J Therm Anal Calorim (2014) 116:1461–1469
DOI 10.1007/s10973-014-3830-3
Evolved gas analysis by single photon ionization-mass
spectrometry
A tool to distinguish different types of coffee
M. Fischer • S. Wohlfahrt • J. Varga •
M. Saraji-Bozorgzad • G. Matuschek •
T. Denner • R. Zimmermann
Received: 14 October 2013 / Accepted: 16 April 2014 / Published online: 13 May 2014
Ó Akadémiai Kiadó, Budapest, Hungary 2014
Abstract The applicability of thermogravimetry (TG)
coupled to single photon ionization time-of-flight mass
spectrometry (TG-SPI-TOFMS) for evolved gas analysis
(EGA) of coffee is demonstrated in this study. Coffee is a
chemically well-known complex food product of large
scientific and commercial interest. The roasting process of
single green coffee beans (Arabica, Robusta) was simu-
lated in the TG-SPI-TOFMS device, and the chemical
composition of the evolved roasting gases was monitored
on-line. Additionally, roasted and ground coffee powders
of different types and brands as well as instant coffee were
successfully investigated. For example, the diterpenes
cafestol and kahweol can be detected among many other
roasting products. These compounds can be of particular
interest for quality control of coffee. It is shown that
M. Fischer
S. Wohlfahrt
J. Varga
G. Matuschek (&)

R. Zimmermann
Joint Mass Spectrometry Centre, Cooperation Group
Comprehensive Molecular Analytics (CMA), Helmholtz
Zentrum München, Ingolstädter Landstr. 1, 85764 Neuherberg,
Germany
e-mail: matuschek@helmholtz-muenchen.de
M. Fischer
S. Wohlfahrt
R. Zimmermann
Joint Mass Spectrometry Centre, Universität Rostock, Chair of
Analytical Chemistry, 18051 Rostock, Germany
J. Varga
Universität Augsburg, Chair of Resource Strategy,
86159 Augsburg, Germany
M. Saraji-Bozorgzad
Photonion GmbH, 85764 Neuherberg, Germany
T. Denner
Netzsch-Gerätebau GmbH, 95100 Selb, Germany
kahweol can be used as a tracer compound to discriminate
arabica coffee species from robusta species.
Keywords Evolved gas analysis
Thermal analysis

