Modification by SiO 2 of Alumina Support for Light Alkane Dehydrogenation Catalysts

Abstract: Due to the continuously rising demand for C3–C5 olefins it is important to improve the performance of catalysts for dehydrogenation of light alkanes. In this work the effect of modification by SiO2 on the properties of the alumina support and the chromia-alumina catalyst was studied. SiO2 was introduced by impregnation of the support with a silica sol. To characterize the supports and the catalysts the following techniques were used: low-temperature nitrogen adsorption; IR-spectroscopy; magic angle spinning 29Si nuclear magnetic resonance; temperature programmed desorption and reduction; UV-Vis-, Ramanand electron paramagnetic resonance (EPR)-spectroscopy. It was shown that the modifier in amounts of 2.5–7.5 wt % distributed on the support surface in the form of SiOx-islands diminishes the interaction between the alumina support and the chromate ions (precursor of the active component). As a result, polychromates are the compounds predominantly stabilized on the surface of the modified support; under thermal activation of the catalyst and are reduced to the amorphous Cr2O3. This in turn leads to an increase in the activity of the catalyst in the dehydrogenation of isobutane.


Introduction
Catalysts with surface chromium species as an active component have large practical importance and are widely used in industrial organic synthesis.Microspherical chromia-alumina catalysts for alkane dehydrogenation in terms of consumption hold the leading position in the petrochemical industry of Russia [1].This process is intended for production of C 3 -C 5 -olefin which is the raw material for synthesis of rubbers, plastics, synthetic films and filaments, high-octane components of fuel, etc. [2].
There are two methods of microspherical chromia-alumina catalysts preparation: the mixing of precursors of the support and active component followed by spray drying of the slurry obtained [3]; and the impregnation of the support by the solutions of the active component and the promoter.The catalysts obtained by the first method have a low mechanical strength under the conditions of a fluidized bed and the continuous circulation between the reactor and regenerator [4].The method of support impregnation seems the most promising.Earlier a method of a gibbsite transformation to boehmite (precursor of γ-Al 2 O 3 ) in the volume of the microgranule by a sequential thermal and hydrothermal treatment was developed [5].The support obtained that way has a high mechanical strength, middle porosity, and a low surface acidity.
Continuously rising demand for light olefins [2,6] as well as the increase in the cost of natural energy resources requires high active and selective catalysts.It is possible to improve the catalyst performance by introduction of the modifiers.The modifiers of chromia dehydrogenation catalyst SiO 2 [7], ZrO 2 [8][9][10], lanthanum [10,11], tin [7,12] and cerium [13] are usually used.Among these modifiers SiO 2 is the most accessible and inexpensive; its use is broadly described in the patent literature [14].However, the mechanism of its positive effect on the performance of the dehydrogenation catalyst has not been described in detail.Therefore, in this article we studied the effect of modification of the structure and properties of the novel alumina support and chromia-alumina catalyst by SiO 2 .

Thermal Treatment of Boehmite Precursor
The microspherical support was used as a precursor of alumina (Figure 1).It was obtained according to the Scheme 1 [5]: performance by introduction of the modifiers.The modifiers of chromia dehydrogenation catalyst SiO2 [7], ZrO2 [8][9][10], lanthanum [10,11], tin [7,12] and cerium [13] are usually used.Among these modifiers SiO2 is the most accessible and inexpensive; its use is broadly described in the patent literature [14].However, the mechanism of its positive effect on the performance of the dehydrogenation catalyst has not been described in detail.Therefore, in this article we studied the effect of modification of the structure and properties of the novel alumina support and chromiaalumina catalyst by SiO2.

Thermal Treatment of Boehmite Precursor
The microspherical support was used as a precursor of alumina (Figure 1).It was obtained according to the Scheme 1 [5]:  According to the X-ray diffraction (Figure 2), the alumina precursor is a well-crystallized boehmite with crystallite size D(020) = 44 nm and D(120) = 47 nm.A high degree of boehmite crystallinity causes its small specific surface area (SBET) and pore volume (Vp) = 27 m 2 •g −1 and 0.07 cm 3 •g −1 respectively (Table 1).
The choice of the boehmite thermal treatment temperature was made based on changes in the specific surface area and the number of strong acid sites (with the energy of ammonia desorption Edes.NH3 more than 150 kJ•mol −1 ) obtained in alumina (Table 1).It is known that strong acid sites are active in hydrocarbon cracking reactions [15,16].
Thermal treatment of the precursor at 750 °C leads to the formation of mesopores (Figure 3), an increase in the specific surface area up to 92 m 2 •g −1 and in the pore volume up to 0.27 cm 3 •g −1 as a result of dehydration and the phase transition of boehmite to γ-Al2O3 (Figure 2).An increase of performance by introduction of the modifiers.The modifiers of chromia dehydrogenation catalyst SiO2 [7], ZrO2 [8][9][10], lanthanum [10,11], tin [7,12] and cerium [13] are usually used.Among these modifiers SiO2 is the most accessible and inexpensive; its use is broadly described in the patent literature [14].However, the mechanism of its positive effect on the performance of the dehydrogenation catalyst has not been described in detail.Therefore, in this article we studied the effect of modification of the structure and properties of the novel alumina support and chromiaalumina catalyst by SiO2.

