CO Poisoning of Ru Catalysts in CO2Hydrogenation under Thermal and Plasma Conditions: A Combined Kinetic and Diffuse Reflectance Infrared Fourier Transform Spectroscopy-Mass Spectrometry Study

Article abstract of DOI:10.1021/acscatal.0c03620

Plasma-catalysis systems are complex and require further understanding to advance the technology. Herein, CO poisoning in CO2 hydrogenation over supported ruthenium (Ru) catalysts in a nonthermal plasma (NTP)-catalysis system was investigated by a combined kinetic and diffuse reflectance infrared Fourier transform spectroscopy-mass spectrometry (DRIFTS-MS) study and compared with the thermal catalytic system. The relevant findings suggest the coexistence of the Langmuir-Hinshelwood and Eley-Rideal mechanisms in the NTP-catalysis. Importantly, comparative study of CO poisoning of the Ru catalyst was performed under the thermal and NTP conditions, showing the advantage of the hybrid NTP-catalysis system over the thermal counterpart to mitigate CO poisoning of the catalyst. Specifically, compared with the CO poisoning in thermal catalysis due to strong CO adsorption and associated metal sintering, in situ DRIFTS-MS analysis revealed that the collisions of reactive plasma-derived species in NTP-catalysis could remove the strongly adsorbed carbon species to recover the active sites for CO2 activation. Thus, the NTP-catalysis was capable of preventing CO poisoning of the Ru catalyst in CO2 hydrogenation. Additionally, under the NTP conditions, the NTP-enabled water-gas shift reaction of CO with H2O (which was produced by CO/CO2 hydrogenation) shifted the equilibrium of CO2 hydrogenation toward CH4 production.

Full text of DOI:10.1021/acscatal.0c03620

pubs.acs.org/acscatalysis
Research Article
CO Poisoning of Ru Catalysts in CO2 Hydrogenation under Thermal
and Plasma Conditions: A Combined Kinetic and Diffuse Reflectance
Infrared Fourier Transform Spectroscopy−Mass Spectrometry Study
Shanshan Xu, Sarayute Chansai, Shaojun Xu, Cristina E. Stere, Yilai Jiao, Sihai Yang,
Christopher Hardacre,* and Xiaolei Fan*
Cite This: ACS Catal. 2020, 10, 12828−12840
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Plasma-catalysis systems are complex and require further understanding to
advance the technology. Herein, CO poisoning in CO2 hydrogenation over supported
ruthenium (Ru) catalysts in a nonthermal plasma (NTP)-catalysis system was investigated by
a combined kinetic and diffuse reflectance infrared Fourier transform spectroscopy−mass
spectrometry (DRIFTS−MS) study and compared with the thermal catalytic system. The
relevant findings suggest the coexistence of the Langmuir−Hinshelwood and Eley−Rideal
mechanisms in the NTP-catalysis. Importantly, comparative study of CO poisoning of the Ru
catalyst was performed under the thermal and NTP conditions, showing the advantage of the
hybrid NTP-catalysis system over the thermal counterpart to mitigate CO poisoning of the catalyst. Specifically, compared with the
CO poisoning in thermal catalysis due to strong CO adsorption and associated metal sintering, in situ DRIFTS−MS analysis revealed
that the collisions of reactive plasma-derived species in NTP-catalysis could remove the strongly adsorbed carbon species to recover
the active sites for CO2 activation. Thus, the NTP-catalysis was capable of preventing CO poisoning of the Ru catalyst in CO2
hydrogenation. Additionally, under the NTP conditions, the NTP-enabled water-gas shift reaction of CO with H2O (which was
produced by CO/CO2 hydrogenation) shifted the equilibrium of CO2 hydrogenation toward CH4 production.
KEYWORDS: nonthermal plasma (NTP) catalysis, CO poisoning, Ru catalyst, CO2 hydrogenation, kinetics, DRIFTS−MS
In addition to the activity, the stability and longevity of the
catalysts are important factors for practical catalysis under both
thermal and NTP conditions. Under thermal conditions during
CO2 hydrogenation, catalyst deactivation is mainly caused by
(i) metal particle sintering due to high reaction temperatures,
(ii) coking caused by carbon deposition, and (iii) catalyst
poisoning resulting from trace impurities in the feed gases such
as carbon monoxide (CO). Due to the low-temperature
activation of catalytic process, NTP-assisted CO2 hydro-
genation intrinsically avoids sintering and coking processes.9
Regarding catalyst poisoning, specifically CO2 hydrogenation,
it is well known that CO poisoning is one of the worst catalyst-
deactivating processes under thermal conditions.11,12 Under
plasma conditions, conversely, previous studies have shown
that the plasma could enable the recovery of poisoned catalytic
sites via dynamic collisions among reactive plasma-derived
species, which lead to the desorption of strongly bound surface
species.13,14 Accordingly, insights into CO poisoning under
1. INTRODUCTION
Hybrid nonthermal plasma (NTP) and catalysis (NTP-
catalysis) systems can activate and convert a variety of stable
molecules, such as carbon dioxide (CO2), methane (CH4), and
nitrogen (N2), into desired products under mild conditions,
e.g., ambient pressure and low bulk gas temperatures (<200
°C),1−3 but the hybrid system is highly complex. NTP-catalysis
is particularly beneficial to enable kinetically and/or
thermodynamically limited reactions, including dry reforming
of methane,4 water-gas shift,5,6 and CO2 hydrogenation.7 In
comparison with the thermal counterparts, NTP-catalysis has
shown the capability of lowering the energy barrier required for
the catalysis and/or changing the reaction pathways on the
catalyst surface.8,9 Recent studies have shown that, being
similar with the thermal catalysis, the intrinsic nature of
heterogeneous catalysts (including the supports), such as metal
dispersion and pore structure, plays a key role in NTP-
catalysis. For example, a series of Ni supported on silicalite-1
(with different pore structures) catalysts was designed to study
CO2 hydrogenation under NTP conditions. It was found that
the pore structure of the silicalite-1 supports determines the
dispersion and location of Ni sites and, hence, the accessibility
of plasma-generated reactive species, thus affecting the
performance of the NTP-catalysis.10
Received: August 18, 2020
Revised: October 8, 2020
https://dx.doi.org/10.1021/acscatal.0c03620
© XXXX American Chemical Society
ACS Catal. 2020, 10, 12828−12840
12828

