Catalytic Ethylene Oligomerization over Ni/Al-HMS: A Key Step in Conversion of Bio-Ethanol to Higher Olefins

Al-modified hexagonal mesoporous silica (HMS) materials were synthesized using dodecylamine as a template according to the methods reported in the literature. FT-IR spectra proved that Al3+ ions entered in the HMS framework in Al-HMS (prepared by sol-gel reaction) but Al3+ ions existed in the extra-framework in Al/HMS (prepared by post-modification). NH3-TPD indicated that either Al-HMS or Al/HMS had solid acid sites on the surface, and the acidic strength of Al/HMS was stronger than that of Al-HMS. For ethylene oligomerization at 200 ◦C under 1 MPa, Ni/Al-HMS showed an ethylene conversion of 96.3%, which was much higher than that over Ni/Al/HMS (45.6%). The selectivity for C4H8, C6H12, C8H16, and C8+ was 37.7%, 24.5%, 24.0%, and 9.1% for ethylene oligomerization over Ni/Al-HMS, respectively. Ni/Al-MCM-41, which has been reported as an effective catalyst for ethylene oligomerization in the literature, showed a high ethylene conversion (95.2%) similar to that of Ni/Al-HMS in this study. However, the selectivity for C8H16 over Ni/Al-MCM-41 (16.3%) was lower than that over Ni/Al-HMS (24.0%) in the ethylene oligomerization. For ethanol dehydration at 300 ◦C under 1 MPa, a commercial H-ZSM-5 catalyst showed a high ethylene yield (91.2%) after reaction for 24 h using a feed containing 90 wt.% ethanol and 10 wt.% water. In this study, a two-step process containing two fixed-bed reactors and one cold trap was designed to achieve the direct synthesis of higher olefins from bio-ethanol. The cold trap was used to collect the water formed from ethanol dehydration. By using H-ZSM-5 as a catalyst for ethanol dehydration in the first reactor and using Ni/Al-HMS as a catalyst for ethylene oligomerization in the second reactor, higher olefins were continuously formed by feeding a mixture containing 90 wt.% ethanol and 10 wt.% water. The yields of higher olefins did not decrease after reaction for 8 h in the two-step reaction system.


Introduction
The conversion of biomass to liquid fuels has become an important research field because biomass utilization has an effect in reducing greenhouse gas emission [1]. Bio-ethanol is produced in a large scale from sugary, starchy, and lignocellulosic biomasses by yeast fermentation [2]. Currently, bio-ethanol accounts for above 90% of worldwide biofuel production. Bio-ethanol is mainly used as fuel for automobiles by blending with gasoline. Although bio-ethanol is an attractive fuel; the needs of transportation fuel for automobiles will be reduced from now due to the popularization of electric vehicle and hydrogen-fueled car. Hence, bio-ethanol is expected to be in surplus in future years. It is necessary to develop new technologies to utilize the surplus bio-ethanol. Production of hydrogen by the steam reforming of bio-ethanol is an available technology [3]. However, hydrogen is expected  Figure 2 shows the FT-IR spectra of various HMS-based samples after calcining at 550 °C for 4 h. In the FT-IR spectrum of HMS, the peak at 1090 cm −1 was assigned to the asymmetric stretching of Si-O-Si and the peak at 810 cm −1 was assigned to the stretching of the tetrahedral SiO4 4− structural units. HMS did not show a peak at 960 cm −1 in the FT-IR spectrum. Al-HMS showed a peak at 960 cm −1 in the FT-IR spectrum. Because the peak at 960 cm −1 was assigned to the stretching Si-O vibration with the neighborhood metal ions, this peak had been used to characterize the incorporation of metal ions in the silica framework [39]. The Al/HMS sample prepared by post-modification did not show a peak at 960 cm −1 in the FT-IR spectrum. Hence, Al 3+ ions entered into the HMS framework in Al-HMS but existed at the outside of the HMS framework in Al/HMS.   In the FT-IR spectrum of HMS, the peak at 1090 cm −1 was assigned to the asymmetric stretching of Si-O-Si and the peak at 810 cm −1 was assigned to the stretching of the tetrahedral SiO 4 4− structural units. HMS did not show a peak at 960 cm −1 in the FT-IR spectrum. Al-HMS showed a peak at 960 cm −1 in the FT-IR spectrum. Because the peak at 960 cm −1 was assigned to the stretching Si-O vibration with the neighborhood metal ions, this peak had been used to characterize the incorporation of metal ions in the silica framework [39]. The Al/HMS sample prepared by post-modification did not show a peak at 960 cm −1 in the FT-IR spectrum. Hence, Al 3+ ions entered into the HMS framework in Al-HMS but existed at the outside of the HMS framework in Al/HMS. corresponding to the (100) plane at about 2.5 degrees and a broad band close to 5.3 degrees. This pattern implied the existence of wormhole mesoporous structure in HMS [16]. By calculating from the degree of (100) plane in the XRD pattern, HMS had a d100 spacing value of 35.3 Å. The value of d100 spacing in Al-HMS was lower than that of HMS because small Al 3+ ions were incorporated in the HMS framework. On the other hand, the value of d100 spacing in Al/HMS was similar to that of HMS because Al 3+ ions did not enter in the HMS framework.  Figure 2 shows the FT-IR spectra of various HMS-based samples after calcining at 550 °C for 4 h. In the FT-IR spectrum of HMS, the peak at 1090 cm −1 was assigned to the asymmetric stretching of Si-O-Si and the peak at 810 cm −1 was assigned to the stretching of the tetrahedral SiO4 4− structural units. HMS did not show a peak at 960 cm −1 in the FT-IR spectrum. Al-HMS showed a peak at 960 cm −1 in the FT-IR spectrum. Because the peak at 960 cm −1 was assigned to the stretching Si-O vibration with the neighborhood metal ions, this peak had been used to characterize the incorporation of metal ions in the silica framework [39]. The Al/HMS sample prepared by post-modification did not show a peak at 960 cm −1 in the FT-IR spectrum. Hence, Al 3+ ions entered into the HMS framework in Al-HMS but existed at the outside of the HMS framework in Al/HMS.   