Abstract. A brief overview of JWST observations of planet-forming disks taken as part of the MINDS guaranteed time program is presented.
Introduction
Over the past decade we have revised our picture of when and how planets form in disks around young stars. The Atacama Large Millimeter Array (ALMA) is measuring how much mass resides in disks in form of solids and gas (Drazkowska et al. 2022, Bae et al. 2022) and showing a plethora of substructure that we generally associate with ongoing planet formation processes. This has now convinced the community that the first steps of forming planetesimals and planetary cores must happen very early, within ~1 Myr. A similar conclusion is also reached by studying the Solar System evidence from Asteroids (e.g. Kruijer et al. 2014, Mezger et al. 2020).
Based on our own Solar System, we think of the inner few au as the terrestrial planet-forming region around a young Solar-type star. For stars with masses of ~1/10th of the Sun, this region could even be closer. With the advent of JWST, we can study the dust and gas composition in these regions around large samples of young stars with unprecedented sensitivity, thus enabling us to confront our theories of disk evolution during the main planet formation phase with detailed observational data. The mid-infrared wavelength range covers rovibrational emission of many simple abundant molecules such as CO, OH, H2O, CO2, HCN, C2H2, which had already been detected by the Spitzer Space Telescope (e.g. Carr & Najita 2008, Salyk et al. 2008, Pascucci et al. 2009, Pontoppidan et al. 2010). However, that wavelength range also contains emission from less abundant molecules such as the rarer isotopologues of these molecules, NH3, CH4, and a large range of simple hydrocarbons. Thermo-chemical disk models suggest a layered molecular structure with OH and CO at the top, followed by H2O and CO2. HCN and C2H2 peak in abundance deeper in the disk (e.g. Woitke et al. 2018, see Fig. 1). Hence, the potential richness of these mid-infrared spectra provides also an excellent testbed to study the warm and dense chemistry in these inner regions of planet forming disks.
The JWST GTO MIRI Mid-INfrared Disk Survey (MINDS, PI: Th. Henning) observed 5 Herbig disks (two of which close to edge-on), 33 T Tauri disks (four of which edge-on), 10 disks around very low-mass stars (VLMS, one close to edge-on, 2MASS- J04381486+2611399), and 5 young debris disks with detected CO sub-mm emission (Henning, Kamp & MINDS team 2024). The ages of our disks span ~1 Myr to several 10 Myr. At these ages, terrestrial planet formation is likely still ongoing, based on Earth studies who show that ~40-50% of its mass was accumulated late (>4-5 Myr, Lammer et al. 2021).
No two disks are alike
Based on Spitzer data, we already knew that the disks around VLMS differ from those around T Tauri stars, having a higher C2H2/HCN ratio (Pascucci et al. 2009, 2013) and lacking clear signs of water emission. With MIRI, we achieve typical S/N ratios of 200-500 within 30-minute exposures for T Tauri stars (e.g. Temmink et al. 2024), leading to mJy noise levels.
Besides confirming this strong dichotomy in spectral appearance (see Fig. 2), the very high S/N and the sensitivity achieved with JWST/MIRI led us to the conclusion that no two disks are the same, even not within the group of T Tauri stars (Fig. 2, van Dishoeck et al. 2023, Kamp et al. 2023).
The analysis of the spectra has so far been done doing a (semi-)manual continuum subtraction and using 0D slab models for the gas emission. An overview of the approach can be found in Kamp et al. (2023). Most notably, for molecules with high column densities, line overlap can lead to a quasi-continuum (Tabone et al. 2023), thus making the placement of the continuum and the analysis of the line emission a coupled problem. On the other hand, 2-layer dust models such as Juhasz et al. (2009) have difficulty retrieving the dust continuum and features in the presence of now visible broad molecular emission bands (Jang et al. in prep.). Solutions have to be found in the simultaneous Bayesian fitting of dust and gas such as proposed by Liu et al. (2019) and Käufer et al. (2024).
