Solvation of CaCl2 at the Water–Vapor Interface Revealed by Ambient Pressure X‑ray Photoelectron Spectroscopy Authors: Andreas Siebert, Kritanjan Polley, Shane W. Devlin, Mauricio Lopez Luna, N. Isaac Zakaria, Nykita Rustad, David T. Limmer, Monika Blum Published: 2025 ABSTRACT The determination of ion concentrations at aqueous interfaces is of paramount importance for the precise modeling of atmospheric reactions. Despite their significant concentrations observed in bulk sea spray aerosol studies, the concentration of Ca2+ ions at the liquid–vapor interface remains unclear. Ambient pressure X-ray photoelectron spectroscopy measurements performed on a liquid flat sheet combined with molecular dynamics simulations reveal the solvation structure and density profiles of Ca2+ and Cl– in the first nanometers of the liquid at the liquid–vapor interface. We report first photoelectron angular distribution measurements on a liquid flat sheet and use them to calculate the photoionization anisotropy parameter of the O 1s, Cl 2p, and Ca 2p core levels. The measurements show the influence of the elastic electron scattering on the anisotropy parameter at low electron kinetic energies, and we include this effect in our Ca2+ and Cl– photoelectron depth profiles. Our measurements and molecular simulations reveal a depletion of the Ca2+ ion concentration and an increased Cl– over Ca2+ ion ratio at the liquid–vapor interface compared to the bulk of the solution. Introduction Inorganic salts are fundamental components of atmospheric chemistry and biological systems, profoundly influencing terrestrial life. Calcium chloride, dissolved in atmospheric aqueous aerosols, exemplifies this significance through its role in cloud condensation nuclei reactions, directly modulating precipitation and Earth’s radiation absorbance. A comprehensive understanding of the ion solvation behavior and interfacial concentration gradients at the liquid–vapor boundaries is necessary for mechanistic insights into these processes. Such mechanistic insights are an essential prerequisite for accurate modeling of atmospheric reactions like ozonolysis and cloud nucleation. − However, investigating these interfacial phenomena presents significant theoretical and experimental challenges due to their fluctuating nature. Although calcium accounts for less than 1% of inorganic salts in seawater, its relative concentration increases substantially with decreasing sea spray aerosol droplet size and enhanced surface-to-volume ratios. − Beyond marine sources, terrestrial mineral dust represents another significant calcium reservoir primarily through CaCO3-bearing particles. Reactions with acidic trace gases like HCl can lead to the formation of CaCl2, an aging product of CaCO3 with a higher hygroscopicity, comparable to (NH4)2SO4. The higher hygroscopicity leads to the ability to form larger aerosol droplets and hence increases the cloud condensation nuclei activity of CaCl2. The surface sensitivity of ambient pressure X-ray photoelectron spectroscopy (APXPS) makes it a perfect technique to deliver insights into the chemical composition and distribution of ions and molecules at interfaces, such as in atmospheric aerosol systems. − The near ambient pressure conditions in the experimental chamber of up to a few millibars are in good agreement with the pressure range observed in the stratosphere, and the tunability of the incoming photon energy allows for depth profiling of the solution. In recent years, developments on liquid jet systems have made the non-destructive investigation of liquid–vapor interfaces more feasible. ,− Beyond the widely utilized circular microjet systems, colliding liquid flat sheet systems are increasingly employed to study phenomena such as solution mixing. The nearly planar liquid surface provides additional advantages, such as a higher achievable electron intensity compared to circular jets (increased by a factor of π/2 for a 90° incident beam compared to the detector axis). The electron intensity might be further increased due to the large target surface of flat sheets, easily exceeding the beam spot size, which is not necessarily the case for circular jets. Several APXPS studies, with varying sample delivery systems, investigated the ion concentration of different salt solutions at the liquid–vapor interface, ,,,− mainly focusing on halide salt ion concentrations. The Cl– ion concentration, for example, was found to be constant or increase at the interface of NaCl solutions in both theoretical and experimental studies. , The Ca2+ concentration at the liquid–vapor interface was, so far, investigated only in mixtures