- Email: [email protected]
Evaluation of MoS2 based catalysts for the conversion of syngas into alcohols: A combinatorial approach
Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
Contents lists available at ScienceDirect
Journal of Environmental Chemical Engineering journal homepage: www.elsevier.com/locate/jece
Evaluation of MoS2 based catalysts for the conversion of syngas into alcohols: A combinatorial approach Arthur José Gerbasi da Silva a,b, *, Paula Claassens-Dekker c , Antônio Carlos Sallarès de Mattos Carvalho b , Antônio Manzolillo Sanseverino b , Cristina Pontes Bittencourt Quitete b , Alexandre Szklo a , Eduardo Falabella Sousa-Aguiar b,d a Energy Planning Program, Graduate School of Engineering, UFRJ – CT, Bloco C, Sala 211, Cidade Universitária, Ilha do Fundão, Rio de Janeiro, RJ 21941-972, Brazil b Petrobras Research Center – Avenida Horácio Macedo, 950, Cidade Universitária, Ilha do Fundão, Rio de Janeiro, RJ 21941-915, Brazil c Avantium Chemicals B.V., Zekeringstraat 29, 1014 BV, Amsterdam, The Netherlands d Department of Organic Processes, School of Chemistry, UFRJ – CT, Bloco E, Cidade Universitária, Ilha do Fundão, Rio de Janeiro, RJ 21941-909, Brazil
A R T I C L E I N F O
A B S T R A C T
Article history: Received 10 April 2014 Received in revised form 14 August 2014 Accepted 3 September 2014
72 MoS2 catalysts were tested in the conversion of syngas to alcohols, using a high-throughput catalyst evaluation unit, to identify the best catalyst, based on CO conversion, both ethanol and higher alcohols and total alcohols selectivity. Catalysts prepared by thermal decomposition of (NH4)2MoS4 at low temperature showed a higher selectivity to total alcohols. The highest selectivity to ethanol and higher alcohols was obtained at 300 C by a catalyst prepared by reacting Mo(CO)6 with sulphur. Catalysts prepared by thermal decomposition of (NH4)2MoS4 at high temperature showed very low activity. Catalysts prepared by thermal decomposition of (NH4)2MoS4 in tridecane/water with hydrogen atmosphere showed low activity and selectivity. There was no significant difference among the alkaline metal promoters K, Cs and Rb regarding total alcohols selectivities. Incorporation of Co and Ni led to catalysts with activity levels equivalent to catalysts that contain Rh. ã 2014 Elsevier Ltd. All rights reserved.
Keywords: Syngas conversion Molybdenum sulphide catalyst Ethanol Higher alcohols
Introduction Increasing awareness about the consequences of global warming and energy dependence has renewed the interest in the development of alternative fuels [1–4]. One option is the production of ethanol from biomass residues through the gasification of biomass and the conversion of the resulting synthesis gas into ethanol and higher alcohols [5–15]. This solution is particularly suited for Brazil, given its growing flex-fuelled (gasoline/ethanol) light vehicle fleet and great availability of sugar cane residues [16,17]. One of the biggest challenges to develop this option is to find a catalyst capable of economically converting syngas into ethanol and higher alcohols. There are various types of catalysts that can be used in this process, but one of the most promising is based on alkali promoted molybdenum sulphide
[18–20]. Various studies have examined the performance of alkali promoted molybdenum sulphide based catalysts, prepared by different methods and used under different operating conditions, for the conversion of syngas into ethanol and higher alcohols [21–34]. The thermodynamics of ethanol and higher alcohols formation have been assessed through theoretical calculations with Aspen software [35], and also by calculations with HSC Chemistry Software [18]. Methanol formation is favoured at low temperature and high pressure. At high pressure the formation of higher alcohols increases with increasing temperature [18]. Ethanol formation from syngas is a highly exothermic reaction (Eq. (1)), and heat dissipation may be a problem for scale-up [36]. 2CO(g) + 4H2(g) ! C2H5OH + H2O (1)
DH 298 = 253.6 kJ/mol of ethanol
* Corresponding author at: Petrobras Research Center – Avenida Horácio Macedo, 950, Cidade Universitária, Ilha do Fundão, Rio de Janeiro, RJ 21941-915, Brazil. Tel.: +55 21 2162 7168; fax: +55 21 2162 1007. E-mail address: [email protected] (A.J. Gerbasi da Silva). http://dx.doi.org/10.1016/j.jece.2014.09.006 2213-3437/ ã 2014 Elsevier Ltd. All rights reserved.
DG 298 = 221.1 kJ/mol of ethanol
(1)
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
Alkali promoted molybdenum sulphide catalysts produce linear alcohols and ethanol and higher alcohols are formed via a classical insertion of CO into the corresponding precursor alcohol [21]. CO hydrogenation to methane is a competitive reaction to the alcohol synthesis with such catalysts. The optimum conditions for this reaction seem to be in the temperature range from 240 to 325 C, pressure around 7 MPa, and GHSV ranging from 2000 to 5000 h1 [5,6,18,35,37]. MoS2 catalysts have water gas shift activity [6,38]. As a result, the preferred ratio of hydrogen to carbon monoxide is 1, although several papers reported a ratio of 2 [18]. To find an MoS2 based catalyst, with high activity and high ethanol and higher alcohols selectivity, capable of economically converting syngas into ethanol and higher alcohols, it is necessary to investigate which is the best preparation procedure and catalyst composition. In this study a total of 72 molybdenum sulphide catalysts have been prepared using different methods, such as thermal decomposition of ammonium tetrathiomolybdate (ATTM) to yield low or high crystalline material (with two different temperatures, named TH1 and TH2), chemical solution reaction of molybdenum hexacarbonyl with sulphur (named HC) and thermal decomposition of ATTM in the presence of water, tridecane and H2, (named WTH). The catalysts were loaded with different alkali promoters and transition metals and, finally tested under different operating conditions using a high-throughput catalyst evaluation unit, to identify the best preparation procedure and catalyst composition, based on CO conversion and selectivity to ethanol, higher alcohols and total alcohols. Briefly, the catalyst preparation procedures were as follows: In TH1 and TH2 methods, MoS2 was obtained by thermal decomposition of ATTM in flowing nitrogen at 450 C and 800 C, respectively. In HC methods, MoS2 was obtained applying a low temperature (140 C) to a solution of Mo(CO)6 and sulphur in p-xylene. In WTH method, MoS2 was obtained by decomposing ATTM in tridecane and water at 275 C, under H2 pressure (6.9 MPa). All transition metals and alkaline promoters were loaded in the catalyst’s active phase (MoS2 particles) by physical mixing or wet impregnation.
rate. To add an alkaline metal, the resulting powder was mixed with K2CO3, Cs2CO3, or Rb2CO3 with the desired atomic ratio (metal/Mo) in a glovebox, by crushing the two powders together, to obtain a homogeneous mixture and dried in a tube oven under nitrogen flow (100 ml/min) for 16 h, at 110 C, with 1 C/min ramp rate. TH2–PM method This method is similar to TH1–PM, except for the decomposition temperature which is 800 C. HC–PM method This method was based on Eq. (3) [40] (the name HC comes from the precursor’s name molybdenum hexacarbonyl and the suffix PM comes from physical mixing): 2S + Mo(CO)6 ! MoS2 + 6CO (3)
(3)
A Schlenk tube containing 100 ml of p-xylene (Fluka, product No. 95682, lot No. 1385554 51,408,139, purity 99 mass%) was cooled in liquid nitrogen until the p-xylene solidified. Vacuum was applied and the p-xylene was warmed up until it returned to liquid phase. This procedure was repeated twice and the tube was filled with nitrogen. An amount of 1.25 g sulphur (Acros, product No. 199930100, lot No. A0276552, purity 99.999 mass%) was weighed into a 3-necked flask and was transferred to an argon containing glovebox, together with the p-xylene containing Schlenk tube. The p-xylene was added to the sulphur and the flask was taken to a fume hood where it was connected to nitrogen and a reflux. The temperature was raised to 140 C in 30 min and this temperature was kept until all sulphur was dissolved (approx. 10 min). Then, the mixture was cooled to room temperature. An amount of 5.15 g of molybdenum hexacarbonyl (Acros, product No. 190390500, lot No. A0282472, purity 98 mass%) was added and the temperature was raised to 140 C in 20 min. After 150 min at 140 C the reaction mixture was cooled to room temperature and transferred to the glovebox. The black powder was filtered, dried with acetone in the glovebox and thermally treated in a tube oven under nitrogen flow (100 ml/min) for 1 h, at 550 C, with 1 C/min ramp rate. To add Co, Ni or Rh to the catalyst, the same procedure described for catalyst TH1–PM was used. The resulting powder was mixed with K2CO3, Cs2CO3 or Rb2CO3 using the same procedure described for catalyst TH1–PM. WTH method This method was based on Eq. (4) [41] (the name WTH comes from the name of some substances present in the reaction medium: water, tridecane and hydrogen):
Material and methods Catalyst preparation Around one or two grams of each catalyst were prepared for each phase of this study. All the preparation steps were performed under inert atmosphere and the catalysts were not oxidized. A list of all catalysts can be seen in Table 1. TH1–PM method This method was based on the thermal decomposition of ATTM at 450 C, according to Eq. (2) [39] (the name TH comes from thermal and the suffix PM comes from physical mixing): NH4)2MoS4 ! 2NH3 + S + H2S + MoS2 (2)
2149
(2)
ATTM (Sigma–Aldrich, product No. 323446, lot No. 00810DJ, purity 99.97 mass%) was heated in a tube oven under nitrogen flow (100 ml/min) for 2 h at 450 C with 2 C/min ramp rate. To add Co, Ni or Rh to the catalyst, aqueous salt (Co(NO3)2, Ni(NO3)2 or Rh(NO3)3) solutions were added to the powder, then, the catalysts were homogenized on a roller bench for 1 h and treated in a tube oven under nitrogen flow (100 ml/min) for 16 h, at 110 C, with 1 C/min ramp rate. The dried catalysts were calcined in a tube oven under nitrogen flow (100 ml/min) for 4 h, at 350 C, with 2 C/min ramp
NH4)2MoS4 + H2 ! MoS2 + 2NH3 + 2H2S (4)
(4)
A 150 ml Premex autoclave was loaded with 2.38 g ATTM (Sigma–Aldrich, product No. 323446, lot No. 00810DJ, purity 99.97 mass%), 100.3 g tridecane (Acros, product No. 139511000, lot No. A0280023, purity 99 mass%) and 4.91 g water. The autoclave was pressurized to 6.9 MPa with hydrogen at room temperature and purged four times. Thereafter, the autoclave was heated to 275 C and kept at this temperature for 3 h. After reaction, the reactor was cooled to 200 C and hot vented for 35 min to release water vapour, NH3 and H2S. After cooling the reactor to room temperature, the contents (black powder) were washed with 20–30 ml of acetone and filtered through a fine filter paper in an argon containing glovebox. The catalysts were dried overnight in a tube oven at 120 C (with 2 C/min ramp rate) under nitrogen flow (100 ml/min). The resulting powder was mixed with K2CO3, Cs2CO3 or Rb2CO3 with the desired atomic ratio (metal/Mo) in a glovebox, by crushing the two powders together, to obtain a homogeneous mixture. After mixing, the catalysts were dried in a tube oven under nitrogen flow (100 ml/min) for 16 h, at 110 C, with 2 C/min ramp rate.
