Detailed understanding of electrochemical reactions at the electrode–electrolyte interface is crucial not only for fundamental electrochemical research, but also for the development of electrochemical devices such as fuel cells, batteries, electrolyzers, supercapacitors and so on (Simon & Gogotsi, 2008; Debe, 2012; Chen et al., 2013). In situ XAS techniques offer a unique method to investigate the local atomic structures, electronic levels and chemical states of the surface under controlled electrochemical conditions using appropriate electrochemical cells (Russell & Rose, 2004; Sasaki & Adzic, 2009).

When designing a three-electrode electrochemical cell for in situ XAS studies, one must consider several points to fulfil the experiments adequately (Sharpe et al., 1990; Russell & Rose, 2004; Watanabe et al., 2007). Some of the requirements are described below.

(i) A relevant amount of specimen material must be loaded to obtain a good signal-to-noise ratio; this depends on whether transmission or fluorescence mode is selected.

(ii) The cell must incorporate windows that are transparent to X-rays and both must be resistant to corrosion by the electrolyte.

(iii) The electrode configuration must (1) ensure a uniform distribution of current density flowing between the working and counter electrodes and (2) minimize the iR (ohmic) drop between the working and reference electrodes.

(iv) The X-ray path through the cell must be minimized to reduce its absorption by the electrolyte since water is an attenuator of low-energy X-rays. This point must be taken into account in both transmission and fluorescence modes when a reference foil is simultaneously measured to calibrate the X-ray energy.

(v) Mass transport of reactants and products must be accounted for in order not to inhibit the process being studied. This factor also applies to processes at the counter electrode, where gas evolution can occur. Generally, the cell design requires provisions for the supply and exit of a gas to purge the electrolyte or assist mass transport.

(vi) The cell must be compact to allow its orientation and alignment to be set up easily.

(vii) All wire contacts must be secured to the cell.

Many electrochemical cells for in situ XAS have been developed. In this chapter, the focus will mainly be on electrochemical cells employed for electrocatalytic studies.

XAS measurements are made in either transmission or fluorescence mode. The former is simpler and usually offers a higher signal-to-noise ratio than the latter. However, the selection depends on other factors such as sample loading, the X-ray energy range in which the target material is examined and the quantity of solution in the X-ray path in an electrochemical cell. We first describe electrochemical cells with ample electrolyte solution (termed `solution cells' here). Owing to the large quantity of solution, measurements are confined to fluorescence mode, which is sensitive to dilute samples (as discussed in the following sections, transmission mode requires concentrated samples).

An example of a solution cell with its working electrode is shown in Fig. 1 (Sasaki et al., 2016). The working electrode consists of a carbon-supported platinum-nanoparticle (Pt/C) catalyst embedded in a carbon-fibre paper that acts as the catalyst support. The loading was approximately 30 µg of platinum per cm². The oxidation/dissolution behaviour of platinum was monitored up to 2.6 V (reversible hydrogen electrode). Gas evolved on the catalyst surfaces can easily be released upwards through the solution and within the carbon paper and does not interfere with the fluorescence signal. The working electrode is placed just behind a thin polyethylene Mylar film (10 µm) that covers a window of the cell. The electrode is in contact with bulk solution (for example ∼50 ml 0.1 M HClO₄) in a plastic container. The incident X-ray beam is directed onto the sample through the Mylar film and a very thin layer of the electrolyte, and the resultant fluorescence intensity is monitored by a passivated implanted planar silicon (PIPS) detector. This cell was designed based on a cell used to study changes in chromium oxidation states on an Al–Cr alloy film under electrochemical control (Davenport et al., 1991). A sample material (for example Al–Cr alloy on a titanium film) sputter-deposited onto a Mylar film was subjected to an electrolyte, and thus no absorption between the Mylar film and the sample took place.

Figure 1 Schematic of an electrochemical cell and electrode configuration. Reprinted with permission from Sasaki et al. (2016). Copyright 2016 American Chemical Society.

