Propane dehydrogenation over extra-framework In(I) in chabazite zeolites

Abstract

Indium on silica, alumina and zeolite chabazite (CHA), with a range of In/Al ratios and Si/Al ratios, have been investigated to understand the effect of the support on indium speciation and its corresponding influence on propane dehydrogenation (PDH). It is found that In₂O₃ is formed on the external surface of the zeolite crystal after the addition of In(NO₃)₃ to H-CHA by incipient wetness impregnation and calcination. Upon reduction in H₂ gas (550 °C), indium displaces the proton in Brønsted acid sites (BASs), forming extra-framework In⁺ species (In-CHA). A stoichiometric ratio of 1.5 of formed H₂O to consumed H₂ during H₂ pulsed reduction experiments confirms the indium oxidation state of +1. The reduced indium is different from the indium species observed on samples of 10In/SiO₂, 10In/Al₂O₃ (i.e., 10 wt% indium) and bulk In₂O₃, in which In₂O₃ was reduced to In(0), as determined from the X-ray diffraction patterns of the product, H₂ temperature-programmed reduction (H₂-TPR) profiles, pulse reactor investigations and in situ transmission FTIR spectroscopy. The BASs in H-CHA facilitate the formation and stabilization of In⁺ cations in extra-framework positions, and prevent the deep reduction of In₂O₃ to In(0). In⁺ cations in the CHA zeolite can be oxidized with O₂ to form indium oxide species and can be reduced again with H₂ quantitatively. At comparable conversion, In-CHA shows better stability and C₃H₆ selectivity (∼85%) than In₂O₃, 10In/SiO₂ and 10In/Al₂O₃, consistent with a low C₃H₈ dehydrogenation activation energy (94.3 kJ mol⁻¹) and high C₃H₈ cracking activation energy (206 kJ mol⁻¹) in the In-CHA catalyst. A high Si/Al ratio in CHA seems beneficial for PDH by decreasing the fraction of CHA cages containing multiple In⁺ cations. Other small-pore zeolite-stabilized metal cation sites could form highly stable and selective catalysts for this and facilitate other alkane dehydrogenation reactions.

---

Introduction

The increasing world demand for propylene and the current low cost of propane derived from shale gas have renewed interest in the propane dehydrogenation reaction (PDH).¹ PDH technologies, i.e., the Catofin process using a CrO_x/Al₂O₃ catalyst, and the Oleflex process using a supported Pt–Sn catalyst, have drawbacks, such as the high cost and rapid deactivation of Pt and the potential environmental impact caused by Cr. Many other catalysts for PDH have been investigated including Ga-,^2–10 Zn-,^11,12 Co-,^13–15 Fe-,^16–18 V-,^19–22 Zr-based materials,^23–25 and nanocarbon catalysts.^26,27 Metal cations (Ga⁺, Zn²⁺, Co²⁺, etc.) exchanged in zeolite H-ZSM-5 ^28–32 have shown promising alkane dehydrogenation properties as well: among them, Ga/H-ZSM-5 exhibits high stability and high propylene selectivity (∼90%) for propane dehydrogenation.^31,32

Shimizu and coworkers^33,34 recently reported that In-CHA is a stable and selective catalyst for ethane dehydrogenation. They have also shown that the CHA zeolite is superior to other zeolite frameworks for this reaction (BEA, MFI, and MOR) and that an Al-rich In-CHA zeolite (Si/Al = 6.85) provided higher dehydrogenation rates. They showed that In₂O₃, initially present on the external surface of CHA, can be reduced and then reacts with the Brønsted acid sites (BASs) of the zeolite to form an indium-exchanged (In⁺) CHA catalyst through the so-called reductive solid-state-ion exchange (RSSIE) process. Isolated [InH₂]⁺ sites were put forward as the active center in these catalysts based on Fourier transform infrared (FTIR) spectroscopy and X-ray absorption fine structure (XAFS) measurements.

The identification of [InH₂]⁺ as the active site for ethane dehydrogenation was primarily based on ex situ FTIR spectroscopy and DFT calculations;^33,34 however, the spectra used in support of the presence of [InH₂]⁺ were collected at low temperature (below 153 K, after treating with H₂ at 773 K), and therefore whether the indium hydride is stable at the reaction temperature (600 °C) remains to be determined. Furthermore, Andrews et al.³⁵ have found that indium hydride is unstable at temperatures above 200 K (indium hydride forms at 3.5 K but disappears above 200–210 K). It is thus uncertain whether [InH₂]⁺ is an intermediate at catalytic reaction temperatures. This point has also been debated in Ga/H-ZSM-5 catalysts with respect to propane dehydrogenation. Bell and coworkers³² believe that single [GaH]²⁺ sites are the active species; however, Schreiber et al.³¹ reported that Ga⁺–H⁺ pair sites are the stable active sites, a model supported by results from our laboratory.³⁶ Phadke et al.^32,37 have pointed out that Ga-exchanged species associated with Al-pairs in Ga/H-MFI catalyze both propane and butane dehydrogenation while isolated Ga species did not. They found, however, that both Ga species exhibited similar properties for ethane dehydrogenation. It is thus possible that the reaction mechanism of propane dehydrogenation with the In-CHA catalyst is different from that of ethane dehydrogenation.

Hart et al.³⁸ reported that MFI-supported catalysts containing high indium loadings (In/Al ratio close to 1 to minimize the presence of BASs) selectively catalyze the propane dehydrogenation reaction, but the reduction of indium species to In(0) during the reaction resulted in catalyst deactivation. Jones and coworkers^39,40 investigated In₂O₃–Ga₂O₃ and ternary In–Ga–Al mixed oxides as catalysts for propane dehydrogenation and found that reducing In₂O₃ domains to In(0) was detrimental to catalytic performance, and that once In(0) is formed from In₂O₃, it cannot be reactivated, indicating that In(0) was inactive for propane activation. These reports show that indium speciation at high temperature is highly dependent on the catalyst support: CHA-supported indium was stable for ethane dehydrogenation for 90 h time-on-stream (TOS), while the Al₂O₃ supported indium catalyst deactivated with time on stream.^33,39 Understanding the support impact on indium speciation and the corresponding influence on catalytic performance is thus important to advance In-based catalyst understanding for alkane dehydrogenation reactions.

In this work, samples of indium on silica, on alumina and exchanged into CHA zeolites with varying In/Al ratios and Si/Al ratios have been used to investigate the effect of the support on the indium speciation and the corresponding impact on propane dehydrogenation rates. On In-CHA, In₂O₃ initially located on the external surface of the H-CHA zeolite is reduced in the presence of dihydrogen displacing the protons of the BAS. Isolated In⁺ in the CHA zeolite was formed upon reduction. The reduced indium species were different on In/SiO₂, In/Al₂O₃ and bulk In₂O₃, where a portion of In₂O₃ was reduced to In(0) species. At comparable conversion, the propane dehydrogenation reaction over In-CHA shows better stability and C₃H₆ selectivity (∼85%) than In/SiO₂, In/Al₂O₃ and bulk In₂O₃. The examination of Si/Al ratios suggests that higher Si/Al ratios are beneficial to propane dehydrogenation. These results all point to the conclusion that the dehydrogenation properties of isolated In⁺ in CHA are superior to those of In(0) on other supports.

