Chemosphere 78 (2010) 672–679

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Experimental study on the effects of H2O on PCDD/Fs formation by de novo

synthesis in carbon/CuCl2 model system
Ke Shao, Jianhua Yan *, Xiaodong Li, Shengyong Lu, Yinglei Wei, Muxing Fu
State Key Laboratory of Clean Energy Utilization, Zhejiang University, Hangzhou 310027, China

a r t i c l e i n f o a b s t r a c t

Article history: The effects of H2O vapor on de novo synthesis of polychlorinated dibenzo-p-dioxins and dibenzofurans
Received 31 August 2009 (PCDD/Fs) and biphenyls (PCB) were investigated at two levels (none and 10 vol.%) in various model sys-
Received in revised form 15 November 2009 tems containing five different carbons, CuCl2, and quartz, exposed to a flow of 10% O2/N2 at 300 °C. The
Accepted 7 December 2009
influence of H2O was studied on (1) speciation and behavior of copper compounds, (2) catalytic oxidation
of carbons of distinct reactivity, and (3) formation of organic chlorine compounds, with the aim to inves-
tigate any effects on de novo synthesis. It is found that H2O converts CuCl2 to CuCl2CuO, and finally to
Keywords:
CuO in a flow of 10% O2/N2 at 300 °C and that it decreases of organic chlorine (C–Cl) formation. When
Cuprous chloride
Model ash
CuCl2 is supported on carbon, the addition of H2O promotes the catalytic oxidation of this carbon. When
Organic chlorine CuCl2 is supported on quartz, however, H2O inhibits carbon oxidation. A decrease in chlorination level of
Catalytic oxidation PCDD/Fs and PCBs with water addition is observed for all (six) model ashes; yet this addition affects the
De novo synthesis yields of PCDD/Fs and PCBs differently. Under the experimental conditions tested H2O does not react with
Cu2Cl2, which is the catalyst of carbon oxidation. On the basis of the experimental results, the following
mechanism is proposed: conversion of CuCl2 into CuO which is less reactive in de novo synthesis and pro-
motion of catalytic oxidation of carbon by Cu2Cl2.
Ó 2009 Published by Elsevier Ltd.

1. Introduction demonstrated that CuCl2 can chlorinate activated carbon according

to the following reaction (Weber et al., 2001):
Incineration of municipal solid waste (MSW) allows a consider-
able volume reduction, as well as safe disposal and energy recov- 2CuCl2 þ R-H ! 2CuCl þ R-Cl þ HCl: ðR1Þ
ery. Unfortunately, toxic compounds, such as polychlorinated
The Cu2Cl2 is then re-chlorinated by gas-phase chlorine species
dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) are generated
(Takaoka et al., 2005). For the second step, Takaoka et al. (2005)
(Olie et al., 1977). Precursor as well as de novo synthesis have both
identify different copper species, and conclude that Cu2Cl2 plays
been proposed as a crucial mechanism in the formation of dioxins
an important role in the catalytic oxidation of carbon and in the for-
in the post-combustion zone of a MSW incinerator (Stieglitz and
mation of PCDD/Fs, PCBs, and CBzs in fly ash within the de novo
Vogg, 1987). Numerous researches have been carried out to inves-
temperature window.
tigate the influence of individual process parameters, such as tem-
It should be noted that water vapor is always amply present in
perature (Addink et al., 1991), carbon structure (Wikstrom et al.,
the flue gas from MSW incineration: it is present in MSW, in com-
2004), chlorine origin (Wikstrom et al., 2003a), oxygen (Wilhelm
bustion air, and it is formed by combustion. In small plants, flue
et al., 1999) and types of catalyst (Stieglitz et al., 1989), on de novo
gases are quenched with water before their cleaning. Therefore,
synthesis and its reaction mechanism.
the influence of water on formation of PCDD/Fs must be clarified.
The basic mechanistic steps for de novo formation have been
The research of Ross et al. (1990) and Sakurai et al. (1996) indicate
proposed by Stieglitz (1998): (a) transfer of chlorine to carbona-
that water enhances PCDD/Fs formation. On the other hand Briois
ceous matter with formation of carbon–chlorine bonds, and (b)
et al. (2007) found that the yield of PCDD/F was considerably re-
oxidation of the (partly chlorinated) macromolecular structure to
duced. However, both studied the heterogeneous precursor route,
carbon dioxide, at the same time releasing PCDD/Fs, PCBs and
using chlorophenols to produce PCDD/F on fly ash or model fly
numerous other compounds as reaction by-products. In both steps,
ash. Contradictory results were also reported on the effect of water
copper is thought to be a key catalyst. For the first step, it has been
on de novo synthesis of PCDD/F. Stieglitz et al. (1990) found that it
enhances the yield of PCDD/F. However, Jay and Stieglitz (1991)
* Corresponding author. Tel.: +86 571 87952629; fax: +86 571 87952438. observed considerably less PCDD and PCDF in the presence of
E-mail address: yanjh@zju.edu.cn (J. Yan). water vapor. Meanwhile, Suzuki et al. (2004) reported that water

