Steam Reforming of Methane and Its Mixtures with Propane in a Membrane Reactor with Industrial Nickel Catalyst and Palladium–Ruthenium Foil

Article abstract of DOI:10.1134/S0965544119040054

Abstract: Steam reforming of methane and its mixtures containing 5 and 10% propane has been studied in a membrane reactor with an industrial nickel catalyst NIAP-03-01 and a membrane in the form of 30-μm foil made of a Pd–Ru alloy. At T = 823 K and a feed space velocity of 1800 h^?1, the almost complete methane conversion is achieved, the selectivity for CO₂ is more than 50%, and about 80% H₂ is recovered from the reaction mixture. High conversion of CH₄ in the membrane reactor under mild conditions allows the steam reforming of its mixtures with C₂₊ alkanes to be conducted in a single process, as shown by the example of model mixtures containing C₃H₈. Under selected conditions (T = 773 or 823 K, a feed space velocity of 1800 or 3600 h^?1, a steam/methane ratio of 3 or 5, atmospheric pressure), almost complete C₃H₈ conversion is observed. The main “undesirable” reaction is methanation, leading to a decrease in the CH₄ conversion. In the system under study, CH₄ is formed with an increase in the feed space velocity. Methanation occurs as a result of C₃H₈ hydrocracking at a steam/feedstock ratio = 3 or the hydrogenation of CO₂ as this ratio is increased to 5. The optimal conditions for steam reforming of methane mixtures containing up to 10% C₃H₈ are T = 823 K, steam/feedstock ratio = 5, and the feed space velocity of 1800 h^?1. Under these conditions involving evacuation of the permeate, the feedstock conversion is complete, the selectivity for CO₂ is 50%, and more than 70% H₂ is recovered from the reaction mixture.

Full text of DOI:10.1134/S0965544119040054

ISSN 0965-5441, Petroleum Chemistry, 2019, Vol. 59, No. 4, pp. 394–404. © Pleiades Publishing, Ltd., 2019.
Russian Text © L.P. Didenko, L.A. Sementsova, P.E. Chizhov, T.V. Dorofeeva, 2019, published in Neftekhimiya, 2019, Vol. 59, No. 3, pp. 271–281.
Steam Reforming of Methane and Its Mixtures with Propane
in a Membrane Reactor with Industrial Nickel Catalyst
and Palladium–Ruthenium Foil
L. P. Didenkoa, *, L. A. Sementsovaa, P. E. Chizhova, and T. V. Dorofeevaa
aInstitute of Problems of Chemical Physics, Russian Academy of Sciences, Chernogolovka, Moscow oblast, Russia
*e-mail: ludi@icp.ac.ru
Received August 11, 2018; revised October 1, 2018; accepted December 10, 2018
Abstract—Steam reforming of methane and its mixtures containing 5 and 10% propane has been studied in a
membrane reactor with an industrial nickel catalyst NIAP-03-01 and a membrane in the form of 30-μm foil
made of a Pd–Ru alloy. At T = 823 K and a feed space velocity of 1800 h−1, the almost complete methane
conversion is achieved, the selectivity for CO2 is more than 50%, and about 80% H2 is recovered from the
reaction mixture. High conversion of CH4 in the membrane reactor under mild conditions allows the steam
reforming of its mixtures with C2+ alkanes to be conducted in a single process, as shown by the example of
model mixtures containing C3H8. Under selected conditions (T = 773 or 823 K, a feed space velocity of 1800
or 3600 h−1, a steam/methane ratio of 3 or 5, atmospheric pressure), almost complete C3H8 conversion is
observed. The main “undesirable” reaction is methanation, leading to a decrease in the CH4 conversion. In
the system under study, CH4 is formed with an increase in the feed space velocity. Methanation occurs as a
result of C3H8 hydrocracking at a steam/feedstock ratio = 3 or the hydrogenation of CO2 as this ratio is
increased to 5. The optimal conditions for steam reforming of methane mixtures containing up to 10% C3H8
are T = 823 K, steam/feedstock ratio = 5, and the feed space velocity of 1800 h−1. Under these conditions
involving evacuation of the permeate, the feedstock conversion is complete, the selectivity for CO2 is 50%,
and more than 70% H2 is recovered from the reaction mixture.
Keywords: steam reforming, methane/propane mixtures, membrane reactor
DOI: 10.1134/S0965544119040054
INTRODUCTION
A typical industrial process for H2 production
includes the high-temperature step of steam reforming
of feedstock (1) followed by two steps of steam reform-
ing of CO by water-gas shift reaction (2) proceeding at
a lower temperature and the subsequent step of H2
purification, which in most cases is carried out using
the pressure swing adsorption technique. The yield of
the products is limited by thermodynamic equilibrium
and, due to the high endothermicity of reactions (1)
and (3), temperatures above 973 K are required to
achieve high methane conversion. For example,
the methane conversion higher than 80% can be
obtained at temperatures above 1123 K. In steam
reforming, natural gas is used as feedstock that con-
tains 3–7 vol % C2+ hydrocarbons in addition to meth-
ane. A particular composition of natural gas depends
on the gas field. For example, natural gas of the Uren-
goy field in Russia has the following composition, %:
methane, 87; ethane, 6.2; propane, 3.4; butane, 1.98;
pentane, 0.76; carbon dioxide, 0.12; and nitrogen, 1.1
[2]. In addition to natural gas, associated petroleum
gases (APG) containing on average, %: 50–70 СН4,
5–10 С2Н6, 5–15 С3+, 1–10 N2, and 1–10 СО2 can be
Hydrogen is the most promising clean energy
source derived from natural raw materials. Due to the
rapid development of hydrogen power engineering,
the demand for these raw materials is increasing. The
need for high-purity hydrogen is especially great, as,
for example, it is used for low-temperature fuel cells
with polymeric proton-exchange membranes, where
the CO concentration should not exceed 10 ppm [1].
