Hydrogenation of CO on interstellar dust: what is the role of water molecules?

Abstract

We performed several co-injection experiments varying the concentration of water molecules within a binary mixture (CO/H₂O) co-condensed with (H/H₂) onto a cold mirror maintained at 10 K. We performed the [CO + H] reaction within a H₂O-rich and a N₂-rich environment and confirmed that the gradual replacement of CO molecules by water molecules greatly enhances the yields of H₂CO and CH₃OH, the main products of CO hydrogenation at 10 K. This trend is observed until the concentration of CO molecules becomes too low and significantly favors the side reaction [H + H] recombination over CO hydrogenation. Irradiation of (CO/H₂O) mixed ices with cold H-atoms highlights the action of water molecules, expressed in the spectra as a decrease in the intensity of water-ice's infrared signatures. The catalytic effect observed can be interpreted by physical, chemical, and energetic considerations. This work is inclined toward extending previously reported results with different experimental conditions.

---

1. Introduction

An assortment of silicate and carbonaceous dust coated with accreted atoms (e.g., H, O, C, N) and molecules (e.g., CO) from the gas phase is found within the cold parts of molecular clouds.¹ These accreted adsorbates (i.e., ices) are primarily made of water^2–7 and can also include a scattered amount of organic compounds, such as CH₄, NH₃, XCN, CH₃OH, and CO molecules.⁸ In space, porous icy dust grains can act as reservoirs as well as catalytic surfaces in providing trapping sites and increased residence time that atoms or molecules can spend on the surface, thus stimulating molecules formation.^9,10 Most ices (except for CO) are predicted to form in situ, onto these icy surfaces, via hydrogenation and oxygenation of atoms and small molecules.¹¹ Methanol formation is one of those example where CO hydrogenation (eqn. (1)) converts CO into CH₃OH in four successive steps:¹²

CO → HCO → H₂CO → H₃CO → CH₃OH (1)

The sequential addition of H-atoms to CO molecules in reaction (eqn. (1)) proceeds in the gas phase through two activation barriers, Ea(1) (H + CO → HCO) and Ea(3) (H + H₂CO → CH₃O) which have been estimated at 12–21 kJ mol⁻¹ and 18–21.9 kJ mol⁻¹, respectively, depending on the methodology and level of theory used.^10,13–20 However, it has been shown experimentally^21,22 that CO hydrogenation occurred up to CH₃OH formation when a cold pure CO ice was exposed to non-energetic H-atoms at 10 K.

The propensity of molecules or atoms adsorbed on a surface to make reactions proceed readily or to enable reactions that need activation energy in the gas phase has been experimentally and theoretically demonstrated for various reaction schemes. Surfaces are able to lower energy barriers,²³ weaken molecular bonds,²⁴ and introduce some steric effects that stimulate some preferential atomic or molecular adsorption orientation.²⁵ Specifically, we demonstrated by performing a dark reaction²⁶ that the [CO + H] reaction proceeded readily with time when both reactants were isolated within an inert matrix with no energy source available for the reaction at 3 K. This experiment showed that the co-injection technique used could trap H-atoms within the solid matrix long enough to allow for diffusion and further reaction with CO molecules. In addition, Andersson, Goumans & Arnaldsson¹⁵ established, mimicking the effect of the solid state, that the effective barrier height Ea(1) dropped by about a factor 15 when tunneling was taken into account at 5 K. These results confirmed that a cold solid medium can enhance the reactivity of reactions that would not take place otherwise by trapping the reactants in the same vicinity long enough to allow for tunneling processes to occur.

Because H₂O molecules are ubiquitous on interstellar grains whereon CO hydrogenation occurs, and are a known catalyst in various heterogeneous (gas–grain) reactions (e.g., ref. 23 and references therein), several experiments investigating the influence of water-ice on CO hydrogenation have been undertaken in the past decade, both theoretically^14,16,20,27 and experimentally.^28–32 Theoretically, early studies demonstrated a moderate effect of water molecules on the barrier to addition in the gas phase.¹⁴ Calculations with up to four explicit water molecules showed no catalytic effect regarding the first step of CO hydrogenation reaction (H + CO → HCO) while only a modest effect (small decrease of the energy barrier to addition) was observed on the third step (H + H₂CO → CH₃O). Goumans^16,27 further showed that the addition of two H₂O molecules during the third reaction step could make addition (H + (H₂CO)(H₂O)₂ → CH₃O(H₂O)₂) significantly more favorable than abstraction (a competitive process in all hydrogenation experiments) in lowering the addition barrier and increasing the abstraction barrier. Thus water molecules can drive the reaction towards more saturated byproducts, probably by stabilizing both formaldehyde and the transition states through dipolar interactions.¹⁶ In addition, providing a competition between trans-HCOH and H₂CO formation upon CO hydrogenation (not specifically observed in experiments so far), a concerted hydrogen atom transfer mechanism was shown to efficiently isomerize trans-HCOH to H₂CO, with an activation energy reduced by 80% in the presence of water molecules, the latter being further transformed up to CH₃OH.²⁰