0045-6535/$ - see front matter Ó 2009 Published by Elsevier Ltd.

doi:10.1016/j.chemosphere.2009.12.011
 K. Shao et al. / Chemosphere 78 (2010) 672–679 673

Table 1
Analysis of carbons used: short analysis, elementary analysis, metal analysis, BET specific surface and pore volume.

Activated carbon CBA CBB CBC Graphite

Moisture (%) 1.73 1.88 1.09 2.74 0.36
Ash (%) 2.09 0.27 0.01 0 0.13
Volatiles (%) 7.38 4.71 11.21 2.92 2.95
Fixed carbon (%) 88.80 93.14 87.69 94.34 96.56
Total short analysis (%) 100 100 100 100 100
C (%) 92.52 95.41 93.21 93.21 98.52
H (%) 1.50 1.30 1.18 1.18 0.09
O (%) 1.67 0.61 3.76 0.64 0.82
N (%) 0.17 0.23 0.22 0.22 0.02
S (%) 0.19 0.30 0.54 0.06 0.06
Cl (%) 0.14 0 0 0 0
Total elementary analysis (%) 96.19 97.85 98.91 95.31 99.51
Al (ppm) 654.7 75.1 66.5 4.1 63.5
Ca (ppm) 591.9 776.1 407.0 471.5 728.2
Cu (ppm) 275.9 0.1 14.7 0.1 1.7
Fe (ppm) 506.9 252.8 642.9 1.8 78.1
Ni (ppm) 81.6 4.9 11.9 0.5 25.4
Zn (ppm) 54.0 114.2 6.9 1.6 71.0
Total metal (ppm) 2165 1223 1150 479.6 967.9
BET-SA (m2 g1) 933.93 71.47 114.85 107.35 6.68
Pore volume (cm3 g1) 0.63 0.54 0.69 0.74 0.03

vapor has little effect on dioxin formation in the dry zone of an iron
ore sintering bed.
In the presence of water, Wikstrom et al. (2003b) reported a
slightly enhanced carbon oxidation and a lower amount of total
Cl and organic chlorine in the ash. Water also slightly reduced
the yields of PCDD/Fs by 17–30% and significantly shifted the PCDF
homologue profile towards the lower chlorinated homologues.
Thus, the presence of water appears to play multiple roles in
PCDD/F formation; viz., enhancing the oxidation of active carbon
to CO and CO2, reducing Cl2 activity by converting it into HCl,
and dechlorinating the PCDD/Fs already-formed. Li et al. (2006)
indicate that water reduces the catalysis of CuCl2 and prevents di-
rect chlorination.
The present work tries to confirm these effects separately by
studying the effect of water on each mechanistic step of de novo
synthesis. Considering the relationship between catalytic oxida-
tion of carbon and PCDD/Fs formation and the important role of
copper chlorides in de novo synthesis, experiments on the effect
of H2O on the chemical speciation of copper, the formation of or-
ganic chlorine, and the catalytic oxidation of carbon were
conducted.
2. Experimental (Shao et al., 2009)
2.1. Chemicals
Reagent-grade quartz (100–120 mesh) was first rinsed twice by
distilled water, and then completely dried at 120 °C. Glass wool
was cleaned by immersion in a diluted HNO3 solution for 12 h,
rinsed by distilled water, and then dried at 500 °C to remove all or-
ganic compounds. Five kinds of carbon were used. Their analysis
and characteristics are given in Table 1. The three kinds of partic-
ulate carbon blacks were obtained from Jinderui Corp. (labeled
CBA and CBB) and Cabot Corp. (labeled CBC). CBA and CBC are fur-
nace blacks, and CBB is a channel black. Three carbon blacks, acti-
vated carbon and graphite were first rinsed twice with distilled
water and acetone, respectively, and then completely dried at
100 °C.
Reagent-grade CuCl22H2O, Cu(NO3)2, KNO3, HNO3, activated
carbon particles, pesticide-grade toluene, methanol, methylene
chloride and acetone, HPLC-grade n-hexane were used. N2 of
99.999% and O2 of 99.995% were supplied by Jingong Corp.
2.2. Sample preparation
The model mixtures used in de novo synthesis were prepared as
follows: acetone and aqueous solution of CuCl2 was added to the
activated carbon and continuously hand-mixed. It was then dried
at 100 °C. The Cu content in this mixture was 2.7%. Then, 3.3 g of
mixture was added to 96.7 g quartz and rotary mixed continuously
for a week to produce model ash [AC]. The model ashes [CBA],
[CBB], [CBC] and [Graphite] were prepared similarly as model ash
[AC].
An aqueous solution of CuCl2 was added to quartz and dried in a
rotary evaporator. It was then dried at 120 °C for 2 h. Its Cu content
in model ash [CuCl2] was also 0.09 wt.%. Finally, 3 g of activated
carbon was added to 97 g [CuCl2] and continuously hand-mixed
to produce model ash [AC].
2.3. Experimental procedure
2.3.1. De novo experiment
Tests were run in a tubular quartz reactor, inserted into a hori-
zontally mounted, three-stage electric furnace. The temperature at
each stage was controlled by a single S type thermocouple. Then
2 g model ash was held in the reactor with a plug of glass wool
and treated at 300 °C in a flow of either 10% O2/N2 or 10% H2O/
10% O2/N2. Total gas flow rate was 360 mL min1. The total reac-
tion time was 30 min, unless otherwise noted.
2.3.2. Sampling and analysis of PCDD/Fs, PCBs
For sampling gas-phase products, the outlet gas from the reac-
tion tube passed through ice-cooled traps consisting of a XAD-2
resinous column and an impinger absorber containing toluene.
After the desired reaction time, acetone, methylene chloride and
toluene were used for rinsing the inside of the reactor, gas tubes
and connectors. The rinses were included in the analysis of the
gas traps. Products remaining in the solid sample residue after
the experiment were also analyzed. The EPA 1613 isotope dilution
method was used for tetra–octa PCDD/Fs determination, and the
EPA 1618A isotope dilution method was used for PCBs determina-
tion. Samples were analyzed by means of high-resolution gas chro-
matography/high-resolution mass spectrometry (HRGC/HRMS
JEOL, JMS-800D) with a DB-5MS (60 m length  0.25 mm
ID  0.25 lm film) column.
674 K. Shao et al. / Chemosphere 78 (2010) 672–679

