(SANITIZED)UNCLASSIFIED SOVIET BLOC PAPERS ON THE SCIENCE OF CATALYSTS(SANITIZED)

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP80-00247A004200070001-6
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 135
• Document Creation Date: December 27, 2016
• Document Release Date: March 4, 2014
• Sequence Number: 1
• Publication Date: January 1, 1964
• Content Type: REPORT

Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
?
INTERACTION OF GASES IN THE SURFACE OF
NICKEL OXIDE
K. ICLIER and M. JIRATOVA
Institute of Physical Chemistry, Czechoslovak Academy of Sciences,
Prague, Czechoslovakia
? Abstract: An investigation has been made of the processes occurring on the sur-
face of nickel oxide in the presence of carbon monoxide carbon dioxide and oxy-
gen under various conditions. Apart from adsorption-desorption and electric
donchictivity measurements, exchange reactions of radio-carbon have been em-
ployed for the analysis of the adsorbate. Radio-carbon is exchanged between the
adsorbate and gases according to the character of the adsorbate: CO(ads) ex-
changes group CO and CO2(ads) or carbonate exchanges group CO2. The results
show that carbon monoxide forms oxidized surface species with lattice oxygen
of the nickel oxide at.20?C, this reaction being enhanced by increasing tempera-
ture and pressure. When oxygen is preadsorbed, oxidized species are preferably
,created by reaction with adsorbed oxygen with simultaneous disappearance of
CO(ads)? Also carbon dioxide reacts with.adsorbed oxygen with charge transfer.
presumably due to the reaction
n?
0(ads) Ni2+ CO r + Ni3.
?
- 1. INTRODUCTION
-
' Interactions of gases CO, CO2 and 02 on the surface of nickel oxide
have hitherto been investigated by several methods, including chemical
analyses of gases 1), adsorption 2-5,9, 13), calorimetric 6,7), infrared
spectra 8), electric conductivity 2,9,10), thermoelectric power 11) and
.optical spectra.12) measurements. As a result of these investigations, it
has been well established that oxygen is adsorbed in anionic form (061, the
negative charge of the adsorbate being compensated by mobile positive
Charge inside nickel oxide (Ni3+). Thus oxidiz,ed nickel oxide can be fully
reduced to its original electronic state (Ni2+) by carbon monoxide adsorp-
:. tion.at room temperature 9). Increased adsorption heats of carbon monoxide
? on surface precovered by oxygen and of-oxygen on surfaces precovered by
CO and CO2, as compared with bare surfaces, have indicated reactions of
adsorbing gas with the adsorbate 6,.?), -Effect of self-poisoning by CO2 of
? Niafor..carbon monoxide oxidation, first.observed by Roginski and Tselins-
ka.ya 1), has later.been. confirmed by other authors 6,14) and explained to
.,:,be due to the occupation by CO2 of sites available for CO and 02 adsorption.
- However, there has been considerable disagreement among different
Workers concerning the adsorbed amounts of CO and 02, formulas of ad-
- Jtorbed complexes, and the possibility of reactions with lattice oxygen at ?
temperature 7,,15). Also the behaviour of CO2 in this system has not
been fully elucidated. Some authors observed localization of electrons in
Declassified in Part -Sanitized Copy Approved for Release 2014/03/04 : TAT.
CIA-RDP80-00247A004200070001-6
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
2
the adsorbate and release of mobile holes into nickel oxide during CO2 ad-
sorption 2,11), others observed no such effect 9,13). These results sug-
gested that a study of the influence due to the various factors (temperature,
pressure, purity of the surface) on the behaviour of this system as well as
the application of new Methods to analysis of adsorbed complexes would be
of value. -
In this work, interactions of gases with and in the surface of nickel
oxide have been investigated by means of adsorption-desorption measure-
ments, analyses of gaseous phase, electric conductivity measurements,
and exchange reactions of radio-carbon between adsorbate and different
gases, and have been exploited for the analysis of the adsorbate. The pre-
sent results provide information about the composition of and the reactions
in the adsorbed layer under various conditions.
2. EXPERIMENTAL PART
STAT
2.1. Apparatus
Apparatus of the type described earlier 16) was used, allowing meas-
urements to be made of amounts adsorbed and desorbed simultaneously
with electric conductivity, measured in situ by a high frequency method16,17)
at 1 Mc/s. Volume of apparatus was 1576 ml for experiments at pressures
lower than-104 mm Hg and 1142 ml for experiments at pressures of about
2 mm Hg. Pressure was measured by Pirani gauges 16,17) up to 10-1 mm
Hg and by a MacLeod gauge up to 2.5 mm Hg. Two freezing traps in the -
measuring part of apparatus were maintained at -780C except when analyses
were made for CO and CO2 by freezing out carbon dioxide at -1900C. In the
latter case, due corrections were made for changes of volume. Installed in
the wall of the reaction vessel space was also a thin micawindow (1 mg/cm2)
of diameter 30 mm sealed by Araldite to a glass opening in the wall. The
-joint was vacuum tight and the apparatus was evacuable to 10-7 mm Hg. The
radioactivity was measured by an alpha G-M counter, set close to the mica
window outside the reaction space. The solid sample was located inside the
reactor in such a position that the radiation of 14C from adsorbate was
completely absorbed by the glass and could not reach the counter window.
Only radioactivity of the gas was thus measured. Due corrections were
made for background radiation and dead time of the counter. Radioactivity
of the labelled gases was proportional to their pressure. Conveniently,
specific activity of a gas is defined as the number of counts per minute
(cpm) per one molecule of this gas in the apparatus. Two specific activities
were used in the present work: 8.2 x 1045-cpm molecule -1 (14CO3 experi-
ments 1-7, and 14CO2, experiment 10 in table 1 below) for low pressure
experiments, and 2.29 x 10-15 cpm molecule-1 for high pressure experi-
ments (14CO3 experiments 8 and 9 in table 1).
2.2. Materials
- Commercial 14C0 (0.61 mC/inl) was used and purified in situ by vacuum
distillation from active charcoal. 14CO2 was prepared in situ by oxidation of
Declassified in Part - Sanitized Copy Approved for Release2014/03/04: .
CIA-RDP80-00247A004200070001-6 ,
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
3 STAT
14C0 by oxygen on a heated platinum filament and subsequent freezing with
evacuation. Preparation of inactive gases CO, CO2 and 02 is described else-
where 4). The carbon oxide contained less than 5 x 10-3 % of 02.
Nickel oxide of high purity and surface area 99 m2/g was prepared ac-
cording to ref. 3 and 17 by vacuum decomposition of nickel hydroxide at
200?C for 15 hours (final pressure 10-6 mm Hg). The nickel oxide contained
some amount of non-decomposed hydroxide 3,18). The following experiment
allowed us to assess the maximum possible contamination of the surface.
Adsorption of 9.2 x 1017 molecules of oxygen on 0.26 g sample caused elec-
tric conductivity to increase from 0.33 x 10-6 ohm-1 to 0.46 x 10-6 ohm-1,
which effect was completely reversed by subsequent adsorption of CO. On
the other hand, when carbon monoxide was adsorbed on a clean nickel oxide,
no change of conductivity occured, similarly as in 9). From estimated error
of our conductivity measurements, 0.02 x 10-6 ohm-1, one is able to see
that no more than 1.7 x 1017 molecules of oxygen could have been adsorbed
on the initial clean nickel oxide, which is less than OA% of surface cover-
age 6).
2.3. Procedure
First, carbon monoxide labelled by 14C was allowed to be adsorbed for
200 minutes (adsorbed amount a), then the gaseous phase with desorbable
part of the adsorbate was removed by pumping for 30 min, the amount
desorbed being measured and analyzed for CO (amount d) and CO2 in a
separate volume behind the gas collection pump. The adsorbate remaining
on the surface was then subjected to exchange reactions at the same tem-
perature as adsorption. First, inactive CO was admitted, its adsorbed
amount (b), amount of CO in gas (g) and CO2 in gas (q) measured and 14C
activity (y) in gaseous carbon monoxide followed with time. After the con-
centration of radioisotope in the gaseous carbon monoxide (y / g) had reached
an equilibrium value, the gaseous mixture was pumped out and analyzed.
Subsequently, non-radioactive CO2 was admitted to the sample, amounts of
CO2 adsorbed (c), CO2 in gas (h) and CO in gas (p) were measured and 14C
activity determined both in gaseous carbon dioxide (activity z) and in gaseous
carbon monoxide until the concentration of isotope (z/h) reached an equilib-
rium value. For some experiments, the sequence of admissions of CO and
- CO2 was reversed.
Each experiment (1-10 in table 1) was carried out with a newly prepared
sample (0.26 g) of reproducible properties.
3. RESULTS
3.1. Carbon monoxide on nickel oxide
3.1.1. Kinetics and extent of adsorption
Fig. 1 shows the courses of adsorption at various temperatures,
pressures and degrees of preadsorption by oxygen. In this graph is not
included CO adsorption on nickel oxide precovered by 118 x 1017 molecules
of oxygen, because both the extent and the rate of adsorption were very high.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 ??,
1
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: CIA-RDP80-00247A.004200070001-6
,
Table 1
, Experimental 'diaa.for adsorption'and exchange reactions
, I,.
2
3 ,
4 '
5
6
7
8
9
10
11
12
, 13
14
Exper-
iment
Temper-
ature
Sequence
Note '
-
d
g cc,
goo
ff.,
h
ac?
heo?
b
c,
m
n
.2*,'
? 3**
5**
6**
7**.
8**
, ? -
.. .
, .
20?C
20?C
55?C
1000C
I50?C'
20?C
20?C
200C
? o 20 C
20?C
CO-0O2
CO2-00
CO-0O2
CO2-CO
CO-CO2
CO-0O2 ,
CO-0O2
'
CO-0O2
,CO2-CO
CO2127.4
' .
.
'
?
preadsorbed
9.35 x 1017
mole pules 02
preadsorbed
118 x 1017
molecules 02
initial
pressure
.2.25 mm Hg
initial
pressure
2.26 ram Hg '
adsorbed
14c02
17.6
. ,
17.3
15.717.0
19.8
30.5
21.5
81.9
152.8
154.0
3.3
(0 .5)
2.9
(0.2)
1.1
(0.3)
13'7
(0.2)
2.15
(0.3)
1'7
(03)
.
??
(00:16)
25.0
(1.1)
21
3.
(1'2
14.2
17.8
17.9
20.4
13.4
4.1
560
*
,
2.05
1.55
1.92
1.00
1.30
1:48
0.83
0.186
3.7
4.4
5.3
5.0
2.6
2.9
3.3
116.7
155.7
2.4
0.085
0.091
0.11
0,26
0.87
0.20
1.31
0,074
0.074
2.18
4.7
2.5
3.7
1.5
2,9
4.9
2.9
5.4
*
448.3
445.6
381.7
272.0
142.0
401.1
374.7
514.5
582.0
290,0
2,2 P
1.6 P
3.0 q
.
6.3
4.3
6:5
2.5
4.1
4.0
0.7
50.2
0
4.7
4.7
5.4
8.9
17.5
9.9
70.0
' 21.0
24.7
106.0
STAT
- ? N
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: CIA-RDP80-00247A004200070001-R
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
LEGEND TO TABLE 1
5 STAT
. a, amount adsorbed in 200 min; expts. 1-9: adsorption of 11/40
expt. 10 -: adsorption of 11/402
d, amount desorbed in 30 mm by pumping after-the 200 min adsorption; upper values:
11/40 desorbed; lower values in brackets: 11/40 in desorbate
gm, amount of CO in gas at the end of exchange with carbon monoxide
hoo, amount of CO2 in gas at the end of exchange with carbon dioxide
Yc?, ,specific activity of CO in equilibrium with the adsorbate, cpm molecule -1 x 1015
-8-**0
zoo specific actiirity of CO2 in equilibrium with the adsorbate, cpm molecule-1 x 1015
/T.D.
b, amount of CO adsorbed during exchange with carbon monoxide
c, amount of CO2 adsorbed during exchange with carbon dioxide
p, amount of CO2 displaced from the adsorbate during exchange with CO
in, amount of adsorbate exchangeable with CO, computed from (2)
n, amount of adsorbate exchangeable with CO2, computed from (3)
The quantities a,d,g,h,b,c,p,q,m and n are in molecules x 10-17
The amounts p and q were pumped out of the system
* Exchange with CO was not carried out in this experiment
** Initial pressure 5.6 x 10-2 mm Hg
?Experiments 1-5 and 8-10 carried out on bare NiO.
Fig. 1. Adsorption qf carbon monoxide
et; adsorbed amount in molecules x 10-17 per 0.26 g sample
t time in min.
Initial pressure 5.6 x 10-2 mm Hg (1-4 clean nickel oxide):
'curve 1: 200, curve 2: 550, curve 3: 1000, curve 4: 1500, .
-curve 5: 21?, 9.35 X.1017 molecules 02 preadsorbed.
curve 6: 20?C, initial pressure 2.25 mm Hg, clean nickel
oxide 0 two independently prepared samples.
? -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 ? ? -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
6
STAT
Data for this experiment are presented in table 1 together with other data
concerning exchange reactions.
Kinetics of adsorption obeyed the well known equation (fig. 2)
20
50
1 ? . . ? L
- 3'
at
2
0
I i if telt(
3 0 S
rot .
?
100
200
Fig. 2. Plots of data from fig. 1 according to equation (1)
t+ to in min, logarithmic scale.
Sower figure: o correspond to curve 1 in fig. 1;
to 2, v, to 3; e, to 4; ?, to 5.
2.3
[log(t + to) - log to]
(a, adsorbed amount in molecules, t, time, b and to constants), the values
of 2.3 / b and to being summarized in table 2 in the discussion. Reproducibility
of results was very good as demonstrated in fig. 1 by a comparison of ad-
sorptions on different samples, including those taken from our earlier
work 4), open circles for 20?C adsorption and triangles for 55?C adsorption.
Since all the adsorptions studied with 14C0 and those- in earlier work with
non-radioactive CO gave the same results, one can exclude isotopic effect
on the extent of adsorption. At 150?C, measuring devices indicated the
presence of hydrogen during CO adsorption. This is most likely a product
_ of the reaction of CO with non-decomposed nickel hydroxide contained in
the sample under the catalytic action of nickel which is being produced by
the reaction of carbon monoxide with the lattice oxygen of NiO. The partial
. Pressure of CO in the gaseous mixture during adsorption was determined at
? .
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
Table 2
Constant of equation (1) and amount of oxidized CO (1+ n)
Temper-
-ature
0C
Conditions of experiment
2 3
b
to
1+ n
20
*
. 8.25
4
8.0
55
*
8.89
5
8.1
100
*
13.4
10
15.0
150
*
18.6
5
24.2
20
9.35 1017 molecules 02
preadsorbed
low pressure adsorption
9.21
3.5
15.8
20
clean NiO, initial pres-
100
20
77.6
'
sure of CO adsorption
2.25 mm Hg
1+ n and! in molecules x 10-17 per 0.26 g of adsorbent,
to in minutes
* Unless otherwise statedr adsorptions were carried out from
the initial pressure 5.6 x 10-2 mm Hg on a_clean nickel oxide.
this temperature from the radioactivity due to a non-condensable gas, which
is associated only with carbon monoxide. No hydrogen was evolved during
exchange experiments.
7 STAT
3.1.2. Exchange reactions
Exchange reactions were carried out after the adsorption of 14C0 lasting,
200 min. The results are presented in fig. 3 (exchange of 14C between ad-
sorbate and gaseous carbon monoxide) and in fig. 4 (exchange between ad-
sorbate and-carbon dioxide). The final values of the concentration of 14C in
CO(g) and in CO2(g) were equilibrium ones and did not change for another
ten hours; they are presented in table 1 (col. 7 and 9) with other data for
adsorption and results of analyses of the gaseous phase. All experiments,
except those showing the effect of pressure (table 1, experiments 8 and 9),
were carried out at low pressures (initial pressure of 14C0 adsorption
5.6 x 10-2 mm Hg, pressure during exchange reactions 10-2 - 10-3 mm Hg).
Effect of temperature was studied in experiments 1-5 for the range 20-150?C.
Effect of oxygen preadsorption on CO adsorption and exchange reactions at
20?C was investigated in experiments 6 and-?. Oxygen was always pread-
sorbed at 150?C because at this temperature all oxygen adsorbs irreversibly
in ionic form 17). Idsome experiments, the exchanging gas displaced a
certain amount of the other gas from the adsorbate into the 'gaseous phase;
this amount was measured and is given in column 12 of table 1.
Experiments with reversed sequence of the two exchanging gases
provide information-about mechanism of both exchange reactions (see
Discussion). The sequences are indicated in column 2 of table 1 as CO-0O2
and CO2-CO according to order of admissions of exchanging gases.
A note has to be added on the maximum on high pressure exchange (0)
in fig. 3. Theory 19) predicts very low maxima on all curves for exchange
_
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 ?
?
?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
' 8
? Fig. 3. Exchange reactions of 14C between adsorbate and carbon monoxide
" ?, concentration of 14C in CO (0, cpm molecule-1 x 1015,
f. time (min). For e scale reduced 10 x.
Pressure below 10-1 mm Hg, clean nickel oxide; 0 20?C -
(experiment 1) A 55?C (experiment 3), V 100?C (experiment 4),
? 1500C (experiment 5).
Pressure below 10-1 mm Hg, oxygen precovered nickel oxide:
??200C, 9.35 x 1017 molecules 02 preadsorbed (experiment 6).
- 4.200c, 118 .k 1017 molecules 02 preadsorbed (experiment?)
0 20?C, pressure about 2 mm, clean nickel oxide (experiment 8).
. , .
with carbon monoxide, which would not be detectable by the experimental
, techniques employed at present. The high maximum on the curve mentioned
is a diffusion effect: at 2 mm-Hg, diffusion coefficient of gas becomes much
lower than that at 5 x 10-2 mm; the' radioactive gas is rapidly exchanged
from the surface into the gaseous phase where its high concentration reaches
first the counter window and then slowly" diffuses into the other parts of
apparatus.
;
-Carbon dioxide on nickel oxide
Carbon dioxide adsorbs to a much higher extent than carbon monoxide.
Kinetics of adsorption were not measured because most of the adsorption is
-
a very rapid process. At low pressures the adsorption amounted to 4.17 x
1019 molecules (amounts a plus c in experiment 10 in table 1) with final
pressure 4.3 x 10-3 mm Hg. The quantity c (col. 11, table 1) shows the in-
fluence of Various factors on CO2 adsorption. There is a little effect when
carbon monoxide is preadsorbed at 20?C (row 1), a slight decrease with
pre:adsorption of oxygen (rows 6 and 7), decrease with increasing tempera-
ture -(rows 1-5)-and a slight enhancement of adsorption with increased pres-
? ? --- I- sure (rows 8,9). ? - ? ,
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 ?
DIA-RDP80-00247A004200070001-6
S TAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
4
?
?
9
io
Fig. 4. Exchange reactions of 14C between adsorbate and carbon dioxide
I, concentration of 14C in CO2(0, cpm molecule-1 x 1017
t, time (hours). For scale reduced 10 times.
The Same symbols as in fig. 3 used for varions experiments.
An experiment on exchange of adsorbed 14CO2 with inactive carbon
dioxide was also carried out and the results are given in table 1, row 10.
The half time of this exchange reaction was about one hour.
Measurements of electric conductivity revealed that on clean nickel
oxide, CO2 adsorbs without changing electric conductivity of NiO, whereas
adsorption of CO2 on NiO with preadsorbed oxygen leads to an increase of
conductivity linearly dependent on the amount of adsorbed CO2 (fig. 5).
so
9 STAT
OS
0
0 0 00 0 .0
0
t5
20
25
Fig. 5. Changes of electric conductivity with adsorption of CO2
a, conductivity, ohm-1 x 108; a, molecules x 10-18 ofCO2 adsorbed
0, clean nickel oxide
nickel oxide with 103.6 x 101:7 preadsorbed oxygen molecules. ?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
IA-RDP80-00247A004200070001-6 - - -
.7
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
10
4. DISCUSSION
STAT
4.1. Interpretation of exchanged amounts
is.evident from figs. 3 and 4 and from table 1 that under comparative
conditions the equilibrium amounts of 14C exchanged with CO decrease with
increasing temperature of 14C0 adsorption and with increased amount of
?preadsorbed oxygen, while reversed effect is observed in exchange with
CO2. This is a clear indication that a part of adsorbing 14C0 becomes
oxidized to carbon dioxide or carbonate which is then able to exchange
group CO2 with gaseous carboddioxide; this part is the greater the higher
is the temperature of 14C0 adsorption and the more oxygen is preadsorbed.
Further evidence is obtained from experiments with reversed sequences
of exchanging gases which reveal mechanism of the two exchange reactions.
If one part (M) of adsorbate exchanges only with CO and not with CO2 and
another part (N), only with CO2 and not with CO (mechanism I), the amounts
exchanged under comparative conditions with the two gases will not depend
on the sequence of the exchanging gases. If, on the other hand, the same
molecules of adsorbate were able to exchange 14C with both CO and CO2
(Mechanism II), one would always expect high radioactivity appearing in
the gas admitted first and low activity in the gas admitted next and, there-
-fore, great differences in the amounts exchanged with CO in the two ex-
periments in which CO is admitted before and after exchange with CO2; the
same would apply for exchange with CO2. ?
Results of experiments 1 and 2, and 8 and 9 in table 1 indicate that
-? mechanism I takes place, since the final isotope concentrations are dependent
to avery small extent on the sequence of exchanging gases. Quantitative
?
analysis of the data fully confirms these conclusions and allows us to deter-
'mine the number of carbon atoms in both parts M and N. At equilibrium, a
uniform distribution of radio-carbon among the mutually exchangeable
species must be attained. Part of exchanging gas, which was originally not
? -radioactive, adsorbs and this adsorbate (amount b in exchange with CO and
? c in_exchange with CO2, table 1) also participates in exchange*.
?? Let m and n be numbers of carbon atoms in parts M and N, and X the
" concentration of radio-carbon in adsorbing 14C0. Then, according to mech-
anism I, amount mX. of 14C labelled Molecules of adsorbate M become
- uniformly distributed, at the end of exchange with CO, among the gaseous
- CO((y??/geo) g, the original adsorbate Mky,?,/gaanz) and the newly adsorbed
CO(y4)b): -
b+ m) = mX . . . (2) exchange of M with CO
-
-Similarly, amount /7.2C of 14C labelled molecules of adsorbate N becomes
.
- 'uniformly distributed at the end of exchange with CO2 am-Ong the gaseous-
.. ?
- .
? * In exchange with CO2, a small fraction' of c molecules of CO2 adsorbate does not
'..'exchange as-is seen from experiment 10 in table 1. We shall assume for the pur-
- pose of discussion that all molecules c are exchangeable, being aware of a slight
inaccuracy introduced by the assumption.
?
?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 :
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
11 STAT
CO2((z..A.,4 hoc), the original adsorbate N n) and newly adsorbed
CO2 ((zoc/h.,) c):
c + = n2C (3) exchange of N with CO2
lue
Formulae (2) and (3) are identical with those obtained from kinetic treat- -
ment of mechanism I 19).
For mechanism II, the conditions for uniform distribution of radio-
carbon are different and depend on the sequence of exchanging CO and CO2.
When the sequence CO-0O2 is employed, one obtains
(ge3+ b.+ m +n) = (m + n) X
g?ce
z ?`'(it + c + m + n) = (m + n) -Y21'
he3
' while with the reversed sequence CO2-CO holds
Em(16+ c+ m+ n) = (m + n) X
co
zoo
= b +-m+-n) =(m+
goo
(4) exchange with CO
(5) exchange with CO2
(6) exchange with CO2
- (7) exchange with CO
Valuei.of m and n can be computed for mechanism I when inserting meas-
ured quantities in (2) and (3), and values m+n for mechanism II from (4) -
(7). These values must be self-consisent for the two reversed experiments,
if the respective mechanism operates.
The following excellent agreement is obtained between amounts m and
n in the two reversed experiments, calculated by means of (2) and (3),
which confirms that mechanism I is operating:
m(expt. 1) = 6.3 x 1017 molec.; (expt. 2) = 6.5 x 1017 molec.
n (expt. 1) = 4.7 x 1017 molec.; n (expt. 2) = 4.5 x 1017 molec.
nt(expt. 8) = 24 x 1017 molec.; n (expt. 9) = 24.7 x 1017 molec.
The value m.' (n') is a sum of amount m (n), determined from exchange reac-
tions (col. 14 table 1), and number of molecules of labelled adsorbate which
were displaced from the surface in preceding exchange (col. 12 table 1) and
removed from the system, so that they could not participate in subsequent
exchange.
Calculations according to (4)-(7) lead to many orders, discrepancies
and even negative values of m n for different sequences of exchanging
gases. Slight interference* of mechanism U with mechanism I would lead to
gross discrepancies and therefore our results show that mechanism I alone
is operating.