Pioneer CO surface hydrogenation experiments (i.e., bombardment of a CO and CO–H₂O mixed-ice surface with H-atoms) were performed in order to assess the effect of the ice structure,^31,33 temperature,³⁴ as well as thickness and medium composition on the formation products and reaction rates of CO hydrogenation, in a temperature range of 8–20 K. Hidaka et al.²⁸ established that [CO + D] effective rate constants for CO–H₂O mixed-ices and CO-capped H₂O-ices were larger than those for pure CO ices. The effective rate constants enhancement is thought to result from a better effective adsorption coefficient of H-atoms on H₂O molecules than on CO molecules.^26,28 While a significant H₂CO yield increase of about 5 times was observed for CO-capped H₂O-ices (i.e., when H₂O-ice is coated with a submonolayer coverage of CO molecules), only a slight H₂CO yield increase was observed on CO–H₂O mixed ices with a H₂O–CO = 4 : 1 ratio (about 1.5 times more) compared to H₂CO yields obtained for pure CO-ice hydrogenation (15 K); this observation is possibly related to the availability of CO molecules within the first few monolayers for hydrogenation.²⁸ Previously, Watanabe et al. observed comparable yields or a modest decrease of CO hydrogenation reaction product yields when a mixed H₂O–CO = 4 : 1 ice was exposed to a H-atom fluence of 10¹⁸ atoms per cm⁻² between 8 and 12 K.³¹

This paper seeks to extend the investigation of the effects of water molecules, by utilizing two techniques: [CO + H₂O + H] co-injection and (CO/H₂O) surface H-atoms irradiation, at 10 K. We showed that performing co-depositions of a cold (H/H₂) mixture and CO molecules was a means of promoting by-product formation in comparison to experiments whereby the CO-ice surface was exposed to H-atoms, and could reveal the formation of all reaction intermediates from CO to CH₃OH.^35,36 The paper is organized as follows: Section 2 illustrates the technical background and the experimental protocol suitable to this study, Section 3 presents and discusses the results of [(CO/H₂O) + H] co-injections and (CO/H₂O) irradiations by (H/H₂) at 10 K. Finally, the conclusion summarizes our discussion about the role of water molecules on CO hydrogenation at 10 K.

2. Experimental method

A detailed description of co-injection experiments can be found in previous works³⁷ Therefore, the experimental setup will be only briefly depicted in order to remind the reader of the basic scheme.

Co-injections involve condensing concurrently CO molecules or (CO + H₂O) mixtures and a mixture of hydrogen atoms and molecular hydrogen onto a cold mirror maintained at 10 K. Samples are prepared using high-purity molecular hydrogen (“Air liquide”; 99.995%), carbon monoxide (Matheson; 99.5%) and distilled water degassed in a vacuum line through several freeze–pump–thaw cycles. Atomic hydrogen is produced by a microwave driven atom source (SPECS PCR-ECS). This source uses a microwave discharge to generate gas plasma in a chamber fed with molecular hydrogen. H₂ gas is injected at 1 bar, and the pressure of the chamber during operation of the atom source is 10⁻⁵ mbar. The gas leaving the plasma chamber is a combination of both atomic and molecular hydrogen (H/H₂) with a dissociation yield of 15%. The H–H₂ mixture is expanded through five 0.2 mm orifices at 295 K into the vacuum chamber. The atomic hydrogen flux is assessed to be about 10¹⁷ atoms cm⁻² s⁻¹ and arrives onto the cold mirror thermally cooled due to numerous H–H₂ collisions while expanded through the orifices. The cold mirror is maintained at 10 K, and the temperature is kept stable by means of a pulsed tube closed-cycle cryogenerator (Cryomech PT405). The setup was evacuated at 10⁻⁷ mbar before cooling of the sample holder. Infrared spectra are recorded in the transmission–reflection mode between 5000 and 500 cm⁻¹ with a resolution of 1 cm⁻¹ using a Bruker 120 FTIR spectrometer. The FTIR spectrometer is equipped with a KBr/Ge beamsplitter and a liquid N₂-cooled narrow band HgCdTe photoconductor detector. Bare mirror backgrounds are recorded from 5000 to 500 cm⁻¹ prior to sample deposition and are used as references in processing sample spectra. Mid-infrared absorption spectra are collected on samples through a KBr window mounted on a flange with an incidence of 5 degrees with respect to the surface normal. This infrared experimental method has also been precisely described in a previous work.³⁸

[CO + (H/H₂)] and [(CO/H₂O) + (H/H₂)] co-injection experiments last for 8 min and spectra are recorded every minute. CO gas and CO–H₂O mixtures are injected with a deposition rate of 45 μmol cm⁻² min⁻¹. Accounting for this maximum dosing time, the dosing rate, and the temperature (no H₂ condensation), the upper thickness estimate assuming unity sticking probability and densities of 0.95 g cm⁻³ and 1.06 g cm⁻³ for pure α-CO and H₂O (Ihda) ices (ref. 39 and references therein), respectively, is of 61–68 μm after 8 min of co-injection, i.e. at the end of the experiment. Due to the high fluxes used in these experiments, a competition between adsorption and prompt desorption of the impinging H-atoms and CO or CO–H₂O gas mixtures may be expected,^40,41 even though some experiments suggested that the impinging H₂O molecules at similar H₂O vapor pressures do not induce the evaporation of surface H₂O molecules through a surface heating mechanism.⁴²

