optimized separation of oc and ec for radiocarbon … · 2018-11-06 · ix summary particulate...
TRANSCRIPT
![Page 1: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/1.jpg)
Research Collection
Doctoral Thesis
Organic Aerosol Source Apportionment on Long-Term, Spatially-Dense Observation Networks Using Novel Mass SpectrometryTechniques
Author(s): Dällenbach, Kaspar Rudolf
Publication Date: 2017
Permanent Link: https://doi.org/10.3929/ethz-b-000238925
Rights / License: In Copyright - Non-Commercial Use Permitted
This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.
ETH Library
![Page 2: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/2.jpg)
DISS. ETH NO. 24200
Organic aerosol source apportionment
on long-term, spatially-dense observation networks
using novel mass spectrometry techniques
A thesis submitted to attain the degree of
DOCTOR OF SCIENCES of ETH ZÜRICH
(Dr. sc. ETH Zürich)
presented by
Kaspar Rudolf Dällenbach
MSc Environmental Sciences, ETH Zürich
born on 12.10.1988
citizen of Oberdiessbach, Switzerland
accepted on the recommendation of
Prof. Dr. Urs Baltensperger (examiner) Prof. Dr. Thomas Peter (co-examiner)
Dr. André Prévôt (co-examiner) Prof. Dr. James Schauer (co-examiner)
2017
![Page 3: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/3.jpg)
![Page 4: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/4.jpg)
iii
![Page 5: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/5.jpg)
iv
![Page 6: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/6.jpg)
v
Content
Summary ................................................................................................................................ ix
Zusammenfassung............................................................................................................... xiii
1 Introduction ........................................................................................................................ 1
1.1 Atmospheric aerosol-definition and properties .......................................................... 1
1.2 Aerosol climate impact .............................................................................................. 3
1.3 Aerosol health impact ................................................................................................ 6
1.4 Chemical composition and sources ............................................................................ 9
2 Motivation and thesis outline .......................................................................................... 13
3 Methodology ..................................................................................................................... 17
3.1 Aerosol Mass Spectrometer ..................................................................................... 17
3.1.1 General ........................................................................................................................... 17
3.1.2 offline AMS ..................................................................................................................... 19
3.2 Laser-Desorption/Ionization mass spectrometer ...................................................... 20
3.3 Source apportionment .............................................................................................. 21
4 Characterization and source apportionment of organic aerosol using offline aerosol
mass spectrometry ........................................................................................................... 25
4.1 Introduction .............................................................................................................. 26
4.2 Methods .................................................................................................................... 28
4.2.1 Aerosol sampling ............................................................................................................ 28
4.2.2 Offline AMS .................................................................................................................... 30
4.3 Results and discussion .............................................................................................. 35
4.3.1 Signal-to-noise, quantification and detection limits ....................................................... 35
4.3.2 OA recovery ................................................................................................................... 37
4.3.3 Mass spectral analysis ................................................................................................... 40
4.3.4 Source apportionment results ......................................................................................... 42
4.4 Conclusions .............................................................................................................. 54
![Page 7: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/7.jpg)
vi
5 Long-term chemical analysis and organic aerosol source apportionment at 9 sites in
Central Europe: source identification and uncertainty assessment ........................... 57
5.1 Introduction ............................................................................................................. 58
5.2 Methods ................................................................................................................... 60
5.2.1 Study area and aerosol sampling ................................................................................... 60
5.2.2 Offline AMS analysis ...................................................................................................... 62
5.2.3 Other chemical analysis ................................................................................................. 63
5.3 Source apportionment .............................................................................................. 63
5.3.1 General principle ........................................................................................................... 63
5.3.2 Preliminary PMF ........................................................................................................... 66
5.3.3 Sensitivity analysis ......................................................................................................... 67
5.3.4 Factor classification ...................................................................................................... 68
5.3.5 Recovery and blank corrections ..................................................................................... 68
5.3.6 Solution selection ........................................................................................................... 69
5.4 Results and discussions ........................................................................................... 72
5.4.1 Interpration of PMF factors ........................................................................................... 72
5.4.2 Uncertainty analysis ...................................................................................................... 80
5.4.3 Factor relative contribution at different sites ................................................................ 86
5.5 Conclusion ............................................................................................................... 88
6 Insights into organic-aerosol sources via a novel laser-desorption/ionization mass
spectrometry technique applied to one year of PM10 samples from nine sites in
central Europe ................................................................................................................. 89
6.1 Introduction ............................................................................................................. 90
6.2 Methods ................................................................................................................... 93
6.2.1 Sample collection and other chemical analysis .............................................................. 93
6.2.2 Laser-desorption/ionization ToF MS analysis ............................................................... 95
6.2.3 Source apportionment / PMF ......................................................................................... 98
6.3 Results ................................................................................................................... 100
6.3.1 Calibration, repeatability, and quantification.............................................................. 100
6.3.2 Combustion source profiles .......................................................................................... 103
6.3.3 Ambient samples .......................................................................................................... 104
6.3.4 Source apportionment results ...................................................................................... 106
6.4 Summary and conclusions ..................................................................................... 119
7 Conclusions and outlook ............................................................................................... 121
8 Bibliography .................................................................................................................. 129
![Page 8: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/8.jpg)
vii
9 List of Figures................................................................................................................. 161
10 List of Tables .................................................................................................................. 167
A Supplementary Material for chapter 5 Long-term chemical analysis and organic
aerosol source apportionment at 9 sites in Central Europe: source identification and
uncertainty assessment .................................................................................................. 169
B Supplementary material for chapter 6 Insights into organic-aerosol sources via a
novel laser-desorption/ionization mass spectrometry technique applied to one year
of PM10 samples from nine sites in central Europe ................................................... 181
11 Acknowledgement .......................................................................................................... 197
12 Curriculum vitae ............................................................................................................ 201
![Page 9: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/9.jpg)
viii
![Page 10: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/10.jpg)
ix
Summary
Particulate matter (PM), liquid or solid particles suspended in the atmosphere,
contributes significantly to air pollution. Besides negative acute and long-term health
effects, PM also affects the Earth’s climate by scattering and absorbing solar radiation
and also by changing cloud properties. A large part of PM consists of organic aerosol
(OA) which is a complex mixture of compounds. OA can be emitted directly as
particleswhich is called primary OA (POA) or can be formed as secondary OA (SOA)
through atmospheric aging of initially emitted volatile organic compounds (VOCs).
Important OA sources are traffic, cooking, combustion of biomass or fossil fuels but
also biogenic VOC emissions from terrestrial and marine ecosystems. For
understanding the many processes involved, measurements with high temporal
resolution are necessary. Similarly, for establishing mitigation goals and strategies also
long-term analyses in spatially dense observation networks are crucial.
In this dissertation, we used two mass spectrometers in a new way for long-term
analyses of OA in order to obtain a deeper understanding of its chemical composition
and the different contributing sources. The Aerodyne aerosol mass spectrometer (AMS)
is designed and used for short-term online measurement campaigns in order to
determine the contribution of the different OA sources. However, the complex
maintenance and the instrument cost hinder long-term and multi-site deployments. In
comparison, collecting PM on filters is easier and is routinely performed for
determining PM concentrations worldwide. In a first study, we developed an analytical
technique, based on analyzing the organic mass spectral signatures of water extracted
filter samples by AMS (termed offline AMS). By comparing offline AMS results to
reference methods like online deployments of AMS or the more robust aerosol chemical
speciation monitor (ACSM), we showed that the analyzed water-soluble OA (WSOA)
covers a large part of OA and also exhibits similar signatures as total OA. While highly
oxygenated organic compound classes seemed to be well represented, hydrocarbons are
strongly underestimated. A statistical source apportionment (SA) analysis pointed out
the value of such offline AMS data. In comparison to SA results from online ACSM
![Page 11: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/11.jpg)
x
analysis, we determined what proportion of the different OA sources was represented in
the offline AMS SA.
In a second study, we determined the contributions of different relevant OA
sources using the offline AMS technique. Thereby, we separated the influence of POA
sources in terms of traffic (HOA), cooking (COA), and biomass burning (BBOA) as
well as secondary OA categories and then validated our results by comparing them to
established state-of-the-art measurements. Moreover also a factor explaining organic
sulfur fragments of yet unknown origin was resolved. This factor is hypothesized to be
related to traffic activity. This dataset of source contributions is unique in its coverage,
both temporal and spatial (covering the entire year 2013 at 9 sites in central Europe).
Our results showed that the influence of BBOA was higher in winter than in summer
and exhibited enhanced concentrations at southern alpine valley sites than the sites in
northern Switzerland (no valley sites). The SOA was separated seasonally in a
component dominant in summer (summer oxygenated organic aerosol, SOOA) and
another component dominant in winter (WOOA). SOOA concentrations increased more
than linearly with rising temperatures, similar to biogenic VOC emissions and biogenic
SOA concentrations in other studies. This suggests that SOOA is strongly influenced by
biogenic emissions. In contrary, WOOA concentrations evolved similarly to
anthropogenically influenced inorganic compounds like ammonium (and also sulfate
and nitrate) suggesting a considerable anthropogenic influence on WOOA.
The benefit of AMS/ACSM data is strongly limited by the loss of chemical
information caused by the fragmentation of the analyzed molecules during the
measurement. This limitation hampers separating sources and is most prominent for
SOA because of the evolution of OA to similar chemical composition during
atmospheric aging. Therefore, we used in a third study in this dissertation a mass
spectrometer equipped with a laser-based ionization (laser-desorption/ionization mass
spectrometer, LDI) inducing less fragmentation. We developed and optimized an
analysis and data analysis framework for LDI analyses of PM collected on quartz-fiber
filters. This framework was applied to samples which were also analyzed by offline
AMS in order to obtain a deeper understanding of OA and its sources. In this study, we
apportioned OA to sources like traffic, wood burning, and also to secondary OA. In
comparison to offline AMS SA results of the same samples (2nd study), LDI SA
suggested higher contributions of primary sources (different response factors for
![Page 12: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/12.jpg)
xi
sources). The influence of traffic is separated into a part interpreted as primary tailpipe
exhaust and another part interpreted as either aged/secondary traffic emissions or
influenced by resuspended road dust. Since also the ratio of fossil organic to fossil
elemental carbon concentration (related to fossil fuel burning, 14C analysis) increased in
summer, the hypothesis of aged/secondary traffic emissions gains further trust.
Furthermore, we separated the influence of biomass burning in different components.
Based on a comparison of the factors’ mass spectral signature to laboratory wood
burning experiments, the factors could be related to the burning conditions. According
to these results, inefficient wood burning POA contributed more to the total wood
burning POA at the southern alpine sites than at the sites north of the alpine crest during
high-pollution winter-days, possibly contributing to the higher OA concentrations
attributed to primary wood burning emissions in the 2nd study.
![Page 13: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/13.jpg)
xii
![Page 14: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/14.jpg)
xiii
Zusammenfassung
Feinstaub, worunter man feste oder flüssige Partikel in Suspension in der
Atmosphäre versteht, trägt substantiell zur Luftverschmutzung in der Atmosphäre bei.
Nebst negativen akuten und Langzeit-Auswirkungen auf die menschliche Gesundheit
beeinflusst der Feinstaub auch das Klima der Erde, durch Streuung und Absorption des
Sonnenlichts und durch die Veränderung von Wolkeneigenschaften. Ein grosser Anteil
des atmosphärischen Feinstaubs (mit aerodynamischem Durchmesser <10µm) besteht
aus organischem Material (OM), einer komplexen Mischung von Molekülen
verschiedener Stoffklassen. Der organische Feinstaub kann direkt als primäres Aerosol
(POA) emittiert werden, oder durch atmosphärische Alterung sekundär aus anfänglich
gasförmigen organischen Stoffen gebildet werden (SOA). Wichtige Quellen sind
Verkehr, Kochen, Verbrennung verschiedener Energieträgern wie Holz und anderer
Biomasse nebst fossilen Brennstoffen, aber auch biogene Emissionen der marinen und
terrestrischen Ökosysteme. Zum Verständnis der Vielzahl involvierter Prozesse sind
Messungen mit einer hohen zeitlichen Auflösung nötig, jedoch sind Langzeitstudien auf
grossen Messnetzwerken ein Kernpunkt bei Bestimmung von
Emissionsreduktionszielen und Strategien zur Verbesserung der Luftqualität.
In dieser Dissertation wurden zwei verschiedene Flugzeitmassenspektrometer in
einer neuartigen Weise für Langzeitmessungen des organischen Feinstaubs zur
Erlangung eines besseren Verständnisses dessen chemischer Zusammensetzung und der
beitragenden Quellen verwendet. Das Aerodyne Aerosolmassenspektrometer (AMS)
wird online für kürzere Messperioden verwendet, um den Anteil verschiedener Quellen
des organischen Feinstaubs zu bestimmen. Jedoch machen es die komplexe Wartung
und der Instrumentpreis schwierig, Langzeitmessungen an verschiedenen Orten
durchzuführen. Vergleichsweise ist die Sammlung von Feinstaub auf Filtern einfacher
und wird weltweit routinemässig zur Bestimmung der totalen Feinstaubkonzentration
(PM) verwendet. Durch den erleichterten Zugang können auf solchen Filtern
Langzeitmessungen durchgeführt werden und zudem sind auch Orte zugänglich, die mit
![Page 15: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/15.jpg)
xiv
dem AMS/ACSM nicht erreichbar sind. Zunächst wurde in einer Studie eine analytische
Methode entwickelt, die auf Messungen der organischen massenspektralen Signatur von
so gesammeltem Feinstaub in wässriger Lösung mithilfe des AMS basiert (offline
AMS). Es wurde anhand von Vergleichen mit Referenzmethoden (AMS, ACSM)
gezeigt, dass der mit dieser Methode analysierte Feinstaubanteil (wasserlöslicher
organischer Feinstaub) einen Grossteil des organischen Feinstaubs abbildet und
grundsätzlich eine ähnliche chemische Zusammensetzung des organischen Anteils
aufweist. Jedoch sind gewisse organische Anteile weniger gut repräsentiert, speziell
unoxidierte Kohlenwasserstoffe im Verhältnis zu hochoxidierten Komponenten. Eine
statistische Quellenzuweisungsanalyse basierend auf offline AMS Daten, zeigte den
Nutzen solcher Daten auf und im Vergleich zu online ACSM Messungen wurde
ermittelt, welcher Massenanteil der verschiedenen aufgelösten Quellen mit der offline
AMS Methode abgebildet werden kann.
In einer Folgestudie wurden basierend auf der offline AMS-Methodik Anteile
verschiedener Feinstaubquellen bestimmt. Dabei wurde der Einfluss primärer
Emissionen des Verkehrs, Kochprozessen und Biomassenverbrennungen nebst
sekundären OA Kategorien bestimmt und mithilfe anderer etablierter Messungen
validiert. Zusätzlich wurde auch ein bis jetzt unbekannter Faktor aufgelöst, der
organische Schwefelfragmente erklärt, und mit der Verkehrsaktivität in Verbindung zu
stehen scheint. Der resultierende Datensatz ist sowohl in der zeitlichen (ganzes Jahr
2013) wie auch geographischen Auflösung (9 Stationen in Zentraleuropa) einzigartig.
Der Einfluss des primären Biomassenverbrennungsfeinstaubs war im Winter höher als
im Sommer und zeigte an den Stationen in den südlichen Alpentälern stark erhöhte
Konzentrationen im Vergleich zu den Stationen nördlich der Alpen (keine Tallagen
vorhanden). Der sekundäre OA-Anteil wurde saisonal aufgetrennt in eine Komponente,
die im Sommer (Sommer-oxidiertes organisches Aerosol, SOOA), und eine andere, die
im Winter (WOOA) dominant war. Die SOOA-Konzentrationen stiegen
überproportional mit steigenden Temperaturen an, ähnlich wie in anderen Studien auch
biogene VOC Emissionen und biogenes SOA. Dies legt den Schluss nahe, dass SOOA
stark von solchen biogenen Emissionen beeinflusst wird. WOOA hingegen zeigte ein
ähnliches zeitliches Verhalten wie Ammonium (und andere anthropogen beeinflusste
![Page 16: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/16.jpg)
xv
inorganische Aerosolkomponenenten wie Sulfat und Nitrat), was wiederum den Schluss
eines starken anthropogenen Einflusses zulässt.
Die Anwendung von AMS/ACSM-Daten ist durch die starke Fragmentierung und
den damit einhergehenden chemischen Informationsverlust limitiert. Aufgrund der
Tendenz, dass während der atmosphärischen Alterung Aerosole zu einer ähnlicheren
chemischen Zusammensetzung konvergieren, ist diese Limitation insbesondere für
SOA, aber auch generell für die Auftrennung chemisch ähnlicher Quellen hinderlich.
Aus diesem Grund wurde in dieser Dissertation auch ein Massenspektrometer mit laser-
basierter Desorption/Ionisation verwendet, welche durch eine sanftere
Ionisierungsmethode zu weniger Fragmentierung führt. Die Anwendung der im Rahmen
dieser Dissertation für die Anwendung auf Feinstaubproben (Quartzfilter) optimierten
Methode wurde auf den gleichen Datensatz wie die offline AMS Messungen
angewendet, um ein tiefergreifendes Verständnis der organischen Feinstaubs und seiner
Quellen zu erlangen. In dieser Studie wurde OA der Einfluss von OA-Quellen wie
Verkehr und Holzfeuerung nebst sekundärem OA aufgelöst. Der Vergleich mit offline
AMS Quellenzuweisungsresultaten legte dar, dass mit LDI-Daten höhere POA-
Konzentrationen als mit offline AMS-Daten abgeschätzt werden (verschiedene
Sensitivitäten zur Messung von Stoffklassen). Der Verkehrseinfluss konnte in einen
Teil, der mit primären Auspuffemissionen in Verbindung steht, und einen Teil, der als
entweder gealterte/sekundäre Auspuffemissionen oder aufgewirbelter Strassenstaub
interpretiert wurde, aufgetrennt werden. Da im Sommer auch das Verhältnis des fossilen
organischen Kohlenstoffs im Vergleich zum fossilen elementaren Kohlenstoff ansteigt
(aus der Verbrennung von fossilen Brennstoffen, 14C Analyse), erhält die Hypothese
gealterter/sekundärer Auspuffemissionen zusätzliches Gewicht. Ausserdem wurde der
mit Biomassverbrennungen einhergehende Feinstaub in Komponenten aufgetrennt, die
aufgrund ihrer massenspektrometrischen Signatur mit der Brenneffizienz in Holzöfen
(Laborexperimente) verbunden werden konnte. Während hochbelasteter Wintertagen
trugen gemäss dieser statistischen Analyse ineffiziente Feuerungen im Süden klar mehr
zum organischen Feinstaub aus Holzfeuerungen bei als an den Stationen nördlich der
Alpen. Diese ineffizienten Feuerungen könnten zu den höheren primären
Biomassenverbrennungs-Konzentrationen in den Alpentälern beitragen.
![Page 17: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/17.jpg)
xvi
![Page 18: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/18.jpg)
1
1 Introduction
1.1 Atmospheric aerosol-definition and properties
Aerosols are solid or liquid particles dispersed in a gas phase and are abundant in
the Earth’s atmosphere (summarized under the term particulate matter, PM), which are
emitted by natural and anthropogenic sources. These sources emit aerosol directly as
particles (primary particles) such as dust, pollen, fossil fuel combustion, wood
combustion. However, aerosol can also be formed by gas-phase oxidation of vapors and
subsequent condensation through new particle formation or particle growth from vapors
from e.g. fossil fuel and wood combustion, biogenic emissions (secondary).
According to the particle size PM is divided generally in a coarse fraction (>2.5
µm) and a fine fraction (<2.5 µm) which also have differing formation and removal
processes (Fig.1.1). The coarse fraction is dominated by resuspended dust, sea spray,
emissions from volcanoes, plant particles, and other emissions. A major sink for coarse
particles is dry deposition. The fine fraction can further be divided into accumulation
(0.1 to 2.5 µm), Aitken (10 to 100 nm), and nucleation mode (up to 10 nm). Particles in
the accumulation mode are either formed through coagulation or particle growth
through condensation of vapors on preexisting particles. For particles in the
accumulation range, removal mechanisms are slowest and thus they accumulate in the
atmosphere. Eventually they are deposited on the Earth’s surface by wet deposition (in-
cloud and below-cloud scavenging). Hot vapors produced during combustion processes
![Page 19: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/19.jpg)
Chapter 1: Introduction
2
condense and produce particles in the Aitken mode while the even smaller particles in
the nucleation mode are formed through gas-to-particle processes. These particles are
lost to a considerable extent through coagulation with accumulation mode particles.
Typical ambient particle number concentrations are dominated by particles in the
nucleation and Aitken mode, while for volume and mass concentrations larger particles
have an increased importance (Seinfeld and Pandis, 2006).
Figure 1.1: Idealized ambient particulate size distribution including formation and removal processes (from Finlayson-Pitts and Pitts, 2000).
![Page 20: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/20.jpg)
3
1.2 Aerosol climate impact
The Earth’s climate is governed by the global radiation budget which is in turn
influenced by the incoming solar radiation, as well as the Earth’s atmosphere and
surface. With the task to regularly assess the scientific understanding of parameters
influencing climate change, the Intergovernmental Panel on Climate Change (IPCC)
was created. The radiative forcing (RF) describes the net change of the global energy
balance (W/m2) caused by changes compared to pre-industrial conditions (after the
reference year 1750, Myhre et al., 2013). The RF is used to assess the magnitude of the
different factors contributing to climate change: a positive (negative) RF leads to more
(less) energy per surface and thus to a warming (cooling) effect of the factor in question.
In this comparison, the natural RF (change in solar irradiance) is small compared to the
total anthropogenic RF (Fig. 1.2).
Figure 1.2: Radiative forcing 2011 (1750 as reference for pre-industrial atmosphere) from Myhre et al. (2013).
Greenhouse gases such as e.g. CO2 strongly impact the radiative budget (RF of
1.68 W/m2), but also other atmospheric constituent have a considerable influence
(Myhre et al., 2013). The aerosol’s impact is separated in a direct aerosol-radiation-
interaction (ARI, related to scattering and absorption) and an indirect category (aerosol-
![Page 21: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/21.jpg)
Chapter 1: Introduction
4
cloud-interaction, ACI). Among the different factors, the uncertainty related to aerosol
is highest of which the biggest part stems from ACI.
Parameters important for ARI (Fig. 1.2) are the total number concentration,
optical properties, and also the particle size. On one hand, aerosol particles scatter light,
reducing the incoming solar radiation at the surface. On the other hand, besides
scattering, a fraction of the aerosol, e.g., black carbon, also absorbs radiation and emits
radiation leading to a positive RF in the respective atmospheric layer. Therefore, the
aerosol source contributions determining the chemical composition of the aerosol are
decisive whether the RF is positive or negative and also define the magnitude of the
effect (Fig. 1.3).
Figure 1.3: Global Radiative forcing of different ambient aerosol constituents evolving over time (from Myhre et al., 2013).
ACI refers to the effect of the aerosol on cloud properties. Higher aerosol
concentrations likely lead to a higher abundance of cloud condensation nuclei (CCN),
which leads to a higher number of cloud droplets lowering the albedo of the cloud
because of the limited amount of water in the atmosphere. Since precipitation is
moreover decreased, the lifetime of the cloud is, thereby, also increased making the
![Page 22: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/22.jpg)
5
effect lasting longer (Albrecht, 1989). However, the presence of black carbon reduces
the formation of clouds because of absorption of radiation and heating. The impact of
anthropogenic aerosol emissions is exemplified by the lower global radiation (Gilgen et
al., 1999) between 1960 and 1990 in e.g., Potsdam (Germany) coinciding with increased
sulfur emissions (Stern et al., 2005) which seemingly also impacted temperature
(Brohan et al., 2006) as illustrated in Fig. 1.4.
Figure 1.4: Surface solar radiation in Potsdam, Germany, (GEBA, Gilgen et al., 1999) and sulfur emissions in Europe (Stern et al., 2005) temperature anomalies over Europe from 1930 to 2010, and annual temperature anomaly in Europe over land (Brohan et al., 2006).
In estimating the anthropogenic (industrial) impact on the radiative forcing, both
an accurate representation of the present and past atmosphere is crucial (Carslaw et al.,
2013). While pre-industrial trace gas levels such as CO2 are largely well defined e.g. by
studying climate archives like ice cores , the pre-industrial aerosol burden is uncertain
(Myhre et al., 2013, Masson-Delmotte et al., 2013, Gordon et al., 2016). Recent results
suggest that under pre-industrial conditions pure biogenic new particle formation, which
can act as CCN after growth, is more pronounced than previously thought and,
therefore, the change in RF caused by aerosol has been overestimated by 27% (0.22
W/m2, Gordon et al., 2016).
![Page 23: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/23.jpg)
Chapter 1: Introduction
6
1.3 Aerosol health impact
Figure 1.5: Mortality and SO2 concentrations during the great London smog (adapted from Bell et
al., 2001 and completed with data from the same publication).
With the great London smog in December 1952, the adverse health effects of air
pollution including particulate matter became apparent to a larger public. The great
London smog in December 1952, caused 5000 excess deaths (acute) and as a long-term
effect further 7000 excess deaths between January and March 1953 which could not be
explained by influenza (Bell et al., 2001). Also sensitivity studies on the impact of
influenza underline the fact that only a fraction of excess deaths can be attributed to this
cause (Bell et al., 2004). These findings triggered further research on the health impact
of PM, especially so-called cohort studies assessing the effect of chronic exposure. In
such studies, mortality was found to scale linearly with chronic particle exposure levels
(e.g., a study period of 14-16 years and a PM concentration between 10 and 30 µg/m3 in
Dockery et al., 1993). A reduction of PM levels in the study region led to a lower
mortality (as described in Laden et al., 2006). Such epidemiological studies relate also
![Page 24: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/24.jpg)
7
an increased risk of cardiopulmonary diseases as well as lung cancer to long-term
exposure of PM (Laden et al., 2006, Pope et al., 2006, 2013).
Figure 1.6: Global distribution of mortality attributed to air pollution (from Lelieveld et al., 2015).
On a global scale, ambient air pollution is estimated to have caused 3.7 million
premature deaths in 2012 (WHO, 2013) and in agreement a recent modelling study links
3.15 million deaths to air pollution, mostly to PM2.5 (PM with an aerodynamic
diameter smaller than 2.5 µm) and ozone. These estimates do however not include the
impact of indoor air pollution causing additional deaths of 4.4 million (Lim et al.,
2013). Environmental regulations regarding particulate air pollution aim to limit the
impact of aerosol on human health by establishing limits in the PM10 (smaller than
10µm) and also other size fractions. These limits consider acute and chronic effects by
separately specifying daily and annual limits (Tab. 1.1).
Table 1.1: PM limits established in Switzerland, EU, U.S.A., and WHO. Country/Organization PM10 daily PM10 annual PM2.5 daily
PM2.5 annual *primary/**secondary
Switzerland (FOEN, 2010)
50 20 - -
EU (European Commission, 2010)
50 40 - 25
US EPA, 2012 150 - 35 12*/15** WHO (2006) 50 20 25 10
![Page 25: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/25.jpg)
Chapter 1: Introduction
8
Given the number of fatalities, the link between the toxicity and the chemical
composition and size of the particles is important, but not yet well understood. Exposure
to PM occurs through inhalation, but not all inhaled particles are deposited in the human
body, e.g., more than 80% of the inhaled 300 nm particles are exhaled again. The size of
the particle does not only influence the overall deposition probability but also the
location of the deposition (Fig. 1.7, Maynard and Kuempel, 2005). While coarse
particles are already deposited in the nose region, smaller particles also can travel as far
as to the tracheobronchial region. The even smaller ultrafine particles even penetrate
into the alveolar region and can pass the air-blood barrier (Kelly and Fussel, 2012).
However, even if coarse particles are mostly deposited in the head region, there is
increasing evidence for adverse health effects of such particles (Brunekreef and
Forsberg, 2005). In order to describe the effect of the chemical composition on human
health, the particles’ ability to create oxidative stress and thus to cause inflammations
and cell death is commonly used (Kelly and Fussel, 2012). Therefore, reactive oxygen
species are a focus in current research aiming to understand the health effects of PM
(Seifried et al., 2006). Among such studies, apportioning ROS to its sources is given
special attention (Zhang et al., 2008, Verma et al., 2015). However, the health effects of
single sources remain unclear and especially the impact organic carbon is uncertain due
to its complex composition (Stanek et al., 2011, Kelly and Fussel, 2012).
![Page 26: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/26.jpg)
9
Figure 1.7: Modelled particle deposition probability upon inhalation in the human respiratory
system as a function of particle size and location (from Maynard and Kuempel, 2005).
1.4 Chemical composition and sources
Aerosol sources have distinctly different composition and their contributions to
PM exhibit high temporal and spatial variability which also depends on the assessed
size fraction. This variability influences also the bulk composition of ambient PM (Fig.
1.8, Putaud et al., 2004). PM10 is composed of organic aerosol (OA), black carbon
(BC), and inorganic aerosol (ammonium, nitrate, sulfate, sea salt, mineral dust).
![Page 27: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/27.jpg)
Chapter 1: Introduction
10
Figure 1.8: Aerosol composition for different environments in Europe in PM10 and coarse PM, with unaccounted mass from water and other aerosol constituents which were not analyzed at all sites (adapted from Putaud et al., 2004).
SO2 (emitted by fossil combustion and volcanic eruptions) and dimethylsulfide
(DMS, C2H6S, by marine microorganisms) are oxidized in the atmosphere and form
sulfuric acid. NOx, emitted during combustion processes, similarly undergoes
atmospheric processing leading to the formation of nitric acid. Ammonia, emitted
mainly by agricultural processes (livestock and use of fertilizers, according to
Bouwman et al., 1997), reacts with sulfuric and nitric acid to ammonium sulfate and
nitrate (secondary aerosol formation). In contrast, BC, mineral dust, and sea salt are
emitted as particles (primary aerosol) and OA can be both primary and secondary.
Increasing PM10 and elevated concentrations of combustion related aerosol components
such as BC reflect the anthropogenic influence (through fossil fuel and biomass
combustion) in urban areas (urban and kerbside, Fig. 1.8). A similar behavior can be
observed for the coarse fraction (PM10-PM2.5) at kerbsides, where PMcoarse is
dominated by mineral dust from resuspension caused by traffic while the concentrations
are lower in rural areas or even dominated by sea salt close to the sea (Visser et al.,
2015). However, a large fraction of OM contributing to PMcoarse might stem from
plant abrasion (Andreae et al., 2007).
![Page 28: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/28.jpg)
11
Figure 1.9: NR-PM1 AMS analyses at multiple sites on the northern hemisphere: a) aerosol composition (organic (Org), sulfate, nitrate, ammonium, chloride), b) oxygenated OA concentrations, c) and primary OA factors as HOA and BBOA (from Zhang et al., 2011).
The composition of the non-refractory-PM1 (NR-PM which is smaller than 1 µm)
is commonly analyzed with the aerosol mass spectrometer (AMS, Canagaratna et al.,
2007, described in Chapter 3). The contribution of OA to PM1 is highest (45%), though
also sulfate (32%), ammonium (13%), and nitrate (10%) contribute substantially (Zhang
et al., 2007). The spatial variability of the NR-PM1 composition as well as the
contributing OA sources along a rural-to-urban gradient is displayed in Fig. 1.9. On
average OA contributes 45% to NR-PM1 at the urban sites, 52% downwind from urban
sites, and 43% at rural sites. The organic aerosol is a complex mixture of compounds of
which only a small fraction is identified (Hoffmann et al., 2011). Anthropogenic sources
such as traffic (Schauer et al., 2002a, Platt et al., 2014, Chirico et al., 2011), cooking
(Schauer et al., 2002b, Klein et al., 2016a, 2016b), residential heating using different
fuels such as wood (Schauer et al., 2001, Bruns et al., 2015 and 2016, Heringa et al.,
![Page 29: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/29.jpg)
Chapter 1: Introduction
12
2011) and natural sources such as the terrestrial (Sanchez-Ochoa et al., 2007; Leaitch et
al., 2011, Guenther et al., 1995) and marine environment (Schmale et al., 2013; Zorn et
al., 2008; Crippa et al., 2013c, Seinfeld and Pandis, 2006) contribute to OA by primary
particle emission (POA) as well as the emission of volatile organic compounds (VOC).
VOCs undergo atmospheric aging (radiation, oxidants) leading to the formation of
lower-volatility compounds which subsequently condense on preexisting particles
(SOA). Source apportionment techniques applied to organic mass spectral signatures
from the AMS enable to quantitatively disentangle the contribution of different POA
sources (traffic, cooking, biomass combustion, coal combustion, e.g. in Elser et al.,
2016b) as well as oxygenated OA (OOA) categories typically linked to SOA (Zhang et
al., 2011; Jimenez et al., 2009). POA and SOA levels increase from rural to urban sites
with increasing anthropogenic influence though the relative enhancement for POA is
higher (from POA/OA of 0.1 at rural to 0.42 at urban sites, Fig. 1.9, Zhang et al., 2011).
![Page 30: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/30.jpg)
13
2 Motivation and thesis outline
The atmospheric aerosol impacts climate and human health (IPCC, 2013; Pöschl
et al., 2005). Since the organic aerosol contributes significantly to PM (up to 90% of
PM1, Jimenez et al., 2009), a detailed characterization and a profound understanding of
the contributing sources and formation processes is essential. OA is a complex mixture
of which only a small proportion of all compounds is identified (Hoffmann et al., 2011).
Estimating source contributions has, therefore, been a focus of numerous studies (Zhang
et al., 2011). Such information enables reducing uncertainties in climate modelling and
designing effective mitigation strategies. However, the lack of long-term analyses limits
developing and testing of tropospheric aerosol prediction models. For obtaining a better
understanding of the health effects of OA, among others, long-term source
apportionment analyses in large networks coupled with epidemiological studies will be
of great value.
In earlier studies, different analytical methods were applied on filter samples.
Organic carbon (OC) was separated into a fossil (anthropogenic fossil fuel emissions)
and non-fossil (wood combustion, cooking biogenic, primary biological, etc)
contribution using 14C analyses (Szidat et al., 2006; Cavalli et al., 2010; Zotter et al.,
2014). However, this approach cannot separate the influence of POA and SOA. The
contribution of specific sources is also estimated based on marker concentrations (e.g.
levoglucosan) using scaling factors to obtain the total contribution of the source. This
approach has the drawback that these scaling factors are uncertain. Combining OC/EC
![Page 31: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/31.jpg)
Chapter 2. Motivation and thesis outline
14
analyses with several organic and inorganic markers is also used for statistical
attribution of OC to its sources (e.g. Waked et al., 2014). With this approach only a
small fraction of OA or OC is resolved and the rest apportioned to sources based on
their variability. Thus, sources not represented by the chosen markers are not resolved.
With the invention of the AMS (DeCarlo et al., 2006, Jayne et al., 2000) highly
time-resolved quantitative mass spectral signatures of NR-OA became available for
source apportionment. However, molecular identification is hindered by electron impact
ionization (EI) and the resulting strong fragmentation. With the help of statistical tools
(e.g., positive matrix factorization, PMF, Paatero and Tapper, 1994) and comparison to
emission studies, the influence of different sources could be separated. Numerous
studies (e.g. Fig. 2.1 from Jimenez et al., 2009) provided quantitative estimates of
primary OA source contributions such as traffic related hydrocarbon emissions (HOA),
biomass burning emission (BBOA) as well as of secondary OA (OOA).
Figure 2.1: NR-PM1 mass concentrations (organic, sulfate, nitrate, ammonium, and chloride) measured by AMS at multiple sites. Source apportionment results of OA are displayed: traffic (HOA), other POA as BBOA (other OA), oxygenated OA linked to SOA (OOA), separated into semi-volatile (SV-OOA) and low-volatility (LV-OOA) from Jimenez et al. (2009).
![Page 32: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/32.jpg)
15
However, the AMS is costly and demanding in maintenance and, therefore,
measurement campaigns are typically limited to 1 to 2 months. For this reason, the
aerosol chemical speciation monitor (ACSM, Aerodyne Inc., Ng et al., 2011c, Fröhlich
et al., 2013) was developed for long-term measurements but in order to cover large
networks, still many instruments are required. Both ACSM and AMS are restricted to
the submicron aerosol defined by the particle transmission through the aerodynamic
lens. Recently a new aerodynamic lens was developed accessing particles of up to a size
of 2.5µm (Williams et al., 2013), but PM10 is yet not accessible. This type of
instrument recovers as mentioned quantitative mass spectral signatures of NR-OA but
yet loses a lot of chemical information. For this reason and due to the evolution of the
organic matter to chemically similar composition during atmospheric ageing,
information on the origin of OOA is limited. In contrast to the AMS, laser-
desorption/ionization mass spectrometry fragments the molecules less and, according to
Samburova et al. (2005a), the molecular weight distribution of water extract PM is not
significantly influenced by fragmentation by LDI and thus adding a matrix assisting the
ionization does not lead to a gain in information. The same study also shows the
possibility to analyze PM collected on filters directly without further treatment. On the
other hand, this instrument is not quantitative because of matrix effects, i.e. the
measured intensity of a compound depends not only on its concentration but also on the
concentration of all other present compounds. With this instrument, earlier studies have
shown that it might be possible to distinguish anthropogenic from biogenic SOA
(Baltensperger et al., 2005). Kalberer et al. (2004) discovered the formation of
oligomers to the organic aerosol in a reaction chamber and concluded that such
oligomers are also important in ambient OA. These results indicate that LDI
measurements capture valuable information on the chemical composition of OA.
In this dissertation, we use different approaches to overcome the above
limitations. On one hand, we develop and characterize an offline application of the
AMS based on liquid extracts of PM on filter samples that can cover long-term
measurements in large networks (Chapter 4). This approach is used to quantify the
sources contributing to the organic aerosol in the PM10 size fraction in Switzerland at 9
sites throughout the year 2013 (Chapter 5). On the other hand, we use offline laser-
![Page 33: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/33.jpg)
Chapter 2. Motivation and thesis outline
16
desorption/ionization mass spectrometry for obtaining a better understanding of the
contributing sources and their chemical characteristics. This measurement approach is
described and applied to the same ambient samples as those analyzed by offline AMS
(Chapter 6).
![Page 34: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/34.jpg)
17
3 Methodology
In this dissertation, we characterize the chemical composition of the ambient
organic aerosol using mainly two different mass spectrometers: the Aerodyne aerosol
mass spectrometer and a laser-desorption/ionization mass spectrometer. In this chapter
both instruments are described in more detail along with a detailed explanation of the
statistical technique employed for source apportionment based on the mass spectral
signatures.
3.1 Aerosol Mass Spectrometer
3.1.1 General
The Aerodyne aerosol mass spectrometer (AMS, Fig. 3.1) analyzes the non-
refractory particulate matter (NR-PM) online and provides quantitative measurements
of organics, ammonium, nitrate, sulfate, and chloride, as well as mass spectral
signatures of the organic aerosol and particle size (Canagaratna et al., 2007).
![Page 35: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/35.jpg)
Chapter 3: Methodology
18
Figure 3.1: Scheme of the high-resolution-time-of-flight aerosol mass spectrometer (HR-ToF-AMS) from DeCarlo et al. (2006).
The instrument is composed of three parts: the aerosol inlet, the particle sizing
chamber, and the particle compositon detection section (DeCarlo et al., 2006). PM
suspended in air is sucked through a critical orifice and focused into a particle beam by
an aerodynamic lens and released into the sizing and detection region. The particles are
sized by measuring the particle time-of-flight using a chopper regulating the
transmission of the particle beam with up to 4% transmission rate (chopped). The
background and air signal is separated from the particle signal by either blocking
completely the particle beam (chopper blocked) or transmitting it continuosly (chopper
open). When impacting on the heated tungsten surface (600°C, vaporizer), the particles
are flash-vaporized. Thus refractory material as e.g., black carbon, are not accessible to
the AMS. The vaporized compounds are ionized by electron impact (EI). This hard
ionization pathway causes significant fragmentation. The charged fragments are
transferred by orthogonal extraction to a time-of-flight unit and detected. The AMS has
a mass resolution of 2’500 (at m/z 43, V-mode, Canagaratna et al., 2007) which enables
separating different ions at the same nominal mass (e.g. C2H3O+ and C3H7+ at m/z 43,
Fig. 3.2). With this information the bulk O/C or OM/OC (organic matter to organic
carbon ratio) can be determined.
![Page 36: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/36.jpg)
19
Figure 3.2: Example of mass resolution of 3 AMS types: Quadrupole, C-ToF-MS, HR-ToF-MS from De Carlo et al. (2006).
3.1.2 offline AMS
In several studies, the AMS, developed as an online instrument, is used as an
offline application for analyzing nebulized liquid extracts of filter samples (Lee et al.,
2011; Sun et al., 2011; Mihara and Mochida, 2011). In this study, we further develop
and characterize this method (offline AMS, Chapter 4) and apply it to environmental
filter samples. The technique involves extracting PM in ultrapure water (20 min in
ultrasonic bath at 30°C and 1 min vortexing) and filtering in order to remove suspended
material. The aqueous solution is nebulized, dried and analyzed by an HR-ToF AMS.
The mass spectral signature of the organic fraction represents WSOM which is of
advantage because the signatures can be related to quantitative WSOC analyses which is
not possible using other solvents. Therefore, the non-quantitative organic offline-AMS
signatures can be quantified using the product of the WSOM/WSOC ratio from the
offline-AMS analyses and the water-soluble OC (WSOC) measurements from a total
OC (TOC) analyzer quantitatively oxidizing WSOM to CO2 which is detected by a non-
dispersive infrared (NDIR) detector. In analogy to online AMS OM signatures, the
offline WSOM signatures are used for source apportionment. The characterization of
the method and performance of offline AMS data in source apportionment in
![Page 37: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/37.jpg)
Chapter 3: Methodology
20
comparison to online ACSM data is assessed in Chapter 4. Main advantages and
disadvantages of this method are lined out in Tab. 3.1.
Table 3.1: Advantages and disadvantages of offline AMS technique.
Advantage Disadvantage
• High temporal and spatial
coverage.
• Low time resolution (defined by
filter sampling time).
• Coarse OM also accessible (if water soluble).
• Filter artifacts.
• Analysis of any nebulizable matrix (e.g. lake sediments, snow, rain, soil water, etc.) possible
• External data needed for quantification (WSOC and/or OC).
• Complementary analysis a posteriori with required external data to study the aerosol sources.
• Low recoveries of less water-soluble sources and thus large uncertainties (as e.g. for traffic emissions).
3.2 Laser-Desorption/Ionization mass spectrometer
Laser-desorption/ioniziation mass spectrometry uses a laser for desorbing liquid
or solid material and producing ions (De Hoffmann and Stroobant, 1999). The laser
emits pulses directly focused on the sample surface with a defined frequency. These
laser pulses create first generation ions and neutral molecules by ablating material from
the sample surface. In the dense plume, ions and neutral molecules react and form
second and further generation ions. Even if the ionization pathway is softer than EI (as
mentioned above), the formed ions are typically fragments of the molecules (De
Hoffmann and Stroobant, 1999). The created ions are in a next step analyzed by a time-
of-flight mass spectrometer. In this study, we use an LDI (Shimadzu Axima
Confidence, Shimadzu-Biotech Corp., Kyoto, Japan) equipped with an N2 laser (λ=337
nm) and a time-of-flight mass spectrometer (axial, without orthogonal injection). Laser-
based ionization instrumentation is known to be affected by matrix effects (Ellis et al.,
2014, Borisov et al., 2013). This means that the recorded intensity of a molecule does
not only depend on its concentration but also on the concentration of all other
compounds present. In complex mixtures, such as ambient PM, not all of these
![Page 38: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/38.jpg)
21
dependencies can be understood. Therefore, this type of analysis is not considered
quantitative. However, using online laser-based instrumentation (ATOF-MS), good
correlations with reference analyses analyses could be achieved for OA and different
inorganic components (Healy et al., 2013).
Earlier studies showed that PM collected on filter samples can be analyzed
without matrix addition (Samburova et al., 2005a) and related this observation to the
possible presence of compounds acting as a matrix in PM. In this study, we apply LDI
measurements directly to PM collected on quartz filter samples in order to avoid losses
that may occur during an intermediate extraction step. Prior to the analysis, aqueous
solution of silver nitrate (SN) is applied to fractions of the sample as an internal
standard for m/z calibration. The dry sample on a stainless steel plate is inserted into the
LDI and analyzed. In Chapter 6, the method development and characterization is
described as well as an application to ambient samples. Table 3.2 presents main
advantages and disadvantages of the offline LDI approach.
Table 3.2:Advantages and disadvantages of LDI technique.
Advantage Disadvantage
• Softer ionization than EI. • High temporal and spatial
coverage.
• Not quantitative. • Matrix effects. • Unknown response factors of
aerosol components. • Low time resolution (defined by
filter sampling time).
• Coarse OM also accessible. • Filter artifacts. • Analyzing any material responding
to laser. • External data needed for
quantification (WSOC and/or OC). • Complementary analysis a
posteriori with required external data to study the aerosol sources.
3.3 Source apportionment
OA is a complex mixture with a multitude of contributing primary and secondary
aerosol sources. Source apportionment aims to identify and quantify sources
![Page 39: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/39.jpg)
Chapter 3: Methodology
22
contributing to OA. Bilinear receptor models, e.g. positive matrix factorization (PMF,
Paatero and Tapper, 1994), are widely used to apportion OA to its sources. PMF
explains the variability in a chosen dataset (x) with linear combinations of factor
contributions varying over time (g) and constant factor profiles (f). The index i
represents the measurement time, j the m/z or ion and k the factor number and the model
residuals are described by eij.
