© ESO, 2014

1. Introduction

Grain surface chemistry is governed by the local H/H₂ ratio. If this ratio is high, hydrogenation is the main process, H₂O is the dominant constituent, and species such as NH₃ and CH₄ are expected to form. On the other hand, if the H/H₂ ratio is substantially lower than unity, CO molecules will be abundant and species such as O₂ and N₂ are easily formed. Therefore, two types of ice mantles can be distinguished, one dominated by polar molecules and another one dominated by nonpolar, or only slightly polar, molecules (see Tielens et al. 1991; Chiar et al. 1995; Gerakines et al. 1996; Whittet et al. 1996). The relative to water abundances of CO₂ and CH₄ are between 4−44% and 0.4−8%, respectively, for the different interstellar environments (Mumma & Charnley 2011, and references therein).

Gas-phase molecular oxygen and nitrogen lack an electric dipole moment, which makes them infrared inactive. These species were therefore not detected in ice mantles and cannot be easily observed in the gas phase. In some clouds the estimated abundance of gaseous N₂ is on the order of 10^-5 with respect to H₂ (Bergin et al. 2002), whereas the coldest and densest cores show a decrease in the gas-phase N₂ by at least a factor of two (Belloche & André 2003). O₂ has a ³Σ ground state and thus a magnetic dipole moment enabling weak transitions in the submillimeter range between finestructure levels. Very low detection rates of O₂ were found toward Orion (Goldsmith et al. 2011) and the ρ-Oph A core (Liseau et al. 2012) with HIFI. The upper limits on O₂ and estimates of N₂ suggest relative abundances much lower than expected from the solar oxygen and nitrogen abundances. Large amounts of oxygen and nitrogen are therefore apparently missing from the gas phase and might be depleted on grains most likely in the form of O₂ and N₂ (Ehrenfreund & van Dishoeck 1998).

Synchrotron radiation has been used as a source of VUV photons to perform VUV-absorption spectroscopy of ice samples. The National Synchrotron Radiation Research Center (NSRRC) in Taiwan (Lu et al. 2005; Cheng et al. 2011; and Wu et al. 2012), ASTRID at the University of Aarhus, and the UK Daresbury Synchrotron Radiation Source (Mason et al. 2006) were used as VUV sources in their measurements. This requires application for synchrotron beamtime, leading to a limited use, and one needs to work with a transportable chamber. The use of a microwave-discharge H₂ flow UV-lamp as the source for VUV-absorption spectroscopy, employed in various astrochemistry laboratories, allows a routine performance of VUV-absorption spectroscopy in the 120−170 nm spectral range. The polar molecular components of interstellar ice were studied in Cruz-Diaz et al. (2013, Paper I). To present a more complete view of the interactions between the interstellar VUV field and icy grain mantles, we report here a similar study on nonpolar ices. Section 2 describes the experimental protocol used in the experiments. Section 3 summarizes the VUV-absorption cross-section measurements of the different nonpolar ices studied: CO₂, CH₄, N₂, and O₂. Gas-phase VUV-absorption cross-section data adapted from other works were used for comparison with our solid-phase data. In addition, the spectra of the different species were fitted with Gaussian profiles, using an in-house IDL code, to provide an individual VUV-absorption cross section for each feature in the spectra. Sections 4 and 5 summarize the astrophysical implications and the conclusions.

2. Experimental protocol

The measurements were conducted using the Interstellar Astrochemistry Chamber (ISAC). This set-up and the standard experimental protocol were described in Muñoz Caro et al. 44. The specific detail on the VUV-measurements of absorption cross sections of ice are provided in Paper I. ISAC mainly consists of an ultra-high-vacuum (UHV) chamber, with pressure typically in the range of P = 3.0−4.0 × 10^-11 mbar, where an ice layer made by deposition of a gas species onto a cold finger at 8 K, achieved by means of a closed-cycle helium cryostat, can be UV-irradiated. The evolution of the solid sample was monitored with in situ transmittance FTIR spectroscopy and VUV-spectroscopy. The chemical components used for the experiments described in this paper were CO₂(gas), Praxair 99.998%; CH₄(gas), Praxair 99.999%; N₂(gas), Praxair 99.999%; and O₂(gas), Praxair 99.8%. The deposited ice layer was photoprocessed with a microwave-discharged hydrogen flow lamp (MDHL), from Opthos Instruments. The source has an UV-flux of ≈2 × 10¹⁴ cm^-2 s^-1 at the sample position, measured by CO₂ → CO actinometry, see Muñoz Caro et al. 44. The Evenson cavity of the lamp is refrigerated with air. The VUV-spectrum is measured routinely in situ during the irradiation experiments with the use of a McPherson 0.2 m focal length VUV monochromator (model 234/302) with a photomultiplier tube (PMT) detector equipped with a sodium salicylate window, optimized to operate from 100−500 nm (11.27−2.47 eV), with a resolution of 0.4 nm. The characterization of the MDHL spectrum was previously reported (Chen et al. 2010, Paper I) and will be discussed in more detail by Chen et al. 6. For more details and a scheme of the experimental set-up we refer to Paper I.