Asakura and coworkers developed a solution cell with a different electrode configuration, in which an extremely dilute platinum-nanoparticle sample (∼0.17 ng of platimum per cm²) was deposited onto a highly oriented pyrolytic graphite (HOPG) substrate (Uehara et al., 2014). The back side of the thin HOPG layer acted as a window for X-ray illumination, while the catalyst side faced an electrolyte solution (0.1 M HClO₄). In the configuration with ultralow platinum loading, however, X-ray scattering from the solution may mask fluorescence signals from the target material; to mitigate this problem, a bent crystal Laue analyzer (BCLA) was used, which removes most of the undesired scattering before it reaches the detector (Zhong et al., 1999; Karanfil et al., 2012). Jaramillo and coworkers investigated the oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) of a bifunctional MnO_x catalyst by measuring in situ XAS at the Mn K edge using a similar solution cell (Gorlin et al., 2013); in their setup, the back side of the Si₃N₄ window faced the X-rays and the front side of the window with electrodeposited MnO_x on a layer of Au/Ti was in contact with an alkaline electrolyte solution (0.1 M potassium hydroxide).

Salmeron, Guo and coworkers developed an electrochemical setup for in situ XAS using soft X-rays (150–2000 eV) to investigate the electronic structure of electrode materials under controlled electrochemical conditions (Fig. 2; Jiang et al., 2010). In the soft X-ray region, features of the L_2,3 edges of 3d metals and the M_2,3 edges of 4s metals can be measured. As the absorption by cross section of most gases and liquids is large in such a low-energy region, the use of ultra-high vacuum (UHV) instrumentation is required. A thin Si₃N₄ membrane (∼100 nm; labelled a in Fig. 2) was used to separate the solution cell from the UHV environment. On the cell side of the membrane, a thin copper film (∼300 nm; labelled b in Fig. 2) was deposited as a working electrode. The thickness of the copper electrode was optimized to allow penetration of the incident X-rays and escape of the fluorescence signal. The electrolyte flowed through two holes located at the back of the body (green arrow in Fig. 2). By measuring XAS signals from the Cu L₃ and L₂ edges (932.7 and 952.3 eV, respectively), the oxidation behaviour of copper in aqueous NaHCO₃ solution under potential cycles was demonstrated. Soft XAS can also provide insight into surface intermediates and their geometry on the surface by probing the K edges of 2p elements such as C, N and O (284.2, 409.9 and 543.1 eV, respectively). In situ soft XAS using the same cell allows identification of the structure of water near gold electrodes and its bias dependence (Velasco-Velez et al., 2014).

Figure 2 Schematic of an electrochemical cell assembly for in situ soft XAS studies. (a) Si3N4 window, (b1) electrical connection to Si3N4 window (working electrode), (b2) reference electrode, (b3) counter electrode, (c) PEEK body, (d) support tube assembly. The green arrow indicates the liquid flow. Reprinted with permission from Jiang et al. (2010). Copyright 2010 Elsevier.

Another type of cell utilizes a thin-layer geometry to minimize the absorption by an electrolyte and thus allow transmission-mode measurements to be made (termed `thin solution-layer cells'). Fig. 3 presents an exploded view of such a thin-layer electrochemical cell for in situ XAS experiments (Sasaki et al., 2010), which is a modified version of the cell designed by McBreen et al. (1987). A catalyst paste made of a catalyst (for example Pt/C) and a small amount of solution is placed uniformly on a carbon cloth. For measurements of the Pt L₃ edge spectrum, the total amount of platinum in the electrode is approximately 5 mg cm⁻² to obtain a good signal-to-noise ratio. The carbon cloth with the catalyst, a proton-exchange membrane and two PTFE gaskets are sandwiched as shown in the figure, and all of the cell components are clamped tightly by two acrylic plastic bodies with an O-ring. Each plastic body has a X-ray window that is glued by a thin acrylic film. A platinum foil on the thicker plastic body acts as a counter electrode, while a platinum ribbon on the thinner plastic body achieves electrical contact for the electrocatalyst (working electrode) through the carbon cloth. An electrolyte (for example 1 M HClO₄) is added to the cell. The quantity of solution has a maximum of 2 ml and the thickness of the compartment housing the catalyst, the membrane and the solution (which is the distance of the X-ray path though the cell) is ∼2 mm. An Ag/AgCl leak-free reference electrode is inserted in a capillary drilled in the thicker plastic block. This electrochemical cell is designed for XAS data acquisition in both transmission and fluorescence modes, although transmission mode is preferred for concentrated samples. The electrochemical cell also allows us to measure X-ray diffraction (XRD) at controlled electrochemical potentials. At high potentials (above 1.6 V), however, oxygen-evolution and hydrogen-evolution reactions may occur on the working and counter electrodes, respectively, and gas evolution induces physical motion of the catalyst layer inside the cell (as the cell components with electrolyte are tightly clamped by two plastic bodies to minimize absorption by the electrolyte), resulting in a serious deterioration in signal quality.

