A Comparative Study of Solvothermal and Sol-Gel-Derived Nanocrystalline Alumina Catalysts for Ethanol Dehydration

The ethanol dehydration to ethylene over alumina catalysts prepared by solvothermal and sol-gel methods was investigated. Also, a commercial alumina was used for comparison purposes.The results showed that the catalytic activity depends on the properties of catalyst derived from different preparation methods and reaction temperature. The alumina synthesized by solvothermal method exhibited the highest activity. This can be attributed to the higher surface area and larger amount of acid site, especially the ratio of weak/strong acid strength as determined by N 2 physisorption and NH 3 -TPD studies. The solvothermal-derived catalyst exhibited an excellent performance with complete ethanol conversion and 100% selectivity to ethylene at 350C in comparison with other ones. In addition, we further studied the catalytic dehydration of alumina catalyst modified with Fe. The presence of 10wt.% Fe decreased both conversion and ethylene selectivity. However, the acetaldehyde selectivity apparently increased. It was related to the dehydrogenation pathway that takes place on Fe species.


Introduction
Nowadays, ethylene is one of the most important compounds for the petrochemical industry.It has been used as feedstock to produce many products such as polyethylene, polystyrene, vinyl chloride, and ethylene oxide, and, therefore, the global demand for ethylene is expected to increase with the continuous increasing of crude oil price [1,2].Ethylene is conventionally produced by thermal cracking of petroleum or natural gas.Since this process requires high temperature (750-900 ∘ C) and the natural resource is limited, much attention has been paid to find the alternative approach to produce ethylene.Recent studies have shown that one effective route by using biomass, especially bioethanol, is considered as the most promising way instead of using petroleum as feedstock.The ethanol dehydration provides many advantages such as lower operating temperature and green manufacturing of ethylene.Hence, the development of dehydration performance has been widely studied in both industry and academia [3][4][5][6][7].The catalysts most commonly used for ethanol to ethylene are based on zeolite [8,9], alumina [10,11], silica [12], and silica-alumina [13].For HZM-5 zeolites, these catalysts possess smaller pore size and have high acid strength.Their properties cause coke deposition resulting in the deactivation of catalyst, so zeolites are inappropriate for dehydration [14,15].Among them, alumina is found as an efficient catalyst because of its high specific surface area, excellent thermal stability, and wide range of chemical properties, especially lower acid strength than zeolites [15,16].It is thus the most common catalyst used to produce ethylene.
Although several studies have been investigated on the relationship between the catalyst structures and performances in ethanol dehydration, an understanding on how the textural properties of catalyst would affect the catalytic activity has not yet been reached.Hosseini and Nikou [17] synthesized -alumina by coprecipitation method with four precipitants and used it as a catalyst for methanol dehydration.The results showed that catalyst prepared by ammonium carbonate showed higher catalytic activity compared to commercial catalyst despite having lower amount of acid sites.The high activity is related to the high surface area and pore volume of catalyst.Akarmazyan et al. [18] 2. Experimental 2.1.Catalyst Preparation.Alumina catalysts were synthesized according to the procedure reported previously [24,25] and is described below.

Solvothermal
Method.An amount of 25 g of aluminium isopropoxide (AIP) as a starting material was suspended in 100 mL of 1-butanol in a test tube and then placed in an autoclave.In the gap between the autoclave wall and test tube, 30 mL of 1-butanol was added.The autoclave was purged with nitrogen gas before heating up to 300 ∘ C and held at that temperature for 2 h.After the autoclave was cooled down to room temperature, the obtained powder was washed with methanol followed by centrifugation.Finally, the white powder product was dried at 120 ∘ C for 24 h and calcined in air at 600 ∘ C for 6 h.

Sol-Gel Method.
The same volume of ethanol and deionized water was added to the flask.The solution was stirred under 20 rpm/min and heated to 80 ∘ C.After the solution was kept for 1 h, 15 g aluminum isopropoxide used as a precursor was added to the solution under continuous stirring.The hydrolysis step was carried out at 80 ∘ C by adding a certain volume of 10 M HCl (37%).The mixture was stirred at that temperature for 8 h.The obtained product was dried at 115 ∘ C for 24 h and further calcined in air at 550 ∘ C for 6 h to obtain white alumina catalyst.

