A continuous etching process for highly-active Pd nanoclusters and their in situ stabilization

Abstract

We report a simple and environmentally-benign method for the synthesis of Pd nanoclusters through a continuous etching process. With the help of L-methionine, newly-formed Pd nanoparticles (larger than 50 nm) will gradually reduce their size, finally resulting in the formation of tiny nanoclusters with a diameter around 1.4 nm. A following sol–gel process can encapsulate these highly-active nanoclusters into a silica matrix. In this way, Pd nanoclusters can be stabilized and can sustain a high temperature treatment up to 600 °C. Moreover, these Pd nanoclusters are proved to be promising for a catalytic Suzuki reaction. Due to the composite structure of nanoclusters and silica, the Pd@SiO₂ catalysts show integrated merits including good catalytic activity, high stability and notable cyclability.

---

For decades, nanoclusters with diameters smaller than 2 nm have received growing attention. Due to their highly-reduced size, nanoclusters can have distinct performance in different areas.^1–3 Specifically, the extremely unsaturated surface of nanoclusters turns out to be active sites for different catalytic reactions.^4–7 In a recent study, it has been identified that Au nanoclusters can be highly active catalysts for the conversion of CO to CO₂ as compared to large sized Au nanoparticles.¹ Similarly, Kitagawa et al.⁸ reveals that supported Ni nanocluster catalyst has a good performance in water gas shift reaction.

Nano-sized palladium plays an essential role in different catalytic reactions such as C–C cross-coupling,^9,10 olefin hydrogenation^11,12 and CO oxidation.¹³ Vigorous efforts have been exerted to control the formation of Pd nanoparticles considering that the catalytic activity is usually size-related.^14–17 However, despite these numerous achievements,^18,19 the syntheses and applications of even smaller species of Pd nanoclusters (<2 nm) still remain challenging.^20–23 In general, nanoclusters are extremely unstable in nature due to the highly unsaturated surface atoms.¹³ Recent progress shows that it is possible to form Pd nanoclusters when their surface is effectively passivated. For example, stable Pd nanoclusters can be prepared when dimethylformamide (DMF) is used as both the reaction solvent and the strong capping agent for metallic Pd.²¹ During our pursuit for a controlled-formation of stable Pd nanoclusters, our synthetic design has followed two preliminary considerations. First of all, the whole synthesis process should be environmentally benign. For this reason, reagents which are green, safe, and cheap become a favourable choice. It is highly demanded that the synthesis be carried out in aqueous solutions other than those usually-used organic solvents such as toluene and DMF. Equally importantly, Pd nanoclusters should be prepared in a way that the product should be easily collected and processed for further applications. However, despite of the successful size control to form tiny Pd nanoclusters, the formerly-reported products are usually dissolved in solutions with the help of different capping agents^21,24 so as to prevent the aggregation of nanoclusters. It turns out to be very hard to harvest these nanoclusters for further usages in practical conditions, typically as heterogeneous catalysts.

In this contribution, we adopted a simple and easy synthetic approach to prepare Pd nanoclusters through a continuous etching process. We identified that L-methionine, a sulphurous α-amino acid essential to the human body, can be used as a cheap but very effective etching agent for the formation of Pd nanoclusters in an aqueous solution. With its help, the preliminary large Pd nanoparticles (>50 nm) can gradually shape into nanoclusters around 1.4 nm. Moreover, after the formation of Pd nanoclusters, a sol–gel process is introduced to form a core–shell structure of Pd@SiO₂, which endows the Pd nanoclusters an extraordinary thermal stability due to the protection of a silica matrix. A demonstrative experiment shows that such a Pd@SiO₂ composite can be highly efficient catalyst in a typical Suzuki reaction, characterized by high catalytic activity, excellent cyclability and extraordinary stability.

Briefly, an aqueous solution of (NH₄)₂PdCl₄ was first reduced in the presence of L-methionine. The addition of NaBH₄ initiated fast reduction process as revealed by the instantaneous colour change into a dark one. A close look on the wall of the reaction flask showed big chunks of dark precipitates right after the reduction process. As shown in Fig. S1a,† the collected sample shows obvious precipitates at the bottom of cuvette, confirming the existence of large particles. Extended reaction time was then applied to achieve a gradual etching of these large particles assisted by L-methionine. The solution could turn into a transparent one with no particle observed by naked eyes (Fig. S1b†). For such Pd species after different etching time, a sol–gel process was introduced to catch these Pd samples and then they can be easily collected for further characterizations.

