Oxychlorination of CO to phosgene in a three-step reaction cycle and corresponding catalytic mechanism

Article abstract of DOI:10.1016/j.jcat.2009.12.008

An improved procedure, three-step reaction cycle procedure, for the continuous preparation of phosgene from CO, air and HCl catalyzed by CuCl₂ was reported for the first time. The corresponding catalytic mechanism of each step was preliminarily disclosed with the powder X-ray diffraction (XRD) analysis: the first step is the oxychlorination of CO to phosgene and simultaneous reduction of CuCl₂ to CuCl; the second step is the oxidation of CuCl with air to Cu₂OCl₂, and the third step is the neutralization of Cu₂OCl₂ with HCl to CuCl₂. The regeneration of catalyst consists of steps 2 and 3, which is called the two-step regeneration of catalyst. The no-simultaneous existence of Cu (I) chloride and water in this three-step reaction procedure prevented effectively copper (I) chloride from the disproportionation. The influence of regeneration conditions, including reaction time, pressure of air or HCl on morphologies and recovery degree of catalyst were investigated and discussed. The degree of recovery for the single-run yield and cumulative yield of phosgene from the two-step regenerated oxychlorination agent can reach, respectively, 87.0% and 97.0% whereas the single-run yield and cumulative yield of phosgene with the one-step regenerated catalyst only can be recovered to 58.8% and 80.5%, respectively. The two-step regeneration method also can result in a higher dispersion of CuCl₂/KCl on silica gel than that of the one-step regeneration. These results not only can offer a quite promising potential for the industrial use, but also can promote our deeply understanding of this important industrial reaction.

Full text of DOI:10.1016/j.jcat.2009.12.008

Journal of Catalysis 270 (2010) 76–85
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Oxychlorination of CO to phosgene in a three-step reaction cycle
and corresponding catalytic mechanism
Tianzhu Zhang a,c, , Carsten Troll , Bernhard Rieger , Juergen Kintrup b,
a
a,*
*
b
b
Oliver F.-K. Schlüter , Rainer Weber
a
Wacker-Lehrstuhl für Makromolekulare Chemie, Technische Universität München, Lichtenbergstraße 4, 85747 Garching, Germany
Bayer Material Science AG, 51373 Leverkusen, Germany
State Key Lab. of Bioelectronics, Jiangsu Key Lab. for Biomaterials and Devices, School of Biological Science and Medical Engineering, Southeast University, Nanjing 210096, China
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
An improved procedure, three-step reaction cycle procedure, for the continuous preparation of phosgene
Received 31 October 2009
Revised 7 December 2009
Accepted 8 December 2009
Available online 6 January 2010
from CO, air and HCl catalyzed by CuCl
mechanism of each step was preliminarily disclosed with the powder X-ray diffraction (XRD) analysis:
the first step is the oxychlorination of CO to phosgene and simultaneous reduction of CuCl to CuCl;
the second step is the oxidation of CuCl with air to Cu OCl , and the third step is the neutralization of
Cu OCl with HCl to CuCl . The regeneration of catalyst consists of steps 2 and 3, which is called the
two-step regeneration of catalyst. The no-simultaneous existence of Cu (I) chloride and water in this
three-step reaction procedure prevented effectively copper (I) chloride from the disproportionation.
The influence of regeneration conditions, including reaction time, pressure of air or HCl on morphologies
and recovery degree of catalyst were investigated and discussed. The degree of recovery for the single-
run yield and cumulative yield of phosgene from the two-step regenerated oxychlorination agent can
reach, respectively, 87.0% and 97.0% whereas the single-run yield and cumulative yield of phosgene with
the one-step regenerated catalyst only can be recovered to 58.8% and 80.5%, respectively. The two-step
2
was reported for the first time. The corresponding catalytic
2
2
2
2
2
2
Keywords:
Oxychlorination
Three-step procedure
Regeneration
Oxidation
Neutralization
Phosgene
2
regeneration method also can result in a higher dispersion of CuCl /KCl on silica gel than that of the
one-step regeneration. These results not only can offer a quite promising potential for the industrial
use, but also can promote our deeply understanding of this important industrial reaction.
Ó 2009 Elsevier Inc. All rights reserved.
1
. Introduction
For example, in the production of bisphenol A polycarbonates,
COCl
gene and recycling of HCl will face a new and huge challenge.
At present, the main Cl production method is still the electrol-
ysis of aqueous solution of sodium chloride, only a small amount of
Cl is produced through the catalytic oxidation of HCl (Deacon
reaction) [4,5]. However, the electrolytic process has still its own
problems. On the one hand, this electrochemical method is rather
electricity-consuming [6,7]; on the other hand, the amount of
caustic soda, which was formed together with chlorine is much lar-
ger than its demand [8–11]. To effectively fulfill the recycle of HCl
in the industrial field is rather pressing and necessary. Therefore, a
large number of researchers are gradually switching their interests
again to the catalytic oxidation of HCl or oxychlorination of CO for
2 2
was degraded to HCl (Eq. (1)). The production of Cl or phos-
Numerous chlorination or phosgenation processes through ele-
mental chlorine (Cl
2
) as a raw material and many metallurgical
2
processes produce a large amount of toxic and corrosive hydrogen
chloride (HCl) stream. Especially, with the rapidly increasing
demand for polyurethanes (PU) and polycarbonates (PC) today
1–3], more and more Cl or phosgene (COCl ) was consumed,
2 2
and consequently more and more by-product HCl was produced.
