Sulfated hafnia as a support for Mo oxide: A novel catalyst for methane dehydroaromatization

Article abstract of DOI:10.1016/j.cattod.2019.02.021

One of the most challenging aspects of modern day catalysis is the conversion of methane. Natural gas is an abundant source of supply of methane and is currently being used mostly for electricity generation, and much of it is simply flared. The potential to convert methane into higher-value hydrocarbons is significant. Among other reactions, direct conversion of methane via dehydroaromatization (MDHA) can be used to produce hydrogen and valuable hydrocarbons like benzene. Mo oxide supported on ZSM-5/MCM-22 has been studied extensively in recent years for MDHA. It has been reported that Mo carbides are responsible for activating methane by forming CHx species. These are dimerized into C₂H_y and then oligomerized on the strong ZSM-5/MCM-22 Br?nsted acid sites to form benzene. Related work has shown that sulfated zirconia (SZ) provides the acid sites for Mo needed to produce benzene in MDHA [1]. The similarity of sulfated hafnia (SH) with sulfated zirconia is a logical novel support for MDHA. Although SH has been used for other acid-catalyzed reactions, we are aware of no systematic study of Mo-SH for the MDHA reaction, despite the appeal of such a catalyst. Here, Mo/SH catalysts have been tested and characterized using SEM-EDS, XPS, XANES, DRIFTS, HR-TEM, BET and temperature programmed techniques. DRIFTS confirms that SH acidity is unaffected by the addition of Mo to the catalyst. SEM-EDS and XPS have been used to characterize the loading of Mo and sulfur on SH support. XANES as well as HR-TEM analysis are used to study the formation of molybdenum carbide/ oxycarbide species, which are generally regarded to be the active sites for MDHA. MDHA runs were carried out to study the effect of Mo loading, temperature and space velocity. 5% Mo-SH is catalyst more active than 1% Mo-SH, as measured by methane conversion. Conversion increased with higher temperature and lower space velocity and gradually deactivated with time. This can be attributed to catalytic surface coking, confirmed with subsequent TPO analysis. Products observed were primarily ethylene and benzene. Ethylene selectivity increased with time for higher Mo loading and lower space velocity. Benzene product selectivity increased with higher Mo loading, lower temperature, and lower space velocity, while gradually decreasing with time. A direct experimental comparison of conventional Mo-HZSM5 synthesized here showed that the Mo-SH catalyst is more active than Mo-HZSM5.

Full text of DOI:10.1016/j.cattod.2019.02.021

Accepted Manuscript
Title: Sulfated hafnia as a support for Mo oxide: A novel
catalyst for methane dehydroaromatization
Authors: Md Ashraful Abedin, Swarom Kanitkar, Srikar
Bhattar, James J. Spivey
PII:
S0920-5861(18)31628-6
DOI:
Reference:
https://doi.org/10.1016/j.cattod.2019.02.021
CATTOD 11973
To appear in:
Catalysis Today
Received date:
Revised date:
Accepted date:
12 October 2018
17 January 2019
12 February 2019
Please cite this article as: Abedin MA, Kanitkar S, Bhattar S, Spivey JJ, Sulfated hafnia
asasupportforMooxide:Anovelcatalystformethanedehydroaromatization, Catalysis
Today (2019), https://doi.org/10.1016/j.cattod.2019.02.021
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Sulfated hafnia as a support for Mo oxide: A novel catalyst for methane
dehydroaromatization
Md Ashraful Abedin, Swarom Kanitkar, Srikar Bhattar, James J. Spivey*
Cain Department of Chemical Engineering, Louisiana State University, Baton Rouge, LA 70803,
USA
*
Corresponding author: J.J. Spivey, email: jjspivey@lsu.edu; Tel: +1-225-578-3690; Fax: +1-
25-578-1476
2

Graphical Abstract
Highlights




Methane aromatization is thermodynamically favored at high temperature (~1000⁰C), but the
reaction occurs at a lower temperature (600⁰C-700⁰C) in presence of acid catalysts.
A novel approach in methane aromatization is introduced using sulfated hafnia (SH) as an acidic
support for the active Mo carbide/oxycarbide sites.
The effect of reaction parameters such as Mo loading, temperature and space velocity have been
investigated for the proposed Mo based sulfated hafnia (Mo-SH) catalyst.
Catalytic deactivation can be primarily attributed to carbon deposition as well as possible loss of
sulfate sites from SH support.