Single photon ionization
Kahweol
Cafestol
Introduction
Thermal analysis (TA) is established in industry and
research as a convenient method to study chemical and
physical temperature dependent properties of materials.
Sophisticated analytical devices often combine thermo-
gravimetry (TG) and differential scanning calorimetry
(DSC) for simultaneous thermal analysis (STA). While TG
and its often used derivative (DTG) describe the mass loss,
DSC describes enthalpy changes as a function of temper-
ature. To generate more detailed information about the
sample composition, evolved gas analysis (EGA) is
required [1]. There are two options for EGA: First, offline,
using a sequentially working analytical techniques such as
gas chromatography [2]. One major drawback is the loss of
temperature resolution by trapping the evolved gases.
Second, on-line techniques such as Fourier-transform
infrared spectroscopy (FTIR) [1] or mass spectrometry
(MS) [1–4] offer real-time detection and, therefore, direct
assignment of mass-loss signals to chemical information. A
connection between TA and MS, depending on the inves-
tigative goal, can be realized with a capillary coupling [5,
6] or a skimmed supersonic expansion [7–9].
The most frequently used ionization method, electron
ionization (EI) with standard kinetic electron energies of
70 eV, leads to high ionization rates but can also cause
intense fragmentation of the molecules. To circumvent this
problem, soft ionization techniques are becoming the
123
1462
method of choice. Broadly used technique is chemical
ionization (CI) [10] or photon ionization (PI) such as res-
onance-enhanced multiphoton ionization (REMPI) [11] or
single photon ionization (SPI) [12]. The vacuum UV
(VUV)-light for SPI process can be generated by laser
processes (coherent VUV-pulses) or by incoherent, lamp-
type sources. The latter class comprehends several types of
VUV-lamps such as the deuterium lamp or the electron
beam pumped rare gas excimer lamp (EBEL) for SPI. The
EBEL source was reported in the past for its rather high
photon densities, optical brilliance and in particular spec-
tral stability over lifetime [13, 14]. SPI causes no or much
less fragmentation than the classical hard EI method and
leads to the predominant appearance of molecular ions
from organic compounds. Furthermore, SPI is a semi-
selective ionization method due to the energetic threshold
selectivity of the photons (8–12 eV). Highly abundant bulk
gases with ionization energies above 12 eV such as nitro-
gen (IE = 15.58 eV), oxygen (IE = 12.06 eV), carbon
dioxide (IE = 13.77 eV), and water (IE = 12.62 eV) will
not be ionized, avoiding detector saturation [15, 16].
Therefore, SPI-MS can be considered for the investigation
of pyrolysis or combustion products of natural samples as
coffee [17, 18] or tobacco [15, 19–21].
TA-SPI-MS using a quadrupole mass spectrometer
(QMS) has already been described in detail [22]. Instead of
a QMS an orthogonally accelerated time-of-flight mass
spectrometer (oa-TOFMS) was applied by Geissler et al.
[13] in order to benefit from the dynamic range of m/z, the
sensitivity as well as the time resolving power of a TOFMS
[23]. Fujii [24] described similar aims using EGA–ion-
attachment mass spectrometry (EGA–IAMS). Recently,
also a hyphenation of TA to a SPI-ion trap mass spec-
trometer is used (TA-SPI-ITMS). The SPI-ITMS approach
allows compound identification by tandem MS [25].
Coffee is one of the largest traded foods worldwide,
with the dominant species being Coffea arabica (Arabica)
and Coffea canephora (Robusta). Thus, research on quality
and optimization of the industrial processes is of great
commercial interest. To efficiently study, improve, or
develop new concepts for the industrial roasting process,
roasting experiments can be done laboratory-based on the
micro-scale using single beans in TG-SPI-TOFMS exper-
iments. The analysis of coffee remains a great challenge.
While in green coffee beans 300 volatile and semi-volatile
organic compounds (VOC/SVOC) have already been
found, there are more than 850 compounds in roasted
coffee. Most of them are formed during the roasting by
Strecker and Maillard reactions [17, 26]. The roasting
process has been investigated in the past with several
techniques using PI: micro-probe sampling PI-MS [17],
thermal desorption (TD)-SPI-ITMS [25], or doing head-
space sampling during roasting and brewing of coffee via
123
M. Fischer et al.
laser ionization (LI)-MS [27]. The formation and on-line
detectability of many roasting products such as phenols,
carbonyls, pyridine and caffeine by photoionization MS
has been reported previously [17, 18, 28, 29]. Interesting
cases are the diterpene lipids cafestol and kahweol. Urgert
and Katan [30] described these two lipids in coffee and
discussed their cholesterol-raising potential in human
serum in detail. Robusta contains small amounts of kah-
weol and less cafestol than Arabica [30–32]. Besides the
potential health effect, cafestol (responsible for more than
80 % of the cholesterol-raising effect [30, 33]) and kah-
weol can be used as tracer substances to discriminate
Arabica from Robusta. The discrimination could be shown
during the simulated roasting process as well as in roasted
and ground coffee powder. Two different SPI sources were
used for the powder samples to obtain knowledge about
their possible advantages and disadvantages. Note, that
isomeric or isobaric components cannot be separated with
the current on-line experimental setup. This would require,
for example, an additional chromatographic separation step
[34, 35].
Experimental
The principal setup of STA–SPI-TOFMS (Photonion
GmbH, Schwerin, Germany) system has been described
previously in detail [13]. Briefly, an STA 449 F3 Jupiter
(Netzsch-Gerätebau GmbH, Selb, Germany) was con-
nected to an oa-TOFMS (C-TOF, Tofwerk AG, Thun,
Switzerland) via a 250 °C heated deactivated fused silica
transfer-capillary with an inner diameter of 180 lm and
3 m length. The pressure changed from ambient pressure in
the STA to a vacuum of 3.2 9 10-5 mbar in the recipient
(ion source) and 1.1 9 10-6 mbar in the TOFMS. The
nitrogen flow through the STA was set to 60 mL min-1
(40 mL min-1 purge gas, 20 mL min-1 protective gas).
The samples were treated with different temperature pro-
grams. Standard coffee roasting processes are performed in
a temperature range from 180 to 240 °C and last for
3–12 min [31]. To simulate the roasting process the green
coffee beans were heated from ambient temperature to
250 °C with a heating rate of 10 K min-1 followed by a
35 min isothermal step at 250 °C. Two coffee powders
were treated with the same parameters. Other powder
samples were heated from ambient temperature to 1,000 °C
with a heating rate of 10 K min-1. The employed VUV-
light sources for SPI were a 150 W, water-cooled L 1835
deuterium lamp (Hamamatsu Photonics Deutschland
GmbH, Herrsching, Germany) and an EBEL. A schematic
setup is shown in Fig. 1.
For the simulated roasting process the sample holder of
the STA was equipped with an Al2O3 beaker (Fig. 2). The
Evolved gas analysis by single photon ionization-mass spectrometry 1463