Ethanol Dehydration
Reaction.Dehydration of ethanol to ethylene was conducted in a glass fixed-bed reactor (length = 0.33 m and inner diameter = 7 mm).A glass reactor was placed into a temperature-programmed tubular furnace.All experiments were performed under atmospheric pressure and in the temperature range 200 to 400 ∘ C using a feed composition consisting of 99.95% ethanol.In a typical run, 0.05 g of alumina catalysts obtained from Sections 2.1.1 and 2.1.2and commercial alumina (Fluka) was charged into the middle zone of reactor tube and pure ethanol as feed was stored in vaporizer.Prior to testing, the catalysts were activated at 200 ∘ C for 1 h under an argon gas flow.Ethanol was conveyed into reactor at an argon gas flow rate of 50 mL/min.The reaction was carried out at each temperature for 1 h.The dehydration products were analyzed by a gas chromatography (Shimadzu GC-14A) with a FID detector using DB-5 capillary column.

X-Ray Diffraction (XRD).
XRD was used to study the bulk crystalline phase of samples.The XRD pattern was performed by SIEMENS D5000 X-ray diffractometer with CuK  radiation.Those samples were scanned in the range of 2 = 10-90.

N 2 -Physisorption.
The surface area and average pore volume of prepared catalysts were determined by Micromeritics ChemiSorb 2750 Pulse instrument.Measurement was performed at −196 ∘ C and calculated according to the BET isotherm equation.

Scanning Electron Microscopy (SEM) and Energy Dispersive X-Ray Spectroscopy (EDX).
Hitachi S-3400N model was used to determine the catalyst morphology.Their elemental distribution and composition over different catalysts were performed by Link Isis Series 300 program EDX.

Ammonia Temperature-Programmed Desorption (NH 3 -TPD).
The acid properties of all catalysts were determined by using Micromeritics ChemiSorb 2750 Pulse Chemisorption System.Under a helium gas flow, the Fe-modified and unmodified catalysts were pretreated at 550 ∘ C for 1 h.Then, the catalyst was saturated with 15% NH 3 /He at 40 ∘ C for 30 min.After chemisorption step, a helium gas was flown over catalyst to remove any adsorbed molecules from catalyst active site from temperature of 40 to 500 ∘ C at heating rate  10 ∘ C/min.The NH 3 TPD profile is a plot of the TCD signal as a function of temperature to report the amount of ammonia, which is related to catalyst acidity.
The samples of 10-20 mg and a temperature range between 30 and 400 ∘ C at 2 ∘ C min −1 were used in the operation with N 2 UHP carrier gas.

