Adsorption of Acetic Acid on Ice Studied by Ambient-Pressure XPS

Article pubs.acs.org/JPCA

Adsorption of Acetic Acid on Ice Studied by Ambient-Pressure XPS and Partial-Electron-Yield NEXAFS Spectroscopy at 230−240 K Adéla Křepelová,† Thorsten Bartels-Rausch,† Matthew A. Brown,‡ Hendrik Bluhm,§ and Markus Ammann*,† †

Laboratory for Radiochemistry and Environmental Chemistry, Paul Scherrer Institute, 5232 Villigen, Switzerland Institute for Chemical and Bioengineering, ETH Zürich, 8093 Zürich, Switzerland § Chemical Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States ‡

ABSTRACT: Ice plays a key role in the environment, and the ice−air interface influences heterogeneous chemical reactions between snowpack or cirrus clouds and the surrounding air. Soluble gases have been suspected to affect the topmost, disordered layer on ice (often referred to as a quasiliquid layer, QLL). Changes are especially expected in the hydrogen-bonding structure of water in the presence of solutes at the ice surface. Here, we used ambient-pressure X-ray photoelectron spectroscopy (XPS) to detect acetic acid at the ice surface at 230−240 K under atmospheric conditions for the first time. Electron-kinetic-energydependent C 1s spectra indicate that acetic acid remains confined to the topmost ice surface layers. Spectral analysis provides information about the protonation state of acetate at the ice surface. Surface-sensitive Auger-electron-yield C-edge near-edge Xray absorption fine structure (NEXAFS) spectra were recorded to probe the molecular state of the adsorbed species. The O-edge NEXAFS spectra show only minor differences between clean ice and ice with adsorbed acetic acid and thus indicate that acetic acid does not lead to an extended disordered layer on the ice surface between 230 and 240 K.

1. INTRODUCTION

Previous experiments directly addressing CH3COOH uptake to ice in the environmental-science context were based on measuring the loss of CH3COOH to an ice surface in a flow tube.9−13 They established that the adsorption of CH3COOH can be reasonably well described by a reversible Langmuir-type adsorption.14 Molecular dynamic simulations and density functional theory (DFT) calculations revealed potential configurations for the adsorption of CH3COOH and its dimers.10,15,16 IR spectroscopy17,18 as well as metastable impact electron spectroscopy and ultraviolet photoelectron spectroscopy (UPS)19 were performed at very low temperature and mostly under high coverages or even with multilayer deposits on ice. These measurements have provided valuable information on hydration and dimerization compatible with theoretical estimates. So far, no direct, local spectroscopic information of

Ice is ubiquitous in the environment in the form of snow, sea ice, or cirrus clouds and can play an important role as a catalyst for reactions between atmospheric trace gases or as a trace-gas scavenger.1 Acetic acid (CH3COOH), together with formic acid, is one of the most abundant oxygenated volatile organic compounds in the atmosphere and is a major contributor to free acidity in precipitation.2 CH3COOH is a weak acid that is fairly soluble in water and occurs at typical mixing ratios of several hundred parts per trillion by volume (pptv) in the free troposphere. The sources of CH3COOH are direct biogenic and anthropogenic emissions and oxidation of other volatile organic compounds. The main sinks are reaction with OH radicals and slow photolysis,3,4 as well as uptake to ice and aqueous particles. CH3COOH is among the oxygenated organic compounds actively involved in the chemistry of polar snow packs that affect the oxidation capacity of the atmosphere and the way such compounds are stored into glacier ice, from which past climates are reconstructed.5−8 © XXXX American Chemical Society

Received: October 16, 2012 Revised: December 13, 2012

A

dx.doi.org/10.1021/jp3102332 | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

an organic acid on the surface of ice at environmentally relevant conditions (i.e., high temperature) has become available. Because hydration of the acidic group of CH3COOH is an important driver of its adsorption properties, the interaction of the partially hydrated CH3 COOH molecule with the quasiliquid layer (QLL) and its effect on the latter are of significant interest. Past studies have suggested that strong acids have a significant impact on this interface.1,20,21 The near-edge X-ray absorption fine structure (NEXAFS) technique has become an established tool for investigating the hydrogenbonding environment in water and ice (see refs 22 and 23 and references therein). When electrons emitted from the surface are detected (rather than fluorescence or transmitted photons), the method becomes surface-sensitive because of the inherently low inelastic mean free path and thus escape depth of lowenergy electrons. Partial-electron-yield O K-edge NEXAFS measurements on ice in equilibrium with its vapor pressure allowed the thickness of the QLL to be probed in the temperature range of 220−270 K.24 At the same time, the NEXAFS technique can also provide essential information about adsorbates, for example, at the carbon edge, and their molecular environment and geometric configuration.25,26 X-ray photoelectron spectroscopy (XPS) is among the most sensitive probes of surface elemental composition that is also chemically specific. Synchrotron-based XPS allows precise elemental ratios to be obtained as a function of depth into the sample. Recent experiments also suggest that XPS can be used to probe the degree of dissociation of acids, including that of acetic acid.27−31 The suitability of using the combination of these two surfacesensitive spectroscopic techniques in the context of ice surface chemistry was demonstrated in our previous studies.25,32 Starr et al.25 probed the adsorption of acetone as a function of its gasphase pressure. The C K-edge NEXAFS spectra were compared to DFT calculations to derive information about the adsorption configuration. At the oxygen K-edge, no changes to clean ice were observed, indicating that no significant changes to the electronic structure of the top few layers of ice occurred upon acetone adsorption. In contrast, HNO3 or nitrate ions induced changes in the pre-edge region as well as the main and postedge features of the O K-edge, which indicated that nitrate ions locally assemble a solutionlike hydration shell under conditions where a macroscopic solution is not stable.32 In the present study, we used XPS and NEXAFS spectroscopy to investigate the electronic structure of CH3COOH, its degree of dissociation, its hydrogen-bonding environment, and its depth profile at the ice surface at 230 and 240 K in equilibrium with the water vapor pressure and with atmospherically relevant pressures of CH3COOH. We explicitly avoided high partial pressures of CH3COOH, where multilayers, solid hydrates, and solutions become likely thermodynamic stable phases.

