Rh single atoms embedded in CeO2 nanostructure boost CO2 hydrogenation to HCOOH

Bo Wu,Xing Yu,Min Huang,Liangshu Zhong,*,Yuhan Sun,*

1 CAS Key Laboratory of Low-Carbon Conversion Science and Engineering,Shanghai Advanced Research Institute,Chinese Academy of Sciences,Shanghai 201210,China

2 University of Chinese Academy of Sciences,Beijing 100049,China

3 School of Physical Science and Technology,ShanghaiTech University,Shanghai 201210,China

Keywords:CO2 hydrogenation Selectivity Single-atom catalysts (SACs)Rh HCOOH

ABSTRACT CO2 hydrogenation to value-added chemicals is a promising pathway to solve CO2-relevant environmental problems but still remains a great challenge.Herein,we report a CeO2 nanostructure supported Rh single atoms (Rh-SAs/CeO2) catalyst and was used for the efficient CO2 hydrogenation to HCOOH.The Rh-SAs/CeO2 exhibited high catalytic activity with turnover numbers (TON) up to 221 at 200 °C,which was 4-fold to that of CeO2 supported Rh nanoparticles (Rh-NPs/CeO2).Moreover,HCOOH selectivity for Rh-SAs/CeO2 reached 85%,much higher than that of Rh-NPs/CeO2(46%).Mechanism studies revealed that Rh single atoms in the Rh-SAs/CeO2 with high metal atoms utilization efficiency not only provided abundant active sites to promote the catalytic activity,but also suppressed the decomposition of HCOOH to CO and benefited the formation of HCOOH.

1.Introduction

Catalytic hydrogenation of CO2to value-added chemicals has attracted much attention,which not only benefits the reduction of CO2-revelant environment problems,but also provides a feasible platform for carbon utilization and hydrogen energy storage [1,2].Catalytic hydrogenation of CO2with H2can be employed to produce various products including alcohol [3],hydrocarbon [4] and formic acid (HCOOH) [5,6].Among them,HCOOH is an important feedstock and can be used as hydrogen source for fuel cells [7].CO2hydrogenation to HCOOH has been widely investigated over various metals-based catalysts such as Pd-,Ru-,Pt-,Au-and Rh[8].As there exist different reaction pathways during the catalytic process of CO2hydrogenation,it is highly desirable to control the reaction network to improve both catalytic activity and HCOOH selectivity.

Single-atom catalysts (SACs) have been proposed as an ideal platform for many catalytic applications including CO2conversion[9].Zeng et al.developed Pt single atoms (Pt1@MIL) and Pt nanoparticles (Ptn@MIL) for CO2hydrogenation.It was found that Pt1@MIL exhibited higher turnover frequency number of 117 h-1and higher CH3OH selectivity of 90.3% via HCOO*intermediate,while Ptn@MIL showed lower catalytic activity (20.9 h-1) with CO (57.5%) as the major product via COOH*intermediate [10].A significant improvement in the activity and selectivity of CO2hydrogenation to CH3OH was achieved by sub-nano Re clusters[11].The Re clusters provided around 82% CH3OH selectivity with product formation rate up to 1.8 mmol.(mmol Re)-1.h-1,while Re single atoms showed lower catalytic activity (0.2 mmol.(mmol Re)-1.h-1) with 94% CO selectivity.Yamashita synthesized a layered double hydroxide (LDH) supported Ru single atoms for the hydrogenation to HCOOH.The unique triads of basic hydroxyl ligands promoted the electron-donating ability to the Ru single atoms and enhanced CO2adsorption capacity in the vicinity of the Ru single atoms,resulting in a significant increase of catalytic activity [12].Shrotri developed ZrO2supported Co single atoms for the hydrogenation to CO2to CO via formate intermediate at 340°C[13].The CoZrOx(1)catalyst showed the highest space time yield (STY) of 51 μmol.(mol Co)-1.s-1with more than 97% CO selectivity.Recently,Gutierrez and co-workers studied the effect of Rh-O-Fe bonds of Rh single atoms and Rh-Fe3O4on CO2hydrogenation to CO[14].Compared with Rh nanoparticles,the stronger CO2adsorption on Rh single atoms with lower activation energies led to higher CO2conversion rates.Generally speaking,the catalytic activity and products selectivity for CO2hydrogenation are directly affected by the structure of the active sites.In order to achieve both high catalytic activity and high HCOOH selectivity for CO2hydrogenation,it is highly necessary to fabricate an efficient catalyst with suitable structure of the active sites.

