EDP Sciences
Free Access
Issue
A&A
Volume 596, December 2016
Article Number L1
Number of page(s) 10
Section Letters
DOI https://doi.org/10.1051/0004-6361/201629913
Published online 22 November 2016

© ESO, 2016

1. Introduction

Conformational isomerism refers to isomers (molecules with the same formula but different chemical structure) having the same chemical bonds but different geometrical orientations around a single bond. Such isomers are called conformers. An energy barrier often limits the isomerization. This barrier can be overcome by light. Photoisomerization (or photoswitching) has been studied in ice IR-irradiation experiments (e.g., Maçôas et al. 2004), in biological processes, and, for large polyatomic molecules, in gas-phase experiments (Ryan & Levy 2001). HCOOH is the simplest organic acid and has two conformers (trans and cis) depending on the orientation of the hydrogen single bond. The most stable trans conformer was the first acid detected in the interstellar medium (ISM; Zuckerman et al. 1971). Gas-phase trans-HCOOH shows moderate abundances towards hot cores (Liu et al. 2001) and hot corinos (Cazaux et al. 2003), in cold dark clouds (Cernicharo et al. 2012), and in cometary coma (Bockelée-Morvan et al. 2000). Solid HCOOH is present in interstellar ices (Keane et al. 2001) and in chondritic meteorites (Briscoe & Moore 1993).

The ground-vibrational state of cis-HCOOH is 1365 ± 30 cm-1 higher in energy than the trans conformer (Hocking 1976). The energy barrier to internal rotation (the conversion from trans to cis) is approximately 4827 cm-1 (Hocking 1976), approximately 7000 K in temperature units. This is much higher than the thermal energy available in molecular clouds (having typical temperatures from approximatley 10 to 300 K). Generalizing this reasoning, only the most stable conformer of a given species would be expected in such clouds. Photoswitching, however, may be a viable mechanism producing the less stable conformers in detectable amounts: a given conformer absorbs a high-energy photon that radiatively excites the molecule to electronic states above the interconversion energy barrier. Subsequent radiative decay to the ground-state would leave the molecule in a different conformer.

In this work we searched for pure rotational lines of the trans- and cis-HCOOH conformers in the 3 millimetre spectral band. We observed three prototypical interstellar sources known to display a very rich chemistry and bright molecular line emission: (i) the Orion Bar photodissociation region (PDR): the edge of the Orion cloud irradiated by ultraviolet (UV) photons from nearby massive stars (e.g., Goicoechea et al. 2016); (ii) the Orion hot core: warm gas around massive protostars (e.g., Tercero et al. 2010); and (iii) Barnard 1-b (B1-b): a cold dark cloud (e.g., Cernicharo et al. 2012). The two latter sources are shielded from strong UV radiation fields. We only detect cis-HCOOH towards the Orion Bar. This represents the first interstellar detection of the conformer.

thumbnail Fig. 1

Detection of cis-HCOOH towards the FUV-illuminated edge of the Orion Bar. Left: 13CO J = 3 → 2 integrated emission image with a HPBW of 8′′ obtained with the IRAM-30 m telescope (Cuadrado et al., in prep.). The cyan contour marks the position of neutral cloud boundary traced by the O i  1.317 μm fluorescent line emission (in contours from 3 to 7 by 2 × 10-4 erg s-1 cm-2 sr-1; Walmsley et al. 2000). Right: Cis- and trans-HCOOH stacked spectra towards the observed positions.

Open with DEXTER

2. Source selection and observations

Because of its nearly edge-on orientation, the Orion Bar PDR is a template source to study the molecular content as the far-UV radiation field (FUV; stellar photons with energies below 13.6 eV, or wavelengths (λ) longer than 911 Å, the hydrogen atom ionisation threshold) is attenuated from the cloud edge to the interior (Hollenbach & Tielens 1999). The impinging FUV radiation field at the edge of the Bar is approximatley 4 × 104 times the mean interstellar radiation field (e.g., Goicoechea et al. 2016, and references therein). We observed three positions of the Bar characterized by a decreasing FUV photon flux.

We have used the IRAM-30 m telescope (Pico Veleta, Spain) and the 90 GHz EMIR receiver. We employed the Fast Fourier Transform Spectrometer (FFTS) backend at 200 kHz spectral resolution (0.7 km s-1 at 90 GHz). Observations towards the Orion Bar are part of a complete millimetre (mm) line survey (80360 GHz, Cuadrado et al. 2015). They include specific deep searches for HCOOH lines in the 3 mm band towards three different positions located at a distance of 14′′, 40′′, and 65′′ from the ionisation front (Fig. 1A). Their offset coordinates with respect to the α2000 = 05h35m20.1s, δ2000 = − 05°25′07.0′′ position at the ionisation front are (+10′′, –10′′), (+30′′, –30′′), and (+35′′, –55′′). The observing procedure was position switching with a reference position at (600′′, 0′′) to avoid the extended emission from the Orion molecular cloud. The half power beam width (HPBW) at 3 mm ranges from ~30.8′′ to 21.0′′. We reduced and analyzed the data using the GILDAS software as described in Cuadrado et al. (2015). The antenna temperature, , was converted to the main beam temperature, TMB, using TMB = , where ηMB is the antenna efficiency (ηMB = 0.870.82 at 3 mm). The rms noise obtained after 5 h integration is ~15 mK per resolution channel.

