Catalysis Today 329 (2019) 44–52

Contents lists available at ScienceDirect

Catalysis Today
journal homepage: www.elsevier.com/locate/cattod

In situ reactivation of spent NiMoP/γ-Al2O3 catalyst for T

hydrodesulfurization of straight-run gas oil
⁎
C.E. Santolalla-Vargasa, Víctor Santesa, , Elizabeth Gómezb, Felipe Sanchez-Mineroc,
Issis Romero-Ibarrad, O. Goiza, L. Lartundo-Rojase, Leonardo Díazf, Rosario Luna-Ramirezf,
J.A. de los Reyesg, O.U. Valdésg
a
Departamento de Biociencias e Ingeniería, Centro Interdisciplinario de Investigaciones y Estudios sobre Medio Ambiente y Desarrollo (CIIEMAD), Instituto Politécnico
Nacional, C.P. 07340, Ciudad de México, Mexico
b
Instituto de Química, Universidad Nacional Autónoma de México, Circuito Exterior S/N, Ciudad Universitaria, Coyoacán, C.P. 04510, Ciudad de México, México
c
Departamento de Ingeniería Química Petrolera, ESIQIE, Instituto Politécnico Nacional, Zacatenco, C.P. 07738, Ciudad de México, México
d
Unidad Profesional Interdisciplinaria en Ingeniería y Tecnologías Avanzadas- Instituto Politécnico Nacional, Av. IPN No. 2580, Gustavo A. Madero, C.P. 07340, Ciudad
de México, México
e
Instituto Politécnico Nacional, Centro de Nanociencias y Micro y Nanotecnologías, UPALM, Zacatenco, C.P. 07738, Ciudad de México, México
f
Instituto Mexicano del Petróleo, Eje Central Lázaro Cárdenas 152, San Bartolo Atepehuacan, Gustavo A. Madero, C.P. 07730, Ciudad de México, México
g
Universidad Autónoma Metropolitana-Iztapalapa, División de Ciencias Básicas e Ingeniería, Av. San Rafael Atlixco 186, Col. Vicentina, C.P. 09340, Ciudad de México,
México

A R T I C LE I N FO A B S T R A C T

Keywords: Washing with xylene (X), 2,6 bis 1 hydroxy 1,1 diphenyl methyl pyridine (A) and in situ reactivation with
Hydrodesulfurization molybdenum acetylacetonate (AceMo) on the NiMoP/Al2O3 spent surface catalyst has been investigated in the
Straight run gas-oil hydrodesulfurization (HDS) of straight-run gas oil. The spent catalyst was washed and dried at 120 °C before in
Nickel-molybdenum situ reactivation. The sulfided catalysts were characterized by temperature programmed reduction (TPR), nu-
Spent catalyst
clear magnetic resonance (NMR), Raman spectroscopy and X-ray photoelectron spectroscopy (XPS). The HDS
Coke
reactions were carried out at 5.5 MPa of H2 and the reaction temperatures at 340 °C, 360 °C and 380 °C. The
CatAXAceMo displayed the highest HDS activity. The TPR results advice that CatAX presented more “rigid” MoS2
edges than CatRef catalyst. NMR results suggest that CatAXAceMo showed lower aromatic or polyaromatic
hydrocarbons surface concentration deposited after HDS reaction than CatAceMo and CatRef. Raman spectro-
scopy revealed the formation of coke crystallites with shorter size for CatAXAceMo than others catalysts. XPS
spectroscopy results exhibited that CatAXAceMo presented smaller superficial carbon and a larger concentration
of MoS2 active phase. A relation between the coke crystallites size and MoS2 superficial concentration was found.
A lower aromatic concentration favors the shorter coke crystallites sizes and a major availability of Mo species
for the formation of MoS2 to HDS reaction.

1. Introduction spent catalyst has several contaminants which limit its use [3].
The lower sulfur conversion in the HDS process leads to increment
The environmental legislation is demanding for low sulfur fuels. In the catalytic bed temperature to maintain the higher sulfur conversion.
this respect, Mexico has to decrease to 15 ppm of sulfur content in Nevertheless, the temperature increment favors the catalyst sintering.
diesel [1]. However, the major reserves are the heavy crudes oil such as In that case, several authors demonstrated the loss of the surface-active
Maya Crude, which is difficult to process. The fresh catalyst market for phase [4,5]. Also, Dufresne P. et al. [4] showed that catalyst deacti-
the hydrotreating process has incremented in order to decrease the vation implies the blockage of catalytic pores, a decrease of metals
sulfur concentration in heavy crudes; consequently, the amount of spent dispersion and the strong adsorption of nitrogen and coke. The main
catalyst has also been growing. The main principles of waste regulation contaminant on the surface catalyst is the carbon deposited after hy-
encourage: minimizing waste production, maximizing reuse and re- drodesulfurization (HDS) reactions, which corresponds to the poly-
cycling [2]. The reuse of spent catalyst is an alternative, although the merization of hydrocarbons such as alkenes, aromatics, and

⁎
Corresponding author.
E-mail address: [email protected] (V. Santes).

https://doi.org/10.1016/j.cattod.2019.03.015
Received 27 July 2018; Received in revised form 11 February 2019; Accepted 8 March 2019
Available online 09 March 2019
0920-5861/ © 2019 Elsevier B.V. All rights reserved.
C.E. Santolalla-Vargas, et al. Catalysis Today 329 (2019) 44–52