Abstract
One of the most challenging aspects of modern day catalysis is the conversion of methane. Natural
gas is an abundant source of supply of methane and is currently being used mostly for electricity
generation, and much of it is simply flared. The potential to convert methane into higher-value
hydrocarbons is significant. Among other reactions, direct conversion of methane via
dehydroaromatization (MDHA) can be used to produce hydrogen and valuable hydrocarbons like
benzene.
Mo oxide supported on ZSM-5/MCM-22 has been studied extensively in recent years for MDHA.
It has been reported that Mo carbides are responsible for activating methane by forming CHx
species. These are dimerized into C
2
y
H and then oligomerized on the strong ZSM-5/MCM-22
Brønsted acid sites to form benzene. Related work has shown that sulfated zirconia (SZ) provides
the acid sites for Mo needed to produce benzene in MDHA[1]. The similarity of sulfated hafnia
(SH) with sulfated zirconia is a logical novel support for MDHA. Although SH has been used for
other acid-catalyzed reactions, we are aware of no systematic study of Mo-SH for the MDHA
reaction, despite the appeal of such a catalyst. Here, Mo/SH catalysts have been tested and
characterized using SEM-EDS, XPS, XANES, DRIFTS, HR-TEM, BET and temperature
programmed techniques. DRIFTS confirms that SH acidity is unaffected by the addition of Mo to
the catalyst. SEM-EDS and XPS have been used to characterize the loading of Mo and sulfur on
SH support. XANES as well as HR-TEM analysis are used to study the formation of molybdenum
carbide/ oxycarbide species, which are generally regarded to be the active sites for MDHA. MDHA
runs were carried out to study the effect of Mo loading, temperature and space velocity. 5% Mo-
SH is catalyst more active than 1% Mo-SH, as measured by methane conversion. Conversion
increased with higher temperature and lower space velocity and gradually deactivated with time.

This can be attributed to catalytic surface coking, confirmed with subsequent TPO analysis.
Products observed were primarily ethylene and benzene. Ethylene selectivity increased with time
for higher Mo loading and lower space velocity. Benzene product selectivity increased with higher
Mo loading, lower temperature, and lower space velocity, while gradually decreasing with time.
A direct experimental comparison of conventional Mo-HZSM5 synthesized here showed that the
Mo-SH catalyst is more active than Mo-HZSM5.
Keywords: Methane activation, Methane dehydroaromatization, Sulfated Hafnia, Bi-functional
catalyst
Introduction
st
The exponential growth of need for energy in the 21 century has been significant. According to
the national Petroleum Council in the USA[2], it has been reported that the energy demand will
increase by 50% by the year 2030. Possible reasons behind energy crisis include increased per-
person consumption, population, limited fossil fuel resources, limits to renewable energy options,
and inefficiency. Around the globe, the demand for nonrenewable resources is increasing,
particularly in the United States[3], and all these resources have limits. To date, oil, natural gas

and coal are the principal fuels for energy production, and natural gas is generally considered to
be a cleaner fuel [4].
U.S. Energy Information Administration recently shows that U.S. natural gas production has set a
record of natural gas production. Conventional use of natural gas includes generation of electricity,
chemical production, and industrial and domestic heating [5]. Reports show that in 2017 alone,
around 90.9 billion cubic feet of natural gas was withdrawn per day from U.S. natural gas
reservoirs. This is the highest amount in the nation’s history.
A major component of natural gas is methane. Its abundance has attracted the scientific community
to develop reactions and processes to convert methane into higher-value products. Despite
significant study, the conversion of methane to large-volume chemicals has not yet fully reached
its potential[6].
Processes for conversion of methane can be categorized into direct and indirect reaction routes[7].
Indirect processes are based on the formation of an intermediate product, such as syngas, which
goes thorough an additional reaction step to form hydrocarbons[7]. These include partial oxidation,
dry reforming[8], bi-reforming[9], steam reforming[10], halogenation[11] and sulfidation[12].
Processes such as steam reforming of methane are commercially practiced and are characterized
by high temperature. Because of the additional intermediate step, these processes are capital
intensive.
Direct conversion of methane include dehydroaromatization[13], isomerization[14],
hydroconversion [15], catalytic pyrolysis[16], and metathesis[17]. The advantage of direct
approaches is to eliminate the intermediate step, which greatly reduce energy requirements and
capital investment.