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
thermal and NTP conditions, especially relevant deactivation
mechanisms, need to be assessed to develop mature NTP-
catalysis technology for potential practical adoptions. NTP-
catalysis is a complex combination of plasma discharge and
surface reactions (and other factors) with multifaceted
interplays between them. Regarding the surface reactions
under NTP conditions, in situ techniques, such as diffuse
reflectance infrared Fourier transform (DRIFTS)7,9 and
extended X-ray absorption fine structure (EXAFS) spectros-
copy,15 have been proved to be powerful tools to gain insights
into the surface dynamics of the catalyst, reaction mechanisms,
and the catalyst state during NTP-catalysis, which can facilitate
the rational design of bespoke catalysts for NTP conditions.
However, relevant in situ studies of NTP-catalysis toward the
understanding of catalyst poisoning are still lacking.
This work presents the comparative study of the effect of
CO on CO2 hydrogenation over a supported Ru catalyst (i.e.,
CO poisoning) under thermal and NTP conditions. The
intrinsic nature of the catalysts on the performance of CO2
hydrogenation was first studied, and the Ru/SiO2 catalyst with
high activity and stability was chosen for further investigation.
To elucidate the mechanism of CO poisoning, the mechanistic
investigation of CO2 hydrogenation including the kinetic and
in situ DRIFTS studies was comparatively performed under
thermal and NTP conditions, which provide useful information
on the intermediates and reaction pathways of CO2 hydro-
genation. Finally, the mechanism of CO poisoning in CO2
hydrogenation over the Ru/SiO2 catalyst was investigated.
Under the thermal conditions, significant catalyst deactivation
due to the strong CO adsorption and metal sintering was
observed; conversely, NTP activation was found to mitigate
the effect of CO on the performance of the catalyst and
regenerate the catalyst efficiently.
have been described elsewhere.9 Parameters of the NTP-
catalysis system were measured using an oscilloscope
(Tektronix TBS1072B) connected with a high-voltage probe
(Tektronix, P6015) and current monitor. NTP-catalysis was
performed at atmospheric pressure without a heating source.
Briefly, ∼100 mg of catalyst (pellet sizes of 250−425 μm) was
packed into a quartz tube (6 mm o.d. × 4 mm i.d.), where an
aluminum foil wrapped outside of the tube served as the high-
voltage electrode and a tungsten rod (1 mm o.d.) in the center
of reactor acted as the ground electrode. Since the catalyst was
exposed to air at RT before being loaded to the DBD reactor, it
was treated in situ by NTP (at 6.5 kV) using 50% H2/Ar as the
discharge gas for 20 min before catalysis. The feed of CO2, H2,
and Ar balance (molar ratio of 1:3:3) was introduced by mass
flow controllers (Bronkhorst, F-201CV-500-RAD-11-V) with
the flowrate of 50 mL min−1. The applied voltage was from 5.5
to 7.5 kV at a constant frequency of 21.0 kHz. The product
was analyzed by using online mass spectrometry (MS, Hiden
HPR-20) and two-channel online gas chromatography (GC)
equipped with a Porapak Q packed column, thermal
conductivity detector (TCD), and flame ionization detector
(FID). An Ar balance was used in the system to avoid the
signal saturation of MS signal. For each measurement, three
samples of gas products were analyzed under steady-state
conditions for an average value and error determination.
Control experiments using the empty reactor (catalyst-free)
and the reactor with the bare supports as a packing were
performed under the same NTP conditions.
CO poisoning study under the NTP condition (at 6.5 kV
and 21.0 kHz) was investigated by varying the inlet molar ratio
of CO/CO2 between 0 and 2. The total gas feed flowrate was
50 mL min−1, corresponding to a space velocity of 30,000 mL
(STP) gcat−1 h−1, which included CO2, CO, H2, and Ar balance
(molar ratio of H2/(CO2 + CO) = 3). Catalyst deactivation
was monitored as a function of time-on-stream (ToS) by
switching the CO on and off in the feed. The average bulk
temperature of the system between 5.5 and 7.5 kV was
measured using an infrared (IR) thermometer and was in the
range of 110−135 °C. Specifically, the average bulk temper-
ature at 6.5 kV was ∼129 °C, which could not activate CO2
conversion thermally, according to a previous study.9
2. EXPERIMENTAL SECTION
2.1. Preparation and Characterization of Catalysts.
Ruthenium (III) chloride trihydrate (RuCl3·3H2O), silicon
dioxide, and γ-Al2O3 were purchased from Sigma-Aldrich and
used without further purification.
Supported Ru catalysts including Ru/SiO2 and Ru/γ-Al2O3
(with the theoretical metal loading of 2 wt %) were prepared
using the wet impregnation method. First, the support (1.5 g)
was suspended in water (30 mL), and then 6.2 mL of RuCl3·
3H2O solution (10 mg mL−1) was added dropwise. The
mixture was vigorously stirred for 3 h and then evaporated
using a rotary evaporator. The resulting precipitate was dried at
70 °C in a convection oven for 12 h. The obtained dry solid
was subsequently reduced in pure H2 at 300 °C for 2 h with a
heating rate of 5 °C min−1. After reduction, the sample was
cooled down to room temperature (RT) naturally under the
H2 flow (at 100 mL min−1). The actual metal loading was
determined by inductively coupled plasma optical emission
spectrometry (ICP-OES; Supporting information). The
prepared catalysts were characterized to understand their
physical and chemical properties by bright-field transmission
electron microscopy (TEM), N2 physisorption (using the
Brunauer−Emmett−Teller (BET) method), hydrogen temper-
ature programmed reduction (H2-TPR), and CO chemisorp-
tion, and the relevant experimental details are provided in the
Supporting Information.
For comparison, thermal catalysis was carried out at 250−
430 °C at atmospheric pressure. Prior to catalysis, the catalyst
(pellets, about 100 mg) was first treated at 300 °C for 1 h in
50% H2/Ar. Then, the feed (CO2/H2/Ar = 1:3:3) was
introduced into the reactor at 50 mL min−1. The temperature
of the catalyst bed was monitored by a K-type thermocouple
embedded in the catalyst bed.
CO poisoning of the catalyst under the thermal condition
(at 330 °C) was studied using the same gas condition as in the
relevant NTP-catalysis. The catalyst deactivation experiment
was performed at 330 °C with the same gas conditions as
described in the NTP-catalysis (for CO poisoning study).
CO2 (XCO ) conversion, CO (XCO) conversion, carbon (XC
2
= XCO + XCO) conversion, selectivity toward CH4 (SCH ), and
2
4
CH4 yield (YCH ) were determined accordingly to evaluate the
4
catalytic performance (all the performance parameters are
defined in the Supporting Information).
2.3. Kinetic Study. The kinetic study of thermal catalysis
was performed at 260−320 °C with ∼30 mg of catalyst
(diluted with inert glass beads to prevent hot spots) to ensure
low CO2 conversions of <20%. The feed mixture containing
2.2. Catalysis. A dielectric barrier discharge (DBD) flow
reactor was used for NTP-activated CO2 hydrogenation, which
is depicted in Figure S1, and the details of the DBD reactor
12829
https://dx.doi.org/10.1021/acscatal.0c03620
ACS Catal. 2020, 10, 12828−12840

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 1. Performance of NTP-activated catalytic CO2 hydrogenation as a function of voltage/input energy over the Ru/SiO2 and Ru/γ-Al2O3
catalysts in reference to the control experiments; (a) CO2 conversion and (b) CH4 yield. Experimental conditions: feed gas composition of CO2/
−1
H2/Ar = 1:3:3 and WHSV of 30,000 mL (STP) gcat h−1.
CO2/H2/Ar (molar ratio = 1:3:3) was fed into the reactor for
the kinetic study. To extract the reaction order with respect to
H2 and CO2 partial pressures, the composition of the feed was
varied; i.e., H2 partial pressure was changed with a constant
partial pressure of CO2 and vice versa.
with the bare γ-Al2O3 and SiO2 supports were only selective to
CO with relevant CO2 conversions of ∼13 and ∼15%,
respectively, at 6.5 kV. In comparison with the system without
a packing, that is, the blank experiment with an empty tube, the
higher CO2 conversions with the bare supports can be
attributed to the enhanced average electric field strength,
benefiting CO2 dissociation.16 Conversely, in NTP-catalysis
with the voltage above 7.0 kV (SIE > 2.5 J mL−1), regardless of
the Ru catalysts used in this work, CO2 conversion and CH4
yield increased significantly to >57%. However, Ru catalysts
based on different supports showed different behaviors under
the NTP conditions, demonstrating the effect of catalyst design
on NTP-catalysis.17 Specifically, the Ru/SiO2 catalyst showed a
higher activity as compared with the Ru/γ-Al2O3 catalyst,
especially at the lower voltage of <6.5 kV (SIE < 2.5 J mL−1).
The highest CO2 conversion (∼65%) and CH4 yield (∼63%)
at 6.5 kV were achieved by Ru/SiO2, while the Ru/γ-Al2O3
catalyst only showed about 35% CO2 conversion and 29% CH4
yield, being less active for NTP-activated CO2 hydrogenation.
Similarly, under thermal conditions (Figure S2), Ru/SiO2
outperformed Ru/γ-Al2O3 as well, suggesting that the intrinsic
nature of catalysts dominated the performance of CO2
hydrogenation regardless of the means of activation. In detail,
the corresponding TEM and CO chemisorption analysis of the
Ru/SiO2 and Ru/γ-Al2O3 catalysts (Figure S3 and Table S3)
showed that the two catalysts presented similar average particle
sizes and Ru dispersions. Additionally, the metal−support
interaction of the two catalysts was also similar, as revealed by
H2-TPR (Figure S5). These findings show that the property of
the supported active Ru phases of the two catalysts under study
is similar, thus suggesting that the supported Ru might not
affect the activity of the two catalysts significantly under NTP
conditions.
Kinetic study of the NTP-catalysis was performed using
similar procedures and conditions as described above (about
30 mg of catalyst diluted with glass beads, at 5.0−6.5 kV and
21.0 kHz). The gas conditions were the same as in the kinetic
study of the thermal catalysis. Due to the low bulk temperature
under the NTP conditions (<129 °C), thermal activation of
CO2 was not possible. Considering the effect of support
packing and discharge volume, control experiments using the
same amount of the bare supports and inert glass beads were
performed to extract the information on the relevant gas phase
and surface (over the bare supports) reactions under NTP,
which was subsequently used to correct the kinetic data of the
NTP-catalysis (Supporting Information).
2.4. In Situ DRIFTS−MS. The experimental setup of
DRIFTS−MS for NTP-catalysis was described elsewhere.9 The
catalyst was pretreated with 50% H2/Ar gas under NTP at 6.0
kV and 27.0 kHz for 20 min in the flow cell. Then, the gas
mixture containing CO2, CO, H2, and Ar balance was fed into
the cell for the reaction. Kr at 10 mL min−1 was also
introduced as the internal standard. The use of Ar balance in
DRIFTS experiments was to avoid the signal saturation of IR
spectra and MS signal. NTP-catalysis in the DRIFTS cell was
performed at a constant peak voltage of 5.5 kV to avoid arcing
between the electrodes. The IR spectra were recorded every 60
s with a resolution of 4 cm−1 and analyzed by OPUS software.
3. RESULTS AND DISCUSSION
3.1. Effect of Catalysts in the NTP-Catalysis. CO2
hydrogenation over the two supported Ru catalysts under
NTP conditions was investigated in reference to the control
experiments (i.e., the empty tube for NTP-alone experiments
and the reactor with the bare γ-Al2O3 and SiO2 support
packing under the NTP conditions) to screen the candidate for
the following study (as shown in Figure 1). Under the NTP
conditions without a catalyst, CO2 was decomposed to CO
with a trivial conversion of ∼6% at 6.5 kV (with the specific
input energy (SIE) of 2.0 J mL−1). Similarly, NTP systems
Under NTP conditions, the effect of dielectric property of
the bare supports on the catalysis was deemed insignificant
since γ-Al2O3 and SiO2 have different dielectric constants
(∼9.1 and ∼4.2, respectively), while the reaction results were
similar.18 In addition to the dielectric constant, the porous
property of the packing material can also influence the plasma
discharge and reaction performance under the NTP con-
ditions.19 N2 physisorption analysis showed that the Ru/SiO2
catalyst and Ru/γ-Al2O3 catalyst had comparable pore volumes
12830
https://dx.doi.org/10.1021/acscatal.0c03620
ACS Catal. 2020, 10, 12828−12840

