Influence of the support on rhodium speciation and catalytic activity of rhodium-based catalysts for total oxidation of methane

Article abstract of DOI:10.1039/d0cy00847h

The role of the support in Rh-based catalysts for total oxidation of CH₄was investigated using both amorphous SiO₂and γ-Al₂O₃as well as ZSM-5 and SSZ-13 zeolites with varying SiO₂/Al₂O₃ratios. The methane oxidation activity was measured in both the presence and absence of 5 vol% H₂O and 20 ppm SO₂. The support had a strong impact on Rh speciation (Rh₂O₃nanoparticlesvs.single Rh atom sites), which was important for the activity, as the nanoparticle sites were found to be far more active than the single atom sites. A 2 wt% Rh/ZSM-5(SiO₂/Al₂O₃= 280) catalyst with Rh exclusively as Rh₂O₃nanoparticles was able to provide oxidation at a 75 °C lower temperature than an ion-exchanged 0.294 wt% Rh/ZSM-5(SiO₂/Al₂O₃= 30) catalyst with Rh exclusively as single atoms despite a similar concentration of active sites in the two samples. All the catalysts were inhibited by the presence of water, but this inhibition was particularly strong for an amorphous SiO₂support and for the most Al-rich zeolites and less severe for Si-rich zeolites. The inhibition from SO₂was relatively stronger for the more Al-rich supports, which was attributed to an uptake of sulfur at Al sites that was detrimental to the performance of the active phase. In a realistic gas atmosphere containing both H₂O and SO₂, Si-rich ZSM-5(SiO₂/Al₂O₃= 280) therefore emerged as the best support. This was because the low acidity ensured minimal loss of Rh to the less active single atom sites, avoidance of the detrimental sulfur uptake by the support seen for more Al-rich supports, and avoidance of the strong water-induced inhibition that occurs for amorphous SiO₂and Al-rich zeolites.

Full text of DOI:10.1039/d0cy00847h

Catalysis
Science &
Technology
View Article Online
View Journal
PAPER
Influence of the support on rhodium speciation
and catalytic activity of rhodium-based catalysts
for total oxidation of methane†
a
Cite this: DOI: 10.1039/d0cy00847h
Yu Zhang, a Peter Glarborg,
Martin Peter Andersson, a Keld Johansen,b
Thomas Klint Torp,b Anker Degn Jensen a and Jakob Munkholt Christensen
a
*
The role of the support in Rh-based catalysts for total oxidation of CH4 was investigated using both
amorphous SiO2 and γ-Al2O3 as well as ZSM-5 and SSZ-13 zeolites with varying SiO2/Al2O3 ratios. The
methane oxidation activity was measured in both the presence and absence of 5 vol% H2O and 20 ppm
SO2. The support had a strong impact on Rh speciation (Rh2O3 nanoparticles vs. single Rh atom sites), which
was important for the activity, as the nanoparticle sites were found to be far more active than the single
atom sites. A 2 wt% Rh/ZSM-5(SiO2/Al2O3 = 280) catalyst with Rh exclusively as Rh2O3 nanoparticles was
able to provide oxidation at a 75 °C lower temperature than an ion-exchanged 0.294 wt% Rh/ZSM-5(SiO2/
Al2O3 = 30) catalyst with Rh exclusively as single atoms despite a similar concentration of active sites in the
two samples. All the catalysts were inhibited by the presence of water, but this inhibition was particularly
strong for an amorphous SiO2 support and for the most Al-rich zeolites and less severe for Si-rich zeolites.
The inhibition from SO2 was relatively stronger for the more Al-rich supports, which was attributed to an
uptake of sulfur at Al sites that was detrimental to the performance of the active phase. In a realistic gas
atmosphere containing both H2O and SO2, Si-rich ZSM-5(SiO2/Al2O3 = 280) therefore emerged as the best
support. This was because the low acidity ensured minimal loss of Rh to the less active single atom sites,
avoidance of the detrimental sulfur uptake by the support seen for more Al-rich supports, and avoidance of
the strong water-induced inhibition that occurs for amorphous SiO2 and Al-rich zeolites.
Received 26th April 2020,
Accepted 18th July 2020
DOI: 10.1039/d0cy00847h
rsc.li/catalysis
especially the combined presence of H2O and SO2 are
strongly detrimental to Pd-based catalysts.9–12 Recent work
1. Introduction
Natural gas engines have the advantage of low NOx, SOx and
particulate emissions, which makes them interesting for
applications such as maritime transport, but the slip of CH4
from lean-burn natural gas engines causes environmental
problems due to the high greenhouse gas potential of CH4 (26–
38 times that of CO2).1–3 Mitigation of the CH4 slip by catalytic
oxidation of CH4 in the exhaust gas is therefore necessary.
The exhaust gas from lean-burn natural gas fired engines
is characterized by intermediate temperatures (350–540 °C),
excess of O2 (∼10 vol%) and the presence of both H2O (5–10
vol%) and SO2 (up to 1 ppm).4,5 Complete removal of CH4
under dry and sulfur-free conditions is facile using Pd-based
catalysts, which are acknowledged as the most active catalysts
for CH4 oxidation.6–8 However, the presence of H2O and
found that Rh-based catalysts show better performance for
CH4 oxidation in atmospheres containing both H2O and
SO2.13 The conversion of CH4 over a Rh/ZSM-5(SiO2/Al2O3
=
280) catalyst could be maintained at 58–79% in the presence
of 5 vol% H2O and 1 ppm SO2 at engine achievable exhaust
gas temperatures (475–500 °C) and a high gas hourly space
−1
velocity (GHSV, 150 000 mL gcat h−1).13–15 The benefit of Rh
was attributed to H2O partly alleviating SO2 poisoning by
destabilizing the Rh sulfate hydrates, which allowed a partial
sulfur release at low temperatures.
Our previous work13 identified Rh2O3 nanoparticles to be
the active phase for CH4 oxidation, which follows the Mars–
van Krevelen type mechanism involving structural oxygen in
the rhodium oxide surface.16 In the optimization of the
catalytic activity, which is of particular importance for
catalysts relying on expensive noble metals, the identification
of the optimal support is of central importance. It is
therefore highly relevant to evaluate how the novel Rh-based
oxidation catalysts depend upon the support. Previously,
zeolite-based support materials with varying SiO2/Al2O3 ratios
have shown good performance in Cu catalysts for NH3-SCR as
a Department of Chemical and Biochemical Engineering, Technical University of
Denmark (DTU), Søltofts Plads 229, 2800 Kgs. Lyngby, Denmark.