Apart from the two adsorbates M and N, there is a part (L) which is
not capable of exchange with either CO or CO2. The number of carbon
atoms in part L is 1 = a - d - m - n. The following diagramme shows the
magnitudes of 1, in, n and d for various conditions employed in the present
research.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
12
;, ?
20% pressure 2mrn (scale reduced VA)
20'
55'
10:0n,
A
AMU
20: osygen preodsorbed
A
so
AN
one per cent surface coverage
Diagramme F part M
LIlpart D
part N
MI part L
a, molecules x10-17
Number of experiment from up down: 8 , 1, 3, 4, 5, 6,7.
A
20
+ID
V
The nature of parts D (desorbable CO, amount d) and M is weakly
chemisorbed CO rwhich is in dynamic equilibrium with the gaseous carbon
monoxide as the rapid courses of exchange reactions indeed show. Part N
Is adsorbed CO2 or carbonate which exchanges whole group CO2 with
gaseous carbon dioxide and does not dissociate under the formation of CO
(no exchange with CO). Its character is the same as that of the product of
- CO2 adsorption, since the half time of its exchange with CO2 is close to
the half time of 14 --2(ads) - CO2(g) exchange, 1 hour. The presence of
carbonate formed with lattice oxygen at some stage of CO2 adsorption has
been proved by exchange of 180 between Ni180 and CO2 at 20?C on the
present preparation of NiO 20). Therefore, once adsorbed carbon dioxide
is formed, it will form carbonates with lattice oxygen and vice versa. The
non-exchangeable part L is a very strongly bonded or sterically hindered
adsorbate. Since its extent increases simultaneously with that of part N,
it also is very likely a carbonate or adsorbed CO2. Further evidence for
this is obtained from the exchange of (a
14C0-z with CO2, which shows
that a:certain part of the adsorbed carbon dioxide does not exchange with
CO2. Mode L is also a product of interaction of adsorbed oxygen with CO
and it may well be true that this very strongly bonded adsorption mode
actually poisons nickel oxide catalysts for the low temperature oxidation of
carbon monoxide.
_Since the starting nickel oxide in experiments 1-5 and 8-9 was free
? fr0ni adsorbed oxygen', the CO2(ads) or carbonate must be formed in these
experiments by the reaction of adsorbing CO with lattice oxygen of the
nickel oxide. This is a reduction process which produce lower valency
states of nickel cations, either Ni? or Ni, e. g.,
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 _
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 13
L
0
CO + 02- + Ni2+ CO ?
- 2( ads) Ni + (8)
lattice
CO2(ads) ?Luce COT- + 0 (9)
Electrons in these states are localized in d-shells of nickel and therefore
no change of electric conductivity occurs, as observed in 9) and also in the
? experiment quoted in the experimental part of this work. However, change
of electronic structure of cation at temperatures as low as 20?C is ,indicated
by the change of visible spectrum of the clean nickel oxide during the ad-
? sorption of carbon monoxide 10). The presence of lower valency states of
nickel will certainly contribute to the adsorptive properties of nickel oxide.
Particularly, considerable amount of adsorbate exchangeable with CO may
be bonded to these lower valency states, e. g., Ni-CO, carbonyl on atomic
scale. Increased amount of weakly bonded carbon monoxide at 150?C as -
compared with that at lower temperatures may well be due to this kind of
adsorption.
? 4.2. Reactions with adsorbed oxygen ions
Oxygen is adsorbed at 150?C in the surface of NiO as atomic anions 17).
Total increase of CO adsorption on nickel oxide with preadsorbed (150?C)
oxygen as compared with clean nickel oxide (fig. 1) is due to a large in-
crease in the amounts of modes N and L, accompanied by a marked decrease
in the amounts of modes M and D. When a sufficient amount of CO is allowed
to react with adsorbed oxygen, electric conductivity of nickel oxide drops
to the value observed on oxygen-free samples 9) and the colour of clean
nickel oxide is restored. The reaction
CO +.0Tado + Ni3+C (10)?2(ads) Ni2+
?explains both electric conductivity and 14C exchange observations.
It is interesting that reaction of CO2 with adsorbed oxygen ions takes
place with charge transfer also. Namely, the increase of conductivity during
CO2 adsorption on oxygenated samples can be interpreted as
+ +
CO2 + 0(ads) + Ni2 CO32- ? + 3 (12)
wittre Ni3+ ions are produced to increase conductivity, while the reaction
of CO2 with lattice oxygen (9) does not lead to any change of conductivity of
the clean nickel oxide.
Only a small part of the oxygen adsorbate participates in reaction (12),
as the effect on conductivity is very small compared with the effect induced
by a comparable amount of adsorbed oxygen.
Relation between adsorption kinetics and ;node of adsorption
For experiments on clean nickel oxide, correlation exists between the
values of the constant b (eq. (1)) and the amounts of CO oxidized by lattice
oxygen, 14- n, as shown in table 2. This means that adsorption decelerates
more slowly when reaction with lattice oxygen is enhanced, which happens
either on increasing the temperature or the pressure. This also explains
the crossing of kinetic curves for various temperatures in fig. 1. If the
. _ ? ? "Z. -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
:3IA-RDP80-00247A004200070001-6
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
?Q
14
constant b contributes to activation energy of adsorption according to
dq r Eo bRTql
dt A exP RT
(13)
STAT
(q, adsorbed amount in time t, Eo and A, constants), then the above men-
tioned correlation means that increased extent of reaction with lattice oxygen
makes the whole process less dependent on adsorbed amount of CO, in
other words, the reaction with lattice oxygen is more uniform with respect
to activation energy than the production of CO(ads). A similar but smaller
effect is observed with CO adsorption on oxygen precovered NiO.
4.4. Comparison with other work
The present results show that carbon monoxide adsorption at 20?C and
5.6 x 10-2 mm Hg on a clean NiO is very low (-0.6% surface coverage in
200 min adsorption) and yet, part of this process is the result of reaction
which has already taken place with lattice oxygen. The effect of increasing
pressure is to increase both the extent of reaction with lattice oxygen and
the amount of weakly adsorbed CO. The latter finding confirms considera-
tions 15) which explain the differences in the amounts of CO adsorbed found
in various works, in terms of low adsorption heat of CO and reactions with
lattice oxygen. There is, however, another factor strongly influencing the
CO adsorption at low pressures, namely, adsorbed oxygen present on the
surface of NiO. This explains why other authors 205) found, at comparable
pressures, amounts adsorbed higher than that in the present work. At
higher pressures (about 40 mm), the effect of oxygen preadsorption on CO
adsorption is small 3) because reaction with lattice oxygen at these pres-
sures becomes very extensive and numbers of adsorption sites on the sur-
-faces with and without adsorbed oxygen are not substantially different.
ACKNOWLEDGEMENT
?
The authors wish to thank Professor R. Brdi6ka, Director of the In-
stitute of Physical Chemistry in Prague, for his constant interest in their
work. They are also indebted to Dr. F. S. Stone from Bristol University
and Dr. J. Haber from Krakow Mining Academy for helpful discussion of
their work.
REFERENCES
1) S. Z.Roginski and T. S. Tselinskaya, J. Phys. Chem. USSR 21 (1947) 919,22 (1948)
1360.
- .
2)14.P.Keier, L,N.Kutseva, lzv. Aka& Nauk USSR, Otd. chim. nauk 1959, 797.
3)- S. J. Teichner, R. P. Marcellini andy.Rue, Advances in Catalysis 9 (1957) 458.
4) K. Klier, Coll. Czech. Chem. Commun. 28 (1963) 2996.
5) YtKubolcawa and O. Toyama, Bull. Chem. Soc. Japan 35 (1962) 1407.
6) R. M. Dell and F. S. Stone, Trans. Farad. Soc. 50 (1954) 501.
? . .
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
15 STAT
7) P. Ch. GraveIle, Dissertation; University of Lyon, May 1963.
8) M. Courtois and S. J. Teichner, J. Chim. Phys. 1962, 272.
9) R. P. Marcellini, R. E.Ranc and S. J. Teichner, Actes du deuxieme Congres inter-
national de catalyse, Technip, Paris 1960, 289.
10) A. Bielanski, ?J. Deren, J. Haber and J. Sloczynski, Zeit. Physik. Chem. N. F. 24
(1960) 345.
J. Deren, J. Haber and J. Sloczynski, Bull. Acad. Sci. Pologne 8 (1960) 391, 8
(1960) 399.
11) G. Parravanosand C. A. Domenicali, J. Chem. Phys. 26(1957) 359.
12) K. Klier, Kinetika i Kataliz 3 (1962) 65.
13) P. Rue, Dissertation, University of Lyon, June 1963.
14) R.Ranc, Dissertation, University of Lyon, March 1963.
15) F. S. Stone and K. Klier, Actes du deuxieme Congres international de catalyse,
Technip,.Paris 1960, 323.
16) T. I. Barry and K. Klier, Discussions Faraday Soc. 31 (1961) 219.
17) K.Kuchynka and.K.Klier, coll. Czech. Chem. Commun. 28 (1963) 148.
18) K. Kuchynka, Dissertation, Czechoslovak Academy of Sciences, PrIgue, October
1963.
19) K. Klier,. Coll. Czech. Chem. Commun. (1964) to be published.
20) K. Klier and Z.Herman, Coll. Czech. Chem. Commun. (1964) to be published.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
ON THE MAGNETIC PROPERTIES AND CATALYTIC
ACTIVITY OF NICKEL OXIDE
D. MEHANDJLEV and G. BLIZNAKOV
'Institute of General and Inorganic Chemistry,
Bulgarian Academy of Sciences, Sofia, Bulgaria
Abstract: An attempt is made to present additional information on the reaction
CO + 402 -? CO2 with NiO as catalyst and to explain the contradictory results ob-
tained by a number of authors, concerning the activation energy and the influence
of various additives. Tc this end, the catalytic activity of pure nickel oxide was
studied simultaneously with its magnetic properties and the content of oxygen in
excess of the stoichiometric amount. It was shown that when the excess of oxygen
is increased, the specimens of nickel oxide pass from antiferromagnetic into pa-
ramagnetic state with traces of ferromagnetism. This leads to a sharp change
of the activation energy of reaction. While for the antiferromagnetic specimens
the activation energy is of the order 4-9 kcal/mole, for the paramagnetic ones
it is about 13-17 kcal/mole. The rate constant changes in an analogous manner.
The magnetic moment calculated for the paramagnetic specimens shows that
.+3
Ni is formed at the expence of oxygen in excess of the stoichiometric amount,
the crystal lattice being preserved, as is evident from the X-ray analysis.
1. INTRODUCTION
The catalytic and adsorption properties of NiO have been studied thor-
oughly. The oxidation of CO to CO2 on a NiO catalyst (pure and with addi-
tives) is investigated in the works 1-5). The values of the activation energy
of reaction presented in these publications are rather contradictory. While
for pure nickel oxide Keier et al. 3), Kutseva 5) and Parravano 1) give low
values of the activation energy (3-5 kcal/mole), Schwab and Block 2) present
about 16 kcal/mole. Further, Keier 3) and Parravano 1) state that introduc-
tion of monovalent ions (Li+, Ag+) increases the activation energy, while
trivalent ions decrease it. Schwab and Block 2) maintain the opposite: tri-
valent ions increase and monovalent ions decrease the activation energy.
The general conclusion from the studies on the magnetic properties of
nickel oxide 8-13) has been that stoichiometric NiO is antiferromagnetic
with an antiferromagnetic temperature of 3770C 7). With increasing content
of oxygen in excess of the stoichiometric amount, nickel oxide becomes
ferromagnetic 8) or displays non-compensated antiferromagnetism 13).
The addition of Li ions, according;to Perakis 8), leads to ferromagnetic
properties, As to the character of the curve: susceptibility-composition of
nickel oxide, there is contradiction only in the region of slight deviations
from the ,stoichiometric compound. While according to Klemm and Haas 6),
specific susceptibility in the above mentioned region increases sharply, the
Alta of Shirokov 13) show a value which is almost ccnstant.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: _ STAT
.-;1A-RDP80-00247A004200070001-6 ? - - "
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
STAT
- Studies on other physical-chemical properties of nickel oxide have also
? been published. The conductivity of nickel oxide with respect to additives is
studied in the works 14,15). The electronic properties of nickel oxide are
investigated in 16-10.
- With the purpose of explaining the mechanism of the reaction CO +
- 02 - CO2, some authors studied the adsorption of fhe initial products on
NiO and that of the products of reaction 20-27). Irrespective of the great
number of studies, the mechanism of this reaction has not been established.
? Still more uncertain is the reason for the different values of activation
energy obtained. While the authors of most of the magnetic studies take into
consideration the excess of oxygen, others, especially authors investigating
the catalytic properties of NiO (the work of Kutseva 5) excepted), only give
the conditions of preparation, considering the nickel oxide obtained to be
stoichiometric NiO. Assuming that the catalytic properties of NiO depend
on its content of oxygen in excess of the stoichiometric amount, this ex-
plains in all cases the differencies between the results of the investigations,
if the nickel oxide obtained contained various amounts of oxygen. Consider-
ing the data, published?for the activation energy, it is evident that they are
grouped about two values: some authors give a high value, about 14-18 kcal/
* mole, 'While others present a lower value of about 3-8 kcal/mole. Accord-
. Ing to the data obtained by Kutseva 5), with changing the content of excessive
- oxygen, the activation energy changes but little. However, small amounts
of lithium or chromium added sharply change the activation energy, but with
the increase of additives, the change of activation energy is insignificant.
For this reason, the present investigation had the purpose of studying
the catalytic activity of nickel oxide together with its magnetic properties,
taking into consideration the deviations from stoichiometry as well.
-1'2: EXPERIMENTAL
. .
_
Specimens of nickel oxide, containing various amounts of an excess of
? .oxygen, were obtained by decomposing basic nickel carbonate in air, .at a
temperature, varied from 3000 up to 1000?C, in the course of 3 to 24 hours.
.The ratio between nickel andnxygen in the specimens was determined corn-
Plexometrically, as well as gravimetrically, through the content of nickel
in each specimen. Magnetic susceptibility was measured in an apparatus
? . for magnetic investigations by Gouy's method, in the temperature interval
? from 0?C to 100?C, with an accuracy of 1%. Catalytic activity was studied
in a 'static apparatus, in the temperature interval from 60? to 100?C and
? pressures from 8)< 10-1 to 4 x 10-1 mmHg. A stoichiometric mixture of
" CO and-02 was used for the experiments and CO2 produced was gathered in
. a trap of liquid nitrogen. Before each experiment, the specimen was heated ,
out-for an hour at 100?C, under a pressure of 10-4 mm Hg. A higher tem-
.
perature of heating out led to a decrease of the excess of oxygen. The acti-
. '-vation energy and the pre-exponential coefficient were calculated after
-'' finding the dependence lg T 1 - 1/T (Here T 1 is the half period and the
absolute temperature). .
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
The specific surfaces of the Specimens were determined by means of
low-temperature-adsorption of nitrogen by the BET method.
3. RESULTS
3 STAT
200 400 603 800 1000 1?C
?
Fig. 1. Atomic ratio (0/Ni) between oxygen and nickel in
dependence with the temperature of preparation of
the specimens. Period of heating for the specimen,
? obtained at 5000; -0- 3 hours; -0- 24 hours;
-0-, 80 hours.
". -
Fig. 1 gives the dependence between the content of oxygen in excess of
the stoichiometric amount and the temperaLure of decomposition of the
specimens. It is evident, that the excess of oxygen decreases with the in-
crease of temperature; at 1000?C stoichiometric nickel oxide was obtained.
The specimen decomposed at 500?C displays, a different character. While
with the rest of the specimens, the period of heating has practically no
effect upon the ratio between nickel and oxygen, in the above mentioned
specimen the-content of oxygen decreases with longer heating. Further-
more, the sample decomposed at 500?C containing oxygen in excess amount-
ing to 0/Ni = 1.19 is paramagnetic, while the remaining two samples de-
composed at the above temperature are antiferromagnetic.
. 'shows the specific susceptibilities of the specimens as functions
of the content of oxygen in excess of the stoichiometric amount. The specific
. .
? susceptibility decreases with the decreasing content of excessive oxygen.
-7
In the interval 0/Ni = 1.15 to 1.12, the functions change their sign. While :
7 :above lA 5 susceptibility decreases with temperature, i. e. the specimens
Declassified in Part - Sanitized Copy Approved for Release 2614/03/04 :
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
104 112 120 128 136 t44.o_
Ni
STAT
Fig. 2. Specific susceptibility (x) of the specimens in
? dependence with the atomic ratio (0/Ni) and
the temperature of measuring the susceptibility.
-0- 250C; -e- 500C; -e- 800C .
?
are paramagnetic with traces of ferromagnetism (their susceptibility is
only slightly influenced by the magnetic field), below 1.12 susceptibility
increases with the increasing temperature, i. e. the specimens are anti-
ferromagnetic. In the interval between 1.12 and 1.15 susceptibility in-
creases slightly with temperature.
For the paramagnetic specimens were calculated the effective magnetic
moments and Weiss' constants, from the dependence 1/g - T, shown in
fig. 3. It is evident that the effective magnetic moment is lower for the
specimens with higher content of oxygen. The latter being in agreement
with the assumption of Ray 28) that Ni with valency higher than 2 has a
low magnetic moment. For the specimen Ni01.19, the magnetic moment
corresponds to the theoretically calculated, under the condition that the
orbital component is frozen.
The change of sign of Weiss' constant is also very characteristic. The
results from the determination of the activation energy of reaction in de-
- pendence with the composition of nickel oxide are shown in fig. 4. While
for the paramagnetic specimens, with 0/Ni above 1.15, the activation
-energy is of the order 13-17 kcal/mole, transition into the antiferromagnetic
state-leads to a sharp decrease of activation energy, down to 4-9 kcal/mole.
? The dependence between the activation energy and the temperature of de-
- composition of the specimens is analogous to the dependence between the
activation energy and the composition of the nickel oxide. In the specimens
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 .
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 5 STAT
I J
120 1.40 1.50
Fig. 3. Dependence between the c`fective magnetic
moment (Peff) in Bohr magnetons, Weiss'
, constant and the atomic ratio (0/Ni) of oxygen
to nickel.
Fig. 4. Dependence between the activation energy of the
reaction CO + 402 -? CO2, (EA) in kcal/mole and
the atomic ratio (0/Ni) of oxygen to nickel.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
STAT
decomposed at 500?C, a change in the ratio 0/Ni by only 0.04, with almost
equal specific surfaces, leads to a change of activation energy amounting to
about 10 kcal/mole. The results obtained for the pre-exponential coefficient
are analogous. Transition from one magnetic state into another calls forth
an analogous sharp change here of the pre-exponential coefficient.
? Fig. 5. Dependence of the specific surface (S) of
the specimens in m4/g and the temperature
r of preparation.
?
Fig. 5 shows the specific surfaces as a function of the temperature of
- decomposition. The change of specific surface at lower temperature is
. rather great (almost twice the initial value) for the specimens decomposed
at 300? and at 400?C. X-ray analysis was carried out, in order to study the
? crystal lattice of the specimen's. The following results were obtained for
the parameter of the lattice in dependence of the temperature of decompo-
sition of the specimens. For the specimen decomposed at 300?C the para-
meter obtained was 4.211 A, for the specimen at 400?C - 4.180 A, 500?C -
4.186 A, 600?C - 4.171 A, 700?C - 4.169 A and 1000?C - 4.184 A. There is
a tendency toward a slight deformation of the lattice. The lines of nickel
oxide are preS'erved in all the specimens. The lines in the specimen de-
composed at 300?C are not very distinct, but they are still there, i. e. the
acceptance of oxygen and likely the formation of Ni203 takes place without
the lattice of NiO being distorted. These results coincide with the data of
Shimomura 8). ? ?
Declassified in Part: Sanitized Copy Approved for Release 2014/03/04 :
CIA-RDP80-00247A004200070001-6 - ?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
4. DISCUSSION
It is evident from the above results that in the oxidation of CO to CO2
and probably in other reactions also, the magnetic state of the catalyst is
an important factor, determining its catalytic activity. When nickel oxide
is antiferromagnetic, the activation energy is low (4-9 kcal/mole); the reac-
tion takes place with a high rate. When nickel oxide is paramagnetic, the
activation energy is of the order 13-17 kcal/mole. The fact that with the
monotonous change of the content of oxygen in excess of the stoichiometric
amount, the catalytic parameters (activation energy, rate constant, etc.)
are sharply changed and this sharp change coincides with the transition
? from paramagnetism to antiferromagnetism, shows that the excess of oxygen
affects mainly through changing the magnetic state of the catalyst and less
by changing the number of active centres or the crystal structure. It is
evident, this is a case of magneto-catalytical effect, which could also be
observed at a constant temperature, if a possibility exists for a definite
way of changing the magnetic state of the substance. The increase of the
excess of oxygen disturbs the strong interaction in the antiferromagnetic
specimens and they become paramagnetic with traces of ferromagnetism,
due to the incomplete suppression of the strong interaction. Transition from
the antiferromagnetic to the paramagnetic state can also take place with
raise of temperature. Therefore, in discussing the results, the history of
nickel oxide must also be taken into consideration, i. e. its composition and
the temperature of determining its catalytic activity. These conclusions
are confirmed by data of other authors and explain the contradictions in
their results. For example, Kutseira 5) has worked with specimens, showing
0/Ni up to 1.12 and temperatures up to 200?C (the data are calculated from
1 the graphs in the publication). Hence, the specimens are antiferromagnetic
?and it should follow that the activation energy would be lower. The experi-
mental data show values from 3 to 5 kcal/mole.
For the specimen of pure nickel oxide Parravano 1) gives two values
for two different temperature intervals. This can be explained, taking into
consideration that, according to the method of preparation, the specimens
should be antiferromagnetic and the activation energy lower. For the tem-
perature interval 100-1800C, it actually is about 3 kcal/mole. The increase
of activation energy in the second temperature interval is due to reaching
the antiferromagnetic temperature and the transition into the paramagnetic
state. A lowered antiferromagnetic point could be expected here, because
it depends upon the dimensions of the particles as well 10). Schwab's speci-
mens are also antiferromagnetic, but because the catalytic measurements
were made at a higher temperature, the activation energy ought also to.be
higher, as is actually observed. -
As to the specimens with various additives, the foretelling on the activ-
- ity of the catalyst could be made, if the ratio between Ni, 0 and the element
added, as well as the magnetic state of the catalyst, are known. We are of
the opinion that when the specimen' is antiferromagnetic and the catalytic
reaction takes place below the antiferromagnetic temperature, every kind
of additive should lead to an increase of the activation energy, if it disturbs
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 -
STAT'
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
the strong interaction. With the paramagnetic state the problem is more
complicated and depends on whether this state has been reached through an
additive or is due to a. content Of oxygen in excess of the stoichiometric
amount, or after increasing the temperature. The data for the magnetic
moments calculated show that Ni+3 is obtained at the expence of the excess
of oxygen. The crystal structure however is preserved almost to the full
transformation of NiO into Ni203 (the specimen with 0/Ni = 1.44). Here, a
deformation of the lattice is observed and the lines of the specimen decom-
posed at 300?C are already slightly diffused. Hence, in this case the exist-
ence of two phases can no more be assumed.
The authors wish to express their thanks to Mrs. K. Hristova for the
complexometric and gravimetric analyses and to Mr. I. Tsolovski for the
X-ray studies.
REFERENCES
STAT
1) G.Parravano, J. Am. Chem. Soc. 72 (1953) 1452.
2) G.M.Schwab and J. Block, Z. Elektrochem. 58 (1954) 756.
3) N.P.Keier, S. Z. Roginsky and N.Sazonova, Doklady Akad. Nauk SSSR 106 (1956)
859.
4) A. Bielanski, Z.Phys. Chem. (DDR) 24 (1960) 345.
5) L.N.Kutseva, Doklady Akad. Nauk SSSR 138 (1961) 409.
6) W. Klemm and K. Haas, Z. anorg. allgem. Chem. 219 (1934) 81.
7) M. Ft5ex and C.H. La Blanchetais, Compt. rend. 228 (1949) 1579.
8) Y.Shimomura, I. Tsubokawa and M.Kojima, J. Phys. Soc. Japan 9 (1954) 521.
9) N.Perakis, A .Serres and G. Parravano, Compt. rend. 242 (1956) 1275.
10) J. T .Richardson and W.0. Milligan, J. Phys. Chem. 60(1956) 1223.
11) W. L.Roth, Phys. Rev. 110 (1958) 1333.
12) T.I.Barry, Atomic Energy Research Establ. (Great Britain) 11-3493 (1960) 1.
13) Yu. G. Shirokov and I. P.Kirilov, Izvest. Ucheb. Vysshikh Zavedenii, Khim. i Khim.
Teldmol. 4 (1961) 599.
14) A. Bielanscki, Y.Deren and D.Haber, Trans. Faraday Soc. No. 1 (1962).
15) G. M.Schwab, J. Appl. Phys. 33 (1962) 426.
