Sources of Volatile Organic Compounds in Industrial, Coastal and Urban Regions

University of Calgary PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2018-12-20 Sources of Volatile Organic Compounds in Industrial, Coastal and Urban Regions

Tokarek, Travis Wade

Tokarek, T. W. (2018). Sources of Volatile Organic Compounds in Industrial, Coastal and Urban Regions (Unpublished doctoral thesis). University of Calgary, Calgary, AB. http://hdl.handle.net/1880/109409 doctoral thesis

University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca UNIVERSITY OF CALGARY

Sources of Volatile Organic Compounds in Industrial, Coastal and Urban Regions

Travis Wade Tokarek

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

GRADUATE PROGRAM IN CHEMISTRY

CALGARY, ALBERTA

DECEMBER, 2018

This thesis describes the application of gas chromatography using direct air injection for the measurement and analysis of volatile organic compounds (VOCs) during three campaigns in unique environments (i.e., industrial, marine and urban). A Griffin 450 gas chromatograph equipped with a cylindrical ion trap mass spectrometer and electron impact ionization (GC-

ITMS) and a Varian 3380CP equipped with an electron capture detector (GC-ECD) were used to acquire speciated measurements of select VOCs (monoterpenes, alkanes and aromatics) and peroxyacyl nitrates (PANs), respectively, which were analyzed to investigate air mass sources. In the first campaign, principal component analysis (PCA) was used on a dataset collected in the

Alberta oil sands to elucidate possible sources of analytically unresolved intermediate volatility organic compounds (IVOCs) that were observed in the GC-ITMS chromatograms. A spectrally similar analytically unresolved peak of IVOCs was observed in the lab from vapours in the headspace of a bitumen sample collected near the measurement site.

In the second campaign, previously-reported nocturnal ozone-depletion events were investigated off the West coast of Vancouver Island. Monoterpenes and their oxidation products were measured to explore the role of biogenic VOCs (BVOC) as a possible chemical loss pathway for ozone in this region. The analysis showed that monoterpenes play a minor role in ozone depletion in this environment and that dry deposition is likely the dominant pathway. During this campaign, the headspace vapours of several local kelp species were measured to probe possible

BVOC sources in the region. Limonene was found to be enhanced above background concentrations by two species (Nereocystis luetkeana and Alaria marginata) making them a previously unrecognized source of a highly reactive monoterpene in this environment. ii

In the third campaign, two PAN species (i.e., peroxyacetyl nitrate (PAN) and peroxypropionic nitrate (PPN)) were measured by GC-ECD during a period when wildfire smoke was transported from California and British Columbia to Calgary, Alberta. The PPN/PAN ratio was calculated and ranged from 0.05 to 0.17 (a typical background value is 0.10) in biomass burning plumes.

Gas chromatography with direct air injection continues to yield new and useful information and should be a component of any comprehensive analysis.

Keywords: VOC emissions, principal component analysis, bitumen, seaweed, kelp,

monoterpenes, limonene, hydrocarbons, ozone depletion, monoterpene emissions

iii

Acknowledgements

Above all I would like to sincerely thank my supervisor Dr. Hans D. Osthoff. The quality and quantity of education I received under your guidance far surpassed what I once believed myself capable of. You have left me equipped with a profound confidence, enthusiasm and curiosity. Thank you for knowing how far to push me without breaking my spirit.

I would also like to thank my supervisory committee Dr. Kevin Thurbide and Dr. Yujun

Shi who helped guide me through the staggered path of my degree. I really value the time and commitment you both put into reading my proposal and thesis and for offering me feedback. I would also like to thank my defense committee who read through these pages. I am sure it wasn’t easy and am grateful for your time and efforts.

Thank you to all fellow Osthoff group members, past and present. A special thank you to

Charles, Connie, Youssef, Duncan, Natasha, Nick J. and Nick G. I can’t wait to see where we all are in 10 years! I would like to also thank my friends and family who understand the social sacrifices I have made to complete my degree. I manage stress well, but a big part of that is having a strong support group of people that I trust and respect.

Finally, thank you Erdem. I think at one time I was concerned that a relationship would make graduate school more difficult, but it had the opposite effect. I have survived and thrived because of you. Thank you for your continued love and support. I am very excited about starting the next Chapter of our lives together. With you, I am confident it will be amazing! iv

Dedication

To my partner Erdem, seni ҫok seviyorum cep ayısı!

Table of Contents

Abstract ...... ii Acknowledgements ...... iv Dedication ...... v Table of Contents ...... v List of Tables ...... ix List of Figures and Illustrations ...... x List of Symbols, Abbreviations and Nomenclature ...... xiv

Chapter One: Introduction ...... 1 1.1 The troposphere and volatile organic compounds ...... 1 1.2 Volatile organic compounds: sources and chemistry ...... 1 1.2.1 Anthropogenic volatile organic compounds: alkanes and aromatics ...... 2 1.2.2 Biogenic volatile organic compounds: terpenes ...... 3 1.3 Volatile organic compound oxidation ...... 5 1.4 Products of volatile organic compound oxidation ...... 7 1.4.1 Ozone production and the peroxy radical ...... 7 1.4.2 Oxygenated volatile organic compounds ...... 9 1.4.3 Formation of peroxyacyl nitrates ...... 12 1.5 Quantification of volatile organic compounds ...... 13 1.5.1 Gas chromatography mass spectrometry ...... 15 1.5.1.1 Sample preconcentration ...... 16 1.6 Thesis motivation and objectives ...... 17

Chapter Two: Instrumentation and Calibration ...... 19 2.1 Gas chromatography ion trap mass spectrometer ...... 19 2.1.1 Electron impact ionization ...... 21 2.1.2 Cylindrical ion trap ...... 21 2.1.3 Injection methodology ...... 22 2.1.4 Low mass, high mass and detector tuning ...... 23 2.2 Peroxyacyl nitrate gas chromatograph specifications ...... 24 2.2.1 Inlet design and schematic ...... 25

Chapter Three: Data Collected at a Ground Site Near Fort McKay During FOSSILS 2013...... 28 3.1 Field campaign details ...... 28 3.2 Instrumentation ...... 30 3.3 Analytically unresolved hydrocarbon signature ...... 32 3.4 Field data summary ...... 34 3.4.1 Qualitative assessment of VOCs ...... 41 3.5 Conclusions ...... 45

Chapter Four: Principal Component Analysis of Intermediate Volatility Organic Compounds Observed During the Summer in the Athabasca Oil Sands Region of Alberta, Canada During FOSSILS 2013 ...... 46 vi

4.1 Introduction ...... 46 4.2 Principal component analysis ...... 47 4.2.1 Selection of variables ...... 48 4.3 PCA analysis with primary variables ...... 49 4.3.1 Extended PCA analysis with added secondary variables ...... 55 4.4 Principal component analysis with a divided IVOC peak ...... 57 4.5 Spatial distribution of IVOC sources ...... 59 4.6 Discussion ...... 61 4.7 Sources associated with IVOCs ...... 62 4.7.1 Component 1: tailings ponds (wet tailings) ...... 62 4.7.2 Component 2: mine fleet and vehicle emissions ...... 65 4.7.3 Component 5: surface-exposed bitumen and hot-water bitumen extraction ...... 68 4.8 Sources not associated with IVOCs ...... 71 4.8.1 Component 3: biogenic emissions and respiration ...... 71 4.8.2 Component 4: Upgrader emissions ...... 73 4.9 Extended PCA with added secondary variables ...... 76 4.10 Conclusions ...... 77

Chapter Five: Bitumen characteristics observed in the lab and field ...... 80 5.1 Introduction ...... 80 5.2 Bitumen characteristics by electron impact fragmentation ...... 81 5.3 Conclusions ...... 86

Chapter Six: Exploring Nocturnal Ozone Depletion Events in the Marine Boundary Layer during ORCA 2015 ...... 87 6.1 Introduction ...... 87 6.2 Methods ...... 90 6.2.1 Study overview ...... 90 6.2.2 Gas phase measurements ...... 92 6.2.3 Particle phase measurements ...... 95 6.2.4 Photolysis frequencies ...... 95 6.2.5 Back trajectory calculations ...... 96 6.3 Results and analysis ...... 96 6.3.1 Identification of ozone depletion events ...... 96 6.3.2 Correlations of ozone depletion events with chemical tracers ...... 98 6.3.3 Role of meteorology ...... 104 6.3.4 Correlations of ODEs with changes in aerosol size distribution ...... 106 6.3.5 Monoterpene ratios: insights into BVOC sources ...... 107 6.3.6 Kelp forests: an unrecognized source of BVOCs ...... 108 6.4 Summary and conclusions...... 109

Chapter Seven: Kelp Headspace Analysis during ORCA 2015...... 111 7.1 Introduction ...... 111 7.2 Methods ...... 113

vii

7.3 Results ...... 117 7.3.1 Overview ...... 117 7.3.2 Identification and quantification of BVOCs ...... 121 7.3.3 Monoterpenes ...... 124 7.3.4 n-Alkanes ...... 125 7.3.5 Oxygenated volatile organic compounds ...... 125 7.3.6 Aromatic hydrocarbons ...... 126 7.4 Discussion ...... 128 7.4.1 Monoterpene emissions from kelp ...... 128 7.4.2 n-Alkanes ...... 130 7.4.3 Oxygenates ...... 131 7.4.4 Toluene ...... 132 7.4.5 Unidentified hydrocarbons ...... 132 7.5 Atmospheric implications and future work...... 133

Chapter Eight: Peroxyacyl Nitrate Chemistry during PANfire 2017 ...... 136 8.1 Summary ...... 136 8.2 Introduction to peroxyacyl nitrates ...... 136 8.2.1 Peroxyacyl nitrates in biomass burning plumes ...... 137 8.2.2 Instrumentation and consumables for measuring peroxyacyl nitrates ..138 8.3 Materials and Methods ...... 138 8.3.1 Data Reduction ...... 141 8.3.2 PAN and PPN standards ...... 142 8.4 Auxiliary measurements...... 143 8.5 Results ...... 144 8.5.1 Trailer PAN Measurements ...... 144 8.5.2 Forest Fire impact ...... 147 8.5.3 Rooftop PAN measurements ...... 150 8.6 Discussion ...... 152 8.6.1 Ambient concentrations and PAN diurnal profiles ...... 152 8.6.2 Air impacted by biomass burning ...... 154

Chapter Nine: Thesis Summary, Conclusions and Future Work ...... 159

viii

List of Tables

Table 2.1 Method details for the GC-ITMS during field campaigns...... 20

Table 3.1 Instruments used to measure ambient gas-phase and aerosol species during the 2013 JOSM intensive study at AMS 13...... 31

Table 3.2 Variables observed at the AMS 13 ground site during the 2013 JOSM campaign...... 36

Table 3.3 Threshold criteria and associated adjunct chemical markers for a qualitative assessment of potential oil sands sources...... 42

Table 4.1 Loadings for the 10-factor, optimal solution (primary variables only). Coefficients with Pearson correlation coefficients r>0.3 are shown in bold font...... 52

Table 4.2 Hypothesized identifications of principal components...... 53

Table 4.3 Loadings for the 11-component solution with the inclusion of variables associated with secondary processes...... 56

Table 4.4 Loadings for the 10-factor, optimal solution (primary variables only) with the IVOCs divided into 2 bins. Coefficients with Pearson correlation coefficients r>0.3 are shown in bold font...... 58

Table 5.1 Time range that a particular ion fragment was observed in sampled vapours of bitumen and an ambient air sample. For bitumen, the cut-off threshold was 1,000,000 counts. For ambient air, the cut-off threshold was 10,000 counts ...... 84

Table 6.1 Summary of ODEs observed...... 98

Table 7.2 Enhancement of VOC mixing ratios in the sampled head spaces. Ambient mixing ratios observed immediately prior to head space analysis (Table S-2) were subtracted prior to presentation. Enhancements of >10 pptv are shown in bold font for clarity. The "error" is measurement precision (±2) of the mixing ratios of four consecutive ambient air chromatograms taken prior to the head space analysis. n/d = not detected. Enhancement deemed significant (>2) are bolded...... 122

Table 8.1 Summary of response factors from calibrations during PANfire 2017 in mV s ppbv-1...... 143

List of Figures and Illustrations

Figure 1.1 A sample of some common terpenes. A) Isoprene is the smallest terpene. B) Monoterpenes are made up of two isoprene subunits, C) sesquiterpenes are made up of 3 isoprene subunits, etc...... 5

Figure 1.2 A summary of VOC oxidation adapted from Koppmann (2007)...... 6

Figure 1.3 A small sample of oxygenated volatile organic compounds (OVOCs) that are commonly observed in the atmosphere. These can be directly emitted of formed through chemical transformations in the atmosphere...... 10

Figure 1.4 ß-pinene oxidation via ozonolysis and HO addition...... 11

Figure 1.5 Schematic showing the oxidation of VOCs to form PANs taken from Roberts et al. (2007)...... 13

Figure 2.1 Schematic of the valve system used by the Griffin 450 GC-ITMS. A) Adsorption mode preconcentrates trace gases in ambient air at 40 °C. When the valve is actuated, B) desorption mode flash heats the preconcentrator to inject the previously adsorbed trace gases onto the analytical column...... 23

Figure 2.2 α-pinene concentration is unaffected by ozone addition when stainless-steel is heated to 125 °C. Blue regions are used to demarcate experimental changes (i.e., bypassing through a Teflon line and adding ozone). The ozone was present at ~ 100 ppbv from 2:00 pm to 5:00 pm...... 26

Figure 2.3 Ozone concentration as a function of temperature. Ozone is completely scrubbed at 110 °C...... 27

Figure 3.1 Map of oil sands facilities showing locations of surface mines and tailings ponds, downloaded from the Oil Sands Information Portal (Alberta, 2017). The red star indicates the location of AMS 13...... 30

Figure 3.2 (Top) Total ion chromatograms of air samples collected on August 27, 2013 from 18:04 to 18:14 UTC (red) and on August 28, 2013 from 13:43 to 13:53 UTC (blue). The TIC of a head space sample of ground-up bitumen collected post-campaign is superimposed (black). The gray area indicates the range over which IVOC signal was integrated. (Bottom) Retention times of n-alkanes, determined after the field campaign by sampling a VOC mixture containing a C10 – C16 n-alkane ladder...... 32

Figure 3.3 Time series of selected pollution tracers observed at the AMS 13 ground site in the Athabasca oil sands during the 2013 JOSM measurement intensive. The gray and yellow backgrounds represent night and day, respectively. (A) Selected non-methane hydrocarbons (NMHCs) and IVOCs. (B) Combustion product tracers: refractory black carbon (rBC), total odd nitrogen (NOy) and particle surface bound polycyclic aromatic hydrocarbons (pPAH), and organic aerosol components: hydrocarbon-like organic aerosol (HOA) and less oxidized oxygenated organic aerosol (LO-OOA). (C) Methane (CH4), carbon dioxide (CO2) and monoxide (CO). (D) Total sulfur (TS), sulfur dioxide (SO2), and total reduced sulfur (TRS) and PM10 particle volume. (E) Biogenic VOCs (α- pinene, ß-pinene and limonene) and ammonia (NH3)...... 35

Figure 3.4 Time series of a qualitative analysis of potential sources of VOCs. Each bar represents a region of time where a particular source may have been sampled based on chemical markers...... 43

Figure 4.1 Images of likely sources associated with each of the principal components. From top left to bottom: (A) Wet tailings ponds (component 1). (B) Mine truck fleet and highway traffic emissions (component 2). (C) Biogenic emissions from vegetation (component 3). (D) Upgrader facilities (component 4). (E) Exposed bitumen on mined surfaces (component 5). (F) Fugitive greenhouse gas emissions from mine faces (component 6). (G) Wind-blown dust from exposed sand (component 7). (H) Fugitive emissions of ammonia from storage tanks (Component 8). (I) Composite (dry) tailings (component 10). No image is shown for production CO from oxidation of VOCs (component 9)...... 54

Figure 4.2 Bivariate polar plots related to IVOCs: (A) IVOCs from the complete data set. (B) Component 5 extracted from the main PCA (Table 4.1). (C) Component 1 extracted from the main PCA. (D) Component 2 extracted from the main PCA analysis. Wind direction is binned into 10° intervals and wind direction into 30° intervals. The polar axis indicates wind speed (m s-1). a.u. = arbitrary units...... 60

Figure 5.1 Side by side comparison of an ambient air GC-ITMS data matrix (left) and a bitumen sampled in the lab (right). Darker pixels represent a greater intensity at a given m/z and time...... 85

Figure 6.1 Map of the study region...... 91

Figure 6.2 Time series of O3 (left hand side) and CO2 mixing ratios (A), monoterpene and nopinone mixing ratios (B), and Ox (=O3+NO2), NO2, NOy mixing ratios (C). The blue lines indicate O3 depletion events identified using the criteria by McKendry et al. (2014). The background colouring indicates the NO2 photolysis frequency. The dates and times are in coordinated universal time (UTC)...... 97

are the median values. Percentiles were calculated over ½ hour bins. The median non- ODE O3 time series is shown as a blue line. (B) O3 and terpene mixing ratios as functions of (solar) time of day during ODEs calculated over 1-hour bins...... 100

Figure 6.4 (A) Median mixing ratios of monoterpenes as functions of time of day. (B) Nocturnal nopinone concentrations during ODE and non-ODE events as functions of time of day...... 102

Figure 6.5 Correlations of O3 mixing ratios with (A) wind direction and speed, (B) CO2 mixing ratios, colour-coded by the sum of α- and β-pinene and limonene mixing ratios, and (C) NO2 mixing ratios, colour-coded by CO2 mixing ratios. (D) Dependence of the limonene to (α- plus β-pinene) ratio on wind direction and speed, plotted as their N-S and E-W components...... 103

Figure 6.6 Time series of submicron aerosol size distributions...... 107

Figure 7.1 Setup for head space sampling (not to scale). GC-ITMS = gas chromatograph ion trap mass spectrometer. MFC = mass flow controller. s.s. = stainless steel...... 117

Figure 7.2 Total ion chromatograms (left axis, blue and black) and selective ion chromatograms of m/z 93±1 which is a fragment associated with monoterpenes (right axis, red and green) of head spaces above kelp species (A) Nereocystis luetkeana (bull kelp).and (B) Alaria marginata (winged kelp) The chromatograms obtained for a sample of sea water are superimposed...... 119

Figure 7.3 Total ion chromatograms (left axis, blue and black) and residual ion chromatograms of m/z 931 (right axis, red and green) of head spaces above seaweed species (A) Fucus gardneri (rock weed), (B) Ulva spp. (sea lettuce), and (C) Callophyllis spp. (red sea fans). The chromatograms obtained for a sample of sea water are superimposed...... 120

Figure 7.4 Comparison of limonene mass spectra from different sources. (A) The headspace of bull kelp, (B) the VOC calibration cylinder used in the field...... 123

Figure 8.1 An example Gaussian peak fit with parameters for a sample taken during the field campaign (September 2, 2017 at 18:00:00 UTC)...... 142

Figure 8.2 Time series of A) PPN/PAN and NOx, B) Total hydrocarbons (THC), PM2.5 and CO, and C) PAN, PPN and O3. PM2.5, PANs, CO and O3 were enhanced from August 30 – September 10, when the site was heavily impacted by biomass burning plumes. The yellow and grey background represent day and night respectively, as determined by the solar elevation angle in Calgary...... 145

Figure 8.3 (A) PAN and (B) PPN diurnal profiles during the forest fire period. The blue squares represent individual measurements with the percentiles and median layered on top...... 146

xii

Figure 8.4 Time of day plot showing a diurnal profile of PAN during the time in the trailer when no forest fires were present. The blue squares represent individual measurements with the percentiles and median layered on top...... 147

Figure 8.5 Time series of (Top) PPN/PAN and NOx, (Middle) total hydrocarbons (THC), PM2.5, and CO and (Bottom) PAN, PPN and O3. The yellow and grey background represent day and night respectively, as determined by the solar elevation angle in Calgary...... 148

Figure 8.6 PPN mixing ratio plotted against PAN mixing ratio during the fire period. The colour scale represents the Julian day the sample was collected on. This allows for visualization of different PPN/PAN ratios on different days...... 149

Figure 8.7 PPN mixing ratio plotted against PAN mixing ratio during the fire period. The colour scale represents the concentration of PM2.5. The center of the color scale (white) -3 represents the median PM2.5 concentration of 22.5 µg m . The red fit represents the correlation of all points less than the median and the blue fit represents the correlation of all points greater than the median...... 150

Figure 8.8 Time series of PAN mixing ratios while making measurements in the rooftop lab. The yellow and grey background represent day and night, respectively...... 151

Figure 8.9 The diurnal profile of PAN mixing ratios while in the rooftop lab. The blue dots represent individual PAN measurements while the median and percentiles are also shown, layered overtop...... 152

xiii

List of Symbols, Abbreviations and Nomenclature

Symbol Definition

AGL Above ground level

AMS Air monitoring station

AN Alkyl nitrates

APO Amphitrite Point Observatory

APS Aerodynamic particle sizer

AVOC Anthropogenic volatile organic compound

BLH Boundary layer height

BTEX Benzene, toluene, ethyl benzene, xylenes

BVOC Biogenic volatile organic compound

C* Saturation vapour concentration

CIMS Chemical ionization mass spectrometry

CIT Cylindrical ion trap

CRAZ Calgary Regional Airshed Zone

CRDS Cavity ring-down spectroscopy

DMS Dimethyl Sulfide

ECD Electron capture detector eV Electron volts

ESRL Earth system research laboratory

FID Flame ionization detector

FOSSILS Fort McMurray Oil Sands Strategic Investigation of Local Sources xiv

FTIR Fourier Transform Infrared Spectrometer

GC Gas chromatography

HOA Hydrocarbon organic aerosol

HPLC High performance liquid chromatography

HYSPLIT Hybrid single particle lagrangian integrated trajectory model

ITMS Ion trap mass spectrometry

IUPAC International union of pure and applied chemistry

IVOC Intermediate volatility organic compound

JOSM Joint Oil Sands Monitoring

LIDAR Light detection and ranging

LIF Laser-induced fluorescence

LO-OOA Less-oxidized oxygenated organic aerosol

LOD Limit of detection

MBL Marine boundary layer

MFC Mass flow controller

MO-OOA More-oxidized oxygenated organic aerosol

MPAN Peroxymethacrylic nitric anhydride

MS Mass spectrometry

NAM North American Meso

NIST National Institute of Standards

NMHC Non-methane hydrocarbon

NOAA National Oceanic and Atmospheric Administration

NRC National Research Council of Canada xv

ODE Ozone depletion event

OOA Oxygenated organic aerosol

ORCA Ozone-depleting Reactions in a Coastal Atmosphere

OVOC Oxidized volatile organic compound

PAH Polycyclic aromatic hydrocarbons

PANs Peroxycarboxylic nitric anhydrides

PAN Peroxyacetic nitric anhydride

PANFire PAN chemistry in forest fires

PCA Principal component analysis

PDA Photo diode array

PEEK Polyether ether ketone

PFA Perfluoroalkoxy

PM2.5 Particulate matter less than 2.5 µm in diameter

PM1 Particulate matter less than 1 µm in diameter

PM10-1 Particulate matter between 10 µm and 1 µm in diameter

PMF Positive matrix factorization pPAH particle-surface bound PAH ppbv Parts per billion by volume (10-9) ppmv Parts per million by volume (10-6)

PPN Peroxypropionic nitric anhydride pptv Parts per trillion by volume (10-12)

PTR-MS Proton transfer reaction mass spectrometry rBC Refractory black carbon xvi

RH Relative humidity

RI Retention indices sccm Standard cubic centimeter

SI Supplementary information

SIC Selected ion chromatogram

SIM Selected ion monitoring slpm Standard liter per minute

SML Syncrude – Mildred Lake

SMPS Scanning mobility particle sizer

SOA Secondary organic aerosol

SP-AMS Soot particle aerosol mass spectrometer

SPME Solid-phase microextraction

SVOC Semi-volatile organic compound

TD Thermal dissociation

TIC Total ion chromatogram

TRS Total reduced sulfur

TS Total sulfur

USB Universal serial bus

UTC Coordinated universal time

VOC Volatile organic compound

WBEA Wood Buffalo Environmental Association

ZA Zero air

xvii

Chapter One: Introduction

1.1 The troposphere and volatile organic compounds

The part of the atmosphere in direct contact with the surface of the Earth is the troposphere which extends to a height of approximately 12 km above sea level. It is this part of the atmosphere that interacts most directly with the biosphere, and in recent centuries, the anthrosphere (2002). The troposphere is comprised primarily of nitrogen (78%), oxygen (21%) and argon (1%) with a small percentage of a complex mixture of trace gases that can vary substantially spatially and temporally (Finlayson-Pitts and Pitts, 2000). Plants, animals and humans emit a plethora of chemical species into the troposphere, but no group is quite as complex and diverse as volatile organic compounds (VOCs).

1.2 Volatile organic compounds: sources and chemistry

Volatile organic compounds are defined as hydrocarbons having a vapour pressure greater than 10 Pa at 25 °C, a boiling point < 260 °C, and 15 or less carbon atoms; all remaining compounds (if present in the gas phase) are classified as semi-volatile organic compounds

(SVOCs) (Koppmann, 2007). Robinson et al. (2007) introduced an alternate classification scheme based on saturation concentration (C*) to improve modeling of organic aerosol formation. In this scheme, VOCs are classified as “nonvolatile” (C* < 0.1 μg m–3), “semi- volatile” (SVOC; 0.1 μg m–3 < C* < 103 μg m–3), or “intermediate-volatility” (IVOC; 103 μg m-3

< C* < 106 μg m–3) organic compounds. Recently, Liggio et al. (2016) showed that IVOCs having a C* in the range 105 µg m-3< C* < 107 µg m-3 are particularly important for aerosol

formation and growth. VOCs and other atmospheric gases have many sources which can be primary (emitted directly) or secondary (formed through chemical reactions). These same gases also have many different losses, or sinks, which include chemical and physical processes (e.g., oxidation, deposition, transport, etc.).

Anthropogenic sources of VOCs include fuel production, distribution and combustion, the largest sources being vehicles and biomass burning (Pachauri and Meyer, 2014). However, biogenic sources of VOCs dominate emissions largely due to forest fires and emissions from plant life (EPA, 2014). After methane, the largest emission of naturally occurring VOCs is the terpenes with isoprene alone contributing ~40% of total VOCs emitted (Fan and Zhang, 2004).

1.2.1 Anthropogenic volatile organic compounds: alkanes and aromatics

Anthropogenic VOCs (AVOCs) are a class of compounds commonly found in urban areas. Due to the large variety of chemicals produced around the world, it is advantageous to focus on classes of AVOCs rather than individual species. These classes include alkanes, alkenes, alkynes, oxygenated VOCs (see section 1.4.2) and aromatics including BTEX (benzene, toluene, ethyl benzene and xylenes) compounds. BTEX compounds are an additive used to increase the combustion efficiency of fuel and is added in up to 40% (by weight) (Dinerman et al., 2011). These compounds are present in the gas phase in polluted urban or industrial areas; hence, these compounds are good markers of anthropogenic pollution (Fujita et al., 2014).

Alkanes are a diverse group which include branched and linear alkanes. Complex mixtures of these gases are found in natural gas, gasoline and other petrochemical fuels (Eneh,

2011). The only oxidant that significantly oxidizes alkanes in the atmosphere is the hydroxyl radical (HO). The reactivity and lifetimes of alkanes can differ dramatically with more linear alkanes and those with a greater carbon number being the most reactive with respect to HO

(Finlayson-Pitts and Pitts, 2000; Carter and Atkinson, 1985).

1.2.2 Biogenic volatile organic compounds: terpenes

Terpenes are a class of biogenic volatile organic compounds (BVOCs) which are unsaturated (contain double bonds or rings). Terpenes are produced in plant tissues as signaling compounds for a variety of purposes including the protection of plant tissue from oxidation, insecticides for protection against predators and some studies have even shown that plants can communicate with each other via their terpene emissions (Kessler et al., 2006; Clark et al., 2012

Dendroctonus ponderosae Hopkins; Niinemets, 2010 priming and consequences). Emissions rates depend on temperature, plant species, soil humidity, atmospheric humidity, age, location and light (Saxton et al., 2007).

Terpenes range dramatically in size and reactivity and are comprised of isoprene (C6H8) subunits (Figure 1.1A). They are important atmospheric species because they are emitted in large quantities and are readily oxidized to form Secondary Organic Aerosol (SOA) which can impact visibility, human health and climate. Generally, deciduous trees emit isoprene and are light dependent, whereas conifers store (and subsequently emit) monoterpenes (C10H16, Figure 1.1B)

in plant tissues resulting in their emissions to be more temperature dependent (Niinemets, 2010 priming and consequences). Monoterpenes are particularly important because their volatility keeps them in relatively high concentrations in the gas phase, yet they are sufficiently large that upon oxidation they can contribute significantly to aerosol formation and growth.

Sesquiterpenes (C15H24) are terpenes that contain 3 isoprene subunits (Figure 1.1C).

These VOCs are so large that they are not volatile enough to appreciate in the gas phase in ambient air to the concentrations typically observed by monoterpenes and isoprene (Duhl et al.,

2008). Often, sesquiterpenes are measured as the sum of all sesquiterpenes by soft-ionization techniques like benzene cluster cation chemical ionization mass spectrometry (CIMS) (Lavi et al., 2018). Further, because of their size and degree of unsaturation, they can be very reactive to atmospheric oxidants. However, due to their reactivity and low boiling point, they can readily cause SOA formation and growth in environments where they are abundant like Africa and

South America; in these locations vegetation is abundant and biomass burning is prevalent (Khan et al., 2017).

Figure 1.1 A sample of some common terpenes. A) Isoprene is the smallest terpene. B) Monoterpenes are made up of two isoprene subunits, C) sesquiterpenes are made up of 3 isoprene subunits, etc.

1.3 Volatile organic compound oxidation

The most common oxidants in the atmosphere include HO, ozone (O3), nitrate (NO3) and chlorine radicals (Cl). These species will readily react with VOCs. A summary of VOC oxidation is shown in Figure 1.2. Once oxidized, VOCs can further oxidize into smaller and smaller organic compounds until the end products CO2 and H2O are formed. Volatile organic compounds can also become heavier by reacting with other species to form larger species or via the addition of oxidants to the VOC structure. These oxidized VOCs (OVOCs) can be transported long distances or react further locally. Heavier OVOCs can condense to cause SOA formation and growth. Secondary organic aerosol itself can be transported and affect chemistry, health and climate along the way. The organic species can also be re-emitted from the aerosol phase into the 5

gas phase and is dependent on each species equilibrium constant. Oxidized VOCs can also deposit to surfaces (dry deposition) or deposit to precipitation (wet deposition) (Seinfeld and

Pandis, 2006). Lastly, the process by which plants take in gases through their stomata can impact the loss of OVOCs, this process depends on the amount of foliage present and a metric known as the leaf area index which is the ratio of the leaf area to the area of the ground.

Figure 1.2 A summary of VOC oxidation adapted from Koppmann (2007).

1.4 Products of volatile organic compound oxidation

Once a VOC is oxidized either via hydrogen abstraction or addition to a double bond, a carbon radical will quickly form and react with oxygen to form a peroxy radical. This has important implications in the production of O3 (see section 1.4.1) but is also important because of the tendency to form highly oxidized, higher molecular weight organic species. As VOCs become heavier with the addition of oxygen (primarily), their boiling point increases causing them to condense more readily onto surfaces. In the atmosphere, small particles can be present which offer a high surface area for VOCs to condense onto, causing particle growth, or SOA formation (Ehn et al., 2014). Secondary organic aerosol is important because larger particles scatter and absorb incoming solar radiation which can affect climate. It can also offer a medium on which heterogeneous reactions can occur. Further, particles smaller than 2.5 µm in diameter

(PM2.5) are small enough to penetrate deep into the alveoli of the lungs. This serves as a vector for any chemical species present in these particles to enter the bloodstream and negatively impact human health (Rohr, 2013).

1.4.1 Ozone production and the peroxy radical

The concentration of O3 depends on the photolysis of NO2. NO2 and nitric oxide (NO), collectively known as NOx, are by-product pollutants from combustion reactions in air. During the day, NO2 will undergo photolysis (at wavelengths < 420 nm) in the presence of oxygen to ultimately form O3 (Reaction 1 and 2).

λ < 420 nm 푁푂2 + ℎ휈 → 푁푂 + 푂 R1 7

푂 + 푂2 + 푀 → 푂3 + 푀 R2

This is the only significant source of O3 in the troposphere (Seinfeld and Pandis, 2006).

Ozone can then oxidize NO to NO2 (Reaction 3) or undergo chemical reactions with other atmospheric species.

푁푂 + 푂3 → 푁푂2 + 푂2 R3

Reaction 1, 2 and 3 all occur at the same time and the cycling between NO and NO2 occur on a timescale of about 1 minute during the day. The end result of these reactions produces no net ozone because it is removed by reaction 3 after it is formed by reaction 1 and 2.

The fate of the majority of VOCs in the atmosphere, when oxidized, is to form an organic peroxy radical (RO2). Subsequently, the primary fate of peroxy radicals is reaction with NO to form NO2 (Reaction 4). This reaction has serious implications for the production of O3.

푁푂 + 푅푂2 → 푁푂2 + 푅푂 R4

Reactions 1, 2 and 3 were shown to result in no net production of ozone. However, when reaction 4 is considered in this chemistry, a new source of NO2 and ultimately O3 (via reactions 1 and 2) is present.

Therefore, it is important to consider VOCs when determining the ozone production in a given region. Due to the vast number of VOCs being emitted into the atmosphere, it is important 8

to focus on the VOCs that are emitted in the highest concentrations or whose atmospheric lifetimes are so long that even small inputs can lead to sizeable concentrations in ambient air. It is important that we better understand how VOCs react in the atmosphere to better understand the repercussions that come with their oxidation.

1.4.2 Oxygenated volatile organic compounds

Oxygenated VOCs are VOCs that contain oxygen. Many OVOCs form due to transformations in the atmosphere via oxidation of VOCs, but they can also occur as a result of biomass burning or be directly emitted into the atmosphere. Many VOC oxidation products have too short a lifetime to accumulate in measurable quantities, whereas others (e.g., ketones) do.

Figure 1.3 shows a small sample of some common OVOCs that have been observed in ambient air (McGraw et al., 1999).

Oxygenated volatile organic compounds are still able to undergo further oxidation reactions. This can result in the formation of very high molecular weight species depending on the VOC that is oxidized and the nature of the oxidation reactions.

One example of VOC oxidation is nopinone, which forms from the oxidation of ß-pinene

(Figure 1.1) by either HO or O3 (Figure 1.4) (Kaminski et al., 2017; Hatakeyama et al., 1989).

Ozone adds to the double bond and undergoes a process called ozonolysis. The results are the formation of a ketone (nopinone) and a criegee biradical (Figure 1.4).

Figure 1.4 ß-pinene oxidation via ozonolysis and HO addition.

Alternatively, HO can add to the double bond forming a tertiary radical and a hydroxyl moiety. In the presence of oxygen, the radical reacts to produce a peroxy radical. After reaction of NO with the peroxy radical to form NO2, the unstable radical species decomposes to form nopinone and a primary alcohol radical (Figure 1.4). Limona ketone (Figure 1.3) is formed via the same pathway on the isopropenyl group of the limonene precursor (Figure 1.1) (Librando and

Tringali, 2005; Donahue et al., 2007).

As is the case with peroxyacyl nitrates (PANs) like peroxyacetic nitric anhydride (PAN,

Figure 1.3), OVOCs can also be reservoir species for NO2. This means that they can chemically bind to NO2 where they can be transported long distances, in particular, downwind of polluted environments where they can thermally dissociate and affect O3 production (Fischer et al., 2010;

Pandey Deolal et al., 2013). In remote regions, PAN can contribute greatly to NOy (=NOx + PAN

+ NO3 + HNO3 + N2O5 + ClNO2 + ...) (Stroud et al., 2003; Alvarado et al., 2010).

1.4.3 Formation of peroxyacyl nitrates

PANs are lachrymators and are toxic to plants in high concentrations (Horvath et al.,

1986; Kleindienst, 1994). Peroxyacyl nitrates are a class of OVOCs that are formed as a product in the photochemical oxidation of VOCs; the same photochemistry forms O3, making them good markers to track which VOCs lead to the formation of O3 (Roberts et al., 2007). They are formed through several pathways summarized in Figure 1.5. The peroxy group that is subsequently formed from the photochemical oxidation of VOCs can undergo reaction with NO to form an organic oxygen radical species. In the presence of oxygen, this species will lose a hydrogen radical which will readily react with O2 to form the hydroperoxyl radical (HO2) and an aldehyde or ketone depending on if the oxidation was on a primary or secondary carbon, respectively. At this point, the aldehyde can undergo several pathways (Figure 1.5), however, if the proton on the aldehyde is removed either by HO or photolytically, the carbon radical will react with atmospheric oxygen to form a peroxyacyl radical. The peroxyacyl radical can either react with

NO to form NO2 or react with NO2 to form a PAN species.

Many PANs have been measured in the laboratory and in the atmosphere (Koppmann,

2007). However, the most abundant species in the atmosphere is PAN (Figure 1.3). PAN is formed from the photochemical oxidation of both AVOCs and BVOCs. Other PANs, such as peroxypropionic nitric anhydride (PPN) and peroxymethacrylic nitric anhydride (MPAN) are

good markers for anthropogenic and biogenic oxidation, respectively. These have been shown to be useful for determining what proportion of O3 is formed from biogenic versus anthropogenic emissions (Roberts et al., 1998).

Figure 1.5 Schematic showing the oxidation of VOCs to form PANs taken from Roberts et al. (2007).

1.5 Quantification of volatile organic compounds

The measurement of atmospheric VOCs is particularly challenging, as there are many

VOCs in the atmosphere. Proton transfer reaction mass spectrometry (PTR-MS) is a type of positive-ion chemical ionization mass spectrometry used for measuring a large array of VOCs in a short amount of time. However, due to its reliance on mass identification, it cannot distinguish between isomers (such as monoterpenes and sesquiterpenes) (Warneke et al., 2003). Further, its

major inherent limitation is that it is unable to detect species that have a proton affinity less than that of water.

An older method for measuring VOCs that is still used ubiquitously today is gas chromatography (GC). Gas chromatography was invented in 1950 by James and Martin (1952) for the separation of volatile fatty acids. Over the years, GC has become a staple in the modern laboratory for its robustness, affordability and flexibility. Gas chromatography is a great method for separating out isomers as well as other VOCs which some more modern instrumentation techniques are unable to do (e.g., PTR-MS). The main principle by which GC separates VOCs is by its boiling point or vapour pressure. The more volatile a compound is, the longer it will spend in the gaseous mobile phase at a given temperature, and the earlier it will elute. For this reason, the retention time of a compound gives a unique perspective into its physical properties.

The short-comings of GC are its poor time resolution as well as its requirement for preconcentration when analyzing trace levels of VOCs in the atmosphere. Further, because of the high volatility of some VOCs, column cooling or the need to change columns is usually required for lighter weight hydrocarbons which can add logistical difficulty when deploying instruments in the field if the GC requires an external cooling unit or requires multiple columns.

Many detection methods can be coupled to GC such as Flame Ionization Detection (FID),

Electron Capture Detection (ECD), or Mass Spectrometry (MS) (Skoog et al., 2007). Mass spectrometry is an excellent tool that has been around since the 1940’s as a method used by the petroleum industry for quantifying complex mixtures of hydrocarbons (Skoog et al., 2007). It is 14

particularly beneficial because it can provide insight into structural features (Majchrzak et al.,

2018). By using electron impact, compounds will fragment into smaller ions which are then detected based on their mass to charge ratio.

