© ESO, 2012

1. Introduction

Star formation begins with the collapse of a prestellar core inside a molecular cloud to form a protostar, which then grows from material coming from its envelope, and later through an accretion disk, where planets and comets may form. This physical evolution from the molecular cloud to a planetary system is accompanied by chemical evolution induced by heat and ultraviolet (UV) or X-rays radiations from the protostar. Much of this chemical evolution takes place in the icy mantle of interstellar grains, and the aim of this work is to consolidate the knowledge of the ice evolution during the star formation process, and especially the influence of thermal reactions as a source of molecular complexity in the interstellar medium (ISM). In this study, we are interested in the thermal reactivity between two interstellar molecules, formaldehyde (H₂CO), and methylamine (CH₃NH₂) in water ice analogues. Formaldehyde absorption features in interstellar ices observed in young stellar objects (YSOs) in the 5.7−6 μm region (the so-called C1 component) indicate abundances around 6% with respect to water (Boogert et al. 2008). Its formation in ice seems to be well understood and is explained by direct hydrogenation of CO on the grain surface (Hiraoka et al. 1994, 2002; Watanabe et al. 2004). Methylamine CH₃NH₂ is the simplest member of the primary alkylamine family and has been detected in gas phase by its 2₀₂ → 1₁₀ A_a state transition in the direction of Sgr B2 and Ori A (Fourikis et al. 1974). In addition, results from the stardust mission support its presence in cometary cores (Elsila et al. 2009). Several pathways of formation either on grain surface or in gas phase have been proposed. Godfrey et al. (1973) suggested its formation on grain surface from the hydrogenation reactions starting on the CN radical (Eq. (1)). This mechanism has recently been experimentally validated starting from HCN (Theule et al. 2011), $\mathrm{CN}\stackrel{\mathrm{H}}{\mathrm{-}0\mathrm{\to }}\mathrm{HCN}\stackrel{\mathrm{H}}{\mathrm{-}0\mathrm{\to }}\mathrm{C}{\mathrm{H}}_{\mathrm{2}}\mathrm{NH}\stackrel{\mathrm{H}}{\mathrm{-}0\mathrm{\to }}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{N}{\mathrm{H}}_{\mathrm{2}}\mathit{.}$(1)Gardner & McNesby (1982) and Ogura et al. (1989) proposed its formation from the UV photolysis of a gaseous mixture containing CH₄ and NH₃. Finally, Herbst (1985) proposed a gas phase methylamine formation from the methylium cation CH$\begin{array}{c}\mathrm{+}\\ \mathrm{3}\end{array}$ and NH₃. However, solid methylamine has never been detected in ice, so far. This could be due to its low abundance, but also to the difficulty to assign this compound from ices in infrared astronomical spectra. Works on ice analogues containing methylamine suggest that a plausible cause of its non-observation in interstellar ices is its high reactivity at low temperature (Bossa et al. 2009a). Indeed, from a chemical point of view, CH₃NH₂ is a better nucleophile and base than ammonia. This molecule can quickly react at low temperature in interstellar ice analogues with acids, such as HNCO, HCN, and HCOOH. In addition, methylamine reacts at low temperature with CO₂, leading to the formation of methylammonium methylcarbamate (CH₃NH$\begin{array}{c}\mathrm{+}\\ \mathrm{3}\end{array}$CH₃NHCOO⁻), which can be isomerised into glycine under UV irradiation (Bossa et al. 2009a). The measured activation barrier for this latter reaction, 3.7 kJ mol^-1, is compatible with low temperature observed in molecular clouds (Bossa et al. 2009a). Among molecules observed in interstellar ices, formaldehyde is of prime interest, since thermally promoted formaldehyde reactions in presence of NH₃ at low temperature have shown that they are an important source of organic molecules in astrophysical environments such as comets and interstellar ices (Bossa et al. 2009b; Schutte et al. 1993; Vinogradoff et al. 2011). The resulting products are the polymer of H₂CO, the polyoxymethylene (POM), and the aminomethanol (NH₂CH₂OH) (Bossa et al. 2009b; Schutte et al. 1993).

In this work, we demonstrate that methylamine and formaldehyde quickly react in water ices at low temperature to form N-methylaminomethanol (CH₃NHCH₂OH). The measured activation energy of this reaction, 1.1 kJ mol^-1, implies that this reaction can easily occur in interstellar ices, when the ices are gently warmed, as it is in YSOs, in photo-dissociation regions (PDRs), or in comets. Therefore, CH₃NHCH₂OH is likely to be found in these objects. This hypothesis is confirmed by numerical simulations, which show that the formation of N-methylaminomethanol is likely at low temperature. Isotopic experiments as well as photochemical studies have also been performed to better characterise ice evolution induced by heat and UV radiation in star-forming regions and cometary environments.

