Solubilization and coordination of the HgCl₂ molecule in water, methanol, acetone, and acetonitrile: an X-ray absorption investigation

Abstract

X-ray absorption spectroscopy (XAS) has been employed to carry out structural characterization of the local environment around mercury after the dissolution of the HgCl₂ molecule. A combined EXAFS (extended X-ray absorption fine structure) and XANES (X-ray absorption near edge structure) data analysis has been performed on the Hg L₃-edge absorption spectra recorded on 0.1 M HgCl₂ solutions in water, methanol (MeOH), acetone and acetonitrile. The Hg–Cl distance determined by EXAFS (2.29(2)–2.31(2) Å) is always comparable to that found in the HgCl₂ crystal (2.31(2) Å), demonstrating that the HgCl₂ molecule dissolves in these solvents without dissociating. A small sensitivity of EXAFS to the solvent molecules interacting with HgCl₂ has been detected and indicates a high degree of configurational disorder associated with this contribution. XANES data analysis, which is less affected by the disorder, was therefore carried out for the first time on these systems to shed light into the still elusive structural arrangement of the solvent molecules around HgCl₂. The obtained results show that, in aqueous and MeOH solutions, the XANES data are compatible with three solvent molecules arranged around the HgCl₂ unit to form a trigonal bipyramidal structure. The determination of the three-body Cl–Hg–Cl distribution shows a certain degree of uncertainty around the average 180° bond angle value, suggesting that the HgCl₂ molecule probably vibrates in the solution around a linear configuration.

---

1 Introduction

The interest in the chemistry of mercury originates from its toxicity: once released into the environment, due to its soft acid nature this element is able to interact with soft bases such as sulfur atoms of enzymes, proteins, and nucleic acids, eventually replacing biological Zn(II) ions and altering the normal activity of these species.^1–4 Furthermore, mercury can enter the food chain, especially through seafood,^5–7 being not expelled by organisms but rather bio-accumulated.⁸ For these reasons, the treatment of mercury and its compounds represents a dramatic environmental and health concern, and its removal from the environment is one of the current challenges of chemistry.^9–12

Despite being a naturally abundant metal, the availability of mercury can even rise because of human activities such as metallurgy and fossil fuel employment,^13,14 the latter one being derived either from coal combustion or natural gas emission. Natural gas can contain mercury in both its elemental form and as mercuric halides, among which HgCl₂ is one of the most common forms.^9–11 However, despite the dramatic impact of these compounds on stream, lake, and ocean pollution,¹⁵ the nature of the interaction of the HgCl₂ molecule with water and with very common molecular solvents seems to deserve further investigation. According to an X-ray absorption spectroscopy (XAS) study,¹⁶ in an aqueous solution, a Hg–Cl distance of 2.29(2) Å was determined, which is thus very similar to the 2.283(9) Å value found for the HgCl₂ crystal by means of X-ray diffraction (XRD) data.¹⁷ This result indirectly suggests that the HgCl₂ molecule dissolves in water without dissociating, at variance with what is observed for other salts like the Hg(ClO₄)₂ and Hg(TfO)₂ (TfO = trifluoromethanesulfonate) ones.^18–21 This evidence is commonly believed to rely on the strong covalent contribution to the Hg–Cl bond, but no explanation has been achieved so far about the reasons why this molecule dissolves in water without traces of precipitates, even if only at low concentrations owing to moderate water solubility.²² According to an XRD investigation,²³ a comparable picture also seems to be obtained in the methanol (MeOH) solution, because a Hg–Cl bond distance of 2.308(3) Å has been determined, even if measurements were carried out under highly concentrated conditions (1.5 M). In light of this, the determination of the dissolution process of the HgCl₂ molecule, i.e., if it is present as neutral species or as dissociated ions in solution, has important implications for biological processes addressing human health, because the conservation of neutrality may imply that it does not need a specific channel to pass through the cell membranes, but rather may undergo passive diffusion through them.²⁴ As concerns the arrangement of the solvent molecules around HgCl₂, according to ab initio calculations, the most stable configuration involves three water molecules on the equatorial plane of the linear HgCl₂ unit, forming an overall trigonal bipyramidal geometry.^25,26 In particular, both density functional theory (DFT) optimizations with an incremental number of water molecules and Monte Carlo simulations with DFT-derived interaction potentials resulted in a Hg–Cl distance in the 2.29–2.32 Å range and a Cl–Hg–Cl angle of 179°.²⁶ Other authors found that for DFT optimizations on increasing hydration clusters, a minimum number of 24 water molecules is needed to complete a first hydration shell and the trigonal bipyramidal HgCl₂(H₂O)₃ unit is produced as a result.²⁵ The step-by-step addition of water molecules around HgCl₂ provokes the lengthening of the Hg–Cl bond from 2.30 to 2.52 Å and the bending of the Cl–Hg–Cl linear configuration, although a bond angle of 176° is recovered once the trigonal bipyramidal structure is obtained. In this structure, the average Hg–O distance with the three coordinating water molecules is 2.48 Å. The formation of the trigonal bipyramidal HgCl₂(H₂O)₃ species is also observed by the same authors from ab initio molecular dynamics simulations, albeit a high number of exchange events occurs between the coordinating and the outer-sphere water molecules. In contrast, some experimental works indicate the association of two ligands to the HgCl₂ molecule, being either water molecules or other O-donor species.^27–31 In these structures, the Cl–Hg–Cl angle is bent, and a distorted tetrahedral geometry is formed with the additional two ligands. However, only a few data are presented in support of this putative structure in solution,²⁷ while this coordination seems to be more limited to the solid state.^28–31 Additional insight into the interaction of the HgCl₂ species with molecular solvents only relies on indirect information from Raman spectroscopy or calorimetric techniques,^32–34 thus leaving a vacant space for a clear experimental determination in support of these observations.

