Removal of dioxins and related aromatic hydrocarbons from flue gas streams by adsorption and catalytic destruction

Article abstract of DOI:10.1016/S0045-6535(00)00235-6

The dioxin removing capacity of the shell dedioxin system (SDDS a - Ti/V oxidative type catalyst) has been tested using the Umefa lab-scale incinerator over the temperature range 100 -230°C and at space velocities of 8000 and 40,000 h^-1. Other analogous organic compounds, such as PCBs, PAHs, chlorobenzenes and chlorophenols have also been investigated. Results show a high degree of dioxin removal already at 100°C (82%), which occurs mainly by adsorption. When the temperature is raised a transition towards destruction is seen and at 150°C, gas hour space velocity (GHSV) 8000 and at 230°C, GHSV 40,000 virtually all removal is by destruction. High PCDD/F destruction efficiencies are reported (> 99.9%, based on I-TEQ); the other dioxin-related species and PAHs are also removed and destroyed to a significant extent. The SDDS has proved to be an effective means of destroying organic compounds in the gas phase, particularly dioxins, at temperatures as low as 150°C.

Full text of DOI:10.1016/S0045-6535(00)00235-6

Chemosphere 42 (2001) 615±623
Removal of dioxins and related aromatic hydrocarbons from
¯ue gas streams by adsorption and catalytic destruction
a,
Per Liljelind , John Unsworth , Onno Maaskant , Stellan Marklund
b
c
a
*
a
Environmental Chemistry, Chemistry Department, Ume ꢀa University, SE-902 87 Ume ꢀa , Sweden
Fuels Business Group, Shell Global Solutions, Shell Research and Technology Centre, Thornton, P.O. Box 1, Chester CH1 3SH, UK
b
c
CRI Technology Services BV, Koninginnegracht 23, 2514AB The Hague, Netherlands
Abstract
The dioxin removing capacity of the shell dedioxin system (SDDS ± a Ti/V oxidative type catalyst) has been tested
using the Ume ꢀa lab-scale incinerator over the temperature range 100±230°C and at space velocities of 8000 and 40,000
�
1
h
. Other analogous organic compounds, such as PCBs, PAHs, chlorobenzenes and chlorophenols have also been
investigated. Results show a high degree of dioxin removal already at 100°C (82%), which occurs mainly by adsorption.
When the temperature is raised a transition towards destruction is seen and at 150°C, gas hour space velocity (GHSV)
8
000 and at 230°C, GHSV 40,000 virtually all removal is by destruction. High PCDD/F destruction eciencies are
reported (> 99:9%, based on I-TEQ); the other dioxin-related species and PAHs are also removed and destroyed to a
signi®cant extent. The SDDS has proved to be an e€ective means of destroying organic compounds in the gas phase,
particularly dioxins, at temperatures as low as 150°C. Ó 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Catalyst; Chlorobenzenes; Chlorophenols; Combustion; Conversion; End-of-pipe; Incineration; Low temperature; PCB;
PCDD; PCDF; PAH
1
. Introduction
(Buekens and Huang, 1998). The principal technology
introduced has been adsorption onto active carbon, of-
ten combined with lime injection to remove acids, and
installed before a bag ®lter. This process yields solid
residue containing carbon contaminated with dioxins,
heavy metals etc., which needs handling as hazardous
waste, and which presents a continuous disposal prob-
lem. In addition to yielding contaminated waste, ®xed
carbon beds can pose ®re hazards (Siret and Gilman,
1996).
The technology of low temperature catalytic removal
and destruction of dioxins and related compounds from
incineration processes is a promising strategy for end-of-
pipe use. Fitting low temperature catalytic converters
after the scrubber and bag house in an incinerator can
avoid costly reheating of ¯ue gases, yet e€ect conversion,
thus minimising investment needed for remediating
measures.