The main industrial method for producing H2 is steam
reforming of methane on nickel catalysts, in which the
following reactions occur simultaneously:
СН4 + Н2О ↔ СО + 3Н2
(1)
ΔН 0 = 206 kJ/mol,
298
СО + Н2О ↔ СО2 + Н2
(2)
ΔН2098 = –41 kJ/mol,
СН4 + 2Н2О ↔ СО2 + 4Н2
(3)
ΔН2098 = 165 kJ/mol.
394

STEAM REFORMING OF METHANE AND ITS MIXTURES
395
used as hydrocarbon feedstock. However, this valuable the reaction process and H2 removal through the
feedstock is used extremely inefficiently. In Russia,
about 15.7 billion m3/year of APG is flared, which is
about 24% of its total amount and causes significant
environmental damage [3]. The steam reforming of a
methane mixture with C2+ hydrocarbons in a single
process is impossible, since catalyst-deactivating car-
bon deposits are rapidly formed from methane homo-
logues at high temperatures. The rate of this reaction
increases with temperature and with an increase in the
number of carbon atoms in a C2+ hydrocarbon. In this
regard, it is necessary to remove C2+ hydrocarbons
from the feedstock. In the industrial process, for this
purpose, a pre-reforming stage is used, consisting in
the steam conversion of methane homologues on
nickel catalysts at temperatures of 673–773 K to yield
a mixture of methane, hydrogen, and carbon oxides
[4, 5]. To prevent the formation of carbon deposits,
pre-reforming is carried out at a steam to methane
membrane. As a result, the equilibrium of the thermo-
dynamically controlled reaction of steam reforming of
methane is displaced, thereby making it possible to
achieve a high conversion at lower temperatures. In
addition to lowering the reaction temperature, the
simultaneous conversion of CH4 and its closest homo-
logues in the membrane reactor eliminates some steps
of the conventional steam reforming process of natural
gas, such as pre-reforming, additional oxidation of CO
by water-gas shift reaction, and fine H2 purification
from CO. In general, this will greatly simplify the pro-
cess flowchart and reduce energy costs. This demand-
ing problem is the subject matter of this study.
The data presented in the literature show that in a
membrane reactor, it is possible to achieve high meth-
ane conversion at temperatures of 723–873 K [10–12].
A nickel catalyst is most frequently used in the reaction
ratio of 2.5–4. The process involves the following irre- because of its high activity, availability, and low cost.
versible reactions of steam reforming of methane
homologues (СnHm, n ≥ 2):
Not only Ni, but also Ru, Rh, Ir, and Pt are active in
the reaction. Palladium alloys are the most promising
as a hydrogen-selective membrane material due to the
unique properties of palladium, such as are the cata-
lytic surface, on which H2 molecules dissociate into
atoms diffusing at a high rate in the metal lattice, and
the ability of palladium to absorb a large volume of
hydrogen exceeding 600 times its own volume. To
obtain high-purity hydrogen, dense palladium mem-
branes (foils) with 100% hydrogen selectivity are the
most suitable. They are inferior to composite palla-
dium membranes in hydrogen permeability because of
a greater thickness, but have significant advantages,
such as ease of fabrication, ease of connection with the
structural elements of the membrane module, and
ease of regeneration. The performance of palladium
foil can be enhanced by alloying Pd with some addi-
tives that increase the H2 permeability, for example,
93% Pd–7% Y, 60% Pd–40% Cu, 94% Pd–6% Ru,
and 93.5% Pd–6% In–0.5% Ru [13, 14]. Preparing a
thin (down to 7 μm) dense foil from such alloys by
magnetron sputtering will make it possible to achieve
СnHm + nH2O → nCO + (m/2 + n)H2
(4)
ΔН2098 > 0,
CnHm + 2nH2O → nCO2 + (m/2 + 2n)H2
(5)
ΔН2098 > 0.
These reactions are accompanied by the water-gas
shift reaction and reversible exothermic methanation
reactions:
СО + 3Н2 ↔ СН4 + Н2О
(6)
ΔН2098 = –205.9 kJ/mol,
СО2 + 4Н2 ↔ СН4 + 2Н2О
(7)
ΔН2098 = –164.7 kJ/mol,
СnHm + (2n – m/2)H2 → nCH4
(8)
ΔН2098 < 0.
The pre-reforming reaction can be used to produce
H2, and this possibility is being actively investigated a further increase in hydrogen permeability [15–17].
