Alcohol promoted methanol synthesis enhanced by adsorption of water and dual catalysts

This is a version of a publication in Please cite the publication as follows: DOI: Copyright of the original publication: This is a parallel published version of an original publication. This version can differ from the original published article. published by Alcohol promoted methanol synthesis enhanced by adsorption of water and dual catalysts Nieminen Harri, Givirovskiy Georgy, Laari Arto, Koiranen Tuomas Nieminen, H., Givirovskiy, G., Laari, A., Koiranen, T. (2018). Alcohol promoted methanol synthesis enhanced by adsorption of water and dual catalysts. Journal of CO2 Utilization, Vol. 24, pp. 180-189. DOI: 10.1016/j.jcou.2018.01.002 Final draft Elsevier Journal of CO2 Utilization 10.1016/j.jcou.2018.01.002 © 2018 Elsevier Ltd. 1 Alcohol promoted methanol synthesis enhanced by adsorption of water and dual catalysts 1 Harri Nieminen*, Georgy Givirovskiy, Arto Laari, Tuomas Koiranen 2 Lappeenranta University of Technology, Laboratory of Process and Product Development, P.O. Box 20, FI-53851 3 Lappeenranta, Finland 4 * Corresponding author Tel.: +358 40 7451800, E-mail address: [email protected] 5 Abstract 6 Alcohol-promoted methanol synthesis uses heterogeneous methanol synthesis catalysts in alcoholic solvents 7 where the alcohols act as a co-catalyst. In the presence of alcohol, the reaction proceeds through alcohol formate 8 ester as an intermediate, allowing methanol synthesis at lower temperatures than conventional gas-phase 9 synthesis. In the present work, alcohol-promoted CO2 hydrogenation to methanol was studied experimentally using 10 a Cu/ZnO catalyst with 1-butanol and 2-butanol as solvents. As water is known to inhibit methanol synthesis on 11 Cu/ZnO catalysts, the alcohol-promoted process was further developed by in-situ adsorption of water using a 3Å 12 molecular sieve. The methanol productivity significantly improved as a result of the lowered concentration of water. 13 The concentration of water was thus identified as a key factor affecting the overall methanol productivity. As the 14 alcohol-promoted methanol synthesis process is characterized by two separate reaction steps, the use of separate 15 catalysts optimized for each step offers an interesting approach for the development of this process. Such a dual16 catalysis concept was tested using a copper chromite catalyst together with Cu/ZnO. Promising results were 17 obtained, as methanol productivity increased with the addition of copper chromite. Catalyst characterization was 18 carried out using XRD and SEM-EDS and potential effects of observed changes in catalyst structure during reaction 19 are discussed. 20 Keywords 21 CO2 hydrogenation, methanol synthesis, Cu/ZnO, liquid-phase, alcohol promoted, dual catalysis, copper chromite, 22 molecular sieve 23 Conflicts of interest: none 24 25 2 1. Introduction 26 Development of efficient and flexible energy storage methods is critical for a global shift from a fossil fuels based 27 economy to a renewable energy based economy [1]. The use of surplus peak electricity generated from fluctuating 28 renewable energy sources, such as wind and solar energy, for the production of chemical compounds would enable 29 energy storage in a highly transportable form at high energy density. Generation of hydrogen by electrolysis of 30 water is the common starting point in chemical energy storage strategies [2]. However, due to the difficulties and 31 hazards associated with large-scale storage and transportation of gaseous hydrogen, further utilization of hydrogen 32 for production of carbon-containing liquid fuels and chemical compounds might be preferable. 33 Methanol is an example of such a potential liquid-phase chemical energy carrier [3]. Methanol is an important and 34 versatile industrial chemical that can also be used as a fuel in power generation and in internal combustion engines 35 and fuel cells [4]. Additionally, methanol is a versatile raw material for synthesis of a variety of chemical products. 36 For instance, methanol can be transformed into gasoline in the methanol-to-gasoline process (MTG) [5] or into 37 olefins in the methanol-to-olefins process (MTO) [6]. 38 Current production of methanol is based on catalytic conversion of synthesis gas generated from fossil sources, 39 commonly natural gas. The syngas is mainly composed of mixtures of hydrogen, carbon monoxide and carbon 40 dioxide. In conventional methanol synthesis, copper and zinc oxide (Cu/ZnO) catalysts are generally employed at 41 reaction temperatures of 200-300 °C and pressures of 50-100 bar [7]. 