IMPROVEMENT OF PROPERTIES OF CATALYST FOR INDUSTRIAL PRODUCTION OF ACROLEIN FROM PROPYLENE

Article abstract of DOI:10.1135/cccc19941536

The catalytic gas-phase oxidation of propylene on a multicomponent Mo-Co-Fe-Bi-K-O catalyst gives up to 85percent yield of acrolein and acrylic acid along with undesirable side products - formaldehyde and acetaldehyde.Effect of four factors have now been studied with the aim of suppressing of formation of the by-product viz. the effect of potassium content in the catalyst, that of final annelating temperature of the catalyst, and those of addition of other components (tungsten, nickel).It has been found that a charge in potassium content in catalyst affect distinctly its activity and also selectivity for products of total oxidation of propylene but does not practically affect the yields of formaldehyde and acetaldehyde.The yields of the said side products decreased with increasing annelating temperature, however, with concominant considerable decrease in the catalytic activity.Whereas a partial substitution of molybdenum atoms in the catalyst by tungsten atoms is accompanied by an increase in production of formaldehyde and acetaldehyde, some catalysts with nickel gave acrolein as the only aldehyde.

Full text of DOI:10.1135/cccc19941536

1536
Machek, Tichy, Martinec, Svachula:
IMPROVEMENT OF PROPERTIES OF CATALYST FOR INDUSTRIAL
PRODUCTION OF ACROLEIN FROM PROPYLENE
a
a
b
a
Jaroslav MACHEK , Josef TICHY , Alexander MARTINEC and Jiri SVACHULA
a Department of Physical Chemistry,
University of Pardubice, 532 10 Pardubice, The Czech Republic,
b Chemical Plants Sokolov,
356 80 Sokolov, The Czech Republic
Received September 28, 1993
Accepted February 10, 1994
The catalytic gas-phase oxidation of propylene on a multicomponent Mo–Co–Fe–Bi–K–O catalyst
gives up to 85% yield of acrolein and acrylic acid along with undesirable side products – formalde-
hyde and acetaldehyde. Effects of four factors have now been studied with the aim of suppressing of
formation of the by-products viz. the effect of potassium content in the catalyst, that of final anneal-
ing temperature of the catalyst, and those of addition of other components (tungsten, nickel). It has
been found that a change in potassium content in catalyst affects distinctly its activity and also selec-
tivity for products of total oxidation of propylene but does not practically affect the yields of formal-
dehyde and acetaldehyde. The yields of the said side products decreased with increasing annealing
temperature, however, with concomitant considerable decrease in the catalyst activity. Whereas a par-
tial substitution of molybdenum atoms in the catalyst by tungsten atoms is accompanied by an in-
crease in production of formaldehyde and acetaldehyde, some catalysts with nickel gave acrolein as
the only aldehyde.
At present acrolein is produced on industrial scale almost exclusively by gas-phase
catalytic oxidation of propylene on multicomponent catalysts. The modern oxide cata-
lysts are usually based on Mo–Bi (Sohio catalyst) which in 1959 was found to produce
acrolein in a higher yield than the copper catalyst used up to that time1. Beside the
mentioned Mo–Bi system the multicomponent catalysts1,2 often contain also Co, Fe, W,
P, Tl, Sb, Sn, Li, K etc. forming thus complex systems both chemically and structu-
rally. The above-mentioned catalysts work at the temperatures of 300 – 400 °C, when
the yield of acrolein reaches up to 85% with 5 – 8% propylene content in the reaction
mixture. The by-products involve acrylic and acetic acids beside (in most cases) for-
maldehyde, acetaldehyde, acetone, allyl alcohol, CO2, and CO. The service life of these
catalysts in industrial reactors reaches 2 – 5 years3.
In the present paper we have focused our attention on improving the properties of the
Mo–Bi–Co–Fe–K–O catalyst4 which at the temperatures of 310 – 340 °C allowed an up
to 95% conversion of propylene with 75% yield of acrolein and 15% yield of acrylic
acid; the yields of formaldehyde, acetaldehyde, and acetone reached up to 3% depend-
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