Herein, we present a study aimed at deepening the knowledge about the solvation structure of the HgCl₂ molecule in an aqueous solution and in common organic solvents, namely MeOH, acetone (Act) and acetonitrile (AN). The difficulty in obtaining an accurate structural description of metal solvation complexes in disordered liquid systems is a well-known task.^35–42 This is particularly true for what concerns the d¹⁰ Hg(II) center, which, like Zn(II) and Cd(II), presents no electronic spectroscopic handle, making the study of its solvation structure even more challenging.¹⁸ In this respect, XAS is an ideal method to provide accurate information on the local arrangement of the solvent molecules around a metal center, owing to its high sensitivity to the closest environment of the photoabsorber.^43,44 In particular, the high energy part of the absorption spectrum (EXAFS, extended X-ray absorption fine structure) is known to be particularly sensitive to the closest coordination shell and allows an accurate determination of bond distances, while the low energy part (XANES, X-ray absorption near-edge structure) can provide complementary information on the three-dimensional arrangement of the scattering atoms.^35,40,45 In principle, a complete recovery of the local structure around the photoabsorber can be achieved by the simultaneous employment of EXAFS and XANES analysis. This combined approach was therefore employed to obtain a unified picture about the coordination of mercury upon dissolution of the HgCl₂ molecule in the studied systems.

2 Materials and methods

2.1 X-ray absorption measurements

Hg L₃-edge XAS spectra were obtained at the 11.1 beamline⁴⁶ of Elettra-Sincrotrone Trieste in the transmission mode. Data were first collected on solid HgCl₂ (Sigma-Aldrich), which was diluted with boron nitride, carefully grained in an agate mortar and pressed into a pellet that was put in 1.5 mm aluminum frames with a Mylar tape covering the sample. The four 0.1 M solutions of HgCl₂ in deionized Milli-Q water, MeOH, Act and AN were prepared by adding stoichiometric amounts of the salt into the corresponding solvent. Cells with Kapton windows were filled with the liquid samples and kept under N₂ flux during data acquisition to avoid contact with the air. All measurements were performed using a Si(111) double crystal monochromator, while the storage ring was operating at 2 GeV with a beam current of 200 mA. At least three spectra were recorded and averaged for each sample. XAS data were also recorded on a 0.1 M solution of Hg(TfO)₂ in water as a comparative system.

2.2 EXAFS data analysis

The analysis of the EXAFS part of the absorption spectra was carried out using the GNXAS code.^47,48 Amplitudes and phase shifts have been calculated from clusters with fixed geometry within the muffin-tin (MT) approximation and advanced models for the exchange–correlation self-energy in the framework of the Hedin–Lundqvist (HL) scheme, which intrinsically allow to take into account the photoelectron inelastic losses in the final state.⁴⁹ Theoretical signals associated with n-body distribution functions have been calculated in accordance with the multiple-scattering (MS) theory and summed in order to reconstruct the total theoretical contribution. Each two-body distribution has been modelled as a Γ-like function depending on four structural parameters, namely the coordination number N, the average distance R, the Debye–Waller factor σ², and the asymmetry index β.

The EXAFS spectrum of solid HgCl₂ was analyzed starting from its crystallographic structure.¹⁷ For this purpose, two-body single scattering (SS) signals have been calculated to take into account the two first-shell chlorine atoms (Hg–Cl^1st), the six chlorine atoms set in the second shell (Hg–Cl^2nd), and the third shell mercury (Hg–Hg) and chlorine (Hg–Cl^3rd) atoms with coordination numbers of four and six, respectively. A three-body MS signal has been also calculated to take into account the Cl–Hg–Cl collinear distribution formed between mercury and the two closest chlorine atoms. Further insights about the treatment of the three body terms are provided in the ESI.† The MT radii for the Hg and Cl centers have been optimized by fitting the overlap between the potential spheres of adjacent atoms.

The EXAFS spectra of the 0.1 M solutions of HgCl₂ in water, MeOH, Act and AN were analyzed considering the Hg–Cl SS and the three-body Cl–Hg–Cl MS signals related to the two first-shell chlorine atoms. MT radii for Hg and Cl have been kept fixed to the values optimized in the fitting procedure of the HgCl₂ crystal. The inclusion of signals connected with the solvent molecules set in the second coordination shell was found not to provide a significant improvement to the quality of the EXAFS data fit (vide infra).

Least-squares minimizations have been carried out directly on the raw data, without preliminary background subtraction and Fourier filtering, by optimizing all the structural parameters. In addition, non-structural parameters have been optimized, namely the L₃-edge ionization energy E₀ and the energy position and amplitude of the double-electron excitation channels. The inclusion of the double-electron excitations allowed us to keep the S₀² amplitude reduction factor constrained between 0.90 and 1.00.

2.3 XANES data analysis

A fitting procedure of the XANES part of the spectra recorded on the HgCl₂ solutions was carried out using the MXAN code.^50,51 The potential has been calculated in the framework of the MT approximation using a complex optical potential, based on the local density approximation of the excited photoelectron self-energy. The real part of the self-energy was calculated using the HL scheme,⁴⁹ while the inelastic losses were accounted for by convolution of the theoretical spectrum with a Lorentzian function having an energy-dependent width of the form Γ_tot(E) = Γ_c + Γ_mfp(E). The constant part Γ_c includes the core–hole lifetime, while the energy-dependent term Γ_mfp(E) represents all the intrinsic and extrinsic inelastic processes. The function Γ_mfp(E) is zero below an onset energy E_s and begins to increase from a value A_s, following the universal functional form of the mean free path in solids. The numerical values of E_s and A_s are derived at each computational step using a Monte Carlo fit. A Gaussian convolution was used to account for the experimental resolution.