42 The methanol synthesis process can be described by the following three equilibrium reactions: 43 CO2+ 3 H2⇌ CH3OH + H2O Δ𝐻0= −49.8 kJ/mol (1) 44 CO + 2 H2⇌ CH3OH Δ𝐻0= −91.0 kJ/mol (2) 45 CO + H2O ⇌ CO2+ H2 Δ𝐻0=41.2 kJ/mol (3) 46 The exothermic reactions (1) and (2) represent, respectively, the hydrogenation of CO2 and CO to methanol. 47 Reaction (3), the water-gas shift (WGS) reaction, is relevant to methanol synthesis as the reaction is also activated 48 by the copper-based methanol synthesis catalysts [8]. As methanol synthesis is exothermic and results in a 49 reduction of molar volume, methanol synthesis is favored by low temperatures and high pressures. However, 50 3 temperatures above 200 °C are required for sufficiently high reaction rates, and thus the thermodynamic equilibrium 51 limits the methanol synthesis to low conversion levels. Hydrogenation of pure CO2 to methanol is also possible but 52 the equilibrium conversions are even lower than for CO. Figure 1 shows the calculated equilibrium conversion of 53 stoichiometric CO and CO2 feeds at different temperatures and pressure. The conversions are modelled by Soave54 Redlich-Kwong equations of state, which have been shown to accurately predict experimental results in methanol 55 synthesis [9]. However, the hydrogenation of CO2 on Cu/ZnO catalysts is highly selective to methanol, with other 56 thermodynamically more favorable products such as methane, ethers and ketones formed only in negligible 57 amounts [10]. 58 59 Figure 1. Effect of temperature and pressure on the equilibrium carbon conversion from stoichiometric 60 CO2:H2 (1:3) and CO:H2 (1:2) mixtures. Calculated with the predictive Soave-Redlich-Kwong 61 (PSRK) [11] equation of state in Aspen Plus. 62 To overcome the thermodynamic limitations in the gas-phase methanol process, liquid-phase synthesis processes 63 have been proposed as an alternative approach to enable lower reaction temperatures in syngas reactions. Early 64 developments utilized highly basic catalyst systems such as alkali alkoxides in combination with copper chromite 65 [12, 13, 14] or nickel-based catalysts [15, 16, 17]. Methanol synthesis from CO/H2 at temperatures as low as 100 66 °C and pressures between 30 and 65 bar were reported [18]. However, the basic catalysts are incompatible with 67 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 150 170 190 210 230 250 Conversion Temperature, °C CO₂, 40 bar CO₂, 100 bar CO, 40 bar CO, 100 bar 4 CO2 or water, the presence of which, even at trace amounts, leads to rapid catalyst deactivation [17]. A method 68 proposed by the Brookhaven National Laboratory (BNL) also utilized a highly basic system for the conversion of 69 CO to methanol at significantly low temperature and pressure [19]. Furthermore, liquid-phase methanol synthesis 70 from CO2-containing synthesis gas in inert hydrocarbon solvent has been demonstrated in the LPMeOH process 71 [20]. 72 CO2 has been identified as the main carbon source in methanol synthesis from syngas [21]. Hence, it may be 73 expected that methanol can also be produced by hydrogenation of pure CO2. Hydrogenation of CO2, captured from 74 point sources or even directly from the atmosphere, would then provide a sustainable source of carbon-based fuels 75 and chemicals while helping to reduce the atmospheric concentration of CO2 [22]. Some pilot-scale methanol 76 processes that can use CO2 as the starting material have been developed. These include the CAMERE process 77 [23], which combines the reverse water-gas-shift reaction and methanol synthesis from syngas, and the Matsui 78 Chemicals process [24], which directly converts CO2 to methanol. Additionally, Carbon Recycling International 79 established commercial methanol production from CO2 in 2011, and the Svartsengi plant is presently operating at 80 a capacity of above 5 million liters per year [25]. The process utilizes geothermal energy readily available in Iceland. 