Atmos. Chem. Phys., 11, 8977–8993, 2011
www.atmos-chem-phys.net/11/8977/2011/
doi:10.5194/acp-11-8977-2011
© Author(s) 2011. CC Attribution 3.0 License.
Atmospheric
Chemistry
and Physics
Technical Note: In-situ derivatization thermal desorption
GC-TOFMS for direct analysis of particle-bound non-polar and
polar organic species
J. Orasche1,2,3, J. Schnelle-Kreis2, G. Abbaszade2, and R. Zimmermann2,3
1Department of Sedimentology & Environmental Geology, Georg-August-University, Go¨ttingen, Germany
2Joint Mass Spectrometry Centre – Cooperation Group “Comprehensive Molecular Analytics”, Helmholtz Zentrum
Mu¨nchen, Neuherberg, Germany
3Joint Mass Spectrometry Centre – Institute of Chemistry, Division of Analytical and Technical Chemistry, University of
Rostock, Rostock, Germany
Received: 21 April 2011 – Published in Atmos. Chem. Phys. Discuss.: 19 May 2011
Revised: 15 August 2011 – Accepted: 29 August 2011 – Published: 2 September 2011
Abstract. An in-situ derivatization thermal desorption
method followed by gas chromatography and time-of-flight
mass spectrometry (IDTD-GC-TOFMS) was developed for
determination of polar organic compounds together with
non-polar compounds in one measurement. Hydroxyl and
carboxyl groups of compounds such as anhydrous sug-
ars, alcohols and phenols, fatty acids and resin acids
are targets of the derivatization procedure. Derivatiza-
tion is based on silylation with N-Methyl-N-trimethylsilyl-
trifluoroacetamide (MSTFA) during the step of thermal des-
orption. The high temperature of 300 ◦C during desorp-
tion is utilized for the in-situ derivatization on the collection
substrate (quartz fibre filters) accelerating the reaction rate.
Thereby, the analysis time is as short as without derivatiza-
tion. At first the filter surface is dampened with derivatization
reagent before insertion of the sample into the thermal des-
orption unit. To ensure ongoing derivatization during ther-
mal desorption the carrier gas is enriched with MSTFA until
the desorption procedure is finished. The precisions of all
studied analytes were below 17 % within a calibration range
from 22 pg (abietic acid) up to 342 ng (levoglucosan). Lim-
its of quantification (LOQ) for polycyclic aromatic hydro-
carbons (PAH) were between 1 pg (fluoranthene) and 8 pg
(indeno[1,2,3-cd]pyrene), for resin acids 37–102 pg and for
studied phenols 4–144 pg. LOQ for levoglucosan was 17 pg.
Correspondence to: J. Schnelle-Kreis
(juergen.schnelle@helmholtz-muenchen.de)
1 Introduction
During the last decade direct thermal desorption (DTD)
(someone like to prefer the term “extraction” which im-
plies the removal of analytes from a native matrix) methods
were developed for quantification of semi volatile organic
compounds (SVOC) adsorbed on ambient particulate matter
(PM) (Falkovich, 2001; Waterman, 2000). The advantage of
DTD is the reduction of analytes losses and memory effects
by desorption of the sample within the GC injector. Desorp-
tion inside of the injector with a short way to the separation
column prevents cold spots. Combined with the possibility of
using a fresh glass liner for every sample memory effects are
avoided and analytes losses are minimized. In recent years an
increasing number of samples had to be analysed for study-
ing aerosol composition. A growing variety of organic com-
pounds becomes more and more important for source appor-
tionment (Dutton, 2009; Sklorz, 2007; Vedal, 2009), aerosol
ageing studies and investigations on characteristics of parti-
cles responsible for climatic effects (Donahue, 2009). There-
fore direct thermal desorption hyphenated with GC-MS can
be a useful tool to deal daily sampling on long time series
(Schnelle-Kreis, 2005a) or with high time resolution. Some
research groups developed methods for direct thermal des-
orption for analysis of organic aerosol compounds in the past
(Bates, 2008; Ding, 2009; Falkovich, 2001; Gil-Molto´, 2009;
Hays, 2003; Ho, 2008; Schnelle-Kreis, 2005b; van Drooge,
2009). Extensive reviews about thermal methods for chemi-
cal characterization of carbonaceous aerosols were published
by Hays and Lavrich (2007) and Chow et al. (2007).
Published by Copernicus Publications on behalf of the European Geosciences Union.
8978 J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS
A recent development in thermal desorption techniques
is the Thermal Desorption Aerosol GC-MS (TAG) (Lambe,
2010; Williams, 2006, 2010). Particles with a cut-point of
PM2 are collected by humidifying and impaction into a ther-
mal desorption cell. After a defined sampling time the or-
ganic matter is transferred to the gas chromatograph by ther-
mal desorption. A high time resolution of one hour per sam-
ple and the chromatographic separation are the advantages of
the TAG system.
Another development for thermal desorption of precip-
itated particulate matter is the methylation by tetramethy-
lammonium hydroxide (TMAH) and its derivatives (Beiner,
2009; Fabbri et al., 2002) for quantification of organic acids.
The applicability for other polar compounds is currently un-
der investigation. Furthermore an in-situ methylation of
carboxylic acids by treatment of filter samples with dia-
zomethane was introduced by Sheesley et al. (2010).
Polar organic substances are playing a major role in char-
acterization of diverse atmospheric processes and also during
formation of aerosols. The main organic combustion prod-
uct of wood is the anhydrous sugar levoglucosan as a de-
composition product of cellulose. This compound as well
as its homologues mannosan and galactosan, decomposi-
tion products of hemi cellulose, are ubiquitous constituents
of the atmosphere (Simoneit, 1999, 2004). Further wood
combustion products originating from lignin breakdown or
colophony can be observed. Most of them are of polar nature
(Bari, 2009; Fine, 2001, 2002 and 2004; Nolte, 2001). An-
other main source of polar organic compounds in the atmo-
sphere are biogenic emissions. Not only sugars from plants,
fungi and bacteria can be observed (Medeiros et al., 2006;
Medeiros and Simoneit, 2007), but also huge amounts of un-
saturated hydrocarbons are continuously evaporated by veg-
etation (Guenther et al., 1995; Kourtchev et al., 2005). Once
organic substances are released to the atmosphere as gas or
particle-bound they are exposed to UV radiation, radicals and
oxidants. Depending on particle properties (e.g. pH value,
possible reaction agents) reaction mechanisms and reaction
rates are affected. In ageing studies these facts are taken
into account to investigate the formation of secondary or-
ganic aerosol (Hallquist et al., 2009). All of these reactions
steadily increase the polarity of the reactants. Hence it is
essential to implement an appropriate analytical method for
observation of the polar organic tracers described and their
behaviour immediately after combustion processes or in the
atmosphere (Edney et al., 2003; Oliveira et al., 2007), e.g.
supported by chamber experiments (Chiappini et al., 2006;
Edney, 2005). Investigation of polar organic compounds is of
great scientific interest due to their effect on particulate prop-
erties impacting climate processes. An increasing oxidation
rate leads to acidification and growing of particles. The on-
going ageing is responsible for transformation of particles to
cloud condensation nuclei (CCN).
In this work an in-situ derivatization thermal desorption
method followed by gas chromatography and time-of-flight
mass spectrometry (IDTD-GC-TOFMS) is introduced com-
bining short sampling time, powerful chromatographic sepa-
ration and determination of polar organic and non-polar or-
ganic substances in one measurement. Like DTD IDTD is a
less time consumption method with relatively low costs com-
pared to solvent extracting methods followed either by HPLC
or GC. Moreover the possibility of in-situ-derivatization is
something like a front-end for already installed DTD sys-
tems. Although it is not necessary to derivatize compo-
nents when working with HPLC it is not necessary to pre-
treat samples in any extensive way when using IDTD. A
very interesting technique without derivatization procedure
is a fast two-dimensional gas chromatography method that
uses heart-cutting and thermal extraction (TE-GC-GC-MS).
Yet it seems to be established for anhydrous sugars, n-
alkanoic acids, substituted phenols, and nitrogen-bearing
heterocyclics (Ma et al., 2008 and 2010). The combination
of a non-polar separation phase with a polar one provides
quantifiable sharp peaks in the chromatograms. Neverthe-
less compared to GC systems with highly polar separation
columns IDTD is a more sensitive way to transfer delicate
molecules from injector to column.
For inter comparison NIST Standard Reference Material
Urban Dust SRM 1649a (National Institute of Standards and
Technology, USA) was employed to validate the method for
analysis of polycyclic hydrocarbons (Bates, 2008; Falkovich,
2001; Gil-Molto´, 2009; Ho, 2008; van Drooge, 2009; Wa-
terman, 2000, 2001). Levoglucosan, mannosan and galac-
tosan were determined in the standard reference material and
compared to results already published by Kuo et al. (2008),
Larsen et al. (2006) and Louchouarn et al. (2009). Com-
parison of different methods for analysis of ambient aerosol
samples was carried out, too, (Ho, 2004, 2008; Ma, 2010;
van Drooge, 2009) concerning polar substances, polycyclic
aromatic hydrocarbons (PAH) and oxidized polycyclic hy-
drocarbons (o-PAH). PAH and even more o-PAH are deli-
cate analytes due to generate artefacts during sampling, sam-
ple preparation and analysis (Liu, 2006). Therefore, analy-
sis methods for these substance classes have to be validated
properly. The direct thermal desorption (DTD) method pre-
sented here as a reference method was already introduced in
former papers by Schnelle-Kreis et al. (2005a, b, 2007). Here
the accuracy of this method is demonstrated briefly.
2 Experimental
2.1 Sampling of ambient aerosol
Prior to sampling quartz fibre filters were baked for at least
eight hours at 550 ◦C to remove all organic matter. The
sampler was located at the aerosol characterization site of
the Helmholtz Zentrum Mu¨nchen at the University of Ap-
plied Sciences, Augsburg, next to the inner city of Augsburg,
Atmos. Chem. Phys., 11, 8977–8993, 2011 www.atmos-chem-phys.net/11/8977/2011/
J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS 8979
Germany (UAS, urban background). Samples were taken
from 1 March 2010 to 14 March 2010.
For chemical analysis PM2.5 samples were collected with
a low volume sequential sampler (Partisol-Plus Model 2025,
Rupprecht & Patashnick, NY, USA) on quartz fibre fil-
ters (T293, Munktell, Grycksbo, Sweden) at a flow rate of
16.7 l min−1. Sampling time was 24 hours, thus airborne par-
ticulate matter of 24 m3 of air was collected on each filter.
The samples were stored in glass containers at −18 ◦C until
analysis. Filters for solvent extraction were extracted directly
after sampling.
2.2 Basics of In-situ Derivatization and Thermal
Desorption (IDTD)
The enhanced IDTD method involves the option to quantify
polar constituents of organic particulate matter. The method
presented allows derivatization of the polar organic fraction
of filter samples in-situ on the filter during thermal desorp-
tion. Derivatization and desorption of polar organic com-
pounds occurs directly from particulate matter on the filters.
