su.diva-portal.orgsu.diva-portal.org/smash/get/diva2:1200210/FULLTEXT02.pdfimprovement in reversed-phase chromatographic separation, ionization efficiency as well as selectivity. For

��

�� !"#$��#"%""��&��'��()��*�� (��+�#,-%

��.��(��/�� %�� 0�� (0�� %�� 0�� 0��/ ��(�� 0�� 1��23%.�� 2 �� %

�� /� �4�� %5 0��(�� /�6��%.�� (�� (�� 0�� /�� ( �� 4�� 0��( ��7�� 8��9��7�� %

.� �� (�� :�� 0�� %�� 6( �� ( � �� 0�� 0 � �� ( �� %� �� 0 �� (�� 0�� :�� (� � �� %& � ��0�� (0��0�� :�� % �� 0�� :��( �� /�� (� ��:�� 0��%; �� (�� 0�� (��:��/0�� 7�� %.�� (�� 0 �� 0�� (��:�� %�� 6�� %

�� 0��(� �� $"<2(� 0�� 1�� (�� 3 �� /=� �� /=��%.��(��0�� 0��(�� %�� (�� (0�� %.�� (�� 0�� (�� %�� >�� :�� %.�� 0�� (0�� (��:�� %

?�� (�� 2�� %@�� 20�� (�� 0/�� %>��(� ��7�� 6�� (�� %

�� !"#$��A44��%��%�4�� B��C��A��A�A�A��/#��!�D

.�-2EF$/E#/FFEF/!=$/E

.�-2EF$/E#/FFEF/!=E/,

��

�� (#",E#��

ANALYTICAL METHODS FOR BIOMOLECULES INVOLVED INATMOSPHERIC AEROSOL FORMATION IN THE ARCTIC

Farshid Mashayekhy Rad

Analytical methods for biomoleculesinvolved in atmospheric aerosolformation in the Arctic

Farshid Mashayekhy Rad

©Farshid Mashayekhy Rad, Stockholm University 2018 ISBN print 978-91-7797-238-9ISBN PDF 978-91-7797-239-6 Printed in Sweden by Universitetsservice US-AB, Stockholm 2018Distributor: Department of Environmental Science and Analytical Chemistry, StockholmUniversity

To my beloved parents andfamily

1

List of papers I. Leck C, Gao Q, Mashayekhy Rad F, Nilsson U (2013). Size-resolved atmospheric

particulate polysaccharides in the high summer Arctic. Atmospheric Chemistry and Physics 13: 12573-12588.The author was partly responsible for the analysis and data processing and evaluation.

II. Mashayekhy Rad F, Zurita J, Gilles P, De Borggraeve W M, Nilsson U, Ilag L L, Leck C (2018) Measurements of atmospheric proteinaceous aerosol in the Arctic using a selective HPLC/ESI-MS/MS strategy. Submitted.The author was responsible for analytical method development, analysis, data processing, evaluation and

partly the writing.

III. Karl M, Leck C, Mashayekhy Rad F, Bäcklund A, Lopez-Aparicio S, Heintzenberg J S (2018) Sources of the sub-micrometer aerosol at Mt. Zeppelin Observatory (Spitsbergen) in the year 2015: Indication for ice-related marine gels. Manuscript.The author was responsible for the sampling, analysis, data processing, evaluation of data and writing the

experimental part regarding polysaccharides.

IV. Mashayekhy Rad F, Leck C, Ilag L L, Nilsson U (2018) Investigation of UHPLC/travelling wave ion mobility/time of flight mass spectrometry for fast profiling of fatty acids in the high Arctic sea surface microlayer. Rapid Communications in Mass Spectrometry, in pressThe author was responsible for generating the idea regarding the analytical method development, all

analysis, data processing and evaluation as well as most of the writing.

V. Mashayekhy Rad F, Spinicci S, Silvergren S, Nilsson U, Westerholm R (2018) Validation of a HILIC/ESI-MS/MS method for the wood burning marker levoglucosan and its isomers in airborne particulate matter. Submitted.The author was responsible for the analytical method development, analyses, data processing and

evaluation as well as part of the writing.

2

Papers not included in the thesis VI. Keshavarzi N, Mashayekhy Rad F, Mace A, Ansari F, Akhtar F, Nilsson U, Berglund L,

Bergström L (2015) Nanocellulose–Zeolite Composite Films for Odor Elimination. ACS applied materials & interfaces 7:14254

VII. Österlund N, Kulkarni Y. S, Misiaszek A. D, Wallin C, Krüger D. M, Liao Qinghua, Mashayekhy Rad F, Jarvet J, Strodel B, Wärmländer S.K.T.S, Ilag L. L, Kamerlin S. C. L, Gräslund A (2018) Amyloid-β peptide interactions with amphiphilic surfactants: electrostatic and hydrophobic effects. ACS Chemical Neuroscience, in press

3

Populärvetenskaplig sammanfattning

Alltmer av isfria förhållanden sommartid och mer näringsämnen från smältande isar har med största sannolikhet en kraftig påverkan på det marina livet i Arktis. I och med ett varmare klimat förväntas också de mikrobiologiska processerna i havet öka, vilket leder till att mer bioorganiskt material bildas på havsytan, som till exempel utsöndringar och nedbrytningsprodukter från alger, plankton och bakterier. I havsytan bildas små partiklar som överförs till atmosfären när havsvågor bryts av vinden. Dessa vattenlösliga luftburna partiklar, aerosoler, kan fungera som molnbildande kondensationskärnor (eng.cloud condensation nuclei, CCN). Tidigare studier har visat på att arktiska CCN bland annat innehåller bioorganiskt material, från havsytan i råkarna mellan isflaken. Koncentrationen av sådana partiklar kan i såfall förväntas öka när temperaturen stiger, vilket i sin tur leder till att mer låga moln kan bildas. Mer låga moln innebär att balansen mellan solljusets in- och utstrålning förändras och därmed klimatet. För att förstå de bakomliggande processerna fullt ut och hur det arktiska klimatet kommer att påverkas av en sådan process, är det en förutsättning att mer noggrant kunna analysera de molnbildande partiklarna, aerosolerna, och även deras källorna.

Ett syfte med den här avhandlingen har varit att utveckla analysmetoder för att kemiskt bestämma biomolekyler i aerosoler, dimdroppar och i det mikrolager av bioorganiskt material som bildas på havsytan i Arktis. Att kunna mäta extremt låga halter av biomolekyler i de komplexa salthaltiga arktiska aerosolerna är en analytiskt kemisk utmaning. Nya metodiker med selektiva uppreningar och separationer med högkänslig masspektrometri har tagits fram i denna avhandling för att kunna bestämma proteinliknande ämnen (summan av aminosyror, peptider och proteiner) och kolhydrater i prover insamlade i den Arktiska luften och havsytan.

Resultaten i avhandlingen från de helårsmätningar som gjordes under 2001 och 2008 visar på ett innehåll av både proteinliknande ämnen och kolhydrater i aerosolerna på mellan 10-12 och 10-9

mol per kubikmeter luft, med de högsta halterna uppmätta under sommartid. En intressant upptäckt är de avsevärt högre halter som uppmättes under 2008 jämfört med 2001, vilket sammanfaller med ett varmare klimat och ett till ytan mindre istäcke. Analys av aerosoler och större dimdroppar visar på samma koncentrationsvariation över året, vilket tyder på att biomolekylerna verkligen ingår i en molnbildningsprocess. Vidare tyder en av studierna på förekomst av två olika typer av marina kolhydratinnehållande geler i aerosolerna som sammanfaller med säsongsvariationer i isarnas smältnings- och infrysningscykler.

4

Även om mycket mer omfattande studier krävs för att fullt förstå de komplexa kemiska och fysikaliska processerna bakom molnbildningen i Arktis, så stödjer resultaten i avhandlingen sammantaget de tidigare antagandena att biomolekyler från den marina miljön fungerar som effektiva CCN och i förlängningen kan påverka klimatet i Arktis. De inom ramen för avhandlingen utvecklade kemiska analysmetoderna finns nu dessutom tillgängliga för fortsatta studier inom detta så viktiga forskningsområde.

5

Abstract In the Arctic, increasing ice-free conditions and nutrients freed from the melting ice must strongly influence the marine life. Aerosol emissions from microbiological marine processes may affect the low clouds and fogs over the summer Arctic, which in turn have effects on the melting of sea ice. The radiative properties of the high Arctic low clouds are strongly dependent on the number concentration of airborne water-soluble particles, known as cloud condensation nuclei (CCN). If the effects of CCN on cloud optical properties are to be fully understood it is important to be able to specify the source and concentrations of the Arctic aerosol particles.

Previous studies in the Arctic have indicated that organic material formed in the uppermost ocean surface is transferred to the atmosphere and plays a potentially very important role in the aerosol-fog/cloud cycle. However, many aspects of this process remain unverified and chemical characterization of targeted groups of biomolecules is still notably fragmentary or non-existing. Investigation of biomolecules, particularly amino acids, peptides and proteins together with mono- and polysaccharides and fatty acids in the airborne aerosol, and their relative contributions to fog/cloud water requires the development of an array of “cutting edge” analytical techniques and methods.

In this thesis, electrospray ionization mass spectrometry was used for all applications and target biomolecules. The measurements in the Arctic turned out to be challenging due to the highly complex, salty matrices, combined with very low concentration and high diversity of the target biomolecules, and each step of the analytical chain needed careful consideration. To increase the detectability of the very low levels of polysaccharides and proteins in aerosols, these compounds were hydrolyzed to their subunits, monosaccharides, and amino acids. Monosaccharides were separated using hydrophilic interaction chromatography, which was beneficial for their detection in electrospray ionization mass spectrometry. Amino acids were derivatized, yielding improvement in reversed-phase chromatographic separation, ionization efficiency as well as selectivity. For fatty acids in a sea surface sample, a novel fast screening method was developed, utilizing travelling-wave ion mobility separation as an orthogonal technique connected to mass spectrometry. In addition, a method for the detection of wood burning as an anthropogenic source of aerosols was developed, utilizing anhydrous monosaccharides as markers. This method can be used in the upcoming expeditions for source apportionment studies.

The results from the analyses of the aerosol and fog water samples, collected over the summer pack ice north of 80°N, show that both total polysaccharides and total proteinaceous compounds (sum of proteins, peptides, and amino acids) occurred at the pmol m-3 to nmol m-3 level.

6

Interestingly, analyses of fog revealed higher levels of proteinaceous material between different years, suggested to be coupled to less ice coverage and thus to a higher biological activity in the ocean surface. The highest concentrations of polysaccharides, as an indication of marine polymer gels, were found during the summer over the pack ice area. In addition, a pilot source apportionment study was carried out combining the measurement of different molecular tracers, used as source markers. This study indicates the seasonality and abundance of marine polymer gels as an important feature of the Arctic Ocean connected to the melting and freezing of sea ice. It should be further studied how the abundance of these gels, which have a high potential for cloud droplet activation, affect the melting and freezing of the perennial sea ice.

Given the successful development of analytical methods for targeted groups of biomolecules, this thesis has supported the importance of biomolecules as CCN and for cloud formation in the Arctic. Less ice coverage may further increase the number of biomolecular CCN which could change the radiative balance, by the formation of more low-level clouds. Overall, more studies are required to further unravel the complex relationship of biogenic sources, atmospheric chemistry, and meteorology to assess the impact of climate change on the Arctic.

7

Contents

Page Abbreviations 91. Introduction 11 1.1 Aerosol, climate and health 111.2 Atmospheric aerosols 121.3 Sources of atmospheric aerosols - present knowledge 131.3.1 Airborne aerosols in the Arctic region 131.3.2 Airborne aerosols, PM, in Swedish suburban areas 141.4 Aims of thesis 141.5 Targeted group of biomolecules in SML, aerosols and fog 141.5.1 Polysaccharides and monosaccharides 141.5.2 Proteins and amino acids 161.5.3 Fatty acids 171.5.4 Anhydrous sugars 181.6 Analytical methods - a survey 191.6.1 Mono- and polysaccharides 191.6.2 Poly- and free amino acids 201.6.3 Fatty acids 212. The analytical workflow in this study 212.1 Sampling methods 232.1.1 Sea SML 232.1.2 Airborne aerosols 232.1.2.1 Multi-stage Berner cascade impactor 232.1.2.2 Two-stage stacked filter unit 242.1.2.3 PM10 sampler 242.1.3 Fog 242.2 Sample preparation 242.2.1 Extraction 242.2.2 Hydrolysis of biomolecular polymers 252.2.2.1 Polysaccharides in aerosol and fog 252.2.2.2 Proteinaceous materials in aerosols and fog 262.2.3 Sample derivatization of free amino acids 282.3. Instrumental methods 282.3.1 Liquid chromatography 282.3.1.1 Hydrophilic interaction liquid chromatography (HILIC) 292.3.1.1.1 HILIC of monosaccharides (from hydrolysed polysaccharides, THNS) 292.3.1.1.2 HILIC of anhydrous monosaccharides 302.3.1.2 Reversed-phase HPLC 312.3.1.2.1 Amino acids (derivatized) 312.3.1.2.2 Fatty acids 342.3.2 Detection (HPLC/MS) 352.3.2.1 Electrospray ionization 352.3.2.2 MS strategies 362.3.2.2.1 Triple-quadrupole mass spectrometer 36

8

2.3.2.2.2 Time-of-flight MS 382.3.3 Ion mobility separation 402.4. Evaluation of the used methods 443. Measurements of the target group of biomolecules in SML, aerosol and fog 443.1 Study areas 443.1.1 The High Arctic north of 80°N 443.1.1.1 The expedition in 2001 453.1.1.2 The expedition in 2008 453.1.2 Aerosol sampling in Svalbard 2015 463.1.3 Airborne PM sampling in suburban areas of Sweden in 2017 463.2 Samples collected in the the High Arctic 473.2.1 The High Arctic north of 80°N 473.2.2. Aerosol sampling in Svalbard 2015 513.3 Anhydrous monosaccharides in airborne PM10 sampled in Swedish suburban areas 564. Conclusions and future perspectives 585. References 64

9

Abbreviations

AQC 6-aminoquinolyl-N-hydroxysuccinimidyl carbamatea.s.l. Above sea levelBCI Berner cascade impactorsCCN Cloud condensation nucleiCCS Collisional cross-sectionDDA Data-dependent acquisitionDOYDp

Day of yearParticle size, measured as EAD

EADFAAs

Equivalent aerodynamic diameterFree amino acids

FWHM Full-width half maximumGC/FID Gas chromatography with flame ionization detectorHCl Hydrochloric acidHILIC Hydrophilic interaction liquid chromatographyHPLC High-performance liquid chromatographyHYSPLIT Hybrid Single Particle Lagrangian Integrated TrajectoryMIZMLR

Marginal ice zone Multi-linear regression

C4-NA-NHS N-butyl nicotinic acid N-hydroxysuccinimide esterPAAs Polyamino acids, such as peptides and proteinsPAHs Polycyclic aromatic hydrocarbonsPIPMPMx

Pack iceParticulate matter, aerosolsParticles= aerosols, up to x-μm EAD

Q-ToFMS Quadrupole-time-of-flight hybrid mass spectrometerS/N Signal-to-noise ratioSFU Stacked filter unitSML Surface microlayerSRM Selected-reaction monitoringTAAs Total amino acidsTHNS Total hydrolysable neutral sugarsTWIMS Travelling-wave ion mobility spectrometryTFA Trifluoroacetic acid UHPLC Ultrahigh-performance liquid chromatography

10

Amino acidsAla AlanineArg ArginineAsn AsparagineAsp AspartateCys CysteineGln GlutamineGlu GlutamateGly GlycineHis HistidineIle IsoleucineLeu LeucineLys LysineMet MethioninePhe PhenylalaninePro ProlineSer SerineThr ThreonineTrp TryptophanTyr TyrosineVal Valine

SugarsAra ArabinoseFuc FucoseGalo GalactosanGal GalactoseGlu GlucoseLevo LevoglucosanMano MannosanMan MannoseRha RhamnoseXyl Xylose

11

1. Introduction

1.1 Aerosol, climate and health

Aerosols, which here refer to liquid or solid particles suspended in air, are ubiquitous components of the Earth´s atmosphere with an important impact on both climate and health. Aerosols contribute directly to the surface energy budget of the Earth, resulting in cooling or warming, through scattering and absorbance of radiation. In addition, they have an important indirect effect by acting as cloud condensation nuclei (CCN) (McCormick, 1967; Charlson and Pilat, 1969). The size, chemical composition, morphology and mixing state of an aerosol population, besides relative humidity in the atmosphere, regulate the ability of the particles to serve as CCN. Depending on the concentration of CCN available, the clouds formed show different characteristics, including variation in reflectivity, lifetime, and thickness (Twomey,1977; Lohmann and Feichter, 2005). In 2013, the fifth Intergovernmental Panel on Climate Change (IPCC) assessed the components affecting the radiative forcing of climate, as illustrated in Figure 1. The panel reported that the total outcome of both direct and indirect aerosol forcing, on a worldwide average, was showing acooling (negative) effect, with the same order of magnitude as the total warming (positive) effects brought about by the anthropogenic greenhouse gases.