Heated 100 0
Heated interface transfer line Heated capillary inlet
–1

DTG/% min–1
Cross beam

Mass/%
ionizer
–2
80
VUV source –3
3.2·10–5 mbar

Reflector
Orthogonal –4
extractor PC

1.1·10–6 mbar
60
DAQ
Preamp
50 150 250 Isotherm
Detector
HV Timing Temperature/°C
pulser controller
STA oa-TOFMS
100 0
Fig. 1 Draft of Photonion’s hyphenation of STA with SPI-TOFMS.
Different VUV-light sources can be used such as a deuterium lamp or
an EBEL –1

DTG/% min–1
Mass/%
–2
80
–3

–4

60
50 150 250 Isotherm
Temperature/°C

Fig. 3 Mass loss (blue, dotted line) and its derivative (green, full
line) of a green Arabica bean (top) and a Robusta bean (bottom)
during roasting. The beans were heated up to 250 °C with
10 K min-1 and held isotherm for 35 min. The cracking of the bean
Fig. 2 Green coffee bean before (left) and after (right) the simulated
causes a vibration which is picked up readily by the microbalance
roasting process in the beaker of the STA
(red arrows). (Colour figure online)

green coffee bean was placed into the beaker to avoid a
possible fall which could occur during the cracking of the
bean. For the coffee powder as well as the instant coffee
crumbs, 30 mg was weighed into Al2O3 crucibles. Initial
masses were normalized thereafter. All samples were
measured without any pre-treatment. All coffee powders
and instant coffee samples used in this study are com-
mercially available products. They were chosen because of
their composition (pure Arabica and pure Robusta) and
randomly out of US supermarkets’ offer.
Results and discussion
The purpose of this work is to demonstrate the applicability
of TG with advanced EGA methods for studying roast
processes in food chemistry. As the properties and chem-
ical composition of natural products such as green coffee
beans vary from bean to bean usually several replicate-
measurement are required for quantitative results [17, 29].
However, the intention of the here presented results is to
demonstrate the potential of the TG-SPI-TOFMS instru-
mental setup in this field. In addition to the EGA data for
every single run/measurement, the thermogravimetric data
are recorded. The mass loss and its derivative (TG and
DTG) of green Arabica and Robusta beans during the
roasting process are shown in Fig. 3. The so called
‘‘cracking‘‘ of the beans due to the internally formed CO2
by decarboxylation of organic acids represents an impor-
tant step in the roasting process (opening of the cellular
structures). The crack can be picked up by the microbal-
ance due to the impulse transferred to the crucible.
Analyzing the time-resolved SPI-mass spectrometric
signature contour plots of the temperature-resolved mass
spectra of the roasting process in Fig. 4, some differences
between the two coffee species are remarkable. At 20 min,
the signals at m/z 296 and 298 are dominant. These signals
can be assigned to diterpene lipids anhydrous cafestol (m/z
298) and anhydrous kahweol (m/z 296) [17]. The difference
between arabica (B) and robusta (D) during isothermal

123
1464 M. Fischer et al.

Mass weighted ion signal/a.u.

m/z 296

Mass weighted ion signal/a.u.

m/z 296
0.004 298 298
0.004
Mass weighted ion signal/a.u.

Mass weighted ion signal/a.u.

0.002 0.002

0.015 0.015
0.0 0.0
0.01 50 150 250 Isotherm 0.01 50 150 250 Isotherm
Temperature/°C Temperature/°C
0.005 0.005

0 0
245 °C 245 °C
350 350
300 300
250 250
20 min 20 min
200 200
m

m
B D
/z

150

/z
150
A C
100 100
C
OH 250 ture/° OH 250
rature
/° C
mpera
50 50 150 150 e
50
Temp
O Te O 50
(a)0.004 (c)
H3C
H3C OH
O OH O H2C
CH3 CH3
Signal intensity/a.u.

OH

Signal intensity/a.u.
OH
0.002 H3C
H2C
H3C O
0.003 O CH3 OH
CH3 H3C
OH O
H3C N
N HO
O OH
OH H 3C O O
HO O H3C
0.001 N N
CH3
N CH3
H 0.001 CH3 N
N O N H
N O N N N O
CH3
CH3 CH3

0.000 0.000
50 100 150 200 250 300 350 50 100 150 200 250 300 350
m/z m/z
(b) (d)
0.0010 O 0.0010 O
HO H3C O
N CH3
N C HO
3C

H3C O H 3
H3

OH H
H

0.0008 CH3 0.0008 N N 3C

C

N N O
Signal intensity/a.u.

OH
Signal intensity/a.u.