Catalyst Characterization.
The XRD patterns of all samples are shown in Figure 1.It can be seen that the characteristic peaks at 2 = 37, 46, 61, and 67 indicating typically -phase of alumina [18,33].The XRD pattern of alumina synthesized by solvothermal (Al-SV) method was broad, because the crystallite sizes were very small, while the catalyst prepared by sol-gel method (Al-SG) and commercial alumina (Al-com) exhibited sharp peaks due to the presence of large crystallite size.The average crystallite size of the catalysts was calculated using Scherrer equation and shown in Table 1.The average crystallite size of alumina prepared by solvothermal method (Al-SV) was the smallest.
Figure 2 shows the nitrogen adsorption/desorption isotherms of alumina catalysts.All samples exhibited type IV isotherms, indicating they are mesoporous materials.For Al-SV, the sample presented hysteresis loop of type H1, occurring at higher relative pressure (/ 0 = 0.7 to 0.9) compared with other catalysts.It indicates larger mesopores and broad pore size distribution with uniform cylindrical shapes.Meanwhile, the isotherm for sol-gel-derived catalyst showed quite different pattern.Two inflection points and hysteresis loop moved toward lower pressure were observed.This suggests that lower porosity and cylindrical mesopores were expected for Al-SG.For Al-com, the isotherm was similar to that for Al-SG, but with slightly decreased BET surface area and pore volume [18,34].The observations are consistent with the value of surface area and sample porosity as shown in Table 1, in which the pore volume of Al-SV, Al-SG, and Al-com was 0.770, 0.221, and 0.212 cm 3 /g, respectively.The surface area of Al-SV was the highest, while the Al-com had the lowest surface area.Also, it was in accordance with those obtained from XRD analysis, where the samples with smaller crystallites exhibit higher surface area.The catalyst with small crystallite size may possess higher catalytic activity, because more active sites on large external surface area are easily reacted with reactants [18].
The pore size distribution (PSD) calculated by BJH method is shown in Figure 3.The Al-SV sample exhibited a bimodal pore size distribution, where pore diameter was in the ranges of 7-9 nm and 23-25 nm.For Al-SG and Al-com catalysts, the samples have a narrow pore size distribution with an average pore diameter around 3 nm, confirming the pore size distribution in the mesoporous range.It can be seen that the calculated pore size distribution was in good agreement with N 2 isotherm as mentioned above.This finding suggests the difference in porous nature greatly affected by preparation method (sol-gel and solvothermal).
To examine the catalyst morphology, SEM technique was performed.The SEM micrographs of different catalysts are shown in Figure 4.For those catalysts synthesized by solvothermal method, agglomeration of primary particles exhibited the porous shape, where the Al-SG and Al-com catalysts consisted of agglomerated particles with primarily irregular shape.The proposed different morphology of alumina obtained from the solvothermal and sol-gel methods is illustrated in Scheme 1.
Considering the distribution of surface acidity and the strength of acid sites in different alumina catalysts, NH 3 -TPD was examined.Although NH 3 -TPD technique has some drawbacks [1], it is fast, simple, and frequently employed method to evaluate the catalyst acidity.The typical NH 3 -TPD profile for catalysts is shown in Figure 5.It can be seen that there were two types of acidic sites for Al-SV, Al-SG, and Alcom.The first peak observed at lower temperature (≈250 ∘ C) was due to the desorption of ammonia chemisorbed at the weak to moderate acid sites, while the higher temperature peak centered at 400 ∘ C was related to the acid site with high strength [35].The amount of acid sites was determined from NH 3 -TPD curve by deconvolution according to the Gauss curve fitting method.The results are provided in Table 2, which demonstrated that the amount of strong acid site was higher than weak to moderate acid sites for all catalysts.The Al-SV has the highest amount of weak to moderate acid sites compared to Al-SG and Al-com.According to Chen et al. [5], higher ethylene selectivity of TiO 2 /-Al 2 O 3 catalysts is consistent with the higher acidity.Thus, the difference in catalytic performance of these catalysts depending on the acid amount will be discussed further.