Fig. 1 shows the transmission electron microscopy (TEM) images of different samples, which have endured different etching time before the sol–gel process. It's not surprising that a short reaction time of 1 h forms large particles as shown in Fig. 1a. It turns out that the fast reduction process forms large quantities of nanoparticles, which show a wide range of size distribution with no specific shapes. The large ones can reach a particle size larger than 100 nm while the smaller ones show a minimum size around 3–4 nm. High resolution TEM (HRTEM) observation on a randomly-picked nanoparticle shows obvious lattice fringes (inset in Fig. 1a) and the inter-planar spacing is measured to be around 0.225 nm, which is in good agreement with the lattice distance of (111) plane for metallic palladium.²⁵ X-ray diffraction (XRD) characterization on the collected powder confirms the formation of crystalline Pd as revealed by the emergence of characteristic peaks for Pd (Fig. 2a line, JCPDS 05-0681).

Fig. 1 TEM images of Pd@SiO₂ samples which have endured different etching time during the formation of Pd nanostructures: (a) 1 h, (b) 5 h, (c) 10 h, and (d) 20 h.

Fig. 2 XRD patterns of the Pd@SiO₂ samples shown in Fig. 1, which are prepared after different etching time: (a) 1 h, (b) 5 h, (c) 10 h (d) 20 h.

Fig. 1b–d shows the TEM images of Pd@SiO₂ samples with longer etching time from 5 h to 20 h. We observed a gradual decrease of the sizes of Pd nanoparticles along with the reaction. When the etching process continues for about 5 h, most of the large particles (>50 nm) disappears. The remaining nanoparticles show a size distribution below 15 nm. HRTEM characterization reveals that these particles are also crystalline Pd. Its XRD pattern in Fig. 2 shows relatively weaker intensity of the characteristic peaks of Pd due to a reduced size.²⁶ When the Pd nanoparticles are etched for an even longer time, for example, 10 h as shown in Fig. 1c, smaller particles with size below 5 nm exist as the major product. The inset in Fig. 1c shows that these particles have a very narrow size distribution around 3.5 nm. A thorough TEM examination on this sample shows that no separate Pd nanoparticles exist away from the silica matrix, indicating that they are actually embedded inside the silica, different from what we have observed in larger particles as shown in Fig. 1a, which seem hard to be encapsulated. Finally, tiny nanoclusters with size centring at 1.4 nm form (Fig. 1d), when the etching time reaches 20 h. Despite the fact that a thick silica background makes the tiny nanoclusters a little hard to be differentiated, we are still able to observe those Pd nanoclusters inside the silica matrix with a rather uniform size (inset in Fig. 1d). The XRD patterns show no distinct peaks at this stage due to the highly reduced size of nanoclusters. The inductively coupled plasma with atomic emission spectroscopy (ICP-AES) tests confirms a 0.85% weight loading of Pd in this nanoclusters sample, which is about 30 wt% of the total Pd species introduced for the reaction.

In addition to the TEM investigations, we also tested the photo-physical property of Pd samples at different etching stages with the help of UV/Vis spectroscopy. As shown in Fig. S2,† Pd NPs shows an exponentially-decreasing absorbance in the UV region, which is similar to what J. Sokolov²⁷ has observed before. The peak around 302 nm can be ascribed to the surface plasmon band of large particles. When particles gradually decrease their size as a result of the etching process, this plasmon band will accordingly reduce its strength. As for the 1.4 nm nanoclusters, this characteristic adsorption at 302 nm will totally disappear due to the highly reduced size.^28,29 Therefore, we can only observe a smooth curve without absorption peak characteristically for Pd nanoclusters.³⁰