of CaCl2 with NaCl or sea salt solution, revealing an increased amount of Ca2+ over Na+ at the interface relative to the expected ratio for the mixed solutions. In other studies, doubly charged ions such as CO3 2– have shown surface activity through cooperative ion pairing effects, as has been measured by APXPS liquid jet and second harmonic generation experiments. , These studies appear to contradict the general belief that multivalent ions tend to prefer the bulk region while large, polarizable, and weakly hydrated ions tend to be surface active. , On the other hand, no surface enrichment of a different double positively charged ion (Fe2+) was observed in FeCl2 solutions studied with APXPS and a circular jet system, in agreement with the general belief. In the case of Ca2+, so far, only the solvated ion ratios have been determined with APXPS; an investigation of the absolute ion concentration relative to water at the interface has not been undertaken. The interpretation of interfacial ion distributions requires careful consideration of the water interface stratification. When APXPS results suggest an increased ion concentration at the interface, within the first nanometers of the solution, it is crucial to understand that the solvated ions are likely not sitting on top of the water surface but that there is a thin water layer with a depleted ion concentration at the surface, followed by a water layer with an increased ion concentration. The interpretation of APXPS measurements conducted on liquids provides a few additional challenges. For instance, the correct photoionization cross section of solutes is still disputed. , Furthermore, the photoelectron angular distribution (PAD) is influenced by elastic electron scattering occurring in the solution. This needs to be accounted for if the measurement geometry is not in the magic angle configuration. Therefore, the anisotropy parameter β can be obtained by measuring and fitting PADs. ,,− By combining this analysis with molecular dynamics simulations using classical polarizable force fields, we are able to determine the interfacial concentrations of Ca2+ and Cl– ions at the air–water interface. Results and Discussion APXPS of Aqueous 0.25 and 0.5 M CaCl2 Solutions To investigate the ion concentration at the liquid–vapor interface, we have measured the survey spectra, valence band, and Ca 2p, Cl 2p, and O 1s core levels of aqueous 0.25 and 0.5 M CaCl2 solutions at different kinetic energies of the photoelectrons (eKEs) between 100 and 800 eV. The eKE correlates with the information depth, i.e., electron effective attenuation length (EAL), and kinetic energies of 100 and 800 eV correspond to EALs of roughly 1–2 and 5–8 nm, respectively (estimated from a water solution). , The survey spectra of the studied solutions, shown in Figure A, measured with an excitation energy (hν) of 900 eV, present all of the expected core levels related to water and the solvated Ca2+ and Cl– ions in the bulk as well as the water O KLL Auger. The peak intensity of the Cl and Ca core levels increases by a factor of two between the 0.25 and 0.5 M CaCl2 solutions, highlighting the sensitivity of APXPS to the number density of the dissolved solutes. Figure B shows examples of the O 1s core level of both solutions with an eKE of 250 eV after linear background subtraction, fitted with two Voigt functions to account for the liquid (O 1s­(l)) and gas (O 1s­(g)) phases of water. In Figure C,D, similar examples of Ca 2p and Cl 2p core levels are presented, fitted with two Voigt functions to account for the p core level doublet splitting. The binding energies have been adjusted relative to the 1b1(l) binding energy (see the Methods section). (A) Survey spectra of aqueous 0.25 and 0.5 M CaCl2 solutions, measured with an incident photon energy of 900 eV. The (B) O 1s, (C) Ca 2p, and (D) Cl 2p core level fits of aqueous 0.25 and 0.5 M CaCl2 solutions measured with an eKE of 250 eV. At an eKE of 500 eV, all of the core levels overlap with the O KLL Auger peak. To account for the influence of the Auger peak on the Cl 2p and Ca 2p spectra, water spectra have been measured under the same experimental conditions and subtracted from the spectra of the salt solution (see section S1 and Figures S1 and S2 in the Supporting Information). An overview of all core level fits can be found in section S2 and Figures S3–S5, and additional details regarding the fitting procedure can be found in the Methods section. Determination of the Anisotropy Parameter with PAD Measurements The measurements in the previous section were performed using horizontally polarized photons. This configuration