Hydrolysis of Sodium Atoms on Water−Ice Films ... - ACS Publications

Dec 5, 2008 - Hydrolysis of Sodium Atoms on Water-Ice Films. Characterization of Reaction Products and Interfacial Distribution of Sodium and Hydroxid...
0 downloads 0 Views 128KB Size
Download PDF
J. Phys. Chem. C 2009, 113, 321–327

321

Hydrolysis of Sodium Atoms on Water-Ice Films. Characterization of Reaction Products and Interfacial Distribution of Sodium and Hydroxide Ions Jung-Hwan Kim, Young-Kwang Kim, and Heon Kang* Department of Chemistry, Seoul National UniVersity, Gwanak-gu, Seoul 151-747, Republic of Korea ReceiVed: September 2, 2008; ReVised Manuscript ReceiVed: NoVember 1, 2008

We have examined the hydrolysis reaction of Na atoms at the surface of thin water-ice films by using the techniques of Cs+ reactive ion scattering and low energy sputtering. Na atoms are adsorbed onto amorphous D2O-ice films, and the reaction products formed at the ice film surfaces are examined for ionic (both positive and negative) species as well as neutral molecules for the temperature range of 95-135 K and Na coverage below the multilayer regime. The hydrolysis reaction produces isolated sodium and hydroxide ions as well as a sodium hydroxide molecule. Efficient solvation of these products by water molecules at high temperature alters their distribution near the ice film surface. Hydroxide ions tend to reside at the film surface, whereas sodium ions migrate to the film interior. NaOD molecules are efficiently solvated by water molecules at high temperature, but a substantial portion of the molecules remains undissociated. 1. Introduction The reaction of sodium metal with liquid water is a wellknown exothermic reaction producing solvated sodium and hydroxide ions and molecular hydrogen, but its reaction mechanism at the molecular level is still not fully understood. The reaction has been investigated not only in liquid water but also for Na atoms interacting with water molecules in various environments such as water-ice films1-9 and gas-phase water clusters,10-13 and by theoretical computation methods as well.14-19 Several pioneering studies were performed for the interactions of alkali metals with water adsorbates or water-ice films grown on cold substrate surfaces.1-9 Thiel et al.1 examined the interaction of K and H2O coadsorbed onto Ru(0001) using temperature-programmed desorption (TPD) and electron energy loss spectroscopy, and observed that KOH is formed on the surface at a temperature of 80 K and exists up to 580 K. Blass et al.2 observed that coadsorbed potassium and water react very slowly at 100 K, and the reaction is complete at 300 K, in their study using TPD and X-ray photoelectron spectroscopy (XPS). The interactions of alkali metals with water-ice films were investigated by Gu¨nster et al.3-8 using metastable impact electron spectroscopy (MIES), ultraviolet photoelectron spectroscopy (UPS), and XPS. MIES spectra suggest a reaction sequence where a Na atom donates a 3s electron to the ice film and then the solvated electron reacts with water molecule to form hydroxyl species. Upon the increase in alkali metal exposure, neutral Na clusters start to form at the film surface.3-8 Studies of Na adsorption on ice films with time-of-flight secondary ion mass spectrometry (TOF-SIMS)9 show that an unhydrated NaOH layer is formed at the film surface, followed by the growth of a metallic Na layer at high Na exposure. The domains of NaOH and metallic Na are created when the film temperature is low (13 K) or when the Na coverage reaches a multilayer regime at high temperature (100 K).9 The interactions of Na atoms with water molecules were also studied in gas phase by crossed-molecular beam studies,10-13 and their results and theoretical studies14-16 indicate that the * To whom correspondence should be addressed. Tel.: +82 2 875 7471. Fax: +82 2 889 8156. E-mail: [email protected].

Na hydrolysis reaction is initiated by the presence of a Na2 dimer in the water cluster leading to the formation of sodium hydroxide. Recent density functional theory calculations17,18 and ab initio molecular dynamics simulations19 suggest that the formation of sodium hydroxide can occur not only with Na2 dimers but also with single Na atoms. According to the investigated mechanism of Na hydrolysis, Na+ and OH- ions can be produced either as a contact pair or as separated species depending on the delocalization of the 3s electron that is released from the Na atom and solvated in the water cluster.17-19 In the present study of Na hydrolysis on water-ice films, we examine several issues that have been left unresolved in previous studies of this system.3-9 We determine the nature of the hydrolysis products by inspecting the chemical species present at the ice film surfaces with low energy sputtering (LES) and reactive ion scattering (RIS) techniques under various conditions for the temperature range of 95-135 K and Na coverage below the multilayer regime. Further, these products are examined for their formation path and interfacial behavior to improve our understanding of the hydrolysis process at ice surfaces. 2. Experimental Section The experiment was carried out in an ultrahigh vacuum surface analysis chamber equipped with instrumentations for LES, RIS, TPD, and Auger spectroscopy.20-22 Ice films were prepared on the (0001) face of a Ru single crystal mounted on a temperature control stage of a sample manipulator. The substrate temperature was variable within the range of 90-1500 K and was monitored by a K-type thermocouple wire spotwelded to the crystal. Ice films were grown by back-filling the chamber with D2O vapor at a partial pressure of 1.0 × 10-8 Torr, as read by a nude ionization gauge without calibration. D2O liquid (99.96+% isotope purity) was degassed by freeze-vacuum-thaw cycles. The thickness of the ice films was typically four bilayers (BLs, 1 BL ) 1.1 × 1015 water molecules cm-2) as deduced from TPD experiments.22,23 D2O deposition onto a Ru(0001) surface at 130 K forms a nonporous, amorphous solid water (ASW) film, and this film was trans-

