[tel-00726959, v1] Caractériser le milieu interstellaire ... - HAL - INRIA

J. R. Goicoechea et al.: Low sulfur depletion in the Horsehead PDR 579tel-00726959, version 1 - 31 Aug 2012still more abundant than neutral sulfur. These results are consistentwith our PdBI detection of CS close to the PDR edge andshow that CS is a PDR tracer. These findings are consistent withobservations of S-bearing species in the diffuse ISM where CSis more abundant than SO 2 ,H 2 S and SO (Lucas & Listz 2002).Between A V ∼ 2and∼8magtheS + abundance smoothly decreases.Since S + is a good source of electrons, the electronicfractionation also decreases accordingly. HCS + , and thus CS,present an abundance minimum in these layers. Neutral atomicsulfur is now the most abundant S-bearing species. Therefore,observations of the [S i]25 µm fine structure line will basicallytrace these intermediate layers of gas where S-bearing moleculeshave not reached their abundance peak. However, the exctitationenergy of the [S i]25 µm line (the upper level energy is ∼570 K)is too high compared to the thermal energy available in the regionswhere the neutral sulfur abundance peaks (T k ≃ 30 K) andno detectable emission is expected. In fact, no Spitzer/IRS linedetection has been reported in the Horsehead (L. Verstraete, privatecom.). However, since most of the neutral atomic sulfurwill remain in the ground-state, the presence of a backgroundIR source (e.g. in face-on PDRs) may allow, with enough spectralresolution and continuum sensitivity, the detection of the[S i]25 µm line in absorption.On the other hand, sulfur in diffuse ionized gas outside themolecular cloud is in the form of sulfur ions. Mid-IR [S iii]fine structure lines have been detected around the Horseheadwith IRS/Spitzer (L. Verstraete, private com.). In the shieldedgas, sulfur is mostly locked in S-bearing molecules togetherwith a smaller fraction in atomic form. Our models predictthat species such as SO will be particularly abundant. Jansenet al. (1995) also noted that the low gas phase sulfur abundanceneeded to explain the CS abundance in the Orion Bar PDR wasincompatible with the observed high H 2 S/CS ∼ 0.5 abundanceratio. Therefore, a complete understanding of the sulfur chemistrywill only be achieved when all the major sulfur moleculescan be explained. In a forthcoming paper we analyse the photochemistry,excitation and radiative transfer of several S-bearingmolecules detected by us in the Horsehead PDR.7. Summary and conclusionWe have presented interferometric maps of the Horsehead PDRin the CS J = 2–1 line at a 3.65 ′′ × 3.34 ′′ resolution together withsingle-dish observations of several rotational lines of CS, C 34 Sand HCS + . We have studied the CS photochemistry, excitationand radiative transfer using the latest HCS + and OCS + dissociativerecombination rates (Montaigne et al. 2005) and CS collisionalcross-sections (Lique et al. 2006). The main conclusionsof this work are as follows:1. CS and C 34 S rotational line emission reveals mean densitiesaround n(H 2 ) = (0.5–1.0) × 10 5 cm −3 .TheCSJ = 5–4 linesshow narrower line widths than the low-J CS lines and requirehigher density gas components, ∼(2–6) × 10 5 cm −3 ,not resolved by a ∼10 ′′ beam. These values are larger thanprevious estimates based on CO observations. It is likely thatclumpiness at scales below ∼0.01 pc and/or a low density envelopeplay a role in the CS line profile formation.2. Nonlocal, non-LTE radiative transfer models of opticallythick CS lines and optically thin C 34 S lines provide an accuratedetermination of the CS abundance, χ(CS) = (7 ± 3) ×10 −9 . We show that radiative transfer and opacity effects playa role in the resulting CS line profiles but not in C 34 S lines.Assuming the same physical conditions for the HCS + molecularion, we find χ(HCS + ) = (4 ± 2) × 10 −11 .3. According to photochemical models, the gas phase sulfurabundance required to reproduce these CS and HCS + abundancesis S/H = (3.5 ± 1.5) × 10 −6 , only a factor ∼4 lessabundant than the solar elemental abundance. Larger sulfurabundances are possible if the gas is significantly warmer.Thus, the sulfur abundance in the PDR is very close to theundepleted value observed in the diffuse ISM. The predictedCS/HCS + abundance ratio is close to the observed valueof ∼175, especially if predicted HCS + peak abundances areconsidered. If not, the HCS + production is underestimatedunless the gas is in a higher ionization phase, e.g. if the cosmicray ionization rate is increased by ∼5.4. High angular resolution PdBI maps reveal that the CS emissiondoes not follow the same morphology shown by thesmall hydrocarbons emission in the PDR edge. In combinationwith previous PdBI C 18 O observations we have modeledthe PDR edge and confirmed that a steep density gradientis needed to reproduce CS and C 18 O observations. The resultingdensity profile qualitatively agrees to that predictedin numerical simulations of a shock front compressing thePDR edge to high densities, n(H 2 ) ≃ 10 5 cm −3 , and high thermalpressures, ≃(5–10) × 10 6 Kcm −3 .5. Conventional PDR heating and cooling mechanisms fail toreproduce the temperature of the warm gas observed in theregion by at least a factor ∼2. Additional mechanical heatingmechanisms associated with the gas dynamics may beneeded to account for the warm gas. The thermal structureof the PDR edge is still not fully constrained from observations.This fact adds uncertainty to the abundances predictedby photochemical models.We have shown that many physical and chemical variations inthe PDR edge occur at small angular scales. In addition, themolecular inventory as a function of the distance from the illuminatingsource can only be obtained from millimeter interferometricobservations. High angular resolution observations containdetailed information about density, temperature, abundanceand structure of the cloud, but only detailed radiative transfer andphotochemical models for each given source are able to extractthe information. A minimum description of the source geometryis usually needed. Future observations with ALMA will allowus to characterize in much more details many energetic surfacessuch as PDRs.Acknowledgements. We are grateful to the IRAM staff at Plateau de Bure,Grenoble and Pico Veleta for the remote observing capabilities and competenthelp with the observations and data reduction. We also thank BASECOL, for thequality of data and information provided, and F. Lique for sending us the CScollisional rates prior to publication. JRG thanks J. Cernicharo, F. Daniel andI. Jiménez-Serra for fruitful discussions. We finally thank John Black, our referee,for useful and encouraging comments. JRG was supported by the frenchDirection de la Recherche and by a Marie Curie Intra-European IndividualFellowship within the 6th European Community Framework Programme, contractMEIF-CT-2005-515340.ReferencesAbergel, A., Bernard, J. P., Boulanger, F., et al. 2002, A&A, 389, 239Abergel, A., Teyssier, D., Bernard, J. P., et al. 2003, A&A, 410, 577Anthony-Twarog, B. J. 1982, A&J, 87, 1213Asplund, M., Grevesse, N., & Sauval, A. J. 2005, in Cosmic Abundances asRecords of Stellar Evolution and Nucleosynthesis, ed. F. N. Bash, & T. G.Barnes, ASP Conf. Ser., 336, 25Batrla, W., & Wilson, T. L. 2003, A&A, 408, 231Bernes, C. 1979, A&A, 73, 67Bogey, M., Demuynck, C., & Destombes, J. L. 1981, Chem. Phys. Lett., 81, 256