Catalyst for Oxidation of Propylene to Acrolein
1537
ing upon the way of controlling the oxidation. As, in most cases, acrolein together with
the other products of propylene oxidation is immediately introduced into another reac-
tor where it is oxidized to acrylic acid (on a different catalyst and at different conditions),
formaldehyde–acetaldehyde–acetone by-products represent an undesirable admixture
lowering the yields of acrylic acid. Therefore, such a composition of catalyst was
looked for as to remove this drawback and to produce acrolein as the only aldehyde.
EXPERIMENTAL
Catalysts
The investigated oxide catalysts based on Mo–Co–Fe–Bi–X–K–O where X represents W or Ni were
prepared from the solutions of ammonium heptamolybdate, nitrates of the respective metals, KOH,
and tartaric acid.
Tungsten was added in the form of 5 [(NH4)2O] . 12 WO3 . 4 H2O. A detailed procedure of prep-
aration of the catalysts is given elsewhere4,5. The application of tartaric acid ensures the final mixture
of all the components to be in the form of a solution, which in turn ensures the homogeneity and
reproducibility of the catalysts prepared. The pH value of the final solutions was adjusted at 6 – 7 by
adding aqueous ammonia, whereafter the solutions were evaporated with continuous stirring at 90 °C
to give a microcrystalline paste which was dried at 120 °C 12 h, annealed at 160 and 180 °C 3 h, at
360 °C 5 h, and at 420 °C 24 h (and at 460 and 500 °C for another 24 h in the case of the catalysts
C1(460) and C1(500), respectively). The samples obtained were ground and sieved to give the frac-
tion of 0.40 – 0.63 mm grain size.
Apparatus
The activity and selectivity of the catalysts produced were investigated with the help of a through-
flow apparatus with integral reactor with solid bed. The reactor was heated by means of an electric
oven with two coilings out of which one was controlled by an automatic controller Radelkis, which
made it possible to maintain the chosen temperature in catalyst bed with the accuracy of ±2 °C. The
reactor had a diameter of 2.5 cm and contained a coaxial tube for a thermocouple. The flow rates of
gases (propylene, O2, N2) were measured by means of capillary flow meters, the water vapour was
dosed into the reaction mixture from a saturator thermostated at a calculated temperature. The com-
position of reaction mixture: C3H6 5, O2 12, H2O 35, N2 48 vol.%. The whole feed of reaction mix-
ture F = 5 dm3 h−1 (p = 1.013 . 105 Pa, T = 273.13 K), weight of catalyst 1 – 5 g. The catalyst was
always diluted with SiC of the same grain size at a volume ratio of 1 : 1, hence the bed height in the
reactor was always greater than 1 cm. It was proved that at this set of conditions the effects of inner
and outer diffusion can be neglected. At the end of the through-flow apparatus there was a water
absorber where the liquid products of oxidation (carbonyl compounds and acids) were absorbed.
Analysis of Reaction Products
The conversion of propylene (the measure of catalyst activity), selectivity, and yields of the individ-
ual oxidation products (for their definition see ref.6) were calculated from analyses carried out on a
GC Chrom 5 apparatus with thermal-conductivity (TCD) or flame-ionization (FID) detectors attached
to the through-flow apparatus. A system of three six-way cocks (Perkin–Elmer) was adopted to pro-
vide the dosing of reaction mixture into the gas chromatograph before and after passing the catalyst
bed of reactor. The columns of 2.5 and 1.2 m length and 3 mm i.d. were packed with Porapak Q for
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

1538
Machek, Tichy, Martinec, Svachula:
determinations of CO, CO2, C3H6, acetone, and aldehydes, and with Chromosorb 101 for determina-
tions of acids.
In addition, the overall yield of acids was checked by the titration of absorbate from the water
absorber with an NaOH solution titrant and with the help of a pH meter. The qualitative composition
of the carbonyl compounds formed was determined via the corresponding 2,4-dinitrophenylhydra-
zones using descending paper chromatography7; the overall yield was determined according to Smith
and Mitchel8. The amount of the acrolein formed was also checked polarographically according to
Moshier9.
RESULTS AND DISCUSSION
For preparing the basic oxide catalyst C1 with the type designation
Mo12Co4Fe1Bi1K0.08OZ the amounts of the respective components were chosen in such
a way as to obtain the molar ratios of the metals Mo : Co : Fe : Bi : K = 12 : 4 : 1 : 1
: 0.08 (the oxygen content in the catalysts was not determined). The study of propylene
conversion and of yields of individual oxidation products in the temperature interval of
308 – 362 °C were carried out for various values of the time factor W/FC H , and the
3
6
results are presented in Table I. It can be seen that the summary yield of useful reaction
products, i.e. acrolein and acrylic acid, is above 80% in most cases. However, in all the
TABLE I
Conversion of propylene and yields of some oxidation products (in %) on the basic catalyst C1
W/FC H
g h mol
Reaction
temperature, °C
3
6
−1
XC H
YAC
YAA
YALD
3
6
194
315
325
334
80.0
83.2
85.8
69.3
71.6
70.1
8.0
9.2
0.5
1.3
2.5
10.4
243
335
345
362
87.0
89.1
92.2
75.4
76.0
70.1
8.6
9.9
1.2
1.7
2.8
11.8
280
320
327
340
94.2
96.0
79.5
75.1
10.7
10.6
2.4
2.9
318
330
336
89.5
93.6
95.5
77.4
75.8
72.6
9.7
12.2
12.3
0.8
1.7
1.6a
404
308
318
327
340
85.3
94.0
95.5
96.8
77.6
76.6
76.4
74.2
5.5
11.0
15.6
16.3
1.3
1.7
2.6
3.0a
a Traces of acetone were formed too.
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