Methane dehydroaromatization (MDHA) is one direct route for methane conversion[18]:
ퟔ 푪푯 ↔ 푪 푯 (품) + ퟗ 푯 ↑ (ΔG298K = +433 kJ/mol, ΔH298K = +530 kJ/mol) (1)
ퟒ
ퟔ
ퟔ
ퟐ
Very high temperatures (700ºC-1000ºC) are needed to drive the equilibrium toward the formation
of benzene and hydrogen. At these high temperatures, thermodynamics can limit the formation of
coke[19], and the formation of benzene rather than coke. However, a catalyst must be active,
selective, and stable at these demanding conditions.
MDHA was first studied by Wang et al. [20] who found that metal activated zeolite supports can
convert methane directly to benzene and hydrogen. The addition of molybdenum and/or tungsten
greatly affected the benzene selectivity at temperature above 700ºC[21]. Further studies showed
that the addition of promoters, including strong Brønsted acid sites and shape-selective a pore
structure, increased activity for MDHA[22]. It was found that a bifunctional catalyst is required
for the direct approach of methane conversion. Studies indicate that transition metal carbide or
oxycarbide act as primary site to activate the stable C-H bond to produce CH
x
species[22], which
are unstable and readily form C dimers. These dimers react on the adjacent Brønsted acid sites.
2
H
y
to oligomerize these dimers to form benzene and hydrogen[7].
Molybdenum carbide, tungsten carbide, and oxycarbide are regarded as the most active catalysts
for the MDHA reaction [23]. ZSM-5 and MCM-22 have been found to be the most active acidic
supports for these catalytic sites[18]. The proposed mechanism on Mo/W carbides/oxycarbides on
ZSM-5/MCM-22 consists of two steps: (a) dimerization of methane on the Mo sites and (b)
oligomerization of the dimers to benzene (and other aromatics) on the Brønsted sites [7].

Figure 1: Bifunctional mechanism of MDHA catalysts (BAS=Bronsted Acid sites)[7]
An important property of these Mo/W bifunctional catalysts is the in-situ formation of metal
carbide/oxycarbide species through the insertion of carbon into the lattice of the active metal[18].
As carbonaceous compounds react with the metal, carbon donates its valence site electron to the
outer d band of the active metal, which resembles an electronic density of states at Fermi level that
is similar to the electronic configuration of noble metals like Pt and Pd. Because of this, the metal
carbides act like noble metals, and this new property can be exploited for a wide range of reactions,
including methane dehydroaromatization, dehydrogenation, oligomerization etc[15].
Solid acids such as sulfated zirconia (SZ), sulfated hafnia (SH), tungstated zirconia (WZ) are
highly acidic compounds that have high surface area that are also stable at elevated
temperatures[24]. These possess strong BA sites[25] that are a key role in the oligomerization of
dimeric species in the MDHA reaction.
Mo doped on solid acids has been reported in the literature for reactions such as alkane
hydroconversion, isomerization etc [15]. Recently, Ahmed [26]reported that the surface sulfate
complex is stabilized if small amount of hafnia is added to the tetragonal phase of zirconia, and
thus enhancing their activity for the isomerization of butane to isobutene. Arata et al. [27] found
that sulfated HfO promotes acidity that catalyzes the isomerization of butane to isobutene; but to
2
the best of our knowledge, no study has yet shown dehydroaromatization of methane on the
proposed Mo-SH catalysts.

The hypothesis here is to replace the conventional ZSM-5/MCM-22 supports with enriched solid
acids containing BA sites and dope with Mo to serve as active metal for MDHA. This suggests
that sulfated hafnia can serve a similar role for MDHA. The aim of this present work is to
understand the effect of these catalysts for MDHA and to study the effect of Mo loading at reaction
conditions of interest.
Experimental Procedure
Materials
Hafnium oxychloride hydrate, HfOCl
tetra hydrate (NH Mo O were purchased from Sigma Aldrich Inc. H
was purchased from Mallinckrodt Chemicals Inc. Ultra-high purity grade H
, and 10% O /He were all purchased from Airgas Inc.
2
.xH
2
O (99% trace metals basis) and ammonium molybdate
SO (95 – 98.0 %)
, CH , 5%NH /He,
4
)
6
7
O
24.4H
2
2
4
2
4
3
N
2
2
Sample Preparation
Literature methods to prepare Zirconium oxide[28] was followed for the synthesis of Hafnium
oxide. The prepared Hafnium oxide was mixed with 500 ml of prepared solution of 0.5 M H SO
The mixture was stirred for 2 hours, followed by vacuum filtration. The retentate was dried at
10ºC overnight and was calcined subsequently at 650ºC for 3 hours to retrieve the final sulfated
2
4
.
1
Hafnia (SH) catalyst.
Mo impregnation was carried out at room temperature. Two catalysts were prepared by adding 1
and 5 weight percent of Mo respectively, in the form of ammonium molybdate, to 8 gm of prepared
SH along with 100 ml DI water mixture. The mixture was stirred for 2 hours, followed by overnight
drying at 110ºC. Subsequent calcination was carried out at 550ºC in presence of dry air. The