Herein,we prepared CeO2nanostructure supported Rh single atoms and nanoparticles to study the effect of active sites on the catalytic activity and products selectivity for CO2hydrogenation.Compared with Rh nanoparticles,the CeO2nanostructure supported Rh single atoms exhibited higher catalytic activity and higher HCOOH selectivity.In addition,the reaction mechanism was also investigated by in-situ spectroscopy characterizations.

2.Experimental

2.1.Catalyst preparation

For the preparation of CeO2nanostructure,27 g Ce(NO3)3.6H2O and 54 g NaOH were dissolved in 120 ml ultrapure water separately,then the two solutions were mixed and transferred to a 500 ml autoclave with polytetrafluoroethylene (PTFE) liner vessel.The autoclave was placed in an electric oven,heated to 120°C and maintained for 12 h.The resulting precipitate was washed with ultrapure water for several times,dried at 80 °C for 12 h.The asobtained powder was further calcined at 400 °C for 4 h.

Rh-SAs/CeO2and Rh-NPs/CeO2were prepared by wet impregnation method.In a typically procedure,3.0 g CeO2powder was dispersed in 30 ml ultrapure water at room temperature for 30 min.Then certain amount volume of RhCl3aqueous (0.0154 g Rh.ml-1) was then dropped into the suspension.After stirring for 3 h,the resulting precipitate was dried at 80 °C.Finally,the asobtained powder was calcined at 300 °C for 4 h.

2.2.Catalyst characterization

X-ray diffraction(XRD) measurements were tested on a Rigaku Ultima IV X-ray instrument with Cu Kα(λ=0.154056 nm)radiation at a beam voltage of 40 kV and 40 mA beam current.The scanning speed was 2(°).min-1at scanning range from 20°to 80°.The ultraviolet–visible (UV–vis) spectra were recorded on a Shimadazu 2600 instrument in the range from 200 to 800 nm.The elemental content of Rh in the CeO2sample was determined using inductively couple plasma optical emission spectroscopy (ICP-OES,Optima 2000DV).The Brunauer-Emmett-Teller (BET) specific surface area for the samples were evaluated by N2adsorption/desorption using a Tristar II system (Micromeritics Instruments,USA).Transmission electron microscopy (TEM) images were obtained on a JEOL JEM-2011 system operating voltage at 100 kV.Aberration-corrected high angle annual dark-field scanning transmission electron microscope (AC-HAADF-STEM) images were obtained JEOL JEM-ARM300F equipped with a CEOS probe corrector.X-ray photoelectron spectroscopy (XPS) analysis was performed on an ESCLAB250 spectrometer,using a monochromatic AlKα radiation source(1486.6 eV).The binding energies(BEs)were calibrated using the C1s peak at 284.8 eV as a reference.Raman measurements were conducted on a confocal Raman microscope(Thermo-Fisher,USA) using Ar+laser 532 nm for the excitation source under ambient conditions.The samples after reduction by hydrogen at 200 °C were also characterized by XPS and TEM characterizations.

Diffuse reflectance infrared Fourier transform spectroscopy(DRIFTS) of CO chemisorption was performed on a Thermo-Fisher Nicolet iS10 spectrometer equipped with an MCT detector and a low temperature cell.The sample was first pretreated with Ar(20 ml.min-1) at 150 °C to remove any contaminant.After cooling to 50 °C under Ar,a background spectrum was collected.Then the sample was exposed to CO at a flow rate of 20 ml.min-1for about 30 min until saturation.Ar was purged at a flow rate of 20 ml.min-1for another 30 min to remove gaseous CO and then the DRIFT spectrum was collected with 64 scans at a resolution of 4 cm-1in the range of 4000–650 cm-1.For the reaction mechanism studies,insitu DRIFTS was performed on the same apparatus for CO-DRIFTS experiments.The background spectrum was obtained after Ar flow under ambient pressure for 30 min at 200°C.For the in-situ DRIFTS experiments under a CO2or H2/CO2(3/1,v/v)atmosphere,the reactant gas was introduced into the cell with a flow rate of 20 ml.min-1and kept at 200 °C for 30 min.

