Geometric effect of Au nanoclusters on room temperature CO oxidation

Yafeng Cai ab, Yun Guo *b and Jingyue Liu *a
aDepartment of Physics, Arizona State University, Tempe, Arizona 85287, USA. E-mail: Jingyue.liu@asu.edu
bKey Laboratory for Advanced Materials and Research Institute of Industrial Catalysis, School of Chemistry and Molecular Engineering, East China University of Science and Technology, 130 Meilong Road, Shanghai 200237, China. E-mail: yunguo@ecust.edu.cn

Received 28th October 2019 , Accepted 28th November 2019

First published on 2nd December 2019


We report the effect of in situ transforming ZnO supported single Au atoms to 3-dimensional Au clusters (Au3D) on room temperature CO oxidation activity. We discovered that an intermediate, highly distorted 2-dimensional Au layer species is much more active than both the Au3D clusters and the Au single atoms. The geometric arrangement of Au clusters and their interactions with support surfaces play a pivotal role in determining their catalytic performance in room temperature CO oxidation on ZnO supported Au catalysts.


The prominent size effect of metal nanoparticles (NPs), due to the change in the coordination number, chemical state and metal support interaction with the particle size, has been widely proposed to account for the experimentally observed catalytic performance of supported metal catalysts for various important chemical transformations.1–6 The most striking example is that supported smaller Au NPs exhibit substantial catalytic activity in many reactions (e.g., CO oxidation) while larger Au NPs are essentially inactive.7–9 Although the origin of highly active nano Au for CO oxidation has been intensively investigated, it still remains a controversial topic on the nature of the most active forms of Au species. To date, the sizes of Au species,10,11 bilayer Au nanostructures,10,11 Au charge states,12,13 and Au–metal oxide interfacial perimeters14–16 were all proposed to account for the unique catalytic properties of nanostructured Au species. Most of the experimental studies focused on Au NPs with sizes ranging from 2 to 5 nm. The study of the catalytic properties of ultra-small (<2 nm) Au species on various supports should be of both scientific interest and practical importance. Theoretical studies show that ultra-small Au clusters may have significant geometric effects.17,18 Mavrikakis et al.19 found that the expansion strain in Au clusters can increase the reactivity of Au and such a strain effect increases rapidly with the decreasing cluster size. In practice, most of the supported metal catalysts often contain metal single atoms, nanoclusters and various sizes of NPs, and such heterogeneity of potential active species hampers the deeper understanding of how each individual species contributes to the observed activity.20 Supported Au single-atom catalysts (SACs) have been synthesized by the wet-chemistry method.5,20,21 However, the facile synthesis of supported Au clusters with narrow size distributions and a better defined morphology is still a challenge.

Very recently, several research groups13,22,23 used supported metal atoms as precursors to obtain supported Au cluster catalysts by reduction and/or calcination treatment. In this work, we report an in situ transformation of Au single atoms to 3D nanoclusters (Au3D) through an intermediate Au monolayer configuration (Aulayer) during room temperature CO oxidation. Our experimental results unambiguously demonstrate that the Aulayer configuration yielded the highest reactivity. Strong support-induced distortion of the Aulayer nanoclusters is proposed to be responsible for the observed high activity for room temperature CO oxidation.

Flower-like 2D ZnO nanosheets with an average thickness of 9 nm were prepared by a hydrothermal method and the subsequent calcination to eliminate organic species (Fig. S1–S3, ESI). The Au1/ZnO SAC with 0.78 wt% of Au (determined by ICP-AES) was prepared by room temperature UV-assisted photochemical deposition of Au single atoms onto the as-synthesized 2D ZnO flower-like powders. The specific area of Au1/ZnO SAC was measured to be 46.6 m2 g−1, according to the Brunauer–Emmett–Teller (BET) method, by the N2 adsorption isotherm (Fig. S4, ESI). The X-ray diffraction pattern of the Au1/ZnO catalyst showed only the characteristic peaks of wurtzite ZnO (Fig. S5, ESI). The representative atomic resolution HAADF-STEM image (Fig. 1a) shows exclusively Au single atoms (marked in red circles) uniformly dispersed on the surfaces of the ZnO support. Careful examination of different sample regions revealed that only Au single atoms were present in Au1/ZnO SAC and these Au atoms occupy the exact positions of the Zn cations on the ZnO surfaces (Fig. S6, ESI).