heterocycles [6]. Likewise, the reactivation of spent catalyst has re- Table 1
quirements such as moderate metals content after the HDS pro- Catalysts and pretreatments before HDS reaction.
cess.Regarding this, the reactivation of spent catalyst with higher Ni Catalyst Synthesis conditions Additive Xylene AceMo
and V content (> 5% wt) is more difficult, and the best results reach
about 80% of catalytic activity [7]. In addition, coke particles could CatRef NiMoP/γ-Al2O3 – – –
CatAX NiMoP/γ-Al2O3 X X –
trap nickel into the coke matrix which leads to strong interaction metal-
CatAceMo NiMoP/γ-Al2O3 – – X
coke and makes difficult metals removal [8]. Therefore, one of the CatAXAceMo NiMoP/γ-Al2O3 X X X
challenges is to remove the carbon on the HDS active phase. Some
authors suggest the use of polar and non-polar solvents [9,10]. Wiwel
et al. [9] exhibited a relation between the coke deposited and the pretreatments to elucidate the effect of the 2,6 bis 1 hydroxy 1,1 di-
paraffin, naphthalene, and aromatics content on the feed. This study phenyl methyl pyridine and molybdenum acetylacetonate on the cata-
showed an increase of aromatic species and a decrease of unsaturated lytic surface with the HDS of straight-run gas oil.
carbons after the spent catalyst was washed with solvents and organic
additive as pentane, methanol-benzene and pyridine. In addition,
2. Experimental
Cortés-Torres et al. [10] pretreated a spent catalyst from petrochemical
refinery with several solvents polar and non-polar as CS2, CH2Cl2,
2.1. Sulfided catalyst preparation
hexane, toluene, acetone, H2O and ethyl acetate. This author suggested
better performance to remove impurities such as V, Fe and coke for CS2
A spent catalyst with 9.0 wt%, 2.8 wt% and 1.5 wt%, of Mo, Ni and
solvent. However, ethyl acetate exhibited similar behavior than CS2 and
P, respectively, was used. The spent NiMoP/Al2O3 was treated by a
it is environmentally friendlier. Hanadi A. Abdullah et al. [11] studied
solution of xylene and 2,6 bis 1 hydroxy 1,1 diphenyl methyl pyridine
new extraction methods using different polar solvents to eliminate the
with a Soxhlet apparatus at 140 °C during 48 h. The details of the
surficial coke. This author found that accelerated solvent extraction
Soxhlet method extraction are reported by Hanadi A. Abdullah et al.
(ASE) exhibited superior extraction in comparison to the Soxhlet
[11]. Then, the catalyst was dried at 120 °C. Before the sulfiding pro-
method. However, the ASE needs several conditions than the Soxhlet
cess, the catalyst was pretreated in situ with molybdenum acet-
method, which are difficult to control. Also, the use of polar solvents
ylacetonate (AceMo) (Table 1). Catalysts were labeled as CatAX, Cat-
could leach some Mo species from the spent catalyst [7,12,13], which
AceMo and CatAXAceMo to relate the pretreatment made it. Also, a
could lead to a lower Mo superficial concentration for the HDS pro-
CatRef without pretreatment was used as a reference.
cess.On the other hand, the use of solvents with organic additives for
rejuvenation/regeneration of hydrotreatment (HDT) catalysts were
studied [14]. The regeneration with rejuvenation presented better 2.2. Characterization techniques
performance in HDT than the conventional reactivation technologies
based on oxidative regeneration [14]. The chemical rejuvenation with UV-Vis spectra of the aqueous solution were recorded in the ab-
solvent and organic additives leads to improve the dispersion of active sorbance mode on a Perkin Elmer Lambda 35.
phase. In this case, the solvent reaches the uniform distribution of the A Leco combustion-IR carbon/sulfur analyzer, Model CS 600 (Leco
organic additive in the pores of the catalyst [14]. The organic additives Corp.,St. Joseph, MI), was used in this work. The instrument consists of
for rejuvenation are molecules with carboxyl, carbonyl or hydroxyl a microprocessor control unit, an electronic balance, resistance furnace,
groups. In this aspect, citric acid, tartaric acid, oxalic acid, malonic and a solid-state infrared detector. The solid sample was combusted in a
acid, maleic acid, butane-diol, methylglyoxal, glycolaldehyde and stream of pure oxygen at up to 1000 °C for approximately 120 s. The
acetaldol were studied as organic additive [14]. Also, the use of ami- evolved SO2 and CO2 was detected and measured in an IR cell after the
nocarboxylic acids such nitrile-triacetic acid (NTA) and diethylene removal of water.
triamine-pentaacetic acid were tested in the rejuvenation process [15]. Temperature programmed reduction (TPR) of the catalysts was
These compounds interact with the transition metals (Ni, Co, Mo and carried out using Altamira Instruments AMI-90 equipped with a
W) and form complexes with significantly stability [16–21]. Medici thermal conductivity detector (TCD). Circa 0.1 g of catalyst sample was
et al. [17,18] showed that the use of NTA forms MoO3(NTA) complex placed in a quartz sample cell (U shaped) and pretreated in situ at 393 K
that shifts the H2S consumption peak to higher temperatures, sug- for 1 h, under pure Ar flow. The reduction of catalysts were carried out
gesting that the Mo structure was more uniform and more stable than in from room temperature to 1273 K (at10 °C/min, heating rate), under
the NTA-Free case. C.E. Santolalla-Vargas et al. [21] found that the 0.84 cm3/s of reducing mixture (10 vol.%H2, balance N2).
interaction between W-CyDTA complex shifts the temperature decom- NMR analysis was conducted on a Bruker 750 MHz spectrometer
position up to 450 °C. Chen et al. [16] displayed that the complex Mo- (Bruker Biospin, Rheinstetten, Germany) using a 4 mm probe at
citric acid has strong stability than Mo without chelating agent. Pyr- 100.621 MHz frequency. 13C NMR spectra were acquired using cross-
idine ligands are excellent chelating agent and they have been studied polarization magic angle spinning mode with a 10 kHz rotation fre-
in the HDS process [20]. In this regard, Elizabeth Gomez et al. [22] quency, a 3 ms contact time and a recycle delay of 4 s. 5120–10,240
suggested the interaction between 2,6-disubstituted pyridine and Sn scans were accumulated per spectrum. The NMR data were processed
where this chelating agent favors the selective formation of monomeric using the TOPSPIN 3.2 software.
five-coordinated tin complexes. Raman spectra were recorded at room temperature on a T64000
On the other part, the catalyst acid properties improve coke pro- triple monochromator (Jobin-Yvon-Horiba) using the 514.5 nm line of
duction. Appleby E.G. et al. [23] presented that the increment of coke an Ar+ laser (Lexel Laser). All spectra of the sulfide solids were ob-
deposited has a relationship with the acidity. Although the acidity fa- tained at a power of 10 mW at the laser head, in the range of 200 –3000
vors the hydrocracking, this author suggested that the catalyst needs an cm−1, using an Olympus microscope with a ×100 objective and 10
acidity equilibrium between the hydrocracking and hydrodesulfuriza- accumulations of 60 s each. The spectrum resolution was 1 cm−1.
tion activity. The type of acidity which favors the coke formation is a X-ray photoelectron spectroscopy (XPS) was carried out using a K-
controversial topic because the active phase could act as Lewis and alpha Thermo Fischer Scientific spectrometer with a monochromatic Al
Bronsted acidity while the alumina support presents Lewis acidity [24]. Ka radiation (1486.6 eV) as an X-Ray source and were micro-focused at
The use of 2,6 bis 1 hydroxy 1,1 diphenyl methyl pyridine as an the source to give a spot size on the sample of 400 μm in diameter. XPS
organic additive to eliminate the coke impregnated in HDS spent cat- high-resolution spectra were collected using analyzer pass energies of
alyst has not been studied. In this work, we studied different 20 eV, respectively. To compensate effects related with charge shift, C1s

