Preparation of unsupported NiMoP catalysts for 4,6-dimethyldibenzothiophene hydrodesulfurization

Article abstract of DOI:10.1007/s10562-014-1290-9

The preparation of unsupported NiMoP catalysts in the presence of citric acid (CA) is reported. Phase pure NiMoP with a surface area ～60 m ²/g was synthesized in the presence of CA at low reduction temperature (550 °C), with a Ni/Mo/P molar ratio of 1/1/1 and a 1.5 CA/Me ratio (Me = Ni + Mo). Depending on the synthesis conditions, small amounts of MoP and NiMoP₂ were also present in the catalysts. The catalyst precursors appeared similar to those identified for MoP and Ni₂P catalysts, also synthesized in the presence of CA. For the phase pure NiMoP catalysts, the TOFs for the HDS of 4,6-dimethyldibenzothiophene were almost identical, despite large differences in NiMoP surface area and crystallite size. Graphical Abstract: [Figure not available: see fulltext.]

Full text of DOI:10.1007/s10562-014-1290-9

Catal Lett
DOI 10.1007/s10562-014-1290-9
Preparation of Unsupported NiMoP Catalysts for 4,6-
Dimethyldibenzothiophene Hydrodesulfurization
•
Rui Wang Kevin J. Smith
Received: 7 March 2014 / Accepted: 27 May 2014
Ó Springer Science+Business Media New York 2014
Abstract The preparation of unsupported NiMoP cata-
lysts in the presence of citric acid (CA) is reported. Phase
pure NiMoP with a surface area *60 m2/g was synthe-
sized in the presence of CA at low reduction temperature
(550 °C), with a Ni/Mo/P molar ratio of 1/1/1 and a 1.5
CA/Me ratio (Me = Ni ? Mo). Depending on the syn-
thesis conditions, small amounts of MoP and NiMoP2 were
also present in the catalysts. The catalyst precursors
appeared similar to those identified for MoP and Ni2P
catalysts, also synthesized in the presence of CA. For the
phase pure NiMoP catalysts, the TOFs for the HDS of 4,6-
dimethyldibenzothiophene were almost identical, despite
large differences in NiMoP surface area and crystallite size.
(CA) to precursor metal salt solutions, prior to TPR [5, 6].
The citric acid acts as a chelating agent, yielding Ni or Mo
citrate complexes that control crystallization during calci-
nation and TPR [7].
Although Ni and Co are effective promoters of MoS2
hydrotreating catalysts [8–10] their impact is quite different
in the case of metal phosphides. Sun et al. [11] reported that
the dibenzothiophene HDS activity of phosphide catalysts
followed the order Ni2P/SiO2 [ NiMoP/SiO2 [MoP/SiO2,
with no synergistic effect observed between the Ni and Mo,
unlike with sulfides, nitrides, and carbides. Similar results
were reported for the hydrodeoxygenation (HDO) of anisole
by Li et al., wherein the HDO activities decreasedin the order
Ni2P/SiO2 [ NiMoP/SiO2 [ MoP/SiO2 [12]. Others have
reported enhanced activities of bimetallic phopshides but the
catalyst activity is very dependent upon composition [13].
For example, in a previous study of NixMoP catalysts, in
which Ni was added to MoP to give a Ni/Mo ratio from 0.125
to 1, the initial conversion of 4,6-dimethyldibenzothiophene
over NixMoP catalysts was lower than that of MoP, and this
was attributed to an increased crystallite size of MoP and
NiMoP [5, 6]. In contrast, the turnover frequency (TOF) of
the NixMoP catalysts increased linearly with increased Ni
content and the NixMoP catalysts were shown to be a mixture
of MoP and NiMoP.
Keywords Heterogeneous catalysis Á
Hydrodesulfurization Á Metal phosphide Á Catalyst
preparation Á Environmental catalysis
1 Introduction
Transition metal phosphides are viewed as an alternative
class of catalysts that have potential for ultra-deep hyd-
rodesulfurization (HDS) and hydrodenitrogenation (HDN)
of petroleum feedstocks [1–3]. Both unsupported and
supported metal phosphides can be prepared by tempera-
ture-programmed reduction (TPR) of the corresponding
metal phosphate salts [1, 4]. Recent studies have reported
on the synthesis of high surface area, unsupported mono-
metallic phosphides (MoP and Ni2P) by adding citric acid
In the present paper, we report on the effect of various
preparation conditions on the properties of unsupported
NiMoP catalysts. As before, the catalysts were prepared in
the presence of CA to obtain high surface area, but in this
case the Ni:Mo molar ratio was fixed at 1:1. Variables
investigated include the amount of CA used in the prepa-
ration of the precursor salts, the precursor composition, and
the reduction temperature. The relationship between these
variables, the resultant NiMoP structure and the HDS
activity, is reported.