In order to investigate the influence of water molecules on CO successive hydrogenation at 10 K, four different mixtures are sampled with varying water concentrations. The first experiment is called the reference experiment and shows the results of pure CO gas solely co-injected with (H/H₂). The relative ratios between reactants before deposition are roughly (CO/H/H₂) = (4/1/3), the limited reactant being H-atoms. The second experiment consists in creating a CO dominant mixture by replacing only 20% of CO molecules by water molecules in the initial gas mixture, (CO/H₂O) = (80/20). The third and fourth experiments contain an excess of water with (CO/H₂O) ratios of about (10/90) and (3/97), respectively before deposition. For these two experiments the CO/H ratio is decreased by a factor 2.5 and 8.3, respectively, compared to the previous environments where CO is dominant. We can get an estimate of the ice film composition from a condensation kinetic law (Dominé, Thibert & Silvente 1996)⁴³ that can determine the proportion of CO in the deposited mixed-ice accounting for the gas' nature before deposition (eqn. (2)):

where P(i) is the partial pressure of the component i in the reservoir before deposition, α(i) is the sticking coefficient, and M(i) is the molar mass. Because of the high deposition rate and the thickness of our ice, we can use the lower limit to the sticking probability of CO, i.e. α(i) = 0.87, calculated by Bisschop et al. (2006)⁴⁴ at 14 K. We assume H₂O's sticking probability to be 1 at 10 K. We calculate approximately the ice composition issued from the three binary gas mixtures (CO/H₂O)_gas = (80/20), (10/90), and (3/97), to be (CO/H₂O)_ice = (84/16), (13/87), and (4/96); respectively. For sake of comparison, the CO hydrogenation reaction was also performed using a starting (CO/N₂) mixture in order to form a mixed ice with a (CO–N₂) ratio of about (10/90). The three different (CO/H₂O) ice samples obtained after 8 minutes of co-injection at 10 K are subsequently subjected to (H/H₂) irradiation during 15 minutes and monitored by the FTIR spectrometer.

In order to compare H₂CO and CH₃OH yields in a quantitative way, their respective abundances have been calculated by integrating their IR bands and dividing them by the appropriate absolute band strength, A. The molecular abundances resulting from this calculation are expressed as column densities with units of molecules per cm². Their column densities have been subsequently normalized by that of CO. The former calculations are performed using two different set of absolute band strengths depending on the ices' composition. Regarding the pure CO environment, the following values were used: (a) CO stretching mode of CO at 2136 cm⁻¹, A = 1.1 × 10⁻¹⁷ cm per molecule,⁴⁵ (b) ν₃ mode of H₂CO at 1500 cm⁻¹, A = 3.9 × 10⁻¹⁸ cm per molecule,⁴⁶ and (c) CO stretching mode of CH₃OH at 1034 cm⁻¹, A = 1.8 × 10⁻¹⁷ cm per molecule.⁴⁷ Regarding the environments involving mixed (CO/H₂O) ices, the following values were used: (a) CO stretching mode of CO at 2136 cm⁻¹, A = 1.1 × 10⁻¹⁷ cm per molecule,⁴⁸ (b) ν₃ mode of H₂CO at 1500 cm⁻¹, A = 4 × 10⁻¹⁸ cm per molecule,⁴⁶ and (c) CO stretching mode of CH₃OH at 1034 cm⁻¹, A = 1.6 × 10⁻¹⁷ cm per molecule.⁴⁹ As in previous works we assumed that H₂CO's absolute band strength at about 1723 cm⁻¹ was the same within a pure H₂CO, CO and N₂ environment.^46,50 The integrated intensity of CO's absorption bands were calculated directly within Beer–Lambert law's application range, typically before absolute intensity reaches 1 and when the isotopic relative ratios are respected as in the early minutes of experiments involving the CO : H₂O ratio of 84 : 16 or the H₂O-rich experiments as the one presented in Fig. 3a. Otherwise, the integrated intensity of CO's absorption band that corresponds to the transition (1 ← 0) was estimated using CO's first overtone (2 ← 0) observed at 4252 cm⁻¹, which is always far below saturation. The two integrated areas increase linearly within Beer–Lambert law's application range. When the transition (1 ← 0) saturates, a linear increase is still observed for the transition (2 ← 0) and the average ratio obtained within Beer–Lambert law's application range can be used to evaluate the integrated intensity of CO's absorption band at 2136 cm⁻¹. We calculated that the ratio between the integrated absorption intensity of the transition (1 ← 0) and that of the transition (2 ← 0), i.e., S(1 ← 0)/S(2 ← 0), equals 150 at 10 K within our CO–H mixed ice. Multiple identical experiments were performed with the same experimental conditions and give a standard deviation of about 7% of this former value. Therefore, when CO's absorption bands (2136 cm⁻¹ or isotopologues) are no longer available for integration, typically for a CO pure environment, CO's first overtone is used to give an estimation of CO's integrated absorption intensity at 2136 cm⁻¹. The resulting column density takes into account the absolute band strength relative to the considered environment.

