Single component gold on protonated titanate nanotubes for surface-charge-mediated, additive-free dehydrogenation of formic acid into hydrogen

Abstract

Surface charge redistribution between gold nanoparticles and protonated titanate nanotubes leads to high catalytic activity of single component gold for additive-free dehydrogenation of formic acid into hydrogen.

---

Hydrogen (H₂) is a promising energy carrier due to its superb conversion efficiency, environmental friendliness, and high energy density. Controlled storage and release of H₂ are among the critical technical obstacles faced by the hydrogen economy. As one of the major products from biomass processing, formic acid (FA) represents a superior hydrogen carrier in fuel cells designed for portable use, since it has an energy content at least five times higher than that of commercially available lithium ion batteries.^1–6 In general, FA can be catalytically decomposed to H₂ and carbon dioxide (CO₂) via a favorable dehydrogenation reaction (HCOOH → H₂ + CO₂; ΔG_298K = −35.0 kJ mol⁻¹). Another undesirable dehydration pathway that produces carbon monoxide (HCOOH → H₂O + CO; ΔG_298K = −14.9 kJ mol⁻¹) can also be involved.

Supported palladium nanoparticles (PdNPs) have been frequently used as catalysts for selective dehydrogenation of FA due to its unique electronic interactions with FA molecules. To improve the overall catalytic efficiency of bimetallic and trimetallic Pd nanocatalysts, gold (Au) is considered to be a superior co-catalyst.^3,7–10 Unfortunately, previously described catalysts based on single component Au for H₂ production from FA were only sporadically reported.¹¹ One possible reason is its much higher energy barrier than Pd for the dehydrogenation of FA in the absence of hydrous additives. It should be noted that the combination of additives (e.g., triethylamine and sodium formate) and FA can dramatically enhance the H₂ generation efficiency,^11–13 while the additive itself can also be catalyzed into H₂. This fact not only brings difficulty in figuring out the real contribution of FA for H₂ production, but also causes problems in determination of the exact dehydrogenation reaction mechanism. Therefore, it is desirable to fabricate efficient single component gold catalysts for catalytic dehydrogenation of additive-free FA solution.

Recent studies have suggested that particle charge state influences the chemical properties, because the binding of molecules to specific sites of a metal, which determines the reactivity of metal/oxide systems, depends on the surface charge changes.¹⁴ In particular, electron-deficient metal is expected to alter the adsorption configuration of molecular adsorbents and subsequently facilitate their bond cleavage, thereby improving the catalytic performance of the system.^15,16 Zhang and co-workers have reported that Pt or Pd single atom catalysts interact strongly with their supports, resulting in the electron flow from metal to support. The positively charged metal catalysts are superior to several hydrocarbon reactions.¹⁶ Although FA dehydrogenation is sensitive to the surface charge state of metal surface, effective approaches for tuning the electron density of Au at the nanoscale to achieve efficient H₂ production from FA are rare.

Herein, we employ protonated titanate nanotubes (TiNTs) as the support for AuNPs and apply this composite catalyst to additive-free FA dehydrogenation at relatively low temperatures. TiNTs consist of a multiwalled scroll-type open-ended structure, which have unique electronic interaction with transitional metals.¹⁷ In contrast with other catalysts, the one calcined at 300 °C represents the optimal catalytic hydrogen production efficiency. It is found that the charge redistribution between Au and TiNTs is the main factor to the enhancement of the dehydrogenation reaction, because proper temperature treatment of TiNTs changes the surface charge state of supported single component AuNPs; the resulting electron-deficient, rather than electron-rich AuNPs are particularly active for FA dehydrogenation into H₂.