Catalyst for Oxidation of Propylene to Acrolein
1539
cases formaldehyde and acetaldehyde were also formed, sometimes with traces of
acetone.
With the aim to suppress these undesirable reactions connected with partial destruc-
tion of carbon chain we studied the effects of four factors upon the oxidation course:
the effect of potassium amount, that of the temperature of annealing of catalyst, and
those of additions of other components (tungsten or nickel).
The effect of amount of potassium: For this purpose we prepared (in the above-men-
tioned way) five oxide catalysts of Mo12Co4Fe1Bi1KNOZ type where N = 0, 0.04, 0.055,
0.07, and 0.12 for the catalysts C2, C3, C4, C5, and C6, respectively. Table II presents
the results of tests of the given catalysts at the temperature of 320 °C and at the time
factor value W/FC H = 171 g h mol−1. The catalyst C2, which does not contain potas-
6
sium, is highly act3ive at the conditions given but, at the same time, also highly selective
for the products of total oxidation. With increasing content of potassium and, hence,
increasing number of basic centres at the catalyst surface both the catalyst activity and
its selectivity for products of total oxidation decrease. The maximum selectivity for
acrolein was reached with the C5 catalyst, of course, it decreased by 10% with increas-
ing the time factor value to W/FC H = 314 g h mol−1.
6
With the catalyst C1 the same 3change in the time factor caused almost no change in
the selectivity for acrolein, and the selectivity for acrylic acid even increased. The
catalyst with the highest potassium content (C6) shows a marked decrease in the con-
version of propylene accompanied by a distinct decrease in selectivity for acrolein. In
all the cases, however, a change in potassium content of the catalysts did not affect the
formation of formaldehyde and acetaldehyde. Hence it is obvious that even slight
TABLE II
Effect of amount of potassium upon activity and selectivity (in %) of catalysts of
Mo12Co4Fe1Bi1KNOZ type (reaction temperature 320 °C, W/FC H = 171 g h mol−1
3
6
Catalyst
N
XC H
SAC
SAA
ST
SALD
3
6
C2
C3
C4
C5
0
95.6
95.0
91.5
81.3
94.0a
65.6
90.5a
54.4
70.3
81.0
78.3
93.5
83.5
91.5
90.1
74.7
10.5
7.6
18.8
9.9
9.1
3.1
6.1
2.6
3.0
2.8
1.5
1.8
2.1
0.7
1.6
0.5
1.4
3.6
0.04
0.055
0.07
9.8
3.0
8.5
C1
C6
0.08
0.12
3.9
5.8
13.5b
a
The measurements were repeated for W/FC H = 314 g h mol−1. b Acetic acid is formed together with
3
6
acrylic acid.
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

