Development of Mo-Modified Pseudoboehmite Supported Ni Catalysts for Efficient Hydrogen Production from Formic Acid.

Formic acid (FA), as a safe and renewable liquid hydrogen storage material, has attracted extensive attention. In this paper, a series of Mo-modified pseudoboehmite supported Ni catalysts were developed and evaluated for efficient hydrogen production from formic acid. Pseudoboehmite (PB) as a catalyst carrier was used for the first time. Ni/PB and NiMo/PB possessed a mesostructure, and the pore size distribution was mainly concentrated between 2 and 20 nm. The oxygen vacancies caused by Mo enhanced Ni anchoring, thus inhibiting Ni sintering. Compared with Ni10/PB (7.62 nm), Ni10Mo1/PB had smaller Ni particles (5.08 nm). The Ni-O-Al solid solutions formed through the interaction of Ni with the PB improved the catalytic performance. Ni10Mo1/PB gave the highest conversion of 92.8% with a H2 selectivity of 98% at 300 °C, and the catalyst activity hardly decreased during the 50 h stability test. In short, Ni10Mo1/PB was a promising catalyst for hydrogen production from formic acid because of the oxygen vacancy anchoring effect as well as the formation of Ni-O-Al solid solutions which could effectively suppress the Ni sintering.

Introduction

To achieve the aim of carbon neutrality, using hydrogen energy to replace fossil fuels is an available method. At present, hydrogen energy is difficult to popularize due to the security risks in hydrogen storage and transportation. The key to solving this problem is the “hydrogen carrier”, which can store and transport hydrogen efficiently.[1]

Formic acid (FA, HCOOH), a biorefinery major subproduct, is a nontoxic and cheap hydrogen carrier.[2−4] Its weight and volume H2 capacity are 4.4 wt % and 53.4 g/L, respectively.[5] The ideal circulation path of formic acid as a hydrogen carrier is shown in Figure . The reaction of H2 and CO2 can produce formic acid, which is transported to a hydrogenation station to extract hydrogen and can be used for fuel cells. The CO2 produced during the reaction can also be separated and recycled.[6] FA can be decomposed into H2 and CO2 or CO and H2O by dehydrogenation (eq ) or dehydration (eq ). However, the dehydration should be avoided for the purpose of hydrogen production.[7] This is because the dehydration reaction produces CO, which will reduce the activity of the catalyst.[8] It is well-known that catalysts have an important influence on the decomposition of formic acid.

Figure 1

Schematic diagram of ideal hydrogen storage and the hydrogen production cycle of formic acid.

FA catalysts for hydrogen production can be divided into homogeneous and heterogeneous catalysts. Homogeneous catalysts based on organometallic complexes such as Ru,[9] Ir,[8] and Rh[10] usually have high selectivity and catalytic activity. However, the recovery of homogeneous catalysts requires special additives and organic solvents, which limits their practical application. It is easy to recover heterogeneous catalysts based on Pt,[11] Au,[12,13] and Pd,[10,14−16] but the noble metals are expensive.[17] Heterogeneous catalysts based on non-noble metals such as Ni,[18,19] Cu,[20] and Co[21] have received much attention. In addition, hydrogen production from formic acid catalyzed by non-noble metal catalysts is usually carried out in the gas phase, but the low concentration of FA has led to a low hydrogen production rate in most of the literature.[22] Therefore, it is urgent to develop a non-noble metal catalyst with high activity and strong stability for hydrogen production from high-concentration FA. Nickel-based catalysts are widely used because of their low cost and high ability to promote the formation of H2 molecules.[23] The activity and stability of the catalyst can be effectively improved by selecting a suitable carrier and promoter. The catalyst carrier can interact with the active metal to change the properties of the active metal.[24] Pseudoboehmite (PB) is a layered material with high specific surface area that can interact with active metals to improve the dispersion and catalytic performance of active metals. However, up to now, there have been few reports about PB as a catalyst carrier for hydrogen production from formic acid. As an excellent promoter, Mo can strengthen the interaction between the active metal and the catalyst carrier and improve the dispersibility of the active metal.[25]

In this paper, new NiMo/PB catalysts were developed for hydrogen production from formic acid, in which Ni, Mo, and PB were an active metal, promoter, and carrier, respectively. The catalysts were characterized by XRF, XRD, BET, SEM/EDX-mapping, TEM, FTIR, H2-TPR, NH3-TPD, and XPS to measure their physicochemical properties. The catalysts were applied to hydrogen production from formic acid in a fixed bed reactor at 300 °C to evaluate the catalytic performance. The relationship between Ni content, Ni/Mo ratio, and catalytic performance was studied, and the mechanism of formic acid decomposition was revealed.

