You are on page 1of 11

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/308879368

Benzothiophene hydrodesulfurization over NiMo/alumina catalysts


modified by citric acid. Effect of addition stage of organic modifier

Article  in  Fuel Processing Technology · February 2017


DOI: 10.1016/j.fuproc.2016.09.028

CITATIONS READS

8 257

4 authors:

Jose Escobar María C. Barrera


Instituto Mexicano del Petroleo Universidad Veracruzana
53 PUBLICATIONS   660 CITATIONS    22 PUBLICATIONS   282 CITATIONS   

SEE PROFILE SEE PROFILE

Ana W. Gutiérrez J. E. Terrazas-Rodríguez


3 PUBLICATIONS   18 CITATIONS    Universidad Veracruzana
6 PUBLICATIONS   59 CITATIONS   
SEE PROFILE
SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Selective hydrodesulphurization catalysts View project

All content following this page was uploaded by Jose Escobar on 24 October 2016.

The user has requested enhancement of the downloaded file.


Fuel Processing Technology 156 (2017) 33–42

Contents lists available at ScienceDirect

Fuel Processing Technology

journal homepage: www.elsevier.com/locate/fuproc

Benzothiophene hydrodesulfurization over NiMo/alumina catalysts


modified by citric acid. Effect of addition stage of organic modifier
José Escobar a,⁎, María C. Barrera b, Ana W. Gutiérrez a, José E. Terrazas b
a
Instituto Mexicano del Petróleo, Eje Central Lázaro Cárdenas 152, San Bartolo Atepehuacan, Gustavo A. Madero, D.F., Mexico, 07730
b
Facultad de Ciencias Químicas-Centro de Investigación en Recursos Energéticos y Sustentables, Universidad Veracruzana, Campus Coatzacoalcos, Av. Universidad km. 7.5, Col. Santa Isabel,
Coatzacoalcos, Veracruz 96538, Mexico

a r t i c l e i n f o a b s t r a c t

Article history: NiMo/Al2O3 catalysts were obtained by one-pot simultaneous impregnation of Mo, Ni and P (at 12, 3 and 1.6 wt%,
Received 3 June 2016 respectively) over alumina. Citric acid (CA) was added (Ni/CA=1 mol ratio) to determine its influence over mo-
Received in revised form 30 August 2016 lybdenum species, as nickel precursor used was complexated Ni acetate. Three different preparation methodol-
Accepted 25 September 2016
ogies were used: (a) modification of calcined (400 °C) NiMo/Al2O3 by CA impregnation, (b) CA deposition
Available online xxxx
directly onto Al2O3 carrier, prior to Ni-Mo-P impregnation, (c) simultaneous Ni-Mo-P-CA deposition on alumina
Keywords:
support. Materials were characterized by N2 physisorption, infrared and UV-vis spectroscopies, temperature-pro-
Citric acid grammed reduction and thermal analysis (TG-DTG), and tested in liquid-phase benzothiophene (BT) conversion
NiMo catalysts in batch reactor. CA deposition over calcined NiMo/Al2O3 resulted in diminished proportion of refractory Mo6+(t)
Hydrodesulfurization tetrahedral species but also in octahedral Mo6+(o) reducible at higher temperature, as to those over non-modified
Benzothiophene NiMo/alumina. That was reflected in lower hydrodesulfurizating (HDS) ability. The highest BT HDS activity (70%
Chelating agent increase, as to the conventional calcined formulation with no organic additive) was found when CA was impreg-
nated over bare alumina prior to Ni-Mo-P deposition. In that case, decreased interaction between deposited mo-
lybdenum and nickel species and the support was evidenced, that due to surface “passivation” from citric
adsorption/decomposition on alumina surface. On the other hand, simultaneous Ni-Mo-P-CA deposition,
where low pH of impregnating solution (pH~1.4) devoided citric acid ionization, (then, no Mo-citrate complex
formation) was ineffective in providing catalyst of enhanced HDS activity.
© 2016 Elsevier B.V. All rights reserved.

1. Introduction The beneficial influence of chelators that could complexate Ni or Co


promoters has been addressed by various groups [2–4]. Most of authors
Continuous tightening of environmental legislations that constrains [5–8] coincide in that stable complexes formed by promoter chelation
sulfur content in oil-derived fuels to extremely low concentration de- could resist low-temperature sulfiding until conditions at which Mo
mands for improved S-removal processes [1]. Posing an extra challenge, (or W) component could be transformed to corresponding sulfides.
low-quality heavy crudes of enhanced concentration of organic precur- Once those MoS2 (or WS2) particles are present, organo-Co or -Ni spe-
sors of atmospheric pollutants have to be used as raw materials in cies could be decomposed allowing promoters sulfidation. The highly
obtaining those fuels, as light-oil resources are progressively worldwide active “CoMo(W)S″ or “NiMo(W)S” phases could be then preferentially
depleted. Thus, more efficient hydrodesulfurization (HDS) processes, formed by efficient Mo (W) sulfide slab edges decoration, precluding
including both optimized industrial facilities set-up and improved cata- segregation of sulfided Co/Ni phases that could result in catalysts of
lytic formulations are clearly needed. minor HDS properties.
Several approaches have been attempted in order to obtain mate- However, other positive roles could be played by chelating agents
rials of enhanced HDS activity as to that of traditional alumina-support- added during hydrotreating catalysts synthesis. Among them, carbon
ed Ni(Co)Mo(W) sulfided formulations. Those strategies have been deposition [9] and decreased deposited phases-support interaction [2]
mainly focused on improving dispersion and sulfidability of supported have been evidenced. Interaction between chelating ligands and car-
phases or in optimizing decoration of MoS2 crystals by promoters (co- riers could strongly influence properties of final hydrotreating catalysts.
balt or nickel) that resulted in improved formation of the highly active Isoelectric point and type of hydroxyls present on a given substrate
“Ni(Co)Mo(W)S” phase. could determine that interaction degree [3]. For instance, carboxylate
groups in various chelators play an important role on those effects
⁎ Corresponding author. through their affinity to OH groups on support surface [10]. For SBA
E-mail address: jeaguila@imp.mx (J. Escobar). 15-supported NiMo catalysts Badoga et al. [2] found that EDTA-support

http://dx.doi.org/10.1016/j.fuproc.2016.09.028
0378-3820/© 2016 Elsevier B.V. All rights reserved.
34 J. Escobar et al. / Fuel Processing Technology 156 (2017) 33–42

