Exploring Platinum Speciation with X-ray Absorption Spectroscopy under High-Energy Resolution Fluorescence Detection Mode

Abstract

Critical to interpreting platinum chemical speciation using X-ray absorption spectroscopy (XAS) is the availability of reference spectra of compounds with known Pt redox and coordination. Here we compare different techniques for Pt L_III-edge X-ray absorption near edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) spectral regions for a large set of Pt-O-Cl-S reference compounds of known structures. The measurements were conducted in HERFD (high-energy resolution fluorescence detection, high-resolution or HR) mode, as well as in two conventional modes such as transmission (TR) and nominal-resolution total fluorescence yield (TFY or NR). Samples analyzed here included Pt⁰ (TR), Pt^IIS (HR), Pt^IVS₂ (TR), K₂Pt^IICl₄ (HR + TR), K₂Pt^IVCl₆ (HR + TR), Pt^IVO₂ (HR + TR), C₆H₁₂N₂O₄Pt^II (HR + TR), and aqueous solutions of K₂Pt^IICl₄ and H₂Pt^IVCl₆ (NR + TR), as well as (NH₄)₂Pt^IV(S₅)₃ (HR + TR). XANES spectra in HERFD mode offer a better energy resolution than in conventional modes, allowing a more accurate identification of Pt redox state and coordination geometry. EXAFS spectra in all three modes for a given compound yield identical within errors values of Pt-neighbor interatomic distances and mean square relative displacement (MSRD, σ²) parameters. In contrast, both TR and NR spectra on the one hand and HR spectra on the other hand yield distinct amplitude reduction factor (S₀²) values, 0.76 ± 0.04 and 0.99 ± 0.07 (1 standard error), respectively. This study contributes to the development of an open-access XAS database SSHADE.

1. Introduction

Platinum forms a variety of minerals (e.g., native platinum, cooperite PtS, sudovikovite PtSe₂, sperrylite PtAs₂) and synthetic phases (e.g., PtO₂, cisplatin Pt(NH₃)₂Cl₂, carboplatin C₆H₁₂N₂O₄Pt, K₂PtCl₄, K₂PtCl₆) largely used in industry [1,2,3] and medicine [4,5,6]. In these solids, platinum commonly exists in three formal oxidation states (0, II, and IV; [7]), forming a variety of structures, from square-planar to dodecahedral (e.g., Figure 1). Platinum, being a trace metal in nature (0.8 ppb in the Earth’s upper continental crust; [8]), also comes as isomorphic substitutions in major minerals (e.g., pyrite, chalcopyrite, pyrrhotite). In natural aqueous fluids and silicate/sulfide melts, platinum forms complexes with the major inorganic ligands (O/OH, Cl, S) [9,10,11,12,13,14,15,16,17,18]. Therefore, knowledge of the Pt redox state and detailed structural environment at the atomic scale in geological materials is required both for understanding the behavior of platinum in natural systems and optimizing its use in numerous technological applications.

This knowledge can be provided by X-ray spectroscopy at synchrotron radiation facilities, which is the most direct probe of Pt coordination, oxidation state, and local atomic environment. X-ray absorption spectroscopy (XAS) can be recorded in transmission and in fluorescence mode as well. An XAS spectrum includes X-ray absorption near edge (XANES) and extended X-ray absorption fine structure (EXAFS) parts. The XANES part provides information about the element oxidation state and coordination geometry. The EXAFS part, being complementary to XANES, helps to identify the ligands around the absorbing element and to quantify their numbers and interatomic distances in chemical complexes formed by the absorbing element in solid, gaseous, liquid, or aqueous phases [19,20,21,22,23]. When applied to low Pt concentrations, conventional fluorescence measurements using solid-state detectors have been widely used to study Pt in its different phases and complexes using both XANES and EXAFS analyses (e.g., [24,25,26,27]). However, conventional fluorescence- as well as transmission-mode XAS [16,28,29,30] are rather limited as to the discrimination between Pt^II and Pt^IV, or among different atomic neighbors (e.g., S vs. Cl, O vs. N) in similar types of solids (e.g., Pt^IIS vs. Pt^IVS₂) or aqueous complexes (e.g., Pt(HS)_n vs. PtCl_n). Compared to conventional fluorescence, HERFD was commonly applied only to XANES analyses, for which it presents numerous benefits: (i) it enables a far better XANES shape definition and may reveal pre-edge and post-edge features that are poorly resolved (if detectable at all) in conventional mode (e.g., [31,32,33,34,35,36]); (ii) it is more sensitive to the presence of light elements whose presence affects HERFD-XANES spectra (e.g., H in Pt catalytic materials; [26]) and more discriminative for neighbors with close atomic numbers or monomeric vs. polymeric ligands of the same element (e.g., HS⁻ vs. S₃⁻; [18,37]); (iii) it allows elimination of unwanted fluorescence signals from other elements with edges close to that of Pt [38], such as tungsten (W-L_II edge; 11,544 eV), gold (Au L_III-edge; 11,919 eV), arsenic (As-K edge; 11,867 eV), tantalum (Ta-L_I edge; 11,682 eV), or rhenium (Re-L_II edge; 11,959 eV). To better estimate the accuracy and limitations of the HERFD-EXAFS, especially for Pt-bearing complexes and materials cited above for which conventional XAS is not sensitive enough, additional studies are required. However, the studies on Pt compounds using HERFD-XAS have so far been mostly devoted to catalytic materials [39,40,41,42,43,44,45,46,47,48,49,50] and remain virtually non-existent for minerals and other geological samples [51]. Furthermore, a straightforward interpretation of HERFD-XAS data requires spectra of reference compounds with a known Pt atomic environment acquired at the same spectral resolution.

The goal of this study is to systematically analyze the XAS spectra of high-quality reference compounds of Pt solid phases of known crystal structures using both high-resolution (HR) and conventional XAS (transmission—TR, and nominal-resolution total fluorescence yield, TFY—NR) in order to compare both techniques and identify their advantages and limitations for studying Pt. The following solid references were analyzed in this study: platinum metal (Pt⁰), platinum(II) sulfide (Pt^IIS), platinum(IV) disulfide (Pt^IVS₂), potassium tetrachloroplatinate (II) (K₂Pt^IICl₄), potassium hexachloroplatinate (IV) (K₂Pt^IVCl₆), platinum (IV) dioxide (Pt^IVO₂), and diamine (cyclobutane-1,1-dicarboxylato(2-)-O,O′) platinum (II) (C₆H₁₂N₂O₄Pt^II). Substantial work was made to synthesize pure references of Pt^IIS and Pt^IVS₂ in this study. Additionally, ammonium tris(pentasulfido) platinum(IV), (NH₄)₂Pt^IV(S₅)₃), representative of Pt^IV in a sulfur-coordinated first and second shell environment, was also synthetized and thoroughly characterized to be examined by XAS using the approaches established on the reference compounds. In particular, we attempted to assess the potential of HR-XAS in detecting second shell sulfur atoms in polysulfide complexes that may be important carriers of chalcophile metals in hydrothermal fluids [18,52,53,54]. The Pt-S solids synthesized and analyzed in this study are particularly important for HERFD-XAS analyses of Pt in natural sulfide minerals of hydrothermal and magmatic origin that represent the PGE major economic resource (e.g., [55,56,57,58]). Our results provide a systematic assessment of the advantages and limitations of HR-XAS for the study of platinum species and contribute to spectroscopic databases of Pt compounds such as the SSHADE database [59] that will help future synchrotron-based studies of this economically critical metal. SSHADE is an open-access spectroscopy database infrastructure (https://www.sshade.eu/ (accessed on 9 December 2022)) created in 2018 and currently gathering 5230 spectra (including XAS, Raman, and FTIR), with 580 XAS spectra for natural and synthetic solids and solutions, acquired and thoroughly described with an exhaustive set of metadata by 20 different groups of users of BM16 and BM30 CRG beamlines at the ESRF.