Results and Discussion

Catalyst Characterization

The X-ray diffraction (XRD) spectra of Ni/PB and NiMo/PB are shown in Figure . The peaks at 14.5°, 28.2°, 38.3°, 49.3°, 67.6°, and 72.3° belong to (020), (021), (130), (002), (220), and (221) planes of pseudoboehmite.[26] The catalysts have two special peak positions of 34.1° and 60.5°, respectively. For Ni/PB series catalysts, the peak intensity at 34.1° increases with the increase in Ni content from 5% to 20%. For NiMo/PB series catalysts, the peak at 34.1° intensity increases with the increase in NiMo ratio from 1:1 to 10:1. The peak at 60.5° shows the same phenomenon. These peaks are nickel species interacting with PB to form the Ni–O–Al solid solutions in the matrices.[27−29]

Figure 2

XRD patterns of Ni/PB and NiMo/PB.

The diffraction peaks of NiO (JCPDS PDF 44-1159) with 2θ values of 37.2° (110), 43.3° (200), and 62.9° (220) were not detected. Diffraction peaks belonging to Ni0 (JCPDS PDF 04-0850) with 2θ values of 44.3° (111) and 51.8° (200) were also not detected. The diffraction peaks associated with Ni and Mo species were not found in the spectra, and the same phenomenon was reported by Qian et al.[30] This can be attributed to two reasons. One is that the crystal structure of NiO and Ni0 is highly amorphous, or the degree of crystallization is poor and cannot be detected by XRD. The other reason is that pseudoboehmite with a high specific surface area can promote a high dispersibility of Ni and Mo, which is beyond the detection range of XRD, indicating that Ni and Mo species with large particle sizes are not formed on pseudoboehmite.[31−33]

The N2 adsorption–desorption isotherms and associated pore size distributions are shown in Figure . According to the classification of IUPAC, the isotherm is determined as type IV, which indicates that Ni/PB and NiMo/PB belong to a typical mesoporous material.[34,35] In addition, adsorption–desorption curves show the characteristics of an H3 type hysteresis loop. The specific surface area (SBET) and volume (Vp) of Ni/PB and NiMo/PB are shown in Table . With an increase in the Ni/Mo ratio, SBET first increases and then decreases for NiMo/PB catalysts, and the Ni8Mo1/PB has the highest SBET (282.8 m2/g). It is worth noting that the changes of SBET in the catalysts are accompanied by the change of the interaction between the metal and support,[36−38] which can suppress the agglomeration and sintering of active metal.[39] Therefore, the Ni/Mo ratio can change the interaction between Ni and PB, resulting in an increase in SBET from 273.8 m2/g for Ni10Mo1-PB to 282.8 m2/g for Ni8Mo1/PB and then a decrease to 199.4 m2/g for Ni1Mo1/PB. Furthermore, the pore size distribution was mainly concentrated between 2 and 20 nm, and the average pore size is 5–6 nm. These results demonstrated that Ni/PB and NiMo/PB have a mesoporous structure with uniform pore size distributions.

Figure 3

(a) N2 adsorption–desorption isotherms and (b) pore distribution of Ni/PB and NiMo/PB.

Table 1

Textural Properties of Ni/PB and NiMo/PB

samples	Ni contenta (wt %)	Mo contenta (wt %)	SBETb (m2/g)	Vpb (cm3/g)	Dpb (nm)
Ni5/PB	5.7		283.2	0.48	6.0
Ni10/PB	10.7		274.9	0.41	5.9
Ni20/PB	20.8		280.8	0.41	5.7
Ni1Mo1/PB	5.3	4.7	199.4	0.30	5.9
Ni4Mo1/PB	8.8	1.9	277.3	0.37	5.6
Ni8Mo1/PB	9.5	1.2	282.8	0.41	5.7
Ni10Mo1/PB	9.7	0.9	273.8	0.39	5.3

Obtained by XRF for catalysts.

Obtained by BET for catalysts.