Figure 1. Schematic representation of the measurement configuration in the XPS end station at beamline 11.0.2 at the Advanced Light Source (ALS) with the gas dosing system.

beamline and of general instrument characteristics is provided in ref 33, and more detailed information about the preparation of ice films and their exposure to trace gases is available in our previous reports.25,32 Ice films were grown from water vapor from a liquid water source that had been degassed by at least three pump−freeze− thaw cycles. The base pressure of the chamber before the experiments was in the 10−8 mbar range. After the vacuum pumps of the analysis chamber had been shut off, water vapor was admitted through one of the calibrated leak valves until the pressure in the chamber reached the vapor pressure of ice (0.13 or 0.24 mbar) at the temperature (233 or 239 K, respectively) planned for the experiment. The pressure was measured using a capacitance manometer with a precision of 0.01 mbar. At constant pressure, the flow rate of water vapor through the leak valve into the chamber exactly matched the flow rate out through the aperture into the electrostatic lens system. To grow ice, the temperature of the Cu substrate was set to within a few kelvin below the equilibrium ice temperature at the given vapor pressure. This first leads to the growth of an optically opaque ice film, which over time (about 1 h) transforms into a transparent polycrystalline ice film; for details, please see refs 25 and 32. We note that, because of the proximity of the sampling aperture of the electron analyzer to the sample, a slight gradient of the pressure field occurs, leading to a slightly reduced pressure at the sample surface of at most 1−2%.34 We never observed a difference between the measured water vapor pressure and that derived from the substrate temperature35 within the experimental uncertainty, probably because the gradient was canceled out through the combination with other effects, such as radiative heating by the aperture cone or the temperature gradient within the ice film. The gas-phase composition was monitored by a differentially pumped quadrupole mass spectrometer (QMS; see Figure 1) with a pressure differential of about 105. For all XPS measurements, the dwell time and pass energy were set to 0.2 s and 10 eV, respectively. Even though the absolute incident photon flux and the photon flux density were minimized, photoelectron emission led to positive charging of the ice surface of several electronvolts, observable as a shift of the spectra to lower kinetic energy (KE) or apparently higher binding energy (BE), especially for the clean ice sample. All XPS BEs were referenced to the BE of the O 1b1 valence-band

2. EXPERIMENTAL SECTION The experiments were performed at the Advanced Light Source (ALS) at Lawrence Berkeley National Laboratory using the ambient-pressure photoelectron spectroscopy end station at beamline 11.0.2, which allows XPS to be performed at pressures up to a few millibar.33 This allows XPS experiments on an ice surface that is in equilibrium with the gas phase at temperatures above 200 K. The setup consists of an XPS chamber (see Figure 1) and a preparation chamber with standard surface analysis tools (not shown). A detailed description of the B



Article

peak at 6.5 eV.25 Such valence photoemission (PE) spectra were recorded for this purpose after each O 1s and C 1s spectrum at the same photon energy. To obtain depth profiles of the chemical composition of the sample, the incident photon energy was varied to measure these spectra with photoelectron kinetic energies between 200 and 800 eV. The stoichiometry as a function of depth was calculated from the ratios of peak areas from the spectra of each element taken at the same kinetic energy. The calibration of C/O elemental ratios was performed by admitting 0.65 mbar of gas-phase CO2 into the chamber and recording C 1s and O 1s gas-phase photoemission spectra (in the absence of a sample in front of the analyzer sampling orifice) at the same photon energies as used for the experiments with ice. Partial-electron-yield NEXAFS spectra were measured at the oxygen and carbon K-edges using kinetic energy windows of 390−410 and 190−210 eV, respectively, which are on the background of the Auger lines of oxygen and carbon, respectively. The dwell time was set to 2 s for both oxygen and carbon K-edge NEXAFS measurements. Background measurements (I0) were taken with the clean gold substrate at base pressure before growing an ice film at the O edge and with the clean ice film at the C edge. All NEXAFS spectra were normalized to their integrated areas. Once stable conditions were established, O 1s and C 1s photoemission (PE) spectra and O and C K-edge NEXAFS spectra of clean ice were recorded at a temperature of 233 K and a pressure of 0.133 mbar. After spectroscopic characterization of the clean ice situation, the temperature was kept at 233 K, and CH3COOH was dosed into the chamber. Gas-phase CH 3 COOH was delivered from a permeation source (Dynacalibrator model 190, VICI Metronics Inc.) in a configuration similar to that of our previous flow-tube experiments.9 For the present experiments, the permeation oven was set to 60 °C, and N2 was used as a carrier gas. The source was calibrated offline by a TEI CO2 analyzer (model) after conversion of CH3COOH to CO2 over CuO at 600°. The CO2 analyzer itself was calibrated with a CO2 standard. The identity of the volatile organic compound emitted from the permeation source was also confirmed offline by chemical ionization mass spectrometry excluding the presence of other organic gases.9 The emission strength was determined to be (1.0 ± 0.24) × 1016 molecules min−1. The N2 flow was set to 4 mL/min at ambient pressure, leading to a CH3COOH/N2 molecular ratio of 1.0 × 10−4 in the gas admitted to the chamber through a calibrated leak valve. N2 in the chamber could be monitored by the QMS at m/z 28 and, at higher pressures, through the total pressure difference measured by the baratron gauge with respect to that measured when only water vapor was present. We aimed at a CH3COOH partial pressure such that subsaturated to saturated surface coverages based on the currently recommended adsorption equilibrium constant (105 cm at 233 K14) were achieved. This leads to partial pressures of at most 10−5 mbar, which is below the detection limit of the QMS. We therefore estimated its partial pressure from the N2 pressure and the CH3COOH/N2 ratio given above. The experiments were complicated by the very slow passivation of the large stainless steel chamber walls at these low CH3COOH pressures. Indication of this was the very slow growth of the C 1s XPS signals (see below) detected at the ice surface after the leak valve had been opened to admit the CH3COOH/N2 mixture. Therefore, the system was allowed to equilibrate for several hours at each total pressure setting before

stable PE and NEXAFS spectra were obtained. Experiments at the two different temperatures were separated by overnight evacuation of the chamber. C K-edge NEXAFS spectra from acetic acid in solution were obtained from separate experiments performed at the PGMU41 beamline of BESSY using a liquid microjet.36,37 A detailed description of the C K-edge X-ray absorption spectroscopy (XAS) measurement was published previously.31