Several technologies for polychlorinated dibenzo-p-
dioxin (PCDD) and dibenzofuran (PCDF) removal have
been developed in response to legislation limiting the
dioxin concentration in ¯ue gas from waste incineration
Not only is the average dioxin removal eciency
important but also the ®nal form of the dioxin hazard,
which in adsorptive solutions, is simply transferred from
the gas phase to a solid stream. Measures to limit the
likelihood of intermittent high dioxin excursions also
need to be considered. Thus, focus should be on solu-
tions which are both fail-safe, and which actually de-
stroy dioxins rather than merely removing them for
recycling elsewhere in the biosphere. Some preventive
*
Corresponding author. Tel.: +46-90-786-9321; fax: +46-90-
28-133.
E-mail address: per.liljelind@chem.umu.se (P. Liljelind).
1
0
045-6535/01/$ - 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 3 5 - 6

6
16
P. Liljelind et al. / Chemosphere 42 (2001) 615±623
measures have proved ecient, such as hot gas
quenching to avoid the low temperature dioxin forming
temperature window (250±450°C). This is an option but
the heat recovery eciency is reduced and it is often still
necessary to clean ¯ue gases further (F a ngmark et al.,
fectiveness of the system for the destruction of PCDDs
and PCDFs were determined, together with that for
additional analogous compounds: polychlorinated bi-
phenyls (PCBs), polycyclic aromatic hydrocarbons
(PAHs) and dioxin precursors, e.g., polychlorobenzenes
(PCBz) and polychlorophenols (PCP).
1
994).
From the available catalysts with varying properties,
the oxidative type consisting of metal oxides with dif-
ferent composition has proved to be e€ective for PCDD/
F removal (Hiraoka et al., 1989; Boos et al., 1992;
Morsbach et al., 1996). Catalytic reactors of the
monolithic and honeycomb type are widely used. In this
study, however, a grain-shaped, highly porous extrudate
is employed, which provides a bed that gases are passed
through. This ensures good contact between gas com-
ponents and catalytic surfaces, and hence gives high
conversion rates. Beds can cause fairly high pressure-
drops unless container design measures are taken to
minimise them. The equivalent construction for use in
full-scale plants is the lateral ¯ow reactor where verti-
cally placed sub-units with a bed thickness of not more
than 3±5 cm are used with good results (Andersson
et al., 1998).
2. Experimental
2.1. The Ume ꢀa reactor
The 5 kW lab-scale incinerator unit (Fig. 1) is
equipped with a ¯uidised bed through which primary air
is fed. There are also two radial inlets of secondary air
above the sand bed. The fuel is fed onto the bed as
pellets (6 mm outer diameter) via a motorised screw. The
fuel used was an arti®cial municipal solid waste (MSW)
designed to re¯ect the contents of real waste, primarily
in Sweden (Wikstr o m and Marklund, 1998). The tem-
perature of the bed and freeboard was around 800°C.
The convector part (boiler) consists of ®ve 3 m long steel
tubes with temperatures from 600°C down to 200°C in
the last section. After this a cyclone is placed, followed
by a coarse metal ®lter for particle removal, and a lime
A test programme was undertaken to study the op-
erational boundary limits of the titanium/vanadium (Ti/
V) catalyst (the shell dioxin destruction system, SDDS)
supplied by CRI Catalysts UK, speci®cally developed
for low temperature utilisation as an end-of-pipe solu-
tion. The combustion unit for simulating full-scale waste
combustion at the Chemistry Department, Environ-
mental Chemistry, Ume ꢀa University, Sweden, is well
suited for doing this. The tests performed were designed
to con®rm removal of dioxins, which had been observed
earlier (Andersson et al., 1998), and to determine
whether removal occurred via adsorption or destruction.
Experiments were carried out at low temperatures and
for medium to high space velocity conditions. The ef-
®lter to reduce SO and HCl ¯ue gas concentrations to
2
levels more typical of post-air pollution control condi-
tions. A summary of operating conditions is shown in
Table 1, and for a more elaborate description of the
reactor and its operation see Wikstr o m et al. (1998). The
main gas stream went through a wet scrubber and ®nally
through a carbon chemical ®lter (activated carbon,
KMnO , Al O ), before entering the atmosphere by way
4
2
3
of a fan.