81 One possible way to influence the reaction kinetics and conditions is to change the reaction route that leads to the 82 formation of methanol. A novel alcohol-promoted liquid-phase methanol synthesis process first proposed by Fan et 83 al. [26] is based on the combination of a conventional Cu/ZnO catalyst and alcohol as a catalytic solvent. The 84 alcohol promotes methanol synthesis by altering the reaction route, allowing operation at lower temperatures. In 85 the presence of the alcohol, the reaction proceeds through the formate ester of the corresponding alcohol as an 86 intermediate. As a result, methanol can be produced from syngas at temperatures starting from 170 °C and 87 pressures in the range of 30 to 50 bar [27]. Importantly, the process does not employ basic catalysts sensitive to 88 deactivation by CO2, allowing direct conversion of CO2. The following reaction steps have been proposed for this 89 process [28], supported by subsequent in-situ IR observations [29]: 90 1. Hydrogenation of carbon dioxide into formic acid 91 CO2+ H2⇄ HCOOH (4) 92 93 5 2. Reaction of formic acid with ethanol, forming ethyl formate 94 95 HCOOH + C2H5OH ⇄ HCOOC2H5+ H2O (5) 96 97 3. Hydrogenation of ethyl formate, forming methanol and ethanol 98 HCOOC2H5+ 2 H2⇄ CH3OH + C2H5OH (6) 99 The net reaction is the hydrogenation of carbon dioxide to methanol (Eq. 1) with a standard reaction enthalpy of - 100 49.8 kJ/mol. Different alcohols have been shown to possess different promoting effect for methanol synthesis. 101 Tsubaki et al. [30] found linear alcohols to be more effective compared to their branched counterparts, with n102 butanol showing the best results. Zeng et al. [31] reported that the yield of both methanol and the corresponding 103 ester decreased with increasing carbon number of the 1-alcohols from ethanol to 1-hexanol. For alcohols with the 104 same carbon number but different structure, 2-alcohols were found to have higher activity, which was explained by 105 a combination of spatial and electronic effects. As a result, 2-propanol showed the highest promotional effect. Later, 106 2-butanol was reported as the most effective solvent for the continuous methanol synthesis in a semibatch reactor 107 [32]. 108 As the alcohol-promoted methanol synthesis process is characterized by two separate reaction steps, the utilization 109 of separate catalysts optimized for each reaction could be beneficial. Such dualor cascade catalytic systems have 110 been considered previously for methanol synthesis. Huff and Sanford [33] reported effective CO2 conversion to 111 methanol at 135 °C using a combination of homogeneous catalysts. Chen et al. [34] used heterogeneous catalysts 112 in 1,4-dioxane solvent: copper chromite for the hydrogenation of CO2 to formate and Cu/Mo2C for the formate 113 hydrogenolysis to methanol. This system was capable of methanol production at rates comparable to conventional 114 gas-phase synthesis at 135 °C and exhibited methanol selectivity above 75%. The methanol synthesis was 115 promoted by the addition of ethanol, with the reaction proceeding through ethyl formate, as reported in the alcohol116 promoted process. On the other hand, copper chromite is known to catalyze the hydrogenolysis of esters to 117 alcohols, i.e. the latter stage in the alcohol-promoted reaction route [35]. As such, copper chromite appears an 118 interesting component of a dual catalytic system for alcohol-promoted methanol synthesis. 119 6 In comparison to CO-containing syngas feed, CO2 hydrogenation to methanol is further complicated by the 120 increased formation rate of water. Water is formed as a byproduct in methanol synthesis, and in the absence of 121 CO, the water-gas shift reaction proceeds in the reverse direction, producing more water. The negative effect of 122 water on methanol synthesis on Cu/ZnO-based catalysts has been well documented [36]. This effect has been 123 explained as a combination of kinetic inhibition effects and structural catalyst deactivation. Water-derived hydroxyl 124 species can block the active sites on the catalyst, resulting in kinetic inhibition. The presence of water can also 125 accelerate the sintering of copper particles [37], resulting in decreased copper dispersion and catalyst deactivation. 126 Removal of methanol and water using membrane reactors [38, 39] and by condensation at high pressures [40] or 127 low temperatures [41] has been previously described for gas-phase methanol synthesis. Reactive distillation [42] 128 provides a further possible approach for continuous product removal, particularly in liquid-phase processes, and 129 has been proposed in literature for the methanol synthesis process [43] and for the Fischer-Tropsch process [44] 130 operating at similar conditions. In addition, selective removal of water by adsorption on zeolite molecular sieves has 131 also been suggested in sorption-enhanced methanol [45] and related dimethyl ether [46] synthesis operated in the 132 gas-phase. 133 In the present work, alcohol-promoted methanol synthesis was investigated experimentally using a commercial 134 Cu/ZnO-based methanol synthesis catalyst with 1-butanol and 2-butanol as the solvents. 