One of the key challenges for VOC quantification by GC is calibration. Separate calibrations are needed for each compound, requiring analytically pure standards which can often come at a large price and be unstable if stored in a cylinder for long periods of time (Jones et al.,

2014). Another difficulty arises when attempting to generate small concentrations of calibration standards, as dilution and sample preparation can introduce large error into the measurement.

1.5.1 Gas chromatography mass spectrometry

The use of mass analyzers coupled to GC greatly improves chemical identification. Most commonly, quadrupole mass analyzers are used because of their simple design and affordability.

An alternative, more recently developed mass analyzer is the cylindrical ion trap (CIT) that allows the user to apply selected-ion monitoring (SIM), which improves the sensitivity of an instrument by allowing only ions of a particular mass to the detector. The CIT is also more compact than a quadrupole allowing for smaller instruments capable of field deployment.

However, this type of mass analyzer is only rarely used in practice (Harold, 1991; Hemberger et al., 1991).

Typically, a sample is introduced into a gas chromatograph ion-trap mass spectrometer

(GC-ITMS) via a syringe. However, direct air sampling without the need of syringes is possible

via the use of a sample loop. In this method, ambient air is pumped through a sample loop of fixed volume. A valve is then actuated, and the contents of the sample loop are introduced via the carrier gas to the column where separation occurs. However, due to the trace levels of VOCs in the atmosphere, sample preconcentration is often required.

1.5.1.1 Sample preconcentration

The two sample preconcentration methods most commonly used in the atmospheric sciences are cryogenic trapping and solid-phase adsorbents (Ras et al., 2009). In cryogenic trapping, a cryogenic liquid (or some other physical means, i.e., electric cooler) is used to dramatically lower the temperature of a small piece of tubing or capillary. Any chemical species in an air sample will condense or desublimate onto the cold surface and preconcentrate based on the volume of air passed over the cooled region. The tube is then flash heated to rapidly evaporate or sublimate the analyte of interest for analysis. Unfortunately, with cryogenic trapping, oxidants and water also condense and can chemically react with analytes or expand and damage the capillary, so the sample needs to be scrubbed and desiccated which can have unwanted side effects on the analyte of interest (Ras et al., 2009).

The more common, practical, and affordable approach is to use a solid-phase adsorbent.

This is a solid material that will (ideally) selectively adsorb the analyte of interest. The adsorption is temperature dependent so that species can adsorb at low temperatures, then desorb at higher temperatures. Many adsorbents are available for a wide range of applications. However, some adsorbents are known to produce artifacts or be dependent on humidity (Ras et al., 2009).

1.6 Thesis motivation and objectives

This thesis was motivated by the need for speciated measurements of VOCs, as more modern mass spectrometry instrumentation allows for the rapid measurement of thousands of compounds while being blind to isomeric species. Terpenes in particular are an incredibly diverse group, containing hundreds of possible isomers each with their own unique reactivity.

For example, limonene has a 3 times higher reaction rate constant with hydroxyl radical than

α-pinene at room temperature (Gill and Hites, 2002). Determining natural ambient mixing ratios of monoterpenes in different regions provide the modelling community with numbers to constrain typical values of the region in order to better understand the chemistry.

This thesis highlights the importance of using speciated measurements to perform analyses that might not otherwise be possible with other instrumentation. Analyses such as source apportionment, source identification and chemical transformations and aging. In this work, I used two primary instruments, a GC-ECD for the detection and quantification of PAN species and a GC-ITMS for the detection and quantification of monoterpenes, n-alkanes, and

BTEX compounds.

In Chapter 2, I discuss the details of the instrumentation including subtleties about the inlet setup, preconcentration details and temperature and pressure considerations. I describe in detail the multiple methods for calibration that were used in this work. I also provide some details regarding auxiliary measurements that aided in the analyses and helped to characterize and calibrate.

In Chapters 3 through 8, I outline three separate field campaigns: Fort McMurray Oil

Sands Strategic Investigation of Local Sources (FOSSILS 2013) in Chapters 3, 4 and 5, Ozone- depleting Reactions in a Coastal Atmosphere (ORCA 2015) in Chapters 6 and 7, and PAN chemistry in forest fires (PANFire 2017) in Chapter 8 and discuss the data sets in detail. For each campaign I highlight the importance of speciated measurements and give a detailed outlook from the conclusions of each study. All three field studies had very unique environments that each came with their own set of challenges. This includes a remote, coastal environment, an environment heavily influenced by petrochemical industries, and an urban environment impacted by forest fires.

In Chapter 9, I conclude and bring together my findings from all Chapters to make a final case for speciated measurement techniques.

Chapter Two: Instrumentation and Calibration

Parts of Chapter 2 have been published by Atmospheric Measurement Techniques, doi: 10.5194/amt-7-3263-2014 (Tokarek et al., 2014) and Atmosphere-Oceans, doi: 10.1080/07055900.2017.1306687 (Tokarek et al., 2017).

2.1 Gas chromatography ion trap mass spectrometer

Throughout this thesis, the primary instrument used to make measurements was the Griffin

450 gas chromatograph equipped with a cylindrical ion trap mass spectrometer and electron impact ionization. The GC-ITMS always sampled from a stainless-steel inlet with an outer diameter (o.d.) of 0.635 cm and an inner diameter (i.d.) of 0.386 cm. The length of the inlet was

3.6 m long and 5 m above ground at FOSSILS (Chapters 3-5) and 6.5 m long and 4.3 m above ground at ORCA (Chapters 6-7). The method details are summarized in Table 2.1 and include

GC oven temperature programs, mass ranges, run times, adsorption/desorption times and temperatures which differed between campaigns. The instrument utilizes a dual sorbent trap containing Tenax TA and Carboxen 1017 for the measurement of trace VOCs with variable adsorb and desorb times and temperatures. The preconcentrated sample is then flash heated using a high-resistance ceramic heater and desorbed onto a 30 m (length) × 0.25 mm (inner diameter) ×

0.25 μm (film thickness) DB-5MS column using helium as a carrier gas (Praxair, ECD grade) purified using a triple trap (Restek 22464). The sampling frequency of the detector was held at a rate of approximately 3 Hz. The instrument weighs 44 kg, and the dimensions are 49 cm × 49 cm

× 54 cm, making it very compact and particularly suitable as a field instrument (Bednar et al.,

2011). Details on calibration can be found in the respective Chapters (3 and 6).

Table 2.1 Method details for the GC-ITMS during field campaigns. Total Adsorb Desorb Mass Oven temperature Field campaign run time Time Temp Time Temp range program (min) (min) (°C) (min) (°C) Start temperature: 40 °C FOSSILS (before hold for 3 min, Ramp to Aug 27, 2013 at 50-425 70 °C at 1.5 °C/min, Ramp 40 15:46 UTC) to 200 °C at 10 °C/min and hold for 4 min 10 Start temperature: 40 °C FOSSILS (after hold for 3 min, Ramp to Aug 27, 2013 at 50-425 70 °C at 1.5 °C/min, Ramp 53 40 5 240 15:46 UTC) to 200 °C at 5 °C/min and hold for 4 min

Start temperature: 40 °C hold for 3 min, Ramp to 70 °C at 3 °C/min hold for ORCA 2015 35-425 26 8 2 min, Ramp to 200 °C at 15 °C/min and hold for 2 min

The GC-ITMS column oven is designed to be small and insulated with little empty space to heat up quickly and efficiently to a maximum of 300 °C. The CIT, ionization source and detector assembly requires a high vacuum (~3 × 10-5 Torr) during operation which is accomplished by a four-stage backing pump and a turbo pump.

The instrument uses icx™ technologies griffin system software (version 3.7.6.0) for data acquisition and peak fitting with a secondary software (GSS Toolbox version 2.0) used to export 20

acquired data into other formats (i.e., .csv). For the purposes of this work, data were exported in

.csv format and copied into Igor Pro 6 (Wavemetrics). This enabled me to more readily select and plot ions of interest and to write macros to fit peaks, save peak areas and save gaussian fitting parameters. The main integration method used for fitting peaks was trapezoidal integration, which calculates the area of the trapezoid between two adjacent points and their respective baseline points and sums all trapezoids within the identified bounds.

2.1.1 Electron impact ionization

Electron impact is a common method used to ionize an analyte. The electron source is a tungsten filament heated using a high current which causes free electrons to be emitted. The electrons collide with the analytes in their path and are then directed into a mass analyzer (in this case a CIT). The collisions between electrons and analytes will be strong enough (energy = 70 electron volts (eV)) to cause fragmentation of the analyte which is detected in this instrument via a dynode/electron multiplier detector. A plot of the number of fragments from chemical species hitting the detector as a function of the fragments’ mass to charge (also called a mass spectrum) can be used as a “ion fingerprint” to help identify an analyte.

2.1.2 Cylindrical ion trap

The CIT is a mass analyzer that was developed out of Purdue University (Patterson et al.,

2002; Riter et al., 2002). The CIT is similar to the more traditional quadrupole ion trap in that it has the same hyperbolic electrode geometry as a cross-section of the quadrupole mass analyzer.

The particular shape of the CIT allows the mass analyzer to take up much less space enabling the

development of a smaller instrument. Like the quadrupole, ions are trapped in an oscillating electric field. The field is generated such that it will only trap ions of interest that have trajectories stable in the field.

2.1.3 Injection methodology

In this work I used a different injection methodology compared with traditional GC injection (i.e., sample injection via syringe). Sampled air was pulled through the inlet system (via pump) and flows over the preconcentrator at 40 °C as shown in Figure 2.1A. When the valve is actuated (after a user-set adsorption time, see Table 2.1), the carrier gas flows over the preconcentrator while it is simultaneously flash heated for a user-set desorption time (Table 2.1).

In these positions, the inert carrier gas, cleaned by a helium-specific inline “triple trap” (Restek

22469), is always flowing over the analytical column into the mass spectrometer such that only a small plug of ambient air generated by desorption of the preconcentrator is injected onto the column. After the set desorption time, the preconcentrator is cooled under a constant flow of purified carrier gas to minimize analyte carrier-over. New preconcentrators were used before each field campaign and were conditioned using a manufacturer method. This method will run several short 5-minute chromatograms in which the preconcentrator is gradually heated to high, adsorbent-stable temperatures to drive off any volatiles that may have adsorbed to the sorbent during manufacture, transport and during replacement of the previous preconcentrator.

2.1.4 Low mass, high mass and detector tuning

The GC-ITMS comes with its own built-in calibrant used regularly to calibrate the CIT and detector. The calibrant is heptacosafluorotributylamine (N(C4F9)3) which is one of the most widely used reference compounds in mass spectrometry (Lide, 1998). The calibrant enables the

user to calibrate the overall mass range of the instrument and was performed regularly during campaigns and always after powering up the instrument.

The detector tune is performed in order to limit mass shifting of the detector by increasing the multiplier bias voltage. This tune was also performed at regular intervals during campaigns and was always run after powering-up the instrument.

2.2 Peroxyacyl nitrate gas chromatograph specifications

The PAN-GC used in this work was a Varian 3380CP equipped with an ECD detector

(model number 02 001972 01) that was converted to measure PANs in a similar fashion to that described by Fischer et al. (2010). A two-position sample valve (VICI Valco EHC10WE, 1/1600

× 0.40 mm with a micro electric actuator) and a sample loop constructed from polyether ether ketone (PEEK, VICI “Cheminert”) were mounted inside the GC oven as shown in Figure 2 of

Tokarek et al. (2014). A 10-port valve was used (in lieu of a 6-port valve) because it was the only valve on hand. The analytical column used was a 15 m, medium polarity “megabore” column (Restek RTX-1701) with a 0.53 mm i.d. and 1 μm film thickness. Sample air was drawn through a 2 mL sample loop using a miniature air compressor (McMaster-Carr 4404K15) at a flow rate of 115 standard cubic centimeters per minute (sccm) set by flow restriction.

Helium is quite expensive to use as a carrier gas relative to hydrogen and the cost can fluctuate dramatically (Reisch, 2017). For the PANFire 2018 field campaign (Chapter 8), hydrogen (as opposed to the typically used helium) was explored as a potential carrier gas for the

measurement of PANs using ECD. Initially, hydrogen was delivered from a commercial hydrogen generator (PEAK Scientific, model Precision Hydrogen 200) at a flow rate of

24 mL min-1 and a make-up gas flow of nitrogen (Praxair, high purity grade, ≥99.998%) at

42 mL min-1, optimized for a minimum background signal. After October 2017, hydrogen was delivered from a gas cylinder (Praxair, high purity grade, ≥99.995%) at a flow rate of

24 mL min-1 and a make-up gas flow of 48 mL min-1. The ECD signal voltage was digitized using a Universal Serial Bus (USB) 2.0 data acquisition module (Omega) connected to a laptop computer running software written in National Instrument’s LABVIEW 2010, which also controlled the main switching valve. The column oven was set to a temperature of 25 °C but was occasionally higher in the afternoons as the trailer housing the instruments warmed up to above this temperature. Calibration and other pertinent details on this instrument can be found in

Chapter 8.

2.2.1 Inlet design and schematic

A 3 m long section of the GC-ITMS inlet was heated to 125 °C (1 m to 110 °C during

FOSSILS) using a custom ordered Clayborn tubing heater to remove interference due to O3 oxidation of unsaturated hydrocarbons during preconcentration (Pollmann et al., 2005; Hellén et al., 2012). This method was tested in the lab prior to the first field campaign. Figure 2.2 shows that when ~ 100 ppbv of ozone and ~ 1 ppm of α-pinene are sampled at the same time, no loss of

α-pinene occurs when the stainless-steel is heated.

A similar experiment as that in Figure 2.2 showed that O3 was indeed scrubbed when the stainless-steel inlet was heated (Figure 2.3). This occurred at ~110 °C and effectively removed all O3. Even low temperatures scrubbed some (~10%) ozone. This scrubbing effect continued up to the maximum temperature tested (150 °C). Based on these experiments, we used an inlet temperature of 110 °C for the FOSSILS campaign (Chapters 3-5) and an inlet temperature of

125 °C for the ORCA campaign (Chapter 6-7). The higher inlet temperature used during the

ORCA campaign was due to frequent precipitation events and relatively (compared with

FOSSILS) cold ambient temperatures which might have affected the temperature of the stainless-steel tubing (which was inserted through the center of heater).

Figure 2.3 Ozone concentration as a function of temperature. Ozone is completely scrubbed at 110 °C.

Chapter Three: Data Collected at a Ground Site Near Fort McKay During FOSSILS 2013

Parts of Chapter 3 have been published by Atmospheric Chemistry and Physics, doi: 10.5194/acp-18-17819-2018 (Tokarek et al., 2018).

3.1 Field campaign details

In August 2013, a comprehensive air quality study was conducted as a part of the Joint Oil

Sands Monitoring (JOSM) plan (JOSM, 2012), referred to here as the 2013 JOSM intensive study. This study was performed in northern Alberta at two ground sites in and near Fort McKay and from a National Research Council of Canada (NRC) Convair 580 research aircraft to characterize oil sands emissions and their downwind physical and chemical transformations

(Gordon et al., 2015; Liggio et al., 2016; Li et al., 2017).

One of the ground sites, located at the Wood Buffalo Environmental Association (WBEA) air monitoring station (AMS) 13 (Figure 3.1), was equipped with a comprehensive set of instrumentation to measure concentrations of trace gases and aerosols (see section 3.2). As part of JOSM, the GC-ITMS described in Chapter 2 was deployed at AMS 13.

The immediate vicinity of the site consisted of mixed-leaf boreal forest with a variety of tree species, including poplar, aspen, pine and spruce trees (Smreciu et al., 2013). The site was accessible via a gravel road; traffic on this road was restricted during the study period (August –

September 2013).The site is impacted by emissions from nearby oil sands facilities (Figure 3.1),

including a large surface mining site operated by Syncrude Canada whose northeastern corner is located 3.5 km to the south of AMS 13 (and which is adjacent to the 5 km long Syncrude –

Mildred Lake (SML) tailings pond) and from a large upgrader stack facility operated by Suncor

Energy Inc. located to the Southeast. Additional oil sands facilities are operated (during the study period) by Canadian Natural Resources Limited, Imperial Oil, and Shell Canada to the N and

NE.

When air masses passing over regions with industrial activities were observed (as judged from a combination of local wind direction and tracer measurements), the total ion chromatogram showed an analytically unresolved hydrocarbon signal associated with intermediate volatile organic compounds (IVOCs) which have a saturation concentration (C*) in the range 105 µg m-3< C* < 107 µg m-3 (Liggio et al., 2016).

Emission estimates for analytically unresolved hydrocarbons range from 5106 kg year-1 to

14106 kg year-1 for the two facilities that reported such emissions (Li et al., 2017). Using aircraft measurements during the 2013 study, Liggio et al. (2016) showed that IVOCs contributed to the majority of the observed secondary organic aerosol (SOA) mass production in a similar fashion as anthropogenic VOCs contributed to SOA production during the Deepwater

Horizon oil spill (de Gouw et al., 2011) and rivaling the magnitude of SOA formation observed downwind of megacities (Liggio et al., 2016), though ultimately it has remained unclear which activities are associated with IVOC emissions.

3.2 Instrumentation

A large number of instruments was deployed for this study; a partial list whose data were utilized in this manuscript is given in Table 3.1. Detailed descriptions of these instruments and operational aspects such as calibrations can be found in the supplementary information associated with the work currently in discussion (Tokarek et al., 2018) . Sample observations of analytically unresolved hydrocarbons by GC-ITMS and how these data were used in the analysis

are described below.

Table 3.1 Instruments used to measure ambient gas-phase and aerosol species during the 2013 JOSM intensive study at AMS 13. Instrument and Time Species measured Operated by Reference Model resolution York University and Picarro CRDS (Chen et al., 2013; CO, CO , CH 1 min G2401 2 4 ECCC Nara et al., 2012) Thermo (Tokarek et al., 2014; Scientific, Model NOy University of Calgary 10 s Odame-Ankrah, 42i 2015) Blue diode (Paul and Osthoff, cavity ring-down NO2 University of Calgary 1 s 2010; Odame- spectroscopy Ankrah, 2015) Thermo (Tokarek et al., 2014; Scientific Model O3 University of Calgary 10 s Odame-Ankrah, 49i 2015) Griffin/FLIR, (Tokarek et al., 2017; model 450 GC- VOCs University of Calgary* 1 hr Liggio et al., 2016) ITMS Thermo Scientific TS ECCC 1 min n/a CON101 Thermo Scientific SO2 ECCC 1 min n/a 43iTLE (Markovic et al., AIM-IC NH , NH + University of Toronto 1 hr 3(g) 4 (p) 2012) University of Toronto 1-5 min Aerodyne rBC, NH + , SO 2 , 4 (p) 4 (p) (variable (Onasch et al., 2012) SP-AMS NO - , Cl , organics and ECCC 3 (p) (p) ) 5-6 min PM size (Peters and Leith, TSI APS 3321 10-1 University of Calgary (variable distribution 2003) ) TSI SMPS (3081 (Wang and Flagan, DMA, 3776 PM size distribution University of Alberta 6 min 1 1990) CPC) EcoChem (Wilson et al., 1994; Analytics PAS pPAH ECCC 1 min Burtscher et al., 2000CE 1982) *This work 31

3.3 Analytically unresolved hydrocarbon signature

The total ion chromatogram of the GC-ITMS occasionally showed elevated and

* analytically unresolved hydrocarbons in the volatility range of C11 – C17 with C ranging from

105 µg m-3< C* < 107 µg m-3. An example is shown in Figure 3.2.

An offline analysis of the headspace above ground-up bitumen gave a similarly unresolved hydrocarbon signal (Figure 3.2, black trace). In this particular case, the ambient air chromatogram also shows enhancements of lower molecular weight hydrocarbons (possibly from naphtha) that were not observed in the bitumen sample. 32

The major ions contributing to the unresolved signals in Figure 3.2 are associated with alkanes (i.e., m/z 55, 57, 67, 69, etc., see Chapter 4). In contrast, counts at masses associated with

+ + aromatics (i.e., m/z 115, C9H7 , and m/z 91, C7H7 ) as reported by Cross et al. (2013) were negligible in both the bitumen head space and polluted day samples. The strong resemblance of the unresolved hydrocarbon feature in ambient air with the bitumen head space sample both in terms of volatility (i.e., elution time) and electron impact mass fragmentation is consistent with bitumen as the source of IVOCs at this site.

In the interpretation of the integrated IVOC signal, it is assumed that it is of primary origin, i.e., emitted directly from point sources in the vicinity of the measurement site. To gain a rough understanding of the frequency and magnitude, the unresolved signal was integrated from a retention time of 25 min to 45 min (gray area in Figure 3.2) in all ambient air chromatograms.

The IVOCs observed in this work likely encompass a portion of the total that is emitted.

For example, IVOCs generated by combustion processes, such as aircraft engine exhaust, are comprised of alkanes, aromatics and oxygenated compounds (Cross et al., 2013). The use of a chromatographic column in this work biases the IVOC signal towards hydrocarbon-IVOCs, since oxygenated compounds (i.e., alcohols and carboxylic acids) and sulfur containing compounds

(i.e., thiols and sulfonic acid) will not elute from the analytical column due to volatility.

Furthermore, the recovery of VOCs from the preconcentration unit, while reproducible and likely complete for n-alkanes which bracket the bulk of IVOC emitted and whose calibration curves were linear, is not known for late-eluting compounds, but is assumed to be sufficiently reproducible to yield a semi-quantitative signal. 33

3.4 Field data summary

Time series of relevant data from the field are presented in Figure 3.3, grouped approximately by source type. Statistics of a subset of the data (i.e., median, average, maxima, minima, etc.) used in a Principal Component Analysis (PCA) in this work (see Chapter 5) is shown in Table 3.2. The majority of the data was above the LOD with the exception of total reduced sulfur (TRS) and sulfur dioxide (SO2) which were both < 81%. The time series show complex and unique air masses which are highly variable. The data have been divided into groups loosely based on their common sources for the purposes of source identification. These groups are: anthropogenic VOCs, biogenic VOCs, combustions tracers, aerosol species, sulfur species and other (Table 3.2).

Time series of VOCs of primarily anthropogenic origin (i.e., o-xylene, 1, 2, 3- and

1, 2, 4-TMB, etc.) as well as the IVOC signature are shown in Figure 3.3A. The abundances of these species, as well as the other compounds, varied as a function of time of day (i.e., boundary layer mixing height) and air mass origin, with higher VOC concentrations generally observed during daytime. The VOC concentrations varied between nearly pristine, remote conditions, with concentrations below detectable limits, to mixing ratios of aromatic species exceeding 100 pptv.

The concentration range of o-xylene is within the extremes reported by WBEA in their 2013 annual report (WBEA, 2013), exemplifying that the data set is representative of typical pollutant levels in this region.

Table 3.2 Variables observed at the AMS 13 ground site during the 2013 JOSM campaign.

Variable Units Mediana Averagea,b Standard LODe Min.a Max.a Fraction deviationa,b

While there is some obvious covariance between variables (i.e., when the mixing ratios of one particular VOC increases, so do others), the ratios of hydrocarbons varied considerably. For example, on August 18, 10:50 UTC, the n-decane to o-xylene ratio was ~22:1, whereas on

August 24, 07:40 UTC it was ~1:5.7. The IVOC magnitude also varied greatly and often increased and decreased in tandem with the other AVOCs (e.g., on Aug 24, 16:30 UTC) but also increased independently from AVOC abundances (e.g., on Aug 30, 01:20 UTC, and on the night of Aug 22). This behaviour suggests the presence of multiple sources with distinct signatures that are being sampled to a varying extent at different times. This, coupled with the intermittency of the highly elevated signals, presents an analysis problem frequently encountered in environmental analysis that is usually investigated through a factor or principal component analysis (Thurston et al., 2011; Guo et al., 2004).

Presented in Figure 3.3B are the time series of NOy, rBC and particle-surface bound PAH

(pPAH) abundances, all of which are combustion by-products. For example, rBC is emitted from combustion of fossil fuels, biofuels, open biomass burning, and burning of urban waste (Bond et al., 2004). Similar to the VOCs, the abundances of these species varied greatly, from very low, continental background levels (i.e., < 100 pptv of NOy, < LOD for rBC and pPAHs) to polluted

-3 -3 concentrations (i.e., > 60 ppbv of NOy, > 1 μg m rBC, > 10 ng m pPAHs) characteristic of polluted urban and industrial areas.

When high concentrations of NOy were observed, its main component was NOx (data not shown), which is a combustion by-product usually associated with automobile exhaust. In the

Alberta oil sands, emissions from off-road mining trucks as well as the upgrading processes are 37

the main contributors to the NOy burden (Percy, 2013; Watson et al., 2013).

Shown in Figure 3.3C are the mixing ratios of the greenhouse gases CH4 and CO2 along with CO. Abundances of CO2 were clearly attenuated by photosynthesis and respiration of the vegetation near the measurement site, as judged from the strong diurnal cycle in its concentration

(not shown). Maxima typically occurred shortly after sunrise, coincident with the expected break-up of the nocturnal boundary layer. In addition to biogenic emissions from vegetation and soil, CO2 originates from a variety of point and mobile sources in this region, including off-road mining trucks (Watson et al., 2013) and the extraction, upgrading, and refining of bitumen and on-road vehicle sources in the area (Nimana et al., 2015b, a). Concentrations of CO2 spiked whenever these emissions were transported to the measurement site.

Concentrations of CH4 also exhibited a diurnal cycle, with higher concentrations generally observed at night and peaking in the early morning hours. While CH4 and CO2 mixing ratios frequently correlated in plumes, their ratios were variable overall, suggesting they often originated from distinct sources. Potential methane point sources in the region include microbial production in tailings ponds (Siddique et al., 2012) and fugitive emissions associated with the mining and processing of bitumen (Johnson et al., 2016). Indeed, a recent analysis shows tailings ponds and open pit mining sources to be the largest sources of CH4 in the region (Baray et al.,

2018).

Similar to the AVOCs, the abundances of CH4 and CO2 were highly variable and ranged from minima of 1.88 and 384 ppmv to maxima of 2.96 and 578 ppmv, corresponding to 38

maximum enhancements of 1.63 and 1.47 relative to tropospheric global monthly means of

1.806 ± 0.001 and 394.3 ± 0.1 ppmv for July, 2013 (Dlugokencky, 2017b, a), respectively.

Mixing ratios of CO also varied with time but generally were not elevated greatly (median

118 ppbv) above background levels (minimum 91 ppbv), except for occasional spikes in concentration (Figure 3.3C). Carbon monoxide is a tracer of biomass burning and fossil fuel combustion, in particular in automobiles with poorly performing or absent catalytic converters, but is also a by-product of the oxidation of VOCs, in particular of methane and isoprene which are oxidized over a wide area upwind of AMS 13 (Miller et al., 2008).

Time series of sulfur species and PM10-1 volume are shown in Figure 3.3D. The TS and

SO2 data are dominated by intermittent plumes containing SO2 mixing ratios exceeding 5 ppbv.

The highest mixing ratio observed was 92.5 ppbv (in between the preconcentration periods of the

GC-ITMS). Mixing ratios of SO2 exhibited the most variability of all pollutants, as judged from the standard deviation of each of the measurements (Table 3.3). TRS levels were generally small

(< 1 ppbv) and variable, except for plumes; TRS abundances in plumes, however, are more uncertain since they were calculated by subtraction of two large numbers. When TS and SO2 abundances were low (< 1 ppbv), TRS abundances were variable and occasionally exhibited spikes that did not show any obvious correlation with other variables, suggesting the presence of one or more distinct TRS sources. PM10 volume concentrations varied a lot as well and, just like

TRS, did not show an obvious correlation with other variables. Fugitive dust emissions likely contributed to much of the PM10 volume in the Athabasca oil sands region (Wang et al., 2015).

Time series of monoterpene mixing ratios are shown in Figure 3.3E. α-Pinene was generally the most abundant monoterpene, followed by β-pinene. Their campaign average ratio was 1:0.85, though occasionally the α- to β-pinene ratio was as low as 1:2 (e.g., on Aug 28,

14:50 UTC and Sept 5, 12:40 UTC). Terpene mixing ratios were generally higher at night than during the day, with maxima of 1.9 and 1.6 ppbv, respectively, a diurnal pattern consistent with what has been observed at other forest locations (Fuentes et al., 1996).

Monoterpenes are emitted by plants via both photosynthetic and non-photosynthetic pathways (Fares et al., 2013; Guenther et al., 2012); at night, their emissions accumulate in a shallow nocturnal boundary layer, whereas during daytime, they are entrained aloft (above the canopy) and oxidized by HO and O3, which are more abundant during the day than at night

(Fuentes et al., 1996). α- and β-Pinene mixing ratios were lowest mid-day (median values at noon of 140 and 133 pptv, respectively). The largest daytime concentrations were observed on

Aug 25, a cloudy day (as judged from spectral radiometer measurements of the NO2 photolysis frequency): on this particular day, mixing ratios at noon were 687 and 850 pptv, respectively.

Also shown in Figure 3.3E is the time series of ammonia. These data were dominated by spikes which were observed sporadically and did not correlate with other variables, suggesting the presence of nearby ammonia point sources. Concentrations of ammonia were not as variable as some of the other pollutants (e.g., the AVOCs, sulfur species) as judged from its standard deviation (Table 3.3), which suggests a geographically more disperse source or sources similar to

CO or CH4, which have a "background". This is consistent with a recent study by

Whaley et al. (2018) that estimated over half (~57%) of the near-surface NH3 during the study 40

period originated from NH3 bi-directional exchange (i.e. re-emission of NH3 from plants and soils), with the remainder being from a mix of anthropogenic sources (~20%) and forest fires

(~23%).

3.4.1 Qualitative assessment of VOCs

A qualitative assessment was conducted on all ambient data collected. In this assessment, chemical markers above a threshold value were used to identify potential sources. These chemical markers include SO2 in mixing ratios greater than 5 ppbv as a marker of upgrader stacks, ortho-xylene in mixing ratios greater than 25 pptv as a marker of tailings ponds emissions, CH4 and CO2 in mixing ratios greater than 1900 ppbv and 400 ppmv, respectively, as a marker of mine face emissions, and rBC in concentrations greater than 30 ng m-3 as a marker for diesel truck emissions. In addition to the above stated criteria, several adjunct chemical markers were used to verify the potential source types by confirming their concurrent elevated mixing ratios. Lastly, the IVOCs chromatograms were visually inspected for unknown peaks elevated above baseline noise. High volatility events occurred when more than 5 unknown peaks were present in the region from 0-10 minutes, medium volatility events occurred when more than

5 unknown peaks were present in the region from 10 minutes - 25 minutes, and low volatility

(aka IVOCs) events occurred when more than 10 unknown peaks were present in the grey region

(25-45 minutes) in Figure 3.2. Figure 3.2 (red trace) is also an example of a chromatogram when all volatility events were observed. All chemical markers mentioned above are summarized in

Table 3.3.

Table 3.3 Threshold criteria and associated adjunct chemical markers for a qualitative assessment of potential oil sands sources. Adjunct Potential source Chemical marker Threshold value markers Upgrader stacks SO2 > 5 ppbv NOx and CO2 Tailings ponds Ortho-xylene > 25 pptv CH4 CH > 1900 ppbv Mine face 4 - CO2 > 400 ppmv -3 Diesel trucks rBC > 30 ng m NOx High volatility 0-10 minutes in > 5 peaks - event chromatogram Medium 10-25 minutes in > 5 peaks - volatility event chromatogram Low volatility 25-45 minutes in > 10 peaks - event chromatogram

A time series of regions where these chemical markers were observed is shown in

Figure 3.4. Many events occurred simultaneously and all, with the exception of facility stacks, were observed both during the day and at night. This suggests that these VOCs may have multiple sources in the Athabasca oil sands region which may or may not include the mining of raw bitumen. It also suggests that the upgrader stack events may only be visible when the boundary layer is high enough so that the stacks emit directly into the mixed layer. At night, when the nocturnal boundary layer is decoupled from the residual layer, the plumes may not be able to mix down to the surface.

Figure 3.4 Time series of a qualitative analysis of potential sources of VOCs. Each bar represents a region of time where a particular source may have been sampled based on chemical markers.

Plumes high in sulfur, likely from upgrader stacks (red bars in Figure 3.4) were primarily characterized by spikes in SO2 concentrations. These spikes were also commonly accompanied by small increases in CO2 and NOx. Six events were identified based on the above criteria in the time frame shown in Figure 3.4 (Aug 17 – Sep 6) always during the day, and usually for a duration of 8 to 12 hours.

Mine faces are known to emit high concentrations of CH4 and CO2, so these were used as markers to explore potential mine face impacted air. Many events containing elevated CH4 and

CO2 were observed and had varying duration from a couple of hours to a full day (24 hours).

Many other sources could be responsible for high CH4 and CO2 including natural gas generators, other combustions sources, and biogenic and bacterial emissions. However, this could only be made clearer using an isotope analysis for which instrumentation during this campaign was not available.

Tailings ponds were tentatively identified by spikes in o-xylene, a common component of naphtha, and CH4. Higher alkanes such as n-decane and n-undecane were also enhanced during these spikes. Possible tailings plumes were observed 8 times with highly varying duration similar to mine faces.

Lastly, diesel trucks as a potential source were explored based on enhancements of refractory black carbon (rBC) and NOx. pPAHs were also often elevated during these plumes.

Sixteen events were determined visually. This is not surprising given the large amount of diesel vehicle traffic in the region.

A qualitative approach is not sufficient to identify sources since many sources were observed at the same time with chemical marker concentrations varying significantly (several orders of magnitude). A more mathematical approach (PCA) was investigated for this data set to elucidate the source of the IVOCs.

3.5 Conclusions

The main objective of this work is to elucidate the origin of the IVOC signature observed at the AMS 13 ground site downwind from the AB oil sands mining operations (Figure 3.1).

Emission inventories show that the facilities that process the mined bitumen are by far the largest anthropogenic point sources in the oil sands region (NPRI, 2013), consistent with recent aircraft measurements (Baray et al., 2018; Howell et al., 2014; Li et al., 2017; Simpson et al., 2010) which have shown substantial emissions of NOy, SO2, CO, VOCs, CO2, and CH4, from these facilities and associated mining activities. As the data presented in this Chapter cannot explicitly point to sources on their own, a more rigorous approach was investigated. PCA was used to elucidate potential source types in the region. These results are discussed in further detail in

Chapter 4.

Chapter Four: Principal Component Analysis of Intermediate Volatility Organic

Compounds Observed During the Summer in the Athabasca Oil Sands Region of Alberta,

Canada During FOSSILS 2013

Parts of Chapter 4 have been published by Atmospheric Chemistry and Physics, doi: 10.5194/acp-18-17819-2018 (Tokarek et al., 2018).

4.1 Introduction

In this Chapter, measurements of the air pollutants discussed in Chapter 3 are analyzed using principal component analysis (PCA) to elucidate the origin of the observed IVOCs in the

Athabasca oil sands. The analysis presented here is a receptor analysis focusing on the normalized variability of pollutants impacting the AMS 13 ground site and hence does not constitute a comprehensive emission profile analysis of the oil sands facilities as a whole, for which aircraft-based measurements and/or direct plume or stack measurements are more suitable.

Details regarding instrumentation, site location and data were discussed in Chapter 3. Principal component analysis was chosen over the more popular positive matrix factorization (PMF) method (Paatero and Tapper, 1994) because it yields a unique solution and is particularly suited as an exploratory tool for identification of components without a priori constraints (Jolliffe and

Cadima, 2016).

The PCA was complemented by bivariate polar plots (Carslaw and Ropkins, 2012;

Carslaw and Beevers, 2013) to show the spatial distribution of sources in the region as a function

of locally measured wind direction and speed. A second PCA was performed to investigate which components correlate with (and generate) secondary pollutants, i.e., pollutants that are formed by atmospheric processes. Potential sources and processes contributing to each of the components identified by PCA are discussed.

4.2 Principal component analysis

Principle component analysis is a multivariate technique that is often applied to arrays of pollution variables to derive information about what type of pollution sources influence the data

(Thurston and Spengler, 1985). Typically, the matrix constructed by pollution variables is approximated by the product of two smaller matrices (Otto, 2007). In other words, the data matrix, where rows are data points and columns are pollutants, is decomposed to form a new matrix (the loading and scoring matrix) that describes the data in terms of a small number of components. Each column in the loading matrix is interpreted as a potential source or source type. To improve the interpretability of the resulting loading matrix, a rotation method is often used (Guo et al., 2004). In this work, varimax rotation was used which maximizes the variance of the loading matrix. The PCA was carried out using the "Statistical Analysis System" (SAS™)

Studio 3.4 software (2015) using a method similar to that described by Thurston et al. (2011;

1985). The source-related components and their associated profiles are derived from the correlation matrix of the input trace constituents. This approach assumes that the total concentration of each "observable" (i.e., input variable) is made up of the sum of contributions from each of a smaller number of pollution sources and that variables are conserved between the points of emission and observation.

4.2.1 Selection of variables

Twenty-two variables whose ambient concentrations are dominated by primary emissions or which are formed very shortly after emission (such as the less oxidized oxygenated organic aerosol (LO-OOA) factor observed by the Soot Particle Aerosol Mass Spectrometer (SP-AMS), see below) were included in the PCA (Table 4.2). These variables included CO2, CH4, NOy, CO, and SO2, which are known to be emitted in the oil sands region from stacks, the mine fleet and faces, tailings ponds, and by fugitive emissions (Percy, 2013). The median NOx to NOy ratio was

0.85, consistent with the close proximity of the measurement site to emission sources and limited chemical processing. Because NOx constituted a large fraction of NOy, its temporal variation was captured by the latter, and it was not included as a separate variable in the PCA analysis.

For this work, mixing ratios of all non-methane hydrocarbons (NMHCs) that were quantified (i.e., o-xylene, the n-alkanes decane and undecane, the aromatics 1, 2, 3- and

1, 2, 4-TMB, as well as limonene and α- and β-pinene) were included as variables. In addition, the aforementioned unresolved signal associated with IVOCs (Chapter 3) was included as a variable by integrating the total GC-ITMS ion counts (m/z 50 –425) over a retention time range of 25 to 45 min (retention index range of 1100 to 1700).

Gas-phase ammonia was included as a variable because elevated reduced nitrogen concentrations have been observed in the region and were linked to the use of ammonia on an industrial scale, for example as a floating agent and for hydrotreating (Bytnerowicz et al., 2010).

Total sulfur and total reduced sulfur were added as tracers of upgrader stack SO2 emissions and of "odours", believed to be emitted from oil sands tailings ponds which continue to be of concern in surrounding communities (Small et al., 2015; Percy, 2013; Holowenko et al., 2000).

Refractory black carbon was added as a variable since it is present in diesel truck exhaust and in biomass burning plumes and, hence, is a combustion tracer (Wang et al., 2016; Briggs and

Long). pPAHs were included because of their association with facility stack emissions and combustion particles in the area (Allen, 2008; Grimmer et al., 1987). Hydrocarbon-like organic aerosol (HOA) was included as a surrogate for fossil fuel combustion by vehicles (Jimenez et al.,

2009). The LO-OOA factor was included as it is unique to the Alberta oil sands and appears to form rapidly after emission of precursors (Lee et al., 2018). Supermicron aerosol volume

(PM10-1, i.e., the volume of particles between PM10 and PM1) was also included as a tracer of coarse particles from primary sources, which are expected to be dominated by dust emissions.

4.3 PCA analysis with primary variables

The loadings of the optimum solution are presented in Table 4.1. The 10-component solution accounts for a cumulative variance of 95.5%. The communalities for the analysis, i.e., the fraction of total pollutant observations accounted for by the PCA are all greater than 85%, with the lowest communality obtained for the IVOCs (0.86).