3. VUV-absorption cross section of interstellar ice analogs

VUV-absorption spectra of pure ices composed of CH₄, CO₂, N₂, and O₂ have been recorded. The column density of the ice sample was measured by FTIR in transmittance. The VUV-spectrum and the column density of the ice were therefore monitored in a single experiment for the same ice sample. This improvement allowed us to estimate the VUV-absorption cross section of the ice more accurately. The column density of the deposited ice obtained by FTIR was calculated according to the formula (1)where N is the column density of the ice, τ_ν the optical depth of the IR band, dν the wavenumber differential, in cm^-1, and is the band strength in cm molecule^-1, see Table 1. The integrated absorbance is equal to 0.43 ×τ, where τ is the integrated optical depth of the IR band. The VUV-absorption cross section was estimated according to the Beer-Lambert law (2)where I_t(λ) is the transmitted intensity for a given wavelength λ, I₀(λ) the incident intensity, σ is the cross section in cm², and N is the average ice column density before (N_i) and after (N_f) exposure to VUV during VUV-spectral acquisition, in cm^-2. We decided to take this average value of N because the ice sample thickness decreases during exposure to VUV photons. If instead of this average value, N_i would be used as N in Eq. (2), this would lead to a lower limit of the VUV-absorption cross section. It is important to notice that the total VUV-flux value emitted by the lamp does not affect the VUV-absorption spectrum of the ice sample, since it is obtained by subtraction of two spectra to obtain the absorbance in the VUV.

Several measurements for different values of the initial ice column density were performed to improve the spectroscopy. Table 1 provides the infrared-band positions and band strengths of CH₄ and CO₂ used to estimate the column density. Solid N₂ and O₂ do not display absorption features in the mid-infrared, therefore their column densities were measured using the expression (3)where N_A is the Avogadro constant (6.022 × 10²³ mol^-1), ρ_i is the density of the ice in g cm^-3, see Table 1, m_i is the molar mass of the species in g mol^-1, and d_i is the ice thickness in cm. The latter was estimated following the classical interfringe relation (4)where n_i is the refractive index of the ice at deposition temperature, and Δν is the wavenumber difference between two adjacent maxima or minima of the fringes observed in the infrared spectrum of the ice. These interference fringes are due to multiple reflections of light within the sample.

Solid O₂ was deposited at a temperature of 8 K. We used the Fulvio et al. 18 values of n_i and ρ at 16 K as an approximation. Error values for the column density in Table 1 result mainly from the selection of the baseline for integration of the IR absorption band and the decrease of the ice column density due to VUV-irradiation during spectral acquisition. The band strengths were adapted from the literature and their error estimates are no more than 10% of the reported values (Richey & Gerakines 2012). The errors in the column density determined by IR spectroscopy were 38%, 22%, and 35% for solid CH₄, CO₂, and O₂. N₂ and O₂ do not present any IR feature, but photoprocessed O₂ ice during VUV-spectral acquisition produces O₃ very readily and the loss of O₂ can be calculated. On the other hand, N₂ ice does not lead to photoproducts and the error in the column density could not be measured. The VUV-absorption cross-section errors result from the error values of the column density determination and the MDHL, photomultiplier tube (PMT), and multimeter stabilities, about 6%, using the expression (5)The VUV-absorption cross-section spectra of CH₄, CO₂, N₂, and O₂ ices were fitted with Gaussian profiles using the band positions reported in the literature (Mason et al. 2006; Lu et al. 2008; Wu et al. 2012) as a starting point, see the red dotted and dashed-dotted traces in Figs. 1, 2, 4, and 5. Table 2 summarizes the Gaussian profile parameters used to fit the spectra of these ices, deposited at 8 K. Gaussian fits of the reported molecules were made with an in-house IDL code. The fits reproduce the VUV-absorption cross-section spectra well.

Fig. 1 VUV-absorption cross section as a function of photon wavelength (bottom X-axis) and VUV-photon energy (top X-axis) of CH4 ice deposited at 8 K, black solid trace. The blue dashed trace is the VUV-absorption cross-section spectrum of gas phase CH4 adapted from Lee et al. 32. The fit of the solid-phase spectrum, red dotted (single bands, see Table 2) and dashed-dotted trace, is the sum of three Gaussians. It has been offset for clarity.