2150
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
Table 1 Catalyst composition and preparation method of each of the tested catalysts. Transition metals loading are in mass% and alkaline metals loading are in metal/Mo atomic ratios. Transition metals loading
Alkaline metal loading
Preparation method
Alkaline metal addition method
0.1% Rh 0.1% Rh 1% Rh 0.06% Co 0.3% Co 0.06% Ni 0.3% Ni 0.25% Rh 0.15% Co 0.25% Rh 0.15% Ni 0.15% Co 0.15% Ni 0.17% Rh 0.1% Co 0.1% 0.1% Rh 1% Rh 0.06% Co 0.3% Co 0.06% Ni 0.3% Ni 0.25% Rh 0.15% Co 0.25% Rh 0.15% Ni 0.15% Co 0.15% Ni 0.17% Rh 0.1% Co 0.1% 0.1% Rh 1% Rh 0.06% Co 0.3% Co 0.06% Ni 0.3% Ni 0.25% Rh 0.15% Co 0.25% Rh 0.15% Ni 0.15% Co 0.15% Ni 0.17% Rh 0.1% Co 0.1% 0.1% Rh 1% Rh 0.06% Co 0.3% Co 0.06% Ni 0.3% Ni 0.25% Rh 0.15% Co 0.25% Rh 0.15% Ni 0.15% Co 0.15% Ni 0.17% Rh 0.1% Co 0.1%
0.7 K/Mo 1 K/Mo 1.4 K/Mo 1.8 K/Mo 1 K/Mo 1.4 K/Mo 1.8 K/Mo 0.7 K/Mo 0.7 K/Mo 0.7 K/Mo 0.3 K/Mo 0.3 Cs/Mo 0.7 Cs/Mo 0.3 Rb/Mo 0.7 Rb/Mo 0.3 K/Mo 0.7 K/Mo 0.3 Cs/Mo 0.7 Cs/Mo 0.3 Rb/Mo 0.7 Rb/Mo 0.3 K/Mo 0.3 Cs/Mo 0.7 Cs/Mo 0.3 Rb/Mo 0.7 Rb/Mo 0.3 K/Mo 0.3 Cs/Mo 0.7 Cs/Mo 0.3 Rb/Mo 0.7 Rb/Mo 0.3 K/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb 0.3 Rb/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo
TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH2 HC WTH TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 HC HC HC HC HC WTH WTH WTH WTH WTH TH1 HC HC HC HC HC HC HC HC HC HC HC HC HC HC HC HC HC HC HC HC TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1
PM PM PM PM WI WI WI PM PM PM PM PM PM PM PM WI WI WI WI WI WI PM PM PM PM PM PM PM PM PM PM WI WI WI WI WI WI WI WI WI WI WI PM PM PM PM PM PM PM PM PM PM WI WI WI WI WI WI WI WI WI WI PM PM PM PM PM PM PM PM PM PM
Ni
Ni
Ni
Ni
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
2151
spectral interferences in the other analytes. Attempts were made to overcome these interferences using a multi-component spectral fitting (MSF) procedure, which is available in PerkinElmer ICP-OES systems. The results improved after using MSF, apart from Rb. Cs was measured by ICP-MS, since ICP-OES has almost no sensitivity for this element. Actually, all of the elements were measured in 10-fold dilution using a PerkinElmer Elan DRC II instrument at analogue mode. This mode of detection avoids fatigue of the detector from ppm concentrations. The results were in good agreement with the MSF corrected ICP-OES results.
TH1–WI and HC–WI methods Alternatively, TH1–PM and HC–PM catalysts were prepared by adding K2CO3, Cs2CO3 or Rb2CO3 through wet impregnation. In these cases they were named TH1–WI and HC–WI, respectively. The procedure was similar until the mixing step of the catalyst with, for instance, K2CO3. The catalyst was sieved to a particle size range of 38–100 mm and, instead of physically mixing K2CO3 with the catalyst, K2CO3 was added by wet impregnation to obtain the desired atomic ratio (K/Mo). This was accomplished by mixing the catalyst with a K2CO3 solution on a roller bench for 2 h and then drying the catalyst in a tube oven under nitrogen flow (100 ml/min) for 32 h, at 110 C, with 1 C/min ramp rate.
Catalytic evaluation
Catalyst characterization
The catalyst screening experiments were performed in a high-throughput parallel catalyst screening unit containing 64 fixed-bed reactors operated in co-current downstream mode, with a catalyst volume between 0.05 and 0.2 ml in each reactor [42]. All reactors were sequentially sampled by an on-line GC. The complete effluent analysis consisted of three methods performed in parallel on three GC channels: CP-porabond-Q column with a FID detector; COX column with a TCD detector, and PPQ column with a TCD detector. Helium was mixed with the gas feed (5 mol% of helium in feed) as an internal standard required to calculate absolute flows. Before the beginning of the experiments, a validation run was carried out with a Cu–ZnO–alumina catalyst for methanol synthesis, to perform a technical validation of the equipment. During this validation, various process parameters were studied: temperature: 260 and 300 C; pressure: 5, 7 and 9 MPa; H2/CO molar ratio: 0.5, 1 and 2; GHSV: 4000, 8000, 16,000 h1. As expected, the main product in this reaction was methanol. Ethanol was mainly detected at higher temperatures and lower H2/CO ratios. The data from this validation run showed that the equipment is capable of producing consistent and reproducible data. The relative standard deviation of the conversion and methanol formation was less than 5% when a GHSV of 8000 h1 or lower was used. The majority of the data showed carbon mass balances higher than 94% at 260 C and higher than 92% at 300 C. Based on the blank data, no cross-contamination was identified and the feed composition was correct.
XRD analysis The diffraction data of powder samples were collected on a D8 Advance diffractometer using CuKa 1 radiation (1.54016 Å) with a germanium monochromator at room temperature. The data were collected from 5 to 35 u with 0.0184 u steps in a solid state Lynx Eye detector. The samples were placed in a flat sample holder and measured under reflection mode using a 90 positions robotic changer with 0.15 rpm rotation and 1 divergence slit. BET-analysis BET analysis was performed using a Micromeritics TriStar 3000. This apparatus provides surface area measurements by using N2 adsorption at 77.4 K. The typical sample weight was between 0.25 and 1.25 g. Before the measurement, the sample was dried for 1 h at 120 C under N2. Elemental analysis An accurately weighed portion (40–50 mg) of each sample was digested in a mixture of hydrochloric and nitric acid using an MARS5 pressurized microwave digestion system. The contents were diluted to 100 ml with purified water and by the addition of scandium solution as internal standard. These solutions were measured in a PerkinElmer Optima 7300DV ICP-OES instrument using multiple emission lines where possible for K, Rb, Rh, Co and Ni. A further 10-fold dilution of these digests was measured by ICP-OES for Mo. It was found that there were significant Mo
14000
12000
Intensity
10000
8000
TH1-WI TH2-PM HC-PM
6000
KMoS2
4000
2000
0 0
10
20
30
40
50
60
70
80
2 Theta (degree)
Fig. 1. XRD pattern of the MoS2 powder (K/Mo = 0.7). KMoS2 peaks are identified by a black square.
2152
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
Table 2 Carbon monoxide hydrogenation results. Author
Catalyst type
Space velocity
P (MPa)
T ( C)
CO Conv. (%)
EtOH Sel. (%)
Gerber et al., 2007 [36]
KMoS2 KMoS2 KMoS2a,b KMoS2b,c
6700 l/Lcat h 6700 l/Lcat h 5100 l/Lcat h 5800 l/Lcat h
8.4 8.4 5 5
325 375 300 300
11.1 20.3 19.5 14.2
22.5 19.3 19.5 21.3
Present work Present work a b c
Prepared by TH1 method. Potassium was added by K2CO3 physical mixture (K/Mo = 0.7). H2/CO molar ratio of 1.0. Prepared by HC method.
At the end of the whole study we have found that, for the data obtained with the MoS2 catalysts, all material balances measured fit to a normal and narrow distribution range of carbon mass balance values with 99% mean and 1.27% standard deviation. Results and discussion Catalysts characterization XRD analysis The diffractograms obtained for the catalysts deployed in the present study display characteristic peaks, which are not very sharp, as can be seen in Fig. 1. Such peaks indicate the presence of crystalline MoS2, according to the well reported JCPDS MoS2 data [43]. Indeed, molybdenum disulphide (MoS2) belongs to a class of group VI dichalcogenides, having C7 type crystal structure. X-ray diffraction data for MoS2 single crystals reveals three main peaks (hkl 003, 009 and 113) with relative intensities of 100, 57.42 and 64.21 [44]. Such peaks may be observed in the samples studied. Since catalysts have been calcined at high temperatures, crystalline MoS2 is likely to occur [45]. However, other components have not been detected, indicating that they are either non-crystalline or in concentrations below the detection limit of the technique. BET analysis The lowest BET surface areas were found for catalysts prepared by method TH1 (under 3 m2/g). The BET surface areas of the other catalysts were much higher (5–20 m2/g). As mentioned by Afanasiev [46], MoS2 usually has the form of hexagonal slabs and Co and Ni have a preference to decorate the edges of the slabs. A low surface area MoS2 has few edge
sites and, thus, can only accommodate low levels of metals at these edges and the remaining of the metal will form separate phases [47]. Elemental analysis The ICP-MS results for Co, Ni and Rh content and Cs/Mo, Rb/Mo and K/Mo atomic ratios are in close agreement with the expected values. Catalytic studies Preliminary tests were carried out in order to identify an adequate set of operating variables. The following conditions were selected: 0.005 mol% H2S in feed, H2/CO molar ratio of 1, 5 MPa, 260–340 C temperature range and 1000–6000 h1 space velocity range. The catalytic studies were divided into 3 phases viz. the effects of preparation methods, the effects of alkali metals (K, Rb and Cs) and those of transition metals (Co, Ni and Rh). The effects of preparation methods Catalysts prepared by TH2, HC and WTH methods were tested with a 0.7 K/Mo atomic ratio,while catalysts prepared by TH1method were tested with 0.7; 1.0; 1.4 and 1.8 K/Mo atomic ratios. The effect of the potassium addition method (physical mixture – PM – or wet impregnation – WI) was also investigated for TH1 method. Results showed that catalysts prepared by TH2 method were completely inactive and this method was abandoned. These results are in line with the findings of Woo et al. [48] who tested the activity of oxidized K2CO3/MoS2 catalysts heated above 1013 K and suggested that their low activity for carbon monoxide hydrogenation is due to the reduction of their surface area by
Fig. 2. CO conversion vs. metals in catalyst – results at 300 C and 3000 h1. When only one transition metal is present Co or Ni mass% is 0.06 or 0.3 and Rh mass% is 0.1 or 1.0 (lower level data points are marked with an inner circle). When two transition metals are present Co or Ni mass% is 0.15 and Rh mass% is 0.25. When three transition metals are present Co mass% is 0.1, Ni mass% is 0.1 and Rh mass% is 0.17.