Least-squares fits have been carried out by optimizing the geometry of the starting models with the minimization of a residual function R_sq defined as:

where m is the number of data points, y^th_i and y^exp_i are the theoretical and experimental values of the absorption, respectively, ε_i is the individual error in the experimental data set, and w_i is a statistical weight. Five non-structural parameters were also optimized, namely the experimental resolution Γ_exp, the Fermi energy level E_F, the threshold energy E₀, and the energy and amplitude of the plasmon E_s and A_s.

3 Results and discussion

3.1 HgCl₂ crystal: EXAFS analysis

To determine the local structure around mercury in the HgCl₂ crystal, we carried out the analysis of the EXAFS part of the absorption spectrum recorded on solid HgCl₂ on the basis of its crystallographic structure, a portion of which is reported in Fig. 1.¹⁷ For this purpose, two-body signals have been calculated for the two first-shell chlorine atoms (Hg–Cl^1st), for the four second shell chlorine atoms (Hg–Cl^2nd), for the four mercury atoms set in the third coordination shell (Hg–Hg), and for the six farther chlorine atoms (Hg–Cl^3rd). The MS signal associated with the collinear Cl–Hg–Cl configuration involving the first-shell chlorine atoms has been also calculated assuming a bond angle of 180°. A least-squares fit was carried out in the 2.6–14.0 Å⁻¹k-range, computing the MT radii as proportional to the atomic number and optimizing the overlap between the potential spheres of adjacent atoms. The final optimized values were 1.6 Å and 0.9 Å for Hg and Cl, respectively. The best-fit results are shown in the upper panel of Fig. 2, where the theoretical two- and three-body signals are depicted together with the total theoretical contribution compared with the experimental data. As can be observed, a very good agreement between the theoretical and experimental data is obtained. The amplitude of the Cl–Hg–Cl MS term has been found to be low but still detectable. The structural parameters obtained for this three-body distribution are listed in Table S1 (ESI†), while the obtained E₀ and S₀² values are listed in Table S2 (ESI†) and the parameters for the double-electron excitations are listed in Table S3 (ESI†). The good agreement between the theoretical and the experimental data is also evident from the corresponding FT spectra shown in the lower panel of Fig. 2. The FT has been calculated in the 2.6–13.6 Å⁻¹ range. Here, it can be observed that the main FT peak is dominated by the Hg–Cl^1st contribution, while the features above ca. 2.5 Å can be associated with the second-shell chlorine atoms, with the Hg–Hg signal, and with the third-shell chlorine atoms, which are found at increasing distances (Table 1). In the region above ca. 3.5 Å, the theoretical signal is of a slightly lower intensity with respect to the experimental one, resulting in a not perfect match between the experimental and the fitting data. However, the reproduction of such features in the FT would require the employment of a larger model cluster to include the signals at longer distances, making the number of the fitted parameters prohibitive with respect to the number of experimental points. The complete list of the optimized structural parameters is reported in Table 1, while the E₀ value resulted to be 2.1 eV above the first inflection point of the experimental spectrum. Note that the Hg–Cl^1st distance obtained by the EXAFS analysis (2.31(2) Å, Table 1) is fully compatible with the value previously determined by means of XRD data (2.283(9) Å)¹⁷ also taking into account the so-called foreshortening effect in the latter technique,⁵² corroborating the goodness of the employed protocol.

Fig. 1 Portion of the HgCl₂ crystallographic structure taken from ref. 17. The Hg and Cl atoms are depicted in gray and green, respectively.

Fig. 2 Upper panel: Analysis of the Hg L₃-edge EXAFS spectrum of crystalline HgCl₂. From the top to the bottom: Hg–Cl^1st, Hg–Cl^2nd, Hg–Hg, and Hg–Cl^3rd SS theoretical signals, Cl–Hg–Cl MS three-body theoretical signal, total theoretical signal (blue line) compared with the experimental spectrum (red dots) and the residuals (blue dots). Lower panel: Non-phase shift corrected Fourier transforms of the best-fit EXAFS theoretical signal (blue line) of the experimental data (red dots) and of the residual curve (blue-dashed line).

Table 1 Structural parameters obtained from the EXAFS data analysis of the Hg L₃-edge adsorption spectrum of crystalline HgCl₂. N is the coordination number, R is the average distance, σ² is the Debye–Waller factor, and β is the asymmetry index

	N	R (Å)	σ ^2 (Å^−2)	β
Hg–Cl^1st	2.0	2.31(2)	0.002(2)	0.1(1)
Hg–Cl^2nd	6.0	3.49(3)	0.025(3)	0.0(2)
Hg–Hg	4.0	4.54(4)	0.011(4)	0.5(3)
Hg–Cl^3rd	6.0	4.77(4)	0.013(4)	0.2(3)