CH3
N N O OH
0.0006 CH3 0.0006 OH
HO
O O
HO

OH

0.0004 OH
0.0004

0.0002 0.0002

0.0000 0.0000
50 100 150 200 250 300 350 50 100 150 200 250 300 350
m/z m/z

Fig. 4 Results of the simulated roasting process of a single Arabica (b and d) after 20 min allow discrimination between Arabica and
coffee bean (left column) and an individual Robusta (right column) Robusta coffee. The mass spectra at 245 °C (a, c) reveal different
coffee bean. The extracted mass traces of the anhydrous diterpenes composition ratios
kahweol m/z 296 and cafestol m/z 298 as well as the mass spectra

treatment is obvious and as Hertz-Schünemann et al. elevation. Previous experiments showed the importance of
describe in ‘‘good accordance to the literature’’; there is no an inert atmosphere to avoid oxidation of the lipids.
anhydrous kahweol in the Robusta cultivar [17]. TG-SPI- Table 1 shows compounds that could be distinguished
TOFMS measurements depict low signal intensities for m/z according to literature.
298, too, and are in line with results on the composition of Robusta is known to be richer in chlorogenic acids
Arabica and Robusta coffees: Robusta coffee contains (CQA) than Arabica coffee (7–10 % in Robusta; 5–7.5 %
considerably less of those lipids. The mass spectra in Fig. 4 in Arabica). Thermal degradation pathways of the CQA
were averaged over 6 s, corresponding to 1 K temperature during coffee roasting are described elsewhere [28, 36].

123
Evolved gas analysis by single photon ionization-mass spectrometry 1465

Table 1 Overview of significant m/z values observed in the TG-SPI- the signals differ between Arabica and Robusta as men-
mass spectra with tentative assignments, tentative sources (precur- tioned before.
sors), and references
Hertz-Schünemann [17] et al. did micro-probe sampling
m/z Tentative assignment Tentative source Reference on-line photo-ionization mass spectrometric research with
single coffee beans during roasting. They figured out the
67 Pyrrole [26]
differences of evolved gases outside of Arabica and
79 Pyridine Trigonelline [18, 26]
Robusta coffee beans by plotting a difference mass spec-
81 Methylpyrrole [26]
trum based on the averaged SPI-mass spectra of several
82 Methylfuran [18, 26]
Arabica and Robusta roasting measurements. Comparing
94 Phenol Chlorogenic acids [18, 26]
Hertz-Schünemann’s results with a difference mass spec-
Methylpyrazine [26]
trum from the single bean Robusta and Arabica TG-SPI-
96 Furfural [18, 26]
TOFMS generated different mass spectra of measurements
C2-Furan [17]
(Fig. 5), the results on the general composition coincidence
98 Furfuryl alcohol Carbohydrates [18, 26]
very well with the literature results. The Robusta bean
110 1,2-Benzenediole Chlorogenic acids [26] roasting simulation shows higher signals on phenol deriv-
Methylfurfural [26] atives (m/z 110, 124, 150, 152) and caffeine. The phenol
124 Guaiacol Chlorogenic acids [18, 26] (m/z 94) signal itself is balanced in this study. The Arabica
Methylbenzenediole [37] bean contains more carbohydrates (celluloses and smaller
150 Vinylguaiacol Chlorogenic acids [18, 26] species) resulting in increased formation of furan deriva-
152 Vanilin [26] tives (m/z 98, 126) and fatty acids such as anhydrous
Ethylguaiacol Chlorogenic acids [26] kahweol (m/z 296) and cafestol (m/z 298).
194 Caffeine Caffeine [26] The roasting simulation results demonstrate the appli-
296 Anhydrous kahweol Kahweol [38] cability of TG-SPI-TOFMS for process investigations.
298 Anhydrous cafestol Cafestol [38] However, the setup can also be used for product and quality
control purposes. In this context further experiments dealt
with the final coffee product, namely roasted and ground
79, 84, 108, 126, coffee (R&G) powders. An extended trail of the work gave
149, 162, 175,
0.5
180, 189, 192, some promising future applications of TG-SPI-TOFMS.
228, 240, 272 The same temperature program at same conditions as for
the roasting was performed in this experiment (Fig. 6). A
15 296
0
298 pure Arabica powder, a 50/50 and 90/10 blend of Arabica/
10 Robusta and a pure Robusta powder were measured. Again
117, 128 280
278 we could confirm the knowledge of the literature with the
5
82 98 –0.5
TG-SPI-TOFMS device. Robusta gives higher signals for
50 100 150 200 250
m/z 194 (caffeine) and m/z 150 (vinylguaiacol). Therefore,
Signal intensity/a.u.

