Research

Overview

Our current research focuses on the modelling of physical and chemical processes of astrophysical interest and its applications. In many cases, such as those encountered in the interstellar medium (ISM), local thermodynamic equilibrium is not reached. Atomic and molecular data then reveal crucial to determine the chemical composition and physical conditions of this medium. The first objective of our work is to obtain accurate data for processes (radiative, collisional and reactive) in which interstellar molecules are involved. The determination of these data also requires the development of appropriate theoretical methods, which is itself a very important complementary axis of research. The skills we have developed during these activities focus on three main areas. The first one relates to the use of quantum chemistry ab initio methods to model the interactions between atoms or molecules involved in the processes studied. The study of the dynamic of the nuclei submitted to these potentials is the second skill. The variety of molecular systems and processes that I faced, led me to use, or develop, a wide range of methods often based on quantum approaches. The third skill consists in the astrophysical modelling and especially in the sensitivity of astrophysical models to molecular data. A very important aspect of our work, which we are very attached to, is the validation of our theoretical approaches by comparison with the available experiments data. Actually, the choice of the molecular systems we are studying is governed by both the astrophysical interest and the availability of experimental data. Such stage is crucial in order to ensure the accuracy of the theoretical calculations.

To make the most of the very high resolution observations and experiments, the astronomers and the experimentalists need to be given very accurate molecular data. In addition, state-to-state data are most of the time required. Hence, the objective of our work is generally the computation of highly accurate cross sections and rate constants for small and medium-sized molecular systems. In our calculations, we generally consider the complex fine/hyperfine structure of the molecules and/or non-adiabatic effects. Hence, the calculation of ab initio potential energy surfaces (PES) between particles is principally done using highly correlated ab initiomethods like configuration interaction or coupled clusters methods. For the dynamics of nuclei, we are generally using and developing quantum close coupling methods that are the most suitable treatment at low energies/temperatures and that also allow state-to-state resolutions in our simulations.

Collisional excitation of interstellar molecules

In recent years, we have studied the collisional excitation of numerous interstellar molecules by H, He and H₂. At first, we provided collisional data for sulfur-bearing molecules such as SO, CS or SiS. More recently, I became interested in the collisional excitation of nitrogen-bearing molecules such as CN, HCN or HNC, which are excellent tracers of dense ISM. I was able to provide the first collisional data for the CN and HNC molecules colliding with He and H₂. These data were strongly requested by the astrophysics community since these molecules are frequently observed. Our research work allows solving the problem of the HCN/HNC abundance ratio in cold molecular clouds (see Part 2). Many of the molecules we have studied (CN, SO, HCN...) show complex structures due to the coupling between the rotational and the spin (electronic and/or nuclear). It was therefore necessary to develop molecular dynamics codes to study these systems.

Comparison between HCN-H₂ and HNC-H₂ O₂(N=1, j=0) + p-H₂→ O₂(N=1, j=1) p-H₂ cross sections:

rate coefficients at 50K Experiments (open circles); theory (solid line)

The results obtained for these molecules have yielded important consequences for astrophysical modelling. Thus, as we were able to compute collisional rate coefficients with He and with para- and ortho-H₂, we have shown that for most of the molecules studied (except hydrides), collisional rate coefficients with He can be a reasonable estimate of the rate coefficients with para-H₂. In contrast, rate coefficients with ortho-H₂clearly differ from those with He and/or para-H₂ so that it appears necessary to have collisional data with both para- and ortho-H₂ for the modelling of warm molecular clouds where ortho-H₂could be abundant. These results are summarized in a review published recently (Roueff & Lique, Chem. Rev. 113, 8906, 2013).

Another crucial aspect of this research work is the validation of our theoretical calculations by comparison with the experiments. When available, we always compared our results with experimental measurements of cross sections and/or rate coefficients. The agreement we generally obtained was very good. The theoretical calculations can reproduce most of the details (including the resonances) seen in the experiments as can be seen in our study of the collisional excitation of O₂ by H₂ that perfectly matches with the experimental findings. Such agreement shows that we can be confident in the accuracy of our calculations.

