Active Pt/CeO₂ catalysts prepared by an alcohol-reduction process for low-temperature CO-PROX reaction

Abstract

Pt/CeO₂ catalysts were prepared with 0.5 and 1 wt% of Pt loadings by an alcohol-reduction process using a solution of ethylene glycol and water as a reducing agent and solvent. The obtained catalysts were characterized by energy-dispersive X-ray spectroscopy, X-ray diffraction, and transmission electron microscopy. Transmission electron micrographs showed Pt nanoparticles with average sizes of 2.2 and 2.4 nm for Pt content of 0.5 and 1 wt%, respectively. The preferential oxidation of carbon monoxide in hydrogen-rich stream (CO-PROX reaction) was studied in the temperature range of 25–150 °C. Pt/CeO₂ catalysts showed maximum CO conversions in the range of 80–98% and CO₂ selectivity in the range of 50–70% at 50 °C.

Introduction

Nowadays, the hydrogen production worldwide is mainly employed in the ammonia synthesis reaction and there is an increasing interest as a clean combustible option for fuel cell technology [1]. Steam reforming of natural gas or the light oil fraction coupled with water gas shift reaction is the most widely used process to produce a H₂-rich gas mixture known as a reformate gas, which contains 15–20 vol% CO₂, 10 vol% H₂O, and ~ 1 vol% (10,000 ppm) of carbon monoxide (CO) [1, 2]. However, the catalysts used in the ammonia production and low-temperature fuel cell devices are very sensitive to CO. Therefore, H₂ must be high purity for both applications and the CO concentration must be decreased to less than 10 ppm [3,4,5]; although in some situations, CO levels between 50 and 100 ppm could be tolerated for use in proton exchange membrane (PEM) fuel cells. [1, 6].

Some processes have been used to remove CO from H₂-rich mixtures like pressure swing adsorption (PSA) that requires large capital investments and employ physic-sorbents to produce a very pure H₂ stream but with H₂ recovery values between 75 and 85% [1, 4, 5]. Another process is the methanation of CO (CO-MET), which operates at 300–400 °C, but causes significant loss of the produced hydrogen (10–15%) because of the unselective methanation of CO₂ present in the reformate gas [1]. In face of this, more efficient processes such as selective methanation of CO (CO-SMET) and the preferential CO oxidation reaction (CO-PROX) have been developed for producing high-purity hydrogen. The CO-SMET process is able to reduce the CO concentration at low levels using the H₂ itself present in the stream producing methane at temperatures ≥ 200 °C. The CO-SMET process tends to be more easily controllable, when compared to CO-PROX, since the CO and CO₂ methanation reactions are less exothermic than H₂ and CO oxidations; however, the H₂ consumption could be higher by about two times compared to the H₂ consumed for H₂O formation during CO-PROX processing [4,5,6,7]. Extensive research has been dedicated to the development of high selective catalysts to reduce the H₂ consume in the unselective CO₂ methanation [6, 7]. The CO-PROX process could avoid the hydrogen and energy loss, oxidizing CO at lower temperatures (in the range of 20–200 °C) with O₂ as oxidant; however, the catalysts for CO-PROX process need to exhibit high CO conversion and CO₂ selectivity, avoiding the oxidation of H₂ to H₂O. Thus, the development of more active and selective catalysts for CO-PROX process continues to be a challenge [1, 2].

Platinum nanoparticles supported on ceria (Pt/CeO₂) catalysts are one of the most active catalysts for CO-PROX reaction, leading to the maximum CO conversions and good CO₂ selectivity in the temperature range between 60 and 150 °C. For these catalysts, the temperatures at which maximum CO conversion occurs, as well as the best CO₂ selectivity values, were strongly dependent on the catalyst preparation method [8,9,10,11,12,13,14,15,16,17,18,19,20,21].

In this work, Pt/CeO₂ catalysts were prepared by an alcohol-reduction process [22], resulting in a very small and highly dispersed Pt nanoparticles on CeO₂ support, showing good CO conversions and CO₂ selectivity at 50 °C.

