The interparticle coupling effect of gold nanoparticles in confined ordered mesopores enhances high temperature catalytic oxidation

Abstract

Based on the experimental results of selective cyclohexanol to cyclohexanone transformation, carbon monoxide total oxidation, and the surface enhanced Raman scattering mapping technique, we show here that the interparticle coupling effect of multiple AuNPs confined within the three dimensional ordered mesoporous scaffold of EP-FDU-12 not only improves their thermal stability but also enhances their catalytic oxidation properties operated at high temperatures.

---

The assembly of metal nanoparticles (MNPs) with controllable interparticle distances, sizes and shapes is of great importance to many frontier areas of research, such as catalysis, sensors, medical diagnostics, information storage, and quantum computation, where precise control of the nanomaterial architecture and component miniaturization is required.^1–3 The interparticle coupling effect (IPCE) has been demonstrated to significantly influence the optical properties of MNPs (e.g., AuNPs), including surface enhanced Raman scattering (SERS),^4–6 because resonant energy transfer between closely spaced MNPs enables transport of electromagnetic energy at length scales below the diffraction limit. Importantly, the IPCE is also able to affect the catalytic properties of small metal particles, which have motivated a large number of researchers concerned with the preparation and characterization of a wide range of new MNPs for catalytic purposes. In particular, multiple MNPs within confined space may have unusual catalytic properties if their size and spatial distribution can be precisely controlled, while it remains a significant challenge by using conventional synthetic routes. Ordered porous materials with controllable structures show great potential in advancing the developments of various fields, such as catalysis, adsorption, separation, drug delivery, sensors, photonics, and nanodevices.^7–10 Precise tuning of pore size distribution and systematic tailoring of pore architecture are critical to control the diffusion and phase equilibrium of entrapped MNPs in restricted nanoscopic geometries, and thus perform their desirable functions. Previous studies relative to the confinement catalysis within mesoporous materials are always based on 2-D platforms, leaving 3-D spatial distributions of multiple MNPs within ordered porous structures not well implemented.

In our recent studies,^11–13 we have demonstrated that supermolecular templating technique allows tuning of the pore size of ordered porous silica (EP-FDU-12) to more than 60 nm through simple control of synthetic variables. More importantly, the unique 3-D porous structure of EP-FDU-12 makes the entrapped AuNPs difficult to migrate and thus prevents direct particle–particle aggregation upon high temperature calcination. Thus multiple AuNPs are permitted to be encapsulated in every cage at high particle concentrations, enabling interparticle interactions via overlapping of the diffusion-spheres of AuNPs. As a result, atom-migration via vapour from cage to cage is restricted and local vapour-particle equilibrium within each cage becomes possible, leading to a thermally stable AuNPs/EP-FDU-12 system. Previous study has reported that the catalytic activity of similarly sized (∼2 nm) TiC-supported Au nanoparticles depends on the interparticle distance, which is a main factor to the enhancement of the catalytic performance as well as stability of gold catalyst.¹⁴ In the present work, considering the facts that (1) the interparticle distribution of encapsulated AuNPs can be finely tuned and (2) the gold vapour (i.e., in the form of gold atoms) and AuNPs confined within the mesopores are supposed to interact strongly with each other, we suppose that organic molecules in gas-phase may also have the potential to interact with AuNPs in the ordered confined mesoscopic space with enhanced IPCE at a proper interparticle distance, leading to significant promotion of the catalytic performance.

The catalyst preparation and characterization processes were given in the ESI.† Catalytic CO combustion into CO₂ and gas-phase selective cyclohexanol oxidation into cyclohexanone using molecular oxygen as the oxidant were selected as typical high temperature reactions for investigation of the IPCE of AuNPs confined within the ordered mesopores. The former reaction is beneficial for the elimination of automobile exhaust, and the latter is a commercially important process in organic synthesis because cyclohexanone is a pivotal precursor to Nylon 6,6 and Nylon 6. As shown in Fig. 1a, for the low gold loading catalyst (1.9 wt%), no CO conversion was achieved even at reaction temperature as high as 310 °C. While high loading catalysts (24.5 wt% and 26.1 wt%) possessed light-off temperature below 100 °C, they showed negligible activity towards CO oxidation throughout the reaction temperature range from 100 to 310 °C; only ∼20% CO conversion was attained for the 26.1 wt% sample at 310 °C. The loss of the catalytic activity could be attributed to the large particle size of a single gold and the inability of silica support to form effective catalytic interfaces with gold. However, we expected that the activity loss problem at high temperature would be alleviated by introducing IPCE within the ordered mesopores, where inter-particle distance may act as a decisive role. For examples, the medium loading catalysts, such as the 13.7 wt% sample, exhibited relatively high turnover frequency (TOF) of approximately 6000 h⁻¹ at high reaction temperature (310 °C), while unfortunately it has a much higher light-off temperature, exceeding 160 °C. Notably, for the 19.5 wt% loading sample, it holds a lower light-off temperature, and increasing reaction temperature has a prominent impact on its catalytic CO conversion efficiency. The CO conversion into CO₂ started around 100 °C and increased steady up to more than 300 °C. At least 7000 h⁻¹ TOF corresponding to 95% conversion is observed at 310 °C for this specific catalyst, which is much more active than the other rest five catalysts.