delivers the highest photon flux, which maximizes the signal-to-noise ratio. On the other hand, the resulting angle (θ) between the detector axis and photon polarization direction (ε) is around 35.3°, which requires consideration of the impact of β on the differential photoionization cross section ( dσidΩ ), where σ i refers to the photoionization cross section of the ith core level. The differential photoionization cross section for linearly polarized light in the dipole approximation can be written as 1 dσidΩ=σi4π(1+β2(3cos2(θ)−1)) By changing the photon polarization direction (ε), the angle θ can be changed (see the inset of Figure D as an example of nearly vertically polarized light). Measuring the same core level under the same experimental conditions but with different photon polarization directions allows for the determination of β using eq . 2 f(θ)=C(1+β2(3cos2(θ)−1)) Fit of the (A) Ca 2p, (B) Cl 2p, (C) O 1s­(l), and (D) O 1s­(g) core level PADs. Shown are normalized peak areas derived from core level fits of spectra measured with different angles (θ) and different eKEs. θ is the angle between the photon polarization direction and the detector axis. The fit delivers β of the solvated Ca2+ and Cl– ions and the liquid (l) and gas (g) phases of the O 1s core level of water. In the insets in (A–C), examples for the recorded core level spectra at an eKE of 400 eV at different angles θ are shown. The data for O 1s­(l) and O 1s­(g) were derived from the same spectra. The inset in (D) shows the liquid flat sheet in front of the analyzer nozzle. Indicated are the incoming light with a nearly vertical polarization direction (ε) and the angle θ between the detector axis and ε. The measured β value of liquids will be smaller compared to the β value obtained for gas phases since the distinct photoelectron distribution is blurred by elastic electron scattering events, changing the electron trajectory. In fact, the ratio of liquid and gas phase β is directly proportional to the elastic scattering mean free path length. With enough elastic scattering events, the photoelectron distribution becomes isotropic and β = 0. In Figure , the measured PADs of Ca 2p3/2, Cl 2p3/2, O 1s­(l), and O 1s­(g) of an aqueous 0.5 M CaCl2 solution are shown for eKEs between 600 and 30 eV. Examples of the core levels measured with an eKE of 400 eV with different polarization directions corresponding to angles θ of 35.3°, 56.2°, and 91.7° are shown in the insets of Figure A–C, while the inset in Figure D shows the geometry of the liquid flat sheet in front of the electron analyzer nozzle. The detailed fits of all core levels at all eKEs and polarization directions can be found in Supporting Information section S3 in Figures S7–S9. All PADs in Figure are reasonably well fitted by eq , underlining that the angular photoelectron distribution is indeed well described by this formula, which is not trivial since the formula is derived by the assumption that the excited atoms/molecules are oriented randomly. At the liquid–vapor interface, this random orientation of water molecules is questionable, ,, but similar to prior measurements conducted on circular jets, a disturbance of the photoelectron angular distribution described in eq by oriented water molecules on the surface could not be observed. The detailed fitting parameters and measures for the quality of the fit are presented in Figure and Tables S1 and S2 in the Supporting Information. Values of β of Ca 2p3/2 and Cl 2p3/2 at 50 and 30 eV could not be determined due to the lower limit of the available beamline photon energy range or the low signal-to-noise ratio of the data. The low concentration of Ca2+ and Cl– is also responsible for the relatively large uncertainty of β compared to the β determined for the liquid and gas phase O 1s PADs. Anisotropy parameter (β) of (A) the liquid (l) and gas (g) phases of the O 1s core level and (B) the Cl 2p3/2 and Ca 2p3/2 core levels obtained from fits in Figure shown over the eKE. For comparison, β of O 1s­(l) and O 1s­(g) of pure water measured by Thürmer et al. on a circular jet (15 μm diameter) and calculated values (without elastic scattering) from Yeh for the Ca 2p3/2 and Cl 2p3/2 core levels are shown. Data reproduced from refs and . As a reference, values of β obtained for the liquid and gas phases of O 1s on pure water with a circular jet by Thürmer et al. are shown in Figure A. The values for the gas phase are generally in very good agreement. At higher eKEs, β of O 1s­(g) is nearly constant with uncertainties overlapping with the values of Thürmer et al. and within error of the expected value of 2. Toward lower eKEs, the