3. Results and discussion

[CO + H + H₂] co-injection at 10 K: reference experiment

The infrared spectra representing each minute of [CO + H + H₂] co-injection at 10 K are displayed on Fig. 1a and b, showing the 4500–2000 cm⁻¹ region and the 1900–950 cm⁻¹ spectral region, respectively. Assignments of the enlarged CO region in Fig. 1a correspond to the resonant V–V transfer of vibrational energy induced by multipolar long range interactions in solid CO for Θ at 2143 cm⁻¹,^7,51 CO molecules at 2136 cm⁻¹, the vibrations of CO molecules that are chemisorbed to the substrate for ε at 2112 cm⁻¹,^{52 13}C¹⁶O at 2092 cm⁻¹, and ¹²C¹⁸O at 2088 cm⁻¹. In Fig. 1b, the infrared signatures of H₂CO (at 1733 and 1500 cm⁻¹) emerge after 1 minute of [CO + H + H₂] co-injection. H₂CO is then the single by-product in reaction (1) characterizable in the spectrum. We observe a linear increase of H₂CO's infrared absorption bands with reaction time as well as the appearance of the transient radical HCO (at 1860 cm⁻¹) at the 2^nd minute of co-injection (Fig. 1b). Furthermore, methanol, the fourth by-product of CO hydrogenation, materializes only at the 7^th minute of [CO + H + H₂] co-injection. The IR intensity of methanol is even less intense than that of HCO and its absorption band feature is quite broad. One can notice that IR patterns of water monomers (at 3707 and 1601 cm⁻¹) are seen in the spectra of Fig. 1a and b. These signatures reveal that water is present in trace quantities due to the high vacuum limitation within the reaction chamber. Impurities such as CH₄ and HO₂ can also be observed in the spectra. CH₄ and O₂ are trace contaminants present in CO gas, the latter forming HO₂ upon hydrogenation. In addition, while CO hydrogenation may also lead to the formation of glycolaldehyde through dimerization of HCO radicals,⁵³ glycolaldehyde is not observed here. In fact, because HCO concentration is so low and spread out within CO ice, the encounter probability between 2 HCO radicals is too low to consider dimerization.

Fig. 1 (a) IR spectra in the 4500–2000 cm⁻¹ region representing [CO + H + H₂] co-injections at 10 K as a function of co-deposition time, from 1 to 8 minutes. (CO)_n represents the combination between the fundamental and the lattice vibration band. (b) IR spectra in the 1900–950 cm⁻¹ region representing [CO + H + H₂] co-injections at 10 K as a function of co-deposition time, from 1 to 8 minutes. Spectra are offset for better visibility.

The trapping probability of H-atoms within the CO bulk is enhanced by the co-injection technique which produces a corrugated amorphous bulk ice with plenty of micro and macro pores as efficient trapping/reactive sites. Because a reasonable amount of CO molecules continuous flow sticks onto the surface, the bottom layers are instantaneously capped by new molecules. Thus, H-atoms desorption is limited because they are “forced” to stay/penetrate on/into the ice that is instantly covered and exhibits a bulk-character almost instantaneously. In addition, the dense ice should dissipate efficiently the heat of condensation. Therefore, the use of the co-injection technique allows to build-up reaction products on a transient surface that are rapidly isolated from the warm vacuum interface. Moreover, CO's desorption rate should be limited due to the intrinsic characteristic of the amorphous ice in which lower desorption rates are observed compared to crystalline ices.³³ In pure CO ice, diffusion within the solid is rather limited, as indicated in Fig. 1, since HCO has not been entirely converted into more saturated products. In addition, because H-atoms react with CO (the reactive medium), H–H recombination reaction are limited compared to experiments conducted within an inert medium or when the ratio CO/H becomes lower than 1.

[CO + H₂O + H + H₂] co-injections at 10 K: hydrated experiment

In order to highlight water's influence on the [H + CO] reaction, new samples were formed under the same experimental conditions by replacing CO molecules by water molecules. The following ice mixtures were then obtained (CO/H₂O)_ice = (100/0, 84/16, 13/87, 4/96). As shown in Fig. 2, the progressive replacement of CO molecules by H₂O molecules (16%) transform the IR signature for CO that now encompasses 5 features located at about 2150, 2148, 2145, 2138, and 2133 cm⁻¹ (Fig. 2 inset). CO molecules are now surrounded by different environments having a polar (≥2138 cm⁻¹) or a non-polar (<2138 cm⁻¹) character, depending on whether CO molecules are in the vicinity of H₂O molecules or not, respectively, and where the peak at 2145 cm⁻¹ is assigned to CO molecules possibly located within the pores of the amorphous water ice.⁵⁴ In addition, Fig. 2 shows that even when CO is still dominant compared to H-atoms, the addition of water molecules progressively enhances the integrated areas of the by-products of [H + CO], to wit: H₂CO and CH₃OH's IR absorption signatures are now more pronounced. Water co-deposition likely creates amorphous ices with many defects, traps and pores and disruption of long range order in the matrix. Within this environment, H-atoms are more likely to diffuse deeper in the ice and hydrogenate bulk CO molecules that would be protected in a pure CO environment. In addition, H-atoms can be trapped in the numerous potential wells available and actively react with the first reaction partner, either CO or H-atoms. H-atoms recombination leads to H₂ formation which can also remain trapped within enclosed macropores (or pockets within the ice) up to about 100 K.⁵⁵ Moreover, each step of the sequential process of H-atoms addition onto CO competes with the reverse processes of H-atoms abstraction. It is very likely that when the reaction [CO + H] takes place in a pure CO environment, the H-atom addition mechanism greatly competes with the H-atom abstraction mechanism and therefore inhibits the formation of by-products. The increasing number of water molecules could progressively change this trend and favor the addition over the abstraction processes as shown by the disappearance of HCO and the increase of H₂CO and CH₃OH in a water-rich environment. This phenomenon has been demonstrated by Goumans¹⁶ with respect to the second step of the reaction process (H + (H₂CO)(H₂O)₂ → H₃CO + (H₂O)₂), whereby the addition of two water molecules to H₂CO had a combined effect of lowering of the hydrogen addition barrier and increasing the abstraction barrier, making addition more favorable than abstraction for the reaction in the condensed phase, regardless of the temperature.

