Formation of glyoxal by oxidative dehydrogenation of ethylene glycol

Article abstract of DOI:10.1246/bcsj.75.375

Iron(III) phosphates doped with a very small amount of molybdenum(VI) were found to be effective as catalysts for a vapor-phase oxidative dehydrogenation of ethylene glycol to glyoxal. The effects of the molybdenum(VI) content and of the reaction variable

Full text of DOI:10.1246/bcsj.75.375

c. Jpn.
© 2002 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 75, 375–381 (2002)
375
e Chemical
ciety of Ja-
n
e Chemical
ciety of Ja-
n
M
Formation of Glyoxal by Oxidative Dehydrogenation of Ethylene Glycol
Glyoxal from Ethylene Glycol
M. Ai
Mamoru Ai
Department of Applied Chemistry and Biotechnology, Niigata Institute of Technology,
1719 Fujihashi, Kashiwazaki 945-1195
02
5
1
(Received June 13, 2001)
SJA8
09-2673
197
Iron(Ⅲ) phosphates doped with a very small amount of molybdenum(Ⅵ) were found to be effective as catalysts for
a vapor-phase oxidative dehydrogenation of ethylene glycol to glyoxal. The effects of the molybdenum(Ⅵ) content and
of the reaction variables on the conversion of ethylene glycol and selectivity to glyoxal were studied. The optimum Mo/
Fe atomic ratio was in the range from 0.002 to 0.01. The highest selectivity to glyoxal was obtained in the temperature
range of 290 to 310 °C. The rate of ethylene glycol consumption increased almost in proportion to the oxygen concen-
tration, while the selectivity to glyoxal increased only slightly as the oxygen concentration was increased. The selectivi-
ty to glyoxal increased markedly as the concentration of water was increased. It reached 37 and 32 mol% at ethylene
glycol conversions of 60 and 80%, respectively. The one-pass yield of glyoxal reached 26.5 mol%.
01
01
.2
Iron(Ⅲ) phosphates have been claimed to be effective as cat-
alysts for oxidative dehydrogenation of isobutyric acid to form
methacrylic acid.1 There have been a number of scientific
studies on iron(Ⅲ) phosphate as catalysts for production of
methacrylic acid.2–13
These findings led us to examine the catalytic performances
for a more difficult reaction, that is, oxidative dehydrogenation
of ethylene glycol to glyoxal (ethandial).
HOCH2–CH2OH + O2 → OHC–CHO + 2H2O
CH3–CH(CH3)–COOH + 0.5 O2 →
CH2wC(CH3)–COOH + H2O
Glyoxal is a raw material of various chemicals. It is gener-
ally produced by a vapor phase oxidation of ethylene glycol
using Ag or Cu catalysts at a high temperature of above 600 °C
in a manner similar to the oxidative dehydrogenation of metha-
nol to formaldehyde. An academic study has been reported by
Gallezot et al.20 It should be noted that formaldehyde is pro-
duced also by another oxidative dehydrogenation of methanol
using mixed metal-oxide catalysts such as MoO3–Fe2O3 at a
low temperature of around 350 °C, while glyoxal is not pro-
duced in the oxidative dehydrogenation of ethylene glycol over
such mixed metal-oxide catalysts. Recently, Chumbhable and
Awasarkar reported that the yield of glyoxal in the oxidation of
ethylene glycol is enhanced by an addition of a small amount
(2.5 wt%) of phosphorus into Fe(MoO4)3.21
Recently, it was found that iron(Ⅲ) phosphate catalysts
show a uniquely high selectivity for several other oxidative de-
hydrogenation reactions, such as formation of pyruvic acid
from lactic acid,14 glyoxylic acid from glycolic acid,15 and
pyruvaldehyde (2-oxopropanal) from hydroxyacetone (ace-
tol).16
CH3–CH(OH)–COOH + 0.5O2 → CH3–CO–COOH + H2O
HOCH2–COOH + 0.5O2 → OCH–COOH + H2O
CH3–CO–CH2OH + 0.5O2 → CH3–CO–CHO + H2O
Iron(Ⅲ) phosphate possesses both acidic and redox func-
tions, as do phosphates of vanadium(Ⅴ) and molybdenum(Ⅵ).
However, it is interesting to note that iron(Ⅲ) phosphate cata-
lysts are effective only for the oxidative dehydrogenation, but
not for oxygen insertion reaction, unlike phosphates of vanadi-
um(Ⅴ) and molybdenum(Ⅵ).15,17,18 It is believed that C–C
bond fission, which is one of the main side-reactions in an oxi-
dative dehydrogenation, is promoted by the oxygen insertion.
This may be the reason why iron(Ⅲ) phosphate shows a
uniquely high selectivity for several oxidative dehydrogenation
reactions.