The NH 3 -TPD measurement was used to evaluate the acidic strength of solid acids. The absorbed NH 3 molecules desorbed from weak solid acids at low temperatures and desorbed from strong solid acids at high temperatures. As shown in Figure 3, no peak could be observed in the NH 3 -TPD profile of HMS, indicating that HMS did not have any acid sites on the surface. Al-HMS had solid acid sites on the surface because a peak at 220 • C was observed in the NH 3 -TPD profile. The maximum temperature of NH 3 desorption of Al/HMS was higher than that of Al-HMS in the NH 3 -TPD profiles, implying that the acid sites in Al/HMS were stronger than those in Al-HMS. Because Al 3+ ions were introduced in Al-HMS at the preparation step, Al 3+ located uniformly in the HMS framework in Al-HMS. On the other hand, because Al 3+ ions were introduced in Al/HMS by post-modification, the Al 3+ ions existed at the extra-framework in Al/HMS. It has been reported that the extra-framework Al 3+ ions had stronger acidity compared with the intra-framework Al 3+ ions in Al-containing MCM-41 catalysts [40]. H-ZSM-5 exhibited two peaks at 220 and 450 • C in the NH 3 -TPD profile. Hence, H-ZSM-5 had both weak solid acid and strong solid acid on the surface. The peak at the highest temperature in the NH 3 -TPD profile corresponds to the strongest acid sites on the solid surface. H-ZSM-5 is a strong solid acid because the strongest acid sites determine the ability of a solid acid catalyst. According to the peak position at the maximum temperature in the NH 3 -TPD profiles, the acidic strength of various samples was in the order of H-ZSM-5 >> Al/HMS > Al-HMS > HMS (no acidity).  Figure 3 shows the NH3-TPD profiles of various samples after calcining at 550 °C for 4 h. The NH3-TPD measurement was used to evaluate the acidic strength of solid acids. The absorbed NH3 molecules desorbed from weak solid acids at low temperatures and desorbed from strong solid acids at high temperatures. As shown in Figure 3, no peak could be observed in the NH3-TPD profile of HMS, indicating that HMS did not have any acid sites on the surface. Al-HMS had solid acid sites on the surface because a peak at 220 °C was observed in the NH3-TPD profile. The maximum temperature of NH3 desorption of Al/HMS was higher than that of Al-HMS in the NH3-TPD profiles, implying that the acid sites in Al/HMS were stronger than those in Al-HMS. Because Al 3+ ions were introduced in Al-HMS at the preparation step, Al 3+ located uniformly in the HMS framework in Al-HMS. On the other hand, because Al 3+ ions were introduced in Al/HMS by post-modification, the Al 3+ ions existed at the extra-framework in Al/HMS. It has been reported that the extra-framework Al 3+ ions had stronger acidity compared with the intra-framework Al 3+ ions in Al-containing MCM-41 catalysts [40]. H-ZSM-5 exhibited two peaks at 220 and 450 °C in the NH3-TPD profile. Hence, H-ZSM-5 had both weak solid acid and strong solid acid on the surface. The peak at the highest temperature in the NH3-TPD profile corresponds to the strongest acid sites on the solid surface. H-ZSM-5 is a strong solid acid because the strongest acid sites determine the ability of a solid acid catalyst. According to the peak position at the maximum temperature in the NH3-TPD profiles, the acidic strength of various samples was in the order of H-ZSM-5 >> Al/HMS > Al-HMS > HMS (no acidity).  Table 1 shows the reaction results of ethylene oligomerization over various catalysts containing Ni and solid acid. Ni/H-ZSM-5 showed a high ethylene conversion of 90.7% at a low reaction temperature of 150 °C due to strong solid acidity of H-ZSM-5. Ni/H-ZSM-5 formed C4H8 as the major product and did not form any olefins larger than C6H12. Also, Ni/H-ZSM-5 formed a relatively large amount of C3H6 from the cracking of C6H12 formed in the reaction. Hence, Ni/H-ZSM-5 is not a suitable catalyst for producing higher olefins from ethylene oligomerization. Three catalysts containing Ni and mesoporous silica used in this study produced higher olefins larger than C6H12 from the ethylene oligomerization at 200 °C. Ni/Al-HMS showed an ethylene conversion of 96.3%, which was much higher than that over Ni/Al/HMS (45.6%) for the reaction. In Al-HMS, Al 3+ ions entered in the positions of Si 4+ ions in the framework and existed in the neighborhood of Si 4+ ion. In Al/HMS, Al 3+ ions did not enter in the positions of Si 4+ ions and existed far from Si 4+ ion. Hence, the existence of Al 3+ ions in the neighborhood of Si 4+ ions is important for improving the catalytic activity  Table 1 shows the reaction results of ethylene oligomerization over various catalysts containing Ni and solid acid. Ni/H-ZSM-5 showed a high ethylene conversion of 90.7% at a low reaction temperature of 150 • C due to strong solid acidity of H-ZSM-5. Ni/H-ZSM-5 formed C 4 H 8 as the major product and did not form any olefins larger than C 6 H 12 . Also, Ni/H-ZSM-5 formed a relatively large amount of C 3 H 6 from the cracking of C 6 H 12 formed in the reaction. Hence, Ni/H-ZSM-5 is not a suitable catalyst for producing higher olefins from ethylene oligomerization. Three catalysts containing Ni and mesoporous silica used in this study produced higher olefins larger than C 6 H 12 from the ethylene oligomerization at 200 • C. Ni/Al-HMS showed an ethylene conversion of 96.3%, which was much higher than that over Ni/Al/HMS (45.6%) for the reaction. In Al-HMS, Al 3+ ions entered in the positions of Si 4+ ions in the framework and existed in the neighborhood of Si 4+ ion. In Al/HMS, Al 3+ ions did not enter in the positions of Si 4+ ions and existed far from Si 4+ ion. Hence, the existence of Al 3+ ions in the neighborhood of Si 4+ ions is important for improving the catalytic activity of Ni-based catalyst in the ethylene oligomerization. Ni/Al-MCM-41 has been reported as an effective catalyst for the synthesis of higher olefins from ethylene oligomerization [34,38]. As shown in Table 1, Ni/Al-MCM-41 showed an ethylene conversion of 95.2% at 200 • C. Ni/Al-MCM-41 showed a C 8 C 16 selectivity of 16.3%, which was lower than that over Ni/Al-HMS (24.0%) for the reaction.  