Formation of PCDD/Fs in the sintering process: Influence of the raw materials

Article abstract of DOI:10.1021/es034679t

The sintering process is among the major sources of PCDD/Fs in the environment. This research studies the influence of the raw materials in this type of industrial plant on the amounts of PCDD/Fs generated. Particular interest is given to coke, which constitutes the principal source of carbon for the de novo synthesis of PCDD/Fs, and to the dust collected in the electrostatic precipitator (E.S.P. dust), usually recycled in the raw materials. The de novo synthesis of PCDD/Fs is simulated at the laboratory scale by thermal treatments of the samples. The use of a particular coke as a fuel does not drastically reduce the formation of PCDD/Fs. Actually, the global amounts of PCDD/Fs generated from the graphite and the two cokes tested are very similar. Only modifications in the fingerprint are observed. On the other hand, the addition of 10 wt % dust collected in the electrostatic precipitator leads to the formation of amounts of PCDD/Fs multiplied by a factor larger than 103. These results imply caution against the recycling of this E.S.P. dust in the raw materials.

Full text of DOI:10.1021/es034679t

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Environ. Sci. Technol. 2004, 38, 4222-4226
coke breeze, and lime and moves slowly. Burners initiate the
process by igniting the feed layer on top, and ambient air
sucked through the layer moves the flame front downward.
The sinter is then cooled, broken, and screened before being
charged into the blast furnace. Separate chambers, called
wind boxes and located below the strand, collect the off-gas
prior to filtering in appropriate dust collectors (usually
electrostatic precipitators). The detailed mechanism(s) as
well as the spot(s) of PCDD/ Fs formation in the sintering
process remains unknown, although all of the necessary
ingredients are present: carbon from the coke, oxygen in
the air sucked through the cake, chlorine, and catalytic metals
available in the ores.
Formation of PCDD/Fs in the
Sintering Process: Influence of the
Raw Materials
C EÄ L I N E X H R O U E T * A N D
E D W I N D E P A U W
Mass Spectrometry Laboratory, University of Lie`ge,
B6c Sart Tilman, B-4000 Lie`ge, Belgium
The European Commission points out this metallurgical
process as the most worrying in terms of PCDD/ Fs emissions
and underlines the necessity of corrective actions (6). The
huge amounts of fumes emitted (more than 106 Nm3/ h) with
high concentration of PCDD/ Fs lead to a real need to reduce
and control the discharge of these pollutants from this
process. This requires the understanding of the formation
spots and formation mechanisms of these pollutants in the
process as well as the knowledge of the factors influencing
this formation.
The sintering process is among the major sources of
PCDD/Fs in the environment. This research studies the
influence of the raw materials in this type of industrial plant
on the amounts of PCDD/Fs generated. Particular interest
is given to coke, which constitutes the principal source
of carbon for the de novo synthesis of PCDD/Fs, and to the
dust collected in the electrostatic precipitator (E.S.P.
dust), usually recycled in the raw materials. The de novo
synthesis of PCDD/Fs is simulated at the laboratory scale by
thermal treatments of the samples. The use of a particular
coke as a fuel does not drastically reduce the formation
of PCDD/Fs. Actually, the global amounts of PCDD/Fs
generated from the graphite and the two cokes tested
are very similar. Only modifications in the fingerprint are
observed. On the other hand, the addition of 10 wt % dust
collected in the electrostatic precipitator leads to the
formation of amounts of PCDD/Fs multiplied by a factor
larger than 103. These results imply caution against the
recycling of this E.S.P. dust in the raw materials.
The goal of this research is to point out critical elements
that influence PCDD/ Fs generation during the sintering
process.
The coke added as a fuel in the feed is the major source
of carbon in the process for the de novo synthesis of PCDD/
Fs. Different authors have shown that the carbon morphology
(amorphous-graphitic structure) is an important factor in
the de novo synthesis of PCDD/ Fs (7-13). The crystal lattice
of graphite is obviously resistant to an attack of chlorine/
oxygen to a greater extent than other carbon structures, which
consist in part of amorphous carbon and of microcrystalline
carbon with a degenerated graphitic structure of disoriented
layers. The way the coke is prepared (temperature and coking
time) can certainly influence the structure of the coke, and
thus perhaps its ability to produce PCDD/ Fs by de novo
synthesis.