16) G.H.Wanmer, Element of solid state (Cambridge, 1959) p. 169.
17) S.Van Houten, Phis. Chem. Soc. 17 (1960) 7.
18) K.Hauffe, Z. Elektrochem. 56 (1952) 366; 57 (1953) 762.
19) K. Hauffe, Semiconductor surface physics.
20) N.P.Keier, Problemy Kinetiki i Kataliza, Akad. Nauk SSSR No. 10 (1960) 73.
21) K.Hauffe, Z. Elektrochem. 65 (1961) 321.
22) Y.Kubokawa, Bull. Univ. Osaka Perfect. A9 (1961) 45.
23) M.Courtois, Chim. mod. 6 (1961) 195.
24) K.Klier, Kinetika i Kataliz 3 (1962) 65.
25) S. J.Teichner, K.P. Marcellini and P. Ru, Adv. Catalysis 9 (1962) 458. -
26) I.Matauure, Technol. Repts. Kansai Univ. No. 3 (1961) 63.
27) K.Kuchynka and K.Klier, Collection Czech. Chem. Commun. 28 (1963) 148.
28) P.Ray, A. Bla.duri and B. Sarme, J. Indian Chem. Soc. 25 (1948) 51.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
3IA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
INTERACTION BETWEEN ADSORBED MOLECULES AND
THE MECHANISM OF THE ELEMENTARY ACTS IN THE
DEHYDRATION OF ALCOHOLS ON OXIDE CATALYSTS
V. E. VASSERBERG
The N.D.Zelinsky Institute of Organic Chemistry, Academy of
Sciences, Moscow, USSR
Abstract: The presence of foreign substances greatly affects the kinetics of al-
cohol dehydration in the adsorption layer over A1203, leading either to retardation
or acceleration (conjugated dehydration) of the reaction. The mutual effect on
the adsorbed molecules can also be observed at ordinary pressures, but at low
temperature, in the "pre-catalytic" region. In particular it is manifested in the
newly discovered reaction of what may be called isotope-radical exchange,
wherein a radioactive label is transferred under conditions of catalytic dehydra-
tion. The degree of transfer depends upon the nature of the catalyst. The charac-
ter of the catalytic heterogeneity of the surface and the number of active centres
have been determined for a ?lumber of A1203 catalysts. Active intermediate com-
plexes have been shown to be formed in the adsorption layer during the dehydra-
tion reaction. These possesb many properties characteristic of free radicals,
catalyzing for instance the para-ortho hydrogen reaction, polymerization of ole-
fins, etc. The existence of two dehydration mechanisms has been suggested: (a)
a high temperature mechanism whereby olefins are formed directly from prima-
ry complexes and (b) a low temperature "collective" or polymolecular mechanism
associated with the formation and decomposition of secondary complexes result-
ing from interaction of the adsorbed molecules.
1. INTRODUCTION
It is generally accepted that all heterogeneous catalytic reactions take
place in the adsorption layer formed by the reactant molecules on the cata-
lyst surface. Hence in order to judge the true mechanism of the reaction and
the nature of its elementary acts one can not merely confine oneself to a
study of the reaction kinetics by determining the changes taking place in the
compositon of the liquid or gaseous phases, but the behavior of the adsorbed
molecules themselves must be accounted for. For this one can utilize IR
spectroscopy (as is now being done, following the pioneer work of Tereninl)
and Eishensz) or direct kinetic determinations of reactions in the adsorption
layer. Despite its unquestionable merits the latter method has not been used
to a great extent. Among the few works in this direction mention should be
made of the classical studies by Dohse et al. 3)carried out as far back as in
? ? 1929-1933 and also some later ones4-9). In recent years this method has
been- systematically employed in A. A. Balandin's laboratory for studying the
dehydration of lower aliphatic alcohols and ethers on a number of A1203 ca-
talysts of different origins.
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: STAT
CIA-RDP80-00247A004200070001-6
. - ,
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
2
? ,
beclassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 _
2. EXPERIMENTAL
STAT
2.1. Mutual effect of the adsorbed molecules and the inhomogeneity of
alumina
Our method, a modification of Dohse's technique, is as follows: A
sample of the catalyst (from 0.1 to 7.0 g) is put on the pan of a springlever
balance 10, placed in the circulating system of a vacuum set-up and degass-
ed at 4500 until a pressure of 10-5 mm Hg has been reached. Then at the
temperature of the run (from 400 for tert. C4H9OH until 2200 for C2H5OH)
a certain amount of reactant vapor (measured both manometrically and di-
rectly from the gain in weight of the catalyst due to adsorption) is introdu-
ced into the system. After that the rate of reaction in the adsorption layer
is followed by the increase in the pressure of the olefin formed, water re-
maining firmly bound to the catalyst.
In this way many pecularities of the reaction playing an essential part
in its mechanism Could be revealed, and those which are frequently hidden
when carrying out the reaction under ordinary conditions, i.e. in a flow
system at atmospheric pressure. For instance, it was found that alcohols
are dehydrated in the adsorption layer at measurable rates (the half decom-
positon period 7-0.5 = 2 - 30 min) already at quite moderate temperatures
(100-1400 for iso-C3H7OH) i. e. when this reaction is not noticeable under
ordinary conditions. Furthermore catalysts most active in the monolayer
reaction may not prove to be optimal under atmospheric pressure 11). Also
there is no regular relation between the corresponding activation energies.
Interesting data have been obtained in the adsorption layer study of the
concurrent dehydration of isopropanol and ethanol. The reaction turned out
to be always hindered for iso-C3H7OH, but accelerated in a number of
cases for ethanol. This phenomenon of conjugated catalytic dehydration has
been explained by the interaction between the active intermediate complexes
formed by the respective alcohols in the catalyst surface under the reaction
conditons 12). Later, together with Balandin and Silskova we decided to see
whether such a conjugated process occurs under ordinary conditions, car-
rying out the reaction at low temperatures, i.e. in the "pre-catalytic" re-
gion. Dehydration of C2H5OH, iso-C3H7OH and their binary mixtures was
studied in a recycling system 13) at 150-2500 and 1 atm pressure, -the gas
being analyzed by partition chromatography in a stream of CO2. It was
found that at '1500 the reaction does not take place at all; at 200-225? the
over-all amount of ethylene formed from an equimolar mixture of the alco-
hols was by 2.3-4.9 times larger than that obtained from pure C2H5OH. At
2500 the amount of C2H4 was less than from pure ethanol. The amount of
C2H4 was just what should have been expected from the trivial assumption
:that dehydration, of both alcohols takes place independently, the relative
yield of the reaction products being determined only by their competition
for the active centers. The amount of ethylene formed is still less for equi-
inolar-mixtures of C2H5OH with CH3OH or H20.
In previous studies 14, 15) examples were presented of the mutual ef-
fect of adsorbed molecules on the rate and direction of reaction for a large
number of binary systems. It was found that even small amounts of, say,
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
3 S TAT
acetone, dioxane, acetonitrile and ethyl acetate can almost completely inhi-
bit the dehydration of 5-7 times the amount of iso-C3H7OH in the adsorption
layer. Further in a number of cases, when the degree of coverage was very
low,polymerisation of the olefins was observed, indicating that surface dif-
fusion of the adsorbed molecules takes place.
In the co-operationwith Balandin and Georgieyskaya the character of
the inhomogeneity of the surface of a number of alumina catalysts and the
shape of the active center distribution curve were determined. In this study,
instead of the usual integral method of Dohse (determination of the over-all
rate of decomposition of the adsorbed alcohols as a function of the degree
of surface coverage a a new, differential method was employed. This was
aimed at determining the dehydration kinetics of a small amount of a given
alcohol adsorbed on various regions of the surface. This is achieved by pre-
liminarily blocking a part of the surface by various amounts of a pread
sorbed inert substance (water) (fig.1.).
? Fig. 1. The half decomposition period of the adsorption layer
dehydration iso-C3H7OH on various alumina catalysts
as function of amount of preadsorbed blocking substance
(curves 1-6: water, curves 7-8: dioxane).
land 7 - Catalyst no. 1, 1400; 2 - Catalyst no. 2, 1100;
3 - Catalyst no. 3, 1200; 4 - Catalyst no. 4, 1200;
5 - Catalyst no. 5,1200; 6 and 8 - Catalyst no. 6, 1300.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 --------
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
4
STAT
On the basis of the "effective specific adsorption area" co of the given cata-
lyst 16) and the value of co ordinarilly accepted in the literature for water
w = 10.2 A2); the number of active sites co and the fraction of the active sur-
face Sad in the various catalysts could be determined from the position of
the point of inflexion of the curves on fig. 1. The following alumina catalysts
were employed: (1) commercial grade A-1; (2) precipitated from sodium
aluminate solution by CO2 at 00; (3) prepared by the hydrolysis of aluminium
isopropilate; (4) precipitated from aluminium nitrate solution by aqueous
ammonia at 100?; (5) ditto, but with Zn(NO3)2 added to the solution to obtain
0.5% (molar) ZnO referred catalyst; (6) ditto, with la ZnO.
It can be seen that Z lies between 8 x 1019/g and 1.7 x 1020/g (which
corresponds to 3.6 - 5.8 x 1017/m2 of the catalyst surface) and Sact from
12% to 30% of the overall surface. Further experimental investigations of
this question led to quite unexpected results. Thus it was found that dehy-
dration of iso-C3H7OH practically stops altogether if dioxane molecules are
present in 15-th to Ath the amounts necessary to cover all the active centers?
(we had determined for the given specimen). Furthermore for the same tem-
perature T0.5 can drop to ith - ilath the former value if the reaction is
carried out at very greatly diminished surface coverage (0.004 - 0.02
mmole/g instead of usually employed values of 0.06 - 0.1 mmole/g). In this
case the break in a - T 0.5 curve (similar to that determined by Dohse),
sets in already at 0.06 mmole/g. This is about ird the value we had obtain-
ed for the same catalyst by the differential method, but on the other hand it
is in very good agreement with that obtained by Dohse for bauxite
(2 - 3 X 1019/g). It thus follows that the lifetime of the intermediate complex
and hence the reaction rate depends not only upon the activity of the center
and the nature of the adsorbate, but also upon the presence or absence of
other molecules in its immediate vicinity. In other words the rate of dehy-
dration of "isolated" molecules is higher than that of the molecules with
neighbors.
-Similar acceleration of the reaction was observed'for iso-C3H7OH in
the "pre;catalytic" region when working in a recycling system under dilution
with argon, nitrogen or CO2.
The low temperature decomposition of ethanol (at 120-1400) may serve
as an example of the inverse mutual interaction of adsorbed molecules we
observed on some alumina catalysts. The reaction in this case assumes a
noticeable rate only after a certain threshold value of the surface concen-
tration of alcohol (fig.2.).
2.2. Radical-like intermediate complexes in dehydration catalysis
The formal similarity of many of the above-mentioned phenomena with
free radical reactions led to the assumption that under the conditions pre-
, vailing in our experiments the reactants forth radical-like intermediates on
thesurface of the catalyst. Such types of intermediates in heterogeneous ca-
talysis have been postulated by numerous workers (Zelinsky and Shchuikin 17),
Eidus 18), Temicin19), Semmv, Voevodsky and Vorkenstein20), Myasni-
kov 21), Kemball 22) et al. "). Direct experimental evidence of their exis-
tence has been provided in the works of the author together with Balandin
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 - ? -?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
5
and Davydova 24) and Davydova and Georgievskaya 25), when it was shown
that the active intermediates formed in the MgSO4 and A1203 catalysed de-
hydration of alcohols and ethers possess such characteristic properties of
free radicals as the ability to induce the para-ortho hydrogen conversion.
Fig. 2. The threshold value of the surface concentration of alcohol in
the low-temperature adsorption layer dehydration of C2H5OH:
1 - Catalyst no. 1, 1400; 2 - Catalyst no. 2, 1400; 3 - Catalyst
no. 3, 1200.
Later together with Georgievskaya it was shown that the adsorbed layer de-
hydration of iso-C41190H on some alumina catalysts was accompanied by i-
nitiated polymerization, also a characteristic property of free radicals. If
propylene or isobutylene is pieliminarily added to the system (at pressures
of 0.05 - 0.1 mm HO dehydration is accompanied not by a monotonous in-
crease in pressure resulting from more of the isobutylene being formed from
the alcohol, but the pressure falls instead below the initial value and begins
to rise after. The polymerization process associated with the pressure
drop is not observed before dehydration sets in and rapidly ceases on its
completion (fig.3.).
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 . . - ?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
6
STAT
2. 3. Catalytic radical-isotope exchange and interaction between adsorbed
molecules
? Interaction between adsorbed molecules in the pre-catalytic region was
demonstrated by Balandin, Levy and the author 26, 27) with the aid of a ra-
diochemical method.
Fig. 3. Initiated polymerization of isobutylene by concurrent adsorption
layer dehydration of iso-C41130H on A1203 + 10% ZnO catalyst
at 1200. 1 - Pure iso-C4H9OH; 2 - Isobutylene + iso-C41190H.
Vertical line - addition of the isobutanol.
We concluded that such interaction if it takes place could lead to an ex-
change of particles or groups between the reacting molecules and hence,
with tracer atoms to a change in the isotopic composition. For this purpose
14H30CH3 (C14H30H in some runs) was chosen as the label donor because
of its stability under the experimental conditions, since, while forming the
same intermediate surface complexes as the other alcohols, it cannot under-
go dehydration. As second components of the mixture, with the dimethyl ether
to be passed over the alumina catalyst , were other alcohols, ethers, ketones,
olefins, etc. In many cases this gave rise to a new type of reaction, radi-
cal-catalytic exchange of the label between the methyl ether and other com-
pounds. Later we showed that the degree of exchange depends upon the na-
ture of the catalyst. Thus on passing a mixture of oct-l-ene and C14H30H
in a ratio 5:1 at 225? under various catalysts, namely A1203 treated with
potassium hydroxide, A1203 + Fe203, untreated A1203, alumosilicate
cracking catalyst and Ca3(1304)2, the specific radioactivity of octene in the
catalysate was 0, 50, 60, 190, 170 and 25 imp/min-mg Bac03, respectively.
3. DISCUSSION OF RESULTS
The data on the surface heterogeneity of the alumina catalyst led to the
following conclusions.
1) It is not at all a generale rule that all active centers should be iden-
tical as found by Dohse. One can see from fig.2 that the initial part of the
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 :
.7,1A-RDP80-00247A004200070001-6 -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
7 STAT
curve is parallel to abscissa axis only for one catalyst.
2) The large differences in activity of the catalysts (the 7-0.5 value at
1200 for isopropanol on various specimens varied between 6 and 27 min)
under our experimental conditions when the entire alcohol was adsorbed and
there was no competition for the active centers, could be explained only by
differences in the activity of the unitary centers, rather than their number.
Under ordinary condtions, the catalytic activity is probably the total result
of the average activity of the working active centers, their number and also
their mobility, i. e. the ease with which they are regenerated after comple-
tion of a unit reaction act.
3) The A1203 surface contains sites that are devoid of catalytic activity.
Thus it can be seen on the curve that after the surface concentration of H20
reached about 3-4 prnole/m2 the reaction stops almost completely, although
the coverage is far from that corresponding to a complete monolayer and the
surface still contains some adsorbed alcohol. The conclusion regarding the
presence of inactive sites could be made only on the basis of the differential
and not the integral approach.
We shall now attempt to treat the available material from a single view-
point, assuming that the adsorbed reactants form labile, radical-like inter-
mediate complexes on the active sites of the catalyst surface. At the limit
such complexes may be regarded as free radicals, stabilized by the sur-
face*. Our studies have demonstrated their existence at the relatively low
temperatures of 150-2500, but it is only natural to assume that they can
occur at all temperatures of the dehydration reaction. However, the direc-
tion of their further transformation depends upon the reaction conditions.
They must have a short lifetime at relatively high temperatures of the order
of 350-450?, the main direction of reaction possibly being immediate break-
down to the ultimate olefin. These strongly adsorbed radical-like forms
have a longer lifetime at 100 - 200?,. under conditions of reaction in the ad-
sorption layer or in the "pre-catalytic" region and by their residual valen-
cies can bind other molecules either of the same or other species migrating
? on the surface. New intermediate forms then arise possessing a different
(frequently higher) stability, and manifest themselves by changing the rate
; ? and even the direction of the reaction. As for the rate of decomposition of
the foreign molecules bound to the initially formed complex, this may even
increase (example of conjugated dehydration).
The proposed mechanism in the case of conjugated dehydration and
transfer of the radioactive label can be represented schematically as follows:
* The existence of such surface stabilized radicals has been repeatedly discussed
In the literature in recent years.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 - ? _
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
\ r, /H
+ H20
H '
H I OH
CH3
*/
- - -C ?H CH3
/
OH 1 ? \H %-112
CH + +2H20
in
H CH 2
H
- - -C
OH
CH3 - CH-= C*I12 + CH2 = CH2 + 21120
STAT
If the conditions induce considerable interaction between the reacting
molecules, an appreciable proportion of the end products can originate not
from the primary "singlet!' active complexes, but from the secondary comp-
lexes resulting from the additon reaction. The" collective", polymolecular
decomposition mechanism will then begin to play a considerable or even pre-
dominant part in the reaction. The differences in the adsorption layer and
ordinary dehydration kinetics can be explained by the possibility of such
dual reaction. The high temperature under the ordinary conditions is detri-
mental to the formation of interaction products so that they cannot be de-
tected in the catalysate.
The collective reaction mechanism is in harmony with the basic prin-
ciple of the multiplet theory of catalysis according to which the active inter-
mediate complex is a molecule with partially ruptured bonds. The bonding
electrons of the atoms comprising the "multiplet reaction index" are par-
tially delocalized due to interaction with the atoms of the catalyst. There
? is therefore no difference between such a formation and a free surface-
stabilized radical, since in this case delocalization of the unpaired electron
of the radical also takes place. Hence such a complex should possess en-
hanced reactivity and form bonds with its neighbors.
Because of this reactivity, the failure of attempts to detect such radi-
cals by EPR when the reactions were carried out under usual conditions be-
comes understandable. In that case the secondary polymolecular complexes
are formed causing considerable loss of the radical-like properties. Thus
when carrying out the dehydration of alcolhols under the usual conditions we
also could not observe the para-ortho hydrogen convertion, in contrast to
the monolayer reaction.
If it is assumed that the concept of the radical-like character of the pri-
mary intermediate complexes here developed is of a general nature, then in
the light of this theory a number of other observed phenomena will become
understandable. This for instance is the case for the effect of the space velo-
city and inert gases on the rate and selectivity of a number of catalytic re-
actions, in particular etherification. It also sheds new light on numerous
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 :
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
9 STAT
data concerning the effect of the solvent species and of various poisons which
as a rule had been explained by selective adsorption of foreign molecules on
the active sites. Attempting to account for the specificity and highly varied
nature of the action of these foreign substances one had then to postulate a
large assortment of active sites responsible for the respective reactions and
selectively sorbing the poisons. We however, believe ti more likely that
these substances form with the adsorbed reacting molecules various secon-
dary complexes, differing in stability and in the direction of further reaction.
The available experimental data does not allow one to judge as yet the
structure of the polymolecular intermediate complexes or the number of
constituent molecules. Neither are we able to tell their mutual arrangement
whether on a single plane or in different layers (sandwich structure), etc.
However the concept of their existence is in conformity with all the data ob-
tained in the present work and can therefore be considered to be sufficient-
ly based on fact to merit further attention.
Table 1
Inhomogeneity of the active surface of alumina catalysts of
different origins
.
No.
of the
catalyst
Sp.monolayer
sur-
face
by N2
ni2/10
capacity of
ole/g -
Surface coy-
erage at the
point of
breaks in the
1
curves fig. ,
in mmole/g
Amount
of di-
oxane***
corn-
at co-
philbeteitioinn:
mmole/g
Number of
active cen-
ters, Z
S ,
x'
Catalytic
activity in
the adsorbed
layer
for
A
.2....
for
.,..._
14,./...
pro-
panol
panol
1018/
m2
1019/
g
E
(kcal/
mole)
Half
period of
decompos
r05(min)
at 120?
for
H20
total**
.
2
3
4
5 -
6
130 .
222
210
236
247
290
2.1
3.6
3.4
3.8
4.0
4.6
0.66
1.14
1.12
1.10
0.91
0.93
0.30
0.20
0.26
0.32
0.30
0.90
0.14
0.13
0.16
0.18
0.13
0.28
0.06
0.06
0.07
0.06
0.04
0.07
0.70
0.36
0.53
0.51
0.33
0.59
9.0
8.0
11.0
12.0
8.0
17.0
23
13
12
14
14
30
17
25
24.5
20.5
19.0
21.0
27
6
8
13
10
24
* Experimentally determined.
** The total effective coverage is expressed in mmole of isopropanol for 1 g of the
catalyst, taking into account the differences in values of the elementary adsorption
areas, i.e. that one molecule of iso-C3H7OH is equivalent approximately to 3
molecules of 1120.
***For dioxane the value of e) on A1203 is equal roughly to a for iso-C3H7OH, ab it
was demonstrated in the paper 28).
REFERENCES
1) A. N. Terenin, in: Poverhnostnye khimicheskije soiedinenia.i ih rol v iavlenash
adsorbzii (Surface chemical compounds and their part in the adsorption pheno-
mena), (Moscow, 1957) p.206.
2) R. P. Eishens, W.A.Pliskin, Adv. Catalysis 9 (19 ) p. 662.
3) H.Dohse and coll. Z.phys. Chem. B5, (1929) 131; B6 (1929) 343; B14 (1931) 349;
B23 (1933) 33; Z. Elektrochem. 36 (1938) 677.
-
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 ?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
1.0
STAT
4) A. A. Balandin, V. E . Vasserberg, Acta Physicochimica URSS 81 (1946) 678.
5) E.Wicke, Z. Elektrochem. 83 (1949) 279.
6) V.I. Levin, in: Problemy kinetiki i kataliza (Problems in kinetics and catalysis)
(Moscow, 1949) Vol. 7, p. 205.
7) N. P. Keyer, in: Problemy kinetiki I kataliza (Problems in kinetics and catalysis)
(Moscow, 1955) Vol. 8, P. 224.
8) E. A. Andreev, in: Problemy kinetiki i kataliza (Problems in kinetics and catalysis)
(Moscow, 1949) Vol. 6, p. 293.
9) I).P.Dobychin, in: Poverhnostnye Ichimicheskije soiedinenia... , (Moscow, 1957)
p.341.
10) V. E. Vasserberg, Kinetika i Kataliz 3 (1962) 556.
11) V. E. Vasserberg and A. A . Balandin, Problemy kinetiki i kataliza 10 (1960) 356.
12) V.E. Vasserberg, A .A. Balandin and T . V. Georgievskaya, Dokl. Akad. Nauk SSSR
134 (1960) 371.
13) G. I. Levy and V. E.Vasserberg, Kinetika i Kataliz 3 (1962) 527.
14) V. E. Vasserberg, A . A. Balandin and T.V. Georgievskaya, Dokl. Akad. Nauk SSSR
140 (1961) 859.
15) V. E. Vasserberg, A. A. Balandin and T. V. Georgievskaya, Dokl. Akad. Nauk SSSR
140 (1961) 1110.
16) V. E .Vasserberg, A.A. Balandin and M.P . Maksirnova, Izvest. Akad. Nauk SSSR,
Otdel. Khirn. Nauk (1959) 363; Zhur. Fiz. Khim. 36 (1961) 858.
17) N.D. Zelinsky and N.I.Shuilcin, Dokl. Akad. Nauk SSSR 3 (1934) 225.
18) Ya. T. Eidus and N.D. Zelinsky, lzvest. Akad. Nauk SSSR, Otdel. Khim. Nauk (1940)
289;
Ya.T.Eidus and N. V.Ershov, Izvest. Akad. Nauk SSSR, Otdel. Khim. Nauk (1959)
1655.
19) M.I.TemIdnand L. O. Apelbaum, in: Problemy fizicheskoi Khimii (Problems in
Physical Chemistry) (Moscow, 1958) p. 1, 94.
20) V.V. Voevodsky, N.N.Semenov and F. F. Vol'kenstein, in: Voprosy Ichimitcheskoi
kinetikl, kataliza I reakzionnoi sposobnosti (Problems in chemical kinetics, catal-
ysis and reactivity (Moscow, 1955) p. 423.
211 M.Y. Myasnikov, Zhur. Fiz. Khim. 34 (1960) 385.
22) C.Kemball, Bull. Soc. Chim. Belg. 63 (1958) 373.
23) E.G., papers in the books Problemy kinetiki i kataliza 10 (Moscow, 1960) Vol. 10, -
p. 369-428 and Voprosy Ichimitcheskoi kinetiki..., (Moscow, 1955) pp. 423-632.
24) V. E.Vasserberg and I.R.Davydova, Kinetika i Katallz 2 (1961) 773.