The combustion gases CO and O were continuously
2
2
measured with IR and zirkonium sensors while SO2,
HCl and particle contents were determined o€-line.
Fig. 1. Schematic view of the Ume ꢀa laboratory scale ¯uidised bed reactor and location of coupling point to the SDDS.

P. Liljelind et al. / Chemosphere 42 (2001) 615±623
2.3. The catalyst
617
Table 1
Typical Ume ꢀa lab-scale incinerator operating conditions during
the di€erent experiments with the SDDS
The catalyst, originally developed for the Shell De-
Combustion eciency
Bed and freeboard temperature
>99.5%
3
nox System for use together with ammonia (NH ) by
780 ‹ 30°C
10.5 ‹ 3.5%
ꢀ170 mg/Nm
ꢀ240 mg/Nm
CRI Catalysts UK, is a titanium dioxide/vanadium
pentoxide type extrudate of 3±6 mm length and 1.0±1.6
mm diameter. This allows ready di€usion of target
molecules to the high internal surface area giving very
high intrinsic activity and removal eciency at operating
temperatures up to 350°C. The amount of catalyst used
in the experiments was either 12 or 60 g yielding a gas
hour space velocity (GHSV) of 8000 or 40,000 with a
constant ¯ow of 8 l/min.
O
2
-level
Particle content
a
3
3
a
HCl concentration
a
SO2
3
ꢀ30 mg/Nm
a
Sampled after coupling point to SDDS as seen in Fig. 1.
2
.2. The SDDS reactor
After the lime ®lter (Fig. 1, coupling point) a side
stream was led ®rst through an additional less coarse
metal ®lter, then via heated tubing to the heated cata-
lytic reactor (SDDS, Fig. 2). This cylindrical container
made of stainless steel has an internal diameter of 23 mm
and is 250 mm long and was mounted vertically. A metal
net in each end acts as a catalyst support. The catalyst
temperature was monitored directly under the net in the
low end. After passing the SDDS gases were cooled and
sampled.
2.4. Experimental plan
The removal and destruction eciencies of PCDD/F
have been determined in six independent experiments by
analysing the ¯ue gas before the reactor (RAW), after
the reactor (CLEAN) and on the catalyst (CAT), see
Table 2. They were performed on separate days and
numbered chronologically. Experiments 1, 2 and 6 were
conducted iso-thermally with catalyst temperatures
Fig. 2. Experimental set-up of the SDDS and location of coupling point to the Ume ꢀa reactor.
Table 2
Levels of PCDD/F expressed as toxic equivalents [I-TEQ] after (CLEAN), before (RAW) and in the catalyst (CAT)
Name
Catalyst
temperature
GHSV
(ml catalyst/
CLEAN
(after catalyst)
RAW
(before catalyst)
CAT
(catalyst)
Measured
catalyst
a
3
3
3
(
°C)
ml h)
(ng/Nm )
(ng/Nm )
(ng/Nm )
temperature (°C)
a
Exp 1
Exp 2
Exp 3
Exp 4
Exp 5
Exp 6
230
150
100
150
230
100
8000
8000
0.03
0.13
10
286
±
233
152
96
55
61
0.05
89
40,000
40,000
40,000
8000
1.5
115
78
9.3
0.05
2.5
148
228
101
0.93
0.36
172
a
Exp 1B was sampled before the last ®lter, possibly yielding a high value in this case due to the particle bound PCDD/F included here.

6
18
P. Liljelind et al. / Chemosphere 42 (2001) 615±623
Table 3
Eciency of the di€erent organic compounds analysed
a
Name
Catalyst
temp.
