Formation of dioxins in the catalytic combustion of chlorobenzene and a micropollutant-like mixture on Pt/γ-Al2O3

Article abstract of DOI:10.1021/es034820y

Catalytic combustion over a 2 wt % Pt/γ-Al₂O₃ catalyst of chlorobenzene (PhCl) and of a micropollutant-like mixture representative for a primary combustion offgas has been investigated. Typical conditions were 1000-1500 ppm of organics in the inflow, contact times ～0.3 s, 16% O₂ in nitrogen at ～1 bar, and temperature range 200-550 °C. PhCl reacts considerably slower than when processing Cl-free compounds such as heptane. At intermediate temperatures-and incomplete conversion-byproducts are formed, especially polychlorobenzenes (PhCl _x). These are accompanied by polychlorinated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) at levels of about 10^-6 relative to PhCl_x. Additional HCl-made by co-reacting PhCl with tert-butylchloride-leads to much higher levels of PhCl_x and PCDD/Fs. Using the micropollutant-like mixture, the total chlorine input is reduced almost 20-fold, but it nevertheless leads to a 30-fold higher PCDD/F output. This is ascribed to reaction of the small amounts of (chloro)phenols in the mixture. The congener/isomer patterns of the PCDD/Fs for the mixture and with PhCl per se are quite comparable with those found in emissions from incinerators. As carbon is not present nor formed on the catalyst surface, de-novo formation therefrom cannot be involved. Rather condensation of phenolic entities or like precursors must have occurred. Consequences and options to ensure safe application are briefly discussed as well.

Full text of DOI:10.1021/es034820y

Environ. Sci. Technol. 2004, 38, 5217-5223
2 3 3
a 12 wt % Cr O / 2 wt % CrO catalyst was tested for the
Formation of Dioxins in the Catalytic
Combustion of Chlorobenzene and a
Micropollutant-like Mixture on
destruction of a chlorine-containing mixture, formation of
volatile chromium oxychlorides was indeed observed (13).
For more recent experience, see the work of Corella and
co-workers (14, 15). It appears that supported noble metals,
especially Pt, are much better suited for the catalytic
combustion of halogen-containing (waste) gases than oxidic
catalysts, certainly when halogen levels are substantial.
Another feature is the possibility that Deacon-type
Pt/γ-Al O3
2
V I N C E N T D E J O N G , †
M A R I U S Z K . C I E P L I K , A N D
‡
reactions occur, which can create substantial amounts of Cl
and hence lead to chlorinated byproducts, such as poly-
chlorobenzenes (PhCl ) from chlorobenzene (PhCl) (16) and
CCl from CH Cl (17-19). These byproducts might not be
2
R O B E R T L O U W *
x
Center for Chemistry and the Environment, Leiden Institute of
Chemistry, Gorlaeus Laboratories,
Leiden University, P.O. Box 9502,
4
2
2
a problem during proper operation with complete miner-
alization of organic chlorine and full conversion of all organics
into CO , but if the feed for instance temporarily contains
2
NL-2300 RA Leiden, The Netherlands
too little organics for the catalyst bed to maintain the desired
working temperature, this can result in a drop in temperature
with the formation of unwanted byproducts. Of course -
gradual - deactivation of the catalyst is a threat as well.
Earlier studies on the combustion of PhCl over a 2 wt %
Pt/ γ-Al O catalyst showed the formation of PhCl with levels
2 3 x
as high as 12% of the total chlorine input at temperatures
around 400 °C at conversions of PhCl of around 90% (16, 20).
Catalytic combustion over a 2 wt % Pt/γ-Al2O3 catalyst of
chlorobenzene (PhCl) and of a micropollutant-like mixture
representative for a primary combustion offgas has been
investigated. Typical conditions were 1000-1500 ppm
of organics in the inflow, contact times ∼0.3 s, 16% O2 in
nitrogen at ∼1 bar, and temperature range 200-550 °C.
PhCl reacts considerably slower than when processing Cl-
free compounds such as heptane. At intermediate
For Pd on several supports, the production of PhCl
to be even higher (21). Varying the support, lowering the Pt
dispersion, and adding H O during the combustion of PhCl
helped to suppress, but could not eliminate, the formation
of PhCl (21). A better “therapy” is to add a fair excess of, for
example, heptane to PhCl; rates of combustion of PhCl are
x
was found
2
temperaturessand incomplete conversionsbyproducts
are formed, especially polychlorobenzenes (PhClx). These
are accompanied by polychlorinated dibenzo-p-dioxins
x
-
6
much larger, and the amount of PhCl is down to trace levels
(
PCDDs) and dibenzofurans (PCDFs) at levels of about 10
relative to PhClx. Additional HClsmade by co-reacting
PhCl with tert-butylchloridesleads to much higher levels
x
only (1). A key feature appears to be that ‘poisoning’ of the
catalyst surface by chlorine is counteracted by its removals
due to water or due to the hydrogen of the alkane (1, 22, 23).