Figure 1. Radiative forcing of climate between 1750 and 2011 [From IPCC (2013). Climate Change 2013: The Physical Science Basis. The contribution of Working Group I to the Fifth Assessment Report].WMGHG = well-mixed greenhouse gases

The forcing agents associated with by the most uncertainty are aerosol-radiation and aerosol-cloud interactions (Figure 1). This has been least studied in the Arctic, where the uncertainty,therefore, is highest. At the same time, the Arctic has been shown to be more sensitive against the climate change with a warming rate approximately twice the global rate since the 1980s (Anisimov et al., 2007; Jeffries and Richter-Menge, 2015). There is thus a need for further

12

studies of aerosols in this region and their effects on the climate, as well as of aerosol sources and sinks.

Moreover, inhalation of airborne aerosols, or particulate matter (PM), may cause adverse health effects on humans, such as cardiovascular and respiratory diseases (Donaldson et al., 2001),asthma and reduction in lung function (Bates, 1992). This is especially the case in urban environments, where the particle concentrations temporarily can be very high. Common and large sources are traffic and heating involving combustion, but also vehicle tires. Particles up to 10μm in diameter size (PM10) can reach the alveolar region of the lung and are therefore considered as the most dangerous, while larger particles are more efficiently “filtered” by the nose and throat (Cheng, 2014). The deposition in the alveolar region implies a longer time of exposure to the particle-associated and potentially harmful compounds. Especially for the smallest particles, particles up to 2.5μm in diameter size (PM2.5), there is a clear correlation between exposure and health effects (Pope et al., 2002). The smoke from the burning of both biomass and fossil fuels can contain PM with highly varying composition of numerous health hazards, both inorganic and organic compounds. Polycyclic aromatic hydrocarbons (PAHs) are one group of organic compounds, many with suspected or proven mutagenic effects as reviewed by Mastrangelo et al. (1996). One example is the carcinogen benzo(a)pyrene (BaP).

1.2 Atmospheric aerosols

The particle size (Dp), measured as equivalent aerodynamic diameter (EAD), of atmospheric aerosols can differ from nanometers to micrometers and are classified as nucleation (Dp < 25 nm), Aitken ( 25 < Dp < 80 nm), accumulation (80 < Dp < 1000 nm) or coarse-mode (Dp > 1000 nm) particles. The atmospheric aerosol can also be referred to as “fine particles” and “coarse particles”, which are usually referring to PM with aerodynamic dp ≤ 1 μm (PM1) or ≤ 2.5 μm (PM2.5), and 1< Dp ≤10 μm (PM10) or 2.5 < Dp ≤10 μm (PM10), respectively. Research within atmospheric aerosol chemistry involves studies of chemical reactions and chemical characterization of size-segregated aerosols, for instance in the context of cloud formation, aerosol transport, lifetimes and deposition. The chemical content, as well as size, is informative for both source apportionment and assessment of health risks associated with inhalation. Natural sources of aerosols include volcanos, fly ash from forest fires, pollen, bacteria, desert dust and sea spray (salt and organic compounds) from the oceans. Anthropogenic sources are also numerous, including biomass burning, and incomplete oil and coal combustions from energy production and transport. Atmospheric PM is categorized as either primary or secondary particles. Primary aerosols are directly liberated to the atmosphere, while secondary aerosols are

13

formed in situ by nucleation of gases, for instance, NOx (nitrogen oxide and nitrogen dioxide) and SO2 (sulfur dioxide) or aggregation of primary particles. Pöschl (2005) has written an informative review of the properties and behavior of atmospheric aerosols.

1.3 Sources of atmospheric aerosols - present knowledge

1.3.1 Airborne aerosols in the Arctic region

Nowadays, it is understood that CCN may consist largely of organic compounds, not only sea salt. Recently, it has been shown that marine gel in the atmosphere originating from the sea ice and surface microlayer (SML) have a role in the cloud formation (Leck et al., 2002; Orellana et al., 2011; Bigg and Leck, 2008; Leck and Bigg, 1999, 2010). A possible outcome of a warmer climate and increased biological activity is an increased formation of CCN, creating clouds of higher albedo. Although it is highly uncertain, this could possibly counteract the warming effect on climate (Myhre et al., 2013). Marine aerosols, which originate from the sea SML, are likely transferred to the atmosphere via film drops and jet drops produced from bubbles that burst due to wind stress in the sea-air interface (Blanchard and Woodcock, 1957; Blanchard, 1975). The sea SML is roughly 100 μm thick, enriched in marine gels with biomolecules originating from the marine environment (Gao et al., 2012). It contains for instance high-molecular-weight soluble, surface-active organic compounds, secretions from phytoplankton and cell-lysed bacteria, which can be scavenged and injected into the atmosphere as described above (Blanchard, 1975; Gershey, 1983).

Sea ice largely covers the High Arctic, at latitudes higher than 80� N, with a maximum and minimum coverage in March and September, respectively. During Arctic haze in wintertime, the transport of aerosols from anthropogenic sources in Eurasia and North America increases (Heintzenberg, 1989), while in the summertime, the arriving air parcels originate from over the nearby marine sources (Stohl, 2006). Thus, in the summertime, local sources such as the open lead SML within the pack ice area may become a prominent source of CCN with implication for fog and cloud formation (Orellana et al., 2011). The chemical and physical properties that indicate the capability of Arctic aerosol to be CCNs are still not fully understood (Leck and Svensson, 2015). Furthermore, the theoretical model prediction has until now just roughly estimated the CCN concentrations (Zhou et al., 2001; Lohmann and Leck, 2005; Martin et al., 2011).

14

1.3.2 Airborne aerosols, PM, in Swedish suburban areas

Airborne PM, as those sampled in Swedish suburban areas in this thesis, originates from a mixture of several sources, such as residential heating using wood and fossil fuels, and vehicle emissions. The portion of PM corresponding to wood burning as the source is usually large in wintertime in the Nordic countries. In 2000, up to 70% of the mass of PM2.5 was estimated to originate from residential wood combustion in Nordic countries (Karvosenoja et al., 2004).

Biomass combustion, including wood burning, is an important a source of air PM worldwide with effects on human health as well as the climate (Bond et al., 2004). The aerosols generated from wood burning contain a complex mixture of organic compounds (Hedberg et al., 2002; Avagyan and Westerholm, 2017; Nyström et al., 2017), many with potentially harmful health effects, such as from PAHs and other different groups of polycyclic aromatic compounds.

1.4 Aims of the thesis

The overall aim of this thesis was to contribute with analytical tools and pieces of new knowledge in the research field of Arctic atmospheric chemistry and cloud formation. Since it has been realized recently that bioorganic constituents, and not only inorganic ions, may be important as CCN and thus for the cloud formation in the High Arctic north of 80º N, and only a minor part of the aerosol content has hitherto been characterized, the goal was to further investigate this in more detail. Another objective was to examine a possible connection between the aerosols and local Arctic sources, such as the SML from the open water of the central Arctic Ocean. To enable this study, advanced chromatography/mass spectrometry (MS) methods of sufficient selectivity and sensitivity had to be developed.

1.5 Targeted groups of biomolecules in SML, aerosols, and fog

1.5.1 Polysaccharides and monosaccharides

The marine gels in the ocean SML can be of different sizes, as illustrated in Figure 2, and are composed of extracellular polymeric substances released by bacteria, ice algae, and phytoplankton, including high-molecular-weight polysaccharides, with a fraction of proteins, non-sugar moieties like uronic acid, acetates, sulphate, esters, and lipids. Marine gels are both highly surface-active and hydrated.

15

Figure 2. Formation of marine gels from extracellular polymers (dissolved organic carbon, DOC), including polysaccharides. The picture is from Verdugo, 2012.

The polysaccharides in gels with molecular weights ranging from 10 to 30 kDa are highly diverse, with different lengths and both linear and branched structures, as illustrated in Figure 2. The polymer structures are composed of monosaccharide units including pentoses, such as D-xylose and D-arabinose, deoxy-hexoses, such as L-rhamnose and L-fucose, and hexoses, like D-glucose, D-mannose, and D-galactose, as shown in Figure 3 (Kenne and Lindberg, 1983). The portion of each different monosaccharide varies with the source of the polysaccharide. For instance, glucose and galactose are typically more abundant in bacteria, whereas the relatively higher levels of rhamnose, xylose, and mannose have been observed in gels from phytoplankton (Hoagland et al., 1993).

The contribution of different functional groups like uronic acids or sulfate in the polymer gels results in a negative net charge. Binding to divalent cations (Ca2+ and Mg2+) affects the physicochemical properties and gives the marine gel a high viscous character (Wei-Chun et al., 1998; Verdugo, 2012). The three-dimensional network of the marine gel can collapse or reassemble by removal or addition of Ca2+.

16

Figure 3. Some common monosaccharide subunits of marine polysaccharides. Picture from PubChem database.

1.5.2 Proteins and amino acids

Free amino acids (FAAs) as well as polyamino acids (PAAs), such as peptides and proteins, have been detected in different atmospheric compartments, including marine aerosols (Gorzelska and Galloway,1990; Milne and Zika, 1993; Mandalakis et al., 2010; Scalabrin et al., 2012; Barbaro et al., 2015), continental aerosols (Zhang et al., 2002; Zhang and Anastasio, 2003; Song et al., 2017), rain (Mopper and Zika, 1987; Mace et al., 2003), fog (Zhang et al., 2003) and dew (Scheller, 2001). In addition, FAAs and PAAs constitute one subclass of organic nitrogen-based substances in the sea SML originating from microorganisms and their debris. Furthermore, it has been assessed by use of TEM (transmission electron microscopy) and epifluorescence microscopy that approximately half of the identified particles collected over the Arctic Ocean are airborne bacteria and microorganisms (Leck et al., 2005). FAAs are liberated from proteins and peptides when they are enzymatically degraded by bacteria in both seawater, aerosols and fog droplets (Fuzzi et al., 1997). The potential of some amino acids (e.g glycine and leucine) to serve as CCN has been examined and the results showed that aerosols containing amino acids at atmospherically relevant mixture ratios were efficient CCN (Kristensson et al., 2009). Discrimination between D- and L-amino acids, with the only difference in rotation of polarized light, has been utilized as biomarker tools to extract information about the source of aerosols

17

(Wedyan and Preston, 2008). The D-forms are considered to originate from bacterial peptidoglycans. The overall profile and levels of the total amino acids (TAAs), i.e. the sum of FAAs and PAAs, are dependent on the emission source. In the continental atmosphere, the main sources are assumed to be plants, pollen, algae and fungi (Milne and Zika, 1993; Scheller, 2001; Zhang et al., 2003). Anthropogenic sources include tobacco smoke (Schmeltz and Hoffmann, 1977), waste and sewage treatment centers (Leach et al., 1999). The TAA profile is also dependent on the atmospheric reactivity of each amino acid, such as the tendency to oxidize or photodegrade. The TAA profile can be informative regarding the distance between the sampling site and emission source (McGregor and Anastasio, 2001). For instance, glycine is an example of a persistent amino acid that can survive long-distance transport, while methionine easily oxidizes. The L-amino acids studied in this thesis with their corresponding structures are presented in Figure 4.

Figure 4. Common proteinogenic amino acids existing in natural proteins and peptides. Picture from PubChem database.

1.5.3 Fatty acids

Similar to other types of biomolecules, fatty acids have shown to be enriched in the sea SML (Marty et al., 1979) and detected worldwide in nascent sea spray particles (Fu et al., 2011;

18

Cochran et al., 2016). Due to their surface-active properties, they could potentially play an important role in the formation of film droplets and sea-spray aerosols (Tseng et al., 1992; Quinn et al., 2015). The long-chain fatty acids released by marine algae and zooplankton serve as a fundamental factor in physiology in marine ecosystems (Sǿreides et al., 2010). The molar ratio of even to odd number of carbons may illustrate whether an aerosol sample is mainly of natural or anthropogenic origin since biological sources yield mainly fatty acids with an even number of carbons (Simoneit, 1989; Mochida et al., 2002; Cochran et al., 2016). Figure 5 shows fatty acids with different degree of unsaturation. Unsaturated fatty acids can easily undergo either microbial oxidation (Lanser, 1993) or photooxidation (Kieber et al., 1997) in samples, such as sea SML or atmospheric aerosols.

Figure 5. Top: a saturated fatty acid (CnH2nO2) Middle: example of an unsaturated fatty acid (CnH2n-2O2) Down: example of a di-unsaturated fatty acid (CnH2n-4O2)

1.5.4 Anhydrous sugars

Cellulose and hemicellulose are constituents of wood, and they both degrade and release anhydrous sugars together with a complex mixture of other compounds into the atmosphere during pyrolysis in the presence of air (Simoneit, 2002). This is illustrated in Figure 6.Levoglucosan is the major isomer formed during combustion, together with usually lower levels of mannosan and galactosan (Young, 1984b; Simoneit, 1999; Schkolnik and Rudich, 2006). While anhydrous sugars are not detectable in emissions from fossil fuel combustion (Kuo, et al., 2008), they occur as stable compounds in PM from wood burning emissions (Fraser and Lakshmanan, 2000). They are not considered hazardous compounds, but can, on the other hand, be useful as long-range transport markers for wood combustion (Fraser and Lakshmanan, 2000; Simoneit, 2002; Hedberg et al., 2006; Hennigan et al., 2010). By using these sugar markers the

19

wood combustion component of PM can be estimated. Such assessments can otherwise be difficult. Furthermore, with the anhydrous sugars as markers, the ratio between isomers can yield some information regarding the type of wood burned, more specifically if the source was mainly of hardwood or softwood (Engling et al., 2006).

Figure 6. Common anhydrous monosaccharides occurring in emissions from wood combustion.

