catalysts Chromium Oxide Supported on Silicalite-1 Zeolite as a Novel E ﬃ cient Catalyst for Dehydrogenation of Isobutane Assisted by CO 2

: The chromium oxide catalysts supported on silicalite-1 zeolite (Cr / S-1) with a Cr content between 0.5% and 7% were synthesized via an incipient wetness method. The catalysts were characterized by XRD, N 2 adsorption, TEM-EDX, UV-vis, DRIFTS, 29 Si MAS NMR, XPS, H 2 -TPR, and NH 3 -TPD. The optimum 3%Cr / S-1 catalyst with 3%Cr is more active and stable than SBA-15-supported one with the same Cr content, which is a consequence of a higher content of Cr 6 + in the fresh 3%Cr / S-1 catalyst and a higher content of Cr 6 + retained on the former catalyst during the reaction. The 3%Cr / S-1 catalyst a ﬀ ords an isobutane conversion of 36.5% with 71.2% isobutene selectivity. The catalytic activity is well correlated with the content of Cr 6 + in the fresh catalysts. Carbon dioxide displays a promoting e ﬀ ect on the dehydrogenation reaction.


Introduction
Isobutene is an important industrial chemical employed to produce butyl rubber, gasoline oxygenates (e.g., ethyl tert-butyl ether), and antioxidants (e.g., butylated hydroxyanisole) [1,2]. Its two main manufacture ways which rely on the source of petroleum, i.e., steam cracking of naphtha as well as fluidized catalytic cracking, cannot meet the increasing requirements. Due to the shortage of petroleum and environmental consideration, the dehydrogenation of small alkanes to alkenes assisted by CO 2 has attracted more attention recently [3][4][5][6][7][8][9][10][11][12][13]. Compared to the oxidative dehydrogenation of light alkanes with O 2 , the beneficial employment of CO 2 as a soft oxidant comprises improving the product selectivity as well as decreasing CO 2 emissions [14,15]. Moreover, this route opens up a new way to utilize greenhouse CO 2 .
The band at 541 cm −1 on the Raman spectra of the Cr/S-1 and 3%Cr/SBA catalysts is characteristic of Cr2O3 crystallites ( Figure 2) [33][34][35]. This band does not appear until the Cr content achieves 3% for the Cr/S-1 catalysts. The stronger intensity of this band found for the 3%Cr/SBA catalyst than 3%Cr/S-1 suggests that chromium oxide is worse dispersed on the former catalyst, which is consistent with the result of STEM mapping. The band at 983 cm −1 and the shoulder at 1006 cm −1 are related to the Cr-O stretching of monochromate and polychromate species, respectively [6,36]. The band at 603 cm −1 occurred in the 3%Cr/SBA catalyst is associated with a tri-siloxane ring in SBA-15 [37]. A surface area of 379 m 2 /g observed for silicalite-1 zeolite (Table 1) is similar to the value reported in the literature [30]. As the Cr content increases from 0.5% to 3%, the surface area, microporosity and mesoporosity (contribution from silicalite-1 intercrystalline voids) of the Cr/S-1 catalysts slightly decrease (Table 1). At a high content of Cr (7%), the microporosity obviously decreases, which can be attributed to the blockage of some micropores by large chromium oxide particles (evidenced by XRD observation).
The band at 541 cm −1 on the Raman spectra of the Cr/S-1 and 3%Cr/SBA catalysts is characteristic of Cr 2 O 3 crystallites ( Figure 2) [33][34][35]. This band does not appear until the Cr content achieves 3% for the Cr/S-1 catalysts. The stronger intensity of this band found for the 3%Cr/SBA catalyst than 3%Cr/S-1 suggests that chromium oxide is worse dispersed on the former catalyst, which is consistent with the result of STEM mapping. The band at 983 cm −1 and the shoulder at 1006 cm −1 are related to the Cr-O stretching of monochromate and polychromate species, respectively [6,36]. The band at 603 cm −1 occurred in the 3%Cr/SBA catalyst is associated with a tri-siloxane ring in SBA-15 [37].