In the following, an overview of the observed components is presented. Associations with r>0.7, r>0.3, and r>0.1 are referred to as "strong", "weak", and "poor", respectively.

Hypothesized identifications are summarized in Table 4.2 and Figure 4.1.

The component accounting for most of the variance of the data, component 1, is strongly associated with the anthropogenic VOCs (r > 0.87), weakly associated with CH4 (r = 0.59), TRS

(r = 0.59), HOA (r = 0.40), LO-OOA (r = 0.45), CO (r = 0.41), and the IVOCs (r = 0.31), and poorly associated with NOy (r = 0.27) and rBC (r = 0.30). Component 2 is strongly associated with the combustion tracers NOy (r = 0.82), rBC (r = 0.77), HOA (r = 0.74), and pPAH (r =

0.94), weakly associated with CH4 (r = 0.39) and IVOCs (r = 0.39), and poorly associated with ammonia (r = 0.20), CO (r = 0.18) and undecane and decane (r = 0.27 and 0.22, respectively).

Component 3 is strongly associated (r > 0.9) with the biogenic VOCs and weakly associated with

CO2 (r = 0.48) and shows poor negative correlations with NOy (r = -0.26), ammonia (r = -0.24), and SO2 (r = -0.15). Component 4 is strongly associated with SO2 and TS (r = 0.97 and 0.93, respectively) and poorly with NOy (r = 0.21) and LO-OOA (r = 0.28).

Components 1 through 4 emerged regardless of the number of components used to represent the data, whereas the structure of components 5 through 10 only fully emerged in the

10-component solution (Tokarek et al., 2018). Hence, components 6 through 10 are somewhat tentative as many (i.e., 7 – 9) are single variable components and have eigenvalues close to or below unity, i.e., account for less variance than any single variable. As a result, the interpretations of these components are subject to more uncertainty and are more speculative but are presented in the SI of the main manuscript for the sake of completeness and transparency.

For the purpose of this manuscript, this is inconsequential as components 6 – 10 are not associated with IVOCs. 50

Table 4.1 Loadings for the 10-factor, optimal solution (primary variables only). Coefficients with Pearson correlation coefficients r>0.3 are shown in bold font.

Commu- 1 2 3 4 5 6 7 8 9 10 nalities Anthropogenic VOCs o-xylene 0.88 0.08 0.02 0.10 0.14 0.13 0.07 -0.04 0.16 0.32 0.95 1,2,3 - TMB 0.93 0.16 0.07 0.05 0.05 0.11 0.04 -0.02 0.18 -0.01 0.95 1,2,4 - TMB 0.94 0.14 0.01 0.10 0.11 0.08 0.07 -0.03 0.18 0.13 0.98 decane 0.92 0.22 -0.02 0.15 0.23 0.01 0.05 0.04 0.04 0.03 0.97 undecane 0.87 0.27 -0.08 0.23 0.20 -0.06 0.12 0.07 -0.04 -0.10 0.96 Biogenic VOCs α-pinene -0.03 -0.08 0.98 -0.11 0.02 0.04 0.01 -0.08 0.02 0.01 0.98 ß-pinene -0.02 -0.08 0.98 -0.12 0.02 0.03 0.02 -0.07 0.00 0.01 0.98 limonene 0.07 -0.03 0.92 -0.08 0.12 0.24 0.05 -0.11 0.03 -0.05 0.95 Combustion tracers

NOy 0.27 0.82 -0.26 0.21 0.22 -0.04 0.02 0.10 -0.08 0.01 0.92 rBC 0.30 0.77 0.03 0.05 0.44 0.10 0.09 0.13 0.12 -0.10 0.94 CO 0.41 0.18 0.04 0.02 0.09 0.09 0.08 0.06 0.87 -0.01 0.99

CO2 0.09 0.08 0.48 -0.12 -0.03 0.77 0.25 -0.14 0.05 -0.08 0.95 Aerosol species pPAH 0.06 0.94 -0.07 -0.13 -0.11 0.07 0.01 0.13 0.10 0.04 0.95

PM10-1 0.18 0.14 0.08 0.09 0.11 0.17 0.93 -0.03 0.07 0.08 0.98 HOA 0.40 0.74 0.02 0.12 0.25 0.15 0.23 -0.06 0.16 0.09 0.90 LO-OOA 0.45 0.11 0.12 0.28 0.72 0.05 0.25 0.00 0.10 0.04 0.91 Sulfur TS 0.25 0.04 -0.16 0.93 0.08 -0.05 0.07 -0.02 0.01 0.12 1.00

SO2 0.12 0.03 -0.15 0.97 0.02 -0.04 0.03 -0.03 0.01 -0.05 0.99 TRS 0.59 0.04 -0.08 0.11 0.26 -0.04 0.16 0.04 -0.04 0.71 0.96 Other IVOCs 0.31 0.39 0.12 -0.08 0.74 -0.02 -0.02 -0.06 0.02 0.20 0.86

NH3 0.01 0.20 -0.24 -0.05 -0.02 -0.08 -0.03 0.94 0.04 0.02 0.99

CH4 0.59 0.39 0.10 -0.05 0.12 0.59 0.11 0.00 0.17 0.14 0.93 Eigenvalues 5.72 3.32 3.23 2.16 1.64 1.13 1.13 0.99 0.96 0.74 % of variance 25.99 15.08 14.69 9.80 7.46 5.14 5.13 4.51 4.36 3.35 Cumulative 25.99 41.07 55.76 65.56 73.02 78.16 83.30 87.81 92.17 95.52 variance

Table 4.2 Hypothesized identifications of principal components. Component Key observations Possible source(s) Relevant references Enhancements of aromatics, (Simpson et al., 2010; n-alkanes, TRS, NO , rBC, Wet tailings ponds and Small et al., 2015; Percy, 1 y HOA, LO-OOA, CO and associated facilities 2013; Holowenko et al., CH4 2000; Howell et al., 2014) (Wang et al., 2016; Enhancements of NO , rBC, y Mine fleet and Grimmer et al., 1987; 2 pPAH and HOA due to operations Allen, 2008; Briggs and engine exhaust Long, 2016) Enhancements of monoterpenes and CO , 2 Biogenic emission and (Guenther et al., 2012; 3 poor anticorrelation with respiration Helmig et al., 1999) NOy and absence of anthropogenic VOCs Enhancements of SO2 and (Simpson et al., 2010; 4 TS, poor correlation with Upgrader facilities Kindzierski and NOy and LO-OOA Ranganathan, 2006) Surface exposed Enhancements of IVOCs, bitumen and hot-water 5 rBC, LO-OOA, NO , and this work y based bitumen TRS extraction Enhancements of CO and 2 (Johnson et al., 2016; 6 CH , absence of combustion Mine face and soil 4 Rooney et al., 2012) tracers 7 Enhancement of PM10-1 Wind-blown dust (Wang et al., 2015) Fugitive emissions from storage tanks and (Bytnerowicz et al., 2010; 8 Enhancement of ammonia natural soil/plant Whaley et al., 2018) emissions Incomplete 9 Enhancement of CO (Marey et al., 2015) hydrocarbon oxidation Enhancements of TRS and (Small et al., 2015; 10 o-xylene, poor association Composite tailings Warren et al., 2016) with CH4

4.3.1 Extended PCA analysis with added secondary variables

The loadings of the optimum solution that includes primary and secondary variables are shown in Table 4.3. In this 11-component solution, the 10 components originally identified were preserved, though their relative order was changed, with the upgrader component moving from the 4th to 2nd position. There was one new component (#6), which encompassed only secondary

- species, including MO-OOA (r = 0.92), Ox (r = 0.33), NO3 (p) (r = 0.36), PM1 (r = 0.31) and

LO-OOA (r = 0.31).

+ 2- - NH4 (p), SO4 (p), and NO3 (p) are associated with the stack emissions component (#2, with r = 0.84, 0.84 and 0.44, respectively), which also weakly correlated with PM1 (r = 0.44) and Ox (r

= 0.36). The association of secondary variables with the primary components suggests rapid formation of these secondary products on a time scale that is similar to the transit time of the pollutants to the measurement site. PM1 correlated strongly with the major IVOC component

- (component 5, r = 0.80), which also weakly associated with LO-OOA (r=0.66) and NO3 (p) (r =

+ 2- 0.59), as well as NH4 (p) and SO4 (p) (r = 0.32 and 0.33, respectively).

Table 4.3 Loadings for the 11-component solution with the inclusion of variables associated with secondary processes.

Commu- 1 2 3 4 5 6 7 8 9 10 11 nalities Anthropogenic oVOCs-xylene 0.89 0.16 0.04 0.04 0.15 0.00 0.10 0.07 -0.04 0.17 0.24 0.94 1,2,3 - 0.91 0.13 0.10 0.16 0.09 0.07 0.11 0.03 -0.03 0.16 -0.08 0.95 1,2,4TMB - 0.93 0.19 0.02 0.13 0.13 0.05 0.06 0.07 -0.03 0.17 0.06 0.99 decaneTMB 0.89 0.25 0.00 0.22 0.26 0.05 -0.01 0.05 0.01 0.00 0.01 0.98 undecane 0.81 0.35 -0.08 0.27 0.21 0.15 -0.07 0.08 0.04 -0.12 -0.10 0.96 Biogenic VOCs α-pinene 0.00 -0.08 0.98 -0.07 0.05 0.03 0.01 0.01 -0.07 0.02 0.01 0.98 ß-pinene 0.01 -0.08 0.98 -0.08 0.05 0.05 0.01 0.03 -0.06 0.01 0.02 0.98 limonene 0.11 -0.02 0.92 -0.02 0.14 0.09 0.21 0.02 -0.10 0.02 -0.03 0.95 Combustion tracers NOy 0.23 0.20 -0.27 0.82 0.21 -0.06 -0.07 0.03 0.10 -0.10 0.01 0.92 rBC 0.22 0.15 0.05 0.80 0.43 0.15 0.10 0.05 0.09 0.07 0.00 0.95 CO 0.40 0.09 0.08 0.20 0.09 0.22 0.08 0.06 0.03 0.83 -0.02 0.97

CO2 0.12 -0.07 0.50 0.08 -0.03 0.09 0.75 0.28 -0.12 0.03 -0.08 0.95 Aerosol species pPAH 0.06 -0.10 -0.06 0.93 -0.07 -0.06 0.07 0.03 0.15 0.13 -0.05 0.94

PM10-1 0.19 0.16 0.08 0.16 0.13 0.08 0.18 0.91 -0.03 0.05 0.07 0.99 PM1 0.24 0.44 0.00 0.17 0.70 0.31 -0.06 0.11 -0.04 0.07 -0.14 0.90 + NH4 (p) 0.28 0.84 0.02 0.12 0.32 0.22 0.06 0.07 -0.04 0.14 -0.04 0.97 2- SO4 (p) 0.29 0.84 0.03 0.12 0.33 0.19 0.06 0.06 -0.05 0.12 -0.05 0.97 - NO3 (p) 0.30 0.44 0.09 0.23 0.59 0.36 0.08 0.15 -0.13 0.02 0.24 0.92 HOA 0.37 0.18 0.02 0.77 0.25 0.10 0.10 0.18 -0.08 0.13 0.14 0.93 LO-OOA 0.37 0.40 0.12 0.16 0.66 0.31 0.03 0.12 -0.06 0.00 0.27 0.97 MO-OOA 0.10 0.15 0.09 0.00 0.10 0.92 0.05 0.07 0.10 0.16 -0.03 0.95 Sulfur TS 0.27 0.90 -0.20 0.03 0.04 -0.04 -0.09 0.07 0.00 -0.04 0.18 0.98

SO2 0.09 0.96 -0.19 0.02 -0.03 -0.01 -0.08 0.03 -0.02 -0.03 0.00 0.98 TRS 0.65 0.14 -0.10 0.05 0.23 -0.08 -0.07 0.17 0.06 -0.04 0.63 0.95 Other IVOCs 0.34 -0.01 0.12 0.33 0.80 -0.23 -0.02 0.02 0.02 0.06 0.06 0.94

NH3 -0.03 -0.08 -0.22 0.21 -0.04 0.09 -0.07 -0.03 0.93 0.02 0.02 0.99

Ox 0.07 0.36 -0.62 0.01 0.27 0.33 -0.41 -0.07 -0.03 -0.14 0.12 0.91

CH4 0.60 0.00 0.14 0.42 0.10 0.08 0.57 0.08 -0.04 0.13 0.16 0.94 Eigenvalues 5.85 4.30 3.71 3.51 2.78 1.58 1.24 1.09 1.01 0.94 0.75 % of 20.9 15.3 13.3 12.5 9.92 5.65 4.43 3.88 3.59 3.37 2.66 variance Cumulative 20.9 36.2 49.5 62.0 71.9 77.6 82.0 85.9 89.5 92.9 95.5 variance

4.4 Principal component analysis with a divided IVOC peak

Another PCA was conducted to investigate if different volatility bins are associated with the same components as the total IVOC signal (Table 4.1). In other words, I probed if higher volatility IVOCs are coming from a separate source as lower volatility IVOCs. To do this, I divided the IVOCs signal into two regions. The first region (bin 1) was integrated between a retention time (tR) = 25 minutes and the tR of tridecane (C13 in Figure 4.2). The second region

(bin 2) was integrated between the tR of tridecane and a tR = 45 minutes. Tridecane was chosen as the dividing point because it was present and distinguishable in all ambient air chromatograms containing the IVOC peak. These areas replaced the total IVOCs variable in the PCA (23 variables in total). The result of the PCA is shown in Table 4.4. The 10 components in the original solution were preserved. However, the variance of components 2 and 5 (IVOC components, see section 4.7.3) changed (14.0 % and 10.9 % from 15.1 % and 7.46 %) and

“switched” with component 3 and 4, respectively. Further, the non-IVOC components 6 and 7 switched places (to 7 and 6, respectively). The two IVOC bins remained with the same components (1,3 (now 2), and 5(now 4)) as the original IVOC variable, albeit with slightly attenuated correlation due to the added variability. This result shows that the observed IVOCs were associated with the same sources throughout the entire volatility range. Lastly, the correlation with IVOCs and what is now component 4 increased from 0.74 to 0.90 and 0.88 for bins 1 and 2, respectively. This also resulted in a decrease in the correlation of LO-OOA with what is now component 4 from 0.72 to 0.55.

Table 4.4 Loadings for the 10-factor, optimal solution (primary variables only) with the IVOCs divided into 2 bins. Coefficients with Pearson correlation coefficients r>0.3 are shown in bold font.

Commu- 1 2 3 4 5 6 7 8 9 10 nalities Anthropogenic VOCs o-xylene 0.88 0.02 0.06 0.18 0.10 0.07 0.12 -0.03 0.16 0.29 0.95 1,2,3 - TMB 0.93 0.06 0.15 0.10 0.04 0.02 0.12 -0.02 0.17 -0.03 0.94 1,2,4 - TMB 0.93 0.00 0.13 0.16 0.10 0.06 0.08 -0.02 0.18 0.11 0.98 decane 0.91 -0.02 0.21 0.25 0.15 0.07 -0.01 0.04 0.03 0.01 0.97 undecane 0.87 -0.08 0.27 0.17 0.23 0.14 -0.10 0.06 -0.05 -0.10 0.96 Biogenic VOCs α-pinene -0.04 0.97 -0.08 0.04 -0.11 0.00 0.03 -0.07 0.02 0.01 0.98 ß-pinene -0.02 0.97 -0.08 0.04 -0.12 0.01 0.02 -0.07 0.00 0.02 0.98 limonene 0.07 0.93 -0.03 0.13 -0.07 0.07 0.22 -0.11 0.03 -0.06 0.95 Combustion tracers

NOy 0.26 -0.26 0.80 0.28 0.22 0.03 -0.04 0.10 -0.08 0.00 0.92 rBC 0.30 0.04 0.76 0.42 0.07 0.16 0.03 0.12 0.13 -0.13 0.93 CO 0.42 0.05 0.18 0.07 0.02 0.09 0.07 0.05 0.87 -0.01 0.98

CO2 0.10 0.49 0.09 -0.02 -0.12 0.27 0.75 -0.14 0.05 -0.09 0.95 Aerosol species pPAH 0.05 -0.08 0.93 0.02 -0.13 -0.04 0.10 0.13 0.09 0.05 0.94

PM10-1 0.17 0.07 0.13 0.07 0.09 0.90 0.19 -0.03 0.06 0.08 0.93 HOA 0.41 0.03 0.74 0.26 0.12 0.26 0.10 -0.07 0.16 0.08 0.91 LO-OOA 0.48 0.15 0.11 0.55 0.31 0.40 -0.11 -0.02 0.12 -0.01 0.85 Sulfur TS 0.26 -0.16 0.04 0.04 0.94 0.08 -0.05 -0.02 0.00 0.12 1.00

SO2 0.12 -0.15 0.03 -0.03 0.97 0.03 -0.03 -0.03 0.01 -0.04 0.99 TRS 0.60 -0.08 0.04 0.28 0.13 0.20 -0.07 0.03 -0.04 0.67 0.96 Other IVOCs (Bin 1) 0.23 0.08 0.28 0.90 -0.04 -0.01 0.01 0.00 0.06 0.03 0.96 IVOCs (Bin 2) 0.32 0.10 0.22 0.88 -0.03 0.06 0.06 -0.03 -0.02 0.15 0.97

NH3 0.01 -0.24 0.20 -0.02 -0.04 -0.03 -0.08 0.94 0.04 0.01 1.00

CH4 0.60 0.11 0.39 0.14 -0.05 0.15 0.55 0.00 0.16 0.12 0.91 Eigenvalues 5.82 3.26 3.21 2.51 2.18 1.27 1.06 0.99 0.96 0.66 % of variance 25.31 14.16 13.95 10.93 9.47 5.51 4.59 4.31 4.16 2.87 Cumulative 25.31 39.48 53.42 64.35 73.83 79.34 83.94 88.24 92.40 95.28 variance

4.5 Spatial distribution of IVOC sources

Bivariate polar plots were generated for all components and their dominant, associated variables and are shown in the supplemental material section of the main manuscript (Figures

S2-S11) (Tokarek et al., 2018). Winds were predominantly from the SW but were also observed often from the S and N. Figure 4.2A shows the plot for IVOCs. The highest concentrations of

IVOCs were observed when the local wind direction was from the NE, where several facilities including the Aurora North, Musket River and Jackpine mines and large swaths of disturbed and cleared land are located in close proximity to each other (Chapter 3, Figure 3.1). The second highest IVOC signal intensity was observed when local wind direction was from the SSE.

The bivariate polar plots of the 3 components associated with IVOCs are shown in

Figures 4.2B-D. These components are associated with winds from the NE, E, SE and S at low to moderate speeds (1-3 m s-1). Component 5 (Figure 4.2B) was the most strongly correlated with

IVOCs and shows the most spatial overlap with the distribution of the IVOC source; however, the intensities differ owing to the association of component 5 with other variables such rBC and

LO-OOA.

4.6 Discussion

The main objective of this work was to elucidate the origin of the IVOC signature observed at the AMS 13 ground site downwind from the AB oil sands mining operations (Figure

4.1) through a principal component analysis. The optimum PCA solution identified 10 components, of which three were associated with the IVOC signature: 1, 2, and 5 (Table 4.1).

The assignments of these components to source types in the oil sands are given in Table 4.2 and are discussed below.

2010) which have shown substantial emissions of NOy, SO2, CO, VOCs, CO2, and CH4 from these facilities and associated mining activities. No single component correlated with all of these variables, suggesting that the PCA is able to distinguish between source types within the facilities such as tailings ponds (component 1), stack emissions (component 4), and mining

(component 2).

Close-up overflights (Howell et al., 2014; Li et al., 2017; Baray et al., 2018) were able to spatially resolve various oil sands facility emission sources (i.e., tailings ponds from upgraders, fluid coking reactors, hydrocrackers and –treaters); the PCA presented in this Chapter is not expected to do this in all cases because some emissions would have frequently merged into a single plume by the time of observation at AMS 13; unless their emissions vary considerably in

time, these sources could be interpreted as originating from a single source in the PCA.

The discussion below focuses on components that are associated with IVOCs (section 4.7), followed by those that are not (section 4.8). The PCA analysis that included 6 secondary products is discussed in section 4.9. Components which are not associated with IVOCs and have only tentatively been identified (i.e., components 6 – 10) are discussed in the SI of

Tokarek et al. (2018).

4.7 Sources associated with IVOCs

4.7.1 Component 1: tailings ponds (wet tailings)

Component 1 is strongly associated with AVOCs (r > 0.87) and weakly with TRS (r =

0.59), and CH4 (r = 0.59). These pollutants originate from tailings ponds (Small et al., 2015), though it is unclear from this analysis how large a source tailings ponds are compared to fugitive emissions of these pollutants from the nearby processing (e.g., bitumen separation and mining) facilities.

Tailings ponds cover large areas of land and are used to slowly (on a time scale of years to decades) separate solid components, or tailings, from water used in bitumen extraction.

Residual bitumen often floats to the top of the settling basins. Most tailings ponds are "wet" (as they contain residual naphtha that is used as a diluent during the transfer of tailings to the ponds) and emit VOCs, CH4, and CO2 (Small et al., 2015). The presence of o-xylene, TMBs and the n- alkanes in component 1 is consistent with the fugitive release of VOCs from residual naphtha, which contains these compounds (Siddique et al., 2008; Siddique et al., 2011; Small et al., 2015).

Furthermore, the observation of TRS and CH4 from this source is consistent with the presence of anaerobic sulfur reducing bacteria and methanogens within the ponds, which degrade not only the residual bitumen (Holowenko et al., 2000; Percy, 2013; Quagraine et al., 2005) but also the various components of naphtha (Shahimin and Siddique, 2017; Small et al., 2015). Overall, tailings ponds emissions explain much of the TRS and CH4 concentration variability in this data set (Table 4.2) and in a recent aircraft study (Baray et al., 2018).

While component 1 correlates with CH4 (r = 0.59), it does not correlate with CO2 (r =

0.09). Emissions of CH4 from tailings ponds due to methanogenic bacterial activity are well- documented (Small et al., 2015; Yeh et al., 2010) and hence the correlation with CH4 is not unexpected. On the other hand, the lack of correlation with CO2 seems inconsistent with emission inventories that generally present tailings ponds as large CO2 sources (Small et al.,

2015). One plausible explanation is that tailings ponds are a relatively small CO2 source overall in the region and that other, larger CO2 sources and sinks (such as photosynthesis and respiration by the vegetation surrounding the site) dominate the variance impacting the PCA results. It may also indicate that, at least on aggregate and for the particular ponds detected in this work, the emissions are in a regime where the release of CH4 dominates over CO2, i.e., the ponds have, perhaps, become more anoxic than believed to be the case in previous studies and hence emit more CH4 (Holowenko et al., 2000). For example, Small et al. (2015) showed that older tailings ponds (those without the addition of fresh froth or thickening treatments) tended to emit more

CH4, while newer ponds are associated with higher VOC emissions. It is likely that component 1 is dominated by the nearest pond (the Mildred Lake settling basin, 6 – 11 km SSE of AMS 13) and other tailings in the SE where the majority of air samples originated from. The Mildred Lake 63

settling basin is one of the oldest in the region and is still actively being used; the correlation with CH4 and VOC emissions is hence expected.

Component 1 is also associated with NOy, rBC, CO, and HOA, though these correlations are relatively modest (r = 0.27, 0.30, 0.41, and 0.40, respectively). These species typically originate from combustion sources, such as generators, motor vehicles, including diesel powered engines powering generators or pumps; it is not obvious if and to what extent these are operated on or near tailings ponds, though. Satellite observations have shown elevated concentrations of

NO2 above on-site upgrader facilities, likely a result of emissions from extraction and transport sources (McLinden et al., 2012). In addition, one of the major highways of the region is located adjacent to the Mildred Lake settling basin and other major ponds in the region; highway traffic emissions (of CO, NOy, rBC, and HOA) may hence also be partially included in component 1.

The bivariate polar plot shows that component 1 was observed when local wind speeds were from the SE and E of the measurement site (Figure 4.2C), which is consistent with the notion that the Mildred Lake settling basin and emissions along Highway 63 and, potentially, more distant facilities are sources contributing to this component.

Component 1 is associated with the IVOC signature, though to a lesser degree than components 2 and 5. The association of the IVOC signal with component 1 is slightly poorer (r =

0.31) than the association with component 2 (r = 0.39), but significantly poorer than component

5 (r = 0.74). The association of IVOCs with tailings ponds vapor can be explained by the presence of bitumen in the ponds that was not separated from the sand during the separation 64

stage (Holowenko et al., 2000). Tailings ponds contain anywhere from 0.5% - 5% residual bitumen by weight (Chalaturnyk et al., 2002; Holowenko et al., 2000; Penner and Foght, 2010).

As illustrated in Figure 4.1A, some of this material floats on the ponds' surfaces, where

IVOCs can partition to the air. Emission of IVOCs from bitumen floating on tailings ponds would be a function of many variables (e.g., diluent composition, extraction methodology, settling rate, temperature, etc.) and is thus not expected to be as persistent as CH4 partitioning from the ponds to the above air or from exposed bitumen on the mine surface, leading to a lower overall correlation.

Component 1 is also weakly associated with the less oxidized oxygenated organic aerosol factor, LO-OOA (r = 0.45). Liggio et al. (2016) found that the observed secondary organic aerosol is dominated by an OOA factor whose mass spectrum was similar to those of aerosols formed from oxidized bitumen vapours. The organic aerosol budget in this study was also dominated by an OOA factor, the LO-OOA (Lee et al., 2018). The association of LO-OOA with component 1 is thus consistent with its association with IVOCs.

4.7.2 Component 2: mine fleet and vehicle emissions

Component 2 strongly correlates with NOy (r = 0.82), rBC (r = 0.77), pPAH (r = 0.94), and HOA (r = 0.74), which suggests a combustion source such as diesel engines. In the AB oil sands, there is a sizeable off-road mining truck fleet consisting of heavy aggregate haulers. In addition, there are diesel engine sources associated with generators, pumps and land moving

equipment, i.e., graders, dozers, hydraulic excavators, and electric rope shovels (Watson et al.,

2013; Wang et al., 2016). Most of these non-road applications have been exempt from highway fuel taxes, on-road fuel formulation requirements and after-engine exhaust treatment (Watson et al., 2013). Emissions from the hauler fleet and the stationary sources would fit the profile of component 2. Other diesel engines operated in the region include a commuter bus fleet, pickup and delivery trucks, tractor-trailers, and privately owned, commonly diesel-powered automobiles used to commute from the work sites to the major residential areas around Fort McMurray, whose emissions are likely captured by component 2 as well, though the magnitude of these relative to the mining truck fleet is not known. Consistent with component 2 being associated with an anthropogenic source is its poor correlation with undecane (r =0.27), likely arising from fugitive fuel emissions.

The bivariate polar plot (Figure 4.2D) for component 2 and NOy in particular (Figure S-

4A in Tokarek et al. (2018)) match the location of Highway 63 which crosses the river to the SE of AMS 13 and bends to the E and is indicative of a line source. At the same time, some of the largest mining operations in the region, the Susan Lake Gravel Pit, Aurora North, Muskeg river, and Millennium mines are located to the NE and SE of AMS 13 as well. NOy, rBC, and HOA

(Figures S-4A, B and D in Tokarek et al. (2018)) all appear to have dominating point sources to the S and E when wind speeds are 1-2 m s-1. These directions are the same as the Fort McKay industrial park to the E and the Syncrude Mildred Lake facility parking lot to the S which would have a higher concentration of vehicles emitting these pollutants in a smaller area, whose emissions would be in addition to those from industrial activities.

Component 2 is associated with the IVOCs signature and CH4 (both r = 0.39). The mining activities bring bitumen to the surface; similar to what we had observed in lab experiments (Figure 4.2, black trace), the surface exposure of bitumen during mining and on-site processing is expected to be associated with fugitive emissions of CH4 (Johnson et al., 2016) and

IVOCs.

Fine-fraction particle-surface bound PAHs (pPAH) are associated strongly with component 2, but no other components. Measurements of individual PAHs in snow and moss downwind from the oil sands facilities have identified multiple sources of PAHs in the

Athabasca oil sands, which include wind-blown petroleum coke dust (also referred to as petcoke for short), a carbonaceous residual product from the upgrading of crude petroleum that is stockpiled on mine sites, and emissions from fine tailings, oil sands ore, and naturally exposed bitumen (Zhang et al., 2016; Jautzy et al., 2015; Parajulee and Wania, 2014). Given this diversity of known sources, the associations of PAHs with only a single component is surprising, though indicates that emissions from the mining fleet (which would include diesel and, perhaps, wind- blown emissions from petcoke that is being transported) gave rise to most of the variability in surface-bound PAH concentrations in this data set. The petcoke emissions identified in the studies mentioned above are likely mainly associated with larger, supermicron sized particles, whose PAH content would not be detected by the pPAH measurement in this data set.

Component 2 is not significantly associated with LO-OOA (r = 0.11), even though

IVOCs are associated with this component. This feature may indicate that the IVOCs emitted in component 2 are qualitatively different from those emitted by components 1 and 5, in that they 67

are less likely to yield organic aerosol on the time scale of transport from emission to observation. One reason for the difference could be that the bitumen that is transported by the mining fleet is relatively freshly exposed, whereas the IVOCs released by bitumen in tailings ponds has been processed by microbes and that released by mine faces (component 5) may have been photochemically oxidized to a greater extent and hence more prone to rapid aerosol formation.

There is little to no association of component 2 with CO2 (r = 0.08). This is somewhat unexpected as the trucks are expected to release CO2 (Wang et al., 2016) but could be due to significantly larger CO2 sources in the area dominating the observed CO2 variability at AMS 13

(e.g., components 3 and 6). Furthermore, one would expect an association of non-road mining truck emissions with aromatics and alkanes. Component 2 exhibited only poor correlations with decane (r = 0.22) and undecane (r = 0.27) and negligible correlation with o-xylene (r = 0.08), suggesting that other components (i.e., component 1) explained most of the variability of their concentrations at this site.

4.7.3 Component 5: surface-exposed bitumen and hot-water bitumen extraction

Component 5 correlates more strongly with the IVOCs (r = 0.74) than with any other component and correlates strongly with LO-OOA (r = 0.72), weakly with rBC (r = 0.44), and poorly with HOA (r = 0.25), NOy (r = 0.22), decane (r = 0.23), undecane (r = 0.20), and TRS (r =

0.26). We interpret this profile as emissions from surface-exposed bitumen which outgases

IVOCs.

One possibility is that these emissions occur on mine faces, where previously unexposed bitumen is brought to the surface as a result of mining. Only a relatively small portion of the mine faces is actively mined; those parts give rise to rBC and NOy emissions from combustion engines in heavy haulers or generators powering equipment. The poor association of component

5 with TRS could be due to sulfur reducing bacteria found on the surface of bitumen. However, most of the variability of TRS at AMS 13 is attributed to composite or “dry” tailings ponds given their more conducive environment to microbial activity.

Component 5 does not correlate with CO2 (r = -0.03) and only poorly with CH4 (r =

0.12), which is somewhat at odds with the notion of mine faces as the main source of IVOCs.

The mine faces give rise to substantial fugitive emissions of CO2 and CH4 (Johnson et al., 2016)

– these emissions are likely captured by component 6 in this analysis (see SI for

Tokarek et al. (2018)). It is unclear to what extent these greenhouse gases are released relatively quickly from "hot spots" (i.e., from a small number of locations) through surface cracks and fissures or by slow release from new material that is exposed and then releases greenhouse gases during material handling, transport and processing (Johnson et al., 2016). IVOCs from surface- exposed bitumen are likely released by the latter mechanism and are temperature-dependent. If the mine faces are indeed the main IVOC source, the analysis results presented here suggest that the IVOCs emissions from surface-exposed bitumen on mine faces are decoupled from CH4 emissions in time and appear as a distinct component and hence corroborate the "hot spots" or fast release hypothesis, though clearly, more work is needed to characterize greenhouse gas emissions from oil sands mine faces. 69

The association of IVOCs with component 5 may also be a result of fugitive emissions during the hot water-based extraction of bitumen sand slurries during the separation phase of bitumen treatment. Generally, bitumen is extracted in a weak alkaline environment by aeration of the solution to optimize the separation of sand and bitumen (Masliyah et al., 2004). Unrecovered bitumen and naphtha then end up in tailings. The recovered bitumen and naphtha are moved to upgrader facilities where they undergo further treatment (such as coking or hydrotreatment). The magnitude of fugitive emissions during these downstream extraction processes could be large, considering the bitumen is heated and actively aerated. Future work should investigate IVOC fluxes near extraction plants and on mine faces.

Finally, it is conceivable that a "natural" background of IVOCs exists in the region (since bitumen can be found at or near the surface in many parts of the region); such a natural background would also be included in component 5. However, this "natural" bitumen would have been exposed at the surface for geological time scales and, unlike unexposed, buried bitumen, likely would have lost most of its volatile content over that period. Furthermore, the mine faces occupy large swaths of land in the region (as evident from satellite imagery). Thus, the IVOCs emissions are more likely due to anthropogenic activity than due to a natural phenomenon.

4.8 Sources not associated with IVOCs

4.8.1 Component 3: biogenic emissions and respiration

Component 3 is strongly correlated with the monoterpenes α-pinene (r = 0.98), ß-pinene

(r = 0.98) and limonene (r = 0.92) and is hence identified as a biogenic emissions source. This component is also weakly associated with CO2 (r = 0.48).

At AMS 13, CO2 and the monoterpenes exhibit a very similar diurnal cycle: they are present in higher concentrations during the night than during the day (Figure 4.3) due to a decrease in the boundary layer height (BLH) at night coupled with plant respiration of CO2 and non-photochemical emission of monoterpenes (Fares et al., 2013; Guenther et al., 2012). During the day, mixing ratios of CO2 are lower due to plant uptake and photosynthesis, and mixing ratios of terpenes are lower due to higher mixing heights and vertical entrainment and due to oxidation by O3 and HO (Fuentes et al., 1996). Hence, the PCA gives a positive correlation of monoterpenes with CO2 even though the physical processes, photosynthesis and respiration, work in opposite direction.

The bivariate polar plots (Figures S-5A-C in the SI of Tokarek et al. (2018)) show that the monoterpenes and CO2 were observed in highest concentrations when the wind speeds were low (< 1 m s-1), consistent with formation of a stable nocturnal boundary layer.

To corroborate this interpretation, the PCA was repeated with BLH estimated by a light detection and ranging (LIDAR) instrument (Strawbridge et al., in prep.) added as a variable

(Table S-9 in the SI of Tokarek et al. (2018)). Since BLH is not "emitted" by any source, it appears as a single variable component (r = 0.90). The only other component that BLH

(anti)correlates with is the biogenic component 3 (r = -0.35).

The dominant monoterpene species observed was α-pinene, followed by ß-pinene and limonene, though occasionally there was twice as much ß-pinene than α-pinene in the sampled air. Some variability of this ratio is expected since emission factors vary considerably between tree species (Geron et al., 2000) which are not homogeneously distributed throughout the region

(e.g., Figure S1 of Rooney et al. (2012)).

Simpson et al. (2010) observed enhancements of α-pinene and, to a greater extent, β- pinene over the oil sands (up to 217 pptv and 610 pptv) compared to background levels of 20±7 and 84±24 pptv, respectively, during mid-day overflights (which occurred between 11:00 and

13:00 local time). Similar enhancements were also reported by Li et al. (2017) who observed emissions of biogenic hydrocarbons in the four facilities sampled, three of which showed a higher β- than α-pinene concentration. The PCA analysis (Table 4.1) showed no significant correlation of α- and β-pinene with any of the anthropogenic sources, which implies that the biogenic source strength is simply too large for any anthropogenic emissions of terpenes to be picked up in the analysis, especially considering that terpenes are relatively short-lived.

The biogenic source shows poor anticorrelations with NOy (r = -0.26), NH3 (r = -0.24), and SO2 (r = -0.15). Many NOy species (i.e., NO2, nitrous acid (HONO), PAN, and nitric acid

(HNO3)) and SO2 deposit to the forest canopy (Hsu et al., 2016; Min et al., 2014; Fenn et al., 72

2015); at night, when mixing heights are lower, their concentrations are expected to decrease faster than during the day and are thus out of phase with the CO2 and terpene concentrations. In addition, there is a time-of-day observation bias for SO2 and, to lesser extent, NOy, which are found in upgrader plumes (section 4.8.2). The poor anticorrelation with NH3 likely arises because the NH3 emissions from plants are mainly stomatal and scale with temperature and are hence larger during the day than at night, anticorrelated with the terpene source (Whaley et al.,

2018).

4.8.2 Component 4: Upgrader emissions

Component 4 is strongly correlated with SO2 (r = 0.97) and total sulfur (r = 0.93). By far

7 the largest source of SO2 in the region are upgrader facilities, which emit as much as 610 kg annually according to emission inventories (ECCC, 2013). Significant SO2 emissions from upgrader facilities have recently been confirmed by aircraft studies (Simpson et al., 2010;

Howell et al., 2014; Liggio et al., 2016). Component 4 is also poorly correlated with NOy (r =

0.21) but not with rBC (r = 0.05), consistent with a non-sooty (i.e., lean) combustion source such as upgrader stacks. Strong enhancements in SO2 were only observed intermittently as "spikes", which is expected when sampling emissions from relatively few and discrete point sources.

Component 4 is poorly anticorrelated with CO2 (r = -0.12), even though inventories indicate that the upgrading facilities are the largest CO2 source in the region (Furimsky, 2003; Englander et al., 2013; Yeh et al., 2010). In this data set, the lack of correlation of component 4 with CO2 (and to some extent with PM10-1 as well) likely arises mainly from a sampling bias as stack emissions were only observed during daytime, likely due to diurnal variability of the atmospheric boundary

layer structure as explained below.

Most of the variability in CO2 concentration at AMS 13 is due to surface-based sources that originate from large areas, especially biogenic processes (photosynthesis during the day and respiration at night, component 3) and anthropogenic surface sources such as those captured by component 6 (see SI of Tokarek et al. (2018)). Other anthropogenic pollutants, such as SO2,

NOy, and CH4, are not subject to large biogenically driven processes and are less affected than

CO2.

In contrast to surface sources, emissions from the > 100 m tall stacks are comparatively under sampled and observed mainly during daytime, when vertical mixing brings elevated plumes to the surface, yet CO2 concentrations are generally much lower than during the night due to uptake by vegetation. At night, pollutants emitted from stacks are injected above the likely very shallow nocturnal surface layer and were hence not observed at the surface. Vertical profile measurements of SO2 stack plumes by a Pandora spectral sun photometer at Fort McKay during daytime have shown considerable vertical gradients and only occasional transport of SO2 all the way to the surface (Fioletov et al., 2016).

The association of component 4 with CO2 is negative because the stack emission source is observed only during the day when the large biogenic sink dominates and effectively masks the relatively small increase due to anthropogenic CO2. In contrast, background concentrations of

SO2 are comparatively low, and the increase in SO2 concentrations is readily picked up by the

PCA. 74

It would be interesting to conduct a future study in winter when biogenic activities decrease; a wintertime PCA analysis of surface measurements might be able to associate CO2 enhancements with upgraders, though boundary layer mixing heights would decrease as well, which would make a PCA analysis using surface data even more challenging.

Component 4 does not correlate with PM10-1 volume (r = 0.09). It is clear that the emitted

SO2 will contribute to secondary aerosol formation downwind, such that a correlation of stack emissions with PM10-1 volume might be expected. However, these secondary contributions will likely mostly be in the submicron aerosol fraction, which adds relatively little to PM10-1 volume.

