Detection of Steady State Multiplicity during Dimethyl Ether Conversion Catalyzed by ZnO/γ-Al2O3 Composite: Effect of Coke and Hydrogen Peroxide

Article abstract of DOI:10.1134/S0965544120070087

Abstract: The heterogeneous catalytic conversion of dimethyl ether (DME) to 1,3-butadiene and butylenes in the presence of a ZnO/γ-Al₂O₃ catalyst in a wide range of temperatures, feed space velocities, and reactant concentrations has been studied. It has been found that temperature regions of 380–420 and 440–480°C, which are conventionally designated as low-temperature and high-temperature regions, differ in the laws governing the occurrence of the process associated with the specificity of coking of the catalyst surface and a redistribution of the surface concentration of Br?nsted and Lewis acid sites. A hypothesis of a change of the catalytic process mechanism upon the transition of the reaction into the high-temperature region has been put forward and confirmed by the occurrence of a hysteresis detected in this temperature range. The effect of hydrogen peroxide on the hysteresis parameters and the steady state stability during DME conversion has been shown.

Full text of DOI:10.1134/S0965544120070087

ISSN 0965-5441, Petroleum Chemistry, 2020, Vol. 60, No. 7, pp. 773–784. © Pleiades Publishing, Ltd., 2020.
Russian Text © The Author(s), 2020, published in Neftekhimiya, 2020, Vol. 60, No. 4, pp. 488–500.
Detection of Steady State Multiplicity during Dimethyl Ether
Conversion Catalyzed by ZnO/γ-Al2O3 Composite:
Effect of Coke and Hydrogen Peroxide
A. L. Maksimova, V. F. Tret’yakova, *, and R. M. Talyshinskiia
aTopchiev Institute of Petrochemical Synthesis, Russian Academy of Sciences, Moscow, 119991 Russia
*e-mail: tretjakov@ips.ac.ru
Received October 18, 2019; revised November 29, 2019; accepted March 12, 2020
Abstract—The heterogeneous catalytic conversion of dimethyl ether (DME) to 1,3-butadiene and butylenes
in the presence of a ZnO/γ-Al2O3 catalyst in a wide range of temperatures, feed space velocities, and reactant
concentrations has been studied. It has been found that temperature regions of 380–420 and 440–480°C,
which are conventionally designated as low-temperature and high-temperature regions, differ in the laws gov-
erning the occurrence of the process associated with the specificity of coking of the catalyst surface and a
redistribution of the surface concentration of Brønsted and Lewis acid sites. A hypothesis of a change of the
catalytic process mechanism upon the transition of the reaction into the high-temperature region has been
put forward and confirmed by the occurrence of a hysteresis detected in this temperature range. The effect of
hydrogen peroxide on the hysteresis parameters and the steady state stability during DME conversion has
been shown.
Keywords: steady state multiplicity, catalysis, DME, 1,3-butadiene, butylenes, carbon deposits, hysteresis,
coke, thermal desorption, acidity
DOI: 10.1134/S0965544120070087
INTRODUCTION
nent. As a consequence, the specific surface area of
the catalyst decreases and the pore structure changes;
these factors, in turn, require a frequent refreshment
of the catalytic charge, which is economically unrea-
sonable [1–3].
In this aspect, a fairly urgent task is the search for
fundamentally new methods to extend the time
between regenerations of deactivating catalysts that
would not lead to undesirable and irreversible changes
in their characteristics. These methods include, for
example, the catalyst system self-regeneration method
based on the introduction of compounds (e.g., hydro-
gen peroxide) into the reaction medium composition
that inhibit the formation of condensation products
and coke on a heterogeneous surface and stabilize the
steady state during reaction [4].
An acute problem of catalytic chemistry is the for-
mation of condensation products (coke) on the sur-
face of heterogeneous catalysts on stream; these prod-
ucts block the active sites of the catalysts and cause a
loss of activity, selectivity, and instability.
The specificity of the coking processes that occur
on a heterogeneous surface of solid catalysts frequently
leads to the occurrence of a steady state multiplicity
(SSM) in the reaction systems and difficulties in con-
trolling the reaction process parameters in an opti-
mum mode.
The key tasks in eliminating these problems to
restore the main performance characteristics of het-
erogeneous catalysts are the stabilization or restoration
of their chemical and phase composition and struc-
tural and textural characteristics.
However, it should be stated that the currently
available methods to regenerate the surface of spent
catalysts do not fully meet all requirements of modern
technologies for the catalytic production of chemicals.
Currently, owing to the shortage in the production
of synthetic rubber monomers, the use of carbon-con-
taining feedstocks alternative to oil, in particular, oxy-
genates, such as ethanol, methanol, and dimethyl
ether (DME), is significantly expanding [5–7].
Earlier [8–10], it has been first shown that one of
the important monomers of heavy organic synthesis—
1,3-butadiene—can be actually produced from DME
in the presence of a modified ZnO/γ-Al2O3 catalyst via
Thus, a commonly used method to regenerate
spent catalysts by a periodic thermal oxidative treat-
ment, i.e., coke burning-off in an air stream (in
the presence or absence of steam), frequently
leads to sintering of the catalytically active compo- the intermediate conversion of DME to ethanol,
773