The contents of Ni and Mo introduced into the as-prepared catalysts were determined by X-ray fluorescence (XRF) and are presented Table , which confirmed that the Ni and Mo were successfully loaded on PB and that the Ni and Mo contents of catalysts were almost similar to the given metal contents during catalyst preparation. Figure shows the energy-dispersive X-ray spectroscopy (EDX) mapping images of Ni and Mo for Ni/PB and Ni10Mo1/PB. Ni and Mo are uniformly loaded on PB. Transmission electron microscopy (TEM) images are shown in Figure . For Ni10/PB, Ni agglomeration occurs, resulting in serious sintering. Compared with Ni10/PB, the addition of Mo inhibits the occurrence of Ni agglomeration in Ni10Mo1/PB, which helps to improve the catalyst stability. From the particle size distribution histogram of Ni, Ni particles in Ni10Mo1/PB have a higher dispersibility and smaller particle size (the average Ni particle sizes in Ni10/PB and Ni10Mo1/PB are 7.62 and 5.08 nm, respectively). The addition of Mo leads to the generation of oxygen vacancies,[40] promotes the anchoring of Ni, and then inhibits the Ni sintering.

Figure 4

SEM images and particle size distributions. (a) Ni10/PB, (b) Ni10Mo1/PB, and (c) Ni10Mo1/PB EDX images. (d) Ni in Ni10/PB, (e) Ni in Ni10Mo1/PB, and (f) Mo in Ni10Mo1/PB

Figure 5

TEM images and particle size distributions. (a) Ni10/PB. (b) Histogram frequency for Ni10/PB. (c) Ni10Mo1/PB. (d) Histogram frequency for Ni10Mo1/PB.

The Fourier transform infrared spectroscopy (FTIR) spectra of Ni/PB and NiMo/PB are shown in Figure . The peak at 1637 cm–1 is related to the bending vibration modes of adsorbed water.[41,42] The peak at 1072 cm–1 is attributed to the stretching vibration modes of the C–O bond. The peak at 3455 cm–1 is related to interacting OH or to bridging OH groups.[43−45] The peak at 1395 cm–1 is assigned to the −OH stretching vibration modes in (Al, Mg)–OH and (Fe, Mg)–OH units,[28,35] and this peak position moves to the left with an increase in Ni content, which may be attributed to the reaction of PB and Ni species to form a new species.[46] OH species can remove carbon on the catalyst surface, which can enhance the coking resistance of the catalysts and subsequently improve catalytic performance.[47] The peaks below 1000 cm–1 are attributed to the interatomic vibrations of Ni–O.[48]

Figure 6

FTIR spectra of Ni/PB and NiMo/PB.

A H2-temperature-programmed reduction (H2-TPR) analysis was used to study the Ni/Mo reducibility and the interaction between Ni/Mo and PB. As shown in Figure , the inverted peak at 280 °C is attributed to the desorption of bound water. The peak at 350 °C for Ni5/PB belongs to an α reduction peak,[29] which is attributed to the insignificant or little interaction of NiO on PB. For Ni/PB, the peak gradually shifts to the right as the Ni content increases. For NiMo/PB, the peak gradually shifts to the right as the Ni/Mo ratio decreases. This indicates that the increase in Mo content enhances the interaction of NiO on PB. The reduction of Mo is usually divided into two steps:[49] one reduction peak located at 600–800 °C is the low-temperature reduction peak of Mo, where Mo6+ is reduced to Mo4+, and the other reduction peak at 800–900 °C is the high-temperature reduction peak of Mo, where Mo4+ is reduced to Mo0. The special peak of Ni1Mo1/PB at 600–900 °C is due to the increased intensity at 600–900 °C caused by more Mo content.

Figure 7

H2-TPR profiles of Ni/PB and NiMo/PB.

The acidity of the catalysts has an important effect on the catalytic performance. NH3 can be absorbed in the acidic position of the catalysts. Therefore, NH3-temperature-programmed desorption (NH3-TPD) was performed to test the acidity of Ni/PB and NiMo/PB. There are three types of NH3 desorption peaks located at 130, 230, and 350 °C, which are sequentially assigned to weak, medium, and strong acidic positions.[48] As shown in Figure , the Ni10Mo1/PB catalysts have an obvious peak at 423 °C, and it belongs to strong acidic positions. With an increase in Mo content, the peak position at 423 °C shifts to the right, which indicates that the addition of Mo forms more acidic sites on the catalyst surface.

Figure 8

NH3-TPD profiles of Ni/PB and NiMo/PB.