1.6 Analytical methods - a survey

1.6.1 Mono- and polysaccharides

The characterization and determination of traces of individual polysaccharides can be analytically very demanding due to their heterogeneity and diversity in order of monomers, the degree of branching, different anomeric and chiral forms. For a total polysaccharide measurement, acidic hydrolysis into monosaccharide subunits is therefore often used, which after analysis yields information about levels and composition of the monosaccharides. Trifluoroacetic acid (TFA) is a commonly used agent for hydrolysis since it is compatible with MS (Gao et al., 2011). However, sulphuric acid (Pakulski and Benner, 1992) and hydrochloric acid (HCl)(Burney and Sieburth, 1977) are also used, but in combination with other detection methods.Measurements of monosaccharides need some consideration since they lack chromophores and fluorophores, they are highly polar and are at the same time of low vapor pressure. For gas chromatography (GC) separations, they need to be derivatized first. Non-destructive spectrometric methods, such as Fourier transform infrared spectroscopy (Russell et al., 2010),scanning transmission X-ray microscopy (Hawkins and Russell, 2010), and nuclear magnetic resonance spectroscopy (Facchini et al., 2008) have been used for identification of airborne non-derivatized monosaccharides. Spectrophotometric methods can be applied for quantification, utilizing the light absorbance from inherent aldehyde and ketone functional groups (Pakulski and

20

Benner, 1992). Chromatography-based methods for monosaccharide derivatives include GC with flame ionization detection (FID) (Skoog and Benner, 1997; Schkolnik and Rudich, 2006), and high-performance anion-exchange liquid chromatography (HPAEC) with pulsed aerometric detection (Borch and Kirchman, 1997). Capillary electrophoresis (CE) (Gao et al., 2010) with UV detection is another option providing high chromatographic resolution for different isomers while suffering from a high detection limit due to the low volumes that can be injected (Gao et al., 2010). MS methods usually yield much higher selectivity and lower detection limits than with other detection techniques. HPAEC has for instance been used with electrospray ionization–tandem mass spectrometry (ESI-MS/MS) (Saarnio et al., 2010; Asakawa et al., 2015), although requiring a rather complex setup with an ion suppressor unit. In ESI, the detection limits of highly polar analytes, for example, monosaccharides, usually benefit from mobile phases with a high content of organic modifier, facilitating the ionization/desorption process. For that reason, hydrophilic interaction liquid chromatography (HILIC)/ESI-MS/MS, utilizing mobile phases with a high content of acetonitrile, was considered a good choice for the determination of monosaccharides (Gambaro et al., 2008; Gao et al., 2011; You et al., 2016) both in Paper I and for the anhydrous monosaccharides in Paper V. MS detection of sugars can be based on either positive or negative adduct ion formation, with for instance Cs+, Na+, or Cl-, et c, (Kohler and Leary, 1995; Zhu and Cole, 2001; Rogatsky et al., 2005) or deprotonation of the analytes (Gao et al., 2011).

1.6.2 Poly- and free amino acids

For the same reason as with the polysaccharides, it is common for the determination of proteins and peptides (PAAs) to hydrolyse them into amino acid units. Both enzymes and acids can efficiently hydrolyse peptide bonds, but acids seem to be most commonly used in atmospheric studies, as mentioned in the review by Matos et al. (2016). The recoveries after hydrolysis vary depending on reaction time, temperature and the properties of the agent used (Matos et al.,2016). One common approach is to add 12 M of HCl to the samples and then incubate them for 24 h at 110°C (Yu and Schauer, 2002). For faster hydrolysis, a mixture of HCl/TFA can be used and the sample is then kept for 23 min at 156°C (Zhang et al., 2002; Zhang and Anastasio, 2003). This method was used in Paper II and for fog samples. Microwave-assisted hydrolysis with 6 M of HCl has shown to be much faster, requiring only 8 min at 160°C (Kuznetsova et al., 2005).

Different variants of HPLC/MS (Scalabrin et al., 2012; Barbaro et al., 2011, 2015), as well as CE(Soga and Heiger, 2000) can be used as quantitative and qualitative methods for underivatized

21

amino acids. Derivatization also improves the detection at analysis with HPLC/UV (Hill et al., 1979) or HPLC/fluorescence (Tapuhi et al., 1981). For GC/MS (Mandalakis et al., 2010, 2011) silylation is common, for instance with N-(tert-butyldimethylsilyl)-N-methyltrifluoroacetamide (MTBSTFA). Use of three different types of stationary phases, including reversed-phase C18, aHILIC column (Zangrando et al., 2010) and mixed-mode columns (e.g. reversed-phase with an embedded anion-exchange or cation-exchange group), is described in the above-mentioned review by Matos et al. Discrimination between L- and D-forms of amino acids has been achieved using specific chiral stationary phases (Barbaro et al., 2014) and also by derivatization into diastereomers (Kaufman and Manley, 1998). Derivatization with the commercially available and common 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC) increases the hydrophobicity and therefore improves the separation in reversed-phase HPLC (Cohen and Michaud, 1993; Mader et al., 2004). The AQC tag works for both fluorescence and MS detection.

1.6.3 Fatty acids

Fatty acids can be determined using different analytical techniques, such as conventional thin-layer chromatography (TLC), as described in a review by Fuchs et al. (2011). The drawback with TLC, as a “basic” method for fatty acid analysis, is a relatively high detection limit. However, a new technique involving TLC, coupled to matrix-assisted laser desorption ionization (MALDI)-time-of-flight (ToF)/MS, with improved detectability was presented in 2012 (Fuchs, 2012). Other methods include GC/FID (Eder, 1995) and GC/MS (Ohshima et al., 1989) with derivatization into methyl esters, HPLC/UV, where double bonds in fatty acids absorb UV-light,fluorescence, refractive index, evaporative light scattering, charge aerosol detection or MS, as described in the review by Dolowy and Pyka (2015). According to a review by Rao et al. (1995), reversed-phase HPLC has been widely used for separation and quantification of both saturatedand unsaturated fatty acids.

2. The analytical workflow in this study

A scheme of the analytical workflow for all samples analyzed in this thesis is shown in Figure 7.Due to the very low levels of the analytes, there was a high risk of contamination. All glassware, such as sample tubes and glass parts of the rotary evaporator, filter substrates were pre-cleaned prior to use. Solvents and acids were checked for contaminants and both field and lab blanks were analyzed.

22

Figu

re 7

. Sc

hem

atic

vie

w o

f th

e w

orkf

low

s in

thi

s th

esis

. FA

As

= fr

ee a

min

o ac

ids,

FAs

= fa

tty a

cids

, TA

As

= to

tal

amin

o ac

ids,

THN

S =

tota

l hy

drol

ysab

le

neut

ral

suga

rs

(pol

ysac

char

ides

), R

PLC

=

reve

rsed

-pha

se

HPL

C,

usin

g C

18,

HIL

IC=

hydr

ophi

lic

inte

ract

ion

liqui

d ch

rom

atog

raph

y. T

he s

chem

e is

div

ided

into

the

diff

eren

t ste

ps o

f th

e an

alyt

ical

cha

in, i

nclu

ding

sam

plin

g an

d st

orag

e, s

ampl

e pr

epar

atio

n,

sepa

ratio

n, a

nd d

etec

tion.

Inte

rnal

sta

ndar

ds (

IS)

wer

e ad

ded

dire

ctly

afte

r sa

mpl

ing,

exc

ept

for

TAA

s,fo

r w

hich

IS

wer

e ad

ded

prio

r to

hy

drol

ysis

.

Aer

osol

SML

Fog

Ber

ner c

asca

de

Stac

ked

filte

r uni

t PM

10 sa

mpl

er

Stor

age:

-80º

CSt

orag

e: -8

0º C

Stor

age:

-20º

C

Dp

< 6μ

mD

p <

40μm

Sampling and storageSt

orag

e: -8

0º C

Stor

age:

-80º

CSt

orag

e: -8

0º C

Sample preparation

Extra

ctio

n

Hyd

roly

sis

THN

S

TFA

5N

, 100

◦C, 2

h

Der

ivat

izat

ion

TAA

s

HC

l/TFA

, 165

◦C, 2

5min

FAA

sA

nhyd

rous

suga

r

Hyd

roly

sis

TAA

s

HC

l/TFA

, 165

◦C, 2

5min

Der

ivat

izat

ion

FAA

sEv

apor

atio

n

Separation Detection

MS

MS

MS

TWIM

S/To

FMS

THN

S

TFA

5N

, 100

◦C, 2

h

RPL

CH

ILIC

RPL

CH

ILIC

FAA

sFA

s

Der

ivat

izat

ion

RPL

CR

PLC

MS

MS

23

2.1 Sampling methods

2.1.1 Sea SML

A radio-controlled vessel, as shown in Figure 8, was employed when collecting the SML of the open leads in the Arctic pack ice, both in 2001 and 2008. The sampler consisted of a rotating drum coated with a thin layer of Teflon, which had first been treated with sodium in liquid ammonia to make it more hydrophilic and able to absorb a thin water layer on its surface. The vessel swept randomly on the surface of the open leads to gather samples as representatively as possible. The SML stuck to the submerged drum surface while rotating on the sea surface and was simultaneously poured into a pre-cleaned glass container placed inside the vessel. The collected samples were kept cool with ice in a dark container until arrival at the laboratory on the icebreaker Oden, where they were immediately frozen to -80°C.

Figure 8. A radio-controlled boat equipped with a rotating drum for sampling of the open lead SML.

2.1.2 Airborne aerosols

2.1.2.1 Multi-stage Berner cascade impactor

Airborne particles were collected on the fourth deck onboard the icebreaker, at 25 m a.s.l., utilizing 5-stage Berner cascade impactors (BCI). Aerosols of Dp > 10μm EAD at a 50-%relative humidity were not sampled, due to a size-segregated inlet. The BCI size ranges corresponded to Dp < 0.161 μm, 0.161 < Dp < 0.665 μm, 0.665< Dp < 2.12 μm, 2.12 < Dp < 5μm and 5 < Dp < 10 μm with a 50-% cut-off efficiency. Field blank samples were obtained with impactors loaded with substrates at the same sampling site and during the same sampling period, but without pumping any air through. Ambient samples and blanks were carefully handled in a glove box (free from particles, sulfur dioxide, and ammonia), both before and after sampling. All substrates for determination of organic compounds were flash-frozen in liquid nitrogen directly after unloading and stored at minus 80°C until further treatment. Sampling flow rate was 77 L/min and sampling time varied between 8 and 40 h. More details are given in apaper by Leck et al. (2001).

24

2.1.2.2 Two-stage stacked filter unit

A two-stage stacked filter unit (SFU) sampler was also employed for collection of airborne particles. The SFU cassette consisted of two Nuclepore polycarbonate filters, one with coarse pore-size and one with fine pore-size (Heidam, 1981). The coarse-size filter had a specified 50% collection efficiency of Dp 2 μm EAD and thus collected aerosols in the size range of 2 to 10 μm EAD, whereas the fine-size filter collected aerosols of Dp < 2 μm EAD (John et al., 1983). The filters in the SFU cassette were pre-cleaned with ethanol and ultrapure water before sampling and were always changed in a glove box (free from particles, sulfur dioxide, and ammonia) similar to the BCI substrates. Field blanks were obtained as for the BCI. Sampled filters were stored at minus 80 °C prior to analyses. Sampling flow rate was 17 L/min and sampling time varied between 45 and 75 h.

2.1.2.3 PM10 sampler

A sequential sampler SEQ47/50 was used, equipped with a PM10 impactor and micro-quartz fiber filters, and appropriate for outdoor sampling under different meteorological conditions. The sampling flow rate was a 1 m3/h and the sampling periods 48 and 72 h in January and July, respectively. Blank samples were obtained as for BCI and SFU. The micro-quartz fiber filters were stored frozen at minus 20 °C before the measurements.

2.1.3 Fog

A cascade impactor for collecting liquid fog water samples was mounted on the seventh deck (about 27 m a.s.l.) on the icebreaker Oden. To separate the fog droplets from the air, the impactor used two jet-impaction stages with cut-offs at 6 μm and 40 μm EAD. The sampler was connected to three pumps, resulting in a total volumetric flow rate of 530 m3/h. Samples were collected from July 5 to August 21 in 2001, and between August 3 and September 7 in 2008.Field blanks were collected as described above for the other samplers. The samples were immediately stored at minus 80 °C after sampling.

2.2 Sample preparation

2.2.1 Extraction

The next step in the analytical chain (see Figure 7) following sampling was the extraction of analytes from the sample. Important parameters when choosing an extraction strategy are, for instance, the complexity of the sample (analytes and matrix), recovery, selectivity, time- and cost-efficiency, as well as environmental aspects. The aerosol samples collected in the pristine

25

Arctic region were generally only very little polluted by anthropogenic sources, in contrast to the particles collected in the Swedish sub-urban areas, but on the other hand, they contained more sea salt and bioorganic substances from marine sources. In both cases, extraction of analytes from the filters was performed by ultrasound-assisted extraction into ultrapure water. The ultrasonication method is a static type of extraction and was therefore repeated with new solvent. Surrogate internal standards were spiked to the filters prior to extraction and all glassware used was pre-cleaned prior to use, including the glass parts of the rotary evaporator. After evaporation, the samples were reconstituted in different solvents depending on the analyte to be determined. The recoveries from the extraction were calculated by comparing standard solutions with extracts from new filters spiked with the compounds, as for the anhydrous monosaccharides in Paper V. For those, filters were spiked with 0.75 μg of each isomer. The extraction yielded recoveries close to 100%, as shown in Figure 9. The rationale for using new blank filters for this test was the expectation of a stronger interaction with the polar polycarbonate filter surface compared with the sample matrix (PM10 from suburban areas in Sweden). The high recoveries can be explained by using water as extractant, which competes with the polar-polar and hydrogen-bonding interactions between sugars and filter surface.

Figure 9. Recoveries obtained for filters spiked with 0.75 μg of each standard compound (n=3). Levoglucosan (blue bar), and the sum of galactosan and mannosan (red bar). The graph is from Paper V. The extraction was performed twice in 2x10 mL of ultrapure water.

2.2.2 Hydrolysis of biomolecular polymers

2.2.2.1 Polysaccharides in aerosol and fog

Hydrolysis was used in Paper I to enable the determination of polysaccharides, total hydrolysable neutral sugars (THNS), in the form of monosaccharide monomers, since the levels in the Arctic samples were expected to be too low for detection of the polymers. The high diversity of the polysaccharide structures, both regarding branching and monomer sequence, makes the detection

0

20

40

60

80

100

Rec

over

y (%

)

26

of a specific polymer difficult. The obtained residue from the extraction was reconstituted in 4M of TFA before incubation at 100 °C for 2 h to hydrolyse the polysaccharides to monosaccharides.TFA is sufficiently volatile to be evaporated and has also shown a high yield of the monosaccharide subunits from hydrolysed non-cellulosic polysaccharides (Neeser andSchweizer, 1984; Gao et al., 2010). The excess of acid was evaporated and the sample residue was then reconstructed in aqueous acetonitrile. In addition to the adopted method, the most important parameters for a high yield were screened by using a full factorial design (not published). Possible practical minimum and maximum settings were chosen for each variable. This design requires 2k experiments, where k is the number of screening parameters, plus triplicate experiments as center points. The significance of each variable can be investigated by this design but does not provide information about the optimum value of each variable. The disaccharide D-trehalose was used as the model compound. The obtained experimental data within each parameter domain, including the concentration of TFA (2- 5M), temperature (100-120 °C) and time (30-120 min), were processed by Modde software (ver. 10; MKS Umetrics, Umeå, Sweden). The fitted linear model shows that both time and temperature significantly improved the hydrolysis recoveries (Figure 10).

Figure 10. Coefficient plot (scaled and centered) of parameters that may affect hydrolysis recovery at a 95-% confidence level. MLR= multilinear regression

2.2.2.2 Proteinaceous materials in aerosols and fog

For the proteinaceous aerosols in Paper II, hydrolysis was chosen for the same reason as for the polysaccharides, since any occurring proteins and peptides are expected to be of highly diverse structures. In this case, only very small volumes of sample were accessible for analysis.