---

3.2 HgCl₂ in solution

One of the central aims of this work is to determine the influence of the solvent on the structural arrangement of the HgCl₂ molecule after its dissolution. The first qualitative insight can be gained by comparing the Hg L₃-edge XANES spectra of the HgCl₂ crystal and of HgCl₂ dissolved in water, MeOH, Act, and AN, together with the XANES spectrum recorded on a Hg(TfO)₂ aqueous solution as a comparative system (Fig. 3). As can be observed, the XANES spectra of the HgCl₂ crystal and of the HgCl₂ solutions present comparable features to one another, albeit small but non-negligible differences that can be better appreciated in the inset of Fig. 3. This is a first clue suggesting that the closest environment of the Hg center is very similar in these systems, regardless of the different nature of the solvent. On the other hand, the XANES spectrum of the Hg(TfO)₂ aqueous solution is very different from the previous ones, both in the amplitude and phase of the main oscillations (Fig. 3). The Hg(TfO)₂ salt is known to be dissociated in an aqueous solution, giving rise to a hydration complex where the Hg(II) ion is completely surrounded by water molecules.^12,18–20 As a consequence, the XANES comparison between these spectra strongly suggests that a different environment around the Hg center occurs after dissolution of the HgCl₂ species with respect to the Hg(TfO)₂ one and is the first qualitative indication that the HgCl₂ molecule is not dissociated in any of the studied solvents. This assumption would also explain the similar shapes of the spectra recorded on the HgCl₂ solutions (Fig. 3), being dominated by the Hg–Cl contribution. Further insights can be gained by the comparison among the EXAFS (Fig. 4(a)) and FT spectra (Fig. 4(b)). Here, small differences in the phase of the EXAFS oscillations appear for the HgCl₂ solutions (Fig. 4(a)), suggesting the presence of slight differences in distances of the first neighbors around the Hg center. The same circumstance can be observed from different positions of the main FT peak (Fig. 4(b)) and can be interpreted with a lengthening of the Hg–Cl distances following the AN < Act < MeOH < H₂O trend.

Fig. 3 Hg L₃-edge normalized XANES experimental spectra recorded on the HgCl₂ crystal and on the 0.1 M HgCl₂ solutions in water, MeOH, Act and AN, together with the XANES spectrum recorded on the 0.1 M Hg(TfO)₂ aqueous solution. The white-line area is magnified in the inset.

Fig. 4 Hg L₃-edge EXAFS (a) and FT (b) experimental spectra collected on the HgCl₂ crystal and on the 0.1 M HgCl₂ solutions in water, MeOH, Act and AN. The FT transforms have been calculated in the 2.6–13.6 Å⁻¹ range.

3.2.1 EXAFS data analysis. To have a definitive proof of these hypotheses and a more quantitative determination of the short-range structural arrangement around mercury in the studied solvents, the analysis of the EXAFS part of the spectra recorded on the HgCl₂ solutions has been carried out. For this purpose, two-body signals accounting for the two first-shell chlorine atoms have been calculated from the linear geometry of the HgCl₂ molecule, optimizing the structural parameters related to the Hg–Cl distribution starting from the values previously determined for the HgCl₂ crystal. In addition, the three-body MS signal associated with the collinear Cl–Hg–Cl configuration has been taken into account. Least-square fits have been performed in the 3.0–17.6 Å⁻¹k-range on the raw spectra, and the best-fit results are shown in the upper panels of Fig. 5. Herein, the theoretical two- and three-body signals are depicted together with the total theoretical contribution compared with the experimental data. As can be observed, the agreement between the theoretical and experimental spectra is very good for all the four solutions, and this is also evident from the corresponding FT spectra shown in the lower panels of Fig. 5. The FTs have been calculated in the 3.0–16.7 Å⁻¹k-range. The complete list of the optimized structural parameters for the Hg–Cl distributions is reported in Table 2, while the obtained E₀ and S₀² values are listed in Table S2 (ESI†), and the parameters for the double-electron excitations are listed in Table S3 (ESI†). From the structural results reported in Table 2, it is evident that the Hg–Cl distance in the solutions is almost identical to the Hg–Cl^1st one determined for the HgCl₂ crystal inside the experimental error range (Table 1). In addition, the optimized Debye–Waller factors σ² for the solutions (Table 2) are comparable with those optimized for the Hg–Cl^1st first-shell chlorine atoms in the HgCl₂ crystal (Table 1). The same is true for the optimized values of the asymmetry index β, which are close to zero for all the solutions, while more pronounced asymmetries are usually typical of disordered solvation complexes.^35,43 These results strongly support the evidence that none of these solvents is able to break the Hg–Cl bond and that HgCl₂ is present as a non-dissociated molecule in water, MeOH, Act, and AN solution. Also, the Hg–Cl distance obtained from the EXAFS data analysis of the aqueous solution (Table 2) is compatible with that found by means of ab initio simulations after the formation of the trigonal bipyramidal HgCl₂(H₂O)₃ species, and the same is true for the Cl–Hg–Cl bond angle.^25,26 Note that although inside the experimental error, the Hg–Cl distances for the solutions show small but detectable differences following the AN < Act < MeOH < H₂O trend. This determination is fully compatible with the shift in the phase of the EXAFS oscillations (Fig. 4(a)) and in the position of the FT main peak (Fig. 4(b)) and could indicate some weakening of the Hg–Cl covalent bond because of a different degree of interaction with the solvent molecules. The Hg–Cl distance has indeed been previously employed as a marker of the interaction of the Hg center with the solvent, i.e., the increase in the covalent bond length goes hand in hand with stronger Hg–solvent interactions.²³

Fig. 5 Upper panels: Analysis of the Hg L₃-edge EXAFS spectra recorded on 0.1 M HgCl₂ solutions in water, MeOH, Act and AN. From the top, the best-fit theoretical signals of Hg–Cl accounting for the two first-shell chlorine atoms and for the three-body Cl–Hg–Cl distributions are shown, together with the total theoretical contributions (blue line) with the experimental spectra (red dots) and the resulting residuals (blue dots). Lower panels: Non-phase shift-corrected Fourier transforms of the best-fit EXAFS theoretical signals (blue line) of the experimental data (red dots) and of the residual curves (blue-dashed line).

