Probing Single Pt Atoms in Complex Intermetallic Al13Fe4 - Journal of

43 mins ago - Because novel catalytic materials may be obtained, it is worthwhile to study the effects of doping catalytic metals into the Al13Fe4 mat...
0 downloads 8 Views 2MB Size
Communication pubs.acs.org/JACS

Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Probing Single Pt Atoms in Complex Intermetallic Al13Fe4 Tsunetomo Yamada,*,† Takayuki Kojima,†,‡ Eiji Abe,§ Satoshi Kameoka,† Yumi Murakami,† Peter Gille,∥ and An Pang Tsai† †

Institute of Multidisciplinary Research for Advanced Materials (IMRAM), Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai 980-8577, Japan ‡ Frontier Research Institute for Interdisciplinary Sciences, Tohoku University, Aramaki aza Aoba 6-3, Aoba-ku, Sendai 980-8578, Japan § Department of Materials Science and Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan ∥ Department of Earth and Environmental Sciences, Ludwig Maximilians University Munich, Theresienstr. 41, Munich 80333, Germany S Supporting Information *

ABSTRACT: The atomic structure of a 0.2 atom % Ptdoped complex metallic alloy, monoclinic Al13Fe4, was investigated using a single crystal prepared by the Czochralski method. High-angle annular dark-field scanning transmission electron microscopy showed that the Pt atoms were dispersed as single atoms and substituted at Fe sites in Al13Fe4. Single-crystal X-ray structural analysis revealed that the Pt atoms preferentially substitute at Fe(1). Unlike those that have been reported, Pt single atoms in the surface layers showed lower activity and selectivity than those of Al2Pt and bulk Pt for propyne hydrogenation, indicating that the active state of a given single-atom Pt site is strongly dominated by the bonding to surrounding Al atoms.

Figure 1. Structure of monoclinic Al13Fe4 along the directions parallel to (a) the b axis and (b) the c axis. The structure consists of 102 atoms (78 Al atoms and 24 Fe atoms) in a C-centered monoclinic lattice with the lattice parameters a = 1.5489 nm, b = 0.80831 nm, c = 1.2476 nm, and β = 107.72° (space group C2/m).5 There are 20 sites, Al(1)− Al(15) and Fe(1)−Fe(5), that result from the stacking of flat (F) and puckered (P) atomic layers perpendicular to the b axis. The site labels in this figure correspond to those provided in ref 5.

C

omplex metallic alloys (CMAs)1,2 are a large family of materials having giant unit cells. The cell volumes in these alloys are typically in excess of several cubic nanometers and possess a large number of inequivalent sites. Recently, Armbrüster et al.3 reported the unexpectedly high catalytic performance of the CMA monoclinic Al13Fe4 during the semihydrogenation of acetylene. The structure of this material is depicted in Figure 1.4,5 The bulk structure is most likely preserved up to the topmost surface layer,6 and the exceptional activity of this alloy is attributed to a combination of site isolation and alternation of the electronic structure as a result of chemical binding.3 Because novel catalytic materials may be obtained, it is worthwhile to study the effects of doping catalytic metals into the Al13Fe4 matrix. Very recently, there was a report of Pt dispersed as single atoms and substituted at Fe sites in a 0.5 atom % Pt-doped Al13Fe4 phase, based on high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) observations.7 X-ray photoelectron spectroscopy (XPS) analyses have also suggested that the Pt atoms substitute at the Fe sites in Al13Fe4.7 However, the crystallographic sites of the Pt atoms could not be determined from the HAADF images since inequivalent Fe sites are overlapped in the two-dimensional projection. In contrast, X-ray diffraction (XRD) may represent a practical means of assessing the Pt sites © XXXX American Chemical Society

and their partial occupancies because the substitutional element Pt, whose atomic number is Z = 78, exhibits good contrast to both Al (Z = 13) and Fe (Z = 26) for X-ray scattering. Therefore, by the combination of HAADF-STEM observations with X-ray structural determination, it is possible to quantitatively characterize the Pt atoms in Al13Fe4 at the atomic level. Herein we report the results of a combined study using HAADF-STEM observation and single-crystal X-ray structural analysis based on the same single crystal of 0.2 atom % Ptdoped Al13Fe4. The HAADF-STEM images indicate that the Pt atoms are dispersed as single atoms and substitute Fe(1) and/ or Fe(2) in the sample, while the X-ray structural analysis reveals that the Pt atoms preferentially substitute at Fe(1) and determine the site occupancy factor (SOF). Finally, we show Received: December 25, 2017

A

DOI: 10.1021/jacs.7b13658 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Communication