Open with DEXTER

Fig. 2 VUV-absorption cross section as a function of photon wavelength (bottom X-axis) and VUV-photon energy (top X-axis) of CO2 ice deposited at 8 K, black solid trace. The blue dashed trace is the VUV-absorption cross-section spectrum of gas phase CO2 adapted from Yoshino et al. 68. The fit, solid red dotted (single bands, see Table 2) and dashed-dotted trace, is the sum of two Gaussians. The inset figure is a CO2 VUV-absorption cross-section close-up in the 124−129 nm range.

Open with DEXTER

The main emission peaks of the MDHL occur at 121.6 nm (Lyman-α), 157.8 nm, and 160.8 nm (molecular H₂ bands). These peaks are thus also present in the secondary VUV photon spectrum generated by cosmic rays in dense interstellar clouds and circumstellar regions where molecular hydrogen is abundant (Gredel et al. 1989). For this reason, the absorption cross-section values measured at these wavelengths are provided for each molecule in the following sections. We also present an average value of the VUV-absorption cross section (a simple arithmetic mean in the spectral range indicated for each species) and an integrated VUV-absorption cross section (an integration of the form ) in the same spectral range. The same measurements were performed using the gas-phase raw data, adapted from other works, for each species.

3.1. Solid methane

The ground state of CH₄ is ¹A₁ and its bond energy is E_b(H−CH₃) = 4.5 eV (Okabe 1978).

Figure 1 shows the VUV-absorption cross section of CH₄ as a function of the wavelength and photon energy. The high absorption of CH₄ ice around and below 120 nm, and the absorption of MgF₂ in the same spectral region only allowed spectroscopic measurements starting from 120 nm instead of 113 nm. As in Wu et al. 67, a broad absorption band extending to 137 nm (9.05 eV) and centered on 124 nm (10.0 eV) was observed. This feature is attributed to the 1t₂–3s (D_2d) Rydberg transition. We detected a small bump near 140 nm (8.85 eV), which was not observed by Wu et al. 67 in the solid phase or by Lee et al. 32 in the gas phase. This feature may be produced by the VUV-absorption of a photoproduct. This phenomenon is more intense in the case of CO₂ studied in detail in Sect. 3.2. CH₄ ice presents high absorption at Ly-α (121.6 nm) and almost no absorption at Lyman band system wavelengths (132−165 nm).

The average VUV-absorption cross section has a value of cm² in the 120−150 nm (10.33−8.26 eV). The total integrated VUV-absorption cross section has a value of cm² nm ( cm² eV) in the same spectral region. The VUV-absorption cross sections of CH₄ ice at 121.6 nm is cm², that is, at the Ly-α position of atomic hydrogen and there is no observable VUV-absorption in the position of the molecular hydrogen emission bands at 157.8 and 160.8 nm. Gas-phase data from Lee et al. 32 were used for comparison with our solid-phase data, see Fig. 1, following Wu et al. 67. The solid CH₄ absorption is shifted to shorter wavelength with respect to the gas data. The VUV-absorption cross section of CH₄ in the gas phase has an average value of 8.2 × 10^-18 cm² in the 120−150 nm range. CH₄ gas data were integrated in the same range, giving a value of 2.9 × 10^-16 cm² nm (2.09 × 10^-17 cm² eV), which is higher than the solid CH₄ value. The VUV-absorption cross section of CH₄ gas at 121.6 nm is 1.8 × 10^-17 cm².

3.2. Solid carbon dioxide

The ground state and bond energy of CO₂ are X and E_b(O−CO) = 5.5 eV (Okabe 1978).

The VUV-absorption cross section of CO₂ as a function of the wavelength and photon energy is shown in Fig. 2. This spectrum is very different from those reported in Lu et al. 35, Mason et al. 38, and Monahan & Walker 41. Mason et al. 38 observed two broad bands centered on 8.8 and 9.9 eV assigned to the and transitions. Lu et al. 35 and Monahan & Walker 41 reported a broad band centered on 9.8 eV. These authors found a vibrational structure in this band with similar positions. The intense absorption feature in the 107−120 nm (11.53−10.33 eV) region reported by Lu et al. 35 is partially observed in our data. Two broad bands in the 120−133 nm (10.33−9.32 eV) and 133−163 nm (9.32−7.60 eV) regions were also reported by Lu et al. 35, which are observed in our spectrum as well.