This study describes the thermal behavior of one graphite
and two cokes differing by their coking time (9 or 15 h at
1325 °C) relating to the de novo synthesis of PCDD/ Fs. The
homologue and isomer distributions are also considered to
attain a better understanding of the formation mechanism
of these compounds.
The sintering process also allows the recycling of various
residues (iron containing dust and sludges). The dust
collected in the electrostatic precipitator (E.S.P. dust) often
constitutes a component of the raw materials. This E.S.P.
dust presents a high capacity to form PCDD/ Fs by de novo
synthesis (14, 15, 17). This complex matrix contains carbon
and chlorine as well as different metals supposed to catalyze
the de novo synthesis of PCDD/ Fs. The presence of recycled
E.S.P. dust in the raw materials may increase the amounts
of PCDD/ Fs generated during the sintering process. In the
present work, the recycling of E.S.P. dust is simulated by
adding E.S.P. dust to the samples (cokes or graphite) before
thermal treatment. Preliminary results of our investigation
have been published previously (16, 17).
Introduction
Polychlorinated dibenzo-p-dioxins (PCDDs) and polychlo-
rinated dibenzofurans (PCDFs) are highly toxic compounds
produced by some natural processes and different human
activities, especially combustion processes. Reducing human
risk toward PCDD/ Fs implies a contamination monitoring
of the food chain, but also the reduction of the emissions of
these pollutants in the environment. Identification and
control of the sources are thus essential.
Anthropogenic sources of PCDD/ Fs include the incinera-
tion of waste and most combustion processes. Since the
discovery of PCDD/ Fs in the flue gas and fly ash of municipal
waste incinerators (1), strict emission limits have been set
at the European level for these facilities, and improvements
of the purification systems in these installations have strongly
reduced PCDD/ Fs emissions from incineration (2). Metal-
lurgical processes, and in particular the sintering process,
are still significant sources of these pollutants in the
environment. In most European countries, the sintering
process is now recognized as the major source of PCDD/ Fs
(3-5).
The sintering process is an essential step in an integrated
iron and steel plant. In this process, the iron ore is converted
to larger lumps able to be charged into the blast furnaces.
The sinter plant consists of a 50-100 m long, 3-5 m wide,
horizontal strand, which supports the feed of hematite ores,
Experimental Section
Materials. The following materials were used: a solution of
2,3,7,8-Cl-substituted 13C12-labeled PCDD/ Fs (EPA 1613 LCS,
Campro Scientific, Veenendaal, The Netherlands); toluene
(p.a., Baker); hexane (p.a., Baker); dichloromethane (p.a.,
Vel); dodecane (Merck); sulfuric acid (95-97%, Baker);
* Corresponding author phone: +32-4-366-34-22; fax: +32-4-
366-43-87; e-mail: c.xhrouet@ulg.ac.be.
9
4 2 2 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 15, 2004
10.1021/es034679t CCC: $27.50
2004 Am erican Chem ical Society
Published on Web 06/29/2004

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
performed. The native concentration was determined by
isotopic dilution using the 2,3,7,8-Cl-substituted labeled
PCDD/ Fs to quantify all of the native isomers within
homologues, assuming equal response for all isomers within
an isomer group and no isomer-selective losses during the
cleanup. Because 13C-OCDF was not present in the solution
standard, this congener was not quantified. The different
congeners were identified according to ref 18.
TABLE 1. Elemental Analysis of the Original E.S.P. Dust
E.S.P. dust com position (in w t %)
Mg
1.04
Cr
Al
Si
3.62
Fe
P
S
Cl
4.07 9.55 9.07 7.83
Zn Pb
0.34 5.98 3.34
K
Ca
2.17
Mn
0.24
Cu
C
0.04
0.31
49.90
0.17
Results and Discussion
Graphite and Coke Exam ination. The graphite and the two
cokes were examined for their PCDD/ Fs content. The
amounts found are very low: between 0.12 ( 0.02 and 0.15
( 0.01 ng/ g for the PCDDs and between 1.15 ( 0.27 and 1.74
( 0.22 ng/ g for the PCDFs. The PCDF/ PCDD ratios, as well
as the homologue and isomer distributions (not shown), do
not differ for the three types of samples.