Fig. 2 IR spectra in the 4500–520 cm⁻¹ region representing (i) [CO + H + H₂] and (ii) [CO + H₂O + H + H₂] co-injections at 10 K after 8 minutes. CO–H₂O proportions in the mixed ice are about 4 : 1. Enlargements of the formaldehyde and methanol regions are also presented. Spectra are offset for better visibility.

When water molecules become by far the main component of the (CO/H₂O) binary mixture co-injected with (H/H₂) at 10 K, the 2ν1 mode of CO (initially at 4252 cm⁻¹) is now red-shifted by 93 cm⁻¹ due to the ubiquitous presence of water molecules (Fig. 3). The water region consists now of a broad band characteristic of amorphous water-ice (Fig. 3a). The stretching region of CO molecules is now divided up into two sharp peaks corresponding to (i) substitutional CO molecules in the amorphous water ice structure or CO molecules interacting with the water's oxygen atoms (2153 cm⁻¹), and (ii) interstitial CO molecules or CO interacting with water's OH dangling groups (2138 cm⁻¹).^45,54,56,57 The most noticeable changes on CO hydrogenation by-product formation occur in the 1900–950 cm⁻¹ region, Fig. 3b. No signal corresponding to HCO is observed in the experiments where water has been added to the mixtures. In contrast, despite the broad absorption band dominating the spectrum around 1650 cm⁻¹ consistent with the IR signature of water-ice, the CH₂ scissoring mode of H₂CO at 1500 cm⁻¹ clearly stands out. While only 13% of CO molecules make up the (CO/H₂O) mixture, the integrated area of H₂CO (1500 cm⁻¹) absorption band is about 6 times greater than in the reference experiment, namely [CO + H + H₂] co-injection. It is worth noticing that the CO stretching mode of H₂CO is significantly overlapped by water-ice. Indeed, only a shoulder feature marks its presence at 1723 cm⁻¹. This absorption band is therefore not exploitable for integration in order to estimate the abundance of H₂CO. Equally important, the IR absorption band of methanol is clearly visible and its integrated peak area at the 8^th minute is about 3.5 times greater than in the reference experiment. However, because in this experiment CO concentration is so small, H-atoms are now dominant in contrast to the first two experiments in which CO was prevailing. Thus, while H₂O molecules may catalyze the reaction, the rise in product yields is enhanced by the smaller CO/H ratio. Therefore, a similar experiment, with the same smaller CO/H ratio was performed with N₂ molecules replacing H₂O molecules (Fig. 4) and showed no increase in H₂CO's production yield in contrast to when H₂O molecules are present after 8 minutes of co-injection for both experiments. The influence of water molecules is highlighted when the yields [H₂CO]/[CO] and [CH₃OH]/[CO] are displayed versus CO proportions within the (CO/H₂O) mixed ices on Fig. 5a and b, respectively. In Fig. 5a and b, Y and X-axes have been cut off for clarity. Fig. 5a shows the increase of H₂CO production with the decrease of CO molecules within the (CO/H₂O) binary mixture co-injected with H–H₂. Indeed, the replacement of only 20% of CO by water molecules (86% CO in the ice) amplifies by almost 2 times H₂CO production, even though the reaction is still limited by the amount of H-atoms. The trend is more remarkable when the (CO/H₂O) mixture reaches a ratio of about (13/87). While CO proportion in the icy mixture is now divided by 7, the amount of H₂CO produced is increased by 118 times. The experiment involving a (CO/N₂) ratio of (10/90) confirmed that the increase of the H₂CO yield is not solely due to the smaller (CO/H) ratio but that in fact H₂O molecules play a role in the rise of CO hydrogenation formation products. Accordingly, the tendency continues even when the amount of CO molecules comes closer to trace quantities. When the (CO/H₂O) mixture reaches a ratio of about 4/96, i.e., when CO proportion is reduced by about 3 times in comparison to the previous mixture, the amount of H₂CO keeps rising until reaching about 2 times the previous quantity. However, Fig. 3b demonstrates that CH₃OH formation does not obey the same rules. Whereas an important increase of the [CH₃OH]/[CO] ratio is observed as the percentage of CO lessens significantly from 86% to 13%, the trend is stopped when CO proportion is again reduced by 3 times. At the 4/96 ratio (CO/H₂O), the [CH₃OH]/[CO] ratio is divided by 3 times as well. This effect demonstrates the limitation of the hydrogenation when there is a low abundance of the primary reagent, i.e. CO molecules. When the proportion of CO molecules becomes too low, the side reaction [H + H] recombination is significantly favored over CO successive hydrogenation. The present experiments suggest that a greater number of H-atoms remain available for the reaction [CO + H] in the presence of water molecules compared to the inert N₂ environment.

Fig. 3 (a) IR spectra in the 4500–2000 cm⁻¹ region representing [CO + H₂O + H + H₂] co-injections at 10 K as a function of co-deposition time, from 1 to 8 minutes. CO and H₂O proportions in the ice are: (CO/H₂O) = (13/87). (b) IR spectra in the 1900–950 cm⁻¹ region representing [CO + H₂O + H + H₂] co-injections at 10 K as a function of co-deposition time, from 1 to 8 minutes. CO and H₂O proportions in the ice are: (CO/H₂O) = (13/87). Spectra are offset for better visibility.

Fig. 4 IR spectra in (a) the 44 260–4230 cm⁻¹, and (b) the CO hydrogenation products regions representing both (i) [CO + N₂ + H + H₂] and (ii) [CO + H + H₂] co-injections at 10 K after 8 minutes. While CO is significantly reduced, H₂CO and CH₃OH integrated intensity remain lower than those obtained in the pure CO environment, in contrast to what is observed with the presence of H₂O molecules.