of PhClx and PCDD/Fs. Using the micropollutant-like mixture,
the total chlorine input is reduced almost 20-fold, but it
nevertheless leads to a 30-fold higher PCDD/F output. This
is ascribed to reaction of the small amounts of (chloro)-
phenols in the mixture. The congener/isomer patterns of the
PCDD/Fs for the mixture and with PhCl per se are quite
comparable with those found in emissions from incinerators.
As carbon is not present nor formed on the catalyst
surface, de-novo formation therefrom cannot be involved.
Rather condensation of phenolic entities or like precursors
must have occurred. Consequences and options to ensure
safe application are briefly discussed as well.
x
When PhCl are formed it is possible, if not likely, that
also higher-chlorinated polyaromaticsssuch as polychlori-
nated dibenzo-p-dioxins (PCDDs) and dibenzofurans
PCDFs)sare produced. These “dioxins”, infamous for their
toxicity, are highly undesired; therefore, we thought it of
prime importance to check the performance of a Pt/ γ-
2 3
alumina (γ-Al O ) catalyst also in this context. In practice
the aim is of course to achieve “complete” destruction of the
target compound(s)sand of possible intermediate or side
productssunder acceptable conditions, for instance at
temperatures which the catalyst can tolerate. However, to
get an insight into the underlying chemistrysand in the
consequences of a malfunctioning catalystsreactions must
be conducted (also) at incomplete conversions, which we
have done.
(
Introduction
Catalytic combustion is an attractive technology for the
complete, clean conversion of hydrocarbon fuels into energy
and for the purification of waste gases with low contents of
organics. Supported noble metals (especially Pt) are well-
suited (1-5), but a number of oxides, mixed oxides, and
perovskites are also used (6-11).
First, PhCl has been studied; as such, and next with tert-
butyl chloride (t-BuCl, 1 mol/ mol PhCl) as an additional
chlorine source. The reason to add t-BuClsfound to be very
rapidly combusted even at the lowest temperatures studied
-
is to obtain a level of HCl common for off-gases of thermal
waste) combustion facilities, and see its effect on the behavior
(
When halogensespecially chlorinesis present in the feed,
several problems can arise, to wit: corrosion due to the
formed HCl, and in some cases formation of volatile and
reactive (oxy)chlorides. For example, chromium oxychloride
of model compound PhCl. Next, the behavior of a complex,
micropollutant-like mixture has been investigated. This is
meant to mimic catalytic combustion of PICs, products of
incomplete combustion, in said off-gases. The composition
of this mixture is based on the analysis of a common
municipal solid waste (MSW) facility off-gas (24). By using
a micro-dosage pump, the PIC concentrations in the feed
(
2 2
CrO Cl ) has a boiling point of only 117 °C (12) and when
*
Corresponding author phone: (+31)71-527-4289; fax: (+31)-
7
1-527-4451; e-mail: r.louw@chem.leidenuniv.nl.
2
gas to the catalyst bed (16% O in nitrogen) could be very
†
Present address: ALcontrol Laboratories, Steenhouwerstraat 15,
low, at even (sub) ppm levels, to approach real conditions
also in this respect. Sampling on polyurethane foam (PUF)
plugs for typically 18 h allowed proper PCDD/ F analysis.
NL-3194 AG Hoogvliet, The Netherlands.
‡
Present address: Energy Research Centre of the Netherlands,
Wetserduinweg 3, NL 1755 LE Petten, The Netherlands.
1
0.1021/es034820y CCC: $27.50

2004 Am erican Chem ical Society
VOL. 38, NO. 19, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5 2 1 7
Published on Web 08/24/2004

FIGURE 1. Schematic overview of the on-line setup.
concentrated to 0.5 mL - first with a rotary evaporator and
next with a Zymark Turbovap II apparatus, and sent for
external HRGC/ HRMS analysis.