	Fig. 3 Comparison between the VUV-absorption cross-section spectra of CO (red dashed-dotted trace) and CO2 (black solid trace, the CO2 ice is mixed with CO formed during the spectral acquisition). CO transitions in a CO2 matrix are shifted to shorter wavelengths with respect to pure CO ice transitions.
Open with DEXTER

A vibrational structure was detected in the 120−133 nm band, see Fig. 2 inlet. This vibrational structure is very faint in our spectrum, but 5 out of the 12 bands reported by Mason et al. 38 are observable, see Table 3. In the 133−167 nm wavelength range of Fig. 2, the spectrum presents a CO-like VUV-absorption. Indeed, the VUV-light cone of our MDHL source processes the entire volume of the CO₂ ice and efficiently leads to the formation of CO by photodissociation of CO₂ molecules. After 9 min, corresponding to the collection time of a spectrum, CO is present with a column density that is 22% of the deposited CO₂ column density, estimated from integration of the IR bands, data not shown. This is enough, for a molecule like CO with a VUV-absorption cross section seven times larger than CO₂, to appear in the VUV-spectrum of CO₂ ice. The measured VUV-spectrum therefore corresponds to a mixture of CO₂ and CO, but because of the interaction of CO molecules with the CO₂ ice matrix, the CO features are shifted to shorter wavelengths than those of pure CO ice reported, for instance, in Paper I. This is shown in Fig. 3. Table 4 summarizes the peak positions of the detected CO features in the CO₂ ice and those reported for pure CO gas and ice. The Δλ- column in Table 4 represents a redshift with respect to pure CO gas and Δλ+ represents a blueshift with respect to the pure CO ice. These shifts gradually increase at higher wavelengths. The Davydov splitting is not observed in the features of CO in a CO₂ matrix where the (2,0) transition is the strongest (as in pure CO gas; see, e.g., Paper I) in contrast to the VUV-absorption spectrum of pure CO ice, where the (1,0) transition dominates. The band positions of CO₂ are not significantly affected by the presence of CO because the broad band centered on 9.8 eV (126.5 nm) and the vibrational structure in the 124−129 nm range of CO₂ present no detectable perturbations compared with other works with a much lower CO₂ photoproduction during spectral acquisition, and the band positions also agree fairly well with Lu et al. 35, Mason et al. 38, and Monahan & Walker 41. The resulting fit and Gaussians involved are displayed in Fig. 2 as a red dashed-dotted trace. The 142.5 nm (8.70 eV) band reported by Lu et al. 35 was not fitted because of the overlap with the features of photoproduced CO in our experiment. The exact peak position of the feature at wavelengths shorter than 120 nm could not be confirmed because it is beyond our spectral range.

An upper limit for the average and the total integrated VUV-absorption cross section was calculated by subtracting CO, they are × 10^-19 cm² and cm² nm ( cm² eV) in the wavelength range of 120−160 nm (10.33−7.74 eV). The VUV-absorption cross section of CO₂ ice at 121.6 nm is cm². The VUV-absorption cross section at the Lyman band system range (132−165 nm) was not measured becuase of the presence of CO. Gas-phase data from Yoshino et al. 68 were used for comparison with the solid phase data, see blue trace in Fig. 2. Mason et al. 38 reported similar values for the gas- and solid-phase VUV-absorption cross sections, but using Yoshino et al. 68 gas phase data, and our solid phase data we can clearly observe that the VUV-absorption cross section of CO₂ gas is lower than the solid-phase value, see Fig. 2. The Mason et al. 38 solid-phase spectra displays a maximum at 9.9 eV with a VUV-absorption cross section of 1.2 × 10^-18 cm², lower than the cm² value deduced from our data. The VUV-absorption cross section of CO₂ in the gas-phase data from Yoshino et al. 68 has an average value of 3.3 × 10^-19 cm² 120−160 nm range. CO₂ gas data were integrated in the spectral range, giving a value of 1.5 × 10^-17 cm² nm (8.2 × 10^-19 cm² eV), which is lower than the value for solid CO₂ in the same range ( cm² nm). The VUV-absorption cross section of CO₂ gas at 121.6 nm is very low, 6.3 × 10^-20 cm².

It should be noted that, despite the photoproduced CO in our VUV-spectra of CO₂, the VUV-absorption cross sections are similar to previously reported values. The data presented in this section provide some evidence on the CO₂:CO mixture effects in the VUV-spectrum. The VUV-spectroscopy of ice mixtures has been poorly studied (Wu et al. 2012), but it is essential for understanding the absorption of UV-photons in icy grain mantles.

3.3. Solid nitrogen

The ground state and bond energy of N₂ are X and E_b(N−N) = 9.8 eV (Okabe 1978).