Chemical reactivity

The study of a chemical reaction is much more expensive in computing time than the study of inelastic collisions. So, we focused only on a few key chemical reactions for the ISM. The F+H₂ → HF+H reaction is notably important for the chemistry of the ISM since it is estimated that almost all the interstellar fluorine is contained in the HF molecule and the HF molecule is mainly formed through the F+H₂reaction.

This reaction is also, in addition to its astrophysical interest, a prototype for exothermic reactions due to its relatively simple experimental access. Thus, using highly accurate quantum treatment including several electronic states, we computed reactive cross sections and rate constants. We are the only researcher to study these reactions with such an approach. The agreement between theoretical results and experiments is excellent. All the experimental results could be reproduced with an absolute accuracy. In addition, the new results differ significantly from those currently used in astrochemical models describing the ISM evolution.

Temperature dependence of the F+H₂ rate constants Temperature dependence of the OH+H rate constants

The new F+H₂rate constants also allow improvement in the determination of the abundance of H₂and, thus of the total mass of interstellar clouds since the HF molecule is considered as a good tracer of H₂. Preliminary calculations show that the total mass of diffuse clouds could be reduced by up to a factor two.

Another key astrophysical reaction we have been studied is the O+OH → H+O₂ reaction. This reaction is the main source of molecular oxygen in cold interstellar clouds. Astrophysicists held as truth the understanding of this reaction mechanism until observations showed that O₂was actually a thousand times less abundant than anticipated by astrophysical models. The rate constant of this reaction is a crucial data for understanding the abundance of the O₂ molecule. Theoretical modelling is very difficult to perform because of the presence of two heavy atoms (oxygen) in the reactants. However, we managed to overcome this difficulty and to obtain data with an accuracy comparable to reactions involving only one heavy atom. The results suggest that O+OH reaction is slower than expected and that the rate constants given in astrophysical databases are incorrect.

Recently, we were interested in the ortho-para-H₂ conversion process in the ISM and in the early universe. Using exact quantum time-independent approach, we modelled the process that takes place by hydrogen exchange. The new results, in excellent agreement with the experiments, will allow better modelling of this key process that was surprisingly not precisely known. We have also investigated the formation and destruction of the HCl molecule in the ISM via the Cl+H₂→ HCl+H reaction. The calculations were performed using a close coupling approach and based on highly accurate PES. The results pushed to a major revision of the chlorine chemistry: contrary to what was expected, HCl is not the main chlorine career in the ISM.

Astrophysical modelling

We were first interested to know the impact of the new collisional data on the radiative transfer models. Then, using a Large Velocity Gradient (LVG) and a nonlocal Monte Carlo radiative transfer code, we have studied the excitation of molecules like SO, CS or NO. We compared the brightness temperatures (line intensities) calculated using our new rate coefficients to those obtained using data previously available. For CS, the brightness temperatures obtained showed significant differences (up to a factor 2) compared to those obtained with the old data. The use of previous rate coefficients can therefore lead to a significant error in the interpretation of CS observations in molecular clouds. For the SO molecule, the differences can reach a factor of 10 and are highly dependent on the molecular transition. Hence, the determination of the abundance of SO in cold molecular clouds can result, according to the observed transition, in an over/under-estimation if the old rates are used. These studies have also allowed us to show that radiative transfer calculations are highly sensitive to the accuracy of the rate coefficients and that the astronomers really benefit from accurate data.

More recently, we worked on the excitation of HNC in cold molecular clouds. As we have computed the first rate coefficients for this important molecule, we have checked the impact of the new data on the astrophysical modelling. We found that the use of our new rates completely changes the determination of the HNC abundance in the ISM. Thus, the new data offer a solution to the unexpected HNC/HCN abundance ratio in the ISM: for a typical cold molecular cloud, the abundance of HNC is 2.5 times lower than what was previously determined using collisional data of HCN (what was done by the astronomers up to our calculations). The HNC/HCN ratio is now ~ 1, a value in better agreement with astrochemical models prediction.