Table 2 Structural parameters for the Hg–Cl distribution obtained from the EXAFS analysis of the 0.1 M HgCl₂ solutions in water, MeOH, Act, and AN. N is the coordination number, R is the average distance, σ² is the Debye–Waller factor and β is the asymmetry index

		H2O	MeOH	Act	AN
Hg–Cl	N	2.0	2.0	2.0	2.0
	R (Å)	2.31(2)	2.30(2)	2.29(2)	2.29(2)
	σ ^2 (Å^−2)	0.003(2)	0.002(2)	0.002(2)	0.003(2)
	β	0.0(1)	0.0(1)	0.0(1)	0.0(1)

---

The structural parameters optimized for the Cl–Hg–Cl distribution, i.e., the bond angle values and standard deviations, are listed in Table S1 (ESI†). Although the EXAFS region possesses a lower sensitivity to the three-body configurations in comparison with the low-energy region of the absorption spectrum, the obtained values could indicate that the HgCl₂ molecule vibrates around a linear structure with a root mean square deviation of 4–5°. Note that some authors have previously inferred about a possible deviation from linearity of mercury(II) halide molecules HgX₂ dissolved in various solvents.^23,27,53,54 This originates from the observation that the ν₃(Hg–X) asymmetric stretching becomes weakly active in the Raman spectrum, suggesting a non-centrosymmetric structure of the HgX₂ unit, even if the authors pointed out that an unambiguous determination of this mode could not be carried out because of an interference pattern of the solvent.^53,54 These studies were later extended to the IR-active modes and compared with XRD data in support of the assertion that the X–Hg–X configuration assumes in solution an angle of 174°.²⁷ Herein, it has to be considered that in a real vibrating system, the probability of finding exactly a linear three-body distribution vanishes.^48,55 A more realistic picture should therefore imply the oscillation of the HgCl₂ molecule around a linear configuration, this being fully compatible with the appearance of the previously discussed spectral contributions^23,27,53,54 and with the mean deviation from linearity that we observed for the Cl–Hg–Cl MS path in the EXAFS data analysis (Table S1, ESI†).

In a further step, we carried out an additional analysis of the EXAFS data collected on the HgCl₂ aqueous solution to test the HgCl₂(H₂O)₃ model. For this purpose, an additional Hg–O SS signal associated with three water molecules has been included, and the results of this fitting procedure are shown in Fig. S1 (ESI†), while the optimized structural parameters are listed in Table S4 (ESI†). The agreement between the experimental spectrum and the theoretical signal including the water contribution (Fig. S1, ESI†) is comparable with that obtained for the HgCl₂ molecule only (Fig. 5). Considering that the fitting procedure has been carried out including three additional fitting parameters and that the frequency of the Hg–O signal is similar to that of the Hg–Cl one (Fig. S1, ESI†), a quantitative determination of the structural contribution associated with the water molecules is extremely difficult to obtain using the EXAFS technique. Moreover, the Debye–Waller factor of the Hg–O path is quite high (0.075 Å⁻² Table S4, ESI†) and this suggests the presence of high configurational disorder of the solvent molecules interacting with the HgCl₂ molecule. As a consequence, the amplitude of the EXAFS signal associated with the solvent molecules is quite low and it is mainly detected in the low-k region of the spectrum.^35,43,56 This circumstance was already evidenced in an XAS study on HgCl₂ in aqueous solution¹⁶ and is in line with the high rate of exchange events observed between the coordinating and the outer-sphere water molecules around the HgCl₂ molecule obtained by ab initio molecular dynamics simulations.²⁵

3.2.2 XANES data analysis. At variance with the EXAFS, the atomic thermal and structural disorder has a smaller impact on the low-energy region of the absorption spectrum, while MS paths provide a much larger contribution.⁵⁷ Therefore, a quantitative analysis of the XANES spectra was carried out to obtain insights into the three-dimensional arrangement of the scattering atoms around mercury. In particular, our aim is to assess the validity of the structures presented in the literature. As concerns the hydration complex, some experimental evidence points out the presence of two water molecules around the HgCl₂ unit,^27–31 while ab initio calculations suggest that three solvent molecules are arranged in the equatorial plane of the HgCl₂ molecule to form a trigonal bipyramidal geometry.^25,26 As a consequence, we carried out the analysis of the XANES spectrum recorded on the 0.1 M HgCl₂ aqueous solution starting from both structures. In addition, we also performed the analysis by considering only the isolated HgCl₂ molecule to check if the XANES technique is sensitive to the second-shell solvent molecules. The results of these fitting procedures are shown in Fig. 6. As can be observed, the starting model without the inclusion of the second-shell water molecules provides a poor fit, resulting in a clear mismatch in the phase and amplitude of the main oscillations between the theoretical and experimental spectra, and the same is true for the structure with two water molecules. Note that in the latter case, the bond angles have been let vary to assess the possible formation of distorted geometries as suggested in the literature,^27–31 but even with this additional degree of freedom, the fit gives back a poor agreement with the experimental data. Conversely, a much better result is obtained by the model with three water molecules, as can also be observed from the much lower value of the residual function R_sq. The best-fit structural and non-structural parameters for this geometry are listed in Table 3. The optimized Hg–Cl bond distance is shorter than that determined from the EXAFS analysis of the same sample (Table 2), but the two values are still in good agreement if the systematic shortening of the coordination distance due to the low-energy behavior of the real part of the HL potential is taken into account.⁴⁵ Finally, the water molecules are found at a Hg–O distance of 2.75(8) Å and are clearly set in the second coordination sphere of the Hg center.

Fig. 6 Comparison between the Hg L₃-edge XANES experimental spectrum recorded on the 0.1 M HgCl₂ aqueous solution (red dots) compared with the theoretical signals (blue line) calculated with three different models: isolated HgCl₂ molecule (left panel) and HgCl₂ plus two (central panel) and three (right panel) water molecules. For each fitting procedure, the obtained residual function R_sq and the optimized clusters are shown as insets.