[6–9]. It was shown that the temperature and
The feasibility of steam reforming of CH4 mixtures
steam/hydrocarbon ratio affect in similar ways the
with C2+ homologues in a membrane reactor was
shown by Shirasaki et al. [18]. They presented a mem-
brane reformer for producing high-purity hydrogen by
steam reforming of natural gas at temperatures of 768–
813 K. Natural gas of a constant composition contain-
ing 88.5% CH4, 4.6% C2H6, 5.4% C3H8, and 1.5%
C4H10 was used in their study, but the effect of change
in the feedstock composition on the parameters of the
membrane process was not investigated. However, the
C2+ content in the feedstock may vary depending on
the gas field and thus can affect the feedstock conver-
sion and the selectivity for steam reforming products
conversion of ethane, propane, and butane on a rho-
dium catalyst and the product selectivity [9]. The con-
version of a mixture containing 86.7% CH4, 8.1%
C2H6, 2.0% C3H8, and 0.4% C4H10 occurs in two tem-
perature regions: ethane, propane, and butane are
almost completely consumed to give a mixture of H2
and carbon oxides below 673–723 K, and the conver-
sion of CH4 occurs with a further increase in tempera-
ture [9]. In principle, in order to conduct steam
reforming of a mixture of methane and its closest
homologues in a single process, it is necessary to lower
the methane conversion temperature. For this pur-
pose, membrane technology can be used, combining in the membrane reactor.
PETROLEUM CHEMISTRY Vol. 59 No. 4 2019

396
DIDENKO et al.
In this study, methane mixtures containing 5 and ization and thermal conductivity detectors. The H2
10% C3H8 were taken as models. The steam conver-
content in the products was determined on a column
sion of the mixtures was investigated by varying the with molecular sieves 13X (2 mm × 2 m, 50°C, carrier
temperature, the feed space velocity, and the gas argon). The hydrocarbon composition of the
steam/methane ratio.
The aim of the work was to optimize the conditions
of the membrane process to obtain the maximum con-
version in the reaction of steam reforming of methane
mixtures with propane.
products was determined on a HP-Al/KCl column
(0.5 mm × 30 m, 80°C, carrier gas helium). The con-
centrations of CO and CO2 were determined on a col-
umn packed with activated carbon (2 mm × 2 m,
100°C, carrier gas helium). The concentrations of the
products were calculated using the method of absolute
calibration.
Gas flow rates were controlled with RRG-12 gas
flow regulators (Eltochpribor, Zelenograd). The reac-
tor was heated by an electric oven. The temperature in
the reactor and in the furnace was monitored with the
use of chromel–alumel thermocouples.
EXPERIMENTAL
Membrane Reactor
A foil with a thickness of 30 μm made of Pd–6% Ru
alloy was kindly provided by staff members of Labora-
tory No. 12 at the Institute of Metallurgy and Materi-
als Science of the Russian Academy of Sciences
(IMET RAS, Moscow). Foil blanks were melted in an
electric arc furnace in an inert atmosphere. A foil of a
given thickness was obtained by cold rolling with inter-
mediate vacuum annealing. The foil membrane had a
disk shape with a diameter of 56 mm (effective surface
area of 15.2 cm2). To maintain mechanical strength, it
was sandwiched between stainless steel fine mesh
sheets. The membrane reactor had a feed (retentate)
compartment and an H2 withdrawal (permeate) com-
partment separated by the membrane. The design of
the reactor is detailed in [19]. In the retentate com-
partment, atmospheric pressure was maintained. The
driving force for the withdrawal of H2 from the reac-
tion mixture through the membrane was created by
reducing the partial pressure of H2 in the permeate
compartment. For this purpose, nitrogen as the sweep
gas was fed to the permeate compartment countercur-
rently to the feedstock at a preset flow rate
(150 cm3/min). In some experiments, vacuum pump-
ing was used to extract H2 from the permeate compart-
ment.
The methane conversion (α, %), the product
selectivity (S, %) and the H2 recovery through the
membrane (ϕ, %) were calculated by the equations:
in
out
СН4
V in X
− V out
X
СН4
α =
× 100%,
V in X
in
СН4
out
prod
X
Sprod
=
× 100%,
out
out
out
СО2
X out + X + X
+ X
Н2
СО
СН4
V perm XНperm
2
φ =
× 100%,
V perm XНperm + V ret XНret
2
2
where Vin is the flow rate of the gas stream at the reac-
in
tor inlet, cm3/min;
is the volume concentration
XCH
4
out
of CH in the gas stream at the reactor inlet, %;
is the volume concentration of methane in the prod-
X
4
CH4
out
ucts at the reactor outlet, %;
is the flow rate of
Vprod
products at the reactor outlet, cm3/min; and Xout is the
volume concentration of the product in the mixture
leaving the reactor, %;
The catalyst used in the form of grains of 0.2–
0.4 mm was pretreated directly in the reactor with a
(60% H2–40% Ar)–steam mixture for 60 min at the
temperature of the experiment (preliminary experi-
ments showed that a change in the catalyst pretreat-
ment temperature in the range of 773 –973 K does not
affect its activity).