The interactions in Ni/PB and NiMo/PB are further investigated by X-ray photoelectron spectroscopy (XPS) characterization, and Ni 2p, Mo 3d, and Al 2p spectra are shown in Figure . Figure b shows the XPS spectra of Ni 2p; the peak near 857 eV is associated with Ni 2p3/2, and the peak near 874 eV is associated with Ni 2p1/2, which have corresponding satellite peaks at 862 and 880 eV, respectively (sat.). According to previous reports, the binding energy of Ni 2p3/2 at about 854 eV is associated with a large amount of isolated NiO, while the binding energy at 857 eV is associated with a strong interaction of Ni2+ and PB.[29] The 854 eV peak is not found in the spectra, indicating that the isolated NiO species has been completely reduced to the Ni species.[27] Ni2+ species found in the range 857–871 eV are Ni species reacting with PB to form Ni–O–Al solid solutions; the peaks associated with Ni2+ species increased significantly with increasing Ni content, indicating that the PB interaction with Ni species formed more Ni–O–Al solid solutions, and Ni–O–Al solid solutions formed through the interplay of Ni with the PB play a key role in catalytic performance.[47] Based on the study of Mo4+ 3d3/2 (231.81 eV), Mo6+ 3d5/2, and Mo6+ 3d3/2 (232.82 and 235.13 eV), Mo 3d spectra (Figure c) can be divided into three subpeaks.[50] To further confirm the formation of Ni–O–Al solid solution, the XPS spectra of Al 2p are shown in Figure d; the peak at about 73 eV belongs to the aluminum metal phase (Al), while the peak at 75 eV belongs to the fully oxidized aluminum substance (Al2O3).[28,51] In Figure d, all catalysts have a main peak at about 74.4 eV. This peak does not belong to the Al metallic phase or Al2O3 phase but belongs to the Al–O linkage formed through the Al3+ species at the octahedral sites of the PB skeleton binding with the O species in PB. The central position of the Al 2p spectra is shifted to high BE values with increasing Ni content. Distinctively, this phenomenon is more predominant in Ni10Mo1/PB, further validating that there are more Ni–O–Al species in Ni10Mo1/PB.[28]

Figure 9

XPS spectra of Ni/PB and NiMo/PB. (a) XPS survey spectrum, (b) Ni 2p, (c) Mo 3d, and (d) Al 2p.

Figure 10

(a) FA conversion rate (XFA), H2 yield (YH), and CH4/CO/CO2 selectivity of Ni/PB and NiMo/PB. (b) Carbon-containing product selectivity. (c) H2 production rate. (d) Ni10Mo1/PB stability test.

Catalytic Performance Test

Figure shows the FA conversion rate, H2 yield, H2 selectivity, and carbon-containing product selectivity for Ni/PB catalysts (Ni5/PB, Ni10/PB, and Ni20/PB) in the hydrogen production experiments from formic acid. With an increase in Ni content from 5 to 10 wt %, the FA conversion rate increases from 27.8% to 82.1%. When the Ni content is 20 wt %, the FA conversion rate decreases slightly to 81.4%. The H2 yields of Ni5/PB and Ni20/PB are 81.8% and 80%, respectively. The H2 yield of Ni10/PB decreased to 75.7%. With an increase in the Ni content, the FA conversion rate increases gradually, while the H2 yield decreases with an increase in FA conversion rate. However, compared with Ni10/PB, the FA conversion rate of Ni20/PB decreases slightly, only changing by 0.7%, while the H2 yield increases from 75.7% to 80%. The Ni10/PB catalyst shows the highest FA conversion rate, mainly because it has the smallest active metal particle size and the largest active metal surface area, which can provide more active sites for the reaction process of hydrogen production from formic acid.[52] Therefore, in the following experiments, Ni10/PB is modified to improve the catalyst performance.