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 2. Dependence of the reaction rate on pH and pCO under (a, c) thermal conditions (at 330 °C) and (b, d) NTP conditions.
2
2
of ∼0.7 cm3 g−1. The average pore sizes of the Ru/SiO2
catalyst and Ru/γ-Al2O3 catalyst were ∼5 and ∼12 nm,
respectively, which are much smaller than the Debye length,
suggesting that the penetration of plasma into the catalyst
pores might be limited. However, a previous study based on
Monte Carlo calculation revealed that microdischarges might
be formed near the pores of mesoporous catalysts with
mesopore sizes of 2−50 nm, and the relatively high surface
area promoted the intensified surface discharge on the
surface.20 This may explain the better catalytic performance
of Ru/SiO2 in CO2 hydrogenation than Ru/γ-Al2O3 since the
high surface area might promote the surface discharge in NTP-
catalysis. The Ru/SiO2 catalyst has a well-developed micro/
mesoporous structure with a higher specific BET surface area
of 557 m2 g−1 than that of the Ru/γ-Al2O3 catalyst (239 m2
g−1). Thus, a high surface area is expected as the key to
determine the catalytic performance of the supported Ru
catalysts under NTP and thermal conditions. The calculated
apparent activation energy (Figure S6) showed that the Ru/
SiO2 catalyst presented lower values, under both conditions,
than the Ru/γ-Al2O3 catalyst (Table S4), e.g., 20 kJ mol−1
versus 71 kJ mol−1 in NTP-catalysis (details of the kinetic
calculations are presented in the Supporting Information).8,9
3.2. Comparative Mechanistic Study of CO2 Hydro-
genation over Ru/SiO2. Preliminary catalytic assessments
have shown that the Ru/SiO2 catalyst presented relatively high
CO2 conversion and CH4 yield for CO2 hydrogenation under
NTP and thermal conditions (in comparison with the Ru/γ-
Al2O3 catalyst); thus, the Ru/SiO2 catalyst was selected for
further investigation. To gain insight into the mechanism of
CO poisoning in CO2 conversions, first, the comparatively
mechanistic study of CO2 hydrogenation over the Ru/SiO2
catalyst was performed. Figure 2 and Table 1 show correlation
Table 1. Reaction Order with Respect to pH and pCO for
Catalytic CO2 Hydrogenation over Ru/SiO2 under the
Thermal (at 330 °C) and NTP Conditions
2
2
NTP
reaction order
6.0 kV
1.40
6.5 kV
1.60
7.0 kV
1.50
thermal
1.0
pH
2
pCO
0.30
0.30
0.25
−0.03
2
between the apparent reaction rate and the CO2/H2 partial
pressures (pH and pCO ) in CO2 hydrogenation under the
2
2
thermal and NTP conditions. Under the thermal condition at
330 °C, the CH4 formation rate over the Ru/SiO2 catalyst
showed a stronger dependence on pH than pCO in the feed.
2
2
Specifically, the reaction order with respect to pH was
2
calculated as 1.0, in line with the Langmuir−Hinshelwood
mechanism.21 A previous study showed that H2 dissociation on
the Ru surface was fast with the produced Had being short-
lived,22 and the reaction order regarding pH indicated that
2
CO2 and H2 were adsorbed on the different active sites on the
Ru surface.23,24 The reaction order regarding pCO was found to
2
be −0.03, which can be approximated as zero order, suggesting
that CO2 concentration has a relatively weak influence on the
formation rate of CH4. This finding suggested (i) the CO2
chemisorption on the catalyst and (ii) the saturation of
relevant active sites on the Ru surface by CO2 molecules at
relatively low CO2 concentrations.22 Therefore, under the
thermal condition, CO2 participated in the reaction via the
Langmuir−Hinshelwood mechanism, i.e., CO2 adsorbed on the
12831
https://dx.doi.org/10.1021/acscatal.0c03620
ACS Catal. 2020, 10, 12828−12840

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 3. In situ DRIFTS spectra of surface species for CO2 hydrogenation over the Ru/SiO2 catalyst under (a) the NTP-off condition with the
feed gas of 3% CO2 + 9% H2 + Ar, (b) NTP-on condition with the feed gas (at 5.5 kV and 27.0 kHz), and (c) NTP-off condition with the feed gas.
(d) Relative intensities of surface species as a function of time-on-stream recorded by in situ DRIFTS from (b) and relative intensity change of
methane recorded in MS (Figure S9) during CO2 hydrogenation by NTP activation (at 5.5 kV and 27.0 kHz).
catalyst, and then dissociated to active intermediates under
heating, which further react with Had to form methane.
In comparison, under the NTP conditions, the reaction
orders with respect to pH and pCO were 1.50 0.10 and 0.30
feed to inert Ar, the COad band decreased slowly (within 10
min, Figure S7b), whereas the intensity of methane decreased
rapidly (Figure S7c), suggesting that the gradual decrease of
COad band was only due to the desorption under the condition
used. Conversely, when the feed was changed to H2, the
intensity of COad band declined fast (within 2 min) with the
associated rapid emergence of peak in the CH4 signal, which
gradually decreased after 2 min. This phenomenon confirms
COad as the active intermediate, which further reacted with H2
to produce CH4, being in line with the kinetic data discussed
above (Figure 2). Under thermal conditions, catalytic CO2
hydrogenation is commonly thought to proceed via a direct
carbon−oxygen bond dissociation mechanism,26,27 which
involves the dissociation of CO2 on the catalyst surface (to
adsorbed COad and surface C) and the subsequent hydro-
genation of surface C. Under the thermal condition used in this
work, DRIFTS only probed COad species on the Ru surface,
confirming the direct carbon−oxygen bond dissociation
mechanism.
Under the NTP condition with a CO2/Ar mixture, the gas-
phase CO2 dissociation was confirmed, as shown in Figure S8.
With the NTP on (Figure S8b), the linearly and bridged
adsorbed COad species (at ∼2092, 2041, and 1881 cm−1) and
bidentate and monodentate carbonate (at ∼1274 and ∼1311
cm−1, respectively) were measured. Since carbonate species
were not observed under the thermal condition, the presence
of carbonate species under the NTP condition was due to the
plasma excitation and could be ascribed to the adsorption of
vibrationally excited CO2 species on the catalyst surface.
2
2
0.05, respectively. Additionally, both reaction orders
remained almost constant as a function of the input power,
suggesting the same surface reaction mechanism at different
input powers. The comparatively strong dependence on pH
2
and pCO under NTP conditions (compared with the thermal
2
condition) indicates the presence of multiple reaction
pathways for CO2 hydrogenation in NTP-catalysis. In addition
to the surface reactions under the thermal catalysis, the
vibrationally activated and dissociated active species (e.g.,
electronically excited H radical) in the gas-phase reaction
under NTP conditions might also participate in the surface
hydrogenation reactions via the Eley−Rideal mechanism.25 To
clarify the relationship between the reaction order and reaction
mechanism under NTP and thermal conditions, in situ
DRIFTS−MS was performed, and the relevant results were
correlated with the kinetic data.
In situ DRIFTS coupled with MS characterization of CO2
hydrogenation over the Ru/SiO2 catalyst was comparatively
performed under thermal and NTP conditions to investigate
the mechanism of CO2 hydrogenation. Under the thermal
condition at 250 °C (Figure S7), characteristic peaks of surface
hydroxyls (OHad, from ∼3596 to ∼3730 cm−1), CHx species
(CH3,ad, at ∼3015 cm−1), carbonyl (COad, at ∼1997 cm−1),
and surface-adsorbed CH4 (at ∼3047 cm−1) were detected on
the catalyst surface. It was observed that after changing the
12832
https://dx.doi.org/10.1021/acscatal.0c03620
ACS Catal. 2020, 10, 12828−12840