E-mail: jmc@kt.dtu.dk
b Haldor Topsoe A/S, Haldor Topsøes Allé 1, 2800 Kgs. Lyngby, Denmark
† Electronic supplementary information (ESI) available: XRD of Rh/ZSM-5(30)-IE
and the pure support; activity of Rh-based catalysts for CH4 oxidation;
characterization of spent catalysts. See DOI: 10.1039/d0cy00847h
This journal is © The Royal Society of Chemistry 2020
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
well as in Pd catalysts for CH4 oxidation.17–20 A high
concentration of zeolite exchange sites, obtained using low
SiO2/Al2O3 ratio zeolites, was found to favor the dispersion of
Pd into smaller nanoparticles, thereby improving the initial
activity.6,21 On the other hand, Pd supported on Si-rich
zeolites was reported to be more stable and more sulfur
tolerant.18,22 The optimal acidity of the zeolite could thus be
the one that strikes a balance between dispersion/initial
activity at higher acidity and better stability/weaker
interaction with inhibitors at lower acidity. The level of
zeolite acidity giving the optimal performance for the Rh-
catalyzed total oxidation of methane should therefore be
identified. On a zeolite support, the active phase can take the
form of both continuous phases and single atoms at the
zeolite exchange sites. Among the continuous phases, it can
be difficult to fully distinguish if the amorphous Rh2O3
structure is in the form of nanoparticles or films. Here we
have mainly used the term nanoparticles throughout the text.
Rh on a zeolite support was studied for partial oxidation of
CH4 to acetic acid and methanol, and single Rh atom sites
were correlated to higher selectivity to the partial oxidation
products.23,24 The superiority of Rh catalysts with single atom
sites was also reported for the water gas shift reaction.25,26
Single Rh atom sites will provide the optimal utilization of
the active phase, but their activity for total oxidation is
unknown. To clarify the importance of single atom sites for
total oxidation, it is important to study the influence of the
support on the distribution between single Rh atoms and
nanoparticles. This distribution can be characterized by
diffuse reflectance infrared Fourier transform spectroscopy
(DRIFTS) measurements on chemisorbed CO, since CO binds
differently to Rh single atom sites and nanoparticle sites.27–29
In this work, a variety of ZSM-5 MFI zeolites with Si/Al =
15–140 and an SSZ-13 CHA zeolite with Si/Al = 12 were
studied as support materials for Rh in order to realize
different Rh site distributions and compared to amorphous
SiO2 and γ-Al2O3 supports. Rhodium was deposited by
impregnation or ion-exchange techniques. The site
distribution (i.e. single atom sites vs. nanoparticle sites) was
analyzed by CO-DRIFTS and the presence of Rh2O3
nanoparticles was also evaluated by X-ray diffraction (XRD)
and transmission electron microscopy (TEM). The activity of
the prepared Rh-based catalysts towards CH4 oxidation was
tested in different reaction atmospheres in a fixed-bed plug
flow reactor. The sulfur tolerance of the catalyst was
correlated to the site distribution to identify the optimal Rh
catalyst for CH4 oxidation under realistic engine exhaust gas
conditions where both H2O and SO2 are present.
8014), 25 (SiO2/Al2O3 = 50, CBV 5524G) and 15 (SiO2/Al2O3 =
30, CBV 3014) were obtained from Zeolyst International. SSZ-
13 with a Si/Al ratio of 12 (SiO2/Al2O3 = 24) was supplied by
Haldor Topsoe. Presently, the zeolites are labelled ZSM-5(X)
and SSZ-13(X) with X denoting the SiO2/Al2O3 ratio. The
amorphous support materials SiO2 (SS 61138) and γ-Al2O3 (SA
6175) were obtained from Saint-Gobain.
2.2 Catalyst preparation
2 wt% Rh-based catalysts were prepared by the incipient
wetness impregnation (IWI) method followed by calcination
in flowing air at 600 °C for 6 h, as described elsewhere.13
A
catalyst with only single atom Rh sites was prepared by an
ion exchange (IE) method with ZSM-5(30) zeolite as the
support material. Here 0.1111 g rhodium(III) nitrate hydrate
(giving 0.04 g Rh) was added to a beaker with 100 ml
deionized water, and 1.96 g ZSM-5(30) powder was then
added. The mixture was stirred at room temperature for 24
hours followed by washing with deionized water until the pH
of the filtrate was neutral. The filter cake was then dried at
80 °C to remove the moisture. The sample was then
subjected to a second, identical ion-exchange step followed
by washing and drying. Finally, the ion exchanged catalyst
was calcined in flowing air at 600 °C for 6 h. Because of
texture changes after impregnation or ion exchange, the
calcined catalysts were ground, pelletized, crushed, and
sieved again to 150–300 μm. The catalysts prepared by IWI
were denoted as 2 wt% Rh/support, e.g. 2 wt% Rh/ZSM-
5(280), and the Rh catalyst prepared by ion exchange was
labelled Rh/ZSM-5(30)-IE. The calcined samples were used for
characterization and catalytic tests without any further
degreening treatments. Tests of stability as a function of time
on stream are presented in the ESI.†
2.3 Catalyst characterization
The XRD patterns of the fresh catalysts were measured using
a Panalytical XPert Pro instrument system in Bragg–Brentano
geometry working in reflectance mode with Cu Kα as the
radiation source (1.54056 Å, 40 kV and 40 mA). The
investigated 2θ range was 5–70° with a step size of 0.017°.