Protonated TiNTs were prepared by a hydrothermal treatment of anatase titanium oxide particles in concentrated NaOH solution (see ESI for detailed information†).¹⁸ AuNPs with uniform size of 3.3 ± 0.5 nm were synthesized according to our previously method.¹⁹ They were readily adsorbed onto TiNTs by colloid deposition, followed by calcination at different temperatures. The resulting catalyst is termed as Au/TiNTs-x, where x indicates the calcination temperature (°C). The XRD patterns for the protonated TiNTs-300 before and after Au loading are illustrated in Fig. 1a. TiNTs have diffraction peaks positioned at 2θ = 10.7°, 24.7°, 28°, and 48°, due to layered protonated titanates, H₂Ti₂O₅·H₂O. The d-spacing was calculated to be 0.68 nm from the 2θ value of 10.7° using the Bragg equation, which is consistent with the interlayer spacing observed in the following TEM analysis. These results suggest that the protonated TiNTs are formed from layered titanates.¹⁷ Au/TiNTs-300 have nearly identical but weaker diffraction peaks than that of the initial TiNTs, indicating that they have similar crystal structures but the periodic layer structure of TiNTs is interrupted to a certain extent by gold loading. In addition, because of the low loading amount and high dispersity, the Au signal cannot be discerned clearly in the XRD pattern, and only small peaks (e.g., Au (200) peak) can be identified. The low-resolution TEM image (Fig. 1b) of protonated TiNTs shows the presence of a high proportion of open-ended nanotubes with diameters of 4 to 7 nm and lengths of several hundred nanometres. High resolution TEM (HRTEM) observation (Fig. 1c) reveals that the walls of the protonated TiNTs consist of two types of lattice fringes. The intervals among the fringes parallel and perpendicular to the tube axis are estimated to be ca. 0.668 and 0.377 nm, respectively. The former and latter correspond to the interlayer space in the layered titanates and the distance between two adjacent O atoms in the topmost corner positions of the TiO₆ edge-sharing octahedra that form a zigzag structure. Fig. 1d shows that AuNPs with a mean particle diameter of ∼4.9 nm are evenly dispersed on the surface of TiNTs, and the morphology of TiNTs underwent limited change after gold loading. The HRTEM image (inset) exhibits lattice fringes with an interplanar spacing of 0.238 nm, corresponding to (111) planes of fcc Au. In addition, the specific area of three selected catalysts, Au/TiNTs-100, Au/TiNTs-300 and Au/TiNTs-500 is determined to be 320, 295, and 210 m² g⁻¹, respectively based on nitrogen sorption analysis. It is seen that the surface area of Au/TiNTs-x is gradually reduced upon higher temperature treatment, but still in the same order of magnitude.

Fig. 1 (a) XRD patterns of TiNTs and AuNPs/TiNTs, (b) TEM image of TiNTs, (c) HRTEM image of TiNTs, (d) TEM image of AuNPs/TiNTs (inset shows the HRTEM image of a single AuNP).

We performed X-ray photoelectron spectroscopy (XPS) spectra to examine the surface charge state of Au species on TiNTs. Here, Au/TiNTs-100, Au/TiNTs-300 and Au/TiNTs-500 samples are selected to determine the high-resolution Au4f XPS signals. As compared to Au/TiNTs-100, a red-shift from 87.2 to 87.4 eV for Au4f_5/2 signals, and 83.5 to 83.7 eV for Au4f_7/2 signals are observed for Au/TiNTs-300 (Fig. 2a), respectively. The ∼0.2 eV red-shifts indicate that positive surface charges are accumulated on Au surface in the Au/TiNTs-300 sample. However, after calcination at 500 °C, the Au4f_5/2 and Au4f_7/2 signals move respectively to 86.9 and 83.1 eV, suggesting that the Au surface becomes negatively charged. The forward and backward electron flow between Au and support is most likely caused by the surface property modification of TiNTs. Thus we checked the XPS Ti2p and O1s spectra of Au/TiNTs with different temperature treatments. As shown in Fig. 2b, the Ti2p binding energies of Au/TiNTs-300 as well as Au/TiNTs-100 locate at 458.8 eV, while it blue-shifts to 458.6 eV for Au/TiNTs-400 and further shifts to 458.3 eV for Au/TiNTs-500. O1s spectra also show the similar trend (Fig. 2c): blue-shifts of 0.2 and 0.5 eV are observed for Au/TiNTs-400 and Au/TiNTs-500, respectively, as compared with other samples. It can be concluded that the electronic structure of Au surface is determined by TiNTs supports with different surface properties.

Fig. 2 High resolution XPS spectra of (a) Au4f, (b) Ti2p, (c) O1s signals, and (d) solid state EPR spectra for the Au/TiNTs samples.

