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

Liquids and gels

Andrea Di Ciccoa*

aSezione di Fisica, Scuola di Scienze e Tecnologie, Università di Camerino, Via Madonna delle Carceri 9, Camerino, Italy
Correspondence e-mail:

The main techniques used to obtain accurate X-ray absorption fine-structure data for liquids and gels are discussed, with an emphasis on the methods used to prepare optimal samples, environmental conditions and data-collection strategies. Selected examples of experiments, including data-acquisition setup and sample design, are briefly reviewed.

Keywords: liquids; gels; samples; EXAFS; temperature; pressure.

1. Introduction

This chapter is devoted to the application of X-ray absorption fine structure (XAFS) to liquids and gels, with the specific aim of briefly describing some techniques for producing samples of adequate quality, as well as the current strategies for collecting reliable data and realizing the necessary sample-environment conditions. Liquids and gels are obviously different aggregation states of matter and call for different strategies in the preparation of XAFS experiments. By definition, a macroscopic liquid is a fluid that conforms to the shape of its container due to its limited viscosity, showing a density near to that of solid-state matter and very limited compressibility. It is one of the fundamental states of matter, and therefore simple substances can be found in thermal equilibrium in selected portions of their phase diagrams. While some substances can be found in liquid form under ordinary conditions (water is the most common example), most need special conditions of pressure and temperature. Both the fluid nature of the sample and the possible need for special thermodynamic conditions represent challenges for successful XAFS experiments. Gels, on the other hand, are dilute cross-linked systems that exhibit negligible flow. Gels can be considered to be mostly liquid, but they behave like solids due to a three-dimensional cross-linked network within the liquid. Gels can thus be considered to be microscopically inhomogeneous systems in which molecules of a liquid are dispersed within a solid, forming a three-dimensional network. Usually, most physical properties of gels can be measured in a restricted region of temperature and pressure, considerably simplifying sample design and the experimental setup. In all cases, particular care must be devoted to the quality of liquid and gel samples and to the sample environment, which in some cases may include the special development of cells for use to confine liquids, maintaining suitable controlled conditions. Details of appropriate experimental techniques used for measuring liquids (and gels) can be found in Section 9.4 of Crozier et al. (1988[link]) and in Filipponi (2001[link]).

2. XAFS of liquids and gels

XAFS spectroscopy is applied to liquid and gels, providing unique information about the short-range structure around selected chemical species in the given substance. Near-edge spectroscopy (X-ray absorption near-edge structure; XANES) is also used to study local coordination and oxidation states. Local structural information is contained in the XAFS structural signal χ(E) = [α(E) − α0(E)]/α0(E), which is the modulation of the X-ray absorption coefficient α(E) around the smooth atomic absorption α0(E). For typical liquids and gels, the χ(E) amplitude for a given atomic edge EE can be limited to less than 1%, reaching values of around 10−3–10−4 at high (50–1000 eV) photoelectron kinetic energies Ekin = E − EE, in the so-called extended X-ray absorption fine-structure (EXAFS) regime.

Due to the limited signal intensity, XAFS measurements typically have to be performed in transmission mode, for which the photon flux provided at modern beamlines and current detection technologies are able to push signal-to-noise ratios into the 104–105 range. For XANES measurements and highly dilute samples XAFS measurements can be performed in fluorescence mode, in general offering lower signal-to-noise performances. A typical setup for measuring XAFS of liquid samples is shown schematically in Fig. 1[link], including different detection modes, sample-environment devices and combined detection of X-ray diffraction, as used to check sample status and possible phase transformations, as described, for example, in Filipponi et al. (2000[link]). Modern energy-scanning beamlines at synchrotron-radiation facilities, equipped with fixed-exit double-crystal monochromators, currently offer a variety of detection modes, optics for beam focusing, cells for realizing suitable sample environments and combined X-ray diffraction data collection. Energy-dispersive beamlines are efficiently used mostly to collect XANES data of liquid samples under rapidly changing and/or extreme conditions, due to their time resolution, which can be lower than 1 ms.

