Single Atomic Pt on SrTiO₃ Catalyst in Reverse Water Gas Shift Reactions

Abstract

Copper catalysts were widely developed for CO₂ conversion, but suffered severe sintering at temperatures higher than 300 °C. Platinum was the most active and stable metal for RWGS reactions. However, the high price and scarcity of platinum restrained its application. Downsizing the metal particles can significantly improve the atom efficiency of the precious metal but the size effect of Pt on RWGS reactions was still unclear. In the present work, the single atomic Pt on SrTiO₃ was prepared using an impregnation leaching method, and the catalyst showed significant activity for an RWGS reaction, achieving a CO₂ conversion rate of 45%, a CO selectivity of 100% and a TOF of 0.643 s⁻¹ at 500 °C. The structures of the catalysts were characterized using XRD, STEM and EXAFS. Especially, the size effect of Pt in RWGS was researched using in situ FTIR and DFT calculations. The results reveal that single Pt atoms are the most active species in RWGS via a “–COOH route” while larger Pt cluster and nanoparticles facilitate the further hydrogenation of CO. The reaction between formate and H* is the rate determination step of an RWGS reaction on a catalyst, in which the reaction barrier can be lowered from 1.54 eV on Pt clusters to 1.29 eV on a single atomic Pt.

1. Introduction

Global warming is becoming a rapid deterioration as a result of the huge release of CO₂ due to the extensive consumption of fossil energy. Upgrading CO₂ into more valuable products is considered as one of the most promising solutions for a CO₂-neutral energy supply [1,2,3,4]. RWGS reaction (Equation (1)) is an important process for converting CO₂ to CO, which can be used as feedstock to produce value-added chemicals and synthetic fuels [5,6]. The key issue of RWGS is how to lower the activation energy barrier of high chemically stabile CO₂ for obtaining high conversion rates and selectivity, and how to improve the long-term stability by reducing the carbon deposition and inhibiting the aggregation of active metal.

$${\mathrm{CO}}_2+{\mathrm{H}}_2\to \mathrm{CO}+{\mathrm{H}}_2\mathrm{O} \Delta H_{298K}^0=41.5kJ/mol$$

(1)

Supported noble metal catalysts are widely investigated for RWGS due to their high activities. However, the high prices and rare resources of these noble metals limit their application in large-scale production. Reducing the particle size to sub-nanoscale, such as single atomic site catalyst, may enhance the atom efficiency of noble metals. Single atomic metal catalysts are featured with isolated metal atoms anchored by the surrounding coordination species of solid supports as single atomic active sites, which are different from Pt nanoparticles where the metal atoms are surrounded by the same metal atoms. Therefore, the catalytic performances and the stabilities of highly dispersed metal species are greatly influenced by the supports because the metal atoms are always coordinated by the atoms on the surface of the support (mainly surface defects or Lewis basic sites) [7,8,9]. Both reducible (TiO₂ [10,11,12,13,14,15,16], CeO₂ [17,18,19,20], etc.) and irreducible supports (SiO₂ [21], Al₂O₃ [16,22,23], zeolite [24], MoS₂ [3], Mo₂C [25], etc.) have been adopted for loading highly dispersed Pt. For the Pt loading on a reducible support, the reverse water gas shift reaction was generally accepted as the redox mechanism via oxidation and reduction at the Pt sites and the metal/oxide interfaces [16]. The reactive area was also proposed over a Pt/CeO₂ catalyst that Pt and ambient ceria (not solely the ceria on the Pt/CeO₂ interface) surrounding the Pt center participated in the RWGS reaction [18]. In addition, CO₂ can associatively adsorb on the surface defect sites to generate CO [18,26]. For the Pt catalysts loading on an irreducible support, H₂ and CO₂ are both activated on a metal center and the hydrogenation route may dominate the RWGS reaction [27]. The CO selectivity of a catalyst in CO₂ hydrogenation reactions is closely related with the CO binding strength that may arise from the strong interactions (e.g., charge transfer) between the noble metal atoms and the oxide support and the particle size of Pt [11,26]. A strong PM-CO bond strength may lead to low CO selectivity. In similar results, a strong Co-CO bond strength was also reported to benefit the CO₂ methanation activity over a Co/CeO₂-ZrO₂ catalyst [27]. However, the particle size effect of Pt on RWGS has seldom been reported.