Results and discussion

Reduction of In-based catalysts

Indium supported on the CHA zeolite with varying In/Al ratios and Si/Al ratios, SiO₂, and Al₂O₃ are prepared as follows: H-CHA(11) samples were obtained via calcination of commercial NH₄-CHA (Si/Al = 11) samples obtained from ACS Material. H-CHA(x) (x = 5, 12 or 25, is the Si/Al ratio) were synthesized based on previous reports^41,42 (details are included in the ESI†). The physical properties of H-CHA samples are summarized in Table S1.† The micropore volume of H-CHA samples is in the range of 0.25–0.30 cm³ g⁻¹, consistent with typical CHA zeolites' properties.^43,44 For the sample H-CHA(11), the ²⁷Al solid-state nuclear magnetic resonance (SS-NMR) spectrum was dominated by a strong sharp signal at 54 ppm, which is assigned to tetrahedrally coordinated framework Al in the zeolite, and a weak feature at ∼0 ppm in the ²⁷Al MAS NMR spectrum (Fig. S2†) is assigned to a small amount of extra-framework Al (Al_EF), as octahedral aluminum (around 15%). Similarly, tetrahedrally coordinated Al is predominant in the spectra of the synthesized H-CHA with varying Si/Al ratios, and almost no Al_EF was observed at Si/Al ratios of 12 and 25.⁴⁴ The SEM images (Fig. S3†) of the CHA sample show approximately 300 nm, 300 nm, 1 μm, and 1.5 μm sized zeolite crystals for H-CHA(11), H-CHA(5), H-CHA(12), and H-CHA(25), respectively. The surface area of SiO₂ and γ-Al₂O₃ investigated was 285 and 79 m² g⁻¹ (Table S2†). The composition of In-based catalysts including In-CHA with varying In/Al and Si/Al ratios and 10In/SiO₂ and 10In/Al₂O₃ is summarized in Table S3.†

The XRD patterns of the In-based catalyst before and after reduction confirm that the indium phase formed after reduction is highly dependent on the support. The expected diffraction peaks of the CHA zeolite (Fig. 1 and S4†) are observed for all H-CHA samples investigated,⁴³ verifying that all zeolite samples are pure in phase and highly crystalline. The peaks of the In₂O₃ phase (JCPDS 06-0416) are clearly observed on the In-CHA(11, 1.0), 10In/Al₂O₃ and 10In/SiO₂ samples, prepared via the incipient wetness impregnation method, followed by calcination at 600 °C (the x and y in In-CHA(x, y) represent the Si/Al ratio and In/Al ratio, respectively). Upon reduction at 600 °C, the In₂O₃ peaks on the impregnated + calcined H-CHA sample disappear (Fig. 1) without another indium-related phase detected by XRD. This shows that In₂O₃ supported on the CHA zeolite is reduced to form highly dispersed species. The decrease in the relative intensity between the diffraction peaks at ∼9° and ∼21° on reduced In-CHA(11, 1.0) was interpreted as the introduction of extra-framework indium cations when exchanged with BASs (results below support the formation of In⁺). Similar observations were made on In-CHA(x, 1.0) (x = 5, 12 and 25, as shown in Fig. S4†). The micropore volumes of In-CHA(11, 1.0), calcined and reduced, are 0.21 and 0.20 cm³ g⁻¹, respectively (Table S1†). Considering the presence of indium species inside the micropores when calculating the micropore volume, by subtracting the mass of indium in In-CHA(11, 1.0) and assuming that In₂O₃ and In⁺ sites were formed in the calcined and reduced samples, the micropore volumes (without the In mass in them) were recalculated to be 0.26 and 0.23 cm³ g⁻¹, respectively. The similar micropore volume of In-CHA(11, 1.0) and H-CHA(11) suggests that In₂O₃ is largely located on the external surface of the H-CHA(11) zeolite particles, rather than inside the micropores, before reduction. The lower micropore volume of reduced In-CHA(11, 1.0) compared to H-CHA(11) suggests that indium species moved into the micropores of the zeolite upon reduction. A significant fraction of micropores were occupied by indium species on reduced In-CHA(5, 1.0) compared with H-CHA(5) (0.27 versus 0.16 cm³ g⁻¹, shown in Table S1†), which is due to a large amount of In₂O₃ that diffused into the micropores of the zeolite, while the void fraction occupation of micropores was low on reduced In-CHA(25, 1.0) samples. The diffusion of In₂O₃ into the zeolite micropores is also supported by the SEM and TEM images of calcined and reduced In-CHA (see Fig. S3 and S5†). The In₂O₃ particles observed on the external surface of the CHA particles (bright spots in SEM images and dark spots in TEM images) nearly disappear after reduction (Fig. S3 and S5†). Note that In₂O₃ particles are larger than 2 nm and thus cannot enter the small micropores of CHA. The exchange process was assisted by H₂ reduction, usually referred to as the RSSIE process.^33,34 The In₂O₃ phase could have been reduced to In₂O, which is volatile and can diffuse into the micropores of CHA where it reacts with BASs forming In⁺ sites. A similar process was also reported by Ga oxide in Ga/H-ZSM-5 systems.^45,46 The ²⁷Al NMR spectrum of reduced In-CHA with different Si/Al ratios showed a broadened peak at 54 ppm (tetrahedrally coordinated Al), assigned to a distorted tetrahedral aluminum environment.^47,48

Fig. 1 XRD patterns of samples investigated after calcination and reduction: (a) H-CHA(11) and In-CHA(11, 1.0); (b) indium on silica and alumina and bulk In₂O₃. The diffraction peaks at 2θ values of 30.6° and 35.5° (marked with asterisks) are assigned to In₂O₃ (JCPDS 060416), and diffraction peaks at 2θ values of 33.0°, 36.3° and 39.2° (marked with #) are associated with In(0) (JCPDS 05-0642).

It follows that upon reduction of In-CHA, the indium species enter the pores of the zeolite and interact with tetrahedrally coordinated Al. When indium is impregnated on SiO₂ and Al₂O₃ supports, In(0) was formed upon reduction (Fig. 1). But note that not all In₂O₃ is reduced since both In₂O₃ and In(0) are detected (Fig. 1). The pristine In₂O₃ sample initially displays only the In₂O₃ phase, and upon reduction In(0) is formed despite the remaining presence of a large amount of In₂O₃. In(0) is the species observed on bulk In₂O₃ and indium supported on silica and alumina although not all In₂O₃ could be reduced under the conditions investigated (10 vol% H₂/N₂ at 600 °C for 30 min), which is very different from the case of CHA. As the In/Al ratio of In-CHA (Si/Al = 12) rises from 0.3 to 1.7, the In₂O₃ diffraction peaks on calcined samples become gradually more clear (Fig. S6†). Upon reduction, In₂O₃ diffraction peaks disappear when the In/Al ratio is at or below 1.0. Both In₂O₃ and In(0) diffraction peaks, however, are detected upon reducing In-CHA(12, 1.7); that is, In₂O₃ is partially reduced to form In(0) when the BASs become scarce, as observed in bulk In₂O₃ and indium loaded on SiO₂ and Al₂O₃.

H₂-TPR measurements were conducted to monitor the reduction process of the In-based catalyst (Fig. 2a). For In-CHA(11, 1.0), the resulting profile shows several peaks with increasing temperature from 200 °C to 600 °C. This complex signal may result from two factors: (1) the distribution of In₂O₃ particle sizes at a high In loading (∼14%) had a variety of reduction temperatures;⁴⁰ (2) an elevated temperature was needed to drive the reaction of In₂O₃ with BASs when the sites had been consumed mainly at lower temperatures. On 10In/SiO₂ and 10In/Al₂O₃ catalysts, the reduction profile shows complete reduction below 600 °C. Bulk In₂O₃, however, shows a much higher reduction temperature with the reduction peak centered at around 700 °C; that is, bulk In₂O₃ is harder to reduce (consistent with the observation of Jones and coworkers).³⁹ The higher reduction temperature explains the large amount of In₂O₃ observed by XRD after reducing the In₂O₃ catalyst (Fig. 1).

Fig. 2 (a) H₂-TPR profiles and (b) results of H₂ pulse experiments on In-based catalyst at 600 °C.

The quantification of the H₂/In ratio during the H₂-TPR experiments provides information about the reaction stoichiometry and composition of the reduced indium species (Table 1). The consumed H₂/In ratio on the In-CHA(11, 1.0) catalyst was 0.98; almost all In₂O₃ species were reduced to In⁺. However, the H₂/In ratio on 10In/SiO₂ and 10In/Al₂O₃ was 0.88 and 0.86, respectively. The fact that the H₂/In ratio is less than unity shows that not all In₂O₃ was reduced. Combined with the XRD measurements, it is concluded that only a fraction of In₂O₃ on 10In/SiO₂ and 10In/Al₂O₃ was reduced to form In(0), which appears to hinder the further reduction of the remaining In₂O₃ and results in the low H₂/In ratios observed. A higher H₂/In ratio is observed on In₂O₃ than on 10In/SiO₂ and 10In/Al₂O₃ although the major consumption of H₂ occurs at elevated temperature (>600 °C).