2
[
O
n HO
O
OH + n Cl
C Cl
bisphenol-A
ð1Þ
O
O
C
O
+ 2n HCl
n
2 2
the production of Cl or COCl out of consideration for economic
polycarbonate
and environmental reasons [12–17]. A lot of effort has been contin-
uously done in finding more efficient methods for recycling HCl by
converting HCl into Cl
Some efficient systems were developed to produce Cl
(Eq. (2)), such as Shell-Chlor process (CuCl –KCl/SiO
2 3 2
Shell Company, Netherlands) [18], MT-Chlor process (Cr O /SiO
2 2
or the relevant chemicals, such as COCl .
*
Corresponding author. Address: Wacker-Lehrstuhl für Makromolekulare Che-
from HCl
2
catalyst)
mie, Technische Universität München, Lichtenbergstraße 4, 85747 Garching,
Germany. Fax: +49 89 28913562.
E-mail addresses: zhangtianzhuglq@yahoo.com.cn (T. Zhang), rieger@tum.de (B.
Rieger).
2
2
(
0
021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.12.008

T. Zhang et al. / Journal of Catalysis 270 (2010) 76–85
77
catalyst) (Mitsui Toatsu Company, Tokyo, Japan) [19,20] and Sumi-
In order to understand the reaction mechanism and improve
further Deacon-type reactions, a lot of studies are correspond-
ingly focused on the characterization of the Cu centers with dif-
ferent techniques, such as powder X-ray diffraction (XRD),
energy-dispersive X-ray (EDX), electron paramagnetic resonance
(EPR), Ultraviolet–visible diffuse reflectance spectroscopy (UV–
Vis DRS), Fourier transform infrared (FT-IR), extended X-ray
absorption fine structure spectroscopy (EXAFS), X-ray absorption
near edge structure spectroscopy (XANES) [25–33]. The detailed
information of catalytic reaction mechanism is still rather little
known [34,35].
tomo process (RuO
Japan) [21].
2
/TiO
2
catalyst) (Sumitomo Chemicals, Tokyo,
HCl þ 1=4O
2
ꢀ 1=2Cl
2
2
þ 1=2H O
ð2Þ
In 1948, Socony-Vacuum Oil Company invented a continuous
process for the production of phosgene from carbon monoxide,
hydrogen chloride and air catalyzed by copper chloride, which in-
cludes two major steps, namely the chlorination step (Eq. (3)) and
the oxidation–neutralization step (Eq. (4)) [22]. One of the major
advantages of this method is that by-product HCl can be directly
recycled to phosgene. This process paved a new way to the effec-
tive recycle of HCl for the production of phosgene. However, be-
cause of problems with the corrosion and catalyst vaporization,
this catalytic process obtained only a limited industrial application.
Obviously, in all of oxychlorination reaction or direct catalytic
oxidation of hydrogen chloride to chlorine, the regeneration of
oxychlorination catalyst is the key step in order to fulfill the re-
cycle of chlorine element. Lamberti et al. reported the experimen-
tal evidences of oxidation of CuCl to oxychloride Cu
neutralization of oxychloride to CuCl in the ethylene oxychlorina-
tion (Eq. (5)) catalyzed by alumina-supported copper chloride at
00 K [36]. Ethylene oxychlorination follows a three-step mecha-
nism: reduction of CuCl by ethylene to CuCl; oxidation of CuCl
by oxygen to oxychloride Cu OCl ; rechlorination of oxychloride
by HCl to CuCl
2 2
OCl and
2
CuCl
CuCl
2
2
þ CO ! COCl
2
þ 2CuCl
! 2CuCl
ð3Þ
ð4Þ
2
2
þ 2HCl þ 1=2O
2
2
2
þ H O
5
In 1973, Rhone-Progil declared an improved process for the pro-
2
duction of phosgene from CO, HCl, air and copper chloride in two
steps: (1) HCl/air (ratio of volume is 0.7) was introduced to the
reactor at 400 °C first to ensure all Cu (I) was converted to Cu
2
2
2
.
(
II); (2) after cleaning with nitrogen CO was fed at 350 °C to pro-
C
2
H
4
þ 1=2O
2
þ 2HCl ! C
2
H
4
Cl þ H
2
O
ð5Þ
duce phosgene. The oxychlorination of CO and the regeneration
of catalyst were carried out in two completely separated stages.
Therefore, phosgene can be formed at lower temperatures, and
the amount of CO and Cl was reduced. The problem with the cor-
2 2
rosion and hydrolysis of phosgene can be overcome. In this pro-
cess, the composition of catalyst was also optimized [23]. In
Opposite to the oxychlorination reaction of ethylene to 1,2-
dichloroethane that has been widely investigated in the last
decades [37–39], very few literature appeared for this analogous
oxychlorination reaction of CO to phosgene although which is
extremely industrially important. The direct experimental evi-
dences of catalytic mechanism are still lacking so far and therefore
seriously prevented us from deeply understanding and further
effectively improving oxychlorination of CO to phosgene.
1
982, Lummus Company improved the first step through using a
mixture of molten cuprous and cupric chlorides to enrich the salt
in the higher valent metal chloride [24].
