Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
FINNISH METEOROLOGICAL INSTITUTE
CONTRIBUTIONS
NO. 74
CHEMICAL MASS CLOSURE AND SOURCE-SPECIFIC COMPOSITION OF
ATMOSPHERIC PARTICLES
Sanna Saarikoski
ACADEMIC DISSERTATION
To be presented, with the permission of the Faculty of Science of the University of Helsinki, for
public criticism in Auditorium A129 of the Department of Chemistry on November 14th, 2008, at
12 o’clock noon
Finnish Meteorological Institute
Helsinki 2008
2
ISBN 978-951-697-679-5 (paperback)
ISSN 0782-6117
Yliopistopaino
Helsinki 2008
ISBN 978-951-697-680-1 (pdf)
http://ethesis.helsinki.fi
Helsinki 2008
3
Series title, number and report code of publication
Published by Finnish Meteorological Institute Finnish Meteorological Institute, (Erik Palménin aukio 1) , P.O. Box 503 Contributions No. 74, FMI-CONT-74
FIN-00101 Helsinki, Finland Date: October 2008 Authors Sanna Saarikoski Title Chemical mass closure and source-specific composition of atmospheric particles Abstract Atmospheric aerosols have significant impacts on human health and climate. It has been known for decades that particles can cause adverse health effects as they are deposited within the respiratory system. Atmospheric aerosol particles influence climate by scattering solar radiation but aerosol particles act also as the nuclei around which cloud droplets form. Both the direct scattering and the indirect effect of aerosols via cloud formation and changing cloud properties cool the climate. However, some aerosol particles such as black carbon (soot) can also warm the climate by absorbing solar radiation or by changing Earth’s surface albedo after deposition on snow. The principal objectives of this thesis were to investigate the chemical composition and the sources of fine particles in different environments (traffic, urban background, remote) as well as during some specific air pollution situations. Quantifying the climate and health effects of atmospheric aerosols is not possible without detailed information of the aerosol chemical composition. Aerosol measurements were carried out at nine sites in six different countries (Finland, Germany, Czech, Netherlands, Greece and Italy). Several different instruments were used in order to measure both the particulate matter (PM) mass and its chemical composition. In the off-line measurements the samples were collected first on a substrate or filter and gravimetric and chemical analysis were conducted in the laboratory. In the on-line measurements the sampling and analysis were either a combined procedure or performed successively within the same instrument. Results from the impactor samples were analyzed by the statistical methods. In addition to atmospheric aerosol studies, this thesis comprises a work where a method for the determination carbonaceous matter size distribution by using a multistage impactor was developed. It was found that the chemistry of PM has usually strong spatial, temporal and size-dependent variability. In the Finnish urban and background sites most of the fine PM consisted of organic matter. However, at a remote site in Crete sulfate dominated the fine PM and in a rural background site in Italy nitrate made the largest contribution to the fine PM. Regarding the size-dependent chemical composition, organic components were likely to be enriched in smaller particles than inorganic ions. Biomass burning and desert dust episodes were found to alter the chemistry of particles significantly. Data analysis showed that organic carbon (OC) had four major sources in Helsinki. Secondary production was the major source in Helsinki during spring, summer and fall, whereas in winter biomass combustion dominated OC. The significant impact of biomass combustion on OC concentrations was also observed in the measurements performed in Central Europe. In this thesis aerosol samples were collected mainly by the conventional filter and impactor methods which suffered from the long integration time. However, by filter and impactor measurements chemical mass closure was achieved accurately, and a simple filter sampling was found to be useful in order to explain the sources of PM on the seasonal basis. The online instruments gave additional information related to the temporal variations of the sources and the atmospheric mixing conditions. Publishing unit Finnish Meteorogical Institute Classification (UDK) Keywords 504.05 504.054 504.064.02 Atmospheric aerosol particles, chemical composition,
chemical mass closure, biomass combustion ISSN and series title 0782-6117 Finnish Meteoorological Institute Contributions ISBN Language 978-951-697-679-5 English Sold by Pages 182 Price
Finnish Meteorological Institute / Library P.O.Box 503, FIN-00101 Helsinki Note Finland
4
Julkaisun sarja, numero ja raporttikoodi Finnish Meteorological Institute, Julkaisija Ilmatieteen laitos, ( Erik Palménin aukio 1) Contributions No. 74, FMI-CONT-74
PL 503, 00101 Helsinki Julkaisuaika: Lokakuu 2008 Tekijä(t) Sanna Saarikoski Nimeke Ilmakehän hiukkasten kemiallinen massasulkeuma ja lähdekohtainen koostumus Tiivistelmä Ilmakehän hiukkaset vaikuttavat ihmisten terveyteen kulkeuduttuaan hengitysteihin. Hiukkaset vaikuttavat myös ilmastoomme sirottamalla auringosta tulevaa säteilyä joko suoraan tai muodostamalla pilviä. Kummatkin mekanismit viilentävät ilmastoa, mutta tietyt hiukkaset, esimerkiksi nokihiukkaset, voivat myös lämmittää ilmastoa absorboimalla auringon säteilyä tai muuttamalla maanpinnasta heijastuvan säteilyn määrää. Tämän väitöskirjan tavoitteena oli tarkastella ilmakehän hiukkasten kemiallista koostumusta ja lähteitä erilaisissa ympäristöissä, sillä hiukkasten terveys- ja ilmastovaikutusten arviointiin tarvitaan tietoa hiukkasten kemiasta. Hiukkasmittauksia tehtiin yhdeksällä eri mittausasemalla kuudessa maassa (Suomessa, Saksassa, Tšekissä, Alankomaissa, Kreikassa ja Italiassa). Mittausasemista suurin osa sijaitsi kaupungissa, mutta mittauksia tehtiin myös tausta-asemilla (Suomessa, Kreikassa ja Italiassa). Lisäksi hiukkasten koostumusta tutkittiin kaukokulkeutumaepisodien aikana sekä uusien hiukkasten syntyessä metsäympäristössä. Hiukkasnäytteet kerättiin impaktoreilla suodattimille tai muille keräysalustoille. Näytteistä määritettiin hiukkasten massa sekä niiden kemiallinen koostumus laboratoriossa. Impaktorikeräysten lisäksi hiukkasten massaa ja kemiallista koostumusta mitattiin käyttämällä online-laitteita, joissa hiukkasten keräys ja analyysi suoritetaan joko samanaikaisesti tai peräkkäin samassa laitteistossa. Lisäksi tässä väitöskirjassa kehitettiin uusi menetelmä hiilihiukkasten kokojakauman analysoimiseksi moniasteimpaktoria käyttäen. Hiukkasten kemiallinen koostumus vaihteli eri mittauspaikkojen välillä, mutta koostumukseen vaikutti myös vuorokaudenaika sekä tutkittu hiukkaskoko. Suomessa olevilla mittausasemilla, Helsingissä ja Hyytiälässä, suurin osa pienhiukkasista koostui orgaanisista yhdisteistä. Tämän tutkimusen mittauskohteessa Kreikassa, suurin osa pienhiukkasista oli sulfaattia, kun taas Italiassa pienhiukkaset koostuivat suurimmaksi osaksi nitraatista. Hiukkasten kemiallisen koostumuksen kokoriippuvuutta tutkittiin lähinnä Hyytiälässä ja Italiassa. Vaikka mittauspaikat ja hiukkasten kemiallinen koostumus yleensä ottaen olivat hyvin erilaisia Hyytiälässä ja Italiassa, kummassakin paikassa orgaaniset yhdisteet olivat pienemmissä hiukkasissa kuin epäorgaaniset ionit. Merkittävästi hiukkasten koostumukseen vaikutti myös kaukokulkeuma. Suomessa hiukkasten koostumus muuttui selvästi, kun Helsinkiin kulkeutui savua Venäjän metsäpaloista. Kreikassa hiukkasten koostumus taas muuttui, kun ilmavirtaukset toivat mukanaan mineraalihiukkasia Saharan hiekkamyrskyistä. Hiukkasten lähteitä selvitettiin analysoimalla näytteiden tuloksia tilastollisilla menetelmillä. Helsingissä tutkittiin erityisesti hiukkasmaisen orgaanisen hiilen lähteitä. Orgaaniselle hiilelle löydettiin neljä lähdettä, joista hiukkasten muodostuminen ilmakehässä oli vallitseva talvikuukausia lukuun ottamatta Talvella biomassan poltto tuotti eniten orgaanista hiiltä. Biomassan polton merkittävä osuus orgaanisen hiilen lähteenä havaittiin Helsingin lisäksi myös Keski-Euroopassa. Suurin osa tämän väitöskirjan hiukkasmitttauksista tehtiin keräämällä näytettä suodattamalla tai impaktoreilla, jolloin keräysaika on tyypillisesti vähintään vuorokauden pituinen. Huolimatta pitkästä keräysajasta näytteet sopivat hyvin kemiallisen massasulkeuman määritykseen sekä hiukkasten lähteiden vuodenaikavaihteluiden tarkasteluun. Online-laitteet taas sopivat nopeasti muuttuvien tilanteiden tarkastelun. Julkaisijayksikkö Ilmatieteen laitos, Ilmanlaatu Luokitus (UDK) Asiasanat 504.05 504.054 504.064.02 llmakehän aerosolihiukkaset, kemiallinen koostumus,
kemiallinen massasulkeuma, biomassan poltto ISSN ja avainnimike 0782-6117 Finnish Meteoorological Institute Contributions ISBN Kieli 978-951-697-679-5 Englanti Myynti Sivumäärä 182 Hinta Ilmatieteen laitos / Kirjasto PL 503, 00101 Helsinki Lisätietoja
5
ACKNOWLEDGEMENTS This study was carried out at the Air Quality Department of the Finnish Meteorological Institute during the years 2002–2008. Funding for the work was provided by the Academy of Finland, the Maj and Tor Nessling Foundation, the Finnish Funding Agency for Technology and Innovation (TEKES) and the European Commission 5th Framework Programme (PAMCHAR-project). I want to thank the former and current Heads of the Department, Professors Yrjö Viisanen and Jaakko Kukkonen for the opportunity to work in the Air Quality Department. I am most grateful to my supervisor, Professor Risto Hillamo, for his guidance and encouragement during the course of my thesis. Professor Marja-Liisa Riekkola is acknowledged for support to my post-graduate studies in the Laboratory of Analytical Chemistry at the University of Helsinki. The official reviewers of the thesis, Doc. Jyrki Mäkelä from the Tampere University of Technology and PhD Fabrizia Cavalli from the Joint Research Center in Ispra, are thanked for reviewing and commenting this thesis. I highly appreciate that Associate Professor Marianne Glasius from the University of Aarhus has kindly promised to be my official opponent in the public examination of this thesis. I wish to express my gratitude to Professor Veli-Matti Kerminen and PhD Markus Sillanpää for their valuable comments on my papers and this thesis manuscript. I also want to thank Professor Douglas Worsnop for giving me an opportunity to enter the wonderful world of Aerosol Mass Spectrometry. I am grateful to all my co-authors from the Finnish Meteorological Institute, University of Helsinki, National Public Health Institute, and University of Crete for the excellent co-operation during the work. Especially I wish to thank all my dear colleagues and ex-colleagues at FMI, Karri, Minna, Hilkka, Anna, and many more, for creating a pleasant and inspiring atmosphere for work. I am indebted to Timo Mäkelä for his assistance with the aerosol instruments. Finally, I warmly thank my parents for their love and support throughout my life. I am also grateful to my sister and her family and to all my friends for showing me that there is life outside the work. Helsinki, October 2008 Sanna Saarikoski
6
ABBREVIATIONS
AMS Aerosol mass spectrometer
APS Aerodynamic particle sizer
BC Black carbon
BLPI Berner low pressure impactor
CMB Chemical mass balance
CPC Condensation particle counter
CPI Carbon preference index
Da Particle aerodynamic diameter
D50 50% aerodynamic cut-off diameter
DL Detection limit
DMA Differential mobility analyzer
DMPS Differential mobility particle sizer
DOS Dioctyl sebacate
EC Elemental carbon
EGA Evolved gas analysis
ELPI Electrical low pressure impactor
EU European union
FDMS Filter dynamics measurement system
HOA Hydrocarbon-like organic aerosol
HPLC-MS High performance liquid chromatograph-mass spectrometer
HULIS Humic-like substances
IC Ion chromatograph/chromatography
ICP-MS Inductively coupled plasma-mass spectrometer
IPCC Intergovernmental panel on climate change
LC Liquid chromatography
LRT Long range transport
MA Monosaccharide anhydrides
MS Mass spectrum/spectrometer
m/z mass-to-charge ratio
7
NDIR Non-dispersive infrared
NPOC Non-purgeable organic carbon
nss Non-sea salt
OC Organic carbon
OOA oxygenated organic aerosol
PAH Polycyclic aromatic hydrocarbon
PILS Particle-into-liquid sampler
PMF Positive matrix factorization
PM Particulate matter
PMx Particle matter with an aerodynamic diameter smaller than x µm
POA Primary organic aerosol
POM Particulate organic matter
PTFE Polytetrafluoroethylene
pToF Particle-time-of-flight
SD Standard deviation
SDI Small deposit area impactor
SES Sample equilibration system
SMEAR Station for measuring forest ecosystem – atmosphere relations
SOA Secondary organic aerosol
TEOM Tapered element oscillating microbalance
TC Total carbon
TOC Total organic carbon
ToF Time-of-flight
TOT Thermal-optical transmittance
VI Virtual impactor
VOC Volatile organic compound
WINSOC Water-insoluble organic carbon
WSOC Water-soluble organic carbon
8
Contents
LIST OF PUBLICATIONS …………………………………………………………….. 10
1. INTRODUCTION ……………………………………………………………………. 12
2. OBJECTIVES OF THE STUDY ……..……………………………………………… 14
3. BACKGROUND …....………………………………………………………………… 15
3.1. Atmospheric aerosols ………………………………………………………………… 15
3.1.1. Particle properties ………………………………………………………..… 16
3.1.2. Health effects …………………………………………………………….… 19
3.1.3. Climate effects …………………..…………………………………………. 20
3.2. Particle chemical composition ……………………………………………………….. 20
3.2.1. Carbonaceous matter …………………………………………………….… 21
3.2.2. Soluble inorganic ions ……………………………………………...........… 24
3.2.3. Trace elements …………………………………………………………...… 25
3.2.4. Chemical mass closure …………………………………………………..… 26
3.3. Source apportionment ……………………………………………………………...… 27
3.3.1 Methods and tracer components ………………………………………….… 27
3.3.2. Resolved sources of atmospheric aerosols …………………………………. 29
4. EXPERIMENTAL …………………………………………………………………… 31
4.1. Measurements sites ………………………………………………………………..…. 31
4.2. Off-line measurements ……………………………………………………………….. 33
4.2.1. Sampling instruments ……………………………………………………… 33
4.2.2. Gravimetric analyses ………………………………………………………. 35
4.2.3. Chemical analyses ……………………………………………………….… 35
4.3. On-line measurements ……………………………………………………………...... 38
4.3.1. Physical properties …………………………………………………………. 38
4.3.2. Chemical composition …………………………………………….……...... 41
9
4.4. Comparison of off-line and on-line measurements ………………………………..… 44
4.4.1. Detection limits ……………………………………………………………. 44
4.4.2. Correlations ……………………………………………………………...…. 46
4.5. Laboratory calibration set-up ……………………………………………………...…. 49
4.6. Data analysis ……………………………………………………………………....…. 50
4.6.1. Data inversion (MICRON) ……………………………………………...…. 50
4.6.2. Receptor models ………………………………………………………...…. 50
5. RESULTS …………………………………………………………………………..…. 53
5.1. Chemical mass closure ………………………………………………………….....…. 53
5.1.1. Urban, background and remote sites ………………………………….....…. 53
5.1.2. Time-resolved mass closure …………………………….…………….……. 56
5.1.3. Size-resolved mass closure ……………………………………………...…. 58
5.1.4. Carbonaceous fraction …………………………………………………..…. 62
5.2. Source apportionment ……………………………………………………………..…. 64
5.2.1. Particulate matter ………………………………………………………..…. 64
5.2.2. Organic carbon ………………………………………………………….….. 66
6. SUMMARY AND CONCLUSIONS ……………………………………………..….. 71
7. REFERENCES ……………………………………………………….…………...….. 73
10
LIST OF PUBLICATIONS This thesis is based on the following six original papers, referred to by their Roman numerals.
I. Saarikoski S., Mäkelä T., Hillamo R., Aalto P.P., Kerminen V.-M. and Kulmala M. (2005)
Physico-chemical characterization and mass closure of size-segregated atmospheric aerosols
in Hyytiälä, Finland. Boreal Env. Res. 10, 385–400.
II. Saarikoski S., Sillanpää M., Sofiev M., Timonen H., Saarnio K., Teinilä K., Karppinen A.,
Kukkonen J. and Hillamo R. (2007) Chemical composition of aerosols during a major
biomass burning episode over northern Europe in spring 2006: experimental and modelling
assessments. Atmos. Environ. 41, 3577–3589.
III. Koulouri E., Saarikoski S., Theodosi C., Markaki Z., Gerasopoulos E., Kouvarakis G.,
Mäkelä T., Hillamo R. and Mihalopoulos N. (2008) Chemical composition and sources of
fine and coarse aerosol particles in the Eastern Mediterranean. Atmos. Environ. 42, 6542–
6550.
IV. Saarikoski S., Frey A., Mäkelä T. and Hillamo R. (2008) Size distribution measurement of
carbonaceous particulate matter using a low pressure impactor with quartz fiber substrates.
Aerosol Sci. Technol. 42, 603–612.
V. Saarikoski S., Sillanpää M., Saarnio K., Hillamo R., Pennanen A.S. and Salonen R. O.
(2008) Impact of biomass combustion on urban fine particulate matter in Central and
Northern Europe. Water Air Soil Pollut. 191, 265–277.
VI. Saarikoski S., Timonen H., Saarnio K., Aurela M., Järvi L., Keronen P., Kerminen V.-M.,
and Hillamo R. (2008) Sources of organic carbon in fine particulate matter in northern
European urban air. Atmos. Chem. Phys. Discuss. 8, 7805–7846.
These articles are reprinted with the kind permission of the respective copyright holders. In
addition, some unpublished material is included.
11
Other publications of the author not included in this thesis:
Timonen H., Saarikoski S., Tolonen-Kivimäki O., Aurela M., Saarnio K., Petäjä T., Aalto P.,
Kulmala M., Pakkanen T. and Hillamo R. (2008) Size distributions, sources and source areas of
water-soluble organic carbon in urban background air. Atmos. Chem. Phys. 8, 5635–5647.
Timonen H., Saarikoski S., Aurela M., Saarnio K. and Hillamo R. (2008) Water-soluble organic
carbon in urban aerosol: concentrations, size distributions and contribution to particulate matter.
Boreal Env. Res. 13, 335–346.
Niemi J.V., Saarikoski S., Tervahattu H., Mäkelä T., Hillamo R., Vehkamäki H., Sogacheva L.
and Kulmala M. (2006) Changes in background aerosol composition in Finland during polluted
and clean periods studied by TEM/EDX individual particle analysis. Atmos. Chem. Phys. 6,
5049–5066.
Sillanpää M., Hillamo R., Saarikoski S., Frey A., Pennanen A., Makkonen U., Spolnik Z., Van
Grieken R., Braniš M., Brunekreef B., Chalbot M.-C., Kuhlbusch T., Sunyer J., Kerminen V.-M.,
Kulmala M. and Salonen, R.O. (2006) Chemical composition and mass closure of particulate
matter at six urban sites in Europe. Atmos. Environ. 40S2, 212–223.
Virkkula A., Teinilä K., Hillamo R., Kerminen V.-M., Saarikoski S., Aurela M., Koponen I.K.
and Kulmala M. (2006) Chemical size distributions of boundary layer aerosol over the Atlantic
Ocean and at an Antarctic site. J. Geophys. Res. 11, D05306, doi:10.1029/2004JD004958.
Virkkula A., Teinilä K., Hillamo R., Kerminen V.-M., Saarikoski S., Aurela M., Viidanoja J.,
Paatero J., Koponen I.K. and Kulmala M. (2006) Chemical composition of boundary layer
aerosol over the Atlantic Ocean and at an Antarctic site. Atmos. Chem. Phys. 6, 3407–3421.
Sillanpää M., Saarikoski S., Hillamo R., Pennanen A., Makkonen U., Spolnik Z., Van Grieken
R., Koskentalo T. and Salonen R. O. (2005) Chemical composition, mass size distribution and
source analysis of long-range transported wildfire smokes in Helsinki. Sci. Total Environ. 350,
119–135.
12
1. INTRODUCTION
Atmospheric aerosols are particles suspended in the atmosphere. They originate from direct
emissions of particles or are formed in the atmosphere by the conversion of gases to particles. There
is a great concern of the health effects of aerosols. Epidemiological data has shown that the
increasing level of particles may cause significant increase in human morbidity and mortality
(Dockery et al., 2003; Samet et al., 2000; Brook et al., 2004). The health effects of particles depend
on various factors, e.g. on their size that defines the location at which they deposit within the
respiratory system (Hinds, 1999), and on their chemical composition. The most susceptible
population groups for the harmful effects of particles are asthmatics, elderly cardio-respiratory
patients and children.