Com parison of the Graphite and the Two Cokes Relating
to de Novo Synthesis of PCDD/Fs. Laboratory experiments
were carried out with the three types of samples to simulate
a de novo synthesis of PCDD/ Fs. The experiments were
performed at 400 °C during 2 h: these are the experimental
conditions corresponding to optimal temperature and reac-
tion time found for the de novo synthesis of PCDD/ Fs on
sinter plant E.S.P. dust (15). In a first set of experiments, the
samples were used directly as such. Because these tests
produce very low amounts of PCDD/ Fs, new experiments
were performed; in a second set of experiments, the samples
were preliminarily mixed with 10 wt % of extracted dust
collected in the electrostatic precipitator of a sinter plant.
This E.S.P. dust, known to be very active in the de novo
synthesis (15), was added to the samples to bring the
necessary catalysts and to approach industrial conditions.
The results are presented in Figure 1.
Figure 1 clearly shows that the global amounts of PCDD/
Fs generated during the thermal treatment are nearly identical
for the three kinds of carbon tested. With or without the
addition of E.S.P. dust, the nature of the carbon source has
no influence on the global amounts of PCDD/ Fs formed in
the different experiments. Note that, in the case of added
E.S.P dust, the amounts produced from the carbons are not
insignificant as compared to those coming from the carbon
in the E.S.P. dust. Indeed, the E.S.P. dust used in these
experiments presents a great capacity to form PCDD/ Fs by
simple thermal treatment. This fact has been demonstrated
previously (15): during thermal treatment of this E.S.P. dust
at 400 °C (2 h), around 741 ( 13 ng/ gE.S.P.dust of PCDD/ Fs is
generated. The thermal treatment of the graphite or the two
cokes in the presence of 10% E.S.P. dust leads to the formation
of 2211 ( 115 ng/ gsample of PCDD/ Fs, corresponding to 22 100
ng/ gE.S.P.dust of PCDD/ Fs. The amounts generated are more
considerable than in the case of E.S.P. dust alone. Thus, the
PCDD/ Fs measured in the experiments with the graphite
and the cokes do not derive exclusively from the E.S.P. dust,
but also from the degradation of the carbon structure
(graphite or cokes). It can thus be concluded that the three
carbons tested present the same ability to produce PCDD/
Fs by de novo synthesis. Moreover, the amounts of PCDD/
Fs formed from the pure samples (pure graphite, coke 9 h,
or coke 15 h) are very low.
sodium chloride (p.a., Merck); potassium hydroxide (p.a.,
Merck); sodium sulfate anhydrous (Baker); aluminum oxide
(activated, neutral, type 507c, Aldrich); glass wool (DMCS
treated, Alltech Europe); and technical dry air (Air Liquide,
Belgium).
Cokes. Two different cokes (Westfalen) and one graphite
(Fluka, purum powder <0.1 mm) were used in the different
experiments. The two cokes differ by their coking time: 9 or
15 h at 1325 °C. They are respectively called “coke 9 h” and
“coke 15 h” in the rest of the paper.
E.S.P. Dust. E.S.P. dust, described in Table 1, was collected
in the electrostatic precipitator of a Belgian sinter plant. This
electrostatic precipitator consists of three fields and operates
at 120-130 °C. The E.S.P. dust used in this study comes from
field 3 and was stored at ambient temperature prior to lab
experiments. Around 72.5 wt % of the E.S.P. dust has a size
of under 40 µm.
All experiments were conducted with extracted E.S.P. dust,
cokes, or graphite to minimize potential interferences from
adsorbed organic precursors and native PCDD/ Fs. Prior to
experiments, all samples were Soxhlet extracted with toluene
(2 × 24 h), rinsed with hexane, and air-dried at room
temperature.