Thermal Treatment of Boehmite Precursor
The microspherical support was used as a precursor of alumina (Figure 1).It was obtained according to the Scheme 1 [5]: According to the X-ray diffraction (Figure 2), the alumina precursor is a well-crystallized boehmite with crystallite size D (020) = 44 nm and D (120) = 47 nm.A high degree of boehmite crystallinity causes its small specific surface area (S BET ) and pore volume (V p ) = 27 m 2 •g −1 and 0.07 cm 3 •g −1 respectively (Table 1).
The choice of the boehmite thermal treatment temperature was made based on changes in the specific surface area and the number of strong acid sites (with the energy of ammonia desorption E des.NH 3 more than 150 kJ•mol −1 ) obtained in alumina (Table 1).It is known that strong acid sites are active in hydrocarbon cracking reactions [15,16].
Thermal treatment of the precursor at 750 • C leads to the formation of mesopores (Figure 3), an increase in the specific surface area up to 92 m 2 •g −1 and in the pore volume up to 0.27 cm 3 •g −1 as a result of dehydration and the phase transition of boehmite to γ-Al 2 O 3 (Figure 2).An increase of treatment temperature to 850 • C resulted in the decrease of the specific surface area to 62 m 2 •g −1 due to pore enlargement (Table 1).It is also accompanied by a decrease in the number of strong acid sites due to the reduction of the support surface, its dehydroxylation, as well as the transformation of the crystal structure of the support.
We believe the optimal temperature of boehmite precursor treatment is 800 • C. γ-Al 2 O 3 support obtained under those conditions has a low concentration of strong acid sites (10.0 µmol•g −1 ) and a medium surface area (S BET = 83 m 2 •g −1 ).A decrease of the temperature to 750 • C is accompanied by a 1.6-fold increase in the concentration of strong acid sites (up to 16.2 µmol•g −1 ) and an increase of surface area by only 9 m 2 •g −1 (up to 92 m 2 •g −1 ).An increase of the temperature to 850 • C leads to a decrease in the specific surface area by 21 m 2 •g −1 and a decrease in the concentration of strong acid sites by just 1.5 µmol•g −1 (Table 1).

SiO 2 -Modification of the Support
SiO 2 -modification of the support was carried out by its impregnation with an aqueous SiO 2 -sol and subsequent thermal treatment at 800 • C. Introduction of silica in an amount of 2.5-7.5 wt % does not affect the phase composition of the support.In the X-ray diffraction patterns of all the samples (not shown in the current paper) the characteristic peaks of only γ-Al 2 O 3 were identified.The absence of changes in the crystalline structure of the supports leads to the conclusion that the SiO 2 is localized on the support's surface.In the IR-spectrum of the SiO 2 -modified support the absorption band of Si-OH at 3741 cm −1 [17,18] appears with a decrease of the intensity of the bands at 3751 and 3771 cm −1 (Al-OH bond vibrations [19,20]) (Figure 4).This also indicates that the SiO 2 is distributed on the support surface.It is also accompanied by a decrease in the number of strong acid sites due to the reduction of the support surface, its dehydroxylation, as well as the transformation of the crystal structure of the support.
We believe the optimal temperature of boehmite precursor treatment is 800 °C.γ-Al2O3 support obtained under those conditions has a low concentration of strong acid sites (10.0 μmol•g −1 ) and a medium surface area (SBET = 83 m 2 •g −1 ).A decrease of the temperature to 750 °C is accompanied by a 1.6-fold increase in the concentration of strong acid sites (up to 16.2 μmol•g −1 ) and an increase of surface area by only 9 m 2 •g −1 (up to 92 m 2 •g −1 ).An increase of the temperature to 850 °C leads to a decrease in the specific surface area by 21 m 2 •g −1 and a decrease in the concentration of strong acid sites by just 1.5 μmol•g −1 (Table 1).

