Understanding reaction processes for n-heptane over 10Mo2C/SZ catalyst

Article abstract of DOI:10.1016/j.cattod.2016.02.036

Supported carbided molybdena catalysts was prepared by treating MoO₃/sulfated zirconia in methane/hydrogen (1:4 volumetric ratio) at 923?K. Characterisation shows the oxide to be different from the carbide phase. Carburisation did not induce phase change of the support with the tetragonal phase remaining dominant but did results in some changes to the textural properties. Lewis acid site density was constant in transforming from oxide to carbide forms while Br?nsted acidity site densities was diminished by ca 10% to give a Br?nsted/Lewis density ratio 2.46 in the carbide form. The ratio of hydrogenation sites to (Br?nsted) acid sites on the carbidic form of the catalyst was 0.12. Increasing temperature and decreasing WHSV augmented heptane conversion but leads to multiple cracking. Analysis of the product distribution as a function of conversion or as a function of temperature implied that the reaction did not simply proceed via a single consecutive reaction pathway Conversion increased the research octane number (RON) due to of the increased fraction of pentane isomers.

Full text of DOI:10.1016/j.cattod.2016.02.036

G Model
CATTOD-10068; No. of Pages11
ARTICLE IN PRESS
Catalysis Today xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Understanding reaction processes for n-heptane over 10Mo C/SZ
2
catalyst
∗
F.F. Oloye, A.J. McCue, J.A. Anderson
Surface Chemistry and Catalysis Group, Materials and Chemical Engineering, School of Engineering, University of Aberdeen, AB24 3UE, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Supported carbided molybdena catalysts was prepared by treating MoO /sulfated zirconia in
3
Received 24 November 2015
Received in revised form 16 February 2016
Accepted 23 February 2016
Available online xxx
methane/hydrogen (1:4 volumetric ratio) at 923 K. Characterisation shows the oxide to be different from
the carbide phase. Carburisation did not induce phase change of the support with the tetragonal phase
remaining dominant but did results in some changes to the textural properties. Lewis acid site density
was constant in transforming from oxide to carbide forms while Brønsted acidity site densities was dimin-
ished by ca 10% to give a Brønsted/Lewis density ratio 2.46 in the carbide form. The ratio of hydrogenation
sites to (Brønsted) acid sites on the carbidic form of the catalyst was 0.12. Increasing temperature and
decreasing WHSV augmented heptane conversion but leads to multiple cracking. Analysis of the product
distribution as a function of conversion or as a function of temperature implied that the reaction did
not simply proceed via a single consecutive reaction pathway Conversion increased the research octane
number (RON) due to of the increased fraction of pentane isomers.
Keywords:
Hydroisomerisation
Molybdena
Carbide catalyst
Sulfated zirconia
©
2016 Elsevier B.V. All rights reserved.
1
. Introduction
dehydrogenation [12,13], hydrogenolysis of butane [14], methane
dehydroaromatisation [15], methane reforming [16,17], methane
capturing [18], methanol oxidation [19], alcohol synthesis [20,21],
alkyne metathesis [22], water-gas shift [23,24], hydroisomerization
[25,26] and hydrotreating [27].
The use of n-alkanes (>hexanes) as fuel is limited by their com-
bustion characteristics at low temperature, in particular, their low
octane rating. These characteristics can be improved by either
additives or chemical modification to the feed. The use of octane
enhancing additives had been phased out because of environ-
mental consideration [1]. The desire for cleaner fuel has led to
increased research activity to develop environmentally benign and
cost effective catalysts for processes such as hydroisomerisation,
which lead to octane number enhancement but without formation
of aromatics [2]. Transition metal carbides, particularly molyb-
denum carbide have received great attention following the work
of Levy and Boudart [3]. This is partly due to their noble metal
like properties [4–8]. Molybdenum carbide has unique physical
and chemical properties, which include mechanical hardness, ther-
mal stability, surface reactivity, tolerance to carbon deposition and
sulfur [9,10]. The carbide can be obtained using different carburi-
sation routes, which involve heat treatment in the presence of a
carbon-containing source. Molybdenum carbide, in either unsup-
ported or supported forms, shows good activity and selectivity
in many catalytic reactions, including hydrogenation of CO [11],
The three common reaction routes for hydroisomerisation of
alkanes over solid acid catalysts are; (1) hydride abstraction
over Lewis acid sites through a monomolecular mechanism that
includes a carbenium ion intermediate, (2) a carbonium intermedi-
ate formed by proton transfer from Brønsted acid sites (bimolecular
mechanism), and (3) a bifunctional mechanism, which involves
hydroisomerisation over a catalyst that has two types of sites,
hydrogenation-dehydrogenation sites and acid sites [28]. A bifunc-
tional mechanism is consistent with experimental data for many
types of catalyst and, when operational, produces more heptane
isomers compared to the former two which produce by-products
with varying amount of carbon atoms [28]. However, for optimal
product distribution over catalysts where a bifunctional mecha-
nism is operative, a balance must exist between the contribution
of acid and metal sites.
Galadima et al. [26] obtained an isomerisation selectivity of
≥85% for n-heptane in the temperature range 623 and 723 K over
supported molybdenum carbide. Multibranched isomers were the
major products with smaller (ca 10%) contributions from mono-
branched isomers, but did not provide a mechanism of the reaction.
Matsuda et al. carried out n-C7 conversion using partly reduced
∗ Corresponding author.
E-mail address: j.anderson@abdn.ac.uk (J.A. Anderson).
http://dx.doi.org/10.1016/j.cattod.2016.02.036
0
920-5861/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: F.F. Oloye, et al., Understanding reaction processes for n-heptane over 10Mo C/SZ catalyst, Catal.
2
Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.036

G Model
CATTOD-10068; No. of Pages11
ARTICLE IN PRESS
F.F. Oloye et al. / Catalysis Today xxx (2016) xxx–xxx
Table 1
2
MoO3 [12], and concluded that it had higher activity compared
with a typical bifunctional catalyst (e.g. 0.5 wt% Pt/USY). As the
product distribution was similar over both catalysts, they con-
cluded that isomerisation occurred via a bifunctional mechanism
Physical and chemical properties of 10MoO3/SZ and 10Mo2C/SZ.