45
C.E. Santolalla-Vargas, et al. Catalysis Today 329 (2019) 44–52

Table 2
Source fuel specification.
Source fuel specification Method

S (ppm) ASTM D5453 22320

N (ppm) ASTM D4629 450
Cetane number ASTM D976 47.2
Aromatics (wt.%) ASTM D5186 60.5
Cinematic viscosity (mm2/s) ASTM D445 5.5
Aniline temperature (°C) ASTM D611 72.2

peak at 284.6 eV was used as internal standard. The details of the XPS
measurements by this spectrometer are reported elsewhere [19].

2.3. Catalytic activity

The catalysts (1.20 mL; 80–100 mesh) were sulfided in virgin SRGO Fig. 2. Conversion in HDS of straight run gas oil for CatRef, CatAX, CatAceMo
(Table 2; Fig. 1) spiked with dimethyl disulfide (DMDS) to afford 2.5 wt and CatAxAceMo catalysts.
% of sulfur in the feedstock and then fed to the reactor at a flow rate
according to a LHSV of 2.5 h−1. The pressure was set to 5.5 MPa and (340 °C, 360 °C and 380 °C) is shown in Fig. 2. Fig. 2 shows that the
the hydrogen rate was adjusted according to a H2/oil ratio of temperature increment favored the sulfur removal for all catalysts. In
500 m3(STP)/m3. The temperature of the catalyst bed was increased to addition, CatAXAceMo exhibits a larger straight-run gas oil conversion
290 °C and held for 15 h. Then for the HDS reactions the SRGO and H2 than CatRef, CatAX and CatAceMo (Fig. 2). The pretreatment with
flow rates were adjusted to LHSV of 1.5 h−1, 340 °C, H2/oil ratio equal additive, xylenes and molybdenum acetylacetonate enhanced the sulfur
to 500 m3 (STP)/m3 and 5.5 MPa. The sulfur content in the liquid removal. However, the treatment only with AceMo disfavored the
product was determined with a Tanaka RX-360SH model sulfur ana- sulfur conversion. This result may be related to the fact that the spent
lyzer according to the ASTM D4294 method. For each test, 10 mL of catalyst still has coke after the treatment with AceMo on the sulfiding
sample was fed to the equipment. The reported instrument sensitivity step. In this regard, the catalyst washed with xylenes and additive
is ± 0.001% of the measured value. (CatAX) showed better performance than the CatAceMo. Nevertheless,
The apparent constant rates (khds(Sw%)−0.5 h−1) were estimated at CatAX displayed the same catalytic activity as the reference catalyst.
steady state conditions (after 18 h ; similar steady state condition was Probably, this fact is associated with the Mo or Ni extraction provoked
used according to Leonardo Diaz et al [25]), and they were calculated by the washed with xylene and additive (Fig. S1) [22,28], and there-
according to the following equation [26,27]: fore, CatAX could have a low concentration of MoS2 active phase than
the reference catalyst.
LHSV ⎛ 1 1
khds = ⎜ − n − 1 ⎞⎟ On the other hand, the CatRef, CatAX and CatAceMo exhibited si-
n − 1 ⎝ Spn − 1 Sf ⎠ milar activation energy (c.a. 26 kcal/mol) while the CatAXAceMo
showed 34.4 kcal/mol (Table 3). This increment is related that Cat-
where:
AXAceMo could convert hindrance sulfur molecules as 4,6 dimethyl
khds: Pseudo 1.5 order HDS kinetic constant (Sw%−0.5 h−1);
dibenzothiophene (Fig. 1) which is less reactive than thiophene and
Sp: Sulfur in the product (wt %);
dibenzothiophene [29]. Also, the novel catalysts have to present a good
Sf: Sulfur in the feedstock (wt %);
performance to convert hindrance sulfur molecules to produce ultra-
LHSV: Liquid Hourly Space Velocity (h−1).
low sulfur diesel. In this sense, the CatAXAceMo could be a potential
catalyst for the HDS process [30].
3. Results

3.1. Catalytic activity measurements 3.2. Textural properties

The straight-run gas oil conversion (%) for CatRef, CatAX, The CatRef and CatAX were analyzed by N2 physisorption in order
CatAceMo and CatAXAceMo under three reactions temperatures to measure the textural properties. The surface area for CatRef and

Fig. 1. GC-SCD analysis of the SRGO feedstock.

46
C.E. Santolalla-Vargas, et al. Catalysis Today 329 (2019) 44–52

Table 3
Activation energy, conversiona and reaction rate constant at steady state (18 h)
in the HDS of straight-run gas oil over NiMoP/γ-Al2O3 sulfided catalysts.
CatRef CatAX CatAceMo CatAXAceMo

Ea (kcal/mol) 27.2 27.1 26.0 34.4

X340 °C (%) 86.0 86.0 81.9 84.2
X360 °C (%) 94.9 94.2 92.4 96.4
X380 °C (%) 98.2 98.2 97.2 98.9
khds340 °C (Sw%)−0.5*h-1 3.4 3.4 2.7 3.1
khds360 °C (Sw%)−0.5*h-1 6.9 6.3 5.3 8.7
khds380 °C (Sw%)−0.5*h-1 13.2 13.2 10.0 17.2

a
Reaction conditions were: P = 5.5 MPa; total reaction time per catalyst ca.
56 h; fixed bed reactor.

CatAX were 118 and 139 m2/g while the pore diameter were 92 and
78 A, respectively. Both catalysts presented the same pore volume (ca.
0.28 cm3/g). According to physisorption N2 results, the pore volume is
similar for CatRef and CatAX. This result suggests that the type of coke
removed is associated with non-refractory coke or soft coke which was
blocking some micropores. Therefore, the CatAX catalyst changed the
average pore diameter, but the pore volume held.