More recently, it was also found that both the catalytic activ-
ity and the selectivity of iron(Ⅲ) phosphate in the oxidative de-
hydrogenation of lactic acid to pyruvic acid are improved dra-
matically by a doping of very small amounts of molybde-
num(Ⅵ) compounds.19
Experimental
1. Catalysts. The original iron(Ⅲ) phosphate catalyst sample
with a P/Fe atomic ratio of 1.0 was prepared according to proce-
dures similar to those described previously.11 Iron(Ⅲ) nitrate
[Fe(NO3)3•9H2O] (122 g, 0.3 mol) was dissolved in about 5 dm3
of water, and a dilute ammonia solution was added to precipitate
iron(Ⅲ) hydroxide gel. Most of the water was removed by decan-
tation; then the obtained precipitate was mixed with a 41.5 g por-
tion of 85% orthophosphoric acid [0.36 mol of H3PO4]. Next, the
mixture was slowly boiled for 3 to 5 h, yielding a slightly brown-
ish-white precipitate. The precipitate was filtered and washed by
water to remove excess phosphoric acid. The obtained cake was
dried in an oven at 120 °C for 6 h. The resulting fine powder of
precursor was pressed under a pressure of 200 Kg/cm2 to make
tablets with a diameter of 20 mm. The tablets were then broken

376 Bull. Chem. Soc. Jpn., 75, No. 2 (2002)
Glyoxal from Ethylene Glycol
Table 1. Comparison of Performances of Catalysts
Catalyst
T
Conv
Yield/mol%
SG total
(atomic ratio)
Fe–Mo
(3–7)
°C
%
GX
4.9
5.5
6.4
GALD GXAD HCHO HCOOH
AcH
2.6
2.0
COx others
mol%
11.3
8.7
260
270
280
64.6
85.4
91.9
2.0
1.5
1.3
0.4
0.5
0.0
25.9
47.6
61.8
0.0
0.0
0.0
6.7
4.0
4.3
22.0
24.5
14.9
3.2
8.4
240
250
260
270
28.6
48.7
63.3
76.3
4.0
6.7
9.1
1.5
2.0
2.0
2.5
0.1
0.5
0.8
1.0
6.3
10.4
15.0
21.8
4.9
5.1
5.9
6.1
4.2
4.8
7.1
8.0
2.1
2.0
3.2
6.0
5.4
17.1
20.2
18.9
19.8
19.1
18.8
19.9
H3PMo12O40
11.7
260
270
280
290
34.9
51.6
77.3
92.4
2.5
3.1
4.6
5.3
1.4
1.6
1.7
1.4
0.2
0.4
0.5
0.6
3.1
6.5
10.5
14.5
2.8
0.0
0.0
0.0
15.24
26.4
43.0
60.4
1.2
3.0
4.3
7.3
8.3
10.7
12.5
2.4
11.9
9.8
9.0
7.9
V–P
(1–1.06)
260
270
280
290
47.0
54.3
64.0
78.2
14.2
18.0
21.0
22.8
1.7
2.5
2.3
1.9
0.9
1.3
1.5
1.8
7.2
10.8
17.1
24.0
2.2
2.3
0.0
0.0
3.4
4.6
6.4
8.5
2.6
3.2
3.8
5.4
6.7
11.7
11.9
13.8
35.7
40.0
38.9
33.8
Fe–P–Mo
(1–1–0.01)
Amount of Catalyst used: 5.0 g; Feed rates: ethylene glycol, oxygen, water, nitrogen = 14.5, 25.0, 954, 500 mmol/h. EG = ethyl-
ene glycol, GX = glyoxal, GALD = glycolaldehyde, GXAD = glyoxylic acid, AcH = acetaldehyde. SG total: selectivity to the sum
of GX, GALD, and GXAD.
up and sieved to a 7–20 mesh size. Finally, the precursor was cal-
cined in a stream of air at 400 °C for 8 h. The product consisted of
tridymite-type FePO4.22 The presence of quartz-type FePO423 was
not detected in the XRD patterns. The BET surface area was 8.9
m2/g and the bulk density was 0.66 g/mL.
The conversion of ethylene glycol was defined as: 100 × [1 −
(moles of unreacted ethylene glycol)/(moles of ethylene glycol
fed)]. The yields of C2 products were defined as: 100 × (moles of
product)/(moles of ethylene glycol fed). The yields of C1 products
were defined as; 50 × (moles of product)/(moles of ethylene gly-
col fed). The selectivity was defined as: 100 × (yield)/(conver-
sion). The selectivity to the sum of glyoxal, glycolaldehyde, and
glyoxylic acid (abbreviated as SG total) was defined as: 100 × (sum
of yields of glyoxal, glycolaldehyde, and glyoxylic acid)/(conver-
sion).