10.1021/jp807774v CCC: $40.75  2009 American Chemical Society Published on Web 12/05/2008

322 J. Phys. Chem. C, Vol. 113, No. 1, 2009 formed to a crystalline phase upon warming to 140-145 K. Both ASW and crystalline ice films will be called “ice films” in the text. Na atoms were deposited onto the ice film by employing a commercial alkali-metal dispenser (SAES Getters, Inc.). The surface coverage of Na atoms and the deposition rate were determined by recording TPD spectra of Na on a bare Ru(0001) surface as a function of deposition time. The TPD features of Na on Ru(0001) were in good agreement with those reported in the literature.24 The Na coverage on ice films is given in units of monolayer equivalents (MLE), where 1 MLE corresponds to the full monolayer coverage of Na on a bare Ru(0001) surface (1 MLE ) 0.53 ML of surface Ru atoms). The Na deposition rate was about 0.02 MLE min-1. The amount of Na exposure was controlled by a shutter installed between the evaporator and the substrate. Neutral and ionic species present at the ice film surfaces were analyzed by the techniques of Cs+ RIS and LES, respectively. In these experiments, a Cs+ beam from a low-energy ion gun (Kimball Physics) was scattered at a sample surface with the incident energy chosen to be between 20 and 65 eV. Both positive and negative ions emitted from the surface were detected by a quadrupole mass spectrometer (ABB Extrel) with its ionizer filament switched off. The beam incidence and detector angles were fixed at 65 and 55°, respectively, from the surface normal. The detected ions were composed of reflected Cs+ primaries, RIS products that were association products of Cs+ with neutral molecules at the surface, and LES ions resulting from collisional desorption of pre-existing ions on the surface. The mechanisms of RIS and LES processes on thin ice films have previously been explained.21,25,26 At an incident Cs+ energy below 35 eV, LES and RIS methods had a probing depth of 1 BL of ice surface.27 The surface contamination or damage by a Cs+ beam was negligible for the typical experimental conditions of a Cs+ current density below 3 nA cm-2 at the sample surface and a spectral acquisition time of 10 s. A Faraday cup was used to measure the actual number of the incident ions, because only a small fraction of Cs+ ions was neutralized on a flat sample surface.26b The incident Cs+ flux thus measured was 2 × 1011 ions cm-2 s-1 at a Faraday cup, which would accumulate a Cs coverage of 0.2% of the water molecular density on ice surface during 10 s. Whenever necessary, fresh samples were prepared for the measurements to reduce the accumulated Cs+ beam dose.

Kim et al.

Figure 1. (a) LES spectrum of positive ions emitted from a D2O-ice film on which Na atoms were deposited at 95 K to coverage of 0.13 MLE. (b) Negative ion LES spectrum from the same film as in (a). The ice films were 4 BL thick and prepared in an amorphous structure by condensing D2O vapor on Ru(0001) at 130 K. The ice film temperature was reduced to 95 K for Na adsorption and LES measurement. The time interval between the LES measurement and the Na adsorption was 3 min. Cs+ beam energy was 35 eV.

3. Results The products of Na hydrolysis formed on ice film surfaces were analyzed by LES and RIS, as shown in Figures 1 and 2. Figure 1 presents LES mass spectra of the ionic species emitted from the Na-adsorbed D2O-ice film. The D2O-ice film was grown at 130 K to a thickness of 4 BL, and Na atoms were deposited onto the film surface at 95 K for coverage of 0.13 MLE. In the positive ion spectrum shown in Figure 1a, Na+ (m/z ) 23 amu/charge), NaD2O+ (m/z ) 43), NaD4O2+ (m/z ) 63), and NaD6O3+ (m/z ) 83) peaks represent the emission of Na+ and its hydrated species from the surface. The NaD2O+ intensity is higher than the Na+ intensity, indicating that Na+ ions are solvated by water molecules at the surface. The K+ and Rb+ peaks are due to the surface scattering of impurity ions contained in the Cs+ beam. The negative ion spectrum from the same surface (Figure 1 b) shows the emission of OD- (m/z ) 18 amu/charge) and its hydrated species (D3O2- and D5O3-) at lower intensities. These signals did not appear without Na adsorption, and therefore indicate Na hydrolysis products.