1540
Machek, Tichy, Martinec, Svachula:
changes in the potassium content in the catalysts affect distinctly the numbers of active
centres at the catalyst surface which are responsible primarily for the total oxidation of
propylene but do not affect the active centres causing the partial destruction of carbon
chain in propylene or acrolein.
The effect of the final temperature of annealing: For these experiments we selected
the basic oxide catalyst C1 of Mo12Co4Fe1Bi1K0.08OZ type which gave 90% selectivity
for acrolein at 90% conversion of propylene (Table II). A part of the catalyst mentioned
was annealed at 460 °C for 24 h and another part at 500 °C for 24 h to give the samples
denoted as C1(460) and C1(500), respectively. These catalysts were tested at the tem-
peratures of 320 and 335 °C and at the time factor value W/FC H = 314 g h mol−1. As
3
6
expected, the increased temperature of annealing resulted in decreased activity of cata-
lyst (Table III). The catalyst C1(500) produces acrolein as the only aldehyde (without
even traces of formaldehyde and acetaldehyde), of course, with distinct drop in activity.
Hence, it can be presumed that formaldehyde and acetaldehyde are probably formed
only from acrolein by a consecutive reaction. The effect of final temperature of anneal-
ing upon the phase composition of the catalyst C1 was studied with the help of a JEOL
JDX-85 diffractograph at the following conditions: CuKα radiation, LiF monochroma-
tor. All the three catalysts contain MoO3, which is indicated by the presence of the
diffraction peaks10 with the values of interplane distances dhkl = 3.82, 3.46, 2.66, and
2.31 10−10 m, and furthermore Bi2O3 . 3 MoO3 (α-phase) with the values11 dhkl = 6.96,
3.19, 3.06, and 2.86 . 10−10 m. Again, intensive diffraction peaks with the values dhkl
=
3.80, 3.38, 3.26, and 2.66 . 10−10 m were detected in all the three catalysts, which
indicates the presence of α-phase CoMoO4 (ref.12). The diffraction peaks of the values
dhkl = 3.26, 3.19, and 3.07 . 10−10 m can also be due to the β-phase of the Bi–Mo
system13 (i.e. Bi2O3 . 2 MoO3). The presence of this phase in the catalysts also seems to
be confirmed by a fairly marked diffraction peak13 with the value dhkl = 4.91 . 10−10 m.
TABLE III
Effect of final annealing temperature upon activity and selectivity (in %) of basic catalyst of
Mo12Co4Fe1Bi1K0.08OZ type (time factor W ⁄ FC H = 314 g h mol−1)
3
6
Reaction temperature,
Catalyst
C1
XC H
SAC
SAA
ST
SALD
3
6
°C
320
335
90.5
94.2
90.1
76.6
5.8
9.6
3.0
8.0
1.4
3.0
C1(460)
C1(500)
320
335
79.2
81.4
91.0
91.6
5.1
6.0
2.5
2.8
1.1
0.2
320
335
22.0
96.5
95.5
3.8
4.1
–
–
–
–
30.4
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

Catalyst for Oxidation of Propylene to Acrolein
1541
Interestingly, no diffraction peak due to Mo–Fe system could be observed. After pre-
cipitation of ammonium heptamolybdate with Fe(NO3)3 and subsequent annealing,
three different Mo–Fe forms are formed14,15: mixed Fe2(MoO4)3 . MoO3, β-phase of
FeMoO4, and Fe2(MoO4)3. The presence of any given crystalline form of iron mo-
lybdate should make itself felt by a more distinct diffraction peak of the value dhkl
=
3.87 . 10−10 m. However, this peak was detected in none of the catalysts. Although the
catalysts C1, C1(460), and particularly C1(500) markedly differ in their activities in
catalytic oxidation of propylene, the X-ray diffraction analysis gave practically identi-
cal results for the three samples.
The effect of tungsten: Two catalysts were prepared for investigating the tungsten
effect, molybdenum being partially replaced by tungsten. The amounts of individual
components were chosen in such a way that the catalysts prepared corresponded to the
types Mo9W3Co4Fe1Bi1K0.08OZ (denoted as C7) and Mo10W2Co4Fe1Bi1K0.08OZ
(referred to as C8). Again the final temperature of annealing was 420 °C. The tests of
catalytic activity and selectivity of these two catalysts were carried out at the value of
time factor W/FC H = 267 g h mol−1, and the results are presented in Table IV. The
3
6
partial replacement of molybdenum atoms by tungsten resulted in a decreased pro-
pylene conversion (as compared with C1), abrupt decrease in acrolein yield, and also
distinct increase in the yields of formaldehyde and acetaldehyde. Hence it is obvious
that introduction of tungsten atoms into the catalyst results in formation of active
centres causing partial destruction of propylene or acrolein to formaldehyde and acetal-
dehyde, i.e. the effects of tungsten upon the properties of the Mo–Co–Fe–Bi–K–O cata-
lyst are quite obviously negative.
The effect of nickel: Four catalysts were prepared for testing the effect of nickel
differing from C1 by gradual replacement of cobalt atoms by nickel. The amounts of
starting components were chosen in such way that the catalysts prepared corresponded
to the types: Mo12Co3Ni1Fe1Bi1K0.08OZ (referred to as C9), Mo12Co2Ni2Fe1Bi1K0.08OZ
(C10), Mo12Co1Ni3Fe1Bi1K0.08OZ (C11), and Mo12Ni4Fe1Bi1K0.08OZ (C12). Moreover,
TABLE IV
Effect of introduction of tungsten upon properties of Mo–Co–Fe–Bi–K–O catalyst (time factor
W/FC H
=
267 g h mol−1, conversion and yields in %)
3
6
Catalyst
C7
Reaction temperature, °C
XC H
YAC
YAA
YALD
3
6
321
337
71.5
82.3
48.2
49.1
12.0
14.6
5.3
8.3
C8
324
335
347
69.8
71.0
73.5
43.2
44.8
41.4
10.1
12.3
14.2
7.7
8.8
10.2
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