interaction (via surface OH groups) weakened the Mo-carrier one favor- In the second methodology the carrier was firstly impregnated by
ing molybdenum sulfiding that was reflected in improved activity in pore-filling with an aqueous CA solution (concentration corresponding
light gas oil HDS. to Ni/CA = 1 mol ratio), then dried at aforementioned conditions. Sub-
Depending on pH and concentration citric acid could complexate sequently, Ni-Mo-P phases were simultaneously deposited over the
both Ni and Mo, contributing in formation of highly dispersed NiMoS modified support by one-pot pore-filling impregnation by using the
active phase of enhanced activity in refractory 4,6-dimethyl- previously described solution. Resulting solid was then dried at 120 °C
dibenzothiophene HDS [11]. In this case, the hydrogenation route was (2 h), sample NMII. In the third methodology CA was directly added
strongly improved in KIT-6 mesoporous silica-supported NiMo catalysts (at Ni/CA = 1 mol ratio) to the already prepared Ni-Mo-P aqueous solu-
by adding citric acid at conditions where both Ni and Mo species could tion that was then used to impregnate the Al2O3 carrier. The impregnat-
be complexated. The organic acid provided efficient bridging between ed material was finally dried at 120 °C (2 h), sample NMIII.
impregnated species and the silica carrier enhancing their interaction,
the deleterious effect of the rather inert SiO2 surface being then 2.2. Materials characterization
precluded.
In the opposite, different organics have also been used to deposit Textural properties of various materials were determined by N2
carbon layers over alumina supports intending to weaken the strong in- physisorption (at −198 °C), in an Autosorb-1 (Quantachrome) appara-
teraction between the carrier and impregnated Mo, W, Ni and Co tus. Studied solids were previously degassed at 300 °C (2 h) at high vac-
phases. In that line, Nikulshin et al. [12] found that depending on con- uum (133.32 × 10−5 Pa) to eliminate any adsorbed molecules. Surface
tent of carbon deposited by pyrolyzing an isopropanol-glycerin (2 area and average pore size were determined though BET equation and
vol.%) mixture over Al2O3 sulfided NiW catalysts of enhanced properties Gurvich's rule [16], respectively.
in dibenzothiophene HDS, naphthalene hydrogenation and quinoline FTIR (Fourier-transform infrared) spectra were acquired by a Perkin
hydrodenitrogenation could be obtained. The best results were found Elmer Frontier FT-IR equipment with an attached Praying Mantis Dif-
at approximate carbon monolayer (C ~5 wt%). fuse Reflection Accessory in the 3750–500 cm−1 range (alumina sup-
In a previous paper [13], the addition stage at which saccharose as port in the 4000–1000 cm−1 range).
non-complexating additive should be added to obtain sulfide NiMo/ Cobalt and molybdenum coordination in oxidic impregnated mate-
Al2O3 catalysts of improved HDS properties was determined. Among rials were studied by UV–vis (diffuse reflectance spectroscopy) by uti-
preparation procedures studied, simultaneous one-pot Ni + Mo + lizing a Cary 100 equipment in the 200–800 nm wavelength range.
P + SA deposition rendered the material of the highest Supported Ni-Mo samples (approx. 100 mg in a quartz cell) were
dibenzothiophene desulfurization activity. studied by temperature-programmed reduction (TPR, in the 25–
Similarly, the present contribution is focused on determining the 900 °C range, 0.166 °C s−1 heating rate) by using an ICID SRyC-2 appa-
stage at which citric acid (complexing agent) should be added (at Ni/ ratus equipped with thermal conductivity detector. A H2/N2 mixture (at
CA = 1 mol ratio) to obtain materials of enhanced activity in organo-S 10 v/v %, 0.33 cm3 s−1 total flow) was used.
compounds removal. One-pot simultaneous Ni-Mo-P impregnation Thermogravimetric analysis (TGA) of impregnated samples were
was used as that preparation method is relevant to commercial catalysts carried out through a TA instruments Q2000 equipment. 15–50 mg
production. This work was focused on CA effect over molybdenum spe- samples were analyzed in the 30–1000 °C temperature range (heating
cies as nickel precursor was complexated Ni acetate. Differently to the ramp of 0.166 °C s−1). Differential thermogravimetric profiles were ob-
work of Wu et al. [11] where CA was added to increase interaction be- tained through derivatives of corresponding TG curves.
tween KIT-6 mesoporous silica and Ni and Mo species, in our work we
tried to tune up the already strong affinity between those deposited 2.3. HDS reaction test
species and the alumina carrier. Prepared solids were characterized
through various instrumental techniques to try to explain the activity Prior to be tested in benzothiophene hydrodesulfurization NiMo/
trend found in benzothiophene HDS. alumina oxidic precursors were submitted to activation. Materials
were ex-situ sulfided in a pyrex glass tubular reactor under H2:H2S
8:1 gas flow (approx. 0.944 ml s−1). Heating rate from room tempera-
2. Experimental ture to 400 °C was 0.166 °C s−1, sulfiding stage taking place for 7200 s
at final conditions. Then, the system was cooled-down to room temper-
2.1. Materials preparation ature and ~ 240 mg of sulfided catalyst were recovered in 20 ml of n-
heptane under inert (N2) atmosphere provided by a glove box, preclud-
Al2O3 was obtained by annealing (500 °C, 5 h under static air) com- ing contact of activated catalysts with atmospheric oxygen to prevent
mercial Pural SB sol-gel pseudo-boehmite (from Sasol) obtained from their oxidation to corresponding sulfates. In order to discard control
Al-alkoxides. Oxidic catalyst precursors were prepared by one-pot im- by internal diffusion/counterdiffusion of reactants/products catalyst
pregnation of Mo, Ni and phosphorous (at 12, 3 and 1.6 wt%, respective- particle size was chosen from the fraction retained between U.S. mesh
ly) by incipient wetness through an aqueous solution where MoO3 was size 80–100 (165 × 10−6 m average particle diameter) which according
digested (at ~ 80 °C) in diluted H3PO4. After ~ 2 h nickel acetate to our extensive experience in HDS reaction tests effectively guarantee
(NiC4H6O4·4H2O) was added, the solution being then digested by addi- testing under kinetic regime [13,14].
tional 12 h (~ 80 °C). Three different methodologies were followed to Sulfided catalysts were tested in benzothiophene (BT, ~0.30 g) HDS
obtain impregnated materials: Method I comprised Al2O3 impregnation in n-heptane in a slurry stainless-steel (T316 SS) Parr batch reactor 4570
with aforementioned solution (at pH ~ 1.4 [14]) followed by drying (2 h, (Scheme 1). BT well-represents less-reactive organo-sulfur species in
120 °C, sample NM120) then calcining (400 °C for 5 h, key NM400). FCC naphtha (although it is also present in light middle distillates),
Right after, part of that calcined material was impregnated with citric whereas n-heptane corresponds to hydrocarbons within naphtha boil-
acid at Ni/CA = 1 mol ratio (sample NMI) being then dried at 120 °C, ing range [17]. S concentration in the mixture was ~ 1040 ppm, value
(2 h) high-temperature annealing being avoided to preserve organic ad- typically found in fluid catalytic cracking naphtha. Reaction volume
ditive integrity. Citric acid decomposition could start at about 165 °C was adjusted to 100 ml by adding the required amount of solvent. Reac-
[15]. Then, in order to effectively determine its influence on HDS cata- tion conditions were 250 °C, 7.2 MPa and ~ 105 rad s−1 mixing speed
lyst properties through various preparation steps it was crucial to (Parr magnetic drive). The last value was carefully chosen based in pre-
keep intact different functionalities (chelating agent, carbonaceous de- vious experience to avoid reaction control by external diffusion phe-
posit precursor, etc.) by precluding annealing at severe conditions. nomena [13]. Reaction temperature was similar to those used in
J. Escobar et al. / Fuel Processing Technology 156 (2017) 33–42 35

naphtha hydrotreating commercial facilities [18]. Also, by operating at Table 1


used pressure it was assured that most of the solvent remains in liquid Textural properties of alumina support and NiMo impregnated materials (dried and cal-
cined), as determined by N2 physisorption (−198 °C).
phase (n-heptane boiling temperature ~264 °C) allowing proper opera-
tion of the slurry tri-phasic reactor. Operating pressure was monitored Sample SBET Vp Dpa
by both T316 stainless steel bourdon gage and stainless steel-housing (m2 g−1) (cm3 g−1) (nm)

transducer. Initial benzothiophene concentration was CBT0 = 0.0223 Al2O3 250 0.51 8.2
mol l−1 at room temperature and atmospheric pressure (0.01546 mol NM120 174 0.30 6.9
NM400 196 0.36 7.3
l−1 under reaction conditions, considering n-heptane density = 0.47 g
cm-3 [19]). a
From 4 × VpS−1
BET.