2. Materials and Methods

2.1. Experimental Samples

2.1.1. Origin of the Investigated Compounds and Solutions

The following Pt-bearing solids were purchased from chemical suppliers and used without further treatment: potassium platinum (II) chloride (K₂PtCl₄, Pt 46.0%, trace metal basis 99.9%, serial number #61601032, ThermoFisher), potassium platinum (IV) chloride (K₂PtCl₆, Pt 39.6%, #61600659, ThermoFisher, Kandel, Germany), platinum (IV) dioxide (PtO₂, Pt 84.4% min, trace metal basis >99.95%, #R12B020, ThermoFisher, Kandel, Germany), platinum (IV) disulfide (PtS₂, Pt 74.8% min, #R22D024, ThermoFisher, Kandel, Germany), and diamine (cyclobutane-1,1-dicarboxylato(2-)-O,O′) platinum “carboplatin” (C₆H₁₂N₂O₄Pt, purity ≥95%, Cayman Chemical, Ann Arbor, MI, USA). Powder X-ray diffraction (XRD) patterns of K₂PtCl₄, K₂PtCl₆, and PtO₂ corresponded to the pure crystalline compounds, but that of PtS₂ was highly problematic (see below). Therefore, the PtS₂ solid along with PtS and (NH₄)₂Pt(S₅)₃ that were commercially unavailable, has been synthesized in this study. The following chemicals were used: platinum black powder (≤20 µm, ≥99.97%, surface area ≥25 m²/g, Aldrich, Saint-Louis, MO, USA), platinum metal powder (>99.9%, 325 mesh, ThermoFisher, Kandel, Germany), aqueous diammonium sulfide solution ((NH₄)₂S, 40–44 wt%), sulfur powder (99.98%, Aldrich, Saint-Louis, MO, USA), chloroplatinic acid (H₂PtCl₆, 0.38 wt% Pt, Aldrich, Saint-Louis, MO, USA), toluene, methanol, N,N-dimethylformamide (DMF). The synthesis and characterization of the platinum sulfide solids are described in the following subsections. Aqueous solutions of K₂Pt^IICl₄ (0.015 mol/kg soln Pt, hereafter K₂PtCl₄ aq) and H₂Pt^IVCl₆ (0.007 mol/kg soln Pt, hereafter H₂PtCl₆ aq) were prepared by dissolution, respectively, of K₂PtCl_4(s) and H₂PtCl₆ in water, and adding NaCl and HCl to stabilize Pt in solution.

Table 1. The solid and aqueous reference compounds analyzed by XAS in this study.

Solid	Origin	Beamline Used	Mode	Concentrations, wt% Pt
Pt metal	GoodFellow (10 µm thickness, 99.95%)	FAME FAME-UHD	TR TR	>99.9 >99.9
PtS-h PtS-s	This study (hydrothermal) This study (solid-state)	FAME-UHD FAME	HR + TR TR	1.1 3.6
PtS2-s PtS2-c *	This study (solid-state) ThermoFisher (not pure)	FAME FAME-UHD	TR HR + TR	3.5 1.5
(NH4)2PtIV(S5)3	This study	FAME-UHD	HR + TR	1.4
K2PtCl4-a	ThermoFisher (99.9%)	FAME-UHD	HR + TR	0.9
K2PtCl4-b	ThermoFisher (99.9%)	FAME	TR	3.1
K2PtCl4 aq	This study	FAME	NR + TR	0.3
K2PtCl6-a	ThermoFisher	FAME-UHD	HR + TR	0.9
K2PtCl6-b	ThermoFisher	FAME	TR	3.0
H2PtCl6 aq	This study	FAME	NR + TR	0.1
C6H12N2O4Pt	Cayman Chemical (≥95%)	FAME-UHD	HR + TR	1.1
PtO2-a	ThermoFisher (99.95%)	FAME-UHD	HR + TR	1.1
PtO2-b	ThermoFisher (99.95%)	FAME	TR	3.7

* mixture of amorphous PtS and elementary sulfur; wt%, weight percent.

summarizes the investigated samples, their origin, and the XAS acquisition modes applied.

2.1.2. Synthesis of PtS and PtS₂

The platinum(II) sulfide (PtS) solid has been reported to be synthesized either hydrothermally at acidic pH at 450 °C from a mixture of elemental Pt and S [11,15] or by the solid-state method in silica tubes at 750 °C [60] and 1000 °C [10]. Platinum (IV) disulfide (PtS₂) has been hydrothermally synthesized at ~100 °C at neutral pH [11], in solid-state in silica tubes at 800 °C [61] and 1000 °C [10], and in a flow of H₂S with (NH₄)₂PtCl₆ as the reactant at 130 °C [62]. In this study, we used both hydrothermal and solid-state syntheses. Hydrothermal PtS (hereafter PtS-h) was obtained from stoichiometric quantities of platinum black powder and sulfur powder in 0.1 mol/L NaOH in deionized water placed in a titanium (grade alloy VT-8) reactor at 450 ± 5 °C and 500 ± 20 bar for 11 days [15]. X-ray diffraction patterns of the solid show a high crystallinity of the corresponding pure phase (Figure A1d). Solid-state PtS (hereafter PtS-s) was obtained by annealing of a slightly non-stoichiometric mixture of elemental platinum and sulfur powders (S:Pt ratio of 1.1 to compensate for the partial sulfur loss in the vapor phase; [63]). The mixture was compacted at the bottom of a fused silica tube sealed under argon atmosphere and placed in a horizontal furnace at 750 °C for 10 days. After the experiment, the tube was cooled down to room temperature at a rate of 100 °C/h by keeping it in the furnace switched off. Platinum disulfide PtS₂ was synthesized by the same solid-state method (with a S:Pt ratio of 2.3:1.0). Note that, compared to previous studies [10,61], our chosen temperature is lower, thereby significantly reducing the risk of the silica tube break due to the pressure of S₂ (gas) that strongly increases with increasing temperature (e.g., 28 bar at 750 °C versus 40 bar at 800 °C; [63]). Scanning electron microscopy (SEM) pictures of the obtained PtS and PtS₂, collected using a Tescan Vega 4 microscope in secondary electron mode, show small crystals (2–10 µm; Figure 2a,b). X-ray diffraction patterns of both solids show a high crystallinity of the corresponding phase (Figure A1b,c), and small amounts (10 ± 5%) of PtS₂ in the PtS solid. This impurity, as quantified by a Rietveld refinement (using a pseudo-Voigt profile function and a zero-shift displacement), is likely due to minor sulfur excess in the synthesis.