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
12
90 TH1-PM 80
TH1-WI
70
HC-PM
60
HC-WI
Total Alcohols Carbon Yield (%)
Total Alcohols Carbon Selectivity (%)
100
2153
50 40 30 20
10
8 TH1-PM TH1-WI HC-PM HC-WI
6
4
2
10 0
0
0
5
10
15
20
25
30
35
40
0
45
5
10
15
20
25
30
35
40
45
CO Conversion (%)
CO Conversion (%)
Fig. 3. Total alcohols carbon selectivity vs. CO conversion.
Fig. 5. Total alcohols carbon yield vs. CO conversion.
sintering. Menart et al. [33] also found that higher temperature treatments (500 C) led to more crystalline and less active CoMoSx catalysts. The most selective catalysts for total alcohols were those prepared by HC method (K/Mo = 0.7), followed by WTH method (K/Mo = 0.7) and by TH1–PM method (K/Mo = 1.0). The low conversions and ethanol selectivity obtained in this phase were probably due to the absence of transition metals in the catalysts’ compositions and were in accordance with literature [36], as can be seen in Table 2.
Catalysts prepared by HC method showed the highest selectivity to ethanol and higher alcohols at low conversions. Contrary to a previous study [49], there was no significant difference among the alkaline metal promoters tested regarding total alcohols selectivity. However, the conditions used in this study were different.
The effects of alkaline metals (K, Rb and Cs) Catalysts prepared by TH1 (with potassium added by physical mixture – PM – or by wet impregnation – WI), HC and WTH methods with 0.3 and 0.7 alkaline metal (K, Rb or Cs) to Mo atomic ratios were tested without transition metals addition. Catalysts prepared by TH1 method were the most active and selective to total alcohols. Catalysts with low alkaline metal-to-Mo atomic ratio (0.3) achieved the best results for ethanol and higher alcohols selectivities. Catalysts prepared by WTH method resulted in somewhat lower activity and selectivity performances when compared to catalysts prepared by TH1 and HC methods. Such catalysts were tested because published literature [41] had reported that this preparation method would provide us with high BET surface areas. Indeed, the surface areas of these catalysts where typically higher than the surface areas of the catalysts obtained by the more usual TH1 methods, but of the same magnitude, as opposed to the literature results, what could in principle justify the lower conversions obtained for such catalysts.
The effects of transition metals (Co, Ni and Rh) HC and TH1 preparation methods (with potassium added by physical mixture – PM – or by wet impregnation – WI) were tested with varying amounts of Co, Ni (0.06, 0.10, 0.15 and 0.30 mass%) and Rh (0.10, 0.17, 0.25 and 1.0 mass%). All catalysts were prepared with a 0.3 alkaline metal-to-Mo atomic ratio. The tests were preceded by a reduction step for 2 h, with the flow of a mixture of H2 and N2 with 3/1 mol L1 ratio, 4000 h1 space velocity, at 350 C and atmospheric pressure. The results showed that incorporation of Co and Ni produced catalysts with activity levels equivalent to catalysts containing Rh (Fig. 2). Based on these results, when at least one transition metal is present, catalysts prepared by TH1 methods are clearly more active for CO overall conversion than those prepared by HC methods. It is interesting to point out that, except for HC–WI preparation method, when only one transition metal is present, catalysts containing Co and Ni display higher CO conversions at low metal content but catalysts containing Rh have an opposite behaviour. The behaviour of the Rh containing catalysts (except for HC–WI) is similar to the behaviour found by Surisetty et al. [21], who showed that CO conversion increased with the metal content of potassium modified molybdenum sulphide catalysts promoted with nickel cobalt and rhodium and supported on alumina or activated carbon.
35 TH1-PM TH1-WI HC-PM HC-WI
25
10 9 8
C2+OH Carbon Yield (%)
C2+OH Carbon Selectivity (%)
30
20 15 10
7 TH1-PM TH1-WI HC-PM HC-WI
6 5 4 3 2
5
1 0
0 0
5
10
15
20
25
30
35
40
CO Conversion (%)
Fig. 4. Ethanol and higher alcohols carbon selectivity vs. CO conversion.
45
0
5
10
15
20
25
30
35
40
45
CO Conversion (%)
Fig. 6. Ethanol and higher alcohols carbon yield vs. CO conversion.
2154
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
Total alcohols carbon selectivity (%)
100
Da Silva et al. data
90
Literature
80
EP 0 119 609 A1 EP 0 172 431 A2
70
US 4675344
60
US 4749724
50
5 MPa
40
5 MPa
30
5 MPa
20 10 0 0
5
10
15
20
25 30 Conversion (%)
35
40
45
50
Fig. 7. Total alcohols carbon selectivity vs. CO conversion. Data for the catalysts tested in this study at 5 MPa and data from literature [18] and from Dow patents reported at 5 MPa (as marked) and higher pressures.
Catalysts prepared by TH1 methods also showed higher selectivities to total alcohols, (Fig. 3). In this case, catalysts prepared by HC methods were less selective, with a higher formation of carbon dioxide. For all the catalysts tested, an increase in CO conversion led to a decrease in total alcohols selectivity confirming the findings of Andersson et al. [22]. When selectivities to ethanol and higher alcohols are considered, catalysts prepared by HC–PM method present the highest selectivities (up to 32.3%). However, such catalysts are less active, and show these higher selectivities at conversions below 20% (Fig. 4). Ferrari et al. [29] reported that catalysts prepared by physical mixing of K2CO3 and MoS2 showed much better ethanol and higher alcohols selectivity when compared to catalysts prepared by incipient wet impregnation. This difference in selectivity was not observed in this study (Fig. 4). However, the catalysts tested by Ferrari et al. did not contain transition metals, as opposed to the catalysts tested in this study. When we look at total alcohols carbon yield, catalysts prepared by TH1 methods present somewhat higher productivities when compared to catalysts prepared by HC methods (Fig. 5). On the other hand, catalysts prepared by HC methods show higher ethanol and higher alcohols carbon yields when compared to catalysts prepared by TH1 methods (Fig. 6). The reasons for the differences between the effect of Co and Ni concentration and the effect of Rh concentration on CO conversion and the reasons for the differences among the activity and selectivity of the catalysts prepared by TH1 methods and HC methods have to be further elucidated. Hence, it was possible to identify a set of promising MoS2 based catalysts for producing higher alcohols from syngas, as can be seen in Fig. 7. As not yet reported, catalysts prepared by HC–PM and HC–WI methods show higher selectivities to ethanol and higher alcohols but lower conversions, when compared to catalysts prepared by TH1–PM and TH1–WI methods. Thus, an economic evaluation has to be performed in order to determine the attractiveness of each specific method for the production of commercial catalysts, aiming at a given industrial application. If, for economic reasons, a higher selectivity to total alcohols is desired, catalysts prepared by TH1 methods should be used, with K as the alkaline promoter (K/Mo atomic ratio of 0.3) and Co, Ni or a mixture of them as the transition metal. The amount of each transition metal that should be added must be further investigated. The best operating temperature is 300 C and the best space
velocity value depends on how much methanol should be recycled to the reactor’s feed. This quantity has yet to be determined by further experiments and simulations. Conclusions Undoubtedly, the concept of high-throughput experimentation (HTE) represents a novel methodology to evaluate catalysts. HTE allows one to establish the main parameters that must be examined in the preparation step without needing to perform extensive and sophisticated methods of characterisation. MoS2 based catalysts for the conversion of syngas into alcohols were evaluated using an HTE approach. Catalysts prepared by TH1 and HC methods showed the most interesting or promising results. In future studies, our results indicate that it is worth focusing on TH1 method, trying to increase the BET surface area and activity. Indeed, previous published literature indicates that higher MoS2 specific surface areas of up to 100 m2/g may be obtained. It would also be interesting to verify the effect of the different preparation methods on MoS2 particle size. Acknowledgements The authors would like to thank Petróleo Brasileiro S.A. (Petrobras), the Brazilian National Council for Scientific and Technological Development (CNPq) and Avantium for the support provided to this research, and Dr. José Luiz Zotin, Dr. C. Martin Lok and anonymous reviewers for their comments and suggestions. References [1] J. Kjärstad, F. Johnsson, Resources and future supply of oil, Energy Policy 37 (2009) 441–464, doi:http://dx.doi.org/10.1016/j.enpol.2008.09.056. [2] A. Demirbas, Political, economic and environmental impacts of biofuels: a review, Appl. Energy 86 (2009) S108–S117, doi:http://dx.doi.org/10.1016/j. apenergy.2009.04.036. [3] A. Eisentraut, Sustainable Production of Second-Generation Biofuels, International Energy Agency, Paris, 2010. [4] International Energy Agency, Technology Roadmap – Biofuels for Transport, International Energy Agency, Paris, 2011. [5] S.D. Phillips, Technoeconomic analysis of a lignocellulosic biomass indirect gasification process to make ethanol via mixed alcohols synthesis, Ind. Eng. Chem. Res. 46 (2007) 8887–8897, doi:http://dx.doi.org/10.1021/ie071224u. [6] S. Phillips, A. Aden, J. Jechura, D. Dayton, T. Eggeman, Thermochemical Ethanol via Indirect Gasification and Mixed Alcohols Synthesis of Lignocellulosic Biomass, National Renewable Energy Laboratory, Golden, 2007.