Journal of the American Chemical Society

projected positions of the Fe atoms at Fe(3) and Fe(4) are not overlapped, and herein these are denoted as wFe. These positions thus result in spots with different intensities, as can be seen in a simulated image (Figure S2c). Comparison of these figures shows that the positions of the spots in the experimental image evidently correspond to the mFe, wFe, and sFe positions. Moreover, the intensity contrast of the spots is well-reproduced in the simulated image; the spots at mFe and sFe are brighter than that at wFe. On closer examination of the experimental HAADF image, it is evident that some spots are slightly brighter than the others, suggesting that Pt atoms are likely to be located at these positions. To find the preferential Fe site for the Pt atoms, we extracted the two-dimensional (2-D) intensity distribution at each mFe, wFe, and sFe from the HAADF image (Figures 2c and S2d,e), and the projected spot intensities are shown by the corresponding one-dimensional (1-D) profiles for each extraction image. It can clearly be seen that significant and discontinuous intensity increases indeed occur at several mFe positions, as denoted by the yellow and red arrowheads in Figure 2c,d, whereas the intensities are essentially constant at the wFe and sFe (Figure S2f,g). These results demonstrate that the Pt atoms do not form clusters (i.e., those composed of several atoms) but are well-dispersed as individual single atoms to substitute preferentially at Fe(1) and/or Fe(2). To determine the preferential sites of the doped Pt atoms in Al13Fe4 and their relative occupancies, we performed singlecrystal XRD intensity measurements. An initial structural model was obtained by ab initio phasing of the diffractions.9−11 The model was assumed to consist solely of Al and Fe atoms. The structure was found to be isostructural to monoclinic Al13Fe4,5 consisting of 20 sites, Al(1)−Al(15) and Fe(1)−Fe(5), with each site occupied by the respective element. After iterations of the refinement12 that included 20 fully occupied sites (refinement 1), difference Fourier syntheses yielded highresidual peaks, as shown in Figure 3. A residual electron peak value of 7.41 e/Å3 was located at Fe(1), demonstrating that the Pt atoms substituted preferentially at Fe(1) sites in Pt-doped Al13Fe4.

results on the catalytic performance of this alloy during propyne (C3H4) hydrogenation. Single-grain crystals of Pt-doped Al13Fe4 (Al76.5Fe23Pt0.5) and nondoped Al13Fe4 were prepared by the Czochralski method.8 Herein these are denoted as PAF and AF, respectively. The phase purity of each sample was assessed using powder XRD. The PAF and AP alloy samples were found to have the compositions Al75.8Fe24.1Pt0.2 and Al76.5Fe23.5, respectively, based on inductively coupled plasma (ICP) results. HAADF-STEM was used to examine the PAF crystal, and Figure 2a shows a HAADF image of PAF over a 15 nm × 11

Figure 2. (a) HAADF-STEM image of Pt-doped Al13Fe4 normal to the b axis. (b) Projected structure of Al13Fe4 along the b axis. Each Fe site is mostly overlapped with other sites at mFe, wFe, and sFe. (c) Extracted 2-D intensity distributions around the mFe in (a). (d) Intensity distribution of the spots in the area enclosed by the yellow lines in (c). The intensities are projected on a 1-D horizontal axis along the vertical direction of the 2-D image in (c). The yellow and red arrowheads in (c) and (d) indicate brilliant spots representing the positions of Pt atoms.

Figure 3. (left) Difference Fourier map of the flat (F) atomic layer in Pt-doped Al13Fe4 from refinement 1 and (right) the corresponding atomic positions.

nm region normal to the b axis. Since the spot intensity is approximately proportional to Z2 in such images, meaning that the Al/Fe intensity ratio will be 0.25, the spots in the image correspond to Fe atoms. When the structural model is projected along the b axis (Figure 2b), each Fe site is seen to overlap with other sites such that (i) the projected positions of two Fe atoms at Fe(1) and Fe(2) are overlapped and (ii) the projected positions of two Fe atoms at Fe(5) and a single Al atom at Al(2) are overlapped. Herein we abbreviate these positions as mFe and sFe, respectively. In contrast, the

In the further refinement (refinement 2), we placed a mixture of Pt and Fe at Fe(1) and refined the relative occupancy of Pt with several restrictions. First, the sum of the SOFs for Fe(1) and Pt(1) had to be equal to 1, and second, the Fe(1) and Pt(1) had common anisotropic atomic displacement parameters. The refined SOF of the Pt(1) was 0.0540(7), giving the composition Al76.47Fe23.32Pt0.21. Excellent agreement was obtained between the experimental and refined compositions. The maximum residual electron density was 2.073 e/Å3 (Figure B

DOI: 10.1021/jacs.7b13658 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Communication