The electron paramagnetic resonance (EPR) spectra of Au/TiNTs representing the surface state of TiNTs are illustrated in Fig. 2d. No obvious signals are found for the samples treated at low temperatures (e.g., Au/TiNTs-100 and Au/TiNTs-200), while an observable signal with a principal g value of 2.006 emerges as the treatment temperature rises to 300 °C (inset). The g value is most likely originated from the surface Ti3p centers that are formed when an electron is trapped at a Ti site. It is similar to those reported for Ti3p in an anatase phase, indicating that calcination of the nanotubes leads to anatase formation. Notably, the specific signal becomes dominated for Au/TiNTs-500, which is in accompany with broad linewidths due to the adsorbed O₂⁻ sites.²⁰ It has been reported that the supported AuNPs accommodate excess electrons that originate from a charge transfer from a local interaction with electron-rich oxide defects that act as Au nucleation centers.^21,22 Consequently, the generation of subsurface Ti³⁺ center favors the electron flow from oxygen vacancy to adjacent metal species, leading to the formation metal centers with negative charges. However, when the concentration of subsurface Ti³⁺ centers are just in a few quantities, such as in the case of Au/TiNTs-300, the electrons tend to transfer from Au to TiNTs due to the effect of surface charge equilibrium, resulting in the formation positive charges on Au (i.e., Au_δ⁺ species). The different electron density in the Au aggregates is supposed to affect both the spatial distribution and the vibrational properties of adsorbed FA species, thus changing the catalytic efficiency, as shown later.

The surface charge state of Au greatly influences the catalytic activity of Au/TiNTs for additive-free FA (1 M) dehydrogenation at 60 °C, giving rise to a volcano curve correlation. In detail, as shown in Fig. 3a, no H₂ is produced when Au/TiNTs-100 is used as the catalyst, while the catalytic performance dramatically increases to 0.08 mol h⁻¹ g⁻¹ for the Au/TiNTs-200 based on the total amount of Au, and reaches an optimal performance at 0.1 mol h⁻¹ g⁻¹ for Au/TiNTs-300. As the calcination temperature further increases to 400 °C or more, a rapid decrease of the H₂ generation rate is observed. The surface-charge-mediated reaction of Au/TiNTs can also be reflected from its significant solvent effect. Fig. 3b shows that FA can be effectively decomposed into H₂ when water or ketone (e.g., acetone, cyclohexanone or acetophenone) is used as the solvent, while limited H₂ is produced when other types of solvents, such as methanol, ethanol, acetonitrile or ethyl acetate are substituted. This solvent effect is most likely derived from the surface-charge-mediated dehydrogenation nature, since water and ketone have bare lone pair electrons that may interact strongly with the surface positive Au_δ⁺ species, contributing to the cleavage of C–H bond in FA. In order to further understand the role of water, we performed deuterated isotopic composition studies. As shown in Table S1 (ESI),† when DCOOD was substitute for HCOOH, a primary deuterium kinetic isotope effect (KIE) of ∼6.37 was observed, indicating that the rate-determining step involves cleavage of C–H bond. The H₂ production rate for H₂O was found to be as large as ∼3.93 times faster than that for D₂O within 3 h reaction, revealing that water plays an important role to the success of the reaction.

Fig. 3 Catalytic FA dehydrogenation performance of (a) Au/TiNTs with different temperature treatment, (b) Au/TiNTs-300 in different ketone solvents, (c) Au on different supports, and (d) Au/TiNTs-300 in different concentration of FA; reaction condition: catalyst, 20 mg; gold loading, 1.0 wt%; reaction temperature, 60 °C; FA concentration for (a)–(c), 1 M.

The surface-charge-mediated reaction depends on the nature of supports. First, the reaction did not proceed at all with Au-free TiNTs, showing that the presence of Au was essential for achieving high activity in the dehydrogenation of FA. Second, as shown in Fig. 3c, on addition of Au/TiNTs-300 into the FA solution, constant H₂ gas was immediately generated without the production of CO, showing its high selectivity for FA dehydrogenation. Next, we evaluated AuNPs supported on TiO₂ NPs, and the resultant catalyst was significantly less active for FA decomposition under similar reaction conditions (Fig. 3c). Meanwhile, Au supported on silica (Au/SiO₂), activated carbon (Au/C), and zirconium dioxide (Au/ZrO₂) are all ineffective for the desired reaction. Although it has been shown that subnanometric gold on ZrO₂ is efficient for FA decomposition into H₂ in the presence of amine adducts (e.g., NEt₃),¹¹ it is much less active without additives. Thus, the involvement of positively charged Au sites is postulated on the basis of the absence of reaction on the Au/SiO₂, Au/C and Au/ZrO₂ catalysts. This observation implies single electron-deficient Au sites as being active in the dehydrogenation reaction. Furthermore, a comparison with Pd, Pt, and Ag NPs supported on TiNTs as reference catalysts shows that Au is uniquely active for selective FA decomposition (Fig. S1, ESI†). For examples, Pt/TiNTs and Pd/TiNTs only show activity within the initial reaction period while gradually lose function with prolonged reaction, because the dehydration of FA occurs on these catalysts, producing detrimental CO molecules. FT-IR spectra demonstrate that the surface of recycled Pd/TiNTs catalyst exhibits a prominent peak at 1795 cm⁻¹ wavenumber corresponding to the bridging CO species, while no observable peak is found within this range for recycled Au/TiNTs catalyst (Fig. S2, ESI†).