GHSV
PCDD/F
(%)
PCB
(%)
PCB DIN
(%)
PAH
(%)
PCBz
(%)
PCPh
(%)
Removal eciency
b
Exp 3
Exp 4
Exp 5
Exp 6
100
40,000
40,000
40,000
8000
82.1
98.6
98.8
99.7
31.6
84.8
93.9
96.9
64.7
93.2
92.1
98.8
75.0
88.2
95.2
95.3
NR
NR
76.6
98.0
99.6
150
230
100
23.1
32.5
86.3
Adsorption eciency
Exp 3
Exp 4
Exp 5
Exp 6
100
150
230
100
40,000
40,000
40,000
8000
100
6.1
0.1
1.1
100
39
18
±
100
100
100
±
19
0.6
0.1
±
±
±
0.9
0.3
±
0.1
0.2
±
Destruction eciency
c
Exp 3
Exp 4
Exp 5
Exp 6
100
150
230
100
40,000
40,000
40,000
8000
ND
ND
46
76
±
ND
ND
ND
±
56
ND
22.3
32.2
±
ND
76.5
97.8
±
92.5
98.7
87.6
95.1
±
98.6%
a
Polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/F), polychlorobiphenyls (PCB) and the speci®c DIN-congeners (IU-
PAC # 28, 52, 101, 138, 153, 180), polychlorophenols (PCP), polychlorobenzenes (PCBz), and polycyclic aromatic hydrocarbons
(
PAH).
NR ± No removal.
ND ± No destruction.
b
c
2
30°C, 150°C and 100°C, respectively, at the lower space
2.5. Sampling
�
1
velocity (8000 h ), whereas experiments 5, 4 and 3
represent the same temperature sequence but at 40,000
The side stream ended going through the dioxin
samplers, each sample has been collected for approxi-
�
1
h
. In addition various related compounds, i.e., PCP,
3
PCBz, polychlorobiphenyls (PCBs) and PAHs, have also
been determined in the experiments 3±6, thus making
some eciency calculations impossible (denoted ``±'' in
Tables 2 and 3).
mately 2 h (ꢀ1 Nm of gas). The cooled probe technique
was used, where the gases are ®rst quenched to < 20°C;
this is followed by a two-stage condensation impinger
system with water and ethylene glycol, respectively, in
the ¯asks. Finally, there is a polyurethane foam adsor-
bent (PUF) and an aerosol ®lter and a second PUF as a
precaution (F a ngmark et al., 1990; Marklund et al.,
1992). The sampler is approved by the CEN (the stan-
dardising committee of the European Union) for dioxin
analysis. It also works well for the related compounds
analysed: PCBs, PAHs, PCBz and PCP.
Samples taken after the catalyst (CLEAN) corre-
spond to emissions to the atmosphere. Samples of ®l-
tered raw gas (RAW) were taken in parallel to the above
(
CLEAN) before the catalyst reactor, except in experi-
ment 2 when it was taken at the exact same location but
before in time with an empty SDDS. In experiment 1,
the raw gas sample was taken before the last ®lter and
NH
the gas stream. The ammonia is added to enable si-
multaneous removal of NO . Its e€ect on dioxin removal
3 3
-inlet. In experiments 1 and 2, NH was added to
2.6. Analysis
x
is not likely to be signi®cant. The catalyst itself was re-
placed between experiments and these samples were
denoted as ``CAT''.
The ¯ue gas samples were worked-up using standard
validated (F a ngmark et al., 1990; Marklund et al., 1992).
First, the water/ethylene glycol solution was ®ltered by
C-18 solid phase extraction (SPE), followed by soxhlet-
Dean±Stark (SDS)-extraction of ®lter, SPE-disc and
PUFÕs. Thereafter, open column liquid chromatography
(LC) was used in three steps with silica, alumina and
activated carbon. The catalyst has also been analysed ±
the whole amount was extracted with soxhlet, and fur-
ther cleaned-up using the same LC columns (Marklund,
1990; van Bavel et al., 1992). The analysis of PCDD/F
was performed on a high resolution gas chromatograph
The incinerator was run under normal operat-
ing conditions with high combustion eciency, see
Table 1. The gases entering the SDDS reactor are not
particle free, the residual ®ne dust escaping up-stream
®lters has not been separated from the catalyst mate-
rial (and would be expected to have contributed to
levels found in the catalyst samples, leading to an
overestimation of the adsorption to an unknown
extent).