2-butanol was selected 135 because of the previously reported high activity for methanol synthesis, and 1-butanol was considered interesting 136 because of the potentially simplified product separation due to the higher boiling point of the alcohol. As novel 137 developments, enhancement of the alcohol-promoted methanol synthesis by in-situ adsorption of water and by the 138 use of dual catalysts were studied. Water adsorption was carried out using a molecular sieve. Methanol synthesis 139 combined with water removal has previously been modelled based on 4Å molecular sieves [45], and the use of 4Å 140 molecular sieves has been modelled for a related dimethyl ether (DME) synthesis [46]. However, experimental work 141 of methanol synthesis promoted by water adsorption has not been published earlier to our knowledge. A dual 142 catalyst system comprising of a combination of Cu/ZnO and copper chromite catalysts was tested with the aim of 143 improving methanol productivity by influencing separately the formate formation and hydrogenolysis reaction steps. 144 7 2. Materials and methods 145 A Parr 4520 autoclave reactor with an inner volume of 450 ml was used for the reaction experiments. The reactor 146 was connected to a Parr 4848 control unit used to control the reaction temperature and mixing speed. A mixing 147 speed of 600 rpm was used in all experiments. Liquid samples from the reaction mixture were collected using a 148 water-cooled sample collection vessel, in which any vapors present in the sample were condensed prior to collecting 149 the sample. 150 Analysis grade 1-butanol and 2-butanol, were used as solvents. A commercial Cu/ZnO-based methanol synthesis 151 catalyst (Alfa Aesar, 65.5 % CuO, 24.7% ZnO, 10.1% Al2O3, 1.3% MgO) was used. The catalyst was ground and 152 sieved to 150-500 µm for each experiment. The 3Å molecular sieve (UOP, beads with diameter of 2 mm), was also 153 ground and sieved to 150-500 µm. An initial experiment with the unground molecular sieve was also performed. 154 The molecular sieve was activated by heating to 250 °C for at least 8 hours under air and subsequent cooling to 155 ambient temperature inside a desiccator prior to use. Powdered copper chromite (Sigma-Aldrich) was used in the 156 dual catalyst experiments. A mixed gas containing 75% hydrogen and 25% carbon dioxide was used as the reaction 157 feed gas, and a mixed gas containing 5% hydrogen in nitrogen was used for activation of the catalysts. A diagram 158 of the experimental setup is presented in Figure 2. 159 8 160 Figure 2. Experimental setup used in the reaction experiments. 161 The ground Cu/ZnO catalyst and the copper chromite catalyst were activated in-situ in the reactor vessel. Catalyst 162 activation was performed under 5 bar of the 5% H2/N2 mixed gas, with the gas inside the reactor replaced every 30 163 minutes. The temperature was 200 °C during the activation. Following catalyst activation, the reactor was cooled 164 and the catalysts were kept under the activation gas until the reaction experiment was executed. 200 ml of the 165 alcohol was quickly poured into the reactor, minimizing the contact time of the catalysts with air. The reactor was 166 purged with nitrogen and heated to the reaction temperature under N2. At the reaction temperature, an initial liquid 167 sample was collected and the reactor was pressurized with the feed gas (CO2:H2 = 1:3) to the set reaction pressure, 168 which was 60 bar unless otherwise noted. Constant pressure was maintained during the experiments by replacing 169 the consumed reaction gas with fresh gas. The total reaction time was 6 hours and liquid samples were collected 170 every 2 hours. 171 An Agilent Technologies 6890N gas chromatograph with a thermal conductivity detector was used for analysis of 172 the liquid samples. A polar Zebron ZB-WAXplus column was used for the 2-butanol samples. An isothermal method 173 with the column temperature at 70 °C and helium (1.1 ml/min) as a carrier gas was used. For the 1-butanol samples, 174 a non-polar HP-1ms column was used due to insufficient separation of butanal and methanol in the ZB-WAXplus 175 15 290 Figure 7. Effect of reaction gas partial pressure on methanol productivity with 10 g of Cu/ZnO catalyst in 2291 butanol at 180 °C. Feed gas (CO2:H2 = 1:3), reaction time 6 h. 292 3.3 Water removal by molecular sieve 293 Continuous removal of water from the reaction mixture was tested by addition of a zeolite molecular sieve. Molecular 294 sieves with a pore diameter of 3 Å can be used for the dehydration of alcohols because of their selective adsorption 295 of water [50]. The selective adsorption is based on size exclusion of molecules larger than water in the inner 296 microporous structure of the zeolite. 