[Figure 1]

Figure 1

(a) Sketch of a typical XAFS energy-scanning synchrotron-radiation beamline, including several important elements for accurate measurements of liquids. Suitable optical elements (optics) can be used to focus the beam on the sample in region (A) (modern beamlines may allow spots of micrometric size). The sample environment (A) includes devices for in situ measurements under variable temperature–pressure conditions. A suitable reference sample (B) under ambient conditions, typically containing edges relevant to the XAFS experiment on sample (A), is measured simultaneously to keep control of tiny energy variations of the energy calibration. X-ray detectors (typically ionization chambers) are used to measure the incoming (I0) and the transmitted (I1 and I2) photon flux. Diffraction patterns are collected at fixed (or variable) energy by means of an X-ray detector (C) covering a range of scattering angles. In the case of dilute samples, X-ray fluorescence is collected by a detector (D). (b) Left: Cu K-edge X-ray absorption data of liquid copper (T = 1400 K). The edge energy EE is indicated by the arrow and dotted line, while the pre-edge background absorption is shown as a dashed line. In the lower left panel the estimated standard deviation of the XAFS data is reported. Right: XAFS signal kχ(k) of liquid copper as a function of the wavevector k (top). In the lower left panel we report both the Fourier transform (FT) of the kχ(k) signal and the reconstructed pair distribution function g(r).

In Fig. 1[link] we report typical XAFS data for liquid copper as an example, showing some of the requirements for reliable studies of these disordered systems at modern synchrotron-radiation facilities. This experiment (Di Cicco et al., 2003[link]) was performed in transmission mode using an energy-scanning beamline (BM29 at ESRF; Filipponi et al., 2000[link]), and particular care was taken to obtain high-quality low-noise XAFS data by optimizing the sample preparation and environment. As shown in Fig. 1[link], the absorption discontinuity Δα at the Cu K edge (see arrow) was about 1.5, and the noise level (the inverse of the signal-to-noise ratio) was around 10−4 in the entire XAFS energy range (and was slightly higher in the region of higher absorption due to the decreased counting statistics). The resulting regular oscillations of the XAFS structural signal χ(k) are shown in the upper right panel {ℏk = [2m(Ekin)]1/2 mimics the photoelectron momentum} and the need for extremely low-noise data is evident. Static structural information in dis­ordered systems is efficiently expressed in terms of pair and higher-order distribution functions, and the XAFS reconstruction of the radial distribution g(r) of liquid copper is shown in the lower right panel. This reconstructed g(r) was obtained by combining diffraction with XAFS data (Di Cicco et al., 2003[link]), with the latter being particularly sensitive to the first-neighbour distribution. The Fourier transform pattern (dashed line) of the kχ(k) oscillations shows a pronounced peak related to the first neighbour (not phase-shift corrected in the figure) and ripples of very low intensity at longer distances. Its shape roughly resembles the first-neighbour radial distribution. As shown by this example, XAFS provides important short-range information in liquids that is complementary to that obtained by X-ray or neutron diffraction techniques.

3. Sample design for XAFS experiments

While most modern beamlines can usually provide a suitable setup for accurate XAFS measurements, the realization of liquid samples requires a specific discussion. The sample design is the most critical aspect to be considered in order to obtain reliable XAFS data in transmission mode for liquid-like systems. In fact, homogeneous samples of uniform thickness are needed for XAFS measurements in transmission mode, and typical sample thicknesses for optimal absorption discontinuity of core levels in the 1–30 keV range are in the range 1–20 µm. Due to their shape stability, gel-like samples are produced using the standard prescriptions used for solid specimens described elsewhere. Different strategies for sample design in liquids are used according to the sample nature and reactivity, the chemical species and core levels involved, and the need for, or the possibility of, variable temperature and pressure conditions.

Liquid solutions containing selected atomic species can often be measured in transmission mode using suitable cells of variable thickness in the millimetre range, depending on the actual solute concentration. For aqueous solutions and gels this can be performed in a wide photon range above 5 keV (covering a variety of core-level energies). Various cell designs compatible with the sketch reported in Fig. 2[link](a) have been widely used in the literature, as shown, for example, in Ertel & Bertagnolli (1993[link]) and Marcos et al. (1994[link]). Different materials for cell windows (Kapton, Teflon and glass) have been used according to the particular experiment under consideration. Solutions are naturally homogeneous and usually provide excellent samples for transmission measurements when the concentration is high enough (care must be taken with regard to possible phase separation, chemical reactions and radiation damage, especially for biological/protein solutions). Concentrations down to the millimolar range are routine, with fluorescence detection being used for very dilute solutions. A sample design substantially similar to that in Fig. 2[link](a) was used to perform several XAFS experiments on metallic fluids at high temperatures and pressures (up to 2000 K and 0.2 GPa). The setup (Tamura & Inui, 2001[link]) included an autoclave system for pressurizing the system, a heater and sapphire windows confining the liquid, with a typical sample thickness of greater than 30–50 µm. This technique was extremely useful for measuring expanded and supercritical fluids (Tamura & Inui, 2001[link]).