Experimental Section
On-Line Measurem ents. The experiments were performed
using an on-line setup with a sampling unitsa polyurethane
foam (PUF) plug adsorbersand the feeding of a liquid input
mixture (PhCl and/ or other organics), as depicted in Figure
Catalyst. A 2 wt % Pt/ γ-Al
2 3
O catalyst was prepared
following the homogeneous deposition precipitation method
as described by Geus (25). For 10 g material, 9.80 g γ-Al
2
O
3
1
. The latter was done by means of a microinjection pump
(
Degussa C) was suspended in 250 mL water; 0.308 g H
2
-
(
CML, Microdialysis 100) connected by a thin capillary (0.28
[Pt(OH)
6
] (Johnson Matthey, 62.5% Pt) was dissolved in 3 mL
mm) to a glass mixing chamber of 1 mL capacity, and filled
with quartz wool. The injected liquid (typical inflow of 0.225
hot concentrated nitric acid and added to the suspension
under continuous stirring. Subsequently a 10-g portion of
urea was added while increasing the temperature to ca. 80
-
1
µL min ) was evaporated in this heated (T ∼300 °C) glass
chamber, and transported to the reactor by the carrier gas,
metered by two electronic mass flow controllers (Brooks 5850
°
C and stirring for 15 h. Due to slow decomposition of urea
the pH increased from 2 to 8, causing the Pt precursor to
precipitate on the γ-Al support. The suspension was
2 2
TR), one for N and one for O . The reactor was a vertical
2
O
3
quartz tube (internal diameter 5 mm) with a quartz filter on
which the catalyst bed rested. For each experiment 0.3 g of
catalyst was used, with a bed height of ca. 10 mm.
filtered off, washed with demineralized water, and dried
overnight at 90 °C. Finally the catalyst was calcined for 3 h
in flowing air at 600 °C. After crushing a sieved fraction of
The outflowing gases were directed by heated stainless
steel tubing (T > 150 °C) to a sampling loop by means of two
six-port sampling valves (Valco C6WT-HC) for on-line analysis
on a Hewlett-Packard 5890 series II gas chromatograph,
utilizing a capillary (CP SIL 5 CB, 50 m, × 0.32 mm, film
thickness 0.4 µm) and a packed column (Carbosphere 80-
1
50-300 µm was collected and used.
Chem icals. The following chemicals were available and
used as such: benzene (Merck Uvasol, >99%), acetone (Baker,
99.5%), naphthalene (Janssen Chimica, >99%) chloroben-
>
zene (Baker Analyzed, >99%), phenol (Aldrich, >99%),
phenanthrene (synthesized, >99%), 1,1,1-trichloroethane
1
00 mesh). The capillary column was at the end split into
(
>
Janssen Chimica, >97%), benzaldehyde (Baker Analyzed,
99%), dibenzofuran (Fluka, >95%), monobenzofuran (syn-
two streams, which were connected to a flame ionization
detector (FID) and an electron capture detector (ECD). The
FID was used for on-line quantification of hydrocarbons etc.
and the ECD was used for (poly)chlorinated organics (up to
thesized, >99%), toluene (Baker Analyzed, >99%), 2-chlo-
rophenol (Fluka, >99%), 2,4,6-trichlorophenol (Fluka, >99%),
tetrachloroethylene (Baker Analyzed, >99%), hydrogen (Air
Products, 99.995%), methane (Air Products, 99.995%), ni-
trogen (Air Products 99.995%). The composition of the model
mixture of products of incomplete combustion (PIC) is given
in Table 1
PhCl
6
). CO and CO
2
were quantified by injection on a packed
in a methanizer;
column followed by hydrogenation to CH
4
the resulting mixture was then analyzed on a FID.
At all relevant temperatures at least a duplicate on-line
analysis was performed. When the reactor reached the set
temperature the first injection on the GC was after 15 min
in order to guarantee a stable situation.
Results
Off-line PCDD/ Fs measurements were performed after
Soxhlet extraction of said PUF plugs with 110 mL of a 10/ 1
heptane/ acetone mixture for approximately 18 h, with 20
Byproduct Form ation during Com bustion of PhCl on Pt/
γ-Al
during the combustion of PhCl on 2 wt % Pt/ γ-Al
of experiments was performed between 150 and 550 °C. Figure
2
O
3
. To get an insight in the formation of byproducts
2
O
3
, a series
-
1
turnovers h . The resulting extract was subsequently
5
2 1 8
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 19, 2004

2
, entry 1. The total of ∼0.13 nmol compared with the output
TABLE 1. Composition of the PIC-like Feed Mixturea
of PhCl
x
at 350 °C (∼2.5% or 0.11 mmol, cf. Figure 2) comprises
-6
ratiob
concnc [mg m-3]
a molar ratio of 1.2 × 10 . The (overall) PCDF/ PCDD ratio
is close to 2; interestingly, this is not unlike the pattern
commonly found for dioxins from (municipal waste) incin-
eration (Table 2, entry 2). Also, the degrees of (poly)
chlorination do not differ much.