Table 1 Quantification of H₂ consumption from H₂-TPR profiles and H₂ pulses and corresponding O₂ consumption from O₂ pulses

Sample	H2/In (H2-TPR)	H2/In (H2-pulse)	O/In (O2 pulse)
In2O3	1.05	—	—
10In/SiO2	0.88	0.25 ± 0.02	0.12 ± 0.01
10In/Al2O3	0.86	0.46 ± 0.01	0.44 ± 0.02
In-CHA(11, 1.0)	0.98	0.86 ± 0.02	0.86 ± 0.02
In-CHA(11, 1.0), (H2-O2 treated)	—	0.92 ± 0.02	0.92 ± 0.02

---

H₂ pulse experiments provide additional details about the reduction process. As shown in Fig. 2b, the initial 5 H₂ pulses were completely consumed on In-CHA(11, 1.0), and occurred in parallel with the formation of H₂O, confirming that the reduction process consumes H₂ and forms H₂O. The quantified H₂ to In content ratio was 0.86, slightly lower than 1. The slightly smaller H₂/In ratio in the H₂-pulse experiments versus the H₂ consumed in the H₂-TPR profiles is probably due to a small fraction of In₂O₃ that was not reduced via H₂ pulses. To test this hypothesis, we reduced In-CHA(11, 1.0) using a 10 vol% H₂ flow for 30 min, followed by O₂ pulses (Fig. S7†) and found that the O/In ratio was 1.01, consistent with the analysis of the H₂-TPR profiles. On the 10In/Al₂O₃, 10In/SiO₂ and In₂O₃ samples, the H₂ consumption in the H₂ pulses decreases gradually with the pulse number (Fig. 2b and Table 1). The H₂/In ratios are lower than the values obtained from H₂-TPR, suggesting that H₂ pulses cannot completely reduce the sample (possibly due to the dilution by the carrier gas) compared with experiments under a constant H₂ flow. The O/In ratio of In-CHA(11, 1.0) was consistent with H₂/In ratios (Table 1), demonstrating the quantitative reduction/oxidation of the In⁺³/In⁺ cations in the In-CHA samples.

On 10In/SiO₂ and 10In/Al₂O₃ catalysts, the lower O/In ratio versus the H₂/In ratio may result from the slow oxidation kinetics during O₂ pulses (Table 1 and Fig. S8†). These experiments demonstrate that over CHA zeolites, the reduction of In₂O₃ to In⁺ prevents the deep reduction of In₂O₃ to In(0). Once In⁺ is formed in the extra-framework positions of the CHA zeolite, it can be oxidized quantitatively to In³⁺ through oxygen pulses consuming 1 oxygen atom per In cation (denoted as H₂-O₂ treated, shown in Table 1 and Fig. S7†), and then reduced quantitatively, consuming one dihydrogen molecule per In cation.

Identification of isolated In⁺ sites

We have employed a pulse-reactor and in situ FTIR spectroscopy to determine that isolated In⁺ sites are the stable species formed upon reduction of the In-CHA catalyst with H₂. In fact, the ratio between the formed H₂O and consumed H₂ at the reaction temperature is consistent with isolated In⁺ sites as the stable indium species: as shown in Fig. 3a, the ratio of formed H₂O/consumed H₂ was 1.5 for the impregnated/calcined In₂O₃/CHA. As a control experiment, the formed H₂O/consumed H₂ ratio on CuO was determined to be ∼1.0, which is consistent with the stoichiometry during reduction (Fig. 3b and S9†). The higher ratio of formed H₂O/consumed H₂ means that another H source contributes to the H₂O formation during the reduction process on In-CHA. It follows that In₂O₃ is reduced to In(I) and then exchanged for protons in CHA. Based on the list of possible reactions (Fig. 3b), it is concluded that In⁺ sites were the product of In₂O₃ reduction, rather than [InH₂]⁺, [InO]⁺ or [HInOH]⁺ (Fig. 3b).

Fig. 3 (a) Determination of the formed H₂O/consumed H₂ ratio from H₂ pulses as a function of pulse number on several catalysts. (b) Stoichiometric analysis during reduction on CuO, In₂O₃, 10In/SiO₂, and 10In/Al₂O₃, and potential reactions on In-CHA catalysts.

On 10In/SiO₂ and 10In/Al₂O₃, the formed H₂O/consumed H₂ ratio in the initial H₂ pulse experiment was 1.0, which means that In₂O₃ reacted with H₂ to form In(0) and H₂O (Fig. 3b). These results are consistent with the identification of In(0) crystals in the XRD patterns of the reduced samples. Schreiber et al.³¹ applied the same technique to determine that Ga⁺ is the dominant species formed after reducing Ga/ZSM-5. Our results provide another example of the application of a pulse reactor to identify the structure and oxidation state of the metal-exchanged zeolite.

The determination of consumed H₂ on In-CHA with varying Si/Al ratios and In/Al ratios demonstrates that In₂O₃ could be completely reduced to In⁺ species when the In/Al ratios ≤ 1.0, independent of Si/Al ratios. For In-CHA (Si/Al = 12) with varying In/Al ratios (Fig. 4a and S10†), the H₂ consumption determined from H₂ pulses and H₂-TPR profiles versus In loading falls on a straight line with a slope of 1.0 when the In/Al ratio is at or below 0.7. This slope suggests that In₂O₃ was chiefly reduced to In⁺.

Fig. 4 (a) Quantification of consumed H₂ from H₂ pulses (black circle point) and H₂-TPR (blue square point) over In-CHA(x, y) (as indicated in the figure). The black dotted line represents the ratio of consumed H₂ to In loading of 1. (b) Quantification of consumed O atoms from O₂ pulses as a function of consumed H₂ from H₂ pulses over In-CHA(x, y) (as indicated in the figure). The blue dotted line represents a ratio of consumed O to consumed H₂ of 1. The corresponding H₂-TPR profiles and H₂ pulse and O₂ pulse curves are shown in Fig. S10 and S11.†

As the In/Al ratio rises to 1.0, the H₂/In ratio determined from H₂-TPR profiles reached 1.0, consistent with the result of the In-CHA(11, 1.0) catalyst. A H₂/In ratio slightly lower than 1.0 from the H₂ pulses is due to the dilution by the carrier gas when using H₂ pulses as the reducing agent. Further increase of the In/Al ratio to 1.7 results in a lower H₂/In ratio as determined from both H₂-TPR profiles and H₂ pulses; that is, the excess of In₂O₃ cannot be reduced completely, a result that is consistent with the detection of In(0) and In₂O₃ on the reduced In-CHA(12, 1.7) in XRD patterns (Fig. S6†). On In-CHA with different Si/Al ratios (Fig. 4a and S11†), the H₂/In ratios are close to 1.0, confirming that In₂O₃ in the CHA zeolite can be reduced to replace BASs and form isolated In⁺ when In/Al ratios ≤ 1.0 (independently of the Si/Al ratio). A ratio between the formed H₂O and consumed H₂ close to 1.5 for these In-CHA samples evidences the stoichiometry of the reaction during reduction (Fig. S12†).

Note that the small increase in the ratio of the formed H₂O to consumed H₂ on In-CHA(12, 1.0) and In-CHA(12, 1.7) may result from the exchange of In₂O₃ with BASs in the presence of H₂, forming [InO]⁺ and H₂O. This process does not consume H₂, but forms H₂O. The exchange of In₂O₃ with BASs is minor at high temperature (600 °C) since the formed [InO]⁺ can be easily reduced to In⁺. Control experiments were conducted introducing H₂ pulses at different temperatures (200 °C to 600 °C, in Fig. S13†). No significant H₂ was consumed at or below 300 °C, but when the temperature was increased to 400 °C, the ratio of formed H₂O to consumed H₂ reached 5.9, much higher than 1.5. This suggests that a significant fraction of In₂O₃ was ion-exchanged with BASs to form [InO]⁺, accompanied by the formation of H₂O without consuming H₂ (third possible reaction of the reduction of In-CHA in Fig. 3b). As the temperature rises to 500 °C, the ratio decreased to 1.39 in the initial H₂ pulses; this is an effect due to two reactions: In₂O₃ + 2H⁺ + 2H₂ → 2In⁺ + 3H₂O and [InO]⁺ + H₂ → In⁺ + H₂O. Further increase in temperature leads to a ratio between 1.0 and 1.5. It is concluded that different indium species were formed at different reduction temperatures. At 600 °C, In⁺ is the stable species upon reduction of the In-CHA catalysts. The subsequent O₂ pulses on reduced In-CHA(x, y) also confirmed that In⁺ is easy to oxidize (Fig. S10 and S11†). The ratio of O/H₂ was 1.0, showing that In³⁺ was reduced to In⁺ and then oxidized to In³⁺ (Fig. 4b).