[Figure 2]

Figure 2

(a) Typical XAFS sample design for liquid solutions. A thin, uniform layer of the liquid specimen is confined by windows of an inert, low X-ray-absorbing material. The green arrow represents the X-ray beam. (b) Scanning electron micrograph of submicrometric liquid-metal droplets (liquid gallium) deposited on a suitable substrate. Films of micrometric thickness are often obtained in this form. (c) Sketch of an XAFS sample suitable for L3-edge measurements of liquid tin. Tin has been deposited directly upon the substrate, showing similar textures as shown in (b). (d) Sketch of another design for liquid XAFS samples. Micrometric grains (droplets on melting; red) are dispersed in a suitable inert and low X-ray-absorbing material. This material can be chosen according to the particular experiment under high-temperature or high-pressure conditions (BN, C, Al2O3, HfO2) and may contain suitable pressure markers. (e) Scanning electron micrograph of an emulsion of submicrometric liquid gallium droplets in epoxy resin, forming a sample suitable for XAFS experiments below 350 K (higher than the melting point of gallium). (f) Single-energy temperature scan near the Ga K edge for an emulsion of gallium droplets. The melting (m) and freezing (f) points of the gallium droplets, determined by the modulation of XAFS at selected energy (see inset and dashed line), confirm the exceptional undercooling properties of the metal droplets.

For many liquid substances, however, simpler strategies have been often adopted to obtain homogeneous specimens of micrometre thickness. In some cases, thin films of chemically pure substances can be deposited (for example by evaporation) onto suitable chemically inert substrates. These films may be covered with a protective layer to prevent leaks or modifications of the shape of the sample upon melting. Usually, this method is preferred for low-melting-point substances because the sample shape and purity are better preserved at lower temperatures. An example of such a sample configuration for measuring liquid tin (Sn L edges) is shown in Fig. 2[link](c), and further details can be found in Di Cicco (1996[link]). It is worth noting that the actual morphology of thin liquid samples is connected to fundamental properties such as surface tension and wettability; therefore, these specimens are often obtained as a distribution of submicrometric droplets, as shown, for example, in Fig. 2[link](b).

When melting is only obtained under high-temperature conditions, a successful strategy for producing XAFS samples is to disperse fine submicrometric grains of the solid specimen within another powdered hosting material, as in Fig. 2[link](d). Samples of this kind have also been used under high-pressure and high-temperature conditions in special furnaces, as reported, for example, by Filipponi & Di Cicco (1994[link]) and Farges et al. (1995[link]), and high-pressure/temperature devices, as reported, for instance, by Filipponi et al. (2000[link]) and references therein. The hosting material must have a low X-ray opacity and be chemically inert at high temperature, so that samples can be produced in the form of pellets that preserve their shape and purity at high temperature. Above the melting point, the droplets of the molten substance are embedded into the solid hosting material. The droplet distribution obtained in this way usually constitutes an excellent sample for XAFS experiments in transmission mode, showing very limited effects associated with the thickness distribution, as shown, for example, in Ottaviano et al. (1994[link]). An important feature related to the droplet distributions of pure substances is the possibility of performing accurate XAFS measurements of metastable states such as undercooled liquids and crystalline structures that cannot be found in ordinary bulk materials. For example, an emulsion of gallium droplets in epoxy resin, as shown in Fig. 2[link](e), is found to melt around 50 K below the melting point Tm of the stable crystalline phase, and can easily be kept in the undercooled liquid state down to 150 K (Di Cicco, 1998[link]). Undercooling rates exceeding (TmT)/T = 0.2 can be obtained in fine droplet distributions of pure substances. X-ray absorption can also be used to underpin melting (m) and solidification (f) of the droplets, as shown in Fig. 2[link](f). The hysteresis of gallium absorption at a selected energy (see inset) above the Ga K edge, associated with deep undercooling, is clearly shown in this figure. The single-energy temperature-scan technique (Filipponi et al., 1998[link]), which is often associated with X-ray diffraction, is of general interest for the study of phase transitions and has often been used to study the melting, crystal nucleation and appearance of metastable phases in metals.

4. XAFS experiments in liquids and gels

Due to limited space, only a few selected examples and references for XAFS experiments in liquids and gels are reported here. Gel substances have been studied by XAFS since the early days of this spectroscopy, and numerous experiments have been carried out on systems of biological (Alagna et al., 1986[link]; Bosco et al., 2002[link]) and geochemical (Combes et al., 1989[link]) interest, precursors of ceramics and other functional materials (Giorgetti et al., 1999[link]; Feth et al., 2002[link]) used as catalysts or electrodes and polymer electrolytes (Linford, 1995[link]). In particular, much effort has been devoted to studying sol–gel processes under variable ex situ and in situ temperature conditions, as reported, for example, in Ali et al. (1997[link]) and Feth et al. (2002[link]).