Iron(Ⅲ) phosphate samples doped with molybdenum(Ⅵ) were
prepared according to the following procedures. A required
amount
of
molybdenum(Ⅵ)
compounds,
such
as
(NH4)6Mo7O24•4H2O or H3PMo12O40, was dissolved in water. Un-
less indicated othwerwise, (NH4)6Mo7O24•4H2O was used as the
source of molybdenum. The solution was added into the above-
mentioned fine powder of iron(Ⅲ) phosphate precursor. Excess
water was evaporated by means of a hot air current; the obtained
paste-like compound was dried in an oven at 120 °C for 6 h. The
other procedures were the same as those of the original iron phos-
phate.
2. Reaction Procedures. The reaction was carried out with a
continuous-flow-system at atmospheric pressure. The reactor was
made of a stainless steel tube, 50 cm long and 1.8 cm i.d., mount-
ed vertically and immersed in a lead bath. The catalyst sample (1
to 20 g) was placed near the bottom of the reactor and porcelain
cylinders, 3 mm long, and 1.5 mm i.d./3.0 mm o.d., were placed
both under and above the catalyst sample. The reaction tempera-
ture was in the range of 230 to 340 °C. A mixture of nitrogen and
oxygen was fed in from the top of the reactor; the feed rate of ni-
trogen was fixed at 200 mL/min. An aqueous solution containing
100 g (1.60 mol) of ethylene glycol per liter was introduced into
the preheating section of the reactor with a feed rate of either 9.04
or 18.08 mL/h by a syringe pump and a fine stainless steel tube.
Unlss indicated otherwise, the feed rates of ethylene glycol, oxy-
gen, water, and nitrogen were 14.5, 15.0, 452, and 500 mmol/h,
respectively. The effluent gas from the bottom of the reactor was
led successively into four chilled scrubbers to recover the water-
soluble compounds. The reaction products and the unreacted eth-
ylene glycol were analyzed by both GC and LC.
The contact time (s) was defined as: (volume of catalyst used)/
(volume of reactant gas mixture per second).
Results
1. Comparison of the Performances of Various Catalysts.
The performances of the following four catalysts in the oxida-
tion of ethylene glycol were compared: (1) iron(Ⅲ) molyb-
deate with a Fe/Mo atomic ratio of 3/7; (2) H3PMo12O40 sup-
ported (10 wt%) on natural pumice; (3) vanadium(Ⅴ) phos-
phate with a P/V atomic ratio of 1.06, consisting of (VO)2P2O7;
(4) iron(Ⅲ) phosphate doped with molybdenum(Ⅵ), with a
Mo/Fe atomic ratio of 0.01 (abbreviated as Fe–P–Mo = 1–1–
0.01). The amount of catalyst used was kept constant at 5.0 g.
The main products were glyoxal, glycolaldehyde (CH2OH–
CHO), glyoxylic acid (HOOC–CHO), acetaldehyde, formalde-
hyde, and carbon oxides. Glycolic acid (CH2OH–COOH) was
not detected in the products. Formic acid was found in some
cases. In the case of the Fe–P–Mo = 1–1–0.01 catalyst, a
small amount (yield of less than 3 mol%) of formic acid was
found, only when the extent of reaction was low. The discrep-
ancy between the amount of consumed ethylene glycol and the
sum of the assigned products corresponds to the amount of un-
identified compounds; it was listed as “others”. The selectivity
to the sum of glyoxal, glycolaldehyde, and glyoxylic acid (ab-

M. Ai
Bull. Chem. Soc. Jpn., 75, No. 2 (2002) 377
breviated as SG total) is important, because these three com-
pounds are not the degradation products. The results are listed
in Table 1.
Activity. Effects of the content of molybdenum(Ⅵ) doped
into iron(Ⅲ) phosphate on the oxidation activity were studied.
The reaction was performed using 5 g-portions of catalysts
with different Mo/Fe atomic ratios (contact time of about 1.2
s). The catalysts with a Mo/Fe atomic ratio of more than 0.2
were prepared using H3PMo12O40 as the source of molybde-
num(Ⅵ). Some representative results obtained from each cata-
lyst are listed in Table 2.
The conversions of ethylene glycol at a fixed temperature of
260 °C are plotted as a function of the content of molybde-
num(Ⅵ) (Mo/Fe atomic ratio) in Fig. 1 (open circles). The
catalytic activity increases markedly by the doping of a very
small amount of molybdenum(Ⅵ) (Mo/Fe atomic ratio of
0.001). Therefore, the conversions obtained from the samples
with a Mo/Fe atomic ratio lower than 0.04 were also studied in
more detail. The results are shown in Fig. 2 (open circles).