The bands at 275 and 359 nm on the diffuse reflectance UV-vis spectra of the Cr/S-1 and 3%Cr/SBA catalysts are attributed to tetrahedral Cr 6+ species, while the ones at 458 and 599 nm are assigned to octahedral Cr 3+ species existing in Cr 2 O 3 or CrO x clusters ( Figure 3) [36,38]. The band at 599 nm cannot be found for the Cr/S-1 catalysts until the Cr content of 7%. 3%Cr/SBA shows the strongest intensity of this band, suggesting the worse dispersion of chromium oxide on 3%Cr/SBA than 3%Cr/S-1. The above result is consistent with those of XRD and STEM mapping.   The bands at 275 and 359 nm on the diffuse reflectance UV-vis spectra of the Cr/S-1 and 3%Cr/SBA catalysts are attributed to tetrahedral Cr 6+ species, while the ones at 458 and 599 nm are assigned to octahedral Cr 3+ species existing in Cr2O3 or CrOx clusters ( Figure 3) [36,38]. The band at 599 nm cannot be found for the Cr/S-1 catalysts until the Cr content of 7%. 3%Cr/SBA shows the strongest intensity of this band, suggesting the worse dispersion of chromium oxide on 3%Cr/SBA than 3%Cr/S-1. The above result is consistent with those of XRD and STEM mapping.
The XPS spectra of the 3%Cr/S-1 and 3%Cr/SBA catalysts are depicted in Figure S3, and the XPS data are listed in Table 2. A remarkable decrease in the Cr 6+ to Cr 3+ ratio after isobutane dehydrogenation is indicative of the reduction of most Cr 6+ to Cr 3+ . More Cr 6+ was reduced to Cr 3+ in the absence of CO2 than in the presence of CO2. The Cr 6+ to Cr 3+ ratio is greater for 3%Cr/S-1 than 3%Cr/SBA, both before and after the dehydrogenation reaction assisted by CO2. After isobutane dehydrogenation over 3%Cr/S-1 in the absence of CO2, followed by the treatment with CO2 at the same temperature, the Cr 6+ to Cr 3+ ratio increases from 0.91 to 1.97, but it is still lower than the value of the fresh catalyst (2.82). This finding implies that CO2 assisted isobutane dehydrogenation proceeds via a redox mechanism (Scheme 1).    The XPS spectra of the 3%Cr/S-1 and 3%Cr/SBA catalysts are depicted in Figure S3, and the XPS data are listed in Table 2. A remarkable decrease in the Cr 6+ to Cr 3+ ratio after isobutane dehydrogenation is indicative of the reduction of most Cr 6+ to Cr 3+ . More Cr 6+ was reduced to Cr 3+ in the absence of CO 2 than in the presence of CO 2 . The Cr 6+ to Cr 3+ ratio is greater for 3%Cr/S-1 than 3%Cr/SBA, both before and after the dehydrogenation reaction assisted by CO 2 . After isobutane dehydrogenation over 3%Cr/S-1 in the absence of CO 2 , followed by the treatment with CO 2 at the same temperature, the Cr 6+ to Cr 3+ ratio increases from 0.91 to 1.97, but it is still lower than the value of the fresh catalyst (2.82). This finding implies that CO 2 assisted isobutane dehydrogenation proceeds via a redox mechanism (Scheme 1).   The peak on the H2-TPR profiles of the Cr/S-1 and 3%Cr/SBA catalysts with peak temperatures between 364 and 424 °C is assigned to the reduction of Cr 6+ to Cr 3+ ( Figure 4) [39][40][41]. An additional Scheme 1. Proposed reaction mechanism of isobutane dehydrogenation assisted by CO 2 over silicalite-1-supported chromium oxide catalysts.