Temperature programmed reduction of hydrogen (H2-TPR)experiments were performed on an AutoChem II 2920 apparatus.30 mg of the sample was loaded on the U-type tubes and pretreated in a gas flow of Helium (30 ml.min-1) at 200 °C for 30 min.After the pretreatment,the sample was cooled to room temperature and then reduced in a flow of 5%H2/N2(40 ml.min-1)from room temperature to 700 °C with a constant ramp rate of 10 °C.min-1.

2.3.Catalytic evaluation

CO2hydrogenation was evaluated in a 50 ml stainless-steel autoclave containing a Teflon liner vessel purchased from Shanghai Yan Zheng Instrument.Firstly,the vessel was charged with 10 mg catalyst and 20 ml H2O,then the vessel was flushed with CO2and charged with a total pressure of 4.0 MPa(1.0 MPa CO2and 3.0 MPa H2).The reaction temperature was maintained at 200 °C.After reaction,the autoclave was quickly cooled down to room temperature before analysis in order to minimize the loss of volatile products.The quantified method for determination of gaseous products(CO and CH4) by GC equipped with a methanizer,and liquid products (CH3OH and HCOOH) was quantified by1H NMR analysis as shown in previous work [15].

The HCOOH selectivity (%) and turnover number (TON) were calculated using the following Eqs.(1)-(2):

3.Results and Discussion

3.1.Structure characterizations

The phase structure for the as-synthesized CeO2and Rh/CeO2samples was studied by X-ray diffraction (XRD) as shown in Fig.1(a).All samples exhibited similar diffraction peaks located at 28.7°,33.2°,47.7°,56.5°,59.1°,69.4°,77.0°,79.0°,88.5°,which can be attributed to the fluorite CeO2phase (JCPDS #34–0394)[16,17].No peaks assigned to metallic Rh and/or Rh oxide can be detected after the introducing of Rh species into CeO2support due to the low loading amount or high dispersion of Rh species.Fig.1(b) shows the ultra-visible (UV–vis) spectrum for CeO2and Rh/CeO2.The absorbance peaks located in the range from 400 to 600 nm can be assigned to the oxidized Rh species.A significant increase of the intensity for Rh-NPs/CeO2revealed the obvious aggregation of Rh species [18].The Rh loading amount was determined by inductively coupled plasma optical emission spectroscopy (ICP-OES),which was 0.15% (mass) and 2.03% (mass) for Rh-SAs/CeO2and Rh-NPs/CeO2,respectively.The Brunauer-Emmett-Teller (BET) surface area decreased with the increase of Rh loading amount,which was 58.6,48.2 and 41.6 m2.g-1for CeO2,Rh-SAs/CeO2and Rh-NPs/CeO2,respectively.

Fig.1.XRD patterns (a) and UV–vis spectra (b) for CeO2,Rh-SAs/CeO2 and Rh-NPs/CeO2.

Fig.2.TEM,HRTEM and AC-HAADF-STEM images(a),(b),(c)for Rh-SAs/CeO2.TEM,HRTEM and AC-HAADF-STEM images(d),(e),(f),HAADF-STEM image and EDS mapping(g) for Rh-NPs/CeO2.

The structure for Rh/CeO2was further characterized by electron microscopy.Transmission electron microscopy (TEM) images in Fig.2(a)and Fig.2(d)show that CeO2mainly exhibited a morphology of nanorods with length about 80–100 nm.The lattice distance for CeO2was 0.31 nm(Fig.2(b)and 2(e)),which can be assigned to the(111)plane of CeO2[19–21].No Rh-based nanoparticles can be detected for the Rh-SAs/CeO2from high resolution TEM (HRTEM,Fig.2(b)) image,suggesting that Rh species in the Rh-SAs/CeO2might exist in subnanometer scale or even in single-atom level.Aberration corrected high angle annular dark field transmission electron microscopy (AC-HAADF-STEM) image in Fig.2(c) shows that the Rh species were atomically dispersed on the surface of CeO2for Rh-SAs/CeO2.For the sample of Rh-NPs/CeO2,Rh species existed in the form of RhO2nanoparticles (Fig.2(e) and 2(f)).The lattice distance of 0.32 nm can be attributed to the (110) plane of RhO2[22].HAADF-STEM image and the corresponding X-ray energy dispersive spectroscopy (EDS) mapping results (Fig.2(g))indicated the obvious aggregation of Rh species for Rh-NPs/CeO2.

In situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) of CO adsorption experiments were performed(Fig.3(a)).For the sample of Rh-SAs/CeO2,two obvious peaks centered at 2095 and 2030 cm-1were observed due to the symmetric and asymmetric vibration of gem-dicarbonyl doublet CO on positively charged Rh atoms,respectively [23–25].As for Rh-NPs/CeO2,four peaks located at 2095,2030,2003 and 1925 cm-1were observed.Comparing with Rh-SAs/CeO2,the new peaks located at 2003 cm-1and 1925 cm-1can be assigned to the linearly bound carbonyl on Rh species and CO adsorbed Rh-CeO2interface,respectively[26].X-ray photoelectron spectroscopy(XPS)measurements show that the Rh 3d5/2binding energy peak for both Rh/CeO2samples was located dominantly at 309.3 eV(Fig.3(b)),indicating that the Rh species were in high valance state[24,27].The structure for Rh-SAs/CeO2and Rh-NPs/CeO2was also investigated by reducing the samples with hydrogen at 200 °C.As shown in Fig.S1a,the Rh 3d5/2for Rh-NPs/CeO2after H2reduction can be resolved into three peaks located at 309.9,308.7 and 307.1 eV,which can be assigned to the Rh3+,positively charged Rh (Rhδ+,0 <δ <3) and metallic Rh species (Rh0),respectively.The presence of Rhδ+(37.1%) and Rh0(18.0%) suggested that the easy reduction for Rh-NPs/CeO2.For Rh-SAs/CeO2after H2reduction,the binding energy for Rh 3d5/2slightly shifted to low binding energy comparing with fresh Rh-SAs/CeO2(Fig.3(b)).The binding energy of Rh 3d5/2was located between Rh3+and Rhδ+,verifying the high oxidation state of Rh species in the Rh-SAs/CeO2.These results suggested that Rh species were mainly atomically dispersed in the Rh-SAs/CeO2with better structure stability comparing with Rh-NPs/CeO2(Fig.S1b,1c and 1d).Fig.3(c) shows the Raman spectra of CeO2and Rh/CeO2.The peaks located at 464 and 600 cm-1were due to the F2gmode of fluorite phase of ceria and defect (D) mode,respectively [28].The peaks of F2gshifted from 464 cm-1for CeO2to 454 cm-1for Rh-NPs/CeO2,suggesting the increase of oxygen defect with Rh doping.The ratio of AD/AF2gwas calculated to study the intrinsic defect concentration.The AD/AF2gratio decreased in the order of CeO2(0.06)

Fig.3.CO-DRIFTS (a),XPS (b) in Rh 3d region,Raman spectra (c) and H2-TPR (d) for various samples.

3.2.Catalytic performance

CO2hydrogenation was performed on a batch reactor with a total pressure of 4.0 MPa with CO2/H2ratio of 3 at 200 °C (Fig.4(a)).We firstly tested the catalytic performance with the same Rh amount (1.97 μmol) for each catalyst,corresponding to 10 mg Rh-NPs/CeO2and 135 mg Rh-SAs/CeO2(Fig.4(a)).It was clear that Rh-SAs/CeO2showed superior catalytic performance compared with Rh-NPs/CeO2.The corresponding TON value was 53 for Rh-SAs/CeO2.In addition,the corresponding HCOOH selectivity reached up to 85%,and only trace CO and CH4was detected.However,the TON value was only 32 for the Rh-NPs/CeO2,and HCOOH selectivity was as low as 40% with more than 49% CO selectivity.