Fig. 1 Catalytic curves of (a) the oxidation of CO and (b) oxidation of cyclohexanol as a function of temperature over AuNPs/EP-FDU-12 catalysts with different loading amount of Au; TOF values were calculated based on the total amount of Au.

Similar catalytic features are also observed during gas-phase oxidation of cyclohexanol into cyclohexanone at high temperatures using the same series of catalysts. Before analyzing the catalytic results, it is noted such an alternative direct gas-phase selective process is of tremendous industrial significance and is more environmentally friendly than the liquid cyclohexane oxidation and phenol hydrogenation processes that are currently employed. However, the cyclohexanol to cyclohexanone transformation is characteristic of low selectivity and low activity because of reduced concentration of cyclohexanol molecule adsorbed on the catalyst surface as well as the difficulty in dissociation of the β-C–H bond arrived from the steric hindrance of cyclohexyl group. In stark contrast, we found in this study that all of the AuNPs/EP-FDU-12 catalysts exhibited very high selectivity towards cyclohexanol to cyclohexanone transformation, and showed superior stability with catalyst exposure to reaction conditions for more than 60 h. For the catalytic performance, as shown in Fig. 1b, very low conversion (5%) and high light-off temperature (220 °C) were obtained when low loading amount (e.g., 1.9 wt%) sample was employed. For the ultrahigh loading amount catalyst (i.e., 26.1 wt%), though low light-off temperature at 160 °C was achieved, it has relatively poor TOF (1000 h⁻¹) as the reaction temperature increased to 270 °C. In contrast, the 19.5 wt% gold loading catalyst, which showed the optimal catalytic activity during CO oxidation, represented low light-off temperature and the best catalytic efficiency with TOF exceeding 1500 h⁻¹ at 270 °C.

Both of the CO and cyclohexanol oxidations indicated the optimal performance of AuNPs/EP-FDU-12 with gold loading amount at 19.5 wt%. Therefore it is desirable to find the structural difference between the 19.5 wt% catalyst and the others. We used transmission electron microscope (TEM) combined with microtomed strategy to provide precise 3-D information on the size, shape and spatial distribution of the entrapped AuNPs. Fig. 2 summarizes the TEM results for the as-prepared AuNPs/EP-FDU-12 catalysts with different loading weight percentage, giving both Au particle size and neighbour-distance distributions. The highly ordered mesoporous structure of EP-FDU-12 is obvious from TEM observation. At low loading amount (1.9 wt% and 5.3 wt%), the AuNPs are uniformly dispersed throughout the porous structures but are separated from each other and individually encapsulated in the extra-large spherical cage (Fig. 2a and b). The average size of the AuNPs is calculated to be 23.3 ± 5.3 nm for the 5.3 wt% gold loading sample, approaching the width of the spherical cage. It indicates the serious sintering of AuNPs without interparticle interaction upon high temperature treatment. As the loading amount increased, the inter particle distance as well as the single particle size decreased. When the loading amount surpassed ∼10 wt%, two or more AuNPs started to be appeared in a single cage, leading to a sharp decrease of the interparticle distance and a medium decrease of the particle size. This feature became rather dominant as the loading amount further increased to more than 20 wt%.

Fig. 2 TEM images of AuNPs/EP-FDU-12 catalysts with gold loading amount of (a) 1.9 wt%, (b) 5.3 wt%, (c) 13.7 wt%, (d) 19.5 wt%, (e) 24.5 wt%, (f) 26.1 wt% and their corresponding size (inset) and neighbour-distance distributions; scale bar = 100 nm.