The ethylene conversion greatly increased by introducing Ni in Al-HMS, suggesting that the Ni sites in Ni/Al-HMS were the main active sites in the ethylene oligomerization. The ethylene conversion increased with increasing Ni content from 0 to 2 wt.%, and then slightly decreased at above 2 wt.% in Ni/Al-HMS. It has been reported that the Ni species exchanged into the pores of mesoporous silica are catalytically active species and the Ni species loaded on the mesoporous silica are inactive for the ethylene oligomerization on Ni/Al-MCM-41 [41]. Hence, about 2 wt.% Ni could be exchanged into the pores of Al-HMS and the excessive Ni was loaded on the Al-HMS surface when the Ni content was larger than 2 wt.% in Ni/Al-HMS. The selectivity for C 4 H 8 decreased with increasing Ni content from 0 to 1.5 wt.% and almost did not change at above 1.5 wt.% in Ni/Al-HMS. On the contrary, the selectivity for C 6 H 12 and the selectivity for C 8 H 16 increased with increasing Ni content from 0 to 1.5 wt.% and almost did not change at above 1.5 wt.% in Ni/Al-HMS. Hence, C 4 H 8 was the primary product and it converted to C 6+ higher olefins during the reaction. of Ni-based catalyst in the ethylene oligomerization. Ni/Al-MCM-41 has been reported as an effective catalyst for the synthesis of higher olefins from ethylene oligomerization [34,38]. As shown in Table  1, Ni/Al-MCM-41 showed an ethylene conversion of 95.2% at 200 °C. Ni/Al-MCM-41 showed a C8C16 selectivity of 16.3%, which was lower than that over Ni/Al-HMS (24.0%) for the reaction.  Figure 4 shows the dependence of Ni content in Ni/Al-HMS for the oligomerization of ethylene. Al-MHS (without Ni) showed a very low ethylene conversion of 2.2%, indicating that the weak acidity of Al-HMS just had a very low catalytic activity for ethylene oligomerization at 200 °C. The ethylene conversion greatly increased by introducing Ni in Al-HMS, suggesting that the Ni sites in Ni/Al-HMS were the main active sites in the ethylene oligomerization. The ethylene conversion increased with increasing Ni content from 0 to 2 wt.%, and then slightly decreased at above 2 wt.% in Ni/Al-HMS. It has been reported that the Ni species exchanged into the pores of mesoporous silica are catalytically active species and the Ni species loaded on the mesoporous silica are inactive for the ethylene oligomerization on Ni/Al-MCM-41 [41]. Hence, about 2 wt.% Ni could be exchanged into the pores of Al-HMS and the excessive Ni was loaded on the Al-HMS surface when the Ni content was larger than 2 wt.% in Ni/Al-HMS. The selectivity for C4H8 decreased with increasing Ni content from 0 to 1.5 wt.% and almost did not change at above 1.5 wt.% in Ni/Al-HMS. On the contrary, the selectivity for C6H12 and the selectivity for C8H16 increased with increasing Ni content from 0 to 1.5 wt.% and almost did not change at above 1.5 wt.% in Ni/Al-HMS. Hence, C4H8 was the primary product and it converted to C6+ higher olefins during the reaction.   Table 1. Figure 5 shows the reaction pathway of C 8 H 16 formation in the oligomerization of ethylene on Ni/Al-HMS. As discussed above, Ni sites are the main active sites for the ethylene oligomerization over Ni/Al-HMS. The oligomerization of ethylene on Ni is a Schulz-Flory type oligomerization. Any intermediate products formed during the reaction undergo further oligomerization (with ethylene) on Ni or desorption from Ni at the same time. Therefore, the products should obey a Schulz-Flory distribution with decreasing in the order of C 4 > C 6 > C 8 > C 10 . However, the selectivity for C 8 H 16 was similar to the selectivity for C 6 H 12 at each Ni content in Ni/Al-HMS (as shown in Figure 4). Hence, some other active species (except Ni) improved the amount of C 8 H 16 from the ethylene oligomerization over Ni/Al-HMS. As discussed above, although Al-HMS had acid sites on the surface (by introduction of Al 3+ ions), the acidity of Al-HMS was weak and thus Al-HMS showed a very low activity for the ethylene oligomerization. Also, C 4 H 8 formed from the ethylene dimerization was the primary product in the ethylene oligomerization over Ni/Al-HMS. Because C 4 H 8 had a higher reactivity than ethylene, it can be assumed that the primary product C 4 H 8 dimerized to C 8 H 16 on the acid site of Al-HMS. On Ni/Al-HMS, C 8 H 16 was formed not only from the reaction of C 6 H 12 with C 2 H 4 on Ni sites but also form the dimerization of C 4 H 8 on acid sites of Al-HMS. The acid sites in Al-HMS improved the amount of C 8 H 16 and caused a non-Schulz-Flory type distribution of the products in the oligomerization of ethylene over Ni/Al-HMS.  