O/CO/H2O ices

Distance-dependent radiation chemistry: Oxidation versus hydrogenation of CO in electron-irradiated H₂O/CO/H₂O ices

Nikolay G. Petrik*, Rhiannon J. Monckton, Sven P. K. Koehler and Greg A. Kimmel*

*Corresponding authors. Email addresses – [e-mail], [e-mail]

Abstract

Electron-stimulated oxidation of CO in layered H₂O/CO/H₂O ices was investigated with infrared reflection-absorption spectroscopy (IRAS) as function of the distance of the CO layer from the water/vacuum interface. The results show that while both oxidation and reduction reactions occur within the irradiated water films, there are distinct regions where either oxidation or reduction reactions are dominant. At depths less than ~ 15 ML from the vacuum interface, CO oxidation to CO₂ dominates over the sequential hydrogenation of CO to methanol (CH₃OH), consistent with previous observations. At its highest yield, CO₂ accounts for ~45% of all the reacted CO. Another oxidation product is identified as the formate anion (HCO₂^-). In contrast, for CO buried more than ~ 35 ML below the water/vacuum interface, the CO-to-methanol conversion efficiency is close to 100%. Production of CO₂ and formate are not observed for the more deeply buried CO layers, where hydrogenation dominates. Experiments with CO dosed on pre-irradiated ASW samples suggest that OH radicals are primarily responsible for the oxidation reactions. Possible mechanisms of CO oxidation, involving primary and secondary processes of water radiolysis at low temperature, are discussed. The observed distance-dependent radiation chemistry results from the higher mobility of hydrogen atoms that are created by the interaction of the 100 eV electrons with the water films. These hydrogen atoms, which are primarily created at or near the water/vacuum interface, can desorb from or diffuse into the water films, while the less-mobile OH radicals remain in the near-surface zone resulting in preferential oxidation reactions there. The diffusing hydrogen atoms are responsible for the hydrogenation reactions that are dominant for the more deeply buried CO layers.

1. 
   Introduction
   

Radiation- and photo- induced (“non-thermal”) processes in water and aqueous systems are very important in a broad range of applications, including nuclear technology,¹ astrochemistry and planetary sciences,² radiation- and photo-catalysis,³ and radiation biology and medicine.⁴ Various types of high-energy radiation (, , , etc.) interacting with condensed matter generate cascades of low-energy electrons, the most abundant of which have energies below 70 eV.⁵ Those electrons are responsible for most of the radiation-induced chemical transformations.^5-6 For this reason, fundamental studies of radiation chemistry often use low energy electrons. However due to their large scattering cross sections and correspondingly short mean-free paths, experiments utilizing low-energy electrons are performed in vacuum with solid forms of water – amorphous solid water (ASW) or crystalline ice – as model systems for condensed water.

Molecular ices also have direct relevance to astrochemistry and planetary sciences.² Molecular clouds, which include dust grains coated with molecular ices, are a birthplace for molecules in space. The dominant components of the ice mantles are H₂O, CO and CO₂,^2b and at the typical temperatures of these clouds (10 – 100 K), the water is predominantly found as ASW.^{2a, 2b} Comets and some moons of the gas giant planets are primarily composed of, or covered by, ices.^{2d, 2g, 2i, 7} All of these objects are exposed to various types of radiation including energetic (keV–MeV) ions, UV-VUV photons and cosmic rays. Radiolysis of ices is an essential component of molecular synthesis / decomposition in space.²

The ices in space are typically not pure but a mixture of water with various simple molecules like CO, CO₂ H₂CO, CH₃OH, CH₄, NH₃ and others.^{2a, 2b, 7b, 8} A substantial number of papers have been published in the last few decades investigating the radiolysis and VUV photolysis of water-molecular ice mixtures for astrochemistry and planetary sciences, including H₂O + CO systems.^{8a, 9} A broad variety of ice compositions, radiation sources, experimental conditions and techniques were employed, which makes comparing the different results difficult. Previous investigations of H₂O + CO ices using infrared (IR) spectroscopy and other methods found two major radiation-induced processes for CO buried in the ASW – hydrogenation/reduction (mainly to methanol) and oxidation (mainly to CO₂).^{8a, 9} Typically, IR spectra revealed that the products of both hydrogenation and oxidation of CO appeared simultaneously within the irradiated films.