Parameters
10MoO3/SZ
10Mo2C/SZ
−1
a
CO uptake (␮mol g
)
74.4
26
25.4
24
[
29–31]. Working at low conversion to ensure differential reactor
Surface area (m2 g−1)b
3
−1
)b
operation, Blekkan et al. obtained high selectivity to iso-heptane
over an oxygen modified molybdenum carbide [32]. However, by
increasing the pressure, and consequently the conversion within
the range ca 30–45%, they were still available to attain iso-C7
selectivities of around 80%. They rejected a mechanism involving
carbocation intermediates and instead, proposed metallocyclobu-
tane intermediates [32]. Blekkan et al. [32] however, found that
the major products were monobranched isomers, hence the mech-
anisms involved over catalysts used by Galadima et al. [26] and
Blekkan et al. [32] may differ. Detailed work had been reported
using bifunctional catalysts such as Pt/SZ but systematic analysis
of the mechanism of hydrocarbon conversion over MoO3 has been
limited [33]. To the best of our knowledge, no detailed studies on
Pore volume (cm
Pore size (nm)
g
0.01
2.56
1.41
24.11
6.01
2.20
0.03
4.89
1.04
26.13
5.43
2.20
b
Sulfur (wt%)c
Mo surface density (␳Mo atoms nm 2)b
Brønsted acid density (nm )
Lewis acid density (nm )
−
2
d
2
d
a
Based on CO chemisorption at 308 K.
Determined from N2 adsorption.
Determined from combustion analysis.
Based on Pyridine adsorption.
b
c
d
determined at 308 K in the pressure range of 1–400 Torr using a
Micromeritics ASAP 2020-C instrument. For the sample which was
previously passivated (10 Mo C/SZ), an in-situ pre-treatment was
2
the reaction pathways of n-C7 conversion over Mo C/SZ catalysts
performed (H2 at 723 K).
2
have been conducted. Since different catalysts may involve dif-
ferent reaction mechanisms, understanding the reaction processes
2.3. Catalytic activity measurement
related to upgrading heptane using low loading of Mo C on sulfated
2
zirconia becomes essential.
Conversion of heptane was carried out with a quartz reactor
(
1.0 cm i.d., 29 cm long) at atmospheric pressure in the presence of
−
1
hydrogen. The liquid flow rate of n-C7 was fixed at 0.5 ␮l min
.
2
. Experimental
In separate experiments, the effect of space velocity (WHSV) was
−
1
varied between 0.029 and 0.101 h by changing catalyst mass in
the range 0.2–0.6 g. Temperature was also independently varied in
the range 723–873 K. The eluent stream was sampled automatically
using an electrically activated ten-port multi-loop VICI Valco sam-
pling valve and samples later analysed using a Perkin-Elmer 8410
GC fitted with a 5.6 m long 15% squalene on chromosorb WHP col-
umn and FIDdetector. The conversion(%X), researchoctane number
2.1. Preparation of materials
Catalyst support (sulfated zirconium hydroxide, MEL chemicals)
was pre-calcined in air at 823 K for 3 h to obtain sulfated zirco-
nia (SZ). A solution of ammonium heptamolybdate tetrahydrate
(
[(NH ) Mo7O ·4H O], Fisher scientific, 1.840 g in 100 ml ultra-
4
6
24
2
pure water, 10 Mo wt%) was added to 9 g of SZ and stirred vigorously
for 2 h. Excess water was removed by rotary evaporation at 353 K
prior to overnight drying at 393 K. The resultant powder was heated
(RON) and selectivity (%S) were calculated as follow:
Xi − Xf
−1
%X =
100
in air at 823 K (50 ml min ) for 3 h and stored in a sample vial
as 10MoO /SZ. Carburisation was achieved in-situ by loading an
Xi
3
amount (0.1–0.6 g) of the previously prepared catalyst between two
quartz wool plugs into a quartz reactor tube. The sample was then
heated in a flow of H2 before switching to a 1:4 mixture of CH :H
i=5
ꢀ
Xn
RON =
RONn
Xi − Xf
4
2
f =7
at 923 K and maintaining this condition for 4 h. Carburised samples
were labelled as 10Mo C/SZ. In order to perform ex-situ character-
2
Xn
isation, carburised sample was passivated by flowing 5% O /He at
2
%S =
100
Xi − Xf
room temperature (298 K) for 1 h to circumvent bulk oxidation.
where, Xi is the initial amount (␮mol) of heptane before reaction, Xf
is the amount of heptane remaining after reaction, Xn is the amount
of the individual product and RONn is the research octane number
of the individual product.
2
.2. Characterisation techniques
Surface area and porosity was determined using a Micromeritics
Tristar instrument after sample pre-treatment in a N flow at 473 K
2
for 4 h. X-ray diffraction patterns were recorded on a Panalytical
X’Pert Pro diffractometer, employing Cu K␣ monochromatic radia-
tion. The patterns were collected at room temperature with steps of
3. Results and discussion
The textural parameters of the synthesised catalyst
(10Mo C/SZ) as well as the precursor (10MoO /SZ) were evaluated
◦ ◦
.02 using a range of 5–80 . Raman spectra were recorded using
0
2
3
a Renishaw spectrometer equipped with an invia Raman micro-
scope RE 02. 514 nm laser line was employed as the exciting source
and the measuring parameters were set as follow; accumulation 5,
exposure time 10 s and laser power of 50%.