catalyst was sent through a sieve to achieve the final powdered form. It was later weighed and
collected in glass ampule to avoid atmospheric air and moisture exposure.
BET
Altamira AMI-200 catalyst characterization system was used for Breauner Emmett Teller (BET)
surface area analysis with N
2
monolayer adsorption. A three-point BET was used with 10 %, 20
%
and 30 % N concentrations for estimation of the catalytic surface areas.
2
XRD measurements
XRD measurements for all the samples was carried out using PANAnalytical EMPYREAN
0
0
diffractometer with Cu K radiation. Sample was collected by scanning the data from 5 to 90
α
with a step size of 0.026 sec/pt. Data analysis of XRD of all samples was done using
PANAnalytical X’Pert software.
Ammonia-TPD
Altamira AMI-200 reactor system in conjunction with Ametek Mass Spectrometer was used to run
Ammonia-TPD. 50 mg of the prepared catalyst was loaded on a quartz tube reactor, followed by
pretreatment. Temperature was ramped up to 200 °C under 30 sccm of He for 30 min to get rid of
any weakly adsorbed particles. Sample was then cooled down to 50°C under He, before
introducing 40 sccm of 5% NH
3
/He to start NH adsorption process for 90 mins. 30 sccm of He
3
was flown for 40 min to remove any residual ammonia. TCD detector was later turned on and the
temperature was ramped up at 10 °C/min. from 50 °C to 650 °C. Based on the signal from TCD,
amounts of ammonia desorbed and peak positions were calculated to quantify the corresponding
acid sites available on the catalyst.

Pyridine DRIFTS
Pyridine was used as a probe molecule in Diffuse Reflectance Infrared Fourier Transform
Spectroscopy (DRIFTS) experiment. A Thermo Scientific Nicolet 6700 FTIR equipped with
Harrick Praying Mantis reaction cell fitted with KBr windows was used to carry out the
experiment.
IR cell was loaded with the catalyst sample inside the glovebox to avoid air and moisture exposure.
Helium was introduced in the cell, followed by pretreatment at 100 °C for 30 min to clean the
surface from adsorbed impurities. Sample was cooled down to 25 °C after pretreatment and a
-1
background spectrum was recorded with a spectral resolution of 4 cm in region going from 4000
-1
–
650 cm . Catalyst was saturated with gaseous pyridine for 180 min at 25 °C. He was later
introduced again in the post saturation step to remove physisorbed pyridine from the catalyst
surface as well as the cell chamber. Sample was then treated at 100ºC for 10 min and cooled back
to room temperature and the actual spectrum was recorded. Similar spectra were recorded at room
temperature after 10 min. Treatments were done at 150 °C, 200 °C, and 300 °C to investigate the
thermal stability of the acid sites on catalysts.
XPS
X-ray Photoelectron Spectroscopy for the samples was performed at Louisiana State University
Shared Instrument Facility (SIF). All the analyses were performed on Scienta Omicron ESCA 2SR
XPS instrument with Al monochromatic X-ray source at 15 kV and at a pass energy of 40. Post
analysis of XPS data was performed using CasaXPS software.
ICP-OES

Inductively coupled plasma – optical emission spectroscopy (ICP-OES) was carried out using
Varian-MPX spectrophotometer in Wetland Biogeochemical Lab at LSU Oceanography. The
digestions of samples were carried out using a mixture of borate flux: Lithium Tetraborate (49.75
wt%), Lithium Metaborate (49.75 wt%), and Lithium Iodide (0.5 wt%). In a typical procedure, 0.2
g of catalyst sample was mixed with 2 g of borate flux mixture. This mixture was fused in a furnace
at 1000⁰C and subsequently dissolved in a warm 10% HNO solution. This solution was diluted
3
subsequently and used for ICP-OES analysis.
HR-TEM
High Resolution Transmission Electron Microscope was used to study differences in Mo phase.
The transition of Mo oxide to the active Mo carbide or oxycarbides could be observed through
HR-TEM analysis. HR-TEM scans of these catalysts were obtained using JEOL JEM-2011
Scanning TEM at an accelerating voltage of 200 kV.
SEM-EDS
SEM-EDS analysis of the samples was carried at LSU Shared Instrument Facilities (SIF) using
FEI quanta 3D FIB/SEM equipped with Ametek EDAX accessory. Samples were analyzed at a
voltage of 5 kV and at a resolution of 3 µm to detect elemental compositions.
XANES
LIII edge X-ray Absorption Near Edge Structure (XANES) spectra of Mo supported on SH were
collected at the Low Energy X-Ray Absorption Spectroscopy (Lexas) beamline of J. Bennett
Johnston, Sr., Center for Advanced Microstructures and Devices (CAMD), Louisiana State
University, Baton Rouge, Louisiana. CAMD operates an electron storage ring at 1.3 GeV, with a
typical ring current of 100 mA to 50 mA. Lexas is a windowless beamline, with only a 13 µm