Figure 5 shows the reaction pathway of C8H16 formation in the oligomerization of ethylene on Ni/Al-HMS. As discussed above, Ni sites are the main active sites for the ethylene oligomerization over Ni/Al-HMS. The oligomerization of ethylene on Ni is a Schulz-Flory type oligomerization. Any intermediate products formed during the reaction undergo further oligomerization (with ethylene) on Ni or desorption from Ni at the same time. Therefore, the products should obey a Schulz-Flory distribution with decreasing in the order of C4 > C6 > C8 > C10. However, the selectivity for C8H16 was similar to the selectivity for C6H12 at each Ni content in Ni/Al-HMS (as shown in Figure 4). Hence, some other active species (except Ni) improved the amount of C8H16 from the ethylene oligomerization over Ni/Al-HMS. As discussed above, although Al-HMS had acid sites on the surface (by introduction of Al 3+ ions), the acidity of Al-HMS was weak and thus Al-HMS showed a very low activity for the ethylene oligomerization. Also, C4H8 formed from the ethylene dimerization was the primary product in the ethylene oligomerization over Ni/Al-HMS. Because C4H8 had a higher reactivity than ethylene, it can be assumed that the primary product C4H8 dimerized to C8H16 on the acid site of Al-HMS. On Ni/Al-HMS, C8H16 was formed not only from the reaction of C6H12 with C2H4 on Ni sites but also form the dimerization of C4H8 on acid sites of Al-HMS. The acid sites in Al-HMS improved the amount of C8H16 and caused a non-Schulz-Flory type distribution of the products in the oligomerization of ethylene over Ni/Al-HMS.   [35,37]. This is the reason that the introduction of Al 3+ ions brought solid acid sites in Al-HMS. The Bronsted acid sites have been assumed as the active sites for the olefin dimerization. Proton-type ion-exchange resin, a type of solid catalyst containing only Bronsted acid (without Lewis acid), has been known as an effective catalyst for the olefin dimerization [42]. Also, the Bronsted acid sites in Ni/Al-MCM-41 have been reported as the active sites for the dimerization of C4H8 [35]. Furthermore, it has been reported that Al-HMS has physical and textural properties similar to those of MCM-41 but with substantially higher Bronsted acidity [19]. Hence, Al-HMS is expected to have a stronger ability for C4H8 dimerization compared with that of Al-MCM-41. This is the reason that the selectivity for C8H16 over Ni/Al-HMS was larger than that over Ni/Al-MCM-41 in the oligomerization of ethylene (as shown in Table 1).    [35,37]. This is the reason that the introduction of Al 3+ ions brought solid acid sites in Al-HMS. The Bronsted acid sites have been assumed as the active sites for the olefin dimerization. Proton-type ion-exchange resin, a type of solid catalyst containing only Bronsted acid (without Lewis acid), has been known as an effective catalyst for the olefin dimerization [42]. Also, the Bronsted acid sites in Ni/Al-MCM-41 have been reported as the active sites for the dimerization of C 4 H 8 [35]. Furthermore, it has been reported that Al-HMS has physical and textural properties similar to those of MCM-41 but with substantially higher Bronsted acidity [19]. Hence, Al-HMS is expected to have a stronger ability for C 4 H 8 dimerization compared with that of Al-MCM-41. This is the reason that the selectivity for C 8 H 16 over Ni/Al-HMS was larger than that over Ni/Al-MCM-41 in the oligomerization of ethylene (as shown in Table 1).  Figure 5 shows the reaction pathway of C8H16 formation in the oligomerization of ethylene on Ni/Al-HMS. As discussed above, Ni sites are the main active sites for the ethylene oligomerization over Ni/Al-HMS. The oligomerization of ethylene on Ni is a Schulz-Flory type oligomerization. Any intermediate products formed during the reaction undergo further oligomerization (with ethylene) on Ni or desorption from Ni at the same time. Therefore, the products should obey a Schulz-Flory distribution with decreasing in the order of C4 > C6 > C8 > C10. However, the selectivity for C8H16 was similar to the selectivity for C6H12 at each Ni content in Ni/Al-HMS (as shown in Figure 4). Hence, some other active species (except Ni) improved the amount of C8H16 from the ethylene oligomerization over Ni/Al-HMS. As discussed above, although Al-HMS had acid sites on the surface (by introduction of Al 3+ ions), the acidity of Al-HMS was weak and thus Al-HMS showed a very low activity for the ethylene oligomerization. Also, C4H8 formed from the ethylene dimerization was the primary product in the ethylene oligomerization over Ni/Al-HMS. Because C4H8 had a higher reactivity than ethylene, it can be assumed that the primary product C4H8 dimerized to C8H16 on the acid site of Al-HMS. On Ni/Al-HMS, C8H16 was formed not only from the reaction of C6H12 with C2H4 on Ni sites but also form the dimerization of C4H8 on acid sites of Al-HMS. The acid sites in Al-HMS improved the amount of C8H16 and caused a non-Schulz-Flory type distribution of the products in the oligomerization of ethylene over Ni/Al-HMS.   [35,37]. This is the reason that the introduction of Al 3+ ions brought solid acid sites in Al-HMS. The Bronsted acid sites have been assumed as the active sites for the olefin dimerization. Proton-type ion-exchange resin, a type of solid catalyst containing only Bronsted acid (without Lewis acid), has been known as an effective catalyst for the olefin dimerization [42]. Also, the Bronsted acid sites in Ni/Al-MCM-41 have been reported as the active sites for the dimerization of C4H8 [35]. Furthermore, it has been reported that Al-HMS has physical and textural properties similar to those of MCM-41 but with substantially higher Bronsted acidity [19]. Hence, Al-HMS is expected to have a stronger ability for C4H8 dimerization compared with that of Al-MCM-41. This is the reason that the selectivity for C8H16 over Ni/Al-HMS was larger than that over Ni/Al-MCM-41 in the oligomerization of ethylene (as shown in Table 1).