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155 [7] D. Koch, Mixed Alcohol, Diesel and Syngas Synthesis from Forest Residues via Gasification – An Economic Analysis, School of Chemical Engineering, Georgia Institute of Technology, Georgia, USA, 2008, pp. 161. [8] T.D. Foust, A. Aden, A. Dutta, S. Phillips, An economic and environmental comparison of a biochemical and a thermochemical lignocellulosic ethanol conversion processes, Cellulose 16 (2009) 547–565, doi:http://dx.doi.org/ 10.1007/s10570-009-9317-x. [9] J.E.A. Seabra, L. Tao, H.L. Chum, I.C. Macedo, A techno-economic evaluation of the effects of centralized cellulosic ethanol and co-products refinery options with sugar cane mill clustering, Biomass Bioenergy 34 (2010) 1065–1078, doi: http://dx.doi.org/10.1016/j.biombioe.2010.01.042. [10] R.E.H. Sims, W. Mabee, J.N. Saddler, M. Taylor, An overview of second generation biofuel technologies, Bioresour. Technol. 101 (2010) 1570–1580, doi:http://dx.doi.org/10.1016/j.biortech.2009.11.046. 19963372. [11] A. Dutta, M. Talmadge, J. Hensley, M. Worley, D. Dudgeon, D. Barton, P. Groenendijk, D. Ferrari, B. Stears, E.M. Searcy, C.T. Wright, J.R. Hess, Process Design and Economics for Conversion of Lignocellulosic Biomass to Ethanol, National Renewable Energy Laboratory, Golden, CO, 2011. [12] J. He, W. Zhang, Techno-economic evaluation of thermo-chemical biomass-toethanol, Appl. Energy 88 (2011) 1224–1232, doi:http://dx.doi.org/10.1016/j. apenergy.2010.10.022. [13] A.L. Villanueva Perales, C. Reyes Valle, P. Ollero, A. Gómez-Barea, Technoeconomic assessment of ethanol production via thermochemical conversion of biomass by entrained flow gasification, Energy 36 (2011) 4097–4108, doi: http://dx.doi.org/10.1016/j.energy.2011.04.037. [14] H. van der Heijden, K.J. Ptasinski, Exergy analysis of thermochemical ethanol production via biomass gasification and catalytic synthesis, Energy 46 (2012) 200–210, doi:http://dx.doi.org/10.1016/j.energy.2012.08.036. [15] C. Reyes Valle, A.L. Villanueva Perales, F. Vidal-Barrero, A. Gómez-Barea, Techno-economic assessment of biomass-to-ethanol by indirect fluidized bed gasification: impact of reforming technologies and comparison with entrained flow gasification, Appl. Energy 109 (2013) 254–266, doi:http://dx.doi.org/ 10.1016/j.apenergy.2013.04.024. [16] L.C. de Freitas, S. Kaneko, Ethanol demand under the flex-fuel technology regime in Brazil, Energy Econ. 33 (2011) 1146–1154, doi:http://dx.doi.org/ 10.1016/j.eneco.2011.03.011. [17] E.S. Lora, R.V. Andrade, Biomass as energy source in Brazil, Renew. Sustain. Energy Rev. 13 (2009) 777–788, doi:http://dx.doi.org/10.1016/j.rser.2007.12.004. [18] V. Subramani, S.K. Gangwal, A review of recent literature to search for an efficient catalytic process for the conversion of syngas to ethanol, Energy Fuels 22 (2008) 814–839, doi:http://dx.doi.org/10.1021/ef700411x. [19] V.R. Surisetty, A.K. Dalai, J. Kozinski, Alcohols as alternative fuels: an overview, Appl. Catal. A 404 (2011) 1–11. [20] K. Fang, D. Li, M. Lin, M. Xiang, W. Wei, Y. Sun, A short review of heterogeneous catalytic process for mixed alcohols synthesis via syngas, Catal. Today 147 (2009) 133–138, doi:http://dx.doi.org/10.1016/j.cattod.2009.01.038. [21] V.R. Surisetty, I. Eswaramoorthi, A.K. Dalai, Comparative study of higher alcohols synthesis over alumina and activated carbon-supported alkalimodified MoS2 catalysts promoted with group VIII metals, Fuel 96 (2012) 77– 84, doi:http://dx.doi.org/10.1016/j.fuel.2011.12.054. [22] R. Andersson, M. Boutonnet, S. Järås, Correlation patterns and effect of syngas conversion level for product selectivity to alcohols and hydrocarbons over molybdenum sulfide based catalysts, Appl. Catal. A 417–418 (2012) 119–128, doi:http://dx.doi.org/10.1016/j.apcata.2011.12.033. [23] J.M. Christensen, P.M. Mortensen, R. Trane, P.A. Jensen, A.D. Jensen, Effects of H2S and process conditions in the synthesis of mixed alcohols from syngas over alkali promoted cobalt-molybdenum sulfide, Appl. Catal. A 366 (2009) 29–43, doi:http://dx.doi.org/10.1016/j.apcata.2009.06.034. [24] J.M. Christensen, P.A. Jensen, A.D. Jensen, Effects of feed composition and feed impurities in the catalytic conversion of syngas to higher alcohols over alkalipromoted cobalt-molybdenum sulfide, Ind. Eng. Chem. Res. 50 (2011) 7949– 7963, doi:http://dx.doi.org/10.1021/ie200235e. [25] S. Zaman, K.J. Smith, A review of molybdenum catalysts for synthesis gas conversion to alcohols: catalysts, mechanisms and kinetics, Catal. Rev. 54 (2012) 41–132, doi:http://dx.doi.org/10.1080/01614940.2012.627224. [26] R. Andersson, M. Boutonnet, S. Järås, Effect of CO2 in the synthesis of mixed alcohols from syngas over a K/Ni/MoS2 catalyst, Fuel 107 (2013) 715–723, doi: http://dx.doi.org/10.1016/j.fuel.2012.11.044. [27] R. Andersson, M. Boutonnet, S. Järås, Higher alcohols from syngas using a K/Ni/ MoS2 catalyst: trace sulfur in the product and effect of H2S-containing feed, Fuel 115 (2014) 544–550, doi:http://dx.doi.org/10.1016/j.fuel.2013.07.057. [28] S.-W. Chiang, C.-C. Chang, J.-L. Shie, C.-Y. Chang, D.-R. Ji, J.-Y. Tseng, C.-F. Chang, Y.-H. Chen, Synthesis of alcohols and alkanes from CO and H2 over MoS2/
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36] [37] [38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47] [48]
[49]
2155
g-Al2O3 catalyst in a packed bed with continuous flow, Energies 5 (2012) 4147– 4164, doi:http://dx.doi.org/10.3390/en5104147. D. Ferrari, G. Budroni, L. Bisson, N.J. Rane, B.D. Dickie, J.H. Kang, S.J. Rozeveld, Effect of potassium addition method on MoS2 performance for the syngas to alcohol reaction, Appl. Catal. A: Gen. 462–463 (2013) 302–309. A. Andersen, S.M. Kathmann, M.A. Lilga, K.O. Albrecht, R.T. Hallen, D. Mei, Effects of potassium doping on CO hydrogenation over MoS2 catalysts: a firstprinciples investigation, Catal. Commun. 52 (2014) 92–97, doi:http://dx.doi. org/10.1016/j.catcom.2014.02.011. K. Atsonios, C. Christodoulou, Koytsoumpa, K.D. Panopoulos, E. Kakaras, Plant design aspects of catalytic biosyngas conversion to higher alcohols, Biomass Bioenergy 53 (2013) 54–64, doi:http://dx.doi.org/10.1016/j.biombioe.2013.01.035. J.-J. Wang, J.-R. Xie, Y.-H. Huang, B.-H. Chen, G.-D. Lin, H.-B. Zhang, An efficient Ni–Mo–K sulfide catalyst doped with CNTs for conversion of syngas to ethanol and higher alcohols, Appl. Catal. A: Gen. 468 (2013) 44–51, doi:http://dx.doi. org/10.1016/j.apcata.2013.08.026. M.J. Menart, J.E. Hensley, K.E. Costelow, Thermal decomposition of bulk K– CoMoSx mixed alcohol catalyst precursors and effects on catalyst morphology and performance, Appl. Catal. A: Gen. 437–438 (2012) 36–43. J.E. Hensley, S. Pylypenko, D.A. Ruddy, Deactivation and stability of K–CoMoSx mixed alcohol synthesis catalysts, J. Catal. 309 (2014) 199–208, doi:http://dx. doi.org/10.1016/j.jcat.2013.10.001. J.J. Spivey, A. Egbebi, Heterogeneous catalytic synthesis of ethanol from biomass-derived syngas, Chem. Soc. Rev. 36 (2007) 1514–1528, doi:http://dx. doi.org/10.1039/b414039g. 17660882. M.A. Gerber, J.F. White, D.J. Stevens, Mixed Alcohol Synthesis Catalyst Screening, Pacific Northwest National Laboratory, Richland, 2007. M.M. Conway, C.B. Murchinson, R.R. Stevens, Method for adjusting methanol to higher alcohol ratios, USPO N US4675344 A, 06/23/1987. Z. Liu, X. Li, M.R. Close, E.L. Kugler, J.L. Petersen, D.B. Dadyburjor, Screening of alkali-promoted vapor-phase-synthesized molybdenum sulfide catalysts for the production of alcohols from synthesis gas, Ind. Eng. Chem. Res. 36 (1997) 3085–3093, doi:http://dx.doi.org/10.1021/ie960648d. H.C. Woo, I.-S. Nam, J.S. Lee, J.S. Chung, K.H. Lee, Y.G. Kim, Room-temperature oxidation of K2CO3/MoS2 catalysts and its effects on alcohol synthesis from CO and H2, J. Catal. 138 (1992) 525–535, doi:http://dx.doi.org/10.1016/0021-9517 (92)90304-Z. D. Duphil, S. Bastide, C. Lévy-Clémént, Chemical synthesis of molybdenum disulfide nanoparticles in an organic solution, J. Mater. Chem. 12 (2002) 2430– 2432, doi:http://dx.doi.org/10.1039/b202162e. Y. Yoneyama, C. Song, A new method for preparing highly active unsupported Mo sulfide. Catalytic activity for hydrogenolysis of 4-(1-naphthylmethyl) bibenzyl, Catal. Today 50 (1999) 19–27, doi:http://dx.doi.org/10.1016/S09205861(98)00477-5. J.K. van der Waal, G. Klaus, M. Smit, C.M. Lok, High-throughput experimentation in syngas based research, Catal. Today 171 (2011) 207–210, doi:http://dx. doi.org/10.1016/j.cattod.2011.02.019. J.A. Wilson, A.D. Yoffe, The transition metal dichalcogenides discussion and interpretation of the observed optical, electrical and structural properties, Adv. Phys. 18 (1969) 193–335, doi:http://dx.doi.org/10.1080/00018736900101307. K. Dartigeas, L. Benoist, D. Gonbeau, G. Pfister-Guillouzo, G. Ouvrard, A. Levasseur, Systematic X-ray photoelectron spectroscopy and theoretical studies of disulfides of groups IVB, VB and VIB transition metals, J. Electron Spectrosc. Relat. Phenom. 83 (1997) 45–58, doi:http://dx.doi.org/10.1016/ S0368-2048(96)03076-9. E.G. Derouane, E. Pedersen, B.S. Clausen, Z. Gabelica, H. Topsøe, EPR studies on the influence of NO adsorption on sulfided Co–Mo hydrodesulfurization catalysts, J. Catal. 107 (1987) 587–588, doi:http://dx.doi.org/10.1016/00219517(87)90325-3. P. Afanasiev, Synthetic approaches to the molybdenum sulfide materials, Compt. Rend. Chim. 11 (2008) 159–182, doi:http://dx.doi.org/10.1016/j. crci.2007.04.009. H. Topsøe, B.S. Clausen, F.E. Massoth, Hydrotreating catalysis, Catal. Sci. Technol. 11 (1996) 1–269. H.C. Woo, J.C. Kim, I.-S. Nam, J.S. Lee, J.S. Chung, Y.G. Kim, Surface species on the oxidized K2CO3/MoS2 and their effects on catalytic carbon monoxide hydrogenation, Appl. Catal. A 104 (1993) 199–214, doi:http://dx.doi.org/ 10.1016/0926-860X(93)85098-A. N. Koizumi, K. Murai, T. Ozaki, M. Yamada, Development of sulfur tolerant catalysts for the synthesis of high quality transportation fuels, Catal. Today 89 (2004) 465–478, doi:http://dx.doi.org/10.1016/j.cattod.2004.02.002.