image file: c9cc08381b-f1.tif
Fig. 1 (a) Representative HAADF-STEM image of Au1/ZnO SAC. Au single atoms are marked in red circles. For clarity, only a few isolated Au atoms are circled although many more are present in the image. (b) Time-dependent CO2 (m/z = 44) signals during CO oxidation over the 0.78 wt% Au1/ZnO SAC (red plot) and the ZnO support (black plot). The CO2 formation rate with the reaction time (blue plot) is displayed in the unit of μmol (gAu−1 s−1).

We tested CO oxidation over Au1/ZnO SAC by measuring the CO2 signal (m/z = 44) with an on-line mass spectrometer (Fig. 1b). It should be noted that no CO2 molecules were released during the first 2 minutes of CO introduction, indicating that Au1/ZnO SAC was either not active or needed a longer incubation time. However, the CO2 signal then increased rapidly with reaction time: the CO2 formation rate increased quickly to a maximum (461.5 μmol gAu−1 s−1) at ∼8 minutes. If we assume that all the Au atoms were accessible we obtain a turnover frequency of 9.14 × 10−2 s−1. When the CO oxidation proceeds further, the CO2 formation rate decreased gradually to 205.1 μmol gAu−1 s−1 after ∼60 minutes. Such a time-dependent reaction rate suggests that the catalyst structure continuously evolved. There was no observable CO2 formation on the pristine ZnO support, suggesting that the observed CO2 molecules exclusively originated from the Au species.

For low temperature CO oxidation, the accumulation of surface carbonate species is usually considered to be one of the main reasons for deactivation.24 To verify if the carbonate species played a major role in our catalyst, temperature programed desorption of CO2 under a He flow environment was conducted immediately after 60 minutes of CO oxidation (Fig. S7, ESI). There were little CO2 desorbed from the used catalyst suggesting that the activity decrease was not caused by strong adsorption of CO2 or accumulation of carbonate species.

A detailed electron microscopy study was conducted on the Au/ZnO catalysts after CO oxidation for 4, 8 and 60 minutes. A gas-flow reaction setup was used to treat the samples (Scheme S2, ESI). As discussed above, the fresh catalyst contained only Au single atoms. However, after room temperature CO oxidation for 4 minutes, Au nanoclusters with an average size of 0.9 nm were formed (Fig. 2) and simultaneously the number density of the single Au atoms significantly decreased (Fig. S8, ESI). Therefore, ZnO supported single Au atoms started sintering as soon as the room temperature CO oxidation started. When CO oxidation proceeded further, statistical analyses of more than 300 Au clusters from each sample indicated that the average size of the Au clusters did not change appreciably (Fig. 2d) with an average size of 0.9 nm after 8 min and a slightly larger average size of 1.1 nm after 60 min. Some single Au atoms were still present in these samples although the number density of the single Au atoms decreased by ∼97.7% after 60 min of CO oxidation (Fig. S8, ESI). The number density of the Au clusters changed appreciably after 8 min of CO oxidation compared with that after 60 min, suggesting that the total number of Au atoms in each Au cluster changed significantly although their sizes did not change much.


image file: c9cc08381b-f2.tif
Fig. 2 Representative HAADF-STEM images of Au/ZnO after CO oxidation for (a) 4 minutes, (b) 8 minutes, and (c) 60 minutes. The size distributions of the Au nanoclusters after different times of the CO oxidation reaction (d) did not show significant changes.