Table 3 Structural and non-structural parameters obtained from the XANES analysis of the Hg L₃-edge experimental spectra recorded on the 0.1 M HgCl₂ aqueous and MeOH solutions for the structures with three solvent molecules. R_Hg–Cl and R_Hg–O are, respectively, the Hg–Cl and Hg–O distances, E₀ is the threshold energy, E_F is the Fermi energy, E_S and A_S are the plasmon energy onset and amplitude, and Γ_exp is the experimental resolution

Solvent	R Hg–Cl (Å)	R Hg–O (Å)	E 0 (eV)	E F (eV)	E S (eV)	A S	Γ exp (eV)
H2O	2.21(8)	2.75(8)	0.5	−3.7	11.0	11.7	3.7
MeOH	2.20(8)	2.74(8)	0.1	−4.0	4.6	12.8	3.3

---

The XANES data analysis for the water case has also been carried out by taking into account a bent HgCl₂ geometry to verify the sensitivity of the XANES data to the Cl–Hg–Cl configuration. For this purpose, fitting procedures have been performed starting from models with 175° and 170° Cl–Hg–Cl angles, plus three water molecules on the equatorial plane. These values have been chosen as some authors previously supported the observation that the HgCl₂ molecule assumes a bent configuration in solution with a bond angle of ∼174°.^23,27,53,54 The results of these fits are shown in Fig. S2 (ESI†), while the optimized parameters are listed in Table S5 (ESI†). The agreement between the theoretical and the experimental oscillations is similar between these two cases and in comparison to the fit carried out assuming a linear configuration (right panel of Fig. 6), as it is also evident from the similar R_sq values. This result may highlight poor sensitivity of the XANES data to such a small deviation from a linear distribution or, as previously discussed for the EXAFS data analysis, this could indicate that the HgCl₂ molecule vibrates in solution around a linear configuration with a root mean square deviation of 4–5°, resulting in an equivalent distribution of all the angle values between ∼180° and 175° in probabilistic terms.

As concerns the HgCl₂ molecule in the MeOH solution, the XANES data fit was carried out using a starting model where three MeOH molecules are set in the equatorial plane of the HgCl₂ unit, because this was the best-fit model for the aqueous system. In fact, both water and MeOH are O-donor species and it is known that MeOH is the organic solvent that most shares its structural properties with water in terms of coordination geometries and formation of hydrogen-bond networks.^58–60 The fit result is shown in Fig. 7, while the optimized parameters are listed in Table 3. As can be observed, this model also provided a theoretical XANES signal in very good agreement with the experimental one for the MeOH solution, as is also testified by the low R_sq value, corroborating the supposition that the solvent molecules around the HgCl₂ unit are arranged with the same geometry both in aqueous and MeOH solutions. The best-fit structural parameters show Hg–Cl and Hg–O distances comparable to those found in the XANES analysis of the aqueous solution (Table 3).

Fig. 7 Comparison between the Hg L₃-edge XANES experimental spectrum recorded on the 0.1 M HgCl₂ MeOH solution (red dots) compared with the theoretical signal (blue line) calculated with a starting structure where three MeOH molecules coordinate the HgCl₂ unit in a trigonal bipyramidal structure. The obtained residual function R_sq and the optimized cluster are shown as insets.

An attempt to fit the XANES spectrum recorded on the 0.1 M HgCl₂ solution in Act has also been carried out using the model involving three solvent molecules normal to the HgCl₂ molecule. The result of this fit is shown in Fig. S3 (ESI†), while the optimized parameters are listed in Table S6 (ESI†). As can be observed, the quality of the fit is acceptable, although it is possible to further improve the agreement between the experimental and the theoretical oscillations. This is because, despite being an O-donor species like water and MeOH, the Act molecule possesses a more complicated structure, where the Hg–O–C MS paths are expected to give a non-negligible contribution due to the so-called “focusing effect”. A systematic study of the HgCl₂ coordination in this solvent would, therefore, require the evaluation of different configurations and the absence of putative starting structures makes it more challenging and beyond the scope of the present work. For the same reasons, the fit of the XANES spectrum recorded on the AN solution was not attempted here, as this solvent has an even more different chemical nature, where more marked MS effects are expected due to the linearity of the molecule.