Com bustion of a t-BuCl/PhCl Mixture. The additional
chlorine source was applied to see if it has an impact on the
byproduct formation. An input mixture of 750 ppm PhCl
and 750 ppm t-BuCl was used, all other flows were kept the
same. Exploratory experiments have shown that t-BuCl gave
a quantitative yield of HCl and CO
The conversion data and the PhCl
temperatures are plotted in Figure 3. The lowest applied
temperature for this series was 350 °C, focusing on the area
with high PhCl
The output of PhCl
per se (Figure 2). Now the highest selectivity to PhCl
observed at 419 °C and found to be 41%. For PhCl per se, the
highest selectivity to PhCl (ca. 7% see Figure 1) was reached
compound
benzene
(-)
7640
4179
1743
568
652
519
329
116
90
92
50
39
44
naphthalene
phenanthrene
acetone
1/3
1/10
1/10
1/20
1/20
1
,1,1-trichloroethane
benzaldehyde
dibenzofuran
m onobenzofuran
toluene
1/50
1/100
1/100
1/100
1/250
1/500
1/250
1/100
phenol
x
at even mild conditions.
production at various
2
2
-chlorophenol
,4,6-trichlorophenol
x
chlorobenzene
tetrachloroethylene
162
x
outputs as in Figure 2.
is clearly higher than that from PhCl
is
a
Inflows: liquid m ixture 0.225 µL m in-1, gases N
2
+ O
, total Cl for 18 h ) 2.4 µm ol. b Relative ratio to benzene
, with 16% O
2
2
) 0.27 m ol
-
1
x
h
(
m ol/m ol). c At room tem perature and 1 bar, with the m ixing ratio (see
footnote a).
x
x
at 397 °C. So, the conversion of PhCl may be a little slower
in the presence of t-BuCl, but the formation of (oxy)-
chlorinated byproducts is greatly enhancedsand about
equally important as oxidative degradation.
2
depicts conversions of PhCl and total PhCl
x
productions
at various temperatures.
Consonant with the earlier observations of Van den Brink
et al. (16), the catalytic oxidation of PhCl is indeed a sluggish
and complex process especially at intermediate conversions.
Increasing the reaction temperature results in higher com-
bustion rates for PhCl, but full conversion is reached only at
Next the influence of t-BuCl on the formation of PCDD/
Fs was investigated. A temperature of 360 °C was used while
the byproducts were again collected for an 18 h period. Data
are in Table 2, entry 3. With a 25% lower PhCl level in the
feed, the total dioxin output is increased from ∼0.13 to ∼0.33
nmol, while the PCDF/ PCDD ratio is up from 1.6 to ∼3.0.
The average degree of (poly)chlorination is quite the same
for the PCDFs but markedly larger for the PCDDssfrom ∼4.9
in entry 1 to ∼6.6 in the present case (entry 3). Altogether,
the selectivity to PCDD/ F remains relatively low, with an
temperatures of 500 °C and above. The selectivity to PhCl
x
strongly depends on the temperature. Already at ca. 250 °C
higher-chlorinated products from PhCl could be detected.
A highest selectivity of 7% (yield on converted PhCl, sum of
(
poly)chlorinated benzenes) is observed at around 400 °C,
which is comparable to previous results with this catalyst
16, 20). The isomer ratio of p-, m-, and o-PhCl was
remarkably constant at 4:3:1 at all temperatures.
(
2
x
output of PhCl at 360 °C of 1.3 mmol, the 0.33 nmol dioxins
-
6
To determine the PCDD/ Fs formed and emitted, an
experiment was performed at a set temperature of 350 °C.
Over a period of 18 h heavier organic byproducts were
collected on a PUF-plug. Results on (tetra- and higher
chlorinated) PCDD/ F isomer groups are collected in Table
are a factor 0.3 × 10 less.
Micropollutant-like Mixture. To mimic the out-flowing
gas of a MSW incinerator, a mixture has been composed
based on the data of Jay and Stieglitz (24), see Table 1. For
practical reasons, a selection of substances has been made
FIGURE 2. PhCl
PhCl ([), PhCl
x
production during combustion of PhCl on 2 wt % Pt/γ-Al
2
O
3
x 2
. PhCl conversion (+), total output of PhCl (9), PhCl (b),
-
1
3
4
(2), PhCl5-6 (×). Inflows: PhCl 1000 ppm, O + N 0.25 mol h ; τvoid ∼ 0.33 s at STP; [O
2
2
2
] ) 16%.