We also searched for HCOOH in regions shielded from strong FUV radiation fields (see Appendix E). We selected two chemically rich sources for which we have also carried out deep mm-line surveys with the IRAM-30 m telescope: towards the hot core in Orion BN/KL (Tercero et al. 2010) and towards the quiescent dark cloud Barnard 1-b (B1-b; Cernicharo et al. 2012).

3. Results

3.1. Line identification

We specifically computed the cis-HCOOH rotational line frequencies by fitting the available laboratory data (Winnewisser et al. 2002) with our own spectroscopic code, MADEX (Cernicharo 2012). The standard deviation of the fit is 60 kHz. For the trans conformer, higher-frequency laboratory data (Cazzoli et al. 2010) were also used in a separate fit. The standard deviation of the fit for trans-HCOOH is 42 kHz. These deviations are smaller than the frequency resolution of the spectrometer we used to carry out the astronomical observations. Formic acid is a near prolate symmetric molecule with rotational levels distributed in different Ka rotational ladders (Ka = 0, 1, 2...). Both a- and b-components of its electric dipole moment μ exist (Winnewisser et al. 2002). The dipole moments of the cis conformer (μa = 2.650 D and μb = 2.710 D, Hocking 1976) are stronger than those of the trans conformer (μa = 1.421 D and μb = 0.210 D, Kuze et al. 1982).

In total, we identify 12 rotational lines of cis-HCOOH and 10 of trans-HCOOH above 3σ towards the FUV-illuminated edge of the Orion Bar, (+10′′, –10′′) position. The detected lines from the cis- and trans-HCOOH are shown in Figs. 2 and E.1, respectively. Lines attributed to HCOOH show a Gaussian line profile centred at the systemic velocity of the Orion Bar (10.4 ± 0.3 km s-1). Lines are narrow, with line widths of 1.9 ± 0.3 km s-1. The large number of detected lines, and the fact that none of the lines correspond to transitions of abundant molecules known to be present in the Bar or in spectroscopic line catalogues, represents a robust detection of the cis conformer. The observational parameters and Gaussian fit results are tabulated in Tables F.1 and F.2 for the cis and trans conformer, respectively.

3.2. Line stacking analysis

Complex organic molecules have relatively low abundances in FUV-irradiated interstellar gas (Guzmán et al. 2014). Indeed, detected trans-HCOOH lines are faint. To improve the statistical significance of our search towards the positions inside the Bar, we performed a line stacking analysis. For each observed position, we added spectra at the expected frequency of several HCOOH lines that could be present within the noise level (sharing similar rotational level energies and Einstein coefficients). The spectra in frequency scale were first converted to local standard of rest (LSR) velocity scale and resampled to the same velocity channel resolution before stacking. We repeated this procedure for trans-HCOOH lines. This method allows us to search for any weak line signal from the two conformers that could not be detected individually.

Figure 1B shows a comparison of the stacking results for cis and trans-HCOOH lines towards the three target positions in the Bar. Although we detect trans-HCOOH in all positions, emission from the cis conformer is only detected towards the position located closer to the cloud edge, (+10′′, –10′′). They demonstrate that cis-HCOOH is detected close to the FUV-illuminated edge of the Bar, but that the emission disappears towards the more shielded cloud interior.

A similar stacking analysis was carried out for the Orion hot core and B1-b spectra. Although we detect several trans-HCOOH lines, the cis conformer was not detected towards the hot core and the cold dark cloud (see Appendix E).

thumbnail Fig. 2

Detected cis-HCOOH rotational lines towards the Orion Bar, (+10′′, –10′′) position. The ordinate refers to the intensity scale in main beam temperature units, and the abscissa to the LSR velocity. Line frequencies (in GHz) are indicated at the top-right of each panel together with the rotational quantum numbers (in blue). The red curve shows an excitation model that reproduces the observations.

Open with DEXTER

3.3. Trans-to-cis abundance ratios

Given the number of HCOOH lines detected towards the Bar, we can determine the column density and rotational temperatures of both conformers accurately (see Appendix D). In particular, we infer a low trans-to-cis abundance ratio of 2.8 ± 1.0. The non-detection of cis-HCOOH towards the Orion hot core and B1-b (see Appendix E) provides much higher trans-to-cis limits (>100 and >60, respectively). This suggests that the presence of cis-HCOOH in the Orion Bar PDR is related to the strong FUV field permeating the region.

thumbnail Fig. 3

Ab initio absorption cross-sections and photoisomerization probabilities computed in this work. Top panel: trans- and cis-HCOOH absorption cross-sections for photons with E< 5 eV (those leading to fluorescence). Middle panels: normalized probabilities of bound-bound decays producing isomerization (trans cis and cis trans). Bottom panel: standard interstellar dust extinction curve (blue). Black and red curves show the effect of an increased PAH abundance.