27

Therefore, FAAs were not determined separately; only TAAs, which is the sum of FAAs and PAAs, was measured. A typical method for hydrolysis of PAAs utilizes 6 M of HCl at 110 °C, for at least 24 h (Hirs et al., 1954). In Paper II, a faster method previously presented by Tsugita and Scheffeler (1982) was used, utilizing a mixture of 12 M of HCl and TFA (>99%) at 165 °C for 25 min. Both methods have been found to be equally efficient for some model proteins, but show discrimination of asparagine and glutamine (due to hydrolysis to aspartate and glutamate, respectively), as well as of tryptophan (Tsugita and Scheffeler, 1982). In reality, it is difficult to know the exact hydrolysis efficiency for the proteins and peptides in the aerosols, since it might be both compound and matrix dependent. In this work, the theoretical composition in percent of each amino acid was compared with the experimentally derived after hydrolysis of HSA. For lysine, alanine, valine, cysteine (measured as cystine), isoleucine, leucine, the sum of glutamate and glutamine, methionine, histidine, arginine and serine, the representation was between 70 and 113% of the theoretical. For phenylalanine and proline, the corresponding values were within 125-130% and for glycine, threonine and the sum of aspartate and asparagine within 51-63%. The values for tyrosine and tryptophan were found to strongly deviate from the theoretical. The reason for this was not further investigated, but these two latter amino acids were not quantified in any of the aerosol samples. Before hydrolysis, deuterium-labelled internal standards were added to the aerosol sample extracts. After hydrolysis, the acid was evaporated and the sample residue was subjected to derivatization (see section 2.3.3) before ultrahigh-performance liquid chromatography (UHPLC)/ESI-MS/MS analysis.

Figure 11. Hydrolysis of HSA. Blue bars show LC/MS response when using HSA with 6M of HCl for 20 h and red bars 12M of HCl:TFA (2:1 v/v) for 25min.

0

300000

area

28

Proteinaceous material was also measured in fog (unpublished results), using the same method for hydrolysis and sample preparation. However, since larger sample volumes were accessible in this case, FAAs (analysis without hydrolysis) and TAAs could be measured separately.

2.2.3 Sample derivatization of free amino acids

Usually, both response and chromatographic separation improve in reversed-phase HPLC/ESI-MS when amino acids are derivatized to more hydrophobic compounds. The improved response is most likely due to an increased ability of the derivatives to gather at the surface of the electrospray droplets, being more easily ionized and desorbed into the gas phase. A common and commercially available derivatization reagent for ESI-MS measurements of amino acids is AQC,as mentioned in section 1.6.2. This reagent forms derivatives that in the MS/MS experiment produce a fragment that is common for all derivatized amino acids, m/z 171, which is specific for the AQC tag. In Paper II, we used another reagent that forms specific fragment for each amino acid (except for isoleucine and leucine derivatives which show identical MS-spectra), namely N-butyl nicotinic acid N-hydroxysuccinimide ester (C4-NA-NHS), which was synthesized by our co-authors from the University of Leuven, Department of Chemistry (Heverlee, Belgium) (Yang et al., 2006). The specific MS fragments are highly advantageous for the identification. The formed derivatives from this tag contain a permanent charge in the form of a quaternary amine function, embedded in a hydrophobic structure, see Figure 21. Tagging with this reagent was shown previously to improve both reversed-phase separation and ESI response compared to non-derivatized amino acids (Ren et al., 2004; Yang et al., 2006). Derivatization was carried out by addition of C4-NA-NHS solution to a mixture of sample and sodium borate buffer at pH 9. The reaction was allowed to proceed for about 1 min, prior to evaporation and reconstitution in aqueous acetic acid followed by UHPLC/ESI-MS/MS.

2.3. Instrumental methods

2.3.1 Liquid chromatography

There are several types of liquid chromatographic methods (Ettre, 1993), for which the basic principle is partitioning of the analytes between a stationary phase inside the chromatographic column and a liquid mobile phase flowing through the column. Analytes, having different partition coefficients between the stationary phase and the mobile phase of a chromatographic system, will migrate with different rates through the column resulting in a physical separation and different elution/retention times. In this work, different HPLC methods were applied

29

depending on the chemical properties of the analytes. For polar compounds and compounds with polar functional groups, it can be useful to utilize normal-phase HPLC This term refers to an HPLC system consisting of a stationary phase with a higher polarity than the mobile phase. The separation mechanisms of such a system are for instance based on dipolar interactions or hydrogen bonding, leading to retention of hydrophilic compounds. One of the early normal-phase HPLC methods was using unmodified silica adsorbent in combination with a lipophilic solvent, such as hexane, as the mobile phase. A more recent type is hydrophilic interaction liquid chromatography (HILIC), which utilizes a polar stationary phase modified with a thin water layer. Such a stationary phase can be used with a mobile phase that is more compatible with ESI-MS. In this case, the partition of analytes occurs between the hydrogen bonded water layer and the mobile phase, the latter usually consisting of acetonitrile with a few percents of water or aqueous ammonium acetate. The interactions can be, as for the older versions of normal-phase HPLC systems, electrostatic interactions, hydrogen bonding or dipolar interactions. A review by Hemström and Irgum on HILIC was published in 2006. Reversed-phase liquid chromatography is better suited for more hydrophobic compounds and is based on the separation between a less polar, more hydrophobic stationary phase (most common a C18-coated silica) and a more polar, usually aqueous, mobile phase. The main retention mechanism is based on dispersive forces (van der Waals forces) giving stronger retention of larger, hydrophobic analytes. The retention time is highly correlated to the lipophilicity (log P) of the analyte and to a minor extent the shape of the molecule.

2.3.1.1 Hydrophilic interaction liquid chromatography (HILIC)

Monosaccharides are highly hydrophilic, polar compounds. As an example, the log P value for glucose is around -3, and thus difficult to retain in conventional reversed-phase HPLC columns,such as C18. For these compounds, HILIC has become a method-of-choice. To ensure a stable, reproducible water layer in the column, the acetonitrile mobile phase should contain a few percents of water, and reproducibility is easier to ensure at isocratic conditions. Using gradient elution usually requires long equilibrium times at the initial solvent composition. In this work, several columns were tested for their performance. Propylamine silica and pentahydroxy silica were selected for the separation of THNS monosaccharides (Paper I, III) and the anhydrous monosaccharides (Paper V), respectively. The other tested columns exhibited lower resolution between diastereomers and peak splitting for anomers.

2.3.1.1.1 HILIC of monosaccharides (from hydrolysed polysaccharides, THNS) The propylamine silica column used in Paper I and III had previously been evaluated for separation of monosaccharides by Gao et al. (2011). This stationary phase exhibited good peak

30

shapes, but not all target analytes could be fully baseline-separated. A possible drawback with this stationary phase is the risk that the column performance deteriorates with time due to the formation of Schiff bases between the aldehyde groups of the monosaccharides and the amino groups of the stationary phase. In this work, this was not a problem and the column performance was satisfactory throughout the project. Figure 12 shows a chromatogram of monosaccharide standards and fog samples from the High Arctic.

Figure 12. Typical LC/ESI-MS/MS chromatograms obtained from monosaccharide left: standards and right: a fog sample collected in the High Arctic at 2nd Aug (DOY 215), 2008.

2.3.1.1.2 HILIC of anhydrous monosaccharidesIn Paper V wood burning markers, i.e. anhydrous monosaccharides, were separated using a pentahydroxyl silica stationary phase. By using this column, the main target compound levoglucosan could be fully separated from mannosan and galactosan with a total analysis time of only 5 min. The two latter compounds could not be separated on any of the tested columns and were therefore quantified together. Figure 13 shows an example of a chromatogram of

31

anhydrous sugars extracted from airborne PM10 particulates collected in January 2017 at Enskede, Stockholm. A good separation was achieved, despite a short time of analysis, and by using isocratic elution a high reproducibility in retention was obtained.

Figure 13. LC/ESI-MS/MS total ion current chromatogram (TICC) and extracted ion chromatograms of anhydrous monosaccharides in an extracted sample from airborne PM10 taken in January 2017 at Enskede, Stockholm. The figure is from Paper V

.

2.3.1.2 Reversed-phase HPLC

2.3.1.2.1 Amino acids (derivatized)

Naturally occurring amino acids constitute a compound class with a large variety of polarities,due to their different side chains, which can be more or less hydrophilic and either ionizable or neutral. Due to this, it is difficult to obtain single, suitable column for their separation, for HILIC, as well as for reversed-phase HPLC. A strategy is to derivatize the amino acids to obtain improved retention and separation on a reversed-phase HPLC system. Initially, two-dimensional HPLC (2D-HPLC) with orthogonal retention mechanisms, was set up in this project in order to obtain as good separation as possible of the derivatized amino acids (unpublished work).

The 2D-HPLC was coupled to an ESI-MS system for detection. C18 was used as stationary phase for chromatographic separation in the first dimension (1D), while a strong cation-exchange (SCX) column was used in the second dimension (2D). During the first three minutes of elution

32

from the first dimension, the fast eluting most hydrophilic amino acids were directed to and loaded on the SCX column (Figure 14, left). After 3.0 min, the two six-port valves were switched, and the flow from the first dimension was directed to the MS for detection of the eluting analytes (Figure 14, right). After 12.5 min run time, the gradient program in the first dimension was finished and Pump 1 stopped. A gradient elution of the compounds loaded on to the second dimension column was then started using Pump 2. After a run time of 13 min, the elution of the second dimension column was completed. The two six-port valves were switched to their initial position and both columns were conditioned for two minutes, which gave a total run time of 27.5 min. The mobile phases were acidified mixtures of Milli-Q water andacetonitrile. A gradient from low to high concentration of acetonitrile was used in the first dimension, and vice versa in the second dimension. On C18 the retention of the amino acid analytes increases with the hydrophobic tag introduced into the molecule, while on the other hand, a positive charge on the analyte gives strong retention on the SCX column. Furthermore, and as described above in section 2.2.3, the tag also increases the ESI responses.

Figure 14. An online 2-D LC/ESI-MS system developed for separation of tagged amino acids (not published).

Figure 15 shows examples of the separation achieved by this 2-D system. Even though a very good separation was achieved with this system, one-dimensional reversed-phase UHPLC on C18

was used in Paper II, the reason being the advantage of using a well-known and commonly

33

available system that is simpler to set up and faster.

Figure 15. TIC obtained from 2-D LC/MS/MS system, the first dimension is until 12.5 min and the second dimension is between 12.5 and 25.5 min. For abbreviations, see p 10.

UHPLC benefits from a higher chromatographic resolution and a higher peak capacity, due to the small particle size (1.7μm), and at the same time runs can be done faster. Also, the solvent consumption is generally lower with UHPLC compared to conventional HPLC, which is beneficial from a green chemistry point of view (Nováková et al., 2006). A drawback is a highpressure needed for the separation, which requires pump systems able to deliver pressures up to 1200 bar. Gradient elution is usually repeatable when using reversed-phase systems and in this case, a C18 column was used for separation of the amino acid derivatives. An MS/MS chromatogram of standard amino acid derivatives is shown in Figure 16. The chromatogram contains 19 peaks of the 20 amino acid derivatives, including the isomers leucine and isoleucine. Almost all of them were well-resolved within a 10 min run. Only threonine and alanine were not separated chromatographically but were easily differentiated by their specific m/z transitions.

34

Figure 16. UHPLC/ESI-MS/MS total ion current chromatogram of a standard solution containing tagged amino acids. For abbreviations, see p 10.

2.3.1.2.2 Fatty acids

A fast screening method for fatty acids in SML from the open leads within the Arctic pack ice was developed in Paper IV. For the non-target/suspect screening in this complex type of sample matrix, a UHPLC system with gradient elution coupled to ToFMS was employed. A review by Gosetti et.al. on non-target screening of contaminants in water using this analytical method has recently been published (2016). The chromatographic separation in Paper IV needed only 11 min and there was no need for optimization of the LC separation; it was mainly used for a rough clean-up step from the salt-rich matrix. For desalting, the first 0.5 min of the chromatographic elution was diverted to waste. To improve the confidence of identification, travelling-wave ion mobility MS (TWIMS) was coupled between the UHPLC and the ToFMS as a second stage of separation. The achieved UHPLC/ToFMS chromatogram is shown in Figure 17.

Figure 17. UHPLC/MS total ion current chromatogram of Arctic sea SML sampled in August 2001. The figure is from Paper IV

35

2.3.2 Detection (HPLC/MS)

2.3.2.1 Electrospray ionization

ESI has been shown to be applicable for a wide range of compounds, from small organic molecules to large biomolecules, and was found to be the most suitable ionization method also for the biomolecules investigated in this project. Atmospheric pressure chemical ionization (APCI) was also tested for the monosaccharides but without any improvement of detection limits. Ion transfer from the electrospray droplets to the gas phase, i. e. ion desorption, is considered likely to occur by either of two different mechanisms, the ion-evaporation mechanism (IEM) (Iribarne and Thomson, 1976) involving field desorption of smaller molecules, and a charge-residue mechanism (CRM) for macromolecules (Dole et al., 1968). ESI is not a universal technique, such as electron ionization, and shows highly compound-dependent response factors. For instance, the hydrophilicity/hydrophobicity plays an important role in the ESI process. When using aqueous mobile phases, hydrophobic compounds often show better responses, due to their tendencies to gather at the droplet surface and therefore more easily become ionized/desorbed. To improve both the ionization efficiency and C18 retention of the amino acids (Paper II), they were derivatized as described above (section 2.2.3 Sample derivatization). The formed derivatives contain both a permanent positive charge and a hydrophobic moiety. The derivatized amino acids showed varying responses (Figure 18).

Figure 18. The variability in response on ESI-MS between the different amino acid derivatives. Each response is normalized to Leu. Cys* was measured as cystine, i.e. two cysteines bound by an S-Sbridge. For abbreviations, see p 10.

Those with hydrophobic side chains, such as derivatives of leucine, isoleucine, valine, andphenylalanine, showed higher signals than derivatives of arginine and histidine. Lysine and

0.00

0.50

1.00

36

cystine are both hydrophilic compounds but were derivatized by two molecules of the tag and thereby achieved two charges, altogether increasing their responses.

2.3.2.2 MS strategies

Either low-resolution or high-resolution mass spectrometry was selected for the projects in this thesis, depending on if either target or suspect screening/profiling was to be performed. Aschematic overview of the two different screening approaches is shown in Figure 19.

Figure 19. A schematic overview of mass spectrometry strategies used in this study.

2.3.2.2.1 Triple-quadrupole mass spectrometer

Triple-quadrupole MS was used for quantification of polysaccharides (THNS) after hydrolysis to the monosaccharide subunits in Papers I and III, for FAA derivatives (unpublished) and also for the proteinaceous matter in Paper II, as well as for anhydrous monosaccharides in PM10 (Paper V). This analyzer exhibits a low resolving power, R=m/�m, with �m values (full-width half maximum, FWHM) between 0.2 and 0.7 over the entire m/z range. A high selectivity is obtained by running MS/MS in selected reaction monitoring (SRM) mode (Figure 20), which usually gives very high S/N ratios, as long as the analyte can be efficiently ionized and there is one or a

Screening

Suspect screening

Suspect listMatching of Theoretical mass Peak detection Matching of isotopic pattern Matching of MS/MS pattern Matching of predicted retention time Matching of predicted drift time

Target screening

List of potential present suspects

Target listReference standards Matching of retention times Matching of MS/MS pattern

Quantification of targets

Triple-quadrupole Ion mobility/Q-TOF

37

few fragments of suitable intensity. The triple-quadrupole is also highly repeatable for quantification and has a very high duty cycle in SRM mode when it measures only the ion transitions of interest. Figure 20 shows the basic principle of SRM in a triple-quadrupole.

Figure 20. A schematic picture of a triple-quadrupole working in SRM mode. The green spheres correspond to the analyte ion of interest, corresponding to the only m/z that is selected to pass Q1. Q2 is an RF-only collision cell, where fragmentation of the accelerated analyte ion occurs (collision-induced dissociation, CID) after collision with an inert gas, such as argon. In Q3, only selected fragment ion(s) can pass to the detector.

Fragment ion transitions and possible fragmentation pathways for monosaccharides using ESI-MS/MS have been described in detail by Gao et al. (2011).

