The graphitic carbon strengthened synergetic effect between Pt and FeNi in CO preferential oxidation in excess hydrogen at low temperature

Abstract

A variety of PtFeNi catalysts supported on carbon materials have been prepared and tested for CO preferential oxidation (PROX) in excess hydrogen. 100% O₂ and CO conversions have been achieved over carbon black (CB) and carbon nanotube (CNT) supported PtFeNi catalysts at room temperature in a feed gas containing 1% CO, 0.5% O₂ (volume ratio) and H₂ balance gas. N₂ adsorption, temperature-programmed desorption (TPD) and transmission electron microscopy (TEM) studies indicate that the carbon textural properties and surface chemistry determine the catalyst particle size distribution and mean size; but the mean particle size does not have a great influence on the catalytic performance within the investigated particle size range. X-ray diffraction (XRD), resistance measurements and the designed catalytic reaction results reveal the ability of graphitic carbon to capture and shuttle electrons from the noble metal to spatially different sites in the FeNi species through the π–π network, enables the indirect interactions between Pt and the FeNi species, leading to a strengthened synergistic effect, enhancing the CO oxidation activity at room temperature, increasing the Pt utilization efficiency, and apparently decreasing the Pt loading level.

---

1. Introduction

Proton exchange membrane fuel cells (PEMFCs) have been considered to be one of the most promising candidates to replace conventional fossil fuels, due to their high energy efficiency, environmentally-friendly characteristics and low operating temperature.^1,2 However, most commercial hydrogen fuels are generally produced through various hydrocarbons or bio-alcohols, from fossil fuels or biomass, via the steam reforming and water-gas shift (WGS) reactions. The resultant H₂-rich gases generally contain 0.5–1.0 vol% of CO due to the thermodynamic limitation of the WGS reaction.^1–22 Unfortunately, a small amount of CO severely poisons the anode catalyst of the PEMFCs, therefore, reducing the concentration of CO to ppm level is required prior to the introduction of hydrogen to the PEMFCs.^1–22 Among the various methods investigated, low temperature preferential oxidation of CO in excess H₂ streams (PROX) is presently regarded as one of the most promising and cost-effective ways.^1–22

Among the various catalysts, Pt-based catalysts especially Fe,^5–8,20,21 Co,^9–14 Ni,^14–19,21 Cu,^11,22 and Ag²³etc. promoted Pt catalysts are the most promising candidates. However, it is extremely difficult to realize the total conversion of CO at ambient temperatures with reasonable costs, which is particularly important for fuel cell applications in transportation.^10,13,17,20 To the best of our knowledge, among all the supported Pt catalysts reported in the literature, there are only a few supported Pt catalysts that can accomplish the total conversion of CO at room temperature or even lower,^4,5,21 mainly including our PtFe/SiO₂ ⁵ and later developed PtFeNi/CNT²¹ catalysts.

In order to find effective, low-cost and highly robust Pt-based catalysts, besides the promotion effects of promoters, the effects of various supports on the catalytic performance have been extensively investigated in the promoted catalytic system, including CeO₂,¹¹ TiO₂,^12,14 SiO₂,^4,5,14 Al₂O₃,^8,11,22 zeolites¹⁰ and carbon materials^6,9,13–21etc. The results indicated that some supports could help to distribute the active component uniformly,¹⁷ and some could interact with the active component²⁴ or influence the chemical state of the active component.¹² Especially, Petkov et al. reported the role of support–nanoalloy (PtCoNi) interactions on CO oxidation.¹⁴ Their investigation indicated that by controlled thermo-chemical treatment, oxygen activation sites on TiO₂ supported nanoalloys could be provided both by the second/third metal sites in the nanoalloy (Type-I sites) and by the anionic oxygen deficiency sites located at the nanoalloy-support perimeter zone (Type-II sites). However, only Type-I sites can be obtained for nanoalloys on carbon supports; neither can be generated on SiO₂ supported nanoalloys, resulting in the lowest activity for the silica-supported nanoalloys.

