Spectroscopic Study of HNO3 Dissociation on Ice - ACS Publications

Nov 6, 2012 - Département de Chimie, Université de Sherbrooke, 2500, boulevard université, .... character of HNO3 ionic dissociation at the surface...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/JPCA

Spectroscopic Study of HNO3 Dissociation on Ice Patrick Marchand, Guillaume Marcotte, and Patrick Ayotte* Département de Chimie, Université de Sherbrooke, 2500, boulevard université, Sherbrooke, Québec J1K 2R1, Canada ABSTRACT: A detailed spectroscopic study of HNO3:H2O binary amorphous mixtures, and of the adsorption of HNO3 onto ice, is reported. Using a classical optics model, the extent of intermixing and of ionic dissociation of adsorbed HNO3, which forms a strong acid with liquid water, is determined as a function of HNO3 coverage and temperature. Even at temperatures as low as 45 K, where intermixing is limited to at most a few molecular layers at the interface, ionic dissociation of adsorbed HNO3 is observed to be extensive. While some amount of molecularly adsorbed HNO3 is observed at the surface of ice at 45 K, its ionic dissociation occurs irreversibly upon heating the ice substrate to 120 K. The molecularly adsorbed state of HNO3 is not restored upon cooling, suggesting HNO3 is a metastable entity at the surface of ice. Therefore, despite ionic dissociation of HNO3 being thermodynamically favored, it appears to be kinetically inhibited at the surface of amorphous solid water at temperatures below 120 K. simple thermodynamics models.13,28 Aqueous protons were reported to display a slight preference for the interface,12,24,25 and nitrates a modest propensity for the bulk,26,27 conditions that are not expected to result in a strong interfacial specificity to the acid−base behavior of nitric acid solutions. Interestingly, observations of molecular HNO3 at the surface of concentrated nitric acid solutions and thin films were reported using sumfrequency generation (SFG)7,8 and infrared29,30 spectroscopies, prompting the authors to propose an unexpected weak acid behavior for nitric acid solutions interfaces. More recently, however, synchrotron-based photoelectron spectroscopic (PES) studies allowed the simultaneous observation of both HNO3 and NO3− at the surface of nitric acid solutions (i.e., in a liquid jet), enabling the degree of dissociation of HNO3 to be determined.31,32 Notwithstanding the fact that caution must be exerted in the comparison of results emerging from these studies due to the different probing depths of SFG and PES techniques, the photoemission investigations showed that HNO3 is extensively dissociated at the surface of liquid water.31,32 It also revealed that nitric acid solutions interfaces indeed only appear to be slightly less acidic than the bulk and that, in agreement with bulk thermodynamics, the ionic dissociation of HNO3 is promoted by decreasing temperatures at the surface of nitric acid solutions.31,32 Given the fact that the pKa of dilute nitric acid solutions is rather small and negative, one is not surprised to find undissociated HNO3 at the surface of liquid water. In fact, beyond the mere thermodynamic aspects of this elementary reaction, the dynamics of proton transfer at aqueous interfaces might present significant kinetics barriers distinct from those

I. INTRODUCTION An increasing interest in the crucial role of heterogeneous reactions at aqueous interfaces in several atmospheric chemistry processes1−6 has uncovered our poor understanding of the interfacial specificity displayed by acid−base equilibria. Observations of strong acids remaining in their molecular form at aqueous solutions7,8 and ice nanoparticles9 surfaces, or weak acids releasing their proton when adsorbed onto amorphous ice10 at cryogenic temperatures, and even propositions that the surface of liquid water may be “acidic”11 display vivid examples of the urgent need to re-examine the physical basis underpinning these counterintuitive behaviors and to craft a new rulebook for interfacial acid−base chemistry. Such baffling of our physical intuition highlights the often dramatic, and certainly nontrivial thermodynamic preference of certain chemical species for aqueous interfaces, certain being based on enthalpic/energetic considerations, while the contribution of a more subtle entropic character to segregation at surfaces is being increasingly recognized.12,13 Of the various model systems used to garner a molecularlevel understanding of these elementary processes, the adsorption, dissolution, and uptake of HNO3 on liquid water and ice display a particularly interesting and important case.14−22 This most intriguing compound forms strongly acidic aqueous solutions at standard temperature and pressure (i.e., pKa ≈ −1.5). Like many strong acids, the ionic dissociation of HNO3 in water displays a strong exothermicity while being also rather strongly entropically inhibited.23 This causes nitric acid solutions to become even more strongly acidic at lower temperatures23 and into their supercooled state.19 The interfacial segregation of the ionic products of HNO3 acid dissolution in water, the aqueous protons (i.e., H3O+(aq))12,24,25 and nitrate anions (i.e., NO3−(aq)),26,27 has been investigated, and an interpretation of this behavior has been proposed using © 2012 American Chemical Society