The peak on the H 2 -TPR profiles of the Cr/S-1 and 3%Cr/SBA catalysts with peak temperatures between 364 and 424 • C is assigned to the reduction of Cr 6+ to Cr 3+ ( Figure 4) [39][40][41]. An additional small peak at 213 • C is observed for the 7%Cr/S-1 catalyst, indicating that large chromia crystals exist on this catalyst in addition to the dispersed chromium oxide [42]. The H 2 -TPR data are presented in Table 1. The 3%Cr/S-1 catalyst exhibits a lower reduction in temperature than 3%Cr/SBA (364 • C vs. 373 • C), showing that the former catalyst displays higher reducibility than the latter one. The higher content of Cr 6+ observed in 3%Cr/S-1 than 3%Cr/SBA (1.19% vs. 1.05%) could be related to the better dispersion of chromium species on silicalite-1 [11], as demonstrated by the STEM mapping result. on this catalyst in addition to the dispersed chromium oxide [42]. The H2-TPR data are presented in Table 1. The 3%Cr/S-1 catalyst exhibits a lower reduction in temperature than 3%Cr/SBA (364 °C vs. 373 °C), showing that the former catalyst displays higher reducibility than the latter one. The higher content of Cr 6+ observed in 3%Cr/S-1 than 3%Cr/SBA (1.19% vs. 1.05%) could be related to the better dispersion of chromium species on silicalite-1 [11], as demonstrated by the STEM mapping result. As revealed in Figure 5, silicalite-1 zeolite has three kinds of hydroxyl groups: Isolated silanol groups (3739 cm −1 ), vicinal silanol groups (3686 cm −1 ), and nest silanol groups (3493 cm −1 ) [27,43,44]. After supporting chromium oxide, the intensities of these -OH groups, particularly the nest silanol groups, diminish. The peak area ratio of nest silanol groups to isolated silanol ones declines from 16 to 5.0 after supporting chromia (3%Cr). This observation suggests that the number of silanol groups decreases via the interaction of Cr species with the -OH groups [29,44,45]. SBA-15 has only isolated silanol groups. Supporting chromium oxide on SBA-15 also leads to a decrease in the number of hydroxyl groups. The nest silanols have a higher local density of hydroxyls than isolated silanols [46], thus resulting in a stronger interaction between chromium species and nest silanols. Therefore, the better dispersion of chromium species on silicalite-1 than SBA-15 can be attributed to the abundant nest silanol groups present on the former support.  As revealed in Figure 5, silicalite-1 zeolite has three kinds of hydroxyl groups: Isolated silanol groups (3739 cm −1 ), vicinal silanol groups (3686 cm −1 ), and nest silanol groups (3493 cm −1 ) [27,43,44]. After supporting chromium oxide, the intensities of these -OH groups, particularly the nest silanol groups, diminish. The peak area ratio of nest silanol groups to isolated silanol ones declines from 16 to 5.0 after supporting chromia (3%Cr). This observation suggests that the number of silanol groups decreases via the interaction of Cr species with the -OH groups [29,44,45]. SBA-15 has only isolated silanol groups. Supporting chromium oxide on SBA-15 also leads to a decrease in the number of hydroxyl groups. The nest silanols have a higher local density of hydroxyls than isolated silanols [46], thus resulting in a stronger interaction between chromium species and nest silanols. Therefore, the better dispersion of chromium species on silicalite-1 than SBA-15 can be attributed to the abundant nest silanol groups present on the former support. on this catalyst in addition to the dispersed chromium oxide [42]. The H2-TPR data are presented in Table 1. The 3%Cr/S-1 catalyst exhibits a lower reduction in temperature than 3%Cr/SBA (364 °C vs. 373 °C), showing that the former catalyst displays higher reducibility than the latter one. The higher content of Cr 6+ observed in 3%Cr/S-1 than 3%Cr/SBA (1.19% vs. 1.05%) could be related to the better dispersion of chromium species on silicalite-1 [11], as demonstrated by the STEM mapping result. As revealed in Figure 5, silicalite-1 zeolite has three kinds of hydroxyl groups: Isolated silanol groups (3739 cm −1 ), vicinal silanol groups (3686 cm −1 ), and nest silanol groups (3493 cm −1 ) [27,43,44]. After supporting chromium oxide, the intensities of these -OH groups, particularly the nest silanol groups, diminish. The peak area ratio of nest silanol groups to isolated silanol ones declines from 16 to 5.0 after supporting chromia (3%Cr). This observation suggests that the number of silanol groups decreases via the interaction of Cr species with the -OH groups [29,44,45]. SBA-15 has only isolated silanol groups. Supporting chromium oxide on SBA-15 also leads to a decrease in the number of hydroxyl groups. The nest silanols have a higher local density of hydroxyls than isolated silanols [46], thus resulting in a stronger interaction between chromium species and nest silanols. Therefore, the better dispersion of chromium species on silicalite-1 than SBA-15 can be attributed to the abundant nest silanol groups present on the former support.    [28]. The ratio of Q 3 to Q 4 decreases from 0.108 for silicalite-1 to 0.096 for 3%Cr/S-1, indicating a loss of silanol groups. In the case of SBA-15 and 3%Cr/SBA, the broad signal can be deconvoluted into a few peaks. The peaks at ca. −90 and −100 ppm are attributed to Q 2 ([HO] 2 -Si-[(OSi) 2 ]) and Q 3 species, respectively, while the ones below −102 ppm are assigned to Q 4 species [23,47]. The lower (Q 2 + Q 3 ) to Q 4 ratio observed for 3%Cr/SBA than SBA-15 (0.272 vs. 0.341) implies a decrease in the number of hydroxyl groups.