0
92 124 150 the blend, consisting of Arabica and Robusta, gives signals
–5
152 of all of the aforementioned masses. Examining the signal
–10 110 intensity of m/z 150 in a blended sample, discrimination
and a mixing ratio determination might be possible
–15
(Fig. 7). The ratios of cafestol and kahweol in these blends
–20 Arabica - Robusta of the first measurements show an obvious trend. To pic-
194 ture the trend, the ratios are plotted in Fig. 8. With an
–25
50 100 150 200 250 300 increasing share of Arabica in the blend the signal of
m/z kahweol also increased. This initial result is promising that
Fig. 5 SPI-difference mass spectrum generated by the subtraction of an a determination of the Robusta and Arabica content in
averaged and normalized Robusta single bean measurement SPI-mass R&G coffee blends is possible. Therefore, an investigation
spectra from of an averaged and normalized Arabica single bean on the ratios and the minimal detectable content of Arab-
measurement SPI-mass spectra ica, or Robusta respectively, in R&G blends will be part of
a future study.
Products of the CQA degradation such as ethylguaiacol Finally, another important class of coffee products was
(m/z 152) vinylguaiacol (m/z 150), guaiacol (m/z 124), 1.2- analyzed, namely regular and decaffeinated instant coffee
benzenediole, (m/z 110), phenol (m/z 94) are dominant (freeze drying process). Here, the EBEL was used as VUV-
components [26, 28] in the mass spectra. The intensities of photon source. In comparison to the broad band emission

123
1466 M. Fischer et al.

0.005
ine
ffe

Mass weighted ion signal/a.u.

a
c O2 Robusta m/z 150
Mass weighted ion signal/a.u.

4 4
z 19 10N
/ 8H
m C
0.01

0.008 0.003
0.006

0.004

0.002
Arabica
0.001
°C
e/
300
t ur
250
250 ra
200
150 150 pe
m
m/z 100
50 50 Te 50 150 250 Isotherm
Temperature/°C

l Fig. 7 Mass trace m/z 150 from pure Arabica (red, low intensity) and
eo Robusta (green, higher intensity) powder allows discrimination of the two
hw
Mass weighted ion signal/a.u.

a
K O2 coffees species according to their signal intensity. (Colour figure online)
2 96 H 24 l
0.01 /z 0 to
m C 2 fes
Ca O 2
Pure robusta Robusta/arabica Robusta/arabica Pure arabica
0.008 50%/50% 10%/90%
8
29 H 26
0.006 /z C 20 K/C = 0.37 K/C = 1.58 K/C = 2.28 K/C = 2.30
m
0.004 K: dehydrated C: dehydrated
kahweol: cafestol:
296 m/z 298 m/z
0.002
H3C H3C

O O

e /°C HO HO
300
250 t ur
200 250 ra
150 150 pe
100 m
m/z 50 50 Te

l
ia co 292 296 300 304 292 296 300 304 292 296 300 304 292 296 300 304
ua m/z
Mass weighted ion signal/a.u.

lg
v iny O 2
0, 10
0.01 15 C 9H 2.5
/z
m
Accumulated signal

0.008
2
intensity/a.u.

0.006
1.5
0.004

0.002 1

°C 0.5
300 e/
ur
at
250
250
er
200 0
150 150 p
m/z
100 m 0 20 40 60 80 100
50 50
Te
Share of Arabica %
Fig. 6 Roasted and ground coffee powder of Arabica (top), Robusta
(bottom) and 50/50, a/b blend (middle). While a shows signals at m/z Fig. 8 Plots of the ratios of cafestol (taken as constant for better
296, 298, b does not but has higher signal intensity at m/z 150. illustration in the top-figure) to kahweol in TG-SPI-TOFMS mea-
Consequently, blend (c) shows all three signals surements of blends of R&G coffees with different Arabica and
Robusta content. The shares of Arabica in the blends increased from
0, 50, and 90 to 100 %. By increasing the share of Arabica, an
of a deuterium lamp, the EBEL emits light with a center increasing trend in the ratios is observable, suggesting the applica-
bility of TG-SPI-TOFMS for analyzing e.g., the Robusta content in
wavelength of 126 nm (about 9.8 eV) and a bandwidth of R&G coffee blends
9 nm [39]. Instant coffee is suspected to have a lower
content of VOC and SVOC due to the losses upon the in the decaffeinated sample. Although the pattern of VOC
vacuum freeze drying process, Fig. 9 shows the TG-SPI- and SVOC is generally depleted in instant coffee m/z 79,
TOFMS results obtained from the instant coffee samples. 82, 94, (82, 95, 96), 98, (108), 110, (122), 124, and 194
As expected, the caffeine peak of m/z 194 peak is missing could be detected.

123
Evolved gas analysis by single photon ionization-mass spectrometry 1467

Mass weighted ion signal/a.u.

× 10–3
× 10–3
5 5

4 4

3 3

2 2

1 1

300 300 /°C

250 250
250 atu
re
200 250 200
150 150 150 150 per
100 m/z 100
m/z 50 50 50 m
50 Te

H3C O
0.00006 CH3 0.00006
N
N
Signal intensity/a.u.