R. Wang Á K. J. Smith (&)
Department of Chemical and Biological Engineering, University
of British, Vancouver, BC V6T 1Z3, Canada
e-mail: kjs@mail.ubc.ca
123

R. Wang, K. J. Smith
2 Experimental
Table 1 List of preparation conditions used in the present study for
high surface area NiMoP synthesis
2.1 Catalyst Preparation
Sample
P/Me molar
ratio
CA/Me molar Reduction
ratio
temperature (°C)
The unsupported NiMoP catalysts were prepared by first
dissolving required amounts of (NH4)6Mo7O24Á4H2O,
(NH4)2HPO4 and citric acid in 15 ml of water. A second
solution was prepared by dissolving an appropriate amount
of Ni(NO3)2Á6H2O and citric acid in 15 ml of deionized
water. The two solutions were then mixed so that the final
solution contained the required CA/Me ratio (Me =
Mo ? Ni), and the synthesis yielded 2–5 g of catalyst
precursor. Various formulations of the precursor solutions
were investigated, but in all cases the Ni/Mo molar ratio
was 1. Catalysts denoted as NiMoP-xP, where x represents
the P/Me molar ratio, were prepared with P/Me ratios of 1,
0.75 and 0.5 and a fixed CA/Me molar ratio of 1. A second
series of catalysts, denoted as NiMoP-xCA, where x rep-
resents the CA/Me molar ratio, were prepared with CA/Me
ratios of 1.5, 1, 0.75 and 0.5 and a fixed P/Me molar ratio
of 1. In each case, the precursor solution was aged in a
covered beaker in a water bath held at 90 °C for 24 h. The
resulting gel was dried in an oven at 120 °C for 24 h and
this material is referred to as the dried catalyst precursor.
Subsequently the dried precursor was calcined at 500 °C
for 5 h, cooled and ground to a powder (dP \ 0.7 mm).
This material is referred to as the calcined catalyst pre-
cursor. The active metal phosphide was subsequently
obtained by temperature-programmed reduction (TPR) of
the calcined precursor in H2 (Praxair, 99.99 %) at a flow
rate of 160 ml(STP)/min. A heating rate of 5 °C/min to
300 °C, followed by a heating rate of 1 °C/min to the final
reduction temperature was used for the TPR, with the final
temperature held for 2.5 h. Final reduction temperatures of
500–650 °C were examined for the NiMoP-0.75CA cata-
lyst, whereas all other catalysts were reduced to 650 °C.
The reduced catalyst was cooled to room temperature in He
and passivated in a flow of 1 % O2/He for 3 h at room
temperature prior to removal from the reactor. This mate-
rial is referred to as the passivated catalyst. Table 1 lists the
catalyst compositions, calcination temperatures and
reduction temperatures examined in the synthesis of
unsupported NiMoP catalysts in the present study. For
comparison, we also prepared an unsupported NiMoP
catalyst without citric acid using the conventional TPR
method to reduce the corresponding phosphate precursor.
NiMoP-1P
1.0
1.0
1.0
650
650
NiMoP-
0.75P
0.75
NiMoP-
0.5P
0.5
0.5
0.5
0.5
0.5
0.5
1.0
1.5
1.0
0.75
0.5
0
650
650
650
650
650
650
NiMoP-
1.5CA
NiMoP-
1CA
NiMoP-
0.75CA
NiMoP-
0.5CA
NiMoP-
noCA
NiMoP-650 0.5
NiMoP-600 0.5
NiMoP-550 0.5
0.75
0.75
0.75
650
600
550
All catalysts calcined at 500 °C
were performed on the dried and calcined catalyst precur-
sors and the passivated catalysts using a Siemens D500
˚
diffractometer with Cu Ka X-rays (k = 1.54 A). Phase
identification was carried out after subtraction of the
background and the grain size of selected phases was cal-
culated from the diffractograms using the Scherrer equa-
tion with correction for instrument broadening.