The results can be summarized as follows. The oxidation
activity increases sharply as the content of molybdenum(Ⅵ)
increases up to a Mo/Fe atomic ratio of around 0.01, it increas-
The results may be summarized as follows:
(1) The Fe–P–Mo (1–1–0.01) catalyst shows clearly a yield
of glyoxal higher than those obtained from the other three cat-
alysts. The yield reaches 22.8 mol% and the selectivity to the
sum of glyoxal, glycolaldehyde, and glyoxylic acid reaches 40
mol% at a conversion of 54%.
(2) The main by-products over the Fe–P–Mo catalyst are
formaldehyde and unidentified compounds.
(3) Oxidation activity of the Fe–P–Mo catalyst is of the
same order of magnitude as that of the other three catalysts.
(4) The main product over the V–P oxide catalyst is acetal-
dehyde.
(5) The main products over the Fe–Mo oxide and
H3PMo12O40 catalysts are formaldehyde and unidentified com-
pounds.
2. Effects of the Molybdenum Content on the Oxidation
Table 2. Effects of the Molybdenum Content on the Catalytic Propertirs
Catalyst
T
Conv
Yield/mol%
HCHO HCOOH
SG total
Conv
°C
%
GX
1.7
2.5
3.5
GALD
1.0
1.3
GXAD
0.4
0.5
AcH
4.4
10.6
21.5
COx
5.8
8.3
others
4.8
2.0
mol%
14.9
14.3
11.0
O2%
10.5
15.2
24.2
Fe–P–Mo
1–1–0.000
(10 g)
260
280
300
20.3
29.4
47.7
1.4
2.4
4.2
1.0
1.9
2.1
1.1
0.6
11.8
3.0
Fe–P–Mo
1–1–0.0012
(10 g)
280
290
300
56.8
70.5
82.3
15.3
17.9
19.8
2.2
2.2
2.1
2.5
2.9
3.3
11.7
14.6
20.6
0.0
0.0
0.0
9.7
11.2
16.6
4.9
6.4
9.4
9.7
15.4
10.5
36.5
32.5
30.6
31.5
45.2
55.6
Fe–P–Mo
1–1–0.005
(5 g)
280
290
300
48.7
66.3
74.2
18.0
22.2
24.0
2.3
2.0
1.8
0.0
0.0
0.0
11.0
17.6
22.4
0.0
0.0
0.0
5.6
8.7
11.4
3.8
5.8
7.0
7.9
10.0
7.9
41.7
36.5
34.8
32.3
41.1
46.8
Fe–P–Mo
1–1–0.01
(5 g)
260
280
300
47.0
64.0
83.8
14.2
21.0
22.9
1.7
2.3
1.6
0.9
1.5
1.7
7.2
17.1
29.9
2.2
0.0
0.0
3.4
6.4
10.7
2.6
3.8
6.9
6.7
11.9
10.1
35.7
38.9
31.2
26.4
36.0
50.4
Fe–P–Mo
1–1–0.02
(5 g)
260
280
300
44.7
65.4
81.4
12.0
19.0
22.6
2.6
2.0
1.6
0.0
0.0
0.0
6.8
17.3
25.2
3.3
0.0
0.0
3.7
6.7
10.0
3.6
6.5
7.3
12.8
13.9
14.8
32.6
32.5
29.6
23.4
39.5
49.6
Fe–P–Mo
1–1–0.04
(5 g)
260
280
300
50.0
75.0
83.8
13.3
18.0
24.3
1.1
2.1
1.4
0.8
1.5
0.0
8.7
21.8
28.3
3.2
0.0
0.0
7.3
6.6
11.1
3.9
6.2
6.7
12.5
18.7
12.0
30.2
28.8
30.7
25.0
41.9
51.6
Fe–P–Mo
1–1–0.075
(5 g)
260
270
280
59.2
66.4
78.6
11.9
13.5
17.0
2.3
1.9
2.0
0.8
1.0
0.0
11.6
15.9
23.1
3.0
0.0
0.0
4.7
5.8
7.9
2.0
2.4
3.2
22.8
26.0
25.3
25.4
24.7
24.3
Fe–P–Mo
1–1–0.5
(5 g)
240
245
255
75.0
79.4
89.2
9.8
10.7
11.8
1.8
1.8
2.0
1.7
1.8
2.0
21.3
30.1
44.0
0.0
0.0
0.0
4.1
4.2
4.7
6.0
7.4
9.7
30.4
23.4
15.0
17.7
18.0
17.7
52.0
53.6
68.0
Fe–P–Mo
1–1–1
(5 g)
240
250
260
55.1
73.8
84.6
6.1
8.0
9.6
1.7
1.6
1.7
0.7
1.0
1.3
12.9
26.8
39.1
0.0
0.0
0.0
3.2
3.9
3.3
2.5
4.4
6.4
28.0
28.1
22.2
15.5
14.4
14.9
26.4
37.6
47.2
Abbreviations are the same as those in Table 1. Conv O2 = conversion of oxygen.