Open with DEXTER

4. Photoisomerization rates and discussion

Photolysis of HCOOH has been widely studied both experimentally (Sugarman 1943; Ioannoni et al. 1990; Brouard & Wang 1992; Su et al. 2000) and theoretically (Beaty-Travis et al. 2002; He & Fang 2003; Maeda et al. 2015). Dissociation of HCOOH takes place after absorption of FUV photons with energies greater than ~5 eV (λ< 2500Å). Recently, Maeda et al. (2015) determined that this dissociation threshold coincides with the crossing of the S0 and T1 electronic states of the molecule. The specific products of the photofragmentation process (of the different photodissociation channels) depend on the specific energy of the FUV photons and on the initial HCOOH conformer. Interestingly, absorption of lower energy photons does not dissociate the molecule but induces fluorescent emission. In particular, HCOOH fluorescence from the S1 excited electronic state has been observed in laser-induced experiments performed in the λ = 2500−2700Å range (Ioannoni et al. 1990; Brouard & Wang 1992). These studies indicate that the geometrical configuration of the two hydrogen atoms is different in the S0 and S1 states. The fluorescence mechanism from the S1 state is a likely route for the transcis isomerization. In addition, the isomerization barrier from the S1 state (~1400 cm-1) is much lower than from the ground.

In order to quantify the role of the photoswitching mechanism, we carried out ab initio quantum calculations and determined the HCOOH potential energy surfaces of the S0 and S1 electronic states as a function of the two most relevant degrees of freedom, φ1 the torsional angle of OH and φ2, the torsional angle of CH (see Appendix A and Fig. A.1). With this calculation we can compute the position of the photon absorptions leading to fluorescence (those in the approximate λ = 2300−2800 Å range), and the probabilities to fluoresce from one conformer to the other (the trans-to-cis and cis-to-trans photoswitching cross-sections and probabilities, see Fig. 3).

With knowledge of Nph(λ), the FUV photon flux in units of photon cm-2 s-1Å-1, we can calculate the number of trans-to-cis and cis-to-trans photoisomerizations per second (ξtc and ξct, respectively, see Appendix B). In the absence of any other mechanism destroying HCOOH, the ξct/ξtc ratio provides the trans-to-cis abundance ratio in equilibrium. The time needed to reach the equilibrium ratio is then (ξtc + ξct)-1. Nph(λ), and thus ξtc and ξct, depend on the FUV radiation sources (type of star) and on the cloud position. Describing the cloud depth position in terms of the visual extinction into the cloud (AV), one magnitude of extinction is equivalent to a column density of approximatley 1021 H2 molecules per cm-2 in the line-of-sight.

In general (for a flat, wavelength-independent FUV radiation field), HCOOH photodissociation will always dominate over fluorescence (photodissociation cross-sections are larger and the relevant photons can be absorbed over a broader energy range, E> 5 eV). The strength and shape of the interstellar FUV radiation field, however, are a strong function of AV and are very sensitive to the dust and gas absorption properties. Because of the wavelength-dependence of the FUV-absorption process, Nph(λ) drastically changes as one moves from the cloud edge to the shielded interior. In particular, the number of low-energy FUV photons (e.g., below 5 eV) relative to the high-energy photons (e.g., those above 11 eV dissociating molecules such as CO and ionising atoms such as carbon) increases with AV. In this work we used a FUV radiative transfer and thermo-chemical model (Le Petit et al. 2006; Goicoechea & Le Bourlot 2007) to estimate Nph(λ) at different positions of the Orion Bar. The well-known 2175 Å bump of the dust extinction curve (absorption of λ = 1700−2500Å photons by PAHs and small carbonaceous grains, Cardelli et al. 1989; Joblin et al. 1992) greatly reduces the number of HCOOH dissociating photons relative to those producing HCOOH fluorescence (Fig. 3, bottom panel). The resulting FUV radiation spectrum, Nph(λ), at different AV is used to calculate ξtc and ξct (Table B.1). We determine that at a cloud depth of approximately AV = 2−3 mag, and if the number of HCOOH dissociating photons is small compared to the number of photons producing photoisomerization (i.e., most photons with E> 5 eV have been absorbed), the cis conformer should be detectable with a trans-to-cis abundance ratio of approximatley 3.54.1. These values are remarkably close to the trans-to-cis ratio inferred from our observations of the Bar.

Closer to the irradiated cloud edge (AV = 0−2 mag), photodissociation destroys the molecule much faster than the time needed for the trans-to-cis isomerization. On the other hand, too deep inside the cloud, the flux of E> 5 eV photons decreases to values for which the isomerization equilibrium would take an unrealistic amount of time (>106 yr for AV = 5 mag). Therefore, our detection of cis-HCOOH in irradiated cloud layers where CO becomes the dominant carbon carrier (a signature of decreasing flux of high-energy FUV photons) agrees with the photoswitching scenario.