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 4. CO2, CO, and carbon conversions as a function of ToS in CO poisoning experiments with different CO/CO2 inlet molar ratios under (a)
thermal condition (at 330 °C) and (b) NTP condition (at 6.5 kV and 21.0 kHz).
Under the NTP-off condition, with the reaction gas feed
(i.e., 3% CO2/9% H2/Ar), in addition to the CO2 gas-phase
peak (at ∼2360 and 2342 cm−1, as shown in Figure 3a),
surface carbon species were not detected by DRIFTS and no
reaction was observed (according to MS). Upon the ignition of
plasma, the MS profile showed the instantaneous appearance
of CH4 signal (Figure S9a), confirming the formation of CH4
over the catalyst under NTP activation. At the same time,
surface formyl species (CHxO, at about 1284, 1270, and 1111
cm−1) and carbonyl species (i.e., linearly adsorbed COad on
Ru0 at 2034 cm−1 and linear form Ruδ+-CO at 2084 cm−1),28 as
shown in Figure 3b, were measured by DRIFTS. Compared
with the COad bands formed under CO2/Ar (Figure S8b), the
two peaks shifted toward lower frequency by about 10 cm−1
and the bridged adsorbed COad peak disappeared, which could
be attributed to the electron donation of Had on the Ru
surface.29 On switching off the plasma, the CH4 concentration
decreased immediately (Figure S9), while the formyl species
decreased slowly (within 4 min), indicating that the system
without plasma discharge was inactive for CH4 formation
(Figure S9b). The gradual decrease of formyl band (within 4
min) reflects its desorption under the NTP-off condition. CO2
hydrogenation can undergo the formyl pathway over Ru-based
catalysts,30,31 which involves the direct CO2 dissociation to
carbonyl (COad) and Oad, followed by the hydrogenation of
COad. The subsequent hydrogenation of COad will form the
formyl species as the intermediates for CH4 production. As
compared with the DRIFTS findings from the thermal system
(i.e., only carbonyl species were observed, Figure S7), the
appearance of carbonyl species and formyl species under NTP
conditions suggested the presence of an alternative reaction
pathway (i.e., formyl pathways) for CO2 hydrogenation under
NTP. Thus, evolution of the surface species as a function of
ToS coupled with the change in CH4 signal intensity (from
MS) was correlated, as presented in Figure 3d. The COad
species increased at a steeper rate than gas CH4, which could
be explained by the plasma-assisted CO2 dissociation in the gas
phase and the dissociation of adsorbed CO2 on the catalyst
surface. The same phenomenon was found between the
formation rates of surface formyl species and CH4, indicating
CHxO species originating from reactions between COad and
Had and as the surface intermediate for CH4 production.30,32
The findings of this work confirmed the presence of the formyl
pathway in CO2 hydrogenation over Ru/SiO2 under NTP
conditions; i.e., CO2 was dissociated to COad and Oad species
on the catalyst surface, then COad was hydrogenated with Had
into formyl intermediate (CHxO) species, and finally, the
formyl group reacted toward CH4 and H2O. In comparison
with the thermally activated CO2 hydrogenation, the vibra-
tionally activated CO2 molecules under NTP conditions could
adsorb on the catalyst surface with lower energy barriers, which
facilitated the formation of COad species.14 This activation
promoted the hydrogenation of CO2 and formation rate of
CH4, leading to the reaction order of pCO increasing slightly.
2
Additionally, the plasma-induced excited/dissociated H
radicals in the gas phase might also interact with the adsorbed
species to form CH4 (i.e., the formyl pathway) via the Eley−
Rideal mechanism in CO2 hydrogenation under NTP.14,25 Due
to the relatively low dissociation energy of H2 molecules (∼4.5
eV),33 the plasma could activate H2 more efficiently, which
produces more H radicals with an increase in H2 concentration
in the feed. Therefore, under NTP, the H2 partial pressure has
a significant influence on the formation rate of CH4, leading to
a much higher reaction order with respect to pH than that in
2
the thermal catalysis.
3.3. Investigation of CO Poisoning on CO2 Hydro-
genation. Catalyst deactivation is complex and significant for
practical catalysis. As expected, under the NTP condition (at
6.5 kV, 21.0 kHz), the Ru/SiO2 catalyst in CO2 hydrogenation
presented excellent stability over 27 h ToS with CO2
conversions and CH4 selectivity maintained at 64.7
0.7%
and 94.1 0.3%, respectively (Figure S10). Comparative TEM
analysis of the catalyst before and after the longevity test
showed no significant change regarding the particle sizes,
neither the sign of metal sintering, which confirmed the
anticoking and antisintering performance of the NTP-catalysis
(Figure S11).
In addition to coking and sintering, catalyst poisoning is
another major factor to deactivate the catalyst. Accordingly, to
understand CO poisoning in the catalysis under the thermal
and plasma conditions, relevant experiments were performed
by varying CO concentration in the gas feed, keeping the H2/C
(i.e., CO2 + CO) inlet molar ratio constant. Figure 4 shows the
thermal and NTP-activated carbon conversions (CO2, CO,
and overall) as a function of ToS with different CO/CO2 ratios
in the feed gas. Under the thermal condition at 330 °C (Figure
4a), the fresh Ru/SiO2 catalyst showed a stable CO2
conversion (∼47%) without CO in the feed gas (CO/CO2 =
12833
https://dx.doi.org/10.1021/acscatal.0c03620
ACS Catal. 2020, 10, 12828−12840

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 5. Long-term deactivation test with the CO2/CO/H2 mixtures, regeneration treatment under Ar, and catalysis in CO2/H2 over the Ru/SiO2
catalyst under (a) the thermal condition (at 330 °C) and (b) NTP condition (at 6.5 kV and 21.0 kHz). Experimental conditions: feed gas
−1
composition of H2/C = 3, CO/CO2 = 0.5, and WHSV of 30,000 mL (STP) gcat h−1.
0, ToS = 0−70 min in Figure 4a). When CO (CO/CO2 =
0.25) was introduced in the gas mixture, CO2 conversion
decreased to 35%, while CO was almost completely consumed,
promoting the overall carbon conversion and CH4 production
due to CO hydrogenation. By switching back to the “CO-free”
feed gas (CO/CO2 = 0, ToS = 165−240 min in Figure 4a),
catalyst deactivation occurred, and the catalyst could not be
fully recovered, as evidenced by the reduced CO2 conversion
(Figure 4a) and CH4 production (Figure S12a), in comparison
with that of the fresh catalyst. By further increasing the CO/
CO2 ratio to 0.5 and 1.0, the decrease in CO2 conversion and
deactivation of catalyst became more significant, while the CO
conversion remained almost complete. The findings showed a
strong inhibiting effect of CO on CO2 conversion. The
condition with CO/CO2 = 0.25 in the feed gas (ToS = 590−
680 min in Figure 4a) was tested again, and carbon conversion
and CH4 production were lower than the previously measured
values, suggesting the permanent deactivation of Ru/SiO2. By
adding more CO in the feed (i.e., CO/CO2 ratio of 2), severe
CO poisoning was measured. Specifically, the CO2 conversion
dropped rapidly below zero, suggesting that CO2 was formed
in the system. This might be attributed to the presence of CO
disproportionation (2CO → C + CO2)34 and/or water-gas
shift reaction (WGSR, CO + H2O → H2 + CO2)35 due to the
excessive CO in the feed and strong adsorption of CO on the
Ru surface. Accordingly, in the thermal catalysis system, a
decrease in CO2 conversion with CO cofeeding was not due to
the kinetic effect, that is, the diluted CO2 concentration in the
gas feed, as discussed above (Figure 2). With the presence of
CO in the feed, competitive adsorption of CO and CO2 on
metal active sites occurs.36 Since the adsorption energy of CO
(−2.3 eV) is much lower than that of CO2 (−0.52 eV),37
preferential adsorption of CO and inhibited CO2 adsorption
on the Ru surface are expected and, thus, a decrease in CO2
conversion. It was proposed that the formation of strongly
adsorbed carbonyl species11,38 due to the presence of CO
might be the dominant factor for catalyst deactivation.
Therefore, the mechanism of CO poisoning of the Ru/SiO2
catalyst was investigated by in situ DRIFTS analysis (to be
discussed later).
0−70 min, in Figure 4b). By introducing CO in the feed (with
CO/CO2 = 0.25), CO2 conversion decreased slightly to ∼49%,
and CO conversion was measured at about 70%, being lower
than that under the thermal condition (which was close to
100%). By increasing the CO concentration in the feed gas
(i.e., CO/CO2 = 0.5, 1.0 and 2.0, respectively), a decrease in
CO2 conversions was measured (due to the change of gas
composition); however, catalyst deactivation was insignificant
since the carbon conversion remained stable in stream during
the cofeeding tests. More importantly, the catalyst activity
regarding CO2 conversion and CH4 production in CO2
hydrogenation can be totally recovered when the system was
switched back to the CO-free feed, regardless of the previous
CO concentration in the feed, confirming that (i) NTP could
be able to completely regenerate the catalyst and (ii) NTP is
able to mitigate CO poisoning on CO2 hydrogenation (Figure
S12b). Interestingly, under NTP conditions, in comparison
with ∼100% CO conversions under the thermal condition, CO
conversions increased from 70 to 92% with an increase in the
inlet CO/CO2 ratio from 0.25 to 2. This suggested that the
reasons for the decrease in CO2 conversion in both systems
may be different. As discussed above, in thermal catalysis,
preferred CO adsorption on the Ru surface and the subsequent
CO hydrogenation prevailed, causing the reduction of the CO2
conversion and almost 100% CO conversion. Conversely,
under NTP conditions, the plasma could activate the CO2
molecules in the gas phase and the vibrationally excited CO2
could adsorb on the catalyst surface with lower energy barriers,
which facilitated the adsorption of CO2 on the catalyst
surface.14 Additionally, the collision of reactive plasma species
(such as the vibrationally excited CO2 and the excited state of
CO, H, OH, and CH in the gas phase according to OES and
FTIR9,25,39) might help remove the strongly adsorbed surface
COad and then release the active sites for adsorption.14,40,41
Thus, NTP alleviated CO adsorption and facilitated CO2
adsorption, which result in lower CO conversions and higher
CO2 conversions than those in the thermal catalysis.
The superiority of the NTP-catalysis over the thermal
counterpart, regarding the maintenance and regeneration of
the catalyst activity, was proved by the long-term CO
poisoning study in Figure 5. At 330 °C, the deterioration of
the catalyst performance with the presence of CO in the feed
was evident during the 9 h test. Specifically, CO2 and CO
As shown in Figure 4b, under the NTP condition at 13.0 kV,
54% CO2 conversion over the Ru/SiO2 catalyst was measured
with the absence of CO in the feed (i.e., CO/CO2 = 0, ToS =
12834
https://dx.doi.org/10.1021/acscatal.0c03620
ACS Catal. 2020, 10, 12828−12840