Petkov's research is based on the assumption that the promotion effects of base transition metals originates from Pt-nanoalloys. However, our extensive investigation demonstrated that the coordinatively unsaturated transition metal cations confined in nanosized Pt matrices were the main oxygen-activation sites.^5–7,9,21 Therefore, Petkov's discovery could not explain the extremely high activity of SiO₂ and carbon materials supported PtM (M = Fe, Ni, Co) catalysts for CO PROX and CO oxidation.^{5–7,9,13,21} Our previous research indicates that a 3 wt% PtFeNi/CNTs catalyst is extremely highly active, and can almost totally remove CO in the gas composition of 1 vol% CO, 0.5 vol% O₂ and H₂ balance even at 6 °C.²¹ Compared with other highly active catalysts,^5,6,9,13 this catalyst consists of an evidently lower Pt loading level, operating with stoichiometric O₂, a much higher concentration of H₂ and nearly the same gas hourly space velocity. Our further investigation indicates that in situ formed coordinatively unsaturated FeO_x and/or NiO_x species confined in Pt matrices are active species and the over oxidation of Fe and Pt species seriously deactivates the catalysts.²¹ These discoveries are similar to those found with our previously reported PtFe/SiO₂ catalysts;⁵ then, what leads to the extremely high activity of PtFeNi/CNTs even with apparently lower Pt loadings? Is it due to the unique properties of CNTs or the general properties of carbon?

Recently, the ability of carbon nanotubes and graphene-based systems to capture and shuttle electrons through the π–π network has been confirmed through spectroscopic studies.^25,26 This feature enables the incorporation of a semiconductor and Pt nanoparticles into a single graphene or reduced graphene oxide sheet, and the resulting multifunctional photocatalyst has been demonstrated for use in photocatalytic H₂ production.^27,28 This photocatalyst can carry out selective catalytic processes at separate sites and tune the selectivity and efficiency of photo-catalytic reduction and oxidation processes, independently. As far as our highly active PtFeNi/CNTs catalyst is concerned, do CNTs possess the ability to capture electrons and shuttle them, and will this ability contribute to the high catalytic activity or not?

In order to answer the above questions and reveal the origin of the extremely high activity of the PtFeNi/CNTs catalyst compared with PtFe/SiO₂ and other catalysts, four types of carbon: CNTs, carbon black (CB) with high graphitization, oxidized carbon black (CB-o), and activated carbon (AC) with less graphitization, were used to prepare PtFeNi catalysts and the effects of their catalytic properties on CO PROX were tested. CB can be utilized to exclude the unique properties of CNTs, for instance, confinement effects; CB-o can be used to investigate the negative effects of carbon surface oxidation on the ability of carbon to capture electrons and shuttle them; AC can facilitate the deliberation of the roles of the graphitization degree and the well-developed pore structure on the catalytic performance. Therefore, the effect of the properties on the catalytic performance were investigated through N₂ adsorption, TPD, TEM, XRD and resistance measurements. The discovery from this research will further direct us to design and prepare more effective, economical and robust catalysts for CO PROX in H₂ rich streams.

2. Experimental

2.1. Materials

Multi-walled carbon nanotubes with a 4–8 nm i.d. and a 10–20 nm o.d. were purchased from Chengdu Organic Chemicals Co., Ltd., China, which contained 0.41 wt% Fe and 0.35 wt% Ni, denoted as CNTs(FeNi). The CNTs(FeNi) purification was conducted using a procedure from the literature,²¹ and the obtained sample was labeled as CNTs-p. Carbon black was purchased from Acros, and marked as CB. CB was oxidized in 5% O₂ balanced in He (volume ratio) at 600 °C for 3 h, and the obtained material was denoted as CB-o. Commercial granulated activated carbon (sieved through a 40–60 mesh) was obtained from Beijing Guanghua Woods Ltd., Beijing, China. The carbon was washed with boiling aqueous nitric acid (0.001 N) and dried at 120 °C overnight. Then, the sample was further treated in He at 600 °C for 3 h and the obtained sample was denoted as AC.

2.2. Catalyst preparation

Pt catalysts were prepared using the conventional wetness impregnation method at room temperature using an ethanol solution of hexachloroplatinic acid (Shenyang Chemical Reagent Company, AR) according to the metal loading. The samples were further dried at 120 °C for 12 h. The Fe and Ni promoted catalysts were prepared using the sequential wetness impregnation method. For example, for the FeNi(H500)Pt catalyst, the carbon support was first impregnated with an ethanol solution of a ferric nitrate and nickel nitrate mixture followed by drying at 120 °C for 12 h and then pretreated in H₂ at 500 °C for 2 h, and finally impregnated with an ethanol solution of hexachloroplatinic acid. For all PtFeNi catalysts, the loadings of Pt, Fe and Ni were 3, 0.41 and 0.35 wt% (atomic ratio, 15 : 7 : 6), respectively; for the PtFe and PtNi catalysts, the loadings of Fe and Ni were 0.75 and 0.81 wt% (the same atomic ratio, 15 : 13), respectively. The catalysts were pretreated in H₂ at 500 °C for 2 h before the catalytic tests.