Catalytic Performance of Ethanol Dehydration
Effect of Reaction Temperatures.In order to study the effect of reaction temperatures on the conversion of ethanol and selectivities to ethylene, diethyl ether, and acetaldehyde, the dehydration was examined over the Al-SV catalyst in the temperature range from 200 to 400 ∘ C. As known, the product distribution of ethanol dehydration strongly depends on  the reaction temperatures.Low temperatures favor the diethyl ether production, while the ethylene formation occurs at high temperature.It is evident from Figure 6 that the conversion increased with an increase in the reaction temperature.The Al-SV exhibited the highest conversion and ethylene selectivity of 100% at 350 ∘ C and kept constant at 400 ∘ C. The enhanced catalytic activity is related to the reaction pathway.Generally, for ethylene production through ethanol dehydration, there are two competitive pathways during reaction.The main path involves the formation of ethylene.This reaction occurring via intramolecular is endothermic.The second one, inter-molecular dehydration to diethyl ether, is exothermic [5,15,26].The results of activity test in Figure 6 show that when increasing the reaction temperature, the selectivity to ethylene continuously increased, whereas the decrease in diethyl ether selectivity was evident over selected catalyst.Chen et al. [5] suggested that at low temperature, not only do the catalysts have poor activities, but also the selectivity of ethylene is low due to a large amount of ethanol that was converted to diethyl ether product.In this research, the reaction temperature of 350 ∘ C was selected as the optimum temperature.
Effect of Preparation Methods.In order to investigate the influence of preparation methods on ethanol dehydration, alumina synthesized by solvothermal and sol-gel methods as well as commercial catalysts was studied.The conversion of ethanol over all catalysts is presented in Figure 7.The conversion was found to be in the order Al-SV > Al-SG > Al-com.The Al-SV showed the best performance for dehydration, giving complete conversion at 350 ∘ C compared to the other ones.From results of NH 3 -TPD analysis, it seems that there is a relationship between the activity and the ratio of weak acid sites to strong acid sites of alumina catalysts.
The sample possessing high ratio of weak to strong acid sites exhibited high catalytic performance on dehydration.The higher activity of Al-SV can be attributed to its higher surface area and higher amount surface acid sites, especially the highest ratio of weak/strong acid strength.Similar trend was found by Hosseini and Nikou [17]; catalyst with high surface area exhibited high activity due to more active sites on large surface area directly to be exposed to ethanol reactant.Xin et al. [1] also reported that the ethanol conversion to ethylene is mainly related to the weak to moderate acid site, while the side reaction such as oligomerization, dehydrocyclization, and the reaction producing higher olefin compounds occurred on the strong acid strength.
In addition, it is generally accepted that ethanol dehydration is an acid-catalyzed reaction.Either Brönsted acid site or Lewis acid site is believed to play an important role in catalytic activity.Different reaction mechanisms of ethanol to ethylene involved the acid sites.One ethanol molecule initially adsorbs on the Brönsted acid site of catalyst and then forms an ethoxide surface species.Subsequently, the ethoxide undergoes deprotonation, losing a proton to Brönsted acid site and forming the ethylene product.On the other hand, the reaction between ethoxide and another ethanol molecule yields diethyl ether (side product).These observations deduced the ethylene selectivity mainly contributed to the Brönsted acidic site [1,36,37].In order to support active Brönsted acidic sites, the dehydration of ethanol over pure siliceous silicate-1 was investigated [1].They reported that a very low activity with ethanol conversion around 2% was observed.Since the sample consists mainly of Si-OH, there is not Al atom in framework.Therefore, it clearly suggested that the Brönsted acid sites play a crucial role in the ethanol conversion and the formation of ethylene.However, Pan et al. [38] observed the opposite trend with the Lewis acid site.They inferred that the near absence of the acidic Brönsted and the presence of strong Lewis site apparently enhanced the ethylene selectivity.In fact, it should be noted that the NH 3 -TPD does not provide the nature of surface Brönsted  and Lewis acid site on catalyst surface.Thus, in this study, the types of acidic site (Brönsted or Lewis sites) are not further discussed here.However, it is generally accepted that the ethanol conversion to ethylene occurs on weak acid sites, while the oligomerization and the alcohol transformation to higher hydrocarbon correspond to strong acid strength [39,40].
From the product selectivity presented in Figures 8, 9, and 10, it was obvious that the Al-SV was able to dehydrate ethanol to ethylene with 100% selectivity, while at 350 ∘ C ethylene selectivity was only 87% and 65% for Al-SG and Al-com, respectively.Table 3 summarizes the catalytic ability  for ethanol dehydration to ethylene over various catalysts reported so far.It was obvious that Al-SV is comparable to those of typical and modified catalysts.Thus, the Al-SV is expected to be applied for the ethanol dehydration to ethylene.