The advantages of a direct thermal desorption system em-
ployed for this study are: (1) direct placement of samples in
the GC liner, so called in-injector thermal desorption avoid-
ing sample transfer lines frequently causing cold spots, (2)
the possibility of automatic liner exchange, (3) rapid heating
rates of the injector and (4) replacement of the GC liner for
every sample.
The derivatization procedure consists of two steps and is
fully automated with an applicable sampling robot. The first
step is the addition of liquid derivatization reagent N-Methyl-
N-trimethylsilyl-trifluoroacetamide (MSTFA) by a syringe
for complete moistening of filter and particulate matter. The
second step is very important for quantitative silylation of
polar organic compounds by MSTFA. The carrier gas is sat-
urated with derivatization reagent by switching to a second
pathway leading over a cartridge filled with MSTFA during
the thermal desorption process (Fig. 1). The high tempera-
ture of 300 ◦C during desorption increases derivatization rate
sufficiently without additional catalyst.
The in-situ derivatization thermal desorption GC-TOFMS
(IDTD-GC-TOFMS) was employed for the main anhydrous
sugars levoglucosan, galactosan and mannosan occurring in
ambient aerosol and further highly polar organic compounds
like resin acids and lignin combustion products. Analy-
sis of levoglucosan is possible by many different methods
published in a variety of papers. An overview of detection
and quantification of levoglucosan in atmospheric aerosols
is given by the review of Schkolnik and Rudich (Schkolnik,
2006). As a conclusion of these studies we can find that
the choice of solvent or solvent mixture is most crucial for
methods using solvent extraction (Louchouarn, 2009). The
polarity of the extraction solvent constitutes a limiting fac-
tor for the extraction yield of levoglucosan. To determine
all substance classes being of interest for studying ambient
GC
Injector
Split
Glass wool
 plug
Goose
neck
liner
TOF-MS
Filter aliquot
Carrier gas
MSTFA-
Cartridge
DAQ
N
O
CF3
CH3
Si(CH3)3 ROH+
N
O
CF3
CH3
H ROSi(CH3)3+
Fig. 1. In-situ derivatization and thermal desorption unit. The car-
tridge with an glas insert filled with MSTFA is localized directly
before the thermal desorption unit. On the top of the figure: chem-
ical equation of the silylation reaction of polar organic compounds
with MSTFA.
aerosol it may be necessary to extract different polar fractions
by different solvents. The advantage of the described in-situ
derivatization thermal desorption method is minimization of
the working steps, avoiding handling with different solvent
mixtures and making even highly polar substances ascer-
tainable. An overview of the compound classes analysed is
shown in Fig. 2 in terms of single ion chromatograms.
2.3 Sample preparation, In-situ Derivatization and
Thermal Desorption
For analysis of particulate matter collected the filters were
cut by a special tool into filter aliquots which were stripes
of the dimension 13.5 mm×2 mm. One filter stripe (sam-
ple aliquot of 0.55 m3 of sampled air) was spiked with two
internal standard mixtures (isotope-labelled reference com-
pounds) for quantification. The first (non polar) internal stan-
dard consisted of fifteen deuterated PAH, two deuterated o-
PAH and four deuterated alkanes. D8-9,10-anthracenedione
and D10-benz[a]anthracene-7,12-dione were synthesized in
our laboratory (Liu, 2006) (see Table S1 in the supplemen-
tary information for detailed description and concentrations).
The second (polar) internal standard mixture contained 13C6-
levoglucosan (Omicron Biochemicals, USA), 13C6-vanillin
(Larodan, Sweden) and D31-palmitic acid (CIL, USA). Sam-
ples were placed into goose-neck glass-liners for thermal
desorption which were sealed with PTFE caps. For each
sample a freshly deactivated GC liner was used. Deactiva-
tion was done by annealing the liners at 550 ◦C for at least
twelve hours followed by a derivatization treatment of the
glass surface with chlorotrimethylsilan (TMCS, Merck, Ger-
many) for further twelve hours. A glass wool plug above
www.atmos-chem-phys.net/11/8977/2011/ Atmos. Chem. Phys., 11, 8977–8993, 2011
8980 J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS
Fluoranthene
Pyrene Chrysene /
TriphenyleneBenz[a]-
anthracene Benzo-
fluoranthenes
Benz[e]-
pyrene
Benz[a]-
pyrene
Perylene
Dibenz[ah]-
anthracene
Indeno-
[1,2,3-cd]
pyrene
Benzo[ghi]-
perylene
Mannosan
Galactosan
Levoglucosan
Dehydroabietic
acid methyl ester
Isopimaric acid Abietic acid
7-Oxodehydroabietic acid /
7-Oxodehydroabietic acid methyl ester
9H-Fluoren-9-one Anthracene-
9,10-dione
1,8-Naphthalic 
anhydride
Benz[a]-
fluoren-
11-one
Benz[b]-
fluoren-
11-one 7H-Benz[de]-
anthracene-7-one
Benz[a]-
anthracene-7,12-dione
Syringaldehyde
Vanillic acid
Syringic acid Divanillyl
Disyringyl
Acetosyringone
Syringylacetone
Shonanin
Matairesinol
a)
b)
c)
d)
e)
PAH
Anhydrous sugars
o-PAH
Resin acids
Lignin
degrading
products
C20H42
C21H44
C22H46
C23H48
C24H50
C25H52C26H54
C27H56 C28H58
C29H60
C30H62
C31H64
C32H66
C33H68
Stearic acid
Palmitic acid
f)
Alkanes and Fatty acids
Dehydroabietic acid
Fig. 2. Multiple ion chromatograms reflecting different groups of polar and non-polar analytes from a sample collected at the aerosol
characterization site at the UAS (urban background): (a) polycyclic aromatic hydrocarbons (PAH) (m/z 178+202+228+252+276+278), (b)
oxidized PAH (o-PAH) (m/z 126+158+180+196+230+258), (c) anhydrous sugars (m/z 204+217), (d) resin acids (m/z 239+241+253), (e)
lignin degradation products (m/z 194+209+224+238+279+297+327), (f) alkanes and fatty acids (m/z 57+117).
Atmos. Chem. Phys., 11, 8977–8993, 2011 www.atmos-chem-phys.net/11/8977/2011/
J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS 8981
the goose-neck prevented particles from the filter to enter the
capillary column.
For analysis of the NIST Standard Reference Material Ur-
ban Dust 1649a portions of 0.06 mg were thermally des-
orbed. Due to the small amount required the urban dust was
diluted and homogenized first with an inert matrix of sodium
sulphate (w/w, 1/1000) which was annealed and ground be-
fore. The GC injection liners were loaded with 60 mg of
this mixture. Isotope-labelled standards were added to this
mixture. The liners were also sealed with PTFE caps until
analysis.
Further treatments were carried out by a sampling robot
(Focus, Atas GL, Netherlands). Prior to analysis liners were
opened and 10 µl MSTFA (Macherey-Nagel, Germany) was
added in each liner to moisten the sample surface. Liners
were placed into a direct (in-injector) thermal desorption unit
(Optic 3, Atas GL, Netherlands) mounted on the gas chro-
matograph. The sampling robot exchanged the complete GC
liners placed in the injector which was automatically closed
and opened by a liner exchanging unit (Linex, Atas GL,
Netherlands). After liners were placed into the cold injec-
tor (60 ◦C) a steady flow of carrier gas (helium) removed
the air. During this venting step of 60 seconds carrier gas
flow was 0.7 ml min−1, split flow was 50 ml min−1. After
180 s carrier gas flow was increased split less to 4 ml min−1.
The carrier gas was lead to a bypass with a cartridge filled
with MSTFA before entering the injector. Immediately af-
ter the flow reached the desired value the bypass was opened
via two 3-port/2-way solenoid valves being controlled by the
Optic software and hardware unit (Atas GL, Netherlands).
The injector was heated up to 300 ◦C with a heating rate
of 2 ◦C s−1directly after bypass was opened. During 16 min
of reaction and desorption time the carrier gas was contin-
uously saturated with derivatization reagent (Fig. 1). Sub-
sequently the bypass was closed and the column flow was
set to 0.7 l min−1 with a split flow of 50 l min−1directly after
thermal desorption was finished.
2.4 Gas chromatography and time-of-flight mass
spectrometry
Desorbed molecules were focused at 70 ◦C on the head of
the capillary column, BPX5, 25 m, 0.22 mm ID, 0.25 µm
film (SGE, Australia) which was installed in an Agilent 6890
gas chromatograph (Agilent, USA). The thermal desorption
step was followed by heating up the GC oven to 130 ◦C with
a rate of 80 ◦C min−1. Then the rate was lowered to 8 ◦C
min−1 until a temperature of 330 ◦C was reached followed
by an isothermal time of 30 min.
Identification and quantification of target compounds were
carried out on a Pegasus III TOFMS using Chroma TOF soft-
ware package (LECO, USA) being capable of peak deconvo-
lution. The data acquisition range was m/z 35 to 500 with
an acquisition frequency of 25 spectra per second which is
necessary for reliable peak deconvolution.
Analytes which were not available as standard in analyti-
cal grade were identified by their mass spectra and the reten-
tion time index. In those cases quantification was achieved
with an adequate surrogate standard on semi quantitative ba-
sis. Surrogates are specified in Table S1.
2.5 Calibration of thermal desorption
The calibration of the thermal desorption method was car-
ried out applying standard addition to reference filters (PM
samples). The standard addition method was used here to
minimize the matrix effects. The reference samples were col-
lected at the aerosol characterization site in Augsburg, Ger-
many, to obtain references with similar matrix to other field
samples. These filters were spiked with internal standard and
derivatization standard respectively. Standard mixtures with
native organic compounds were added to the reference fil-
ters in different concentration levels (for calibration ranges
see Table 1). Further filter treatment was identical to that
described in section 2.3.
2.6 Direct Thermal Desorption (DTD)
DTD was described in detail elsewhere (Schnelle-Kreis,
2005a, b, 2007). The pre-treatment of filters and liners for
DTD was identical as described above without adding the
isotope-labelled standard mixture of polar compounds. The
used instrument is identical to that used for IDTD. The vent
time before thermal desorption was shorter (one minute) than
in IDTD method and the carrier gas was not treated with
derivatization reagent during thermal desorption process. In
that case the valves stayed closed and no MSTFA was added.