2.3. Catalytic reaction test

The catalytic reaction tests were performed in a fixed bed flow reactor. A small quartz tube containing a thermocouple was placed in the middle of the catalyst bed. Generally, 60 mg of the catalyst sample was used. A gas mixture containing 1% CO and 0.5% O₂ (volume ratio) in H₂ was fed at a flow rate of 25 ml min⁻¹, with a gas hourly space velocity of 25 000 ml g⁻¹ h⁻¹. The composition of the effluent gas was monitored on-line with a gas chromatograph (Agilent Technologies GC-6890N) equipped with PN and TDX-01 columns. The CO conversion was calculated from the change in CO concentration, and the selectivity towards CO₂ was obtained from the O₂ mass balance. The details were elaborated in our previous publications.^29,30

2.4. Characterization

The texture properties of the carbon supports were determined using N₂ adsorption–desorption isotherms at 77 K using an AS-1-MP adsorption instrument. Samples were degassed at 383 K overnight before the measurements.²⁹ The specific surface areas were calculated using the BET equation, and the t-plot method was used to calculate the micropore volumes and areas.²⁹ The temperature-programmed desorption (TPD) experiments were carried out in a quartz U type microreactor, connected with an on-line quadruple mass spectrometer (Balzers, OmniStar GSD300 O). The samples (60 mg) were first flushed with He (30 ml min⁻¹) at room temperature for 2 h. Subsequently, the temperature was increased to 990 °C at a rate of 5 °C min⁻¹. MS intensities for 2 (H₂), 4 (He), 28 (CO), and 44 (CO₂) were measured as a function of temperature.

Transmission electron microscopy (TEM, TECNAI Spirit) was used to study the size of the metal particles. XRD measurements were carried out on a Rigaku D/Max 2500 diffractometer with a Cu Kα monochromatized radiation source at 40 kV and 250 mA. The sheet resistivities were measured by a conventional SZ-82 four-probe instrument (Suzhou Tel. Apparatus Co., China).

3. Results and discussion

3.1. Reaction results

3.1.1. The catalytic performance of different carbon supported Pt catalysts. Fig. 1 presents the variation in the O₂ conversion (A), CO conversion (B) and selectivity (C) over different carbon supported 3 wt% Pt catalysts as a function of the reaction temperature. Pt/CNTs-p exhibited the highest catalytic activity with the highest O₂ and CO conversions among the studied catalysts. For example, the O₂ and CO conversions were 35.7% and 37.1% at 23 °C, respectively. At around 125 °C, O₂ was completely consumed and the CO conversion reached the maximum of 77.8% for the Pt/CNts-p sample. The other three catalysts (supported over CB, CB-o, and AC) showed very low reactivity below 120 °C. The full O₂ conversion was obtained at 200 °C over the Pt/CB catalyst and the corresponding maximum CO conversion was 66.6%. For the Pt/CB-o catalyst, the full O₂ conversion was obtained at 240 °C and the corresponding maximum CO conversion was 56.7%. For the Pt/AC catalyst, the full O₂ conversion was obtained at 210 °C and the corresponding maximum CO conversion was only 51.2%. The O₂ reactivity of the four catalysts followed the order of Pt/CNTs-p > Pt/CB > Pt/AC > Pt/CB-o, whereas the maximum CO conversion followed the order of Pt/CNTs-p > Pt/CB > Pt/CB-o > Pt/AC.

Fig. 1 Variation in the O₂ conversion (A), CO conversion (B), and selectivity (C) over different carbon supported 3 wt% Pt catalysts as a function of the reaction temperature.