For standard grain properties and neglecting HCOOH photodissociation, we calculate that the time needed to achieve a low trans-to-cis abundance ratio and make cis-HCOOH detectable at AV = 2−3 mag is 104−105 yr (see Table B.1). This is reasonably fast, and shorter than the cloud lifetime. In practice, it is not straightforward to quantify the exact contribution of HCOOH photodissociation and photoisomerization at different cloud positions. The above time-scales require that the flux of E> 5 eV dissociating photons is small compared to those producing fluorescence. This depends on the specific dust absorption properties, that sharply change with AV as dust populations evolve (Draine 2003), on the strength and width of the 2175 Å extinction bump, and on the role of molecular electronic transitions blanketing the FUV spectrum. The similar trans-HCOOH line intensities observed towards the three positions of the Bar (Fig. 1) suggest that even if the HCOOH photodestruction rate increases at the irradiated cloud edge, the HCOOH formation rate (from gas-phase reactions or desorbing directly from grain surfaces, Garrod et al. 2008) must increase as well. The inferred HCOOH abundances are not particularly high, (0.63.0) × 10-10 with respect to H. Hence, modest HCOOH photodestruction and formation rates are compatible with the photoswitching mechanism occurring in realistic times.

Although the observed abundances of trans- and cis-HCOOH in the Orion Bar are compatible with gas-phase photoisomerization, we note that photoswitching may also occur on the surface of grains covered by HCOOH ices. In a similar way, solid HCOOH (mostly trans) can absorb FUV photons that switch the molecule to the cis form before being desorbed. Once in the gas, both conformers will continue their photoisomerization following absorption of λ 2500 Å photons. Laboratory experiments are needed to quantify the mechanisms leading to HCOOH ice photoswitching by FUV photon absorption.

Searching for further support to the FUV photoswitching scenario, we qualitatively explored two other possibilities for the trans-to-cis conversion. First, the isomerization of solid HCOOH after IR irradiation of icy grain surfaces (as observed in the laboratory, Maçôas et al. 2004; Olbert-Majkut et al. 2008) and subsequent desorption to the gas-phase. Second, the gas-phase isomerization by collisions of HCOOH with energetic electrons (~0.5 eV). We concluded that if these were the dominant isomerization mechanisms, emission lines from cis-HCOOH would have been detected in other interstellar sources (see Appendix C).

Isomerization by absorption of UV photons was not considered as a possible mechanism to induce structural changes of molecules in interstellar gas. The detection of cis-HCOOH towards the Orion Bar opens new avenues to detect a variety of less stable conformers in Space. This could have broad implications in astrochemistry and astrobiology.


1

MOLPRO (Werner et al. 2012), version 2012, is a package of ab initio programs for advanced molecular electronic structure calculations, designed and maintained by H.-J. Werner and P. J. Knowles, and with contributions from many other authors (see http://www.molpro.net).

Acknowledgments

We thank N. Marcelino for helping with the observations of B1-b. We thank the ERC for support under grant ERC-2013-Syg-610256- NANOCOSMOS. We also thank Spanish MINECO for funding support under grants AYA2012-32032 and FIS2014-52172-C2, and from the CONSOLIDER-Ingenio program ASTROMOL CSD 2009-00038. IRAM is supported by INSU/CNRS (France), MPG (Germany), and IGN (Spain).

References

Appendix A: Ab initio estimation of fluorescence cross-sections and photoisomerization probabilities

In this appendix we demonstrate that the detected cis-HCOOH towards the Orion Bar can be produced by a gas-phase photoswitching mechanism. To estimate the FUV photon absorption cross-sections and probabilities of the trans-cis photoisomerization process, we start calculating the potential energy surfaces of the HCOOH S0 and S1 electronic states as a function of the two most relevant degrees of freedom, the torsional angle of OH (φ1), and the torsional angle of CH (φ2) (see Fig. A.1).

We performed ic-MRCI-F12 ab initio calculations using the MOLPRO suite of programs1 with the VDZ-F12 basis set. The obtained results agree with the stationary points previously reported by Maeda et al. (2012, 2015). The molecular orbitals and reference configurations were determined with a CASSCF calculation using 16 active orbitals. The optimized equilibrium geometries in the S0 and S1 electronic states are in agreement with previous results, corresponding to planar and bent trans-HCOOH conformers, respectively. They are listed in Table A.1. For trans-HCOOH, the normal modes in the S0 state have the following frequencies: 628.59, 662.86, 1040.64, 1117.90, 1316.0, 1416.22, 1792.32, 3083.01, and 3749.88 cm-1. The two lowest frequencies correspond to the torsional angles of the OH and CH bonds, respectively.

For the two lower singlet states, S0 and S1, we calculate a two-dimensional (2D) grid composed of 37 equally spaced points for φ1 and φ2, fixing the remaining coordinates to the corresponding values listed in Table A.1. These points are interpolated using a two-dimension splines method to get the potential energy surfaces, S0 and S1, at any desired geometry, including the two conformers.

The potential energy surface of the S0 electronic state presents two minima for φ2 = 0°, one at φ1 = 0° or 360° (trans), and a second at φ1 = 180° (cis). As illustrated in Fig. A.1, both minima correspond to a planar geometry. The potential for the S1 excited state presents two equivalent wells for the trans-conformer (φ1 = 300°, φ2 = 120° or φ1 = 60°, φ2 = 240°). Therefore, the minimum geometrical configuration in the S1 excited state is no longer planar. The cis conformer minimum transforms into a shoulder of the potential. This is shown in the one-dimensional (1D) cut shown in Fig. A.1 for the case of φ2 = 120°. In this case, the potential energy surface as a function of φ1 is relatively flat, while it shows a double-well structure as a function of φ2, corresponding to geometries above and below the molecular plane.

thumbnail Fig. A.1

1D potential energy surfaces of HCOOH as function of the OH torsional angle φ1. Bottom panel: ground S0 electronic state. Top panel: excited S1 state. 1D cuts were obtained from the 2D grid (see text) by setting φ2 = 180° and φ2 = 300° for S0 and S1, respectively. We also show the vibrational-wave functions obtained from a 1D model. The different geometrical structures of the HCOOH molecule in each energy minimum are shown.