The composition of the H₂O + CO ice is an important parameter for the experiments. In many cases, H₂O + CO gas mixtures with comparable concentrations of both components are dosed on a cold substrate. In these systems, there is a high probability of direct radiation impact on the CO. Moore et al.,^{8a, 9a} Kaiser et al.,^9b Parent et al.,^9h and Watanabe et al.^9c irradiated H₂O + CO mixtures with H₂O:CO ratios ranging from 0.5:1 to 5:1 at 10 – 20 K with 0.8 MeV protons, 5 keV electrons, soft X rays (< 1 keV), and VUV photons (8 – 10 eV), respectively. For these experiments, conversion of CO to CO₂ was the dominant reaction, while smaller amounts of formic acid (HCOOH), the hydrocarboxyl radical (HOCO), the formyl radical (HCO), formaldehyde (H₂CO) and methanol (CH₃OH) were also observed.

For dilute CO solutions in H₂O dosed on a cold substrate, the CO – H₂O reactions are primarily initiated by the radiolysis of water, with the water radiolysis products subsequently attacking the CO. Moore, et al.^8a and Valero, et al.^9f irradiated H₂O + CO mixtures (10:1 to 20:1) at 10 – 20 K with 0.8 MeV protons and VUV photons (8 – 10 eV), respectively. With the exception of HCOOH and HOCO, which were not observed in the dilute systems, the reaction products were similar for the dilute and concentrated systems. Several experiments have also investigated layered H₂O – CO – H₂O films, where pure components were dosed sequentially.^{9e, 10} For example, Yamamoto et al.^9e dosed CO on 5 nm thick water films, then capped the CO with 30 nm thick water layers, and irradiated the ices with 200 eV electrons at 25 – 140 K. They also observed CO₂ and reduction products (HCO, CH₂O and CH₃OH). All the product yields increased with temperature of irradiation.

Due to the short penetration depth of low-energy electrons in water, the approach of Yamamoto, et al. utilizing layered films of H₂O and CO is attractive. While not directly applicable for many astrophysical environments, this approach allows for precise control of the system, opening up the possibility of studying the spatial distribution of reactions and products, and thereby providing important fundamental insights. Previously, experiments with layered films of H₂O and D₂O have proven useful in studies of radiation-induced processes in neat ASW films.¹¹ We used this approach in a recent study of the radiation-induced hydrogenation of CO buried in ASW.¹⁰ Layered films were grown so as to position the molecules of interest precisely ( 2 ML) at pre-determined locations within the ASW films. For CO layers buried 60 ML or deeper within the ASW films, we observed only hydrogenation products (HCO, CH₂O and CH₃OH) after irradiation with 100 eV electrons.¹⁰ D₂O/H₂O layering experiments revealed that hydrogen atoms, produced by water radiolysis near the ASW/vacuum interface, were responsible for the hydrogenation reactions. The majority of the hydrogen atoms reacted in the near-surface region or desorbed into the vacuum. However, some of the hydrogen atoms diffused into the film and reached the buried CO layer. Sequential reactions with these hydrogen atoms then converted CO to methanol (CH₃OH):

(1)
Here we investigate the electron-stimulated reactions in layered H₂O/CO/H₂O ices for CO layers that are within 60 ML of the ASW/vacuum interface. Placing CO at different locations within the ASW film maps out zones with very different radiation chemistry. For water cap layer coverages, _Cap, less than 15 ML, the CO is preferentially oxidized to CO₂. For _Cap > 15 ML, more of the CO is reduced to methanol than is oxidized to CO₂. The observation of both oxidation and reduction of CO is consistent with earlier investigations.^29,30,32-34 For _Cap = 10 ML, CO₂ accounts for ~45% of all the converted CO. Another product of radiation-induced oxidation of CO is identified as the formate anion (HCO₂^-). For pure ASW samples irradiated at low temperatures and then dosed with CO, CO₂ is produced, suggesting that OH radicals are the main oxidants. The results presented here provide new insights on the distribution of oxidation and reduction reactions within irradiated water films and the processes leading to these different zones of reactivity within the films.

