Enhanced ammonia-borane decomposition by synergistic catalysis using CoPd nanoparticles supported on titano-silicates

Abstract

Pd and Co_xPd_1−x nanoparticles (NPs), synthesized using the reduction by solvent method, were loaded on SiO₂ and Ti–SiO₂ supports. The resulting catalysts were tested in the ammonia-borane decomposition reaction under dark and UV-vis conditions. The synergistic promotion by Co (in the NPs) and Ti (in the support), combined with the UV-vis light irradiation, enhanced the catalytic activity showing very promising TOFs values for this kind of catalysis, from 1.53 to 49.5 mol H₂ per mol Pd per min.

---

The decomposition of small molecules, such as formic acid or ammonia-borane (AB), is one of the most promising alternatives for in situ H₂ generation, which could eventually lead to implementation of H₂-based technology.^1–3 AB is claimed as one of the inorganic compounds with the highest hydrogen content (19.6 wt%). This, added to its high reactivity with noble metals, such as Ru, Pd or Pt, makes it one of the best candidates for its use as H₂ feedstock in a PEMFC.^3–8 The total decomposition of this compound in liquid phase using water as solvent produces 3 mol of H₂ per mol of AB, according to the following reaction equation.⁹

NH₃BH₃ + 2H₂O → NH_4(aq.)⁺ + BO_2(aq.)⁻ + 3H_2(g)

Consequently, ammonia borane is a liquid-phase chemical hydrogen storage material of great current interest. Its decomposition by a wide range of catalysts has been extensively reported in literature. Among these, noble metals, such as Rh, Ir, Ru, and Pt, have shown interesting catalytic activities,¹⁰ but they are unsuitable for widespread practical applications due to their availability and price. Previous studies have demonstrated that bimetallic nanoparticles combining a noble metal and a first-row transition metal to form an alloy structure could be promising candidates for the design of catalysts for the hydrolysis of ammonia borane.¹¹ In order to assess the beneficial effect of these noble metal/first-row transition metal combinations for this application, we have chosen to assess Pd, due to its moderate activity among noble metals¹² and Co because it shows the highest activity among non-noble metal catalysts.^11,13,14

Enhancement of the catalytic activity of the supported noble metal can be attained by different routes; (i) increasing the metal dispersion on the support by the synthesis of very small NPs,¹⁵ (ii) alloying the noble metal with a transition metal, (which is also attractive due to the cost reduction of the catalysts)^11,16,17 and (iii) supporting the NPs on a UV-vis active support to upgrade the electron-transfer from the support to the NPs.^1,18 In the present study, the synthesis of Pd and Co–Pd NPs by a reduction-by-solvent methodology has been applied with successful results in the production of H₂ by AB decomposition. The synthesis conditions allow a perfect control over size and morphology of the NPs (both mono- and bimetallic). Pure and alloyed cobalt (as oxide or reduced form) have been addressed in the recent literature as promising catalysts for AB decomposition due to their low cost compared with noble metals and their high activity in relevant reactions.^{12,16,19–21} Due to the difference in reduction potential as compared with Pd alone, when M–Pd catalyst (i.e. Cu or Co) exhibits charge transfer from the M to Pd that increases the electron density on the Pd atom and, consequently, its activity in the AB decomposition is enhanced.^1,18 On the other hand, the use of an active support under UV-vis conditions, such as titania or a titano-silicate (Ti–SiO₂), is another available strategy to enhance the catalytic activity of the system.^22,23

In the present work, Pd and Co_xPd_1−x NPs were prepared by the well-established procedure of reduction by solvent method using PVP as a capping agent and following an already reported procedure.^15,24 After their purification, the NPs were loaded on SiO₂ and Ti–SiO₂ (UV-vis inactive and active supports, respectively) to study their catalytic activity in AB decomposition in liquid-phase reactions (see all the experimental details in the ESI†).

From the textural characterization of the catalysts (see ESI†), it must be mentioned that there were no significant changes in the isotherm shape and specific surface properties between the two raw supports (SiO₂ and Ti–SiO₂) and the resulting catalysts after NPs impregnation. Regarding the NPs preparation and deposition on the supports, two main characteristics must be highlighted: both the nominal Co/Pd ratio and the nominal metal loading (1 wt%) were approximately achieved in all the catalysts (calculated from ICP-OES results). The NPs size was very similar for the two series of catalysts (SiO₂ and Ti–SiO₂-supported) upon deposition on the supports (3.4–3.5 nm, see TEM micrographs and histograms in Fig. 1). Therefore, any change in the catalytic activity may be ascribed to the chemical composition of the supports or modifications in the bimetallic NPs' size with respect to the monometallic counterpart. From CO adsorption experiments after H₂ pre-reduction of the catalysts, it is possible to obtain information about the Co and Pd distribution in the bimetallic NPs. To this end, the adsorbed CO moles were normalized per mol of Pd and per total mol of metal (Co and Pd) (see ESI†). The Pd-normalized results showed an approximately constant CO adsorption of about 0.23 mol CO/mol Pd, while the total metal results indicated that the adsorbed CO decreased proportionally as the Co content in the NPs increased. These results, together with the surface Co/Pd ratio as determined by XPS analysis, suggested a homogeneous distribution of both elements in the alloyed NP structure. In order to study the UV-vis light response, solid phase UV-vis analysis was mandatory (see ESI†). SiO₂ did not present any absorption. A small fraction of octahedral Ti(IV) was observed at 300 nm in the raw Ti–SiO₂, but the main absorption of the support was displayed in the 200 nm range corresponding to tetrahedral Ti(IV) dispersed in the SiO₂ framework.^25,26 After loading NPs on the supports, a broad absorption at around 300 nm was observed even in the SiO₂-supported catalysts due to absorption by the Pd NPs.²⁷

Fig. 1 TEM images of Co_xPd_1−x NPs deposited on Ti–SiO₂ and their corresponding histograms and average particle size.

Pd, Co and Ti XPS analysis offered information about the electronic state of these elements in order to determine the influence of the Co in the NPs and the Ti in the support. Binding Energy (BE) values, as well as the most relevant peaks of their corresponding XPS spectra, are presented in Table S1.† The signals corresponding to Pd(0) and Pd^δ+ could be clearly observed for all the catalysts. The appearance of the latter signal may be due to the electron deficient Pd directly interacting with PVP. Cobalt species, such as electron deficient Co or CoO_x, and electron deficient Ti(IV) due to its dispersion into the SiO₂ support are also observed in the Co_xPd_1−x/Ti–SiO₂ catalysts. In this respect, it must be noted that evidence of the presence of Co(0) (778.0 eV) was not found in our analysis.^26,28

For the SiO₂-supported catalysts, a slight decrease in the BE for Pd(0) and Pd^δ+ as measured by XPS was observed when Co was incorporated into the NPs due to Pd electron density enrichment. An increase in Pd(0) content due to the addition of Co was also observed, from 73.7% for Pd/SiO₂ to 86.0% and 83.5% for Co_0.25Pd_0.75/SiO₂ and Co_0.5Pd_0.5/SiO₂, respectively. However, when the Co content increased to Co_0.75Pd_0.25/SiO₂, the Pd(0) content decreased to 69.6%. This might be due to the presence of different Co species in the NPs formed at this ratio. Thus, it seems that for low Co contents (i.e. Co_0.25Pd_0.75 and Co_0.5Pd_0.5) there is an efficient charge transfer from the Co species to Pd, which is supported by the absence of metallic Co signals in the XPS spectra; however, when the Co ratio increases, more oxidized Co species are formed and the charge transfer is hindered. Furthermore the PVP–metal interaction and its effect on the final electronic features of the resulting NPs should also be considered. Along this line, the high surface PVP/Pd ratio in the Co_0.75Pd_0.25 NPs together with the strong PVP–Pd interaction and the electron-withdrawing property of PVP through the C=O groups,^29,30 might also be responsible for the higher Pd^δ+ content detected in Co_0.75Pd_0.25/SiO₂.

