- Email: [email protected]
γ-Al2O3 washcoated monoliths
Fuel 163 (2016) 180–188
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Hydrodesulfurization of dibenzothiophene on NiMo/c-Al2O3 washcoated monoliths Rupesh Singh, Deepak Kunzru ⇑ Department of Chemical Engineering, Indian Institute of Technology Kanpur, Kanpur 208016, UP, India
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Optimum metal loading was
determined for hydrodesulfurization of dibenzothiophene on NiMo/cAl2O3 catalyst. Higher catalyst size showed pore diffusional resistances. Catalyst dispersion decreased with metal loading. NiMo/c-Al2O3 washcoated on monoliths was fully utilized with no external or internal mass transfer resistances.
a r t i c l e
i n f o
Article history: Received 16 April 2015 Received in revised form 18 August 2015 Accepted 23 September 2015 Available online 1 October 2015 Keywords: Monolith reactor Hydrodesulfurization NiMo/c-Al2O3 Dibenzothiophene
a b s t r a c t Hydrodesulfurization of dibenzothiophene, dissolved in n-hexadecane, was conducted in the temperature range of 548–593 K at 10 MPa pressure in a conventional packed bed reactor for two different catalyst sizes and on monoliths washcoated with different metal loadings. The monolith catalysts were made by washcoating using NiMo/c-Al2O3 catalysts prepared by incipient wetness and uniform welladhered washcoats were obtained by calcination after each dipping–air blowing–drying cycle. The MoO3 loading of the powder catalyst was varied from 12.0 to 24.3 wt.%, keeping the MoO3/NiO wt. ratio fixed at 6. With an increase in metal loading, the surface area, pore volume and the concentration of weak acidic sites decreased, whereas the metal particle size and the concentration of acidic sites of intermediate strength increased. The highest conversion was obtained with the catalyst containing 18.6 wt.% MoO3 and 3.21 wt.% NiO. The effectiveness factor of the larger size catalyst (average size: 1.5 mm) varied from 0.61 to 0.71 at 593 K. The rate constants (calculated on the basis of metal loading) for the monoliths were in good agreement with those obtained in a packed bed, implying that all the catalyst was utilized in the monoliths. After a 72 h run at 593 K and 10 MPa, the concentration of strong acidic sites increased but there was no change in the activity or product distribution. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction In recent years, the availability of low sulfur crude oils is diminishing and the remaining hydrocarbon reserves consist mostly of ⇑ Corresponding author. Tel.: +91 512 2597193; fax: +91 512 2590104. E-mail address: [email protected] (D. Kunzru). http://dx.doi.org/10.1016/j.fuel.2015.09.058 0016-2361/Ó 2015 Elsevier Ltd. All rights reserved.
heavy crudes that contain relatively higher concentrations of polluting sulfur, oxygen and nitrogen compounds. On the other hand, the allowable limit of sulfur in transportation fuels is being continuously reduced. Presently, in USA, the sulfur content of diesel should be below 15 ppm and the allowable limit is expected to be lowered further [1].
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
The most commonly used process for reducing the sulfur content of petroleum fractions is hydrodesulfurization (HDS). Though several catalysts have been reported, NiMo/Al2O3 and CoMo/Al2O3 are the catalysts of choice [2–4]. HDS is a multiphase reaction which requires high temperature and pressure and is usually conducted in co-current, downflow packed bed reactors. The advantages of such a reactor include high volumetric catalyst loading and the ability to handle high throughputs without flooding. However, it suffers from several disadvantages such as high pressure drop, flow maldistribution, pore diffusional limitation and partial wetting of the catalyst. Moreover, with such a flow arrangement, the H2S concentration increases along the reactor length and inhibits the removal of the more refractory sulfur compounds remaining near the reactor exit. To overcome the shortcomings of packed bed reactors, monoliths have emerged as a promising option. Monoliths are assembly of large number of parallel channels, separated from each other by a thin wall. Depending on the requirement, the channel cross-section can be circular, square or hexagonal [5]. Monoliths possess certain advantages in comparison to packed bed reactors, which includes lower pressure drop, higher mass transfer rates, negligible pore diffusional limitation, high tolerance to plugging and easier scale-up. Moreover, monoliths permit countercurrent operation without flooding problems [6,7]. Out of the different types of sulfur compounds present in diesel, dibenzothiophene and alkyl substituted dibenzothiophenes are the most difficult to desulfurize because of their bulky molecular structure and the steric hindrance associated with alkyl substituted groups [8]. To reduce the sulfur content to below the allowable limit, it is necessary to develop HDS catalysts and reactors with improved activity, selectivity, and stability. Very limited published information is available for HDS on monoliths, especially at high pressures. The kinetics of HDS of dibenzothiophene was investigated in the temperature range of 543–573 K and a hydrogen pressure of 6–8 MPa on CoMo deposited on a c-Al2O3 monolith but no comparison was made with a packed bed reactor [9]. Ismagilov et al. [10] used Pt supported on zeolite-containing (65% zeolite; 35 wt.% Ca-monlmorillonite) monoliths for the HDS of diesel fuels at a temperature of 573 K and a total pressure of 6 MPa. The crushed catalyst had a higher hydrogenation activity that was attributed to intragranular diffusional limitations on the monolith catalyst. It should be mentioned that monoliths made of c-Al2O3 and zeolites are not available commercially. Several modeling studies have shown the advantages of monolith reactors over conventional reactors for HDS and hydrogenation reactions [11–13]. In this study, the HDS of dibenzothiophene (DBT) has been investigated in the temperature range of 548–593 K and a pressure of 10 MPa on ceramic monoliths coated with Ni–Mo/c-Al2O3 and on powdered catalysts of two different sizes. Moreover, the effect of metal loading (at a constant Ni/Mo ratio) on the conversion of DBT has been evaluated for the monolith and trickle bed reactor. The catalysts were characterized using X-ray diffraction (XRD), surface area using nitrogen adsorption, transmission electron microscopy (TEM), X-ray fluorescence spectroscopy (XRF), temperature-programmed reduction (TPR), temperature programmed desorption of NH3 (NH3-TPD) and Raman spectroscopy.
2. Experimental details 2.1. Preparation of NiMo/c-Al2O3 catalysts The NiMo/c-Al2O3 catalysts of three different metal loadings were prepared by incipient wetness impregnation method. The MoO3 loading was varied from 12 to 24.3 wt.% whereas the MoO3/NiO weight ratio in all the catalysts was kept the same at
181
6. The metal precursor salts used were ammonium heptamolybdate and nickel nitrate. The catalysts were prepared by dissolving 3 g of citric acid in 14 ml of water followed by addition of the required amounts of (NH4)6Mo7O244H2O and Ni(NO3)26H2O. The contents were stirred till a clear solution was obtained. The solution thus prepared was slowly added to 20 g of Al2O3 placed in a rotary evaporator. The catalyst was dried at 353 K under vacuum till the support was free-flowing. The catalyst sample was then dried at 393 K for 1 h followed by calcination of the dried sample at 823 K for 2 h. The powder thus obtained was pelletized, crushed, sieved and the fractions between 0.25 and 0.3 mm (denoted as NiMo 1, NiMo 2 and NiMo 3) and 1.4–1.6 mm (denoted as NiMo 4, NiMo 5 and NiMo 6) retained for use. 2.2. Washcoating of monoliths Deposition of the catalyst on the monoliths can be done either by first depositing the active metal on the support and then washcoating the monolith with the slurry of the supported catalyst, or by a two-step procedure, in which the monolith is first washcoated with only the support and then the active metal is deposited on the washcoated support. Mogalicherla and Kunzru [14] found that for Pd/Al2O3 catalysts, the catalyst dispersion reduced with washcoated loading using the two step procedure and there was a possibility of nonuniform metal deposition. Therefore, for this study, the monoliths were washcoated with slurry of NiMo/c-Al2O3. NiMo/c-Al2O3 catalysts, prepared using the procedure detailed in Section 2.1 were used for washcoating the 400 cpsi (cells per square inch) cordierite monoliths (Corning, USA). The monolith pieces were approximately 30 cm in length containing 16 channels and a square cross section of 0.5 cm 0.5 cm. These pieces were machined from larger diameter monolith blocks. The preparation of monolith catalyst was done in two steps. First, the desired composition of supported NiMo/c-Al2O3 catalyst was prepared using the incipient impregnation method followed by deposition of the NiMo/c-Al2O3 on the channel walls using slurry washcoating method [15]. Initially, the as-received monolith was heated for 6 h at 773 K followed by ultra-sonication for 1 h to remove any adsorbed impurities. Then, the initial weight of monolith was measured. The average size of the NiMo/c-Al2O3 powder catalyst was reduced from 50 lm to 3 lm by wet ballmilling for 24 h in a Planetary mono mill (Pulverisette 6, Fritsch, Germany). Concentrated nitric acid was added to adjust the pH of the slurry to 3.0 during milling. The milled alumina was dried in an oven at 393 K for 10 h and the soft agglomerates stored for further experiments. The washcoating slurry was prepared by adding the milled alumina powder to water containing colloidal alumina (used as a binder) and again milled for 2 h. The monolith was then dipped vertically in this slurry and the slurry rose through the channels of the monolith due to capillary action. Dipping time was fixed at 4 min. The monolith was removed from the slurry and the excess slurry removed from the channels by blowing compressed air. Then the monolith was dried at 393 K for 4 h followed by calcination at 773 K for 6 h. The change in weight of the monolith was measured to determine the amount of washcoat deposited on the channels. For achieving higher loadings, the dipping–air blowing–drying-cal cination cycle was repeated till the desired loading was obtained. The final washcoat loading of the supported catalyst on the monolith was between 10.2 and 13.0 wt.% of bare monolith weight. The adherence of the washcoat was measured by subjecting the coated monolith to ultrasonication in acetone for 1 h at an intensity of 33 kHz. In all cases, the weight loss was less than 0.5 wt.% of the washcoat loading. Since the pH during washcoating is maintained around 3, there is a possibility of the metals leaching out in the slurry. The leaching
182
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
of metals (if any) was checked. For this, the metal contents of the powder catalyst and the monolith catalyst prepared using this powder catalyst were compared. The XRF analysis of the crushed monolith confirmed that metal loading on the c-Al2O3 was the same in both cases.
