On-line real-time measurements at incineration plants: PAHs and a PCDD/F surrogate compound at stationary combustion conditions and during transient emission puffs

Article abstract of DOI:10.1016/S0045-6535(00)00243-5

Laser mass spectrometry has been applied for on-line monitoring of traces of aromatic compounds from flue gas of incineration plants. The experiments have been carried out at two sampling sites in an industrial hazardous-waste incinerator. With laser mass spectrometry resonance-enhanced multiphoton ionization (REMPI) with time-of-flight mass spectrometry (TOFMS) (REMPI-TOFMS), using the group selective multi-component monitoring approach, aromatic compounds are selectively ionized from the complex flue-gas matrix. In this case, the result of an REMPI-TOFMS on-line measurement is a distinct pattern of aromatic compounds. These patterns are dependent on: (i) the point of measurement, (ii) the incineration plant, (iii) the temperature, and (iv) the fuel. This contribution focuses on the fuel dependence of the pattern. The most transient behavior can be observed when containers filled with hazardous waste are burnt, leading to puffs. Real-time monitoring results of puffs are given. Furthermore, as an approach towards on-line monitoring of the TEQ (PCDD/F toxicity equivalent), REMPI-TOFMS on-line analysis results of chlorobenzene are presented.

Full text of DOI:10.1016/S0045-6535(00)00243-5

Chemosphere 42 (2001) 691±696
On-line real-time measurements at incineration plants: PAHs
and a PCDD/F surrogate compound at stationary combustion
conditions and during transient emission pu€s
a
H.J. Heger , R. Zimmermann
a,b,
a
, M. Blumenstock , A. Kettrup
a,b
*
a

GSF Forschungszentrum f u r Umwelt und Gesundheit GmbH, Institut f u r Okologische Chemie, Ingolst a dter Landstr. 1, D-85764,
Neuherberg, Germany
b

Technische Universit a t M u nchen, Lehrstuhl f u r Okologische Chemie und Umweltanalytik, D-85748, Freising, Germany
Abstract
Laser mass spectrometry has been applied for on-line monitoring of traces of aromatic compounds from ¯ue gas of
incineration plants. The experiments have been carried out at two sampling sites in an industrial hazardous-waste
incinerator. With laser mass spectrometry resonance-enhanced multiphoton ionization (REMPI) with time-of-¯ight
mass spectrometry (TOFMS) (REMPI±TOFMS), using the group selective multi-component monitoring approach,
aromatic compounds are selectively ionized from the complex ¯ue-gas matrix. In this case, the result of an REMPI±
TOFMS on-line measurement is a distinct pattern of aromatic compounds. These patterns are dependent on: (i) the
point of measurement, (ii) the incineration plant, (iii) the temperature, and (iv) the fuel. This contribution focuses on the
fuel dependence of the pattern. The most transient behavior can be observed when containers ®lled with hazardous
waste are burnt, leading to pu€s. Real-time monitoring results of pu€s are given. Furthermore, as an approach towards
on-line monitoring of the TEQ (PCDD/F toxicity equivalent), REMPI±TOFMS on-line analysis results of chloro-
benzene are presented. Ó 2001 Elsevier Science Ltd. All rights reserved.
Keywords: REMPI; TOFMS; Online; PAH; Dioxin; Monitoring; Combustion
1
. Introduction
generate hazardous waste themselves. In fact, a major
part of the investment and operational costs of modern
incineration plants is due to the emission-reduction de-
vices (Thom ꢀe -Kozmiensky, 1994).
The toxic by-products emitted from combustion
processes in trace amounts are of still growing public
interest, since more than 90% of the total energy
worldwide (including trac) is generated by combustion
processes (Warnatz et al., 1996). In industrial combus-
tion plants, several ¯ue gas-cleaning measures ensure
that the legal emission limits are not exceeded. Flue gas-
cleaning units are expensive to build and run, and often
Recent developments focus on the optimization of
the primary combustion process (Beckmann et al.,
1997). For improving the knowledge about the process,
time-resolved ¯ue-gas analysis has to be performed. So
far, on-line analysis methods existed only for small
molecules in high concentrations, e.g. carbon monoxide.
The measurement of the organic pollutants in trace
amounts requires a sensitive and selective method, as
otherwise, interferences would prevent a peak allocation
or quanti®cation.
*
Corresponding author. Tel.: +49-89-3187-4544; fax: +49-
8
9-3187-3371.
E-mail address: ralf.zimmermann@gsf.de (R. Zimmer-
mann).
In general, laser-based methods can provide a
high degree of selectivity. Several methods have been
0045-6535/01/$ - see front matter Ó 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 5 - 6 5 3 5 ( 0 0 ) 0 0 2 4 3 - 5

