 196. 194 Caseiro, A., I. L. Marr, M. Claeys, A. Kasper-Giebl, H. Puxbaum and C. A. Pio (2007). "Determination of saccharides in atmospheric aerosol using anion-exchange high-performance liquid chromatography and pulsed-amperometric detection." Journal of Chromatography A 1171(1-2): 37-45. Chandrasekhar, I. and B. P. Gaber (1988). "STABILIZATION OF THE BIO-MEMBRANE BY SMALL MOLECULES - INTERACTION OF TREHALOSE WITH THE PHOSPHOLIPID-BILAYER." Journal of Biomolecular Structure & Dynamics 5(6): 1163-&. Chang, S. and D. T. Allen (2006). "Chlorine chemistry in urban atmospheres: Aerosol formation associated with anthropogenic chlorine emissions in southeast Texas." Atmospheric Environment 40(Supplement 2): 512-523. Chaturvedi, V., A. Bartiss and B. Wong (1997). "Expression of bacterial mtlD in Saccharomyces cerevisiae results in mannitol synthesis and protects a glycerol-defective mutant from high-salt and oxidative stress." Journal of Bacteriology 179(1): 157-162. Chen, L.-W. A., B. G. Doddridge, R. R. Dickerson, J. C. Chow and R. C. Henry (2002). "Origins of fine aerosol mass in the Baltimore-Washington corridor: implications from observation, factor analysis, and ensemble air parcel back trajectories." Atmospheric Environment 36(28): 4541-4554. 195 Chen, L. W. A., J. G. Watson, J. C. Chow and K. L. Magliano (2007). "Quantifying PM2.5 source contributions for the San Joaquin Valley with multivariate receptor models." Environmental Science & Technology 41(8): 2818-2826. Cheng, C, J. T. Chen and H. R. Huang (1992). "High-Performance Liquid-Chromatography of Carbohydrates in Enzymatic Hydrolysate of Waste Paper." Chromatographic! 34(9-10): 534-536. Cheng, C, H. R. Tsai and K. C. Chang (2006). "On-line cut-off technique and organic modifier addition aided signal enhancement for trace analysis of carbohydrates in cellulase hydrolysate by ion exclusion chromatography-electrospray ionization mass spectrometry." Journal of Chromatography A 1119(1-2): 188-196. Cheshire, M. V. (1979). Nature and origin of carbohydrates in soils. London, Academic Press. Chizhov, O. S., N. V. Molodtsov and N. K. Kochetkov (1967). "Mass spectrometry of trimethylsilyl ethers of carbohydrates." Carbohydr. Res. FIELD FullJournal Title:Carbohydrate Research 4(3): 273-6. Chow, J. C, J. G. Watson, H. Kuhns, V. Etyemezian, D. H. Lowenthal, D. Crow, S. D. Kohl, J. P. Engelbrecht and M. C. Green (2004). "Source profiles for industrial, mobile, and area sources in the Big Bend Regional Aerosol Visibility and Observational study." Chemosphere 54(2): 185-208. Chow, J. C, J. G. Watson, D. H. Lowenthal, P. A. Solomon, K. L. Magliano, S. D. Ziman and L. W. Richards (1992). "PmlO Source Apportionment in California 196 San-Joaquin Valley." Atmospheric Environment Part a-General Topics 26(18): 3335-3354. Chow, P. S. and S. M. Landhausser (2004). "A method for routine measurements of total sugar and starch content in woody plant tissues." Tree Physiology 24(10): 1129-1136. Clarke, A. J., V. Sarabia, W. Keenleyside, P. R. Maclachlan and C. Whitfield (1991). "The Compositional Analysis of Bacterial Extracellular Polysaccharides by High-Performance Anion-Exchange Chromatography." Analytical Biochemistry 199(1): 68-74. Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. New Jersey, Hillsdale. Cooper, J. E. and E. E. Bray (1963). "A postulated role of fatty acids in petroleum formation." Geochimica et Cosmochimica Acta 27(11): 1113-1127. Coulter, C. T. (2004). EPA-CMB8.2 Users Manual. Air Quality Modeling Group, Office of Air Quality Planning & Standards. Research Triangle Park, NC. Cowie, G. L. and J. I. Hedges (1984). "Carbohydrate Sources in a Coastal Marine-Environment." Geochimica Et Cosmochimica Acta 48(10): 2075-2087. Cox, C. S. and C. M. Wathes (1995). Bioaerosols Handbook, Lewis Publishers. Crowe, J. H., L. M. Crowe and D. Chapman (1984). "Preservation of Membranes in Anhydrobiotic Organisms - the Role of Trehalose." Science 223(4637): 701-703. 197 Dahlman, L., J. Persson, T. Nasholm and K. Palmqvist (2003). "Carbon and nitrogen distribution in the green algal lichens Hypogymnia physodes and Platismatia glauca in relation to nutrient supply." Planta 217(1): 41-48. Debosz, K., L. Vognsen and R. Labouriau (2002). "Carbohydrates in hot water extracts of soil aggregates as influenced by long-term management." Communications in Soil Science and Plant Analysis 33(3-4): 623-634. • Decesari, S., M. C. Facchini, S. Fuzzi and E. Tagliavini (2000). "Characterization of water-soluble organic compounds in atmospheric aerosol: A new approach." Journal of Geophysical Research-Atmospheres 105(D1): 1481-1489. Deguillaume, L., M. Leriche, P. Amato, P. A. Ariya, A. M. Delort, U. Poschl, N. Chaumerliac, H. Bauer, A. I. Flossmann and C. E. Morris (2008). "Microbiology and atmospheric processes: chemical interactions of primary biological aerosols." Biogeosciences 5(4): 1073-1084. DeJongh, D. C, T. Radford, J. D. Hribar, Hanessia.S, M. Bieber, G. Dawson and C. C. Sweeley (1969). "Analysis of Trimethylsilyl Derivatives of Carbohydrates by Gas Chromatography and Mass Spectrometry." Journal of the American Chemical Society 91(7): 1728-40. Dey, P. M. and J. B. Harborne (1997). Plant Biochemistry, Academic Press. Dijksterhuis, J. and R. A. Samson (2007). Food Mycology: A Multifaceted Approach to Fungi and Food, CRC Press. Dockery, D. W. and C. A. Pope (1994). "Acute Respiratory Effects of Particulate Air-Pollution." Annual Review of Public Health 15: 107-132. 