Steam Reforming of Hydrocarbon Feedstock
The feedstock used was either 100% CH4 or its
mixtures containing 5 and 10 vol % C3H8. The reten-
tate compartment was charged with 2 cm3 (3.35 g) of
the NIAP-03-01 industrial nickel catalyst for steam
reforming of natural gas (manufactured by NIAP-
KATALIZATOR, Novomoskovsk). The feedstock was
mixed with steam in the required ratio and fed at a
given rate to the catalyst through holes located around
the periphery of the retentate compartment, and the
resulting products were withdrawn through the central
hole and passed through a condenser to condense
unreacted water. The volumetric flow rate of “dry”
gases containing H2, CO, CO2, and unreacted CH4
was measured at the reactor outlet using a soap bubble
flow meter, and the gas was fed to a chromatograph.
The product composition was determined on-line
RESULTS AND DISCUSSION
Experiments with 100% Methane
Initially, experiments were performed in the
absence of C3H8. Their goal was to determine reaction
conditions that allow the temperature to be kept as low
as possible while maintaining high rates of methane
conversion in the membrane reactor. The effect of H2
recovery through the membrane on the methane con-
using a Kristall-5000 chromatograph with flame ion- version and the product selectivity was revealed on the
PETROLEUM CHEMISTRY
Vol. 59
No. 4
2019

STEAM REFORMING OF METHANE AND ITS MIXTURES
397
basis of comparison with experiments without H2 ently, in the system under study using the sweep gas,
this balance of rates is not reached, the imbalance
being most pronounced at 773 K, at which an
increase in the methane conversion relative to the
“nonmembrane” value is not observed at all (Fig. 1a,
curves 1, 2). In the experiments with permeate evacu-
ation, the rate of H2 withdrawal from the reaction mix-
removal (“nonmembrane” reaction), in which a gas-
tight stainless steel partition was installed instead of
the membrane. In the majority of studies reported in
the literature, an increased pressure is used in the
retentate compartment of membrane reactor. Despite
the adverse effect on the position of thermodynamic
equilibrium of the reaction under study, an increase in ture increases, thereby leading to an increase in meth-
pressure allows an increase in the H2 flux through the ane conversion over the entire temperature range. The
selectivity for CO is higher than the “nonmembrane”
value, and this difference increases with an increase in
temperature (Fig. 1a, curves 4, 5), which is explained
by a decrease in CO consumption in exothermic
water-gas shift reaction (2). When the permeate is sub-
jected to vacuum pumping, the CO2 selectivity
membrane, which generally has a positive effect on the
feedstock conversion. However, atmospheric pressure
is more preferable as far as the use of steam reforming
of hydrocarbon feedstock in small-scale units with fuel
cells for generating electricity is concerned, where the
use of high-pressure compressor units is problematic
[20]. Thus, the reaction was investigated at atmo- increases and the concentration of CO significantly
spheric pressure in this work.
decreases (Fig. 1a, curve 6) by its consumption in the
water-gas shift reaction.
Effect of temperature. The effect of temperature on
the methane conversion and the selectivity for CO at a
feed space velocity of 3600 h−1 and a steam/methane
ratio of 3 is illustrated in Fig. 1a. It can be seen that
with an increase in temperature, the conversion of
CH4 and the CO selectivity increase (curves 1, 4),
exceeding their values for the “nonmembrane” reac-
tion represented by curves 2 and 5, respectively. These
relationships reflect the favorable effect of tempera-
ture on both endothermic reaction (1) and the flux of
H2 through the membrane. Compared with the exper-
iments without H2 withdrawal, the methane conver-
sion at 973 K is increased by 15.5%. The increase in
conversion relative to the “nonmembrane” value
becomes smaller with a decrease in temperature, and
it does not exist at all at 773 K. This may be due to a
decrease in the H2 uptake through the membrane with
decreasing temperature. For example, it is 64–68% at
923–973 K and 57% at 773 K (Fig. 1b, curve 1). The
reason for the lower recovery of H2 may be a decrease
in its partial pressure in the reaction mixture with a
decrease in temperature. It is possible to increase the
withdrawal of H2 by pumping it out of the permeate
section using a vacuum pump, which can reduce the
H2 pressure in the permeate to 3 mmHg. It can be seen
that during evacuation, the withdrawal of H2 increases
to 79–80% in the interval of 923–973 K, and it is 70%
at 773 K (Fig. 1b, curve 2). Due to an increase in the
H2 removal, the methane conversion increases to
100%, in the temperature range of 923–973 K and to
38% at 773 K (Fig. 1a, curve 3).
Effect of feed space velocity. The effect of the space
velocity of the methane–steam mixture on the meth-
ane conversion and the selectivity for CO and CO2 at
823 K and a steam/methane ratio = 3 is shown in
Fig. 2a. It can be seen that the methane conversion
increases with a decrease in the feed space velocity
from 8.0% at 9600 h−1 to 78% at 1800 h−1 (curve 1).
This is due to the increase in both the time of contact
of feedstock with the catalyst and the H2 recovery from
40% at 9600 h−1 to 70% at 1800 h−1.