In situ FTIR spectroscopy reveals the consumption of BASs upon reduction of the In-CHA sample and clarifies the redox cycle of In-CHA catalysts in the presence of H₂ or O₂. For the H-CHA sample (without indium, Fig. 5a(i)), two peaks at 3728 cm⁻¹ and 3575 cm⁻¹ are assigned to the external silanol group (Si–OH) and bridge-OH groups (BAS), respectively. As expected, there is no significant change before and after H₂ treatment (Fig. 5a(i) red and black traces). For In-CHA(11, 1.0), the spectrum of the dehydrated sample shows the same two peaks at 3728 cm⁻¹ and 3575 cm⁻¹, but upon reduction at 550 °C, there is a clear reduction of the intensity of the peak at 3575 cm⁻¹, evidence of the consumption of BASs after reducing In-CHA(11, 1.0). Additionally, there is a significant increase in the intensity of the zeolite structure Si–O–Si band at 1900–2100 cm⁻¹ (Fig. S14†), which might be due to the diffusion of a large amount of In₂O₃ located initially on the outer surface into the micropores and strong bonding of In(I) to the zeolite oxygen atoms.

Fig. 5 FTIR spectra of In-containing catalysts and H-CHA collected during the H₂ reduction process at 550 °C: (a) indium on SiO₂ and Al₂O₃, H-CHA(11) and In-CHA(11, 1.0), (b) In-CHA(12, y) with varying In/Al ratios ranging from 0 to 1.7, and (c) H-CHA(x) and In-CHA(x, 1.0) with varying Si/Al ratios ranging from 5 to 25 (see figure legends). Black and red traces represent spectra before and after reduction with H₂ (1 atm) at 550 °C, respectively. The background spectrum was collected in the IR cell, without a sample pellet, at room temperature under vacuum.

Interestingly, the peak at 3575 cm⁻¹ of the In-CHA(11, 1.0) catalyst did not disappear after reduction, and D₂ was used to determine if this peak is associated with OH groups. We found that the silanol group and the extra-framework Al–OH groups were converted to the corresponding –OD groups after coming into contact with D₂; however, the peak at 3575 cm⁻¹ remains unmodified after D₂ treatment at 550 °C (Fig. S15†); that is, this peak after reducing In-CHA(11, 1.0) is not from bridged-OH groups (i.e., BAS). Even for H-CHA(11), the peak at 3575 cm⁻¹ still exists upon D₂ treatment (Fig. S16†), suggesting that the observed peak may be unrelated to –OH groups. In the –OD region, there is no effect of D₂versus H₂ treatment between 1600 cm⁻¹ and 2000 cm⁻¹ (Fig. S15c†). The lack of indium hydride vibrations also suggests that the isolated In⁺ site—rather than [InH₂]⁺—was the dominant species formed upon reducing In-CHA at 550 °C (Fig. S15c†). This is different from Ga/H-ZSM-5, where a significant GaH_x band is formed upon reduction at 550 °C.⁴⁵

In the samples of 10In/SiO₂ and 10In/Al₂O₃, the hydroxyl group observed before reduction is largely consumed after reduction at 550 °C, a consequence of the replacement of the proton in these sites by In species. The examination of reduced In-CHA with varying In/Al ratios and Si/Al ratios by FTIR spectroscopy also supported the consumption of BASs by the indium species (Fig. 5b, c and S17†). Meanwhile, no InH_x bands were observed on In-CHA(x, y) with varying Si/Al and In/Al ratios (x = 5, 12, and 25 and y = 0.3, 0.7 and 1.0, shown in Fig. S18–S22†), confirming that the only isolated In⁺ species were the stable In species in all In-CHA samples investigated.

The lower H⁺ concentration in reduced In-CHA samples compared with H-CHA is additional evidence showing that In⁺ species rather than InH_x were formed at 550 °C. Note that the sample was first reduced in a 10 vol% H₂/N₂ flow, followed by the purge of a pure N₂ flow. Afterward, a 10 vol% D₂/N₂ flow was introduced to measure the H concentration of the sample (including H₂ and HD, 2 amu and 3 amu, respectively). For H-CHA(12) (Fig. 6a), significant H₂ (2 amu) and HD (3 amu) MS signals were observed upon switching to 10 vol% D₂/N₂, due to the H–D exchange of the surface H species when introducing D₂ into H₂-reduced H-CHA (see FTIR spectra in Fig. S15–S22†). Thus, the peak area of H₂ (2 amu) and HD (3 amu) could be regarded as an indicator of the total H species on the samples.

Fig. 6 (a) H₂ (2 amu), HD (3 amu), and D₂ (4 amu) signals as a function of time when switching the N₂ flow to a 10 vol% D₂ flow on In-CHA(12, y) with In/Al ratios ranging from 0 to 1.7 (see figure legends). The samples were reduced using a 10 vol% H₂/N₂ flow for 30 min and purged with a N₂ flow for another 30 min before switching to a 10 vol% D₂ flow. (b) Integrated H₂ and HD peak area in (a) and relative BAS densities as a function of In/BAS ratios (normalized to the H₂ and HD peak area or BAS density of H-CHA(12)).

The magnitude of the H₂ and HD MS signals decreases gradually as the In/BAS ratio increases and almost disappears when the In/BAS ratio is 2.3 (corresponding to an In/Al ratio of 1.7, in Fig. S23†); that is, the addition of In to the CHA zeolite replaces the –OH group associated with BASs without forming indium hydrides (such as [InH₂]⁺) since more H species should be observed if one mol BAS was converted to one mol [InH₂]⁺. The subsequent switching from 10 vol% D₂/N₂ to 10 vol% H₂/N₂ shows consistent results: the D concentration decreases as the In/Al ratio rises (Fig. S23†). We measured the BAS density via ethylamine-TPD of In-CHA(12, y) (details are included in the ESI†). The relative BAS density vs. In/BAS ratio shows that one indium atom replaces one BAS, as concluded in the above discussions. The decreased H₂/HD concentration is similar to the relative BAS concentration with the increase of the In/BAS ratio, suggesting that the former may be used as an estimate for the relative BAS concentration on the In-CHA system.

FTIR spectroscopy of d₃-acetonitrile (CD₃CN) adsorption provides information about the indium speciation of the reduced samples. In these experiments, the catalyst was reduced at 550 °C for 30 min in pure H₂ and cooled down to 50 °C under vacuum, followed by the dosing of CD₃CN. As shown in Fig. 7a, two peaks, at 2324 and 2297 cm⁻¹, were observed on H-CHA(11), which were assigned to CD₃CN adsorbed on extra-framework Al and BASs, respectively.⁴⁹ This is consistent with the aluminum speciation determined from the ²⁷Al MAS NMR spectra of the samples (Fig. S2†). Upon the introduction of H₂ into In-CHA(11, 1.0), the peaks at 2324 and 2297 cm⁻¹ almost disappeared, confirming the consumption of BASs after reduction. Instead, a peak at 2270 cm⁻¹ appears, a signal attributed to the interactions of the nitrile group with Lewis acid sites (LASs). In the case of reduced In-CHA(11, 1.0), the LAS should be isolated In⁺ sites. For 10In/SiO₂ and 10In/Al₂O₃, no well-defined peaks were observed upon CD₃CN adsorption on the reduced sample. This is attributed to the formation of In(0), as observed in the corresponding XRD patterns and the pulse reactor experiments. For In-CHA (Si/Al = 12) with varying In/Al ratios, the peak intensity of isolated In⁺ species increases linearly as the In/Al ratio rises from 0.3 to 1.0 and stays constant with a further increase of the In/Al ratio to 1.7 (Fig. 7d); that is, the In⁺ density is related to the In loading and no more In⁺ species are formed when the BASs were replaced completely. A higher In⁺ species concentration is also observed on In-CHA(5, 1.0) (Fig. 7c and d) due to the higher In loadings.