elastic scattering effects increase, and β is reduced, following the decline observed by Thürmer et al. This underlines the quality of our measurement results and confirms that potential depolarization effects due to the angular acceptance of the detector and/or depolarization of the beam are small. At higher eKEs, similar to the β of O 1s­(g), the values of β of O 1s­(l) are in good agreement with the values determined by Thürmer et al., which hints at a negligible impact of the solved Ca2+ and Cl– ions acting as additional elastic scattering centers. This seems plausible, considering that at a 0.5 M CaCl2 concentration, every Ca2+ and Cl– ion competes with roughly 100 or 50 water molecules, respectively. However, there is a discrepancy between the β values of O 1s­(l) at low eKEs (50 and 100 eV), i.e., a later but steeper decline for the CaCl2 solution compared to the reference measurements on water. We observed a similar steeper decline for similar measurements of DI water with a trace amount of NaCl conducted with the liquid flat sheet and a circular jet with 80 and 20 μm diameter nozzle apertures, shown in Figure S10 in the Supporting Information. This leads to the conclusion that the difference at low eKE is likely not caused by the different sample geometry, i.e., by the difference between the flat sheet and circular jet measurements. Simulations of the different sample geometries have indeed concluded that while the cylindrical geometry increases the surface sensitivity of the experiment, it does not lead to significant differences for XPS experiments. , Further investigations, beyond the scope of this article, are required and planned to elucidate the discrepancy between our measurements and the reference results. Values of β of the two 2p3/2 core levels of the solvated Cl– and Ca2+ ions shown in Figure B behave very similarly to each other. To the authors’ knowledge, these are the first experimentally determined values of the anisotropy parameter of Ca 2p3/2 and Cl 2p3/2. We sought to validate these unprecedented findings with theoretical results. , The calculations in the dipole approximation do not account for elastic scattering and, therefore, should be an approximation for β of the gas phase ions of Cl 2p3/2 and Ca 2p3/2. Nevertheless, the observed values of β of around 1.4 for eKE between 400 and 600 eV are in good agreement with literature results, showing a maximum discrepancy of less than 10%. , The experimental values are only slightly smaller, and the uncertainty overlaps with the calculated values. This might be due to the fact that values of β of the p core levels are generally smaller compared to those of the s core levels. Assuming a similar decrease of roughly 20% from β of the calculated gas phase of water to experimental liquid water at higher eKEs, we would expect values of around 1.2 for β of Cl and Ca 2p3/2 in the liquid phase at around 600 eV. This value is still in good agreement with our experimental observed β values of Cl 2p3/2 and Ca 2p3/2. At lower eKEs, the experimental values are generally smaller compared to the calculated values, which again shows the increasing influence of the elastic scattering, similar to the β values of O 1s. Additionally, our data hint at slightly lower β values for Cl 2p3/2 compared to Ca 2p3/2 at low eKEs, but generally, most of the uncertainties are overlapping. Determination of the anisotropy parameter using PADs of liquids is rare in the literature, and only a few core levels have been investigated, among them O 1s, C 1s, and S 2p. ,,− With the lack of other experimental data for comparison, it is still reassuring that β of S 2p of various sulfur solutions (0.6 ≤ β ≤ 0.75) is close to our Cl 2p3/2 and Ca 2p3/2 values at an eKE of 100 eV. The determined values of β of Cl 2p3/2, Ca 2p3/2, O 1s­(l), and O 1s­(g) have been used to correct the peak areas derived in the previous section in order to calculate the Ca2+ and Cl– ion concentrations at the liquid–vapor interface. Ca2+ and Cl– Ion Concentration Ratio at the Liquid–Vapor Interface Molecular dynamics (MD) simulations, detailed in the Methods section, have been used to simulate the density profiles of a 0.28 M CaCl2 solution (see the inset of Figure A) at the liquid–vapor interface. Water, Cl– ion, and Ca2+ ion densities are divided by their density in the bulk and are shown over the distance z from the liquid–vapor interface. The interface position was defined to be at the inflection point of the water density. , An APXPS experiment represents the integrated density of the probed elements from the surface to the maximum escape depth of the excited