The activity of Au catalysts for CO oxidation is, in general, predominantly affected by the size of the Au species and the redox properties of the support. In this work, the ZnO support does not possess high redox capability, especially at ambient temperatures. Our experimental data, however, clearly show that the Au clusters with similar sizes yielded very different activity. These results suggest that the geometric configuration of Au clusters may have an important role in CO oxidation.

To understand the relationship between the Au geometric configuration and the catalytic performance, we analyzed the atomic arrangement of Au atoms in the Au nanoclusters. After 4 min, monolayers of Au species (Aulayer) were observable (Fig. 3a) with ∼0.25 and ∼0.28 nm lattice spacings, similar to the spacings of ZnO(101) and ZnO(100), in contrast to the bulk Au lattice spacings of dAu(111) = 0.24 nm and dAu(200) = 0.20 nm. After 8 min, in addition to Aulayer, Au3D clusters appeared (indicated by the yellow squares in Fig. 3b and c). The measured lattice spacings of these Aulayer (∼0.25 and ∼0.26 nm) still match well with those of ZnO(101) and ZnO(002), but are much larger than those of the bulk Au. After 60 min, the Aulayer species were not detectable and only Au3D nanoclusters and a tiny amount of single Au atoms were detected, suggesting that the transformation of Au single atoms to Au3D nanoclusters completed. The few Au atoms that still persisted on the ZnO surfaces could have occupied the Zn cation vacancy positions. The monolayer Aulayer species seem to be the transient intermediate state of transforming single Au atoms to Au3D nanoclusters during the CO oxidation reaction. The Au3D nanoclusters (Fig. 3f) showed lattice spacings of 0.20 and 0.24 nm, matching well with the bulk Au{200} and Au{111} spacings.


image file: c9cc08381b-f3.tif
Fig. 3 Representative atomic-resolution HAADF-STEM images of the Au/ZnO catalyst after CO oxidation for (a and d) 4 minutes, (b and e) 8 minutes, and (c and f) 60 minutes. Au single atoms, Aulayer and Au3D nanoclusters are marked in red circles, cyan squares and yellow squares, respectively. The magnified images (d–f) correspond to the cyan squares in (a and b) and the yellow square in (c), respectively.

The atomic resolution HAADF studies clearly show that (1) the majority of the as-prepared Au single atoms are not stable under a CO environment, (2) the stabilized Au clusters take a 3D shape with lattice spacings close to those of the bulk Au, and (3) the transient 2D monolayer species are significantly distorted in order to form an epitaxial relationship with the ZnO support. Austin and co-workers17 conducted DFT calculations of the shape effect on the adsorption behavior of small Au nanoclusters (∼1 nm) and found that planar Au nanoclusters would transit to Au3D at high CO coverage. Our experimental results seem to support this conclusion for ZnO supported Au species.

Based on the experimental results, we can deduce the following structural change in the Au atoms during room temperature CO oxidation: under a CO environment, single Au atoms assemble to initially form highly distorted Aulayer that aligns with the support lattice spacings and then the Aulayer species transform to Au3D when the Aulayer/ZnO interfacial energy becomes larger. If we assume that the CO conversion rate on the ZnO supported single Au atoms is small or negligible and that the highly distorted 2D Aulayer species are most active then we can fully understand the time dependent behavior of the CO2 formation rate. The reactivity of these ZnO supported Au species follows the sequence Aulayer > Au3D ≫ Au1.