25)-V.E.Vasserberg, A. A.Balandin and I. R.Davydova, Dokl. Akad. Nauk SSSR 136
(1961) 376.
26) V.E. Vasserberg, A. A. Balandin and G.I. Levy, Kinetika Kataliz 2 (1961) 61.
27) G. I. Levy and V. E. Vasserberg, Kinetika i Kataliz 2 (1961) 758.
28) V. E. Vas serberg, A. A. Balandin and M. P . Maks imova, Izvest. Akad. Nauk SSSR,
Otdel. Khim. Nauk (1962) 1865.
?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
.7,1A-RDP80-00247A004200070001-6 -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
Jtr
MOLECULAR MECHANISM OF SOME
CATALYTICAL REACTIONS AS REVEALED BY
MEANS OF ISOTOPIC KINETICAL EFFECTS AND
EXPERIMENTS WITH TRACER MOLECULES
S. Z .ROGINSKY
Institute for Chemical Physics, Moscow, USSR
Abstract: The measurement of hydrogen and oxygen isotopic kinetical effects in
the oxidation of hydrogen on platinum points to the.participation of oxygen and
the non-participation of hydrogen in the limiting step of the oxidation of hydro-
gen. The step-wise scheme of this process is considered; the promoting effect
of oxygen and the heterogeneity of the surface are taken in account. Results of
the investigation of the T mechanism of Fisher-Tropsh process on cobalt catal-
ysts by means of the distribution of the radiocarbon mark in the products and
the isotopical kinetical effect are exposed. The possibility of a considerable
simplification and selectivity of the hydrogenation processes by conducting them
In a chromatographic separating method was shown, as well as the perspectivity
of a simultaneous application of radiochromatography and a chromatographic
column in the study of reactions and catalysts.
1. INTRODUCTION
The participation of solid surfaces in heterogenous catalysiS- involves
? intermediate forms and reaction mechanisms which are very difficult to
elucidate. For a considerable number of the homogeneous chemical reac-
tions of gases and liquids the molecular mechanism is unambiguously stated.
The mechanism of heterogenous reactions remains open to discussion, even
for the most investigated and simple gaseous contact reactions, proceeding
with the formation of one stable product and the breaking or formation of
only 2 or 3-chemical bonds. Still less clear is the mechanism of the com-
plicated contact reactions, where the main function of the catalyst is a ki-
bernetic regulator of reaction routes and a control of chemical and spacial
structure of the complex end products.
This situation shows, that the orthodox methods of research in cataly-
sis are not sufficient for an elucidation of the deep reaction mechanism. In
this report I shall expose some results, obtained by investigating the me-
chanism by means of other (mainly isotopic) methods.
2. OXIDATION OF HYDROGEN ON PLATINUM
For this system the main features are: a low activity of the degased
? 'metal, the possibility of strong activation of Pt by means of oxygen uptake
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: STAT
CIA-RDP80-00247A004200070001-6 -
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: ?
CIA-RDP80-00247A004200070001-6
:
from 02 or from the reacting mixture 1, 2) and a strong catalytic corro-
sion 3).- The reaction kinetics are different on Pt-specimens with different
histories.
For stoichimetric mixtures the reaction order is usually (z4f. kpl).
In some cases this order results from a true first order for oxygen and a
- I order for hydrogen w ?[02]/[112]r, in other cases there exists a true
order for oxygen and a zero order for hydrogen. These peculiarities of the
reaction kinetics led us to a probable stepwise reaction scheme. The pro-
cess starts with chemisorption of oxygen, which is charged negatively by
the metal's electrons. The limiting step involves oxygen, but not hydrogen 2)."
To check this conclusion the magnitudes of isotopical kinetic effects
(a) of hydrogen and oxygen were measured.
? For reactions, consisting of??-? 2 steps it is important to distinguish be-
tween the valifes ai - (obtained by the comparison of velocity constants of the
two separately Conducted isotopic reactions) and as (obtained from the mag-
nitude of isotopic separation, accompanying reactions of a mixture of isoto-
pic molecules). For a stationary process in the simplest case ai equals the
ratio of velocity constants of the limiting step. In the presence of auxiliary
equilibria, ai can -contain (as factors) thermodynamic isotopical effects.
? Accumulation of a of different kinetic steps is impossible. Owing to
competition between the isotopic molecules a3 can be observed on each bi-
furcation for the element, whose isotopes do not participate in the limiting
'step. For the same reason -in multi-step reactions, 'as-values may differ
from those of the single steps 4). If there is no redistribution provoked by
- isotopic exchange, or if this is strictly taken in account, the as-values can
be measured with greater reliability and precision, than the ai-values.
. Therefore for oxygen measurements of only as were real. For hydrogen
both values of a were determined with deuterium (mass-spectrometrically
..and by water density measurements). In a large temperature range for ow
? total pressures (0.1 mm Hg) and for atmospheric pressure, ai F.., 1. The
? data, obtained it low temperatures and low pressures (see fig. 1) are the
-? Most convincing.
Evidently the assumption 5) about-the value at ;1=-. 1 for hydrogen is true.
The Coincidence of the curves of fig.2 to a great extent shows that the reac-
tion velocity is independent of [H2]. This fact, together with ai = 1, almost
unambiguously excludes a participation of free or bonded hydrogen in the li-
miting step.of the hydrogen oxidation.'
For platintim, poisoned with respect to the isotopic exchange:
DH + H20 ? DHO + HD, there was found earlier an as-value, changing from
1.3 to 1.1 in the temperature range 950-3oo?C 6). When the temperature is
'lowered to 20?C this a-:value increases up to 1.8 (fig.3), which is greater
.than (mHD/mH). This fact indicates the presence of an isotopic bifurcation-
_
With A chemical mechanism. Under different conditions the as-values for
oxygen are found to vary from 1.01 to 1.05. Experiments were performed
' with a gas, containing 13 atom % 018; so only the molecules 016 018 and
:". :0218 have to be taken in account. The values a5> 1 for oxygen and the ex-
plicit dependance of the stationary velocity of the process upon oxygen con-
:tentration indicate that oxygen participates in the limiting step. This result,
STAT
, '
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
3
in conjuction with other data for the same system leads to the following step
scheme: Scheme I
_ 2 ,..t.,.,,(-) ,_*,_,(-) 3 ?4, ?(-)
I. Pt + 02--LIA ; it -i: ...12----rw2chs ; rty2chs ---21-tuchs ;
Pt + H2-2--Pt11(2+-2,1s. ;
'?,..r(") 4 .?, + L-12?
rinchs--ri- %-, ;_,
1 ..(-)
-11 Pt + I112?Ptnchs
* - 3 *? (-) ?
Pt + 1112?Ptlichs ;
" M. 0(-h) + 144215-1- H20 ; 01-1s + litd) H202-1- Oc(-h)s + H20,
c s
(--)? 4
-?2chs + OH ?
We should like to emphasize the introductiop of a preparatory step -
the formation of the active surface Of platinum (Pt).
mm
aim
STAT
. , 0, , q 8 re MI ' 20 nun
Fig.-1. Reaction kinetics of stoichiometric mixtures
?. of oxygen with protium and deuterium at - 780.
? - (1o)mixtV 0.107 mmHg
x - protium
o?- deuterium.
_ ? , ?
This step is a result of the oxygen uptake by the metal. This activation can
- bp prevented by a firm uptake of other gases and in particular, of hydrogen.
Pt probably consists of sites, of a two-dimensional oxide Pt0. After long
- work a 3-dimensional oxide layer is formed. This layer can be detected by
- electron diffraction (fig.4); it was identified as Pt304 7). According to Shis-
? hakof this stable peroxide may be Pt208, with 0-0-bonds in the crystal lat-
tice 8). Platinum, covered with this oxide, is especially active.
. Depending on external conditions and the state of the surface, the cata-
lysis can be controlled either by the chemisorption of oxygen in the form of
02ads12, or by its transiton in dissociated state 13. According to electrO-
, ? chemical data both forms can coexist 9,10) and 02 and can react with water
forming peroxide. Under special conditions the formation of peroxide can
Declassified in Pad- Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 crj:
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
t
occur also at the oxidation of H2.
AP mm Hg
0200
0100
0 10 20 30 rto rmin
Fig. 2. Reaction kinetics in mixtures with constant oxygen volume
and different amounts of hydrogen at - 780C.
? -02 : H2 = 1: 1 (P0)1 = 0.140 mmHg (P0)2 = 0.104 mmHg
o -02: H2 = 1: 2 (P0)3. = 0.198 mmHg (Po)2 = 0.157 mmHg
X -02: H2 = 1: 3 (P0)1 = 0.258 mmHg (P0)2 = 0.214 mm Hg
The upper bundle of curves corresponds to mixtures at high
Initial pressure.
o 1 I I I I I
a so 100 ISO 200 250 300 t T
Fig. 3. Temperature dependence of the separation degree Si,
corresponding to 10% conversion in a mixture of H2 and
,HD (atmospheric pressure, dynamical conditions),
-
(Po = 0.2 mmHg statics).
I - 412;HD, )4 - S10%
- Under usual conditions the peroxide is unstable in contact with platinum,
therefore as a rule only water is formed in a gaseous medium.
[02] when the velocity and kinetics are controlled by step 12; and
w-02T when by step -13. It is very difficult to give a more concrete defi-
i
,
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 5
nition of the adsorbed forms of hydrogen and oxygen, participating in the
catalytic process, because the adsorption of both gases is complex as re-
vealed by a study of the adsorption on surfaces which have been degased in
an ultra-vacuum, and they affect one another. Let us suppose, as a working
hypothesis, that 02 or 0- react with H2+, which, according to Mignolet 11)
and others can be formed at adsorption on the first negatively charged hy-
dride (or oxide) layer. Similarly Oi or 0- reacting on Pt, lie in the second
layer.
Fig. 4. a. Electron diffraction picture of a fresh platinum plate.
b. Electron diffraction picture of platinum after catalytic
oxidation of hydrogen.
Supplementary complications are introduced into the mechanism and
kinetics of hydrogen oxidation and of the majority of other contact reaction
by the catalytical and chemosorbtional heterogeneity of the surface and by
the modifying interaction of adsorbed molecules. In catalysis and adsorption
it is impossible to distinguish between the influence of heterogeneity or inter-
action on the grounds of external manifestations. It is possible to do so by
means of contemporary physical methods. Some isotopic methods, elabora-
ted in our laboratory 12), are effective for the investigation and detection of
heterogeneity.
To the study of the mechanism .of catalytic isotopic hydrogen exchange
on active platinum these methods were recently applied by Boreskov and
Wasilewitch 13). The kinetics of the low-temperature isotopic exchange of
adsorbed tritium with D2 and H2 indicates a large inhomogeneity of the Pt-
..:
,
Declassified in Part- Sanitized Copy Approved for Release 2014/03/04 : .
CIA-RDP80-00247A004200070001-6 -= - 11.? ? ?
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
surface relative to the velocities of isotopic exchange. The distribution of
sites relative to activation energies of exchange is found to be nearly uni-
form. The part of highly active centres is small, as for homomolecular
H2 + D2 - exchange.
The nature of these sites probably (tigers from those of the other sur-
face regions. It is interesting to note the absence of an isotopic effect
(ai = 1) in the exchange of H2 and D2 with the adsorbate. This indicates
that H2 and D2 do not participate in the formation of transition complexes
of the limiting step of exchange. It follows from the same data that chemo-
sorbed H and T participate at this step.
,
3. SYNTHESES OF ALKANES FROM CO + H2. HYDROPOLYMERLSATION OF
ETHYLENE.
STAT
The former processes is characterised by a simultaneous formation
of several (many) members of the homological series. Depending on the
choice of the catalyst and on the conditions of the synthesis there are formed
chiefly CnH2n+2 or CnH212_40H from CO + H2. Other homological series are
represented in a much lesser degree. For the chemosorption of hydrogen on
these catalysts there exist still more possible forms, than on Pt. For CO
on metals the spectroscopical investigations of Eistens and others 14) indi-
cate the possibility of several forms of chemosorption without and a disso-
ciation. Besides that, the possibility of carbidising Me and of other react-
ions must be taken in account. Unfortunately the study of independent and
combined chemisorption of CO and 112 cannot alone give a clue to the spe-
cificity of these processes, because their direction and the composition of
their products depend upon the following steps.
Of great principal importance was the pioneering work of Cummer and
Emmett 15). In their experiments, adding alcohols, marked with C14, to.
the CO + 112 - mixture, they found that carbon from the alcohol entered in
products of synthesis. The molecular radioactivity (am), appearing in al-
kanes of different molecular weights, was constant. This result, obtained
on Fe-catalysts, was considered by the authors as an indication, that the
reactions are initiated by the alcohol or by a surface radical, formed from
the alcohol. At the same time it was considered as a confirmation of the
Storch and others 16) dehydration-condensation mechanism of the carbon
chain growth. In our investigations the data 17) on the constance of the mo-
lar radioactivity of products were confirmed also for syntheses in the pre-
sence of alcohols on a cobalt-thorium-catalyst. But at the same time it was
shown, that a similar effect is observed with the introduction of marked
organic compounds of other functions, including olefins (fig. 5). Under the
same conditions the participation of olefins in the initiation is higher, than
that of alcohols. Hence the assumption about the role of olefins or their di-
rect convertion products (for instance methylenes) in the initiation of the
methylene chain growth. Other authors later arrived at similar conclusions
for cobalt and iron catalysts. The results of a comparison of the distribu-
tion of radioactivity in products of reactions, initiated by ketene
. , .
'Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 - -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
7 STAT
H2c = c = o, marked with C14 in the methylene or the carbonyl groups on-
ly 18) are very convincing.
10
CO2 loft num is
po
IGO
ISO
Fig. 5. a) Molar (I) and specific (II) radioactivity of hydrocarbons
In the hydropolymerisation of ethylene in the presence
of 2 at % of marked CO (8.6 pCuimmol).
b) Specific (III) and molar (IV) radioactivity of hydrocarbons
In the synthesis from a CO + H2 - mixture, containing
1.45 at % of marked C2H2 (12.5 ?Cu/mmol).
Fig.6 shows that only for low atomic concentrations (? 1%) of ethylene its
role can be reducted to initiation. When its concentration is. increased up to
several percents an increase of am with the molecular weight of product is
observed. In this case ethylene participates also in the chain growth. Hence
a natural passage to hydrocondensation and Eidus hydropolymerisation in
H2 + C2114 - mixtures containing additions of CO 19). In this reaction the
majorpart of methylene groups of the carbon chain is supplied by ethylene,
so that, if- the ethylene is marked, am must increase linearly with molecular
weight.
Less clear is the rOle of CO in the hydropolymerisation. Besides the -
Initiation, it can be a poisoning of harmful reactions - hydrogenation of eth-
ylene, etc. At the hydropolymerisation of mixtures, containing 2% of
marked CO, the specific radioactivity of hydrocarbons grows nearly line-
arly (quadratic increase of am) and the prepondering part of light hydrocar-
,
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 ' - ?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
15
bons is formed without a participation of CO *.
1 41 lo j1.1 NI 18
caliort rtunitaz
Fig. 6. Molar (I) and specific (II) radioactivity of hydrocarbons in
the synthesis of hydrocarbons from a mixture of CO and
H2, containing 4.8 at % of marked C2H4 (1.53 mCu/mrnol).
STAT
The constant nature of am in conjunction with the non-participation of
marked CnH2n4.2 in the synthesis confirms that in reactions of this group
chemisorption is followed by steps of 3 main types:
a) Formation of the active form (priming), able to start a carbon chain of a
hydrocarbons or alcohols:
b) Growth of the hydrocarbon chain, by means of a series of consecutive u-
niforms processes.
These processes proceed with the participation of CO and H2.
c) Stopping of the single reaction, when the active form is converted in a
molecule of the final product;
d) Removal of this molecule so that the active center gets free to repeat the
same step sequence.
The whole process of formation of every separate molecule with any
number of C proceeds without disengagement from the surface; according
to the fixed chains mechanism without "relay-race" 17, 21). Small and big
molecules of the products of every homological series are formed indepen-
dently. It is striking, that even additions, which participate strongly at the
initiation, do not change considerably the velocity of the total process or the
? * There is in this question a certain discrepancy in the data of Gibson et al."), who
found a linear increase of ani7
Declassified in Part: Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 I- - -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 9
compositon of the products. This shows, that even in synthesis reactions
without additions, the initiation phase, including the formation of the pri-
ming of a chain, able to grow and firmly bond on to the surface is short.
The participation of additions at the initiation can be used as a relative mea-
_ sure of the probability of the priming formation. Assuming the mechanism
of fixed chains, it seems natural to consider the priming (in the kinetic re-
spect) as a whole with its active center. The surface is probably hetero-
genous relative to these centers in respect to the strength of the bond with
the priming and the ability of chain growth. This can be an additional factor,
defining the chemical and molecular composition of the products.
Every center must consist of several atoms, because during the pro-
- longed exclusion of atoms, bonded with the priming; there must be available
sites of an other type for the chemisorption of CO and H2, (which take part
in the growth), and perhaps also for the slipping down of the product and
the "step over", postulated by some authors.
Considering detailed atomic models, we show here one of the possible
schemes of the process prematurely.
* Let us design active sites, on which the growing chain is fixed, with
S, and the other sites, participating at the process, with S. Scan differ
from S by its structure and by content of modifying additons. S can be me-
tallic atoms, adjoining to dislocations or to microcrystalls of Th02. The
growing chain priming can get fixed at the surface, as a result of the direct
bonding of the active part of the molecule with bare Me-atoms. Our experi-
ments with olefins, the experiments of Emmett with ketene 18) lead to the
assumption, that the priming consist p 2f a radical, not containing oxygen.
In scheme 2 it is supposed that it is S-CH2, - some sort of simplest sur-
face carbonium - ion. The chain grows by the following steps 111-117. Some
indications-of the character of the limiting step are given in the work of
Sakkaroff and Dolcukina 22). For the reaction of CO they got a i (H2, D2) =
0.77. The value: ai 0 + HO-CH2-CH2- [0C2H4L OR
N /
MeoMe... cs MeoMe...
The first version of termination by means of solvent or impurities as
said above is more probable than the second version, i. e., termination by
polymer.
Kinetic data obtained earlier 3-5) corroborate the proposed scheme.
REFERENCES
1) F. N. Hill, F. E. Bailey and J. T. Fitzpatrick, Ind. Eng. chem. 50 (1958) 5.
2) J. Furukawa, T. Saegusa, T. Tsuruta and G. Kakogawa, Makromolek. Chem. 36
(1959) 25,
3) 0. V.Krylov and Y. E. Sinyak, Vysokomolek. sojedinenija 3 (1961) 898.
4) O. V.Krylov and Y. E. Sinyak, Neftechimija 2 (1962) 688.
5) O. V.Krylov, M. J.Kushnerev and E. A. Fokina, Neftechimija 2 (1962) 697.
6) Z. A. Markova, Kinetyka i kataliz 3 (1962) 366.
7) D. A. Dowden and D.Wells, Actes 2e Congr. intern. catalyse, Paris, 2(1961) 1499.
8) T. V. Antipina and G. N.Avdonina, Zhurnal fisitcheskoi chimii 31 (1959) 192.
9) I.E. Neumark and I. B. Slinyakova, Ultrainskii chImitcheskii zhurnal 27 (1961) 196.
10) R. M. Hexter, J. Opt. Soc. Am. 48 (1958) 770.
11) I. M. Hunt, M. P. Wishard and L. C. Bonham, Analyt. Chem. 22 (1950) 1478,
12) P. J. Flory, j. Am. Chem. Soc. 62 (1940) 1561.
13) J. Furukawa, Polymer 3 (1962) 487.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
-2,IA-RDP80-00247A004200070001-6
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
THE STRUCTURE OF POWDER CATALYSTS, MADE OF
Ni-Co ALLOYS, AND THEIR CATALYTIC ACTIVITY AND
KINETICS OF REAGENTS ADSORPTION IN THE PROCESS
OF BENZENE HYDROGENATION
- Zdzislaw SOICALSKI and Jozef PODKOWKA
Physica-Chemical Department of the Silesian Institute of Technology,
Gliwice, Poland
.? ? :
1. INTRODUCTION .
? - .
. _
Powder catalysts, made of Ni-Co alloys, have been investigated by
Lihl, Wagner and Zemsch 1), as well as by Hund 2). This type of contact
, catalysts is of special importance in investigations, the aim of which is to
determine the dependency between the catalytical properties and the crys-
tal structures of the catalysts. They offer much possibility of choosing
1
1 such a,series of catalysts, which are characterized by continual changes
f - - . of their structural properties. The method of preparing these catalysts is
i'. . based on the reduction of suitable salts of mixed crystals.
1. According to Grim's theory 3), two metals can form a continuous series
!_,A- ?
of: mixed crystals, if the difference of their ionic radii does not exceed 7%
.
?
1,- _ (in cases when the electrical load and their chemical nature is the same):
*-.. The ionic radii of Ni and Co amount to 0.78 and 0.82 A respectively. Thus
? _ the difference is only 5%. Because the ionic radii of both metals, founding -
- - the structural lattice, of the catalyst are so very similar, there exists a
rather' broad range of concentrations of aqueous solutions of salts, which
_ may be used for the production of catalysts. In this case one cation of the
. metal A is to be wedged into the: crystal' lattice of the metal B. This case
in the interchangeability of the cations forms the basis for a wide series of
experiments, the more so, because when suitable proportions of the initial
components are used, there may be produced crystals of various crystalline
phases. The above mentioned authors 1,2) have limited their investigations
to the examination of the catalytical properties and cristalline structure of
powder catalysts, made up of Ni-Co alloys. Today it is well-known that the
? : chemisorption of the reagents of a catalic reaction is one of the elementary: '
, -prOcesses of contact analysis, so that the investigations into the kinetics-, ? .. .
' --of chemisorption.render it possible to describe the heterogeneity of.the .-: ?
catalyst' surface quantitatively and it seems to be worth-while to supplement..
, . .
.,. ? both the kinetic and structural investigations with measurements of the
-: a6orptiotr kinetics of the reagents of benzene hydrogenation in. conjunction.f ...,
'
with the activity of-the contacts is such a.reaction.
....:. - -
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: _
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
2. EXPERIMENTAL PART
STAT
2.1. The preparation of the catalysts
For the experiments there were used powder catalysts, which had been
produced by means of a thermal decomposition of mixed crystals composed
of Ni and Co formates. In such a way, too, catalysts have been produced by
Hund, Lihl and Wagner 1,2). The initial reagents were Ni and Co formates,
got by means of eleetrolysis; they have been dissolved in suitable proportions
in a 10%-HCOOH. Then this solution was evaporated at its boiling tempera-
ture until a saturated solution has been received. After that the solution was
cooled down by means of water at a temperature of 14?C. The crystals pre-
pared in this way, were dried at a temperature of 110?C.
2.2. Catalytical investigations
The reduction of the catalysts and the cata.lytical reaction of C6H6-
hydrogenation have been carried out in an apparatus which is shown in fig. 1.
^
Fit
ContociUmnnomoter
Coftilyst toyer
t:InlansO goy: et falter
Cooling noir. ?
Fig. Li Apparatus for reduction and catalytical hydrogenation of C6H6.
The. electrolytical hydrogen, taken from a.pressure bottle, has been passing
?_ . successively through a fleometer 1), then through absorbers with H2SO4 2)
and with KOH 3), through a filter with active carbon 4), an absorber with
-pyrogallol solved in KOH 5), next through an oven containing the catalysts
? --- for removing the last traces of oxygen 6), and finally through pipes filled
with anhydrous CaC12 and Mg(C104)2 7). The reduction of all the catalysts
? has-been executed at a temperature of 330?C with hydrogen flowing past at
a rate of 40 1/h.- The time of reduction amounted to 2 hours. Afterwards the
? ? ? ?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 4'"
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
IA-RDP80-00247A004200070001-6
I I 3 STAT
oven was cooled down.to reaction temperature, while the hydrogen flow was
kept low. The cooling time was the same in all the experiments. The amount
of Ni in the catalysts, which have been prepared in this way, increased suc-
cessively from 0 to 10, 30, 70 and up to 100%. The nickel content was deter-
mined by precipitating a -dimethyloglyoxime and weighing it.
The reduction value was determined by solving the weighed amount of
the reduced catalyst in 20% H2SO4 and by measuring the volume of the evolv-
ed 112. The reduction values of various catalysts are presented in table 1.
Table 1
Reduction values of catalysts with various Ni-contents.
NI-content
in weight %
Reduction
value
0
97.5
10
99.0
30
98.1
70
96.5
100
98.5
It follows from this table that the reduction values for all catalysts amount
nearly up to 100%.
Having set up the temperature, benzene was dropped down from a micro-
burette at a rate of 0.1 ml/min, while the hydrogen was flowing at 6 1/h.