Experimental

Catalyst preparation

Pt/CeO₂ catalysts were prepared by an alcohol-reduction process [22] with Pt contents of 0.5 and 1 wt%, named, respectively, as Pt0.5/CeO₂ and Pt1/CeO₂. In this process, CeO₂ support (nanopowder with particle size ≤ 25 nm from Sigma-Aldrich) was dispersed in a solution of ethylene glycol/water (3/1, vol/vol) followed by the addition of the Pt precursor (H₂PtCl₆ 6H₂O—Sigma-Aldrich). The resulting mixture was placed in an ultrasonic bath for 5 min, and then, it was immersed in an oil bath and refluxed at about 150 °C under magnetic stirring. After 2 h under stirring at 150 °C, the mixture was cooled to room temperature and the solid was separated by centrifugation and washed with distilled water six times to remove chloride ions and reaction by-products. The obtained catalysts were dried overnight at 85 °C in an oven.

Characterization

The semi-quantitative chemical analysis of Pt/CeO₂ catalysts were performed by the energy-dispersive X-ray spectroscopy (EDX) using a Philips scanning microscope (model Jeol with electron beam of 20 keV) equipped with EDAX microanalyzer (model DX-4). The solid samples were deposited on the grids and the results refer to an average of four random points collected from each sample. For the study of morphological and structural properties, X-ray diffraction (XRD) and transmission electron microscopy (TEM) techniques were used. The crystalline structure of the catalysts was obtained in a Rigaku diffractometer model Miniflex II from 2θ = 20 to 90° with 0.05 step and 2 s count using Cu Kα radiation source (λ = 1.54 Å). The micrographs analysis was performed on transmission electron microscope model JEM-2100 (200 kV). For this, an aliquot of a suspension of the catalyst in 2-propanol was deposited on a copper grid (0.3 cm in diameter) with a carbon film. Eight micrographs were taken, allowing counting of about 200 Pt nanoparticles for each sample, to measure the sizes and particle-size distributions.

The temperature-programmed reduction (TPR) with H₂ measurements was performed on ChemBET Pulsar TPR/TPD chemisorption analyzer with a thermal conductivity detector (TCD). The catalysts (50 mg) in a U-shaped quartz cell were treated in a flow of N₂ (50 mL min⁻¹) at 200 °C for 1 h. After cooling to room temperature, the catalysts were exposed to 10% vol H₂/N₂ gas flow (30 mL min⁻¹) and heated to 900 °C with a heating rate of 10 °C min⁻¹.

Catalytic tests

CO-PROX reaction measurements were performed in gas phase, using a fixed bed reactor, containing 100 mg of catalyst. No catalyst activation treatment was performed prior to the catalytic experiments. The catalytic testing was conducted at atmospheric pressure and in two runs (cycles), each run in the temperature range from 20 to 150 °C. The experiments were performed with a gas stream consisting of 1 v% CO, 0.5–1 v% O₂, and 98–98.5 v% H₂, and flow rates of 25 mL min⁻¹ (O₂/CO ratio of 1, space velocity of 15,000 mL  g ⁻¹_cat   h⁻¹) and 50 mL min⁻¹ (O₂/CO ratio of 0.5, space velocity of 30,000 mL  g ⁻¹_cat   h⁻¹.) The reaction products and unconverted reagents were quantified by gas chromatography with TCD and FID (methanation of CO and CO₂) detectors. The CO and O₂ conversions and CO₂ selectivity were calculated according to the following equations:

$${\text{CO conversion}} = 100 \times \left( {\left[ {\text{CO}} \right]_{\text{in}} {-}\left[ {\text{CO}} \right]_{\text{out}} } \right)/\left[ {\text{CO}} \right]_{\text{in}} ,$$

(1)

$${\text{O}}_{ 2} \, {\text{Conversion}} = 100 \times \left( {\left[ {{\text{O}}_{{ 2 {\text{in}}}} } \right]{-}\left[ {{\text{O}}_{{ 2 {\text{out}}}} } \right]/\left[ {{\text{O}}_{{ 2 {\text{in}}}} } \right]} \right),$$