774
MAKSIMOV et al.
composition of the synthesized catalysts was analyzed
chromatographically as described in [8].
11
12
1
Dimethyl ether conversion rate was calculated by
the formula
2
5
WDΣME
=
Wo,
α
100
4
(1)
6
3
where α is the DME conversion (%) and Wo is the
DME feed rate (mol/(L h)).
The tests were repeated to determine reproducibil-
ity; the data were averaged. The standard deviations
from the average rate values for the system compo-
nents did not exceed Δ = 7%. Reaction rate was nor-
malized to 1 L of catalyst per hour; it is shown in terms
of mol/(L h).
7
Gaseous products
For analysis
9
8
For analysis
10
The catalyst was synthesized by the diffusion
impregnation of γ-alumina with zinc and aluminum
nitrate solutions at 200–500°C and subsequent drying
and calcining of the precursors under conditions
described in [9, 10].
Aluminum nitrate taken in an amount to provide
10 wt % in the final formulation was introduced into
the system to form a strong bond between the zinc and
aluminum oxides.
Liquid products
Fig. 1. Diagram of a flow system for testing the catalyst
activity: (1, 2) cylinders containing DME and nitrogen,
(3) dosing pump for feeding liquid reactants, (4) oven,
(5) catalyst bed, (6) reactor, (7) refrigerator, (8) condenser,
(9, 10) Kristallyuks-4000M chromatographs to analyze gas
and liquid phases, (11) load thermocouple in the catalyst
bed, and (12) control thermocouple of the oven.
Prior to using, γ-alumina was calcined at 600°C for
6 h and then placed in a desiccator, where it was cooled
to room temperature. The support was impregnated
with active component solutions at 80°C for 3 h. The
amount of the active component solution was calcu-
lated by the incipient wetness impregnation method
with the addition of a required amount of distilled
water.
The impregnated wet support was emulsified at
20°C for 12 h; after that, the sample was heated to
120°C and dried under stirring for 2 h to prevent the
precursor particles from coalescence. Next, after the
addition of an aqueous solution of tartaric acid at a pH
of 2.5 by the incipient wetness impregnation method,
the catalyst was further dried in a desiccator for 10 h.
The synthesized sample was placed in a muffle fur-
nace, where it was calcined at 200 and 300°C for 2 h at
each temperature; after that, it was cooled to room
temperature in a desiccator. Ammonia water was
added by the incipient wetness impregnation method.
After stripping nitrogen oxides in a muffle furnace at
350°C, the catalyst precursor in a required amount
was placed into the reactor, where the sample was cal-
cined and simultaneously activated at 400 and 550°C
in alternating streams of nitrogen, air, and hydrogen
(2 L/h per 5 mL of sample) for 3 h. The catalyst for-
mulation was calculated in terms of oxide content:
22.5 wt % ZnO and 77.5 wt % γ-Al2O3.
which is further converted to butadiene in accordance
with the known mechanism discovered by Lebedev
[11]. A kinetic model to satisfactorily describe the
revealed relationships in the considered temperature
range was proposed in [10].
However, it should be noted that, unlike the above
conversion of ethyl alcohol to butadiene, DME con-
versions are accompanied by a relatively high degree of
coking of the catalyst surface, which leads to a rapid
loss of initial activity, the formation of a variable-com-
position catalyzate, and a significant decrease in the
time between regenerations.
This study is focused on the relationship between
the characteristics of the proposed ZnO/γ-Al2O3 cata-
lyst for DME conversion to divinyl and the coking of
the catalyst surface, the effect of hydrogen peroxide
additives to the feed stream on the regeneration of the
active sites of the catalyst to remove carbon deposits,
and the establishment of a respective steady state
during reaction.
EXPERIMENTAL
The effect of DME conversion mode parameters
was studied in a quartz reactor with a catalyst load of
10 mL on the system shown in Fig. 1 in a temperature
range of 380–480°C at a DME gas hourly space veloc-
Special tests on variation in the linear velocity of
ity (GHSV) of 100–190 h−1 with dilution of the feed- the feed stream [8] in a range of 5–10 cm/s and the
stock with nitrogen in a molar ratio of 2.0 : 1 to 11.5 : 1. fractional composition of the catalyst granules in a
The catalyst-bed temperature gradient at a load of 5– range of (0.2–0.3)–(0.8–1.0) mm showed that exter-
10 mL was 5°C. The qualitative and quantitative nal and internal diffusion does not have any effect
PETROLEUM CHEMISTRY
Vol. 60
No. 7
2020