2.3. Catalytic activity test 2.3.1. Experimental set up The hydrodesulphurization of dibenzothiophene was performed in a high pressure tubular reactor with downflow of hydrogen and liquid. A schematic diagram of the experimental set up is shown in Fig. 1. Hydrogen and the liquid were preheated separately and then mixed just before the reactor inlet. The pressure of the inlet gases from the cylinders was boosted by two separate air-driven gas boosters, whereas the liquid was fed through a high pressure liquid pump (Model series III, Lab Alliance, USA). The gas– liquid mixture passed through the reactor placed in a three-zone furnace. The reactor housed an insert of square cross-section (6 mm 6 mm) and 30 cm length. This insert served as the monolith holder and was placed inside a cylindrical tube of length 50 cm with an inner diameter of 12 mm and outer diameter of 19 mm. The reactor effluent passed through a condenser that was cooled by a refrigerated circulator. From the condenser, the gas–liquid mixture was sent to a gas liquid separator via a back pressure regulator. The liquid samples were collected from the bottom of the gas–liquid separator and the noncondensables vented. The liquid sample was analyzed on a gas chromatograph equipped with Petrocol DH capillary column (100 m 0.25 mm) with a flame ionization detector.
2.3.2. Procedure Before use, the monolith holder was cleaned by acetone and dried at 393 K. After that the monolith, was inserted inside the monolith holder. The holder containing the monolith was then placed inside the outer cylindrical tube. This monolith holder was also used for the packed bed runs. In this case, after cleaning and drying, approximately 2 cm of the reactor from the top was filled with quartz wool. After that, a weighed amount of alumina supported Ni–Mo catalyst was packed in the holder. Again, some quartz wool was placed below the catalyst bed in order to support the catalyst followed by bare alumina pellets. At the exit of the holder, a small piece of 16 channel monolith was cemented by high temperature cement (Omega, USA) for giving support to the whole catalyst bed. The sulfided form of the NiMo/c-Al2O3 catalyst is known to be the active form of the hydrodesulfurization catalyst. Therefore, before evaluating the catalytic activity of the catalyst, the catalysts were sulfided in situ using 10 wt.% of dimethyl disulfide (DMDS) dissolved in toluene. The sulfiding protocol was as follows: (i) the catalyst temperature was first increased at atmospheric pressure from 298 K to 503 K in 210 min under a helium flow; (ii) when the temperature reached 503 K, the helium flow was reduced to zero and the reactor pressure was gradually increased to 1.36 MPa by feeding hydrogen at a flow rate of 400 ml/min. The system was kept at this condition for 2 h, and then the sulfiding feed was started. Under the combined flow of liquid and hydrogen, the reactor pressure was further increased to 5.5 MPa. The catalyst was kept at this temperature for 255 min; (iii) the temperature was then increased from 503 K to 616 K in 3 h and the catalyst kept at this temperature for 2 h. After completion of sulfiding, H2 flow was stopped and He flow at a rate of 300 ml/min was started. Then,
Fig. 1. Schematic diagram of the experimental set-up: (1) air compressor, (2) H2 gas cylinder, (3) He gas cylinder, (4) air driven gas booster, (5) valve, (6) gas storage tank, (7) pressure gauge, (8) mass flow controller, (9) gas preheater, (10) liquid preheater, (11) check valve, (12) high pressure liquid pump, (13) liquid storage tank, (14) furnace, (15) reactor, (16) thermocouple, (17) insulation, (18) condenser, (19) back pressure regulator, (20) gas liquid separator.
183
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
the sulfiding feed was stopped. It is important that the catalyst should not be in contact with H2 without the presence of any liquid otherwise, hydrogen can reduce the sulfided catalyst. The system was kept at this condition for 6 h. Then the temperature was slowly increased to the desired reaction temperature at a ramp rate of 2 K/min. The HDS activity of the catalyst was investigated at 10 MPa pressure in the temperature range of 548–593 K. The liquid feed was 1.6 wt.% DBT in n-hexadecane. For all the runs, the hydrogen to liquid ratio was 400 Nm3/m3 and W/FA0 was 91.0 kgcat.h/kmol of DBT for the packed bed runs and varied from 105.9 to 136.2 kgcat.h/kmol of DBT for the monoliths.
samples were dispersed in n-hexane. 5 ll from the sample was loaded on a carbon-coated copper grid which was dried under vacuum. The Raman spectra of the fresh and spent catalysts were recorded using a Raman spectrometer (Sentera, Bruker Optik GmbH). An Ar-ion laser having a wavelength of 532 nm was used for excitation. 2 mW of laser power was used to obtain the Raman spectra. The uniformity and thickness of the washcoat was examined on a scanning electron microscope (Tescan Mira3).
2.4. Catalyst characterization
3.1. Catalyst characterization
The compositions of the prepared NiMo/c-Al2O3 samples were determined using the X-ray fluorescence technique (XRF). The analysis was done on a Rigaku make ZSX Primus II XRF spectrometer. For XRF analysis, the calcined samples were pelletized using a hydraulic press at a pressure of 15–20 tons. For the analysis of Al, the sample was excited using a current 100 mA and a voltage of 30 kV. For Ni and Mo, these values were fixed at 60 mA and 50 kV. X-ray Diffraction (XRD) was performed on a PAN analytical make X-ray Diffractometer using Ni filtered Ka radiation from a Cu target (k = 1.541841 Å). The sample was scanned between the angles 10–80° at a scan rate of 3°/min. The surface area of each NiMo/c-Al2O3 catalyst was determined based on the amount of N2 that was adsorbed and desorbed at 77 K, using Autosorb-iQ (Quantachrome, USA). The TPR analysis of the prepared samples was performed to determine the reducibility of the material, using the Quantachrome instrument, with the help of ASiQwin software. TPR was performed in a U-shaped quartz tube using 5% (v/v) H2 in N2. Prior to the analysis, the test sample was degassed in helium flow at 473 K for 2 h followed by introduction of 5% H2 in N2 at 10 ml/ min for 30 min at 313 K. The sample temperature was then continuously raised from 313 K to 1273 K at 15 K/min and the amount of hydrogen consumed measured by TCD. NH3 TPD analysis was performed for the measurement of acidic properties of the catalysts. NH3 TPD of the prepared samples was carried out on a QuantaChrome instrument, with the help of ASiQwin software. The analysis was performed in a U shaped quartz tube using 20% (v/v) NH3 in He. For the analysis, the sample (150 mg) was heated with a flow of 20% (v/v) NH3 in He at the flow rate of 30 ml/min, at heating rate of 10 K/min up to 573 K. Sample was kept at this temperature for 75 min, followed by cooling of the sample to room temperature using the gas mixture at same flow rate. TPD was run from room temperature to 1273 K at a heating rate of 10 K/min with He flow rate of 30 ml/min. The amount of NH3 desorbed was measured using a TCD. The TEM analysis of the unsupported and c-Al2O3 supported NiMo samples was carried out on FEI make TecnaiTM G2 U-Twin (200 kV) Transmission Electron Microscope. For TEM analysis, the
The MoO3 and NiO content (in wt.%) of the prepared samples were determined using XRF and the results are shown in Table 1. The balance amount was Al2O3. The XRF results confirm that the MoO3/NiO weight ratio was approximately 6 for all the samples. The surface area and pore volume of the catalysts are also shown in Table 1. In comparison to the unsupported alumina (surface area = 200 m2/g; pore volume = 0.65 cm3/g), the surface area and pore volume of the supported catalyst decreased significantly with an increase in metal loading, most probably due to the partial blockage of the pores of the support due to deposition of metal. Fig. 2 shows the XRD analysis of NiMo/c-Al2O3 samples. The two prominent broad diffraction peaks that appeared at 2h = 46° and 67° correspond to c-Al2O3 which was used as the support. For NiMo 1, there was a small diffraction peak at 26.64° corresponding to the (0 1 1) plane of orthorhombic MoO3 phase [JCPDS 47-1081]. As the metal loading increased, the diffraction peaks of MoO3 became more prominent. NiMo 2 and NiMo 3 showed three diffraction peaks at 2h = 26.79° and 29.12° and 30.87° corresponding to (0 1 1), (2 0 0) and (1 1 1) planes of MoO3 orthorhombic phase. No separate diffraction peak for Ni or any of its oxide was observed during XRD analysis which confirms that NiO was well dispersed over the Al2O3 support. The TEM images of the NiMo/c-Al2O3 catalysts are shown in Fig. 3. With an increase in metal loading, there was not only an increase in the metal particle size but the polydispersity of the metal particles also increased. For NiMo 1, the size was in the range of 2–5 nm; for NiMo 2 in the size range of 2–6 nm whereas for NiMo 3 the metal particle size range was from 3 to 9 nm. The TPR results for all the NiMo/c-Al2O3 catalysts are shown in Fig. 4. The TPR results show that there was no change in the reduction behavior of the catalyst with an increase in metal loading. The reduction profile of the NiMo/c-Al2O3 catalyst shows hydrogen consumption in a broad temperature interval between 673 and 1273 K, with three main reduction peaks observed in the temperature ranges of 760–761 K, 831–839 K and 1125–1169 K. The low-temperature peak (760 K) can be attributed to the reduction of Mo6+ to Mo4+. The shoulder observed at the intermediate temperature (831–839 K) could be assigned to the reduction of Ni2+
3. Result and discussion
Table 1 Composition, surface area and pore volume of NiMo/c-Al2O3 catalysts. Sample code NiMo NiMo NiMo NiMo NiMo NiMo a b
1 4 2 5 3 6
(avg. (avg. (avg. (avg. (avg. (avg.