3.1.2. The catalytic performance of different carbon supported Pt-based catalysts. Fig. 2 presents the O₂ conversion (A), CO conversion (B) and selectivity (C) over different carbon supported Fe and Ni promoted Pt catalysts at 23 °C after reacting for a certain time. Since the PtFeNi/CNTs-p catalyst after activation at 500 °C in H₂ can nearly completely remove CO at 6 °C in a feed gas containing 1% CO, 0.5% O₂ (volume ratio) and H₂ balance,²¹ and H₂ oxidation occurs at a temperature higher than 6 °C, the produced water apparently displays promotion effects. Therefore, in order to compare the catalytic performance at the same reaction temperature, the Pt/CNTs(FeNi), PtFeNi/CB, PtFeNi/CB-o and PtFeNi/AC catalysts were selected. All of the catalysts lost activity to some extent. The Pt/CNTs(FeNi) and PtFeNi/CB catalysts had better catalytic reactivities and stabilities, especially the Pt/CNTs(FeNi) catalyst, and both of them showed nearly complete O₂ conversion and CO conversion at 23 °C. The PtFeNi/CB-o and PtFeNi/AC catalysts lost activity very quickly. Furthermore, PtFeNi/AC exhibited the worst reactivity and stability, and displayed only 40% and 30% initial O₂ and CO conversions at 23 °C, respectively. The initial catalytic activity decreased in the following order: Pt/CNTs(FeNi) ≥ PtFeNi/CB > PtFeNi/CB-o > PtFeNi/AC.

Fig. 2 Variation in the O₂ conversion (A), CO conversion (B), and selectivity (C) at 23 °C over different carbon supported Fe and Ni promoted Pt catalysts as a function of the reaction time.

These results indicate that the addition of Fe and Ni can dramatically enhance the catalytic activity of Pt catalysts; additionally, the catalysts can be deactivated to some extent, which has been extensively investigated in our previous publications.^5,21 In addition, carbon supports also play important roles in the catalytic performance. On one hand the same PtFeNi catalysts with different carbon supports exhibited different catalytic performances; on the other hand, Pt and PtFeNi catalysts with the same carbon supports gave different catalytic activity trends. For example, the Pt/AC catalyst presented a significantly higher activity (O₂ conversion at the same temperature) than Pt/CB-o; on the contrary, the PtFeNi/AC catalyst showed a significantly lower activity (O₂ conversion at 23 °C) than PtFeNi/CB-o. Furthermore, it is interesting that all PtFeNi catalysts exhibit relatively stable selectivities although there is some loss in O₂ and CO reactivity. In order to elucidate the effects of the carbon supports on the catalytic performance, the catalysts have been investigated by a number of characterization methods (vide infra).

3.2. Characterization results

3.2.1. Textural properties and surface chemistry of the carbon supports

The nitrogen adsorption/desorption isotherms of the different carbon supports are shown in the ESI† (Fig. S1) and the corresponding textural parameters are listed in Table 1. The isotherm profile of AC is a typical IV type with rich micropores and mesopores; the calculated specific surface area and mean pore diameter (MPD) are 1158.0 m² g⁻¹ and 2.1 nm, respectively. The isotherm profile of CB-o is similar to that of AC but has many more mesopores and fewer micropores. For CB-o, the calculated specific surface area and MPD are 210.2 m² g⁻¹ and 11.2 nm, respectively. In comparison with CB-o, CB has much fewer mesopores and micropores and the corresponding specific surface area and MPD are 75.8 m² g⁻¹ and 6.5 nm, respectively. For CNTs, the calculated specific surface area and inner pore size are 172.2 m² g⁻¹ and 4–8 nm, respectively.

Table 1 The pore texture parameters and the integration areas of the CO and CO₂ desorption from TPD

Sample	Surface area (m^2 g^−1)	Mean pore size (nm)	CO^a	CO2^a	Total^a^b
a Unit: ×10^−7. b The total amounts of released CO and CO2.
AC	1158.0	2.1	4.08	0.49	4.57
CB-o	210.2	11.2	1.35	0.25	1.60
CB	75.8	6.5	0.04	0.03	0.07
CNTs	172.2		0.74	0.17	0.91

---

Temperature-programmed desorption (TPD) in an inert gas (He, Ar, N₂) is effective in determining the different oxygen-containing groups of carbon materials.^29–34 The TPD curves of CO and CO₂ from different carbon supports are shown in the ESI† (Fig. S2) and the corresponding integration areas of the CO and CO₂ desorption from TPD are listed in Table 1. It is clear that AC has many more acidic oxygen containing functional groups, CB-o possesses a moderate amount of oxygen containing functional groups, CNTs hold fewer oxygen containing functional groups compared with CB-o and AC, and CB hardly bears any oxygen containing functional groups.