DETECTION OF STEADY STATE MULTIPLICITY
775
during DME conversion in the studied range of varia- couple (PP-1) with an accuracy of 2°C in a tempera-
tion in the reaction process parameters.
ture range of 22–900°C.
X-ray diffraction patterns were recorded on a
DRON-3 automated X-ray diffractometer with a
graphite monochromator. Measurements were con-
ducted using CuKα radiation in the step-by-step scan-
ning mode in increments of 2θ = 0.1°. The acquisition
time was 3 s per point. The diffraction patterns were
processed using the PHAN and PHAN (%) software
programs for qualitative and quantitative X-ray dif-
fraction analyses.
The textural characteristics of the catalyst samples
were studied on a unified Sorbi-MS system by the low-
temperature nitrogen adsorption method and calcu-
lated by the Brunauer–Emmett–Teller (BET)
method. Total sorption pore volume V (total volume
Ʃ
of micro- and mesopores) was determined from the
amount of nitrogen adsorbed at a relative pressure
close to unity on the assumption that all the pores were
filled with nitrogen in the condensed state due to cap-
illary condensation. In this case, the adsorbed nitro-
Infrared spectra were recorded at an adsorption
temperature on a Shimadzu 8300 Fourier transform
IR spectrometer at a resolution of 4 cm–1 and a spec-
trum accumulation factor of 50. The freshly prepared
samples were precompressed into pellets with a den-
sity of 7–17 mg/cm2, calcined in an IR cell at 450°C in
a vacuum (0.013–0.001 MPa) for 1 h, and then cooled
to 110°C with liquid nitrogen. Acid sites were identi-
fied using ammonia and carbon monoxide as probe
molecules.
gen volume (V ) was converted into the liquid nitrogen
Ʃ
volume (Vliq) using the following equation:
PA VΣVM
RT
(2)
Vliq
=
,
where РА and T are the ambient pressure and tempera-
ture, respectively; VM is the molar volume of liquid
nitrogen (VM = 34.7 cm3/g); and R is the gas constant.
The acidic properties of the catalyst surface were
determined by the temperature programmed desorp-
tion (TPD) of ammonia and infrared (IR) spectral
analysis of adsorbed ammonia and carbon monoxide
molecules as probe molecules.
The total surface acidity of the samples was deter-
mined under the assumption of single-site ammonia
adsorption from the number of chemisorbed mole-
cules, the desorption of which was almost completed
upon an increase in the temperature in the column
with the catalyst to 600°C.
The total number of acid sites NΣ (unit/m2) was
calculated from the values of the area under the
desorption curves by the formula
RESULTS AND DISCUSSION
Table 1 shows characteristics of the specific surface
area and porosity of the ZnO/γ-Al2O3 catalyst samples
synthesized for testing in DME conversion. It is evi-
dent that their final textural characteristics are deter-
mined by the parameters of the selected support.
However, the introduction of zinc oxide into the
matrix of the support leads to a certain decrease in the
specific surface area and total pore volume. An
increase in the zinc oxide content in the catalyst sam-
ples leads to a redistribution of pore volume with
respect to size, namely, a decrease in the number of
micro- and mesopores and an increase in the number
of macropores.
According to the diffraction patterns shown in
Fig. 2, the observed change in the textural characteris-
tics of the ZnO/γ-Al2O3 catalysts with an increase in
the zinc oxide content is apparently attributed to the
formation of a zinc aluminate phase and the increase
in the relative content of this phase, which affects the
degree of development of the active surface of the
samples.
6.02 ×1023
T
maxi
(3)
NΣ =
,
22400 × SspG
where 6.02 × 1023 is the Avogadro number; Ssp is the
specific surface area of catalyst samples, m2/g; G is the
weighed portion of the catalyst, g; and
is the
T
maxi
total peak area on the thermal desorption curve, mm2,
which is proportional to the desorbed volume of gas-
eous ammonia V, mL. The number of moles of
desorbed ammonia was calculated as V/22400, where
22400 is the volume of one mole of NH3.
The initial data used to calculate the total and par-
tial DME conversion rates were a sample of results of
the chromatographic analysis of the component com-
position of the reaction products at a reaction tem-
perature varied in a range of 380–480°C, a DME
GHSV of 100–190 h−1, and a nitrogen : feedstock
molar ratio of (1–10) : 1 (Table 2).
The differential thermal analysis (DTA) of the sup-
port and catalyst samples (fresh and coked) and their
precursors was conducted on a Q-1500 D derivato-
graph (F. Paulik, J. Paulik, L. Erdey; MOM, Hun-
gary); the linear heating rate was 7.5–10.0°C/min.
According to the above data and the results of
Thermogravimetric analysis (TGA) was conducted studying the kinetic laws of the reaction described in
in an air atmosphere or a nitrogen atmosphere. The [10], the most probable sequence of DME conversion
weighed portions of the samples were 120–250 mg to the target reaction product—1,3-butadiene—that
(weighing error of 0.4 mg). Temperature was mea- occurs in the high temperature region of 420–480°С is
sured using a platinum/platinum–rhodium thermo- the following:
PETROLEUM CHEMISTRY Vol. 60 No. 7 2020