Perovskite catalysts are becoming increasingly significantly advantageous in many reactions, such as dry reforming of methane [28], Suzuki couplings [29], water splitting [30], oxygen reduction [31] and automotive emissions control [32]. It is interesting that palladium (Pd) could reversibly move into and out of the perovskite lattice in the cycling between oxidative and reductive atmospheres. This movement helped to suppress the growth of metallic Pd particles [32]. A similar conclusion was also found on the Pd/SrTiO₃ in Suzuki coupling reactions [29]. These results suggested that SrTiO₃ may be a good support for a Pt catalyst in redox reactions. Furthermore, the surface alkaline-earth-metal Sr component was observed to enhance the adsorption and activation of CO₂ [33]. Orlov and Chen et al. reported that Rh/SrTiO₃ achieved good activity and CO selectivity for CO₂ hydrogenation and the superior activity was suggested to originate from the cooperative effect between the highly dispersed sub-nanometer Rh clusters for the efficient dissociation of H₂/C₂H₆ and the reconstructed SrTiO₃ with oxygen vacancies for preferential adsorption/activation of CO₂ [34]. Recently, the size effects of Pt particles in Pt/CeO₂ catalysts among Pt single atoms, sub-nanoclusters (ca. 0.8 nm) and nanoparticles (3.8–9.3 nm) for the WGSR were reported [35]. However, the size effects of Pt from single atomic to cluster in RWGS reactions have not been discussed; for example, the reaction routes and the states of intermediates on subnano metal catalysts may be different from nanocatalysts. In the present work, single atomic Pt and Pt nanoparticles were prepared using impregnation method followed by leaching. By means of in situ FTIR, the intermediate processes of Pt single atoms and Pt nanoparticles in an RWGS reaction were compared. Then, the reason for the different activities of Pt single atoms and Pt nanoparticles were explained from energy using DFT calculation.

2. Results and Discussions

2.1. Catalytic Performances

The catalytic performances of catalysts in RWGS reactions are shown in Figure 1. The true Pt content of catalysts determined using ICP-OES was as follows: Pt in Pt/SrTiO₃ = 0.39% and Pt in Pt/SrTiO₃-L = 0.17%. Although the Pt content was significantly reduced by leaching, the Pt/SrTiO₃-L catalyst presented much higher CO₂ conversions than pristine Pt/SrTiO₃ at different temperatures.

Table 1. TOFs of Pt/SrTiO₃-L catalyst in RWGS reaction at different temperatures.

	300	400	500
TOF (s−1)	0.154	0.455	0.643

shows the TOF of a Pt/SrTiO₃-L catalyst in an RWGS reaction at different temperatures. Chen [36] studied an RWGS reaction over the Pt/CeO₂ catalyst; in his work, the best TOF of the sample with 1% Pt was 0.056 s⁻¹ at 300 °C. The Pt/SrTiO₃-L catalyst showed a much higher TOF than the Pt/CeO₂ catalyst; the main reason is the difference of metal dispersion. Large Pt particles make many Pt sites exist as invalid sites, so the TOF of Pt/CeO₂ catalyst is smaller than Pt/SrTiO₃-L catalyst that loaded Pt in the form of a single atom and a cluster.

2.2. Textural Properties

In order to further characterize the dispersion state of Pt species in a catalyst, high-resolution STEM images of the samples were taken. Figure 2a shows the HAADF-STEM images of a Pt/SrTiO₃ catalyst, in which the Pt nanoparticles are 2.72 nm in size. Figure 2b shows the HAADF-STEM images of a Pt/SrTiO₃-L catalyst. Single atomic Pt and small Pt clusters composed of 2–8 atoms (atoms) can be seen. No Pt nano particles can be found on the Pt/SrTiO₃-L catalyst.

The crystal phase of SrTiO₃ and the Pt/SrTiO₃ catalyst were measured using XRD, and the results are shown in Figure 2d. The synthesized SrTiO₃ is highly consistent with the standard diffraction pattern of SrTiO₃ (JCPDS# 35-0734). Pt related species cannot be found in all of the samples, indicating that the Pt species were well dispersed on SrTiO₃. The extended X-ray absorption fine structure (EXAFS) measurements are shown in Figure 2e. Compared with the standard sample of Pt foil and PtO₂, no obvious Pt-Pt bond was observed in either of the samples. The peak at 1.58 Å belonging to Pt-O coordination in the Pt/SrTiO₃-L catalyst, was sharper than that of Pt/SrTiO₃. The second peak at 2.8–3.2 Å, belonging to PtO₂ in Pt/SrTiO₃-L catalyst, does not match the reference indicating that Pt species are mainly in sub-nanoscale and in highly undercoordinated state [37,38]. It should be mentioned that the EXAFS data were obtain ex situ. The valance state of Pt may not be the true state of Pt in RWGS reactions.