Experim ental Apparatus. First, 5 g of sample was packed
into a horizontal glass tube reactor (16 cm long, 3 cm
diameter) with glass wool as plugs. The tube was placed in
a chromatographic furnace, and the samples were heated
under a flow of technical air (100 mL/ min). The experiments
were performed at 400 °C during 2 h. Products evaporating
from the sample were collected using two cold-traps in series
(100 mL of toluene cooled with ice). The sample bed was
Soxhlet extracted with toluene, and the extract was combined
with the toluene of the cold-traps to measure the global
amount of PCDD/ Fs formed. Each experiment was performed
in duplicate or triplicate.
Cleanup. The slightly modified EPA-8280 was followed
for classical PCDD/ Fs analysis. The detailed method has been
previously described (15).
Analysis. All analyses were performed by HRGC/ HRMS
using a Mat95-XL high-resolution mass spectrometer and a
Hewlett-Packard 6890 Series gas chromatograph. The GC
conditions were optimized to separate most of the PCDD/ Fs
as follows: column, SP2331 capillary column (Supelco, 60 m
× 0.25 mm i.d., 0.2 µm film thickness); splitless injection of
2 µL of the extract at 275 °C; initial oven temperature, 150
°C; temperature programming, 150 °C, held for 1 min, then
increased at 15 °C/ min to 200 °C, then increased at 1.2 °C/
min to 273 °C, and held during 18 min. Helium was used as
the carrier gas. The mass spectrometer was operated in the
electron impact ionization mode using selected ion moni-
toring. The mass spectrometer was tuned to a minimum
resolution of 10 000 (10% valley) and was operated in a mass
drift correction mode using FC5311 to provide lock masses.
The two most abundant ions in the chlorine clusters of the
molecular ion were recorded for each congener of native
and labeled PCDD/ Fs. The source temperature was set at
270 °C.
The ratios of PCDFs and PCDDs have been calculated for
all of the experiments performed for each chlorofamily
(TCDFs/TCDDs,PeCDFs/PeCDDs,HxCDFs/HxCDDs,HpCDFs/
HpCDDs). These ratios do not present a particular tendency
for the experiments performed without the addition of E.S.P.
dust. These results are not presented. On the other hand,
Figure 2 presents the PCDFs/ PCDDs ratios calculated for
the different chlorofamilies in the case of the experiments
performed at 400 °C, 2 h in the presence of E.S.P. dust. The
graphite and the coke 9 h present the same trend: TCDFs/
Identification and Quantification. The TCDD-OCDD and
TCDF-HpCDF congeners were analyzed. No analyses of the
species without chlorine or less than 4 chlorines were
9
VOL. 38, NO. 15, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 2 2 3

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
FIGURE 1. Total amounts of PCDD/Fs measured from annealing the three types of carbon ((A) 400 °C, 2 h; (B) 400 °C, 2 h + 10 wt % E.S.P.
dust), mean value ( range.
FIGURE 2. PCDFs/PCDDs ratios for the different chlorofamilies as
a function of the carbon source (400 °C, 2 h + 10 wt % E.S.P. dust),
mean value ( range.
FIGURE 3. Homologue distributions of PCDDs (A) and PCDFs (B)
TCDDs around 28, PeCDFs/ PeCDDs around 17, HxCDFs/
HxCDDs around 10, and HpCDFs/ HpCDDs around 7. For
the coke 15 h, the proportion of PCDFs produced is much
higher, especially for the lower chlorinated families. TCDFs/
TCDDs rise from 28 in the graphite and coke 9 h, and reach
up to 54 in the coke 15 h. PeCDFs/ PeCDDs rise from 17 for
the graphite and coke 9 h, and up to more than 35 for the
coke 15 h. A possible difference in the concentration of
chemically bound oxygen or chemisorbed precursors to
PCDDs of the three samples may be responsible for these
different PCDF/ PCDD ratios.
The homologue distributions have been analyzed for all
of the experiments performed. These distributions are
dependent on the carbon source. An example of distribution
is presented in Figure 3. The tendencies are not always clear,
but it seems that coke 9 h leads to the formation of lower
chlorinated species, in opposition to graphite and coke 15
h. This tendency can be visualized in Figure 3, as well as in
Table 2, which summarizes the mean degree of chlorination.