The change in particles size provides a straightforward evidence of etching effect for the formation of Pd nanoclusters. After the reduction of Pd, L-methionine is the only additive we use to control the formation of Pd species and it plays a critical role in our synthesis design. Without the addition of L-methionine, those large Pd nanoparticles which form right after the reduction are quite robust and no significant change in the shape are observed as the reaction continues. Fig. S3a† shows the TEM image of those large Pd nanoparticles sampled after 20 h reaction at the absence of L-methionine. No obvious etching process can be observed and the Pd species exist as large nanoparticles. Our further investigation shows that a minimum concentration of L-methionine should be guaranteed if these Pd nanoclusters are to be expected in good quality. For example, at a low L-methionine concentration of 2.25 mM, nanoparticles rather than nanocluster will form as shown in Fig. S3b.†

Although the use of an etching process to achieve Pd nanoclusters has scarcely been discussed before, the strong thiol–metal interaction is known to be crucial for the formation of nanoclusters of Ag and Au.^31,32 Considering the importance of L-methionine in the formation of Pd nanoclusters, it's therefore proposed that such a sulfur-containing agent should play a similar role as what thiol usually does on Au nanoclusters: the fast reduction of Pd leads to the formation of irregular nanoparticles and then the sulfur atom can bond strongly with Pd surface atoms.^33,34 Through this reaction, a gradual etching process takes effect and is able to transform large Pd nanoparticles into small nanoclusters when a reservoir of sulphur atoms exists in the solution. It is worthy of noting that the selection of L-methionine is highly favourable in our pursuit of an environmentally friendly synthetic protocol. First, L-methionine itself is known as a proteinogenic amino acid which is what the body uses to make protein and peptides. It is of minimum hazard as compared to the other mercapto-group containing capping agents like mercaptoethanol. Second, L-methionine is commercially available and can be massively produced at a relatively low cost. For example, its price is 1/40 of that of mercaptoethanol. Third, L-methionine is hydrophilic in nature due to the amine (–NH₂) and the carboxylic acid (–COOH) functional groups of amino acid. We consider it an important character which not only makes an environment-friendly process possible by avoiding organic solvents, but also makes it easy to further functionalize the newly-formed Pd nanoclusters. Specifically, a subsequent sol–gel process can be easily applied to encapsulate nanoclusters. In this way, a core–shell structure of Pd@SiO₂ can form as shown in Fig. 1c and d.

The encapsulation of Pd nanoclusters into the silica matrix can not only make the product much easier be centrifuged out for further applications, but also endow these nanoclusters a much higher thermal stability due to the protection of a thermally-stable silica matrix.³⁵ As shown in Fig. 3a, we observe no substantial morphology change of the nanoclusters after the sample being heating at 600 °C for 2 h. The nanoclusters only observe a slight size increase from 1.4 nm to 2.0 nm. Even after a high temperature treatment at 700 °C, Pd nanoparticles can still maintain at around 2.8 nm (Fig. 3b). It is noteworthy that nanoclusters are well-known for their vulnerability to high temperature treatments due to melting point depression.^35,36 The existence of silica turns out to be an effective shield to endow the nanoclusters good sustainability against the harsh environment. For comparison, when these nanoparticles are just supported onto a preformed silica matrix through impregnation, (see the Experimental section for preparation details), they are easy to sinter into large particles when no encapsulation exists and the nanoclusters are directly exposed to the high temperature. As shown in Fig. 3c, large-sized nanoparticles around 40 nm will form when this impregnated sample is heated at 700 °C for 2 h.

Fig. 3 (a and b) TEM images of the nanoclusters sample which has been treated at 600 °C (a) and 700 °C (b), respectively. (c) TEM image of a contrast sample which is prepared by an impregnation method and then is subjected to 700 °C heat treatment.

As inspired by the small size of Pd nanoclusters and the composite structure of Pd@SiO₂, we consider it a good model system with promising potentials in different areas, especially heterogeneous catalyst cross-coupling.^10,37–40 We therefore carried out demonstrative experiments by using the Suzuki C–C coupling reaction as an example. First, the nitrogen adsorption–desorption isotherms (Fig. S4†) show typical curves for mesoporous structures.^26,41 The surface area of the Pd@SiO₂ composites is measured to be as high as 815.1 m² g⁻¹. It reveals that the silica matrix is mesoporous and it will be easy for the reactants to diffuse through. Scheme 1 shows a typical Suzuki C–C coupling reaction in which iodobenzene and phenylboronic acid will form biphenyl with the assistance of catalyst. In this reaction, the turn over frequency (TOF) which refers to the turnover number of iodobenzene per unit time has been adopted to measure the catalytic activity.