198 Dye, C. and K. E. Yttri (2005). "Determination of monosaccharide anhydrides in atmospheric aerosols by use of high-performance liquid chromatography combined with high-resolution mass spectrometry." Analytical Chemistry 77(6): 1853-1858. Eberly, S. (2005). EPA PMF 1.1 User's Guide. US EPA, National Exposure Research Laboratory, Research Triangle Park Elbe in, A. D. (1974). "The metabolism of alpha,alpha-trehalose." Adv Carbohydr Chem Biochem 30: 227-56. Elbert, W., P. E. Taylor, M. O. Andreae and U. Poschl (2007). "Contribution of fungi to primary biogenic aerosols in the atmosphere: wet and dry discharged spores, carbohydrates, and inorganic ions." Atmospheric Chemistry and Physics 7(17): 4569-4588. Eleutherio, E. C. A., P. S. Araujo and A. D. Panek (1993). "Protective Role of Trehalose During Heat-Stress in Saccharomyces-Cerevisiae." Cryobiology 30(6): 591-596. Eleutherio, E. C. A., P. S. Dearaujo and A. D. Panek (1993). "Role of the Trehalose Carrier in Dehydration Resistance of Saccharomyces-Cerevisiae." Biochimica Et Biophysica Acta 1156(3): 263-266. Elias, V. O., B. R. T. Simoneit, A. S. Pereira, J. A. Cabral and J. N. Cardoso (1999). "Detection of high molecular weight organic tracers in vegetation smoke samples by high-temperature gas chromatography-mass spectrometry." Environmental Science & Technology 33(14): 2369-2376. 199 Engling, G., C. M. Carrico, S. M. Kreldenweis, J. L. Collett, D. E. Day, W. C. Malm, E. Lincoln, W. M. Hao, Y. Iinuma and H. Herrmann (2006). "Determination of levoglucosan in biomass combustion aerosol by high-performance anion-exchange chromatography with pulsed amperometric detection." Atmospheric Environment 40: S299-S311. Feofilova, E. P., V. M. Tereshina, N. S. Khokhlova and A. S. Memorskaya (2000). "Different mechanisms of the biochemical adaptation of mycelial fungi to temperature stress: Changes in the cytosol carbohydrate composition." Microbiology 69(5): 504-508. Fetzer, J. C. (2007). "The chemistry and analysis of large PAHs." Poly cyclic Aromatic Compounds 27'(2): 143-162. Fiehn, O., J. Kopka, R. N. Trethewey and L. Willmitzer (2000). "Identification of uncommon plant metabolites based on calculation of elemental compositions using gas chromatography and quadrupole mass spectrometry." Analytical Chemistry 72(15): 3573-3580. Fine, P. M., G. R. Cass and B. R. T. Simoneit (2004). "Chemical characterization of fine particle emissions from the wood stove combustion of prevalent United States tree species." Environmental Engineering Science 21(6): 705-721. Fraser, M. P., G. R. Cass and B. R. T. Simoneit (1998). "Gas-phase and particle-phase organic compounds emitted from motor vehicle traffic in a Los Angeles roadway tunnel." Environmental Science & Technology 32(14): 2051-2060. 200 Fraser, M. P., G. R. Cass and B. R. T. Simoneit (1999). "Particulate organic compounds emitted from motor vehicle exhaust and in the urban atmosphere." Atmospheric Environment 33(17): 2715-2724. Fraser, M. P., G. R. Cass, B. R. T. Simoneit and R. A. Rasmussen (1997). "Air quality model evaluation data for organics .4. C-2-C-36 non-aromatic hydrocarbons." Environmental Science & Technology 31(8): 2356-2367. Fraser, M. P. and K. Lakshmanan (2000). "Using levoglucosan as a molecular marker for the long-range transport of biomass combustion aerosols." Environmental Science & Technology 34(21): 4560-4564. Fraser, M. P., Z. W. Yue and B. Buzcu (2003). "Source apportionment of fine particulate matter in Houston, TX, using organic molecular markers." Atmospheric Environment 37(15): 2117-2123. Fraser, M. P., Z. W. Yue, R. J. Tropp, S. D. Kohl and J. C. Chow (2002). "Molecular composition of organic fine particulate matter in Houston, TX." Atmospheric Environment 36(38): 5751-5758. Fuzzi, S., S. Decesari, M. C. Facchini, F. Cavalli, L. Emblico, M. Mircea, M. O. Andreae, I. Trebs, A. Hoffer, P. Guyon, P. Artaxo, L. V. Rizzo, L. L. Lara, T. Pauliquevis, W. Maenhaut, N. Raes, X. G. Chi, O. L. Mayol-Bracero, L. L. Soto-Garcia, M. Claeys, I. Kourtchev, J. Rissler, E. Swietlicki, E. Tagliavini, G. Schkolnik, A. H. Falkovich, Y. Rudich, G. Fisch and L. V. Gatti (2007). "Overview of the inorganic and organic composition of size-segregated 201 aerosol in Rondonia, Brazil, from the biomass-burning period to the onset of the wet season." Journal of Geophysical Research-Atmospheres 112(D1). Garcia-Raso, A., M. I. Paez, I. Martinezcastro, J. Sanz and M. M. Calvo (1992). "Gas-Chromatographic Retention of Carbohydrate Trimethylsilyl Ethers .4. Disaccharides." Journal of Chromatography 607(2): 221-225. Gelencser, A. (2004). Carboaceous Aerosol. Dordrecht, The Netherlands, Springer. Gogou, A., N. Stratigakis, M. Kanakidou and E. G. Stephanou (1996). "Organic aerosols in Eastern Mediterranean: components source reconciliation by using molecular markers and atmospheric back trajectories." Organic Geochemistry 25(1-2): 79-96. Goldberg, M. S., J. C. Bailar, 3rd, R. T. Burnett, J. R. Brook, R. Tamblyn, Y. Bonvalot, P. Ernst, K. M. Flegel, R. K. Singh and M. F. Valois (2000). "Identifying subgroups of the general population that may be susceptible to short-term increases in particulate air pollution: a time-series study in Montreal, Quebec." Res Rep Health Eff Inst{91): 7-113; discussion 115-20. Gorny, R. L., G. Mainelis, S. A. Grinshpun, K. Willeke, J. Dutkiewicz and T. Reponen (2003). "Release of Streptomyces albus propagules from contaminated surfaces." Environmental Research 91(1): 45-53. Gorny, R. L., T. Reponen, K. Willeke, D. Schmechel, E. Robine, M. Boissier and S. A. Grinshpun (2002). "Fungal fragments as indoor air biocontaminants." Applied and Environmental Microbiology 68(7): 3522-3531. 