Atmospheric aerosol particles influence climate and visibility by scattering and absorbing solar
radiation (Cabada et al., 2004; Andreae et al., 2005). Aerosol particles act also as the nuclei around
which cloud droplets form (Andreae and Rosenfeld, 2008), so without aerosols clouds could not be
formed at atmospheric conditions. However, if aerosol number concentration is substantially
increased, the number concentration of cloud droplets may also increase. This can lead to the
enhanced reflectivity and lifetime of the clouds (Seinfeld and Pandis, 1998). Both the direct
scattering and the indirect effect of aerosols via cloud formation and changing cloud properties cool
the climate. Despite extensive studies on atmospheric aerosol properties and modeling efforts, the
Intergovernmental Panel on Climate Change (IPCC) has stated that the radiative forcing caused by
aerosols remains the dominant uncertainty in total radiative forcing (IPCC, 2007).
Those two current topics in atmospheric science – climate change and health effects – are coupled
with each other. Most European Union (EU) member countries still count fossil fuel as their
primary source of energy (Taylor et al., 2005). EU’s aim is to reduce fossil fuel dependency and
energy-related emissions by promotion of renewable energy sources and alternative transport fuels–
in particular biofuels. Other important priority of the EU’s environmental policy is the reduction of
air pollution and related damages to human health and environment. However, although the
displacement of fossil fuel by bioenergy may decrease carbon dioxide emissions, the traditional air
13
pollutants are likely to increase due to higher emission factors associated with biomass burning
(Szarka et al., 2008).
This thesis presents the chemical composition of particles and their sources measured in the past
few years in Finland. In addition to Finland, aerosol measurements were carried out in central and
southern Europe in order to obtain a larger perspective for the air quality issues. Although the
chemical composition of particles has been rather constant in the recent few years, the expanding
consumption of biofuels and the increasing number of diesel automobiles is expected to change the
future aerosol composition in Europe. Therefore the results of this thesis can be considered to
represent the situation at present, against which the prospective changes in aerosol chemistry are
evaluated. For the Mediterranean site Finokalia and for Helsinki, this thesis will provide the most
comprehensive data on atmospheric aerosol composition published so far.
14
2. OBJECTIVES OF THE STUDY
The overall purpose of this thesis was to investigate the chemical composition and sources of fine
particles in the lower troposphere. Aerosol measurements were conducted in different environments
(traffic, urban background, remote), as well as during some specific air pollution situations such as
episodes of long-range transported emissions and particle formation events. Several complementary
experimental methods, both on-line and off-line, were deployed in order to obtain various
perspectives for the chemistry and the sources of atmospheric particles.
The specific objectives of this thesis were:
• to compare on-line instruments with filter-based methods and to illustrate their advantages
and disadvantages in aerosol chemistry studies (Papers II and VI)
• to find out how the chemical composition varies with particle size and time of the day
(Papers I, III and IV)
• to develop a method enabling one to determine the mass size distribution of organic carbon
(OC) and elemental carbon (EC) using size-segregating sampling and subsequent thermal
carbon analysis (Paper IV)
• to assess the contribution of wildfire and fuel biomass burning to fine particle aerosol mass
concentrations (Papers II, V and VI)
15
3. BACKGROUND
3.1. Atmospheric aerosols
Aerosol is a system that consists of solid and liquid particles suspended in a gas. One typical feature
of the atmospheric aerosol is that it undergoes constant physical and chemical transformation cycles
during its lifetime which is typically of the order of one week in the lower troposphere. Aerosol
particles originate from both natural and anthropogenic sources. Natural aerosol sources include
windblown soil from arid regions, natural weathering of rock, volcanic emissions, sea spray,
biomass burning from wildfires and biogenic aerosol formation. Anthropogenic aerosol sources
have four major categories: fuel combustion in the energy production, industrial processes, non-
industrial fugitive sources and transportation. Non-industrial fugitive sources include dust from
paved and unpaved roads, agricultural operations, construction and fires, whereas transportation
sources refer to vehicle exhaust and vehicle-related particles from tires, clutch and brake wear
(Seinfeld and Pandis, 1998). Primary particles are introduced directly into the atmosphere, whereas
secondary particles are formed in the atmosphere from gaseous components (Hinds, 1999).
Atmosphere is divided into five regions: troposphere, stratosphere, mesosphere, thermosphere and
exosphere (Seinfeld and Pandis, 1998). The two lowest layers, troposphere and stratosphere, are the
most important regions in terms of air pollution. Troposphere extends up to 10–15 kilometers from
the Earth’s surface and most of the aerosol mass is concentrated in the troposphere. The
temperature of troposphere decreases with increasing altitude resulting in rapid vertical mixing of
pollutants in the troposphere. The troposphere can be divided into the planetary boundary layer (0–2
km from the Earth’s surface) and free troposphere above it. Basically all of the water vapor, clouds
and precipitation are found in the troposphere. The tropospheric water cycle forms an important
removal mechanism of aerosols (Finlayson-Pitts and Pitts, 2000). Stratosphere covers the region on
the top of the troposphere extending to 50 kilometers from the Earth’s surface. The temperature
profile of the stratosphere is reverse to the troposphere, increasing with the increasing altitude. In
the stratosphere there is a permanent ozone layer which is essential for life because it absorbs the
short wavelength radiation of the sun. Stratosphere has also a natural aerosol layer at altitudes of 12
to 30 km. Stratospheric aerosol is composed of aqueous sulfuric acid and it is periodically increased
by volcanic eruptions.
16
3.1.1. Particle properties
The most commonly-measured aerosol property is the mass concentration that is the mass of
particulate matter (PM) in a unit volume of gas. For atmospheric aerosol the unit used for mass
concentration is usually µg/m3. The aerosol concentration can also be expressed as a number
concentration that is number of particles per unit volume of aerosol. The unit for number
concentration is typically number/cm3.
The size of a particle is the most important quantity for characterizing the behavior of particles
because almost all of the properties of aerosols depend on particle size. The diameter of
atmospheric aerosol particles ranges from a few nanometers to several hundreds of micrometers.
The smallest aerosol particles approach the size of gas molecules and have many properties similar
to those of gas molecules. Large aerosol particles are visible grains whose properties can be
depicted by the Newtons physics of baseball and automobiles. Particle aerodynamic diameter (Da)
is defined as the diameter of the spherical particle with a unit density (1 g/cm3) that has the same
settling velocity as the actual particle. Particle electrical mobility diameter is a diameter of particles
with a specific electrical mobility. Impactors classify particles according to their aerodynamic
diameters, whereas in a typical size distribution measurement made using electrical instruments the
electrical mobility diameter of particles is measured.
Figure 1 shows a typical size range for atmospheric aerosols. Particles below 2.5 µm in diameter are
generally called as “fine” particles and those above 2.5 µm in diameter are called “coarse” particles.
Also the terms “submicron” and “supermicron” particles are used, referring to particles smaller and
greater than 1 µm in diameter, respectively. Ultrafine particles cover the size range from large gas
molecules up to about 0.1 µm, whereas particles less than 50 nm in diameter are called
nanoparticles (not shown in Figure 1). Particles of different sizes typically originate from different
sources, are transported separately, and have different removal mechanisms, chemical composition,
optical properties and deposition in a respiratory tract.
17
dm/d
logD
p (μg
m-3
)
Ultrafine particles
Fine particles Coarse particles
Coarse-particle mode Accumulation mode
100 10 1 0.01 0.1 Particle diameter (μm)
Mechanically produced particles: wind blown dust sea spray volcanoes plant particles
Sedimentation
Rainout &
Washout
Nucleation
Low volatility vapor
Primary particles
Chain aggregates
Coagulation
Secondary particles
Hot vapors
Condensation
Condensation
Condensation
Coagulation
Nucleation and Aitken modes
Figure 1. The distribution of particle mass of an atmospheric aerosol and principal modes, sources and particle formation and removal mechanisms. Adapted from Seinfeld and Pandis (1998).
The size distribution of atmospheric aerosol particles can be presented in terms of four modes: the
nucleation, Aitken, accumulation and coarse particle mode (Figure 1). Ultrafine particles, i.e.
nucleation mode (Dp < 20–25 nm) and Aitken mode (20 nm < Dp < 0.1 µm) particles, have the
majority of particles by number (typically >90%), but because of the small size of particles they
seldom accounts for more than a few percent of total mass of atmospheric particles. Ultrafine
particles consist of combustion particles formed by the condensation of hot vapors or fresh particles
formed in the atmosphere by nucleation. They are usually found near combustion sources e.g. close
to highway or in environments where the nucleation of biogenic volatile organic compounds
(VOCs) takes place (Kulmala et al., 2001). Because of large number of particles in the ultrafine
18
mode, they can coagulate rapidly with each other and with particles in the accumulation mode.
Nucleation mode particles have short lifetimes (typically a few hours) while Aitken mode particles
can be found far away from their sources.
The accumulation mode, that roughly covers a size range from 0.1 to 2.5 µm, has usually a
substantial fraction of the total particulate mass. The source of particles in the accumulation mode is
the coagulation of ultrafine mode particles but also particles that have grown to accumulation size
by condensation of vapors onto existing particles and by cloud processing. As the mode name
suggests, the particle removal mechanisms are least efficient in this size range, and therefore
accumulation mode particles remain airborne for many days and consequently can travel long
distances in the atmosphere. Accumulation mode particles can be removed from the atmosphere by
rainout or washout. Coagulation to coarse mode is slow as well, hence there is relatively little mass
shift from accumulation mode to coarse mode (Hinds, 1999). Accumulation mode particles cause
most of the visibility effects of atmospheric particles since the accumulation size range includes the
wavelength range of visible light (Hinds, 1999).
Coarse mode particles (> 2.5 µm) are formed by mechanical processes and usually include both
man-made and natural dust particles. In addition to dust, coarse mode consists of large sea salt
particles and particles from volcanic eruptions. Because of their large size, coarse particles are
removed from the lower atmosphere by sedimentation or inertial impaction over a reasonably short
time scale (typically from hours to one day). Although coarse particles have a limited life time in
the atmosphere, they can cause exposure to people close to the particle source.
Particles that are liquid have usually a spherical shape. Also some solid particles formed by
condensation can be spherical, even though most of them have complex shapes. Some of solid
particles have regular geometric shapes like cubic (sea salt particles), cylindrical (fibers), single
crystals, or clusters of spheres. Crushed material or agglomerated particles have irregular shapes
(Hinds, 1999, see Figure 2). Also the particle density varies. The particle density refers to the mass
per unit volume of the particle itself, not of the aerosol. For liquid particles and crushed solid
particles, the density is equal to that of their parent material. Combustion particles may have low
19
densities because of their highly agglomerated structure (Figure 2a). Particle density is generally
given in g/cm3. The standard density (ρ0) is the density of water (1.0 g/m3).
500 nm
(b)
Figure 2. Transmission electron microscope image of soot particle (a) and Na-S-K-O rich particle (Niemi et al., 2006).
3.1.2. Health effects
Exposure to particles has been shown to increase morbidity and mortality. The health effects of
particles depend on the location at which they deposit within the respiratory system and on their
chemical composition. The deposition in the respiratory tract is influenced by the size of a particle
and its shape. Because of a high relative humidity in the respiratory tract, hygroscopic properties of
particles affect their growth and subsequent deposition. A large fraction of coarse particles deposit
already in head airways (e.g. nose, mouth) but smaller particles can be carried into lung airways or
alveolar region. The deposition is also affected by the respiratory frequency and volume, which
vary within individual persons. The removal of particles from the lungs depends on the location in
the lung they are deposited, and on their solubility. The head and lung airways are covered with a
layer of mucus that carries the deposited particles to the gastrointestinal tract, whereas in the
alveolar region the soluble material from particles can pass through the thin alveolar membrane and
transfer into the bloodstream (Hinds, 1999). Potential chemical components that can cause health
effects are metals, acids, organic components, soluble salts, peroxides and black carbon (Lighty et
al., 2000). However, no single chemical species seems to dominate the health effects. On the
20
contrary, the health effects appear to be a combined effect of different species (Davidson at al.,
2005). Regarding the particle sizes the ultrafine particles (<0.1 µm in diameter) are potentially most
dangerous to human health (Nel, 2005). The most susceptible people to particle exposure are those
weakened by illnesses, such as cardiovascular disease and asthma, as are also elderly and very
young people.
3.1.3. Climate effects
Particles in the troposphere have an influence on the global climate. They scatter incoming solar
radiation back into space by two mechanisms. The first mechanism is the direct scattering by
particles themselves. The second mechanism arises from the increased concentration of cloud
condensation nuclei, creating higher number of smaller cloud droplets and amplifying cloud
reflectivity. Both scattering mechanisms modify the Earth’s radiation balance and cool the Earth’s
surface (Hinds, 1999). Some particles also affect the global climate by absorbing solar radiation or
by changing the surface albedo after deposition in the Polar areas. In a recent study of Ramanathan
and Carmichael (2008), it was estimated that light absorbing black carbon (BC) particles may have
a global warming effect which is comparable to that of CO2.
3.2. Particle chemical composition
The chemical composition of fine and coarse particles differ substantially from each other (Figure
3). Atmospheric fine particles are typically acidic and are composed mainly of sulfate, ammonium,
nitrate, organic and elemental carbon and water. Coarse particles are basic and account for most of
the crustal material and their oxides and large sea-salt particles. Trace metals, nitrate and organic
compounds are found in both fractions, but the majority of organic carbon is generally in fine
particles. In the ultrafine mode (< 0.1 µm), most of the particulate matter consists of organic
components (Jimenez et al., 2003) and a small amount of inorganic ions (mostly sulfate) (Matta et
al., 2003). In urban areas elemental carbon may have a significant contribution to ultrafine mode
(Berner et al., 1996). In the following sections (Sections 3.2.1 to 3.2.4), the chemistry of particles is
discussed in more detail.
21
Figure 3. Typical composition distribution in ultrafine, accumulation and coarse size fractions.
dm
/dlo
gDp (μg
m-3
)
100 10 1 0.1 Particle diameter (μm)
Fe Ca Si Na+ Cl- Al NO3
-
OC
SO42-
NH4+
NO3-
Cd Pb OC EC H2O OC
EC, SO42-
3.2.1. Carbonaceous matter
Carbonaceous matter in particles is comprised of two components: elemental carbon, also called
graphitic carbon, black carbon or soot, and organic matter. EC has a chemical structure similar to
impure graphite and it is produced only in the combustion processes. EC (or BC) particles are
efficient light absorption species in the atmosphere. Organic matter is a complex mixture of many
classes of organic compounds. Organic carbon is emitted directly from sources (primary organic
aerosol, POA) or formed in the atmosphere from gaseous precursors (secondary organic aerosol,
SOA). Sources for primary OC include meat cooking, paved road dust, fireplaces, vehicles, forest
fires and cigarettes (Rogge et al., 1996), but also natural biogenic detritus (e.g. leaf wax, microbes,
pollen) contributes to primary organic carbon in aerosol particles (Medeiros et al., 2006). In
addition to OC and EC, small quantities of carbon may exist as carbonates (e.g. CaCO3) or CO2
adsorbed onto particulate matter.
The contributions of primary and secondary OC are difficult to determine in a quantitative way, but
several methods have been used to assess the share of POA and SOA fractions in OC. EC has been
used as a tracer for POA by assuming that EC and POA have same sources, and hence there is a
22
representative ratio of OC/EC for primary aerosol (Turpin and Huntzicker, 1995; Viidanoja et al.,
2002). One limitation of the EC-tracer method is that the OC/EC ratio varies with source type but is
also affected by the meteorology, diurnal and seasonal changes in the emissions and local sources
(Castro et al., 1999, Robinson et al., 2007; Kroll and Seinfeld, 2008). The contributions of POA and
SOA have also been predicted using the models that couple the formation, transport and deposition
with atmospheric dynamics (Strader et al., 1999; Simpson et al., 2007).
The water-solubility of organic carbon affects the chemical and physical properties of aerosols,
including their hygroscopic behavior (the ability of particles to act as condensation cloud nuclei),
acidity and radiative properties (Jacobson et al., 2000). Therefore OC in atmospheric aerosols is
often divided into water-soluble organic carbon (WSOC) and water-insoluble organic carbon
(WISOC). WSOC is assumed to contain the more oxygenated and polar fraction of OC, whereas
WINSOC includes e.g. non-polar hydrocarbons. A large fraction of WSOC is suggested to originate
from secondary formation (Kondo et al., 2007). The fraction of WSOC in OC is usually small near
the sources, such as near a fossil fuel combustion sources in urban areas, to increase in more aged
aerosol at remote locations (Ruellan and Cachier, 2001; Pio et al., 2007). According to Sullivan and
Weber (2006a), WSOC can be further fractionated into hydrophobic and hydrophilic fraction based
on their retention to XAD-8 resin. That method was further developed to separate acid, neutral and
basic functional groups of hydrophobic and hydrophilic WSOC (Sullivan and Weber, 2006b). In
addition to water, the solubility of organic fraction to other solvents, such as n-hexane (Zappoli et
al., 1999), dichloromethane (Carvalho et al., 2003) and methanol (Polidori et al., 2008), has been
studied.
Besides solubility, particulate organic matter (POM) can be characterized by measuring its
volatility. The volatility is examined by studying the thermal evolution of carbon mass during
evolved gas analysis (EGA) (Sillanpää et al., 2005a; Pio et al., 2007). The fraction of OC
evaporated at low temperatures (<150 °C) represents semivolatile OC, even though some
semivolatile OC may char during the heating process and subsequently evaporate at higher
temperatures. Also the presence of inorganic salts changes the thermal evolution of organic
components (Yu et al., 2002). Extracting aerosol samples with water has been found to change the
23
thermal profile (Gelencsér et al., 2000), indicating different volatilities of WSOC and WINSOC
fractions.
By using an Aerosol Mass Spectrometer (AMS, see Chapter 4.3.2.), the organic matter can be
characterized further. Organic matter can be separated into hydrocarbon-like organic aerosol (HOA)
and oxygenated organic aerosol (OOA) (Zhang et al., 2005a; Zhang et al., 2005b). That method
uses mass-to-charge ratios (m/z) of 57 and 44 as traces for HOA and OOA, respectively. OOA can
be further divided into the highly-aged fraction with a low volatility, OOA I, and the more volatile
and probably less processed fraction, OOA II (Lanz et al., 2007).
Several hundreds of organic species have been identified from atmospheric samples, including
alkanes, polycyclic aromatic hydrocarbons (PAHs), alkanoic, dicarboxylic and aromatic acids.
Nevertheless, the organic speciation using recent analytical methods has succeeded to assign only
10–40% of the total carbon to individual organic compounds (Rogge et al., 1993b; Alves et al.,
2002). One reason for the low identified fraction can be that a considerable fraction of organic
matter is suggested to be composed of polymers or other macromolecules not analyzed by
chromatography techniques routinely. Kalberer et al. (2004) discovered in laboratory tests that the
reaction of carbonyls and their hydrates resulted in organic polymers that made half of the particles
mass after aging for more than 20 hours. Humic-like substances (HULIS) are found to be one of the
largest classes of organic compounds extracted from atmospheric particles (Graber and Rudich,
2006). HULIS consists of organic macromolecular compounds with average molecular weight
values in the range of 200–300 Da. These macromolecules possess similar properties to those of
fulvic and humic acids (Kiss et al., 2003). HULIS originates most likely from various secondary
anthropogenic and biogenic sources (e.g. biomass burning, reactive biogenic emissions) (Puxbaum
et al., 2007). The fraction of HULIS in fine particle WSOC may vary largely from 15–36% in
Amazon biomass burning aerosol (Mayol-Bracero et al., 2002) to 55–60% in European fine aerosol
(Krivácsy et al., 2001).
24
3.2.2. Soluble inorganic ions
Inorganic ions include secondary inorganic ions like sulfate, ammonium and nitrate, as well as sea
salt and soluble fraction of soil dust. Depending on the location, ammonium salt of either nitrate or
sulfate dominates the inorganic fine particle composition (Chow et al., 1994; Matta et al., 2003;
Paper I). In fine particles, nitrate is usually the result of the reaction between gaseous nitric acid and
ammonia to form ammonium nitrate (Seinfeld and Pandis, 1998):
(s) NONH(g) HNO(g) NH 3433 ↔+ (1)
Ammonium nitrate is formed in areas where ammonia and nitric acid concentrations are high and
the sulfate concentration is low. Depending of the ambient relative humidity, ammonium nitrate can
exist as a solid salt crystal or as a droplet comprising aqueous solution of NH4+ and NO3
-.