To gain detailed information about the chemical states of the different Au species, fresh Au1/ZnO SAC and after 4, 8 and 60 minutes of CO oxidation were immediately examined by XPS (Fig. S9 (ESI) and Fig. 4a). Although most of the Au 4f peaks were buried under the strong Zn 3p3/2 peak (Fig. S10, ESI), a small shoulder peak is clearly observable at ∼85 eV in the fresh Au1/ZnO. This peak can be assigned to the positively charged Au species (Auδ+).25–28 Phala and co-workers29 found, by DFT calculations, that oxidized Au can be stabilized at Zn2+ positions in the bulk terminated surface sites. Our STEM imaging results also showed that Au single atoms are located on the positions of Zn2+ sites (not necessarily the Zn2+ vacancies) (Fig. 1a and Fig. S6, ESI). The interaction between the Au atoms and the ZnO surface oxygen induces electron transfer from the Au atoms to the ZnO support, yielding the high content of Auδ+. When CO oxidation started the content of the Auδ+ species drastically decreased while the amount of the Au0 species increased. The XPS data clearly show that the change in the Au oxidation state followed the structural change of the Au species. The in situ DRIFTS study of the as-prepared Au1/ZnO under CO oxidation conditions (Fig. S11, ESI) shows that the relative CO2/CO adsorption peak ratio increased dramatically during the first several minutes of the reaction, reached the maximum at ∼8 min, and then slowly decreased, corroborating the activity results shown in Fig. 1b.


image file: c9cc08381b-f4.tif
Fig. 4 (a) High resolution XPS spectra of the Au 4f core-level of Au/ZnO for CO oxidation after different time intervals (0, 4, 8 and 60 minutes). (b) IR spectra of CO adsorbed on Au/ZnO at room temperature after reaction for different time intervals (4, 8, 30 and 60 minutes). The band of CO adsorbed on Aulayer species at different time intervals is highlighted in red in each XPS spectrum.

To further understand the nature of the active Au species we conducted CO-DRIFTS experiments on the Au1/ZnO SAC. After each specified reaction time, the residual gas was purged with pure Ar and then the CO-DRFITS spectra were recorded (Fig. 4b). Three adsorption bands were identified at around 2200, 2110, and 2165 cm−1, respectively. The band at ∼2110 cm−1 can be assigned to CO adsorption on Au0 sites.30 The band at ∼2200 cm−1 can be assigned to CO adsorption on Au3+ sites31,32 or CO adsorption on ZnO.33,34 The CO adsorption band on Au3+ sites is considered to be extremely weak and is difficult to be even detected since Au3+ is normally coordinately saturated and can be easily reduced to Auδ+ by CO.23,31,35 Both the XPS and the STEM imaging data suggest that with reaction time the Au0 species should increase and the high valence Au species should decrease. Since Fig. 4b shows that the intensity of the 2200 cm−1 band increases with reaction time, the assignment of this band to Au3+ is not plausible. It was reported that CO could adsorb on ZnO at 2200 cm−1 in the presence of CO2.33,34 We detected CO adsorption on the treated ZnO at about 2190 cm−1 under similar conditions (Fig. S12, ESI). Therefore, we assigned the 2200 cm−1 band to CO adsorption on Zn cations. As discussed above, the isolated Au atoms move to form Aulayer and Au3D as the reaction proceeds. The CO molecules that adsorb on the active Au clusters react with surface oxygen species on ZnO to produce CO2 and simultaneously create more exposed Zn2+ sites around the Au species. Both the migration of Au atoms and the oxidation of the adsorbed CO molecules lead to increased intensity of CO adsorption on the exposed Zn cation sites.

The shoulder band centered at ∼2165 cm−1 can be assigned to the CO adsorption on Auδ+ sites.36 After deconvolution of the adsorption bands, one can observe that the peak area of the 2165 cm−1 Auδ+ band first increased, reached a maximum at ∼8 min, and then decreased gradually. This time dependent change in the amount of Auδ+ sites coincides with the time dependent change in the CO oxidation activity (Fig. 1b) and the formation of the Aulayer (Fig. 3). The STEM images showed that the Aulayer species possess a longer Au–Au bond which should possess a stronger adsorption strength of CO than that on Au3D due to strain induced effects.18,19 The strong interaction between the Aulayer and the ZnO surface facilitates electron transfer from the Au species to the ZnO support, resulting in a larger amount of positively charged Au species. When Au3D were formed the perimeter Au atoms may behave similar to the Aulayer species and CO adsorption at Auδ+ sites still exists but in a smaller amount. The excellent match between the amounts of the Auδ+ sites with the CO oxidation activity further confirms that the Aulayer species with Auδ+ sites exhibit the highest activity.