MoO3 size: size: size: size: size: size:
0.275 mm) 1.5 mm) 0.275 mm) 1.5 mm) 0.275 mm) 1.5 mm)
a
wt.%
NiO
a
wt.%
Surface area
b
(m2/g)
Pore volume (cm3/g)
12.0
2.0
164.0
0.50
18.6
3.2
140.6
0.45
24.3
4.0
109.1
0.41
Measured by XRF. Determined by N2 adsorption–desorption at 77 K.
184
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
The SEM images of the cross-section of bare monolith and washcoated monolith with a catalyst loading of 13 wt.% are shown in Fig. 6. The rounding of the corners for the coated monolith can be clearly seen. This has also been reported by others [18] and occurs because the corners are affected more by the viscous forces as compared to the region near the center. The maximum and minimum thickness of the washcoat were measured in several channels, and the average minimum and maximum thickness were 21 lm and 78 lm, respectively. 3.2. Catalyst activity in packed beds
Fig. 2. X-ray diffraction spectra of NiMo/c-Al2O3 catalysts.
species on the alumina surface. The high temperature peak can be ascribed to the reduction of Mo4+ to Mo0 [16]. The ammonia TPD profiles of the fresh catalysts are shown in Fig. 5. As can be seen from this figure, the desorption of NH3 occurred over three broad temperature ranges of 323–573 K, 673–973 K and 973–1173 K, corresponding to weak, intermediate and strong acid sites. With an increase in metal content, the number of weak acid sites decreased whereas the number of intermediate acid sites increased. For the NH3-TPD of NiMo/Al2O3, Ferdous et al. [17] also reported that with increasing content of Ni or Mo in the catalyst, the number of intermediate and strong acid sites increased.
The catalyst activity for the HDS of DBT was tested for two different sizes and three different metal loadings at a constant Ni/Mo ratio for the same W/FA0. The contact time was intentionally kept low so that the differences in the conversion of the catalysts would be more noticeable. The W/FA0 used corresponds to a WHSV of 126.3 kg feed/kgcat.h. The smaller catalyst size (average diameter: 0.275 mm) was chosen such that pore diffusional resistance did not affect the rate. Preliminary runs were conducted at an average catalyst size of 0.225 mm and, at identical conditions, the conversion for the 0.225 mm size catalyst was the same as for the 0.275 mm size catalyst, confirming the absence of diffusion limitations for the 0.275 mm size catalysts. The larger size (average diameter: 1.5 mm) was chosen such that the characteristic length (volume of catalyst particle/external surface area) of 250 lm was similar to the characteristic length of 1/1600 (1.6 mm) extrudates that are used commercially. The effect of MoO3 loading for the two sizes at 548 and 593 K is shown in Fig. 7. For both the sizes, conversion of DBT passed through a maximum with an increase in MoO3 loading. The highest conversion of 60.8% at 593 K was obtained at a MoO3 loading of 18.6 wt.%. At higher MoO3 loading,
Fig. 3. TEM analysis of NiMo/c-Al2O3 (a) NiMo 1, (b) NiMo 2 and (c) NiMo 3.
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
Fig. 4. TPR profiles of NiMo/c-Al2O3 cataysts.
conversion of DBT decreased most probably due to the lower metal dispersion in the NiMo 3 catalyst. Thus, even though the metal loading of NiMo 3 is nearly twice that of NiMo 1, the overall conversion is only marginally higher. The activities of the catalysts were compared by assuming the HDS of DBT to be a pseudo-first-order reaction. Others [19,20] have also assumed this reaction to be pseudo first-order. The rate equation for first order kinetics for a plug flow reactor can be expressed as follows:
ln
1 1 XA
¼ kapp
W Q
185
size, it can be seen that the effectiveness factors at 548 K varied between 0.68 and 0.76 and were slightly lower (0.61–0.71) at 593 K. Comparing the rate constants calculated based on the amount of active metal, it can be seen that rate constants decreased with an increase in metal loading due to the decrease in the metal dispersion with increasing metal content. As can be seen from Table 2, the rate constant for NiMo 3 was only 59% of the rate constant for NiMo 1. Therefore, although the total metal loading of NiMo 3 was significantly higher than NiMo 1, the corresponding increase in conversion was only marginal. The main products obtained during HDS were biphenyl (BP) and cyclohexyl benzene (CHB). It has been proposed that hydrodesulfurization of DBT follows two pathways: (i) direct desulfurization of DBT in which the hydrogenolysis of sulfur bond takes place leading to the formation of biphenyl (BP); and (ii) desulfurization after hydrogenation of one of the aromatic rings followed by hydrogenolysis of the sulfur bond, leading to formation of cyclohexylbenzene (CHB) [21,22]. The variation of the product distribution with temperature is shown in Fig. 8. For all catalyst samples, the major product was BP. With an increase in temperature there was a shift in product selectivity toward CHB. The CHB/BP selectivity ratio varied from 0.36 ± 0.02 at 548 K to 0.45 ± 0.05 at 593 K. It is difficult to correlate the activity of the catalysts with any single physico-chemical characteristics of the catalyst. The activity will depend not only on the surface area of the catalyst but also on the metal dispersion and the distribution of the acidic sites. There
ð1Þ
where kapp is the apparent first order rate constant, Q is the inlet volumetric flow rate of liquid, W is the mass of the catalyst and XA is the conversion of DBT. The rate constants were calculated on the basis of the total mass of catalyst as well as on the basis of the metal present, and the results are shown in Table 2. For a first order reaction, the effectiveness factor is defined as
g ¼ kapp =kintrinsic
ð2Þ
where kintrinsic is the rate constant measured in the absence of diffusional limitations. For this study, kintrinsic corresponds to the rate constants determined for the catalysts of 0.275 mm average diameter. Comparing the rate constants (based on total mass of catalyst) of the smaller size catalyst with the corresponding catalyst of larger
Fig. 5. NH3-TPD profile of fresh NiMo/Al2O3 and spent NiMo 2 catalysts. (a) NiMo 1, (b) NiMo 2, (c) NiMo 3 and (d) spent NiMo 2.
Fig. 6. SEM images of (a) bare monolith and (b) monolith with 13.0 wt.% washcoat loading.
186
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
dispersion (higher metal particle size) and lower surface area. Thus, it seems that the activity is affected by surface area, catalyst acidity and metal dispersion.
3.3. Catalyst activity in monoliths Monoliths with three different metal loadings were prepared by respectively using NiMo 1, NiMo 2 and NiMo 3 milled powders. Thus, a direct comparison between the monolith results and the powdered catalysts could be made. Due to the difficulty in controlling the washcoat amount deposited in each dipping, the catalyst loading was not identical in the three monoliths and varied between 10.2 and 13.0 wt.% of bare monolith weight. The liquid velocity in the monolith channels was kept fixed at 0.1 cm/s, and therefore the W/FA0 was not the same for the three monoliths. The W/FA0 for Monoliths 1, 2 and 3 was 136.2, 107.6 and 105.9 kgcat.h/kmol DBT, respectively. At 548 K, the conversions obtained with Monoliths 1, 2 and 3 were 31.1%, 37.0% and 33.2%, respectively; whereas at 593 K, the conversions were 52.1%, 66.8% and 58.7%, respectively. The rate constants obtained with the three monoliths are tabulated in Table 2. Comparing the rate constants obtained with Monoliths 1, 2 and 3 with those obtained on the corresponding 0.275 mm size catalysts (NiMo 1, NiMo 2 and NiMo 3, respectively), it can be seen that the rate constants for Monoliths 2 and 3 at 548 and 593 K are nearly the same as those obtained on the powdered catalysts confirming that the catalyst is fully utilized in the monoliths and there are no diffusional resistances or maldistribution of the liquid in the monolith channels. However, it should be noted that for all the monoliths, the rate constants are
Fig. 7. Effect of MoO3 loading on conversion of DBT in the packed bed (W/FA0 = 91.04 kgcat.h/kmol DBT, P = 10 MPa, MoO3/NiO = 6 wt/wt) (a) 548 K, (b) 593 K.
was no effect of MoO3 loading on the reducibility of the catalyst (refer Fig. 4). Even though the metal dispersion and surface area of the NiMo 2 was lower than NiMo 1, its activity was higher, most probably due to the higher number of acidic sites of intermediate strength. On the other hand, NiMo 3 catalyst having the highest number of acidic sites of intermediate strength had an activity lower than NiMo 2. This was most likely due to the lower metal
Table 2 Pseudo-first order constants for HDS of DBT on monoliths and powdered catalysts. Catalyst
NiMo 1 NiMo 4 g = kNiMo 4/kNiMo NiMo 2 NiMo 5 g = kNiMo 5/kNiMo NiMo 3 NiMo 6 g = kNiMo 6/kNiMo Monolith 1 Monolith 2 Monolith 3
1
2
3
Rate constant m3/kgcat.s 105 Temperature
Rate constant m3/kg (NiO + MoO3).s 105 Temperature
548 K
593 K
548 K
593 K
1.27 0.97 0.76 1.64 1.11 0.68 1.51 1.02 0.68 1.09 1.71 1.52
2.69 1.90 0.71 4.11 2.51 0.61 3.38 2.22 0.66 2.16 4.09 3.34
9.07 6.93
19.25 13.56
7.53 5.09
18.85 11.53
5.33 3.60
11.94 7.85
7.80 7.87 5.38
15.4 18.76 11.80
Fig. 8. Selectivity to biphenyl for NiMo catalysts (a) 548 K and (b) 593 K.