Nitrogen (at 7.5 MPa) was fed into the rector to remove air traces
then to check hermeticity of the experimental set-up. After confirming BT conversion (x) at a given time (t) was determined by (see
airtight system (0.5 h), N2 was evacuated and heating through corre- Nomenclature section):
sponding cylindrical oven (Scheme 1) started meanwhile stirring (at ~
105 rad s−1) uniformized radial thermal profile. Temperature monitor- AEB
xBT ¼ ð1Þ
ing and control was carried out through a type-J (iron-constantan) ther- ABT þ AEB
mocouple attached to corresponding 4848 Parr Instrument Company
controller. Once reaction temperature (250 °C) was reached, H2 was HDS kinetic constants (k) were calculated assuming pseudo-first
fed until getting 7.2 MPa total pressure. Zero time was considered at order kinetics referred to BT concentration:
that moment. A deep tube allowed periodically obtaining liquid sam-
−RBT ¼ k1 C BT C H2 ð2Þ
pling at various reaction times (from 0 to 7200 s, each 600 s, 11 samples
in total) to determine BT conversion. Operating pressure slightly de-
creased after sampling, H2 being then fed to recover original conditions. Constant H2 concentration in solvent (CH2) was assumed due to both
Liquid samples were analyzed in an Agilent 6890 N gas chromato- very high excess fed and negligible hydrogen consumption under reac-
graph provided with flame ionization detector and HP-5 (5% phenyl- tion conditions.
95% methyl-polysiloxane) non-polar capillary column (50 m, Then,
0.20 mm, 0.50 μm, length, diameter and film thickness, respectively).
−RBT ¼ kC BT ð3Þ
Good repeatability was confirmed by analyzing each liquid sample at
least twice. Under our reaction conditions, the only product detected
The mass balance of a batch reactor at constant volume,
was ethylbenzene (EB) which saturation to ethylcyclohexane was just
observed after total BT consumption. Although decompression (from Z
dC BT
7.2 MPa to atmospheric pressure) that took place during sampling RBT ¼ ð4Þ
dt
could provoke components evaporation we considered that due to its
higher volatility and corresponding required phase-change enthalpy
Integral reactor regime was considered as at high conversions HDS
[20] n-heptane rapid vaporization resulted in liquid samples cooling
reaction rate was dependent on BT transformation degree. As CBT =
fast enough to avoid BT and EB significant losses. Considerable higher
CBT0 (1-x), by combining Eqs. (3) and (4):
vaporization enthalpy of the latter [21] could also contribute to that. Al-
though due to solvent evaporation BT and EB concentration in analyzed Z
dC BT0 ð1−xÞ
liquid samples could differ from the actual one in the reactor, BT conver- −kC BT0 ð1−xÞ ¼ ð6Þ
dt
sion determinations could be considered accurate as they were calculat-
Z Z
ed from the proportion of those species in a given sample. dx
k dt ¼ ð7Þ
ð1−xÞ

− ln ð1−xÞ
k¼ ð8Þ
t

Fig. 1. Room temperature Fourier-transform infrared spectra (4000–1000 cm−1 region) of


Scheme 1. Tri-phasic slurry batch reactor used during catalytic tests (benzothiophene alumina-supported NiMo oxidic materials with and without citric acid prepared through
hydrodesulfurization) of various sulfided catalysts prepared. various protocols. Infrared spectrum of alumina support also included as reference.
36 J. Escobar et al. / Fuel Processing Technology 156 (2017) 33–42

Fig. 2. Room temperature Fourier-transform infrared spectra (2000–1000 cm−1 region) of


alumina-supported NiMo oxidic materials with and without citric acid prepared through Fig. 4. Visible region DRS spectra (Schuster-Kubelká-Munk function) of alumina-
various protocols. Infrared spectrum of alumina support also included as reference. supported NiMo oxidic materials prepared with and without citric acid through various
protocols.

Kinetic constants were normalized by mass of catalyst (k in m3 kg−1


cat
s−1). Undissociated water adsorbed on the carrier originated an intense
broadband in between 1500 and 1700 cm−1 [25], Fig. 2. For samples
3. Results and discussion prepared with CA the absorption at ~ 1570 cm−1 (more intense in
NMII) could be assigned to υas stretch of carboxylate groups [26] pre-
3.1. Textural properties served after impregnation. The absorption at ~1460 cm−1 (more nota-
ble in NMI) was also related to carboxylates (υs OCO vibrations) [27]. A
Lower Sg of NM400 as to that of bare alumina support, nicely band due to interactions between COO– and OH groups on the carrier
corresponded to textural diminution by deposition of ~25 wt% of non- was recorded at ~ 1415 cm−1 [28]. That signal could correspond to υs
porous Ni-Mo-P phases, Table 1. That suggested well-dispersed impreg- COO– (originally found at 1390 cm−1 in free citrate groups [29]) that
nated species as significant textural losses by partial plugging of support blue-shifted after deposition over the alumina substrate. The highest in-
porous network was not observed. tensity of that absorption in NMI suggested stronger interaction be-
tween carboxylate groups and acidic hydroxyls (type III, triply bonded
3.2. FTIR (Fourier-transform infrared) spectroscopy to Al3 + atoms in either octahedral or tetrahedral coordination [22])
remnant on Al2O3 carrier after Ni-Mo-P phases deposition-calcining.
From FTIR spectra in Fig. 1 bands due to stretching vibrations of sur- Disappearance of shoulder at ~ 3550 cm−1 in NM400 spectrum after
face OH groups at 3780 (type Ia, basic, single-bonded to one octahedral CA impregnation (sample NMI in Fig. 1) supported that observation.
aluminum [22]) and 3696 cm−1 (type IIb, neutral, bridged to two octa- Also, absorptions due to C_O in carboxylates complexating Ni2+ [26]
hedral Al3+ [23]) on Al2O3 were not evident in impregnated materials were observed at 1621 cm− 1 for CA-modified solids (shoulder in
due to preferential deposition of molybdate anions over those types of NMIII spectrum).
surface OH groups [24]. The absorption at 1660 cm−1 which intensity decreased following
Absorptions at 2970 and 2857 cm−1 related to –CH3 degenerated the order NMII N NMI N NMIII was related to Mo complexation. The
stretching mode and –CH2 stretching symmetrical vibrations, respec- presence of [(MoO2)2O(H2cit)2]2−·4H2O anions with infrared absorp-
tively, appeared in all spectra but in NM400 (no organic additive). In tion at 1657 cm− 1 [30] was then suggested in our CA-impregnated
general, in the 400–3000 cm−1 region studied samples spectra were solids. From spectra in Fig. 2 it seemed that CA deposition over bare alu-
pretty similar to those of alike prepared with saccharose as modifier mina prio to Ni-Mo-P impregnation (sample NMII) was the most effec-
[13]. tive methodology to obtain chelated Mo species. By contacting alumina
with CA aqueous solutions at pH b 5 citrate anions could react with
Lewis acid sites on carrier [29] resulting in highly negative charged sur-
face then shifting the point of zero charge (usually between 7 and 9 for
Al2O3 [31]) to lower pH. Clearly, that could occur during NMII prepara-
tion when impregnating the carrier with CA solution [~1 M, pH ~ 2.95].
Formation of other complexated Mo species could also be consid-
ered. In this regard, the possible presence of [Mo4(C6H5O7)2O11]4− an-
ions could originate the signal at 1736 cm− 1 (in NMII, shoulder in

Table 2
Band-gap energy (BGE) and average number of nearest Mo6+ neighbors (NMo) in depos-
ited clusters, as determined from UV spectra of NiMo impregnated materials prepared
with and without citric acid.

Sample BGE NMo


(eV)

NM400 3.56 2.47


NMI 3.76 1.70
NMII 3.89 1.20
Fig. 3. UV DRS spectra (Schuster-Kubelká-Munk function) of alumina-supported NiMo
NMIII 3.83 1.46
oxidic materials prepared with and without citric acid through various protocols.
J. Escobar et al. / Fuel Processing Technology 156 (2017) 33–42 37

Fig. 5. Temperature-programmed reduction profiles of alumina-supported NiMo oxidic Fig. 6. Thermogravimetrical analysis profiles of alumina-supported NiMo materials
materials prepared with and without citric acid through various protocols. prepared with and without citric acid through various protocols.