On the other hand, the Pd(0) signal was shifted to a higher BE value by 0.3 eV when the Pd NPs were deposited on the Ti–SiO₂ support and its Pd^δ+ content was reduced. This is in good agreement with the Ti 2p(3/2) BE displacement from 460.48 eV in the raw support to 459.38 eV after the NPs loading.^26,31 This observation confirms a charge transfer from the Pd NPs to the Ti-based support. When Ti–SiO₂ was used as a support for the Co_xPd_1−x NPs, there was also a decrease in the BE at which the peaks appeared, as well as an increase in the Pd(0) content (higher in value than the SiO₂-catalysts, although for high Co loadings this trend is no longer observed, vide supra). The Ti(IV) BE is also significantly reduced, to 458.93 eV when the Co_0.25Pd_0.75 NPs were deposited. For this catalyst series, the same trend was observed in Pd(0) content when the Co ratio increased in the NPs.

The catalysts were studied in the AB decomposition reaction, using a metal/AB molar ratio of 0.02 and analyzing the production of H₂ every 2.5 min (see full details of the procedure in the ESI†). The H₂ evolution profiles indicated that no induction period was necessary, but important differences in the catalytic activities were observed. For the least active catalyst total AB conversion was achieved after approximately 25 minutes of reaction while for the most active samples less than 10 minutes were necessary (results of the AB conversions in terms of n(H₂)/n(AB) ratios are shown in the ESI†). To compare the activity of all the prepared catalysts, the TOF values at 2.5 min under dark and UV-vis light conditions were calculated with respect to Pd and total metal content; these are presented in Fig. 2. It must be noted that when the supports were used without any nanoparticles impregnated on their surface, no or negligible activity in the AB decomposition reaction was observed.

Fig. 2 TOFs values based on Pd and M (total metal) vs. Co molar ratio for AB decomposition under dark and UV-vis light irradiation conditions. (A and B) Results presented for the catalysts prepared using SiO₂ as support, (C and D) when Ti–SiO₂ is used as support.

As the Co ratio increased in the NPs supported on SiO₂ (Fig. 2A) the specific activity of the Pd increased proportionally, from the initial value of 1.5 to 34.4 mol H₂ per mol Pd per min, as measured under dark conditions. This behavior was in good agreement with our XPS results and other reported works, where the electron enrichment of the noble metal (as Pd) from a transition metal (such as Ni or Cu) was studied.^24,32 However, when the catalysts supported on SiO₂ were tested under UV-vis irradiation a small decrease of the activity was observed. This activity loss can be assigned to the partial degradation of PVP and consequent blocking of Pd sites on the NPs surface under UV-vis light irradiation.³³ For the same catalytic tests, but considering Co content for the TOF calculation (Fig. 2B), there was a maximum in the activity of 13.9 mol H₂ per mol Co per min for the catalyst containing Co_0.5Pd_0.5 NPs.

Regarding the catalytic behavior of Ti–SiO₂-based catalysts (Fig. 2C and D), a noticeable enhancement in the catalytic activity with respect to the SiO₂ samples was obtained for all catalysts under both conditions (dark and UV-vis). Even for the pure Pd based catalysts under dark conditions, there is a considerable increase from 1.5 to 5.5 mol H₂ per mol Pd per min in the TOF value when Ti is incorporated to the support. In this sense, the Ti-based silicate catalysts might present strong interactions between TiO₂-based supports and AB which favors its decomposition, as previously reported.^34,35 Comparing the Pd-normalized results, the TOF drastically increased from 5.5 mol H₂ per mol Pd per min for Pd/Ti–SiO₂ to ∼40 mol H₂ per mol Pd per min for all Co-containing species. Additionally, the Pd-normalized activity of the samples increased up to 49.5 mol H₂ per mol Pd per min when the catalysts were tested under UV-vis light irradiation. On the other hand, the highest value for the total metal normalized TOFs was obtained for Co_0.25Pd_0.75/Ti–SiO₂, with values of 31.7 and 39.7 mol H₂ per mol M per min under dark and UV-vis light conditions, respectively. Upon consideration of these catalytic tendencies, it seems that the activity of the Co_xPd_1−x/Ti–SiO₂ catalysts under UV-vis light irradiation per Pd content is the same regardless of the Co content, (49.5 mol H₂ per mol Pd per min) corroborating the synergistic effect of Co when alloyed in the NPs and, especially, the Ti incorporated into the support. The role of Ti in the photocatalytic enhancement of AB decomposition via accumulation of electrons and holes in the metal deposited on its surface and their transfer to the AB molecules has recently been reported.³⁶

Fig. 3 shows the H₂ production in μmol after 2.5 minutes of reaction with the Ti–SiO₂ based catalysts under dark and UV-vis light conditions. It can be observed that there was a general positive effect of UV-vis light irradiation on the H₂ production for all samples under study. However, the outstanding enhancement displayed by the Co_0.25Pd_0.75/Ti–SiO₂ catalyst should be noted. This paramount catalytic behavior confirms the suitability of the present catalytic system and makes it a promising candidate for its possible implementation in H₂-fed devices.