297 The limiting effect of water on the alcohol-promoted methanol synthesis process was first confirmed by performing 298 an experiment with approximately 1.4 mol/dm3 of water added to 2-butanol. This concentration is slightly above the 299 maximum concentration range of water found in the experiments (Figure 4). At 180 °C and 60 bar of total pressure, 300 the methanol production rate was approximately 74% lower than in the base experiment with no water added. The 301 concentration of water did not significantly increase during this experiment but rather remained relatively constant 302 at the apparent equilibrium level. 303 Next, the effect of in-situ adsorption of water by the addition of a 3Å molecular sieve was tested. The relative 304 amounts of the catalyst and the molecular sieve were varied, maintaining a total solids mass of 50 g. The results of 305 these experiments are presented in Figure 8. A base experiment with 20 g of catalyst and no molecular sieve is 306 also presented for comparison. 307 0 5 10 15 20 25 30 35 40 30 35 40 45 50 Methanol productivity, g / kg / h CO2+ H2partial pressure, bar 16 308 Figure 8. Effect of catalyst and molecular sieve mass on methanol and water formation in 2-butanol. 309 Temperature 180 °C, feed gas CO2:H2 = 1:3, total pressure 60 bar. 310 Compared to the base case with 20 g of Cu/ZnO catalyst and no molecular sieve, the addition of the unground 311 molecular sieve increased the methanol productivity from 8.2 g/kg/h to 11.2 g/kg/h. A more significant improvement 312 was found with the molecular sieve ground into 150-300 µm particle size range. Due to the clear effect of the particle 313 size, the adsorption of water appears to be significantly diffusion-limited for the unground molecular sieve. With 20 314 g of catalyst, the addition of 30 g of the ground molecular sieve increases the methanol productivity to 33.6 g/kg/h, 315 an increase of over 300% over the Cu/ZnO catalyst used without a molecular sieve. Keeping the total amount of 316 solids (catalyst + molecular sieve) at 50 g, the methanol productivity increased with increasing amounts of molecular 317 sieve. For instance, the productivity increased to 54.4 g/kg/h using 10 g of the catalyst and 40 g of the molecular 318 sieve. These results clearly show that the catalyst is most effectively utilized for methanol synthesis when larger 319 relative amounts of the molecular sieve to the catalyst are used. This observation can be explained by the increased 320 water adsorption capacity of the larger amount of the molecular sieve, leading to decreased concentrations of water, 321 as shown in Figure 8. 322 323 0 0.2 0.4 0.6 0.8 1 1.2 1.4 20 g Cu/ZnO 20 g Cu/ZnO, 20 g MS (unground) 25 g Cu/ZnO, 25g MS 20 g Cu/ZnO, 30 g MS 10 g Cu/ZnO, 40 g MS 0 10 20 30 40 50 60 70 Concentration, mol/dm3 Methanol productivity, g / kg / h Methanol productivity Methanol concentration Water concentration 17 3.4 Dual catalysts 324 To test the dual catalysis concept for alcohol-promoted methanol synthesis, copper chromite (CuCr) was used in 325 combination with the Cu/ZnO catalyst. The ratios of the two catalysts were varied: 20 g of the Cu/ZnO catalyst was 326 used with 10 g of CuCr, and vice versa. The experiments were carried out in 2-butanol at 180 °C and 60 bar of total 327 pressure, corresponding to a CO2 + H2 partial pressure of 50.1 bar. The results of these experiments are presented 328 in Figure 9. A base experiment with 20 g of Cu/ZnO catalyst and no copper chromite is also presented for 329 comparison. 330 331 Figure 9. Effect of different amounts of Cu/ZnO and copper chromite (CuCr) catalysts on the formation of 332 methanol and water in 2-butanol. Reaction time 6 hours. Temperature 180 °C, feed gas CO2:H2 = 333 1:3, total pressure 60. An experiment with 20 g of Cu/ZnO catalyst and no copper chromite is 334 included for comparison. 