Open with DEXTER

We solved the two-dimension Shrödinger equation for φ1 and φ2 and obtained the vibrational eigenfunctions. The first six vibrational levels of the S0 electronic state correspond to the trans conformer, the seventh energy level corresponds to the ground-vibrational state of cis-HCOOH. In the S1 excited electronic state, the presence of a double well as a function of φ2 implies that two degenerate vibrational states appear. The two well depths are different in geom-S0 and geom-S1 which means that their nodal structure changes significantly.

In a second step, we calculate the transition dipole moments for the 2D grids of geom-S0 and geom-S1, and determine the transitions between the S0 state and the S1 state. We derive the absorption spectra starting from both trans-HCOOH (ν = 0) and cis-HCOOH (ν = 7) in the S0 electronic ground-state, to the first 200 vibrational levels of the S1 excited-state. The use of different geometries in the two electronic states allows us to approximately reproduce the experimental frequencies (Beaty-Travis et al. 2002). The absorption spectrum is obtained using the transition dipole moments obtained for geom-S0.

Table A.1

Optimized geometries for trans-HCOOH in the ground (S0) and excited electronic state (S1).

The calculated radiative lifetimes of the different vibrational levels of the S1 electronic excited-state vary from 75 × 10-6 s to 375 × 10-6 s, but each level has a different probability to decay towards the trans of cis well of the S0 ground electronic state. We explicitly determine the probability to fluoresce into each conformer by calculating: (A.1)and (A.2)where we separate the contributions of the ν levels corresponding to the trans or cis conformers and normalize the sum to 1. We then normalize the above values and compute Pcis(ν)/(Pcis(ν)+Ptrans(ν)) and Ptrans(ν)/(Pcis(ν)+Ptrans(ν)) for ν levels corresponding to absorption energies below approximatley 40 000 cm-1 (E< 5 eV), approximately the energy for which the dominant photodissociation channel opens and fluorescence starts to become negligible.

In summary, with these ab initio calculations we estimate the cis- and trans-HCOOH cross-sections σλi for absorption of photons with energies lower than approximatley 40 000 cm-1 (those producing fluorescense). These absorptions radiatively excite the molecule to the S1 electronic excited-state. We explicitly compute the σλi values for each photon energy as well as the probabilites to fluoresce back to a specific cis or trans state (i.e., we determine the normalized probabilities of the different trans cis, trans trans, cis cis, cis trans bound-bound transitions). The σλi(trans) and σλi(cis) cross-sections and the Ptranscis and Pcistrans probabilities are plotted in Fig. 3.

Appendix B: Estimation of the photoisomerization rate in the Orion Bar

The number of photoisomerizations per second depends on the flux of FUV photons with energies below 5 eV. The trans-to-cis and cis-to-trans photoisomerization rates (ξtc and ξct) are derived from the discrete sums: (B.1)and (B.2)where σλi is the absorption cross-section from a given conformer (in cm2 photon-1) and P is the probability to fluoresce from one isomer to the other. Both quantities are determined from our ab initio calculations (previous section). Nph,λi (photon cm-2 s-1) is the flux of photons at each wavelength producing absorption.

In order to estimate the most realistic ξtc and ξct rates for the FUV-irradiation conditions in the Orion Bar, we used the Meudon PDR code (Le Petit et al. 2006) and calculate Nph(λ) at different cloud depth AV values. Following our previous studies of the Bar (Cuadrado et al. 2015; Goicoechea et al. 2016) we run a model of an isobaric PDR (Pth/k = 108 K cm-3) illuminated by χ = 4 × 104 times the mean interstellar radiation field (Draine 1978). For photons in the λ = 2000−3000 Å range, we adopt Nph(λ) = 4 × 104 × 732 × λ0.7 photon cm-2 s-1Å-1 at the PDR edge (AV = 0) (van Dishoeck & Black 1982). We use a constant dust composition and size distribution that reproduces a standard interstellar extinction curve (Cardelli et al. 1989).

Table B.1 shows the resulting photoisomerization rates at different cloud depths, the expected trans-to-cis HCOOH abundance ratio at equilibrium, and the time needed to reach the equilibrium ratios (neglecting photodissociation).

Table B.1

Photoisomerization rates for the irradiation conditions in the Orion Bar.