Figure 2. Cs+ RIS spectra obtained on D2O-ice films adsorbed with Na atoms for 0.13 MLE. Na deposition temperature was 95 K in (a) and 135 K in (b), with all other experimental conditions being the same as those given in the caption of Figure 1. The inset magnifies the mass region for CsNaOD+ signal.

Collisional ionization of D2O molecules to OD- by Cs+ impact requires a higher energy on a pure ice film (Figure S1 in the Supporting Information). The appearance of Na hydrolysis products at such low Na coverage indicates that the hydrolysis reaction can occur with single Na atoms at the ice surfaces. Figure 2 shows RIS spectra obtained from the same surface as that for Figure 1. The spectrum in Figure 2a shows a peak of reflected Cs+ primary at m/z ) 133 amu/charge and a series

Hydrolysis of Sodium Atoms on Water-Ice Films of RIS products at m/z ) 133 + 20n (n ) 1-3), which are the association products of Cs+ with surface water molecules [Cs(D2O)n+]. A CsNa+ signal (m/z ) 156 amu/charge) is not detected. This observation indicates that all Na atoms deposited on the surface at 95 K are ionized by the hydrolysis reaction, and neutral metal clusters are not formed. The CsNaOD+ signal at m/z ) 174 indicates the formation of the hydrolysis product NaOD. This signal has a very weak intensity (∼30 cps) as shown in the inset of Figure 2a at a magnified scale. The weak CsNaOD+ intensity suggests that the surface population of NaOD species is small or their RIS detection efficiency is low. RIS efficiency can be significantly reduced for an adsorbate with strong bonding to the surface; for example, chemisorbed adsorbates exhibit RIS intensities comparable to CsNaOD+.28 Figure 2b shows that, by raising the film temperature to 135 K, the CsNaOD+ signal disappears. We can summarize the information deduced from the spectral features of Na hydrolysis products in Figures 1 and 2 as follows. Single Na atoms react with water molecules at the ice film surfaces to produce Na+, OD-, and NaOD. Neutral Na atoms and Nan (n > 1) clusters are absent from the surface at these temperatures (95-135 K). Na+ ions are efficiently hydrated, as inferred from the larger LES signals of Na(D2O)n+ (n ) 1-2) compared to that of Na+. The presence of molecular NaOD is indicated by the RIS signal of CsNaOD+ observed at 95 K, though the surface population of NaOD is difficult to estimate from spectral intensity alone because the RIS detection efficiency for NaOD is unknown. There is ambiguity in that the Na+ and OD- signals can be produced from isolated Na+ and OD- ions as well as from other sources including the collisioninduced dissociation of NaOD molecules and the collisioninduced hydrolysis reaction promoted by solvated electrons. The strong emission of the hydrated species Na(D2O)n+ and OD(D2O)n- in LES suggests that Na+ and OD- ions are efficiently hydrated, and thus they probably exist as isolated ions. At this stage, however, the other possibilities also exist. To further characterize the hydrolysis products observed in LES and RIS spectra, we varied the experimental conditions of ice film preparation and surface analysis. Figure 3 shows the observed variations in signal intensities of a sodium ion, hydroxide ion, and CsNaOD+ measured as a function of ice film temperature for T ) 95-135 K. Here, the intensity of the “sodium ion” represents the summation of all the intensities of Na+ and its hydrated ion signals, and the same applies for the “hydroxide ion”. These notations will be used for the other results presented hereafter. A fresh film was prepared for each measurement at a different temperature to reduce possible sample contamination from Cs+ beams. In Figure 3a, the sodium ion intensity gradually decreases as the temperature increases, whereas the hydroxide ion intensity increases. We assume that sodium and hydroxide ions have equal populations in the top surface layer at 95 K, because, as a previous study shows,22 ions are usually immobile in ice films at this temperature (the same absolute intensity of LES signals for two species in this result is fortuitous). The reduced LES intensity of sodium ions at high temperature indicates that Na+ ions move from the ice surface toward the interior. On the other hand, the increased intensity of hydroxide ions indicates an increasing surface population of OD- ions at high temperature. Figure 3b shows the variation in CsNaOD+ intensity as a function of film temperature. There is a decrease in CsNaOD+ intensity at temperatures above 110 K, and the CsNaOD+ signal eventually disappears from the surface at 135 K. This indicates

J. Phys. Chem. C, Vol. 113, No. 1, 2009 323

Figure 3. (a) Temperature-dependent variation in LES intensities for sodium and hydroxide ions measured from Na-adsorbed ice films. The sodium ion intensity measured from an ice film adsorbed with NaF salt is also shown (explained in section 4B). (b) The variation in RIS intensity of CsNaOD+. The ice films were 4 BL thick and Na coverage was 0.13 MLE. LES and RIS analyses were made at the same temperature as Na adsorption, indicated on the abscissa, 3 min after Na adsorption. Cs+ beam energy was 35 eV.