2. Experimental and theoretical section

Methylamine is commercially available as a 99.9995% pure gas from Air Liquide. Methylamine ¹⁵N is formed from methylamine (¹⁵N)-HCl commercially available from Cambridge Isotope (98%). The H₂O/CH₃NH₂/H₂CO gas mixtures are prepared at room temperature into two mixing lines under primary vacuum. The first one for the CH₃NH₂ gas mixture and the second one for H₂CO and H₂O, to prevent any gas phase reaction. They are then co-deposited with a chosen ratio onto a gold-plated metal surface cooled to 20 K by a Model 21 CTI cold head within a high vacuum chamber (ca 10^-8 mbar). The infrared spectra are recorded between 4000 and 650 cm^-1 in the reflexion mode using a Fourier transform infrared spectrometer (FTIR). A typical spectrum has a 1 cm^-1 resolution and is averaged over one hundred interferograms. The sample is warmed using a heating resistance, and the temperature is controlled using a Lakeshore Model 331 temperature controller. The mass spectra are recorded using an MKS quadrupole mass spectrometer as the products are being desorbed. The ionisation source is a 70 eV impact electronic source, and the mass spectra are recorded between 1 and 80 amu. The UV radiations (λ > 120 nm) are generated from a microwave discharge hydrogen flow lamp (Opthos instruments) separated from the vacuum chamber by a magnesium fluoride window. Lamp settings during irradiation are P_H2 = 300 × 10^-3 mbar, microwave forward power 70%, reflected power less than 10%. The distance between the lamp and the ice sample is about 10 cm. The flux has been calibrated with the actinometry method described by Cottin et al. (2003) using the photolysis of methanol. The flux of the hydrogen lamp is 2    ×    10¹⁵ photons cm^-2 s^-1. The concentration ratio of the different mixtures is first set before deposition using standard manometric techniques, and then derived from the infrared spectra by integrating vibrational bands to estimate the column density N_x (molecule cm^-2) of CH₃NH₂, and H₂CO according to their band strengths (cm molecule^-1) provided by the literature using ${\mathit{N}}_{\mathit{x}}\mathrm{=}\frac{\mathrm{2.3}{\mathrm{\times }}^{\mathrm{\int }}{\mathit{A}}_{\mathit{v}}\mathrm{d}\mathit{v}}{\mathrm{2}\mathrm{\times }{\mathrm{𝒜}}_{\mathit{x}}}$. For CH₃NH₂, we used the band strength of the wagging mode at 889 cm^-1 of 1.3 × 10^-17 cm molecule^-1 (Holtom et al. 2005), and for H₂CO, we used the values given by Schutte et al. (1993): 9.6 × 10^-18 cm molecule^-1 for the C=O stretching mode at 1725.5 cm^-1 and 4 × 10^-18 cm molecule^-1 for the CH₂ scissoring mode at 1494 cm^-1. The thickness of the deposited solid films, assuming a density of 0.92 g cm^-3 for H₂O, is estimated to be around 0.1 μm. We also used the band integration strengths to monitor the decrease in the reacting infrared band intensities as reactants are consumed during the reaction, and to estimate how much of each product is formed. We estimated the N-methylaminomethanol band strengths assuming that all formaldehyde has been converted into it. Accordingly, knowing the amount of N-methylaminomethanol and using the relation ${\mathit{N}}_{\mathit{x}}\mathrm{=}\frac{\mathrm{2.3}{\mathrm{\times }}^{\mathrm{\int }}{\mathit{A}}_{\mathit{v}}\mathrm{d}\mathit{v}}{\mathrm{2}\mathrm{\times }{\mathrm{𝒜}}_{\mathit{x}}}$, we calculated the band strengths of the N-methylaminomethanol. We performed a vibrational analysis of the N-methylaminomethanol (CH₃NHCH₂OH) to compute the harmonic vibrational frequencies for both the ¹⁴N and ¹⁵N isotopologues. Nevertheless, to simplify this analysis we did not take into account the effect of environment on the calculations. Calculations were performed using the Gaussian 98 package (Frisch et al. 1998; Lee et al. 1988) at the B3LYP/6-31G** level, which is known to supply reliable predictions of vibrational wavenumbers (Duvernay et al. 2004, 2007).

3. Results

3.1. Warming of an H₂CO/CH₃NH₂ ice mixture: N-methylaminomethanol formation

Fig. 1 FTIR spectrum of (a) H2O/H2CO/CH3NH2 = 1/0.1/2 ice mixture deposited at 30 K; (b) pure H2O ice deposited at 30 K; (c) pure H2CO ice deposited at 30 K; (d) pure CH3NH2 deposited at 20 K.

Thermally promoted formaldehyde reactions at low temperature in H₂O/NH₃/H₂CO ices have been studied by Schutte et al. (1993) and Bossa et al. (2009b). These authors distinguished at least two different products: polyoxymethylene (POM) HO(CH₂O)_nH, the H₂CO polymer, and aminomethanol (NH₂CH₂OH). The nature of the products and their relative abundances depend strongly on the initial composition of the initial ice mixture. Schutte et al. (1993) have shown that NH₂CH₂OH is the dominant product when NH₃ is in excess with respect to H₂CO. We chose to adopt the same argument in working with an excess of methylamine to avoid formaldehyde polymerisation. Figure 1 displays the infrared spectrum of an ice mixture of H₂O/H₂CO/CH₃NH₂ (1/0.1/2) deposited at 30 K compared with the spectra of pure H₂O, H₂CO, and CH₃NH₂ ices. The ice mixture spectrum is dominated by a strong and broad bands caused by the OH and NH stretching modes of H₂O and CH₃NH₂. Formaldehyde is easily characterised from two characteristic bands located at 1715 and 1497 cm^-1 caused by the C=O stretching mode and CH bending modes. The infrared absorption bands observed in Fig. 1 can be assigned to initial species only, which indicates the lack of reactivity at 30 K. These infrared absorption bands are listed in Table 1 along with their vibrational assignments and band strengths. This initial ice analogue was then warmed from 30 K to 240 K using different heating rates of 10, 4, 2, 1, and 0.5 K min^-1.

Fig. 2 FTIR spectra of (a) H2O/H2CO/CH3NH2 = 1/0.1/2 ice mixture deposited at 30 K and recorded at 30 K; (b) recorded at 70 K; (c) recorded at 190 K; (d) pure aminomethanol (NH2CH2OH) recorded at 190 K.