(2)

$${\text{CO}}_{ 2} \,{\text{selectivity}} = 100 \times \left( {0. 5* \left[ {{\text{CO}}_{ 2} } \right]_{\text{out}} } \right)/\left( {\left[ {{\text{O}}_{ 2} } \right]_{\text{in}} {-}\left[ {{\text{O}}_{ 2} } \right]_{\text{out}} } \right),$$

(3)

for CO₂ selectivity, the number 0.5 refers to stoichiometric coefficient, where 1 mol of CO reacts with 0.5 mol of O₂ to produce 1 mol of CO₂.

Results and discussion

The EDX analysis of the Pt/CeO₂ catalysts (Table 1) showed that the obtained Pt values (wt%) were similar to the nominal values, suggesting that all Pt(IV) ions were reduced and deposited on the CeO₂ support.

Table 1 Chemical composition of Pt/CeO₂ catalysts

The X-ray diffractograms of Pt/CeO₂ catalysts are shown in Fig. 1. In these diffractograms, it was observed diffraction peaks at 2θ = 28.3°, 32.9°, 47.3°, 56.2°, 58.9°, 69.2°, 76.5°, 78.9°, and 88.3°, referring to the diffraction pattern of the ceria cubic crystalline phase (ICSD # 72155). On the other hand, no peaks relative to face-centered cubic (fcc) phase of Pt nanoparticles at 2θ = 40°, 47°, 67°, 82°, and 87° were observed [22]. Since the EDX analyses showed the presence of Pt in these catalysts, the absence of the Pt (fcc) peaks could be associated with a very small size of the nanoparticles. Thus, the small size and low contents (wt%) of Pt result in broad and low-intensity peaks that were too small to be clearly detected by XRD [9, 16].

Fig. 1

X-ray diffractograms of CeO₂ support and Pt/CeO₂ catalysts

The transmission electron micrographs and the histograms of Pt/CeO₂ catalysts are shown in Fig. 2. The TEM micrographs of Pt0.5/CeO₂ and Pt1/CeO₂ catalysts, Fig. 2a, b, respectively, showed spherical Pt nanoparticles highly dispersed on CeO₂ support with average sizes of 2.2 and 2.4 nm, respectively. This shows that this methodology is adequate to produce very small Pt nanoparticles highly dispersed on CeO₂ support as already observed in the preparation of Pt/C catalysts (20 wt% of Pt) for low-temperature fuel cells [22].

Fig. 2

Transmission electron micrographs and histograms of (a) Pt0.5/CeO₂ and (b) Pt1/CeO₂

The H₂-TPR profiles of CeO₂ support and Pt/CeO₂ catalysts are shown in Fig. 3. As described in the literature [21, 23, 24], H₂-TPR profiles of Pt/CeO₂ catalysts showed three regions: the first one, in the range of 100–300 °C due to the reduction of PtO_x species to Pt(0) and/or to CeO_x species directly bonded to Pt reducing at lower temperatures; the second in the range of 300–600 °C assigned to the reduction of superficial CeO_x species; and the third in the region above 600 °C due to the bulk reduction of CeO₂. The H₂-TPR profile of CeO₂ support showed two reduction peaks: one at 456 °C assigned to the reduction of CeO₂ species on the surface and one at 790 °C within the bulk. Comparing H₂-TPR profiles of Pt/CeO₂ catalysts with the CeO₂ support, it can be seen the peak observed at 456 °C for CeO₂ support was shifted to 415 °C and 377 °C for Pt0.5/CeO₂ and Pt1/CeO₂ catalysts, respectively. This shift to lower temperatures indicated the existence of metal-support interactions, increasing the mobility of oxygen inside CeO₂ crystal lattice and improving the redox properties of the catalysts [21]. On the other hand, the peak at 790 °C observed for the CeO₂ support was practically in the same position for Pt/CeO₂ catalysts, which indicated that Pt did not influence the reduction of oxygen species within bulk of CeO₂ [23]. In addition, the H₂-TPR profiles of Pt/CeO₂ catalysts showed two small peaks in the range of 100–250 °C. Under the synthesis conditions of Pt/CeO₂ catalysts prepared by an alcohol-reduction process, the Pt(IV) ions are reduced to Pt(0), leading to the Pt nanoparticle formation [22], which suggest that these peaks could be attributed principally to CeO_x species directly bonded to Pt reducing at lower temperatures [21]. Thus, these results indicate that for Pt0.5/CeO₂ and Pt1/CeO₂ catalysts a strong metal-support interaction occurs preferentially on the surface of the CeO₂ than in the bulk.