VOL. 38, NO. 19, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5 2 1 9

a
2 3
TABLE 2. PCDD/F (tetra through octa) Production during Combustion of Various Organics on 2 wt % Pt/γ-Al O
b
c
d
e
PhCl (1)
flue gas (2)
PhCl/t-BuCl (3)
PIC mixture (4)
PCDD/F
ng]
PCDD/F
[10 mol]
PCDD/F
[ng]
PCDD/F
[10 mol]
PCDD/F
[ng]
PCDD/F
[10 mol]
-
9
-9
-9
[
PCDD/F
T4CDF
P5CDF
H6CDF
H7CDF
O8CDF
T4CDD
P5CDD
H6CDD
H7CDD
O8CDD
sum PCDF
sum PCDD
2.8 (6)
0.009
0.016
0.025
0.014
0.016
0.025
0.012
0.005
0.003
0.003
0.079
0.049
(25)
(17)
(12)
(8)
(4)
(4)
(6)
(9)
(9)
(9)
10.2 (8)
20.0 (16)
20.8 (16)
20.6 (16)
22.5 (18)
1.5 (1)
4.6 (4)
6.9 (5)
9.1 (7)
0.033
0.058
0.055
0.050
0.051
0.005
0.013
0.018
0.021
0.025
0.247
0.081
324 (27)
261 (22)
138 (11)
84 (7)
1.1
5.3 (11)
9.4 (20)
5.7 (12)
7.2 (15)
8.9 (19)
4.3 (9)
2.0 (4)
1.3 (3)
1.4 (3)
29 (64)
18 (36)
0.76
0.37
0.20
0.05
0.37
0.38
0.22
0.09
0.03
2.4
24 (2)
115 (10)
129 (11)
84 (7)
36 (3)
15 (1)
11.4 (9)
94 (74)
34 (26)
(64)
(36)
831 (70)
379 (30)
1.1
a
b
Total PCDD/F production in 18 h; conditions, see Figure 2; values in parentheses are wt % of total PCDD/Fs. PhCl (1000 ppm ) ) 4.5 m m ol
in 18 h, at 350 °C. c Data from ref 26. d PhCl (750 ppm )/t-BuCl (750 ppm ) ) 6.75 m m ol of Cl in 18 h, at 360 °C. PIC m ixture ) 2 µm ol of Cl in 18
h, at 350 °C.
e
FIGURE 3. Combustion of PhCl with t-BuCl on 2 wt % Pt/γ-Al
PhCl5-6 (×). Inflows: PhCl 750 ppm, t-BuCl 750 ppm, O + N
2
O
3
: PhCl conversion (2), total PhCl
x
(9), PhCl
2
(b), PhCl
3
([), PhCl
4
(2),
-
1
2
2
0.25 mol h ; τvoid ∼0.1 s at 250 °C.
since over 200 different compounds were identified. Those
chosen represent all important categories of PICs with the
exception of S-containing derivatives. By applying a low
inflow rate, concentrations have been obtained that in total
are ∼1500 ppm C again. However, these are about 2 orders
of magnitude larger than those of the PICs in real combustion
off-gases. Due to the complex mixture and the low level of
most of the individual compounds, we have refrained from
analyzing the conversions of the individual compounds and/
or identifying any intermediate (Cl-free) products. The
combustion could however be followed through the total
around 60%, based on CO
ics were collected on a PUF plug kept at -25 °C. Data are
collected in Table 2, entry 4.
x
production. (Semi)volatile organ-
Compared with the results from PhCl per se (entry 1),
levels of PCDD/ Fs have greatly increased from 0.13 to 3.5
nmol, notwithstanding that the concentrations of key
aromatics in the feed are much lower than in the model
experiments: benzene ∼220 ppm, and phenol only 2 ppm
-6
(∼10 M). Note also that the PCDF:PCDD ratio again is close
to 2 (Table 1, entry 4).
Dioxin and Furan Patterns. Figure 4 depicts relative
amounts of the tetra- to octa-isomers together with data
representative for incinerator flue gases. It underlines that
the PCDD/ F ratios (here as “% PCDDs”) are strikingly similar
despite the large variation in type and/ or concentration of
the organics employed. The variations in chlorination
patterns, degrees, are limited too despite the large differences
in chlorine (input) levels.
CO
x x
production. CO was already detected at 200 °C and
reached 90% at 400 °C.
By on-line ECD analysis some (poly)chlorinated products
could be seen but also due to the low concentrations it was
not possible to identify any of these. On the basis of retention
times, it appears that both chlorinated aliphatics/ olefins and
x
traces of PhCl were present in the effluent gas.
To quantify PCDD/ Fs formed with this realistic inflow
mixture, an 18 h run was performed with the reactor
temperature set at 350 °C. This implies a total conversion of
Further insight into conformity or difference in patterns
can be obtained by considering the distributions within
homologuessespecially those with a sizable toxic equivalence
5
2 2 0
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 19, 2004

FIGURE 4. Homologue distributions, data for flue gas samples from ref 26.