6
92
H.J. Heger et al. / Chemosphere 42 (2001) 691±696
successfully applied to on-line measurements of ¯ue gas
Dreizler et al., 1997). For larger, aromatic molecules at
(
ppb concentrations and lower, laser mass spectrometry
has proven to be a successful method for on-line analysis
from the complex ¯ue-gas matrix (Zimmermann et al.,
1
997b). In the following, laser mass spectrometry results
are presented and analyzed with respect to the in¯uence
of the fuel composition.
Additionally, an indirect monitoring of the interna-
tional toxicity equivalent (I-TEQ, summed toxicity of
the polychlorinated dioxin and furan congeners) by us-
ing chlorobenzene as the indicator parameter compound
is performed. By this, the formation of chlorinated or-
ganic species is also observable.
Fig. 1. Setup of the hazardous waste incineration plant (ESP).
The thermal power of the plant is about 22 MW. The two
sampling points for the on-line measurement are marked as A
and B.
2
. Materials and methods
measurement have been recorded (i.e. every single shot
mass spectra is available for data processing). Further,
data processing allows di€erent presentations of the re-
sults as e.g., time-concentration pro®le or contour plot.
A detailed description of the instrument and the data
analysis including quanti®cation is already published
elsewhere (Heger et al., 1999b).
Laser mass spectrometry combines the selective ion-
ization method resonance-enhanced multiphoton ion-
ization (REMPI) with time-of-¯ight mass spectrometry
(
TOFMS). The method is described in detail in several
publications, e.g. Lubman (1990), Boesl et al. (1994),
Zimmermann et al. (1994), Gittins et al. (1997) and
Heger et al. (1999b). REMPI can be carried out in two
degrees of selectivity, depending on the spectroscopic
properties of the molecules which can be in¯uenced by
the molecular beam-inlet system into the mass spec-
trometer. The two achievable selectivities are (i) group-
selective ionization of a group of molecules out of a
substance class (e€usive beam sample inlet, (Heger et al.,
The measurements have been performed at an in-
dustrial hazardous-waste incinerator (22 MW). The
setup of the plant is depicted in Fig. 1. This plant con-
sists of rotary kilns and a separate post-combustion
chamber. A boiler is applied for steam generation. These
parts of the primary combustion process are followed by
¯ue-gas cleaning devices (dust precipitator, washer, etc.)
In previous on-line monitoring experiments, it turned
out that the dust precipitator will cause extensive
memory e€ects in the plant, i.e. a short concentration
peak due to a short deviation of the primary combustion
process will cause an increased concentration after the
dust precipitator for a longer time (Zimmermann et al.,
1997b). Thus, the points of measurement have been
chosen close to the exit of the post combustion chamber
(A) at a ¯ue-gas temperature of 950°C and in the boiler
in front of the ®rst dust precipitator (B) at a ¯ue-gas
temperature of 350°C.
1
999b)), and (ii) species-selective ionization of one single
species (supersonic beam sample inlet, (Zimmermann
et al. (1994)). In the experiments described in this con-
tribution, the ®rst approach has been applied. Note that
both degrees of REMPI selectivity are highly selective
compared to, e.g. direct inlet mass spectrometry with
electron-impact ionization.
The substance class of molecules is chosen by the
spectral range of the laser light utilized for ionization.
Here, UV light in the range of absorption of the p
electron system of aromatic compounds has been used
(
Nd:YAG Laser, 266 nm). A group of aromatic mole-
The sampling has been performed by drawing ¯ue gas
through a heated quartz-glass tube and a specially de-
veloped glass-wool dust ®lter into the MS. Details of the
sampling system are described in Heger et al. (1999b). It
should be emphasized here that the amount of analyte
molecules adsorbed on particles can be neglected, since
all the target compounds are nearly completely in the gas
phase due to the high ¯ue-gas temperatures.
cules is thus selectively ionized, each molecule with an
individual cross-section depending on its UV spectrum.
Emphasis can be laid on a target compound by using a
tunable laser system and tuning the wavelength to the
maximum REMPI cross-section of this compound to
enhance sensitivity. This combination of group-selective
analysis and target compound monitoring has been de-
veloped and applied to the simultaneous measurement
of PAHs and chlorobenzene (Zimmermann et al.,
1999a). A detailed discussion on the two concepts is
given by Heger et al. (1999a).
3. Results and discussion
With a novel data-acquisition concept, the complete
information about the group-selective REMPI±TOFMS
Results of two conceptionally di€erent on-line
experiments are presented. The aim of the ®rst experi-