Received: September 25, 2012 Revised: November 5, 2012 Published: November 6, 2012 12112

dx.doi.org/10.1021/jp309533f | J. Phys. Chem. A 2012, 116, 12112−12122

The Journal of Physical Chemistry A

Article

the ensuing local heat dissipation dynamics, should be included in a more realistic description of HNO3 adsorption onto ice from the gas phase. Tolbert and Middlebrook38 were first to report spectroscopic observations of the interaction of HNO3 vapors with ice films. Under the high temperature and HNO3 partial pressures used in their investigations (i.e., 183 K and 1.8 × 10−6 Torr),38 significant diffusive uptake of HNO3 by ice was noticed. The resulting mixed HNO3:H2O interfacial layer crystallized rapidly to a ∼45 nm thick superficial layer of nitric acid trihydrate (while evidence for formation of the nitric acid dihydrate38,39 and monohydrate38 was reported to occur at greater exposures). Unfortunately, the chemical state of HNO3 adsorbed onto the surface of ice was not investigated in these studies. Devlin and co-workers later reported observations of a small fraction of molecularly adsorbed HNO3 on ice nanoparticles in the 70−135 K range using FTIR spectroscopy.40 However, as these samples were prepared either by solute exclusion from crystallizing (≤5 mole percent HNO 3 concentration) supercooled nitric acid solutions, or by the sublimation of HNO3 nanoparticles within three-dimensional ice nanoparticle arrays at 135 K, they do not address the chemical state of sub monolayer HNO3 adsorbed from the gas phase onto the ice surface at cryogenic temperature directly.40 Furthermore, the HNO3 coverage at the surface of the ice nanoparticle arrays is rather poorly controlled and characterized using this methodology. A subsequent and related FTIRkinetics study of HNO3 adsorption on ice films at 145 K by Purcell and co-workers41 reported the observation of much more abundant molecular HNO3 at the surface of an ice film; however, this study appears to have been conducted under conditions where multilayer HNO3 had condensed onto ice. Therefore, the kinetics reported therein should rather pertain to the convoluted processes of diffusive intermixing, uptake, dissolution, and absorption of solid HNO3 overlayers within the bulk of the ice film, rather than those of a putative ionic dissociation reaction of isolated molecularly adsorbed HNO3 at the surface of ice. Finally, a very recent PES spectroscopic study (XPS/NEXAFS) concluded that HNO3, although formed indirectly through in situ NO2 hydrolysis at the surface of ice, appears to be extensively dissociated on ice at 230 K.42 In summary, the important differences in the conditions used in these various investigations unfortunately preclude any definitive conclusion regarding the chemical state of HNO3 adsorbed onto ice, and of its temperature and coverage dependencies, to be drawn. At the outset, it thus appears that microsolvation considerations should enhance the interfacial stability of HNO3 (with respect to ionic dissociation and relative to the bulk), while thermodynamic arguments would exacerbate the tendency for ionic dissociation at the surface of condensed water at low temperature (where its entropic inhibition is increasingly suppressed). It is therefore intriguing to ponder: Which of these counteracting forces should prevail under the conditions relevant to the natural environment, for the H2O(ads) + HNO3(ads) ↔ H3O+(ads) + NO3−(ads) equilibrium at the air/ quasi-liquid layer interface on ice? On this problem of HNO3 adsorption, dissolution and uptake in ice hinges a broad array of atmospherically important processes such as the formation and lifetime of cirrus clouds,21,22 the denitrification of the stratosphere,43−45 the renoxification of the troposphere,30,46,47 as well as through nitrates photolysis being a recognized source of the intense photochemical NOx fluxes observed to emanate