3.2.2. Particle sizes of different carbon supported PtFeNi catalysts

Fig. 3 presents the transmission electron microscopy (TEM) images and the corresponding particle size distributions of the different carbon supported PtFeNi catalysts after activation in H₂ at 500 °C for 2 h. The PtFeNi particle size distribution on AC (1.5–6 nm) is a little broader than that on the other three supports. The average particle size of the PtFeNi/AC catalyst is about 3.7 nm, which is much bigger than that of the others. The particle size distribution of the PtFeNi/CB-o catalyst is very narrow with an average particle size of 1.6 nm. The PtFeNi/CB catalyst mainly consists of particles with a diameter from 2 to 3 nm, giving an average particle size of about 2.5 nm. The Pt/CNTs(FeNi) catalyst exhibits a very narrow particle size distribution, giving an average particle size of about 1.6 nm. The average particle size of the four catalysts decreased in the following order: PtFeNi/AC > PtFeNi/CB > PtFeNi/CB-o ≈ Pt/CNTs(FeNi).

Fig. 3 TEM images (left) and particle size distributions (right) of different carbon supported Fe and Ni promoted Pt catalysts after activation in H₂ at 500 °C for 2 h. (A) AC; (B) CB-o; (C) CB, and (D) CNTs.

3.3. Discussion

3.3.1. The relationship between catalyst particle size and catalytic performance. It has been well established that the textural properties and surface chemistry of carbon play vital roles in determining the catalyst particle size.^29,30,34 Large numbers of mesopores tend to increase the dispersion of the catalyst particles, while instead, microporous structures have a negative effect on the dispersion of the catalyst particles; rich surface functional groups not only make the size distributions of the catalyst particles bimodal, but also promote the formation of very fine metal particles. This is in good agreement with our previous study.³⁰ In most cases, the particle size is crucial to the catalytic performance.^29,34 The mean particle sizes and catalytic activities of the PtFeNi catalysts at 23 °C as a function of the different carbon supports are displayed in Fig. 4. It is clear that the PtFeNi/CB-o and Pt/CNTs(FeNi) catalysts possess similar mean particle sizes but present quite different CO conversions at 23 °C. On the contrary, the PtFeNi/CB catalyst has an apparently bigger mean particle size than Pt/CNTs(FeNi) but has comparable O₂ and CO conversions. This means the mean particle size is not the only factor to govern the catalytic performance and there should be other factors, other than the textural properties and surface chemistry of carbon, used to determine the catalytic performance.

Fig. 4 The mean particle sizes and catalytic activities of Fe and Ni promoted Pt catalysts at 23 °C as a function of the different carbon supports.

3.3.2. The graphitic carbon strengthened synergetic effect between Pt and FeNi. As discussed above, the particle size is not crucial to the catalytic performance within the investigated particle size range. Inspired by the multifunctional photocatalyst with spatially separate active components on a graphene sheet,^27,28 the authors hypothesise that the graphitic carbon may tune the interactions between Pt and FeNi, strengthening the synergetic effect and causing the extremely high catalytic performance. In order to confirm this bold assumption, the ability of carbon to capture and shuttle electrons through the π–π network (the graphitization degree and conductivity) and its effect on the catalytic properties have been investigated in detail below.

The graphitization degree of carbon materials can be reflected using XRD characterization, as shown in Fig. 5. The diffraction peaks at 25.78, 42.88, 53.58 and 78.18° are characteristic of graphitic carbon (JCPDS 656212). The strong and sharp diffraction peaks for the CNTs verify the high graphitization degree of the CNT sample, which is essentially composed of coaxial graphitic cylinders. CB and CB-o show similar but slightly broadened diffraction peaks with a slightly lower intensity compared to that of the CNTs. In comparison, the diffraction peaks of AC are broadened obviously (Fig. 5d), indicating an evidently poorly crystallized graphitic structure. Generally, the high graphitization degree means that there is a complete π–π network, indicating the free movement of delocalized π-electrons on the carbon surface, i.e. the high electrical conductivity of the carbon materials.³⁵ As far as CB-o is concerned, it displays similar diffraction features to CB; however, it bears more surface functional groups, suppressing the movement of π-electrons, resulting in a lower electrical conductivity. Therefore, the conductivity of the carbon materials was measured according to the resistivity measurement, as shown in Fig. 6. The resistivities of CNTs, CB, CB-o and AC are 0.04, 0.45, 1.35 and 141.80 Ω cm, respectively; which implies that the electron conductivity decreases in the order: CNTs > CB > CB-o > AC.

Fig. 5 XRD patterns of carbon supports (a) CNTs, (b) CB, (c) CB-o, (d) AC.

Fig. 6 The resistivities of the supports and the catalytic activities of the Fe and Ni promoted Pt catalysts at 23 °C as a function of the different carbon supports.