3. RESULTS AND DISCUSSION 3.1. Characterization of the Clean Ice Surface. Photemission (PE) survey spectra recorded at 490 and 735 eV did not show elements other than oxygen; specifically, no carbon contamination was detected before the first admission of CH3COOH. Examples of an O 1s PE spectrum at a kinetic energy of 200 eV and an O K-edge NEXAFS spectrum of clean ice measured at 233.4 K and 0.13 mbar are shown in Figure 2.

Figure 2. (a) O 1s PE spectrum (symbols) measured at a photon energy of 735 eV and single Gaussian fitted to the data (solid line). (b) Oxygen K-edge partial-electron-yield NEXAFS spectrum of clean ice at 233 K.

The O 1s peak is at a BE of 533.4 eV, consistent with previous studies,32 and shows some broadening (1.7 eV fwhm), which could be due to inhomogeneous charging of the ice surface. The oxygen K-edge NEXAFS spectrum of clean ice (Figure 2) agrees well with that reported in previous studies by us25,32 and others.24,38 The pre-edge peak at 535 eV is characteristic for broken hydrogen bonds in single donor configuration, and the broad band around 543 eV is assigned to the contribution of 4fold-coordinated water molecules.38 The absence of any peaks below 535 eV in Figure 2b, which would be characteristic of a 1s−π* transition in compounds with CO or NO bonds, confirms the clean conditions of our experiment before the first admission of CH3COOH to the chamber. A representative C C



Article

appearance of the sample were observed when compared to the ice film before admission of CH3COOH. The experimental conditions and results of XPS and NEXAFS measurements are summarized in Table 1. As explained in the Experimental Section, the pressure of CH3COOH in the chamber was estimated from the measured N2 pressure and the calibrated CH3COOH/N2 ratio delivered by the permeation source, assuming that full equilibration with the chamber walls occurred. Examples of O 1s and C 1s PE spectra of ice after CH3COOH admission to the chamber recorded at photon energies of 735 and 490 eV, respectively, are shown in Figure 3. Figure 3a illustrates the O 1s PE spectrum obtained at a temperature of 233.4 K and a pressure of 1.33 mbar (experiment II). The BE is at around 533.2 eV, that is, similar to that of clean ice, but the peak has broadened. The O 1s spectrum of the surface in the presence of CH3COOH is characterized by the strongly overlapping contributions of oxygen in ice (533.1 eV) and the two oxygens in the COOH group of CH3COOH, expected at 532.3 and 533.7 eV, respectively, based on data from carboxyl-terminated thiol monolayers on Au(111).39 Because of the strong overlap, we do not derive quantitative information about the composition from the O 1s spectrum. When taking the COOH-to-H2O ratio from the measured C/O ratios (see below and Table 1) and assuming the shape of the ice O 1s peak to be the same as that observed for pure ice shown in Figure 2a, a reasonable fit is obtained (Figure 3a), with the higher-BE peak from COOH shifted to a slightly higher value, 534.3 eV. We did not consider elaborating on differentiating protonated and deprotonated forms of acetic acid in the O 1s spectra. The O 1s PE spectra taken at other positions on the sample surface, at other N2/ CH3COOH pressures, and at 239.1 K did not exhibit significant differences in peak shape or BE within the uncertainty of this analysis. For further analysis, only the total O 1s peak area was used to calculate elemental C/O ratios. The C 1s PE spectrum measured at 233.4 K (total pressure 1.33 mbar) and at a kinetic energy of 200 eV before (gray dots) and after admission of CH3COOH to the chamber is shown in Figure 3b. The spectrum associated with CH3COOH on ice has two main features at BEs of around 285.5 and 289.3 eV, consistent with BEs for aliphatic carbon and carboxyl carbon.39−42 The peak separation is comparable to that of gas-phase CH3COOH (3.83 eV43). Especially in the carboxyl region, two contributions are clearly apparent. Recently, Ottosson et al.29 reported C 1s spectra from acetate solutions

1s spectrum taken under these clean conditions is also included in Figure 3, along with the acetic acid spectra.

Figure 3. Measured (symbols) O 1s and C 1s spectra of ice exposed to acetic acid at 233.4 K. (a) O 1s. The red line is the result of fitting Gaussian peaks to represent O in H2O (blue, 533.1 eV) and O in COOH (green, 532.3 and 534.3 eV). (b) C 1s. The blue line is the sum of adjusted Gaussians representing protonated acetic acid (green) and deprotonated acetate (red) constrained by data obtained by Ottosson et al.29 in aqueous solution. The gray dots represent a spectrum of the C 1s region before the admission of acetic acid to the chamber.