Figure 2 displays the infrared spectrum of an ice mixture of H₂O/H₂CO/CH₃NH₂ (1/0.1/2) deposited at 30 K and then warmed at 4 K min^-1. From 45 K, the bands caused by formaldehyde start to decrease and fully disappear at 70 K (Fig. 2b). This range of temperature is lower than expected for formaldehyde sublimation, which in our experimental conditions is observed between 100 and 110 K. At 190 K, it remains on the substrate the reaction product (Fig. 2c). Thus the disappearance of H₂CO can only be explained by reactivity. Changing the heating rate from 10 to 0.5 K min^-1 does not alter the reactivity. Nevertheless, heating rates in the ISM are expected to be lower than those usually used in laboratory. Heating rates as low as this cannot be investigated in the laboratory, so it is hard to predict in which way they will affect the chemistry. The thermal reactivity of H₂CO in ice with NH₃ has previously been investigated experimentally and theoretically (Bossa et al. 2009b; Chen & Woon 2011; Rimola et al. 2010; Schutte et al. 1993). This reaction leads to the formation of the aminoalcohol NH₂CH₂OH, which occurs in a low barrier nucleophilic addition between NH₃ and H₂CO. The infrared spectrum of the observed product is consistent with an aminoalcohol structure, which in this experiment should be the N-methylaminomethanol (CH₃NHCH₂OH) (Figs. 2c, and d). Indeed, the spectrum is dominated by a strong band in the NH stretching region at 3260 cm^-1, and in the C−O, C−N stretching region (Fig. 3). To confirm the N-methylaminomethanol formation, we performed an isotopic substitution experiment using ¹⁵N methylamine, and compared these results with those obtained from our quantum calculation using the B3LYP 6-311G++(2d, 2p) basis set. The theoretical and experimental frequencies for ¹⁴N-methylaminomethanol are listed in Table 2 along with theoretical and experimental frequency shifts between ¹⁴N and ¹⁵N isotopologues. Although harmonic calculations overestimate frequencies corresponding to X-H stretching modes (Table 2), we observe a good agreement between the experimental and calculated frequencies of ¹⁴N-methylaminomethanol (Table 2, Fig. 3), especially at frequencies lower than 1600 cm^-1. We also compared experimental frequency shifts between ¹⁴N and ¹⁵N isotopologues with theoretical values. Calculated frequency shifts are small (smaller than 15 cm^-1). The largest calculated shift of 13 cm^-1 is for the C−N stretching mode, which agrees well with the observed experimental shift of 11 cm^-1 (Table 2). The same agreement is also observed for others modes. For example, the calculated isotopic shift for the NH stretching mode is 8 cm^-1, while we experimentally observed 7 cm^-1.

Fig. 3 FTIR spectrum of pure CH3NHCH2OH formed from the warming of an H2O/H2CO/CH3NH2 ice mixture compared with the corresponding B3LYP infrared spectrum.

Fig. 4 Mass spectrum of CH3NHCH2OH 14N and 15N recorded at 210 K.

Fig. 5 FTIR spectra of (a) H2O/H2CO/CH3NH2 = 1/0.1/0.05 ice mixture deposited at 30 K and recorded at 30 K; (b) recorded at 125 K; (c) recorded at 185 K; (d) recorded at 200 K.

The CH₃NHCH₂OH formation is also confirmed by mass spectrometry as shown in Fig. 4. A clear peak at m/z 61 is observed when the ¹⁴N-methylaminomethanol sublimates from the substrate around 210 K. This peak is assigned to the CH₃NHCH₂OH molecular ion and is labelled M in Fig. 4. In addition, the peak at m/z 43 (Fig. 4) corresponding to the loss of H₂O (M-18) is consistent with a molecule-bearing OH group. The mass spectrum of the ¹⁵N isotopologue is similar, although most of the peaks are shifted forward by 1 mass unity due to the isotopic substitution (Fig. 4), which confirms the previous assignments. The ice composition we used to form and characterise the N-methylaminomethanol is not astrophysically relevant since we worked with an excess of methylamine to avoid the formaldehyde polymerisation. To check if the N-methylaminomethanol formation is still possible in a more realistic ice, we investigated the warming of an ice mixture containing H₂O/H₂CO/CH₃NH₂ = 1/0.1/0.05. Figure 5 shows the thermal evolution of this ice mixture. The infrared spectrum recorded at 30 K after deposition is dominated by the strong infrared band of water at 3200 cm^-1, while the infrared features of formaldehyde and methylamine are weak due to their low abundances in the ice mixture (Fig. 5). The ice mixture was then warmed at 4 K min^-1. At 125 K, both formaldehyde and methylamine have been consumed while the bands of N-methylaminomethanol are observed at 1481, 1454, 1423, 1388, 1208, 1143, 1064, 1032, and 1004 cm^-1. The position of the bands slightly differs from their positions observed for the pure solid (Table 2) because of the large amount of water. Hydrogen interaction between water and the N-methylaminomethanol could be at the origin of these slight frequency shifts. At 185 K, while the water is sublimating from the substrate, we notice a decrease of the infrared features of the N-methylaminomethanol, indicating that a part of the product is driven away by the water desorption. This is confirmed by the infrared spectrum recorded at 200 K. At this temperature, only a trace of the N-methylaminomethanol remains on the substrate. This is an important result from an astrophysical point of view. Indeed, the formation of the N-methylaminomethanol, from the heating of a realistic ice analogue, is still possible and the majority of the thermal product is released in the gas phase during the water desorption, allowing its detection by radio astronomy.

Fig. 6 Formaldehyde thermal evolution as a function of temperature in water ice containing CH3NH214N (triangle), and 15N (square) (H2O/CH3NH2/CH2O = 1/2/0.1). The rate coefficients are given in the form k = ν0exp( − E/RT), where T is the ice temperature (K), ν0 the pre-exponential factor (s-1), and E the activation energy (J mol-1). The corresponding rate coefficients are k14N = 1.5 ± 0.1 × 10-2exp( − 1100 ± 50/RT) and k15N = 2.6 ± 0.3 × 10-2exp( − 1800 ± 80/RT).