As mentioned in the Introduction section, carbon with a graphitic structure can capture and shuttle electrons to different sites through π–π networks. Furthermore, recent spectroscopic studies have proved the ability of graphene oxide to accept electrons from excited semiconductor nanoparticles and then reduce Ag⁺ at another site spatially different from that of the semiconductor nanoparticles.²⁵ Based on this observation, a Pt-loaded graphene-Sr₂Ta₂O_7−xN_x multifunctional photocatalyst with selective catalytic processes at separate sites has been successfully demonstrated.²⁷ Our previous research demonstrated that coordinatively unsaturated FeO_x and/or NiO_x confined in nanosized Pt matrices are active species for the PROX of CO and that there are interactions between Pt and the FeO_x/NiO_y species.^5–7,21 All investigations suggest that the ability of graphitic carbon to capture electrons and shuttle them to spatially different sites through π–π networks may tune the interactions between spatially separate Pt and FeO_x/NiO_y species, i.e. indirect interactions between Pt and the FeO_x/NiO_y species. This speculation means that the higher the conductivity of the carbon materials, the stronger the indirect interaction is. Furthermore, this indirect interaction between Pt and the FeO_x/NiO_y species may weaken the CO adsorption on Pt and facilitate oxygen adsorption, enhancing the CO oxidation activity. On the other hand, this indirect interaction may also enable optimum CO adsorption and O₂ adsorption on Pt and the FeO_x/NiO_y species, respectively; then the CO oxidation reaction can proceed through a chemical reaction–diffusion wave, as proposed in the previous publications.^23,36 There are also other possibilities, such as metal modified carbon catalyzed CO PROX reactions, extensively studied in electro-catalysis recently.^37–40 Either way will enhance the Pt utilization efficiency, leading to better or similar catalytic performances with relatively low Pt loadings. This concept is illustrated in Fig. 7 in detail.

Fig. 7 Schematic representation of the CO oxidation process on the PtFeNi catalysts tuned by the ability of graphitic carbon to capture and shuttle electrons from the noble metal to the FeO_x/NiO_y species through π–π networks.

In order to verify the above concept, the relationship between the conductivity and the catalytic performance was investigated, as shown in Fig. 6. It is interesting that the electron conductivity of the carbon materials agrees very well with the catalytic activities of the corresponding catalysts, which can be seen from the plots of the CO and O₂ conversions. The PtFeNi catalysts supported on the highly electronically conductive supports, such as CNTs, CB and CB-o, exhibit high CO conversions; however, when supported on the carbon materials with low electron conductivities, such as AC, they display low catalytic activities. Besides the high CO oxidation activity of the PtFeNi catalysts, the deactivation rate of the different carbon supported catalysts follows the order AC > CB-o > CB > CNTs. The results indicated that both the CO oxidation activity and stability agree well with the electronic conductivity of the carbon support. This implies that strong indirect interactions may lead to good catalytic performance. This observation confirms that the graphitic carbon strengthens the synergetic effect between Pt and FeNi, leading to high catalytic performance with low Pt loading levels.

It is well known that X-ray photoelectron spectroscopy (XPS) is a good way to characterize the interactions between metals and supports and/or metals and promoters. Therefore, XPS characterization was conducted. However, there is a very low signal for Fe 2p and Ni 2p due to their low loading level. Generally, different catalyst preparation methods will also influence the metal–support interactions; sometimes even changing the structure of the active phase. Thus, different catalyst preparation methods have been designed to adjust the graphitic carbon induced indirect interactions to investigate their effects on the catalytic performance.

Fig. 8 compares the catalytic performance of CB supported FeNi promoted Pt catalysts with different preparation methods. The FeNi(H500)Pt catalyst exhibits an apparently higher catalytic activity and stability compared with the FeNiPt and PtFeNi catalysts. Before being impregnated with the platinum precursor, FeNi/CB was pre-reduced with H₂ which will produce elemental iron/nickel and obtain stronger metal–carbon interactions simultaneously. When impregnated with the platinum precursor, the nucleation is prone to occur around the FeNi particles, which is helpful to create active species (it shortens the distance between the Pt and FeNi nanoparticles). Therefore, this preparation method contributes to stronger indirect interactions, resulting in better catalytic performance. This example verifies that the graphitic carbon induced indirect interactions between Pt and the FeNi species play an important role in the synergistic effects. And the catalytic performance difference between CNTs and CB supported PtFeNi may be due to the unique properties of CNTs.

Fig. 8 Variation in the O₂ conversion (A), CO conversion (B), and selectivity (C) at 23 °C over CB supported Fe and Ni promoted Pt catalysts as a function of the reaction time.