3.3. Carbon and sulfur content for CatRef and CatAX catalysts before HDS
reaction

The CatRef and CatAX catalysts were characterized by carbon

/sulfur analyzer in order to quantify the superficial carbon. Table 4
Fig. 3. TPR profiles after HDS reaction for CatRef (a) and CatAX (b).
shows the sulfur and carbon content for CatRef and CatAX. The CatRef
showed 7.92 wt% and 14.04 wt% for S and C, respectively. The CatAX
exhibited 11.14 wt% and 6.17 wt% for C and S, respectively. The washing with xylene and additive. The higher weakly bounded sulfur
sample that was treated with xylene and 2,6 bis 1 hydroxy 1,1 diphenyl temperature strongly suggests that the CatAX catalyst has more “rigid”
methyl pyridine removed 20.65% of carbon in comparison with CatRef. MoS2 edges than the catalyst reference. This phenomenon has been
On the other hand, to verify the interaction between the spent catalyst already observed in the MoS2 catalysts vs. sulfiding temperature
with xylene solvent, the solution was measured by Uv Vis spectroscopy [32,33]. In addition, the peak around 390 °C – 500 °C corresponding to
(supplementary information). The spectrum showed a wide band be- sulfur species with strong interaction on the catalytic surface is more
tween 250 nm and 370 nm corresponding to O2− →Mo6+ charge markedly for CatAX than the CatRef. In this sense, the sulfur species
transfer [31]. This result suggests that the xylene could leach the Mo with strong interaction leads to lower reducible species or sulfur va-
species from the spent catalyst. The decrement of Mo on the surface cancies [34–36]. This result could suggest lower sulfur vacancies for
disfavored the Mo active phase concentration, and therefore, it leads to CatAX in comparison with CatRef. Extrapolating this effect to HDS re-
a similar catalytic activity than the catalyst reference. This result is in action, it could be suggested that the CatAX would have lower sulfur
accordance with the catalytic activity seen in Table 3. vacancies than the CatRef. Consequently, a lower concentration of ac-
tive phase and catalytic activity in HDS of straight run gas oil.
3.4. Temperature programmed reduction for CatRef and CatAX catalysts
after HDS reaction
3.5. Nuclear magnetic resonance (NMR) of CatRef, CatAceMo and
CatAXAceMo catalysts after HDS reaction
The CatRef and CatAX catalysts were analyzed by temperature-
programmed reduction in order to get information about the metal-
The CatRef, CatAceMo and CatAXAceMo catalysts were character-
sulfur interaction. The TPR profiles of the CatRef and CatAX samples
ized by NMR in order to evaluate the superficial carbon remains after
after the reaction is shown in Fig. 3. As seen in this figure, the TPR
the HDS reactions. Fig. 4 shows the NMR 13C for CatRef before reaction
profile CatRef (Fig. 3a) shows one peak at 255 °C ascribed to weakly
(a), CatRef (b), CatAceMo (c) and CatAXAceMo (d) catalysts. Regarding
bounded sulfur [32,33]. The broad shoulder at 151 °C corresponding to
the pretreatment, all catalysts showed peaks about 0–60 ppm,
dehydration reactions while the peaks at 393 °C and 497 °C could relate
60–100 ppm 100–165 ppm and > 165 ppm corresponding to aliphatic,
to the reducible sulfur species with strong interaction on the catalytic
quaternary, aromatic and carbonyl carbons, respectively [37]. In this
surface [32,33]. For TPR profile of CatAX (Fig. 3b) exhibits two peaks at
sense, Table 5 exhibits the semi-quantitative area of aromatic and ali-
307 °C and 475 °C corresponding to weakly bounded sulfur desorption
phatic superficial carbon after the HDS reaction of straight run gas oil.
and sulfur species with strong interaction, respectively. As compared
The CatRef before reaction, CatRef, CatAceMo and CatAXAceMo pre-
with CatRef sample, there is an important upward shift of weakly
sented 46%, 86%, 85% and 77% aromatic superficial carbons, respec-
bounded sulfur temperature (from 255 °C to 307 °C) originated by
tively. The results of the aliphatic carbons percentage showed that the
CatRef before reaction, CatRef, CatAceMo and CatAXAceMo catalysts
Table 4
displayed 54%, 14%, 15% and 23%, respectively.
Sulfur and carbon content on the solid materials.
According to the literature, the presence of aromatic or polyaro-
Catalyst Sulfur (%) Carbon (%) matic hydrocarbons on the catalytic surface leads to coke formation
[38,39]. In this regard, the aliphatic to aromatic ratio shows that the
CatRef 7.92 14.04
CatAX 6.17 11.14 CatRef presented the lowest ratio (0.16) while the CatAXAceMo ex-
hibited the highest ratio (0.30). This result suggests that the catalyst

47
C.E. Santolalla-Vargas, et al. Catalysis Today 329 (2019) 44–52

Fig. 4. 13C-NMR for CatRef before reaction (a), CatRef (b), CatAceMo (c) and Fig. 5. Raman spectra for CatRef (a), CatAceMo (b), CatAXAceMo (c) and
CatAXAceMo (d) catalysts. CatRef before reaction (d) catalysts.