The selectivity for CO2 increases from 7% at
9600 h−1 to 34% at 1800 h−1 (curve 2). In this case, the
selectivity for CO varies only slightly, in the range of
7.1–6.2% (curve 3). With an increase in the removal of
H2 by evacuating the permeate, the selectivity for CO2
increases to 50% at 1800 h−1, and the conversion
of methane increases to 100% (shown by arrows in
curves 1 and 2). Carbon dioxide is formed as a result of
consumption of CO, the selectivity for which
decreases (shown by the arrow in curve 3). These data
show that the selectivity of the formation of CO2 by
water-gas shift reaction (2) is significantly affected by
an increase in the rate of H2 removal through the
membrane. Even with an increase in temperature to
973 K, which is unfavorable for exothermic reaction
(2), at low feed space velocities of 1800 and 3600 h−1,
evacuation of the permeate leads to a decrease in
selectivity for CO due to its consumption for the for-
mation of CO2 (Fig. 3).
These results reflect the peculiarity of the mem-
Effect of steam/methane ratio. The methane steam
brane process, in which the condition for maximizing reforming reaction is usually carried out at a
the conversion in a membrane reactor is close values of steam/methane ratio exceeding the stoichiometric
the rates of formation of H2 and its removal through ratio. First, this is necessary for increasing the conver-
sion of CH4 and, second, for preventing the formation
shown in many papers, for example [21–24]. Appar- of carbon deposits. The possibility of using a lower
the membrane, a fact that has been experimentally
PETROLEUM CHEMISTRY Vol. 59 No. 4 2019

398
DIDENKO et al.
(a)
100
90
80
70
60
50
40
30
20
10
0
(b)
100
90
80
70
60
50
40
30
20
10
0
100
90
80
70
60
50
40
30
20
10
0
3
1
2
1
2
4
5
6
800
850
900
950
1000
800
850
900
950
1000
750
750
Temperature, К
Temperature, К
–1
Fig. 1. (a) Temperature dependence of (1–3) methane conversion and (4–6) CO selectivity at a feed space velocity of 3600 h
3
and steam/methane = 3 (1, 4) with H withdrawal (sweep-gas flow rate 150 cm /min), (2, 5) without H withdrawal, and (3, 6)
with permeate evacuation used instead of sweep gas. (b) Influence of temperature on the H recovery through the membrane
2
2
2
using (1) sweep gas and (2) permeate evacuation.
temperature in the membrane reactor makes it possi- under investigation. An increase in the H2 flux
ble to significantly reduce the yield of carbon deposits,
so that a lower steam/methane ratio can be used.
However, a higher ratio is required to increase the
selectivity for CO2. Therefore, there is an optimum of
2.0–4.0, which is commonly used in the membrane
process. The effect of the steam/methane ratio on the
methane conversion and the selectivity for carbon
oxides in the membrane reactor at T = 823 K and a
space velocity of the methane–steam mixture of
3600 h−1 is shown in Fig. 2b. It can be seen that with
an increase in the ratio from 2 to 5, the methane con-
version increases from 45 to 80% (curve 1), the selec-
tivity for CO decreases slightly from 5.0 to 4.7%
(curve 3), and the selectivity for CO2 increases from 20
to 40% (curve 2). Despite the fact that the reaction
mixture is dilute and the partial pressure of H2
decreases with an increase in the steam excess, the
increasing methane conversion compensates for the
negative effect of dilution and, therefore, the H2
recovery generally increases from 45% at a ratio of 2–
3 to 50% at a ratio of 4–5.
through the membrane during the evacuation of per-
meate leads to the almost complete conversion of CH4
and a CO2 selectivity of about 50%. In connection
with this, the mixtures of methane with propane were
studied at temperatures of 773 and 823 K, space veloc-
ities of 1800 and 3600 h−1, and steam/feedstock ratios
of 3 and 5.
Experiments with Mixtures of Methane and Propane
The steam reforming of C2+ alkanes, as well as their
mixtures with methane, was studied predominantly in
the “nonmembrane” reactions. The alkane conver-
sion and the product selectivity are mainly affected by
the temperature and steam/feed ratio. At a low ratio
(0.39–0.58) in the temperature range of 423–598 K,
along with alkane steam reforming, reactions (6) and
(7) occur, leading to the formation of СН4 [25].
During the conversion of a mixture containing
17 vol % C3H8, 34 vol % CH4, and 49 vol % H2O, both
the steam reforming of propane and hydrocracking
according to reaction (8) occur on the nickel catalyst
NIAP-18 in the temperature range of 523–723 K,
leading to an increase in the CH4 content, with the
selectivity for CH4 increasing with decreasing tem-
perature [26]. The complete conversion of C2+ alkanes
In general, the results presented above show that
under the mild conditions (atmospheric pressure and
823 K) in the membrane reactor, an 80% methane
conversion can be achieved at methane–steam space
velocities of 1800–3600 h−1 in the catalytic system
PETROLEUM CHEMISTRY
Vol. 59
No. 4
2019

STEAM REFORMING OF METHANE AND ITS MIXTURES
399
(a)
80
60
40
20
0
(b)
(evacuation)
100
80
60
40
20
0
80
60
40
20
0
100
80
60
40
20
0
1
(evac)
1
2
2
3
3
(evac)
3
4
5
2
4000 6000 8000 10000
Space velocity, h–1
2000
Steam/СH4
Fig. 2. Dependence of (1) methane conversion and selectivity for (2) CO and (3) CO on (a) the feed space velocity at a
2
−1
steam/methane ratio of 3 and (b) the steam/methane ratio at a space velocity of 3600 h . T = 823 K.