SiO2-Modification of the Support
SiO2-modification of the support was carried out by its impregnation with an aqueous SiO2-sol and subsequent thermal treatment at 800 °C.Introduction of silica in an amount of 2.5-7.5 wt% does not affect the phase composition of the support.In the X-ray diffraction patterns of all the samples (not shown in the current paper) the characteristic peaks of only γ-Al2O3 were identified.The absence of changes in the crystalline structure of the supports leads to the conclusion that the SiO2 is localized on the support's surface.In the IR-spectrum of the SiO2-modified support the absorption band of Si-OH at 3741 cm −1 [17,18] appears with a decrease of the intensity of the bands at 3751 and 3771 cm −1 (Al-OH bond vibrations [19,20]) (Figure 4).This also indicates that the SiO2 is distributed on the support surface.On the surface of the modified supports silica is distributed in the form of SiOx-fragments Si(OSi)3O and Si(OSi)4.This is indicated by the occurrence of the signals at −101 and −125 ppm on the 29 Si Magic Angle Spinning Nuclear Magnetic Resonance (MAS NMR) spectra (Figure 5) [18,[21][22][23][24].
The distribution of the modifier on the surface leads to changes in the porous system of the supports.The volume of the pores with diameter less than 10 nm decreases from 0.08 to 0.06 cm 3 •g −1 .SiOx-islands form additional porosity in the range of diameters 10-30 nm; the volume of these pores On the surface of the modified supports silica is distributed in the form of SiO x -fragments Si(OSi) 3 O and Si(OSi) 4 .This is indicated by the occurrence of the signals at −101 and −125 ppm on the 29 Si Magic Angle Spinning Nuclear Magnetic Resonance (MAS NMR) spectra (Figure 5) [18,[21][22][23][24].
The distribution of the modifier on the surface leads to changes in the porous system of the supports.The volume of the pores with diameter less than 10 nm decreases from 0.08 to 0.06 cm 3 •g −1 .SiO x -islands form additional porosity in the range of diameters 10-30 nm; the volume of these pores increases from 0.13 to 0.16 cm 3 •g −1 (Figure 6).At the same time the specific surface area does not change significantly (Table 2).increases from 0.13 to 0.16 cm 3 •g −1 (Figure 6).At the same time the specific surface area does not change significantly (Table 2).increases from 0.13 to 0.16 cm 3 •g −1 (Figure 6).At the same time the specific surface area does not change significantly (Table 2).
-     SiO 2 -modification of the support causes changes in the surface acidity (Figure 7, Table 3).It increases from 102.3 to 109.0-125.3µmol•g −1 due to the formation of weak (E des.NH 3 < 100 kJ•mol −1 ) and medium (E des.NH 3 = 100-150 kJ•mol −1 ) acid sites (Table 3).This is indicated by the growth of the NH 3 -TPD (temperature programmed desorption) profiles in the range 160-350 • C (Figure 7).Along with the formation of an additional amount of the weak and medium sites the concentration of strong acid sites (with E des.NH 3 > 150 kJ•mol −1 ) decreases (Table 3).This is indicated by the decrease in intensity of the high temperature component (400-500 • C) of the NH 3 -TPD profile.The maximal temperature of ammonia desorption shifts to lower values (Figure 7) indicating a decrease in strength of the acid sites.SiO2-modification of the support causes changes in the surface acidity (Figure 7, Table 3).It increases from 102.3 to 109.0-125.3μmol•g −1 due to the formation of weak (Edes.NH3 <100 kJ•mol −1 ) and medium (Edes.NH3 = 100-150 kJ•mol −1 ) acid sites (Table 3).This is indicated by the growth of the NH3-TPD (temperature programmed desorption) profiles in the range 160-350 °C (Figure 7).Along with the formation of an additional amount of the weak and medium sites the concentration of strong acid sites (with Edes.NH3 >150 kJ•mol −1 ) decreases (Table 3).This is indicated by the decrease in intensity of the high temperature component (400-500 °C) of the NH3-TPD profile.The maximal temperature of ammonia desorption shifts to lower values (Figure 7) indicating a decrease in strength of the acid sites.It is likely that the acid sites formed are hydroxyl groups bonded to the silicon atoms and the Lewis acid sites of various structures [21,24].For example, Iengo et al. [21] suggested, that the new  It is likely that the acid sites formed are hydroxyl groups bonded to the silicon atoms and the Lewis acid sites of various structures [21,24].For example, Iengo et al. [21] suggested, that the new acid sites are formed at the border of the SiO x -islands, where aluminum and silicon are arranged relative to each other in such a way, that the aluminum atoms are positively charged and partially neutralized by the negative charges of the silicates.The decrease in number of the strong acid sites upon modification is probably due to them being covered by SiO x -islands.