776
MAKSIMOV et al.
Table 1. Dependence of textural characteristics of the γ-Al2O3 support and the ZnO/γ-Al2O3 catalysts on heat treatment
conditions
Heat treatment conditions Т,
Pore volume distribution by pore size, nm
Ssp, Vpore
,
ZnO/Al2O3 catalyst
sample
°С/time, min
m2/g cm3/g
10–102
102–103
thermolysis
calcining
5–10
γ-Al2O3*
–
–
188 0.58
178 0.55
168 0.52
165 0.50
153 0.47
0.23
0.21
0.18
0.16
0.13
0.27
0.24
0.22
0.21
0.15
0.08
0.10
0.12
0.13
0.19
320/120
320/150
320/180
320/150
430/240
450/240
450/300
450/300
ZnO (5.0%)
ZnO (10%)
ZnO (15%)
ZnO (20%)
* The support before the deposition of the active mass.
CH3CHO
(4)
CH3OCH3
C2H5OH
CH2 CH CH CH2.
+
CH2 CH2
In the low-temperature reaction region, the temperature regions of the process, which are conven-
sequence of formation of divinyl is apparently imple- tionally designated as low-temperature (380–420°C)
mented in accordance with the Lebedev mechanism and high-temperature regions (440–480°C). In addi-
[11]:
tion, upon transition from the boundary point of the
low-temperature region (420°C) to the high-tempera-
ture region (480°C), the relative decrease in the total
DME conversion rate and the reaction product yield is
more pronounced.
2CH3OCH3 → 2C2H5OH → 2CH3CHO
→ CH3CHOH−CH3CHO
(5)
→ CH3CH=CHCHO → CH2=CH−CH=CH2.
It is noteworthy that at the maximum reaction tem-
perature in the studied range (480°C), the introduc-
tion of hydrogen peroxide in an amount of 1.5 wt %
Comparison of the activity of the freshly prepared
catalyst samples and the samples after use on-stream
for 7 h and more (Table 3) revealed a change in the
DME conversion rate and routes, which was not with respect to DME into the feed stream contributes
observed in the early period of manifestation of activ- to the restoration of the catalyst activity; therefore, the
ity. It was found that the most significant differences rates of the total and partial DME conversions almost
in these behaviors are observed with a change of the approach the values listed in Table 2; in this case, a
ZnO
ZnO
ZnAl2O4
2500
ZnO
ZnO
Al2O3
Al2O3
ZnAl2O4
Al2O3
ZnAl2O4
2000
1500
1000
500
Al2O3
2
ZnO
ZnAl2O4
Al2O3
ZnO
ZnO
Al2O3
Al2O3
ZnAl2O4
Al2O3
ZnAl2O4
1
ZnO
Al2O3
20
30
40
50
60
70
80
10
2θ, deg
Fig. 2. X-ray diffraction patterns of ZnO/Al O catalyst samples. Zinc content in the samples: (1) 10 and (2) 20 wt % in terms of
2
3
oxides.
PETROLEUM CHEMISTRY Vol. 60 No. 7 2020

DETECTION OF STEADY STATE MULTIPLICITY
777
PETROLEUM CHEMISTRY Vol. 60 No. 7 2020

778
MAKSIMOV et al.
The processing of the experimental data revealed
mol
L h
Σ
WDME
,
the presence of a hysteresis in the WDME/TPEAK coor-
dinates (Fig. 3); the hysteresis is apparently due to a
change of the mechanism of surface DME conversion,
which is determined by a change in the pattern of the
active surface of the catalyst.
4.0
1
2
3.5
3.0
2.5
2.0
3
During the tests, the total DME conversion rate in
the presence of a freshly prepared catalyst at 380°C
was initially determined (first point in curve 1). After
1 h, the rate was measured again (second point in
curve 1). Upon achieving a reaction temperature of
440°C, at the fourth point in curve 1 (transition of the
reaction to the high-temperature region), the tem-
perature dependence of the DME conversion rate
abruptly deviates from the nearly exponential behavior
and the further increase in the rate decreases. It is
obvious that, within the final time of the reaction run
(sixth point in curve 1), coke accumulates in a certain
amount providing the blocking of the active surface
sites of the catalyst; it is this factor that causes the
decrease in the catalyst activity. It is characteristic
that, with a stepwise decrease in the reaction tempera-
ture (curve 2) and reversion to the points identified in
temperature curve 1, a significant discrepancy in the
DME conversion rate is observed; that is, an S-shaped
temperature hysteresis takes place.
550
550
600
650
700
Incline, deg
500
Fig. 3. Temperature dependence of the total DME conver-
sion rate: (1) the original ZnO/γ-Al O catalyst sample,
2
3
(2) the catalyst sample after the completion of the reaction
run, and (3) DME conversion after the completion of the
reaction run with the introduction of 1.5 wt % of hydrogen
−1
peroxide. Reaction conditions: GHSV = 190 h ; time of
DME conversion to the next measurement of the DME
conversion rate, 1 h.
It was found that the introduction of H2O2 in an
amount of 1.5 wt % into the reaction mixture and the
subsequent stepwise decrease in temperature leads to a
partial restoration of the catalyst activity (curve 3). In
this case, the difference in the activity of the freshly
prepared and coked samples becomes less significant;
that is, in this case, hydrogen peroxide exhibits
the regenerating function as an oxidizer of carbon
deposits.
small amount of carbon dioxide is identified in the
gaseous portion of the catalyzate. In addition, upon
transition from the boundary point of the low-tem-
perature region (420°C) to the high-temperature
region (480°C), the relative decrease in the total DME
conversion rate and the reaction product yield is more
pronounced.
A direct evidence of the formation of carbon
deposits during DME conversion of in the presence of
ZnO/Al2O3 catalysts was obtained by DTA of the cat-
alyst samples after a 10-h continuous run of the system
(Fig. 4).
The results show that, during the heating of the
samples, their weight gradually decreases; this
decrease is accompanied by several exothermic effects
associated with the burnout of the surface carbon
deposits. The occurrence of two maxima in the DTA
curves in a temperature range of 450–500 and 500–
550°C (sample 1) and 350–400 and 400–450°C (sam-
ple 2) indicates the presence of two types of deposits
that differ in the strength of the bond to the surface of
the catalyst samples.
It is noteworthy that at the maximum reaction tem-
perature in the studied range (480°C), the introduc-
tion of hydrogen peroxide in an amount of 1.5 wt %
with respect to DME into the feed stream contributes
to the restoration of the catalyst activity; therefore, the
rates of the total and partial DME conversions almost
approach the values listed in Table 2; in this case, a
small amount of carbon dioxide is identified in the
gaseous portion of the catalyzate.
In the high-temperature reaction mode, the transi-
tion of DME conversions into the diffusion region or
the occurrence of heterogeneous–homogeneous post-
catalytic volumetric conversions were eliminated after
special tests; therefore, the most probable cause of the
change in the catalytic activity is the specificity of cok-
ing of the catalyst surface.
According to DTA of the catalyst samples, after a
10-h run of the system with the addition of hydrogen
peroxide to the feed stream, the exothermic peaks cor-
responding to the two types of carbon deposits
decreased; this finding confirms that H2O2 exhibits a
To confirm the above assumption on the cause of
the change in the mechanism of DME conversions
upon transition to the high-temperature region, com-
parative tests of the activity of the catalysts were con-
ducted after a certain time of contacting with the reac- coke formation regenerating function during DME
tion medium.
conversion.
PETROLEUM CHEMISTRY Vol. 60 No. 7 2020