In coarse particles nitrate, is the product of the reaction between nitric acid and the crustal material
or nitric acid with the sea-salt particle (Finlayson-Pitts and Pitts, 2000):
(l) OH(g) CO(s) )Ca(NO(g) 2HNO(s) CaCO 222333 ++→+ (2)
(s) NaNO(g) HCl(g) HNO(s) NaCl 33 +↔+ (3)
Sulfuric acid (H2SO4) dissociates to bisulfate (HSO4-) in aerosol phase that subsequently dissociates
to sulfate by the reaction:
(aq)SO(aq)H(aq)HSO 244−+− +↔ (4)
If there is a low amount of ammonium available, sulfuric acid exists in the aerosol phase. As the
ammonium concentration increases, sulfuric acid is converted to HSO4- and its salts. If there is an
abundance of ammonium, sulfuric acid is converted to ammonium sulfate by the reaction (Seinfeld
and Pandis, 1998):
(s) SO)(NH(aq) SO(aq) NH 2 424244 ↔+ −+ (5)
25
Sodium and chloride (sea salt) are present in aerosol with substantial concentrations in marine
environments. Besides with nitric acid (equation 3), sodium chloride can react with sulfuric acid
resulting in the chloride deficiency in aerosol particles (Seinfeld and Pandis, 1998):
(g) HCl 2(s) SONa(g) SOH(s) NaCl 2 4242 +↔+ (6)
(g) HCl(s) NaHSO(g) SOH(s) NaCl 442 +↔+ (7)
Potassium and calcium in the coarse fraction originate from soil dust. However, only 51–86% of the
total amount of calcium in aerosol has been found to be in the form of water-soluble ions, whereas
for potassium the corresponding percentage range was 15–79% (Pakkanen et al., 2001; Sillanpää,
2006a). Besides soil dust, water-soluble potassium has been associated with other sources such as
sea salt (Brewer, 1975), biomass burning (Andreae, 1983; Paper II) and biogenic origins (Artaxo
and Hansson, 1995). Potassium from biomass burning can be found mainly in the fine size fraction.
3.2.3. Trace elements
More than 40 trace elements have been identified in atmospheric aerosol particles. Their sources
include coal and oil combustion, wood burning, steel furnaces, boilers, smelters, dust, waste
incineration and break wear (Seinfeld & Pandis, 1998). Because they have a primary origin, local
sources influence strongly in concentration levels, which causes large temporal variations. At
certain locations some trace elements may be dominated by long-range transport (Hueglin et al.,
2005; Sillanpää et al., 2005b).
In case of metals it is often necessary to quantify specific metallic forms since their bioavailability,
solubility and geochemical transport depend on the physical-chemical speciation. Particularly the
knowledge of the chemical forms of metals is essential when the effects of metals on human health
are investigated. Elements emitted during combustion typically exist as oxides (e.g. Fe2O3, Fe3O4,
Al2O3) but their specific chemical form is usually uncertain (Seinfeld and Pandis, 1998).
26
3.2.4. Chemical mass closure
By comparing the sum of the masses of the individually identified chemical species to the
gravimetric particulate matter, aerosol chemical mass closure can be constructed. Usually chemical
components and gravimetric mass are analyzed from filter samples. In order to achieve better time-
resolution for the chemical mass closure, continuous or semi-continuous instruments can be used in
lieu of filter or impactor samples (Jeong et al., 2004; Park et al., 2006; Grover et al., 2006).
A majority of particle mass can be determined by analytical methods. For fine particles the extent of
mass closure is usually greater than that for coarse particles (Brook, et al., 1997; Hueglin et al.,
2005; Sillanpää et al., 2006b; Paper III). The reason for this difference is the large contribution of
dust in coarse particles. While multielement techniques, such as X-ray fluorescence and Particle
induced X-ray emission, are able to analyze the elemental composition of dust, assumptions
regarding the unaccounted presence of heteroatoms (e.g. oxygen) are still needed. The contribution
of soil dust in particles can be assessed by using the concentrations of aluminum, silicon, calcium,
iron, titanium and potassium and assuming that they appear as oxides (Brook et al., 1997). In
simpler approaches only iron (Paper III), calcium (Putaud et al., 2004) or calcium and iron
(Harrison et al., 2003; Yin and Harrison, 2008) concentrations were utilized in order to estimate the
contribution of dust to aerosol mass. Water-soluble and water-insoluble soil dust fractions have
been estimated separately (Graney et al., 2004; Sillanpää et al., 2006b), since they have different
health and climatic effects.
In addition to the dust content, chemical mass closure studies generally include an estimate of the
contribution of particulate organic matter. Since there are limited number of instruments available
to measure directly POM (only AMS), the organic carbon content of particles is typically
determined instead of POM. Thus a conversion of the mass of OC to that of POM is required in
order to take into account other atoms than carbon in POM (mostly hydrogen and oxygen). Turpin
et al. (2001) recommended the conversion factor of 1.6 ± 0.2 and 2.1 ± 0.2 for an urban and non-
urban aerosol, respectively, calculated from the POM-to-OC ratios of organic species identified at
various measurement sites. Aerosol heavily impacted by wood smoke can have a higher ratio (2.2–
2.6), as WSOC is assumed to have higher POM-to-OC ratio than that of WINSOC. Russel (2003)
obtained lower ratios (1.2–1.6) than those suggested by Turpin et al. (2001) for samples collected in
the Caribbean and northeastern Asia by using Fourier transform infrared spectroscopy that
27
quantifies OM by using functional groups. The POM-to-OC ratio can also be determined by using
an AMS. By comparing POM from the AMS to OC measured by the semicontinuous OC/EC
analyzer, Kondo et al. (2007) measured a POM-to-OC ratio of 1.72 and 2.22 for winter and summer
aerosol in Tokyo, respectively. Zhang et al. (2005b) obtained an average POM-to-OC ratio of 1.8 in
fall in Pittsburgh. Zhang et al. (2005b) also separated HOA and OOA and obtained POM-to-OC
ratios of 1.2 and 2.2 for HOA and OOA, respectively. A high-resolution time-of-flight (ToF) AMS
enables the determination of the POM-to-OC ratio without other instruments (Aiken et al., 2008). In
Mexico City an average POM-to-OC ratio of 1.71 was obtained by using a ToF-AMS.
A part of unaccounted particulate matter mass can consists of water associated with particles.
Aerosol water content is either directly measured (Khlystov et al., 2005), estimated based on the
chemical composition of particles (Temesi et al., 2001: Sellegri et al., 2003) or modeled by
employing a mass transfer concept (Kajino et al., 2006). Nonetheless, the contribution of water is
frequently ignored in mass closure studies though it has been suggested that particle bound water
can constitute up to 20–35% of the annual mean PM10 and PM2.5 concentrations (Tsyro, 2005). If
compounds not crystallizing at low relative humidities, like ammonium bisulfate, contribute to
aerosol mass significantly, water content in particles may be substantial.
3.3. Source apportionment
3.3.1. Methods and tracer components
The origin of particulate matter is attributed to different sources by using statistical methods. Two
approaches are employed in order to assess the source contributions: source-oriented models and
receptor-oriented models (Schauer et al., 1996). Source-oriented models use data from emissions
and their transport calculations to predict pollutant concentrations at specific receptor sites
(Eldering and Cass, 1996). Receptor-oriented models assess the source contributions by
determining the best-fit linear combination for the chemical composition profiles of the emission
sources which are needed to reconstruct the measured chemical composition of ambient aerosol
(Watson, 1984). The most common receptor models used for ambient aerosols are Positive matrix
factorization (PMF; Paatero 1997; Paatero 1999) and chemical mass balance (CMB; Schauer et al.,
1996). In the PMF the chemical composition of PM measured at a given site is exploited to identify
28
the potential sources of PM at that site. In contrast, in the CMB the chemical composition of PM
measured at its sources (e.g. traffic, wood combustion) is used to model the contribution of these
sources to the PM concentrations at the site.
Certain elemental and organic compounds are specific for particular sources. For the identification
and quantification of aerosol sources, many organic and elemental molecular tracers have been
employed. Hopanes and steranes have been used as tracers for traffic emissions (Rogge at al.,
1993a; Zielinska et al., 2004; Brook, et al., 2007). There are also a number of organic tracers for
smoke from incomplete combustion (Simoneit, 2002), such as levoglucosan (Simoneit et al., 1999;
Nolte et al., 2001; Paper VI) and retene (Schauer et al., 1996). Sugars have suggested as tracers for
resuspension of soil from agricultural activities (Simoneit et al., 2004) and airborne fungal spores
(Bauer et al., 2008). Steroids have been used as tracers for the dust from open lot dairies and cattle
feedlots as well as for fungal biomass in atmospheric aerosols (Lau et al., 2006; Rogge et al., 2006).
However, the reliability of organic tracers suffers from the limited atmospheric lifetimes due to
their chemical reactivity (Gao et al., 2003; Robinson et al., 2006) and highly variable emission
factors (Fine et al., 2001; Fine et al., 2002; Fine et al., 2004; Hedberg et al., 2006).
Regarding elemental tracers, V and Ni are indicative of oil combustion, whereas elevated
concentrations of As and Se are associated with coal burning and smelter operations (Chow et al.,
1994). Cu and Zn have been utilized to indicate the impact of road traffic (Sillanpää et al., 2005a).
The concentrations of elements commonly associated with mineral dust are aluminum, silicon, iron,
titanium, calcium, magnesium and manganese (Graham et al., 2003). As mentioned previously,
potassium has been frequently used as a tracer for biomass burning (Jaffrezo et al., 1998; Duan et
al., 2004; Paper II).
The estimates of the contribution of biomass combustion to atmospheric aerosols have been
calculated based solely on the amount of levoglucosan in particles (Zraháhal et al., 2002; Yttri et
al., 2005; Wang et al., 2007; Puxbaum et al., 2007; Paper V). The characteristic levoglucosan-to-
PM or levoglucosan-to-OC ratios for biomass combustion were obtained from the laboratory
experiments published previously (e.g. Fine et al., 2001; Schauer et al., 2001; Rogge et al., 1998).
Similar approach was used for the investigation of the motor vehicle exhaust contribution to
29
primary OC in Toronto (Brook et al., 2007). However, they used several tracer compounds and
measured their emission profiles by themselves. The origin of organic compounds has also been
studied by calculating a Carbon preference index (CPI) which is the sum of the odd carbon number
components over the sum of the even carbon number components over a specific range of C atoms.
For example CPI>4 for C24–C33 n-alkenes indicates a major incorporation of biogenic components
to the n-alkene fraction whereas the presence of n-alkenes from fossil fuel combustion reduces CPI
(Plaza et al., 2006).
Besides tracers, sources of atmospheric aerosol have been evaluated by using accurate
measurements of radiocarbon 14C present in the atmosphere and transferred to plants by CO2 uptake
(Sheffield et al., 1994: Lemire et al., 2002: Szidat et al., 2004). 14C is absent in fossil fuels, since it
decays with a half-life of 5730 years, but it is present in living and contemporary plant materials.
Hence, measuring the ratio of 14C to stable carbon isotope, either 13C or 12C, provides information
about the contributions of fossil fuel, biomass burning and biogenic emissions to carbonaceous
aerosols. In addition to 14C the sources of carbonaceous matter have been investigated by measuring
the 13C/12C ratio of OC and EC (Huang et al., 2006). The determination of 13C has also a potential
to trace the extent the atmospheric photochemical processing evidenced by the investigation of
photochemical processing of isoprene (Rudolph et al., 2003).
3.3.2. Resolved sources of atmospheric aerosols
In Fresno, California, the origin of fine particulate matter (PM2.5) was attributed to 13 sources:
paved road dust, gasoline vehicles, diesel vehicles, hardwood combustion, softwood combustion,
smoked chicken, charbroiled chicken, propane chicken, charbroiled hamburger, meat cooking, sea
salt, ammonium nitrate and ammonium sulfate (Chow et al., 2007). Of these sources the largest
contributions to PM2.5 were obtained for ammonium nitrate (23%), wood combustion (sum of hard-
and softwood 22%) and vehicle emissions (sum of gasoline and diesel vehicles 9%). In addition to
those sources, PM2.5 has attributed to vegetative detritus, natural gas combustion, fuel oil
combustion and cigarette smoke, but their contributions were small (Schauer et al., 1996; Fraser et
al., 2003).
30
Most of the sources associated with PM2.5 have emissions of carbonaceous matter. For primary
organic aerosol eight source classes was applied in Pittsburgh: diesel vehicles, gasoline vehicles,
road dust, biomass combustion, cooking emissions, coke production, vegetative detritus, and
cigarette smoke (Subramanian et al., 2007). Of these the major contributions were attributed to
gasoline vehicles and cooking. In Fresno the major sources for primary OC were wood smoke (sum
of hard- and softwood 56–66%), meat cooking (5.8–14%), diesel exhaust (7.5–11%) and gasoline
vehicles (5.9–6.2%) (Schauer et al., 2000). Secondary organic aerosol originates from
anthropogenic and biogenic sources of which the biogenic sources have been found to exceed
anthropogenic ones even in an urban environment (Szidat et al., 2006). Additionally, biogenic SOA
was suggested to be a major source for biogenic OC whereas primary biogenic sources for OC (e.g.
vegetation debris) were much smaller. All EC sources have been found to be anthropogenic, of
which the fossil fuel usage was by far the prevailing (~75–94%) with biomass burning contributing
the remaining of EC (Szidat et al., 2006).
For some of the sources the seasonal variation is evident. In Houston Texas the contributions of
wood combustion, vehicle exhaust and vegetative detritus to PM2.5 were larger in winter than in
summer but that of paved road dust was the largest in springtime (Fraser et al., 2003). In Beijing
also coal combustion had larger contribution to PM2.5 in winter than in summer (Song et al., 2006).
Similar to PM2.5, biomass combustion had significantly greater contribution to OC in winter than in
summer (Zdráhal, et al., 2002; Puxbaum et al., 2007; Paper V), whereas in summer the biogenic
emissions dominated OC (Szidat et al., 2004; Paper VI). In contrast, the contribution of motor
vehicles did not exhibit a seasonal pattern in Toronto and Vancouver, Canada (Brook et al., 2007).
31
4. EXPERIMENTAL
The field measurement sites of this thesis varied from polluted urban to clean remote environments.
Several different instruments were used in order to measure both the PM mass and its chemical
composition. Measurements can be divided into two categories. In the off-line measurements the
samples were collected first on a substrate or filter and gravimetric and/or chemical analysis were
conducted in the laboratory. In the on-line measurements the sampling and analysis were either a
combined procedure (e.g. AMS) or performed successively within the same instrument(s) (e.g.
semicontinuous OC/EC analyzer). In addition to atmospheric aerosol studies, this thesis comprises a
work where a method for the determination carbonaceous matter size distribution by using a
multistage impactor was developed. At the end of Section 4 the methods used in analyzing the data
are presented.
4.1. Measurements sites
Measurements were made at nine sites in six different countries. In Finland, measurements were
carried out at three sites in Helsinki and at a site in Hyytiälä. Two of the Helsinki sites represent
urban background environment, whereas one of the sites was located at a roadside. In Hyytiälä the
measurement station was in the boreal forest in a rural area. The central European sites (in
Germany, Czech and Netherlands) were all urban background sites, but in southern Europe, in
Greece and Italy, the measurements were carried out at remote coastal and rural background areas,
respectively. Measurement sites and the measurement periods are summarized in Table 1 and
shown in the map in Figure 4.
32
Table 1. Measurement sites and periods used in this thesis.
Site Site type Measurement period Paper
Finland Helsinki, Itäväylä Roadside Jan–Mar 2004 IV
Hyytiälä, (SMEAR* II)
Boreal forest
7–31 May 2004 Feb 2007–Feb 2008
I, IV
Helsinki, Kallio Urban background 2 Oct–3 Nov 2003 9 Jan–2 Feb 2004 1 Apr–3 May 2004 1 Jul–2 Aug 2004
V
Helsinki, (SMEAR III) Urban background 1 Mar 2006–28 Feb 2007 Feb 2007–Feb 2008
II, VI
Germany Duisburg Urban background 4 Oct–21 Nov 2002 V
Czech Prague Urban background 29 Nov 2002–16 Jan 2003 V
Netherlands Amsterdam Urban background 24 Jan–13 Mar 2003 V
Greece Finokalia Remote coastal Jul 2004–Jul 2006 III
Italy San Pietro Capofiume Rural background 31 Mar–20 Apr 2008 -
* Station for Measuring forest Ecosystem – Atmosphere Relations
●●
●
●
●
●
● Helsinki
Hyytiälä Amsterdam
Duisburg
Prague San Pietro Capofiume
Finokalia
Figure 4. Measurement sites of this study (Map © Genimap).
33
4.2. Off-line measurements
4.2.1. Sampling instruments
Size classification in the off-line aerosol sampling was done by using inertial impactors. Inertial
impactors separate particles according to their aerodynamic diameter into two or more size
fractions. The aerosol is passed trough a nozzle and then directed against a flat plate. The flat plate
deflects the flow to form 90° turn in the streamlines. Particles whose inertia exceeds certain value
can not follow the curved streamlines and they collide on the flat plate, whereas smaller particles
can follow the streamlines and avoid impacting the plate. In the cascade (or multistage) impactors
several impactor stages with different cut-off sizes are operated in series in order to classify the PM
into different size ranges. In the following text the sampling instruments are described briefly.
VI. In Papers I and III–V the samples were collected with a virtual-impactor (VI; Loo and Cork,
1988), in which the impaction plate is replaced by a nozzle followed by the filter for the sample
collection. In the VI the flow is divided into two flows after the accelerating nozzle: a major flow
(90%) and minor flow (10%). Similar to the conventional impactors, smaller particles follow the
curved main flow while larger particles with enough inertia continue the straight stream line and are
transported to the filter by the minor flow. Because particles from the main flow are also filtrated,
the VI divides particles into two size fractions, typically into fine and coarse particles. In Paper V
the cut-off size of the VI was at 2.5 µm, whereas in Papers I, III and IV the modified version of the
VI with a cut-off at 1.3 µm was used. The lower cut-off diameter of the VI enables more accurate
comparison of the aerosol mass concentration with the instruments measuring the submicron
fraction only (Putaud et al. 2000). The flow rate of the VI was 16.7 L/min. The time-resolution for
the VI samples varied from 17 hours to four days.
PM1 impactor and BLPI. In Papers II and VI PM1 particles were collected using a filter cassette
system (Gelman Sciences). In order to discriminate particles with an aerodynamic diameter larger
than 1 µm, the four upper stages (8–11) of the Berner low pressure impactor (BLPI; Berner and
Lürzer, 1980) were used prior to the filter cassette using a flow rate of 80 L min-1. The 50%
aerodynamic cut-off diameter (D50) of the preimpactor is determined by the lowest stage, the other
stages are used to divide the PM loading to several stages and to ensure removal of particles well
34
above the lowest cut-size. The nominal D50 value for stage 8 is 1.8 µm with a flow rate of 24.5 L
min-1, but since the flow rate of the BLPI was increased to 80 L min-1, the D50 value for the stage 8
decreased to 1 µm. A similar sampling system consisting of the BLPI and filter cassette was also
used in Hyytiälä from February 2007 to February 2008. Besides using the BLPI as a preimpactor, it
was used in its original design in Paper V in order to collect size-segregated aerosol samples. The
aerodynamic cutoff diameters for the BLPI stages are 0.035, 0.067, 0.093, 0.16, 0.32, 0.53, 0.94,
1.8, 3.5 and 7.5 µm. The time-resolution for the PM1 samples was typically one day, but in case of
the BLPI the sampling time was from three to four days.
SDI. Size segregated samples were collected also using a Small Deposit area low pressure Impactor
(SDI; Papers I and IV). The SDI is described in detail by Maenhaut et al. (1996) and in Paper IV. In
short, the SDI is a 12-stage, low-pressure, multinozzle inertial impactor that operates at a flow rate
of 11 L/min. The nozzles of the impactor stages are in a cluster close to each other, so that the
diameter of aerosol deposition area remains smaller than 8 mm for each stage. The aerodynamic
cutoff diameters of the stages are 0.045, 0.088, 0.142, 0.235, 0.380, 0.580, 0.800, 1.06, 1.61, 2.60,
4.07 and 8.40 µm. In Paper IV the SDI was calibrated for the quartz fiber substrates used when the
samples are analyzed for OC and EC by the thermal methods. By using the quartz substrates in the
SDI, the cut-off diameters for stages 1–8 were equal to 0.009, 0.037, 0.063, 0.145, 0.277, 0.449,
0.608 and 0.878 µm, respectively (Paper IV). The sampling period for SDI varied from 17 hours to
five days.
Filters/substrates. Filter or impactor substrate material used for the sampling depended on the
subsequent analytical method. In the VI both polytetrafluoroethylene (PTFE) and quartz fiber filters
were used, but in the PM1 impactor particles were collected only on quartz fiber filters. SDI
substrates were cut from either quartz filter or aluminum foil. In the BLPI the substrate material was
a polycarbonate film supported by an aluminum foil in order to reduce static charges and to make
the substrate handling and weighing easier. The polycarbonate film and aluminum foil were greased
with a thin layer of Apiezon L-vacuum grease to prevent particle bouncing.