Fume hood 1
Fume hood 2
Max 12 bar
PM
Max 5 bar
PM
Off-gas
PM
GR 1
GR 2
GR 3
HCl
CO
CV 2
CV 3
BV 3
EBV 1
CV 1
IR
Gasflow from Regeneration
Off-gas
Reactor
TWBV 2
BV 2
Gasflow from
Phosgene-Synthesis
BV 1
CV 4
EBV 2
Reactiongas-
Dilution
Ar 0,3 bar
1
L
PM
GR
CV
= Pressure meter
= Gas regulator
= Check valve
Phosgene Sensor
Monitoring
BV
= Ball valve
EBV
= Electronic ballvalve
TWBV = Three-way-ballvalve
Fig. 1. The schematic view of the experiment set-up for oxychlorination of CO and regeneration of catalyst.

7
8
T. Zhang et al. / Journal of Catalysis 270 (2010) 76–85
In our recent work, the kinetics of oxychlorination of carbon
lytic mechanism of each step with XRD analysis. The regeneration
conditions in the three-step reaction cycle were optimized.
The comparison of two-step and one-step regenerations indi-
cates that the two-step regeneration process has some remark-
able advantages in oxychlorination of CO to phosgene.
Therefore, on the one hand, this three-step procedure offers a
quite promising potential for the industrial use; on the other
hand, the preliminary disclosure of catalytic mechanism will
strongly promote further our understanding of oxychlorination
monoxide was investigated, and composition of catalyst was opti-
mized systematically [40,41]. Here, we reported for the first time
an effectively improved procedure, three-step reaction procedure,
for the continuous preparation of phosgene from CO, air and HCl
catalyzed by copper chloride and probed the corresponding cata-
2
CuCl2
of CO to phosgene catalyzed by CuCl
an improvement to this reaction.
2
and contribute to make
H O
2
CO
2
. Experiment
(3)
(1)
2
HCl
COCl2
2 CuCl
2
.1. Apparatus
A special Cr–Fe alloy autoclave of 45 cm3 (6.4 cm (length) ꢀ
(
2)
3
.0 cm (diameter)) was charged with 0.3361 g catalyst (containing
Cu OCl
2
2
1
/2 O2
2
0.5 mmol CuCl ) enclosed by an electrically heated tape. The prod-
uct effluent was first diluted with l L Ar (0.25 bar) then was re-
leased to a Cyclone™ Gas cell (with base pathlength 12.5 cm,
Scheme 1. Schematic flow chart of three-step reaction cycle procedure.
a
b
001
fresh catalyst
chlorine-depleted
−
2
00
−
2
01
−
1
12
3
12
112
CuCl2
111
−
−
−
220
0
01 020
111 112 122
CuCl
−
021
102
311
200
400
KCuCl3
10
20
30
40
50
60
70
10
20
30
40
50
60
70
2
θ (°)
2
θ (° )
c
d
neutralized by HCl
CuCl2
111
oxidized by Air
Cu OCl
2
22
220
1
31
133
313
2
2
044
KCuCl
3
10
20
30
40
50
60
70
10
20
30
40
50
60
70
2
θ (°)
2
θ (° )
Fig. 2. Comparison of measured XRD patterns of the different kinds of copper species and calculated XRD patterns for crystallites CuCl
fresh catalyst, CuCl and KCuCl ; (b) the chlorine-depleted catalyst and CuCl; (c) the air-oxidized chlorine-depleted catalyst and Cu OCl
depleted catalyst, CuCl and KCuCl
2
, KCuCl
3 2 2
, CuCl and Cu OCl : (a) the
2
3
2
2
; (d) the HCl-neutralized chlorine-
2
3
.

T. Zhang et al. / Journal of Catalysis 270 (2010) 76–85
79
volume 0.19 L) installed on a Perkin Elmer Spectrum100 Series
2.2. Catalyst preparation
spectrometer to analyze the components of the product effluent
online. The reactor off-gas was finally destroyed by a 25% aqueous
solution of sodium hydroxide. The schematic experimental set-up
is illustrated in Fig. 1.
The catalysts were prepared by the incipient-wetness tech-
nique. 2.1524 g (16.0 mmol) CuCl (Acros, 99%) and 0.2404 g
(3.2 mmol) KCl (Merck, 99%) were dissolved in a minimum quan-
2
1
st run of 1st cycle
1.2
1st run of 1st cycle
1.2
0.8
0.4
0.0
CO
CO
0
0
.8
.4
COCl2
COCl2
COCl2
COCl2
CO2
CO2
CO2
HCl
0.0
3
600 3000 2400 1800 1200
600
3
600 3000 2400 1800 1200
600
-1
Wavenumbers (cm )
-1
Wavenumbers (cm )
1
0
.2
.8
1st run of 2nd cycle
1st run of 2nd cycle
1
0
0
0
.2
.8
.4
.0
CO
CO
COCl2
0.4
COCl2
COCl2
COCl2
0
.0
HCl
CO2
3
600 3000 2400 1800 1200
600
-
1
3
600 3000 2400 1800 1200
600
Wavenumbers (cm )
-1
Wavenumbers (cm )
Fig. 5. Comparison of IR absorbance of the one-step regenerated catalyst and the
fresh catalyst.