Owing to its very low VUV-absorption cross section, a deposition of 2.36 μm, nearly 4.7 × 10¹⁸ molecules cm^-2, for N₂ ice was required to detect the absorption features. The VUV-absorption cross section as a function of wavelength and photon energy is shown in Fig. 4. It presents a vibrational structure in the 114−147 nm (10.87−8.43 eV) region. Two systems (attributed to a X and w X) can be observed. These systems have been reported by Boursey et al. 3, Mason et al. 38, and Wu et al. 67. In agreement with Boursey et al. 3 and Wu et al. 67, we did not observe the two broad continua between 133−163 nm (9.30−7.60 eV) and 113−133 nm (9.30−11.0 eV) that Mason et al. 38 reported. The VUV-absorption cross-section spectrum of N₂ is not as well resolved as in other works, but this probably does not affect the VUV-absorption cross-section scale we provide because no integration of the band area is involved. Table 5 summarizes the peak position and the band area integration of all the features presented. Even though our UV-lamp flux decreases below 120 nm, three features are observed in that spectrum range.

The average VUV-absorption cross section has a value of 7.0 × 10^-21 cm² in the 114.6−146.8 nm (10.82−8.44 eV) range. The total integrated VUV-absorption cross section has a value of 2.3 × 10^-19 cm² nm (1.8 × 10^-20 cm² eV) in the same spectral region. The latter value is indeed affected by the low resolution of our VUV-absorption spectrum. A spectrum with better resolved bands would result in a lower value. Comparing our value of VUV-absorption cross section with data from Wu et al. 67, we estimated an increase of 34% in the total integrated VUV-absorption cross-section value, therefore this integrated value would be ~1.5 × 10^-19 cm² nm. The VUV-absorption cross section at Ly-α (121.6 nm) is very low; we estimated an upper limit value of 1.0 × 10^-21 cm². There is no observable VUV-absorption at molecular hydrogen-band wavelengths (157.8 and 160.8 nm). Gas-phase VUV-absorption data adapted from Mason et al. 38 are plotted in Fig. 4 as a blue dashed trace. The gas data of Mason et al. 38 are not in cross section units, this is why we did not calculate the average and the total integrated VUV-absorption cross sections.

Fig. 4 VUV-absorption cross section as a function of photon wavelength (bottom X-axis) and VUV-photon energy (top X-axis) of N2 ice deposited at 8 K, black solid trace. The blue dashed trace is the VUV-absorption cross-section spectrum of gas phase N2 adapted from Mason et al. 38. The fit, red dotted (single bands, see Table 2) and dashed-dotted trace, is the sum of twenty Gaussian profiles. It has been offset for clarity.

Open with DEXTER

Fig. 5 VUV-absorption cross section as a function of photon wavelength (bottom X-axis) and VUV-photon energy (top X-axis) of O2 ice deposited at 8 K, black solid trace. The blue dashed trace is the VUV-absorption cross-section spectrum of gas phase O2 adapted from Lu et al. 36. The fit, solid red dotted (single bands, see Table 2) and dashed-dotted trace, is the sum of two Gaussians. It has been offset for clarity.

Open with DEXTER

3.4. Solid oxygen

The ground state of O₂ is X and its bond energy is E_b(O−O) = 5.1 eV (Okabe 1978).

The VUV-absorption cross section as a function of wavelength and photon energy is shown in Fig. 5. The O₂ ice thickness estimation based on the fringes method, described in Sect. 3, requires a thick ice of about 60 ML. But the VUV-absorption cross-section measurement only works well for thin ices below 200 ML. Therefore, several experiments were performed to ensure that the ice thickness estimation was correct in the case of O₂. Solid O₂ presents a broad band centered on 141 nm (8.79 eV) in the 118−162 nm (10.50−7.65 eV) region (attributed to the X transition, named Schumann-Runge band). It also presents a feature centered on 177 nm (7.0 eV), according to Mason et al. 38 and Lu et al. 35, which is beyond our spectral range. The resulting fit and Gaussians involved are represented in Fig. 5 by a red dashed-dotted trace. Following Lu et al. 35, there is no obvious requirement for a broad line centered near 161 nm (7.70 eV), but without this band the resulting fit in the 150−162 nm range would not be as good.