Fig. 7 FTIR spectra of the reduced In-based catalyst and H-CHA catalyst collected upon CD₃CN adsorption at 50 °C: (a) In on SiO₂ and Al₂O₃, H-CHA(11) and In-CHA(11, 1.0), (b) In-CHA(12, y) with In/Al ratios ranging from 0 to 1.7, and (c) H-CHA(x) and In-CHA(x, 1.0) with Si/Al ratios ranging from 5 to 25 (see figure legends). The background spectrum was collected in the IR cell without a sample pellet at room temperature under vacuum. (d) Integrated In⁺ peak area of In-CHA(x, y) catalysts with varying Si/Al ratios and In/Al ratios determined from infrared spectra (normalized to the In⁺ peak area of In-CHA(12, 1.0)) as a function of In loading.

In situ FTIR spectroscopy of the reduced and oxidized samples shows quantitative and reproducible oxidation/reduction of the In cation in In-CHA(11, 1.0). We introduced D₂ and D₂O to facilitate the investigation of the reduction and oxidation of the sample. As shown in Fig. 8, two peaks at 2747 and 2643 cm⁻¹ were observed upon D₂ introduction (see Fig. S15†). When introducing D₂O, there are no significant changes on the spectrum; that is, reduced In⁺ sites do not react with D₂O to form In-OD, clearly different from Ga/H-ZSM-5 where reduced Ga⁺ was easy to oxidize to form Ga-OH at 550 °C with H₂O.⁴⁵ Upon coming into contact with dry air, the BAS OD group (2631 cm⁻¹) re-appears, a result of the regeneration of BASs when isolated In⁺ was oxidized by air. From the difference spectrum (Fig. 8a(iv)), a peak at 2689 cm⁻¹ is now observed and accompanied by the regeneration of BASs. The peak at 2689 cm⁻¹ could be assigned to the In-OD group considering that the peak position is higher than the BAS O-D group and extra-framework Al-OD group. The introduction of D₂O again regenerates BASs, likely due to the hydrolysis of [InO]⁺ and the formation of BASs. Subsequently, the introduction of D₂ and quantification of the number of consumed In-OD and BAS OD groups (Fig. 8(vii)) show reversibility between the reduced and oxidized states of the sample.

Fig. 8 (a) FTIR spectra of reduced In-CHA(11, 1.0) treated under different conditions at 550 °C: (i) D₂; (ii) D₂O; (iii) air; (iv) (i) subtracted from the spectrum (iii); (v) D₂O; (vi) (i) subtracted from the spectrum (v); (vii) D₂O. Background spectrum was collected in the spectral cell with the H₂ reduced sample pellet at 550 °C. (b) Proposed structures of exchanged In species in In-CHA upon reduction and subsequent oxidation.

It follows that on a calcined sample, In₂O₃ located on the external surface of the CHA zeolite was reduced and diffused into the micropores in the presence of H₂ at high temperature, displacing BASs and forming isolated In⁺ sites. The O₂ treatment of the reduced In species leads to the formation of [InO]⁺ and [In(OH)₂]⁺ and regeneration of BASs (Fig. 8b). H₂ pulses on the H₂-O₂ treated In-CHA(11, 1.0) are consistent with the proposed oxidation state of isolated In⁺ sites (Fig. S24†). The low ratio of formed H₂O/consumed H₂ (∼0.25) is interpreted as the result of the reaction of H₂O—formed in the early portion of the catalyst bed—with [InO]⁺ in the back end of the catalyst bed, a result consistent with the in situ FTIR spectra (Fig. 8a). Overall, the ratio of formed H₂O/consumed H₂ during the entire experiment was estimated to be 1.1; that is, the main oxidized species was [InO]⁺, as supported by XRD that revealed no In₂O₃ diffraction peaks detected on H₂-O₂ treated In-CHA(11, 1.0) (Fig. S25†).

Propane dehydrogenation mechanism on isolated In⁺ sites

PDH catalytic rates and selectivity were measured on the reduced In-based catalyst. Calcined catalysts were reduced in 10 vol% H₂ for 30 min at 600 °C before the introduction of 2.54 kPa C₃H₈ (balancing N₂). At comparable conversion (<6%), In-CHA shows better stability than In₂O₃, 10In/SiO₂ and 10In/Al₂O₃. There is no significant decrease of the C₃H₆ yield over 14 hours (Fig. 9a). A long-term test of the 10In/Al₂O₃ sample (Fig. S26†) shows that the C₃H₆ yield keeps decreasing within a time on stream of 65 hours, evidencing that indium supported on Al₂O₃ exhibited poor stability for propane dehydrogenation. Isolated In⁺ sites in In-CHA(11, 1.0) show good C₃H₆ selectivity (∼85%) at 600 °C (Fig. 9b), higher than the selectivity exhibited by the other samples. The higher reaction rates on 10In/Al₂O₃ might be due to bare γ-Al₂O₃ since it has been shown that (bare) γ-Al₂O₃ catalyzes propane dehydrogenation.⁵⁰ A control experiment conducted on bare γ-Al₂O₃ found that indeed this sample displays higher catalytic rates and C₃H₆ selectivity than 10In/Al₂O₃ (Fig. S27†). Compared with the inert SiO₂ support, 10In/SiO₂ shows a higher C₃H₆ yield and similar C₃H₆ selectivity (∼55%) (Fig. S28†). The reduction in the C₃H₆ yield with time is attributed to the gradual reduction of In₂O₃ to In(0), evidenced by the negligible reactivity exhibited by In₂O₃ after being reduced for 8 hours (Fig. S29†). H₂O formation quantification and XRD measurements support that In₂O₃ was completely reduced to In(0) after 8 hours of reduction (Fig. S29†). Considering the occurrence of the thermal cracking reaction of propane, we tested the empty tube for propane dehydrogenation (C₃H₈ partial pressure 2.54 kPa with balancing N₂). The C₃H₆ yield and selectivity were 0.39% and 60%, respectively (see Fig. 9, dotted line). We then corrected the propane dehydrogenation reaction rate and cracking rate by subtracting the product derived from thermal cracking (see Fig. S30†). As expected, a small reduction in reaction rates was observed when considering the thermal cracking. The C₃H₆ selectivity increased from 85% to 88% for the In-CHA catalyst, while the C₃H₆ selectivity of 10In/Al₂O₃ increased from 77% to 79%. For 10In/SiO₂ and In₂O₃ catalysts, there is no significant change in the C₃H₆ selectivity when subtracting the thermal cracking products, which might result from the similar C₃H₆ selectivity (∼60%) of those catalysts with the thermal cracking reaction. Thus, we conclude that the CHA zeolite shows better stability and C₃H₆ selectivity than indium on the other supports investigated.

Fig. 9 (a) C₃H₆ yield and (b) C₃H₆ selectivity as a function of time on stream over In-CHA(11, 1.0), 10In/SiO₂, 10In/Al₂O₃ and In₂O₃. Reaction conditions: 600 °C, C₃H₈ partial pressure 2.54 kPa with balancing N₂, and space time 806 400 g_Cat s mol_C3H8⁻¹. The propane conversions are below 6% in the rate measurements. The dotted line in the figure represents the C₃H₆ yield (0.39%) in (a) and C₃H₆ selectivity (60%) in (b) without a catalyst under the same conditions.

We also examined the propane dehydrogenation reaction rate on the parent H-CHA (without indium); this catalyst shows higher catalytic rates than In-CHA(11, 1.0), and however, the rates decrease monotonically over time (Fig. S31†). In addition, the C₃H₆ selectivity of H-CHA(11) is much lower, approximately 20% (Fig. S31†), than that on In-CHA(11, 1.0) (∼85%). This observation confirms that In-CHA is stable and selective for the non-oxidative dehydrogenation of propane to propylene reaction while H-CHA favors propane cracking and has poor stability.