Figure 3 Schematic diagram of an electrochemical cell used for in situ XAS. Reprinted with permission from Sasaki et al. (2010). Copyright 2010 Elsevier.

Also, based on the McBreen cell, a new flowthrough-type electrochemical cell was fabricated (Arruda et al., 2010); this cell configuration is advantageous as it allows the user to oxygenate/deoxygenate the electrolyte or introduce poisons without the need to disassemble the cell during the experiment. Another flowthrough-type cell has been reported to monitor the reduction and oxidation of iron corrosion precipitates at different pH values by measuring Fe K edge XANES spectra in transmission mode (Monnier et al., 2014).

Crooks and coworkers have assembled an electrochemical cell which can provide continuous gas purging in an electrolyte solution without interfering with XAS measurements by the formation of bubbles either on the electrode or near the X-ray window (Weir et al., 2010). Using this in situ XAS cell, they investigated the structural properties of dendrimer-encapsulated platinum nanoparticles during the CO oxidation reaction.

One of limitations of the electrochemical cells described above is that it is difficult to study a single-crystal surface using these cells. The cell shown in Fig. 4 has a unique electrode configuration that provides the ability to perform in situ measurements on planar model electrocatalysts prepared under UHV conditions (Renner et al., 2007; Friebel et al., 2011; Merte et al., 2012). In these cell designs, the electrolyte is brought into contact with the front surface of a single crystal with/without adatoms via a meniscus formed by a droplet from a capillary positioned just above the surface (the gap is typically ∼2 mm). This cell is termed a `meniscus solution cell' here. The counter and reference electrodes are in contact with the electrolyte solution upstream of the capillary and a stable meniscus is possible with flowing electrolyte. XAS measurements are made by directing the incident X-ray beam at grazing incidence to the electrocatalyst surface and collecting X-ray fluorescence (arrows in Fig. 4). Nilsson and coworkers explored the atomic structures and oxidation behaviour of a platinum monolayer or nanoclusters on a single-crystal surface such as Rh(111) or Au(111) using this cell (Friebel et al., 2011; Merte et al., 2012).

Figure 4 Schematic drawing showing a meniscus-type cell for performing in situ XAS measurements on planar model electrocatalysts. Reprinted with permission from Merte et al. (2012). Copyright 2012 American Chemical Society.

Different types of cells have also been developed to investigate electrochemical reactions and the resultant adsorbed species on a single crystal or polycrystalline surface by in situ X-ray techniques. A side view of a cell employed by Tamura and coworkers for the structural study of electrochemically deposited copper on p-GaAs(100) is illustrated in Fig. 5 (Tamura et al., 2000). The cell was sealed with a thin Mylar film (6 µm), apart from a tube connected to a syringe filled with 0.1 M H₂SO₄ solution containing 0.1 mM CuSO₄. By pushing the syringe plunger in, the solution was introduced into the cell and inflated the polyethylene film (the motion is illustrated from A to B at the syringe and the Mylar window in Fig. 5). Copper was then electrodeposited onto the GaAs(100) surface by applying a potential of −0.5 V versus Ag/AgCl for a certain time period, while the thickness of the electrolyte between the electrode surface and the window was maintained at more than 5 mm. The coverage of the copper layer on the electrode surface can be controlled by varying the deposition time. To terminate the deposition, the potential was changed to −0.35 V versus Ag/AgCl, and at the same time the solution was drained off from the cell to deflate the membrane, leaving a thin electrolyte layer (less than several tens of micrometres), by pulling the plunger out. XAS measurements were then made in fluorescence mode with grazing-incidence X-ray geometry. The thin electrolyte-layer geometry can minimize the scattering of X-rays by the solution. Repeated deposition–interruption–XAS measurement cycles permit detailed characterization of the structure of electrodeposited copper layers with various coverages to be carried out.

Figure 5 Side view of an inflated/deflated solution cell for grazing-angle XAS measurements. Reprinted with permission from Tamura et al. (2000). Copyright 2002 American Chemical Society.

These cells are termed `inflated/deflated solution cells', the principle of which was originally proposed by Toney et al. (1992). Similar cells have been employed for in situ XANES measurements to monitor the formation of chromate conversion coatings on aluminium and its alloys (Isaacs et al., 2005) and in situ grazing-incidence EXAFS/XRD to examine the structure of underpotentially deposited (UPD) copper and cadmium layers on a Pt(533) single crystal (Prinz & Strehblow, 2002), as well as in situ X-ray scattering studies to investigate Au(111)/electrolyte interfaces (Wang et al., 1992) and bismuth adlayers on Au(111) (Tamura et al., 2002).