378 Bull. Chem. Soc. Jpn., 75, No. 2 (2002)
Glyoxal from Ethylene Glycol
Fig. 3. Relation between the selectivity and conversion. (ꢀ):
glyoxal, (ꢂ): formaldehyde, (ꢃ): acetaldehyde, (ꢁ): car-
bon oxides, (ꢄ): unidentified compounds, (ꢅ): sum of gly-
colaldehyde and glyoxylic acid, (ꢆ) sum of glyoxal, gly-
colaldehyde, and glyoxylic acid.
Fig. 1. Effects of the content of molybdenum(Ⅵ) on the con-
version and selectivity (Part 1). (ꢀ): Conversion of ethyl-
ene glycol at 260 °C. Selectivity at a conversion of 60%;
(ꢁ): glyoxal, (ꢂ): formaldehyde, (ꢃ): sum of glyoxal, gly-
colaldehyde, and glyoxylic acid.
dentified compounds are relatively low. The selectivity to car-
bon oxides increases with the conversion, but the relation be-
tween the conversion and the selectivity to acetaldehyde and
that to unidentified compounds are not clear.
4. Effects of the Molybdenum Content on the Selectivity.
Since the selectivity varies depending on the extent of reaction,
the values of selectivity obtained from the catalysts with differ-
ent molybdenum(Ⅵ) contents were compared at a fixed level
of conversion. The selectivity to glyoxal, that to formalde-
hyde, and that to the sum of glyoxal, glycolaldehyde, and gly-
oxylic acid at a conversion of 60% are shown in Figs. 1 and 2
(closed symbols).
As the content of molybdenum(Ⅵ) increases, the selectivity
to glyoxal increases sharply, shows a broad maximum at a Mo/
Fe atomic ratio of 0.01, and then falls gradually. The selectivi-
ty to the sum of glyoxal, glycolaldehyde, and glyoxylic acid
also shows a broad maximum at a Mo/Fe atomic ratio of 0.005.
It is likely that the catalysts with a Mo/Fe atomic ratio of lower
than 0.05 are beneficial to the formation of glycolaldehyde and
glyoxylic acid. On the other hand, the selectivity to formalde-
hyde increases by the doping of a very small amount of molyb-
denum(Ⅵ) (Mo/Fe atomic ratio of 0.001), but it levels off with
a further increase in the molybdenum(Ⅵ) content. That is, it is
kept constant at around 25 mol% over the whole range of Fe/
Mo atomic ratio.
5. Effects of the Reaction Temperature on the Selectivity.
Several series of reaction runs were performed with the Fe–P–
Mo = 1–1–0.01 catalyst by changing both the amount of cata-
lyst used from 1.25 to 20.7 g and the reaction temperature from
230 to 340 °C. The feed rates of ethylene glycol, oxygen, wa-
ter, and nitrogen were 14.5, 25.0, 954, and 500 mmol/h, re-
spectively. The values of selectivity to each product at an eth-
ylene glycol conversion of 60% are plotted as a function of the
reaction temperature in Fig. 4.
Fig. 2. Effects of the content of molybdenum(Ⅵ) on the con-
version and selectivity (Part 2). The symbols are the same
as those in Fig. 1.
es more dully in a Mo/Fe atomic ratio ranging from 0.01 to
0.3, and then it levels off with a further increase in the molyb-
denum(Ⅵ) content.
3. Relation between the Conversion and the Selectivity.
The relations between the conversion and the selectivity were
studied. Since the oxidative dehydrogenation of ethylene gly-
col consists of several consecutive reactions, it is reasonable
that the selectivity changes depending on the variation in the
extent of reaction. The reaction was performed over a 5 g-por-
tion of the Fe–P–Mo = 1–1–0.01 catalyst. The conversion
was changed by changing the temperature from 260 to 310 °C,
while fixing the rate of feed. The values of selectivity to each
product are plotted as a function of the conversion in Fig. 3.
As the conversion increases, the selectivity to glyoxal and
that to the sum of glycolaldehyde and glyoxylic acid decrease
a little, while that to formaldehyde clearly increases. The se-
lectivity to acetaldehyde, that to carbon oxides, and that to uni-
The selectivity to glyoxal shows a broad maximum at 290 to
310 °C, while the selectivity to the sum of glycolaldehyde and
glyoxylic acid decreases steadily as the temperature is raised.