CatAXAceMo could have lower coke formation in comparison with the Table 6
CatRef. Also, the use of additive, xylene and AceMo desorbed the soft Raman deconvoluted results.
superficial carbon as aliphatic and quaternary carbons in contrast with
Catalyst ID IG IG’ La (Å) I 1599/I 1317
the catalyst reference before reaction. Furthermore, the pretreatment
with additive, xylene and AceMo decrement the coke deposited after CatRef before reaction 426000 179300 64910 17.80 1.72
HDS reaction on the spent catalyst surface. CatRef 54070 15500 7166 20.82 0.89
CatAceMo 1156 799 264 14.05 0.74
CatAXAceMo 1001 735 277 13.41 0.77
3.6. Raman spectroscopy of CatAceMo and CatAXAceMo catalysts after
HDS reaction
intensity Raman ratio I1599/I1314 were measured to acquire the gra-
The CatRef, CatAceMo and CatAXAceMo catalysts were measured phitization degree (Table 6). The CatRef before reaction and CatRef
by Raman spectroscopy. Also, the crystallized of carbon degree after the presented the higher Raman ratio (1.72 and 0.89, respectively) while
HDS reactions was studied. The formation of graphite well-crystallized the CatAceMo and CatAXAceMo presented the lower ratio, 0.74 and
shows two Raman vibration modes: at 42 cm−1 corresponding to an 0.77 respectively. This fact suggests that the pretreatment and in situ
interlayer mode while the intralayer mode is at 1581 cm−1. The pre- reactivation could decrease the graphitization degree or coke crystal
graphitic carbons exhibit another wide band with a maximum centered size.
at 1355 cm−1. On the other hand, the ratio intensity with the vibrations To get information about the crystal size, the Raman spectra were
at 1355 cm−1 and 1581 cm−1 is related to the graphitization degree deconvoluted (Fig. 6) according to reported by Guichard et al. [8]. The
[40]. In this sense, Raman spectroscopy was recorded on defective band at 1317 cm−1 is assigned to D1 vibration while the main band
graphite region (1200 cm−1 -1700 cm−1) to get information about the centered at about 1599 cm−1 were deconvoluted in three bands at
coke deposited after HDS reaction. The Raman spectra for CatRef before 1575 cm−1, 1599 cm−1 and 1613 cm−1 corresponding to G’, G and D2
reaction, CatRef, CatAceMo and CatAXAceMo catalysts are presented in vibrations, respectively. These bands intensity were measured with the
Fig. 5. All samples showed two wide bands centered at 1314 cm−1, equations reported by Robertson et al. [41]. Table 6 shows the Raman
1599 cm−1 and a shoulder at 1210 cm−1 approximately. According to deconvolution results and the crystal size of coke (La) for the catalysts
above, the band at 1314 cm−1 correspond to pre-graphitic carbon while with different pretreatments. Table 6 exhibits that the crystal size of
the vibration at 1599 cm−1 corresponds to intralayer mode. The coke for CatAceMo is 14.05 Å while the CatRef after and before HDS
shoulder at 1210 cm−1 may ascribe to C–H vibrations [8]. Also, the reaction show a maximum between 20.82- 17.80 Å, then decreases for

Table 5
Semiquantitative area of aromatic and aliphatic superficial carbon after HDS reaction of straight run gas oil.
Catalyst Aromatic Aliphatic Ali/Aro ratio Aromatic (%) Aliphatic (%)

CatRef before reaction 268,373.00 234294.00 0.87 46 54

CatRef 84.31 13.91 0.16 86 14
CatAceMo 79.01 14.34 0.18 85 15
CatAXAceMo 73.68 22.17 0.30 77 23

48
C.E. Santolalla-Vargas, et al. Catalysis Today 329 (2019) 44–52

Fig. 6. Gaussian deconvolution example for CatRef.

Fig. 8. X-ray photoelectron spectra of sulfided NiMoP/γ-Al2O3 catalyst: For
CatAXAceMo (13.41 Å). Besides, the CatAXAceMo showed the lower CatAXAceMo (a), CatAceMo (b) and CatRef (c) samples, deconvolution in Ni 2p
coke crystal size. This result indicates that coke crystal size is influenced emission line.
by the pretreatments before the HDS reaction. Also, these results sug-
gest that the shortest crystal size of coke provoked by washed with core level presented two doublets, indicating the presence of at least
additive and xylene is significantly modified compared to the reference two species. They could be deconvoluted with two sets of doubles
catalyst before and after the HDS reaction (CatRef). Besides, Guichard correspond to Mo3d5/2 and Mo3d3/2 components coming from the spin-
et al. [8] suggested that coke localizations exist over MoS2 particles and orbit splitting. The first set at 229.6 eV and 232.0 eV is related to the
others cover the Al2O3 support. In this sense, the pretreatment with MoS2 species while the second set at 233.0 and 236.2 eV correspond to
additive and xylenes may influence on the formation of MoS2 pre- MoO3 species [20,42–45].
venting the formation of large carbons. This result is in accordance with Table 7 summarizes the relative concentration of each species. For
the surface concentration of aromatic carbons evidenced by NMR CatRef catalyst, 9.3% of Mo species was converted to MoS2 species
spectroscopy for the catalyst CatAXAceMo. Therefore, the washed with (Table 7). This superficial concentration increase with the pretreat-
additive and xylenes could unblock the coke deposited on the active ments before HDS reaction, changing to 19.5% and 33.2% for CatA-
phase and increased the MoS2 surface concentration. ceMo and CatAXAceMo, respectively. This result involves that the spent
catalyst washed with xylene and additive before the in situ reactivation
improve the MoS2 active phase.
3.7. X-ray photoelectron spectroscopy of CatRef, CatAceMo and Fig. 8 shows the XPS spectra of the Ni 2p emission line for NiMoP
CatAXAceMo catalysts after HDS reaction spent catalyst. As seen in this figure, the spectrum exhibited six peaks
correspond to NiMoS, NiS and NiO. Since the line-shape of the Ni 2p3/2
The CatRef, CatAceMo and CatAXAceMo catalysts were analyzed by peak, and especially its satellite, was unsymmetrical and broad, we
X-ray photoelectron spectroscopy to analysis the superficial sulfided applied an empirical method to roughly fit the curve using Gaussian/
species. The evolution of MoO3, MoS2, NiMoS, NiS2 and NiO species Lorentzian distributions. For instance, the deconvolution of the Ni 2p3/2
after the HDS reactions was investigated. The high-resolution spectrum profile made for the CatAXAceMo sample is shown in Fig. 8. Each Ni
for the Mo 3d emission line and deconvolution (220 eV–240 eV) was 2p3/2 profile was resolved in three components at 852.5 eV, 853.6 eV
recorded to quantify the MoS2 and MoO3 (Fig. 7). To Ni 2p emission and 855.9 eV, ascribing to NiS, NiMoS and NiOx, respectively
line (888 eV-850 eV) was recorded for NiS, NiMoS and NiO (Fig. 8). [20,42–45]. The BE at 852.5 eV is close value reported in the literature
Regarding the pretreatment, all catalysts exhibited similar profile for NiS supported on MoS2 while the BE at 853.6 is close value to re-
spectra, therefore, the deconvolution was measured by the same para- ported for NiMoS species formed after sulfidation of the NiMo sup-
meters (FWHM, position and area construction). As expected, the Mo 3d ported on Al2O3. According to the binding energy position and the re-
lative area related to the three Ni species, the percentages of each
species were measured. Table 7 shows that CatRef presented 7.9% of
NiMoS species, while the CatAXAceMo increased approximately 50%
more NiMoS species than the catalyst reference. Also, Table 7 exhibits
that the NiO species incremented 10% for CatAceMo catalyst in com-
parison with CatRef and CatAXAceMo. On the other hand, Table 8
shows the C, Ni and Mo surface concentration and the sulfidation

Table 7
Percentage of the surface species of freshly sulfided CoMoP/γ-Al2O3 catalysts.
Catalyst Ni 2p core level (%) Mo 3d core level (%)

NiMoS NiS NiOx MoS2 MoSxOy Mo6+

CatRef 7.9 10.7 81.5 9.3 37.8 52.9

CatAceMo 5.4 5.7 88.9 19.5 27.1 53.4
Fig. 7. X-ray photoelectron spectrum of sulfided NiMoP/γ-Al2O3 catalyst: For
CatAXAceMo 11.6 7.7 80.7 33.2 16.7 50.1
the CatAXAceMo sample, deconvolution in Mo 3d emission line.