2. 
   Experimental Procedure
   

The experiments were carried out in an ultrahigh vacuum (UHV) system that has been described previously,^{10, 12} using procedures very similar to our previous investigation of the electron-stimulated reactions in layered H₂O/CO/H₂O systems.¹⁰ Here we give a brief overview of the main components and procedures. The system is equipped with a molecular beam line for dosing water, CO and other adsorbates on the sample, a closed-cycle helium cryostat for sample cooling, a low-energy electron gun (Kimball Physics, model ELG-2), a quadrupole mass spectrometer (Extrel, model EXM720), and a Fourier-transform infrared spectrometer (Bruker, Vertex 70) for infrared reflection absorption spectroscopy (IRAS).

As discussed in the introduction, the key aim of these experiments is to investigate the non-thermal reactions of water and CO versus the position of a thin CO layer within the water films. Thus, controlling and characterizing the distribution of CO within the water films is crucial for the success of the endeavor.¹³ Since electron-stimulated reactions that can occur at the water/Pt(111) interface^{11a, 11b, 11e, 14} are not the focus of the current experiments, a smooth, dense ASW layer of at least 30 ML was initially deposited on the Pt(111) at 100 K.¹⁵ Various control experiments (data not shown) showed such films are sufficiently thick that reactions occurring at or near the water/Pt(111) interface did not influence the CO-water reactions. The CO was deposited on the ASW spacer layer at T < 30 K and buried under ASW “cap” layers of varying coverages, _Cap. The ASW cap layer was constructed by adsorbing 4 ML of ASW at T < 30 K, then setting the temperature to 100 K and adding additional water to achieve the desired total cap layer coverage. While heating the sample to 100 K to add any additional water to the cap layer, approximately 1/3 of the adsorbed CO desorbs, but the remaining CO is trapped within the film and, in the absence of electron-irradiation, will not desorb on the timescale of the experiments. The coverage of the trapped CO, _CO, was ~2.5 × 10¹⁴ cm^-2. This procedure was developed based on several control experiments that were conducted to determine the distribution of CO within the ASW films, which are described in previous paper and its supporting information.¹⁰ Since previous research has shown that the diffusion of atoms and small molecules in ASW is negligible below ~120 K,¹⁶ the CO remains trapped within the first 4 ML when additional water is dosed at 100 K.

The incident energy of the electrons, E_i, used to irradiate the films was 100 eV, and the instantaneous current densities were ~1.5 × 10¹⁵ cm^-2s^-1. The electron beam was smaller than the molecular beam spot size on the sample (~1.5 mm and 7.5 mm, respectively). To produce a uniform electron fluence across the films, the electron beam was rastered over the surface.^11b Low-energy electrons, such as the 100 eV electrons used here, efficiently sputter water films.^11c Therefore, to maintain approximately constant ASW cap layer coverages during the experiments, additional water was dosed during the experiments to compensate for the electron-stimulated sputtering. After initially preparing the layered ASW films with the trapped CO, as described above, IRAS spectra were recorded prior to irradiating the films. After electron irradiation, additional water was dosed to account for the amount sputtered during the irradiation, and an IRAS spectrum for the irradiated film was obtained. The sequence of electron irradiation, water dosing and IRAS was then repeated the desired number of times to obtain a series of IRAS spectra for increasing electron fluences while maintaining an approximately constant water coverage.