Further, PM10-1 volume is dominated by coarse particles from other primary sources, mostly wind-blown emission of sand from the mine surfaces, roadways and, perhaps, bioaerosol

(component 7, see SI of Tokarek et al. (2018)). These effects make PM10-1 volume from stacks appear comparatively small, such that the variability of the larger, surface-based sources likely masks the contribution of stacks emissions to PM10-1 variability.

The bivariate polar plot of component 4 (Figures S-6D in the SI of Tokarek et al. (2018)) shows that the largest magnitudes were observed when local winds were from the SE. The corresponding plot of SO2 (Figure S-6A in the SI of Tokarek et al. (2018)) reveals two more distinct sources: a larger one from the E and a smaller one from the SSE. However, only two facilities (Sunrise and Firebag) are located to the E at relatively large distances of 37 km and 47 km respectively. The largest known upgraders and SO2 sources in the area (i.e., upgraders located at the Mildred Lake and Suncor base plants) are located to the S and SE of AMS 13.

Considering that the stack emissions are only observed intermittently, we speculate that there 75

exists a mesoscale transport pattern in the Athabasca river valley which channel emissions, such that the local wind direction and speed may be misleading as to the true location of these sources.

For more extensive data sets, such phenomena may very well average out but perhaps did not in this case.

4.9 Extended PCA with added secondary variables

The extended analysis (Table 4.3) qualitatively preserves the structure (with the exception of an added “Aged” component, # 6) of the original 10-component solution but allows an assessment of which components most result in formation of secondary products such as SOA, which has implications for health (Bernstein, 2004 ) and climate (Charlson et al., 1992).

Secondary products vary considerably as a function of air mass chemical age (which depends, amongst other components, on time of day and synoptic conditions, including wind speed) and are hence expected to add considerable noise and scatter to the results leading to lower correlations. On the other hand, the distance between the measurement site and sources is fixed, such that this variability should average out over time. This indeed appears to have happened in this data set in spite of the relatively low sample size.

The analysis indicates that the component with the strongest IVOC source (Component 5) also has the highest association with PM1 (r = 0.7; Table 4.1). Aircraft measurements combined with a modelling study have required a group of IVOC hydrocarbons to explain the significant

SOA formation and growth downwind of the oil sands region (Liggio et al., 2016). The

association of IVOCs with PM1 volume suggests a link to fine particulate matter formation, for example, from the oxidation of the IVOCs observed at AMS 13.

The second component influencing PM1 is that from stack emissions (Component 4 in the primary PCA; Component 2 in the secondary PCA) (Tables 4.1 and 4.3). It is well established that the oxidation of SO2 to sulfate will lead to formation of fine particulate matter. This apparently occurs, at least partially, on the time scale between the point of emission and the

AMS 13 site (assuming a wind speed of 3 m/s and a distance of 11 km, the transit time is 1

2- hour), though some fraction of SO4 (p) is likely directly emitted.

4.10 Conclusions

In this Chapter a PCA was applied to continuous measurements of 22 primary pollutant tracers at the AMS 13 ground site in the Athabasca oil sands during the 2013 JOSM intensive study to elucidate the origins of airborne analytically unresolved hydrocarbons that were observed by GC-ITMS. The analysis identified 10 components. Three components correlated with the IVOC signature and were assigned to mine faces and, potentially, hot-water bitumen extraction facilities, the mine hauler fleet, and wet tailings ponds emissions. All three are anthropogenic activities that involve the handling of raw bitumen, i.e., the unearthing, mining and transport of crude bitumen, and the disposal of processed material that contains residual bitumen in wet tailings ponds. The PCA results are consistent with our previous interpretation that the unresolved hydrocarbons originate from bitumen, which was based on the similarity of

the chromatograms with those obtained in a head space vapor analysis of ground-up bitumen in the laboratory.

Liggio et al. (2016) showed that these hydrocarbons constitute a group of IVOCs in the saturation vapor concentration (C*) range 105 µg m-3< C* < 107 µg m-3 that contribute significantly to secondary organic aerosol formation and growth downwind of the oil sands facilities. The correlation of LO-OOA with two of the three IVOC components in the main PCA analysis and with PM1 in the extended analysis corroborates the high SOA formation potential of

IVOCs and suggests that further differentiation may be needed and stresses the need for IVOCs to be routinely monitored. In particular, direct measurements of emissions throughout the processing of raw bitumen are needed to pinpoint source contributions more accurately and aid in the development of potential mitigation strategies.

The PCA analysis in this study suffered from several limitations. For instance, PCA does not provide insight into emission factors of individual facilities, though it does capture what conditions change ambient concentrations the most. Further, the receptor nature of PCA did not always discern between large source areas that may have many individual point sources coming together at the point of observation. For example, component 1 contains an obvious tailings pond signature because of its high correlations with anthropogenic VOCs, methane and TRS, but also includes several combustion sources, making interpretation of this IVOC source location more challenging. A longer continuous data set with a greater number of variables would have perhaps been able to resolve these different sources, including the various tailings ponds, of which there are 19 in the region, all with slightly different emission profiles (Small et al., 2015) . 78

Another limitation is the bias of this (and most) ground site data set towards surface-based emissions and the under sampling of stack emissions. Facility stacks were only observed in the daytime because at night the mixing height is so low that the stacks are emitting directly into the residual layer. These emissions could be quantified using aircraft-based platforms (Howell et al.,

2014; Li et al., 2017; Baray et al., 2018). The PCA struggled most with the allocation of greenhouse gases. Mixing ratios of CO2, in particular, were difficult to reconcile in this analysis due to a high background and large attenuation by biogenic activity and boundary layer meteorology. Forests greatly affected CO2 levels in the region because it is taken up during the day when plants are photosynthetically active and emitted at night when plants undergo cellular respiration. This CO2 source and sink appears to dominate the PCA, effectively masking relatively small emissions from tailings ponds, facilities, and tail pipes in particular from the mine hauling fleet.

Finally, there is a need for improved monitoring of IVOCs. For instance, future studies should focus on characterizing the VOCs in the above mentioned volatility range using a greater mass and time resolution instrument, such as a time-of-flight mass spectrometer (TOF-MS) or higher resolution separation methods (e.g., multi-dimensional gas chromatography), and also include measurement of speciated aerosol organic composition by, for example, thermal desorption aerosol GC (TAG) analysis (Williams et al., 2006). Future studies should also investigate how IVOC volatility distributions vary with source type and chemical age.

Chapter Five: Bitumen characteristics observed in the lab and field

Parts of Chapter 5 have been published by Atmospheric Chemistry and Physics, doi: 10.5194/acp-2017-1026 (Tokarek et al., 2018).

5.1 Introduction

Bitumen is a substance that has a long history as a raw material. In Canada, first nations would use the hydrophobic properties of bitumen to waterproof their canoes (Mair and

MacFarlane, 1908). In modern society, petroleum is extracted from mined bitumen for energy and chemical feedstocks. In the 20th century, scientists began researching the chemical composition of bitumen and discovered that it contains aromatic ring structures and that it is high in carbon and hydrogen with lower quantities of oxygen, sulfur, nitrogen, vanadium, nickel and iron (Glozman and Akhmetova, 1970; Selucky et al., 1977; Mossman and Nagy, 1996). Later research showed that many of the organic compounds in bitumen are not particularly volatile and will not resolve on analytical chromatography columns concluding that the material is highly complex (Subramanian et al., 1996). Highly volatile gases are also found in bitumen and include:

CO2, CO, hydrogen sulfide (H2S), carbonyl sulfide (COS), carbon disulfide (CS2) and sulfur dioxide (SO2) (Strausz et al., 1977).

Bitumen is most commonly recovered by mining (and Steam Assisted Gravity Drainage

(SAGD), also known as fracking). Large equipment is used to scrape and shovel the bitumen from the mine face for further processing via caustic, hot-water based extraction (Masliyah et al.,

2004). These mine faces are known to emit large quantities of CO2 and CH4 (Johnson et al.,

2016). Further, bitumen extraction experiments have shown bitumen contains many complex organic species with atmospherically relevant volatility (meaning they can be observed in the gas phase) (Pereira et al., 2013; Ajaero et al., 2018). However, little is known about what organic species are emitted as a result of mining, as most of these studies are concerned with the liberation and fate of naphthenic acids and their impact on soil and water contamination (Ahad et al., 2013).

A recent aircraft based study has shown that surface mining facilities report VOC emission lower than what was observed downwind of the facility (Li et al., 2017). The same study concluded that IVOC emissions from an oil sands source is contributing significantly to the formation and growth of SOA (Liggio et al., 2016). More research was needed to characterize the gas phase emissions of raw, mined bitumen to better understand the role these emissions play on the chemistry in the region.

5.2 Bitumen characteristics by electron impact fragmentation

A sample of bitumen was collected from the Athabasca oil sands region and stored in a freezer until the analysis. The sample was ground up by mortar and pestle and poured into the bottom of a 12 L glass bulb. After the sample vapours were allowed to equilibrate (2-3 hours), nitrogen was used to carry the headspace sample to the inlet of the GC-ITMS. The CIT mass analyzer monitored m/z 35- 425, details of the operating temperature can be found in the SI of

Tokarek et al. A total ion chromatogram of the sampled bitumen vapours is shown in Chapter 3 in Figure 3.2.

The chromatogram is dominated in the region between 25 minutes and 45 minutes. This

* region is characterized by IVOCs ranging in volatility from C11 – C17 with C ranging from

105 µg m-3 < C* < 107 µg m-3 (Chapter 3). The region is comprised of many unresolved peaks and resulted in the inability to quantify individual molecules in this region.

Major observed ions were m/z 55, 57, 67, 69, 82, 84, 96, 98, 110. Unfortunately, acylium and carbenium ions are isobars of one another and the CIT lacks the mass resolution to distinguish between these fragments.

Ambient air was sampled hourly as a 10 minute, preconcentration (~10 min average) at

AMS13 in the Athabasca oil sands region. An example of a total ion chromatogram of a polluted sample is shown in Chapter 3 in Figure 3.2.

The TIC shows high intensity (i.e., greater than 25,000 counts) peaks from 3.5 minutes to

43 minutes. This includes the presence of several species confirmed in the lab using analytical standards including toluene, o-xylene, m- and p- xylene, ethylbenzene and several monoterpene species (α-pinene, ß-pinene, limonene, 3-carene, camphene) which were most likely a component of the background air as confirmed by their presence in unpolluted samples.

The data matrix of both the bitumen sample and ambient air sample show high intensities of ion fragments at unique masses, these are summarized in Table 5.1. In order to define a time range and to focus only on relevant fragment ions, a threshold value was arbitrarily chosen to identify only the top 9 ion fragments observed in the bitumen sample. For bitumen, the threshold was 1,000,000 counts and for ambient air the threshold was 10,000 counts. The relatively high counts for the bitumen sample are due to the concentrated effect of the headspace analysis in the laboratory.

The data show that many of the same prominent ion fragments appear in both the bitumen sample and an ambient air sample. In particular, ion fragments associated with alkanes

(m/z 55, 57, 67, 69) have the highest intensities, suggesting that bitumen, and the IVOCs sampled in ambient air contain primarily alkane substituents. The minimal ion fragments detected at m/z

91 and 115 suggest few aromatic species present, as these are common markers for aromatic substituents in electron impact fragmentation (Cross et al., 2013).

Time range (min)

m/z Bitumen Ambient

55 24-40 25-39

57 26-37 *

67 24-38 *

69 24-39 25-38

82 24-40 25-40

84 25-39 *

96 24-40 25-40

98 27-38 *

110 25-38 *

* Present at high intensity but below threshold value

Many of the ion fragments that are observed at maximum intensity occur in both samples

(Figure 5.1). Figure 5.1 also shows that ion fragments typical of aromatic species (i.e., m/z 91 and 115) were absent in both samples suggesting minimal aromatic contribution to the composition of the air samples. Ion fragments appeared in groupings of approximately ±15 (i.e.,

+ + were associated with the loss of a carbenium ion (CH3 ) or an acylium ion (HCO )). An example of this is shown in Figure 5.1 where we observe high intensity at m/z 110, 95 and 80 (and surrounding fragments).

Figure 5.1 Side by side comparison of an ambient air GC-ITMS data matrix (left) and a bitumen sampled in the lab (right). Darker pixels represent a greater intensity at a given m/z and time.

The absence of high intensity ion fragments typical of aliphatic oxonium ions (m/z 17 +

14n, e.g., 59, 73, 87, etc.), immonium ions (m/z 16 + 14n, e.g., 58, 72, 86, etc.) and sulfonium ions (m/z 33 + 14n, e.g., 61, 75, 89, etc.) where n is the number of carbon, in both the bitumen and ambient samples suggests minimal contributions from other elements (Gross, 2011). This is consistent with research that has shown bitumen from the Athabasca oil sands region is relatively low in oxygen, nitrogen and sulfur (wt% = 1.81 %, 1.63 % and 4.78 %, respectively) (Yoon et al., 2009). However, these species may be present and are either more volatile and have already off-gassed or are less volatile such that they do not appreciate in ambient air at significant concentrations. More research is needed into the presence of heteroatoms in bitumen vapours.

5.3 Conclusions

The ambient air samples observed at AMS13 in August 2013 show consistent ion fragments compared with those from a ground up bitumen sample taken from the region and analyzed in the lab. Due to many co-eluting peaks, quantification of species in the intermediate volatility region (25 minutes to 45 minutes) was not possible, though bitumen characteristics were assessed by investigating particular ion fragments. This particular sample appears to be comprised primarily of alkanes and did not emit volatile aromatic species. Oxygen, nitrogen and sulfur were also explored via typical ion fragments for aliphatic species. Very little oxygen, nitrogen, and sulfur appear to be present, consistent with research that shows these elements are present in relatively low quantities (Yoon et al., 2009). More research is needed to confirm the absence or presence of heteroatoms in bitumen, perhaps via condensation of volatile gases followed by elemental analysis of the condensate.

Chapter Six: Exploring Nocturnal Ozone Depletion Events in the Marine Boundary Layer

during ORCA 2015

Parts of this Chapter have been published by Atmosphere-Oceans, doi: 10.1080/07055900.2017.1306687 (Tokarek et al., 2017).

6.1 Introduction

Processes controlling the abundances of O3 and PM2.5 continue to be of interest because of the association of high O3 and PM2.5 concentrations with health impacts such as chronic obstructive pulmonary disease, heart disease, and lung cancer (e.g., (Pope et al., 2009; Peng et al., 2013; Lelieveld et al., 2015) and references therein). On the west coast of North America, the budgets of O3 and, indirectly, HO and PM2.5 depend critically on the “background O3” concentration, i.e., the concentration of O3 entering from the Pacific Ocean (Ambrose et al.,

2011; McDonald-Buller et al., 2011). In recent decades, this concentration has slowly been increasing due to long-range transport of continuously rising emissions in Asia (Jaffe et al.,

2003; Parrish et al., 2004; Parrish et al., 2010). On the other hand, the concentration of O3 in the layer immediately above the ocean surface, the so-called marine boundary layer (MBL), is frequently depleted relative to atmospheric layers aloft. The principal O3 depletion mechanisms in the MBL are dry deposition to the sea surface (Gallagher et al., 2001a; Gallagher et al., 2001b;

Ganzeveld et al., 2009) and photochemical pathways, such as the HOx driven O3 destruction in low NOx environments and catalytic cycles involving reactive halogen species (Singh et al.,

1996; Galbally et al., 2000; Lee et al., 2009; Watanabe et al., 2005; Read et al., 2008).

The Amphitrite Point Observatory (APO) was established on the west coast of Vancouver

Island, British Columbia, Canada, in 2010 to serve as a background monitoring site. Recently,

McKendry and co-workers (2014) reported observations of rapid nocturnal O3 depletion events

(ODEs) that have not been fully explained to date. Conventionally, nocturnal loss of surface O3 is interpreted as being due to a combination of dry deposition and chemical titration of NO to

NO2 and by reaction of O3 with unsaturated biogenic hydrocarbons (Neu et al., 1994; Kleinman et al., 1994; Trainer et al., 1987; Logan, 1989). At APO, the ODEs are generally associated with along- or onshore flow, stable boundary layer conditions, and an increase in the concentration of

CO2 but not with anthropogenic pollution tracers, i.e., NOx, CO, and SO2 (McKendry et al.,

2014). Carbon dioxide has sinks (photosynthesis) and sources (respiration and combustion) that are much stronger in the continental compared to the marine environment (which is a net CO2 sink (Gruber et al., 2009)). Hence, elevated concentrations of CO2 trace continentally influenced air at coastal sites (Parrish et al., 2009). The correlation of ODEs with CO2 and the absence of correlations with anthropogenic tracers at APO suggest that the upwind air had been in contact with vegetation where O3 would have been more rapidly deposited than to the ocean surface

(Gallagher et al., 2001a; Wesely and Hicks, 2000). Yet, back trajectories calculated using the

Hybrid Single Particle Lagrangian Integrated Trajectory Model (HYSPLIT) (Draxler and Hess,

1998; Draxler and Rolph, 2011) and the "North American Meso" (NAM) meteorological field at

12 km resolution were inconsistent with this interpretation. The back trajectories showed that

ODEs were characterized by light onshore or alongshore winds and that the air masses generally originated in the MBL, i.e., had a marine origin. Overall, the evidence indicated that the ODEs at

APO are caused by a non-photochemical coastal zone process that extends over a spatial scale of tens of kilometres (McKendry et al., 2014).

One such process could be the natural nocturnal O3 deposition on near-shore forest vegetation coupled with a nocturnal land breeze that mixes continentally influenced air against the predominant westerly synoptic flow. This process is in all likelihood poorly captured by the meteorological fields and, hence, HYSPLIT trajectories. Evidence to corroborate this hypothesis would be an enhancement of BVOC concentrations, in particular of monoterpenes whose emissions have both a light-dependent (photosynthetic) as well as a light-independent fraction which (unlike those of isoprene) continue at night (Fares et al., 2013; Guenther et al., 2012). The tree species in the region are predominantly coniferous and include Western Red Cedar (Thuja plicate), Coastal Douglas Fir (Psudotsuga menziesii ssp. menziesii) and Western Hemlock

(Tsuga mertensiana) (Mason et al., 2015). Monoterpenes, in particular α- and β-pinene, dominate

BVOC emissions from this type of vegetation (Helmig et al., 2013; Copeland et al., 2014;

Drewitt et al., 1998; Sakulyanontvittaya et al., 2008a; Lagalante and Montgomery, 2003; Geron et al., 2000; Burney and Jacobs, 2012).

In this work, mixing ratios of monoterpenes (i.e., α- and β-pinene, limonene, carene, and terpinolene), as well as the β-pinene oxidation product nopinone (C9H14O) were quantified at

APO for the first time to provide further insight into the cause (or causes) of the nocturnal ODEs during summertime. Nopinone is produced from the reaction of O3 with β-pinene in ~50% yield

(Jenkin, 2004) and is thus a direct tracer of O3-monoterpene chemical reactions, though it is also produced (and destroyed) during the daytime from reactions initiated by HO (Barthelmie and 89

Pryor, 1999; Larsen et al., 2001; Jaoui and Kamens, 2003; Kavouras et al., 1999). The data set includes measurements of aerosol size distributions as an indirect probe of O3-monoterpene chemistry which results in the production of non-volatile material and manifests itself by either new particle formation or growth (Griffin et al., 1999; Sakulyanontvittaya et al., 2008b; Ehn et al., 2014). Correlations of ODEs with CO2 and BVOC concentrations are presented. Sources of

BVOCs in the region and implications of the data for the interpretation of ODEs at APO are discussed.

6.2 Methods

6.2.1 Study overview

The APO is located on the west coast of Vancouver Island south of the town of Ucluelet at 48.92° N and 125.54° W (Figure 6.1) less than 100 m from the high tide line of the Pacific

Ocean. It was the location of several recent field campaigns, and the immediate vicinity of the measurement site has been described comprehensively in the associated manuscripts (McKendry et al., 2014; Yakobi-Hancock et al., 2014; Wilson et al., 2015; Mason et al., 2015). For this study, named the "Ozone-depleting Reactions in a Coastal Atmosphere" (ORCA) campaign, instrumentation was housed in two mobile laboratories separated by a distance of approximately

15 m: one was operated by Environment Canada, British Columbia's Ministry of the

Environment, and Metro Vancouver, whose measurement suite (which include a Licor 820 CO2 monitor) has been described elsewhere (McKendry et al., 2014; Yakobi-Hancock et al., 2014;

Wilson et al., 2015; Mason et al., 2015). The other was operated by the University of Calgary,

whose measurements are described below. Meteorological parameters were measured at the

Amphitrite Lighthouse (located halfway between the main field site and the ocean) and by a weather station (Vaisala WXT520) operated at a height above ground of 5.2 m atop the

University of Calgary mobile laboratory. Off-shore meteorological data were collected at a moored buoy (C46206, located at La Perouse Bank 35 km WSW of APO) as described by

Mason et al. (2015). Data were collected from July 8 - 31, 2015.

Figure 6.1 Map of the study region.

6.2.2 Gas-phase measurements

Mixing ratios of BVOCs and nopinone were quantified using the GC-ITMS equipped with a cylindrical ion-trap mass analyzer (Riter et al., 2002; Patterson et al., 2002) and electron impact ionization (see Chapter 2). This instrument sampled from a 6.5 m long and 0.635 o.d. stainless-steel inlet from a height of 4.3 m above ground. A 3 m long section of the inlet was heated to 125 °C to remove interference due to O3 reactions with alkenes during preconcentration

(Pollmann et al., 2005; Hellén et al., 2012). Air samples were preconcentrated for 8 minutes at a flow rate of 213 mL min-1 on a dual sorbent trap containing Tenax TA and Carboxen 1017 held at 40 °C and were desorbed at a temperature of 240 °C onto a 30 m (length)  0.25 mm (inner diameter)  0.25 μm (film thickness) DB-5MS column operated with helium carrier gas (Praxair,

ECD grade) purified using a triple trap (Restek 22464). The GC oven was programmed as follows: hold at 40° C for 3.00 min, heat at 3° C min-1 to 70° C and hold for 2.00 min, and heat at 15° C min-1 to 200 °C and hold for 2.00 min (total 25.67 min). This was followed by a 5.00 min recovery time to allow the oven and preconcentration trap to cool back to 40 °C. Full mass spectra (m/z from 35 to 425 Da) were collected at a rate of ~ 3 Hz. Retention times and indices, quantification ions, and limits of detection of the monoterpenes monitored are summarized in

Table S-1 in the supplemental information of the original work (Tokarek et al., 2017).

The instrument was calibrated using mass flow controllers and a gas cylinder filled with a standard VOC mixture prepared in-house on July 2, 2015, a week prior to the campaign. This working standard was prepared by quantitatively transferring a 50 VOC component calibration mixture (Supelco 49148-U) containing α-pinene, β-pinene and limonene in methanol to an 92

evacuated aluminum cylinder (supplied by Scott-Marrin; internal volume 29.500.33 L) which was subsequently pressurized with nitrogen gas to a pressure of 8500250 kPa. The accuracy of this standard was compound dependent and calculated from the concentration uncertainties provided by Supelco and uncertainty of the dilution volume and was 9.6% on average (range

8.3% to 12.0%). The uncertainties in the amounts delivered from this cylinder are increased further due to uncertainties in the flow delivered by the mass flow controllers to a total of ~11% on average. This error analysis assumes quantitative delivery of the VOCs through the 2-stage cylinder pressure regulator and the all-metal mass flow controller (a reasonable assumption following a several-day-long break-in period). Not included in this uncertainty is the potential loss (or production) of terpenes within the cylinder over the course of this campaign, which are expected to be small in a one month old aluminum cylinder but can be substantial, especially over longer time periods (e.g., (Apel et al., 1999; Jones et al., 2014; Rhoderick and Lin, 2013;

Rhoderick, 2010)).

Between July 16 and 31, 2015, five "full" calibrations, i.e., with ≥ 5 points on the calibration curve measured in random order (including a "blank"), were performed. For the monoterpenes, the slopes of the calibration curves had relative standard deviations in the range of 3.2% to 22.6% (average 8.0%) and intercepts whose value  1 standard error encompassed zero. The response factors were linearly interpolated between calibration points.

The precision of the GC-ITMS was determined from the standard deviation of repeated measurements of a small amount of β-pinene evaporated into a 4,000 L smog chamber to an approximate mixing ratio of ~100 pptv and was 3.4%. 93

It has been reported that monoterpenes can undergo rearrangement reactions during the thermal desorption process following preconcentration on Tenax (Arnts, 2010). In post-campaign experiments, approximately 1.5% of trapped β-pinene was found to isomerize to limonene,

α-pinene, and camphene during desorption on this instrument. Because this process happens during both calibration and ambient measurement of β-pinene, its effect is included in the instrumental response factor. In contrast to β-pinene, no rearrangement products were observed for either limonene or α-pinene.

Assuming that the above uncertainties are independent from each other and neglecting the potential degradation of monoterpenes within the standard gas cylinder, the total uncertainty

(at the 1σ level) of the monoterpene measurements (not including carene) presented is 15%, driven by the systematic uncertainty in the preparation of (and delivering a calibrated gas flow from) the standard gas cylinder (11%), the standard uncertainties in the slopes of the calibration curves (8%), and the measurement precision (3.4%).

Not included in this uncertainty estimate are potential systematic errors arising from the interpolation between calibration points.

The response factors for carene and nopinone were determined prior to and after the campaign relative to that of α-pinene; the uncertainty of their mixing ratios was estimated at

30%.

Other gas-phase measurements in the University of Calgary mobile laboratory included a

UV absorption O3 analyzer (Thermo Scientific 49i), a total odd nitrogen (NOy) chemiluminescence analyzer (Thermo Scientific 42i), and a multi-channel cavity ring-down spectrometer for quantification of NO2 and NOx constructed in-house (Paul and Osthoff, 2010;

Odame-Ankrah, 2015).

6.2.3 Particle phase measurements

Submicron and supermicron size distributions were monitored using a scanning mobility particle sizer (SMPS, TSI 3936L75) operated with butanol and an aerodynamic particle sizer

(APS, TSI 3321). The SMPS sampled through a 1 μm cut-off impactor (0.071 cm internal diameter) via 3/8" o.d. conductive silicon tubing secured next to the GC-ITMS inlet. The APS was operated from the container located on top of the trailer and sampled from a 1.6 m tall ½ o.d. aluminum tube (4.4 m AGL) whose tip was bent into a U-shape to prevent precipitation from entering the inlet. Both instruments were calibrated by the manufacturer but not in the field (e.g., inlet transmission efficiencies were not characterized), such that the aerosol size distributions are semi-quantitative.

6.2.4 Photolysis frequencies

Photolysis frequencies were determined by solar actinic flux spectroradiometry

(Hofzumahaus et al., 1999) using a commercial radiometer with 2π receptor optics and photo diode array (PDA) detector (Metcon; 512 pixels, wavelength range 285 nm - 690 nm) which had been calibrated by the manufacturer. The spectrometer was mounted facing up (zenith view) to

measure the down-dwelling radiation. The spectrometer was inverted periodically to determine the up-dwelling radiation. Photolysis frequencies were calculated using reference spectra and quantum yields from the NASA-JPL evaluation (Sander et al., 2010) and updated ClNO2 cross- sections (Ghosh et al., 2011).

6.2.5 Back trajectory calculations

Twelve-hour back trajectories were calculated using the National Oceanic and

Atmospheric Administration (NOAA) HYSPLIT online model and the NAM meteorological field at 12 km resolution. The trajectory calculations were initiated every hour, starting 1 hr prior to the ODEs and ending 1 hr after the ODEs ended, with heights of 10 m, 50 m, and 100 m above ground.

6.3 Results and analysis

6.3.1 Identification of ozone depletion events

Time series of O3 and CO2, the monoterpenes α- and β-pinene, limonene, and carene as well as the β-pinene oxidation product nopinone, NO2 and odd oxygen (Ox = O3 + NO2) for the period of July 16 to 31, 2015 are shown in Figure 6.2. Time series of meteorological parameters are shown in Figure S-1 in the original work (Tokarek et al., 2017). McKendry et al. (2014) outlined the following criteria to identify ODEs: a drop in O3 mixing ratio from a starting value of no higher than 32 ppbv (the third monthly quartile O3 mixing ratio in July), by more than 15

ppbv, and occurring over a time period of 12 hours or less. Using these criteria for the 16 nights shown, 9 nocturnal ODEs were identified. These are indicated as blue bars in Figure 6.2 and have been summarized in Table 6.1. The high frequency of ODEs confirms that they are indeed a frequent phenomenon at APO.

Table 6.1 Summary of ODEs observed.

start end wind Date wind direction (°) ΔO3 ΔOx ΔCO2 ΔTerp. time time speed

mo/day hr:min hr:min median HYSPLIT m/s ppbv ppbv ppmv ppbv 2015 UTC UTC local NAM12

07/16 09:14 14:00 WNW WNW 1.1 -15.9 - 20.7 0.10

07/18 03:41 14:46 ESE E 0.9 -16.0 -15.2 39.0 0.40

W W 07/19 02:50 15:15 initially, initially, 1.7 -23.3 -19.5 12.0 0.12 then E then E

07/23 00:02 15:15 WNW W 1.7 -19.5 -17.1 15.7 0.10

07/24 03:00 09:02 W W 1.6 -18.1 -17.4 13.8 0.41

W 07/27 01:39 13:59 initially, W 1.7 -17.6 -15.6 15.5 0.32 then E

07/28 01:59 14:01 WNW WNW 2.1 -24.5 -23.8 14.3 0.07

07/30 02:23 10:03 W W 2.5 -16.8 -16.0 51.7 0.64

07/31 07:04 10:09 WNW WNW 2.2 -17.2 -16.8 32.4 0.19

6.3.2 Correlations of ozone depletion events with chemical tracers

The ODEs were accompanied by an increase in CO2 mixing ratios (Figures 6.2A), as had been reported by McKendry et al. (2014). At the beginning of the ODEs, CO2 mixing ratios were usually below 380 ppmv. This abundance is less than what was measured at Trinidad Head,

California in July, 2015 (3961 ppmv) (Dlugokencky et al., 2016). Plausible explanations for the lower than expected CO2 mixing ratios are either a prolonged period of CO2 oceanic uptake in a

decoupled marine boundary layer and/or photosynthetic uptake by vegetation upwind of the measurement location.

Once the ODEs started and O3 mixing ratios decreased, CO2 mixing ratios increased.

Data from all nine ODEs were lumped together in one-hour bins as a function of time of day, with 12:00 (noon) set to the solar zenith angle minimum. The 90th, 75th, 25th, and 10th percentile

th and median (i.e., 50 percentile) O3 and CO2 mixing ratios as functions of local time of day during the ODEs are shown in Figure 6.3A. The median O3 and CO2 mixing ratios are anticorrelated with a linear correlation coefficient of r = -0.92 and slope of (-0.3900.005) ppbv O3 / ppmv CO2.

Figure 6.3 (A) O3 (left hand side) and CO2 (right hand side) mixing ratios as functions of (solar) time of day during ODEs. The dashed lines indicate the 90th and 10th percentiles, whereas the shaded areas encompass the 75th and 25th percentiles. The thick solid lines are the median values. Percentiles were calculated over ½ hour bins. The median non-ODE O3 time series is shown as a blue line. (B) O3 and terpene mixing ratios as functions of (solar) time of day during ODEs calculated over 1-hour bins.

It is well known that forest vegetation represents a sink of O3 and, because of the absence of O3 production, results in a net loss of O3 at night (Turner et al., 1974; Trainer et al., 1987).

The observed anti-correlation can be rationalized by respiration by vegetation, which releases

CO2 into a shallow nocturnal boundary layer; at the same time, O3 deposition continues, though with reduced deposition velocity (Fan et al., 1990; Ro-Poulsen et al., 1998; Zeller and Nikolov,

2000; Monson and Holland, 2001).

Since trees continue to emit monoterpenes at night, the nocturnal release of CO2 from trees is expected to be accompanied by an increase in monoterpene concentration. Accordingly, mixing ratios of monoterpenes were also elevated during ODEs, coinciding with the increase in 100

CO2 (Figures 6.2B, 6.3B and 6.5B). The highest monoterpene mixing ratios were observed during the ODEs of the nights of July 18/19 and 19/20. Carene and nopinone mixing ratios were elevated on those nights (Figure 6.2B). The HYSPLIT back trajectories showed that the air masses sampled on those nights originated from the east (Table 6.1), i.e., had travelled over land covered by dense forests, explaining the higher than average monoterpene mixing ratios and the presence of the -pinene oxidation product nopinone.

The linear correlation coefficient of CO2 and total monoterpene concentrations during

ODEs was r = 0.93, whereas that of O3 and total monoterpenes was r = -0.76. Carbon dioxide is a more conserved tracer than monoterpenes which are oxidized by O3 (and the nitrate radical,

NO3). The nopinone mixing ratios (Figure 6.2B) did not increase during ODEs (Figure 6.4B), with exception of the nights of July 18/19 and 19/20, which indicates that titration of O3 in the gas-phase was a minor O3 depletion process and suggests that O3 is lost more by dry deposition than by titration in the gas-phase. This conclusion is supported by the absence of new particle formation and aerosol growth during the ODEs (see section 6.3.4 below).

101

Figure 6.4 (A) Median mixing ratios of monoterpenes as functions of time of day. (B) Nocturnal nopinone concentrations during ODE and non-ODE events as functions of time of day.

In polluted areas, O3 mixing ratios are often depleted due to titration of O3 by NO, which manifests itself by an anti-correlation of O3 and NO2 mixing ratios (Neu et al., 1994; Kleinman et al., 1994; Trainer et al., 1987; Logan, 1989). In contrast to CO2 and monoterpenes, O3 mixing ratios only correlated weakly with those of the anthropogenic pollution tracer NO2 (r = -0.44). A scatter plot of O3 against NO2 mixing ratios is shown in Figure 6.5C, where the data points collected during ODEs are colour-coded by CO2 mixing ratio. The majority of data points during

ODEs are associated with NO2 mixing ratios < 2 ppbv. Hence, the contribution of NO-O3 titration to ozone depletion was minor.

102

This conclusion is corroborated by considering trends in Ox which parallel those of O3

(Figure 6.2C). When Ox (instead of O3) is correlated with total monoterpenes during ODEs, the anti-correlation becomes more pronounced (r = -0.78 compared to -0.76). In contrast, the anti-

103

correlation of Ox with CO2 changes to r = -0.81 (from -0.92), likely because NO2 and CO2 are both present in combustion exhaust where they correlate.

Overall, the correlations of O3 with chemical tracers corroborate the notion that ODEs are not driven primarily by anthropogenic pollutants and that ODEs occur in air masses that have been in contact with respiring vegetation.

6.3.3 Role of meteorology

The ODEs were generally marked by a reduction in local wind speed. The lowest O3 mixing ratios were observed when the air stagnated (Figure 6.5A). This suggests that ODEs are caused by a local mechanism and are not representative of a process occurring over a wider scale. There was no correlation of ODEs with tide height (data not shown).

Synoptic conditions during the ODEs, judged from composite sea level pressure and geopotential height and their anomaly maps obtained from the NOAA's Physical Sciences

Division's web site (Figure S-2 in the original manuscript (Tokarek et al., 2017) were similar to those reported by McKendry et al. (2014), in that they show a high pressure region centered north of the Hawaiian Islands in the Pacific Ocean, with a developing high-pressure ridge aligned along the Pacific Northwest coast. Though not as pronounced as shown by McKendry et al. (2014), this high pressure region extends above the measurement site on Vancouver Island.

The anticyclonic (i.e., clockwise rotation) conditions are expected to be associated with predominantly WNW to NW winds in the study region as well as downward advection of

104

warmer, dry air from the upper troposphere in the anticyclone's center, leading to greater temperature inversion and capping of the boundary layer throughout the high-pressure system

(Stull and Ahrens, 2000).

McKendry et al. (2014) had based much of their analysis on HYSPLIT back trajectories with the NAM met field. Tables S-2 through S-10 in the supplemental information for the original publication summarize the hourly HYSPLIT back trajectory data for this data set

(Tokarek et al., 2017). By and large, the HYSPLIT results agreed with local wind direction, with exception of July 27, when the flow reversed (from on-shore to off-shore flow) in the middle of the night which was not captured by the HYSPLIT trajectories. The ends of the ODEs were generally also marked by a change (often a reversal) of local wind direction, which also were not captured by HYSPLIT trajectories (e.g., on July 16, 23, 30). This suggests that the meteorological field only poorly captures the land breeze (which goes against the synoptic westerly flow), and that the HYSPLIT back trajectories do not capture injections of continentally influenced air into the coastal MBL.

Corroborating this are the relative humidity (RH) data. Often, the RH in HYSPLIT was inconsistent with local observations. For example, during the ODEs on July 18, 19 and 30, local

RH were >80%, whereas those in the simulations were <65%. A lower RH indicates an air mass that has travelled more over (dry) land than over the ocean. Interestingly, the nights of July 18,

19, and 30 had the highest CO2 and total monoterpene abundances, which is consistent with the air having been in contact with land upwind. In general, RH decreased at the end of the ODEs as the air masses shifted. 105

The majority of the ODEs were associated with winds from the WNW, i.e., with air that has travelled in parallel to the coast line, consistent with the synoptic conditions. The nights of

July 21 and 22 are interesting cases in that on both occasions, local winds and HYSPLIT back trajectories were from the WNW sector and wind speeds decreased (to < 1 m/s on July 22), yet

O3 was conserved on those nights. One can conclude that the HYSPLIT back trajectories by themselves are poor predictors of ODEs. The NOAA ESRL maps show that the high pressure region was located significantly further from the coast on those days (Figure S-3 in the supplemental information for the original publication (Tokarek et al., 2017)) than during ODE events and that a low pressure system had formed to the North of the study region.

6.3.4 Correlations of ODEs with changes in aerosol size distribution

Reactions of O3 with monoterpenes are known to yield SOA (Griffin et al., 1999;

Sakulyanontvittaya et al., 2008b; Ehn et al., 2014). The time series of submicron aerosol size distributions is shown in Figure 6.6. During the ODEs, the accumulation mode (diameter range

80 to 100 nm) dominated. The highest counts were observed when the predominant wind direction was from the east (nights of July 18/19 and 19/20), whereas much lower aerosol loadings were observed when the winds were from the West. With exception of the night of July

19/20, there was neither noticeable particle growth nor nucleation event during the ODEs. The absence of aerosol growth and new particle formation corroborates that dry deposition is the primary O3 loss mechanism and that titration of O3 by gas-phase monoterpenes (which would have yielded SOA and nopinone) was minor. However, new particles were frequently observed after the ODEs ended, usually associated with a wind shift to the east at or after sunrise.

106

Figure 6.6 Time series of submicron aerosol size distributions.

6.3.5 Monoterpene ratios: insights into BVOC sources

An interesting feature of this data set is that the limonene mixing ratio was often greater than those of α- and β-pinene (and sometimes even greater than their sum). Limonene is a factor of 2.5 more reactive than the other monoterpenes with respect to oxidation by O3 and NO3

(Draper et al., 2015). The high relative concentrations of limonene observed suggest that the air had not been significantly processed ("aged") chemically, which implies that the source (or sources) of limonene was (were) close to the sampling site. Carene and terpinolene were less abundant than the other monoterpenes and accounted for only a small fraction of the total concentration of terpenes observed (Figure 6.4A). Only one of the common tree species in the area, the Western Red Cedar, emits a greater amount of limonene than other monoterpenes

(Drewitt et al., 1998); its emission ratios are most consistent with the observed monoterpene mixtures. In contrast, emissions from Coastal Douglas Firs are expected to have a higher α- and

β-pinene content and relatively little limonene (Geron et al., 2000; Burney and Jacobs, 2012), and emissions from Western Hemlock trees have a high α - and β-pinene as well as phellandrene content (Lagalante and Montgomery, 2003), which are less consistent with the relative terpene concentrations observed.