49
C.E. Santolalla-Vargas, et al. Catalysis Today 329 (2019) 44–52

Table 8
Mo and Ni surface atomic ratios of NiMoP/γ-Al2O3 sulfided catalysts.
Sample Mo/Al Ni/Al Ni + Mo/Al S/Ni + Mo C/Al

CatRef 1.41 0.11 1.52 0.24 7.60

CatAceMo 2.67 0.49 3.15 0.09 5.48
CatAXAceMo 2.13 0.43 2.56 0.07 3.14

degree as C/Al, (Ni + Mo)/Al and S/(Ni + Mo), respectively. The Ca-
tRef exhibits a lower Mo species surface exposure than CatAceMo and
CatAxAceMo. Take into account the in situ reactivation with mo-
lybdenum acetylacetonate, the lower concentration is expected. The
CatRef presented a higher C surface concentration in comparison with
CatAceMo and CatAXAceMo. The decreasing is related to the washed
with xylene, organic additive and the in situ reactivation. Finally, there
is a significant variation of sulfidation degree of Mo and Ni species for
Fig. 9. Relation between the coke crystallite size and the aromatics superficial
CatAceMo and CatAxAceMo. However, this is related to the larger Ni
concentration for CatRef, CatAceMo and CatAXAceMo after HDS straight run
and Mo surface concentration than CatRef (Table 7).
gas oil reaction.

4. Discussion

The experiments with pretreatments as washed with xylene and an

additive or in situ reactivation were carried out to increase the catalytic
activity of NiMoP/Al2O3 spent catalyst. The CatAXAceMo exhibited a
better performance in HDS straight run gas oil than CatAX, CatAceMo
and CatRef.
On the one hand, the TPR results suggested that CatAX presented
higher sulfur species with strong interaction than CatRef which suggest
lower sulfur vacancies (Fig. 3). Also, the lower sulfur vacancies are
related to lower catalytic activity in HDS reactions. However, the CatAX
exhibited similar conversion than the CatRef; 94.2% vs 94.9%, re-
spectively. This fact may be related to xylene and the organic additive
that could be leaching the Ni and Mo metals from the spent catalyst
(Figure S1), and therefore, the catalytic activity keeps the same beha-
vior in spite the washed remove superficial carbon.
On the other hand, the catalyst with in situ reactivation without Fig. 10. Relation between the coke crystallite size and superficial carbon for
washed with xylene and additive (CatAceMo) showed lower catalytic CatRef, CatAceMo and CatAXAceMo catalysts after HDS straight run gas oil
activity than CatRef. This effect could be related to the double incre- reaction.
ment of Mo surface concentration for CatAceMo in comparison with
CatRef. This increasing could lead to forming MoO3 three-dimensional
and MoS2 with larger size wherein both species result in a lower Mo
dispersion [46]. This result is evidenced by XPS (Table 8). In this re-
gard, Iwamoto et al. [47] suggested that the regeneration with milder
conditions favored the formation of active metal and increased the
catalytic activity in HDS of thiophene.
To elucidate the effect of coke on the catalytic activity, the CatRef,
CatAceMo and CatAXAceMo were compared. Fig. 9 showed the relation
between the coke crystallite size and the aromatics superficial con-
centration for CatRef, CatAceMo and CatAXAceMo after HDS straight
run gas oil reaction. Fig. 9 suggests that the concentration of aromatics
carbon increased the formation of coke with a larger size. In this sense,
Fig. 10 presents the relation between elemental analysis of carbon and
the size of coke crystallites. Fig. 10 points out that the catalyst with the
highest carbon superficial forms highest coke crystallites (CatRef). On
the other part, Bui et al. [48] studied the sulfide state of CoMo/Al2O3
regenerated with maleic acid. This author suggested that the chelating Fig. 11. Relation between the coke crystallite size and MoS2 superficial con-
agent favored a more advanced initial sulfidation of Mo (MoS2). In this centration for CatRef, CatAceMo and CatAXAceMo catalysts after HDS straight
regard, the formation of MoS2 species is related to the responsible for run gas oil reaction.
catalytic activity in HDS of straight-run gas oil and 4,6-DMDBT [48]. In
this aspect, Fig. 11 displays the formation of MoS2 with the coke for-
mation. Fig. 11 shows that the MoS2 species are incremented when the
coke crystallites size decreased. This effect may be related that the
shorter coke size unblocks Mo species and increment the sulfidation
degree to the formation of Mo active phase. The formation of MoS2
leads to better performance in HDS reactions. Indeed, this evidence is in