3.2. Rate of CH₃NHCH₂OH formation

We measured the CH₃NHCH₂OH formation rate in a H₂O/CH₃NH₂/H₂CO ice mixture in an 1/2/0.1 concentration ratio. We chose to work with a methylamine excess to avoid any formaldehyde polymerisation and to ensure a first-order decay for the formaldehyde. In a first approximation we can state that the methylamine concentration in the ice is constant during the ice warming. The ice mixture was deposited at 19 K and then warmed at 4 K min^-1. During the warming, we followed the evolution of the formaldehyde using its band located at 1715 cm^-1 ( cm molecule^-1). This thermal evolution is displayed in Fig. 6. Experimental data were fitted using Eq. (2). In this equation, T is the temperature (K), ν₀ the pre-exponential factor (s^-1), E_a the activation energy (kJ mol^-1), and β the temperature ramp (K min^-1). We integrated the formaldehyde variation over the time and then substituted the time by the temperature according to the relation T = T₀+βt. We finally obtained a relation between the amount of formaldehyde as a function of temperature (Eq. (2)), $\{\begin{array}{c}\\ \frac{\mathrm{d}\mathrm{[}{\mathrm{H}}_{\mathrm{2}}\mathrm{CO}{\mathrm{]}}_{\mathit{t}}}{\mathrm{d}\mathit{t}}\mathrm{=}\mathrm{-}{\mathit{k}}_{\mathit{T}}\mathrm{[}{\mathrm{H}}_{\mathrm{2}}\mathrm{CO}{\mathrm{]}}_{\mathit{t}}\\ \mathit{T}\mathrm{=}{\mathit{T}}_{\mathrm{0}}\mathrm{+}\mathit{\beta t}\mathrm{and}{\mathit{k}}_{\mathit{T}}\mathrm{=}{\mathit{\nu }}_{\mathrm{0}}{\mathrm{e}}^{\mathrm{-}\frac{{\mathit{E}}_{\mathrm{a}}}{\mathit{RT}}}\\ {\left[{\mathrm{H}}_{\mathrm{2}}\mathrm{CO}\right]}_{\mathit{T}}\mathrm{=}{\left[{\mathrm{H}}_{\mathrm{2}}\mathrm{CO}\right]}_{\mathrm{0}}{\mathrm{e}}^{\frac{{\mathit{\nu }}_{\mathrm{0}}\mathrm{(}{\mathit{T}}_{\mathrm{0}}\mathrm{-}\mathit{T}\mathrm{)}}{\mathit{\beta }}{\mathrm{e}}^{\mathrm{-}\frac{{\mathit{E}}_{\mathrm{a}}}{\mathit{RT}}}}\mathit{.}\end{array}$(2)The derivated activation energy of 1.1 ± .05 kJ mol^-1 (i.e. 133 K) is lower than the one obtained for the aminomethanol (4.5 kJ mol^-1 or 541 K) (Bossa et al. 2009b). This is not surprising since methylamine is a better nucleophile than ammonia. Moreover, a barrier this low is compatible with the low temperature observed in interstellar ices, meaning that the addition of methylamine on formaldehyde is likely in interstellar grains. We also investigated the influence of isotopic substitution on the reaction rate. We used the same experimental procedure as previously on an ice mixture containing H₂O/CH$\begin{array}{c}\mathrm{15}\\ \mathrm{3}\end{array}$NH₂/CH₂O in a 1/2/0.1 concentration ratio. This ice mixture was deposited at 19 K and then warmed at 4 K min^-1 as for the ¹⁴N experiment. The corresponding thermal evolution is displayed in Fig. 6. The time evolution of formaldehyde was then fitted with Eq. (2). The derived activation energy is in this case, 1.8 ± .08 kJ mol^-1 (i.e. 217 K). The isotopic substitution modifies the reaction rate (as seen in Fig. 6) and the activation barrier. This is consistent with a primary isotopic effect, since the bonds that are formed during the reaction involves the nitrogen atom. However, the difference between the measured barrier for the ¹⁴N and ¹⁵N is bigger than the value we would expect for nitrogen substitution. This could be due to the mechanism of N-methylaminomethanol formation in ice at low temperature. Indeed, the measured activation barriers are effective barriers that include diffusion processes of the reactants, nucleophilic addition (i.e. attack of the methylamine on the formaldehyde), and finally the hydrogen transfer from the nitrogen to the oxygen of the N-methylaminomethanol. All these processes could be sensitive to the isotopic substitution, leading to an isotopic kinetic effect stronger than it would appear in liquid phase at room temperature where diffusion and hydrogen transfers are not necessary the limiting rate of nucleophilic addition reaction.

3.3. Photolysis of CH₃NHCH₂OH

Fig. 7 Absorbance of CH3NHCH2OH as a function of UV irradiation time. Experimental data are fitted against a first-order decay.