107

Oddly, there was a dependence of the monoterpene ratios on local wind direction during the ODEs; when local wind speeds were strongest and from the WNW, the ratios of limonene to

α- and β-pinene were highest (Figure 6.5D). The ocean surface is a known (though small) source of a variety of volatile compounds (Carpenter et al., 2012), including unsaturated compounds such as isoprene (Yokouchi et al., 1999; Meskhidze and Nenes, 2006) and monoterpenes (Yassaa et al., 2008). However, one local peculiarity at APO is the presence of kelp forests along the coast line (Watson and Estes, 2011).

6.3.6 Kelp forests: an unrecognized source of BVOCs

There is currently very little literature on BVOC emissions from near-shore vegetation.

Yassaa and co-workers recently reported evidence for monoterpene emissions from phytoplankton (Dunaliella tertiolecta), in which p-ocimene, limonene, and camphene were the dominant monoterpenes emitted (Yassaa et al., 2008). Kelp is known to emit molecular iodine and small molecule halocarbons (e.g., (Dixneuf et al., 2009; Ball et al., 2010; Weinberg et al.,

2013; Nitschke et al., 2015)). These iodine emissions and their subsequent chemistry have been linked to near-coastal new particle formation (O'Dowd and Hoffmann, 2005; O'Dowd et al.,

2002) though uncertainties remain (e.g., (McFiggans, 2005; O'Dowd et al., 2005).

To investigate whether local sea weed species contributed to BVOC emissions, samples were collected from the intertidal zone, placed in buckets along with sea water, and the head spaces were analyzed by GC-ITMS. The results of this analysis are described in detail in Chapter 7.

108

6.4 Summary and conclusions

In this Chapter, I have shown that the ODEs observed at APO occurred in air masses that had been in contact with near-coastal vegetation including kelp forests and trees located near the shore line. Even though this data set was limited to one month during summer, it can be expected that the O3-terpene anticorrelation will also be observed during other seasons as terpene emissions continue throughout the year and peak in spring (Helmig et al., 2013). The absence of aerosol growth and new particle formation associated with ODEs indicates that dry deposition was likely the primary O3 loss mechanism and that titration of O3 by gas-phase monoterpenes was a minor pathway. At the same time, more measurements, in particular of terpenoids towards which O3 is reactive such as the sesquiterpenes (Sakulyanontvittaya et al., 2008a), are needed to better constrain their roles in O3 depletion, especially within the forest canopy. Speciated VOC measurements in particular would help to shed light on O3 loss reactions by giving a more complete picture of oxidant reactivity in the region. However, the absence of organic aerosol formation during ODEs suggests that reactions of O3 with biogenic VOCs were of minor importance during this campaign. Back trajectories calculated using HYSPLIT and the NAM meteorology field poorly represented local coastal land-sea breeze effects and hence fell short in explaining how the upwind air at APO came in contact with continentally influenced air masses.

Future studies should focus on a better description of the regional flow patterns and should aim to quantify the relative contributions of terrestrial and oceanic surfaces to O3 dry deposition, the potential for halogen-catalyzed photochemical destruction of O3 (e.g., (Huang et al., 2010)), and the exchange of CO2 with surfaces and vegetation in the region. Future studies should also

109

elucidate whether near-costal oceanic uptake of O3 is enhanced in the region, for example by elevated iodide surface concentrations (Reeser and Donaldson, 2011; Carpenter et al., 2013).

110

Chapter Seven: Kelp Headspace Analysis during ORCA 2015

Parts of this Chapter are currently under review in Atmospheric Environment.

7.1 Introduction

Emissions of BVOCs constitute the largest source of organic matter in the troposphere, affecting cloud formation and climate (Jimenez et al., 2009; Goldstein and Galbally, 2007;

Hewitt et al., 2011). In contrast to terrestrial vegetation, the biotas of the Earth’s oceans are a minor atmospheric source of BVOCs in the troposphere. Nonetheless, oceanic emissions of small molecular weight hydrocarbons such as the methyl halides and dimethyl sulfide (DMS) are recognized as important components of many global biogeochemical element cycles, for example, carbon, oxygen, sulfur, nitrogen, and the halogens (Lovelock et al., 1972; Andreae and

Raemdonck, 1983; Carpenter et al., 2012). Large quantities of organic carbon and molecular oxygen are produced in the Earth's oceans by phytoplankton, microscopic organisms too small to be individually seen with the unaided eye (Falkowski et al., 1998). The biological activity of oceanic phytoplankton yields isoprene (C5H8) (Yokouchi et al., 1999; Sinha et al., 2007;

Wingenter et al., 2004; Bonsang et al., 2010) and, to a lesser extent, the monoterpenes (C10H16)

(Yassaa et al., 2008; Meskhidze et al., 2015) that partition to the marine boundary layer. In spite of the relatively low levels emitted, marine-derived isoprene (Gantt et al., 2009; Broadgate et al.,

1997; Palmer and Shaw, 2005; Arnold et al., 2009) and monoterpenes (Luo and Yu, 2010; 2013) contribute to the organic content of marine aerosols and influence cloud formation and climate in

111

tropical and remote regions. Isoprene can also be produced on the ocean surface through abiotic photochemical pathways (Ciuraru et al., 2015).

In addition to phytoplankton, macroalgae, broadly defined as algae visible to the naked eye, i.e., kelp, and, to a lesser extent, marine flowering plants, i.e., sea grasses, produce BVOCs in the marine environment (Broadgate et al., 2004; Bravo-Linares et al., 2010; Laturnus, 1996;

Giese et al., 1999; Sartin et al., 2001; Garcia-Jimenez et al., 2013; Ekdahl et al., 1998; Leedham et al., 2013; Sartin et al., 2002). While emissions of a select few volatile, brominated (Carpenter and Liss, 2000) and iodinated hydrocarbons (Giese et al., 1999) by marine macroalgae have a broader impact (e.g., on O3 budgets), generally, they are only of regional importance, because of their existence mainly along the coastlines of continents in the intertidal and subtidal zones and

'hot spots' in the tropics (Exton et al., 2015). However, the production of BVOCs in coastal marine ecosystems remains poorly constrained by ambient observations. It is known that macroalgae emit isoprene and other C≤5 non-methane hydrocarbons (Broadgate et al., 2004).

Sartin et al. (2002; 2001) showed that seaweed in the inter-tidal zone on the coast of Ireland emit

C8-C15 n-alkanes and C8-C10 aldehydes; the latter are associated with biotic and abiotic stress factors (Goulitquer et al., 2009).

In July 2015, our group conducted ambient air measurements at the Amphitrite Point

Observatory (APO) on the west coast of Vancouver Island near the town of Ucluelet, British

Columbia (see Chapter 6) (Tokarek et al., 2017). On the west coast of North America, large kelp

"forests" are located from the Californian Coast to Aleutian Islands due to a favorable balance of available nutrients (from the upwelling of ocean currents), sunlight, and oceanic temperature 112

(Steneck et al., 2002). Genera found in these forests include Laminaria (Devil's apron, sea colander), Ecklonia (brown algae), Macrocystis (giant kelp), Nereocystis (bull kelp), and

Pterygophora (stalked kelp) (Wilmers et al., 2012; Reed and Brzezinski, 2009).

During on-shore flow, the sampled air was enriched in limonene relative to - and β- pinene, in contrast to when winds originated from land (Tokarek et al., 2017). A similar enhancement was observed when head space vapors above Nereocystis luetkeana (bull kelp) were analyzed, suggesting that near-shore vegetation was the source of the observed monoterpenes (Tokarek et al., 2017).

In this Chapter, measurements of C9-C16 hydrocarbons in the head spaces of several macroalgae species collected near the APO site, Fucus gardneri (rock weed), Ulva spp. (sea lettuce), Callophyllis spp. (red sea fans), Alaria marginata (winged kelp), and Nereocystis leutkeana (bull kelp), are presented. Potential implications of these observations are discussed.

7.2 Methods

For details related to the measurement site, instruments operated, and calibrations, see

Chapter 6. The GC-ITMS described in Chapter 2 sampled from a stainless-steel inlet whose tip was located at a height of 4.3 m above ground (Figure 7.1). A section of this inlet (3.0 m) was heated to 125 °C to scrub ozone to prevent oxidation of unsaturated hydrocarbons during preconcentration (Hellén et al., 2012). Ambient air was sampled at a by-pass flow rate of 500 sccm in addition to (during preconcentration) an instrument flow rate of 213 mL min-1 and was

113

preconcentrated for 8 min onto the dual sorbent trap, which was held at 40 °C. After preconcentration, the sorbent trap was heated to 240 °C and held at this temperature for 5 min, while the trap was flushed with helium carrier gas (Praxair, ECD grade) onto a 30 m (length),

0.25 mm (inner diameter) and 0.25 µm (film thickness) DB-5MS analytical column (20% split).

The GC was operated with the following temperature program: hold at 40 °C for 3 min, heat to

70 °C at a rate of 3 °C min-1, hold at 70 °C for 2 min, heat to 200 °C at a rate of 15 °C min-1, and hold at 200 °C for 2 min, for a total GC runtime of 25.7 min. This was followed by a 5 min recovery period for a total run time of ~38.7 min.

Retention times and indices and quantification ions are summarized in Table 7.1. For quantification of monoterpenes, counts at m/z 92 and 94 were added to those of the dominant

+ fragment at m/z 93 (C7H9 ) to form the SIC because it improved the shape of the chromatographic peaks in some cases. The likely cause is (random) jitter of the electromagnetic fields that control ion selection in the ion trap mass spectrometer (Bell et al., 2015).

The relative response factors of VOCs not present in the standard gas cylinder (e.g., camphene) were determined after the campaign using ~10 mM standard solutions in HPLC grade methanol which contained α-pinene, decane, and undecane as internal standards. Typically, a

15 μL aliquot was injected into a 4200 L Teflon chamber that had been filled with scrubbed air, from which the GC-ITMS sampled using the same inlet as had been used in the field.

Samples of near-shore seaweed and kelp species were collected in the intertidal and shallow subtidal zones off the shore of the APO and were placed in a polypropylene container 114

and filled with sea water to a constant height. This container was covered with polyethylene plastic wrap to allow for gases to collect in the head space for a period of ~1 hour. The air temperature and time of day varied between samples. A ~1 m long, 0.635 cm (1/4") outer diameter and 0.476 cm (3/16") inner diameter Teflon tube was attached to the inlet, and one end was placed inside the head space for ~ 5 minutes prior to the sample cycle (Figure 7.1). Mixing ratios of VOCs in ambient air were measured immediately prior to each head space analysis.

Chromatograms were collected for the empty container (rinsed with tap or deionized water), the container filled with seawater (which likely contained organisms invisible to the naked eye, i.e., phytoplankton), and seawater filled with samples of rock weed, sea lettuce, red sea fan, winged kelp and bull kelp collected in the intertidal zone. Peaks were identified by comparison to authentic standards, via their retention index (RI), and National Institute of

Standards (NIST) mass spectra library searches.

115

Table 7.1 Quantification and retention information for observed VOCs. RT = retention time. RI = retention index based on an n-alkane hydrocarbon ladder on a DB-5MS column. n/d = not determined. RT RI RI Major ion(s) VOC (min) (observed) (literature)* (m/z) 8.33 n-Nonane 900 900 71, 85 9.71 α-Pinene 936 939 93 10.32 Camphene 950 951 93 11.41 β-Pinene 978 979 93 12.51 n-Decane 1000 1000 71, 85 12.76 Carene 1010 1017 80, 91 13.82 Limonene 1033 1033 93 16.34 Terpinolene 1093 1088 121, 136 16.86 n-Undecane 1100 1100 71, 85 16.98 Nonanal 1108 1106 98 18.85 n-Dodecane 1200 1200 71, 85 20.02 n-Tridecane 1300 1300 71, 85 21.28 n-Tetradecane 1400 1400 71, 85 22.22 n-Pentadecane 1500 1500 71, 85 * Averaged value of retention indices reported by (Xu et al., 2003; Slavkovska et al., 2005;

Mevy et al., 2006; Miyazaki et al., 2011; Maia et al., 2005; Lucero et al., 2006; Angioni et al.,

2004; Angioni et al., 2006; Silva et al., 2010; Sartin et al., 2001; Tuberoso et al., 2005)

116

Figure 7.1 Setup for head space sampling (not to scale). GC-ITMS = gas chromatograph ion trap mass spectrometer. MFC = mass flow controller. s.s. = stainless steel.

7.3 Results

7.3.1 Overview

A total ion chromatogram (TIC) of the headspace vapors above a sample of Alaria marginata (winged kelp) and that above sea water void of visible macro organisms are shown in

Figure 7.2A. The total ion counts of the kelp sample (blue trace) are considerably (by a factor of

~102 to ~103) elevated compared to those for the sea water sample (black trace) over the range of retention times shown, indicating that the kelp outgassed a large number of hydrocarbons, which were not (baseline) resolved in the TIC.

117

Selected ion chromatograms (SICs) were used to resolve, identify and, when possible, quantify the compounds emitted (section 7.3.2). An example SIC showing counts at m/z 93±1 for

+ a headspace sample of Nereocystis luetkeana (bull kelp), an ion (C7H9 ) associated with

(amongst others) monoterpenes, is superimposed in Figure 7.2A (right-hand axis). It shows baseline resolved peaks, notably of α-pinene and limonene.

Shown in Figure 7.2B are TICs and SICs of headspace vapors above a sample containing

Alaria marginata (winged kelp). These chromatograms are qualitatively similar in that peaks were enhanced by similar magnitudes over the range of retention times shown, although in different relative proportions.

118

Shown in Figures 7.3A, 7.3B and 7.3C are TICs and SICs for headspace vapors above samples of Fucus gardneri (rock weed), Ulva spp. (sea lettuce), and Callophyllis spp. (red sea fans), respectively. In contrast to the data shown in Figure 7.2, these chromatograms show little to no enhancement of hydrocarbons relative to the seawater sample.

119

120

7.3.2 Identification and quantification of BVOCs

The observed BVOCs broadly fell into three categories.

The first category of BVOCs includes those compounds for which authentic standards, i.e., reference mass spectra, were available; these species were quantified and their mixing ratios are summarized in Table 7.2. Compounds in this category include n-alkanes, nonanal, and monoterpenes. Cylinder and head space mass spectra for limonene (retention time (RT) 13.82 min) are shown in Figure 7.4. Compounds in this category were identified in nearly all samples, though only the Alaria marginata and Nereocystis luetkeana samples showed significant enhancements (up to factors of 50) compared to the head space above sea water.

The second category includes compounds that were identified on the basis of retention index (RI) and from a match of major ions with NIST mass spectra library entries but were not quantified; these molecules are listed in Table 7.3. Some of these compounds were identified only because we looked for them either because of chemical similarity to compounds listed in

Table 7.2 (e.g., octanal and decanal) and/or because their presence had been reported in previous head space analyses of kelp species e.g., octanal, 2-ethyl-1-hexanol, geranyl acetone, or 6- methyl-5-hepten-2-one (Sartin et al., 2002; Sartin et al., 2001; Goulitquer et al., 2009). Other molecules of note (i.e., not previously reported in the literature) in this category are toluene and limona ketone.

121

Empty Fucus Callophyllis Alaria Nereocystis Seawater Ulva spp. Container gardneri spp. marginata luetkeana α-pinene -5±6 3±6 0±1 -1±5 -1±6 58±5 -34±25 (pptv) Camphene n/d n/d n/d n/d n/d 25±2 4±1 (pptv) β-pinene -2±4 2±4 -1±1 1±2 0±4 133±8 31±7 (pptv) Carene (pptv) n/d n/d n/d n/d n/d n/d 0±1 Limonene -12±21 21±21 13±4 17±9 19±21 2220±135 1779±115 (pptv) Terpinolene n/d n/d n/d n/d n/d n/d 10±2 (pptv) Nonanal 9.5±1.8 11.4±2.1 5.5±1.0 9.2±1.7 10.4±2.0 7.8±0.6 61.1±4.5 (ppbv) n-Dodecane 31±3 12±1 9±1 6±1 11±1 156±28 247±44 (pptv) n-Tridecane 94±10 49±5 14±2 13±2 45±5 293±24 473±38 (pptv) n-Tetradecane 56±8 32±5 18±3 16±3 31±5 760±31 1289±49 (pptv) n-Pentadecane 126±19 162±24 11±2 26±4 149±22 507±171 2641±256 (pptv)

122

Figure 7.4 Comparison of limonene mass spectra from different sources. (A) The headspace of bull kelp, (B) the VOC calibration cylinder used in the field.

A large number of BVOCs, many of which did not resolve on the chromatographic column, were not identified. The common ions observed for these molecules (m/z 57, 71, 85, 99,

105, and 111) suggest that they consist mainly of branched alkanes (Cross et al., 2013). What

123

follows is a closer examination of the monoterpene (section 7.3.3), n-alkane (section 7.3.4), oxygenated (section 7.3.5), and aromatic hydrocarbon data (section 7.3.6).

7.3.3 Monoterpenes

The Alaria marginata and Nereocystis luetkeana head space samples were enriched in a number of monoterpenes, foremost in limonene, which accumulated to a mixing ratio of ~2.1 ppbv and 1.8 ppbv, respectively. In contrast, emissions of α- and β-pinene and camphene were lower than those of limonene (Table 7.2). The observation of terpene emissions is in contrast to

Sartin et al. (2001; 2002) who found no evidence of terpene emissions by seaweed species at

Mace Head, Ireland. On the other hand, the head space samples of Fucus Gardneri, Ulva spp., and Callophyllis spp., showed little to no monoterpene enhancements.

The high limonene concentrations were unexpected and were hence scrutinized further.

The observed RI (1033) matches a literature average of 1033 (Table 7.1), and the mass spectra are also excellent matches (Figure 7.4). However, it is possible that another monoterpene,

ß-phellandrene which has a similar RI to limonene, co-elutes with limonene and hence would interfere with its measurement. The average literature RI reported for ß-phellandrene is 1038, and if the values of (Miyazaki et al., 2011; Maia et al., 2005) of 1052 are excluded, the average literature RI changes to 1031. Electron impact ionization of ß-phellandrene would give rise to mass fragments identical to limonene, but not yield m/z 67 and 68 (Linstrom and Mallard, 2001) which were observed (Figure 7.4). It is hence unlikely that interference from ß-phellandrene was significant, confirming that the emitted terpene was, in fact, limonene.

124

7.3.4 n-Alkanes

All analyzed samples contained quantifiable amounts of C12-C15 n-alkanes that were in excess of ambient air concentrations measured prior to head space analysis (Table 7.2). Mixing ratios of n-pentadecane were enhanced the most, followed by n-tetradecane and n-tridecane. The highest mixing ratios were observed for the Nereocystis luetkeana and Alaria marginata samples, with head space mixing ratios of n-tetra- and n-pentadecane peaking at 1.3 ppbv and 4.4 ppbv, respectively. These observations are consistent with those by Sartin et al. (2001; 2002) who showed that seaweed species can emit n-alkanes (n-pentadecane in particular) via biochemical routes.

Surprisingly large mixing ratios (Table 7.2) were observed for the sea water sample and the "empty" polypropylene container (which had been rinsed with water). This suggests that the n-alkanes outgassed from the inner walls of the container and/or the water itself, possibly due to in-field contamination followed by temperature- and/or light-driven volatilization. While similar mixing ratios were observed for Callophyllis spp., lower concentrations were measured for

Fucus gardneri and Ulva spp. (Table 7.2) for reasons that are unclear.

7.3.5 Oxygenated volatile organic compounds

Sartin et al. (2001; 2002) quantified a number of OVOCs emitted from Fucus Spiralis

(Spiral wrack), seaweed in the same genus as Fucus gardneri (rock weed) used in this study.

These OVOCs included octanal, nonanal, decanal, 2-ethyl-1-hexanol, geranyl acetone, and di- tert-butyl-p-benzoquinone. In this work, the same list of oxygenated species was observed in two

125

seaweed samples, those of Alaria marginata and Nereocystis leutkeana, including octanal, nonanal, decanal, 6-methyl-5-hepten-2-one, geranyl acetone, 2-ethyl-1-hexanol (Table 7.3), with exception of di-tert-butyl-p-benzoquinone, which was not detected. Nonanal was present in all samples in unreasonably high abundance, including the seawater and container blanks, which suggests that its abundance was affected by emissions from the container walls and/or the plastic wrap.

Limona ketone is an oxidation product of limonene (Jaoui et al., 2006; Rossignol et al.,

2012); it was tentatively identified only in the head space of Nereocystis luetkeana on the basis of its retention index and m/z 67, 95 and 123 (Table 7.3) which are the dominant ions for this compound in the NIST database. The other chromatograms did not show a discernible peak associated with the above ions in the expected window of retention times.

7.3.6 Aromatic hydrocarbons

Toluene was the only aromatic compound detected. Its dominant electron impact mass fragments are m/z 91 and 92 (with a minor isotopomeric contribution at m/z 93, Figures 7.2A and

2B) eluted at ~4.6 min and was detected in all samples. Its concentration was enhanced by orders of magnitudes in the head spaces above all of the seaweed samples.

126

Table 7.2 VOCs identified (but not quantified) by retention information and major ions. The number for each sample is the logarithm of the peak area for the tabulated major ions. Cases where peaks were absent are indicated with “-“. RT = retention time. RI = retention index based on an n-alkane hydrocarbon ladder on a DB-5MS column. Alaria RT RI RI Major ions Empty Sea- Fucus Ulva Callo- Nereocystis VOC margi- (min) (observed) (literature)* (m/z) Container water gardneri spp. hyllis spp. luetkeana nata 4.70 Toluene 765 771 91, 92 4 3 6 5 5 8 7 2,3 - 11.82 986 991 43, 71, 99 5 5 5 5 5 † 6 octanedione 6-methyl-5- 55, 69, 93, 11.88 hepten-2- 987 992 6 5 5 6 5 6 6 108 one 12.68 n-octanal 1006 1008 70, 83, 98 6 5 5 5 5 7 6 2-ethyl-1- 13.57 1028 1028 93, 121, 136 - - - - - 6 - hexanol 70, 82, 95, 16.38 Terpinolene 1093 1088 - 3 - - 3 - 5 112 Limona 17.39 1119 1121 67, 95, 123 ------5 ketone 69, 125, 18.95 n-decanal 1205 1204 5 5 5 5 5 5 6 136, 151 Geranyl 125,136, 21.73 1450 1450 5 5 5 5 5 5 5 Acetone 151 * Averaged value of retention indices reported by (Xu et al., 2003; Slavkovska et al., 2005; Mevy et al., 2006; Miyazaki et al., 2011;

Maia et al., 2005; Lucero et al., 2006; Angioni et al., 2004; Angioni et al., 2006; Silva et al., 2010; Sartin et al., 2001; Tuberoso et al.,

2005)

†Peak unresolved

127

7.4 Discussion

7.4.1 Monoterpene emissions from kelp

This study was motivated by observations of monoterpenes in ambient air at the APO, which suggested a marine source (Tokarek et al., 2017). This work has shown that monoterpenes, and limonene in particular, are emitted from two near-shore marine macroalgae species, Nereocystis luetkeana (bull kelp) and Alaria marginata (winged kelp), which are in the taxonomic order Laminariales and families Alariaceae and Laminariaceae, respectively.

Marine algae produce monoterpenes as secondary metabolites (Wise, 2003), possibly as a defense against herbivores or as a response to wounding, though at a high metabolic cost (Hay and Fenical, 1988; Wise, 2003). Extraction studies have shown that monoterpenes and monoterpenoids, i.e., chemically modified monoterpenes including many halogenated variants, are present within macroalgae (Wise et al., 2002; Shaw et al., 2010; Wise, 2003; Naylor et al.,

1983). Relevant to the results of this study, limonene has been detected in many, but not all, algae variants, including Laminaria sp. (Naylor et al., 1983), which belongs to the same family as Nereocystis luetkeana. Using the cultured seaweed Ochtodes secundiramea as a model system, Wise et al. showed limonene is produced from neryl diphosphate (Wise et al., 2002;

Wise, 2003). Ochtodes secundiramea belongs to the phylum Rhodophyta, and it is unclear to what extent monoterpene biosynthesis differs between it and the species investigated in this study. Our results indicate that there may be some difference between species, since we found no evidence for terpene emissions from Ulva spp. (sea lettuce), Callophyllis spp. (red sea fan) or 128

Fucus gardneri (rock weed) in contrast to Nereocystis luetkeana and Alaria marginata

(Table 7.2).

In general, very little is known about the emissions of long-chain hydrocarbons by macroalgae in the marine environment from ambient observations, such that it is challenging to put our observations in context. To our knowledge, the only studies having investigated BVOC emissions from species related to Nereocystis luetkeana and Alaria marginata are those by

Broadgate et al. (2004), Bravo-Linares et al. (2010), and Goulitquer et al. (2009). They investigated BVOC emissions from Laminaria digitata, which belongs to the order

Laminariales, family Alariaceae and genus Laminaria. Bravo-Linares et al. (2010) measured several short-chain hydrocarbons but did not comment on isoprene or terpene emissions, and it is unclear if their analytical method would have detected these species, if they had been present.

Goulitquer et al. (2009) limited their analysis to aldehydes (see below). Broadgate et al. (2004) limited their analysis to C≤5 hydrocarbons and observed emissions of isoprene from Laminaria digitata. Since isoprene (and by inference, monoterpene) production rates can vary between species (Shaw et al., 2010) and is linked to metabolic activity, it is not clear if isoprene emissions

(not detectable by our GC-ITMS) also occurred from Nereocystis luetkeana or Alaria marginata.

There is more relevant information in the literature with respect to Fucus gardneri. Bravo-

Linares et al. (2010) investigated emissions from Fucus serratus, Broadgate et al. (2004) studied

C≤5 hydrocarbon emissions from Fucus serratus and Fucus vesiculosus, and Sartin et al. (2002;

2001) investigated long-chain hydrocarbons (C8-C16) emissions from Fucus spiralis and Fucus

129

vesiculosus. All of these species are in the taxonomic family of Fucaceae and are in the same genus (Fucus).

The only study in the literature to date, where measurements of terpenes from related species were attempted is that of Sartin et al. (2002; 2001), who found no evidence of terpenes in the head space above Fucus spiralis. Our analysis of emissions from Fucus gardneri is consistent with this earlier finding, as the headspace above the sample of Fucus gardneri did not contain α-pinene, camphene, carene, or β-pinene, though was slightly elevated in limonene

(Table 7.2; Figure 7.3A). This suggests that members of the Fucus genus are an insignificant source of gas-phase monoterpenes in general. However, environmental factors (see section 7.5) may have affected emissions.

7.4.2 n-Alkanes

Sartin et al. (2002; 2001) observed C8-C15 n-alkanes in their analyses and reported biochemical, photosynthetically active radiation (PAR) controlled emissions of n-pentadecane from Fucus spiralis as well as temperature-driven volatilization of n-alkanes, even in the absence of seaweed (i.e., when the head space above a combination of sand and seawater was examined).

Our results are consistent with those of Sartin et al. (2002; 2001), in that concentrations of C12 to

C15 n-alkanes were enhanced in all samples (relative to ambient air), including the sea water and the empty container (Table 7.2). The latter was likely an artifact, since the container was rinsed only with water, which would not adequately remove any residual alkanes. Interestingly, mixing ratios of n-alkanes were greatly enhanced in the presence of Alaria marginata and Nereocystis 130

luetkeana, suggesting that these species share similar biochemical routes for n-alkane synthesis, as identified by Sartin et al. (2002; 2001), in that for some reason these pathways were particularly active. In contrast, the VOCs and levels above the Ulva spp. (sea lettuce) and

Callophyllis spp. (red sea fan) samples were similar to those of Fucus gardneri.

7.4.3 Oxygenates

Goulitquer et al. (2009) showed that aldehydes are released from Laminaria digitata in response to certain stresses such as emersion at low tide. Bravo-Linares et al. (2010) and Sartin et al. (2002; 2001) reported emissions of oxygenates from Fucus serratus and spiralis, i.e., aldehydes such as octanal, nonanal and decanal, and more unusual oxygenated compounds such as 2-ethyl-1-hexanol and geranyl acetone. These molecules were also observed in this work

(Table 7.3), though their abundances were small (compared to the n-alkanes) such that they might have gone unnoticed if we had not specifically searched for them. The only exception was nonanal, which was very abundant in all samples, including the blanks, suggesting that it was at least partially emitted by the container. Aldehydes were observed in all samples such that its plausible source was seawater and the inner walls of the container.

There was no evidence for the presence of 2,6-di-tert-butyl-p-benzoquinone, or 3- hydroxy-2,4,4-trimethylpentyl-ester-2-methyl-propanoic acid, species that had been reported by

Sartin et al. (2002; 2001). Another compound found in the majority of the seaweed headspaces was 6-methyl-5-hepten-2-one; however, this compound was found in the empty container, and was likely an artifact. 131

7.4.4 Toluene

Our data indicated that toluene out-gassed from all seaweed species examined. We found no indications for the presence of toluene in, or emissions from, seaweed in the literature, including recent comprehensive studies (Lopez-Perez et al., 2010). We considered that the source of toluene may have been anthropogenic (e.g., leaked fuel from watercrafts), rather than biogenic. However, an anthropogenic source is inconsistent with the absence of other aromatic species (e.g., o-xylene, ethylbenzene, etc.) that are present in fuel and that our GC-ITMS is sensitive to, which suggests the source of toluene is biogenic.

Biogenic emissions of toluene would likely be small and easily masked by anthropogenic emissions and may have therefore gone unnoticed in other studies. If a biogenic toluene source were to exist, its oxidation would yield glyoxal, whose budget in the marine boundary layer is currently not reconciled, leading to a speculation with respect to its sources (Mahajan et al.,

2014). Clearly, more observations of toluene emissions from seaweed are needed.

7.4.5 Unidentified hydrocarbons

Marine organisms are known to form a plethora of natural products (Moore, 1977), many of which are volatile in nature. Not surprisingly, the chromatograms contained numerous species that could not be identified. In addition, there were many molecules which did not resolve on the chromatographic column and appeared as a broad background from RI of ~1000 to ~1400 and whose mass spectra were dominated by m/z 57, 71, and 85. These ions are typical of branched

132

alkanes (or highly alkylated species) such as methylundecane, dimethyldecane, dimethylundecane, etc. of which there are many structural isomers. Many such compounds are found in seaweed extracts (Lopez-Perez et al., 2010) and may have also included oxygenated

+ + molecules: counts at m/z 57, for example, may be due to C4H9 (m/z 57.0699) or C2H5CO (m/z

57.0335). Additionally, we might have expected to see some of the aldehydes identified by

Goulitquer et al. (2009) but whose retention times and mass spectra are not known, e.g., (E)-hex-

2-enal, (E)-non-2-enal, dodecadienal, (E)-4-hydroxyhex-2-enal , and (E)-4-hydroxynon-2-enal.

Furthermore, we would expect to observe halogenated terpenes, whose mass spectra and retention times are also not known. Clearly, additional studies with more advanced analytical instrumentation than used in this study are needed to resolve and identify the various hydrocarbons emitted by kelp.

7.5 Atmospheric implications and future work

The goal of the presented study was the identification of the source of limonene emissions due to observed elevated mixing ratios relative to α-pinene and ß-pinene in ambient air. Large emissions of limonene and other BVOCs were observed for two kelp species,

Nereocystis luetkeana and Alaria marginata. This is an important result because these kelp species comprise a large portion of the marine vegetation around the northern Pacific coastline.

Specifically, Nereocystis luetkeana is the most prominent kelp species on the northern west coast of North America and makes up part of the vast kelp forests in the region (Antrim et al., 1995), and Alaria marginata constitutes a large portion of the kelp in the area near the APO (Paine,

133

2002). It is hence likely that these species emit terpenes (and other BVOCs) along the entire northern Pacific coastline.

The next step is to identify how these emissions vary with environmental factors, such as temperature, PAR, plant life cycle and age, and stress factors such as emersion at low tide. A particularly important (and often neglected) variable is season, known to affect metabolic activity of seaweeds (Graiff et al., 2015) and hence production of secondary metabolites such as limonene. The species sampled in this work were collected within the intertidal or shallow subtidal zones off the coast of the APO in July. Due to the nature of the intertidal region, some of the species collected may have been subjected to battering, herbivory, exposure to heat and light, changes to water salinity, and tissue death prior to collection and sampling. Hence, a potential mechanism by which monoterpenes are volatilized may be microbial decomposition of kelp matter as has been observed, for example, in landfills (Wang et al., 2008). The potential role of microbes in BVOC emissions from kelp will need to be examined.

Emissions of BVOCs from kelp will affect HO reactivity and hence the rates of photochemical production of O3 in coastal regions. Furthermore, much of this material will add to the organic content of marine derived aerosol, potentially affecting numbers of cloud and ice nucleating (CN and IN) aerosol, believed to impact the atmosphere as a whole (Wilson et al.,

2015). Emission of limonene is of particular interest since it has a greater potential to form SOA in the presence of O3 than any of the other common monoterpenes (Draper et al., 2015).

134

BVOC emissions from seaweed species are not currently included in chemical transport model simulations investigating the budgets of O3 formation, SOA, or climate, as relevant emission factor data are lacking such that their impact is currently not known. However, their impact may be significant considering that the majority of the world's population lives in the proximity of an ocean and sea grasses, and kelp marine forest and algae are increasing due to fertilizer run-off.

135

Chapter Eight: Peroxyacyl Nitrate Chemistry during PANfire 2017

8.1 Summary

This Chapter describes field work in northwest Calgary, Alberta relating to the measurement of PANs during periods of biomass burning plume impact from forest fires throughout western

USA and Canada. Peroxyacyl nitrates are photooxidation products of VOCs and can provide insight into the history of an airmass (i.e., biogenic, anthropogenic, industrial). Auxiliary measurements made by the Calgary Regional Airshed Zone (CRAZ) provided a useful data set to explore the chemistry affecting peroxyacyl nitrates that may occur in biomass burning plumes. In particular, PAN loss mechanisms are explored in an attempt to describe a wide range (0.05 – 0.17) of PPN/PAN ratios.

8.2 Introduction to peroxyacyl nitrates

The oxidation of volatile organic compounds in the presence of NOx produces O3 and, as a by-product, PANs (general formula RC(O)O2NO2, IUPAC name peroxycarboxylic nitric anhydrides) (Fischer et al., 2014; Mielke et al., 2011; Tokarek et al., 2014; Roberts et al., 2004).

The PANs can be the largest component of odd nitrogen (NOy = NOx + HONO + HNO3 + HNO4

+ NO3 + 2N2O5 + ΣAN + ΣPAN + …) in ambient air (Roberts et al., 2004; Roberts, 1990). So far, 43 PANs have been discovered including the most commonly observed species PAN and

PPN, depicted in Figure 1 of Tokarek et al. (2014).

136

PAN mixing ratios range from pptv in warm, remote areas to ppbv in polluted areas

(Tokarek et al., 2014; Fischer et al., 2014). It is formed from the peroxyacetyl radical (PA,

CH3C(O)OO), which in turn is produced by numerous photochemical pathways (Figure 1.5).

Since the PA radical is formed photochemically, PAN production occurs mostly during the day.

Thermal decomposition is the most important PAN loss mechanism in the lower troposphere during the day (Roberts, 2007). As a result, minimum PAN mixing ratios usually occur overnight, and a maximum occurs in the early to mid-afternoon (Tokarek et al., 2014). Since

PAN is formed from NOx, produces NOx upon thermal decomposition, and has a long lifetime in the mid- and upper troposphere, it is considered a long-lived reservoir species for NOx (Fischer et al., 2014).

8.2.1 Peroxyacyl nitrates in biomass burning plumes

In biomass burning plumes from forest fires, a large fraction (51% to 74%) of NOx is converted to PAN within the first hour from the point of emission as the plumes travel away from the fire (Liu et al., 2016 reactive nitrogen, and organic aerosol). Due to the buoyancy of biomass burning plumes, they enter the free troposphere where the cold temperatures stabilize

PAN and allow for its long-range transport (Fischer et al., 2014). In clean air, the ratio of

PPN/PAN is usually ~0.10. In biomass burning plumes, however, this ratio increases to about

0.15-0.20 (Zaragoza et al., 2017). PAN formed from fires at high northern latitudes is longer lived than that formed from lower latitudes. In high latitude regions, therefore, it is important to consider PAN as a component of downwind photochemical smog formation (Fischer et al.,

2014). 137

8.2.2 Instrumentation and consumables for measuring peroxyacyl nitrates

GC-ECD remains one of the most popular methods used to measure PANs due to its high sensitivity and separation capabilities (Roberts, 2007). Most commonly, GC-ECDs are operated with helium as a carrier gas. However, due to its limited reserves and fluctuating supply on global markets, more sustainable options are needed (Nnaji et al., 2015; Benedetti et al., 2017;

Reisch, 2017). Hydrogen is often considered to be a good replacement carrier gas for helium because of their similar properties (i.e., non-reactive to analytes of interest) (Fiorini and Boarelli,

2016; Benedetti et al., 2017). Due to its abundance and cost, hydrogen is a more economic option as a carrier gas for GC work (Nnaji et al., 2015; Benedetti et al., 2017). Furthermore, hydrogen has a faster optimum linear velocity due to lower viscosity than helium, allowing for shorter runtimes (Fiorini and Boarelli, 2016; Benedetti et al., 2017; Nnaji et al., 2015).

Luxenhofer et al. (1996) successfully used hydrogen as a carrier gas for the detection of C6-C17 alkyl nitrates in 1996 (Luxenhofer et al., 1996). Since then, few studies using hydrogen as a carrier gas for GC have emerged. These applications include GC-MS and GC-FID, but none have been used to measure PANs (Fiorini and Boarelli, 2016; Benedetti et al., 2017; Watanabe et al., 2016; Nnaji et al., 2015).

8.3 Materials and Methods

Field campaign measurements were made at the University of Calgary campus in

Calgary, Alberta. The university is located in the northwest quadrant of the city, where it is surrounded by residential communities, roadways and construction zones. Measurements in 2017

138

were made from August 26 until October 22 in a trailer on the northwest end of campus (latitude

= 51.0793°, longitude = -114.1418°). The site was ~500 m away from a major road as well as a construction site, which occasionally created large amounts of dust. While at this site, forest fires were prominent throughout BC, resulting in fore biomass burning plumes impacting the site most days from August 30 until September 10. A second set of measurements was made in 2018 from

January 18 until March 4 using a rooftop lab in the Science B building, located in the northeast quadrant of campus. This site was in close proximity to the trailer (< 1 km, latitude = 51.0795 °, longitude = -114.1298 °) and is described in greater detail elsewhere (Woodward-Massey et al.,

2014; Mielke et al., 2016; Mielke et al., 2011).

Flows and consumable details are described in Chapter 2. For the first portion of this campaign, hydrogen from a PEAK Scientific hydrogen generator (version PH200) was used as a carrier gas. During this part of the experiment, the hydrogen flow rate was maintained at ~24 mL min-1 and the nitrogen flow rate was maintained at ~42 mL min-1. Flow rates in the trailer increased on rare occasions moisture freezing in the lines causing an increase in the pressure of the system. During the second part of this experiment in the rooftop lab, hydrogen from a cylinder was used as a carrier gas. Here, the hydrogen flow rate was maintained at ~24 mL min-1 and the nitrogen flow rate was maintained at ~48 mL min-1. Samples were injected onto the column via a 2 mL PEEK sample loop during trailer measurements and a 2 mL cheminert stainless-steel sample loop used for the rooftop measurements. A 10-port valve was used to inject the sample onto the column and is described in Figure 2 of Tokarek et al. (2014). A midpolarity, fused silica analytical column (Restek RTX-200) with an inside diameter of

139

0.53 mm, a film thickness of 1 μm, and a length of 15 m was used. The inlet was composed of

Teflon tubing and Teflon fittings while tubing from the gas cylinder was PEEK with stainless-steel swage fittings.