202 Graham, B., P. Guyon, P. E. Taylor, P. Artaxo, W. Maenhaut, M. M. Glovsky, R. C. Flagan and M. O. Andreae (2003). "Organic compounds present in the natural Amazonian aerosol: Characterization by gas chromatography-mass spectrometry." Journal of Geophysical Research-Atmospheres 108(D24). Graham, B., O. L. Mayol-Bracero, P. Guyon, G. C. Roberts, S. Decesari, M. C. Facchini, P. Artaxo, W. Maenhaut, P. Koll and M. O. Andreae (2002). "Water-soluble organic compounds in biomass burning aerosols over Amazonia - 1. Characterization by NMR and GC-MS." Journal of Geophysical Research-Atmospheres 107(D20): -. Griffin, R. J., D. R. Cocker, R. C. Flagan and J. H. Seinfeld (1999). "Organic aerosol formation from the oxidation of biogenic hydrocarbons." Journal of Geophysical Research-Atmospheres 104(D3): 3555-3567. Hackley, B., A. Feinstein and J. Dixon (2007). "Air Pollution: Impact on Maternal and Perinatal Health." Journal of Midwifery & Women's Health 52(5): 435-443. Harborne, J. B. (1998). Phytochemical methods: a guide to modern techniques of plant analysis, Springer. Hildemann, L. M., G. R. Cass, M. A. Mazurek and B. R. T. Simonelt (1993). "Mathematical-Modeling of Urban Organic Aerosol - Properties Measured by High-Resolution Gas-Chromatography." Environmental Science & Technology 27(10): 2045-2055. 203 Hildemann, L. M, M. A. Mazurek, G. R. Cass and B. R. T. Simoneit (1991). "Quantitative Characterization of Urban Sources of Organic Aerosol by High-Resolution Gas-Chromatography." Environmental Science & Technology 25(7): 1311-1325. HON Foundation (2004). "European Pollen Calendar - Data for Austria." 2009, from http://www.hon.ch/Library/Theme/Allergy/Glossary/austria.html Hunt, J. M. (1979). Petroleum Geochemistry and Geology. San Francisco, W.H. Freeman & Co. Iinuma, Y., G. Engling, H. Puxbaum and H. Herrmann (2009). "A highly resolved anion-exchange chromatographic method for determination of saccharidic tracers for biomass combustion and primary bio-particles in atmospheric aerosol." Atmospheric Environment 43(6): 1367-1371. Jacobson, M. C, H. C. Hansson, K. J. Noone and R. J. Charlson (2000). "Organic atmospheric aerosols: Review and state of the science." Reviews of Geophysics 38(2): 267-294. Jaeckels, J. M., M. S. Bae and J. J. Schauer (2007). "Positive matrix factorization (PMF) analysis of molecular marker measurements to quantify the sources of organic aerosols." Environmental Science & Technology 41(16): 5763-5769. Jaenicke, R. (2005). "Abundance of cellular material and proteins in the atmosphere." Science 308(5718): 73-73. 204 Jaffe, R., A. I. Rushdi, P. M. Medeiros and B. R. T. Simoneit (2006). "Natural product biomarkers as indicators of sources and transport of sedimentary organic matter in a subtropical river." Chemosphere 64(11): 1870-1884. Jia, Y. (2008). Characterization of Sugars in Fine Particles - Biomass Burning and Soil Organic Matter as Two Major Sources. Civil and Environmental Engineering. Houston, TX, Rice University. Master of Science: 78. Jia, Y., A. L. Clements and M. P. Fraser (2010). "Saccharide composition in atmospheric particulate matter in the southwest US and estimates of source contributions." Journal of Aerosol Science 41(1): 62-73. Kadowaki, S. (1994). "Characterization of Carbonaceous Aerosols in theNagoya Urban Area .2. Behavior and Origin of Particulate N-Alkanes." Environmental Science & Technology 28(1): 129-135. Kanakidou, M., J. H. Seinfeld, S. N. Pandis, I. Barnes, F. J. Dentener, M. C. Facchini, R. Van Dingenen, B. Ervens, A. Nenes, C. J. Nielsen, E. Swietlicki, J. P. Putaud, Y. Balkanski, S. Fuzzi, J. Horth, G. K. Moortgat, R. Winterhalter, C. E. L. Myhre, K. Tsigaridis, E. Vignati, E. G. Stephanou and J. Wilson (2005). "Organic aerosol and global climate modelling: a review." Atmospheric Chemistry and Physics 5: 1053-1123. Kavouras, I. G. and E. G. Stephanou (2002). "Direct evidence of atmospheric secondary organic aerosol formation in forest atmosphere through heteromolecular nucleation." Environmental Science & Technology 36(23): 5083-5091. 205 Kawamura, K. and R. B. Gagosian (1987). "Implications of Omega-Oxocarboxylic Acids in the Remote Marine Atmosphere for Photooxidation of Unsaturated Fatty-Acids." Nature 325(6102): 330-332. Kawamura, K., H. Kasukabe and L. A. Barrie (1996). "Source and reaction pathways of dicarboxylic acids, ketoacids and dicarbonyls in arctic aerosols: One year of observations." Atmospheric Environment 30(10-11): 1709-1722. Kawamura, K. and F. Sakaguchi (1999). "Molecular distributions of water soluble dicarboxylic acids in marine aerosols over the Pacific Ocean including tropics." Journal of Geophysical Research-Atmospheres 104(D3): 3501-3509. Kawamura, K. and K. Usukura (1993). "Distributions of low molecular weight dicarboxylic acids in the North Pacific aerosol samples." Journal of Oceanography 49(3): 271-283. Kiem, R. and I. Kogel-Knabner (2003). "Contribution of lignin and polysaccharides to the refractory carbon pool in C-depleted arable soils." Soil Biology & Biochemistry 35(1): 101-118. Kim, E. and P. K. Hopke (2007). "Source identifications of airborne fine particles using positive matrix factorization and US environmental protection agency positive matrix factorization." Journal of the Air & Waste Management Association 57(7): 811-819. Kleeman, M. J., J. J. Schauer and G. R. Cass (1999). "Size and composition distribution of fine particulate matter emitted from wood burning, meat 206 charbroiling, and cigarettes." Environmental Science & Technology 33(20): 3516-3523. Kleinman, M. T., D. K. Bhalla, W. J. Mautz and R. F. Phalen (1995). "Cellular and immunologic injury with PM-10 inhalation." Inhalation Toxicology 7(5): 589-602. Krewski, D., R. T. Burnett, M. S. Goldberg, B. K. Hoover, J. Siemiatycki, M. Jerrett, M. Abrahamowicz and W. H. White (2003). "Overview of the reanalysis of the Harvard Six Cities Study and American Cancer Society Study of particulate air pollution and mortality." Journal of Toxicology and Environmental Health-Part A 66(16-19): 1507-1551. Langer, S. H., S. Connell and I. Wender (1958). "Preparation and Properties of Trimethylsilyl Ethers and Related Compounds." The Journal of Organic Chemistry 23(1): 50-58. Lewis, D. H. and D. C. Smith (1967). "Sugar Alcohols (Polyols) in Fungi and Green Plants .1. Distribution Physiology and Metabolism." New Phytologist 66(2): 143-&. Li, M., S. R. McDow, D. J. Tollerud and M. A. Mazurek (2006). "Seasonal abundance of organic molecular markers in urban particulate matter from Philadelphia, PA." Atmospheric Environment 40(13): 2260-2273. Limbeck, A. and H. Puxbaum (1999). "Organic acids in continental background aerosols." Atmospheric Environment 33(12): 1847-1852. 