2.7 Solvent extraction of non-polar compounds
The method for solvent extraction of non-polar compounds
was described elsewhere (Liu, 2006; Sklorz, 2007). Ex-
traction was carried out utilizing soxhlet extraction with
dichloromethane as solvent. Therefore, filters or a mixture
of SRM 1649a and sodium sulphate were placed in fritted
soxhlet sleeves made of glass. Soxhlet extraction time was
16 h with at least six cycles per hour. Extracts were dried
over pre-baked sodium sulphate and filtrated to remove filter
residues. Extracts were concentrated to 1 ml and additionally
cleaned up and fractionated on a liquid chromatography col-
umn packed with silica (Promochem, Germany) deactivated
with three percent of water. First alkanes were eluted from
the column by a solvent mixture of hexane/dichloromethane
(9:1, v/v). In the second and third fraction PAH and o-PAH
were eluted after each other by hexane/dichloromethane (1:1,
v/v) and dichloromethane/methanol (19:1, v/v) respectively
(all solvents: Merck, Germany). The mixtures were concen-
trated to adjust the concentrations of the analytes to the same
range as applied for DTD (one injection equates to 1 m3 of
sampled air). Samples were analysed by GC-HRMS.
www.atmos-chem-phys.net/11/8977/2011/ Atmos. Chem. Phys., 11, 8977–8993, 2011
8982 J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS
Table 1. Validation parameters obtained from the calibration curves of IDTD-GC-TOF MS. Calibration range means lowest and highest
standard concentrations used for the calibration curve; a = interception, b = slope (area ratio/mass ratio). The limit of quantification (LOQ)
of the method is defined as the minimum amount of substance that is according to the minimum reliable signal plus nine times (three
times for LOD) the standard deviation of this underground signal. The complementary information of SE-GC-TOFMS can be found in the
supplementary information (Table S2).
Analyte Calibration range [ng] a b R2 Precision LOQ per sample [ng]
Phenanthrene 0.790–39.5 0.044 1.148 0.999 3 % 0.003
Anthracene 0.212–10.6 −0.132 1.789 0.983 15% 0.003
Fluoranthene 0.283–14.1 0.050 1.403 0.998 5 % 0.001
Pyrene 0.392–19.6 0.007 1.315 1.000 1 % 0.005
Benzo[a]anthracene 0.290–14.5 −0.217 1.871 0.997 7 % 0.003
Chrysene 0.277–13.8 0.018 0.534 0.998 5 % 0.004
sum Benzofluoranthenes 0.098–4.89 0.017 0.805 1.000 2 % 0.002
Benzo[e]pyrene 0.134–6.71 −0.021 1.523 1.000 2 % 0.003
Benzo[a]pyrene 0.130–6.49 −0.010 1.384 0.999 4 % 0.005
Perylene 0.070–3.52 −0.092 1.510 0.996 7 % 0.004
Indeno[1,2,3-cd]pyrene 0.069–3.44 0.139 0.850 0.996 6 % 0.008
Dibenzo[ah]anthracene 0.072–3.60 0.859 1.288 0.986 7 % 0.006
Benzo[ghi]perylene 0.073–3.63 0.030 1.833 0.999 3 % 0.004
Coronene 0.072–3.60 -0.065 1.175 0.996 7 % 0.002
9H-Fluoren-9-one 0.362–18.1 −0.100 0.826 0.977 17 % 0.005
9,10-Anthracenedione 0.280–14.0 −0.012 0.485 1.000 2 % 0.028
1,8-Naphthalic anhydride 0.414–20.7 0.215 0.613 0.993 9 % 0.026
Cyclopenta[def]phenanthrenone 0.075–3.74 −0.058 0.479 0.987 17% 0.012
11H-Benzo[a]fluoren-11-one 0.081–4.06 0.016 0.157 0.988 12 % 0.033
11H-Benzo[b]fluoren-11-one 0.079–3.93 0.010 0.638 0.990 11 % 0.008
Benzo[a]anthracene-7,12-dione 0.37–18.5 −0.704 3.026 0.995 9 % 0.229
Malic acid 0.976–97.6 0.143 0.033 0.931 10 % 1.59
Vanillin 0.057–5.67 −0.016 1.077 0.985 10 % 0.144
Galactosan 0.221–22.1 0.046 1.288 0.992 11 % 0.048
Mannosan 1.05–105 0.101 0.575 0.981 14 % 0.033
Levoglucosan 3.42–342 0.405 0.495 0.994 6 % 0.017
Phthalic acid 0.604–60.4 −0.003 0.239 0.983 14% 0.054
Vanillic acid methyl ester 0.054–5.44 −0.040 2.539 0.986 17 % 0.067
Vanillic acid 0.054–5.38 0.016 1.170 0.987 13 % 0.047
Syringic acid 0.053–5.34 −0.002 2.371 0.980 16 % 0.025
Syringyl aldehyde 0.036–3.64 0.019 2.081 0.992 10 % 0.088
Acetosyringone 0.036–3.64 0.019 2.081 0.992 10 % 0.025
Syringylacetone 1.17–117 −0.050 2.951 0.996 9% 0.004
Retene 0.290–14.5 −0.217 1.871 0.997 7 % 0.008
Isopimaric acid 0.142–14.2 0.043 0.155 0.997 12 % 0.037
Dehydroabietic acid methyl ester 0.392–19.6 0.007 1.315 1.000 1 % 0.005
Dehydroabietic acid 0.142–14.2 0.074 0.267 0.997 9 % 0.030
Abietic acid 0.022–2.20 0.023 0.557 0.983 14 % 0.102
7-Oxodehydroabietic acid 0.142–14.2 0.043 0.155 0.997 12 % 0.037
Divanillyl 0.392–19.6 0.007 1.315 1.000 1 % 0.014
The chromatographic separation for solvent extracted (SE)
samples was carried out on a Varian GC 3400 (Varian, USA)
assembled with a retention gap (guard column), deactivated
fused silica, 2.5 m, 0.22 mm ID (SGE) and a BPX5 column,
25 m, 0.22 mm ID, 0.25 µm film (SGE). The sector field mass
spectrometer MAT95 (Thermo Scientific, Germany) was op-
erated in multiple ion detection mode (MID) for target anal-
ysis.
2.8 Solvent extraction of polar compounds
Solvent extraction of polar compounds was carried out with
dichloromethane/methanol (1:1, v/v) in an ultrasonic bath.
Prior to extraction the samples were spiked with internal
standard mixtures. Ultrasonication was carried out three
times with five millilitres of solvent for fifteen minutes each.
The three extracts were combined and filtered over PTFE
Atmos. Chem. Phys., 11, 8977–8993, 2011 www.atmos-chem-phys.net/11/8977/2011/
J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS 8983
syringe membrane filters (0.2 µm, Sartorius, Germany). Sol-
vent was evaporated to dryness. Derivatization was started
by adding MSTFA to the samples. Reaction time was 3 h at
80 ◦C. Samples were measured with the same GC-TOFMS
equipment as described above for IDTD-GC-TOFMS which
was also used for thermal desorption. The injector em-
ployed was also capable of liquid injection (here one injec-
tion equates to 0.5–1 m3 of sampled air). The GC method
was programmed as follows: Injector temperature: 300 ◦C,
Oven temperature: 100 ◦C for 1 min. Heating rates were
25 ◦C min−1 to 175 ◦C followed by a rate of 5 ◦C min−1 to
330 ◦C with an isothermal of 15 min at 330 ◦C at the end of
the run.
Calibration for liquid extraction samples was carried
out by evaporating different standard dilutions to dryness.
MSTFA was added directly to the dried standard. The fur-
ther procedure was like the sample treatment.
3 Results and discussion
3.1 Derivatization
MSTFA as silylation reagent for the IDTD method has the
advantage that it is not necessary to be removed from the
sample before GC-MS analysis. MSTFA itself and the prod-
ucts of MSTFA reactions exhibit high vapour pressures.
Even at room temperature a significant enrichment of carrier
gas with MSTFA is possible. On non-polar capillary columns
MSTFA has characteristics like many other solvents being
used for solvent injection in GC-MS. Thus short retention
times with sharp solvent peaks without carry over are achiev-
able. At usual conditions (e.g. temperature of 80 ◦C) MSTFA
is reacting relatively slow. Sterically hindered molecules ex-
hibit poor reaction yields with MSTFA. For that reason a cat-
alyst like chlorotrimethylsilan (TMCS) which acts as Lewis
acid may be used to accelerate the reaction. Combinations
such as BSTFA (N,N’-bistrimethylsilyl-trifluoroacetamide)
and TMCS (99/1, v/v) or MSTFA and TMCS (99/1, v/v) are
widely used for derivatization particularly for determination
of levoglucosan and other polar compounds in atmospheric
aerosol (Nolte, 2001; Simoneit, 1999; Zdrahal, 2002). Due
to a large polar organic fraction and inorganic salts ambient
aerosol samples also contain water. Water is also involved in
reactions with silylation reagents. This leads to two essen-
tials for derivatization reaction with ambient aerosol: first,
a sufficient surplus of derivatization agent is necessary and
second, TMCS has to be excluded for in-situ reactions in the
gas chromatography system. Poor column life time could
be a result of high reaction yields of HCl formed by wa-
ter and TMCS. To avoid possible degradation reactions of
derivatized compounds with water a surplus of MSTFA is
necessary. This is ensured when working with dampen of
filters followed by enrichment of carrier gas with MSTFA.
A high water content of samples collected on quartz fibre
filters was visible when working without solvent delay dur-
ing mass spectrometric detection. The mass signals dur-
ing thermal desorption were indicated by a large signal at
m/z 147 caused by hexamethyldisiloxane. Without a sur-
plus of MSTFA the water content could be responsible for
a competitive reaction next to derivatization of analytes and
water can also react with already derivatized analytes during
derivatization/desorption process.
To accelerate the reaction an increase of temperature can
be applied instead of catalysts. In the presented method
the high temperatures of thermal desorption increase reac-
tion rate when working with MSTFA. IDTD utilized only
16 minutes of desorption time for derivatization reaction.
The same reaction yields were obtained with the SE method
when derivatization was performed for 3 h. A verification of
derivatization yields was done by comparison of responses of
isotope-labelled standards generated with the IDTD method
and with directly derivatized standard solution as used for the
here described calibration curves of SE methods. Therefore,
a recovery efficiency standard (D42-eicosane) was added to
the samples prior to injection to normalise abundance of
polar standard compounds by this non-polar standard com-
pound which was even not affected by sample pre-treatment.
Assuming that derivatization reactions in solution were com-
plete (100 %) the yields for IDTD were calculated relative to
yields of derivatization in solution. The derivatization yields
of vanillin (111 %), levoglucosan (97 %) and palmitic acid
(107 %) suggest that reactions of both methods were compa-
rable.
It was found that reactions take place in multiple ways.
Therefore the two step derivatization procedure was devel-
oped. For the silylation of multifunctional molecules or poly-
cyclic polar compounds like anhydrous sugars on the one
side or sterols on the other side the soak of sample with
MSTFA is necessary. Some phenols (e.g. sinapinic acid)
were only accessible when working with enrichment of car-
rier gas with MSTFA. Resin acids were found to be only ac-
cessible when working with both steps of the derivatization
procedure. Experiments with a direct inlet to a mass spec-
trometer will provide clearness about formation of reaction
products and possible competitive or degradation reactions.
