Hiroyuki
Asakura
*ab,
Naomi
Kawamura
c,
Masaichiro
Mizumaki
c,
Kiyofumi
Nitta
c,
Kenji
Ishii
d,
Saburo
Hosokawa
ab,
Kentaro
Teramura
ab and
Tsunehiro
Tanaka
*ab
aDepartment of Molecular Engineering, Graduate School of Engineering, Kyoto University, Kyotodaigaku Katsura, Nishikyo-ku, Kyoto 615-8510, Japan. E-mail: asakura@moleng.kyoto-u.ac.jp; tanakat@moleng.kyoto-u.ac.jp
bElements Strategy Initiative for Catalysts & Batteries (ESICB), Kyoto University, 1-30 Goryo-Ohara, Nishikyo-ku, Kyoto 615-8245, Japan
cJapan Synchrotron Radiation Research Institute, SPring-8, 1-1-1 Kouto, Sayo, Hyogo 679-5198, Japan
dSynchrotron Radiation Research Center, National Institutes for Quantum and Radiological Science and Technology, 1-1-1 Kouto, Sayo, Hyogo 679-5148, Japan
First published on 27th October 2017
A feasibility study of a “range-extended” EXAFS measurement was conducted on the Pt L3-edge of Pt/Al2O3 in the presence of Au2O3. The Pt L3-edge EXAFS spectrum of the Pt/Al2O3 sample was successfully obtained without any interference at the Au L3-edge by means of high energy resolution fluorescence detection (HERFD) of the Pt Lα1 line with a Bragg type X-ray emission analyser crystal. A newly observed significant enhancement of the EXAFS oscillation amplitude was discussed on the points of the core-hole lifetime width and X-ray energy width. We also successfully obtained reasonable physical parameters by a conventional EXAFS curve fitting on the HERFD-EXAFS data. The HERFD-EXAFS technique can broaden the scope of application of EXAFS spectroscopy for complicated samples with many different elements.
In this paper, we evaluate the feasibility of the range-extended EXAFS technique to obtain a Pt L3-edge EXAFS spectrum of Pt/Al2O3 in the presence of Au, which has the L3-edge X-ray absorption edge (11919 eV) close to the Pt L3-edge (11564 eV). For example, Kaito et al. reported a mixed material of Pt and Au for the oxygen reduction reaction of a fuel cell. Because of the apparent interruption by the Au L3-edge feature in the Pt L3-edge EXAFS, they used a Pt K-edge and Au K-edge XAS measurement to investigate the local structure.8 However, such high-energy X-rays are only available at a limited number of synchrotron radiation facilities.
Our initial interest in the application of “range-extended” EXAFS measurement is analyses on the local structure of Pt species of dilute Pt catalysts in the presence of a W element and the local structure of Pt and Au atoms of Pt–Au clusters. In such cases, high energy XAS measurement at the Pt K-edge (78395 eV), W K-edge (69525 eV), and Au K-edge (80725 eV) must be very difficult because of very weak X-ray fluorescence intensity. In addition, common Pt L3-edge (11564 eV) measurement suffers from the existence of the W L2-edge (11544 eV) or Au L3-edge (11919 eV). Thus, we first conducted a feasibility study of the “range-extended” EXAFS measurement on a Pt–Au mixture and examined the applicability of the conventional XAS analysis technique including EXAFS curve fitting analysis.
Hereafter, for clarity in this paper, we call the fluorescence mode measurement with a Pilatus detector as the total fluorescence yield (TFY) mode, the fluorescence mode measurement with an SDD as the partial fluorescence yield (PFY) mode, and the fluorescence mode measurement with an analyser crystal using a Pilatus detector as the HERFD mode. Data reduction was carried out with the Athena and Artemis ver. 0.9.25 and FEFF 6L,11 which is freely distributed and included in the Demeter package.12 The back-scattering amplitude or phase shift parameters were simulated with FEFF 6L and used to perform the curve fitting procedure on the obtained EXAFS spectra, as will be discussed in the following section.