Fig. 3. Comparison of IR absorbance of the two-step regenerated catalyst and the
fresh catalyst.
fresh
15
1
6
2
8
4
0
12
1
fresh
regenerated
9
6
3
0
regenerated
0
1
2
3
4
5
run
6
7
8
9
0
1
2
3
4
5
Run
6
7
8
9
Fig. 4. Comparison of run curves of the two-step regenerated catalyst and the fresh
Fig. 6. Comparison of run curves of the one-step regenerated catalyst and the fresh
catalyst.
catalyst.

8
0
T. Zhang et al. / Journal of Catalysis 270 (2010) 76–85
tity of distilled water and then added to 8.3650 g silica gel (Grace,
surface area 25 m g , pore volume 0.9 mL g ). The obtained
2.3. EDX and SEM measurements
2
ꢁ1
ꢁ1
paste was first rotary evaporated then dried at 120 °C in an oven
The energy-dispersive X-ray (EDX) and scanning electron
microscopy (SEM) experiments were conducted on a Hitachi TM-
1000 instrument operated at 200 kV or on EDAX instrument from
for 2 h. The composition of the prepared catalyst is 20 wt.% CuCl
2
,
molar ratio of CuCl :KCl is 5:1, which is denoted as 20Cu:4K/SiO
2
2
.
Fig. 7. The morphologies of the fresh catalyst (left: SEM, right: EDX).
Fig. 8. The morphologies of the chlorine-depleted catalyst (300 °C, 5 bar CO, 6 min).

T. Zhang et al. / Journal of Catalysis 270 (2010) 76–85
81
EDAX Inc. According to the principle of EDX, heavy atoms, like Cu
A ¼ 1 þ 42:3X
ð6Þ
and K, have a stronger ability to reflect the X-ray than light atoms
Si, O, C and Cl. Therefore, in SEM pictures, the bright area indicates
the dispersion of metal atoms. According to the shape and size of
the bright area of silica surface, the dispersion information of
where A is the absorption area (at 1827 cmꢁ1) of phosgene, and X is
molar percentage of phosgene (mol%) in the product effluent. The
yield of phosgene was calculated based on the amount of copper
chloride (0.5 mmol) because CO was fed to excess (4.7 mmol) (Eq.
2
CuCl /KCl can be obtained.
(
7)). The yield of each run is added together to obtain the cumula-
tive yield (cy), which indicates the effectively utilized amount of
CuCl in the catalyst. The yield of the first run is considered as the
single-run yield (sy).
2.4. X-ray powder diffraction
2
The copper species of catalyst in the different stages were iden-
tified on a STOE STADI powder diffractometer equipped with a lin-
ear position sensitive detector, using Cu-K radiation (k = 1.5406 Å,
a
2
.6. Regeneration of catalyst with the two-step or one-step procedure
1
0° 6 2h 6 100°, step width = 0.01°) in transmission mode. Micro-
crystalline diamond was used as the internal standard. The pat-
terns were indexed by using the software package WinXPOW
After the catalyst was exhausted, the regeneration reaction was
immediately performed following
procedure:
a two-step or one-step
(
Stoe) and refined with the WinPLOTR program [42]. The crystal-
line structure of the present phases was identified through com-
parison with the powder XRD patterns calculated from crystal
structure data of the American Mineralogist Crystal Structure Data-
base (AMCSD) [43].
Two-step regeneration procedure: First, air (pressure 1–5 bar)
was fed to the autoclave at 300 °C and kept t min. After the oxida-
tion of t min, air was released, and then the autoclave was flushed
with Ar. Then, dry HCl (pressure 1–3 bar) gas was fed and kept t
min. After the neutralization of t min, unreacted HCl was released
to an aqueous solution of sodium hydroxide. The total regeneration
time is sum of oxidation time and neutralization time, t + t min.
o
o
n
n
2
2.5. Reaction of silica-supported CuCl with CO
o
n
To the autoclave charged with 0.3361 g fresh catalyst (20 Cu:4
K/SiO ) (including 0.5 mmol CuCl ) at 300 °C, 5 bar CO (4.7 mmol)
2
2
(
1) Oxidation step (Eq. (8)): temperature 300 °C, air 1–5 bar,
time t min
2) Neutralization step (Eq. (9)): temperature 300 °C, HCl 1–
bar, time t min
was fed. The reaction time is 6 min. Following the earlier men-
tioned reaction conditions, the catalyst was continuously run sev-
o
(
2
eral times until all active CuCl is depleted. The concentration of
3
n
phosgene in the product effluent was calculated according to the
ꢁ1
IR absorption area (at 1827 cm ) following the Eq. (6) [40]:
One-step regeneration procedure (Eq. (4)): First, 3.0 bar HCl was
fed to the autoclave, and then immediately 5.0 bar air was fed.
Fig. 9. The morphologies of the two-step regenerated catalyst. (a) 300 °C, 5 bar air, 6 min and (b) 300 °C, 3 bar HCl, 6 min.

8
2
T. Zhang et al. / Journal of Catalysis 270 (2010) 76–85
Fig. 9 (continued)
The pressure in the autoclave is finally 8.0 bar. After 6 min regen-
eration, unreacted air and HCl were released to the aqueous solu-
tion of sodium hydroxide. This one-step regeneration was carried
out as a control experiment of the two-step regeneration.