50
C.E. Santolalla-Vargas, et al. Catalysis Today 329 (2019) 44–52

agreement with the highest catalytic activity for CatAXAceMo in com- Hydrotreating Catalyst, the Resulting Restored Catalyst, and Method of
parison with CatRef and CatAceMo. Hydroprocessing, (2004).
[16] J. Chen, F. Maugé, J. El Fallah, L. Oliviero, IR spectroscopy evidence of MoS2
morphology change by citric acid addition on MoS2/Al2O3 catalysts – a step for-
5. Conclusions ward to differentiate the reactivity of M-edge and S-edge, J. Catal. 320 (2014)
170–179.
[17] L. Medici, R. Prins, The influence of chelating ligands on the sulfidation of Ni and
The in situ reactivation with molybdenum acetylacetonate pre- Mo in NiMo/SiO2Hydrotreating catalysts, J. Catal. 163 (1996) 38–49.
viously washed with xylene and 2,6 bis 1 hydroxy 1,1 diphenyl methyl [18] L. Medici, R. Prins, Structure of oxidic NiMo/SiO2Hydrotreating catalyst pre-
pyridine on spent catalyst, has been investigated in the hydro- cursors, J. Catal. 163 (1996) 28–37.
[19] C.E. Santolalla-Vargas, V. Santes, C. Ortega-Niño, A. Hernández-Gordillo,
desulfurization (HDS) of straight-run gas oil. Xylene and 2,6 bis 1 hy- F. Sanchez-Minero, L. Lartundo-Rojas, R. Borja-Urby, J.C. López-Curiel, O. Goiz,
droxy 1,1 diphenyl methyl pyridine removed the aromatic carbons and I.I. Padilla-Martinez, Effect of trimesic acid as chelating agent in sulfided CoMoP/γ-
coke with higher crystallites size from the spent catalyst. Nevertheless, Al2O3 catalyst for hydrodesulfurization of straight-run gas oil, Catal. Today (2018).
[20] E.C. Santolalla-Vargas, V. Santes, E. Meneses-Domínguez, V. Escamilla,
the Mo superficial concentration could decrease by the leaching pro-
A. Hernández-Gordillo, E. Gómez, F. Sánchez-Minero, J. Escobar, L. Díaz, O. Goiz,
cess. On the contrary, the in situ reactivation with molybdenum acet- Effect of 2,6-Bis-(1-hydroxy-1,1-diphenyl-methyl) pyridine as organic additive in
ylacetonate favors the deposition of Mo metals which incremented the sulfide NiMoP/γ-Al2O3 catalyst for hydrodesulfurization of straight-run gas oil,
Mo surface catalyst content. However, this increment of Mo content Molecules 22 (2017).
[21] C.E. Santolalla-Vargas, V.A. Suárez Toriello, J.A. de los Reyes, D.K. Cromwell,
may promote Mo segregates. According to catalytic activity results, Mo B. Pawelec, J.L.G. Fierro, Effects of pH and chelating agent on the NiWS phase
segregates no promoted the sulfur conversion in HDS of straight run gas formation in NiW/γ-Al2O3 HDS catalysts, Mater. Chem. Phys. 166 (2015) 105–115.
oil. In contrast, the use of xylene and 2,6 bis 1 hydroxy 1,1 diphenyl [22] E. Gómez, R. Flores, G. Huerta, C. Alvarez-Toledano, R.A. Toscano, V. Santes, amp,
x, ctor, N. Nava, P. Sharma, Dimethyltin(IV) 2,6-disubstituted pyridine complexes,
methyl pyridine with molybdenum acetylacetonate as rejuvenation/ J. Organomet. Chem. 672 (2003) 115–122.
regeneration method enhanced the MoS2 species, the lower carbon [23] W.G. Appleby, J.W. Gibson, G.M. Good, Coke formation in catalytic cracking, Ind.
concentration and shorter coke crystallites size which resulted in a Eng. Chem. Process. Des. Dev. 1 (1962) 102–110.
[24] V.H.J. de Beer, F.J. Derbyshire, C.K. Groot, R. Prins, A.W. Scaroni, J.M. Solar,
higher gas oil HDS activity. Therefore, the use of both pretreatments Hydrodesulphurization activity and coking propensity of carbon and alumina
leads to suggest as a novel process in the reuse of spent catalyst for supported catalysts, Fuel 63 (1984) 1095–1100.
hydrotreating in comparison with the regeneration of aged catalyst [25] L. Díaz, J. Herbert, M.T. Cortez, R. Zárate, V. Santes, G. Hernández, Heavy gas oil
hydrotreating over NiMo supported on alumina and alumina–silica, Pet. Sci.
with a conventional reactivation process. Technol. 22 (2004) 141–155.
[26] S.T. Sie, Miniaturization of hydroprocessing catalyst testing systems: theory and
Acknowledgment practice, Aiche J. 42 (1996) 3498–3507.
[27] L.F. Ramírez, J. Escobar, E. Galván, H. Vaca, F.R. Murrieta, M.R.S. Luna, Evaluation
of diluted and undiluted trickle-bed hydrotreating reactor with different catalyst
The authors acknowledge financial support from SENER CONACYT- volume, Pet. Sci. Technol. 22 (2004) 157–175.
136363 and Instituto Politécnico Nacional (Proyecto SIP 20181105 SIP [28] C.P. Mane, S.V. Mahamuni, S.S. Kolekar, S.H. Han, M.A. Anuse, Hexavalent chro-
20181076 and SIP 20182337). Also, we acknowledge financial support mium recovery by liquid–liquid extraction with 2-octylaminopyridine from acidic
chloride media and its sequential separation from other heavy toxic metal ions,
from CONACYT-México (Grant 237857). Arab. J. Chem. 9 (2016) S1420–S1427.
[29] L.E. Kallinikos, A. Jess, N.G. Papayannakos, Kinetic study and H2S effect on re-
Appendix A. Supplementary data fractory DBTs desulfurization in a heavy gasoil, J. Catal. 269 (2010) 169–178.
[30] A. Stanislaus, A. Marafi, M.S. Rana, Recent advances in the science and technology
of ultra low sulfur diesel (ULSD) production, Catal. Today 153 (2010) 1–68.
Supplementary data associated with this article can be found, in the [31] J.A. Bergwerff, T. Visser, B.M. Weckhuysen, On the interaction between Co- and
online version, at https://doi.org/10.1016/j.cattod.2019.03.015. Mo-complexes in impregnation solutions used for the preparation of Al2O3-sup-
ported HDS catalysts: a combined Raman/UV–vis–NIR spectroscopy study, Catal.
Today 130 (2008) 117–125.
References [32] B. Baubet, E. Devers, A. Hugon, E. Leclerc, P. Afanasiev, The influence of MoS2 slab
2D morphology and edge state on the properties of alumina-supported molybdenum
sulfide catalysts, Appl. Catal. A: Gen. 487 (2014) 72–81.
[1] Norma Oficial Mexicana NOM-016-CRE-2016, (2016).
[33] P. Afanasiev, On the interpretation of temperature programmed reduction patterns
[2] G. Ertl, H. Knözinger, J. Weitkamp, Handbook of Heterogeneous Catalysis, VCH,
of transition metals sulphides, Appl. Catal. A: Gen. 303 (2006) 110–115.
1997.
[34] M. Makarova, Y. Okawa, M. Aono, Selective adsorption of thiol molecules at sulfur
[3] M. Marafi, A. Stanislaus, E. Furimsky, Chapter 4 - catalyst deactivation, Handbook
vacancies on MoS2(0001), followed by vacancy repair via S–C dissociation, J. Phys.
of Spent Hydroprocessing Catalysts, Elsevier, Amsterdam, 2010, pp. 51–92.
Chem. C 116 (2012) 22411–22416.
[4] P. Dufresne, Hydroprocessing catalysts regeneration and recycling, Appl. Catal. A
[35] F. Dumeignil, J.-F. Paul, E.W. Qian, A. Ishihara, E. Payen, T. Kabe, Elucidation by
Gen. 322 (2007) 67–75.
computer simulations of the CUS regeneration mechanism during HDS over MoS2
[5] M. Marafi, A. Stanislaus, E. Furimsky, Chapter 4 - catalyst deactivation, Handbook
in combination with 35S experiments, Res. Chem. Intermed. 29 (2003) 589–607.
of Spent Hydroprocessing Catalysts, second edition, Elsevier, 2017, pp. 67–140.
[36] A. Ishihara, F. Dumeignil, D. Wang, X. Li, H. Arakawa, E.W. Qian, S. Inoue, A. Muto,
[6] E. Furimsky, F.E. Massoth, Deactivation of hydroprocessing catalysts, Catal. Today
T. Kabe, Investigation of sulfur behavior on CoMo-based HDS catalysts supported
52 (1999) 381–495.
on high surface area TiO2 by 35S radioisotope tracer method, Appl. Catal. A: Gen.
[7] M. Marafi, A. Stanislaus, E. Furimsky, Chapter 5 - regeneration, Handbook of Spent
292 (2005) 50–60.
Hydroprocessing Catalysts, second edition, Elsevier, 2017, pp. 141–220.
[37] B. Behera, P. Gupta, S.S. Ray, Structure and composition of hard coke deposited on
[8] B. Guichard, M. Roy-Auberger, E. Devers, B. Rebours, A.A. Quoineaud, M. Digne,
industrial fluid catalytic cracking catalysts by solid state 13C nuclear magnetic
Characterization of aged hydrotreating catalysts. Part I: coke depositions, study on
resonance, Appl. Catal. A: Gen. 466 (2013) 123–130.
the chemical nature and environment, Appl. Catal. A Gen. 367 (2009) 1–8.
[38] Y.T. Chua, P.C. Stair, An ultraviolet Raman spectroscopic study of coke formation in
[9] P. Wiwel, P. Zeuthen, A.C. Jacobsen, Initial coking and deactivation of hydro-
methanol to hydrocarbons conversion over zeolite H-MFI, J. Catal. 213 (2003)
treating catalysts by real feeds, in: C.H. Bartholomew, J.B. Butt (Eds.), Studies in
39–46.
Surface Science and Catalysis, Elsevier, 1991, pp. 257–264.
[39] H. Schulz, M. Wei, Deactivation and thermal regeneration of zeolite HZSM-5 for
[10] R. Cortés-Torres, E.Y. Nolasco-Terrón, O. Olea-Mejia, V. Varela-Guerrero,
methanol conversion at low temperature (260–290°C), Microporous Mesoporous
C.E. Barrera-Díaz, E. Cuevas-Yañez, Solvent mediated impurity removal process for
Mater. 29 (1999) 205–218.
a spent hydroprocessing catalyst and its use in alcohol oxidations, Catal. Today 305
[40] D. Espinat, H. Dexpert, E. Freund, G. Martino, M. Couzi, P. Lespade, F. Cruege,
(2018) 126–132.
Characterization of the coke formed on reforming catalysts by laser raman spec-
[11] H.A. Abdullah, A. Hauser, F.A. Ali, A. Al-Adwani, Optimal conditions for coke ex-
troscopy, Appl. Catal. 16 (1985) 343–354.
traction of spent catalyst by accelerated solvent extraction compared to soxhlet,
[41] J. Robertson, Diamond-like amorphous carbon, Mater. Sci. Eng. R Rep. 37 (2002)
Energy Fuels 20 (2006) 320–323.
129–281.
[12] J.A. Marinsky, Y. Marcus, Ion Exchange and Solvent Extraction: A Series of
[42] L. Díaz-García, V. Santes, T. Viveros-García, A. Sánchez-Trujillo, J. Ramírez-
Advances, Taylor & Francis, 1997.
Salgado, C. Ornelas, E. Rodríguez-Castellón, Electronic binding of sulfur sites into
[13] G. Ertl, H. Knözinger, J. Weitkamp, Handbook of Heterogeneous Catalysis, VCH,
Al2O3-ZrO2 supports for NiMoS configuration and their application for
1997.
Hydrodesulfurization, Catalysis Today 282 (Part 2) (2017) 230–239.
[14] R. Pascal, T. Hervé, Catalysis by Transition Metal Sulphides: From Molecular
[43] C.E. Scott, M.J. Perez-Zurita, L.A. Carbognani, H. Molero, G. Vitale, H.J. Guzmán,
Theory to Industrial Application, Editions Technip, 2013.
P. Pereira-Almao, Preparation of NiMoS nanoparticles for hydrotreating, Catal.
[15] J.M. Ginestra, K.S. Lee, J.D. Seamans, A Method for Restoring Activity of a Spent
Today 250 (2015) 21–27.