Recently, we also worked on the impact of using H₂-rate coefficients instead of He ones. The calculations of H₂-rate coefficients is far more complex than the calculations of He-rate coefficients and this is why, up to recently, most of the collisional data were computed using He as a collisional partner. In our studies, we found that this approach may be valid for heavy molecules but it is much more questionable when dealing with molecular hydrides. For example, for the HCl molecule, we found that the use of H₂-rate coefficients compared to He-rate coefficients leads to a decrease of the HCl abundance by a factor 10 in star forming regions. Then, we strongly recommend the computation of H₂ rate coefficients for hydrides when it is feasible.

Line intensity (in K) of the 1-0 transitions of HCl (T= 50 K; First detection of the CN^- molecule in IRC+10216

n(H₂)= 10⁶ cm⁻³) using different sets of rate coefficients

Simultaneously, we have interpreted millimeter observations of SO and HCl from the TMC-1 molecular cloud and OMC-2 protostellar core, respectively. We have determined the physical conditions in these clouds. As expected, significant differences, mainly explained by the use of new rate coefficients, exist with previous studies. In particular, the abundance of HCl and SO molecules in these media is significantly lower than previously found. Such examples show the clear impact of data accuracy in astrophysical modelling.

We participated in the first detection of the CN^- molecule in the ISM by providing data for interpreting the observations and determining its abundance. We were also implied in the first detection of OH⁺ in planetary nebulae. Finally, we were also interested in the abundance of metal-bearing molecules in the circumstellar gas and we have shown the evidence of selective cyanide chemistry.

Main collaborations

• M. H. Alexander, J. Klos, University of Maryland (USA)

• P. J. Dagdigian, Johns Hopkins University, Baltimore (USA)

• S. Marinakis, Queen Mary University, London (England)

• D. R. Flower, Dunham University (England)

• H.-J. Werner, University of Stuttgart (Germany)

• S. Y. T. van de Meerakker, Nijmegen University (Netherlands)

• G. Chalasinski, University of Varsovie (Poland)

• O. Roncero, M.-L. Senent,CSIC, Madrid (Spain)

• S. Gómez-Carrasco, D. González-Sánchez, University of Salamanca (Spain)

• N. Bulut, Firat University (Turkey)

• N. Jaidane, D. Ben Abdallah, K. Hammami, O. Yazidi, University of Tunis (Tunisia)

• F. Dumouchel, I. F. Schneider, Université du Havre (France)

• L. Tchang-Brillet, N. Feautrier, F. Dayou, Observatoire de Meudon (France)

• T. Stoecklin, P. Halvick, M. Costes, C. Naulin, A. Bergeat, Université de Bordeaux (France)

• M. Hochlaf, Université de Marne la Vallée (France)

• A. Faure, L. Wiesenfeld, Observatoire de Grenoble (France)

• M. Monnerville, Université de Lille (France)

• P. Honvault, B. Honvault, Université de Dijon (France)

• I. R. Sims, S. D. Le Picard, L. Biennier, S. Carles, R. Georges, Université de Rennes (France)

• J. Cernicharo, M. Agundez, CSIC, Centro de Astrobiologia, Madrid (Spain)

• F. F. van de Tak, University of Groningen(Netherlands)

• R. Loughnane, Universidad Nacional Autónoma de México(Mexico)

• M. Kama, Leiden Observatory (Netherlands)

• L. Pagani, E. Roueff, Observatoire de Paris (France)

• C. Ceccarelli, A. Bacman, S. Maret, P. Hily-Blant, F. Daniel, Observatoire de Grenoble (France)

• E. Caux, S. Bottinelli, Université de Toulouse (France)