Typically, for the 19.5 wt% sample (Fig. 2d), a surface-averaged Au particle size of 5.7 ± 2.0 nm was derived from the corresponding size histogram. The uneven distribution of the catalytic particles in the pore system resulted in distance distributions centered at 18.9 and 24.7 nm for the nearest and second-nearest neighbours, respectively. Such irregular spatial distributions and relatively short interparticle distances are quite common for technical catalysts.¹⁵ The high loading catalysts (Fig. 2e and f) comprised AuNPs with a narrow size distribution (4.0 ± 1.4 nm) that, in marked contrast to the 19.5 wt% sample, were evenly distributed throughout the entire pore system. This homogeneous distribution is further illustrated by the narrow neighbour-distance distributions. The average distances of 8.3 and 9.5 nm to the nearest and second-nearest neighbours, respectively, were 2–3-fold lower than for AuNPs/EP-FDU-12 with 19.5 wt% loading percentage. The most significant difference between 19.5 wt% and higher loading samples is that only two AuNPs were entrapped in a single cage for the former sample while more than two AuNPs (perhaps 4–6) were found for the latter samples.

In general, gold catalysis is sensitive to Au particle size, the Au–Au distance and the nature of the support, as has been observed with many oxidation reactions.^16–21 Herein, the support effect can be ruled out because the same mesoporous silica was used for loading AuNPs in different concentrations. Therefore, both of the particle size and IPCE may play important roles in determining the catalytic properties of entrapped AuNPs. However, it has been demonstrated that the catalytic potential of Au increased with decreasing particle size and increased dramatically below about 2 nm, partially due to the alteration of electronic structure of Au at the quantum scale.¹⁹ Considering the facts that (1) the average Au particle size is in the range of 5 to 23 nm and (2) the smallest AuNPs (∼5 nm) obtained in the 26.1 wt% sample showed unsatisfactory catalytic properties, we supposed that the particle size effect may not play a decisive role, while on the other hand, IPCE does. Along this line, in order to provide the hidden correlations between IPCE and the catalytic properties of AuNPs, we employed SERS as a powerful technique to study the interactions between organic molecules and AuNPs.

4-Mercaptobenzoic acid (4-MBA) was selected as the probe molecule for SERS detection, which can be freely adsorbed on the AuNPs surface due to the large window size of EP-FDU-12. In Fig. 3a, the integrated intensity of SERS signals for 4-MBA adsorbed on the surface of AuNPs in the region of 940–1200 cm⁻¹ was compared for samples with different Au loading amount, in which the samples with loading amount ranging from 10–20 wt% exhibited the optimal enhancement of Raman scattering (Fig. S1, ESI†). To obtain more information on the SERS property, a hexagonal plate of the 10.5 wt% loading sample is selected to conduct Raman mapping at wavenumber of 1080 cm⁻¹ to determine the specific sites with the strongest SERS response (Fig. 3b). Obviously, most of SERS signals occurred within the hexagonal plate and some noise signals at its periphery as well most likely derived from the smaller particles that beyond the detection level of the microscope. Unlike powder samples, the SERS intensity at different sites of the individual plate fluctuated. As can be seen in Fig. 3b, the strongest responding occurred at the coordinate (1, 0) in Raman map. The weak responding nearest to this site occurred at the points with coordinate (0, −1), (1, 1) and (2, 1). The intensity at (1, 0) site was over ten times higher than those at (0, −1), (1, 1) and (2, 1) sites. This result indicated that the largest electromagnetic enhancement occurred in the internal area of the hexagonal plate, where the central AuNPs suffer from the strongest integral electromagnetic interaction with nearby AuNPs. The SERS enhancement factor (G) could be evaluated by adopting the definition as follows:

G = (I_SERS/N_MPS)/(I_RS/N_Sol)

where N_Sol = CV_RS is the average number of molecules in the scattering volume (V_RS) for the normal Raman (non-SERS) measurement at a certain concentration of solution and N_MPS = d_MPSV_SERS is the average number of adsorbed molecules in the scattering volume (V_SERS) for the SERS experiments with a certain molecular density (d_MPS) in AuNPs/EP-FDU-12. I_SERS and I_RS represent the intensity of SERS and normal Raman scattering, respectively. Through a rough estimation, the SERS enhancement factor in AuNPs/EP-FDU-12 could be 10²⁻³ for AuNPs/EP-FDU-12 with 10.5 wt% loading, which is almost one order larger than other loading samples.