The distribution of Rh between Rh nanoparticle sites and
single atom sites was determined by DRIFTS measurements
of CO chemisorbed at 25 °C on the oxidized form of the
catalyst. The details of the instrument have been described
elsewhere.13,30 The spectra were obtained by averaging 76
scans at a resolution of 4 cm−1. Catalyst samples in the fresh
oxide form were diluted with KBr (FTIR grade, Sigma-Aldrich)
particles and loaded into the chamber. Before dosage of CO,
the background spectrum was collected in He at 25 °C. Then
CO was introduced to the sample by mixing 9.5 vol% CO in
Ar (AGA Industrial Gas Denmark) and He (Air Liquide
Denmark), giving a CO concentration in the range of 0–5000
ppm. The spectra were collected during both CO
chemisorption and He flush. The concentration of CO was
increased stepwise until no changes in the spectra were
2. Experimental methods
2.1 Materials
Rhodium(III) nitrate hydrate Rh(NO3)3·xH2O (∼36 wt% Rh
basis, Sigma Aldrich) was used as a precursor for the catalyst
preparation. ZSM-5 zeolites with Si/Al molar ratios of 140
(SiO2/Al2O3 = 280, CBV 28014), 40 (SiO2/Al2O3 = 80, CBV
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Catalysis Science & Technology
Paper
Table 1 List of reaction atmospheres for oxidation of CH4
Concentrations
observed, indicating saturation of the surface. After
chemisorption the CO flow was switched off, the spectra were
collected during He flush until no further change in the
spectra occurred. On both isolated single atom sites and
Rh2O3 oxide nanoparticles, CO adsorption is expected to
result in Rh(CO)2 dicarbonyl species.24,31–33 The symmetric
and asymmetric CO stretches of Rh carbonyls Rh(CO)2 on
isolated Rh single atoms are labelled as as and aa, and the
symmetric and asymmetric CO stretches of Rh(CO)2 on
Rh2O3 nanoparticles are labelled as bs and ba throughout this
work. The last spectrum during the He flush was used to
determine the ratio between single atom Rh sites and
nanoparticle Rh sites according to eqn (1).
CH4
O2
H2O
SO2
N2
(vol%)
Atmospheres
(ppm)
(vol%)
(vol%)
(ppm)
Rea-1
Rea-2
Rea-3
2500
2500
2500
10
10
10
—
5
5
—
—
20
Rest
Rest
Rest
was measured using a fixed-bed plug-flow quartz reactor placed
in a furnace with three heating zones (top, middle, and bottom)
described in more detail elsewhere.13 A fresh catalyst sample
was loaded for the measurement in each gas atmosphere. For
each test, 0.12 g of catalyst was weighed and diluted with 1.08
g inert sand to give a total bed mass of 1.20 g. The catalyst and
sand mixture was loaded between two layers of quartz wool.
The gas flow rate was 300 N mL min−1 (normal conditions for
temperature and pressure: 0 °C, 1 atm), giving a gas hourly
RhðCOÞ2;SA
A
as
¼
(1)
RhðCOÞ2;NP Ab
s
Here Aa is the peak area of the symmetric stretch of Rh
−1
space velocity (GHSV) of 150 000 N mL gcat h−1. The effluent
s
single atom bonded carbonyls, Rh(CO)2,SA, and Ab is the
S
gas was measured using an online IR NGA 2000 gas analyzer
from Rosemount to monitor the concentrations of CO2, CO
and O2 and a Micro GC 3000A from Agilent Technologies for
CH4, CO2, and O2 concentrations. The CH4 concentration
measured by the Micro GC was used to calculate the conversion
of CH4 as shown in eqn (2).
peak area of the symmetric stretch of Rh2O3 nanoparticle
bonded CO Rh(CO)2,NP. This analysis is based on the two
strongest bands, which lowers uncertainty, and as the
extinction coefficients of the two IR modes are not too
dissimilar (less than a factor of 3),25 this ratio should give a
reasonable estimate of the site ratio.
The Rh loading of the Rh/ZSM-5(30)-IE catalyst was
determined by inductively coupled plasma optical emission
spectroscopy (ICP-OES) with an Agilent 720 ES ICP-OES. Here
0.1 g finely ground catalyst and 4 g finely ground Na2O2 (pro
analysis quality) were mixed thoroughly and transferred
quantitatively into a zirconium crucible in duplicate. The
sample was heated to 500 °C for 4 hours in a muffle furnace.
After cooling, the fusion cake was dissolved in pure water
(18.2 MΩ) and the zirconium crucible was leached with 10 ml
37% HCl (pro analysis quality) three times with gentle
heating. The sample solution was then transferred to a 200
ml volumetric flask and filled to the mark with pure water
(18.2 MΩ). The sample solution was analyzed by ICP-OES
with two to five times further dilution and the emission
signals from several Rh specific emission lines were
compared to the signal from certified and matrix matched
calibration standards containing 0–1 mg L−1 Rh. The relative
precision of the analysis is estimated to be 5%.
The morphology and distribution of Rh on fresh and spent
Rh catalysts were analyzed with a Talos F200X analytical (S)
TEM equipped with a new advanced energy dispersive X-ray
spectrometer (EDS). A high angle annular dark field (HAADF)
detector was used to acquire pictures. The EDS was
complemented by an electron energy loss spectroscopy (EELS)
system with 0.7 eV energy resolution for image filtering
(EFTEM) and quantitative analysis of the sample elements.
½CH4ꢀin − ½CH4ꢀ
XCH
¼
out × 100%
(2)
4
½CH4ꢀin
Here [CH4]in is the inlet CH4 concentration measured during
bypass of the reactor and [CH4]out is the outlet CH4
concentration. In these experiments, the only products of CH4
oxidation were CO2 and H2O and the carbon balance (CB) was
within 2% for all the activity tests.
The temperature programs for the activity measurement
for CH4 oxidation in different atmospheres are shown in
Fig. 1. Measurements were made during both heating and
subsequent cooling, and throughout the text the heating is
termed the 1st run, whereas the subsequent cooling is
termed the 2nd run. Under dry and SO2-free conditions (Rea-
1), the activity during the 1st run was measured under
steady-state conditions at each temperature and during the
2nd run it was measured under transient conditions as
shown in Fig. 1(a). In the presence of 5 vol% H2O (Rea-2), an
extra stability test at 475 °C for 15 h was carried out in the
2nd run during cooling (Fig. 1(b)). With both 5 vol% H2O
and 20 ppm SO2 present (Rea-3), both the 1st and 2nd runs
were measured under steady-state conditions, and extra 15 h
stability tests were carried out at 450 °C during heating and
at 500 °C during cooling (Fig. 1(c)).