Another positive feature of the Au/TiNTs-300 catalyst is its high tolerance toward FA concentration. As shown in Fig. 3d, the H₂ evolution rate increases significantly with increasing FA concentration to a maximum at 10 M, and the optimal turnover frequency (TOF) is measured to be 100.3 h⁻¹, corresponding to a H₂ generation rate of 0.51 mol h⁻¹ g⁻¹. However, even at a concentration of 20 M, which is close to pure FA (26.6 M), Au/TiNTs-300 still shows substantial catalytic activity towards FA decomposition with TOF = 22.3 h⁻¹. In addition, Au/TiNTs-300 catalyzed FA dehydrogenation process is sensitive to the reaction temperatures (Fig. S3, ESI†), and the apparent activation energy of the reaction is calculated to be 108 kJ mol⁻¹ (Fig. S4, ESI†). Since no additive is added into the reaction system, this TOF value is relatively large as compared with other catalysts, such as Pd, thus higher reaction temperature is required to initiate the dehydrogenation reaction. In addition, the TOF and activation energy values for a range of Au/TiNTs-x catalysts during HCOOH dehydrogenation are illustrated in Table S2 (ESI).† Recycling of Au/TiNTs-300 was also performed, and the catalyst exhibited high activity and excellent selectivity over 5 cycles (Fig. S5, ESI†). The stability is not good enough, and two main reasons that potentially lead to the decrease of catalytic activity. One is the gradual growth of AuNPs during catalysis that reduces the surface area of catalytically active species, and the other is the surface contaminant by carbon deposition.

The elucidation of the reaction mechanism benefits further designing of more efficient catalysts. Considering that dehydrogenation reaction often involves the participation of reactive radicals, room-temperature spin trapping technique are thus employed as an in situ analysis to detect potential radicals originated from the catalytic system in open air. As shown in Fig. 4, no EPR signals are observed in the FA solution containing DMPO spin trap reagent and bare TiNTs in open air. A triple signal (as indicated by the red circle) due to the C=N bond broken of DMPO appears as TiNTs is replaced by Au/TiNTs (blue line) or Pd/TiNTs (red line). In accompany with the triple signal, a six-line signal (as indicated by the stars) emerges in the system, which is attributed to the peroxide species, most likely ˙OOH (hyperfine splitting constants: α_N = 7.98 G and α_H = 3.61 G). In addition, a prominent 1:2:2:1 quadruple EPR peaks of DMPO–˙OH adduct (α_N = α_H = 14.9 G) appears in the Pd/TiNTs. The formation of ˙OH radical in Pd/TiNTs system indicates the strong oxidation ability of Pd, because ˙OOH can be easily decomposed into ˙OH. However, on Au/TiNTs surface, ˙OOH remains intact without any destruction. This evidence indirectly proves that the strong C–O bond in FA may be cleaved on Pd surface to ultimately form CO and H₂O, while Au is unable to break this bond. Therefore, CO molecules, which are noxious to metal catalysts, are not formed on Au/TiNTs surface.

Fig. 4 Room temperature EPR spin-trapping experiments: black line, DMPO + FA solution + TiNTs; red line: DMPO + FA solution + Pd/TiNTs; blue line, DMPO + FA solution + Au/TiNTs.

On the basis of the aforementioned results, we propose the following fundamental chemical steps responsible for the FA dehydrogenation with Au/TiNTs-300 sample. First, charge redistribution occurs by electron transfer from Au to nearby TiNTs, giving rise to Au_δ⁺ species. Next, since FA spontaneously undergoes ionization to produce HCOO⁻ anions in water or ketone, the positively charged gold surface tends to readily attract these anions. It is expected that C–H bond cleavage is preferentially occurred on the gold surface with much lower energy barriers than neutral or negatively charged surface. At last, the resulting hydride species reacts with the ionized proton to generate H₂. During the whole processes, C–O bond cleavage over the catalyst is not preferred, leading to the high selectivity of FA dehydrogenation.

Conclusions

In summary, we found that single component AuNPs supported on protonated TiNTs are able to dehydrogenize additive-free FA solution into H₂ with promising efficiency at low temperatures. The surface charge state of gold depends on the physiochemical properties of TiNTs with different high temperature treatment, and positively-charged Au/TiNTs-300 is demonstrated to be the most efficient catalyst. This surface-charge-mediated, additive-free FA decomposition method, using an efficient heterogeneous gold catalyst composed of single AuNPs dispersed on TiNTs, can make a significant contribution not only to disclose the intrinsic catalytic potential of gold but also to establish a more practical and convenient H₂ production process of industrial importance.