Catalytic Performance
The catalytic results of CO 2 assisted isobutane dehydrogenation over the Cr/S-1 catalysts are shown in Figure 6 and Table 3. The effect of content of Cr in the Cr/S-1 catalysts on the initial conversion of isobutane and initial product selectivity is depicted in Figure S5. The initial isobutene selectivity declines from 79.7% to 71.2%, with increasing the Cr content from 0.5% to 3%, followed by a slight diminishment with further increasing the Cr content to 7%. The initial selectivities to C1-C3 (alkanes and alkenes) and butenes (except isobutene) follow the opposite variation trend. The activity is strongly dependent on the Cr content. The initial isobutane conversion improves markedly from 20.8% to 36.5% with increasing the Cr content from 0.5% to 3%, followed by a very slight decline in the conversion with further increasing the Cr content to 7%. The 3%Cr/S-1 catalyst displays the optimum activity, giving 36.5% isobutane conversion and 71.2% selectivity toward isobutene. As shown in Table 3, the catalyst which is more active for CO 2 assisted dehydrogenation of isobutane displays higher activity for the conversion of CO 2 to CO. The very small conversion of isobutane (ca. 3%) observed on silicalite-1 zeolite suggests that the dispersed chromium oxide on silicalite-1 is primarily responsible for the catalytic activity. Figure S4 presents the 29 Si MAS NMR spectra of some selected samples. For silicalite-1 and 3%Cr/S-1 samples, two resonances at −113 and −102 ppm correspond to Q 4 and Q 3 species representing (Si-[(OSi)4]) and (HO-Si-[(OSi)3]), respectively [28]. The ratio of Q 3 to Q 4 decreases from 0.108 for silicalite-1 to 0.096 for 3%Cr/S-1, indicating a loss of silanol groups. In the case of SBA-15 and 3%Cr/SBA, the broad signal can be deconvoluted into a few peaks. The peaks at ca. −90 and −100 ppm are attributed to Q 2 ([HO]2-Si-[(OSi)2]) and Q 3 species, respectively, while the ones below −102 ppm are assigned to Q 4 species [23,47]. The lower (Q 2 + Q 3 ) to Q 4 ratio observed for 3%Cr/SBA than SBA-15 (0.272 vs. 0.341) implies a decrease in the number of hydroxyl groups.

Catalytic Performance
The catalytic results of CO2 assisted isobutane dehydrogenation over the Cr/S-1 catalysts are shown in Figure 6 and Table 3. The effect of content of Cr in the Cr/S-1 catalysts on the initial conversion of isobutane and initial product selectivity is depicted in Figure S5. The initial isobutene selectivity declines from 79.7% to 71.2%, with increasing the Cr content from 0.5% to 3%, followed by a slight diminishment with further increasing the Cr content to 7%. The initial selectivities to C1-C3 (alkanes and alkenes) and butenes (except isobutene) follow the opposite variation trend. The activity is strongly dependent on the Cr content. The initial isobutane conversion improves markedly from 20.8% to 36.5% with increasing the Cr content from 0.5% to 3%, followed by a very slight decline in the conversion with further increasing the Cr content to 7%. The 3%Cr/S-1 catalyst displays the optimum activity, giving 36.5% isobutane conversion and 71.2% selectivity toward isobutene. As shown in Table 3, the catalyst which is more active for CO2 assisted dehydrogenation of isobutane displays higher activity for the conversion of CO2 to CO. The very small conversion of isobutane (ca. 3%) observed on silicalite-1 zeolite suggests that the dispersed chromium oxide on silicalite-1 is primarily responsible for the catalytic activity.    a The values outside and inside the bracket are the data obtained at 10 min and 6 h, respectively; b Butenes excluding isobutene; c Molar ratio of H 2 to CO. Reaction conditions are the same as given in caption to Figure 6.