NMI) ascribed to non-bonding non-dissociated υ(C_O) in carboxylates addition provoked both increased intensity (much more notable ab-
[27]. Those molybdenum complexes could also produce bands at 1567 sorption in NMII and NMIII) and hypsochromic shift in that signal
and 1414 cm−1 due to υas(COO) in monodentated species and to with maximum at 258, 263, 267 and 275 nm for NMII, NMIII, NMI and
υs(COO) bridiging molybdenum atoms, respectively. NM400, respectively. As blue-shifted low-energy absorption edge has
The slight shoulder at approximately 1685 cm−1 in NMII and NMI, been related to smaller MoxOy domains [33] better dispersed oxo-mo-
associated to “free” noncoordinatively bonded\\CO\\OH groups [26], lybdates were strongly suggested in CA-modified formulations. CA im-
was more defined in NMIII pointing out to carboxylates not pregnation onto calcined sample NM400 (sample NMI) could not
complexating metallic atoms. That could be explained by the thermody- significantly improve molybdenum dispersion. Even more, the intense
namically non-favored CA dissociation in the highly acidic (pH ~ 1.4, signal in the 276–310 nm range in NMI spectrum strongly suggested
[14]) simultaneous one-pot Ni-Mo-P impregnating solution. Based on less-dispersed Mo6+ domains.
diagrams of Ni(II) and Mo(VI) species in CA aqueous solutions at differ- Regarding visible region spectra in Fig. 4, absorption at 401 nm in
ent pH Wu et al. [11] found no Mo complexes at conditions similar to NMII (shoulder in NMIII) was associated to octahedral Ni2+ [34]. That
those used during NMIII preparation (pH ~ 2, [Cit] = 1.3). Respecting signal was merged to low-energy Mo absorption edge in NM400 and
nickel, it could remain as corresponding acetate (from reactant used NMI spectra. Absorptions around 670 and 740 nm (3A2g → 3T1g (F) elec-
during impregnation, Section 2.1) under those impregnating conditions. tronic transitions) were also related to octahedral Ni2+ [35]. The former
Taking into account the CA amount added during catalysts synthesis nickel d-d transition blue-shifted from 676 nm in NM400 and NMIII to
(Section 2.1 Materials preparation) and the stoichiometry of aforemen- 669 nm in NMI and NMII pointing out to stronger coordinating ligands
tioned chelated Mo species, about 40–80% of molybdenum could be in the latter solids thus suggesting presence of complexated Ni2+ spe-
complexated whereas the rest could exist as polyoxomolybdates. cies [25]. Conversely, NM400 and NMIII could be associated to weaker
ligands (probably water molecules) in their coordination sphere.
3.3. UV–vis DRS spectroscopy For all samples but NMIII, the higher intensity of band at ~670 nm as
to that at 740 nm could indicate weak exchange interaction among Ni2+
From DRS UV–vis spectra of Fig. 3, all samples prepared with organic species suggesting well-dispersed supported nickel phases [35].
additive had high-energy absorption related to chromophore carbonyl The small shoulder at 588 nm for NM400 appearing just as hump in
in carboxylate groups (at 204 nm for NMI and NMII, at 208 nm for NMI suggested traces of tetrahedral Ni2+ thus pointing out to incipient
NMIII). The bathochromic shift in the latter suggested close interaction “surface spinel-like species” formation when submitting those solids at
of C_O moiety with OH groups [32] of undissociated carboxylate annealing (400 °C) [36].
under the strongly acidic conditions of one-pot Ni-Mo-P simultaneous
impregnation. 3.4. Band gap energy determinations
A broad inflexion which intensity decreased in the order NMII ≈
NMI N NMIII N NM400, suggested Mo species in different interaction de- To get better insight on oxidic molybdenum coordination in impreg-
grees with the alumina support. Position of that band precluded clear nated materials band-gap energy (BGE) of Mo6+ domains (from data
discrimination of tetrahedral and octahedral Mo6 + oxo-species. CA obtained through UV–vis spectra, Fig. 3) were determined. BGE values
were then correlated to molybdenum species of various coordination
states [33], Table 2.
Table 3
Maxima of signals from temperature-programmed reduction profiles of alumina-support-
ed NiMo materials (Fig. 5). Table 4
Weight losses at various temperature ranges from thermal analysis profiles of alumina-
6+ 4+ 2+ 0
Sample Mo (o) to Mo Ni to Ni Mo6+(t) to Mo 0
supported NiMo materials (Fig. 6).
(°C) (°C) (°C)
Sample b90 °C 175–310 °C 310–600 °C N800 °C Total
NM400 469 685 892c
(wt%) (wt%) (wt%) (wt%) (wt%)
NMI 498 750 892c
NMII 409, 462a 547b, 624b, 717 840b NM120 4.5 2.7 2.4 8.0 17.6
NMIII 451 669b, 825 891c NM400 3.4 2.3 1.8 9.6 17.1
a NMI 5.0 3.1 4.7 8.9 21.7
Shoulder.
b NMII 5.7 3.1 4.6 7.4 20.8
Hump.
c NMIII 3.5 3.6 3.8 6.2 17.1
And beyond.
38 J. Escobar et al. / Fuel Processing Technology 156 (2017) 33–42

that mentioned, NMI had enhanced Mo6 + dispersion as to that in


NM400 (Table 2). Very recently, others [39] have also reported molyb-
dates re-dispersion by citric acid addition over calcined (400 °C)
CoMo/alumina. In those cases, deposited Mo6 + redistribution was
more pronounced at CA/Co ratios of 1 and beyond.
Following similar trend to that of Mo, nickel reduction peak original-
ly observed at 685 °C in NM400 shifted to higher temperature in NMI
(750 °C), suggesting Ni-citrate complexes formation. The very wide sig-
nal related to Ni2+ reduction (~700 to 820 °C) in the latter pointed out
to supported species of various interaction degrees with the carrier. The
signal at ~890 °C and beyond from reduction of both Mo4+ (from octa-
hedral Mo6+ low-temperature partial reduction) and tetrahedral Mo6+
species strongly interacting with the support was significantly dimin-
ished by CA deposition over calcined NM400. That suggested partial car-
rier dissolution that could occur at the acidic pH (~ 2.95) of CA
impregnating solution resulting in extraction of Mo6+ species in high
interaction with the support. In full agreement, the shoulder at
588 nm in visible region spectrum of NM400 (Fig. 4) related to tetrahe-
dral Ni2+ in surface spinel-like species almost disappeared after CA ad-
dition over that formulation (sample NMI).
Fig. 7. Differential thermogravimetrical analysis profiles alumina-supported NiMo
materials prepared with and without citric acid through various protocols. Low-temperature (~100–285 °C) signals in NMI profile could be re-
lated to evolution of products from decomposition of CA loosely bound
to deposited species, Fig. 5. Oxidic impregnated species present over
CA addition contributed in obtaining augmented BGE, NMII showing NM400 could catalyze organic additive degradation at some extent. In
the maximum value. By analyzing UV–vis spectra of defined Mo species, this case, the broad signal in the aforementioned temperature range
Weber [33] found that larger iso-polymolybdates aggregates had lower could be due to alteration of the thermal conductivity detector by
BGE. Conversely, decreased BGE corresponded to increased estimated evolved species, discarding H2 consumption events. In this line, low-
average aggregation degree of oxidic molybdenum clusters. Thus, en- temperature signals (below 350 °C) in H2-TPR profiles of CA-modified
hanced BGE of CA-modified samples suggested higher Mo dispersion. CoMo/alumina calcined at various conditions have been attributed
The average number of nearest neighbors in Mo clusters (NMo) could [38] to organics decomposition. Also, according to Ramírez et al. [40]
be determined through: NMo = 16–3.8 × BGE [33]. CA addition over alu- low-temperature TPR signal (~117 °C) in CA/SiO2/Al2O3 could be related
mina prior to Ni-Mo-P deposition rendered the material of the lowest to organics reduction/decomposition.
NMo related to smaller Mo6 + domains (Table 2). The low NMo values Signals attributed to octahedral Mo6 + reduction shifted to lower
of CA-modified catalysts suggested that the complexated Mo species temperature (409 °C) in NMII, as to those in NM400 and NMI. That sug-
on those materials was [(MoO2)2O(H2cit)2]2− (see 3.2 FTIR (Fourier- gested molybdenum species in weaker interaction with the carrier thus
transform infrared) spectroscopy section). pointing out to support surface “passivation” due to organics deposition
Finally, it appeared that more efficient Mo chelation in NMII (Fig. 2) [40]. The shoulder at 462 °C could indicate non-complexated Mo con-
contributed in obtaining the highest molybdenum dispersion. Also, Ni sidering that amount of CA used was not sufficient to chelate totality
species could be well-dispersed over that solid (Fig. 4). Expectably, of impregnated molybdenum. Nickel species in various interaction de-
those properties could be reflected in improved HDS performance of grees with Al2O3 were suggested as corresponding reduction events
corresponding sulfided catalyst. originated several humps in ample temperature range. In this case, the
main signal due to octahedral Ni2 + was observed at 717 °C. As NMII
3.5. Supported phase characterization by temperature-programmed reduc- was not submitted to high-temperature annealing formation of Mo6+
tion (TPR) species in high interaction with the carrier was not expected. Thus,
the hump at 840 °C could be rather related to Mo4+ species (from for-
Maxima of signals from TPR profiles of Fig. 5 are summarized in mer octahedral Mo6+ reduction) conversion to metallic Mo.
Table 3. The signal around 469 °C in NM400 profile was attributed to re- Octahedral Mo6+ reduction at 460 °C in simultaneously impregnat-
duction of octahedral Mo6+ in oxomolybdates (identified by UV–vis, ed NMIII was accompanied by Ni2+ transformation to corresponding
Section 3.3) to Mo4+ [37,38]. The shift to higher temperature (498 °C) metallic phase in a wide range of temperature (slight hump centered
in NMI could be related to enhanced Mo6 + dispersion by CA- at 669 °C) again suggesting species at various interactions degrees
complexated species formation [9] (Section 3.2 FTIR (Fourier-transform with the support. At the highly acidic pH (~1.5 [14]) of impregnating so-
infrared) spectroscopy). Indeed, Rinaldi et al. [9] found that CA-post- lution used the weak organic acid could not be dissociated existing then
treatment of calcined (500 °C) CoMo/Al2O3-B2O3 dissolved both depos- as protonated acid ruling out complexation of either nickel or molybde-
ited CoMoO4 and polymolybdates, increasing then Co and Mo disper- num. In the same line, Li et al. [41] identified no Ni2+ chelation in acidic
sion. Even more, enhanced amount of CoMoS sites after sulfidation (pH = 1.5) impregnating solutions obtained from nickel nitrate, ammo-
resulted in augmented activity in thiophene HDS. In full agreement to nium metatungstate and CA.