In summary, our work shows highly distorted monolayer Au species as the most active sites for room temperature CO oxidation on ZnO supported Au catalysts. Both ZnO supported Au atoms and Aulayer species, however, are not stable: they evolve into 3D Au clusters. ZnO supported single Au atoms are not active for room temperature CO oxidation. For small Au clusters, the cluster geometry and their electronic interaction with support surfaces are critical. This work provided deeper insights into the structure–performance relationship of supported Au species.

This work was supported by the US National Science Foundation under CHE-1465057 (YC and JL), the National Key Research and Development Program of China (2016YFC0204300), the NSFC of China (21571061 and 21908079), and the Pujiang Program of the Shanghai Municipal Human Resources and Social Security Bureau (18PJD011). The authors gratefully acknowledge the use of facilities within the John M. Cowley Center for High Resolution Electron Microscopy at Arizona State University.

Conflicts of interest

There are no conflicts to declare.

Notes and references

  1. B. Roldan Cuenya and F. Behafarid, Surf. Sci. Rep., 2015, 70, 135–187 CrossRef CAS.
  2. K. An and G. A. Somorjai, Catal. Lett., 2015, 145, 233–248 CrossRef CAS.
  3. S. Cao, F. Tao, Y. Tang, Y. Li and J. Yu, Chem. Soc. Rev., 2016, 45, 4747–4765 RSC.
  4. B. R. Cuenya, Thin Solid Films, 2010, 518, 3127–3150 CrossRef CAS.
  5. L. Liu and A. Corma, Chem. Rev., 2018, 118, 4981–5079 CrossRef CAS.
  6. T. Takei, T. Akita, I. Nakamura, T. Fujitani, M. Okumura, K. Okazaki, J. H. Huang, T. Ishida and M. Haruta, Adv. Catal., 2012, 55, 1–126 CAS.
  7. M. Haruta, T. Kobayashi, H. Sano and N. Yamada, Chem. Lett., 1987, 405–408 CrossRef CAS.
  8. M. Haruta, Catal. Today, 1997, 36, 153–166 CrossRef CAS.
  9. G. J. Hutchings, ACS Cent. Sci., 2018, 4, 1095–1101 CrossRef CAS PubMed.
  10. M. Valden, X. Lai and D. W. Goodman, Science, 1998, 281, 1647–1650 CrossRef CAS PubMed.
  11. A. A. Herzing, C. J. Kiely, A. F. Carley, P. Landon and G. J. Hutchings, Science, 2008, 321, 1331–1335 CrossRef CAS PubMed.
  12. A. Y. Klyushin, M. T. Greiner, X. Huang, T. Lunkenbein, X. Li, O. Timpe, M. Friedrich, M. Hävecker, A. Knop-Gericke and R. Schlögl, ACS Catal., 2016, 3372–3380 CrossRef CAS.
  13. S. Wei, X.-P. Fu, W.-W. Wang, Z. Jin, Q.-S. Song and C.-J. Jia, J. Phys. Chem. C, 2018, 122, 4928–4936 CrossRef CAS.
  14. M. Haruta, Faraday Discuss., 2011, 152, 11–32 RSC.
  15. Y. Y. Wu, N. A. Mashayekhi and H. H. Kung, Catal. Sci. Technol., 2013, 3, 2881 RSC.