Table 2
Summary of the experimental conditions and results.

Test Experiment duration (min) Model ash H2O (%) Inorganic chlorine content (lg g1) EOCl content (lg g1) NEOCl content (lg g1)
Gas-phase Solid-phase
1 30 [CuCl2] None 518 ± 6 N/A N/A N/A
2 60 [CuCl2] None 457 ± 17 N/A N/A N/A
3 30 [AC] None 349 ± 12 8±3 6±2 272 ± 34
4 30 [CBA] None 241 ± 8 62 ± 8 2 311 ± 21
5 30 [CBB] None 196 ± 3 49 ± 4 2 372 ± 23
6 30 [CBC] None 305 ± 8 74 ± 17 7±3 326 ± 23
7 30 [Graphite] None 268 ± 40 48 ± 11 3 14 ± 5
8 30 [AC] None 431 ± 12 4±2 4±2 156 ± 17
9 30 [CuCl2] 10 92 N/A N/A N/A
10 30 [AC] 10 248 ± 12 7±3 6±3 268 ± 7
11 30 [CBA] 10 178 ± 3 30 ± 8 3±1 242 ± 11
12 30 [CBB] 10 150 ± 6 21 ± 5 2 307 ± 28
13 30 [CBC] 10 230 ± 25 29 ± 4 3±2 213 ± 18
14 30 [Graphite] 10 89 ± 25 18 ± 3 n.d. 22 ± 5
15 30 [AC] 10 95 ± 13 2 n.d. 64 ± 17

All experiments: 2.0 g model ash; Tests 1–7 in triplicate, Tests 8–15 in duplicate. Mean value ± standard deviation is shown, no standard deviation for Cl is given if negligible
(<1 lg g1); N/A = not analyzed; n.d. = non-detected, i.e. below detection limits (=1 lg g1).

2.3.3. Measurement of organic chlorine and C–Cl
For sampling gas phase organic chlorine, the outlet gas was sent
through ice-cooled sampling columns of activated carbon. After the
reaction was finished, the activated carbon was washed by distilled
water and further by an aqueous solution of potassium nitrate (0.8
mass% KNO3) to remove inorganic chlorides. Then, the activated
carbon bulk was burnt at 1000 °C in 80% O2/20% H2O mixture. Dis-
tilled water was used to absorb the HCl in the outlet gas. Chlorine
ion concentrations were measured in the resulting solutions by ion
chromatography (Metrohm, 729 Basic IC). Also the organic chlo-
rides in the model ash were analyzed. Briefly, the residual ash
was mixed with 50 mg activated carbon, and extracted with a di-
luted nitric acid solution (pH = 2) at 70 °C for 30 min followed by
a rapid quench. After 60 min of mixing via a mechanical shaker,
the sample was filtered and washed with a nitrate solution to dis-
place inorganic chlorides. Chlorine ion and sulfate ion concentra-
tions in the eluate were measured by ion chromatography. The
ash plus 50 mg of activated carbon was dried at 100 °C. The total
extractable organic chlorine (EOCl) remaining in the solid was ex-
tracted by toluene for 24 h. Then the amounts of EOCl and non-
extractable organic chlorine (NEOCl) in the model ash were mea-
sured similarly as for outlet gas analysis.
2.3.4. Measurement of CO and CO2
Through the experiments the CO2, CO, HCl concentrations in the
flue gas were measured continuously using an infrared gas ana-
lyzer (FTIR). An infrared spectrum was collected automatically
every 20 s. Each experimental condition was tested two times, to
check the reproducibility of the experiments.
3. Results and discussion
3.1. Influence of H2O on organic chlorine formation
A summary of the experiments performed and of the results ob-
tained is presented in Table 2; the values are averages of more than
two experiments. After 30 min reaction in a flow of 10% O2/N2, half
of the original chlorine content in model ash [CuCl2] was reduced
(Test 1). However, further de-chlorination of CuCl2 in this model
ash was much slower (Test 2). 46% of the original chlorine was left
after 60 min reaction. This suggests that CuCl2CuO is formed. Then
CuCl2CuO decomposes to form CuO and starting material, CuCl2
(Takaoka et al., 2005).
4CuCl2 þ O2 ! 2CuCl2  CuO þ 2Cl2 ðR2Þ
CuCl2  CuO $ CuCl2 þ CuO ðR3Þ
The addition of 10% H2O to the gas phase resulted in an appar-
ent reduction of the inorganic chlorine in the model ashes. In the
model ash [CuCl2], only 10% of the original chlorine was left after
30 min reaction (Test 9). It has been reported that CuCl2 can react
with H2O to form CuCl2CuO by reaction (R4) (Lewis et al., 2009).
This result also suggests that the rate of R4 is much faster than that
of R2.
H2 O þ 2CuCl2 ! CuCl2  CuO þ 2HCl ðR4Þ
The inorganic chlorine in model ashes also decreased, compared
with the results without H2O. With the exception of [Graphite]
and [AC], the extent of decrease was much less than the results
of [CuCl2]. This distinction may be attributed to the different chem-
ical speciation of copper between the model ashes.
The addition of H2O also decreased organic chlorine formation,
in line with the results of Wikstrom et al. (2003b). However, Cl2
was not used in our experiments. Therefore, the decrease of organ-
ic chlorine was ascribed to the conversion of CuCl2 to CuCl2CuO by
R4. For model ash [AC], the addition of H2O has little influence on
organic chlorine formation, suggesting that R1 is very fast in this
model ash. However, for model ash [AC], a remarkable decrease
(60%) in C–Cl is observed when 10% H2O is introduced. In model
ash [AC], the support of CuCl2 is quartz. H2O converted the CuCl2
to CuCl2CuO. This inhibited the CuCl2 transfer to carbon and its
chlorination. For model ashes [CBA], [CBB], [CBC] and [Graphite],
addition of H2O decreases the formation of C–Cl, by 27%, 22%,
40% and 38%, respectively. For all six model ashes, the inhibiting ef-
fect of H2O is substantially more pronounced for EOCl than for
NEOCl, suggesting that the release of organic chlorine from carbon
matrix is also inhibited.
3.2. Effect of H2O on catalytic oxidation of carbon
The effect of H2O on catalytic oxidation of carbon is illustrated
in Fig. 1, shown for CuCl2 with AC, CBC and graphite; the values
are averages of two experiments. The results of model ash [CBA]
and [CBB] (not shown) are similar to model ash [CBC]. The red lines
represent the CO2 produced from model ashes in 10% O2/N2 at
300 °C. The black lines represent the CO2 produced from model
ashes without CuCl2. The green and blue lines represent the CO2
produced when water is added from 0 to 90 min (0–30 min for
 K. Shao et al. / Chemosphere 78 (2010) 672–679 675

1500 AC
Model ash [AC]
Model ash [AC] + H 2O (0-90min)
1200 Model ash [AC] + H2O (30-60min)

CO2 (ppm)
900

600

300

0
0 15 30 45 60 75 90

600 AC
Model ash [AC*]
Model ash [AC*] + H 2O (0-90min)
500
Model ash [AC*] + H2O (0-30min)

400
CO2 (ppm)

300

200

100

0
0 15 30 45 60 75 90

3500
CBC
3000 Model ash [CBC]
Model ash [CBC] + H 2O (0-90min)

2500 Model ash [CBC] + H 2O (30-60min)

CO2 (ppm)

2000

1500

1000

500

0
0 15 30 45 60 75 90

250
Graphite
Model ash [Graphite]
200 Model ash [Graphite] + H 2O (0-90min)
CO2 (ppm)

150

100

50

0
0 15 30 45 60 75 90
Time (min)

Fig. 1. Effect of H2O on catalytic oxidation of carbon (2 g of model ash, 300 °C, 10% O2, 10%H2O, balance N2).

model ash [AC]) and 30–60 min, respectively. The results show
that CuCl2 dramatically catalyzes the oxidation of carbon. Takaoka
et al. (2005) have reported that only low-valence copper com-
pounds were observed in model fly ash containing activated car-
bon at 300 °C under 10% O2/90% N2. This suggests that Cu2Cl2 is
the catalyst in oxidation of carbon.
For [AC], when 10% H2O was added during a time period from
30 to 60 min, the concentration of CO2 increased sharply at the
time of addition, and then dropped slowly with time (Fig. 1). After
stopping the addition, the CO2 concentration rapidly returned to
the same level as in the case without H2O. Similar results were also
observed when carbon blacks were used as carbon sources. These
676 K. Shao et al. / Chemosphere 78 (2010) 672–679

18 180
PCDDs in the gas phase

Amount of PCDFs in the gas sample

Amount of PCDDs in the gas sample
PCDFs in the gas phase 150
15
120
12 90

(ng g sample)
(ng g sample)
60
9

-1
-1

6
6
4
3
2

0 0
[AC] [CBA] [CBB] [CBC] [Graphite] [AC*]
[AC]+H2O [CBA]+H2O [CBB]+H2O [CBC]+H2O [Graphite]+H2O [AC*]+H2O

180 1600
PCDDs in the solid phase
Amount of PCDDs in the solid sample

Amount of PCDFs in the solid sample

PCDFs in the solid phase
150
1200
120

(ng g sample)
(ng g sample)

90 800

-1
-1

60
400
30

0 0
[AC] [CBA] [CBB] [CBC] [Graphite] [AC*]
[AC]+H2O [CBA]+H2O [CBB]+H2O [CBC]+H2O [Graphite]+H2O [AC*]+H2O

Fig. 2. Effect of H2O on PCDD/Fs formation by de novo synthesis. All tests with two or three repeats, mean value with standard deviation is shown (2 g of model ash with
0.19%CuCl2, 3%C, bulk quartz; 300 °C, 10% O2, 10%H2O, balance N2, 30 min).

results suggest that H2O can promote the catalytic oxidation of car-
bon and does not react with Cu2Cl2. This result confirms that de-
chlorination by H2O has less influence on copper chloride sup-
ported on activated carbon or carbon blacks. Wikstrom et al.
(2003b) proposed that chemisorption of water to free carbon sites
or to oxygen in the macromolecular carbon structure in the fly ash
change the reactivity of the carbon and enhance carbon oxidation.
In another case, 10% H2O was added from the beginning and un-
til the end at 90 min (Fig. 1). For model ashes [AC], [CBC] and
[Graphite], the addition of H2O promoted the oxidation of carbons.
However, for model ash [AC], the concentration of CO2 was much
lower, compared with results without H2O. This distinction may be
attributed to the difference of support of copper catalyst. In model
ash [AC], the support of CuCl2 is activated carbon. CuCl2 is reduced
by carbon to form Cu2Cl2. According to the results above, H2O does
not react with Cu2Cl2 but it promotes the catalytic oxidation of car-
bon. These results suggest that the chlorination of carbon by CuCl2
is very fast, compared with the reaction between CuCl2 and H2O.
For [AC], the support of CuCl2 is quartz. H2O converts the CuCl2
to CuCl2CuO and finally to CuO which is less reactive in de novo
synthesis (Lewis et al., 2009).
3.3. Effect of H2O on PCDD/Fs, PCB formation
The influence of H2O on PCDD/Fs, PCBs formation by de novo
synthesis is illustrated in Figs. 2–5. The composition of model
ashes [AC] and [AC] is the same. Surprisingly, the amounts of
PCDD/Fs and PCBs produced in model ash [AC] are 2.8 and 2.2
times more than those in model ash [AC], respectively. The support
of CuCl2 also influences the congener distribution. For model ash
[AC], OCDF and OCDD both accounted for approximately 70% of
the total Te-OCDF and Te-OCDD formed, respectively. However,
the formation of moderately chlorinated PCDD/F (Te-Hx) was fa-
vored in the model ash [AC]. The chlorination level of PCBs formed
in [AC] was also higher than that in [AC]. The reason for this dis-
tinction is that the nature of the support of CuCl2 changes the
chemical form of the copper catalyst in model ash. In [AC], only
part of CuCl2 was transferred to carbon. The copper chloride left
on the quartz remained at the bivalent state and could chlorinate
the PCDD/Fs and PCBs formed from de novo synthesis. The
[PCDD]:[PCDF] ratios for model ashes [AC], [CBA], [CBB], [Graph-
ite], and [AC] were comparable, at 0.08, 0.08, 0.12, 0.09, 0.19,
and 0.21, respectively.
The addition of 10% H2O exhibits multiple effects on the
yields of PCDD/Fs when different model ashes are used. For mod-
el ashes [AC], [CBB] and [CBC], H2O promoted the PCDD/Fs for-
mation. For model ashes [AC], [CBA] and [Graphite], H2O
exhibited lower yields of PCDD/Fs. The inhibiting effects in
[CBA] and [Graphite] were both slight (a decrease by 8% and
11%, respectively). In [AC], a remarkable decrease in PCDD/Fs
was observed when 10% H2O was introduced. This result is in
agreement with the results of influence of H2O on catalytic oxi-
dation of carbon in model ashes [AC]. These results were not
unexpected, considering that H2O has two aspect influences on
de novo synthesis: Converting CuCl2 to CuCl2CuO to decrease
the C–Cl and Cu2Cl2 formation and promoting catalytic oxidation
of carbon. It is also found that for all six model ashes, H2O in-
creases the [PCDD]:[PCDF] ratios, to 0.09, 0.10, 0.15, 0.10, 0.29
and 0.41, respectively.
While the effect of H2O on the amounts of PCDD/F produced ap-
peared to depend on carbon source and mixing manner of model
ashes, the effect on congener distribution was not. For all kinds
of model ashes, the introduction of 10% H2O decreased the chlori-
nation level of PCDD/F congeners. This result is in agreement with
 K. Shao et al. / Chemosphere 78 (2010) 672–679 677

100
OCDD
HpCDD
80 HxCDD

PCDD homologue distribution%

PeCDD
TeCDD
60

40

20

0
[AC] [CBA] [CBB] [CBC] [graphite] [AC*]
[AC]+H2O [CBA]+H2O [CBB]+H2O [CBC]+H2O [graphite]+H2O [AC*]+H2O

100
OCDF
HpCDF
80 HxCDF
PCDF homologue distribution%

PeCDF
TeCDF
60

40

20

0
[AC] [CBA] [CBB] [CBC] [graphite] [AC*]
[AC]+H2O [CBA]+H2O [CBB]+H2O [CBC]+H2O [graphite]+H2O [AC*]+H2O

Fig. 3. Percentage of PCDD and PCDF homologue mass distribution of the samples with and without H2O during thermal treatment of model ashes in laboratory scale
experiments. All tests with two or three repeats, mean value is shown (2 g of model ash with 0.19%CuCl2, 3%C, bulk quartz; 300 °C, 10% O2, 10% H2O, balance N2, 30 min).

400
Amount of PCBs in the gas phase and solid sample

PCBs in the gas phase

PCBs in the solid phase

300
(ng g sample)

200
-1

100

0
[AC] [CBA] [CBB] [CBC] [Graphite] [AC*]
[AC]+H2O [CBA]+H2O [CBB]+H2O [CBC]+H2O [Graphite]+H2O [AC*]+H2O

Fig. 4. Effect of H2O on PCBs formation by de novo synthesis. All tests with two or three repeats, mean value with standard deviation is shown (2 g of model ash with
0.19%CuCl2, 3%C, bulk quartz; 300 °C, 10% O2, 10%H2O, balance N2, 30 min).