To explore the differences in catalytic properties between Rh-SAs/CeO2and Rh-NPs/CeO2,we investigated the catalytic performance at different reaction time.As shown in Fig.4(b),for Rh-SAs/CeO2,the TON value increased from 125 to 221 with the reaction time prolonged from 5 h to 12 h,while HCOOH selectivity slightly increased from 60% to 67%.However,the TON value for Rh-NPs/CeO2only increased from 32 to 50,and HCOOH selectivity remained less than 50% within 12 h reaction time.The Rh-SAs/CeO2with Rh single atoms as active sites exhibited 4-fold catalytic activity to that of Rh-NPs/CeO2,which may be attributed the high metal utilization efficiency comparing with Rh NPs as active sites.

The effect of catalyst amount on the CO2hydrogenation for Rh-SAs/CeO2was also investigated(Fig.4(c)).The TON value increased from 53 to 145 with the decrease of catalyst mass from 135 mg to 10 mg,while HCOOH selectivity declined from 85% to 65%.Moreover,the effect of mole ratio of CO2/H2was also explored (Fig.4(d)) over Rh-SAs/CeO2.No significant change was found when the CO2/H2ratio was tuned,and the TON value kept around at 130 with 60% HCOOH selectivity.Stability test in Fig.S2a showed that the yield for oxygenates (HCOOH and CH3OH,with HCOOH as the major liquid product)was 39.5 μmol for the first cycle.About 33%catalytic activity decreased (26.4 μmol) when the catalyst was used for three times.AC-STEM and EDS mapping results (Fig.S2b and S2c) revealed that Rh mainly existed in the form of single atoms,and the decrease of catalytic activity may be due to the partial aggregation of Rh single atoms.

Some recent works for CO2hydrogenation to HCOOH have been summarized and showed in Table 1.For the selective hydrogenation of CO2to HCOOH,the additive of bases(NEt3,NaOH,NaHCO3)or alcohol(CH3OH)will significantly promote the catalytic activity.However,our Rh-SAs/CeO2catalyst showed similar catalytic activity without the adding of bases or alcohol compared with reported works.

Table 1 Catalytic performance for CO2 hydrogenation to HCOOH over various catalysts

Fig.4.Catalytic performance for CO2 hydrogenation.(a) Comparison of TON for Rh-SAs/CeO2 and Rh-NPs/CeO2.Reaction conditions:200 °C,5 h,1.0 MPa CO2,3.0 MPa H2,20 ml H2O.Catalyst weight:135 mg for Rh-SAs/CeO2 and 10 mg for Rh-NPs/CeO2,corresponding to 1.97 μmol Rh;(b) Effect of reaction time for Rh-SAs/CeO2 and Rh-NPs/CeO2.Reaction conditions:200°C,10 mg of the catalyst,1.0 MPa CO2,3.0 MPa H2,20 ml H2O.(c)Effect of catalysts amount for Rh-SAs/CeO2.Reaction conditions:200°C,5 h,1.0 MPa CO2,3.0 MPa H2,20 ml H2O.(d) Effect of CO2/H2 ratio for Rh-SAs/CeO2.Reaction conditions:200 °C,10 mg of the catalyst,5 h,20 ml H2O.

3.3.Mechanism studies

The CO2adsorption experiments at 200 °C were performed to investigate the adsorption behaviour on the Rh-SAs/CeO2and Rh-NPs/CeO2.As shown in Fig.5(a),upon CO2adsorption on Rh-SAs/CeO2,three peaks located at 1646,1249 and 1020 cm-1can be clearly observed,which can be attributed to the vibrations of bidentate bridged carbonate.Whereas the peak located at 1078 cm-1can be assigned to the monodentate carbonate [41].Similar peaks were found at 1635,1236 and 1036 cm-1for Rh-NPs/CeO2,verifying the same adsorption model for CO2on Rh-SAs/CeO2and Rh-NPs/CeO2(Fig.5(b)).Especially,a new peak located at 1888 cm-1can be observed with time on stream,which was attributed to the bridge adsorption of CO.There were no peaks located in the region of 1450–1400 cm-1,which excluded the possible existence of bicarbonate [41,42].The appearance of three peaks located at 1646,1249 and 1036 cm-1confirmed the existence of bidentate bridge carbonate.