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 6. In situ DRIFTS spectra of surface species collected at 250 °C in the thermally activated CO hydrogenation over the Ru/SiO2 catalyst. (a)
Initial feed composition: 3% CO + 9% H2+ Ar; (b) change to inert Ar; (c) variations of COad intensity from in situ DRIFTS and CH4 intensity
from MS after switching to Ar at 250 °C; (d) change back to the feed: 3% CO + 9% H2 + Ar; (e) change to H2/Ar; (f) variations of COad intensity
from in situ DRIFTS and CH4 intensity from MS after switching the feed to H2/Ar at 250 °C.
conversions (Figure 5a) and CH4 formation (Figure S13a)
dropped by about 42, 7, and 15%, respectively. By removing
CO from the feed (ToS = 640−790 min in Figure 5a), CO2
conversion and CH4 production over Ru/SiO2 were recovered
to ∼93 and ∼87% only. Considering that flowing inert gases at
high temperature could be used to recover the catalyst
reactivity, the deactivated catalyst was regenerated in situ at
330 °C by sweeping with Ar for 3.5 h, trying to remove the
strongly adsorbed surface species from the catalyst surface.
However, as shown in Figure 5a and Figure S13a, the
deactivation of catalyst due to CO poisoning under the thermal
condition was permanent. Previous theoretical and exper-
imental studies42,43 suggested that the CO molecule could
block the active sites for CO2 and H2 adsorption, thus
decreasing the dissociated Had on the Ru surface and
consequently leading to the deposition of surface carbon
species and metal sintering. Conversely, in NTP-catalysis (at
6.5 kV), the catalyst presented stable performance over 9 h,
with the constant CO2 conversion at about 38% and decreased
CO conversion (by about 6%). Furthermore, after returning to
the CO-free feed (ToS = 640−790 min, Figure 5b), the CO2
conversion was recovered to ∼53% slowly (being comparable
with that of the fresh catalyst at ToS = 0−100 min). During
the same period (ToS = 640−790 min in Figure S13b), the
corresponding CH4 production decreased to the initial level,
confirming that NTP could recover the performance of the
catalyst. The recovery trend of CO2 conversion and CH4
production was attributed to the consumption of residual
adsorbed carbon species under NTP, thus regenerating active
sites available for CO2 hydrogenation. The catalyst was further
treated in situ under Ar and NTP (at 4.0 kV) for 30 min (ToS
= 790−820 min, as shown in Figure 5b). After that, NTP-
activated CO2 hydrogenation was performed again with the
CO-free feed (ToS = 829−945 min in Figure 5b), and the
NTP-catalysis system showed the fully recovered performance.
The corresponding TEM analysis of the catalysts after the
long-term deactivation test (Figure S14) showed the metal
sintering of the catalyst in thermal catalysis; that is, the Ru
particle size increased from ∼1.6 to ∼3.1 nm after the thermal
catalysis. Conversely, the Ru particle size showed no significant
changes after the NTP-catalysis, confirming the antisintering
ability of the hybrid system.
3.4. Mechanisms of CO Poisoning. To understand CO
poisoning in the catalysis, comparative in situ DRIFTS−MS
studies were carried out and compared with the DRIFTS study
of CO2 hydrogenation in Figure 3 and Figure S7. Under the
thermal condition, the DRIFTS spectra measured with CO/H2
mixture (Figure 6a,c) showed that the intensity of the carbonyl
bands was significantly enhanced compared with the case of
CO2 hydrogenation (Figure S7), suggesting the relatively
strong CO binding with the Ru surface. Specifically, in addition
to the gas-phase CO band (at ∼2143 cm−1), the broad
carbonyl bands in a range of 2140−1770 cm−1 can be
deconvoluted into three kinds of COad bonds, i.e., the bands at
1775 and 1950−1980 cm−1 (for the bridged carbonyls), 2005
cm−1 (for the linearly adsorbed CO with monobinding
configuration), and 2030−2075 cm−1 (for the linearly
adsorbed CO with multiple-binding configuration).30 After
changing the feed to inert Ar (Figure 6b,c), the COad bands
decreased much slower than that in CO2 hydrogenation
(Figure S7b), indicating that more strongly adsorbed carbonyl
species formed on the surface when CO was in the feed. By
switching the feed to H2 (Figure 6e,f), surface carbonyl species
disappeared within 10 min, and the CH4 concentration at the
outlet of the DRIFTS cell showed a maximum (at ∼1.3 min,
which was followed by a continuous decline until zero),
12835
https://dx.doi.org/10.1021/acscatal.0c03620
ACS Catal. 2020, 10, 12828−12840

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 7. (a) Corresponding MS signals collected simultaneously from the DRIFTS cell as a function of time during the NTP-assisted CO
hydrogenation over the Ru/SiO2 catalyst. In situ DRIFTS spectra of surface species for CO hydrogenation over the Ru/SiO2 catalyst under (b) the
NTP-off condition with the feed gas of 3% CO + 9% H2 + Ar, (c, d) NTP-on condition with the feed gas (at 5.5 kV and 27.0 kHz), and (e) NTP-
off condition with the feed gas. (f) Relative intensities of surface species as a function of ToS recorded in the in situ DRIFTS from (c) and (d)
during CO hydrogenation under NTP (at 5.5 kV and 27.0 kHz).
showing that the adsorbed CO was converted to CH4 in the
presence of H2. The evolution of the respective surface species
as a function of time (Figure 6f) showed that the intensity of
the carbonyl group at 2030−2050 cm−1 quickly decayed
(within 2 min) under H2, being the most reactive surface
species, while the bridged carbonyl at 1775 cm−1 and linear
monocarbonyl at 2005 cm−1 disappeared completely with
comparatively slow rates. In contrast, the intensity of the peak
at 1950−1980 cm−1, corresponding to geminal dicarbonyls
adsorbed on the low coordination Ru sites, remained constant
within the initial 2 min and then decreased slowly, being
relatively stable on the Ru surface and less reactive for
hydrogenation.44 The presence of these stable and less reactive
surface species might block the active sites and hence
contributed to the catalyst deactivation. Based on the findings
from in situ DRIFTS−MS, one can conclude that, under the
thermal condition, CO hydrogenation proceeded with similar
pathways to those of the catalytic CO2 hydrogenation.45
However, the strong adsorption of CO on the catalyst surface
could saturate the active sites, inhibiting CO2 and H2
adsorption.
With the CO2 + CO + H2 feed under the thermal condition,
the associated DRIFTS spectra showed the combined features
of the CO2-/CO-alone hydrogenation system (as shown in
Figure S15a,c), which was substantiated by the presence of
strongly adsorbed COad and CxHy species on the surface. After
the introduction of CO into the feed, CO coverage increased
significantly and could not be completely removed by Ar
sweeping (Figure S15d), indicating that the strongly adsorbed
COad occupied the active sites for CO2 and H2 adsorption.
Accordingly, based on the findings obtained from the thermal
catalysis and relevant in situ DRIFTS characterization, it was
plausible that the presence of CO in the system produced
strongly adsorbed CO species on the Ru sites, which inhibited
both CO2 and H2 adsorption, thus suppressing CO2 hydro-
genation. Due to the limited concentration of surface Had
species, the relatively stable and inactive carbon-containing
species, such as carbonyl deposition, were encouraged to be
formed on the catalyst surface, and they might progressively
block the active sites. Thus, the associated carbonaceous
species deposition and metal sintering lead to the permanent
catalyst deactivation,46 which confirms the results in Figures 4a
and 5a.
Without plasma discharge at RT, the Ru/SiO2 catalyst
showed no activity for CO hydrogenation, that is, (i) no CO
conversion by MS as shown in Figure 7a and (ii) the only
presence of gas-phase CO (at ∼2143 cm−1) according to
DRIFTS (Figure 7b). Upon the ignition of NTP, the MS
profiles (Figure 7a) showed the instant decrease of CO signal
and simultaneous increase of CO2 and CH4 signals, confirming
the production of CH4 and CO2 in the NTP-catalysis. CO2
formation was due to WGSR, which could be activated by
NTP.5 Water was the product from the NTP-activated
catalytic CO hydrogenation. As discussed above (Figure 4b),
when CO was introduced into the feed for NTP-activated CO2
hydrogenation, a decrease in CO2 conversion was measured,
which might be partly caused by the water-gas shift reaction.
Simultaneously, the gas-phase CO2 peak at about 2350 cm−1
was measured by DRIFTS (Figure 7c), in line with the
intensity change from MS. In the OCO region (Figure 7c,d),
the peak for gas-phase CO at 2143 cm−1 disappeared, and the
continuous development of the IR bands at 2095 and 2160
cm−1 could be attributed to the linearly adsorbed carbonyl
species on Ruδ+ with Ruδ+-CO and Ruδ+-(CO)n configurations,
12836
https://dx.doi.org/10.1021/acscatal.0c03620
ACS Catal. 2020, 10, 12828−12840