(3.1)
The entries of g and f are constrained to be positive, as also the name PMF
specifies. In OA applications based on AMS data, x reflects measured mass spectra, f
the chemical composition of the source, g the time series of the source and eij not
resolved/residual mass. In solving Eq. 3.2, also the measurement uncertainties (sij) of
the mass spectra are considered by minimizing eij/sij, through the object function Q:
(3.2)
PMF can be implemented by the multilinear engine 2 (ME-2, Paatero, 1999)
which allows exploration of the solution space. This is possible through using a priori
information on e.g. chemical composition of sources. In the so called a-value approach,
it is possible also to allow the solver to deviate from the a priori information defined by
the relative deviation, a (Eq. 3.3).
(3.3)
In Eq. 3.3 n and m represent any two entries (m/z, ion) in the constrained factor
profile. Different representations of x by PMF, rotations of the linear combinations of f
and g, can have a similar mathematical quality (Q) which is referred to as rotational
ambiguity. Therefore, it is crucial to explore the solution space in order to estimate the
model uncertainty. Uncertainty in PMF results also arise from random fluctuations in
the input data (measurement), which is termed statistical uncertainty. Resampling the
![Page 40: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/40.jpg)
23
input matrices in order to assess the influence of the choice of the model input on the
PMF solution, called bootstrapping, is an efficient approach to estimate the statistical
uncertainty (Paatero et al., 2014, general atmospheric description in Wilks, 2006).
Other statistical tools like principal component analysis (PCA, described in detail
for example in Wilks, 2006), chemical mass balance (CMB, Watson et al., 1998),
cluster analysis (detailed in Wilks, 2006) have similar goals. CMB requires a priori
information on the source profiles (chemical signature) and their uncertainty (Watson et
al., 1998). In general, whenever a priori information is needed, knowledge on the related
uncertainty is crucial. In contrast, PMF does not require a priori information on the
chemical source signature although such information can be used optionally. Moreover,
PMF uses also the measurement uncertainty (see Eq. 3.2). Thus datapoints are given
weight related to their respective uncertainty in the partial least square method. Thereby,
especially outliers with a high uncertainty are given less weight. In comparison to PMF,
CMB requires smaller datasets since it can be theoretically solved for any single
datapoint. CMB, PMF, and PCA separate the influence of different sources rather than
grouping similar spectra as for example cluster analysis (Wilks, 2006). Unlike CMB,
PMF (without a priori information) is limited in separating highly correlated time
sources. However, by using a priori information (as possible using ME-2 as a PMF
implementation), separating such sources is facilitated (Paatero, 1999). As stated earlier,
PMF outputs are constrained to be positive which makes the factorizations (g) directly
interpretable as concentration. This is not the case for other related techniques such as
PCA. Also PCA assumes orthogonal components/factors which is hindering the
separation of highly correlated components, and negative values also hinder the
interpretation in an environmental sense as concentrations. However, PCA is of great
value to assess the influence of environmental parameters as e.g., temperature on PM
levels. CMB is traditionally used to separate different POA and SOA was treated as a
residual. Nevertheless, there were some attempts to use similar approaches using
markers of secondary products of defined precursors to estimate their SOA contribution,
e.g., aromatics, alpha-pinene, isoprene (Kleindienst et al., 2007). Moreover, when
performing PMF using AMS data, SOA is separated by its variability and not modelled
as a residual because the bulk chemical composition and not only selected markers are
used. However, the use of markers can also be advantageous for resolving correlating
POA due to of their high specificity.
![Page 41: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/41.jpg)
Chapter 3: Methodology
24
![Page 42: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/42.jpg)
25
4 Characterization and source
apportionment of organic aerosol using offline aerosol mass spectrometry
Kaspar R. Daellenbach1, Carlo Bozzetti1, Adela Křepelová1, Francesco Canonaco1, Robert Wolf1, Peter Zotter1,a, Paola Fermo4, M. Crippa1,b, Jay G. Slowik1, Yuliya Sosedova1, Yanlin Zhang1,2,3,c, Ru-Jin Huang1, Laurent Poulain5, Sönke Szidat2, Urs Baltensperger1, Imad El Haddad1, and André S. H. Prévôt1 1Laboratory of Atmospheric Chemistry, Paul Scherrer Institute, 5232 Villigen, Switzerland 2Department of Chemistry and Biochemistry & Oeschger Centre for Climate Change Research, University of Bern, 3012 Bern, Switzerland 3Laboratory of Radiochemistry and Environmental Chemistry, Paul Scherrer Institute, 5232 Villigen PSI, Switzerland 4Department of Chemistry, University of Milan, 20133 Milan, Italy 5Leibniz Institute for Troposphärenforschung, Leipzig, Germany anow at: Lucerne School of Engineering and Architecture, Bioenergy Research, Lucerne University of Applied Sciences and Arts, 6048 Horw, Switzerland bnow at: EC Joint Research Centre, Institute for Environment and Sustainability, 21027 Ispra, Italy cnow at: Yale-NUIST Center on Atmospheric Environment, Nanjing University of Information Science and Technology, Nanjing 10044, China
Pulished in Atmospheric Measurement and Technique DOI: 10.5194/amt-9-23-2016
![Page 43: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/43.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
26
Abstract
Field deployments of the Aerodyne Aerosol Mass Spectrometer (AMS) have
significantly advanced real-time measurements and source apportionment of non-
refractory particulate matter. However, the cost and complex maintenance requirements
of the AMS make its deployment at sufficient sites to determine regional characteristics
impractical. Furthermore, the negligible transmission efficiency of the AMS inlet for
supermicron particles significantly limits the characterization of their chemical nature
and contributing sources. In this study, we utilize the AMS to characterize the water-
soluble organic fingerprint of ambient particles collected onto conventional quartz
filters, which are routinely sampled at many air quality sites. The method was applied to
256 particulate matter (PM) filter samples (PM1, PM2.5, and PM10, i.e., PM with
aerodynamic diameters smaller than 1, 2.5, and 10 µm, respectively), collected at 16
urban and rural sites during summer and winter. We show that the results obtained by
the present technique compare well with those from co-located online measurements,
e.g., AMS or Aerosol Chemical Speciation Monitor (ACSM). The bulk recoveries of
organic aerosol (60-91 %) achieved using this technique, together with low detection
limits (0.8 µg of organic aerosol on the analyzed filter fraction) allow its application to
environmental samples. We will discuss the recovery variability of individual
hydrocarbon ions, ions containing oxygen, and other ions. The performance of such data
in source apportionment is assessed in comparison to ACSM data. Recoveries of
organic components related to different sources as traffic, wood burning, and secondary
organic aerosol are presented. This technique, while subjected to the limitations inherent
to filter-based measurements (e.g., filter artifacts and limited time resolution) may be
used to enhance the AMS capabilities in measuring size-fractionated, spatially resolved
long-term data sets.
4.1 Introduction
Aerosols affect climate, air quality, ecosystems, and human health (Kelly et al.,
2012; Griggs and Noguer, 2002). Organic aerosol (OA), a significant fraction of the dry
particle mass (Jimenez et al. (2009) and references therein), is either directly emitted
(primary organic aerosol, POA), or formed in the atmosphere through gas-phase
oxidation of anthropogenic and biogenic volatile organic compounds and subsequent
![Page 44: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/44.jpg)
27
condensation or nucleation (secondary organic aerosol, SOA). Characterization of OA
chemical composition and sources is necessary for understanding the corresponding
atmospheric processes and mitigating the adverse effects of aerosols. Previous studies
have shown that OA contains a variety of organic species, including hydrocarbons,
alcohols, aldehydes, and carboxylic acids. However, only about 10- 30 % of OA has
been chemically speciated so far (Hoffmann et al., 2011).
The High-Resolution Time-of-Flight Aerosol Mass Spectrometer (HR-ToF-AMS,
Aerodyne Research, Inc.) has been widely used for characterizing OA in both field and
laboratory studies. This instrument couples thermal vaporization with electron
ionization (EI, 70 eV) and provides quantitative mass spectra of non-refractory aerosol
components including OA, NH4+, NO3
- , SO42-, and Cl-. Application of advanced factor-
based receptor models such as positive matrix factorization (PMF, Paatero and Tapper,
1994) to these spectra has been proven effective in apportioning OA into different
factors (e.g. Lanz et al., 2007,2010; Jimenez et al., 2009; Zhang et al., 2011; Ng et al.,
2010b; Crippa et al., 2014). These factors are subsequently related to primary sources
like biomass burning (BBOA), traffic (HOA), and cooking (COA), as well as
oxygenated organic aerosol (OOA), which is often attributed to SOA.
The cost and complex operation required by the AMS makes its simultaneous
long-term deployment at many sites impractical. Consequently, current data sets are
typically limited to few weeks and specific sites or measurements from mobile
platforms (Mohr et al., 2011). Recently, a robust, less expensive, Aerosol Chemical
Speciation Monitor (ACSM, Ng et al., 2011c) and a time-of-flight aerosol chemical
speciation monitor (ToF-ACSM, Fröhlich et al., 2013, 2015) were developed to
overcome some of these shortcomings. However, the low mass resolution of these
instruments reduces their utility. Meanwhile, other studies have proposed the use of the
HR-ToF-AMS for the analysis of aqueous or organic solvent extracts of filter samples,
which are already routinely collected at many sites worldwide, offering a greater
coverage than with ACSMs (Mihara and Mochida, 2011; Lee et al., 2011; Sun et al.,
2011). While such methodologies may greatly extend the ability of the AMS to measure
spatially re-solved long-term data sets, the results obtained only pertain to a sub-fraction
of the total organic aerosol and are subject to inherent artifacts of filter-based
![Page 45: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/45.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
28
measurements. It is not clear whether this fraction adequately reflects the chemical
nature of the entire bulk OA and whether these results may be used for OA source
apportionment. Here, we have adopted such an approach based on measurements of the
water-soluble organic fraction. We present a methodology to generalize the results to
bulk OA, based on the analysis of 256 filter samples from 16 urban to rural sites during
different seasons and its comparison to online measurements. These results are expected
to significantly broaden the spatial, temporal, and particle size ranges accessible to
AMS measurements of organic aerosol.
Table 4.1: Filter samples and available supporting measurements used in this study. Location Campaing period Sampling
period (h)
Samples Size Supporting measurements
Zürich
(urban
background)
Apr 2011 12 11 PM1 PM2.5 fingerprints and OA,
SO42-a, gas-phase
measurements (CO)
Aug 2008-
Jul 2009
24 42 PM10 OC/EC, ions
Feb 2011-
Feb 2012
24 41 PM10 PM1 fingerprints and OA,
SO42-b, eBC, WSOC
Paris (urban core) Jul 2009
Jan-Feb 2010
12 12 PM1 OC/EC, PM1 fingerprints
and OA, SO42-a
15 NABEL
stations
in Switzerlandc
Dec 2007-
Feb 2008
Dec 2008-
Feb 2009
24 150 PM10 OC/EC, ions
aHR-ToF-AMS; bQuadrupole ACSM; cNABEL (Swiss National Air Pollution Monitoring Network): the stations represented in the study are Basel, Bern, Chiasso, St. Gallen, Magadino, Massongex, Moleno, Payerne, Reiden, Roveredo, Sissach, Solothurn, San Vittore, Vaduz, and Zürich Zotter et al. (2014).
4.2 Methods
4.2.1 Aerosol sampling
Particulate matter (PM) of different sizes (PM1, PM2.5, and PM10, representing
aerodynamic particle sizes smaller than 1, 2.5, and 10 µm, respectively) were collected
onto pre-heated (800 °C, 12 h) Pall quartz filters (diameter 14.7 cm) using HiVol
samplers (500 Lmin-1). Field blanks were collected using the same method as for the
![Page 46: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/46.jpg)
29
exposed filters. The filters were stored in sealed bags at 18 °C and only transported
cooled. Before handling the filters, they were left at room temperature in the sealed bags
in order to avoid condensation of volatile compounds on the cold surface (> 15 min).
The offline AMS analysis of the filters collected in April 2011 in Zürich were
conducted in October 2011. The other filters were analyzed between April and October
2012. While samples were collected at different seasons at 16 sites including urban,
suburban, and rural sites (Table 4.1), we will mainly focus on the Zürich data sets
(filters evenly distributed in the years 2011-2012) because of the extensive supporting
measurements performed there. Measurements on the remaining samples are used for
the assessment of the bulk OA water solubility.
The urban background site Kaserne in Zürich is located in a park in the middle of
the urban core of a densely populated area (1.2 million inhabitants, including
surrounding communities). In addition to filter sampling, an Aerosol Chemical
Speciation Monitor (ACSM) was operated with a PM1 (standard) aerodynamic lens in
Zürich from February 2011 to February 2012 (Canonaco et al., 2013, 2015, 2018). The
ACSM provides quantitative unit mass resolution (UMR) mass spectra with a time
resolution of 30 min. These mass spectra can be used to determine the concentration of
species such as OA and SO42- , while the OA mass spectra are suitable for source
apportionment (Ng et al., 2011b). At the same site, equivalent black carbon (eBC) was
monitored using an aethalometer, AE 31 (Magee Scientific Inc.) (Hansen et al., 1984;
Herich et al., 2011), and CO by non-dispersive Fourier-transform infrared spectroscopy
(APNA 360, Horiba, Kyoto, Japan). During spring 2011, PM2.5 filter samples were
also collected and a HR-ToF-AMS equipped with a (PM2.5) lens (Williams et al., 2013)
was operated at the same site.
During the winters of 2007/08 and 2008/09, offline AMS measurements (PM10)
were conducted for 15 sites spread over Switzerland including flatland and alpine sites
with varying population density and local emissions. A yearly cycle (August 2008-July
2009) from the urban background station in Zürich described above completes this data
set. For this campaign, measurements of organic carbon (OC), elemental carbon (EC)
(Zotter et al., 2014) and the most com-mon ions are available. Finally, we have also
analyzed 12 filter samples collected in Paris during summer 2009 and winter 2010,
![Page 47: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/47.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
30
where concomitant online HR-ToF-AMS PM1 measurements are available (Crippa et
al., 2013a,b,c; Freutel et al., 2013).
4.2.2 Offline AMS
4.2.2.1 Sample extraction
Sample fractions (2 cm2 or 1.2 % of the entire filter sample, may be increased for
low filter loadings) are collected from each filter sample and extracted in 10 mL
ultrapure water (18.2 MΩ cm, total organic carbon .TOC/ < 5 ppb, 25 °C) by means of
an ultrasonic generator for 20 min at 30 °C. Samples are then briefly vortexed (1 min),
to ensure their homogeneity. The extracts are subsequently filtered with 0.45 µm nylon
membrane syringe filters, prior to AMS analysis.
4.2.2.2 Offline AMS analysis
The water extracts are aerosolized using a custom-built nebulizer designed to
work with small liquid volumes (5-15 mL). When passing through the nebulizer nozzle,
an air stream is accelerated. Simultaneously, liquid is sucked into the nebulizer. The
high velocity air stream breaks up the solution and forms particles. The resulting
particles are dried by a silica gel diffusion dryer, and subsequently analyzed by the HR-
ToF-AMS (V-mode). For each sample, spectra are recorded in the range of 12-300 amu,
with a collection time for each spectrum of 30-60 s. To reduce memory effects,
ultrapure water is nebulized before every sample measurement. This information is used
as a system blank. Raw data depicting the measurement procedure are presented in Fig.
4.1 Field blanks are analyzed using the same procedure as the sample filters, and the
retrieved signals are statistically equal to those obtained from the direct nebulization of
ultrapure water. During each experiment, the nebulizer air is also filtered and measured
with the AMS to remove gas-phase contributions from the mass spectra (Allan et al.,
2004).
The HR-ToF-AMS operating principles, calibration procedures, and analysis
protocols are described in detail else-where (DeCarlo et al., 2006). The instrument
provides quantitative mass spectra of non-refractory PM1 (vacuum aero-dynamic
diameter (Dva) 60-600 nm) components, at 600 °C and 107 Torr (1.3×105 Pa). These
include organic aerosol and ammonium nitrate and sulfate. Data are analyzed using
![Page 48: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/48.jpg)
31
high-resolution analysis fitting procedures, Squirrel v1.52L (SeQUential Igor data
RetRiEvaL) and Pika v1.10C (Peak Integration by Key Analysis, D. Sueper), in the
IGOR Pro software package (Wavemetrics, Inc., Portland, OR, USA).
4.2.2.3 Other chemical analysis
Cations (e.g., K+, Na+, Mg2+, Ca2+, NH4+) and anions (e.g. SO4
2-, NO3-, Cl-) were
analyzed using an ion chromatographic system (850 Professional, Metrohm, Switzer-
land) equipped with a Metrosep C4 cation column and a Metrosep A anion column,
respectively. For this analysis, 1 cm2 filter fractions were extracted in 15 mL ultrapure
water (18.2 M • cm). Filters (1.5 cm2) were also analyzed for EC and OC content by a
thermo-optical transmission method on a Sunset OC/EC analyzer (Birch and Cary,
1996), following the EUSAAR-2 thermal-optical transmission protocol (Cavalli et al.,
2010). Replicate analysis shows a good analytical precision with relative standard
deviations of 7.7, 14.8, and 8.1 % for OC, EC, and TC (total carbon), respectively
(Zotter et al., 2014). The water-soluble organic carbon (WSOC) estimates from the
offline AMS analyses are compared to WSOC measured using a standard method.
Following this method, filter samples are extracted in ultrapure water, they are gently
shaken for 24 h, and the extracts are subsequently analyzed with a TOC analyzer. The
bulk of these offline measurements are used as reference methods to assess the offline
AMS approach.
4.2.2.4 PMF using ME-2
The ability of the offline AMS analysis to characterize the organic aerosol sources
compared with other online techniques (i.e., ACSM) is evaluated by analyzing the
obtained mass spectra from online and offline measurements using positive matrix
factorization (PMF, Paatero and Tapper, 1994) for the case of the yearly cycle from
Zürich (2011-2012). PMF is a bilinear unmixing receptor model used to describe
measurements (in this case AMS or ACSM organic mass spectra time series) as a linear
combination of static factor profiles and their time-dependent source contributions, as
expressed in Eq. (4.1):
(4.1)
![Page 49: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/49.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
32
Here xij, fkj, gik, and eij are matrix elements of the measurement, factor profile,
factor time series, and residual matrices, respectively. The subscript j corresponds to a
measured ion or m/z, i corresponds to a measured time stamp, and k to a discrete factor.
The user determines the number of factors, p, returned by the PMF algorithm. PMF
requires non-negative entries for fkj and gik, suitable for environmental measurements
such as OA mass concentrations. The PMF algorithm solves Eq. (4.1) by iteratively
minimizing the object function Q, defined as
(4.2)
where σij are the elements of the error matrix (measurement uncertainties), which
together with xij and p are provided as model inputs. Measurement uncertainties
considered in the error matrix include electronic noise, ion-to-ion variability at the
detector, and ion counting statistics (Allan et al., 2003). For offline AMS analyses, both
sample and blank uncertainties are incorporated. Following the recommendation of
Paatero and Hopke (2009) variables with low signal-to-noise (SNR < 0.2) are removed
(no variables affected), whereas “weak” variables (0.2 < SNR < 2) are downweighted
by a factor of 3 (26 variables in the PMF input affected). Further, 19 variables were not
considered in PMF because they were not present in the reference spectra used.
In this study, PMF is solved using the multi-linear engine (ME-2) (algorithm
Paatero (1999) and references therein), with the toolkit Source Finder (SoFi version 4.7,
Canonaco et al., 2013) for IGOR Pro (Wavemetrics, Inc., Portland, OR, USA)
employed as a front end for the model. PMF was operated using the robust outlier
treatment mode, in which outliers were dynamically downweighted. Most published
analyses using PMF are limited in their ability to explore rotational ambiguity in the
solution space, which is typically accessible only in a single, random dimension (Zhang
et al., 2011). As a consequence, these analyses do not guarantee the access to
environmentally optimal solutions. In contrast, the ME-2 implementation of PMF
allows efficient exploration of the entire solution space, including approximate matrix
rotations. In the present study, the solutions were directed towards environmentally
meaningful rotations by constraining the elements of one or more profiles in the factor
![Page 50: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/50.jpg)
33
profiles matrix (fkj) to a predetermined range defined by a centervalue (fkj’) and a scalar
defining the width of range (a), such that the returned profile satisfies fkj=fkj’±a×fkj’.
This approach has previously been utilized for AMS data sets to separate distinct
sources with correlated mass spectra profiles or time series (Lanz et al., 2008; Crippa et
al., 2014) and shown to provide improved factor separation compared to conventional
PMF (Canonaco et al., 2013).
In the case of the offline AMS, the HR data matrices were arranged as follows: in
the measurement matrix, each filter sample is represented by on average eight high-
resolution mass spectra (see description above and Fig. 4.1), corrected for the
corresponding average blank measured before the sample. Each mass spectrum is
composed of 154 HR ions (m/z 12-96). 41 samples were considered in this analysis
(total of 41 time points and matrix total dimension of 334×154=51 436). The
corresponding error matrix has the same dimensions. The elements of the error matrix,
σij, include the uncertainties related to the AMS measurements as discussed above
(computed according to Allan et al., 2003; Ulbrich et al., 2009), denoted by δij, added in
quadrature to the variability of the preceding blank βij, which includes the AMS
measurement precision but also accounts for possible drifts in the nebulization:
(4.3)
In order to allow comparisons with external data, the offline AMS data and error
matrices are converted to ambient concentrations. The contribution of δij and βij to σij
depends on the ion in question, but in general δij dominates (98 %, first and third
quartiles of 89 and 100 %). Since the measured data points are not averaged prior to the
ME-2 analysis, but rather used individually, their variability is not included in the error
matrix, but instead directly reflected in the results. This also provides a metric for the
mathematical stability of the ME-2 solution and thus a part of the uncertainties of the
source apportionment results.
In order to assess the performance of offline AMS data in source apportionment
we compare the obtained results to source apportionment results using online ACSM
![Page 51: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/51.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
34
data. For an ideal offline/online comparison, the online data set should resemble the
offline data set as closely as possible. However, the low mass resolution of the ACSM
spectra prevents satisfactory factor resolution, when using 24 h averages of ACSM data
for the selected days. Moreover, running ME-2 on the selected days for the entire year
(retaining 30 min time resolution) results in biases between winter and summer residual
distributions, which was not the case for the offline data. That is, the model tends to
explain the diurnal variation of the online ACSM data, rather than seasonal differences.
For these reasons, a rolling window ME-2 approach was developed to perform source
apportionment analysis on the yearly online UMR ACSM data (Zürich 2011-2012)
(Canonaco et al., 2018). The approach can be described as a controlled bootstrap
technique applied to sorted data, which would help represent summer and winter data
and provide an estimate of the uncertainties (Paatero et al., 2014). In this approach, a
rolling window is capable of capturing seasonal variations in the aerosol factors and/or
variations driven by meteorology. Within a window, which is considerably shorter than
the yearly data set, the ME-2 model is applied, allowing the factors to adapt to the
measured data. A rolling window corresponds to 4 weeks of measurements and rolls
over the whole set of data with a 1 day time step. The PMF window was rolled over the
temperature-sorted Zürich data (by daily average temperature). By sorting the data with
respect to temperature, days with similar conditions in terms of SOA formation and
dominant primary sources (e.g., BBOA at lower temperature) are grouped together. For
every window the solution was optimized using criteria based on correlations between
the time series and the diurnal cycles of the factors and those of the markers. This novel
approach was compared to classical source apportionment results for the winter part of
this data set presented in Canonaco et al. (2013). The rolling window solution presents
an improved representation of OOA (R2 with NH4+ 0.69 vs. on average 0.53 for the
PMF solution in Canonaco et al., 2013) for the overlapping period, which is consistent
with the variable character of OOA. The correlation of HOA and BBOA with their
respective markers is comparable to Canonaco et al. (2013). For the reasons described
above and with the lack of standard techniques to apply PMF to yearly organic mass
spectral data, the rolling window source apportionment results are chosen as reference.
![Page 52: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/52.jpg)
35
4.3 Results and discussion
4.3.1 Signal-to-noise, quantification and detection limits
Figure 4.1: Data recorded with HR-ToF-AMS of filter samples collected in Zürich (2011-2012).
Data from a typical measurement cycle are underlaid in gray. (a) Raw signals obtained for organic aerosol (OA, green), nitrate (NO3
- , blue), sulfate (SO42-, red), and ammonium (NH4
+, orange), where AMS filter air as well as blank and sample measurements are indicated. (b) OA average signal for samples and blanks (logarithmic scale), blank correction curve and the noise (smoothed standard deviation of the blank) associated with the signal of different species used for the calculation of errors. On the y axes, a.u. denotes arbitrary units.
Figure 4.1a shows a typical time pattern of OA, NO3-, SO4
2-, and NH4+ from
offline AMS measurements. The signal intensity of offline AMS measurements can be
expressed in μgm-3 (of nebulized aerosol), but for simplicity we denote this as arbitrary
units (a. u.) to avoid confusion with concentrations in ambient air (μgm-3). This
conversion between AMS signals and real concentrations is explained below. The
intensity of OA is typically 1-2 orders of magnitude higher than that of the
measurement blanks (see Fig. 4.1b). The blank offline AMS signal is typically below
2.1 a. u., with interday and intraday variation (standard deviation) of 0.3 and 1.5 a. u.,
![Page 53: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/53.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
36
respectively. The nebulization efficiency assessedbased on the SO42- signal is
3.8mLsolution mair-3 (first and third quartile of 1.2 and 7.3 mL solution mair
-3). The
particles generated with this nebulizer have a mode diameter of ~200 nm (dV /dlogDp).
The SO42- detected by the offline AMS is related to SO4
2- loadings on the filter
area (calculated from ACSM measurements) analyzed by a power relationship (Fig.
4.2). However, the offline AMS measurements described herein cannot be directly
quantified, without external measurements of e.g., OC, due to variability in the
nebulization process. Another significant source of uncertainty is the ACSM cutoff
(Dva 600 nm).
Figure 4.2: Offline AMS SO4
2- blank corrected concentrations compared to theoretical SO42-
loadings of the filter fractions (μg). The theoretical SO42- loadings are calculated based on ambient
SO42- concentrations measured by the ACSM for the Zürich yearly cycle and the volume of air sampled
through the analyzed filter fraction. Results are fitted using a power function (ln(y) = 2.3 × ln(x) - 5.2).
![Page 54: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/54.jpg)
37
The detection limit (dlj) of species j (e.g., SO42-), in μg on the analyzed filter
fraction, is evaluated based on the blank variability of j in comparison to the signal in
the sample, σblank,j . We define dlj as the mass of j in μg required to produce a signal
equal to 3 σblank,j. dlj is inferred using the existing relationship between blank corrected
offline AMS signals, Ij in a. u., and the mass concentration of j , Mj in μg, in the
analyzed filter fractions (e.g., Fig. 4.2). We estimated the detection limits for OA and
SO42- as 0.80 and 0.25 μg on the analyzed filter area, corresponding to 80 and 25 μgL-1,
respectively.
4.3.2 OA recovery
The loss of hydrophobic or volatile organic species during sample collection,
handling, extraction, and nebulization may significantly hinder the applicability of the
offline AMS technique. In the following, the organic aerosol signals are normalized to
the sulfate mass, in order to evaluate OA recoveries. This is based on the assumption
that sulfate is quantitatively extracted and measured by the AMS, which is expected
since sulfate is mostly bonded to ammonium (watersoluble and non-refractory). This is
not given at all sites, e.g., in the strong presence of potassium. We also assume that the
fractional composition in the size range sampled by both ACSM and filter samples is
the same. Accordingly, the comparison of Ij and Mj both normalized to SO42- yields the
recovery Rj:
(4.4)
The extraction time does not have a statistically significant effect on OA/SO42-
ratios and fingerprints when increasing the extraction time from 20 to 60 min. Likewise,
multiple extractions did not significantly enhance the recovery of the particulate
compounds as the OA signal from the second extraction was below 8% of that from the
first extraction and only 3 times higher than the blank signal. Therefore, we have
concluded that a single extraction step was sufficient in our case to obtain the water-
extractable material.
![Page 55: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/55.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
38
Figure 4.3: Estimated recoveries of organic compounds based on the comparison of OA/SO4
2- ratios using the offline AMS method to reference measurements for different days. The error bars represent the variability of the offline OA/SO4
2- ratio within a sample and were obtained from different runs during the same measurement of the same sample. (a) The reference OA/SO4
2- ratio is obtained by OC filter measurements (Sunset OC/EC analyzer) scaled to OA using OM/OC values from the HR offline AMS data and SO4
2- from IC. (b) OA/SO42- ratios from online measurements were used as
reference values. For both Paris campaigns and the Zürich spring campaign, the online measurements were conducted using HR-ToF-AMS and for the yearly cycle in Zürich by a quadrupole ACSM. (c) For Zürich (2011-2012), probability density functions of Rbulk are presented both using the offline AMS measurements as well as using WSOC from the Sunset OC=EC Analyzer (in combination with OM=OC ratios from offline AMS).
![Page 56: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/56.jpg)
39
We have evaluated the recovery of the bulk OA (Mj here representing OA), by
comparing the offline AMS OA/SO42- ratios with OA/SO4
2- from reference
measurements using the Sunset OC/EC analyzer and ion chromatography (IC)
(described in Sect. 2.3). The recovery of complex mixtures such as ambient OA
depends on the water solubility of its numerous compounds. Figure 4.3a compares
offline AMS and reference measurements for 15 stations in Switzerland where the
reference measurements were performed on the same filters (150 PM10 samples, Table
4.1), using IC and the Sunset OC/EC analyzer for SO42- and OA measurements,
respectively. The latter were calculated by multiplying the Sunset OC/EC analyzer OC
with the OM/OC ratios from the HR analysis of the AMS spectra. While we
acknowledge that also OM/OC from offline AMS is subjected to errors caused by
compound-dependent extraction efficiencies and filter sampling artifacts, such errors do
not significantly affect the results and the OM/OC range found here (median of 1.84,
first quartile of 1.80 and third of 1.87) compare well with previously measured online
ratios (e.g., 1.80 provided by Favez et al. (2010) for Grenoble, January 2009, 1.66 by
Crippa et al. (2013a) for Paris, and 1.6 and 2.0 by Minguillón et al. (2011) for Barcelona
and Montseny, respectively). From this, we estimate a median Rbulk of 0.60 (first and
third quartiles of 0.49 and 0.80), which suggests that the technique can capture a large
part of the organic fraction.
Similar comparisons between offline AMS results and reference measurements
were also performed for other data sets where online AMS data were available (Zürich
spring and Paris campaigns). Offline AMS measurements of filter samples collected in
Paris (summer 2009 and winter 2010) and Zürich (spring 2011) were compared with
online HR-ToF AMS with the same size cutoff (PM1) (Fig. 4.3b). For these data sets,
OA recoveries range between 64 and 76 %. For Zürich, it should be noted that PM1
filter samples are not available and therefore offline PM10 HR-ToF-AMS
measurements are compared to online data from ACSM (yearly cycle) and HR-ToF-
AMS equipped with a PM2.5 lens (spring). We show that for both campaigns the
overall Rbulk is in the same range as values obtained for the other data sets inspected
here, despite the potential contribution of coarse mode OA (median = 0.91; first and
third quartiles of 0.66 and 1.32, respectively for the yearly cycle and median = 1:05;
first and third quartiles of 0.99 and 1.26, respectively for the spring campaign). This
![Page 57: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/57.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
40
implies that the contribution of the latter is not dominant, consistent with previous
measurements at this site, suggesting that the fine particle mass constitutes on average
75 % of the PM10 mass (Putaud et al., 2010). Note that outliers in Rbulk higher than 1
are associated with very low, and therefore highly uncertain sulfate concentrations. For
the Zürich yearly cycle campaign (2011-2012), we validated the Rbulk calculation
approach adopted here to a more conventional approach for the determination of WSOC
(Fig. 4.3c; water-soluble organic aerosol, WSOA = WSOC × (OM/OC)offline AMS). We
show that both approaches give similar estimates (based on the WSOC median Rbulk =
0.74 compared to Rbulk = 0.91 if the calculation is based on the ACSM), suggesting that
offline AMS measurements are related to WSOA and that a great part of the organic
mass is accessible by the analysis procedure followed here.
4.3.3 Mass spectral analysis
Results above raise the question as to whether the offline AMS analysis maintains
the mass spectral signature of the ambient OA. We have addressed this question by
comparing online and offline OA mass spectra in Fig. 4.4, illustrating an example of the
results obtained from Zürich winter and spring campaigns. Such a comparison
implicitly assumes that the mean organic composition across the entire size range
collected by the filter (up to PM10) is identical to that of the approximately 60-600 nm
particles measured by the online ACSM. Although this assumption will not hold for all
conditions, the comparison is nonetheless useful for characterization of the offline AMS
technique. The comparison of offline and online spectra shows a high correlation (R2 >
0.97) irrespective of the seasonal variation in aerosol composition. More importantly, it
can be observed that this method is also able to capture variations in the aerosol
fingerprints between the two seasons. For instance, both online and offline methods
show higher contributions from BBOA and HOA-related fragments (e.g., m/z 60, 73,
and m/z 55, 57, 69, respectively) for the winter samples and higher contributions from
OOA-related m/z values (e.g., m/z 43, 44) for the spring sample.
![Page 58: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/58.jpg)
41
Figure 4.4: Comparison between 24 h average online and offline AMS (both PM2:5) spectra for
winter (a) and spring (b) samples, collected in Zürich. Fragments (m/z) commonly considered as source-specific markers are explicitly labeled with their nominal mass.
HR-ToF-AMS data enable the analysis of individual ions at the same integer m/z,
which in turn provides better assessment of the recovery of the initial parent organic
compounds or ion families. For this analysis, we use Eq. (4.4) to describe the recovery
of individual ions Rfrag with Mj and Ij defined as the concentration of an individual ion
(cfrag). We have grouped the fragments into five different families, based on their
heteroatom content and degree of unsaturation, including N-containing hydrocarbon
ions (CHN), monooxygenated (CHOz=1) and poly-oxygenated (CHOz>1) ions and pure
hydrocarbons (CH) divided into saturated (CHsat) and unsaturated hydrocarbons
(CHunsat)
Figure 4.5 presents Rfrag for the Zürich spring (2011) campaign (see Fig. 4.3b,
green points for Rbulk = 0.65 (first and third quartiles of 0.62 and 0.70). Results show
that highly oxygenated fragments (CHOz>1, mainly organic acids) are well recovered,
RCHOz>1 67% (first and third quartiles of 65 and 72 %). This proportion slightly
decreases to 64% (first and third quartiles of 63 and 71 %) for the CHOz=1 (RCHOz=1)
family, which could mainly be composed of alcohols, aldehydes, and ketones. In
contrast, Rfrag for non-oxygenated species are in general lower, i.e., 55% (first and third
quartiles of 51 and 60 %) for the CHN family (RCHN), and 61% (first and third quartiles
![Page 59: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/59.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
42
of 55 and 64 %) for the CH family (RCH). Within the CH family, the saturated
hydrocarbon fragments (CnH2n+1+), which stem at least in part from the fragmentation of
hydrophobic normal and branched alkanes (Alfarra et al., 2004), are especially strongly
underestimated (RC2H2n+1+=44%, first and third quartiles of 42 and 48%). Note that a
higher variability in the Rfrag value is also observed for the CH fragments, probably due
to the variability in the water solubility of their parent molecules. This may lead to
higher uncertainties in the source apportionment of hydrocarbon-like OA and even to an
underestimation of such sources, using the offline AMS technique, as will be shown
below.
Figure 4.5: Median recovery of single organic fragments, and chemical families for the Zürich
spring campaign (offline vs. online PM2.5 AMS). The first and third quartiles of the inter-sample variability are shown as error bars. A ratio of 1 indicates a recovery of 100 %. The fragments are color-coded with the family (CH (hydrocarbon fragments, split into saturated and unsaturated), CHOz=1 and CHOz>1 (oxygenated fragments), and CHN, nitrogen-containing hydrocarbon fragments). Numbers across the top of the plot indicate the fragments’ nominal mass. Families include all respective fragments weighted by their mass contribution.
4.3.4 Source apportionment results
Differences between offline and online HR-ToF-AMS spectra caused by e.g.,
compound-dependent recoveries may also influence source apportionment results.
Therefore, we assess the ability of the offline AMS in the apportionment of OA sources,
by analyzing the offline Zürich yearly data set using ME-2 and comparing the source
apportionment results to those obtained by applying ME-2 to online ACSM data.
![Page 60: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/60.jpg)
43
4.3.4.1 ME-2 output evaluation
A key consideration for PMF analysis is the number of factors selected by the
user. As mathematical criteria alone are insufficient for choosing the right number of
factors, this selection must be evaluated through comparisons of factor and tracer time
series, analysis of the factor mass spectra, and the evolution of the residual time series
as a function of the number of resolved factors. As described below, a five-factor
solution was selected as the best representation of the offline AMS data. To improve the
resolution of the POA sources by the model, literature profiles were used to define the
range of acceptable profiles (using the a value approach - Sect. 2.4). SOA factors are
not constrained because of the complex dependence of SOA composition on source,
atmospheric age, processing mechanisms, and meteorological conditions. This is
consistent with the approach of Crippa et al. (2014). After determining the optimal
number of factors (and their identity) required for explaining the variability in the data
set, we thoroughly assessed the sensitivity of the PMF results to the selection of the a
values.
Figure 4.6: Change in the time-dependent contribution of Q/Qexp as a function of the number of
factors Δ(Qi,cont/Qexp,i,cont) for a chosen offline solution (for aHOA = 0.0 and aCOA = 0.0).
Previous studies at this site have shown the influence of traffic, cooking, biomass
burning, and secondary organic aerosol (Lanz et al., 2008; Canonaco et al., 2013). Here,
we have constrained HOA and COA (profiles adapted from Mohr et al., 2012) and
optimized the solution by investigating different combinations of a values for the
![Page 61: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/61.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
44
constrained factors. In the selected five-factor solution, the non-constrained factors
extracted by ME-2 were related to BBOA, a highly oxygenated (OOA1) and moderately
oxygenated (OOA2) organic aerosol; the sum of OOA1 and OOA2 will be henceforth
considered as a proxy for secondary organic aerosol (referred to as OOA) which can,
though, be mixed with aged primary organic aerosol. These designations are based on
the correlation between OOA time series and that of secondary inorganic species (i.e.,
SO42- and NH4
+) and the comparison of OOA profile mass spectra with those extracted
from previous AMS data sets.
Number of factors
Figure 4.6 shows the change in the time-dependent Q/Qexp when increasing the
number of factors for the offline data set Δ(Qi,cont/Qexp,i,cont): contribution to Qi,cont for the
(p)- factor solution minus that of the (p+1)-factor solution. A significant decrease in
Δ(Qi,cont/Qexp,i,cont) signifies that structure in the residuals disappeared with the additional
factor. Removed structure is evident up to five factors. This behavior indicates that
while the ME-2 solution is clearly enhanced when increasing the number of factors to
five, addition of further factors does not improve the model description of the input
data. For this solution, there is no statistically significant difference in the residual
distributions of most variables between winter and summer (Fig. 4.7), indicating that the
modeled profiles represent the sources over the entire year well. Lower order solutions
(three and four factors) show one or two OOA factors besides the constrained HOA and
COA. Higher order solutions were explored but yielded additional OOA factors, which
could not be clearly attributed to a distinct source or process. Given this lack of
improvement in Δ(Qi,cont/Qexp,i,cont) and in the understanding of aerosol sources and
formation processes, and the absence of external tracers supporting the additional OOAs
in the high order solutions, the five-factor solution was considered as being optimal.
Furthermore, we consider only the sum of OOAs to facilitate the inter-method
comparison (as explained below). Note that PMF model uncertainties, i.e., imperfect
mathematical unmixing of sources, propagate into this comparison. This setting allows a
direct comparison between the offline and online methodologies, as the same set of
factors are obtained.
![Page 62: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/62.jpg)
45
Figure 4.7: Residuals weighted with the uncertainty (residuals/uncertainty) of the offline solutions
for the periods April-September and October-March (example shown for one chosen solution, aHOA = 0.0; aCOA =0.0). Panels (a, b) show residuals as a function of m/z averaged over the whole periods color-coded with the probability that the residuals for April-September are the same as for October-March (Wilcoxon-Mann-Whitney test). Panels (c, d) shown the probability distribution function (pdf) of R/U during the same periods.
a-value optimization
The a values are independently varied for all constrained factors within a wide
range (a values from 0 to 1 with a step size of 0.1) for offline data in order to find an
optimal solution. Amongst the different solutions obtained, we selected those with
factor time series with the strongest correlation with those of the corresponding tracers.
The a value combinations of the chosen solution are specific for the data set used herein
and the selected reference profiles used, i.e., they may not be directly applicable to other
cases.
For this selection, the approach described above is adopted for the offline data
(illustrated in Fig. 4.8). For each set of a values selected as ME-2 input parameters (two
a values to constrain HOA and COA) five-factor time series are first generated by ME-
![Page 63: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/63.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
46
2. The ratios between factor and marker time series are then displayed as probability
density functions (pdf). The width of this distribution is used as a quality criterion, since
the narrower it is, the closer the linear relation of the factor is to the marker. Here, eBC,
CO, and NH4+ are used as markers for HOA, BBOA, and OOA, respectively. For COA,
no specific marker has yet been identified and studies presenting online data validate
this factor using the daily pattern of its concentration, which typically peaks at lunch-
and dinner-time (e.g. Crippa et al., 2014). For 24 h integrated filter data, this diagnostic
cannot be used and therefore the optimization of COA separation by ME-2 is not used
as a quality criterion. In practice, the solution with the narrowest factor-to-marker
distributions is defined as the best solution with respect to its interpretability in the
environment. For the other factors, we have examined the variability in the ratio, x,
between factor and corresponding marker: . CO0 is the
background concentration, which is estimated to be 100 ppb. This is both in agreement
with measurements at this site and also literature presenting measurements of
background air masses (e.g. Griffiths et al., 2014, for Jungfraujoch). In practice, the best
solution is obtained by minimizing the sum of the ratios of the logarithmic geometric
standard deviations (σg) to the logarithmic geometric averages (µg) of
. Besides using eBC as an HOA marker, the quality of the solution
was also checked using eBCtr. For both markers, the same a value combination was
considered best according the overall criterion. All solutions, for which none of the
single distributions showed a different relative variance than the best solution, were also
accepted (this comparison was performed using an F test). Note that the determination
of the a value ranges resulting in the most satisfactory solutions for the offline data set
is performed independently from the online measurements. The comparison between
source apportionment from offline and online data sets provides, therefore, a direct
measure of the ability of the offline AMS technique to resolve aerosol sources and
formation processes. Systematically stepping through the multi-dimensional a value
space, as opposed to most published analyses using one-dimensional a-value-based ME-
2 or PMF, offers a more effective and objective exploration of the solution space.
Environmentally reasonable factors are obtained by selecting the subset of solutions that
optimizes factor-to-marker relationships.
![Page 64: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/64.jpg)
47
Figure 4.8: Relative width of the distributions cj/cmarker displayed as a function of aHOA and aCOA.
Panel (a) shows the sum of the criteria for HOA, BBOA, and OOA being the sum of OOA1 and OOA2. (The chosen solutions are pointed out in the white area.) Panels (b-d) show the individual criteria as a function of the a values of HOA and COA aHOA, aCOA.
The chosen offline solutions lie in general in a part of the solution space with low
a values for HOA and COA (the single chosen a value combinations are shown as white
boxes in Fig. 4.8a). The relative variability of the signatures of the constrained factors
among the different accepted solutions for the offline source apportionment is below 6
and 3% for HOA and COA, respectively. Note that this is an incomplete exploration of
the rotational ambiguity and thus does not describe the complete model uncertainty.
Online ACSM solution
Like the offline solution, the online ACSM solution yielded a five-factor solution
representing HOA (constrained using the profile reported by Crippa et al., 2013c), COA
(constrained using the profile reported by Crippa et al., 2013c), BBOA (constrained
using the profile reported by Ng et al., 2011b), SV-OOA, and LV-OOA. In contrast to
the offline solution, ME-2 could not extract BBOA independently; thus this factor was
constrained as suggested by Crippa et al. (2014).
![Page 65: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/65.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
48
Comparing the offline to online PMF source apportionment results obtained with
the approaches described earlier has the obvious drawback that we compare OOA
factors extracted in winter-only or summer-only (online) vs. combined winter and
summer (offline). However, this is mitigated by two factors. First, the discrimination
between OOAs for the offline solution is largely driven by seasonal differences (average
relative contributions: in winter OOA2 9 %, OOA1 91% and in summer 87, 13 %,
respectively), indicating only small biases in the composition. Second, the residuals for
both winter and summer are normally distributed and centered around zero (Fig. 4.7),
indicating negligible seasonally dependent bias in the apportioned mass. Therefore,
while this comparison method may contribute somewhat to the overall uncertainties, it
is unlikely to significantly affect the conclusions or values reported below.
Figure 4.9: Comparison of overall factor profiles obtained for the chosen solutions both from the
offline (left, for HOA and COA, spectra from Mohr et al. (2012) were used as reference) and the retrieved factor profiles from the online source apportionment (right).
![Page 66: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/66.jpg)
49
Factor profiles
The averages of factor profiles of the selected ME-2 online and offline solutions
are presented in Fig. 4.9. Apart from the good correlations between factors and external
markers time series used as an acceptance criteria, our results show that the factors
retrieved by ME-2 exhibit spectral profiles are consistent with previous studies. The
BBOA profile extracted from the offline data set closely resembles those reported in the
literature for other locations (Crippa et al., 2013c), characterized by the contribution of
oxygenated fragments at m/z 29 (CHO+), 60 (C2H4O2+), and 73 (C3H5O2
+), from
fragmentation of anhydrous sugars (Ng et al., 2011b). The OOA mass spectra retrieved
by ME-2 for both online and offline data sets is characterized by a typical fingerprint,
dominated by oxygenated fragments at m/z 43 (C2H3O+) and 44 (CO2+) characteristic of
secondary compounds. The consistency of these spectral profiles with previously
reported profiles from online measurements provides additional support to the source
apportionment results presented here.