photoelectrons. Therefore, to make the simulations comparable with APXPS results, it is necessary to integrate over the simulated density profiles (additional details on the method can be found in the literature). ,− Specifically, the intensity of element A (I A) can be calculated from 3 IA(Ekin)=αFσT∫0∞ρA(z)exp(−1EAL(Ekin)∫0zρtot(y)ρ0dy)dz which depends on the total density of all elements within the solution (ρ tot) divided by the total density in the bulk (ρ tot(∞) = ρ 0), and the density of element A (ρ A). Additionally, the intensity needs to be corrected for an experimental alignment factor (α), the photon flux (F(hν)), the photoionization cross section (σ(hν)), and the detector transmission function (T(hν)). EAL­(E kin) refers to the depth of the probed solution for photoelectrons with kinetic energy E kin. (A) Cl–/Ca2+ ion ratio derived from the integrated density profiles of MD calculated density profiles of an aqueous 0.28 M CaCl2 solution. The MD calculated density profiles are shown in the inset. (B) Cl–/Ca2+ ion ratio of aqueous 0.25 and 0.5 M CaCl2 solutions derived from corrected Cl 2p3/2 and Ca 2p3/2 peak areas measured with different excitation energies corresponding to eKEs between 100 and 800 eV. From left to right, the information depth into the solution is increasing, showing the most interface sensitive measurement on the left. Dashed lines indicate the expected bulk ion concentration. The experimental alignment factor stays the same within the experiment and cancels out when intensity ratios are formed. For sufficiently large values of EAL­(E kin), the Cl–/Ca2+ ratio has been calculated by dividing I Cl with I Ca, see Figure A. The ion ratio from the MD simulations clearly suggests an increase of the Cl– ion concentration over the Ca2+ ion concentration toward the liquid–vapor interface, relative to the expected ratio of 2 in the bulk of the solution. This trend is experimentally confirmed with the APXPS measurements of the aqueous 0.25 and 0.5 M CaCl2 solutions. By dividing the corrected peak areas of Cl 2p3/2 and Ca 2p3/2 measured with the same eKE, as shown in Figure B, a similar increase of the Cl– ion concentration can be observed, starting at an eKE of around 200 eV. For higher eKEs, i.e., more bulk sensitive measurements, the observed ion ratio of both solutions is close to the expected ratio of 2. The error bars correspond to the uncertainty of the peak areas derived from fits using fityk and the uncertainty of the determined β values. A significant difference in the ion concentration ratio between the two solutions could not be observed experimentally. Interestingly, both solutions suggest a small drop in the ion concentration ratio below values of 2 between 300 and 450 eV, which is not reflected by the MD simulations. Though in the inset of Figure A, there is a region around 5 Å where the Cl– ion concentration drops significantly and the Ca2+ ion concentration increases. Apart from the ion ratios shown in Figure , MD simulations and APXPS can also be used to calculate the overall molar concentrations of Cl– and Ca2+ ions in the solution as shown in Figure by dividing the corrected Cl 2p and Ca 2p peak areas by the corrected O 1s­(l) peak area and by multiplying by the molar volume of water. Here, it becomes unambiguously clear that the Ca2+ ions are repelled from the liquid–vapor interface. MD calculations, as well as both probed solutions, show a decreased Ca2+ concentration toward the interface. The results regarding the Cl– ion concentration are less definite. While simulations suggest an increase of the Cl– ion concentration at the liquid–vapor interface, the APXPS measurements of the 0.25 M CaCl2 solution show a slight decrease, and the 0.5 M CaCl2 solution shows an unchanged concentration at the interface. Toward the bulk, MD simulations and APXPS results show molar ion concentrations close to the expected ion concentrations in the solution indicated with red (for Cl–) and black (for Ca2+) dashed lines in Figure . In general, the APXPS results show larger fluctuations regarding the Cl– ion concentration, which makes the interpretation of the results more challenging. As a caveat, we note that the biggest uncertainties of the measurements result from slight changes in the flat sheet position, which change the overall measured intensity. If such a change occurs in between a measurement of different core levels with the same eKE, the measured molar concentration will be slightly off, and this uncertainty is not reflected in the presented error bars. Overall molar Cl– and Ca2+ ion concentrations derived from APXPS measurements of (A) an aqueous 0.25 M CaCl2 