The impact of the In/Al ratio and Si/Al ratio of In-CHA catalysts on the PDH performance has also been examined (Fig. 10). With the increase of the In/Al ratio to 1.7 on the In-CHA (Si/Al = 12) catalyst, several interesting trends were observed: (1) the initial activity including C₃H₈ dehydrogenation and C₃H₈ cracking decreases; (2) the ratio between the C₃H₈ dehydrogenation and C₃H₈ cracking increases monotonically; (3) the catalyst exhibits better stability over 10 hours of testing (Fig. S32†). These observations are consistent with the results on H-CHA(11) and In-CHA(11, 1.0), showing that the addition of In could increase the C₃H₆ selectivity by replacing BASs. The lower reactivity of In-CHA compared with H-CHA is due to the weak intrinsic activity of isolated In⁺ species than BASs. Interestingly, the C₃H₈ dehydrogenation rate normalized to In loading declines as the Si/Al ratio decreases (Fig. 10b and S33†). In particular, the C₃H₈ dehydrogenation rate decreased from 88 μmol_C3H8 mol_In⁻¹ s⁻¹ for In-CHA(25, 1.0) to 7.9 μmol_C3H8 mol_In⁻¹ s⁻¹ for In-CHA(5, 1.0).

Fig. 10 (a) C₃H₈ dehydrogenation rate and C₃H₈ cracking rate as well as the ratio between the two (D/C) as a function of In/Al ratios on In-CHA (Si/Al = 12) catalysts. (b) C₃H₈ dehydrogenation rate and C₃H₆ selectivity as a function of Si/Al ratios on the In-CHA catalysts investigated (In/Al = 1). Reaction rates were collected at TOS = 21 min. Reaction conditions: 600 °C and C₃H₈ partial pressure 2.54 kPa with balancing N₂. The propane conversion is below 6% in the rate measurements.

Earlier we showed that for In/Al ratios ≤ 1.0, In₂O₃ could be reduced to isolated In⁺ species, independent of Si/Al ratios. This excludes the possibility that a low concentration of In⁺ in In-CHA(5, 1.0) results in the decrease of In-normalized PDH rates. Considering that a cage of zeolite CHA contains more BASs as the Si/Al ratios decrease, it is more likely to form multi-nuclear In⁺ sites in one cage (or In⁺–In⁺ sites) at low Si/Al ratios. These multi-nuclear species are detrimental to propane conversion and thus decrease the In-normalized rates compared with the case of one cage containing one In⁺ at high Si/Al ratios. The significant decrease of the micropore volume of In-CHA(5, 1.0) upon reduction compared with the reduced In-CHA(25, 1.0) catalyst (0.15 versus 0.29 cm³ g⁻¹, Table S1†) shows that In species occupy a good fraction of the micropore space on reduced In-CHA(5, 1.0). For H-CHA with varying Si/Al ratios, the BAS-normalized C₃H₈ dehydrogenation rates are close to each other (approximately 280–370 mol_C3H8 mol_BAS⁻¹ s⁻¹) and C₃H₆ selectivity is between 30% and 50% (Fig. S34 and S35†), showing similar C₃H₈ dehydrogenation performance on CHA with varying BAS densities.

We prepared as a benchmark a PtSn/Al₂O₃ catalyst (1%Pt and 2.6%Sn) using incipient wetness impregnation.⁵¹ As shown in Fig. S36,†the PtSn/Al₂O₃ catalyst showed a high initial C₃H₆ yield (30%) and deactivated quickly within a time on stream of 5 hours, although the C₃H₆ yield decreased slowly in the following 30 hours. The C₃H₆ selectivity decreased from 97% to 91% within a time on stream of 35 hours. The deactivation trend is consistent with the results reported in ref. 51. We calculated the Pt normalized reaction rate (TOF_Pt) based on the C₃H₆ yield at TOS = 10 h (TOF_Pt = 6.9 mmol_C3H8 mol_Pt⁻¹ s⁻¹). This value of TOF_Pt is 78 times higher than TOF_In determined on In-CHA(12, 1.0). Although In⁺ sites show a lower intrinsic reaction rate, it has better stability than PtSn/Al₂O₃.

Apparent activation energies (E_app) for the PDH reaction on the reduced catalyst were estimated using an Arrhenius plot in the temperature range of 580–620 °C. On H-CHA(11), the E_app of the C₃H₈ dehydrogenation and cracking reactions was 172 kJ mol⁻¹ and 130 kJ mol⁻¹ (Table 2). The lower E_app for C₃H₈ cracking is typical for BASs since these sites favor cracking.⁴¹

Table 2 The activation energies for propane dehydrogenation and cracking reactions on In-based and H-CHA catalysts

Sample	C3H8 dehydrogenation Eapp (kJ mol^−1)	C3H8 cracking Eapp (kJ mol^−1)
In2O3	143 ± 4.0	133 ± 3.8
10In/SiO2	179 ± 8.5	135 ± 7.4
10In/Al2O3	69.5 ± 11.4	200 ± 3.9
In-CHA(11, 1.0)	94.3 ± 0.6	206 ± 11
H-CHA(11)	172 ± 2.7	130 ± 6.0
H-CHA(12)	190 ± 1.5	177 ± 1.8
In-CHA(12, 0.3)	188 ± 16	177 ± 12
In-CHA(12, 0.7)	149 ± 5.6	212 ± 3.0
In-CHA(12, 1.0)	109 ± 0.8	198 ± 1.5
In-CHA(12, 1.7)	125 ± 14	167 ± 19
H-CHA(25)	223 ± 2.0	206 ± 3.0
In-CHA(25, 1.0)	109 ± 1.5	265 ± 28
H-CHA(5)	180 ± 0.7	174 ± 3.2
In-CHA(5, 1.0)	187 ± 8.8	233 ± 6.4

---

The In-containing samples have, in contrast, much lower propane dehydrogenation E_app and higher propane cracking E_app (Table 2). The gradual decrease of E_app for C₃H₈ dehydrogenation on In-CHA(12, 1.0) from 190 to 109 kJ mol⁻¹ as the In/Al ratio increases from 0 to 1.0 is consistent with the fact that In⁺ dominates the reaction since the BAS density decreases, although it slightly increases as the In/Al ratio is increased to 1.7. A similar E_app for C₃H₈ dehydrogenation on In-CHA(25, 1.0) was observed and an even higher E_app for C₃H₈ dehydrogenation was found on In-CHA(5, 1.0), which evidenced that multi-nuclear In⁺ sites in one cage (or In⁺–In⁺ sites) are less effective than single sites.

In summary, isolated In⁺ sites in the CHA zeolite favor dehydrogenation over cracking under all the conditions investigated. These observations are consistent with the high C₃H₆ selectivity on the In-CHA(11, 1.0) catalyst. On In₂O₃ and 10In/SiO₂ catalysts, both samples show higher E_app of C₃H₈ dehydrogenation than the C₃H₈ cracking reaction (Table 2), which is consistent with the lower C₃H₆ selectivity (∼50%) on the catalyst. Although lower E_app of C₃H₈ dehydrogenation and higher C₃H₈ cracking reaction were observed on 10In/Al₂O₃, the measured reaction rates were mainly the contribution of the γ-Al₂O₃ support.

The turnover frequency (TOF) and pre-exponential factor were calculated over the In-CHA catalyst for propane dehydrogenation at 600 °C assuming that all the In is equally reactive. For In-CHA(11, 1.0), entropy changes of formation of the transition state for propane dehydrogenation are estimated based on transition state theory,¹⁸ and determined to be −207 J mol⁻¹ K⁻¹ (Table S4†). The obtained entropy changes are larger than the values reported on H-[Fe]ZSM-5 (−155 J mol⁻¹ K⁻¹) towards propane dehydrogenation.¹⁸ This large negative entropy reflects that the transition state is geometrically constrained due to the small cage of the CHA zeolite. Similar trends are observed on In-CHA(25, 1.0) and In-CHA(12, 1.0); however, the quite distinct entropy changes of the transition state formation indicate the presence of a different active site on In-CHA(5, 1.0).