Table 2. The BET results of samples.

Sample	SrTiO3	Pt/SrTiO3	Pt/SrTiO3-L
BET(${\mathrm{m}}^2$/g)	17	11	9

shows the BET surface areas of the samples. After calcination at 1000 °C for 3 h, the specific surface area of SrTiO₃ was 17 m²/g. The Pt/SrTiO₃ and Pt/SrTiO₃-L catalysts have a much smaller BET because loading Pt may block some of the pores in the catalyst and acid treatment may damage the pore structure of the catalyst.

2.3. H₂-TPR

H₂-TPR was used to characterize the redox properties of catalysts and the results are shown in Figure 3. There is only one H₂ consumption peak centered at 373 °C on SrTiO₃, ascribed to the surface oxygen [39]. This peak is obviously widened on the Pt/SrTiO₃ catalyst arising from the activation effect of Pt on the oxide support surface [40,41]. As for the reduction in the Pt related species on the Pt/SrTiO₃ catalyst, two peaks at 68 and 201 °C, appearing that can be assigned to the reduction in the Pt nanoparticles and the oxygen on interfaces between Pt and the SrTiO₃ support. The significantly weakened H₂ consumption peak of Pt over Pt/SrTiO₃-L catalyst is also located at 68 °C, but has a much smaller peak area, suggesting that a part of Pt atoms strongly interact with the support. Two overlapped H₂ consumption peaks at 300–400 °C of the Pt/SrTiO₃-L catalyst can be ascribed to the reduction of the interface oxygen of Pt-O-SrTiO₃ and the surface oxygen of SrTiO₃ without interaction with Pt [42].

2.4. In Situ FTIR

2.4.1. CO₂ and H₂ Co-Adsorption on SrTiO₃ Support

Figure 4a,b shows the in situ FTIR of CO₂ and the H₂ co-adsorption on SrTiO₃. The peaks at 2400–2300 cm⁻¹ are assigned to the linear adsorption of CO₂ [2,18,25]. The peaks at 1650–1680 cm⁻¹ and 1300cm⁻¹ are assigned to asymmetric and symmetric stretching of ${\mathrm{CO}}_2^{\mathsf{\delta }-}$, respectively [18]. All of the species desorb with the elevation in the temperature. At temperatures higher than 200 °C, CO₂ molecules mainly physically and weakly adsorb on the SrTiO₃ support. The formate species (1559 cm⁻¹) on SrTiO₃ can only be found at temperatures higher than 150 °C, indicating that SrTiO₃ also participates in the formation of formate intermediates.

When the in situ FTIR characterization was performed at first for CO₂ adsorption and then for purging by H₂, the peaks of all of the CO₂ derived species reduced quickly and could be detected clearly at temperatures higher than 200 °C, suggesting that a bare SrTiO₃ support cannot adsorb and activate CO₂ efficiently.

2.4.2. CO₂ and H₂ Adsorption on Pt/SrTiO₃ and Pt/SrTiO₃-L

(1)

CO₂ pre-adsorption and purging by H₂

The in-situ DRIFT spectra of adsorbed species on the Pt/SrTiO₃ catalyst under CO₂ pre-adsorption and then purging by H₂, are shown in Figure 5a–c. Physically adsorbed CO₂ molecules on both catalysts at 50 °C can be found as indicated by the peaks at 2200–2400 cm⁻¹. The faster CO₂ desorption behavior on Pt/SrTiO₃ indicates that CO₂ adsorbing on Pt nanoparticles is more stable than those on Pt single atoms and clusters. The peaks at 2000–2100 cm⁻¹ were assigned to CO adsorbing atop metallic Pt atoms [43,44]. Both catalysts show similar behavior in CO adsorption with the elevation of temperature: the peak intensity becomes weaker because of desorption and the peak position redshifts due to a lower surface coverage. The peaks of Pt (CO) species on Pt/SrTiO₃-L are much sharper and narrower than those on the Pt/SrTiO₃ catalyst, indicating the narrower size distribution of Pt species. The peaks at 1914–1929 cm⁻¹ are ascribed to CO on Pt nanoparticles or clusters [45]. The peaks in this region on a leached sample are significantly lower than those on pristine Pt/SrTiO₃, indicating the removal of Pt nanoparticles by leaching. After excluding the influence of support and CO₂ products, the peaks at 1566–1539 and 1357–1395 cm⁻¹ can be assigned to the formate species produced in the co-adsorption reaction [2,13,25], as shown in Figure 5c,f. Compared with those on SrTiO₃ in Figure 4b,d, the signals of formate on Pt/SrTiO₃ and Pt/SrTiO₃-L present at 50 °C, suggesting that Pt is the more active species, rather than the SrTiO₃ support for CO₂ hydrogenation to formate.