Fig. 3

H₂-TPR profiles of CeO₂ support and Pt/CeO₂ catalysts

The catalytic performances of Pt/CeO₂ are shown in Fig. 4a–c. No previous treatments were done in these samples and the results shown correspond to the second cycle of the catalytic reactions. In a general manner, for all catalysts, it was observed after the first cycle (from 25 to 150 °C) an increase of the performance of the catalysts principally at low temperatures (below 75 °C). Probably, this increase of performance at low temperature in the second cycle could be due to a decrease of Pt cationic species and/or to a surface cleaning of the catalysts which lead to a better accessibility of the Pt surface under reaction conditions. Above 75 °C, the values of CO and O₂ conversion and CO₂ selectivity practically did not change from the first to the second cycle.

Fig. 4

Catalytic performance of Pt/CeO₂ catalysts: (a) CO conversion, (b) CO₂ selectivity, and (c) O₂ conversion (filled square Pt1/CeO₂ 1% O₂, filled triangle Pt1/CeO₂ 0.5% O₂, empty square Pt0.5/CeO₂ 1% O₂, and empty triangle Pt0.5/CeO₂ 0.5% O₂)

Initially, the catalysts Pt0.5/CeO₂ and Pt1/CeO₂ were tested using a volumetric O₂/CO ratio of 1 (λ = 2 − excess of O₂). For both catalysts, the CO and O₂ conversions were very low (around 25%) at 25 °C. When the temperature was raised to 50 °C both the CO conversion and the O₂ consumed increased to 98% for Pt0.5/CeO₂ catalyst while for Pt1/CeO₂ catalyst, a CO conversion increased to 83% and 65% of the O₂ was consumed. For both catalysts, the CO₂ selectivity values were in the range of 45–55%. Above 75 °C, the O₂ was totally consumed for both catalysts; however, the CO conversion and CO₂ selectivity values begin to decrease continuously up to 150 °C due to the undesirable water formation. According to the literature [9], at low temperature, there is a complete coverage of the Pt surface by CO, which is oxidized to a greater extent as the O₂ content increases in the gas stream. When the temperature increases, desorption of CO from Pt surface becomes important and it is partially replaced by H₂, favoring its oxidation and consequently decreasing CO₂ selectivity.

Pt0.5/CeO₂ and Pt1/CeO₂ were also tested using a volumetric O₂/CO ratio of 0.5 (λ = 1). When tested under stoichiometric conditions, the maximum CO conversion continues to occur at 50 °C for both catalysts; however, an increase of CO₂ selectivity values was observed for Pt0.5/CeO₂ catalyst. Although CO conversion at 50 °C decreased from 98 to 81% when the O₂/CO ratio decreased from 1 to 0.5, the CO₂ selectivity increased from 46 to 73%. Wootsch et al. [8] using Pt/CeO₂ prepared by impregnation and O₂/CO ratios of 0.5 and 1 observed similar values of maximum CO conversion and CO₂ selectivity; nevertheless, these values were obtained at 100 °C, while for our catalysts, it occurred at 50 °C.

To evaluate the stability of the catalysts after the second cycle, long-term experiments were performed at 50 °C (Fig. 5a–c). In all cases, the CO conversion, O₂ conversion, and CO₂ selectivity values were very similar to the ones observed in the second cycle at 50 °C, and these values remained stable throughout the period evaluated.