The first derivatization step (the moistening of the fil-
ter with MSTFA) was introduced as described before to
start the reaction and to protect the compounds until the
MSTFA-saturated helium enters the injection port of the
gas chromatograph. The enrichment of carrier gas with
volatile derivatization reagent has several advantages. (1) A
steady flow of MSTFA during thermal desorption protects
derivatized polar organic compounds from degradation un-
til the transfer to the gas chromatograph is completed. (2)
Derivatization products are removed immediately from the
reaction medium. (3) Equilibrium can be prevented and a
high yield of conversion is possible. (4) The surplus of
MSTFA is maintained until derivatization and thermal des-
orption are finalized. (5) An improvement of response of
www.atmos-chem-phys.net/11/8977/2011/ Atmos. Chem. Phys., 11, 8977–8993, 2011
8984 J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS
2620 2640 2660 2680 2700 2720 2740 2760 2780 2800
2500
5000
7500
10000
12500
15000
17500
20000
22500
Time (s)
264 FHA_100217_07_08_2s_1 264 FHA_100217_07_08_2s_2
2620 2640 2660 2680 2700 2720 2740 2760 2780 2800
1000
2000
3000
4000
5000
6000
7000
8000
9000
Time (s)
a
b
c
d
e
m/z 264
Isotope labeled
standard compounds
m/z 252
Target compounds
c
d
e
b
a f
without MSTFA
with MSTFA
Fig. 3. The effect of filter matrix and MSTFA on sensitivity
of PAH analysis. Orange line: 108 mm2 filter of sampled PM1
(equivalent to 200 l aerosol) analysed without MSTFA. Green line:
108 mm2 filter of sampled PM1 (equivalent to 200 l aerosol)
analysed with MSTFA-saturated carrier gas during thermal des-
orption. The mass trace m/z 264 is shown for the isotope-
labelled internal standards of (a) benzo[b]fluoranthene (0.802 ng),
(b) benzo[k]fluoranthene (1.04 ng), (c) benzo[e]pyrene (0.802 ng),
(d) benzo[a]pyrene (1.04 ng), (e) perlyene (1.12 ng). Below the
mass trace m/z 252 is shown for the corresponding native com-
pounds with the exception that (a) benzo[b]-, (f) benzo[j]- and (b)
benzo[k]fluoranthene could not be separated.
components suspected to be sensitive for artefact formation
at thermal desorption conditions can be recognized. For in-
stance, low volatile PAH like benzopyrenes and perylene
showed a higher response when the MSTFA saturated car-
rier gas flow was switched on during thermal desorption even
without moistening of the filters with MSTFA before thermal
desorption. The explanation for this effect is the deactivation
of active surfaces e.g. on the quartz fibres. Figure 3 shows
an example for a low volume ambient aerosol sample (200 l
of sampled air) spiked with isotope labelled standard sub-
stances when measured with and without MSTFA-saturated
carrier gas. The example was taken from a series of PM1 fil-
ter samples with sampling on hourly basis. The PM deposits
on the filters were quite low. Therefore, a larger part of the
filter sample was applied for thermal desorption (108 mm2 in
contrast to 27 mm2 for 24 h samples). A two-fold response
improvement was found for some compounds when working
with MSTFA saturated carrier gas. This was especially the
case for the reactive compounds benz[a]pyrene and perylene
without affecting quantification results. The positive effect
(resulting in a lower limit of quantification (LOQ)) was ver-
ified by eighteen additional samples of PM1 filter samples
collected hourly.
Nevertheless, MSTFA may also be responsible for some
artefacts like adducts with aldehydes (Blau, 1977; Halket,
2003; Little, 1999). Derivatization adducts are formed
by enols originating from aldehydes with an α-hydrogen
atom. Electron impact mass spectra therefore show ions
with fragments at m/z 228 (C7H13O2NF3Si+) and m/z 184
(C5H9ONF3Si+). For aromatic aldehydes like vanillin,
coniferyl aldehyde and syringyl aldehyde, however, no for-
mation of adducts was observed.
3.2 Calibration
Calibration curves of IDTD were very similar to the solvent
extraction method. Precision is better for calibration curves
of SE due to the fact that standard solutions were directly
used for calibration. Whereas for thermal desorption meth-
ods calibration reasonably is carried out by standard addition
to reference filters. Thus response differences due to ma-
trix effects are minimized. For that reason some calibration
regressions exhibit a considerable offset. The intercept de-
pends on the concentration of components on the reference
filters. Regression curve data are specified in Table 1.
Calibrations were done with the according isotope-
labelled standards and the native compounds with high purity
as far as possible (Table S1). Only three different isotope-
labelled polar substances were applicable, namely 13C6-
vanillin, 13C6-levoglucosan and D31-palmitic acid. As most
substances analysed here were available in sufficient purity,
only an isotope-labelled internal surrogate standard had to
be applied. However, some substances described here (e.g.
dehydroabietic acid, syringyl aldehyde, divanillyl) were not
available in sufficient purity. Moreover, no isotope-labelled
analogues were available. Thus, only semi-quantitative re-
sults using surrogates were obtained for these compounds.
Applied surrogates are specified in Table S1, too. Due to
different responses of compounds caused by different ex-
traction and derivatization yields the range of application of
isotope-labelled compounds as internal standard for further
substances is limited. Moreover, those compounds should
exhibit equal chromatographic properties and similar frag-
mentation characteristics when treated by electron-impact
ionization. For these reasons only compounds very similar
to the anhydrous sugars levoglucosan, mannosan and galac-
tosan can be quantified properly with 13C6-levoglucosan as
internal standard. D31-palmitic acid was used mainly for
acids, aldehydes and ketones. Even some deuterated PAH
were used as internal standard for some polar compounds
characterized by a two-ring (divanillyl) or three-ring struc-
ture (isopimaric acid). 13C6-vanillin was used as standard
only for its native form. Experiments with vanillic acid and
vanillic acid methyl ester were not successful in receiving
an adequate calibration curve with 13C6-vanillin as internal
standard.
Atmos. Chem. Phys., 11, 8977–8993, 2011 www.atmos-chem-phys.net/11/8977/2011/
J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS 8985
1874 1878 1882 18861888
0
1000
2000
3000
4000
5000
6000
7000
8000
Time (s)
333 338
40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340
50
100
150
200
 45 
 204 
 217  147 
 103  129  59  191 
 117 
 333  161 
73
Fig. 4. Levoglucosan: Structural differences of levoglucosan and its isomers (top right). Left side: coelution of silylated native levoglucosan
(orange line) and the silylated isotope labelled 13C6-levoglucosan (green line) in the chromatogram. Right side: The mass spectra of the
chromatogram at maximum of the isotope labelled levoglucosan.
Levoglucosan showed good linearity in calibration curves
over the whole working range of 3.5–350 ng Levoglucosan
per analysis. The slope is similar to that of the calibration
curve of the solvent extraction method shown in Table S2. It
also shows the intercept of the IDTD method caused by stan-
dard addition to a reference sample. The precision of both
methods is quite good: 6 % for thermal desorption and 4 %
for SE. The limit of detection is somewhat higher for IDTD
compared to SE (0.6 ng and 0.1 ng, respectively). Other stud-
ies using GC-MS methods obtained precisions of 20 % (Gra-
ham et al., 2002 and 2003), 5 % (Simpson et al., 2004) and
2–5 % (Pashynska, 2002). In the presented study mass frag-
ments at m/z 217 and 220 (13C6-levoglucosan) were used
for quantification. Quantification with the base peak pair
at m/z 204 and 206 is not recommended due to the rela-
tively small retention time shift for 13C6-levoglucosan. High
resolution mass spectrometers are necessary to separate iso-
topes in that case. Due to its ubiquity in the atmosphere with
even high concentrations during summer time also fragments
at m/z 333 respectively 338 can be used for quantification
(Fig. 4).
Other polar compounds than levoglucosan exhibit good
linear fittings, too. However, sensitivity of some polar com-
pounds is quite poor, especially for small molecules like suc-
cinic acid, malic acid, or guaiacol. Therefore, Table 1 shows
the calibration data of malic acid as an example for these
compounds with a relative low slope of 0.033 compared to
0.228 for the solvent extraction method. This short chain
multifunctional acid is also discussed in the comparison of
the methods applied for ambient aerosol samples. The cor-
relation of quantitative results from IDTD with the solvent
extraction method was quite good with a slope of the cor-
relation regression near one (see Sect. 3.4). Precision for
malic acid for both methods was near 10 %. Some com-
pounds in low concentrations exhibit higher variation coeffi-
cients up to 17 %, especially syringic acid (16 %) and vanillic
acid methyl ester (17 %).
Usually PAH and oxidized PAH (o-PAH) are not affected
by the derivatization procedure. Although an improvement
of sensitivity was observed like described in Sect. 3.1 deriva-
tization had no influence on quantification. Most calibra-
tion regression curves are showing good linearity with simi-
lar sensitivities for IDTD and SE. No generation of artefacts
were observed neither when comparing the in-situ derivati-
zation procedure with the usual thermal desorption proce-
dure nor when comparing both thermal desorption methods
with a solvent extraction procedure. Moreover, the calibra-
tions of the thermal desorption methods were almost ap-
plicable for solvent extraction methods. Only anthracene
shows variation coefficients higher than 10 % induced by
the poor response of anthracene in GC-MS. Other PAH ex-
hibit variation coefficients below 10 %. This is because of
the application of the respective isotope-labelled standards
for each PAH (Table S1). For o-PAH similar observations
were made. 9,10-Anthracenedione and benzo[a]anthracene-
7,12-dione have low variation coefficients (2 % and 9 %, re-
spectively). For these two o-PAH the according isotope-
labelled standards were applied. Especially 9H-fluoren-9-
one and cyclopenta[def]phenanthrenone show higher varia-
tions though adequate isotope-labelled surrogates were used.
D10-phenanthrene and D12-benzo[a]anthracene were used in
these cases due to best regression fittings.
3.3 Comparison of methods based on the analysis of
SRM 1649a
The DTD method was employed in studies already published
by Schnelle-Kreis et al. (2005a, b, 2007). The standard de-
viation of results (n=5) for all compounds with DTD was
www.atmos-chem-phys.net/11/8977/2011/ Atmos. Chem. Phys., 11, 8977–8993, 2011
8986 J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS
Table 2. Concentration of PAH in NIST Standard Reference Material 1649a (Urban Dust).