In order to further prove the important effects of the graphitic carbon induced indirect interactions on the synergistic effects, highly conductive flake graphite (FG) supported Pt and FeNi catalysts were physically mixed by grinding to prepare PtFeNi/FG catalysts and their catalytic properties for CO PROX were evaluated as shown in the ESI† (Fig. S3). This series of catalysts can effectively exclude the interfacial effects of Pt–FeNi and the alloying effects of PtFeNi on the performances; and the promotion effects should come from the graphitic carbon induced indirect interactions if they exist. It is very interesting that both the CO and O₂ conversions are enhanced sharply compared with those of the Pt/FG and FeNi/FG catalysts, as shown in Fig. S3.† This is a good example to confirm the significant roles of the graphitic carbon induced indirect interactions in the synergistic effects.

Thus, carbon materials with graphitic structures can capture electrons and shuttle them to spatially different sites, which induces indirect interactions between Pt and the FeNi species, consolidating the synergistic effects, leading to high catalytic activities of PtFeNi/CNTs catalysts with much lower Pt loadings. The roles of the graphitic carbon induced indirect interactions between Pt and the transition metal species on the thermal catalytic processes have rarely been recognized, although their roles in electrochemical activity and durability of the fuel cell catalytic layers are well understood. This discovery may pave the way to further reduce the loading level of the noble metal and search for more optimal and economical catalysts for the CO PROX reaction and other thermal catalytic processes.

4. Conclusions

CNTs, CB, CB-o and AC have been selected to load PtFeNi catalysts and their catalytic properties in the PROX of CO were evaluated. Through the comparative study, it is found that the ability of graphitic carbon to capture electrons and shuttle them to different sites induces indirect interactions between Pt and the FeNi species, strengthening the synergistic effect. This graphitic carbon amplified synergistic effect leads to extremely high catalytic activity with much lower Pt loadings, when operating with stoichiometric O₂, a much higher concentration of H₂ and nearly the same gas hourly space velocity. This discovery will help to design more effective, economical and robust catalysts for the CO PROX reaction and other thermal catalytic processes.

Acknowledgements

The authors gratefully thank Dr. Prof. Xinhe Bao and Dr. Prof. Xiulian Pan at Dalian Institute of Chemical Physics, Chinese Academy of Sciences, for their enthusiastic supervision and helpful discussions. This work is financially supported by the National Natural Science Foundation of China (21207039, 51108187), the Fundamental Research Funds for the Central Universities (Grant no. 2015zz052) and Guangdong Natural Science Foundation, China (Grant no. S2011010000737).

References

1. E. D. Park, D. Lee and H. C. Lee, Catal. Today, 2009, 139, 280.
2. K. Liu, A. Q. Wang and T. Zhang, ACS Catal., 2012, 2, 1165.
3. B. T. Qiao, A. Q. Wang, L. Li, Q. Q. Lin, H. S. Wei, J. Y. Liu and T. Zhang, ACS Catal., 2014, 4, 2113.
4. S. Huang, K. Hara, Y. Okubo, M. Yanagi, H. Nambu and A. Fukuoka, Appl. Catal., A, 2009, 365, 268.
5. Q. Fu, W. X. Li, Y. X. Yao, H. Y. Liu, H. Y. Su, D. Ma, X. K. Gu, L. M. Chen, Z. Wang, H. Zhang, B. Wang and X. H. Bao, Science, 2010, 328, 1141.
6. H. Xu, Q. Fu, Y. X. Yao and X. H. Bao, Energy Environ. Sci., 2012, 5, 6313.
7. X. G. Guo, Q. Fu, Y. X. Ning, M. M. Wei, M. R. Li, S. Zhang, Z. Jiang and X. H. Bao, J. Am. Chem. Soc., 2012, 134, 12350.
8. A. Tomita, K. Shimizu, K. Kato and Y. Tai, Catal. Commun., 2012, 17, 194.
9. H. Xu, Q. Fu, X. G. Guo and X. H. Bao, ChemCatChem, 2012, 4, 1645.
10. C. X. Wang, L. H. Zhang and Y. T. Liu, Appl. Catal., B, 2013, 136, 48.
11. J. Kugai, T. Moriya, S. Seino, T. Nakagawa, Y. Ohkubo, H. Nitani and T. A. Yamamoto, Int. J. Hydrogen Energy, 2013, 38, 4456.
12. H. R. Zheng, H. Y. Yang, R. R. Si, W. X. Dai, X. Chen, X. X. Wang, P. Liu and X. Z. Fu, Appl. Catal., B, 2011, 105, 243.
13. C. Wang, B. D. Li, H. Q. Lin and Y. Z. Yuan, J. Power Sources, 2012, 202, 200.
14. L. Yang, S. Y. Shan, R. Loukrakpam, V. Petkov, Y. R. Bridgid, N. Wanjala, M. H. Engelhard, J. Luo, J. Yin, Y. S. Chen and C. J. Zhong, J. Am. Chem. Soc., 2012, 134, 15048.
15. K. Tanaka, M. Shou and Y. Z. Yuan, J. Phys. Chem. C, 2010, 114, 16917.
16. H. W. Yang, C. Wang, B. D. Li, H. Q. Lin, K. Tanaka and Y. Z. Yuan, Appl. Catal., A, 2011, 402, 168.
17. S. H. Lu, C. Zhang and Y. Liu, Int. J. Hydrogen Energy, 2011, 36, 1939.
18. S. H. Lu and Y. Liu, Appl. Catal., B, 2012, 111–112, 492.
19. C. Zhang, W. Lv, Q. H. Yang and Y. Liu, Appl. Surf. Sci., 2012, 258, 7795.
20. H. Zhang, D. R. Lin, G. T. Xu, J. B. Zheng, N. W. Zhang, Y. H. Li and B. H. Chen, Int. J. Hydrogen Energy, 2015, 40, 1742.
21. L. M. Chen, D. Ma, Z. Zhang, Y. Y. Guo, D. Q. Ye and B. C. Huang, Catal. Lett., 2012, 142, 975.
22. T. Komatsu, M. Takasaki, K. Ozawa, S. Furukawa and A. Muramatsu, J. Phys. Chem. C, 2013, 117, 10483.
23. L. M. Chen, D. Ma, Z. Zhang, Y. Y. Guo, D. Q. Ye and B. C. Huang, ChemCatChem, 2012, 4, 1960.
24. Y. H. Kim, E. D. Park, H. C. Lee, D. Lee and K. H. Lee, Catal. Today, 2009, 146, 253–259 (Pt, Rh, Ru/Al₂O₃).
25. I. V. Lightcap, T. H. Kosel and P. V. Kamat, Nano Lett., 2010, 10, 577.
26. P. V. Kamat, J. Phys. Chem. Lett., 2012, 3, 663.
27. A. Mukherji, B. Seger, G. Q. Lu and L. Wang, ACS Nano, 2011, 5, 3483.
28. Q. J. Xiang, J. G. Yu and M. Jaroniec, J. Am. Chem. Soc., 2012, 134, 6575.
29. L. M. Chen, D. Ma and X. H. Bao, J. Phys. Chem. C, 2007, 111, 2229.
30. L. M. Chen, D. Ma, X. Y. Li and X. H. Bao, Catal. Lett., 2006, 111, 133.
31. B. W. Zhong, H. Y. Liu and X. M. Gu, ChemCatChem, 2014, 6, 1553.
32. S. Kundu, Y. M. Wang, W. Xia and M. Muhler, J. Phys. Chem. C, 2008, 112, 16869.
33. Z. L. Fan, W. Chen, X. L. Pan and X. H. Bao, Catal. Today, 2009, 147, 86.
34. L. Perini, C. Durante and M. Favaro, Appl. Catal., B, 2014, 144, 300.
35. D. Sebastian, I. Suelves, R. Moliner and M. J. Lazaro, Carbon, 2010, 48, 4421.
36. W. X. Huang, X. H. Bao, H. H. Rotermund and G. Ertl, J. Phys. Chem. B, 2002, 106, 5645.
37. G. K. Ramesha, J. F. Brennecke and P. V. Kamat, ACS Catal., 2014, 4, 3249.
38. J. Deng, P. J. Ren, D. H. Deng, L. Yu, F. Yang and X. H. Bao, Energy Environ. Sci., 2014, 7, 1919.
39. D. H. Deng, L. Yu. X. Chen, G. X. Wang, L. Jin, X. L. Pan, J. Deng, G. Q. Sun and X. H. Bao, Angew. Chem., Int. Ed., 2013, 52, 371.
40. W. J. Zhou, J. Zhou, Y. C. Zhou, J. Lu, K. Zhou, L. J. Yang, Z. H. Tang, L. G. Li and S. W. Chen, Chem. Mater., 2015, 27, 2026.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5cy01091h

---