Diffuse reflectance infrared Fourier-transform (DRIFT)
spectroscopy measurements were performed on a Thermo
Scientific Nicolet iS10 Fourier transform IR spectrometer
equipped with a MCT detector and a high temperature/high
pressure chamber. The sample was diluted with KBr and
DRIFTS spectra were collected from 650 to 4,000 cm-1 at
a resolution of 2 cm-1 averaged over 35 scans, in a N2
atmosphere at room temperature.
N2 adsorption–desorption isotherms, measured at
-196 °C using a Micromeritics ASAP 2020 unit, were used
to determine the BET surface area and pore volume of the
catalysts. The passivated catalysts were degassed at 250 °C
for 10 h prior to the measurement. TPR experiments were
carried out using a Micromeritics AutoChem II 2920 sys-
tem. Prior to TPR the calcined catalyst was pretreated in Ar
at 500 °C for 1 h, and then cooled to room temperature.
Subsequently, the Ar flow was switched to 10 % H2/Ar and
the sample was heated from room temperature to 1,000 °C
at a rate of 5 °C/min. The H2 consumption was measured by
a thermal conductivity detector (TCD).
2.2 Catalyst Characterization
The chemical composition of the catalysts was determined
using inductively coupled plasma-atomic emission spec-
troscopy (ICP-AES, done by Cantest Laboratories, Van-
couver, BC). Powder X-ray diffraction (XRD) analyses
The CO uptake was measured by pulsed chemisorption
also using the Micromeritics Autochem II 2920. The
123

Preparation of Unsupported NiMoP Catalysts
passivated catalyst was pretreated to remove the passiv-
ation layer by passing 50 ml(STP)/min of 10 % H2/Ar
while heating from 40 to 650 °C at a rate of 5 °C/min,
maintaining the final temperature for 1 h. The 10 % H2/Ar
flow was then switched to He (50 ml(STP)/min) at 650 °C
for 1 h in order to remove adsorbed species. The reactor
was subsequently cooled to 25 °C, and 0.5 ml pulses of CO
were injected into a flow of He (50 ml(STP)/min) and the
CO uptake was measured using a TCD. CO pulses were
repeatedly injected until no further CO uptake was
observed after consecutive injections.
NiMoP-0.5P
NiMoP-0.75P
NiMoP-1P
A Leybold Max200 X-ray photoelectron spectrometer
was used for XPS analysis of the catalysts. Al Ka was used
as the photon source generated at 15 kV and 20 mA. The
pass energy was set at 192 eV for the survey scan and
48 eV for the narrow scan. All catalyst samples were
analysed after passivation at room temperature. Exposure
of the samples to ambient atmosphere was minimized by
transferring the samples either in vacuum or under N2. All
XPS spectra were corrected to the C1s peak at 285.0 eV.
NiMoP
PDF:31-0873
NiMoP2
PDF:30-0862
20
30
40
50
60
70
80
2 Theta
2.3 Catalyst Activity
Fig. 1 X-ray diffractograms of NiMoP passivated catalysts prepared
with varying P/Me ratios
The HDS of 4,6-dimethyldibenzothiophene (4,6-DMDBT)
was measured in a fixed bed reactor (i.d. = 9 mm) at
310 °C and 3.0 MPa H2. The passivated catalysts
(dp \ 0.7 mm) were activated at 600 °C for 1 h in H2 at a
flow rate of 160 ml(STP)/min prior to the reaction. The
concentration of 4,6-DMDBT was 3,000 ppm in toluene.
The liquid feed was evaporated into a stream of flowing H2.
Gas and liquid flows and catalyst charged to the reactor
were chosen to give a WHSV = 26 h-1. The product was
collected periodically and analyzed using a Varian Star
3400 Gas Chromatograph equipped with a flame ionization
detector (FID). Component separation was achieved using
a capillary column (CP-Sil 19 CB, 25 m length and
0.53 mm i.d.). Component identification was confirmed
NiMoP-0.5CA
NiMoP-0.75CA
NiMoP-1CA
NiMoP-1.25CA
NiMoP-1.5CA
using the same column and
6890/5973N).
a GC–MS (Agilent
NiMoP PDF:31-0873
For all the activity data reported herein, the carbon
balance across the reactor was[95 % and a number of the
experiments were repeated to ensure repeatability of the
data.