SE-GC-MS DTD-GC-MS IDTD-GC-MS Certified Values
Mean value mass of urban dust [µg] 14.8 62.5 46.3
Mean conc. SD Mean conc. SD Mean conc. SD Mean conc. Confidence (95 %)
[pg µg−1] [pg µg−1] [pg µg−1] [pg µg−1]
Phenanthrene 4.52 0.52 5.10 0.64 5.36 0.53 4.14 0.37
Anthracene 0.53 0.04 1.99 0.40 0.98 0.17 0.43 0.08
Fluoranthene 6.21 0.33 6.19 1.24 6.07 0.93 6.45 0.18
Pyrene 5.32 0.26 5.47 0.99 5.15 0.81 5.29 0.25
Benzo[a]anthracene 1.97 0.32 2.21 0.50 1.93 0.25 2.21 0.07
Chrysene 4.70 0.53 4.47 0.40 4.28 0.69 4.41 0.08
Benzo[b]fluoranthene 8.65 0.62 12.51a 1.12 11.71a 1.18 6.45 0.64
Benzo[k]fluoranthene 2.16 0.13 1.93 0.03
Benzo[e]pyrene 3.07 0.25 3.35 0.57 2.83 0.75 3.09 0.19
Benzo[a]pyrene 2.16 0.29 3.51 0.61 2.60 0.41 2.51 0.09
Perylene 0.68 0.11 1.27 0.53 0.44 0.10 0.65 0.08
Indeno[1,2,3-cd]pyrene 2.73 0.31 3.11 0.52 2.87 0.77 3.18 0.72
Dibenzo[a,h]anthracene 0.62 0.09 b 0.88 0.23 0.29 0.02
Benzo[g,h,i]perylene 4.11 0.28 3.79 0.28 3.95 0.47 4.01 0.91
a sum of benzo[b]fluoranthene, benzo[j]fluoranthene and benzo[k]fluoranthene; b not quantified.
Table 3. Concentration of levoglucosan, mannosan and galactosan in NIST Standard Reference Material 1649a (Urban Dust). A comparison
with other publications.
Levoglucosan Mannosan Galactosan
Method n Mean conc. SD Mean conc. SD Mean conc.& SD
[pg µg−1] [pg µg−1] [pg µg−1]
This study IDTD-GC-TOF MS 3 165 1.4 20.5 1.0 10.8 1.8
Larsen et al. (2006) PFEa GC-MS 3 162 8 – – – –
Louchouarn et al. (2009) PFEb GC-MS 4 160.5 4.7 17.3 1.0 5.0 0.3
Kuo et al. (2008) PFEb GC-MS 11 163.9 11.8 – – – –
a Pressurized fluid extraction with ethyl acetate, derivatization with BSTFA/TMCS, derivatization with BSTFA/TMCS (99/1, v/v), Larsen (2006); b Pressurized fluid extraction with
CH2Cl2/CH3OH (9/1, v/v), derivatization with BSTFA/TMCS (99/1, v/v), Kuo (2008).
smaller than 20 % except for perylene and anthracene which
were near the limits of detection. This is due to their low con-
tent in the reference material and the small amount of about
60 µg which was used for thermal desorption (this means a
total mass of anthracene of 26 pg and 39 pg of perylene, re-
spectively). Similar effects were described by van Drooge
et al. (2009). More crucial seems the fact that the con-
centrations of benzo[a]pyrene, an important marker for sev-
eral emission sources and with high carcinogenic potential,
were overestimated by 40 % of the certified value. This phe-
nomenon was not observed when using the IDTD method.
Especially benzo[a]pyrene and perylene seem to be suscepti-
ble for degradation of spiked internal standards.
Nevertheless, the results are satisfying with respect to the
low amounts of SRM in analysis. Other studies used at least
500 µg (Gil-Molto´, 2009), 900 µg up to 1.9 mg (Ho, 2008), or
3 mg (Waterman, 2000) of urban dust for thermal desorption.
Only van Drooge et al. analysed lower SRM 1649a quantities
(van Drooge, 2009) ranging from 60 µg to 550 µg. Results
are visualized in Table 2 and Fig. S1.
The concentration of levoglucosan contained in SRM
1649a was already measured and published (Kuo, 2008;
Larsen, 2006; Louchouarn, 2009). Although the standard
reference materials were not certified for levoglucosan it is
possible to use them for laboratory inter comparison. In
this study a determination of levoglucosan and its analogues
mannosan and galactosan was carried out the same way as for
filter samples with IDTD. Measured values for levoglucosan
were comparable with values already published (Table 3).
Concentrations of mannosan and galactosan have only been
published by Louchouarn et al. (Louchouarn, 2009). Our
results for mannosan were comparable with the published,
whereas the concentrations of galactosan were more than
two times higher (10.8 µg g−1 in this study and 5.0 µg g−1
Atmos. Chem. Phys., 11, 8977–8993, 2011 www.atmos-chem-phys.net/11/8977/2011/
J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS 8987
Phenanthrene
y = 0.887x - 0.024
R2 = 0.994
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0 1 2 3 4 5
SE ng m -3
ID
TD
 n
g 
m
 -3
Fluoranthene
y = 1.042x - 0.028
R2 = 0.991
0
0.5
1
1.5
2
2.5
3
3.5
4
0 1 2 3 4
SE ng m -3
ID
TD
 n
g 
m
 -3
Pyrene
y = 1.091x + 0.067
R2 = 0.981
0
0.5
1
1.5
2
2.5
3
0 1 2 3
SE ng m -3
ID
TD
 n
g 
m
 -3
Benz[a]anthracene
y = 1.241x - 0.033
R2 = 0.945
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 0.2 0.4 0.6 0.8
SE ng m -3
ID
TD
 n
g 
m
 -3
Chrysene and Triphenylene
y = 1.219x + 0.042
R2 = 0.954
0
0.5
1
1.5
2
2.5
3
3.5
0 0.5 1 1.5 2 2.5
SE ng m -3
ID
TD
 n
g 
m
 -3
sum of Benzofluoranthenes
y = 1.255x - 0.109
R2 = 0.965
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0 1 2 3 4
SE ng m -3
ID
TD
 n
g 
m
 -3
Benz[e]pyrene
y = 1.139x + 0.039
R2 = 0.864
0
0.2
0.4
0.6
0.8
1
1.2
0 0.2 0.4 0.6 0.8 1
SE ng m -3
ID
TD
 n
g 
m
 -3
Benz[a]pyrene
y = 1.153x + 0.058
R2 = 0.936
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 0.2 0.4 0.6 0.8 1
SE ng m -3
ID
TD
 n
g 
m
 -3
Perylene
y = 0.723x + 0.015
R2 = 0.694
0
0.02
0.04
0.06
0.08
0.1
0.12
0 0.05 0.1 0.15
SE ng m -3
ID
TD
 n
g 
m
 -3
Indeno[1,2,3-cd]pyrene
y = 0.999x + 0.130
R2 = 0.862
0
0.5
1
1.5
2
2.5
3
3.5
0 1 2 3 4
SE ng m -3
ID
TD
 n
g 
m
 -3
Benzo[ghi]perylene
y = 0.985x - 0.107
R2 = 0.828
0
0.5
1
1.5
2
2.5
3
0 1 2 3
SE ng m -3
ID
TD
 n
g 
m
 -3
Retene
y = 1.068x - 0.010
R2 = 0.624
0
0.05
0.1
0.15
0.2
0.25
0.3
0 0.05 0.1 0.15 0.2 0.25
SE ng m -3
ID
TD
 n
g 
m
 -3
Fig. 5a. Correlation of IDTD and SE for analysis of PAH in ambient aerosol samples (blue: linear regression, red: line through origin, slope
1).
found by Louchouarn et al. respectively). The main differ-
ence of the methods is the manner of extraction. Louchouarn
et al. employed a pressurized fluid extraction with an ac-
celerated solvent extraction system. Due to the similarity
of the anhydrous sugars we doubt that this difference could
be an extraction artefact. The observed ratio of levoglu-
cosan/mannosan/galactosan was similar as found for ambient
aerosol by Ma et al. (2010) (15:3:1 v/v/v).
3.4 Comparison of the methods based on ambient
aerosol samples
A comparison of the in-situ derivatization thermal desorp-
tion method and an ultrasonic extracting method followed by
a derivatization step in solution (SE) was carried out. Cor-
relations of quantitative results are demonstrated. This is vi-
sualized by plotting the result of IDTD analysis against con-
centrations determined by solvent extraction (x-axis). Both
methods show good comparability for most PAH (Fig. 5a).
The correlations indicate good linearity with a slope of nearly
one or slightly higher with the exception of phenanthrene
and perylene (11 % respectively 31 % below results of SE).
These compounds showed higher concentrations when anal-
ysed by solvent extraction. Van Drooge et al. (2009) found
values for phenanthrene 35 % over a liquid extraction method
when working with a thermal desorption method. Also Ho
and Yu (2004) found 33 % higher values when analyzing
www.atmos-chem-phys.net/11/8977/2011/ Atmos. Chem. Phys., 11, 8977–8993, 2011
8988 J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS
9H-Fluoren-9-one
y = 1.024x + 0.047
R2 = 0.858
0
0.5
1
1.5
2
2.5
3
3.5
4
0 1 2 3 4
SE ng m -3
ID
TD
 n
g 
m
 -3
9,10-Anthracenedione
y = 1.649x - 0.564
R2 = 0.855
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 1 2 3 4
SE ng m -3
ID
TD
 n
g 
m
 -3
1,4-Naphthalic anhydride
y = 1.318x + 0.203
R2 = 0.746
0
2
4
6
8
10
12
14
16
18
0 5 10 15
SE ng m -3
ID
TD
 n
g 
m
 -3
Cyclopenta[def]phenanthren-
4-one
y = 1.810x - 0.008
R2 = 0.789
0
0.2
0.4
0.6
0.8
1
1.2
0 0.1 0.2 0.3 0.4 0.5
SE ng m -3
ID
TD
 n
g 
m
 -3
11H-Benzo[a]fluoren-11-one
y = 1.434x - 0.190
R2 = 0.783
0
0.5
1
1.5
2
2.5
3
3.5
0 0.5 1 1.5 2
SE ng m -3
ID
TD
 n
g 
m
 -3
11H-Benzo[b]fluoren-11-one
y = 1.113x - 0.067
R2 = 0.827
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 0.2 0.4 0.6 0.8
SE ng m -3
ID
TD
 n
g 
m
 -3
y = 1.024x + 0.047
R2 = 0.858
0
1
2
3
4
5
6
0 1 2 3 4 5
SE ng m -3
ID
TD
 n
g 
m
 -3
y = 1.810x - 0.0 8
R2 = 0.789
0.2
0.4
0.6
8
1
1.2
1.4
6
0.2 0.4 6
ID
TD
 n
g 
m
 -3
y = 1.649x - 0.564
R2 = 0.855
0
1
2
3
4
5
6
7
0 1 2 3
SE ng m -3
ID
TD
 n
g 
m
 -3
y = 1.318x + 0.203
R2 = 0.746
5
10
15
20
25
ID
TD
 n
g 
m
 -3
y = 1.43 x - 0.190
R2 = 0.783
0.5
1
1.5
2
2.5
3
3.5
4
4
ID
TD
 n
g 
m
 -3
y = 1.113x - 0.067
R2 = 0.827
0.2
4
0.6
8
1
1 2
0.2 0.4 0.6 0.8 1
ID
TD
 n
g 
m
 -3
Dehydroabietic acid, methyl 
ester
y = 0.861x + 0.134
R2 = 0.706
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 0.2 0.4 0.6
SE ng m -3
ID
TD
 n
g 
m
 -3
Fig. 5b. Correlation of IDTD and SE for analysis of o-PAH and dehydroabietic acid methyl ester in ambient aerosol samples (blue: linear
regression, red: line through origin, slope 1).
phenanthrene in aerosol samples with an in-injector thermal
desorption technique. Usually phenanthrene and anthracene
exhibit higher variations due to their low boiling point and
low concentration in the particulate phase. Nevertheless, cor-
relation of IDTD and SE was quite good. The data of fluo-
ranthene and pyrene exhibited the highest correlation of all
PAH analysed, whereas for benzo[a]anthracene 24 %, chry-
sene (and triphenylene) 22 % and the benzofluoranthenes
26 % higher concentrations were determined with IDTD.