(2)

CO₂ and H₂ co-adsorption

The DRIFT spectra of the adsorbed species were collected under H₂ and CO₂ co-adsorption, and the results are shown in Figure 6. From Figure 6b,e, the Pt/SrTiO₃-L catalyst has stronger and narrower peaks of CO atop Pt at 2071 cm⁻¹ [44,45]. The peaks of CO on Pt clusters (1900–2000 cm⁻¹) are very small. The Pt/SrTiO₃ catalyst has obviously broad peaks at 2072 cm⁻¹ (CO atop Pt atoms) and 1927 cm⁻¹ (CO on Pt nanoparticles), indicating the multiple CO-derived species on the Pt/SrTiO₃ catalyst. From Figure 6c,f, the peaks of formate (1800–1200 cm⁻¹) on the Pt/SrTiO₃-L catalyst are very weak suggesting that the formate species could convert to CO immediately. However, those species on the Pt/SrTiO₃ catalyst are still robust, even at 400 °C, indicating the strong adsorption of formate on the Pt nanoparticles.

According to the above three groups of in situ FTIR results, the reaction path and gas activation site can be reasonably speculated. The SrTiO₃ support has no activity for an RWGS reaction, but intermediate species appeared on its surface when CO₂ was introduced. This shows that CO₂ is activated to generate intermediate species. After the addition of Pt, H₂ can be cracked on the surface of the metal, thereby making the catalyst active. As for the two catalysts, the Pt/SrTiO₃-L catalyst showed a more obvious consumption of intermediate products than the Pt/SrTiO₃ catalyst, which might lead to the higher activity of the Pt/SrTiO₃-L catalyst.

2.5. DFT Results and Reaction Mechanisms

According to the in situ FTIR results, sub-nano Pt is more active than Pt nano particles for the formation of formate and CO₂ can only be physically adsorbed on Pt/SrTiO₃. Formate species may convert to CO immediately over the Pt/SrTiO₃-L catalyst, which only has single Pt atoms and Pt clusters. To consolidate these conclusions, we further conducted DFT calculations to explore the possible reaction mechanism of RWGS over a catalyst.

The SrTiO₃ [110] surface (Figure 7) can provide six isolated sites to anchor six Pt atoms. While, further increasing the number of Pt atoms, there were not enough sites on the surface to isolate the Pt atoms, which made the formation of Pt clusters occur gradually. The Pt atoms tend to occupy the surface vacancy preferentially and hate to form Pt-Pt bonds when the Pt content is low. From the in situ FTIR results, metallic Pt dominates the Pt species. Therefore, the Pt1 single atom and Pt7 metal cluster are used for the RWGS reaction as representatives. For comparison, we chose the models of the SrTiO₃ [110] surface loaded with isolated Pt atoms (Pt₁/SrTiO₃) and Pt₇ cluster (Pt₇/SrTiO₃) to investigate the size effect of Pt on the RWGS’s performance. In the work of Chunyang Dong et al., the density functional theory of Pt nanoparticles was performed. The results show that it is difficult to form CO on Pt nanoparticles [46].