DETECTION OF STEADY STATE MULTIPLICITY
779
PETROLEUM CHEMISTRY Vol. 60 No. 7 2020

780
MAKSIMOV et al.
800
700
600
500
400
300
200
100
0
Т
TGA
1
2
DTA
1
2
20
30
40
50
60
70
10
Time, min
Fig. 4. Thermogravimetric analysis and DTA curves of ZnO/γ-Al О catalysts after 10-h runs of the system (1) in the absence of
2
3
H O additives and (2) in the case of introduction of 1.5 wt % hydrogen peroxide into the reaction zone.
2
2
0.10
0.08
0.06
0.04
0.02
Tmax1
Tmax2
Tmax3
3
2
1
0
100
200
300
T, °C
400
500
600
Fig. 5. Ammonia TPD spectra of the surface of (1) the γ-Al О support and ZnO/Al О catalyst samples with a ZnO content in
2
3
2 3
the support matrix of (2) 10 and (3) 20 wt %.
The described results suggest that the formation of
The observed maxima correspond to the desorp-
tion of ammonia from surface acid sites, which are
conventionally classified as weak, medium-strength,
and strongly acidic sites.
carbon deposits on the catalyst surface involves at least
two types of active—presumably acid—sites localized
in the γ-Al2O3 support matrix.
The above data suggest that the introduction of
zinc oxide into the alumina support matrix leads to a
quantitative redistribution of acid sites.
Thus, an increase in the content of the active mass
(ZnO) leads to the blocking of acid sites of all three
types and a decrease in their number; these effects are
also evident as a shift of the maxima of the ammonia
The TPD spectra of ammonia adsorbed on the sur-
face of the γ-Al2O3 support of the active mass of poten-
tial ZnO/γ-Al2O3 catalysts (Fig. 5, Table 4) exhibit
three desorption maxima that change their position on
the temperature coordinate depending on the experi-
mental conditions, °C:
(145–235),
(240–
T
T
max1
max2
290), and
(500–550).
desorption peaks (
) to the low temperature region
T
T
max3
maxi
PETROLEUM CHEMISTRY
Vol. 60
No. 7
2020