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 8. In situ DRIFTS spectra of surface species for hydrogenation of CO2/CO over the Ru/SiO2 catalyst under (a, b) the NTP-on condition
with the feed gas of 1.5% CO2 + 1.5% CO + 9% H2 + Ar (at 5.5 kV and 27.0 kHz), (c) relative intensities of surface species as a function of ToS
recorded in the in situ DRIFTS from (a) and (b), and (d) NTP-off condition with the feed gas.
respectively. Another characteristic peak at 2040 cm−1 was
assigned to the CO linearly adsorbed on Ru0, while the
gradually increased peaks at about 1272 and 1306 cm−1
corresponded to formyl species (CHxO). The evolution of
the surface carbon species recorded by DRIFTS as a function
of time is correlated (Figure 7f). The increasing rate of
carbonyl bands at 2095 and 2160 cm−1 was comparable with
that of formyl species, which supported the fact that CO is the
intermediate toward formyl species. Also, the formation rate of
COad band at 2040 cm−1 increased relatively fast, which might
be due to CO2 dissociation (formed by water-gas shift
reaction) and CO adsorption. By switching off NTP, the
peak of the gas-phase CO2 decreased quickly, and the system
was not active again for CO hydrogenation, which was in good
agreement with the MS profile. Regarding the carbon species,
the formyl species disappeared gradually due to desorption
after the extinction of plasma (Figure S16), while COad band
intensity barely changed, indicating the strong interaction
between the COad species and catalyst surface. Furthermore,
when the feed was switched to H2/Ar (from 3% CO + 9% H2
+ Ar), DRIFTS characterization (Figure S17a) showed that the
surface COad species at 2090 cm−1 (due to gas-phase CO
adsorption) decreased immediately, while the intensity of
formyl species increased initially and then decreased slowly.
The initial increase of the formyl species on the catalyst surface
could be ascribed to the reaction between COad and Had (to
form the formyl), while the subsequent decrease of the formyl
species was due to the consumption of residual formyl species
to form CH4. As shown in Figure S17b, the rate of decrease of
formyl species and CH4 concentration (at the outlet of the
DRIFTS cell by MS) was similar, confirming that the formyl
species originated from the reaction between COad and Had
and were the active intermediate for CH4 formation. DRIFTS
analysis of CO hydrogenation under the NTP condition
showed that CO hydrogenation to CH4 proceeded via CO
adsorption and the formyl pathway, being similar with that of
CO2 hydrogenation (Figure 3).
NTP-catalysis with the CO2/CO/H2 feed was examined by
DRIFTS−MS (Figure 8 and Figure S18). Being different from
CO hydrogenation, the COad peak at 2080 cm−1, due to CO2
dissociation, appeared first and then combined with the peak at
2097 cm−1 (originating from the gas-phase CO adsorption).
Accordingly, the evolution profile of the surface carbon species
(Figure 8c) showed that the COad species had higher
increasing rates than that of formyl species initially (within 8
min) due to CO2 dissociation and CO adsorption on the Ru
sites. The subsequent change in the increasing rate was due to
saturation of relevant active sites on the Ru surface by CO2/
CO adsorption.30 This finding suggested that CO2 and CO
coadsorption existed in the NTP-catalysis. In contrast, the
formyl species presented a constant formation rate, confirming
that the formyl species originated from the reaction between
COad and Had. In addition, by switching CO feed on and off
alternatively, DRIFTS−MS characterization of the catalysis
12837
https://dx.doi.org/10.1021/acscatal.0c03620
ACS Catal. 2020, 10, 12828−12840

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
(Figure S19) showed that the CO2 MS signal increased with
CO in the feed (i.e., production of CO2), which confirms the
presence of WGSR under the NTP condition with the CO2/
CO/H2 mixture. Therefore, under NTP conditions, the
presence of CO in the feed affected CO2 conversions, which
was due to (i) the occurrence of WGSR in the system for CO2
formation and (ii) the relatively strong adsorption of CO, in
line with the result in Figure 4b. Based on the in situ DRIFTS
characterization and relevant discussion, the presence of CO in
the feed did not alter the reaction pathways for CO2
hydrogenation under thermal and NTP conditions. However,
in comparison with the CO poisoning under the thermal
conditions (as discussed before, i.e., due to strong CO
adsorption and associated metal sintering of the catalyst), the
collisions between reactive plasma-derived species in NTP
could recover the active sites by removing the adsorbed carbon
species effectively, which lead to the sites available for CO2
adsorption.40,41,47 This is confirmed by the comparison of the
relevant IR spectra (Figure S20), which showed the
comparatively low intensity of the adsorbed COad on the Ru
catalyst under NTP. Therefore, NTP-catalysis promoted the
adsorption of CO2 and alleviated CO adsorption on the
catalyst surface in the presence of CO, mitigating the CO
poisoning effect on the performance of CO2 hydrogenation
and being opposite to that experienced by the thermal
catalysis. More importantly, according to the literature,48−50
H2O molecules will occupy the active sites and present an
inhibiting effect on the CO2 hydrogenation. Conversely, NTP
enabled WGSR of CO with the produced H2O, which shifted
the equilibrium of CO2 hydrogenation toward CH4 produc-
tion. The phenomenon observed in the system under
investigation showed the interesting effect of CO on NTP-
catalytic CO2 hydrogenation, that is, as a reaction promoter
rather than a catalyst poison, due to the copresence of WGSR,
CO2 hydrogenation, and CO hydrogenation under NTP
conditions.
which facilitated CO2 adsorption and, hence, CO2 hydro-
genation. Therefore, NTP-catalysis could alleviate the CO
effect on CO2 hydrogenation and regenerate the catalyst in situ
in the presence of CO during the catalysis. Importantly, the
NTP-induced WGSR of CO with the produced H2O also
promoted the equilibrium shift of CO2 hydrogenation toward
CH4 production. This work demonstrates that, under NTP
conditions, the role played by CO in Ru-catalyzed CO2
hydrogenation is fundamentally different from its positioning
role in thermal catalysis, showing the potential of NTP-
catalysis to address some of the challenges in conventional
heterogeneous catalysis, specifically, the development of
advanced hybrid NTP-catalysis systems to solve the chemical
deactivation issues for practical catalysis.
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acscatal.0c03620.
Detailed characterization of catalysts; relevant catalyst
assessment for catalytic CO2 hydrogenation; kinetic
parameters of the thermal and NTP systems; relevant in
situ DRIFTS data of the thermal and NTP-catalysis
systems (PDF)
AUTHOR INFORMATION
■
Corresponding Authors
Christopher Hardacre − Department of Chemical Engineering
and Analytical Science, School of Engineering, The University of
Manchester, Manchester M13 9PL, United Kingdom;
orcid.org/0000-0001-7256-6765; Email: c.hardacre@
manchester.ac.uk
Xiaolei Fan − Department of Chemical Engineering and
Analytical Science, School of Engineering, The University of
Manchester, Manchester M13 9PL, United Kingdom;
orcid.org/0000-0002-9039-6736; Email: xiaolei.fan@
manchester.ac.uk
4. CONCLUSIONS
In this study, the NTP-catalysis system was demonstrated to be
efficient for CO2 hydrogenations under atmospheric con-
ditions, in which 65% CO2 conversion and 63% CH4 yield can
be achieved over the Ru/SiO2 catalyst. Also, the intrinsic
nature of catalyst such as surface area is crucial under both
thermal and NTP conditions. The comparative kinetic and in
situ DRIFTS−MS study revealed that the NTP-catalysis could
lower the energy barrier required for catalysis and enable both
Langmuir−Hinshelwood and Eley−Rideal mechanisms.
The effect of CO on the catalysis under both thermal and
NTP conditions was investigated to understand CO poisoning
comparatively. In the thermal catalysis, the catalyst suffered
from a significant decrease in CO2 conversion and deactivation
due to CO poisoning, while in the NTP-catalysis, the CO
played a different role in the system, and the catalyst showed
comparatively good stability and regenerability by NTP. In situ
DRIFTS−MS study of the thermal catalysis showed that (i)
CO preferred to adsorb on the Ru surface strongly to inhibit
CO2 and H2 adsorption and decrease CO2 conversion
significantly and (ii) the formation of less reactive and strongly
adsorbed carbon species (e.g., COad) due to CO strong
adsorption and metal sintering deactivates the catalyst
permanently. Conversely, in NTP-catalysis, collisions of
reactive plasma-derived species contributed to the recovery
of the active sites by removing the strongly adsorbed COad,
Authors
Shanshan Xu − Department of Chemical Engineering and
Analytical Science, School of Engineering, The University of
Manchester, Manchester M13 9PL, United Kingdom
Sarayute Chansai − Department of Chemical Engineering and
Analytical Science, School of Engineering, The University of
Manchester, Manchester M13 9PL, United Kingdom
Shaojun Xu − UK Catalysis Hub, Research Complex at Harwell,
Didcot OX11 0FA, United Kingdom; Cardiff Catalysis Institute,
School of Chemistry, Cardiff University, Cardiff CF10 3AT,
United Kingdom; orcid.org/0000-0002-8026-8714
Cristina E. Stere − Department of Chemical Engineering and
Analytical Science, School of Engineering, The University of
Manchester, Manchester M13 9PL, United Kingdom;
orcid.org/0000-0001-8604-0211
Yilai Jiao − Shenyang National Laboratory for Materials Science,
Institute of Metal Research, Chinese Academy of Sciences,
Shenyang 110016, China
Sihai Yang − Department of Chemistry, School of Natural
Science, The University of Manchester, Manchester M13 9PL,
United Kingdom; orcid.org/0000-0002-1111-9272
Complete contact information is available at:
https://pubs.acs.org/10.1021/acscatal.0c03620
12838
https://dx.doi.org/10.1021/acscatal.0c03620
ACS Catal. 2020, 10, 12828−12840