These findings are comparable to those of Ho and Yu (2004)
who determined 36 % higher values for benz[a]anthracene
and 19 % higher values for chrysene in their study. Ho et
al. (2008) found a higher correlation coefficient for PAH than
in their previous study (Ho and Yu, 2004) with the same des-
orption technique. The concentration ranges in their study,
2004, were in the same order of magnitude like in this study.
In their recent study from 2008 concentration values were
about ten times higher.
Benzo[a]pyrene (15 % higher values with IDTD) which
is known to be more reactive showed no difference in be-
haviour in this study compared to benzo[e]pyrene (14 %
higher). These benzopyrenes showed higher concentra-
tions when analysed by thermal desorption. The lower
concentrations of perylene when analysed by IDTD may
be a result of low concentrations of perylene in ambi-
ent aerosol samples being also indicated by a poor corre-
lation coefficient (R2 = 0.69). Indeno[1,2,3-cd]pyrene and
Benzo[ghi]perylene showed good comparability but with
somewhat higher variations (R2 = 0.86 respectively 0.83).
Comparability of the analytical methods is also demon-
strated for oxidized PAH (Fig. 5b). For o-PAH only two
isotope-labelled standards were applied. The higher varia-
tion coefficients of calibration curves indicate a lower pre-
cision of analysis of o-PAH for both methods. An influ-
ence on analysis of field samples is therefore not avoid-
able. 9H-fluoren-9-one and 11H-benzo[b]fluoren-11-one
showed good correlation and a slope near unity for the
comparison of IDTD and SE. 9H-fluoren-9-one showed 2 %
higher values and 11H-benzo[b]fluoren-11-one 11 % higher
values when analyzed by IDTD. Ho et al. (2008) also
found good results for 9H-fluoren-9-one (13 % lower values
with TD) and for benz[a]anthracene-7,12-dione (not demon-
strated in this study due to low concentrations). On the
other hand higher values of other o-PAH determined were
observed when being analysed by the in-situ derivatization
thermal desorption. 9,10-Anthracenedione, 1,4-naphthalic
Atmos. Chem. Phys., 11, 8977–8993, 2011 www.atmos-chem-phys.net/11/8977/2011/
J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS 8989
Malic acid
y = 1.088x + 3.425
R2 = 0.905
0
20
40
60
80
100
120
140
0 50 100
SE ng m -3
ID
TD
 n
g 
m
 -3
Phthalic acid
y = 1.140x - 0.303
R2 = 0.837
0
2
4
6
8
10
12
14
0 5 10
SE ng m -3
ID
TD
 n
g 
m
 -3
Galactosan
y = 1.550x - 0.822
R2 = 0.830
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 1 2 3 4
SE ng m -3
ID
TD
 n
g 
m
 -3
Mannosan
y = 0.832x + 0.610
R2 = 0.828
0
5
10
15
20
25
30
35
40
45
50
0 10 20 30 40 50
SE ng m -3
ID
TD
 n
g 
m
 -3
Levoglucosan
y = 1.142x - 17.035
R2 = 0.947
0
50
100
150
200
250
0 50 100 150 200 250
SE ng m -3
ID
TD
 n
g 
m
 -3
Acetosyringone
y = 1.795x - 0.064
R2 = 0.843
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
0 0.2 0.4 0.6 0.8 1
SE ng m -3
ID
TD
 n
g 
m
 -3
Vanillic acid
y = 2.441x - 0.101
R2 = 0.835
0
1
2
3
4
5
6
0 0.5 1 1.5 2 2.5
SE ng m -3
ID
TD
 n
g 
m
 -3
Syringic acid
y = 0.991x + 0.070
R2 = 0.895
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 0.5 1 1.5
SE ng m -3
ID
TD
 n
g 
m
 -3
Divanillyl
y = 2.099x + 0.025
R2 = 0.932
0
0.5
1
1.5
2
2.5
3
0 0.5 1 1.5
SE ng m -3
ID
TD
 n
g 
m
 -3
Isopimaric acid
y = 1.255x - 0.102
R2 = 0.819
0
0.5
1
1.5
2
2.5
0 0.5 1 1.5 2
SE ng m -3
ID
TD
 n
g 
m
 -3
Dehydroabietic acid
y = 2.084x + 6.523
R2 = 0.810
0
5
10
15
20
25
30
35
40
45
50
0 5 10 15 20
SE ng m -3
ID
TD
 n
g 
m
 -3
Abietic acid
y = 1.431x - 0.162
R2 = 0.920
0
0.5
1
1.5
2
2.5
0 0.5 1 1.5 2
SE ng m -3
ID
TD
 n
g 
m
 -3
Fig. 5c. Correlation of IDTD and SE for analysis of polar substances in ambient aerosol samples (blue: linear regression, red: line through
origin, slope 1).
anhydride, cyclopenta[def]phenanthren-4-one and 11H-
benzo[a]fluoren-11-one showed similar correlation coeffi-
cients but concentrations were at least 30 % higher when
analysed with IDTD compared to SE. Neither a loss of an-
alytes by solvent extraction nor a generation of o-PAH from
PAH at thermal desorption conditions can be excluded. An
underestimation of analyte concentrations determined by sol-
vent extraction seems to be possible as a result of a different
extraction efficiency of spiked standards and native sample
compounds.
For selected polar organic compounds, especially wood
combustion tracers, IDTD was applied to ambient aerosol
samples, the results being compared with SE. For all com-
pounds with concentrations above the limit of detection
(LOD) in most collected ambient aerosol samples the cor-
relation to solvent extracted ones is demonstrated (Fig. 5c).
A good correlation and a slope near unity were found for lev-
oglucosan. Ma et al. (2010) showed also good comparability
of their fast two-dimensional GC-MS technique with SE for
levoglucosan without derivatization by analyzing biomass
burning aerosol.
The correlations of galactosan and mannosan were not as
good by far. Due to advantages of levoglucosan in form-
ing silylation products it can be assumed that derivatiza-
tion yields were higher for levoglucosan compared to those
of mannosan and galactosan. Levoglucosan provides three
www.atmos-chem-phys.net/11/8977/2011/ Atmos. Chem. Phys., 11, 8977–8993, 2011
8990 J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS
functional hydroxyl groups which are located in alternating
planes. Its isomers mannosan and galactosan have each two
neighbouring hydroxyl groups which are in the same plane
(Fig. 4). These hydroxyl groups maybe steric hindered when
substituted by trimethylsilyl groups. Therefore the reaction
rate and yield of the three anhydrous sugars could be rather
different due to their stereo isomeric differences. The con-
centrations of these sugars in the ambient air samples were
all in the linear range of their calibration curve using 13C6-
levoglucosan as internal standard. Hence no differences in
correlation of comparison of IDTD/SE were observed.
The resin derived compounds dehydroabietic acid and de-
hydroabietic acid methyl ester were other compounds being
present in all ambient aerosol samples. These compounds
are reaction products of wood combustion generated by de-
hydrogenation and oxidation of resin acids originating from
colophony of conifers (Leithead, 2006; Nolte, 2002). Both
substances were quantified using D10-pyrene as internal stan-
dard. Isopimaric acid was used as surrogate for external cali-
bration of dehydroabietic acid. Pyrene was used for external
calibration of dehydroabietic acid methyl ester already ex-
hibiting a protected carboxylic group and therefore not be-
ing affected by derivatization. The results for dehydroabi-
etic acid, dehydroabietic acid methyl ester and abietic acid
showed up to two times higher values when analysed by
IDTD. Correlations are not even roughly as good as for sub-
stances with according isotope-labelled internal standard but
were still acceptable. But correlation coefficients were in
the same range like the other resin acids abietic acid and
isopimaric acid. These acids were calibrated using respec-
tive standards. Here the lower correlation coefficients were a
result of low concentrations in ambient aerosol.
Beside levoglucosan especially for the relatively small
acids syringic acid, phthalic acid and malic acid a good con-
sistency of the two methods could be observed. On the other
hand twofold higher concentrations of the lignin combustion
products acetosyringone, vanillic acid and divanillyl were
found when analysed by IDTD-GC-TOFMS, as was the case
for dehydroabietic acid. An interesting fact is the lack of
consistency for syringic acid and vanillic acid. These simi-
lar molecules exhibited different results when analysed with
IDTD and SE. Both acids were calibrated and analysed by
applying D31-palmitic acid as internal standard (Table S1).
Experiments with different isotope-labelled standards sub-
stantiate the suspicion that in some cases extraction effi-
ciency was poor when using ultrasonic-assisted extraction in
dichloromethane/methanol (1/1, v/v). The use of this sol-
vent mixture is a compromise to achieve sufficient extraction
yields of all analytes. Otherwise different fractions have to be
extracted by different solvents requiring additional work for
sample work-up. Despite of better precision of calibration
curves of SE as shown in Table S2 it must be noticed that
calibration regression fittings were calculated on two differ-
ent ways. The standard addition as used for IDTD for calcu-
lation of regression curves already comprises the extraction
efficiencies, whereas the calibration data of the SE method
are based on derivatized standard compounds. Indeed, IDTD
calibration curves for vanillic acid, acetosyringone, isopi-
maric acid and dehydroabietic acid showed less sensitivity
compared to SE. In these cases a correction by recovery cal-
culation is not possible due to non reliable isotope-labelled
standards. PAH also showed a tendency to be underestimated
in concentration when analysed by the SE method. This devi-
ation increased with increasing boiling point of the PAH. Due
to a high number of according isotope-labelled standards it is
not as present as for other substances by far.
The correlation experiments showed a good consistency
of the analysis methods for malic acid, mannosan, lev-
oglucosan, phthalic acid, syringic acid and isopimaric acid,
9H-fluoren-9-one, 11H-benzo[b]fluoren-11-one and the most
PAH. A good consistency for other substances was opposed
by the use of internal surrogate standards which were not
isotope-labelled analogues of the analytes. Dehydroabietic
acid methyl ester demonstrates that this is not only a draw-
back of polar substances even though measured concentra-
tions of the methyl ester were somewhat lower than those of
the free acid. On the other hand derivatization yields were
similar for both methods.