from N2 physisorption isotherms. Both isotherms and pore size
distributions for 10MoO /SZ and 10Mo C/SZ were different (Fig. 1)
suggesting transformation during the carburisation step. The pore
size distributions are majorly in the mesoporous range. Hysteresis
3
2
The sulfur content was determined using a LECO CS744 analyser
using combustion and infrared detection of the evolved gases. The
morphology of the sample was investigated using Field emission
scanning electron microscope, under high-vacuum mode operated
at 10 kV. The sample was carbon coated to prevent reflection. The
strength, type and nature of acid sites were determined by pyri-
dine adsorption study using a Perkin Elmer FT100 series Fourier
Transform Infrared Spectrometer. CO chemisorption studies were
in the isotherm for 10Mo C/SZ was more pronounced than in
2
the precursor suggesting that mesoporosity had been enhanced
by carburisation. The results from the isotherm and pore size
distribution using BET equation and Barrett-Joyner Halenda (BJH)
equation, respectively are summarised in Table 1. The BET surface
area (SBET) decreased very slightly on transforming from the oxide
to carbide. This relatively insignificant decrease could be indicative
of changes to the interactions between carbide and the support
Please cite this article in press as: F.F. Oloye, et al., Understanding reaction processes for n-heptane over 10Mo C/SZ catalyst, Catal.
2
Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.036

G Model
CATTOD-10068; No. of Pages11
ARTICLE IN PRESS
F.F. Oloye et al. / Catalysis Today xxx (2016) xxx–xxx
3
Fig. 1. N2 adsorption–desorption isotherms of (a) 10MoO3/SZ and (b) 10Mo2C/SZ. Insert: Pore size distribution of (c) 10MoO3/SZ (d) 10Mo2C/SZ.
[
34] or merely to change in sample density. This may arise due to
of the two samples, the EDX mapping confirms that the two major
elements present were Zr and Mo.
the loss of sulfate species during carburisation, therefore reducing
competition for zirconium sites. The apparent Mo surface density
was calculated using the equation below [13,34].
Mo features were also detected by Raman spectroscopy (Fig. 4).
Bands were observed at 115, 126, 147, 156, 196, 218, 243, 281, 290,
ꢁ
ꢂ
atoms
mol
−1
%
wt∗Avogadrosconstant
337, 378, 465, 640, 664, 750, 820, 943, 983 and 994 cm in the spec-
␳Mo =
ꢁ
ꢂ
g
∗BETSurfacearea(nm2/g)
trum of the oxide as well as that of the carbide but with significant
differences in intensity. Peaks at 117, 129, 158, 197, 217, 246, 283,
Momolecularweight
mol
The density increased in proportion to the loss in BET surface
area, consistent with expectation based on the use of such a calcu-
lation.
−
1
2
91, 338, 366, 378, 471, 666, 819 and 995 cm have been assigned
to different molybdenum oxide species [39] and many of these are
consistent with those observed here. Generally, bands located at ca
The XRD pattern for both the catalyst and the precursor are
−
1
◦
◦
◦
820 cm correspond to bridging Mo-O-Mo bonds and peaks at ca
shown in Fig. 2. The diffraction lines at 2ꢀ = 30.3 , 34.5 , 50.5 ,
−
1
◦
◦
◦
◦
910–980 cm are assigned to terminal Mo O in two dimensional
polymolybdates [19,35,39,40]. It is widely reported in the literature
that Raman spectra of monoclinic zirconia show features 180, 188,
5
0
2
9.1 60.5 , 63.1 and 74.8 , correspond to tetragonal zirconia (ICDD
1-079-1769 and 01-079-1764) [25]. However, additional lines at
◦
◦
◦
ꢀ = 23.6 , 25.8 and 27.5 were observed for the precursor (insert
−
1
2
21, 307, 335, 380, 476, 538, 560 and 617 cm while tetragonal
Fig. 2), and these additional peaks can be ascribed to MoO [35].
Peaks at 2ꢀ = 12.8 , 23.3 , 25.7 , 27.5 and 39.1 , correspond to
reflections from the (020), (110), (040), (021) and (131) planes of
MoO3 (ICDD 01-074-7382). The diffraction lines at, 23.6 , 25.8 ,
and 27.5 for the precursor disappeared following carburisation
and a new peak at 40.6 appeared. This new peak may be due to
ZrC-MoC (i.e., carburisation of both ZrO and Mo). Diffraction lines
at 2ꢀ = 23.9 , 30.6 and 50.1 associated with Zr(MoO4)2 hexago-
nal phase [36] were not observed here but may have overlapped
with features due to ZrO . The disappearance of diffraction lines
in the pattern for the oxide and appearance of other peaks in the
pattern for the carbide suggest formation of a new material as a con-
sequence of the carburisation process. Peaks at 2ꢀ = 34.5 , 37.9 ,
3
◦
◦
◦
◦
◦
zirconia has characteristic features at 148, 263, 325, 472 608 and
−
1
−
1
6
40 cm [41,42]. Hence, peaks here at 147 and 640 cm may be
◦
◦
assigned to tetragonal zirconia while other peaks are due to MoO3
related species. The intensity of all the bands decreased following
conversion of the oxide (Fig. 4a) to carbide (Fig. 4b). The band due
◦
◦
−
1
to bridging Mo
feature in the spectrum of the precursor was almost absent in spec-
O Mo species (820 cm ) which was the strongest
2
◦
◦
◦
−1
tra of 10Mo C/SZ, The high intensity of the 820 cm band in the
2
precursor state may be due to interaction of the oxide with the sur-
face of the support or with residual moisture. It has been reported
that mono or polymolybdate species often distort irregularly due
to their interaction with surface of the support [35].
Acid sites for 10MoO3/SZ and 10Mo2C/SZ were probed by
pyridine adsorption using FTIR spectroscopy. The samples were
2
◦
◦
◦
◦
◦
◦
3
(
9.5 , 52.2 , 61.6 and 69.6 due to reflections at (100), (002), (101),
102), (110) and (103) planes are indicative of ␤-Mo C [37,38].