3
ble salt KCuCl was formed [44,45]. The peaks at 2h = 15.4°, 26.0°,
30.8°, 38.1°, 48.9° and 50.5° may be assigned to the reflection of
planes of (0 0 1), (2 0ꢁ1), (2 0 0), (1 1ꢁ2), (3 1ꢁ2) and (1 1 2) in
2
crystallite CuCl (space group C2/m), respectively [46]. The peaks
at 2h = 12.1°, 12.8°, 16.4°, 24.2°, 28.4°, 29.1° and 31.2° may be as-
signed to the reflection of planes of (0 0 1), (0 2 0), (0 2 1),
3
. Results and discussion
(
1 1ꢁ1), (1 0ꢁ2), (1 1ꢁ2) and (1 2ꢁ2) in crystallite KCuCl
group P2 /c), respectively [47].
In Fig. 2b, the chlorine-depleted catalyst mainly includes CuCl,
and double salt KCuCl disappeared. The peaks at 2h = 28.6°,
3.1°, 47.5°, 56.4° and 69.5° may be assigned to the reflection of
3
(space
1
The oxychlorination of CO to phosgene in a three-step reaction
cycle was illustrated in Scheme 1.
In Scheme 1, there are three separate steps: the first step is the
oxychlorination of CO to phosgene and simultaneous reduction of
CuCl
air to Cu
Cu OCl with HCl to CuCl
2
depleted catalyst consists of steps 2 and 3, which is called the
two-step regeneration procedure. In the earlier reaction cycle,
the masses containing copper chloride cannot be qualified as cata-
lyst. For convenience, we still call the mass as ‘‘catalyst”.
3
3
planes of (1 1 1), (2 0 0), (2 2 0), (3 1 1) and (4 0 0) in crystallite
CuCl (space group F-43 m (2 1 6)), respectively [48].
2
to CuCl (Eq. (7)); the second one is the oxidation of CuCl with
OCl (Eq. (8)), and the third one is the neutralization of
(Eq. (9)). The regeneration of chlorine-
2
2
In Fig. 2c, CuCl was oxidized to Cu
h = 17.6°, 30.2°, 32.0°, 35.5°, 41.7°, 46.7° and 53.9° may be as-
signed to the reflection of planes of (1 1 1), (2 2 0), (1 3 1), (2 2 2),
2 2
OCl by air. The peaks at
2
2
2
(
(
1 3 3), (3 1 3) and (0 4 4) in crystallite Cu
7 0)), respectively [49].
2
OCl
2
(space group Fddd
In Fig. 2d, Cu
2
OCl
2
was neutralized to CuCl
2
by HCl, and KCuCl
3
was formed again; The XRD patterns were recovered to the original
one similar to patterns shown in Fig. 2a.
Although the relative intensity of peaks is difficult to calculate
accurately due to the strong background from the amorphous silica
gel, on the whole, the measured XRD patterns are in excellent
2
CuCl
CuCl þ 1=2O
OCl
þ 2HCl ¼ 2CuCl
2
þ CO ¼ COCl
2
þ 2CuCl
ð7Þ
ð8Þ
ð9Þ
2
2
¼ Cu
2
OCl
2
Cu
2
2
2
þ H
2
O
The earlier proposed mechanism of each step in Scheme 1 was
agreement with the calculated patterns for crystallites CuCl
2
,
preliminarily confirmed by XRD measurements (Fig. 2).
Fig. 2 shows the comparison of the measured XRD patterns of
copper species in the different stages and the calculated XRD pat-
terns for crystallites CuCl
In Fig. 2a, two phases were found: one is from CuCl
other is from KCuCl in the fresh catalyst. In order to improve
the activity of CuCl , the promoter KCl was added, therefore, dou-
KCuCl , CuCl and Cu OCl . It is very necessary to point out here that
3
2
2
the earlier XRD measurements were not carried out under reaction
conditions online due to the limited experimental set-up, therefore,
potential hydration/restructuring of substance possibly occurred.
In order to avoid hydration/restructuring and obtain a real-time
state as well as non-crystalline phases of materials, the combined
2 3 2 2
, KCuCl , CuCl and Cu OCl .
2
, and an-
3
2

T. Zhang et al. / Journal of Catalysis 270 (2010) 76–85
83
XRD/XAFS methods will be employed in the further investigation.
Therefore, the proposed mechanism of each step in three-step reac-
tion cycle was proved preliminarily by XRD analysis here.
and regenerated catalyst almost overlap completely. The degree
of recovery for the cumulative yield is 97.0% (Fig. 4).
However, for the result from the one-step regeneration proce-
dure, the degree of recovery for the single-run yield is only 58.8%
(
8
Fig. 5), the degree of recovery for the cumulative yield is about
0.5% (Fig. 6).
The reasons why the degree of recovery for the single-run yield
3
.1. Two-step regeneration procedure versus one-step regeneration
procedure
or the cumulative yield in the three-step procedure is remarkably
higher than that in the one-step regeneration procedure can be
summarized as follows.
Based on the analysis of the thermodynamical data and the
mechanism of the reaction, two main drawbacks can be overcome
effectively using this three-step procedure, and therefore, the de-
gree of recovery of catalyst was greatly improved.
3
.1.1. The degree of recovery of the catalyst
After the regeneration, the degree of recovery of the catalyst
from the two-step and one-step procedure was compared. Here,
two-step regeneration conditions are (1) 300 °C, 5 bar air, 6 min;
(
2) 300 °C, 3 bar HCl, 6 min.