3.2. Adsorption of Acetic Acid at the Ice Surface. CH3COOH was admitted to the chamber to lead to total pressures between 0.133−1.333 mbar, that is, to a maximum of 1 mbar of the N2/CH3COOH mixture in the chamber while keeping the water vapor pressure constant. The ice film remained stable over many hours, and no visible changes in the

Table 1. Experimental Conditions and Summary of Results of XPS Measurementsa experiment

T (K)

P (mbar)

PAA+N2 (mbar)

I II III IV V VI VII

233.4 233.4 233.4 233.4 233.4 239.1 239.1

0.13 1.33 0.67 0.27 0.14 0.24 1.33

− 1.20 0.53 0.13 0.003 0.24 1.20

PAA (mbar) − 1.01 4.49 1.13 2.52 8.99 1.01

× × × × × ×

10−5 10−6 10−6 10−8 10−7 10−5

IC/IO (200 eV)

n1C/n1O

nAA/nH2O

− 0.19 0.10 0.22 0.15 0.077 0.097

0.25 0.14 0.28b 0.18b 0.081 0.12b

0.17 0.083 0.19b 0.11b 0.044 0.069b

NAA (cm−2) 5.4 3.3 5.8 4.1 1.9 2.8

× × × × × ×

1014 1014 1014 b 1014 b 1014 1014 b

a

P, total pressure in the chamber; PAA+N2, partial pressure of the mixture of acetic acid and nitrogen calculated as the difference between the total pressure in the chamber and the vapor pressure of ice before admission of acetic acid; PAA, pressure of acetic acid calculated from PAA+N2 and the partial pressure ratio of acetic acid to nitrogen (1.0 × 10−4); IC/IO, C 1s (490 eV) to O 1s (735 eV) photoemission signal ratio, referenced to gasphase CO2; n1C/n1O, carbon-to-oxygen atomic ratio in top layer from fit to depth profile using eq 5; NAA, surface coverage of acetic acid in molecules cm−2. bValue calculated assuming n1C/n1O = 1.25 × IC/IO (average for experiments II, III, and VI). D



Article

as a function of pH, indicating that the protonated and deprotonated forms of acetic acid can be differentiated through significant shifts of the apparent C 1s binding energies. In particular, the shift between protonated and deprotonated forms was larger at the carboxyl carbon than at the methyl carbon. Based on these results, we fit the C 1s spectrum with four Gaussians using constraints similar to those reported by Ottosson et al., namely, equal full widths at half-maximum and areas for the two associated peaks and shifts of −1.2 and −0.7 eV between protonated and deprotonated forms for carboxyl carbon and methyl carbon, respectively. Our fits showed a somewhat smaller separation between carboxyl and methyl carbon (3.6 eV) compared to 3.9 eV as reported in Ottosson et al.29 However it is within the range observed for acetic acid on other substrates.26,39,41,42 Even though this fit catches the overall shape, there is still some additional intensity in the methyl carbon region at around 285 and 288 eV. The residual between the measured and adjusted spectra in this energy region varied randomly during the course of the experiments. We consider three potential contributions to this unexplained intensity: We hypothesize that a preferential orientation of the molecules with the methyl group toward the gas phase and the carboxyl group into the ice phase could lead to different photoemission intensities, that is, the photoelectrons from carboxyl C 1s would experience more scattering losses, which would lead to methyl-to-carboxyl C 1s peak ratios deviating from 1. Because the ice surface is not flat owing to its polycrystalline nature, measurements at different positions probe at slightly differing angles to the surface normal of individual ice grains so that the ratio of the two contributions can vary. Another option is a contaminating species from the acetic acid source, the lines to the leak valve, or secondary chemistry on the chamber walls. A carbonyl species deriving from these hypothetical sources would have a binding energy of around 288 eV and could therefore contribute to intensity in that region. There is also a possibility that some of the CH3COOH is decomposing in the beam. In our previous investigation with acetone, a small shoulder on the low-bindingenergy side of the methyl carbon peak was assigned to such contamination.25 There is only a very small residual between the fitted and measured spectra in that binding energy range. Direct radiolysis of acetic acid (decarboxylation) at the ice surface could also lead to a carbonyl, but HCHO has a very weak propensity for the ice surface14 and would desorb and remain undetected. Further information about the electronic structure and identity of adsorbed acetic acid and potential other species present can be expected from the C K-edge NEXAFS spectra discussed next. The carbon-edge partial-electron-yield spectra shown in Figure 4 for experiments II, IV, and VI at different CH3COOH pressures and temperatures are very similar to each other. No significant difference was found for the other conditions listed in Table 1 as well. The sharp resonance at 288.6 eV is due to the C 1s-to-π* transition of the carboxyl carbon. The broader spectral feature in the region between 292 and 300 eV is assigned to the 1s → σ* transition. The dips at about 285 eV and slightly above 290 eV are likely due to imperfect normalization to I0, which could only be measured prior to any admission of CH3COOH to the chamber (up to several hours before). There is a minor feature at about 287 eV, which could be assigned to a carbonyl44 that we mentioned above as a contamination possibly contributing to the C 1s PE spectrum. The spectra are very similar to those known from the literature

Figure 4. Carbon K-edge partial-electron-yield NEXAFS spectra of ice exposed to acetic acid for experiments II (red), IV (green), and VI (blue) and of acetic acid (pH 1, dark red) and acetate (pH 9, black) aqueous solution. For details, see Table 1 and the text.

for CH3COOH adsorption on Si;40 the similarity in position and width of the resonant line with the corresponding gasphase spectrum45 would, at first glance, indicate only weak adsorbate−adsorbate or adsorbate−substrate interactions. Based on theoretical studies,10,15 acetic acid monomers and open dimers (hydrogen-bonded at one end) are the more stable adsorbates on ice compared to closed dimers (hydrogenbonded at both ends). The NEXAFS spectra are likely not sufficiently sensitive to differentiate between open dimers and monomers. Note that, at the low partial pressures of acetic acid in the chamber, the monomer form clearly dominates in the gas phase, as discussed in detail by Kerbrat et al.9 Surprisingly, the NEXAFS spectrum does not display any splitting of the resonant line as might be expected from the significant shifts observed in the PE spectra that differentiate protonated and deprotonated acetate at the ice surface. To confirm this point, reference C K-edge partial-electron-yield NEXAFS spectra of acetic acid in aqueous solution at high and low pH are also shown in Figure 4. These spectra were collected using a liquid microjet31,37 in separate experiments. The idea that a single C 1s−π* resonance is recorded when both protonation states are present would at first thought be rather surprising. However, Brown et al.31 made the same observation for formic acid solution in a more detailed study that included XPS, XAS, and DFT calculations. Similarly to the case of acetic acid, the C 1s binding energy of formate decreases relative to that of formic acid. Their theoretical calculations suggest that the deprotonation process to form formate results in a destabilization of the π* orbital (lowest unoccupied molecular orbital, LUMO) that increases the energy of the LUMO. This increase in energy exactly offsets the decrease in C 1s binding energy so that the absorption process occurs at the same energy as for neutral formic acid. The same holds for acetic acid solutions and apparently also for acetic acid at the ice surface. Within the noise level of these NEXAFS spectra, no indication of other carbon species is apparent. We now use the tentative interpretation of the XPS data to obtain a rough estimate of the degree of protonation of acetic acid at the ice surface. The average fraction of the higherbinding-energy species (i.e., protonated acetic acid) of the total intensity was around 60% based on the fits as presented in Figure 3b. For a hypothetical aqueous solution at 233 K, the acid dissociation constant would be Ka = e(−ΔG/RT) = 8.2 × 10−7,46 and the degree of dissociation of acetic acid, α ≈ E