2
activated in-situ for 1 h in air or in H at 773 K for oxide and carbide,
2
While some of these peaks are absent (possibly due to overlap with
reflections associated with zirconia), a peak at 40.6 may indicate
−
6
◦
respectively and then outgassed under vacuum (2 × 10 mbar)
for 1 h. Fig. 5 shows the spectra obtained after adsorption and
evacuation at 373 K. The resulting absorbance bands indicate
only chemisorbed pyridine molecules were retained at this stage.
Spectra reveal bands characteristic of coordinative adsorption of
the presence of carbide. However, it is possible that the carbide
phase formed was less crystalline than the parent oxide phase and
as such was not readily observed by XRD.
Fig. 3 shows the SEM imaging along with EDX profile. There
appears to be relatively little difference between the morphology
−
1
pyridine on Lewis acid sites (1610 and 1448 cm ) and of the
Please cite this article in press as: F.F. Oloye, et al., Understanding reaction processes for n-heptane over 10Mo C/SZ catalyst, Catal.
2
Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.036

G Model
CATTOD-10068; No. of Pages11
ARTICLE IN PRESS
4
F.F. Oloye et al. / Catalysis Today xxx (2016) xxx–xxx
Fig. 2. XRD pattern of (a) 10MoO3/SZ and (b) 10Mo2C/SZ.
pyridinium ion formed by interaction with Brønsted acid sites
the (Brønsted) acid site density may be closely related to the num-
ber of sulfate groups on the zirconia support. Note however that
this assumes a constant value of the 1.8 cm ␮mol for pyridine at
Brønsted acid sites while it has been reported that changes in sul-
fate loading may lead to variation in the value of this parameter
(see ref 46 and references within).
−1
−1
(
1638 and 1537 cm ) [43]. The absorption bands at1488 cm
−
1
corresponds to a vibration common to both Brønsted and Lewis
bound pyridine [43,44]. On comparing the two spectra, absorption
−
1
bands at 1537 and 1448 cm for the carbided sample appear more
intense than that for the oxide. However, difference in intensities
may be related to the mass of disc employed and/or be a con-
sequence of the carburisation process and/or sulfur content. The
amount of sulfur in the precursor (1.41%) was greater than the
amount contained in the catalyst (1.04%). Triwahyono et al. [44]
Fig. 6, shows CO isotherms on 10MoO /SZ and10Mo C/SZ with
3
2
the derived uptake data shown in Table 1. Both reach maxima
at ca 200 mmHg. CO uptake by the oxide was nearly three times
that of the carbide. This is almost certainly a reflection on the
changing adsorption stoichiometry rather than an indication of
changes in dispersion. Molybdenum in an oxidised state may form
dimeric and trimeric carbonyl species depending upon the oxida-
tion state. A correlation coefficient of 0.999 was obtained on fitting
the isotherm data to Langmuir equation for both oxide and carbide.
observed differences in total acidity on comparing MoO -ZrO2 and
3
Pt/MoO -ZrO and concluded that the difference was not due to Mo
3
2
content since they had similar Mo loading but may be due to mod-
ifications to the crystallinity of ZrO2 during calcination in which
the tetragonal phase decreased with addition of Pt [44]. Bachiller-
Baeza and Anderson [43] also observed difference in the intensity of
−
1
The Langmuir constant was 0.1808 mmHg for CO on the oxide
−
1
−1
absorbance bands at 1540 cm for heteropoly acid (HPA)/SiZr and
SiZr with the former having higher intensity and concluded that
the density and strength of the Lewis and Brønsted sites depend
on factors such as preparation conditions and Si/Zr ratio [43]. The
carburisation process affected the surface area, sulfur content, and
phases of the catalysts, all of which may have affected the acidity of
the catalyst. To understand the effect of carburisation on the acidity
of the catalyst, quantification of the amount adsorbed derived from
but 0.1477 mmHg on the carbide showing that the strength of
adsorption on the oxide was greater than on the carbide.
Most catalyst suffers from deactivation, often during the first
few minutes of reaction. To determine the stability of the syn-
thesised catalyst, the reaction was monitored during the first 3 h
on stream. It can be seen (Fig. 7) that the activity remains stable
throughout this period of reaction. Previous studies of carbided
molybdena supported on sulfated zirconia have been reported to be
stable at 723 K for hydroisomerisation of n-heptane [26]. It has been
widely reported that sulfated zirconia catalyses the conversion
of butane and other alkanes at low temperature with a reason-
able selectivity to isomers, but suffers from deactivation [48–50].
Yashima et al. [51] had associated the stability of H␤ zeolite to
the medium acid strength of the catalyst and the three dimen-
sional pore structures. The availability of appropriately sized pores
−
1
the absorbance bands at 1537 and 1448 cm was performed and
converted to acid site density using the molar absorption coeffi-
−
1
cients of 1.8 and 1.5 cm ␮mol for pyridine at Brønsted and Lewis
acid sites, respectively [45,46]. The carburisation has no effect on
the Lewis site density, but affected the Brønsted acid site density
(Table 1). Davis et al. [47] have shown that there was a decrease in
Brønsted-Lewis ratio as the sample is reactivated at high temper-
ature for Pt promoted SZ. The reduction in the Brønsted acid site
density may, in part, be related to sulfur loss. However, Blekkan
et al. [32] report that MoO3 also contributes to surface acidity,
therefore, acidity may be generated on a non-support phase. The
relationship between the number of Brønsted acid sites formed and
the quantity of retained sulfur was 0.6 sites: S atom and 0.7 sites: S
atom in the oxidic and carbidic forms of the sample suggesting that
in 10Mo C/SZ may lower its sensitivity to coke laydown. García-
2
Gutiérrez et al. [33] reported that MoO required an induction time
3
before reaching steady state. However, this induction period has
been associated with the time required to form the active carbide
phase. The catalyst was stable after reaching steady state [33]. A
similar observation was observed for n-heptane over an oxygen
modified molybdenum carbide catalyst [32]. It is assumed here
Please cite this article in press as: F.F. Oloye, et al., Understanding reaction processes for n-heptane over 10Mo C/SZ catalyst, Catal.