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
(13) Barboun, P.; Mehta, P.; Herrera, F. A.; Go, D. B.; Schneider, W.
F.; Hicks, J. C. Distinguishing Plasma Contributions to Catalyst
Performance in Plasma-Assisted Ammonia Synthesis. ACS Sustainable
Chem. Eng. 2019, 7, 8621−8630.
(14) Whitehead, J. C. Plasma−Catalysis: The Known Knowns, the
Known Unknowns and the Unknown Unknowns. J. Phys. D: Appl.
Phys. 2016, 49, 243001.
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript.
Notes
The authors declare no competing financial interest.
(15) Gibson, E. K.; Stere, C. E.; Curran-McAteer, B.; Jones, W.;
Cibin, G.; Gianolio, D.; Goguet, A.; Wells, P. P.; Catlow, C. R. A.;
Collier, P.; Hinde, P.; Hardacre, C. Probing the Role of a Non-
Thermal Plasma (NTP) in the Hybrid NTP Catalytic Oxidation of
Methane. Angew. Chem., Int. Ed. 2017, 56, 9351−9355.
(16) Mei, D.; Zhu, X.; He, Y.-L.; Yan, J. D.; Tu, X. Plasma-Assisted
Conversion of CO2 in a Dielectric Barrier Discharge Reactor:
Understanding the Effect of Packing Materials. Plasma Sources Sci.
Technol. 2014, 24, No. 015011.
(17) Mehta, P.; Barboun, P.; Go, D. B.; Hicks, J. C.; Schneider, W. F.
Catalysis Enabled by Plasma Activation of Strong Chemical Bonds: A
Review. ACS Energy Lett. 2019, 4, 1115−1133.
(18) Zhang, Y.; Wang, H.-y.; Jiang, W.; Bogaerts, A. Two-
Dimensional Particle-in Cell/Monte Carlo Simulations of a Packed-
Bed Dielectric Barrier Discharge in Air at Atmospheric Pressure. New
J. Phys. 2015, 17, No. 083056.
ACKNOWLEDGMENTS
■
S.X. thanks the financial support from the Dean’s Doctoral
Scholar Awards from the University of Manchester. The UK
Catalysis Hub is kindly thanked for resources and support
provided via our membership of the UK Catalysis Hub
Consortium and funded by EPSRC grant EP/R026939/1, EP/
R026815/1, EP/R026645/1, EP/R027129/1, or EP/
M013219/1(biocatalysis).
REFERENCES
■
(1) Whitehead, J. C. Plasma-Catalysis: Is It Just a Question of Scale?
Front. Chem. Sci. Eng. 2019, 13, 264−273.
(2) Neyts, E. C.; Ostrikov, K. K.; Sunkara, M. K.; Bogaerts, A.
Plasma Catalysis: Synergistic Effects at the Nanoscale. Chem. Rev.
2015, 115, 13408−13446.
(3) Bogaerts, A.; Tu, X.; Whitehead, J. C.; Centi, G.; Lefferts, L.;
Guaitella, O.; Azzolina-Jury, F.; Kim, H.-H.; Murphy, A. B.;
Schneider, W. F.; Nozaki, T.; Hicks, J. C.; Rousseau, A.; Thevenet,
F.; Khacef, A.; Carreon, M. The 2020 Plasma Catalysis Roadmap. J.
Phys. D: Appl. Phys. 2020, 53, 443001.
(4) Vakili, R.; Gholami, R.; Stere, C. E.; Chansai, S.; Chen, H.;
Holmes, S. M.; Jiao, Y.; Hardacre, C.; Fan, X. Plasma-Assisted
Catalytic Dry Reforming of Methane (DRM) over Metal-Organic
Frameworks (MOFs)-Based Catalysts. Appl. Catal., B 2020, 260,
118195.
(5) Stere, C. E.; Anderson, J. A.; Chansai, S.; Delgado, J. J.; Goguet,
A.; Graham, W. G.; Hardacre, C.; Taylor, S. F. R.; Tu, X.; Wang, Z.;
Yang, H. Non-Thermal Plasma Activation of Gold-Based Catalysts for
Low-Temperature Water-Gas Shift Catalysis. Angew. Chem., Int. Ed.
2017, 56, 5579−5583.
(6) Xu, S.; Chansai, S.; Stere, C.; Inceesungvorn, B.; Goguet, A.;
Wangkawong, K.; Taylor, S. F. R.; Al-Janabi, N.; Hardacre, C.; Martin,
P. A.; Fan, X. Sustaining Metal−Organic Frameworks for Water-Gas
Shift Catalysis by Non-Thermal Plasma. Nat. Catal. 2019, 2, 142−
148.
(7) Chen, H.; Mu, Y.; Shao, Y.; Chansai, S.; Xu, S.; Stere, C. E.;
Xiang, H.; Zhang, R.; Jiao, Y.; Hardacre, C.; Fan, X. Coupling Non-
Thermal Plasma with Ni Catalysts Supported on Beta Zeolite for
Catalytic CO2 Methanation. Catal. Sci. Technol. 2019, 9, 4135−4145.
(8) Kim, J.; Go, D. B.; Hicks, J. C. Synergistic Effects of Plasma-
Catalyst Interactions for CH4 Activation. Phys. Chem. Chem. Phys.
2017, 19, 13010−13021.
(9) Xu, S.; Chansai, S.; Shao, Y.; Xu, S.; Wang, Y.-c.; Haigh, S.; Mu,
Y.; Jiao, Y.; Stere, C. E.; Chen, H.; Fan, X.; Hardacre, C. Mechanistic
Study of Non-Thermal Plasma Assisted CO2 Hydrogenation over Ru
Supported on MgAl Layered Double Hydroxide. Appl. Catal., B 2020,
268, 118752.
(19) Zhang, K.; Zhang, G.; Liu, X.; Phan, A. N.; Luo, K. A Study on
CO2 Decomposition to CO and O2 by the Combination of Catalysis
and Dielectric-Barrier Discharges at Low Temperatures and Ambient
Pressure. Ind. Eng. Chem. Res. 2017, 56, 3204−3216.
(20) Zhang, Y.; Wang, H.-y.; Zhang, Y.-r.; Bogaerts, A. Formation of
Microdischarges inside a Mesoporous Catalyst in Dielectric Barrier
Discharge Plasmas. Plasma Sources Sci. Technol. 2017, 26, No. 054002.
(21) Weatherbee, G. D.; Bartholomew, C. H. Hydrogenation of CO2
on Group VIII Metals: II. Kinetics and Mechanism of CO2
Hydrogenation on Nickel. J. Catal. 1982, 77, 460−472.
(22) Prairie, M. R.; Renken, A.; Highfield, J. G.; Thampi, K. R.;
̈
Gratzel, M. A Fourier Transform Infrared Spectroscopic Study of CO2
Methanation on Supported Ruthenium. J. Catal. 1991, 129, 130−144.
̈
(23) Solymosi, F.; Erdohelyi, A.; Kocsis, M. Methanation of CO2 on
Supported Ru Catalysts. J. Chem. Soc., Faraday Trans. 1 1981, 77,
1003−1012.
́
(24) Kusmierz, M. Kinetic Study on Carbon Dioxide Hydrogenation
over Ru/γ-Al2O3 Catalysts. Catal. Today 2008, 137, 429−432.
(25) Azzolina-Jury, F.; Thibault-Starzyk, F. Mechanism of Low
Pressure Plasma-Assisted CO2 Hydrogenation over Ni-USY by
Microsecond Time-Resolved FTIR Spectroscopy. Top. Catal. 2017,
60, 1709−1721.
(26) Miao, B.; Ma, S. S. K.; Wang, X.; Su, H.; Chan, S. H. Catalysis
Mechanisms of CO2 and CO Methanation. Catal. Sci. Technol. 2016,
6, 4048−4058.
̈
(27) Lalinde, J. A. H.; Roongruangsree, P.; Ilsemann, J.; Baumer, M.;
Kopyscinski, J. CO2 Methanation and Reverse Water Gas Shift
Reaction. Kinetic Study Based on in situ Spatially-Resolved
Measurements. Chem. Eng. J. 2020, 124629.
(28) Garbarino, G.; Bellotti, D.; Finocchio, E.; Magistri, L.; Busca,
G. Methanation of Carbon Dioxide on Ru/Al2O3: Catalytic Activity
and Infrared Study. Catal. Today 2016, 277, 21−28.
(29) Fisher, I. A.; Bell, A. T. A Comparative Study of CO and CO2
Hydrogenation over Rh/SiO2. J. Catal. 1996, 162, 54−65.
(30) Eckle, S.; Anfang, H.-G.; Behm, R. J. r. Reaction Intermediates
and Side Products in the Methanation of CO and CO2 over
Supported Ru Catalysts in H2-Rich Reformate Gases. J. Phys. Chem. C
2011, 115, 1361−1367.
(10) Chen, H.; Goodarzi, F.; Mu, Y.; Chansai, S.; Mielby, J. J.; Mao,
B.; Sooknoi, T.; Hardacre, C.; Kegnæs, S.; Fan, X. Effect of Metal
Dispersion and Support Structure of Ni/Silicalite-1 Catalysts on Non-
Thermal Plasma (NTP) Activated CO2 Hydrogenation. Appl. Catal.,
B 2020, 272, 119013.
(11) Gupta, N. M.; Londhe, V. P.; Kamble, V. S. Gas-Uptake,
Methanation, and Microcalorimetric Measurements on the Coad-
sorption of CO and H2 over Polycrystalline Ru and a Ru/TiO2
Catalyst. J. Catal. 1997, 169, 423−437.
́
(31) Navarro-Jaen, S.; Navarro, J. C.; Bobadilla, L. F.; Centeno, M.
A.; Laguna, O. H.; Odriozola, J. A. Size-Tailored Ru Nanoparticles
Deposited over γ-Al2O3 for the CO2 Methanation Reaction. Appl.
Surf. Sci. 2019, 483, 750−761.
(32) Fajín, J. L. C.; Gomes, J. R. B.; Cordeiro, M. N. D. S.
Mechanistic Study of Carbon Monoxide Methanation over Pure and
Rhodium- or Ruthenium-Doped Nickel Catalysts. J. Phys. Chem. C
2015, 119, 16537−16551.
(12) Falbo, L.; Visconti, C. G.; Lietti, L.; Szanyi, J. The Effect of CO
on CO2 Methanation over Ru/Al2O3 Catalysts: A Combined Steady-
State Reactivity and Transient DRIFT Spectroscopy Study. Appl.
Catal., B 2019, 256, 117791.
12839
https://dx.doi.org/10.1021/acscatal.0c03620
ACS Catal. 2020, 10, 12828−12840