3. Results and discussion
2.4 CH4 oxidation activity measurement
3.1 Characterization of the fresh catalysts
The activity of the Rh-based catalyst in the temperature range
The results of XRD analyses for the supported Rh catalysts
of 250–600 °C in different atmospheres (as listed in Table 1)
and the pure support materials are shown in Fig. 2(a)–(f).
This journal is © The Royal Society of Chemistry 2020
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Fig. 1 Temperature programs for the CH4 oxidation activity test in (a): 2500 ppm CH4 + 10 vol% O2, (b): 2500 ppm CH4 + 10 vol% O2 + 5 vol%
H2O, and (c): 2500 ppm CH4 + 10 vol% O2 + 5 vol% H2O + 20 ppm SO2, balanced with N2, GHSV = 150 000 N mL gcat h−1. Figure reproduced
−1
with permission from Zhang et al.13 Copyright 2020, American Chemical Society.
Fig. 2 XRD patterns of the pure supports and 2 wt% Rh catalysts on (a): SiO2, (b): γ-Al2O3, (c): ZSM-5(280), (d): ZSM-5(80), (e): ZSM-5, and (f): SSZ-
13(24). Data for Rh/SiO2, Rh/γ-Al2O3, and Rh/ZSM-5(280) from Zhang et al.13
The XRD patterns of the pure supports confirmed the MFI
structure of all the ZSM-5 samples, the CHA structure of SSZ-
13(24), and the amorphous nature of SiO2 and γ-Al2O3. The
impregnation of Rh did thus not change the structure of the
support. The Rh catalysts contain well-dispersed Rh2O3,
which is largely amorphous on all the supports. Previous
analyses of spent catalysts have also suggested that this oxide
form is the working state of the catalyst under the presently
used oxygen-rich reaction conditions.13 The XRD pattern of
Rh/ZSM-5(30)-IE is shown in Fig. S1 in the ESI† and it is
identical to that of the pure ZSM-5(30) support, but ICP-OES
continuous Rh2O3 phase could be clearly identified on the Rh/
ZSM-5(30)-IE sample, as shown in Fig. 3(b). The EDS mapping
does reveal a Rh content of 0.294
0.015 wt%, indicating
that Rh is in a highly dispersed state invisible to XRD.
The HAADF-STEM images of 2 wt% Rh/ZSM-5(280) and
Rh/ZSM-5(30)-IE are shown in Fig. 3 and the corresponding
EDS maps are shown in Fig. 4. Rh2O3 nanoparticles (∼20 nm)
as a separate phase outside the zeolite structure could be
observed for 2 wt% Rh/ZSM-5(280) (Fig. 3(a)). Oppositely, no
Fig. 3 HAADF-STEM images of (a): 2 wt% Rh/ZSM-5(280) and (b): Rh/
ZSM-5(30)-IE. The red lines in (a) represent 20 nm. Fig. 3(a)
reproduced with permission from Zhang et al.13 Copyright 2020,
American Chemical Society.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Catalysis Science & Technology
Paper
Fig. 4 HAADF-STEM-EDS of (a): 2 wt% Rh/ZSM-5(280) and (b): Rh/ZSM-5(30)-IE. Fig. 4(a) reproduced with permission from Zhang et al.13
Copyright 2020, American Chemical Society.
of Rh/ZSM-5(30)-IE shown in Fig. 4(b) also confirmed the high
dispersion of Rh in the ion exchanged sample. The Rh loading
determined from EDS of the measured crystal is 0.44 0.21
wt% which is in reasonable agreement with the result from
dispersed as single Rh atoms at exchange sites with small
amounts of nanoparticles as shown in Fig. 5(g). To quantify
ICP-OES (0.294
0.015 wt%). This suggests that the highly
dispersed state seen in Fig. 3(b) and 4(b) is representative of
the entire Rh/ZSM-5(30)-IE sample.
The CO-DRIFTS results illustrating the distribution of the
Rh sites (i.e. Rh2O3 nanoparticles or single Rh atoms) on the
various supports are shown in Fig. 5. Fig. 5 shows that
dicarbonyls, Rh(CO)2, were formed on both isolated single
Rh atoms (as and aa at 2115 cm−1 and 2048 cm−1 24,31
and
)
oxidized Rh2O3 nanoparticles (bs and ba, at 2084–2089 cm−1
and 2010–2016 cm−1).29,32,33 Fig. 5(a) shows that the ion
exchanged Rh/ZSM-5(30)-IE catalyst only contains isolated
single Rh atoms at the exchange sites, in good agreement
with the high dispersion evident from the XRD and TEM
results. Oppositely, Fig. 5(b–d) show that the fresh Rh/SiO2
and Rh/γ-Al2O3 catalysts and the Rh/ZSM-5(280) catalyst on a
ZSM-5 support with a SiO2/Al2O3 ratio of 280 only contain
Rh2O3 particles as evidenced by the symmetric and
asymmetric stretches of Rh dicarbonyl Rh(CO)2 on oxide
nanoparticles (bs and ba). This is in good agreement with the
XRD results in Fig. 2(c) and the STEM results in Fig. 3(a) and
4(a). For ZSM-5 zeolites with lower SiO2/Al2O3 ratios and
hence more exchange sites, the IR data in Fig. 5(e–g) show
that more Rh is dispersed as single atoms at the exchange
sites and that Rh exists as a mixture of nanoparticles and
isolated Rh atoms. Fig. 5(d–f) show an increased relative
intensity of single Rh atoms at exchange sites (as and aa)
relative to Rh2O3 nanoparticle sites (bs and ba) with
decreasing SiO2/Al2O3 ratio of the ZSM-5 zeolites. By using an
Al-rich CHA structure zeolite, SSZ-13(24), Rh was also mainly
Fig.