DETECTION OF STEADY STATE MULTIPLICITY
781
Table 4. Comparison of the surface acidity characteristics of the γ-Al2O3 support and ZnO/γ-Al2O3 catalysts calculated from
ammonia TPD spectra: initial temperature of NH3 adsorption, 50–65°C; programmed heating rate, β = 10–20 °C/min; and
NH3 dosage, 0.1–0.5 mmol
Support and
ZnO/Al2O3 catalyst
samples*
Temperature
Acidity corresponding to
T
maxi
T
maxi
× 1019 unit/m2
i
N
N1
N2
N3
T
max1
T
max2
T
max3
1.20 × 1019 6.72 × 1018 9.58 × 1018
1.26 × 1019 7.08 × 1018 3.52 × 1018
1.25 × 1019 6.25 × 1018 2.75 × 1018
1.12 × 1019 5.75 × 1018 1.25 × 1018
1.28 × 1019 6.65 × 1018 3.25 × 1018
γ-Al2O3
145–235 240–290 500–550
140–225 233–285 490–545
133–210 230–280 485–540
135–220 225–275 470–520
138–220 235–280 492–547
2.83
2.32
2.15
1.82
2.27
ZnO, 10%
ZnO, 20%
ZnO, 10%**
ZnO, 10%***
* The support before the deposition of the active mass.
** The catalyst after a 10-h run of the system.
*** The catalyst after a 10-h run of the system with the introduction of Н О into the feed stream.
2
2
(see Fig. 5). In this case, the number of strong acid
sites undergoes the most significant decrease.
It is evident from the spectrum parameters of the
support that before the deposition of the active mass of
zinc oxide (1), the OH groups peaking at the highest
frequencies (νОН = 3775 cm−1), which are apparently
bound to tetrahedral aluminum ions, and the OH
groups (νОН = 3725 cm−1) bound to octahedral alumi-
num ions can be classified as terminal groups.
A similar acid site redistribution was observed
during the thermal desorption of ammonia from the
coked surface of ZnO/Al2O3 catalysts that were in
contact with the reaction medium after a 10-h run of
the system.
The other OH groups giving bands in the low-fre-
quency region of the spectrum (νОН = 3670–
To determine the Brønsted acidity of the original
γ-Al2О3 support and ZnO/γ-Al2О3 catalyst samples,
the absorption spectra of OH groups that occur in the
region of 3400–3800 cm–1 (bridging and terminal OH
groups) were analyzed (Fig. 6).
3570 cm−1) are bridging groups localized between dif-
ferently coordinated aluminum ions.
An increase in the amount of the deposited ZnO
active mass (spectra 2, 3) leads to a shift of the absorp-
130
120
110
3695
3730
4
100
90
3
2
80
70
60
1
3500
3600
3700
3800
3900
3400
ν, сm–1
Fig. 6. Fragments of IR spectra of (1) the γ-Al О support and ZnO/Al О catalyst samples in a wave number region of 3900–
2
3
2 3
3
−1
3400 cm (absorption by OH groups). Zinc oxide content in the γ-Al О matrix: (2) 10 and (3) 20 wt %; (4) after feeding ammo-
2
nia into the cell.
PETROLEUM CHEMISTRY Vol. 60 No. 7 2020

782
MAKSIMOV et al.
are two temperature regions of the reaction, namely,
60
40
380–420 and 440–480°C. In each of these regions, a
certain sequence of formation of the target product—
1,3-butadiene—is dominant in accordance with the
above schemes (3) and (4); that is, depending on reac-
tion conditions, the presence of an SSM is observed;
in the case under consideration, two steady states take
place.
The presence of two steady states during the DME
conversion reaction suggests that the reaction mecha-
nism in the above temperature regions is also multi-
channel; it is not improbable that the slow stages
change with temperature variation. In this case, it is
assumed that the active Brønsted and Lewis acid sites
are mostly involved in the reaction in the low-tem-
perature and high-temperature regions, respectively.
3
2
20
0
1
2100
2150
2200
2250
2050
ν, сm–1
Fig. 7. Infrared spectra of carbon monoxide adsorbed on
the surface of (1) the γ-Al О support and ZnO/Al О cat-
2
3
2 3
A similar mechanism most probably takes place
despite the following fact revealed by X-ray diffraction
analysis: in the entire studied temperature range, the
phase composition of the samples is constant regard-
less of the degree of coking of the catalyst surface after
reaction.
alysts with a ZnO content in the sample of (2) 10 and
(3) 20 wt %.
tion bands of OH groups to the low-frequency region
and a simultaneous increase in their relative intensity,
which is attributed to the partial neutralization of
stronger Brønsted acid sites on the surface of the sam-
ples.
CLOSING REMARKS
Along with studying the kinetic mechanism for the
implementation of processes, it is necessary to thor-
oughly study their dynamics associated with both
macroscopic factors and critical phenomena in catal-
ysis that are caused by SSM, surface coking, self-
vibration modes, etc. This study showed that macroki-
netic factors do not significantly affect the DME con-
version process. The hysteresis detected in the tem-
perature range examined indicates the presence of an
SSM caused by the presence of two types of sites
(Brønsted and Lewis).
The results make it possible to accelerate the pre-
diction of the catalyst design and effectively imple-
ment the process on an industrial scale. Note that
most reports in the field of heterogeneous catalysis do
not provide data on nonlinear effects and SSM; the
lack of these data impedes the development of science
and production in the chemical industry.
The evolution of heterogeneous catalytic conver-
sions in porous media is associated with multiphase
processes [12–15], which are evident in the dynamics
of reactions apparently attributed to the interplays of
the catalyst and the initiator, which forms the catalyst
system under the conditions of catalyst surface coking.
A deeper understanding of the catalytic process mech-
anism makes it possible to predict the catalyst genesis
determined by the chemical composition and synthe-
sis algorithm of the catalyst.
After feeding a dosed amount of ammonia into the
measuring cell, the IR spectra of the sample contain-
ing 20 wt % ZnO in the support matrix show a
decrease in the intensity of the absorption bands of
free hydroxyl groups at 3695 and 3730 cm−1 and the
simultaneous appearance of bands at 3560 and
3410 cm−1, which is characteristic of the formation of
a hydrogen bond between NH3 molecules and proton-
donating Brønsted acid sites on the catalyst surface.
The absorption spectra of carbon monoxide
adsorbed on the surface of the γ-Al2О3 support and
ZnO/Al2О3 catalysts precalcined at 500–550°C indi-
cate the presence of two types of Lewis acid sites that
differ in strength; the sites are characterized by CO
absorption bands in the regions of 2150–2170 and
2180–2210 cm−1, respectively (Fig. 7).
It is evident that, with an increase in the zinc cation
concentration, the absorption bands corresponding to
both strong and weak Lewis acid sites are shifted to the
low-frequency region. In this case, a decrease in the
intensity of the absorption bands characteristic of the
strongly acidic sites of the original γ-Al2O3, a redistri-
bution of the surface concentration of acid sites, and
the appearance of new absorption bands correspond-
ing to weaker acceptor sites (2155 and 2147 cm−1) are
observed.
Thus, the data on DME conversion in the presence
of freshly prepared ZnO/γ-Al2О3 catalysts and the cat-
alysts coked during reaction, the IR data of probe mol-
ecules adsorbed on the catalyst surface (NH3 and CO),
The relationship between the catalyst genesis and
the kinetic parameters of reactions was studied in [17–
21], where it was shown that by varying catalyst syn-
and the DTA data for the samples suggest that there thesis parameters, it is possible to affect the activation
PETROLEUM CHEMISTRY
Vol. 60
No. 7
2020