As already mentioned in the introduction, interstellar ices in a star-forming region or in pre-cometary ices are submitted to various physical processes such as UV irradiation. Therefore, we investigated the UV degradation of the N-methylaminomethanol. We first formed CH₃NHCH₂OH as described above. We then desorbed the starting materials (H₂O, H₂CO, and CH₃NH₂) at 180 K to obtain pure CH₃NHCH₂OH. We then irradiated it at 20 K with a UV hydrogen lamp and followed its photolysis by infrared spectroscopy as displayed in Fig. 7. The UV flux is estimated to be 2  ×  10¹⁵ photons cm^-2 s^-1. After 120 min of UV irradiation, 60% of CH₃NHCH₂OH was consumed, as shown in Fig. 7, and its decay was fitted by a first-order kinetic rate. The corresponding kinetic constant is found to be around 7  ±  0.8 × 10^-4 s^-1. With our flux of photons, the photo-destruction cross-section can be estimated to be σ_photo = 3 ± 0.4 × 10^-19 photon^-1 cm². This photo-destruction cross-section refers at least to two different processes: photo-dissociation (i.e. bond-breaking) and photo-desorption (Fayolle et al. 2011, and references therein). However, in our vacuum condition, it is not possible to separate those two processes.

In addition to the infrared signatures of CH₃NHCH₂OH (see Fig. 8), the irradiated sample displays through new bands at 1670 cm^-1 the formation of an amide, which could be the N-methylformamide CH₃NHCHO (Shin et al. 2006), as shown in Fig. 8. After irradiating N-methylaminomethanol at 10 K, the sample was warmed to 250 K and the desorption of the photo-product was monitored by mass spectrometry (Fig. 9). At 220 K, a clear peak is observed at m/z 59, corresponding to the molecular ion of N-methylformamide. This peak is not present when CH₃NHCH₂OH is not irradiated. It has been noted that UV irradiation induces a modification of temperature desorption through changes in the ice structure induced by photons (Fig. 9). The N-methylformamide formation from N-methylaminomethanol may be the result of a dehydrogenation process as follows:

$\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{NHC}{\mathrm{H}}_{\mathrm{2}}\mathrm{OH}\mathrm{+}\mathrm{h}\mathit{\nu }\mathrm{\to }\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{NHCHO}\mathrm{+}{\mathrm{H}}_{\mathrm{2}}\mathit{.}$The same dehydrogenation mechanism has been observed in the UV photolysis of methanol and aminomethanol that give formaldehyde and formamide, respectively (Bossa et al. 2009b; Öberg et al. 2009). The direct production of N-methylformamide from the photolysis of our product is an additional evidence of the N-methylaminomethanol formation.

Fig. 8 Infrared spectra of the UV photolysis of pure CH3NHCH2OH 14N at 20 K.

Fig. 9 Temperature-programmed desorption of pure N-methylaminomethanol with and without UV irradiation. The m/z 61 signal refers to the N-methylaminomethanol while the m/z 59 signal refers to the photo-product: the N-methylformamide.

4. Astrophysical implications

We experimentally demonstrated that formaldehyde and methylamine quickly react in laboratory conditions to form the N-methylaminomethanol. To know if this reaction could be likely in more astrophysical conditions in terms of temperature, time scale, abundances, and UV flux, we simulated its formation using a reduced chemical reaction network (Eq. (3)),