XAFS experiments on liquids and solutions have been reviewed in Filipponi (2001[link]) and Filipponi & D'Angelo (2016[link]), where the reader can find details of the numerous applications of the technique.

Most experiments on liquid solutions have been carried out under standard or moderate conditions of pressure and temperature. Recent experiments also include XAFS analysis of liquid solutions under high pressure (in the gigapascal range) using an adapted Paris–Edinburgh cell; see, for example, Filipponi et al. (2003[link]) and Migliorati et al. (2013[link]). XAFS experiments on elemental molecular liquids (such as, for example, Br2 and I2) have been performed using special cells in order to obtain samples of suitable thickness and to avoid chemical reactions (Filipponi, 2001[link]).

Many XAFS experiments on molten salts as well as liquid metals and alloys have been performed using high-temperature devices in the temperature range 300–3000 K, as described in the preceding section, although in some cases, such as liquid gallium (Di Cicco, 1998[link]), they were performed with a cryostat.

Generally speaking, low-noise XAFS measurements in an extended wavevector window (Δk ≃ 10–20 Å−1) were obtained in transmission mode on energy-scanning beamlines. XAFS data were found to provide reliable information on the short-range structure complementary to that obtained by diffraction techniques, as shown, for example, in Di Cicco et al. (1997[link], 2003[link]). Most available XAFS measurements of liquids have been obtained using energy-scanning beamlines, which presently allow controlled conditions over a wide range of temperatures and pressures using different devices. The Paris–Edinburgh press, used in a special anvil–gasket configuration, has become very popular for XAFS measurements in the 300–1500 K temperature and 0–10 GPa pressure ranges, as reported, for example, in Katayama et al. (1997[link]) and Filipponi (2001[link]). Installation and usage of the diamond anvil cell (DAC), pushing the pressure limits to the 100 GPa range, has now become feasible due to the improved focusing capabilities of the X-ray beam and the adoption of nanocrystalline diamonds, avoiding the presence of spurious (Bragg) peaks in XAFS data collection (Baldini et al., 2011[link]). Energy-dispersive beamlines provide XAFS data of narrower spectral extension (Δk ≃ 5–10 Å−1) and higher noise, but have often been used to identify phase transitions at high pressure and in XANES analysis, in view of the relatively simple setup to use the DAC. There are a few examples of applications to simple liquid systems at high pressures using the DAC; for example, krypton (Di Cicco et al., 1996[link]) and gallium (Comez et al., 2002[link]). Laser heating, coupled with DAC devices, and short acquisition times have been found to be extremely useful for monitoring phase transitions and the occurrence of chemical reactions under extreme conditions. An example is the melting curve of iron up to the megabar range (Aquilanti et al., 2015[link]).

It is also worth mentioning that the collection of XAFS-like data from light elements in liquids is feasible using the X-ray Raman (inelastic scattering) technique. An application to the most common liquid, water, is described in Bergmann et al. (2007[link]) and shows that reliable and unique information about the oxygen distribution can be obtained. Finally, measuring XAFS spectra of transient states in liquids (with subpico­second time resolution) is now becoming possible using free-electron laser facilities, as shown, for example, by Obara et al. (2014[link]).