Contents lists available at ScienceDirect
Journal of Environmental Chemical Engineering journal homepage: www.elsevier.com/locate/jece
Evaluation of MoS2 based catalysts for the conversion of syngas into alcohols: A combinatorial approach Arthur José Gerbasi da Silva a,b, *, Paula Claassens-Dekker c , Antônio Carlos Sallarès de Mattos Carvalho b , Antônio Manzolillo Sanseverino b , Cristina Pontes Bittencourt Quitete b , Alexandre Szklo a , Eduardo Falabella Sousa-Aguiar b,d a Energy Planning Program, Graduate School of Engineering, UFRJ – CT, Bloco C, Sala 211, Cidade Universitária, Ilha do Fundão, Rio de Janeiro, RJ 21941-972, Brazil b Petrobras Research Center – Avenida Horácio Macedo, 950, Cidade Universitária, Ilha do Fundão, Rio de Janeiro, RJ 21941-915, Brazil c Avantium Chemicals B.V., Zekeringstraat 29, 1014 BV, Amsterdam, The Netherlands d Department of Organic Processes, School of Chemistry, UFRJ – CT, Bloco E, Cidade Universitária, Ilha do Fundão, Rio de Janeiro, RJ 21941-909, Brazil
A R T I C L E I N F O
A B S T R A C T
Article history: Received 10 April 2014 Received in revised form 14 August 2014 Accepted 3 September 2014
72 MoS2 catalysts were tested in the conversion of syngas to alcohols, using a high-throughput catalyst evaluation unit, to identify the best catalyst, based on CO conversion, both ethanol and higher alcohols and total alcohols selectivity. Catalysts prepared by thermal decomposition of (NH4)2MoS4 at low temperature showed a higher selectivity to total alcohols. The highest selectivity to ethanol and higher alcohols was obtained at 300 C by a catalyst prepared by reacting Mo(CO)6 with sulphur. Catalysts prepared by thermal decomposition of (NH4)2MoS4 at high temperature showed very low activity. Catalysts prepared by thermal decomposition of (NH4)2MoS4 in tridecane/water with hydrogen atmosphere showed low activity and selectivity. There was no significant difference among the alkaline metal promoters K, Cs and Rb regarding total alcohols selectivities. Incorporation of Co and Ni led to catalysts with activity levels equivalent to catalysts that contain Rh. ã 2014 Elsevier Ltd. All rights reserved.
Keywords: Syngas conversion Molybdenum sulphide catalyst Ethanol Higher alcohols
Introduction Increasing awareness about the consequences of global warming and energy dependence has renewed the interest in the development of alternative fuels [1–4]. One option is the production of ethanol from biomass residues through the gasification of biomass and the conversion of the resulting synthesis gas into ethanol and higher alcohols [5–15]. This solution is particularly suited for Brazil, given its growing flex-fuelled (gasoline/ethanol) light vehicle fleet and great availability of sugar cane residues [16,17]. One of the biggest challenges to develop this option is to find a catalyst capable of economically converting syngas into ethanol and higher alcohols. There are various types of catalysts that can be used in this process, but one of the most promising is based on alkali promoted molybdenum sulphide
[18–20]. Various studies have examined the performance of alkali promoted molybdenum sulphide based catalysts, prepared by different methods and used under different operating conditions, for the conversion of syngas into ethanol and higher alcohols [21–34]. The thermodynamics of ethanol and higher alcohols formation have been assessed through theoretical calculations with Aspen software [35], and also by calculations with HSC Chemistry Software [18]. Methanol formation is favoured at low temperature and high pressure. At high pressure the formation of higher alcohols increases with increasing temperature [18]. Ethanol formation from syngas is a highly exothermic reaction (Eq. (1)), and heat dissipation may be a problem for scale-up [36]. 2CO(g) + 4H2(g) ! C2H5OH + H2O (1)
DH 298 = 253.6 kJ/mol of ethanol
* Corresponding author at: Petrobras Research Center – Avenida Horácio Macedo, 950, Cidade Universitária, Ilha do Fundão, Rio de Janeiro, RJ 21941-915, Brazil. Tel.: +55 21 2162 7168; fax: +55 21 2162 1007. E-mail address: [email protected] (A.J. Gerbasi da Silva). http://dx.doi.org/10.1016/j.jece.2014.09.006 2213-3437/ ã 2014 Elsevier Ltd. All rights reserved.
DG 298 = 221.1 kJ/mol of ethanol
(1)
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
Alkali promoted molybdenum sulphide catalysts produce linear alcohols and ethanol and higher alcohols are formed via a classical insertion of CO into the corresponding precursor alcohol [21]. CO hydrogenation to methane is a competitive reaction to the alcohol synthesis with such catalysts. The optimum conditions for this reaction seem to be in the temperature range from 240 to 325 C, pressure around 7 MPa, and GHSV ranging from 2000 to 5000 h1 [5,6,18,35,37]. MoS2 catalysts have water gas shift activity [6,38]. As a result, the preferred ratio of hydrogen to carbon monoxide is 1, although several papers reported a ratio of 2 [18]. To find an MoS2 based catalyst, with high activity and high ethanol and higher alcohols selectivity, capable of economically converting syngas into ethanol and higher alcohols, it is necessary to investigate which is the best preparation procedure and catalyst composition. In this study a total of 72 molybdenum sulphide catalysts have been prepared using different methods, such as thermal decomposition of ammonium tetrathiomolybdate (ATTM) to yield low or high crystalline material (with two different temperatures, named TH1 and TH2), chemical solution reaction of molybdenum hexacarbonyl with sulphur (named HC) and thermal decomposition of ATTM in the presence of water, tridecane and H2, (named WTH). The catalysts were loaded with different alkali promoters and transition metals and, finally tested under different operating conditions using a high-throughput catalyst evaluation unit, to identify the best preparation procedure and catalyst composition, based on CO conversion and selectivity to ethanol, higher alcohols and total alcohols. Briefly, the catalyst preparation procedures were as follows: In TH1 and TH2 methods, MoS2 was obtained by thermal decomposition of ATTM in flowing nitrogen at 450 C and 800 C, respectively. In HC methods, MoS2 was obtained applying a low temperature (140 C) to a solution of Mo(CO)6 and sulphur in p-xylene. In WTH method, MoS2 was obtained by decomposing ATTM in tridecane and water at 275 C, under H2 pressure (6.9 MPa). All transition metals and alkaline promoters were loaded in the catalyst’s active phase (MoS2 particles) by physical mixing or wet impregnation.
rate. To add an alkaline metal, the resulting powder was mixed with K2CO3, Cs2CO3, or Rb2CO3 with the desired atomic ratio (metal/Mo) in a glovebox, by crushing the two powders together, to obtain a homogeneous mixture and dried in a tube oven under nitrogen flow (100 ml/min) for 16 h, at 110 C, with 1 C/min ramp rate. TH2–PM method This method is similar to TH1–PM, except for the decomposition temperature which is 800 C. HC–PM method This method was based on Eq. (3) [40] (the name HC comes from the precursor’s name molybdenum hexacarbonyl and the suffix PM comes from physical mixing): 2S + Mo(CO)6 ! MoS2 + 6CO (3)
(3)
A Schlenk tube containing 100 ml of p-xylene (Fluka, product No. 95682, lot No. 1385554 51,408,139, purity 99 mass%) was cooled in liquid nitrogen until the p-xylene solidified. Vacuum was applied and the p-xylene was warmed up until it returned to liquid phase. This procedure was repeated twice and the tube was filled with nitrogen. An amount of 1.25 g sulphur (Acros, product No. 199930100, lot No. A0276552, purity 99.999 mass%) was weighed into a 3-necked flask and was transferred to an argon containing glovebox, together with the p-xylene containing Schlenk tube. The p-xylene was added to the sulphur and the flask was taken to a fume hood where it was connected to nitrogen and a reflux. The temperature was raised to 140 C in 30 min and this temperature was kept until all sulphur was dissolved (approx. 10 min). Then, the mixture was cooled to room temperature. An amount of 5.15 g of molybdenum hexacarbonyl (Acros, product No. 190390500, lot No. A0282472, purity 98 mass%) was added and the temperature was raised to 140 C in 20 min. After 150 min at 140 C the reaction mixture was cooled to room temperature and transferred to the glovebox. The black powder was filtered, dried with acetone in the glovebox and thermally treated in a tube oven under nitrogen flow (100 ml/min) for 1 h, at 550 C, with 1 C/min ramp rate. To add Co, Ni or Rh to the catalyst, the same procedure described for catalyst TH1–PM was used. The resulting powder was mixed with K2CO3, Cs2CO3 or Rb2CO3 using the same procedure described for catalyst TH1–PM. WTH method This method was based on Eq. (4) [41] (the name WTH comes from the name of some substances present in the reaction medium: water, tridecane and hydrogen):
Material and methods Catalyst preparation Around one or two grams of each catalyst were prepared for each phase of this study. All the preparation steps were performed under inert atmosphere and the catalysts were not oxidized. A list of all catalysts can be seen in Table 1. TH1–PM method This method was based on the thermal decomposition of ATTM at 450 C, according to Eq. (2) [39] (the name TH comes from thermal and the suffix PM comes from physical mixing): NH4)2MoS4 ! 2NH3 + S + H2S + MoS2 (2)
2149
(2)
ATTM (Sigma–Aldrich, product No. 323446, lot No. 00810DJ, purity 99.97 mass%) was heated in a tube oven under nitrogen flow (100 ml/min) for 2 h at 450 C with 2 C/min ramp rate. To add Co, Ni or Rh to the catalyst, aqueous salt (Co(NO3)2, Ni(NO3)2 or Rh(NO3)3) solutions were added to the powder, then, the catalysts were homogenized on a roller bench for 1 h and treated in a tube oven under nitrogen flow (100 ml/min) for 16 h, at 110 C, with 1 C/min ramp rate. The dried catalysts were calcined in a tube oven under nitrogen flow (100 ml/min) for 4 h, at 350 C, with 2 C/min ramp
NH4)2MoS4 + H2 ! MoS2 + 2NH3 + 2H2S (4)
(4)
A 150 ml Premex autoclave was loaded with 2.38 g ATTM (Sigma–Aldrich, product No. 323446, lot No. 00810DJ, purity 99.97 mass%), 100.3 g tridecane (Acros, product No. 139511000, lot No. A0280023, purity 99 mass%) and 4.91 g water. The autoclave was pressurized to 6.9 MPa with hydrogen at room temperature and purged four times. Thereafter, the autoclave was heated to 275 C and kept at this temperature for 3 h. After reaction, the reactor was cooled to 200 C and hot vented for 35 min to release water vapour, NH3 and H2S. After cooling the reactor to room temperature, the contents (black powder) were washed with 20–30 ml of acetone and filtered through a fine filter paper in an argon containing glovebox. The catalysts were dried overnight in a tube oven at 120 C (with 2 C/min ramp rate) under nitrogen flow (100 ml/min). The resulting powder was mixed with K2CO3, Cs2CO3 or Rb2CO3 with the desired atomic ratio (metal/Mo) in a glovebox, by crushing the two powders together, to obtain a homogeneous mixture. After mixing, the catalysts were dried in a tube oven under nitrogen flow (100 ml/min) for 16 h, at 110 C, with 2 C/min ramp rate.