Degree of recovery was calculated as follows:
sy
r
Degree of recovery for the single-run yield (sy): sy ꢀ 100%
f
cy
r
Degree of recovery for the cumulative yield (cy): cyf ꢀ 100%
3.2.1. Prevention of the disproportionation of Cu (I)
sy and cy : the single-run yield and the cumulative yield for the
In the one-step regeneration, oxidation and neutralization al-
most simultaneously proceed, the formed water will contribute
to the disproportionation of Cu (I) to Cu (II) and Cu (0) (Eq. (10)).
According to the redox potentials, the equilibrium constant is
f
f
fresh catalyst;
sy and cy : the single-run yield and cumulative yield for the
r
r
regenerated catalyst.
6
K = 1.3 ꢀ 10 at 298 K. So the reaction is highly favorable [50]. Un-
der this regeneration conditions, the oxidation of Cu (0) to Cu (II)
will not take place easily because the oxide film was formed and
the reaction time is too short.
3
.2. The degree of recovery of the single-run yield or cumulative yield
from two- or one-step regeneration
ꢂ
m
2
Cu ðaqÞ ! Cu2þðaqÞ þ CuðsÞ
2
D
r
H
¼ ꢁ42:1 kJ=mol
ð10Þ
The Fig. 3 shows the IR spectra comparison of results from the
fresh catalyst and two-step regenerated catalyst. For the result
from two-step regeneration, the degree of recovery for the sin-
gle-run yield is 87.0%. The two run curves of the fresh catalyst
However, when the oxidation and neutralization is carried out
in two separate steps, the disturbance of water can be prevented
effectively. In the oxidation step, according to the thermodynami-
Fig. 10. The morphologies of the one-step regenerated catalyst.

8
4
T. Zhang et al. / Journal of Catalysis 270 (2010) 76–85
cal calculation, the disproportionation of Cu (I) at the absence of
water is very difficult (Eq. (11)). It is an endothermal reaction with
1
00
recovery of sy (%)
recovery of cy (%)
0
ꢁ1
an enthalpy
D
r
H
of 872.1 kJ mol . This should lead to a more de-
m
fined reaction pathway where Cu (I) is completely oxidized to Cu
9
8
7
0
0
0
(
II) and then converted to CuCl
2
in the neutralization.
ꢂ
m
þ
2
Cu ðgÞ ! Cu2þðgÞ þ CuðsÞ
D
r
H
¼ 872:1 kJ=mol
ð11Þ
3
.2.2. Prevention of the undesirable Deacon reaction
Another problem for the one-step regeneration is that in a Dea-
con reaction chlorine can be produced at 650–750 K. In the pres-
ence of CuCl , HCl can be oxidized to Cl with oxygen, which also
2
2
decreases the efficiency of the regeneration to some extent [51].
In the two-step regeneration procedure, air and HCl cannot meet
in the reactor; therefore, the Deacon reaction here can be
suppressed.
60
1
2
3
4
5
PAir (bar)
3.3. The influence of the regeneration on the morphologies of catalyst
Fig. 12. The influence of pressure of air on the degree of recovery (pressure of HCl is
bar).
3
The dispersion of CuCl /KCl on the surface of silica in the fresh
2
catalyst is not homogeneous due to the interface tension [52],
namely, the silica surface cannot be uniformly covered by Cu–chlo-
recovery of sy (%)
recovery of cy (%)
1
00
90
ride species. Some lumps of CuCl
the surface of silica (Fig. 7). During the reaction, CuCl
2
/KCl crystal were accumulated on
/CuCl/KCl is
2
in a molten state. After the reaction was finished and cooled to r.t.,
CuCl/KCl solidified in the different shapes (Fig. 8). However, after
the oxidation, these CuCl/KCl forms were broken into smaller
pieces (Fig. 9a). The further neutralization leads to numerous short
needle-like crystals on the surface of silica (Fig. 9b). Therefore,
8
0
after the two-step regeneration, the dispersion of CuCl
2
/KCl on sil-
70
ica was slightly improved. For the one-step regeneration, a lot of
thicker and curly extrudates were formed that probably result
from extrusion through silica pores (Fig. 10).
6
5
0
0
3.4. The influence of regeneration time (t
o n
:t = 1:1) on the degree of
1
.0
1.5
2.0
2.5
3.0
recovery
PHCl(bar)
In this experiment, the ratio of oxidation t
was kept 1:1, the total regeneration time t (=t
min to 20 min with a interval of 2 min. Fig. 11 shows the influ-
o
to neutralization t
+ t ) increase from
n
n
Fig. 13. The influence of pressure of HCl on the degree of recovery (pressure of air is
bar).
o
5
1
ence of regeneration time on the degree of recovery of the sin-
gle-run yield or cumulative yield. When the regeneration time is
very short, say, 1 min, the degree of recovery for the single-run
yield and cumulative yield is only 74.7% and 77.5%, respectively.
When the regeneration was increased to 4–16 min, the degree of
recovery of the cumulative yield is about 96.0%. If the regeneration
time is too long, say, 20 min, the recovery is inversely decreased.