It is well accepted that CO 2 assisted dehydrogenation of small alkanes over metal oxides with redox property (e.g., chromia and vanadia) proceeds through a redox mechanism [18,22,48,49]: The XPS results (Table 2) confirm that CO 2 assisted isobutane dehydrogenation follows the redox mechanism (Scheme 1). According to this redox mechanism, one might envisage that the Cr/S-1 catalyst having a higher amount of Cr 6+ species would favor the dehydrogenation reaction. This hypothesis is further demonstrated by a good correlation between the initial activity and the content of Cr 6+ in the fresh Cr/S-1 catalysts as measured by H 2 -TPR (Figure 7). On the other hand, H 2 was detectable in the products. The H 2 to CO molar ratio for the Cr/S-1 catalysts is presented in Table 3. This finding suggests that besides the redox mechanism (i.e., a one-step pathway, i-C 4 H 10 + CO 2 = i-C 4 H 8 + CO + H 2 O), a two-step pathway also occurs during the reaction, i.e., a simple dehydrogenation of isobutane coupled with the reverse water-gas shift (RWGS) reaction (Scheme 1). The results of a separate RWGS reaction carried out at 570 • C verify that the Cr/S-1 catalysts are indeed active for this reaction ( Figure S6). It is well accepted that CO2 assisted dehydrogenation of small alkanes over metal oxides with redox property (e.g., chromia and vanadia) proceeds through a redox mechanism [18,22,48,49]: The XPS results (Table 2) confirm that CO2 assisted isobutane dehydrogenation follows the redox mechanism (Scheme 1). According to this redox mechanism, one might envisage that the Cr/S-1 catalyst having a higher amount of Cr 6+ species would favor the dehydrogenation reaction. This hypothesis is further demonstrated by a good correlation between the initial activity and the content of Cr 6+ in the fresh Cr/S-1 catalysts as measured by H2-TPR (Figure 7). On the other hand, H2 was detectable in the products. The H2 to CO molar ratio for the Cr/S-1 catalysts is presented in Table 3. This finding suggests that besides the redox mechanism (i.e., a one-step pathway, i-C4H10 + CO2 = i-C4H8 + CO + H2O), a two-step pathway also occurs during the reaction, i.e., a simple dehydrogenation of isobutane coupled with the reverse water-gas shift (RWGS) reaction (Scheme 1). The results of a separate RWGS reaction carried out at 570 °C verify that the Cr/S-1 catalysts are indeed active for this reaction ( Figure S6). We chose the best 3%Cr/S-1 to investigate its catalytic performance under a CO2 or N2 atmosphere ( Figure 8). In the case of CO2 atmosphere, this catalyst gives a 36.5% isobutane conversion with 71.2% isobutene selectivity after 10 min of reaction. In the case of N2 atmosphere, the initial isobutane conversion and isobutene selectivity are 30.8% and 67.5%, respectively. This finding suggests that CO2 displays a promoting effect on isobutane dehydrogenation. CO2 enhances the dehydrogenation reaction via a redox mechanism in which the catalyst undergoes reduction (by isobutane) and reoxidation (by carbon dioxide) cycles as well as the reaction coupling between a simple dehydrogenation of isobutane and the RWGS reaction, as illustrated in Scheme 1. We chose the best 3%Cr/S-1 to investigate its catalytic performance under a CO 2 or N 2 atmosphere (Figure 8). In the case of CO 2 atmosphere, this catalyst gives a 36.5% isobutane conversion with 71.2% isobutene selectivity after 10 min of reaction. In the case of N 2 atmosphere, the initial isobutane conversion and isobutene selectivity are 30.8% and 67.5%, respectively. This finding suggests that CO 2 displays a promoting effect on isobutane dehydrogenation. CO 2 enhances the dehydrogenation reaction via a redox mechanism in which the catalyst undergoes reduction (by isobutane) and reoxidation (by carbon dioxide) cycles as well as the reaction coupling between a simple dehydrogenation of isobutane and the RWGS reaction, as illustrated in Scheme 1.