This makes a 20% excess of a, stoichiometric proportion. In this way the
benzene was dropping down for 10 minutes, after which for 5 minutes pure
hydrogen was passed through in order to remove from the catalyst all C6H6
and C61112. Then again benzene was dropped in for 10 minutes (after Lihl
and Wagner). For each catalyst six experiments have been carried out in
order to get reliable information as to the activity changes during the pro-
cess. The reaction products were being cooled down in a water condenser 8),
the liquid products being investigated by-using a refractometer for their
C6H6- and C6142-content. For this purpose a calibration curve was designed
determining in this way the content of C6H6 and C61112. Experiments have
.also been carried out, the aim of which was to show whether C6H6 can be
hydrogenated by using simply a pure glass pumice. It has been proved,
that in such conditions no hydrogenation does occur. The kinetics of hy-
drogenation has been examined at temperatures of 1400 and 1800 using
Catalysts with 0, 10, 30, 70 and 100% Ni content. The dependence of the
yield (in %) of the hydrogenation of C6H6 to C61112 upon the temperature is
represented in fig. 2, whereas figs. 3 and 4 show the dependence of the
yield upon the duration of the experiment at temperatures of 140 and 180?C
respectively.
For these same catalysts the kinetics of the dehydrogenation of C6H6
have been carried out at temperatures of 215, 260, 290 and 310?C. The
results of those experiments are set up in table 2.
2.3. Survey of the adsorption kinetics
It is much easier to interprete quantitatively the results of the meas-
urements-in the kinetics of adsorption, if these measurements are taken
- under isobaric conditions. In these conditions the constants of the equation,
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
-2,IA-RDP80-00247A004200070001-6 - - -
T
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
4
Win%
A
?
0
???
?
?
I
I
BS M
V% Ni
JO% M
A% IN
DAM
.
?
rm
za 120
f
Fig. 2. The dependence of the yield of a C6H6 to C61112-
. reaction upon the temperature.
.Table 2
Ni-content in
the catalyst
Yield of the reaction C61112 ?C6H6 at temp.
215
260
290
310
- in%
_
0
0
4.3
5.2
-
10-
0
6.0
6.4
7.1
30
0
6.5
8.4
9.8
70,
0
6.4
.
7.6
9.5
?
100
-
9.3
14.4
16.3
STAT
conterning the distribution function of the active spots according to the ac-
tivation energy, e. g. in the equation
p (E) = exp kot exp (- EMT)] ,
, the constant kc, is independent of the pressure, so that a less complicated
equation can be used.
To avoid ,such difficulties, the apparatus, measuring the adsorption
s kinetics, has been equipped with a manostat and thus the isobaric conditions
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 ?-v? - ?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 5
W in%
re
a
N. TempflOV ?
NUMber SIXOSCSIVO sample
? Fig. 3. The dependence of the yield of a C6H6 to
Cain-reaction at 140?C upon the duration
IV of the experiment.
re/nix NOV
4
'2Alorn: ber of siiivissiiv sample
Fig. 4. The dependence of the yield of a C6146
; -tci C61112-reaction at 180?C upon the.
duratiolk of the experiment.
' -
- . -
Declassified in Part - Sanitized dopy Approved for Release 2014/03/04 :
CIA-RDP80-00247A004200070001-6
? ?
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: STAT
CIA-RDP80-00247A004200070001-6
are fulfilled. The scheme of a modified apparatus, developed by Taylor and
Strother 4) for taking measurements of adsorption kinetics is shown in fig. 5.
-
Fig. 5. Apparatus for measuring the adsorption kinetic by p = most.
. The samples were degased at a temperature of 3500C; in this operation a
vacuum was used of about 10 mm Hg during 4 hours. The hydrogen was
developed in an electrolyser, as shown in fig. 5, and introduced into the
apparatus through the mercury lock D. Liquid benzen and cyclohexane was
passed into the vessel E, where a high vacuum was formed, and from there
- brought into the main apparatus.
All measurements were performed under a pressure of 25 mm Hg. The
first measurement was taken 1 minute after the cock 3 has been opened. As
sinmediate measurement is impossible because of the aerodynamical perturb-
-ances, Occurring when the gas is passed into the adsorption vessel, and
because of the initial instability in working of the manostat. The last meas-
urement was performed after 90 minutes. Roginski has stated in his paper 5)
basing on an analysis of a great number of experiments - that between the
-first measurement and the last one there should be observed that /2 > 40 t1,
--what in our case is fulfilled as /2 = 90 t1.
? Kinetical measurements have been carried through at temperatures of
140? and 180?C, using catalysts of 0, 10, 30, 70 and 100% Ni-content. The
results of those experiments are represented in figs. 6 to 10.
2.4: RoentgenographiC measurements
The roentgenographs were made with cameras of the Debye and Scherrer
type. The experimental results of these measurements are to'be seen in
-I table 3 as well as in fig. 11. The roentgenographs of the investigated catalyst
:are given in the plates 1 - 5.
-Basing on the analysis of Ni-Co systems, the appearance of two lattice
-phases has been proved. They are the hexagonal a -phase and the Cubic 13 -
phase. - -. , .
In standard conditions for 05; of Ni content the a -phase is stable. The
-appearance of the (3-phase in these conditions in a quantitative prevailance
'can be ascribed to the stabilizing properties of dissolved atomic hydrogen
-into a cristalline lattice of Co.
?? .
:
, .
Declassified in Part: Sanitized Copy Approved for Release 2.6.14/0/04 ?
3IA-RDP80-00247A004200070001-6 ? :=2.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
.2 ? ;my
* ? 2%16
' MX*
r
50 10 70 10 20
In /Odes
? Fig. 6. Kinetic-adsorption isotherms of H2 at 140?C.
. .
STAT
? ?0005I I Z I
? ? IZI
.12 ,
111I
?? .NZ I
?
-
tfc--adSorption fsotherms, of C6116 at 1400C.
1
a
. ?
. Declassified in Part - Sanitized Copy Approved for Release 2014/03/04::
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
ICIA-RDP80-00247A004200070001-6
1 . 8
?Catirt 0% Ni
f . h
- NO% h
1
? ID 31 ? 770 77 ?
Fig. 8. Kinetic-adsorption isotherms of H2 at 180?C.
tin Ninnies
STAT
_ 60
. h &Wes
?
Fig 9 Kinetic-adsorption isotherms" of C6H6 at 180?C.
. ,
. ?
Declassified in Part - Sanitized Copy Approved for Release 201-4/03/04 :
CIA-RDP80-00247A004200070001-6 -
_
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
L
.Citokst sZ
* ?
.- ?
In
311%i
X %IR
li 1'
-t--:--.
--
-- -
?
'1 ?
,
o
?
to
A
7 . (
.
7 ,
i
r
'
ip lirlirr.
sr
4-158
-
JO
P il
II
ro
JO JO
_Fig 10. Kinetic-adsorption isotherms of C61112 at 140?C. fin Knutes
-
M111
-
UM
'.......'*"........L?................,.?...........,?,,,
ISO
.
J a
UN
-
/
.
?
P JO
.
I ?
1
a
a
a -
a ?
Xm
Fig, 11. Dependence of the lattic constant g on the composition
of the catalysts.
_
9
STAT
A catalyst with 10% nickel content develop's a higher amount of the 13 -
_
phase. This is caused by building into the Co-lattice some amount of nickel
atoms (there appears a greater ability of dissolving Ni-atoms). Together
with this taere has been observed a decrease of the lattice constant a. (cf. .
fig. 11) and a deviation from the additive properties of the lattice constants
in regard to its Chemical consistance. (The deviation from Vegard's rule in
caused by lattice contraction.)
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 :
CIA-RDP80-00247A004200070001-6 - ?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
10
Plate 1. Roentgenograph of a contact with 0% Ni-content.
STAT
Plate 2. Roentgenograph of a contact with 10% Ni-content.
Plate 3. Roentgenograph of a contact with 30% Ni-content.
Plata 4. Roentgenograph of a contact with 70% Ni-content.
. -
Plate 5. Roentgenograph of a contact with 100% Ni-content.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
ICIA-RDP80-00247A004200070001-6
Table 3
Roentgenographic-structure investigation results for Ni-Co powder-catalysts
11 STAT
Ni -
content
in % of
weight
Phase composition
Dispersivity of the system
Morfological structure of
the crystalits
Phase 0
Phase a.
Phase 0
Phase o.
0
a. and13 , phase 8
average size
non isometric
has an iso-
appears in
prevailing
of the
shape of the
metric shape
form of ex-
crystalits,
range 0.141.
crystalits:
thickness 0.01u
width: 0.11.4
of the
crystalits
tremely thin
sheets, ar-
ranged in
layer packets
.10
phase composi-
tion as above -
mixed crystals
as above
_
as above
as above
as above
30
Gradual decay
of phase a.
mixed crystals
slight in-
crease of the
average size
of the crys-
talits in
phase 13
as above ?
as above
as above
70 -
phase 0
mixed crystals
as above
as above
as above
as above
100
phase 0
mixed crystals
0.1 ?
as above_
as above
The roentgenograph of a catalyst with 30% Ni-content is characteristic
for its further decrease of the lattice constant of the 13 -p ha se, for the decay
of the a-phase and the sharpening of the bands (reflexions). This phenomenon
may be attributed to the increase of the regularity in the lattice structure.
Catalysts with 70% and 100% nickel content are made only of the 13 -p ha se ,
which is in standard conditions the only stable structural form of nickel.
.Very sharp bands of the catalysts with 100% Ni-content prove that the system
is characteristic for its well-formed structure of the crystal lattice.
The investigated series of powder catalysts 4) has been assumed to be
structurally modified by forming a solid solution of two substitutive metals,
viz. Ni and Co. As it is shown in fig. 11, in this series there is maintained
a continuity of changes of the crystal lattice constant, this continuity being
dependent upon the chemical composition of the catalysts.
According to Wolkensteijn's theory of catalysis, the contact catalysts
are modified due to their contact with the reagents of the catalytic reaction.
If those reagents are gases, the contact modification is called a gaseous
modification 5). The contacts are submitted to such a gaseous modification
even while the reagents of the investigated reaction are being absorbed.
Especially the initial period of adsorption, characteristic for its larger
numerical values of the adsorption potentials, exerts a dominant influence
upon the modification. In this case both the catalyst and the adsorbed reagents
are treated as being one quantum mechanic system 5). The process of the
gaseous modification may be forwarded on the basis of the following assump-
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: STAT
CIA-RDP80-00247A004200070001-6
12
tion. The priority of covering the adsorbed molecules belongs to those sur-
face elements, which expose the highest potentials of adsorption, characteriz-
ing the investigated system by stronger binding forces between the atoms of
the reagents and the catalyst surface. The covering of these surface elements
is executed during the initial stages of the adsorption time tx, the degree of
covering the reagent x being equal to O. The remaining fraction of the sur-
face, not covered by any reagents, i. e. (1 - Os), the surface elements of
which are characteristic for their medium or low values of adsorption
potentials, is being covered at later stages of the adsorption time tix.
- In result of the surface reaction the products, formed within the range
of active surface concentrations, are being desorbed and the substrats
again adsorbed.
It may be assumed that the covered surface 0 is constantly blockaded
throughout the whole time tx. Thus the surface (1 - e, which has been
covered during the time interval rx, possesses active surface elements,
.allowing in the case of a dynamic adsorption to achieve a continuous repe-
tition of the reaction cycles, viz, the adsorption of the reagents, the sur-
face reaction and the desorption of the products. In this case on the surface
fraction (1 - 0x), covered within the period of time t3c, there are gathered
the adsorbed reagents with an active surface concentration. For the purpose
of a quantitative, comparative comprehension of the kinetics of the adsorp-
tion processes, taking place for the investigated series of catalysts, it is
_1 desirable to distinguish between the two time intervals tx and tx'. Such a
distinction is made rather easy in cases when the transfer from the extent
? .8 to the extent (1 - Ox) is characterized by a discontinuity of the function.
U has been assumed that the longest times of all the possible times of initial
adsorption is in the case of our series of contacts the time of modification,
after which the process kinetics for all the contacts follows the functional
_ dependency y = a In t, this dependency remaining constant. The position of
the straight lines repsresenting a nickel content of 30% and 70% proves that
according to the above-mentioned assumptions it is impossible to choose a
shorter period of time than 10 minutes, i. e. log t = 1. The straight lines
represented in figs. 1 - 5 may be expressed by means of the equation
y = a log x + b. They are practically continual functions from log t = 1 up-
-wards, whereas the quantity b is characteristic for the initial period of
adsorption, which is a period of spontaneous gaseous modification. The
- time tx, . extending from 0 to 10 minutes, is assumed to be the time of the
spontaneous gaseous modification, while G > 10 minutes is assumed to be
the time of adsorption taking place at the surface fraction (1 - Ox). In this
time there occurs the covering of the adsorption places with active concen-
trations. If the straight lines in figs. 13-17 cut the axis of log tx = 1, the
_contact surface is not modified (gaseous modification).
Basing on these principles there have been received straight lines in
the coordinates system q = f (log t), as shown in figs. 13-17.
_ In order to determine the characteristic coefficients as being a quanti-
tative expression of_the adsorption process we may take up the following
considerations, basing upon the kinetics of adsorption:
?-?
,.. The adsorption rate constant k is determined by means of eq. (1).
i?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
;IA-RDP80-00247A004200070001-6 ? -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 13
? 0.0 ?C
? NT
iS-
(Ms
?
3WO 3-* IWO 3.
Lc//ice.c-onstant .cr'/?.? A 15#2
Fig.. 12. Dependence of the reaction efficience of
the hydrogenation of C6116 on the lattice
constant a.
INS
_ I .
STAT
. t
Fig. 13. Diagram q (logl) of H2 adsorption at -140?C.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: -
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
14
. i
I', is to It
Fig, 14. Diagram q:= f(log t) of C6H6-adsorption at 140?C.
? : :. 1 - - dqx. .. ?kx
Sx dt ? no(x)
STAT
(1)
where:
dqx - differential of the adsorbed molecule of the reagent x,
? cit - differential of the adsorption time,
- Sx - surface at which the adsorption process is taking place,
concentration of the reagent x in the gaseous phase. For isobaric
conditions no(x) = const.
? 'On the other hand the rate of the adsorption constant can be represented
by means of Arrhenius' equation
kx = Ax exp
- R
where:
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
(2) '
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
1 15 STAT
-
-Fig. .15. Diagram q = f(log t) of H2-adsorption at 180?C.
A- coefficient referring to the reagent x,
x
Ex - activation energy of adsorption of the reagent x,
R - molar gas constant,
T --absolute temperature.
? ? .The coefficient Ax depends upon the number of collisions of the gaseous
_molecules occuring at one unit of the surface Sx per second. Thus
'Zx
? Px
where Zx/Sx represents the number of molecules striking against the sur-
face unit within the period of one unit of time. px is a quantity dependent on
the arrangement'of the adsorbed gas molecules and their orientation in
? respect to the lattice parameters at the- surface of the adsorbent. Comparing
* kxln both equations (1) and (2) we get a new dependency (4):
(3)
?
??
For our particular case', where hydrogen and benzen aie the
(4) may be written in the following form:
. ?
reagents-, eq.
_ .
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 :
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
16
Fig. 16. Diagram q =f(log!) of C6$12-adsorption at 140?C.
' 1%1 -
lif
'
Fig: 17. Di rain q = j(log t) of C6H6-adsorption at /80?C:
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 -
STAT
_1.
I: ? ? ???:. 4.4 ? ?
? ..? . ?. "
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
ICIA-RDP80-00247A004200070001-6 - 17
and
.44(-12) ? EH2.? ?
dt
= ZB2pH2 exp. (
? no. H2
dqc 6146 - E c6H6\
& ? oc6H6
= Zc6H6pc6H6 exp RT )
n
(5)
(6)
The -empirical eq. (7) expresses the adsorption kinetics of the reagents
of the citalytical hydrogenation process of benzen to cyclohexane:
qx = ax1n tx + bx . . (7)
To this process corresponds the heterogenetic distribution function 6)
whereas
. , .
,1 dNE
?- Nx (8a)
in which NE denotes the number of adsorption spots at one surface u.nit with
an activation energy of Ex,. while Nx. is the general number. of all adsorption
spots at the surface unit. The expression dNEx/Nx represents the probability
that at. one surface Unit there occur dNE adsorption spots with an activation
energy of E.
Differentiating eq. (7) against the time-we obtain:
dqg2 af/2
VH2 7 at t
. ,
STAT
r: -;c1qC6116 ? -'aC6H6
,Y.C6116 . & ?77?17?
Henc.e_fpr. equal adsorption times we shall get the .fpllowing equation:
r, (10)
where Vx is the adsorption rate of the reagent x.
. Inserting into eqs. (5) and (6) the corresponding expressions from eqs.
(9).:?.nd (10) we get to another dependency, viz,:
2
zil pH ? exP
, = H PC6I1
exp'c RT ). ? (12)-
t noc6g6 6 6
Declassified in Part - Sanitized Copy Approved for Release 2014/-03/04 :
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
Pli
18
STAT
Dividing both these latter equations by sides we obtain:
all, Z112 pH2 + Ec6H6\
g
aC6H6 ZC6H6 PC6..6
eV RT )
where = nor12 x const.
(13)
ncC6116
- Equation (13) is valid for equal times of adsorption of the reagents i. e.
.for such theoretical conditions which form the basis for expressing the ad-
-- sorption by the name of a quasiconsecutive adsorption of two reagents. In
our case they are hydrogen and benzene.
Making use of eq. (13) we can set up the algebraic sum Ex - E. if only
the respective numerical values of ax and ay are known. The latter values
can be evaluated from the kinetical equation of adsorption p = f(ln t) where
for our experimental measurements a = tg a.
Setting up a respective equation of type (13) for the reagents C6H6, 112,
C61112 by multiplying both sides of the equations by the coefficients ac6H6,
a112 and then dividing both sides of the equations by another equation with
the coefficient ac6H12, we come to the following dependency:
ace% aH2= Ze6H6 pc6H6 ZH2pH2
RT1n
5C6812RT In f3' z
C6H12 PC61112
? (14)
- EC6H12 EC6H6 E112 '
- ?
This equation is valid for equal times of adsorption of the reagents C6H6,
and C61112 in case when the adsorption takes place for each reagent sepe-
rately. ?
Thi algebraic sum of activation energies of adsorption of the respective
-reagents in a catalytical. reaction may be measured in a similar way, if only
In (ax/ay) is known. We base in this case, of course, on experimental in-
vestigations.
In fig. 18 we find represented the efficiency of the reaction of hydro-
genizing benzene into cyclohexane at a temperature of 180?C; it is drawn in
form of the function In (ax/ay):
la
1.47 Osk71).. ?
?
(14a)
The appearance of two parallel straight lines in fig. 18 may be explained
?-when basing on fig. 17.
The straight lines appearing in fig. 17 and concerning the content of 30%
and 70% Ni refer to the adsorption of benzene steams, for which, as we see
? from the drawing, there is no preliminary adsorption procoess. In our case
this is synonymous with-depriving the -catalyst surfaces of their gaseous
modification.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 .
. .
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: ?
IA-RDP80-00247A004200070001-6 19
Win X.
? . ? an2 .
Fig. 18. 'Diagram w f (log .) for the kinetics of adsorption at 180?C.
, ace.H6
As the parallel straight lines I and II may be displaced with regard to
one another, it must be concluded, basing on eq. (14), that the multiplier
occuring in front of the exponent, as given in eq. (14) is larger for straight
line II then for straight line I. Hence the mathematical product zH2 pH2 for
30% and 70% catalysts, the surfaces of which have not been modified by
benzene steams, Is larger than in case of modified surfaces.
As a 'temperature of 1400Csthe.-diagrams in fig. 19 comprising the
characteristic value dw/d(ln ax/dyr 1,, have a negative slope, whereas
for a temperature of 180?C the slope, is a positive one.
- In the stationary state of a catalytical process there establishes at the
catalyst surface a simultaneous adsorption process of the substrates, as
well as a process of desorption of the reaction products. In such conditions
the simultaneous adsorption must agree with the kinetics of the catalytical
process.
? For the-stationary state we apply the following equation:
x x .
-C6.11642i
H2 -ix C6H6 PH2 PC6H6 exp(
-ExC6116 EH2
exp ) (15)
x
C6H12 Colli2 :PC61-1f2 ? . ? ? ? ? RT ? - ?
-
. _
'where both values 46116 and 42 refer to-the adsorption ratio at the active
. .
Declassified in Pari- Sanitized Copy Approved for Release 2014/03/04:
DIA-RDP80-00247A004200070001-6
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
20
ti A ? lif! *MO
fy ,
a
Fig. 19. Diagram w f (log 112) for the kinetics of adsorption at 140?C.
aC6H6
STAT
spots, whereas at,H12 expresses the ratio of desorption in cases when
tx = const. Ex are The values of the activation energy in a real desorption,
connected with the active spots of adsorption, in contract to the value Ex
denoting the apparent activation energy of the process. In such conditions
the value expressed by static parameters equals to 1. The right hand side
of the expression in eq. (15) bases on the conception of Van 't Hoff's box,
the box being replaced by a porous catalyst. In order not to obscure the
mechanism of this process, in which the cycles of adsorption and desorption
occur again and again one after the other on various, so called "mosaical"
areas of the catalyst surface, this successive occurence depending on the
activation energy of these areas, we have confined our observations to
stating the most general form of these phenomena.
Equation (15) may be expressed in the following way:
?
? aC6H6 7C6116 aH2 VH2
aC61-112-1T6H12
The values yx denote the coefficients, by which the numerical values of ax,
? resulting from the measurements of a separate adsorption process, are to
-be multiplied in order to obtain the expression ax. Thus eq. (16) expresses
the stationary state of a catalytic process in case of a quasi-simultaneous
adsorption, i. e. in case of an adsorption (on active spots) determined on
, ?
the basis of kinetical measurements of a separate adsorption.
Exr. - Exr. - Exu
??6..0 6 ..2\
-
- const.x exp
RT (16)
?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
-.;IA-RDP80-00247A004200070001-6 -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
I I21 STAT
Because of the fact that for the stationary state the condition must be
fulfilled, demanding that
aceili 'toe.
= 1 , (17)
aC6H12 YC6/112
- hence
EH2\
? a112 = const. I?-cH2 )
x
RT (18)
-
. where ?H2 EH2 = 42.
? In case when the selectivity of the hydrogenization process of benzene
to cyclohexane for a stationary state amounts to 100%,. then
' -4/2 1/2 a112 a112 yH2
w -
dt dt .const. const. (19)
? -
where w denotes the efficiency of a detailed process and -(dli2/dt) the hy-
drogen consumption "during the catalytic process.
- Taking into consideration eq. (19) as well as eq. (18) we obtain the
- following formula:
_
'/-i112-EH2N
= const.. exp RT
(20)
? The experimental data for w, obtained for a series of catalysts, which
have been modified ata temperature of 140?C, are presented in form of a
' function log aH2 in fig. 20 (curve I). _
..? ?
? . This curve I fulfills the following equation:
._ . ? , .
const. exp In (21)
?
?
where k is the negative directivity index.
- Because the quantity w is connected with the adsorption of the active
spots, eq. (21) may be expressed thus:
. aH2 const ? exp (-k ln a/12) . (22)
?
It results from this equation, that on the surface of catalysts with rising
values of (dg112/dt) = a112 of a separate adsorption', there appears in a
? quisi-simultaneous adsorption for the stationary state a decrease of active
spots: These facts correspond to the comparative results of the efficiency
of the catalytical process, as we may see from the experimental curve
presented in fig. 20. The efficiency of the process shrinks with the decrease
of the number of active adsorption spots.
? Dividing both sides of eqs. (20) and (21) we get:
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
22
STAT
00
02AI OX OZNi
100 20
80 40
5O2
40 Q8
20Q4
002Ni a 100ZWi
44. aa Q8? 0 2 fl
..
'const.x- -E, H2
const.exp ( RT k 'nal/2.
It results from eq. (23) that function w has only one meaning if
.10owing, then, from the diagram the value of k (according to eq. (21),
and from the experiments the value of ati2, we can thus determine
:EH2 EH2 = E2 for a quasi-simultaneous process.
The Curve in fig.- 20 corresponding to the temperature of 140?C and to
n ait2;: quite fulfills the experimental equation
.ccinit' ? e40 (-k inaH2) ?
-For the &lice of comparison, Curve 111(140?C) has been plotted in the cci-
Cirdinate system; too, viz.:
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 ?
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 23
r
= (ln ac6R6 aH2)
aC6H12
It is easy to make out, that the curve in the latter case does not corre-
spond to eq. (21). This is proved by the scatter of points of the straight line
presented in this system:
'log w = (log Ca 6116 aH2)
aC6H12
The failure of -curve II is a result of the lack of the coefficients yx;
taking into account these coefficients, we should obtain the relation:
_ aC6116 VC6116
aC61-112 7C611i2
The experimental curve II for a temperature of 180?C in the coordinate
system w = cp(ln all2) fulfills eq. (21) quite well.