MoP PDF:24-0771
Ni2P PDF:3-953
20
30
40
50
60
70
80
3 Results and Discussion
2 Theta
3.1 Catalyst Characterization
Fig. 2 X-ray diffractograms of NiMoP passivated catalysts prepared
with varying CA/Me ratios
The XRD analysis of the passivated NiMoP catalysts
(Figs. 1, 2, 3) showed the presence of a NiMoP phase
mixed with varying amounts of MoP. For the NiMoP cat-
alysts prepared with different P/Me ratios (see Fig. 1) a
new P-rich NiMoP2 phase appeared at increased P con-
centrations. The NiMoP2 was dominant for the NiMoP-1P
catalyst. Lee et al. [14] also reported the presence of
123

R. Wang, K. J. Smith
Table 2 Textural and crystalline properties of NiMoP catalysts pre-
pared by TPR of calcined precursor salts mixed with citric acid (CA)
NiMoP-550
NiMoP-600
Sample
XRD
BET Pore
Average
(m2/ volume pore
g)
(cm3/
size
(nm)
Phase
Grain
size
(nm)
g)
NiMoP-1P
MoP, NiMoP2
MoP, NiMoP2
14, 44
26, 45
66
78
0.04
0.09
3.3
4.6
NiMoP-
0.75P
NiMoP-0.5P MoP, NiMoP
32, 24
16
9
0.03
0.02
70
NiMoP-
1.5CA
MoP, NiMoP,
Ni2P
31, 25,
60
8.8
NiMoP-
1CA
MoP, NiMoP
MoP, NiMoP
MoP, NiMoP
MoP, NiMoP
32, 24
31, 16
24, 12
31, 19
16
40
24
8
0.03
0.04
0.05
0.02
70
43
86
8.7
NiMoP-650
NiMoP-
0.75CA
NiMoP
PDF:31-0873
NiMoP-
0.5CA
NiMoP-
noCA
20
30
40
50
60
70
80
NiMoP-650 MoP, NiMoP
NiMoP-600 NiMoP
NiMoP-550 NiMoP
31, 16
40
35
58
0.04
0.05
0.05
43
59
35
2 Theta
11
9
Fig. 3 X-ray diffractograms of NiMoP passivated catalysts prepared
with varying reduction temperatures
NiMoP2 in excess P (Ni/Mo/P = 1/1/2.5). For the NiMoP-
xP catalysts, the grain size of MoP, as determined from the
XRD analysis, decreased from 32 to 14 nm as the P content
increased, while the grain size of NiMoP2 (45 nm) was not
affected by increased P content (Table 2). Moreover, the
surface areas significantly increased with P content. These
data suggest that growth of MoP domains was inhibited by
the Ni and excess P, likely due to the decoration of Ni and
P on the edges of MoP crystallites.
the different metal citrate precursors formed from the
(NH4)6Mo7O24Á4H2O and Ni(NO3)2Á6H2O salts, as dis-
cussed previously [7].
The X-ray diffractograms of the NiMoP catalysts, pre-
pared using different final reduction temperatures, are
shown in Fig. 3. For the NiMoP-650 sample, a mixture of
MoP and NiMoP was observed, similar to the mixed MoP
and NiMoP phase reported by Stinner et al. [4], although
Ma et al. [16] reported that phase pure NiMoP was
obtained from a precursor with Ni/Mo/P = 1 when a high
H2 flow rate (1,000 ml/min) was used in the TPR. Figure 3
shows that a lower final reduction temperature favored the
production of phase pure NiMoP with broad diffraction
peaks, indicative of smaller crystallites. The NiMoP grain
size increased from 9 to 16 nm with increased reduction
temperature from 550 to 650 °C, while the surface area
decreased from 58 to 40 m2/g (Table 2). From these results
one concludes that phase pure, high surface area NiMoP
can be obtained at lower reduction temperature (550 °C),
with a Ni/Mo/P ratio of 1/1/1 and a 1.5 CA/Me ratio
(Me = Ni ? Mo).