The measurements of amino acid derivatives in this work were done using two transitions; one specific for each analyte, and one specific for the tag, which implies a higher selectivity compared with the commonly used AQC tag. Figure 21 shows spectra using product ion scan with triple-quadrupole MS for the serine derivative and levoglucosan, respectively.

In an attempt to differentiate between co-eluting galactosan and mannosan, one specific MS/MS transition was used for each of them, m/z 161>113 for galactosan and m/z 161>129 for mannosan (Dye and Yttri, 2005). However, for unknown reasons, these transitions did not yield repeatable responses with our MS system. The reason for this was not further investigated and it was decided to quantify galactosan and mannosan together.

Detector

m/z

Q1 Q2 Q3

38

Figure 21. Spectrum of a standard solution of derivatized serine (upper), and levoglucosan in PM10 (lower)obtained with HPLC/ESI-MS/MS in product-ion scan mode.

2.3.2.2.2 Time-of-flight MS

Q-ToFMS was used for suspect screening of fatty acids in open lead SML from the High Arctic (Paper IV). A simplified scheme of the used instrument is shown in Figure 22. The ToF analyzer, which records the flight time of accelerated ions as a measure of m/z, generates accurate mass (at calibration with an error less than 1 ppm) and the isotopic pattern of the precursor ions (with <6% deviation from the theoretical values) and is also fast “scanning”.

The resolving power in “V” mode is between 10000 and 20000 FWHM over the entire m/zrange. “V” mode, refers to the use of one of the two reflectrons (Figure 22). When in “W” mode(not shown), R can reach 40000 FWHM.

39

Figure 22. A schematic overview of a Q-TOFMS working in “V” mode, providing a resolving power R=m/�m from 10000 to 20000 FWHM.

Initially, the total ion current chromatogram was extracted for ions within the entire mass range corresponding to negative ions of suspected fatty acids. The obtained spectra were then processed by smoothing, centering and lock mass correction using the function available in the Masslynx 4.1 package. For instance, m/z 283.2719 was corrected to 283.2638, by use of “lock mass” calibration with the negative ion of leucine enkephalin at m/z 554.2615. The corrected m/zvalues were then utilized for elemental composition assignment. Figure 23 shows the calibrated spectrum of octadecanoic acid and the raw spectrum, respectively.

Figure 23. Spectra of octadecanoic acid. Top panel: calibrated mass spectrum, with lock mass spectrum inserted,both in centroid data mode. Lower panel: the raw mass spectrum obtained in profile mode.

40

In Q-ToFMS, the quadrupole first isolates a certain precursor ion to be accelerated and fragmented in the collision cell. The formed fragments are then transferred, “pushed”, into the high-resolution ToFMS analyzer. The tandem MS contributes with more possibilities in non-target screening, due to the additional information from product-ion spectra (Gosetti et al., 2016).Figure 24 shows the product-ion spectrum of octadecanoic acid obtained by fast data-dependent acquisition (DDA). Fast DDA scan includes a low-energy pre-run as part of the scan to find suitable precursor ions for fragmentation. DDA uses intensity threshold, charge state, isotopic pattern and exclusion/inclusion lists and then performs MS/MS scan on each screened precursor.

Figure 24. Product-ion spectrum of octadecanoic acid obtained in DDA mode with collision energy ramped between 10 and 50V.

2.3.3 Ion mobility separation

Ion mobility separation (IMS) is an orthogonal dimension bringing additional information complementary to retention time, m/z values and fragmentation pattern given by the HPLC/ESI-Q-ToFMS analysis. The ion separation occurs within milliseconds, on the contrary to the very fast microsecond separation in the ToF, and the drift time of an ion is based on its collisional cross-section. Commercial ion mobility instruments today include field asymmetric waveform IMS (FAIMS), drift tube IMS (DTIMS), ion trap IMS and travelling-wave IMS (TWIMS). TWIMS was used in the fatty acid project described in Paper IV and is a compartment between the collision cell and the flight tube of the Q-ToFMS described above, see Figure 22. The TWIMS drift tube consists of stacked ring electrodes applied with a combination of RF voltage, which radially confines the ions, and DC voltage, which axially transfers the ions through the cell (Figure 25). As the countercurrent flow of inert gas molecules collides with the analyte ions,

41

ions with a higher collisional cross section (CCS) decrease in mobility, resulting in longer drift times. The theoretical relation between instrumental settings, CCS and TWIMS drift time is not yet fully understood, but CCS determination is performed in practice by calibration with compounds of well-known CCS (Giles et al., 2011).

Figure 25. A schematic overview of a TWIMS. Wave refers to RF-voltage, wave height to RF amplitude.

Mannose and glucose (Paper I, and unpublished results for fog samples) could not be chromatographically separated on the aminopropyl silica column, and therefore separation of these two monosaccharides was tested in TWIMS (not published). In an attempt to increase the difference in CCS, adduct formation was evaluated with Li+, Na+, Co+, K+, Zn+, Mg2+ and Cs+

.

Metal adducts with disaccharides and oligosaccharides in IMS have previously been studied (Huang and Dodds, 2013). The metal ions were added to the reference solutions before injections and the metal-sugar adducts were investigated in full scan mode. Nitrogen was used as drift gas in the TWIMS cell. Adducts were observed for most of the metal ions, except Mg 2+ and Cs+.TWIMS parameters such as wave velocity (RF frequency), wave height (amplitude) and different IMS gas pressure settings were evaluated. The best separation was altogether observed for the Na+ adducts (Figure 26). However, the difference in drift time was not considered sufficient for further use for the monosaccharides. For the future, it would be of interest to evaluate the performance of other types of drift gases.

42

Figure 26. Upper horizontal: Overlaid drift time plots of (glucose+Na)+ and (mannose+Na)+.

Vertical: ToFMS spectrum. Both isomers have the same m/z at 203.1.

In Paper IV, a combination of UHPLC, TWIMS, and MS was used for quickly screening fatty acids in Arctic sea SML. The TWIMS drift times, as orthogonal information to the data from UHPLC/ESI-QToFMS were used to increase the confidence in the identification of the fatty acids. Predicted drift times, obtained by calibration with only a few fatty acid standards, versusmeasured drift times of assigned fatty acids in the sample, are shown in the diagram in Figure 27.

Figure 27. Correlation between the predicted drift times versus measured drift times. The figure is from Paper IV.

43

No sample preparation other than evaporation and centrifugation was required. UHPLC, with common C18 as stationary phase, was used mostly for desalting the sea SML sample and for estimation of retention time, without any optimization for best possible separation of the fatty acids. Instead, TWIMS was used to further separate them, and the TWIMS peaks were integrated in the driftograms. Since a method for profiling was the goal, yielding relative responses only, we performed no absolute quantification of the compounds and used no internal standards for calibration of the LC/MS responses. Both C18 retention time and TWIMS drift times were calibrated with external standard solutions of saturated and non-saturated fatty acids. Predicted log P values, correlated with C18 retention, and TWIMS drift times showed both a very high correlation with measured values. ToFMS further contributed with accurate mass data, elemental composition, double bond equivalent values (DBE), isotopic pattern and fragmentation pattern.Overall, 25 peaks could be assigned as fatty acid candidates from C8 to C24, both saturated and unsaturated forms and the entire profiling required only an 11-min run.

The UHPLC/ESI-TWIMS/ToFMS method was compared with UHPLC/ESI-ToFMS without the use of the TWIMS cell. When quantification was performed with the latter, peak integration was difficult for low abundant fatty acids since the noise level in the UHPLC run was too high, as well as due to peak tailing. The additional use of TWIMS yielded detection of more fatty acid peaks, as shown in Figure 28, pie chart to the left. For comparison, a rough estimation of the relative signals with UHPLC/ESI-ToFMS is also shown (pie chart to the right).

Figure 28. The profile, i.e. relative intensities of fatty acids in an Arctic sea SML sample collected in the high Arctic (89°N) in August 2001. Left: relative intensities obtained by integration of assigned peaks in the TWIMS/ToFMS driftograms. Right: relative intensities obtained by integration of the chromatographic peaks in the UHPLC/ToFMS method without TWIMS. The figure is from Paper IV.

44

As can be observed in Figure 28, the relative signals, achieved from integration of UHPLC/ToFMS and TWIMS/ToFMS peaks, respectively, showed some similarities in the pattern. However, the decrease of the noise level by use of TWIMS made it possible also to integrate and obtain relative responses for the low abundant fatty acid peaks in driftogram. Lanucara et al. (2014) have in a recent review described the usefulness of different ion mobility techniques, including TWIMS, for an increased detectability of peptides and other biomolecules in Q-ToFMS.

2.4. Evaluation of the used methods

Different methods were used for the applications in this thesis. The performance of each, in terms of instrumental detection limits (injected amount on HPLC column), method LOD (amount per sampled air volume), measured linear ranges, precision and accuracy were tested.The calculated instrumental LOD values are presented in Table 1.

Analyte His Asn Ser Gly Arg Asp Gln

Instrumental LOD (pg) 3.97 5.89 1.30 0.86 36.92 4.02 2.14

Analyte Pro Glu Thr Ala Tyr Val Met


Analyte Lys Cys Ile Leu Phe Try Xyl


Analyte Ara Rha Fuc Glu+Man Gal Levo Mano+galo

Instrumental LOD (pg) 1.5 1.0 4.3 5.7 0.12 4 10

Table1. Instrumental LOD of each analyte measured in this thesis is calculated based on the injected amount on the HPLC column. For abbreviations, see p 10.

3. Measurements of the target group of biomolecules in SML, aerosol and fog

3.1 Study areas

3.1.1 The High Arctic north of 80°N

Amino acids were measured in atmospheric aerosols (Paper II) and fog (unpublished results), and total hydrolysable polysaccharides, THNS, in both aerosols (Paper I, III) and fog (unpublished results). The open lead SML was also screened for fatty acids, utilizing a novel analytical method (Paper IV). All these samples were collected in the summertime when the

45

biological activity in the pack ice area is high. The studies included in this thesis were part of the Arctic Ocean Expedition (AOE-2001) and the Arctic Summer Cloud Ocean Study (ASCOS) in 2008, respectively, both performed on the Swedish icebreaker Oden. One common goal of these multidisciplinary expeditions was to increase the understanding of low-level cloud formation over the central Arctic Ocean and related effects on the surface radiation budget and thus melting or freezing of the ice.

3.1.1.1 The expedition in 2001

This expedition started from Gothenburg, Sweden, on June 29, i. e. day 180 of the year (DOY180) and continued northwards through the “marginal ice zone” (MIZ), covered by 20-70% of ice. The icebreaker then moved further towards the “pack ice region” (PI) with 80 to 95 % of ice coverage in the central Arctic Ocean. Some of the samplings and measurements started on July 5(DOY 186) near Svalbard in the direction towards the North Pole, where sampling continued from August 2 (DOY 215) when Oden was anchored to an ice floe ~ 89N, 1W (Figure 29). Both samplings and meteorologically observations were then carried out during nearly twenty-one days when the sun was always above the horizon (Tjernström et al., 2004). Paper IV presents results from the analyzed samples collected in this expedition.

Figure 29. Map of Oden’s track 2001 toward the High Arctic with given dates and locations for the atmospheric monitoring stations. The inset shows the track while drifting with a large ice floe during the sampling period. The picture is from Tjernström et al. (2004).

3.1.1.2 The expedition in 2008

The expedition in 2008 departed from Longyearbyen, Svalbard, on August 2 (DOY 215) and traveled to the north through the MIZ until reaching the PI in the central Arctic Ocean. The first measurement took place at an open water station on August 3 and then at the MIZ on August 4,with both stations located in the Greenland Sea-Fram Strait region. The samplings, atmospheric

46

measurements and observations then continued when Oden was moored to a large ice floe (referred to PI-drift), slightly north of 87N (87.4N, 1.5W), until early September 2 when the icebreaker started to return southwards (Figure 30). As for the 2001 expedition, the sun was always above the horizon (Tjernström et al., 2014). The analyses of samples collected in this expedition are described in Paper I.

Figure 30. Map of Oden’s track toward the High Arctic 2008. The left-hand part of the track presents the initial track to the north and the right-hand track shows return track to the south. The picture is from Tjernström et al. (2014).

3.1.2 Aerosol sampling in Svalbard 2015

The aerosol sampling took place at the Zeppelin Mountain, near Ny Ålesund (79° N, 12° E) at 474 m a.s.l. from August 2014 to December 2016. The sampling was part of a multidisciplinary program that studied atmospheric chemistry and meteorology in the Arctic region to raise the knowledge regarding remote marine aerosols originating from within the marginal ice zone or in the open waters south thereof. The Norwegian Institute for Air Research (NILU) coordinated the program. Paper II and III present analyses and results from the study.

3.1.3 Airborne PM sampling in suburban areas of Sweden in 2017

Air measurements were simultaneously performed at three different suburban sampling stations, namely Ytterjärna (59°05'N, 17°34'E) near Södertälje, Enskede (59°16'N, 18°02'E) in Stockholm and Delsbo (61°48'N, 16°33'E) near Hudiksvall, during January and July, respectively. Paper V describes this work.

47

3.2 Samples collected in the High Arctic

3.2.1 The High Arctic north of 80°N

One of the aims of this thesis was to investigate the content of polysaccharides in the atmospheric aerosol (Paper I) and thereby gain further knowledge on the probability of these compounds to act as CCN, with subsequent formation of fog and cloud droplets. Aerosols and fog, simultaneously sampled during the summer of 2008, were screened for THNS content. As can be observed in Figure 31, the temporal trends were in agreement (not published). However, ahigher concentration was observed for THNS in aerosols compared to fog (not shown). This difference can be explained by that only a fraction of the total number of aerosols, the accumulation (0.08 -1μm) and coarse (1-10μm) modes, are able to act as CCN and form fog droplets.

Figure 31. The relative levels of total hydrolysable neutral sugars (THNS) in aerosols and fogs simultaneously collected in the High Arctic in the summer of 2008. FW(40) = fog water droplets < 40μm.

Sea spray was generated in situ during the ASCOS 2008 expedition (Paper I). This was performed by artificial bubble bursting at an open lead site. The relative content of THNS subunits, classified as pentoses, hexoses and deoxy sugars, in sea spray, in airborne aerosols, and in fog samples is illustrated in the pie charts in Figure 32. The overall pattern is very similar in

48

all three matrices, except that the fog samples showed the lowest relative content of deoxy sugars (Figure 32, middle panel, to the right). This is most likely related to a relatively low abundance of deoxy sugars in both accumulation- (~10-1 to 1 μm) and coarse-modes (1 to 10 μm) of the atmospheric aerosols over the pack ice (Figure 32, lowest panel).

Figure 32. Top and middle panels: monosaccharide composition in SML*, sea spray, aerosols, and fog collected simultaneously in 2008 in the High Arctic. Lowest panel: size distribution of particulate pentoses,hexoses, and deoxy-sugars collected over the pack ice. * The pie chart is calculated based on the results from Gao et al. (2012).

The developed UHPLC/ESI-MS/MS method, described in Paper II and used for analyzing Svalbard aerosols for TAAs in 2015, was also used to determine proteinogenic compounds in fog

49

samples collected over the pack ice area in the High Arctic in 2001 and 2008. In fog, however, FAAs and PAAs could be quantified separately, since larger sample volumes were accessible than for the Svalbard aerosols in 2015. Derivatization was performed directly on one fraction of the sample, yielding the FAA concentration, while the other fraction was first hydrolysed and thus yielded the TAA level. The PAA levels were then obtained by subtracting the FAA levels from the TAA concentrations. Significant higher levels of PAAs compared to FAAs were found,indicating that proteins and peptides are the major sources of the amino acids also in fog. According to data summarized in the review by Matos et al. (2016), these results seem to be in accordance with other studies of marine aerosols sampled in remote areas around the world, which also show higher levels of PAAs compared to FAAs. On the contrary, areas more polluted from anthropogenic sources generally show higher relative levels of FAAs.