P. Liljelind et al. / Chemosphere 42 (2001) 615±623
HRGC) coupled to a high resolution mass spectrome-
619
(
ter, the other organic compound groups were analysed
on a HRGC/low resolution mass spectrometer. The
compounds analysed were: tetra- to octaCDD/F, tri- to
decaCB, di- to pentaCPh, di- to hexaCBz and 3 to 5-
ringed PAHs. The 2, 3, 7, 8-substituted PCDD/Fs and
PCB DIN-congeners (IUPAC # 28, 52, 101, 138, 153,
1
80) were isomer speci®cally determined. The homo-
logue sums for PCDD/F and PCB were determined us-
ing the response of the 2, 3, 7, 8-substituted PCDD/Fs
and PCB DIN-congeners. The PAHs determined were:
phenanthrene (Ph), anthracene (A), ¯uoranthene (Fl),
pyrene (Py), benzo(a)anthracene (BA), chrysene (C),
benzo(b)¯uoranthene + benzo(k)¯uoranthene
(BFl),
benzo(a)pyrene (BPy), indeno(1,2,3-cd)pyrene (IPy),
benzo(g,h,i)perylene (BPe), dibenso(a,h)anthracene
(
DBA). In experiments 1 and 2, no sums or other
compounds than PCDD/F (I-TEQ) were determined.
The content of organic compounds in the ¯ue gas can
vary over time during an experiment due to ¯uctuations
in combustion conditions and parallel sampling was
used in order to obtain good eciency estimates. Apart
from the sampling and analytical error (ꢀ20%), cost
constraints dictated that only one sample of each type
could be taken during each experiment. Clearly, more
statistically robust eciency estimates would require
additional experiments and samples.
Fig. 3. PCDF homologue removal (R) by destruction (D) and
adsorption (A) in per cent ± comparison of space velocity dif-
ferences at constant temperatures. The lines represent levels in
3
ng/Nm (log scale), raw gas (RAW) ± normal line, after cat
(
(
CLEAN) ± stretched line, and catalyst (CAT) ± dotted line.
a) 100°C; (b) 150°C.
2
.7. Calculation
(
b), respectively. At the lower temperature, the e€ect of
The eciency estimates expressed as percentages
have been calculated as follows:
�
1
increasing space velocity from 8000 to 40,000 h is a
switch from removal mainly by destruction to one by
adsorption only. Increasing the temperature to 150°C
increases the destruction mode such that removal at
both space velocities is now predominantly via destruc-
tion. The results for PCDDs are very similar, see overall
eciences in Tables 2 and 3.
Removal, R ˆ (1)[CLEAN]/[RAW])*100
Adsorption, A ˆ ([CAT]/[RAW])*100
Destruction, D ˆ (1)([CLEAN]+[CAT])/[RAW])*100
3
. Results
Temperature e€ects: Results for the higher space ve-
locity showing the e€ect of temperature increase from
100°C to 150°C to 230°C are given in Figs. 4(a)±(f) for
PCDFs, PCDDs, PCBs, PCBz and PCPh homologues,
and for individual PAH compounds, respectively. Thus:
Dioxin levels in experiment 1, sample A had a TEQ-
3
value less than 0.1 ng/Nm and experiment 2, CLEAN,
3
was a similar level, 0.13 ng/Nm , indicating very satis-
factory dioxin removal with respect to the statutory limit
value, despite the high raw gas concentration of PCDD/
F, see Table 2. In experiment 1, RAW, the dioxin level
was especially high, probably because sampling before
the last ®lter included more particles in this sample.