Catalytic Ethanol Dehydration over H-ZSM-5
Bio-ethanol produced by the fermentation of biomass is a mixture of ethanol and water. Because ethanol and water form a binary azeotrope system containing 95.6 wt.% ethanol and 4.4 wt.% water, the conventional distillation cannot produce high-purity ethanol above 95.6 wt.%. The direct use of hydrous ethanol to produce ethylene is an environmentally friendly method by avoiding a strenuous separation process. The development of catalyst for the dehydration of ethanol to ethylene has been researched for many years [43]. The recent researches focus on improvement of catalyst stability in diluted bio-ethanol and improvement of catalyst activity at low reaction temperatures [6,7].
The dehydration of ethanol to ethylene is an acid-catalytic reaction which needs the existence of acid sites on the catalyst surface. Acidic zeolites, transition metal oxides, acidic clays, and heteropolyacids have been used in the dehydration of ethanol to ethylene [43]. Among various catalysts, acidic zeolite H-ZSM-5 is known as a highly active catalyst for the dehydration of ethanol [44]. Hence, we investigated a commercial H-ZSM-5 for the dehydration of ethanol in this study. Table 2 shows the effect of reaction temperature in the dehydration of ethanol over H-ZSM-5. The data were obtained after reaction for 1 h at various reaction temperatures. The C 2 H 5 OH conversion was 64.8% and the selectivity for C 2 H 4 was 75.6% for the reaction at a low temperature of 250 • C. The selectivity for the by-product C 2 H 5 OC 2 H 5 was high (23.8%) at 250 • C but it greatly decreased with increasing reaction temperature from 250 to 325 • C. Both the C 2 H 5 OH conversion and the selectivity for C 2 H 4 increased with increasing reaction temperature from 250 to 300 • C. When the reaction was carried at a high temperature of 325 • C, the C 2 H 5 OH conversion continuously increased to 98.7% but the selectivity for C 2 H 4 begun to decrease due to CH 3 CHO formation. In the dehydration of ethanol at 300 • C, the C 2 H 5 OH conversion was 95.4% and the selectivity for C 2 H 4 was 96.9%, which gave a C 2 H 4 yield of 92.4% for the dehydration of ethanol over H-ZSM-5. Equations (1)-(4) list the main reactions that occurred in the dehydration of ethanol on the solid acid catalysts. C 2 H 4 was formed through the dehydration from one C 2 H 5 OH molecule (Equation (1)). C 2 H 5 OC 2 H 5 was formed through the dehydration between two C 2 H 5 OH molecules (Equation (2)) [39]. At high reaction temperatures, the by-product C 2 H 5 OC 2 H 5 decomposed to C 2 H 5 OH and C 2 H 4 (Equation (3)) [8]. CH 3 CHO was formed at high temperatures through the dehydrogenation of ethanol (Equation (4) Figure 7 shows the influence of water amount in feed for the dehydration of ethanol over H-ZSM-5. When a pure ethanol (without water) was fed for the reaction, the C 2 H 4 yield almost did not decrease after reaction at 300 • C for 24 h. Using a feed containing 90 wt.% ethanol and 10 wt.% water, the C 2 H 4 yield was 92.4% after reaction for 1 h and the C 2 H 4 yield decreased to 91.2% after reaction for 24 h. By increasing water amount in feed from 10 wt.% to 50 wt.%, the C 2 H 4 yield after reaction for 1 h slightly decreased, and the C 2 H 4 yield after reaction for 24 h obviously decreased during the reaction. A large amount of water in feed caused the catalyst deactivation in the dehydration of ethanol over H-ZSM-5.  The purpose of this study is designing a continuous process for the direct synthesis of higher olefins from bio-ethanol by combining ethanol dehydration and ethylene oligomerization, but is not developing a highly active catalyst for the dehydration of ethanol. Because an ethylene yield of 91.2% could be obtained after reaction for 24 h using a feed containing 90 wt.% ethanol and 10 wt.% water (as shown in Figure 7), commercial H-ZSM-5 had a high catalytic performance for the dehydration of hydrous ethanol at 300 °C. Hence, we used the commercial H-ZSM-5 as a catalyst for the dehydration of hydrous ethanol in the following part.