Denuders. Quartz material collects an unknown fraction of gaseous organic compounds in addition
to aerosol particles (Turpin et al., 2000). To reduce this positive artifact, gas-phase organic
35
compounds were removed from the air stream before the quartz filters/substrates (SDI and VI) with
three multi-annular denuders (URG-2000, 30×242 mm, Chapel Hill, NC) (Papers I, II, IV). The
denuders were coated with XAD-4 (polystyrene-divinylbenzene) adsorbent according to Gundel et
al. (1995). After the use the collected gas-phase compounds were extracted from the adsorbent with
acetonitrile and hexane. The extract of the denuders was not analyzed.
4.2.2. Gravimetric analyses
PM mass of the aerosol samples was determined by pre- and post-weighing the PTFE-filters (VI),
polycarbonate substrates (SDI) and polycarbonate films supported by aluminum foils (BLPI). In
Papers I and III the samples were weighed in Hyytiälä using a MT5 microbalance (Mettler-Toledo
Inc. Hightstown, NJ), whereas in Helsinki (Paper V) the weighing was performed by a M3
microbalance (Mettler Instrumente AG, Zurich, Switzerland). The samples collected at Finokalia
site were weighed at the University of Heraklion with a CA-27 microbalance (ATI Cahn, Analytical
Technology Inc., Boston, MA).
4.2.3. Chemical analyses
IC. Concentrations of ions were analyzed by an ion chromatography (IC). IC is the application of
liquid chromatography (LC) in which the separation of ions is based on the interactions between the
sample ions, stationary phase (analytical column) and mobile phase (eluent). Typically anions and
cations are analyzed separately because in anion analysis the eluent is basic and for cations the
eluent is acidic. After the separation of ions in an analytical column, the mobile phase is neutralized
by an ion suppressor in order to improve the sensitivity of the IC method. The detection of ions in
the IC is usually based on the conductivity of the eluent, but in the recent applications, also a mass
spectrometer has been coupled with the IC system (Fisseha et al., 2006). In this thesis the IC
instruments used were Dionex DX500 (Papers I -VI) and Dionex ICS (Paper VI). The ions analyzed
were methanesulfonate, Cl-, NO3-, SO4
2-, oxalate, Na+, NH4+, K+, Mg2+ and Ca2+. In Paper III also
acetate, propionate, formate pyruvate and HPO42- were determined. Methane sulfonic acid was the
eluent for cations, whereas NaOH was used for anions analysis, except in Paper III when
Na2CO3/NaHCO3 was used as the eluent for anions. Analytical columns used for anion analyses
were AS11 (Papers I–VI) AS4A-SC (Paper III) and AS17 (Paper VI), and for cations CS12A
(Paper I–II, IV–VI) and CS12-SC (Paper III). Prior to the IC analyses filters were extracted with 0.5
36
mL of methanol and 4.5 mL deionized water (Paper I, IV–V), 5 mL of water (Paper II, VI), or 20
mL of water (Paper III).
OC/EC analyzer. Organic and elemental carbon were determined using a thermal optical
transmittance (TOT) method. The instrument was a carbon analyzer developed by Sunset
Laboratory Inc., Oregon. For the analysis a 1–1.5 cm2 sample piece was punched from the quartz
filters/substrates. The thermal method had two phases to determine OC and EC. In the first phase
the sample was kept in a pure helium atmosphere and heated in four consecutive steps in order to
evaporate OC and carbonate carbon from the sample. In the second phase, the helium was replaced
by a mixture of oxygen and helium (1:49) and, similar to the first phase, the over temperature was
raised stepwise. In the second phase the OC remaining on the sample as well as EC was volatilized.
Part of OC is pyrolyzed into compounds resembling EC in the thermal analysis. Because of that an
optical correction, i.e. the measurement of laser light transmittance through the sample, was used in
order to separate charred OC from EC. Charring of OC decreases the transmittance of laser light, so
the charred OC was determined as the carbon evolved in the second phase before the transmittance
had reached its initial value. After this value the carbon evolved was considered as EC. The
temperatures and durations of the temperature steps varied slightly between the methods used in
Papers I–VI. The temperature program affects the split between OC and EC, and subsequently their
concentrations. The effect is greater on EC than on OC concentrations (Schauer et al., 2003;
Subramanian et al., 2006). The results of OC were corrected for the positive artifact, i.e. for the
adsorption of organic gases on the filter, by subtracting OC collected on the back-up filter from that
on the front filter. However, if denuders were used prior to the impactors, OC on the back-up filter
was added to OC on the front filter, since OC found on the back-up filter was assumed to result
from the negative artifact, i.e. evaporation from the front filter.
TOC analyzer. WSOC was analysed with a Total Organic Carbon Analyzer equipped with a high
sensitive catalyst (TOC-VCPH, Shimadzu). The method used was the Non-Purgeable Organic
Carbon (NPOC) method. In the NPOC method the sample solution is first drawn to syringe where
inorganic carbon (carbonates, hydrogen carbonates and dissolved carbon dioxide) is converted to
carbon dioxide, and subsequently evaporated from the sample, by adding HCl (1%) to the sample
and bubbling it with helium. After that sample is injected into an oven, where it is catalytically
37
oxidized to carbon dioxide at 680 °C and detected by non-dispersive infra red (NDIR) detector.
Prior to WSOC analysis samples were extracted by shaking the filter piece with 15 ml of Milli-Q
water for 15 minutes. Similar to the OC results, the concentrations of WSOC on the back-up filter
were subtracted from those on the front filter. TOC method is described in detail in Timonen et al.
(2008).
HPLC-MS. Monosaccharide anhydrides (MA) consist of levoglucosan and its two isomers
galactosan and mannosan. MA were analysed with a high performance liquid chromatograph
coupled with an ion trap mass spectrometer (HPLC-MS, Agilent Series 1100 LC/MSD Trap SL;
Agilent Technologies, Waldbronn, Germany). In HPLC-MS system the sample solution is injected
onto an HPLC column that consists of a narrow stainless steel tube packed with fine, chemically
modified silica particles. Compounds are separated on the basis of their relative interaction with the
chemical coating of the particles (stationary phase) and the solvent eluting through the column
(mobile phase). Components eluting from the HPLC column are introduced to the mass
spectrometer by an electrospray ionization. In the electrospray ionization the analyte solution passes
through the electrospray needle that has a high potential difference that forces a spraying of charged
droplets from the needle. Solvent evaporates from the droplets and the droplet shrinks until it
reaches the point at which the droplet is ripped apart. These charged molecules are analyzed on the
basis of their mass-to-charge ratio by ion trap mass spectrometer. For MA the monitored ion was
m/z 161. Prior to the HPLC-MS analysis MA were extracted with a 2-mL of a mixture of
tetrahydrofuran and water (1:1) (Papers II and VI) or with a mixture of 2 mL of methanol and 0.5
mL of water (Paper V) in an ultrasonic bath for 30 minutes. Basically the HPLC-MS method was
similar to that of Dye and Yttri (2005). However, in Papers II and VI the eluent was Milli-Q water
with a flow rate of 0.1 mL min-1 and two HPLC-columns (Atlantis, 4.6 × 150 mm, 3 µm, Waters)
were used at 7°C. In Paper V the HPLC-column was Zorbax® SB-C18 (4.6 × 250 mm, 5 µm,
Agilent Technologies) and the eluent methanol and deionized water (4:1) with a flow rate of 0.4 mL
min-1. In addition to MA, the HPLC-MS has been used for analyzing e.g. benzo[a]pyrene diones
(Koeber et al., 1999) and organic acids (Glasius et al., 1999; Anttila et al., 2005) from the aerosol
samples.
38
ICP-MS. Elements were analyzed only from the VI samples collected at Finokalia site (Paper III).
Al, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, Cd, Pb were determined by using an inductively coupled plasma-
mass spectrometry (ICP-MS; X-series, Thermo Electron Elemental Analysis, Winsford, UK). In the
ICP-MS samples are decomposed to neutral elements in a high temperature argon plasma and
subsequently analyzed on the basis of their mass to charge ratios. Before the ICP-MS analyses
aerosol samples were extracted by using acid digestion. Analytical methods used in this thesis are
summarized in Table 2.
Table 2. Off-line chemical analyses used in this study.
Chemical component Analytical instrument Company Paper
Ions IC Dionex I–VI
OC, EC OC/EC analyzer Sunset Laboratory Inc. I–VI
WSOC TOC Shimadzu II, VI
MA LC-MS Agilent Technologies II, V, VI
Elements ICP-MS Thermo Electron Elemental Analysis
III
4.3. On-line measurements
4.3.1. Physical properties
TEOM. PM2.5 mass concentration was measured with a Tapered Element Oscillating Microbalance
(TEOM© 1400a; Rupprecht & Patashnick; Allen et al., 1997; Patashnick, 1991). In the TEOM
particles are collected to a filter which is attached to the narrow end of a hollow, tapered glass stalk.
The glass stalk vibrates and the resonant frequency of the stalk decreases with the accumulated
mass deposited on the filter. The size range of the TEOM (PM2.5) was selected using a virtual
impactor prior to the TEOM and measuring PM from the VI major flow (15 L/min). The TEOM
was also equipped with a Filter Dynamics Measurement System (FDMS). The FDMS had a Sample
Equilibration System (SES) dryer which reduced relative humidity of the ambient air. The FDMS
enabled also the quantification of the volatilization losses in the PM concentration measurements.
After the SES the sample air was passed either directly to the TEOM or it went through a filter in
39
the FDMS in order to get a particle-free sample flow into the TEOM. During the latter period the
mass loss due to the volatilization of the collected particles was measured. For the TEOM a time-
resolution of 30 minutes was selected.
Eberline FH 62 I-R. In Paper III the PM10 mass was measured with an Eberline FH 62 I-R (Eberline
Instruments GmbH) Particulate Monitor, designed to measure the mass concentration of the
suspended particles in the ambient air based on β-attenuation (Gerasopoulos et al., 2006). The
Eberline FH 62 I-R collected particles on a glass fiber filter and simultaneously measured the PM
mass by the ability of particles to attenuate beta radiation. The air flow rate was 1 m3 h−1 and the
time-resolution was 5 minutes.
ELPI. In Paper I the PM1 and the mass size distribution were measured by using an Electrical Low
Pressure Impactor (ELPI; Outdoor Air ELPI, manufactured by Dekati Ltd, Tampere, Finland). In
the ELPI a multistage low pressure impactor classifies the aerosol sample into 12 size fractions.
Prior to the impactor stages the aerosol is charged in a corona charger and the charge brought by the
impacted particles is measured as a current by the electrometers connected to each impactor stage.
The size range of the ELPI was from 29 nm to 10 µm (aerodynamic diameter), but by adding an
electrical filter stage (ELA 650) the size range of the ELPI was extended to the particles smaller
than 29 nm. Prior to the ELPI, there was a conditioning unit (The Dekati Ambient Sampler, DAS
3100), in which the sample aerosol was first heated up to 30 °C and then led through a Nafion drier
(Permapure LCC, USA). The ELPI was used only in the electrical mode, and the impactor
substrates were not used for the gravimetric or chemical analysis. The time resolution of the ELPI
was one second but the data was averaged over ten minutes.
DMPS-APS. In Paper I particle number size distribution was measured with a Differential Mobility
Particle Sizer (DMPS) and a Aerodynamic Particle Sizer (APS). The DMPS system consists of a
differential mobility analyzer (DMA) and a condensation particle counter (CPC). Briefly, the DMA
classifies particles according to their electrical mobility. Each voltage used in the DMA allows a
certain narrow particle size range with same electrical mobility to penetrate the DMA. The number
of classified particles is then detected using a CPC. The CPC measures the number concentration of
particles by saturating the aerosol by water or alcohol vapor and cooling it to enlarge particles so
40
that they can be detected easily by optical methods. The DMPS system consisted of two parallel
DMPS devices: one classifying particles between 3 and 10 nm and the other between 10 and 500
nm (mobility diameter). Both instruments used a Hauke-type DMA (Winklmayr et al., 1991) and a
closed loop sheath flow arrangement (Jokinen and Mäkelä, 1997). The CPCs used were TSI Model
3025 and TSI Model 3010, respectively. For larger particles an APS (TSI Model 3320) was used. In
the APS particles are accelerated by the sheath airflow in the nozzle. Small particles (<0.3 µm) can
be kept up with the accelerating air, but larger particles accelerate more slowly than the air. The
larger or heavier the particles are the more they lag behind the air. From the magnitude of the lag
the particles aerodynamic diameter can be determined. The APS measured the number size
distribution from 0.5 µm to 20 µm (aerodynamic diameter). The time resolution of the DMPS and
the APS was ten minutes and one minute, respectively, but the APS data was averaged to ten
minutes.
The SDI, VI, APS and ELPI measured the particle aerodynamic diameter (Da) but the DMPS
measured the electrical mobility diameter (Db). The relation between these two quantities is (Hinds,
1999): 1/2
0
p1/2
ac
bc
b
a
ρρ
)(DC)(DC
DD
⎟⎟⎠
⎞⎜⎜⎝
⎛⎟⎟⎠
⎞⎜⎜⎝
⎛= , (8)
where Cc(Di) is a slip correction factor for a particle of diameter Di, ρ0 is the unit density (1 g/cm3)
and ρp is the particle density.
The DMPS and the APS measured the particle number size distribution nN(Da) which was
transformed to particle mass size distribution nM(Da) using the following equation (Seinfeld and
Pandis, 1998):
)(Dn)(DD6πρ)(Dn aN
3abpaM = . (9)
Here nN(Da) is given in units particles/cm3 and nM(Da) in units µg/m3. In Paper I the particle density
of 1.5 g/cm3 was used to convert the particle number size distribution to the mass size distribution.
41
4.3.2. Chemical composition
PILS-IC. Ion composition of the aerosol particles was determined continuously using a Particle-
into-Liquid Sampler (PILS; Orsini et al., 2003) combined with two ICs (Dionex 2000). In the PILS
particles were collected by growing them in a supersaturated steam atmosphere, after which they
were impacted in to the liquid phase. From the liquid phase the ions were analyzed by two IC
systems, one for cations and one for anions. Similar to the TEOM system (Sect. 4.3.1.), a virtual
impactor with a cut-off at 1.0 µm was placed prior to the PILS in order to remove particles larger
than 1.0 µm from the sample air. The VI major flow (15 L/m), including only particles smaller than
1 µm in diameter, was fed into the PILS. Annular denuders (URG) in series upstream of the PILS
were used to remove acidic gases and ammonia from the sample air. Two denuders were coated
with a KOH (1%) solution and one denuder was coated with a H3PO4 (3%) solution. Cation analysis
(sodium, ammonium, potassium, magnesium, calcium) were made using 4mm CG12A/CS12A
columns with an electrochemical suppressor (CSRS ULTRA II 4 mm). Anion analyses (chloride,
nitrate, sulfate, oxalate) were made using 4 mm AG11/AS11 columns with an electrochemical
suppression (ASRS ULTRA II 4 mm). The sample loop volume was 1 mL in both ion
chromatographs. Lithium fluoride was used as an internal standard. For the used sample loops and
sample flow rate, the detection limits (DLs) for the measured ions were 0.005-0.010 µg m-3,
depending on the analyzed ion. Time resolution for the samples was 15 minutes. The PILS-IC
system has been described in detail by Kuokka et al. (2007).
Semicontinuous OC/EC analyzer. OC and EC concentrations were measured online using a
semicontinuous OCEC Carbon Aerosol Analyzer (Sunset Laboratory Inc., Portland, OR). The
semicontinuous OC/EC analyzer resembled the carbon analyzer designed for laboratory analyses
(Sect. 4.2.3.) but, in addition to that analyzer, it included also the aerosol sampling system. In the
semicontinuous OC/EC analyzer particles were collected on two quartz filters located in the front
oven. After the sampling the temperature of the oven was increased similar to the laboratory
analyzer in order to evaporate carbonaceous components. As in the laboratory instrument, the
temperature program consisted of a helium phase and helium-oxygen phase, but the number of
temperature steps was smaller in order to keep the duration of the thermal-optical analysis short. By
using a tuned diode laser (red 660 nm) the charred OC was separated from EC. Carbon components
evaporated from the sample were oxidized to carbon dioxide in the MnO2 oven similarly to the
42
laboratory model of the OC/EC analyzer. However, in the semicontinuous OC/EC analyzer carbon
dioxide was detected by a NDIR detector. Because of that, the methanator, used in the laboratory
carbon analyzer to reduce carbon dioxide to methane, was not needed in the semicontinuous OC/EC
analyzer. Similarly, as in the laboratory carbon analyzer, a known amount of methane gas was
injected in each analysis cycle as an external standard. In addition to thermally determined OC and
EC, described above, the semicontinuous OC/EC analyzer measured also optical EC with the laser
(660 nm).
In the semicontinuous OC/EC analyzer organic gas-phase components were removed from sample
air by a parallel plate carbon filter denuder (Sunset Laboratory Inc., Portland, OR). The instrument
had originally a sample flow rate of 5 L min-1, but it was raised to 9.2 L min-1 in order to increase
the sensitivity of the measurements. Also the original cyclone with a cut-off at 2.5 µm in particle
aerodynamic diameter was changed into a cyclone with a cut-off at 1 µm at the flow rate of 9.2 L
min-1. The time-resolution of the semicontinuous OC/EC analyzer was three hours with the
measurements starting every day at 0:00, 3:17, 6:00, 9:00, 12:00, 15:00, 18:00 and 21:00 at local
time.
Aethalometer. Black carbon was measured with a 5-minute time resolution using a single
wavelength aethalometer (AE-16, Magee Scientific, wavelength 880 µm; Hansen et al., 1984). The
aethalometer is an instrument that uses optical absorption to determine the mass concentration of
black carbon particles collected on a filter. The aethalometer was equipped with a cyclone that
removed particles larger than 2.5 µm in aerodynamic diameter. Black carbon equivalent
concentrations were calculated from absorption measurements of the aethalometer using a mass
absorption efficiency of 16.6 m2g-1.
Aerosol mass spectrometer. In Po Valley the chemical composition of particles was measured by
using a High-Resolution Time-of-Flight aerosol mass spectrometer (HR-ToF-AMS, Aerodyne
Research Inc., USA). The AMS had three main sections: an aerosol inlet, particle sizing chamber
and particle detection section. The aerosol inlet sampled sub-micron particles into the AMS through
an aerodynamic lens forming a narrow particle beam. The beam was transmitted into the detection
chamber, in which non-refractory components of aerosol were flash vaporized upon impact on hot
43
surface (~560 °C) under high vacuum (~10-5 Pa). After that the components were ionized by
electron impact ionization and the ions were detected by a mass spectrometer. The transmission of
the particle beam to the detector was modulated with a mechanical chopper. The chopper had three
positions. An “open” position transmitted the beam continuously, “closed” position blocked the
beam completely and “chopped” position modulated the beam transmission with a 1–4% duty cycle
which was determined by the width of the slit in chopper.
The AMS alternated between two modes of operation: mass spectrum and particle-Time-of-Flight
(pToF) mode. In the MS mode the chopper was in open position to obtain an ensemble-average MS
of the sampled air. Signal from the background gases was accounted for by subtracting the MS
obtained with the chopper in closed position. The particle size was determined in the pToF mode.
When the chopper was operating in the chopped mode, the particle velocity was measured from its
flight time between a chopper and the vaporizer surface. The HR-ToF-AMS included also ion
optics for two modes of operation: V- and W-mode. In the V-mode ions followed the traditional
reflection path, whereas in the W-mode the ions exiting the reflector were directed into a hard
mirror that focused them back into the reflector for a second time before travelling to a
multichannel plate detector. The mass resolving power of the ToFMS increased as the flight path
was lengthened, but the lateral broadening of the ions increased over a longer flight path and
reduced the total signal as fewer ions struck the detector. Therefore the V-mode was more
sensititive, but the W-mode offered a higher mass resolution. The resolutions for the V-mode and
W-mode are typically ~2000 and 4000, respectively (DeCarlo et al., 2006). In this thesis only the
data from the V-mode was used. The time-resolution for the AMS was 15 min. Note also that the
AMS results shown in this thesis are preliminary and therefore they are discussed only qualitatively.
All on-line instruments used in this study are summarized in Table 3.
44
Table 3. On-line instruments used in this study.