Composition, Crystal and Pore Structure, Acidity of Catalysts
On the basis of the initial alumina support it was established that surface concentration 10 atoms•nm −2 provides both high activity and the absence of crystalline α-Cr 2 O 3 in the catalyst (Figure 8), which accelerate catalyst sintering and decreases its thermal stability (Table 4).
On the basis of the initial alumina support it was established that surface concentration 10 atoms•nm −2 provides both high activity and the absence of crystalline α-Cr2O3 in the catalyst (Figure 8), which accelerate catalyst sintering and decreases its thermal stability (Table 4).In the synthesized catalysts an optimum mass ratio of chromium to potassium was used (Cr/K = 8) which was established by Kataev [25] for alumina supports obtained by the sequential thermal-hydrothermal treatment of gibbsite.Potassium as a promoter performs several functions in chromia-alumina catalysts.It neutralizes the strong acid sites [15] and thus decreases the activity of the catalyst in hydrocarbon cracking reactions (Table 4).Potassium in catalyst forms potassium chromates at the expense of the low-active crystalline Cr2O3 phase.These chromates are reduced to catalytically active Cr(III) phase in dehydrogenation conditions [15,26].However, an excessive  In the synthesized catalysts an optimum mass ratio of chromium to potassium was used (Cr/K = 8) which was established by Kataev [25] for alumina supports obtained by the sequential thermal-hydrothermal treatment of gibbsite.Potassium as a promoter performs several functions in chromia-alumina catalysts.It neutralizes the strong acid sites [15] and thus decreases the activity of the catalyst in hydrocarbon cracking reactions (Table 4).Potassium in catalyst forms potassium chromates at the expense of the low-active crystalline Cr 2 O 3 phase.These chromates are reduced to catalytically active Cr(III) phase in dehydrogenation conditions [15,26].However, an excessive amount of potassium poisons the catalyst due to the coverage of the active Cr(III) sites [26] and neutralization of the weak and medium acid sites which adsorb alkane molecules [15].Optimal potassium content depends not only on the surface area and acidity of the support but also on the concentration of chromium because potassium as aluminate and chromates is distributed both on the open areas of alumina and on the Cr 2 O 3 particles [26].
The chromia-alumina catalysts with a surface concentration of chromium 10 atoms•nm −2 were synthesized on the SiO 2 -modified supports (Table 5).As a result of the chromium and potassium distribution in the pores of initial and SiO 2 -modified support (Figure 9), specific surface area and pore volume decrease slightly-by 6-17 m 2 •g −1 and 0.04-0.06cm 3 •g −1 respectively (Tables 1 and 5).
amount of potassium poisons the catalyst due to the coverage of the active Cr(III) sites [26] and neutralization of the weak and medium acid sites which adsorb alkane molecules [15].Optimal potassium content depends not only on the surface area and acidity of the support but also on the concentration of chromium because potassium as aluminate and chromates is distributed both on the open areas of alumina and on the Cr2O3 particles [26].
The chromia-alumina catalysts with a surface concentration of chromium 10 atoms•nm −2 were synthesized on the SiO2-modified supports (Table 5).As a result of the chromium and potassium distribution in the pores of initial and SiO2-modified support (Figure 9), specific surface area and pore volume decrease slightly-by 6-17 m 2 •g −1 and 0.04-0.06cm 3 •g −1 respectively (Tables 1 and 5).Introduction of 6.5 wt% chromium and 0.8 wt% potassium to supports decreases the number of acid sites by 15.7-29.9μmol•g −1 (Tables 3 and 6).SiO2 affects the acidity of the catalysts similar to those of the supports: introduction of 2.5-7.5 wt% SiO2 increases the total acidity of catalyst due to the formation of an additional weak and medium acid sites; it also decreases the number of strong acid sites.Introduction of 6.5 wt % chromium and 0.8 wt % potassium to supports decreases the number of acid sites by 15.7-29.9µmol•g −1 (Tables 3 and 6).SiO 2 affects the acidity of the catalysts similar to those of the supports: introduction of 2.5-7.5 wt % SiO 2 increases the total acidity of catalyst due to the formation of an additional weak and medium acid sites; it also decreases the number of strong acid sites.