Production of Higher Olefins from Hydrous Ethanol by Two-Step Process
A two-step reaction system was designed in this study to achieve the direct synthesis of higher olefins from bio-ethanol by combining ethanol dehydration and ethylene oligomerization. H-ZSM-5 was used as a catalyst for the first step of ethanol dehydration, and Ni/Al-HMS was used as a catalyst for the second step of ethylene oligomerization.
Because the binary azeotrope system of ethanol and water contains 4.4 wt.% of water, a mixture of ethanol and water is usually used as a feed in the ethanol dehydration. Moreover, a large amount of water (steam) is formed form the first step of ethanol dehydration (as shown in Equations (1) and (2)). The water certainly causes the catalyst deactivation in the following step of ethylene oligomerization over Ni/Al-HMS. It is necessary to remove the steam from the mixed gas before introducing the gas into the second reactor of ethylene oligomerization. A cold trap has been used for separating steam from other gas products in some reaction systems, such as in Fischer-Tropsch synthesis and in methanol dehydrogenation [45,46]. Figure 8 shows an illustration of the reaction process for the production of higher olefins from hydrous ethanol. The system contained two fixed-bed reactors and one cold trap. The two fixed-bed reactors were used to achieve ethanol dehydration and ethylene oligomerization, respectively. The cold trap was used to eliminate the steam from the mixed gas before the second step of ethylene The purpose of this study is designing a continuous process for the direct synthesis of higher olefins from bio-ethanol by combining ethanol dehydration and ethylene oligomerization, but is not developing a highly active catalyst for the dehydration of ethanol. Because an ethylene yield of 91.2% could be obtained after reaction for 24 h using a feed containing 90 wt.% ethanol and 10 wt.% water (as shown in Figure 7), commercial H-ZSM-5 had a high catalytic performance for the dehydration of hydrous ethanol at 300 • C. Hence, we used the commercial H-ZSM-5 as a catalyst for the dehydration of hydrous ethanol in the following part.

Production of Higher Olefins from Hydrous Ethanol by Two-Step Process
A two-step reaction system was designed in this study to achieve the direct synthesis of higher olefins from bio-ethanol by combining ethanol dehydration and ethylene oligomerization. H-ZSM-5 was used as a catalyst for the first step of ethanol dehydration, and Ni/Al-HMS was used as a catalyst for the second step of ethylene oligomerization.
Because the binary azeotrope system of ethanol and water contains 4.4 wt.% of water, a mixture of ethanol and water is usually used as a feed in the ethanol dehydration. Moreover, a large amount of water (steam) is formed form the first step of ethanol dehydration (as shown in Equations (1) and (2)). The water certainly causes the catalyst deactivation in the following step of ethylene oligomerization over Ni/Al-HMS. It is necessary to remove the steam from the mixed gas before introducing the gas into the second reactor of ethylene oligomerization. A cold trap has been used for separating steam from other gas products in some reaction systems, such as in Fischer-Tropsch synthesis and in methanol dehydrogenation [45,46]. Figure 8 shows an illustration of the reaction process for the production of higher olefins from hydrous ethanol. The system contained two fixed-bed reactors and one cold trap. The two fixed-bed reactors were used to achieve ethanol dehydration and ethylene oligomerization, respectively. The cold trap was used to eliminate the steam from the mixed gas before the second step of ethylene oligomerization. The reaction pressure in the system was controlled at 1 MPa by a back-pressure regulator. oligomerization. The reaction pressure in the system was controlled at 1 MPa by a back-pressure regulator. During the reaction, a mixture of ethanol and water (containing 90 wt.% ethanol and 10 wt.% water) was fed from a tank by a high-pressured micro-pump to reactor A (packed with H-ZSM-5 catalyst) to achieve the dehydration of ethanol to ethylene at 300 °C. At the same time, Ar was also introduced in reactor A from a cylinder as a carry gas and an inner standard material for calculation. Then, the gas which flowed out from the outlet of reactor A was introduced into a cold trap (cooled by ice-water) to collect the formed water, unreacted ethanol, by-products diethyl ether and acetaldehyde from the first step of ethanol dehydration. Finally, the gas which flowed out from the cold trap was introduced into reactor B (packed with Ni/Al-HMS catalyst) to achieve the oligomerization of ethylene at 200 °C. Figure 9 shows the time course of olefin yield obtained from the system containing two fixedbed reactors and one cold trap. A mixture of 90 wt.% C2H5OH and 10 wt.% H2O was fed in the reactor A with a rate of 0.13 g min −1 . By assuming the yield of C2H4 was 92% over H-ZSM-5 (Table 2) in the reactor A and water was collected in the cold trap, the gas which introduced in the reactor B contained 52.4 mL min −1 of C2H4 and 10 mL min −1 of Ar. After reaction for 1 h in the two-step process, the yields of C2H4, C4H8, C6H12, C8H16, and C8+ were 1.8%, 35.7%, 22.5%, 22.1%, and 7.6%, respectively. Moreover, the yield of each olefin formed from the two-step process almost did not change after reaction for 8 h under a reaction pressure of 1 MPa. During the reaction, a mixture of ethanol and water (containing 90 wt.% ethanol and 10 wt.% water) was fed from a tank by a high-pressured micro-pump to reactor A (packed with H-ZSM-5 catalyst) to achieve the dehydration of ethanol to ethylene at 300 • C. At the same time, Ar was also introduced in reactor A from a cylinder as a carry gas and an inner standard material for calculation. Then, the gas which flowed out from the outlet of reactor A was introduced into a cold trap (cooled by ice-water) to collect the formed water, unreacted ethanol, by-products diethyl ether and acetaldehyde from the first step of ethanol dehydration. Finally, the gas which flowed out from the cold trap was introduced into reactor B (packed with Ni/Al-HMS catalyst) to achieve the oligomerization of ethylene at 200 • C. Figure 9 shows the time course of olefin yield obtained from the system containing two fixed-bed reactors and one cold trap. A mixture of 90 wt.% C 2 H 5 OH and 10 wt.% H 2 O was fed in the reactor A with a rate of 0.13 g min −1 . By assuming the yield of C 2 H 4 was 92% over H-ZSM-5 (Table 2) in the reactor A and water was collected in the cold trap, the gas which introduced in the reactor B contained 52.4 mL min −1 of C 2 H 4 and 10 mL min −1 of Ar. After reaction for 1 h in the two-step process, the yields of C 2 H 4 , C 4 H 8 , C 6 H 12 , C 8 H 16 , and C 8+ were 1.8%, 35.7%, 22.5%, 22.1%, and 7.6%, respectively. Moreover, the yield of each olefin formed from the two-step process almost did not

Catalyst Syntheses
HMS was prepared using dodecylamine as a template according to the method in the literature [16]. A clear solution of Si(OC2H5)4 (1.00 mol) in ethanol (6.54 mol) was added to a stirring solution of dodecylamine (0.27 mol) and HCl (0.02 mol) in water (36.3 mol). Then, the resulting gel was aged for 18 h at room temperature to form a crystalline templated product. The solid sample obtained by filtration was dried at 110 °C for 24 h. Finally, the sample was calcined in air at 550 °C for 4 h to remove the template.