While the recovery of the single-run yield always fluctuates
around about 75.0% with the regeneration time. Therefore, the
regeneration time of 2–6 min is better.
for cumulative yield
3.5. The influence of pressure of air or HCl on the degree of recovery
1
00
In order to fully convert Cu (I) chloride to Cu (II) chloride, the
excess of air and HCl is necessary. The recovery of the single-run
yield and cumulative yield approximately increases linearly with
the pressure of air (pressure of HCl was 5 bar) (Fig. 12). The similar
case can be found when the pressure of HCl increases from 1 bar to
8
6
4
0
0
0
3
bar (pressure of air was 5 bar) (Fig. 13). The reason is understand-
able. Since the dispersion of CuCl /KCl is heterogeneous, with the
increasing pressure of either air or HCl, air or HCl can penetrate in-
side the crystal of CuCl /KCl, therefore more complete oxidation
and neutralization can occur.
2
for single-run yield
2
0
4
8
12
16
20
24
4
. Conclusions
Regeneration time (min)
In this work, for the first time, a three-step reaction procedure
was proposed to produce continuously phosgene from HCl, air
Fig. 11. The influence of CO pressure on the yield of phosgene with the reaction
time of 10 min at 345 °C.

T. Zhang et al. / Journal of Catalysis 270 (2010) 76–85
85
and CO catalyzed by CuCl
2
where the regeneration of catalyst con-
[15] H.-U. Dummersdorf, F. Gestermann, H. Härle, F.-R. Minz, H. Waldmann, H.
Judat, Z. Kricsfalussy, G. Wiechers, DE Patent 4440642 A1, to Bayer AG, 1996.
sists of the oxidation and neutralization. The two-step regenera-
tion conditions were optimized. Compared with the one-step
regeneration procedure, the two-step regeneration can effectively
prevent the disproportionation of Cu (I) and a type of Deacon reac-
[
[
16] H.W. Schussler, US Patent 6174834 B1, to PPG Industrial Ohio Inc., 2001.
17] H.E. Tuinstra, H.D. Myers, US Patent 5336750, to The Dow Chemical Company,
1994.
[
[
18] E.G. Foster, US Patent 4394367, to Shell Oil Co., 1983
19] K.K. Tadamitsu, K.Y. Yasuo, N.Y. Tokio, K.K. Kamakura, US Patent 4822589, to
Mitsui Toatsu Chemicals, Inc. (Tokyo, JP), 1989.
2
tion. At the same time, the dispersion of CuCl /KCl on the silica sur-
face from the two-step regeneration is better than that from one-
step regeneration. Finally, the degree of recovery of the single-
run yield and cumulative yield for the two-step regeneration can
reach up to 87.0% and 97.0%, respectively when the regeneration
conditions are 5 bar air (6 min) and 3 bar HCl (6 min) at 300 °C.
Therefore, this three-step reaction cycle provides a potential ap-
proach to the industrial production of phosgene. Furthermore,
the preliminary disclosure of reaction mechanism by XRD mea-
surements can help us effectively to deepen the understanding of
this important oxychlorination of carbon monoxide to phosgene.
Without doubt, these results are quite interesting for both indus-
trial field and academic research.
[20] K. Miyata, J. Morisaki, T. Hirayama, H. Kamachi, K. Yamada, US Patent 5707919,
to Mitsui Toatsu Chemicals, Inc. (Tokyo, JP), 1998.
[
[
21] K. Iwamoto, WO/2005/075346, to Sumitomo Chemicals, 2005.
22] E. Gorin, C.B. Miles, US Patent 2444289, to Socony-Vacuum Oil Company, 1948.
[23] F. Wuesthoff, E. Frhr. von. Pechmann, D. Behrens, R. Goetz, DE Patent 2232538,
to Rhone-Progil, 1973.
[
[
24] R. Herbert, US Patent 4346047, to The Lummus Company, 1982.
25] N. López, J. Gómez-Segura, R.P. Marín, J. Pérez-Ramírez, J. Catal. 255 (2008) 29.
[26] R.L. Dotson, J. Catal. 33 (1974) 210.
[
27] C. Zipelli, J.C.B. Bart, G. Petrini, S. Galvagno, C. Cimino, Z. Anorg. Allg. Chem. 502
1983) 199.
28] P.S. Sai Prasad, P. Kanta Rao, J. Chem. Soc. Chem. Commun. (1987) 951.
(
[
[29] E. Finocchio, N. Rossi, G. Busca, M. Padovan, G. Leofanti, B. Cremaschi, A.
Marsella, D. Carmello, J. Catal. 179 (1998) 606.
[
30] G. Leofanti, M. Padovan, M. Garilli, D. Carmello, G.L. Marra, A. Zecchina, G.
Spoto, S. Bordiga, C. Lamberti, J. Catal. 189 (2000) 105.
[
31] G. Leofanti, M. Padovan, M. Garilli, D. Carmello, A. Zecchina, G. Spoto, S.
Bordiga, G. Turnes Palomino, C. Lamberti, J. Catal. 189 (2000) 91.
32] E.M. Fortini, C.L. García, D.E. Resasco, J. Catal. 99 (1986) 12.
33] C. Prestipino, S. Bordiga, C. Lamberti, S. Vidotto, M. Garilli, B. Cremaschi, A.
Marsella, G. Leofanti, P. Fisicaro, G. Spoto, A. Zecchina, J. Phys. Chem. B 107
(2003) 5022.