$\{\begin{array}{c}\\ \frac{\mathrm{d}\mathrm{[}{\mathrm{H}}_{\mathrm{2}}\mathrm{CO}\mathrm{]}}{\mathrm{d}\mathit{t}}\mathrm{=}\mathrm{-}{\mathit{k}}_{\mathrm{1}}\mathrm{\left[}{\mathrm{H}}_{\mathrm{2}}\mathrm{CO}{\mathrm{]}}_{\mathit{t}}\mathrm{\right[}\mathrm{N}{\mathrm{H}}_{\mathrm{3}}{\mathrm{]}}_{\mathit{t}}\mathrm{-}{\mathit{k}}_{\mathrm{2}}\mathrm{\left[}{\mathrm{H}}_{\mathrm{2}}\mathrm{CO}{\mathrm{]}}_{\mathit{t}}\mathrm{\right[}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{N}{\mathrm{H}}_{\mathrm{2}}{\mathrm{]}}_{\mathit{t}}\\ \mathrm{-}{\mathit{k}}_{\mathrm{3}}\mathrm{[}{\mathrm{H}}_{\mathrm{2}}\mathrm{CO}{\mathrm{]}}_{\mathit{t}}\mathrm{-}{\mathit{k}}_{\mathrm{4}}\mathrm{[}{\mathrm{H}}_{\mathrm{2}}\mathrm{CO}{\mathrm{]}}_{\mathit{t}}\\ \frac{\mathrm{d}\mathrm{[}\mathrm{N}{\mathrm{H}}_{\mathrm{3}}{\mathrm{]}}_{\mathit{t}}}{\mathrm{d}\mathit{t}}\mathrm{=}\mathrm{-}{\mathit{k}}_{\mathrm{1}}\mathrm{\left[}{\mathrm{H}}_{\mathrm{2}}\mathrm{CO}{\mathrm{]}}_{\mathit{t}}\mathrm{\right[}\mathrm{N}{\mathrm{H}}_{\mathrm{3}}{\mathrm{]}}_{\mathit{t}}\mathrm{-}{\mathit{k}}_{\mathrm{5}}\mathrm{[}\mathrm{N}{\mathrm{H}}_{\mathrm{3}}{\mathrm{]}}_{\mathit{t}}\mathrm{-}{\mathit{k}}_{\mathrm{6}}\mathrm{[}\mathrm{N}{\mathrm{H}}_{\mathrm{3}}{\mathrm{]}}_{\mathit{t}}\\ \frac{\mathrm{d}\mathrm{[}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{N}{\mathrm{H}}_{\mathrm{2}}{\mathrm{]}}_{\mathit{t}}}{\mathrm{d}\mathit{t}}\mathrm{=}\mathrm{-}{\mathit{k}}_{\mathrm{2}}\mathrm{\left[}{\mathrm{H}}_{\mathrm{2}}\mathrm{CO}{\mathrm{]}}_{\mathit{t}}\mathrm{\right[}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{N}{\mathrm{H}}_{\mathrm{2}}{\mathrm{]}}_{\mathit{t}}\\ \mathrm{-}{\mathit{k}}_{\mathrm{7}}\mathrm{[}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{N}{\mathrm{H}}_{\mathrm{2}}{\mathrm{]}}_{\mathit{t}}\mathrm{-}{\mathit{k}}_{\mathrm{8}}\mathrm{[}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{N}{\mathrm{H}}_{\mathrm{2}}{\mathrm{]}}_{\mathit{t}}\\ \end{array}$(3)The abundance of CH₃NHCH₂OH was calculated using a numerical resolution of a system of first-order ordinary differential equations that model the formation and destruction of each species, the water abundance being kept constant (Eq. (3)). Terms in these equations can be sorted into three different classes: thermal formation, thermal desorption, and photo-destruction. The photo-destruction term includes two different processes: photo-dissociation and photo-desorption. However, due to our experimental setup, we cannot separate those two phenomena. The reduced chemical reaction network, parameters, and the ordinary differential equation system are shown in Eq. (3), and Table 3. For two-body thermal reactions (k₁, k₂, and k₁₃) and thermal desorptions (k₄, k₅, k₈, k₉, k₁₁), the rate coefficients are given in the form k = ν₀exp( − E/T), where T is the ice temperature, ν₀ the pre-exponential factor, and E the energy (activation or desorption). Two sets of simulation were carried out. The first one neglects diffusion processes, while the second one includes them. In the latter case, we substituted the pre-exponential factor by a new parameter (ν₀(dif)), which takes into account the diffusion processes according to the method developed by Hasegawa et al. (1992) (see Table 3). Thus for a two-body thermal reaction including the diffusion between the i and j species, the rate coefficients are given in the form k = ν₀(dif)exp(− E/T). Photo-dissociation rates (k₃, k₆, k₇, k₁₀, k₁₂, k₁₄) are written as k = σf, where σ represents the photo-dissociation cross-section for a given species, and f the ultraviolet radiation field depending on the astrophysical environment (Woodall et al. 2007). Kinetic data were used as well as the photo-dissociation cross-section obtained in this work (σ_photo (CH₃NHCH₂OH) = 3 ± 0.4 × 10^-19 photon^-1 cm²) to estimate kinetic rates of the formation (k₁) and photo-dissociation reactions (k₁₂) of CH₃NHCH₂OH. The photo-dissociation rates of CO₂, H₂CO, NH₃, and NH₂CH₂OH were also taken into account using the kinetic rates given in the literature. We also included the desorption rates for all reactants and all products. We included thermal reactions that could compete with the N-methylaminomethanol formation: methylcarbamate (Eq. (4)) and aminomethanol formations (Eq. (5)) (Bossa et al. 2009a,b). All thermal reactions, photo-dissociation, and desorption rates are given in Table 3. $\begin{array}{ccc}& & \mathrm{2}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{N}{\mathrm{H}}_{\mathrm{2}}\mathrm{+}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}\stackrel{\mathrm{\Delta }}{\mathrm{-}0\mathrm{\to }}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{N}{\mathrm{H}}_{\mathrm{3}}^{\mathrm{+}}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{NHCO}{\mathrm{O}}^{\mathrm{-}}\\ & & {\mathrm{H}}_{\mathrm{2}}\mathrm{CO}\mathrm{+}\mathrm{N}{\mathrm{H}}_{\mathrm{3}}\stackrel{\mathrm{\Delta }}{\mathrm{-}0\mathrm{\to }}\mathrm{N}{\mathrm{H}}_{\mathrm{2}}\mathrm{C}{\mathrm{H}}_{\mathrm{2}}\mathrm{OH}\mathit{.}\end{array}$Other reactions that can occur in the ice between CH₃NH₂ and acids such as HNCO, HCOOH, and HCN are neglected in our simulation since they are supposed to be in lower abundances than NH₃ and CO₂. The initial ice composition H₂O/NH₃/CH₃NH₂/H₂CO/CO₂ = 100/10/1/5/7 was chosen according to the average abundances observed in interstellar ices. The temperature and the UV flux were kept constant at 10 K and 10³ photons cm^-2 s^-1 during the simulation. Simulated abundances of reactants (namely NH₃, H₂CO, CH₃NH₂, and CO₂) are displayed in Fig. 10, while the abundances of the related products are displayed in Fig. 11. At 10 K, without taking into account the diffusions processes, the abundances of NH₃, and CO₂ are not affected by thermal activation (Fig. 10). The slow decrease observed for ammonia after 10⁵ years is only caused by photo-degradation process. In contrast, in the same conditions, formaldehyde and methylamine quickly react in less than one year (see Fig. 10 left panel) to form the N-methylaminomethanol (see Fig. 11). The maximum concentration (1% with respect to water) is reached after one year and slowly decreases after 10⁶ years through UV degradation. Finally it is fully photolysed, forming the N-methylformamide after 10⁸ years (Fig. 11 left panel). In this case methylamine is fully consumed (Fig. 10).

In interstellar ices, molecules are diluted in water, and diffusion of molecules is an important process that will slow down two-body thermal reactions since molecules have to diffuse before reacting. Therefore, we included a diffusion parameter (Hasegawa et al. 1992) in our simulations by changing the pre-exponential factor (ν₀) to a new constant parameter ν₀(dif) = 3.3 × 10^-9 s^-1. Results are shown in the right panel of Figs. 10 and 11. In this case formaldehyde and methylamine still react but at higher temperature (30 K) which can be observed on astrophysical time scales. The methylamine is fully consumed and is not observed after 10⁵ years, meaning that in the presence of formaldehyde, methylamine is not likely to be detected in warmed objects. As previously, the N-methylaminomethanol thermally formed was photolysed and fully disappears after 10⁸ years, giving the corresponding amide (CH₃NHCHO).