  16. B. Shao, J. Zhang, J. Huang, B. Qiao, Y. Su, S. Miao, Y. Zhou, D. Li, W. Huang and W. Shen, Small Methods, 2018, 1800273 CrossRef.
  17. N. Austin, J. K. Johnson and G. Mpourmpakis, J. Phys. Chem. C, 2015, 119, 18196–18202 CrossRef CAS.
  18. J.-X. Liu, I. A. W. Filot, Y. Su, B. Zijlstra and E. J. M. Hensen, J. Phys. Chem. C, 2018, 122, 8327–8340 CrossRef CAS PubMed.
  19. M. Mavrikakis, P. Stoltze and J. K. Nørskov, Catal. Lett., 2000, 64, 101–106 CrossRef CAS.
  20. J. Liu, ACS Catal., 2017, 7, 34–59 CrossRef CAS.
  21. Z. Li, D. Wang, Y. Wu and Y. Li, Natl. Sci. Rev., 2018, 5, 673–689 CrossRef.
  22. J. Wang, H. Y. Tan, S. Z. Yu and K. B. Zhou, ACS Catal., 2015, 5, 2873–2881 CrossRef CAS.
  23. L.-W. Guo, P.-P. Du, X.-P. Fu, C. Ma, J. Zeng, R. Si, Y.-Y. Huang, C.-J. Jia, Y.-W. Zhang and C.-H. Yan, Nat. Commun., 2016, 7, 13481 CrossRef CAS.
  24. Y. Hao, M. Mihaylov, E. Ivanova, K. Hadjiivanov, H. Knözinger and B. C. Gates, J. Catal., 2009, 261, 137–149 CrossRef CAS.
  25. D. Boyd, S. Golunski, G. R. Hearne, T. Magadzu, K. Mallick, M. C. Raphulu, A. Venugopal and M. S. Scurrell, Appl. Catal., A, 2005, 292, 76–81 CrossRef CAS.
  26. M. P. Casaletto, A. Longo, A. Martorana, A. Prestianni and A. M. Venezia, Surf. Interface Anal., 2006, 38, 215–218 CrossRef CAS.
  27. K. Qian, W. Huang, J. Fang, S. Lv, B. He, Z. Jiang and S. Wei, J. Catal., 2008, 255, 269–278 CrossRef CAS.
  28. M. Han, X. Wang, Y. Shen, C. Tang, G. Li and R. L. Smith, J. Phys. Chem. C, 2009, 114, 793–798 CrossRef.
  29. N. S. Phala, G. Klatt, E. van Steen, S. A. French, A. A. Sokol and C. R. A. Catlow, Phys. Chem. Chem. Phys., 2005, 7, 2440–2445 RSC.
  30. M. Manzoli, A. Chiorino and F. Boccuzzi, Appl. Catal., B, 2004, 52, 259–266 CrossRef CAS.
  31. M. Mihaylov, H. Knözinger, K. Hadjiivanov and B. C. Gates, Chem. Ing. Tech., 2007, 79, 795–806 CrossRef CAS.
  32. M. Mihaylov, B. C. Gates, J. C. Fierro-Gonzalez, K. Hadjiivanov and H. Knözinger, J. Phys. Chem. C, 2007, 111, 2548–2556 CrossRef CAS.
  33. Y. Wang, X. Xia, A. Urban, H. Qiu, J. Strunk, B. Meyer, M. Muhler and C. Wöll, Angew. Chem., Int. Ed., 2007, 46, 7315–7318 CrossRef CAS PubMed.
  34. H. Noei, C. Wöll, M. Muhler and Y. Wang, Appl. Catal., A, 2011, 391, 31–35 CrossRef CAS.
  35. X. P. Fu, L. W. Guo, W. W. Wang, C. Ma, C. J. Jia, K. Wu, R. Si, L. D. Sun and C. H. Yan, J. Am. Chem. Soc., 2019, 141, 4613–4623 CrossRef CAS PubMed.
  36. M. Mihaylov, E. Ivanova, Y. Hao, K. Hadjiivanov, H. Knözinger and B. C. Gates, J. Phys. Chem. C, 2008, 112, 18973–18983 CrossRef CAS.

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/c9cc08381b

This journal is © The Royal Society of Chemistry 2020