Table 5
Maxima of signals from differential thermal analysis profiles of alumina-supported NiMo materials (Fig. 7).

Sample Low T Low T organics decomposition Medium T organics decomposition (°C) High T carbon residua combustion (°C) Mo6+
evaporation (°C) sublimation
(°C) (°C)

NM400 84 – – – 874
NMI 64 376 376, 453 – 857
NMII 51 296 – 503 842
NMIII 50 296 – 503 869
J. Escobar et al. / Fuel Processing Technology 156 (2017) 33–42 39

Table 6 organic moieties adsorption on the substrate, prior to Ni-Mo-P phases


Pseudo first-order kinetic constant (benzothiophene HDS) over various sulfided NiMo/ deposition.
alumina catalysts prepared with and without citric acid through various protocols. R-
squared values (coefficient of determination) of corresponding linear least squares fitting
also included. Batch reactor, n-heptane as solvent, P = 7.2 MPa, T = 250 °C, ~107 rads−1 3.6. Materials characterization by thermal analysis (TG-DTG)
(1030 rpm) mixing speed.
Impregnated material with no CA additive and dried at 120 °C (2 h,
Catalyst k × 10−5 R-squared
(m3 kg−1
cat s
−1
) sample NM120) was included as reference in thermogravimetrical anal-
ysis profiles shown in Fig. 6. Weight losses at temperature below 90 °C
NM400 7.94 0.95
NMI 4.20 0.91 (water boiling point in Mexico City) were related to physisorbed solvent
NMII 13.57 0.93 evaporation (Table 4). Losses between 175 and 310 °C could be originat-
NMIII 7.61 0.96 ed by support dehydroxylation (surface and structural [44]) and or-
ganics decomposition [15], including elimination of coordination
water from Ni acetate [45]. In full agreement, weight losses in that
A well-defined peak accompanied by an incipient signal (at 825 and range were slightly enhanced in CA-containing samples. Between 310
891 °C, respectively), could include contributions from Mo4+ transfor- and 600 °C rather similar weight losses for materials with CA strongly
mation to metallic phase and tetrahedral Mo6 + reduction as well. In suggested that in addition to nickel acetate decomposition burning of
our molybdophosphates impregnating solutions prepared from MoO3 carbon residua from citrates took place at that temperature range [15].
digestion in diluted H3PO4 (see Section 2.1 Materials preparation) Daw- Thermal phenomena beyond 800 °C were attributed to MoO3 sublima-
son heteropolyanions could be formed (P2Mo18O6− 62 ) [42], in spite of the tion [46].
large P/Mo ratio used (1/2.5 instead of 1/9). According to Blanchard et Temperature maxima of weight loss rate (from DTG profiles of Fig.
al. [42] those anions could co-exist with H6PMo9O3−34 entities accompa- 7) are summarized in Table 5. Signal around 300 °C in NMII and NMIII
nied by large phosphates excess. After impregnation over alumina the profiles nicely corresponded to Ni acetate decomposition [45]. The ab-
pH in the pores could be high enough (out of the stability domain of sence of that inflexion in NM400 and NMI could be expected as they
those species) to provoke Dawson anions decomposition, PMo9O934− were submitted to high-temperature calcining conditions at which ace-
species and [Al(OH)6Mo6O18]3 − Anderson-type heteropolyanions tates combustion could take place. Ni acetate decomposition under H2
being then formed. Alumina partial dissolution is possible even at pH atmosphere could occur at similar temperature as in air [45], meaning
values (between 4 and 6) at which that phenomenon could be consid- that thermal phenomena occurring under either air combustion or cat-
ered as improbable [43], due to the effect of polymolybdates acting as alyst activation by sulfiding (see Section 2.3 HDS reaction test) could
inorganic ligands. It seemed that by using our simultaneous impregna- present alike characteristics.
tion solution Al2O3 network disruption could contribute to tetrahedral Main DTG signals related to CA residua combustion were observed at
Mo6+ species in impregnated solids not submitted to high-temperature 376 and 453 °C in NMI whereas those inflexions were registered at more
annealing (see peak at 891 °C, NMIII profile in Fig. 5). The absence of severe conditions in NMII and NMIII. That pointed out to oxidic molyb-
that signal in NMII profile suggested that alumina dissolution could denum catalyzing combustion of organic residua originally located on
probably be precluded at some extent by surface “passivation” due to deposited metals surface. In this line, Al2O3-supported Mo6+ materials

Fig. 8. Linear regression fit of ln(1-x) against time plots for various tested materials. (a): NM400; (b): NMI; (c): NMII; (d) NMIII. Batch reactor, n-heptane as solvent, P = 7.2 MPa, T =
250 °C, ~107 rads−1 (1030 rpm) mixing speed.
40 J. Escobar et al. / Fuel Processing Technology 156 (2017) 33–42

Hamrin [49], no intermediate product from direct sulfur extrusion (sty-


rene) was registered under used reaction conditions probably due to its
rapid transformation to ethylbenzene (EB) [50]. According to Wang and
Prins [51] EB could be formed from BT beginning with breaking of the
aryl C\\S bond, then hydrogenation of the vinyl one, followed by alkyl
C\\S bond scission. Last two steps could occur much faster than desorp-
tion of corresponding intermediates resulting in EB kinetically behaving
as primary product although it actually requires of three distinctive
steps to be formed. Interestingly, in some cases aromatic ring hydroge-
nation of EB to ethylcyclohexane was observed after total BT conversion
over our sulfided catalysts. That was alike to biphenyl hydrogenation to
cyclohexylbenzene that coul take place right after total
dibenzothiophene (DBT) conversion [52], suggesting stronger adsorp-
Scheme 2. Benzothiophene HDS reaction pathways (from [48]). tion of BT as to that of EB over active sites of tested catalysts. EB could
adsorb on sulfur vacancies available after total benzothiophene elimina-
tion, being then hydrogenated to corresponding alkylated cycloalkane.
have been tested as catalysts for diesel soot combustion where required Benzothiophene and ethylbenzene concentration (adjusted at actual
temperature was decreased by about 100 °C, as to the case of non-cata- operating conditions) against time plots shown in Fig. 9 confirms that
lyzed oxidation [46,47]. essentially total organo-S species conversion was reached in the test du-
Finally, the high-temperature DTG signal attributed to MoO3 subli- ration for the sample of the highest activity (NMII). After hydrotreating,
mation in Fig. 7 clearly shifted to lower values in CA-modified samples. the liquid product had about 31 ppm S. Considering that FCC naphtha
The maximum for that inflexion was observed at 842 and 874 °C for comprises about 40% of the gasoline pool in refineries and that various
NMII and NM400, respectively (see Table 5), again pointing out to de- fractions composing the commercial product are essentially sulfur-free
creased interaction between molybdenum species and carrier surface (those from alkylation and reforming, for instance) the final mixed
“passivated” by CA deposition prior to Ni-Mo-P phases impregnation. stream with around 13 ppm S could fulfill stringent environmental reg-
ulations currently enforced in developed countries dictating 10–30 ppm
3.7. Hydrodesulfurization (HDS) activity test maximum S content.
Modification of calcined NM400 by CA impregnation (NMI sample)
Benzothiophene HDS activity (on pseudo first order kinetic constant was not effective in enhancing HDS activity, Table 6. That could proba-
basis, k) of sulfided catalysts prepared are shown in Table 6. k values bly be due to molybdenum content in prepared catalysts. Okamoto et
were determined from slope of corresponding –ln(1-x) against time al. [53] found that by post-treatment of calcined alumina-supported
plots shown in Fig. 8. Reaction pathways for BT transformation are CoMo catalyst (20 wt% Mo) with citric acid (at CA/Mo = 2) the activity
shown in Scheme 2 (from [48]). In agreement with Morooka and of corresponding sulfided materials in DBT HDS could be enhanced by a