First citationAlagna, L., Prosperi, T., Tomlinson, A. A. G. & Rizzo, R. (1986). J. Phys. Chem. 90, 6853–6857.Google Scholar
First citationAli, F., Chadwick, A. V. & Smith, M. E. (1997). J. Mater. Chem. 7, 285–291.Google Scholar
First citationAquilanti, G., Trapananti, A., Karandikar, A., Kantor, I., Marini, C., Mathon, O., Pascarelli, S. & Boehler, R. (2015). Proc. Natl Acad. Sci. USA, 112, 12042–12045.Google Scholar
First citationBaldini, M., Yang, W., Aquilanti, G., Zhang, L., Ding, Y., Pascarelli, S. & Mao, W. L. (2011). Phys. Rev. B, 84, 014111.Google Scholar
First citationBergmann, U., Di Cicco, A., Wernet, P., Principi, E., Glatzel, P. & Nilsson, A. (2007). J. Chem. Phys. 127, 174504.Google Scholar
First citationBosco, M., Miertus, S., Turchini, S., Prosperi, T., Ascone, I. & Rizzo, R. (2002). Carbohydr. Polym. 47, 15–26.Google Scholar
First citationCombes, J. M., Manceau, A., Calas, G. & Bottero, J. Y. (1989). Geochim. Cosmochim. Acta, 53, 583–594.Google Scholar
First citationComez, L., Di Cicco, A., Itié, J.-P. & Polian, A. (2002). Phys. Rev. B, 65, 014114–014123.Google Scholar
First citationCrozier, E. D., Rehr, J. J. & Ingalls, R. (1988). X-ray Absorption: Principles, Applications, Techniques of EXAFS, SEXAFS, and XANES, edited by D. C. Koningsberger & R. Prins, pp. 375–384. New York: Wiley.Google Scholar
First citationDi Cicco, A. (1996). Phys. Rev. B, 53, 6174–6185.Google Scholar
First citationDi Cicco, A. (1998). Phys. Rev. Lett. 81, 2942–2945.Google Scholar
First citationDi Cicco, A., Filipponi, A., Itié, J.-P. & Polian, A. (1996). Phys. Rev. B, 54, 9086–9098.Google Scholar
First citationDi Cicco, A., Minicucci, M. & Filipponi, A. (1997). Phys. Rev. Lett. 78, 460–463.Google Scholar
First citationDi Cicco, A., Trapananti, A., Faggioni, S. & Filipponi, A. (2003). Phys. Rev. Lett. 91, 135505.Google Scholar
First citationErtel, T. & Bertagnolli, H. (1993). Nucl. Instrum. Methods Phys. Res. B, 73, 199–202.Google Scholar
First citationFarges, F., Itié, J.-P., Fiquet, G. & Andrault, D. (1995). Nucl. Instrum. Methods Phys. Res. B, 101, 493–498.Google Scholar
First citationFeth, M. P., Weber, A., Merkle, R., Reinöhl, U. & Bertagnolli, H. (2002). J. Non-Cryst. Solids, 298, 43–52.Google Scholar
First citationFilipponi, A. (2001). J. Phys. Condens. Matter, 13, R23–R60.Google Scholar
First citationFilipponi, A., Borowski, M., Bowron, D. T., Ansell, S., Di Cicco, A., De Panfilis, S. & Itié, J.-P. (2000). Rev. Sci. Instrum. 71, 2422–2432.Google Scholar
First citationFilipponi, A., Borowski, M., Loeffen, P. W., De Panfilis, S., Di Cicco, A., Sperandini, F., Minicucci, M. & Giorgetti, M. (1998). J. Phys. Condens. Matter, 10, 235–253.Google Scholar
First citationFilipponi, A. & D'Angelo, P. (2016). X-ray Absorption and X-ray Emission Spectroscopy: Theory and Applications, edited by J. A. van Bokhoven & C. Lamberti, pp. 745–771. Chichester: John Wiley & Sons.Google Scholar
First citationFilipponi, A., De Panfilis, S., Oliva, C., Ricci, M. A., D'Angelo, P. & Bowron, D. T. (2003). Phys. Rev. Lett. 91, 165505.Google Scholar
First citationFilipponi, A. & Di Cicco, A. (1994). Nucl. Instrum. Methods Phys. Res. B, 93, 302–310.Google Scholar
First citationGiorgetti, M., Passerini, S., Smyrl, W. H., Mukerjee, S., Yang, X. Q. & McBreen, J. (1999). J. Electrochem. Soc. 146, 2387–2392.Google Scholar
First citationKatayama, Y., Mezouar, M., Itié, J.-P., Besson, J.-M., Syfosse, G., Le Fèvre, P. & Di Cicco, A. (1997). J. Phys. IV Fr. 7, C2-1011–C2-1012.Google Scholar
First citationLinford, R. G. (1995). Chem. Soc. Rev. 24, 267–277.Google Scholar
First citationMarcos, E. S., Gil, M., Martínez, J. M., Muñoz-Páez, A. & Sánchez Marcos, A. (1994). Rev. Sci. Instrum. 65, 2153–2154.Google Scholar
First citationMigliorati, V., Mancini, G., Tatoli, S., Zitolo, A., Filipponi, A., De Panfilis, S., Di Cicco, A. & D'Angelo, P. (2013). Inorg. Chem. 52, 1141–1150.Google Scholar
First citationObara, Y., Katayama, T., Ogi, Y., Suzuki, T., Kurahashi, N., Karashima, S., Chiba, Y., Isokawa, Y., Togashi, T., Inubushi, Y., Yabashi, M., Suzuki, T. & Misawa, K. (2014). Opt. Express, 22, 1105–1113.Google Scholar
First citationOttaviano, L., Filipponi, A. & Di Cicco, A. (1994). Phys. Rev. B, 49, 11749–11758.Google Scholar
First citationTamura, K. & Inui, M. (2001). J. Phys. Condens. Matter, 13, R337–R368.Google Scholar

to end of page
to top of page