Figure shows the catalyst performance with different Ni/Mo ratios (1:1, 4:1, 8:1, and 10:1). An increase in the Ni/Mo ratio can improve the FA conversion rate and H2 yield. The FA conversion rate of Ni10Mo1/PB is 92.8%, and the H2 yield first increases and then decreases with an increase in the Ni/Mo ratio. The H2 yield of Ni8Mo1/PB reaches a maximum of 83.4%. The increase in the FA conversion rate and H2 yield is mainly due to the decrease in Ni particle size, which provides more active sites for hydrogen production from formic acid.[53] The above characterization methods have demonstrated that the addition of Mo significantly reduces the Ni particle size and increases the interaction between Ni and PB. In addition, the effects of various catalysts on the carbon-containing product selectivity are shown in Figure b. The CH4 selectivity gradually increases with an increase in Ni/Mo ratio, and the CH4 selectivity reaches a minimum of 0.095% for the Ni1Mo1/PB catalyst; the CH4 selectivity of the Ni10Mo1/PB catalyst reaches a maximum of 0.83%. The CH4 in the product is produced by a methanation reaction (CO + 3H2 = CH4 + H2O). Therefore, it can be assumed that the increase in Mo content can inhibit the methanation reaction. The change trend of CO2 selectivity is consistent with that of H2 selectivity, while the change trend of CO selectivity is to decrease and then increase. The H2 production rate increases with increases in the Ni/Mo ratio (Figure c). The average H2 production rate of the Ni10Mo1/PB catalyst for 6 h reaches 99 mL/min. In addition, the hydrogen production experiment from formic acid is carried out without catalyst, and formic acid shows no decomposition phenomenon. The catalytic phenomenon of PB and Mo10/PB is studied: the PB catalyst shows low activity (Figure a), and the FA conversion rate is 17.5% and CO selectivity 93.1%. Mo10/PB has a higher activity than PB, and the FA conversion rate is 38.9% and CO selectivity 85.5%; this phenomenon is also observed by Chiang et al.[54] Therefore, it can be inferred that Ni can promote formic acid decomposition toward the dehydrogenation path.

The currently reported research progress of nonprecious metal catalysts for the hydrogen production reaction from formic acid is summarized in Table . Carrales-Alvarado et al. prepared a series of Ni/Cu-based catalysts using different nanostructural carbon materials as carriers, investigated the effect of N doping on the catalyst performance, and found that formic acid reached complete conversion at 180–210 °C and that hydrogen selectivity was 94–98%.[19] Bulusheva et al. reported the effect of Ni/carbon on hydrogen production from formic acid and found that, at 300 °C, the conversion rate of FA was 60%, and the hydrogen selectivity was 96%.[55] Junli Wang et al. synthesized a high-performance Mo-based catalyst using soybean and molybdate. When the soybean and molybdate ratio was 1:0.1, the obtained catalyst had an excellent performance, and formic acid could be completely transformed at 110 °C.[24] Bulushev et al. deposited MoS2 on graphene flakes by decomposition of MoS3 in vacuum conditions at different temperatures (500–800 °C) as a catalyst for hydrogen production from formic acid and found that the conversion rate of formic acid was 92%, and hydrogen selectivity was 42% at 235 °C.[56] Rodríguez-Ramos et al. investigated the catalysts of Ni and Ca supported SiO2 for hydrogen production from formic acid and found that the hydrogen selectivity was 94.5%.[57] Faroldi et al. synthesized Ni, Cu, and Ni–Cu catalysts loaded with high-specific-surface-area graphite by the impregnation method and investigated the effect of alkali metal (Li, Na, and K) doping. The bimetallic Ni–Cu catalyst had the best catalytic performance; the formic acid reached complete conversion, and hydrogen selectivity was 95% at 130 °C.[58] Kazakova et al. prepared catalysts of Co loaded on carbon nanotubes for hydrogen production from formic acid. At 300 °C, they found that the conversion rate of FA was close to 100% for of 7.5% Co/MWCNT, but the hydrogen selectivity was low (85%).[21] Bulushev et al. prepared Ni/C single-atom catalysts; the conversion rate of FA was 40% at 350 °C, and the hydrogen selectivity was 95–97%.[59] The conversion rate of FA was significantly lower than the results in this paper, which may be due to the low active metal content for the single-atom catalyst. Compared with the above catalysts, the Ni10Mo1/PB catalyst in this paper exhibited an excellent catalytic performance. Since the catalyst stability is an important factor in assessing the feasibility of industrial applications for hydrogen production reactions from formic acid, the catalyst stability was tested. Figures –13 show the gas yield, gas content, and gas selectivity of Ni/PB and NiMo/PB in the hydrogen production reaction from formic acid. Ni/PB and NiMo/PB always maintain an excellent catalytic performance in the 6 h catalytic reaction. In addition, the stability of Ni10Mo1/PB is tested for 50 h; the FA conversion, H2 yield, and carbon-containing product selectivity almost remain constant, indicating that Ni10Mo1/PB possesses excellent stability during 50 h of reaction.