2.1.3. Commercial PtS₂

Commercially purchased platinum disulfide (see Section 2.1.1, hereafter PtS₂-c) was analyzed to check for its purity and crystallinity. It was found to exhibit the XRD pattern of an amorphous compound (Figure A1e) that looked suspiciously similar to the XRD pattern of amorphous sulfur [64]. The S:Pt ratio in this solid was determined by dissolution: 25 mg of PtS₂-c powder was dispersed in 5 mL of hot aqua regia (~70 °C; 2/3 HCl of 37 wt% + 1/3 HNO₃ of 69 wt%) for 3 h to achieve complete dissolution; then aqua regia was evaporated to a wet residue (<0.3 g) and diluted with an acidic aqueous solution (HCl 0.5 wt% − HNO₃ 1.5 wt%), which was analyzed by Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES). The obtained S:Pt molal concentration ratio of 3 (vs. the stoichiometric ratio 2) clearly indicates that the “PtS₂“ solid as claimed by the chemical company is in reality highly non-stoichiometric and likely consists of amorphous PtS with a large fraction of elemental sulfur.

2.1.4. Synthesis of (NH₄)₂Pt^IV(S₅)₃

The synthesis of the (NH₄)₂Pt^IV(S₅)₃ solid has been a subject of several studies using different protocols of variable complexity [65,66,67,68]. We have adopted the method of [68], which is the most straightforward way of avoiding the use of highly toxic carbon disulfide CS₂. Our synthesis included the following steps. First, polysulfide anions were generated by saturating 5 mL of (NH₄)₂S (40–44 wt%) aqueous solution with an excess of powder sulfur (of the 3 g of sulfur added, ~2 g were dissolved after 10–15 min). This sulfur-saturated solution contains a mixture of polysulfides among which the pentasulfide anion S₅²⁻ is the most abundant one [69]. The remaining solid sulfur was removed by filtration. Then, to the produced polysulfide solution, 10 mL of H₂PtCl₆ (0.0198 mol/kg soln) aqueous solution was added dropwise while vigorously stirring. Assuming mainly the pentasulfide ion was formed, approximately 100 moles of S₅²⁻ were available per 1 mole of platinum. The resulting mixture was quickly filtered and the filtrate was stored at 5 °C for 24 h. The solution was then centrifuged (3800 rpm for 15 min). Finally, the obtained brick-red to maroon-colored precipitate was washed twice with a small amount of cold water and purified three times with a few mL of methanol (CH₃OH) and toluene (C₇H₈) to remove the remaining sulfur. The obtained product was re-dispersed in 10 mL of N,N-dimethylformamide (DMF). The DMF solvent was evaporated for 1 day at 70 °C, while letting the (NH₄)₂Pt^IV(S₅)₃ precipitate in the condensed phase to avoid its potential volatilization or decomposition that may occur when using more elevated temperatures (>100–150 °C; [70]).

Scanning electron microscopy (SEM) coupled with energy dispersive X-ray spectroscopy (EDS) was also used to quantify the amount of sulfur and platinum in the synthesized solid (Figure 2c) using a Bruker Nano GmbH microscope with a Quantax EDS unit. An atomic ratio of S:Pt = 14.4 was found, which is close to the stoichiometric S:Pt ratio of 15.0 in this compound. The X-ray diffraction (XRD) analysis of the produced precipitate showed an amorphous pattern (Figure A1a). The XRD pattern shows the presence of crystalline impurities of ammonium sulfate, (NH₄)₂SO₄, and elemental sulfur (Figure A1a). An infrared spectrum for this powder sample was collected using a Nicolet 6700 Thermoscientific spectrometer, in the range 100–600 cm⁻¹, with a resolution of 8 cm⁻¹, and an attenuated total reflectance (ATR) iTX^TM for a total of 16 scans (Figure A2a). Despite the small amount of available solid and the resulting weak signal, the spectrum shows 4 peaks (3 peaks at 551, 487, and 462 cm⁻¹ and a broad feature at ~278 cm⁻¹;

Table 2. Major absorbance frequency positions in the far-infrared (IR) and the UV-visible spectrum of (NH₄)₂Pt^IV(S₅)₃ obtained in this study and its comparison with literature values.

Infrared	Frequency, cm−1
	278	461 w	487 w	551 s	This study
	286 s 294	450 w	490 w	568 b	[65]
UV-visible	Wavelength, nm
	<200	290	385	This study
	<190	290	390	[65]

s—strong; b—broad; w—weak.

), which are in decent agreement with those reported by [65] for the analogous crystalline compound (568 cm⁻¹ for an NH₄⁺ vibration mode, 490 and 450 cm⁻¹ for the S-S mode, and a shoulder at 294 cm⁻¹ for the Pt-S mode). A UV-Vis spectrum of (NH₄)₂Pt^IV(S₅)₃ aqueous solution was collected using a Specord S600 spectrophotometer at 25 ± 1 °C in the range 184–1020 nm with a 0.5 nm resolution and an integration time of 60 ms (Figure A2b, Table 2). The two identified peaks at 290 and 385 nm are similar both in their energy position and amplitude ratio (1.2) to those reported by Wickenden and Krause (1969) [65]. The obtained solid was analyzed as a sample by XAS using the approach established in this study on the Pt reference compounds.

2.2. X-ray Absorption Spectroscopy (XAS)

2.2.1. Acquisition Setup

The X-ray absorption spectra were recorded at the Pt L_III-edge (~11,564 eV) at beamlines BM16-FAME-UHD [38] and BM30b-FAME [71] of the European Synchrotron Radiation Facility (ESRF), Grenoble, France. Both beamlines have the same X-ray optics setup, with a Si (220) double-crystal monochromator with sagittal focusing for choosing the energy of X-rays delivered from a bending magnet, and two Rh-coated mirrors for harmonic rejection and vertical focusing, yielding a beam spot on the sample of 200 × 300 and 200 × 200 μm² (full width at half maximum, FWHM) at FAME and FAME-UHD, respectively. Both beamlines also use the same setup for acquisition in transmission mode, consisting of silicon diodes collecting scattered radiation from a Kapton foil placed in the incident and transmitted X-ray beam. The differences between the beamlines are in the fluorescence acquisition mode. At FAME, a solid-state (SS) multi-element detector is placed in the right-angle geometry to the incident beam (Canberra solid-state 13-element germanium detector). The dead time for all samples was around 2%. At such a dead time value, the signal is linear and there is no need for dead time correction. The dead time was not registered during scan acquisition, but the distance between the fluorescence detector and each sample was optimized to always remain within the linear part of fluorescence signal acquisition. At FAME-UHD, a high-energy resolution fluorescence detection (HERFD) setup is used [72], with a recently developed crystal analyzer spectrometer [73]. For Pt L_III-edge HERFD-XANES measurements, the spectrometer was tuned to the maximum Lα₁ (9.442 keV) X-ray emission line using the (660) reflection of a Ge (220) analyzer at a Bragg angle of 80°. Photons diffracted by the crystal analyzers were collected using a silicon drift mono-element detector (SDD).

The difference in energy resolution of TFY and HERFD is due to different widths of the energy levels involved in the absorption process. In conventional XAS, for an L_III absorption edge, the final state has a 2p_3/2 core hole. In HERFD-XAS, the Lα₁ fluorescence line corresponds to the 3d to 2p_3/2 electronic transition with a shorter core-hole lifetime. Fundamentally, the improvement in the HERFD-XANES spectral resolution compared to conventional spectra is a direct consequence of the difference between these final-state widths [31]. Because the energy resolution of the incident beam (beamline resolution) is always better than the energy bandwidth of the final state, the intensity in HERFD increase due to the transition will occur in a narrower energy range [31]. For Pt, the apparent core-hole lifetime broadening of the HERFD Lα₁ measurement is only 1.94 eV, whereas for the TFY and transmission, the core-hole broadening is 5.39 eV [38]. Please note that the resolution of the monochromator (0.65 eV; [38]) and the emission spectrometer (0.65 ± 0.05 eV as measured at the elastic peak of the Lα₁ Pt fluorescence line) are both below the apparent core-hole lifetime broadening (1.94 eV). As a result, the apparent core-hole lifetime broadening is the main factor determining the energy resolution of Pt HERFD spectra. In practice, due to selective filtering of the Pt fluorescence line by the crystal analyzers, the apparent energy resolution of HERFD spectra compared to conventional resolution provided by the SS detector is 1–2 eV versus >~5 eV [38].