The use of constant dust grain properties through the PDR is likely the most important simplification for the calculation of the photoisomerization rates ξtc and ξct. Grain populations are known to evolve in molecular clouds, especially in PDRs where the sharp attenuation of a strong FUV field results in a stratification of the dust and PAH properties with AV (Draine 2003). Therefore, although the varying optical properties of grains are difficult to quantify and include in PDR models, they likely play a role on how FUV photons of different energies are differentially absorbed as a function of AV (Goicoechea & Le Bourlot 2007). For the particular case of HCOOH, the strength and width of the 2175 Å extinction bump (Cardelli et al. 1989) naturally divides the range of photons producing HCOOH photodissociation (those with E> 5 eV) from those producing fluorescence (E< 5 eV). The extinction bump has been related with the absorption of FUV photons by PAH mixtures and small carbonaceous grains (Joblin et al. 1992; Draine 2003). Although it is not known how the bump evolves with AV, it clearly determines how the lower-energy FUV photons are absorbed. In Fig. 3 (bottom panel) we show different extinction curves for different PAH abundances (Goicoechea & Le Bourlot 2007). Optical properties are taken from Li & Draine (2001) and references therein. In addition, and as in most PDR models, our predicted FUV spectrum does not include the absorption produced by hundreds of molecular electronic transitions blanketing the FUV spectrum (other than H2 and CO lines). Altogether, our assumption that the detected cis-HCOOH arises from PDR layers in which the flux of photons with λ> 2500 Å dominates over the higher-energy photodissociating photons is very plausible.

Appendix C: Alternative mechanisms for trans-to-cis isomerization in the ISM

Searching for further support to the photoswitching scenario, we qualitatively explored other possibilities that may apply in interstellar conditions. In the laboratory, trans-to-cis isomerization has been observed in molecular ices irradiated by near-IR photons (Maçôas et al. 2004; Olbert-Majkut et al. 2008). Hence, isomerization of solid HCOOH and subsequent desorption to the gas-phase might also be responsible of the cis-HCOOH enhancement. However, owing to the short lifetime of the cis conformer observed in ices (a few minutes if the irradiation is stopped, Maçôas et al. 2004), a very strong flux of IR photons would be needed to maintain significant abundances of solid cis-HCOOH. In addition, near-IR photons penetrate molecular clouds much deeper than FUV photons, and one would have expected to detect cis-HCOOH in all positions of the Bar, and towards the Orion hot core, a region irradiated by intense IR fields. Alternatively, the trans-to-cis isomerization might be triggered by collisions with electrons. Electrons are relatively abundant in FUV-irradiated environments (with ionisation fractions up to approximatley ne/nH ≈ 10-4) compared to shielded cloud interiors (ne/nH ≈ 10-9). Simple calculations show that electrons with energies of approximatley 0.5 eV would be needed to overcome the energy barrier to HCOOH isomerization, and to produce a trans-to-cis abundance ratio of approximatley 3. Such suprathermal electrons could be provided by the photoionisation of low ionisation potential atoms (C, S, Si...), but their abundance sharply decreases with AV (Hollenbach & Tielens 1999). We estimate that at a cloud depth of AV = 2 mag, HCOOH collisional isomerization, if effective, could compete with photoswitching only if the elastic collisional rate coefficients were very high; of the order of >10-6 cm3 s-1 for a typical electron density of ne< 1 cm-3 in PDRs. However, the detection of trans-HCOOH, but not cis-HCOOH, towards other PDRs such as the Horsehead (Guzmán et al. 2014), with similar electron densities but much lower FUV photon flux (>100 times less than the Bar), supports a photoswitching mechanism in the Orion Bar (i.e., high ξtc and ξct rates), but makes it too slow for the Horsehead and other low FUV-flux sources. Either way, we encourage laboratory and theoretical studies of the possible role of electron collisions, as well as more detailed investigation of the photoswitching mechanism of HCOOH and other species.

thumbnail Fig. C.1

Rotational population diagrams from the observed HCOOH lines towards the Orion Bar, (+10′′, –10′′) position. Left: diagram for the cis conformer (measurements lie along a single component). Right: diagram for the trans conformer showing how different Ka rotational ladders split in different components. Fitted values of the rotational temperature, Trot, and column density, N, are indicated in each panel (see also Table C.1).

Open with DEXTER

Table C.1

Rotational temperatures (Trot), column densities (N), and abundances towards the Orion Bar PDR, (+10′′,–10′′) position.

Appendix D: Rotational diagrams and column density calculation

Owing to the large number of detected HCOOH lines, we calculated rotational temperatures (Trot) and column densities (N) from rotational population diagrams. The standard relation for the rotational diagram (Goldsmith & Langer 1999) is: (D.1)with Nu/gu given by (D.2)In the above relation, Nu is the column density of the upper level in the optically thin limit [cm-2], N is the total column density [cm-2], gu is the statistical weight of the upper state of each level, QTrot is the rotational partition function evaluated at a rotational temperature Trot, Aul is the Einstein coefficient [s-1], Eu/k is the energy of the upper level of the transition [K], νul is the frequency of the ul transition [s-1], dv is the velocity-integrated line intensity corrected from beam efficiency [K km s-1], and ηbf is the beam filling factor. Assuming that the emission source has a 2D Gaussian shape, ηbf is equal to , with θB the HPBW of the telescope [arcsec] and θS the diameter of the Gaussian source [arcsec]. The values for νul, Eu/k, gu, and Aul are taken from the MADEX spectral catalogue.