Signal intensity/a.u.
N N O
0.00004 CH3 0.00004

0.00002 0.00002

0.00000 0.00000
100 200 300 400 500 100 200 300 400 500
m/z m/z

Fig. 9 Regular instant coffee on the left and decaffeinated instant coffee on the right. In the mass spectra at 245 °C the missing caffeine peak (m/z
194) points out the decaffeinated product

Conclusions Quality and product control and studies for improve-

ment of the roasting process in the micro-scale were
The application of TG-SPI-TOFMS for simulated coffee shown to be possible by TG-SPI-TOFMS in the field of
roasting experiments (individual coffee beans) gave coffee. However, it is necessary to build up a bigger
promising results and valuable insights in the chemical database to make coffee species and varieties distinctive
composition of the evolved roasting gases. Furthermore, due to their chemical SPI-fingerprints. Ongoing work will
every coffee product sample (roast and ground coffee include the development of an additional separation via a
powder, instant coffee) was successfully analyzed by TG- rapid modulated gas chromatography step followed by a
SPI-TOFMS. The possibility to discriminate between the combination of SPI and EI operated in fast alternation
raw coffee species Arabica or Robusta employing TG-SPI- switching [40] to be able to confirm unknown substances
TOFMS was established. The attention hereby focuses on by their fragment patterns. Finally, the approach is cer-
the discrimination between the diterpenes cafestol and tainly transferable to other ‘‘high aroma’’ food stuffs
kahweol. Further measurements and statistical analyses are including e.g., cocoa or tea.
underway to confirm the broad applicability of the
approach (i.e., to elucidate different varieties or planting Acknowledgements Funding from the Bavarian Science Founda-
tion (Bayerische Forschungsstiftung, BFS) and support from Netzsch
regions) and to determine the precision and, therefore, the
Gerätebau GmbH, Selb, Germany as well as Photonion GmbH,
limit of reliable quantification. It bears the potential for Neuherberg, Germany is gratefully acknowledged.
quality control of pure and blended coffees.