4.3.4.2 Recoveries of different OA categories (HOA, COA, BBOA, OOA)
While results above show that the bulk OA recovery lies between 60 and 91%
(Sect. 3.2), for the current analysis we assess the recovery of the different factors as
representative of ambient compound classes from various sources/processes determined
by ME-2. This is based on the comparison between online and offline source
apportionment results. However, for this comparison, the approach presented in Eq.
(4.4) cannot be adopted because of the noisy data and/or low sulfate content during
periods critical for recovery determination of a specific factor. We therefore perform a
selfconsistent calculation of factor-dependent recoveries, which, when applied to the
offline data, yield (1) fractional composition consistent with online measurements, and
(2) calculated bulk recovery consistent with the measured bulk recovery using WSOA
measurements. The implementation is described below (Eq. 4.5). Let i denote the time
index and k the factor index. We define the time-dependent recovery of a factor k (Ri,k)
as the time-dependent ratio of the contribution of this factor in the offline (offi,k) and in
the online (oni,k) solution multiplied by the time-dependent bulk recovery of WSOA,
Ri,WSOA (Eq. 4.5):
![Page 67: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/67.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
50
(4.5)
Figure 4.10: Recoveries Rk for HOA, COA, BBOA, and OOA (OOA1COOA2) obtained from the
intercomparison of source apportionment results of offline AMS to online ACSM data (Zürich 2011-2012). 100 000 random combinations of offline and online solutions and randomly chosen offline repeats result in the same amount of time-independent Rk, which are expressed as probability density functions (pdf).
Finally, Rk, the median of Ri,k over time, is computed. Rk reflects not only the bias
caused by the water extraction but also filter sampling/storage effects and differences
between the individual ME-2 solutions. The uncertainty of Rk depends both on the
uncertainty related to the single offline solution point in time as well as on their spread
in comparison to the online solution. The first can be quantified by assessing the model
error for the offline and online using the variability of the solution for different model
runs. The offline approach adopted here, including several measurements of the same
sample in ME-2 (in general eight spectra, called repeats, per sample), enables assessing
the performance of the ME-2 solution for different samples and different factors. For
this reason, we have repeatedly calculated Rk using randomly chosen combinations of
(1) different ME-2 offline solutions (selected in Sect. 3.4.1 and reference online
solutions (due to the rolling window approach providing individually optimized
periods) and (2) different repeats of the offline AMS measurements for the same
![Page 68: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/68.jpg)
51
samples. The result is an ensemble of Rk (for each factor 100 000 Rk values are
calculated) displayed in Fig. 4.10 as probability density functions. The range of these
distributions reflects both model and measurement uncertainties. Note that this range
does not reflect the variability in time of Ri,k. The retrieved factor recoveries are
consistent with our understanding of the chemical nature of the different OA
components, with primary hydrophobic species less efficiently extracted than secondary
oxygenated species. As expected, the most hydrophobic component, HOA, has the
lowest recovery with a median RHOA of 11% (first and third quartiles of 10 and 12 %).
We note that RHOA is lower than the RCH,sat (see Sect. 3.3), which seems to indicate that
these ions can also originate from more hydrophilic molecules than those in traffic
emissions. COA appears to be moderately soluble, with RCOA = 54% (first and third
quartiles of 48 and 60 %). BBOA and OOA species were largely recovered with RBBOA
= 65% (first and third quartiles of 62 and 68 %) and ROOA = 89% (first and third
quartiles of 87 and 91 %). Uncertainties in RWSOA are not included in the calculation.
Further, online measurements have a lower size cutoff than the offline data (600 nm vs.
10 μm), and large accumulation mode particles are expected to preferentially contain
OOA, due to their extended aging. This might provide a positive bias to the OOA Rk
and a negative bias to Rk of the other factors.
4.3.4.3 Quantitative comparison of offline and online OA factors
We assume that the Rk values calculated in the previous section are characteristic
properties of the retrieved OA components, i.e., that they can be applied throughout the
analyzed offline data sets. This allows us to quantitatively compare the mass
concentrations of offline and online OA factors retrieved throughout the year. By
applying the Rk obtained to the offline data set, the source apportionment results
(relative composition) can be corrected. In a second step the results can be scaled to
ambient concentrations. Here OA concentrations from the ACSM measurements are
used. The Rk corrected offline source apportionment results are compared to results
from the ACSM analysis and respective marker concentrations (Fig. 4.11). The
displayed error bars reflect the variability of a factor’s contribution for one offline
sample assessed by the repeats, and also using the different chosen ME-2 solutions
provides an estimate of the measurement and (partial) model uncertainties. Overall the
variability of the offline factor contributions for an individual sample is 0.1 μgm-3. The
![Page 69: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/69.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
52
factors with a lower recovery Rk (HOA and COA), reveal also bigger differences
between the time series for the offline and online data and thus more scattering. In
general, the variability of the offline factor contributions for an individual sample
increases when moving away from the 1 : 1 line. This is especially apparent for BBOA,
where outliers with low offline concentrations are much more uncertain than the points
matching the contribution in the online solution (Fig. 4.11c, f). All factors but COA (for
which no marker is known) show similar relationships with their marker for both offline
and online data (Fig. 4.11e-g).
Figure 4.12 presents the ratio of factor contributions and their respective marker
concentrations for the online and corrected offline solutions. The medians and spreads
of the distributions are comparable between offline and online solutions. Only for
BBOA is the distribution wider (also seen in Fig. 4.11). Chirico et al. (2011) and El
Haddad et al. (2013a) report HOA/EC ratios of 0.4, which is close to the median found
in this study (HOAoff/eBC = 0.57 (first and third quartiles of 0.42 and 0.74),
HOAon/eBC = 0.64; first and third quartiles of 0.42 and 0.79). The ratio BBOA/(CO-
CO0) is 6.1 (first and third quartiles of 2.2 and 7.8) and 5.7 μgm-3 ppm-1 (first and third
quartiles of 4.5 and 8.4) for offline and online, respectively. However, this ratio has to
be considered as a lower limit, as CO may also be emitted by non-biomass-burning
sources (e.g., traffic). While this ratio is significantly lower than values reported for
prescribed/open burns (De Gouw and Jimenez, 2009), values found here are within the
same range as those measured for modern stoves used in Switzerland (Heringa et al.,
2011). Crippa et al. (2014) reported OOA/NH4+ ratios for 25 sites, with an average of
2.0 (0.3 for the site with the lowest ratio and 7.3 for the one with the highest). Lanz et
al. (2010) reported values of 5.6 and 1.5 for Zürich in July 2005 and January 2006,
respectively. The values for Zürich during the period analyzed here are 5.1 (offline, first
and third quartiles of 3.0 and 11.50) and 5.1 (online, first and third quartiles of 3.0 and
10.5). The examination of these ratios and their comparison with previously reported
values provide additional support to the offline AMS methodology and resulting source
apportionments. Indeed, the application of this methodology to additional filters from
other locations where accompanying online AMS measurements are available may aid
the further constraint of the Rk estimates presented here.
![Page 70: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/70.jpg)
53
Figure 4.11: Comparison of factor contributions from separate offline (PM10 AMS, two constrained factors: HOA, COA) and online (PM1 ACSM) source apportionment using ME-2 (traffic (HOA), cooking (COA), biomass burning (BBOA), and oxygenated organic aerosol; OOADOOA1COOA2). Factor-specific recoveries (Rk) are applied to the offline contributions. Error bars (in gray) denote the variability between the different ME-2 solutions and for different recorded spectra per sample for offline and for online only the first of the two. Panels (a-d) show scatter plots comparing the absolute contribution of the respective source/OA category for offline AMS and online ACSM measurements. The color code distinguishes all factor contributions (bullets, saturated colors) from winter points (open circle, light colors). The gray dashed line indicates the 1 V 1 line. Panels (e-g) show the correlation with the respective markers: black symbols represent the absolute contribution of the respective source for the online ACSM measurements and the colored symbols represent the absolute contribution of the same source for the offline AMS measurements.
![Page 71: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/71.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
54
Figure 4.12: Ranges of ratios of the contribution of different factors to their markers for the offline
(corrected with Rk) and online ACSM source apportionment results. Note that OOA is the sum of OOA1 and OOA2.
4.4 Conclusions
In this study, we developed and evaluated an offline method using an HR-ToF-
AMS for the characterization of the chemical fingerprints of aerosol collected on filters
(Pall quartz filters for the current study). Particulate matter on filters is extracted in
water and introduced into the HR-ToF-AMS using a nebulizer. The method was applied
to more than 250 filters from different seasons in different environments in Europe. The
detection limits depend on the nebulizer and species; for the current setup for OA and
![Page 72: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/72.jpg)
55
SO42- they are 80 and 25 μgL-1, respectively. External data are needed for quantification.
We recommend the use of OC analysis from a Sunset OC/EC analyzer and ion
chromatography data for the determination of the inorganic fraction. Estimates of the
recovery of bulk OA using different reference measurements show that OA is largely
captured (60-91% depending on the data set). The obtained organic mass spectra are
comparable to online HR-ToF-AMS spectra, although hydrocarbons are
underestimated. Rbulk also shows a good agreement with the WSOA fraction.
Source apportionment on offline AMS data is conducted with positive matrix
factorization, implemented using the ME-2 algorithm. We investigate a set of PMF
solutions, for which the different OA components show tight relationships with their
respective markers. Thereby, we demonstrate that organic mass spectral data generated
using this method are suitable for identifying different OA sources as HOA, COA,
BBOA, and OOA. By comparing the results for offline AMS and ACSM data, we
retrieved recoveries of the different OA components (Rk): traffic (RHOA = 0.11), cooking
(RCOA = 0.54), biomass burning (RBBOA = 0.65), and secondary OA (ROOA = 0.89).
Qualitatively, Rk also relates to the water solubility of the respective source, e.g.,
primary OA related to hydrocarbons (e.g., HOA) shows a low Rk caused by its low
water solubility. Such Rk should be determined at other sites where also additional
sources might be important, providing an assessment of site-to-site variability.
Nevertheless, these best estimates of Rk may be used to correct source apportionment
results from offline AMS measurements (as in Huang et al., 2014). When combined
with WSOC measurements, one might also be able to assess the applicability of these
values at the site in question by comparing overall modelled Rbulk to RWSOA. Even
though the offline AMS approach might poorly capture sources exhibiting fast changes,
this method broadens the applicability of the AMS to longterm size-segregated (PM1,
PM2.5, PM10) measurements (in contrast to online campaigns of typically 1 month) for
extended monitoring networks.
![Page 73: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/73.jpg)
Chapter 4: Characterization and source apportionment of OA using offline AMS
56
Acknowledgements
This work was supported by the Swiss Federal Office of Environment, inNet
Monitoring AG, Ostluft, the country Liechtenstein, the Swiss cantons Basel-Stadt,
Basel-Landschaft, Graubünden, Solothurn, Ticino, Thurgau, Valais, the Lithuanian-
Swiss Cooperation Programme “Research and Development” project AEROLIT (Nr.
CH-3-ŠMM-01/08), and the IPR-SHOP SNSF starting grant.
![Page 74: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/74.jpg)
57
5 Long-term chemical analysis and organic
aerosol source apportionment at 9 sites in Central Europe: source identification
and uncertainty assessment
K. R. Daellenbach1, G. Stefenelli1, C. Bozzetti1, A. Vlachou1, P. Fermo2, R. Gonzalez2, A. Piazzalunga3,a, C. Colombi4, F. Canonaco1, C. Hueglin5, A. Kasper-Giebl6, J.-L. Jaffrezo7, F. Bianchi1,b, J. G. Slowik1, U. Baltensperger1, I. El-Haddad1, and A. S. H. Prévôt1 1Paul Scherrer Institute (PSI), 5232 Villigen-PSI, Switzerland. 2Università degli Studi di Milano, 20133 Milano, Italy. 3Università degli Studi di Milano-Bicocca, 20126 Milano, Italy. 4ARPA Lombardia, Regional Centre for Air Quality Monitoring, 20122 Milan, Italy. 5Swiss Federal Laboratories for Material Sciences and Technology, 8600, Dübendorf, Switzerland 6Institute of Chemical Technologies and Analytics, Vienna University of Technology, 1060 Wien, Austria. 7Université Grenoble Alpes, CNRS, IGE, 38000 Grenoble, France. anow at: water and soil lab, 24060 Entratico, Italy. bnow at: Department of Physics, University of Helsinki, 00014 Helsinki, Finland.
Pulished in Atmospheric Chemistry and Physics DOI: 10.5194/acp-17-13265-2017
![Page 75: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/75.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
58
Abstract Long-term monitoring of organic aerosol is important for epidemiological studies,
validation of atmospheric models, and air quality management. In this study, we apply a
recently developed filter-based offline methodology using an aerosol mass spectrometer
(AMS) to investigate the regional and seasonal differences of contributing organic
aerosol sources. We present offline AMS measurements for particulate matter smaller
than 10 μm at nine stations in central Europe with different exposure characteristics for
the entire year of 2013 (819 samples). The focus of this study is a detailed source
apportionment analysis (using positive matrix factorization, PMF) including in-depth
assessment of the related uncertainties. Primary organic aerosol (POA) is separated in
three components: hydrocarbon-like OA related to traffic emissions (HOA), cooking
OA (COA), and biomass burning OA (BBOA). We observe enhanced production of
secondary organic aerosol (SOA) in summer, following the increase in biogenic
emissions with temperature (summer oxygenated OA, SOOA). In addition, a SOA
component was extracted that correlated with an anthropogenic secondary inorganic
species that is dominant in winter (winter oxygenated OA, WOOA). A factor (sulfur-
containing organic, SC-OA) explaining sulfur-containing fragments (CH3SO2+, which
has an event-driven temporal behaviour, was also identified. The relative yearly average
factor contributions range from 4 to 14% for HOA, from 3 to 11% for COA, from 11 to
59% for BBOA, from 5 to 23% for SC-OA, from 14 to 27% forWOOA, and from 15 to
38% for SOOA. The uncertainty of the relative average factor contribution lies between
2 and 12% of OA. At the sites north of the alpine crest, the sum of HOA, COA, and
BBOA (POA) contributes less to OA (POA/OA=0.3) than at the southern alpine valley
sites (0.6). BBOA is the main contributor to POA with 87% in alpine valleys and 42%
north of the alpine crest. Furthermore, the influence of primary biological particles
(PBOAs), not resolved by PMF, is estimated and could contribute significantly to OA in
PM10.
5.1 Introduction
The development and field deployment of the Aerodyne aerosol mass
spectrometer (AMS; Canagaratna et al., 2007) have greatly improved air quality
monitoring by providing real-time measurements of the non-refractory (NR) submicron
![Page 76: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/76.jpg)
59
aerosol (PM1) components. The application of factor analysis on the collected organic
aerosol (OA) mass spectra enabled the efficient disentanglement of aerosol factors,
which could be subsequently related to specific aerosol sources and processes (Lanz et
al., 2007, 2008; Jimenez et al., 2009; Ulbrich et al., 2009, Zhang et al., 2011; Ng et al.,
2011b; Crippa et al., 2014). Factors typically extracted include directly emitted primary
OA (POA) from biomass burning (BBOA) or traffic (HOA), and oxygenated OA
(OOA) that is typically associated with secondary OA (SOA), formed through the
oxidation of organic vapour precursors or heterogeneous processes. The model is not
capable of identifying the main SOA precursors, but often differentiates OOA based on
its volatility and degree of oxygenation (semivolatile fraction and low-volatility
fraction) due to the available highly time-resolved data.
However, the cost and operational requirements of the AMS make its deployment
impractical throughout a dense monitoring network and over longer time periods. As a
result, most available datasets are often limited to a few weeks of measurements, and
factors are extracted mainly based on diurnal variations in POA emission strength and
SOA oxygen content (Zhang et al., 2011; El Haddad et al., 2013b). Highly mobile
measurements on platforms as aircrafts (e.g. DeCarlo et al., 2008) or vehicles (e.g.
Mohr et al., 2011) are designed for regional studies, but are even more limited by cost,
availability, and time than stationary studies. This hinders the determination of the
aerosol regional and seasonal characteristics and evaluation of long-term emission
trends, limiting the information required for model validation and development of
efficient mitigation strategies. Furthermore, the negligible transmission efficiency of the
AMS inlet for coarse particles prevents the characterization of their chemical nature and
contributing sources.
The recent development of the aerosol chemical speciation monitor (ACSM; Ng
et al., 2011c, Fröhlich et al., 2013) has enabled the establishment of dense networks of
long-term AMS-type measurements and source apportionment of the organic aerosol
(e.g. Crippa et al., 2014, using AMSs for shorter campaigns within the EUCAARI
project or EMEP/ACTRIS projects for longer multi-season campaigns using ACSMs).
However, the mass spectrometers used by the ACSMs have far lower mass resolution
than the AMS, reducing their performance for OA characterization and source
![Page 77: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/77.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
60
apportionment. An alternate monitoring strategy involves extending AMS spatial and
temporal coverage by measuring the nebulized water extracts of filter samples
(Daellenbach et al., 2016; Mihara and Mochida, 2011). This approach allows the
retroactive investigation of specific events, e.g. haze events in China (Huang et al.,
2014), as well as AMS measurements of coarse-mode aerosol (Bozzetti et al., 2016) and
long-term source apportionment studies (Bozzetti, 2017a, b). Such an approach was also
used in recent studies for identifying the different types of water-soluble chromophores
(Chen et al., 2016). Additionally, such filters are routinely collected and are already
available over multi-year periods at many air quality monitoring stations around the
world for years and/or decades. Unlike single-season online AMS studies, the offline
AMS analysis of filter samples may reveal seasonal and long-term variations in the
emissions of POA and SOA precursors required for model validations and the
establishment of efficient mitigation strategies.
Here, we present offline AMS measurements of PM10 (particulate matter with an
aerodynamic diameter smaller than 10 μm) at nine stations in central Europe with
different exposure characteristics for the entire year of 2013 (819 samples). The sites
cover rural and urban locations, including urban background and traffic and wood-
burning-influenced stations. Such long-term multi-site analyses allow the quantitative
description of the temporal and spatial variability in the main OA sources and may
provide further insights into SOA precursors and formation pathways. This paper
focuses on the identification of the main factors influencing the OA concentrations at
the different sites and the assessment of the associated uncertainties. In a second paper,
we will investigate the site-to-site differences and general trends in the factor time series
and their relationship with external parameters.
5.2 Methods
5.2.1 Study area and aerosol sampling
PM10 samples were collected at nine sites in Switzerland and Liechtenstein
(Table 5.1 and Fig. 5.1). Seven of the sites (Basel, Bern, Payerne, Zürich, Frauenfeld,
St. Gallen, Vaduz) are located in northern Switzerland and Liechtenstein and two
(Magadino and San Vittore) in southern Switzerland. Aerosol was sampled at the
![Page 78: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/78.jpg)
61
selected sites every fourth day for 24 h throughout the year 2013 on quartz fibre filters
(14.7 cm diameter) using high-volume samplers (500 L min-1). Filters were then
wrapped in aluminium foil or lint-free paper and stored at -20 °C. Field blanks were
collected following the same approach.
Figure 5.1: Map of study area with locations of sites indicating their characteristics. The topography is displayed as meters above sea level.
Table 5.1: Study sites with geographical location and classification Site (station code) Classification General location altitude Basel, St. Johann (bas) Urban/background North of Alps/Swiss plateau 308 m. Bern, Bollwerk (ber) Urban/traffic North of Alps/Swiss plateau 506 m. Frauenfeld, (fra) Suburban/background North of Alps/Swiss plateau 403 m. Payerne (pay) Rural/background North of Alps/Swiss plateau 539 m. St. Gallen,(gal) Urban/traffic North of Alps/Swiss plateau 457 m. Zürich, Kaserne (zue) Urban/background North of Alps/Swiss plateau 457 m. Vaduz, Austrasse (vad) Urban/traffic North of Alps/alpine valley 706 m. Magadino, Cadenazzo (mag) Rural/background South of Alps/alpine valley 254 m. San Vittore, (vi) Rural/traffic South of Alps/alpine valley 330 m.
![Page 79: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/79.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
62
5.2.2 Offline AMS analysis
The offline AMS analysis summarized below was carried out following the
methodology developed by Daellenbach et al. (2016). For each analysed filter sample,
four 16mm diameter filter punches were sonicated together in 10mL ultrapure water
(18.2MΩcm, total organic carbon TOC < 5 ppb, 25 °C) for 20 min at 30 °C. Liquid
extracts were then filtered (0.45 μm) and nebulized in synthetic air (80 % volume N2,
20% volume O2; Carbagas, Gümligen CH-3073 Switzerland) using a customized Apex
Q nebulizer (Elemental Scientific Inc., Omaha, USA) operating at 60°C. The resulting
droplets were dried using a Nafion® dryer and then injected and analysed using the
high-resolution time-of-flight AMS (HR-ToF-AMS). Three types of measurements
were performed: (i) filter samples, (ii) field blanks (collected and treated in the same
way as the exposed filters), and (iii) measurement blanks (nebulized ultrapure water
without filter extract). The measurement blank was determined before and after every
filter sample. Each sample was recorded for 480 s (AMS V-mode, m/z 12-447), with a
collection time for each spectrum of 30 s. Ultrapure water was measured for 720 s.
Once per day, ultrapure Milli-Q water was nebulized with a particle filter interposed
between the nebulizer and the AMS for the determination of the gas-phase contribution
to the measured mass spectrum, which was then subtracted during analysis from both
blanks and filter samples. The filters from Zürich were analysed twice with a time
difference of approximately 5 months to assess the measurement repeatability. High-
resolution mass spectral analysis was performed for each m/z (mass-to-charge) in the
range of 12-115. The measurement blank was subtracted from the sample spectra. In a
previous study, it has been shown that the measurement blank is comparable to the
organic blanks obtained from the nebulization of NH4NO3 (Bozzetti et al., 2017b). The
interference of NH4NO3 in the CO2+ signal described by Pieber et al. (2016) was
corrected as follows (Eq. 5.1):
(5.1)
![Page 80: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/80.jpg)
63
The correction factor was determined based on
measurements of aqueous NH4NO3 conducted regularly during the entire measurement
period and varied between <1% and up to approximately ~5% (Pieber et al., 2016).
5.2.3 Other chemical analysis
Organic and elemental carbon (OC, EC) content were measured using a thermo-
optical transmission method with a Sunset OC/EC analyser (Birch and Cary, 1996),
following the EUSAAR-2 thermal-optical transmission protocol (Cavalli et al., 2010).
Water-soluble carbon was measured with water extraction followed by catalytic
oxidation, nondispersive infrared detection of CO2 using a total organic carbon analyser,
only for the samples from Magadino and Zürich. Water-soluble ions (K+, Na+, Mg2+,
Ca2+, and NH4+ and SO4
2-, NO3-, and Cl- and methane sulfonic acid were analysed using
ion chromatography (Piazzalunga et al., 2013 and Jaffrezo et al., 1998). Levoglucosan
measurements were performed with a high-performance anion exchange
chromatographer (HPAEC) with pulsed amperometric detection (PAD) using an ion
chromatograph (Dionex ICS-1000) following Piazzalunga et al. (2010 and 2013). Free
cellulose was determined using an enzymatic conversion to D-glucose (Kunit and
Puxbaum, 1996) and subsequent determination of glucose with an HPAEC (Iinuma et
al., 2009). Online measurements of gas-phase compounds and meteorology were also
performed at selected sites.
5.3 Source apportionment
5.3.1 General principle
Source apportionment of the organic aerosol is performed using positive matrix
factorization (PMF; Paatero and Tapper, 1994). PMF is a statistical un-mixing model
explaining the variability in the organic mass spectral data (xi,j as linear combinations of
static factor profiles (fj,k and their timedependent contributions (gi,k); see Eq. (5.2)
(where p is the number of factors). The index i represents a specific point in time, j an
ion, and k a factor. The elements of the model residual matrix are termed ei,j.
![Page 81: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/81.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
64
(5.2)
In the input data matrix, each filter sample was represented on average by 11 mass
spectral repetitions to examine the influence of the AMS measurement repeatability on
the PMF outputs. A preceding blank from nebulized ultrapure water was subtracted
from each mass spectrum. The input errors si,j required for the weighted least-squares
minimization by the model consist of the blank variability (σi,j and the uncertainty
related to ion counting statistics and ion-to-ion signal variability at the detector (δi,j
Allan et al., 2003; Ulbrich et al., 2009). We applied a minimum error according to
Ulbrich et al. (2009) and a down-weighting factor of 3 to all fragments with an average
signal-to-noise ratio lower than 2 (Ulbrich et al., 2009). Input data and the
corresponding error matrices consisted of 202 organic ions. The organic fragments, x’i,j,
obtained from offline AMS analyses do not directly represent ambient concentrations.
Therefore, the signal of each fragment was converted to such an ambient concentration
(xi,j in µgm-3) by multiplying the fraction of this signal with the estimated organic matter
(OM) concentration. The latter was calculated as the product of the OC concentrations
measured by the Sunset OC/EC analyser and the OM/OC from the offline AMS
measurements (OM/OC)oAMS (Eq. 5.3). Note that such scaling does not change the
outcome of Eq. (5.2) since both data and error matrices are scaled in the same manner
and the fingerprints (fk,j) are not changed.
(5.3)
The Source Finder toolkit (SoFi v.4.9; Canonaco et al., 2013) for the Igor Pro
software package (Wavemetrics, Inc., Portland, OR, USA) was used to configure the
PMF model and for post-analysis. The PMF algorithm was solved using the multilinear
engine-2 (ME-2; Paatero, 1999). Normalization of the PMF solution during the iterative
minimization process is disabled as implemented in SoFi (Canonaco et al., 2013). ME-2
enables an efficient exploration of the solution space by constraining the fk,j elements a
![Page 82: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/82.jpg)
65
priori within a certain range defined by the scalar a (0 ≤ a ≤ 1) from a starting value f’k,j,
such that the modelled fk,j in the solution satisfies Eq. (5.4):
(5.4)
f’k,j is the starting value used as a priori knowledge from previous studies and fk,j is
the resulting value in the solution. In all PMF runs (unless mentioned otherwise), we
used the high-resolution mass spectra for HOA and COA (cooking OA) from Crippa et
al. (2013c) as constraints, i.e. two rows of f’k,j were set equal to the mass spectra of
HOA and COA. Ions that were present in our datasets but not in the reference profiles
for HOA and COA were inferred from published unit mass resolution (UMR) profiles
(Ng et al., 2011b and Crippa et al., 2013b). For this purpose, the fraction of signal at a
specific m/z in the UMR reference spectrum (fUMR,m/z was compared to the fraction of
signal of all ions at this m/z in the HR reference spectrum (fHR,m/z). The difference
fUMR,m/z - fHR,m/z was used as an entry in f’k,j for such missing ions. For these ions, an a
value of unity was set. For the other factors, the factor elements were fitted using ME-2.
Alternatively, such missing ions can be also treated as ordinary factor elements, to be
fitted using ME-2 with all other ordinary factor elements.
Source apportionment analysis was performed following the scheme shown in
Fig. 5.2 and discussed below. Unconstrained and constrained exploratory PMF runs
provided information on the number of interpretable factors (Sect. 5.3.2). Multiple
constrained PMF runs were then performed to assess the model sensitivity to the chosen
a value, the model starting point and input matrix (entire dataset: PMFblock; only Zürich:
PMFzue,isol; one filter per site and month: PMF1filter/month; repeated measurements for
Zürich: PMFzue,reps and repeated measurements (Sect. 5.3.3). The factors obtained were
then classified and corrected for their recovery (Sect. 5.3.4 and 5.3.5). Finally, the
different solutions were evaluated and only the solutions that satisfied a set of
predefined criteria (Sect. 5.3.6 and Supplement A) were considered.
![Page 83: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/83.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
66
5.3.2 Preliminary PMF
Figure 5.2: Step-by-step outline of adopted source apportionment approach (factor recoveries Rk).
aHOA and aHOA represent the a value applied for HOA and COA, respectively.
We explored constrained PMF solutions, ranging from 1 to 10 factors. This
investigation is performed on the entire dataset, including all stations and seasons
(details in the Supplement A). The impact of the number of factors on the residuals is
examined in the Supplement A. The introduction of two factors, in addition to HOA and
COA, resulted in a significant reduction in the residuals and the separation of BBOA
and OOA contribution. BBOA exhibited a prominent seasonal variation with a
significant increase during winter and contributed most to the explained variation in the
fragment C2H4O2+, originating from the decomposition of anhydrous sugars, i.e. from
cellulose pyrolysis. OOA was identified based on its mass spectral fingerprint, with
high contribution from oxygenated ions at m/z 43 and 44. A further increase in the
number of factors did not significantly contribute to the reduction in the residuals.
However, the introduction of a fifth factor allowed the separation of the OOA into two
different factors, with distinct seasonal variability and different relative contributions
from oxygenated fragments at m/z 43 and 44. The two OOA factors will be referred to
as winter and summer OOA (WOOA and SOOA) according to their seasonality. The
introduction of a sixth factor allowed the resolving of a factor with a distinct time series
explaining the variability of sulfur-containing fragments (e.g. CH3SO2+. This factor will
be referred to as sulfur-containing organic aerosol (SC-OA). We explored higher-order
solutions, but could not interpret the resulting factor separations. Therefore, we further
consider a six-factor solution below.
![Page 84: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/84.jpg)
67
5.3.3 Sensitivity analysis
We assessed the model sensitivity to the chosen a value for HOA and COA and
the model starting point (independently for all four PMF inputs, as described below).
The a values were independently varied for HOA and COA (a value from 0 to 1 with
increments of 0.1, giving 121 a-value combinations). For every a -value combination,
the model was initiated from five different pseudo-random starting points (seeds),
yielding 605 total runs. As the selection of the a value combination was randomized, the
process was repeated four times in order to ensure that every a-value combination was
represented at least once (2420 runs), which in turn provided an assessment of the seed
effect on the results.
While this approach has been proven very effective in selecting a range of
environmentally relevant solutions (Elser et al., 2016a, b, and Daellenbach et al., 2016),
the resulting uncertainties may be underestimated. Paatero et al. (2014) compared the
effectiveness in estimating uncertainties of factor elements using two different
approaches: the displacement (DISP) and bootstrap analysis (BS). BS involves applying
the model to input matrices consisting of a subset of the entire dataset. DISP involves
running PMF several times using systematically perturbed factor profile elements of a
reference solution, but allowing a defined difference in Q from the reference solution.
Both approaches are computationally intensive, especially DISP. Because of such
computational limitations, the combination of BS and DISP was not feasible for the
dataset presented here, especially in combination with a-value sensitivity tests.
Therefore, we chose to perform four sensitivity tests performing PMF runs using four
different input datasets, presented in the following. These sensitivity tests allow
conclusions on the stability of PMF analysis when reducing the temporal or spatial
resolution as well as the influence of the measurement repeatability.
1. PMFblock: PMF was performed on data from all seasons and all sites
combined (all measured in October 2014). The corresponding data and error
matrices involved 819 samples from nine sites with 202 ions and, on average,
11 spectra per sample. This represents the base case.
![Page 85: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/85.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
68
2. PMFzue,isol: PMF was performed on data from Zürich alone (isolated from
PMFblock input). The corresponding data and error matrices involved 91
samples with 202 ions and on average 11 spectra per sample.
3. PMF1filter/month: PMF was performed on data from all sites but only
considering the 1st filter collected for every month (12 filters per site), as for
these samples levoglucosan and cellulose data was available. The
corresponding data and error matrices involved 108 samples with 202 ions
and on average 11 spectra per sample.
4. PMFzue,reps: PMF was performed on data from the repeated measurements of
Zürich samples. The corresponding data and error matrices comprised 91
samples with 196 ions and on average 14 spectra per sample.
For each of the four PMF datasets, 2420 PMF runs were performed for evaluating
the sensitivity of the model to the chosen a value and the seed. The quality of each of
the 2420 PMF runs was individually assessed using criteria lined out in Sect. 5.3.6.
5.3.4 Factor classification
From the sensitivity analysis, a large number of solutions were generated.
Systematic analysis of these solutions required automatic identification and
classification of the retrieved factors within each solution. We applied a sequential
classification algorithm as follows. Since HOA and COA were initially constrained on
preselected rows of fk,j, they did not need to be identified. In a second step, the factor
showing the highest explained variation for C2H4O2+ among the four remaining factors
was identified as BBOA. In a third step, the factor with the highest explained variation
for CH3SO2+ among the three remaining factors was identified as SC-OA. From the last
two factors, the one with the highest explained variation in CO2+ was identified as
WOOA and the other as SOOA.
5.3.5 Recovery and blank corrections
After factor identification, factor time series are corrected using factor-specific
recoveries (Eq. 5.5, resulting in OAi,k) determined in Daellenbach et al. (2016) for
HOA, COA, BBOA, and OOA.
![Page 86: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/86.jpg)
69
(5.5)
where gi,k values are the concentrations of factor k at the time point i, Rk the
recoveries of the respective factor, and OAi the OA concentration. For a limited number
of PMF runs (PMFblock, the field blank analyses were also included in the PMF input
data. This provides the contributions of different factors to the field blanks, which were
used to correct the output factor time series. Uncertainties induced by the blank
subtraction were propagated.
5.3.6 Solution selection
Each of the 2420 PMF solutions per PMF dataset (PMFblock, PMFzue,isol,
PMFzue,reps, PMF1filter/month was evaluated based on their factor profiles, time series, and
the OC mass closure. Solutions were selected if they satisfied the following set of
criteria:
1. fCO2+ < 0.04 in HOA and COA factor profiles (HOA based on Aiken et al.,
2009, Mohr et al., 2012, Crippa et al., 2013c, 2014 and COA based on Crippa
et al., 2013b, 2013c, Mohr et al., 2012);
2. fC2H4O2+ < 0.004 and 0.01 in HOA and COA, respectively (HOA based on
Aiken et al., 2009, Mohr et al., 2012, Crippa et al., 2013b, 2014 and COA
based on Crippa et al., 2013b, 2013c, Mohr et al., 2012);
3. HOA correlates significantly with NOx being the sum of NO and NO2
(defined below);
4. HOA correlates significantly better with NOx than COA; BBOA correlates
significantly with levoglucosan (defined below);
5. SC-OA correlates significantly with CH3SO2+ (defined below);
6. for samples from Zürich and Magadino, where watersoluble organic carbon
(WSOC) data are available, modelled and measured OC mass are comparable
for a set of different conditions (see below and in Supplement A).
The first two criteria (1-2) ensure an appropriate separation of HOA and COA
from OOA and BBOA, respectively. Criteria 3-5 relate to the evaluation of the
![Page 87: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/87.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
70
correlation between factor and marker time series. This was achieved by computing the
Fisher-transformed correlation coefficient z at different stations (Eq. 5.6):
(5.6)
where r is the correlation coefficient between the factor and marker at a given
station. The z values obtained at the different stations are subsequently averaged and
transformed back to ravg before further analysis. A t test is then used to verify the
significance (α = 0.5) of the average correlation coefficient between factor and marker
time series, ravg (Eq. 5.7):
(5.7)
Here, ravg is the correlation coefficient averaged over the different stations,
derived from the average z value, tavg is the corresponding t value, and N is the average
number of samples at the different stations. Results with a significance level of α = 0.05
are summarized in Fig. SI.A.8 in the Supplement A.
To evaluate whether HOA correlated significantly better with NOx than COA did,
the average z values obtained between HOA and NOx and between COA and NOx (Eq.
5.6) were compared using a standard error on the z distribution of (Zar,
1999). The last criterion (6) relates to OC mass closure. A Monte Carlo approach was
applied to evaluate whether a combination of water-soluble factor time series and
recovery parameters would achieve OC mass closure, as described in the following. For
the samples from Zürich and Magadino, for which WSOC concentrations were available
(in contrast to the other samples), offline AMS measurements were scaled to the water-
soluble organic matter (WSOM) calculated using the WSOC measurements and
OM/OC from the HR AMS analysis. The water-soluble contributions from an identified
aerosol source in a sample i were rescaled to their total organic matter concentrations
(OAi,k, where k represents a given factor, using combinations of factor recoveries as
![Page 88: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/88.jpg)
71
determined by Daellenbach et al. (2016, medians of the combinations used being RHOA:
0.11, RCOA: 0.54, RBBOA: 0.65, and ROOA: 0.89, used for WOOA and SOOA). For SC-
OA, whose recovery was not previously determined, a recovery value was stochastically
generated between 0 and 1. The OAi,k concentrations obtained were then converted to
OC concentrations OCi,k, using factor-specific OM/OC determined from the factor
profiles. The sum of OCi,k from all factors k (mod-OCi was then evaluated against the
measured OC (meas-OCi. For this, the residual OC mass (res-OCi for each sample was
calculated (meas-OCi - mod-OCi), and the residual distributions were examined for
different conditions that are specified in the Supplement A. In summary, a solution was
only accepted if res-OCi values were normally distributed around 0 considering all
points and subsets of points: (a) summer, (b) winter, (c) Magadino, (d) Zürich, and (e)
low and high concentrations of the single factors (see Table SI.A.1 in the Supplement
A).
For each of the Monte Carlo simulations, criteria 1-6, which satisfy the water-
soluble factor time series, were used together with a combination of factor recoveries
from Daellenbach et al. (2016) as input data. The WSOC used for scaling the Gi,k matrix
and the meas-OCi used for residual calculation were varied within their uncertainties
(5%) and biases (5%) assuming a normal distribution of the errors. Likewise, constant
biases were also introduced into the initial recovery distributions from Daellenbach et
al. (2016). Monte Carlo simulations were performed and simulations for which res-OCi
distributions were significantly different from 0 (Q25<0<Q75, details in the Supplement
A) were discarded until 500 acceptable simulations were found. Thereby, 331 PMF runs
were selected for PMFblock (230 for PMFzue,isol, 99 for PMFzue,reps, and 269 for
PMF1filter/month. Median factor time series and recovery parameters from all retained
simulations were then determined and the interquartile range (IQR) represents our best
estimate of the uncertainties for the single PMF datasets. The Monte Carlo process was
repeated for the four different PMF datasets described above and the resulting median
time series of their estimated uncertainties were compared. The resulting uncertainty
estimates and the method are described in Sect. 5.4.2.1 and in the Supplement A.
![Page 89: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/89.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
72
5.4 Results and discussions
In this section, the final source apportionment results are presented and validated.
The source signatures are presented in Fig. 5.3 for PMFblock colour-coded with the ion
family. Figure 5.4 shows the time series for Zürich obtained from all PMF approaches
and Table 5.2 summarizes the correlation coefficients between factor and marker time
series for Zürich (all PMF runs) and the other sites in the study area (PMFblock, while the
relation between factor and marker time series is displayed in Figs. 5.4 and 5.5.
Presented are median (and quartile) results for all PMF runs accepted following the
criteria described above.
5.4.1 Interpration of PMF factors
HOA: HOA profile elements were constrained using the reference profile from
Crippa et al. (2013c). The final factor profile (Fig. 5.3) maintains the same features,
characterized by high contributions from hydrocarbon fragments. The fraction of
oxygenated organic fragments that were missing in the initial reference profile, which
were added based on UMR spectra, show an increased contribution to the ions above
m/z=100 (see Sect. 5.3.1). While this indicates a possible overestimation of the
contribution of these fragments, using this methodology, this increase does not
substantially affect the results: e.g. the HOA OM/OC remains low (1.32, IQR 1.30-
1.33). The HOA time series follows an expected pattern that matches the NOx yearly
cycle (Fig. 5.4a), except for San Vittore, which is very likely due to the extremely high
contribution of biomass burning at this site during winter, which may result in
additional NOx inputs and/or may affect the separation of HOA by PMF. The
HOA/NOx (Fig. 5.5a) at the different sites (0.015 - 0.011 µgm-3ppb-1) lies within the
range of literature values (0.001 to 0.028 µgm-3ppb-1, Lanz et al., 2007 and Kirchstetter
et al., 1999). A similar average ratio was obtained for Zürich from the different
sensitivity tests, but with high variability (0.013 - 0.009 µgm-3ppb-1) similar to that
obtained between the different sites. This implies that the observed site-to-site
differences are not statistically significant given our uncertainty in extracting HOA
contributions.
![Page 90: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/90.jpg)
73
Figure 5.3: PMF factor profiles of HOA, COA, BBOA, SOOA, WOOA, SC-OA, color-coded with
ion family of PMFblock (average). fm/z is the relative intensity at a specific mass-to-charge ratio (m/z).
COA: COA profile elements were constrained using the COA profile from
Crippa et al. (2013c) and the obtained factor profile maintains the same features
(OM/OC of 1.32, IQR 1.30-1.33, Fig. 5.3). For COA, no molecular marker is available
for validation purposes. Daellenbach et al. (2016) demonstrated that COA
concentrations can be estimated with offline AMS (in Zürich at the same site) by
constraining its signatures, but only with a high uncertainty. This was performed by
comparing offline AMS results to those from a collocated ACSM, which, owing to its
higher time resolution, enabled the identification of cooking emissions based on their
diurnal cycles (Canonaco et al., 2013). Here, while no ACSM data were available, we
followed the same methodology used in Daellenbach et al. (2016) to estimate the
contribution of COA. The average COA contributions estimated here and their yearly
variability are similar to those from previous studies at the same sites, but as expected
have high uncertainties (Fig. 5.4b).
![Page 91: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/91.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
74
Figure 5.4: HOA, COA, BBOA, SC-OA, SOOA, and WOOA and their respective marker concentrations as a function of time for Zürich in 2013. Depicted are the median factor time series results for the different PMF datasets (median) including the uncertainties for PMFblock (first and third quartile) (green: PMFblock, black: PMFzue,isol, red: PMFzue,reps, pink bullets: PMF1filter/month).
BBOA: BBOA is identified based on its spectral fingerprint (OM/OC of 1.74,
IQR 1.74-1.75; Fig. 5.3), which, similar to previously extracted BBOA factors at other
locations (Daellenbach et al., 2016; Lanz et al., 2007; Crippa et al., 2014), exhibits high
contributions from oxygenated fragments (CHO+, C2H4O2+, C3H5O2
+ from anhydrous
sugar fragmentation (see comparison to nebulized levoglucosan in Supplement A Fig.
SI.A.6). Similar to levoglucosan, the BBOA time series shows an expected seasonal
variation with high concentrations in winter, supporting the identification of this factor
(Fig. 5.4c). Except for Bern and Magadino (7.5 and 11.2), a similar ratio of BBOA to
levoglucosan is found at all other sites (3.9 to 5.7), despite apparent site-to-site
differences in the model residuals during winter due to significantly higher
contributions of BBOA at the southern stations (Fig. 5.5c). The ratios obtained are
within the range of values reported in literature (between 4 and 18 assuming OM/OC)
between 1.6 and 1.8 for the non-AMS analyses; Zotter et al., 2014; Herich et al., 2014;
Minguillón et al., 2011; Crippa et al., 2013a; and Favez et al., 2010). We note that a
similar ratio is also found for the different PMF datasets performed for the case of
Zürich (BBOA/levoglucosan between 3.9 and 12.1). Taken together, the high (for most
![Page 92: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/92.jpg)
75
sites) correlation (R2 = 0.78 for all sites, single sites in Table 5.2) between levoglucosan
and BBOA and their consistent ratios at different sites and between the different PMF
datasets indicates that BBOA is well resolved by PMF at all sites, despite potential site-
to-site differences in BBOA composition.
Figure 5.5: Scatter-plots for the different extreme sensitivity tests for Zürich and for all sites for PMFblock median concentrations): a) HOA vs NOx, b) BBOA vs levoglucosan, c) SOOA vs temperature, d) WOOA vs NH4
+.
SC-OA: Sulfur-containing fragments (e.g. CH3SO2+ are predominantly
apportioned to this factor, which also has a high OM/OC (1.82, IQR 1.80-1.93; Fig.
5.3). As mentioned in Sect. 5.3.6, the recovery of SC-OA was unknown and had to be
determined by mass closure, while the recoveries of the other factors were determined
by comparison to their online counterparts (albeit for a different dataset; Daellenbach et
al., 2016). In the lack of specific constraints (like an online counterpart), the recovery of
SC-OA is highly uncertain and thus the factor time series is also highly uncertain. A
similar factor profile had been extracted from previous online AMS datasets and was
related to the fragmentation of methane sulfonic acid (MSA) present in PM1 particles, a
secondary product of marine origin (Crippa et al., 2013c; Zorn et al., 2008). However,
the SC-OA factor extracted here did not seem to be related to marine emissions because
neither its variability nor its levels matched those of MSA (Fig. 5.4d). First we
compared the MSA levels measured in Zürich using ion chromatography to those
![Page 93: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/93.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
76
estimated based on the concentration of sulfur-containing fragments from offline AMS
measurements in SC-OA (Eq. 5.8), based on Crippa et al. (2013c):
(5.8)
Here, MSAi,est is the estimated MSA concentration, SC-OAi the factor concentration
of the sulfur-containing factor, fSC-OA(CH2SO2+) and the following summands the
fractional contributions of the respective organic fragment to SC-OA, and 0.147 is a
scaling factor from Crippa et al. (2013c). The estimated MSA levels are 6 times higher
than the measured MSA, indicating the presence of another source of sulfur-containing
species. Second, unlike marine OA factors from previous online datasets (lower size
cut-off, typically PM1), the SC-OA time series does not correlate with MSA (R2=0.02).
While MSA concentrations show a clear enhancement during summer, the SC-OA time
series exhibit a very weak seasonal variability with slightly higher concentrations in
winter. SC-OA instead exhibits low background levels episodically intercepted by
remarkable 10-fold enhancements, especially at urban sites affected by traffic emissions
(e.g. the SC-OA contribution is significantly higher at sites with higher yearly NOx
average levels). The hypothesis of an influence of traffic activity on SC-OA is provided
by the correlation of the yearly average concentrations with NOx (Rs,SC-OA,NOx=0.65,
n=9, p<0.06), which is, however, comparable to the correlation of HOA and COA (e.g.