A comparison of 3%Cr/S-1 and 3%Cr/SBA catalysts indicates that the initial isobutane conversion is higher on 3%Cr/S-1 than 3%Cr/SBA (Figure 9, 36.5% vs. 33.5%), which is caused by the fact that the former catalyst possesses a higher content of Cr 6+ than the latter one (1.19% vs. 1.05%). The higher isobutene selectivity observed for 3%Cr/SBA than 3%Cr/S-1 is due to the fact that the former catalyst has lower acidity and weaker acid sites. Two desorption peaks on the NH 3 -TPD profiles of both 3%Cr/S-1 and 3%Cr/SBA catalysts correspond to the weak and strong acid sites of the catalysts ( Figure S7). The higher peak temperature observed for 3%Cr/S-1 than 3%Cr/SBA (393 • C vs. 320 • C) suggests that the former catalyst has stronger acid sites than the latter one. Moreover, the 3%Cr/S-1 catalyst has more acid sites than 3%Cr/SBA (0.445 vs. 0.266 mmol/g). Moreover, the 3%Cr/S-1 catalyst exhibits higher stability than 3%Cr/SBA. After 6 h of the reaction, the isobutane conversion for 3%Cr/S-1 and 3%Cr/SBA is 28.4% and 20.2%, respectively. Coking and the reduction of Cr 6+ to Cr 3+ are two causes responsible for the catalyst deactivation [11,50]. An in situ pretreatment of the 3%Cr/S-1 catalyst with 10% H 2 /Ar (30 mL/min) at 450 • C for 1 h leads to a decline in the initial activity from 36.5% to 32.3% (Figure 10), further confirming that the reduction of Cr 6+ to Cr 3+ is one of the causes for the deactivation of catalyst. The amount of coke on 3%Cr/S-1 and 3%Cr/SBA catalysts after 6 h of the reaction is 3.2% and 1.5%, respectively. The TPO profiles of both spent 3%Cr/S-1 and 3%Cr/SBA catalysts are shown in Figure S8. The peak temperature is lower for the former catalyst than the latter one (359 • C vs. 375 • C), suggesting that coke deposited on 3%Cr/S-1 is more easily burnt off. The NH 3 -TPD result accounts for the higher amount of coke on 3%Cr/S-1 than 3%Cr/SBA. The XPS result shows that after the reaction, the Cr 6+ to Cr 3+ ratio is higher for 3%Cr/S-1 than 3%Cr/SBA ( Table 2, 1.19 vs. 0.72). This result implies that the reduced chromium species which interact with the nest silanols could be more easily reoxidized to Cr 6+ species by CO 2 during the dehydrogenation reaction. Hence, the higher catalytic stability of 3%Cr/S-1 than 3%Cr/SBA is ascribed to a higher amount of Cr 6+ species retained during the reaction, which could be related to the abundant nest silanol groups present on the silicalite-1 support. There are no differences in the XRD patterns for the spent and fresh 3%Cr/S-1 catalysts ( Figure S9), indicating the good maintenance of the catalyst structure. However, an evident diminishment in the intensity of XRD patterns for the spent 3%Cr/SBA catalyst in comparison with the fresh one was observed, suggesting a degradation of the ordered hexagonal arrangement of the SBA-15 mesopores. This is another cause for the lower catalytic stability of the 3%Cr/SBA catalyst. A comparison of 3%Cr/S-1 and 3%Cr/SBA catalysts indicates that the initial isobutane conversion is higher on 3%Cr/S-1 than 3%Cr/SBA (Figure 9, 36.5% vs. 33.5%), which is caused by the fact that the former catalyst possesses a higher content of Cr 6+ than the latter one (1.19% vs. 1.05%). The higher isobutene selectivity observed for 3%Cr/SBA than 3%Cr/S-1 is due to the fact that the former catalyst has lower acidity and weaker acid sites. Two desorption peaks on the NH3-TPD profiles of both 3%Cr/S-1 and 3%Cr/SBA catalysts correspond to the weak and strong acid sites of the catalysts ( Figure S7). The higher peak temperature observed for 3%Cr/S-1 than 3%Cr/SBA (393 °C vs. 320 °C) suggests that the former catalyst has stronger acid sites than the latter one. Moreover, the 3%Cr/S-1 catalyst has more acid sites than 3%Cr/SBA (0.445 vs. 0.266 mmol/g). Moreover, the 3%Cr/S-1 catalyst exhibits higher stability than 3%Cr/SBA. After 6 h of the reaction, the isobutane conversion for 3%Cr/S-1 and 3%Cr/SBA is 28.4% and 20.2%, respectively. Coking and the reduction of Cr 6+ to Cr 3+ are two causes responsible for the catalyst deactivation [11,50]. An in situ pretreatment of the 3%Cr/S-1 catalyst with 10% H2/Ar (30 mL/min) at 450 °C for 1 h leads to a decline in the initial activity from 36.5% to 