123
1468
References
1. Raemakers KGH, Bart JCJ. Applications of simultaneous thermo-
gravimetry-mass spectrometry in polymer analysis. Thermochim
Acta. 1997;295(1–2):1–58. doi:10.1016/s0040-6031(97)00097-x.
2. Chiu J. Polymer characterization by coupled thermogravimetry-
gas chromatography. Anal Chem. 1968;40(10):1516–20. doi:10.
1021/ac60266a037.
3. Chiu J, Beattie AJ. Techniques for coupling mass spectrometry to
thermogravimetry. Thermochim Acta. 1980;40(2):251–9. doi:10.
1016/0040-6031(80)87226-1.
4. Chiu J, Beattie AJ. A universal interface for coupling mass spectrometry
to thermogravimetry. Thermochim Acta. 1981;50(1–3):49–56. doi:10.
1016/0040-6031(81)85042-3.
5. Kaisersberger E, Post E. Practical aspects for the coupling of gas
analytical methods with thermal-analysis instruments. Thermochim
Acta. 1997;295(1–2):73–93. doi:10.1016/s0040-6031(97)00099-3.
6. Várhegyi G, Czégény Z, Jakab E, McAdam K, Liu C. Tobacco
pyrolysis. Kinetic evaluation of thermogravimetric–mass spec-
trometric experiments. J Anal Appl Pyrol. 2009;86(2):310–22.
doi:10.1016/j.jaap.2009.08.008.
7. Kaisersberger E, Post E. Applications for skimmer coupling systems,
combining simultaneous thermal analysers with mass spectrometers.
Thermochim Acta. 1998;324(1–2):197–201. doi:10.1016/s0040-
6031(98)00536-x.
8. Saraji-Bozorgzad MR, Streibel T, Kaisersberger E, Denner T,
Zimmermann R. Detection of organic products of polymer pyrolysis
by thermogravimetry-supersonic jet-skimmer time-of-flight mass
spectrometry (TG-Skimmer-SPI-TOFMS) using an electron beam
pumped rare gas excimer VUV-light source (EBEL) for soft photo
ionisation. J Therm Anal Calorim. 2011;105(2):691–7. doi:10.1007/
s10973-011-1383-2.
9. Tsugoshi T, Ito N, Nagaoka T, Watari K. Evolved gas analysis
with skimmer interface and ion attachment mass spectrometry for
burnout monitoring of organic additives in ceramic processing.
Talanta. 2006;70(1):186–9. doi:10.1016/j.talanta.2006.01.035.
10. Lindinger W, Hirber J, Paretzke H. An ion/molecule-reaction mass
spectrometer used for on-line trace gas analysis. Int J Mass Spectrom
Ion Process. 1993;129:79–88. doi:10.1016/0168-1176(93)87031-m.
11. Cao L, Muhlberger F, Adam T, Streibel T, Wang HZ, Kettrup A,
et al. Resonance-enhanced multiphoton ionization and VUV-single
photon ionization as soft and selective laser ionization methods for
on-line time-of-flight mass spectrometry: investigation of the pyro-
lysis of typical organic contaminants in the steel recycling process.
Anal Chem. 2003;75(21):5639–45. doi:10.1021/ac0349025.
12. Arii T, Otake S. Study on thermal decomposition of polymers by
evolved gas analysis using photoionization mass spectrometry
(EGA-PIMS). J Therm Anal Calorim. 2008;91(2):419–26. doi:10.
1007/s10973-007-8416-x.
13. Geissler R, Saraji-Bozorgzad MR, Gröger T, Fendt A, Streibel T,
Sklorz M, et al. Single photon ionization orthogonal acceleration
time-of-flight mass spectrometry and resonance enhanced multi-
photon ionization time-of-flight mass spectrometry for evolved
gas analysis in thermogravimetry: comparative analysis of crude
oils. Anal Chem. 2009;81(15):6038–48. doi:10.1021/ac900216y.
14. Mühlberger F, Wieser J, Ulrich A, Zimmermann R. Single photon
ionization (SPI) via incoherent VUV-excimer light: robust and
compact time-of-flight mass spectrometer for on-line, real-time
process gas analysis. Anal Chem. 2002;74(15):3790–801. doi:10.
1021/ac0200825.
15. Adam T, Ferge T, Mitschke S, Streibel T, Baker RR, Zimmermann
R. Discrimination of three tobacco types (Burley, Virginia and
Oriental) by pyrolysis single-photon ionisation–time-of-flight mass
spectrometry and advanced statistical methods. Anal Bioanal Chem.
2005;381(2):487–99. doi:10.1007/s00216-004-2935-0.
123
M. Fischer et al.
16. Streibel T, Mitschke S, Adam T, Weh J, Zimmermann R. Ther-
mal desorption/pyrolysis coupled with photo ionisation time-of-
flight mass spectrometry for the analysis and discrimination of
pure tobacco samples. J Anal Appl Pyrol. 2007;79(1–2):24–32.
doi:10.1016/j.jaap.2006.12.017.
17. Hertz-Schünemann R, Streibel T, Ehlert S, Zimmermann R.
Looking into individual coffee beans during the roasting process:
direct micro-probe sampling on-line photo-ionisation mass
spectrometric analysis of coffee roasting gases. Anal Bioanal
Chem. 2013;405:7083–96. doi:10.1007/s00216-013-7006-y.
18. Dorfner R, Ferge T, Yeretzian C, Kettrup A, Zimmermann R. Laser
mass spectrometry as on-line sensor for industrial process analysis:
process control of coffee roasting. Anal Chem. 2004;76(5):
1386–402. doi:10.1021/ac034758n.
19. Fischer M, Wohlfahrt S, Saraji-Bozorgzad M, Matuschek G, Post E,
Denner T, et al. Thermal analysis/evolved gas analysis using single
photon ionization. J Therm Anal Calorim. 2013;113:1667–73.
doi:10.1007/s10973-013-3143-y.
20. Mitschke S, Adam T, Streibel T, Baker RR, Zimmermann R.
Application of time-of-flight mass spectrometry with laser-based