Article

(KacCH3COOH)1/2/cCH3COOH, would be about 10−4 at 1 M and 2 × 10−2 at 10−3 M, each about a factor of 10 lower than at room temperature. The Henry’s law constant for acetic acid extrapolated to 233 K is about 106 M atm−1,47 so that, at CH3COOH pressures of 10−6 mbar as in the present experiment, the equilibrium concentration in this hypothetical supercooled solution would be about 10−3 M. Therefore, the degree of dissociation estimated from the PE spectra is rather higher than the degree of dissociation in a dilute aqueous solution in equilibrium with the same CH3COOH partial pressure and far from the expected dissociation for a locally more concentrated aqueous solution at the ice surface. O 1s and C 1s PE spectra obtained at different photoelectron kinetic energies (200−800 eV) were measured to obtain depth profiles of the O/C ratio, relying on the kinetic energy dependence of the inelastic electron mean free path (∼1.4−3.5 nm).48 The measured peak area ratios were normalized to corresponding gas-phase peak area ratios measured for CO2 at the same photon energies to correct for cross sections, photon flux, and energy-dependent electron analyzer sensitivity. The results are depicted in Figure 5 as atomic C/O ratios for three

atoms per unit volume and n1O O atoms per unit volume is sitting on top of an infinitely thick pure ice layer 2 containing n2O O atoms per unit volume. Because the density of acetic acid solutions deviates by only a few percent from that of pure water, we assume that C and O atoms contribute by equal amounts to the volume of the layer. We normalize the concentrations to that of the pure ice layer 2, that is, n2O = 1. Mass balance then requires n1C + n1O = 1, or n2O = n1C + n1O. The two layers are assumed to have the same density and thus similar inelastic mean free path. The partial photoemission signal, I1Cx, contributed by C atoms contained in layer 1 at depth x can be given by I1Cx(E kin) = I1C(0, E kin)e−x /[λ cos(θ)]

(1)

where λ is the inelastic mean free path, which depends on Ekin, and θ is the take-off angle at which the electrons were measured, namely, 42° in our experiments. Integration to depth d gives the total signal from C atoms contained in layer 1 I1C(Ekin) ∝ n1C(1 − e−d /[λ cos(θ)])

(2)

The photoemission signal attributed to oxygen has two contributions, one from layer 1 and one from layer 2. The contribution from layer 1 is analogous to that from C atoms I1O(E kin) ∝ λ cos(θ )n1O(1 − e−d /[λ cos(θ)])

(3)

For the contribution from layer 2, eq 1 is integrated from d to infinity, yielding I2O(E kin) ∝ λ cos(θ )n2Oe−d /[λ cos(θ)]

(4)

Therefore, the ratio of the total photoemission signals from C and O atoms out of the two layers becomes n

1C (1 − e−d /[λ cos(θ)]) IC(E kin) n1O = n IO(E kin) 1 + n1C e−d /[λ cos(θ)]

(5)

1O

Figure 5. Measured ratios of C 1s to O 1s photoemission signals (referenced to 0.5 for CO2) as a function of electron kinetic energy (symbols) for experiments II, III, and VI (red, blue, and green, respectively). The colored lines represent fits of eq 5 to the data for d = 1.5 nm (thick solid), d = 0.5 nm (thin solid line), and d = 2.5 nm (dashed lines). The black line is the calculated inelastic mean free path taken from ref 48.

Fits to the data with d and the n1C/n1O ratio as variables are also shown in Figure 5. The slope is reasonably well reproduced with d = 1.5 nm (thick solid lines) and n1C/n1O ratios of 0.25, 0.14 and 0.08, indicating that acetic acid might enter the top few H2O bilayers to some degree. Thin solid and dashed lines display simulated profiles for 0.5 and 2.5 nm, respectively. Therefore, the profiles indicate that acetic acid is unlikely to reach depths more than 2 nm. This is consistent with coatedwall flow-tube experiments that showed reversible adsorption with relatively short desorption life times.9,12 For those cases for which we measured the profiles, the fits directly return the n1C/n1O ratios, listed in Table 1. The n1C/n1O atomic ratios can be converted to acetic acid-to-water molar ratios, X, by considering the fact that the CH3COOH molecule contains two carbon and two oxygen atoms nCH3COOH n 1 X= = 1C n H 2O n1O 2 1 − n1C n

experiments (II, VI, VII) with different amounts of CH3COOH on the ice surface and for two different temperatures. The error bars represent the overall error, including systematic errors due to uncertainties of the inelastic mean free path or the C/O ratio calibration. For the case with the lowest surface coverage, the signal at 800 eV kinetic energy was too weak to be safely quantified. Because the location of adsorbed gases at the ice surface is a debated issue in the environmental science community, especially for acidic gases, it has been suggested that they enter the ice surface to some degree and thereby induce a thicker QLL (refs 1 and 49 and references therein). Here, we did not make the a priori assumption that acetic acid remains at the top of the ice surface to quantify the surface coverage, as in our earlier study of acetone adsorption,25 but instead retrieved the depth profile of CH3COOH from the measured kinetic energy dependence. A simple two-layer model was applied to fit the measured data. A surface layer 1 of thickness d (nm) containing n1C C