DETECTION OF STEADY STATE MULTIPLICITY
783
energy and the factors of interaction of reactants with
active surface sites.
(iv) It has been found that, during DME conver-
sion, an SSM takes place; it is attributed to the pres-
ence of coke on the surface of a catalyst containing a
ZnO/γ-Al2O3 binary composite and controlled by the
ratio of Lewis and Brønsted sites depending on the
process temperature.
Note that aluminum was used as a binder in the
ZnO + Al2O3 binary system in [5–7]. There are com-
mon approaches in the synthesis that are in accor-
dance with the results for the catalyst samples synthe-
sized in this study. For example, impregnation is prac-
ticed by the diffusion of a mixed solution of active
components containing zinc and aluminum, which
are deposited on alumina, and air-drying at room
temperature for 12 h. This procedure is aimed at
strengthening the bond between zinc and aluminum
oxides. The desired final result of strengthening the
bond between the metals would be the formation of
the ZnAl2O4 spinel. Typically, spinels are formed at
fairly high temperatures above 1000°C. However,
Elam et al. [6] reported the presence of a spinel despite
that the highest calcining point did not exceed 500°C.
At the same time, regardless of the presence of a spinel
in the system, according to [5–7, 22], zinc dissolves in
a zinc–aluminum mixture and strengthens the bond
between the metal oxides.
FUNDING
This work was performed under a state task of the Top-
chiev Institute of Petrochemical Synthesis of the Russian
Academy of Sciences.
CONFLICT OF INTEREST
A.L. Maksimov is the editor-in-chief of the Neftekh-
imiya (Petroleum Chemistry) journal. The other authors
declare that there is no conflict of interest to be disclosed
this paper.
REFERENCES
1. A. A. Lamberov, I. N. Mukhambetov, and R. F. Zaly-
Thus, the current fundamental ideas on heteroge-
neous catalysis can be expanded and developed not
only owing to solving the practical problem of catalyst
synthesis, but also owing to a deeper understanding of
the mechanism including the kinetics and dynamics of
the process [21].
aliev, Catal. Ind. 6, 128 (2014).
2. R. Hughes, Deactivation of Catalysts (Academic, Lon-
don, 1984).
3. N. V. Ostrovskii, Catalyst Deactivation Kinetics: Mathe-
matical Models and Their Application (Nauka, Moscow,
2001) [in Russian].
4. V. F. Tret’yakov and R. M. Talyshinskii, Kinetics and
Dynamics of Heterogeneous Catalyzed Petrochemical
Processes (MITKhT, Moscow, 2012) [in Russian].
Note that the chemistry of catalytic DME conver-
sion in terms of the effect on the SSM mechanism was
not studied in [8–10]; therefore, the revealed relation-
ships required a more detailed study.
5. D. Liu, Y. Men, J. Wang, et al., Am. J. Anal. Chem. 7
(7), 568 (2016).
The results of this study suggest that the dynamics
of processes catalyzed by composites based on alumi-
num and zinc oxides used in various reactions involv-
ing oxygenates should be studied in detail [7, 22–24].
Typical objects for studying these effects are multir-
oute reactions that occur in accordance with nonlinear
mechanisms complicated by a significant yield of
byproducts and coking of the catalyst surface [25–33];
this fact should be taken into account to implement
industrial petrochemical processes.
6. J. W. Elam, A. D. Routkevitch, and S. M. George, J.
Electrochem. Soc. 150, 339 (2003).
7. M. A. Kipnis, I. A. Belostotskii, E. A. Volnina, and
G. I. Lin, Catal. Ind. 11, 53 (2019).
8. S. N. Khadzhiev, A. L. Maksimov, V. F. Tret’yakov,
et al., Pet. Chem. 58, 405 (2018).
9. A. L. Maksimov, V. F. Tret’yakov, R. M. Talyshinskii,
et al., Neftegazokhimiya 59 (2), 44 (2019).
10. A. L. Maksimov, V. F. Tret’yakov, Yu. N. Litvishkov,
et al., Khim. Probl., No. 1, 135 (2019).
11. S. V. Lebedev, Life and Works (ONTI Khim. Teoret.,
Leningrad, 1938) [in Russian].
CONCLUSIONS
12. V. I. Bykov, Modeling of Critical Phenomena in Chemical
(i) It has been found that DME conversion occurs
in two temperature regions exhibiting different ten-
dencies in affecting the mode parameters; this finding
indicates the existence of different mechanisms of the
reaction and the presence of an SSM.
(ii) The presence of an SSM is evidenced by the
occurrence of a hysteresis loop between the detected
temperature regions of the reaction in the temperature
range of 380–420°C.
Kinetics (KomKniga, Moscow, 2006) [in Russian].
13. G. S. Yablonskii, V. I. Bykov, and A. N. Gorban’, Ki-
netic Models of Catalytic Reactions (Nauka, Novosi-
birsk, 1983) [in Russian].
14. B. S. Gudkov, A. N. Subbotin, and V. I. Yakerson, Tem-
perature Hysteresis in Heterogeneous Catalysis (Khimi-
ya, Moscow, 2001) [in Russian].
15. A. I. Boyarinov, Multiplicity of Steady States in the Mix-
er–Reactor–Separation Unit System (Khimiya, Mos-
cow, 1982) [in Russian].
16. S. N. Khadzhiev, V. F. Tret’yakov, A. M. Ilolov, and
R. M. Talyshinskii, Nanotekhnol., Nauka Proizvod. 1
(10), 33 (2015).
(iii) The revealed steady states are characterized by
stability, which is enhanced upon the initiation of the
process by hydrogen peroxide.
PETROLEUM CHEMISTRY Vol. 60 No. 7 2020