5 CO-DRIFTS spectra obtained in He flush after CO
chemisorption at 25 °C. (a): Rh/ZSM-5(30)-IE; (b): 2 wt% Rh/γ-Al2O3;
(c): 2 wt% Rh/SiO2; (d): 2 wt% Rh/ZSM-5(280); (e): 2 wt% Rh/ZSM-
5(80); (f): 2 wt% Rh/ZSM-5(50); (g): 2 wt% Rh/SSZ-13(24). Data for Rh/
SiO2, Rh/γ-Al2O3, and Rh/ZSM-5(280) from Zhang et al.13
This journal is © The Royal Society of Chemistry 2020
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
the ratio of single atoms to nanoparticle sites, the area ratios
between the IR band for the symmetric CO stretch at single
atom sites (Rh(CO)2,SA, as) and the band for the symmetric
CO stretch at nanoparticles (Rh(CO)2,NP, bs) were calculated
according to eqn (1) and are shown in Fig. 6. The linear
relationship between the SiO2/Al2O3 ratio of the zeolite and
the Rh(CO)2,SA/Rh(CO)2,NP ratio in Fig. 6 illustrates that the
dispersion of Rh atoms to exchange sites is directly
correlated to the availability of exchange sites in the zeolite
which scales with the Al content of the zeolites.
The characterization of the Rh-based catalysts by XRD,
HAADF-STEM-EDS, and CO-DRIFTS thus provides a relatively
consistent picture. For SiO2, γ-Al2O3 and the most Si-rich
ZSM-5(280) zeolite, rhodium only exists in the form of Rh2O3
nanoparticles. For the ZSM-5 samples with higher Al-content,
rhodium exists as a mixture of oxide nanoparticles and single
Rh atoms at exchange sites, and the relative amount of Rh at
exchange sites scales with the number of available exchange
sites. Based on the absence of a continuous Rh2O3 phase in
STEM-EDS investigations (Fig. 4(b)) and the exclusive
identification of single atom sites in DRIFTS measurements
(Fig. 5(a)), it is concluded that Rh/ZSM-5(30)-IE only contains
single Rh atoms at the exchange sites of the zeolite. Using
the active site density on Rh2O3 estimated from methanol
chemisorption by Badlani and Wachs,34 and the Rh2O3
nanoparticle size of 10–20 nm in the 2 wt% Rh/ZSM-5(280)
sample as determined from the STEM results in Fig. 3(a), the
concentration of active sites in this sample should be in the
order of 15–29 μmol gcat−1. The ion-exchanged Rh/ZSM-5(30)
frequencies of the two types of sites – i.e. Rh2O3 nanoparticle
sites and single atom sites.
3.2 Activity for CH4 oxidation
The catalytic activity of the Rh-based catalysts for CH4
oxidation in different reaction atmospheres is shown in the
form of light-off curves in Fig. 7, and the temperature for 50%
CH4 conversion (T50
) for the catalysts under different
conditions is listed in Table 2. A fresh catalyst sample was
loaded for the activity measurement in each atmosphere. For
the feeds without SO2, only the conversion during heating
(the so-called 1st run) is shown in Fig. 7, while the conversion
during cooling is shown in Fig. S2 and S3 in the ESI.† For the
SO2-containing feed, the data for both heating (1st run) and
cooling (2nd run) are shown. Fig. 7(a and b) and Table 2 show
the catalytic activity in the absence of SO2 and here the most
active sample is Rh/ZSM-5(280), which according to Fig. 5 only
has Rh in the form of Rh2O3 particles. The least active sample
is the ion-exchanged Rh/ZSM-5(30)-IE sample, where rhodium
initially is exclusively present in the form of Rh atoms at
zeolite exchange sites. As Rh/ZSM-5(280) and Rh/ZSM-5(30)-IE
are estimated to have a similar concentration of surface sites,
this strongly suggests that the sites on Rh2O3 particles are far
more active than single Rh atoms. The remaining zeolite
supported samples, which initially have rhodium in a mixture
of nanoparticle and single atom forms, exhibit intermediate
activity. The difference in activity between particle and single
atom sites is likely to arise from the underlying differences in
electronic structure. The single atom sites are likely to be
oxygen-deficient under reaction conditions, possibly because
they are occupied by stable CO species,35 and this is likely to
lower their activity in a Mars–van Krevelen type reaction
mechanism. The lower activity of the Rh single atom sites
may contribute to their higher selectivity to partial oxidation
products such as methanol under certain conditions.24 The
significantly lower activity of single atom sites compared to
nanoparticle sites for CH4 oxidation was also reported for Pd
catalysts.18,36,37 The conversion for the Rh/ZSM-5(30)-IE
catalyst is improved significantly during the subsequent
cooling (2nd run), and the conversion is shifted ca. 25 °C
down in temperature (Fig. S2 in the ESI†). Characterization of
the spent Rh/ZSM-5(30)-IE catalyst by HAADF-STEM (Fig. 8
and S6†) clearly illustrates that agglomeration of the Rh single
atoms into small (<5 nm) nanoparticles has occurred during
operation. The higher activity during cooling (2nd run)
compared to heating (1st run) for this sample is attributed to
a higher activity of the Rh2O3 particles that have formed by
agglomeration of single Rh atoms during the heating (1st
catalyst contains only 0.294
0.015 wt% Rh, but Rh is
exclusively present as single atom sites. Taking each single
Rh atom as an active site, this translates into a concentration
of active sites of 29 μmol gcat−1. Given the similarity in site
concentration, the differences in activity between the two
samples should therefore reflect the differences in turnover
run). These results indicate
a
tendency to partial
agglomeration during operation that can explain why the
activity of the zeolite supported samples with a mixture of
particle and single atom sites does not correlate fully to the
initial dispersion illustrated in Fig. 6.