FIGURE 5. Relative abundances of 2,3,7,8-substituted PCDD/Fs; data for MSW incinerator from ref 28.
factor (TEF) (27). Data are shown in Figure 5. While part of
the differences may be caused by the different levels of
chlorination, it is worth noting that some isomers are at a
rather constant level/ ratio, such as the (relative toxic)
the latter species result from chemisorption of PhCl by C-H
bond fission and are subject to oxidation as the major
pathway (16, 20). Chlorine at the surface also acts as a poison
for the overall conversion, which can be relieved by co-
reacting suitable hydrocarbons (1, 22).
2
1
,3,4,7,8-P
,2,3,7,8-P
5
CDF, whereas especially 2,3,7,8-TCDD and
CDD show important variations (see Figure 5)
5
Dioxins (PCDD/ Fs) are also clearly formed; under condi-
tions of partial conversion of the ∼1000 ppm of PhCl at levels
much higher than those in flue gases of classical MSW
incinerators.This is further discussed below.
and are rather prominent in the reaction with PhCl per se.
Discussion
Catalytic Com bustion of Chlorobenzene. The model reac-
tions with PhCl underscore that Pt is suited for the combus-
tion of organochlorine compounds, but with care. The
oxidative degradation is relatively slow, and substantial levels
Added tert-butylchloride is very rapidly converted to give
∼100% HCl; discrete reaction to (sorbed, or even free)
isobutene, by catalyzed HCl elimination is likely. Catalytic
combustion of isobutene is expectedly very rapid (22).
Anyway, organic intermediate products could not be de-
of higher chlorinated benzenes PhCl
process. The isomer distribution of the dichlorobenzenes
with some 40% meta) precludes a common electrophilic
x
are formed in a parallel
x
tected. It firmly promotes the formation of PhCl , however,
(
meaning that its hydrogen is of no value for removal of Cl
from the surface and, hence, suppression of chlorination. It
appears that a hydrocarbon co-fuel must have a reactivity
comparable to that of PhCl to be effective (1). Altogether,
substitution mechanism. The chemistry is best explained by
combination of surface-bound Cl and an (o-, m-, or p-)-
chlorophenyl entity (in a reductive elimination reaction);
VOL. 38, NO. 19, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5 2 2 1

t-BuCl has acted as a “perfect” additional source for HCl.
It led to nearly thrice as much PCDD/ F compared with
PhCl per se under like conditions, notwithstanding a ∼25%
lower PhCl input level. Apparently, the population on the
catalyst surface alters in a way that there is more Cl (lowering
the overall rate of conversion of PhCl) that accelerates
chlorination but also leads to more PCDD/ F (see below).
Micropollutant Mixture. Refraining from a discussion of
the fate of individual constituents, we may note that the
cocktail contained little aliphatics which could have “helped”
in keeping the catalyst clean and highly active. Given that,
let us focus at the PCDD/ Fs observed. With an input level
of the major constituent benzene ∼5 times lower than that
of PhCl, the dioxin output nevertheless increased ∼25-fold.
This is notwithstanding the much lower chlorine level: in a
standard PhCl experiment the total input in a 18 h run was
TABLE 3. Toxicity Levels
TEQ [ng Nm-3
]
PhCl
PhCl + t-BuCl
PIC m ixture
93a
31
571
a
For PhCl, the m ain com ponents that contribute to the total TEQ are
CDD (1)*71.5 ) 71.5 ng
(
with I-TEF factors in parentheses): 2,3,7,8-T
Nm ; 1,2,3,7,8-P
0.1)*52 ) 5.2 ng Nm ; other com ponents give a neglible contribution.
4
-
3
-3
5
CDD (0.5)*31.6 ) 15.8 ng Nm ; 1,2,3,4,7,8-H CDF
6
-3
(
meant for levels in food, etc., it is worthwhile to express the
outputs of PCDD/ Fsat the conditions mentioned in Table
3
1
sas ng I-TEQ/ Nm (Table 3). These can be compared with
values, common for “raw” MSW incinerator combustion off-
-
3
gases, viz., ∼0.1-0.75 µg m (36), for “full” toxic 17 PCDD/
∼
4.5 mequiv of Cl and in the runs with the cocktail only
Fs; when expressed in TEQ, these numbers will be about a
about 0.24 mequiv. Not surprising, the chlorination pattern
is somewhat biased to the lower chlorinated congeners, but
the average degree of chlorination of the PCDFs is still close
to 5.0.