335 The addition of the copper chromite catalyst clearly increases the methanol productivity. Both the absolute methanol 336 production rate, as measured by the methanol end concentration, and the specific productivity of the catalyst 337 increase with addition of copper chromite. The increased productivity can be explained either by a synergistic effect 338 between the two catalysts or by higher methanol synthesis activity of CuCr compared to Cu/ZnO. However, a higher 339 intrinsic activity of copper chromite appears unlikely, as the activity of Cu/ZnO for methanol synthesis is well-known 340 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20 g Cu/ZnO 20 g Cu/ZnO, 10 g CuCr 10 g Cu/ZnO, 20 g CuCr Concentration, mol/dm3 Methanol productivity, g / kg / h Methanol productivity Methanol concentration Water concentration 18 and industrially applied. Fan et al. [26] also reported higher methanol yield and selectivity of Cu/ZnO compared to 341 CuCr in alcohol promoted methanol synthesis. Fan et al. also found similar CO selectivity, or RWGS activity, for 342 both of the catalysts. This is supported by the present results, as the concentration of water was not significantly 343 affected by the changed ratio of Cu/ZnO and CuCr (Figure 9, columns 2 and 3), supporting similar RWGS activity 344 of the two catalysts. The overall methanol selectivity appears to be higher with the combined catalysts, as the ratio 345 of methanol to water produced is increased compared to Cu/ZnO used alone. 346 3.5 Characterization of Cu/ZnO catalyst before and after reaction 347 The structural features of the Cu/ZnO catalyst before and after reaction were investigated by the means 348 of XRD and SEM-EDS in order to assess the catalyst stability. Figure 10 presents the X-ray 349 diffractograms of the catalyst as supplied in the calcined form, following reduction in 5% hydrogen, and 350 following use in alcohol-promoted methanol synthesis in 1-butanol at 180 °C. It is noted that the same 351 batch of catalyst was analyzed prior to reduction and following the reaction, while the reduced catalyst 352 was prepared and analyzed separately. 353 354 Figure 10. X-ray diffractograms of the unused Cu/ZnO catalyst (A), the reduced catalyst (B), and the 355 catalyst following methanol synthesis from CO2 and H2 (1:3) in 1-butanol at 180 °C (C). 356 19 The calcined catalyst is largely amorphous, showing a minor pattern corresponding to copper(II)oxide 357 (CuO) typical to Cu/ZnO catalysts [51]. The patterns are identified based on the PDF 4+ 2018 358 crystallography database. The reduced catalyst presents with a clearly defined pattern consistent with 359 crystalline, copper(I)oxide (Cu2O), and metallic copper. Weak crystalline features of zinc oxide are also 360 evident, consistent with previous studies [52]. As the reduction of copper proceeds stepwise from CuO 361 to Cu via Cu2O [53], the presence of Cu2O may imply incomplete reduction, possibly due to insufficient 362 reduction time or temperature. However, as the reduced catalyst sample was transferred and analyzed 363 in contact with air, re-oxidation of copper crystallites during this process cannot be ruled out. 364 Only metallic copper and zinc oxide is found present in the used catalyst. Cu/ZnO catalysts are known to 365 show dynamic structural changes depending on the oxidation potential of the gas phase [54, 55] and 366 ongoing reduction of the catalyst at the reaction conditions is possible. As the reduced and used catalyst 367 analyzed here are not from the same batch of ground and prepared catalyst, batch-to-batch variation 368 cannot be eliminated as a cause of the observed structural differences. 369 The peaks corresponding to zinc oxide are more clearly defined compared to the reduced catalyst, 370 potentially indicating continuing crystallization of ZnO at the reaction conditions. Lunkenbein et al. [56] 371 identified zinc oxide as the more dynamic phase compared to metallic copper under reaction conditions, 372 and found that crystallization of ZnO and the resulting loss of reactive Cu-ZnO interfaces is the main 373 mechanism of initial catalyst deactivation. The SEM-EDS elemental maps of copper and zinc presented 374 in Figure 11 indicate that such a process may have initiated in the catalyst used here. The unused 375 (calcined) catalyst shows a relatively homogeneous distribution of both copper and zinc. However, a 376 degree of segregation of these elements can be observed in the used catalyst, with the elemental map 377 showing distinct areas with high content of zinc (oxide) that are relatively poor in copper. 378 20 379 Figure 11. SEM-EDS elemental maps of copper and zinc in the unused Cu/ZnO catalyst (upper), 380 and the catalyst following methanol synthesis from CO2 and H2 (1:3) in 1-butanol at 180 381 °C (lower). Composition scales in weight percent. 382 383 Further insight is provided by the SEM images presented in Figure 12. Distinct crystals in the 384 micrometer dimension can be observed, identified as zinc oxide by the EDS analysis. No such features 385 were found in the unused catalyst. It is concluded that agglomeration and crystallization of zinc oxide 386 during reaction has occurred, acting as a potential deactivation mechanism for the catalyst. However, 387 as long-term stability tests were not performed here, the actual effect of these structural changes on the 388 activity of the catalyst cannot be discussed. 