Al-HMS which incorporated Al 3+ ions in the HMS framework was prepared according to the method in the literature [19]. A clear solution containing 0.02 mol of Al(iso-OC3H7)3 in 35 mL isopropyl alcohol and 0.2 mol of Si(OC2H5)4 in 80 mL ethanol (Al:Si = 1/10 (molar ratio)) was heated at 70 °C for 4 h with vigorous stirring. Then, the solution was added to a solution containing 0.05 mol dodecylamine in 80 mL of water and 120 mL of ethanol. The solid sample obtained by filtration was dried at 110 °C for 24 h. Finally, the sample was calcined in air at 550 °C for 4 h to remove the template.
Al/HMS was prepared by grafting Al 3+ ions on HMS external surface according to the method in the literature [18,29]. The HMS sample which calcined at 550 °C for 4 h was impregnated with Al(iso-OC3H7)3 in isopropyl alcohol (Al:Si = 1/10 (molar ratio)), following by adding 25 mL H2O to precipitate aluminum oxide. After stirring at room temperature for 3 h, the solid product was filtrated, dried at 110 °C for 24 h, and calcined in air at 550 °C for 4 h.
Al-MCM-41 was prepared using cetyltrimethylammonium bromide (CTMABr) (25% in water) as a template according to the method in the literature [14,15]. Tetramethylammonium silicate solution (with a molar ratio of OH − to Si of 0.26) was used as a silica source. The gel with a molar composition of 4 Si:0.4 Al:CTMABr:250 H2O was heated at 100 °C for 4 days. As-synthesized Al-MCM-41 was obtained by filtration. The as-synthesized Al-MCM-41 was dried at 110 °C for 24 h, and then calcined in air at 550 °C for 4 h to remove the structurally incorporated template.
Ni/Al-HMS, Ni/Al/HMS, and Ni/Al/MCM-41 were prepared by cationic exchange with NH4NO3 solution according to the method in the literature [34]. Al-HMS, Al/HMS, or Al/MCM-41 was treated with an aqueous solution of NH4NO3 to form NH4-type sample. Then, the NH4-type sample was

Catalyst Syntheses
HMS was prepared using dodecylamine as a template according to the method in the literature [16]. A clear solution of Si(OC 2 H 5 ) 4 (1.00 mol) in ethanol (6.54 mol) was added to a stirring solution of dodecylamine (0.27 mol) and HCl (0.02 mol) in water (36.3 mol). Then, the resulting gel was aged for 18 h at room temperature to form a crystalline templated product. The solid sample obtained by filtration was dried at 110 • C for 24 h. Finally, the sample was calcined in air at 550 • C for 4 h to remove the template.
Al-HMS which incorporated Al 3+ ions in the HMS framework was prepared according to the method in the literature [19]. A clear solution containing 0.02 mol of Al(iso-OC 3 H 7 ) 3 in 35 mL isopropyl alcohol and 0.2 mol of Si(OC 2 H 5 ) 4 in 80 mL ethanol (Al:Si = 1/10 (molar ratio)) was heated at 70 • C for 4 h with vigorous stirring. Then, the solution was added to a solution containing 0.05 mol dodecylamine in 80 mL of water and 120 mL of ethanol. The solid sample obtained by filtration was dried at 110 • C for 24 h. Finally, the sample was calcined in air at 550 • C for 4 h to remove the template.
Al/HMS was prepared by grafting Al 3+ ions on HMS external surface according to the method in the literature [18,29]. The HMS sample which calcined at 550 • C for 4 h was impregnated with Al(iso-OC 3 H 7 ) 3 in isopropyl alcohol (Al:Si = 1/10 (molar ratio)), following by adding 25 mL H 2 O to precipitate aluminum oxide. After stirring at room temperature for 3 h, the solid product was filtrated, dried at 110 • C for 24 h, and calcined in air at 550 • C for 4 h.