2150
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
Table 1 Catalyst composition and preparation method of each of the tested catalysts. Transition metals loading are in mass% and alkaline metals loading are in metal/Mo atomic ratios. Transition metals loading
Alkaline metal loading
Preparation method
Alkaline metal addition method
0.1% Rh 0.1% Rh 1% Rh 0.06% Co 0.3% Co 0.06% Ni 0.3% Ni 0.25% Rh 0.15% Co 0.25% Rh 0.15% Ni 0.15% Co 0.15% Ni 0.17% Rh 0.1% Co 0.1% 0.1% Rh 1% Rh 0.06% Co 0.3% Co 0.06% Ni 0.3% Ni 0.25% Rh 0.15% Co 0.25% Rh 0.15% Ni 0.15% Co 0.15% Ni 0.17% Rh 0.1% Co 0.1% 0.1% Rh 1% Rh 0.06% Co 0.3% Co 0.06% Ni 0.3% Ni 0.25% Rh 0.15% Co 0.25% Rh 0.15% Ni 0.15% Co 0.15% Ni 0.17% Rh 0.1% Co 0.1% 0.1% Rh 1% Rh 0.06% Co 0.3% Co 0.06% Ni 0.3% Ni 0.25% Rh 0.15% Co 0.25% Rh 0.15% Ni 0.15% Co 0.15% Ni 0.17% Rh 0.1% Co 0.1%
0.7 K/Mo 1 K/Mo 1.4 K/Mo 1.8 K/Mo 1 K/Mo 1.4 K/Mo 1.8 K/Mo 0.7 K/Mo 0.7 K/Mo 0.7 K/Mo 0.3 K/Mo 0.3 Cs/Mo 0.7 Cs/Mo 0.3 Rb/Mo 0.7 Rb/Mo 0.3 K/Mo 0.7 K/Mo 0.3 Cs/Mo 0.7 Cs/Mo 0.3 Rb/Mo 0.7 Rb/Mo 0.3 K/Mo 0.3 Cs/Mo 0.7 Cs/Mo 0.3 Rb/Mo 0.7 Rb/Mo 0.3 K/Mo 0.3 Cs/Mo 0.7 Cs/Mo 0.3 Rb/Mo 0.7 Rb/Mo 0.3 K/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 Rb 0.3 Rb/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 Rb/Mo 0.3 Rb/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 K/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo 0.3 Cs/Mo
TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH2 HC WTH TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 HC HC HC HC HC WTH WTH WTH WTH WTH TH1 HC HC HC HC HC HC HC HC HC HC HC HC HC HC HC HC HC HC HC HC TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1 TH1
PM PM PM PM WI WI WI PM PM PM PM PM PM PM PM WI WI WI WI WI WI PM PM PM PM PM PM PM PM PM PM WI WI WI WI WI WI WI WI WI WI WI PM PM PM PM PM PM PM PM PM PM WI WI WI WI WI WI WI WI WI WI PM PM PM PM PM PM PM PM PM PM
Ni
Ni
Ni
Ni
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
2151
spectral interferences in the other analytes. Attempts were made to overcome these interferences using a multi-component spectral fitting (MSF) procedure, which is available in PerkinElmer ICP-OES systems. The results improved after using MSF, apart from Rb. Cs was measured by ICP-MS, since ICP-OES has almost no sensitivity for this element. Actually, all of the elements were measured in 10-fold dilution using a PerkinElmer Elan DRC II instrument at analogue mode. This mode of detection avoids fatigue of the detector from ppm concentrations. The results were in good agreement with the MSF corrected ICP-OES results.
TH1–WI and HC–WI methods Alternatively, TH1–PM and HC–PM catalysts were prepared by adding K2CO3, Cs2CO3 or Rb2CO3 through wet impregnation. In these cases they were named TH1–WI and HC–WI, respectively. The procedure was similar until the mixing step of the catalyst with, for instance, K2CO3. The catalyst was sieved to a particle size range of 38–100 mm and, instead of physically mixing K2CO3 with the catalyst, K2CO3 was added by wet impregnation to obtain the desired atomic ratio (K/Mo). This was accomplished by mixing the catalyst with a K2CO3 solution on a roller bench for 2 h and then drying the catalyst in a tube oven under nitrogen flow (100 ml/min) for 32 h, at 110 C, with 1 C/min ramp rate.
Catalytic evaluation
Catalyst characterization
The catalyst screening experiments were performed in a high-throughput parallel catalyst screening unit containing 64 fixed-bed reactors operated in co-current downstream mode, with a catalyst volume between 0.05 and 0.2 ml in each reactor [42]. All reactors were sequentially sampled by an on-line GC. The complete effluent analysis consisted of three methods performed in parallel on three GC channels: CP-porabond-Q column with a FID detector; COX column with a TCD detector, and PPQ column with a TCD detector. Helium was mixed with the gas feed (5 mol% of helium in feed) as an internal standard required to calculate absolute flows. Before the beginning of the experiments, a validation run was carried out with a Cu–ZnO–alumina catalyst for methanol synthesis, to perform a technical validation of the equipment. During this validation, various process parameters were studied: temperature: 260 and 300 C; pressure: 5, 7 and 9 MPa; H2/CO molar ratio: 0.5, 1 and 2; GHSV: 4000, 8000, 16,000 h1. As expected, the main product in this reaction was methanol. Ethanol was mainly detected at higher temperatures and lower H2/CO ratios. The data from this validation run showed that the equipment is capable of producing consistent and reproducible data. The relative standard deviation of the conversion and methanol formation was less than 5% when a GHSV of 8000 h1 or lower was used. The majority of the data showed carbon mass balances higher than 94% at 260 C and higher than 92% at 300 C. Based on the blank data, no cross-contamination was identified and the feed composition was correct.
XRD analysis The diffraction data of powder samples were collected on a D8 Advance diffractometer using CuKa 1 radiation (1.54016 Å) with a germanium monochromator at room temperature. The data were collected from 5 to 35 u with 0.0184 u steps in a solid state Lynx Eye detector. The samples were placed in a flat sample holder and measured under reflection mode using a 90 positions robotic changer with 0.15 rpm rotation and 1 divergence slit. BET-analysis BET analysis was performed using a Micromeritics TriStar 3000. This apparatus provides surface area measurements by using N2 adsorption at 77.4 K. The typical sample weight was between 0.25 and 1.25 g. Before the measurement, the sample was dried for 1 h at 120 C under N2. Elemental analysis An accurately weighed portion (40–50 mg) of each sample was digested in a mixture of hydrochloric and nitric acid using an MARS5 pressurized microwave digestion system. The contents were diluted to 100 ml with purified water and by the addition of scandium solution as internal standard. These solutions were measured in a PerkinElmer Optima 7300DV ICP-OES instrument using multiple emission lines where possible for K, Rb, Rh, Co and Ni. A further 10-fold dilution of these digests was measured by ICP-OES for Mo. It was found that there were significant Mo
14000
12000
Intensity
10000
8000
TH1-WI TH2-PM HC-PM
6000
KMoS2
4000
2000
0 0
10
20
30
40
50
60
70
80
2 Theta (degree)
Fig. 1. XRD pattern of the MoS2 powder (K/Mo = 0.7). KMoS2 peaks are identified by a black square.
2152
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
Table 2 Carbon monoxide hydrogenation results. Author
Catalyst type
Space velocity
P (MPa)
T ( C)
CO Conv. (%)
EtOH Sel. (%)
Gerber et al., 2007 [36]
KMoS2 KMoS2 KMoS2a,b KMoS2b,c
6700 l/Lcat h 6700 l/Lcat h 5100 l/Lcat h 5800 l/Lcat h
8.4 8.4 5 5
325 375 300 300
11.1 20.3 19.5 14.2
22.5 19.3 19.5 21.3
Present work Present work a b c
Prepared by TH1 method. Potassium was added by K2CO3 physical mixture (K/Mo = 0.7). H2/CO molar ratio of 1.0. Prepared by HC method.
At the end of the whole study we have found that, for the data obtained with the MoS2 catalysts, all material balances measured fit to a normal and narrow distribution range of carbon mass balance values with 99% mean and 1.27% standard deviation. Results and discussion Catalysts characterization XRD analysis The diffractograms obtained for the catalysts deployed in the present study display characteristic peaks, which are not very sharp, as can be seen in Fig. 1. Such peaks indicate the presence of crystalline MoS2, according to the well reported JCPDS MoS2 data [43]. Indeed, molybdenum disulphide (MoS2) belongs to a class of group VI dichalcogenides, having C7 type crystal structure. X-ray diffraction data for MoS2 single crystals reveals three main peaks (hkl 003, 009 and 113) with relative intensities of 100, 57.42 and 64.21 [44]. Such peaks may be observed in the samples studied. Since catalysts have been calcined at high temperatures, crystalline MoS2 is likely to occur [45]. However, other components have not been detected, indicating that they are either non-crystalline or in concentrations below the detection limit of the technique. BET analysis The lowest BET surface areas were found for catalysts prepared by method TH1 (under 3 m2/g). The BET surface areas of the other catalysts were much higher (5–20 m2/g). As mentioned by Afanasiev [46], MoS2 usually has the form of hexagonal slabs and Co and Ni have a preference to decorate the edges of the slabs. A low surface area MoS2 has few edge
sites and, thus, can only accommodate low levels of metals at these edges and the remaining of the metal will form separate phases [47]. Elemental analysis The ICP-MS results for Co, Ni and Rh content and Cs/Mo, Rb/Mo and K/Mo atomic ratios are in close agreement with the expected values. Catalytic studies Preliminary tests were carried out in order to identify an adequate set of operating variables. The following conditions were selected: 0.005 mol% H2S in feed, H2/CO molar ratio of 1, 5 MPa, 260–340 C temperature range and 1000–6000 h1 space velocity range. The catalytic studies were divided into 3 phases viz. the effects of preparation methods, the effects of alkali metals (K, Rb and Cs) and those of transition metals (Co, Ni and Rh). The effects of preparation methods Catalysts prepared by TH2, HC and WTH methods were tested with a 0.7 K/Mo atomic ratio,while catalysts prepared by TH1method were tested with 0.7; 1.0; 1.4 and 1.8 K/Mo atomic ratios. The effect of the potassium addition method (physical mixture – PM – or wet impregnation – WI) was also investigated for TH1 method. Results showed that catalysts prepared by TH2 method were completely inactive and this method was abandoned. These results are in line with the findings of Woo et al. [48] who tested the activity of oxidized K2CO3/MoS2 catalysts heated above 1013 K and suggested that their low activity for carbon monoxide hydrogenation is due to the reduction of their surface area by
Fig. 2. CO conversion vs. metals in catalyst – results at 300 C and 3000 h1. When only one transition metal is present Co or Ni mass% is 0.06 or 0.3 and Rh mass% is 0.1 or 1.0 (lower level data points are marked with an inner circle). When two transition metals are present Co or Ni mass% is 0.15 and Rh mass% is 0.25. When three transition metals are present Co mass% is 0.1, Ni mass% is 0.1 and Rh mass% is 0.17.