At both beamlines, the energy calibration of each scan was achieved using a 25 µm Pt metal foil recorded in transmission mode whose L_III-edge energy was set to 11,564.0 eV as the maximum of the spectrum first derivative. The acquisition time for each XAS scan was ~40 min at both beamlines. For fluorescence measurements, the solids were diluted with boron nitride (BN) powder to have concentrations close to 0.5–1.5 wt% Pt in the sample both on FAME and FAME-UHD, well below the potential saturation level of the fluorescence detector (>5 wt%) and possible self-absorption effects (>2–3 wt%). The absence of self-absorption issues was carefully checked using the Athena software corrections as a well as by direct comparisons of the white-line intensities between conventional fluorescence and transmission spectra that were always identical. Powders were pressed in 5 mm diameter pellets and affixed to a sample holder. Aqueous solutions (K₂PtCl₄ aq, H₂PtCl₆ aq) were placed in the glassy-carbon inner cell of the hydrothermal apparatus developed at the Néel Institute ([74], see also Pokrovski et al., 2006 [75] for details). The spectra of the solid references were recorded at ambient conditions (25 °C, 1 bar), whereas those of aqueous solutions at 30 °C and ~600 bar to avoid potential beam-induced phenomena such as the formation of air bubbles common at ambient pressure.

2.2.2. EXAFS Spectra Modeling

The acquired XAS spectra¹ were analyzed using the Athena and Artemis programs [76], based on the IFFEFIT program [77], and following previously established protocols (e.g., [78]), according to the EXAFS function $\chi \left(\mathrm{k}\right)$:

$$\chi \left(\mathrm{k}\right)={\mathrm{S}}_0^2{{\displaystyle \sum }}_{\mathrm{shell} \mathrm{i}}\frac{{\mathrm{N}}_{\mathrm{i}}{\mathrm{A}}_{\mathrm{i}}\left(\mathrm{k}\right)}{{\mathrm{kR}}_{\mathrm{i}}^2}{\mathrm{e}}^{\frac{-2{\mathrm{R}}_{\mathrm{i}}}{\mathsf{\lambda }\left(\mathrm{k}\right)}}\sin \left(2{\mathrm{kR}}_{\mathrm{i}}+{\mathsf{\Phi }}_{\mathrm{i}}\left(\mathrm{k}\right)\right){\mathrm{e}}^{-2{\mathrm{k}}^2{\mathsf{\sigma }}_{\mathrm{i}}²}$$

(1)

where S₀² is the amplitude reduction factor, ${\mathrm{N}}_{\mathrm{i}}$ is the coordination number in shell i, ${\mathrm{A}}_{\mathrm{i}}$ is the backscattering amplitude, ${\mathrm{R}}_{\mathrm{i}}$ is the interatomic distance, ${\mathsf{\Phi }}_{\mathrm{i}}\left(\mathrm{k}\right)$ is a phase shift, and ${\mathsf{\sigma }}_{\mathrm{i}}^2$ is the mean square relative displacement (MSRD, so-called Debye–Waller factor). Spectra were normalized to the absorption edge height, and atomic background subtracted. Care was taken to process the spectra as consistently as possible, and to evaluate the effect of the different Athena tuning parameters for atomic background subtraction. This effect was checked by fits (see below) of EXAFS spectra extracted using different Athena parameters but was found to be very minor. Therefore, the following parameters have been chosen here for consistency, which are also in agreement with common EXAFS extraction recommendations (The Athena Users’ Guide, [76,79]). The k-weight of the background polynome spline was chosen at 2 for all solids and at 1 for aqueous solutions, with a normalization order of 3 for all samples. The spline ranges in k and E were 0–14 Å⁻¹ and 0–740 eV (where 0 is the relative position of the absorption edge), respectively. The normalization ranges were from −115 to −30 eV for the pre-edge and from 150 to 730 eV for the post-edge. The E₀ parameter was chosen for all samples as the energy of the first maximum of the first derivative of µ(E) (

Table A1. Positions of the absorption edge (edge jump EJ; maximum of the first derivative) and the maximum of the white line (WL) of Pt L_III-edge XANES spectra of the Pt compounds investigated in this study in different modes (HR vs. TR and NR). Spectra were calibrated with a Pt foil. Uncertainty of the energy values is ±0.3 eV.

Sample/Standard	Feature	Position, eV		WL Amplitude (Normalized Spectra)
		This Study	Literature	HR	TR or NR
Pt0	EJ WL	11,564.0 11,565.5	11,564.0 a 11,565.5 a	–	1.29
PtIIS-h	EJ WL	11,565.6 11,566.9	11,564.5 a 11,567.5 a	2.66	1.35
(NH4)2PtIV(S5)3	EJ WL	11,566.7 11,567.8	– –	3.55	1.71
K2PtIICl4-a	EJ WL	11,565.0 11,566.1	11,563.5 b 11,566.0 b	3.35	1.48
K2PtIVCl6-a	EJ WL	11,567.2 11,568.0	11,565.5 b 11,567.7 b	5.62	2.24
PtIVO2-a	EJ WL	11,567.5 11,568.4	11,565.6 c 11,567.8 c	5.88	2.49
C6H12N2O4PtII	EJ WL	11,566.2 11,567.5	– –	3.20	1.65

High-resolution mode (HR), and conventional mode analogues (transmission, TR, and nominal-resolution total fluorescence yield, TFY—NR). ^a Ref. [28], calibrated with a Pt foil (EJ: 11,564.0 eV and WL: 11,566.5 eV). ^b Ref. [16], calibrated with a Pt foil (EJ: 11,564 eV). ^c Ref. [30], calibrated with a Pt foil (EJ: 11,564 eV).

). The frequency cutoff of the background, determined by the Rbkg parameter, was set to a value of 1.2 Å, which was found to be the most optimal for all samples because it provided Fourier transforms (FTs) clean from unphysical low-distance features while not affecting the higher-distance tail of the nearest neighbor shell, in particular for PtO₂ and C₆H₁₂N₂O₄Pt compounds. An FT was applied to extract the frequency-dependent contributions from the backscattering atoms, using a Kaiser–Bessel window with a dk value of 3.0. Exploitable k-ranges of our EXAFS spectra were between 3–11 Å⁻¹ and 3–13 Å⁻¹, depending of the signal-to-noise ratios (S/N), with neither influence on S₀² values, nor on other EXAFS structural parameters, within their respective uncertainties.

Least-square EXAFS fits were performed in R-space on both real and imaginary parts of FT to obtain the structural information. In addition, a non-structural parameter, Δe, accounts for energy shift between experimental spectrum and FEFF calculations (FEFF6 ab initio code integrated in the Artemis program; [80]) that were used to calculate the amplitude and phase electron scattering factors for neighboring atoms. To derive S₀² from the spectra of well-known compounds, the coordination numbers N were fixed in the fit to the crystallographic values for each Pt-bearing compound. Fits were performed with k-weighting of 1, 2, and 3 to diminish correlations between (N × S₀²) and σ², and R and Δe, and to improve fit robustness [76]. The uncertainties of the derived structural parameters reported by the Artemis program were further assessed by comparing fits of the same EXAFS spectrum with different numbers of independent points (defined by the fitted k and R-ranges and the S/N ratio) and variables (by setting, constraining, or releasing some of them). The correlations intrinsic between specific EXAFS parameters (such as S₀² or N vs. σ², and R vs. Δe) have been carefully examined and their effect on the S₀² and R values assessed (see below). The uncertainties reported in this study typically correspond to 1 standard error (1 SE). The multiple scattering (MS) contributions were considered for all samples and, except for Pt metal, the only significant contribution was the forward through absorber linear scattering. However, because the EXAFS signal is in most cases strongly dominated by the first shell single scattering, MS contributions had almost negligible effect on the determination of S₀².

3. Results

3.1. XANES Results

It can be seen in Figure 3a and Table A1 that the XANES spectrum of the K₂Pt^IICl₄ solid has a white-line energy 0.6 eV lower than Pt^IIS, which is also in agreement with previous studies [16,28,29]. The Pt^IV hexa-coordinated compounds (K₂Pt^IVCl₆ and Pt^IVO₂) have distinctly higher white-line amplitudes (by a factor of 2 to 3) and energy positions of white-line maximum (by 1 to 2 eV) than their Pt^II square-planar counterparts (K₂Pt^IICl₄, Pt^IIS), in good agreement with the general tendencies for metal coordination and oxidation state observed in XANES spectra in previous studies, e.g., [81]. The differences among the Pt^IV-S and Pt^II-S solids are revealed both in HR and TR modes (Figure 3b and Figure A3), with a Pt^IIS white-line amplitude and energy position distinctly lower than those of Pt^IVS₂. Platinum disulfide of commercial origin (PtS₂-c) presents a spectrum in between those of Pt^IIS and crystalline Pt^IVS₂ (Figure A3b), illustrating its impure composition, which should include some Pt^IIS. Aqueous solutions of K₂PtCl₄ and H₂PtCl₆ show similar normalized XANES spectra as their respective solid references K₂PtCl₄ and K₂PtCl₆, both in TR and NR modes, illustrating the spectral similarity between the two conventional acquisition modes (Figure A4).

The Pt^IV hexa-coordinated compound ((NH₄)₂Pt^IV(S₅)₃) has a similar white-line amplitude and energy position of the white-line maximum as the other Pt^IV references investigated here (Figure 3a). Both (NH₄)₂Pt^IV(S₅)₃ and Pt^IVS₂ have very similar white-line shapes, positions, and amplitudes (Figure 3b), demonstrating that their XANES signals are largely dominated by a similar first shell Pt-S₆ clusters (Figure 1). Therefore, (NH₄)₂Pt^IV(S₅)₃ may also serve as a Pt^IV-S XANES reference for sulfide-bearing minerals and fluids. The presence of a second shell of sulfur atoms has no visual impact on the XANES spectra in HR-fluorescence mode, which show only one wide post-edge peak right to the white line (Figure A3a). The presence of more distant shells (Pt-Pt and Pt-S₂) is not directly detectable by XANES and would require more advanced ab initio modelling of XANES spectra, e.g., [18,36], which is beyond the subject of our exploratory study.

The HR-XANES acquisition mode offers a gain in resolution for the white-line compared to nominal-resolution XANES (Figure 4 and Figure A5 and Table A1). This gain allows a more precise determination (±0.3 eV) of the white-line maximum energy position. This gain enables, for example, a more accurate distinction between the Pt^II and Pt^IV redox states in sulfide and chloride solids, which was difficult in nominal-resolution spectra in some previous studies, e.g., [28,29]. The HR mode also results in a significant increase in amplitude and decrease in width of the post-edge XANES resonances, with each feature neatly emphasized (Figure 4), in particular for compounds having pronounced second shell signals evidenced by EXAFS (see below). These resolution improvements offered by HR mode allow more in-depth analyses of spectral differences among Pt-S and Pt-Cl compounds and better distinction between Pt redox states.

3.2. EXAFS Results

3.2.1. Reference Compounds

For all studied compounds (Pt⁰, Pt^IIS, Pt^IVS₂, K₂Pt^IICl₄, K₂Pt^IVCl₆, Pt^IVO₂, and C₆H₁₂N₂O₄Pt^II) whose spectra were acquired in HR (or NR) and TR modes, the EXAFS-derived interatomic distances are in good agreement with those from XRD crystallographic information (

Table 3. Parameters of the Pt local atomic structure in the Pt solids and solutions obtained by fitting Pt L_III-edge EXAFS spectra and their comparison with crystallographic values.

	EXAFS Fit										Crystallographic Value a
Atom	N	R, Å	σ2 × 103, Å−2	Δe, eV	R-Factor × 103	S02	R-Range, Å	k-Range, Å−1	Number of Scans	Mode	R, Å
Pt metal, FAME
Pt	12	2.762 (4)	6 (1) b	8 (1)	18	0.75 (6)	1.2–6.0	3.0–12.0	7	TR	2.773
Pt	6	3.89 (2)	10 (5) b								3.922
Pt	24	4.8 (1)	9 (2) b								4.803
Ptfs	24	5.525 d	24 (10)
Ptdfs	12	5.525 d	13 (3)
Ptot	96	5.14 (4)	13 (10)
Pt metal, FAME-UHD
Pt	12	2.767 (3)	4.3 (6) b	8 (1)	9	0.79 (5)	1.2–6.0	3.0–12.0	13	TR	2.773
Pt	6	3.91 (1)	6 (2) b								3.922
Pt	24	4.807 (7)	7 (1) b								4.803
Ptfs	24	5.525 d	22 (6)
Ptdfs	12	5.525 d	10 (2)
Ptot	96	5.14 (3)	14 (10)
PtS-h, FAME-UHD
S	4	2.301 (8)	4 (1)	6 (1)	33	1.06 (12)	1.2–5.0	3.0–13.0	3	HR	2.312
Pt	4	3.47 (3)	9 (5)								3.470
Pt	8	3.89 (4)	9 (5)								3.918
S	8	4.10 (3)	8 (7)								4.169
Sfwa	4	4.61 c	7 d
PtS-h, FAME-UHD
S	4	2.303 (7)	3 (1)	10 (1)	29	0.85 (8)	1.2–5.0	3.0–13.0	3	TR	2.312
Pt	4	3.47 (3)	8 (5)								3.470
Pt	8	3.93 (3)	9 (5)								3.918
S	8	4.12 (4)	14 (10)								4.169
Sfwa	4	4.61 c	6 d
PtS-s, FAME
S	4	2.310 (4)	2.1 (8)	7 (1)	15	0.72 (4)	1.2–5.0	3.0–13.0	2	TR	2.312
Pt	4	3.47 (2)	8 (3)								3.470
Pt	8	3.92 (2)	7 (3)								3.918
S	8	4.13 (3)	8 (6)								4.169
Sfwa	4	4.62 c	12 (10)
PtS2-s, FAME
S	6	2.385 (6)	2.4 (1.2)	9 (1)	28	0.77 (6)	1.2–5.0	3.0–13.0	2	TR	2.344
Pt	6	3.53 (1)	3 (2)								3.543
S	12	4.24 (2)	9 (5)								4.324
Sfwa	6	4.77 c	7 (7)
K2PtCl4-a, FAME-UHD
Cl	4	2.301 (7)	3 (1)	7 (1)	20	1.03 (10)	1.2–5.0	3.0–11.0	2	HR	2.308
K	8	4.06 (4)	6 (6)								4.078
Clfwa	4	4.61 c	5 (4)
K2PtCl4-a, FAME-UHD
Cl	4	2.293 (7)	1 (1)	9 (1)	42	0.8 (1)	1.2–5.0	3.0–11.0	2	TR	2.308
K	8	4.08 (3)	3 (1)								4.078
Clfwa	4	4.61 c	2 d
K2PtCl4-b, FAME
Cl	4	2.303 (5)	2 (1)	8 (1)	15	0.77 (1)	1.2–5.0	3.0–11.0	1	TR	2.308
K	8	4.09 (4)	23 (10)								4.078
Clfwa	4	4.61 c	5 (4)
K2PtCl4 aq, FAME
Cl	4	2.304 (5)	1 (1)	8 (1)	20	0.68 (5)	1.2–5.0	3.0–11.0	1	NR	2.308
Clfwa	4	4.61 c	17 d
K2PtCl4 aq, FAME
Cl	4	2.305 (6)	14 (14)	8 (1)	30	0.79 (7)	1.2–5.0	3.0–11.0	1	TR	2.308
Clfwa	4	4.61 c	28 d
K2PtCl6-a, FAME-UHD
Cl	6	2.311 (8)	3 (1)	6 (1)	23	1.0 (1)	1.2–5.0	3.0–11.0	2	HR	2.323
K	8	4.08 (6)	20 (10)								4.222
Clfwa	6	4.62 c	7 (7)
K2PtCl6-a, FAME-UHD
Cl	6	2.30 (1)	17 (15)	10 (2)	58	0.8 (1)	1.2–5.0	3.0–11.0	2	TR	2.323
K	8	4.2 (2)	40 (40)								4.222
Clfwa	6	4.61 c	33 d
K2PtCl6-b, FAME
Cl	6	2.312 (6)	1.5 (1.0)	8 (1)	16	0.78 (7)	1.2–5.0	3.0–11.0	1	TR	2.323
K	8	4.13 (6)	20 (15)								4.222
Clfwa	6	4.62 c	7 (7)
H2PtCl6 aq, FAME
Cl	6	2.314 (6)	1.2 (1.0)	8 (1)	20	0.71 (6)	1.2–5.0	3.0–11.0	2	NR	2.323
Clfwa	6	4.63 c	2 d
H2PtCl6 aq, FAME
Cl	6	2.315 (7)	1.2 e	9 (1)	44	0.78 (4)	1.2–5.0	3.0–11.0	2	TR	2.323
Clfwa	6	4.63 c	2.4 d
C6H12N2O4Pt, FAME-UHD
N	2	2.03 (2)	2 (1)	8 (3)	27	1.0 f	1.2–2.5	3.0–13.0	1	HR	1.995 g
O	2	2.36 (8)	23 (11)								2.317 g
PtO2-a, FAME-UHD
O	6	2.019 (6)	2 (1)	9 (1)	23	0.88 (6)	1.2–5.0	3.0–13.0	1	HR	2.070
Pt	6	3.104 (6)	2.4 (1.2)								3.100
O	12	3.68 (3)	9 (6)								3.598
Ofwa	6	4.04 c	5 d
PtO2-a, FAME-UHD
O	6	2.010 (6)	1 (1)	12 (1)	27	0.74 (5)	1.2–5.0	3.0–13.0	1	TR	2.070
Pt	6	3.106 (6)	1.0 (8)								3.100
O	12	3.82 (5)	24 (15)								3.598
Ofwa	6	4.02 c	3 d
PtO2-b, FAME
O	6	2.012 (5)	2 (1)	9 (1)	24	0.76 (5)	1.2–5.0	3.0–13.0	2	TR	2.070
Pt	6	3.100 (5)	3 (1)								3.100
O	12	3.67 (5)	10 (8)								3.598
Ofwa	6	4.02 c	4 d

S₀² is the amplitude reduction factor determined by fixing the number of atoms N to their crystallographic values in the fit; uncertainties are calculated by the ARTEMIS program. All fits are performed in R-space. The number of variables (3–10) is always lower than the number of independent points (7–23) for all fits. The sample concentrations were optimized for fluorescence mode acquisition and not for transmission mode for all solids analyzed on FAME-UHD, except for Pt metal. Multiple scattering contributions: forward scattering (fs); double forward scattering; obtuse triangle (ot); forward through absorber (fwa) Pt-X-Pt-X-Pt (X = S or Cl or O). ^a Crystallographic references: Pt—[88]; PtS—[82]; PtS₂—[86]; K₂PtCl₆—[84]; K₂PtCl₄—[85]; PtO₂—[87]. ^b σ² values for Pt metal on FAME beamline are slightly higher than those on FAME-UHD because of the lower number of scans and, consequently, lower spectral S/N statistics at FAME. ^c R(X) is set to 2 × R(X)_{1st shell} (X = S or Cl or O). ^d σ² (S) is set to 2 × σ² (X)_{1st shell} (X = S or O). ^e σ² is set to that of H₂PtCl₆ aq in NR mode. ^f S₀² is set to that of PtS-h in HR mode. ^g Values are from Pt^II(NH₃)₂Cl₂ “cisplatin” crystallographic reference [89] because of the lack of data for C₆H₁₂N₂O₄Pt^II.

and Figure A6), and the EXAFS features, which could be adequately fitted within the given S/N ratio, account for the nearest (and next-nearest when enough signal) coordination structures (Figure 5). Platinum–sulfur distances in the Pt^IIS solids derived by EXAFS fitting (Table 3) are in agreement within <0.01 Å with the crystallographic value (2.312 Å; [82]). For the C₆H₁₂N₂O₄Pt^II solid, the EXAFS-derived Pt-N and Pt-O distances, 2.03 ± 0.02 Å and 2.36 ± 0.08 Å, respectively, (Table 3), are somewhat larger on average than the corresponding crystallographic values for cisplatin Pt^II(NH₃)₂Cl₂ with Pt-N and Pt-Cl distances of 1.995 Å and 2.317 Å, respectively [83]. More distant shells were too weak and composed of too light atoms to be properly fitted. Note that carboplatin is an unstable compound, rapidly degrading at ambient conditions, that brings additional noise to signal acquisition and may explain the differences found with the more accurate crystallographic values. In platinum chloride solids and solutions, Pt-Cl distances derived by EXAFS fitting (2.30–2.31 Å, Table 3) are in agreement within <0.01 Å with the crystallographic values [84,85]. For PtS₂ and PtO₂, the EXAFS-derived Pt-S₁ and Pt-O₁ first shell and Pt-S₂, Pt-O₂, and Pt-Pt₂ second shell distances are comparable with the crystallographic values within 0.08 Å for PtS₂ [86] and 0.05 Å for PtO₂ [87]. This agreement is believed to be reasonable in light of the complex second shell structures of these solids. Both fluorescence and transmission were recorded simultaneously to compare HR (or NR) and TR spectra from several samples (HR + TR, NR + TR; Table 1). Thus, as sample concentrations were optimized for fluorescence mode for these samples (0.5–1.5 wt% Pt, Table 1), requiring much lower Pt concentrations to avoid self-absorption and detector saturation effects [36], the S/N ratios of their transmission spectra were lower than those of their fluorescence spectra. However, these S/N differences, which are rather difficult to properly quantify in EXAFS, had only a very small impact on the resulting EXAFS-derived parameters (Table 3). Overall, all our samples yield the same interatomic distances as derived from transmission and fluorescence spectra. For example, for the most concentrated samples (~3 to 6 wt% Pt; PtS-s, K₂PtCl₄-b, K₂PtCl₆-b, PtO₂-b; Table 1) optimized for transmission measurements, both the interatomic distances and MSRD factors are similar with the same uncertainty magnitudes as those for samples specifically optimized for fluorescence mode (~1 wt% Pt; PtS-h, K₂PtCl₄-a, K₂PtCl₆-a, PtO₂-a; Table 1). In addition, no self-absorption effects were found in our fluorescence spectra (see Section 2.2.1). As a result, our data demonstrate good consistency in the EXAFS-derived structural parameters, which are the distances and MSRD factors, that are independent of the S/N ratio, Pt concentration, and acquisition mode. The major differences concern the amplitude reduction factor S₀², highlighted by the increased amplitude of the HR-EXAFS oscillation at low k compared to conventional EXAFS oscillations (Figure 6). In HR mode, the value of the S₀² parameter averaged over all investigated Pt references (Table 3) is 0.99 ± 0.07 (1 SE), whereas in TR and NR modes, this value is systematically lower on average, of 0.76 ± 0.04 (1 SE), as will be discussed in Section 4.2.

3.2.2. (NH₄)₂PtS₅ Sample

The EXAFS fitting approach and the S₀² values derived above were validated on the (NH₄)₂PtS₅ sample. Modeling of EXAFS spectra of (NH₄)₂Pt^IV(S₅)₃ in HR and TR modes was performed with the two S₀² values determined from the other Pt references (see below and

Table 4. Parameters of the Pt local atomic structure in the (NH₄)₂Pt^IV(S₅)₃ compound obtained by fitting Pt L_III-edge EXAFS spectra and the comparison with crystallographic values.

	EXAFS Fit							Crystallographic Value a
Atom	Mode	S02	N	R, Å	σ2 × 103, Å−2	Δe, eV	R-Factor × 103	N	R, Å
(NH4)2PtIV(S5)3, FAME-UHD
S1 b	HR	0.99	5.7 (4)	2.370 (5)	6 (1)	8 (1)	12	6	2.358–2.436
S2			7 (3)	3.56 (2)	20 (10)			6	3.39–3.81
SMS			6 (3)	4.78 (2)	10 (10)			6
S1 b	TR	0.76	5.8 (3)	2.366 (4)	4 (1)	9 (1)	15	6	2.358–2.436
S2			7 (2)	3.56 (2)	16 (9)			6	3.39–3.81
SMS			5 (1)	4.76 (2)	8 c			6

S₀² is the amplitude reduction factor set at the values of 0.76 and 0.99 for the HR and TR modes, respectively. R-ranges and k-ranges for all fits were, respectively, 1.2–5.0 Å and 3.0–13.0 Å⁻¹. Uncertainties are calculated by the ARTEMIS program and further evaluated by comparison of different fit models. All fits are performed in R-space. The number of variables (9–10) is always lower than the number of independent points (24) for all fits. Multiple scattering paths (MS) correspond to Pt-S₁-Pt-S₁-Pt. ^a values from [90]. ^b The potential distance resolution in our EXAFS fits for a k-range of 10 Å⁻¹ is 0.16 Å [91,92], which does not allow discrimination of the different S distances ranging only over 0.08 Å in the distorted PtS₆ first shell octahedron according to the crystallographic data. ^c σ² (S_MS) is set to 2 × σ² (S₁).

and Figure 7). In both cases, the results are in good agreement with the known crystallographic structure of the solid [90]. The number of S atoms in the first shell is 5.7 ± 0.4 when S₀² is set to 0.99 in HR mode, and 5.8 ± 0.3 when S₀² is set to 0.76 in TR mode, which is within the uncertainties of nominal six S atoms from crystallographic data. Inversely, if S₀² is set to 0.76 in HR mode or to 0.99 in TR mode, for similar σ² values, the numbers of S atoms are either significantly overestimated (7.4 ± 0.5) or underestimated (4.5 ± 0.2), respectively. The number of S atoms in the second shell has larger uncertainties (±2–3), but yet converges to an average number of six atoms in all fits. The EXAFS-derived mean Pt-S₁ (1st shell) distance for (NH₄)₂Pt^IV(S₅)₃ is 2.370 ± 0.005 Å in HR mode and 2.366 ± 0.004 Å in TR mode, which is in the range of the XRD-derived values (2.358–2.436 Å; [90]). However, statistical resolution of our EXAFS spectra (0.16 Å, ΔR = π/2Δk; e.g., [91,92]) does not allow the different Pt-S₁ crystallographic distances to be distinguished within the first shell (≤0.08 Å) reporting only a mean value (2.370 ± 0.005 Å in HR mode and 2.366 ± 0.004 Å in TR mode). The derived distance is significantly longer than those of other Pt-S compounds, such as PtS₂ (2.34 Å; [86]) and PtS (2.31 Å; [82]). The EXAFS-derived mean Pt-S₂ (2nd shell) distance, 3.56 ± 0.02 Å both in TR and HR modes, is coherent with the crystallographic value (3.39–3.81 Å; [90]). The multiple scattering contributions within the first shell S atoms are identical within uncertainties in both modes (4.78 ± 0.02 Å in HR and 4.76 ± 0.02 Å in TR mode). A wavelet transform analysis, based on the Continuous Cauchy Wavelet Transform (CCWT; [93,94]), was performed for the (NH₄)₂Pt^IV(S₅)₃ (in HR and TR modes) and Pt^IVS₂-s solids (in TR mode), and confirmed the presence of both relatively light atoms (S₂) and MS in beyond-the-first shells of (NH₄)₂Pt^IV(S₅)₃, in contrast to Pt^IVS₂ whose second shell is dominated by heavy Pt atoms (Figure 8).

4. Discussion

4.1. Resolution Improvement in HERFD Mode

Our results show that Pt L_III-edge XANES spectra acquired in HERFD mode (HR) show a significant gain both in the energy-position resolution and the shape and amplitude of white-line and post-absorption features, compared to conventional transmission (TR) or fluorescence (NR) modes, as has been already recognized in previous studies [39,40,41,44]. Therefore, the Pt^II and Pt^IV formal redox state and atomic coordination may be more clearly distinguished in Pt chloride- and sulfide-bearing systems in HR-XANES mode than in some previous works that used conventional acquisition modes [16,28,29,95]. This resolution improvement may add, in particular, more precision to the characterization of Pt-S and Pt-Cl complexes in hydrothermal fluids where a mixture of Pt^IV and Pt^II of different coordination may be expected, e.g., [18,96], as well as in natural mineral phases where Pt can be present in different coordination and oxidation states (e.g., Pt^II and Pt^IV in pyrite and pyrrhotite; [28,29]). It can be seen in

Table A2. Ratio between the maximum amplitude absorption at the white line (WL) of the normalized XANES spectra for the indicated pairs of compounds. Note that ratios in HR-fluorescence mode are always higher than in transmission mode (TR), demonstrating an improved spectral contrast in HR mode.