Figure 4. LES intensities of sodium and hydroxide ions measured from Na-adsorbed ice films as a function of incident Cs+ energy. Solid symbols represent the signals from a sample at temperature 95 K and open symbols at 135 K. The ice films were 4 BL thick, and the adsorbed Na coverage was 0.13 MLE.

that NaOD species dissociate at the ice surface at 135 K or they are buried underneath the surface. We examined the sodium and hydroxide ion intensities as a function of incident Cs+ energy, and the result is shown in Figure 4 for two extreme temperatures, 95 and 135 K. If Na+ and OD- emission in LES comes from the collisional fragmentation of surface NaOD species located in random molecular orientations, then Na+ and OD- intensities will concurrently increase upon the increase of Cs+ impact energy. Figure 4 shows that sodium and hydroxide ion intensities do indeed increase with the impact energy. However, the increase in the hydroxide ion intensity is greater than that of the sodium ion. In addition, the difference between sodium and hydroxide ions is magnified at a higher temperature (135 K). These behaviors favor the other interpretation, that is, the sputtering of isolated Na+ and ODspecies which have different distributions along the vertical direction from the surface at high temperatures, as the result in Figure 3a shows. The steep increase of the hydroxide ion curve

324 J. Phys. Chem. C, Vol. 113, No. 1, 2009

Kim et al.

Figure 5. RIS signal intensity for NaOD as a function of incident Cs+ energy measured at 95 and 135 K. The curves show the summation of CsNaOD+ and Cs(NaOD)(D2O)+ intensities. All other experimental conditions were the same as those in Figure 4.

at 135 K may result from the increased sputtering yield of the surface-segregated OD- ions. On the other hand, the suppressed sodium ion intensity reflects the relatively inefficient sputtering of subsurface Na+ ions. Hydrolysis of Na atoms can produce solvated (or delocalized) electrons in an ice film, and the solvated electrons may promote the collision-induced reaction and OD- emission during the Cs+ impact especially at high energy. The presence of solvated electrons is suggested by theoretical calculations17,18 and the Na(3s) feature in MIES experiments,7,8 although the Na(3s) feature can equally be assigned to the 3s electron of neutral Na atom.7,8 The Na(3s) peak disappears at high temperatures (>115 K),7 which suggests that solvated electrons (and neutral Na atoms also) are absent from the present surface under high temperature conditions. The solvated electrons disappear by reacting with water molecules to form OD- ions or may be trapped underneath the surface. Under these conditions, preformed OD- ions may be considered as a primary source of OD- emission, but the possibility of solvated electrons contributing to collision-induced reaction is not excluded. The latter process is expected to occur at somewhat higher energy because the collision must induce the multiple-step processes of OD- formation and desorption. The threshold energies of two channels are not distinguishable in the curves shown in Figure 4. In Figure 5, the RIS signal intensity for NaOD is plotted as a function of Cs+ energy for the samples at 95 and 135 K. The plots show the summation of CsNaOD+ and Cs(NaOD)(D2O)+ intensities, where the latter is smaller by about one order. As Figure 1 shows, the RIS signal for NaOD disappears when the sample is warmed from 95 and 135 K. However, Figure 5 shows that the NaOD signal reappears from the 135 K sample when the Cs+ energy is raised higher, with its intensity continuously increasing with increase in Cs+ energy. The result indicates that NaOD molecules still exist at 135 K, but they are just undetectable by RIS at low Cs+ energies. The effect of increasing the incident Cs+ energy is to increase RIS yield for strongly adsorbed species28-30 and to increase RIS probing depth. The observation, therefore, suggests that, upon warming the ice film, NaOD becomes more strongly bound to the surface or occupies a relatively subsurface position. These situations may result when NaOD is solvated by an increasing number of water molecules at elevated temperature. The strong solvation of NaOD is also consistent with the fact that the CsNaOD+ signal is very weak even when it is detected.

Figure 6. (a) Variation in LES intensities of sodium and hydroxide ions as a function of Na coverage on ice film at 135 K. The logarithmic intensity scale is used on the ordinate. (b) RIS intensity of CsNaOD+ measured as a function of Na coverage. All other experimental conditions were the same as those in Figure 4.