H.J. Heger et al. / Chemosphere 42 (2001) 691±696
693
ment was to gain information about the combustion
process using the PAHs as products of incomplete
combustion. The latter experiments aimed at achieving
an indirect on-line measurement of the I-TEQ and thus,
to measure the dioxin/furan-related toxicity on-line.
Compared to the electron impact, these fragment peaks
are very small. No fragmentation into a mass smaller
than 78 amu has been observed. In summary, frag-
mentation only makes a small contribution to the mass
spectra.
The upper mass spectrum was recorded when the
plant used heating oil as fuel for at least 3 h, the lower
one was recorded after using hazardous waste as fuel for
at least 3 h too. A shift in the di€erent aromatic skele-
tons can be clearly observed: with heating oil, the peaks
of the alkylated naphthalenes are more abundant in the
spectrum, while with hazardous waste, there is a slight
predominance of phenanthrenes. Furthermore, a di€er-
ence in alkylation can be observed for benzenes: with
heating oil, xylene is the most prominent peak among
the alkylated benzenes, while with hazardous waste, the
highest peak is that of toluene. This result shows that the
pattern of the products of incomplete combustion
depends on the fuel used. A process optimization for a
speci®c plant has to consider the fuel properties and,
especially for inhomogeneous fuels such as, for e.g.,
waste, on-line monitoring is an important tool for pro-
cess understanding and optimization.
3.1. Incomplete combustion: Patterns of polycyclic aro-
matic compounds
For a group-selective measurement of aromatic
compounds, REMPI at 266 nm was performed using the
7
�2
fourth harmonic of a Nd:YAG laser (10 W cm ).
Fig. 2 shows two mass spectra obtained with REMPI
at 266 nm at sampling point 1. The mass range covers
the one to ®ve-ring aromatics; the dominant peaks are
naphthalene and phenanthrene, and their alkylated de-
rivatives. Next to these, benzene, styrene and phenol are
the most prominent peaks. In case of ambiguous peak
allocation, the REMPI cross-section and the abun-
dances of isobaric substances in waste-incineration ¯ue
gas (taken from Jay and Stieglitz (1995)) was taken into
account. In the following, peaks are assigned according
to the most contributing substance. A note on the
fragmentation observed under laser-analytical condi-
tions should be added: alkylated aromatic compounds
A prominent example for inhomogeneous fuel feed-
ing are containers ®lled with liquid hazardous waste that
are directly put into the rotary kiln leading to strong
pu€s. The problem is well known and special containers
have been suggested to reduce the pu€s (Wendt, 1994).
However, a research on the exact behavior of the con-
tainers during incineration is dicult without highly
time-resolved on-line data for more complex products of
incomplete combustion.
(
mass M) show little fragmentation to M �14 amu and
M �15 amu, as well as to M �1 amu fragments.
In Fig. 3, REMPI±TOFMS on-line measurements of
a transient pu€ due to barrel uptake are presented. The
mass spectra of the single measurement are arranged in a
contour plot to show the development of the pattern
with time. The measurements have been performed with
1
0 Hz repetition rate, allowing a time resolution of 200
m. The main pu€ is observed at about 50 s measurement
time with a strong peaking of the aromatic skeletons
from benzene (78 amu) to naphthalene (128 amu), and
phenanthrene (178 amu). Of these, only phenanthrene
shows a broad emission peak. This is in accordance with
previous measurements, where larger PAHs showed a
memory e€ect in the plant that was interpreted as an
e€ect of the wall surface. An e€ect of the sampling sys-
tem can be excluded (Heger et al., 1999b). The alkylated
benzenes show, furthermore,
a delayed maximum
emission output. This is a hint that there are di€erent
temperature zones involved in the formation of these
compounds, as degrees of alkylation vary with temper-
ature (Blumer, 1976).
Besides the degree of alkylation, there are further
di€erences between the ``on-peak'' chemistry and the
``after peak'' chemistry of formation of aromatic trace
compounds: during the pu€, a strong peak of an
Fig. 2. Two mass spectra are recorded at the sampling point A.
The upper one is recorded with heating oil as fuel, and the lower
one with hazardous waste as fuel.