Accordingly, a maximum value of the selectivity to the sum of

M. Ai
Bull. Chem. Soc. Jpn., 75, No. 2 (2002) 379
Fig. 4. Effects of the temperature on the selectivity at a con-
version of 60%. The symbols are the same as those in
Fig. 3.
Fig. 5. Effects of the feed rate of oxygen on the conversion
and selectivity. (ꢀ): Conversion of ethylene glycol at 260
°C. Selectivity at a conversion of 80%; (ꢁ): glyoxal, (ꢄ):
sum of glycolaldehyde and glyoxylic acid, (ꢃ): sum of
glyoxal, glycolaldehyde, and glyoxylic acid.
glyoxal, glycolaldehyde, and glyoxylic acid is achieved at a
lower temperature ranging from 260 to 280 °C. The selectivity
to formaldehyde falls markedly as the temperature is raised,
while that to unidentified compounds (others) and that to ace-
taldehyde increase. The selectivity to carbon oxides increases
slightly with the temperature.
6. Effects of Oxygen Concentration. Five series of reac-
tion runs were performed using a 5 g portion of the Fe–P–Mo
= 1–1–0.01 catalyst and by changing both the feed rate of ox-
ygen from 5 to 25 mmol/h and the reaction temperature from
260 to 320 °C, while fixing the feed rate of ethylene glycol as
14.5 mmol/h.
The selectivity to glyoxal, that to the sum of glycolaldehyde
and glyoxylic acid, and that to the sum of glyoxal, glycolalde-
hyde, and glyoxylic acid obtained at a fixed ethylene glycol
conversion of 80% are also plotted in Fig. 5 (closed symbols).
The values of selectivity increase only slightly as the oxygen
feed rate is increased.
7. Effects of the Concentration of Water on the Selectivi-
ty. Three series of reaction runs were performed over a 5 g
portion of the Fe–P–Mo = 1–1–0.01 catalyst by changing the
feed rate of water. The feed rate of oxygen was fixed as 20
mmol/h. The results are listed in Table 3. The selectivity to
each product and that to the sum of glyoxal, glyxolaldehyde,
and glyoxylic acid obtained at a fixed ethylene glycol conver-
sion of 60% are plotted as a function of the feed rate of water
The conversions of ethylene glycol at 260 °C are plotted as a
function of the feed rate of oxygen in Fig. 5 (open circles).
The conversion increases almost proportionally to the feed rate
of oxygen.
Table 3. Effect of the Water Concentration on the Performances
Water feed
(mmol/h)
T
Conv
Yield/mol%
HCHO HCOOH
SG total
°C
%
GX
16.3
17.6
21.1
21.7
22.0
GALD
3.4
2.8
2.4
2.2
GXAD
0.0
0.0
0.0
0.0
AcH
5.4
6.8
9.0
9.9
COx
2.9
4.1
6.0
6.6
83
others
27.1
30.1
25.4
25.3
19.2
mol%
30.7
27.1
28.2
27.1
26.8
260
270
280
285
290
63.9
75.0
83.4
88.2
89.8
8.9
13.7
19.5
22.5
27.5
0.0
0.0
0.0
0.0
0.0
200
ct = 1.55 s
2.2
0.0
10.8
270
280
285
290
295
59.0
72.0
76.7
79.3
84.6
18.5
22.2
23.2
24.6
25.8
2.3
2.8
2.6
1.9
2.2
1.3
2.1
0.0
0.0
2.3
10.7
16.5
18.4
24.5
25.1
0.0
0.0
0.0
0.0
0.0
4.4
7.0
8.3
8.4
11.3
3.1
4.2
5.2
6.5
6.8
18.7
19.4
19.0
11.3
11.1
37.5
34.7
33.6
36.0
35.8
452
ct = 1.15 s
270
280
290
300
305
310
45.9
54.7
64.4
76.0
83.0
86.4
17.8
20.7
23.5
25.9
26.3
26.5
3.3
2.6
2.3
2.0
1.7
1.6
1.4
1.6
1.9
2.3
2.3
2.4
11.2
13.2
19.0
25.6
29.7
33.2
0.0
0.0
0.0
0.0
0.0
0.0
5.9
4.5
6.3
.6
10.5
11.7
2.0
2.9
3.8
5.1
6.1
7.9
4.3
9.2
7.7
6.6
6.4
3.1
48.9
45.6
42.8
39.7
36.6
35.3
954
ct = 0.763 s
Catalyst: Fe–P–Mo (1–1–0.01), 5.0 g. Feed rates of ethylene glycol, oxygen, nitrogen = 14.5, 20, 500 mmol/h. Abbreviations are
the same as those in Table 1.