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
higher than the rates obtained with catalysts of average size corresponding to the catalyst size used commercially. The rate constants at 548 and 593 K for Monolith 1 were somewhat lower than the corresponding values for NiMo 1. A possible reason for this could be some non uniformity in the washcoat along the length of the coated monolith. As shown in Fig. 8, the selectivity to biphenyl was similar to that obtained in a packed bed. Generally, for relatively slow reactions such as hydrodesulfurization of dibenzothiophene and the hydrogenation of acetylene, the measured rate of reaction is not affected by changes in either the superficial liquid or gas velocities [23,24] that affect the external mass transfer coefficients. This was checked for Monolith 1. The conversion of DBT at a hydrogen/liquid ratio of 800 Nm3/m3 and 548 K was 29.0% as compared to 31.1% obtained at a hydrogen/liquid ratio of 400 Nm3/m3. Similarly, there was no significant effect of doubling the liquid velocity on the rate constant. This confirms that at these conditions the external mass transfer coefficients did not affect the rate of reaction. Due to the relatively low gas and liquid flow rates used in this study, the pressure drops for both the packed bed and monolith reactors were so small that these were not measurable by the pressure gauges used at the inlet and the outlet. However, as has been reported by several investigators [25,26] the pressure drop in monoliths is one to two orders of magnitude lower in comparison to trickle bed reactors. It should be mentioned that at these conditions the volume productivity of monoliths was lower than that of the packed bed reactor. The difference between the monolith and a conventional trickle bed reactor would be reduced for monoliths with a higher washcoat loading or a more active catalyst or higher temperature
187
Fig. 10. Raman spectra for fresh and spent NiMo 2.
where the catalyst effectiveness factors would be still lower in a packed bed. We have deposited monoliths with a loading of 50 wt.% with well-adhered washcoats and these are being further tested. 3.4. Characterization of spent catalyst To check the changes (if any) on the conversion, product distribution and catalyst characteristics, a longer duration (72 h) run was conducted using NiMo 2 catalyst at 593 K and 10 MPa. The variation in the conversion and product distribution with run time is shown in Fig. 9. There was no noticeable change in either the conversion or the product distribution. The spent catalyst was analyzed by NH3-TPD, Raman spectroscopy and surface area measurement. The surface area of the spent catalyst was 122 m2/g in comparison to 140.6 m2/g (refer Table 1) for the fresh catalyst. Comparing the NH3-TPD profile of the fresh and spent catalysts (Fig. 5), it can be clearly seen that the number of weak acid sites significantly decreased whereas the strong acid sites increased. This is in agreement with the reported results [27]. The Raman spectrum of fresh and spent catalysts is shown in Fig. 10. Two broad peaks at 1350 and 1590 cm1, corresponding to amorphous carbon, were observed for the spent catalyst, whereas the fresh catalyst did not show any peak. These analysis show that, during the run time of 72 h, although there was no effect on the activity and selectivity, but there were changes in the physico-chemical characteristics of the catalyst. 4. Conclusions
Fig. 9. Effect of runtime on (a) conversion of DBT and (b) selectivity of biphenyl (Catalyst: NiMo 2; W/FA0 = 182.08 (kgcat.h/kmol of DBT).
For the hydrodesulphurization of dibenzothiophene on NiMo/cAl2O3 catalysts with a MoO3/NiO weight ratio of 6, the most active catalysts are obtained with a MoO3 loading of 18.6 wt.%. Higher metal loading has an adverse effect on the conversion, due to the lower metal dispersion. The rate constants (calculated on the basis of metal loading) decrease with metal loading. In a packed bed reactor, the effectiveness factors of catalysts of 1.5 mm average diameter are in the range of 0.67–0.77 at 548 K and from 0.61 to 0.70 at 593 K. The rate constants for the different catalysts obtained on washcoated monoliths are in good agreement with those obtained with powdered catalysts of 0.275 mm average size confirming that the active metal was fully utilized in the monoliths. The concentration of strong acidic sites increased on the spent catalyst but there was no effect of this on the activity or product distribution in a 72 h run.
188
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
Acknowledgement The financial support provided by Chevron Corporation, USA for this project is gratefully acknowledged.
References
[13]
[14] [15] [16]
[1] Kim JH, Ma X, Song C, Lee Y-K, Oyama ST. Kinetics of two pathways for 4,6dimethyldibenzothiophene hydrodesulfurization over NiMo, CoMo sulfide, and nickel phosphide catalysts. Energy Fuels 2005;19:353–64. [2] Li X, Chai Y, Liu B, Liu H, Li J, Zhao R, et al. Hydrodesulfurization of 4,6dimethyldibenzothiophene over CoMo catalysts supported on c-alumina with different morphology. Ind Eng Chem Res 2014;53:9665–73. [3] Liu F, Xu S, Cao L, Chi Y, Zhang T, Xue D. A comparison of NiMo/Al2O3 catalysts prepared by impregnation and coprecipitation methods for hydrodesulfurization of dibenzothiophene. J Phys Chem C 2007;111: 7396–402. [4] Ferdous D, Dalai AK, Adjaye J. Hydrodenitrogenation and hydrodesulfurization of heavy gas oil using NiMo/Al2O3 catalyst containing boron: experimental and kinetic studies. Ind Eng Chem Res 2006;45:544–52. [5] Roy S, Bauer T, Al-Dahhan M, Lehner P, Turek T. Monoliths as multiphase reactors: a review. AlChE J 2004;50:2918–38. [6] Kapteijn F, Heiszwolf JJ, Nijhuis TV, Moulijn JA. Monoliths in multiphase catalytic processes: aspects and prospects. Cattech 1999;3:24–41. [7] Lebens PJM, Stork MM, Kapteijn F, Sie ST, Moulijn JA. Hydrodynamics and mass transfer issues in a countercurrent gas–liquid internally finned monolith reactor. Chem Eng Sci 1999;54:2381–9. [8] Kallinikos LE, Jess A, Papayannakos NG. Kinetic study and H2S effect on refractory DBTs desulfurization in a heavy gasoil. J Catal 2010;269:169–78. [9] Edvinsson R, Irandoust S. Hydrodesulfurization of dibenzothiophene in a monolithic catalyst reactor. Ind Eng Chem Res 1993;32:391–5. [10] Ismagilov ZR, Yashnik SA, Startsev AN, Boronin AI, Stadnichenko AI, Kriventsov VV, et al. Deep desulphurization of diesel fuels on bifunctional monolithic nanostructured Pt-zeolite catalysts. Catal Today 2009;144:235–50. [11] Iliuta I, Iliuta MC. Sulfur removal in monolithic three-phase reactors: model and simulation. Sep Purif Technol 2012;92:64–76. [12] van Hasselt BW, Lebens PJM, Calis HPA, Kapteijn F, Sie ST, Moulijn JA, et al. A numerical comparison of alternative three-phase reactors with a conventional
[17]
[18] [19]
[20]
[21]
[22]
[23] [24] [25] [26] [27]
trickle-bed reactor. The advantages of countercurrent flow for hydrodesulfurization. Chem Eng Sci 1999;54:4791–9. Kapteijn F, Nijhuis TA, Heiszwolf JJ, Moulijn JA. New non-traditional multiphase catalytic reactors based on monolithic structures. Catal Today 2001;66:133–44. Mogalicherla AK, Kunzru D. Effect of method of preparation on activity of Pd/ Al2O3 monolith catalysts. Can J Chem Eng 2010;88:367–75. Nijhuis TA, Beers AEW, Vergunst T, Hoek I, Kapteijn F, Moulijn JA. Preparation of monolithic catalysts. Cat Rev – Sci Eng 2001;43:345–80. Badoga S, Sharma RV, Dalai AK, Adjaye J. Hydrotreating of heavy gas oil on mesoporous zirconia supported NiMo catalyst with EDTA. Fuel 2014;128:30–8. Ferdous D, Dalai AK, Adjaye J. A series of NiMo/Al2O3 catalysts containing boron and phosphorus: Part I. Synthesis and characterization. Appl Catal, A 2004;260:137–51. Kolb WB, Papadimitriou AA, Cerro RL, Leavitt DD, Summers JC. The ins and outs of coating monolithic structures. Chem Eng Prog 1993:61–7. Laredo GC, Altamirano En, De los Reyes JA. Inhibition effects of nitrogen compounds on the hydrodesulfurization of dibenzothiophene: Part 2. Appl Catal, A 2003;243:207–14. Wang Y, Sun Z, Wang A, Ruan L, Lu M, Ren J, et al. Kinetics of hydrodesulfurization of dibenzothiophene catalyzed by sulfided CoMo/ MCM-41. Ind Eng Chem Res 2004;43:2324–9. Lamure-Meille V, Schulz E, Lemaire M, Vrinat M. Effect of experimental parameters on the relative reactivity of dibenzothiophene and 4methyldibenzothiophene. Appl Catal, A 1995;131:143–57. Houallia M, Broaderick DH, Sapre AV, Naga NK, de Beer VJH. BC gates hydrodesulfurization of methyl-substituted dibenzothiophenes catalyzed by sulfided Co-Mo/c-Al2O3. J Catal 1980;61:523–7. Irandoust S, Gahne O. Competitive hydrodesulfurization and hydrogenation in a monolithic reactor. AlChE J 1990;36:746–52. Edvinsson RK, Holmgren AM, Irandoust S. Liquid-phase hydrogenation of acetylene in a monolithic catalyst reactor. Ind Eng Chem Res 1995;34:94–100. Cybulski A, Moulijn JA. Monoliths in heterogeneous catalysis. Catal Rev Sci Eng 1994;36:179–270. Stankiewicz A. Process intensification in in-line monolithic reactor. Chem Eng Sci 2001;56:359–64. Ferdous D, Dalai AK, Adjaye J. A series of NiMo/Al2O3 catalysts containing boron and phosphorus: Part II. Hydrodenitrogenation and hydrodesulfurization using heavy gas oil derived from Athabasca bitumen. Appl Catal, A 2004;260:153–62.