An interesting observation regarding the fog samples is that there was a higher concentration of both FAAs and PAAs in 2008 compared with 2001. The reason could be the High Arctic atmosphere being more exposed to open waters in 2008, with a 22-% shrinkage of the ice cover between the years (Figure 33).

Figure 33. Ice coverage in the High Arctic during August in 2001 (left) and 2008 (right), respectively. Orange area: trajectory frequencies of arriving air parcels to the sampling point overlying the ice concentration map, for the period when Oden was moored in the High Arctic. The ice coverage maps were taken from the website of National Snow & Ice data center (NSIDC).

50

The amino acids glycine, serine, and alanine were the most abundant amino acids among both FAAs and PAAs in the fog samples (unpublished) collected over the PI. This pattern is most likely due to plankton and marine suspended particulate matter as main sources, which both have been reported to contain relatively high levels of alanine and glycine (Lee, 1988).

When the temporal trends of the measured concentrations of THNS (Paper I) and TAA in the fog samples (unpublished) collected in 2008 are compared, similarities can be observed, as shown in Figure 34. Altogether, this suggests common sources of biogenic origin, such as from the marine materials mentioned above. Both polysaccharides and proteins can promote fog droplet activation due to surface-active properties and most likely constitute an important factor in cloud droplet activation and cloud formation in the Arctic.

Figure 34. Variation in the relative levels of atmospheric THNS and TAAs sampled in the High Arctic during August 2008. Blue line: THNS in aerosols. Green line: TAAs in fog, Dp < 6μm. Black line: TAAs in fog, Dp < 40 μm.

Fatty acids were screened for in sea SML from the PI area (Figure 28), which is described in Paper IV. One result from the SML screening was more intense signals from fatty acids with an even number of carbon atoms. Although the exact individual response factors are not known

51

without reference compounds, it is reasonable to assume they are roughly similar within the same compound class having similar molecular weights. The highest signals were obtained for hexadecanoic acid and octadecanoic acid. The carbon preference index (CPI) number obtained, i.e. the ratio between even-numbered and odd-numbered saturated fatty acids detected in the SML sample was approximately 55. Mixtures with CPI numbers above 1 are considered to be of mainly biogenic origin (Cochran et al., 2016). The obtained CPI seems reasonable for sea SML from the High Arctic, reflecting almost no anthropogenic pollutants but predominantly biomolecules from bacteria, algae, phytoplankton, and secretions, etc.

The developed method is novel and should be useful for profiling fatty acids also in aerosols and fog. A less laborious and time-consuming analytical method is valuable when there is a need for many analyses, such as when studying temporal trends. The CPI index, as well as the overall profile of fatty acids, may also be applied to atmospheric samples and add useful information in source apportionments. Furthermore, the oxidation of fatty acids may be of interest to study, yielding information regarding the aging of aerosols.

3.2.2. Aerosol sampling in Svalbard 2015

For the aerosols sampled at the Zeppelin station in Svalbard during the entire 2015, one of the aims was to identify the sources and to investigate if fingerprints of each of those could be obtained. The tool used for this was positive matrix factorization (PMF) (Paatero, 1997). The results from this study are presented in Paper III.

Aerosols of Dp < 2 μm EAD (PM1) were collected, using the two-stage SFU sampler, as described in section 2.1.2.2. Several types of analysis apart from THNS determination were performed, such as ion chromatography for the inorganic salts and visible photometric detection for black carbon. The PMF model could identify a number of sources (factors). For instance, sea salt (ss) could be resolved as one factor with a high relative contribution from Na+, Ca+, ss-K+,ss-SO4

2-, i.e. a high variability in percentage of their total concentration in PM1 (see Figure 35a). Differentiation between ions of marine origin and non-sea salt (nss) ions in the aerosols uses the known Na+/SO4

- ratio, so that [nssSO4-] = [SO4

-] - (0.25 x [Na+]).

52

Figure 35. Factorization with PMF for source apportionment of atmospheric aerosols showing the variability in percentage of their total concentration in PM1. Error bars indicate the ranges of uncertainty range from the bootstrapping to 5th percentile as the lower and 95th percentile as the upper limit. Red boxes denote the range where 50% of the bootstrap values are. Clustered plus signs near to a box indicate highreproducibility.ssK and ssSO4= potassium and sulphate ions from sea salt, nssK and nssSO4= non-sea salt potassium and sulphate ions, BC= black carbon, MSA= methane sulphonate, Oxal = oxalic acid, Xly= xylose, Ara= arabinose, Rha= rhamnose, Fuc= fucose, GluMa= glucose/mannose, Gal= galactose.

The presence of marine gels in this fine-mode aerosol was identified as two types of gel factors,with fingerprints as shown in Figure 35b and c. These both factors showed remarkably higher contributions from THNS, suggested to reflect a marine origin, such as phytoplankton or bacteria. Relative to the gel Type 1 factor, the gel Type 2 factor occurred less frequently and mainly during autumn and winter when the Arctic sea is frozen. This Type 2 factor showed a higher relative level of xylose (Figure 35c), compared to the other factors which could, for instance,, be due to an origin of ice-algae, either from the pack ice or ice floes in the open leads (see Paper III).

a) Sea salt factor

c) Marine gel factor (type 2)

b) Marine gel factor (type 1)

53

Samples from September to December 2015 were further subjected to determination of proteinaceous material. Because of limited volumes of the samples, it was decided to hydrolyse them directly, without differentiation between FAAs and PAAs. Thus, the reported levels for these samples are for TAAs. The temporal concentrations are shown in Figure 36. One maximum was observed in early/mid-October (week 41-42), 5th – 19th, while lower levels were detected in November, 16th – 30th (week 46-48). The main source of these proteinogenic compounds detected in the autumn may be marine gel type 2, as discussed above. The air parcel reaching the Zeppelin station had previously been travelling for at least 3 of the last days at a low altitude, within the atmospheric boundary layer (<500 m a.s.l), thus being able to bring proteinaceous material into the atmosphere, from the sea and ice surfaces. In contrast, the sampled air during November had been mostly in the free troposphere (at ca 1500 m a.s.l), with limited contact with surface sources during the last 5 days.

Figure 36. The total concentrations of the sum of free and combined amino acids, TAAs, in aerosols sampled in Svalbard from September to December in 2015. The blue line shows the trend for particles of Dp< 2μm and the red line for the 2< Dp <10μm size fraction. Each data point corresponds to the average concentrations during two weeks. The insets show two-week backwards calculated trajectories for October 19th and November 30th corresponding to the maximum and minimum levels, respectively.

54

Our results show significantly higher levels of TAAs compared to reported levels of FAAs (about 10-fold) in previous studies of aerosols, both in the Arctic (Scalabrin et al., 2012) and in the Antarctic (Barbaro et al., 2015). This indicates that most of the amino acid content in our investigation was from PAAs, such as proteins and peptides. This is also in accordance with the results from the fog samples described in 3.2.1.

The synoptic-scale systems have ensured that heat, moisture, and aerosols have been transported from a variety of sources, and for a variable length of time, to the Zeppelin observatory before sampling. This should have affected the chemical and physical transformations of the sampled airborne compounds, also the proteinaceous matter, and hence the relative levels of each amino acid in the size-resolved fractions. Consequently, the determined amino acid fingerprint could ideally be useful as an indicator of the origin and age of the aerosols.

In Paper II, possible sources of individual amino acids in the aerosols were investigated. Backward trajectories (5 to 10-days and in three dimensions), were computed for the air parcels, both in the free troposphere and in the boundary layer, before their arrival at the Zeppelin sampling station, see Figure 37 a and b. The fine-mode aerosols are known to have a longer residence time in the atmosphere and therefore they reflect long-range transport, while coarse-mode aerosols usually spend less than one day in the atmosphere before their deposition. Thus, the coarse-mode particles should indicate mostly local sources, such as the open sea or MIZ close to Svalbard. The collected air samples that had travelled mostly in the free troposphere correspond to Figure 37 c and e, while Figure 37 d and f reflect air that had spent most time closer to the surface within the boundary layer. The amino acids were divided according to their bioavailability and reactivity into three groups. Group A includes alanine, asparagine, glutamic acid, and glycine, which due to their relatively low reactivity in the atmosphere likely survive long-range transport. Group B includes proline, valine and serine, all associated with terrestrial and marine sources, such as plants, pollen, bacteria, and phytoplankton. Group C comprises isoleucine, leucine, and threonine with possible sources as coastal and marine phytoplankton and bacteria.

While the geographic distributions of the two calculated trajectory clusters (free troposphere and boundary layer, Fig. 37a, and b, respectively) were largely complementary, the overlap of the MIZ and the open waters in the vicinity of Svalbard made the coarse-mode fingerprints of Fig. 37e and f look very similar. At the 50th percentile level about 55%, 15% and 30% of TAA mass came from Groups A, B, and C, respectively. Thus the TAA fingerprints of the coarse-mode

55

aerosols were suggested to represent a freshly (< 1 day atmospheric residence time) generated aerosol for which local marine sources are predominant.

For the fine-mode aerosols (see fingerprints in Fig. 37 c and d), the higher median total mass concentration of group A (dominated by glycine) in the free troposphere cluster (Figure 37 c) relative to the boundary layer (Figure 37 d) indicates long-distant aged aerosols. Those could originate from anthropogenic sources, such as residential heating. Group C, which is more abundant in the boundary layer cluster (Figure 37 d), is indicating local or medium distant aerosols of coastal origin having distinct marine phytoplankton /microbial contributions.

Figure 37. The median relative abundance of TAA in the fine- (EAD < 2 μm) and coarse-mode (2 < EAD <10μm) fractions within Group A, B, and C for the trajectory cluster free troposphere (FT) and boundary layer (BL), respectively.

56

3.3 Anhydrous monosaccharides in airborne PM10 sampled in Swedish suburban areas

The goal with the study in paper V was to develop and validate a reliable method for wood smoke markers, i.e. levoglucosan, mannosan, and galactosan, to be used in future source apportionment studies, such as for Arctic aerosols. Paper V utilized HILIC/ESI-MS/MS for the measurements of anhydrous monosaccharides, as for the THNS, but here the analytes were measured directly without any previous hydrolysis of the samples. Levoglucosan has previously been shown to be stable for long periods in the atmosphere and can survive long-range transports (Fraser and Lakshmanan, 2000). The HILIC method was validated by use of standard reference materials, SRM1649a and SRM1649b, which correspond to atmospheric PM10 collected in an urban area of Washington D.C., U.S. The detected levels were found to be very close those reported by other authors (Larsen et al., 2006; Kuo et al., 2008; Louchouarn et al., 2009). Acomparison of the detected levels can be found in Figure 38. The repeatability was very high, with RSDs below 5%, and LODs were comparable with the lowest reported. An advantage,besides low LODs, is the isocratic elution, which makes retention times more reproducible. In my experience, a mobile phase/stationary phase equilibrium between the runs in HILIC is difficult to achieve within a reasonable time.

Figure 38. Measured concentrations of levoglucosan (blue bars) and the sum of galactosan+mannosan (not

chromatographically separated, red bars) in SRM 1649a. (a) This study= HILIC/ESI-MS/MS, (b) other studies using GC/MS, n/a = not applicable, i.e. mannosan and galactosan were not measured. The figure is from Paper V.

The validated method was used to follow the seasonal variation of wood combustion contribution in PM10 sampled at three suburban sampling stations in Sweden in 2017. The results

0

50

100

150

200

This study(a) Larsen et al.,2006(b)

Kuo et al.,2008(b)

Louchouran etal., 2009(b)

This study(a) Louchouran etal., 2009(b)

μg/g

SRM 1649a SRM 1649b

n/a n/a

57

showed, as expected, the highest concentrations in the wintertime (Figure 39). These increased levels, as shown for January, are due to frequent local wood combustion occasions for heating purposes. Levoglucosan was the major anhydrous monosaccharide detected in all PM10 samples. The calculated ratio between levoglucosan and the sum of mannosan and galactosan was around 9 in the samples taken in Enskede wintertime, indicating that the firewood combusted had been mainly of hardwood types, such as oak and birch. Interestingly, it came later to our knowledge that an old rotten cut-down oak (a hardwood type) had been combusted during this month in this specific area, which most likely explains this result. On the other hand, the corresponding ratios for Delsbo and Ytterjärna sampling sites were between 4 and 6, which rather suggests mixed softwood/hardwood as firewood. Examples of softwood are pine and spruce.

Figure 39. Determined PM10 concentrations of levoglucosan (blue bars) and the sum of mannosan and galactosan (red bars) in Enskede (Stockholm), Delsbo (Hudiksvall) and Ytterjärna (Söderälje) in January and July 2017, respectively. N=3, RSD < 3%. The figure is from Paper V.

0

50

100

150

200

Enskede Delsbo Ytterjärna Enskede Delsbo Ytterjärna

January July

ng/m

3

58

4. Conclusions and future perspectives

This thesis supports the hypothesis that polysaccharides and proteinaceous materials act as CCN and are important for the cloud formation in the Arctic. A transfer from local sources, likely the sea surface microlayer from the Arctic Ocean and open leads and/or the pack ice, via aerosols to fog was shown in this study. This was reflected as similar time trends in concentrations in aerosols and fog during the sampling periods. In addition, a source apportionment study, based on statistical analysis of results obtained from chemical analyses of several different compound groups, indicated marine gels as some of the major sources. A higher level of proteins compared to free amino acids in the aerosols and fog was shown, which in turn reflects natural sources rather than anthropogenic.

To my knowledge, this is the first time that proteinaceous material, not only free amino acids, has been measured in aerosols and fog in the Arctic. Interestingly, the levels in fog were found higher in 2008 compared to 2001, correlating with less ice coverage and thus a higher biological activity. A less ice coverage may further increase the number of biomolecular CCN that could change the radiative balance, by the formation of more low-level clouds. This will be important to study in the coming years.

Novel methods, useful for further investigation of the bioorganic aerosol chemistry in the Arctic, were developed in this thesis. The analytical method for amino acid derivatives is fast and at the same time of high specificity, beneficial for identification of amino acids at trace levels in highly complex samples. The screening method for fatty acids utilizing reversed-phase UHPLC/ESI-TWIMS/ToFMS is also fast and can be applied to investigate their occurrence in aerosols. These compounds exhibit surface-active properties and could, therefore, be important for the formation of CCN. The high tendency of unsaturated fatty acids to oxidize might be a useful marker of aerosol aging. For an absolute quantification, the method should be evaluated against quality control samples and by use of isotope-labelled surrogate internal standards. Another method, HILIC/ESI-MS/MS, was validated for the measurement of wood burning as one of several possible anthropogenic sources of atmospheric aerosols, by utilizing anhydrous sugars as markers. In future Arctic expeditions, measurements of both fatty acids and anhydrous sugars could add to the aerosol source traceability.

59

For proteins, the methods for hydrolysis should be further explored. Experimental design should help to screen the critical parameters and optimal settings for more model proteins, including those of both hydrophilic and hydrophobic properties. For future investigations, intact proteins could also be determined, however requiring sufficient amounts of sample. For instance, the occurrence of ice-nucleating proteins would be of interest to study. For the future, it would also be desirable to analyze each individual sample comprehensively, using all methods developed in this thesis, to obtain a larger picture of the role of the different biomolecules. In addition, a larger number of air samples, collected within shorter time intervals, would make the interpretation of 3-D backward trajectories and source identification less complicated. Non-target analytical strategies could reveal new types of organic compounds, possibly also important as CCN for cloud formation in the Arctic. Furthermore, additional experiments on artificial bubble bursting in situ would be of interest, in order to investigate the kinetics of biomolecules in atmospheric reactions, like oxidation and photodegradation.