The average VUV-absorption cross section has a value of cm² in the 120−162 nm (10.33−7.65 eV) range. The total integrated VUV-absorption cross section has a value of cm² nm ( cm² eV) in the same spectral region. The VUV-absorption cross sections of the O₂ ice at Ly-α (121.6 nm) and the molecular hydrogen band emissions (157.8 and 160.8 nm) are cm², cm², and cm². Gas-phase data from Lu et al. 36 were used for comparison with our solid phase data, see Fig. 5. Solid- and gas-phase data have a broad absorption band, but only the gas-phase spectrum presents discrete transitions. The VUV-absorption cross section of O₂ in the gas phase has an average value of 4.0 × 10^-18 cm². The O₂ gas spectrum was integrated in the 118−162 nm range, giving a value of 3.3 × 10^-16 cm² nm (2.0 × 10^-17 cm² eV), which is higher than the solid-O₂ value. The VUV-absorption cross sections of O₂ gas at 121.6 nm, 157.8 nm and 160.8 nm are 2.5 × 10^-18 cm², 6.5 × 10^-18 cm², and 4.7 × 10^-18 cm², which is also higher than the solid-phase measurements, with the exception of the Ly-α wavelength (121.6 nm).

3.5. Comparison between all the species

Figure 6 shows a comparison of the VUV-absorption cross section for all the species represented in the same linear scale. Because of its low VUV-absorption cross section, the N₂ spectrum was multiplied by a factor of 200 for better appreciation. It can be observed that the most highly absorbing molecule around the hydrogen Ly-α (121.6 nm) position is CH₄. Solid O₂ is most highly absorbing around the Lyman band system. N₂ gives the lowest absorption in both cases. All the species absorb at the Ly-α wavelength, with the exception of N₂, for which an upper limit was given in Sect. 3.3. CH₄ and N₂ absorb VUV-photons with wavelengths lower than 150 nm, while O₂ absorbs throughout the scanned range. Table 6 compares the VUV-absorption cross sections of all the species in the gas and solid phase as well as an average VUV-absorption cross section in the 120.8−122.6 nm and 132−162 nm range normalized by the Ly-α and Lyman band system photon flux of our MDHL.

	Fig. 6 VUV-absorption cross section as a function of wavelength (lower X-axis) and photon energy (top X-axis) of all the species in the solid phase at 8 K. The CO2 absorption only occurs at wavelengths shorter than 133 nm, the absorption at longer wavelengths is due to the photoproduced CO during spectral adquisition.
Open with DEXTER

4. Astrophysical implications

Interstellar and circumstellar ice mantles are composed of several species such as H₂O, CO, NH₃, CO₂, CH₄, and CH₃OH (Mumma & Charnley 2011, and ref. therein). In some cases, these molecular ice components are mixed in the ice mantle.

The VUV-absorption cross-section spectrum of CO₂ presented in this work has the disadvantage of being mixed with the CO produced by the VUV-irradiation of the MDHL used to acquire the VUV-spectroscopy of the species. But on the other hand, our data show that a binary ice mixture containing CO₂ and CO can lead to a mixed VUV-absorption spectrum where mixture effects can be appreciated, using this new way of performing VUV-spectroscopy (cf. Wu et al. 2010, 2012), where the band profiles and positions are altered to some extent by mixture effects.

As is commonly observed for example in infrared spectra, multicomponent ice mantles most likely display VUV-absorption features that correspond to the various species, where CO is the most highly absorbing molecule in the VUV among the common ice components. VUV-irradiation of CO2 ice led to a 22% of CO with respect to the starting CO₂ column density. CH₃OH ice led to 16% of CO for the same UV-fluence, see Paper I. The two production rates are similar but only the VUV-absorption cross-section spectrum of CO₂ ice presents CO features. On the other hand, the average VUV-absorption cross section of CH₃OH ice is almost ten times larger than the average VUV-absorption cross section of CO₂ in the 133−160 nm range.

The VUV-spectrum of CO in a CO₂ ice matrix is different from that of pure CO ice, with shifted positions that occur half-way between the gas and solid CO values, see Table 4 and Fig. 3. Therefore, CO in a solid CO₂ matrix can be treated as a highly porous CO ice mantle. Infrared spectroscopy of CO₂:CO = 20:1 ice shows a similar trend, the CO band shifts from 2138.6 cm^-1 (for pure CO ice) to 2139.7 cm^-1 and the FWHM increases from 2.5 cm^-1 (pure CO ice) to 5.8 cm^-1 (Sandford et al. 1988).

Formation of N₃ from N₂ irradiation with VUV-photons is extremely difficult (Hudson & Moore 2002). Proton and electron bombardment is more effective in the production of the azide radical than VUV-photons (Hudson & Moore 2002; Jamieson & Kaiser 2007). The likely explanation is the low VUV-absorption cross section of N₂. Detection of N₃ in space might be indicate a cosmic-ray field, or alternatively VUV-photons with energies higher than Ly-α (10.2 eV) that we were unable to study in this work. The latter possibility is very unlikely because much more intense absorption is not expected in the short interstellar radiation field range, from 91.2 to 114 nm, which was not measured in the laboratory.