1542
Machek, Tichy, Martinec, Svachula:
two catalysts were prepared in which nickel was used as additional component of the
catalyst C1: Mo12Co4Ni2Fe1Bi1K0.08OZ (C13) and Mo12Co4Ni1Fe1Bi1K0.08OZ (C14).
From data in Table V it follows that the presence of nickel in catalyst increases ab-
ruptly its activity in the catalytic oxidation of propylene, whether nickel is used instead
of cobalt (C9 – C12) or in addition to it (C13 and C14). It was possible to attain 90%
TABLE V
Effect of introduction of nickel on properties of Mo–Co–Fe–Bi–K–O catalyst (conversion and yields in %)
W/FC H
g h mol
Reaction
temperature, °C
3
6
Catalyst
C9
XC H
YAC
YAA
YT
YALD
−1
3
6
179
340
355
95.0
96.2
80.1
76.3
10.2
14.2
4.6
5.8
–
–
202
305
325
340
350
86.3
92.7
94.0
95.1
78.5
77.0
77.1
75.0
6.6
9.2
1.2
4.8
5.7
7.2
–
–
–
–
11.0
13.1
269
179
265
275
90.2
96.0
80.3
77.5
6.7
3.2
8.4
–
–
10.0
C10
275
300
320
84.7
94.1
96.3
75.8
76.9
76.0
5.4
9.3
3.8
7.8
8.5
–
–
–
12.0
C11
269
269
250
265
80.2
94.0
70.3
74.5
4.4
9.1
5.0
–
10.0
0.3
C12a
250
281
295
62.8
72.6
92.0
52.3
60.9
65.8
3.7
5.5
7.2
6.9
–
0.5
0.8
14.7
10.6
C13
C14
269
265
280
301
77.0
87.4
90.2
66.2
68.9
70.8
3.3
5.0
8.9
6.8
11.9
12.0
–
0.2
0.2
179
269
305
325
82.2
92.1
75.4
82.6
4.5
6.8
2.1
3.0
–
–
255
265
281
290
301
53.7
75.0
92.1
93.2
95.0
46.8
65.4
73.6
76.6
73.4
4.0
4.8
8.0
9.5
8.8
3.7
4.9
–
–
8.8
0.1
0.1
0.6
8.0
12.7
a In the catalyst C12 all cobalt atoms are replaced by nickel.
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