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
12
90 TH1-PM 80
TH1-WI
70
HC-PM
60
HC-WI
Total Alcohols Carbon Yield (%)
Total Alcohols Carbon Selectivity (%)
100
2153
50 40 30 20
10
8 TH1-PM TH1-WI HC-PM HC-WI
6
4
2
10 0
0
0
5
10
15
20
25
30
35
40
0
45
5
10
15
20
25
30
35
40
45
CO Conversion (%)
CO Conversion (%)
Fig. 3. Total alcohols carbon selectivity vs. CO conversion.
Fig. 5. Total alcohols carbon yield vs. CO conversion.
sintering. Menart et al. [33] also found that higher temperature treatments (500 C) led to more crystalline and less active CoMoSx catalysts. The most selective catalysts for total alcohols were those prepared by HC method (K/Mo = 0.7), followed by WTH method (K/Mo = 0.7) and by TH1–PM method (K/Mo = 1.0). The low conversions and ethanol selectivity obtained in this phase were probably due to the absence of transition metals in the catalysts’ compositions and were in accordance with literature [36], as can be seen in Table 2.
Catalysts prepared by HC method showed the highest selectivity to ethanol and higher alcohols at low conversions. Contrary to a previous study [49], there was no significant difference among the alkaline metal promoters tested regarding total alcohols selectivity. However, the conditions used in this study were different.
The effects of alkaline metals (K, Rb and Cs) Catalysts prepared by TH1 (with potassium added by physical mixture – PM – or by wet impregnation – WI), HC and WTH methods with 0.3 and 0.7 alkaline metal (K, Rb or Cs) to Mo atomic ratios were tested without transition metals addition. Catalysts prepared by TH1 method were the most active and selective to total alcohols. Catalysts with low alkaline metal-to-Mo atomic ratio (0.3) achieved the best results for ethanol and higher alcohols selectivities. Catalysts prepared by WTH method resulted in somewhat lower activity and selectivity performances when compared to catalysts prepared by TH1 and HC methods. Such catalysts were tested because published literature [41] had reported that this preparation method would provide us with high BET surface areas. Indeed, the surface areas of these catalysts where typically higher than the surface areas of the catalysts obtained by the more usual TH1 methods, but of the same magnitude, as opposed to the literature results, what could in principle justify the lower conversions obtained for such catalysts.
The effects of transition metals (Co, Ni and Rh) HC and TH1 preparation methods (with potassium added by physical mixture – PM – or by wet impregnation – WI) were tested with varying amounts of Co, Ni (0.06, 0.10, 0.15 and 0.30 mass%) and Rh (0.10, 0.17, 0.25 and 1.0 mass%). All catalysts were prepared with a 0.3 alkaline metal-to-Mo atomic ratio. The tests were preceded by a reduction step for 2 h, with the flow of a mixture of H2 and N2 with 3/1 mol L1 ratio, 4000 h1 space velocity, at 350 C and atmospheric pressure. The results showed that incorporation of Co and Ni produced catalysts with activity levels equivalent to catalysts containing Rh (Fig. 2). Based on these results, when at least one transition metal is present, catalysts prepared by TH1 methods are clearly more active for CO overall conversion than those prepared by HC methods. It is interesting to point out that, except for HC–WI preparation method, when only one transition metal is present, catalysts containing Co and Ni display higher CO conversions at low metal content but catalysts containing Rh have an opposite behaviour. The behaviour of the Rh containing catalysts (except for HC–WI) is similar to the behaviour found by Surisetty et al. [21], who showed that CO conversion increased with the metal content of potassium modified molybdenum sulphide catalysts promoted with nickel cobalt and rhodium and supported on alumina or activated carbon.
35 TH1-PM TH1-WI HC-PM HC-WI
25
10 9 8
C2+OH Carbon Yield (%)
C2+OH Carbon Selectivity (%)
30
20 15 10
7 TH1-PM TH1-WI HC-PM HC-WI
6 5 4 3 2
5
1 0
0 0
5
10
15
20
25
30
35
40
CO Conversion (%)
Fig. 4. Ethanol and higher alcohols carbon selectivity vs. CO conversion.
45
0
5
10
15
20
25
30
35
40
45
CO Conversion (%)
Fig. 6. Ethanol and higher alcohols carbon yield vs. CO conversion.
2154
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155
Total alcohols carbon selectivity (%)
100
Da Silva et al. data
90
Literature
80
EP 0 119 609 A1 EP 0 172 431 A2
70
US 4675344
60
US 4749724
50
5 MPa
40
5 MPa
30
5 MPa
20 10 0 0
5
10
15
20
25 30 Conversion (%)
35
40
45
50
Fig. 7. Total alcohols carbon selectivity vs. CO conversion. Data for the catalysts tested in this study at 5 MPa and data from literature [18] and from Dow patents reported at 5 MPa (as marked) and higher pressures.
Catalysts prepared by TH1 methods also showed higher selectivities to total alcohols, (Fig. 3). In this case, catalysts prepared by HC methods were less selective, with a higher formation of carbon dioxide. For all the catalysts tested, an increase in CO conversion led to a decrease in total alcohols selectivity confirming the findings of Andersson et al. [22]. When selectivities to ethanol and higher alcohols are considered, catalysts prepared by HC–PM method present the highest selectivities (up to 32.3%). However, such catalysts are less active, and show these higher selectivities at conversions below 20% (Fig. 4). Ferrari et al. [29] reported that catalysts prepared by physical mixing of K2CO3 and MoS2 showed much better ethanol and higher alcohols selectivity when compared to catalysts prepared by incipient wet impregnation. This difference in selectivity was not observed in this study (Fig. 4). However, the catalysts tested by Ferrari et al. did not contain transition metals, as opposed to the catalysts tested in this study. When we look at total alcohols carbon yield, catalysts prepared by TH1 methods present somewhat higher productivities when compared to catalysts prepared by HC methods (Fig. 5). On the other hand, catalysts prepared by HC methods show higher ethanol and higher alcohols carbon yields when compared to catalysts prepared by TH1 methods (Fig. 6). The reasons for the differences between the effect of Co and Ni concentration and the effect of Rh concentration on CO conversion and the reasons for the differences among the activity and selectivity of the catalysts prepared by TH1 methods and HC methods have to be further elucidated. Hence, it was possible to identify a set of promising MoS2 based catalysts for producing higher alcohols from syngas, as can be seen in Fig. 7. As not yet reported, catalysts prepared by HC–PM and HC–WI methods show higher selectivities to ethanol and higher alcohols but lower conversions, when compared to catalysts prepared by TH1–PM and TH1–WI methods. Thus, an economic evaluation has to be performed in order to determine the attractiveness of each specific method for the production of commercial catalysts, aiming at a given industrial application. If, for economic reasons, a higher selectivity to total alcohols is desired, catalysts prepared by TH1 methods should be used, with K as the alkaline promoter (K/Mo atomic ratio of 0.3) and Co, Ni or a mixture of them as the transition metal. The amount of each transition metal that should be added must be further investigated. The best operating temperature is 300 C and the best space
velocity value depends on how much methanol should be recycled to the reactor’s feed. This quantity has yet to be determined by further experiments and simulations. Conclusions Undoubtedly, the concept of high-throughput experimentation (HTE) represents a novel methodology to evaluate catalysts. HTE allows one to establish the main parameters that must be examined in the preparation step without needing to perform extensive and sophisticated methods of characterisation. MoS2 based catalysts for the conversion of syngas into alcohols were evaluated using an HTE approach. Catalysts prepared by TH1 and HC methods showed the most interesting or promising results. In future studies, our results indicate that it is worth focusing on TH1 method, trying to increase the BET surface area and activity. Indeed, previous published literature indicates that higher MoS2 specific surface areas of up to 100 m2/g may be obtained. It would also be interesting to verify the effect of the different preparation methods on MoS2 particle size. Acknowledgements The authors would like to thank Petróleo Brasileiro S.A. (Petrobras), the Brazilian National Council for Scientific and Technological Development (CNPq) and Avantium for the support provided to this research, and Dr. José Luiz Zotin, Dr. C. Martin Lok and anonymous reviewers for their comments and suggestions. References [1] J. Kjärstad, F. Johnsson, Resources and future supply of oil, Energy Policy 37 (2009) 441–464, doi:http://dx.doi.org/10.1016/j.enpol.2008.09.056. [2] A. Demirbas, Political, economic and environmental impacts of biofuels: a review, Appl. Energy 86 (2009) S108–S117, doi:http://dx.doi.org/10.1016/j. apenergy.2009.04.036. [3] A. Eisentraut, Sustainable Production of Second-Generation Biofuels, International Energy Agency, Paris, 2010. [4] International Energy Agency, Technology Roadmap – Biofuels for Transport, International Energy Agency, Paris, 2011. [5] S.D. Phillips, Technoeconomic analysis of a lignocellulosic biomass indirect gasification process to make ethanol via mixed alcohols synthesis, Ind. Eng. Chem. Res. 46 (2007) 8887–8897, doi:http://dx.doi.org/10.1021/ie071224u. [6] S. Phillips, A. Aden, J. Jechura, D. Dayton, T. Eggeman, Thermochemical Ethanol via Indirect Gasification and Mixed Alcohols Synthesis of Lignocellulosic Biomass, National Renewable Energy Laboratory, Golden, 2007.