Acknowledgments
[
[
The authors thank the financial support from Bayer Material
Science AG and T. Zhang thanks Dr. Peter Reuter and Dr. Christian
Hanisch for EDX measurements at Universität Ulm, Dr. Florian
Kraus and Dr. Bernhard Wahl for XRD measurements, Dr. Tobias
Förstern for CO adsorption experiment at TU München.
[
[
[
34] M.W.M. Hisham, S.W. Benson, J. Phys. Chem. 99 (1995) 6194.
35] D.P. Ashton, T.A. Ryan, Appl. Catal. 12 (1984) 263.
36] G. Leofanti, A. Marsella, B. Cremaschi, M. Garilli, A. Zecchina, G. Spoto, S.
Bordiga, P. Fisicaro, C. Prestipino, F. Villain, C. Lamberti, J. Catal. 205 (2002)
3
75.
37] P.S. Sai Prasad, K.B.S. Prasad, P. Kenta Rao, V.K. Kaushik, J. Mater. Sci. 32 (1997)
479.
[
[
References
1
38] J.S. Naworski, E.S. Evil, in: B.E. Leach (Ed.), Applied Industrial Catalysis, vol. 1,
Academic Press, New York, 1983, p. 239.
[
1] T.A. Ryan, C. Ryan, E.A. Seddon, K.R. Seddon, Phosgene and Related Carbonyl
Halides, Elsevier, Amsterdam, 1996.
[39] M.N. Newmann, Encyclopedia of Polymer Science and Engineering, vol. 17,
Wiley, New York, 1985 (and references therein).
[40] T. Zhang, C. Troll, B. Rieger, J. Kintrup, O. F.-K. Schlüter, R. Weber, Appl. Catal. A:
Gen. 357 (2009) 51.
[41] T. Zhang, C. Troll, B. Rieger, J. Kintrup, O.F.-K. Schlüter, R. Weber, Appl. Catal. A:
Gen. 365 (2009) 20.
[42] T. Roisnel, J. Rodriguez-Carvajal, FULLPROF, Version 3.20, France, 2005.
[43] R.T. Downs, M. Hall-Wallace, Am. Mineral. 88 (2003) 247.
[44] C.L. García, D.E. Resasco, J. Catal. 122 (1990) 151.
[45] M.P. Vorobei, O.V. Skiba, Russ. J. Inorg. Chem. 15 (1970) 725.
[46] P.C. Burns, F.C. Hawthorne, Am. Mineral. 78 (1993) 187.
[47] A. Oosawa, T. Takamasu, K. Tatani, H. Abe, N. Tsujii, O. Suzuki, H. Tanaka, G.
Kido, K. Kindo, Phys. Rev. B 66 (2002) 104405.
[48] R.W.G. Wyckoff, E. Posnjak, J. Am. Chem. Soc. 44 (1922) 30.
[49] S.V. Krivovichev, S.K. Filatov, P.C. Burns, Can. Mineral. 40 (2002) 1185.
[50] P. Atkins, T. Overton, J. Rourke, M. Weller, F. Armstrong, Inorganic Chemistry,
fourth ed., Oxford University Press, Oxford, 2006. p. 151.
[51] W.V. Bauer, Fr 1526958, to The Lummus Company, 1966.
[52] J.M. Gallardo Amores, V.S. Escribano, G. Busca, V. Lorenzelli, J. Mater. Chem. 4
(1994) 965.
[
[
[
2] M. Braddock, Chem. Eng. News, vol. 83, 2005, p. 67.
3] ICIS Chemical Business source, vols. 15–21, 2008, pp. 16 (September).
4] S. Motupally, D.T. Mah, F.J. Freire, J.W. Weidner, Electrochem. Soc. Interf. 7 (3)
(
1998) 32.
[
[
5] H. Deacon, US Patent 016802, 1875.
6] F.J. Freire, K.B. Keating, E.K. Sakata, US Patent 5891319, to E.I. du Pont de
Nemours and Company, 1999.
[
[
7] F. Gestermann, J. Dobbers, H.-N. Rindefleisch, WO97/19205, to Bayer AG, 1997.
8] H.Y. Pan, R.G. Minet, S.W. Benson, T.T. Tsotsis, Ind. Eng. Chem. Res. 33 (1994)
2
996.
9] P.-T. Chang, G.-S. Hu, M.-H. Han, Q. Wu, F. Wei, Y. Jin, Chinese J. Process. Eng. 6
2006) 47.
[
(
[
[
10] Y.-L. Wu, F. Wei, M.-H. Han, Y. Jin, Chinese J. Process. Eng. 4 (2004) 269.
11] L.-W. Wang, M.-H. Han, Y.-L. Wu, F. Wei, Y. Jin, Chinese J. Process. Eng. 3 (2003)
3
40.
12] M. Mortensen, R.G. Minet, T.T. Tsotsis, S. Benson, Chem. Eng. Sci. 51 (1996)
031.
13] F.J. Freier, K.B. Keating, E.K. Sakata, J.A. Trainham, C.G.J. Law, J.S. Newman, D.J.
Eames, WO97/24473, to E.I. du Pont de Nemours and Company, 1997.
14] H.R. Maplewood, V.A. Strangio, G. Ridge, US Patent 4119705, to The Lummus
Company, 1978.
[
[
[
2