Fig. 7(b) and Table 2 show that in the presence of 5 vol%
H2O there is a lower activity for all the supports, and T50 is
Fig. 6 The area ratio of the symmetric stretch of the CO bonded to
the single atom Rh site Rh(CO)2,SA to the symmetric stretch of the CO
bonded to the Rh2O3 nanoparticle site Rh(CO)2,NP as a function of the
SiO2/Al2O3 ratio of the zeolite support in the fresh 2 wt% Rh samples
determined from Fig. 5.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Catalysis Science & Technology
Paper
Fig. 7 Conversion of CH4 on Rh-based catalysts in different atmospheres. (a): 1st run in CH4 + O2; (b): 1st run in CH4 + O2 + H2O; (c): 1st run in
CH4 + O2 + H2O + SO2; (d): 2nd run in CH4 + O2 + H2O + SO2. 1: Rh/ZSM-5(30)-IE (orange), 2: 2 wt% Rh/γ-Al2O3 (blue), 3: 2 wt% Rh/SiO2 (black),
4: 2 wt% Rh/ZSM-5(280) (red), 5: 2 wt% Rh/ZSM-5(80) (light green), 6: 2 wt% Rh/ZSM-5(50) (gray), 7: 2 wt% Rh/SSZ-13(24) (dark green). 2500 ppm
−1
CH4, 10 vol% O2, 5 vol% H2O when present, 20 ppm SO2 when present, balanced with N2, GHSV = 150 000 N mL gcat h−1. Data for 2 wt% Rh/
ZSM-5(280) from Zhang et al.13
Table 2 T50 of Rh-based catalyst in different atmospheres
T50b (°C)
CH4 + O2
1st run
CH4 + O2 + H2O
1st run
CH4 + O2 + H2O + SO2
1st run
Catalysts
2nd run
Rh/ZSM-5(30)-IE(0.294 wt%)a
2 wt% Rh/γ-Al2O3
2 wt% Rh/SiO2
2 wt% Rh/ZSM-5(280)
2 wt% Rh/ZSM-5(80)
2 wt% Rh/ZSM-5(50)
2 wt% Rh/SSZ-13(24)
404
358
361
326
355
330
357
445
418
460
395
412
401
433
531
509
>600
478
570
495
567
515
>600
483
574
513
520
540
a
b
The weight loading of Rh/ZSM-5(30)-IE was determined by ICP-OES. The T50 was calculated by interpolation between the CH4 conversions at
two adjacent temperatures.
shifted ca. 60 °C up in temperature. Such inhibition by H2O
has also been observed previously for both Rh and Pd
catalysts.13,14,38 This is partly attributable to a competitive
adsorption by OH formed from H2O. For Si-rich Rh/ZSM-
5(280), the water inhibition is completely reversible,13 and no
gradual deactivation was seen here during a stability test at 475
°C for 15 h (Fig. S4 in the ESI†). However, for the Al-rich
zeolites the stability is also affected (Fig. S4†) with a clear
continuous deactivation over time, especially for the Rh/ZSM-
5(30)-IE and Rh/SSZ-13(24) catalysts with the highest Al-content
in the zeolite. The stronger inhibition for Al-rich zeolite
supports has also been observed for zeolite-supported Pd
catalysts and has been attributed to changes in hydrophobicity
with proton content, which makes more proton-rich zeolites
more hydrophilic and thus more susceptible to the adverse
effects of water.18–20,22 Given that the single atom sites are
This journal is © The Royal Society of Chemistry 2020
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
explanation could be that some sulfation occurs for the more
Al-rich zeolites in the same way as for the amorphous γ-Al2O3
and that this sulfation has negative implications for the
performance of the active phase.18,45 Previous EDS mapping
of the elemental distribution in Rh/ZSM-5 after operation in
a SO2-containing atmosphere did indicate some correlation
between Al and sulfur, which could suggest that sulfur
uptake occurs at the Al sites.13 More Al-rich zeolites are thus
likely to take up greater amounts of sulfur on the support.
This creates a situation, where the most silica-rich zeolites,
particularly the 2 wt% Rh/ZSM-5(280) catalyst, provide the
best performance. The activity of 2 wt% Rh/ZSM-5(280) was
found to be stable over time in the presence of SO2 and H2O
as indicated by a limited (5 °C) shift for T50 between the 1st
and 2nd runs after time on stream in SO2 for around 20 h
(Table 2 and Fig. 7(c) and (d)) as well as the stable conversion
of CH4 during the 15 h stability test at 450 °C and 500 °C (Fig.
S5 in the ESI†). The relatively sulfur-tolerant behavior of Rh-
catalysts for methane oxidation has previously been correlated
to the ability of moist, bulk Rh sulfates to release part of their
sulfur at lower temperatures.13
Fig. 8 HAADF-STEM images of spent Rh/ZSM-5(30)-IE. (a): after CH4
oxidation in CH4 + O2 atmosphere; (b) after CH4 oxidation in CH4 + O2
+ H2O atmosphere. EDS analysis can be seen in Fig. S6 and S7 in the
ESI.† The red lines represent 5 nm. The STEM picture of the fresh
catalyst can be seen in Fig. 3(b).
found to be less active and favored by a higher Al content, this
gradual loss of activity for Al-rich zeolites could be due to a
water-mediated dispersion into single atom species. The results
in Fig. 7(a and b) show that Rh/γ-Al2O3 and Rh/SiO2 are of
intermediate activity despite having rhodium entirely in the
more active nanoparticle form, which would indicate that there
is a beneficial support effect of the structured zeolite support.
Rh/SiO2 exhibits a modest activity in the absence of H2O
(Fig. 7(a)), but is strongly inhibited by the presence of H2O in
the feed and becomes the least active catalyst when water is
present. The strong inhibition by H2O for SiO2 as a support
was also reported for CH4 oxidation on Pd.39 This suggests the
existence of specific interactions between water and the
amorphous silica support that are highly detrimental to the
catalytic activity. The Rh/SiO2 catalyst also experienced a
continuous gradual deactivation in the presence of H2O, which
was not seen for the Si-rich zeolite supports (ESI† Fig. S4). A Si-
rich zeolite support is thus able to avoid the stronger inhibition
occurring for both amorphous SiO2 and Al-rich zeolites.
In a realistic reaction atmosphere containing H2O and
SO2, the Rh/ZSM-5(280) catalyst thus emerges as the most
active sample. As discussed, this can be attributed to the
silica-rich zeolite, which minimizes the formation of less
active Rh single atom sites, minimizes the water inhibition
that occurs for amorphous SiO2 and the more Al-rich
supports and avoids the negative effects of SO2 that occur for
both Al-rich zeolites and γ-Al2O3.
4. Conclusion
The influence of the support has been investigated for Rh-
based catalysts employed for total oxidation of CH4 using
both amorphous SiO2 and γ-Al2O3 and ZSM-5 (SiO2/Al2O3
=
In the presence of both 5 vol% H2O and 20 ppm SO2, the
catalytic activity is lowered further, and the oxidation is
shifted to a higher temperature as shown in Fig. 7(c) and (d).