Fig. 5 Column density of (a) H₂CO and (b) CH₃OH normalized by that of CO after 8 minutes of [CO + H₂O + H + H₂] co-injection at 10 K as a function of the percentage of CO within the (CO/H₂O) mixed-ice.

Surface irradiations by H–H₂

The former mixed-ice surfaces previously monitored are finally exposed to the (H/H₂) beam for 15 minutes at 10 K. Difference spectra of the CO stretching region are depicted in Fig. 6. Negative and positive peaks correspond to a decrease and increase in absorbance, respectively, after 15 minutes of H-atoms exposure. Spectra b–d evidence the efficiency of H₂CO formation on surfaces containing H₂O molecules. A small increase of H₂CO's absorption bands (1723 cm⁻¹, 1500 cm⁻¹) is observed on spectrum a, when an initially CO-ice devoid of water is subjected to the H–H₂ beam. However, the water-rich ice instantly increases the H₂CO concentration as suggested by the intense IR signatures that arise after 15 min of H-atoms exposure. In parallel with the growth of H₂CO, a depletion of CO, (CO)_n, and H₂O is observed in Fig. 4b. Likewise, in water-rich environments, CO(H₂O)_n, and (H₂O)_n depletions are observed in Fig. 4c and d. The depletion of the IR signature of water molecules/water ice is observed both in the OH stretching region (3630 cm⁻¹) and in the HOH bending region (1647 cm⁻¹). Beyond the growth of H₂CO, Spectra c and d also confirm the significant formation of methanol despite the small CO amount in the ice. In addition, one can notice the consistent IR intensity of CO₂ ν3 mode at 2344 cm⁻¹ in Fig. 4a and b, and in the 13/97 water-rich environment, spectrum c. However, likewise the by-product of CO hydrogenation, the IR absorption band of CO₂ lessens when the CO proportion is as low as 4%, spectrum d.

Fig. 6 IR difference spectra in the 2400–950 cm⁻¹ region after 15 min of surface irradiation by H–H₂ at 10 K. Surfaces correspond to those obtained after 8 minutes of [(CO/H₂O) + H + H₂] co-injections at 10 K (a) CO ice, (b) (CO/H₂O) = (84/16), (c) (CO/H₂O) = (13/87), (d) (CO/H₂O) = (4/96). Spectra are offset for better clarity.

Fig. 6 shows the significant growth of the two main products H₂CO and CH₃OH while desorption or delocalization of water molecules/water-ice from the sample region probed occurs. This phenomenon has been noted previously in the OH stretching region by Watanabe, Shiraki & Kouchi.³² The authors suggested that the depletion of the IR signature of water molecules was likely due to a change in the interaction between H₂O and CO molecules. In other words, the decrease observed in the OH stretching region (very sensitive to the environment) would be due to the consumption of a CO molecule localized in close vicinity of a water molecule. We propose that the negative IR signature of water upon hydrogen exposure reflects additionally water desorption – either while fragmenting or intact – or random hydrogen network reorganization after hydrogen distribution. The first possibility is that water molecules fragment upon H-atom exposure and produce OH radicals and H₂ molecules. However, H-atoms impinge the sample with a very low energy and are instantly thermalized on the surface. Thus, H-atoms do not possess sufficient energy to dissociate water molecules. In addition, OH radicals are known to react with CO molecules to produce CO₂.^37,58 Therefore, because the IR signature of CO₂ is consistent in every environment, from the “CO pure” one to the (CO/H₂O) = (13/87) one, and a bit smaller in the latest environment in which water molecules are by far the more predominant species (CO/H₂O) = (04/96), we know that water molecules are not fragmented upon H-atoms exposure and that CO₂ is not formed through OH radicals but is an impurity in our reaction chamber. The second assumption is that water molecules/water-ice could be desorbing without fragmentation upon H-atom exposure. However, the decrease observed in water's IR signature in the OH stretching region by Watanabe, Shiraki & Kouchi³² was under experimental conditions involving the use of a cooling device coupled to the microwave source from which stem the hydrogen atoms. This device allows H-atoms to impinge the cold surface with an energy equivalent to 0.8 kJ mol⁻¹. This energy is not sufficient to fragment H₂O molecules. The last hypothesis that could explain water desorption or delocalization from the spectral region probed with the infrared beam is related to a mechanism that would assist the reaction of CO hydrogenation though the formation of H⋯(H₂O)_n complexes. The H-atoms engaged in this type of bonding are extremely reactive and would release significant energy when encountering a reaction partner (H, CO, or a by-product previously formed).⁵⁹ Buch & Czerminski⁶⁰ and Devlin & Buch⁶¹ showed that the binding energy is strongly governed by the topological distortion of the surface potential. The ice studied presently has been formed using a co-injection technique at 10 K and is therefore very porous and distorted (presence of many defects), exhibiting probably various binding sites with different potential well depths available for H-atoms. Studies of [H + D] recombination reactions⁶² showed that the greatest part of the energy released during recombination was in equilibrium with the surface, i.e., released to the surrounding ice before desorption or delocalization.⁶³ Accordingly, the energy released after hydrogen distribution to another reaction partner (CO or H) may be sufficiently high to displace water molecules present in the vicinity of the event. The water displacement expressed by the depletion of the IR signal in the HOH bending region could subsequently induce a local and random restructuration of the hydrogen bonding network and favor a water-mediated hydrogen transfer toward the first few layers of the bulk. Consequently, more hydrogen would be available for the hydrogenation process. The mechanism depicted here would be a “cooperative effect” of the hydrogen bonding upon exposure to hydrogen in surface experiments and hydrogen and CO molecules in co-injection experiments. This mechanism would be worth investigating using molecular dynamics simulations. The results observed in our study suggest that when CO is surrounded by water molecules in cold dense clouds (in polar ices or hydrogen-rich ices dominated by H₂O),⁶⁴ CH₃OH formation via CO hydrogenation could be enhanced, depending on the H/CO/H₂O ratio.