Active Component of Catalysts
Formation on the support surface of SiO x -islands results in changes in the distribution of the active component.Modification of the support by 2.5-7.5 wt % SiO 2 leads to a decrease of Cr(VI) content in the catalysts based on it.
By diffuse reflectance UV-Vis spectroscopy it was found that chromium deposition on the SiO 2 -modified supports results in an increase of Cr(III) content in comparison to the initial support.It can be seen by the increase in intensity of the signal at ~17,000 cm −1 (Figure 10) which corresponds to an electronic transition 4 A 2g → 4 T 2g in the Cr(III) oct ion [26,27].Increasing the Cr(III) content is consistent with the data of the iodometric titration, showing that the concentration of Cr(VI) (calculated as CrO 3 ) decreases from 2.5 to 1.4 wt % with an increase of SiO 2 content from 0 to 7.5 wt % (Table 5).UV-Vis spectra also demonstrate a shift of the Cr(VI) signal at ~27,000 cm −1 [26,28,29] to the short-wavelength region and an increase in intensity of the complex Cr(III)-Cr(VI) signal at ~22,000 cm −1 [26-29] (Figure 10).In combination these indicate a decrease in symmetry of surface chromates due to oligomerization of Cr(VI) compounds [26].By diffuse reflectance UV-Vis spectroscopy it was found that chromium deposition on the SiO2modified supports results in an increase of Cr(III) content in comparison to the initial support.It can be seen by the increase in intensity of the signal at ~17,000 cm −1 (Figure 10) which corresponds to an electronic transition 4 A2g→ 4 T2g in the Cr(III)oct ion [26,27].Increasing the Cr(III) content is consistent with the data of the iodometric titration, showing that the concentration of Cr(VI) (calculated as CrO3) decreases from 2.5 to 1.4 wt% with an increase of SiO2 content from 0 to 7.5 wt% (Table 5).UV-Vis spectra also demonstrate a shift of the Cr(VI) signal at ~27,000 cm −1 [26,28,29] to the short-wavelength region and an increase in intensity of the complex Cr(III)-Cr(VI) signal at ~22,000 cm −1 [26-29] (Figure 10).In combination these indicate a decrease in symmetry of surface chromates due to oligomerization of Cr(VI) compounds [26].The increase in Cr(III) content observed by UV-Vis spectroscopy and the oligomerization of surface Cr(VI) compounds are consistent with the Raman-spectroscopy data.Here, the increase in SiO2 content results in the growth in intensity of the signal at ~550 cm −1 , corresponding to the vibrations of the Cr(III)oct-O bond [26,30,31]; the intensity of hydrated dichromate signals at 906 and 948 cm −1 [28,30] decreases; the typical signals of dehydrated polychromates at 850-900 and 1000 cm −1 [28,[30][31][32] also appear (Figure 11).
Decreasing Cr(VI) content and the simultaneous oligomerization of chromates indicates a decrease of stabilization of the chromates by the support surface through X-O-Cr (X = Al, Si) bonds.This is also supported by the temperature-programmed reduction of catalysts.All H2-TPR profiles are decomposed into 3-4 Gaussian components which correspond to the chromates with different degrees of binding to the surface.The higher the temperature of chromate reduction the greater is the number of X-O-Cr bonds per one chromium atom [26,33].Therefore, the low-temperature components (at 300-450 °C and 400-450 °C) most likely correspond to polychromates, and the hightemperature components (at 450-600 °C and 500-650 °C) correspond to mono-and dichromates.The The increase in Cr(III) content observed by UV-Vis spectroscopy and the oligomerization of surface Cr(VI) compounds are consistent with the Raman-spectroscopy data.Here, the increase in SiO 2 content results in the growth in intensity of the signal at ~550 cm −1 , corresponding to the vibrations of the Cr(III) oct -O bond [26,30,31]; the intensity of hydrated dichromate signals at 906 and 948 cm −1 [28,30] decreases; the typical signals of dehydrated polychromates at 850-900 and 1000 cm −1 [28,[30][31][32]] also appear (Figure 11).
Decreasing Cr(VI) content and the simultaneous oligomerization of chromates indicates a decrease of stabilization of the chromates by the support surface through X-O-Cr (X = Al, Si) bonds.This is also supported by the temperature-programmed reduction of catalysts.All H 2 -TPR profiles are decomposed into 3-4 Gaussian components which correspond to the chromates with different degrees of binding to the surface.The higher the temperature of chromate reduction the greater is the number of X-O-Cr bonds per one chromium atom [26,33].Therefore, the low-temperature components (at 300-450 • C and 400-450 • C) most likely correspond to polychromates, and the high-temperature components (at 450-600 • C and 500-650 • C) correspond to mono-and dichromates.The introduction of 2.5-7.5 wt % SiO 2 leads to the redistribution of the hydrogen consumption temperature: the low temperature component of the H 2 -TPR profile (in the region 300-400 • C) increases and the high temperature component (in the range 500-600 • C) decreases.The Gaussian component of the H 2 -TPR profile in the region 550-600 • C, corresponding to monochromates with the highest degree of interaction with the support surface, disappears.The maximum on the H 2 -TPR profiles sequentially shifts from 436 to 419 • C (Figure 12).
Catalysts 2016, 6, 162 10 of 20 introduction of 2.5-7.5 wt% SiO2 leads to the redistribution of the hydrogen consumption temperature: the low temperature component of the H2-TPR profile (in the region 300-400 °C) increases and the high temperature component (in the range 500-600 °C) decreases.The Gaussian component of the H2-TPR profile in the region 550-600 °C, corresponding to monochromates with the highest degree of interaction with the support surface, disappears.The maximum on the H2-TPR profiles sequentially shifts from 436 to 419 °C (Figure 12).The decrease in interaction between the support and the active component is due to the changes in hydroxyl cover of the support.The isoelectronic point (IEP) of alumina is 7.2-8.6,and the IEP of silica is 2.0-3.9 [34][35][36].This is caused by the predominance of the following reactions on the surface of the hydrated supports [34,36]: according to Reaction (2) the part of the inner Helmholtz plane becomes negatively charged which results in the electrostatic attraction between support surface and chromate ions diminishing.
A decrease in the number of Cr-O-X (X = Al, Si) bonds leads to the stabilization of the polychromate ions and polynuclear [-Cr(III)-O-Cr(III)-] n ions (also called amorphous Cr 2 O 3 [26] or Cr 2 O 3 -clusters [28]) (Scheme 4) by the SiO 2 -modified support surface after thermal activation of the catalyst.A similar effect of SiO 2 has also been reported by Weckhuysen et al. [29]; at the transition from Al 2 O 3 to aluminosilicate (with 40 wt % SiO 2 ) the chromate/dichromate band intensity ratio on the UV-Vis spectra of the samples calcined at 550 • C decreased from ∞ to 2.18.Weckhuysen et al. [38] reported that in SiO 2 with 0.2 wt % Cr calcined at 720 • C, Cr(III) ions were identified by UV-Vis spectroscopy, whereas for the same amount of chromium on alumina, Cr(III) ions were not observed.Increasing interaction between Cr(III) ions was also confirmed by EPR-spectroscopy.Upon introduction of SiO2 the intensity of the δ-signal component at g = 3.8-4.0,corresponding to isolated Cr(III) ions [39], decreases while the intensity of β-signal with g = 1.98 and ΔH = 1400-1600 G, corresponding to magnetically concentrated Cr(III) ions [28,39], increases (Figure 13).Increasing interaction between Cr(III) ions was also confirmed by EPR-spectroscopy.Upon introduction of SiO 2 the intensity of the δ-signal component at g = 3.8-4.0,corresponding to isolated Cr(III) ions [39], decreases while the intensity of β-signal with g = 1.98 and ∆H = 1400-1600 G, corresponding to magnetically concentrated Cr(III) ions [28,39], increases (Figure 13).Intensity (a.u.) Figure 13.Electron paramagnetic resonance (EPR) spectra of catalysts.