Rs,HOA,NOx=0.68, n=9, p<0.05; Rs,COA,NOx=0.68, n=9, p<0.05). In addition, the SC-OA
time series also correlates with that of NOx (overall R2=0.32, for sites in Table 5.2).
While HOA and BBOA also correlate with NOx, both of the secondary factors, WOOA
and SOOA, do not, supporting the hypothesis that SC-OA consists of locally emitted
anthropogenic (primary) OA. The site-to-site differences in SC-OA concentrations and
temporal behaviour suggest that this factor, which to the best of our knowledge is
reported here for the first time, is influenced by primary sources.
![Page 94: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/94.jpg)
77
Table 5.2: Comparison of factor time series to reference data for different PMF input datasets runs (by Pearson and Spearman correlation coefficient, Rp
2 and Rs). Displayed are the results for PMFblock unless stated otherwise. R2
(number of points)
HOA vs NOx, Rp
2 BBOA vs levo, Rp
2 WOOA vs NH4
+, Rp2
SOOA vs T, Rs
SC-OA vs NOx, Rp
2
Basel 0.31 (91) 0.91 (11) 0.66 (91) 0.70 (91) 0.17 (91) Bern 0.22* (90) 0.48 (12) 0.53 (90) 0.63 (90) 0.17 (90) Frauenfeld 0.40 (89) 0.73 (12) 0.77 (90) 0.63 (90) 0.28 (89) St. Gallen 0.23 (91) 0.39 (12) 0.78 (91) 0.72 (91) 0.50 (91) Magadino 0.18 (91) 0.55 (12) 0.54 (91) 0.72 (91) 0.63 (91) Payerne 0.48 (91) 0.65 (12) 0.44 (91) 0.68 (91) 0.17 (91) Vaduz 0.38 (91) 0.90 (12) 0.77 (91) 0.68 (91) 0.46 (91) San Vittore 0.02 (90) 0.99 (12) 0.36 (90) 0.76 (68) 0.01 (90) Zürich PMFblock 0.35 (91) 0.43 (12) 0.79 (90) 0.65 (91) 0.40 (91) PMFzue,iso 0.29 (91) 0.59 (12) 0.82 (90) 0.66 (91) 0.27 (91) PMFzue,rep (only 12 points)
0.32 (12) 0.23 (12) 0.84 (12) 0.85 (12) 0.01 (12)
PMF1filter/month 0.30 (91) 0.44 (12) 0.77 (90) 0.59 (91) 0.53 (91) * 1 outlier removed.
Oxygenated OA factors: Unlike oxygenated OA factors from limited-
duration intensive online campaigns characterized by a high temporal resolution in
which factor variability is thought to be primarily driven by volatility and/or local
oxidation reactions, OOA factors are resolved based on differences in their seasonal
behaviour: SOOA (in summer) and WOOA (in winter). The SOOA and WOOA mass
spectral signatures (Fig. 5.3) show similarities with OOA from earlier measurements
(Ng et al., 2011b; Canonaco et al., 2013, 2015), with high contributions of C2H3O+ and
CO2+ and high OM/OC, though SOOA (OM/OC=1.89, IQR 1.88-1.89) is less oxidized
than WOOA (OM/OC=2.12, IQR 2.11-2.14). The mass spectral fingerprints (Fig. 5.3),
the temporal behaviour (Fig. 5.4e and f), and the relation to markers (Fig. 5.5c and d) of
the two factors are in agreement with those from earlier work at other locations,
including Zürich (Daellenbach et al., 2016), Payerne (Bozzetti et al., 2016), and
Lithuania (Bozzetti et al., 2017b). This OOA separation appears to be typical for PMF
analysis of long-term, low-time resolution OA mass spectra of filter samples.
![Page 95: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/95.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
78
SOOA correlates significantly among the different sites (also south and north of
the alpine crest) and with local temperature (Fig. 5.5c). The SOOA exponential increase
with average daily temperatures from 5 to 30°C is consistent with the exponential
increase in terpene emissions, which are dominant biogenic SOA precursors (Guenther
et al., 2006). This is also consistent with the mass spectral fingerprint of this factor,
characterized by an fC2H3O+ of 0.10 and an fCO2+ of 0.13, which are similar to values
reported for chamber SOA from terpenes or at an urban location (Zürich) during
summer (Canonaco et al., 2015). A similar temperature dependence of biogenic SOA
concentrations has been observed for a terpene-dominated Canadian forest (Leaitch et
al., 2011) and for the case in Switzerland, using a similar source apportionment model
(Daellenbach et al., 2016; Bozzetti et al., 2016). Taken together, these observations
suggest that SOOA principally derives from the oxidation of biogenic precursors during
summer. Site-to-site SOOA concentrations were not statistically different within our
model errors, assessed from the different sensitivity tests for the case of Zürich.
Therefore, even though the behaviour of SOOA at the different sites studied here might
be controlled by various parameters, including tree cover, available OA mass, air mass
photochemical age, and oxidation conditions (e.g. NOx concentrations), temperature
seems to be the main driver of the SOOA concentrations. Indeed, the aforementioned
parameters may contribute, together with model and measurement uncertainties, to the
observed scatter in the data. Biogenic volatile organic compound emissions might even
be non-negligible in winter (Oderbolz et al., 2013; Schurgers et al., 2009; Holzke et al.,
2006). Therefore, significant wintertime SOOA concentrations are not in disagreement
with the hypothesized biogenic origin. The lower SOOA concentrations in the
temperature range between 7 and 12°C might be explained by often-occurring
precipitation in this temperature range. We note that relative uncertainties related with
SOOA increase with decreasing concentrations (Fig. 5.7). A small error in modelling
sources with high contributions (BBOA, WOOA) in winter can result in a large error of
SOOA with its small contribution during winter. Furthermore, some other sources like
primary biological OA (PBOA; see Sect. 5.4.2.2) might also mix into SOOA.
Compared to SOOA, the WOOA profile can be distinguished by a higher
contribution from CO2+ and a lower contribution from C2H3O+ (Fig. 5.3), similar to
OOA factors previously extracted in this region during winter based on ACSM
![Page 96: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/96.jpg)
79
measurements. This fingerprint is characteristic of highly oxidized SOA from non-
biogenic precursors with low H/C (e.g. aromatic compounds from wood combustion
emissions; Bruns et al., 2016).WOOA is well correlated with NH4+ (Fig. 5.5d; overall
R2=0.65 for all sites, overall R2=0.81 for all PMF runs for Zürich in Table 5.2), which is
in agreement with earlier studies (e.g. Zürich in Lanz et al., 2008). This is probably
explained by its correlation with other inorganic secondary ions NO3- and SO4
2- (driven
like WOOA by meteorological factors including boundary layer height and
temperature), which govern the NH4+ concentration in the aerosol. Here, we have used
ammonium as a proxy for aged aerosols affected by anthropogenic emissions, as
WOOA correlates better with ammonium than with nitrate sulfate.We note that in
winter, whenWOOA is highest, 56% of ammonium can be attributed to nitrate, whereas
in summer ammonium sulfate dominates (97% of ammonium can be attributed to
sulfate). Therefore, WOOA correlates more with nitrate (R2= 0.64) than sulfate
(R2=0.48). WOOA exhibits a regional behaviour and its concentrations are correlated at
all sites on the Swiss plateau. The WOOA mass spectral fingerprint, its seasonal
variability, and its high correlation with long-range transported anthropogenic inorganic
secondary ions suggest that this factor is characteristic of a highly aged OOA influenced
by wintertime anthropogenic emissions (e.g. biomass burning).
![Page 97: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/97.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
80
5.4.2 Uncertainty analysis
5.4.2.1 Model uncertainties
PMF uncertainties depend on the factor contribution. According to Ulbrich et al.
(2009), reliable interpretation of factors with a low relative contribution is challenging.
However, the specificity of the time series and factor profile (caused by rotational
ambiguity), and in this sense also solution acceptance criteria, influence the uncertainty
as well. In our analysis, we correct our results from WSOM to OM using Rk and thereby
introduce additional uncertainties (caused by the uncertainty of Rk or an unknown Rk.
The more uncertain Rk, the higher is the additional uncertainty in the extrapolation (Eq.
5.5). As mentioned in Sect. 5.3.5, Rk constraints (recovery combinations for different
factors) are available for RHOA, RCOA, RBBOA, and ROOA but not for RSC-OA and not for
individual OOA factors (Daellenbach et al., 2016). With the available constraints of
mass closure (for Magadino and Zürich), RSC-OA can only be determined with a high
uncertainty (Fig. 5.6).
Figure 5.6: Distributions of Rk for HOA, COA, BBOA, OOA (WOOA plus SOOA) and SC-OA (500 pairs). A priori information for HOA, COA, BBOA, and OOA on Rk is used from Daellenbach et al., 2016, with propagated errors and biases, while RSC-OA is determined in this study. Distributions of all factors have a resolution of dRk=0.01 except for dRSC-OA=0.05.
![Page 98: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/98.jpg)
81
Figure 5.7: Relative σa (a) and err‘tot (b) for factor concentrations > 0.1 µg/m3 as a function of factor concentration. err‘tot includes the uncertainties from a-value, seed variability and Rk, and the different PMF datasets.
The variability in the factor time series for the single PMF sensitivity tests
(PMFblock, PMFzue,isol, PMF1filter/month, PMFzue,reps is used as an uncertainty estimate
(shaded area in Fig. 5.4). This estimate (σa) depends on the measurement repeatability
(10 single mass spectra included for each sample) and on the selected PMF solution and
Rk combinations, and therefore also on the a value. However, the variability depending
(1) on the choice of input points (time and site; PMFblock, PMFzue,isol, PMF1filter/month and
(2) on the instrumental reproducibility (PMFzue,reps of the offline AMS measurements is
not accounted for. The contribution of (1) and (2) to the uncertainty is assessed through
the sensitivity tests by examining the variability of the median factor time series (σb). σb
![Page 99: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/99.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
82
is the variability of the median factor concentrations from the PMF sensitivity tests
using PMFblock, PMFzue,isol, PMF1filter/month, PMFzue,reps for the 12 samples common to all
4 PMF datasets. For the 12 filters common in all PMF datasets (PMFblock, PMFzue,isol,
PMF1filter/month, PMFzue,reps), we calculate a best estimate of the overall uncertainty (errtot,
by propagating both error terms: σa and σb. As σb is not available for all datapoints, we
parametrized σb as a function of the factor concentration (details in the Supplement A)
and subsequently used this parameterized σb, σ’b, to calculate an approximated overall
error, err’tot. err’tot is displayed in Fig. 5.7b in comparison with σa (Fig. 5.7a). For all
factors, err’tot are in general high, especially for low factor concentrations (~ a factor of
2). It is worthwhile to note that for major factors exhibiting a similar seasonality, i.e.
WOOA and BBOA, a great part of the uncertainty arises from σb. Thus the variability
between the PMF solutions using PMFblock, PMFzue,isol, PMF1filter/month, PMFzue,reps (σb)
and, therefore, the sensitivity of the factor concentrations on the chosen PMF dataset,
significantly contribute to the uncertainty. By contrast, for moderately soluble fractions
constrained in the PMF, COA and HOA, the major part of err’tot is related to σa.
5.4.2.2 Influence of unresolved primary biological OA.
Unresolved sources in PMF are an inherent uncertainty of source apportionment
analyses. As Bozzetti et al. (2016) show, PBOA can present considerable contributions
to OA in PM10 (constituting a large part of coarse OA). In the present analysis, PBOA
could not be separated by PMF (neither unconstrained nor using the mass spectral
signature from Bozzetti et al., 2016). This inability might be caused by the low water
solubility and the absence of PM2.5 filters in the dataset. Since these coarse particles are
only abundant in PM10 and not in PM2.5 or PM1, the presence of both PM10 and
PM2.5 samples, exhibiting a large gradient in PBOA, might allow an unambiguous
separation of PBOA. The aim of this section is to estimate the influence of PBOA on
the source apportionment results. A quantification of this fraction is, however, beyond
the scope of this paper. In the following, we estimate the influence of PBOA in three
alternative ways.
![Page 100: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/100.jpg)
83
Based on factor profiles: Bozzetti et al. (2016) identified the AMS fragment
C2H5O2+ as a possible tracer ion for PBOA. Based on the seasonality of SOOA (high in
summer and low in winter), one can assume that SOOA in this study is a linear
combination of PBOA and SOOA identified in PM2.5 and PM1. Based on the relative
contribution of the ion C2H5O2+ to the factor profiles of SOOA from this analysis and
literature profiles of PBOA and SOOA from Bozzetti et al. (2016, study site: Payerne),
we estimate that 17% of the watersoluble SOOA is in fact PBOA (between 2 and 23%
for the different sensitivity tests). Using this approach, we estimate that PBOA
contributes 0.30 µgm-3 during the warm months (site-to-site variability computed as the
standard deviation of the average concentration of all sites of 0.03 µgm-3). During the
same period, SOOA concentrations are 1.78 µgm-3 (site-to-site variability of 0.18µgm-3)
and OA concentrations are 4.32 µgm-3 (site-to-site variability of 0.44 µgm-3). This
approach is very uncertain, mainly due to the uncertainty in PBOA and SOOA profiles,
the assumption of a constant PBOA contribution to SOOA, and also the uncertainty of
RPBOA.
Based on coarse OC: Bozzetti et al. (2016) showed that coarse OC
(OCcoarse=OCPM10-OCPM2.5) in summer is dominated by PBOA for samples collected at a
rural site in Switzerland (Payerne). For a subset of the samples used in the present work,
OC in the PM2.5 fraction was also analysed (Basel, Bern, Magadino, Payerne, Zürich
accounting for 149 samples in total). For these samples, the OCcoarse contribution to OC
in the PM10 fraction is 16% higher in summer than in winter (site-to-site variability of
4%). This part of OC might be related to resuspension caused by traffic or emissions of
primary biological particles. The ion C2H5O2+ (indicator for PBOA) shows higher
concentrations with increasing OCcoarse concentrations. Therefore, this increment can
tentatively be ascribed to PBOA, which leads to a contribution of 0.55 µgm-3 to OC in
summer (site-to-site variability 0.16 µgm-3). This results in an average summer PBOA
concentration of 1.21 µgm-3 with a site-to-site variability of 0.39 µgm-3 when assuming
an OM/OC of 2.2 (or 0.66 ± 0.21 µgm-3, for OM/OC=1.2); OM/OC range according to
Bozzetti et al., 2016). ForMagadino (2014, Vlachou et al., 2017), OCcoarse represents 8%
of OC in PM10 in winter while this ratio is 25% in summer. It can be assumed that the
difference of 17% in summer can be attributed to PBOA. Extrapolating this estimate to
the overall dataset from 2013 considered in this study and assuming an OM/OC of 2.2,
![Page 101: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/101.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
84
PBOA contributes on average 0.97 µgm-3 to OA in PM10 in summer, with a site-to-site
variability of 0.13 µgm-3 (or 0.63 ± 0.07 µgm-3 OA with for OM/OC of 1.2).
Figure 5.8: Cellulose concentrations as a function of the season and site. For comparison literature data from other years is added European sites: Payerne (Bozzetti et al., 2016, error bars representingthe standard deviation of the measurements in June and July), Puy de Dôme, Schauinsland, Sonnblick, K-Puszta (Sanchez-Ochoa et al., 2007), Birkenes, Hyytiälä, Lille Valby, and Vavihill (Yttri et al., 2011).
Based on cellulose: It has previously been shown that free cellulose contributes
strongly to PBOA (25% of PBOA mass, for measurements made in Payerne during
summer 2012 and winter 2013; Bozzetti et al., 2016). Therefore, we can attempt to use
cellulose analyses on a subset of samples (the same one as for levoglucosan but Bern;
see Sect. 5.2.3) to estimate PBOA concentrations (Fig. 5.8). As seen for the case of
OCcoarse, cellulose concentrations also increase with higher C2H5O2+ concentrations. For
![Page 102: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/102.jpg)
85
the sites with cellulose measurements available (all sites in the study but Bern),
cellulose average concentrations of 0.17 µgm-3 (site-tosite variability of 0.08 µgm-3, in
the warm season 0.18 ± 0.07 µgm-3) are observed, which corresponds to 0.69 µgm-3
PBOA with a site-to-site variability of 0.34 µgm-3 (in the warm season
0.77 ± 0.29 µgm-3), using the cellulose/PBOA from Bozzetti et al. (2016). In this last
study conducted during summer (15 days in June-July 2012), PBOA concentrations of 3
µgm-3 on average (with cellulose concentrations of 0.8 µgm-3) were estimated, which is
clearly above the observation made here. However, Bozzetti et al. (2016) assessed a
shorter time period with diurnal resolution, instead of one sample per month as in the
present work. Cellulose concentrations from other European sites during other years are
consistent with the results in this study (Sanchez-Ochoa et al., 2007; Yttri et al., 2011).
In general, the background cellulose concentrations at the southern alpine sites are
higher and also the temporal behaviour deviates from that observed at the northern sites:
the maximal concentrations are not reached in July-August but rather in May or
October-November. The different seasonality might be caused by different agricultural
procedures. The higher background concentrations of cellulose for the southern Alpine
sites might be caused by interferences from wood burning, which in the absence of
glucose analyses cannot be excluded.
All these PBOA estimates (between 0.3 and 1.0 µgm-3 during the warm season)
are consistently lower than reported in Bozzetti et al. (2016), with a factor 3 to 10 times
lower depending on the site. One should keep in mind that these estimates are based on
limited datasets in both studies (30 samples in Bozzetti et al. (2016) and 12 samples
from the same site in this study).
![Page 103: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/103.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
86
Figure 5.9: Map of Switzerland with yearly cycles. Negative concentrations were set to 0 prior to normalization for display. The OA mass explained by the source apportionment analysis is termed OAexpl.
5.4.3 Factor relative contribution at different sites
In general, the seasonality of the factor time series is consistent for all nine sites in
the entire study area (Fig. 5.9). In summer, SOOA is the main contributor to OA, while
in winter POA (HOA+COA+BBOA) becomes more important, although WOOA still
contributes significantly. In comparison to the sites in northern Switzerland, OA in the
southern alpine valleys is dominated by BBOA in winter, while in the north WOOA
also plays a role. The different factors contribute 0.47±0.12 (HOA, average and site-to-
site variability), 0.31±0.13 (COA), 1.37±1.77 (BBOA), 0.67±0.31 (SC-OA), 1.11±0.23
(WOOA), and 1.31±0.13 (SOOA) µgm-3 for all sites during the entire year (Table 5.3).
In northern Switzerland, POA contributes less to OA (POA/OA=0.3) than in the
southern alpine valleys, where POA/OA is equal to 0.6. Among POA, BBOA is the
most important, with 87% of POA in the south and 42% in the north. The higher
relative contribution of BBOA to POA in the southern alpine valleys than at the
![Page 104: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/104.jpg)
87
northern sites supports the conclusion that the high BBOA concentrations (e.g. 2.45
µgm-3 in Magadino compared to 0.62 µgm-3 in Vaduz) are not only a consequence of
the meteorological situation in the valleys (strong thermal inversion close to the valley
ground) but mainly reflect the emission strength. SC-OA, which is possibly linked to a
local source of rather primary origin, shows clear site-to-site differences, with high
concentrations at a traffic site in Bern (1.25 µgm-3) and low concentrations at a rural site
in Payerne (0.26 µgm-3), for example. SOOA, believed to have strong influences from
biogenic SOA, shows consistently low concentrations at all sites for low temperatures
(0.76 ± 0.67 µgm-3 at 5-15°C) and clearly increased concentrations under warmer
conditions (4.85 ± 1.51 µgm-3 at 25-35 °C).
Table 5.3: Yearly average contribution and uncertainty of resolved factors for PMFblock run for the different sites and the average for all sites. The uncertainty is calculated based on the variability in the yearly averages from PMFblock and the variability between the 4 sensitivity tests. Factor contribution and uncertainty µg/m3 (%)
HOA COA BBOA SC-OA WOOA SOOA
Basel 0.65±0.23 (14)
0.35±0.19 (8)
0.72±0.15 (16)
0.51±0.24 (11)
1.08±0.24 (24)
1.21±0.30 (27)
Bern 0.61±0.23 (11)
0.59±0.29 (11)
0.64±0.14 (12)
1.25±0.45 (23)
1.21±0.28 (22)
1.11±0.29 (21)
Frauenfeld 0.56±0.22 (12)
0.28±0.19 (6)
0.64±0.14 (14)
0.96±0.35 (20)
0.98±0.22 (21)
1.30±0.32 (27)
St. Gallen 0.40±0.20 (11)
0.15±0.16 (3)
0.42±0.09 (1)
0.71±0.27 (19)
0.83±0.19 (22)
1.22±0.30 (33)
Magadino 0.41±0.20 (6)
0.27±0.21 (4)
2.45±0.50 (37)
0.41±0.20 (6)
1.53±0.32 (23)
1.54±0.35 (24)
Payerne 0.34±0.19 (9)
0.15±0.16 (4)
0.54±0.12 (15)
0.26±0.16 (7)
1.00±0.22 (27)
1.41±0.33 (38)
Vaduz 0.43±0.20 (10)
0.27±0.19 (6)
0.62±0.14 (14)
0.84±0.30 (20)
0.93±0.22 (22)
1.22±0.30 (28)
S. Vittore 0.33±0.18 (4)
0.28±0.22 (3)
5.78±1.16 (59)
0.51±0.23 (5)
1.39±0.30 (14)
1.45±0.33 (15)
Zürich 0.54±0.22 (12)
0.41±0.21 (9)
0.51±0.11 (12)
0.62±0.28 (14)
1.01±0.23 (23)
1.35±0.33 (30)
Average 0.47±0.21 (9)
0.31±0.20 (6)
1.37±0.28 (26)
0.67±0.28 (13)
1.11±0.25 (21)
1.31±0.32 (25)
![Page 105: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/105.jpg)
Chapter 5. Long-term organic source apportionment in Central Europe using offline AMS
88
5.5 Conclusion
Sources contributing to OA are quantitatively separated and their uncertainty
estimated statistically at nine sites in central Europe throughout the entire year 2013
(819 samples). Thereby, three primary (HOA, COA, BBOA) OA sources are separated
from two secondary (WOOA, SOOA) categories and a yet unknown source explaining
sulfur-containing fragments (SC-OA). BBOA exhibits clearly higher concentrations at
the alpine valley sites in southern Switzerland than at the sites in northern Switzerland.
SOOA, characterized by high concentrations in summer, shows a more than linear
increase with rising temperatures as is observed from biogenic volatile organic
compound emissions and biogenic SOA concentrations. WOOA, the dominant SOA
category during winter, closely correlates with NH4+. The influence of PBOA, not
resolved by PMF, is estimated using, among others, cellulose analyses and could be an
important contributor. Cellulose’s temporal behaviour suggests maximal PBOA
contributions in northern Switzerland during summer, while at the southern alpine sites
maximal concentrations are reached in spring and autumn.
Acknowledgements
This work was supported by the Swiss Federal Office of Environment;
Liechtenstein; Ostluft; the Swiss cantons Basel, Graubünden, and Thurgau; the
Lithuanian-Swiss Cooperation Programme “Research and Development” project
AEROLIT (no. CH-3-ŠMM-01/08); and the IPR-SHOP SNSF starting grant. The
authors at IGE-Grenoble would like to thank the LABEX OSUG@2020 (ANR-10-
LABX-56) for funding analytical instruments.
![Page 106: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/106.jpg)
6 Insights into organic-aerosol sources via a novel laser-desorption/ionization mass
spectrometry technique applied to one year of PM10 samples from nine sites in
central Europe
K. R. Daellenbach1, I. El-Haddad1, L. Karvonen1, A. Vlachou1, J. C. Corbin1,
J. G. Slowik1, M. F. Heringa1, E. A. Bruns1, S. M. Luedin2,a, J.-L. Jaffrezo3, S.
Szidat4, A. Piazzalunga5,b, R. Gonzalez6, P. Fermo6, V. Pflueger2, G. Vogel2,
U. Baltensperger1, A. S. H. Prévôt 1Laboratory of Atmospheric Chemistry, Paul Scherrer Institute, 5232 Villigen, Switzerland 2MABRITEC AG, Riehen, Switzerland 3Université Grenoble Alpes, CNRS, IGE, 38000 Grenoble, France 4Department of Chemistry and Biochemistry & Oeschger Centre for Climate Change Research,
University of Bern,
3012 Bern, Switzerland 5Università degli Studi di Milano-Bicocca, 20126 Milan, Italy 6Università degli Studi di Milano, 20133 Milan, Italy anow at: University of Geneva, 1211 Geneva, Switzerland bnow at: Water and Soil Lab, 24060 Entratico, Italy
Accepted in Atmospheric Chemistry and Physics
![Page 107: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/107.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
90
Abstract We assess the benefits of offline laser-desorption/ionization mass spectrometry
(LDI-MS) in understanding ambient particulate matter (PM) sources. The technique was
optimized for measuring PM collected on quartz-fiber filters using silver nitrate as an
internal standard for m/z calibration. This is the first application of this technique to
samples collected at nine sites in central Europe throughout the entire year 2013 (819
samples). Different PM sources were identified by positive matrix factorization (PMF)
including also concomitant measurements (such as NOx, levoglucosan, and
temperature). By comparison to reference mass spectral signatures from laboratory
wood burning experiments as well as samples from a traffic tunnel, three biomass-
burning factors and two traffic factors were identified. The wood-burning factors could
be linked to the burning conditions; the factors related to inefficient burns had a larger
impact on air quality in southern Alpine valleys than in northern Switzerland. The
traffic factors were identified as primary tailpipe exhaust and most possibly
aged/secondary traffic emissions, respectively. The latter attribution was supported by
radiocarbon analyses of both the organic and elemental carbon. Besides these sources,
also factors related to secondary organic aerosol were separated. The contribution of the
wood burning emissions based on LDI-PMF correlates well with that based on AMS-
PMF analyses, while the comparison between the two techniques for other components
is more complex.
6.1 Introduction
Climate and health are strongly affected by atmospheric aerosols (Kelly et al.,
2007; IPCC, 2013), a substantial fraction of which is organic (Jimenez et al., 2009 and
reference therein). This organic aerosol (OA) is a complex mixture of thousands of
compounds (Goldstein and Galbally, 2007), of which only 10-30% have been speciated
by modern techniques (Hoffmann et al., 2011; Simoneit et al., 2004). Therefore, the
chemical composition of OA, its emission sources and formation processes are under
ongoing investigation. OA can be directly emitted as primary particulate matter
(primary OA, POA) or formed through the oxidation of gas-phase precursors with
subsequent condensation or nucleation (secondary OA, SOA). SOA dominates
![Page 108: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/108.jpg)
91
submicron OA at remote sites (90%) and is also a substantial contributor (30-80%) in
the urban environment (Zhang et al., 2011).
Mass spectrometry has significantly advanced the chemical characterization and
quantification of OA. Instruments equipped with electron ionization (EI), such as the
aerosol mass spectrometer (AMS, Canagaratna et al., 2007) and aerosol chemical
speciation monitor (ACSM, both Aerodyne Research, Inc., Ng et al., 2011c; Fröhlich et
al., 2013) provide quantitative online measurements of OA (Jimenez et al., 2016). The
application of positive matrix factorization (PMF) to AMS and ACSM mass spectra has
allowed the separation of different POA sources such as traffic, cooking, and wood
burning, as well as oxygenated OA factors representing SOA (e.g. Jimenez et al., 2009;
Lanz et al., 2007; Lanz et al., 2008; Lanz et al., 2010; Crippa et al., 2014, Canonaco et
al., 2013). While different SOA factors identified by AMS-PMF have been separated
according to their degrees of oxygenation and volatility, information on the different
origins of this fraction is limited due to significant fragmentation of the molecules by
EI. Moreover, SOA from different sources converges to a chemically similar
composition during oxidation (Kroll et al., 2011, Ng et al., 2011a).
Other strategies for aerosol mass spectrometry have provided complementary
information. Ionization by electrospray (ESI) avoids significant analyte fragmentation.
The coupling of ESI to ultra-high-resolution Fourier-transform mass spectrometers thus
provides detailed information on the chemical composition of a sample. However, using
ESI not all compound classes can be detected efficiently due to ion suppression (Furey
et al., 2013; Trufelli et al., 2011; Kourtchev et al., 2013). In addition, high costs and
labor intensity restrict the application typically to smaller sets of samples. Traditional
techniques such as gas- or liquid-chromatography (Hoffmann et al, 2011) coupled to
MS suffer from similar restrictions.
Laser-based mass spectrometers such as laser-desorption/ionization mass
spectrometers (LDI-MS) and two-step laser mass spectrometers (L2MS) are, similarly
as ESI, less affected by fragmentation than EI. Haefliger et al. (2000a and 2000b) show
that such instrumentation allows for identification of the dominant primary sources
using mass spectral fingerprints (tracer m/z’s identified using emission samples and
![Page 109: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/109.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
92
principle component analysis of ambient measurements). LDI-MS instruments may be
subject to matrix effects, i.e. ion formation from a given compound does not only
depend on its abundance but also on the abundance of all other compounds (De
Hoffmann and Stroobant, 1999). Additional matrix sensitivities may arise if ion-neutral
reactions take place in the desorbed plume, which means that MS design (e.g. time-of-
flight vs. ion trap) also plays a role (Murphy, 2007). Although the importance of matrix
effects is evident, studies with a single-particle LDI-MS (ATOFMS, laser wavelength
266 nm) have achieved good correlations of major aerosol components with well-
established reference measurements such as elemental carbon (EC, Sunset Analyzer,
Sunset Laboratory Inc.), OA (AMS), NH4 (AMS), SO4 (AMS), NO3 (AMS), and K+
(collected with a particle-into-liquid sampler, PILS), suggesting that underlying matrix
effects did not dominate measurement reproducibility (Healy et al., 2013).
Similar to the AMS, online single-particle LDI-MS instruments such as the
ATOFMS also yield extensive fragmentation. However, such fragmentation can be
avoided by measuring offline aerosol samples (filters) using other systems. Samburova
et al. (2005a) showed by comparing measurements with and without matrix addition,
that fragmentation was negligible in their instrument (wavelength 337 nm, LDI-MS,
Shimadzu/Kratos, Axima CFR). Overall, offline LDI-MS may therefore provide quick
access to additional chemical information at near-to-molecular level, potentially
allowing differentiation between several primary organic aerosol sources and even
different precursor-related SOA categories (Kalberer et al., 2004; Samburova et al.,
2005a). Based on offline LDI-MS measurements performed for an urban background
site in Zürich, Switzerland (same site as Haefliger et al., 2000a; 2000b), Baltensperger
et al. (2005) suggested that SOA from biogenic precursors is more important than SOA
from anthropogenic precursors. However, such LDI-MS analyses are rare, typically
focusing on pattern analysis in the mass spectrum since LDI-MS signal quantification is
difficult due to the variability in ionization efficiencies and chemistry for different
compounds. In addition, LDI-MS has not been applied to extensive datasets. In
comparison to online analyses, analyzing offline filter samples allows covering longer
time periods and larger observation networks.
![Page 110: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/110.jpg)
93
In this work, we evaluated the use of offline LDI-MS for the direct measurements
of PM collected on filters, without the addition of an ionization matrix. We assessed the
contributions of different aerosol sources to the total PM. To control for instrumental
differences, we measured all filters on the same instrument. We developed novel
procedures for instrument calibration and uncertainty assessment. We measured three
source reference samples (traffic, wood burning, and cooking) in addition to 819
ambient samples. The ambient samples included filters from the entire year of 2013 at
nine sites in central Europe with different emission conditions (Alpine valleys strongly
influenced by wood burning, as well as urban and rural regions). Based on positive
matrix factorization (PMF) using LDI-MS mass spectral data, the ability to resolve OA
sources in source apportionment was assessed and used for obtaining a deeper
understanding of sources contributing to the organic aerosol in central Europe.
6.2 Methods
6.2.1 Sample collection and other chemical analysis
Ambient samples were collected at nine sites in Switzerland and Liechtenstein
(described in more detail in Daellenbach et al., 2017) covering different atmospheric
conditions (urban/rural and background/curbside). Seven sites were located on a plateau
north of the Alpine crest (Basel, Bern, Payerne, Frauenfeld, St. Gallen, Zürich, Vaduz);
the remaining two sites were located in Alpine valleys south of the Alps (Magadino,
San Vittore). Samples were collected on quartz filters (Pall Corp.) by local air quality
monitoring networks every 4th day during the entire year 2013 (819 filters) and used for
gravimetric PM10 quantification. The samples were stored at -18°C and transported in
cooling boxes. Before any further handling steps, the samples were allowed to warm up
at room temperature for 60 minutes in order to avoid condensation of ambient humidity.
The organic (OC) and elemental (EC) carbon content was determined by a thermo-
optical transmission method using a Sunset OC/EC analyzer (Sunset Laboratory Inc.,
Birch and Cary, 1996), following the EUSAAR-2 thermal-optical transmission protocol
(Cavalli et al., 2010). For 33 samples from Magadino, radiocarbon (14C) analyses were
performed (2013; 10 samples, 2014; 23 samples, Vlachou et al., in prep). The
radiocarbon analyses were conducted at the University of Bern at the Laboratory for the
Analysis of Radiocarbon with the Accelerator MS (LARA; Szidat et al., 2014) to
![Page 111: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/111.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
94
determine the contribution of fossil and non-fossil OC (OCf and OCnf, respectively) and
EC (ECf and ECnf, respectively) content (Zhang et al., 2012). Major ion concentrations
were measured by ion chromatography (IC) by a Dionex ICS1000 instrument
(Piazzalunga et al., 2013; Jaffrezo et al., 1998). Levoglucosan was measured by high-
performance anion exchange chromatography (HPAEC) with pulsed amperometric
detection (PAD) using an ion chromatograph (Dionex ICS1000) following the method
of Piazzalunga et al. (2010; 2013a). Gas-phase NOx was measured online using a
chemiluminescence method, and meteorological parameters (e.g., temperature) were
monitored. Further, equivalent black carbon (eBC) was measured with a multi-
wavelength Aethalometer AE 31 (Magee Scientific Inc.) (Hansen et al., 1984; Herich et
al., 2011) in Magadino, Payerne and Zürich. eBC was separated into a wood burning
influenced (eBCwb) and traffic influenced (eBCtr) fraction based on the enhanced
absorption of eBCwb in the ultraviolet range. For this computation, we used the
Ångström exponents for wood-burning (αwb) of 1.7 and for traffic (αtr) 0.9 (Zotter et al.,
2017; Sandradewi et al., 2008). On all the samples, offline AMS analyses (Daellenbach
et al., 2016) were conducted. This involved the analysis of the water-soluble organic
matter (WSOA) by a high resolution time-of-flight AMS. The offline AMS data were
used for source apportionment (Daellenbach et al., 2017; Bozzetti et al., 2016) and used
in this study for comparison; the source mass concentrations were corrected for the
missing water-insoluble mass fraction using the method described in Daellenbach et al.
(2016).
Reference source samples for traffic (representing PM10) were collected in a
tunnel (Islisberg tunnel on Swiss highway A4, exit Wettswil am Albis). The PM10
samples were collected on Sunday, 2014-05-18, and Tuesday, 2014-05-20, from
midnight to noon with 2 hour intervals (rush hour in the morning). Additional source
samples were collected on quartz filters for beech-wood burning through laboratory
experiments of whole cycle burns and stable flaming phase only. These samples did not
only include the primary aerosol, but also two different levels of aging, i.e., after
simulated aging in a smog chamber for 1 hour (equivalent to an OH dose of 107 cm-3 h )
and 4 hours (OH dose 3·107 cm-3 h) (Bruns et al., 2015; 2016). Additionally, also filter
samples collected during cooking experiments were analyzed (Klein et al., 2016a,
2016b).
![Page 112: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/112.jpg)
95
6.2.2 Laser-desorption/ionization ToF MS analysis
6.2.2.1 Instrument and measurement settings
In laser desorption/ionization mass spectrometry, LDI-MS, liquid or solid material is
simultaneously desorbed and ionized by pulsed laser irradiation. The laser is focused on
the surface of the sample. Ions are produced through two major pathways
(Knochenmuss et al., 2000; Zenobi et al., 1998). In the first pathway, ions are formed by
interaction with the laser beam (primary ions). In the second pathway, in the expanding
primary plume of desorbed molecules and primary ions, ion-molecule reactions produce
secondary ions (Knochenmuss et al., 2000). The detailed ionization mechanisms are still
under investigation (Knochenmuss et al., 2002; 2003; 2006; 2016). Although this type
of ionization is considered softer than electron ionization, the observed ions are
typically still fragments of their parent molecules (De Hoffmann and Stroobant, 1999).
Furthermore, the measured intensity of one compound does not only depend on its
concentration but also on the concentration of all other compounds present (the so-
called matrix effect) making quantification challenging (Ellis et al., 2014; Borisov et al.,
2013).
We recorded the mass spectra of 819 filter samples in an m/z range 65-500 thomsons
(Th; 1 Th =1 Da e-1, where e is the elementary charge, ion gate at m/z 60 Th) using a
laser-desorption/ionization-ToF MS (Shimadzu Axima Confidence, Shimadzu-Biotech
Corp., Kyoto, Japan) equipped with an N2 laser (wavelength 337 nm, frequency 50 Hz,
laser pulse width 3 ns, 130-180 µJ/pulse) in the positive reflectron mode. All the
accessible instrumental parameters were kept constant during the whole period of
measurements taking place from November 2015 to mid-March 2016. Specifically, the
laser intensity was adjustable by means of a rotating wheel of filters with varying
transmissivity (0 being blocked and 180 being completely open). We set the wheel
parameter to 105 of 180 which would result in an estimated laser energy of ~6-9
µJ/pulse or 2.8-4.2*108 W/cm2 with a 3ns pulse and 30µm laser beam diameter. While
this wheel the laser energy was initially set (wheel parameter 105 of 180 equaling
roughly 6-9 µJ/pulse or 2.8-4.2*108 W/cm2 with a 3ns pulse and 30µm laser beam
diameter, 0 being blocked 180 being completely open) and kept constant, the aging of
the laser during the given time period was also expected to reduce its intensity. We
![Page 113: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/113.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
96
monitored and assessed changes in laser power and other instrumental parameters, as
well as possible sources of uncertainty/contamination from sample preparation and intra
and inter-day reproducibility, by repeated measurements of a subset of our samples.
Quartz filter punches of 8 mm were attached to a custom-made stainless steel sample
holder (32 slots). Each of the samples was additionally spiked with a droplet of dilute
aqueous silver nitrate solution (AgNO3, Sigma Aldrich, >99.8%, 500 ppt to 20 ppm), to
provide Ag cations as an internal standard for the m/z calibration. After drying under
ambient conditions, filter punches were inserted into the sampling chamber and
analyzed by the LDI-MS. Blanks were measured according to the same procedure. Both
intra-day and delayed repeated measurements were conducted for the same filters to
assess instrument performance and uncertainty/contamination from sample preparation.
The intra-day repeatability was assessed by measuring 3 filters 10 times on 3 different
occasions as outlined in Section 6.3.1. Overall, 96% of the available ambient filter
samples (785 filters) provided usable data (defined below).
6.2.2.2 Data treatment
While most atmospheric LDI-MS studies present raw mass spectra (Samburova et
al., 2005a; 2005b; Kalberer et al., 2004; Baltensperger et al., 2005), in the present study,
we introduced data treatment techniques in order to perform further mass spectral
analysis on stick-integrated spectra. The techniques we employed are described in the
following.
The m/z calibration parameters were highly variable between different samples.
Therefore, to determine unit mass resolution integration regions it was necessary to
perform an m/z calibration on every single sample. However, unlike in the Aerodyne
AMS (N2+, O2
+, W+), there are no dominant anchor ions present in these spectra that
could be used for an m/z calibration. In absence of such ions, we performed a two-step
calibration procedure as described in the following. Each sample was spiked with silver
nitrate (AgNO3, Sigma Aldrich, >99.8%) as an internal standard (approach illustrated in
Fig. 6.1). In order to avoid the suppression of the sample signal, the internal standard
(aqueous solution 500 ppt to 20 ppm) was only placed (as a droplet) on a small part of
the sample. The 499 spectra from all positions on the measurement grid were separated
![Page 114: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/114.jpg)
97
and defined as (1) silver-spiked (lower panel in Fig. 6.1a), (2) silver-free (upper panel in
Fig. 6.1a), and (3) intermediate-silver. Intermediate-silver cases were defined using the
signal intensity in the regions of the mass spectrum where silver was expected, in
comparison to adjacent silver-free regions of the spectrum, and were discarded. To
calculate the first calibration, we calibrated the average silver-spiked spectrum of each
filter sample, using the peaks of the silver monomer (m/z 107, 109 Th), dimer (m/z 214,
216, 218 Th), and trimer (m/z 321, 323, 325, 327 Th). We found that this calibration of
silver-spiked was not directly applicable to the average silver-free mass spectrum.
Possibly, spiking the filter region with aqueous AgNO3 caused enough of a change to
the surface of the filter sample to influence the ionization physics. This could affect the
m/z calibration, since the delayed-pulse ion extraction in our instrument is not
orthogonal to the ablation plume but nearly parallel. Therefore, a second calibration was
obtained for the averaged silver-free spectra using prominent non-silver peaks present in
both spectra. Such a two-step calibration is necessary to achieve accurate m/z
calibrations of the silver-free spectra.
After calibration, mass spectra were baselined using the following custom algorithm.
A window of width 1 Th (the detector bins were approximately 0.04 Th in width) was
applied to identify the lowest signal intensity of that range, which was defined as the
baseline. The window was moved across the mass spectrum with steps of 0.25 Th to
obtain a baseline spectrum. After linear interpolation, the baseline was subtracted from
the spectrum. Subsequently the spectra were integrated to unit mass resolution sticks
(UMR, 1 stick per unit mass). The UMR integration window was not centered at integer
masses, but rather at integer mass plus the mass defect of an alkyl ion (R-CH2+). The
width of this window was defined as extending to the minimum signal intensity above
and below the center using the first derivative of the signal.
To enable a direct comparison of LDI-MS and HR-AMS results, the intensity of the
calibrated and baseline-subtracted spectra (I’m/z) with arbitrary units were scaled to OA
(µg/m3) using the respective OA contents of the filter samples (see Eq. 6.1, with
rescaled intensities termed Im/z thereafter). OCSunset was determined using the Sunset
OC/EC analyzer and (WSOA/WSOC)oAMS through WSOA measurements (described in
Section 6.2.1, WSOC being water-soluble organic carbon). Note that an additional
![Page 115: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/115.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
98
contribution of non-organic material to the signal could not be fully excluded. However,
we consider the detected signal mainly to be related to organics based on the
measurements of ambient aerosol with offline LDI-MS systems using different laser-
wavelengths (193 nm and 355 nm, Aubriet et al., 2010) and pure components (337 nm,
Goheen et al., 1997) indicating that sulfate and nitrate respond only in the negative
mode. Additionally, molecular nitrate and ammonium ions are too small to be detected
in this study (cutoff at m/z 65 Th). We also do not expect clusters between inorganic
cations and organics because we did not observe significant amounts of silver/organic
clusters in silver-spiked spectra (Fig. 6.1a).
(6.1)
6.2.3 Source apportionment / PMF
6.2.3.1 General and input data
Positive matrix factorization (PMF, Paatero and Tapper, 1994; Paatero, 1997) is a
widely used algorithm for source apportionment. PMF (Eq. 6.2) explains the variability
in a dataset (xi,j, here LDI-MS mass spectra scaled to OA, Im/z) as a linear combination
of constant factor profiles (fk,j, here mass spectral signatures) and their time-dependent
contributions (gi,k, here concentrations of this factor). The index i represents a specific
point in time, j the signal at a specific m/z, and k a factor (up to the number of factors p).
(6.2)
PMF is solved by minimizing the residuals ei,j weighed by the measurement
uncertainty, here σPMF,i,j.. In this study, PMF was solved by the multilinear-engine 2
(Paatero, 1999 and references therein) using the front-end user interface SoFi 4.9
(Canonaco et al., 2013) developed for Igor Pro v6 (Wavemetrics). We performed PMF
without constraints of e.g. reference mass spectra of aerosol sources (unconstrained
PMF). The data matrix consisted of the stick-integrated LDI-MS spectra (m/z 65 to 485
Th scaled to OA) of all 785 filter samples. We excluded all m/z’s for which we expected
![Page 116: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/116.jpg)
99
silver signals (listed in section 6.2.2.2) even though only silver-free spectra (according
to the definition given above) were considered.
We determined the PMF uncertainty matrix (σPMF,i,j) based on the instrument
repeatability (described in Section 6.3.1) according to the two-component model for
measurement uncertainty described by Rocke and Lorenzato (1995) and Wilson et al.
(2004). This approach ultimately means that an absolute-error term (e.g., due to noise)
is combined in quadrature with a relative-error term (e.g., due to scaling or calibration
uncertainties) as shown Fig. 6.2 and in Eq. 6.3 (σ’PMF,m/z, for a single mass spectrum).
The need for and applicability of such an error model has been demonstrated by Corbin
et al. (2015) for mass-spectrometry-PMF. We obtained the two error terms by fitting
Eq. 6.3 to the relative standard deviation of replicate measurements for each filter and
experiment, and applied the averaged fitted parameters to the entire data set (constant
error term, , and an error term proportional to the measured peak intensity,
).
(6.3)
The error was scaled to OA in the same way as the LDI-MS mass spectra (Eq. 6.4).