Fig. 3 H₂ production (μmol) of the Ti-based catalysts after 2.5 minutes of reaction under dark and UV-vis irradiation conditions.

The results in TOF and AB conversion to H₂ production for the Co_0.25Pd_0.75/Ti–SiO₂ under UV-vis light (our system) are low in comparison with highly complex silica-coated cobalt ferrite loaded with Pd NPs,²⁰ but higher than related works that also use Co–Pd based catalysts and similar metal/AB ratios^11,37,38 loaded onto high performance carbon materials with significantly higher metal loadings.

Conclusions

In summary, highly active catalysts with less than 1 wt% of noble metal content for AB decomposition were synthesized, taking advantage of a two-fold strategy for promotion of the initial Pd/SiO₂ catalytic system; increasing the electron density of Pd by alloying with Co, and deposition of these alloyed NPs on a very simple and UV-vis-active support doped with Ti. As a result, a very promising TOF (49.5 mol H₂ per mol Pd per min) and a very fast H₂ delivery of more than 160 μmol of H₂ in less than 2.5 min has been obtained using a sample containing only 0.8 wt% of Pd. These promising catalysts benefit from the catalytic synergy between Pd and Co in the NPs that results from electron density transfer from Co to Pd (as observed by XPS) and the incorporation of Ti into the support, making it active to the UV-vis radiation.

Acknowledgements

We thank the Spanish Ministry of Economy and Competitiveness (MINECO), Generalitat Valenciana and FEDER (Projects CTQ2015-66080-R MINECO/FEDER and PROMETEOII/2014/010) for financial support. J. G. A. thanks the MINECO for his fellowship (BES-2013-063678), with special thanks for the mobility grant of MINECO (EEBB-I-15-10219) at Osaka University.