6
94
H.J. Heger et al. / Chemosphere 42 (2001) 691±696

a fast screening came up, ®rst proposed by Oberg and
Bergstr o m (1989) and investigated in detail by Kaune et
al. (1994). With this concept, I-TEQ analysis is reduced
to the analysis of one target compound, which should be
easily accessible with conventional analysis, e.g. HPLC
methods. In these investigations, mainly highly chlori-
nated benzenes have been considered to be TEQ indi-
cator compounds.
A further development is to analyse an indicator
compound for I-TEQ on-line. To apply REMPI for se-
lective ionization of chlorinated aromatics in combustion
¯ue gases had been proposed a decade ago (Rohl®ng,
1
988); REMPI detection limits for a variety of candidates
for on-line measurement in a combustion ¯ue gas have
been determined by Cool and coworkers (Williams et al.,
1
992; Tanada et al., 1994). Recent developments have
improved the detection limit to the pptv range (Oser et
al., 1996; Heger et al., 1999b). The REMPI spectra of
chlorinated aromatics have been investigated in detail
Fig. 3. Contour plots recorded at the sampling point B. The
peaks at 50 s are due to a pu€ from container incineration.
Following the pu€, the concentration of methylated aromatics
was raised for a few minutes. Peaks recorded after 100 s are
enhanced ®ve times.
(
Zimmermann et al., 1994, 1996) and it turned out that,
for spectroscopic reasons, mono and dichlorobenzenes
are the best suited for REMPI on-line measurement.
Recent conventional GC-MS measurements revealed
that monochlorobenzene is a good indicator parameter
for the 1-TEQ (Blumenstock et al., 1999, 2000; Zim-
mermann et al., 1999a).
oxygenated compound, phenol, is observed and a series
of products that can be interpreted as intermediates in
the formation of higher aromatics from benzene show
concentration peaks as well. Examples are styrene and
phenylacetylene. The most prominent peaks are marked
in the contour plot.
The REMPI spectrum of monochlorobenzene (e€u-
sive inlet) has been recorded (Heger et al., 1999a), and
0
the maximum UV absorption of the 0 -transition of the
1
0
1
B
2
ꢀS
1
†
A ꢀS †band at 269.82 nm has been identi-
ed as a suited intermediate state for REMPI±TOFMS
1
0
®
detection. A dye laser was tuned to the respective
wavelength for the following on-line measurement,
performed at sampling point 2.
The variety in the behavior of individual compounds
observed during and after the pu€ shows the complexity
of the process, and the ability of on-line REMPI±
TOFMS measurements to monitor such complex
processes using multiple parameters that are measured
simultaneously with every laser shot.
3.2. On-line measurement of MCB as ``PCDD/F''
indicator
The polychlorinated dioxin and furan congeners are in
the ppqv-concentration range and thus, are far beyond
the detection limits of todayÕs on-line methods. Further-
more, the determination of the most interesting value, the
I-TEQ, would require a simultaneous but selective mea-
surement of all toxic congeners. Although REMPI±
TOFMS selectivity would help solving a task like this, and
might be applied to improve analysis time by using it as a
selective detector for HRGC±MS (Zimmermann et al.,
1
997a, 1999b), a real-time on-line measurement of the I-
TEQ requiring the simultaneous measurement of the 17
toxic PCDD/F congeners is not possible.
Owing to the complexity of TEQ-analysis, the idea of
using indicator parameter compounds for the I-TEQ for
Fig. 4. Monochlorobenzene (a surrogate compound of the
I-TEQ) measured on-line with REMPI±TOFMS.