(

1O

)

(6)

Values for X are also included in Table 1. NCH3COOH, the total number density of CH3COOH molecules in the layer normalized to the unit surface area, can then be obtained as NCH3COOH =

ρd

(mCH COOH + 3

F

mH2O X

)

(7)



Article

where ρ is the density (g cm−3) and mCH3COOH and mH2O are the masses (g) of CH3COOH and H2O molecules, respectively. The resulting coverages for experiments II, III, and VI are 1.9 × 1014, 3.3 × 1014, and 5.4 × 1014 molecules cm−2, respectively. We can also estimate the surface coverages for the other experiments, for which only the signal ratio at 200 eV was measured, by assuming that the proportionality factor between n1C/n1O and IC/IO (1.25 on average for experiments II, III and VI) remains constant (numbers in italics in the last three columns of Table 1). It should be kept in mind that this analysis carries some systematic uncertainties, because it does not include any variation of the inelastic mean free path due to the changing composition or any effect of a preferred orientation of the acetic acid molecules at the interface. The coverages reported in the table tend to decrease with decreasing CH3COOH pressure, indicating subsaturated surface coverage under those conditions. In view of the uncertainties and in the absence of a direct acetic acid measurement in the gas phase, we refrain from quantitatively estimating isotherm data. We only compare the maximum coverage observed here with that obtained from flow-tube experiments11 (2.8 × 1014 molecules cm−2), which is in the range of our measured coverages. The oxygen K-edge NEXAFS spectra of ice are sensitive to the local hydrogen-bonding structure and, thus, can help to assess whether acetic acid adsorption has an effect on the extent of the QLL or surface premelting. Partial-electron-yield NEXAFS spectra measured in the presence of CH3COOH (experiments II−VII in Table 1) are shown in Figure 6,

broad feature between 538 and 550 eV.45 If we take into account the contribution of oxygen in acetic acid to the residual in Figure 6a, a substantial reduction in intensity between 539 and 544 eV remains. The extent of this residual is comparable for the differential between the spectra of the other experiments and that of clean ice. This reduction in intensity in the postedge region is reminiscent of the difference between the ice and liquid-water absorption spectra. In the latter, the ratio of the postedge to the main edge is reduced in comparison to that of ice.24 This would indicate that, in comparison to clean ice, a small fraction of the water molecules on the surface are hydrogen-bonded similarly to those in an aqueous solution, presumably those that are engaged in hydrating the COOH group. Overall, the contribution by aqueous-solution-like water molecules observed here is substantially lower than that observed previously for the case of nitrate at the ice surface,32 where we estimated that about 20% of the water molecules over the probe depth of around 2.4 nm are engaged in hydrating the nitrate ion. The present results demonstrate that acetic acid does not perturb the surface of ice strongly and does not induce an extension of the QLL thickness at the temperature of the experiments. The behavior observed here (minor amounts of hydrating H2O molecules) in relation to that of nitrate (20% of interfacial H2O molecules engaged in hydrating nitrate32) and to that of acetone (no change reported at the O K-edge,25) can be compared to the room-temperature solubilities of the three compounds, which differ by orders of magnitude (2.6 × 106 M atm−1 for HNO3,50 4.1 × 103 M atm−1 for CH3COOH,47 and 3.0 × 101 M atm−1 for CH3COCH351). The likely large degree of dissociation of HNO3 adds further demand for water to hydrate the dissociated ions. Similarly, the standard free energies of hydration differ significantly (−28.5 kJ/mol for acetic acid and −16.12 kJ/mol for acetone in the gas phase52). The two properties might not be entirely representative for the specific environment at the ice surface, but can still provide an explanation for the order of the amount of hydrating water associated with adsorbed molecules of these species at the ice surface. Kuo et al.49 developed a thermodynamic model of liquid brine layers in equilibrium with ice for soluble gases. For the case of HNO3 at low partial pressures, this model does not predict formation of a brine layer, consistent with the HNO3− ice phase diagram and with our earlier study.32 However, that model does not address the thermodynamics of a quasiliquid layer, for which other concepts are required.53 In view of the general lack of spectroscopic studies probing the ice−air interface with nanometer resolution, we hope that this work adds to a so far small database that will allow for a more quantitative description of gas−ice interactions at a molecular level.

Figure 6. Oxygen K-edge partial-electron-yield NEXAFS spectra of clean ice (I) and ice with different CH3COOH surface coverages measured at two temperatures. For details, see Table 1 and the text.

together with the clean-ice (experiment I) measurement. Each spectrum was intensity-corrected for I0 and area-normalized over the range of 534−560 eV for the purpose of comparison. The most obvious difference occurs at 532 eV, indicating the contribution of oxygen of adsorbed CH3COOH (O 1s 1a1 → 2b1* transition). There are, furthermore, small differences in the intensities observable at the pre-edge (535 eV), main edge (537 eV), and postedge (542 eV) and in the region of the intensity drop around 548 eV. As an example, the difference between the spectrum obtained for experiment III and that for clean ice is also shown on a magnified scale in Figure 6. To our knowledge, the solution-phase oxygen K-edge absorption spectrum of acetic acid has not been measured. In the gas phase, apart from the already mentioned transition at 532 eV, a second resonant transition occurs at 535.5 eV, followed by a

4. CONCLUSIONS We have studied acetic acid adsorption on ice under environmental conditions using a combination of PE and NEXAFS spectroscopies. C 1s XPS data indicate the presence of two species, assigned to protonated and deprotonated acetic acid, that allow an estimation of the degree of protonation of acetic acid at the ice surface, about 60%, which is higher than that expected for an aqueous solution in equilibrium with the same partial pressure of CH3COOH. For both the aqueous solution and the ice surface, the partial-electron-yield C-edge NEXAFS spectra are insensitive to the degree of protonation and similar to the corresponding gas-phase spectrum. Depth profiling information obtained from the kinetic-energy-dependG