The effect of Na surface coverage on the hydrolysis reaction is examined. Figure 6a shows the LES intensities of sodium and hydroxide ions measured on ice films at temperature 135 K as a function of Na coverage up to 1.36 MLE. The LES signal intensities are plotted on a logarithmic scale. The hydroxide ion intensity increases almost exponentially with the increase in Na coverage from 1 × 103 cps at 0.13 MLE to 2 × 105 cps at 1.36 MLE. Here, the intensities for both OD- and its hydrated ion (D3O2-) increase rapidly. On the other hand, the sodium ion intensity rather decreases after Na coverage exceeds ∼0.3 MLE. The nearly exponential increase in hydroxide ion intensity and the accompanying decrease in sodium ion intensity cannot be explained by a stoichiometric increase in the hydroxide ion population at the surface at high Na coverage. The result implies a certain change in the surface properties, which greatly enhances OD- emission but suppresses Na+ emission. Possible interpretations of this observation will be discussed in section 4. Figure 6b plots CsNaOD+ intensity on the linear scale as a function of Na coverage. It shows that the NaOD surface population increases with increasing Na coverage. However, this increase is only mild compared to the rapid increase of OD- intensity and approximately follows the behavior of Langmuir adsorption kinetics. A CsNa+ signal is not detected on the ice films shown in Figure 6. This indicates that Na atoms and neutral clusters are absent from the surface for Na coverage up to 1.36 MLE. It has been shown31 that the RIS technique can detect alkali metal atoms and clusters adsorbed onto Si surfaces. Failure to detect the LES signals of Nan+ (n > 1) ions from the present surfaces, in comparison to their detection by TOF-SIMS under the conditions of high Na coverage and low temperature,9 further supports the view that neutral Na clusters do not form under

Hydrolysis of Sodium Atoms on Water-Ice Films these conditions. It might be suspected that neutral Na species exist in the interior of the films. To examine this possibility, we increased the Cs+ energy up to 60 eV to probe the subsurface species, but neither CsNa+ nor Nan+ (n > 1) signals were detected. 4. Discussion A. Nature of Na Hydrolysis Products. The results in the previous section indicate that Na hydrolysis on the ice films produces Na+, OD-, and NaOD. These observations will be discussed in comparison with the results of previous studies of alkali metal (Li, Na, K, and Cs) adsorption on ice surfaces.3-9 Evidence for the formation of Na+ is clearly given in the TOFSIMS study of Na-adsorbed ice surfaces at 13 K, which observes the emission of Na+ and its hydrated signals.9 MIES studies3-8 observed the disappearance of neutral Na clusters deposited on ice films upon warming the films above 132 K, indicating the ionization of Na adsorbates. Since the hydrolysis reaction generates Na+ ions, hydroxide ions may be expected to form as a counteranion in the ice films. Nevertheless, evidence of this species has not been unambiguously shown before. MIES and UPS studies observed hydroxyl bands (1π and 3σ) after Na hydrolysis on ice films,3-8 but these features could be attributed to molecular NaOH as well as isolated OH- ions. (In the following discussion, the hydroxide ion is represented by either OH- or OD- depending on which species is used in the specified work, but the distinction here is only isotopic, without any chemical significance.) The LES signals of OD- and its hydrated ions in Figure 1 suggest the presence of isolated OD- species at the ice surface. At the higher temperature at which the diffusion of water molecules and hydration processes can be activated, the hydroxide ion signals increase while the sodium ion signals decrease (Figure 3a). This observation, revealing the independent motions of OD- and Na+ ions, supports the view that they are created as separated species rather than a combined molecular species. Molecular NaOD is detected by the CsNaOD+ signal with weak intensity in RIS spectra. This signal disappears from the surface by increasing the temperature above 135 K (Figure 3), but it reappears upon the increase of Cs+ energy (Figure 5). Thus, we think that not all NaOD species dissociate to Na+ and OD- even at high temperature, but rather they populate the subsurface region or become effectively covered by solvating water molecules. The weak intensity of the CsNaOD+ signal is consistent with the strong solvation of NaOD. According to the investigation by Souda,9 an “unhydrated NaOH” layer forms at an ice film surface at low temperature (13 K) or in the multilayer Na coverage regime. The “unhydrated NaOH” layer disappears when the film temperature is raised above 100 K.9 These conditions suggest that the “unhydrated NaOD” domains will not form in the present study, even under the conditions of the highest coverage examined (1.36 MLE at T ) 135 K). Thus, we consider that the CsNaOD+ signal comes from undissociated, hydrated NaOD species rather than from the “unhydrated NaOD” layer. Neutral Na species have been observed in previous studies6-9 on ice films at low temperature, as they are detected by Na(3s) electron emission in MIES experiments6-8 and by the emission of Na2+ ions in TOF-SIMS experiments.9 The neutral Na clusters form only in the multilayer Na coverage regime at 100 K, and they disappear as the temperature is raised above ∼135 K in both studies.6-9 These observations are consistent with our LES and RIS results that Na atoms are hydrolyzed completely under the present conditions (T g 95 K, θNa e 1.36 MLE).