2
Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.036

G Model
CATTOD-10068; No. of Pages11
ARTICLE IN PRESS
F.F. Oloye et al. / Catalysis Today xxx (2016) xxx–xxx
5
Fig. 3. SEM profile of (a) 10MoO3/SZ and (b) 10Mo2C/SZ. EDX profile of (c) 10MoO3/SZ (d) 10Mo2C/SZ.
Table 2
version up to maximum irrespective of operating parameter and
then consumed to produce other products [32]. At low conversion,
ethane and pentane isomers (C2 and iC5) were the dominant prod-
ucts followed by propane (C3), iso-heptane (iC7) and iso-butane
(iC4). Blomsma et al., reported that selectivity to cracked products
increased as conversion was increased with conversion between
The effect of varying space velocity on conversion and RON at 823 K and hydrogen
flowrate of 18 ml/min.
WHSV/h 1
−
Conversion (%)
RON
0.101
0.068
0.051
0.041
0.029
22
31
39
45
50
30
30
37
66
61
1
0 and 50% over 0.1 to1 Pd loading on H-beta [54]. In a similar
report, they observed that butane and C3 selectivity dominates and
that pentane and hexane (C5 and C6) were formed only in small
quantities while C2 and methane were absent [55]. However, C4
selectivity was in excess of the C3, which they explained to be as a
result of other reaction routes which differ from the classic hydro-
cracking pathway, where equal amounst of C3 and C4 are expected
that the carburisation process overcame the need to activate the
catalysts in the reagent stream.
The effect of the space velocity (WHSV) on conversion and RON
is summarised in Table 2. As expected, the conversion decreased
with an increase in WHSV (decrease in contact time) while RON
increased with conversion, as previously observed [52]. Consistent
with expectation [53], low space velocity gives high conversion,
but there is higher probability of formation of secondary products
which arise from cracking [47].
Fig. 8 shows the selectivity as a function of conversion while
varying WHSV. In the classical view of isomerisation of hydro-
carbon over typical bifunctional catalysts, the yield of isomers is
a function of the conversion [32]. The isomer yield follows con-
[
55]. Here, C2 and iC5 were the major products at the lowest con-
version, with C2 in excess of iC5, and thus it could be said that C2
formation occurs via more than just one route. García-Gutiérrez
et al. [33] also observed hydrocracked isomers at low conversion
which increased with conversion. They concluded in their study
of selectivity of C6-C7 hydroisomerisation over MoO3 that n-C7
hydroisomerisation-hydrocracking were not only taking place as
consecutive reactions, but also as parallel reactions and that hydroi-
somerisation of n-C7 leads to predominantly C4, C3 and C2 in a
mole ratio of 1:2.2:2 [33]. Some authors have used dimerization
Please cite this article in press as: F.F. Oloye, et al., Understanding reaction processes for n-heptane over 10Mo C/SZ catalyst, Catal.
2
Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.036

G Model
CATTOD-10068; No. of Pages11
ARTICLE IN PRESS
6
F.F. Oloye et al. / Catalysis Today xxx (2016) xxx–xxx
Fig. 4. Raman spectra of (a) 10MoO3/SZ and (b) 10Mo2C/SZ.
Fig. 5. FTIR spectra of (a) 10MoO3/SZ and (b) 10Mo2C/SZ following exposure to pyridine followed by evacuation at 373 K for 1 h.
and Brønsted acid sites (216 ␮mol g 1 pyridine retained). The ratio
of metal-like to Brønsted acid sites was 0.12, which falls in the
range of 0.03 and 0.17 that were reported to be indicative of low
and high hydrogenating/acid site density ratios, respectively, for
hydroisomerisation of n-alkanes over PtHFAU [56]. The challenge
in manipulating this ratio for this type of catalysts is that MoO3
itself may also contributes to surface acidity [32] and data here
for acid sites shows no simple relationship with the amount of
sulfur present (Table 1). It had been reported that catalyst with
high degree of hydrogenation activity and low degree of acidity
are best for maximising the extent of hydroisomerisation versus
−
cracking mechanisms to explain this kind of hydrocracked product
distribution from n-C7 [55]. An ideal hydroisomerisation reaction
is believed to occur on a bifunctional catalyst with an appropri-
ate balance between acid and metal sites [26,52]. When n-C7 is
converted over a catalyst with a mismatch between acid/metals
sites, the cracked products ranged between C3 to C6, whereas
on a well-balanced catalyst, the cracked products is limited to
only C3 and C4 [55]. Hence, the range of hydrocracked product
obtained in this research may be because the ratio of metal site
to acid site was not fully optimised. Catalyst here contained both
−
1
metal-like hydrogenation sites (25.4 ␮mol g from CO adsorption)
Please cite this article in press as: F.F. Oloye, et al., Understanding reaction processes for n-heptane over 10Mo C/SZ catalyst, Catal.
2
Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.036

G Model
CATTOD-10068; No. of Pages11
ARTICLE IN PRESS
F.F. Oloye et al. / Catalysis Today xxx (2016) xxx–xxx
7
Fig. 6. Adsorption isotherms for CO on (a) 10MoO3/SZ and (b) 10Mo2C/SZ at 308 K.
Fig. 7. Plot of nC7 conversion as a function of time on stream reaction over 0.2 g 10Mo2C-SZ at 873 K with a hydrogen flow of 9 ml/min showing the catalyst stability.
hydrocracking [57] and for avoiding deactivation due to coking [54].
The lack of deactivation through coking (Fig. 7) would suggest that
a significant imbalance with excess acid sites was not operative.
However, these ratios also fail to distinguish between the relative
rates on these sites and so the relatively high density of (Brønsted)
acid sites (216 ␮mol g ) compared to carbide sites (25.4 ␮mol g
CO adsorbed) may not fully explain the low yield of the heptane
isomers.