4 Conclusions
A fast in-situ derivatization thermal desorption technique
was developed for GC-MS. This method is able to deal
with daily ambient aerosol sampling. The feature of
analysing polar compounds gets more and more important
to characterize ambient aerosol supporting source apportion-
ment studies and aerosol ageing studies, even those em-
ploying chamber experiments (Bo¨ge, 2006; Chandramouli,
2003; Edney, 2005). As an amplification of this method
GCxGC-techniques (Goldstein, 2008; Ma, 2008; Schnelle-
Kreis, 2005b) could deal with enlarged data quantities
and reduces peak co-elution significantly. Although the
LecoCromaTOF®-Software has a powerful deconvolution
algorithm peak separation becomes a problem for chro-
matograms including additional silylation products. Never-
theless, one dimensional gas chromatography in combination
with derivatization techniques provides more relevant infor-
mation of the chemical properties of ambient aerosol than
without derivatization.
A positive side effect demonstrated is the minimization of
matrix effects by deactivation of active quartz fibres caused
by the use of MSTFA during thermal desorption. An im-
provement of LOD/LOQ for PAH was shown and could also
be suggested for further analytes.
The advantages of isotope-labelled standards in GC-MS
are well known. They gain an even higher importance for
thermal desorption methods. The response of organic com-
pounds depends on the composition of the particle matrix
and on the quartz fibres. Compared to solvent extraction not
Atmos. Chem. Phys., 11, 8977–8993, 2011 www.atmos-chem-phys.net/11/8977/2011/
J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS 8991
only a fraction of this matrix but the whole sample composi-
tion affects the analysis. This always should be kept in mind
when applying thermal desorption techniques. It is strictly
recommended to employ a large set of isotope-labelled
standards. Even compounds like PAH and dehydroabietic
methyl ester do not exhibit responses proportional to reason-
able surrogates, as shown in this paper. Nevertheless, the
measured concentration of these compounds may be more
affected by sampling artefacts than by the analysis methods
employed. We clearly demonstrated that in-situ derivatiza-
tion thermal desorption gas chromatography time-of-flight
mass spectrometry shows a good linearity and sensitiv-
ity over nearly one order of magnitude for analysis of the
important and relatively stable biomass marker levoglucosan.
Edited by: A. Kiendler-Scharr
Supplement related to this article is available online at:
http://www.atmos-chem-phys.net/11/8977/2011/
acp-11-8977-2011-supplement.pdf.
References
Bari, M. A., Baumbach, G., Kuch, B., and Scheffknecht, G.: Wood
smoke as a source of particle-phase organic compounds in resi-
dential areas, Atmos. Environ., 43, 4722–4732, 2009.
Bates, M., Bruno, P., Caputi, M., Caselli, M., de Gennaro, G.,
and Tutino, M.: Analysis of polycyclic aromatic hydrocarbons
(PAHs) in airborne particles by direct sample introduction ther-
mal desorption GC/MS, Atmos. Environ., 42, 6144–6151, 2008.
Beiner, K., Plewka, A., Haferkorn, S., Iinuma, Y., Engewald, W.,
and Herrmann, H.: Quantification of organic acids in pariculate
matter by coupling of thermally assisted hydrolysis and methyla-
tion with thermodesorption-gas chromatography-mass spectrom-
etry, J. Chromatogr. A, 1216, 6642–6650, 2009.
Blau, K. and King, G. S.: Handbook of Derivatives for Chromatog-
raphy, Heyden & Son Ltd., London, UK, 1977.
Bo¨ge, O., Miao, Y., Plewka, A., and Herrmann, H.: Formation of
secondary organic particle phase compounds from isoprene gas-
phase oxidation products: An aerosol chamber and field study,
Atmos. Environ., 40, 2501–2509, 2006.
Chandramouli, B., Jang, M., and Kamens, R. M.: Gas-particle par-
titioning of semivolatile organic compounds (SOCs) on mixtures
of aerosols in a smog chamber, Environ. Sci. Technol., 37, 4113–
4121, 2003.
Chiappini, L., Perraudin, E., Durand-Jolibois, R., and Doussin,
J.: Development of a supercritical fluid extraction–gas
chromatography–mass spectrometry method for the identifica-
tion of highly polar compounds in secondary organic aerosols
formed from biogenic hydrocarbons in smog chamber experi-
ments, Analyt. Bioanalyt. Chem., 386, 1749–1759, 2006.
Chow, J. C., Yu, J. Z., Watson, J. G., Hang Ho, S. S., Bohannan,
T. L., Hays, M. D., and Fung, K. K.: The application of thermal
methods for determining chemical composition of carbonaceous
aerosols: A review, J. Environ. Sci. Health, 42, 1521–1541, 2007.
Ding, L. C., Ke, F., Wang, D. K. W., Dann, T., and Austin, C. C.:
A new direct thermal desorption-GC/MS method: Organic spe-
ciation of ambient particulate matter collected in Golden, BC,
Atmos. Environ., 43, 4894–4902, 2009.
Donahue, N. M., Robinson, A. L., and Pandis, S. N.: Atmospheric
organic particulate matter: From smoke to secondary organic
aerosol, Atmos. Environ., 43, 94–106, 2009.
Dutton, S. J., Williams, D. E., Garcia, J. K., Vedal, S., and Han-
nigan, M. P.: PM2.5 characterization for time series studies:
Organic molecular marker speciation methods and observations
from daily measurements in denver, Atmos. Environ., 43, 2008–
2013, 2009.
Edney, E. O., Kleindienst, T. E., Conver, T. S., McIver, C. D., Corse,
E. W., and Weathers, W. S.: Polar organic oxygenates in PM2.5
at a southeastern site in the United States, Atmos. Environ., 37,
3947–3965, 2003.
Edney, E. O., Kleindienst, T. E., Jaoui, M., Lewandowski, M., Of-
fenberg, J. H., Wang, W., Claeys, M.: Formation of 2-methyl
tetrols and 2-methylglyceric acid in secondary organic aerosol
from laboratory irradiated isoprene/NOx/SO2/air mixtures and
their detection in ambient PM2.5 samples collected in the eastern
United States, Atmos. Environ., 39, 5281–5289, 2005.
Fabbri, D., Prati, S., and Vassura, I.: Molecular characterisation of
organic material in air fine particles (PM10) using conventional
and reactive pyrolysis-gas chromatography-mass spectrometry,
J. Environ. Monitor., 4, 210–215, 2002.
Falkovich, A. H. and Rudich, Y.: Analysis of semivolatile or-
ganic compounds in atmospheric aerosols by direct sample in-
troduction thermal desorption GC/MS, Environ. Sci. Technol.,
35, 2326–2333, 2001.
Fine, P. M., Cass, G. R., and Simoneit, B. R. T.: Chemical char-
acterization of fine particle emissions from fireplace combustion
of woods grown in the northeastern United States, Environ. Sci.
Technol., 35, 2665–2675, 2001.
Fine, P. M., Cass, G. R., and Simoneit, B. R. T.: Chemical charac-
terization of fine particle emissions from the fireplace combus-
tion of woods grown in the southern United States, Environ. Sci.
Technol., 36, 1442–1451, 2002.
Fine, P. M., Cass, G. R., and Simoneit, B. R. T.: Chemical charac-
terization of fine particle emissions from the fireplace combus-
tion of wood types grown in the midwestern and western United
States, Environ. Eng. Sci., 21, 387–409, 2004.
Gil-Molto´, J., Vareaa, M., Galindoa, N., and Crespo, J.: Appli-
cation of an automatic thermal desorption-gas chromatography-
mass spectrometry system for the analysis of polycyclic aromatic
hydrocarbons in airborne particulate matter, J. Chromatogr. A,
1216, 1285–1289, 2009.
Goldstein, A. H., Worton, D. R., Williams, B. J., Hering, S. V.,
Kreisberg, N. M., Panic, O., Gorecki, T.: Thermal desorption
comprehensive two-dimensional gas chromatography for in-situ
measurements of organic aerosols, J. Chromatogr. A, 1186, 340–
347, 2008.
Graham, B., Mayol-Bracero, O. L., Guyon, P., Roberts, G. C.,
Decesari, S., Facchini, M. C., Artaxo, P., Maenhaut, W., Ko¨ll,
P., and Andreae, M. O.: Water-soluble organic compounds
in biomass burning aerosols over amazonia – 1. Characteri-
zation by NMR and GC-MS, J. Geophys. Res., 107, 8047,
doi:10.1029/2001jd000336, 2002.
Graham, B., Guyon, P., Taylor, P. E., Artaxo, P., Maenhaut, W.,
www.atmos-chem-phys.net/11/8977/2011/ Atmos. Chem. Phys., 11, 8977–8993, 2011
8992 J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS
Glovsky, M. M., Flagan, R. C., and Andreae, M. O.: Organic
compounds present in the natural amazonian aerosol: Character-
ization by gas chromatography-mass spectrometry, J. Geophys.
Res., 108, 4766, doi:10.1029/2003jd003990, 2003.
Guenther, A., Hewitt, C. N., Erickson, D., Fall, R., Geron, C.,
Graedel, T., Harley, P., Klinger, L., Lerdau, M., McKay, W. A.,
Pierce, T., Scholes, B., Steinbrecher, R., Tallamraju, R., Taylor,
J., and Zimmerman, P.: A global model of natural volatile or-
ganic compound emissions, J. Geophys. Res., 100, 8873–8892,
1995.
Halket, J. M., Zaikin, V. G.: Derivatization in mass spectrometry –
1. Silylation, Europ. J. Mass Spectrom., 9, 1–21, 2003.
Hallquist, M., Wenger, J. C., Baltensperger, U., Rudich, Y., Simp-
son, D., Claeys, M., Dommen, J., Donahue, N. M., George, C.,
Goldstein, A. H., Hamilton, J. F., Herrmann, H., Hoffmann, T.,
Iinuma, Y., Jang, M., Jenkin, M. E., Jimenez, J. L., Kiendler-
Scharr, A., Maenhaut, W., McFiggans, G., Mentel, T. F., Monod,
A., Pre´voˆt, A. S. H., Seinfeld, J. H., Surratt, J. D., Szmigiel-
ski, R., and Wildt, J.: The formation, properties and impact of
secondary organic aerosol: Current and emerging issues, At-
mos. Chem. Phys., 9, 5155–5236, doi:10.5194/acp-9-5155-2009,
2009.
Hays, M. D. and Lavrich, R. J.: Developments in direct thermal ex-
traction gas chromatography-mass spectrometry of fine aerosols,
Trends Analyt. Chem., 26, 88-102, 2007.