Scheme 1 Suzuki cross-coupling reaction of iodobenzene with phenyl boronic acid.

Table 1 shows the values of the TOFs for these size-controlled samples and the Pd nanoclusters turn out to have the highest catalytic activity. For those large particles (100 nm), they show a low activity during the formation of biphenyl with a TOF value of 31.6 min⁻¹. We observe a gradual increase in the catalytic activity and the nanoclusters reaches a high TOF value of 98.01 min⁻¹, which is about 3 times as much as that of the original larger particles. It has been expected that smaller particles will have more surface atoms and can be more active as compared to those larger ones.⁴² In our experiments, these nanoclusters show an obvious advantage as far as the catalytic activity is concerned. The Pd@SiO₂ nanocatalyst we produced can have a good cyclability in benefit of the protection from the silica matrix. In comparison, we also measured the performance of a Pd(II) complex as Pd(II) acetate which has been used as a typical homogeneous catalyst for the same Suzuki reaction.³⁸ A TOF value of 75 min⁻¹ is achieved for Pd(II) acetate under the same testing conditions. As compared to the catalyst of Pd(II) acetate, which has been dissolved in the reaction media and can hardly be separated for continuous uses, the solid powders of Pd@SiO₂ catalysts can be easily collected after each test cycle by centrifuging, and then be regenerated by heating at a high temperature at 450 °C, which has been proved to have no impact on Pd nanoclusters. In our experiments, we didn't observe an obvious fading on the catalytic activity (Fig. 4), which makes the Pd@SiO₂ truly reliable catalysts for their continuous use.

Table 1 Catalysis tests the size-controlled samples. The yields are measured after 5 minutes' reaction. For the first two samples collected at the early stage of etching, they have broad size distributions as shown in Fig. 1a and b

Size (nm)	Loading (%)	Yield	TOF (min^−1)
Up to 100	1.27	64.03%	31.6
Up to 15	1.28	66.47%	47
3.5	0.66	54.14%	73.13
1.4	0.85	91.93%	98.01

---

Fig. 4 Five continuous tests on the performance of Pd nanoclusters. The yields are measured after 5 minutes' reaction.

In summary, a green, environmentally benign process is developed to prepare Pd nanoclusters in aqueous solution. The cheap and innocuous L-methionine is selected as capping agent to gradually etch Pd nanoparticles into small nanoclusters with a diameter around 1.4 nm. A following sol–gel process is used to encapsulate the nanoclusters into a silica matrix, resulting in a core shell structure of Pd@SiO₂. Such a composite structure endows the nanoclusters a high thermal stability so that they can sustain a high temperature treatment up to 700 °C. Moreover, the nanoclusters are proven to be promising for a catalytic Suzuki reaction. Due to the composite structure of nanoclusters and silica, the Pd@SiO₂ catalysts show integrated merits including good catalytic activity, high stability and notable cyclability.

Experimental section

180 μl (0.5 M) aqueous (NH₄)₂PdCl₄ was added into 20 ml water at 60 °C, and then a suitable amount of methionine (such as L-methionine: (NH₄)₂PdCl₄ = 0.5, 1, 10, 20) was introduced. The solution turns clear accompanied by vigorous stirring. NaBH₄ powder was then quickly added (10 : 1 for molar ratio between NaBH4 and Pd). The solution turned black immediately and flocculated precipitations could be observed. As the reaction continued, the suspension gradually turned homogeneous and large particles disappeared. Finally a transparent solution formed without signs of visible particles. Next, 0.1 g cetyltrimethylammonium bromide (CTAB) was added together with 0.2 ml (1 M) NaOH aqueous solution. After that, 1.0 ml tetraethoxylsilicate (TEOS) was dropped into the solution to initiate the hydrolysis step. The reaction was allowed to continue for another 2 hours and the solution eventually turned opaque. The Pd@SiO₂ sample was collected as solid product by a continuous centrifugation/washing cycle for 5 times. The powder was then dried at 80 °C overnight and then reduced in an H₂/Ar flow for 2 h.