Contents lists available at ScienceDirect
Fuel journal homepage: www.elsevier.com/locate/fuel
Hydrodesulfurization of dibenzothiophene on NiMo/c-Al2O3 washcoated monoliths Rupesh Singh, Deepak Kunzru ⇑ Department of Chemical Engineering, Indian Institute of Technology Kanpur, Kanpur 208016, UP, India
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Optimum metal loading was
determined for hydrodesulfurization of dibenzothiophene on NiMo/cAl2O3 catalyst. Higher catalyst size showed pore diffusional resistances. Catalyst dispersion decreased with metal loading. NiMo/c-Al2O3 washcoated on monoliths was fully utilized with no external or internal mass transfer resistances.
a r t i c l e
i n f o
Article history: Received 16 April 2015 Received in revised form 18 August 2015 Accepted 23 September 2015 Available online 1 October 2015 Keywords: Monolith reactor Hydrodesulfurization NiMo/c-Al2O3 Dibenzothiophene
a b s t r a c t Hydrodesulfurization of dibenzothiophene, dissolved in n-hexadecane, was conducted in the temperature range of 548–593 K at 10 MPa pressure in a conventional packed bed reactor for two different catalyst sizes and on monoliths washcoated with different metal loadings. The monolith catalysts were made by washcoating using NiMo/c-Al2O3 catalysts prepared by incipient wetness and uniform welladhered washcoats were obtained by calcination after each dipping–air blowing–drying cycle. The MoO3 loading of the powder catalyst was varied from 12.0 to 24.3 wt.%, keeping the MoO3/NiO wt. ratio fixed at 6. With an increase in metal loading, the surface area, pore volume and the concentration of weak acidic sites decreased, whereas the metal particle size and the concentration of acidic sites of intermediate strength increased. The highest conversion was obtained with the catalyst containing 18.6 wt.% MoO3 and 3.21 wt.% NiO. The effectiveness factor of the larger size catalyst (average size: 1.5 mm) varied from 0.61 to 0.71 at 593 K. The rate constants (calculated on the basis of metal loading) for the monoliths were in good agreement with those obtained in a packed bed, implying that all the catalyst was utilized in the monoliths. After a 72 h run at 593 K and 10 MPa, the concentration of strong acidic sites increased but there was no change in the activity or product distribution. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction In recent years, the availability of low sulfur crude oils is diminishing and the remaining hydrocarbon reserves consist mostly of ⇑ Corresponding author. Tel.: +91 512 2597193; fax: +91 512 2590104. E-mail address: [email protected] (D. Kunzru). http://dx.doi.org/10.1016/j.fuel.2015.09.058 0016-2361/Ó 2015 Elsevier Ltd. All rights reserved.
heavy crudes that contain relatively higher concentrations of polluting sulfur, oxygen and nitrogen compounds. On the other hand, the allowable limit of sulfur in transportation fuels is being continuously reduced. Presently, in USA, the sulfur content of diesel should be below 15 ppm and the allowable limit is expected to be lowered further [1].
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
The most commonly used process for reducing the sulfur content of petroleum fractions is hydrodesulfurization (HDS). Though several catalysts have been reported, NiMo/Al2O3 and CoMo/Al2O3 are the catalysts of choice [2–4]. HDS is a multiphase reaction which requires high temperature and pressure and is usually conducted in co-current, downflow packed bed reactors. The advantages of such a reactor include high volumetric catalyst loading and the ability to handle high throughputs without flooding. However, it suffers from several disadvantages such as high pressure drop, flow maldistribution, pore diffusional limitation and partial wetting of the catalyst. Moreover, with such a flow arrangement, the H2S concentration increases along the reactor length and inhibits the removal of the more refractory sulfur compounds remaining near the reactor exit. To overcome the shortcomings of packed bed reactors, monoliths have emerged as a promising option. Monoliths are assembly of large number of parallel channels, separated from each other by a thin wall. Depending on the requirement, the channel cross-section can be circular, square or hexagonal [5]. Monoliths possess certain advantages in comparison to packed bed reactors, which includes lower pressure drop, higher mass transfer rates, negligible pore diffusional limitation, high tolerance to plugging and easier scale-up. Moreover, monoliths permit countercurrent operation without flooding problems [6,7]. Out of the different types of sulfur compounds present in diesel, dibenzothiophene and alkyl substituted dibenzothiophenes are the most difficult to desulfurize because of their bulky molecular structure and the steric hindrance associated with alkyl substituted groups [8]. To reduce the sulfur content to below the allowable limit, it is necessary to develop HDS catalysts and reactors with improved activity, selectivity, and stability. Very limited published information is available for HDS on monoliths, especially at high pressures. The kinetics of HDS of dibenzothiophene was investigated in the temperature range of 543–573 K and a hydrogen pressure of 6–8 MPa on CoMo deposited on a c-Al2O3 monolith but no comparison was made with a packed bed reactor [9]. Ismagilov et al. [10] used Pt supported on zeolite-containing (65% zeolite; 35 wt.% Ca-monlmorillonite) monoliths for the HDS of diesel fuels at a temperature of 573 K and a total pressure of 6 MPa. The crushed catalyst had a higher hydrogenation activity that was attributed to intragranular diffusional limitations on the monolith catalyst. It should be mentioned that monoliths made of c-Al2O3 and zeolites are not available commercially. Several modeling studies have shown the advantages of monolith reactors over conventional reactors for HDS and hydrogenation reactions [11–13]. In this study, the HDS of dibenzothiophene (DBT) has been investigated in the temperature range of 548–593 K and a pressure of 10 MPa on ceramic monoliths coated with Ni–Mo/c-Al2O3 and on powdered catalysts of two different sizes. Moreover, the effect of metal loading (at a constant Ni/Mo ratio) on the conversion of DBT has been evaluated for the monolith and trickle bed reactor. The catalysts were characterized using X-ray diffraction (XRD), surface area using nitrogen adsorption, transmission electron microscopy (TEM), X-ray fluorescence spectroscopy (XRF), temperature-programmed reduction (TPR), temperature programmed desorption of NH3 (NH3-TPD) and Raman spectroscopy.
2. Experimental details 2.1. Preparation of NiMo/c-Al2O3 catalysts The NiMo/c-Al2O3 catalysts of three different metal loadings were prepared by incipient wetness impregnation method. The MoO3 loading was varied from 12 to 24.3 wt.% whereas the MoO3/NiO weight ratio in all the catalysts was kept the same at
181
6. The metal precursor salts used were ammonium heptamolybdate and nickel nitrate. The catalysts were prepared by dissolving 3 g of citric acid in 14 ml of water followed by addition of the required amounts of (NH4)6Mo7O244H2O and Ni(NO3)26H2O. The contents were stirred till a clear solution was obtained. The solution thus prepared was slowly added to 20 g of Al2O3 placed in a rotary evaporator. The catalyst was dried at 353 K under vacuum till the support was free-flowing. The catalyst sample was then dried at 393 K for 1 h followed by calcination of the dried sample at 823 K for 2 h. The powder thus obtained was pelletized, crushed, sieved and the fractions between 0.25 and 0.3 mm (denoted as NiMo 1, NiMo 2 and NiMo 3) and 1.4–1.6 mm (denoted as NiMo 4, NiMo 5 and NiMo 6) retained for use. 2.2. Washcoating of monoliths Deposition of the catalyst on the monoliths can be done either by first depositing the active metal on the support and then washcoating the monolith with the slurry of the supported catalyst, or by a two-step procedure, in which the monolith is first washcoated with only the support and then the active metal is deposited on the washcoated support. Mogalicherla and Kunzru [14] found that for Pd/Al2O3 catalysts, the catalyst dispersion reduced with washcoated loading using the two step procedure and there was a possibility of nonuniform metal deposition. Therefore, for this study, the monoliths were washcoated with slurry of NiMo/c-Al2O3. NiMo/c-Al2O3 catalysts, prepared using the procedure detailed in Section 2.1 were used for washcoating the 400 cpsi (cells per square inch) cordierite monoliths (Corning, USA). The monolith pieces were approximately 30 cm in length containing 16 channels and a square cross section of 0.5 cm 0.5 cm. These pieces were machined from larger diameter monolith blocks. The preparation of monolith catalyst was done in two steps. First, the desired composition of supported NiMo/c-Al2O3 catalyst was prepared using the incipient impregnation method followed by deposition of the NiMo/c-Al2O3 on the channel walls using slurry washcoating method [15]. Initially, the as-received monolith was heated for 6 h at 773 K followed by ultra-sonication for 1 h to remove any adsorbed impurities. Then, the initial weight of monolith was measured. The average size of the NiMo/c-Al2O3 powder catalyst was reduced from 50 lm to 3 lm by wet ballmilling for 24 h in a Planetary mono mill (Pulverisette 6, Fritsch, Germany). Concentrated nitric acid was added to adjust the pH of the slurry to 3.0 during milling. The milled alumina was dried in an oven at 393 K for 10 h and the soft agglomerates stored for further experiments. The washcoating slurry was prepared by adding the milled alumina powder to water containing colloidal alumina (used as a binder) and again milled for 2 h. The monolith was then dipped vertically in this slurry and the slurry rose through the channels of the monolith due to capillary action. Dipping time was fixed at 4 min. The monolith was removed from the slurry and the excess slurry removed from the channels by blowing compressed air. Then the monolith was dried at 393 K for 4 h followed by calcination at 773 K for 6 h. The change in weight of the monolith was measured to determine the amount of washcoat deposited on the channels. For achieving higher loadings, the dipping–air blowing–drying-cal cination cycle was repeated till the desired loading was obtained. The final washcoat loading of the supported catalyst on the monolith was between 10.2 and 13.0 wt.% of bare monolith weight. The adherence of the washcoat was measured by subjecting the coated monolith to ultrasonication in acetone for 1 h at an intensity of 33 kHz. In all cases, the weight loss was less than 0.5 wt.% of the washcoat loading. Since the pH during washcoating is maintained around 3, there is a possibility of the metals leaching out in the slurry. The leaching
182
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
of metals (if any) was checked. For this, the metal contents of the powder catalyst and the monolith catalyst prepared using this powder catalyst were compared. The XRF analysis of the crushed monolith confirmed that metal loading on the c-Al2O3 was the same in both cases.