Conclusion

We performed several co-injection experiments varying the concentration of water molecules within the binary mixture (CO/H₂O) co-condensed with (H/H₂) onto a cold mirror maintained at 10 K. We showed that the progressive replacement of CO molecules by water molecules greatly enhances the yields of H₂CO and CH₃OH, the main products of CO hydrogenation at 10 K. This trend is observed until the concentration in CO molecules becomes too low and significantly favors the side reaction [H + H] recombination over CO hydrogenation. We confirmed the highest product yield observed in the presence of water as previously reported by other groups. The catalytic effect of water molecules observed in this work can be explained by physical, chemical, and energetic considerations: (1) water molecules could distort the original CO ice and create bigger voids allowing H-atoms to reach for deeper buried reaction partners, (2) water molecules could form a hydrogen-bonded network that would be able to transfer hydrogen atoms towards the bulk through cooperative effects, (3) water molecules favor the addition process over the abstraction process for the second step of CO hydrogenation, supporting CH₃OH production. Based on the catalytic effect observed, we performed an experiment to express the role of water molecules on the reaction. We exposed the ice obtained through the co-injections experiments to the [H + H₂] beam. While the experiment confirmed the efficient growth of H₂CO and CH₃OH in the ices containing water molecules, it also showed a simultaneous consumption of CO molecules with desorption or delocalization of water molecules. This result suggests that the action of water molecules induces either a local desorption without involving significant co-desorption of CO molecules or surrounding by-products or results in a reorganization of the hydrogen bonding network after hydrogen distribution to another reaction partner.