Instrument Measured aerosol property/chemical components
Company Paper
TEOM© 1400a PM2.5 Rupprecht & Patashnick II
Eberline FH 62 I-R PM10 Eberline Instruments GmbH III
Outdoor Air ELPI Size-segregated mass conc. Dekati Ltd I
DMPS (DMA: Hauke type CPCs: TSI 3025 & 3010)
Size-segregated number conc. CPCs: TSI Inc. I
APS (TSI 3320) Size-segregated number conc. TSI Inc. I
PILS-IC (ICs: Dionex 2000)
Cl-, NO3-, SO4
2-, Ox2-, Na+, NH4
+, K+Methromn (PILS), Dionex (IC)
II, VI
Semicontinuous OC/EC analyzer
OC, EC Sunset Laboratory Inc. VI
Aethalometer AE-16 Black carbon Magee Scientific II
HR-ToF-AMS Cl-, NO3-, SO4
2-, NH4+, Org. Aerodyne Research Inc. -
4.4. Comparison of off-line and on-line measurements
4.4.1. Detection limits
The detection limit (DL) is the lowest concentration that can be determined to be statistically
different from the background. The DLs for aerosol chemical components (ions, OC and EC)
measured by off-line and on-line methods are given in Table 4. The way to calculate DLs depends
on the analytical technique used. For the ion chromatography methods (VI, PM1 and PILS-IC), the
DLs were determined as the concentration that gave a peak (signal) three times the peak noise level
in ion chromatogram, whereas the detection limits for the AMS were obtained from the paper of
DeCarlo et al. (2006). In case of OC and EC, the detection limits for the VI and PM1 were
calculated from the DL of 0.15 µgC per cm-2 of filter given in Birch and Cary (1996) and for the
semicontinuous OC/EC the DLs were calculated by using the minimum quantifiable level of 500
ngC m-3 for 30-min sample provided by the manufacturer.
45
The DLs for the off-line measurements were generally much smaller than those for the on-line
methods. Since the DLs related to analytical instruments were similar for off-line and on-line
methods, the difference in the DLs arose from the amount of sample collected (duration of sampling
and flow rate) as well as the fraction of sample used for the analysis. By comparing the on-line
methods the AMS had lower DLs for nitrate and sulfate than the PILS-IC but the PILS-IC was more
sensitive to chloride and ammonium than the AMS. The semicontinuous OC/EC had a slightly
lower DL for OC than the VI but much higher than the AMS. Furthermore, instead of OC, the AMS
measured POM, which is roughly 1.4–2.2 times OC.
Another way to determine the detection limits is to calculate the concentration that gives a signal
equal to three times the standard deviation (SD) of the filter blanks. In Table 4 the DLs calculated
from the blanks are shown in parenthesis. The filter material for the VI was PTFE, except for OC
and EC for which it was quartz. In the PM1 only quartz filters were used. For the PTFE filters the
DLs from the blanks were reasonable, but for quartz the DLs were particularly large for sodium and
calcium compared to the concentrations measured typically in ambient air. Also for sulfate the DL
using quartz was large however it was much smaller than the ambient concentrations.
46
Table 4. Detection limits for the ions, OC and EC measured using the off-line and on-line instruments (ng m-3). The time-resolution was 24 hours for the VI and PM1, 15 minutes for the PILS-IC and AMS and 3 hours for the semicontinuous OC/EC.
Off-line methods On-line methods Component
VI PM1 PILS-IC Semicontinuous-OC/EC AMS
Chloride 0.093 (0.73) 0.043 (0.67) 4 12
Nitrate 0.37 0.17 (2.1) 16 2.9
Sulfate 0.28 0.13 (27) 12 5.2
Oxalate 0.46 0.21 (2.7) 20
Sodium 0.038 (0.78) 0.018 (410) 1.6
Ammonium 0.075 (0.67) 0.035 3.2 38
Potassium 0.10 (0.36) 0.046 (0.67) 4.3
Magnesium 0.040 (1.2) 0.019 (3.3) 1.7
Calcium 0.14 (5.6) 0.066 (35) 6.2
Organic carbon 96 (130) 13 (77) 83 (290) 22
Elemental carbon 96 13 83
4.4.2. Correlations
The correlations between off-line and on-line measurements for ions, OC and EC are shown in
Table 5. For the ions the PILS-IC data, and for OC and EC the semicontinuous OC/EC analyzer
data, were compared with data from the PM1 samples. For potassium, sulfate and nitrate the
concentrations from the PILS-IC correlated well with those obtained from the filter samples,
however, for nitrate the concentrations from the PILS-IC were 149% of those from the PM1. Larger
concentrations for the PILS-IC suggest that nitrate might evaporate from the filter samples
(Pakkanen and Hillamo, 2002; Schaap et al., 2004). The smallest correlation coefficient was
obtained for oxalate together with the smallest slope. This is indicative of the decomposition of
oxalate in the PILS-IC but needs to be confirmed by further laboratory tests. For OC and EC the
correlation between PM1 and semicontinuous OC/EC was good (Table 5). Additionally, the
difference between semicontinuous OC/EC and the PM1 was only 5 and 4% for OC and EC,
respectively.
47
Table 5. Correlation between the PM1 filter samples and the PILS-IC (ions), and between the PM1 filter samples and the semicontinuous OC/EC (OC and EC). The intercepts and correlation coefficients were calculated from the least-square plot. The results from the PILS-IC and semicontinuous OC/EC have been averaged to 24-hour periods corresponding the sampling periods of the PM1 sampling.
Component R2 Slope Intercept Number of samples
Nitrate 0.81 1.49 0.28 135
Sulfate 0.83 0.96 -0.11 135
Oxalate 0.38 0.28 0.0025 127
Ammonium 0.69 0.86 0.22 135
Potassium 0.84 0.79 -0.0015 120
Organic carbon 0.95 1.05 0.030 118
Elemental carbon 0.90 1.04 0.060 118
In addition to aerosol chemical components, the PM1 mass concentrations measured with the
impactors were compared with those obtained from the instruments measuring aerosol continuously
(Paper I). The PM1 concentrations were calculated for the VI, SDI, ELPI and DMPS-APS
combination, of which the VI was selected to serve as a reference against which the other
instruments were compared. For the VI the fine particle concentration (Da < 1.3 µm) was used as a
PM1 concentration, while for the SDI the seven lowest stages were summed up (Da < 1.06 µm) and
for the ELPI the filter stage plus seven lowest stages were summed up (Da < 0.963 μm) to obtain
the PM1 mass concentration. For the DMPS and APS, the mass concentration (Da < 1.3 μm) was
calculated from the number concentration by using equations 8 and 9.
Measured PM1 concentrations are presented in Figure 5a. From the figure it can be seen that all the
instruments followed a similar time-evolution with only minor differences. The concentration of the
SDI was smaller than that of the VI (SDI-to-VI ratio of 0.70), being indicative of bouncing and
interstage losses of particles in the SDI. In contrast, the ELPI to VI ratio was above unity (1.11)
perhaps because the ELPI measured also semivolatile compounds which might have been lost in the
impactors. The PM1 concentration of the DMPS-APS was slightly smaller than that of the VI
48
(0.92). In addition to PM1, the mass size distributions from the ELPI, DMPS and APS were
compared with those measured by the SDI. An example of the mass size distributions is shown in
Figure 5b. The mass size distributions of the DMPS-APS and ELPI were quite similar, with one
accumulation mode peaking at nearly the same size (~400 µm), but the mass size distribution of the
SDI was slightly different. The SDI had two modes in the submicron size range, the dominant one
peaking at around 0.5 µm and the minor mode peaking at around 0.15 µm. The dominant mode of
the SDI was centered at larger particle size than that of the DMPS-APS and ELPI, but the mode was
clearly smaller in magnitude than that of the DMPS-APS and ELPI. This shift may have been
caused by the difference in the relative humidity during the sampling, since the SDI mass size
distribution was measured at ambient relative humidity while for the ELPI and DMPS the drier was
used. The minor mode of the SDI (at 0.15 µm) could result from the large uncertainties associated
with the weighing of small mass concentrations. In addition, it cannot be ruled out that semivolatile
mass was lost in the SDI low pressure stages (below 0.5 µm) during long sampling durations. In the
coarse size range the difference between the ELPI, DMPS-APS and SDI was even larger than in the
fine size range.
Date
07/05
/04
09/05
/04
11/05
/04
13/05
/04
15/05
/04
17/05
/04
19/05
/04
21/05
/04
23/05
/04
25/05
/04
27/05
/04
29/05
/04
31/05
/04
µg/m
3
0
5
10
15
20
Da (µm)
0.01 0.1 1 10
dm/d
logD
p (µg
/m3 )
0
10
20
30
40
50b) mass size distributiona) PM1 concentration
ELPI VI
SDIDMPS-APS
ELPIDMPS-APSSDI
Figure 5. PM1 concentrations measured with the ELPI, VI, SDI and DMPS-APS (a) and the mass size distributions obtained by the ELPI, DMPS-APS and SDI (b) in Hyytiälä (Paper I).
49
4.5. Laboratory calibration set-up
Particle generation. The SDI was calibrated for the quartz filter substrates in Paper IV. Calibrations
were performed using monodisperse dioctyl sebacate particles (DOS, density 0.914 g/cm3). The
primary aerosol consisting of submicrometer particles (0.01–1 µm) was generated with a Constant
Output Atomizer (TSI Model 3068, St. Paul, Minnesota, USA) from 0.002% to 2% solutions of
DOS in 2-propanol. The aerosol obtained from the atomizer was transported into an evaporation-
condensation aerosol generator in order to get a narrow particle number size distribution (Liu and
Lee, 1975). After that a desired monodisperse particle size fraction was selected using a DMA (TSI
Model 3080, St. Paul, Minnesota, USA).
Detection methods. The collection efficiency curves were measured in two ways, using either
electric detection or two CPCs. The electric measurement method has been described in detail by
Hillamo and Kauppinen (1991) and Maenhaut et al. (1996). Briefly, charged monodisperse test
particles from the DMA were fed into the impactor stage and the current carried by these particles
was measured simultaneously from the impaction plate and back-up filter using electrometers
(Keithley 617 Programmable Electrometer, Cleveland, OH, USA). The quartz filter was attached to
the metal impaction plate with a conductive tape and the impaction plate was electrically isolated
from the rest of the impactor body by a plate made of PTFE. The collection efficiency was
calculated as the ratio of impacted portion to the sum of impacted and penetrated portions.
In the second calibration system the detection of the particles was made using two CPCs (TSI
Model 3010), one before and another after an impactor stage. The test aerosols were generated
similarly to the electrical detection, but they were neutralized by using a Kr-85 radioactive source
before the CPC detection. The concentrations were measured from the CPC directly or on the basis
of detected pulses of which the latter was made using a computer. Before the calibration, a
measurement was made without an impaction plate in order to determine the factor that took into
account the differences in the flow rate and absolute pressure upstream and downstream of the stage
as well as possible losses in the CPC sampling lines and inside the impactor. The base line was
measured by filtering the aerosols off, but also the small systematic difference between the two
CPCs was measured on a daily-basis by measuring parallel with two instruments. The collection
efficiency was calculated from the concentration data by assuming that the difference between the
50
upstream and downstream number concentrations was due to particle impaction. As a result, the
collection efficiency was equal to the ratio of this difference to the upstream concentration.
4.6. Data analysis
4.6.1. Data inversion (MICRON)
For the SDI samples collected in Hyytiälä (Paper I), the raw concentration data obtained for each
impactor stage were run through the inversion code MICRON (Wolfenbarger and Seinfeld, 1990) in
order to obtain continuous mass size distribution (see Figure 5b for the SDI). By fitting a lognormal
curve to the inverted data (Winklmayr et al., 1990), the mode parameters, e.g. mass median
aerodynamic diameter, mode mass concentration, were obtained. The inversion procedure requires
information on the uncertainties related to the analyses as well as the collection efficiency
characteristics of the impactor. Because of that the inversion code is different for the aluminium
foils, used for measuring the PM mass and ions, and for the quartz substrates, used for OC and EC.
For aluminium foils the collection efficiency curves are well-known and subsequent inversion
reliable. In Paper IV the collection efficiency curves were achieved also for the quartz substrates
(stages 1–8). However, the subsequent data inversion by MICRON was unsuccessful. The reason
why the data inversion failed was not clear, but it might be because of the overlapping collection
efficiency curves.
4.6.2. Receptor models
Results from the impactor samples were analyzed by the receptor methods in order to find out (1)
common sources and/or (2) similar generation and/or removal mechanisms and (3) transport
patterns of chemical components. The general receptor model assumes that there are p sources.
Source types or source regions (factors) impact a receptor, and linear combinations of the impacts
from the p factors generate the observed concentrations of the various species. Mathematically it
can be expressed as
ij
p
kkjikij efgx += ∑
=1, (10)
51
where xij is the concentration of species j in sample i, gik is the contribution of the kth factor to
sample i, fkj is the fraction of species j in the kth factor, and eij is the residual for the species j in
sample i.
Varimax-rotated Factor Analysis. Varimax-rotated Factor Analysis was applied to the samples
collected in Finokalia (Paper III). Data set consisted of 92 samples with 20 variables, including 12
ions, OC, EC and 11 elements. Five factors were obtained for the fine particles and four factors for
the coarse particles. These factors explained 81.9 and 80.6% of the total system variance for the fine
and coarse particles, respectively.
PMF. In Paper VI the sources of organic carbon were investigated by using the PMF (EPA PMF
1.1). For the PMF only the concentrations of species (xij’s) need to be known and the target is to
estimate the contributions (gik) and the fractions (or profiles) (fkj). It is assumed that the
contributions and the fractions are all non-negative, hence the “constrained” part of the least
squares. EPA PMF also allows one to take into account how much uncertainty there is in each xij.
The samples with a lot of uncertainty are not allowed to influence the estimation of the
contributions and profile as much as those with a small uncertainty, hence the “weighed” part of the
least squares.
EPA PMF minimizes the sum of squares (Q) using the equation:
2
1 1
1∑∑∑
= =
=
⎟⎟
⎠
⎞
⎜⎜
⎝
⎛ −=
n
i
m
jij
p
k kjikij
sfgx
Q, (11)
where sij is the uncertainty in the species j in sample i.
Data consisted of 230 samples and 11 variables because the two major biomass burning episodes
were excluded from the data set. Variables were OC, EC, WSOC, levoglucosan, ammonium,
potassium, sulfate, oxalate and gaseous ozone, nitrogen monoxide and nitrogen dioxide. PMF was
run with the factor numbers of 3–6 giving the best fit for four factors. The factor solution explained
52
88–100% of the variation of each component, with the lowest percentage obtained for ozone. The
uncertainties loaded for the PMF analysis were in range 5–15%, the lowest percentage being
estimated for sulfate and the largest for WSOC and levoglucosan.
53
5. RESULTS
5.1. Chemical mass closure
In this thesis the chemical mass closure was examined from different perspectives. The mass
closure was investigated at urban, background and remote sites. Year-round measurements made it
possible to study the seasonal variation, but the mass closure was also studied during some special
cases, such as nucleation events, biomass smoke and dust episodes, that took place during the
measurement periods. Those results were obtained mostly by using impactors with the sampling
times ranging from one day to a few days. By using on-line measurements the time-resolution of the
chemical mass closure was improved significantly, which allowed to investigate variations during a
single day. Both impactors and online instruments were utilized to construct the particle size-
resolved chemical mass closure. In Section 5.1.4. the focus of the chemical mass closure has been
moved from the total particulate mass to the carbonaceous fraction of the particles.
5.1.1. Urban, background and remote sites
An average chemical composition of fine particles at an urban (Helsinki), background (Hyytiälä)
and remote site (Finokalia) is presented in Table 6. In Helsinki the majority of the PM consisted of
POM (~40%). Of the water-soluble ions, the dominant compound was sulfate with an average
contribution of 20% to the PM. The contributions of the other major water-soluble ions, ammonium
and nitrate, were significantly smaller than that of sulfate. Also in Hyytiälä most of the PM was
made of POM (~48%) and sulfate (~30%). In Finokalia the chemical composition of the fine
particles was different from the Finnish sites, the contribution of water-soluble ions being higher
than that of POM with large fractions of sulfate (37%) and ammonium (12%). The most significant
difference between the sites was in the contribution of EC to PM which was clearly higher in
Helsinki (7.6%) than in Hyytiälä (3.9%) or Finokalia (2.2%). Trace elements were determined only
from the Finokalia samples. In Finokalia the analyzed trace elements constituted 4% of the PM1,
whereas in an earlier study conducted in Helsinki the total contribution of elements to PM2.5 was
found to be around 2% (Pakkanen et al., 2001).
54
Table 6. Concentrations of the major chemical components in fine particles (µg m-3) and the chemical mass closure. The conversion factor of POM (see Section 3.2.4.) varied from 1.4 to 1.8.
Site PM1 POMfactor EC Water-soluble ions
Elements Mass closure
Paper
Helsinki 21a 8.41.6 1.6 4.8 n.a. 0.71 II
Hyytiälä 4.4 2.11.4 0.17 1.8 n.a. 0.98 I
Finokalia 12 3.31.8 0.27 6.2 0.47 0.89 III aPM2.5, measured by the TEOM n.a. not analyzed
In Hyytiälä and Finokalia most of the fine particulate matter was explained by the analytical
methods used, as seen from the high percentages of the chemical mass closure (Table 6). For
Helsinki samples the agreement between chemically analyzed PM and the mass measured by the
TEOM was poorer. This was partly due to the difference in the cut-off sizes for the particle mass
(Da < 2.5 µm) and for the filter samples (Da < 1 µm), from which the chemical composition of
particles was determined. The factor for converting OC to POM was different for all the sites
ranging from 1.4 in Hyytiälä to 1.8 in Finokalia.
The chemical composition of particles presented in Table 6 is an average over measurement periods
ranging from 24 days (Hyytiälä) to 2 years (Finokalia). However, the chemical composition
changed from season to season. The seasonal variation was most evident in Finokalia (Paper III). In
summer the contribution of secondary ions, sulfate and ammonium, to PM was larger than in
winter, whereas in winter the contribution of POM was slightly larger than in summer. In Helsinki
the chemical composition was more stable over the course of the year than in Finokalia. However,
there was an observable diurnal variation in Helsinki, which will be discussed in Section 5.1.2.
Overall, the seasonal or diurnal changes in the chemical composition of particles were small
compared to those resulting from the long-range transported emissions.
In spring and summer, 2006, there were several periods when dense smoke plumes reached
Helsinki from forest or wild land fires in Russia. During those smoke episodes the PM2.5
concentration raised significantly, but also the chemical composition of the particles changed
(Figure 6; Paper II). The contribution of POM to PM2.5 increased up to 65% during the episode
55
while that of ions decreased. For EC the difference between the episodic and non-episodic period
was not notable, and also the extent of the chemical mass closure did not change. The smoke
episodes detected in spring 2006 in Helsinki are characterized in detail in Paper II.
Mar-06 Apr-06 May-06 Jun-06 Jul-06 Aug-06 Sep-06 Oct-06
µg m
-3
0
10
20
30
40
50
60
ECPOMNO3
-
NH4+
SO42-
PM2.5
Figure 6. Chemical composition of PM2.5 from March 2006 to September 2006 at SMEAR III in Helsinki. POM was obtained by multiplying OC by the factor of 1.6.
In Finokalia the PM was frequently influenced by the episodes of Sahara dust. In addition to the
higher PM concentration, also the chemical composition of PM changed during the dust episodes.
The contributions of POM, ions and EC to PM1 decreased slightly while those of elements
increased when the dust plume arrived at Finokalia. However, more than on the PM1 concentration
the dust episodes influenced on coarse particles. The dust episode elevated PM1–10 to ten times
larger than in non-dust periods, on average. The time-evolution of the chemical composition of
PM1–10 was similar to that of PM1, with a lower fraction of POM and ions but a higher fraction of
elements. The chemical mass closure decreased during the dust episodes. The mass closure for PM1
was only 0.50 during the dust episodes, but the difference between the dust and non-dust periods
was more pronounced for PM1–10. An average mass closure for PM1–10 in non-dust period was 0.78,
decreased to 0.24 during the dust episodes. More about the differences in the chemical composition
of fine and coarse particles are discussed in Section 5.1.3.
56
In Hyytiälä, the beginning of the measurement campaign was dominated by the polluted air from
south-east, after which the air masses came mostly from the clean areas in north and west (Paper I).
During the polluted air masses PM1, POM, ion and EC concentrations all increased. However, the
contributions of the chemical components were quite similar during the whole campaign, although
the air masses came from the different areas and the concentration levels varied significantly. The
chemical mass closure was slightly lower during the high-concentration period. More than air
pollution episodes, Hyytiälä is known as a place where aerosol formation and subsequent particle
growth has been observed regularly (Mäkelä et al., 1997; Kulmala et al., 2001). Several nucleation
events were detected during the campaign in spring 2004 (Paper I) but no clear differences between
the chemical composition of the event and non-event samples were observed. However, the
previous studies conducted in Hyytiälä have shown that PM organic composition changes from the
non-event to event periods. During the clean nucleation events WSOC has been dominated by
aliphatic biogenic species, whereas the non-events periods have had anthropogenic oxygenated
species (Cavalli et al., 2006). Additionally, Mäkelä et al. (2001) found that the concentration of
dimethylamine was larger during the particle formation and/or particle growth periods.