solution and (B) an aqueous 0.5 M CaCl2 solution. (C) The concentrations derived from integrated density profiles of MD simulations of an aqueous 0.28 M CaCl2 solution. From left to right, the information depth into the solution is increasing, showing the most interface sensitive measurement on the left. Another important aspect is the length scale of the MD simulations. By comparing the MD simulations with the APXPS results in Figures and , it seems that our measurements are more surface sensitive than the literature suggests (EAL of ∼1–2 nm at 100 eV). , This discrepancy, at least partially, is related to the translation of the MD calculated density profiles using eq and has similarly been observed in the literature. On a first view, the observed Ca2+ ion depletion at the water interface seems to contrast literature results of calcium enrichment in small sea spray aerosol particles, which originate from the seawater surface. − However, the solutions investigated here contain higher salt concentrations, and other organic or inorganic components are absent compared to seawater. Furthermore, the aerosol particles are exposed to numerous chemical and physical effects like evaporation and condensation cycles that alter their composition after emerging from seawater. Ion pairing of Ca2+ with organic surfactants like carboxylates and/or biological activity in seawater are discussed as the main contributors for the calcium enrichment in sea spray aerosol particles, which is not in conflict with the intrinsic Ca2+ ion depletion at the water interface detected in this work. ,,, In other literature, higher amounts of Ca2+ ions in mixed solutions have been observed compared to the amounts of other solvated ions. However, our results show that this is not necessarily related to higher Ca2+ ion concentrations at the interface since the other ion concentrations are influenced by the interface as well. Experimental and theoretical results have shown that the ion propensity at the interface is related to the Hofmeister series, i.e., large polarizable ions usually favor an enrichment at the interface, while multiply charged ions are repelled from the interface. , The reason is that the electrostatic interactions of multiply charged ions are usually stronger and hence dominate their polarization effects. The electrostatic interactions are better shielded in the bulk, making the ion solvation in the bulk energetically favorable. , The depletion of Ca2+, a doubly positively charged ion with low polarizability, from the interface is in good agreement with these findings. Also, the increased ratio of Cl– over Ca2+ ions observed with MD simulations and APXPS measurements agrees well with these findings since the single negatively charged Cl– ion is more easily polarized compared to the double positively charged Ca2+ ion. Conclusion We have investigated the ion concentration of aqueous CaCl2 solutions at the liquid–vapor interface experimentally by liquid flat sheet APXPS depth profiling and measurements of the photoelectron angular distributions of the solvated ions and water. The photoelectron angular distribution measurements of water and the solvated Ca2+ and Cl– ions show decreased values of β at low eKEs, reflecting the elastic scattering in the solution. At higher eKEs, the β values of the 2p3/2 core levels of the solvated ions are in good agreement with calculated values for the gas phase from the literature, showing only a small impact of elastic scattering effects. A comparison with MD simulations shows a good agreement between theory and experiment, for instance, an unanimously decreased Ca2+ ion concentration at the liquid–vapor interface compared to the expected ratio in the bulk of the solution. On the other hand, the Cl– concentration at the interface is less clear, with APXPS results suggesting a lesser decrease or unchanged Cl– concentration, while MD simulations suggest a small increase. Nevertheless, MD simulations and APXPS results agree on an increased Cl– over Ca2+ ratio at the interface relative to the bulk. Our results underline the importance of the accurate determination of the chemical compositions at liquid–vapor interfaces and their difference from the bulk for a deep understanding of atmospheric aerosol chemistry. Methods Liquid Flat Sheet Setup Aqueous 0.25 and 0.5 M CaCl2 solutions have been prepared by mixing calcium chloride hexahydrate (CaCl2 + 6H2O, Sigma-Aldrich, BioUltra, ≥99.0%) with DI water. The used jet nozzles had an aperture diameter of 33 μm, and a Knauer Blue Shadow 40P HPLC pump was operated with a flow rate of 3.8 mL/min and pressure of around 65 bar to form