2.3. Catalytic activity test 2.3.1. Experimental set up The hydrodesulphurization of dibenzothiophene was performed in a high pressure tubular reactor with downflow of hydrogen and liquid. A schematic diagram of the experimental set up is shown in Fig. 1. Hydrogen and the liquid were preheated separately and then mixed just before the reactor inlet. The pressure of the inlet gases from the cylinders was boosted by two separate air-driven gas boosters, whereas the liquid was fed through a high pressure liquid pump (Model series III, Lab Alliance, USA). The gas– liquid mixture passed through the reactor placed in a three-zone furnace. The reactor housed an insert of square cross-section (6 mm 6 mm) and 30 cm length. This insert served as the monolith holder and was placed inside a cylindrical tube of length 50 cm with an inner diameter of 12 mm and outer diameter of 19 mm. The reactor effluent passed through a condenser that was cooled by a refrigerated circulator. From the condenser, the gas–liquid mixture was sent to a gas liquid separator via a back pressure regulator. The liquid samples were collected from the bottom of the gas–liquid separator and the noncondensables vented. The liquid sample was analyzed on a gas chromatograph equipped with Petrocol DH capillary column (100 m 0.25 mm) with a flame ionization detector.
2.3.2. Procedure Before use, the monolith holder was cleaned by acetone and dried at 393 K. After that the monolith, was inserted inside the monolith holder. The holder containing the monolith was then placed inside the outer cylindrical tube. This monolith holder was also used for the packed bed runs. In this case, after cleaning and drying, approximately 2 cm of the reactor from the top was filled with quartz wool. After that, a weighed amount of alumina supported Ni–Mo catalyst was packed in the holder. Again, some quartz wool was placed below the catalyst bed in order to support the catalyst followed by bare alumina pellets. At the exit of the holder, a small piece of 16 channel monolith was cemented by high temperature cement (Omega, USA) for giving support to the whole catalyst bed. The sulfided form of the NiMo/c-Al2O3 catalyst is known to be the active form of the hydrodesulfurization catalyst. Therefore, before evaluating the catalytic activity of the catalyst, the catalysts were sulfided in situ using 10 wt.% of dimethyl disulfide (DMDS) dissolved in toluene. The sulfiding protocol was as follows: (i) the catalyst temperature was first increased at atmospheric pressure from 298 K to 503 K in 210 min under a helium flow; (ii) when the temperature reached 503 K, the helium flow was reduced to zero and the reactor pressure was gradually increased to 1.36 MPa by feeding hydrogen at a flow rate of 400 ml/min. The system was kept at this condition for 2 h, and then the sulfiding feed was started. Under the combined flow of liquid and hydrogen, the reactor pressure was further increased to 5.5 MPa. The catalyst was kept at this temperature for 255 min; (iii) the temperature was then increased from 503 K to 616 K in 3 h and the catalyst kept at this temperature for 2 h. After completion of sulfiding, H2 flow was stopped and He flow at a rate of 300 ml/min was started. Then,
Fig. 1. Schematic diagram of the experimental set-up: (1) air compressor, (2) H2 gas cylinder, (3) He gas cylinder, (4) air driven gas booster, (5) valve, (6) gas storage tank, (7) pressure gauge, (8) mass flow controller, (9) gas preheater, (10) liquid preheater, (11) check valve, (12) high pressure liquid pump, (13) liquid storage tank, (14) furnace, (15) reactor, (16) thermocouple, (17) insulation, (18) condenser, (19) back pressure regulator, (20) gas liquid separator.
183
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
the sulfiding feed was stopped. It is important that the catalyst should not be in contact with H2 without the presence of any liquid otherwise, hydrogen can reduce the sulfided catalyst. The system was kept at this condition for 6 h. Then the temperature was slowly increased to the desired reaction temperature at a ramp rate of 2 K/min. The HDS activity of the catalyst was investigated at 10 MPa pressure in the temperature range of 548–593 K. The liquid feed was 1.6 wt.% DBT in n-hexadecane. For all the runs, the hydrogen to liquid ratio was 400 Nm3/m3 and W/FA0 was 91.0 kgcat.h/kmol of DBT for the packed bed runs and varied from 105.9 to 136.2 kgcat.h/kmol of DBT for the monoliths.
samples were dispersed in n-hexane. 5 ll from the sample was loaded on a carbon-coated copper grid which was dried under vacuum. The Raman spectra of the fresh and spent catalysts were recorded using a Raman spectrometer (Sentera, Bruker Optik GmbH). An Ar-ion laser having a wavelength of 532 nm was used for excitation. 2 mW of laser power was used to obtain the Raman spectra. The uniformity and thickness of the washcoat was examined on a scanning electron microscope (Tescan Mira3).
2.4. Catalyst characterization
3.1. Catalyst characterization
The compositions of the prepared NiMo/c-Al2O3 samples were determined using the X-ray fluorescence technique (XRF). The analysis was done on a Rigaku make ZSX Primus II XRF spectrometer. For XRF analysis, the calcined samples were pelletized using a hydraulic press at a pressure of 15–20 tons. For the analysis of Al, the sample was excited using a current 100 mA and a voltage of 30 kV. For Ni and Mo, these values were fixed at 60 mA and 50 kV. X-ray Diffraction (XRD) was performed on a PAN analytical make X-ray Diffractometer using Ni filtered Ka radiation from a Cu target (k = 1.541841 Å). The sample was scanned between the angles 10–80° at a scan rate of 3°/min. The surface area of each NiMo/c-Al2O3 catalyst was determined based on the amount of N2 that was adsorbed and desorbed at 77 K, using Autosorb-iQ (Quantachrome, USA). The TPR analysis of the prepared samples was performed to determine the reducibility of the material, using the Quantachrome instrument, with the help of ASiQwin software. TPR was performed in a U-shaped quartz tube using 5% (v/v) H2 in N2. Prior to the analysis, the test sample was degassed in helium flow at 473 K for 2 h followed by introduction of 5% H2 in N2 at 10 ml/ min for 30 min at 313 K. The sample temperature was then continuously raised from 313 K to 1273 K at 15 K/min and the amount of hydrogen consumed measured by TCD. NH3 TPD analysis was performed for the measurement of acidic properties of the catalysts. NH3 TPD of the prepared samples was carried out on a QuantaChrome instrument, with the help of ASiQwin software. The analysis was performed in a U shaped quartz tube using 20% (v/v) NH3 in He. For the analysis, the sample (150 mg) was heated with a flow of 20% (v/v) NH3 in He at the flow rate of 30 ml/min, at heating rate of 10 K/min up to 573 K. Sample was kept at this temperature for 75 min, followed by cooling of the sample to room temperature using the gas mixture at same flow rate. TPD was run from room temperature to 1273 K at a heating rate of 10 K/min with He flow rate of 30 ml/min. The amount of NH3 desorbed was measured using a TCD. The TEM analysis of the unsupported and c-Al2O3 supported NiMo samples was carried out on FEI make TecnaiTM G2 U-Twin (200 kV) Transmission Electron Microscope. For TEM analysis, the
The MoO3 and NiO content (in wt.%) of the prepared samples were determined using XRF and the results are shown in Table 1. The balance amount was Al2O3. The XRF results confirm that the MoO3/NiO weight ratio was approximately 6 for all the samples. The surface area and pore volume of the catalysts are also shown in Table 1. In comparison to the unsupported alumina (surface area = 200 m2/g; pore volume = 0.65 cm3/g), the surface area and pore volume of the supported catalyst decreased significantly with an increase in metal loading, most probably due to the partial blockage of the pores of the support due to deposition of metal. Fig. 2 shows the XRD analysis of NiMo/c-Al2O3 samples. The two prominent broad diffraction peaks that appeared at 2h = 46° and 67° correspond to c-Al2O3 which was used as the support. For NiMo 1, there was a small diffraction peak at 26.64° corresponding to the (0 1 1) plane of orthorhombic MoO3 phase [JCPDS 47-1081]. As the metal loading increased, the diffraction peaks of MoO3 became more prominent. NiMo 2 and NiMo 3 showed three diffraction peaks at 2h = 26.79° and 29.12° and 30.87° corresponding to (0 1 1), (2 0 0) and (1 1 1) planes of MoO3 orthorhombic phase. No separate diffraction peak for Ni or any of its oxide was observed during XRD analysis which confirms that NiO was well dispersed over the Al2O3 support. The TEM images of the NiMo/c-Al2O3 catalysts are shown in Fig. 3. With an increase in metal loading, there was not only an increase in the metal particle size but the polydispersity of the metal particles also increased. For NiMo 1, the size was in the range of 2–5 nm; for NiMo 2 in the size range of 2–6 nm whereas for NiMo 3 the metal particle size range was from 3 to 9 nm. The TPR results for all the NiMo/c-Al2O3 catalysts are shown in Fig. 4. The TPR results show that there was no change in the reduction behavior of the catalyst with an increase in metal loading. The reduction profile of the NiMo/c-Al2O3 catalyst shows hydrogen consumption in a broad temperature interval between 673 and 1273 K, with three main reduction peaks observed in the temperature ranges of 760–761 K, 831–839 K and 1125–1169 K. The low-temperature peak (760 K) can be attributed to the reduction of Mo6+ to Mo4+. The shoulder observed at the intermediate temperature (831–839 K) could be assigned to the reduction of Ni2+
3. Result and discussion
Table 1 Composition, surface area and pore volume of NiMo/c-Al2O3 catalysts. Sample code NiMo NiMo NiMo NiMo NiMo NiMo a b
1 4 2 5 3 6
(avg. (avg. (avg. (avg. (avg. (avg.