This is attributed to the formation of sulfate-type species on
the active phase that block the active sites.13 In the initial
period of SO2 exposure during the 1st run, the more Al-rich
samples, particularly Rh/γ-Al2O3 and Rh/ZSM-5(50), are more
active. It is well-established that the surface of Al2O3 takes up
SO2 and becomes sulfated.12,40–45 The higher initial activity
on the Al2O3 support is attributed to this initial SO2 uptake
by the Al2O3 sites that protects the active phase from SO2
until the support is saturated. During cooling (2nd run in
Fig. 7(d)), a stable activity level (Fig. S5(b)†) indicates that the
γ-Al2O3 support has been saturated, whereby the full impact
of SO2 on the active phase becomes visible, and here a strong
deactivation is also observed for the Rh/γ-Al2O3 sample.
Fig.7(d) shows that the more Al-rich zeolites are relatively
more affected by the presence of SO2. This is most clearly
seen for the Rh/ZSM-5(50) sample, which is almost as active
as Rh/ZSM-5(280) in the SO2-free atmospheres, but
significantly less active in the presence of SO2. An
280, 80, and 50) and SSZ-13 (SiO2/Al2O3 = 24) zeolites. A series
of 2 wt% Rh catalysts on the various supports were prepared
by incipient wetness impregnation and a catalyst containing
0.294
0.015 wt% Rh was prepared by ion exchange. The
support was found to have a strong impact on both the
speciation of Rh and the catalytic activity, as the support
influences the distribution of Rh between Rh2O3 nanoparticle
sites and single atom sites. On SiO2, γ-Al2O3 and the most Si-
rich ZSM-5(280) zeolite, Rh is only present as Rh2O3 particles,
but on the more Al-rich zeolite supports a fraction of Rh is
dispersed as single atoms on the zeolite exchange sites. The
fraction of Rh at exchange sites in the fresh catalyst scales
with the Al-content of the zeolites as expected from the
increased number of exchange sites. However, the ion-
exchanged sample clearly shows that there is a tendency for
agglomeration of single atom sites into particles during
methane oxidation. This means that the site distribution of
the working catalyst may differ from that of the fresh catalyst.
For CH4 oxidation with and without the presence of 5
vol% H2O, the highest catalytic activity is obtained for the
Rh/ZSM-5(280) sample containing only rhodium in the
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Catalysis Science & Technology
Paper
nanoparticle form and the lowest activity occurs for the ion
exchanged sample with Rh exclusively in the single atom
form. The zeolite-supported samples containing a mixture of
nanoparticle and single atom sites exhibit an intermediate
activity. This tendency clearly illustrates that the oxide
nanoparticles are the most active form of Rh for total
oxidation of methane. The presence of water is inhibiting to
all the catalysts, but most inhibiting for Rh2O3 on an
amorphous SiO2 support, where the T50 is shifted 100 °C up
in temperature by the presence of 5 vol% H2O. The lower
activity of Rh/SiO2 despite Rh being in the more active
nanoparticle form is attributed to this particularly strong
inhibition from water that must be associated with the SiO2
support. With both amorphous SiO2 and Al-rich supports,
water also affects the stability resulting in slow protracted
deactivation, whereas Si-rich ZSM-5 zeolite supports do not
yield this instability in the presence of water.
In the presence of both H2O and SO2, the most Al-rich
supports are initially more active, but once a steady state has
been reached the inhibition is stronger for the more Al-rich
supports. This is most likely because the more Al-rich
supports take up sulfur. This uptake delays the poisoning of
the active phase, but at the steady state the sulfur uptake by
the more Al-rich supports appears to be detrimental to the
performance of the active phase. These findings provide a
basis for design of catalysts to be used under real engine
exhaust gas conditions where both H2O and SO2 are present.
The results illustrate that the optimal support is one that
minimizes the interaction with both the reaction atmosphere
and the active phase. Too strong interactions with the active
phase causes more Rh to be lost to the less active single atom
sites and too strong interaction with the atmosphere causes
the support to undergo interactions with H2O or SO2 that are
detrimental to the performance of the active phase. The Si-
rich ZSM-5(280) zeolite emerges as the best support because
the low number of exchange sites ensures that Rh is present
in the more active oxide nanoparticle form, the low Al content
prevents the sulfur uptake by the support that inhibits the
active phase at the steady state and the structuring of the
support prevents the particularly strong water-induced
inhibition that occurs with an amorphous SiO2 support.
Lundegaard, Ramchandra R. Tiruvalam and Lars Frøsig
Østergaard from Haldor Topsoe. We thank Saint Gobain for
providing the γ-Al2O3 and SiO2 supports, and Haldor Topsoe
for providing the SSZ-13 zeolite support.
References
1 A. Raj, Johnson Matthey Technol. Rev., 2017, 60, 228–235.
2 H. Cai, A. Burnham, R. Chen and M. Wang, Energy Policy,
2017, 109, 565–578.
3 M. I. Khan, T. Yasmeen, M. I. Khan, M. Farooq and M.
Wakeel, Renewable Sustainable Energy Rev., 2016, 66,
702–741.
4 P. Gélin and M. Primet, Appl. Catal., B, 2002, 39, 1–37.
5 Information on dual-fuel engine in natural gas mode
provided by MAN Energy Solutions, 2017.
6 Y. Lou, J. Ma, W. Hu, Q. Dai, L. Wang, W. Zhan, Y. Guo,
X.-M. Cao, Y. Guo, P. Hu and G. Lu, ACS Catal., 2016, 6,
8127–8139.
7 D. Ciuparu, M. R. Lyubovsky, E. Altman, L. D. Pfefferle and
A. Datye, Catal. Rev.: Sci. Eng., 2002, 44, 593–649.
8 M. Cargnello, J. J. D. Jaen, J. C. H. Garrido, K. Bakhmutsky,
T. Montini, J. J. C. Gamez, R. J. Gorte and P. Fornasiero,
Science, 2012, 337, 713–717.
9 A. T. Gremminger, H. W. Pereira De Carvalho, R. Popescu,
J. D. Grunwaldt and O. Deutschmann, Catal. Today,
2015, 258, 470–480.
10 M. Monai, T. Montini, M. Melchionna, T. Duchon, P. Kúš, C.
Chen, N. Tsud, L. Nasi, K. C. Prince, K. Veltruská, V.
Matolín, M. M. Khader, R. J. Gorte and P. Fornasiero, Appl.