Fig. 3 (a) The integrated SERS intensity in the range of 940–1200 cm⁻¹ varying with loading amount of AuNPs, and (b) Raman mapping of the hexagonal particle (inset) at wavenumber of 1080 cm⁻¹ for the 10.5 wt% loading AuNPs/EP-FDU-12.

Based on the above results, it is clear that the AuNPs/EP-FDU-12 with gold loading amounts ranging from 10–20 wt% outperformed other samples in both of catalysis and SERS. Since two adjacent AuNPs have been found to be confined in a single cage for this range samples, which is in contrast to the lower or higher loading samples, where only one AuNP or more than two AuNPs are entrapped in a single cage, respectively, we suggest that the two neighbouring AuNPs may have a sort of electronic interactions with each other within the mesopores, leading to the optimal enhancement of the SERS as well as the catalytic properties. Importantly, it is suggested that the catalytic “hot” spots are present on the contiguous surface of the two entrapped AuNPs where the strongest SERS signals have been detected. However, considering the lack of available instruments to study the interactions between gas-phase molecules and solid-state catalytic AuNPs, the detailed mechanism is still controversial and further investigation including DFT calculations are requisite to provide more evidence on this aspect.

Conclusions

On the basis of catalytic CO oxidation, selective cyclohexanol transformation, and the SERS examination, we found that the IPCE is not only able to enhance the thermal stability but also the catalytic performance of AuNPs confined within EP-FDU-12, especially when two AuNPs are entrapped in a single mesocage. We believe that our approach to improve the catalytic properties of MNPs via IPCE provides a clear and generic conceptual leap forward. The outcome of this study will have practical implications and advantages for catalytic applications using well-defined nanoporous structures operating under realistic conditions.

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. A. T. Bell, Science, 2003, 299, 1688–1691.
2. M. C. Daniel and D. Astruc, Chem. Rev., 2004, 104, 293–346.
3. G. M. Whitesides, Small, 2005, 1, 172–179.
4. S. K. Ghosh and T. Pal, Chem. Rev., 2007, 107, 4797–4862.
5. M. K. Kinnan and G. Chumanov, J. Phys. Chem. C, 2010, 114, 7496–7501.
6. H. Okamoto, K. Imura, T. Shimada and M. Kitajima, J. Photochem. Photobiol., A, 2011, 221, 154–159.
7. A. Corma, Chem. Rev., 1997, 97, 2373–2419.
8. F. Hoffmann, M. Cornelius, J. Morell and M. Froba, Angew. Chem., Int. Ed., 2006, 45, 3216–3251.
9. C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and J. S. Beck, Nature, 1992, 359, 710–712.
10. J. Y. Ying, C. P. Mehnert and M. S. Wong, Angew. Chem., Int. Ed., 1999, 38, 56–77.
11. G. C. Ma, X. Q. Yan, Y. L. Li, L. P. Xiao, Z. J. Huang, Y. P. Lu and J. Fan, J. Am. Chem. Soc., 2010, 132, 9596–9597.
12. X. Q. Yan, X. J. Wang, Y. Tang, G. C. Ma, S. H. Zou, R. H. Li, X. G. Peng, S. Dai and J. Fan, Chem. Mater., 2013, 25, 1556–1563.
13. X. Q. Yan, X. J. Wang, Y. Tang, G. C. Ma, S. H. Zou, R. H. Li, X. G. Peng, S. Dai and J. Fan, Chem. Commun., 2013, 49, 7274–7276.
14. L. K. Ono and B. Roldán-Cuenya, Catal. Lett., 2007, 113, 86–94.
15. G. Prieto, J. Zecevic, H. Friedrich, K. P. de Jong and P. E. de Jongh, Nat. Mater., 2013, 12, 34–39.
16. G. C. Bond and D. T. Thompson, Catal. Rev.: Sci. Eng., 1999, 41, 319–388.
17. C. T. Campbell, Science, 2004, 306, 234–235.
18. T. V. Choudhary and D. W. Goodman, Top. Catal., 2002, 21, 25–34.
19. M. Haruta, Catal. Today, 1997, 36, 153–166.
20. A. S. K. Hashmi and G. J. Hutchings, Angew. Chem., Int. Ed., 2006, 45, 7896–7936.
21. G. J. Hutchings, Catal. Today, 2005, 100, 55–61.

---

Footnote

† Electronic supplementary information (ESI) available: Catalyst preparation and characterization; catalytic process; SERS characterization and results. See DOI: 10.1039/c6ra18726a

---