32.3% (Figure 10), further confirming that the reduction of Cr 6+ to Cr 3+ is one of the causes for the deactivation of catalyst. The amount of coke on 3%Cr/S-1 and 3%Cr/SBA catalysts after 6 h of the reaction is 3.2% and 1.5%, respectively. The TPO profiles of both spent 3%Cr/S-1 and 3%Cr/SBA catalysts are shown in Figure S8. The peak temperature is lower for the former catalyst than the latter one (359 °C vs. 375 °C), suggesting that coke deposited on 3%Cr/S-1 is more easily burnt off. The NH3-TPD result accounts for the higher amount of coke on 3%Cr/S-1 than 3%Cr/SBA. The XPS result shows that after the reaction, the Cr 6+ to Cr 3+ ratio is higher for 3%Cr/S-1 than 3%Cr/SBA (

Catalyst Preparation
Silicalite-1 zeolite was synthesized as follows [51]. Tetrapropylammonium hydroxide (TPAOH, 25% aqueous solution), tetraethyl orthosilicate (TEOS), and distilled water were mixed to obtain a clear suspension (9TPAOH:25SiO2:480H2O, molar composition). The above mixture was stirred at room temperature to hydrolyze TEOS for 4 h, followed by being transferred into an autoclave and crystallized at 170 °C for 72 h. The obtained product was filtered, washed, and dried at 100 °C overnight, followed by calcination at 550 °C for 4 h in air.
The silicalite-1-supported chromium oxide catalysts were prepared through an incipient wetness method employing Cr(NO3)3·9H2O as the precursor. The impregnated samples were dried at 100 °C overnight, followed by calcination in air at 600 °C for 6 h. The obtained catalysts were designated as x%Cr/S-1, where x% represents the weight percentage of Cr in the catalysts. For the purpose of comparison, the 3%Cr/SBA catalyst (3%Cr) was prepared in the same way using SBA-15 as the support. SBA-15 was prepared according to the literature [52].

Catalyst Preparation
Silicalite-1 zeolite was synthesized as follows [51]. Tetrapropylammonium hydroxide (TPAOH, 25% aqueous solution), tetraethyl orthosilicate (TEOS), and distilled water were mixed to obtain a clear suspension (9TPAOH:25SiO 2 :480H 2 O, molar composition). The above mixture was stirred at room temperature to hydrolyze TEOS for 4 h, followed by being transferred into an autoclave and crystallized at 170 • C for 72 h. The obtained product was filtered, washed, and dried at 100 • C overnight, followed by calcination at 550 • C for 4 h in air.
The silicalite-1-supported chromium oxide catalysts were prepared through an incipient wetness method employing Cr(NO 3 ) 3 ·9H 2 O as the precursor. The impregnated samples were dried at 100 • C overnight, followed by calcination in air at 600 • C for 6 h. The obtained catalysts were designated as x%Cr/S-1, where x% represents the weight percentage of Cr in the catalysts. For the purpose of comparison, the 3%Cr/SBA catalyst (3%Cr) was prepared in the same way using SBA-15 as the support. SBA-15 was prepared according to the literature [52].

Catalyst Characterization
X-ray diffraction (XRD) measurements were performed with a D2 PHASER X-ray diffractometer (Brucker, Madison, WI, USA) at 40 mA and 40 kV. Small-angle X-ray scattering (SAXS) measurements were performed with a Nanostar U SAXS system (Brucker, Madison, WI, USA) using Cu Kα radiation at 35 mA and 40 kV. The surface areas and pore volumes of the catalysts were measured by N 2 adsorption on a Tristar 3000 instrument (Micromeritics, Atlanta, GA, USA). The HAADF-STEM images and elemental mapping were acquired with a Tecnai G 2 F20 S-TWIN instrument (FEI, Hillsboro, TX, USA). Diffuse reflectance ultraviolet-visible (UV-vis) spectra were collected with a Lambda 650S spectrometer (Perkin-Elmer, Waltham, MA, USA). Raman spectra were collected with an XploRA spectrometer (HORIBA Jobin Yvon, Paris, France). The exciting light wavelength was 532 nm. X-ray photoelectron spectroscopy (XPS) measurements were performed on a PHI 5000C spectrometer with Mg Kα radiation (Perkin-Elmer, Waltham, MA, USA). 29 Si MAS NMR characterization was carried out with an AVANCE III 400WB instrument (Brucker, Rheinstetten, Germany). To analyze the amount of deposited coke on the catalyst after reaction, thermogravimetric (TG) analysis was carried out in flowing air with a TGA8000 apparatus (Perkin-Elmer, Waltham, MA, USA).