Platinum or platinum-based nanoparticles are known to be the most efficient catalysts for both the anodic hydrogen oxidation reaction (HOR) and the cathodic oxygen reduction reaction (ORR) in low-temperature proton-exchange membrane fuel cells (PEMFCs, also known as polymer electrolyte membrane fuel cells). The core of the PEMFC is a membrane electrode assembly (MEA), in which a membrane is sandwiched between anode and cathode catalytic layers. In situ XAS measurements using a single MEA fuel cell under realistic operating conditions were first demonstrated in 2002 (Roth et al., 2002; Viswanathan, Hou et al., 2002; Viswanathan, Liu et al., 2002). Fig. 6 depicts a schematic and a photograph of the fuel cell for in situ XAS designed by Smotkin and coworkers (Viswanathan, Hou et al., 2002; Viswanathan, Liu et al., 2002). The graphite blocks on each side of the cell were machined to provide flow channels for gases, and a rectangular hollow was milled behind the flow channels to provide a path for the X-ray beam. The two graphite blocks were clamped tightly to provide a sufficiently high compression on the MEA and a good electrical contact. The fuel cell was operated with humidified pure H₂ or reformate gas fed to the anode and humidified air or O₂ gas fed to the cathode, and in situ X-ray absorption data were obtained in transmission mode at various gas flow rates, operating temperatures and cell voltages. The measurement requires no liquid solution, so the problems associated with absorption by electrolytes and interference by the evolution of bubbles can be avoided.

Figure 6 Schematics and photograph of a single MEA fuel cell designed by Smotkin and coworkers. Reprinted with permission from Viswanathan, Hou et al. (2002) and Viswanathan, Liu et al. (2002). Copyright 2002 Elsevier and AIP.

However, the following two points must be taken into account. Firstly, the use of transmission mode requires a minimum loading of catalyst in the MEA (Wiltshire et al., 2005). Lower loadings reduce the signal-to-noise ratio, thus making EXAFS analysis more difficult. To gain the optimal quality of transmission data, the change in absorption at the edge should be at least 0.3, which requires 1.7 mg of platinum per cm² at the Pt L₃ edge (Wiltshire et al., 2005). We note that typical platinum loadings for real operating fuel cells are ∼0.1–0.4 mg of platinum per cm², and thus higher loadings are necessary to obtain reasonably good data from XAS transmission measurements. However, an increase in loading may lower the utilization of the catalyst, which leads to misinterpretation since the number of electrochemically unreacted nanoparticles in the catalyst layer increases. In order to keep the catalyst loading as low as appropriate while maintaining a satisfactory signal-to-noise ratio, the fluorescence acquisition mode can be operated with a slight modification of the MEA cells (Roth et al., 2002; Wiltshire et al., 2005; Principi et al., 2007, 2009).

Secondly, transmission mode simultaneously measures the catalysts on both the anode and cathode sides of the MEA; for instance, if an XAS investigation focuses on a platinum catalyst at an anode or cathode, the material at the other electrode should be a nonplatinum catalyst without significant influence on the absorption of the target platinum catalyst. For instance, a carbon-supported palladium nanoparticle (Pd/C) catalyst can be used at the other electrode since palladium is reasonably active in both the ORR and the HOR and its absorption coefficient is smooth around the Pt L₃ edge (Viswanathan et al., 2002; Wiltshire et al., 2005; Tada et al., 2007; Principi et al., 2009; Ishiguro et al., 2012). An alternative way to avoid this problem is to remove part of the platinum catalyst at the other side of the target electrode in a beam window region in order to confine the signal contribution to the target electrode side (Roth et al., 2002, 2005). However, care must be taken since such removal of the catalyst may modify the current distribution in the region of the target catalyst probed by the X-rays.

We have given a brief overview of electrochemical cells for in situ XAS studies in this chapter, together with the advantages and disadvantages of the existing cells and some requirements for new cell designs to perform experiments. In situ XAS with appropriate electrochemical cells provides information on the atomic structures, oxidation/electronic states and adsorbate species of the catalytic surface during the course of an electrochemical reaction, and is thus an indispensable tool for electrocatalytic/corrosion studies. The data that it furnishes can enhance our understanding of the mechanisms of electrocatalytic reactions at surfaces and may provide a rational approach to the design of new highly active and corrosion-resistant materials. With the development of new synchrotrons worldwide, which provide high photon fluxes and small beam sizes, the in situ technique has the promise to become widespread in the scientific community.