Fig. 1 shows the Pt L3-edge XANES spectra of 0.1 wt% Pt/Al2O3 in the TFY and HERFD modes. The energy resolution of the Pt L3-edge XANES spectrum measured in the HERFD mode is astonishingly improved. To the best of our knowledge, this phenomenon was first reported by Hämäläinen et al. in 1991 for a few XANES spectra of Dy compounds14 based on the discussion by Tulkki and Aberg.2 This impressive improvement in the energy resolution can be understood because of the difference in the overall energy resolution in the TFY and HERFD modes. In general, the energy resolution of the XAS spectra is determined by the light source properties, instrumental properties (e.g., stability, aperture of the slits, and perfection and parallelism of the double crystal monochromator), and inherit physical phenomena (e.g., lifetime broadening effect15). In the present case, the energy resolution of the incident X-ray at the BL39XU was estimated from the full width half maximum of the X-ray elastic scattering spectrum at around 9.5 keV and found to be ca. 0.8 eV. The energy bandwidth accepted by the Pilatus detector is much wider than the natural energy width of both the L3-edge and the M5-edge. However, the analyser crystal used in the HERFD mode enables us to separate the Pt Lα1 (L3–M5, 9442 eV) fluorescence X-ray from the Lα2 line (L3–M4, 9362 eV). Based on the discussion of the energy resolution of the Pb L3-edge HERFD XANES by Swarbrick et al.,15 the overall energy resolution of the Pt L3-edge HERFD XAS can also be estimated to be 2.9 eV, which is much smaller than that in the TFY mode (5.2 eV), from the core-hole lifetime of the initial/intermediate and final states, along with the energy bandwidth of the incident X-ray and emitted X-ray, that is, the Pt L3 state (5.1 eV) and the M5 state (2.4 eV) for the Lα1 line (L3–M5), 0.8 eV (incident X-ray) and 1.5 eV (emitted X-ray), respectively, with the convolution formula given by de Groot et al.16 and Olivero et al.17 Therefore, as also discussed by Hämäläinen et al.,14 the energy resolution in the TFY mode was restrained to the core-hole lifetime of the L3 state; however, in the HERFD mode, it was only restrained to one narrower than that of the M5 state, but not the L3 state.
The high-resolution XANES measurement has already been utilized in various scientific fields, such as environmental science,15 biochemistry,18 and the fundamental science of the molecular orbitals of Os19 and La.20 This is a very interesting topic, but in this paper, we focus on the EXAFS region of the XAS spectra measured in the HERFD mode as follows.
The comparison between the Pt L3-edge EXAFS oscillations of 0.1 wt% Pt/Al2O3 measured in the TFY and HERFD modes is shown in Fig. 2. Even though the X-ray flux density at the sample position is considerably high (i.e., 1012 photons per s), with access to a highly brilliant X-ray undulator at SPring-8, a small solid angle and imperfection of the analyser crystal decreased the total fluorescence X-ray reaching the X-ray detector. The EXAFS oscillation is extracted from the averaged XAS spectra from six measurements (ca. 1 h per spectrum). The two EXAFS oscillations are fundamentally equal. However, the amplitude of the EXAFS oscillation measured in the HERFD mode is significantly larger than that in the TFY mode. The amplitude enhancement of the EXAFS oscillation might be ascribed to the difference in the overall energy resolution, as discussed above regarding the high-resolution XANES spectra. Thus, the enhancement of the EXAFS oscillation amplitude can be attributed to the relatively narrow natural width at the M5 state.
Then, we tried to perform quantitative analysis on the EXAFS spectra by using a curve fitting technique from a practical point of view. First, we obtained an amplitude reduction factor (S02) from EXAFS curve fitting analysis of the first shell of the Pt foil and found it to be 0.821.
Next, we performed a curve fitting analysis on the first shell of the 0.1 wt% Pt/Al2O3 EXAFS spectrum measured in the TFY mode, as summarized in Table 1. The coordination number and atomic distance of the Pt–O shell were found to be 6.4 ± 1.4 and 2.00 ± 0.01 Å, respectively, which are reasonable for the generation of β-PtO2-like species on Al2O3. This result implies two facts: the XAS measurements are performed well and the amplitude reduction factor is chemically transferable. Third, we moved to the 0.1 wt% Pt/Al2O3 EXAFS spectrum measured in the HERFD mode. The first shell of the Fourier transformed EXAFS spectrum looks significantly higher than that of the one measured in the TFY mode (Fig. S2†). Then, we conducted EXAFS curve fitting analysis on the 0.1 wt% Pt/Al2O3 HERFD EXAFS spectrum with a fixed coordination number of the Pt–O first shell equal to 6.4. Finally, we found the amplitude reduction factor for the HERFD mode to be 1.00. The fitting result is provided in the ESI (Fig. S3†). All the fitting procedures had an R factor below 0.003, indicating the goodness of the fit.