Fig. 10 H2CO (full triangle), NH3 (full circle), CO2 (full diamond), CH3NH2 (full square) simulated abundances in ice as a function of time at 10 K and in a UV flux of 103 photons cm-2 s-1 without taking into account diffusion processes (left panel), and at 30 K in a UV flux of 103 photons cm-2 s-1 with diffusion processes included (right panel).

Fig. 11 N-methylaminomethanol (CH3NHCH2OH) (full square) and N-methylformamide (CH3NHCHO) (open square) simulated abundances in ice as a function of time at 10 K and in a UV flux of 103 photons cm-2 s-1 without taking into account diffusion processes (left panel), and at 30 K in a UV flux of 103 photons cm-2 s-1 with diffusion processes included (right panel).

Fig. 12 Infrared spectra of CH3OH, CH3NHCH2OH, NH2CH2OH recorded at 10 K. The shaded areas indicate the regions of the strong interstellar H2O and silicate absorptions.

Fig. 13 Comparison between W 33A and laboratory spectra of CH3OH, CH3NHCH2OH, and NH2CH2OH in the 5.5−7.7 μm range.

According to our simulations, the formation of N-methylaminomethanol from the reaction between formaldehyde and methylamine seems to be possible in interstellar ices when they are gently warmed, as occurs in YSOs, or in comets. However, solid-state identification in interstellar ices from infrared spectra is far from obvious, since some regions are blended by the strong infrared features of both water and silicate (see shaded areas in Fig. 12), and also by the low abundance of these compounds. Aminoalcohols have absorption bands in a clear region in the 5−7 μm wavelength range (Fig. 12), and thus could contribute to the 5−7 μm complex, and more specially to the 6, 6.85 and 7.41 μm bands that are observed in many astronomical infrared spectra such as W 33A (Fig. 13). W 33A is a high-mass YSO with particularly rich solid-state molecular features (Dartois et al. 1999; Gibb et al. 2000; Keane et al. 2001), with abundant quantities of methanol, which is a signature of an advanced solid-state chemistry. In the bottom panel of Fig. 13 we display the laboratory infrared spectrum of NH₂CH₂OH in the 5−8 μm range. NH₂CH₂OH has been proposed to be a possible carrier to the 6 μm and 6.85 μm bands. In particular its 6.15 μm band helps filling in the low-energy half of the 6 μm band (Bossa et al. 2009b). The 7 μm to 7.5 μm regions are also better explained. The same arguments can be used for the N-methylaminomethanol. The shape of the CH₃NHCH₂OH 7 μm band is similar to the CH₃OH band, but its 7.20 μm feature reproduces the 7.24 μm feature in the astronomical spectrum much better (Fig. 13). Although it is not possible to extract reliable abundances for N-methylaminomethanol by deconvolving the astronomical infrared spectra with so many carriers, we can propose that aminoalcohols must be accounted as a carrier of the 5−8 μm complex, and that their presence is the signature of a thermal processing of the ice. Furthermore, as many other complex molecules, their detection after sublimation in the hot core region by radio astronomy should be possible. The non-detection so far of these compounds in the gas phase comes from the lack of rotational spectra for these molecules. From an astrobiological point of view, aminoalcohols are interesting molecules because they are considered as amino acid precursors in the Strecker synthesis (Bossa et al. 2009b; Danger et al. 2011; Duvernay et al. 2010; Rimola et al. 2010; Vinogradoff et al. 2011). In our case the N-methylaminomethanol will form through the Strecker synthesis the corresponding amino acid, the sarcosine (Eq. (6)), which has been detected in meteorites (Burton et al. 2012), $\mathit{}\mathrm{}\mathit{}$(6) Another interesting aspect that has been underlined in this work is the possibility to induce isotopic fractionation through thermal reaction in ice as is observed in some astrophysical objects. Indeed, some meteoritic materials and interstellar dust particles possess enhancements and anomalies in their $\frac{{\mathrm{15}}^{}\mathit{N}}{{\mathrm{14}}^{}\mathit{N}}$ ratios. The variability of the $\frac{{\mathrm{15}}^{}\mathrm{N}}{{\mathrm{14}}^{}\mathrm{N}}$ ratio in these materials suggest, as with deuterium, a formation induced by low-temperature interstellar chemistry (Charnley & Rodgers 2002; Kerridge et al. 1987; Messenger & Walker 1997). To estimate the possible isotopic fractionation between the ¹⁴N methylamine and the corresponding ¹⁵N isotopologue in interstellar ices, we calculated with a reduced kinetic model (Eq. (7)) the $\frac{{\mathrm{15}}^{}\mathrm{N}}{{\mathrm{14}}^{}\mathrm{N}}$ isotopic ratio at 30 K. We chose for the initial conditions an ice mixture of H₂O/CH₂O/CH₃¹⁴NH₂/CH₃¹⁵NH₂ in a 100/5/1/2.8 × 10^-3 concentration ratio. The $\frac{{}^{\mathrm{15}}\mathrm{N}}{{}^{\mathrm{14}}\mathrm{N}}\mathrm{=}\mathrm{2.8}\mathrm{\times }{\mathrm{10}}^{-3}$ isotopic ratio is the value reported for ammonia in dense molecular clouds (Pizzarello & Williams 2012). The simulation is displayed in Fig. 14. At 30 K, the isotopic ratio for methylamine in ice exponentially increases and reaches the value of 9 × 10^-1 after 0.5 × 10⁶ years. This result clearly shows that isotopic fractionation at low temperature, induced by thermal reaction, is efficient and has to be taken into account in more sophisticated astrophysical models. Moreover, the ¹⁵N enrichment observed in this simulation is consistent with the increase observed in the ¹⁵N/¹⁴N ratio between interstellar grains and cometary and meteoritic grains, $\begin{array}{ccc}\{\begin{array}{c}\\ \\ \\ \\ \end{array}& & \end{array}$(7)

Fig. 14 Evolution of the $\frac{{\mathrm{15}}^{}\mathrm{N}}{{\mathrm{14}}^{}\mathrm{N}}$ isotopic ratio in methylamine as a function of time at 30 K, considering only the methylamine and formaldehyde reaction.