Rotational diagrams were built considering two limiting cases: (i) that the detected HCOOH emission is extended, with ηbf = 1; and (ii) that the emission is semi-extended, with θS = 9″ (Cuadrado et al. 2015). In a plot of ln(Nu/gu) versus the energy of the upper level of each rotational transition, Eu/k, the population distribution approximatley follows a straight line with a slope –1/Trot. The total column density of the molecule, N, is obtained from the y-intercept and the partition function. Figure C.1 shows the resulting rotational diagrams assuming extended emission. Table C.1 lists the Trot and N obtained by linear least squares fits. The uncertainties shown in Table C.1 indicate the uncertainty obtained in the fit. The uncertainties obtained in the determination of the fit line parameters with CLASS are included in the error bars at each point of the rotational diagram.

To crosscheck that the relative intensities of the detected cis- and trans-HCOOH rotational lines are those expected according to their inferred rotational temperatures (i.e., that the assigned lines are not blended with lines from other molecules), we carried out a simple excitation and radiative transfer calculation using MADEX. We assumed that the cis- and trans-HCOOH rotational levels are populated following a Boltzmann distribution at a single rotational temperature (obtained from the rotational diagrams). For a given column density N, the model computes each line opacity (optically thin for the observed HCOOH lines) assuming a Gaussian line profile (for a linewidth of 2 km s-1) and simulates the output mm spectrum at a given spectral resolution. Figures 2 and E.1 show the observed spectra (black histograms) and the modeled lines (red curves) for the Trot and N values obtained assuming extended emission. The good agreement of the fits, and lack of any other candidate line from a different molecule in our catalogue, confirms that all detected lines belong to cis- and trans-HCOOH.

Appendix E: Non-detection of cis-HCOOH towards the Orion BN/KL hot core and Barnard-B1

We searched for cis-HCOOH in regions shielded from strong FUV radiation fields. We selected chemically rich sources for which we have also carried out deep mm-line surveys with the IRAM-30 m telescope. In particular, we searched for HCOOH towards the hot core in Orion BN/KL (Tercero et al. 2010) and towards the quiescent dark cloud Barnard 1-b (B1-b; Cernicharo et al. 2012). Although we clearly detected lines from trans-HCOOH towards both sources, we did not find lines from cis-HCOOH above the detection limit of these deep surveys. Using the MADEX excitation code, we derived lower limits to the trans-to-cis abundance ratio towards these sources. Below we summarize the main results from these observations:

Orion BN/KL hot core: the hot core is embedded in the Becklin-Neugebauer/Kleinmann-Low massive star-forming region, at ~4 North-West from the Orion Bar, and ~0.5 North-West from the Trapezium stars. Relatively narrow lines (ΔvFWHM ≈ 3 km s-1) corresponding to a-type transitions of trans-HCOOH, with upper level energies up to Eu/k 300 K, are detected at a LSR velocity of ~8 km s-1. The observed line parameters are consistent with emission from the hot core itself. This is dense, nH of a few 107 cm-3, and hot gas at approximatley 200 K (Blake et al. 1987; Tercero et al. 2010), and also from the more extended warm gas (approximatley 60 K) in the ambient molecular cloud, the so-called the extended ridge (Blake et al. 1987; Tercero et al. 2010). Using MADEX and our accumulated knowledge of the source structure (see Tercero et al. 2010; Cernicharo et al. 2016, and references therein), we determine Trot(trans) = 100 ± 35K and N(trans) = (1.0−0.3) × 1015 cm-2 in the hot core, and Trot(trans) = 40 ± 15 K and N(trans) = (1.0−0.3) × 1014 cm-2 in the extended ridge. We note that the extended ridge is the main cloud component responsible for the observed trans-HCOOH line emission in the 3 mm band. Although we obtain much higher trans-HCOOH column densities compared to the Orion Bar, lines from cis-HCOOH are not detected towards the hot core. Assuming Trot(trans) = Trot(cis), we compute a lower limit to the trans-to-cis abundance ratio of >100 in the hot core, and >30 in the extended ridge.

thumbnail Fig. E.1

Detected trans-HCOOH rotational lines towards the edge of the Bar, (+10′′, –10′′) position. The ordinate refers to the intensity scale in main beam temperature units, and the abscissa to the LSR velocity. Line frequencies (in GHz) are indicated at the top-right of each panel together with the rotational quantum numbers (in blue). The red curve shows an excitation model that reproduces the observations. Cis-HCOOH lines are shown in Fig. 2.

Open with DEXTER

B1-b cold cloud: Barnard 1 is a low-mass star-forming region located in the Perseus cloud. The cold core B1-b harbors two submillimetre continuum sources (B1-bN and B1-bS) identified as first hydrostatic core candidates (Gerin et al. 2015), and B1b-W, an infrared source detected with Spitzer (Jørgensen et al. 2006). Complex organic molecules such as CH3OCOH, CH3SH, and CH3O have been identified (Marcelino 2007; Öberg et al. 2010; Cernicharo et al. 2012). We detect nine lines from trans-HCOOH in the 3 mm band. A rotational diagram provides Trot(trans) = 12 ± 4 K and N(trans) = (1.5 ± 0.5) × 1012 cm-2. Figure F.1 shows the detected lines together with our best model fit (red curve). Lines are very narrow (ΔvFWHM 0.5 km s-1), consistent with emission from quiescent cold gas (approximatley 20 K) shielded from FUV radiation. Although the inferred trans-HCOOH column density is similar to that obtained towards the Orion Bar, we do not detect lines from cis-HCOOH at the noise level of the B1-b data. Assuming Trot(trans) = Trot(cis), we determine a lower limit to the trans-to-cis abundance ratio of >60. This is similar to that of the extended molecular Ridge of Orion, but significantly higher than towards the Bar.