According to the results of structural investigations, the increase of
nickel in the contact series of 0, 30 and 100% of Ni-content causes the
appearance of a more regular structure of the crystals. The more regular
structure of the crystal lattice comes forth in the kinetics of hydrogen ad-
sorption, and the expressions 42 and aH2 represept the quantitative ex-
pressions of its kinetics. The numerical value of aH2 decreases as the
amount of nickel content in the catalysts increases, in spite of the fact that
aH2 increases, too.
The quantity -tg a = (d42/daH2) on the other hand, is a constant value,
which proves that the changes of the structural properties of the investigated
catalysts are .continuous ones.
?
REFERENCES
STAT
1) F. Lihli H. Wagner and P. Zerrisch, Z. Elektrochem. 56 (1952) 612, 56(1952) 619.
2) F.Hund, Z. Elektrochem. 56(1952) 609.
_3) W. Langenbeck, Angew. Chem. 68 (1956) 453.
4) S. Strother and H. Taylor, J. Am. Chem. Soc. 56 (1934) 586.
5) S. Z.Roginski, Problemy kinetiki i kataliza, Vol. VI (Moslcwa - Leningrad, 1948).
6) S. T.Roginski, Adsorpcja i kataliz na nieodnorodnych powierchnostiach (Moskwa,
1948).
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 ?
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
IA-RDP80-00247A004200070001-6
STAT
ABOUT MECHANISM OF CATALYTIC CONVERSIONS
AND STRONG ADSORPTION OF UNSATURATED CYCLIC
HYDROCARBONS ON PLATINUM AND PALLADIUM
V. M. GRYAZNOV, V. I. SHIMULLS, V. D. YAGODOVSKH
Peoples' Friendship University,
named after Patrice Lumumba, Moscow, USSR
Abstract: The mass-spectrometric investigation of the gaseous products of ben-
zene vapour adsorption on platinum film at room temperature has shown the for-
mation of hydrogen and methane. in the uncondensable at -196?C products of
cyclohexene conversion on palladium film there has been found only hydrogen.
These results confirm the idea that the redistribution of hydrogen in the unsatu-
rated hydrocarbons is the combination of the dehydrogenation and hydrogenation
reactions. Besides these data reveal some peculiarities of the strong adsorption
of benzene on metal films.
1. INTRODUCTION
The investigations of cyclohexene, dyclohexadiene, benzene and cyclo-
hexane conversions on platinum and palladium films 1-12) have shown that
catalytic redistribution of hydrogen at room temperature and pressures less
than 10-2 torr may occur as the combination of the dehydrogenation and hy-
drogenation reactions of the original hydrocarbon. In the course of these
catalytic processes there has been observed the educing of gas which did not
condense at the temperature of liquid nitrogen and contained hydrogen. The
hydrogen is also splitted at the strong adsorption which takes place during
the first contact of any studied hydrocarbon with fresh platinum or palladium
film. The splitting of hydrogen at the adsorption of benzene on the films of
nickel, iron and platinum was observed earlier 13) but the data of gas anal-
ysis are not presented in that paper. For further elucidation of the interac-
tions between hydrocarbons and metal surfaces the mass-spectrometric in-
vestigation of the forming gases was carried out.
2. EXPERIMENTAL
The gaseous products of the strong adsorption of benzene vapour on the
platinum film were analysed with the help of radio frequency mass spectrom-
eter (omegatron) which had the glass bulb *. The benzene vapour introduced
into the adsorption bulb and omegatron by molecular flow through the capil-
* These experiments were made with participation of T.N.Ovchinnikova.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 '
'
STAT..
_ _
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
??
2
lary cut off with gallium valve. The pressure of residual gases before the
experiment was about 2 X 10-7 torr on the air scale of ionisation gauge. The
peaks of residual gases, methane, benzene and krypton for calibrating of
the omegatron have been measured. It has been found that the relation M/ e =
constif between the We-mass-to-charge ratio of resonance ions and the
radio frequency f is fulfilled satisfactorily till M/ e = 84.
. The mass-spectrum of benzene has been recorded at radio frequency
voltage 2 V which was used in the work 14). This spectrum had peaks M/ e =
74, 52, 39 except molecular ion peak. The decrease of radio frequency volt-
age to 0.8V allowed to observe the peaks 51 and 50 too. Peaks 74, 52, 51, 50
and 39 are-the most intensive ones according to data 15), but the relative
intensity of each of these peaks respective to the molecular peak is much
more in the experiments with the omegatron (see table 1). It seems likely
that peak 77 was not separated from peak 78. The relative intensities of the
peaks of methane mass spectrum have been found close to the data 15) (see
table 2). The intensity.of peak 14 has not been estimated because of the
presence of nitrogen admixture in methane.
The mass spectra of benzene and methane have been recorded during
the pumping of analysed vapour out of omegatron and adsorption bulb. How-
Table 1
The most intensive peaks of the benzene mass spectrum.
_
M/e
Relative intensities
- ?
Roberts and
This work
immediately
after admission
10 min
after admission
?
Johnsen i5)
r
78
- 100.0 ?
100
100
77-
18.4
_
-0
74
3.79 -
.-
8
1.7 -
52
- 17.5
70
-
51
14.7
80
460
50
12.0
63
? -
_ -39
. 11.3
-- 60
428
?
- Table 2
_
The intensities of the methane mass spectrum.
Mix
? ' -'
Relative intensities
-
Roberts and Johnsen 15)
This work
16 -
.15.
14.-.
.13.
100.0 ?
.. .
? 80.1. ??83
. - -8.28...-
.
_ -2.90
100
- '
2
STAT
Declassified in Part - Sanitized Copy Approved for Release 4/03/04 :
3IA-RDP80-00247A004200070001-6 ? ?-?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 - 3
ever, it was desirable to make the analyses of small quantities of gaseous
products of benzene chemisorption without pumping out the omegatron. In
this connection experiments have been carried out with introduction of the
definite quantities of benzene and methane in the omegatron.
The necessary quantity of vapour was introduced with a help of capil-
lary. The speed of molecular flow through the capillary was measured for
krypton whose adsorption on the glass walls of the apparatus at room tem-
perature was negligible. The corresponding 'data for benzene and methane
have been calculated as W= Wicr,/M/MKr.
It has been found that the iraensities of peaks 16 and 15 rise up linearly
with the quantity of methane admitted in omegatron. The other result has
been obtained by letting in benzene vapour in the omegatron out off from the
pump. Peak 78 dropped swiftly and peaks 1, 2, 14, 15, 16 rose. This indi-
cated the decomposition of benzene probably caused by the contact with hot
metal parts of the working omegatron. This undesirable effect was dimin-
ished by freezing the benzene in front of the omegatron at temperature of
liquid nitrogen. The introduction of benzene vapour into uncooled adsorption
bulb without platinum film brought about the appearance of peaks 15 and 16
in the mass spectrum of uncondensable gases. Special conditions for cooling
the low part of the adsorption bulb have been selected, and they prevent the
appearance of peaks 15 and 16 in the mass spectrum. Under these conditions
the opening of the magnet operated non-greased glass valve between the
omegatron and the adsorption bulb with benzene vapour increased the inten-
- sity of peak I alone. Therefore this procedure is good for revealing methane
if it appeared at the chemisorption of benzene on platinum film.
The platinum film was sublimed into the adsorption bulb by electric
heating of platinum wire. The bulb walls had the room temperature, the
.pressure during the sublimation was about 5 X 10-7 torr and dropped to
2 X 10-7 torr after cooling the wire. The platinum film was transparent as
the ones used in previous works 5-9), its geometrical surface was equal to
100 cm2. The benzene vapour was introduced into the adsorption bulb im-
mediately after the film sublimation. During the benzene introduction and
5 min after that the adsorption bulb had the room temperature. For the next
40 min the low part of the bulb was cooled by liquid nitrogen. The cooled
part of the bulb was not covered with platinum film. The cooling continued
for 27 min more after opening non-greased glass valve between the adsorp-
tion bulb and omegatron for freezing the benzene vapour which could pene-
trate into omegatron. through the glass valve.
?The mass spectrum of uncondensable at -196?C gaseous products of
benzene adsorption on platinum film is presented in table 3. The pressure
In the system was 1.5 X 10-7 torr. Peaks 15 and 16 indicate the formation
of methane. It is interesting to note that peak 15 is somewhat higher than
peak 16 in spite of reverse intensity correlation in the methane mass spec-
trum (see table 2). The following records of mass spectra of the benzene
,.adsorption products have shown the regular change of some peak-intensities
(see fig. 1). The extrapolation of straight lines for M/e= 15 and 16 in fig. 1
-with the help of the data of table 1 permit to evaluate the amount of methane
which has been formed before switching on the omegatron as 1014 molecules.
,
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 ?
:3IA-RDP80-00247A004200070001-6
-7
1
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
4
3
Table 3
The mass spectrum of uncondensable at -1960C gaseous
products of benzene adsorption on the platinum film.
mie
Intensity
(my)
Ave
Intensity
(mV)
m ie
Intensity
(m11)
1
64
8
12
16.5
7
2
23
12
2
17
6
6
1
14
42
18
2
6.5
3
15
37
28
48
7
15
15.5
6
29
15
7.5
16
16
33
40
2
STAT
The detection of methane in gaseous products of strong benzene adsorp-
tion on platinum film made usverify whether methane formed during the
catalytic redistribution of hydrogen in cyclohexene and cyclohexadiene. Pal-
ladium is a better catalyst for these reactions than platinum and the mirror
palladium film was used in the experiments with cyclohexene. The proce-
dure of film sublimation was similar to the described above, the reaction
bulb had the same size as the adsorption bulb used in the experiments with
platinum film and benzene. The cyclohexene vapour was admitted in the re-
action bulb with fresh palladium film till the pressure was about 0.1 torr and
was heated during 15 min at 50?C. In accordance With the previous data 6,9)
the uncondensable at -1960 gas was found in the reaction bulb. The pressure
of that gas was-close to 3.5 x 10-3 torr. The gas was introduced into the
ionisation region of MX-1302 mass spectrometer through the needle valve.
The speed of gas introduction was permanent in all the analyses and in the
course of calibration on hydrogen before sublimation on the palladium film
on the walls of reaction bulb. Mass spectrum of uncondensable products of
? cyclohexene conversion coincided with the hydrogen mass spectrum, the
pressure dependence of peak 2 intensity (black circles on fig. 2) was very
close to that for pure hydrogen (light circles). Thus methane has not been
found in the gaseous products of cyclohexene conversion on palladium film.
Fig. 1. The aependence of peaks intensities of mass spectrum of
gaseous products of benzene adsorption on platinum film on
the time after this products admission into omegatron.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 : ?
:31A-RDP80-00247A004200070001-6 - ?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 -
5
I I
8
taimm
Fig. 2. The pressure dependence of peak 2 intensity for gaseous
products of cyclohexene conversion on palladium film
(black circles) and for hydrogen (light circles).
3. DISCUSSION
The detection of methane in products of benzene adsorption on fresh
platinum film throws light upon the mechanism of interaction between ben-
zene molecules and metal surface. There are some data 16) about formation
of methane at high temperature hydrocracking of hydrocarbons on iron, co-
balt and nickel catalysts. For high temperature the mechanism of methane
formation from the elements was proposed in paper 17). The main role in
that mechanism was ascribed to the external C atoms of graphite cristal
lattice. It was supposed that the hydrogenation of double bonds between such
atoms led to breaking these bonds and forming of pairs of CH3 groups linked
to neighbouring "benzene" elements of graphite net. This scheme was
used 18) for explanation of kinetics of methane formation from the elements.
The strong adsorption of benzene molecules on platinum with partial
splitting of H atoms is probably connected with appearance on the metal
surface of peculiar radicals which are bound with metal atoms. The hydrogen
being released may join adsorbed benzene molecules. As a result of such
redistribution of hydrogen among adsorbed molecules there appears a whole
series of products, and it is quite possible that tensions occurring in rings,
connected, in situ of splitted H atoms, with several metal atoms, condition
a deep destruction. Among the products of this destruction one may expect
methane and CH3 radicals. The last assumption conforms with reverse in-
tensity correlation of lines 15 and 16 than the usual one for the methane
mass-spectrum.
The increase of methane and hydrogen formation with contact time of
-benzene and platinum film may be caused by displacement of the adsorbed
benzene molecules along the metal surface. The molecules previously ad-
sorbed in the places of low adsorption potential pass to the strong adsorp-
tion centres where the destruction can occur. Such displacement of benzene
molecules takes about 10 mm at room temperature 4). The cooling of the
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
4
low part of the adsorption vessel by liquid nitrogen may decrease the tem-
perature of platinum film under 200C and therefore moderate the mobility of
adsorbed benzene molecules. This explanation is not the only possible one.
The question about the methane formation at the strong adsorption of
cyclohexene is to be solved by special investigation. In the above exper-
ments the cyclohexene pressure was comparatively high and the catalytic
dehydrogenation could make the detection of strong adsorption products dif-
ficult.
In considering the mechanism of catalytic redistribution of hydrogen in
unsaturated hydrocarbons on carbons 9,11 12) the essential point was the
evaluation of quantities of uncondensable at -1960C gas which was taken for
hydrogen. This assumption is confirmed in this paper. It is a new argument
in favour of the idea that the hydrogen redistribution in hydr.ocarbons on
metal catalysts is the combination of dehydrogenation and hydrogenation re-
actions but not a one-stage process.
STAT
REFERENCES
? 1) V.M. Grjaznov and V .D . Yagodovskii, DolcLady Alcad.Nauk SSSR 116 (1957) 81.
2) V. M. Grajznov, V.D . Yagodovskii and A. M. Bogomolnii,- Ho Dju ok, Doklady Akad.
Nauk SSSR 121 (1958) 499.
3) V.M. Grajznov, V.D . Yagodovskii and M.K. Charkviani, Vestnik Moskovsk.
Univers.Ser II, Chemistry (1960) no. 1, p.11.
'4) y .M. Grajznov, V. I.Shimulis and V.D .Yagodovskii, Doklady Alcad.Nauk SSSR 132
? (1960) 1132.
5) V.M.Grajznov and V.I. Shimulis , Kinetika Kataliz 2 (1961) 534.
6) V. M. Grajznov and V. I.Shimulis, Kinetilca Kataliz 2 (1961) 894.
7) V.M.Grajznov and V.I. Shimulis, Vestnik Moskovsk. Univers. Ser. 11, Chemistry
- (1961) no. 6, p.25.
'.8) V:M.Grajznov, V. L Shirnulis and T . V .Dillingerova, Vestnik Moskovsk. Univers .
_ Ser. II, Chemistry (1962) no. 2, p.26.
9) V. M. Grajznov and V. I.Shimulis, Doklady Alcad.Nauk SSSR 139 (1961) 870.
10) V. M. Grajznov, V. D ? Yagodovskii, E . A. Savel' eva and V . I. Shimulis , Kinetilca
Kataliz 3 (1962) 99. ?
11) V. M. Grajznov, Investigation of Kinetics and Mechanism of Conversions of Some
Hydrocarbons on Metals, D. Sci. Thesis Moscow (1962).
12) V. M. Grajznov, Uspekhi Khim. 32 (1963) 433.
13) R.Suhrman, B.Hahn and G.Wedler, Naturwissenschaften 44 (1957) 60.
14) A. P. Averina, Pribori Technilca Eicsperim. (1962) no. 3, p.123.
15) R.H.Roberts and S. E. J. Johnsen, Analyt. Chem. 20 (1948) 1225.
16) H.Koelbel, H. B. Ludwig and H. Hammer, J. Catalysis 1 (1962) 156.
17) C. W. Zielke and E. Gorin, Ind. Eng. Chem. 47 (1955) 820.
18) K. Hedden, Z. Eleictrochem. 66 (1962) 652..
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 :
CIA-RDP80-00247A004200070001-6 ? %
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
SYSTEMATIC STUDY OF FACTORS INFLUENCING THE
SELECTIVITY OF DECOMPOSITION REACTIONS OF
ALCOHOLS ON PURE AND SUPPORTED MIXED OXIDES
_ I. BATTA, S. BORCSOK, F. SOLYMOSI and Z G. SZABO
Institute of Inorganic and Analytical Chemistry,
University of Szeged, Szeged, Hungary
Abstract: To contribute to the problem of selectivity, kinetical parameters of
the decomposition reactions of isopropyl alcohol, depending on the procedure of
preparation and the poisoning of the catalyst, were measured as accurately as
possible. The poisoning was achieved not only with cations, but with anions,
too. A very marked change in selectivity was caused already by a relatively
small amount of sulphate ions, since in their presence the dehydration reaction
became dominating. Special care was taken to establish a correlation between
the electric properties and catalytic activity. Considering the other experi-
mental data also it can be stated as a rule that the more ionic a character a
catalyst oxide possesses, the more it dehydrogenates, and the nearer the oxide
approaches. the covalent character, the greater is the dehydration.
1. INTRODUCTION
In this paper we summarize the results of experiments on the catalytic
decomposition of isopropyl alcohol on zinc oxide. The catalytic decomposi-
tion of isopropyl alcohol may be twofold: dehydrogenation and dehydration.
The-process - as we shall see - goes under certain conditions in both
directions on zinc oxide, too. Thus the study of this catalytic reaction
renders a suitable model for the investigation of the problem of selectivity.
The reaction belongs to type U of Wheeler's classification.
- Literature contains numerous experimental data on the catalytic de-
composition of alcohols on zinc oxide. The overwhelming majority of
authors regards zinc oxide as a dehydrogenating catalyst. However, there
are diverse opinions concerning the mechanism of dehydrogenation and selec-
tivity, 1;e?pectiyely. None of these have so far rendered a generally accepted
theory. The mechanism as suggested by Eucken 1) and Wicke 2) is based
on geometrical yiewpoints and really a great many of the experimental
facts are in accordance with-the deductions drawn from this view. For
example, the deuterium content of water formed in the dehydration of al-
coholts increased when the surface of the oxide was covered with deuteri-
um 3). Further the geometrical view is supported by the experimental
facts, too, referred to by Marsh 4), namely that with several oxide cata-
lysts, whether previously preheated or not, the same selectivity factor
Declassified in Part - Sanitized Copy Approved for Release 20/04: _ STAT
STAT
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
.C1
h
STAT
Another standpoint is taken by Schwab 5). According to him catalysts
treated at high temperatures are dehydrogenating while those treated at low
temperatures dehydrating. The difference comes from the surface condi-
tions. Catalysts formed at high temperatures are dense, have smooth sur-
faces and dehydrogenation takes place on them in one-point adsorption,
while catalysts prepared at low temperatures have large surfaces and de-
hydration occurs in the pores according to a two-point adsorption.
Wheeler 8) has given considerable support to Schwab's theory when pointing
out theoretically the correlation between the pore structure and the activity
of the catalyst, what influences the temperature coefficient, the kinetic
order and the poisoning characteristics, and hence the selectivity of the
catalyst. For selectivity type II, studied by us, the rule is generally valid
that reactions taking place at low pressures are accelerated by catalysts
with small pores, those operating at high pressures by catalysts with large
pores.
However, since Wheeler in his theory considers geometrical factors
only, such as pore size, his results, though being very interesting and
valuable for experimental purposes, do not give satisfactory solution con-
cerning the mechanism, i.e., the sequence of elementary processes of the
chemical conversion. Namely, it is not substantiated how the reaction is
influenced by the material of the catalyst (supposing the same surface
characteristics).
Another possible interpretation of selectivity is rendered by Hauffe's
theory when discussing the role of the electronic factor in the alterations of
selectivity, so successfully applied in the theory of catalytic processes on
metals and semiconductors.
' In our own experiments we have started from the observation that the
selectivity of A1203, applied widely in the dehydrogenation of alcohols, is
different in the case of oxides prepared from nitrate and from sulphate by
precipitation. Our experiments were aimed at studying how the activity of
pure ticonductor zinc oxide, a dehydrogenating catalyst, can be altered so
that it also operates as a dehydrating agent.
2. EXPERIMENTAL
2.1. Apparatus
Catalytic decomposition reactions were carried out in the so called open
Schwab reactor modified by us 8). The evaporation of alcohol at a constant
rate, i.e., the constant feed concentration was assured by constant tem-
perature adjusted by an ultrathermostate. Catalysts were applied in the
form of small grains of pellets. The stream of alcohol vapour penetrates
the catalyst layer from top downwards. The rate of the catalysed reaction
was controlled by soap-bubble flow-meters after condensing the products
boiling above 0?C in a water cooler. The first flow-meter controlled the
total amount of gas. Thereafter propylene was frozen out by liquid air,
thus the second flow-meter indicated only the amount of hydrogen, i.e.,
F
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
.7,1A-RDP80-00247A004200070001-6 _
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
the rate of the dehydrogenation reaction.
? 2.2. Preparation of catalysts
'Catalysts used in our experiments were all prepared from the same
:sample of zinc oxide. Zinc oxide was dissolved in either c.p. nitric acid
or in a mixture of C. p. nitric acid and C. p. sulphuric acid of various ratios.
Then the complex [Zn(NH3)612+ salt was formed with excess ammonium
hydroxide. Zinc hydroxide was Separated by boiling this solution and by
expelling the ammonia. The precipitate was dried, decomposed at 400?C
while slowly raising the temperature, and tempered for five hours at 500
or 800?C, respectively. These were the catalysts prepared by coprecipita-
lion. In case of catalysts prepared by suspension, zinc oxide was suspended
?in a solution of certain amount of zinc sulphate, phosphorus pentoxide, boric
acid and zinc chloride, then dried and treated at 500 or 800?C, respectively.
The sintered catalysts were powdered and compressed into pellets. The
' surface of the ready catalysts was determined by BET method from the
adsorption measured at the temperature of liquid nitrogen. Electric
resistance was measured on tablets pressed between platinum plates, in a
vacuum apparatus at different temperatures, first in oxygen, then in vacuo,
in hydrogen and in oxygen again.
.3. RESULTS
. Our experimental data da not allow us to make a decision between the
? two points of view referred to in the Introduction as they support both ? -
-considerations in. certain aspects. . _
? ?The poisoning with cations of higher or lower valency, and this is what
? _increase? ordecreases the electron conductivity, only influences the dez?
hydrogenation reactions so that it changes the energies of activation but not
the. selectivity 9). When different amounts of sulphate ions were added to
zinc. oxide and the catalyst and treated exactly in the same manner..as in
case of pure zinc.okide, catalysed the dehydration reaction to a marked ..
extent: This experimental result may give rise to the problem of pOisoning
:With anions, To this end we attempted to poison zinc' oxide with other anions.
In the 'case of phosphate and borate,- although a slight decrease of the .?
..selectivity factor was observable (by selectivity factor we mean the ratio
.of conversion towards dehydrogenation and of total conversion) the effect
Was far -weaker than that due to the sulphate. Ni) effect was observed when
?silicic acid and chloride ions were added: .The latter-can be ascribed to the
fact that.at the:temperature of the eXperiment.zinc:".chloride is too volatile
.."
and so the chloride content diminishes appreciably:during pre-treatment.
Table I' shows that by increasing the S03-Content the energy .of activation of
? the total conversion changes considerably, the energy of activation of the
it dehydrogenation reaction increases, while that of the dehydration reaction
-decreases.
BET measurements showed that by the addition of sulphates the surface
-T. of zinc oxide is markedly enlarged. The compact catalyst changes into one
? .
Declassified in Part-- Sanitized Copy Approved for Release 2014/03/04
CIA-RDP80-00247A004200070001-6
L:
? Declassified in Part- Sanitized Copy Approved for Release 2014/03/04: CIA-RDP80-00247A004200070001-6
Table 1 ? ?
Values of ostalytlo aotivtty, anemic, area and resistivity for various specimens.
,
?ml
? ?
Catalyst
?
'
,
? analysis;
. mole% of
poisoning
-
'
total 1110+ ,
? ?ity
tivit/
V(H24.po
min-1
g1 at
Soec
sur-
face
ma/g
Specific
softy-
ity
V(H. 24?Pr) 2
ml min-
m
eeleotiv-
VH2 .
Et?Aai
kcal
.e -.1
moi .
kcal
mole-1.
EH2
koosile-1
m
p(ohm Cm)'
,at 340?C'
.
- Ep
kcal .
mole-1
V(H2+1,r)
:
100-2.
coprecipitated, treatment at 500?C
-
for 5h.
: 4
? ZnO
-
1.08
7.6
0.14
93-97
22.2
22.4
-
8. 5. 106(02)
17.1
8
? Zn0?14303
0.2% 803
0.32
6.0
0. 085
36-45
34.0
33.6
33. 4
3. 9 (H2)
148
ZnO-+603
2.2% sos
0.85 .
25.9
0.034
29-50
31.9
27.6
32.8
5
'ZnO-+603
3.8% 303
1.99
21.10.094
8-15
31.1
' -
21.2
3.1. 106(02)
20.4
,
copra? pitated, treatment at 800?C
.193 (H2)
for 5h.
? 8-
ZnO .
-
1.27
6.2
0.21
90-98
21.0
21.4
29.2
0. 63. 106(02)
11.2
.