in the reaction with steam is observed in the tempera- nificantly higher feed space velocities of 1800 and
3600 h−1 and temperatures of 773 and 823 K, selected
as a result of the above-described experiments with
100% CH4. Tables 1 and 2 present the results of the
study of the mixtures containing 5 and 10% C3H8 at
steam/feedstock ratios of 3 and 5, respectively. To
remove H2 from the reaction mixture through the
membrane, both sweep gas and permeate evacuation
were used. It can be seen that the C3H8 conversion was
complete under all the conditions.
As shown above in the general form for СnНm
alkanes, their interaction with steam can occur via
reactions (4)–(8). Using the example of the “non-
membrane” reaction of steam reforming of a methane
mixture containing 17% C3H8 and 49% H2O in the
temperature range of 523–753 K, it was shown that
depending on temperature, C3H8 is converted by the
following reactions [26]:
ture range of 723–773 K; that is, the reaction is almost
irreversible. However, simultaneously with steam
reforming, the hydrogenation reactions of carbon
oxides occur to give CH4; these reactions are limited
by thermodynamic equilibrium, which impedes the
achievement of high conversion of a C2+ alkane.
Therefore, it is reasonable to carry out these reactions
in a membrane reactor, where the removal of H2
through the membrane reduces the contribution of
these reactions and, thus, increases the alkane conver-
sion and the yield of H2. This suggestion was con-
firmed by Rarib et al. [28], who studied the steam
reforming of propane in a fluidized bed membrane
reactor using the Haldor Topsoe nickel-containing
industrial catalyst RR-212 and a membrane in the
form of Pd77Ag23-alloy foil at temperatures of 723–823
K and showed a significant increase in propane con-
version compared with the “nonmembrane” reaction.
The reaction was investigated at a low feed flow rate,
which was 0.595 mL/min, and a temperature of 798 K.
Under these conditions, the yield of H2 was 9.26
moles, which is close to the theoretically possible
maximum of 10 moles per mole of propane. However,
during the study, the H2 selectivity of the membrane
decreased due to the appearance of defects. In this
work, we studied the steam reforming of methane
mixtures with propane in the membrane reactor at sig-
at temperatures less than 623 K:
(5а)
(8а)
(7а)
С3Н8 + 6Н2О → 3СО2 + 10Н2,
С3Н8 + 2Н2 → 3СН4,
СО2 + 4Н2 ↔ СН4 + 2Н2О
at temperatures above 623 K, C3H8 steam reforming
predominantly yields CO, which subsequently enters
into reactions with H2 and H2O:
PETROLEUM CHEMISTRY Vol. 59 No. 4 2019

400
DIDENKO et al.
(4а)
2
1
С3Н8 + 3Н2О → 3СО + 7Н2,
СО + 3Н2 ↔ СН4 + Н2О,
СО + Н2О ↔ СО2 + Н2.
100
90
80
70
60
50
40
30
20
10
0
100
(6а)
(2а)
80
60
40
20
0
It can be seen that the propane transformation
scheme includes methanation reactions (6a), (7a),
and (8a) along with steam reforming reactions (4a)
and (5a). The formation of methane is undesirable
because it reduces the conversion parameter. In this
context, the analysis of the data included in Tables 1
and 2 has focused on determining the conditions
under which these undesirable reactions may occur.
The effect of C3H8 was established by comparing the
data for the mixtures and 100% CH4 under the same
conditions. The data presented in Table 1 show that
with the steam/feedstock ratio of 3, an increase in the
CH4 content of the products is observed only at T =
4
3
823 K and the feed space velocity of 3600 h−1, and this
leads to a decrease in methane conversion from 64.4%
for 100% CH4 to 56.8% for its mixture containing 10%
C3H8. An increase in the rate of H2 removal from the
reaction products by evacuating the permeate does not
have a noticeable effect on the methane conversion,
and the selectivities for CO and CO2 also change little.
These data suggest that the formation of CH4 leads to
irreversible reaction (8a). Apparently, as a result of a
decrease in the time of feedstock contact with the cat-
alyst with an increase in the feed space velocity, the
conversion of C3H8 occurs not only by the reaction of
steam reforming, but also by the hydrocracking reac-
tion. With a decrease in space velocity to 1800 h−1,
C3H8 hydrocracking is not observed and methane and
propane are converted by the steam reforming and
water-gas shift reactions, as indicated by closeness of
product compositions and conversion values. The
results of similar studies at the steam/feedstock ratio of
5 (see Table 2) that both at 773 and 823 K, products
also have a higher CH4 content at the higher feed space
4000
6000
8000
10000
2000
Space velocity, h–1
Fig. 3. Influence of feed space velocity on (1, 2) methane
conversion and (3, 4) CO selectivity at T = 973 K and a
steam/methane ratio of 3 in experiments with (1, 3) sweep
gas and (2, 4) permeate evacuation.
close; that is, the formation of CH4 does not occur.