Due to the small amount of CO initially buried with the ASW films, the electron-stimulated reaction products can be difficult to discern in the raw IRAS spectra. However, the reaction products are more readily observed in differential absorbance spectra obtained by subtracting the irradiated ASW films without buried CO from the corresponding spectra with the buried CO (for the same electron fluences). The difference spectra are particularly useful since electron irradiation leads to spectral changes in the absorbance features of the ASW, even in the absence of CO. These changes can make it difficult to observe the small signals associated with the CO-H₂O reaction products. However, even using these differential absorbance spectra, there are some residuals left from the major water bands after subtraction.

3. 
   Results
   

Figure 1a shows the IRAS spectra for a CO/ASW film prior to electron irradiation (black lines) and after irradiation at 100 K with 100 eV electrons and a fluence of 1.6 × 10¹⁵ e^-/cm² (red lines). The CO was deposited on a 30 ML ASW layer and was capped with a 10 ML ASW layer as described in the experimental section. The loss of the CO peak due to electron irradiation, the appearance of a CO₂ peak and changes in the OH-stretch region are readily apparent in the IRAS spectra (see expanded scale inset), while other changes can only be seen in differential absorbance spectra such as in Figure 1b. Figure 1a also shows the spectrum of an electron-irradiated neat water film deposited using the same procedure but without CO (i.e. 30 ML ASW plus a 10 ML ASW cap) (green lines).

The black, red, blue and green lines in Figure 1b show a series of differential absorption spectra from irradiated ASW films with and without trapped CO for increasing electron fluences. For comparison, we also show experimental IR spectra for formic acid (HCOOH, pink line) and formaldehyde (H₂CO, orange line) buried in the ASW layer using the same procedure that was used to trap the CO layers. As seen in the figure, new features at 1018, 1352, 1387, 1594, ~ 1700 and 2341 cm^-1 appear in the irradiated sample. Peaks at 1018, 2831 (not shown) and 2856 cm^-1 (not shown) are due to methanol, and the peak at 2341 cm^-1 corresponds to CO₂. These two molecules were reported earlier as main products of irradiated H₂O/CO ices.^{8a, 9}

Figure 1. IRAS spectra for electron-irradiated ¹²C¹⁶O / H₂¹⁶O films with _cap = 10 ML and T_irr = 100 K. a) Absorbance spectra for the samples before and after irradiation (black and red lines), as well for a sample irradiated without any co-adsorbed CO (green line). b) Differential absorbance spectra versus irradiation fluence from 0 to 6.9 x 10¹⁵ e^-/cm² (black, red, blue and green lines). The IR spectra of formic acid (HCOOH, pink line) and formaldehyde (H₂CO, orange line) buried in H₂O are shown for comparison. The spectra for formic acid and formaldehyde have been arbitrarily scaled. The IR spectrum of the formyl anion (HCOO^-) in liquid water¹⁷ is also shown (light blue line).

For the irradiated films, a new peak at 1594 cm^-1 (see Fig. 1b) is quite similar to the asymmetric stretch of CO in the formate anion HCO₂^- (Fig. 1b, light blue line).^17-18 The formate anion has a characteristic doublet due to the CO symmetric stretch and CH bend modes, which we also see in the spectra of irradiated CO at 1352 and 1387 cm^-1. A similar triplet of 1589, 1384, and 1353 cm^-1 peaks was also attributed to the formate anion in IR spectra of proton irradiated H₂O + CH₃OH ice.¹⁹ The formate anion was also discussed as a possible candidate for the small non-resolved shoulder around 1550–1600 cm⁻¹ and a small peak at 1358 cm⁻¹ in an electron irradiated, equimolar mixture of H₂O + CO.^9b