Acknowledgements

We are grateful for financial supports from the NSFC (21503189 and 21403197), Zhejiang Provincial Natural Science Foundation of China (LY15B030009), and China Postdoctoral Science Foundation (2014M550333 and 2015T80636).

Notes and references

1. C. Fellay, P. J. Dyson and G. Laurenczy, Angew. Chem., Int. Ed., 2008, 47, 3966.
2. M. Grasemann and G. Laurenczy, Energy Environ. Sci., 2012, 5, 8171.
3. X. J. Gu, Z. H. Lu, H. L. Jiang, T. Akita and Q. Xu, J. Am. Chem. Soc., 2011, 133, 11822.
4. T. C. Johnson, D. J. Morris and M. Wills, Chem. Soc. Rev., 2010, 39, 81.
5. B. Loges, A. Boddien, H. Junge and M. Beller, Angew. Chem., Int. Ed., 2008, 47, 3962.
6. K. Tedsree, T. Li, S. Jones, C. W. A. Chan, K. M. K. Yu, P. A. J. Bagot, E. A. Marquis, G. D. W. Smith and S. C. E. Tsang, Nat. Nanotechnol., 2011, 6, 302.
7. D. A. Bulushev, S. Beloshapkin and J. R. H. Ross, Catal. Today, 2010, 154, 7.
8. Y. J. Huang, X. C. Zhou, M. Yin, C. P. Liu and W. Xing, Chem. Mater., 2010, 22, 5122.
9. Z. L. Wang, J. M. Yan, Y. Ping, H. L. Wang, W. T. Zheng and Q. Jiang, Angew. Chem., Int. Ed., 2013, 52, 4406.
10. X. C. Zhou, Y. J. Huang, W. Xing, C. P. Liu, J. H. Liao and T. H. Lu, Chem. Commun., 2008, 3540.
11. Q. Y. Bi, X. L. Du, Y. M. Liu, Y. Cao, H. Y. He and K. N. Fan, J. Am. Chem. Soc., 2012, 134, 8926.
12. K. Jiang, K. Xu, S. Z. Zou and W. B. Cai, J. Am. Chem. Soc., 2014, 136, 4861–4864.
13. Z. L. Wang, J. M. Yan, H. L. Wang, Y. Ping and Q. Jiang, Sci. Rep., 2012, 2, 1.
14. X. Lin, B. Yang, H. M. Benia, P. Myrach, M. Yulikov, A. Aumer, M. A. Brown, M. Sterrer, O. Bondarchuk, E. Kieseritzky, J. Rocker, T. Risse, H. J. Gao, N. Nilius and H. J. Freund, J. Am. Chem. Soc., 2010, 132, 7745.
15. T. K. Ghosh and N. N. Nair, ChemCatChem, 2013, 5, 1811.
16. B. T. Qiao, A. Q. Wang, X. F. Yang, L. F. Allard, Z. Jiang, Y. T. Cui, J. Y. Liu, J. Li and T. Zhang, Nat. Chem., 2011, 3, 634.
17. M. Kitano, E. Wada, K. Nakajima, S. Hayashi, S. Miyazaki, H. Kobayashi and M. Hara, Chem. Mater., 2013, 25, 385.
18. T. Kasuga, M. Hiramatsu, A. Hoson, T. Sekino and K. Niihara, Adv. Mater., 1999, 11, 1307.
19. R. H. Li, H. Kobayashi, J. W. Tong, X. Q. Yan, Y. Tang, S. H. Zou, J. B. Jin, W. Z. Yi and J. Fan, J. Am. Chem. Soc., 2012, 134, 18286.
20. S. Ribbens, I. Caretti, E. Beyers, S. Zamani, E. Vinck, S. Van Doorslaer and P. Cool, J. Phys. Chem. C, 2011, 115, 2302.
21. H. Y. Kim, H. M. Lee and G. Henkelman, J. Am. Chem. Soc., 2012, 134, 1560.
22. M. A. Brown, F. Ringleb, Y. Fujimori, M. Sterrer, H. J. Freund, G. Preda and G. Pacchioni, J. Phys. Chem. C, 2011, 115, 10114.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental, deuterated isotopic composition studies, catalytic performance of different catalysts and life cycles. See DOI: 10.1039/c6ra19703e

---