Table 2

Research Progress of Non-Noble Metal Catalysts for Hydrogen Production from Formic Acid

	reaction conditions
catalyst type	carrier gas flow rate (mL/min)	formic acid concentration (%)	reaction temperature (°C)	formic acid conversion (%)	hydrogen selectivity (%)	ref
NiCu/NCNT	25	5.5	180–210	100	94–98	(19)
Ni/CN-1073	67	2.5	300	61	97	(55)
Soy–Mo (0.1)	50		110	100		(24)
MoS2/graphene	51	1.8	235	92	42	(56)
Ni/Ca(19.3)–SiO2	25	6	160	100	92	(57)
Ni–Cu/K	25	6	130	100	95	(58)
7.5% Co/MWCNT	20	5	300	100	85	(21)
1Ni/C	67	2.5	300	40	95–97	(59)
Ni10Mo1/PB	200	41	300	92.8	98

Figure 11

Gas yield of Ni/PB and NiMo/PB catalysts: (a) hydrogen, (b) carbon dioxide, (c) carbon monoxide, and (d) methane.

Figure 13

Gas selectivity of Ni/PB and NiMo/PB: (a) hydrogen, (b) carbon dioxide, (c) carbon monoxide, and (d) methane.

Gas content of gas yield of Ni/PB and NiMo/PB: (a) hydrogen, (b) carbon dioxide, (c) carbon monoxide, and (d) methane.

Mechanism Analysis of Formic Acid Decomposition

The mechanism of formic acid decomposition is shown in Figure . The C=O double bond energy in the FA molecule is much larger than the C—H and O—H bonds,[60−62] so there are generally two paths for FA decomposition. One is that the O—H bond of FA breaks to produce HCOO* and H*, and the C—H bond of HCOO* breaks to produce CO2 and H*. Then, H* combines with H* to form H2. The other path is that the C—H bond of FA breaks to produce COOH* and H*, and the C—O bond of COOH* breaks to produce CO and OH*; then, OH* combines with H* to form H2O. COOH* may also break the O—H bond to form CO2 and H*, and then, H2 is produced. However, the O—H bond of COOH* has a larger bond energy relative to the C—H bond,[63] so it is easier for COOH* to produce CO. Therefore, the ideal direction for the hydrogen production reaction from FA is HCOOH → HCOO* + H* → CO2 + H* + H* → CO2 + H2. The CH4 is obtained from the methanation reaction of H2 and CO.

Figure 14

Reaction mechanisms of HCOOH decomposition.

Conclusions

In conclusion, Mo-modified pseudoboehmite supported Ni catalysts have been prepared and evaluated for hydrogen production from formic acid. It was found that both the PB textural properties and Mo-promoter had significant roles in the catalytic performance. The PB textural structure was beneficial to obtain small Ni particles, and Mo led to the generation of oxygen vacancies, enhancing Ni anchoring and thus inhibiting Ni sintering. The Ni/PB and NiMo/PB had mesoporous structures, and the pore size distributions were mainly concentrated between 2 and 20 nm. Compared with Ni10/PB (7.62 nm), Ni10Mo1/PB had a smaller Ni particle size (5.08 nm). The Ni–O–Al solid solutions formed through the interaction of Ni with PB, which improved the catalytic performance. Ni10Mo1/PB gave the highest conversion of 92.8% with a H2 selectivity of 98% at 300 °C, and the catalyst activity hardly decreased during the 50 h stability test. Therefore, Ni10Mo1/PB was a promising catalyst for hydrogen production from formic acid.

Experimental Section

Catalyst Preparation

A series of Ni/PB catalysts with different Ni contents (5%, 10%, and 20%) were prepared by the chemical reduction method. 3 g of PB and a certain amount of Ni (NO3)2·6H2O (0.78, 1.65, and 3.72 g) were added into 300 mL of deionized water and stirred for 3 h. A certain amount of NaBH4 (1.01, 2.15, and 4.83 g) was dissolved in 100 mL of deionized water. After it was completely dissolved, NaBH4 aqueous solution was dropped into the above suspension at a rate of 2 mL/min. After that, the solution was stirred for 1 h. Then, the catalysts were centrifuged and washed three times with deionized water. Finally, the catalysts were dried at 60 °C in a vacuum oven for 12 h. The catalysts are named Ni5/PB, Ni10/PB, and Ni20/PB according to the different Ni contents.