found in the bulk and that remain challenging to investigate in such complex systems. Some of these dynamical effects were captured through first-principle molecular dynamics studies, which identified several microscopic characters of the local Hbond topology that may inhibit ionic dissociation of HNO3.31−34 The microsolvation structures and dynamics were found to display distinctively different attributes for surface adsorption of HNO3 onto condensed water versus bulk solvation environments: (i) decreasing [H2O·HONO2](aq) special pair proton donor−acceptor H-bonded O−O distance; (ii) decreasing coordination number of the acidic proton accepting solvating water molecule; (iii) increasing solvation of the proton accepting water molecule; and (iv) increasing hydration of the incipient nitrate group.34 These features all display rather distinct behaviors in the confinement of condensed water interfaces, and their overall effect was reported to enhance the stabilization of molecularly adsorbed HNO3 over the products of its ionic dissociation as compared to the bulk hinting at a weaker acidity at the surface.31−34 Particularly intriguing, however, was the observation that solvation of HNO3 in the bulk of liquid water displays a local H-bond topology reminiscent of the hydration of hydrophobic solutes.31,32 Therefore, in addition to the reported enhanced propensity for H-bonding to the NO3 moiety of HNO3 at the surface of liquid water (as compared to that observed in the bulk),31−34 a significant entropic contribution may also exist, which could further enhance the stability of molecularly adsorbed HNO3 at the air−water interface relative to the products of its ionic dissociation.31,32 It remains unclear, however, whether the magnitude of these effects is sufficient to provide a rationale for the proposed weaker acidity of nitric acid solutions interfaces7,8,29,30,33 and for a putative enhanced stability of molecularly adsorbed HNO3 relative to the products of its ionic dissociation at the surface of ice.35 Indeed, this controversy surrounding the “interfacial acidity” of HNO3 aqueous solutions was echoed by an intense effort to identify the microsolvation aspects that may convey a similar “weak acid” character HNO3 adsorbed to the surface of ice. Quantum chemical studies, based on a QM/MM methodology,35−37 identified local H-bonding architectures at model ice surfaces, that were either thermodynamically unfavorable to ionic dissociation, or that presented significant barriers inhibiting this elementary reaction. Interestingly, the QM/ MM investigations35,37 concluded to a very clearly endothermic character for the ionic dissociation of molecularly adsorbed HNO3 onto ice surface models. Decisively positive reaction free energies were reported, even down to the very lowest temperatures where the entropic inhibition is mostly suppressed using, however, partition functions for the zero Kelvin structures.35 However, the reliability of this QM/MM approach, upon which conclusions of the endothermic character of HNO3 ionic dissociation at the surface of ice were drawn,35 was later questioned by these same authors.34 Indeed, they identified a combination of potential weaknesses in their original investigation: the known propensity of the theoretical methodology to enhance the stability of strongly Hbonded molecularly adsorbed HNO3, deficiencies in the description of electronic correlation effects in the QM methodology, and/or inadequate selection of structural isomers as models to describe the adsorption of HNO3 onto the ice surface. Furthermore, one might argue that a proper account of the enthalpy of condensation that is released upon adsorption of such a strong H-bond donor, and a complete description of 12113

dx.doi.org/10.1021/jp309533f | J. Phys. Chem. A 2012, 116, 12112−12122

The Journal of Physical Chemistry A

Article

from the sunlit snowpack and into the polar boundary layer.48 In this Article, we report a spectroscopic study of HNO3 adsorption onto amorphous solid water (ASW) at cryogenic temperatures. Our objective is to identify the conditions under which HNO3 ionically dissociates when adsorbed onto ASW, our model substance for the quasi-liquid layer that exists at the surface of ice above about 200 K, and up to its melting point.