Isobutane Dehydrogenation Performance of Catalysts
Chromia-alumina catalysts were tested in isobutane dehydrogenation at 570 °C.The curves of dehydrogenation and cracking rate versus time on stream are shown on Figure 14a,b, respectively.All the curves on Figure 14a show the maxima as a function of time on stream.In all the cases at the initial stage of the dehydrogenation cycle (for the first 30-55 min) an increase of catalyst dehydrogenation activity by 8%-9.5% is observed (Figure 14a); this is the catalyst development period.As we know [40,41] during catalyst development the surface stabilized Cr(VI) compounds are reduced by hydrocarbons to additional catalytically active Cr(III) compounds.Decrease of the dehydrogenation rate at the final stage of the cycle is due to the catalyst coking.The decrease of the cracking rate during the first 30 min (Figure 14b) is also caused by covering of the acid sites by coke deposits [15].

Isobutane Dehydrogenation Performance of Catalysts
Chromia-alumina catalysts were tested in isobutane dehydrogenation at 570 • C. The curves of dehydrogenation and cracking rate versus time on stream are shown on Figure 14a,b, respectively.All the curves on Figure 14a show the maxima as a function of time on stream.In all the cases at the initial stage of the dehydrogenation cycle (for the first 30-55 min) an increase of catalyst dehydrogenation activity by 8%-9.5% is observed (Figure 14a); this is the catalyst development period.As we know [40,41] during catalyst development the surface stabilized Cr(VI) compounds are reduced by hydrocarbons to additional catalytically active Cr(III) compounds.Decrease of the dehydrogenation rate at the final stage of the cycle is due to the catalyst coking.The decrease of the cracking rate during the first 30 min (Figure 14b) is also caused by covering of the acid sites by coke deposits [15].
The increase in the content of amorphous Cr 2 O 3 phase, which is the most active in the dehydrogenation of light alkanes [26], and the formation of polychromates which also reduce to amorphous Cr 2 O 3 , causes an increase in dehydrogenation activity of catalysts based on SiO 2 -modified supports (Figure 14a).The average dehydrogenation activity increases by 10%-14%-from 2.9 to 3.2-3.3µmol C 4 H 10 •g cat −1 •s −1 .At the same time the catalyst development period diminishes due to a decrease in the Cr(VI) content.The rise in dehydrogenation activity with the increase in the polynuclear Cr(III) ions content is consistent with the results of the kinetic studies by Airaksinen et al. [42] and Carra et al [43].According to these works the dehydrogenation of one molecule of butane proceeds at two active sites.
A decrease in the number of strong acid sites upon SiO 2 -modification of support leads to a decrease in cracking rate on catalysts with 2.5-7.5 wt % SiO 2 (Figure 14b).The average cracking rate decreases by 18%-34%-from 0.29 to 0.19-0.24µmol The increase in the content of amorphous Cr2O3 phase, which is the most active in the dehydrogenation of light alkanes [26], and the formation of polychromates which also reduce to amorphous Cr2O3, causes an increase in dehydrogenation activity of catalysts based on SiO2-modified supports (Figure 14a).The average dehydrogenation activity increases by 10%-14%-from 2.9 to 3.2-3.3μmol C4H10•gcat −1 •s −1 .At the same time the catalyst development period diminishes due to a decrease in the Cr(VI) content.The rise in dehydrogenation activity with the increase in the polynuclear Cr(III) ions content is consistent with the results of the kinetic studies by Airaksinen et al. [42] and Carra et al [43].According to these works the dehydrogenation of one molecule of butane proceeds at two active sites.

Preparation of Supports and Catalysts
Microspherical (40-200 µm) boehmite support prepared by consecutive thermal and hydrothermal treatment of aluminum trihydroxide (GD00 grade, produced by Bogoslovsk Aluminum Smelter, Krasnoturyinsk city, Russia) under industrial conditions [11] in an autoclave for 3 h at 195 • C (Chemical Plant Karpov, Mendeleyevsk city, Russia), was used for the synthesis of alumina supports and chromia-alumina catalysts.According to thermogravimetry the phase composition of the support precursor is the following: boehmite-96.0wt %, gibbsite-3.0 wt%, Al 2 O 3 -1.0wt %.SiO 2 was introduced into the support by incipient wetness impregnation with a water silica sol ("Leiksil" brand produced by "Compas" Scientific and Technical Center, Kazan city, Russia).
The catalyst was prepared by incipient wetness impregnation of the support with an aqueous solution of chromic acid and potassium carbonate, followed by drying under vacuum.Then the catalyst was activated by thermal treatment in a muffle furnace at 750 • C for 4 h.