Fig. 9. Concentration (at operating conditions) against time plots for BT and EB over various tested catalysts. Filled squares: benzothiophene. Open squares: ethylbenzene. Symbols:
experimental values. Lines: pseudo first order kinetics fitting. (a): NM400; (b): NMI; (c): NMII; (d) NMIII. Batch reactor, n-heptane as solvent, P = 7.2 MPa, T = 250 °C, ~107 rads−1
(1030 rpm) mixing speed.
J. Escobar et al. / Fuel Processing Technology 156 (2017) 33–42 41

factor of more than two. They proposed that significant improvements operations (P N 10 bar, as in our case) the most important role of CA
in organo-S removal activity could be related to transformation of crys- consisted in increasing the sulfiding degree of CoMo phases by weaken-
talline MoO3 and CoMoO4 into well-dispersed Mo-CA surface com- ing the interaction between deposited oxidic precursors and the Al2O3
plexes that resulted in enhanced MoS2 particles edge dispersion after carrier.
sulfidation. However, in formulations of molybdenum content close to It was demonstrated that the preparation stage where the organic
monolayer (approximately 3 Mo atoms nm−2, as in our case) CA depo- modifier was added significantly affected the HDS properties of sulfided
sition over annealed Mo phase could be even detrimental on HDS prop- alumina-supported NiMo catalysts. Others have also found that the
erties (as tested in thiophene conversion) [54]. It is worth mentioning preparation step where chelating agent was contacted with metallic
that those authors found positive effect of postreating with citric acid components was crucial in determining HDS ability of corresponding
at CA/Mo ratio much larger as to that used in the present work (2 and sulfided materials. For instance, it was recently reported [56] that
0.4, respectively). thioglycolic acid (TGA) effect on hydrotreating properties of Nb-HMS-
Although from UV data better dispersed Mo phases were found in supported NiMo strongly depended on the stage at which it was
NMI as to those in NM400 (Table 2), TPR profiles (Fig. 5 and Table 3) added (very negative when impregnated on calcined material, slightly
suggested that on the former deposited Mo phases were more difficult detrimental when simultaneously impregnated with the NiMo precur-
to reduce, as to those over the latter. Thus, oxidic phases dispersion itself sors). The nature of the impregnated phases-support interaction is
was not determining on desulfurizating properties of corresponding also crucial regarding the effect of organic modifiers on HDS activity of
sulfided formulations. corresponding sulfided catalysts. In this line, by varying the carrier
HDS activity of simultaneously impregnated NMIII was inferior to from Nb-HMS to Nb SBA-15 [56], the TGA influence over sulfided
that of NM400. Due to low pH of acidic impregnating solution used dur- NiMo phases was strongly modified (good when on calcined materials,
ing NMIII preparation CA could not be ionized precluding then Mo com- very good when simultaneously impregnated with NiMo precursors).
plexation. Li et al. [41] recently reported that citric (at CA/Ni = 1.2, Even more, in some cases the effect of organic additives could still be ev-
alumina-supported NiW materials not calcined after modifier addition) ident after their removal by calcining, as found for polyethylene glycol-
simultaneously added in aqueous solution of nickel and tungsten pre- modified CoMo/alumina materials [57].
cursors at pH alike to that in our NMIII impregnating solution did not re-
sulted in improved HDS activity (4,6-dimethyl-dibenzothiophene 4. Conclusions
conversion) of final sulfided catalyst. According to their XPS data that
could probably be attributed to decreased W and Ni sulfides dispersion NiMo/Al2O3 formulations obtained by one-pot simultaneous im-
in corresponding formulation. In the same line, Okamoto et al. [54] pregnation of Mo, Ni and P were modified by citric acid addition at Ni/
found negative effect of simultaneous CA addition in CoMo/alumina at CA = 1 mol ratio at various preparation stages to assess the influence
Mo loading alike to that used in the present work. of that parameter on HDS activity of final sulfided catalysts. Nickel ace-
The catalyst obtained by impregnating CA on the bare carrier prior to tate was used as precursor to focus the investigation on molybdenum
Ni-Mo-P phases deposition had the highest BT HDS activity (70% in- complexation. CA deposition over calcined alumina-supported NiMo re-
crease as to that of the reference calcined formulation with no organic sulted in diminished proportion of refractory tetrahedral Mo6+(t) spe-
additive), Table 6. From UV data that sample also had the highest cies but also in octahedral ones (Mo6 +(o)) reducible at higher
Mo6+ dispersion (Table 2). According to TPR profiles of Fig. 5 it seemed temperature, as to those over non-modified NiMo/Al2O3. That was
that the “passivated” alumina surface provided by CA adsorption origi- reflected in lower HDS ability. The highest BT HDS activity (70% in-
nated molybdenum species of enhanced reducibility due to their weak- crease, as to the calcined formulation with no organic additive) was
er interaction with the carrier. Even more, refractory tetrahedral Mo6+ found for the catalyst obtained by impregnating CA on bare alumina car-
entities strongly interacting with the Al2O3 substrate were absent in rier prior to Ni-Mo-P deposition. In that case, molybdenum and nickel
NMII. Thus, the enhanced HDS properties of that sample seemed to be species of decreased interaction with the support were evidenced due
consequence of those positive characteristics. In full agreement, Li et to surface “passivation” from citric adsorption/decomposition on the
al. [41] found enhanced activity of alumina-supported NiW catalysts Al2O3 support. In the opposite, simultaneous Ni-Mo-P-CA deposition
(at CA/Ni = 1.2) where part of the organic acid (equivalent to CA/ where low pH of acidic impregnating solution (pH ~ 1.4) devoided citric
Ni = 0.3) was used firstly to modify the carrier being the rest added acid ionization (then, no citrate complex formation) was ineffective in
during Ni-W-simultaneous impregnation. In that case, the CA-modified improving HDS activity.
Al2O3 was submitted to high-temperature pyrolysis resulting in carrier
surface passivation by carbonaceous residua deposition, prior to active
phases precursors impregnation. In our approach, our CA-modified Nomenclature
Al2O3 was just dried (120 °C) high-temperature annealing being ABT Integrated area of benzothiophene chromatographic peak
avoided in order to preserve the complexing ability of the adsorbed or- (a.u.)
ganic modifier. AEB Integrated area of ethylbenzene chromatographic peak (a.u.)
It has been proposed [29] that during CA impregnation over alumina CBT0 Initial BT concentration (mol l−1)
carboxylate groups and deprotonated –OH groups as well could be CBT BT concentration at given time (mol l−1)
bonded to Lewis coordinatively unsaturated sites (Al3+) on the Al2O3 CH2 H2 concentration in saturated solvent (n-heptane) (mol l−1)
surface. However, significant number of carboxylate moieties in k Pseudo first order HDS kinetic constant (s−1)
adsorbed CA molecules could remain unbound being then available to k1 Pseudo first order HDS kinetic constant (l mol−1 s−1)
be coordinated to other atoms (presumably, molybdenum in our RBT BT HDS reaction rate (mol l−1 s−1)
case). Indeed, our FTIR data (see Section 3.2 FTIR (Fourier-transform in- t Reaction time (s)
frared) spectroscopy) strongly suggested that Mo-citrate complexes x Benzothiophene conversion at time t
formation were favored in NMII.
It is worth mentioning that the amount of CA used in our work (CA/ Acknowledgements
Ni = 1) could not be enough to complexate all impregnated Mo6 +
pointing out to the possibility of optimizing the CA/Mo ratio to obtain A. Gutiérrez acknowledges CONACYT (Mexico) for a graduate stu-
further improvements in HDS activity. dent (Masters in Science) fellowship. The authors appreciate comments
In full agreement with our findings, others have very recently pro- and suggestions from the reviewers that significantly contributed in im-
posed [55] that under conditions relevant to commercial HDS proving this paper.
42 J. Escobar et al. / Fuel Processing Technology 156 (2017) 33–42