The nature of the broad feature near ~1700 cm^-1 in Figure 1b is not as clear. It is close to the 1715 cm^-1 peak from the C=O stretch of formic acid (see Fig. 1b, pink line).^{9c, 20} However, formic acid should also have a strong peak near 1230 cm^-1, which is not seen in the spectrum of irradiated CO. The frequency of the _S(C=O) stretch mode of formaldehyde buried in ASW is at 1734 cm^-1 (Fig. 1b, orange line), but the formaldehyde peak is much narrower than the broad peak observed in the irradiated films. Formaldehyde also has the CH₂ scissors peak at 1496 cm^-1, which we do not see in the spectra of irradiated CO + H₂O. (Formaldehyde is observed for CO buried deeper in the ASW layer, where the hydrogenation reactions dominate.¹⁰) Changing the isotopic composition of the CO (¹²C¹⁶O and ¹³C¹⁶O) and the water (H₂¹⁶O, H₂¹⁸O and D₂¹⁶O) results in red shifts for the formate and the ~1700 cm^-1 peaks relative to ¹²C¹⁶O in H₂¹⁶O (see Fig. S1). The ~1700 cm^-1 feature can be seen in D₂O, but it appears relatively smaller and sharper.

There is no measurable peak of the trans- (cis-) HOCO (hydrocarboxyl) radical observed in the spectra of the irradiated samples around 1848 (1797) cm^-1.²¹ HOCO is considered as an intermediate in the CO oxidation with OH radicals to CO₂,²² but the published data on this product are controversial. HOCO was reported in the spectrum of electron-irradiated equimolar mixture of H₂O + CO,^9b and among the reaction products of OH radicals with solid CO.^21b On the other hand, the HOCO peak was not observed in IR spectra of proton irradiated H₂O + CO ice (20:1),^8a or in the layered H₂O/CO/H₂O films irradiated with electrons,^9e while a big CO₂ peak was detected in both cases. It is possible that the HOCO radical can only be observed in the systems with relatively high concentrations of CO, which minimizes the probability for this radical to react with the products of water radiolysis.

Figure 2. CO (black circles), CO₂ (red circles) and CH₃OH coverages (green squares) versus electron fluence in irradiated CO/ASW films for _Cap = 10 ML and T_irr = 100 K. The purple triangles and blue diamonds show the (arbitrarily scaled) integrated absorbances of formate (HCO₂^-) and an unidentified species with a peak at 1700 cm^-1, respectively.

Figure 2 shows the coverages of CO, CO₂ and methanol (derived from the integrated IRAS signals) versus electron fluence for a film where a CO layer was dosed on a 30 ML ASW film, capped with a 10 ML ASW film, and then irradiated at 100 K. The amount of CO decreases exponentially as the fluence increases (Fig. 2, black circles). In this experiment, CO₂ is the major product. At their maximum yields, about 45% of the CO has been converted to CO₂ and 18% to methanol. At higher electron fluences, the CO₂ decreases indicating radiation-induced decomposition of CO₂ in the ASW. The CO₂ yield versus fluence can be nicely fit to two exponentials: _CO2() = K[exp(-k₁) – exp(-k₂)], describing the sequential reaction CO k₁ CO₂ k₂ products , where K = _CO(0) k₁/(k₂-k₁) and  is the electron fluence. No new products are observed in the IRAS spectra at higher doses, when CO₂ decays.

The integrated IRAS signals of formate (1594 cm^-1) and the 1700 cm^-1 feature versus electron fluence are also shown in Fig. 2 (purple triangles and blue diamonds, respectively). Since we do not have coverage calibrations for these peaks, the intensities have been arbitrarily scaled. The formate grows with fluence in a fashion quite similar to the CO₂, maximizing near  ~1.5 x 10¹⁵ electron/cm², but then decays at a rate that is slower than for CO₂. The ~ 1700 cm^-1 feature evolves differently: It maximizes at a lower fluence near  ~ 0.7 x 10¹⁵ electron/cm², and then decays.