Characterization of Supports and Catalysts
Thermogravimetric analysis was performed with a STA-449C (Netzsch, Selb, Germany) combined thermogravimetric and differential scanning calorimetric (DSC) analyzer coupled with an Aeolos QMS 403 quadruple mass spectrometer (Netzsch, Selb, Germany) in a temperature range of 30-1000 • C at a heating rate of 10 • C. min −1 in a flow of argon.The concentrations of aluminum hydroxide phases were calculated from the amount of water released in their dehydration.
The elemental composition of the catalysts was determined by X-ray fluorescence spectroscopy on a Clever C31 instrument (ELERAN, Elektrostal, Russia).
Powder X-ray diffraction measurements were carried out using a DRON-2 diffractometer (Burevestnik, Saint Petersburg, Russia).The patterns were obtained using CuKα radiation and graphite monochromator (λ = 1.54187Å) at 30 kV and 15 mA.The identification of different crystalline phases in the samples was performed by comparing the data with the Joint Committee for Powder Diffraction Standards (JCPDS) files.The crystallite size of the boehmite phase was calculated using the Selyakov-Scherrer Equation.The error in determining the crystallite size was 10%.
Specific surface (S sp ) and pore volume (V p ) of samples were determined from the N 2 physisorption measurements at 196 • C using an universal Autosorb-iQ analyzer (Quantachrome, Boynton Beach, FL, USA).Prior to measurement, the sample was outgassed for 1 h at 150 • C (for boehmite precursor) or for 3 h at 300 • C (for alumina supports and catalysts).S sp and V p were calculated according to the Brunauer-Emmett-Teller method.The pore diameter distribution was calculated by the desorption branch of isotherm using the standard Barrett-Joyner-Halenda method.
The 29 Si MAS NMR spectra of supports were recorded at room temperature on an Avance 400 spectrometer (Bruker, Ettlingen, Germany) operating at frequencies of 79.5 MHz with spectral resolution 48.83 Hz.The sample rotation frequency was 5 kHz.
The IR spectra of supports were measured on a VERTEX 70 (Bruker, Ettlingen, Germany) instrument fitted with a mercury−cadmium−telluride detector.The measurements were done in transmission mode using a Harrick high temperature cell.A background spectrum and the spectra were measured at 480 • C and a residual pressure of less than 10 −3 mbar with a resolution of 1 cm −1 and averaged by 128 scans.For the IR analysis, samples were prepared in a tablet-shape of 20 mg; optical density was 20 mg•cm −2 .
Hexavalent chromium concentration in the catalyst was determined by dissolution of Cr(VI) in sulfuric acid and subsequent volumetric titration with the iodometric method.
UV-Vis diffuse reflectance spectra of the catalysts were recorded using a V-650 spectrophotometer (Jasco, Tokyo, Japan) equipped with an integrating sphere ISV-722 (Jasco, Tokyo, Japan).A BaSO 4 plate was used as the reference.Spectra were recorded in the wavenumber range 12,500-50,000 cm −1 with the spectral resolution 2 nm.UV-Vis-spectra were deconvoluted into Gaussian bands to determine the positions and intensities of the bands' maximums.
Raman spectra of catalysts were recorded using a dispersion Raman-microspectrophotometer Nicolet Almega XR (Thermo Fisher Scientific, Waltham, MA, USA).The 532 nm line of a Nd-YAG laser was used as an excitation.The spectra were recorded in the wavenumber range 100-1100 cm −1 with the spectral resolution 2 cm −1 .Each spectrum was received by averaging 10 exposures on 10 s.
EPR measurements were made at the temperature of −196 • C on a RE-1306 EPR-spectrometer (Institute of Analytical Instrumentation of Russian Academy of Sciences, Saint Petersburg, Russia) with a working frequency 9.37 GHz and 100 kHz magnetic field modulation.Diphenylpicrylhydrazyl (g = 2.0036) was used as reference for g-value determination.
NH 3 -TPD and H 2 -TPR measurements were carried out on the ChemBET Pulsar TPR/TPD (Quantachrome, Boynton Beach, FL, USA).Before NH 3 -TPD measurement the sample was degassed at 600 • C for 2 h in a helium flow.The adsorption step is carried out in an ammonia flow at 100 • C for 30 min.Then the physically sorbed ammonia was removed with helium at 100 • C for 30 min and the sample was cooled to a room temperature in the helium flow.Temperature programmed desorption was performed from room temperature to 700 • C at a heating rate of 10 • C•min −1 .The strength of the acid sites was evaluated by the temperature of ammonia desorption [44].A temperature of 175 • C corresponds to ammonia desorption energy of 100 kJ•mol −1 and a temperature of 380 • C to the desorption energy 150 kJ•mol −1 .Acid sites with ammonia desorption energy lower than 100 kJ•mol −1 were attributed to weak ones, while the sites with desorption energies of 100-150 kJ•mol −1 and higher than 150 kJ•mol −1 were attributed to medium and strong sites respectively.The number of weak, medium and strong acid sites was calculated from the area under the NH 3 -TPD profiles in the temperature ranges <175 • C, 175-380 • C and >380 • C respectively.
Before H 2 -TPR measurement the catalyst was heated to 650 • C and held at this temperature for 60 min in a flow of a gas mixture (5 vol% O 2 + 95 vol% N 2 ).Then the catalyst was cooled down to room temperature in the helium flow.Temperature programmed reduction was performed from room temperature to 700 • C at a heating rate of 10 • C•min −1 .NH 3 -TPD and H 2 -TPR profiles were Gauss fitted using the TPRWin software (version 3.52, Quantachrome, Boynton Beach, FL, USA, 2012).