5. Conclusion

We have investigated the thermal formation of N-methylaminomethanol (CH₃NHCH₂OH) formation in water ices from formaldehyde (H₂CO) and ¹⁴N, or ¹⁵N methylamine (CH₃¹⁴NH₂, and CH₃¹⁵NH₂). We showed that the N-methylaminomethanol formation in interstellar or cometary ices is likely and proposed this molecule as a possible candidate for some infrared features observed in W 33A. In addition, the kinetics study performed with ¹⁵N methylamine shows that these thermal reactions with a low barrier (1.1 ± 0.05 kJ mol^-1) are efficient in producing ¹⁵N/¹⁴N isotopic fractionation. We also investigated the N-methylaminomethanol stability in UV radiation and showed that this molecule leads to the formation of the N-methylformamide. This work provides additional evidences for thermal reactions in interstellar ices as a source of molecular complexity in the ISM. N-methylaminomethanol should be detected in the gas phase when the ices have sublimated, for instance in a hot core region.

Acknowledgments

This work has been founded by the French national program Physique Chimie du Milieu Interstellaire (PCMI), the Centre National des Études Spatiales (CNES), and the Provence-Alpes-Côte d’Azur region (PACA).

References

All Tables

All Figures

	Fig. 1 FTIR spectrum of (a) H2O/H2CO/CH3NH2 = 1/0.1/2 ice mixture deposited at 30 K; (b) pure H2O ice deposited at 30 K; (c) pure H2CO ice deposited at 30 K; (d) pure CH3NH2 deposited at 20 K.
In the text

	Fig. 2 FTIR spectra of (a) H2O/H2CO/CH3NH2 = 1/0.1/2 ice mixture deposited at 30 K and recorded at 30 K; (b) recorded at 70 K; (c) recorded at 190 K; (d) pure aminomethanol (NH2CH2OH) recorded at 190 K.
In the text

	Fig. 3 FTIR spectrum of pure CH3NHCH2OH formed from the warming of an H2O/H2CO/CH3NH2 ice mixture compared with the corresponding B3LYP infrared spectrum.
In the text

	Fig. 4 Mass spectrum of CH3NHCH2OH 14N and 15N recorded at 210 K.
In the text

	Fig. 5 FTIR spectra of (a) H2O/H2CO/CH3NH2 = 1/0.1/0.05 ice mixture deposited at 30 K and recorded at 30 K; (b) recorded at 125 K; (c) recorded at 185 K; (d) recorded at 200 K.
In the text

Fig. 6 Formaldehyde thermal evolution as a function of temperature in water ice containing CH3NH214N (triangle), and 15N (square) (H2O/CH3NH2/CH2O = 1/2/0.1). The rate coefficients are given in the form k = ν0exp( − E/RT), where T is the ice temperature (K), ν0 the pre-exponential factor (s-1), and E the activation energy (J mol-1). The corresponding rate coefficients are k14N = 1.5 ± 0.1 × 10-2exp( − 1100 ± 50/RT) and k15N = 2.6 ± 0.3 × 10-2exp( − 1800 ± 80/RT).

In the text

	Fig. 7 Absorbance of CH3NHCH2OH as a function of UV irradiation time. Experimental data are fitted against a first-order decay.
In the text

	Fig. 8 Infrared spectra of the UV photolysis of pure CH3NHCH2OH 14N at 20 K.
In the text

	Fig. 9 Temperature-programmed desorption of pure N-methylaminomethanol with and without UV irradiation. The m/z 61 signal refers to the N-methylaminomethanol while the m/z 59 signal refers to the photo-product: the N-methylformamide.
In the text

	Fig. 10 H2CO (full triangle), NH3 (full circle), CO2 (full diamond), CH3NH2 (full square) simulated abundances in ice as a function of time at 10 K and in a UV flux of 103 photons cm-2 s-1 without taking into account diffusion processes (left panel), and at 30 K in a UV flux of 103 photons cm-2 s-1 with diffusion processes included (right panel).
In the text

	Fig. 11 N-methylaminomethanol (CH3NHCH2OH) (full square) and N-methylformamide (CH3NHCHO) (open square) simulated abundances in ice as a function of time at 10 K and in a UV flux of 103 photons cm-2 s-1 without taking into account diffusion processes (left panel), and at 30 K in a UV flux of 103 photons cm-2 s-1 with diffusion processes included (right panel).
In the text

	Fig. 12 Infrared spectra of CH3OH, CH3NHCH2OH, NH2CH2OH recorded at 10 K. The shaded areas indicate the regions of the strong interstellar H2O and silicate absorptions.
In the text

	Fig. 13 Comparison between W 33A and laboratory spectra of CH3OH, CH3NHCH2OH, and NH2CH2OH in the 5.5−7.7 μm range.
In the text

	Fig. 14 Evolution of the $\frac{{\mathrm{15}}^{}\mathrm{N}}{{\mathrm{14}}^{}\mathrm{N}}$ isotopic ratio in methylamine as a function of time at 30 K, considering only the methylamine and formaldehyde reaction.
In the text