Kapton™ tape separating the ring from the beamline. The University of Bonn designed double
crystal monochromator uses InSb (111) crystals in this energy range, has a resolution of ca. 0.5
eV. The monochromator was calibrated to 2481.44 eV with the peak of zinc sulfate white line.
Samples were prepared by spreading the finely-ground powder on silicon and sulfur free Kapton™
tape. All the measurements of the in transmission due to severe overlap of sulfur K and Mo L
fluorescence lines. For Mo L III edge measurement, the sample chamber was evacuated to 50 Torr.
The parameters for the scans were from 2505 eV to 2530 eV with 0.1 eV steps and from 2530 eV
to 2550 eV with 0.2 eV steps. Multiple spectra were averaged for better signal to noise ratio. Mo
standards used in the analysis were: MoO
metals basis, Alfa Aesar).
3
(99.9% metals basis, Alfa Aesar) and Mo C (99.9%
2
Catalyst activation and Methane dehydroaromatization (MDHA)
Activation of methane using bi-functional catalyst has been reported previously in literature[7].
This mechanism suggests that for methane to be activated for aromatization, two separate active
sites are necessary to work simultaneously. The first site refers to molybdenum carbide or
oxycarbide species, where methane dehydrogenates and produce C
2
H
y
dimers that go through
oligomerization stage later. The Mo C or Mo sites are generated in situ as DHA reaction
2
2
O C
x y
takes place[18]. The catalyst was loaded in an Altamira AMI 200HP reactor system equipped with
quartz tube reactor. In situ catalytic pretreatment was initiated by reducing the sulfated Hafnia
supported MoO
x
catalyst under dry hydrogen gas at a ramp rate of 10K/min till the reaction
:H gas flow
temperature of 650ºC was reached. After the reaction temperature was reached, a CH
4
2
ratio of 1:4 was introduced which led to carburization of the molybdenum oxide species. This
mixture of flow was maintained for 4 hours to initiate the formation of molybdenum

carbide/oxycarbide species. This was followed by turning off H
2
and CH gas and purging the
4
system with inert gas. CH flow was reintroduced afterwards to carry out DHA reaction.
4
Downstream reaction product gas was analyzed using Shimadzu GC2014 (FID, 2 TCDs) equipped
with Restek RT-Q-Bond column (30 m x 0.53 mm x 20 µm) in conjunction with Shimadzu QP2010
GC-MS system.
For a comparative study, Mo supported HZSM-5 was also studied under similar reaction
conditions for methane DHA. Synthesis of Mo-HZSM-5 included similar impregnation method
that was used to prepare Mo-SH catalyst.
Following formulas were used for the activity calculation.
푚표푙 퐶퐻4푖푛− 푚표푙 퐶퐻4ꢀ푢푡
%
CH
4
conversion =
x 100
푚표푙 퐶퐻4푖푛
Selectivity was calculated based on products observed. Carbon deposition was excluded from this
calculation as the coke formed at any instant of the reaction was not measured. Another reason for
the exclusion was because the rate of coke formation differs over a period and will not be discussed
in this work.
푚표푙 푃푟표푑ꢁ푐ꢂ
%
selectivity =
푥 100
푚표푙 ꢂ표ꢂ푎푙 푝푟표푑ꢁ푐ꢂ푠
Results and Discussion
Physico-chemical properties
Table 1 shows the physico-chemical properties of the 1 and 5 wt.% of Mo-SH. Baseline hafnium
2
oxide is reported in the literature to have surface area of 30-60 m /g[29, 30]. BET analysis on the
2
fresh sulfated hafnia was found to be 49 m /g, well within the expected surface area. Addition of

Mo to SH decreases the surface area only slightly, suggesting that Mo species do not impact the
surface area of SH, which is basically a monolayer structure[29]. XPS and SEM-EDX analyses
show that Mo content is close to the expected loading.
Table 1: Physico-chemical properties of Mo-SH catalysts
Sample
Mo wt.%
Mo wt.%
Mo wt.%
BET surface
2
(
intended)
(SEM-EDS)
(XPS)
-
area (m /g)
SH
-
-
49
44
45
1
5
% Mo-SH
% Mo-SH
1
5
1.6
5.2
1.1
4.5
DRIFTS
DRIFTS spectroscopy using pyridine adsorption was used to investigate the nature of Brønsted
and Lewis acid sites as a function of Mo loading. Figure 2 represents DRIFTS spectra of SH
catalysts for various loading of Mo. Literature suggests that sulfated hafnia has both Brønsted and
Lewis strong acid sites[29]. Based on this, identification of the observed IR absorbance bands after
pyridine desorption at the range of 373 K-673 K can be made. Characteristic bands at 1490, 1540,
-
1
1
1
607 and 1640 cm represent pyridine adsorbed on Brønsted acid sites [31, 32]. Bands at 1447,
-1
490, 1610 cm are assigned to pyridine adsorbed on Lewis acid sites [29, 31, 32]. Literature
suggests that bands at 1490 and 1610 represent both the Lewis as well as Brønsted acid sites.[31]