References

1. E. Herbst and E. F. van Dishoeck, Annu. Rev. Astron. Astrophys., 2009, 47, 427–480.
2. V. F. Petrenko and R. W. Whitworth, Physics of ice, Oxford University Press, 1999.
3. G. Consolmagno and J. Lewis, IAU Colloq. 30: Jupiter: Studies of the Interior, Atmosphere, Magnetosphere and Satellites, 1976.
4. D. A. Rothery, Satellites of the outer planets: worlds in their own right, Oxford University Press, 1999.
5. S. A. Sandford and L. J. Allamandola, Icarus, 1988, 76, 201–224.
6. M. S. Westley, R. A. Baragiola, R. E. Johnson and G. A. Baratta, Nature, 1995, 373, 405–407.
7. G. Zumofen, International Conference on Matrix Isolation Spectroscopy, West-Berlin, Germany, 1977.
8. K. I. Öberg, A. C. A. Boogert, K. M. Pontoppidan, S. van den Broek, E. F. van Dishoeck, S. Bottinelli, G. A. Blake and N. J. Evans II, Astrophys. J., 2011, 740, 109.
9. O. Biham, I. Furman, V. Pirronello and G. Vidali, Astrophys. J., 2001, 553, 595.
10. T. P. M. Goumans, A. Wander, C. R. A. Catlow and W. A. Brown, Mon. Not. R. Astron. Soc., 2007, 382, 1829–1832.
11. A. G. G. M. Tielens and W. Hagen, Astron. Astrophys., 1982, 114, 245–260.
12. A. G. G. M. Tielens, IAU Symp. 135: Interstellar Dust, Kluwer, Dordrecht, 1989.
13. D. E. Woon, J. Chem. Phys., 1996, 105, 9921–9926.
14. D. E. Woon, Astrophys. J., 2002, 569, 541.
15. S. Andersson, T. Goumans and A. Arnaldsson, Chem. Phys. Lett., 2011, 513, 31–36.
16. T. P. M. Goumans, Mon. Not. R. Astron. Soc., 2011, 413, 2615–2620.
17. T. P. M. Goumans, C. R. A. Catlow and W. A. Brown, J. Chem. Phys., 2008, 128, 1347091–1347096.
18. T. G. Robin, L. W. W. Susanna and H. Eric, Astrophys. J., 2008, 682, 283.
19. O. Yoshihiro, R. Helen and H. Eric, Astrophys. J., 2005, 621, 348.
20. P. S. Peters, D. Duflot, A. Faure, C. Kahane, C. Ceccarelli, L. Wiesenfeld and C. l. Toubin, J. Phys. Chem. A, 2011, 115, 8983–8989.
21. N. Watanabe and A. Kouchi, Astrophys. J., Lett., 2002, 571, L173.
22. G. W. Fuchs, H. M. Cuppen, S. Ioppolo, C. Romanzin, S. E. Bisschop, S. Andersson, E. F. van Dishoeck and H. Linnartz, Astron. Astrophys., 2009, 505, 629–639.
23. H.-B. Xie, Y.-H. Din and C.-C. Sun, Astrophys. J., 2006, 643, 573.
24. T. P. M. Goumans, M. A. Uppal and W. A. Brown, Mon. Not. R. Astron. Soc., 2008, 384, 1158–1164.
25. S. M. Madzunkov, J. A. MacAskill, A. Chutjian, P. Ehrenfreund, M. R. Darrach, G. Vidali and B. J. Shortt, Astrophys. J., 2009, 697, 801.
26. C. Pirim and L. Krim, Phys. Chem. Chem. Phys., 2011, 13, 19454–19459.
27. T. P. M. Goumans, Mon. Not. R. Astron. Soc., 2012, 423, 3775–3775.
28. H. Hidaka, A. Kouchi and N. Watanabe, J. Chem. Phys., 2007, 126, 2047071–2047079.
29. N. Watanabe and A. Kouchi, Prog. Surf. Sci., 2008, 83, 439–489.
30. N. Watanabe, O. Mouri, A. Nagaoka, T. Chigai, A. Kouchi and V. Pirronello, Astrophys. J., 2007, 668, 1001.
31. N. Watanabe, A. Nagaoka, T. Shiraki and A. Kouchi, Astrophys. J., 2004, 616, 638.
32. N. Watanabe, T. Shiraki and A. Kouchi, Astrophys. J., Lett., 2003, 588, L121.
33. H. Hidaka, N. Miyauchi, A. Kouchi and N. Watanabe, Chem. Phys. Lett., 2008, 456, 36–40.
34. H. Hidaka, N. Watanabe, T. Shiraki, A. Nagaoka and A. Kouchi, Astrophys. J., 2004, 614, 1124.
35. C. Pirim and L. Krim, Chem. Phys., 2011, 380, 67–76.
36. C. Pirim, L. Krim, C. Laffon, P. Parent, F. Pauzat, J. Pilmé and Y. Ellinger, J. Phys. Chem. A, 2010, 114, 3320–3328.
37. E.-L. Zins, P. R. Joshi and L. Krim, Astrophys. J., 2011, 738, 175.
38. L. Manceron, B. Tremblay and M. E. Alikhani, J. Phys. Chem. A, 2000, 104, 3750–3758.
39. M. P. Collings, J. W. Dever, H. J. Fraser, M. R. S. McCoustra and D. A. Williams, Astrophys. J., 2003, 583, 1058.
40. A. Al-Halabi, H. Fraser, G. Kroes and E. Van Dishoeck, Astron. Astrophys., 2004, 422, 777–791.
41. A. Al-Halabi, E. Van Dishoeck and G. Kroes, J. Chem. Phys., 2004, 120, 3358–3367.
42. D. R. Haynes, N. J. Tro and S. M. George, J. Phys. Chem., 1992, 96, 8502–8509.
43. F. Dominé and E. Thibert, Geophys. Res. Lett., 1996, 23, 3627–3630.
44. S. E. Bisschop, H. J. Fraser, K. I. Öberg, E. F. van Dishoeck and S. Schlemmer, Astron. Astrophys., 2006, 449, 1297–1309.
45. M. E. Palumbo, J. Phys. Chem. A, 1997, 101, 4298–4301.
46. W. A. Schutte, L. J. Allamandola and S. A. Sandford, Icarus, 1993, 104, 118–137.
47. D. Hudgins, S. Sandford, L. Allamandola and A. Tielens, Astrophys. J., Suppl. Ser., 1993, 86, 713–870.
48. P. A. Gerakines, W. A. Schutte, J. M. Greenberg and E. F. van Dishoeck, Astron. Astrophys., 1995, 296, 810.
49. O. Kerkhof, W. A. Schutte and P. Ehrenfreund, Astron. Astrophys., 1999, 346, 990.
50. R. B. Bohn, S. A. Sandford, L. J. Allamandola and D. P. Cruikshank, Icarus, 1994, 111, 151–173.
51. H. Dubost, International Conference on Matrix Isolation Spectroscopy, West-Berlin, Germany, 1977.
52. C. S. Jamieson, A. M. Mebel and R. I. Kaiser, Astrophys. J., Suppl. Ser., 2006, 163, 184.
53. P. M. Woods, B. Slater, Z. Raza, S. Viti, W. A. Brown and D. J. Burke, Astrophys. J., 2013, 777, 90.
54. S. Sandford, L. Allamandola, A. Tielens and G. Valero, Astrophys. J., 1988, 329, 498–510.
55. G. A. Grieves and T. M. Orlando, Surf. Sci., 2005, 593, 180–186.
56. J. P. Devlin, J. Phys. Chem., 1992, 96, 6185–6188.
57. P. Jenniskens, D. Blake, M. Wilson and A. Pohorille, Astrophys. J., 1995, 455, 389.
58. Y. Oba, N. Watanabe, A. Kouchi, T. Hama and V. Pirronello, Astrophys. J., Lett., 2010, 712, L174.
59. S. K. Chulkov, N. F. Stepanov and Y. V. Novakovskaya, Russ. J. Phys. Chem., 2009, 83, 798–808.
60. V. Buch and R. Czerminski, J. Chem. Phys., 1991, 95, 6026.
61. J. P. Devlin and V. Buch, J. Phys. Chem., 1995, 99, 16534–16548.
62. L. Hornekær, A. Baurichter, V. V. Petrunin, D. Field and A. C. Luntz, Science, 2003, 302, 1943–1946.
63. E. Congiu, E. Matar, L. E. Kristensen, F. Dulieu and J. L. Lemaire, Mon. Not. R. Astron. Soc.: Lett., 2009, 397, L96–L100.
64. P. Ehrenfreund, E. Dartois, K. Demyk and L. d'Hendecourt, Astron. Astrophys., 1998, 339, L17–L20.

---