Space velocity e€ects: The average residence time of
the combustion gases in the catalyst bed was much
longer (450 ms) at GHSV 8000 ml/[ml cat and h] (ex-
periments 1, 2 and 6) compared to that at GHSV 40,000
·
For PCDFs and PCDDs, removal eciency increases
from approximately 80% to > 98% as temperature
increases from 100°C to 150°C, with only a marginal
further increase from 150°C to 230°C. At 100°C re-
moval is by adsorption, at 150°C it is mainly by de-
struction, and by 230°C it is entirely via
destruction. Fig. 5 shows a contour plot of the depen-
dence of PCDD/F (as I-TEQ) removal eciency on
temperature and ¯ow rates.
(
90 ms ± experiments 3, 4 and 5). Except for PCDD/F,
·
For PCBs removal eciency increases progressively
from 32% to 94% as the temperature increases. At
100°C removal is by adsorption only, at 150°C it is
by both adsorption and destruction, and by 230°C
relatively few experiments allow comparison of space
velocity di€erences at constant temperatures. Results for
PCDFs at 100°C and 150°C are shown in Figs. 3(a) and

6
20
P. Liljelind et al. / Chemosphere 42 (2001) 615±623
Fig. 4. Temperature e€ects on overall removal (R) by destruction (D) and adsorption (A) in per cent of (a) PCDFs, (b) PCDDs,
-
1
(
levels in ng/Nm , log scale, raw gas (RAW) ± normal line, after cat (CLEAN) ± stretched line, and catalyst (CAT) ± dotted line.
c) PCBs, (d) PCBz and (e) PCPh homologues, and for (f) individual PAH compounds (space velocity 40,000 h ). The lines represent
3
it is mainly via destruction. Destruction appears to be
more e€ective for more highly chlorinated homo-
logues.
·
For PAHs removal eciency increases progressively
from 75% to 98% as temperature increases. At
100°C removal is already at a high level but by both
adsorption and destruction, at 150°C it is mainly by
destruction, and by 230°C it is almost entirely via de-
struction. No clear compound selectivity was ob-
served.
·
·
For PCBz removal eciency is low, increasing pro-
gressively from nil to 33% as the temperature increas-
es. Whatever removal is detected is mainly by
destruction, which seems to be less e€ective for more
highly chlorinated benzenes ± little or no adsorption
is observed.
Thus, the chloro-homologue pro®les for almost all
compound classes remained virtually una€ected when
levels decreased signi®cantly after the catalyst, the ex-
ception being the PCBs where there was a clear di€er-
ence in behaviour between overall sums and the DIN-
congeners. Some of the DIN-congeners seem to accu-
mulate or even form over the catalyst under certain
conditions. The total PCB amounts illustrate how well
For PCPh removal eciency increases progressively
from nil to 98% as temperature increases. Although
no removal is detected at 100°C, removal eciency
rises rapidly, but only as destruction ± again little
or no adsorption is observed. No clear homologue se-
lectivity was observed.

P. Liljelind et al. / Chemosphere 42 (2001) 615±623
621
Fig. 4. (continued).
Fig. 5. Contour plot of PCDD/F [I-TEQ] removal eciency
percentage (%) estimated from the six experiments performed,
see Fig. 3.
Fig. 6. Mode of removal, adsorption or destruction for poly-
chlorobiphenyls (PCB) depending on temperature (Temp) at a
GHSV of 40,000.

6
22
P. Liljelind et al. / Chemosphere 42 (2001) 615±623
the behaviour gradually shifted from adsorption to de-
struction when the temperature was raised (Fig. 6)
whereas dioxins predominantly were destructed as soon
as the temperature was higher than 100°C.
where the DIN-congeners resemble PCDD/Fs more in
behaviour, but only being adsorbed not destructed.
Using as high space velocity as 40,000 in the exper-
iments was warranted by ®nding the limits of e€ective
catalysis rather than by regular dedioxin situations for
standard operation and long-term use and the results
showed that it is indeed possible to operate this high
and still have a large degree of removal at 150°C and
above.
4
. Discussion
These results con®rm the dioxin removing capacity of
Possibilities to perform periodic heating exist; the
catalyst system can be operated under adsorption con-
ditions and then heated to a temperature sucient to
ensure destruction at time intervals depending on con-
centration and adsorption capacity.
the catalyst system and have distinguished between di-
oxin removal by adsorption and by catalytic destruction.