Based on the DRIFTS results, Figure 2 displays the characteristic Brønsted and Lewis acid sites
were present at 373 K after Mo was loaded in SH. As Mo loading increased, bands corresponding
-1
to Brønsted acid sites at 1490, 1540 and 1640 cm shifted slightly to lower wavenumbers,
suggesting a decrease in strength for Brønsted acidity[31]. This can be attributed to the interaction
of Lewis acidic Mo oxides [33, 34] with the available hydroxyl groups and oxygens on sulfated
hafnia surface. Thermal stability of the SH catalysts was tested up to 473 K, and despite of the
high temperature, all the catalysts showed stable acid sites and strength.
Figure 2: Pyridine-DRIFTS comparison data for sulfated hafnia and Mo-SH catalysts. Samples were desorbed in
0
pyridine at 100 C. (a) SH (b) 1% Mo-SH (c) 5% Mo-SH ; B- BrØnsted, L= Lewis
NH -TPD
3

A comparison of NH
Note that at temperature above ~600⁰C-700⁰C, the available surface sulfates start to decompose,
causing sulfur oxides to get released in gaseous form [27, 32, 35]. This causes a false signal in
3
-TPD for SH and the three Mo doped SH catalysts is shown in Figure 3.
[
TCD. Because of this, NH
signal].
3
TPD was run on all SH catalysts up to 520⁰C to measure the TCD
3
2
1
0
.5
3
.5
2
c
b
.5
1
a
.5
0
70
120
170
220
270
320
370
420
470
520
0
Temperature ( C)
Figure 3: NH3-TPD comparison data for (a) SH, (b) 1% Mo-SH, (c) 5% Mo-SH
Table 2 shows the concentration of acid sites (mmol NH3/g cat). A desorption peak at around
30⁰C was observed for all SH catalysts, due to relatively weak acidity. The results for SH and
low Mo loaded SH catalyst were quite similar, suggesting that most of the acidity was from SH
1

alone. The acid concentration decreased slightly with increase in total metal loading. This is likely
due to an exchange between the protonated sulfate sites on hafnia and metal species during
calcination. This also suggests that the addition of Mo on SH may have covered sufficient catalyst
to affect the access of the probe ammonia to the available acidic sites.
Table 2: Amount of NH desorbed/consumed per gram of catalyst
3
Catalyst
Total micromoles of NH desorbed/
3
consumed per gram catalyst
Sulfated hafnia
888.6
840.7
819.8
1
5
% Mo-SH
% Mo-SH
XANES
The formation of the reported active phase Mo
identified using Mo LIII edge XANES spectra. Figure 4 shows the absorption versus energy of the
III edge spectra for the 5% Mo-SH catalyst, along with MoO and Mo C standards for absorption
peak comparison. Fresh Mo-SH catalyst resembles the MoO standard, with two split peaks (B
and C). These unique splits refer to t2g and e splitting of Mo 4d orbitals[36] from tetrahedral
2
C or MoO
x
C generated in-situ and can be
y
L
3
2
3
g
coordination. Similar sets of peaks are observed for carburized samples, where a decrease in
intensity in the second peak (C) was observed, in comparison with the fresh sample. The spent

catalyst showed absorption at lower energy[37], including a loss of split peaks, which has uniquely
resembles the absorption edge of Mo C standard sample [37, 38]. This supports the claims made
2
in the literature [7, 33, 38] that Mo oxides are transformed to Mo
as active sites for the MDHA reaction.
2
C or MoO
x
C in-situ, which act
y
Figure 4: LIII edge XANES spectra for catalysts and reference samples: (a) MoO
Mo-SH (oxidized state), (c) Carburized 5% Mo-SH, (d) Spent 5% Mo-SH, (e) Mo
3
standard, (b) Fresh 5%
C standard
2

HR-TEM
High resolution transmission electron microscopy was used to understand the state of molybdenum
at different steps of the reaction. 5% Mo-SZ showed higher activity, hence it was chosen for HR-
TEM analysis. Catalysts after three different steps were analyzed: fresh, carburized and spent. HR-
TEM analysis of sulfated hafnia identified two different phases, primarily consisting of monoclinic
hafnia planes (111) along with some orthorhombic hafnium oxide planes (211).
Before initiating the MDHA reaction, HR-TEM was carried on fresh 5% Mo-SH samples.
Monoclinic hafnia with (111) planes was most commonly observed in fresh sample (figure 5).
Along with hafnium oxide, several planes corresponding to molybdenum (III) oxide were
identified, as those were finely dispersed over the hafnium oxide support. The most common plane
found for MoO
3
was (040), followed by (021) planes. These planes confirm the presence of Mo in
oxide phase before the beginning of the reaction[39, 40].