H.J. Heger et al. / Chemosphere 42 (2001) 691±696
695
Fig. 4 shows the time-concentration pro®le of the
and chlorobenzene in ¯ue gases of combustion processes.
Eur. Mass Spectr. 5, 51±57.
3
7
chlorobenzene molecular peak (112 amu) and the Cl
isotope (114 amu). The development in time, of the
signals of the mass range 110±115 amu is depicted in 3-d
presentation. The chlorobenzene concentration has been
quanti®ed using an external gas standard and is in the
range of 500 ppt in the depicted trace. A transient be-
havior at about 300 s measurement time can be observed
at both 112 and 114 amu. Via the indicator parameter
relation (Blumenstock et al., 1999, 2000)), an estimation
of the I-TEQ is possible on-line, in real-time.
Heger, H.J., Zimmermann, R., Dorfner, R., Beckmann, M.,
Griebel, H., Kettrup, A., Boesl, U., 1999b. On-line emission
analysis of polycyclic aromatic hydrocarbons down to pptv
concentration levels in the ¯ue gas of an incineration pilot
plant with a mobile resonance-enhanced multiphoton ion-
ization time-of-¯ight mass spectrometer. Anal. Chem. 71,
46±57.
Jay, K., Stieglitz, L., 1995. Identi®cation and quanti®cation of
volatile organic compounds in emissions of waste inciner-
ation plants. Chemosphere 30, 1249±1260.
Kaune, A., Lenoir, D., Nikolai, U., Kettrup, A., 1994.
Estimating concentrations of polychlorinated dibenzo-p-
dioxins and dibenzofurans in the stack gas of a hazardous
waste incinerator from concentrations of chlorinated benz-
enes and biphenyls. Chemosphere 29, 2083±2096.
In summary, the results show that a comprehensive
characterization of combustion processes can be
achieved by REMPI±TOFMS on-line monitoring.
Lubman, D.M. (Ed.), 1990. Lasers and Mass Spectrometry.
Oxford University Press, New York.
Acknowledgements