4 Conclusions

Structural characterization of the local environment around mercury after dissolution of the HgCl₂ molecule has been carried out by means of a combined approach between EXAFS and XANES data analysis. The analysis of the EXAFS part of the Hg L₃-edge absorption spectra recorded on HgCl₂ solutions in water, MeOH, Act and AN showed that the Hg–Cl distance (2.29(2)–2.31(2) Å) is always comparable to that determined from the EXAFS analysis of crystalline HgCl₂ (2.31(2) Å). The HgCl₂ molecule is, therefore, found not to dissociate after dissolution in the studied solvents. Small but detectable differences in the Hg–Cl distance obtained for the solutions follow the AN < Act < MeOH < H₂O trend and suggest a different degree of interaction between the solvent molecules and the HgCl₂ unit. A small sensitivity of the EXAFS analysis to the solvent molecules interacting with HgCl₂ was detected and it is probably derived from the high degree of configurational disorder associated with this contribution. The XANES analysis, which is less affected by thermal and structural broadening, was, therefore, employed for the first time on these systems to shed light on the still elusive arrangement of the solvent around the HgCl₂ molecule. Our analysis showed that among the various coordinations reported in the literature, the model where three water molecules are set on the equatorial plane of the HgCl₂ unit to form a trigonal bipyramidal geometry is the most compatible with the XANES data. This model was also extended to the MeOH solution and the XANES results support the hypothesis that the MeOH molecules are arranged around HgCl₂ in a similar manner compared to what was found for the water case. As concerns the three-body distribution in the HgCl₂ molecule, the EXAFS analysis determined a Cl–Hg–Cl bond angle value of 180° with a 4–5° standard deviation. In addition, XANES data analysis for the water case carried out with Cl–Hg–Cl bond angles of 175° and 170° showed similar results with respect to the linear case. Altogether these findings suggest that the HgCl₂ molecule vibrates in the solution around a linear configuration and explain the data previously presented in the literature in support of a possible deviation of HgCl₂ from linearity. The considerations reported here therefore represent a step forward in the understanding of the dissolution mechanism of the HgCl₂ compound in the inspected media and have the potentiality of better clarifying the structural arrangement of the solvent molecules around HgCl₂, an issue that sees a proliferation of putative structures previously presented in the scientific literature.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We acknowledge financial support from the Italian Ministry of University and Research (MIUR) through Grant “PRIN 2017, 2017KKP5ZR, MOSCATo” and from the University of Rome “La Sapienza” Grant RG11916B702B43B9. Elettra Sincrotrone Trieste S. C. p. A. and its staff are acknowledged for synchrotron radiation beam time and laboratory facilities (proposal number 20180157).