Catal., B, 2017, 202, 72–83.
11 M. S. Wilburn and W. S. Epling, Emiss. Control Sci. Technol.,
2018, 4, 78–89.
12 N. Sadokhina, G. Smedler, U. Nylén, M. Olofsson and L.
Olsson, Appl. Catal., B, 2018, 236, 384–395.
13 Y. Zhang, P. Glarborg, K. Johansen, M. P. Andersson, T. K.
Torp, A. D. Jensen and J. M. Christensen, ACS Catal.,
2020, 10, 1821–1827.
14 Y. Zhang, P. Glarborg, M. P. Andersson, K. Johansen, T. K.
Torp, A. D. Jensen and J. M. Christensen, Appl. Catal., B,
2020, 277, 119176.
15 Y. Zhang, PhD Thesis, Technical University of Denmark,
2019.
16 O. V. Buyevskaya, D. Wolf and M. Baerns, Catal. Lett.,
1994, 29, 249–260.
Conflicts of interest
There are no conflicts of interest to declare.
17 I. Nova and E. Tronconi, Urea-SCR Technology for DeNOx
After Treatment of Diesel Exhausts, Springer-Verlag, New York,
2014.
Acknowledgements
This work is part of the Danish societal partnership, Blue
INNOship and partly funded by Innovation Fund Denmark
(IFD) under File No.: 155-2014-10 and the Danish Maritime
Fund. Additional funding is provided by the Technical
University of Denmark (DTU) and Haldor Topsoe. The engine
exhaust gas conditions were provided by MAN Energy
18 I. Friberg, N. Sadokhina and L. Olsson, Appl. Catal., B,
2019, 250, 117–131.
19 A. W. Petrov, D. Ferri, F. Krumeich, M. Nachtegaal, J. A. Van
Bokhoven and O. Kröcher, Nat. Commun., 2018, 9, 2545.
20 Q. Dai, Q. Zhu, Y. Lou and X. Wang, J. Catal., 2018, 357,
29–40.
Solutions.
The
XRD,
HAADF-STEM-EDS,
ICP-OES
21 J. Bin Lim, D. Jo and S. B. Hong, Appl. Catal., B, 2017, 219,
155–162.
measurements and analyses were carried out by Lars Fahl
This journal is © The Royal Society of Chemistry 2020
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
22 P. Losch, W. Huang, O. Vozniuk, E. D. Goodman, W.
Schmidt and M. Cargnello, ACS Catal., 2019, 9,
4742–4753.
33 A. Theolier, A. K. Smith, M. Leconte, J. M. Basset, G. M.
Zanderighi, R. Psaro and R. Ugo, J. Organomet. Chem.,
1980, 191, 415–424.
23 Y. Tang, Y. Li, V. Fung, D. Jiang, W. Huang, S. Zhang, Y.
Iwasawa, T. Sakata, L. Nguyen, X. Zhang, A. I. Frenkel and
F. F. Tao, Nat. Commun., 2018, 9, 1231.
34 M. Badlani and I. E. Wachs, Catal. Lett., 2001, 75, 137–149.
35 R. J. Bunting, J. Thompson and P. Hu, Phys. Chem. Chem.
Phys., 2020, 22, 11686–11694.
24 J. Shan, M. Li, L. F. Allard, S. Lee and M. Flytzani-
Stephanopoulos, Nature, 2017, 551, 605–608.
25 J. C. Matsubu, V. N. Yang and P. Christopher, J. Am. Chem.
Soc., 2015, 137, 3076–3084.
26 H. Guan, J. Lin, B. Qiao, S. Miao, A.-Q. Wang, X. Wang and
T. Zhang, AIChE J., 2017, 63, 2081–2088.
27 J. T. Yates, T. M. Duncan, S. D. Worley and R. W. Vaughan,
J. Chem. Phys., 1979, 70, 1219–1224.
36 E. D. Goodman, A. C. Johnston-Peck, E. M. Dietze, C. J.
Wrasman, A. S. Hoffman, F. Abild-Pedersen, S. R. Bare, P. N.
Plessow and M. Cargnello, Nat. Catal., 2019, 2, 748–755.
37 K. Okumura, E. Shinohara and M. Niwa, Catal. Today,
2006, 117, 577–583.
38 R. Gholami, M. Alyani and K. Smith, Catalysts, 2015, 5, 561–594.
39 K. Muto, N. Katada and M. Niwa, Appl. Catal., A, 1996, 134,
203–215.
28 C. Yang and C. W. Garl, J. Phys. Chem., 1957, 61, 1504–1512.
29 C. A. Rice, S. D. Worley, C. W. Curtis, J. A. Guin and A. R.
Tarrer, J. Chem. Phys., 1981, 74, 6487–6497.
30 N. D. Nielsen, J. Thrane, A. D. Jensen and J. M. Christensen,
Catal. Lett., 2020, 150, 1427–1433.
31 C. Y. Fang, S. Zhang, Y. Hu, M. Vasiliu, J. E. Perez-Aguilar,
E. T. Conley, D. A. Dixon, C. Y. Chen and B. C. Gates, ACS
Catal., 2019, 9, 3311–3321.
32 A. K. Smith, F. Hugues, A. Theolier, J. M. Basset, R. Ugo,
G. M. Zanderighi, J. L. Bilhou, V. Bilhou-Bougnol and W. F.
Graydon, Inorg. Chem., 1979, 18, 3104–3112.
40 T. Hamzehlouyan, C. Sampara, J. Li, A. Kumar and W.
Epling, Appl. Catal., B, 2014, 152–153, 108–116.
41 T. Hamzehlouyan, C. S. Sampara, J. Li, A. Kumar and W. S.
Epling, Appl. Catal., B, 2016, 181, 587–598.
42 M. S. Wilburn and W. S. Epling, ACS Catal., 2019, 9, 640–648.
43 X. Zi, L. Liu, B. Xue, H. Dai and H. He, Catal. Today,
2011, 175, 223–230.
44 L. S. Escandón, S. Ordóñez, A. Vega and F. V. Díez, J. Hazard.
Mater., 2008, 153, 742–750.
ˇ
45 I. C. Marcu and I. Sandulescu, J. Serb. Chem. Soc., 2004, 69,
563–569.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2020