380 Bull. Chem. Soc. Jpn., 75, No. 2 (2002)
Glyoxal from Ethylene Glycol
iron(Ⅲ) phosphate is not suitable as a catalyst for the forma-
tion of glyoxal.
However, both the oxidation activity and the selectivity to
glyoxal are improved dramatically by doping of a very small
amount of molybdenum, as in the case of oxidative dehydroge-
nation of lactic acid to pyruvic acid.19 The oxidation activity
of the catalyst with Mo/Fe atomic ratio of 0.001 is of the same
order of magnitude as those of iron(Ⅲ) molybdete,
H3PMo12O40, and vanadium(Ⅴ) phosphate.
As seen in Figs. 1 and 2, the oxidative C–C bond fission to
form formaldehyde always takes place in parallel to the oxida-
tive dehydrogenation to form glyoxal. On the other hand, as
the extent of reaction increases, the selectivity to glyoxal falls,
while that to formaldehyde increases (Fig. 3). This finding
suggests that a part of glyoxal is transformed consecutively to
formaldehyde. Since the formation of CO was much smaller
than that of CO2, the decomposition reaction is assumed to be
oxidative, as follows:
Fig. 6. Effects of the feed rate of water on the selectivity.
Selectivity at a conversion of 60%. The symbols are the
same as those in Fig. 3.
OCH–CHO + 0.5 O2 → HCHO + CO2
in Fig. 6.
It is clear that the selectivity to glyoxal and that to formalde-
hyde increase as the concentration of water is increased, while
that to unidentified compounds falls.
The selectivity to glyoxal may be governed by two factors:
(1) side-reactions parallel to glyoxal formation, (2) consecu-
tive degradations of glyoxal. Therefore, it is reasonable that
one ascribes the enhancement of selectivity by the molybde-
num(Ⅵ) doping to the enhancement of the rate of glyoxal for-
mation relative to the rate of C–C bond fission which takes
place in parallel to the glyoxal formation and to the rate of con-
secutive degradation of glyoxal. To understand the role of mo-
lybdenum(Ⅵ) in more detail, further information seems to be
necessary.
Discussion
As seen in Table 1, the iron(Ⅲ) molybdate, which is known
as a typical catalyst for oxidative dehydrogenation of methanol
to formaldehyde, is very active, but it is not suitable for the ox-
idative dehydrogenation of ethylene glycol to glyoxal, because
it promotes mainly an oxidative C–C bond fission to form two
moles of formaldehyde.
The rate of consumption of ethylene glycol increases almost
in proportion to the feed rate of oxygen (Fig. 5). Similar re-
sults were observed in the cases of oxidation of other com-
pounds over iron phosphate catalysts.2,7,14–16 On the other
hand, the selectivity to glyoxal is not much affected by the feed
rate of oxygen (Fig. 5). The finding suggests that the oxygen
dependency of side-reactions is the same as that of glyoxal for-
mation reaction.
The selectivity increases markedly as the concentration of
water increases (Table 3 and Fig. 6). Similar results were usu-
ally observed in the oxidative dehydrogenation reactions per-
formed over iron phosphate catalysts.2,9,12–17 It has been pro-
posed that the crystalline water of iron phosphate takes part in
promoting the oxidative dehydrogenation,13,25 and that the re-
dox cycles in iron(Ⅲ) phosphate are performed in a manner
different from that in oxides or phosphates of molybdenum(Ⅵ)
and vanadium(Ⅴ), since iron(Ⅲ) phosphate has no metal-oxy-
gen double bond species, unlike oxides or phosphates of mo-
lybdenum and vanadium.17
It is interesting to note that the selectivity to formaldehyde
and that to unidentified compounds vary in opposite directions.
As the temperature is raised, the selectivity to formaldehyde
falls, while that to unidentified compounds increases (Fig. 4).
On the other hand, as the feed rate of water is increased, the se-
lectivity to formaldehyde increases, while that to unidentified
compounds falls (Fig. 6). Lower temperatures and higher wa-
ter concentrations are beneficial to the formaldehyde forma-
tion, while higher temperatures and lower water concentrations
HOCH2–CH2OH + 0.5 O2 → 2HCHO + H2O
The supported H3PMo12O40, which is a typical acidic oxida-
tion-catalyst, shows a sufficient oxidation activity, but the se-
lectivity to the sum of glyoxal, glycolaldehyde, and glyoxylic
acid is less than 20 mol%. The main products are formalde-
hyde and unidentified compounds. It is interesting to note that
the catalyst produces uniquely a relatively large amount of for-
mic acid. This finding suggests that the catalyst is inactive for
decomposition of formic acid. Possibly, the catalyst is poor in
the basic properties which would activate acidic compounds.24
The vanadium(Ⅴ) phosphate is not effective for the forma-
tion of glyoxal. Interestingly, it promotes mainly the forma-
tion of acetaldehyde. It is likely that ethylene glycol is dehy-
drated to form ethylene oxide (C2H4O) and the obtained ethyl-
ene oxide is then transformed into acetaldehyde by isomeriza-
tion. These two steps are acid-catalyzed steps, but not oxida-
tion steps. The dehydration of ethylene glycol may be the rate-
determining step. It is still hard to understand why the vanadi-
um(Ⅴ) phosphate promotes the dehydration more strongly than
the H3PMo12O40 catalyst does.