The almost complete conversion of CH4 from a mix-
ture containing 10% C3H8 was attained by evacuating
the permeate at 823 K and a space velocity of 1800 h−1.
Under these conditions, the selectivity for CO2 is 53%
and more than 70% of H2 is removed from the mix-
ture.
Catalyst and Membrane Stabilities
Catalytic reactions of conversion of alkanes are
complicated by the formation of carbon deposits. In a
membrane reactor, their yield generally increases,
with the increase being due to a decrease in the stabi-
lizing effect of H2 as its content in the reaction mixture
decreases. Carbon deposits can accumulate in the cat-
alyst bed, resulting in catalyst deactivation. In addi-
tion, as a result of the buildup of carbon deposits on
the surface of the membrane, its hydrogen permeabil-
ity may decrease and defects may appear, leading to a
decrease in H2 permselectivity. The stability of the
given catalytic system in the membrane reactor was
studied using a methane mixture containing 5% C3H8
velocity of 3600 h−1. As a result, the conversion of CH4
from mixtures with C3H8 is lower than in the absence
of C3H8 to be 46.9% at T = 773 K and 62.9% at 823 K.
In the case of permeate evacuation, the values of
these parameters increase significantly and the CO2
selectivity also increases. The effect of the increase in
the H2 withdrawal rate, which is observed when the
permeate is evacuated, on the methane conversion
indicates the reversible nature of a reaction by which
methane can be produced. This reaction can be
reversible reaction (7a) of CO2 hydrogenation. This
assumption is supported by an increase in CO2 selec-
tivity with increasing methane conversion. At a space
velocity of 1800 h−1, the product compositions and the
as feedstock at T = 823 K, a space velocity of 3600 h−1,
and a steam/feedstock ratio of 3. The sweep gas was
used to remove H2. It was found that within 50 h, the
methane conversion varied within 60–62% and the
conversions of methane and its mixtures with C3H8 are selectivites for CO and CO2 were 7.0– 8.5 and 21–
PETROLEUM CHEMISTRY
Vol. 59
No. 4
2019

STEAM REFORMING OF METHANE AND ITS MIXTURES
401
PETROLEUM CHEMISTRY Vol. 59 No. 4 2019

402
DIDENKO et al.
PETROLEUM CHEMISTRY
Vol. 59
No. 4
2019

STEAM REFORMING OF METHANE AND ITS MIXTURES
403
23%, respectively, indicating that the catalyst is resis- containing 5% C3H8 for 50 h at 823 K, a steam/feed-
stock ratio of 3, and a feed space velocity of 3600 h−1
has show that the catalytic system is not deactivated
and the membrane retains H2 permeability and
remains defect-free. The results show that the most
favorable conditions for the steam reforming of meth-
ane mixtures containing up to 10% C3H8 are the tem-
perature of 823 K, the steam/feedstock ratio of 5, and
the space velocity of 1800 h−1. Under these conditions,
when the permeate is subjected to vacuum pumping,
the complete conversion of the feedstock is observed,
the selectivity for CO2 is 50%, and about 70% of H2 is
removed from the reaction mixture.
tant to deactivation by carbon deposits. The recovery
of H2 through the membrane varied within 65–67%;
that is, the membrane permeability to hydrogen was
constant. According to chromatographic analysis, the
permeate contained H2 only, indicating the absence of
defects in the membrane. Regarding the service life of
the membrane, it retained the hydrogen permeability
and was defect-free during all the experiments, the
results of which are presented in this paper, lasting for
about 350 h.
CONCLUSIONS
The reaction of steam reforming of methane and its
mixtures containing 5 and 10% C3H8 on the industrial
nickel catalyst NIAP-03-01 was investigated at
steam/feedstock ratios 3 and 5 and space velocities of
1800 and 3600 h−1. The results of the experiments with
100% CH4 showed that by carrying out the reaction in
a membrane reactor, high feedstock conversion can be
achieved under mild conditions. At atmospheric pres-
sure, T = 823 K, a space velocity of 1800 h−1, and
steam/methane ratios of 3–5, the methane conversion
is 80–90%. With an increase in the rate of H2 removal
from the reaction mixture by evacuating the permeate,
almost complete conversion of CH4 is observed, the
selectivity for CO2 is more than 50%, and about 80%
of H2 is removed from the reaction medium. This
result shows that in the membrane reactor, it is possi-
ble to convert mixtures of methane with the closest
homologues, as demonstrated by the example of
model mixtures containing 5 and 10% C3H8. The
steam reforming of the mixtures was studied at atmo-
spheric pressure, temperatures of 773 and 823 K, space
velocities of 1800 and 3600 h−1, and steam/feedstock
ratios of 3 and 5. Under these conditions, the complete
conversion of propane was observed. Simultaneously
with the steam reforming of C3H8, methanation reac-
tions can occur. Their occurrence is undesirable
because it leads to a decrease in the conversion param-
eter. It has been established that under the chosen
conditions, the formation of CH4 occurs at a higher
ACKNOWLEDGMENTS
This work was carried out under the 2013–2020
Program of Basic Research of State Academies of Sci-
ences, theme code 0089-2014-0032.