Others researchers have concluded that both monomolecular
and bimolecular mechanisms contribute to the formation of C3–C6
products and branched isomers [54]. Iglesia et al. [53] observed
oligomerisation-cracking for n-alkanes and that hydride transfer
limits the reaction rates. Direct cracking of n-heptane can occur
by ␤-scission of the reactive carbenium ion intermediate. Also, iso-
merisation occurs by direct reaction of alkanes on acid sites assisted
by the presence of metal sites or of hydride transfer agents [53]. This
infers that both hydrocracking and hydroisomerisation products
are derived from the same intermediate, a protonated cyclopropane
[58,59]. This is illustrated in Schemes 1 and 2 and suggests that both
cracking and isomerisation occur simultaneously. The presence of
C2, C3, iC4, C4, iC5, iC6 and iC7 in reaction products here suggest
that the reaction routes are consistent with the steps described in
Schemes 1 and 2 and consistent with the observations of others
researchers [33,53,54,58].
−
1
−1
Please cite this article in press as: F.F. Oloye, et al., Understanding reaction processes for n-heptane over 10Mo C/SZ catalyst, Catal.
2
Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.036

G Model
CATTOD-10068; No. of Pages11
ARTICLE IN PRESS
F.F. Oloye et al. / Catalysis Today xxx (2016) xxx–xxx
8
Fig. 8. The effect of varying space velocity on conversion and selectivity at 823 K and 18 ml min−1. Insertion shows selectivity to products below 10% (solid circle, ethane;
solid upward triangle, propane; open downward triangle, isobutane; solid diamond, butane; open square, isopentane; solid cross, pentane; open hexagon, isohexane; solid
star, hexane; open pentagon, isoheptane).
Scheme 1. Proposed reaction pathway for formation of ethane and isopentane from n-heptane conversion.
Please cite this article in press as: F.F. Oloye, et al., Understanding reaction processes for n-heptane over 10Mo C/SZ catalyst, Catal.
2
Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.036

G Model
CATTOD-10068; No. of Pages11
ARTICLE IN PRESS
F.F. Oloye et al. / Catalysis Today xxx (2016) xxx–xxx
9
Scheme 2. Proposed reaction pathway for formation of propane and isobutane from n-heptane.
Scheme 3. Proposed reaction pathway for direct formation of hydrocracked products from n-heptane.
Table 3
Relatively minor amounts of C5 and C6 were formed suggesting
that dimerisation did not occur readily. This is also supported by
the observation that any butadecane (C14) obtained from heptane
dimerisation should also give rise to C5 and nonane (C9) and/or C6
and octane (C8). No C9 or C8 products were detected. Surprisingly,
the amount of iC5 increased with increased conversion while C2
was depleted above ca 35% conversion. This increase in iC5 selectiv-
ity may be explained either by cracking of dimerisation products or
by coupling of ethene with propene then hydrogenated. Formation
of such olefins via dehydrogenation become more thermodynam-
ically favoured at higher temperatures [60]. Lower alkanes can
readily be hydroisomerised to their corresponding isomers, so C5
may isomerise to iC5 thereby leading to increase of iC5 at the
expense of C2, C3 and C5. Oligomerisation of C2 to form a surface
bound product may also explain the loss in C2 for conditions above
−
1
The effect of temperature on conversion and RON at WHSV of 0.101 h and hydro-
gen flowrate of 18 ml/min.
Temperature/K
Conversion/%
RON
723
823
9
22
44
26
30
46
873
of C2/iC5 and C3/iC4 were far greater than their equilibrium pre-
dicted values [33] because of the possibility of iC5 cracking to give
C3 and C2. Scheme 1, route f shows the formation of C3 from C5.
Similarly, from Scheme 3 routes a and b, it is possible to form C2–C5.
Therefore, one cannot rule out C2, C3, iC4 and iC5 as part of the pri-
mary products because they were all detected, even at low levels of
heptane conversion. Since iC7 was present at low conversion then
decreased with conversion, it might be concluded that iC7 was one
of the primary products. The increase in iC5 as a function of conver-
sion, led to an increase in RON, since the RON was estimated based
on the fractions of C5–C7 in the products.
The relationship between n-heptane conversion and RON yield
with reaction temperature is shown in Table 3. Both RON and
conversion increased with temperature. Similar results have been
reported by other researchers [28,51,52]. The RON increases with
conversion due to the increase in iC5. iC5 selectivity is increased
at high conversion (high temperature) as there is increased proba-
bility of the olefins produced by the metal site reacting on the acid
3
5% conversion, although this would be expected to lead to some
form of deactivation. The stability test (Fig. 7) was conducted at
ca 48% conversion, where C2 product is expected to be consumed
(Fig. 8) yet did not provide evidence of deactivation (and, it might
be assumed, formation of significant amounts of oligomer based
product). Carbonaceous deposits could occur on a catalyst through
oligomerisation and/or alkylation, and these products could crack
to form lower alkanes such as iC5 [33], although one might expect
some evidence of deactivation should carbonaceous laydown be
significant and this was not observed (Fig. 7). At low conversion,
selectivity to C2 and iC5 was greater than for C3 and iC4. The ratios
Please cite this article in press as: F.F. Oloye, et al., Understanding reaction processes for n-heptane over 10Mo C/SZ catalyst, Catal.
2
Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.036

G Model
CATTOD-10068; No. of Pages11
ARTICLE IN PRESS
F.F. Oloye et al. / Catalysis Today xxx (2016) xxx–xxx
1
0
−
1
Fig. 9. The effect of varying temperature on selectivity and conversion at WHSV of 0.102 h and hydrogen flowrate of 18 ml/min. (solid square, methane; solid circle, ethane;
solid upward triangle, propane; open downward triangle, isobutane; solid diamond, butane; open square, isopentane; solid cross, pentane; open hexagon, isohexane; solid
star, hexane; open pentagon, isoheptane).
sites rather than being hydrogenated to form the corresponding
alkane [60]. The formation of olefins is less likely at low temper-
ature due to the relatively slower dehydrogenation rates over the
metal site [60]. Hence, the difference in the iC5 and C2 selectivity
at low and high temperature could be associated with the rela-
tive amount of olefins generated. Giannetto et al. [61] observed C3
and iC4 as the main cracked products from C7 hydrocracking. The
selectivity to C3 and iC4 decreased from 45 and 40.6 to 10.4 and 9.5,
respectively, when the Pt loading was increased from 0.1 to 1.5 wt%
over Pt/HY3 [61]. This suggests that in Schemes 1 and 2, routes d
are more favourable when the number of available metal sites is
low. Equimolar amounts of C3 and C4 and a high iC4/nC4 ratio over
Pt/H␤ [62] was interpreted as bifunctional hydrocracking. The reac-
tion pathway followed route a in Scheme 3. As equimolar amounts
of C3 and C4 were not observed here (Fig. 9), it can be assumed that
n-C7 conversion did not follow direct bifunctional hydrocracking,
but also included a secondary pathway leading to cracking of iC7.