Appendix F: Detected cis- and trans-HCOOH lines towards the Orion Bar PDR

thumbnail Fig. F.1

Detected trans-HCOOH rotational lines towards the cold cloud Barnard 1-b. The ordinate refers to the intensity scale in main beam temperature units and the abscissa to the Doppler velocity. Line frequencies (in GHz) are indicated at the top of each panel together with the rotational quantum numbers. The red curve shows an excitation model that reproduces the rotational population diagram. The bottom panel shows the stacked spectra for cis-HCOOH.

Open with DEXTER

Table F.1

Line parameters for cis-HCOOH towards the Orion Bar, (+10′′, –10′′) position.

Table F.2

Line parameters for trans-HCOOH towards the Orion Bar, (+10′′, –10′′) position.

All Tables

Table A.1

Optimized geometries for trans-HCOOH in the ground (S0) and excited electronic state (S1).

Table B.1

Photoisomerization rates for the irradiation conditions in the Orion Bar.

Table C.1

Rotational temperatures (Trot), column densities (N), and abundances towards the Orion Bar PDR, (+10′′,–10′′) position.

Table F.1

Line parameters for cis-HCOOH towards the Orion Bar, (+10′′, –10′′) position.

Table F.2

Line parameters for trans-HCOOH towards the Orion Bar, (+10′′, –10′′) position.

All Figures

thumbnail Fig. 1

Detection of cis-HCOOH towards the FUV-illuminated edge of the Orion Bar. Left: 13CO J = 3 → 2 integrated emission image with a HPBW of 8′′ obtained with the IRAM-30 m telescope (Cuadrado et al., in prep.). The cyan contour marks the position of neutral cloud boundary traced by the O i  1.317 μm fluorescent line emission (in contours from 3 to 7 by 2 × 10-4 erg s-1 cm-2 sr-1; Walmsley et al. 2000). Right: Cis- and trans-HCOOH stacked spectra towards the observed positions.

Open with DEXTER
In the text
thumbnail Fig. 2

Detected cis-HCOOH rotational lines towards the Orion Bar, (+10′′, –10′′) position. The ordinate refers to the intensity scale in main beam temperature units, and the abscissa to the LSR velocity. Line frequencies (in GHz) are indicated at the top-right of each panel together with the rotational quantum numbers (in blue). The red curve shows an excitation model that reproduces the observations.

Open with DEXTER
In the text
thumbnail Fig. 3

Ab initio absorption cross-sections and photoisomerization probabilities computed in this work. Top panel: trans- and cis-HCOOH absorption cross-sections for photons with E< 5 eV (those leading to fluorescence). Middle panels: normalized probabilities of bound-bound decays producing isomerization (trans cis and cis trans). Bottom panel: standard interstellar dust extinction curve (blue). Black and red curves show the effect of an increased PAH abundance.

Open with DEXTER
In the text
thumbnail Fig. A.1

1D potential energy surfaces of HCOOH as function of the OH torsional angle φ1. Bottom panel: ground S0 electronic state. Top panel: excited S1 state. 1D cuts were obtained from the 2D grid (see text) by setting φ2 = 180° and φ2 = 300° for S0 and S1, respectively. We also show the vibrational-wave functions obtained from a 1D model. The different geometrical structures of the HCOOH molecule in each energy minimum are shown.

Open with DEXTER
In the text
thumbnail Fig. C.1

Rotational population diagrams from the observed HCOOH lines towards the Orion Bar, (+10′′, –10′′) position. Left: diagram for the cis conformer (measurements lie along a single component). Right: diagram for the trans conformer showing how different Ka rotational ladders split in different components. Fitted values of the rotational temperature, Trot, and column density, N, are indicated in each panel (see also Table C.1).

Open with DEXTER
In the text
thumbnail Fig. E.1

Detected trans-HCOOH rotational lines towards the edge of the Bar, (+10′′, –10′′) position. The ordinate refers to the intensity scale in main beam temperature units, and the abscissa to the LSR velocity. Line frequencies (in GHz) are indicated at the top-right of each panel together with the rotational quantum numbers (in blue). The red curve shows an excitation model that reproduces the observations. Cis-HCOOH lines are shown in Fig. 2.

Open with DEXTER
In the text
thumbnail Fig. F.1

Detected trans-HCOOH rotational lines towards the cold cloud Barnard 1-b. The ordinate refers to the intensity scale in main beam temperature units and the abscissa to the Doppler velocity. Line frequencies (in GHz) are indicated at the top of each panel together with the rotational quantum numbers. The red curve shows an excitation model that reproduces the rotational population diagram. The bottom panel shows the stacked spectra for cis-HCOOH.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.