REFERENCES
1. S. J. Peighambardoust, S. Rowshanzamir, and
M. Amjjadi, J. Hydrogen Energy 35, 9349 (2010).
2. S. A. Skrylev, Energ. Strateg., No. 2, 15 (2012).
3. I. V. Kozodoev, N. A. Kuzin, Yu. I. Amosov, et al., Ind.
Ecol. North. 11, 40 (2011).
4. G. S. Fal’kovich, N. N. Rostanin, L. M. Vilenskii,
et al., Katal. Prom-sti, No. 3, 10 (2003).
5. V. S. Arutyunov and O. V. Krylov, Oxidative Conversion
of Methane (Nauka, Moscow, 1998) [in Russian].
6. K. Aasberg-Petersen, I. Dybkjaer, C. V. Ovesen, et al.,
J. Nat. Gas Sci. Eng. 3, 423 (2011).
7. S. Yoon and J. Bae, Catal. Today 156, 49 (2010).
8. X. Wang, X. Zou, X. Wang, et al., Int. J. Hydrogen
Energy 36, 4908 (2011).
9. B. T. Schadel, M. Duisberg, and O. Deutschmann,
Catal. Today 142, 42 (2009).
10. Y. Chen, Y. Wang, H. Xu, and G. Xiong, Appl. Catal.,
B 80, 283 (2008).
11. M. Saric, Y. C. Delft, R. Sumbharaju, et al., Catal.
Today 193, 74 (2012).
12. J. Y. Matsumura and J. Tong, Top. Catal. 51, 283
(2008).
feed space velocity, which is 3600 h−1. When the
steam/feedstock ratio is 3, the formation of CH4 is
observed at 823 K. It is assumed that under these con-
ditions, CH4 is formed as a result of the irreversible
hydrocracking reaction of C3H8. With an increase in
this ratio to 5, methane is produced both at 773 and
823 K, as a result of the reversible CO2 hydrogenation
reaction. With an increase in the rate of H2 removal by
evacuating the permeate, the contribution of this reac-
tion decreases, which leads to an increase in conver-
sion. In general, the presented results show that it is
possible to perform steam reforming of CH4 mixtures
containing up to 10% C3H8 in a membrane reactor.
The study of steam reforming of a mixture of methane
13. G. Burkhanov, N. Gorina, N. Kolchugina, and
N. Roshan, Platinum Met. Rev. 55, 3 (2011).
14. L. Didenko, V. Savchenko, L. Sementsova, and
L. Bikov, Int. J. Hydrogen Energy 41, 307 (2016).
15. V. M. Ievlev, K. A. Solntsev, A. I. Dontsov, et al., Tech.
Phys. 61, 467 (2016).
16. V. M. Ievlev, A. A. Maksimenko, A. I. Sitnikov, et al.,
Materialovedenie, No. 2, 37 (2016).
17. V. M. Ievlev, G. S. Burkhanov, A. A. Maksimenko,
et al., Kondens. Sredy Mezhfaz. Granitsy 15, 121
(2013).
18. Y. Shirasaki, T. Tsuneki, Y. Ota, et al., Int. J. Hydrogen
Energy 34, 4482 (2009).
19. L. P. Didenko, L. A. Sementsova, P. E. Chizhov, et al.,
Russ. Chem. Bull. 65, 1997 (2016).
PETROLEUM CHEMISTRY Vol. 59 No. 4 2019

404
DIDENKO et al.
20. S. Campanari, E. Macchi, and G. Manzolini, Int. J. 25. M. V. Zyryanova, P. V. Snytnikov, A. B. Shigarov, et al.,
Hydrogen Energy 33, 1361 (2008).
Fuel 135, 76 (2014).
26. M. M. Zyryanova, S. D. Badmaev, V. D. Belyaev, Yu. I.
Amosov, P. V. Snytnikov, V. A. Kirillov, V. A. Soby-
anin, Catal. Ind. 5, 312 (2013).
21. P. Ciavarella, D. Casanave, H. Moueddeb, et al., Catal.
Today 67, 177 (2001).
22. L. van Dyk, S. Miachon, L. Lorenzen, et al., Catal.
27. B. T. Schadel, M. Duisberg, and O. Deutschmann,
Today 82, 167 (2003).
Catal. Today 142, 42 (2009).
23. S. Miachon and J.-A. Dalmon, Top. Catal. 29, 59
28. M. A. Rarib, J. R. Grace, C. J. Lim, and S. S. E. H. El-
(2004).
nashaie, Int. J. Hydrogen Energy 35, 6276 (2010).
24. L. P. Didenko, V. I. Savchenko, L. A. Sementsova, and
Translated by S. Zatonsky
P. E. Chizhov, Pet. Chem. 56, 459 (2016).
PETROLEUM CHEMISTRY
Vol. 59
No. 4
2019