To study the amount of coke deposition on catalyst after reaction, TGA measurement was performed.As seen in Figure 11, the weight loss below 150 ∘ C was attributed to the removal of physically adsorbed water.The weight loss at higher temperature (200-800 ∘ C) was attributed to the burning of coke deposited on the used sample surface.It was observed that the amount of coke present on various catalysts was in the order of Al-SV (4.0 wt.%) > Al-SG (3.9 wt.%) > Al-com (2.6 wt.%), corresponding to the amount of acidity as determined by NH 3 -TPD analysis.Accordingly, the textural and acidic properties, the amount of acidic sites, and distribution of acid strength as well as the reaction temperature are important parameters that influenced the catalytic activity.It is thus recommended that alumina catalyst prepared by solvothermal method (Al-SV) was the most suitable catalyst for using in the production of ethylene from ethanol and 350 ∘ C was selected as the optimum temperature.
In addition, we extend our study on how the Fe-modified alumina catalyst affects the catalytic activity for ethanol dehydration.We selected solvothermal alumina impregnated with Fe as a catalyst and denoted it as Fe/Al-SV.
The XRD pattern exhibited the presence of Fe characterized by peaks at 24 ∘ , 33 ∘ , 35.5 ∘ , 49 ∘ , 54 ∘ , and 62 ∘ .The average crystallite size calculated by Scherrer equation was 4.43 nm.The crystallite size increased after modified with Fe.For the BET surface areas, it can be observed that the addition of Fe leads to a decrease in the surface area.The result changes from 215 to 145 m 2 /g, which is due to the pore blockage with Fe.It is consistent with pore size distribution results that it showed a slight shift of pore diameter to small size obtained for Fe/Al-SV.
For NH 3 -TPD results, after the introduction of Fe onto alumina catalyst, the first peak of NH 3 -TPD slightly shifted to higher temperature, while the second peak became broader.It suggested that the weak to moderate acid sites were decreased, but the amount of strong acid strength was increased.The results were probably due to the changes made by alternation of some strong acidic sites to the weak to moderate ones in modified catalysts [20].A similar result was carried out by Li et al. [41], which reported that the presence of Mn 4+ in molecular sieve catalyst caused the higher strong acidity, but lower weak acid site.
Table 4 displays the catalytic dehydration of Al-SV and Fe/Al-SV catalysts under the optimized reaction temperature ( = 350 ∘ C).It is evident that the activity of Al-SV is higher than Fe-modified catalyst, which is in good agreement with the results of N 2 adsorption-desorption and NH 3 -TPD techniques.The Al-SV has high surface area and possesses large amount of weak to moderate acid sites.This shows the importance of textural property and distribution of acid sites as well as the amount of acid site in catalysts.Interestingly, the Fe/Al-SV showed higher selectivity to acetaldehyde compared to unmodified one.An improvement in acetaldehyde selectivity is related to the dehydrogenation pathway promoted by Fe.According to the work done by Li et al. [41], they found that the formation of acetaldehyde over MnO  /molecular sieve catalyst involves (i) the adsorption of ethanol on manganese cations as Lewis acid sites forming ethoxide species; (ii) hydrogen atom of ethoxide abstraction by lattice oxygen and forming acetaldehyde product; (iii) a reoxidation process at metal site and the replenishment of lattice oxygen vacancies to complete catalytic cycle.Also, Michorczyk et al. [42] revealed that the dehydrogenation of propane proceeded via oxidative pathway over catalyst.
Propane is oxidized to propene by lattice oxygen from the iron oxide phase (Fe 2 O 3 or Fe 3 O 4 ).However, in the oxygenfree experiment, acetaldehyde was found to a small extent because of the limited supply of oxygen.Based on these results and previous studies, a plausible reaction for the formation of acetaldehyde is proposed in Scheme 2.
Although the addition of Fe decreased the ethylene selectivity, the enhanced selectivity to acetaldehyde, which is a valuable compound in petrochemical industry, was obtained.The results in this research suggested a new chance to produce acetaldehyde.The ethanol reaction on alumina catalyst with and without Fe modification is demonstrated in Scheme 3.

Conclusion
The catalytic performance for ethanol dehydration over alumina catalysts prepared by solvothermal and sol-gel methods as well as commercial catalyst was investigated.It revealed that solvothermal-derived alumina showed the highest catalytic activity among other ones because it has the highest surface area and highest ratio of weak acid sites to strong acid sites.From N 2 adsorption-desorption and NH 3 -TPD results, it can be concluded that these textural and acidic properties significantly affect catalytic dehydration.In addition, in the part of Fe modification, the Fe loading was found to improve the acetaldehyde selectivity.It is due to the ethanol dehydrogenation over Fe species.

Figure 6 :
Scheme 1: Proposed different morphology of alumina obtained from the solvothermal and sol-gel methods.

Figure 11 :
Figure 11: TGA curves of the different alumina catalysts.

Table 1 :
Physical properties of alumina catalysts prepared by different methods.
a Crystallite size of alumina: determined by XRD results using Scherrer equation.

Table 3 :
Summary of catalysts for ethylene synthesis and their catalytic ability.

Table 4 :
NH 3 -TPD data and catalytic activity of Fe-modified alumina catalyst.