In control experiments, we have irradiated CO₂ buried in ASW under a 10 ML cap. In that case, the CO₂ signal decays with a rate comparable to the CO₂ decay in Fig. 2 (see Fig. S2a). The methanol yield from irradiated CO₂/ASW films is approximately half its yield in CO/ASW films with the same _Cap. On the other hand, both the formate peak at 1594 cm^-1 and the ~1700 cm^-1 feature are detected in the irradiated CO₂/ASW films (see Fig. S1 and S2b). The formate yield from the CO₂/ASW films is comparable to the formate yield from CO/ASW films. No other products are seen in the IR spectra. From the data in Fig. 2 and S1, one might assume that formate (and the 1700 cm^-1 feature) is produced from CO₂, which is itself produced when irradiating the CO/ASW films, e.g. CO  CO₂  HCOO^-. However if this was the case, then we would expect a kinetic signature of such a reaction scheme. Specifically, the production of the formate would be delayed relative to CO₂. Instead, the initial formate and CO₂ kinetics are similar (see Fig. 2). This indicates that in the CO + H₂O system, most of the formate is produced directly from CO and not from CO₂.

From the data presented in Fig. 2, the initial reaction yields, which we define as the number of reactions per incident electron, can be determined for a cap layer coverage of 10 ML. Similar experiments were performed for a range of cap layer coverages from 4 ML to 60 ML. Figure 3 shows the initial yields for the CO decomposition, and the CO₂ and CH₃OH accumulation as function of the H₂O cap layer coverage. Because the number of reactions per incident electron decrease as _Cap increases, the maximum electron fluence used for each coverage increased from 4  10¹⁴ for 4 ML to 1.2  10¹⁷ electron/cm² for 60 ML. The CO reaction yield decreases dramatically from ~2 to ~ 0.0015 reactions per electron for 4 ML ≤ _Cap ≤ 60 ML (Fig. 3, black circles). Among the products, CO₂ has the largest yield for _Cap < 15 ML with its yield decreasing from ~ 0.4 to ~ 0.0013 reactions per electron for 4 ≤ _Cap ≤ 30 ML (Fig. 3, red circles). For the thinnest ASW cap layers, electron-stimulated desorption (ESD) of the CO is significant (see Fig. S3), limiting the amount of CO₂ that is produced. For CH₃OH, the initial yield decreases from ~ 0.05 to ~ 0.0015 reaction per electron for 4 ≤ _Cap ≤ 60 ML (green squares). For _Cap > 15 ML, methanol is the most common reaction product and for _Cap > 30 ML nearly all the CO is converted to methanol. Since we do not have a coverage calibration for the formate peak, we have calculated the initial slope of the HCOO^- integrated IR signal versus electron fluence, which is also shown in Fig. 3 in purple triangles (arbitrary scale). The formate yield decreases by ~ 300 times for 4 ≤ _Cap ≤ 40 ML.

We have previously investigated the electron-stimulated reactions for films with cap layer coverages as large as 400 ML.¹⁰ Those experiments showed that for _Cap > 60 ML, the CO reaction rate was proportional to 1/_Cap while for _Cap < 60 ML, the CO reaction rate increased significantly. The inset to Figure 3 shows the CO reactions per incident electron from the earlier results (open circles) along with the results presented here (solid circles). The results of both experiments are consistent and again show a change in behavior for the cap layer coverage of ~60 ML.

Figure 3. CO (black circles), CO₂ (red circles) and CH₃OH (green squares) reactions per incident electron versus _Cap. The IRAS signal for formate (arbitrarily scaled) is also shown (purple triangles). CO is lost and the other species are produced during electron irradiation. Inset: The CO reactions per incident electron for larger range of cap layer coverages. The open circles show the results from Ref. ¹⁰ .