A.J. Gerbasi da Silva et al. / Journal of Environmental Chemical Engineering 2 (2014) 2148–2155 [7] D. Koch, Mixed Alcohol, Diesel and Syngas Synthesis from Forest Residues via Gasification – An Economic Analysis, School of Chemical Engineering, Georgia Institute of Technology, Georgia, USA, 2008, pp. 161. [8] T.D. Foust, A. Aden, A. Dutta, S. Phillips, An economic and environmental comparison of a biochemical and a thermochemical lignocellulosic ethanol conversion processes, Cellulose 16 (2009) 547–565, doi:http://dx.doi.org/ 10.1007/s10570-009-9317-x. [9] J.E.A. Seabra, L. Tao, H.L. Chum, I.C. Macedo, A techno-economic evaluation of the effects of centralized cellulosic ethanol and co-products refinery options with sugar cane mill clustering, Biomass Bioenergy 34 (2010) 1065–1078, doi: http://dx.doi.org/10.1016/j.biombioe.2010.01.042. [10] R.E.H. Sims, W. Mabee, J.N. Saddler, M. Taylor, An overview of second generation biofuel technologies, Bioresour. Technol. 101 (2010) 1570–1580, doi:http://dx.doi.org/10.1016/j.biortech.2009.11.046. 19963372. [11] A. Dutta, M. Talmadge, J. Hensley, M. Worley, D. Dudgeon, D. Barton, P. Groenendijk, D. Ferrari, B. Stears, E.M. Searcy, C.T. Wright, J.R. Hess, Process Design and Economics for Conversion of Lignocellulosic Biomass to Ethanol, National Renewable Energy Laboratory, Golden, CO, 2011. [12] J. He, W. Zhang, Techno-economic evaluation of thermo-chemical biomass-toethanol, Appl. Energy 88 (2011) 1224–1232, doi:http://dx.doi.org/10.1016/j. apenergy.2010.10.022. [13] A.L. Villanueva Perales, C. Reyes Valle, P. Ollero, A. Gómez-Barea, Technoeconomic assessment of ethanol production via thermochemical conversion of biomass by entrained flow gasification, Energy 36 (2011) 4097–4108, doi: http://dx.doi.org/10.1016/j.energy.2011.04.037. [14] H. van der Heijden, K.J. Ptasinski, Exergy analysis of thermochemical ethanol production via biomass gasification and catalytic synthesis, Energy 46 (2012) 200–210, doi:http://dx.doi.org/10.1016/j.energy.2012.08.036. [15] C. Reyes Valle, A.L. Villanueva Perales, F. Vidal-Barrero, A. Gómez-Barea, Techno-economic assessment of biomass-to-ethanol by indirect fluidized bed gasification: impact of reforming technologies and comparison with entrained flow gasification, Appl. Energy 109 (2013) 254–266, doi:http://dx.doi.org/ 10.1016/j.apenergy.2013.04.024. [16] L.C. de Freitas, S. Kaneko, Ethanol demand under the flex-fuel technology regime in Brazil, Energy Econ. 33 (2011) 1146–1154, doi:http://dx.doi.org/ 10.1016/j.eneco.2011.03.011. [17] E.S. Lora, R.V. Andrade, Biomass as energy source in Brazil, Renew. Sustain. Energy Rev. 13 (2009) 777–788, doi:http://dx.doi.org/10.1016/j.rser.2007.12.004. [18] V. Subramani, S.K. Gangwal, A review of recent literature to search for an efficient catalytic process for the conversion of syngas to ethanol, Energy Fuels 22 (2008) 814–839, doi:http://dx.doi.org/10.1021/ef700411x. [19] V.R. Surisetty, A.K. Dalai, J. Kozinski, Alcohols as alternative fuels: an overview, Appl. Catal. A 404 (2011) 1–11. [20] K. Fang, D. Li, M. Lin, M. Xiang, W. Wei, Y. Sun, A short review of heterogeneous catalytic process for mixed alcohols synthesis via syngas, Catal. Today 147 (2009) 133–138, doi:http://dx.doi.org/10.1016/j.cattod.2009.01.038. [21] V.R. Surisetty, I. Eswaramoorthi, A.K. Dalai, Comparative study of higher alcohols synthesis over alumina and activated carbon-supported alkalimodified MoS2 catalysts promoted with group VIII metals, Fuel 96 (2012) 77– 84, doi:http://dx.doi.org/10.1016/j.fuel.2011.12.054. [22] R. Andersson, M. Boutonnet, S. Järås, Correlation patterns and effect of syngas conversion level for product selectivity to alcohols and hydrocarbons over molybdenum sulfide based catalysts, Appl. Catal. A 417–418 (2012) 119–128, doi:http://dx.doi.org/10.1016/j.apcata.2011.12.033. [23] J.M. Christensen, P.M. Mortensen, R. Trane, P.A. Jensen, A.D. Jensen, Effects of H2S and process conditions in the synthesis of mixed alcohols from syngas over alkali promoted cobalt-molybdenum sulfide, Appl. Catal. A 366 (2009) 29–43, doi:http://dx.doi.org/10.1016/j.apcata.2009.06.034. [24] J.M. Christensen, P.A. Jensen, A.D. Jensen, Effects of feed composition and feed impurities in the catalytic conversion of syngas to higher alcohols over alkalipromoted cobalt-molybdenum sulfide, Ind. Eng. Chem. Res. 50 (2011) 7949– 7963, doi:http://dx.doi.org/10.1021/ie200235e. [25] S. Zaman, K.J. Smith, A review of molybdenum catalysts for synthesis gas conversion to alcohols: catalysts, mechanisms and kinetics, Catal. Rev. 54 (2012) 41–132, doi:http://dx.doi.org/10.1080/01614940.2012.627224. [26] R. Andersson, M. Boutonnet, S. Järås, Effect of CO2 in the synthesis of mixed alcohols from syngas over a K/Ni/MoS2 catalyst, Fuel 107 (2013) 715–723, doi: http://dx.doi.org/10.1016/j.fuel.2012.11.044. [27] R. Andersson, M. Boutonnet, S. Järås, Higher alcohols from syngas using a K/Ni/ MoS2 catalyst: trace sulfur in the product and effect of H2S-containing feed, Fuel 115 (2014) 544–550, doi:http://dx.doi.org/10.1016/j.fuel.2013.07.057. [28] S.-W. Chiang, C.-C. Chang, J.-L. Shie, C.-Y. Chang, D.-R. Ji, J.-Y. Tseng, C.-F. Chang, Y.-H. Chen, Synthesis of alcohols and alkanes from CO and H2 over MoS2/
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36] [37] [38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47] [48]
[49]
2155
g-Al2O3 catalyst in a packed bed with continuous flow, Energies 5 (2012) 4147– 4164, doi:http://dx.doi.org/10.3390/en5104147. D. Ferrari, G. Budroni, L. Bisson, N.J. Rane, B.D. Dickie, J.H. Kang, S.J. Rozeveld, Effect of potassium addition method on MoS2 performance for the syngas to alcohol reaction, Appl. Catal. A: Gen. 462–463 (2013) 302–309. A. Andersen, S.M. Kathmann, M.A. Lilga, K.O. Albrecht, R.T. Hallen, D. Mei, Effects of potassium doping on CO hydrogenation over MoS2 catalysts: a firstprinciples investigation, Catal. Commun. 52 (2014) 92–97, doi:http://dx.doi. org/10.1016/j.catcom.2014.02.011. K. Atsonios, C. Christodoulou, Koytsoumpa, K.D. Panopoulos, E. Kakaras, Plant design aspects of catalytic biosyngas conversion to higher alcohols, Biomass Bioenergy 53 (2013) 54–64, doi:http://dx.doi.org/10.1016/j.biombioe.2013.01.035. J.-J. Wang, J.-R. Xie, Y.-H. Huang, B.-H. Chen, G.-D. Lin, H.-B. Zhang, An efficient Ni–Mo–K sulfide catalyst doped with CNTs for conversion of syngas to ethanol and higher alcohols, Appl. Catal. A: Gen. 468 (2013) 44–51, doi:http://dx.doi. org/10.1016/j.apcata.2013.08.026. M.J. Menart, J.E. Hensley, K.E. Costelow, Thermal decomposition of bulk K– CoMoSx mixed alcohol catalyst precursors and effects on catalyst morphology and performance, Appl. Catal. A: Gen. 437–438 (2012) 36–43. J.E. Hensley, S. Pylypenko, D.A. Ruddy, Deactivation and stability of K–CoMoSx mixed alcohol synthesis catalysts, J. Catal. 309 (2014) 199–208, doi:http://dx. doi.org/10.1016/j.jcat.2013.10.001. J.J. Spivey, A. Egbebi, Heterogeneous catalytic synthesis of ethanol from biomass-derived syngas, Chem. Soc. Rev. 36 (2007) 1514–1528, doi:http://dx. doi.org/10.1039/b414039g. 17660882. M.A. Gerber, J.F. White, D.J. Stevens, Mixed Alcohol Synthesis Catalyst Screening, Pacific Northwest National Laboratory, Richland, 2007. M.M. Conway, C.B. Murchinson, R.R. Stevens, Method for adjusting methanol to higher alcohol ratios, USPO N US4675344 A, 06/23/1987. Z. Liu, X. Li, M.R. Close, E.L. Kugler, J.L. Petersen, D.B. Dadyburjor, Screening of alkali-promoted vapor-phase-synthesized molybdenum sulfide catalysts for the production of alcohols from synthesis gas, Ind. Eng. Chem. Res. 36 (1997) 3085–3093, doi:http://dx.doi.org/10.1021/ie960648d. H.C. Woo, I.-S. Nam, J.S. Lee, J.S. Chung, K.H. Lee, Y.G. Kim, Room-temperature oxidation of K2CO3/MoS2 catalysts and its effects on alcohol synthesis from CO and H2, J. Catal. 138 (1992) 525–535, doi:http://dx.doi.org/10.1016/0021-9517 (92)90304-Z. D. Duphil, S. Bastide, C. Lévy-Clémént, Chemical synthesis of molybdenum disulfide nanoparticles in an organic solution, J. Mater. Chem. 12 (2002) 2430– 2432, doi:http://dx.doi.org/10.1039/b202162e. Y. Yoneyama, C. Song, A new method for preparing highly active unsupported Mo sulfide. Catalytic activity for hydrogenolysis of 4-(1-naphthylmethyl) bibenzyl, Catal. Today 50 (1999) 19–27, doi:http://dx.doi.org/10.1016/S09205861(98)00477-5. J.K. van der Waal, G. Klaus, M. Smit, C.M. Lok, High-throughput experimentation in syngas based research, Catal. Today 171 (2011) 207–210, doi:http://dx. doi.org/10.1016/j.cattod.2011.02.019. J.A. Wilson, A.D. Yoffe, The transition metal dichalcogenides discussion and interpretation of the observed optical, electrical and structural properties, Adv. Phys. 18 (1969) 193–335, doi:http://dx.doi.org/10.1080/00018736900101307. K. Dartigeas, L. Benoist, D. Gonbeau, G. Pfister-Guillouzo, G. Ouvrard, A. Levasseur, Systematic X-ray photoelectron spectroscopy and theoretical studies of disulfides of groups IVB, VB and VIB transition metals, J. Electron Spectrosc. Relat. Phenom. 83 (1997) 45–58, doi:http://dx.doi.org/10.1016/ S0368-2048(96)03076-9. E.G. Derouane, E. Pedersen, B.S. Clausen, Z. Gabelica, H. Topsøe, EPR studies on the influence of NO adsorption on sulfided Co–Mo hydrodesulfurization catalysts, J. Catal. 107 (1987) 587–588, doi:http://dx.doi.org/10.1016/00219517(87)90325-3. P. Afanasiev, Synthetic approaches to the molybdenum sulfide materials, Compt. Rend. Chim. 11 (2008) 159–182, doi:http://dx.doi.org/10.1016/j. crci.2007.04.009. H. Topsøe, B.S. Clausen, F.E. Massoth, Hydrotreating catalysis, Catal. Sci. Technol. 11 (1996) 1–269. H.C. Woo, J.C. Kim, I.-S. Nam, J.S. Lee, J.S. Chung, Y.G. Kim, Surface species on the oxidized K2CO3/MoS2 and their effects on catalytic carbon monoxide hydrogenation, Appl. Catal. A 104 (1993) 199–214, doi:http://dx.doi.org/ 10.1016/0926-860X(93)85098-A. N. Koizumi, K. Murai, T. Ozaki, M. Yamada, Development of sulfur tolerant catalysts for the synthesis of high quality transportation fuels, Catal. Today 89 (2004) 465–478, doi:http://dx.doi.org/10.1016/j.cattod.2004.02.002.