Journal of the American Chemical Society

noted in a previous report7 were well-reproduced in this study (Figure S4), which leads to the same conclusion that the doped Pt atoms preferentially substituted at the Fe sites rather than the Al sites in Al13Fe4. Moreover, the Pt 3d5/2 peak position of Al2Pt is also shifted from that of bulk Pt in Figure 4b, so the chemical shift of Pt 3d5/2 in the case of PAF is attributed to electron transfer from Pt to Al atoms. This may be due to the strong binding of the Al−Pt−Al groups in Al13Fe4. Although the surface Pt environments will be different from those in the bulk, a correlation is seen when the areal reaction rate during the hydrogenation of C3H4 is plotted versus the chemical shift of Pt 3d5/2 (Figure 4c). Since a Pt atom is surrounded by only Al atoms in PAF and Al2Pt, it can be assumed that the chemical shifts of Pt atoms in the surface layer will be proportional to those in the bulk. Therefore, this result indicates that the active state of a given Pt atom is strongly dominated by the bonding to surrounding Al atoms. This means that the nature of the single-atom catalytic interactions with the intermetallic matrix is crucial, as it controls the electronic structure of the atom, its charge, the coordination pattern, and the overall catalytic ensemble. Recently, it has been reported that Pt single atoms on different supports, e.g., metal oxide nanocrystals15,16 and graphene,17,18 maximize the efficiency of precious-metal use.19,20 Unlike those reports, Pt single atoms in the Al13Fe4 surface show a lower reaction rate than those of Al2Pt and bulk Pt for the hydrogenation of C3H4. However, the present study provides quantitative information for the Pt single atoms, i.e., the preferential site and its SOF, which has been lacking in previous studies. Therefore, Al13Fe4 can be used as a model system to investigate single-atom catalysts. In summary, we have examined the CMA monoclinic Al13Fe4 as a new platform for single dispersion of the precious metal Pt. A combination of STEM and single-crystal XRD provided evidence that the doped Pt atoms are dispersed as single atoms and preferentially substituted at Fe(1) in Al13Fe4. The Pt atoms dispersed as single atoms in the surface layers enhance the catalytic performance during the hydrogenation of C3H4. Finally, we have shown that the active state of a given Pt single atom is strongly dominated by its surrounding Al atoms.

S3). This value is less than that observed in nondoped Al13Fe4, and we thus conclude that the doped Pt atoms substituted preferentially at Fe(1). The details are provided in the Supporting Information. It is not unexpected that the Pt atoms would preferentially substitute at Fe(1), where the volume of the Voronoi cell is the largest (Table S1), since the atomic radius of Pt (0.1387 nm) is larger than that of Fe (0.1274 nm).13 Moreover, Fe(1) has a different coordination number from the other Fe sites. Indeed, a recent study using 57Fe Mössbauer spectroscopy14 showed that the local environment of Fe(1) is different from that of the other Fe sites, confirming that Fe(1) has unique local environment. Therefore, the strong site preference of the doped Pt atoms observed in this study is related to the unique local configurations at Fe(1). Catalytic trials were conducted using PAF to promote the hydrogenation of C3H4. For comparison, the same trails were also conducted using AF, Al2Pt, and bulk Pt. Figure 4a shows

Figure 4. (a) Conversion of propyne as a function of temperature for Pt-doped and nondoped Al13Fe4. (b) Pt 3d5/2 XPS spectra acquired from a single crystal of Pt-doped Al13Fe4, a polycrystal of Al2Pt, and bulk Pt. (c) Correlation between the chemical shift of Pt 3d5/2 and the areal reaction rate during C3H4 hydrogenation.

data for the conversion of C3H4 between ambient temperature and 523 K. The conversion with PAF rapidly increased at approximately 400 K and reached around 13% at 523 K, whereas the conversion with AF increased monotonically to a maximum of 3% as the temperature was increased to 523 K. (The product selectivity of propene was larger than 80% during the reaction in both cases.) These results indicate that the Pt atoms exposed at the topmost surface layer enhanced the catalytic performance during hydrogenation of C3H4 by a factor of 4. Taking into account the fact that the Pt atoms are dispersed as single atoms in PAF, as evidenced by the HAADF observations, this result leads us to the conclusion that the Pt atoms in the surface layers are also dispersed as single atoms. The reaction rate of PAF per Pt atom is estimated to be approximately one-seventh that of bulk Pt on the basis of the refined SOFs for Fe(1)/Pt(1), e.g., 1.3 × 10−23, 1.1 × 10−23, and 8.8 × 10−23 molC3H4·min−1/Pt atom for PAF, Al2Pt, and bulk Pt, respectively, at 373 K. This result indicates that the character and catalytic interactions of the Pt atoms in the Al13Fe4 surface layers differ from those in others. XPS data were acquired using the BL15UX beamline at the SPring-8 facility (E = 5.950 keV, ΔE ≤ 250 meV). A Pt 3d5/2 peak was clearly generated by PAF (Figure 4b), providing evidence that the Pt was soluble in the sample. Characteristics



ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b13658. Methods, powder and single-crystal XRD data, and HAADF-STEM images (PDF) X-ray crystallographic data for Pt-doped Al13Fe4 (CIF)



AUTHOR INFORMATION

Corresponding Author

*[email protected] ORCID

Tsunetomo Yamada: 0000-0003-0138-9778 Takayuki Kojima: 0000-0002-9162-2758 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors thank Dr. Y. Matsushita and Mr. A. Sato for assistance during single-crystal XRD experiments at the C

DOI: 10.1021/jacs.7b13658 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Communication

Journal of the American Chemical Society National Institute for Materials Science (NIMS). The XPS experiments were performed at the SPring-8 facility (Project 2016B4907). A.P.T. acknowledges support from the Dynamic Alliance for Open Innovation Bridging Human, Environment and Materials.



REFERENCES

(1) Urban, K.; Feuerbacher, M. J. Non-Cryst. Solids 2004, 334-335, 143−150. (2) Complex Metallic Alloys: Fundamentals and Applications; Dubois, J.-M., Belin-Ferré, E., Eds.; Wiley-VCH: Weinheim, Germany, 2010. (3) Armbrüster, M.; Kovnir, K.; Friedrich, M.; Teschner, D.; Wowsnick, G.; Hahne, M.; Gille, P.; Szentmiklosi, L.; Feuerbacher, M.; Heggen, M.; Girgsdies, F.; Rosenthal, D.; Schlögl, R.; Grin, Y. Nat. Mater. 2012, 11, 690−693. (4) Black, P. J. Acta Crystallogr. 1955, 8, 43−48. (5) Grin, J.; Burkhardt, U.; Ellner, M.; Peters, K. Z. Kristallogr. - Cryst. Mater. 1994, 209, 479−487. (6) Ledieu, J.; Gaudry, É.; Loli, L. S.; Villaseca, S. A.; de Weerd, M. C.; Hahne, M.; Gille, P.; Grin, Y.; Dubois, J.-M.; Fournée, V. Phys. Rev. Lett. 2013, 110, 076102. (7) Kameoka, S.; Wakabayashi, S.; Abe, E.; Tsai, A. P. Catal. Lett. 2016, 146, 1309−1316. (8) Gille, P.; Bauer, B. Cryst. Res. Technol. 2008, 43, 1161−1167. (9) Palatinus, L.; Chapuis, G. J. Appl. Crystallogr. 2007, 40, 786−790. (10) Oszlányi, G.; Sütő , A. Acta Crystallogr., Sect. A: Found. Crystallogr. 2004, 60, 134−141. (11) Oszlányi, G.; Sütő , A. Acta Crystallogr., Sect. A: Found. Crystallogr. 2005, 61, 147−152. (12) Sheldrick, G. M. Acta Crystallogr., Sect. C: Struct. Chem. 2015, 71, 3−8. (13) Pearson, W. B. The Crystal Chemistry and Physics of Metals and Alloys; Wiley: New York, 1972. (14) Albedah, M. A.; Nejadsattari, F.; Stadnik, Z. M.; Przewoźnik, J. J. Alloys Compd. 2015, 619, 839−845. (15) Qiao, B.; Wang, A.; Yang, X.; Allard, L. F.; Jiang, Z.; Cui, Y.; Liu, J.; Li, J.; Zhang, T. Nat. Chem. 2011, 3, 634−641. (16) Wei, H.; Liu, X.; Wang, A.; Zhang, L.; Qiao, B.; Yang, X.; Huang, Y.; Miao, S.; Liu, J.; Zhang, T. Nat. Commun. 2014, 5, 5634. (17) Sun, S.; Zhang, G.; Gauquelin, N.; Chen, N.; Zhou, J.; Yang, S.; Chen, W.; Meng, X.; Geng, D.; Banis, M. N.; Li, R.; Ye, S.; Knights, S.; Botton, G. A.; Sham, T.-K.; Sun, X. Sci. Rep. 2013, 3, 1775. (18) Yan, H.; Cheng, H.; Yi, H.; Lin, Y.; Yao, T.; Wang, C.; Li, J.; Wei, S.; Lu, J. J. Am. Chem. Soc. 2015, 137, 10484−10487. (19) Yang, X. F.; Wang, A.; Qiao, B.; Li, J.; Liu, J.; Zhang, T. Acc. Chem. Res. 2013, 46, 1740−1748. (20) Liu, J. ACS Catal. 2017, 7, 34−59.

D

DOI: 10.1021/jacs.7b13658 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX