Tables for
Volume I
X-ray absorption spectroscopy and related techniques
Edited by C. T. Chantler, F. Boscherini and B. Bunker

International Tables for Crystallography (2022). Vol. I. Early view chapter

Combined approaches and challenges: XAS and UV–visible and vibrational spectroscopy

Stephen P. Besta*

aSchool of Chemistry, University of Melbourne, Victoria 3010, Australia
Correspondence e-mail:

An overview of the strategies applied for the collection of X-ray absorption near-edge spectroscopy (XANES) and extended X-ray absorption fine-structure (EXAFS) measurements in combination with UV–visible and vibrational spectroscopy is developed in the context of the sampling requirements of the relevant techniques. Case studies highlight cases in which the sample is homogeneous but undergoing reaction, inhomogeneous and the short-lived products of chemical reactions. Sampling approaches which use optical fibres or attenuated total reflection allow coincident measurement of the X-ray and complementary spectroscopic measurements and this is complemented by strategies which allow the study of spectroscopy-validated, freeze-quenched ex situ samples.

Keywords: UV–visible spectroscopy; vibrational spectroscopy.

1. Introduction and scope

There are few, if any, investigations that can rely on a single technique, and the development of a scientific enquiry has an allied timeline of the application of the different measurements that are brought in to resolve the problem. The utility of the different observations will in many cases depend critically on the temporal relationship between the different observations, for example serial or parallel, and the development of sampling strategies or experimental techniques which allow the simultaneous measurement of different spectroscopic or physical probes of the sample can give multiplicative advantages for interpretation. This chapter focuses on the application of X-ray absorption spectroscopy (XAS) in applications where long-wavelength (IR and UV–visible) complementary techniques are deployed immediately preceding, following or concurrently with the XAS measurement.

The research areas that have been most active in the development of XAS measurements incorporating complementary characterization techniques can be categorized as being focused on catalysis or reaction chemistry (Agistini et al., 2023[link]). The majority of studies of catalytic systems have involved heterogeneous catalysis (Patlolla et al., 2012[link]; Rochet et al., 2017[link]; Müller et al., 2018[link]; Newton, 2009[link]; Beale et al., 2005[link]) where X-ray scattering or diffraction (Frenkel et al., 2012[link]), Raman and UV–visible (Patlolla et al., 2012[link]; Müller et al., 2018[link]) spectroscopy have been shown to be compatible sampling requirements. An excellent review of the difficulties, benefits and prospects has been published by Newton & van Beck (2010[link]). Reaction chemistry is used as a broad label covering the study of rapid chemical reactions (Olivo et al., 2017[link]; Bauer et al., 2010[link]), the growth of nanoparticles (Nayak et al., 2017[link], 2018[link]; Stötzel et al., 2010[link]), crystallization (Zalden et al., 2012[link]) and metal nanoparticle-stabilized block copolymer growth (Nayak et al., 2016[link]). In a number of applications the rate of reaction has demanded the rapid collection of spectroscopic information, and this has been facilitated by the development of rapid scanning and/or dispersive XAS capabilities (Newton, 2009[link]; Rochet et al., 2017[link]; Smolentsev et al., 2009[link]; Pascarelli & Mathon, 2022[link]; Bauer et al., 2010[link]; Briois et al., 2005[link]; Olivo et al., 2017[link]). In materials science a different set of problems are more common. In this case the sample may be stable at the time of measurement, but heterogeneous. Independent characterization of the micro-domain under XAS investigation is needed to support interpretation of the results (Vantelon et al., 2009[link]; Briois et al., 2007[link]). Heterogeneity introduced by surface reactions of pigment grains can destroy artworks, and the delineation of degradation pathways depends on spatially resolved spectroscopic interrogation at micrometre resolution (Van der Snickt et al., 2012[link]). Each of these cases involves measurements where the XAS measurements are conducted simultaneously with at least one other means of characterization of the sample.

Photoreduction, particularly for measurements conducted using focused high-intensity beamlines, is a potential issue for XAS investigations and the reproducibility of the spectra provides the most common measure of the stability of the sample. However, there are cases in which complementary techniques can provide a more sensitive, and informative, independent measure of the integrity of the sample. The reliability of this measure will, of course, depend on factors such as the time required for acquisition of a single spectrum, whether the sample in the beam is static and the sensitivity of the XAS spectrum to photodegradation. Clearly, the availability of alternate spectroscopic measurements can play a crucial role in establishing the state of the sample and the utility of the XAS measurement.

The preceding discussion highlights the importance of coupled or simultaneous measurement of XAS and complementary techniques as it relates to different fields of application. Situations where XAS measurements alone are insufficient to define the state of the sample are summarized in the following section. It is these cases where the incorporation of complementary techniques into the experimental plan can significantly enhance the value of the measurements. Section 3[link] showcases the different approaches to the coupled measurement of XAS spectra with complementary spectroscopic techniques. Attention will be paid both to the strategies used to study different types of sample, the timescale needed for the measurement and the value added by coupling of the spectroscopic results.

2. Cases where the internal consistency of the XAS measurements is not a sufficient measure of sample integrity

2.1. Sample stability over extended scan times

Notwithstanding the advances in rapid-scan XAS and energy-dispersive XAS (EDXAS), which have allowed the acquisition of spectra on timescales of tens of milliseconds to ∼10 s, spectra acquired over a period of 60–90 min are common, and in many applications are advantageous. Generally, the consistency of 2–3 spectra is used as a measure of the stability of the sample in the beam, where fresh sample positions are irradiated for each spectrum. Ideally, measurement in the pre-edge and X-ray absorption near-edge structure (XANES) regions before and after each extended X-ray absorption fine-structure (EXAFS) scan for the same sample point should be used as a criterion for accepting or rejecting the measurement. At least two limitations of this approach present themselves. Firstly, the variable sensitivity of the XANES profile to photodamage. XANES is most sensitive to a change in the oxidation state of the absorber, particularly when this is associated with a pronounced structural change. Where this is not the case it is difficult to monitor the state of the sample during XANES data collection. Secondly, despite the sensitivity of many samples to photodegradation, there is no widely accepted sample-validation protocol for the publication of XAS/XAFS. Issues related to this second point clearly lie outside the scope of this contribution.

2.2. Reactive or unstable samples

This is a general problem and has implications both for the preparation of the sample and the presentation of the sample to the interrogating radiation. The strategies which best ameliorate sample degradation before, and during, the measurement of spectra are, of course, sample-specific and commonly employ freeze-quench techniques.

2.3. Samples undergoing chemical reaction

XAS will be sensitive to the composition of the reaction mixture and can be used to monitor the extent of the reaction, provided that there is no photodegradation of the sample. When combined with other spectroscopic techniques, XAS can provide insights into the redox states and structure of intermediates formed during the reaction.

2.4. Samples that are inhomogeneous on the micrometre length scale

In cases where the beam dimensions are commensurate with the scale of inclusions/patterning/mosaic structure in the sample, association of XAS with the different components of the sample can greatly benefit from independent spectral characterization of the component sample irradiated by the X-ray beam. Clearly, this is only an issue when the XAS spectrum cannot be matched to known spectra, as may occur for inclusions at the boundary of well defined materials.

3. Integration of complementary spectroscopic methods with XAS measurements

It may be argued that the most effective approach to the acquisition of complementary information during XAS measurements is for the X-ray beam to be the source of both measurements. Elegant examples of such an approach are provided by measurements of anomalous wide-angle X-ray scattering and XAS, where an early study of zeolite-supported platinum–molybdenum clusters provided simultaneous measurements of their size, spatial distribution and structure (Samant, Bergeret, Meitzner, Gallezot et al., 1988[link]; Samant et al., 1988[link]). Information related to both the molecular and the electronic structure is of paramount importance for understanding catalytic systems and this underlies the value of such systems in XANES and EXAFS investigations. A problem associated with such measurements is in cases where the catalytically important element(s) have multiple environments which impact differently on the catalytic process. This situation is common in metalloenzymes, where iron–sulfur clusters often facilitate electron transport to the active site, which itself can utilize iron. Site-selective XANES has been obtained with sufficient resolution to allow the assignment of iron redox states in metalloproteins by crystallographic refinement at multiple X-ray wavelengths (Einsle et al., 2007[link]). A more recent study has explored questions relating to the formal assignment of oxidation state in clusters and the relationship between the results of multiple-wavelength anomalous diffraction, 57Fe Mössbauer and XANES measurements in more detail (Bartholomew et al., 2019[link]). These examples, while involving a nonspectroscopic companion technique, highlight the synergies that can be obtained by using a multi-probe approach.

The incorporation of complementary techniques is, of course, achieved without disruption of the beamline when the measurements are sequential. The choice of spectroscopic or physical technique will be determined by the objectives of the study and the availability of on-site instrumentation, and the challenge is associated with the reliable transfer of the sample, unchanged, to the beamline. XAS measurements may be complicated by photodamage and assessment of the state of the sample is an important aspect of any study. While the XANES will often provide the basis for such an assessment, in some cases different spectroscopic methods provide a more sensitive measure of the state of the sample. For reactive samples it is necessary to quench chemical reactions, with this being achieved by freeze-;quenching of the sample and low-temperature measurement. This approach is described in more detail in Section 3.4[link].

Simultaneous measurements of Raman and/or UV–visible spectroscopy at synchrotron beamlines have been reported since about 1998, in which the initial measurements were coupled with SAXS/WAXS (Bras et al., 1995[link]; Bryant et al., 1998[link]) or diffraction (Ran et al., 2003[link]; Davies et al., 2005[link]; Boccaleri et al., 2007[link]) beamlines. UV–visible and Raman spectroscopy are particularly well suited to measurements of this type. In addition to providing results that are sensitive to the electronic and molecular structure of the material, by the use of optical fibre technology both methods can accommodate highly flexible sampling geometries. The selection of these complementary methods is based on the their sensitivity to both the valence electronic structure and the vibrational spectra. A range of experimental approaches have been adopted and these have been optimized to the constraints of the chemistry and the limitations associated with the beamline or the spectroscopic equipment available for the experiment. An indication of the range of approaches is given in the following examples.

3.1. Solution XAS measurements with fibre-optic coupled Raman and UV–Vis spectroscopy

The integration of Quick XANES, UV–Vis and Raman spectroscopy for the examination of solution species is most easily achieved by the modification of XAS solution cells to accommodate fibre-optic dip-probe sampling methods. An example of this strategy is provided by the investigation of the oxidation of ethanol by cerium(IV) using the D2 bending-magnet beamline of the DCI storage ring (LURE, Orsay, France; Briois et al., 2005[link]). For the system under study the XAS measurement was conducted in transmission mode with a path length of 0.5 mm, and the thermostatted solution cell included a reservoir designed to accept UV–Vis and Raman dip probes (Fig. 1[link]a). For both Raman and UV–Vis the optical fibre bundles include separate fibres for the incident and transmitted, or scattered, radiation. XANES measurements were conducted at the Ce L3 edge and extended from 5.68 to 5.88 keV. The need to match spectral collection times to the rate of chemical reaction and the rapid-scan monochromator developed by Frahm et al. (2005[link]) allowed spectra to be collected on a timescale of ∼5 s (Fig. 1[link]b). The scanning UV–Vis spectrometer allowed the collection of spectra with a duty cycle of 18 s. The multichannel Raman spectrometer can support rapid spectral acquisition, although this must be matched to the strength of Raman scattering by the sample: for the system under study [2.5 mM cerium(IV)] an integration time of 5 s gave well defined spectra.

[Figure 1]

Figure 1

(a) Experimental setup for simultaneous XAS, Raman and UV–Vis measurements of a solution sample. A Raman dip-probe tip is shown in the inset. (b) Comparison between step-scan (20 min) and rapid-scan (5 s) XAS measurements. Reprinted with permission from Briois et al. (2005[link]). Copyright 2005 American Chemical Society.

While the sampling points for the different forms of spectroscopy are not coincident, and a comparatively large volume of solution is required for operation of the cell, the kinetics of the reaction solution could be followed by UV–Vis simultaneously with the fate of the cerium(IV)–ethanol complex by Raman spectroscopy (Figs. 2[link]a and 2[link]b) and XANES, showing the interconversion of cerium(IV) (peaks A and B) to cerium(III) (peak A′) (Fig. 2[link]c).

[Figure 2]

Figure 2

Simultaneous measurements of the UV–Vis (a), Raman (b) and Ce L3-edge XAS (c) spectra recorded during cerium(IV) oxidation of ethanol. Spectra were recorded over a period of ∼2000 s from different points within the same thermostatted sample cell (Fig. 1[link]a). For clarity, every 20th XAS spectrum is shown, where the conversion from cerium(IV) to cerium(III) is reflected by the replacement of the intense features (A and B) by a single dominant white-line peak (A′). Reprinted with permission from Briois et al. (2005[link]). Copyright 2005 American Chemical Society.

It is instructive to consider the modification of the experimental setup to permit the simultaneous measurement of XANES and UV–Vis spectra from the same sample volume. In this case, a flow-cell arrangement allowed the presentation of a continuous stream of fresh solution to the sample point and crossed X-ray and UV–Vis beams permitted transmission XAS measurements of the sample (Bauer et al., 2010[link]). A schematic diagram of the experimental design is shown in Fig. 3[link] and details of the XAS flow solution cell are given in Bauer et al. (2010[link]). These experiments were completed using the XAS beamline at ANKA in the QEXAFS mode (1 min scan time), and a multichannel UV–Vis spectrometer with optical fibre light guides allowed rapid spectral acquisition (<1 s per averaged spectrum). The quality of the XANES spectra are reflected by the observation of clean isosbestic points through the early stages of the reaction.

[Figure 3]

Figure 3

(a) Schematic diagram showing the experimental design for simultaneous UV–Vis and XAS measurements from the same sample volume in a flow solution cell. (b) Ce L3-edge XAS measurements during the oxidation of ethanol by cerium(IV). The stability of the XAS spectra is reflected by the observation of well defined isosbestic points (circled) during the reaction. (a) was redrawn and (b) was reproduced from Bauer et al. (2010[link]).

While the implementation of a flow cell has significant advantages in terms of monitoring the progress of the reaction while minimizing the effects of ionizing radiation, including photodegradation of the sample, there are significant challenges, which are mostly associated with the damping of the pressure waves introduced by the pump. The cell design (clamping of the Kapton windows), pump design (magnet-coupled gear wheel) and flow rate contribute to the stability of the resulting spectra. As evident from the spectra shown in Fig. 3[link](b), careful attention to these details can deliver high-quality XANES.

EDXAS coupled with UV–Vis spectroscopy has also been applied to the study of TiO2 nanoparticle growth (using the SAMBA beamline at the SOLEIL synchrotron-radiation source; Stötzel et al., 2010[link]), the growth of block copolymers stabilized by metal nanoparticles (on BL-8 at the Indus-2 synchrotron source, India; Nayak et al., 2016[link]) and the growth of Au-Pt nanoparticles (on BL-8 at Indus-2; Nayak et al., 2017[link]).

3.2. Solid-state XAS measurements with simultaneous Raman and UV–Vis spectroscopy

The simultaneous implementation of energy-dispersive XAS, UV–Vis and Raman spectroscopies in an operando study of a heterogeneous catalyst was demonstrated by Weckhuysen and coworkers in 2005 (Beale et al., 2005[link]). XAS was measured on the ID24 beamline at the ESRF using the experimental setup shown schematically in Fig. 4[link](a). The XAS study was conducted at the Mo K edge (∼20 keV) and this allowed the use of a quartz tube reactor operating at 550°C during successive propane dehydrogenation cycles. The performance of the catalyst was monitored by inline mass spectrometry. In this example the sample-cell design allows the different spectroscopic measurements to be made simultaneously from the same sampling point, albeit with different dimensions of the interrogating beams and measurement strategies: transmission (EDXAS), back-scattering (Raman) and diffuse reflectance (UV–Vis). The minimum time for the collection of spectra ranges from 400 ms (EDXAS) to 2 s (UV–Vis), with this matching the 10 s sampling time for the mass-spectrometry measurements well. Signal-to-noise considerations may require the co-addition of individual scans; however, the successive reduction/regeneration cycles operated on a timescale of ∼20 min and the time evolution of changes to the catalyst could be characterized well using this approach. The synchronous measurements allowed changes in the oxidation state and aggregation of the molybdenum species to be linked to the deposition of a coke layer, which in turn can be related to the efficiency of the catalyst and the ability to regenerate the active species during the reoxidation cycle.

[Figure 4]

Figure 4

(a) Schematic diagram of the experimental setup for simultaneous EDXAS, Raman and UV–Vis spectroscopy. Time-resolved UV–Vis (b), Raman (c) and Mo K-edge EDXAS (d) recorded during the first propane dehydrogenation cycle for Mo/SiO2. The time between spectra is 60 s and the time course of the spectral changes is indicated by an arrow. Redrawn (a) and reproduced (b, c, d) with permission from Beale et al. (2005[link]). Copyright 2005 the Royal Society of Chemistry.

3.3. XAS measurements integrated with IR spectroscopy

Raman and IR spectroscopy provide information regarding the vibrational structure of the sample: IR by direct absorption and Raman by a less probable inelastic scattering event using shorter-wavelength radiation. While simpler from a spectroscopic perspective, IR spectroscopy presents more challenges for simultaneous XAS measurements. This mostly arises from the more limited characteristics of IR-transmitting optical fibres, the larger diffraction-limited focus dimension and less flexible sampling geometries. Despite these issues, diffuse-reflectance sampling methodologies have allowed synchronous XAS/IR measurements of supported Rh/Al2O3 catalysts (Newton, 2009[link]; Newton et al., 2004[link], 2007a[link],b[link]). The strategy used in these experiments is to study the surface region of a sample using diffuse-reflectance Fourier transform spectroscopy (DRIFTS) with the transmission XAS path lying close to the IR-probed surface (Fig. 5[link]a). The IR beam penetration will depend on the particle size of the sample, and the incident beam will have a focus at the sample surface with significantly different dimensions when compared with the X-ray beam; this will translate either into lower sensitivity (if the sample is masked) or noncoincidence of the IR and X-ray interrogated sample.

[Figure 5]

Figure 5

(a) Schematic of the experimental arrangement allowing DRIFTS/EXAFS/mass spectrometry from 0.5% Rh/Al2O3 under gas-flow conditions. (b) IR spectra: (i) He flow post NO exposure, (ii) He flow, (iii) 1 s after a switch to 5% CO/He using spectrum (ii) as a reference, (iv) 60 s of 5% CO/He, (v) 140 s of 5% CO/He then 60 s He flow. (c) Rh K-edge XAS. (i) 0.5% Rh/Al2O3 `as prepared', reduced form. (ii) Following 5% H2/He flow. (iii) After 5% NO/He flow. (iv) After exposure of (iii) to 5% CO/He at 373 K. Reproduced with permission from Newton (2009[link]). Copyright 2009 Springer Nature.

IR spectra are typically complex spectra which can be used as a fingerprint of the composition of the sample, but in cases where there is strong bonding between the atoms, such as for groups such as CO, CN and NO, the stretching mode has a wavenumber that is separated from the other vibrations of the molecule and is very sensitive to the strength of bonding. Of more general importance is the impact of binding such molecules to metals, where this is marked by a large increase in the intensity of the vibration (by one to two orders of magnitude), facilitating the identification of the metal-bound species. More importantly, the CO, CN or NO stretching frequency is highly sensitive to the identity and oxidation state of the metal to which the diatomic is bound. The IR spectra recorded from the Rh/Al2O3 catalyst (Newton, 2009[link]) after prior exposure to 5% NO/He at 323 K has a single dominant NO-bound species [ν(NO) = 1725 cm−1] as shown in spectrum (i) in Fig. 5[link](b). Exposure of the sample to helium (373 K) leads to a significant increase in the relative abundance of a minor species with ν(NO) at 1920 cm−1 [spectrum (ii) in Fig. 5[link]b]. The introduction of 5% CO/He (373 K) results in rapid displacement of NO by CO [spectrum (iii) in Fig. 5[link]b], where the spectral changes are associated with the depletion of NO [negative bands for the spectra calculated using spectrum (ii) in Fig. 5[link]b as a reference] and rhodium-bound CO (2092 and 2020 cm−1). The gas-phase CO is readily identified by the comparison of spectra (iv) and (v) in Fig. 5[link](b), where the latter was recorded with a pure helium gas stream. The Rh K-edge XANES spectra recorded from the Rh/Al2O3 catalyst are shown in its pristine form [spectrum (i) in Fig. 5[link]c], following reduction by 5% H2/He [spectrum (ii) in Fig. 5[link]c] and then after exposure to 5% NO/He [spectrum (iii) in Fig. 5[link]c] and 5% CO/He [spectrum (iv) in Fig. 5[link]c]. It is important to note that a distinction between Rh-(NO+), Rh-(NO), mixed Rh-(NO)/(CO), mono and dicarbonyl, terminally bound and bridged species can be inferred from the spectra and this can inform the identification of the conditions needed for the collection of XAS spectra from a sample dominated by a single species or inform on the structural models used to model the EXAFS.

3.4. Combination of micro-XAS and micro-Raman spectroscopy

In cases where the sample is inhomogeneous it is important to have high spatial resolution for both spectroscopic techniques, and for the measuring beams to be aligned to interrogate the same sample volume. These capabilities can be important for the study of interfaces, structural order/defects and inclusions. The implementation of these capabilities is demonstrated by simultaneous micro-XAS and micro-Raman measurements from a heterogeneous mica sample in experiments conducted on the LUCIA beamline at SOLEIL (Briois et al., 2007[link]). The 7 × 5 µm X-ray beam was obtained from a double-crystal Si(111) monochromator focused using two mirrors in a Kirkpatrick–Baez configuration. The resulting focused X-ray beam has dimensions which match the size of the 785 nm Raman exciting laser after focusing its output from a single-mode optical fibre. The back-scattered radiation was collected using an optical fibre bundle. The optical demands of the two measurements were accommodated by inclining the sample by 45° to both beams. Translators both on the sample and the optical Raman probehead allowed the optical beams to be aligned; this was checked using a 16 µm titanium dot (Fig. 6[link]a). XAS measurements were conducted at the Si K edge with the sample environment maintained under vacuum.

[Figure 6]

Figure 6

(a) Vertical scan of the focused X-ray and Raman microprobe beams over a 16 µm diameter titanium dot on a silicon substrate. A similar alignment was obtained in the horizontal direction. (b) XRF map of a gneiss mineral section. (c) Raman spectra from two points of interest (POI) and a reference muscovite sample. (d) Micro-XAS measurements from POI 2 and POI 3 and the polarization-dependent muscovite XAS. Reproduced from Briois et al. (2007[link]).

A thin section of gneiss rock contains quartz, feldspar and mica as major elements, where these have a distribution which has structure on the micrometre level, as reflected by the variation in elemental distribution as measured by X-ray fluorescence (XRF) cartography (for example aluminium; Fig. 6[link]b). Based on the XRF measurements, several `points of interest' (POI) were identified and analysed by a combination of Raman and Si K-edge XANES. An example of the advantage gained from access to a second technique is given by the spectra from POI 2 and POI 3 (Fig. 6[link]c), where the Raman spectra suggest the presence of the same mineral but the XANES profile is significantly different for the two sampling points. Phyllosilicates such as muscovite have a lamella structure and will have different orientations in the gneiss. XANES with the photon electric field in and normal to the plane of the muscovite sheets gives distinct spectra which match the spectra recorded from POI 2 and POI 3 well and resolve the apparent inconsistency between the Raman and XANES measurements.

3.5. Fast XAS, UV–Vis and XAS measurements to follow rapid chemical reactions

The path of chemical reactions is charted through the identification of comparatively stable intermediates formed during the progress of the reaction. Elucidation of the details of important reactive species depends on matching the reaction conditions with techniques having sufficient time resolution to allow spectroscopic measurement on a timescale where the composition of the sample is both well defined and stable. The temporal resolution of simultaneous XAS measurements with UV–Vis and/or Raman spectroscopy is in most cases limited by the X-ray technique. The development of EDXAS techniques with millisecond time resolution at ID24 at the ESRF (Pascarelli et al., 2016[link]; Smolentsev et al., 2009[link]) has allowed details of the redox-state and structure changes in the species formed during iron-catalysed reactions with peroxide and peroxyacetic acid to be obtained (Olivo et al., 2017[link]).

The short time interval between mixing the reactant solutions and the collection of spectra requires the application of stopped-flow techniques, with the attendant pressure wave associated with the rapid change in flow. This places significant constraints on the design of the solution cell. Geometric stability of the cell path length has been achieved using a quartz capillary. EDXAS measurements in transmission mode could be obtained from 35–70 mM solutions of [FeII(TPA)(CH3CN)2]2+ in acetonitrile. While the high solute concentration needed for good-quality EDXAS results in saturation absorption over the main absorption bands in the UV–Vis spectra, the kinetics of the reaction could be followed using the weak absorbance at the wings of these bands. The reproducibility of the experiment was checked by repeated experiments in which the 40 ms time resolution of the XAS measurement was demonstrated by comparison between conventional and rapid-scanning measurements (Fig. 7[link]a). The Fe K-edge XANES shows well defined shifts over the first second of the reaction which differ for the reactions with peroxide and peroxyacetic acid. Whereas the reaction with peroxide gives only iron(II) and iron(III) species, chelating peroxyacetate iron(III) species and ferryl [oxo iron(IV)] species are proposed for reactions involving peroxyacetate (Fig. 7[link]b). The structures of the starting complex, intermediate species and long-lived final product are shown in Fig. 7[link](c). The technical challenges of the experiment are significant, but this can be justified by the high specificity of the XAS spectra for the electronic and molecular structure of the absorbing transition-metal cation, i.e. the catalytic centre of the molecule.

[Figure 7]

Figure 7

(a) Transmission EDXAS compared with step-scan XAS (40 min) measurements from a 35 mM solution of [Fe(TPA)(CH3CN)2]2+. (b) Time evolution of XAS of (a) following the addition of acetic acid. (c) Proposed structures of the intermediates formed during the reaction. Reprinted with permission from Olivo et al. (2017[link]). Copyright 2017 American Chemical Society.

3.6. Sequential application of complementary techniques

The simultaneous measurement of XAS and Raman/UV–Vis/IR spectroscopy will in many cases either be impossible or introduce complications into the measurement that cannot be justified in terms of the additional information. Issues related to sample stability, concentration, volume or sampling geometry may dictate that sample characterization be better managed through the sequential application of spectroscopic techniques. In this case the critical aspect of the experimental design is the generation of the sample in the required form, validation using an appropriate technique and controlled transfer to the sample holder ready for XAS measurement, most commonly with the sample maintained at cryogenic temperature.

The insights that can be gained from the application of such an approach relate to the reaction chemistry that develops following the reduction of diiron compounds which are structurally related to the active site, or H-cluster, of the [FeFe]-hydrogenase enzyme (Nicolet et al., 1999[link]; Peters et al., 1998[link], 2015[link]; Lubitz et al., 2014[link]). An intrinsic component of the H-cluster is the presence of CO and CN ligands strongly bound to the iron (De Lacey et al., 2000[link]); these groups give intense bands in the IR spectrum in a region that is largely devoid of protein absorptions. In fact, IR spectroscopy allowed the identification of CO and CN as components of the H-cluster before the determination of the X-ray structure (Pierik et al., 1998[link]). The stretching vibrations of both the CO and CN groups are sensitive to the oxidation state and local bonding interactions to the iron, and characteristic shifts in the wavenumbers of these bands are used to identify the different states of the H-cluster during turnover (Best, 2005[link]; Lubitz et al., 2014[link]). Small-molecule dithiolato-bridged diiron carbonyl model compounds have been shown to catalyse proton reduction and, owing to their higher symmetry, facile chemical modification and the prospect that they may provide useful catalysts, have become a subject of intense investigation in their own right. For both the parent compounds and the products derived therefrom, the ν(CO) bands in the IR spectrum provide a clear signature which can be used to identify the different products and product mixtures that are formed during reaction. Electrochemical or chemical reduction of the parent propanedithiolato-bridged diironhexa­carbonyl compound (3S) is marked by the generation of a short-lived one-electron reduction product which undergoes further reduction to give three distinct products which can be shown to give distinctive signatures in the IR stretching region (Fig. 8[link]; Borg et al., 2004[link]). Chemical reduction by hydride leads to initial attack at one of the carbonyl ligands and provides an alternate path to the comparatively stable end member [{\bf 3D}^{\bf red}_{\bf B}].

[Figure 8]

Figure 8

IR spectra in the ν(CO) region of 3S and products obtained by electrochemical or chemical reduction. The reduced species were generally obtained as mixtures and spectral subtraction was used to obtain the spectral signatures shown above. For [{\bf 3D}^{\bf red}_{\bf B}] the IR spectrum is shown from the fragment of the molecule enclosed in the box.

While IR spectroscopy provides a good signature of the speciation, the structure of the individual species can not reliably obtained from these measurements alone. As demonstrated in Section 3.3[link], for metal carbonyl complexes the combination of IR with XAS can provide the structural information needed to give unambiguous identification of the structure. The challenge in this case is the preparation of a sample with a well defined composition which can be maintained in that form over the duration of the XAS measurements. IR spectroscopy can be used to monitor the product stream from a flow reactor (chemical or electrochemical), where the solution flow can be diverted to an XAS solution cell when the desired solute composition is attained. Rapid freezing of the XAS solution cell in liquid nitrogen (LN) quenches the chemical reaction and provides a sample that can be transferred to the beamline cryostat. While this latter task is simple in principle, it is generally necessary to develop purpose-built sample holders that are suitable for transfer of the frozen sample to the cryostat sample stick. 3D printer technology can be used to produce highly optimized cells at low cost. CAD drawings of a system recently used in solution XAS measurements are shown in Fig. 9[link] (Streltsov et al., 2018[link]). The solution cell consists of a Kapton-covered slot with a square end section which is connected to 1/16′′ Teflon tubing using liquid-chromatography (LC) fittings incorporating ferrule seals, where the 1/4′′ UNF fittings are screwed into an aluminium block, allowing a press seal of the ferrule both against the cell body and the Teflon tubing (Fig. 9[link]a). Anaerobic transfer of the solution can be achieved using standard methods. Once the sample has been loaded into the sample space, the assembly is immersed in liquid nitrogen and, once equilibrated at LN temperature, the LC fittings can be removed and the solution cell released. While under LN the solution cells can be loaded into a holder connected to the sample stick of the beamline cryostat. The holder shown in Fig. 9[link] can accept three solution cells which are loaded through the bottom opening and made secure using a baseplate (Fig. 9[link]b). The sample cassette is then attached to the cryostat stick using the coupling shown in grey. The sample cassette allows simultaneous transmission and fluorescence measurements and a clear path is available above and below the set of three sample holders. While not a feature of the published studies, the availability of inexpensive 3D-printed sample cassettes can allow the preparation of multiple samples in advance of the experiment and can avoid the need to establish specialized sample-handling capabilities at the synchrotron.

[Figure 9]

Figure 9

Solution cell holder for low-temperature XAS measurements. (a) Schematic representation of the approach used for XAS measurements on IR-validated, electrochemically generated freeze-quenched samples under anaerobic conditions and (b) orthogonal CAD views of a three-position sample cell holder for simultaneous transmission and fluorescence XAS measurements.

Electrochemical generation and XAS measurements of chemically and electrochemically reduced 3S (Fig. 8[link]) were conducted on a bending-magnet beamline, BL18B, at the Photon Factory, where the lower flux contributed to the high stability of the spectra over repeated sample scans. A schematic diagram of the approach used to prepare IR-validated, electrochemically generated samples of reactive reduced species is shown in Fig. 9[link](a). The output of a flow electrochemical cell (Yeo et al., 2012[link]) can be directed to either an IR solution cell or the XAS cell. Initially, the output stream is directed to the IR cell and an N2 gas stream is used to purge the XAS cell. When the product is in the required form the output stream is diverted to the XAS cell and the sample is freeze-quenched in LN, where the XAS cell is maintained under LN until transfer to a suitable holder for insertion into the beamline cryostat (for example as shown in Fig. 9[link]b). A key requirement of the design of the experiment is to control the quantity of sample needed for sample generation. For approaches of the sort described a minimum sample volume of above 5 ml was required for electrochemistry, IR sample validation and XAS sample collection. Approaches involving XAS measurement from the solution in contact with a porous working electrode either in the frozen (Yeo et al., 2012[link]) or solution (Best et al., 2016[link]) state can allow the sample volume to be decreased to ∼1 ml; however, in these cases the state of the solute is only able to be monitored using electrochemical and XAS methods.

EXAFS analysis of complex molecular species is more effective in providing metric data on particular scattering interactions than in giving reliable structural assignments. This limitation can be resolved by the integration of IR and EXAFS with density-functional theory calculations of the alternate proposed structures (Borg et al., 2007[link]; Bondin et al., 2006[link]). The complementarity of the three techniques ensures that there is effective triangulation to the assigned structure. The application of this strategy and the use of the resulting structural information to better understand the details of electrocatalytic proton reduction has been demonstrated for ethanedithiolato- and propanedithiolato-bridged diiron compounds (Borg et al., 2007[link]; Cheah & Best, 2011[link]).

4. Summary and concluding remarks

While the case for the development of experimental programs which make full, or at least an appropriate, use of coupled or simultaneous measurements using appropriate techniques is clear, the logistics associated with X-ray measurements can present significant impediments. These are associated with the physical layout of the experiment, the dimensions and location of the detectors and the X-ray beam, the time required for measurement of XAS spectra, the penetration depth of the radiation, limited access to the experimental safety hutch and the intense time pressures associated with XAS measurements. A significant investment of human and beamline resources is needed to allow successful measurements, and this explains the comparatively narrow base of successful experiments reported in the literature. Advances in beamline automation, together with improvements in fibre-optic coupled spectroscopies (principally Raman and UV–Vis) should decrease the threshold for a broadening of this base. Further, as demonstrated by the examples discussed in this contribution, there are established solutions to many of these problems and there are strong arguments for a carefully considered suite of ancillary methods to be available at XAS beamlines. In terms of the time resolution of the measurement, it is important to note that EDXAS and rapid-scan XAS measurements can allow the study of reacting systems, where the prize of these studies is understanding the structural changes along the reaction coordinate. Of course, these studies must focus on relatively stable intermediates, but advances in the theory and practice of XAS will continue to increase the range of problems that are able to be usefully studied.


First citationAgistini, G., Gianolio, D. & Lamberti, C. (2023). Int. Tables Crystallogr. I, .Google Scholar
First citationBartholomew, A. K., Teesdale, J. J., Hernández Sánchez, R., Malbrecht, B. J., Juda, C. E., Ménard, G., Bu, W., Iovan, D. A., Mikhailine, A. A., Zheng, S.-L., Sarangi, R., Wang, S. G., Chen, Y.-S. & Betley, T. A. (2019). Proc. Natl Acad. Sci. USA, 116, 15836–15841.Google Scholar
First citationBauer, M., Heusel, G., Mangold, S. & Bertagnolli, H. (2010). J. Synchrotron Rad. 17, 273–279.Google Scholar
First citationBeale, A. M., van der Eerden, A. M. J., Kervinen, K., Newton, M. A. & Weckhuysen, B. M. (2005). Chem. Commun. 2005, 3015–3017.Google Scholar
First citationBest, S. P. (2005). Coord. Chem. Rev. 249, 1536–1554.Google Scholar
First citationBest, S. P., Levina, A., Glover, C., Johannessen, B., Kappen, P. & Lay, P. A. (2016). J. Synchrotron Rad. 23, 743–750.Google Scholar
First citationBoccaleri, E., Carniato, F., Croce, G., Viterbo, D., van Beek, W., Emerich, H. & Milanesio, M. (2007). J. Appl. Cryst. 40, 684–693.Google Scholar
First citationBondin, M. I., Borg, S. J., Cheah, M. H., Foran, G. & Best, S. P. (2006). Aust. J. Chem. 59, 263–272.Google Scholar
First citationBorg, S. J., Behrsing, T., Best, S. P., Razavet, M., Liu, X. & Pickett, C. J. (2004). J. Am. Chem. Soc. 126, 16988–16999.Google Scholar
First citationBorg, S. J., Tye, J. W., Hall, M. B. & Best, S. P. (2007). Inorg. Chem. 46, 384–394.Google Scholar
First citationBras, W., Derbyshire, G. E., Bogg, D., Cooke, J., Elwell, M. J., Komanschek, B. U., Naylor, S. & Ryan, A. J. (1995). Science, 267, 996–999.Google Scholar
First citationBriois, V., Lützenkirchen-Hecht, D., Villain, F., Fonda, E., Belin, S., Griesebock, B. & Frahm, R. (2005). J. Phys. Chem. A, 109, 320–329.Google Scholar
First citationBriois, V., Vantelon, D., Villain, F., Couzinet, B., Flank, A.-M. & Lagarde, P. (2007). J. Synchrotron Rad. 14, 403–408.Google Scholar
First citationBryant, G. K., Gleeson, H. F., Ryan, A. J., Fairclough, J. P. A., Bogg, D., Goossens, J. G. P. & Bras, W. (1998). Rev. Sci. Instrum. 69, 2114–2117.Google Scholar
First citationCheah, M. H. & Best, S. P. (2011). Eur. J. Inorg. Chem. 2011, 1128–1137.Google Scholar
First citationDavies, R. J., Burghammer, M. & Riekel, C. (2005). Appl. Phys. Lett. 87, 264105.Google Scholar
First citationDe Lacey, A. L., Stadler, C., Cavazza, C., Hatchikian, E. C. & Fernandez, V. M. (2000). J. Am. Chem. Soc. 122, 11232–11233.Google Scholar
First citationEinsle, O., Andrade, S. L. A., Dobbek, H., Meyer, J. & Rees, D. C. (2007). J. Am. Chem. Soc. 129, 2210–2211.Google Scholar
First citationFrahm, R., Richwin, M. & Lützenkirchen-Hecht, D. (2005). Phys. Scr. 2005, 974.Google Scholar
First citationFrenkel, A. I., Rodriguez, J. A. & Chen, J. G. G. (2012). ACS Catal. 2, 2269–2280.Google Scholar
First citationLubitz, W., Ogata, H., Rüdiger, O. & Reijerse, E. (2014). Chem. Rev. 114, 4081–4148.Google Scholar
First citationMüller, R. J., Lan, J. G., Lienau, K., Moré, R., Triana, C. A., Iannuzzi, M. & Patzke, G. R. (2018). Dalton Trans. 47, 10759–10766.Google Scholar
First citationNayak, C., Bhattacharyya, D., Bhattacharyya, K., Tripathi, A. K., Bapat, R. D., Jha, S. N. & Sahoo, N. K. (2017). J. Synchrotron Rad. 24, 825–835.Google Scholar
First citationNayak, C., Bhattacharyya, D., Jha, S. N. & Sahoo, N. K. (2016). J. Synchrotron Rad. 23, 293–303.Google Scholar
First citationNayak, C., Bhattacharyya, D., Jha, S. N. & Sahoo, N. K. (2018). ChemistrySelect, 3, 7370–7377.Google Scholar
First citationNewton, M. A. (2009). Top. Catal. 52, 1410–1424.Google Scholar
First citationNewton, M. A., Dent, A. J., Fiddy, S. G., Jyoti, B. & Evans, J. (2007a). Catal. Today, 126, 64–72.Google Scholar
First citationNewton, M. A., Dent, A. J., Fiddy, S. G., Jyoti, B. & Evans, J. (2007b). J. Mater. Sci. 42, 3288–3298.Google Scholar
First citationNewton, M. A., Jyoti, B., Dent, A. J., Fiddy, S. G. & Evans, J. (2004). Chem. Commun. 2004, 2382–2383.Google Scholar
First citationNewton, M. A. & van Beek, W. (2010). Chem. Soc. Rev. 39, 4845–4863.Google Scholar
First citationNicolet, Y., Piras, C., Legrand, P., Hatchikian, C. E. & Fontecilla-Camps, J. C. (1999). Structure, 7, 13–23.Google Scholar
First citationOlivo, G., Barbieri, A., Dantignana, V., Sessa, F., Migliorati, V., Monte, M., Pascarelli, S., Narayanan, T., Lanzalunga, O., Di Stefano, S. & D'Angelo, P. (2017). J. Phys. Chem. Lett. 8, 2958–2963.Google Scholar
First citationPascarelli, S. & Mathon, O. (2023). Int. Tables Crystallogr. I, .Google Scholar
First citationPascarelli, S., Mathon, O., Mairs, T., Kantor, I., Agostini, G., Strohm, C., Pasternak, S., Perrin, F., Berruyer, G., Chappelet, P., Clavel, C. & Dominguez, M. C. (2016). J. Synchrotron Rad. 23, 353–368.Google Scholar
First citationPatlolla, A., Carino, E. V., Ehrlich, S. N., Stavitski, E. & Frenkel, A. I. (2012). ACS Catal. 2, 2216–2223.Google Scholar
First citationPeters, J. W., Lanzilotta, W. N., Lemon, B. J. & Seefeldt, L. C. (1998). Science, 282, 1853–1858.Google Scholar
First citationPeters, J. W., Schut, G. J., Boyd, E. S., Mulder, D. W., Shepard, E. M., Broderick, J. B., King, P. W. & Adams, M. W. W. (2015). Biochim. Biophys. Acta, 1853, 1350–1369. Google Scholar
First citationPierik, A. J., Hulstein, M., Hagen, W. R. & Albracht, S. P. J. (1998). Eur. J. Biochem. 258, 572–578.Google Scholar
First citationRan, S. F., Fang, D. F., Sics, I., Toki, S., Hsiao, B. S. & Chu, B. (2003). Rev. Sci. Instrum. 74, 3087–3092.Google Scholar
First citationRochet, A., Baubet, B., Moizan, V., Devers, E., Hugon, A., Pichon, C., Payen, E. & Briois, V. (2017). J. Phys. Chem. C, 121, 18544–18556.Google Scholar
First citationSamant, M. G., Bergeret, G., Meitzner, G. & Boudart, M. (1988). J. Phys. Chem. 92, 3542–3546.Google Scholar
First citationSamant, M. G., Bergeret, G., Meitzner, G., Gallezot, P. & Boudart, M. (1988). J. Phys. Chem. 92, 3547–3554.Google Scholar
First citationSmolentsev, G., Guilera, G., Tromp, M., Pascarelli, S. & Soldatov, A. V. (2009). J. Chem. Phys. 130, 174508.Google Scholar
First citationStötzel, J., Lützenkirchen-Hecht, D., Frahm, R., Santilli, C. V., Pulcinelli, S. H., Kaminski, R., Fonda, E., Villain, F. & Briois, V. (2010). J. Phys. Chem. C, 114, 6228–6236.Google Scholar
First citationStreltsov, V. A., Ekanayake, R. S. K., Drew, S. C., Chantler, C. T. & Best, S. P. (2018). Inorg. Chem. 57, 11422–11435.Google Scholar
First citationVan der Snickt, G., Janssens, K., Dik, J., De Nolf, W., Vanmeert, F., Jaroszewicz, J., Cotte, M., Falkenberg, G. & Van der Loeff, L. (2012). Anal. Chem. 84, 10221–10228.Google Scholar
First citationVantelon, D., Lagarde, P., Flank, A. M., Berrier, E., Secordel, X., Cristol, S., La Fontaine, C., Villain, F. & Briois, V. (2009). Phase Transit. 82, 322–335.Google Scholar
First citationYeo, J., Cheah, M. H., Bondin, M. I. & Best, S. P. (2012). Aust. J. Chem. 65, 241–253.Google Scholar
First citationZalden, P., Aquilanti, G., Prestipino, C., Mathon, O., André, B., Wuttig, M. & Coulet, M.-V. (2012). J. Synchrotron Rad. 19, 806–813.Google Scholar

to end of page
to top of page