Article

(16) Compoint, M.; Toubin, C.; Picaud, S.; Hoang, P. N. M.; Girardet, C. Chem. Phys. Lett. 2002, 365, 1−7. (17) Gadermann, M.; Vollmar, D.; Signorell, R. Phys. Chem. Chem. Phys. 2007, 9, 4535−4544. (18) Gao, Q.; Leung, K. T. J. Phys. Chem. B 2005, 109, 13263− 13271. (19) Bahr, S.; Borodin, A.; Hofft, O.; Kempter, V.; Allouche, A.; Borget, F.; Chiavassa, T. J. Phys. Chem. B 2006, 110, 8649−8656. (20) McNeill, V. F.; Geiger, F. M.; Loerting, T.; Trout, B. L.; Molina, L. T.; Molina, M. J. J. Phys. Chem. A 2007, 111, 6274−6284. (21) McNeill, V. F.; Loerting, T.; Geiger, F. M.; Trout, B. L.; Molina, M. J. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 9422−9427. (22) Huse, N.; Wen, H. D.; Nordlund, D.; Szilagyi, E.; Daranciang, D.; Miller, T. A.; Nilsson, A.; Schoenlein, R. W.; Lindenberg, A. M. Phys. Chem. Chem. Phys. 2009, 11, 3951−3957. (23) Nordlund, D.; Odelius, M.; Bluhm, H.; Ogasawara, H.; Pettersson, L. G. M.; Nilsson, A. Chem. Phys. Lett. 2008, 460, 86−92. (24) Bluhm, H.; Ogletree, D. F.; Fadley, C. S.; Hussain, Z.; Salmeron, N. J. Phys.: Condens. Matter 2002, 14, L227−L233. (25) Starr, D. E.; Pan, D.; Newberg, J. T.; Ammann, M.; Wang, E. G.; Michaelides, A.; Bluhm, H. Phys. Chem. Chem. Phys. 2011, 13, 19988− 19996. (26) Lee, H.-K.; Kim, K.-j.; Han, J.-h.; Kang, T.-H.; Chung, J. W.; Kim, B. Phys. Rev. B 2008, 77, art. no.115324. (27) Lewis, T.; Winter, B.; Stern, A. C.; Baer, M. D.; Mundy, C. J.; Tobias, D. J.; Hemminger, J. C. J. Phys. Chem. B 2011, 115, 9445− 9451. (28) Lewis, T.; Winter, B.; Stern, A. C.; Baer, M. D.; Mundy, C. J.; Tobias, D. J.; Hemminger, J. C. J. Phys. Chem. C 2012, 115, 21183− 21190. (29) Ottosson, N.; Wernersson, E.; Soderstrom, J.; Pokapanich, W.; Kaufmann, S.; Svensson, S.; Persson, I.; Ohrwall, G.; Bjorneholm, O. Phys. Chem. Chem. Phys. 2011, 13, 12261−12267. (30) Parent, P.; Lasne, J.; Marcotte, G.; Laffon, C. Phys. Chem. Chem. Phys. 2011, 13, 7142−7148. (31) Brown, M. A.; Vila, F.; Sterrer, M.; Thürmer, S.; Winter, B.; Ammann, M.; Rehr, J. J.; van Bokhoven, J. A. J. Phys. Chem. Lett. 2012, 3, 1754−1759. (32) Krepelova, A.; Newberg, J.; Huthwelker, T.; Bluhm, H.; Ammann, M. Phys. Chem. Chem. Phys. 2010, 12, 8870−8880. (33) Ogletree, D. F.; Bluhm, H.; Hebenstreit, E. D.; Salmeron, M. Nucl. Instrum. Methods A 2009, 601, 151−160. (34) Bluhm, H. J. Electron Spectrosc. Relat. Phenom. 2010, 177, 71−84. (35) Murphy, D. M.; Koop, T. Q. J. R. Meteorol. Soc. 2005, 131, 1539−1565. (36) Brown, M. A.; Seidel, R.; Thurmer, S.; Faubel, M.; Hemminger, J. C.; van Bokhoven, J. A.; Winter, B.; Sterrer, M. Phys. Chem. Chem. Phys. 2011, 13, 12720−12723. (37) Brown, M. A.; Huthwelker, T.; Beloqui Redondo, A.; Janousch, M.; Faubel, M.; Arrell, C. A.; Scarongella, M.; Chergui, M.; van Bokhoven, J. A. J. Phys. Chem. Lett. 2012, 3, 231−235. (38) Myneni, S.; Luo, Y.; Naslund, L. A.; Cavalleri, M.; Ojamae, L.; Ogasawara, H.; Pelmenschikov, A.; Wernet, P.; Vaterlein, P.; Heske, C.; Hussain, Z.; Pettersson, L. G. M.; Nilsson, A. J. Phys.: Condens. Matter 2002, 14, L213−L219. (39) Ketteler, G.; Ashby, P.; Mun, B. S.; Ratera, I.; Bluhm, H.; Kasemo, B.; Salmeron, M. J. Phys.: Condens. Matter 2008, 20, art. no. 184024. (40) Lee, H. K.; Kim, K. J.; Han, J. H.; Kang, T. H.; Chung, J. W.; Kim, B. Phys. Rev. B 2008, 77, art. no.115324. (41) Filler, M. A.; Van Deventer, J. A.; Keung, A. J.; Bent, S. F. J. Am. Chem. Soc. 2006, 128, 770−779. (42) Haley, R. D.; Tikhov, M. S.; Lambert, R. M. Catal. Lett. 2001, 76, 125−130. (43) Smith, S. R.; Thomas, T. D. J. Am. Chem. Soc. 1978, 100, 5459− 5466. (44) Hopkins, R. J.; Tivanski, A. V.; Marten, B. D.; Gilles, M. K. J. Aerosol Sci. 2007, 38, 573−591.