A small amount (~100 g) of CuSO4(H2O)5 (99.995%, Aldrich) in a stainless-steel sample cylinder with 150 cm3 internal volume capped by a 2 µm pore size in-line filter (Swagelok 316L-

HDF4-150 and SS-4FW5-2) was initially used to humidify the carrier gas. This was done as H2O has been reported to increase reproducibility of results when using continuous runs for long periods of time (months) by decreasing on-column PAN losses (Tokarek et al., 2014; Flocke et al., 2005). The water trap became contaminated while in the trailer and therefore was removed for rooftop measurements.

The PAN-GC was run with automated injections every 6 or 10 minutes. Injections lasted

6 or 30 seconds before switching to a position used to fill the sample loop for the remainder of the run (Figure 2 of Tokarek et al. (2014)). Throughout the campaign, the oven temperature was set to 25 °C and the detector temperature to 50 °C. While in the trailer, occasionally the oven temperature would rise above 25 °C as the external summer temperatures would heat the inside of the trailer. The ECD was set to the “CAP” (480 pA) sensitivity setting with a contact potential of 330 mV.

The setup used for measurement and calibration is similar to the setup in Figure 3 of

Tokarek et al. (2014). The TD-CRDS is fully depicted by Paul et al. and Thaler et al. (2009;

140

Thaler et al., 2011) but briefly, it is a spectroscopic technique that uses a heated inlet to dissociate organic nitrates into peroxyacyl radicals and NO2. Nitrogen dioxide concentrations are then measured by absorption. Simultaneously, a second ambient inlet measures the background mixing ratio of NO2. By subtracting the background mixing ratio from the total mixing ratio, the mixing ratio of organic nitrates delivered to the PAN-GC is calculated. PANs were delivered to the PAN-GC via oxygen flowing through a glass trap containing a synthetic standard (see section

8.3.2).

8.3.1 Data Reduction

Data reduction was carried out using the method described by Tokarek et al. (2014).

Briefly, a macro written in Igor Pro 6 (Wavemetrics) fits the chromatographic parameters of a

Gaussian peak (Equation 8.1). In this equation, V is the observed signal voltage, A is peak area,

σ is the standard width of the peak, t is the amount of time from injection until elution, tₒ is the retention time, Vₒ is the offset of the background, and m is the slope of the baseline. An example fit from the campaign is shown in Figure 8.1

푡−푡 2 퐴 −0.5( 0) 푉 = 푉 + 푒 휎 + 푚푡 (8.1) 0 √2휋휎2

141

Figure 8.1 An example Gaussian peak fit with parameters for a sample taken during the field campaign (September 2, 2017 at 18:00:00 UTC).

8.3.2 PAN and PPN standards

PAN and PPN were synthesized according to the procedure communicated by J. Roberts

(personal communication, 2009) which are also explained in published work (Williams et al.,

2000 PPN, and MPAN). The standards were characterized by TD-CRDS to determine check for alkyl nitrate and nitric acid impurities. Calibrations were conducted regularly to determine drift in the detector response over the duration of the campaign. A summary of these calibrations is shown in Table 8.1. Calibrations were linear below 3 ppbv, which is approximately 2 times higher than the highest value measured during the campaign.

142

Table 8.1 Summary of response factors from calibrations during PANfire 2017 in mV s ppbv-1. PAN Response PPN Response Date factor factor October 5 365±10

October 6 350±10 465±15

October 13 413±4

October 20 250±10

January 23 255±25

January 26 365±10* *Response factors appeared to be non-linear above ~3 ppbv, so only those points below 3 ppbv were included in this fit

8.4 Auxiliary measurements

The mobile lab was located next to (~10 m) a CRAZ monitoring site, which captured meteorology (i.e., relative humidity, pressure, temperature, wind speed, and wind direction) as well as criteria contaminants such as PM2.5 (particulate matter below 2.5 µm in diameter), NOx, total hydrocarbons (THC), CO, and O3 during the measurement period. During the winter, the

CRAZ trailer was located < 1 km from the rooftop lab. Instrument details including sampling height and concentration range for these measurements can be found at the CRAZ website

(http://craz.ca/monitoring/info-calgary-nw/).

143

8.5 Results

8.5.1 Trailer PAN Measurements

Figure 8.2 depicts a time series of PAN mixing ratios and temperature from August 26,

2017 until October 22, 2017. The days when smoke plumes from BC were in the area range from August 30 – September 10. Fire periods were identified on the basis of enhancements in

PM 2.5 and CO, which are tracers of biomass burning plumes (Aruffo et al., 2016). During the fire period, the background was higher on average than during the times when no forest fires were observed. At this time, PAN mixing ratios never dropped below 60 pptv at night. During the time period when no forest fires were observed, PAN mixing ratios dropped as low as 20 pptv overnight. Throughout the campaign the ambient concentration of PAN never dropped below the LOD (1 pptv). However, PPN regularly dropped below the LOD (2 pptv), in particular during the period when forest fires were not observed. For this reason, PPN was only quantified during the forest fire events (Figure 8.2). Interestingly, the calibrations show that the detector response varied somewhat over the duration of the campaign (Table 8.1) due to reasons that remain unclear but may be related to the hydrogen carrier gas since this detector variability was not observed with the use of helium as a carrier gas (Tokarek et al., 2014).

144

Figure 8.3 A and B shows diurnal profiles of PAN and PPN mixing ratios, respectively, in the trailer. Due to the large effect forest fires had on the PAN profile, the diurnal profile has been split into two profiles, one during the fire period (Figure 8.3A) and one for during the non- fire period (Figure 8.4).

During the fire period, PAN mixing ratios began increasing around 8 am and achieved a maximum mixing ratio around noon. PPN at in this period showed a similar profile, beginning to

145

increase around 8 am and reaching a maximum around 1 pm. During the non-fire period, the

PAN mixing ratio began to increase around 10 am while the maximum PAN mixing ratio was observed around 1 pm. Some of the individual measurements (i.e., the blue squares) don’t align with the maximum as they represent individual events when the maximum PAN mixing ratio occurred at a different time or when a smoke plume affected the site (during the forest fire period).

Figure 8.3 (A) PAN and (B) PPN diurnal profiles during the forest fire period. The blue squares represent individual measurements with the percentiles and median layered on top. 146

8.5.2 Forest Fire impact

Figure 8.5 (top) shows the PPN/PAN ratio time series for the time period when biomass burning impacted the site. The PPN/PAN ratio fluctuated throughout this period from a minimum of 0.046 to a maximum of 0.177 with no obvious dependence on meteorological parameters (i.e., time of day, temperature, relative humidity, wind direction/speed, etc.) consistent with a non-local source. Enhancement (or absence) of other measured pollutants (i.e.,

NOx, CO, PM2.5, O3) did not always occur with changes in the PPN/PAN ratio suggesting multiple sources (i.e., different forest fires). To investigate the potential for multiple sources further, PPN mixing ratio was plotted as a function of PAN mixing ratio and were color coded by

Julian day (Figure 8.6).

147

The data show that different ratios were observed on different days. The highest ratios were observed on September 4th and the lowest were observed on September 8th with the average ratio during this period being 0.104. Typical background air is expected to have a PPN/PAN ratio of ~0.10 (Zaragoza et al., 2017).

148

In Figure 8.7, PPN mixing ratio was plotted as a function of PAN mixing ratio and were

-3 color coded by PM2.5 concentration. The figure shows that PM2.5 concentrations > 22.5 µg m

(the median PM2.5 concentration for the forest fire period) are associated with PPN/PAN ratios

-3 that are higher than background ratios (0.10) and that PM2.5 concentrations > 22.5 µg m are associated with PPN/PAN ratios that are lower than background ratios.

149

8.5.3 Rooftop PAN measurements

Figure 8.8 shows the time series of PAN mixing ratios from the rooftop portion of this work with several other monitored pollutants < 1 km away (at the trailer site) from January 18,

2018 until March 4, 2018. The other pollutants plotted include NOx, O3, PM2.5, CO and THC.

Figure 8.8 shows the background PAN mixing ratio to be high (minimum of ~0.30 excluding a

150

couple outliers on the night of January 30th). This is in contrast with the minimum of 20 pptv observed at the trailer site (section 8.4.1).

Figure 8.8 Time series of PAN mixing ratios while making measurements in the rooftop lab. The yellow and grey background represent day and night, respectively.

Figure 8.9 shows the diurnal profile of PAN mixing ratios during the rooftop part of this work. A PAN mixing ratio daily maximum was determined to occur around 5 pm. This profile has a less pronounced diurnal trend than that observed at the trailer site.

151

Figure 8.9 The diurnal profile of PAN mixing ratios while in the rooftop lab. The blue dots represent individual PAN measurements while the median and percentiles are also shown, layered overtop.

8.6 Discussion

8.6.1 Ambient concentrations and PAN diurnal profiles

Ambient PAN and PPN mixing ratios during the forest fire period followed the expected diurnal profile (Figure 8.2 and 8.3), with a maximum occurring in the middle of the day (around

1 pm) and a minimum occurring overnight. This is in contrast with the diurnal profile in the winter months (Figure 8.9) which was less prominent and had a maximum that occurred later in the day around 5 pm. This lower maximum is expected due to the decrease in solar radiation in the winter (compared with fall) which limits PAN formation. The shift from noon to 5 pm is consistent with an expected decrease in the primary loss mechanism of PAN, which is typically

152

thermal decomposition, due to the colder winter temperatures. In summer, the formation and loss of PAN are occurring simultaneously around noon which results in a sharper, more prominent maximum.

Interestingly, after the forest fires period ended, the maximum still occurred around 1 pm but was much less sharp (Figure 8.4). This could be a coincidence because the forest fires occurred during warmer, summer-like temperatures and after the forest fires the season was beginning to change (decreasing thermal losses as a result). Though, it seems more likely that the sharp peak in PAN concentration during the forest fires occurred due to a well-mixed boundary layer with local formation and transported (from forest fires) PANs contributing to higher overall concentration.

Background mixing ratios were also higher while on the roof in the winter months.

Again, this is most likely because colder temperatures limit thermal decomposition of PAN which occurs in tandem with a reduction in daytime photochemical production due to a seasonal decrease in solar radiation. Conversely, while in the summer, ambient measurements during the non-forest fire periods showed a decrease in PAN mixing ratio back to roughly the same mixing ratio every night as the temperature was still warm enough for PAN to undergo thermal decomposition at night.

153

8.6.2 Air impacted by biomass burning

PAN mixing ratios were greatly increased during the forest fire period. This was expected due to NOx produced by forest fires which can then react to form PANs in the presence of peroxy radicals (an eventual photoxidation product of biomass burning emissions) (Akagi et al., 2012; Akagi et al., 2011). Since this event occurred near the end of the summer, when solar radiation is still somewhat high (relative to winter months), it is possible that some of the observed increase in PAN (and PPN) mixing ratio is due to increased local photochemical production and not only from biomass burning plumes.

The PPN/PAN ratio was shown to fluctuate from as low as 0.046 to as high as 0.177

(Figures 8.5 and 8.6). The concentrations of PPN and PAN were also elevated during times of forest fire impact which is expected as forest fires are known to emit many precursors to PAN

(e.g., ethane, propane, acetone, isoprene, acetaldehyde, etc.) and PPN (propanal) as well as being a source of NOx (Figure 8.5) (Ridley et al., 1990; Roberts, 1990). Typically, the PPN/PAN ratio increases when a source of the PPN precursor (propanal) is present, which most often occurs in urban environments. However, while many studies have measured PAN in biomass burning plumes (Liu et al., 2017; Mauzerall et al., 1998; Yokelson et al., 2013; Alvarado et al., 2010), there are much fewer observations of PPN in biomass burning plumes in the literature (Andreae and Merlet, 2001; Trentmann et al., 2003; Christian et al., 2003; Roberts et al., 2004; Zaragoza et al., 2017).

154

Since PPN is most often a product of anthropogenic emissions, a higher PPN/PAN ratio

(i.e., > 0.10) suggests that the air was in contact with an urban environment. This is in contrast with a low PPN/PAN ratio (i.e., < 0.10) that is usually observed in background continental air.

This is caused by the differences in precursor reactivity (Ridley et al., 1990). The precursor to

PPN is propanal and other higher alkanes (like butane and pentane) which are shorter lived with respect to HO than the precursors for PAN (e.g., methane, propane, acetone, etc.). Over long distances, this results in a faster decrease in the production of PPN compared with PAN and, thus, an overall decrease in their ratio.

Other reactions can also affect this ratio. Peroxy acyl nitrates all have similar thermal dissociation rates, ruling out the possibility that PAN is lost faster thermally than PPN (Schurath and Wippercht, 1980). However, PPN does have a reaction rate with respect to HO greater than an order of magnitude compared with PAN. This is only significant when thermal losses are not as dominant (i.e., colder regions/seasons) and is expected to lower the PPN/PAN ratio (Carter and Atkinson, 1985). Lastly, it is possible that heterogeneous uptake could be playing a role in this chemistry, though few studies have investigated the effect of aerosol composition on the uptake of peroxyacyl nitrates (Zhao et al., 2017).

Figure 8.6 shows that different days during the forest fire period had different PPN/PAN ratios, ranging from about 0.05 to 0.15. This ratio does not appear to have a dependence on NOx.

These ratios were less than those observed by Zaragoza et al. (2017), who observed ratios ranging from 0.15 to 0.20 in the Colorado Front Range, mostly due to the large contributions of

155

anthropogenic VOCs known to be abundant in the region. Because different ratios were observed on different days and there were multiple fires in BC throughout this time, it is expected that the different ratios represent plumes from different fires or from the same fires with varying transit times.

In Figure 8.7, PPN is again plotted as a function of PAN, this time with the color scale

-3 representing PM2.5 concentration. The median PM2.5 concentration (22.5 µg m ) was selected as the middle of the scale and is colored white. Interestingly, when correlating only points at concentrations greater than the median the PPN/PAN ratio is higher than when concentrations less than the median are correlated. In other words, “cleaner”, less polluted (with PM2.5) air exhibits a higher ratio of PPN to PAN. This is important because these ratios are used as markers to point to the cause of ozone production in aged airmasses and high ratios generally point to anthropogenic causes. Further, while high, medium, and low ratios are observed when PM2.5 concentrations are greater than the median, primarily only high ratios were observed below the median PM2.5 concentration, in particular when PAN concentrations are higher (i.e., greater than

~400 pptv). This suggests that PM2.5 may be playing a complex role in PAN chemistry, presumably via heterogenous reactions and uptake. However, more evidence is needed to corroborate this premise. Only one study that we know of has measured uptake coefficients of

PAN on particulates (in this case soot) (Zhao et al., 2017). However, it is unknown to what degree PPN is taken up onto particles.

156

Further, the aerosol phase (organic vs. aqueous) and composition would almost certainly have an impact the uptake coefficient. Several studies have measured Henry’s law constants for

PAN in n-octanol and water (Kames et al., 1991; Sander, 2015; Roberts, 2005). However, few have measured PPN Henry’s law constants and so their relative uptake rates cannot be rigorously compared. One study that compiled all Henry’s law constants for PAN and PPN in water show that the solubility is very similar (PAN = 2.2 - 4.9×10−2 M m-3 Pa-1, PPN = 2.9×10−2 M m-3 Pa-1)

(Sander, 2015). However, one would expect that in n-octanol, PPN would be more soluble than

PAN. It therefore seems reasonable that the liquid water content of aerosol would affect the relative uptake of PANs. Regardless, this is clearly a gap in this field that would provide modelers with important information.

To investigate the sources of the airmasses affecting the PPN/PAN ratio I used back trajectories from HYSPLIT and forward trajectories from firesmoke.ca. Unfortunately, even on the finest resolution the grid size (12 km) was too coarse and the fires too far away to glean any useable information on which fires impacted the site. For example, many of the times when forest fires were observed (i.e., high PM2.5 and CO) the lagrangian models showed that multiple fires had merged or mixed significantly making it impossible to tell which fires dominated and thus determine an approximate age for the airmass.

5. Conclusions

In summary, PAN was observed to follow a diurnal profile with an expected maximum in the early to mid-afternoon. PAN mixing ratios were lower during the winter months than the

157

summer months due to less solar radiation. PAN mixing ratios were also found to increase during times when biomass burning plumes were in the area. These events were found to be identified by PPN/PAN ratios greater than 0.10, while decreased PPN concentrations possibly due to more rapid photochemical losses of propanal and PPN than PAN and its precursors was attributed to days with PPN/PAN ratios less than 0.10. This would vary depending on transit time and active photochemistry. Further, thermal losses will likely not play a role but over large distances photochemical oxidation of PAN and PPN may affect the PPN to PAN ratio by decreasing it.

Lastly, it seems likely that the liquid water content of aerosol would affect the relative uptake rates of PAN and PPN, possibly affecting their ambient concentrations. More work needs to be done to characterize these rates to gain a better understanding of PAN and PPN chemistry.

158

Chapter Nine: Thesis Summary, Conclusions and Future Work

My research has focused on utilizing direct air injection chromatography to measure

VOCs and their oxidation products. Volatile organic compounds and their oxidation products are of great interest to the atmospheric chemistry community for many reasons, in particular, their direct effects on ozone production, human health and aerosol formation as well as their indirect effects on visibility and climate. Because of these effects, it is important that we understand how

VOCs are emitted, how they react and how they are lost in the troposphere. Modifying commercial instrumentation to make these measurements enabled me to take an inventory of particular VOCs in air. These chemical inventories can provide useful information about where an airmass came from and how it has aged. In addition, supplementary measurements allowed me to take my analysis further.

These sensitive techniques allowed me to make speciated trace measurements (pptv to ppbv range) on the time scale of minutes to hours. The development of an inlet system allowed for a preconcentration method free of interfering oxidants. Lastly, the lightweight instrumentation provided a means to transport these measurements to challenging locations.

In this thesis, results from three measurement campaigns are presented. The first campaign was the Fort McMurray Oil Sands Strategic Investigation of Local Sources (FOSSILS

2013) (Tokarek et al., 2018). This work included the measurement of monoterpenes and BTEX compounds in the Athabasca oil sands using the Griffin 450 GC-ITMS (Chapter 2). During this

159

campaign I observed monoterpenes to have a regularly occurring diurnal trend with maximum concentration at night. I also observed high concentrations (maximum = ~ 600 pptv) of o-xylene occurring intermittently when intercepting anthropogenic plumes from industrial activities

(Chapter 3). The most important observation from this work was the presence of a large, analytically unresolved hydrocarbon peak at the end of the chromatograms of intermediate volatility (i.e., C11-C16) which often occurred with enhancements of other anthropogenic species (e.g., aromatics, PAHs, rBC, etc.). However, concentrations varied independently from each other suggesting multiple sources of the observed IVOCs in this region. Post-campaign laboratory measurements of bitumen vapours produced a consistent IVOC profile at the end of the chromatogram, suggesting that the mining and processing of bitumen may be a possible source of these IVOCs.

To elucidate the origin of the IVOCs observed during FOSSILS 2013, a principal component analysis was conducted to deconstruct this data set mathematically (Chapter 4). From this work, three components were identified as possible sources of IVOCs. The component that most correlated with IVOCs was component 5 and was interpreted as surface exposed bitumen.

This component did not exhibit high correlation with many other species with the exception of rBC and a less-oxidized oxygenated organic aerosol. The two other components that were correlated with IVOCs were consistent with tailings ponds (component 1) and the mining vehicle fleet (component 2). This is consistent with the observation that IVOCs are outgassed from bitumen because the mining fleet dig up and transport bitumen to processing facilities and the tailings ponds contained processed bitumen.

160

An assessment of sources would greatly benefit from high mass resolution techniques for measuring VOCs and IVOCs to get a better sense of the molecular formula and oxidation states of these IVOCs throughout different stages in the process of bitumen extraction. In particular,

+ + being able to distinguish between HCO (m/z 29.0022) and C2H5 (m/z 29.0386) would provide unique insight into the level of oxidation. Further, measurements should be made closer to these sources to confirm that IVOCs in this volatility range are indeed emitted.

Despite the poor mass resolution of the Griffin GC-ITMS, some conclusions can be drawn from the spectral information contained in the IVOCs peak (Chapter 4). For example, an image plot of the mass to charge ratio as a function of the retention time shows that the bitumen vapours observed in the lab contain similar ion fragments compared with the IVOCs peak observed in the field. These fragments are consistent with relatively unoxidized hydrocarbons and non-aromatic species. Further, there is little evidence of other elements in these ion fragments (i.e., N, O, S) for both the lab bitumen and the IVOCs observed in the field which is consistent with the current body of knowledge on bitumen composition. However, the GC-ITMS is somewhat blind to oxidized species as they may not adsorb to the tenax preconcentrator or are lost to the analytical column. Future work should address this gap by using instrumentation that can simultaneously observe oxidized and non-oxidized species.

The second field study was the Ozone-depleting Reactions in a Coastal Atmosphere

(ORCA 2015) campaign (Tokarek et al., 2017). In this campaign, our groups’ aim was to investigate nocturnal ozone depletion events that occurred off the west coast of Vancouver Island

161

at a remote monitoring site set up by Environment and Climate Change Canada (Chapter 6).

These ozone depletion events were characterized by an accompanying increase in CO2 but no increase in other anthropogenic species (e.g., NOx, SO2, CO) and along- or onshore flow. The

GC-ITMS was used to quantify monoterpenes and specific ozone oxidation products of the most abundant terpenes (i.e., nopinone and limona ketone) in the region. The aim was to determine if ozone titration by terpenes could explain the large decrease in ozone concentrations at night. The results suggested that terpenes played only a minor role in ozone depletion events and the more likely explanation is that dry deposition is responsible for the sharp decrease in ozone. This is because the air masses generally were in contact with near shore ocean water where kelp forests and terrestrial vegetation are located providing a large surface area for deposition. Further, the aerosol size distributions generally showed limited aerosol growth which would be significant if ozone oxidation of terpenes was the major contributor of the ozone depletion events. Future studies should focus on better characterizing air flow in the region as well as utilize instrumentation that is capable of measuring larger VOCs (i.e., sesquiterpenes) which may play a role in ozone loss in this region.

During the ORCA 2015 campaign, I also observed that air masses with a marine (as opposed to continental) origin tended to be higher in limonene (Chapter 7). The vegetation in the area (e.g., Western Hemlock, Western Red cedar, Coastal Douglas Fir) is known to primarily emit α-pinene and ß-pinene suggesting that limonene may be emitted from other sources in the region (Mason et al., 2015). This led our group to investigate marine vegetation as a possible source of the limonene. We sampled the headspace vapours of 5 kelp species: Fucus gardneri

162

(rock weed), Ulva spp. (sea lettuce), Callophyllis spp. (red sea fans), Alaria marginata (winged kelp), and Nereocystis leutkeana (bull kelp) (Tokarek et al., In review). The results showed that kelp may be a significant source of terpenes, in particular, the highly reactive limonene.

Nereocystis luetkeana and Alaria marginata showed significant enhancement of limonene

(above ambient concentrations and sea water) and are particularly abundant in this region. These results are important as kelp was not recognized as a source of monoterpenes and modelling studies may benefit from including the effects that reactive terpenes may have on ozone production and loss in coastal regions.

How important these terpene emissions from kelp are in coastal regions is still largely uncertain. The next step is to identify how these emissions vary with environmental factors, such as temperature, PAR, plant life cycle and age, and stress factors such as emersion at low tide.

Further, it is not known what the potential role of microbes in BVOC emissions from kelp is and will need to be examined. Emission of limonene is of particular interest since it has a greater potential to form SOA in the presence of O3 than any of the other common monoterpenes.

The last field study discussed in this thesis are local measurements in Calgary in the summer of 2017, called PAN chemistry in forest fires (PANFire) campaign. In this work, a PAN

GC (Chapter 2) was deployed in our groups’ mobile lab through the fall of 2017 and to our rooftop lab in the winter of 2018. The original intent of this campaign was to investigate if hydrogen is a suitable carrier gas for the quantifications of PANs using GC. During the fall, we observed many biomass burning plumes transported from forest fires in western USA and

163

Canada which provided a unique opportunity to study the chemistry of PANs in these conditions.

The research showed that the PPN/PAN ratio, typically used as an indicator for the anthropogenic contribution toward ozone formation, ranged from 0.05 to 0.17 in biomass burning plumes. Of particular interest is an apparent dependence of this ratio on PM2.5 concentration, which has yet to be reported in the literature. While many mechanisms are known to describe the production and loss of PANs (i.e., thermal loss, HO oxidation, precursor abundance), few studies have investigated how PAN uptake onto particles may affect the PPN to

PAN ratio. Future work should investigate the solubility of PANs in n-octanol and characterize the uptake of PANs onto organic aerosol, as this gap in the literature may help to shed light on the wide range of PPN/PAN ratios observed in biomass burning plumes.

Throughout my thesis I have demonstrated the value of using direct air injection gas chromatography for speciated measurements of VOCs. New instrumentation, in particular high- resolution mass spectrometry, is becoming more common today as technology advances.

However, these techniques are unable to distinguish between isomers which may cause researchers to overlook important information. As I have shown, gas chromatography with direct air injection continues to yield new and useful information and should be used in conjunction with newer, high resolution techniques for comprehensive analyses.

164

References

Ahad, J. M. E., Pakdel, H., Savard, M. M., Calderhead, A. I., Gammon, P. R., Rivera, A., Peru, K. M., and Headley, J. V.: Characterization and quantification of mining-related "naphthenic acids" in groundwater near a major oil sands tailings pond, Environ. Sci. Technol., 47, 5023- 5030, 10.1021/es3051313, 2013.

Ajaero, C., Peru, K. M., Simair, M., Friesen, V., O'Sullivan, G., Hughes, S. A., McMartin, D. W., and Headley, J. V.: Fate and behavior of oil sands naphthenic acids in a pilot-scale treatment wetland as characterized by negative-ion electrospray ionization Orbitrap mass spectrometry, Sci. Total Environ., 829-839, 10.1016/j.scitotenv.2018.03.079, 2018.

Akagi, S. K., Yokelson, R. J., Wiedinmyer, C., Alvarado, M. J., Reid, J. S., Karl, T., Crounse, J. D., and Wennberg, P. O.: Emission factors for open and domestic biomass burning for use in atmospheric models, Atmos. Chem. Phys., 11, 4039-4072, 10.5194/acp-11-4039-2011, 2011.

Akagi, S. K., Craven, J. S., Taylor, J. W., McMeeking, G. R., Yokelson, R. J., Burling, I. R., Urbanski, S. P., Wold, C. E., Seinfeld, J. H., Coe, H., Alvarado, M. J., and Weise, D. R.: Evolution of trace gases and particles emitted by a chaparral fire in California, Atmos. Chem. Phys., 12, 1397-1421, 10.5194/acp-12-1397-2012, 2012.

Government of Alberta, oil sands information portal: http://osip.alberta.ca, access: 23-FEB-2017, 2017.

Allen, E. W.: Process water treatment in Canada's oil sands industry: II. A review of emerging technologies, J. Environ. Eng. Sci., 7, 499-524, 10.1139/s08-020, 2008.

Alvarado, M. J., Logan, J. A., Mao, J., Apel, E., Riemer, D., Blake, D., Cohen, R. C., Min, K. E., Perring, A. E., Browne, E. C., Wooldridge, P. J., Diskin, G. S., Sachse, G. W., Fuelberg, H., Sessions, W. R., Harrigan, D. L., Huey, G., Liao, J., Case-Hanks, A., Jimenez, J. L., Cubison, M. J., Vay, S. A., Weinheimer, A. J., Knapp, D. J., Montzka, D. D., Flocke, F. M., Pollack, I. B., Wennberg, P. O., Kurten, A., Crounse, J., Clair, J. M. S., Wisthaler, A., Mikoviny, T., Yantosca, R. M., Carouge, C. C., and Le Sager, P.: Nitrogen oxides and PAN in plumes from boreal fires during ARCTAS-B and their impact on ozone: an integrated analysis of aircraft and satellite observations, Atmos. Chem. Phys., 10, 9739-9760, 10.5194/acp-10-9739-2010, 2010.

Ambrose, J. L., Reidmiller, D. R., and Jaffe, D. A.: Causes of high O3 in the lower free troposphere over the Pacific Northwest as observed at the Mt. Bachelor Observatory, Atmos. Environ., 45, 5302-5315, 10.1016/j.atmosenv.2011.06.056, 2011.

Andreae, M. O., and Raemdonck, H.: Dimethyl sulfide in the surface ocean and the marine atmosphere - a global view, Science, 221, 744-747, 10.1126/science.221.4612.744, 1983.

165

Andreae, M. O., and Merlet, P.: Emission of trace gases and aerosols from biomass burning, Glob. Biogeochem. Cycle, 15, 955-966, 10.1029/2000gb001382, 2001.

Angioni, A., Barra, A., Cereti, E., Barile, D., Coïsson, J. D., Arlorio, M., Dessi, S., Coroneo, V., and Cabras, P.: Chemical composition, plant genetic differences, antimicrobial and antifungal activity investigation of the essential oil of rosmarinus officinalis L, J. Agric. Food Chem., 52, 3530-3535, 10.1021/jf049913t, 2004.

Angioni, A., Barra, A., Coroneo, V., Dessi, S., and Cabras, P.: Chemical composition, seasonal variability, and antifungal activity of lavandula stoechas L. ssp. stoechas essential oils from stem/leaves and flowers, J. Agric. Food Chem., 54, 4364-4370, 10.1021/jf0603329, 2006.

Antrim, L. D., Thom, R. M., Gardiner, W. W., Cullinan, V. I., Shreffler, D. K., and Bienert, R. W.: Effects of petroleum-products on bull kelp (Nereocystis-luetkeana), Mar. Biol., 122, 23- 31, 10.1007/bf00349274, 1995.

Apel, E. C., Calvert, J. G., Gilpin, T. M., Fehsenfeld, F. C., Parrish, D. D., and Lonneman, W. A.: The nonmethane hydrocarbon intercomparison experiment (NOMHICE): task 3, J. Geophys. Res., 104, 26,069–026,086., 10.1029/1999jd900793, 1999.

Arnold, B., Kelly, L., Oleske, J. B., and Schill, A.: Standoff detection of nitrotoluenes using 213- nm amplified spontaneous emission from nitric oxide, Anal. Bioanal. Chem., 395, 349-355, 10.1007/s00216-009-2990-7, 2009.

Arnts, R. R.: Evaluation of adsorbent sampling tube materials and Tenax-TA for analysis of volatile biogenic organic compounds, Atmos. Environ., 44, 1579-1584, 10.1016/j.atmosenv.2010.01.004, 2010.

Aruffo, E., Biancofiore, F., Di Carlo, P., Busilacchio, M., Verdecchia, M., Tomassetti, B., Dari- Salisburgo, C., Giammaria, F., Bauguitte, S., Lee, J., Moller, S., Hopkins, J., Punjabi, S., Andrews, S. J., Lewis, A. C., Palmer, P. I., Hyer, E., Le Breton, M., and Percival, C.: Impact of biomass burning emission on total peroxy nitrates: fire plume identification during the BORTAS campaign, Atmos. Meas. Tech., 9, 5591-5606, 10.5194/amt-9-5591-2016, 2016.

Ball, S. M., Hollingsworth, A. M., Humbles, J., Leblanc, C., Potin, P., and McFiggans, G.: Spectroscopic studies of molecular iodine emitted into the gas phase by seaweed, Atmos. Chem. Phys., 10, 6237-6254, 10.5194/acp-10-6237-2010, 2010.

Baray, S., Darlington, A., Gordon, M., Hayden, K. L., Leithead, A., Li, S. M., Liu, P. S. K., Mittermeier, R. L., Moussa, S. G., O'Brien, J., Staebler, R., Wolde, M., Worthy, D., and McLaren, R.: Quantification of methane sources in the Athabasca Oil Sands Region of Alberta by aircraft mass balance, Atmos. Chem. Phys., 18, 7361-7378, 10.5194/acp-18-7361- 2018, 2018.

166

Barthelmie, R. J., and Pryor, S. C.: A model mechanism to describe oxidation of monoterpenes leading to secondary organic aerosol - 1. alpha-pinene and beta-pinene, J. Geophys. Res.- Atmos., 104, 23657-23669, 10.1029/1999JD900382, 1999.

Bednar, A. J., Russell, A. L., Georgian, T., Splichal, D., Hayes, C. A., Tackett, P., Jones, W. T., Justes, D., Parker, L., Kirgan, R. A., and Wells, M.: Field-portable gas chromatograph mass spectrometer (GC-MS) unit for semi-volatile compound analysis in groundwater, Department of Defense's Environmental Research Programs, 2011.

Bell, R. J., Davey, N. G., Martinsen, M., Short, R. T., Gill, C. G., and Krogh, E. T.: The effect of the earth's and stray magnetic fields on mobile mass spectrometer systems, J. Am. Soc. Mass Spectrom., 26, 201-211, 10.1007/s13361-014-1027-4, 2015.

Benedetti, P., Guerriero, E., Mosca, S., and Rotatori, M.: Analysis of polychlorodibenzo-p- dioxins and polychlorodibenzofurans in stationary source emissions in GC-MS/MS using hydrogen as the carrier gas, J. Sep. Sci., 40, 3469-3478, 10.1002/jssc.201700026, 2017.

Bernstein, D. M.: Increased mortality in COPD among construction workers exposed to inorganic dust, Eur. Resp. J., 24, 512-512, 10.1183/09031936.04.00044504, 2004.

Bond, T. C., Streets, D. G., Yarber, K. F., Nelson, S. M., Woo, J. H., and Klimont, Z.: A technology-based global inventory of black and organic carbon emissions from combustion, J. Geophys. Res.-Atmos., 109, 1-43, 10.1029/2003JD003697, 2004.

Bonsang, B., Gros, V., Peeken, I., Yassaa, N., Bluhm, K., Zoellner, E., Sarda-Esteve, R., and Williams, J.: Isoprene emission from phytoplankton monocultures: the relationship with chlorophyll-a, cell volume and carbon content, Environ. Chem., 7, 554-563, 10.1071/EN09156, 2010.

Bravo-Linares, C. M., Mudge, S. M., and Loyola-Sepulveda, R. H.: Production of volatile organic compounds (VOCs) by temperature macroalgae. The use of solid phase microextraction (SPME) coupled to GC-MS as method of analysis, J. Chil. Chem. Soc., 55, 227-232, 10.4067/S0717-97072010000200018, 2010.

Briggs, N. L., and Long, C. M.: Critical review of black carbon and elemental carbon source apportionment in Europe and the United States, Atmos. Environ., 144, 409-427, 10.1016/j.atmosenv.2016.09.002, 2016.

Broadgate, W. J., Liss, P. S., and Penkett, S. A.: Seasonal emissions of isoprene and other reactive hydrocarbon gases from the ocean, Geophys. Res. Lett., 24, 2675-2678, 10.1029/97gl02736, 1997.

167

Broadgate, W. J., Malin, G., Kupper, F. C., Thompson, A., and Liss, P. S.: Isoprene and other non-methane hydrocarbons from seaweeds: a source of reactive hydrocarbons to the atmosphere, Mar. Chem., 88, 61-73, 10.1016/j.marchem.2004.03.002, 2004.

Burney, O. T., and Jacobs, D. F.: Terpene production and growth of three Pacific Northwest conifers in response to simulated browse and nutrient availability, Trees, 26, 1331-1342, 10.1007/s00468-012-0709-4, 2012.

Burtscher, H., Scherrer, L., Siegmann, H. C., Schmidtott, A., and Federer, B.: Probing aerosols by photoelectric charging, J. Appl. Phys., 53, 3787-3791, 10.1063/1.331120, 1982.

Bytnerowicz, A., Fraczek, W., Schilling, S., and Alexander, D.: Spatial and temporal distribution of ambient nitric acid and ammonia in the Athabasca Oil Sands Region, Alberta, J. Limnol., 69, 11-21, 10.3274/jl10-69-s1-03, 2010.

Carpenter, L. J., and Liss, P. S.: On temperate sources of bromoform and other reactive organic bromine gases, J. Geophys. Res.-Atmos., 105, 20539-20547, 10.1029/2000jd900242, 2000.

Carpenter, L. J., Archer, S. D., and Beale, R. l.: Ocean-atmosphere trace gas exchange, Chem. Soc. Rev., 41, 6473-6506, 10.1039/c2cs35121h, 2012.

Carpenter, L. J., MacDonald, S. M., Shaw, M. D., Kumar, R., Saunders, R. W., Parthipan, R., Wilson, J., and Plane, J. M. C.: Atmospheric iodine levels influenced by sea surface emissions of inorganic iodine, Nat. Geosci., 6, 108-111, 10.1038/NGEO1687, 2013.

Carslaw, D. C., and Ropkins, K.: openair - An R package for air quality data analysis, Environ. Modell. Softw., 27-28, 52-61, 10.1016/j.envsoft.2011.09.008, 2012.

Carslaw, D. C., and Beevers, S. D.: Characterising and understanding emission sources using bivariate polar plots and k-means clustering, Environ. Modell. Softw., 40, 325-329, 10.1016/j.envsoft.2012.09.005, 2013.

Carter, W. P. L., and Atkinson, R.: Atmospheric chemistry of alkanes, J. Atmos. Chem., 3, 377- 405, 10.1007/bf00122525, 1985.

Chalaturnyk, R. J., Don Scott, J., and Ozum, B.: Management of oil sands tailings, Petrol. Sci. Technol., 20, 1025-1046, 10.1081/lft-120003695, 2002.

Charlson, R. J., Schwartz, S. E., Hales, J. M., Cess, R. D., Coakley, J. A., Hansen, J. E., and Hofmann, D. J.: Climate forcing by anthropogenic aerosols, Science, 255, 423-430, 10.1126/science.255.5043.423 1992.

Chen, H., Karion, A., Rella, C. W., Winderlich, J., Gerbig, C., Filges, A., Newberger, T., Sweeney, C., and Tans, P. P.: Accurate measurements of carbon monoxide in humid air using 168

the cavity ring-down spectroscopy (CRDS) technique, Atmos. Meas. Tech., 6, 1031-1040, 10.5194/amt-6-1031-2013, 2013.

Christian, T. J., Kleiss, B., Yokelson, R. J., Holzinger, R., Crutzen, P. J., Hao, W. M., Saharjo, B. H., and Ward, D. E.: Comprehensive laboratory measurements of biomass-burning emissions: 1. Emissions from Indonesian, African, and other fuels, J. Geophys. Res.-Atmos., 108, 13, 4719, 10.1029/2003jd003704, 2003.

Ciuraru, R., Fine, L., Pinxteren, M. v., D’Anna, B., Herrmann, H., and George, C.: Unravelling new processes at interfaces: photochemical isoprene production at the sea surface, Environ. Sci. Technol., 49, 13199-13205, 10.1021/acs.est.5b02388, 2015.

Clark, E. L., Huber, D. P. W., and Carroll, A. L.: The legacy of attack: implications of high phloem resin monoterpene levels in lodgepole pines following mass attack by mountain pine beetle, Environ. Entomol., 41, 392-398, 10.1603/en11295, 2012.

Copeland, N., Cape, J. N., Nemitz, E., and Heal, M. R.: Volatile organic compound speciation above and within a Douglas fir forest, Atmos. Environ., 94, 86-95, 10.1016/j.atmosenv.2014.04.035, 2014.