For In-CHA, the C₃H₈ reaction order was 0.77 for dehydrogenation and 0.93 for cracking, close to the first order for both reaction channels (Fig. 11a). In addition, the reaction order on H₂ is close to 0 (Fig. 11b); that is, H₂ is likely not part of the rate-determining step. These values are consistent with the reaction order values in In-CHA for ethane dehydrogenation.³³ These reaction orders can be rationalized by the following mechanisms: the propane C–H bond is initially dissociated through an oxidative addition process by isolated In⁺ to form [HIn(III)-C₃H₇]⁺ species. This step is followed by H-beta elimination with the formation of a metastable [H₂In(III)-C₃H₆]⁺ intermediate. The catalytic cycle is closed by the desorption of C₃H₆, and the reductive elimination of H₂ to form In⁺ (Scheme 1). The reaction order of ∼1 for C₃H₈ and ∼0 for H₂ point to the first step as the rate-determining step.

Fig. 11 The dependence of propane dehydrogenation and cracking reactions on (a) propane partial pressure and (b) H₂ partial pressure over the In-CHA(11, 1.0) catalyst at 600 °C.

Scheme 1 Schematic of proposed mechanisms for PDH on isolated In⁺ sites.

TG analysis of the spent catalyst illustrates the high stability of the In-CHA(11, 1.0) catalyst. After a time-on-stream of 14 hours, the samples show only a small amount of coke formation (∼2.4%, Fig. S37†), a value lower than that on the H-CHA(11) catalyst, confirming that the deactivation of the H-CHA catalyst is mainly due to the formation of coke and that an isolated In⁺ site is less prone to coke formation reactions. There may be a steric contribution to catalysis on In-CHA compared with H-CHA, considering that the available micropore volume on reduced In-CHA is lower than that on the parent H-CHA (Table S1†). Another possibility is that propylene adsorbs more strongly on H⁺ sites than In⁺ sites (i.e., has higher heats of adsorption). As shown in Fig. S38,† upon introducing C₃H₆ on the H-CHA catalyst surface, significant consumption of BASs (3575 cm⁻¹) is observed. Upon evacuation of C₃H₆, a strong signal associated with hydrocarbon species centered at 3055 cm⁻¹ is detected,³⁶ showing that C₃H₆ has a strong interaction with BASs and forms stable hydrocarbon species, which could lead to coke formation. No stable hydrocarbon species are observed when introducing C₃H₈ and C₃H₆ on the reduced In-CHA(12, 1.0), showing that In⁺ has a weak interaction with C₃H₈ and C₃H₆, which is also a reason for the low formation rate of coke. XRD measurements on the spent catalyst confirmed that no significant structural change occurred on the CHA catalysts, or on the reduced samples (Fig. S39†). It is worth noting that the relative intensities between the diffraction peaks at ∼9° and ∼21° decreased on In-CHA samples compared with H-CHA, suggesting that the formed extra-framework In⁺ sites are stable after the propane dehydrogenation reaction. On the spent In₂O₃ sample, there is an overall increase in the sample weight during TGA, attributed to the oxidation of In(0) to indium oxide. This is consistent with the results presented above showing that In(0) is present on the reduced In₂O₃ catalyst. The slight decrease of the weight on 10In/SiO₂ and 10In/Al₂O₃ results from the combined effect of the two processes: oxidation of In(0) to indium oxide and removal of formed coke.

Main-group compounds containing a group 13 metal (I) compound (M = Al, Ga, In) have been prepared for oxidative addition and reductive elimination reactions of strong single bonds such as X–H σ bonds (X = H, C, Si, P, etc.). These are key steps in catalytic cycles and are frequently observed in organometallic transition-metal chemistry.⁵² These highly reactive compounds with low oxidation states have been shown to possess high-energy lone pairs and available empty (usually π-antibonding) orbitals, thereby mimicking the state of electrons observed in transition-metal compounds. For instance, Seifert et al.⁵³ found that the gallium(I) complex [Ga({N(dipp)CMe}₂CH)] (dipp = 2,6- diisopropylphenyl) undergoes facile oxidative addition reactions with various element–hydrogen bonds at relatively low temperature (−78 to 25 °C). Beachley et al.⁵⁴ reported that organoindium compounds NaIn(CH₂SiMe₃)₂ with indium in the +1 oxidation state could be prepared from In(CH₂SiMe₃)₃ and NaH via a reductive elimination reaction with Si(CH₃)₄ as another product. These examples suggest that main-group compounds with low oxidation states can catalyze C–H bond scission via oxidative addition and reductive elimination processes. In this case, In⁺ in the CHA zeolite could be regarded as an analogue of an In(I) molecular complex, which catalyzes propane dehydrogenation first through the oxidative addition of propane to form [HIn(III)-C₃H₇] species, and then proceeds via H-beta elimination with the formation of an [H₂In(III)-C₃H₆] intermediate, and the cycle is closed by the desorption of C₃H₆ and the reductive elimination of H₂ to form In⁺ (Scheme 1). Our findings are a good example of the connection between the chemistry of metals in extra-framework positions and the molecular chemistry of main-group elements like In and Ga. This line of thought may be a fruitful avenue to identify novel catalytic applications of these materials and other main-group elements supported on acidic zeolites.

Conclusions

This report shows that CHA zeolites stabilize In in the +1 oxidation state as isolated species in extra-framework positions, independent of In/Al ratios and Si/Al ratios. In contrast, In/SiO₂, In/Al₂O₃ and In₂O₃ form substantial amounts of In(0) upon reduction in H₂. In-CHA shows quantitative and reproducible redox cycles by sequentially exposing the sample to sets of O₂ and H₂ pulses producing first mixtures of InO⁺ and (In(OH)₂)⁺ and then In⁺ upon reduction. In⁺ in CHA zeolites catalyzes propane dehydrogenation with a C₃H₆ selectivity of ∼85%, with minor propane cracking to methane and ethylene. Coke formation is minimal (<2.5% w/w) over a period of 14 h. The dehydrogenation selectivity over In-CHA is much higher than the selectivity observed over In/SiO₂, In/Al₂O₃ and bulk In₂O₃. The mechanism of the propane dehydrogenation reaction seems to proceed through a sequence of oxidative addition, beta-H elimination, and reductive elimination of a (postulated) dihydride indium intermediate, based on the reaction rate orders with respect to propane and hydrogen obtained at low conversions. There are clear analogies between the observed chemistry of In⁺ in CHA zeolite and In⁺ organometallic complexes.

Data availability

All data needed to evaluate the conclusions in the paper are present within the article or in the ESI file.† Additional data related to this paper may be requested from the authors.

Author contributions

Y. Y. and R. F. L. conceived the research project. Y. Y. conducted the experiments. Y. Y. and R. F. L. analyzed data and wrote the paper. All authors have given approval of the final version of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We acknowledge the support by the RAPID manufacturing institute, USA, supported by the Department of Energy (DOE) Advanced Manufacturing Office (AMO), award number DE-EE0007888-6.5. Y. Y. also acknowledges the support of the US National Science Foundation under the grant number of CBET-1803246. We thank Prof. B. J. Xu (Peking U.) for helpful suggestions on the manuscript.