Hays, M. D., Smith, N. D., Kinsey, J., Dong, Y., Kariher, P.: Poly-
cyclic aromatic hydrocarbon size distributions in aerosols from
appliances of residential wood combustion as determined by di-
rect thermal desorption-GC/MS, J. Aerosol Sci., 34, 1061–1084,
2003.
Ho, S. S. H. and Yu, J. Z.: In-injection port thermal desorption and
subsequent gas chromatography-mass spectrometric analysis of
polycyclic aromatic hydrocarbons and n-alkanes in atmospheric
aerosol samples, J. Chromatogr. A, 1059, 121–129, 2004.
Ho, S. S. H., Yua, J. Z., Chow, J. C., Zielinskaa, B., Watson, J. G.,
Leung Sit, E. H., and Schauer, J. J.: Evaluation of an in-injection
port thermal desorption-gas chromatography/mass spectrometry
method for analysis of non-polar organic compounds in ambient
aerosol samples, J. Chromatogr. A, 1200, 217–227, 2008.
Kourtchev, I., Ruuskanen, T., Maenhaut, W., Kulmala, M.,
and Claeys, M.: Observation of 2-methyltetrols and related
photo-oxidation products of isoprene in boreal forest aerosols
from Hyytia¨la¨, Finland, Atmos. Chem. Phys., 5, 2761–2770,
doi:10.5194/acp-5-2761-2005, 2005.
Kuo, L.-J., Herbert, B. E., and Louchouarn, P.: Can levoglucosan
be used to characterize and quantify char/charcoal black carbon
in environmental media?, Org. Geochem., 39, 1466–1478, 2008.
Lambe, A. T., Chacon-Madrid, H. J., Nguyen, N. T., Weitkamp,
E. A., Kreisberg, N. M., Hering, S. V., Goldstein, A. H., Don-
ahue, N. M., and Robinson, A. L.: Organic aerosol speciation:
Intercomparison of thermal desorption aerosol GC/MS (TAG)
and filter-based techniques, Aerosol Sci. Technol., 44, 141–151,
2010.
Larsen III, R. K., Schantz, M. M., and Wise, S. A.: Determination of
levoglucosan in particulate matter reference materials, Aerosol
Sci. Technol., 40, 781–787, 2006.
Leithead, A., Li, S.-M., Hoff, R., Cheng, Y., and Brook, J.: Lev-
oglucosan and dehydroabietic acid: Evidence of biomass burn-
ing impact on aerosol in the lower fraser valley, Atmos. Environ.,
40, 2721–2734, 2006.
Little, J. L.: Artifacts in trimethylsilyl derivatization reactions and
ways to avoid them, J. Chromatogr. A, 844, 1–22, 1999.
Liu, Y., Sklorz, M., Schnelle-Kreis, J., Orasche, J., Ferge, T.,
Kettrup, A., and Zimmermann, R.: Oxidant denuder sampling
for analysis of polycyclic aromatic hydrocarbons and their oxy-
genated derivates in ambient aerosol: Evaluation of sampling
artefact, Chemosphere, 62, 1889–1898, 2006.
Louchouarn, P., Kuo, L.-J., Wade, T. L., and Schantz, M.: Determi-
nation of levoglucosan and its isomers in size fractions of aerosol
standard reference materials, Atmos. Environ., 43, 5630–5636,
2009.
Ma, Y. and Hays, M. D.: Thermal extraction-two-dimensional gas
chromatography-mass spectrometry with heart-cutting for nitro-
gen heterocyclics in biomass burning aerosols, J. Chromatogr. A,
1200, 228–234, 2008.
Ma, Y., Hays, M. D., Geron, C. D., Walker, J. T., and Gatari
Gichuru, M. J.: Technical Note: Fast two-dimensional GC-MS
with thermal extraction for anhydro-sugars in fine aerosols, At-
mos. Chem. Phys., 10, 4331–4341, doi:10.5194/acp-10-4331-
2010, 2010.
Medeiros, P. M. and Simoneit, B. R. T.: Analysis of sugars in en-
vironmental samples by gas chromatography-mass spectrometry,
J. Chromatogr. A, 1141, 271–278, 2007.
Medeiros, P. M., Conte, M. H., Weber, J. C., and Simoneit, B. R.
T.: Sugars as source indicators of biogenic organic carbon in
aerosols collected above the howland experimental forest, maine,
Atmos. Environ., 40, 1694–1705, 2006.
Nolte, C. G., Schauer, J. J., Cass, G. R., and Simoneit, B. R. T.:
Highly polar organic compounds present in wood smoke and in
the ambient atmosphere, Environ. Sci. Technol., 35, 1912–1919,
2001.
Nolte, C. G., Schauer, J. J., Cass, G. R., and Simoneit, B. R. T.:
Trimethylsilyl derivatives of organic compounds in source sam-
ples and in atmospheric fine particulate matter, Environ. Sci.
Technol., 36, 4273–4281, 2002.
Oliveira, C., Pio, C., Alves, C., Evtyugina, M., Santos, P.,
Gonc¸alves, V., Nunes, T., Silvestre, A. J. D., Palmgren, F.,
Wa˚hlin, P., and Harrad, S.: Seasonal distribution of polar or-
ganic compounds in the urban atmosphere of two large cities
from the north and south of europe, Atmos. Environ., 41, 5555–
5570, 2007.
Pashynska, V., Vermeylen, R., Vas, G., Maenhaut, W., and
Claeys, M.: Development of a gas chromatographic/ion trap
mass spectrometric method for the determination of levoglu-
cosan and saccharidic compounds in atmospheric aerosols. Ap-
plication to urban aerosols, J. Mass Spectrom., 37, 1249–1257,
doi:10.1002/jms.391, 2002.
Schkolnik, G. and Rudich, Y.: Detection and quantification of lev-
oglucosan in atmospheric aerosols: A review, Anal. Bioanal.
Chem., 385, 26–33, 2006.
Schnelle-Kreis, J., Sklorz, M., Peters, A., Cyrys, J., Zimmermann,
R.: Analysis of particle-associated semi-volatile aromatic and
aliphatic hydrocarbons in urban particulate matter on a daily ba-
sis, Chemosphere, 39, 7702–7714, 2005a.
Schnelle-Kreis, J., Welthagen, W., Sklorz, M., Zimmermann, R.:
Application of direct thermal desorption gas chromatography and
comprehensive two-dimensional gas chromatography coupled to
time of flight mass spectrometry for analysis of organic com-
Atmos. Chem. Phys., 11, 8977–8993, 2011 www.atmos-chem-phys.net/11/8977/2011/
J. Orasche et al.: In-situ derivatization thermal desorption GC-TOFMS 8993
pounds in ambient aerosol particles, J. Sep. Sci., 28, 1648–1657,
2005b.
Schnelle-Kreis, J., Sklorz, M., Orasche, J., Sto¨lzel, M., Peters, A.,
and Zimmermann, R.: Semi volatile organic compounds in am-
bient PM2.5. Seasonal trends and daily resolved source contribu-
tions, Environ. Sci. Technol., 41, 3821–3828, 2007.
Sheesley, R. J., DeMinter, J. T., Meiritz, M., Snyder, D. C., and
Schauer, J. J.: Temporal trends in motor vehicle and secondary
organic tracers using in situ methylation thermal desorption
GCMS, Environ. Sci. Technol., 44, 9398–9404, 2010.
Simoneit, B. R. T., Schauer, J. J., Nolte, C. G., Oros, D. R., Elias, V.
O., Fraser, M. P., Rogge, W. F., and Cass, G. R.: Levoglucosan, a
tracer for cellulose in biomass burning and atmospheric particles,
Atmos. Environ., 33, 173–182, 1999.
Simoneit, B. R. T., Elias, V. O., Kobayashi, M., Kawamura, K.,
Rushdi, A. I., Medeiros, P. M., Rogge, W. F., Didyk, Borys M. :
Sugars dominant water-soluble organic compounds in soils and
characterization as tracers in atmospheric particulate matter, En-
viron. Sci. Technol., 38, 5939–5949, 2004.
Simpson, C. D., Dills, R. L., Katz, B. S., and Kalman, D. A.: Deter-
mination of levoglucosan in atmospheric fine particulate matter,
J. Air Waste Manage. Assoc., 54, 689–694, 2004.
Sklorz, M., Schnelle-Kreis, J., Liu, Y., Orasche, J., and Zimmer-
mann, R.: Daytime resolved analysis of polycyclic aromatic hy-
drocarbons in urban aerosol samples – impact of sources and me-
teorological conditions, Chemosphere, 67, 934–943, 2007.
van Drooge, B. L., Nikolova, I., and Ballesta, P. P.: Thermal des-
orption gas chromatography–mass spectrometry as an enhanced
method for the quantification of polycyclic aromatic hydrocar-
bons from ambient air particulate matter, J. Chromatogr. A, 1216,
4030–4039, 2009.
Vedal, S., Hannigan, M. P., Dutton, S. J., Miller, S. L., Milford,
J. B., Rabinovitch, N.,Kim, S.-Y., and Sheppard L.: The den-
ver aerosol sources and health (dash) study: Overview and early
findings, Atmos. Environ., 43, 1666–1673, 2009.
Waterman, D., Horsfield ,B., Leistner, F., Hall, K., and Smith, S.:
Quantification of polycyclic aromatic hydrocarbons in the nist
standard reference material (SRM1649a) urban dust using ther-
mal desorption GC/MS, Anal. Chem., 72, 3563–3567, 2000.
Waterman, D., Horsfield, B., Hall, K., and Smith, S.: Ap-
plication of micro-scale sealed vessel thermal desorption–gas
chromatography–mass spectrometry for the organic analysis of
airborne particulate matter: Linearity, reproducibility and quan-
tification, J. Chromatogr. A, 912, 143–150, 2001.
Williams, B. J., Goldstein, A. H., Kreisberg, N. M., and Hering, S.
V.: An in-situ instrument for speciated organic composition of
atmospheric aerosols: Thermal desorption aerosol GC/MS-FID
(TAG), Aerosol Sci. Technol., 40, 627–638, 2006.
Williams, B. J., Goldstein, A. H., Kreisberg, N. M., Hering, S. V.,
Worsnop, D. R., Ulbrich, I. M., Docherty, K. S., and Jimenez, J.
L.: Major components of atmospheric organic aerosol in south-
ern California as determined by hourly measurements of source
marker compounds, Atmos. Chem. Phys., 10, 11577–11603,
doi:10.5194/acp-10-11577-2010, 2010.
Zdrahal, Z., Oliveira, J., Vermeylen, R., Claeys, M., and Maenhaut,
W.: Improved method for quantifying levoglucosan and related
monosaccharide anhydrides in atmospheric aerosols and applica-
tion to samples from urban and tropical locations, Environ. Sci.
Technol., 36, 747–753, 2002.
www.atmos-chem-phys.net/11/8977/2011/ Atmos. Chem. Phys., 11, 8977–8993, 2011