The temperature dependence of the electron-stimulated reactions in the buried CO layers depends on the position of the CO layer within the ASW films. For example, the electron-stimulated reactions per incident electron for CO with _Cap = 14 ML decreases by a factor of ~ 3 when the irradiation temperature decreases from 100 K to 20 K, as depicted in Figure 4a (black symbols). Accumulation of the main product, CO₂, also slows down correspondingly (red symbols). The subsequent radiation-induced CO₂ decomposition is suppressed even more significantly at 20 K. The slower loss of CO₂ at 20 K is consistent with control experiments where the loss of buried CO₂, irradiated at 100 and 20 K, was measured directly (see Figure S4). The modest temperature dependence of the reactions for _Cap = 14 ML (Fig. 4a) is in sharp contrast to the temperature dependence for _Cap = 45 ML, where the CO decay probability decreases by more than two orders of magnitude in the same temperature range (Figure 4b). Here, CO hydrogenation is the dominant process, and methanol is the main product. The results in Figure 4b are consistent with the dramatic temperature dependence of hydrogenation reactions reported earlier for CO deeply buried in ASW.¹⁰

Figure 4. CO coverage (black circles) and CO₂ coverage (red circles) versus electron fluence in irradiated CO/ASW films. The filled and open symbols are for irradiations performed at 100 K and 20 K, respectively. a) _Cap = 14 ML. b) _Cap = 45 ML.

For thin ASW cap layers, the other products – formate and the species responsible for the peak at ~1700 cm^-1 – are also observed in the samples irradiated at 20 K (Figure S5). For formate, the initial reaction yield is significantly lower at 20 K, while the other species is similar at both temperatures. As a result, the peak at ~1700 cm^-1 appears before the formate at 20 K, while the two species appear simultaneously at 100 K. The distinct kinetics for the two peaks shows that they are associated with different products.

To explore the mechanism(s) by which CO₂ is produced, we have performed a series of experiments where an ASW film was irradiated with 100 eV electrons at 20 K and then annealed to various temperatures, or irradiated at 100 K. Next, CO was adsorbed onto the pre-irradiated ASW films and capped with a 15 ML ASW film. The amount of CO₂ produced was then monitored via IRAS (Fig. 5a), or by measuring the amount of CO₂ that desorbed concomitantly with the water during temperature programed desorption (Fig. 5b). For an ASW film pre-irradiated at 20 K without any further annealing, a substantial amount of CO₂ is produced (Fig. 5a, red line), with a corresponding decrease in the CO signal (Fig. 5a, inset). In contrast, very little CO₂ is produced for a film pre-irradiated at 100 K (Fig. 5a, blue line). For the ASW films pre-irradiated at 20K and then annealed prior to the CO dose, the CO₂ production decreases to zero in the annealing temperature range between 50 and 100 K (Fig. 5b). However, annealing the pre-irradiated films after the CO is dosed and capped with water results in very little change to the CO and CO₂ peaks (data not shown).

Figure 5. a) IRAS spectra for ¹³CO dosed at 20K on H₂¹⁶O film pre-irradiated with electrons at 20 K (middle red trace) and 100 K (upper blue trace), as well as without irradiation (lower black trace) and then capped with _cap = 15 ML at 20 K Inset: Expanded ¹³CO peak showing the loss of CO for the film irradiated at 20 K. b) Amount of CO₂ produced versus annealing temperature for ASW films pre-irradiated at 20 K prior to the CO dose. For this experiment, the CO₂ yield was measured by monitoring the CO₂ desorption during temperature programed desorption (TPD) of the ASW films.

The results in Figure 5 provide important information regarding the mechanism of radiation-induced oxidation of CO buried in an ASW matrix. In these experiments, the CO reacts with products of water radiolysis that remain on the ASW surface after the end of irradiation. Therefore, the data indicate that CO oxidation can proceed by an indirect process, i.e. it does not necessarily involve any direct electron-induced impact on the carbon monoxide. The water radiolysis products that oxidize the CO are thermalized and stable at 20 K. Furthermore, the reaction of CO  CO₂ is efficient at 20 K indicating it is practically barrierless. These observations are consistent with earlier studies indicating that the radiation-induced oxidation of CO in low-temperature ices occurs via a barrierless reaction between thermalized CO and OH radicals: CO + OH  CO₂ + H.^{9b, 9c, 9h, 21b, 23} As we will discuss later, these results are also in agreement with earlier experiments on the mobility and reactions of OH radicals in ASW films in the temperature range ~ 60 – 100 K.^{11e, 24}

4.