Figure 5: HR-TEM analysis for 5% Mo-SH fresh catalyst
After the catalyst was carburized for 4 hours, it was collected to run HR-TEM (figure 6) to identify
the reduced phases of molybdenum oxide. Along with the abundant monoclinic and some
orthorhombic hafnium oxides, new planes were observed. Reduced molybdenum oxides were
found to be formed because of carburization step, along with planes (211) corresponding to the
formation of Mo carbides. Matus et al.[41]confirmed the presence of Mo C planes at similar d-
2
spacing of 0.23 nm. Apart from these detectable Mo planes, planes corresponding to intermediate
phases in between Mo oxides and Mo carbides were also observed. These can be due to the
formation of Mo oxycarbides [42, 43]with d-spacing of 0.24 nm.

Figure 6: HR-TEM analysis for 5% Mo-SH carburized catalyst
After the 16 hour-run, 5% Mo-SH spent catalyst was collected to analyze by HR-TEM to determine
if the active Mo
hafnium oxides were detected in the process. Molybdenum was found in planes corresponding to
211) Mo C, as well as some unreduced MoO . This confirmed the formation of molybdenum
carbides, which is generally accepted as the active sites for dimerization of methane in MDHA
reaction[18]. Mo C planes were mostly found around the edges of the catalyst planes, surrounded
by graphitic carbon monolayers[42].
2
C phase is formed (Figure 7). Various planes (110, 020, -111) of monoclinic
(
2
x
2

Figure 7: HR-TEM analysis for 5% Mo-SH spent catalyst
Reaction data
Effect of Mo content
To compare the effect of the Mo content on sulfated hafnia, 1 wt % and 5 wt % of Mo catalysts
were tested.

Deactivation. Figure 8 shows that 5% Mo-SH was initially more active than 1% Mo-SH (13%
versus 7% methane conversion). However, deactivation was significant for both catalysts, and
after 650 min, methane conversion was the same, ~4%. Several reasons are likely responsible.
The primary reason is the formation of coke, as shown below by TPO. A second reason was also
found to be inevitable in these catalysts--the loss of sulfate groups from the surface of sulfated
hafnia. During the NH3-TPD of the fresh catalysts, oxide compound decomposition was detected
in the mass spec signal, which can only be attributed to sulfur oxides, because hafnia is a very
stable compound (melting point of 2758ºC). For the similar catalyst sulfated zirconia, using TGA
analysis, Srinivasan et al.[44]reported that decomposition of O
2
from the ZrO structure is highly
2
unlikely, leaving oxygen decomposition from SO -2
to be the only possible explanation.
4
Selectivity. Primary products included ethylene, ethane, propylene, propane, and aromatics
including benzene, toluene and xylene, along with hydrogen. For 1% Mo-SH, ethylene selectivity
initially increased with time (Figure 9), then decreased. For 5% Mo-SH, ethylene selectivity
increased monotonically with time, and was lower than 1% Mo-SH until the two selectivities were
essentially equal after ~700 min. This suggests that the sites that are initially responsible for
ethylene formation are different. To a first approximation, ethylene selectivity was relatively
constant on 1% Mo-SH, after an initial period of time. By comparison, sites responsible for
ethylene selectivity increase monotonically with time. This behavior may be attributed an
induction period of time in which the Mo is carbided. If so, the time needed to reach the more
fully carbided catalyst takes longer for the higher loaded 5% Mo-SH than for the 1% Mo-SH. This
is also in accordance with the general observation that Mo is responsible for the production of
C
2
y
H groups from dimerization of methane.