A supported sample as Pd–SiO₂ synthesized via an impregnation method. The silica matrix firstly prepared from a separated sol–gel process, and then a certain amount of aqueous solution containing those Pd nanoclusters dropped onto the silica powder. After a continuous stirring for 2 h, the slurry dried by heating on a hot plate. The sample is then collected for further use as Pd–SiO₂.

Characterizations

Transmission electron microscopy (TEM) was carried out on a JEOL 2100 F instrument operating at 200 kV. XRD was performed on a Rigaku D/max 2500 type instrument using CuKa radiation with a step size of 0.01°. N₂ adsorption–desorption isotherms were obtained at liquid nitrogen temperature on a Micrometrics ASAP 2010 apparatus. The sample was outgassed at 350 °C prior to adsorption. Pore size distribution was obtained with the N₂ desorption branch, using the Barrett–Joyner–Halenda (BJH) method. The species surface area of the samples was determined by the Brunauer–Emmett–Teller (BET) method.

Acknowledgements

This work was supported by the major State Basic Research Program of China (973 program: 2013CB934000, 863 program: 2013AA030800), the National Natural Science Foundation of China (Grant no. 21373238), and the Chinese Academy of Sciences. Rongwen Lu is supported by the National Natural Science Foundation of China (Grants no. 20976023 and no. 21176038).