References [29] P.C. Hidbert, T.J. Graule, L.J. Gauckler, Citric acid-a dispersant for aqueous alumina
suspensions, J. Am. Ceram. Soc. 79 (1996) 1857–1867.
[1] C. Song, X. Ma, New design approaches to ultra-clean diesel fuels by deep desulfur- [30] R.-H. Zhang, X.-W. Zhou, Y.-C. Guo, M.-L. Chen, Z.-X. Cao, Y.L. Chow, Z.-H. Zhou, Crys-
ization and deep dearomatization, Appl. Catal. B 41 (2003) 207–238. talline and solution chemistry of tetrameric and dimeric molybdenum(VI) citrato
[2] S. Badoga, K. Chandra Mouli, K.K. Soni, A.K. Dalai, J. Adjaye, Beneficial influence of complexes, Inorg. Chim. Acta 406 (2013) 27–36.
EDTA on the structure and catalytic properties of sulfided NiMo/SBA-15 catalysts [31] M. Kosmulski, The pH-dependent surface charging and points of zero charge V. Up-
for hydrotreating of light gas oil, Appl. Catal. B 125 (2012) 67–84. date, J. Colloid Interface Sci. 353 (2011) 1–15.
[3] Y. Ohta, T. Shimizu, T. Honma, M. Yamada, Effect of chelating agents on HDS and ar- [32] A. Assabane, Y.A. Ichou, H. Tahiri, C. Guillard, J.-M. Herrmann, Photocatalytic degra-
omatic hydrogenation over CoMo-and NiW/Al2O3, Stud. Surf. Sci. Catal. 127 (1999) dation of polycarboxylic benzoic acids in UV-irradiated aqueous suspensions of tita-
161–168. nia. Identification of intermediates and reaction pathway of the photomineralization
[4] P. Moradi, M. Parvari, Hydrodesulfurization reaction on CoMo catalysts: effect of of trimellitic acid (1,2,4-benzene tricarboxylic acid), Appl. Catal. B 24 (2000) 71–87.
preparation method, J. Chem. Eng. Jpn 44 (2011) 643–648. [33] R.S. Weber, Effect of local structure on the UV–visible absorption edges of molybde-
[5] M. Sun, D. Nicosia, R. Prins, The effects of fluorine, phosphate and chelating agents num oxide clusters and supported molybdenum oxides, J. Catal. 151 (1995)
on hydrotreating catalysts and catalysis, Catal. Today 86 (2003) 173–189. 470–474.
[6] J. Escobar, M.C. Barrera, J.A. de los Reyes, J.A. Toledo, V. Santes, J.A. Colín, Effect of [34] P. Torres-Mancera, J. Ramírez, R. Cuevas, A. Gutiérrez-Alejandre, F. Murrieta, R. Luna,
chelating ligands on Ni-Mo impregnation over wide-pore ZrO2–TiO2, J. Mol. Catal. Hydrodesulfurization of 4,6-DMDBT on NiMo and CoMo catalysts supported on
A 287 (2008) 33–40. B2O3-Al2O3, Catal. Today 107-108 (2005) 551–558.
[7] L. Peña, D. Valencia, T. Klimova, CoMo/SBA-15 catalysts prepared with EDTA and [35] K. Hadjiivanov, M. Mihaylov, D. Klissurski, P. Stefanov, N. Abadjieva, E. Vassileva, L.
citric acid and their performance in hydrodesulfurization of dibenzothiophene, Mintchev, Characterization of Ni/SiO2 catalysts prepared by successive deposition
Appl. Catal. B 147 (2014) 879–887. and reduction of Ni2+ ions, J. Catal. 185 (1999) 314–323.
[8] M.S. Rana, J. Ramírez, A. Gutiérrez-Alejandre, J. Ancheyta, L. Cedeño, Support effects [36] J. Abart, E. Delgado, G. Ertl, H. Jeziorowski, H. Knözinger, N. Thiele, X.Z.H. Wang, E.
in CoMo hydrodesulfurization catalysts prepared with EDTA as a chelating agent, J. Taglauer, Surface structure and reduction behaviour of NiO-MoO3/Al2O3 - catalysts,
Catal. 246 (2007) 100–108. Appl. Catal. 2 (1982) 155–176.
[9] N. Rinaldi, T. Kubota, Y. Okamoto, Effect of citric acid addition on Co-Mo/B2O3/Al2O3 [37] R. López Cordero, A. López Agudo, Effect of water extraction on the surface proper-
catalysts prepared by a post-treatment method, Ind. Eng. Chem. Res. 48 (2009) ties of Mo/Al2O3 and NiMo/Al2O3 hydrotreating catalysts, Appl. Catal. A 202 (2000)
10414–10424. 23–36.
[10] J. Ryczkowski, IR studies of EDTA alkaline salts interaction with the surface of inor- [38] S.L. González-Cortés, Y. Qian, H.A. Almegren, T. Xiao, V.L. Kuznetsov, P.P. Edwards,
ganic oxides, Appl. Surf. Sci. 252 (2005) 813–822. Citric acid-assisted synthesis of g-alumina-supported high loading CoMo sulfide cat-
[11] H. Wu, A. Duan, Z. Zhao, D. Qi, J. Li, B. Liu, G. Jiang, J. Liu, Y. Wei, X. Zhang, Preparation alysts for the hydrodesulfurization (HDS) and hydrodenitrogenation (HDN) reac-
of NiMo/KIT-6 hydrodesulfurization catalysts with tunable sulfidation and disper- tions, Appl. Petrochem. Res. 5 (2015) 181–197.
sion degrees of active phase by addition of citric acid as chelating agent, Fuel 130 [39] Y. Zhang, W. Han, X. Long, H. Nie, Redispersion effects of citric acid on CoMo/γ-Al2O3
(2014) 203–210. hydrodesulfurization catalysts, Catal. Commun. 82 (2016) 20–23.
[12] P.A. Nikulshin, P.P. Minaev, A.V. Mozhaev, K.I. Maslakov, M.S. Kulikova, A.A. [40] A. Villarreal, J. Ramírez, L. Cedeño Caero, P. Castillo Villalón, A. Gutiérrez-Alejandre,
Pimerzin, Investigation of co-effect of 12-tungstophosphoric heteropolyacid, nickel Importance of the sulfidation step in the preparation of highly active NiMo/SiO2/
citrate and carbon-coated alumina in preparation of NiWcatalysts for HDS, HYD and Al2O3 hydrodesulfurization catalysts, Catal. Today 250 (2015) 60–65.
HDN reactions, Appl. Catal. B 176 (2015) 374–384. [41] H. Li, M. Li, Y. Chu, F. Liu, H. Nie, Essential role of citric acid in preparation of efficient
[13] J. Escobar, A.W. Gutiérrez, M.C. Barrera, J.A. Colín, NiMo/alumina NiW/Al2O3 HDS catalysts, Appl. Catal. A 403 (2011) 75–82.
hydrodesulfurization catalyst modified by saccharose. Effect of addition stage of or- [42] P. Blanchard, C. Lamonier, A. Griboval, E. Payen, New insight in the preparation of
ganic modifier, Can. J. Chem. Eng. 94 (2016) 66–74. alumina supported hydrotreatment oxidic precursors: a molecular approach, Appl.
[14] J. Escobar, M.C. Barrera, J.A. Toledo, M.A. Cortés-Jácome, C. Angeles-Chávez, S. Núñez, Catal. A 322 (2007) 33–45.
V. Santes, E. Gómez, L. Díaz, E. Romero, J.G. Pacheco, Effect of ethyleneglycol addition [43] X. Carrier, J.F. Lambert, M. Che, Ligand-promoted alumina dissolution in the prepa-