adjusting the relative condensation rates of H2O and HNO3 vapors from the two independent sources. In this work, we used infrared interferometry and molecular beam dosing to devise a methodology enabling absolute coverage and composition to be determined. Amorphous solid water (ASW) films coverages (i.e., H2O molecular beam flux) were first calibrated to within 10% using TDMS of precisely controlled amounts of H2O deposited at 80 K and 45° incidence with the molecular beam doser. This allowed a convenient determination of the exposure required to produce the saturation coverage on Pt(111), hereby defined as θH2O = 1 ML.50,51 A 50 ML coverage dense ASW film (n∞ = 1.285 at λ = 6328 Å yielding a density of ρ = 0.875 g/cm3 and a film thickness h = 50 ML·3.67 Å/ML = 183.5 Å as confirmed by infrared interferometry)52,53 was selected as a compromise between improving signal-to-noise ratios, afforded by thicker films, and their increasing deposition times. An HNO3 film having the same optical thickness, β = 2πn∞h cos θt/λ, where θt is the incidence angle of the infrared beam transmitted/ refracted within the sample as calculated from θIR and n∞ using Snell’s law, was then prepared using the same molecular beam source following an infrared interferometry prescription described previously.54 Briefly, a neat amorphous HNO3 film was grown to a coverage where its absorbance at ω > 3700 cm−1 coincided with that of the 50 ML coverage ASW film (i.e., in the spectral range where both substances are transparent and where their RAIRS spectra are thus dominated by optical interference effects). Using the optical index of refraction for amorphous HNO3 films reported in the literature (i.e., n∞ = 1.47 at λ = 6328 Å yielding a density of ρ = 1.75 g/cm3),55 we could obtain the physical thickness of a neat amorphous HNO3 film (h = 138 Å) prepared to have the same nominal coverage (i.e., optical thickness, β) as that of a 50 ML coverage dense ASW film. Binary films having a nominal coverage of 50 ML were then created by dosing HNO3 to a certain predetermined coverage (e.g., θHNO3 = 25 ML nominal HNO3 coverage) with the molecular beam source while simultaneously adjusting the H2O condensation rate with the leak valve (i.e., backfilling method) to achieve the requisite complementary coverage (e.g., θH2O = 25 ML) yielding a predefined nominal composition [i.e., defined as monolayer fractions: θHNO3/(θH2O + θHNO3) = 0.5]. Eleven samples were thereby grown to explore the full nominal composition range from neat HNO3 to neat H2O in 0.1 increments. However, because nonporous amorphous solid HNO3 is significantly more optically dense than nonporous amorphous solid H2O, a sample composed of θHNO3 = 25 ML and θH2O = 25 ML nominal coverages displays a HNO3 mole fraction significantly smaller than 50%. The absolute mole fractions of binary films were therefore determined a posteriori from the absolute H2O and HNO3 coverages measured optically through infrared interferometry. This was achieved by using the Lorentz−Lorenz relationship55 to provide areal densities from the neat H2O and HNO3 samples densities and absolute thickness. A 50 ML H2O film coverage was thus determined to correspond to 5.35 × 1016 H2O molecules/cm2, using an absolute coverage of 1.07 × 1015 H2O molecules/cm2 per ML for the √37 × √37R25.3° monolayer structure on Pt(111).51 Using the density of amorphous solid HNO3 that of a neat 50 ML HNO3 nominal (i.e., optical) coverage film was determined to correspond to 2.31 × 1016 HNO3 molecules/cm2. While the backfilling method straightforwardly allowed control over the H2O

II. EXPERIMENTAL SECTION Stratified and homogeneous binary composite nanoscopic films were grown using molecular beam deposition methods and studied spectroscopically in an apparatus described in detail previously.49 Briefly, it consisted of an ultra high vacuum (UHV) chamber (turbomolecularly pumped to a base pressure 3700 cm−1).53,54,56 Indeed, the RAIRS spectra of thin films measured at grazing incidence display a smoothly decreasing baseline where the absorbance becomes increasingly negative with increasing wavenumbers.53,54,56 Negative absorbance arises from the increase in reflectivity of the film/substrate system as compared to that of the bare Pt(111) substrate and results from trivial optical interference effects.53,54,56 Spectra are color coded to their absolute mole percent as calculated a posteriori from their absolute H2O and HNO3 areal densities (see Experimental Section for details). In Figure 1, the spectra are grouped in three distinct families according to similarities in their spectral features. In the waterrich samples (i.e., 0−15 HNO3 mole percent), the amplitude of the Zundel continuum, as well as that of the spectral features arising from dissolved nitrate anions NO stretching vibrational modes (labeled by NO3− on the top series of spectra), all increase continuously with HNO3 concentration. This is accompanied by a decreasing intensity of the spectral features due to the intramolecular OH stretching modes of water molecules (centered near 3400 cm−1) and of the intermolecular H-bond deformation modes of ASW (centered near 900 cm−1). The sharp doublet centered near 2400 cm−1 belongs to ubiquitous CO2(g) contamination in the purge system external to the UHV analysis chamber. In the intermediate compositional range (i.e., 15−40 HNO3 mole percent), the intensity of the Zundel continuum and of the nitrate anions spectral features does not increase significantly with HNO3 concentration. However, spectral features that can be attributed to molecular HNO3 (labeled by HNO3 on the bottom series of spectra) begin to develop, and their intensity is observed to increase continuously with HNO3 concentration. Finally, in the acid-rich mixtures (i.e., 40−100 HNO3 mole percent), the amplitude of the spectral features attributed to molecular HNO3 continues to increase, while that of the Zundel continuum and of the spectral features due to nitrate anions decreases continuously with increasing HNO3 concentration. One notices that the spectrum of the “neat” HNO3 film (maroon trace, 100 HNO3 mole percent) displays a small band centered near 2700 cm−1, which is attributed to the expected small amount of water present (i.e.,