MoO3 size: size: size: size: size: size:
0.275 mm) 1.5 mm) 0.275 mm) 1.5 mm) 0.275 mm) 1.5 mm)
a
wt.%
NiO
a
wt.%
Surface area
b
(m2/g)
Pore volume (cm3/g)
12.0
2.0
164.0
0.50
18.6
3.2
140.6
0.45
24.3
4.0
109.1
0.41
Measured by XRF. Determined by N2 adsorption–desorption at 77 K.
184
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
The SEM images of the cross-section of bare monolith and washcoated monolith with a catalyst loading of 13 wt.% are shown in Fig. 6. The rounding of the corners for the coated monolith can be clearly seen. This has also been reported by others [18] and occurs because the corners are affected more by the viscous forces as compared to the region near the center. The maximum and minimum thickness of the washcoat were measured in several channels, and the average minimum and maximum thickness were 21 lm and 78 lm, respectively. 3.2. Catalyst activity in packed beds
Fig. 2. X-ray diffraction spectra of NiMo/c-Al2O3 catalysts.
species on the alumina surface. The high temperature peak can be ascribed to the reduction of Mo4+ to Mo0 [16]. The ammonia TPD profiles of the fresh catalysts are shown in Fig. 5. As can be seen from this figure, the desorption of NH3 occurred over three broad temperature ranges of 323–573 K, 673–973 K and 973–1173 K, corresponding to weak, intermediate and strong acid sites. With an increase in metal content, the number of weak acid sites decreased whereas the number of intermediate acid sites increased. For the NH3-TPD of NiMo/Al2O3, Ferdous et al. [17] also reported that with increasing content of Ni or Mo in the catalyst, the number of intermediate and strong acid sites increased.
The catalyst activity for the HDS of DBT was tested for two different sizes and three different metal loadings at a constant Ni/Mo ratio for the same W/FA0. The contact time was intentionally kept low so that the differences in the conversion of the catalysts would be more noticeable. The W/FA0 used corresponds to a WHSV of 126.3 kg feed/kgcat.h. The smaller catalyst size (average diameter: 0.275 mm) was chosen such that pore diffusional resistance did not affect the rate. Preliminary runs were conducted at an average catalyst size of 0.225 mm and, at identical conditions, the conversion for the 0.225 mm size catalyst was the same as for the 0.275 mm size catalyst, confirming the absence of diffusion limitations for the 0.275 mm size catalysts. The larger size (average diameter: 1.5 mm) was chosen such that the characteristic length (volume of catalyst particle/external surface area) of 250 lm was similar to the characteristic length of 1/1600 (1.6 mm) extrudates that are used commercially. The effect of MoO3 loading for the two sizes at 548 and 593 K is shown in Fig. 7. For both the sizes, conversion of DBT passed through a maximum with an increase in MoO3 loading. The highest conversion of 60.8% at 593 K was obtained at a MoO3 loading of 18.6 wt.%. At higher MoO3 loading,
Fig. 3. TEM analysis of NiMo/c-Al2O3 (a) NiMo 1, (b) NiMo 2 and (c) NiMo 3.
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
Fig. 4. TPR profiles of NiMo/c-Al2O3 cataysts.
conversion of DBT decreased most probably due to the lower metal dispersion in the NiMo 3 catalyst. Thus, even though the metal loading of NiMo 3 is nearly twice that of NiMo 1, the overall conversion is only marginally higher. The activities of the catalysts were compared by assuming the HDS of DBT to be a pseudo-first-order reaction. Others [19,20] have also assumed this reaction to be pseudo first-order. The rate equation for first order kinetics for a plug flow reactor can be expressed as follows:
ln
1 1 XA
¼ kapp
W Q
185
size, it can be seen that the effectiveness factors at 548 K varied between 0.68 and 0.76 and were slightly lower (0.61–0.71) at 593 K. Comparing the rate constants calculated based on the amount of active metal, it can be seen that rate constants decreased with an increase in metal loading due to the decrease in the metal dispersion with increasing metal content. As can be seen from Table 2, the rate constant for NiMo 3 was only 59% of the rate constant for NiMo 1. Therefore, although the total metal loading of NiMo 3 was significantly higher than NiMo 1, the corresponding increase in conversion was only marginal. The main products obtained during HDS were biphenyl (BP) and cyclohexyl benzene (CHB). It has been proposed that hydrodesulfurization of DBT follows two pathways: (i) direct desulfurization of DBT in which the hydrogenolysis of sulfur bond takes place leading to the formation of biphenyl (BP); and (ii) desulfurization after hydrogenation of one of the aromatic rings followed by hydrogenolysis of the sulfur bond, leading to formation of cyclohexylbenzene (CHB) [21,22]. The variation of the product distribution with temperature is shown in Fig. 8. For all catalyst samples, the major product was BP. With an increase in temperature there was a shift in product selectivity toward CHB. The CHB/BP selectivity ratio varied from 0.36 ± 0.02 at 548 K to 0.45 ± 0.05 at 593 K. It is difficult to correlate the activity of the catalysts with any single physico-chemical characteristics of the catalyst. The activity will depend not only on the surface area of the catalyst but also on the metal dispersion and the distribution of the acidic sites. There
ð1Þ
where kapp is the apparent first order rate constant, Q is the inlet volumetric flow rate of liquid, W is the mass of the catalyst and XA is the conversion of DBT. The rate constants were calculated on the basis of the total mass of catalyst as well as on the basis of the metal present, and the results are shown in Table 2. For a first order reaction, the effectiveness factor is defined as
g ¼ kapp =kintrinsic
ð2Þ
where kintrinsic is the rate constant measured in the absence of diffusional limitations. For this study, kintrinsic corresponds to the rate constants determined for the catalysts of 0.275 mm average diameter. Comparing the rate constants (based on total mass of catalyst) of the smaller size catalyst with the corresponding catalyst of larger
Fig. 5. NH3-TPD profile of fresh NiMo/Al2O3 and spent NiMo 2 catalysts. (a) NiMo 1, (b) NiMo 2, (c) NiMo 3 and (d) spent NiMo 2.
Fig. 6. SEM images of (a) bare monolith and (b) monolith with 13.0 wt.% washcoat loading.
186
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
dispersion (higher metal particle size) and lower surface area. Thus, it seems that the activity is affected by surface area, catalyst acidity and metal dispersion.
3.3. Catalyst activity in monoliths Monoliths with three different metal loadings were prepared by respectively using NiMo 1, NiMo 2 and NiMo 3 milled powders. Thus, a direct comparison between the monolith results and the powdered catalysts could be made. Due to the difficulty in controlling the washcoat amount deposited in each dipping, the catalyst loading was not identical in the three monoliths and varied between 10.2 and 13.0 wt.% of bare monolith weight. The liquid velocity in the monolith channels was kept fixed at 0.1 cm/s, and therefore the W/FA0 was not the same for the three monoliths. The W/FA0 for Monoliths 1, 2 and 3 was 136.2, 107.6 and 105.9 kgcat.h/kmol DBT, respectively. At 548 K, the conversions obtained with Monoliths 1, 2 and 3 were 31.1%, 37.0% and 33.2%, respectively; whereas at 593 K, the conversions were 52.1%, 66.8% and 58.7%, respectively. The rate constants obtained with the three monoliths are tabulated in Table 2. Comparing the rate constants obtained with Monoliths 1, 2 and 3 with those obtained on the corresponding 0.275 mm size catalysts (NiMo 1, NiMo 2 and NiMo 3, respectively), it can be seen that the rate constants for Monoliths 2 and 3 at 548 and 593 K are nearly the same as those obtained on the powdered catalysts confirming that the catalyst is fully utilized in the monoliths and there are no diffusional resistances or maldistribution of the liquid in the monolith channels. However, it should be noted that for all the monoliths, the rate constants are
Fig. 7. Effect of MoO3 loading on conversion of DBT in the packed bed (W/FA0 = 91.04 kgcat.h/kmol DBT, P = 10 MPa, MoO3/NiO = 6 wt/wt) (a) 548 K, (b) 593 K.
was no effect of MoO3 loading on the reducibility of the catalyst (refer Fig. 4). Even though the metal dispersion and surface area of the NiMo 2 was lower than NiMo 1, its activity was higher, most probably due to the higher number of acidic sites of intermediate strength. On the other hand, NiMo 3 catalyst having the highest number of acidic sites of intermediate strength had an activity lower than NiMo 2. This was most likely due to the lower metal
Table 2 Pseudo-first order constants for HDS of DBT on monoliths and powdered catalysts. Catalyst
NiMo 1 NiMo 4 g = kNiMo 4/kNiMo NiMo 2 NiMo 5 g = kNiMo 5/kNiMo NiMo 3 NiMo 6 g = kNiMo 6/kNiMo Monolith 1 Monolith 2 Monolith 3
1
2
3
Rate constant m3/kgcat.s 105 Temperature
Rate constant m3/kg (NiO + MoO3).s 105 Temperature
548 K
593 K
548 K
593 K
1.27 0.97 0.76 1.64 1.11 0.68 1.51 1.02 0.68 1.09 1.71 1.52
2.69 1.90 0.71 4.11 2.51 0.61 3.38 2.22 0.66 2.16 4.09 3.34
9.07 6.93
19.25 13.56
7.53 5.09
18.85 11.53
5.33 3.60
11.94 7.85
7.80 7.87 5.38
15.4 18.76 11.80
Fig. 8. Selectivity to biphenyl for NiMo catalysts (a) 548 K and (b) 593 K.