Regarding the two homonuclear molecules studied, O₂ and N₂, the former presents a broad absorption band in the 118−160 nm range with a VUV-absorption cross section on the order of 10^-18 cm^-2, while N₂ presents a vibrational structure in the 118−150 nm range with a VUV-absorption cross section on the order of 10^-20 cm^-2. Using the Pontoppidan 50 approximation of the abundances of nitrogen and oxygen, these two orders of magnitude differences can determine different photochemistry efficiencies for N₂ and O₂. This supports the hypothesis that a photochemistry richer in O than in N is expected in icy grain mantles, because a significant fraction of the N atoms might be locked as solid N₂.

The ice penetration depth of photons with a given wavelength, or the equivalent absorbing ice column density of a species in the solid phase, can be calculated from the VUV-absorption cross section following (6)which can be converted into an ice thickness using Eq. (3), inserting the volumetric density of the ice provided in Table 1: (7)where I_t(λ) is the transmitted intensity for a given wavelength λ, I₀(λ) the incident intensity, N(λ) is the absorbing column density in cm^-2, and σ(λ) is the cross section in cm². Following the estimation for polar ice molecules provided in Paper I, Table 7 provides the column density values of nonpolar ice species that correspond to an absorbed flux of 95% and 99% using the cross-section value at Ly-α, the average cross section in the 120−160 nm range, and the maximum cross section in the same range.

Fayolle et al. 24 reported the photodesorption of N₂ and O₂ ice as a function of photon wavelength, see Figs. 1 and 3 of that work. It can be observed that the photodesorption in the 10.8−8.4 eV spectral range is very low compared with the photodesorption in the 12.0−13.8 eV range for N₂. This is mainly due to the low absorption cross section that we report in that region, see Fig. 4. For O₂, Fig. 5 resembles the photodesorption profile of O₂ for different photon wavelengths in Fayolle et al. 24.

Using the expression developed in Paper I, (8)where I₀ and I_abs are the incident and the absorbed photon intensities at a certain wavelength, we can estimate the photodesorption rate per absorbed photon, , which differs significantly from the photodesorption rate per incident photon, , see Table 8.

For N₂ ice, Fayolle et al. 24 reported an upper limit of molecules per incident photon in the spectral range below 12.4 eV for N = 30 ML. The average absorption cross section for N₂ ice that we measured in that range is σ = 7 × 10^-21 cm². The resulting photodesorption rate is quite high, molecules per absorbed photon, meaning that a very small fraction of the incident photons are absorbed in the ice, but each absorbed photon led to the photodesorption of about 19 molecules on average (this in fact is a maximum value because Fayolle et al. 24 measured photodesorption rates that do not exceed 4 × 10^-3 molecules per incident photon).

A more direct comparison between N₂ and O₂ ice photodesorption rates could be made if the number of monolayers closer to the ice surface that truly contribute to the photodesorption were known. This value has not been estimated for other ices diferent from CO (about 5 ML, Muñoz Caro et al. 2010; Fayolle et al. 2011; and Chen et al. 2013). The above values of and correspond to the total ice column density of 30 ML in the experiment of Fayolle et al. (2013). With this uncertainty still remaining, we can conclude that when the VUV-absorption cross section of each specific ice composition is taken into account, it is possible to estimate the efficiency of the photodesorption per absorbed photon; in the case of N₂, for VUV photon energies that do not lead to direct dissociation of the molecules in the ice, these values are higher than unity. This observation and the fact that photons absorbed in ice monolayers deeper than the top monolayers (up to five for CO) can lead to a photodesorption event, indicate that the excess photon energy is transmitted to neighboring molecules in the ice within a certain range (Rakhovskaia et al. 1995; this range may correspond to about 5 ML in the case of CO ice, e.g. Muñoz Caro et al. 2010). In the case of O₂ the VUV photons have enough energy to dissociated the molecule, which makes it harder to measure the photodesorption rate because the dissociation dominates between the two processes.

5. Conclusions

This work adds to Paper I to complete our set of nine molecular ice components (CO, H₂O, CH₃OH, NH₃, H₂S, CH₄, CO₂, N₂, and O₂) selected to perform VUV-spectroscopy in the 120−160 nm spectral range. Some key aspects of this work are summarized below.