207 Lipsett, M. J., F. C. Tsai, L. Roger, M. Woo and B. D. Ostro (2006). "Coarse particles and heart rate variability among older adults with coronary artery disease in the Coachella Valley, California." Environmental Health Perspectives 114(8): 1215-1220. Loos, H., R. Kramer, H. Sahm and G. A. Sprenger (1994). "Sorbitol Promotes Growth of Zymomonas-Mobilis in Environments with High-Concentrations of Sugar - Evidence for a Physiological-Function of Glucose-Fructose Oxidoreductase in Osmoprotection." Journal of Bacteriology 176(24): 7688-7693. Lough, G. C, Christensen, C.G., Schauer, J.J., Tortorelli, J., Mani, E., Lawson, D.R., Clark, N.N., Gabele, P.A. (2007). "Development of molecular marker source profiles for emissions from on-road gasoline and diesel vehicle fleets." Journal of the Air & Waste Management Association 57(10): 1190-1199. Lough, G. C, J. J. Schauer, J. S. Park, M. M. Shafer, J. T. Deminter and J. P. Weinstein (2005). "Emissions of metals associated with motor vehicle roadways." Environmental Science & Technology 39(3): 826-836. Lough, G. C, Schauer, J.J. (2007). "Sensitivity of source apportionment of urban particulate matter to uncertainty in motor vehicle emissions profiles." Journal of the Air & Waste Management Association 57(10): 1096-2247. Lugo, A. E. and M. J. Sanchez (1986). "LAND-USE AND ORGANIC-CARBON CONTENT OF SOME SUBTROPICAL SOILS." Plant and Soil 96(2): 185-196. 208 Madelin, T. and M. Madelin (1995). Biological analysis of fungi and associated molds. Bioaerosols Handbook. C. S. Cox and C. M. Wathes, Lewis Publishers: 361-386. Mader, B. T., J. Z. Yu, J. H. Xu, Q. F. Li, W. S. Wu, R. C. Flagan and J. H. Seinfeld (2004). "Molecular composition of the water-soluble fraction of atmospheric carbonaceous aerosols collected during ACE-Asia." Journal of Geophysical Research-Atmospheres 109(D6). Maenhaut, W., A. Onjia, X. Chi, J. Cafmeyer, V. Pashynska, R. Vermeylen, G. Vas and M. Claeys (2002). Speciation of water-soluble organic aerosol at Skukuza, South Africa, during SAFARI 2000. 7th Scientific Conference, International Global Atmospheric Chemistry Project (IGAC). Heraklion, Crete, Greece. Mandels, G. R., R. Vitols and F. W. Parrish (1965). "Trehalose as an Endogenous Reserve in Spores of Fungus Myrothecium Verrucaria." Journal of Bacteriology 90(6): 1589-&. Martin, F., M. Ramstedt, K. Soderhall and D. Canet (1988). "Carbohydrate and Amino-Acid Metabolism in the Ectomycorrhizal Ascomycete Sphaerosporella-Brunnea During Glucose-Utilization - a C-13 Nmr-Study." Plant Physiology 86(3): 935-940. Martinez-Castro, I., M. I. Paez, J. Sanz and A. Garciaraso (1989). "Gas-Chromatographic Behavior of Carbohydrate Trimethylsilyl Ethers .2. Aldohexoses." Journal of Chromatography 462: 49-60. 209 Mazurek, M. A., G. R. Cass and B. R. T. Simoneit (1989). "Interpretation of High-Resolution Gas-Chromatography and High-Resolution Gas-Chromatography Mass-Spectrometry Data Acquired from Atmospheric Organic Aerosol Samples." Aerosol Science and Technology 10(2): 408-420. Mazurek, M. A., B. R. T. Simoneit, G. R. Cass and H. A. Gray (1987). "Quantitative High-Resolution Gas-Chromatography and High-Resolution Gas-Chromatography Mass-Spectrometry Analyses of Carbonaceous Fine Aerosol-Particles." International Journal of Environmental Analytical Chemistry 29(1-2): 119-139. McMurry, P. H. and X. Q. Zhang (1989). "Size Distributions of Ambient Organic and Elemental Carbon." Aerosol Science and Technology 10(2): 430-437. Medeiros, P. M., M. H. Conte, J. C. Weber and B. R. T. Simoneit (2006). "Sugars as source indicators of biogenic organic carbon in aerosols collected above the Howland Experimental Forest, Maine." Atmospheric Environment 40(9): 1694-1705. Medeiros, P. M., M. F. Fernandes, R. P. Dick and B. R. T. Simoneit (2006). "Seasonal variations in sugar contents and microbial community in a ryegrass soil." Chemosphere 65(5): 832-839. Medeiros, P. M. and B. R. T. Simoneit (2007). "Analysis of sugars in environmental samples by gas chromatography-mass spectrometry." Journal of Chromatography A 1141(2): 271-278. 210 Medeiros, P. M. and B. R. T. Simoneit (2007). "Gas chromatography coupled to mass spectrometry for analyses of organic compounds and biomarkers as tracers for geological, environmental, and forensic research." Journal of Separation Science 30(10): 1516-1536. Mellor, R. B. (1992). "IS TREHALOSE A SYMBIOTIC DETERMINANT IN SYMBIOSES BETWEEN HIGHER-PLANTS AND MICROORGANISMS." Symbiosis 12(2): 113-129. Miguel, A. H., T. W. Kirchstetter, R. A. Harley and S. V. Hering (1998). "On-Road Emissions of Particulate Poly cyclic Aromatic Hydrocarbons and Black Carbon from Gasoline and Diesel Vehicles." Environ. Sci. Technol. 32(4): 450-455. Modzeleski, J. E., W. A. Laurie and B. Nagy (1971). "Carbohydrates from Santa-Barbara Basin Sediments - Gas Chromatographic-Mass Spectrometric Analysis of Trimethylsilyl Derivatives." Geochimica Et Cosmochimica Acta 35(8): 825-&. Murelli, C, V. Adamo, P. V. Finzi, F. M. Albini, A. Bochicchio and A. M. Picco (1996). "Sugar biotransformations by fungi on leaves of the resurrection plant Sporobolus stapfianus." Phytochemistry 43(4): 741-745. NARSTO (2004). Particulate Matter Science for Policy Makers - ANARSTO Assessment, Cambridge University Press. NCDC (2007, January 11, 2007). "National Climatic Data Center: Climate of 2006 - Wildfire Season Summary." 