J. Phys. Chem. C, Vol. 113, No. 1, 2009 325 The reaction paths leading to the hydrolysis products may be considered in light of the results of theoretical calculations.17,19 Upon the adsorption of a Na atom onto the ice surface, a 3s electron released from the Na atom may be solvated near the Na+ ion or be spread over a considerably large distance. The solvated electron located close to Na+ can attack a proton in a water molecule to produce a hydrogen atom and a sodium hydroxide molecule.17,19 On the other hand, the reaction of a solvated electron far away from Na+ can produce an isolated hydroxide ion.19 At the initial stage of the hydrolysis reaction, the ratio of the branching into the products of sodium hydroxide and isolated sodium and hydroxide ions will be determined by the probabilistic spatial distribution of solvated electrons in an ice film. At high temperature, the diffusion of water molecules and proton transport processes may separate the sodium and hydroxide ions farther in ice film, as observed in Figure 3. B. Distribution of Na+ and OD- Ions near the Ice Surface. Figure 3a shows an increase in hydroxide ion intensity at high temperature, accompanied by a decrease in sodium ion intensity, which suggests that OD- ions thermally migrate to the surface of an ice film and Na+ to the interior. The interior migration of Na+ ions at high temperature agrees with the observations from adsorption studies22,32-35 of sodium halide salts (NaX for X ) F, Cl, Br, and I) on ice films. Other possibilities than the surface segregation of OD- may include the hydrolysis of unreacted Na atoms or solvated electrons trapped beneath the surface, producing additional OD- ions. However, Na atoms are not detected even in the subsurface region, as mentioned above. Moreover, the accompanying decrease of sodium ion intensity may exclude these possibilities. Sodium ion intensity does not drop all the way to zero even at the highest temperature examined (Figure 3a). This behavior differs from that of Na+ ions produced by the ionization of sodium halide salts adsorbed on ice films;22,32-35 as Figure 3a also shows, the sodium ion signal from the ionization of NaF disappears completely from the surface at T > 130 K. For the case of Na hydrolysis, therefore, the remaining portion of the sodium ion intensity at high temperature is attributed to other species than isolated Na+ ions. Since NaOD is also observed, it is quite possible that this species gives rise to extra sodium ion intensity by collisional dissociation during LES process. Because this effect will add extra intensity to the observed hydroxide ion signal also, one can suspect that it may alter the temperature-dependent variation of the hydroxide ion curve. However, as Figure 5 shows, the population of NaOD changes only little upon a temperature increase from 95 to 135 K. Thus, the contribution of NaOD dissociation to sodium and hydroxide ion intensities will be rather uniform in the examined temperature range, rather than affecting only one particular ion curve in a specific temperature range. This is supported by the parallel behavior of the two sodium ion curves resulting from NaF ionization and Na hydrolysis in Figure 3a. Oriented NaOD molecules are unlikely to form on the present surfaces, because the “unhydrated NaOD” layer is absent and only hydrated NaOD species exist. For this reason, the separation of sodium and hydroxide ion curves observed at high temperature is most likely due to the migration of isolated Na+ ions to the film interior and OD- ions to the surface. In Figure 6a, the LES intensity of hydroxide ion increases almost exponentially with increasing coverage of Na, whereas the sodium ion intensity concurrently decreases. Such a drastic change in the hydroxide ion intensity seems to be incompatible with any simple explanation based on the increase of hydroxide ion population at the surface, whether it is a result of reaction