19. 3(H2)
12
ZnO-4803
0.2% 803
0.57 -
5.9
0.097 ?
80-92
24.3
25.4
28.3.
.
147 '
? Zn0+303
19% 803
0.38
27.8
0.014
33-50
37.0
30.2
44.7
.145 ?
Zn04803
3.3% 303
0.26
18.8
0.014
20-28
41. 6 ?
36.8
37.4
52.9. 106(02)
21.8
?
?
suspended, treatment at 500?C for 5 h
365 (H2)
.
152.
.
ZnO
- ,
3.57
12.9
0.28
89-96
26.0
23.8
-
3'
Zn04603
0.1% 803
0.88 '
20-30
37.4
24.5
41.2
.
? 1
Zn0+603
1.0% 303
1.17
10.4
0.11
8-10
31.7
31.7
. 9
' Zn0+13203
1.0% B203
3. 8589-94
27.9
29.5
13
Zn0+8203 .
5.0%B203
3.58
113.4
0.22
93-99
21.3
24.0
11
Zn04CI
0.24% Cl
3.2
88-92
30.7
31.5
- .
.160
' ZnO+P205
1.0% 1'205
1.37 .
94-98
32.5
31.3
?
suspended, treatment at 800?C for 5 h. -
? 10
Zn0+13203
1.0% B203
2.81
95-100
-
-
151
ZnO+P205 .?
1.0% P205
1.01
86-98
34.1
28.0
-
154
Zn0+13,02
5.0% P205
1.64
92-100
28.6
27.3
153
Zn0461102
2.0% S12
2.91
96-100
29.8
29.3
-
54
MgO (50ec)
-
0.10
96-100
27.0
26.6
15
240+603 (50fP C)
0. 6% 303
0.23
91-94
27.4
26.6
155
Z n2 P207 (400?C)
-
1.0
o
-
-
53.4
17
? Zn2 P207 (800?C)
- .
0.19
11.0
0.01,1
o
-
-
37.0
18
ZnSO4 (400?C)
-
9.3
_
0
-
-
.41. 0
.?
? x?determinad in 100 mm Hg oxygen or hydrogen, reap. ?
Ix determined in 100 mm Hg oxygen
Pr - propylene .
btu
STAT
DeClassified in Part - Sanitized Copy Approved for Release 2014/03/04: CIA-RDP80-00247A0047nnn7nnn1_R
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
the rate of the dehydrogenation reaction.
2.2. Preparation of catalysts
Catalysts used in our experiments were all prepared from the same
sample of zinc oxide. Zinc oxide was dissolved in either c. p. nitric acid
or in a mixture of C. p. nitric acid and c. p. sulphuric acid of various ratios.
Then the complex [Zn(NH3)6]2+ salt was formed with excess ammonium
hydroxide. Zinc hydroxide was separated by boiling this solution and by
expelling the ammonia. The precipitate was dried, decomposed at 4000C
while slowly raising the temperature, and tempered for five hours at 500
or 800?C, respectively. These were the catalysts prepared by coprecipita-
tion. In case of catalysts prepared by suspension, zinc oxide was suspended
in a solution of certain amount of zinc sulphate, phosphorus pentoxide, boric
acid and zinc chlroride, then dried and treated at 500 or 800?C, respectively.
The sintered catalysts were powdered and compressed into pellets. The
surface of the ready catalysts was determined by BET method from the
adsorption measured at the temperature of liquid nitrogen. Electric
resistance was measured on tablets pressed between platinum plates, in a
vacuum apparatus at different temperatures, first in oxygen, then in vacuo,
in hydrogen and in oxygen again.
3. RESULTS
Our experimental data do not allow us to make a decision between the
two points of view referred to in the Introduction as they support both
considerations in certain aspects.
The poisoning with cations of higher or lower valency, and this is what
increases or decreases the electron conductivity, only influences the de-
hydrogenation reactions so that it changes the energies of activation but not
the selectivity 9). When different amounts of sulphate ions were added to ?
zinc oxide and the catalyst and treated exactly in the same manner as in
case of pure zinc oxide, catalysed the dehydration reaction to a marked
extent. This experimental result may give rise to the problem of poisoning
with anions. To this end we attempted to poison zinc oxide with other anions.
In the case of phosphate and borate, although a slight decrease of the
selectivity factor was observable (by selectivity factor we mean the ratio
of conversion towards dehydrogenation and of total conversion) the effect
was far weaker than that due to the sulphate. No effect was observed when
silicic acid and chloride ions were added. The 'latter can be ascribed to the
fact that at the temperature of the experiment zinc chloride is too volatile
and so the chloride content diminishes appreciably during pre-treatment.
Table 1 shows that by increasing the SO3 content the energy of activation of
the total conversion changes considerably, the energy of activation of the
dehydrogenation reaction increases, while that of the dehydration reaction
'decreases.
BET measurements showed that by the addition of sulphates the surface
of zinc oxide is markedly enlarged. The compact catalyst changes into one
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 -
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
6.
Fig. 1. Arrhenius plot of one exp. of isopropyl alcohol
'decomposition on ZnO + 2.2 mole % SO3 catalyst
sintered at 5000C for 5 h.
???..
STAT
? with looser structure. This supports the Schwab-Wheeler theory. But a very
Important difference must be emphasized here, namely that the marked
's increase of dehydration, i. e. the decrease of the selectivity factor due to
_ 'poisoning with SO3, .is not wholly the result of poisoning of dehydrogenating
centres. At the same tithe the dehydration centres remain active since the
specific activity of catalysts poisoned with sulphate ions increases for the
dehydration reaction in absolute value. Therefore the case is not that the
- dehydrogenation centres being covered no longer take part in the process
and dehydration becomes more apparent.
Variations in the electric conductivity due to poisoning with sulphate
ions have also been studied, namely in vacuo, in oxygen and hydrogen
? atmospheres, and at the same time the energy of activation of conductivity
has also been taken in each ambient atmosphere. No very marked differ-
- ences were found. ?
This statement is against the effect of the electronic factor, but it is
? in accordance'with Schwab's conception. Anyway, we are of the opinion
that the electronic factor cannot be entirely left out of consideration, as it-
Is shown by changes in the selectivity factor of the decomposition of formic
aCid on a 'number of other oxides, poisoned with various cations 1%. On
:..n-conductor titanic dioxide the selectivity of decomposition of formic acid
-- is shifted towards dehydration when as a result of poisoning the electron
-pincentration decreases. On the contrary, at p-conductor chromic oxide
vtith.a decrease in the number of electron holes the rate of dehydration
? reaction also inereased. Thus both on p- and n-conductor oxides, when the
Declassified in Part - Sanitized dopy Approved for Release 2014/03/04 :
;IA-RDP80-00247A004200070001-6
?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
lot
a
6
4
2
to'
6 ?
4 ?
- to zo
_ 4. mi.-
2,2
Fig. 2. The dependence of specific resistivity on temperature.
1. ZnO (sintered at 5000C for 5 h)
2. ZnO + 3.8 mole % 503 (sintered at 500?C for 5h)
x, A upwards in oxygen
?, o downwards in oxygen
conductivity of the catalyst decreases, the dehydrogenation reaction is
shifted to the second plane.
The structure of catalysts and the building-in of poisoning ions, respec
tively, was controlled by X-ray examinations.
The sulphate ions penetrate the zinc oxide lattice very easily but other
poisoning anions do not. This is in accordance with the observation that
the latter ions do not affect the selectivity.
4. DISCUSSION
. Taking the experimental data above into accOunt we'think that it would
be erroneous to regard the geometrical factor to be the only significant one
in the determination of selectivity. The function suggested by Eucken 1)
(radius of the cation)3
(mole volume per cation)(charge of cation)
_
_
? ?
?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 :
.7,1A-RDP80-00247A004200070001-6 .
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
is, perhaps, correct basically, but the deductions made from it are wrong,
if the size of the cation, as in Balandin's multiplet theory, is regarded as
an independent variable of the catalytic efficiency. The scattered experi-
mental data does not allow that Eucken's function should be regarded valid
strictly, with,respect to the exponents. However, regarding that the ratio
of the valency of a cation and its radius gives the polarizing power, then the
changes of selectivity can be related to the character of bonds in the cata-
lyst. Namely, the greater the polarizing power, the more covalent will be
the bond formed and the greater the dehydration. In case of low ionic charge
and large size the bond between the closely or almost closely packed oxygen
ions and the cation is more ionic in character. Oxides of more positive
metals are dehydrogenating. Therefore it can generally be said that the
rnore ionic a bond an oxide possesses, the more strongly dehydrogenating
it is and the nearer it approaches the covalent bond character, the more
dehydrating it is. Knowing this it is obvious that the selectivity factor is
not determined by the n- orp- conducting characteristics of semiconductor
oxides, but by the nature of oxygen-cation bonds.
The variation of selectivity due to the basic or acidic character of
oxides can also be interpreted so that the strength of the hydrogen coupling
to non-deformed (oxides of more positive metals) and to deformed (acidic
oxides) oxygen ions is considered. In the first case the hydrogen coupling
is strong, i.e., we shall have dehydrogenation and in the second the hydro-
, gen coupling is weak, but a strong hydroxyl coupling becomes possible,
consequently dehydration takes place.
Although data on the selectivity factor are rather scattered the
following series of oxides can be written:
" Dehydrogenation
? --
CaO MgO ZnO ?Fe0 ? F-e203 Cr203 Ti02 A1203 Si02 W03
Dehydration
This classification, although from a different point of view, to some
? extent has a similar content to Schwab and his coworkers' statement on the
role of Lewis-acids and Broensted-acids 11).
That here the electronic state of the oxygen ions is a controlling factor
Is proved also by the data obtained on ZnSO4 and Zn2P07. In these salts
? the oxygens exhibit covalent character and thus the catalytic effect is com-
pletely a dehydrating one.
? ? The Authors express their thanks to Mr.K.JAky and Mrs.E. Takacs
for their valuable assistance.
STAT
'7".? ?-- , . .
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: ?'
CIA-RDP80-00247A004200070001-6 ? . ?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 9
STAT
REFERENCES_
1) A. Eucken, Naturwiss. 36 (1949) 48;
A. Eucken and K. Heuer, Z. physik. Chem. (Leipzig) 196 (1950) 40.
2) E. Wicke, Z. Elektrochem. 53 (1949) 279.
3) H. J. Leugering, Heterogener Wasserstoffisotopen-Austausch an A1203 -Oberfla -
chen mit verschiedener Hydroxylgruppenbelegung. Thesis (GOttingen, 1948).
4) P. Marsh, The Mechanism of Heterogeneous Catalysis, ed. by J. H. de Boer
.(Elsevier, 1960) p. 50.
5) G. M. Schwab and Elly Schwab -Agalliclis, J. Am. Chem. Soc. 71 (1949) 1806.
6) A. Wheeler , Catalysis, ed. by P. H. Emmett , Vol. 2. (New York , 1955) p.105.
7) K. Hauffe, Angewandte Chemie 67 (1955) 189.
8) Z. G. Szab6 and F. Solymosi, Z. Elektrochem. 63 (1959) 1177.
9) G. M. Schwab, XVIIth International Congreis of Pure and Applied Chemistry
(1959).
10) Z. G. Szab6 and F. Solymosi, Acta Chim. Hung. 25 (1960) 145.
11) G. M. Schwab, 0. Jenkner and W. Leitenberger, Z. Elektrochem. 63 (1959) 461.
.1 ? .!
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 :
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
SELECTIVE HETEROGENEOUS CATALYTIC OXIDATION
OF HYDROCARBONS
t. I. IOFFE
Institute for Organic Sgmi-Products and Dyes, Moscow
and
L. Ya. MARGOLIS
Institute of Chemical Physics, Academy of Sciences, Moscow, USSR
Abstract: The kinetics of selective heterogeneous catalytic oxidation of hydro-
carbons with molecular oxygen is a function of two factors: of the site where a
hydrocarbon molecule is attacked with oxygen, which depends on the sum of
crystalline, complexing and semiconducting properties of the catalyst, and of
the ratio of adiorption to description rates for all intermediate and end-pro-
ducts of the reaction.
Investigations were made on the vapour- and liquid-phase oxidation of ole-
fines and aromatics, on simple and complex oxide systems containing vana-
dium, molybdenum, tungsten, copper, chromium and bismuth oxides, and also
on silver and platinum catalysts. The phase composition of catalysts was stu-
died by X-ray analysis, ESR spectra were interpreted, and the electrophysical
properties of catalysts were investigated in vacuum and in the reactant medium.
The oxide catalysts were found to be a - and rr- active. The former are re-
sponsible for breaking of CH bonds, and the latter of C=C bonds. The type of
activity is a function of electronic structures both of individual cations and of
? the whole solid crystal.
For all systems studied oxidation proceeded by a parallel-consecutive
scheme.
STAT
1. INTRODUCTION
Successive development of the catalytic oxidation of hydrocarbons depends
on the working out of a proper theory, the determination of the nature, of
elementary acts and the relation of these to the catalyst surface properties.
Intermediate steps of heterogeneous reactions proceed on the catalyst
surface. The catalytic process of organic oxidation may be conceived as a
chain of conversions the route of which is determined by the site of oxygen
attack on the molecule. Consequently the problem of selective catalytic oxi-
dation is essentially that of determining the probable site of oxygen inclusion
into organic molecules of various solids.
Hydrocarbon oxidation involves the following steps:
1. Transier of electrons from the hydrocarbon molecule to the catalyst (che-
misorption of the hydrocarbon).
2. Transfer of electrons from the catalyst to the oxygen molecule (chemi-
sorption of oxygen).
3. Interaction of the charged particles formed (a radical, an ion-radical or
a complex) yielding the oxidation products.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: STAT
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
4. Desorption of oxidation products.
_ The oxidation catalysts should possess two properties: that of forming
acceptor-donor bonds with organic compounds and oxygen, and that of trans-
ferring electrons from one reacting particle to another. The mechanism of
electron transfer may be different for various types of solids.
2. METAL OXIDES
In order to establish the dependence of certain hydrocarbon oxidation
rates on the electronic properties of a surface and on metal structures, the
authors, together with Lyubarskii, Ezhkova, Krylova, Kazanskii avidksaev,
have investigated a number of oxide catalysts in various reactions
The characteristic data on catalysts is given in table 1.
Table 1
Characteristic data on oxide catalysts.
Catalyst
Composition
atomic
percentage
Surface
In%
V205 + Mo03
0 to 25 of Mo
1 to 2.0
30 to 80 of Mo
1 to 2
V205 +Cr203
0 to 80 of Cr
1 to 2
V205 +Co203
0 to 80 of Co
1 to 2
V205 + P205
0 to 70 of P
1 to 2
V205 + Li2O
0 to 50 of Li
1 to 2
Mo03 + B1203
0 to 80 of Bi
1.4 to 1.5
W03 + Bi203
0 to 80 of Bi
7.0 to 8.0
V205.+ B1203
0 to 43 of Bi
2.6 to 2.8
V205 + P205
0 to 43 of P
2.6 to 2.8
Mo03 + P205
0 to 43 of P
1.0 to 1.2
? W03 + P205
0 to 43 of P
2 to 12
STAT
It was -shown by X-ray analysis that mixed catalysts (excluding oxides
involving pure V205 and Mo03 may be divided into two groups: solid solu-
tions and compounds. ,The results obtained on the phase composition of the
metal oxides studied are summarized in table 21/4. Certain oxide systems
-were studied earlier by other investigators 4-7), but their results are not
always consistentwith those obtained by the authors of this paper.
? The phase composition of catalysts was determined by X-ray analysis
, in a PKD chamber, using a chromium anode (ka-irradiation).
The activity of catalysts was studied from the kinetics of benzene and
propene oxidation in flow-circular and flow apparatus 8,9) with respect to
the benzene yield in maleic anhydride, and from the constants of acrolein ?
formation rates with respect to propene. Selectivity was determined from
the amount of the main oxygenated product yielded by the hydrocarbon.
The structures were investigated from ESP spectra, using a technique
described before 10).
- ? The electronic properties of a surface (the electron work function ch
1- ? = -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
I I
.71
3
were determined from the contact potential difference catalysts and standard
electrodes measured by means of the vibrating condensor technique using an
apparatus the schematic view of which was given previously 11).
:Table 2
Phase composition of mixed oxide catalysts.
Catalyst
Composition
atomic
percentage
Phase composition
V205 + Mo03
0 to 30 of Mo
solid solution
V205 + Mo03
30 to 50 of Mo
V MoOx
V205 + Cr203
0 to 50 of Cr
VCr04
V205 + Co203
0 to 30 of Co
V205 + CO2V207
V205 + Co203
30 to 50 of Co
V205 + Co304 + Co2V207
V205 + P205
0 to 3.5 of P
solid solution
V205 + P205
3.0 to 50 of P
VP0x
V205 + Li20
0 to 50 of Li
LiVOx
Bi203 Mo03
Mo03 + B1203
0 to 50
B1203. 2Mo03
Bi203 3Mo03
Bi203 WO3
W03 + Bi203
0 to 50
43W03
Bi203
The activity and selectivity of mixed oxide catalysts were correlated
with their electrophysical properties and structures. Following were the
results obtained.
2.1. Oxidation OY benzene to maleic anhydride
The maximum activity of a mixed vanadium-molybdenum catalyst corres-
ponds to the solubility limit for molybdenum oxide in vanadium pentoxide.
The selectivity, the work function variations and the intensity of ESR signals
characteristic of the concentration of V4+ ions in the V205 lattice were ob-
served to change in the same sense (fig. 1).
A small increase in the catalytic activity of V205 is observed for a
vanadium-chromium catalyst in the presence of minor amounts of Cr203
(about one molecular per cent of Cr203) the activity and selectivity of the
catalyst changing again in the same sense (as in the case of the V205 +
+ Mo03 system).
Addition of more than one percent of Cr203 will decrease the activity
and selectivity of the catalyst (fig. 2).
With mixed vanadium-chromium catalysts two ESR signals are observed
simultaneously: a broad signal (5 = 400 gauss) and on the background of it a
narrow one (5 = 150 gauss) (fig. 3). The narrow line may be ascribed, by
analogy with the V + Mo system, to V4t ions appearing in the solid solution
of Cr203 in V205. The intensity of this signal is consistent with the cata-
lytic activity and selectivity.
The catalytic activity, ?selectivity and electron work function of this
system change in the same sense as for the V205 + Mo03 system. When the
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
STAT
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
4
mot%
MR
70
60
50
40
30
20
#0
10
ditfey
4123
20 40 60 BO 100
nvi?6
- Fig. 1. V205+ Mo03 system.
I. Activity. Molar percentage of maleic anhydride
with respect to amount of benzene.
2.Selectivity, Percentage with respect to the benzene reacted.
3.6cp in vacuum at 20?C (eV).
4. Acp after treatment with a benzene-air mixture at 300?C (eV).
5. Content in V4+ after contact with the catalyst.
STAT
catalyst surface is treated with a benzene-air mixture, the maximum Ay)
value corresponds to the maximum selectivity of the process.
Addition of cobalt oxide to V205 will decrease the activity and selecti-
vity of the catalyst; the work function value will also drop, though a small
increase inAcp would be observed with addition of 1% of Co203, the reaction
selectivity remaining unaffected (fig. 4). The V205 + Co203 system shows
no ESR signals.
Investigations on the activity of vanadium-phosphorus catalysts in the
oxidation of benzene to maleic anhydride have shown that the influence of
P205 is negative. With an increasing content of P205 in V205 the activity
and selectivity of a sample show a monotonic decrease.
ESR signals from vanadium-phosphorus catalysts involving more than
5% of P205 represent a single line of about 200 gauss. With increase in
P205 the intensity of ESR signals becomes higher. A correlation with X-ray
data shows? that the signal is due to formation of a chemical V205 + 13205
compound.
The addition of P205 also decreases the electron work function, as well
i?as the activity and selectivity of the catalyst. Only at phosphorus concentra-
tions up to 5%, when a solid solution is possible, the Ay) value will increase
r
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
1
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
I I
STAT
V205 + Cr203 system.
1. Activity. Percentage of maleic anhydride with
' respect to the amount of benzene.
? 2. Selectivity. Percentage with respect to the benzene reacted.
3. Acp in vacuum at 20?C (eV).
4. ap after treatment with a benzene-air mixture at 300?C.
,
Fig. 3. ESR spectra of vanadium-chromium catalysts.
' a) Composition: 5 mol% of Cr203 + 95 mol% of V203.
b) Composition: 50 mol% of Cr203 + 50 mol% of V203.
by 0..1'eV without changing the characteristics of the catalytic process.
. A change in the catalytic activity of V205 with addition of Li20 in benzene
oxidation to maleic anhydride resulted in a monotonic decrease in activity and
_
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 :
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
6
40 60 80 400
mire Co,03
STAT
Fig. 4. V205 + CO203 system.
1. Activity. 'Percentage of maleic anhydride with respect to
the amount of benzene.
2. Selectivity. Percentage of maleic anhydride with respect
to the benzene reacted.
-3; Acp in vacutun at 20?C (eV).
4. cp-after treatment with a benzene-air mixture at 300?C (eV).
selectivity. ESR signals increasing in intensity with a greater amount of the
substance added are observed for the V,05 + Li20 system. These are due
to ri+ ions in the chemical compounds formed by interaction between V205
and oxides of alkali metals.
- A monotonic decrease in the work function with increasing amount of Li20
is observed for the V205 + Li20 system, and, consequently, the activity and
- work function will change in the same sense.
2.2. Oxidation of propene to acrolein
The oxidation ofpropene to acrolein 12) over vanadium and molybdenum
oxides was investigated earlier and the rate of acrolein fromation was shown
to decrease continuously with increasing concentration of Mo03 in V205.
Variations in the selectivity of propene oxidation and in the electron work
function depending upon the composition of mixed catalysts, Mo03 + Bi203
.and W03 + Bi203., are shown in fig. 5. The maximum selectivity corresponds
to maximum increase in the work function. The electronconductivity builds
up in a way similar to that obser/eVor the work function. It was shown by
X-ray analysis of these systems u' i that Bi20aMo03 , Bi 20a2Mo03 and
and Bi203 3Mo03 compounds were formed within the range of 0 .to 40 % of
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 :
CIA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
I I
svic
75
50
25
102
44
a
61
co at %&.
0 at p w,m0
(22
at;
af% Z31
7 STAT
Fig. 5. Variations in the electron work function (69, eV) and selectivity
percentage (S) as a function of the composition of bismuth molyb-
denum and tungsten catalysts
a) S1 Bi + Mo 2 - Bi +-W
aq71 - Bi + Mo 2 - Bi + W.
the Bi203 concentration in Mo03. Propene oxidation over V205 yields acro-
lein, saturated aldehydes and acids, Co2, and H20. The addition of bismuth
oxide results in a lower catalytic activity, but the relation between the rates
of formation of acrolein and saturated aldehydes from propene remains al-
most unchanged. With a bismuth-vanadium catalyst the rate of formation of
acids is lower (43% at Bi in V205). Vanadates of various structure were found
to form in the presence of Bi203.
The formation of new compounds was established by X-ray analysis of
V205, Mo03, and W03 systems in the presence of P205. The selectivity of
propene oxidation over vanadium, molybdenum, and tungsten phosphates in-
creased with decreasing specific catalytic activity. The activation energies
for CO2 formation were by 8 to 10 kcal/mole lower than for pure oxides. An
interesting feature of the V203 + P205 system is a change of the oxidation
route, due to suppression of reactions induced by breaking of the double bond
In a proiglie molecule. Similar observations were made by Kernos and Mol-
davskii "'for the oxidation of butenes to maleic anhydride over phosphorus-
vanadium catalysts.
The addition of various compounds to oxide catalysts is known to change
the selectivity of oxidation of unsaturated hydrocarbons. Modifying additions
increasing the work function and displaying a highed electronegativity than
that of Cu20 will increase the selectivity of propene oxidation, while an oppo-
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 :
CIA-RDP80-00247A004200070001-6 _ .
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 ?
8
3
site effect would be observed with additions decreasing the work function.
Catalysts with a higher co value will show greater coverage with positively
charged propene, as compared with Cu20, while increased coverage with
negatively charged oxygen will be characteristic of catalysts with a lower
value.
STAT
3. METALS
Platinum, palladium, copper and silver are the most widespread metal
catalysts for oxidation. Complete destruction of the hydrocarbon molecule
skeleton to CO2 and H20 will be observed in the gas-phase oxidation of hydro-
carbons over these catalysts. Silver is the only active catalyst inducing in-
complete oxidation of ethylene to ethyleng?oxide. Investigation of oxygen ad-
sorption and of oxygen isotope exchange 6) over these metals has shown that
chemisorption of oxygen yielded molecular and atomic oxygen ions, the ratio
of which is a. function of temperature and the surface condition. Vol and
4)
Shishakov 1 have shown by electron diffraction that the interaction be-
tween silver and oxygen at 100 to 2000 yields silver peroxide which is decom-
posed by ethylene. With platinum and palladium catalysts the mobility of
oxygen at these temperatures will be considerably lower. The interaction
between saturated hydrocarbons and an oxygen-covered platinum surface re-
sults in the formation of an intermediate complex readily decomposed by
oxygen of the gas phase, the reaction continuing in the gas phase at 50 to
70?. The oxidation of unsaturated hydrocarbons on silver was not found to
continue in the gas phase.