The fact that dioxins and other organic compounds
could be extracted from the catalyst showed that the
initial adsorption is reversible. There might be a risk that
the adsorption characteristics change with increasing
temperature giving irreversible binding rather than de-
struction. This needs to be ruled out and provisional
Further studies are planned to investigate removal in
more detail, also how eciency depends on other fac-
tors, such as ¯ue gas dioxin concentration, SO
HCl concentration, NH -injection, particle loading etc.
x
and
3
1
4
results of C labelled dioxin experiments have detected
1
With a better practical as well as theoretical under-
standing a reaction mechanism can be proposed. Sus-
tained long-term actitivity will also be studied and has
been demonstrated by the experience of the Watco
waste incineration plant in Roosendaal, Netherlands,
where the dioxin removing capacity of an SDS-DeNOx
reactor unit has been monitored for the last 3 yr
(Unsworth, 1997). Supporting evidence for dioxin de-
struction has been obtained from the absence of any
accumulation of dioxins on the SDS catalyst after 3 yr
operation of the Watco incinerator (Andersson et al.,
1998).
4
the formation of
destruction.
CO
2
in support of complete
Destruction eciency trends with temperature and
gas space velocity obtained for dioxins were similar
whether expressed as TEQs or when considering the sum
of the PCDD/F homologues. The chloro-homologue
pro®les from tetra to octa were also similar before and
after the catalyst passage, no relative enrichment of any
homologue group can be seen. Weber et al. (1999) saw
2 2 5 3
that on a TiO -based V O ±WO catalyst more highly
chlorinated PCDD/Fs were destroyed to a lesser extent
due to their electron withdrawal e€ect, this is something
we cannot see. Whether the mechanism of destruction
means C±C bond breakage directly or is via loss of
substituents needs more study even though we do not see
dechlorination of PCDD/Fs here. Weber et al. (1999)
also saw little destruction at 150°C, 59±75% adsorption,
while we at 8000 SV virtually only have destruction
5. Summary
The shell dedioxin system (SDDS) has been shown
to be a highly ecient way of removing gas phase
PCDD/F, even at temperatures as low as 100°C and
under relatively severe ¯ue gas composition condi-
tions.
(
99.7%). These di€erences are probably due to the dif-
ferent type of catalyst material and experimental pro-
cedures used.
In the transition from adsorption to destruction of
PCBÕs, seen in Fig. 6, what might be a rearrangement of
chlorine atoms or a dechlorination of the biphenyl back-
bone occurs prior to ring cleavage. This is supported by
the increased amount of speci®c congeners adsorbed in
the CAT samples compared to RAW case. More data,
however, are needed in order to formulate a mechanism
for destruction.
By measuring dioxin levels in the catalyst, as well as
¯ue gas composition pre- and post-catalyst, it has been
possible to distinguish between dioxin removal by ad-
sorption and by catalytic destruction for a range of
temperatures and space velocities.
Dioxin destruction by SDDS could proceed via for-
mation of other aromatic structures like chlorinated
phenols or benzenes, however, these species are not ac-
cumulated but instead converted themselves almost as
e€ectively.
The temperature dependence for e€ective removal is
more pronounced for chlorophenols than for dioxins
and PAH. This could be caused by the molecular size
where the larger PAHs and dioxins have a lower vola-
tility and are more easily adsorbed onto catalytic sur-
faces, and the higher activation energy for conversion
for smaller molecules like PCPhs. The total PCB re-
moval is intermediately dependent on temperature
The ratio of gas ¯ow rate to amount of catalyst
(GHSV) is more important for the removal eciency
than temperature over the range tested. Destruction ef-
®ciencies higher than 98% were obtained for the PCDD/
F at the lower gas space velocity at all three tempera-
tures from 100°C to 230°C, however, for the higher

P. Liljelind et al. / Chemosphere 42 (2001) 615±623
623
GHSV destruction eciencies were temperature depen-
dent.