5.1.2. Time-resolved mass closure
Figure 7 shows the average chemical composition of fine particles and chemical mass closure in
Helsinki with a time-resolution of 3-hours. The chemical mass closure ranged from 0.64 to 0.75
with a high standard deviation associated with all the measurement periods. Similar to the filter
measurement performed in Helsinki (Table 6), the on-line the chemical composition (OC, EC,
major ions) was determined from the particles below 1 µm while the PM obtained from the TEOM
had a cut-off at 2.5 µm.
57
Local time0 3 6 9 12 15 18 21 24
Frac
tion
0.2
0.4
0.6
0.8
1.2
0.0
1.0NO3
-
SO42-
NH4+
POM
EC
Mass closure
Figure 7. Chemical composition of particles and the mass closure measured with a time-resolution of 3 hours at SMEAR III in Helsinki from June to November 2006 (average ± SD). PM2.5 was determined by using the TEOM, ions by the PILS-IC and OC and EC by the semicontinuous OC/EC. POM was obtained by multiplying OC by the factor of 1.6. The diurnal variation was clearest for EC. The contribution of EC to PM2.5 was largest during the
morning rush hour from 06:00 to 09:00 (13%), after which it decreased until the evening rush from
15:00 to 18:00. The lowest EC contribution was measured from 00:00 to 03:00, being equal to
6.8%. For EC there was also a difference between weekdays and weekends (Paper VI). Besides EC,
nitrate displayed a diurnal trend. The contribution of nitrate to PM2.5 was lowest in the afternoon
from 12:00 to 03:00, indicating the transfer of semivolatile nitrate from particle phase to gas phase
when the temperature increased. However, for the contribution of ammonium there was no clear
diurnal pattern, which can be explained by a higher fraction of ammonium associated with sulfate
than with nitrate.
In contrast to the small diurnal variation observed in Helsinki, the chemical composition in Po
Valley changed more dramatically (Figure 8). In Po Valley the nitrate and ammonium
concentrations were typically high in the morning and decreased around midday. A similar diurnal
pattern was also observed for organic components, but the change between morning and evening
was smaller than for nitrate and ammonium. For sulfate there was no clear diurnal variation. The
58
diurnal pattern for nitrate, ammonium and POM can be explained mostly by the boundary layer
height. The low boundary layer together with the local or regional emissions (traffic, agriculture)
caused the high concentrations in the morning. The sharp drop around noon was due to the sudden
increase in the boundary layer height, inducing an efficient mixing of emissions. On the other hand,
also the wind direction changed often at that time of the day, so the different source areas cannot be
excluded from discussion before the trajectory analysis. Most of sulfate was probably transported to
Po Valley region and to the site from further away, since sulfate concentration was not affected by
the boundary layer height. There were no continuous PM2.5 or PM1 measurements in Po Valley so
the chemical mass closure could not be constructed.
01-04-08 02-04-08 03-04-08 04-04-08 05-04-08
µg m
-3
0
5
10
15
20
OrganicsNO3
-
NH4+
SO42-
Figure 8. Composition of the fine particles in Po Valley, Italy, 1–5 April, 2008 with a time-resolution of 10-min. Measurements were performed with the AMS.
5.1.3. Size-resolved mass closure
In addition to spatial and temporal variation, the chemical composition varied with the particle size.
Figure 9 illustrates the distribution of the chemical components between two size fractions in
Finokalia: fine and coarse particles. There were components of which 90% were in fine particles
(ammonium) and components which were primary in coarse particle (sea salt; sodium and chloride)
(Paper III). Regarding the major components in particles on mass basis, 80, 62 and 67% of sulfate,
OC and EC, were in fine mode, respectively.
59
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
NH4+
pyruv
ateMS-
SO42-
Cd EC K+OC Pb V
oxala
te Cu
HPO42- Ni Ti Zn Cr Al
Mn Ca FeNO3-
Ca2+ Cl-
Na+
Con
tribu
tion
Coarse
Fine
Figure 9. Contribution of trace metals, water-soluble ions, OC and EC to the fine and coarse fractions in Finokalia (Paper III). A more detailed size resolved mass closure study was conducted in Hyytiälä (Paper I). By using
two SDIs in parallel, both gravimetric and chemical mass were obtained. However, by replacing
the thin film substrates (e.g. made of aluminum or polycarbonate) normally used in multi-stage
impactors by thick and porous quartz substrates, the collection characteristics of impactor stages
changed. Because of that, the SDI needed to be recalibrated for the quartz substrates (Paper IV).
By using the quartz substrates the impactor stage cut-off sizes (D50 values) shifted to smaller
particle sizes and the shapes of the collection efficiency curves changed. This can be explained
by the collection mechanisms other than inertial impaction, e.g. filtration, because the high
velocity jets penetrate the porous quartz filter surface. The change in the collection efficiency
was most severe for small particles (lower stages). In Figure 10a the collection efficiency curves
are shown for a plate (or thin film) and for the quartz substrates. Figure 10b illustrates the
difference in the mass size distributions of OC calculated by using the new and the original D50
values.
60
a) Stage 4
Da (µm)
0.1 1
effic
ienc
y
0.00
0.25
0.50
0.75
1.00
Da (µm)
0.01 0.1 1 10
dm/d
logD
a, µg
m-3
0.0
0.5
1.0
1.5
b) Mass size distribution
quartzplate
new D50 valuesoriginal D50 values
Figure 10. Collection efficiency curves for stage 4 measured using a plate and a quartz fiber substrate (a) and the effect of new D50 values for the mass size distribution of OC measured at SMEAR II in Hyytiälä 26–28 May 2004 (Paper IV). An example of the size-segregated chemical mass closure in Hyytiälä is represented in Figure 11. It
was constructed by subtracting the analyzed inorganic ions from the gravimetric mass, and by
comparing the remaining mass with the measured OC and EC concentrations calculated using the
D50 values obtained for the quartz substrates (Paper IV). The white area between OC and ions
(unanalyzed) consists mainly of oxygen and hydrogen atoms in organic compounds not analyzed
(the difference between OC and POM). The accumulation mode of total carbon (TC; OC+EC) fitted
well to the area that remained when the inorganic ions were subtracted from mass. However, the
right side of the accumulation mode was steeper for unanalyzed mass than for TC. Also, in the
particle size range of 0.02–0.1 µm there was a mode for OC that was not found for mass. That mode
might be due to gaseous OC evaporated from the particles collected in the upper stages of the SDI
and subsequently collected on stages 1 and 2 by adsorption. However, it shows that there are more
organic compounds in small particles compared to sulfate and ammonium. Nitrate was found only
in the coarse mode.
61
HyytiäläMay 7-9 2004
Da, µm0.01 0.1 1 10
dm/d
logD
a, µg
/m3
0
5
10
15
20
NO3-
NH4+
SO42-
UnanalyzedOCEC
Mass
Figure 11. The size-segregated chemical composition of particles at SMEAR II in Hyytiälä (Paper IV).
Besides with multistage impactors, the size-resolved chemical composition can be measured by the
AMS. An example of the size-distribution measured in Po Valley, is shown in Figure 12. Compared
with the chemical composition of particles in Hyytiälä (Figure 10), there were much larger
contributions of nitrate and ammonium in Po Valley while the fraction of sulfate was much smaller.
The size distribution of POM was quite similar between Po Valley and Hyytiälä, demonstrating that
there were significant amount of organics at particle size below 150 nm also in Po Valley. The
AMS could not determine the total mass of particles because it was not able to measure EC, some
ions (e.g. sodium, potassium) and elements. However, compared to the TOT analysis made for the
impactor samples, the AMS measured directly the mass of POM and the uncertainty associated with
the conversion of OC to POM could be avoided.
62
Po Valley 15-16 April, 2008
Dvac, nm
10 100 1000
dm/d
logD
va, µ
g m
-3
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
NO3NH4 SO4 Org
Figure 12. Size-resolved chemical composition of fine particles in Po Valley, Italy, measured with the AMS.
5.1.4. Carbonaceous fraction
Carbonaceous matter in particles can be separated into organic and elemental fractions. In Helsinki
30% of total carbon was made of EC on average, whereas in Hyytiälä and Finokalia the fraction of
EC was lower being equal to 14% at both sites. In Helsinki the split between OC and EC had no
seasonal variation, whereas in Hyytiälä the ratio of OC to EC increased significantly during summer
(Figure 13). This indicates a prominent formation of secondary organic aerosol in Hyytiälä in
summer. During the biomass burning episodes observed in Helsinki (Section 5.1.1.), the fraction of
EC decreased.
OC/EC ratio
Feb-07 Mar-07 Apr-07 May-07 Jun-07 Jul-07 Aug-07 Sep-07 Oct-07 Nov-07 Dec-07 Jan-08 Feb-08
ratio
0
10
20
30 HelsinkiHyytiälä
Figure 13. The ratio of OC-to-EC for fine particles in Helsinki and in Hyytiälä from February 2007 to February 2008.
63
The organic carbon fraction was investigated further by determining how much of its mass is water-
soluble. The OC concentration, divided into water-soluble and water-insoluble portions, is shown in
Figure 14 for the Helsinki data (Paper VI). On an annual basis an average of 56% of OC was water-
soluble in Helsinki. The percentage varied during the year but, quite surprisingly, it was not
different during the biomass burning episodes from that during the other times of the year.
However, the contribution of WSOC to OC decreased in October 2007 and stayed low until the end
of the measurements. Based on these results the fraction of WSOC in OC was slightly higher in
summer than in winter which is line with the previous studies conducted in Europe (Decesari et al.,
2001; Jaffrezo et al., 2005; Viana et al., 2007).
Mar-06
Apr-06
May-06
Jun-0
6
Jul-0
6
Aug-06
Sep-06
Oct-06
Nov-06
Dec-06
Jan-0
7
Feb-07
Mar-07
WIN
SO
C, W
SO
C; µ
g m
-3
02468
10121416
WSO
C-to
-OC
ratio
0.0
0.2
0.4
0.6
0.8
1.0
WINSOCWSOC
WSOC/OC ratio
Figure 14. The concentrations of WSOC and WINSOC, and the ratio of WSOC-to-OC in Helsinki from March 2006 to February 2007. The AMS enables one to separate the organic matter into hydrocarbon-like and oxygenated organic
aerosol. By using the mass-to-charge ratio of 57 for HOA (mainly C4H9+) and 44 for OOA (mainly
CO2+) (Zhang et al., 2005a), the ratio of HOA-to-OOA can be obtained (Figure 15). From Figure 15
it can be seen that the ratio of HOA-to-OOA followed the time-evolution of total organics
concentration in Po Valley. That suggests that the high concentrations of organics in Po Valley
were caused mainly by the increased concentrations of HOA, whereas the concentrations of OOA
were more stable during the measurement period.
64
Po Valley
02-04-08 03-04-08 04-04-08 05-04-08
Org
anic
s µg
m-3
0
2
4
6
8m
z57-
to-m
z44
ratio
0,0
0,1
0,2
0,3
0,4
Organicsmz57-to-mz44 ratio
Figure 15. The ratio of m/z 57-to-m/z 44 and the concentration of total organics in Po Valley, Italy, 2–4 April, 2008. 5.2. Source apportionment
In this thesis the sources of the particles were studied by using two data analysis techniques: factor
analysis and PMF. The factor analysis was used for the particulate matter data (PM1 and PM1–10)
collected in Finokalia, whereas the PMF was used to reveal the sources of organic carbon in
Helsinki. Additionally, the sources were assessed by less quantitative methods. The impact of
traffic on the diurnal OC concentration was studied by applying the OC/EC ratio obtained from the
PMF analyses. Solely based on the levoglucosan concentration, the impact of biomass combustion
on PM2.5 and OC concentrations was assessed for four northern and central Europe urban sites.
5.2.1. Particulate matter
The sources of fine particulate matter in Finokalia were investigated in Paper III. Five factors were
found for PM1 particles (Table 7), of which two were due to the natural sources and the remaining
three were attributed to anthropogenic sources. Because of high loadings of Ca, Al, Fe, Mn and
Ca2+, the first factor was attributed to crustal components. The second factor was identified as
heavy oil combustion because it had high loading of V and Ni. The presence of MS- in the factor 2
was explained by the marine source which was the prevailing sector in summer, often associated
with heavy pollution episodes. OC and EC were associated mostly with the third factor. Because of
that it was ascribed to (automotive and fuel) combustion aerosols. Marine factor (the fourth factor)
had high loading of nitrate, sodium and magnesium. The fifth factor was assigned to secondary
aerosol, yet the presence of potassium suggested also a biomass burning origin. Crustal component
65
explained the largest fraction of the total variance (28.4%), whereas combustion accounted for the
smallest fraction of variance (9.5%).
Table 7. Varimax rotated factor matrix and probable source types for PM1 in Finokalia. Values above 0.6 are bolded. Based on Paper III.
Variable Factor 1 Factor 2 Factor 3 Factor 4 Factor 5
MS- -0.066 0.789 0.033 0.309 0.097
NO3- 0.426 0.225 0.182 0.674 -0.333
nss-SO42- -0.145 0.573 0.026 0.052 0.698
Oxalate -0.130 0.528 0.333 0.255 0.590
Na+ 0.152 -0.084 -0.119 0.880 0.188
NH4+ -0.189 0.518 0.075 0.004 0.724
K+ -0.055 -0.013 0.288 -0.037 0.796
Mg2+ 0.744 -0.127 -0.132 0.514 -0.049
Ca2+ 0.815 0.015 -0.138 0.085 -0.121
OC 0.008 0.098 0.703 0.011 0.314
EC 0.066 0.042 0.873 -0.066 0.063
Al 0.837 0.132 0.213 0.117 -0.255
Ca 0.900 0.001 0.116 0.012 -0,190
V 0.072 0.919 0.058 -0.149 0.139
Mn 0.939 0.008 0.084 0.113 0.153
Fe 0.960 -0.101 -0.053 0.049 -0.015
Ni 0.083 0.919 0.057 -0.143 0.110
Variance % 28.4 19.5 9.5 10.2 14.3 Probable source Crustal Heavy oil
combustion Combustion Marine Secondary
66
For the PM1–10, four sources were found (Paper III). Crustal component explained almost half of the
total variance (43.1%), marine factor accounted for 17.1% and photochemical sources accounted for
13.1%. Combustion sources explained only 7.3% of the total variance. Unlike for PM1 coarse OC
was associated with the crustal component. In the previous studies it has been shown that mineral
dust particles can serve as reaction surfaces for different species (Aymoz et al., 2004; Putaud et al.,
2004). This concerns especially semivolatile organic components (Falkovich et al., 2004), which
can explain the correlation of crustal species and OC.
5.2.2. Organic carbon
The sources of organic carbon in PM1 in Helsinki were studied in Paper VI. Four sources were
found: traffic, secondary formation, biomass combustion and long-range transport (LRT). Traffic
factor had high contributions of EC, NO and NO2, whereas SOA had a large contribution from
oxalate and ozone. The third factor was identified as biomass combustion because nearly all
levoglucosan was loaded into this factor. Also a large fraction of potassium, another tracer for
biomass combustion, was associated with the third factor. However, almost an equal contribution of
potassium was loaded into the fourth factor. Due to high loadings of sulfate and ammonium, the
fourth factor was identified as long-range transport.
The OC concentrations associated with the four sources are presented in Figure 16. In winter the
largest fraction of OC originated from biomass combustion (Table 8), but regarding the annual
averages, the largest fraction was attributed to SOA (34%) while the contributions of other sources
were nearly equal (21–23%). The contribution of SOA to OC was especially high in summer.
Traffic had the lowest contribution in summer probably due to low traffic volume during summer
holidays. The contribution of LRT to OC varied throughout the year, giving the highest contribution
in spring. All four sources also influenced the WSOC concentrations, however, the contribution of
SOA was significantly larger for WSOC than OC (Figure 16b).
67
Figure 16. The source-related concentrations of OC (a) and WSOC (b) at SMEAR III in Helsinki. Sources were identified by the PMF. Two major biomass burning episodes have been excluded. (Paper VI)
Table 8. The source contributions for OC in winter (Dec-Feb), spring (Mar-May), summer (Jun-Aug) and fall (Sep-Nov) (average ± SD) at SMEAR III in Helsinki.
Winter Spring Summer Fall
% of OC Traffic 26 ± 12 25 ±13 15 ± 13 27 ± 15
SOA 16 ± 11 34 ± 17 64 ± 12 32 ± 18
Biomass combustion 41 ± 15 12 ± 8.9 3.4 ± 6.0 20 ± 14
LRT 17 ± 13 29 ± 10 19 ± 11 21 ± 14
68
The diurnal variation of the impact of traffic on OC was estimated by using the OC/EC ratio of 0.71
for the 3-hour data. The contribution of traffic to OC was largest on weekdays during morning rush
hours from 06:00 to 09:00, being equal to 57%. After that the contribution decreased, having the
lowest value from 03:00 to 06:00 (18%). On weekends the contribution of traffic to OC ranged
from 19 to 30%. On Saturdays the contribution increased slightly in the morning, whereas on
Sunday the highest contribution was measured in the evening (18:00–21:00), caused obviously by
people returning home from their summer cabins. Similarly to the weekdays, the lowest
contribution on weekends was measured at 03:00–06:00. On average the contribution of traffic to
OC was 36% at weekdays and 24 and 25% on Saturdays and Sundays, respectively.
An example of the influence of local sources on OC concentration in Helsinki is shown in Figure
17. On Saturday, 10 February, the OC concentration increased significantly at 18:00 and remained
elevated until 06:00 on Sunday, 11 February. This concentration peak was probably caused by a
decreased boundary layer height accompanied by a decreasing temperature and low wind speed
(Figure 17d), causing a meteorological situation that favored poor dilution of emissions. The
highest OC concentration was measured on Saturday evening at 18:00–21:00 concurrently with the
peak concentration of potassium (Figure 17b), suggesting an impact by biomass combustion. In
addition to extensive wood combustion for domestic heating because of low ambient temperature,
the saunas are traditionally heated in Finland on Saturday evenings, and also the use of fireplace for
pleasure is most common on Saturday evenings. The chloride concentration followed closely that of
potassium (Figure 17b).
Besides biomass combustion the OC concentration was influenced by traffic. Traffic was obviously
partly responsible for the elevated EC concentrations, especially on Sunday, (Figure 17a) as well as
for elevated NO, NO2 and CO concentrations (not shown). Other ions analyzed, including sodium,
sulfate, ammonium and nitrate (Figure 17b–c), were not influenced by the temperature inversion,
which indicates that they did not originate from the local sources.
69
Saturday, 10 Feb
Local time 03 06 09 12 15 18 21 03 06 09 12 15 18 21 00 00 00
OC
, EC
; µg
m-3
0
5
10
15
20
25
OCEC
µg m
-3
0
1
2
3
4
5
6
K, C
l; µg
m-3
0.0
0.4
0.8
1.2
1.6
Na;
µg
m-3
0.00
0.05
0.10
0.15
0.20
Local time 03 06 09 12 15 18 21 03 06 09 12 15 18 21 00 00 00
Tem
pera
ture
; °C
-20
-15
-10
-5
0w
ind
spee
d; m
s-1
0
1
2
3
4
5
Twind speed
Sunday, 11 Feba)
b)
c)
d)
SO42-
NH4+
NO3-
Cl-
Na+
K+
Figure 17. The concentrations of OC and EC (a), sodium, potassium and chloride (b), sulfate, ammonium and nitrate (c) and temperature and wind speed (d) on 10–11 February, 2007, at SMEAR III in Helsinki (Paper IV).
70
The impact of biomass combustion on OC concentrations was also investigated at four urban sites
in central and northern Europe (Paper V). Besides Helsinki, the cities were Duisburg (Germany),
Prague (Czech) and Amsterdam (Netherlands). In Helsinki the measurements were performed in
four seasons. The largest contribution of biomass combustion to OC was obtained in Prague with a
campaign mean of 79%. In Duisburg and Amsterdam the campaign-mean contributions were 49
and 55%, respectively. Regarding the seasonal campaigns in Helsinki, the campaign-mean
contributions of biomass combustion to OC were 47, 48, 21 and 58% in winter, spring, summer and
fall, respectively. They were all larger than those obtained for Helsinki in 2006–2007 by using the
PMF (Table 8). One reason for the larger contributions is that the method of using only the ratio of
levoglucosan-to-OC obtained from the laboratory test is very uncertain. The emitted levoglucosan
concentration is highly dependent on combustion conditions (Hedberg et al., 2006) and the wood
type (Fine et al., 2001; Fine et al., 2002; Fine et al., 2004). Additionally, wood is not the only
biomass material burnt and the emissions from non-wood biomass burning can differ from those of
wood. Also the sampling methods used in the emission studies and in the field measurements are
not congruent, which can cause different sampling artifacts. However, also the PMF had some
issues. Although 92% of levoglucosan was associated with the biomass combustion factor by the
PMF, potassium was split in two factors: biomass combustion (39%) and long-range transport
(35%). The difference between levoglucosan and potassium suggested that a large portion of
biomass combustion particles might be long-range transported, and that levoglucosan had
disappeared during the transport.