The results of the 0D slab model analysis are the column densities, excitation temperatures and equivalent emitting areas of each molecule. In general, degeneracies are high, but temperatures (shapes of Q-branches) are often more robust than column densities. Thermo-chemical disk models clearly show that many of these molecules are not expected to be co-spatial. However, we can look for first trends in the retrieved data of T Tauri disks, i.e. which molecules share the same temperature regime and which ones are different. Even though it is still small number statistics, there could be a trend that CO2 tends to be somewhat cooler (100-400 K) than the other molecules and that HCN is often very warm (>500 K). Thermo-chemical models without inner disk substructure often show that CO2 emission extends to larger radii (up to ~10 au) compared to e.g. mid-IR water emission at the same wavelengths, and that HCN tends to be more compact (e.g. Bosman et al. 2017, Woitke et al. 2018, Anderson et al. 2021, Kamp et al. 2023). This very preliminary comparison paves the way to the use of more complex thermo-chemical disk models to be applied to JWST/MIRI data.
A first study was recently done by Woitke et al. (2024) on the disk around EX Lup. Two-dimensional radiation thermo-chemical disk models assume a radial gas and dust structure, allow the dust grains to settle depending on grain size, solve the continuum radiative transfer based on the dust opacities and then solve the gas chemistry (in this case using a network of 235 species connected by ~3000 reactions) and energy balance (heating/cooling) iteratively. These models take hours instead of milliseconds (like 0D slabs, Bayesian retrievals); hence, we have to use alternative ways to match observed mid-IR spectra. Based on the DIANA project (Woitke et al. 2019), we use an evolutionary algorithm, define the free parameters of the 2D disk model, and define the c2 to be minimalized. This approach can provide more context compared to the 0D slab modelling as we can study whether or not the physical conditions retrieved by slab models can exist based on our current understanding of the physical and chemical structure of such disks. Figure 3 shows the resulting fit after ~200 generations (~60000 CPU hours) for a model that uses the solar C/O ratio (0.46). Compared to the results achieved with slab model fitting, where each molecule is fitted separately, the emission shown here is from a ray tracing of the final thermo-chemical disk model using the Fast Line Tracer FLiTs (developed by M. Min, Woitke et al. 2018). In such models, the molecular emission of CO, H2O, CO2, C2H2 originates spatially in very different regions of the disk dictated by a combination of density, UV field (e.g. scattering by dust, dust+gas shielding) and temperature (chemistry and heating/cooling are intertwined). Based on such an approach, one cannot rule out that there are not equally well-fitting alternative disk structures, but it is an amazing achievement to have 2D disk models that do reproduce the observations at this level. This would not have been possible without many of the insights of past years highlighting the importance of dust processing, settling, rounded inner rims, molecular self-shielding and mutual shielding and the wealth of molecular data now available through databases such as HITRAN2020 (Gordon et al. 2022) and Geisa (Delahaye et al. 2021).
The dichotomy found earlier with Spitzer between the T Tauri and VLMS disks is even more clear now with JWST/MIRI spectra. The four VLMS objects analysed so far by MINDS (Tabone et al. 2023, Arabhavi et al. 2024, Kanwar et al. submitted, Morales-Calderon et al. in prep.) span spectral types M5-M7.5 and show an incredibly rich spectrum of small hydrocarbon molecules such as CH4, C2H2, 13CCH2, C4H2, HCN and in some cases also C2H4, C2H6, C3H4, C6H6, CH3, as well as HC3N, while water and OH are likely absent (or very weak); the only oxygen bearing molecule routinely found is CO2 and its isotopologue 13CO2. In one case (Tabone et al. 2023), we detected CO, but often the short wavelength regions (5-7 mm) are dominated by the molecular absorption from the VLMS, making the search for CO disk emission very difficult. Such a rich hydrocarbon chemistry clearly points to an elemental abundance ratio C/O > 1 in the gas inside ~1 au of these VLMS. Interestingly, Xie et al. (2023) analysed another M5 VLMS disk and detected clear H2O emission next to C2H2, CO, CO2 and HCN. So, it could be that we still have a strong selection bias in our samples and that eventually we will find a gradual change in the inner disk composition from oxygen dominated to carbon dominated. This has recently been proposed based on models of ice transport in viscously evolving disks by Mah et al. (2023).