(6.4)
6.2.3.2 Uncertainty estimate of PMF results
A well-established statistical tool for estimating the uncertainty is the bootstrap
technique, which consists of randomly resampling the input data, with replacement, to
create input matrices with the same dimensions as the initial data matrix (Davison and
Hinkley, 1997; Brown et al., 2015). We performed 1’000 bootstrap PMF runs (also
using the replicate measurements) with different initial guesses (“seeds”) using input
matrices of the same size (785 filter samples and 412 ions), resulting in the uncertainty
estimate σbs. We also assessed and parameterized the uncertainty arising from the intra-
day repeatability for three filters analyzed repeatedly by LDI-MS on 3 different days, 10
![Page 117: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/117.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
100
times each day (σintraday, see Supplement B). The reported uncertainty for each PMF
factor is the quadratic sum of σbs and σintraday, as explained in more detail in the
Supplement B. Potential long-term drifts in instrumental response were evaluated
(σinterday, see Supplement B, Fig. SI.B.10) but were not accounted for numerically.
6.3 Results
6.3.1 Calibration, repeatability, and quantification
An example of a sample spectrum after m/z calibration and baseline subtraction,
separated into silver-spiked and silver-free spectra and averaged over the entire filter
punch, is presented in Fig. 6.1a. The silver-spiked spectrum features an intensive signal
at m/z 214 Th related to the 107Ag2 dimer, in contrast to the silver-free spectrum.
The operational mass resolution, m/Δm, based on the measurements conducted on
one sample tray per month and the accuracy of both calibration steps are shown in Fig.
6.1b and 6.1c, respectively. The operational resolution of the instrument was only
determined for the silver mono- (resolution 1100), di- (1700), and trimer (2100) as we
are confident of the absence of interfering ions for these peaks. In comparison to the
LDI-MS, the HR-ToF AMS in V-mode (and W-mode) has a resolution of ca. 2’000
(4’000) at m/z 100 Th and an m/z calibration accuracy < 20 ppm (< 10 ppm, DeCarlo et
al., 2006). With this accuracy and resolution, neither distinguishing different ions at the
same nominal mass nor estimating properties such as O/C and H/C for a nominal mass
following Stark et al. (2015) was possible for the quartz-filter-LDI-MS measurements.
Instead, the spectra were integrated to UMR sticks. The presence of considerable signal
at odd masses might indicate that significant fragmentation occurs during desorption
and ionization of the organic aerosol in the LDI-MS (Fig. 6.1a, 6.3, 6.4). However,
Samburova et al. (2005a) suggested that fragmentation in this instrument is negligible.
The mass spectral signatures acquired in this study are similar to the ones in Samburova
et al. (2005a) and the laser energy applied in this study is lower than in the formed
study. This suggests that also during our measurements fragmentation was not a
prominent process. Overall, the extent of fragmentation remains unclear.
![Page 118: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/118.jpg)
101
The repeatability of the mass spectral signature was assessed in detail for 3 selected
filters (collected in Basel on 2013-06-21, 2013-09-21, and 2013-12-22). Ten punches
from each of these 3 samples were prepared on the sample holder in the same time and
subsequently measured. The given procedure was repeated on 3 occasions (on 2015-11-
25, 2015-12-17, 2016-01-28). The relative error was related to the absolute signal
intensity for each m/z for each sample and experiment (Fig. 6.2), with an asymptotic
value of only ~9 ± 7 % at high signals (average and standard deviation of fitting Eq.
6.3). The average absolute error for small signals was 102 ± 48 a.u. (from Eq. 6.3).
These parameters did not show a temporal trend which suggests that the measurements
are repeatable, despite filter inhomogeneity and laser instability. Analyses of field
blanks (also spiked with AgNO3) exhibited low signal with 99% of peaks below
detection limit (defined as ). The other 1% of the peaks (m/z 197, 249, 251,
322,324 Th) exhibited high variability among the field blank analyses (see details in
Supplement B, Fig. SI.B.1 and SI.B.2). Therefore, no blank subtraction was performed.
The total measured intensity did not show a relation with the filter
loading (OC, OC+EC, PM10, see Supplement B, Fig. SI.B.3), indicating that factors
such as the composition, size, and/or mixing state of the collected aerosol particles had
a larger influence on the measured ion intensities than the mass of PM. Moreover,
replicate measurements indicated that the Itot of a single sample measured on different
instances was affected by an instrumental drift (details in Supplement B, Fig. SI.B.4).
![Page 119: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/119.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
102
Figure 6.1: m/z calibration of LDI-MS analysis of aerosol collected on quartz-fiber filter: a) example of calibrated silver (Ag-spec, blue) and no-silver-containing average mass spectra (noAg-spec, green) of a filter sample. The insert in a) displays a zoom-in of the Ag-spec and noAg-spec. b) operational resolution determined based on silver-mono-, di-, and trimer. c) m/z calibration accuracy for both steps of the m/z calibration.
![Page 120: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/120.jpg)
103
Figure 6.2: Error model parametrization based on 3 samples (Basel, 2013-06-21, 2013-09-21, 2013-12-22) measured on 3 instances with each time 10 repeats on 1 sample holder.
6.3.2 Combustion source profiles
Filters from cooking processes (frying, related largely to oil pyrolysis, Klein et al.
2016a; Allan et al., 2010) did not yield measurable mass spectra in our LDI-MS. This
observation is in agreement with a prior study which performed in-situ single-particle
LDI-MS measurements at 266 nm (using the ATOFMS) and did not observe cooking
particles (Healy et al., 2013). The lack of either graphene-like black carbon or
polycyclic aromatic hydrocarbons (PAHs) to absorb the LDI-MS laser may explain
these observations (Healy et al., 2013), which suggests that our measurements are more
sensitive towards certain PM components. We speculate that it may be possible in some
cases to observe cooking-related ions in LDI-MS of atmospheric samples, if such ions
are secondary ions formed in the ablation plume or cooking particles are coated by
absorbing SOA.
The mass spectra of tunnel filter samples were complex and characterized by distinct
patterns (Fig. 6.3). Some m/z fragments (e.g. 84, 94, 101, 177 Th) were present in the
weekday and weekend tunnel samples, but not in the biomass smoke. Between the two
tunnel samples, large differences could also be identified. The tunnel sample collected
on a Tuesday during rush hour (06:00-08:00) showed many more peaks (among others
163 and 177) which were less prominent during the weekend at the same time and in the
same location (where m/z 101 and 143 Th dominated the weekend spectrum).
![Page 121: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/121.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
104
Figure 6.3: samples representing different combustion sources a) and b) traffic emissions from samples collected in Islisberg tunnel (exit, Wettswil, Switzerland), and c) primary wood burning using whole cycle emissions d) and stable flaming phase emissions d). In absence of a measurable mass spectrum, no spectrum for cooking emissions is displayed.
Previous studies have identified lubrication oil as a major component of diesel- and
gasoline-vehicle exhaust (Gentner et al., 2017; Chirico, et al., 2011), which consists
largely of aliphatic hydrocarbons. Cooking particles similarly consist of such
hydrocarbons (Schauer et al., 2002; Allan et al., 2010; Crippa et al., 2013c). Since we
did not observe a mass spectrum for the cooking sample, but do observe one for the
tunnel sample, our detection of tunnel particles may rely either on the presence of black
carbon or PAHs on the filter samples. To the extent that these species are
heterogeneously distributed within the sampled aerosol particles, our technique may be
specifically sensitive to certain components of traffic emissions.
The mass spectrum of the wood burning sample from the whole burning cycle
shows a complex pattern with a bimodal envelope of relatively-intense ions from m/z
200 Th onwards. By contrast, the sample from the stable-flaming phase shows fewer
peaks (e.g. m/z 85, 140, 213 Th, etc.) with a less significant envelope of ions.
6.3.3 Ambient samples
Typical mass spectra from 4 different locations, known to be heavily influenced by
specific sources at different times of the year are displayed in Figure 6.4. The fragments
are color-coded by their correlation coefficients with 3 different markers or proxies for
aerosol types, i.e., NOx for traffic exhaust (Fig. 6.4a), levoglucosan for biomass burning
emissions (Fig. 6.4c and 4d) and temperature for SOA production (as temperature is
![Page 122: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/122.jpg)
105
expected to exponentially enhance biogenic emissions, Leaitch et al., 2011, Fig. 6.4b).
In general, the mass spectra from winter-time were characterized by higher
contributions from higher-molecular-weight fragments compared to summer. For the
winter sample from San Vittore, a large “hump” was present, on top of which signals
appeared with high intensities and a highly regular pattern of m/z differences of 14 Th.
These fragments, albeit not all detected in smog chamber biomass burning aerosols,
were highly correlated with levoglucosan, a unique marker of biomass smoke.
Figure 6.4: LDI-MS mass spectra from summer and winter samples (dates given in the legend) from a traffic-influenced (Bern), a rural (Payerne), a wood burning-influenced (San Vittore) and an urban background (Zürich) site. Spectra are color-coded with the correlation coefficient (r) between the m/zs and a specific environmental parameter for the whole dataset: for a) the correlation with NOx during summer, for b) the correlation with temperature, for c) and d) the correlation with levoglucosan.
The m/z 84 94, 101, 120, 143, 165, and 177 Th, also detected in tunnel samples,
showed high relative intensities in Bern in summer, of which 94, 120, and 177 Th
correlated with NOx, which suggests a high contribution of traffic emissions at this
location, consistent with previous observations (Zotter et al., 2014). These ions were
also observed, but to a lesser extent, at the rural site of Payerne. At the latter site, the
highest intensities were found for m/z 74 and 104 Th. Only these 2 ions showed a clear
relation with temperature and a clear increase during summer. Overall, this spatial and
temporal variability observed in the spectral fingerprints and the correlation of specific
fragments with environmental parameters suggest that LDI-MS data contain mass
![Page 123: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/123.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
106
spectral information that may be used to separate the contribution from different aerosol
sources. Source apportionment analysis using these data in combination with PMF was
therefore explored in the following.
6.3.4 Source apportionment results
6.3.4.1 PMF setup
Residual analysis of preliminary PMF runs showed that structure in the residuals
was removed when increasing the number of factors up to 5 but not when further
increased. However, when allowing for 7 factors, further environmentally interpretable
separations could be achieved. Increasing to 8 factors led to the separation of a third
traffic related factor, which however contributed little to the overall mass. Therefore, we
opted for 7 factors (see details in Supplement B, Fig. SI.B.6, SI.B.7, and SI.B.8).
6.3.4.2 Interpretation of PMF factors
6.3.4.2.1 Traffic related factors
Two factors could be related to traffic (Fig. 6.5a and 6.5b, termed traffic1 and
traffic2, respectively) based on patterns in the factor profiles similar to patterns in the
samples measured in the tunnel on a weekday (Fig. 6.3a). Since the signatures obtained
from the tunnel filters represent both tailpipe exhaust and resuspended dust, the two
factors identified could be a mixture of both sources. In order to elucidate the reasons
behind the separation of traffic related aerosols into two factors by PMF, we inspected
their relationship with NOx and eBCtr, typical markers of traffic emissions. Both
traffic1 and traffic2 showed increasing concentrations with increasing NOx levels.
However traffic1/NOx and traffic2/NOx were seasonally variable (Rp,traffic1,NOx=-0.17,
Rp,traffic2,NOx=0.24, Fig. 6.6, Fig. SI.B.9). While traffic2 correlated with eBCtr , traffic1
did not show such a dependency (Tab. SI.B.1, Fig. 6.6, Fig. SI.B.9). The correlation
with eBCtr suggests that traffic2 is related to primary emissions from the combustion
process. Based on the lack of correlation between traffic1 and eBCtr, traffic1 might also
be influenced by other processes, such as e.g. secondary production. Additionally,
![Page 124: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/124.jpg)
107
tailpipe exhaust cannot be distinguished from other traffic related emissions in the
tunnel samples and might also contribute to traffic1.
Figure 6.5: PMF factor profiles (colored sticks) and their uncertainty (grey shaded areas, variability among PMF runs): a) traffic1, b) traffic2, c) efficient wood burning (BBeff), d) inefficient wood burning (BBineff1), e) inefficient wood burning 2 (BBineff2), f) lower molecular weight OA (LMW-OA), and g) biogenic OA (bio-OA).
The ratio of both factors to NOx exhibited a clear seasonality, increasing during
summer (Fig. 6.6a and 6.6b). Such a change can either be caused by a change in the
emission patterns, by an enhanced photochemical production of these factors or due to a
change in the lifetime of e.g. NOx. In Fig. 6.6c and d, we compared the contribution of
the traffic related factors with those of eBCtr, another tracer of traffic emissions whose
lifetime is similar to OA (see Fig. SI.B.9 for summer points only). The comparison
shows that while the ratio of traffic2 to eBCtr is not season-dependent, the one of
traffic1 to eBCtr increases during summer. The seasonal variability in the relative
contribution of traffic1 might thus be related to a seasonal change in fleet composition
![Page 125: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/125.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
108
or combustion conditions, or an enhancement of the photochemical production of this
fraction. An enhanced dust resuspension in the warm season (e.g., more dust on the road
due to less precipitation) could also contribute to an increased traffic1 concentration in
summer.
Figure 6.6: Scatterplots between factor-time series and respective markers (traffic1, traffic2, BBeff, BBineff1, BBineff2, bio-OA, LMW-OA, eBCtr, levoglucosan, potassium, ammonium are displayed in µg/m3, NOx in ppm, and temperature in °C).
![Page 126: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/126.jpg)
109
6.3.4.2.2 Wood burning related factors
Three factors could be related to wood burning emissions (Fig. 6.5c, 6.5d, 6.5e). The
first (Fig. 6.5c) showed a similar mass spectral pattern as samples from laboratory
experiments from stable flaming wood burning exhaust from a log wood burner and
thus was termed efficient wood burning (BBeff). The fingerprint of the second (Fig.
6.5d) resembled that of laboratory wood smoke aerosols from the entire burning cycle
including the inefficient starting and burnout phase and, therefore, was related to
inefficient wood burning (BBineff1). BBineff1 showed also a similar signature as
spectra from wood burning haze episodes in San Vittore identified by high levoglucosan
concentrations. The third (Fig. 6.5e) mostly explained the masses above m/z 300 Th.
The signature of the whole cycle wood burning aerosols from a log wood burner (Fig.
6.3c) also showed high relative contributions at high m/z, similar to this factor. For this
reason this factor was termed BBineff2. Both BBeff and BBineff1 correlated with
levoglucosan (Fig. 6.6e, 6.6f, Tab. SI.B.1). Similar to BBeff and BBineff1, BBineff2
correlated with levoglucosan and thus could be related to wood burning emissions (Fig.
6.6g, Tab. SI.B.1). However, BBineff2 also correlated with NH4+, a marker of aged
aerosols, at the northern sites (Fig. 6.6m), which may also indicate that this fraction can
include aged and/or secondary components (Tab. SI.B.1).
Comparing the wood burning related factor time series to potassium (K+), an
inorganic wood burning marker mostly present in ash, provides further insight into the
separation of the 3 wood burning related factors. Among the factors related to wood
burning emissions, BBeff has a lower BB/K+ ratio (2.4, IQR 1.25-4.2) than BBineff1
(6.0, IQR 2.7-12.0) and BBineff2 (11.6, IQR 6.0-21.0). In wood burning experiments, it
was found that OC, EC, PM10, and PAH emissions increase relative to the potassium
output during non-ideal burning conditions (Lamberg et al., 2011). Zotter et al. (2014)
found a north-south gradient in levoglucosan/K+ and OCnf/K+ in Switzerland and
hypothesized that it might be linked to the burning conditions. Thus the higher the LDI-
BB/K+ ratios (north: 16.6, IQR 10.4-30.8 south: 30.8, IQR 22.5-44.2), the less efficient
the burning conditions which in turn supports the hypothesis that BBeff represents the
most efficient burning conditions among the three factors. All wood burning related
factors showed higher median BB/K+ ratios at the southern Alpine valley site (BBeff:
3.0 with IQR 1.1-5.2, BBineff1: 10.9 with IQR 7.5-17.3, BBineff2: 15.8 with IQR 9.9-
![Page 127: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/127.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
110
24.0) than in northern Switzerland (BBeff: 2.1 with IQR 1.2-3.4, BBineff1: 4.3 with
IQR 2.2-8.6, BBineff2: 9.6 with IQR 5.2-18.3) as also visible in Fig. 6.6h/i/j. The north-
south gradient in the BB/K+ ratios might be caused by imperfections of the separation of
the burning conditions by PMF, but other effects such as the age of the stove population
and the used technology could contribute as well. The wood consumption (BFE, 2013)
of automatic burners (>50kW), was higher in northern Switzerland (in the respective
regions, 48m3 wood/km2 area) than in southern Switzerland (8m3 wood/km2 area). Since
the OM, POA, and SOA emissions of pellet burners during the stable phase are
drastically reduced compared to modern logwood burners in a stable flaming phase (e.g.
Heringa et al., 2011), they might contribute over-proportionally to potassium in
northern Switzerland but only little to OA, leading to the higher BB/K+ ratios at the
southern Alpine sites.
6.3.4.2.3 Biogenic-OA, low-molecular-weight OA, and other factor
A factor characterized by high contributions of low-molecular weight ions (Fig. 6.5f,
LMW-OA) correlates with NH4+ suggesting secondary processes as origin (Fig. 6.6l,
Tab. SI.B.1). At the southern Alpine valley sites, the LMW-OA / NH4+ ratio was higher
than in northern Switzerland. As NH4+ is mostly associated with the secondary
inorganic species sulfate and nitrate this variability will mostly be related to differences
in the VOC versus SO2 and NOx emissions, along with temperature differences
influencing the partitioning of nitrate to the particle phase.
The contribution of the last remaining factor showed an exponential increase with
temperature, similar to terpene emissions and biogenic SOA (Leaitch et al., 2011),
suggesting this factor to be strongly influenced by biogenic SOA production. Therefore,
this factor was termed bio-OA. The relationship between bio-OA and temperature was
similar both at the northern and southern sites (Fig. 6.6k, Tab. SI.B.1). Bio-OA had a
highest relative contribution at the most rural site in the dataset (Payerne: yearly average
bio-OA 11%, yearly average NOx concentration: 9 ppb) and the lowest at the most
trafficated site (Zürich: yearly average bio-OA 3%, yearly average NOx concentration:
48 ppb). The chemical signature of bio-OA was dominated by fragments at m/z 74 and
104 Th (Fig. 6.5g). The nature of these fragments remained unidentified, but could not
![Page 128: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/128.jpg)
111
be related to monoterpene or sesquiterpene SOA because of their very low m/z. Further,
the composition of secondary aerosols is expected to be much more complex, showing a
series of fragments distributed over a wide m/z range. We note that the abundance of the
fragments at m/z 74 and 104 Th relative to the total signal should not be directly related
to their absolute concentrations and their ionization efficiency might be superior to that
of other molecules, which may increase their apparent contribution.
6.3.4.3 Comparison to offline AMS and assessment of LDI-MS response factors
The LDI-MS source apportionment results were compared to those based on offline
AMS (oAMS) analyses performed on the same samples (Daellenbach et al., 2017). We
note that this comparison is not straightforward as different sources are separated by the
two methods. For this purpose traffic1 and traffic2 were summed up to LDI-traffic, and
BBeff, BBineff1, and BBineff2 to LDI-BB (Fig. 6.7).
Figure 6.7: Comparison of LDI-MS to reference offline AMS source apportionment results for the sum of traffic related factors (LDI-traffic, R2
LDI,oAMS=0.04), sum of wood burning related factors (LDI-BB, R2
LDI,oAMS =0.83), bio-OA (R2LDI,oAMS =0.62), and LMW-OA (R2
LDI,oAMS =0.45).
As was the case for the comparison to NOx and eBCtr (Section 6.3.4.2), also in the
comparison to the offline AMS traffic (HOA), a higher LDI-traffic/HOA was observed
![Page 129: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/129.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
112
in summer which contributes to the low correlation coefficient (R2=0.04). This might be
related to the fact that HOA is only primary without major contributions of traffic SOA
or dust resuspension while LDI-traffic is potentially also influenced by aged/secondary
traffic aerosol and resuspension. In contrast to summer, we observe a good agreement
between LDI-traffic and HOA in winter. Thus even though a varying relative response
factor (rRF) of the LDI-MS might contribute to these differences, these biases are not
systematic but season dependent. As stated earlier, LDI-traffic is thought to be a
mixture of primary tailpipe exhaust, aged/secondary tailpipe exhaust, and resuspended
dust (as well as tyre break and engine wear). This will be further elucidated in Section
6.3.4.5.
LDI-BB was highly correlated with offline AMS-BBOA (R2=0.83), yet the LDI-BB
concentrations were higher than BBOA from oAMS, especially in northern Switzerland
(with an LDI-BB:AMS-BBOA ratio between 1.4 and 4.1 for the different sites). A
possible reason is also here the mixing of secondary components into LDI-BB when
comparing to primary BBOA from oAMS. However, we cannot exclude that this effect
is due to different rRFs of the LDI-MS for different compound classes.
The identified secondary components, LMW-OA (R2=0.45) and bio-OA (R2=0.62),
also correlate with the corresponding OOA factors from the oAMS analysis, WOOA
and SOOA, yet the correlation coefficients are smaller than for LDI-BB and BBOA. For
these factors, differences between the two methodologies could be related to differences
in the PMF performance or to differences in the response factors for different
components in the LDI-MS (the LMW-OA: WOOA ratio is between 0.7 and 1.1 and the
bio-OA: SOOA ratio between 0.2 and 0.4).
At some of the sites analyzed with LDI-MS in this study, OA was monitored in
previous years with state-of-the-art online aerosol mass spectrometry (AMS or ACSM).
Earlier campaigns with quantitative online AMS analyses show a higher SOA
contribution at those sites: e.g. in Zürich, July 2005 (66% vs 25% for LDI-MS summer)
and December 2005 (55% vs 14% for LDI-MS winter), in Roveredo (close to San
Vittore), in March 2005 (53% vs 46% for LDI-MS summer), and in December 2005
(43% vs 11% for LDI-MS winter), in Payerne, in July 2005 (94% vs 47% for LDI-MS
![Page 130: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/130.jpg)
113
summer) and in December 2005 (71% vs 20% for LDI-MS winter) (Lanz et al., 2010).
Canonaco et al. (2013) presented source apportionment results for the winter of 2011
and 2012 for Zürich (PM1 ACSM), also with higher SOA contributions (71%) than the
LDI-MS in 2013 (20%).
Overall, the source apportionment results based on LDI-MS data provide source
separations with similar temporal behaviors as offline AMS and online AMS and
ACSM analysis, yet seem to overestimate combustion related primary particle sources
and underestimate secondary OA. However, we assume for this comparison that all
factors give an equal response at a given concentration, i.e. the relative response factor
(rRFLDI) of all factors to be 1. The relative contribution of a factor k to the total signal
observed with the LDI-MS for a specific measurement (i) depends on the
factor concentration ( as well as the sum of all factors separated
for the LDI-MS data. Assuming that differences between the LDIr-PMF and the AMS-
PMF arise solely from different rRFLDIs for different factors, is a function of the
AMS factor concentrations and the rRFLDI of this factor (Eq. 6.5):
(6.5)
In order to determine rRFLDI for the LDI-MS several strong assumptions are
required: (1) the sum of traffic1 and traffic2 represents HOA, (2) the sum of WBeff,
WBineff1 and WBineff2 represents BBOA, (3) LMW-OA represents WOOA, (4) bio-
OA represents SOOA and (5) the AMS factors for which there is an LDI-MS
equivalent, (1) - (4), are the only contributors to OA. In scenario 1, all the above
assumption are considered to be true. For scenario 1, we estimate rRFLDI-BB as 1.84,
rRFLDI-LMW-OA as 0.29 and rRFLDI-bio-OA as 0.40 (using LDI-traffic as reference factor,
rRFLDI-traffic = 1.00). The LDI-MS factor concentrations corrected using rRFLDI show a
close relation to the uncorrected LDI factor concentrations (R2LDI-traffic = 0.78,
R2LDI-BB=0.94, R2
LDI-LMW-OA=0.80, R2LDI-bio-OA=0.85, Fig. 6.8).
![Page 131: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/131.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
114
Table 6.1: relative response factors (rRF) for LDI-MS analyses for 3 different scenarios. scenario 1 scenario 2 scenario 3
rRFLDI-traffic = 1.00±0.00 rRFLDI-traffic2 = 1.00±0.00 rRFLDI-traffic = 1.00±0.00
rRFLDI-BB = 1.84±0.06 rRFLDI-BBeff+LDI-BBineff1 = 1.26±0.06 rRFLDI-BB = 5.67±0.13
rRFLDI-LMW-OA = 0.29±0.02 rRFLDI-LMW-OA+LDI-BBineff2 = 1.67 ±0.07 rRFLDI-LMW-OA = 0.30±0.02
rRFLDI-bio-OA = 0.14±0.01 rRFLDI-bio-OA+LDI-traffic1 = 0.39±0.02 rRFLDI-bio-OA = 0.65±0.02
We tested the sensitivity of the rRFLDI estimates (Tab. 6.1) to the assumptions (1)
and (2) being wrong (scenario 2) and only assumption (5) being wrong (scenario 3). In
scenario 2, we alter scenario 1 by comparing the sum of traffic1 and bio-OA to SOOA
and the sum of BBineff2 and LMW-OA to WOOA when computing rRFLDI. For
scenario 2, LDI-MS factor concentrations corrected using rRFLDI show also a close
relation to uncorrected LDI-MS factor concentrations (R2LDI-traffic2=0.96,
R2BBeff+BBineff1=1.00, R2
LDI-BBineff2+LMW-OA=0.99, R2traffic1+bio-OA=0.96). In scenario 3, we
alter scenario 1 by considering also AMS factors without an equivalent in the LDI-MS
PMF. Therefore, we compare LDI-traffic to the sum of HOA, COA, and SC-OA. Under
these conditions LDI-MS factor concentrations corrected using rRFLDI show a close
relation to uncorrected LDI-MS factor concentrations (R2LDI-traffic=0.75, R2
LDI-BB=0.90,
R2LMW-OA=0.74, R2
bio-OA=0.85). The differences in rRFLDI between scenarios 1, 2, and 3
highlight the uncertainties caused by plausible violations of the underlying assumption
when determining rRFLDI. Overall the estimates rRFLDI are highly uncertain and their
accurate determination needs to be the focus of future work. Given the high uncertainty
of rRFLDI and the good correlation between rRFLDI corrected and uncorrected LDI-MS
factor concentrations, we present uncorrected results without considering rRFLDI.
![Page 132: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/132.jpg)
115
Figure 6.8: Comparison of LDI-MS factor concentrations corrected using relative response factors (rRF) to uncorrected LDI factor concentrations (Scenario 1).
6.3.4.4 Uncertainty of PMF results
The uncertainty estimate, σtot, includes both the statistical uncertainty (σbs) and the
uncertainty arising from the intraday variability of the measurements (σintrad, see Fig.
6.9, Section 6.2.3.2 and Supplement B). In order to assess the impact of the intraday
variability consideration (σintrad) in estimating the uncertainty besides the statistical
uncertainty obtained from the bootstrapping approach (σbs), we compared the ratios
σbs/σtot for the different factors: for traffic1 the ratio was 0.83, for traffic2 0.75, for
BBeff 0.86, for BBineff1 0.80, for BBineff2 0.72, for LMW-OA 0.68, and for bio-OA
0.68. Thus, it was important to propagate σintrad. The relative uncertainties for the
median factor concentrations ranged within 0.15 (for traffic1), 0.16 (bio-OA), 0.17
(LMW-OA), 0.18 (BBineff2), 0.20 (traffic2), 0.22 (BBeff, and 0.28 (BBineff1). Unlike
the relative error of σtot of traffic1 (0.63 at the 10th percentile concentration and 0.13 at
the 90th percentile concentration), traffic2 (0.68 and 0.11), BBeff (0.55 and 0.18),
BBineff1 (0.47 and 0.20), BBineff2 (0.37 and 0.14), and LMW-OA (0.27 and 0.16) the
![Page 133: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/133.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
116
relative error of bio-OA (0.20 and 0.15) only depended weakly on the factor
concentration.
Figure 6.9: Relative for the different PMF factors as a function of the factor concentration.
Throughout the measurement campaign, subsets of filters were analyzed
repeatedly in order to assess the repeatability over longer time periods (see details in
Supplement B, Fig. SI.B.10). Most factors did not show significant changes of the
attributed concentration during the measurement campaign. However, BBeff showed
decreasing and LMW-OA increasing concentrations as a function of the measurement
time. This could suggest an uncertain separation of these 2 factors. However, for these
long time delays the variability increased and only few samples were repeated with such
long time delays. Furthermore, the intra-day variance largely explains the total variance
(traffic1 97%, traffic2 94%, BBeff 85%, BBineff1 89%, BBineff2 82%, LMW-OA
79%, bio-OA 97%, details in Supplement B Fig. SI.B.10).
6.3.4.5 Factor variability and contribution
The time series of all factors are shown for the 9 sites in Fig. 6.10 as relative
contributions to OA and summarized as yearly averages in Tab. 6.2. In the yearly
average for all sites, traffic1 contributes 7%, traffic2 12%, BBeff 7%, BBineff1 17%,
BBineff2 32%, LMW-OA 21%, and bio-OA 5% to OA as measured by LDI-MS.
![Page 134: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/134.jpg)
117
Figure 6.10: Relative factor time series of 7 identified factors for all nine sites in study area.
At the southern Alpine Swiss sites, the wood burning influenced categories
(WB=BBeff+BBineff1+ BBineff2) contribute more than at the northern sites (70% vs
50%). Moreover, in winter WB explains 81% and 61% of OA in the south and the
north, respectively. This difference is mostly caused by the higher relative contribution
of BBineff1 (34% in the south vs 14% in the north) since BBineff2 (28% and 30%) and
BBeff (7% and 5%) do not show strong geographical differences. The ratio
BBineff1/BB (Fig. 6.11b) shows enhanced contributions of BBineff1 at the southern
Alpine sites, especially during high pollution episodes in winter. This suggests different
wood burning regimes in the 2 regions, as already discussed above.
![Page 135: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/135.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
118
Figure 6.11: a) time series of traffic1 normalized to eBCtr and ECf in comparison to OCf /ECf in Magadino, b) influence of inefficient wood burning emissions (BBineff1) in comparison to the sum of wood burning influenced factors (BB=BBineff1+BBineff2+BBeff) for the entire datasets.
After scaling of the mass spectra to OA, absolute factor concentrations have still to
be interpreted with caution. One reason for this is that the relative response factors of
the sources/factors are not known, another is that certain species (in analogy to the
cooking-aerosol sample) may not be detected by LDI-MS if externally mixed. In
support of the quantitative interpretation of our results, we note that the sum of the
wood burning related LDI-MS factors correlates well with the offline AMS counterpart
(R2LDI,oAMS =0.82, described in Section 6.3.4.3). On the yearly average 24% of the
measured OA is apportioned to secondary OA. In summer, a bigger fraction is attributed
to secondary OA (35% compared to 16% in winter). 35% (summer) and 1% (winter) of
the secondary OA is attributed to biogenic sources. For some samples from the same
period in Magadino, also the fossil and non-fossil content of OC and EC was
determined (Vlachou et al., in prep) based on the method of Zhang et al. (2012).
Vlachou et al. (in prep.) observed increased OCf/ECf ratios in summer which suggests
![Page 136: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/136.jpg)
119
other fossil POA sources in summer than in winter, or secondary formation of fossil
OA. The higher traffic1 / eBCtr and traffic1/ECf ratios in summer are in agreement with
enhanced OCf /ECf ratios (Fig. 6.11a). However, a part of traffic can also be mixed into
BB leading to an underestimation of the traffic concentrations in winter. Overall this
suggests that traffic1 represents aged or secondary traffic OA.
Table 6.2: yearly averages of the relative factor contributions and NOx concentrations. yearly
average, % traffic1 traffic2 BBeff BBineff1 BBineff2 bio-OA LMW-OA NOx, ppb
Bern 11 24 7 13 30 3 12 48 Zürich 9 27 8 9 27 5 15 24 St. Gallen 8 16 4 17 27 7 21 24 Basel 6 12 7 17 32 5 21 21 Frauenfeld 8 16 6 13 30 7 20 21 Vaduz 7 12 7 20 27 7 20 20 Payerne 5 15 12 9 25 11 23 9 Magadino 3 11 7 27 27 8 17 20 S. Vittore 2 4 3 40 32 5 14 18
6.4 Summary and conclusions
In this study, we advanced a known method for the chemical characterization of
particulate matter collected on quartz-fiber filters by LDI-MS and applied the method to
819 samples. The method included the use of silver nitrate for m/z calibration and the
automated peak integration of the mass spectra at unit-mass resolution. The benefit of
LDI-MS measurements for the chemical characterization and a better understanding of
the sources contributing to the ambient PM10 was assessed at nine sites in central
Europe throughout the entire year 2013.
Wood combustion smog chamber experiments revealed an influence of the
burning conditions on the mass spectral signature. Tunnel samples used as a reference
for traffic related emissions show mass spectral signatures distinctly different from
wood combustion. Key m/z’s identified in the wood burning and traffic signatures
showed links to expected markers as e.g., levoglucosan and NOx, respectively. The
ambient mass spectral information was further used for source apportionment by PMF.
Thereby, the influence of efficient and inefficient wood burning was separated. The
![Page 137: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/137.jpg)
Chapter 6. LDI-MS for understanding of OA sources in Central Europe
120
extracted wood burning emissions correlated with the results from offline AMS source
apportionment. Other components are more difficult to compare quantitatively because
of different source separations in PMF as well as differences in the relative response
factors (rRF) of OA components. rRF determined in this study are uncertain and,
therefore, not used for correcting the LDI-PMF results. The influence of traffic
emissions was represented by 2 factors. One of these could clearly be linked to BC-
related traffic (eBCtr) and NOx, and thus to primary emissions. The other, when
normalized to eBCtr, showed a similar behavior as OCf /ECf, and was therefore
attributed to aged/secondary traffic OA. A factor was attributed to biogenic SOA based
on its concentration exponentially increasing with temperature. Another OA factor was
characterized by low-molecular-weight ions and was correlated with NH4+ and was
attributed to SOA from an unknown source.
Acknowledgements
This work was supported by the Swiss Federal Office of Environment,
Liechtenstein, Ostluft, the Cantons Basel, Graubünden, and Thurgau. We also thank
AWEL Zürich for providing us with samples collected in Islisbergtunnel. JCC received
financial support from the ERC under grant ERC-CoG-615922-BLACARAT.
![Page 138: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/138.jpg)
121
7 Conclusions and outlook
Linking long-term particulate matter air pollution exposure and the contributing
sources to health effects is important to develop effective mitigation strategies. In this
dissertation, we contribute to this research topic by developing analytical approaches
capable of providing long-term source contributions. In particular, we characterized an
offline application of HR-ToF-AMS based on particulate matter extracted from filter
samples. We applied this technique to create long-term source contribution estimates to
OA at several sites in central Europe. Furthermore, we optimized a method using laser-
desorption/ionization mass spectrometry for measuring PM collected on filter samples.
This type of analysis was applied to the same set of samples as offline AMS and was
used in order to obtain a better understanding of the chemical nature of the contributing
PM sources.
In Chapter 4, we present the method development and characterization of the
offline AMS application using water-extracted PM from filter samples. While
oxygenated fragments are well recovered, hydrocarbon-like fragments are not well
captured. The key element of this study is the comparison of offline AMS source
apportionment results to collocated source apportionment results based on online
ACSM analyses. This comparison yields recovery factors of the different resolved
factors: traffic, cooking, biomass burning, and oxygenated OA. They are essential for
estimating the source contributions to OA based on the offline AMS analysis of WSOA.
![Page 139: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/139.jpg)
Conclusions and outlook
122
For 9 sites in central Europe, sources contributing to OA are estimated using offline
AMS (Chapter 5). For quantification and validation of the source apportionment results,
offline AMS analyses are coupled with state-of-the art measurements of water-soluble
carbon, organic carbon, elemental carbon, water-soluble ions and levoglucosan as an
organic marker for wood burning. In this study, the contribution of primary OA from
traffic, cooking, and biomass burning, besides two seasonally separated OOA factors
could be separated. BBOA shows the highest geographical variability among all factors:
at the alpine valley sites concentrations are clearly higher than in northern Switzerland
and Liechtenstein. This is hypothesized to be related to differences in the burning
conditions between the two geographical regions. Different sensitivity tests also
including the measurement reproducibility allow a detailed examination of the
uncertainty of the factor contributions. Furthermore, cellulose analyses (enzymatic
degradation) are used among other approaches to estimate the contribution of primary
biological OA which could not be resolved based on the offline AMS analyses.
Together with results presented in other studies (Fig. 7.1), the offline AMS approach
yielded results for 13 sites in Europe and for a high pollution episode in China (4 sites).
These wide applications show the potential of the offline AMS approach.
Figure 7.1: Average contributions of OA sources resolved using offline AMS data discussed in this thesis and other studies (this dissertation, and from Bozzetti et al., 2016, 2017a, 2017b, Huang et al., 2014).
![Page 140: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/140.jpg)
123
In a third study (Chapter 6), we optimized the application of laser-
desorpion/ionization mass spectrometry for analyzing PM. Subsequently this approach
was applied to the same samples as analyzed by offline ASM (in Chapter 5) and the
benefit of LDI-MS in understanding OA sources was assessed. In this study, the
influence of traffic (two factors), wood buring (three factors) was separated from two
factors related to secondary OA. Comparison of the cumulative influence of wood
burning resolved by LDI to offline AMS and collocated levoglucosan give trust in the
separation of biomass burning OA. The three wood burning factors could be related to
the burning condition and inefficient burns show a higher contribution at the alpine
valley compared to the site on the northern swiss plateau. This might explain the higher
BBOA concentrations observed with offline AMS. Furthermore, one of the traffic-
related factors showed higher factor concentrations per eBCtr during the warm season
which coincides with higher OCf to ECf ratios at the same site making it possible that
this factor resolves aged traffic emissions.
While routine air quality monitoring encompasses measurements of some gas-
phase constituents, typically PM10 (alternatively PM in other size fractions as PM2.5 or
PM1) concentration is the only parameter assessing PM. Such samples can be accessed
worldwide, also a posteriori. This makes the assessment of past long-term exposure
possible. Moreover, information on extreme pollution episodes, which are only partly
predictable, is accessible through a posteriori analysis. Furthermore, such offline
datasets can be completed with any necessary complementary filter-based chemical
analysis specific to the research question. In contrast to online, offline measurements
also allow testing the measurement repeatability allowing detailed uncertainty
assessments. Centralized analysis of OA sources allows comparing results from
different sites and also limits the influence of instrumental differences by only using one
instrument in the lab, while similar online datasets are much more subjected to such
influences.
Performing measurements online has the major advantage of a better time-
resolution which helps understanding variabilites on a diurnal scale. The low time-
resolution limits analyzing episodes with offline approaches. Increasing the temporal
![Page 141: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/141.jpg)
Conclusions and outlook
124
resolution (several filters per day) would result in a better understanding of the
variabilities and would provide further constraints based on diurnal patterns of the
resolved sources. The impossibility to resolve PBOA (Chapter 5) is closely related to
the absence of size resolution (in contrast to Bozzetti et al., 2016). Therefore, for future
studies it is recommended to include PM1 (or PM2.5) filter samples at the cost of
reducing the number of sites which enables the further study of sources contributing to
coarse OA. Including PM1 samples, furthermore, allows a direct comparison to online
AMS/ACSM results and, thus, the determination (confirmation) of the source-specific
recoveries at other sites. Expanding the temporal coverage (to e.g., 4.5 years) using
PM1 and PM10 samples while keeping the temporal resolution constant (1 sample
every 4 days) and only focusing on 1 site, also the number of samples remains constant.
In Dockery et al. (1993), data from 14-16 years is compiled. When reducing the
temporal resolution to 1 sample every 12 days and still keeping the number of samples
constant, the temporal coverage can be further expanded to 14 years. This way it might
be possible to apportion health effects to chronic exposure to single OA sources (given
quantitative source contributions from offline AMS).
The recent development of a time-of-flight unit for the AMS with double the
resolution allows a better separation of ions at the same nominal mass. This is especially
helpful for nitrogen containing species, typically shadowed by other more prominent
peaks. Such ions are possibly important in the coarse PM fraction (Bozzetti et al.,
2016). Yet the AMS OA measurements suffer from an artifact which increases OA
concentrations (CO2+) when injecting nitrate (Pieber et al., 2016). As routinely
performed during offline AMS measurement, regular ammonium nitrate calibrations
should also be performed during online AMS/ACSM campaigns. Offline AMS
measurements also access larger particles than AMS/ACSM. Thus also the interference
with carbonates (decomposing thermally leading to the detection of CO2+) is important
and leading to further inorganic interference at this fragment.
Interference with air (N2+) makes it under most conditions impossible to resolve
the fragment CO+ in AMS analyses (online and offline). As already shown by Bozzetti
(2017b) it is possible to overcome this limitation in offline AMS by changing the carrier
gas to argon making valuable information accessible. As mentioned in Chapter 4,
![Page 142: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/142.jpg)
125
offline AMS analyses are not quantitative alone but need WSOC and/or OC
measurements. By spiking the liquid extracts with isotopically labelled compounds the
main inorganic (15NO3-, 34SO4
2-) can be retrieved quantitatively. Establishing similar
approaches for WSOA might make collocated WSOC and ion chromatography
obsolete.
The sulfur-containing OA factor separated in Chapter 5 is hypothetically linked to
traffic activity which might stem from dust resuspension. For a better understanding of
this influence, separated only in PM10 analyses, further work is required which could
involve the analysis of tunnel samples in different size fractions. Based on the
quantitative source apportionment results obtained from offline AMS, further studies
can be conducted. By performing further measurements of health related parameters
(e.g., ROS, exposing cells to filter extracts) or collecting information on the number of
hospital entries related to respiratory diseases, the implications of the different sources
might be quantified. Completing the dataset with trace metal analyses might also allow
the separation of the impact of organic sources from the one of inorganic species.
Coupled offline AMS and UV-VIS absorption measurements, potentially allow
estimating the direct RF forcing caused by different OA sources. As stated earlier, the
lack of information on the preindustrial aerosol contributes largely to the uncertainty in
estimating RF. Model validation requires datasets like the one presented in Chapter 6
(comparable to the intercomparison in Ciarelli et al., 2016, based on Crippa et al.,
2014). Since any liquid matrix is potentially accessible to offline AMS, it is tempting to
attempt to access information on the bulk OA from preindustrial times using ice and
sediment cores. Environmental matrices in general can be analyzed, such as rain, cloud
and runoff or also ground water.
The LDI based technique does not require an extraction step and thus does not
suffer from losses during such a step, but only matrices desorbing and ionizing upon
interaction with the laser can be detected. Therefore, a better characterization of the
laser (power density) is crucial. The instrument used in Chapter 6 does not extract the
produced ions orthogonallay to the ToF unit which potentially reduces both the
accuracy and precision of the m/z calibration. Another factor potentially playing a role
in the quality of the m/z calibration and reproducibility is the sample surface, i.e.
![Page 143: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/143.jpg)
Conclusions and outlook
126
collecting aerosol on a smoother surface like an aluminium plate instead of a quartz-
fiber filter might improve the chemical resolution (attribution of signal at a nominal
mass to different ions/fragments). Similarly also an extraction step and subsequent
deposition of the extract on a steel plate could be applied at the cost of losses during the
extraction. Such improvements would potentially provide further insight into the
precursors of SOA. Moreover, the effect of matrix effects needs to be further studied
explicitly, e.g. by comparing LDI source apportionment results in different
environments (e.g., marine, minimal biogenic emissions) to results from other
techniques. Even if factors like LDI-BB correlate with the corresponding offline AMS
factor, further research is required to understand response factors of different sources
before achieving quantitative relative source contributions. The approach outlined for
LDI analyses of ambient PM samples (Chapter 6) could also be applied to ambient
microorganism samples (where different internal standards are required to account for
the different m/z range) to separate the different species/families/etc.
While ca. 35% of the global city population experiences decreasing PM levels
(mostly in Europe and America), 30% of the city global population suffers from
increasing annual average PM levels between 2008 and 2015 (Eastern Mediterrannean,
Southeast Asia). For Africa similar information is not available. In the Eastern
Mediterrannean, Ryiadh (Saudi Arabia), Ma’amer (Bahrain), Greater Cairo (Egypt),
Doha (Qatar), Abu Dhabi (U.A.E.) are pointed out and in Southeast Asia Delhi (India)
and Dhaka (Bangladesh), but also in other regions there are cities with extreme PM
levels such as Ulaanbaatar (Mongolia), Beijing (China) and Dakar (Senegal) (WHO,
2016, sorted by decreasing PM10 levels within the geographical region). Since not all
these sites are easily accessible to online instrumentation, filter-based offline analysis
approaches can be pointed out as a key tool for analyzing PM. Certain cities are
suspected to be strongly affected by dust from adjacent desertous regions. In such
situations, interference with carbonates are important when using offline AMS, making
it necessairy to resolve (or measure) carbonates or to use alternative measurement
approaches (e.g., acid fumigation of samples). Some sites might also be strongly
affected by several non-water-soluble sources (e.g., in proximity of oil refineries)
making the bulk OA only little water-soluble (as e.g. observed for Baghdad, Hamad et
al., 2015). In such situations, it might be advantageous to use different solvents (offline
![Page 144: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/144.jpg)
127
AMS) or to switch to approaches without an extraction step such as the LDI technique
presented in Chapter 6.
Source apportionment results based on offline AMS and LDI data allowed
separating SOA seasonally, but both approaches could not identify SOA from different
precursors by mass spectral comparison. Offline AMS is limited by the hard ionization
leading to the loss of chemical information required for distinguishing SOA types based
on their precursor and LDI among other factors by the poor chemical resolution and not
well understood ionization pathways. To deepen our understanding of SOA types,
instrumentation equipped with soft ionization and a high chemical resolution is needed.