326 J. Phys. Chem. C, Vol. 113, No. 1, 2009 stoichiometry or the surface segregation of OD- ions. The observation is more likely to indicate a change in the electronic property or structure of the film surface, which affects the efficiency of ion emission from the surface. Na adsorption reduces the work function (φ) of a thin ice film,8 as it does for clean metal surfaces. The yield of thermionic emission from a surface varies exponentially with surface work function according to the Saha-Langmuir equation, which can be written as f+ ∝ exp[(φ - IE)/kBT] or f- ∝ exp[(EA - φ)/kBT], where f+ is the ratio of positive ion to neutral flux, f- is the ratio of negative ion to neutral flux, IE is the ionization energy of neutral species, EA is the electron affinity, and kB is Boltzmann constant. Therefore, the reduced work function of the ice surface will increase the OD- emission yield in the LES process, whereas the Na+ emission yield will decrease. Structural changes of the surface may also be considered. Suppose that molecular orientation of NaOD species is aligned such that the OD group points toward the vacuum, when an “unhydrated NaOH” layer is formed at high Na exposures. In this case, the OD moiety may be efficiently detached, as an OD- ion, by the Cs+ impact. However, the “unhydrated NaOH” domains are probably absent from the present surfaces at 135 K, as mentioned above, where water solvation is still quite effective, as indicated by the concurrent increases of OD- and its hydrated signals (D3O2-). Thus, we consider that the assumption of aligned NaOD domains at the surface lacks additional supporting experimental evidence or chemical rationale. It is interesting to consider whether the trend observed for Na+ and OD- distributions near the ice film surfaces can be extended to liquid water surfaces. The distribution of hydroxide ions at the air/water interface is a subject of active investigation and controversy in recent years; air bubble and oil droplet experiments indicate the accumulation of hydroxide ions at the interface,36,37 whereas molecular dynamics simulations suggest no appreciable enhancement in the hydroxide ion surface population.38 Vibrational sum frequency generation (VSFG) studies of NaOH solutions38,39 have suggested that hydroxide ions tend to reside in the interior of solution, but re-examination of the system by using the phase-sensitive VSFG technique40 suggests the opposite behavior of hydroxide ions residing near the solution interface. Reports to date seem to indicate that ASW and liquid water exhibit qualitatively similar trends for the interfacial distributions of hydronium ion,27,38,41 alkali metal ions,22,32-35,42 and halide ions.22,32-35,42 The hydroxide ion behavior may also be in this category, but one may need to adequately account for the dynamically different properties of ASW and liquid water surfaces. 5. Conclusions The present study examines the hydrolysis reaction of Na atoms at ice film surfaces for the temperature range 95-135 K and Na coverage up to 1.36 MLE. Through the inspection of the products at the surfaces with LES and RIS techniques and the interpretation of the results in light of the hydrolysis mechanism suggested in previous works, we have been able to reach the following conclusions: (i) Na atoms adsorbed on a D2O-ice film undergo a hydrolysis reaction to produce Na+ and OD- ions as well as NaOD molecules on the surface. (ii) The hydrolysis reaction can occur with single Na atoms. The adsorbed Na atoms all react away under the present conditions without forming neutral Na clusters on the surface. (iii) Na+ and OD- ions are efficiently solvated by water molecules to exist as isolated species.