Cross, E. S., Hunter, J. F., Carrasquillo, A. J., Franklin, J. P., Herndon, S. C., Jayne, J. T., Worsnop, D. R., Miake-Lye, R. C., and Kroll, J. H.: Online measurements of the emissions of intermediate-volatility and semi-volatile organic compounds from aircraft, Atmos. Chem. Phys., 13, 7845-7858, 10.5194/acp-13-7845-2013, 2013. de Gouw, J. A., Middlebrook, A. M., Warneke, C., Ahmadov, R., Atlas, E. L., Bahreini, R., Blake, D. R., Brock, C. A., Brioude, J., Fahey, D. W., Fehsenfeld, F. C., Holloway, J. S., Le Henaff, M., Lueb, R. A., McKeen, S. A., Meagher, J. F., Murphy, D. M., Paris, C., Parrish, D. D., Perring, A. E., Pollack, I. B., Ravishankara, A. R., Robinson, A. L., Ryerson, T. B., Schwarz, J. P., Spackman, J. R., Srinivasan, A., and Watts, L. A.: Organic aerosol formation downwind from the Deepwater Horizon oil spill, Science, 331, 1295-1299, 10.1126/science.1200320, 2011.

Dinerman, E., Dubowski, Y., and Friedler, E.: Fuel derived pollutants and boating activity patterns in the Sea of Galilee, J. Environ. Manage., 92, 3002-3010, 10.1016/j.jenvman.2011.07.015, 2011.

Dixneuf, S., Ruth, A. A., Vaughan, S., Varma, R. M., and Orphal, J.: The time dependence of molecular iodine emission from Laminaria digitata, Atmos. Chem. Phys., 9, 823-829, 10.5194/acp-9-823-2009, 2009.

Trends in atmospheric methane: www.esrl.noaa.gov/gmd/ccgg/trends_ch4/, access: April 11, 2017, 2017a.

169

Trends in atmospheric carbon dioxide: www.esrl.noaa.gov/gmd/ccgg/trends/, access: April 11, 2017, 2017b.

Atmospheric Carbon Dioxide Dry Air Mole Fractions from the NOAA ESRL Carbon Cycle Cooperative Global Air Sampling Network, 1968-2015, Version: 2016-08-30: ftp://aftp.cmdl.noaa.gov/data/trace_gases/co2/flask/surface/, 2016.

Donahue, N. M., Tischuk, J. E., Marquis, B. J., and Hartz, K. E. H.: Secondary organic aerosol from limona ketone: insights into terpene ozonolysis via synthesis of key intermediates, Phys. Chem. Chem. Phys., 9, 2991-2998, 10.1039/b701333g, 2007.

Draper, D. C., Farmer, D. K., Desyaterik, Y., and Fry, J. L.: A qualitative comparison of secondary organic aerosol yields and composition from ozonolysis of monoterpenes at varying concentrations of NO2, Atmos. Chem. Phys., 15, 12267-12281, 10.5194/acp-15- 12267-2015, 2015.

Draxler, R. R., and Hess, G. D.: An overview of the HYSPLIT_4 modelling system for trajectories, dispersion and deposition, Aust. Meteorol. Mag., 47, 295-308, 1998.

HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) Model access via NOAA ARL READY Website http://ready.arl.noaa.gov/HYSPLIT.php, 2011.

Drewitt, G. B., Curren, K., Steyn, D. G., Gillespie, T. J., and Niki, H.: Measurement of biogenic hydrocarbon emissions from vegetation in the Lower Fraser Valley, British Columbia, Atmos. Environ., 32, 3457-3466, 10.1016/S1352-2310(98)00043-0, 1998.

Duhl, T. R., Helmig, D., and Guenther, A.: Sesquiterpene emissions from vegetation: a review, Biogeosciences, 5, 761-777, 10.5194/bg-5-761-2008, 2008.

National pollutant release inventory (NPRI): http://open.canada.ca/data/en/dataset/e40099ae- b116-4c48-9475-f3806fe5a6a6, access: October 5, 2016, 2013.

Ehn, M., Thornton, J. A., Kleist, E., Sipila, M., Junninen, H., Pullinen, I., Springer, M., Rubach, F., Tillmann, R., Lee, B., Lopez-Hilfiker, F., Andres, S., Acir, I.-H., Rissanen, M., Jokinen, T., Schobesberger, S., Kangasluoma, J., Kontkanen, J., Nieminen, T., Kurten, T., Nielsen, L. B., Jorgensen, S., Kjaergaard, H. G., Canagaratna, M., Maso, M. D., Berndt, T., Petaja, T., Wahner, A., Kerminen, V.-M., Kulmala, M., Worsnop, D. R., Wildt, J., and Mentel, T. F.: A large source of low-volatility secondary organic aerosol, Nature, 506, 476-479, 10.1038/nature13032, 2014.

Ekdahl, A., Pedersen, M., and Abrahamsson, K.: A study of the diurnal variation of biogenic volatile halocarbons, Mar. Chem., 63, 1-8, 10.1016/s0304-4203(98)00047-4, 1998.

170

Eneh, O. C.: A review on petroleum: source, uses, processing, products and the environment, J. appl. sci., 11, 2084-2091, 10.3923/jas.2011.2084.2091, 2011.

Englander, J. G., Bharadwaj, S., and Brandt, A. R.: Historical trends in greenhouse gas emissions of the Alberta oil sands (1970-2010), Environ. Res. Lett., 8, 10.1088/1748-9326/8/4/044036, 2013.

Data from the 2011 National Emissions Inventory: https://cfpub.epa.gov/roe/, 2014.

Exton, D. A., McGenity, T. J., Steinke, M., Smith, D. J., and Suggett, D. J.: Uncovering the volatile nature of tropical coastal marine ecosystems in a changing world, Glob. Change Biol., 21, 1383-1394, 10.1111/gcb.12764, 2015.

Falkowski, P. G., Barber, R. T., and Smetacek, V.: Biogeochemical controls and feedbacks on ocean primary production, Science, 281, 200-206, 10.1126/science.281.5374.200, 1998.

Fan, J. W., and Zhang, R. Y.: Atmospheric oxidation mechanism of isoprene, Environmental Chemistry, 1, 140-149, 10.1071/en04045, 2004.

Fan, S.-M., Wofsy, S. C., Bakwin, P. S., Jacob, D. J., and Fitzjarrald, D. R.: Atmosphere- biosphere exchange of CO2 and O3 in the central Amazon Forest, J. Geophys. Res. Atmos., 95, 16851-16864, 10.1029/JD095iD10p16851, 1990.

Fares, S., Schnitzhofer, R., Jiang, X., Guenther, A., Hansel, A., and Loreto, F.: Observations of diurnal to weekly variations of monoterpene-dominated fluxes of volatile organic compounds from mediterranean forests: implications for regional modeling, Environ. Sci. Technol., 47, 11073-11082, 10.1021/es4022156, 2013.

Fenn, M. E., Bytnerowicz, A., Schilling, S. L., and Ross, C. S.: Atmospheric deposition of nitrogen, sulfur and base cations in jack pine stands in the Athabasca Oil Sands Region, Alberta, Canada, Environ. Pollut., 196, 497-510, 10.1016/j.envpol.2014.08.023, 2015.

Finlayson-Pitts, B. J., and Pitts, J. N.: Chemistry of the upper and lower atmosphere: theory, experiments, and applications, Academic Press, San Diego, Calif., 969 pp., 2000.

Fioletov, V. E., McLinden, C. A., Cede, A., Davies, J., Mihele, C., Netcheva, S., Li, S. M., and O'Brien, J.: Sulfur dioxide (SO2) vertical column density measurements by Pandora spectrometer over the Canadian oil sands, Atmos. Meas. Tech., 9, 2961-2976, 10.5194/amt- 9-2961-2016, 2016.

Fiorini, D., and Boarelli, M. C.: Solid-phase microextraction may catalize hydrogenation when using hydrogen as carrier in gas chromatography, J. Chromatogr. A, 1453, 134-137, 10.1016/j.chroma.2016.05.045, 2016.

171

Fischer, E. V., Jaffe, D. A., Reidmiller, D. R., and Jaegle, L.: Meteorological controls on observed peroxyacetyl nitrate at Mount Bachelor during the spring of 2008, J. Geophys. Res., 115, D03302, 10.1029/2009JD012776, 2010.

Fischer, E. V., Jacob, D. J., Yantosca, R. M., Sulprizio, M. P., Millet, D. B., Mao, J., Paulot, F., Singh, H. B., Roiger, A., Ries, L., Talbot, R. W., Dzepina, K., and Pandey Deolal, S.: Atmospheric peroxyacetyl nitrate (PAN): a global budget and source attribution, Atmos. Chem. Phys., 14, 2679-2698, 10.5194/acp-14-2679-2014, 2014.

Flocke, F. M., Weinheimer, A. J., Swanson, A. L., Roberts, J. M., Schmitt, R., and Shertz, S.: On the measurement of PANs by gas chromatography and electron capture detection, J. Atmos. Chem., 52, 19-43, 10.1007/s10874-005-6772-0, 2005.

Fuentes, J. D., Wang, D., Neumann, H. H., Gillespie, T. J., DenHartog, G., and Dann, T. F.: Ambient biogenic hydrocarbons and isoprene emissions from a mixed deciduous forest, J. Atmos. Chem., 25, 67-95, 10.1007/BF00053286, 1996.

Fujita, E. M., Campbell, D. E., Arnott, W. P., Johnson, T., and Ollison, W.: Concentrations of mobile source air pollutants in urban microenvironments, J. Air Waste Manage. Assoc., 64, 743-758, 10.1080/10962247.2013.872708, 2014.

Furimsky, E.: Emissions of carbon dioxide from tar sands plants in Canada, Energy Fuels, 17, 1541-1548, 10.1021/ef0301102, 2003.

Galbally, I. E., Bentley, S. T., and Meyer, C. P.: Mid-latitude marine boundary-layer ozone destruction at visible sunrise observed at Cape Grim, Tasmania, 41°S, Geophys. Res. Lett., 27, 3841-3844, 10.1029/1999gl010943, 2000.

Gallagher, M. W., Beswick, K. M., and Coe, H.: Ozone deposition to coastal waters, Q. J. R. Meteorol. Soc., 127, 539-558, 10.1002/qj.49712757215, 2001a.

Gallagher, M. W., Beswick, K. M., McFiggans, G., Coe, H., and Choularton, T. W.: Ozone dry deposition velocities for coastal waters, Water. Air. Soil. Poll., 1, 233-242, 10.1023/a:1013119524952, 2001b.

Gantt, B., Meskhidze, N., and Kamykowski, D.: A new physically-based quantification of marine isoprene and primary organic aerosol emissions, Atmos. Chem. Phys., 9, 4915-4927, 10.5194/acp-9-4915-2009, 2009.

Ganzeveld, L., Helmig, D., Fairall, C. W., Hare, J., and Pozzer, A.: Atmosphere-ocean ozone exchange: A global modeling study of biogeochemical, atmospheric, and waterside turbulence dependencies, Glob. Biogeochem. Cycle, 23, GB4021, 10.1029/2008gb003301, 2009.

172

Garcia-Jimenez, P., Brito-Romano, O., and Robaina, R. R.: Production of volatiles by the red seaweed Gelidium arbuscula (Rhodophyta): emission of ethylene and dimethyl sulfide, J. Phycol., 49, 661-669, 10.1111/jpy.12083, 2013.

Geron, C., Rasmussen, R., Arnts, R. R., and Guenther, A.: A review and synthesis of monoterpene speciation from forests in the United States, Atmos. Environ., 34, 1761-1781, 10.1016/S1352-2310(99)00364-7, 2000.

Ghosh, B., Papanastasiou, D. K., Talukdar, R. K., Roberts, J. M., and Burkholder, J. B.: Nitryl 3 chloride (ClNO2): UV/Vis absorption spectrum between 210 and 296 K and O( P) quantum yield at 193 and 248 nm, J. Phys. Chem. A, 116, 5796-5805, 10.1021/jp207389y, 2011.

Giese, B., Laturnus, F., Adams, F. C., and Wiencke, C.: Release of volatile iodinated C1−C4 hydrocarbons by marine macroalgae from various climate zones, Environ. Sci. Technol., 33, 2432-2439, 10.1021/es980731n, 1999.

Gill, K. J., and Hites, R. A.: Rate constants for the gas-phase reactions of the hydroxyl radical with isoprene, α- and ß-pinene, and limonene as a function of temperature, J. Phys. Chem. A, 106, 2538-2544, 10.1021/jp013532q, 2002.

Glozman, E. P., and Akhmetova, R. S.: Chemical composition of bitumen components, Chem. Tech. Fuels. Oil., 6, 364-367, 10.1007/bf01171681, 1970.

Goldstein, A. H., and Galbally, I. E.: Known and unexplored organic constituents in the earth's atmosphere, Environ. Sci. Technol., 41, 1514-1521, 10.1021/es072476p, 2007.

Gordon, M., Li, S. M., Staebler, R., Darlington, A., Hayden, K., O'Brien, J., and Wolde, M.: Determining air pollutant emission rates based on mass balance using airborne measurement data over the Alberta oil sands operations, Atmos. Meas. Tech., 8, 3745-3765, 10.5194/amt- 8-3745-2015, 2015.

Goulitquer, S., Ritter, A., Thomas, F., Ferec, C., Salaun, J. P., and Potin, P.: Release of volatile aldehydes by the brown algal kelp Laminaria digitata in response to both biotic and abiotic stress, ChemBioChem, 10, 977-982, 10.1002/cbic.200900004, 2009.

Graiff, A., Bartsch, I., Ruth, W., Wahl, M., and Karsten, U.: Season exerts differential effects of ocean acidification and warming on growth and carbon metabolism of the seaweed Fucus vesiculosus in the Western Baltic Sea, Front. Mar. Sci., 2, 112, 10.3389/fmars.2015.00112, 2015.

Griffin, R. J., Cocker, D. R., Flagan, R. C., and Seinfeld, J. H.: Organic aerosol formation from the oxidation of biogenic hydrocarbons, J. Geophys. Res., 104, 3555-3567, 10.1029/1998jd100049, 1999.

173

Grimmer, G., Brune, H., Deutschwenzel, R., Dettbarn, G., Jacob, J., Naujack, K. W., Mohr, U., and Ernst, H.: Contribution of polycyclic aromatic-hydrocarbons and nitro-derivatives to the carcinogenic impact of diesel-engine exhaust condensate evaluated by implantation into the lungs of rats, Cancer Lett., 37, 173-180, 10.1016/0304-3835(87)90160-1, 1987.

Gross, J. H.: Mass spectrometry: a textbook, 2nd ed., edited by: Springer-Verlag, Berlin, 734 pp., 2011.

Gruber, N., Gloor, M., Mikaloff Fletcher, S. E., Doney, S. C., Dutkiewicz, S., Follows, M. J., Gerber, M., Jacobson, A. R., Joos, F., Lindsay, K., Menemenlis, D., Mouchet, A., Müller, S. A., Sarmiento, J. L., and Takahashi, T.: Oceanic sources, sinks, and transport of atmospheric CO2, Glob. Biogeochem. Cycle, 23, GB1005, 10.1029/2008gb003349, 2009.

Guenther, A. B., Jiang, X., Heald, C. L., Sakulyanontvittaya, T., Duhl, T., Emmons, L. K., and Wang, X.: The model of emissions of gases and aerosols from nature version 2.1 (MEGAN2.1): an extended and updated framework for modeling biogenic emissions, Geosci. Model Dev., 5, 1471-1492, 10.5194/gmd-5-1471-2012, 2012.

Guo, H., Wang, T., and Louie, P. K. K.: Source apportionment of ambient non-methane hydrocarbons in Hong Kong: Application of a principal component analysis/absolute principal component scores (PCA/APCS) receptor model, Environ. Pollut., 129, 489-498, 10.1016/j.envpol.2003.11.006, 2004.

Harold, F. H.: A backpack-portable mass spectrometer for measurement of volatile compounds in the environment, Rev Sci Instrum., 62, 1420-1425, 10.1063/1.1142461, 1991.

Hatakeyama, S., Izumi, K., Fukuyama, T., and Akimoto, H.: Reactions of ozone with alpha- pinene and beta-pinene in air: Yields of gaseous and particulate products, J. Geophys. Res. Atmos., 94, 13013-13024, 10.1029/JD094iD10p13013, 1989.

Hay, M. E., and Fenical, W.: Marine plant-herbivore interactions: The ecology of chemical defense, Annu. Rev. Ecol. Syst., 19, 111-145, 10.1146/annurev.es.19.110188.000551, 1988.

Hellén, H., Kuronen, P., and Hakola, H.: Heated stainless steel tube for ozone removal in the ambient air measurements of mono- and sesquiterpenes, Atmos. Environ., 57, 35-40, 10.1016/j.atmosenv.2012.04.019, 2012.

Helmig, D., Klinger, L. F., Guenther, A., Vierling, L., Geron, C., and Zimmerman, P.: Biogenic volatile organic compound emissions (BVOCs) I. Identifications from three continental sites in the U.S, Chemosphere, 38, 2163-2187, 10.1016/S0045-6535(98)00425-1, 1999.

Helmig, D., Daly, R. W., Milford, J., and Guenther, A.: Seasonal trends of biogenic terpene emissions, Chemosphere, 93, 35-46, 10.1016/j.chemosphere.2013.04.058, 2013.

174

Hemberger, P. H., Alarid, J. E., Cameron, D., Leibman, C. P., Cannon, T. M., Wolf, M. A., and Kaiser, R. E.: A transportable turnkey gas chromatography ion trap detector for field analysis of environmental samples, Int. J. Mass Spectrom. Ion Process., 106, 299-313, 10.1016/0168- 1176(91)85025-H, 1991.

Hewitt, C. N., Ashworth, K., Boynard, A., Guenther, A., Langford, B., MacKenzie, A. R., Misztal, P. K., Nemitz, E., Owen, S. M., Possell, M., Pugh, T. A. M., Ryan, A. C., and Wild, O.: Ground-level ozone influenced by circadian control of isoprene emissions, Nature Geoscience, 4, 671-674, 10.1038/ngeo1271, 2011.

Hofzumahaus, A., Kraus, A., and Muller, M.: Solar actinic flux spectroradiometry: a technique for measuring photolysis frequencies in the atmosphere, Appl. Optics, 38, 4443-4460, 10.1364/AO.38.004443, 1999.

Holowenko, F. M., MacKinnon, M. D., and Fedorak, P. M.: Methanogens and sulfate-reducing bacteria in oil sands fine tailings waste, Can. J. Microbiol., 46, 927-937, 10.1139/cjm-46-10- 927, 2000.

Horvath, S. M., Bedi, J. F., and Drechslerparks, D. M.: Effects of peroxyacetyl nitrate alone and in combination with ozone in healthy-young women, J. Air Pollut. Control. Assoc., 36, 265- 270, 10.1080/00022470.1986.10466067, 1986.

Howell, S. G., Clarke, A. D., Freitag, S., McNaughton, C. S., Kapustin, V., Brekovskikh, V., Jimenez, J. L., and Cubison, M. J.: An airborne assessment of atmospheric particulate emissions from the processing of Athabasca oil sands, Atmos. Chem. Phys., 14, 5073-5087, 10.5194/acp-14-5073-2014, 2014.

Hsu, Y. M., Bytnerowicz, A., Fenn, M. E., and Percy, K. E.: Atmospheric dry deposition of sulfur and nitrogen in the Athabasca Oil Sands Region, Alberta, Canada, Sci. Total Environ., 568, 285-295, 10.1016/j.scitotenv.2016.05.205, 2016.

Huang, R. J., Seitz, K., Buxmann, J., Pohler, D., Hornsby, K. E., Carpenter, L. J., Platt, U., and Hoffmann, T.: In situ measurements of molecular iodine in the marine boundary layer: the link to macroalgae and the implications for O-3, IO, OIO and NOx, Atmos. Chem. Phys., 10, 4823-4833, 10.5194/acp-10-4823-2010, 2010.

Jaffe, D., Price, H., Parrish, D., Goldstein, A., and Harris, J.: Increasing background ozone during spring on the west coast of North America, Geophys. Res. Lett., 30, 1613, 10.1029/2003gl017024, 2003.

James, A. T., and Martin, A. J. P.: Gas-liquid partition chromatography: the separation and micro-estimation of volatile fatty acids from formic acid to dodecanoic acid, Biochem. J., 50, 679-690, 10.1042/bj0500679 1952.

175

Jaoui, M., and Kamens, R. M.: Gaseous and particulate oxidation products analysis of a mixture of alpha-pinene plus beta-pinene/O3/air in the absence of light and alpha-pinene plus beta- pinene/NOx/air in the presence of natural sunlight, J. Atmos. Chem., 44, 259-297, 10.1023/A:1022977427523, 2003.

Jaoui, M., Corse, E., Kleindienst, T. E., Offenberg, J. H., Lewandowski, M., and Edney, E. O.: Analysis of secondary organic aerosol compounds from the photooxidation of d-limonene in the presence of NOx and their detection in ambient PM2.5, Environ. Sci. Technol., 40, 3819- 3828, 10.1021/es052566z, 2006.

Jautzy, J. J., Ahad, J. M. E., Gobeil, C., Smirnoff, A., Barst, B. D., and Savard, M. M.: Isotopic evidence for oil sands petroleum coke in the peace-Athabasca delta, Environ. Sci. Technol., 49, 12062-12070, 10.1021/acs.est.5b03232, 2015.

Jenkin, M. E.: Modelling the formation and composition of secondary organic aerosol from alpha- and beta-pinene ozonolysis using MCM v3, Atmos. Chem. Phys., 4, 1741-1757, 10.5194/acp-4-1741-2004, 2004.

Jimenez, J. L., Canagaratna, M. R., Donahue, N. M., Prevot, A. S. H., Zhang, Q., Kroll, J. H., DeCarlo, P. F., Allan, J. D., Coe, H., Ng, N. L., Aiken, A. C., Docherty, K. S., Ulbrich, I. M., Grieshop, A. P., Robinson, A. L., Duplissy, J., Smith, J. D., Wilson, K. R., Lanz, V. A., Hueglin, C., Sun, Y. L., Tian, J., Laaksonen, A., Raatikainen, T., Rautiainen, J., Vaattovaara, P., Ehn, M., Kulmala, M., Tomlinson, J. M., Collins, D. R., Cubison, M. J., E., Dunlea, J., Huffman, J. A., Onasch, T. B., Alfarra, M. R., Williams, P. I., Bower, K., Kondo, Y., Schneider, J., Drewnick, F., Borrmann, S., Weimer, S., Demerjian, K., Salcedo, D., Cottrell, L., Griffin, R., Takami, A., Miyoshi, T., Hatakeyama, S., Shimono, A., Sun, J. Y., Zhang, Y. M., Dzepina, K., Kimmel, J. R., Sueper, D., Jayne, J. T., Herndon, S. C., Trimborn, A. M., Williams, L. R., Wood, E. C., Middlebrook, A. M., Kolb, C. E., Baltensperger, U., and Worsnop, D. R.: Evolution of organic aerosols in the atmosphere, Science, 326, 1525-1529, 10.1126/science.1180353, 2009.

Johnson, M. R., Crosland, B. M., McEwen, J. D., Hager, D. B., Armitage, J. R., Karimi- Golpayegani, M., and Picard, D. J.: Estimating fugitive methane emissions from oil sands mining using extractive core samples, Atmos. Environ., 144, 111-123, 10.1016/j.atmosenv.2016.08.073, 2016.

Jolliffe, I. T., and Cadima, J.: Principal component analysis: a review and recent developments, Philosophical transactions of the royal society A: Mathematical, physical and engineering sciences, 374, 10.1098/rsta.2015.0202, 2016.

Jones, C. E., Kato, S., Nakashima, Y., and Kajii, Y.: A novel fast gas chromatography method for higher time resolution measurements of speciated monoterpenes in air, Atmos. Meas. Tech., 7, 1259-1275, 10.5194/amt-7-1259-2014, 2014.

176

Kames, J., Schweighoefer, S., and Schurath, U.: Henry's law constant and hydrolysis of peroxyacetyl nitrate (PAN), J. Atmos. Chem., 12, 169-180, 10.1007/bf00115778, 1991.

Kaminski, M., Fuchs, H., Acir, I. H., Bohn, B., Brauers, T., Dorn, H. P., Haseler, R., Hofzumahaus, A., Li, X., Lutz, A., Nehr, S., Rohrer, F., Tillmann, R., Vereecken, L., Wegener, R., and Wahner, A.: Investigation of the beta-pinene photooxidation by OH in the atmosphere simulation chamber SAPHIR, Atmos. Chem. Phys., 17, 6631-6650, 10.5194/acp- 17-6631-2017, 2017.

Kavouras, I. G., Mihalopoulos, N., and Stephanou, E. G.: Formation and gas/particle partitioning of monoterpenes photo-oxidation products over forests, Geophys. Res. Lett., 26, 55-58, 10.1029/1998GL900251, 1999.

Kessler, A., Halitschke, R., Diezel, C., and Baldwin, I. T.: Priming of plant defense responses in nature by airborne signaling between Artemisia tridentata and Nicotiana attenuata, Oecologia, 148, 280-292, 10.1007/s00442-006-0365-8, 2006.

Khan, M. A. H., Jenkin, M. E., Foulds, A., Derwent, R. G., Percival, C. J., and Shallcross, D. E.: A modeling study of secondary organic aerosol formation from sesquiterpenes using the STOCHEM global chemistry and transport model, J. Geophys. Res.-Atmos., 122, 4426-4439, 10.1002/2016jd026415, 2017.

Kindzierski, W. B., and Ranganathan, H. K. S.: Indoor and outdoor SO2 in a community near oil sand extraction and production facilities in northern Alberta, J. Environ. Eng. Sci., 5, S121- S129, 10.1139/s06-022, 2006.

Kleindienst, T. E.: Recent developments in the chemistry and biology of peroxyacetyl nitrate, Res. Chem. Intermed., 20, 335-384, 10.1163/156856794X00379, 1994.

Kleinman, L., Lee, Y.-N., Springston, S. R., Nunnermacker, L., Zhou, X., Brown, R., Hallock, K., Klotz, P., Leahy, D., Lee, J. H., and Newman, L.: Ozone formation at a rural site in the southeastern United States, J. Geophys. Res.-Atmos., 99, 3469-3482, 10.1029/93jd02991, 1994.

Koppmann, R.: Volatile organic compounds in the atmosphere, 1st ed., Blackwell Pub., Oxford ; Ames, Iowa, 2007.

Lagalante, A. F., and Montgomery, M. E.: Analysis of terpenoids from hemlock (Tsuga) species by solid-phase microextraction/gas chromatography/ion-trap mass spectrometry, J. Agric. Food Chem., 51, 2115-2120, 10.1021/jf021028s, 2003.

Larsen, B. R., Di Bella, D., Glasius, M., Winterhalter, R., Jensen, N. R., and Hjorth, J.: Gas- phase OH oxidation of monoterpenes: Gaseous and particulate products, J. Atmos. Chem., 38, 231-276, 10.1023/A:1006487530903, 2001. 177

Laturnus, F.: Volatile halocarbons released from Arctic macroalgae, Mar. Chem., 55, 359-366, 10.1016/s0304-4203(97)89401-7, 1996.

Lavi, A., Vermeuel, M. P., Novak, G. A., and Bertram, T. H.: The sensitivity of benzene cluster cation chemical ionization mass spectrometry to select biogenic terpenes, Atmos. Meas. Tech., 11, 3251-3262, 10.5194/amt-11-3251-2018, 2018.

Lee, J. D., Moller, S. J., Read, K. A., Lewis, A. C., Mendes, L., and Carpenter, L. J.: Year-round measurements of nitrogen oxides and ozone in the tropical North Atlantic marine boundary layer, J. Geophys. Res., 114, D21302, 10.1029/2009jd011878, 2009.

Leedham, E. C., Hughes, C., Keng, F. S. L., Phang, S. M., Malin, G., and Sturges, W. T.: Emission of atmospherically significant halocarbons by naturally occurring and farmed tropical macroalgae, Biogeosciences, 10, 3615-3633, 10.5194/bg-10-3615-2013, 2013.

Lelieveld, J., Evans, J. S., Fnais, M., Giannadaki, D., and Pozzer, A.: The contribution of outdoor air pollution sources to premature mortality on a global scale, Nature, 525, 367-371, 10.1038/nature15371, 2015.

Li, S.-M., Leithead, A., Moussa, S. G., Liggio, J., Moran, M. D., Wang, D., Hayden, K., Darlington, A., Gordon, M., Staebler, R., Makar, P. A., Stroud, C. A., McLaren, R., Liu, P. S. K., O'Brien, J., Mittermeier, R. L., Zhang, J., Marson, G., Cober, S. G., Wolde, M., and Wentzell, J. J. B.: Differences between measured and reported volatile organic compound emissions from oil sands facilities in Alberta, Canada, Proc. Natl. Acad. Sci., 114, 3756- 3765, 10.1073/pnas.1617862114, 2017.

Librando, V., and Tringali, G.: Atmospheric fate of OH initiated oxidation of terpenes. Reaction mechanism of alpha-pinene degradation and secondary organic aerosol formation, J. Environ. Manage., 75, 275-282, 10.1016/j.jenvman.2005.01.001, 2005.

Lide, D. R.: CRC handbook of chemistry and physics: a ready-reference book of chemical and physical data, 79th ed., CRC Press, Boca Raton, FL, 1998.

Liggio, J., Li, S.-M., Hayden, K., Taha, Y. M., Stroud, C., Darlington, A., Drollette, B. D., Gordon, M., Lee, P., Liu, P., Leithead, A., Moussa, S. G., Wang, D., O'Brien, J., Mittermeier, R. L., Brook, J., Lu, G., Staebler, R., Han, Y., Tokarek, T. W., Osthoff, H. D., Makar, P. A., Zhang, J., Plata, D., and Gentner, D. R.: Oil sands operations as a large source of secondary organic aerosols, Nature, 534, 91-94, 10.1038/nature17646, 2016.

Linstrom, P. J., and Mallard, W. G.: The NIST chemistry webBook: a chemical data resource on the internet, J. Chem. Eng. Data, 46, 1059-1063, 10.1021/je000236i, 2001.

Liu, X., Huey, L. G., Yokelson, R. J., Selimovic, V., Simpson, I. J., MÃ¼ller, M., Jimenez, J. L., Campuzano-Jost, P., Beyersdorf, A. J., Blake, D. R., Butterfield, Z., Choi, Y., Crounse, J. D., 178

Day, D. A., Diskin, G. S., Dubey, M. K., Fortner, E., Hanisco, T. F., Hu, W., King, L. E., Kleinman, L., Meinardi, S., Mikoviny, T., Onasch, T. B., Palm, B. B., Peischl, J., Pollack, I. B., Ryerson, T. B., Sachse, G. W., Sedlacek, A. J., Shilling, J. E., Springston, S., St. Clair, J. M., Tanner, D. J., Teng, A. P., Wennberg, P. O., Wisthaler, A., and Wolfe, G. M.: Airborne measurements of western U.S. wildfire emissions: Comparison with prescribed burning and air quality implications, J. Geophys. Res. -Atmos., 122, 6108-6129, 10.1002/2016JD026315, 2017.

Liu, X. X., Zhang, Y., Huey, L. G., Yokelson, R. J., Wang, Y., Jimenez, J. L., Campuzano-Jost, P., Beyersdorf, A. J., Blake, D. R., Choi, Y., St Clair, J. M., Crounse, J. D., Day, D. A., Diskin, G. S., Fried, A., Hall, S. R., Hanisco, T. F., King, L. E., Meinardi, S., Mikoviny, T., Palm, B. B., Peischl, J., Perring, A. E., Pollack, I. B., Ryerson, T. B., Sachse, G., Schwarz, J. P., Simpson, I. J., Tanner, D. J., Thornhill, K. L., Ullmann, K., Weber, R. J., Wennberg, P. O., Wisthaler, A., Wolfe, G. M., and Ziemba, L. D.: Agricultural fires in the southeastern US during SEAC(4)RS: Emissions of trace gases and particles and evolution of ozone, reactive nitrogen, and organic aerosol, J. Geophys. Res.-Atmos., 121, 7383-7414, 10.1002/2016jd025040, 2016.

Logan, J. A.: Ozone in rural areas of the United States, J. Geophys. Res.-Atmos., 94, 8511-8532, 10.1029/JD094iD06p08511, 1989.

Lopez-Perez, O., Picon, A., and Nunez, M.: Volatile compounds and odour characteristics of seven species of dehydrated edible seaweeds, Food Res. Int., 99, 1002-1010, 10.1016/j.foodres.2016.12.013, 2010.

Lovelock, J. E., Maggs, R. J., and Rasmusse.Ra: Atmospheric dimethyl sulfide and natural sulfur cycle, Nature, 237, 452-453, 10.1038/237452a0, 1972.

Lucero, M. E., Fredrickson, E. L., Estell, R. E., Morrison, A. A., and Richman, D. B.: Volatile composition of Gutierrezia sarothrae (Broom Snakeweed) as determined by steam distillation and solid phase microextraction, J. Essent. Oil Res., 18, 121-125, 10.1080/10412905.2006.9699039, 2006.

Luo, G., and Yu, F.: A numerical evaluation of global oceanic emissions of alpha and beta- pinene and isoprene, Atmos. Chem. Phys., 10, 2007-2015, 10.5194/acp-10-2007-2010, 2010.

Luxenhofer, O., Schneider, M., Dambach, M., and Ballschmiter, K.: Semivolatile long chain C6- C17 alkyl nitrates as trace compounds in air, Chemosphere, 33, 393-404, 10.1016/0045- 6535(96)00205-6, 1996.

Mahajan, A. S., Prados-Roman, C., Hay, T. D., Lampel, J., Pöhler, D., Groβmann, K., Tschritter, J., Frieß, U., Platt, U., Johnston, P., Kreher, K., Wittrock, F., Burrows, J. P., Plane, J. M. C., and Saiz-Lopez, A.: Glyoxal observations in the global marine boundary layer, Journal of Geophysical Research, 119, 6160-6169, 10.1002/2013jd021388, 2014. 179

Maia, J. G. S., Andrade, E. H. A., da Silva, A. C. M., Oliveira, J., Carreira, L. M. M., and Araújo, J. S.: Leaf volatile oils from four Brazilian Xylopia species, Flavour Fragr. J., 20, 474-477, 10.1002/ffj.1499, 2005.

Mair, C., and MacFarlane, R. R.: Through the Mackenzie Basin; a narrative of the Athabasca and Peace River treaty expedition of 1899, 1908.

Majchrzak, T., Wojnowski, W., Lubinska-Szczygeł‚, M., Różańska, A., Namieśnik, J., and Dymerski, T.: PTR-MS and GC-MS as complementary techniques for analysis of volatiles: A tutorial review, Anal. Chim. Acta., 10.1016/j.aca.2018.06.056, 2018.

Marey, H. S., Hashisho, Z., Fu, L., and Gille, J.: Spatial and temporal variation in CO over Alberta using measurements from satellites, aircraft, and ground stations, Atmos. Chem. Phys., 15, 3893-3908, 10.5194/acp-15-3893-2015, 2015.

Markovic, M. Z., VandenBoer, T. C., and Murphy, J. G.: Characterization and optimization of an online system for the simultaneous measurement of atmospheric water-soluble constituents in the gas and particle phases, J. Environ. Monit., 14, 1872-1884, 10.1039/c2em00004k, 2012.

Masliyah, J., Zhou, Z. J., Xu, Z., Czarnecki, J., and Hamza, H.: Understanding water-based bitumen extraction from Athabasca oil sands, Can. J. Chem. Eng., 82, 628-654, 10.1002/cjce.5450820403, 2004.

Mason, R. H., Si, M., Li, J., Chou, C., Dickie, R., Toom-Sauntry, D., Pohlker, C., Yakobi- Hancock, J. D., Ladino, L. A., Jones, K., Leaitch, W. R., Schiller, C. L., Abbatt, J. P. D., Huffman, J. A., and Bertram, A. K.: Ice nucleating particles at a coastal marine boundary layer site: correlations with aerosol type and meteorological conditions, Atmos. Chem. Phys., 15, 12547-12566, 10.5194/acp-15-12547-2015, 2015.

Mauzerall, D. L., Logan, J. A., Jacob, D. J., Anderson, B. E., Blake, D. R., Bradshaw, J. D., Heikes, B., Sachse, G. W., Singh, H., and Talbot, B.: Photochemistry in biomass burning plumes and implications for tropospheric ozone over the tropical South Atlantic, J. Geophys. Res.-Atmos., 103, 8401-8423, 10.1029/97JD02612, 1998.

McDonald-Buller, E. C., Allen, D. T., Brown, N., Jacob, D. J., Jaffe, D., Kolb, C. E., Lefohn, A. S., Oltmans, S., Parrish, D. D., Yarwood, G., and Zhang, L.: Establishing policy relevant background (PRB) ozone concentrations in the United States, Environ. Sci. Technol., 45, 9484-9497, 10.1021/es2022818, 2011.

McFiggans, G.: Atmospheric science: Marine aerosols and iodine emissions, Nature, 433, E13, 10.1038/nature03372, 2005.

180

McGraw, G. W., Hemingway, R. W., Ingram, L. L., Canady, C. S., and McGraw, W. B.: Thermal degradation of terpenes: camphene, delta(3)-carene, limonene, and alpha-terpinene, Environ. Sci. Technol., 33, 4029-4033, 10.1021/es9810641, 1999.

McKendry, I. G., Christensen, E., Schiller, C., Vingarzan, R., Macdonald, A. M., and Li, Y.: Low ozone episodes at Amphitrite Point Marine Boundary Layer Observatory, British Columbia, Canada, Atmos. Ocean, 52, 271-280, 10.1080/07055900.2014.910164, 2014.

McLinden, C. A., Fioletov, V., Boersma, K. F., Krotkov, N., Sioris, C. E., Veefkind, J. P., and Yang, K.: Air quality over the Canadian oil sands: A first assessment using satellite observations, Geophys. Res. Lett., 39, 8, 10.1029/2011gl050273, 2012.

Meskhidze, N., and Nenes, A.: Phytoplankton and cloudiness in the Southern Ocean, Science, 314, 1419-1423, 10.1126/science.1131779, 2006.

Meskhidze, N., Sabolis, A., Reed, R., and Kamykowski, D.: Quantifying environmental stress- induced emissions of algal isoprene and monoterpenes using laboratory measurements, Biogeosciences, 12, 637-651, 10.5194/bg-12-637-2015, 2015.

Mevy, J. P., Bessiere, J. M., Rabier, J., Dherbomez, M., Ruzzier, M., Millogo, J., and Viano, J.: Composition and antimicrobial activities of the essential oil of Triumfetta rhomboidea Jacq, Flavour Fragr. J., 21, 80-83, 10.1002/ffj.1511, 2006.

Mielke, L. H., Furgeson, A., and Osthoff, H. D.: Observation of ClNO2 in a mid-continental urban environment, Environ. Sci. Technol., 45, 8889-8896, 10.1021/es201955u, 2011.

Mielke, L. H., Furgeson, A., Odame-Ankrah, C. A., and Osthoff, H. D.: Ubiquity of ClNO2 in the nocturnal boundary layer of Calgary, AB, Canada, Can. J. Chem., 94, 414-423, 10.1139/cjc-2015-0426, 2016.

Miller, S. M., Matross, D. M., Andrews, A. E., Millet, D. B., Longo, M., Gottlieb, E. W., Hirsch, A. I., Gerbig, C., Lin, J. C., Daube, B. C., Hudman, R. C., Dias, P. L. S., Chow, V. Y., and Wofsy, S. C.: Sources of carbon monoxide and formaldehyde in North America determined from high-resolution atmospheric data, Atmos. Chem. Phys., 8, 7673-7696, 10.5194/acp-8- 7673-2008, 2008.