2009, from http://www.ncdc.noaa.gov/oa/climate/research/2006/fire06.html. 211 Neter, J., M. H. Kutner, C. J. Nachtsheim and W. Wasserman (1996). Applied Linear Statistical Models. Chicago, Irwin. Niederer, M., W. Pankow and A. Wiemken (1989). "Trehalose Synthesis in Mycorrhiza of Norway Spruce - an Indicator of Vitality." European Journal of Forest Pathology 19(1): 14-20. Nolte, C. G., J. J. Schauer, G. R. Cass and B. R. T. Simoneit (1999). "Highly polar organic compounds present in meat smoke." Environmental Science & Technology 33(19): 3313-3316. Nolte, C. G., J. J. Schauer, G. R. Cass and B. R. T. Simoneit (2001). "Highly polar organic compounds present in wood smoke and in the ambient atmosphere." Environmental Science & Technology 35(10): 1912-1919. Nolte, C. G., J. J. Schauer, G. R. Cass and B. R. T. Simoneit (2002). "Trimethylsilyl derivatives of organic compounds in source samples and in atmospheric fine particulate matter." Environmental Science & Technology 36(20): 4273-4281. O'Rourke, M. K. (1990). "Comparative pollen calendars from Tucson, Arizona: Durham vs. Burkard samplers." Aerobiologia 6: 136-140. Oades, J. M. (1984). "SOIL ORGANIC-MATTER AND STRUCTURAL STABILITY - MECHANISMS AND IMPLICATIONS FOR MANAGEMENT." Plant and Soil 76(1-3): 319-337. Oberdorster, G. and M. J. Utell (2002). "Ultrafine particles in the urban air: To the respiratory tract - And beyond?" Environmental Health Perspectives 110(8): A440-A441. 212 Ogulei, D., P. K. Hopke, L. Zhou, P. Paatero, S. S. Park and J. M. Ondov (2005). "Receptor modeling for multiple time resolved species: The Baltimore supersite." Atmospheric Environment 39(20): 3751-3762. Oros, D. R. and B. R. T. Simoneit (1999). "Identification of molecular tracers in organic aerosols from temperate climate vegetation subjected to biomass burning." Aerosol Science and Technology 31(6): 433-445. Oros, D. R. and B. R. T. Simoneit (2001). "Identification and emission factors of molecular tracers in organic aerosols from biomass burning Part 1. Temperate climate conifers." Applied Geochemistry 16(13): 1513-1544. Oros, D. R. and B. R. T. Simoneit (2001). "Identification and emission factors of molecular tracers in organic aerosols from biomass burning Part 2. Deciduous trees." Applied Geochemistry 16(13): 1545-1565. Paatero, P. (1997). "Least squares formulation of robust non-negative factor analysis." Chemometrics and Intelligent Laboratory Systems 37(1): 23-35. Paatero, P. and P. K. Hopke (2003). "Discarding or downweighting high-noise variables in factor analytic models." Analytica Chimica Acta 490(1-2): 277-289. Paatero, P. and U. Tapper (1994). "Positive Matrix Factorization - aNonnegative Factor Model with Optimal Utilization of Error-Estimates of Data Values." Environmetrics 5(2): 111-126. Pacini, E. (2000). "From anther and pollen ripening to pollen presentation." Plant Systematics and Evolution 222(1): 19-43. 213 Pacini, E., M. Guarnieri and M. Nepi (2006). "Pollen carbohydrates and water content during development, presentation, and dispersal: a short review." Protoplasma 228(1-3): 73-77. Pandis, S. N., R. A. Harley, G. R. Cass and J. H. Seinfeld (1992). "Secondary Organic Aerosol Formation and Transport." Atmospheric Environment Part a-General Topics 26(13): 2269-2282. Pashynska, V., R. Vermeylen, G. Vas, W. Maenhaut and M. Claeys (2002). "Development of a gas chromatographic/ion trap mass spectrometric method for the determination of levoglucosan and saccharidic compounds in atmospheric aerosols. Application to urban aerosols." Journal of Mass Spectrometry Yi'(12): 1249-1257. Paul, E. A. and F. E. Clark (1996). Soil Microbiology and Biochemistry, Academic Press, San Diego. Penner, J. E., M. O. Andreae, H. Annegarn, L. Barrie and et al. (2001). Aerosols, their direct and indirect effects. In: Houghton, J.T., et al. (Eds.), Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press. Peters, A., E. Liu, R. L. Verrier, J. Schwartz, D. R. Gold, M. Mittleman, J. Baliff, J. A. Oh, G. Allen, K. Monahan and D. W. Dockery (2000). "Air pollution and incidence of cardiac arrhythmia." Epidemiology 11(1): 11-17. 214 Peters, K. E. and J. M. Moldowan (1993). The biomarker guide: Interpreting molecular fossils in petroleum and ancient sediments. Englewood Cliffs, NJ, Prentice Hall. Pfyffer, G. E. and D. M. Rast (1980). "The Polyol Pattern of Some Fungi Not Hitherto Investigated for Sugar Alcohols." Experimental Mycology 4(2): 160-170. Pigman, W. and D. Horton (1980). The Carbohydrates - Chemistry and Biochemistry. New York, Academic Press. Pio, C. A., M. Legrand, C. A. Alves, T. Oliveira, J. Afonso, A. Caseiro, H. Puxbaum, A. Sanchez-Ochoa and A. Gelencser (2008). "Chemical composition of atmospheric aerosols during the 2003 summer intense forest fire period." Atmospheric Environment 42(32): 7530-7543. Pitchford, M. L., Tombach, I.H., Barna, M.G., Gebhart, K.A., Green, M.C., Knipping, E., Kumar, N., Malm, W.C., Pun, B., Schichtel, B.A., Seigneur, C. (2004). Big Bend Regional Aerosol and Visibility Observational (BRAVO) Study Pope, C. A. (2000). "Epidemiology of fine particulate air pollution and human health: Biologic mechanisms and who's at risk?" Environmental Health Perspectives 108: 713-723. Pope, C. A., R. T. Burnett, M. J. Thun, E. E. Calle, D. Krewski, K. Ito and G. D. Thurston (2002). "Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution." Jama-Journal of the American Medical Association 287(9): 1132-1141. 