Fig. 5

Long-term tests of Pt/CeO₂ catalysts at 50 °C: (a) CO conversion, (b) CO₂ selectivity, and (c) O₂ conversion (filled square Pt1/CeO₂ 1% O₂, filled triangle Pt1/CeO₂ 0.5% O₂, empty square Pt0.5/CeO₂ 1% O₂, and empty triangle Pt0.5/CeO₂ 0.5% O₂)

Comparing our results with those reported in the literature (Table 2), it could be seen that most of Pt/CeO₂ catalysts showed maximum CO conversions at temperatures ≥ 80 °C. Only a few results [14, 16] described the maximum CO conversion at 60 °C but with CO₂ selectivity values ≤ 40%. Kugai et al. [14] prepared Pt/CeO₂ catalyst (2.2 wt% of Pt) by radiolytic process and obtained a maximum CO conversion of 100% at 60 °C using a high O₂/CO ratio of 2 (λ = 4), however, with a low CO₂ selectivity of 25%. Gao et al. [16] prepared Pt/CeO₂ catalysts with different Pt loadings and different morphologies of CeO₂ support by incipient wetness impregnation. The material prepared with 0.5 wt% of Pt and supported on rods CeO₂ showed a maximum CO conversion of 85% at 60 °C and CO₂ selectivity of 40% using an O₂/CO ratio of 1 (λ = 2).

Table 2 Comparison of the catalytic performance over Pt/CeO₂ catalysts for CO-PROX reaction reported in the literature

The good performance of our Pt/CeO₂ catalysts (Pt nanoparticles supported on nanosized CeO₂) for CO-PROX reaction at 50 °C, perhaps, could be explained by the following factors: the small sizes of Pt nanoparticles and the strong metal-support interactions. Gatla et al. [25] described that nanosized CeO₂ has an important role as a catalyst support due to the great amount of surface oxygen defects and in particular oxygen ion vacancies that are fundamental for reactions like CO oxidation and CO-PROX reaction. Besides that, also the presence of noble metals can change the CeO₂ surface properties, weakening Ce–O bond, and making the surface more reducible. The study of Pt nanoparticles dispersed on nanosized CeO₂ active at room temperature for CO oxidation using different techniques revealed that elongated Pt–O distance resulting from the interaction of Pt species with CeO₂ in the form of low-temperature active species–support interaction. Gao et al. [16] also observed for Pt/CeO₂ prepared by impregnation that the Pt precursor interacts in different ways with CeO₂ supports with different morphologies resulting in catalysts with different fractions of metallic Pt and Pt²⁺ species. Moreover, the promoting effect of Pt on the reducibility of CeO₂ supports and the concentration of oxygen vacancies varied with CeO₂ morphology and Pt–CeO₂ interaction. Polster et al. [13] studied CO and H₂ oxidation over a series of Pt/CeO₂ catalysts with different Pt loadings and dispersion, and showed that interfacial Pt–O–Ce sites are responsible for Mars and van Krevlen redox activity. Recently, Gänzler et al. [26] described that the Pt/CeO₂ interface can be tuning by the variation of the Pt nanoparticle sizes, which determines the number of interfacial sites between Pt nanoparticles and CeO₂. It was demonstrated that the formation of small Pt nanoparticles (in the range of 1 and 2 nm) induces variations in CeO₂ reducibility and this is a prerequisite for CeO₂ reduction at low temperatures. Besides that, the importance of an intimate and optimal interaction between Pt and CeO₂ activates the redox chemistry that is important for applications involving high oxygen storage capacity and enhanced low-temperature CO oxidation.

Conclusions

Pt/CeO₂ catalysts with Pt contents of 0.5 and 1wt% and average nanoparticles sizes of 2.2 and 2.4 nm could be prepared by a facile methodology. The obtained catalysts showed to be active, selective, and stable at 50 °C for CO-PROX reaction. This activity at low temperature could be a result of the optimal interaction of the small-sized Pt nanoparticles and CeO₂ nanoparticles used as support leading to a weakening of the CO adsorption on Pt sites and favoring O₂ adsorption/activation on active redox sites at Pt/CeO₂ interfaces.