The PCDD/ Fs generated from coke 9 h are always lower
chlorinated than the species formed from the two other
samples (coke 15 h and graphite). This could hardly be
explained by the fact that less chlorine is available or from
differences in the chlorination/ dechlorination catalysts
present because the same E.S.P. dust was added to the three
carbons. On the other hand, differences in interaction
between the catalyst and carbons are a more likely explana-
tion. These differences could be due to difference in H-atoms,
O-atoms, functional groups, etc., present in the carbon
structure (19, 20).
observed after thermal treatment of the samples (400 °C, 2 h, air 100
mL/min, 10% E.S.P. dust), mean value ( range.
TABLE 2. Mean Degree of Chlorination Observed in the
Different Samples after Thermal Treatment (400 °C, 2 h, Air
100 mL/min, 10% E.S.P. Dust)
mean degree of chlorination
nature of the sample
PCDDs
PCDFs
graphite
coke 9 h
coke 15 h
7.2
6.2
7.5
5.8
5.3
5.8
In conclusion, concerning the global PCDD/ Fs amounts
produced, the three kinds of samples do not present
important differences. Although only one graphite and two
cokes are tested in this study, it seems that the use of a
particular coke in the industrial plant cannot drastically
reduce PCDD/ Fs emissions. Indeed, the pure carbons
produce very low amounts of PCDD/ Fs, and fine variations
in the cokes structure will certainly not modify this fact. The
presence of catalysts is obviously a more crucial factor, leading
to the formation of large amounts of PCDD/ Fs, even from
graphite, a no-active matrix in terms of de novo synthesis.
Nevertheless, the PCDD/ Fs produced during the thermal
treatment present differences in the fingerprint, and thus a
deeper study of the results, combined with a fine analysis of
the composition and the structure of the matrixes, will
perhaps bring information concerning the mechanisms of
formation of the PCDD/ Fs by de novo synthesis.
The isomer distributions (not shown) are not influenced
by the nature of the carbon source for the experiments
performed. Actually, the differences found between the
different samples tested are small and in the same order of
magnitude as the variation between replicates.
Recycling of the E.S.P. Dust in the Raw Materials. The
recycling of the dust collected in the electrostatic precipitator
in the raw materials is usually used at industrial plants. A
9
4 2 2 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 15, 2004

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
FIGURE 5. Homologue distributions of PCDDs observed after thermal
treatment of coke 9 h (400 °C, 2 h, air 100 mL/min), in the presence
or not of 10% E.S.P. dust, mean value ( range.
measured in the experiments with the graphite and the cokes
do not derive exclusively from the E.S.P. dust, but also from
the degradation of the carbon structure.
FIGURE 4. Global amounts of PCDD/Fs produced as a function of
The recycling of the E.S.P. dust in the raw materials risks
increasing the PCDD/ Fs emissions of the industrial process.
This increase will be the result of the addition of E.S.P. dust
able to generate PCDD/ Fs by simple thermal treatment: the
passage of this E.S.P. dust in zones around 300-400 °C can
result in the formation of great amounts of PCDD/ Fs. On the
other hand, the recycling of this E.S.P. dust will also provide
catalysts and chlorine atoms, necessary for the formation of
PCDD/ Fs from other carbon sources such as coke. These
results should caution against recycling the E.S.P. dust in the
feed of the sintering process. However, the recycling of E.S.P.
dust is economically favorable and avoids an expensive and
polluting dumping. The ideal would be to combine the E.S.P.
dust recycling with the reduction of the PCDD/ Fs emissions
from the sintering process. The inhibition technique, studied
elsewhere (17, 22), seems to be adequate for this industrial
process and gives excellent results in industrial tests.
the thermal treatment, mean value ( range.
part of the experiments presented in the first section of this
paper has clearly demonstrated that the presence of the E.S.P.
dust greatly increased the amounts of PCDD/ Fs formed by
the degradation of the three carbons tested (graphite, coke
9 h, and coke 15 h). This fact implies caution against the
recycling of the E.S.P. dust in the raw materials at the
industrial scale and needs further consideration. The experi-
ments performed to study the ability of the three carbons to
form PCDD/ Fs can be looked at as a way to simulate the
recycling of the E.S.P. dust in the raw materials. Is the
proportion of 10 wt % of E.S.P. dust realistic to study the
recycling? Reference 21 gives the following data: 17-313 g
of E.S.P. dust/ kg of breeze coke are generally used corre-
sponding to 1.7-31.3 wt % E.S.P dust. Taking the dilution by
the other raw materials into account, the % of E.S.P. dust in
the global raw materials is of course less, but the total dusts
and recycled materials are also more important: 11-154
kg/ t sinter (sum of all dusts and recycled materials), thus
between 1.1 and 15.4 wt %. The mixture of 90 wt % carbon
and 10 wt % E.S.P. dust is thus arbitrary but realistic.