Variations in the electron work function, in the electroconductivity of
the compounds added, and the selectivity of ethylene oxidation to ethylene
oxide on modified silver are shown in fig. 6.
Fig. 6. Variations in the selectivity (A St%) of ethylene oxidation to
s ethylene oxide over silver as a function of variations in the
electron work function (Acp, eV).
- ?
.?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 _
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
9
I
Additions involving alkali and earth-alkali metals will decrease, while
metalloids will increase the electron work function. Elements displaying a
lower electroconductivity than that of silver will decrease cp and the selecti-
vity, while higher electronegativity results in increased and selectivity of
ethylene oxidation.
Variations in rp in the adsorption of various atoms on metals are consi-
dered as due to the formation of a dipole layer on the surface. Dipoles are
formed also by chemisorption of oxygen and hydrocarbons, and the inter-
action between these dipoles and those formed by additives seems to induce
changes in the surface coverage with oxygen and ethylene.
-4. LIQUID-PHASE OXIDATION OF HYDROCARBONS
The liquid-phase oxidation of hydrocarbons has been investigated less
extensively than that in the gas phase. It was shown by experiment that
transient metal oxides and certain metals may be used as catalysts for liquid-
phase oxidation. Oxidation of toluene by air over a mixed vanadium-tungsten
catalyst at 2000 and a pressure of 80 atm yields 71 mol % of benzoic acid and
about 10 mol % of benzoic aldehyde. With the same catalyst and under similar
conditions toluene will yield 7.2 mol % of acid and 1.5 mol % of aldehyde. In
a number of cases, depending on the hydrocarbon structure and oxidation con-
ditions, the part played by oxide and metal catalysts comes to initiation of
a, homogeneous chain reaction 16,17). The products yielded by catalytic re-
actions are similar in composition with those formed in liquid-phase reactions
involving homogeneous initiators. In other cases reactions on solid catalysts
proceed in a,more specific way 15). The main products of n-heptane oxidation
on vanadium and tungsten oxides are butyric and valeric acids, while mainly
acetic and propionic acids are formed with soluble catalysts. In the presence
of heterogeneous catalysts the yield in carbonyl compounds is several times
higher. In this case liquid-phase oxidation will proceed mainly on the surfaCe,
and not in the gas phase.
STAT
5. DISCUSSION OF THE RESULTS OBTAINED
It was shown by analysis of the results obtained that the mixed and modified
catalysts studied may be divided into two groups: solid inclusion and substi-
tution solutions, and chemical compounds (vanadates; molybdates, etc).
The introduction of minor amounts of molybdenum and chromium oxides to
vanadium pentoxide results in the formation of solid solutions containing V4+
ions.
The addition of minor amounts of molybdenum and chromium oxide to
solid vanadium pentoxide results in the. formation of solid solutions containing
V4+ ions. These catalysts display increased activity and selectivity with
respect to be,nzene oxidation. This is no so for butene oxidation to maleic
anhydride 13) or of propene to acrolein `); the addition of Mo03 to V205 will
decrease the activity and selectivity. A relation was found to exist between
the electron work function and the selectivity of both reactions.
The formation of chemical compounds from vanadium pentoxide and from
-_
Declassified in Part Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 - , .
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
10
STAT
chromium, molybdenum, and cobalt oxides resulsts in decreased activity
and selectivity of benzene oxidation, while molybdenum and bismuth tungstate
increase the selectivity of propene oxidation to acrolein. Vanadium phosphates
accelerate butene oxidation to maleic anhydride and slow down its formation
from benzene. A relation between the electronic properties of the catalyst
and the selectivity is observed for chemical compounds, same as with solid
solutions.
All systems of the oxidation catalysts studied contain transient metals,
and surface formation of various complexes with the lattice cations seems to
be possible in the adsorption of hydrocarbons. Hydrocarbon molecules involving
a double bond with the transient metal atoms may form 7r-complexes 16). The
ir-complexing capacity is a function of the electronic structures of cations and
of organic molecules. Possible is also a-complexing involving the breaking
of C-H bonds in the hydrocarbon.
At the catalyst surface the complexes seem to convert into ion radicals
forming peroxide ion radicals by reaction with oxygen 2).
The first group of catalysts may be called IT-active, and the second a-
active.
Similarly to homogeneous catalysts, reduction of transient metal ions
may occur in the solid lattice, following the scheme
RCH + Me) RCH(+) +
The superposition of d-orbitals with the formation of a generalized elec-
tron conduction band is not observed for metal oxides with open d-shells 16),
and displacement of current caff\iers proceeds by charge exchange between
? ions, as suggested by Verwey ). The redox potential value characteristic
of the electron or hole transfer from one ion to another is a measure for
determining the redox capacity of ions in solutions. When electronic levels
form no band, as is the case for the systems studied, the redox potentials
will be a function of the activity and selectivity of catalysts in hydrocarbon
oxidation. For these catalysts the work function (cp) will be related to the
distribution of electronic levels in the lattice of the solid. The acceptor
capacity of the surface, thus the coverage with electron donors, that is with
hydrocarbons, will increase with increasing cp. Additions to metals and to
oxide catalysts result in the formation of solid solutions of heterogeneous
systems. With metals (in contrast to oxides) the main part will be played
by oxygen activation at the surface resulting in charged molecular (oxide)
and atomic forms.
The reaction selectivity is a function not only of the chemical composi-
' tion and electronic properties of the catalyst, but also of the ratios of rates
for individual steps. Selectivity is a function either of the relation between
formation rates for various products, or of the destruction rate of the final
product. This question was widely discussed and it was shown that oxidation
of hydrocarbons proceeds by a parallel-consecutive mechanism. When the
time of a reaction product desorption is comparable with that of its formation
a subsequent surface conversion of this product becomes possible. Interesting
results were obtained for the liquid-phase oxidation of o-xylene on a vanadium
catalyst. Instead of the? expected phtaleic anhydride, o-toluene acid was formed.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
11
The reason for it probably is the shortage of time the products remain on the
surface, and desorption of these into the gas phase. A similar effect is ob-
served for gas-phase oxidation in the presence of water vapour. An examRlc
of it would be the increase in selectivity in Rropene oxidation to acrolein 42),
or in that of furfurole to maleic anhydride 2.5).
-
REFERENCES
1) LI.Ioffe, Z. I. Ezhkova and A. G. Lyubarskii, Kinetilca, i Kataliz 2 (1962) 194.
2) L.Ya.Margolis, Geterogennoye Katalyticheskoye Okislenyi Uglevodorodov (Gos-
toptekhizdat, 1962).
3) Z. I. Ezhkova, I. I. Ioffe, V. B. Kazanskii, A. V. Krylova, A. G. Lyubarskii and
L.Ya. Margolis, Kinetika I Kataliz 4 (1963) 187.
4) A.Magnalli and B. M.Oughton, Acta Chem. Scandinavica 5 (1951) 581.
: 5) R. Tarama, Sh. Teranishi and T. Jasui, J. Chem. Soc. Japan, Industr. Chem. 60
(1957) 1222.
6) I.N.Belyaev and N. P .Smolyaninov, Zhurn. Neorgan. Khirnii 7 (1962) 1126, 2521.
7) G.Gattow, Z. anorgan. Chem. 298 (1959) 64.
8) I.I.Ioffe and A. C. Lyubarskii, Kinetika i Kataliz 2 (1962) 261.
9) 0.V. Isaev and L.Ya.Margolis, Kinetika i Kataliz 1 (1960) 237.
10) V. B.KazansIdi, Z.I.Ezhkova, A.G. Lyubarskii, V. V. Voevodskii and I.I.Ioffe,.
Kinetika i Kataliz 2 (1961) 862.
11) E .Kh. Enikeev, L. Ya.Margolis and S. Z. Roginskii, Dokl. Akad. Nauk SSSR 130
(1960) 807.
12) L.N. Kutzeva and L. Ya.Margolis, Zhur. Obshch. Khim. 32 (1962) 102.
13) Yu.D.Kernos and B. L. Moldavskii, Kinetika i Kataliz 2(1962) 271.
14) Yu.U.Vol and N.A.Shisbakov, lzv. Alcad. Nauk SSSR,Otdel. Khim. Nauk, No. 4
(1962) 586. ?
15) I. I.Ioffe and N.V.Klimova, Kinetika i Kataliz, 4(1963) No. 5.
16).J.Burger, C. Meyer, G.Clement and J. C. Balaceanu, Compt. Rend. 252 (1961)
2235.-
17) E.A.Blumberg, M.G. Bulygin, L.Ya. Margolis and N.N. Emanuel, Dokl. Akad.
. ? Nauk SSSR 150 (1963) 1066.
18) Ya.K.Syrkin, Zhur. Strulct. IChim. 1 (1960) 189.
19) LI:Ioffe, Kinetika i Kataliz 2 (1962) 175.
20) F.Y. Morin. Semiconductors (ed. by M.B. Hanney, N.Y., 1961).
21) E.Y.W.Verwey,- Semiconducting Material (1951) p. 151.
22) A. G. Polkovnikova, A. N.Shatalova and L. L. Tzeitina, Neftekhimya 3 (1963) 246.
23) V.A.Slavinskaya, M.A.Shimanskaya, C.A. Giller and I.I.Ioffe, Kinetika i Ka-
taliz 2 (1961) 252.
STAT
_... ? _
-
Declassified in Part - Sanitized CopiApproved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
.. ?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
STAT
THE INFLUENCE OF SELECTIVE PROPERTIES
OF MOLECULAR SIEVES ON
CATALYTIC DEHYDRATION OF ETHANOL
Milo? RALEK and Oto GRUBNER
Institute of Physical Chemistry, Czechoslovak Academy of Sciences, Prague
Abstract: The catalytic dehydration of ethanol and diethylether has been studied.
Model catalysts used were molecular sieves. It has been found, that the rate of
transport of reaction components in the microporous structure of molecular
sieves substantially influences the selectivity of the reaction. Results obtained
confirm the simultaneous mechanism of the reaction.
1. INTRODUCTION
The catalytic dehydration of ethanol and diethylether, taking place on
alumina and aluminosilicate catalysts, is a complex reaction which includes
equilibria and chemical reactions between the individual components (ethanol,
diethylether, water, ethylene) 1). -
The conversion of ethanol to diethylether, in dependance on time of
contact, runs through a maximum. Conversion to ethylene increases with
increasing time of contact. Conversion of diethylether to ethanol shows a
similar maximum.
The conversion curves have been described by Brey and Krieger 2).
These authors have studied the dehydration of ethanol on aluminas, treated
In different ways, using kinetic equations derived under the assumption
that water and ethanol are sorbed strongly on the catalysts and that the
rate of reaction is determined by the rate of the chemical conversion of the
adsorbed intermediary complex to ethylene. They have made use of Whit-
more's conception 3) assuming that the intermediary complex, forming on
the catalyst surface, is a carbonium ion. The carbonium ion concentration
is proportional to the surface concentration of ethanol and its formation is
conditioned by interactions of active centers of the catalyst, where non-
saturated valence forces exist, with molecules of the substrate. The reac-
tion takes place simultaneously according to the scheme:
+ ? ?
C2H5vri (-v?15 -H+ ---1..,21-14
r fly +C2H50-
(C2H5)20 _
Balaceanu and Jungers 4). have studied the dehydration of ethanol and
ether at low temperatures on alumina. In agreement with opinions expressed
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04: STAT
C1A-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
2 ?
STAT
earlier in the literature 5), they have come to the conclusion, that this is a
consecutive reaction and they have proposed the following scheme:
2 C2H50H --(C2H5)20 + 1120 C2H4 + C2H5OH + H20
? Topshieva and Yun-Pin 6) have derived kinetic equations for the dehydra-
tion of ethanol and diethylether on alumina and aluminosilicate catalysts of
various composition. The equations apply to the consecutive scheme of the
reaction under the condition that sorption of water on the catalyst prevails
appreciably over the sorption of the other components. The rate constants
in the temperature range of 375 to 450?C are comparable and directly pro-
portional to the A1203-content of the catalysts. The activation energy for the
dehydration of ethanol and ether to ethylene is the same for all the catalysts
(14.5 kcal/mole). At a reaction temperature of 2500C the main product of
ethanol dehydration is diethylether. Ethylene is formed in slight amounts
(2%).
When later analysing the results, the authors have come to the con-
clusion that the reaction proceeds by a simultaneous mechanism, through
a surface compound of the type of aluminium alcoholate, which is formed
on the catalyst surface according to the scheme
C2H50H + 1(C2H5)20 + HO-Al
lli
c2H4 + OH-Al
Reactions I and II are reversible, the decomposition of the inter-
mediary complex (HI) is non-reversible. During a short contact time,
equilibrium between the reactions I and II is not obtained on the catalyst
surface. In ethanol dehydration a reaction is preferred in this case, which
supports the establishment of equilibrium, i.e., ether formation, ether
passing by desorption into a gaseous phase. When the time of contact is in-
creased, decomposition of the intermediary complex to ethylene takes
place. in a greater measure. The equilibrium concentration of ethylene is
renewed by the reaction of ether from the gaseous phase, resulting in de.-
creased ether production. A similar process takes place in the case of
diethylether dehydration, the dependance of ethanol production on the time
of contact passing through a maximum.
The kinetic relations and assumptions of the reaction mechanism de-
rived from these are based on measurements of catalytic activity in the
kinetic region, where conversions measured are independent of the particle
size of the catalysts.
? Boreskov 7) has found, that at particle size of the alumina of 3 mm,
-inner diffusion becomes important and the relatic amounts of products
change. Reactions in the region of inner diffusion take place at a lower rate.
With the same amount of reacting alcohol, ether production is roughly half
of that, which is produced when the reaction takes place in the kinetic
region. Qualitatively the results agree with theoretical conclusions on the
influence of inner diffusion on catalyst selectivity.
.,C 4.1,4 ltrnr?Ir lune +ha annliratinn nf the shame selectivity of
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 : ,
:3IA-RDP80-00247A004200070001-6 ?
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
3 STAT
molecular sieves to the confirmation of the proposed mechanism, and an
elucidation of the influence of pore size on catalytic selectivity.
The molecular sieves are crystalline aluminosilicates with defined
pore size, comparable with the size of the molecules of some of the react-
ing substances. They are therefore suitable for use as model catalysts.
-2. EXPERIMENTAL PART
The molecular sieves have been synthesised in the sodium form 8,9)
and their purity was checked by means of X-ray methods 10,11). The cal-
cium and potassium forms were prepared from the sodium form by means
of ion exchange. In the dehydrated state, the molar composition of the
catalysts was:
molecular sieve 3A
molecular sieve 4A
molecular sieve 5A
molecular sieve 10X
molecular sieve 13X
0.81 K20 . 0.18 Na20 . A1203 . 1.98 Si02
0.98 Na20 . A1203 . 1.98 Si02
0.72 CaO . 0.26 Na20 . A1203 . 1.98 Si02
0.87 CaO . 0.12 Na20 . A1203 . 2.6 Si02
1.0 Na20. A1203 . 2.6 Si02
Alumina, prepared by the hydrolysis of aluminium alcoholate, was
used as comparison catalyst. Its specific surface was 200 m2/g, the diam-
eter of the most frequent pores was 40 A (sample B I, having the proper-
ties described in ref. 12).
Microcrystalline catalyst samples were pelletised by pressure and
grains of 0.2 to 0.4 mm size were used for the experiments.
A registering microbalance was used to measure the rate of ethanol
and ether sorption on molecular sieves 5A and 10X at temperatures at
which no reaction occurs yet. Values of the apparent diffusion energy Ea
for ethanol and ether into these catalysts were calculated from these mea-
surements. The calculations were performed by means of the relation 13)
3 d(Qt/003)
R In -1Ea
a (1/T) d,it -
(Qt is the amount sorbed in the time tat the temperature VIC, Q?,0 corre-
sponds to the equilibrial sorption value at the same temperature). The de-
pendance of Qt on,it was linear up to Qt = 0.6 Q.. It has been found, that
the apparent activation energy of ethanol diffusion in molecular sieves 5A
and 10X is the same and is equal to 1.4 kcal/mol. The apparent activation
energy of ether diffusion in molecular sieve 10X is equal to 2.1 kcal/mol.
The apparent activation energy of ether diffusion in molecular sieve 5A
was substantially higher (6.4 kcal/mol).
The catalytic dehydration of ethanol has been studied in a circulating
flow apparatus with.differential reactor 14). Pure nitrogen was used as
carrier gas. Ethanol feed was controlled by the saturator temperature.
Reaction components have been analysed by means of gas chromatography.
? - 'Overall catalytic activity values, expressed as fraction of reacted al-
cohol amounts, in dependance on the reciprocal value of the feed, are shown
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6 -
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
4
moi C.,14.
moi(C,H,10 4
2
0
06
conversion
C28.08 04
02
0 2 4 6 80 2 4 8.)0'
g. soe/mor CAM
Fig. 1. Overall catalytic activity of molecular sieves 5A, 10X and ','-alumina,
and the value of the molar ratio of ethylene-diethylether, in depend-
ance on the reciprocal value of the feed.
STAT
in the lower part of fig. 1. Values of the molar ratio of ethylene and diethyl-
ether, formed by the reaction, are shown in the upper part of fig. I. The
overall activity of both types of molecular sieves is practically identical in
the temperature range of 240 to 2600C. On the small-pore molecular sieve
5A, ethylene is mainly formed, whereas ether is the main product obtained
on the molecular sieve 10X, which has wider pores. With decreasing feed
total ethanol conversion increases. The value of the molar ratio ethylene/
ether increases much more rapidly in the case of molecular sieve 5A than
in the case of molecular sieve 10X.
Alumina is a more active catalyst than both types of molecular sieves
described. Its activity at 240?C corresponds roughly to the activity of molec-
ular sieves at 260?C. Alumina produces mainly ether, similarly as sieve
1(1X. -
Molecular sieve 13X has, in the temperature range studied, approxi-
mately 40% of the activity of molecular sieve 10X, and forms mainly ether.
Molecular sieves 3A and 4A are non-active in the temperature range
studied. A measurable catalytic reaction takes place only at higher tempera-
tures (--3000C).
The dehydration of ethanol and diethylether on sieves 5A and 10X and
alumina has been studied furthermore by means of the microcatalytic pulse
technique ?15). In the dehydration of ethanol molecular sieve 5A formed
mainly ethylene: Under comparable conditions, molecular sieve 10X and
alumina formed mainly ether.
200 250 300 350
- ? T
. Fig. 2. Dehydration of ethanol and diethylether on molecular sieves
, 5A and 10X using the pulse microcatalytic technique, in
? dependance on temperature.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
::',IA-RDP80-00247A004200070001-6
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
5
STAT
The dehydration of ether took place easier on molecular sieve 10X and
on alumina than on molecular sieve 5A. Comparing the behaviour of both
types of molecular sieves on the basis of the amount of ethylene formed by
catalytic reaction, we obtain the dependance shown in fig. 2. The figure
shows, that molecular sieve 5A dehydrates diethylether relatively with
great difficulty.
3. DISCUSSION
From the results of this work follows:
The activity of the model catalysts, used for the dehydration of ethanol,
decreased in the series: alumina, molecular sieves 5A and 10X, molecular
sieve 13X. Molecular sieves 3A and 4A were non-active, due to the small
dimensions of their pores, which made it impossible to utilise the inner
pore surfaces.
On molecular sieve 5A, ethanol dehydration proceeded selectively to
ethylene in a temperature range, where alumina and molecular sieve 10X
gave mainly ether.
Diethylether has been dehydrated easily by alumina and by molecular
sieve 10X, with difficulty by molecular sieve 5A.
In catalytic action, alumina and molecular sieve 10X are more similar
to each other, although these two substances are chemically and crystallo-
graphically different, than both molecular sieves, which have similar chem-
ical compositions and character of the arrangement of the inner surface.
The cause for this difference in behaviour must be searched for in the
differing pore size, which influences the rate of mass transport into the
catalysts. This effect is most distinct in the case of ether diffusion into
molecular sieve 5A, whose activation energy is substantially higher than
for the other components, corresponding roughly to one half the value of
activation energy of the dehydration of ether to ethylene. The high activa-
tion energy of the diffusion of ether into molecular sieve 5A is the direct
cause of the fact, that the dehydration of ether in the pulse microcatalytic
arrangement takes place more slowly than the dehydration of ethanol,
because the relative frequency of ether molecules which have access to the
inner catalytic surface, is decreased. In the case of catalysts with wider
pores, the reaction is not slowed down by transport effects, and dehydra-
tion of ether is easy. Steric effects, influencing ether transport, may ex-
plain the increased selectivity of molecular sieve 5A in the dehydration
of ethanol to ethylene. Schemes of the simultaneous mechanisms of dehydra-
tion reactions are based on the establishment of equilibria between the
intermediary surface complex and ethanol or ether. In the case of the nar-
row-pore sieve 5A, transport of ether formed by ethanol, from catalyst
cavities into the gaseous phase is retarded by slow diffusion. The equilibrium
concentration of the ether-complex is therefore established easily. The in-
termediary complex decomposes to ethylene, and its equilibrium concentra-
tion is constantly amplified by ethanol, coming from the gaseous phase. Due
to these processes the catalytic reactions produce mainly ethylene.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
::',IA-RDP80-00247A004200070001-6 "-
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04:
CIA-RDP80-00247A004200070001-6
5
STAT
Among the two dehydration mechanisms proposed, the carbonium
mechanism seems to be less probable, because no parallel has been found
between the acidity of catalysts - which increases in the series 5A, 10X,
13X 1.6) - and their catalytic activity. Our results show, that there exists a
relatfontetween the content of A1203 and the overall catalytic activity, which
has been found in the work of Topshieva and Boreskoy.
The explanation of the observed catalytic effects also agrees with the
work of Brey and Krieger, from which follows that aluminas, structurally
: modified by the influence of steam at elevated temperatures (leading-to
- decrease of the specific surface values and thus also to the loss of the very
small pores), have produced under otherwise equal conditions a larger
amount of ether than the initial catalysts.
It has been observed in the present work, that on the molecular sieves
appreciable ether" sorption takes place even at temperatures close tO the
reaction temperatures. A similar observation has also been made by
Balaceanu and Jungers in the course of determining the adsorption coef-
ficients of ethanol, ether and water on alumina. Kinetic equations given
in the literature do not respect sufficiently this fact, and therefore it cannot
be assumed that they describe the kinetics of these dehydration reactions
? fully.
. It seems probable, that selective catalytic effects of molecular sieves,
caused by transport phenomena, will be obserVed in a number of other
reactions also.
REFERENCES1
1) 1VCE.Winfield. in: Catalysis (Reinhold. New York. 196b) Vol. 7, p. 93.
2) W.S.Brby and K.A.Krieger. J. Am. Chem. Soc. 71 (1949) 3657. _
3) F.C.Whitmore, J. Am. Chem. Soc. 54 (1932) 3274.
4)- J.C. Balaceanu and J.C. Jungers , Bull. Soc. Chim. Belg. 60 (1951) 476.
5) Alvarado, J. Am. Chem. Soc. 50 (1928) 790.
6) K.V.Topshieva and K.Yuit-Pin, Zhur. Fiz. Khim. 29 (1955) 1678, 1854, 2976.
7) G.K.Boreskov, V.A.Dzisko and M. C. Borisova, Zhur. Fiz. Ithim. 28 (1954) 1055.
8) O. Grubner. M.RAlek and P. Jiru, Chem. prim. 11 (1961) 521.
- 9) P.Jiru, O. Grubner and M.Ralek, Chem. prim. 11 (1961) 20.
? 10) R.M.Barrer, J.W. Baynham, F.W.Bultitude and W. M.Meier. J. Chem. Soc.
(1959) 195.
? 11) T. B. Reed and D.W.Breck, J. Am. Chem. Soc. 78 (1956) 5972.
12) V.13ostteek. R.Polak,.E.Ku6era and V.DaneA, Coll. Czech. Chem. Commun. 27
(1962) 2575.
:13) R.M. Barrer, Trans. Faraday Soc. 45 (1949) 358.
' 14) M. I. Temkin, C.-L.Kiperma.n and Lullanova, Doklady Akad. Nauk SSSR 74
, (1950) 763.
- 15) R. C Stein, J. J. Feeman. G. P . Thomson, J. F. Shultz, L. J. E . Hofer and R. B.
? Anderson, Ind. Eng. Chem. 52 (1960) 671.
?--'16) C.J.Norton, Chem. Ind. (1962) 258.
Declassified in Part - Sanitized Copy Approved for Release 2014/03/04 :
CIA-RDP80-00247A004200070001-6
? .