Temperature-programmed desorption of NH 3 (NH 3 -TPD), temperature-programmed reduction of H 2 (H 2 -TPR), and temperature-programmed oxidation (TPO) characterizations were performed on an AutoChem II instrument (Micromeritics, Atlanta, GA, USA). We pretreated 0.1 g of sample (40-60 mesh) in situ at 550 • C in N 2 flow for 1 h before the measurement. In the case of NH 3 -TPD experiment, the temperature was cooled to 80 • C, and the flow was changed to 10 vol.% NH 3 /He (30 mL/min) and maintained at 80 • C for 2 h, followed by being swept with He (30 mL/min) for 2 h. Then, the sample was heated in He (30 mL/min) to 600 • C at a ramp rate of 10 • C/min. In the case of H 2 -TPR experiment, the temperature was cooled to 100 • C, and the flow was switched to 10% H 2 /Ar (30 mL/min), followed by heating to 650 • C at a ramp rate of 10 • C/min. In the case of TPO experiment, the temperature was cooled to 150 • C, and the flow was changed to 3% O 2 /He (30 mL/min), followed by heating to 650 • C at a ramp rate of 10 • C/min. Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) spectra were recorded at 300 • C on a Nicolet 6700 spectrometer (ThermoFisher, Waltham, MA, USA) equipped with an MCT detector and a heating accessory. The sample was pretreated in situ in flowing He (30 mL/min) at 450 • C for 1 h before the measurement.

Catalytic Evaluation
Catalytic performance in CO 2 assisted dehydrogenation of isobutane was carried at 570 • C in a fixed-bed quartz tube reactor (6 mm internal diameter) under ambient pressure. The catalyst loading was 0.1 g. The catalyst (40-60 mesh) was activated in situ in N 2 flow at 570 • C for 1 h before the reaction. The feed gas contained 50 vol.% CO 2 and 50 vol.% isobutane (2.9 mL/min of isobutane), i.e., the weight hourly space velocity of 4.1 h −1 for isobutane. In the case of isobutane dehydrogenation in the absence of CO 2 , CO 2 was replaced by N 2 , while keeping the other reaction conditions the same. In the case of reverse water-gas shift reaction, the feed gas contained 50 vol.% H 2 and 50 vol.% CO 2 (2.9 mL/min of CO 2 ), while keeping the other reaction conditions the same. The hydrocarbon products were in-situ analyzed with a GC (FID and HP-AL/S capillary column). The column temperature was 100 • C. CO and CO 2 were in-situ analyzed with another GC (TCD and carbon molecular sieve 601 packed column). The column temperature was 70 • C. The conversion and selectivity were calculated using the standard normalization method.

Conclusions
In this work, we have explored the silicalite-1-supported chromium oxide catalysts for isobutane dehydrogenation assisted by CO 2 . This family of catalysts is shown to be effective for the reaction. A good correlation between the initial activity of the Cr/S-1 catalysts and content of Cr 6+ in the fresh catalysts is established. CO 2 promotes the isobutane dehydrogenation via a redox mechanism and two-step pathway. The best 3%Cr/silicalite-1 catalyst gives an isobutane conversion of 36.5% with 71.2% isobutene selectivity. The greater initial activity observed for 3%Cr/S-1 than 3%Cr/SBA is attributed to a higher content of Cr 6+ in the fresh 3%Cr/S-1 catalyst, and the better catalytic stability for the former catalyst than the latter one is associated with a higher content of Cr 6+ retained on the former catalyst during the reaction. This finding could be related to the abundant nest silanol groups present on the silicalite-1 support.