We screened possible intermediates and reaction routes for CO₂ hydrogenation over Pt₁/SrTiO₃ and Pt₇/SrTiO₃. The results are shown in Figure 8. On both surfaces, our calculations support the COOH intermediates mechanism in the RWGS reaction, CO₂→COOH→CO, which is consistent with the results of in the situ FTIR. Taking the reaction on Pt₁ as the example, the reaction starts by H₂ dissociated adsorption with an exothermic energy of 1.43 eV. Next, CO₂ shows a physical adsorption with an adsorption energy of −0.05 eV, suggesting the physically adsorption of CO₂ on Pt. In the presence of dissociated H atoms on Pt sites, the CO₂ can react with the H atoms to form COOH. The reaction barrier and reaction energy of this hydrogenation reaction over two catalysts are 0.72 (Pt₁/SrTiO₃) and 0.71 eV (Pt₇/SrTiO₃), respectively. The COOH intermediate over Pt₁/SrTiO₃ then further reacts with a nearby H atom to produce CO and H₂O with an energy barrier of 1.29 eV. The generated water desorbs into the gas phase, leaving behind the adsorbed CO molecule. However, this process on Pt₇/SrTiO₃ needs much the higher energy of 1.54 eV, indicating that the reaction between formate and H* is the rate determination step of the RWGS reaction and that Pt₁ is more active than Pt₇. The strong adsorption of formate (in situ FTIR results) on the Pt cluster and nanoparticle may retard the reaction between the COOH intermediate and nearby H.

To understanding the CO behavior of desorption or hydrogenation on the catalysts, we further calculate the CO hydrogenation to CHO. The reaction barriers of CO hydrogenation are 1.88 and 1.07 eV on Pt₁/SrTiO₃ and Pt₇/SrTiO₃, respectively. The high reaction barrier on Pt₁/SrTiO₃ means that the generated CO molecule tends to desorb rather than hydrogenate. In contrast, on Pt₇/SrTiO₃ the same reaction barrier of CO hydrogenation and RWGS indicates that the CO prefers to be hydrogenated to produce other hydrogenated products (e.g., CH₂O, CH₃OH, CH₄ etc.).

3. Experimental Method

3.1. Catalyst Synthesis

The SrTiO₃ supports were prepared using a hydrothermal method. In 20 mL anhydrous ethanol, 4.39 g (0.015 mol) titanium isopropoxide (Ti[OCH(CH₃)₂]₄, 97%, Sigma-Aldrich) (St Louis, MO, USA) was dissolved then 1.2 g of sodium hydroxide dissolved in 10mL deionized water was added into the above solution. The obtained mixture was treated at 80 °C for 2 h to obtain precursor A. In 10mL deionized water, 3.165 g of (0.015 mol) strontium nitrate was dissolved in 20 mL of deionized water, and then 1.2 g of sodium hydroxide was added dropwise to obtain precursor B. Precursors A and B were mixed and sealed in a Teflon container and kept at 150 °C for 2 h. The obtained product was centrifuged and washed several times with deionized water, then dried at 80 °C for 48 h. After that, the product was grinded to 80 mesh and heated at 1000 °C in a muffle furnace for 3 h to obtain well crystalized SrTiO₃.

The Pt/SrTiO₃ catalyst was prepared using impregnation method. A certain amount of tetraammineplatinum dinitrate (H₁₂N₆O₆Pt, Macklin, 99%, Newport Beach, CA, USA) was dissolved in 25 mL of an Ethylenediamine (Aladdin, 99%, Seattle, WA, USA) solution (0.001 g/mL of Pt) and then mixed with 5 g SrTiO₃ under stirring. After a 24-h immersion at room temperature, the catalyst was dried at 80 °C for 48 h. The obtained product was grinded to 80 mesh and heated at 350 °C at 5 °C/min for 3 h.

The Pt/SrTiO₃-L catalyst was prepared by leaching Pt/SrTiO₃ catalyst. A total of 2.5 g of a Pt/SrTiO₃ catalyst was put into 50 mL HCl solution with a pH = 2 for 3 h. The product was centrifuged and washed several times with deionized water, then dried at 80 °C for 48 h.

3.2. Characterizations

3.2.1. Powder X-ray Diffraction (XRD)

XRD patterns were collected on a Bruker D8 Advance X-ray diffractometer (Billerica, MA, USA) using Cu Kα (λ = 1.5406 Å) radiation. Samples were tested on 2θ = 10°–90° at a scan rate of 10° min⁻¹.

3.2.2. BET and Inductively Coupled Plasma-Atomic Emission Spectrometry (ICP-OES)

Specific surface areas and pore structures of the catalysts were measured using ASAP 2020 Plus physisorption micromeritics (Micromeritics, Norcross, GA, USA). The true loading amount of Pt in a catalyst was measured on a Thermo IRIS Intrepid II (Waltham, MA, USA).

3.2.3. H₂-Temperature-Programmed Reduction (H₂-TPR)

Temperature-programmed-reduction (TPR) experiments were performed in a packed-bed flow microreactor (L = 22 inch, O.D. = 1/2 inch) in AutoChem II 2920 (Micromeritics). Samples of 50 mg mixed with 200 mg of SiO₂ were first pretreated under agron at 100 °C for 1 h. The sample then was cooled to 25 °C and heated from 25 °C to 700 °C in 10 °C/min under 10% H₂/He.