Figure 10 shows that the benzene selectivity decreases rapidly with both catalysts, at roughly the
same rate. This is likely due to the loss of Brønsted acid sites that are responsible for aromatics
formation [7, 22, 41]. As benzene selectivity decreases, more of the products are attributable to the
formation of coke, as shown below for the TPO results. The 5% Mo-SH catalyst seems to have
more selective sites than 1% Mo-SH, but they are catalytically very similar, at least as measured
by benzene selectivity.
1
1
1
4
2
0
8
6
4
2
1
5
% Mo-SH
% Mo-SH
0
100
200
300
400
500
600
700
800
900
1000
TIME (MIN)
Figure 8: Reaction data of different loadings of Mo-SH catalysts for conversion of methane (650⁰C, 1 atm, 600
ml/gcat-hr)

70
60
50
40
30
20
10
1
5
% Mo-SH
% Mo-SH
0
100
200
300
400
500
600
700
800
900
1000
TIME (MIN)
Figure 9: Reaction data of different loadings of Mo-SH catalysts for Ethylene selectivity (650⁰C, 1 atm, 600 ml/gcat-
hr)
60
50
40
30
20
10
0
1
5
% Mo-SH
% Mo-SH
0
100
200
300
400
500
600
700
800
900 1000
TIME (MIN)
Figure 10: Reaction data of different loadings of Mo-SH catalysts for benzene selectivity (650⁰C, 1 atm, 600 ml/gcat-
hr)

Effect of temperature
The effect of temperature on methane conversion on 5% Mo-SZ catalyst for MDHA is shown in
Figures 11-13. Three different temperature reactions (600ºC, 650ºC and 700ºC) were run at
otherwise the same reaction conditions.
Fig. 11 shows that methane conversion increased expectedly with higher temperature. A maximum
of 21% conversion of methane was observed at 700ᵒC, which includes carbon formation on
catalyst bed.
The methane conversion increases for the three temperatures studied here, and decreased with time
at all three temperatures, reaching a near-steady state conversion 4%-7%. As discussed above,
deactivation is attributable to both coke deposition and loss of surface sulfates.
Ethylene and benzene selectivity were measured as a function of time at each temperature. For all
three temperatures, ethylene selectivity generally increased with time but the result at 700⁰C was
different. Unlike the monotonic increase with ethylene selectivity for 600⁰C and 650⁰C, at 700⁰C
the ethylene initially increased rapidly, then decreases more slowly, reaching a maximum at ~350
mins. These results show that at all three temperatures, the ethylene selectivity appears to
approaching a common steady state value of ~45% at 900 min. As discussed above, this can be
attributed to the gradual Mo carbide formation of these active sites.
Benzene selectivity decreased rapidly with time for all three temperatures, but much more rapidly
at 700⁰C than at 600⁰C and 650⁰C. Selectivity as high as 65%, at 600⁰C, and that was at an initial
value. As discussed above, deactivation may be due to both coke and the loss of sulfate sites form
SZ surface at a temperature higher than 630ᵒC-650ᵒC[45]. The loss of sulfate sites result in a

corresponding loss of acid sites for the aromatization of the C
available Mo carbide/oxycarbide species.
2
y
H dimers that are generated by the
25
20
15
10
5
0
6
6
7
00 degC
50 degC
00 degC
0
100
200
300
400
500
600
700
800
900
TIME (MIN)
Figure 11: Reaction data for methane conversion of MDHA over 5% Mo-SH catalyst (600⁰C-700⁰C, 1 atm, 600
ml/gcat-hr)

70
60
50
40
30
20
10
0
6
6
7
00 degC
50 degC
00 degC
0
100
200
300
400
500
600
700
800
900
TIME (MIN)
Figure 12: Reaction data for ethylene selectivity in MDHA over 5% Mo-SH catalyst (600⁰C-700⁰C, 1 atm, 600
ml/gcat-hr)
70
60
50
40
30
20
10
0
600 degC
50 degC
700 degC
6
0
100
200
300
400
500
600
700
800
900
TIME (MIN)
Figure 13: Reaction data for benzene selectivity in MDHA over 5% Mo-SH catalyst (600⁰C-700⁰C, 1 atm, 600
ml/gcat-hr)

Effect of space velocity
Effect of space velocity (SV) on methane conversion, ethylene selectivity, and benzene ethylene
selectivity for the 5% Mo-SH was studied using three different SVs: 600, 1200 and 1800 GHSV
(Figures 14-16). Methane conversion decreases with time at all SVs, and conversion decreases as
the space velocity increases, as expected. Ethylene selectivity was an anomaly for all three space
velocity runs. For both 1800 and 1200 SV, there is a maximum in selectivity with time, while for
6
00 SV, selectivity increases monotonically, reaching 54%, essentially the same maxima of that
of 1800 and 1200 SV. These results represent the relative rates of ethylene formation on the Mo
sites and that of the conversion of ethylene to both benzene and coke.
Benzene selectivity for all three SVs decreases with time, and decreases with higher SVs. At the
highest 1800 SV, benzene selectivity is lower while ethylene selectivity is higher at shorter times
(Fig. 15), reflecting the lower rate of benzene formation compared that of ethylene, at initial and
intermediate times on stream. At 600 SV, benzene selectivity is greater than 1200 or 1800,
suggesting that relatively less coke is selectively produced at 1200 or 1800 SV.