Catalyst Testing
The catalysts were tested in the reaction of isobutane dehydrogenation in a steel fixed bed reactor of 10 mm internal diameter under atmospheric pressure.An amount of 2 g of fresh catalyst (sieve fraction 40-200 µm) was filled into the reactor.The catalyst was heated at 5 • C•min −1 to 650 • C in an air flow (60 mL•min −1 ) followed by flushing with air for 30 min at the same temperature.The catalyst was cooled in air to 570 • C and flushed with argon for 15 min at that temperature.Then a mixture of 30 vol% C 4 H 10 in Ar was fed at a rate of 60 mL•min −1 at the same reaction temperature.The reaction was run for 130 min followed by catalyst regeneration in an air flow for 60 min at 650 • C. The regeneration/dehydrogenation cycles were repeated three times.
The hydrocarbon composition of feed and reaction products were analyzed by gas chromatography on a GH-1000 instrument (Chromos, Dzerzhinsk, Russia) with a flame-ionization detector and a capillary VP-Alumina/KCl column (VICI Valco, Houston, TX, USA).The concentrations of H 2 , CH 4 , and CO were determined with the use of a column filled by 13× molecular sieves on a GH-1000 apparatus with a thermal conductivity detector.
Based on of the results of chromatographic analysis the rates of i-C 4 H 10 dehydrogenation and cracking of hydrocarbons were calculated using Equations ( 6) and (7)

Conclusions
The distribution of SiO 2 and its effect on the structure and acidity of the alumina support, as well as the effect of SiO 2 on the active component and the performance in isobutane dehydrogenation of chromia-alumina catalyst were investigated.It was shown that SiO 2 in an amount of 2.5-7.5 wt % is distributed on the surface of alumina in the form of Si(OSi) 4 and Si(OSi) 3 (O-) compounds.These SiO x -islands on the support produce additional porosity with a pore range 10-30 nm, as well as additional weak and medium acid sites.At the same time SiO 2 -modification causes a decrease in the number of strong acid sites.
Upon introduction of the active component, the SiO x -islands diminish the interaction of chromate ions with the support surface; as a consequence polychromates are the compounds predominantly stabilized on the support.During thermal activation of the catalyst these polychromates are reduced to a phase of amorphous Cr 2 O 3 .This results in the increase of the catalyst activity in the isobutane dehydrogenation.At the same time, the decrease in the number of strong acid sites in SiO 2 -modified catalysts leads to a diminution of its activity in hydrocarbon cracking reactions.

Figure 2 .Figure 3 .
Figure 2. X-ray diffraction patterns of precursor of alumina and alumina support, obtained at 800 °C.

Figure 3 .
Figure 3. Pore size distribution for precursor of alumina and alumina supports, obtained at 750 °C and 850 °C.

Figure 4 .
Figure 4. IR spectra of initial alumina support and a support containing 4.5 wt% SiO2.

Figure 4 .
Figure 4. IR spectra of initial alumina support and a support containing 4.5 wt % SiO 2 .

Figure 6 .
Figure 6.Pore size distribution for initial and SiO 2 -modified supports.

Figure 9 .
Figure 9. Pore size distribution for support with 4.5 wt% SiO2 and for catalyst based on it.

Figure 9 .
Figure 9. Pore size distribution for support with 4.5 wt % SiO 2 and for catalyst based on it.

Scheme 4 .
Scheme 4. Hydration of alumina and silica surface, their interaction with chromate ions.

Scheme 4 .
Scheme 4. Hydration of alumina and silica surface, their interaction with chromate ions.

Figure 14 .
Figure 14.Dehydrogenation rate (a) and cracking rate (b) versus time on stream.

Table 1 .
Integral parameters of the porous system and the acidity of alumina supports. 2

Table 1 .
Integral parameters of the porous system and the acidity of alumina supports.temperatureto 850 °C resulted in the decrease of the specific surface area to 62 m 2 •g −1 due to pore enlargement (Table1 treatment ).

Table 1 .
Integral parameters of the porous system and the acidity of alumina supports.

Table 2 .
Porous system parameters of supports.

Table 2 .
Porous system parameters of supports.

Table 2 .
Porous system parameters of supports.

Table 4 .
The composition of catalysts and their performance in isobutane dehydrogenation.

Table 4 .
The composition of catalysts and their performance in isobutane dehydrogenation.

Table 5 .
Composition, Cr(VI) content, specific surface area and pore volume of the initial and SiO 2 -modified catalysts.

Table 5 .
Composition, Cr(VI) content, specific surface area and pore volume of the initial and SiO2modified catalysts.

Table 6 .
Results from NH 3 -TPD measurement of catalysts.
Formation on the support surface of SiOx-islands results in changes in the distribution of the active component.Modification of the support by 2.5-7.5 wt% SiO2 leads to a decrease of Cr(VI) content in the catalysts based on it.
, respectively.CH 4 + C C 2 H 6 + C C 2 H 4 + C C 3 H 8 + C C 3 H 6 ) • V outlet 22400 • 3600 • 100 • m cat (7)where X(iC 4 H 10 ) is the isobutane conversion, %; F is the feed rate of isobutane, mole•h −1 ; m cat is the weight of catalyst, g; C CH 4 , C C 2 H 6 , C C 2 H 4 , C C 3 H 8 , C C 3 H 6 are the concentrations of methane, ethane, ethylene, propane, propylene respectively in the reaction products, vol%; V outlet is the volumetric flow of the reaction products, mL•h −1 .