389 These observations can be compared to other findings discussed in literature. Previously, the stability 390 of Cu/ZnO catalyst in alcohol promoted methanol synthesis has been explored by Reubroycharoen et 391 al. [32] who found the performance stable during 40 hours of continuous methanol synthesis (at 170 392 °C), and by Jeong et al. [57] who found no decline in activity during 60 hours of reaction (150 °C). In 393 contrast to our results, Jeong et al. found no changes in the XRD profile of the catalyst before and after 394 reaction. Other than the lower reaction temperature, the differing findings might be explained by 395 different feed gas composition, as a CO-rich syngas was used in these studies opposed to the CO2:H2 396 21 mixture used here. Therefore, it is possible that the detected differences might be caused by the large 397 amount of water present in the reaction system in the present study. 398 399 Figure 12. SEM micrographs of the Cu/ZnO catalyst following methanol synthesis from CO2 and H2 400 (1:3) in 1-butanol at 180 °C. Zinc oxide crystals are highlighted. 401 4. Conclusions 402 Methanol synthesis from CO2 was studied in an alcohol-promoted liquid-phase process using conventional Cu/ZnO 403 and copper chromite as catalysts. 1-butanol and 2-butanol were found to act as catalytic solvents, allowing methanol 404 synthesis at lower temperatures than conventional gas-phase processes. Although it was not possible to determine 405 the exact reaction route, it is expected that the promoting effect of the alcohols is based on a reaction route 406 proceeding through the intermediate of formate ester of the alcohol. 407 The effect of continuous water removal using molecular sieve adsorption was explored. The addition of a 3Å 408 molecular sieve significantly enhanced methanol productivity. Grinding of the molecular sieve resulted in improved 409 results due to the shorter diffusion path compared to the granular material. The maximum methanol productivity of 410 22 54.4 g/kg/h was found when the maximum relative amount of the molecular sieve (40 g) to the catalyst (10 g) was 411 used. The final methanol concentration after 6 hours of reaction time reached 0.5 mol/dm3. The catalyst was most 412 effectively used for methanol synthesis when the amount of molecular sieve was maximized, which minimized the 413 concentration of water. The water concentration was found to significantly affect the rate of methanol synthesis. 414 The overall methanol production rate in this process appears to be limited by the concentration of water and its 415 effects on the catalyst surface. To prevent the negative effects of water, continuous water removal or development 416 of more water resistant catalysts is vital for further development of this process. Based on the results, the use of a 417 3Å molecular sieve for water removal appears a promising approach. 418 The methanol productivity obtained in the current research can be compared to results reported in other studies. 419 Yang et al. [49] found an even higher methanol productivity of up to 167 g/kg/h for alcohol-promoted methanol 420 synthesis at 170 °C and 50 bar using an optimized Cu/ZnO catalyst composition. The difference to the results 421 presented here can be explained mainly by the different feed gas composition in their experiments (CO/CO2/H2/Ar 422 = 32.4/5.1/59.5/3.9). For gas-phase CO2 hydrogenation to methanol, productivity values even up to 1200 g/kg/h 423 have been achieved [58]. However, these results were obtained at a relatively high temperature of 240 °C and at 424 high space velocities giving relatively low CO2 conversions. 425 Dual catalysis by the combination of Cu/ZnO with copper chromite was also studied in this work. A remarkable 426 increase in catalytic activity was found for the dual catalyst. When 20 g of copper chromite and 10 g of Cu/ZnO was 427 used, the productivity increased by 80% compared to the use of 20 g of the Cu/ZnO catalyst alone. A synergistic 428 effect between the two catalysts is suggested, which is possibly based on an increased formation rate of the formate 429 ester intermediate by the copper chromite catalyst. The two catalysts appeared to have similar reverse water-gas 430 shift activity, as the concentration of water did not change when the relative amounts of Cu/ZnO and copper chromite 431 were varied. 