The product distribution resulting for temperature (Fig. 9) is similar
to that of varying space velocity (Fig. 8), showing that iC5 selec-
tivity increased with conversion while iC7, C2 and iC6 decreased.
Busto et al. [52] observed a decrease in the yield of iso-alkane—with
increased temperature. Galadima et al. [26] reported a slight loss
in hydroisomerisation selectivity due to secondary reactions when
the temperature was varied between 623 and 723 K. However, they
obtained a significant decrease in hydroisomerisation selectivity
when 2-methylheptane was reacted over carbided molybdena on
sulfated zirconia with isomer selectivity low at 673 K, dropping
to zero at 723 K. This provides an indication of the relative ease
with which alkane isomers are cracked. This suggests a key role of
route f in Scheme 1. Blekkan et al. [32] also reported decreases in
isomer selectivity as the conversion increased. Therefore, hydroc-
racked products (iC5, iC4, C3, C2) as well as heptane isomer were
formed as primary reaction products, but heptane isomers readily
crack to give hydrocracked product, thereby increasing iC5 selec-
tivity.
4. Conclusion
No induction period was observed over 10Mo C/SZ due to car-
2
burisation before reaction. n-C7 conversion occurred through a
combination of consecutive and parallel reactions. C2 and iC5 were
the major products at low conversion, but were not formed in
equimolar quantities. Selectivity to C3 was greater than to iC4. The
formation of more C2 than iC5 and C3 than iC4 confirmed that the
reaction followed additional pathways rather than just consecutive
reaction routes. As conversion increased, iC5 selectivity increased,
leading to increased RON. The high selectivity of the catalyst to C2
and C3 suggests a relationship between carbide (hydrogenation)
and acid (isomerisation) sites were not optimised in this sample
although no deactivation was observed, suggesting that the rela-
tionship did not lead to undesired oligomer or coked products. Also,
high iC5 selectivity at high conversion may arise as a consequence
−
1
of the relative amounts of the metal-like sites (25.4 ␮mol g ) and
Brønsted acid sites (216 ␮mol g ).
−
1
Acknowledgements
We are grateful to Tertiary Education Trust Fund (TET-
FUND)Nigeria and Adekunle Ajasin University Nigeria for
studentship (to FFO). We would like to thank Mr John Still and
Mr Colin Taylor, University of Aberdeen for their kind support and
assistance with SEM and sulfur analysis, respectively.
a
References
[1] M. Busto, V. Benítez, C. Vera, J. Grau, J. Yori, Appl. Catal. A: Gen. 347 (2008) 117.
[
2] A.S. Khder, E.A. El-Sharkawy, S.A. El-Hakam, A.I. Ahmed, Catal. Commun. 9
2008) 769.
3] R. Levy, M. Boudart, Science 181 (1973) 547.
(
[
[4] S.T. Oyama, J. S. S. Chem. 96 (1992) 442.
[
5] M.J. Ledoux, C. Pham-Huu, A.P.E. York, E.A. Blekkan, P. Delporte, P. Del Gallo,
In: Oyama S.T. (Ed.) Blackie, New York (1996) 373.
[
6] S.T. Oyama, Catal. Today 15 (1992) 179.
[
7] J. Lee, S. Oyama, M. Boudart, J. Catal. 106 (1987) 125.
Please cite this article in press as: F.F. Oloye, et al., Understanding reaction processes for n-heptane over 10Mo C/SZ catalyst, Catal.
2
Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.036

G Model
CATTOD-10068; No. of Pages11
ARTICLE IN PRESS
F.F. Oloye et al. / Catalysis Today xxx (2016) xxx–xxx
11
[
[
8] X. Wang, H. Hao, M. Zhang, W. Li, K. Tao, J. S. S. Chem. 179 (2006) 538.
9] C. Liu, M. Lin, K. Fang, Y. Meng, Y. Sun, RSC Adv. 4 (2014) 20948.
[35] K.V.R. Chary, K.R. Reddy, G. Kishan, J.W. Niemantsverdriet, G. Mestl, J. Catal.
226 (2004) 283.
[
10] L. Leclercq, M. Provost, H. Pastor, J. Grimblot, A. Hardy, L. Gengembre, G.
[36] J.R. Sohn, M.Y. Park, Langmuir 14 (1998) 6140.
[37] P. Heo, M. Nagao, M. Sano, T. Hibino, J. Electrochem. Soc. 154 (2007) B53.
Leclercq, J. Catal. 117 (1989) 371.
[
[
[
[
[
11] T. Mo, J. Xu, Y. Yang, Y. Li, Catal. Today 261 (2016) 101–115.
12] T. Matsuda, H. Sakagami, N. Takahashi, Appl. Catal. A: Gen. 213 (2001) 83.
13] K. Chen, S. Xie, E. Iglesia, A.T. Bell, J. Catal. 189 (2000) 421.
14] J.S. Lee, S. Locatelli, S.T. Oyama, M. Boudart, J. Catal. 125 (1990) 157.
15] T.E. Tshabalala, N.J. Coville, J.A. Anderson, M.S. Scurrell, Appl. Catal. A: Gen.