71
6. SUMMARY AND CONCLUSIONS
In this thesis the chemical composition and the properties of atmospheric aerosol particles were
investigated in various environments by conducting campaign-based or long term measurements
and by deploying several experimental and analytical methods. It was found that the particle mass
concentration is rarely an adequate measure of their potential effects, since the chemistry of
particulate matter has usually strong spatial, temporal and size-dependent variability. In Finnish
urban and background sites most of the fine particle mass consisted of organic matter. However, at
a remote site in Crete sulfate dominated the fine PM and in a rural background site in Italy nitrate
made the largest contribution to the fine PM. In Helsinki the diurnal variation was clearest for
elemental carbon with its concentration cycle following the traffic intensity. In Italy the temporal
variation in particle chemistry was probably associated with the mixing of local emissions in the
boundary layer. Regarding the size-dependent chemical composition, organic components were
likely to be enriched in smaller particles than inorganic ions. Besides the different locations, the
chemical composition of particles was studied in the course of specific air pollution situations,
biomass burning and dust episodes, which were shown to alter the chemistry of particles
significantly.
The sources of organic carbon in Helsinki were investigated by using the PMF. This analysis
revealed a clear seasonal pattern for two of the particulate matter sources: secondary production and
biomass combustion. SOA was the major source in Helsinki during spring, summer and fall,
whereas in winter biomass combustion dominated OC. The significant impact of biomass
combustion on OC concentrations was also observed in the measurements performed in central
Europe.
In this thesis the aerosol samples were collected mainly by the conventional filter and impactor
methods. A new method was developed for determining size-segregated OC and EC by using a
multi-stage impactor, but similarly to other filter and impactor measurements, it suffered from the
long integration time. However, by filter and impactor measurements chemical mass closure was
achieved accurately and a simple filter sampling of submicrometer particles was found to be useful
in order to explain the sources of PM on the seasonal basis. The online instruments gave additional
72
information related to the temporal variations of the sources and the atmospheric mixing conditions.
Particularly the AMS proved to be a valuable aerosol instrument as it provided size-segregated
chemical composition of particles with a reasonable good time-resolution and sensitivity.
The focus of this thesis was on the major chemical components found in particles. Regarding
chemical mass closure most of the organic matter remained unresolved. Organic aerosol fraction
can be characterized more detailed by analyzing a variety of organic compounds from filter samples
by analytical methods not deployed in this study, such as by gas chromatography-mass
spectrometry. Also the high resolution mass spectrometry, used in Po Valley in this thesis, enables
to study the chemical composition of particles in species level. Moreover, new efforts to combine
AMS data with the PMF analysis will provide a promising tool for time-resolved source
apportionment, but the procedure was not yet available for this study.
73
7. REFERENCES
Aiken A.C, DeCarlo P.F., Kroll J.H., Worsnop D.R., Huffman J.A., Docherty K.S., Ulbrich I.M.,
Mohr C., Kimmel J.R., Sueper D., Sun Y., Zhang Q., Trimborn A., Nirthway M., Ziemann P.J., Canagaratna M.R., Onasch T.B., Alfarra M.R., Prevot A.S.H., Dommen J., Duplissy J., Metzger A., Baltensberger U. and Jimenez J.F. (2008) O/C and OM/OC ratios of primary, secondary, and ambient organic aerosols with high-resolution time-of-flight Aerosol mass spectrometry. Environ. Sci. Technol. 42, 4478–4485.
Allen G., Sioutas C., Koutrakis P., Reiss R., Lurmann F.W. and Roberts P.T. (1997) Evaluation of the TEOM® method for measurement of ambient particulate mass in urban areas. J. Air & Waste Manage. Assoc. 14, 682–689.
Alves C., Carvalho A. and Poi C. (2002) Mass balance of organic carbon fraction in atmospheric aerosols. J. Geophys. Res. 107, doi:10.1029/2001JD000616.
Andreae M.O. (1983) Soot carbon and excess fine potassium: long-range transport of combustion derived aerosols. Science 220, 1148–1151.
Andreae M.O., Jones C.D. and Cox P.M. (2005) Strong present-day aerosol cooling implies a hot future. Nature 435, 1187–1190.
Andreae M.O. and Rosenfeld D. (2008). Aerosol-cloud-precipitation interactions. Part 1. The nature and sources of cloud-active aerosols. Earth-Sci. Rev. 89, 13–41.
Anttila P., Rissanen T., Shimmo M., Kallio M., Hyötyläinen T., Kulmala M. and Riekkola M.-L. (2005) Organic compounds in atmospheric aerosols from a Finnish coniferous forest. Boreal Env. Res. 10, 371–384.
Artaxo P. and Hansson H.-C. (1995) Size distribution of biogenic aerosol particles from the Amazon basin. Atmos. Environ. 29, 393–402.
Aymoz G., Jaffrezo J.L., Jacob V., Colomb A. and George C. (2004) Evolution of organic and inorganic components of aerosol during a Saharan dust episode observed in the French Alps. Atmos. Chem. Phys. 4, 2499–2512.
Bauer H., Claeys M., Vermeylen R., Schueller E., Weinke G., Berger A. and Puxbaum H. (2008) Arabitol and mannitol as tracers for a quantification of airborne fungal spores. Atmos. Environ. 42, 588–593.
Berner A. and Lürzer C. (1980) Mass size distributions of traffic aerosols at Vienna. J. Phys. Chem. 84, 2079–2083.
Berner A., Sidla S., Galambos Z., Kruisz C., Hitzenberger R., ten Brink H.M. and Kos G.P.A. (1996) Modal character of atmospheric black carbon size distributions. J. Geophys. Res. 101, 19,559–19,565.
Birch M.E., and Cary R.A. (1996) Elemental carbon-based method for monitoring occupational exposures to particulate diesel exhaust. Aerosol Sci. Technol. 25, 221–241.
Brewer P. G. (1975) Minor elements in sea water. In: Chester R. (Ed.), Chemical Oceonography. Vol. 1. Academic, San Diego, California, pp. 417–425.
Brook J.R., Dann T.F. and Burnett R.T. (1997) The relationship among TSP, PM10, PM2.5 and inorganic constituents of atmospheric particulate matter at multiple Canadian locations. J. Air & Waste Manage. Assoc. 47, 2–19.
Brook J.R., Graham L., Charland J.P., Cheng Y., Fan X., Lu G., Li S.M., Lillyman C., MacDonald P., Caravaggio G. and MacPhee J.A. (2007) Investigation of the motor vehicle exhaust contribution to primary fine particle organic carbon in urban air. Atmos. Environ. 41, 119–135.
74
Brook R.D., Franklin B., Cascio W., Hong Y., Howard G., Lipsett M., Luepker R., Mittleman M., Samet J., Smith S. and Tager I. (2004) Air pollution and cardiovascular disease. Circulation 109, 2655–2671.
Cabada J. C., Khlystov A., Wittig A. E., Pilinis C. and Pandis S. (2004) Light scattering by fine particles during Pittsburgh air quality study: measurements and modeling. J. Geophys. Res. 109, doi:10.1029/2003JD004155.
Carvalho A., Pio C. and Santos C. (2003) Water-soluble hydroxylated organic compounds in German and Finnish aerosols. Atmos. Environ. 37, 1775–1783.
Castro L.M., Pio C.A., Harrison R. and Smith D.J.T. (1999) Carbonaceous aerosol in urban and rural European atmospheres: estimation of secondary organic carbon concentrations. Atmos. Environ. 33, 2771–2781.
Cavalli F., Facchini M.C., Decesari S., Emblico L., Mircea M., Jensen N.R. and Fuzzi S. (2006) Size-segregated aerosol chemical composition at a boreal site in southern Finland, during the QUEST project. Atmos. Chem. Phys. 6, 993–1002.
Chow J.C., Watson J.G., Fujita E.M., Lu Z., Lawson D.R. and Ashbaugh L.L. (1994) Temporal and spatial variations of PM2.5 and PM10 in the southern California air quality study. Atmos. Environ. 28, 2061–2080.
Chow J.C., Watson J.G., Lowenthal D.H., Chen L.W.A, Zielinska B., Mazzoleni L.R. and Magliano K.L. (2007) Evaluation of organic markers for chemical mass balance source apportionment at the Fresno supersite. Atmos. Chem. Phys. 7, 1741–1754.
Davidson C.I., Phalen R.F. and Solomon P.A. (2005) Airborne particulate matter and human health: a review. Aerosol Sci. Technol. 39, 737–749.
DeCarlo P.F., Kimmel J.R., Trimborn A., Northway M.J., Jayne J.T., Aiken A.C., Gonin M., Fuhrer K., Horvath T., Docherty K.S., Worsnop D.R. and Jimenez J.L. (2006) Field-deployable, high-resolution, time-of-flight mass spectrometer. Anal. Chem. 78, 8281–8289.
Decesari S., Facchini M. C., Matta E., Lettini F., Mircea M., Fuzzi S., Tagliavini E. and Putaud J.-P. (2001) Chemical features and seasonal variation of fine aerosol water-soluble organic compounds in the Po Valley, Italy. Atmos Environ. 35, 3691–3699.
Dockery D. W., Pope C.A:, Xu X., Spengler J.D., Ware J. H., Fay M. E., Ferris B. G. and Speizer F. E. (1993) A association between air pollution and mortality in six U.S. cities. N. Engl. J. Med. 329, 1753-1759.
Duan F., Liu X., Yu T. and Cachier H. (2004) Identification and estimate of biomass burning contribution to the urban aerosol organic carbon concentration in Beijing. Atmos. Environ. 38, 1275–1282.
Dye C. and Yttri K.E. (2005) Determination of monosaccharide anhydrides in atmospheric aerosols by use of high-performance liquid chromatography combined with high-resolution mass spectrometry. Anal. Chem. 77, 1853–1858.
Eldering A. and Cass G.R. (1996) Source-oriented model for air pollutant effects on visibility. J. Geophys. Res. 101, 19,343–19,369.
Falkovich A.H., Schkolnik G., Ganor E. and Rudich Y. (2004) Adsorption of organic compounds to urban environment onto mineral dust particles. J. Geophys. Res. 109, D02208, doi 10.1029/2003JD003919.
Fine P.M., Cass G.R. and Simoneit B.R.T. (2001) Chemical characterization of fine particle emissions from fireplace combustion of woods grown in the Northeastern United States. Environ. Sci. Technol. 35, 2665–2675.
75
Fine P.M., Cass G.R. and Simoneit B.R.T. (2002) Chemical characterization of fine particle emissions from fireplace combustion of woods grown in the Southern United States. Environ. Sci. Technol. 36, 1442–1451.
Fine P.M., Cass G.R. and Simoneit B.R.T. (2004) Chemical characterization of fine particle emissions from the fireplace combustion of wood types grown in the Midwestern and Western United States. Environ. Eng. Sci. 21, 387–409.
Finlayson-Pitts B.J. and Pitts J.N. Jr (2000) Chemistry of upper and lower atmosphere: theory, experiments, and applications. Academic press, San Diego.
Fisseha R., Dommen J., Gaeggeler K., Weingartner E., Samburova V., Kalberer M. and Baltensperger U. (2006) Online gas and aerosol measurement of water soluble carboxylic acids in Zurich. J. Geophys. Res. 111, D12316.
Fraser M.P., Yue Z.W. and Buzcu B. (2003) Source apportionment of fine particulate matter in Houston, TX, using organic molecular markers. Atmos. Environ. 37, 2117–2123.
Gao S., Hegg D. A., Hobbs P.V., Kirchstetter T.W., Magi B.I. and Sadilek M. (2003) Water-soluble organic components in aerosols associated with savanna fires I southern Africa: Identification, evolution and distribution. J. Geophys. Res. 108, D13, 8491, doi:10.1029/2002JD002324.
Gelencsér A., Hoffer A., Molnár A., Krivácsy Z., Kiss G. and Mészáros E. (2000) Thermal behaviour of carbonaceous aerosol from a continental background site. Atmos. Environ. 34, 823–831.
Gerasopoulos E., Kouvarakis, G., Babasakalis P., Vrekoussis, M., Putaud J.P. and Mihalopoulos N. (2006) Origin and variability of particulate matter (PM10) mass concentrations over the Eastern Mediterranean, Atmos. Environ. 40, 4679-4690.
Glasius M., Duane M. and Larsen B.R. (1999) Determination of polar terpene oxidation products in aerosols by liquid chromatography-ion trap mass spectrometry. J. Chromatogr. A 833, 121–135.
Graber E.R. and Rudich Y. (2006) Atmospheric HULIS: How humic-like are they? A comprehensive and critical review. Atmos. Chem. Phys. 6, 729–753.
Graham B., Falcovich A.H., Rudich Y., Maenhaut W., Guyon P. and Andreae M. (2003) Local and regional contributions to the atmospheric aerosol over Tel Aviv, Israel: a case study using elemental, ionic and organic tracers. Atmos. Environ. 38, 1593–1604.
Graney J.R., Landis M.S. and Norris G.A. (2004) Concentrations and solubility of metals from indoor and personal exposure PM2.5 samples. Atmos. Environ. 38, 237–247.
Grover B.D., Eatough N.L., Eatough D.J., Chow J.C., Watson J.G., Ambs J.L., Meyer M.B., Hopke P.K., Al-Horr R., Later D.W. and Wilson W.E. (2006) Measurement of both nonvolatile and semi-volatile fractions of fine particulate matter in Fresno, CA. Aerosol Sci. Techonol. 40, 811–826.
Gundel L.A., Lee V.C., Mahanama K.R.R., Stevens R.K. and Daisey, J.M. (1995) Direct Determination of the Phase Distributions of Semi-volatile Polycyclic Aromatic Hydrocarbons Using Annular Denuders. Atmos. Environ. 29, 1719-1733.
Hansen A.D.A., Rosen H. and Novakov T. (1984) The aethalometer: an instrument for the real-time measurement of optical absorption by aerosol particles. Sci. Total Environ. 36, 191–196.
Harrison R.M., Jones A.M. and Lawrence R.G. (2003) A pragmatic mass closure model for airborne particulate matter at urban background and roadside sites. Atmos. Environ. 37, 4927–4933.
Hedberg E., Johansson C., Johansson L., Swietlicki E. and 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, 1669–78.
76
Hillamo R.E. and Kauppinen E.I. (1991) On the performance of the Berner Low Pressure Impactor, Aerosol Sci. Technol. 14, 33–47.
Hinds W.C. (1999) Aerosol Technology: Properties, behaviour, and measurement of airborne particles. New York, Wiley.
Huang L., Brook J.B., Zhang W., Li S.M., Graham L., Ernst D., Chiculescu A. and Lu G. (2006) Stable isotope measurements of carbon fractions (OC/EC) in airborne particulate: A new dimension for source characterization and apportionment. Atmos. Environ. 40, 2690–2705.
Hueglin C., Gehrig R., Baltensberger U., Gysel M., Monn C. and Vonmont H. (2005) Chemical characterization of PM2.5, PM10 and coarse particles at urban, near-city and rural Switzerland. Atmos. Environ. 39, 637–651.
IPCC (the Intergovernmental Panel on Climate Change) (2007) IPCC Fourth assessment report: The Physical science basis. Summary for policy makers. Available online at http://www.ipcc.ch/pdf/assessment-report/ar4/wg1/ar4-wg1-spm.pdf
Jacobson M.C., Hansson H.-C., Noone K.L. and Charlson R.J. (2000) Organic atmospheric aerosols: Review and state of the science. Rev. Geophys. 38, 267–294.
Jaffrezo J.-L., Davidson C.I., Kuhns H.D., Bergin M.H., Hillamo R., Maenhaut W., Kahl J.W. and Harris J.M., (1998) Biomass burning signatures in the atmosphere of central Greenland. J. Geophys. Res. D23, 31,067–31,078.
Jaffrezo J.-L., Aymoz G., Delaval C. and Cozic J. (2005) Seasonal variation of the water soluble organic carbon mass fraction of aerosol in two valleys of the French Alps. Atmos. Chem. Phys. 5, 2809–2821.
Jeong C.-H., Lee D.-W., Kim E. and Hopke P.K. (2004) Measurement of real-time PM2.5 mass, sulfate and carbonaceous aerosols at the multiple monitoring sites. Atmos. Environ. 38, 5247–5256.
Jimenez J-L., Jayne J.T., Shi Q., Kolb C.E., Worsnop D.R., Yourshaw I., Seinfeld J.H., Flagan R.C., Zhang X., Smith K. A., Morris J.W. and Davidovits P. (2003) Ambient aerosol sampling using the Aerodyne Aerosol Mass Spectrometer. J. Geophys. Res. 108, 8425. doi:10.1029/2001JD001213.
Jokinen V. and Mäkelä J.M. (1997) Closed loop arrangement with critical orifice for DMA sheath/excess flow system. J. Aerosol Sci. 28, 643–648.
Kajino M., Winiwarter W. and Ueda H. (2006) Modeling retained water content in measured aerosol mass. Atmos. Environ. 40, 5202–5213.
Kalberer M., Paulsen D., Sax M., Steinbacher M., Dommen J., Prevot A.S.H., Fisseha R., Weingartner E., Frankevich V., Zenobi R. and Baltensberger U. (2004) Identification of polymers as major components of atmospheric organic aerosols. Science 303, 1659–1662.
Khlystov A., Stanier C.O., Takahama S. and Pandis S.N. (2005) Water content of ambient aerosol during the Pittsburgh Air Quality Study. J. Geophys. Res. 110, D07S10, doi:10.1029/2004JD004651.
Kiss G., Tombácz E., Varga B., Alsberg T. and Persson L. (2003) Estimation of the average molecular weight of humic-like substances isolated from fine atmospheric aerosol. Atmos. Environ. 37, 3783–3794.
Koeber R., Bayona J.M. and Niessner R. (1999) Determination of benzo[a]pyrene diones in air particulate matter with liquid chromatography mass spectrometry. Environ. Sci. Technol. 33, 1552–1558.
77
Kondo Y., Miyazaki Y., Takegawa N., Miyakawa T., Weber R.J., Jimenez J.L., Zhang Q. and Worsnop D.R. (2007) Oxygenated and water-soluble organic aerosols in Tokyo. J. Geophys. Res. 112, D01203, doi:10.1029/2006JD007056, 2007.
Krivácsy Z., Hoffer A., Sárvári Zs., Temesi D., Baltensberger U., Nyeki S., Weingartner E., Kleefeld S. and Jennings S.G. (2001) Role of organic and black carbon in the chemical composition of atmospheric aerosol at European background sites. Atmos. Environ. 35, 6231–6244.
Kroll J.H. and Seinfeld J.H. (2008) Chemistry of secondary organic aerosol: formation and evolution of low-volatility organics in the atmosphere. Atmos. Environ. 42, 3593–3624.
Kulmala M., Hämeri K., Aalto P.P., Mäkelä J.M., Pirjola L., Nilsson E.D., Buzorius G., Rannik Ü., Dal Maso M., Seidl W., Hoffmann. T., Janson R., Hansson H.-C., Viisanen Y., Laaksonen A. and O’Dowd C.D. (2001) Overview of the intenational project on Biogenic aerosol formation in the boreal forest (BIOFOR). Tellus 53B, 324–343.
Kuokka S., Teinilä K., Saarnio K., Aurela M., Sillanpää M., Hillamo R., Kerminen V.-M., Pyy K., Vartiainen E., Kulmala M., Skorokhod A.I., Elansky N.F. and Belikov I.B. (2007) Using a moving measurement platform for determining the chemical composition of atmospheric aerosols between Moscow and Vladivostok. Atmos. Chem. Phys. 7, 4793–4805.
Lanz V.A., Alfarra M.L., Baltensberger U., Buchmann B., Hueglin C. and Prévôt A.S.H. (2007). Source apportionment of submicron organic aerosols at an urban site by factor analytical modelling of aerosol mass spectra. Atmos. Chem. Phys. 7, 1503–1522.
Lau A.P.S., Lee A.K.Y., Chan C.K. and Fang M. (2006) Ergosterol as a biomarker for the quantification of the fungal biomass in atmospheric aerosol. Atmos. Environ. 40, 249–259.
Lemire K.R., Allen D.T., Klouda G.A. and Lewis C.W. (2002) Fine particulate matter source attribution for Southesat Texas using 14C/13C ratios. J. Geophys. Res. 107, doi:10.1029/2002JD002339.
Lighty J. S., Veranth J. M. and Sarofim A. F. (2000) Combustion aerosols: factors governing their size and composition and implications to human health. J. Air Waste Manage. Assoc. 50, 1565–1618.
Loo B.W. and Cork C.P. (1988) Development of high efficiency virtual impactor. Aerosol Sci. Technol. 9, 167-170.
Maenhaut W., Hillamo R., Mäkelä T., Jaffrezo J.-L., Bergin M.H.and Davidson C.I. (1996) A New Cascade Impactor for Aerosol Sampling with subsequent PIXE analysis. Nucl. Inst. Meth. Phys. Res. B. 109/110: 482-487.