Notes and references

1. K. W. Jennette, Environ. Health Perspect., 1981, 40, 233–252.
2. K. Gwozdzinski, Arch. Environ. Contam. Toxicol., 1992, 23, 426–430.
3. L. M. Barone, C. Shih and B. P. Wasserman, J. Biol. Chem., 1997, 272, 30672–30677.
4. R. Ynalvez, J. Gutierrez and H. Gonzales-Cantu, Biometals, 2016, 29, 781–788.
5. P. Grandjean, P. Weihe, R. F. White, F. Debes, S. Araki, K. Yokoyama, K. Murata, N. Sorensen, R. Dahl and P. J. Jorgensen, Neurotoxicol. Teratol., 1997, 19, 417–428.
6. G. J. Myers, P. W. Davidson, C. Cox, C. F. Shamlaye, E. Cernichiari and T. W. Clarkson, Environ. Res., 2000, 83, 275–285.
7. G. J. Myers, P. W. Davidson, C. Cox, C. F. Shamlaye, D. Palumbo, E. Cernichiari, J. Sloane-Reeves, G. E. Wilding, J. Kost, L. S. Huang and T. W. Clarkson, Lancet, 2003, 361, 1686–1692.
8. F. M. M. Morel, A. M. L. Kraepiel and M. Amyot, Annu. Rev. Ecol. Syst., 1998, 29, 543–566.
9. L. Ji, S. W. Thiel and N. G. Pinto, Ind. Eng. Chem. Res., 2008, 47, 8396–8400.
10. M. Abai, M. P. Atkins, A. Hassan, J. D. Holbrey, Y. Kuah, P. Nockemann, A. A. Oliferenko, N. V. Plechkova, S. Rafeen, A. A. Rahman, R. Ramli, S. M. Shariff, K. R. Seddon, G. Srinivasan and Y. Zou, Dalton Trans., 2015, 44, 8617–8624.
11. S. E. E. Warrag, E. O. Fetisov, D. J. G. P. van Osch, D. B. Harwood, M. C. Kroon, J. I. Siepmann and C. J. Peters, Ind. Eng. Chem. Res., 2018, 57, 9222–9230.
12. V. Migliorati, G. Fazio, S. Pollastri, A. Gentili, P. Tomai, F. Tavani and P. D'Angelo, J. Mol. Liq., 2021, 329, 115505.
13. J. E. Gray and M. E. Hines, Appl. Geochem., 2006, 21, 1819–1820.
14. C. T. Driscoll, R. P. Mason, H. M. Chan, D. J. Jacob and N. Pirrone, Environ. Sci. Technol., 2013, 47, 4967–4983.
15. P. Seller, C. A. Kelly, J. W. M. Rudd and A. R. MacHutchon, Nature, 1996, 380, 694–697.
16. R. Åkesson, I. Persson, M. Sandström and U. Wahlgren, Inorg. Chem., 1994, 33, 3715–3723.
17. V. Subramanian and K. Seff, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1980, 36, 2132–2135.
18. G. Chillemi, G. Mancini, N. Sanna, V. Barone, S. Della Longa, M. Benfatto, N. V. Pavel and P. D’Angelo, J. Am. Chem. Soc., 2007, 129, 5430–5436.
19. G. Mancini, N. Sanna, V. Barone, V. Migliorati, P. D'Angelo and G. Chillemi, J. Phys. Chem. B, 2008, 112, 4694–4702.
20. I. Persson, L. Eriksson, P. Lindqvist-Reis, P. Persson and M. Sandström, Chem. – Eur. J., 2008, 14, 6687–6696.
21. W. Lu, P. S. Barber, S. P. Kelley and R. D. Rogers, Dalton Trans., 2013, 42, 12908–12916.
22. W. Y. Shiu, K. C. Ma, D. Mackay, J. N. Seiber and R. D. Wauchope, in Solubilities of Pesticide Chemicals in Water Part II: Data Compilation, ed. G. W. Ware, H. N. Niggs and A. Bevenue, Springer, New York, New York, NY, 1990, pp. 15–187.
23. M. Sandström, Acta Chem. Scand., Ser. A, 1978, 32, 627–641.
24. J. Gutknecht, J. Membr. Biol., 1981, 61, 61–66.
25. L. Castro, A. Dommergue, A. Renard, C. Ferrari, A. Ramírez-Solís and L. Maron, Phys. Chem. Chem. Phys., 2011, 13, 16772–16779.
26. J. Hernández-Cobos, A. Ramírez-Solís, L. Maron and I. Ortega-Blake, J. Chem. Phys., 2012, 136, 014502.
27. I. Persson, M. Sandström and P. L. Goggin, Inorg. Chim. Acta, 1987, 129, 183–197.
28. C.-I. Brändén, Acta Chem. Scand., 1963, 17, 1363–1374.
29. K. Balasubramani, P. Thomas, G. Bocelli and A. Cantoni, J. Coord. Chem., 2005, 58, 1689–1694.
30. A. Crochet and K. Fromm, Z. Anorg. Allg. Chem., 2010, 636, 1484–1496.
31. I. Galkina, A. Tufatullin, D. Krivolapov, Y. Bakhtiyarova, D. Chubukaeva, V. Stakheev, V. Galkin, R. Cherkasov, B. Büchner and O. Kataeva, CrystEngComm, 2014, 16, 9010–9024.
32. I. Persson, C. D. Kailash and K. Yoshiaki, Acta Chem. Scand., 1990, 44, 433–442.
33. S. Ahrland, I. Persson and R. Portanova, Acta Chem. Scand., Ser. A, 1981, 35, 49–60.
34. S. Ahrland, E. Hansson, A. Iverfeldt and I. Persson, Acta Chem. Scand., Ser. A, 1981, 35, 275–285.
35. M. Busato, A. Melchior, V. Migliorati, A. Colella, I. Persson, G. Mancini, D. Veclani and P. D’Angelo, Inorg. Chem., 2020, 59, 17291–17302.
36. V. Migliorati, A. Gibiino, A. Lapi, M. Busato and P. D’Angelo, Inorg. Chem., 2021, 60, 10674–10685.
37. M. Busato, P. D'Angelo, A. Lapi, M. Tolazzi and A. Melchior, J. Mol. Liq., 2020, 299, 112120.
38. M. Busato, A. Lapi, P. D’Angelo and A. Melchior, J. Phys. Chem. B, 2021, 125, 6639–6648.
39. M. Busato, P. D’Angelo and A. Melchior, Phys. Chem. Chem. Phys., 2019, 21, 6958–6969.
40. P. D'Angelo, V. Migliorati, F. Sessa, G. Mancini and I. Persson, J. Phys. Chem. B, 2016, 120, 4114–4124.
41. V. Migliorati, A. Caruso and P. D’Angelo, Inorg. Chem., 2019, 58, 14551–14559.
42. V. Migliorati, F. Sessa, G. Aquilanti and P. D’Angelo, J. Chem. Phys., 2014, 141, 044509.
43. A. Filipponi and P. D’Angelo, XAS in Liquid Systems, John Wiley & Sons, Ltd, 2016, ch. 25, pp. 745–771.
44. A. Di Cicco, XAS Res. Rev., 2016, 15, 1–7.
45. P. D’Angelo, M. Benfatto, S. Della Longa and N. V. Pavel, Phys. Rev. B: Condens. Matter Mater. Phys., 2002, 66, 064209.
46. A. D. Cicco, G. Aquilanti, M. Minicucci, E. Principi, N. Novello, A. Cognigni and L. Olivi, J. Phys.: Conf. Ser., 2009, 190, 012043.
47. A. Filipponi, A. Di Cicco and C. R. Natoli, Phys. Rev. B: Condens. Matter Mater. Phys., 1995, 52, 15122–15134.
48. A. Filipponi and A. Di Cicco, Phys. Rev. B: Condens. Matter Mater. Phys., 1995, 52, 15135–15149.
49. L. Hedin and S. Lundqvist, Solid State Phys., 1969, 23, 1–181.
50. M. Benfatto and S. Della Longa, J. Synchrotron Radiat., 2001, 8, 1087–1094.
51. M. Benfatto, S. Della Longa and C. R. Natoli, J. Synchrotron Radiat., 2003, 10, 51–57.
52. P. Moeck, O. Čertik, B. Seipel, R. Groebner, L. Noice, G. Upreti, P. Fraundorf, R. Erni, N. D. Browning, A. Kiesow and J.-P. Jolivet, Two- and Three-Dimensional Methods for Inspection and Metrology III, 2005, pp. 206–217.
53. J. Smith and T. Brill, Inorg. Chim. Acta, 1976, 18, 225–229.
54. D. N. Waters and Z. Kantarci, J. Raman Spectrosc., 1977, 6, 251–256.
55. A. Filipponi, A. Di Cicco, R. Zanoni, M. Bellatreccia, V. Sessa, C. Dossi and R. Psaro, Chem. Phys. Lett., 1991, 184, 485–490.
56. M. Busato, A. Tofoni, G. Mannucci, F. Tavani, A. Del Giudice, A. Colella, M. Giustini and P. D’Angelo, Inorg. Chem., 2022, 61, 8843–8853.
57. M. Benfatto, C. R. Natoli, A. Bianconi, J. Garcia, A. Marcelli, M. Fanfoni and I. Davoli, Phys. Rev. B: Condens. Matter Mater. Phys., 1986, 34, 5774–5781.
58. V. Migliorati and P. D'Angelo, RSC Adv., 2013, 3, 21118–21126.
59. V. Migliorati and P. D’Angelo, Chem. Phys. Lett., 2015, 633, 70–75.
60. M. Busato, A. Del Giudice, V. Di Lisio, P. Tomai, V. Migliorati, A. Gentili, A. Martinelli and P. D’Angelo, ACS Sustainable Chem. Eng., 2021, 9, 12252–12261.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2cp02106d

---