To ensure destruction and avoid adsorption and in
F a ngmark, I., Wikstr o m, L.-E., Marklund, S., Rappe, C., 1990.
Studies on sampling methods for PCDD and PCDFs in
stack emission. Chemosphere 20, 1333±1340.
F a ngmark, I., Str o mberg, B., Berge, N., Rappe, C., 1994. The
in¯uence of post combustion temperature pro®les on the
formation of PCDDs, PCDFs, PCBzs and PCBs in a pilot
reactor. Environ. Sci. Technol. 28, 624±629.
3
order to achieve less than 0.1 ng I-TEQ per Nm in the
o€-gas a temperature of approximately 150°C is pre-
ferred. The precise temperature needed is dependent
both on the amount of catalyst and on the level of dioxin
in the ¯ue gas.
Hiraoka, M., Takeda, N., Okajima, S., 1989. Catalytic
destruction of PCDDs in ¯ue gas. Chemosphere 19, 361±
366.
Marklund, S., S o derstr o m, G., Ljung, K., Rappe, C., Kraft,
M., Hagenmaier, H.-P., 1992. Parallel sampling using
various sampling techniques at a Swedish MSW incinerator.
Waste Management Res. 10, 21±36.
Acknowledgements
The authors thank Gunilla L o fvenius and Maria
Hjelt for the invaluable help in the chemical analysis.
The project MINIDIP ± Minimization of Dioxins in
Thermal Industrial Procedures, funded by the European
Union, is acknowledged for ®nancial support.
Marklund, S., 1990. Dioxin emissions and environmental
imissions. Ph.D. Thesis, Ume ꢀa University, Sweden ISBN
9
1-7174-496-7.
Morsbach, B., Odermatt, P., Spahl, R., 1996. BASF catalysts
for the removal of NO and dioxin in various applications.
x
In: Proceedings of the VTT Symposium 163, pp. 77±94.
Siret, B., Gilman, K., 1996. A survey of post-combustion
PCDD/PCDF control technologies. In: Proceedings of the
International Conference Incineration and Thermal Treat-
ment Technology, pp. 583±586.
References
Andersson, P., Rappe, C., Maaskant, O., Unsworth, J.F.,
Marklund, S., 1998. Low temperature catalytic destruction
of PCDD/F in ¯ue gas from waste combustion. Organo-
halogen Compound 36, 109±112.
van Bavel, B., F a ngmark, I., Marklund, S., S o derstr o m, G.,
Ljung, K., Rappe, C., 1992. Quantitative analysis of organic
compounds by isotope dilution: formation of non-ortho
PCBs 77, 126, 169 during incineration. Organohalogen
Compound 8, 225±228.
Unsworth, J.F., 1997. Dioxin removal by CRIÕs shell denox
system. Heeren Waste Incineration trials, Roosendaal, The
Netherlands. Report, CRI Catalyst Co. UK Ltd., Woking,
GU21 5BH, UK.
Weber, R., Sakurai, T., Hagenmaier, H., 1999. Low tempera-
ture decomposition of PCDD/PCDF, chlorobenzenes and
PAHs by TiO
20, 249±256.
2 2 5 3
-based V O ±WO catalysts. Appl. Catal. B
Boos, R., Budin, R., Hartl, M., Stock, M., Wurst, F., 1992.
PCDD- and PCDF-destruction by a SCR-unit in a munic-
ipal waste incinerator. Chemosphere 25, 375±382.
Buekens, A., Huang, H., 1998. Comparative evaluation of
techniques for controlling the formation and emission of
chlorinated dioxins/furans in municipal waste incineration.
J. Hazard. Mater. 62, 1±33.
Wikstr o m, E., Marklund, S., 1998. Combustion of an arti®cial
municipal solid waste in a laboratory ¯udised bed reactor.
Waste Management Res. 16, 342±350.
Wikstr o m, E., Andersson, P., Marklund, S., 1998. Design of a
laboratory scale ¯uidized bed reactor. Rev. Sci. Instr. 69,
1850±1859.