60

Acknowledgments

Actually, I would need to write a book to express my gratitude to the many skilled, kind and friendly people I have met during my studies at Stockholm University. However, below is a summary and I hope I haven´t forgotten someone.

First, I would like to thank both the Analytical Chemistry unit of ACES and the Department of Meteorology, both at Stockholm University, for financing my research and also for providing me a high level of education.

I would like to express my deepest gratitude and appreciation to my dear supervisors, Prof. Ulrika Nilsson, Prof. Caroline Leck, and Assoc. Prof. Leopold L Ilag, for giving me this great opportunity to work with you and also learn a lot from you. I am very happy to have had such amazing, caring and experienced supervisors. I will always think of you as part of my achievements in my professional carrier.

I am very thankful to you, Ulrika, for being such a nice and helpful supervisor. You are always my hero in the field of mass spectrometry and I became motivated when I took the MS course with you. Working as your assistant in the MS course during the past years was a great honor for me. I learned a lot from you about both science and humanity. Thank you for everything and I wish you all the best.

Caroline, thank you so much for your support during my Ph.D. studies. You kindly and patiently explained to me the meteorological events in order to make them understandable. Your efforts to understand the complex parameters involving the Arctic changes and your scientific contribution to this field has been inspiring to me.

Leopold, I appreciate your help and support during these years. I have always been impressed by your generosity. I learned a lot from you how to think scientifically and find a way to solve the problems. I have always been surprised by your general knowledge about cultures, history, and art. I have enjoyed a lot talking to you during the coffee breaks.

I also want to thank all of my teachers during the master program, for giving me knowledge and skills within the different fields of analytical chemistry: Ulrika Nilsson, Roger Westerholm, Conny Östman, Anders Colmsjö, Leopold L Ilag, Magnus Åberg, Ingrid Granelli, Mohamed

61

Abdel Rehim and Gunnar Thorsen. Also thanks to Carlo Crescenzi for nice kitchen conversation and many sweet souvenirs from Italy, and to Jan Holmbäck for kindly giving me some reference compounds.

I would like to express my deepest gratitude to Prof. Natalia Tretyakova for offering me a great postdoc opportunity in her lab and for your kindness.

Sincere thanks to our administrative staff Lena Elfver and Jonas Rutberg. You have been very helpful and nice and also very patient, considering numerous practical questions. I also want to welcome Emilia Eklund to the unit and I am sure that the instruments and also the MS lab will be in good order under your management.

I would like to thank our former colleagues Anne-Marie Hagelroth for training me in nice short chats in Swedish and Christoffer Bergvall for nice conversations during lunchtime.

I would like to offer my sincere thanks to Agneta Öhrström for your support in the Meteorology department and all nice conversations about sailing.

Thank you, Qiuju Gao, for answering questions and helping me to start up the project. Many thanks to Evelyne Hamacher-Barth for the interesting conversations in the lab while we were working, and for inviting me to a delicious homemade German food.

Huge thanks to Karin Englund for helping me to fill out the tricky Swedish forms.

I would like to thank both Conny Östman and Cynthia de Wit a lot for your support, helpful comments and revisions of my thesis.

Also, my deepest gratitude to all Ph.D. students at the unit for your friendship. Francesco Iadaresta, you are a real friend and I definitely will miss your comments, your delicious pasta and your sense of humor. Javier Zurita, thank you for our nice collaboration and for being such a nice friend. Pedro Sousa, thank you for your valuable help in statistics and opposition at my seminar, despite the last minute notice; Ahmed Ramzy for being such a great friend and thank you for motivating me to do some exercise. I will always remember the fun bike trip to the conference in Djurönäset; Hwanmi Lim, I am always learning something new from you and thank you for all your help. I will always remember the nice Korean dinner that you cooked for us; Hatem Elmongy, thank you for all nice conversations in the kitchen; Nadia Zguna, thank you

62

for sharing the Vantage instrument and your patience when it was busy; Alessandro Quaranta, thank you for letting me use the gel electrophoresis instrumentation in my project and also for nice conversations in the kitchen; Nicolo Riboni, thank you for the nice trip to Umeå. It was really fun. Wish you the best luck with your remaining Ph.D. studies. I would also like to acknowledge my former office mates, Rozanna Avagyan for all kind discussions and advice; Jonas Fyrestam for nice conversations and your sense of humor; Erik Alm for being very helpful; Aljona Saleh for being a fantastic friend and supervisor who gave me infinitive support.

I would also like to thank the former Ph.D. students at the department for their support when I started my studies; Trifa Ahmed for being a very helpful colleague in the MS course; Peter Olsson for scientific discussions; Ioannis Sadiktsis for your unlimited IT support; Liying Jiang for being such a nice colleague; Giovanna Luongo for your sense of humor; Erik Tengstrand for helping me with Matlab; Tran Thi Thuy for nice chats; Silvia Masala for being a very supportive head of the PhD council.

Not to forget our neighbor corridor people: Thanks to Professor Margareta Törnqvist for being very nice to me and for talking to me in Swedish; Lillemor Asplund for always saying hello in the mornings; Hitesh Motwani for being a great friend; Isabella Karlsson for always being helpful and for all nice conversations in the kitchen; Jenny Aasa for helping me with my thesis; Henrik Carlsson for nice conversation during lunchtime; Jana Weiss for our conversations during numerous weekends at the department; Johan Eriksson for all your technical advice and help with instruments; Ioannis Athanassiadis for technical helps in the MS lab; Johan Gustafsson for being a nice colleague in the MS course and you are really a perfect lab teacher.

I would like to show my gratitude to some fantastic diploma workers: Linnea Axelsson, Aram Ismail, Vasileios Tsioutsios and Silvia Spinicci for your enthusiasm, performance, and contribution to my projects. I wish you great future carriers.

I would like to thank also my former office mate in Meteorology: Erik Svensson for helping me with Matlab and for the scientific discussions.

Thanks to all my friends in the Meteorology department and also the crews of Oden Swedish Icebreaker.

63

My sincere thanks to my collaborators: Qiuju Gao, Javier Zurita, Matthias Karl, Are Bäcklund, Susana Lopez-Aparicio, Jost Heintzenberg, Sanna Silvergren, Silvia Spinicci and Roger Westerholm.

I would like to express my appreciation also to the engineers from different companies for their great technical skills regarding the MS instruments and from which I have had the possibility to learn a lot.

Huge thanks also to my fantastic Persian friends in ACESk, MMK, DDB, Meteorology and Sociology departments of Stockholm University and everywhere, for being nice, caring, supportive and kind.

Infinitive thanks to my mother and father for supporting and motivating me during these years. No words are sufficient to express my appreciation. I would also like to thank my brother, my sisters and their families for all your support and motivation.

Finally, thanks to Sweden and the Swedish people for giving me the opportunity to study here.

64

5. References Anisimov OA, Vaughan DG, Callaghan TV, Furgal C, Marchant H, Prowse TD, Vilhjálmsson H, Walsh JE (2007) Polar regions (Arctic and Antarctic). Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Parry ML, Canziani OF, Palutikof JP,van der Linden PJ, Hanson CE, Eds., Cambridge University Press, Cambridge, p 653.

Asakawa D, Furuichi Y, Yamamoto A, Oku Y, Funasaka K (2015) Rapid and sensitive quantification of levoglucosan in aerosols by high-performance anion-exchange chromatography with positive electrospray ionization mass spectrometry (HPAEC-positive ESI-MS). Atmos Environ 122:183

Avagyan R, Westerholm R (2017) Target and suspect screening of OH-PAHs in air particulates using liquid chromatography-orbitrap high resolution mass spectrometry. Talanta 165:702

Barbaro E, Zangrando R, Moret I, Barbante C, Cescon P, Gambaro A (2011) Free amino acids in atmospheric particulate matter of Venice, Italy. Atmos Environ 45:5050

Barbaro E, Zangrando R, Vecchiato M, Turetta C, Barbante C, Gambaro A (2014) D-and L-amino acids in Antarctic lakes: assessment of a very sensitive HPLC-MS method. Anal Bioanal Chem 406:5259

Barbaro E, Zangrando R, Vecchiato M, Piazza R, Cairns W, Capodaglio G, Barbante C, Gambaro A (2015) Free amino acids in Antarctic aerosol: potential markers for the evolution and fate of marine aerosol. Atmospheric Chem Phys15:5457

Bates DV (1992) Health indices of the adverse effects of air pollution: the question of coherence. Environ Res 59:336

Bigg EK, Leck C (2008) The composition of fragments of bubbles bursting at the ocean surface. J Geophys Res 113:D11209

Blanchard D, Woodcock A (1957) Bubble formation and modification in the sea and its meteorological significance. Tellus 9:145

65

Blanchard D (1975) Bubble Scavenging and the Water-to-Air Transfer of Organic Material in the Sea, in: Applied Chemistry at Protein Interfaces; Advances in Chemistry 145:360

Bond TC, Streets DG, Yarber KF, Nelson SM, Woo J, Klimont Z (2004) A technology based global inventory of black and organic carbon emissions from combustion. J Geophys Res 109: D14203

Borch NH, Kirchman DL (1997) Concentration and composition of dissolved combined neutral sugars (polysaccharides) in seawater determined by HPLC-PAD. Mar Chem 57:85

Burney CM, Sieburth JM (1977) Dissolved carbohydrates in seawater. II, A spectrophotometric procedure for total carbohydrate analysis and polysaccharide estimation. Mar Chem 5:15

Charlson R, Pilat M (1969) Climate: The influence of aerosols. J Appl Meteorol 8:1001

Cheng YS (2014) Mechanisms of pharmaceutical aerosol deposition in the respiratory tract. AAPS PharmSciTech 15:630

Cochran RE, Laskina O, Jayarathne T, Laskin A, Laskin J, Lin P, Sultana C, Lee C, Moore KA, Cappa CD (2016) Analysis of organic anionic surfactants in fine and coarse fractions of freshly emitted sea spray aerosol. Environ Sci Technol 50:2477

Cohen SA, Michaud DP (1993) Synthesis of a fluorescent derivatizing reagent, 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate, and its application for the analysis of hydrolysate amino acids via high-performance liquid chromatography. Anal Biochem 211:279

Dole M, Mack L, Hines R, Mobley R, Ferguson L, Alice Md (1968) Molecular beams of macroions. J Chem Phys 49:2240

Dołowy M, Pyka A (2015) Chromatographic methods in the separation of long-chain mono-and polyunsaturated fatty acids. Journal of Chemistry 2015:1

Donaldson K, Stone V, Seaton A, MacNee W (2001) Ambient particle inhalation and the cardiovascular system: potential mechanisms. Environ Health Perspect 109 Suppl 4:523

66

Dye C, Yttri KE (2005) Determination of monosaccharide anhydrides in atmospheric aerosols by use of high-performance liquid chromatography combined with high-resolution mass spectrometry. Anal Chem 77:1853

Eder K (1995) Gas chromatographic analysis of fatty acid methyl esters. J Chromatogr B Biomed Appl 671:113

Engling G, Carrico CM, Kreidenweis SM, Collett Jr JL, Day DE, Malm WC, Lincoln E, Hao WM, Iinuma Y, Herrmann H (2006) Determination of levoglucosan in biomass combustion aerosol by high-performance anion-exchange chromatography with pulsed amperometric detection. Atmos Environ 40:299

Ettre LS (1993) Nomenclature for chromatography (IUPAC Recommendations 1993). Pure &Appl, Chem 65:819

Facchini MC, Rinaldi M, Decesari S, Carbone C, Finessi E, Mircea M, Fuzzi S, Ceburnis D, Flanagan R, Nilsson ED (2008) Primary submicron marine aerosol dominated by insoluble organic colloids and aggregates. Geophys Res Lett 35:L17814

Fraser MP, Lakshmanan K (2000) Using levoglucosan as a molecular marker for the long-range transport of biomass combustion aerosols. Environ Sci Technol 34:4560

Fu P, Kawamura K, Miura K (2011) Molecular characterization of marine organic aerosols collected during a round the world cruise. J Geophys Res 116:D13302

Fuchs B, Süß R, Teuber K, Eibisch M, Schiller J (2011) Lipid analysis by thin-layer chromatography a review of the current state. J Chromatogr A 1218:2754

Fuchs B (2012) Analysis of phospolipids and glycolipids by thin-layer chromatography–matrix-assisted laser desorption and ionization mass spectrometry. J Chromatogr A 1259:62

Fuzzi S, Mandrioli P, Perfetto A (1997) Fog droplets an atmospheric source of secondary biological aerosol particles. Atmos Environ 31:287

67

Gambaro A, Zangrando R, Gabrielli P, Barbante C, Cescon P (2008) Direct determination of levoglucosan at the picogram per milliliter level in Antarctic ice by high-performance liquid chromatography/electrospray ionization triple quadrupole mass spectrometry. Anal Chem 80:1649

Gao Q, Araia M, Leck C, Emmer Å (2010) Characterization of exopolysaccharides in marine colloids by capillary electrophoresis with indirect UV detection. Anal Chim Acta 662:193

Gao Q, Nilsson U, Ilag LL, Leck C (2011) Monosaccharide compositional analysis of marine polysaccharides by hydrophilic interaction liquid chromatography-tandem mass spectrometry. Anal Bioanal Chem 399:2517

Gao Q, Leck C, Rauschenberg C, Matrai PA (2012) On the chemical dynamics of extracellular polysaccharides in the high Arctic surface microlayer. Ocean Sci 8:401

Gershey RM (1983) Characterization of seawater organic matter carried by bubble generated aerosols. Limnol Oceanogr 28:309

Giles K, Williams JP, Campuzano I (2011) Enhancements in travelling wave ion mobility resolution. Rapid Commun Mass Spectrom 25:1559

Gorzelska K, Galloway JN (1990) Amine nitrogen in the atmospheric environment over the North Atlantic Ocean. Global Biogeochem Cycles 4:309

Gosetti F, Mazzucco E, Gennaro MC, Marengo E (2016) Contaminants in water: non-target UHPLC/MS analysis. Environ Chem Lett 14:5177

Hawkins LN, Russell LM (2010) Polysaccharides, proteins, and phytoplankton fragments: Four chemically distinct types of marine primary organic aerosol classified by single particle spectromicroscopy. Adv Meteorol 2010: Article ID 612132.

Hedberg E, Kristensson A, Ohlsson M, Johansson C, Johansson P, Swietlicki E, Vesely V, Wideqvist U, Westerholm R (2002) Chemical and physical characterization of emissions from birch wood combustion in a wood stove. Atmos Environ 36:4823

68

Hedberg E, Johansson C, Johansson L, Swietlicki E, Brorström-Lundén E (2006) Is levoglucosan a suitable quantitative tracer for wood burning? Comparison with receptor modeling on trace elements in Lycksele, Sweden. J Air Waste Manage Assoc 56:1669

Heidam NZ (1981) Aerosol fractionation by sequential filtration with Nuclepore filters. Atmos Environ (1967) 15:891

Heintzenberg J (1989) Arctic Haze: Air pollution in polar regions. Ambio 18:50

Hemström P, Irgum K (2006) Hydrophilic interaction chromatography. J Sep Sci 29:1784

Hennigan CJ, Sullivan AP, Collett JL, Robinson AL (2010) Levoglucosan stability in biomass burning particles exposed to hydroxyl radicals. Geophys Res Lett 37:L09806

Hill DW, Walters FH, Wilson TD, Stuart JD (1979) High performance liquid chromatographic determination of amino acids in the picomole range. Anal Chem 51:1338

Hirs CH, Stein WH, Moore S (1954) The amino acid composition of ribonuclease. J Biol Chem 211:941

Hoagland KD, Rosowski JR, Gretz MR, Roemer SC (1993) Diatom extracellular polymeric substances: function, fine structure, chemistry, and physiology. J Phycol 29:537

Huang Y, Dodds ED (2013) Ion mobility studies of carbohydrates as group I adducts: isomer specific collisional cross section dependence on metal ion radius. Anal Chem 85:9728

Iribarne J, Thomson B (1976) On the evaporation of small ions from charged droplets. J Chem Phys 64:2287

Jeffries MO, Richter-Menge J (2015) The Arctic [in “State of the Climate in 2014”]. Bull Amer Meteor Soc 96:S127

John W, Hering S, Reischl G, Sasaki G, Goren S (1983) Characteristics of Nuclepore filters with

69

large pore size II. Filtration properties. Atmos Environ 17:373

Karvosenoja N., Johansson M., Kindbom K., Lükewille A., Jensen D., Sternhufvud C. and Illerup J.B. 2004. Fine particulate matter emissions from residential wood combustion and reduction potential in the Nordic countries. Proceedings of the 13th World Clean Air and Environmental Protection Congress and Exhibition, London 22-27 August 2004.