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
(33) Liu, C.-j.; Xu, G.-h.; Wang, T. Non-Thermal Plasma
Approaches in CO2 Utilization. Fuel Process. Technol. 1999, 58,
119−134.
(34) Gupta, N. M.; Kamble, V. S.; Rao, K. A.; Iyer, R. M. On the
Mechanism of CO and CO2 Methanation over Ru/Molecular-Sieve
Catalyst. J. Catal. 1979, 60, 57−67.
(35) Utaka, T.; Okanishi, T.; Takeguchi, T.; Kikuchi, R.; Eguchi, K.
Water Gas Shift Reaction of Reformed Fuel over Supported Ru
Catalysts. Appl. Catal., A 2003, 245, 343−351.
(36) Gupta, N. M.; Kamble, V. S.; Iyer, R. M.; Thampi, K. R.;
Gratzel, M. FTIR Studies on the CO, CO2 and H2 Co-Adsorption
over Ru-RuOx/TiO2 Catalyst. Catal. Lett. 1993, 21, 245−255.
(37) Zhang, S.-T.; Yan, H.; Wei, M.; Evans, D. G.; Duan, X.
Hydrogenation Mechanism of Carbon Dioxide and Carbon Monoxide
on Ru(0001) Surface: A Density Functional Theory Study. RSC Adv.
2014, 4, 30241−30249.
(38) Mukkavilli, S.; Wittmann, C.; Tavlarides, L. L. Carbon
Deactivation of Fischer-Tropsch Ruthenium Catalyst. Ind. Eng.
Chem. Process Des. Dev. 1986, 25, 487−494.
(39) Mikhail, M.; Wang, B.; Jalain, R.; Cavadias, S.; Tatoulian, M.;
́
Ognier, S.; Galvez, M. E.; Da Costa, P. Plasma-Catalytic Hybrid
Process for CO2 Methanation: Optimization of Operation Parame-
ters. React. Kinet., Mech. Catal. 2018, 126, 629−643.
(40) Wang, L.; Zhao, Y.; Liu, C.; Gong, W.; Guo, H. Plasma Driven
Ammonia Decomposition on a Fe-Catalyst: Eliminating Surface
Nitrogen Poisoning. Chem. Commun. 2013, 49, 3787−3789.
(41) Zhu, B.; Li, X.-S.; Liu, J.-L.; Liu, J.-B.; Zhu, X.; Zhu, A.-M. In-
Situ Regeneration of Au Nanocatalysts by Atmospheric-Pressure Air
Plasma: Significant Contribution of Water Vapor. Appl. Catal., B
2015, 179, 69−77.
(42) Zhao, P.; He, Y.; Cao, D.-B.; Wen, X.; Xiang, H.; Li, Y. W.;
Wang, J.; Jiao, H. High Coverage Adsorption and Co-Adsorption of
CO and H2 on Ru(0001) from DFT and Thermodynamics. Phys.
Chem. Chem. Phys. 2015, 17, 19446−19456.
(43) Diemant, T.; Rauscher, H.; Bansmann, J.; Behm, R. J.
Coadsorption of Hydrogen and CO on Well-Defined Pt35Ru65/
Ru(0001) Surface Alloys-Site Specificity Vs. Adsorbate-Adsorbate
Interactions. Phys. Chem. Chem. Phys. 2010, 12, 9801−9810.
(44) Wang, X.; Hong, Y.; Shi, H.; Szanyi, J. Kinetic Modeling and
Transient DRIFTS−MS Studies of CO2 Methanation over Ru/Al2O3
Catalysts. J. Catal. 2016, 343, 185−195.
(45) Cant, N. W.; Bell, A. T. Studies of Carbon Monoxide
Hydrogenation over Ruthenium Using Transient Response Techni-
ques. J. Catal. 1982, 73, 257−271.
(46) Carballo, J. M. G.; Finocchio, E.; García-Rodriguez, S.; Ojeda,
M.; Fierro, J. L. G.; Busca, G.; Rojas, S. Insights into the Deactivation
and Reactivation of Ru/TiO2 During Fischer−Tropsch Synthesis.
Catal. Today 2013, 214, 2−11.
(47) Saoud, W. A.; Assadi, A. A.; Guiza, M.; Bouzaza, A.;
Aboussaoud, W.; Ouederni, A.; Soutrel, I.; Wolbert, D.; Rtimi, S.
Study of Synergetic Effect, Catalytic Poisoning and Regeneration
Using Dielectric Barrier Discharge and Photocatalysis in a Continuous
Reactor: Abatement of Pollutants in Air Mixture System. Appl. Catal.,
B 2017, 213, 53−61.
(48) Aziz, M. A. A.; Jalil, A. A.; Triwahyono, S.; Saad, M. W. A. CO2
Methanation over Ni-Promoted Mesostructured Silica Nanoparticles:
Influence of Ni Loading and Water Vapor on Activity and Response
Surface Methodology Studies. Chem. Eng. J. 2015, 260, 757−764.
(49) Falbo, L.; Martinelli, M.; Visconti, C. G.; Lietti, L.; Bassano, C.;
Deiana, P. Kinetics of CO2 Methanation on a Ru-Based Catalyst at
Process Conditions Relevant for Power-to-Gas Applications. Appl.
Catal., B 2018, 225, 354−363.
́
(50) Bacariza, M. C.; Biset-Peiro, M.; Graca̧ , I.; Guilera, J.; Morante,
J.; Lopes, J. M.; Andreu, T.; Henriques, C. DBD Plasma-Assisted CO2
Methanation Using Zeolite-Based Catalysts: Structure Composition-
Reactivity Approach and Effect of Ce as Promoter. J. CO2 Util. 2018,
26, 202−211.
12840
https://dx.doi.org/10.1021/acscatal.0c03620
ACS Catal. 2020, 10, 12828−12840