A series of Ni–Mo/PB catalysts with different Ni/Mo ratios (1:1, 4:1, 8:1, and 10:1) were prepared by the chemical reduction method, in which the total mass of Ni and Mo accounted for 10% of the total mass of the catalysts. 3 g of PB and certain amounts of Ni (NO3)2·6H2O and (NH4)6Mo7O24·4H2O were added into 300 mL of deionized water and stirred for 3 h. A certain concentration of an aqueous solution of NaBH4 (1.70, 1.96, 2.03, and 2.05 g; 100 mL of deionized water) was dropped into the above suspension at a rate of 2 mL/min. The following steps were similar to that of Ni/PB catalysts. The catalysts are named Ni1Mo1/PB, Ni4Mo1/PB, Ni8Mo1/PB, and Ni10Mo1/PB according to the different Ni/Mo ratios.

Catalytic Test

Catalytic activity tests were carried out in a fixed-bed flow reactor with continuous feeding, as shown in Figure . First, the catalyst wrapped in quartz cotton was placed in the center of the reaction tube, and the oxidizing gas in the reaction tube was removed by a 200 mL/min N2 flow. Then, the reaction tube was heated to 300 °C, and the feed rate of formic acid (0.23 mL/min) was controlled by a constant flow peristaltic pump using 200 mL/min N2 as the carrier gas. The produced gas is cooled by a flowing water cooler to remove the condensable components. The composition and content of the gas were detected by a portable flue gas analyzer (Gasboard-3100). The FA conversion rate (XFA), H2 yield (YH), H2 selectivity (SH), and carbon-containing products selectivity (Si) are determined by the following formula:where i stands for CO, CO2, or CH4.

Figure 15

Diagram of the experimental setup.

The amount of the active metal was analyzed with an XRF spectrometer (ZSX-100e, Rigaku Corporation). The determination conditions were as follows: Rh target; element range B–U; 7–148° scan; tube voltage of 60 kV; and tube current of 150 mA. The XRD patterns were recorded by a Rigaku Corporation SmartLab (3 kW) X-ray diffractometer utilizing Cu Kα radiation (λ = 0.154 056 nm) over 5–80° at 40 kV and 40 mA. Fourier transform infrared spectroscopy (FTIR) was conducted on a Nicolet 5700 spectrometer in 4000–400 cm–1 with a resolution of 2 cm–1.

SEM images were obtained on a Quanta 250 FEI microscope operating at 20.0 kV. TEM images were carried out by an FEI Tecnai G2 F20 transmission electron microscope at an operating voltage of 200 kV. The particle size of the active metal was obtained using Nano Measurer 1.2 software (Beijing Zhongke Baice Technology Service Co., LTd, Beijing, China). Additionally, the average particle size was statistically calculated by measuring more than 100 uniform particles from several selected TEM images.

The N2 adsorption–desorption isotherms were recorded by a surface area and porosity analyzer (JW-BK200C). Prior to the measurements, all catalysts were preheated to 250 °C for 4 h under a vacuum. The reduction behavior and metal–carrier interactions of the catalysts were evaluated by H2-TPR experiments on an AutoChemII 2920 instrument. First, about 10 mg of sample was pretreated at 200 °C for 2 h under a 40 mL/min Ar flow to remove the adsorbed moisture and then cooled down to 50 °C. Second, the sample was heated up to 900 °C at 10 °C/min under a 40 mL/min 5% H2/Ar flow. The effluent gas was forced into the thermal conductivity detector (TCD) to detect the hydrogen consumption. NH3-TPD was conducted to evaluate the surface acidity of the catalysts on an AutoChem II 2920 instrument. The catalysts were heated up to 300 °C at 10 °C/min under a 30 mL/min He flow and kept at 300 °C for 1 h to remove the adsorbed volatile substances. Then, a mixture of 5% NH3/He was administered to saturate the surface. The physically adsorbed NH3 was removed by a 30 mL/min He flow for 1 h at 50 °C. Then, the sample was heated to 800 °C at 10 °C/min under a He flow to desorb the chemisorbed NH3.

All XPS measurements were performed by a Thermo Fisher Scientific ESCALAB 250Xi spectrometer using monochrome Al Kα (1486.68 eV) radiation as the excitation source under a vacuum of 0 mbar. The spectrometer was operated at 150 W and utilized a 500 μm beam spot. The obtained binding energies (BEs) were corrected using the 284.8 eV signal of adventitious hydrocarbon C 1s.