photoionization methods for time-resolved on-line analysis of
mainstream cigarette smoke. Anal Chem. 2005;77(8):2288–96.
doi:10.1021/ac050075r.
21. Adam T, Streibel T, Mitschke S, Mühlberger F, Baker RR,
Zimmermann R. Application of time-of-flight mass spectrometry
with laser-based photoionization methods for analytical pyrolysis
of PVC and tobacco. J Anal Appl Pyrol. 2005;74(1–2):454–64.
doi:10.1016/j.jaap.2004.11.021.
22. Saraji-Bozorgzad M, Geissler R, Streibel T, Muhlberger F, Sklorz
M, Kaisersberger E, et al. Thermogravimetry coupled to single
photon ionization quadrupole mass spectrometry: a tool to
investigate the chemical signature of thermal decomposition of
polymeric materials. Anal Chem. 2008;80(9):3393–403. doi:10.
1021/ac702599y.
23. Guilhaus M, Selby D, Mlynski V. Orthogonal acceleration time-of-
flight mass spectrometry. Mass Spectrom Rev. 2000;19(2):65–107.
doi:10.1002/(sici)1098-2787(2000)19:2\65:aid-mas1[3.0.co;2-e.
24. Fujii T. A new method for thermal analysis Ion-attachment mass
spectrometry (IAMS). J Therm Anal Calorim. 2012;110(1):17–25.
doi:10.1007/s10973-011-2165-6.
25. Schramm E, Kurten A, Holzer J, Mitschke S, Muhlberger F,
Sklorz M, et al. trace detection of organic compounds in complex
sample matrixes by single photon ionization ion trap mass
spectrometry: real-time detection of security-relevant compounds
and online analysis of the coffee-roasting process. Anal Chem.
2009;81(11):4456–67. doi:10.1021/ac900289r.
26. Flament I, Bessière-Thomas Y. Coffee flavor chemistry. New
York: John Wiley & Sons; 2002.
27. Zimmermann R, Heger HJ, Yeretzian C, Nagel H, Boesl U. Application
of laser ionization mass spectrometry for on-line monitoring of volatiles
the headspace of food products: roasting and brewing of coffee. Rapid
Commun Mass Spectrom. 1996;10(15):1975–9. doi:10.1002/(sici)
1097-0231(199612)10:15\1975:aid-rcm786[3.0.co;2-x.
28. Dorfner R, Ferge T, Kettrup A, Zimmermann R, Yeretzian C.
Real-time monitoring of 4-vinylguaiacol, guaiacol, and phenol
during coffee roasting by resonant laser ionization time-of-flight
mass spectrometry. J Agric Food Chem. 2003;51(19):5768–73.
doi:10.1021/jf0341767. already read
29. Hertz-Schünemann R. On-line process monitoring of coffee roasting
by resonant laser ionisation time-of-flight mass spectrometry:
bridging the gap from industrial batch roasting to flavour formation
inside an individual coffee bean. J Mass Spectrom (accepted).
30. Urgert R, Katan MB. The cholesterol-raising factor from coffee
beans. Annu Rev Nutr. 1997;17:305–24. doi:10.1146/annurev.
nutr.17.1.305.
Evolved gas analysis by single photon ionization-mass spectrometry
31. Viani R. Coffee. Ullmann’s Encyclopedia of Industrial Chemis-
try. Weinheim: VCH Verlag; 1986.
32. Speer K, Hruschka A, Kurzrock T, Kölling-Speer I. Diterpenes in
coffee. Caffeinated beverages. ACS Symposium Series, vol 754.
American Chemical Society; 2000. p. 241-51.
33. de Roos B, Katan MB. Possible mechanisms underlying the
cholesterol-raising effect of the coffee diterpene cafestol. Curr
Opin Lipidol. 1999;10(1):41–6.
34. Wohlfahrt S, Fischer M, Saraji-Bozorgzad M, Matuschek G,
Streibel T, Post E, et al. Rapid comprehensive characterization of
crude oils by thermogravimetry coupled to fast modulated gas
chromatography–single photon ionization time-of-flight mass
spectrometry. Anal Bioanal Chem. 2013;405:7107–16. doi:10.
1007/s00216-013-7029-4.
35. Saraji-Bozorgzad MR, Eschner M, Groeger TM, Streibel T,
Geissler R, Kaisersberger E, et al. Highly resolved online
organic-chemical speciation of evolved gases from thermal ana-
lysis devices by cryogenically modulated fast gas chromatogra-
phy coupled to single photon ionization mass spectrometry. Anal
Chem. 2010;82(23):9644–53. doi:10.1021/ac100745h.
1469
36. Semmelroch P, Grosch W. Studies on character impact odorants
of coffee brews. J Agric Food Chem. 1996;44(2):537–43. doi:10.
1021/jf9505988.
37. Maier HG. Kaffee. Grundlagen und fortschritte der lebensmitte-
luntersuchung und lebensmitteltechnologie. Berlin und Hamburg:
Verlag Paul Parey, 1981.
38. Lercker G, Frega N, Bocci F, Rodriguez-Estrada MT. High resolu-
tion gas chromatographic determination of diterpenic alcohols and
sterols in coffee lipids. Chromatographia. 1995;41(1–2):29–33.
doi:10.1007/bf02274191.
39. Muhlberger F, Wieser J, Morozov A, Ulrich A, Zimmermann R.
Single-photon ionization quadrupole mass spectrometry with an
electron beam plumped excimer light source. Anal Chem.
2005;77(7):2218–26. doi:10.1021/ac048319f.
40. Eschner MS, Gröger TM, Horvath T, Gonin M, Zimmermann R.
Quasi-simultaneous acquisition of hard electron ionization and
soft single-photon ionization mass spectra during GC/MS ana-
lysis by rapid switching between both ionization methods: ana-
lytical concept, setup, and application on diesel fuel. Anal Chem.
2011;83(10):3865–72. doi:10.1021/ac200356t.
123