215 Pdschl, U., T. Fehrebach, T. Franze, E. Mikhailov, U. McKeon, C. Schauer and A. Zerrath (2005). "Carbonaeceous aerosol components: chemical composition, reactivity, and hygroscopicity (CARBAERO)." Geophysical Research Abstracts 7(05164). Puxbaum, H. and M. Tenze-Kunit (2003). "Size distribution and seasonal variation of atmospheric cellulose." Atmospheric Environment 37(26): 3693-3699. Ramadan, Z., X. H. Song and P. K. Hopke (2000). "Identification of sources of Phoenix aerosol by positive matrix factorization." Journal of the Air & Waste Management Association 50(8): 1308-1320. Ravindra, K., R. Sokhi and R. Van Grieken (2008). "Atmospheric polycyclic aromatic hydrocarbons: Source attribution, emission factors and regulation." Atmospheric Environment 42(13): 2895-2921. Rogge, W. F., L. M. Hildemann, M. A. Mazurek, G. R. Cass and B. R. T. Simoneit (1993). "Sources of Fine Organic Aerosol .2. Noncatalyst and Catalyst-Equipped Automobiles and Heavy-Duty Diesel Trucks." Environmental Science & Technology 21 {A): 636-651. Rogge, W. F., L. M. Hildemann, M. A. Mazurek, G. R. Cass and B. R. T. Simoneit (1993). "Sources of Fine Organic Aerosol .3. Road Dust, Tire Debris, and Organometallic Brake Lining Dust - Roads as Sources and Sinks." Environmental Science & Technology 21 {9): 1892-1904. Rogge, W. F., L. M. Hildemann, M. A. Mazurek, G. R. Cass and B. R. T. Simoneit (1993). "Sources of Fine Organic Aerosol .4. Particulate Abrasion Products 216 from Leaf Surfaces of Urban Plants." Environmental Science & Technology 27(13): 2700-2711. Rogge, W. F., L. M. Hildemann, M. A. Mazurek, G. R. Cass and B. R. T. Simoneit (1993). "Sources of Fine Organic Aerosol .5. Natural-Gas Home Appliances." Environmental Science & Technology 27(13): 2736-2744. Rogge, W. F., L. M. Hildemann, M. A. Mazurek, G. R. Cass and B. R. T. Simoneit (1997). "Sources of fine organic aerosol .8. Boilers burning No. 2 distillate fuel oil." Environmental Science & Technology 31(10): 2731-2737. Rogge, W. F., L. M. Hildemann, M. A. Mazurek, G. R. Cass and B. R. T. Simoneit (1991). "Sources of Fine Organic Aerosol .1. Charbroilers and Meat Cooking Operations." Environmental Science & Technology 25(6): 1112-1125. Rogge, W. F., M. A. Mazurek, L. M. Hildemann, G. R. Cass and B. R. T. Simoneit (1993). "Quantification of Urban Organic Aerosols at a Molecular-Level - Identification, Abundance and Seasonal-Variation." Atmospheric Environment Part a-General Topics 27(8): 1309-1330. Rogge, W. F., P. M. Medeiros and B. R. T. Simoneit (2006). "Organic marker compounds for surface soil and fugitive dust from open lot dairies and cattle feedlots." Atmospheric Environment 40(1): 27-49. Rogge, W. F., P. M. Medeiros and B. R. T. Simoneit (2007). "Organic marker compounds in surface soils of crop fields from the San Joaquin Valley fugitive dust characterization study." Atmospheric Environment 41(37): 8183-8204. 217 Ruijter, G. J. G., M. Bax, H. Patel, S. J. Flitter, P. J. I. van de Vondervoort, R. P. de Vries, P. A. vanKuyk and J. Visser (2003). "Mannitol is required for stress tolerance in Aspergillus niger conidiospores." Eukaryotic Cell 2(4): 690-698. Rushdi, A. I., K. Al-Mutlaq and B. R. T. Simoneit (2005). "Sources of organic compounds in fine soil and sand particles during winter in the metropolitan area of Riyadh, Saudi Arabia." Archives of Environmental Contamination and Toxicology 49(4): 457-470. Rushdi, A. I., S. Al-Zarban and B. R. T. Simoneit (2006). "Chemical compositions and sources of organic matter in fine particles of soils and sands from the vicinity of Kuwait city." Environmental Monitoring and Assessment 120(1-3): 537-557. Russell, M. and D. T. Allen (2004). "Seasonal and spatial trends in primary and secondary organic carbon concentrations in southeast Texas." Atmospheric Environment 38(20): 3225-3239. Saxena, P. and L. M. Hildemann (1996). "Water-soluble organics in atmospheric particles: A critical review of the literature and application of thermodynamics to identify candidate compounds." Journal of Atmospheric Chemistry 24(1): 57-109. Schauer, J. J., M. P. Fraser, G. R. Cass and B. R. T. Simoneit (2002). "Source reconciliation of atmospheric gas-phase and particle-phase pollutants during a severe photochemical smog episode." Environmental Science & Technology 36(17): 3806-3814. 218 Schauer, J. J., M. J. Kleeman, G. R. Cass and B. R. T. Simoneit (1999). "Measurement of emissions from air pollution sources. 1. C-l through C-29 organic compounds from meat charbroiling." Environmental Science & Technology 33(10): 1566-1577. Schauer, J. J., M. J. Kleeman, G. R. Cass and B. R. T. Simoneit (1999). "Measurement of emissions from air pollution sources. 2. C-l through C-30 organic compounds from medium duty diesel trucks." Environmental Science & Technology 33(10): 1578-1587. Schauer, J. J., M. J. Kleeman, G. R. Cass and B. R. T. Simoneit (2001). "Measurement of emissions from air pollution sources. 3. C-l-C-29 organic compounds from fireplace combustion of wood." Environmental Science & Technology 35(9): 1716-1728. Schauer, J. J., M. J. Kleeman, G. R. Cass and B. R. T. Simoneit (2002). "Measurement of emissions from air pollution sources. 4. C-l-C-27 organic compounds from cooking with seed oils." Environmental Science & Technology 36(4): 567-575. Schauer, J. J., M. J. Kleeman, G. R. Cass and B. R. T. Simoneit (2002). "Measurement of emissions from air pollution sources. 5. C-l-C-32 organic compounds from gasoline-powered motor vehicles." Environmental Science & Technology 36(6): 1169-1180. 219 Schauer, J. J., W. F. Rogge, L. M. Hildemann, M. A. Mazurek and G. R. Cass (1996). "Source apportionment of airborne particulate matter using organic compounds as tracers." Atmospheric Environment 30(22): 3837-3855. Schichtel, B. A., K. A. Gebhart, M. G. Barna and W. C. Malm (2006). "Association of airmass transport patterns and particulate sulfur concentrations at Big Bend National Park, Texas." Atmospheric Environment 40(5): 992-1006. Schobert, B. (1977). "IS THERE AN OSMOTIC REGULATORY MECHANISM IN ALGAE AND HIGHER-PLANTS." Journal of Theoretical Biology 68(1): 17-26. Schumacher, M. J. (2008, 7/2008). "Allergy and Asthma in the Southwestern United States." 