Kaufman DS, Manley WF (1998) A new procedure for determining DL amino acid ratios in fossils using reverse phase liquid chromatography. Quat Sci Rev17:987

Kenne L, Lindberg B (1983) Bacterial polysaccharides. In The Polysaccharides; Aspinall, G.O., Ed.; Academic Press: New York, NY, USA; 2:287

Kieber RJ, Hydro LH, Seaton PJ (1997) Photooxidation of triglycerides and fatty acids in seawater: Implication toward the formation of marine humic substances. Limnol Oceanogr 42:1454

Kohler M, Leary JA (1995) LC/MS/MS of carbohydrates with postcolumn addition of metal chlorides using triaxial electrospray probe. Anal Chem 67:3501

Kristensson A, Rosenørn T, Bilde M (2009) Cloud droplet activation of amino acid aerosol particles. J Phys Chem A 114:379

Kuo L, Herbert BE, Louchouarn P (2008) Can levoglucosan be used to characterize and quantify char/charcoal black carbon in environmental media? Org Geochem 39:1466

Kuznetsova M, Lee C, Aller J (2005) Characterization of the proteinaceous matter in marine aerosols. Mar Chem 96:359

Larsen III RK, Schantz MM, Wise SA (2006) Determination of levoglucosan in particulate matter reference materials. Aerosol Sci Technol 40:781

Lanser A (1993) Conversion of oleic acid to 10-ketostearic acid by a Staphylococcus species. J Am Oil Chem Soc 70:543

70

Lanucara F, Holman SW, Gray CJ, Eyers CE (2014) The power of ion mobility-mass spectrometry for structural characterization and the study of conformational dynamics. Nat Chem 6:281

Leck C, Bigg EK (1999) Aerosol production over remote marine areas A new route. Geophys Res Lett 26:3577

Leck C, Nilsson ED, Bigg EK, Bäcklin L (2001) Atmospheric program on the Arctic Ocean Expedition 1996 (AOE 96): An overview of scientific goals, experimental approach, and instruments. J Geophys Res 106:32051.

Leck C, Norman M, Bigg EK, Hillamo R (2002) Chemical composition and sources of the high Arctic aerosol relevant for cloud formation. J Geophys Res 107:D124135

Leck C, Bigg EK (2005) Biogenic particles in the surface microlayer and overlaying atmosphere in the central Arctic Ocean during summer. Tellus B 57:305

Leck C, Bigg EK (2010) New particle formation of marine biological origin. Aerosol Sci Technol 44:570

Leck C, Svensson E (2015) Importance of aerosol composition and mixing state for cloud droplet activation over the Arctic pack ice in summer. Atmosperic Chem Phys 15:2545

Lee C (1988) Amino acid and amine biogeochemistry in marine particulate material andsediments, p. 125–141. In T. H. Black- burn and J. Sørensen [eds.], Nitrogen cycling in coastal marine environments. John Wiley & Sons.

Lohmann U, Feichter J (2005) Global indirect aerosol effects: a review. Atmospheric Chem Phys 5:715

Lohmann U, Leck C (2005) Importance of submicron surface active organic aerosols for pristine Arctic clouds. Tellus B 57:261

Louchouarn P, Kuo L, Wade TL, Schantz M (2009) Determination of levoglucosan and its isomers in size fractions of aerosol standard reference materials. Atmos Environ 43:5630

71

Mace KA, Duce RA, Tindale NW (2003) Organic nitrogen in rain and aerosol at Cape Grim, Tasmania, Australia. J Geophys Res 108:D114338

Mader B, Yu J, Xu J, Li Q, Wu W, Flagan R, Seinfeld J (2004) Molecular composition of the water soluble fraction of atmospheric carbonaceous aerosols collected during ACE Asia. JGeophys Res 109:D06206.

Mandalakis M, Apostolaki M, Stephanou EG (2010) Trace analysis of free and combined amino acids in atmospheric aerosols by gas chromatography–mass spectrometry. J Chromatogr A1217:143

Mandalakis M, Apostolaki M, Tziaras T, Polymenakou P, Stephanou EG (2011) Free and combined amino acids in marine background atmospheric aerosols over the Eastern Mediterranean. Atmos Environ 45:1003

Martin M, Chang R, Sierau B, Sjogren S, Swietlicki E, Abbatt J, Leck C, Lohmann U (2011) Cloud condensation nuclei closure study on summer arctic aerosol. Atmospheric Chem Phys 11:11335

Marty J, Saliot A, Buat Ménard P, Chesselet R, Hunter K (1979) Relationship between the lipid compositions of marine aerosols, the sea surface microlayer, and subsurface water. J Geophys Res 84:5707

Mastrangelo G, Fadda E, Marzia V (1996) Polycyclic aromatic hydrocarbons and cancer in man. Environ Health Perspect 104:1166

Matos JT, Duarte RM, Duarte AC (2016) Challenges in the identification and characterization of free amino acids and proteinaceous compounds in atmospheric aerosols: A critical review. TrAC 75:97

McCormick RA, Ludwig JH (1967) Climate modification by atmospheric aerosols. Science 156:1358

McGregor KG, Anastasio C (2001) Chemistry of fog waters in California's Central Valley: 2. Photochemical transformations of amino acids and alkyl amines. Atmos Environ 35:1091

72

Milne PJ, Zika RG (1993) Amino acid nitrogen in atmospheric aerosols: Occurrence, sources and photochemical modification. J Atmos Chem 16:361

Mochida M, Kitamori Y, Kawamura K, Nojiri Y, Suzuki K (2002) Fatty acids in the marine atmosphere: Factors governing their concentrations and evaluation of organic films on sea salt particles. J Geophys Res 107:D174325

Mopper K, Zika RG (1987) Free amino acids in marine rains: evidence for oxidation and potential role in nitrogen cycling. Nature 325:246

Myhre G, Shindell D, Bréon F.M, Collins W, Fuglestvedt J, Huang J, Koch D, Lamarque J.F, Lee D, Mendoza B, Nakajima T, Robock A, Stephens G, Takemura T, Zhang H (2013) Anthropogenic and Natural Radiative Forcing. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Neeser J, Schweizer TF (1984) A quantitative determination by capillary gas-liquid chromatography of neutral and amino sugars (as O-methyloxime acetates), and a study on hydrolytic conditions for glycoproteins and polysaccharides in order to increase sugar recoveries. Anal Biochem 142:58

Nováková L, Matysová L, Solich P (2006) Advantages of application of UPLC in pharmaceutical analysis. Talanta 68:908

Nyström R, Lindgren R, Avagyan R, Westerholm R, Lundstedt S, Boman C (2017) Influence of wood species and burning conditions on particle emission characteristics in a residential wood stove. Energy Fuels 31:5514

Ohshima T, Yoon H, Koizumi C (1989) Application of selective ion monitoring to the analysis of molecular species of vegetable oil triacylglycerols separated by open-tubular column GLC on a methylphenylsilicone phase at high temperature. Lipids 24:535

Orellana MV, Matrai PA, Leck C, Rauschenberg CD, Lee AM, Coz E (2011) Marine microgels as a source of cloud condensation nuclei in the high Arctic. Proc Natl Acad Sci USA 108:13612

73

Pakulski JD, Benner R (1992) An improved method for the hydrolysis and MBTH analysis of dissolved and particulate carbohydrates in seawater. Mar Chem 40:143

Pakulski JD, Benner R (1992) An improved method for the hydrolysis and MBTH analysis of dissolved and particulate carbohydrates in seawater. Mar Chem 40:143

Pope III CA, Burnett RT, Thun MJ, Calle EE, Krewski D, Ito K, Thurston GD (2002) Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA 287:1132

Pöschl U (2005) Atmospheric aerosols: composition, transformation, climate and health effects. Angew Chem Int Ed 44:7520

Quinn PK, Collins DB, Grassian VH, Prather KA, Bates TS (2015) Chemistry and related properties of freshly emitted sea spray aerosol. Chem Rev 115:4383

Rao MS, Hidajat K, Ching C (1995) Reversed-phase HPLC: the separation method for the characterization and purification of long chain polyunsaturated fatty acids a review. J Chromatogr Sci 33:9

Ren D, Julka S, Inerowicz HD, Regnier FE (2004) Enrichment of cysteine-containing peptides from tryptic digests using a quaternary amine tag. Anal Chem 76:4522

Rogatsky E, Jayatillake H, Goswami G, Tomuta V, Stein D (2005) Sensitive LC MS quantitative analysis of carbohydrates by Cs attachment. J Am Soc Mass Spectrom 16:1805

Russell LM, Hawkins LN, Frossard AA, Quinn PK, Bates TS (2010) Carbohydrate-like composition of submicron atmospheric particles and their production from ocean bubble bursting. Proc Natl Acad Sci USA 107:6652

Saarnio K, Teinilä K, Aurela M, Timonen H, Hillamo R (2010) High-performance anion-exchange chromatography–mass spectrometry method for determination of levoglucosan, mannosan, and galactosan in atmospheric fine particulate matter. Anal Bioanal Chem 398:2253

Scalabrin E, Zangrando R, Barbaro E, Kehrwald N, Gabrieli J, Barbante C, Gambaro A (2012) Amino acids in Arctic aerosols. Atmospheric Chem Phys 12:10453

74

Scheller E (2001) Amino acids in dew–origin and seasonal variation. Atmos Environ 35:2179

Schkolnik G, Rudich Y (2006) Detection and quantification of levoglucosan in atmospheric aerosols: A review. Anal Bioanal Chem 385:26

Schmeltz I, Hoffmann D (1977) Nitrogen-containing compounds in tobacco and tobacco smoke. Chem Rev 77:295

Leach J, Blanch A, Bianchi A (1999) Volatile organic compounds in an urban airborne environment adjacent to a municipal incinerator, waste collection centre and sewage treatment plant. Atmos Environ 33:4309

Simoneit BR (1989) Organic matter of the troposphere V: application of molecular marker analysis to biogenic emissions into the troposphere for source reconciliations. J Atmos Chem 8:251

Simoneit BR (2002) Biomass burning a review of organic tracers for smoke from incomplete combustion. Appl. Geochem. 17:129

Skoog A, Benner R (1997) Aldoses in various size fractions of marine organic matter: Implications for carbon cycling. Limnol Oceanogr 42:1803

Soga T, Heiger DN (2000) Amino acid analysis by capillary electrophoresis electrospray ionization mass spectrometry. Anal Chem 72:1236

Song T, Wang S, Zhang Y, Song J, Liu F, Fu P, Shiraiwa M, Xie Z, Yue D, Zhong L (2017) Proteins and Amino Acids in Fine Particulate Matter in Rural Guangzhou, Southern China: Seasonal Cycles, Sources and Atmospheric Processes. Environ Sci Technol 51:6773

Sǿreides J.E., Leu E, Berge J, Graeve M, Falk Petersen S (2010) Timing of blooms, algal food quality and Calanus glacialis reproduction and growth in a changing Arctic. Glob Change Biol 16:3154

75

Stohl A (2006) Characteristics of atmospheric transport into the Arctic troposphere. J GeophysRes 111:D11306

Tapuhi Y, Schmidt DE, Lindner W, Karger BL (1981) Dansylation of amino acids for high-performance liquid chromatography analysis. Anal Biochem 115:123

Tjernström M, Leck C, Persson POG, Jensen ML, Oncley SP, Targino A (2004) The summertime Arctic atmosphere: meteorological measurements during the Arctic Ocean Experiment 2001. Bull. Am. Meteorol. Soc. 85:1305

Tjernström M, Leck C, Birch CE, Bottenheim JW, Brooks BJ, Brooks IM, Bäcklin L, Chang R, de Leeuw G, Di Liberto L (2014) The Arctic Summer Cloud Ocean Study (ASCOS): overview and experimental design. Atmospheric Chem Phys 14:2823

Tseng R, Viechnicki JT, Skop RA, Brown JW (1992) Sea to air transfer of surface active organic compounds by bursting bubbles. J Geophys Res 97:5201

Tsugita A, Scheffeler J (1982) A rapid method for acid hydrolysis of protein with a mixture of trifluoroacetic acid and hydrochloric acid. The FEBS Journal 124(3):585

Twomey S (1977) The influence of pollution on the shortwave albedo of clouds. J Atmos Sci 34:1149

Verdugo P (2012) Marine microgels. Annu. Rev. Mar. Sci 4:375

Wang X, Sultana CM, Trueblood J, Hill TC, Malfatti F, Lee C, Laskina O, Moore KA, Beall CM, McCluskey CS (2015) Microbial control of sea spray aerosol composition: a tale of two blooms. ACS Cent Sci 1:124

Wedyan MA, Preston MR (2008) The coupling of surface seawater organic nitrogen and the marine aerosol as inferred from enantiomer-specific amino acid analysis. Atmos Environ 42:8698

Wei-Chun C, Orellana MV, Verdugo P (1998) Spontaneous assembly of marine dissolved organic matter into polymer gels. Nature 391:568

76

Yang W, Mirzaei H, Liu X, Regnier FE (2006) Enhancement of amino acid detection and quantification by electrospray ionization mass spectrometry. Anal Chem 78:4702

You C, Song L, Xu B, Gao S (2016) Method for determination of levoglucosan in snow and ice at trace concentration levels using ultra-performance liquid chromatography coupled with triple quadrupole mass spectrometry. Talanta 148:534

Young RA (1984) The chemistry of solid wood. In: Pettersen RC, The chemical composition of wood. ACS Publications, Washington 57:126. DOI: 10.1021/ba-1984-0207.ch002

Yu J Z (2002) Chemical characterization of water soluble organic compounds in particulate matters in Hong Kong. A Final Report Submitted for the Provision of Service to the Environmental Protection Department, HKSAR

Zangrando R, Piazza R, Cairns WR, Izzo FC, Vianello A, Zendri E, Gambaro A (2010) Quantitative determination of un-derivatised amino acids in artistic mural paintings using high-performance liquid chromatography/electrospray ionization triple quadrupole mass spectrometry. Anal Chim Acta 675:1

Zhang Q, Anastasio C, Jimenez Cruz M (2002) Water soluble organic nitrogen in atmospheric fine particles (PM2. 5) from northern California. J Geophys Res 107:D114112:AAC 3

Zhang Q, Anastasio C (2003) Free and combined amino compounds in atmospheric fine particles (PM 2.5) and fog waters from Northern California. Atmos Environ 37:2247

Zhou J, Swietlicki E, Berg OH, Aalto PP, Hämeri K, Nilsson ED, Leck C (2001) Hygroscopic properties of aerosol particles over the central Arctic Ocean during summer. J Geophys Res 106:32111

Zhu J, Cole RB (2001) Ranking of gas-phase acidities and chloride affinities of monosaccharides and linkage specificity in collision-induced decompositions of negative ion electrospray-generated chloride adducts of oligosaccharides. J Am Soc Mass Spectrom 12:1193

Documents

su.diva-portal.orgsu.diva-portal.org/smash/get/diva2:1200210/FULLTEXT02.pdfimprovement in reversed-phase chromatographic separation, ionization efficiency as well as selectivity. For