Notes and references

1. J. J. H. B. Sattler, J. Ruiz-Martinez, E. Santillan-Jimenez and B. M. Weckhuysen, Chem. Rev., 2014, 114, 10613–10653.
2. Y. He, Y. Song and S. Laursen, ACS Catal., 2019, 9, 10464–10468.
3. T. Bauer, S. Maisel, D. Blaumeiser, J. Vecchietti, N. Taccardi, P. Wasserscheid, A. Bonivardi, A. Görling and J. Libuda, ACS Catal., 2019, 9, 2842–2853.
4. N. Raman, S. Maisel, M. Grabau, N. Taccardi, J. Debuschewitz, M. Wolf, H. Wittkämper, T. Bauer, M. Wu, M. Haumann, C. Papp, A. Görling, E. Spiecker, J. Libuda, H. Steinrück and P. Wasserscheid, ACS Catal., 2019, 9, 9499–9507.
5. Y. He, Y. Song, D. A. Cullen and S. Laursen, J. Am. Chem. Soc., 2018, 140, 14010–14014.
6. K. Searles, K. W. Chan, J. A. Mendes Burak, D. Zemlyanov, O. Safonova and C. Copéret, J. Am. Chem. Soc., 2018, 140, 11674–11679.
7. M. Raad, A. Astafan, S. Hamieh, J. Toufaily, T. Hamieh, J. D. Comparot, C. Canaff, T. J. Daou, J. Patarin and L. Pinard, J. Catal., 2018, 365, 376–390.
8. K. Searles, G. Siddiqi, O. V. Safonova and C. Copéret, Chem. Sci., 2017, 8, 2661–2666.
9. W. Kim, J. So, S. Choi, Y. Liu, R. S. Dixit, C. Sievers, D. S. Sholl, S. Nair and C. W. Jones, Chem. Mater., 2017, 29, 7213–7222.
10. J. J. H. B. Sattler, I. D. Gonzalez-Jimenez, L. Luo, B. A. Stears, A. Malek, D. G. Barton, B. A. Kilos, M. P. Kaminsky, T. W. G. M. Verhoeven, E. J. Koers, M. Baldus and B. M. Weckhuysen, Angew. Chem., Int. Ed., 2014, 53, 9251–9256.
11. N. M. Schweitzer, B. Hu, U. Das, H. Kim, J. Greeley, L. A. Curtiss, P. C. Stair, J. T. Miller and A. S. Hock, ACS Catal., 2014, 4, 1091–1098.
12. S. M. T. Almutairi, B. Mezari, P. C. M. M. Magusin, E. A. Pidko and E. J. M. Hensen, ACS Catal., 2012, 2, 71–83.
13. C. Chen, S. Zhang, Z. Wang and Z.-Y. Yuan, J. Catal., 2020, 383, 77–87.
14. Y. Dai, J. Gu, S. Tian, Y. Wu, J. Chen, F. Li, Y. Du, L. Peng, W. Ding and Y. Yang, J. Catal., 2020, 381, 482–492.
15. M. Moselage, J. Li and L. Ackermann, ACS Catal., 2016, 6, 498–525.
16. S. Tan, B. Hu, W. Kim, S. H. Pang, J. S. Moore, Y. Liu, R. S. Dixit, J. G. Pendergast, D. S. Sholl, S. Nair and C. W. Jones, ACS Catal., 2016, 6, 5673–5683.
17. J. H. Yun and R. F. Lobo, J. Catal., 2014, 312, 263–270.
18. J. H. Yun and R. F. Lobo, J. Phys. Chem. C, 2014, 118, 27292–27300.
19. Z. Zhao, T. Wu, C. Xiong, G. Sun, R. Mu, L. Zeng and J. Gong, Angew. Chem., Int. Ed., 2018, 57, 6791–6795.
20. Y. Xie, R. Luo, G. Sun, S. Chen, Z. Zhao, R. Mu and J. Gong, Chem. Sci., 2020, 11, 3845–3851.
21. N. Kong, X. Fan, F. Liu, L. Wang, H. Lin, Y. Li and S. Lee, ACS Nano, 2020, 14, 5772–5779.
22. Y. Gu, H. Liu, M. Yang, Z. Ma, L. Zhao, W. Xing, P. Wu, X. Liu, S. Mintova, P. Bai and Z. Yan, Appl. Catal., B, 2020, 274, 119089.
23. Y. Zhang, Y. Zhao, T. Otroshchenko, H. Lund, M. Pohl, U. Rodemerck, D. Linke, H. Jiao, G. Jiang and E. V. Kondratenko, Nat. Commun., 2018, 9, 3794.
24. T. Otroshchenko, V. A. Kondratenko, U. Rodemerck, D. Linke and E. V. Kondratenko, J. Catal., 2017, 348, 282–290.
25. T. Otroshchenko, S. Sokolov, M. Stoyanova, V. A. Kondratenko, U. Rodemerck, D. Linke and E. V. Kondratenko, Angew. Chem., Int. Ed., 2015, 54, 15880–15883.
26. Z.-P. Hu, C. Chen, J.-T. Ren and Z.-Y. Yuan, Appl. Catal., A, 2018, 559, 85–93.
27. Z.-P. Hu, H. Zhao, C. Chen and Z.-Y. Yuan, Catal. Today, 2018, 316, 214–222.
28. S. Y. Yu, G. J. Yu, W. Li and E. Iglesia, J. Phys. Chem. B, 2002, 106, 4714–4720.
29. W. Li, S. Y. Yu, G. D. Meitzner and E. Iglesia, J. Phys. Chem. B, 2001, 105, 1176–1184.
30. J. A. Biscardi, G. D. Meitzner and E. Iglesia, J. Catal., 1998, 179, 192–202.
31. M. W. Schreiber, C. P. Plaisance, M. Baumgärtl, K. Reuter, A. Jentys, R. Bermejo-Deval and J. A. Lercher, J. Am. Chem. Soc., 2018, 140, 4849–4859.
32. N. M. Phadke, E. Mansoor, M. Bondil, M. Head-Gordon and A. T. Bell, J. Am. Chem. Soc., 2019, 141, 1614–1627.
33. Z. Maeno, S. Yasumura, X. Wu, M. Huang, C. Liu, T. Toyao and K. Shimizu, J. Am. Chem. Soc., 2020, 142, 4820–4832.
34. Z. Maeno, X. Wu, S. Yasumura, T. Toyao, Y. Kanda and K. Shimizu, Catalysts, 2020, 10, 807.
35. L. Andrews and X. Wang, Angew. Chem., Int. Ed., 2004, 43, 1706–1709.
36. Y. Yuan, C. Brady, R. F. Lobo and B. Xu, ACS Catal., 2021, 11, 10647–10659.
37. N. M. Phadke, E. Mansoor, M. Head-Gordon and A. T. Bell, ACS Catal., 2021, 11, 2062–2075.
38. V. I. Hart, M. B. Bryant, L. G. Butler, X. Wu and K. M. Dooley, Catal. Lett., 1998, 53, 111–118.
39. S. Tan, L. B. Gil, N. Subramanian, D. S. Sholl, S. Nair, C. W. Jones, J. S. Moore, Y. Liu, R. S. Dixit and J. G. Pendergast, Appl. Catal., A, 2015, 498, 167–175.
40. S. Tan, S. Kim, J. S. Moore, Y. Liu, R. S. Dixit, J. G. Pendergast, D. S. Sholl, S. Nair and C. W. Jones, ChemCatChem, 2016, 8, 214–221.
41. J. H. Yun and R. F. Lobo, Catal. Sci. Technol., 2015, 5, 264–273.
42. B. Ipek, M. J. Wulfers, H. Kim, F. Göltl, I. Hermans, J. P. Smith, K. S. Booksh, C. M. Brown and R. F. Lobo, ACS Catal., 2017, 7, 4291–4303.
43. T. D. Pham, M. R. Hudson, C. M. Brown and R. F. Lobo, ChemSusChem, 2014, 7, 3031–3038.
44. J. A. Loiland and R. F. Lobo, J. Catal., 2014, 311, 412–423.
45. Y. Yuan, C. Brady, L. Annamalai, R. F. Lobo and B. Xu, J. Catal., 2021, 393, 60–69.
46. G. D. Meitzner, E. Iglesia, J. E. Baumgartner and E. S. Huang, J. Catal., 1993, 140, 209–225.
47. R. Dupree, M. H. Lewis and M. E. Smith, J. Appl. Crystallogr., 1988, 21, 109–116.
48. A. Kentgens, K. Scholle and W. S. Veeman, J. Phys. Chem., 1983, 87, 4357–4360.
49. H. J. Cho, D. Kim and B. Xu, ACS Catal., 2020, 10, 4770–4779.
50. D. Zhao, H. Lund, U. Rodemerck, D. Linke, G. Jiang and E. V. Kondratenko, Catal. Sci. Technol., 2021, 11, 1386–1394.
51. A. H. Motagamwala, R. Almallahi, J. Wortman, V. O. Igenegbai and S. Linic, Science, 2021, 373, 217–222.
52. T. Chu and G. I. Nikonov, Chem. Rev., 2018, 118, 3608–3680.
53. A. Seifert, D. Scheid, G. Linti and T. Zessin, Chem.–Eur. J., 2009, 15, 12114–12120.
54. O. T. Beachley Jr and R. N. Rusinko, Inorg. Chem., 1981, 20, 1367–1370.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1sc05866e

---