on the properties of P-doped NiMo/Al2O3 HDS Catalysts: part I. Materials prepara- ration of MoOX/g-Al2O3 catalysts: evidence for the formation and deposition of an
tion and characterization, Appl. Catal. B 88 (2009) 564–575. Anderson-type alumino heteropolymolybdate, J. Am. Chem. Soc. 119 (1997)
[15] M.M. Barbooti, D.A. Al-Sammerrai, Thermal decomposition of citric acid, 10137–10146.
Thermochim. Acta 98 (1986) 119–126. [44] J. Escobar, J.A. De los Reyes, T. Viveros, Influence of the Synthesis Additive on the
[16] L. Gurvich, Physico-chemical attractive force, Zh. Russ. Fiz-Khim. Obshchestva Chem. Textural and Structural Characteristics of Sol-Gel Al2O3-TiO2, Ind. Eng. Chem. Res.
47 (1915) 805–827. 39 (2000) 666–672.
[17] C. Fontaine, Y. Romero, A. Daudin, E. Devers, C. Bouchy, S. Brunet, Insight into sul- [45] J.C. De Jesus, I. Gonzalez, A. Quevedo, T. Puerta, Thermal decomposition of nickel ac-
phur compounds and promoter effects on molybdenum-based catalysts for selec- etate tetrahydrate: an integrated study by TGA, QMS and XPS techniques, J. Molec.
tive HDS of FCC gasoline, Appl. Catal. A 388 (2010) 188–195. Catal. A 228 (2005) 283–291.
[18] K.A. Nadeina, O.V. Klimov, V.Y. Pereima, G.I. Koryakina, I.G. Danilova, I.P. Prosvirin, [46] S. Braun, L.G. Appel, M. Schmal, Molybdenum species on alumina and silica supports
E.Y. Gerasimov, A.M. Yegizariyan, A.S. Noskov, Catalysts based on amorphous alumi- for soort combustion, Catal. Commun. 6 (2005) 7–12.
nosilicates for selective hydrotreating of FCC gasoline to produce Euro-5 gasoline [47] I.C. Leite, S. Braun, M. Schmal, Diesel soot combustion on Mo/Al2O3 and V/Al2O3 cat-
with minimum octane number loss, Catal. Today 271 (2016) 4–15. alysts: investigation of the active catalytic species, J. Catal. 223 (2004) 114–121.
[19] I.M. Abdulagatov, N.D. Azizov, Volumetric properties of the binary mixture n-hep- [48] I.V. Babich, J.A. Moulijn, Science and technology of novel processes for deep desul-
tane + ethylbenzene mixtures at high temperatures and high pressures, J. Thermal furization of oil refinery streams: a review, Fuel 82 (2003) 607–631.
Anal. Cal. 87 (2007) 483–492. [49] S. Morooka, C.E. Hamrin Jr., Desulfurization of model coal sulfur compounds by coal
[20] J. Dykyj, J. Svoboda, R.C. Wilhoit, M.L. Frenkel, K.R. Hall, Vapor Pressure of Chemicals: mineral matter and a cobalt molybdate catalyst-II: Benzothiophene, Chem. Eng. Sci.
Part A. Vapor Pressure and Antoine Constants for Hydrocarbons and Sulfur, Seleni- 34 (1979) 521–525.
um, Tellurium and Hydrogen Containing Organic Compounds, Springer, Berlin, [50] I.A. Van Parljs, L.H. Hosten, G.F. Froment, Kinetics of hydrodesulfurization on a
1999 373. CoMo/g- Al2O3 catalyst. 2. Kinetics of the hydrogenolysis of benzothiophene, Ind.
[21] V. Majer, V. Svoboda, Enthalpies of Vaporization of Organic Compounds: A Critical Eng. Chem. Prod. Res. Dev. 25 (1986) 437–443.
Review and Data Compilation, Blackwell Scientific Publications, Oxford, 1985 300. [51] H. Wang, R. Prins, HDS of benzothiophene and dihydrobenzothiophene over
[22] H. Knözinger, P. Ratnasamy, Catalytic aluminas: surface models and caracterization sulfided Mo/g - Al2O3, Appl. Catal. A 350 (2008) 191–196.
of surface sites, Catal. Rev. Sci. Eng. 17 (1978) 31–70. [52] E.J.M. Hensen, P. Kooyman, Y. van der Meer, A.M. van der Kraan, V.H.J. de Beer, J.A.R.
[23] M. Digne, P. Sautet, P. Raybaud, P. Euzen, H. Toulhoat, Hydroxyl groups on γ-alumi- van Veen, R.A. van Santen, The relation between morphology and hydrotreating ac-
na surfaces: a DFT Study, J. Catal. 211 (2002) 1–5. tivity for supported MoS2 particles, J. Catal. 199 (2001) 224–235.
[24] O. Yasuaki, I. Toshinobu, Interaction chemistry between molybdena and alumina: [53] N. Rinaldi, M. Yoshioka, T. Kubota, Y. Okamoto, Hydrodesulfurization activity of Co–
infrared studies of surface hydroxyl groups and adsorbed carbon dioxide on alu- Mo/Al2O3 catalysts prepared with citric acid: post-treatment of Calcined catalysts
minas modified with Molybdate, sulfate, or fluorine anions, J. Phys. Chem. 92 with high Mo loading, J. Jpn. Petrl. Inst. 53 (2010) 292–302.
(1988) 7102–7112. [54] N. Rinaldi, T. Kubota, Y. Okamoto, Effect of citric acid addition on the
[25] V.A. Suárez-Toriello, C.E. Santolalla-Vargas, J.A. de los Reyes, A. Vázquez-Zavala, M. hydrodesulfurization activity of MoO3/Al2O3 catalysts, Appl. Catal. A 374 (2010)
Vrinat, C. Geantet, Influence of the solution pH in impregnation with citric acid 228–236.
and activity of Ni/W/Al2O3 catalysts, J. Mol. Catal. (2015) 36–46 A 404–405. [55] L. van Haandel, G.M. Bremmer, E.J.M. Hensen, T. Weber, Influence of sulfiding agent
[26] R.E. Sievers, J.C. Bailar Jr., Some metal chelates of ethylenediaminetetraacetic acid, and pressure on structure and performance of CoMo/Al2O3 hydrodesulfurization
diethylenetriaminepentaacetic acid, and triethylenetetraminehexaacetic acid, catalysts, J. Catal. 342 (2016) 27–39.
Inorg. Chem. 1 (1962) 174–182. [56] R. Palcheva, L. Kaluža, L. Dimitrov, G. Tyuliev, NiMo catalysts supported on the Nb
[27] O.V. Klimov, A.V. Pashigreva, M.A. Fedotov, D.I. Kochubey, Y.A. Chesalov, G.A. modified mesoporous SBA-15and HMS: effect of thioglycolic acid addition on
Bukhtiyarova, A.S. Noskov, Co-Mo catalysts for ultra-deep HDS of diesel fuels pre- HDS, Appl. Catal. A 520 (2016) 24–34.
pared via synthesis of bimetallic surface compounds, J. Molec. Catal. A 322 (2010) [57] R. Iwamoto, N. Kagami, A. Iino, Effect of polyethylene glycol addition on
80–89. Hydrodesulfurization activity over CoO-MoO3/Al2O3 catalyst, J. Jpn. Pet. Inst. 48
[28] S. Badoga, A.K. Dalai, J. Adjaye, Y. Hu, Combined effects of EDTA and heteroatoms (Ti, (2005) 237–242.
Zr, and Al) on catalytic activity of SBA-15 supported NiMo catalyst for Hydrotreating
of heavy gas oil, Ind. Eng. Chem. Res. 53 (2014) 2137–2156.

View publication stats

You might also like