• For the first time, to our knowledge, the VUV-absorption cross sections of CH₄, N₂, and O₂ were measured for the solid phase, with average VUV-absorption cross sections of 5.7 +0.5/−1.1 × 10^-18 cm², 7.0 +0.6/−1.4 × 10^-21 cm², and 4.8 +0.4/−1.0 × 10^-18 cm². The total integrated VUV-absorption cross sections are 2.0 +0.1/−0.4 × 10^-16 cm² nm, 2.3 +0.2/−0.5 × 10^-19 cm² nm, and 2.4 +0.2/−0.5 × 10^-15 cm² nm. Our estimated value of the average VUV-absorption cross section of CO₂ ice, 2.6 +0.2/−0.3 × 10^-17 cm², is comparable with that reported by Mason et al. 38, which was measured using a synchrotron as the monochromatic VUV-emission source.

• The ice samples made of N₂ or CO₂ display discrete VUV-absorption bands, the latter were not well resolved, while samples containing O₂ or CH₄ present a continuum VUV-absorption band, see Fig. 6, where N₂ presents by far the lowest VUV-absorption cross section.

• The ice sample is inevitably UV-irradiated for a few minutes during the spectral acquisition in our experimental configuration, which can lead to photoproduct formation. This effect is clearly observed in the case of CO₂ because the CO photoproduced has a large VUV-absorption cross section compared with CO₂ in the solid phase. In this experiment, the CO bands are shifted in position and their profiles changed with respect to the pure CO ice spectrum reported in Paper I. These mixture effects were not significantly studied and have clear implications for the absorption of multicomponent ice mantles in space.

• The VUV-absorption cross sections of the two homonuclear molecules studied, N₂ and O₂, differ by two orders of magnitude. This affects their photodesorption rates and the formation of photoproducts in the ice matrix.

Acknowledgments

This research was financed by the Spanish MICINN under projects AYA2011-29375 and CONSOLIDER grant CSD2009-00038. This work was partially supported by NSC grants NSC99-2112-M-008-011-MY3 and NSC99-2923-M-008-011-MY3, and the NSF Planetary Astronomy Program under Grant AST-1108898.

References

All Tables

All Figures

	Fig. 1 VUV-absorption cross section as a function of photon wavelength (bottom X-axis) and VUV-photon energy (top X-axis) of CH4 ice deposited at 8 K, black solid trace. The blue dashed trace is the VUV-absorption cross-section spectrum of gas phase CH4 adapted from Lee et al. 32. The fit of the solid-phase spectrum, red dotted (single bands, see Table 2) and dashed-dotted trace, is the sum of three Gaussians. It has been offset for clarity.
Open with DEXTER
In the text

	Fig. 2 VUV-absorption cross section as a function of photon wavelength (bottom X-axis) and VUV-photon energy (top X-axis) of CO2 ice deposited at 8 K, black solid trace. The blue dashed trace is the VUV-absorption cross-section spectrum of gas phase CO2 adapted from Yoshino et al. 68. The fit, solid red dotted (single bands, see Table 2) and dashed-dotted trace, is the sum of two Gaussians. The inset figure is a CO2 VUV-absorption cross-section close-up in the 124−129 nm range.
Open with DEXTER
In the text

	Fig. 3 Comparison between the VUV-absorption cross-section spectra of CO (red dashed-dotted trace) and CO2 (black solid trace, the CO2 ice is mixed with CO formed during the spectral acquisition). CO transitions in a CO2 matrix are shifted to shorter wavelengths with respect to pure CO ice transitions.
Open with DEXTER
In the text

	Fig. 4 VUV-absorption cross section as a function of photon wavelength (bottom X-axis) and VUV-photon energy (top X-axis) of N2 ice deposited at 8 K, black solid trace. The blue dashed trace is the VUV-absorption cross-section spectrum of gas phase N2 adapted from Mason et al. 38. The fit, red dotted (single bands, see Table 2) and dashed-dotted trace, is the sum of twenty Gaussian profiles. It has been offset for clarity.
Open with DEXTER
In the text

	Fig. 5 VUV-absorption cross section as a function of photon wavelength (bottom X-axis) and VUV-photon energy (top X-axis) of O2 ice deposited at 8 K, black solid trace. The blue dashed trace is the VUV-absorption cross-section spectrum of gas phase O2 adapted from Lu et al. 36. The fit, solid red dotted (single bands, see Table 2) and dashed-dotted trace, is the sum of two Gaussians. It has been offset for clarity.
Open with DEXTER
In the text

	Fig. 6 VUV-absorption cross section as a function of wavelength (lower X-axis) and photon energy (top X-axis) of all the species in the solid phase at 8 K. The CO2 absorption only occurs at wavelengths shorter than 133 nm, the absorption at longer wavelengths is due to the photoproduced CO during spectral adquisition.
Open with DEXTER
In the text