432 Structural changes in the catalyst during alcohol-promoted methanol synthesis were found by the means of XRD 433 and SEM-EDS investigations. EDS elemental analysis showed that segregation of copper and zinc oxide had taken 434 place, and both XRD analysis and SEM imaging provided evidence that crystallization of zinc oxide occurred. Such 435 phenomena has previously been identified as cause of catalyst deactivation due to the loss of reactive Cu-ZnO 436 interfaces [56]. However, comprehensive catalyst stability tests were not performed in the current study, and thus 437 23 the effect of the observed changes on catalytic activity cannot be determined conclusively. It is clear that stability 438 tests at different reaction temperatures and, importantly, at different feed gas compositions are necessary to further 439 characterize the alcohol-promoted methanol synthesis process. 440 Acknowledgements 441 The Authors are grateful for Finnish Academy of Science for “Microand millistructured reactors for catalytic 442 oxidation reactions” MICATOX project funding, number: 269896. Funding provided by the Lappeenranta 443 University of Technology Doctoral School is also gratefully acknowledged. 444 24 References 445 [1] R. Schlögl, "The solar refinery," in Chemical Energy Storage, Berlin/Boston, Walter de Gruyter GmbH, 2013, pp. 1-34. [2] F. Schüth, "Energy storage strategies," in Chemical Enery Storage, Berlin/Boston, Walter de Gruyter GmbH, 2013, pp. 35-48. [3] G. Olah, "Beyond oil and gas: The methanol economy," Angew. Chem. Int. Ed., vol. 44, pp. 2636-2639, 2005. doi:10.1002/anie.200462121 [4] H. Offermanns, L. Plass and M. Bertau, "From raw materials to methanol, chemicals, and fuels" in Methanol: The Basic Chemical and Energy Feedstock of the Future, Berlin Heidelberg, Springer-Verlag, 2014, pp. 1-7. doi:10.1007/978-3-642-39709-7 [5] L. Reichelt and F. Schmidt, "Methanol-to-gasoline process," in Methanol: The Basic Chemical and Energy Feedstock of the Future, Berlin Heidelberg, Springer-Verlag, 2015, pp. 440-453. doi:10.1007/978-3-642-39709-7 [6] F. Schmidt and C. Pätzold, "Methanol-to-olefins processes," in Methanol: The Basic Chemical and Energy Feedstock of the Future, Berlin Heidelberg, Springer-Verlag, 2015, pp. 454-472. doi:10.1007/978-3-642-39709-7 [7] J. Ott, V. Gronemann, F. Pontzen, E. Fiedler, G. Grossmann, Kersebohm, D.B., G. Weiss and C. Witte, "Methanol," in Ullmann's Encyclopedia of Industrial Chemistry, Weinheim, Wiley-VCH Verlag GmbH & Co. KGaA, 2012, pp. 1-27. doi:10.1002/14356007 [8] S. Lee, "Methanol synthesis from syngas," in Handbook of Alternative Fuel Technology, Taylor & Francis Group, LLC, 2007, pp. 297-321. [9] J. Skrzypek, M. Lachowska, M. Grzesik, J. Sloczynski and P. Nowak, "Thermodynamics and kinetics of low pressure methanol synthesis," Chem. Eng. J., vol. 58, pp. 101-108, 1995. doi:10.1016/0923-0467(94)02955-5 [10] E. Kunkes and M. Behrens, "Introduction to methanol synthesis and steam reforming," in Chemical Energy Storage, Berlin, Walter de Gruyter GmbH, 2013, pp. 415-417. [11] T. Holderbaum and J. Gmehling, "PSRK: A group contribution equation of state based on UNIFAC," Fluid Phase Equilib., vol. 70, pp. 251-265, 1991. doi:10.1016/0378-3812(91)85038-V [12] V. Palekar, "Alkali compounds and copper chromite as low-temperature slurry phase methanol catalysts," Appl. Catal., A, vol. 103, no. 1, pp. 105-122, 1993. doi:10.1016/0926860X(93)85177-Q [13] Y. Zhao, L. Bai, Y. Hu, B. Zhong and S. Peng, "Catalytic performance of CuCr/CH3ONa used for low temperature methanol synthesis in slurry phase," J. Nat. Gas Chem., vol. 8, no. 3, pp. 181-187, 1999. [14] W. Chu, T. Zhang, C. He and Y. Wu, "Low-temperature methanol synthesis (LTMS) in liquid phase on novel copper-based catalysts," Catal. Lett., vol. 79, pp. 129-132, 2002. doi:10.1023/A:1015384015528 [15] M. Marchionna, L. Basini, A. L. M. Aragno and F. Ancillotti, "Mechanistic studies on the homogeneous nickel-catalyzed low temperature methanol synthesis," J. Mol. Catal., vol. 75, no. 2, pp. 147-151, 1992. doi:10.1016/0304-5102(92)80116-X [16] E. Lee and K. Aika, "Low-temperature methanol synthesis in liquid-phase with a Raney Nickel– alkoxide system: Effect of Raney Nickel pretreatment and reaction conditions," J. Mol. Catal. A: Chem., vol. 141, pp. 241-248, 1999. doi:10.1016/S1381-1169(98)00267-2 [17] S. Ohyama, "Low-temperature methanol synthesis in catalytic systems composed of nickel compounds and alkali alkoxides in liquid phases," Appl. Catal. A, vol. 180, pp. 217-225, 1999. doi:10.1016/S0926-860X(98)00338-X