The recent application of extractive electrospray ionization (EESI) provides more
chemical information (given a high enough chemical resolution defined by the ToF unit
or addition of other instrumentation). In a similar way as the AMS, mass spectrometry
coupled with EESI can be used offline for nebulized filter extracts. Solvent impurities
can be thought only to affect single ions because of the soft ionization pathway which
makes it possible to use, other than in the current offline AMS framework, also organic
solvents like methanol. The use of alternative solvents makes more compound classes
accessible than for the current offline AMS setup (e.g., traffic emissions, compare
Chapter 4). Analyses with ultra-high-resolution mass spectrometry (e.g., Orbitrap)
provide even more chemical information than EESI MS but require extensive lab work
and are, therefore, probably not applicable for studies with a similar amount of samples
as in the presented dissertation.
![Page 145: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/145.jpg)
Conclusions and outlook
128
![Page 146: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/146.jpg)
129
8 Bibliography
Aiken, A. C., Salcedo, D., Cubison, M. J., Huffman, J. A., DeCarlo, P. F., Ulbrich, I.
M., Docherty, K. S., Sueper, D., Kimmel, J. R., Worsnop, D. R., Trimborn, A.,
Northway, M., Stone, E. A., Schauer, J. J., Volkamer, R. M., Fortner, E., de Foy, B.,
Wang, J., Laskin, A., Shutthanandan, V., Zheng, J., Zhang, R., Gaffney, J., Marley, N.
A., Paredes-Miranda, G., Arnott, W. P., Molina, L. T., Sosa, G., and Jimenez, J. L.:
Mexico City aerosol analysis during MILAGRO using high resolution aerosol mass
spectrometry at the urban supersite (T0) - Part 1: Fine particle composition and organic
source apportionment, Atmos. Chem. Phys., 9, 6633-6653, doi:10.5194/acp-9-6633-
2009, 2009.
Albrecht, B. A.: Aerosols cloud microphysics, and fractional cloudiness, Science,
245, 1227-1230, 1989.
Alfarra, M. R., Coe, H., Allan, J. D., Bower, K. N., Boudries, H., Canagaratna, M.
R., Jimenez, J. L., Jayne, J. T., Garforth, A. A., Li, S.-M., and Worsnop, D. R.:
Characterization of urban and rural organic particulate in the Lower Fraser Valley using
two Aerodyne Aerosol Mass Spectrometers, Atmos. Environ., 38, 5745-5758,
doi:10.1016/j.atmosenv.2004.01.054, 2004.
![Page 147: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/147.jpg)
Bibliography
130
Allan, J. D., Jimenez, J. L., Williams, P. I., Alfarra, M. R., Bower, K. N., Jayne, J. T.,
Coe, H., and Worsnop, D. R.: Quantitative sampling using an aerodyne aerosol mass
spectrometer - 1. Techniques of data interpretation and error analysis, J. Geophys. Res.-
Atmos., 108, 4090, doi:10.1029/2002JD002358, 2003.
Allan, J. D., Delia, A. E., Coe, H., Bower, K. N., Alfarra, M. R., Jimenez, J. L.,
Middlebrook, A. M., Drewnick, F., Onasch, T. B., Canagaratna, M. R., Jayne, J. T., and
Worsnop, D. R.: A generalised method for the extraction of chemically resolved mass
spectra from aerodyne aerosol mass spectrometer data, J. Aerosol Sci., 35, 909-922,
doi:10.1016/j.jaerosci.2004.02.007, 2004.
Allan, J. D.,Williams, P. I., Morgan,W. T., Martin, C. L., Flynn, M.J., Lee, J.,
Nemitz, E., Phillips, G. J., Gallagher, M.W., and Coe, H.: Contributions from transport,
solid fuel burning and cooking to primary organic aerosols in two UK cities, Atmos.
Chem. Phys., 10, 647-668, doi:10.5194/acp-10-647-2010, 2010.
Andreae, M. O.: Aerosols before pollution, Science, 315, 50,
doi:10.1126/science.1136529, 2007.
Aubriet, F., Carré, V.: Potential of laser mass spectrometry for the analysis of
environmental dust particles - a review, Anal. Chim. Act., 659, 34-54,
doi:10.1016/j.aca.2009.11.047, 2010.
Baltensperger, U., Kalberer, M., Dommen, J., Paulsen, D., Alfarra, M. R., Coe, H.,
Fisseha, R., Gascho, A., Gysel, M., Nyeki, S., Sax., M., Steinbacher, M., Prevot, A. S.
H., Sjögren, S., Weingartner, E., Zenobi, R.: Secondary organic aerosols from
anthropogenic and biogenic precursors, Faraday, Discuss., 130, 265-278,
doi:10.1039/b417367h ,2005.
Bell, M. L., Davis, D. L.: Reassessment of the lethal London fog of 1952: novel
indicators of acute and chronic consequences of acute exposure to air pollution,
environmental health perspectives, 109, no. Suppl. 3, 389, doi:10.1289/ehp.01109s3389,
2001.
![Page 148: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/148.jpg)
131
Bell, M. L., Davis, D. L., and Fletcher, T.: A retrospective of mortality from the
London smog episode of 1952: the role of influenza and pollution, Environmental
health perspectives, 112, doi:10.1289/ehp.6539, 2004.
BFE (Bundesamt für Energie Schweiz): Schweizerische Holzenergiestatistik:
Erhebung für das Jahr 2013, Bundesamt für Energie Schweiz, Ittigen, 2013.
Birch, M. E. and Cary, R. A.: Elemental carbon-based method for monitoring
occupational exposures to particulate diesel exhaust, Aerosol Sci. and Tech., 25, 221-
241, doi:10.1080/02786829608965393, 1996.
Borisov, R. S., Polokov, N. Y., Zhilyaev, D. I., and Zaikin, V. G.: Matrix effect in
matrix assisted laser desorption/ionization mass spectra of derivatized oligometric
polyols, Rapid. Commun. Mass. Spectrom., 27, 333-338, doi:10.1002/rcm.6449, 2013.
Bouwman, A. F., Lee, D. S., Asman, W. A. H., Dentener, F. J., Van der Hoek, K.
W., and Olivier, J. G. J.: A global high resolution emission inventory for ammonia,
Glob. Biogeochem. Cycles, 11, 561-587, doi:10.1029/97gb02266, 1997.
Bozzetti, C., Daellenbach, K., R., Hueglin, C., Fermo, P., Sciare, J., Kasper-Giebl,
A., Mazar, Y., Abbaszade, G., El Kazzi, M., Gonzalez, R., Shuster Meiseles, T., Flasch,
M., Wolf, R., Křepelová, A., Canonaco, F., Schnelle-Kreis, J., Slowik, J. G.,
Zimmermann, R., Rudich, Y., Baltensperger, U., El Haddad, I., and Prévôt, A. S. H.:
Size-resolved identification, characterization, and quantification of primary biological
organic aerosol at a European rural site, Environ. Sci. Technol., 50, 3425-3434, doi:
10.1021/acs.est.5b05960, 2016.
Bozzetti, C. El Haddad, I., Salameh, D., Daellenbach, K. R., Fermo, P., Gonzalez, R.,
Minguillón, M. C., Iinuma, Y., Poulain, L., Müller, E., Slowik, J. G., Jaffrezo, J.-L.,
Baltensperger, U., Marchand, N., and Prévôt, A. S. H.: Organic aerosol source
apportionment by offline-AMS over a full year in Marseille, Atmos. Chem. Phys., 17,
8247-8268, doi:10.5194/acp-17-8247-2017, 2017a.
![Page 149: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/149.jpg)
Bibliography
132
Bozzetti, C., Sosedova, Y., Xiao, M., Daellenbach, K. R., Ulevicius, V., Dudoitis, V.,
Mordas, G., Byčenkienė, S., Plauškaitė, K., Vlachou, A., Golly, B., Chazeau, B.,
Besombes, J.-L., Baltensperger, U., Jaffrezo, J.-L., Slowik, J. G., El Haddad, I., and
Prévôt, A. S. H.: Argon offline-AMS source apportionment of organic aerosol over
yearly cycles for an urban, rural and marine site in Northern Europe, Atmos. Chem.
Phys., 17, 117-141, doi:10.5194/acp-17-117-2017, 2017b.
Brohan, P., Kennedy, J. J., Harris, I., Tett, S. F. B., and Jones, P. D.: Uncertainty
estimates in regional and global observed temperature changes: A new data set from
1850, J. Geophys. Res., 111, D12106, doi:10.1029/2005JD006548, 2006.
Brown, S. G., Eberly, S., Paatero, P., and Norris, G. A.: Methods for estimating
uncertainty in PMF solutions: Examples with ambient air and water quality data and
guidance on reporting PMF results, Sci. Tot. Environ., 518-519, 626-635,
doi:10.1016/j.scitotenv.2015.01.022, 2015.
Brunekreef, B., and Forsberg, B.: Epidemiological evidence of effects of coarse
airborne particles on health, Eur. Respir. J., 26, 309-318, doi:
10.1183/09031936.05.00001805, 2005
Bruns, E. A., Krapf, M., Orasche, J., Huang, Y., Zimmermann, R., Drinovec, L.,
Močnik, G., El-Haddad, I., Slowik, J. G., Dommen, J., Baltensperger, U. and Prévôt, A.
S. H.: Characterization of primary and secondary wood combustion products generated
under different burner loads, Atmos. Chem. Phys., 15, 2825-2841, doi:10.5194/acp-15-
2825-2015, 2015.
Bruns, E. A., El Haddad, I., Slowik, J. G., Kilic, D., Klein, F., Baltensperger, U., and
Prévôt, A. S. H.: Identification of significant precursor gases of secondary organic
aerosols from residential wood combustion, Sci. Rep., 6, 27781, doi:10.1038/srep27881,
2016.
![Page 150: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/150.jpg)
133
Canagaratna, M. R., Jayne, J. T., Jimenez, J. L., Allan, J. D., Alfarra, M. R., Zhang,
Q., Onasch, T. B., Drewnick, F., Coe, H., Middlebrook, A., Delia, A., Williams, L. R.,
Trimborn, A. M., Northway, M. J., DeCarlo, P. F., Kolb, C. E., Davidovits, P. and
Worsnop, D. R.: Chemical and microphysical characterization of ambient aerosols with
the Aerodyne aerosol mass spectrometer, Mass Spectrom. Rev. 26:185-222,
doi:10.1002/mas.20115, 2007.
Canonaco, F., Crippa, M., Slowik, J. G., Baltensperger, U., and Prévôt, A. S. H.:
SoFi, an IGOR-based interface for the efficient use of the generalized multilinear engine
(ME-2) for the source apportionment: ME-2 application to aerosol mass spectrometer
data, Atmos. Meas. Tech., 6, 3649-3661, doi:10.5194/amt-6-3649-2013, 2013.
Canonaco, F., Slowik, J. G., Baltensperger, U., and Prévôt, A. S. H.: Seasonal
differences in oxygenated organic aerosol composition: implications for emissions
sources and factor analysis, Atmos. Chem. Phys. 15, 6993-7002, doi:10.5194/acp-15-
6993-2015, 2015.
Canonaco, F., Daellenbach, K. R., Crippa, M., ElHaddad, I., Bozzetti, C., Huang, R.-
J., Baltensperger, U., Hüglin, C., Herich, H., and Prévôt, A. S. H.: A novel strategy for
the source apportionment of long-term ACSM data based on ME-2 with SoFi:
Automatic Rolling PMF window (AuRo SoFi), in preparation, 2018.
Carslaw, K. S., Lee, L. A., Reddington, C. L., Pringle, K. J., Rap, A., Forster, P. M.,
Mann, G. W., Spracklen, D. V., Woodhouse, M. T., Regayre, L. A., Pierce, J. R.: Large
contribution of natural aerosols to uncertainty in indirect forcing, Nature, 503,67-71,
doi: 10.1038/nature12674, 2013.
Cavalli, F., Viana, M., Yttri, K. E., Genberg, J., and Putaud, J. P.: Toward a
standardised thermal-optical protocol for measuring atmospheric organic and elemental
carbon: the EUSAAR protocol, Atmos. Meas. Tech., 3, 79-89, doi:10.5194/amt-3-79-
2010, 2010.
![Page 151: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/151.jpg)
Bibliography
134
Chen, Q, Miyazaki, Y., Kawamura, K., Matsumoto, K., Coburn, S. C., Volkamer, R.,
Iwamoto, Y., Kagami, S., Deng, Y., Ogawa, S., Ramasamy, S., Kato, S., Ida, A., Kajii,
Y., and Mochida, M.: Characterization of chromophoric water-soluble organic matter in
urban, forest, and marine aerosols by HR-ToF-AMS analysis and excitation emission
matrix spectroscopy, Environ. Sci. Technol., 50, 10351-10360,
doi:10.1021/acs.est.6b01643, 2016.
Chirico, R., Prevot, A. S. H., DeCarlo, P. F., Heringa, M. F., Richter, R.,
Weingartner, E., and Baltensperger, U.: Aerosol and trace gas vehicle emission factors
measured in a tunnel using an Aerosol Mass Spectrometer and other online
instrumentation, Atmos. Environ., 45, 2182-2192, doi:10.1016/j.atmosenv.2011.01.069,
2011.
Ciarelli, G., Aksoyoglu, S., Crippa, M., Jimenez, J.-L., Nemitz, E., Sellegri, K.,
Äijälä, M., Carbone, S., Mohr, C., O'Dowd, C., Poulain, L., Baltensperger, U., and
Prévôt, A. S. H.: Evaluation of European air quality modelled by CAMx including the
volatility basis set scheme, Atmos. Chem. Phys., 16, 10313-10332, doi:10.5194/acp-16-
10313-2016, 2016.
Corbin, J. C., Othman, A., Allan, J. D., Worsnop, D. R., Haskins, J. D., Sierau, B.,
Lohmann, U., and Mensah, A. A.: Peak-fitting and integration imprecision in the
Aerodyne aerosol mass spectrometer: effects of mass accuracy on location-constrained
fits, Atmos. Meas. Tech., 8, 4615-4636, doi:10.5194/amt-8-4615-2015, 2015.
Crippa, M., Canonaco, F., Slowik, J. G., El Haddad, I., De-Carlo, P. F., Mohr, C.,
Heringa, M. F., Chirico, R., Marchand, N., Temime-Roussel, B., Abidi, E., Poulain, L.,
Wiedensohler, A., Baltensperger, U., and Prévôt, A. S. H.: Primary and secondary
organic aerosol origin by combined gas-particle phase source apportionment, Atmos.
Chem. Phys., 13, 8411-8426, doi:10.5194/acp-13-8411-2013, 2013a.
Crippa, M., DeCarlo, P. F., Slowik, J. G., Mohr, C., Heringa, M. F., Chirico, R.,
Poulain, L., Freutel, F., Sciare, J., Cozic, J., Di Marco, C. F., Elsasser, M., Nicolas, J.
B., Marchand, N., Abidi, E., Wiedensohler, A., Drewnick, F., Schneider, J., Borrmann,
![Page 152: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/152.jpg)
135
S., Nemitz, E., Zimmermann, R., Jaffrezo, J.-L., Prévôt, A. S. H., and Baltensperger, U.:
Wintertime aerosol chemical composition and source apportionment of the organic
fraction in the metropolitan area of Paris, Atmos. Chem. Phys., 13, 961-981,
doi:10.5194/acp-13-961-2013, 2013b.
Crippa, M., El Haddad, I., Slowik, J. G., DeCarlo, P. F., Mohr, C., Heringa, M. F.,
Chirico, R., Marchand, N., Sciare, J., Baltensperger, U., and Prevot, A. S. H.:
Identification of marine and continental aerosol sources in Paris using high resolution
aerosol mass spectrometry, J. Geophys. Res.-Atmos., 118, 1950-1963,
doi:10.1002/jgrd.50151, 2013c.
Crippa, M., Canonaco, F., Lanz, V. A., Äijälä, M., Allan, J. D., Carbone, S., Capes,
G., Ceburnis, D., Dall’Osto, M., Day, D. A., De-Carlo, P. F., Ehn, M., Eriksson, A.,
Freney, E., Hildebrandt Ruiz, L., Hillamo, R., Jimenez, J. L., Junninen, H., Kiendler-
Scharr, A., Kortelainen, A.-M., Kulmala, M., Laaksonen, A., Mensah, A. A., Mohr, C.,
Nemitz, E., O’Dowd, C., Ovadnevaite, J., Pandis, S. N., Petäjä, T., Poulain, L.,
Saarikoski, S., Sellegri, K., Swietlicki, E., Tiitta, P., Worsnop, D. R., Baltensperger, U.,
and Prévôt, A. S. H.: Organic aerosol components derived from 25 AMS data sets
across Europe using a consistent ME-2 based source apportionment approach, Atmos.
Chem. Phys., 14, 6159-6176, doi:10.5194/acp-14-6159-2014, 2014.
Cubison, M. J. and Jimenez, J. L.: Statistical precision of the intensities retrieved
from constrained fitting of overlapping peaks in high-resolution mass spectra, Atmos.
Meas. Tech., 8, 2333-2345, doi:10.5194/amt-8-2333-2015, 2015.
Daellenbach, K. R., Bozzetti, C., Krepelova, A., Canonaco, F., Huang, R.-J., Wolf,
R., Zotter, P., Crippa, M., Slowik, J., Zhang, Y., Szidat, S., Baltensperger, U., Prévôt,
A. S. H., and El Haddad, I.: Characterization and source apportionment of organic
aerosol using offline aerosol mass spectrometry, Atmos. Meas. Tech., 9, 23-39,
doi:10.5194/amt-9-23-2016, 2016.
Daellenbach, K. R., Stefenelli, G., Bozzetti, C., Vlachou, A., Fermo, P., Gonzalez,
R., Piazzalunga, A., Colombi, C., Canonaco, F., Hueglin, C., Kasper-Giebl, A., Jaffrezo,
![Page 153: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/153.jpg)
Bibliography
136
J.-L., Bianchi, F., Slowik, J. G., Baltensperger, U., El-Haddad, I., and Prévôt, A. S. H.:
Long-term chemical analysis and organic aerosol source apportionment at nine sites in
central Europe: source identification and uncertainty assessment, Atmos. Chem. Phys.,
17, 13265-13282, doi:10.5194/acp-17-13265-2017, 2017.
Davison, A. C. and Hinkley, D. V.: Bootstrap methods and their Application,
Cambridge University Press, Cambridge, UK, 582 pp., doi:
10.1017/CBO9780511802843, 1997.
De Gouw, J. and Jimenez, J. L.: organic aerosols in the earth’s atmosphere, Environ.
Sci. Technol., 43, 7614-7618, doi:10.1021/es9006004, 2009.
De Hoffmann and, E., and Stroobant, V.: Mass spectrometry: principles and
applications, second edition, John Wiley and Sons, Inc., New York, 1999.
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.: Field-deployable, high-resolution, time-of-flight aerosol mass spectrometer, Anal.
Chem., 78, 8281-8289, doi:10.1021/ac061249n, 2006.
DeCarlo, P. F., Dunlea, E. J., Kimmel, J. R., Aiken, A. C., Sueper, D., Crounse, J.,
Wennberg, P. O., Emmons, L., Shinozuka, Y., Clarke, A., Zhou, J., Tomlinson, J.
Collins, D. R., Knapp, D., Weinheimer, A. J., Montzka, D. D., Campos, T., and
Jimenez, J. L.: Fast airborne aerosol size and chemistry measurements above Mexico
city and Central Mexico during the MILAGRO campaign, Atmos. Chem. Phys., 8,
4027-4048, doi:10.5194/acp-8-4027-2008, 2008.
Dockery, D. W., Pope, C. A., Xu, X., Spengler, J. D., Ware, J. H., Fay, M. E., Ferris,
B. G., and Speizer, F. E.: An association between air pollution and mortality in six U.S.
cities, New England Journal of Medicine, 329, 1753-1759,
doi:10.1056/NEJM199312093292401, 1993.
![Page 154: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/154.jpg)
137
El Haddad, I., D’Anna, B., Temime-Roussel, B., Nicolas, M., Boreave, A., Favez,
O., Voisin, D., Sciare, J., George, C., Jaffrezo, J.-L., Wortham, H., and Marchand, N.:
Towards a better understanding of the origins, chemical composition and aging of
oxygenated organic aerosols: case study of a Mediterranean industrialized environment,
Marseille, Atmos. Chem. Phys., 13, 7875-7894, doi:10.5194/acp-13-7875-2013, 2013a.
El Haddad, I., Marchand, N., D’Anna, B., Jaffrezo, J.-L., and Wortham, H.:
Functional group composition of organic aerosol from combustion emissions and
secondary processes at two contrasted urban environments, Atmos. Environ., 75, 308-
320, doi: 10.1016/j.atmosenv.2013.04.019, 2013b.
Ellis, S. R., Bruinen, A. L., and Heeren, R. M.: A critical evaluation of the current
state-of-the-art in quantitative imaging mass spectrometry,. Anal. Bioanal. Chem., 406,
1275-1289, doi: 10.1007/s00216-013-7478-9, 2014.
Elser, M., Bozzetti, C., El-Haddad, I., Maasikmets, M., Teinemaa, E., Richter, R.,
Wolf, R., Slowik, J. G., Baltensperger, U., and Prévôt, A. S. H.: Urban increments of
gaseous and aerosol pollutants and their sources using mobile aerosol mass
spectrometry measurements, Atmos. Chem. Phys., 16, 7117-7134, doi:10.5194/acp-16-
7117-2016 , 2016a.
Elser, M., Huang, R.-J., Wolf, R., Slowik, J. G., Wang, Q., Canonaco, F., Li, G.,
Bozzetti, C., Daellenbach, K. R., Huang, Y., Zhang, R., Li, Z., Cao, J., Baltensperger,
U., El-Haddad, I., and Prévôt, A. S. H.: New insights into PM2.5 chemical composition
and sources in two major cities in China during extreme haze events using aerosol mass
spectrometry, Atmos. Chem. Phys., 16, 3207-3225, doi:10.5194/acp-16-3207-2016,
2016b.
European Commission, Air quality standards:
http://ec.europa.eu/environment/air/quality/standards.htm, access: 2016/05/12.
Favez, O., El Haddad, I., Piot, C., Boréave, A., Abidi, E., Marchand, N., Jaffrezo, J.-
L., Besombes, J.-L., Personnaz, M.-B., Sciare, J., Wortham, H., George, C., and
![Page 155: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/155.jpg)
Bibliography
138
D’Anna, B.: Inter-comparison of source apportionment models for the estimation of
wood burning aerosols during wintertime in an Alpine city (Grenoble, France), Atmos.
Chem. Phys., 10, 5295-5314, doi:10.5194/acp-10-5295-2010, 2010.
Federal Office for the Environment (FOEN): Immissionsgrenzwerte der
Luftreinhalte-Verordnung (LRV), 2010.
Finlayson-Pitts, B. J., and Pitts, J. N. J.: Chemistry of the Upper and Lower
Atmosphere, 39, 327-328,2000.
Freutel, F., Schneider, J., Drewnick, F., von derWeiden-Reinmüller, S.-L., Crippa,
M., Prévôt, A. S. H., Baltensperger, U., Poulain, L., Wiedensohler, A., Sciare, J., Sarda-
Estève, R., Burkhart, J. F., Eckhardt, S., Stohl, A., Gros, V., Colomb, A., Michoud, V.,
Doussin, J. F., Borbon, A., Haeffelin, M., Morille, Y., Beekmann, M., and Borrmann,
S.: Aerosol particle measurements at three stationary sites in the megacity of Paris
during summer 2009: meteorology and air mass origin dominate aerosol particle
composition and size distribution, Atmos. Chem. Phys., 13, 933-959, doi:10.5194/acp-
13-933-2013, 2013.
Fröhlich, R., Cubison, M. J., Slowik, J. G., Bukowiecki, N., Prévôt, A. S. H.,
Baltensperger, U., Schneider, J., Kimmel, J. R., Gonin, M., Rohner, U., Worsnop, D. R.,
and Jayne, J. T.: The ToF-ACSM: a portable aerosol chemical speciation monitor with
TOFMS detection, Atmos. Meas. Tech., 6, 3225-3241, doi:10.5194/amt-6-3225-2013,
2013.
Fröhlich, R., Crenn, V., Setyan, A., Belis, C. A., Canonaco, F., Favez, O., Riffault,
V., Slowik, J. G., Aas, W., Aijälä, M., Alastuey, A., Artiñano, B., Bonnaire, N.,
Bozzetti, C., Bressi, M., Carbone, C., Coz, E., Croteau, P. L., Cubison, M. J., Esser-
Gietl, J. K., Green, D. C., Gros, V., Heikkinen, L., Herrmann, H., Jayne, J. T., Lunder,
C. R., Minguillón, M. C., Močnik, G., O’Dowd, C. D., Ovadnevaite, J., Petralia, E.,
Poulain, L., Priestman, M., Ripoll, A., Sarda-Estève, R., Wiedensohler, A.,
Baltensperger, U., Sciare, J., and Prévôt, A. S. H.: ACTRIS ACSM intercomparison -
Part 2: Intercomparison of ME-2 organic source apportionment results from 15
![Page 156: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/156.jpg)
139
individual, co-located aerosol mass spectrometers, Atmos. Meas. Tech., 8, 2555-2576,
doi:10.5194/amt-8-2555-2015, 2015.
Furey, A., Moriarty, M., Bane, V., Kinsella, B., and Lehane, M.: Ion suppression: A
critical review on causes, evaluation, prevention and applications, Talanta, 115, 104-
122, doi: 10.1016/j.talanta.2013.03.048, 2013.
Gentner, D. R., Jathar, S. H., Gordon, T. D., Bahreini, R., Day, D. A., El Haddad, I.,
Hayes, P. L., Pieber, S. M., Platt, S. M., de Gouw, J., Goldstein, A. H., Harley, R. A.,
Jimenez, J. L., Prévôt, A. S. H., and Robinson, A. L.: Review of urban secondary
organic aerosol formation from gasoline and diesel motor vehicle emissions, Environ.
Sci. Tech., 51, 1074-1093, doi: 10.1021/acs.est.6b04509, 2017.
Gilgen, H. and Ohmura, A.:: The Global Energy Balance Archive (GEBA). Bull.
Am. Met. Soc., 80,831-850, 10.1175/1520-0477(1999)080<0831:tgeba>2.0.co;2, 1999.
Goheen, S. C., Wahl, K. L., Campbell, J. A., and Hess, W. P.: Mass spectrometry of
low molecular mass solids by matrix-assisted laser desorption/ionization, J. mass.
Spectr., 32, 820,828, doi:10.1002/(sici)1096-9888(199708)32:8<820::aid-
jms538>3.0.co;2-q, 1997
Goldstein, A. H., and Galbally, I. E.: Known and unexplored organic constituents in
the earth’s atmosphere, Environ. Sci. Technol., 41, 1514-1521, doi: 10.1021/es072476p,
2007.
Gordon, H., Sengupta, K., Rap, A., Duplissy, J., Frege, C., Williamson, C., Heinritzi,
M., Simon, M., Yan, C., Almeida, J., Tröstl, J., Nieminen, N., Ortega, I. K., Wagner, R.,
Dunne, E. M., Adamov, A., Amorim, A., Bernhammer, A.-K., Bianchi, F.,
Breitenlechner, M., Sophia Brilke, S., Chen, X., Craven, J. S., Dias, A., Ehrhart, S.,
Fischer, L., Flagan, R. C., Franchin, A., Fuchs, C., Guida, R., Hakala, J., Hoyle, C. R.,
Jokinen, T., Junninen, H., Kangasluoma, J., Kim, J., Kirkby, J., Krapf, M., Kürten, A.,
Laaksonen, A., Lehtipalo, K., Makhmutov, V., Mathot, S., Molteni, U., Monks, S. A.,
Onnela, A., Peräkylä, O., Piel, F., Petäjä, T., Praplan, A. P., Pringle, K. J., Richards, N.
![Page 157: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/157.jpg)
Bibliography
140
A. D., Rissanen, M. P., Rondo, L., Sarnela, N., Schobesberger, S., Scott, C. E., Seinfeld,
J. H., Sharma, S., Sipilä, M., Steiner, G., Stozhkov, Y., Stratmann, F., Tomé, A.,
Virtanen, A., Vogel, A. L., Wagner, A. C., Wagner, P. E., Weingartner, E., Wimmer,
D., Winkler, P. M., Ye, P., Zhang, X., Hansel, A., Dommen, J., Donahue, N. M.,
Worsnop, D. R., Baltensperger, U., Kulmala, M., Curtius, J., and Carslaw, K. S.:
Reduced anthropogenic aerosol radiative forcing caused by biogenic new particle
formation, Proc. Natl. Acad. SCi U.S., 113, 12053-12058,
doi:10.1073/pnas.1602360113, 2016.
Griffiths, A. D., Conen, F., Weingartner, E., Zimmermann, L., Chambers, S. D.,
Williams, A. G., and Steinbacher, M.: Surfaceto-mountaintop transport characterised by
radon observations at the Jungfraujoch, Atmos. Chem. Phys., 14, 12763-12779,
doi:10.5194/acp-14-12763-2014, 2014.
Griggs, D. J. and Noguer, M.: Climate Change 2001: the Scientific Basis,
Contribution of working group I to the third assessment report of the Intergovernmental
Panel on Climate Change,Weather, 57, 267-269, 2002.
Guenther, A., Karl, T., Harley, P., Wiedinmyer, C., Palmer, P. I., and Geron, C.:
Estimates of global terrestrial isoprene emissions using MEGAN (Model of Emissions
of Gases and Aerosols from Nature), Atmos. Chem. Phys., 6, 3181-3210,
doi:10.5194/acp-6-3181-2006, 2006.
Haefliger, O. P., Bucheli, T. D., Zenobi, R.: Laser mass spectrometry analysis of
organic atmospheric aerosols. 1. Characterization of emission sources, Environ. Sci.
Technol., 34, 2178-2183, doi:10.1021/es991122y, 2000a.
Haefliger, O. P., Bucheli, T. D., Zenobi, R.: Laser mass spectrometry analysis of
organic atmospheric aerosols. 2. Elucidation of urban air pollution processes, Environ
Sci. Technol., 34, 2184-2189, doi:10.1021/es991123q: 2000b.
![Page 158: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/158.jpg)
141
Hamad, S. H., Schauer, J. J., Jongbae, H., and Kadhim, A. K. H.: Source
apportionment of PM2.5 carbonaceous aerosol in Baghdad, Iraq, Atmos. Res., 156, 80-
90, doi: 10.1016/j.atmosres.2014.12.017, 2015.
Hansen, A. D. A., Rosen, H., and Novakov, T.: The Aethalometer - an instrument for
the real-time measurement of opticalabsorption by aerosol-particles, Sci. Total Environ.,
36, 191-196, doi:10.1016/0048-9697(84)90265-1, 1984.
Healy, R. M., Sciare, J., Poulain, L., Crippa, M. Wiedensohler, A., Prevot, A. S. H.,
Baltensperger, U., Sarda-Estève, R., McGuire, M. L., Jeong, C.-H., McGillicuddy, E.,
O’Connor, I. P., Sodeau, J. R., Evans, G. J., Wenger, J. C.: Quantitative determination
of carbonaceous particle mixing state in Paris using single-particle mass spectrometer
ande aerosol mass spectrometer measurements, Atmos. Chem. Phys., 13, 9479-9496,
doi:10.5194/acp-13-9479-2013, 2013.
Herich, H., Hueglin, C., and Buchmann, B.: A 2.5 year’s source apportionment study
of black carbon from wood burning and fossil fuel combustion at urban and rural sites
in Switzerland, Atmos. Meas. Tech., 4, 1409-1420, doi:10.5194/amt-4-1409-2011,
2011.
Herich, H., Gianini, M. F. D., Piot, C., Močnik, G., Jaffrezo, J.-L., Besombes, J.-L.,
Prévôt, A. S. H., Hueglin, C.: Overview of the impact of wood burning emissions on
carbonaceous aerosols and PM in large parts of the Alpine region, Atmos. Env., 89, 64-
75, doi:10.1016/j.atmosenv.2014.02.008, 2014
Heringa, M. F., DeCarlo, P. F., Chirico, R., Tritscher, T., Dommen, J.,Weingartner,
E., Richter, R.,Wehrle, G., Prévôt, A. S. H., and Baltensperger, U.: Investigations of
primary and secondary particulate matter of different wood combustion appliances with
a high-resolution time-of-flight aerosol mass spectrometer, Atmos. Chem. Phys., 11,
5945-5957, doi:10.5194/acp-11-5945-2011, 2011.
Hoffmann, T., Huang, R.-J., and Kalberer, M.: Atmospheric analytical chemistry,
Anal. Chem., 83, 4649-4664, doi:10.1021/ac2010718, 2011.
![Page 159: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/159.jpg)
Bibliography
142
Holzke, C., Hoffmann, T., Jaeger, L., Koppmann, R., and Zimmer, W.: Diurnal and
seasonal variation of monoterpene and sesquiterpene emissions from Scots pine (Pinus
sylvestris L.), Atmos. Environ., 40, 3174-3185,
doi:org/10.1016/j.atmosenv.2006.01.039, 2006.
Huang, R.-J., Zhang, Y., Bozzetti, C., Ho, K.-F., Cao, J.-J., Han, Y., Daellenbach, K.
R., Slowik, J. G., Platt, S. M., Canonaco, F., Zotter, P., Wolf, R., Pieber, S. M., Bruns,
E. A., Crippa, M., Ciarelli, G., Piazzalunga, A., Schwikowski, M., Abbaszade, G.,
Schnelle-Kreis, J., Zimmermann, R., An, Z., Szidat, S., Baltensperger, U., El Haddad,
I., and Prévôt, A. S. H.: High secondary aerosol contribution to particulate pollution
during haze events in China, Nature, 514, 218-222, doi:10.1038/nature13774, 2014.
Iinuma, Y., Engling, G., Puxbaum, H., and Herrmann, H.: A highly resolved anion-
exchange chromatographic method for determination of saccharidic tracers for biomass
combustion and primary bio-particles in atmospheric aerosol, Atmos. Environ., 43,
1367-1371, 2009.
IPCC, 2013: Climate Change 2013:The Physical Science Basis. Contribution of
Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on
Climate Change; Stocker, T. F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S. K.,
Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P. M., Eds., Cambridge University
Press: Cambridge, United Kingdom and New York, NY, USA, 2013.
Jaffrezo, J. L., Calas, T., and Bouchet, M.: Carboxylic acids measurements with ionic
chromatography, Atmos. Environ., 32, 2705-2708, doi:10.1016/S1352-2310(98)00026-
0, 1998.
Jimenez, J. L., Canagaratna, M. R., Donahue, N. M., Prévôt, A. S. H., Zhang, Q.,
Kroll, J. H., DeCarlo, P. F., Allan, J. D., Coe, H., Ng, N. L., Aiken, A. C., Docherty, K.
S., Ulbrich, I. M., Grieshop, A. P., Robinson, A. L., Duplissy, J., Smith, J. D., Wilson,
K. R., Lanz, V. A., Hueglin, C., Sun, Y. L., Tian, J., Laaksonen, A., Raatikainen, T.,
Rautiainen, J., Vaattovaara, P., Ehn, M., Kulmala, M., Tomlinson, J. M., Collins, D. R.,
![Page 160: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/160.jpg)
143
Cubison, M. J., Dunlea, E. J., Huffman, J. A., Onasch, T. B., Alfarra, M. R., Williams,
P. I., Bower, K., Kondo, Y., Schneider, J., Drewnick, F., Borrmann, S., Weimer, S.,
Demerjian, K., Salcedo, D., Cottrell, L., Griffin, R., Takami, A., Miyoshi, T.,
Hatakeyama, S., Shimono, A., Sun, J. Y., Zhang, Y. M., Dzepina, K., Kimmel, J. R.,
Sueper, D., Jayne, J. T., Herndon, S. C., Trimborn, A. M., Williams, L. R., Wood, E. C.,
Middlebrook, A. M., Kolb, C. E., Baltensperger, U., and Worsnop, D. R.: Evolution of
organic aerosols in the atmosphere, Science, 326, 1525-1529,
doi:10.1126/science.1180353, 2009.
Jimenez, J. L., Canagaratna, M. R., Drewnick, F., Allan, J. D., Alfarra, M. R.,
Middlebrook, A. M., Slowik, J. G., Zhang, Q., Coe, H., Jayne, J. T., and Worsnop, D.
R.: Comment on “The effects of molecular weight and thermal decomposition on the
sensitivity of a thermal desorption aerosol mass spectrometer, Aerosol Sci. Technol.,
50:9, i-xv, doi:10.1080/02786826.2016.1205728, 2016.
Kalberer, M., Paulsen, D., Sax, M., Steinbacher, M., Dommen, J., Prevot, A. S. H.,
Fisseha, R., Weingartner, E., Frankevich, Zenobi, R., Baltensperger, U.: Identification
of polymers as major components of atmospheric organic aerosols, Science, 303, 1659-
1662, doi:10.1126/science.1092185, 2004.
Kelly, F. J., and Fussell, J. C.: Size, source and chemical composition as
determinants of toxicity attributable to ambient particulate matter, Atmos. Env., 60,
504-526, doi:10.1016/j.atmosenv.2012.06.039, 2012.
Kirchstetter, T., Harley, R., Kreisberg, N., Stolzenburg, M., and Hering, S.: On-road
measurement of fine particle and nitrogen oxide emissions from light- and heavy-duty
motor vehicles, Atmos. Environ., 33(18), 2955-2968, , 1999.
Klein, F., Platt, S. M., Farren, N. J., Detournay, A., Bruns, E. A., Bozzetti, C.,
Daellenbach, K. R., Kilic, D., Kumar, N. K., Pieber, S. M., Slowik, J. G., Temime-
Roussel, B., Marchand, N., Hamilton, J. F., Baltensperger, U., Prévôt, A. S. H., and El
Haddad, I.: Characterization of gas-phase organics using proton transfer reaction time-
![Page 161: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/161.jpg)
Bibliography
144
of-flight mass spectrometry: cooking emissions, Environ. Sci. Tech., 50, 1243-1250,
doi: 10.1021/acs.est.5b04618, 2016a.
Klein, F., Farren, N. J., Bozzetti, C., Daellenach, K. R., Kilic, D., Kumar, N. K.,
Pieber, S. M., Slowik, J. G., Tuthill, R. N., Hamilton, J. F., Baltensperger, U., Prévôt, A.
S. H., El and Haddad, I.: Indoor terpene emissions from cooking with herbs and pepper
and their secondary organic aerosol production potential, Sci. Rep., 6, 36623,
doi:10.1038/srep36623, 2016b.
Kleindienst, T. E. Jaoui, M., Lewandowski, M., Offenberg, J. H., Lewis, C. W.,
Bhave, P. V., Edney, E. O.: Estimates of the contributions of biogenic and
anthropogenic hydrocarbons to secondary organic aerosol at a southern US location,
Atmos. Env., 37, 8288-8300, doi:10.1016/j.atmosenv.2007.06.045, 2007.
Knochenmuss, R., Stortfelder, A., Breuker, K., Zenobi, R.: Secondary ion-molecule
reactions in matrix-assisted laser desorption/ionization, J. Mass. Spectrom, 35, 1237-
1245, doi: 10.1002/1096-9888(200011)35:11<1237::AID-JMS74>3.0.CO;2-O, 2000.
Knochenmuss, R.: A quantitative model of ultraviolet matrix-assisted laser
desorption/ionization including analyte ion generation, Anal. Chem., 75, 2199-2207,
doi:10.1021/ac034032r, 2003.
Knochenmuss, R.: A quantitative model of ultraviolet matrix-assisted laser
desorption/ionization, J. Mass. Spectrom., 37, 867-877, doi: 10.1002/jms.349, 2002.
Knochenmuss, R.: Ion formation mechanism in UV-MALDI, Analyst, 131(9), 966-
86, doi: 10.1039/b605646f, 2006.
Knochenmuss, R.: The coupled chemical and physical dynamics model of MALDI,
Annu. Rev. Anal. Chem., 9, 365-385, doi:10.1146/annurev-anchem-071015-041750,
2016
![Page 162: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/162.jpg)
145
Kourtchev, I., Fuller, S., Aalto, J., Ruuskanen, T. M., McLeod, M. W., Maenhau, W.,
Jones, R., Kulmala, M., Kalberer, M.: Molecular composition of boreal forest aerosol
from Hyytiälä, Finland, using ultrahigh resolution mass spectrometry, Environ. Sci.
Technol., 2013, 47, 4069-4079, doi:0.1021/es3051636, 2013.
Kroll, J. H., Donahue, N. M., Jimenez, J. L., Kessler, S. H., Canagaratna, M. R.,
Wilson, K. R., Altieri, K. E., Mazzoleni, L. R., Wozniak, A. S., Blum, H., Mysak, E. R.,
Smith, J. D., Kolb, C. E., and Worsnop, D. R.: Carbon oxidation state as a metric for
describing the chemistry of atmospheric organic aerosol, Nat. Chem., 3, 133-139,
doi:10.1038/NCHEM.948, 2011
Kunit, M. and Puxbaum, H.: Enzymatic determination of the cellulose content of
atmospheric aerosols, Atmos. Environ., 30, 1233-1236, doi:10.1016/1352-
2310(95)00429-7, 1996.
Laden, F., Schwartz, J., Spelzer, F. E., and Dockery, D. W.: Reduction in fine
particulate air pollution and mortality, Am. J. Respir. Crit. Care Med., 173, 667-672,
doi:10.1164/rccm.200503-443OC, 2006.
Lamberg, H., Nuutinen, K., Tissari, J., Ruusunen, J., Yli-Pirilä, P., Sippula., O.,
Tapanainen, M., Jalava, P., Makkonen, U., Teinilä, K., Saarnio, K., Hillamo, R.,
Hirvonen, M.-R., Jininiemi, J.: Physicochemical characterization of fine particles from
small-scale wood combustion, Atmos. Environ., 45, 7635-7643,
doi:10.1016/j.atmosenv.2011.02.072, 2011.
Lanz, V. A., Alfarra, M. R., Baltensperger, U., Buchmann, B., Hueglin, C., and
Prévôt, A. S. H.: Source apportionment of submicron organic aerosols at an urban site
by factor analytical modelling of aerosol mass spectra, Atmos. Chem. Phys., 7, 1503-
1522, doi:10.5194/acp-7-1503-2007, 2007.
Lanz, V. A., Alfarra, M. R., Baltensperger, U., Buchmann, B., Hueglin, C., Szidat,
S., Wehrli, M. N., Wacker, L., Weimer, S., Caseiro, A., Puxbaum, H., and Prévôt, A. S.
H.: Source attribution of submicron organic aerosols during wintertime inversions by
![Page 163: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/163.jpg)
Bibliography
146
advanced factor analysis of aerosol mass spectra, Environ. Sci. Technol., 42, 214-220,
doi:10.1021/es0707207, 2008.
Lanz, V. A., Prévôt, A. S. H., Alfarra, M. R., Weimer, S., Mohr, C., DeCarlo, P. F.,
Gianini, M. F. D., Hueglin, C., Schneider, J., Favez, O., D'Anna, B., George, C., and
Baltensperger, U.: Characterization of aerosol chemical composition with aerosol mass
spectrometry in Central Europe: an overview, Atmos. Chem. Phys., 10, 10453-10471,
doi:10.5194/acp-10-10453-2010, 2010.
Leaitch, W. R., Macdonald, A. M., Brickell, P. C., Liggio, J., Sjostedt, S. J.,
Vlasenko, A., Bottenheim, J. W., Huang, L., Li, S.-M., Liu, P. S. K., Toom-Sauntry, D.,
Hayden, K. A., Sharma, S., Shantz, N. C., Wiebe H. A., Zhang, W., Abbatt, J. P. D.,
Slowik, J. G., Chang, R., Y.-W., Russell, L. M., Schwartz, R. E., Takahama, S., Jayne,
J. T., and Ng, N. L.: Temperature response of the submicron organic aerosol from
temperate forests, Atmos. Environ., 45, 6696-6704,
doi:10.1016/j.atmosenv.2011.08.047, 2011.
Lee, A. K. Y., Herckes, P., Leaitch, W. R., Macdonald, A. M., and Abbatt, J. P. D.:
Aqueous OH oxidation of ambient organic aerosol and cloud water organics: formation
of highly oxidized products, Geophys. Res. Lett., 38, L11805,
doi:10.1029/2011GL047439, 2011.
Lelieveld, J., Evans, J. S., Fnais, M., Giannadaki, D., and Pozzer, A.: The
contribution of outdoor air pollution sources to premature mortality on a global scale,
Nature, 525, 367-371, doi:10.1038/nature15371, 2015.
Lim, S. S., Vos, T., Flaxman, A. D., Danaei, G., Shibuya, K., Adair-Rohani, H.,
AlMazroa, M. A., Amann, M., Anderson, H. R., Andrews, K. G., Aryee, M., Atkinson,
C., Bacchus, L. J., Bahalim, A. N., Balakrishnan, K., Balmes, J., Barker-Collo, S.,
Baxter, A., Bell, M. L., Blore, J. D., Blyth, F., Bonner, C., Borges, G., Bourne, R.,
Boussinesq, M., Brauer, M., Brooks, P., Bruce, N. G., Brunekreef, B., Bryan-Hancock,
C., Bucello, C., Buchbinder, R., Bull, F., Burnett, R. T., Byers, T. E., Calabria, B.,
Carapetis, J., Carnahan, E., Chafe, Z., Charlson, F., Chen, H., Chen, J. S., Cheng, A. T.-
![Page 164: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/164.jpg)
147
A., Child, J. C., Cohen, A., Colson, K. E., Cowie, B. C., Darby, S., Darling, S., Davis,
A., Degenhardt, L., Dentener, F., Jarlais, D. C. D., Devries, K., Dherani, M., Ding, E.