previous findings. Two reasons can be used to explain this result: of PCDD/F; (2) Converting copper (II) chloride to CuCl2CuO to inhi-
(1) decreasing the C–Cl formation to lower the chlorination level bit the further chlorination of PCDD/Fs.
678
100
80
PCB homologue distribution%
60
40
20
0
[AC]
[AC]+H2O [CBA]+H2O [CBB]+H O [CBC]+H O [Graphite]+H O [AC*]+H O
Fig. 5. Percentage of PCDD and PCDF homologue mass distribution of the samples with and without H2O during thermal treatment of model ashes in laboratory scale
experiments. All tests with two or three repeats, mean value is shown (2 g of model ash with 0.19%CuCl2, 3%C, bulk quartz; 300 °C, 10% O2, 10% H2O, balance N2, 30 min).
The influence of H2O on PCB formation by de novo synthesis is
illustrated in Figs. 4 and 5. With the exception of [AC], the addition
of H2O increases the yields of PCBs formation by de novo synthesis,
by 3.3, 1.5, 2.1, 1.6 and 1.6 times, respectively. The PCBs increase
more than that PCDD/Fs, yet only Te-OCDD/Fs were measured in
our experiments. For model ash [AC], a decrease of 9.5% in Di-
DeCB can be seen in Fig. 5. Addition of H2O also decreases the chlo-
rination level of PCBs.
In de novo synthesis of PCDD/Fs, copper is a key catalyst. Copper
catalyst oxychlorinates the carbon matrix to form dioxin-like
structures. CuCl2 also chlorinates carbon, forming C–Cl bonds,
while Cu2Cl2 transmits the gaseous oxygen to solids forming C–
O; Cu2Cl2 also catalyzes the oxidation of carbon decomposing the
macromolecular carbon and releasing PCDD/Fs. Then PCDD/Fs are
further chlorinated by copper (II) chloride.
According to these experimental results, the mechanism by
which H2O influences upon PCDD/Fs de novo synthesis is as fol-
lows: (1) Promoting the conversion of CuCl2 into CuO inhibits the
C–Cl bond formation as well as further chlorination of PCDD/Fs;
(2) Promoting carbon oxidation catalysis by Cu2Cl2, thus decom-
posing macromolecular carbon, while releasing PCDD/F, PCBs, etc.
4. Conclusion
Model mixtures containing cupric chloride and five different
carbon sources was used to study the effects of H2O on PCDD/Fs
formation by de novo synthesis, i.e. those on chemical speciation
of copper, organic chlorine (C–Cl) formation, catalytic oxidation
of carbon, and on PCDD/Fs and PCBs formation. It was found that
H2O could convert CuCl2 to CuCl2CuO and promote conversion of
CuCl2 to CuO, as confirmed by the decrease of organic chlorine
(C–Cl) formation by H2O. For model ashes where CuCl2 was sup-
ported by carbon ([AC], [CBA], [CBB], [CBC] and [Graphite]), the
addition of H2O promoted catalytic oxidation of this carbon. The
oxidation rate of carbon increased by 14–110% when 10% H2O
was introduced. After stopping H2O addition, the oxidation rate
dropped sharply to the level as in the cases without H2O. This re-
sult suggests that H2O does not react with Cu2Cl2 that catalyzes
carbon oxidation. The addition of H2O decreased the chlorination
level of both PCDD/Fs and PCBs. For model ashes [AC], [CBA] and
[Graphite], addition of H2O decreased the amounts of PCDD/Fs pro-
K. Shao et al. / Chemosphere 78 (2010) 672–679
DeCB
NCB
OCB
HpCB
HxCB
PeCB
TeCB
TrCB
DiCB
[CBA] [CBB] [CBC] [Graphite] [AC*]
2 2 2 2
duced. For other ashes, addition of H2O exhibited the opposite
effect.
According to the experimental results, the mechanism by which
H2O influences upon PCDD/Fs de novo synthesis is as follow: (1)
Promoting the conversion of CuCl2 into CuO inhibits the C–Cl bond
formation as well as further chlorination of PCDD/Fs. (2) Promoting
carbon oxidation catalysis by Cu2Cl2, thus decomposing macromo-
lecular carbon, while releasing the PCDD/Fs, PCBs, etc.
Acknowledgements
The authors acknowledge the support of the National High
Technology Research and Development Key Program of China
(2007AA061302) and (2007AA06Z336), National Natural Science
Foundation of China (50576082), Important Project on Science
and Technology of Zhejiang Province of China (2007C13084), Zhe-
jiang Provincial Natural Science Foundation of China (R107532)
and Project on Science and Technology of Zhejiang Province of Chi-