CN | R/Å | DW/Å2 | ΔE/eV | |
---|---|---|---|---|
a CN: coordination number, R: atomic distance, DW: XAFS Debye–Waller factor, and ΔE: energy correction during the fitting procedure. | ||||
Pt–O | 6.4 ± 1.4 | 2.00 ± 0.01 | 0.0024 ± 0.0017 | 11.1 ± 2.9 |
For example, the k3-weighted EXAFS spectra at the Mn K-edge of PS II samples in the conventional PFY and HERFD modes reported by Yano et al. appear to be almost identical to each other in the k range up to the Fe K-edge derived from the Fe concomitant.4 This result can be explained in the same manner. The natural widths of the Mn K-edge and L2-edge for the Kα2 (K–L2) line, and the L3-edge for the Kα1 (K–L3) line are 1.2, 1.5, and 0.42 eV, respectively. Based on the aforementioned Swarbrick paper,15 the overall lifetime broadening effect on the Mn K-edge HERFD EXAFS measurement was estimated to be 1.3 eV, which is close to the energy resolution of the overall energy bandwidth of the X-ray. Therefore, the amplitudes of the EXAFS oscillation, in both conventional PFY and HERFD modes, were identical to each other, which differs from our case. In our case, the natural widths of the Pt L3-edge and M4-edge for the Lα2 (L3–M4) line, and the M5-edge for the Lα1 (L3–M5) line are 5.1, 3.1, and 2.4 eV, respectively. Then, the overall lifetime broadening effect on the Pt L3-edge HERFD EXAFS measurement was estimated to be 2.9 eV, which is significantly smaller than that of the transmission mode (5.2 eV). At present, we assume that the difference in the amplitude reduction factors between the TFY and HERFD modes comes from the total energy resolution and is mainly dependent on the overall core-hole lifetime broadening discussed above. Indeed, the HERFD EXAFS spectrum of the 0.1 wt% Pt/Al2O3 convoluted using a lifetime broadening Lorentz function (γ = 5.0 eV) is identical to the EXAFS spectrum measured in TFY mode (Fig. S4†). However, the difference might arise from other factors, such as the difference in the many-body effect on the second-order optical process in the HERFD method.
Pt L3-edge XAS spectra of 0.1 wt% Pt/Al2O3 in the presence of Au2O3 in the PFY mode and in the HERFD mode as well as their magnified view around the Au L3-edge are shown in Fig. 3. As clearly seen in the magnified view (Fig. 3 (bottom)), the Au L3-edge absorption edge of Au2O3 was found at around 11920 eV in the EXAFS region of the Pt L3-edge of 0.1 wt% Pt/Al2O3. Then, the extracted Pt L3-edge EXAFS oscillation measured in the conventional PFY mode with an SDD is severely distorted by the interference of the Au L3-edge (k = 9.6 Å−1), as shown in Fig. 4 (top), because of the relatively low energy resolution of the SDD and the spectral tails of Au L lines. In contrast, there is no interference from the Au L3-edge on the EXAFS oscillation measured in the HERFD mode. This result demonstrates that the HERFD method enables us to perform EXAFS analysis on a target element even if some other absorption edges are in the EXAFS region of interest. Indeed, the Pt L3-edge EXAFS oscillations of both 0.1 wt% Pt/Al2O3 and its mixture with Au2O3 look essentially identical to each other, as shown in Fig. 4 (bottom).
Fig. 3 Pt L3-edge XAS spectra of 0.1 wt% Pt/Al2O3 in the presence of Au2O3 in the conventional PFY mode and in the HERFD mode (top), and their magnified view around the Au L3-edge (bottom). |
The X-ray emission spectra of the Pt Lα1 line of PtO2 and 0.1 wt% Pt/Al2O3 are shown in Fig. 5. The electronic structure of the Pt atom in these samples can be different because of the smallness of the PtO2-like species on Al2O3. However, both XES are identical to each other, which means that the Pt Lα1 line is not sensitive to the chemical state of the Pt atoms. In addition, the FWHM of the Pt Lα1 line was about 7.3 eV. As mentioned before, the energy resolution of the incident X-ray and crystal analyser are 0.8 eV and 1.5 eV, respectively. Thus, the XAS measurement in the HERFD mode can be made versatile using chemically insensitive X-ray emission lines (at least in the Pt L3-edge study).
The HERFD-EXAFS method realizes EXAFS analysis on the arbitrary target element in the presence of other elements, which have their absorption edges close to that of the target element. We have already measured a Pt L3-edge EXAFS spectrum of some catalysts in the presence of W, which was our initial interest in the application of EXAFS spectroscopy to a mixture of several elements with their X-ray absorption edges close to each other. This result will be discussed in a separate paper.
Footnote |
† Electronic supplementary information (ESI) available: Pt L3-edge XAS spectra of 0.1 or 0.5 wt% Pt/Al2O3 in the transmission, TFY, or HERFD modes and a curve fitting result. See DOI: 10.1039/c7ja00309a |
This journal is © The Royal Society of Chemistry 2018 |