Figure 2 shows the XRD patterns of the NiMoP cata-
lysts prepared with different CA/Me ratios (NiMoP-xCA
catalysts prepared with P/Me = 0.5 and a final reduction
temperature of 650 °C). In all cases NiMoP was the
dominant phase and as the CA content increased, the peaks
associated with MoP and NiMoP sharpened, indicative of
increased crystallization. For NiMoP prepared without
citric acid (NiMoP–noCA), similar diffractograms were
obtained [5]. The grain size of NiMoP-xCA, reported in
Table 2, increased from 12 to 25 nm as the CA content
increased. The surface areas, also shown in Table 2, were
higher than that of NiMoP–noCA, and reached a maximum
of 40 m2/g at a CA/Me ratio of 0.75. The maximum in
surface area with CA content was likely due to residual
carbon in the catalyst [7] and is in agreement with similar
trends observed for Ni2P and MoP catalysts prepared with
CA [7, 15]. Note that for the NiMoP-1.5CA, a new Ni2P
phase appeared (Fig. 2), implying that excess citric acid
restricts Ni–Mo interaction. This effect is likely related to
For the NiMoP–noCA catalyst, a green precipitate was
formed immediately that the Ni or Mo salt was mixed with
the (NH4)2HPO4 solution. However, for the NiMoP-xCA
samples, a green gel-like product formed during mixing
and the viscosity of the gel increased with increased CA
content. During drying at 120 °C, water was further
removed and a foam-like maroon material was produced.
123

Preparation of Unsupported NiMoP Catalysts
854
B
NiMoP-noCA
NiMoP-0.75CA
1720
1420
923
1184
A
1560
NiMoP-1CA
NiMoP-1.5CA
NiMoP-1CA
NiMoP-1.5CA
NiMoP-0.75P
754
937
1060
NiMoP-0.75CA
NiMoP-0.5P
NiMoP-0.5P
NiMoP-noCA
1383
NiMoP-0.75P
830
800
1000
1200
1400
1600
1800
2000
800
1000
1200
1400
1600
1800
2000
Wavenumber, cm-1
Wavenumber, cm-1
Fig. 4 DRIFT spectra of NiMoP dried precursors (a) and calcined precursors (b)
Figure 4 shows the DRIFT spectra of the dried NiMoP-
xCA, NiMoP-xP and NiMoP–noCA catalyst precursors.
For the dried NiMoP–noCA (Fig. 4a), three bands at 920,
989 and 1,053 assigned to metal ammonia phosphate
complexes (i.e., PO43-, HPO42-) are apparent, and a
decrease in intensity of these bands was observed with an
increase in the amount of CA used in the preparation. The
band at 830 cm-1 is assigned to NiMoO4, suggesting that a
portion of the Ni and Mo did not interact with the P [14].
The calcined NiMoP–noCA and NiMoP-xCA precursors
were quite different in appearance, with the former a dense
green precipitate, whereas the latter was brown in colour
with low density and high porosity, apparently due to
organic residues [7]. The FTIR spectra of the calcined
NiMoP-xCA and NiMoP-xP catalyst precursors are shown
in Fig. 4b. Compared with the spectra of the dried samples,
the bands corresponding to phosphates, citrate groups and
citric acid completely disappeared after calcination. The
Mo–O–Mo vibration at 754 cm-1 of the calcined NiMoP-
xCA was weaker than that of the calcined NiMoP–noCA
suggesting that the precursors prepared in the presence of
CA reduced the degree of Mo aggregation. The broad
region from 930 to 1,100 cm-1 is composed of Ni(Mo)–
P–O heteropoly compounds containing P–O bonded with
Mo and/or Ni.
The sharp band at 1,383 cm-1 is ascribed to NO3 [17].
-
For the NiMoP-xCA catalysts, the characteristic peak of
bimetallic oxides or nitrate at 830 and 1,383 cm-1 both
disappeared, demonstrating that chelating of the Mo or Ni
atom with citric acid could break the polyanionic cluster.
New bands at 854, 1,184 and 1,420 cm-1 assigned to metal
citrates were observed, indicating that carboxyl hydrogen
was replaced by Ni or Mo to form citrate complexes [7]. In
addition, a new band at 1,720 cm-1 assigned to the free
C=O of carboxylic acid [7, 18] in citric acid, appeared and
its intensity increased with increased CA content. It follows
therefore that only a fraction of the H atoms associated
with –COOH were replaced by the metal. Hence we con-
clude that the dried NiMoP-xCA precursors were com-
posed of a metal citrate complex, phosphomolybdate, and
nonbonded citric acid, similar to what was observed during
the synthesis of the mono-metal phosphides MoP and Ni2P
using CA [7, 15]. For the NiMo-xP catalysts (Fig. 4a), the
The DRIFTS features following drying and calcination
of the NiMoP catalyst precursors were very similar to those
reported previously for the Ni2P and MoP samples, also
prepared with CA [7, 15]. These data suggest that the
formation of NiMoP follows from chemically similar pre-
cursors as were identified for the synthesis of Ni2P and
MoP in the presence of CA [7, 15].
The TPR profiles of the NiMoP-xCA calcined catalyst
precursors are shown in Fig. 5. The reduction peak tem-
perature of all samples (*620 °C) was much lower than
that of MoP (712 °C) or Ni2P (721 °C) prepared with CA,
as reported previously [5]. Similar TPR profiles containing
one shoulder and one main peak assigned to Ni–Mo–O and
Ni–Mo–O–P precursors, respectively, were observed in all
samples. These data show the mixed metal oxides present
in the calcined precursors are more easily reduced than the
corresponding single-metal species present in the calcined
precursors of MoP or Ni2P prepared in the presence of CA.
intensity of the bands at *923, 1,080 and 1,180 cm-1
3-
associated with phosphate complexes (i.e., PO4
,
HPO42-) [19], increased with increased P content. The
intensity of a series of characteristic peaks of the metal
citrates (854 and 1,420 cm-1
) also increased with
increased P content, suggesting that the interaction between
phosphates and metal citrate increased with P content.
123

R. Wang, K. J. Smith
Table 3 Summary of XPS analysis of the reduced metal phosphide
catalysts
NiMoP-0.5CA
NiMoP-0.75CA
NiMoP-1CA
Catalyst
Binding energy
Surface composition
Nominal Surface
P/Me Ni/Mo
Atomic ratio
Ni2p
Mo3d P2p
Surface
P/Me
eV
NiMoP-
1P
854.2 228.7 129.8
1
1.08
1.54
1.52
0.79
0.72
0.79
0.86
0.88
0.79
0.89
0.97
NiMoP-
0.75P
853.8 228.3 130.1 0.75
853.9 228.2 130.2 0.5
853.8 228.2 130.2 0.5
853.9 228.2 130.2 0.5
853.8 228.4 130.4 0.5
853.9 228.3 130.3 0.5
853.8 228.4 130.4 0.5
853.9 228.3 130.3 0.5
854.0 228.2 130.4 0.5
2.86
0.67
0.85
0.67
0.67
0.67
0.67
0.65
0.62
NiMoP-
0.5P
NiMoP-
1.5CA
NiMoP-1.5CA
200
NiMoP-
1CA
0
400
600
800
Temperature, oC
NiMoP-
0.75CA
Fig. 5 TPR profiles of catalyst precursors prepared with different
CA/Ni ratios
NiMoP-
0.5CA
NiMoP-
650
NiMoP-
600
The XPS spectra of the Ni 2p, the Mo 3d and the P 2p
regions of the passivated NiMoP catalysts were very sim-
ilar (not shown) with the major peak binding energies
(BEs) summarized in Table 3. In all cases, the spectra were
indicative of reduced metal phosphide species, with some
oxidized Ni, Mo and P species also apparent due to the
presence of the passivation layer on the NiMoP. Peaks at
853.8, 228.2 and 130.4 eV were ascribed to the Nid?
(0 \ d \ 2), Mod? (0 \ d \ 6) and Pd- (0 \ d \ 1),
respectively, and no significant shift in BE was observed as
the preparation conditions for the reduced, passivated metal
phosphide catalysts were varied. The measured catalyst
surface compositions determined by XPS (Table 3) show P
rich surfaces in all cases, as has been reported previously
for most metal phosphides [5, 7, 20]. Also, the data show
that the surface P/Me ratio (Me = Ni ? Mo) increased
gradually with decreased reduction temperature, a conse-
quence of more P loss at higher reduction temperatures.
Most significant, however, is the increased P/Me ratio and
Ni/Mo ratio as the P content of the precursors increased,
and this increase follows the formation of the NiMoP2
phase identified by XRD.
NiMoP-
550
catalyst was 79.1 %, similar to that of Ni2P prepared with
CA, as reported previously [6]. For the phase pure NiMoP
catalysts (NiMoP-500 and NiMoP-600) the TOFs were
almost identical, despite large differences in NiMoP sur-
face areas and crystallite size. Among the catalysts tested,
they were the most active in terms of 4,6-DMDBT con-
version per unit mass of catalyst.
Table 4 also shows that the initial HDS activity
increased with increased P/Me ratio from 0.5 to 1, corre-
sponding to an increase in CO uptake, despite a lower BET
area. However, in this case, there was a significant phase
change in the prepared catalysts. At a P/Me = 1, NiMoP2
was the dominant phase rather than NiMoP. The higher
TOF observed with increased P content is likely related to
the presence of the NiMoP2 phase. The crystal structure of
NiMoP is similar to that of Ni2P, with the Ni atoms near
the center of the side faces of the unit cell replaced by Mo.
However, NiMoP2 has a different crystal structure and a
higher P content compared to NiMoP [21], yielding a
higher metal-P coordination, that should reduce adsorption
of the reactant molecule. Nagai et al. [22] reported that Ni
and Ni2? species in Ni–Mo–P/Al2O3 increased during
reduction and nickel migrated to the surface and promoted
the HDS activity. The high surface Ni/(Ni ? Mo) ratio of
the NiMoP-1P catalyst of the present study also contributes
to the high TOF compared to the NiMoP-0.5P catalyst. The
3.2 HDS of 4,6-DMDBT
The hydrodesulfurization (HDS) of 4,6-DMDBT was per-
formed at 310 °C, 3.0 MPa, and Table 4 reports the HDS
conversions obtained over the NiMoP catalysts initially
and after 5 h time-on-stream. The 4,6-DMDBT initial
conversion decreased significantly with increased reduction
temperature and the initial conversion of NiMoP-550
123

Preparation of Unsupported NiMoP Catalysts
Table 4 CO uptake and 4,6-DMDBT HDS performance over different catalysts
Catalyst
CO uptake Initial
conv.
(%)
Final
conv.
(%)
TOF
(10-3 s-1
Selectivity (TH ? HH)/ MCHT/
MCHT DMBP
THDMDT ? HHDMDBT DMBP MCHT DMBCH Ratio
(lmol/g)
)
%
NiMoP-
2P
8.3
7.1
0.5
7.1
19.6
–
38.3
31.9
37.3
31.9
44.1
45.5
44.1
43.0
28.9
27.2
28.9
35.1
21.0
35.1
7.0
5.5
2.4
5.5
3.1
–
41.7
16.0
19.0
16.0
10.4
23.2
10.4
12.9
16.0
16.7
16.0
18.1
18.4
18.1
39.0
51.9
64.4
51.9
67.4
58.4
67.4
6.5
6.1
–
1.07
0.31
0.3
3.00
3.24
3.87
3.24
3.72
3.23
3.72
NiMoP-
1P
NiMoP-
3CA
NiMoP-
2CA
6.1
–
0.31
0.15
0.4
NiMoP-
1.5CA
NiMoP-
1CA
–
NiMoP-
1.5CA-
650
19.6
3.1
0.15
NiMoP-
1.5CA-
600
19.1
32.6
58.8
79.1
22.5
26.4
4.7
4.9
20.8
12.3
18.8
22.2
61.9
52.3
4.3
0.34
0.24
3.3
NiMoP-
1.5CA-
550
12.5
2.34
Ni2P
from [5]
30.1
42.4
87.7
74.7
80.6
34.3
7.1
3.3
4.2
8.5
66.6
52.3
19.3
22.8
0.06
0.23
7.83
4.28
MoP
12.2
12.2
from [5]
Reaction conditions: temperature 310 °C, pressure 3.0 MPa H2, 3,000 ppm 4,6-DMDBT and WHSV 26 h-1
THDMDBT 4,6-tetrahydrodimethyldibenzothiphene, HHDMDBT hexahydrodimethyldibenzothiphene
data of Table 4 also show that the TOF of the most active
NiMoP preparation was comparable to that reported pre-
viously for Ni2P prepared in the presence of CA [5].
The reaction network of 4,6-DMDBT involves two
parallel reactions that lead to desulfurized products [23].
One is denoted as direct desulfurization (DDS) which in
the case of 4,6-DMDBT results in the formation of 3,3-
dimethylbiphenyl (3,3-DMBP). The second, referred to as
the hydrogenation pathway (HYD), involves the pre-
hydrogenation of 4,6-DMDBT to 4,6-tetrahydro- and
hexahydro-dimethyldibenzothiophenes (THDMDBT and
HHDMDBT), which are subsequently desulfurized to
methylcyclohexyltoluene (MCHT) and dimethylbicylco-
hexane (DMBCH). In the present study, the high ratio of
the principal products of the HYD to DDS routes (MCHT/
DMBP) shows that the dominant pathway is via the HYD
route for all the catalysts.
NiMoP phase. The possible contribution of the Brønsted
acidity of metal phosphides that facilitate the DDS pathway
of HDS of 4,6-DMDBT [9, 10, 20] is less important for the
NiMoP-xP catalysts since the DMBP selectivity was only
slightly affected by
P
content. However, the
(THDMDBT ? HHDMDBT)/MCHT ratio increased as
the P content increased, suggesting that the NiMoP2 phase
decreased the rate of sulfur elimination from hydrogenated
intermediates of the HYD route.
4 Conclusion
The effect of preparation conditions on the structure and
catalytic performance of unsupported NiMoP catalysts in
the presence of CA has been investigated. The chemical
composition and surface area of the NiMoP catalysts was
varied by the synthesis conditions, with small amounts of
MoP and NiMoP2 present in some of the catalysts. Che-
lating Mo and Ni with citric acid suppressed the
It is also found that the MCHT/DMBP ratio of NiMoP
catalysts decreased with increase in reduction temperature,
consistent with a decrease in the crystallite size of the
123

R. Wang, K. J. Smith
8. Topsøe H, Clausen BS, Massoth FE (1996) In: Anderson JR,
Boudart M (eds) Catalysis science and technology, vol 11.
Springer, Berlin, p 162
aggregation of the catalyst precursors and yielded precur-
sors chemically similar to those identified for MoP and
Ni2P. Phase pure, high surface area NiMoP was synthe-
sized in the presence of CA at low reduction temperature
(550 °C), with a Ni/Mo/P ratio of 1/1/1 and a 1.5 CA/Me
ratio (Me = Ni ? Mo), resulting in a high initial conver-
sion of 4,6-DMDBT, similar to that of Ni2P prepared with
CA. For the phase pure NiMoP catalysts, the TOFs for the
HDS of 4,6-DMDBT were almost identical, despite large
differences in NiMoP surface areas and crystallite size.
9. Abu II, Smith KJ (2006) J Catal 241:356–366
10. Abu II, Smith KJ (2007) Catal Today 125:248–255
11. Sun F, Wu W, Wu Z, Guo J, Wei Z, Yang Y, Jiang Z, Tian F, Li
C (2004) J Catal 228:298–310
12. Li K, Wang R, Chen J (2011) Energy Fuels 25:854–863
13. Gaudette AF, Burns AW, Hayes JR, Smith MC, Bowker RH,
Seda T, Bussell ME (2010) J Catal 272:18–27
14. Lee SY, Lee YK, Oyama ST, Lee SH, Woo HC (2007) Solid
State Phenom 124–126:1765–1768
15. Whiffen VL, Smith K (2012) Top Catal 55:981–990
16. Ma D, Xiao T, Xie S, Zhou W, Gonzalez-Cortes S, Green MLH
(2004) Chem Mater 16:2697–2699
Acknowledgments Financial support from the Natural Sciences and
Engineering Research Council of Canada is gratefully acknowledged.
17. Takahashi R, Sato S, Sodesawa T, Suzuki M, Ichikuni N (2003)
Microporous Mesoporous Mater 66:197–208
18. Cheng R, Shu Y, Li L, Zheng M, Wang X, Wang A, Zhang T
(2007) Appl Catal A 316:160–168
References
19. Arai Y, Sparks DL (2001) J Colloid Interface Sci 241:317–326
20. Prins R, Bussell M (2012) Catal Lett 142:1413–1436
1. Li W, Dhandapani B, Oyama ST (1998) Chem Lett 27:207–208
2. Stinner C, Prins R, Weber T (2000) J Catal 191:438–444
3. Wang X, Clark P, Oyama ST (2002) J Catal 208:321–331
4. Stinner C, Prins R, Weber T (2001) J Catal 202:187–194
5. Wang R, Smith KJ (2009) Appl Catal A 361:18–25
6. Wang R, Smith KJ (2010) Appl Catal A 380:149–164
7. Whiffen VML, Smith KJ, Straus SK (2012) Appl Catal A
419–420:111–125
´
´
´
21. Oryshchyn SV, Zhak OV, Guerin P, Deputier S, Kuzma YB
(2002) Inorg Mater 38:1195–1202
22. Nagai M, Fukiage T, Kurata S (2005) Catal Today 106:201–205
23. Whitehurst DD, Isoda T, Mochida
42:345–471
I (1998) Adv Catal
123