ent XPS data indicates that acetic acid resides within the topmost bilayers of the ice surface. Within the experimental uncertainties, the surface coverages observed are consistent with those of flow-tube experiments performed under comparable conditions. The electron-yield O K-edge NEXAFS spectrum from the ice surface in the presence of acetic acid indicates only minor perturbations of the hydrogen-bonding environment of water at the ice surface.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The Advanced Light Source and beamline 11.0.2 are supported by the Director, Office of Energy Research, Office of Basic Energy Sciences, and Chemical Sciences Division of the U.S. Department of Energy under Contract DE-AC02-05CH11231. This work was also supported by the Office of Science, Biological and Environmental Research, Environmental Remediation Sciences Division (ERSD), U.S. Department of Energy, under Contract DE-AC02-05CH11231. M.A. acknowledges support by the Swiss National Science Foundation (Project 140540). M.A.B. thanks Martin Sterrer and Bernd Winter for their continued support and for their help during the BESSY experiments.

■

REFERENCES

(1) Huthwelker, T.; Ammann, M.; Peter, T. Chem. Rev. 2006, 106, 1375−1444. (2) Chebbi, A.; Carlier, P. Atmos. Environ. 1996, 30, 4233−4249. (3) Huang, Y.-w.; Dransfield, T. J.; Anderson, J. G. J. Phys. Chem. A 2010, 114, 11538−11544. (4) Khamaganov, V. G.; Bui, V. X.; Carl, S. A.; Peeters, J. J. Phys. Chem. A 2006, 110, 12852−12859. (5) Domine, F.; Shepson, P. B. Science 2002, 297, 1506−1510. (6) Grannas, A. M.; Jones, A. E.; Dibb, J.; Ammann, M.; Anastasio, C.; Beine, H. J.; Bergin, M.; Bottenheim, J.; Boxe, C. S.; Carver, G.; Chen, G.; Crawford, J. H.; Dominé, F.; Frey, M. M.; Guzmán, M. I.; Heard, D. E.; Helmig, D.; Hoffmann, M. R.; Honrath, R. E.; Huey, L. G.; Hutterli, M.; Jacobi, H. W.; Klán, P.; Lefer, B.; McConnell, J.; Plane, J.; Sander, R.; Savarino, J.; Shepson, P. B.; Simpson, W. R.; Sodeau, J. R.; von Glasow, R.; Weller, R.; Wolff, E. W.; Zhu, T. Atmos. Chem. Phys. 2007, 7, 4329−4373. (7) Legrand, M.; Preunkert, S.; Wagenbach, D.; Cachier, H.; Puxbaum, H. J. Geophys. Res. Atmos. 2003, 108, art. no.4788. (8) McNeill, V. F.; Grannas, A. M.; Abbatt, J. P. D.; Ammann, M.; Ariya, P.; Bartels-Rausch, T.; Domine, F.; Donaldson, D. J.; Guzman, M. I.; Heger, D.; Kahan, T. F.; Klán, P.; Masclin, S.; Toubin, C.; Voisin, D. Atmos. Chem. Phys. 2012, 12, 9653−9678. (9) Kerbrat, M.; Huthwelker, T.; Bartels-Rausch, T.; Gaggeler, H. W.; Ammann, M. Phys. Chem. Chem. Phys. 2010, 12, 7194−7202. (10) Picaud, S.; Hoang, P. N. M.; Peybernes, N.; Le Calve, S.; Mirabel, P. J. Chem. Phys. 2005, 122, art. no. 194707. (11) Sokolov, O.; Abbatt, J. P. D. J. Phys. Chem. A 2002, 106, 775− 782. (12) von Hessberg, P.; Pouvesle, N.; Winkler, A. K.; Schuster, G.; Crowley, J. N. Phys. Chem. Chem. Phys. 2008, 10, 2345−2355. (13) Symington, A.; Cox, R. A.; Fernandez, M. A. Z. Phys. Chem. 2010, 224, 1219−1245. (14) Crowley, J. N.; Ammann, M.; Cox, R. A.; Hynes, R. G.; Jenkin, M. E.; Mellouki, A.; Rossi, M. J.; Troe, J.; Wallington, T. J. Atmos. Chem. Phys. 2010, 10, 9059−9223. (15) Allouche, A.; Bahr, S. J. Phys. Chem. B 2006, 110, 8640−8648. H



Article

(45) Robin, M. B.; Ishii, I.; McLaren, R.; Hitchcock, A. P. J. Electron Spectrosc. Relat. Phenom. 1988, 47, 53−92. (46) Goldberg, R. N.; Kishore, N.; Lennen, R. M. J. Phys. Chem. Ref. Data 2002, 31, 231−370. (47) Johnson, B. J.; Betterton, E. A.; Craig, D. J. Atmos. Chem. 1996, 24, 113−119. (48) NIST Chemistry WebBook; NIST Standard Reference Database 69; Linstrom, P. J., Mallard, W. G., Eds.; National Institute of Standards and Technology: Gaithersburg, MD, 1995; available at http://webbook.nist.gov (accessed January 2013). (49) Kuo, M. H.; Moussa, S. G.; McNeill, V. F. Atmos. Chem. Phys. 2011, 11, 9971−9982. (50) Chameides, W. L. J. Geophys. Res. 1984, 89, 4739−4755. (51) Staudinger, J.; Roberts, P. V. Crit. Rev. Environ. Sci. Technol. 1996, 26, 205−297. (52) Cabani, S.; Gianni, P.; Mollica, V.; Lepori, L. J. Solution Chem. 1981, 10, 563−595. (53) Wettlaufer, J. S. Phys. Rev. Lett. 1999, 82, 2516−2519.

I


Adsorption of Acetic Acid on Ice Studied by Ambient-Pressure XPS

Recommend Documents