Al-MCM-41 was prepared using cetyltrimethylammonium bromide (CTMABr) (25% in water) as a template according to the method in the literature [14,15]. Tetramethylammonium silicate solution (with a molar ratio of OH − to Si of 0.26) was used as a silica source. The gel with a molar composition of 4 Si:0. Acidic zeolite H-ZSM-5 was prepared from Na-ZSM-5 (Tosho Co., Tokyo, Japan, SiO 2 /Al 2 O 3 = 23.2, surface area: 322 m 2 g −1 ) by ion-exchange. Na-ZSM-5 was treated with an aqueous solution of NH 4 NO 3 to form NH 4 -ZSM-5. The obtained NH 4 -ZSM-5 was dried at 110 • C for 24 h and calcined in air at 550 • C for 4 h to form H-ZSM-5.
Ni/H-ZSM-5 was prepared by an wetness impregnation of H-ZSM-5 in Ni(NO 3 ) 2 aqueous solution. The water solvent was removed by evaporating at 95 • C. After impregnation, the obtained solid sample was dried at 110 • C for 24 h and calcined in air at 550 • C for 4 h.

Catalyst Characterization
The crystalline structure was characterized by an X-ray diffractometer (MAC Science MXP-18, MAC Science Co. Japan, Tokyo, Japan) operated at 40 kV and 50 mA using Cu Kα radiation. Fourier transform infrared spectra (FT-IR) measurement were recorded using a JASCO FT/IR spectrometer (JASCO Co., Osaka, Japan) under atmospheric conditions. A KBr pellets technology was used in the FT-IR measurement and the mass ratio of sample to KBr was 1:100. Temperature-programmed desorption of ammonia (NH 3 -TPD) measurement was carried out using a BELCAT-B automatic system (Bell Co. Japan, Tokyo, Japan). A thermal conductivity detector (TCD) (Shimadzu Co., Kyoto, Japan) and an Omnistar Q-mass (Shimadzu Co., Kyoto, Japan) were used for detecting ammonia. In a typical NH 3 -TPD measurement, 0.05 g sample was pretreated in a He flow (50 mL min −1 ) at 400 • C for 1 h. After the furnace was cooled to 100 • C, ammonia was adsorbed onto the sample's surface. The sample was evacuated at 100 • C for 1 h to eliminate the weakly physical adsorbed ammonia, and then NH 3 -TPD was recorded from 100 to 600 • C by heating the furnace with a rate of 8 • C min −1 .

Reactions and Instruments
The oligomerization of ethylene was performed in a fixed-bed catalytic reaction system. The catalyst (24-60 meshes) was packed in a stainless-steel reactor (i.d.: 1 cm; length: 40 cm), and was pretreated in a N 2 flow at 500 • C for 2 h. After the reactor was cooled down to reaction temperature, a mixed gas containing 90% C 2 H 4 and 10% Ar was introduced in the reactor from a cylinder. The reaction pressure was 1 MPa and the main reaction temperature was 200 • C.
The dehydration of ethanol was performed in a fixed-bed catalytic reaction system. The catalyst (24-60 meshes) was packed in a stainless-steel reactor (i.d.: 1 cm; length: 40 cm), and was pretreated in a N 2 flow at 500 • C for 2 h. After the reactor was cooled down to the reaction temperature, a mixture of ethanol and water was fed in the reactor from a tank by a high-pressured micro-pump (Shimadzu LC-10, Shimadzu Co., Kyoto, Japan). At the same time, Ar was introduced in the reactor from a cylinder as a carrier gas and an inner standard material for calculation.
During the reaction, the products were continuously analyzed using an on-line GC analysis system. Inorganic gases were analyzed using a Shimadzu 14B TCD-GC (Shimadzu Co., Kyoto, Japan) equipped with a packed column filled by SHINCARBON. C 1 -C 4 gas hydrocarbons were analyzed using a Shimadzu GC-2014 FID-GC (Shimadzu Co., Kyoto, Japan) equipped with a RT-QPLOT capillary column (Agilent Technologies Inc., Santaclara, CA, USA). Liquid organic compounds were analyzed using a Shimadzu GC-2014 FID-GC equipped a Stabilwax capillary column (Restek Co., Bellefont, PA, USA). A standard mixed gas (from a cylinder) with known concentration for each component was used to calculate the factors of gas samples. A standard mixed solution with known concentration for each component was prepared to calculate the factors of liquid organic compounds.
The carbon mass balance (before and after reaction) had an error less than ±5% for the reactions carried out in this study.

Conclusions
Ni/Al-HMS was an effective catalyst for the oligomerization of ethylene to higher olefins. The selectivity for C 8 H 16 over Ni/Al-HMS was larger than that over the reported Ni/Al-MCM-41 catalyst. The existence of Al 3+ ion in the neighborhood of Si 4+ ion was important to improve the catalytic activity of Ni/Al-HMS in the oligomerization of ethylene. Ni sites were the main active sites in the ethylene oligomerization over Ni/Al-HMS. The primary product was C 4 H 8 and the formed C 4 H 8 was converted to larger olefins in the oligomerization of ethylene. The weak acid sites in Al-HMS had activity for the dimerization of the primary product C 4 H 8 (formed on Ni sites) to C 8 H 16 , but they almost had no activity for the dimerization of C 2 H 4 to C 4 H 8 . H-ZSM-5 was an effective catalyst for the dehydration of ethanol to ethylene, but a large amount of water in the feed caused the catalyst deactivation during the reaction. Using a reaction system containing two fixed-bed reactors (to achieve ethanol dehydration on H-ZSM-5 and ethylene oligomerization on Ni/Al-HMS) and one cold trap (to eliminate water), higher olefins were continuously produced by feeding a mixture containing 90 wt.% ethanol and 10 wt.% water.