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
higher than the rates obtained with catalysts of average size corresponding to the catalyst size used commercially. The rate constants at 548 and 593 K for Monolith 1 were somewhat lower than the corresponding values for NiMo 1. A possible reason for this could be some non uniformity in the washcoat along the length of the coated monolith. As shown in Fig. 8, the selectivity to biphenyl was similar to that obtained in a packed bed. Generally, for relatively slow reactions such as hydrodesulfurization of dibenzothiophene and the hydrogenation of acetylene, the measured rate of reaction is not affected by changes in either the superficial liquid or gas velocities [23,24] that affect the external mass transfer coefficients. This was checked for Monolith 1. The conversion of DBT at a hydrogen/liquid ratio of 800 Nm3/m3 and 548 K was 29.0% as compared to 31.1% obtained at a hydrogen/liquid ratio of 400 Nm3/m3. Similarly, there was no significant effect of doubling the liquid velocity on the rate constant. This confirms that at these conditions the external mass transfer coefficients did not affect the rate of reaction. Due to the relatively low gas and liquid flow rates used in this study, the pressure drops for both the packed bed and monolith reactors were so small that these were not measurable by the pressure gauges used at the inlet and the outlet. However, as has been reported by several investigators [25,26] the pressure drop in monoliths is one to two orders of magnitude lower in comparison to trickle bed reactors. It should be mentioned that at these conditions the volume productivity of monoliths was lower than that of the packed bed reactor. The difference between the monolith and a conventional trickle bed reactor would be reduced for monoliths with a higher washcoat loading or a more active catalyst or higher temperature
187
Fig. 10. Raman spectra for fresh and spent NiMo 2.
where the catalyst effectiveness factors would be still lower in a packed bed. We have deposited monoliths with a loading of 50 wt.% with well-adhered washcoats and these are being further tested. 3.4. Characterization of spent catalyst To check the changes (if any) on the conversion, product distribution and catalyst characteristics, a longer duration (72 h) run was conducted using NiMo 2 catalyst at 593 K and 10 MPa. The variation in the conversion and product distribution with run time is shown in Fig. 9. There was no noticeable change in either the conversion or the product distribution. The spent catalyst was analyzed by NH3-TPD, Raman spectroscopy and surface area measurement. The surface area of the spent catalyst was 122 m2/g in comparison to 140.6 m2/g (refer Table 1) for the fresh catalyst. Comparing the NH3-TPD profile of the fresh and spent catalysts (Fig. 5), it can be clearly seen that the number of weak acid sites significantly decreased whereas the strong acid sites increased. This is in agreement with the reported results [27]. The Raman spectrum of fresh and spent catalysts is shown in Fig. 10. Two broad peaks at 1350 and 1590 cm1, corresponding to amorphous carbon, were observed for the spent catalyst, whereas the fresh catalyst did not show any peak. These analysis show that, during the run time of 72 h, although there was no effect on the activity and selectivity, but there were changes in the physico-chemical characteristics of the catalyst. 4. Conclusions
Fig. 9. Effect of runtime on (a) conversion of DBT and (b) selectivity of biphenyl (Catalyst: NiMo 2; W/FA0 = 182.08 (kgcat.h/kmol of DBT).
For the hydrodesulphurization of dibenzothiophene on NiMo/cAl2O3 catalysts with a MoO3/NiO weight ratio of 6, the most active catalysts are obtained with a MoO3 loading of 18.6 wt.%. Higher metal loading has an adverse effect on the conversion, due to the lower metal dispersion. The rate constants (calculated on the basis of metal loading) decrease with metal loading. In a packed bed reactor, the effectiveness factors of catalysts of 1.5 mm average diameter are in the range of 0.67–0.77 at 548 K and from 0.61 to 0.70 at 593 K. The rate constants for the different catalysts obtained on washcoated monoliths are in good agreement with those obtained with powdered catalysts of 0.275 mm average size confirming that the active metal was fully utilized in the monoliths. The concentration of strong acidic sites increased on the spent catalyst but there was no effect of this on the activity or product distribution in a 72 h run.
188
R. Singh, D. Kunzru / Fuel 163 (2016) 180–188
Acknowledgement The financial support provided by Chevron Corporation, USA for this project is gratefully acknowledged.
References
[13]
[14] [15] [16]
[1] Kim JH, Ma X, Song C, Lee Y-K, Oyama ST. Kinetics of two pathways for 4,6dimethyldibenzothiophene hydrodesulfurization over NiMo, CoMo sulfide, and nickel phosphide catalysts. Energy Fuels 2005;19:353–64. [2] Li X, Chai Y, Liu B, Liu H, Li J, Zhao R, et al. Hydrodesulfurization of 4,6dimethyldibenzothiophene over CoMo catalysts supported on c-alumina with different morphology. Ind Eng Chem Res 2014;53:9665–73. [3] Liu F, Xu S, Cao L, Chi Y, Zhang T, Xue D. A comparison of NiMo/Al2O3 catalysts prepared by impregnation and coprecipitation methods for hydrodesulfurization of dibenzothiophene. J Phys Chem C 2007;111: 7396–402. [4] Ferdous D, Dalai AK, Adjaye J. Hydrodenitrogenation and hydrodesulfurization of heavy gas oil using NiMo/Al2O3 catalyst containing boron: experimental and kinetic studies. Ind Eng Chem Res 2006;45:544–52. [5] Roy S, Bauer T, Al-Dahhan M, Lehner P, Turek T. Monoliths as multiphase reactors: a review. AlChE J 2004;50:2918–38. [6] Kapteijn F, Heiszwolf JJ, Nijhuis TV, Moulijn JA. Monoliths in multiphase catalytic processes: aspects and prospects. Cattech 1999;3:24–41. [7] Lebens PJM, Stork MM, Kapteijn F, Sie ST, Moulijn JA. Hydrodynamics and mass transfer issues in a countercurrent gas–liquid internally finned monolith reactor. Chem Eng Sci 1999;54:2381–9. [8] Kallinikos LE, Jess A, Papayannakos NG. Kinetic study and H2S effect on refractory DBTs desulfurization in a heavy gasoil. J Catal 2010;269:169–78. [9] Edvinsson R, Irandoust S. Hydrodesulfurization of dibenzothiophene in a monolithic catalyst reactor. Ind Eng Chem Res 1993;32:391–5. [10] Ismagilov ZR, Yashnik SA, Startsev AN, Boronin AI, Stadnichenko AI, Kriventsov VV, et al. Deep desulphurization of diesel fuels on bifunctional monolithic nanostructured Pt-zeolite catalysts. Catal Today 2009;144:235–50. [11] Iliuta I, Iliuta MC. Sulfur removal in monolithic three-phase reactors: model and simulation. Sep Purif Technol 2012;92:64–76. [12] van Hasselt BW, Lebens PJM, Calis HPA, Kapteijn F, Sie ST, Moulijn JA, et al. A numerical comparison of alternative three-phase reactors with a conventional
[17]
[18] [19]
[20]
[21]
[22]
[23] [24] [25] [26] [27]
trickle-bed reactor. The advantages of countercurrent flow for hydrodesulfurization. Chem Eng Sci 1999;54:4791–9. Kapteijn F, Nijhuis TA, Heiszwolf JJ, Moulijn JA. New non-traditional multiphase catalytic reactors based on monolithic structures. Catal Today 2001;66:133–44. Mogalicherla AK, Kunzru D. Effect of method of preparation on activity of Pd/ Al2O3 monolith catalysts. Can J Chem Eng 2010;88:367–75. Nijhuis TA, Beers AEW, Vergunst T, Hoek I, Kapteijn F, Moulijn JA. Preparation of monolithic catalysts. Cat Rev – Sci Eng 2001;43:345–80. Badoga S, Sharma RV, Dalai AK, Adjaye J. Hydrotreating of heavy gas oil on mesoporous zirconia supported NiMo catalyst with EDTA. Fuel 2014;128:30–8. Ferdous D, Dalai AK, Adjaye J. A series of NiMo/Al2O3 catalysts containing boron and phosphorus: Part I. Synthesis and characterization. Appl Catal, A 2004;260:137–51. Kolb WB, Papadimitriou AA, Cerro RL, Leavitt DD, Summers JC. The ins and outs of coating monolithic structures. Chem Eng Prog 1993:61–7. Laredo GC, Altamirano En, De los Reyes JA. Inhibition effects of nitrogen compounds on the hydrodesulfurization of dibenzothiophene: Part 2. Appl Catal, A 2003;243:207–14. Wang Y, Sun Z, Wang A, Ruan L, Lu M, Ren J, et al. Kinetics of hydrodesulfurization of dibenzothiophene catalyzed by sulfided CoMo/ MCM-41. Ind Eng Chem Res 2004;43:2324–9. Lamure-Meille V, Schulz E, Lemaire M, Vrinat M. Effect of experimental parameters on the relative reactivity of dibenzothiophene and 4methyldibenzothiophene. Appl Catal, A 1995;131:143–57. Houallia M, Broaderick DH, Sapre AV, Naga NK, de Beer VJH. BC gates hydrodesulfurization of methyl-substituted dibenzothiophenes catalyzed by sulfided Co-Mo/c-Al2O3. J Catal 1980;61:523–7. Irandoust S, Gahne O. Competitive hydrodesulfurization and hydrogenation in a monolithic reactor. AlChE J 1990;36:746–52. Edvinsson RK, Holmgren AM, Irandoust S. Liquid-phase hydrogenation of acetylene in a monolithic catalyst reactor. Ind Eng Chem Res 1995;34:94–100. Cybulski A, Moulijn JA. Monoliths in heterogeneous catalysis. Catal Rev Sci Eng 1994;36:179–270. Stankiewicz A. Process intensification in in-line monolithic reactor. Chem Eng Sci 2001;56:359–64. Ferdous D, Dalai AK, Adjaye J. A series of NiMo/Al2O3 catalysts containing boron and phosphorus: Part II. Hydrodenitrogenation and hydrodesulfurization using heavy gas oil derived from Athabasca bitumen. Appl Catal, A 2004;260:153–62.