The results previously presented in Figure 1, combined
with the initial PCDD/ Fs content of the carbons, are again
presented in Figure 4 but in a different fashion (note the
logarithmic scale). The PCDD/ Fs content of the three
samples, as well as the PCDD/ Fs produced with the pure
samples during the thermal treatment, are very low in
comparison with the huge amounts of PCDD/ Fs produced
when E.S.P. dust is added to the carbon sources. The amounts
are multiplied by a factor greater than 103 because of a simple
addition of extracted dust collected in the electrostatic
precipitator of a sinter plant. These results confirm the great
dependence of the de novo synthesis on the catalysts.
The addition of E.S.P. dust also modifies the homologue
distributions. An example is presented in Figure 5. The
presence of E.S.P. dust moves the distribution toward the
formation of more chlorinated species. In Figure 5, the
homologue distribution observed after the thermal treatment
of pure coke 9 h is dominated by TCDDs (53%) and PeCDDs
(22%). In the presence of E.S.P. dust, the homologue profile
is dominated by HxCDDs (30%) and HpCDDs (29%). The
mean degree of chlorination of the PCDDs formed from the
pure coke 9 h is 4.8, but this value rises to 6.2 with the addition
of E.S.P. dust to the sample. The presence of chlorine in the
E.S.P. dust as well as catalysts of chlorination certainly
explains this tendency.
Acknowledgments
We would like to thank Mr. J.-M. Brouhon for interesting
discussions, as well as C. Dohmen for performing some of
the experiments. C.X. was funded as a fellow by the F.N.R.S
(Fonds National de la Recherche Scientifique Belge).
Literature Cited
(1) Olie, K.; Vermeulen, P. L.; Hutzinger, O. Chemosphere 1977, 8,
455-459.
(2) Buekens, A.; Huang, H. J. Hazard. Mater. 1998, 62, 1-33.
(3) Lahl, U. Chemosphere 1994, 29, 1939-1945.
(4) Quass, U.; Fermann, M. European Dioxin Emission Inventory-
Stage I, Report prepared by the North Rhine Westphalia State
Environment Agency on behalf of the European Commission
1997.
(5) Quass, U.; Fermann, M.; Bro¨ ker, G. European Dioxin Emission
Inventory-Stage II, Report prepared by the North Rhine West-
phalia State Environment Agency on behalf of the European
Commission 2000.
(6) Communication from the Commission to the Council, the
European Parliament and the economic and social Committee,
Community strategy for dioxins, furans and polychlorinated
biphenyls, COM(2001) 593 final.
(7) Stieglitz, L.; Zwick, G.; Beck, J.; Roth, W.; Vogg, H. Chemosphere
1989, 18, 1219-1226.
(8) Milligan, M. S.; Altwicker, E. Environ. Sci. Technol. 1993, 27,
1595-1601.
(9) Huang, H.; Buekens, A. Chemosphere 1995, 31, 4099-4117.
(10) Huang, H.; Buekens, A. Sci. Total Environ. 1996, 193, 121-
141.
(11) Schwarz, G.; Stieglitz, L. Chemosphere 1992, 25, 277-282.
(12) Schwarz, G. Thesis, Kernforschungszentrum Karlruhe.
(13) Stieglitz, L.; Zwick, G.; Beck, J.; Schleihauf, J.; Will, R. Chemosphere
1993, 27, 343-351.
As explained previously in this paper, although the E.S.P.
dust used in these experiments presents a great capacity to
form PCDD/ Fs by simple thermal treatment, the PCDD/ Fs
9
VOL. 38, NO. 15, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 2 2 5