L., Dorsey, E. R., Driscoll, T., Edmond, K., Ali, S. E., Engell, R. E., Erwin, P. J.,
Fahimi, S., Falder, G., Farzadfar, F., Ferrari, A., Finucane, M. M., Flaxman, S., Fowkes,
F. G. R., Freedman, G., Freeman, M. K., Gakidou, E., Ghosh, S., Giovannucci, E.,
Gmel, G., Graham, K., Grainger, R., Grant, B., Gunnell, D., Gutierrez, H. R., Hall, W.,
Hoek, H. W., Hogan, A., III, H. D. H., Hoy, D., Hu, H., Hubbell, B. J., Hutchings, S. J.,
Ibeanusi, S. E., Jacklyn, G. L., Jasrasaria, R., Jonas, J. B., Kan, H., Kanis, J. A.,
Kassebaum, N., Kawakami, N., Khang, Y.-H., Khatibzadeh, S., Khoo, J.-P., Kok, C.,
Laden, F., Lalloo, R., Lan, Q., Lathlean, T., Leasher, J. L., Leigh, J., Li, Y., Lin, J. K.,
Lipshultz, S. E., London, S., Lozano, R., Lu, Y., Mak, J., Malekzadeh, R., Mallinger,
L., Marcenes, W., March, L., Marks, R., Martin, R., McGale, P., McGrath, J., Mehta, S.,
Memish, Z. A., Mensah, G. A., Merriman, T. R., Micha, R., Michaud, C., Mishra, V.,
Hanafiah, K. M., Mokdad, A. A., Morawska, L., Mozaffarian, D., Murphy, T., Naghavi,
M., Neal, B., Nelson, P. K., Nolla, J. M., Norman, R., Olives, C., Omer, S. B., Orchard,
J., Osborne, R., Ostro, B., Page, A., Pandey, K. D., Parry, C. D., Passmore, E., Patra, J.,
Pearce, N., Pelizzari, P. M., Petzold, M., Phillips, M. R., Pope, D., Pope, C. A. 3rd,
Powles, J., Rao, M., Razavi, H., Rehfuess, E. A., Rehm, J. T., Ritz, B., Rivara, F. P.,
Roberts, T., Robinson, C., Rodriguez-Portales, J. A., Romieu, I., Room, R., Rosenfeld,
L. C., Roy, A., Rushton, L., Salomon, J. A., Sampson, U., Sanchez-Riera, L., Sanman,
E., Sapkota, A., Seedat, S., Shi, P., Shield, K., Shivakoti, R., Singh, G. M., Sleet, D. A.,
Smith, E., Smith, K. R., Stapelberg, N. J., Steenland, K., Stöckl, H., Stovner, L. J.,
Straif, K., Straney, L., Thurston, G. D., Tran, J. H., Van Dingenen, R., van Donkelaar,
A., Veerman, J. L., Vijayakumar, L., Weintraub, R., Weissman, M. M., White, R. A.,
Whiteford, H., Wiersma, S. T., Wilkinson, J. D., Williams, H. C., Williams, W.,
Wilson, N., Woolf, A. D., Yip, P., Zielinski, J. M., Lopez, A. D., Murray, C. J., Ezzati,
M., AlMazroa, M. A., and Memish, Z. A.: A comparative risk assessment of burden of
disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions,
1990-2010: a systematic analysis for the Global Burden of Disease Study 2010, The
Lancet, 380, 2224-2260, doi: 10.1016/S0140-6736(12)61766-8, 2012
Masson-Delmotte, V., Schulz, M., Abe-Ouchi, A., Beer, J., Ganopolski, A.,
González Rouco, J.F., Jansen, E., Lambeck, K., Luterbacher, J., Naish, T., Osborn, T.,
![Page 165: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/165.jpg)
Bibliography
148
Otto-Bliesner, B., Quinn, T., Ramesh, R., Rojas, M., Shao, X., and Timmermann, A.:
Information from Paleoclimate Archives. in: Climate Change 2013: The Physical
Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change edited by Stocker, T.F., Qin, D., Plattner,
G.-K., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex V., and Midgley,
P.M., Cambridge University Press, Cambridge, United Kingdom and New York, NY,
USA, 2013.
Maynard, A., and Kuempel, E.: Airborne nanostructured particles and occupational
health, J. Nanopart. Res., 7, 587-614, doi:10.1007/s11051-005-6770-9, 2005.
Mihara, T. and Mochida, M.: Characterization of solvent extractable organics in
urban aerosols based on mass spectrum analysis and hygroscopic growth measurement,
Environ. Sci. Technol., 45, 9168-9174, doi:10.1021/es201271w, 2011.
Minguillón, M. C., Perron, N., Querol, X., Szidat, S., Fahrni, S. M., Alastuey, A.,
Jimenez, J. L., Mohr, C., Ortega, A. M., Day, D. A., Lanz, V. A., Wacker, L., Reche,
C., Cusack, M., Amato, F., Kiss, G., Hoffer, A., Decesari, S., Moretti, F., Hillamo, R.,
Teinilä, K., Seco, R., Peñuelas, J., Metzger, A., Schallhart, S., Müller, M., Hansel, A.,
Burkhart, J. F., Baltensperger, U., and Prévôt, A. S. H.: Fossil versus contemporary
sources of fine elemental and organic carbonaceous particulate matter during the
DAURE campaign in Northeast Spain, Atmos. Chem. Phys., 11, 12067-12084,
doi:10.5194/acp-11-12067-2011, 2011.
Mohr, C., Richter, R., DeCarlo, P. F., Prévôt, A. S. H., and Baltensperger, U.: Spatial
variation of chemical composition and sources of submicron aerosol in Zürich during
wintertime using mobile aerosol mass spectrometer data, Atmos. Chem. Phys., 11,
7465-7482, doi:10.5194/acp-11-7465-2011, 2011.
Mohr, C., DeCarlo, P. F., Heringa, M. F., Chirico, R., Slowik, J. G., Richter, R.,
Reche, C., Alastuey, A., Querol, X., Seco, R., Penuelas, J., Jimenez, J. L., Crippa, M.,
Zimmermann, R., Baltensperger, U., and Prevot, A. S. H.: Identification and
quantification of organic aerosol from cooking and other sources in Barcelona using
![Page 166: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/166.jpg)
149
aerosol mass spectrometer data, Atmos. Chem. Phys., 12, 1649-1665, doi:10.5194/acp-
12-1649-2012, 2012.
Murphy, D. M.: The design of single particle laser mass spectrometers, Mass Spectr.
Rev., 26, 150-165, doi:10.1002/mas.20113, 2007.
Myhre, G., Shindell, D., Bréon, F.-M., Collins, W., Fuglestvedt, J., Huang, J., Koch,
D., Lamarque, J.-F., Lee, D., Mendoza, B., Nakajima, T., Robock, A., Stephens, G.,
Takemura, T., and Zhang, H.: Anthropogenic and Natural Radiative Forcing. in:
Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to
the Fifth Assessment Report of the Intergovernmental Panel on Climate Change edited
by Stocker, T.F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S.K., Boschung, J., Nauels,
A., Xia, Y., Bex, V., and Midgley, P.M., pp. 659-740, Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA, 2013
Ng, N. L., Canagaratna, M. R., Zhang, Q., Jimenez, J. L., Tian, J., Ulbrich, I. M.,
Kroll, J. H., Docherty, K. S., Chhabra, P. S., Bahreini, R., Murphy, S. M., Seinfeld, J.
H., Hildebrandt, L., Donahue, N. M., DeCarlo, P. F., Lanz, V. A., Prévôt, A. S. H.,
Dinar, E., Rudich, Y., and Worsnop, D. R.: Organic aerosol components observed in
Northern Hemispheric datasets from Aerosol Mass Spectrometry, Atmos. Chem. Phys.,
10, 4625-4641, doi:10.5194/acp-10-4625-2010, 2010.
Ng, N. L., Canagaratna, M R., Jimenez, J. L., Chhabra, P. S., Seinfeld, J. H., and
Worsnop, D. R.: Changes in organic aerosol composition with aging inferred from
aerosol mass spectra, Atmos. Chem. Phys., 11, 6465-6474, doi: 10.5194/acp-11-6465-
2011, 2011a.
Ng, N. L., Canagaratna, M. R., Jimenez, J. L., Zhang, Q., Ulbrich, I. M., and
Worsnop, D. R.: Real-time methods for estimating organic component mass
concentrations from aerosol mass spectrometer data, Environ. Sci. Technol., 45, 910-
916, doi:10.1021/es102951k, 2011b.
![Page 167: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/167.jpg)
Bibliography
150
Ng, N. L., Herndon, S. C., Trimborn, A., Canagaratna, M. R., Croteau, P. L., Onasch,
T. B., Sueper, D., Worsnop, D. R., Zhang, Q., Sun, Y. L., and Jayne, J. T.: An aerosol
chemical speciation monitor (ACSM) for routine monitoring of the composition and
mass concentrations of ambient aerosol, Aerosol Sci. Tech., 45, 780-794,
doi:10.1080/02786826.2011.560211, 2011c.
Oderbolz, D. C., Aksoyoglu, S., Keller, J., Barmpadimos, I., Steinbrecher, R., Skjøth,
C. A., Plaß-Dülmer, C., and Prévôt, A. S. H.: A comprehensive emission inventory of
biogenic volatile organic compounds in Europe: improved seasonality and land-cover,
Atmos. Chem. Phys., 13, 1689-1712, doi:10.5194/acp-13-1689-2013, 2013.
Paatero, P. and Tapper, U.: Positive matrix factorization - a nonnegative factor model
with optimal utilization of error-estimates of data values, Environmetrics, 5, 111-126,
doi:10.1002/env.3170050203, 1994.
Paatero, P.: Least squares formulation of robust non-negative factor analysis,
Chemometr. Intell. Lab., 37, 23-35, doi: 10.1016/S0169-7439(96)00044-5, 1997.
Paatero, P.: 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., 8, 854-888, doi:10.2307/1390831, 1999.
Paatero, P.: User’s guide for positive matrix factorization programs PMF2 and
PMF3, 2004.
Paatero, P. and Hopke, P. K.: Rotational tools for factor analytic models, J.
Chemometr., 23, 91-100, doi:10.1002/cem.1197, 2009.
Paatero, P., Eberly, S., Brown, S. G., and Norris, G. A.: Methods for estimating
uncertainty in factor analytic solutions, Atmos. Meas. Tech., 7, 781-797,
doi:10.5194/amt-7-781-2014, doi:10.5194/amt-7-781-2014, 2014.
![Page 168: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/168.jpg)
151
Piazzalunga, A., Fermo, P., Bernardoni, V., Vecchi, R., Valli, G., and De Gregorio,
M. A.: A simplified method for levoglucosan quantification in wintertime atmospheric
particulate matter by high performance anion-exchange chromatography coupled with
pulsed amperometric detection, Intern. J. Environ. Anal. Chem., 90, 12, 934-947,
doi:10.1080/03067310903023619, 2010.
Piazzalunga, A., Bernardoni, V., Fermo, P., and Vecchi, R.: Optimisation of
analytical procedures for the quantification of ionic and carbonaceous fractions in the
atmospheric aerosol and applications to ambient samples, Anal., Bioanal. Chem., 56,
30-40, doi:10.1007/s00216-012-6433-5, 2013
Pieber, S. M., El Haddad, I., Slowik, J. G., Canagaratna, M. R., Jayne, J. T., Platt, S.
M., Bozzetti, C., Daellenbach, K. R., Fröhlich, R., Vlachou, A., Klein, F., Dommen, J.,
Miljevic, B., Jimenez, J. L., Worsnop, D. R., Baltensperger, U., and Prévôt, A. S. H.:
Inorganic salt interference on CO2+ in aerodyne AMS and ACSM organic aerosol
composition studies, Environ. Sci. Technol, 50, 10494-10503, doi:
10.1021/acs.est.6b01035,2016.
Platt, S. M., El Haddda, I., Pieber, S. M., Huang, R.-J., Zardini, A. A., Clairotte, M.,
Suarez-Bertoa, R., Barmet, P., Pfaffenberger, L., Wolf, R., Slowik, J. G., Fuller, S. J.,
Kalberer, M., Chirico, R., Dommen, J., Astorga, C., Zimmermann, R., Marchand, N.,
Hellebust, S., Temime-Roussel, B., Baltensperger, U., Prévôt, A. S. H.: two-stroke
scooters are a dominant source of air pollution in many cities, Nat. Comm., 5, 3749,
doi:10.1038/ncomms4749, 2014.
Pope, C. A., and Dockery, D. W.: Health effects of fine particulate air pollution:
Lines that connect, J. Air Waist Manage. Assoc., 56, 709-742,
doi:10.1080/10473289.2006.10464485, 2006.
Pope, D., III, C. A. P., Powles, J., Rao, M., Razavi, H., Rehfuess, E. A., Rehm, J. T.,
Ritz, B., Rivara, F. P., Roberts, T., Robinson, C., Rodriguez-Portales, J. A., Romieu, I.,
Room, R., Rosenfeld, L. C., Roy, A., Rushton, L., Salomon, J. A., Sampson, U.,
Sanchez-Riera, L., Sanman, E., Sapkota, A., Seedat, S., Shi, P., Shield, K., Shivakoti,
![Page 169: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/169.jpg)
Bibliography
152
R., Singh, G. M., Sleet, D. A., Smith, E., Smith, K. R., Stapelberg, N. J., Steenland, K.,
Stöckl, H., Stovner, L. J., Straif, K., Straney, L., Thurston, G. D., Tran, J. H., Dingenen,
R. V., van Donkelaar, A., Veerman, J. L., Vijayakumar, L., Weintraub, R., Weissman,
M. M., White, R. A., Whiteford, H., Wiersma, S. T., Wilkinson, J. D., Williams, H. C.,
Williams, W., Wilson, N., Woolf, A. D., Yip, P., Zielinski, J. M., Lopez, A. D., Murray,
C. J., and Ezzati, M.: A comparative risk assessment of burden of disease and injury
attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a
systematic analysis for the Global Burden of Disease Study 2010, The Lancet, 380,
2224 - 2260, doi:10.1016/S0140-6736(12)61766-8 2013.
Putaud, J. P., Van Dingenen, R., Baltensperger, U., Brüggemann, E., Charron, A.,
Facchini, M. C., Decessari, S., Fuzzi, S., Gehrig, R., Hansson, H.-C., Harrison, R. M.,
Jones, A. M., Laj, P., Lorbeer, G., Maenhaut, W., Mihalopoulos, N., Müller, K.,
Palmgren, F., Querol, X., Rodriguez, S., Schneider, J., Spindler, G., ten Brink, H.,
Tunved, P., Törseth, K., Wehner, B., Weingartner, E., Wiedensohler, A., Wahlin, P.,
Raes, F.: A European aerosol phenomology, Joint Research Center, EUR 20411, 2004.
Putaud, J. P., Van Dingenen, R., Alastuey, A., Bauer, H., Birmili, W., Cyrys, J.,
Flentje, H., Fuzzi, S., Gehrig, R., Hansson, H. C., Harrison, R. M., Herrmann, H.,
Hitzenberger, R., Hueglin, C., Jones, A. M., Kasper-Giebl, A., Kiss, G., Kousa, A.,
Kuhlbusch, T. A. J., Loeschau, G., Maenhaut, W., Molnar, A., Moreno, T., Pekkanen,
J., Perrino, C., Pitz, M., Puxbaum, H., Querol, X., Rodriguez, S., Salma, I., Schwarz, J.,
Smolik, J., Schneider, J., Spindler, G., ten Brink, H., Tursic, J., Viana, M.,
Wiedensohler, A., and Raes, F.: A European aerosol phenomenology-3: physical and
chemical characteristics of particulate matter from 60 rural, urban, and kerbside sites
across Europe, Atmos. Environ., 44, 1308-1320, doi:10.1016/j.atmosenv.2009.12.011,
2010.
Putaud, J. P., Van Dingenen, R., Alastuey, A., Bauer, H., Birmili, W., Cyrys, J.,
Flentje, H., Fuzzi, S., Gehrig, R., Hansson, H. C., Harrison, R. M., Herrmann, H.,
Hitzenberger, R., Hueglin, C., Jones, A. M., Kasper-Giebl, A., Kiss, G., Kousa, A.,
Kuhlbusch, T. A. J., Loeschau, G., Maenhaut, W., Molnar, A., Moreno, T., Pekkanen,
J., Perrino, C., Pitz, M., Puxbaum, H., Querol, X., Rodriguez, S., Salma, I., Schwarz, J.,
![Page 170: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/170.jpg)
153
Smolik, J., Schneider, J., Spindler, G., ten Brink, H., Tursic, J., Viana, M.,
Wiedensohler, A., and Raes, F.: A European aerosol phenomenology-3: physical and
chemical characteristics of particulate matter from 60 rural, urban, and kerbside sites
across Europe, Atmos. Environ., 44, 1308-1320, doi:10.1016/j.atmosenv.2009.12.011,
2010.
Rocke, D. M., and Lorenzato, S.: A two-component model for measurement error in
analytical chemistry, Technometrics, 37, 2, 176-184, doi:
10.1080/00401706.1995.10484302, 1995.
Samburova, V., R. Zenobi, Kalberer, M.: Characterization of high molecular weight
compounds in urban atmospheric particles Atmos. Chem. Phys., 5.8 2163-2170,
doi:10.5194/acp-5-2163-2005, 2005a.
Samburova, V.,Szidat, S., Hueglin, C., Fisseha, R., Baltensperger, U., Zenobi, R.,
Kalberer, M.: Seasonal variation of high-molecular-weight compounds in the water-
soluble fraction of organic urban aerosols, J. Geo. Res., 110, D23,
doi:10.1029/2005jd005910, 2005b.
Sanchez-Ochoa, A., Kasper-Giebl, A., Puxbaum, H., Gelencser, A., Legrand, M., and
Pio, C.: Concentration of atmospheric cellulose: A proxy for plant debris across a west-
east transect over Europe, J.Geo. Res., 112, D23, doi:10.1029/2006jd008180, 2007.
Sandradewi, J., Prévôt, A. S. H., Szidat, S., Perron, N., Alfarra, M. R, Lanz, V. A.,
Weingartner, E., and Baltensperger, U.: Using aerosol light absorption measurements
for the quantitative determination of wood burning and traffic emission contributions to
particulate matter, Environ. Sci. Technol., 42, 3316-3323, doi:10.1021/es702253m,
2008.
Schauer, J. J., Kleeman, M. J., Cass, G. R., and Simoneit, B. R. T.: Measurement of
emissions from air pollution sources. 3. C1-C29 organic compounds from fireplace
combustion of wood, Environ. Sci. Technol., 35, 1716-1728, doi: 10.1021/es001331e,
2001.
![Page 171: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/171.jpg)
Bibliography
154
Schauer, J. J., Kleeman, M. J., Cass, G. R., and Simoneit, B. R. T.: Measurement of
emissions from air pollution sources. 5. C-1-C-32 organic compounds from gasoline-
powered motor vehicles, Environ. Sci. Technol., 36, 1169-1180, doi:
10.1021/es0108077, 2002a.
Schauer, J. J., Kleeman, M. J., Cass, G. R., and Simoneit, B. R. T.: Measurement of
emissions from air pollution sources. 4. C-1-C-27 organic compounds from cooking
with seed oils, Environ. Sci. Technol., 36, 567-575, doi: 10.1021/es002053m, 2002b.
Schmale, J., Schneider, J., Nemitz, E., Tang, Y. S., Dragosits, U., Blackall, T. D.,
Trathan, P. N., Phillips, G. J., Sutton, M., and Braban, C. F.: Sub-Antarctic marine
aerosol: dominant contributions from biogenic sources, Atmos. Chem. Phys., 13, 8669-
8694, doi:10.5194/acp-13-8669-2013,2013.
Schurgers, G., Arneth, A., Holzinger, R., and Goldstein, A. H.: Process-based
modelling of biogenic monoterpene emissions combining production and release from
storage, Atmos. Chem. Phys., 9, 3409-3423, doi:10.5194/acp-9-3409-2009, 2009.
Seifried, H. E., Anderson, D. E., Fisher, E. I., Milner, J. A.: A review of the
interaction among dietary antioxidants and reactive oxygen species, J. Nutr. Biochem.,
18, 567-579, doi:10.1016/j.jnutbio.2006.10.007, 2006.
Seinfeld, J. H., and Pandis, S. N.: Atmospheric chemistry and physics: from air
pollution to climate change, 2nd ed., John Wiley and Sons, Inc., New York, 2006.
Simoneit, B. R. T., Kobayashi, M., Mochida, M., Kawamura, K., Lee, M., Lim, H.-
J., Turpin, B. J., Komazaki, Y.: Composition and major sources of organic compounds
of aerosol particulate matter sampled during the ACE-Asia campaign, J. Geophys. Res.,
109, D19S10, doi:10.1029/2004jd004598, 2004.
![Page 172: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/172.jpg)
155
Stanek, L. W., Sacks, J. D., Dutton, S. J., Dubois, J.-J. B.: Attributing health effects
to apportioned components and sources of particulate matte: An evaluation of collective
results, Atmos. Environ., 45, 5655-5663, doi:10.1016/j.atmosenv.2011.07.023, 2011.
Stark, H., Yatavelli, R. L. N., Thompson, S. L., Kimmel, J. R., Cubison, M. J.,
Chhabra, P. S., Canagaratna, M. R., Jayne, J. T., Worsnop, D. R., and Jimenez, J. L.:
Methods to extract molecular and bulk chemical information from series of complex
mass spectra with limited mass resolution, Int. J. Mass Spectrom., 389, 26-38,
doi:10.1016/j.ijms.2015.08.011, 2015.
Stern, D. I.: Global sulfur emissions from 1850 to 2000, Chemosphere, 58, 163-175,
doi: 10.1016/j.chemosphere.2004.08.022, 2005.
Sun, Y., Zhang, Q., Zheng, M., Ding, X., Edgerton, E. S., and Wang, X.:
Characterization and source apportionment of watersoluble organic matter in
atmospheric fine particles (PM2:5) with high-resolution aerosol mass spectrometry and
GC-MS, Environ. Sci. Technol., 45, 4854-4861, doi:10.1021/es200162h, 2011.
Szidat, S., Jenk, T. M., Synal, H.-A., Kalberer, M., Wacker, L., Hajdas, I., Kasper-
Giebl, A., Baltensperger, U.: Contributions of fossil fuel, biomass-burning, and biogenic
emissions to carbonaceous aerosols in Zürich as traced by 14C, J. Geophys. Res.-
Atmos., 111, D07206, 10.1029/2005JD006590, 2006.
Szidat, S., Salazar, G. A., Battaglia, M., Wacker, L., Synal, H.-A., Vogel, E., and
Türler, A.: 14C analyses and sample preparation at the new Bern Laboratory for the
Analyses of Radiocarbon with AMS (LARA), Radiocarbon, 56, No. 2, 561-566,
doi:10.2458/56.17457, 2014.
Trufelli, H., Palma, P., Famiglini, G., and Cappiello, A.: An overview of matrix
effects in liquid chromatography-mass spectrometry, Mass. Spectrom, Rev., 30, 491-
509, doi:10.1002/mas.20298, 2011.
![Page 173: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/173.jpg)
Bibliography
156
Ulbrich, I. M., Canagaratna, M. R., Zhang, Q., Worsnop, D. R., and Jimenez, J. L.:
Interpretation of organic components from Positive Matrix Factorization of aerosol
mass spectrometric data, Atmos. Chem. Phys., 9, 2891-2918, doi:10.5194/acp-9-2891-
2009, 2009.
US Environmental Protection Agency, National Ambient Air Quality Standards
(NAAQS): https://www.epa.gov/criteria-air-pollutants/naaqs-table, access: 2016/12/05.
Verma, V., Fang, T., Xu, L., Peltier, R. E., Russell, A. G., Ng, N. L., and Weber, R.
J.: Organic aerosols associated with the generation of reactive oxygen species (ROS) by
water-soluble PM2.5, Environ. Sci. Technol., 49, 4646-4656, doi:10.1021/es505577w,
2015.
Vlachou, A., Daellenbach, K. R., Bozzetti, C., Szidat, S., Salazar, G., Agrios, K.,
Jaffrezo, J. L., Baltensperger, U., El Haddad, I., and Prévôt, A. S. H.: Radiocarbon
based source apportionment of carbonaceous aerosols in Magadino, Switzerland, in
preparation, 2017.
Visser, S., Slowik, J. G., Furger, M., Zotter, P., Bukowiecki, N., Canonaco, F.,
Flechsig, U., Appel, K., Green, D. C., Tremper, A. H., Young, D. E., Williams, P. I.,
Allan, J. D., Coe, H., Williams, L. R., Mohr, C., Xu, L., Ng, N. L., Nemitz, E., Barlow,
J. F., Halios, C. H., Fleming, Z. L., Baltensperger, U., and Prévôt, A. S. H.: Advanced
source apportionment of size-resolved trace elements at multiple sites in London during
winter, Atmos. Chem. Phys., 15, 11291-11309, doi:10.5194/acp-15-11291-2015, 2015.
Vlachou, A., Daellenbach, K. R., Bozzetti, C., Chazeau, B., Salazar, G. A., Szidat,
S., Jaffrezo, J.-L., Hueglin, C., Baltensperger, U., El Haddad, I., and Prévôt, A. S. H.:
Advanced source apportionment of carbonaceous aerosols by coupling offline AMS and
radiocarbon size segregated measurements over a nearly two-year period, Atmos.
Chem. Phys. Discuss., doi:10.5194/acp-2017-1102, in review, 2017.
Waked, A., Favez, O., Alleman, L. Y., Piot, C., Petit, J.-E., Delaunay, T., Verlinden,
E., Golly, B., Besombes, J.-L., Jaffrezo, J.-L., and Leoz-Garziandia, E.: Source
![Page 174: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/174.jpg)
157
apportionment of PM10 in a north-western Europe regional urban background site
(Lens, France) using positive matrix factorization and including primary biogenic
emissions, Atmos. Chem. Phys., 24, 3325-3346, doi:acp-14-3325-2014, 2014.
Watson, J. G., Robinson, N. F. Fujita, E. M., Chow, J. C., Pace, T. G. Lewis, C.,
Coulter, T.: CMB8 applications and validation protocol for PM2.5 and VOCs, US EPA,
USA, 1998.
Wilks, D. S.: Statistical methods in the atmospheric sciences, second edition, edited
by Dmowska, R., Hartmann, D., and Rossby, H. T., Academic Press, 2006.
Williams, L. R., Gonzalez, L. A., Peck, J., Trimborn, D., McInnis, J., Farrar, M. R.,
Moore, K. D., Jayne, J. T., Robinson, W. A., Lewis, D. K., Onasch, T. B., Canagaratna,
M. R., Trimborn, A., Timko, M. T., Magoon, G., Deng, R., Tang, D., de la Rosa
Blanco, E., Prévôt, A. S. H., Smith, K. A., and Worsnop, D. R.: Characterization of an
aerodynamic lens for transmitting particles greater than 1 micrometer in diameter into
the Aerodyne aerosol mass spectrometer, Atmos. Meas. Tech., 6, 3271-3280,
doi:10.5194/amt-6-3271-2013, 2013.
Wilson, M. D., Rocke, D. M., Durbin, B., and Kahn, H. D.: Detection limits and
goodness-of-fit measures for the two-component model of chemical analytical error,
Anal. Chim. Acta., 509, 197-208, doi:10.1016/j.aca.2003.12.047, 2004.
World Health Organization (WHO): WHO air quality guidelines for particulate
matter, ozone , nitrogen dioxide and sulfur dioxide. Global update 2005. Summary of
risk assessment, WHO/SDE/PHE/OEH/06.02, 2006.
World Health Organization (WHO): Review of evidence on health aspects of air
pollution - REVIHAAP Project, 2013.
World Health Organization (WHO): WHO’s urban ambient air pollution database -
Update 2016, version 2.0, 2016.
![Page 175: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/175.jpg)
Bibliography
158
Yttri, K. E., Simpson, D., Nøjgaard, J. K., Kristensen, K., Genberg, J., Stenström, K.,
Swietlicki, E., Hillamo, R., Aurela, M., Bauer, H., Offenberg, J. H., Jaoui, M., Dye, C.,
Eckhardt, S., Burkhart, J. F., Stohl, A., and Glasius, M.: Source apportionment of the
summer time carbonaceous aerosol at Nordic rural background sites, Atmos. Chem.
Phys., 11, 13339-13357, doi:10.5194/acp-11-13339-2011, 2011.
Zar, J.H., Biostatistical Analysis, 4th ed., 929 pp., Prentice Hall, Englewood Cliffs,
New Jersey, 1999.
Zenobi, R and Knochenmuss, R.: Ion formation in MALDI mass spectrometry, Mass
Spectr. Rev., 17, 337-366, doi:10.1002/(sici)1098-2787(1998)17:5<337::aid-
mas2>3.3.co;2-j, 1998.
Zhang, Q., Jimenez, J. L., Canagaratna, M. R., Allan, J. D., Coe, H., Ulbrich, I.,
Alfarra, M. R., Takami, A., Middlebrook, A. M., Sun, Y. L., Dzepina, K., Dunlea, E.,
Docherty, K., DeCarlo, P. F., Salcedo, D., Onasch, T., Jayne, J. T., Miyoshi, T.,
Shimono, A., Hatakeyama, S., Takegawa, N., Kondo, Y., Schneider, J., Drewnick, F.,
Borrmann, S., Weimer, S., Demerjian, K., Williams, P., Bower, K., Bahreini, R.,
Cottrell, L., Griffin, R. J., Rautiainen, J., J. Y., Zhang, Y. M., and Worsnop, D. R.:
Ubiquity and dominance of oxygenated species in organic aerosols in
anthropogenically-influenced Northern Hemisphere midlatitudes, Geophys. Res. Lett.,
34, L13801, 2007.
Zhang, Y., Schauer, J. J., Shafer, M. M., Hannigan, M. P., and Dutton, S. J.: Source
apportionment of in vitro reactive oxygen species bioassay activity from atmospheric
particulate matter, Environ. Sci. Technol., 42, 7502-7509, doi:10.1021/es800126y,
2008.
Zhang, Q., Jimenez, J. L., Canagaratna, M. R., Ulbrich, I. M., Ng, N. L., and
Worsnop, Y. S.: Understanding atmospheric organic aerosols via factor analysis of
aerosol mass spectrometry: a review, Anal. Bioanal. Chem., 401, 3045-3067,
doi:10.1007/s00216-011-5355-y, 2011.
![Page 176: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/176.jpg)
159
Zhang, Y. L., Perron, N., Ciobanu, V. G., Zotter, P., Minguillón, M. C, Wacker, L.,
Prévôt, A. S. H., Baltensperger, U., and Szidat, S: On the isolation of OC and EC and
the optimal strategy of radiocarbon-based source apportionment of carbonaceous
aerosols, Atmos. Chem. Phys., 12, 10841-108556, doi:10.5194/acp-12-10841-2012,
2012.
Zorn, S. R., Drewnick, F., Schott, M., Hoffmann, T., and Borrmann, S.:
Characterization of the South Atlantic marine boundary layer aerosol using an aerodyne
aerosol mass spectrometer, Atmos. Chem. Phys., 8, 4711-4728, doi:10.5194/acp-8-
4711-2008, 2008.
Zotter, P., Ciobanu, V. G., Zhang, Y. L., El-Haddad, I., Macchia, M., Daellenbach,
K. R., Salazar, G. A., Huang, R.-J., Wacker, L., Hueglin, C., Piazzalunga, A., Fermo, P.,
Schwikowski, M., Baltensperger, U., Szidat, S., and Prévôt, A. S. H.: Radiocarbon
analysis of elemental and organic carbon in Switzerland during winter-smog episodes
from 2008 to 2012 - Part 1: Source apportionment and spatial variability, Atmos. Chem.
Phys., 14, 13551-13570, doi:10.5194/acp-14-13551-2014, 2014.
Zotter, P., Herich, H., Gysel, M., El Haddad, I., Zhang, Y., Močnik, Hüglin, C.,
Baltensperger, U., Szidat, S., and Prévôt, A. S. H.: Evaluation of the absorption
Ångström exponents for traffic and wood burning in the Aethalometer based source
apportionment using radiocarbon measurements of ambient aerosols, Atmos. Chem.
Phys., 17, 4229-4249, https://doi.org/10.5194/acp-17-4229-2017, 2017.
![Page 177: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/177.jpg)
Bibliography
160
![Page 178: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/178.jpg)
161
9 List of Figures
Figure 1.1: Idealized ambient particulate size distribution including formation and removal processes (from Finlayson-Pitts and Pitts, 2000). ................................................................................................. 2
Figure 1.2: Radiative forcing 2011 (1750 as reference for pre-industrial atmosphere) from Myhre et al. (2013). ............................................................................................................................................. 3
Figure 1.3: Global Radiative forcing of different ambient aerosol constituents evolving over time (from Myhre et al., 2013). .................................................................................................................... 4
Figure 1.4: Surface solar radiation in Potsdam, Germany, (GEBA, Gilgen et al., 1999) and sulfur emissions in Europe (Stern et al., 2005) temperature anomalies over Europe from 1930 to 2010, and annual temperature anomaly in Europe over land (Brohan et al., 2006). ............................................. 5
Figure 1.5: Mortality and SO2 concentrations during the great London smog (adapted from Bell et al., 2001 and completed with data from the same publication). ........................................................... 6
Figure 1.6: Global distribution of mortality attributed to air pollution (from Lelieveld et al., 2015). . 7
Figure 1.7: Modelled particle deposition probability upon inhalation in the human respiratory system as a function of particle size and location (from Maynard and Kuempel, 2005). ................................. 9
Figure 1.8: Aerosol composition for different environments in Europe in PM10 and coarse PM, with unaccounted mass from water and other aerosol constituents which were not analyzed at all sites (adapted from Putaud et al., 2004). .................................................................................................... 10
Figure 1.9: NR-PM1 AMS analyses at multiple sites on the northern hemisphere: a) aerosol composition (organic (Org), sulfate, nitrate, ammonium, chloride), b) oxygenated OA concentrations, c) and primary OA factors as HOA and BBOA (from Zhang et al., 2011). .............. 11
Figure 2.1: NR-PM1 mass concentrations (organic, sulfate, nitrate, ammonium, and chloride) measured by AMS at multiple sites. Source apportionment results of OA are displayed: traffic (HOA), other POA as BBOA (other OA), oxygenated OA linked to SOA (OOA), separated into semi-volatile (SV-OOA) and low-volatility (LV-OOA) from Jimenez et al. (2009). ........................ 14
![Page 179: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/179.jpg)
List of figures
162
Figure 3.1: Scheme of the high-resolution-time-of-flight aerosol mass spectrometer (HR-ToF-AMS) from DeCarlo et al. (2006). ................................................................................................................ 18
Figure 3.2: Example of mass resolution of 3 AMS types: Quadrupole, C-ToF-MS, HR-ToF-MS from De Carlo et al. (2006). ........................................................................................................................ 19
Figure 4.1: Data recorded with HR-ToF-AMS of filter samples collected in Zürich (2011-2012). Data from a typical measurement cycle are underlaid in gray. (a) Raw signals obtained for organic aerosol (OA, green), nitrate (NO3
- , blue), sulfate (SO42-, red), and ammonium (NH4
+, orange), where AMS filter air as well as blank and sample measurements are indicated. (b) OA average signal for samples and blanks (logarithmic scale), blank correction curve and the noise (smoothed standard deviation of the blank) associated with the signal of different species used for the calculation of errors. On the y axes, a.u. denotes arbitrary units. ............................................................................. 35
Figure 4.2: Offline AMS SO42- blank corrected concentrations compared to theoretical SO4
2- loadings of the filter fractions (μg). The theoretical SO4
2- loadings are calculated based on ambient SO4
2- concentrations measured by the ACSM for the Zürich yearly cycle and the volume of air sampled through the analyzed filter fraction. Results are fitted using a power function (ln(y) = 2.3 × ln(x) - 5.2). ......................................................................................................................................... 36
Figure 4.3: Estimated recoveries of organic compounds based on the comparison of OA/SO42- ratios
using the offline AMS method to reference measurements for different days. The error bars represent the variability of the offline OA/SO4
2- ratio within a sample and were obtained from different runs during the same measurement of the same sample. (a) The reference OA/SO4
2- ratio is obtained by OC filter measurements (Sunset OC/EC analyzer) scaled to OA using OM/OC values from the HR offline AMS data and SO4
2- from IC. (b) OA/SO42- ratios from online measurements were used as
reference values. For both Paris campaigns and the Zürich spring campaign, the online measurements were conducted using HR-ToF-AMS and for the yearly cycle in Zürich by a quadrupole ACSM. (c) For Zürich (2011-2012), probability density functions of Rbulk are presented both using the offline AMS measurements as well as using WSOC from the Sunset OC=EC Analyzer (in combination with OM=OC ratios from offline AMS). ................................................. 38
Figure 4.4: Comparison between 24 h average online and offline AMS (both PM2:5) spectra for winter (a) and spring (b) samples, collected in Zürich. Fragments (m/z) commonly considered as source-specific markers are explicitly labeled with their nominal mass. ........................................... 41
Figure 4.5: Median recovery of single organic fragments, and chemical families for the Zürich spring campaign (offline vs. online PM2.5 AMS). The first and third quartiles of the inter-sample variability are shown as error bars. A ratio of 1 indicates a recovery of 100 %. The fragments are color-coded with the family (CH (hydrocarbon fragments, split into saturated and unsaturated), CHOz=1 and CHOz>1 (oxygenated fragments), and CHN, nitrogen-containing hydrocarbon fragments). Numbers across the top of the plot indicate the fragments’ nominal mass. Families include all respective fragments weighted by their mass contribution. .............................................. 42
Figure 4.6: Change in the time-dependent contribution of Q/Qexp as a function of the number of factors Δ(Qi,cont/Qexp,i,cont) for a chosen offline solution (for aHOA = 0.0 and aCOA = 0.0). .................. 43
Figure 4.7: Residuals weighted with the uncertainty (residuals/uncertainty) of the offline solutions for the periods April-September and October-March (example shown for one chosen solution, aHOA = 0.0; aCOA =0.0). Panels (a, b) show residuals as a function of m/z averaged over the whole periods color-coded with the probability that the residuals for April-September are the same as for October-
![Page 180: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/180.jpg)
163
March (Wilcoxon-Mann-Whitney test). Panels (c, d) shown the probability distribution function (pdf) of R/U during the same periods. ................................................................................................ 45
Figure 4.8: Relative width of the distributions cj/cmarker displayed as a function of aHOA and aCOA. Panel (a) shows the sum of the criteria for HOA, BBOA, and OOA being the sum of OOA1 and OOA2. (The chosen solutions are pointed out in the white area.) Panels (b-d) show the individual criteria as a function of the a values of HOA and COA aHOA, aCOA. .................................................. 47
Figure 4.9: Comparison of overall factor profiles obtained for the chosen solutions both from the offline (left, for HOA and COA, spectra from Mohr et al. (2012) were used as reference) and the retrieved factor profiles from the online source apportionment (right). ............................................. 48
Figure 4.10: Recoveries Rk for HOA, COA, BBOA, and OOA (OOA1COOA2) obtained from the intercomparison of source apportionment results of offline AMS to online ACSM data (Zürich 2011-2012). 100 000 random combinations of offline and online solutions and randomly chosen offline repeats result in the same amount of time-independent Rk, which are expressed as probability density functions (pdf). ................................................................................................................................... 50
Figure 4.11: Comparison of factor contributions from separate offline (PM10 AMS, two constrained factors: HOA, COA) and online (PM1 ACSM) source apportionment using ME-2 (traffic (HOA), cooking (COA), biomass burning (BBOA), and oxygenated organic aerosol; OOADOOA1COOA2). Factor-specific recoveries (Rk) are applied to the offline contributions. Error bars (in gray) denote the variability between the different ME-2 solutions and for different recorded spectra per sample for offline and for online only the first of the two. Panels (a-d) show scatter plots comparing the absolute contribution of the respective source/OA category for offline AMS and online ACSM measurements. The color code distinguishes all factor contributions (bullets, saturated colors) from winter points (open circle, light colors). The gray dashed line indicates the 1 V 1 line. Panels (e-g) show the correlation with the respective markers: black symbols represent the absolute contribution of the respective source for the online ACSM measurements and the colored symbols represent the absolute contribution of the same source for the offline AMS measurements. .................................. 53
Figure 4.12: Ranges of ratios of the contribution of different factors to their markers for the offline (corrected with Rk) and online ACSM source apportionment results. Note that OOA is the sum of OOA1 and OOA2. .............................................................................................................................. 54
Figure 5.1: Map of study area with locations of sites indicating their characteristics. The topography is displayed as meters above sea level. ............................................................................................... 61
Figure 5.2: Step-by-step outline of adopted source apportionment approach (factor recoveries Rk). aHOA and aHOA represent the a value applied for HOA and COA, respectively. ................................. 66
Figure 5.3: PMF factor profiles of HOA, COA, BBOA, SOOA, WOOA, SC-OA, color-coded with ion family of PMFblock (average). fm/z is the relative intensity at a specific mass-to-charge ratio (m/z). ............................................................................................................................................................ 73
Figure 5.4: HOA, COA, BBOA, SC-OA, SOOA, and WOOA and their respective marker concentrations as a function of time for Zürich in 2013. Depicted are the median factor time series results for the different PMF datasets (median) including the uncertainties for PMFblock (first and third quartile) (green: PMFblock, black: PMFzue,isol, red: PMFzue,reps, pink bullets: PMF1filter/month). ..... 74
Figure 5.5: Scatter-plots for the different extreme sensitivity tests for Zürich and for all sites for PMFblock median concentrations): a) HOA vs NOx, b) BBOA vs levoglucosan, c) SOOA vs temperature, d) WOOA vs NH4
+. ....................................................................................................... 75
![Page 181: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/181.jpg)
List of figures
164
Figure 5.6: Distributions of Rk for HOA, COA, BBOA, OOA (WOOA plus SOOA) and SC-OA (500 pairs). A priori information for HOA, COA, BBOA, and OOA on Rk is used from Daellenbach et al., 2016, with propagated errors and biases, while RSC-OA is determined in this study. Distributions of all factors have a resolution of dRk=0.01 except for dRSC-OA=0.05. ........................ 80
Figure 5.7: Relative σa (a) and err‘tot (b) for factor concentrations > 0.1 µg/m3 as a function of factor concentration. err‘tot includes the uncertainties from a-value, seed variability and Rk, and the different PMF datasets. ...................................................................................................................... 81
Figure 5.8: Cellulose concentrations as a function of the season and site. For comparison literature data from other years is added European sites: Payerne (Bozzetti et al., 2016, error bars representingthe standard deviation of the measurements in June and July), Puy de Dôme, Schauinsland, Sonnblick, K-Puszta (Sanchez-Ochoa et al., 2007), Birkenes, Hyytiälä, Lille Valby, and Vavihill (Yttri et al., 2011). ......................................................................................................... 84
Figure 5.9: Map of Switzerland with yearly cycles. Negative concentrations were set to 0 prior to normalization for display. The OA mass explained by the source apportionment analysis is termed OAexpl. ................................................................................................................................................ 86
Figure 6.1: m/z calibration of LDI-MS analysis of aerosol collected on quartz-fiber filter: a) example of calibrated silver (Ag-spec, blue) and no-silver-containing average mass spectra (noAg-spec, green) of a filter sample. The insert in a) displays a zoom-in of the Ag-spec and noAg-spec. b) operational resolution determined based on silver-mono-, di-, and trimer. c) m/z calibration accuracy for both steps of the m/z calibration. ................................................................................................ 102
Figure 6.2: Error model parametrization based on 3 samples (Basel, 2013-06-21, 2013-09-21, 2013-12-22) measured on 3 instances with each time 10 repeats on 1 sample holder. ............................. 103
Figure 6.3: samples representing different combustion sources a) and b) traffic emissions from samples collected in Islisberg tunnel (exit, Wettswil, Switzerland), and c) primary wood burning using whole cycle emissions d) and stable flaming phase emissions d). In absence of a measurable mass spectrum, no spectrum for cooking emissions is displayed. .................................................... 104
Figure 6.4: LDI-MS mass spectra from summer and winter samples (dates given in the legend) from a traffic-influenced (Bern), a rural (Payerne), a wood burning-influenced (San Vittore) and an urban background (Zürich) site. Spectra are color-coded with the correlation coefficient (r) between the m/zs and a specific environmental parameter for the whole dataset: for a) the correlation with NOx during summer, for b) the correlation with temperature, for c) and d) the correlation with levoglucosan. .................................................................................................................................... 105
Figure 6.5: PMF factor profiles (colored sticks) and their uncertainty (grey shaded areas, variability among PMF runs): a) traffic1, b) traffic2, c) efficient wood burning (BBeff), d) inefficient wood burning (BBineff1), e) inefficient wood burning 2 (BBineff2), f) lower molecular weight OA (LMW-OA), and g) biogenic OA (bio-OA). .................................................................................... 107
Figure 6.6: Scatterplots between factor-time series and respective markers (traffic1, traffic2, BBeff, BBineff1, BBineff2, bio-OA, LMW-OA, eBCtr, levoglucosan, potassium, ammonium are displayed in µg/m3, NOx in ppm, and temperature in °C)................................................................................ 108
Figure 6.7: Comparison of LDI-MS to reference offline AMS source apportionment results for the sum of traffic related factors (LDI-traffic, R2
LDI,oAMS=0.04), sum of wood burning related factors (LDI-BB, R2
LDI,oAMS =0.83), bio-OA (R2LDI,oAMS =0.62), and LMW-OA (R2
LDI,oAMS =0.45). ......... 111
![Page 182: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/182.jpg)
165
Figure 6.8: Comparison of LDI-MS factor concentrations corrected using relative response factors (rRF) to uncorrected LDI factor concentrations (Scenario 1). .......................................................... 115
Figure 6.9: Relative for the different PMF factors as a function of the factor concentration. ... 116
Figure 6.10: Relative factor time series of 7 identified factors for all nine sites in study area. ........ 117
Figure 6.11: a) time series of traffic1 normalized to eBCtr and ECf in comparison to OCf /ECf in Magadino, b) influence of inefficient wood burning emissions (BBineff1) in comparison to the sum of wood burning influenced factors (BB=BBineff1+BBineff2+BBeff) for the entire datasets. ....... 118
Figure 7.1: Average contributions of OA sources resolved using offline AMS data discussed in this thesis and other studies (this dissertation, and from Bozzetti et al., 2016, 2017a, 2017b, Huang et al., 2014). ................................................................................................................................................ 122
SI Figure A.1: Qi,j as a function of the number of factors for a reference experiment with all data used in PMF (9 sites, full year 2013, HOA and COA constrained with a=0.0 (b and d). Δ(median(Qi,j))max is evaluated for the different periods during the year 2013 (January-February-March, April-Mai-June, July-August-September, October-November-December) and for all sites (a and c). The grey line depicts the difference between the category (geographical or season) with the highest and the lowest median Qi,j. ................................................................................................... 170
SI Figure A.2:Qi,j as a function of the day of the week. ................................................................... 171
SI Figure A.3: a) Average Qi,j of ions in PMFblock as a function of their mass-to-charge ratio (m/z). The ions are color-coded with their composition (CH: ions consisting only of C and H; CHO1: ions consisting of C, H, and 1 O; CHOgt1: ions consisting of C, H, and more than 1 O; CHN: ions consisting of C, H, N, (and O); CS: ions consisting of C, H, S, (and O)). b) Average Qi,j of the ions in PMFblock as a function of their mass defect (exact mass - nominal mass) as well as a histogram of the number of ions with a certain mass defect. The mean Qi,j of the ion families is displayed separately. ......................................................................................................................................... 172
SI Figure A.4: Cumulative density functions of a-values for HOA and COA for the accepted solutions. ........................................................................................................................................... 173
SI Figure A.5: Histograms of yearly average factor concentrations of all selected PMFblock solutions (after Rk correction). ......................................................................................................................... 174
SI Figure A.6: mass spectral fingerprints of BBOA (PMFblock) and nebulized levoglucosan. fion is the fraction of signal of a respective ion to the sum of the total signal. ................................................. 176
SI Figure A.7: SOOA concentrations compared to temperature, ozone, and Ox (O3+NO2) for Zürich. .......................................................................................................................................................... 177
SI Figure A.8: number density functions of source apportionment results obtained using a significance level of 0.05 normalized to results obtained using a significance level of 0.5: a) Comparison of factor concentrations b) Comparison of uncertainty estimate (σa). .......................... 179
SI Figure B.1: Cumulative density function (CDF) of the peak intensity in field blank analyses (Iavg,m/z) normalized to the minimal absolute error (σabs). ................................................................. 182
![Page 183: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/183.jpg)
List of figures
166
SI Figure B.2: mass spectral signature of field blank analysis: a) colour-coded with the ratio of peak intensity (Iavg,m/z) divided by the minimal absolute error (σabs), b) colour-coded with the ratio of the measurement of the peak intensity (Istdev,m/z) divided by Iavg,m/z. ...................................................... 183
SI Figure B.3: measured total intensity plotted against filter loadings in OC, OC+EC, and PM10 (color-coded with measurement date). ............................................................................................. 184
SI Figure B.4: Difference of total measured signal intensity (Itot) of samples repeated on different occasions to the average of Itot normalized to the average of Itot (lines with dots with varying colors for the different samples). The data is summarized using the median (black diamond) and quartiles (black horizontal lines), binned as intra-day repeats, repeats within 30, 30-60,60-90,90-120 days, respectively. ..................................................................................................................................... 185
SI Figure B.5: Scatterplots of selected m/zs (scaled to OM, µg/m3) and NOx (ppm), levoglucosan (µg/m3), and temperature (°C). ......................................................................................................... 186
SI Figure B.6:Qavg as a function of the number of factors for the single sites (color-coded with geographical region). ........................................................................................................................ 187
SI Figure B.7: change in time-dependent contribution Qavg,i as a function of the number of factors, ΔQavg,i. .............................................................................................................................................. 188
SI Figure B.8: Qavg as a function of m/z when allowing for 7 factors. ............................................. 189
SI Figure B.9: Scatterplots of selected traffic 1 / 2 (µg/m3) and NOx (ppm), and temperature (°C). Correlation coefficients are computed on all summer points together. ............................................ 192
SI Figure B.10: Difference in factor concentration of samples measured on different instances to the average factor concentration normalized to the average concentration as a function of measurement time delay. The data is summarized using the median (diamond) and quartiles (horizontal lines) of the samples, binned as intra-day repeats, repeats within 30, 30-60,60-90,90-120 days, respectively (195, 240, 46, 53, 28 points per bin, respectively). .......................................................................... 194
SI Figure B.11: relative yearly average factor contributions for the different sites (grey: traffic1, black: traffic2, orange: BBeff, brown: BBineff1, violet: BBineff2, light green: bio-OA, dark green: LMW-OA. ........................................................................................................................................ 195
![Page 184: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/184.jpg)
167
10 List of Tables
Table 1.1: PM limits established in Switzerland, EU, U.S.A., and WHO. ........................................... 7
Table 3.1: Advantages and disadvantages of offline AMS technique. ............................................... 20
Table 3.2:Advantages and disadvantages of LDI technique. .............................................................. 21
Table 4.1: Filter samples and available supporting measurements used in this study. ....................... 28
Table 5.1: Study sites with geographical location and classification ................................................. 61
Table 5.2: Comparison of factor time series to reference data for different PMF input datasets runs (by Pearson and Spearman correlation coefficient, Rp
2 and Rs). Displayed are the results for PMFblock unless stated otherwise. ...................................................................................................................... 77
Table 5.3: Yearly average contribution and uncertainty of resolved factors for PMFblock run for the different sites and the average for all sites. The uncertainty is calculated based on the variability in the yearly averages from PMFblock and the variability between the 4 sensitivity tests. ...................... 87
Table 6.1: relative response factors (rRF) for LDI-MS analyses for 3 different scenarios. ............. 114
Table 6.2: yearly averages of the relative factor contributions and NOx concentrations. ................ 119
SI Table A.1:set of acceptance criteria used. r is the correlation coefficient between a factor time series and the respective marker. Q25 is the 1st quartile and Q75 the 3rd quartile. .............................. 175
SI Table A.2: σminimal and k for the different factors including their uncertainty.............................. 179
SI Table B.1: correlation coefficients between factor and marker time series (Rp: pearson correlation coefficient, Rs: Spearman correlation coefficient). ........................................................................... 191
SI Table B.2: error model coefficients for parameterized σintrad. ...................................................... 193
![Page 185: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/185.jpg)
List of tables
168
![Page 186: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/186.jpg)
169
A
Supplementary Material for chapter 5
Long-term chemical analysis and organic aerosol source apportionment at 9 sites in Central Europe:
source identification and uncertainty assessment
![Page 187: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/187.jpg)
A: Supplementary material
• Number of factors:
Qi,j is computed using the PMF residuals (eij) and the PMF input errors (si,j):
(SI.A.1)
SI Figure A.1: Qi,j as a function of the number of factors for a reference experiment with all data used in PMF (9 sites, full year 2013, HOA and COA constrained with a=0.0 (b and d). Δ(median(Qi,j))max is evaluated for the different periods during the year 2013 (January-February-March, April-Mai-June, July-August-September, October-November-December) and for all sites (a and c). The grey line depicts the difference between the category (geographical or season) with the highest and the lowest median Qi,j.