Oberg, T., Bergstr o m, J., 1989. Indicator parameters for
PCDD/PCDF. Chemosphere 19, 337±344.
We thank Dr. U. Boesl, TU M u nchen, for the sup-
port and helpful discussions. The research was funded
by the ``Deutsche Bundesstiftung Umwelt'', Osnabr u ck.
HJH thanks the ``Max-Buchner-Stiftung der DECHE-
MA'', Frankfurt, for the research scholarship.
Oser, H., Thanner, R., Grotheer, H.-H., 1996. Jet-REMPI for
the detection of trace gas compounds in complex gas
mixtures, a tool for kinetic research and incinerator process
control. Combust. Sci. Technol. 116±117, 567±582.
Rohl®ng, E.A., 1988. Resonantly enhanced multiphoton ion-
ization for the trace detection of chlorinated aromatics.
In: Proceedings of the 22nd International Symposium
on Combustion. The Combustion Institute, Pittsburgh,
pp. 1843±1850.
References
Beckmann, M., Scholz, R., Wiese, C., Busch, M., Peppler, E.,
Tanada, T.N., Velazquez, J., Hemmi, N., Cool, T.A., 1994.
Surrogate detection for continuous emission monitoring by
resonance ionization. Combust. Sci. Technol. 101, 333±348.
Thom ꢀe -Kozmiensky, K.J., 1994. Thermische Abfallbehand-
lung, EF Verlag f u r Energie-und Umwelttechnik, Berlin.
Warnatz, J., Maas, U., Dibble, R.W., 1996. Combustion.
Springer, Berlin.
1
1
997. In: Fifth International Furnace and Boiler Conference
997 (INFUB), Porto-Espinho, Portugal.
Blumenstock, M., Zimmermann, R., Schramm, K.-W., Kaune,
A., Nikolai, U., Lenoir, D., Kettrup, A., 1999. Estimation
of the dioxin emission (PCDD/F I-TEQ) from the concen-
tration of low chlorinated aromatic compounds in the ¯ue
and stack gas of a hazardous waste incinerator. J. Anal.
Appl. Pyrolysis 49, 179±190.
Wendt, J.O.L., 1994. Combustion science for incineration
technology. In: Proceedings of the 25th International
Symposium on Combustion. The Combustion Institute,
Pittsburgh, pp. 277±298.
Blumenstock, M., Zimmermann, R., Schramm, K.-W., Kettrup,
A., 2000. Identi®cation of surrogate compounds for the
emission of PCDD/F (I-TEQ value) and evaluation of their
on-line real-time detectability in ¯ue gases by REMPI±
TOFMS laser mass spectrometry. Chemosphere 42, 507±518.
Blumer, M., 1976. Polycyclic aromatic compounds in nature.
Sci. Am. 234, 34±45.
Williams, B.A., Tanada, T.N., Cool, T.A., 1992. Resonance
ionization detection limits for hazardous emissions.
In: Proceedings of the 24th International Symposium
on Combustion. The Combustion Institute, Pittsburgh,
pp. 1587±1596.
Boesl, U., Weinkauf, R., Weickhardt, C., Schlag, E.W., 1994.
Laser ion sources for time-of-¯ight mass spectrometry. Int.
J. Mass Spectr. Ion Processes 131, 87±124.
Zimmermann, R., Boesl, U., Heger, H.J., Rohwer, E.R.,
Ortner, E.K., Schlag, E.W., Kettrup, A., 1997a. Hyphen-
ation of gas chromatography and resonance-enhanced laser
mass spectrometry (REMPI±TOFMS): A multidimensional
analytical technique. J. High Resolution Chromatogr. 20,
461±470.
Dreizler, A., Sick, V., Wolfrum, J., 1997. Applied laser
spectroscopy in technical combustion systems. Ber. Bun-
sen-Ges. Phys. Chem. 101, 771±782.
Gittins, C.M., Castaldi, M.J., Senkan, S.M., Rohl®ng, E.A.,
1
Zimmermann, R., Heger, H.J., Kettrup, A., Boesl, U., 1997b. A
mobile resonance-enhanced multiphoton ionization time-of-
¯ight mass spectrometry device for on-line analysis of
aromatic pollutants in waste incinerator ¯ue gases: First
results. Rapid Commun. Mass Spectr. 11, 1095±1102.
Zimmermann, R., Heger, H.J., Blumenstock, M., Dorfner, R.,
Schramm, K.-W., Boesl, U., Kettrup, A., 1999a. On-line
monitoring of chlorobenzene in the ¯ue gas of a full-scale
hazardous waste incinerator with a mobile REMPI-time-of-
997. Real-time quantitative analysis of combustion-gener-
ated polycyclic aromatic hydrocarbons by resonance-en-
hanced multiphoton ionization time-of-¯ight mass
spectrometry. Anal. Chem. 69, 286±293.
Heger, H.J., Boesl, U., Zimmermann, R., Dorfner, R., Kettrup,
A., 1999a. On-line REMPI±TOFMS laser mass spectrom-
etry for combined multi-component-pattern analysis and
target-compound monitoring: Non-chlorinated aromatics

6
96
H.J. Heger et al. / Chemosphere 42 (2001) 691±696
¯
ight mass spectrometer indirect on-line measurement of the
by laser-induced resonance-enhanced multi-photon ioniza-
tion/mass Spectrometry. In: Proceedings of the 26th Inter-
national Symposium on Combustion. The Combustion
Institute, Pittsburgh, pp. 2859±2868.
emission of chlorinated dioxins (I-TEQ value). Rapid
Commun. Mass Spectr. 13, 307±314.
Zimmermann, R., Rohwer, E.R., Heger, H.J., 1999b. In-line
catalytic derivatization method for selective detection of
chlorinated aromatics with a hyphenated gas chromatogra-
phy laser mass spectrometry technique: A concept for
comprehensive detection of isomeric ensembles. Anal.
Chem. 71, 4148±4153.
Zimmermann, R., Weickhardt, C., Boesl, U., Lenoir, D.,
Schramm, K.-W., Kettrup, A., Schlag, E.W., 1994. Isomer-
selective ionization of chlorinated aromatics with lasers for
analytical time-of-¯ight mass spectrometry; ®rst results for
polychlorinated dibenzo-p-dioxins (PCDD), biphenyls
(PCB) and benzenes (PCBz). Chemosphere 29, 1877.
Zimmermann, R., Lenoir, D., Kettrup, A., Nagel, H., Boesl,
U., 1996. On-line emission control of combustion processes