Kim et al. (iv) NaOD molecules are also efficiently hydrated at high temperature to occupy subsurface positions, but a substantial portion of the molecules remains undissociated. (v) OD- ion tends to reside at the ice surface at elevated temperature, whereas Na+ ion migrates toward the film interior. Acknowledgment. This work was supported by the Korea Science and Engineering Foundation Grant funded by the Korean government (MEST) (R11-2007-012-02001-0). Supporting Information Available: Variation in LES intensity of hydroxide ions measured on a pure D2O film as a function of incident Cs+ energy (Figure S1). This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Thiel, P. A.; Hrbek, J.; Depaola, R. A.; Hoffmann, F. M. Chem. Phys. Lett. 1984, 108, 25. (2) Blass, P. M.; Zhou, X.-L.; White, M. J. J. Phys. Chem. 1990, 94, 3054. (3) Gu¨nster, J.; Krischok, S.; Stultz, J.; Goodman, D. W. J. Phys. Chem. B 2000, 104, 7977. (4) Gu¨nster, J.; Krischok, S.; Kempter, V.; Stultz, J.; Goodman, D. W. Surf. ReV. Lett. 2002, 9, 1511. (5) Gu¨nster, J.; Souda, R. Chem. Phys. Lett. 2003, 371, 534. (6) Krischok, S.; Ho¨fft, O.; Gu¨nster, J.; Souda, R.; Kempter, V. Nucl. Instrum. Methods Phys. Res., Sect. B 2003, 203, 124. (7) Borodin, A.; Ho¨fft, O.; Kahnert, U.; Kempter, V.; Allouche, A. Vacuum 2004, 73, 15. (8) Gu¨nster, J.; Kempter, V.; Souda, R. J. Phys. Chem. B 2005, 109, 17169. (9) Souda, R. J. Chem. Phys. 2006, 125, 044706. (10) Bewig, L.; Buck, U.; Rakowsky, S.; Reymann, M.; Steinbach, C. J. Phys. Chem. A 1998, 102, 1124. (11) Buck, U.; Steinbach, C. J. Phys. Chem. A 1998, 102, 7333. (12) Bobbert, C; Schulz, C. P. Eur. Phys. J. D 2001, 16, 95. (13) Steinbach, C.; Buck, U. Phys. Chem. Chem. Phys. 2005, 7, 986. (14) Ramaniah, L. M.; Bernasconi, M.; Parrinello, M. J. Chem. Phys. 1998, 109, 6839. (15) Mundy, C. J.; Hutter, J.; Parrinello, M. J. Am. Chem. Soc. 2000, 122, 4837. (16) Mercuri, F.; Mundy, C. J.; Parrinello, M. J. Phys. Chem. A 2001, 105, 8423. (17) Ferro, Y.; Allouche, A. J. Chem. Phys. 2003, 118, 10461. (18) Ferro, Y.; Allouche, A.; Kempter, V. J. Chem. Phys. 2004, 120, 8683. (19) Chan, K. W.; Siu, C.-K; Wong, S. Y.; Liu, Z.-F. J. Chem. Phys. 2005, 123, 124313. (20) Han, S.-J.; Lee, C.-W.; Hwang, C.-H.; Lee, K.-H.; Yang, M. C.; Kang, H. Bull. Korean Chem. Soc. 2001, 22, 883. (21) Kang, H. Acc. Chem. Res. 2005, 38, 893. (22) Kim, J.-H.; Kim, Y.-K.; Kang, H. J. Phys. Chem. C 2007, 111, 8030. (23) Denzler, D. N.; Wagner, S.; Wolf, M.; Ertl, G. Surf. Sci. 2003, 532-535, 113. (24) Doering, D. L.; Semancik, S.; Madey, T. E. Surf. Sci. 1983, 133, 49. (25) Hahn, J.-R.; Lee, C.-W.; Han, S.-J.; Lahaye, R. J. W. E.; Kang, H. J. Phys. Chem. A 2002, 106, 9827. (26) (a) Lahaye, R. J. W. E.; Kang, H. ChemPhysChem 2004, 5, 697. (b) Kim, J.-H.; Lahaye, R. J. W. E.; Kang, H. Surf. Sci. 2007, 601, 434. (27) (a) Jung, K.-H.; Park, S.-C.; Kim, J.-H.; Kang, H. J. Chem. Phys. 2004, 121, 2758. (b) Park, S.-C.; Jung, K.-H.; Kang, H. J. Chem. Phys. 2004, 121, 2765. (28) Yang, M. C.; Hwang, C.-H.; Kang, H. J. Chem. Phys. 1997, 107, 2611. (29) Kim, C.-M.; Hwang, C.-H.; Lee, C.-W.; Kang, H. Angew. Chem., Int. Ed. 2002, 41, 146. (30) Kim, K.-Y.; Kim, J.-H.; Cho, J.-H.; Kleinman, L.; Kang, H. J. Chem. Phys. 2003, 118, 6083. (31) Han, S.-J.; Park, S.-C.; Lee, J.-G.; Kang, H. J. Chem. Phys. 2000, 112, 8660. (32) Borodin, A.; Ho¨fft, O.; Kahnert, U.; Kempter, V.; Poddey, A.; Blochl, P. E. J. Chem. Phys. 2004, 121, 9671. (33) Borodin, A.; Ho¨fft, O.; Kempter, V. J. Phys. Chem. B 2005, 109, 16017. (34) Kim, J.-H.; Shin, T.; Jung, K.-H.; Kang, H. ChemPhysChem 2005, 6, 440.

Hydrolysis of Sodium Atoms on Water-Ice Films (35) Ho¨fft, O.; Borodin, A.; Kahnert, U.; Kempter, V.; Dang, L. X.; Jungwirth, P. J. Phys. Chem. B 2006, 110, 11971. (36) Graciaa, A.; Creux, P.; Lachaise, J. Electrokinetics of Gas Bubbles. In Interfacial Electrokinetics and Electrophoresis; Delgado, A. V., Ed.; Marcel Dekker: New York, 2002; Vol. 106, p 825. (37) Beattie, J. K. Phys. Chem. Chem. Phys. 2008, 10, 330. (38) Mucha, M.; Frigato, T.; Levering, L. M.; Allen, H. C.; Tobias, D. J.; Dang, L. X.; Jungwirth, P. J. Phys. Chem. B 2005, 109, 7617.

J. Phys. Chem. C, Vol. 113, No. 1, 2009 327 (39) Tarbuck, T. L.; Ota, S. T.; Richmond, G. L. J. Am. Chem. Soc. 2006, 128, 14519. (40) Shen, Y. R. To be submitted for publication. (41) Buch, V.; Milet, A.; Vacha, R.; Jungwirth, P.; Devlin, J. P. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 7342. (42) Jungwirth, P.; Tobias, D. J. J. Phys. Chem. B 2002, 106, 6361.

JP807774V