	PtIV-Cl/PtII-Cl	PtIV-S/PtII-S	PtII-Cl/PtII-S	PtIV-Cl/PtIV-S
HR-XANES	1.68	1.34	1.26	1.58
TR-XANES	1.52	1.28	1.10	1.31

Pt^IV-Cl = K₂Pt^IVCl₆-a; Pt^II-Cl = K₂Pt^IICl₄-a; Pt^IV-S = (NH₄)₂Pt^IV(S₅)₃; Pt^II-S = Pt^IIS-h.

that the amplitude ratios between the main XANES resonances of the spectra of Pt^II-Cl versus Pt^IV-Cl, or Pt^II-S versus Pt^IV-S, types of compounds are significantly amplified in HR mode, allowing the use of these ratios to quantify the fractions of the different complexes in a mixture. Until now, such an analysis was limited to Pt-O vs. Pt-Cl complexes [81], whose spectra have relatively strong contrasts in TR or NR modes.

4.2. Amplitude Reduction Factor (S₀²)

We have not found any significant differences in the EXAFS spectra and the resulting fitted interatomic distances for the same compounds passed in HR and NR/TR modes (Table 3). Both beamlines provide identical S₀² values within errors in conventional modes (TR and NR), as illustrated here by Pt metal EXAFS fit results yielding 0.75 ± 0.06 on FAME and 0.79 ± 0.05 on FAME-UHD, similar to the values obtained at other synchrotrons in previous studies (0.82, [97]; 0.84 ± 0.03, [98]). Even though S₀² values in HR and NR/TR modes are all within the typical range of 0.65–1.10 reported for many elements at different beamlines (e.g., [99,100]), there is a systematic difference of >0.2 on average between the two sets of values (Figure 9). For example, Asakura et al. (2018) [97] reported the same gap with an S₀² value of 1.00 in HR mode versus 0.82 in NR mode from EXAFS fitting of platinum metal at Pt L_III-edge. Indeed, spectra in HR mode were reported to have larger amplitudes than in TR mode due to a smaller apparent core hole (e.g., [97,101]), yielding an increase in the amplitude of the EXAFS signal in the k-range 2–6 Å⁻¹ compared to the spectra in TR or NR modes (Figure 6). However, the disorder, represented by the mean square relative displacement, σ², is known to contribute to the EXAFS signal more significantly at higher k values (>8–10 Å⁻¹), and might also influence the determination of S₀². This effect is discussed below.

4.3. Modeling Disorder Using Mean Square Relative Displacement, σ²

To investigate the potential influence of σ² on the S₀² determination, these parameters that are strongly correlated in the EXAFS formula must be decoupled. To do so, we performed EXAFS fits on several samples (Figure 10), whose spectra were acquired both in HR and TR modes, by fixing the σ² value to the mean value derived from HR and TR mode spectra for the same Pt solid compound (Table 3). The newly obtained S₀² values still present similar systematic differences between HR and TR modes (~0.15;

Table A3. Parameters of the Pt local atomic structure in the Pt solids obtained by fitting Pt L_III-edge EXAFS spectra, by setting the σ² value to the mean value between those obtained in HR-fluorescence (HR) and transmission (TR) modes from Table 3.

Compound	σ2 × 103, Å−2	Mode	S02	Δe, eV	Pt-X a, Å	R-Factor × 103
K2PtCl4-a	2.0	TR	0.82 (4)	9 (1)	2.293 (7)	43
K2PtCl4-a	2.0	HR	0.94 (4)	7 (1)	2.300 (6)	23
PtS-h	3.1	TR	0.87 (4)	10 (1)	2.303 (6)	39
PtS-h	3.1	HR	1.02 (6)	6 (1)	2.301 (7)	34
PtO2-a	1.5	TR	0.74 (3)	12 (1)	2.009 (6)	33
PtO2-a	1.5	HR	0.83 (4)	9 (1)	2.018 (6)	28

^a X = Cl or S or O in the 1st shell.

, Figure 10b). In contrast, for analogous fits, performed by fixing the mean S₀², the obtained σ² values overlap within their uncertainties between HR and TR modes (Figure 10c). Thus, despite the apparent correlation between S₀² and σ² (Figure 9b and Figure 10b), the S₀² values are distinctly different for a given solid between the two modes, in contrast to the largely overlapping σ² values. This analysis demonstrates a relatively minor effect of σ² compared to S₀² on the EXAFS fits between TR/NR and HR modes. The increase in EXAFS amplitude at low k values clearly highlights the gain of amplitude in HR-EXAFS, while the influence of σ² is less apparent, being partly obscured by the growing noise at higher k values. The greater amplitude is therefore compensated by a significantly greater S₀² value in HR-EXAFS fits.

4.4. Implications

The systematic difference in the S₀² values revealed between HR and NR/TR acquisition modes highlights the necessity of the accurate determination of S₀² in different acquisition modes. This necessity may be particularly crucial when determining coordination numbers, as we have demonstrated using the (NH₄)₂Pt^IV(S₅)₃ sample. For example, using a “wrong” S₀² value (0.76) established by fitting conventional transmission spectra of Pt references, yields a significant overestimation of the first shell Pt-S coordination number, N(S₁) = 7.4 ± 0.5 instead of the right value of 6, when EXAFS fits HR fluorescence spectra, while neither the R nor σ² value would be significantly affected. These results demonstrate the much larger influence of S₀² than σ² on the determination of coordination numbers, at least for EXAFS spectra with comparable signal-to-noise ratios at relatively low k values (<10 Å⁻¹, Figure 7). More generally, the necessity of the S₀² value correctly calibrated in the right mode may be particularly crucial when determining ligand coordination numbers in a mixture of aqueous complexes from EXAFS data, for example, in the case of Pd-O/Cl complexes [14] or Mo-O/S complexes [102]. Note that, an error of 0.2 in S₀² would lead to an error of 0.4 to 0.8 sulfur atoms for the average number of HS⁻ ligands around Pt in an aqueous solution containing Pt(HS)₂⁰, Pt(HS)₃⁻, and Pt(HS)₄²⁻ complexes, making it difficult to identify the major solubility-controlling complex [96]. This identification is necessary for accurate predictions of metal transport by hydrothermal fluids, e.g., [18,35].

5. Concluding Remarks

In this study, we attempted to develop robust and safe protocols of laboratory synthesis of different Pt-S-bearing solid compounds (Pt^IIS, Pt^IVS₂, and (NH₄)₂PtS₅) that may be used as XAS references for studying Pt in natural sulfide minerals, hydrothermal fluids, and technological materials. XANES and EXAFS spectra, recorded in this study on these solids and additional Pt chloride and oxide compounds, contribute to the extension of an open-access XAS reference database (SSHADE). Spectra recorded in HERFD mode present a significant gain in resolution compared to conventional fluorescence spectra, both in energy position and amplitude/width of the white-line and post-edge XANES features. This gain may be crucial when quantifying Pt^II/Pt^IV redox and Cl/S ligand ratios in samples that contain a mixture of different Pt complexes and redox states. EXAFS spectra in all modes (transmission, fluorescence, and HR-fluorescence) yield identical structural parameters in terms of interatomic distances, in good agreement with available crystallographic values. However, the amplitude reduction factor (S₀²) derived by fitting HR fluorescence spectra, 0.99 ± 0.07, is systematically higher than the value derived from transmission and conventional fluorescence spectra, 0.76 ± 0.04. This difference is explained by the larger EXAFS amplitudes in HR-fluorescence than in transmission or NR-fluorescence, and highlights the importance of using the right acquisition mode to be able to better constrain the S₀² parameter and to use it for more accurate derivations of coordination numbers in unknown samples. The HR mode appears, in addition, to be particularly interesting for complex natural or industrial Pt samples containing elements with close absorption edge energies (W, Au, As, Ta, Re).