Notes and references

1. H. Cheng, T. Kamegawa, K. Mori and H. Yamashita, Angew. Chem., Int. Ed. Engl., 2014, 53, 2910–2914.
2. K. Mori, M. Dojo and H. Yamashita, ACS Catal., 2013, 3, 1114–1119.
3. K. Mori, K. Miyawaki and H. Yamashita, ACS Catal., 2016, 6, 3128–3135.
4. S. Enthaler, J. von Langermann and T. Schmidt, Energy Environ. Sci., 2010, 3, 1207–1217.
5. C. W. Hamilton, R. T. Baker, A. Staubitz and I. Manners, Chem. Soc. Rev., 2009, 38, 279–293.
6. F. H. Stephens, V. Pons and R. Tom Baker, Dalton Trans., 2007, 2, 2613–2626.
7. Z.-L. Wang, J.-M. Yan, Y. Ping, H. L. Wang, W.-T. Zheng and Q. Jiang, Angew. Chem., Int. Ed., 2013, 52, 4406–4409.
8. U. Eberle, M. Felderhoff and F. Schüth, Angew. Chem., Int. Ed., 2009, 48, 6608–6630.
9. T. Kamegawa and T. Nakaue, Chem. Commun., 2015, 51, 16802–16805.
10. M. Yadav and Q. Xu, Energy Environ. Sci., 2012, 5, 9698.
11. D. Sun, V. Mazumder, Ö. Metin and S. Sun, ACS Nano, 2011, 5, 6458–6464.
12. Ö. Metin, Ş. Şahin and S. Özkar, Int. J. Hydrogen Energy, 2009, 34, 6304–6313.
13. J.-M. Yan, X.-B. Zhang, H. Shioyama and Q. Xu, J. Power Sources, 2010, 195, 1091–1094.
14. Ö. Metin and S. Özkar, Int. J. Hydrogen Energy, 2011, 36, 1424–1432.
15. S. Domínguez-Domínguez, Á. Berenguer-Murcia, D. Cazorla-Amorós and Á. Linares-Solano, J. Catal., 2006, 243, 74–81.
16. J.-M. Yan, X.-B. Zhang, T. Akita, M. Haruta and Q. Xu, J. Am. Chem. Soc., 2010, 132, 5326–5327.
17. D. Sun, V. Mazumder, Ö. Metin and S. Sun, ACS Catal., 2012, 2, 1290–1295.
18. Y. Kuwahara, K. Nishizawa, T. Nakajima, T. Kamegawa, K. Mori and H. Yamashita, J. Am. Chem. Soc., 2011, 133, 12462–12465.
19. J. Hu, Z. Chen, M. Li, X. Zhou and H. Lu, ACS Appl. Mater. Interfaces, 2014, 6, 13191–13200.
20. S. Akbayrak, M. Kaya, M. Volkan and S. Özkar, Appl. Catal., B, 2014, 147, 387–393.
21. A. Rossin and M. Peruzzini, Chem. Rev., 2016, 116, 8848–8872.
22. F. X. Llabrés i Xamena, P. Calza, C. Lamberti, C. Prestipino, A. Damin, S. Bordiga, E. Pelizzetti and A. Zecchina, J. Am. Chem. Soc., 2003, 125, 2264–2271.
23. Y. Shiraishi, N. Saito and T. Hirai, J. Am. Chem. Soc., 2005, 127, 8304–8306.
24. J. García-Aguilar, I. Miguel-García, Á. Berenguer-Murcia and D. Cazorla-Amorós, Carbon, 2014, 66, 599–611.
25. A. Prieto, M. Palomino, U. Díaz and A. Corma, Catal. Today, 2014, 227, 87–95.
26. A. K. Sinha, S. Seelan, S. Tsubota and M. Haruta, Angew. Chem., Int. Ed., 2004, 43, 1546–1548.
27. M. Navlani-García, K. Mori, M. Wen, Y. Kuwahara and H. Yamashita, Bull. Chem. Soc. Jpn., 2015, 88, 1500–1502.
28. A. Bulut, M. Yurderi, İ. E. Ertas, M. Celebi, M. Kaya and M. Zahmakiran, Appl. Catal., B, 2016, 180, 121–129.
29. Y. Borodko, S. M. Humphrey, T. D. Tilley, H. Frei and G. A. Somorjai, J. Phys. Chem. C, 2007, 111, 6288–6295.
30. I. Miguel-García, Á. Berenguer-Murcia and D. Cazorla-Amorós, Appl. Catal., B, 2010, 98, 161–170.
31. M. C. Capel-Sanchez, G. Blanco-Brieva, J. M. Campos-Martin, M. P. de Frutos, W. Wen, J. A. Rodriguez and J. L. G. Fierro, Langmuir, 2009, 25, 7148–7155.
32. K. Mori, H. Tanaka, M. Dojo, K. Yoshizawa and H. Yamashita, Chem.–Eur. J., 2015, 21, 12085–12092.
33. C. Aliaga, J. Y. Park, Y. Yamada, H. S. Lee, C. Tsung, P. Yang and G. a. Somorjai, J. Phys. Chem. C, 2009, 113, 6150–6155.
34. O. V. Komova, V. I. Simagina, N. L. Kayl, G. V. Odegova, O. V. Netskina, Y. A. Chesalov and A. M. Ozerova, Int. J. Hydrogen Energy, 2013, 38, 6442–6449.
35. R. Fernandes, N. Patel, A. Miotello, R. Jaiswal and D. C. Kothari, Int. J. Hydrogen Energy, 2012, 37, 2397–2406.
36. A. Yousef, N. A. M. Barakat and H. Y. Kim, Appl. Catal., A, 2013, 467, 98–106.
37. J. Wang, Y.-L. Qin, X. Liu and X.-B. Zhang, J. Mater. Chem., 2012, 22, 12468.
38. B. Kılıç, S. Şencanlı and Ö. Metin, J. Mol. Catal. A: Chem., 2012, 361–362, 104–110.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra21302b

---