784
MAKSIMOV et al.
17. A. M. Gyul’maliev, V. F. Tret’yakov, R. M. Taly- 27. L. G. Bruk and O. N. Temkin, Kinet. Catal. 57, 277
shinskii, et al., Pet. Chem. 59, 979 (2019).
(2016).
18. Yu. N. Litvishkov, P. A. Muradova, M. R. Efendiev,
28. P. Debye, Selected Works (Nauka, Leningrad, 1987) [in
et al., Khim. Probl., No. 2, 290 (2009).
Russian].
19. V. F. Tret’yakov, R. M. Talyshinskii, K. V. Tret’yakov,
et al., Neftepererab. Neftekhim., No. 12, 44 (2011).
20. S. L. Kiperman, Foundations of Chemical Kinetics in
Heterogeneous Catalysis (Khimiya, Moscow, 1979) [in
Russian].
29. A. V. Myshlyavtsev and M. D. Myshlyavtseva, Kinet.
Catal. 48, 541 (2007).
30. V. Kh. Fedotov, N. I. Kol’tsov, and B. V. Alekseev,
Dokl. Akad. Nauk SSSR 302, 126 (1988).
21. A. L. Maksimov, V. F. Tret’yakov, and R. M. Taly-
31. L. I. Kheifets and A. V. Neimark, Multiphase Processes
in Porous Media (Khimiya, Moscow, 1982) [in Rus-
sian].
shinskii, Neftegazokhimiya, No. 4, 34 (2018).
22. Kh. M. Kadiev, M. Kh. Kadieva, L. A. Zekel’, et al.,
Colloid J. 81, 90 (2019).
23. M. Cai, V. Subramanian, V. V. Sushkevich, et al., Appl.
32. I. Ya. Mittova, A. A. Samsonov, and B. V. Sladkopevt-
sev, Basic Concepts and Mechanisms of Catalytic and
Coupled Reactions: Nonequilibrium Processes in Cataly-
sis (Classical Systems) (Izd. Voronezhskogo Gosudarst-
vennogo Univ., Voronezh, 2012) [in Russian].
Catal., A 502, 370 (2015).
24. V. L. Sushkevich, I. I. Ivanova, and E. Taarning, Green
Chem. 17, 2552 (2015).
25. O. N. Temkin, Lectures Notes in the Theory of Mecha-
nisms of Complex Reactions and Catalysis (Moscow,
2003) [in Russian].
26. T.G. Alkhazov, and L.Ya. Margolis, High-Level Cata-
lytic Oxidation of Organic Substances (Khimiya, Mos-
cow, 1985) [in Russian].
33. F. A. Yandieva, M. V. Tsodikov, A. V. Chistyakov, et al.,
Kinet. Catal. 51, 548 (2010).
Translated by M. Timoshinina
PETROLEUM CHEMISTRY
Vol. 60
No. 7
2020