-
3
factor 50 less or 2-15 ng Nm , which is still much higher
than an aimed or regulated emission level of <0.1 ng TEQ
-
3
Nm . Our result with the PIC mixturesalbeit, fed at much
higher concentrations than present in real combustion off-
gasessalso shows that a malfunctioning catalytic bed may
lead to much higher levels of dioxins than originally present.
Next to the option of co-feeding “soft” hydrocarbons (could
for example be naphtha), proper process control is required
to guarantee operation according to specification.
Pathways to PCDD/F. Without neglecting some sub-
stantial differences in the congener/ isomer compositions,
the outputs from PhCl and the PIC cocktail are quite
comparable, also with the PCDD/ F profiles common for MSW
facilities, where catalysis by fly ash is important. In the latter
case, catalysis cannot be due to Ptsor any other noble metals
but is based on copper (and maybe iron also) (29). Never-
theless, the similarities suggest common underlying mech-
anisms. In contrast with fly ash, our catalyst does not contain
nor forms any carbon, not so strange under the very “lean”
conditions employed. Therefore, de novo formation (30)s
common term for making dioxins from macromolecular
structures such as carbon and sootsdoes not have to be
taken into account. Instead, “precursor” pathways must be
involved. Consider the result with the PIC cocktail (Table 1,
entry 4). The output of 3.5 nmol means nearly 0.02 mol %
based on the total input of dibenzofuran, of ∼20 µmol in 18
h. Repeated chlorination of dibenzofuran may have con-
tributed to at least the PCDF production. More likely
candidates are the “trace” amounts of phenols in the cocktail
Acknowledgments
Financial support from the European Union within the
framework of the ‘MINIDIP’ project (Contract ENV4-CT97-
0
587). Our thanks are due to Dr. Kees Olie of the University
of Amsterdam as well as to Dr. Per Liljelind and Dr. Stellan
Marklund, Chemistry Department, Environmental Chem-
istry, Umeå University for performing the dioxin analyses.
Literature Cited
(
(
(
1) Van den Brink, R. W.; Mulder, P.; Louw, R. Catal. Today 1999,
4, 101-106.
2) Burch, R.; Crittle, D. J.; Hayes, M. J. Catal. Today 1999, 47, 229-
34.
5
2
3) Au-Yeung, J.; Chen, K.; Bell, A. T.; Iglesia, I. J. Catal. 1999, 188,
(
31-33) expressed on the intake of phenol itself the PCDD/ F
132-139.
outputsunder the used conditionssis 0.07%. In the reactions
with PhCl per se, chlorinated phenoxy entitiessnot neces-
sarily free chlorinated phenolssare likely intermediates. That
PCDD/ F patterns (from precursors) are a function of catalyst
and conditions, and pertinent mechanistic aspects will be
discussed separately (34). Which compounds in a realistic
cocktail are really involved in dioxin formation is worth further
study (35)snot only for oxidation on well-defined catalysts
but also in fly ash mediated processes in combustors.
Noteworthy is that the PIC cocktail, in reaction over MSW
incinerator fly ash, also leads to a “common” PCDD/ F product
with a DF:DD ratio of ∼4 (33, 35, 36).
(4) Burch, R.; Loader, P. K. Appl. Catal. B 1994, 5, 149-164.
(
(
5) Spivey, J. J. Ind. Eng. Chem. Res. 1987, 26, 2165-2180.
6) Lago, R. M.; Green, M. L. H.; Tsang, S. C.; Odlyha, M. Appl.
Catal. B 1996, 8, 107-121.
(
(
7) Imamura, S. Catal. Today 1992, 11, 547-567.
8) Stenger, H. G.; Buzan, G. E.; Berty, J. M. Appl. Catal. B 1993, 2,
117-130.
(9) Krishnamoorthy, S.; Baker, J. P.; Amiridis, M. D. Catal. Today
998, 40, 39-46.
1
(
(
10) Van den Brink, R. W.; Mulder, P.; Louw, R.; Sinquin, G.; Petit,
C.; Hindermann, J.-P. J. Catal. 1998, 180, 153-160.
11) Jang, B. H.; Lee, S. S.; Yeon, T. H.; Yie, J. E. Kor. J. Chem. Eng.
1998, 15, 516-526.
(12) NIST Standard Thermodynamic Database; available at http:/ /
webbook.NIST.gov.
Perform ance and Safety. As mentioned in the Introduc-
tion, in practice, a successful catalytic combustion should
be complete and durable, the more so if chlorine is involved.