na (2008C23090). The authors also thank Ms. Mao Lijuan, the tech-
nician of 985-Institute of Agrobiology and Environmental Sciences
of Zhejiang University, for her contribution in carrying through the
analysis of carbon.
References
Addink, R., Drijver, D.J., Olie, K., 1991. Formation of polychlorinated dibenzo-p-
dioxins/dibenzofurans in the carbon/fly ash system. Chemosphere 23 (8–10),
1205–1211.
Briois, C., Ryan, S., Tabor, D., Touati, A., Gullett, B.K., 2007. Formation of
polychlorinated dibenzo-p-dioxins and dibenzofurans from a mixture of
chlorophenols over fly ash: influence of water vapor. Environ. Sci. Technol.
41, 850–856.
Jay, K., Stieglitz, L., 1991. On the mechanism of formation of polychlorinated
aromatic compounds with copper(II) chloride. Chemosphere 22, 987–996.
Lewis, M.A., Masin, J.G., O’Hare, P.A., 2009. Evaluation of alternative
thermochemical cycles. Part I: the methodology. Int. J. Hydrogen Energy 34,
4115–4124.
Li, X.-D., Zhang, J., Yan, J.-H., Chen, T., Lu, S.-Y., Cen, K.-F., 2006. Effect of water on
catalyzed de novo formation of polychlorinated dibenzo-p-dioxins and
polychlorinated dibenzofurans. J. Hazard. Mater. 137, 57–61.
Olie, K., Vermeulen, P.L., Hutzinger, O., 1977. Chlorodibenzo-p-dioxins and
chlorodibenzofurans are trace components of fly ash and flue gas of some
municipal incinerators in the Netherlands. Chemosphere 6 (8), 455–459.
Ross, R.J., Lacombe, D., Naikwadi, K.P., Karasek, F.W., 1990. Investigation of the effect
of water, acids, and bases in the gas stream in the catalytic formation of PCDD
and PCDF over MSW fly ash. Chemosphere 20 (10–12), 1967–1972.
 K. Shao et al. / Chemosphere 78 (2010) 672–679
Sakurai, T., Kobayashi, T., Watanabe, T., 1996. Formation of PCDD/Fs from
chlorophenols (CPs) on fly ash produced by municipal solid waste
incinerators. Organohalogen Compd. 27, 183–186.
Shao, K., Yan, J.-H., Li, X.-D., Lu, S.-Y., Wei, Y.-L., Fu, M.-X., 2009. Inhibition of de novo
synthesis of PCDD/Fs by SO2 in a model system. Chemosphere, accepted for
publication.
Stieglitz, L., Vogg, H., 1987. On formation conditions of PCDD/PCDF in fly ash from
municipal waste incinerators. Chemosphere 16 (8–9), 1917–1922.
Stieglitz, L., Zwick, G., Beck, J., Roth, W., Vogg, H., 1989. On the de-novo synthesis of
PCDD/PCDF on fly ash of municipal waste incinerators. Chemosphere 18 (1–6),
1219–1226.
Stieglitz, L., Zwick, G., Beck, J., Bautz, H., Roth, W., 1990. The role of particulate
carbon in the de-novo synthesis of polychlorinated dibenzodioxins and furans
in fly ash. Chemosphere 20, 1953–1958.
Stieglitz, L., 1998. Selected topics on the de novo synthesis of PCDD/PCDF on fly ash.
Environ. Eng. Sci. 15 (1), 5–18.
Suzuki, K., Kasai, E., Aono, T., Yamazaki, H., Kawamoto, K., 2004. De novo formation
characteristics of dioxins in the dry zone of an iron ore sintering bed.
Chemosphere 54, 97–104.
679
Takaoka, M., Shiono, A., Nishimura, K., Yamamoto, T., Uruga, T., Takeda, N., Tanaka,
T., Oshit, K., Matsumoto, T., Harada, H., 2005. Dynamic change of copper in fly
ash during de novo synthesis of dioxins. Environ. Sci. Technol. 39, 5878–5884.
Weber, P., Dinjus, E., Stieglitz, L., 2001. The role of copper (II) chloride in the
formation of organic chlorine in fly ash. Chemosphere 42, 579–582.
Wilhelm, J., Stieglitz, L., Dinjus, E., 1999. The determination of the role of gaseous
oxygen in PCDD/F formation on fly ash by the use of oxygen-18. Organohalogen
Compd. 41, 83–86.
Wikstrom, E., Ryan, S., Touati, A., Telfer, M., Tabor, D., Gullett, B.K., 2003a.
Importance of chlorine speciation on de novo formation of polychlorinated
dibenzo-p-dioxins and polychlorinated dibenzofurans. Environ. Sci. Technol. 37,
1108–1113.
Wikstrom, E., Ryan, S., Touati, A., Gullett, B.K., 2003b. Key parameters for de novo
formation of polychlorinated dibenzo-p-dioxins and dibenzofurans. Environ.
Sci. Technol. 37, 1962–1970.
Wikstrom, E., Ryan, S., Touati, A., 2004. Origin of carbon in polychlorinated dioxins
and furans formed during soot combustion. Environ. Sci. Technol. 38, 3778–
3784.