Min, K. E., Pusede, S. E., Browne, E. C., LaFranchi, B. W., and Cohen, R. C.: Eddy covariance fluxes and vertical concentration gradient measurements of NO and NO2 over a ponderosa pine ecosystem: observational evidence for within-canopy chemical removal of NOx, Atmos. Chem. Phys., 14, 5495-5512, 10.5194/acp-14-5495-2014, 2014.

Miyazaki, T., Plotto, A., Goodner, K., and Gmitter, F. G.: Distribution of aroma volatile compounds in tangerine hybrids and proposed inheritance, J. Sci. Food Agric., 91, 449-460, 10.1002/jsfa.4205, 2011. 181

Monson, R. K., and Holland, E. A.: Biospheric trace gas fluxes and their control over tropospheric chemistry, Annu. Rev. Ecol. Evol. Syst., 32, 547-576, 10.1146/annurev.ecolsys.32.081501.114136, 2001.

Moore, R. E.: Volatile compounds from marine-algae, Accounts Chem. Res., 10, 40-47, 10.1021/ar50110a002, 1977.

Mossman, D. J., and Nagy, B.: Solid bitumens: An assessment of their characteristics, genesis, and role in geological processes, Terr. Nova, 8, 114-128, 10.1111/j.1365- 3121.1996.tb00736.x, 1996.

Nara, H., Tanimoto, H., Tohjima, Y., Mukai, H., Nojiri, Y., Katsumata, K., and Rella, C. W.: Effect of air composition (N-2, O-2, Ar, and H2O) on CO2 and CH4 measurement by wavelength-scanned cavity ring-down spectroscopy: calibration and measurement strategy, Atmos. Meas. Tech., 5, 2689-2701, 10.5194/amt-5-2689-2012, 2012.

Naylor, S., Hanke, F. J., Manes, L. V., and Crews, P.: Chemical and Biological Aspects of Marine Monoterpenes, in: Fortschritte der Chemie organischer Naturstoffe / Progress in the Chemistry of Organic Natural Products, Springer Vienna, Vienna, 189-241, 1983.

Neu, U., Kunzle, T., and Wanner, H.: On the relation between ozone storage in the residual layer and daily variation in near-surface ozone concentration — A case study, Bound.-Layer Meteor., 69, 221-247, 10.1007/bf00708857, 1994.

Niinemets, U.: Mild versus severe stress and BVOCs: thresholds, priming and consequences, Trends Plant Sci., 15, 145-153, 10.1016/j.tplants.2009.11.008, 2010.

Nimana, B., Canter, C., and Kumar, A.: Energy consumption and greenhouse gas emissions in upgrading and refining of Canada's oil sands products, Energy, 83, 65-79, 10.1016/j.energy.2015.01.085, 2015a.

Nimana, B., Canter, C., and Kumar, A.: Energy consumption and greenhouse gas emissions in the recovery and extraction of crude bitumen from Canada's oil sands, Appl. Energy, 143, 189-199, 10.1016/j.apenergy.2015.01.024, 2015b.

Nitschke, U., Dixneuf, S., Schmid, M., Ruth, A. A., and Stengel, D. B.: Contribution of living and degrading kelp to coastal iodine fluxes, Mar. Biol., 162, 1727-1738, 10.1007/s00227- 015-2699-4, 2015.

Nnaji, C. N., Williams, K. C., Bishop, J. M., and Verbeck, G. F.: Hydrogen as a GC/MS carrier and buffer gas for use in forensic laboratories, Sci. Justice, 55, 162-167, 10.1016/j.scijus.2015.01.003, 2015.

182

Detailed facility information: http://www.ec.gc.ca/inrp-npri/donnees- data/index.cfm?do=facility_information&lang=En&opt_npri_id=0000002274&opt_report_y ear=2013, access: April 13, 2017, 2013.

O'Dowd, C. D., Jimenez, J. L., Bahreini, R., Flagan, R. C., Seinfeld, J. H., Hameri, K., Pirjola, L., Kulmala, M., Jennings, S. G., and Hoffmann, T.: Marine aerosol formation from biogenic iodine emissions, Nature, 417, 632-636, 10.1038/nature00775, 2002.

O'Dowd, C. D., and Hoffmann, T.: Coastal new particle formation: A review of the current state- of-the-art, Environ. Chem., 2, 245-255, 10.1071/EN05077, 2005.

O'Dowd, C. D., Jimenez, J. L., Bahreini, R., Flagan, R. C., Seinfeld, J. H., Hameri, K., Pirjola, L., Kulmala, M., Jennings, S. G., and Hoffmann, T.: Marine aerosols and iodine emissions - Reply, Nature, 433, E13-E14, 10.1038/nature03373, 2005.

Odame-Ankrah, C. A.: Improved detection instrument for nitrogen oxide species, Ph.D., Chemistry, University of Calgary, http://hdl.handle.net/11023/2006, Calgary, 2015.

Onasch, T. B., Trimborn, A., Fortner, E. C., Jayne, J. T., Kok, G. L., Williams, L. R., Davidovits, P., and Worsnop, D. R.: Soot particle aerosol mass spectrometer: Development, validation, and initial application, Aerosol Sci. Technol., 46, 804-817, 10.1080/02786826.2012.663948, 2012.

Paatero, P., and Tapper, U.: Positive matrix factorization - a nonnegative factor model with optimal utilization of error-estimates of data values, Environmetrics, 5, 111-126, 10.1002/env.3170050203, 1994.

Pachauri, R. K., and Meyer, L. A.: IPCC climate change 2014: Synthesis report. contribution of working groups I, II and III to the fifth assessment report of the intergovernmental panel on climate change, Geneva, 2014.

Paine, R. T.: Trophic control of production in a rocky intertidal community, Science, 296, 736- 739, 10.1126/science.1069811, 2002.

Palmer, P. I., and Shaw, S. L.: Quantifying global marine isoprene fluxes using MODIS chlorophyll observations, Geophys. Res. Lett., 32, L09805, 10.1029/2005gl022592, 2005.

Pandey Deolal, S., Staehelin, J., Brunner, D., Cui, J., Steinbacher, M., Zellweger, C., Henne, S., and Vollmer, M. K.: Transport of PAN and NOy from different source regions to the Swiss high alpine site Jungfraujoch, Atmos. Environ., 64, 103-115, 10.1016/j.atmosenv.2012.08.021, 2013.

183

Parajulee, A., and Wania, F.: Evaluating officially reported polycyclic aromatic hydrocarbon emissions in the Athabasca oil sands region with a multimedia fate model, Proc. Natl. Acad. Sci., 111, 3344-3349, 10.1073/pnas.1319780111, 2014.

Parrish, D. D., Dunlea, E. J., Atlas, E. L., Schauffler, S., Donnelly, S., Stroud, V., Goldstein, A. H., Millet, D. B., McKay, M., Jaffe, D. A., Price, H. U., Hess, P. G., Flocke, F., and Roberts, J. M.: Changes in the photochemical environment of the temperate North Pacific troposphere in response to increased Asian emissions, J. Geophys. Res., 109, D23S18, 10.1029/2004jd004978, 2004.

Parrish, D. D., Millet, D. B., and Goldstein, A. H.: Increasing ozone in marine boundary layer inflow at the west coasts of North America and Europe, Atmos. Chem. Phys., 9, 1303-1323, 10.5194/acp-9-1303-2009, 2009.

Parrish, D. D., Aikin, K. C., Oltmans, S. J., Johnson, B. J., Ives, M., and Sweeny, C.: Impact of transported background ozone inflow on summertime air quality in a California ozone exceedance area, Atmos. Chem. Phys., 10, 10093-10109, 10.5194/acp-10-10093-2010, 2010.

Patterson, G. E., Guymon, A. J., Riter, L. S., Everly, M., Griep-Raming, J., Laughlin, B. C., Ouyang, Z., and Cooks, R. G.: Miniature cylindrical ion trap mass spectrometer, Anal. Chem., 74, 6145-6153, 10.1021/ac020494d, 2002.

Paul, D., Furgeson, A., and Osthoff, H. D.: Measurements of total peroxy and alkyl nitrate abundances in laboratory-generated gas samples by thermal dissociation cavity ring-down spectroscopy, Rev. Sci. Instrum., 80, 8, 114101, 10.1063/1.3258204, 2009.

Paul, D., and Osthoff, H. D.: Absolute measurements of total peroxy nitrate mixing ratios by thermal dissociation blue diode laser cavity ring-down spectroscopy, Anal. Chem., 82, 6695- 6703, 10.1021/ac101441z, 2010.

Peng, R., Samoli, E., Pham, L., Dominici, F., Touloumi, G., Ramsay, T., Burnett, R., Krewski, D., Tertre, A., Cohen, A., Atkinson, R., Anderson, H. R., Katsouyanni, K., and Samet, J.: Acute effects of ambient ozone on mortality in Europe and North America: results from the APHENA study, Air Qual Atmos Health, 6, 445-453, 10.1007/s11869-012-0180-9, 2013.

Penner, T. J., and Foght, J. M.: Mature fine tailings from oil sands processing harbour diverse methanogenic communities, Canadian Journal of Microbiology, 56, 459-470, 10.1139/w10- 029, 2010.

Percy, K. E.: Ambient air quality and linkage to ecosystems in the Athabasca oil sands, Alberta, Geosci. Can., 40, 182-201, 10.12789/geocanj.2013.40.014, 2013.

184

Pereira, A. S., Bhattacharjee, S., and Martin, J. W.: Characterization of oil sands process-affected waters by liquid chromatography Orbitrap mass spectrometry, Environ. Sci. Technol., 47, 5504-5513, 10.1021/es401335t, 2013.

Peters, R. W.: Environmental chemistry, 7th ed., Environmental Progress, 2, CRC Press, J13-J14 pp., 2002.

Peters, T. M., and Leith, D.: Concentration measurement and counting efficiency of the aerodynamic particle sizer 3321, J. Aerosol. Sci., 34, 627-634, 10.1016/s0021- 8502(03)00030-2, 2003.

Pollmann, J., Ortega, J., and Helmig, D.: Analysis of atmospheric sesquiterpenes: Sampling losses and mitigation of ozone interferences, Environ. Sci. Technol., 39, 9620-9629, 10.1021/es050440w, 2005.

Pope, C. A., III, Ezzati, M., and Dockery, D. W.: Fine-particulate air pollution and life expectancy in the United States, N Engl J Med, 360, 376-386, 10.1056/NEJMsa0805646, 2009.

Quagraine, E. K., Headley, J. V., and Peterson, H. G.: Is biodegradation of bitumen a source of recalcitrant naphthenic acid mixtures in oil sands tailing pond waters?, J. Environ. Sci. Health Part A-Toxic/Hazard. Subst. Environ. Eng., 40, 671-684, 10.1081/ese-200046637, 2005.

Ras, M. R., Borrull, F., and Marc, R. M.: Sampling and preconcentration techniques for determination of volatile organic compounds in air samples, Trends Analyt. Chem., 28, 347- 361, 10.1016/j.trac.2008.10.009, 2009.

Read, K. A., Mahajan, A. S., Carpenter, L. J., Evans, M. J., Faria, B. V. E., Heard, D. E., Hopkins, J. R., Lee, J. D., Moller, S. J., Lewis, A. C., Mendes, L., McQuaid, J. B., Oetjen, H., Saiz-Lopez, A., Pilling, M. J., and Plane, J. M. C.: Extensive halogen-mediated ozone destruction over the tropical Atlantic Ocean, Nature, 453, 1232-1235, 10.1038/nature07035, 2008.

Reed, D. C., and Brzezinski, M. A.: Kelp Forests, The management of natural coastal carbon sinks, edited by: Laffoley, D., and Grimsditch, G., International Union of Conservation for Nature (IUCN), Gland, Switzerland, 2009.

Reeser, D. I., and Donaldson, D. J.: Influence of water surface properties on the heterogeneous reaction between O3(g) and I(aq), Atmos. Environ., 45, 6116-6120, 10.1016/j.atmosenv.2011.08.042, 2011.

185

Rhoderick, G. C.: Stability assessment of gas mixtures containing terpenes at nominal 5 nmol/mol contained in treated aluminum gas cylinders, Anal. Bioanal. Chem., 398, 1417- 1425, 10.1007/s00216-010-4058-0, 2010.

Rhoderick, G. C., and Lin, J.: Stability assessment of gas mixtures containing monoterpenes in varying cylinder materials and treatments, Anal. Chem., 85, 4675-4685, 10.1021/ac400324v, 2013.

Ridley, B. A., Shetter, J. D., Walega, J. G., Madronich, S., Elsworth, C. M., Grahek, F. E., Fehsenfeld, F. C., Norton, R. B., Parrish, D. D., Hubler, G., Buhr, M., Williams, E. J., Allwine, E. J., and Westberg, H. H.: The behavior of some organic nitrates at Boulder and Niwot Ridge, Colorado, J. Geophys. Res.-Atmos., 95, 13949-13961, 10.1029/JD095iD09p13949, 1990.

Riter, L. S., Peng, Y. A., Noll, R. J., Patterson, G. E., Aggerholm, T., and Cooks, R. G.: Analytical performance of a miniature cylindrical ion trap mass spectrometer, Anal. Chem., 74, 6154-6162, 10.1021/ac0204956, 2002.

Ro-Poulsen, H., Mikkelsen, T. N., Hovmand, M. F., Hummelsehoj, P., and Jensen, N. O.: Ozone deposition in relation to canopy physiology in a mixed conifer forest in Denmark, Chemosphere, 36, 669-674, 10.1016/s0045-6535(97)10105-9, 1998.

Roberts, J. M.: The atmospheric chemistry of organic nitrates, Atmospheric Environment Part A - General Topics, 24, 243-287, 10.1016/0960-1686(90)90108-Y, 1990.

Roberts, J. M., Williams, J., Baumann, K., Buhr, M. P., Goldan, P. D., Holloway, J., Hubler, G., Kuster, W. C., McKeen, S. A., Ryerson, T. B., Trainer, M., Williams, E. J., Fehsenfeld, F. C., Bertman, S. B., Nouaime, G., Seaver, C., Grodzinsky, G., Rodgers, M., and Young, V. L.: Measurements of PAN, PPN, and MPAN made during the 1994 and 1995 Nashville Intensives of the Southern Oxidant Study: Implications for regional ozone production from biogenic hydrocarbons, J. Geophys. Res.-Atmos., 103, 22473-22490, 10.1029/98JD01637, 1998.

Roberts, J. M., Flocke, F., Chen, G., de Gouw, J., Holloway, J. S., Hubler, G., Neuman, J. A., Nicks, D. K., Nowak, J. B., Parrish, D. D., Ryerson, T. B., Sueper, D. T., Warneke, C., and Fehsenfeld, F. C.: Measurement of peroxycarboxylic nitric anhydrides (PANs) during the ITCT 2K2 aircraft intensive experiment, J. Geophys. Res.-Atmos., 109, 10.1029/2004JD004960, 2004.

Roberts, J. M.: Measurement of the Henry's law coefficient and first order loss rate of PAN in n- octanol, Geophys. Res. Lett., 32, 10.1029/2004GL022327, 2005.

Roberts, J. M.: PAN and Related Compounds, in: Volatile Organic Compounds in the Atmosphere, edited by: Koppmann, R., Blackwell Publishing, Oxford, UK, 221-268, 2007. 186

Roberts, J. M., Marchewka, M., Bertman, S. B., Sommariva, R., Warneke, C., de Gouw, J., Kuster, W., Goldan, P., Williams, E., Lerner, B. M., Murphy, P., and Fehsenfeld, F. C.: Measurements of PANs during the New England Air Quality Study 2002, J. Geophys. Res.- Atmos., 112, D20306, 10.1029/2007jd008667, 2007.

Robinson, A. L., Donahue, N. M., Shrivastava, M. K., Weitkamp, E. A., Sage, A. M., Grieshop, A. P., Lane, T. E., Pierce, J. R., and Pandis, S. N.: Rethinking organic aerosols: Semivolatile emissions and photochemical aging, Science, 315, 1259-1262, 2007.

Rohr, A. C.: The health significance of gas- and particle-phase terpene oxidation products: A review, Environ. Int., 60, 145-162, 10.1016/j.envint.2013.08.002, 2013.

Rooney, R. C., Bayley, S. E., and Schindler, D. W.: Oil sands mining and reclamation cause massive loss of peatland and stored carbon, Proc. Natl. Acad. Sci. U. S. A., 109, 4933-4937, 10.1073/pnas.1117693108, 2012.

Rossignol, S., Chiappini, L., Perraudin, E., Rio, C., Fable, S., Valorso, R., and Doussin, J. F.: Development of a parallel sampling and analysis method for the elucidation of gas/particle partitioning of oxygenated semi-volatile organics: a limonene ozonolysis study, Atmos. Meas. Tech., 5, 1459-1489, 10.5194/amt-5-1459-2012, 2012.

Sakulyanontvittaya, T., Duhl, T., Wiedinmyer, C., Helmig, D., Matsunaga, S., Potosnak, M., Milford, J., and Guenther, A.: Monoterpene and sesquiterpene emission estimates for the United States, Environ. Sci. Technol., 42, 1623-1629, 10.1021/es702274e, 2008a.

Sakulyanontvittaya, T., Guenther, A., Helmig, D., Milford, J., and Wiedinmyer, C.: Secondary organic aerosol from sesquiterpene and monoterpene emissions in the United States, Environ. Sci. Technol., 42, 8784-8790, 10.1021/es800817r, 2008b.

Sander, R.: Compilation of Henry's law constants (version 4.0) for water as solvent, Atmos. Chem. Phys., 15, 4399-4981, 10.5194/acp-15-4399-2015, 2015.

Sander, S. P., Abbatt, J. P. D., Barker, J. R., Burkholder, J. B., Friedl, R. R., Golden, D. M., Huie, R. E., Kolb, C. E., Kurylo, M. J., Moortgat, G. K., Orkin, V. L., and Wine, P. H.: Chemical kinetics and photochemical data for use in atmospheric studies, Evaluation No. 17, JPL Publication 10-6, Jet Propulsion Laboratory, Pasadena, CA, 2010.

Sartin, J. H., Halsall, C. J., Davison, B., Owen, S., and Hewitt, C. N.: Determination of biogenic volatile organic compounds (C8–C16) in the coastal atmosphere at Mace Head, Ireland, Anal. Chim. Acta., 428, 61-72, 10.1016/S0003-2670(00)01214-9, 2001.

187

Sartin, J. H., Halsall, C. J., Hayward, S., and Hewitt, C. N.: Emission rates of C8–C15 VOCs from seaweed and sand in the inter-tidal zone at Mace Head, Ireland, Atmos. Environ., 36, 5311-5321, 10.1016/S1352-2310(02)00639-8, 2002.

Saxton, J. E., Lewis, A. C., Kettlewell, J. H., Ozel, M. Z., Gogus, F., Boni, Y., Korogone, S. O. U., and Serça, D.: Isoprene and monoterpene measurements in a secondary forest in northern Benin, Atmos. Chem. Phys., 7, 4095-4106, 10.5194/acp-7-4095-2007, 2007.

Schurath, U., and Wippercht, V.: Reactions of peroxyacyl radicals, Proceedings of the first european symposium on the physicochemical behavior of atmospheric pollutant, Luxembourg, 1980, 157-166,

Seinfeld, J. H., and Pandis, S. N.: Atmospheric chemistry and physics: from air pollution to climate change, 2nd ed., Wiley, Hoboken, N.J., 2006.

Selucky, M. L., Chu, Y., Ruo, T., and Strausz, O. P.: Chemical composition of Athabasca bitumen, Fuel, 56, 369-381, 10.1016/0016-2361(77)90061-8, 1977.

Shahimin, M. F. M., and Siddique, T.: Sequential biodegradation of complex naphtha hydrocarbons under methanogenic conditions in two different oil sands tailings, Environ. Pollut., 221, 398-406, 10.1016/j.envpol.2016.12.002, 2017.

Shaw, S. L., Gantt, B., and Meskhidez, N.: Production and emissions of marine isoprene and monoterpenes: A review, Adv. Meterol., 2010, 24, 10.1155/2010/408696, 2010.

Siddique, T., Gupta, R., Fedorak, P. M., MacKinnon, M. D., and Foght, J. M.: A first approximation kinetic model to predict methane generation from an oil sands tailings settling basin, Chemosphere, 72, 1573-1580, 10.1016/j.chemosphere.2008.04.036, 2008.

Siddique, T., Penner, T., Semple, K., and Foght, J. M.: Anaerobic biodegradation of longer-chain n-alkanes coupled to methane production in oil sands tailings, Environ. Sci. Technol., 45, 5892-5899, 10.1021/es200649t, 2011.

Siddique, T., Penner, T., Klassen, J., Nesbo, C., and Foght, J. M.: Microbial communities involved in methane production from hydrocarbons in oil sands Tailings, Environ. Sci. Technol., 46, 9802-9810, 10.1021/e53022024, 2012.

Silva, D. B., Pott, A., and Oliveira, D. C. R.: Analyses of the headspace volatile constituents of aerial parts (leaves and stems), flowers and fruits of Bidens gardneri Bak. and Bidens sulphurea (Cav.) Sch.Bip. using solid-phase microextraction, J. Essent. Oil Res., 22, 560- 563, 10.1080/10412905.2010.9700400, 2010.

Simpson, I. J., Blake, N. J., Barletta, B., Diskin, G. S., Fuelberg, H. E., Gorham, K., Huey, L. G., Meinardi, S., Rowland, F. S., Vay, S. A., Weinheimer, A. J., Yang, M., and Blake, D. R.: 188

Characterization of trace gases measured over Alberta oil sands mining operations: 76 speciated C2–C10 volatile organic compounds (VOCs), CO2, CH4, CO, NO, NO2, NOy, O3 and SO2, Atmos. Chem. Phys., 10, 11931-11954, 10.5194/acp-10-11931-2010, 2010.

Singh, H. B., Gregory, G. L., Anderson, B., Browell, E., Sachse, G. W., Davis, D. D., Crawford, J., Bradshaw, J. D., Talbot, R., Blake, D. R., Thornton, D., Newell, R., and Merrill, J.: Low ozone in the marine boundary layer of the tropical Pacific Ocean: Photochemical loss, chlorine atoms, and entrainment, Journal of Geophysical Research, 101, 1907-1917, 10.1029/95jd01028, 1996.

Sinha, V., Williams, J., Meyerhofer, M., Riebesell, U., Paulino, A. I., and Larsen, A.: Air-sea fluxes of methanol, acetone, acetaldehyde, isoprene and DMS from a Norwegian fjord following a phytoplankton bloom in a mesocosm experiment, Atmos. Chem. Phys., 7, 739- 755, 10.5194/acp-7-739-2007, 2007.

Skoog, D. A., J., H. F., and R., C. S.: Principles of instrumental analysis, 6th ed., David Harris, Belmont, CA, 2007.

Slavkovska, V., Couladis, M., Bojovic, S., Tzakou, O., Pavlovic, M., Lakusic, B., and Jancic, R.: Essential oil and its systematic significance in species of Micromeria Bentham from Serbia & Montenegro, Plant Syst. Evol., 255, 1-15, 10.1007/s00606-005-0303-y, 2005.

Small, C. C., Cho, S., Hashisho, Z., and Ulrich, A. C.: Emissions from oil sands tailings ponds: Review of tailings pond parameters and emission estimates, J Petrol. Sci. Eng., 127, 490-501, 10.1016/j.petrol.2014.11.020, 2015.

Steneck, R. S., Graham, M. H., Bourque, B. J., Corbett, D., Erlandson, J. M., Estes, J. A., and Tegner, M. J.: Kelp forest ecosystems: biodiversity, stability, resilience and future, Environ. Conserv., 29, 436-459, 10.1017/s0376892902000322, 2002.

Strausz, O. P., Jha, K. N., and Montgomery, D. S.: Chemical composition of gases in Athabasca bitumen and in low-temperature thermolysis of oil sand, asphaltene and maltene, Fuel, 56, 114-120, 10.1016/0016-2361(77)90128-4, 1977.

Stroud, C., Madronich, S., Atlas, E., Ridley, B., Flocke, F., Weinheimer, A., Talbot, B., Fried, A., Wert, B., Shetter, R., Lefer, B., Coffey, M., Heikes, B., and Blake, D.: Photochemistry in the arctic free troposphere: NOx budget and the role of odd nitrogen reservoir recycling, Atmos. Environ., 37, 3351-3364, 10.1016/S1352-2310(03)00353-4, 2003.

Stull, R. B., and Ahrens, C. D.: Meteorology for scientists and engineers, 2nd ed., Brooks/Cole, Pacific Grove, CA, 502 pp., 2000.

Subramanian, M., Deo, M. D., and Hanson, F. V.: Compositional analysis of bitumen and bitumen-derived products, J. Chromatogr. Sci., 34, 20-26, 10.1093/chromsci/34.1.20, 1996. 189

Thaler, R. D., Mielke, L. H., and Osthoff, H. D.: Quantification of nitryl chloride at part per trillion mixing ratios by thermal dissociation cavity ring-down spectroscopy, Anal. Chem., 83, 2761-2766, 10.1021/ac200055z, 2011.

Thurston, G. D., and Spengler, J. D.: A quantitative assessment of source contributions to inhalable particulate matter pollution in metropolitan Boston, Atmos. Environ., 19, 9-25, 10.1016/0004-6981(85)90132-5, 1985.

Thurston, G. D., Ito, K., and Lall, R.: A source apportionment of U.S. fine particulate matter air pollution, Atmos. Environ., 45, 3924-3936, 10.1016/j.atmosenv.2011.04.070, 2011.

Tokarek, T. W., Huo, J. A., Odame-Ankrah, C. A., Hammoud, D., Taha, Y. M., and Osthoff, H. D.: A gas chromatograph for quantification of peroxycarboxylic nitric anhydrides calibrated by thermal dissociation cavity ring-down spectroscopy, Atmos. Meas. Tech., 7, 3263-3283, 10.5194/amt-7-3263-2014, 2014.

Tokarek, T. W., Brownsey, D. K., Jordan, N., Garner, N. M., Ye, C. Z., Assad, F. V., Peace, A., Schiller, C. L., Mason, R. H., Vingarzan, R., and Osthoff, H. D.: Biogenic Emissions and Nocturnal Ozone Depletion Events at the Amphitrite Point Observatory on Vancouver Island, Atmos. Ocean, 1-12, 10.1080/07055900.2017.1306687, 2017.

Tokarek, T. W., Odame-Ankrah, C. A., Huo, J. A., McLaren, R., Lee, A. K. Y., Adam, M. G., Willis, M. D., Abbatt, J. P. D., Mihele, C., Darlington, A., Mittermeier, R. L., Strawbridge, K., Hayden, K. L., Olfert, J. S., Schnitzler, E. G., Brownsey, D. K., Assad, F. V., Wentworth, G. R., Tevlin, A. G., Worthy, D. E. J., Li, S. M., Liggio, J., Brook, J. R., and Osthoff, H. D.: Principal component analysis of summertime ground site measurements in the Athabasca oil sands: Sources of IVOCs, Atmos. Chem. Phys., 18, 17819-17841, 10.5194/acp-18-17819- 2018, 2018.

Tokarek, T. W., Brownsey, D. K., Jordan, N., Garner, N. M., Ye, C. Z., and Osthoff, H. D.: Emissions of C9-C16 hydrocarbons from kelp species on Vancouver Island: Alaria marginata (winged kelp) and Nereocystis luetkeana (bull kelp) are an atmospheric source of limonene, Atmos. Environ., In review.

Trainer, M., Williams, E. J., Parrish, D. D., Buhr, M. P., Allwine, E. J., Westberg, H. H., Fehsenfeld, F. C., and Liu, S. C.: Models and observations of the impact of natural hydrocarbons on rural ozone, Nature, 329, 705-707, 10.1038/329705a0, 1987.

Trentmann, J., Andreae, M. O., and Graf, H. F.: Chemical processes in a young biomass-burning plume, J. Geophys. Res.-Atmos., 108, 21, 10.1029/2003jd003732, 2003.

Tuberoso, C. I. G., Kowalczyk, A., Coroneo, V., Russo, M. T., Dessì, S., and Cabras, P.: Chemical composition and antioxidant, antimicrobial, and antifungal activities of the

190

essential oil of Achillea ligustica All, J. Agric. Food Chem., 53, 10148-10153, 10.1021/jf0518913, 2005.

Turner, N. C., Waggoner, P. E., and Rich, S.: Removal of ozone from the atmosphere by soil and vegetation, Nature, 250, 486-489, 10.1038/250486a0, 1974.

Wang, H., Chen, C., Huang, C., and Fu, L.: On-road vehicle emission inventory and its uncertainty analysis for Shanghai, China, Sci. Total Environ., 398, 60-67, 10.1016/j.scitotenv.2008.01.038, 2008.

Wang, S. C., and Flagan, R. C.: Scanning electrical mobility spectrometer, Aerosol Sci. Technol., 13, 230-240, 10.1080/02786829008959441, 1990.

Wang, X. L., Chow, J. C., Kohl, S. D., Percy, K. E., Legge, A. H., and Watson, J. G.: Characterization of PM2.5 and PM10 fugitive dust source profiles in the Athabasca Oil Sands Region, J. Air Waste Manage. Assoc., 65, 1421-1433, 10.1080/10962247.2015.1100693, 2015.

Wang, X. L., Chow, J. C., Kohl, S. D., Percy, K. E., Legge, A. H., and Watson, J. G.: Real-world emission factors for Caterpillar 797B heavy haulers during mining operations, Particuology, 28, 22-30, 10.1016/j.partic.2015.07.001, 2016.

Warneke, C., de Gouw, J. A., Kuster, W. C., Goldan, P. D., and Fall, R.: Validation of Atmospheric VOC Measurements by Proton-Transfer- Reaction Mass Spectrometry Using a Gas-Chromatographic Preseparation Method, Environ. Sci. Technol., 37, 2494-2501, 10.1021/es026266i, 2003.

Warren, L. A., Kendra, K. E., Brady, A. L., and Slater, G. F.: Sulfur biogeochemistry of an oil sands composite tailings deposit, Front. Microbiol., 6, 14, 10.3389/fmicb.2015.01533, 2016.

Watanabe, A., Watanabe, C., Freeman, R. R., Teramae, N., and Ohtani, H.: Hydrogenation reactions during pyrolysis-gas chromatography/mass spectrometry analysis of polymer samples using hydrogen carrier gas, Anal. Chem., 88, 5462-5468, 10.1021/acs.analchem.6b00892, 2016.

Watanabe, K., Nojiri, Y., and Kariya, S.: Measurements of ozone concentrations on a commercial vessel in the marine boundary layer over the northern North Pacific Ocean, Journal of Geophysical Research, 110, D11310, 10.1029/2004jd005514, 2005.

Watson, J., and Estes, J. A.: Stability, resilience, and phase shifts in rocky subtidal communities along the west coast of Vancouver Island, Canada, Ecological Monographs, 81, 215-239, 10.1890/10-0262.1, 2011.

191

Watson, J. G., Chow, J. C., Wang, X., Zielinska, B., Kohl, S. D., and Gronstal, S.: Characterization of real-world emissions from nonroad mining trucks in the athabasca oil sands region during september, 2013.

WBEA: WBEA annual report 2013, Wood Buffalo Environmental Association, 2013.

Weinberg, I., Bahlmann, E., Michaelis, W., and Seifert, R.: Determination of fluxes and isotopic composition of halocarbons from seagrass meadows using a dynamic flux chamber, Atmos. Environ., 73, 34-40, 10.1016/j.atmosenv.2013.03.006, 2013.

Wesely, M. L., and Hicks, B. B.: A review of the current status of knowledge on dry deposition, Atmos. Environ., 34, 2261-2282, 10.1016/S1352-2310(99)00467-7, 2000.

Whaley, C. H., Makar, P. A., Shephard, M. W., Zhang, L. M., Zhang, J. H., Zheng, Q., Akingunola, A., Wentworth, G. R., Murphy, J. G., Kharol, S. K., and Cady-Pereira, K. E.: Contributions of natural and anthropogenic sources to ambient ammonia in the Athabasca Oil Sands and north-western Canada, Atmos. Chem. Phys., 18, 2011-2034, 10.5194/acp-18- 2011-2018, 2018.

Williams, B. J., Goldstein, A. H., Kreisberg, N. M., and Hering, S. V.: An in-situ instrument for speciated organic composition of atmospheric aerosols: Thermal desorption aerosol GC/MS- FID (TAG), Aerosol Sci. Technol., 40, 627-638, 10.1080/02786820600754631, 2006.

Williams, J., Roberts, J. M., Bertman, S. B., Stroud, C. A., Fehsenfeld, F. C., Baumann, K., Buhr, M. P., Knapp, K., Murphy, P. C., Nowick, M., and Williams, E. J.: A method for the airborne measurement of PAN, PPN, and MPAN, J. Geophys. Res.-Atmos., 105, 28943- 28960, 10.1029/2000JD900373, 2000.

Wilmers, C. C., Estes, J. A., Edwards, M., Laidre, K. L., and Konar, B.: Do trophic cascades affect the storage and flux of atmospheric carbon? An analysis of sea otters and kelp forests, Front. Ecol. Environ., 10, 409-415, 10.1890/110176, 2012.

Wilson, N. K., Barbour, R. K., Chuang, J. C., and Mukund, R.: Evaluation of a real-time monitor for fine particle-bound PAH in air, Polycycl. Aromat. Compd., 5, 167-174, 10.1080/10406639408015168, 1994.

Wilson, T. W., Ladino, L. A., Alpert, P. A., Breckels, M. N., Brooks, I. M., Browse, J., Burrows, S. M., Carslaw, K. S., Huffman, J. A., Judd, C., Kilthau, W. P., Mason, R. H., McFiggans, G., Miller, L. A., Najera, J. J., Polishchuk, E., Rae, S., Schiller, C. L., Si, M., Temprado, J. V., Whale, T. F., Wong, J. P. S., Wurl, O., Yakobi-Hancock, J. D., Abbatt, J. P. D., Aller, J. Y., Bertram, A. K., Knopf, D. A., and Murray, B. J.: A marine biogenic source of atmospheric ice-nucleating particles, Nature, 525, 10.1038/nature14986, 2015.

192

Wingenter, O. W., Haase, K. B., Strutton, P., Friederich, G., Meinardi, S., Blake, D. R., and Rowland, F. S.: Changing concentrations of CO, CH4, C5H8, CH3Br ,CH3I, and dimethyl sulfide during the southern ocean iron enrichment experiments, Proc. Natl. Acad. Sci. U. S. A., 101, 8537-8541, 10.1073/pnas.0402744101, 2004.

Wise, M. L., Rorrer, G. L., Polzin, J. J., and Croteau, R.: Biosynthesis of marine natural products: Isolation and characterization of a myrcene synthase from cultured tissues of the marine red alga Ochtodes secundiramea, Arch. Biochem. Biophys., 400, 125-132, 10.1006/abbi.2002.2780, 2002.

Wise, M. L.: Monoterpene biosynthesis in marine algae, Phycologia, 42, 370-377, 10.2216/i0031-8884-42-4-370.1, 2003.

Woodward-Massey, R., Taha, Y. M., Moussa, S. G., and Osthoff, H. D.: Comparison of negative-ion proton-transfer with iodide ion chemical ionization mass spectrometry for quantification of isocyanic acid in ambient air, Atmos. Environ., 98, 693-703, 10.1016/j.atmosenv.2014.09.014, 2014.

Xu, X., Stee, L. L. P., Williams, J., Beens, J., Adahchour, M., Vreuls, R. J. J., Brinkman, U. A., and Lelieveld, J.: Comprehensive two-dimensional gas chromatography (GC × GC) measurements of volatile organic compounds in the atmosphere, Atmos. Chem. Phys., 3, 665-682, 10.5194/acp-3-665-2003, 2003.

Yakobi-Hancock, J. D., Ladino, L. A., Bertram, A. K., Huffman, J. A., Jones, K., Leaitch, W. R., Mason, R. H., Schiller, C. L., Toom-Sauntry, D., Wong, J. P. S., and Abbatt, J. P. D.: CCN activity of size-selected aerosol at a Pacific coastal location, Atmos. Chem. Phys., 14, 12307- 12317, 10.5194/acp-14-12307-2014, 2014.

Yassaa, N., Peeken, I., Zöllner, E., Bluhm, K., Arnold, S., Spracklen, D., and Williams, J.: Evidence for marine production of monoterpenes, Environ. Chem., 5, 391-401, 10.1071/EN08047, 2008.

Yeh, S., Jordaan, S. M., Brandt, A. R., Turetsky, M. R., Spatari, S., and Keith, D. W.: Land use greenhouse gas emissions from conventional oil production and oil sands, Environ. Sci. Technol., 44, 8766-8772, 10.1021/es1013278, 2010.

Yokelson, R. J., Burling, I. R., Gilman, J. B., Warneke, C., Stockwell, C. E., de Gouw, J., Akagi, S. K., Urbanski, S. P., Veres, P., Roberts, J. M., Kuster, W. C., Reardon, J., Griffith, D. W. T., Johnson, T. J., Hosseini, S., Miller, J. W., Cocker Iii, D. R., Jung, H., and Weise, D. R.: Coupling field and laboratory measurements to estimate the emission factors of identified and unidentified trace gases for prescribed fires, Atmos. Chem. Phys., 13, 89-116, 10.5194/acp-13-89-2013, 2013.

193

Yokouchi, Y., Li, H. J., Machida, T., Aoki, S., and Akimoto, H.: Isoprene in the marine boundary layer (Southeast Asian Sea, eastern Indian Ocean, and Southern Ocean): Comparison with dimethyl sulfide and bromoform, J. Geophys. Res.-Atmos., 104, 8067- 8076, 10.1029/1998JD100013, 1999.

Yoon, S., Bhatt, S. D., Lee, W., Lee, H. Y., Jeong, S. Y., Baeg, J.-O., and Lee, C. W.: Separation and characterization of bitumen from Athabasca oil sand, Korean J. Chem. Eng., 26, 64-71, 10.1007/s11814-009-0011-3, 2009.

Zaragoza, J., Callahan, S., McDuffie, E. E., Kirkland, J., Brophy, P., Durrett, L., Farmer, D. K., Zhou, Y., Sive, B., Flocke, F., Pfister, G., Knote, C., Tevlin, A., Murphy, J., and Fischer, E. V.: Observations of acyl peroxy nitrates during the front range air pollution and photochemistry experiment (FRAPPE), J. Geophys. Res.-Atmos., 122, 12416-12432, 10.1002/2017jd027337, 2017.

Zeller, K. F., and Nikolov, N. T.: Quantifying simultaneous fluxes of ozone, carbon dioxide and water vapor above a subalpine forest ecosystem, Environ. Pollut., 107, 1-20, 10.1016/S0269- 7491(99)00156-6, 2000.

Zhang, Y., Wang, Y., Chen, G., Smeltzer, C., Crawford, J., Olson, J., Szykman, J., Weinheimer, A. J., Knapp, D. J., Montzka, D. D., Wisthaler, A., Mikoviny, T., Fried, A., and Diskin, G.: Large vertical gradient of reactive nitrogen oxides in the boundary layer: Modeling analysis of DISCOVER-AQ 2011 observations, J. Geophys. Res.-Atmos., 1922–1934, 10.1002/2015jd024203, 2016.

Zhao, X. M., Gao, T. Y., and Zhang, J. B.: Heterogeneous reaction of peroxyacetyl nitrate (PAN) on soot, Chemosphere, 177, 339-346, 10.1016/j.chemosphere.2017.03.001, 2017.

194