a liquid flat sheet. Using a liquid nitrogen filled cooling trap, the pressure was kept in the low 10–4 Torr region in the experimental chamber during all of the measurements, but the local pressure in the proximity of the liquid flat sheet was likely higher. Ambient Pressure Photoelectron Spectroscopy The APXPS measurements were conducted using a colliding micro liquid flat sheet system at the Advanced Light Source (ALS) at beamline 11.0.2.1 in the LARaXS endstation described elsewhere. The depth profiling measurements were conducted with horizontally linear polarized light, which overall delivers the highest achievable photon flux, resulting in the best signal-to-noise ratio. In this setup, the electric field vector of the linear polarized light is approximately at an angle θ = 35.3° with respect to the analyzer direction. Peak areas have been derived from core level fits with Voigt functions and a linear background function using the program fityk 0.9.3. All core level fits can be found in the Supporting Information in Figures S3–S5 and Figures S7–S9. The Voigt functions were restricted to use the same Gaussian full width at half-maximum (fwhm) for each excitation energy and core level (see Figure S6). The Lorentzian fwhm of the fits was restricted to be the same for each core level. The fit delivered a value of 0.11 eV for Ca 2p, 0.17 eV for Cl 2p, and 0.25 eV for O 1s compared to calculations from core hole lifetimes of 0.17 ± 0.04 eV for the Ca 2p, 0.09 ± 0.02 eV for the Cl 2p, and 0.10 ± 0.04 eV for the O 1s core level. , The area of the 2p1/2 peak was restricted to be half of the 2p3/2 peak area for all of the Cl and Ca 2p core level fits. The derived peak areas were divided by the differential photoionization cross section, and the photon flux was derived from electric currents measured with a photodiode in measurement position for each excitation energy prior to the measurements of the solutions. Values for the photoionization cross section have been taken from Trzhaskovskaya et al. calculated in the quadrupole approximation. , The anisotropy parameter of the O 1s, Cl 2p, and Ca 2p core levels used for all solutions was derived from measurements of an aqueous 0.5 M CaCl2 solution using horizontally and vertically polarized light and light that was polarized at an angle of 45°, i.e., right between horizontally and vertically polarized light. These polarizations correspond to angles θ = 35.3°, θ = 56.2°, and θ = 91.7°. Whether the photoionization values for the elements in the gas phase are also correct for solvated ions in liquids is still up for debate, but due to the lack of other alternatives, it is the best approach. , All spectra were corrected with the analyzer transmission function before any fitting procedure was applied. The binding energies were adjusted with valence band measurements performed at each excitation energy after measuring the core levels and adjusting the energy of the 1b1(l) peak to 11.33 eV. Additional details regarding the fit of the valence band can be found in the Supporting Information in section S4 with an example shown in Figure S11. MD Simulations MD simulations were carried out using a slab geometry system containing 768 water molecules and CaCl2 at a concentration of 0.28 M, maintained at 300 K in a box of size 22.8 × 24.8 × 111.8 Å3. The system was modeled using a polarizable force field based on the symmetrized Drude oscillator approach. , The equations of motion were integrated using the velocity Verlet algorithm within an extended Lagrangian dynamics framework with a low-temperature (1 K) thermostat for the Drude oscillators. After a 5 ns equilibration, production runs of 15 ns were conducted to generate density profiles. These profiles were obtained by binning molecular positions across the slab at 1 fs intervals and subsequently averaging the data. All molecular dynamics simulations were carried out in LAMMPS. We used a polarizable classical force field for the MD simulation presented in this article. Periodic boundary conditions were applied along all directions. The SWM4-NDP water model was used. The ion force fields were adapted from the literature. − A Drude oscillator model was used to replicate the polarization. The non-bonded interactions were modeled with a Lennard-Jones potential. A rigid body dynamics for water was used with a time step of 1 fs. The Drude oscillators were fixed with a harmonic spring constant of 1000 kcal/mol/Å2. Thole damping was added to modulate the electrostatic interaction between the particles and induced dipoles. The Lennard-Jones interactions were truncated and shifted at 12 Å. Supplementary Material