3.2.4. Scanning Transmission Electron Microscopy (STEM)

Samples were characterized by FEI Tecnai G2 F30 Field Emission Transmission Electron Microscope (Hillsboro, OR, USA) and Titan ETEM Themis (Bangalore, India). The particle size distribution was obtained by measuring no less than 100 particles via ImageJ software (Shenzhen, China).

3.2.5. In Situ Fourier-Transformed Infrared Spectroscopy

In-situ DRIFTS experiments were performed on a Thermo Scientific Nicolet iS50. The following two kinds of intake experiments were carried out:

(1) The samples were firstly exposed to 10%CO₂ in He at room temperature for 30 min and then 10%H₂ in He treatment for 10 min.

(2) The samples were treated under a mixed gas of 5%H₂+5%CO in He.

During these two processes, the IR spectra were collected at various temperatures under gas purging.

3.2.6. Density Functional Theory (DFT) Calculation Details

All of the density functional theory (DFT) calculations were performed using the Vienna ab-initio simulation package (VASP). The projector-augmented wave (PAW) method was used to represent the core-valence interaction [45,46]. The plane wave energy cutoff was set to 400 eV. The generalized gradient approximation (GGA) with the Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional and local Hubbard term U correction was used in our calculations. The values of U are 3.5 for Ti. The Brillouin zone was sampled at Gamma point with the 1 × 1 × 1 k-point meshes for the SrTiO₃ [110] surface. The energy and force criterion for the convergence of the electron density are set at 10⁻⁵ eV and 0.05 eV/Å, respectively. In order to research the size effect of Pt on SrTiO₃, the size of single Pt atoms and clusters (sever atoms) on the SrTiO₃ [110] surface were both calculated, considering the RWGS reaction on large Pt particles were studied extensively.

3.2.7. Extended X-ray Absorption Fine Structure

The k-edge extended X-ray absorption fine structure spectrum (EXAFS) of Pt was collected at the synchrotron radiation source (SSLS) center in Singapore, where the X-ray beam was monochromated with channel-cut Si [111] crystals. The k-edge absorption data of Pt was recorded in the transmission mode. Pt foil and PtO₂ were used as references. The working energy of the storage ring was 2.5 GeV, and the average electron current was less than 200 mA. The ATHENA module implemented in the IFEFFIT software package used to extract and process the acquired EAXFS data according to standard procedures. Using the Bessel window function, the k3-weighted Fourier transform (FT) of χ(k) in R space was obtained in the range of 0–14.0 Å⁻¹.

3.3. Activity Tests

All of the catalysts were performed in a fixed-bed reactor under atmospheric pressure. An electromagnetic flowmeter was used to control the flow rate of the feed gas. For the activity and stability test, 200 mg of a catalyst (80 mesh) mixed with 300mg of quartz sand (Macklin, AR, 50–80 mesh) was loaded in the center of a quartz tube (i.d. 10 mm). The gas flow of the reactants into the reactor was kept at 50 mL min⁻¹ (gas hourly space velocity = 6000 h⁻¹). After purging by He for 30 min at 200 °C, the temperature of the tubular furnace was elevated to the reaction temperature for testing. Descending temperature was used for the activity test. The gas products were analyzed using a gas chromatograph equipped with a flame ionization detector (FID) and a TCD (Fuli, GC9790plus, Taizhou, China). The RWGS activity of the catalyst was measured by the ratio of converted CO₂ to total CO₂.

4. Conclusions

In the present work, the single platinum atoms and nanoparticles on SrTiO₃ were prepared. The single atomic Pt catalyst showed excellent activity at temperatures higher than 300 °C, reaching a TOF of 0.643 s⁻¹ and 100% of CO selectivity at 500 °C in the RWGS reaction. The size effects of Pt in the RWGS reaction were researched using the DRIFTs and DFT methods. The results revealed that single Pt atoms have the highest activity and CO selectivity in RWGS via a “–COOH route”, while Pt metals of a larger particle size facilitate the further hydrogenation of CO. The reaction between formate and H* is the rate determination step of the RWGS reaction over Pt/SrTiO₃. The strong adsorption of formate (in situ FTIR results) on Pt clusters and nanoparticles may retard this reaction step.