[38] W. Wu, Z. Wu, C. Liang, X. Chen, P. Ying, C. Li, Phys. Chem. B 107 (2003) 7088.
[39] L. Seguin, M. Figlarz, R. Cavagnat, J. Lassègues, Spectrochim. Acta Part A: Mol.
Biomol. Spec. 51 (1995) 1323.
[40] J.R. Sohn, E.W. Chun, Y.I. Paet, Bull. Korean Chem. Soc. 24 (2003) 1785.
[41] J. Miciukiewicz, T. Mang, H. Knözinger, Appl. Catal. A: Gen. 122 (1995) 151.
[42] S. Xie, K. Chen, A.T. Bell, E. Iglesia, J. Phys. Chem. B 104 (2000) 10059.
[43] B. Bachiller-Baeza, J.A. Anderson, J. Catal. 228 (2004) 225.
[44] S. Triwahyono, A.A. Jalil, N.N. Ruslan, H.D. Setiabudi, N.H.N. Kamarudin, J.
Catal. 303 (2013) 50.
5
03 (2015) 218.
[
[
16] K. Oshikawa, M. Nagai, S. Omi, J. Phy. Chem. B 105 (2001) 9124.
17] A.J. Brungs, A.P. York, J.B. Claridge, C. Márquez-Alvarez, M.L. Green Catal,
Letters 70 (2000) 117.
[
18] S. Posada-Pérez, dos Santos Politi, José Roberto, F. Vi n˜ es, F. Illas, RSC Adv. 5
[45] S. Khabtou, T. Chevreau, J.C. Lavalley, Microporous Mater. 3 (1994) 133.
[46] A.J. McCue, G.A. Mutch, A.I. McNab, S. Campbell, J.A. Anderson, Catal. Today
259 (2015) 19–26.
[47] B.H. Davis, R.A. Keogh, S. Alerasool, D.J. Zalewski, D.E. Day, P.K. Doolin, J. Catal.
183 (1999) 45.
(
2015) 33737.
[19] A.S.C. Brown, J.S.J. Hargreaves, S.H. Taylor, Catal. Lett. 57 (1999) 109.
[20] T. Tatsumi, A. Muramatsu, H. Tominaga, Appl. Catal. 34 (1987) 77.
[21] X. Li, L. Feng, Z. Liu, B. Zhong, D.B. Dadyburjor, E.L. Kugler, Ind. Eng. Chem. Res.
3
7 (1998) 3853.
[48] K. Arata, M. Hino, Appl. Catal. 59 (1990) 197.
[
[
[
[
[
[
22] A. Fürstner, C. Mathes, C.W. Lehmann, Chem. Eur. J. 7 (2001) 5299.
23] M. Saito, R.B. Anderson, J. Catal. 63 (1980) 438.
24] M. Łaniecki, M. Małecka-Grycz, F. Domka, Appl. Catal. A: Gen. 196 (2000) 293.
25] A. Galadima, R.K.P. Wells, J.A. Anderson, Top. Catal. 55 (2012) 931.
26] A. Galadima, R.P.K. Wells, J.A. Anderson, Appl. Petrochem. Res. 1 (2012) 35.
27] M. Vrinat, D. Hamon, M. Breysse, B. Durand, T. Des Courieres, Catal. Today 20
[49] M. Hino, K. Arata, J. C. S. Chem. Commun. (1979) 1148.
[50] K. Arata, H. Matsuhashi, M. Hino, H. Nakamura, Catal. Today 81 (2003) 17.
[51] T. Yashima, T. Komatsu, Z. Wang, A. Kamo, T. Yoneda, Appl. Catal. A: Gen. 159
(1997) 119.
[52] M. Busto, J. Grau, C. Vera, Appl. Catal. A: Gen. 387 (2010) 35.
[53] E. Iglesia, S.L. Soled, G.M. Kramer, J. Catal. 144 (1993) 238.
[54] E. Blomsma, J. Martens, P. Jacobs, J. Catal. 159 (1996) 323.
[55] E. Blomsma, J. Martens, P. Jacobs, J. Catal. 155 (1995) 141.
[56] M. Guisnet, Catal. Today 218–219 (2013) 123.
(
1994) 273.
[
[
28] Y. Ono, Catal. Today 81 (2003) 3.
29] T. Matsuda, T. Ohno, Y. Hiramatsu, Z. Li, H. Sakagami, N. Takahashi, Appl.
Catal. A: Gen. 362 (2009) 40.
[57] H. Deldari, Appl. Catal. A: Gen. 293 (2005) 1.
[
[
[
30] T. Matsuda, H. Sakagami, N. Takahashi, Catal. Today 81 (2003) 31.
31] T. Matsuda, H. Sakagami, N. Takahashi, Catal. Lett. 42 (1997) 51.
32] E.A. Blekkan, C. Pham-Huu, M.J. Ledoux, J. Guille, Ind. Eng. Chem. Res. 33
[58] S.T. Sie, Ind. Eng. Chem. Res. 32 (1993) 403.
[59] S.T. Sie, Ind. Eng. Chem. Res. 31 (1992) 1881.
[60] M. Busto, C. Vera, J. Grau, Fuel Process. Technol. 92 (2011) 1675.
[61] G.E. Giannetto, G.R. Perot, M.R. Guisnet, Ind. Eng. Chem. Prod. Res. Dev. 25
(1986) 481.
[62] Z. Wang, A. Kamo, T. Yoneda, T. Komatsu, T. Yashima, Appl. Catal. A: Gen. 159
(1997) 119.
(
1994) 1657.
33] J.L. García-Gutiérrez, J.L. Cano, L.O. Alemán-Vázquez, F. Jiménez-Cruz, Fuel 94
2012) 532.
34] E.A. El-Sharkawy, A.S. Khder, A.I. Ahmed, Microporous Mesoporous Mater.
02 (2007) 128.
[
[
(
1
Please cite this article in press as: F.F. Oloye, et al., Understanding reaction processes for n-heptane over 10Mo C/SZ catalyst, Catal.
2
Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.02.036