Fig.5.In-situ DRIFTS results of CO2 adsorption for Rh-SAs/CeO2(a)and Rh-NPs/CeO2(b).CO2 hydrogenation(CO2 and H2 co-adsorption)for Rh-SAs/CeO2(c)and Rh-NPs/CeO2(d) at 200 °C from 1000 to 3200 cm-1.

To further study the catalytic mechanism of CO2hydrogenation over Rh-SAs/CeO2and Rh-NPs/CeO2,in situ DRIFTS measurements under the atmosphere of CO2and H2at 200°C were also performed to analyze the surface species.A set of peaks appeared at 1020,1078,1250,1505,1643,1796,1908,2027,2177,2701,2806 and 2948 cm-1for Rh-SAs/CeO2(Fig.5(c)).The peaks at 2948,2806,2701 and 1796 cm-1were assigned to formate,whereas the peaks located at 1643,1250 and 1020 cm-1were assigned to bidentate bridged carbonate,and the peaks located at 1505 and 1078 cm-1were assigned to monodentate carbonate.The peaks located at 1908,2027 and 2177 cm-1can be assigned to CO adsorbed to the Rh-CeO2interface,the linear-CO adsorption on Rh species and gaseous CO,respectively,indicating that the partial aggregation of Rh single atoms for Rh-SAs/CeO2[25].The gradually disappearance of 1250 and 1020 cm-1with the gradually appearance of monodentate (1078 and 1505 cm-1,1 min,red line),then the appearance of formate (2948,2806,2701,1796 cm-1,3 min) and carbonyl(2177,2027 and 1908 cm-1,3 min)suggested the gradually transformation of bidentate bridged carbonate to monodentate,which was then converted to formate and CO (Scheme 1(a)).For Rh-NPs/CeO2(Fig.5(d)),however,no peaks around 1250 cm-1and 1036 cm-1can be observed under the coadsorption of CO2and H2compared with the sole CO2adsorption on Rh-NPs/CeO2(Fig.5(b)),verifying the fast conversion of bidentate bridged carbonate on Rh-NPs/CeO2.Meanwhile,the peaks of bicarbonate (1625 and 1403 cm-1,1 min) [42],monodentate(1509 and 1075 cm-1,1 min),formate (2840 cm-1,1 min),formyl(1744 cm-1,1 min) [43] and CO (2038 and 1932 cm-1,1 min)appeared once the CO2and H2reacted on the Rh-NPs/CeO2.Therefore,it can be reasonably inferred that the formed bidentate bridged carbonate on Rh-NPs/CeO2would be quickly converted into bicarbonate and monodentate,and then to formyl group and formate,which will finally cause the appearance of CO species(Scheme 1(b)).

Scheme 1.Proposed reaction mechanism for CO2 hydrogenation over Rh-SAs/CeO2 (a) and Rh-NPs/CeO2 (b).

Therefore,we draw the conclusion that the Rh SA can suppress the fast decomposition of bidentate bridged carbonate to monodentate and hinder the decomposition of formate to CO,thus resulting in high HCOOH selectivity.

4.Conclusions

In summary,two kinds of CeO2supported Rh-based catalysts including single atoms and nanoparticles were prepared for CO2hydrogenation.Rh-SAs/CeO2with Rh single atoms as the active sites exhibited higher catalytic activity with TON values of 221,which was 4-fold to that of Rh-NPs/CeO2with Rh nanoparticles as active sites.In addition,HCOOH selectivity reaches as high as 85% for Rh-SAs/CeO2,while HCOOH selectivity was only 46% with CO selectivity of 49% for Rh-NPs/CeO2.It was suggested that CeO2supported Rh single atoms (Rh-SAs/CeO2) with high atom efficiency greatly promoted the catalytic activity.In addition,Rh single atoms suppressed the decomposition of HCOOH to CO,thus favored the formation of HCOOH with high selectivity.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by Natural Science Foundation of China (91945301),Program of Shanghai Academic/Technology Research Leader(20XD1404000),Key Research Program of Frontier Sciences of the Chinese Academy of Sciences (QYZDB-SSWSLH035),the ‘‘Transformational Technologies for Clean Energy and Demonstration”and Strategic Priority Research Program of CAS(XDA21020600)and the Youth Innovation Promotion Association of CAS.

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.11.018.