2009, from http://allergy.peds.arizona.edu/southwest/prevalence.html. Schwartz, J. (1994). "Air-Pollution and Daily Mortality - a Review and Meta Analysis." Environmental Research 64(1): 36-52. Schwartz, J., D. W. Dockery and L. M. Neas (1996). "Is daily mortality associated specifically with fine particles?" Journal of the Air & Waste Management Association 46(10): 927-939. Seinfeld, J. H. and S. N. Pandis (1998). Atmospheric Chemistry and Physics: From Air Pollution to Climate Change, Wiley, New York. Shafizadeh, F., R. H. Furneaux, T. G. Cochran, J. P. Scholl and Y. Sakai (1979). "Production of Levoglucosan and Glucose from Pyrolysis of Cellulosic Materials." Journal of Applied Polymer Science 23(12): 3525-3539. 220 Shrivastava, M. K., R. Subramanian, W. F. Rogge and A. L. Robinson (2007). "Sources of organic aerosol: Positive matrix factorization of molecular marker data and comparison of results from different source apportionment models." Atmospheric Environment 41(40): 9353-9369. Sillje, H. H. W., J. W. G. Paalman, E. G. ter Schure, S. Q. B. Olsthoorn, A. J. Verkleij, J. Boonstra and C. T. Verrips (1999). "Function of trehalose and glycogen in cell cycle progression and cell viability in Saccharomyces cerevisiae." Journal of Bacteriology 181(2): 396-400. Simoneit, B. R. T. (1984). "Organic-Matter of the Troposphere .3. Characterization and Sources of Petroleum and Pyrogenic Residues in Aerosols over the Western United-States." Atmospheric Environment 18(1): 51-67. Simoneit, B. R. T. (1985). "Application of Molecular Marker Analysis to Vehicular Exhaust for Source Reconciliations." International Journal of Environmental Analytical Chemistry 22(3-4): 203-233. Simoneit, B. R. T. (1989). "Organic-Matter of the Troposphere .5. Application of Molecular Marker Analysis to Biogenic Emissions into the Troposphere for Source Reconciliations." Journal of Atmospheric Chemistry 8(3): 251-275. Simoneit, B. R. T. (1999). "A review of biomarker compounds as source indicators and tracers for air pollution." Environmental Science and Pollution Research 6(3): 159-169. Simoneit, B. R. T. (2002). "Biomass burning - A review of organic tracers for smoke from incomplete combustion." Applied Geochemistry 17(3): 129-162. 221 Simoneit, B. R. T. and V. O. Elias (2000). "Organic tracers from biomass burning in atmospheric particulate matter over the ocean." Marine Chemistry 69(3-4): 301-312. Simoneit, B. R. T., V. O. Elias, M. Kobayashi, K. Kawamura, A. I. Rushdi, P. M. Medeiros, W. F. Rogge and B. M. Didyk (2004). "Sugars - Dominant water-soluble organic compounds in soils and characterization as tracers in atmospheric particulate matter." Environmental Science & Technology 38(22): 5939-5949. Simoneit, B. R. T., M. Kobayashi, M. Mochida, K. Kawamura and B. J. Huebert (2004). "Aerosol particles collected on aircraft flights over the northwestern Pacific region during the ACE-Asia campaign: Composition and major sources of the organic compounds." Journal of Geophysical Research-Atmospheres 109(D19). Simoneit, B. R. T., M. Kobayashi, M. Mochida, K. Kawamura, M. Lee, H. J. Lim, B. J. Turpin and Y. Komazaki (2004). "Composition and major sources of organic compounds of aerosol particulate matter sampled during the ACE-Asia campaign." Journal of Geophysical Research-Atmospheres 109(D19). Simoneit, B. R. T. and M. A. Mazurek (1981). "Air-Pollution - the Organic-Components." Crc Critical Reviews in Environmental Control 11(3): 219-276. 222 Simoneit, B. R. T. and M. A. Mazurek (1982). "Organic-Matter of the Troposphere .2. Natural Background of Biogenic Lipid Matter in Aerosols over the Rural Western United-States." Atmospheric Environment 16(9): 2139-2159. Simoneit, B. R. T., J. J. Schauer, C. G. Nolte, D. R. Oros, V. O. Elias, M. P. Fraser, W. F. Rogge and G. R. Cass (1999). "Levoglucosan, a tracer for cellulose in biomass burning and atmospheric particles." Atmospheric Environment 33(2): 173-182. Song, X. H., A. V. Polissar and P. K. Hopke (2001). "Sources of fine particle composition in the northeastern US." Atmospheric Environment 35(31): 5277-5286. Spielvogel, S., J. Prietzel and I. Kogel-Knabner (2008). "Soil organic matter stabilization in acidic forest soils is preferential and soil type-specific." European Journal of Soil Science 59(4): 674-692. Standley, L. J. and B. R. T. Simoneit (1987). "Characterization of Extractable Plant Wax, Resin, and Thermally Matured Components in Smoke Particles from Prescribed Burns." Environmental Science & Technology 21(2): 163-169. Steerenberg, P. A., C. E. T. Withagen, J. Dormans, W. J. van Dalen, H. van Loveren and F. R. Casee (2003). "Adjuvant activity of various diesel exhaust and ambient particles in two allergic models." Journal of Toxicology and Environmental Health-Part A 66(15): 1421-1439. 223 Stoop, J. M. H., J. D. Williamson and D. M. Pharr (1996). "Mannitol metabolism in plants: A method for coping with stress." Trends in Plant Science 1(5): 139-144. Sweeley, C. C, R. Bentley, M. Makita and W. W. Wells (1963). "Gas-Liquid Chromatography of Trimethylsilyl Derivatives of Sugars and Related Substances." Journal of the American Chemical Society 85(16): 2497-&. Taylor, P. E., R. C. Flagan, A. G. Miguel, R. Valenta and M. M. Glovsky (2004). "Birch pollen rupture and the release of aerosols of respirable allergens." Clinical and Experimental Allergy 34(10): 1591-1596. Taylor, P. E., R. C. Flagan, R. Valenta and M. M. Glovsky (2002). "Release of allergens as respirable aerosols: A link between grass pollen and asthma." Journal of Allergy and Clinical Immunology 109(1): 51-56. Tereshina, V. M., A. S. Memorskaya, E. V. Morozova, V. P. Kozlov and E. P. Feofilova (2000). "Alterations in the carbohydrate composition of the cytosol of fungal spores caused by temperature variations and the storage process." Microbiology 69(4): 422-427. Turpin, B. J. and J. J. Huntzicker (1995). "Identification of Secondary Organic Aerosol Episodes and Quantitation of Primary and Secondary Organic Aerosol Concentrations During Scaqs." Atmospheric Environment 29(23): 3527-3544. Turpin, B. J., J. J. Huntzicker, S. M. Larson and G. R. Cass (1991). "LOS-ANGELES SUMMER MIDDAY PARTICULATE CARBON - PRIMARY AND 224 SECONDARY AEROSOL." Environmental Science & Technology 25(10): 1788-1793. U.S.EPA (2004). Air Quality Criteria for Particulate Matter. II. U.S.EPA (2005). Review of the National Ambient Air Quality Standards for Particulate Matter: Policy Assessment of Scientific and Technical Information; OAQPS Staff Paper. U.S.EPA (2006). "US EPA Green Book - Federal Register Notice." 