Catalyst for Oxidation of Propylene to Acrolein
1543
conversion of propylene at the value of time factor W/FC H = 269 g h mol−1 and at the
6
temperatures below 300 °C. The best results were achiev3ed with the catalyst C9. This
catalyst is sufficiently stable in a relatively wide temperature interval (265 – 355 °C)
which is indicated by the selectivity for the valuable oxidation products (acrolein and
acrylic acid) reaching 96.6% at 265 °C and at 90% conversion of propylene, and 94%
at 355 °C and at 96% conversion of propylene. The most important experimental result,
however, consists in the finding that the catalysts C9 and C10 only produce acrolein
and no other aldehydes (not even traces of CH2O and CH3CHO).
The service life of catalyst C9 was tested in a long interval in one tube of an indus-
trial reactor for two-step production of acrylic acid (Mitsubishi Chemical Company)
using the following set of conditions: The tube diameter 27 mm, tube length 5 m,
reaction temperature 300 – 340 °C, composition of reaction mixture: C3H6 – 8, O2 – 14,
H2O – 25.5, N2 – 52.5 vol.%, total feed of reaction mixture F = 870 dm3 h−1. The
catalyst in the form of pellets of 4 mm diameter and height was diluted with inert
particles. The composition of reaction mixture and the temperature somewhat varied
according to actual situation in the reactor. The experiment was performed in such a
way that the analysis of reaction components was carried out at the outlet of the tube in
the second reactor (oxidation of acrolein to acrylic acid). From data of Table VI it is
TABLE VI
Long-term tests of C9 catalyst in a tube of industrial reactor
Time of working load
XC H , %
Nonreacted acrolein, %
YAA, %
3
6
of catalyst, h
500
1 000
1 500
2 000
3 000
3 500
4 000
4 500
5 000
5 500
6 000a
6 500
99.5
99.5
100.0
99.2
99.4
99.0
99.1
99.0
99.0
99.2
98.2
99.0
0.2
0.3
1.8
0.5
0.2
0.5
1.5
3.0
2.4
1.5
8.4
1.6
83.5
83.0
82.0
87.0
86.0
85.0
85.0
82.5
83.5
84.2
80.0
83.9
a When the technology was put out of action, the catalyst of the second step was reactivated with air
23 h.
Collect. Czech. Chem. Commun. (Vol. 59) (1994)

1544
Machek, Tichy, Martinec, Svachula:
obvious that even after 6 500 h working load in an industrial reactor, the catalyst C9
retains its excellent parameters. The yield of acrylic acid lower by ca 1% than that
guaranteed by the Japanese producer (84.6%) is most likely due to the somewhat lower
selectivity of the catalyst of the second step (oxidation of acrolein to acrylic acid).
SYMBOLS
dhkl
F
W/FC H
interplane distance, m
total feed of reaction mixture, dm3 h−1 (p = 1.013 . 105 Pa, T = 273.13 K)
time factor, g h mol−1, where W is catalyst weight, g, and FC H is propylene feed,
3
6
3 6
mol h−1
Si
XC H
Yi
selectivity for the i-th product, %
conversion of propylene, %
yield of the i-th product, %
3
6
Abbreviations
AA
AC
ALD
T
acrylic acid
acrolein
formaldehyde plus acetaldehyde
products of total oxidation (CO2 and CO)
REFERENCES
1. Keulks G. W., Krenzke L. D., Notermann T. N.: Adv. Catal. 27, 183 (1978).
2. Hucknall D. J.: Selective Oxidation of Hydrocarbons, p. 23. Academic Press, London 1974.
3. Ullmann’s Encyclopedia of Industrial Chemistry, A1, p. 152. VCH Verlagsgesellschaft,
Weinheim 1985.
4. Tichy J., Machek J., Tockstein A., Opatril P., Majer I.: Czech. 185 016 (1980); Chem. Abstr. 95,
80170 (1981).
5. Martinec A., Moravek J., Havlicek W., Berka Z., Slavik J., Tichy J., Machek J., Svachula J.,
Rosenthal J.: Czech. 242 277 (1987); Chem. Abstr. 109, 150213 (1988).
6. Machek J., Tichy J., Svachula J.: Collect. Czech. Chem. Commun. 58, 2867 (1993).
7. Hais I. M., Macek K. (Eds): Papirova chromatografie, p. 238. CSAV Publishing House, Praha
1959.
8. Smith G. J., Mitchel J. T.: Chem. Anal. 22, 750 (1950).
9. Moshier R. W.: Ind. Eng. Chem. Anal. 15, 107 (1943).
10. Mirkin L. I.: Spravochnik po rentgenostrukturnomu analizu polikristallov, p. 504. Gos. Izd.
Fiz.-Mat. Literatury, Moscow 1961.
11. Aykan K.: J. Catal. 12, 281 (1968).
12. Trifiro F., Caputo G., Villa P. L.: J. Less-Common Met. 36, 305 (1974).
13. Erman L. Ya., Galperin E. L.: Zh. Neorg. Khim. 11, 221 (1966).
14. Pasquon I., Trifiro F., Caputo G.: Chim. Ind. (Milan) 55, 168 (1973); Chem. Abstr. 79, 46143
(1973).
15. Carbucicchio M., Trifiro F., Villa P. L.: J. Phys. (Paris), Colloq. 6, 253 (1976); Chem. Abstr. 86,
189089 (1977).
Translated by J. Panchartek.
Collect. Czech. Chem. Commun. (Vol. 59) (1994)