51
C.E. Santolalla-Vargas, et al. Catalysis Today 329 (2019) 44–52

[44] T.A. Zepeda, B. Pawelec, R. Obeso-Estrella, J.N. Díaz de León, S. Fuentes, G. Alonso- [46] U. Usman, M. Takaki, T. Kubota, Y. Okamoto, Effect of boron addition on a MoO3/
Núñez, J.L.G. Fierro, Competitive HDS and HDN reactions over NiMoS/HMS-Al Al2O3 catalyst: physicochemical characterization, Appl. Catal. A: Gen. 286 (2005)
catalysts: diminishing of the inhibition of HDS reaction by support modification 148–154.
with P, Appl. Catal. B: Environ. 180 (2016) 569–579. [47] I. Iwamoto, J. Grimblot, J. Adv. Catal. Sci. Technol. 44 (2000) 417.
[45] T.K.T. Ninh, L. Massin, D. Laurenti, M. Vrinat, A new approach in the evaluation of [48] N.-Q. Bui, C. Geantet, G. Berhault, Maleic acid, an efficient additive for the acti-
the support effect for NiMo hydrodesulfurization catalysts, Appl. Catal. A: Gen. 407 vation of regenerated CoMo/Al2O3 hydrotreating catalysts, J. Catal. 330 (2015)
(2011) 29–39. 374–386.

52