Matta E., Facchini M.C., Decesari S., Mircea M., Cavalli F., Fuzzi S., Putaud J.-P. and Dell’Acqua A. (2003) Mass closure on the chemical species in size-segregated atmospheric aerosol collected in an urban area of the Po Valley, Italy. Atmos. Chem. Phys. 3, 623–637.
Mayol-Bracero O.L., Guyon P., Graham B., Roberts G., Andreae M.O., Decesari S., Facchini M.C., Fuzzi S. and Artaxo P. (2002) Water-soluble organic compounds in biomass burning aerosols over Amazonia 2. Apportionment of the chemical composition and importance of the polyacidic fraction. J. Geophys. Res. 107, D20 8091, doi:10.1029/2001JD000522.
Medeiros P.M., Conte M.H., Weber J.C. and Simoneit B.R.T. (2006) Sugars as source indicators of biogenic organic carbon in aerosols collected above the Howland Experimental forest, Maine. Atmos. Environ. 40, 1964–1705.
Mäkelä J.M., Aalto P., Jokinen V., Pohja T., Nissinen A., Palmroth S., Markkanen T., Seitsonen K., Lihavainen H. and Kulmala, M. (1997) Observations of utrafine particle formation and growth in boreal forest. Geophys. Res. Lett. 24, 1219–1222.
78
Mäkelä J.M., Yli-Koivisto S., Hiltunen V., Seidl W., Swietlicki E., Teinilä K., Sillanpää M., Koponen I., Paatero J. and Rosman K. (2001) Chemical composition of aerosol during particle formation events in boreal forest. Tellus 53B, 380–393.
Nel A. (2005). Air pollution-related illness: effects of particles. Science 308, 804–806. Niemi J.V., Saarikoski S., Tervahattu H., Mäkelä T., Hillamo R., Vehkamäki H., Sogacheva L. and
Kulmala M. (2006) Changes in background aerosol composition in Finland during polluted and clean periods studied by TEM/EDX individual particle analysis. Atmos. Chem. Phys. 6, 5049-5066.
Nolte C.G., Schauer J. J, Cass G.R. and Simoneit B.R.T. (2001) Highly polar organic compounds present in wood smoke and in the ambient atmosphere. Environ. Sci. Technol. 35, 1912–1919.
Orsini D.A., Ma Y., Sullivan A., Sierau B., Baumann K. and Weber R.J. (2003) Refinements to the particle-into-liquid sampler (PILS) for ground and airborne measurements of water soluble aerosol composition. Atmos. Environ. 37, 1243–1259.
Paatero P. (1997) Least squares formulation of robust non-negative factor analysis, Chemometr. Intell. Lab. 37, 23–35.
Paatero P. (1999) The multilinear engine – a table-driven, least squares program for solving multilinear problems, including the n-way parallel factor analysis model, J. Comput. Graph. Stat. 1, 854–888.
Pakkanen T.A., Loukkola K., Korhonen C.H., Aurela M., Mäkelä T., Hillamo R., Aarnio P., Koskentalo T., Kousa A. and Maenhaut W. (2001) Sources and chemical composition of atmospheric fine and coarse particles in the Helsinki area. Atmos. Environ. 35, 5381–5391.
Pakkanen T.A. and Hillamo R.E. (2002) Comparison of sampling artifacts an ion balances for a Berner low-pressure impactor and virtual-impactor. Boreal Env. Res. 7, 129–140.
Park S.S., Kleissl J., Harrison D., Kumar V., Nair N.P., Adam M. and Parlange M. (2006) Characteristics of PM2.5 episodes revealed by semi-continuous measurements at the Baltimore Supersite at Ponca St. Aerosol Sci. Technol. 40, 845–860.
Patashnick H., Rupprecht E.G. (1991) Continuos PM-10 measurements using the Tapered Element Oscillating MicroBalance. J. Air Waste Manage. Assoc. 41, 1079–1083.
Pio C.A., Legrand M., Oliveira T., Afonso J., Santos C., Caseiro A., Fialho P., Barata F., Puxbaum H., Sanchez-Ochoa A., Kasper-Giebl A., Gelencsér A., Preunkert S. and Schock M. (2007) Climatology of aerosol composition (organic versus inorganic) at nonurban sites on a west-east transect across Europe. J. Geophys. Res. 112, doi:10.1029/2006JD008038.
Plaza J., Gómez-Moreno F. J., Núñez L., Pujadas M. and Artíñano B. (2006) Estimation of secondary organic aerosol formation from semi-continuous OC-EC measurements in a Madrid suburban area. Atmos. Environ. 40, 1134–1147.
Polidori A., Turpin B.J., Davidson C.I., Rodenburg L.A. and Maimone F. (2008) Organic PM2.5: fractionation by polarity, FTIR spectroscopy, and OM/OC ratio for the Pittsburgh aerosol. Aerosol. Sci. Technol. 42, 233–246.
Putaud J.-P., van Dingenen R., Mangoni M., Virkkula A., Raes F., Maring H., Prospero J.M., Swietlicki E., Berg O.H., Hillamo R. & Mäkelä T. (2000) Chemical mass closure and assessment of the origin of the submicron aerosol in the marine boundary layer and the free troposphere at Tenerife during ACE-2. Tellus 52B: 141–168.
Putaud J.P., Raes F., Van Dingenen R., Brüggemann E., Facchini M.C., Decesari S., Fuzzi S., Gehrig R., Hüglin C., Laj P., Lorbeer G., Maenhaut W., Mihalopoulos N., Müller K., Querol X., Rodriguez S., Schneider J., Spindler G., ten Brink H., Torseth K. and Wiedensohler A. (2004) A European aerosol phenomenology—2: chemical characteristics of particulate matter at kerbside,
79
urban, rural and background sites in Europe. Atmos. Environ. 38, 2579–2595. Puxbaum H., Caseiro A., Sánchez-Ochos A., Kasper-Giebl A., Claeys M., Legrand M., Preunkert S.
and Pio C. (2007) Levoglucosan levels at background sites in Europe for assessing the impact of biomass combustion on the European aerosol background. J. Geophys. Res. 112, D23S05, doi:10.1029/2006JD008114.
Ramanathan V. and Carmichael G. (2008) Global and regional climate changes due to black carbon. Nature Geosci. 1, 221 - 227.
Robinson A.L., Donahue N.M. and Rogge W.F. (2006) Photochemical oxidation and changes in molecular composition of organic aerosol in the regional context. J. Geophys. Res. 111, D03302, doi:10.1029/2005JD006265.
Robinson A.L., Donahue N.M., Shrivastava M.K., Weitkamp E.A., Sage A.M., Grieshop A.P., Lane T.E., Pierce J.R. and Pandis S. N. (2007) Rethinking organic aerosols: semivolatile emissions and photochemical aging. Science 315, 1259–1262.
Rogge W.F., Hildemann L.M., Mazurek M.A., Cass G.R. and Simoneit B.R.T. (1993a) Sources of fine organic aerosol 2. noncatlyst and catalyst-equipped automobiles and heavy duty diesel trucks. Environ. Sci. Technol. 27, 636–651.
Rogge W. F., Mazurek M. A., Hildemann L. M., Cass G. R. and Simoneit B. R. T. (1993b) Quantification of urban organic aerosols at molecular level: identification, abundance and seasonal variation. Atmos. Environ. 27A, 1309–1330.
Rogge W.F., Hildemann L.M., Mazurek M.A., Cass G.R. and Simoneit B.R.T. (1996) Mathematical modeling of atmospheric fine particle-associated primary organic compound concentrations. J. Geophys. Res. 101, 19,379–19,394.
Rogge W.F., Hildemann L.M., Mazurek M.A. and Cass G.R. (1998) Sources of fine organic aerosol. 9. pine, oak, and synthetic log combustion in residential fireplaces. Eviron. Sci. Technol. 32, 13-22.
Rogge W.F., Medeiros P.M. and Simoneit B.R.T. (2006) Organic marker compounds for surface soil and fugitive dust from open lot dairies and cattle feedlots. Atmos. Environ. 40, 27–49.
Ruellan S. and Cachier H. (2001) Characterization of fresh particulate vehicular exhausts near a Paris high flow road. Atmos. Environ. 35, 453–468.
Rudolph J., Anderson R.S., Czapiewski K.V., Czuba E., Ernst D., Gillespie T., Huang L., Rigby C. and Thompson A.E. (2003) The stable carbon isotope ratio of biogenic emisissions of isoprene and the potential use of stable isotope ratio measurements to study photochemical processing of isoprene in the atmosphere. J. Atmos. Chem. 44, 39–55.
Russel L.M. (2003) Aerosol Organic-Mass-to-Organic-Carbon Ratio Measurements. Environ. Sci. Technol. 37, 2982–2987.
Samet J. M., Dominici F., Curriero F. C., Coursac I. and Zeger S. L. (2000) Fine particulate air pollution and mortality in 20 U.S. cities, 1987–1994. N. Eng. J. Med. 343, 1742-1749.
Schaap M., Spindler G., Schulz M., Acker K., Maenhaut W., Berner A., Wieprecht W., Streit N., Müller K., Brüggemann E., Chi X., Putaud J-P., Hitzenberger R., Puxbaum H., Baltensperger U. and ten Brink H. (2004) Artefacts in the sampling of nitrate studied in the “INTERCOMP” campaigns of EUROTRAC-AEROSOL. Atmos. Environ. 38, 6487–6496.
Schauer J.J., Rogge W.F., Hildemann L.M., Mazurek M.A., Cass G.R. and Simoneit B.R.T. (1996) Source apportionment of airborne particulate matter using organic compounds as tracers. Atmos. Environ. 22, 3837–3855.
Schauer J.J. and Cass G. (2000) Source apportionment of wintertime gas-phase and particle-phase air pollutants using organic compounds as tracers. Environ. Sci. Technol. 34, 1821–1832.
80
Schauer J.J., Kleeman M.J., Cass G.R. and Simoneit B.R.T. (2001) Measurement of emissions from air pollution sources. 3. C1–C29 organic compounds from fireplace combustion of wood. Environ. Sci. Technol. 35, 1716–1728.
Schauer J.J., Mader B.T., Deminter J.T., Heidemann G., Bae M.S., Seinfeld J.H., Flagan R.C., Cary R.A., Smith D., Huebert B.J., Bertram T., Howell S., Kline J.T., Quinn P., Bates T., Turpin B., Lim H.J., Yu J.Z., Yang H. and Keywood M.D. (2003) ACE-Asia intercomparison of a thermal-optical method for the determination of particle-phase organic and elemental carbon. Environ. Sci. Technol. 37, 993–1001.
Seinfeld J.H. and Pandis S.N. (1998) Atmospheric Chemistry and Physics, from Air Pollution to Climate Change. New York, Wiley.
Sellegri K., Laj P., Peron F., Dupuy R., Legrand M., Preunkert S., Putaud J.-P., Cachier H. and Ghermandi G. (2003) Mass balance of free tropospheric aerosol st the Puy de Dôme (France) in winter. J. Geophys. Res. 108, D11 2 1–17.
Sheffield A.E., Gordon G.E., Currie L.A. and Riederer G.E. (1994) Organic, elemental and isotopic tracers of air pollution sources in Albuquerque, NM. Atmos. Environ. 28, 1371–1384.
Sillanpää M., Frey A., Hillamo R., Pennanen A. and Salonen R.O. (2005a) Organic, elemental and inorganic carbon in particulate matter of six urban environments in Europe. Atmos. Chem. Phys. 5, 2869–2879.
Sillanpää M., Saarikoski S., Hillamo R., Pennanen A., Makkonen U., Spolnik U., Van Grieken R., Koskentalo T. and Salonen R.O. (2005b) Chemical composition, mass size distribution and source analysis of long-range transported wildfire smokes in Helsinki. Sci. Total Environ. 350, 119–135.
Sillanpää M. (2006a) Chemical and source characterization of size-segregated urban air particulate matter. PhD Thesis. Finnish Meteorological Institute Contributions NO. 58, pp.50.
Sillanpää M., Hillamo R., Saarikoski S., Frey A., Pennanen A., Makkonen U., Spolnik Z., Van Grieken R., Braniš M., Brunekreef B., Chalbot M.-C., Kuhlbusch T., Sunyer J., Kerminen V.-M., Kulmala M. and Salonen R.O. (2006b) Chemical composition and mass closure of particulate matter at six urban sites in Europe. Atmos. Environ. 40S2, 212–223.
Simoneit B.R.T., Schauer J.J., Nolte C.G., Oros D.R., Elias V.O., Fraser M.P., Rogge W.F. and Cass G.R. (1999) Levoglucosan, a tracer for cellulose in biomass burning and atmospheric particles. Atmos. Environ. 33, 173–182.
Simoneit B.R.T. (2002) Biomass burning — a review of organic tracers for smoke from incomplete combustion. Appl. Geochem. 17, 129–162.
Simoneit B.R.T., Elias V.O., Kobayashi M., Kawamura K., Rushdi A.I., Medeiros P.M., Rogge W.F. and Didyk B.M. (2004) Sugars–Dominant water-soluble organic compounds in soils and characterization as tracers in atmospheric particulate matter. Env. Sci. Technol. 38, 5939–5949.
Simpson D., Yttri K.E., Klimont Z., Kupiainen K., Caseiro A., Gelencsér A., Pio C., Puxbaum H. and Legrand M. (2007) Modeling carbonaceous aerosol over Europe: Analysis of the CARBOSOL and EMEP EC/OC campaigns. J. Geophys. Res. 112, doi:10.1029/2006JD008158.
Song Y., Zhang Y., Xie S., Zeng L., Zheng M., Salmon L.G., Shao M. and Slanina S. (2006) Source apportionment of PM2.5 in Beijing by positive matrix factorization. Atmos. Environ. 40, 1526–1537.
Strader R., Lurmann F. and Pandis P.N. (1999) Evaluation of secondary organic aerosol formation in winter. Atmos. Environ. 33, 4849–4863.
Subramanian R., Khlystov A.Y. and Robinson A.L. (2006) Effect of peak inert-mode temperature on elemental carbon measured using thermal optical analysis. Aerosol Sci.Technol. 40, 763–780.
81
Subramanian R., Donahue N.M., Bernardo-Bricker A., Rogge W.F. and Robinson, A.L. (2007) Insight into the primary-secondary and regional-local contributions to organic aerosol and PM2.5 mass in Pittsburgh, Pennsylvania. Atmos. Environ. 41, 7414–7433.
Sullivan A.P. and Weber R.J. (2006a) Chemical characterization of the ambient organic aerosol soluble in water: 1. Isolation of hydrophobic and hydrophilic fractions with a XAD-8 resin. J. Geophys. Res. 111, doi:10.1029/2005JD006485.
Sullivan, A.P. and Weber R.J. (2006b) Chemical characterization of the ambient organic aerosol soluble in water: 2. Isolation of acid, neutral and basic fractions by modified size-exclusion chromatography. J. Geophys. Res. 111, doi:10.1029/2005JD006486.
Szarka N., Kakucs O., Wolfbauer J. and Bezama A. (2008) Atmospheric emissions modeling of energetic biomass alternatives using system dynamics approach. Atmos. Environ. 42, 403–414.
Szidat S., Jenk T.M., Gäggeler H.W., Synal H.-A., Fisseha R., Baltensberger U., Kalberer M., Samburova V., Wacker L., Saurer M., Schwikowski M. and Hajdas I. (2004) Source apportionment of aerosols by 14C measurements in different carbonaceous particle fractions. Radiocarbon 46, 475–484.
Szidat S., Jenk T.M., Synal H.-A., Kalberer M., Wacker L., Hajdas I., Kasper-Giebl A. and Baltensperger U. (2006) Contributions of fossil fuel, biomass-surning, and biogenic emissions to carbonaceous aerosols in Zurich as traced by 14C. J. Geophys. Res. 111, D07206.
Taylor P.G., Wiesenthal T. and Mourelatou A. (2005) Energy and environment in European Union: an indicator-based analysis. Natural Resources Forum, November 2005, vol. 29(4), pp. 360–380.
Temesi D., Molnár A., Mészáros E., Feczkó F., Gelencsér A., Kiss G. and Krivácsy Z. (2001) Size- resolved chemical mass balance of aerosol particles over rural Hungary. Atmos. Environ. 35, 4347–4355.
Timonen, H.J., Saarikoski, S.K., Aurela, M.A., Saarnio, K.M. and Hillamo R.E.J. (2008) Water-soluble organic carbon in urban aerosol: concentrations, size distributions and contribution to particulate matter. Boreal Env. Res. 13, 335–346.
Tsyro S.G. (2005) To what extent can aerosol water explain the discrepancy between model calculated and gravimetric PM10 and PM2.5? Atmos. Chem. Phys. 5, 515–532.
Turpin B.J. and Huntzicker J.J. (1995) Identification of secondary organic aerosol episodes and quantitation of primary and secondary organic aerosol concentrations during SCAQS. Atmos. Environ. 29, 3527–3544.
Turpin B.J., Saxena P. and Andrews, E. (2000) Measuring and Simulating Particulate Organics in the Atmosphere: Problems and Prospects. Atmos. Environ. 34, 2983-3013.
Turpin B.J. and Lim H.-J. (2001) Species Contributions to PM2.5 Mass Concentrations: Revisiting Common Assumptions for Estimating Organic Mass. Aerosol Sci. Technol. 35, 602–610.
Viana M., Maenhaut W., ten Brink H.M., Chi X., Weijers E., Querol X., Alastuey A., Mikuška P. and Večeřa Z. (2007) Comparative analysis of organic and elemental carbon concentrations in carbonaceous aerosols in three European cities, Atmos. Environ. 41, 5972–5983.
Viidanoja J., Sillanpää M., Laakia J., Kerminen V.-M., Hillamo R., Aarnio P. and Koskentalo T. (2002) Organic and black carbon in PM2.5 and PM10: 1 year of data from an urban site in Helsinki, Finland. Atmos. Environ. 36, 3183–3193.
Wang Q., Shao M., Liu Y., William K., Paul G., Li X., Liu Y. and Lu S. (2007) Impact of biomass burning on urban air quality estimated by organic tracers: Guangzhou and Beijing as cases. Atmos. Environ. 41, 8380–8390.
Watson J. G. (1984) Overview of receptor model principles. JAPCA 34, 619–623. Winklmayr W., Wang H.-C. and John W. (1990) Adaptation of the Twomey algorithm to the
82
inversion of cascade impactor data. Aerosol Sci. Technol. 13, 322–331. Winklmayr W., Reischl G.P., Lindner A.O. and Berner A. (1991) A new electromobility
spectrometer for the measurement of aerosol size distribution in the size range from 1 to 1000 nmn. J Aerosol Sci. 22, 289–296.
Wolfenbarger J.K. and Seinfeld J.H. (1990) Inversion of aerosol size distribution data. J. Aerosol Sci. 21, 227–247.
Yin J. and Harrison R.M. (2008) Pragmatic mass closure study for PM1.0, PM2.5 and PM10 at roadside, urban background and rural sites. Atmos. Environ. 42, 980–988.
Yttri K.E., Dye C., Siørdal L.H. and Braathen O.-A. (2005) Quantification of monosaccharide anhydrides by liquid chromatography combined with mass spectrometry: application to aerosol samples from an urban and a suburban site influenced by small-scale wood burning, J. Air & Waste Manage. Assoc. 55, 1169–1177.
Yu J.Z., Xu J. and Yang, H. (2002) Charring characteristics of atmospheric organic particulate matter in thermal analysis. Environ. Sci. Technol. 36, 754–761.
Zappoli S., Andracchio A., Fuzzi S., Facchini M.C., Gelencsér A., Kiss G., Krivácsy Z., Molnár Á., Mészáros E., Hansson H.-C., Rosman K. and Zebühr Y. (1999) Inorganic, organic and macromolecular components of fine aerosol in different areas of Europe in elation to their water-solubility. Atmos. Env. 33, 2733–2743.
Zdráhal Z., Oliveira J., Vermeylen R., Claeys M. and Maenhaut W. (2002) Improved method for quantifying levoglucosan and related monosaccharide anhydrides in atmospheric aerosols and application to samples from urban and tropical locations. Environ. Sci. Technol. 36, 747–753.
Zhang Q., Alfarra M.R., Worsnop D.R., Allan J.D., Coe H., Canagaratna M.R. and Jimenez J.L. (2005a) Deconvolution and quantification of hydrocarbon-like and oxygenated organic aerosols based on Aerosol Mass Spectrometry, Environ. Sci. Techol. 39, 4938–4952.
Zhang Q., Worsnop D.R., Canagaratna M.R. and Jimenez J.L. (2005b) Hydrocarbon-like and oxygenated organic aerosols in Pittsburg: insights into sources and processes of organic aerosol, Atmos. Chem. Phys. 5, 3289–3311.
Zielinska B., Sagebiel J., McDonald J.D., Whitney K. and Lawson, D.R. (2004) Emission rates and comparative chemical composition from selected in-use diesel and gasoline-fueled vehicles, J. Air Waste Manage. Assoc. 54, 1138–1150.