Fig. SI.A.1 shows Qi,j s as a function of the number of factors for different sites (b)
and seasons (d) and the difference between the highest (a) and lowest (c) median to
evaluate the maximal difference in the mathematical quality of the solutions. As
expected, forcing PMF to explain the variability in the dataset only with the 2
constrained factors (p=2), results in very high median Qi,j . Δ(median(Qi,j))max shows the
difference in the median Qi,j between groups of points like sites or season. The smaller
the Δ(median(Qi,j))max, the smaller are the differences in the mathematical quality of the
PMF solution for the different seasons/sites. To explain the temporal and geographical
variability at least 5 factors are required. However, the difference between the site that
is best explained and the site that is least explained is approximately 6 when using 5 or
![Page 188: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/188.jpg)
171
6 factors. When increasing to 6 factors, also a factor explaining the variability of sulfur-
containing organic ions (especially, CH3SO2+) is resolved. Therefore, we opted to
perform PMF using 6 factors. Using 6 factors, there is also no difference between the
average Qi,j on week-days and weekend (Fig. SI.A.2).
SI Figure A.2:Qi,j as a function of the day of the week.
However, for PMFblock also with 6 factors, the average Qi,j is clearly larger (7 only
the Zürich data points) than the ideal value of 1, i.e. the PMF residuals are larger than
the measurement uncertainties. In comparison to PMFblock, the average Qi,j for Zürich is
slightly reduced for the same number of factors when only including 1 site in PMF
(PMFzue,isol, PMFzue,reps, average Qi,j 6). In this study, we analyse yearly cycles and,
thereby, assume constant factor profiles throughout the year which can contribute to
Q>1.
Another possible reason for Q>1 is an underestimation of the measurement
uncertainty. A main contributor in high-resolution AMS data treatment (attribution of
the signal at a nominal mass to several ions) stems from errors in the m/z calibration
which could not be incorporated in the current data analysis. Recent studies demonstrate
that for overlapping peaks (ions) the measurement uncertainties are strongly
underestimated (Cubison et al., 2015; Corbin et al., 2015). For PMFblock using 6 factors,
average Qi,j do not depend on m/z but rather on the ion family (Fig. SI.A.3): ions
consisting of C, H, S, (and O) summarized under the name (CS) and ions consisting of
C, H, N, (and O) summarized under the name CHN have a higher Qi,j than hydrocarbon
![Page 189: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/189.jpg)
A: Supplementary material
ions (CH, only C and H) and oxygenated ions (CHOz=1 with 1 oxygen and CHOz>1 with
more than 1 oxygen). Since the time series of CH3SO2+ is event-driven, the high Qi,j of
this ion hints to the fact that PMF is unable to accurately resolve all of these events.
The average Qi,j for ions with a mass defect (nominal mass - exact ion mass) around
0.03 a.m.u. is higher than for the other ions (Fig SI.A.3). Mass defects in this range are
most common in our dataset. This makes these peaks prone to overlap with other ions
and thus their error prone to an underestimation because this effect is not considered in
the sij calculation (described above).
SI Figure A.3: a) Average Qi,j of ions in PMFblock as a function of their mass-to-charge ratio (m/z). The ions are color-coded with their composition (CH: ions consisting only of C and H; CHO1: ions consisting of C, H, and 1 O; CHOgt1: ions consisting of C, H, and more than 1 O; CHN: ions consisting of C, H, N, (and O); CS: ions consisting of C, H, S, (and O)). b) Average Qi,j of the ions in
![Page 190: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/190.jpg)
173
PMFblock as a function of their mass defect (exact mass - nominal mass) as well as a histogram of the number of ions with a certain mass defect. The mean Qi,j of the ion families is displayed separately.
SI Figure A.4: Cumulative density functions of a-values for HOA and COA for the accepted solutions.
Cumulative density functions for the a-values of HOA and COA are presented for
the accepted solutions in Fig. SI.A.4. We found that 80% of the accepted solutions have
an a-value≤0.3 for HOA and an a-value≤0.5 for COA. The output HOA and COA
factor profiles are therefore not significantly variable and very similar to the input
profiles, indicating that similar solutions were selected. Furthermore, the yearly average
factor concentrations of all selected PMFblock solutions after Rk correction are shown for
the case of Zürich as an illustration in Fig. SI.A.5. The distributions of each of the
different factors do not show more than 1 distinct mode, indicating that we do not have
several populations of solutions.
![Page 191: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/191.jpg)
A: Supplementary material
The yearly average factor concentrations of all selected PMFblock solutions after Rk
correction areshown for the case of Zürich as an illustration (Fig. SI.A.5). The
distributions of each of the different factors do not show more than 1 distinct mode.
SI Figure A.5: Histograms of yearly average factor concentrations of all selected PMFblock solutions (after Rk correction).
![Page 192: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/192.jpg)
175
• Quality assessment of solutions:
Set of criteria used when assessing quality of a single PMF run:
SI Table A.1: set of acceptance criteria used. r is the correlation coefficient between a factor time series and the respective marker. Q25 is the 1st quartile and Q75 the 3rd quartile. criteria on profile f(CO2
+) f(C2H4O2+)
HOA <0.4 <0.004 COA <0.4 <0.01 Criteria on time series HOA r(HOA,NOx)> 0 & r(HOA,NOx)> r(COA,NOx) BBOA r(BBOA,levo)> 0 SC-OA r(SC-OA,CH3SO2
+)>0 Mass closure criteria
OCres total Q25(res-OC)<0 & Q75(res-OC)>0 Magadino winter, Magadino summer Q25(res-OC)<0 & Q 75(res-OC)>0 Zürich winter, Zürich summer Q25(res-OC)<0 & Q 75(res-OC)>0 Magadino, Zürich Q25(res-OC)<0 & Q 75(res-OC)>0 HOC<median, HOC>median Q25(res-OC)<0 & Q 75(res-OC)>0 COC<median, COC>median Q25(res-OC)<0 & Q 75(res-OC)>0 BBOC<median, BBOC>median Q25(res-OC)<0 & Q 75(res-OC)>0 SC-OC<median, SC-OC>median Q25(res-OC)<0 & Q 75(res-OC)>0 WOOC<median, WOOC>median Q25(res-OC)<0 & Q 75(res-OC)>0 SOOC<median, SOOC>median Q25(res-OC)<0 & Q 75(res-OC)>0 for PMF with 12 filters per site summer Magadino and Zürich
Q25(res-OCi)<0 & Q 75(res-OCi)>0
for PMF with 12 filters per site winter Magadino and Zürich
Q25(res-OCi)<0 & Q 75(res-OCi)>0
![Page 193: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/193.jpg)
A: Supplementary material
• Comparison of mass spectral signature of BBOA and nebulized levoglucosan:
Fig. SI.A.6 demonstrated the high similarity between the retrieved BBOA
signature and the mass spectrum of nebulized levoglucosan.
SI Figure A.6: mass spectral fingerprints of BBOA (PMFblock) and nebulized levoglucosan. fion is the fraction of signal of a respective ion to the sum of the total signal.
![Page 194: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/194.jpg)
177
• Comparison of SOOA to ozone and Ox:
In Figure SI.A.7, we compare the SOOA concentrations to ozone and Ox (O3+NO2)
for Zürich. The SOOA concentrations follow best the temperature (Rs,SOOA,temp=0.65,
Fig. SI.A.7.a) but show also some correlation to ozone Rs,SOOA,O3=0.33, Fig. SI.A.7.b)
and Ox (Rs,SOOA,Ox=0.38, Fig. SI.A.7.c).
SI Figure A.7: SOOA concentrations compared to temperature, ozone, and Ox (O3+NO2) for Zürich.
![Page 195: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/195.jpg)
A: Supplementary material
• Uncertainty estimation and propagation:
The uncertainty described by the interquartile range from the a-value sensitivity
assessment (σa) does not fully explain the variability between the 4 sensitivity tests. In
the following, we use the source apportionment results of the 12 filters common to all 4
sensitivity tests for achieving a better estimate of the uncertainty of the factor
concentrations. For these 12 filters the uncertainty is estimated by propagating the
variability between the median concentrations for the 4 sensitivity tests (σb) and half the
interquartile range of PMFblock (σa, Eq. SI.A.2):
(SI.A.2)
In absence of σb for all other points, we parametrize σb. We express σb as a function
of a minimal uncertainty (σminimal) and an uncertainty proportional (k) to the factor
concentration and fit the equation using the 12 points in common to all datasets (Eq.
SI.A.3):
( SI.A.3)
The uncertainty (σa and σb) is propagated for all points using the parameters from Eq.
SI.A.3 in order to obtain the total uncertainty for all points in the dataset (Eq. SI.A.4):
(SI.A.4)
The resulting coefficients of the error model are presented in table SI.A.2:
![Page 196: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/196.jpg)
179
SI Table A.2: σminimal and k for the different factors including their uncertainty.
factor σminimal k
HOA 0.16±0.06 0.39±0.24
COA 0.09±0.01 0.52±0.09
BBOA 0.06±0.01 0.48±0.05
SC-OA 0.30±0.00 0.32±0.27
WOOA 0.28±0.08 0.42±0.27
SOOA 0.05±0.01 0.24±0.05
• Sensitivity to significance level of statistical tests in PMFblock:
For PMFblock, a sensitivity test with significance level of 0.05 instead of 0.5 as in the
base case was performed. The factor concentrations and their corresponding
uncertainties (σa) are compared and displayed as number density functions (Fig. SI.A.8).
Changes in the estimated factor concentrations are within 10% of the factor
concentrations for SCOA and smaller for all other factors. The uncertainty related to
COA is decreased when lowering the significance level to 0.05, while the other factors
remain largely unaffected.
SI Figure A.8: number density functions of source apportionment results obtained using a significance level of 0.05 normalized to results obtained using a significance level of 0.5: a) Comparison of factor concentrations b) Comparison of uncertainty estimate (σa).
![Page 197: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/197.jpg)
A: Supplementary material
![Page 198: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/198.jpg)
181
B
Supplementary material for chapter 6
Insights into organic-aerosol sources via a novel laser-desorption/ionization mass spectrometry
technique applied to one year of PM10 samples from nine sites in central Europe
![Page 199: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/199.jpg)
A: Supplementary material
• Field blank analyses
Analyses of field blanks (also spiked with AgNO3) exhibited low signal with 99% of
peaks below detection limit (defined as 3*σabs, Fig SI.B.1).
SI Figure B.1: Cumulative density function (CDF) of the peak intensity in field blank analyses (Iavg,m/z) normalized to the minimal absolute error (σabs).
Mass spectra from field blanks exhibit low signals throughout the spectrum (Fig
SI.B.1 and SI.B.2). 1% of all peaks (m/z 197, 249, 251, 322,324) show signals which
are higher than the detection limit (Fig SI.B.2a, highlighted in black). These peaks are
also subjected to a high variability between the repeated analyses (Figure SI.B.2b,
highlighted in black).
![Page 200: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/200.jpg)
183
SI Figure B.2: mass spectral signature of field blank analysis: a) colour-coded with the ratio of peak intensity (Iavg,m/z) divided by the minimal absolute error (σabs), b) colour-coded with the ratio of the measurement of the peak intensity (Istdev,m/z) divided by Iavg,m/z.
![Page 201: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/201.jpg)
A: Supplementary material
• Comparison of total LDI-MS signal and external measurements:
The total LDI-MS-intensity measured for the filter samples does not show a
relation to the filter loading in terms of OC, sum of OC and EC, or PM10 (Fig SI.B.3).
SI Figure B.3: measured total intensity plotted against filter loadings in OC, OC+EC, and PM10 (color-coded with measurement date).
![Page 202: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/202.jpg)
185
• Influence of measurement time delays on total LDI-MS signal recorded:
In order to assess the repeatability of the total intensity, all repeats are normalized
to the first measurement of the respective sample (Fig. SI.B.4). The intra-day variability
(without the ones used to parametrize the error model) ranges from -27% (first quartile)
to +33% (third quartile) of the average measurement. The inferred instrumental drift has
an effect on the measured intensity (Kruskal-Wallis-test, p-value<0.05, Kendall-Tau-
Test, p-value<0.05). Therefore, for quantification purposes we use external data like
OM.
SI Figure B.4: Difference of total measured signal intensity (Itot) of samples repeated on different occasions to the average of Itot normalized to the average of Itot (lines with dots with varying colors for the different samples). The data is summarized using the median (black diamond) and quartiles (black horizontal lines), binned as intra-day repeats, repeats within 30, 30-60,60-90,90-120 days, respectively.
![Page 203: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/203.jpg)
A: Supplementary material
• Comparison of selected m/zs and external measurements:
For visual comparison certain m/zs are plotted as a function of external
measurements (Fig. SI.B.5).
SI Figure B.5: Scatterplots of selected m/zs (scaled to OM, µg/m3) and NOx (ppm), levoglucosan (µg/m3), and temperature (°C).
• Preliminary source apportionment analysis and factor identification:
A preliminary source apportionment analysis is performed by unconstrained PMF
(no a priori information on source fingerprints used) in order to determine the number
of factors for the source apportionment analysis. Based on Qavg, defined as
![Page 204: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/204.jpg)
187
(n: number of spectra, m:number of m/zs), the model explains the sites
north and south of the alpine crest equally well when allowing for at least 4 factors (Fig.
SI.B.6). Furthermore, we assess the change in time-dependent Qavg,i, ,
when increasing the number of factors, (Fig. SI.B.7). A significant
signifies that structure in the residuals disappeared when adding an additional factor. Up
to 5 factors, removed structure is evident but not when adding further factors. When
increasing the number of factors up to 7, a factor with a similar signature as stable phase
wood burning and a second traffic factor appear. A further increase leads to the
separation of another possibly traffic related factor which contributes less than 5%.
Therefore, we consider 7 factors to be optimal for this dataset.
SI Figure B.6:Qavg as a function of the number of factors for the single sites (color-coded with geographical region).
![Page 205: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/205.jpg)
A: Supplementary material
SI Figure B.7: change in time-dependent contribution Qavg,i as a function of the number of factors, ΔQavg,i.
For the 7 factor solution, we assess how well the different m/zs are explained by
PMF using the quantity Qavg,j, (Fig. SI.B.8). Qavg,j shows that even with 7
factors not all m/zs are explained within their measurement uncertainty. This might be
linked to an underestimation of the measurement uncertainty itself. Additionally, factor
profiles, which are assumed to be constant, might vary with changing seasons
contributing to increased Qavg,j of certain m/zs. The m/zs 322, 324, 326 might be affected
by the signal from silver-trimers, even though these peaks were removed prior to PMF.
![Page 206: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/206.jpg)
189
SI Figure B.8: Qavg as a function of m/z when allowing for 7 factors.
The explained variation (EV, Paatero et al., 2004, Canonaco et al., 2013) describes
how much of the measured variation (time or variable) is explained by each PMF-factor
(Eq. SI.B.1, for variable j):
for k=1,…,p (SI.B.1)
fk,j are the constant factor profiles and gi,k their time-dependent contributions. The
index i represents a specific point in time (up to the number of points in time n, j the
signal at a specific m/z, and k a factor (up to the number of factors p). ei,j are the PMF-
residuals and σ’PMF,i,j the measurement uncertainties. Using the factor identifications
based on the preliminary source apportionment analysis, the factors of the bootstrap
runs are identified sequentially based on the explained variation as follows:
• traffic1: factor with maximal explained variation of m/z 177. • traffic2: factor with maximal explained variation of m/z 163 of the
remaining factors. • BBeff: factor with maximal average explained variation of the m/zs
85, 124, 140, 197, and 213 of the remaining factors. • BBineff: factor with maximal average explained variation of the m/zs
284 and 298 of the remaining factors.
![Page 207: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/207.jpg)
A: Supplementary material
• BBineff2: factor with maximal average explained variation of ions bigger than 300 a.m.u. of the remaining factors.
• bio-OA: factor with maximal average explained variation of m/z 74 and 104 of the remaining factors.
• LMW-OA: factor with maximal average explained variation of ions smaller than 150 a.m.u. of the remaining factors.
Bootstrap runs that showed mixing between factors described above were not further
considered for the analysis. Criteria used for minimizing the effect of mixing based on
the explained variation of groups of m/zs are lined out in the following:
1. EV(traffic2,m/zs>300) < EV(BBineff 2,m/zs>300) 2. EV(traffic2, m/zs 85, 124, 140, 197, 213) <
EV(BBeff, m/zs 85, 124, 140, 197, 213) 3. EV(traffic2, m/zs 284, 298) <
EV(BBineff1, m/zs 284, 298) 4. EV(LMW-OA, m/zs 284, 298) < EV(BBeff, m/zs 284, 298)
![Page 208: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/208.jpg)
191
• Factor time series vs marker time series
Tab. SI.B.1 summarizes the correlation coefficients of all the factor time series
and respective markers for all sites and as an average.
SI Table B.1: correlation coefficients between factor and marker time series (Rp: pearson correlation coefficient, Rs: Spearman correlation coefficient).
Correlation bas* ber* fra* gal* mag* pay* vad* vi* zue* avg
Rp traffic1 NOx -0.18 -0.31 -0.08 0.13 -0.23 -0.26 -0.1 -0.29 -0.19 -0.18
Rp traffic2 NOx 0.10 -0.06 0.29 0.35 0.54 0.5 0.24 -0.23 0.32 0.24
Rp traffic1 eBCtr
-0.31 0.15
-0.01 -0.08
Rp traffic2 eBCtr
0.73 0.68
0.63 0.69
Rp BBeff levo 0.71 0.65 0.78 0.44 0.83 0.88 0.42 0.18 0.9 0.70
Rp BBineff1 levo 0.88 0.11 0.64 0.91 0.59 0.93 0.93 0.98 0.58 0.83
Rp BBineff2 levo 0.9 0.85 0.77 0.44 0.96 0.81 0.65 0.9 0.79 0.83
Rp BBeff K+ 0.29 0.61 0.78 0.32 0.67 0.1 0.35 -0.02 0.77 0.48
Rp BBineff1 K+ 0.08 0.08 0.85 0.26 0.75 0.11 0.5 0.9 0.4 0.52
Rp BBineff2 K+ 0.19 0.59 0.75 0.4 0.76 0.09 0.48 0.81 0.68 0.57
Rp BBeff BCwb 0.76 0.42 0.75 0.67
Rp BBineff1 BCwb 0.91 0.49 0.55 0.71
Rp BBineff2 BCwb 0.69 0.4 0.82 0.67
Rs bio-OA temp 0.73 0.69 0.74 0.76 0.69 0.78 0.78 0.72 0.75 0.74
Rp LMW-OA NH4+ 0.7 0.46 0.75 0.78 0.78 0.31 0.75 0.76 0.75 0.69
*ber: Bern, bas: Basel, fra: Frauenfeld, gal: St. Gallen, mag: Magadino, Pay: Payerne: vi: S. Vittore, zue, Zürich
![Page 209: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/209.jpg)
A: Supplementary material
Fig. SI.B.9 illustrates the relation between traffic1, traffic2 and NOx and eBCtr
when only using the summer points.
SI Figure B.9: Scatterplots of selected traffic 1 / 2 (µg/m3) and NOx (ppm), and temperature (°C). Correlation coefficients are computed on all summer points together.
• Uncertainty estimation of PMF results:
The variability (standard deviation) of the apportioned factor concentration for the
same measurement in the bootstrap runs is the base for the uncertainty estimate ( ).
Besides , we also account for the effect of the repeatability on the source
apportionment results (assessed by the standard deviation of the apportioned factor
concentration, σintraday). For the latter purpose, 3 filters were measured on 3 instances
throughout the measurement campaign (10 times during 1 day). For these filters, both
types of uncertainty can be readily propagated to (Eq. SI.B.2):
(SI.B.2)
For the other filters, cannot be estimated directly. Therefore, we parameterized
σintraday as a function of two error terms, one absolute and one relative, similarly to the
PMF input uncertainty described in Section 2.3.2: (Eq. SI.B.3):
![Page 210: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/210.jpg)
193
(SI.B.3)
Eq. SI.B.3 was fitted using the repeatability tests for the 3 filters, to obtain the
parameters and for each of the PMF factors (Tab. SI.B.2) and
the factor concentration (conc). These parameters were then extrapolated to the
other filter samples, such that could be applied to all filter samples.
SI Table B.2: error model coefficients for parameterized σintrad.
, µg/m3 , -
traffic1 0.01±0.01 0.08±0.02 traffic2 0.09±0.13 0.00±0.21 BBeff 0.01±0.0 0.10±0.02
BBineff1 0.04±0.04 0.13±0.06 BBineff2 0.11±0.05 0.08±0.07 LMW-OA 0.02±0.01 0.12±0.04
bio-OA 0±0.01 0.12±0.04
![Page 211: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/211.jpg)
A: Supplementary material
• Influence of instrumental drifts on source apportionment results:
Fig SI.B.10 illustrates the difference in the apportioned factor concentrations as a
function of the measurement time. All factors are affected by instrumental changes
occurring during the measurement campaign (Kruskal-Wallis-test, p-value<0.05) and
for all factors but traffic1 (p-value=0.58) and BBineff1 (p-value=0.23) a trend is
identified consistent with an impact of an instrumental drift (ranked Mann-Kendall test,
p-value<0.05). The intra-day variance explains largely the total variance, which is
defined as the sum of intra-day (σintrad2) and inter-day (σinter
2) variance (for traffic1 97%,
traffic2 94%, BBeff 85%, BBineff1 89%, BBineff2 82%, LMW-OA 79%, bio-OA
97%).
SI Figure B.10: Difference in factor concentration of samples measured on different instances to the average factor concentration normalized to the average concentration as a function of measurement time delay. The data is summarized using the median (diamond) and quartiles (horizontal lines) of the samples, binned as intra-day repeats, repeats within 30, 30-60,60-90,90-120 days, respectively (195, 240, 46, 53, 28 points per bin, respectively).
![Page 212: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/212.jpg)
195
• Relative yearly average contributions of PMF factors for all nine sites: Fig. SI.B.11 presents the relative yearly average contributions of the PMF factors for all nine sites as presented in Tab. 6.2 in the manuscript.
SI Figure B.11: relative yearly average factor contributions for the different sites (grey: traffic1, black: traffic2, orange: BBeff, brown: BBineff1, violet: BBineff2, light green: bio-OA, dark green: LMW-OA.
![Page 213: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/213.jpg)
A: Supplementary material
![Page 214: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/214.jpg)
197
11 Acknowledgement
During my time as a Ph.D. student at the Laboratory of Atmospheric Chemistry, I
could pursue my curiosity and met great people. I wish all of you who were part of this
journey all the best:
In particular I would like to thank:
… My doctoral advisor Prof. Dr. Urs Baltensperger for being my doctoral father
and giving me the opportunity to work at LAC.
… My direct supervisor Dr. André Prévôt for all the interesting discussion and
giving me the freedom also to follow side projects.
… Prof. Dr. Thomas Peter and Prof. Dr. James Schauer for accepting to be my co-
referees.
… Prof. Dr. Harald Bugmann for being the chairman at the Ph.D. exam.
… Dr. Guido Vogel and Dr. Valentin Pflüger for introducing my to the world of
LDI, helping me with all the issues, and always warmly welcoming me at MABRITEC.
![Page 215: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/215.jpg)
Acknowledgement
198
… René Richter, Günther Wehrle for helping me set up the autosampler and all the
other times I didn’t know how to approach a problem.
… Thomas Attinger for providing me with a big screen, and especially his help
whenever I had major disagreements with my laptop.
… Lola Schmid for helping me when I was in need for advice or needed something
in the lab and especially for taking such good care of the milliQ system. Without this
water I wouldn’t have been able to do any of my measurements.
… Jay and Maarten for the help and insights related to AMS.
… Dr. Rolf Siegwolf for reminding me that plants are not only aerosol emitters and
all the discussions we had.
… My office was chameleon: in the mornings the smell of past centuries pointed
out the century-long history of PSI. But sometimes, a cinema-like atmosphere was
created by the odors of popcorn. Carlo’s popcorn, well, we all lived on it. Sometimes it
also felt like in a circus with all the madness. I highly enjoyed your company.
… With all that popcorn, how good that Imad “lent” me his squash racket. You lent
me much more than that. Thanks for all your time.
… Dogushan and Manuel who did their best in helping me burn the popcorn when
playing squash. Outstanding.
… Felix for appointing me (besides others) to ensure the edibility of the dishes he
prepared during his cooking emission campaigns.
… Manolis who shared my love for cheese and who never would let me it eat all
alone.
… Stephen helping me eat my fruits despite the danger of sugar.
![Page 216: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/216.jpg)
Acknowledgement
199
… All you offline people: How many filters did we punch? How many liters of
milliQ water did we use? How many hours did we spend besides our dear old manual
nebulizer waiting to switch sample. Thank you, modern world, for the invention of the
auto-sampler. We could create quite some amount of data and had the chance to think in
detail about it.
… Giulia, Nassia, Lassi, Kevin, Emmanuel for all the efforts in the lab.
... Rotational ambiguity, we spoke often about it and accepted it as the reason for
headaches, but Francesco showed us how to enjoy all the craziness related to rotational
ambiguity.
… Joel, one day the people will understand the beauty of playing PMF residuals as
music, we shouldn’t give up on that because it is not crazy and will become a thing at
some point.
… Marco who shared my passion for naming vectors in funny ways.
… the LAC house inhabitants accepting my frequent, self-invited visits
… Andrin, Johanna, Ju, Christoph, Samuel, and all of you, in a lot of ways I grew
up with you, thanks for all the time since 2007.
… I remember I was so annoyed that I had to practice reading, I always thought that
I would just be able to read, thanks Christine for forcing me to learn it, I think it was
worth it. And lastly, yes Hans, I admit it, monkeys can indeed be blue.
Merci öich aune wo sone wichtige Teil vo denä vier Jahr syt gsi.
![Page 217: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/217.jpg)
Acknowledgement
200
![Page 218: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/218.jpg)
12 Curriculum vitae
Kaspar Rudolf Dällenbach
born on October 12, 1988
in Bern, Switzerland
Education
04/2013- 03/2017 PhD, Laboratory of Atmospheric Chemistry, Villigen PSI and
Department of Environmental Systems Science, ETH Zürich.
Dissertation advisor: Prof. Urs Baltensperger (ETH)
2007-2013 Bachelor’s and Master’s Program in Environmental Science
with a subject focus on Atmosphere and Climate at ETH
Zürich, Zürich, Switzerland.
• Master thesis on “Sensitivity Study on the Chlorine
Activation by Volcanic Aerosols” supervised by Prof.
Thomas Peter.
• Internship at Laboratory for Atmospheric Chemistry, Paul
Scherrer Institute, Villigen, Switzerland.
• ERASMUS program: Studies of Biology and Earth
Sciences at the University Uppsala, Sweden
![Page 219: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/219.jpg)
Curriculum vitae
202
2001-2007 Gymnasium at Freies Gymnasium Bern, Bern, Switzerland
1995-2001 Elementary School in Schliern b. Köniz, Switzerland
Platform presentations at conferences
American Association for Aerosol Research 35th Annual Conference, Portland, U.S.A.,
2016 (oral presentation).
17th Annual AMS Users Meeting, Portland, U.S.A., 2016.
European Aerosol Conference, Milan, Italy, 2015 (two oral presentations).
16th Annual AMS Users Meeting, Milan, Italy, 2015.
European Geoscience Union General Assembly, Vienna, Austria, 2014 (poster).
European Aerosol Conference, Prague, Czech Republic, 2013 (oral presentation).
14th Annual AMS Users Meeting, Prague, Czech Republic, 2013.
Invited presentation
Laboratory for environmental chemistry, Paul Scherrer Institute, Villigen PSI.
Publications
Daellenbach, K. R., Bozzetti, C., Krepelova, A., Canonaco, F., Huang, R.-J., Wolf, R.,
Zotter, P., Crippa, M., Slowik, J., Zhang, Y., Szidat, S., Baltensperger, U., Prévôt, A. S.
H., and El Haddad, I.: Characterization and source apportionment of organic aerosol
using offline aerosol mass spectrometry, Atmos. Meas. Tech., 9, 23-39,
doi:10.5194/amt-9-23-2016, 2016.
Daellenbach, K. R., Stefenelli, G., Bozzetti, C., Vlachou, A., Fermo, P., Gonzalez, R.,
Piazzalunga, A., Colombi, C., Canonaco, F., Hueglin, C., Kasper-Giebl, A., Jaffrezo, J.-
L., Bianchi, F., Slowik, J. G., Baltensperger, U., El-Haddad, I., and Prévôt, A. S. H.:
Long-term chemical analysis and organic aerosol source apportionment at nine sites in
central Europe: source identification and uncertainty assessment, Atmos. Chem. Phys.,
17, 13265-13282, doi:10.5194/acp-17-13265-2017, 2017.
![Page 220: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/220.jpg)
203
Daellenbach, K. R., El-Haddad, I., Karvonen, L., Vlachou, A., Corbin, J. C., Slowik, J.
G., Heringa, M. F., Bruns, E. A., Luedin, S. M., Jaffrezo, J.-L., Szidat, S., Piazzalunga,
A., Gonzalez, R., Fermo, P., Pflueger, V., Vogel, G., Baltensperger, U., and Prévôt, A.
S. H.: Insights into organic-aerosol sources via a novel laser-desorption/ionization mass
spectrometry technique applied to one year of PM10 samples from nine sites in central
Europe, Atmos. Chem. Phys., in press, 2018.
Bozzetti, C., Daellenbach, K., R., Hueglin, C., Fermo, P., Sciare, J., Kasper-Giebl, A.,
Mazar, Y., Abbaszade, G., El Kazzi, M., Gonzalez, R., Shuster Meiseles, T., Flasch, M.,
Wolf, R., Křepelová, A., Canonaco, F., Schnelle-Kreis, J., Slowik, J. G., Zimmermann,
R., Rudich, Y., Baltensperger, U., El Haddad, I., and Prévôt, A. S. H.: Size-resolved
identification, characterization, and quantification of primary biological organic aerosol
at a European rural site, Environ. Sci. Technol., 50, 3425-3434,
doi:10.1021/acs.est.5b05960, 2016.
Bozzetti, C. El Haddad, I., Salameh, D., Daellenbach, K. R., Fermo, P., Gonzalez, R.,
Minguillón, M. C., Iinuma, Y., Poulain, L., Müller, E., Slowik, J. G., Jaffrezo, J.-L.,
Baltensperger, U., Marchand, N., and Prévôt, A. S. H.: Organic aerosol source
apportionment by offline-AMS over a full year in Marseille, Atmos. Chem. Phys., 17,
8247-8268, https://doi.org/10.5194/acp-17-8247-2017, 2017.
Bozzetti, C., Sosedova, Y., Xiao, M., Daellenbach, K. R., Ulevicius, V., Dudoitis, V.,
Mordas, G., Byčenkienė, S., Plauškaitė, K., Vlachou, A., Golly, B., Chazeau, B.,
Besombes, J.-L., Baltensperger, U., Jaffrezo, J.-L., Slowik, J. G., El Haddad, I., and
Prévôt, A. S. H.: Argon offline-AMS source apportionment of organic aerosol over
yearly cycles for an urban, rural, and marine site in northern Europe, Atmos. Chem.
Phys., 17, 117-141, https://doi.org/10.5194/acp-17-117-2017, 2017.
Elser, M., Huang, R.-J., Wolf, R., Slowik, J. G., Wang, Q., Canonaco, F., Li, G.,
Bozzetti, C., Daellenbach, K. R., Huang, Y., Zhang, R., Li, Z., Cao, J., Baltensperger,
U., El-Haddad, I., and Prévôt, A. S. H.: New insights into PM2.5 chemical composition
and sources in two major cities in China during extreme haze events using aerosol mass
![Page 221: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/221.jpg)
Curriculum vitae
204
spectrometry, Atmos. Chem. Phys., 16, 3207-3225, doi:10.5194/acp-16-3207-2016,
2016.
Klein, F., Platt, S. M., Farren, N. J., Detournay, A., Bruns, E. A., Bozzetti, C.,
Daellenbach, K. R., Kilic, D., Kumar, N. K., Pieber, S. M., Slowik, J. G., Temime-
Roussel, B., Marchand, N., Hamilton, J. F., Baltensperger, U., Prévôt, A. S. H., and El
Haddad, I.: Characterization of gas-phase organics using proton transfer reaction time-
of-flight mass spectrometry: cooking emissions, Environ. Sci. Tech., 50, 1243-1250,
doi:10.1021/acs.est.5b04618, 2016
Klein, F., Farren, N. J., Bozzetti, C., Daellenach, K. R., Kilic, D., Kumar, N. K., Pieber,
S. M., Slowik, J. G., Tuthill, R. N., Hamilton, J. F., Baltensperger, U., Prévôt, A. S. H.,
El and Haddad, I.: Indoor terpene emissions from cooking with herbs and pepper and
their secondary organic aerosol production potential, Sci. Rep., 6, 36623,
doi:10.1038/srep36623, 2016.
Krapf, M., El Haddad, I., Bruns, E. A., Molteni, U., Daellenbach, K. R., Prévôt, A. S.
H., Baltensperger, U., Dommen, J.: Labile peroxides in secondary organic aerosol,
Chem, 1, 603-616, doi:10.1016/j.chempr.2016.09.007, 2016.
Pieber, S. M., El Haddad, I., Slowik, J. G., Canagaratna, M. R., Jayne, J. T., Platt, S. M.,
Bozzetti, C., Daellenbach, K. R., Fröhlich, R., Vlachou, A., Klein, F., Dommen, J.,
Miljevic, B., Jimenez, J. L., Worsnop, D. R., Baltensperger, U., and Prévôt, A. S. H.:
Inorganic salt interference on CO2+ in Aerodyne AMS and ACSM organic aerosol
composition studies, Environ. Sci. Technol, 50, 10494-10503,
doi:10.1021/acs.est.6b01035, 2016.
Platt, S. M., El Haddad, I., Pieber, S. M., Zardini, A. A., Suarez-Bertoa, R., Clairotte,
M., Daellenbach, K. R., Huang, R.-J., Slowik, J. G., Hellebust, S., Temime-Roussel, B.,
Marchand, N., de Gouw, J., Jimenez, J. L., Hayes, P. L., Robinson, A. L.,
Baltensperger, U., Astorga, C., and Prévôt, A. S. H.: Gasoline cars produces more
carbonaceous particulate matter than modern filter-equipped diesel cars, 7, 4926,
doi:10.1038/s41598-017-03714-9, 2017.
![Page 222: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/222.jpg)
205
Ulevicius, V., Byčenkienė, S., Bozzetti, C., Vlachou, A., Plauškaitė, K., Mordas, G.,
Dudoitis, V., Abbaszade, G., Remeikis, V., Garbaras, A., Masalaite, A., Blees, J.,
Fröhlich, R., Dällenbach, K. R., Canonaco, F., Slowik, J. G., Dommen, J.,
Zimmermann, R., Schnelle-Kreis, J., Salazar, G. A., Agrios, K., Szidat, S., El Haddad,
I., and Prévôt, A. S. H.: Fossil and non-fossil source contributions to atmospheric
carbonaceous aerosols during extreme spring grassland fires in Eastern Europe, Atmos.
Chem. Phys., 16, 5513-5529, doi:10.5194/acp-16-5513-2016, 2016.
Zhang, Y.-L., Huang, R.-J., El Haddad, I., Ho, K.-F., Cao, J.-J., Han, Y., Zotter, P.,
Bozzetti, C., Daellenbach, K. R., Canonaco, F., Slowik, J. G., Salazar, G.,
Schwikowski, M., Schnelle-Kreis, J., Abbaszade, G., Zimmermann, R., Baltensperger,
U., Prévôt, A. S. H., and Szidat, S.: Fossil vs. non-fossil sources of fine carbonaceous
aerosols in four Chinese cities during the extreme winter haze episode of 2013, Atmos.
Chem. Phys., 15, 1299-1312, doi:10.5194/acp-15-1299-2015, 2015.
Huang, R.-J., Zhang, Y., Bozzetti, C., Ho, K.-F., Cao, J., Han, Y., Dällenbach, K. R.,
Slowik, J. G., Platt, S. M., Canonaco, F., Zotter, P., Wolf, R., Pieber, S. M., Bruns, E.
A., Crippa, M., Ciarelli, G., Piazzalunga, A., Schwikowski, M., Abbaszade, G.,
Schnelle-Kreis, J., Zimmermann, R., An, Z., Szidat, S., Baltensperger, U., Haddad, I. E.,
and Prévôt, A. S. H.: High secondary aerosol contribution to particulate pollution during
haze events in China, Nature, 514, 218-222, doi:10.1038/nature13774, 2014.
Wang, Y. C., Huang, R.-J., Ni, H. Y., Chen, Y., Wang, Q. Y., Li, G. H., Tie, X. X.,
Shen, Z. X., Huang, Y., Liu, S. X., Dong, W. M., Xue, Fröhlich, R., Canonaco, F.,
Elser, M., Daellenbach, K. R., Bozzetti, C., El Haddad, I., A. S. H. Prévôt, Canagaratna,
M. R., Worsnop, D. R., and Cao, J. J.: Chemical composition, sources and secondary
processes of aerosols in Baoji city of northwest China, Atmos. Environ. 158, 128-137,
doi: 10.1016/j.atmosenv.2017.03.026, 2017.
Zotter, P., Ciobanu, V. G., Zhang, Y. L., El-Haddad, I., Macchia, M., Daellenbach, K.
R., Salazar, G. A., Huang, R.-J., Wacker, L., Hueglin, C., Piazzalunga, A., Fermo, P.,
Schwikowski, M., Baltensperger, U., Szidat, S., and Prévôt, A. S. H.: Radiocarbon
![Page 223: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/223.jpg)
Curriculum vitae
206
analysis of elemental and organic carbon in Switzerland during winter-smog episodes
from 2008 to 2012 - Part 1: Source apportionment and spatial variability, Atmos. Chem.
Phys., 14, 13551-13570, doi:10.5194/acp-14-13551-2014, 2014.
![Page 224: OPTIMIZED SEPARATION OF OC AND EC FOR RADIOCARBON … · 2018-11-06 · ix Summary Particulate matter (PM), liquid or solid particles suspended in the atmosphere, contributes significantly](https://reader030.vdocument.in/reader030/viewer/2022040907/5e7d8ec03764cd507759a395/html5/thumbnails/224.jpg)