2009, from http://www.epa.gOv/air/oaqps/greenbk/7161143.html. van Doom, W. G. (2004). "Is petal senescence due to sugar starvation?" Plant Physiology 134(1): 35-42. Venkataraman, S. and M. B. Low (1994). "The Effects of Liabilities of Age and Size on Autonomous Sub-Units of Established Firms in the Steel Distribution Industry." Journal of Business Venturing 9(3): 189-204. Wallaart, R. A. M. (1980). "Chemotaxonomy of Rosaceae .1. Distribution of Sorbitol in Rosaceae." Phytochemistry 19(12): 2603-2610. Wan, E. C. H. and J. Z. Yu (2006). "Determination of sugar compounds in atmospheric aerosols by liquid chromatography combined with positive electrospray ionization mass spectrometry." Journal of Chromatography A 1107(1-2): 175-181. Wan, E. C. H. and J. Z. Yu (2007). "Analysis of sugars and sugar polyols in atmospheric aerosols by chloride attachment in liquid 225 chromatography/negative ion electrospray mass spectrometry." Environmental Science & Technology 41(7): 2459-2466. Wang, G., K. Kawamura, M. Xie, S. Hu, S. Gao, J. Cao, Z. An and Z. Wang (2009). "Size-distributions of n-hydrocarbons, PAHs and hopanes and their sources in the urban, mountain and marine atmospheres over East Asia." Atmos. Chem. Phys. Discuss. 9(3): 13859-13888. Wang, X. K., A. Emelyanov, I. Sleptsova-Friedrich, V. Korzh and Z. Y. Gong (2001). "Expression of two novel zebrafish iroquois homologues (zirol and ziro5) during early development of axial structures and central nervous system." Mechanisms of Development 105(1-2): 191-195. Watson, J. G., Fujita, E.M, Chow, J.C., Zielinska, B., Richards, L.W., Neff, W.D., Dietrich, D. (1998). Northern Front Range Air Quality Study. Desert Research Institute Weather Underground (2007). "Big Bend National Park History Weather Data." 2009, from http://www.wunderground.eom/history/airport/K6R6/2007/5/l/CustomHistory .html?dayend=30&monthend=6&yearend=2007&req_city=NA&req_state=N A&req_statename=NA. Wei, Y. and M. Y. Ding (2000). "Analysis of carbohydrates in drinks by high-performance liquid chromatography with a dynamically modified amino column and evaporative light scattering detection." Journal of Chromatography A 904(1): 113-117. 226 Weitkamp, E. A., A. T. Lambe, N. M. Donahue and A. L. Robinson (2008). "Laboratory Measurements of the Heterogeneous Oxidation of Condensed-Phase Organic Molecular Makers for Motor Vehicle Exhaust." Environmental Science & Technology 42(21): 7950-7956. Wiemken, A. (1990). "Trehalose in Yeast, Stress Protectant Rather Than Reserve Carbohydrate." Antonie Van Leeuwenhoek International Journal of General and Molecular Microbiology 58(3): 209-217. Wilson, R. and J. Spengler (1996). Particles in Our Air: Concentrations and Health Effects. Cambridge, MA, Harvard University Press. Winiwarter, W., H. Bauer, A. Caseiro and H. Puxbaum (2009). "Quantifying emissions of primary biological aerosol particle mass in Europe." Atmospheric Environment 43(7): 1403-1409. Wong, B., J. S. Murray, M. Castellanos and K. D. Croen (1993). "D-Arabitol Metabolism in Candida-Albicans - Studies of the Biosynthetic-Pathway and the Gene That Encodes Nad-Dependent D-Arabitol Dehydrogenase." Journal of Bacteriology 175(19): 6314-6320. Woodruff, T. J., J. Grillo and K. C. Schoendorf (1997). "The relationship between selected causes of postneonatal infant mortality and particulate air pollution in the United States." Environmental Health Perspectives 105(6): 608-612. Yttri, K. E., C. Dye and G. Kiss (2007). "Ambient aerosol concentrations of sugars and sugar-alcohols at four different sites in Norway." Atmospheric Chemistry and Physics 7(16): 4267-4279. 227 Yu, J. Z., R. J. Griffin, D. R. Cocker, R. C. Flagan, J. H. Seinfeld and P. Blanchard (1999). "Observation of gaseous and particulate products of monoterpene oxidation in forest atmospheres." Geophysical Research Letters 26(8): 1145-1148. Yue, Z. W. and M. P. Fraser (2004). "Characterization of nonpolar organic fine particulate matter in Houston, Texas." Aerosol Science and Technology 38: 60-67. Zdrahal, Z., J. Oliveira, R. Vermeylen, M. Claeys and W. Maenhaut (2002). "Improved method for quantifying levoglucosan and related monosaccharide anhydrides in atmospheric aerosols and application to samples from urban and tropical locations." Environmental Science & Technology 36(4): 747-753. Zheng, M., G. R. Cass, J. J. Schauer and E. S. Edgerton (2002). "Source apportionment of PM2.5 in the southeastern United States using solvent-extractable organic compounds as tracers." Environmental Science & Technology 36(11): 2361-2371. Zheng, M., M. Fang, F. Wang and K. L. To (2000). "Characterization of the solvent extractable organic compounds in PM2.5 aerosols in Hong Kong." Atmospheric Environment 34(17): 2691-2702. Zielinska, B., J. Sagebiel, J. D. McDonald, K. Whitney and D. R. Lawson (2004). "Emission Rates and Comparative Chemical Composition from Selected In-Use Diesel and Gasoline-Fueled Vehicles." Journal of the Air & Waste Management Association (1995) 54(9): 1138-1150.