References

1. A. A. Herzing, C. J. Kiely, A. F. Carley, P. Landon and G. J. Hutchings, Science, 2008, 321, 1331–1335.
2. L. A. Peyser, A. E. Vinson, A. P. Bartko and R. M. Dickson, Science, 2001, 291, 103–106.
3. Y. Cui, Q. Wei, H. Park and C. M. Lieber, Science, 2001, 293, 1289–1292.
4. W. E. Kaden, T. Wu, W. A. Kunkel and S. L. Anderson, Science, 2009, 326, 826–829.
5. A. Samanta and R. N. Devi, ChemCatChem, 2013, 5, 1911–1916.
6. M. Pagliaro, V. Pandarus, F. Beland, R. Ciriminna, G. Palmisano and P. D. Cara, Catal. Sci. Technol., 2011, 1, 736–739.
7. L. Yang, C. Guan, W. Yue, J. Wu, H. Yan, X. Zhang, Z. Wang, X. Zhan, Y. Li, D. Wang and L. Wan, Sci. China: Chem., 2012, 56, 124–130.
8. H. Kitagawa, N. Ichikuni, H. Okuno, T. Hara and S. Shimazu, Appl. Catal., A, 2014, 478, 66–70.
9. M. Pérez-Lorenzo, J. Phys. Chem. Lett., 2012, 3, 167–174.
10. N. Kim, M. S. Kwon, C. M. Park and J. Park, Tetrahedron Lett., 2004, 45, 7057–7059.
11. N. Marín-Astorga, G. Pecchi, T. J. Pinnavaia, G. Alvez-Manoli and P. Reyes, J. Mol. Catal. A: Chem., 2006, 247, 145–152.
12. D. Teschner, J. Borsodi, A. Wootsch, Z. Revay, M. Havecker, A. Knop-Gericke, S. D. Jackson and R. Schlogl, Science, 2008, 320, 86–89.
13. W. E. Kaden, T. Wu, W. A. Kunkel and S. L. Anderson, Science, 2009, 326, 826–829.
14. O. M. Wilson, M. R. Knecht, J. C. Garcia-Martinez and R. M. Crooks, J. Am. Chem. Soc., 2006, 128, 4510–4511.
15. Y. Takasu, T. Akimaru, K. Kasahara, Y. Matsuda, H. Miura and I. Toyoshima, J. Am. Chem. Soc., 1982, 104, 5249–5250.
16. W. P. Zhou, A. Lewera, R. Larsen, R. I. Masel, P. S. Bagus and A. Wieckowski, J. Phys. Chem. B, 2006, 110, 13393–13398.
17. Y. Li, E. Boone and M. A. El-Sayed, Langmuir, 2002, 18, 4921–4925.
18. E. Ozensoy, B. K. Min, A. K. Santra and D. W. Goodman, J. Phys. Chem. B, 2004, 108, 4351–4357.
19. T. Prasomsri, D. Shi and D. E. Resasco, Chem. Phys. Lett., 2010, 497, 103–107.
20. H. Zhang, T. Watanabe, M. Okumura, M. Haruta and N. Toshima, Nat. Mater., 2012, 11, 49–52.
21. M. Hyotanishi, Y. Isomura, H. Yamamoto, H. Kawasaki and Y. Obora, Chem. Commun., 2011, 47, 5750–5752.
22. G. Li and R. C. Jin, Acc. Chem. Res., 2013, 46, 1749–1758.
23. G. Schmid, M. Baumle, M. Geerkens, I. Heim, C. Osemann and T. Sawitowski, Chem. Soc. Rev., 1999, 28, 179–185.
24. Y. Li and M. A. El-Sayed, J. Phys. Chem. B, 2001, 105, 8938–8943.
25. V. Mazumder and S. Sun, J. Am. Chem. Soc., 2009, 131, 4588–4589.
26. I. Yuranov, P. Moeckli, E. Suvorova, P. Buffat, L. Kiwi-Minsker and A. Renken, J. Mol. Catal. A: Chem., 2003, 192, 239–251.
27. C. K. Yee, R. Jordan, A. Ulman, H. White, A. King, M. Rafailovich and J. Sokolov, Langmuir, 1999, 3486–3491.
28. D. E. Cliffel, F. P. Zamborini, S. M. Gross and R. W. Murray, Langmuir, 2000, 9699–9702.
29. J. A. Creighton and D. G. Eadon, J. Chem. Soc., Faraday Trans., 1991, 3881–3891.
30. M. Chen, J. Falkner, W. H. Guo, J. Y. Zhang, C. Sayes and V. L. Colvin, J. Colloid Interface Sci., 2005, 287, 146–151.
31. Z. Wu, J. Suhan and R. Jin, J. Mater. Chem., 2009, 19, 622.
32. O. Toikkanen, V. Ruiz, G. R. nnholm, N. Kalkkinen, P. Liljeroth and B. M. Quinn, J. Am. Chem. Soc., 2008, 130, 11049–11055.
33. M. Ganesan, R. G. Freemantle and S. O. Obare, Chem. Mater., 2007, 19, 3464–3471.
34. M. Cargnello, N. L. Wieder, P. Canton, T. Montini, G. Giambastiani, A. Benedetti, R. J. Gorte and P. Fornasiero, Chem. Mater., 2011, 23, 3961–3969.
35. S. H. Joo, J. Y. Park, C.-K. Tsung, Y. Yamada, P. Yang and G. A. Somorjai, Nat. Mater., 2008, 8, 126–140.
36. P. M. Arnal, M. Comotti and F. Schuth, Angew. Chem., Int. Ed., 2006, 45, 8224–8227.
37. Y. Li, J. H. Liu, C. A. Witham, W. Huang, M. A. Marcus, S. C. Fakra, P. Alayoglu, Z. Zhu, C. M. Thompson, A. Arjun, K. Lee, E. Gross, F. D. Toste and G. A. Somorjai, J. Am. Chem. Soc., 2011, 133, 13527–13533.
38. C. A. Witham, W. Huang, C.-K. Tsung, J. N. Kuhn, G. A. Somorjai and F. D. Toste, Nat. Chem., 2009, 2, 36–42.
39. R. Narayanan and M. A. El-Sayed, J. Am. Chem. Soc., 2003, 125, 8340–8347.
40. V. V. Pushkarev, Z. Zhu, K. An, A. Hervier and G. A. Somorjai, Top. Catal., 2012, 55, 1257–1275.
41. H. Xia, Y. Wang, J. Lin and L. Lu, Nanoscale Res. Lett., 2012, 7, 33.
42. S. Nie, Science, 1997, 275, 1102–1106.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra02047b

---