HOCH2–CH2OH → C2H4O + H2O → CH3CHO + H2O
The pure iron(Ⅲ) phosphate is very low both in the oxida-
tion activity and in the selectivity to glyoxal (Table 2). The
main products are acetaldehyde and carbon oxides. The pure

M. Ai
Bull. Chem. Soc. Jpn., 75, No. 2 (2002) 381
9
M. Dekiouk, N. Boisdron, S. Pietrzyk, Y. Barbaux, and J.
are beneficial to the formation of unidentified compounds. At
present it is hard to explain fully the results.
Grimblot, Appl. Catal., 90, 61 (1992).
10 M. Ai, E. Muneyama, A. Kunishige, and K. Ohdan, J. Cat-
al., 144, 632 (1994).
11 M. Ai, E. Muneyama, A. Kunishige, and K. Ohdan, Appl.
Catal., 109, 135 (1994).
12 D. Rouzies, J. M. M. Millet, D. Siew Hew Sam, and J. C.
Védrine, Appl. Catal., 124, 189 (1995).
13 J. M. M. Millet, D. Rouzies, and J. C. Védrine, Appl. Cat-
al., 124, 205 (1995).
14 M. Ai and K. Ohdan, Appl. Catal., 150, 13 (1997).
15 M. Ai and K. Ohdan, Stud. Surf. Sci. Catal., 110, 527
(1997).
The selectivity to acetaldehyde increases as the temperature
is raised (Fig. 4). It is likely that a higher temperature is bene-
ficial to the dehydration of ethylene glycol to ethylene oxide.
Iron(Ⅲ) phosphate is considered to be inactive for promot-
ing oxygen insertion reactions. However, ethylene glycol is
very susceptible to C–C bond fission by oxygen insertion, ac-
cordingly the formation of glyoxal from ethylene glycol seems
to be one of very difficult reactions. At present it is still hard to
suppress completely the C–C bond fission of ethylene glycol
even with the promoted iron(Ⅲ) phosphate catalysts, though
clearly improved performances are achieved.
16 M. Ai and K. Ohdan, Bull. Chem. Soc. Jpn., 72, 2143
(1999).
References
17 M. Ai, Catal. Today, 52, 65 (1999).
18 M. Ai, Stud. Surf. Sci. Catal., 110, 527 (1997).
19 M. Ai, Preprints, Europacat V, Limerick, Sept., 2001, 16-
P-01.
1
W. C. Atkins, (to Eastman Kodak Co.), U.S. Patent 3 855
279 (1974).
2
C. Virely, O. Fabregue, and M. Forissier, Bull. Soc. Chim.
20 P. Gallezot, S. Tretjak, Y. Christidis, G. Mattioda, and A.
Schouteeten, J. Catal., 142, 729 (1993).
21 V. R. Chumbhale and P.A. Awasarkar, Appl. Catal., 205,
109 (2001).
22 M. Ronis and F. d’Yoire, C. R. Acad. Sci. Paris, 269c, 1388
(1969).
23 A. Goiffon, J. C. Jumas, and E. Philippot, Rev, Chim.
Minér., 24, 99 (1986).
24 M. Ai, J. Catal., 50, 291 (1977).
Fr., 3, 457 (1988).
3
J. C. Védrine, J. M. M. Millet, and J. C. Volta, Faraday
Discuss. Chem. Soc., 87, 207 (1989).
4
J. M. M. Millet, C. Virely, M. Forissier, P. Bussière, and J.
C. Védrine, Hyperfine Interact., 46, 619 (1989).
J. M. M. Millet, J. C. Védrine, and G. Hecquet, Stud. Surf.
Sci. Catal., 55, 833 (1990).
5
6
(1991).
7
J. M. M. Millet and J. C. Védrine, Appl. Catal., 76, 209
25 E. Muneyama, A. Kunishige, K. Ohdan, and M. Ai, Catal.
Lett., 31, 209 (1995).
C. Virely, M. Forissier, P. Bussière, J. M. M. Millet, and J.
C. Védrine, J. Mol. Catal., 71, 199 (1992).
Y. Barbaux and M. Dekiouk, Appl. Catal., 90, 51 (1992).
8