When processing “recalcitrant” chlorinated aromatics, co-
feeding of a proper alkane (such as heptane with PhCl) is
highly beneficial, as it increases rates for chlorobenzene(s)
and suppresses (oxy)chlorinationsand almost certainly also
dioxin formationsas side reactions. A safe margin in working
temperature is of course also advisable. For a safe application
of (in this case, supported Pt) in the destruction ofseven,
dioxin-containingswaste gases, this approach should be
further tested, also including the effect of increased HCl levels,
and checked for compliance with regulation.
(
(
13) Agarwal, S. K.; Spivey, J. J.; Butt, J. B. Appl. Catal. A 1992, 82,
59-275.
14) Padilla, A. M.; Corella, J.; Toledo, J. M. Appl. Catal. B 1999, 22,
2
1
07-121.
(15) Corella, J.; Toledo, J. M. Ind. Eng. Chem. Res. 2002, 41, 1171-
1181.
(
(
16) Van den Brink, R. W.; Louw, R.; Mulder, P. Appl. Catal. B 1998,
6, 219-226.
1
17) Van den Brink, R. W. Ph.D. Thesis, Leiden University, Leiden,
The Netherlands, 1999 (in English).
(18) Green, M. L. H.; Lago, R. M.; Tsang, S. C. J. Chem. Soc. Chem.
Commun. 1995, 365-366.
(
19) Schneider, R.; Kiessling, D.; Herzschuh, R.; Wendt, G. React.
Kinet. Catal. Lett. 1997, 61, 245-251.
20) De Jong, V.; Cieplik, M. K.; Reints, W. A.; Fernandez-Reino, F.;
Louw, R. J. Catal. 2002, 211, 355-365.
(
Reactions conducted at incomplete conversion also serve
as a model for malfunctioning by deactivation of the catalyst
or by temperatures lower than required. Although primarily
(21) Van den Brink, R. W.; Krzan, M.; Feijen-Jeurissen, M. M. R.;
Louw, R.; Mulder, P. Appl. Catal. B 1999, 24, 255-264.
5
2 2 2
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 19, 2004

(
(
22) Van den Brink, R. W.; Louw, R.; Mulder, P. Appl. Catal. B 2000,
5, 229-237.
23) Van den Brink, R. W.; De Jong, V.; Louw, R.; Maggi, P.; Mulder,
P. Catal. Lett. 2001, 71, 15-20.
24) Jay, K.; Stieglitz, L. Chemosphere 1995, 30, 1249-1260.
25) Geus, J. W. Dutch Patent 6,705,259; 1967.
(31) Tuppurainen, K.; Ruokoj a¨ rvi, P.; Ruuskanen, J. Acc. Chem. Res.
2003, 36, 652-658.
(32) Born, J. P.; Mulder, P.; Louw, R. Environ. Sci. Technol. 1993, 27,
1849-1863.
(33) Born, J. P. Ph.D. Thesis, Leiden University, Leiden, The
Netherlands, 1992 (in English).
(34) Cieplik, M. K.; De Jong, V.; Louw, R. Environ. Sci. Technol.
(manuscript in preparation).
2
(
(
(
26) Wikstr o¨ m, E.; Marklund, S. Environ. Sci. Technol. 2000, 34, 604-
6
09.
(
27) Van den Berg, M.; Birnbaum, L.; Bosveld, B. T. C.; Brunstr o¨ m,
B.; Cook, P.; Feeley, M.; Giesy, J.; Hanberg, A.; Hasegawa, R.;
Kennedy, S. W.; Kubiak, T.; Larsen, J. C.; Van Leeuwen, F. X. R.;
Liem, A. K. D.; Nolt, C.; Peterson, R. E.; Poellinger, L.; Safe, S.;
Schrenk, D.; Tillit, D.; Tysklind, M.; Younes, M.; Waern, F.;
Zacharewski, T. Environ. Health Perspect. 1998, 106, 775-792.
28) Wikstr o¨ m, E. Ph.D. Thesis, Umeå University, Umeå, Sweden,
(35) Cieplik, M. K. Ph.D. Thesis, Leiden University, Leiden, The
Netherlands, 2003 (in English).
(36) Kanters, M. J. Ph.D. Thesis, Leiden University, Leiden, The
Netherlands, 1996 (in English).
(37) Huang, H.; Buekens, A. Chemosphere 2001, 44, 1505-1510.
(
Received for review July 24, 2003. Revised manuscript re-
ceived June 23, 2004. Accepted July 12, 2004.
1
999 (in English).
(
(
29) Addink, R.; Olie, K. Environ. Sci. Technol. 1995, 29, 1425-1435.
30) Stieglitz, L.; Bautz, H.; Roth, W.; Zwick, G. Chemosphere 1997,
3
4, 1083-1090.
ES034820Y
VOL. 38, NO. 19, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5 2 2 3