Warm inner disk chemistry
In the inner disks (~1 au) around young stars, we find densities up to 1015 cm-3 and temperatures between a few 100 and a few 1000 K. The conditions do overlap thus with warm planetary atmospheres. Woods & Willacy (2007) already demonstrated that the simplest cyclic hydrocarbon, benzene (C6H6) can form under these conditions through ion-molecule gas phase chemistry. Recently, Kanwar et al. (2024a) revisited the formation of small hydrocarbons expanding the large DIANA chemical network (Kamp et al. 2017) with 92 hydrocarbons with up to eight carbon atoms. We show that there are two main pathways to form the simple hydrocarbon C2H2 in the disk surface layers of T Tauri stars, the ion-molecule pathway starting with C+ (unlocked from CO via He+, cosmic rays/X-rays) and CH+, and the neutral-neutral pathway unlocking C from CO via UV photons. This surface layer drives an efficient hydrocarbon chemistry, because it coincides with the H/H2 transition and contains non-negligible abundances of atomic/ionized C. Its location is hence tied to the details of the warm H2 formation on dust grains (Cazaux & Tielens 2010). Benzene forms inside 0.2 au, in a reservoir that is well below the AV~1 mag line and extends down to the midplane. This means that under normal C/O circumstances, it can only be exposed if the dust opacities differ from the canonical ones.
Inspired by the richness of the disks around VLMS, we investigated the impact of a C/O ratio larger than 1 in such disks and found that many hydrocarbons will be enhanced in the optically thin surface layers. This happens because for a C/O larger than 1, all oxygen is locked in CO and the remaining C can drive additional carbon chemistry. In fact, we find all hydrocarbons detected in JWST/MIRI spectra to reside in our model inside a few au above the AV~1 mag layer (Kanwar et al. 2024a, Fig. 4). The temperatures fall into the same range as the observations (~150-500 K). This demonstrates that if there is a mechanism enhancing the elemental abundance ratio inside 1 au, gas-phase chemistry will proceed on short timescales and form the entire observed range of small hydrocarbons, including benzene. The models also show the presence of a deeper hydrocarbon reservoir which exists independent on the C/O ratio. In case dust opacities are very low due to e.g. efficient dust growth in the inner disk, that warm reservoir inside 0.05 au could become observable.
The JWST/MIRI spectra still show a few unidentified features and our thermo-chemical disk models would predict species such as C2, C2H, C3, CH2CCH, both cyclic and linear isotopomers of C3H2, C5H2, CH3C4H, C6, C6H2. None of them have spectral data in HITRAN2020 or Geisa. More work and collaboration with our colleagues from physical chemistry and/or laboratory spectroscopy is required to expand the existing molecular databases.
Unseen substructure, radiative transfer, and/or transport processes
We are just at the beginning of the interpretation of JWST/MIRI spectra of disks around young stars. Other contributions at this conference put forward detailed evolutionary scenarios, which will be further tested once larger unbiased samples of disk spectra have been analysed. Many theoretical works have already investigated how specific ‘local’ effects change the mid-IR appearance of disks, e.g. the C/O ratio (Najita et al. 2011, Woitke et al. 2018, Anderson et al. 2021), disk (sub-)structure and dust transport (Antonellini et al. 2015, Greenwood et al. 2019, Antonellini et al. 2023, Kamp et al. 2023, Vlasblom et al. 2024), photochemistry and shielding (Bethell & Bergin 2009, Bosman et al. 2022 a, b). This has increased the complexity and number of parameters in thermo-chemical disk models. The natural next step to couple these to transport models is being explored right now (e.g. Krijt et al. 2020), but more work is required to capture the aforementioned complexity also in these transport models. Figure 5 shows a sketch of the complexity that arises when trying to build a coherent modeling framework for the interpretation of JWST data. In fact, the outer disk, which is in many cases well characterised by ALMA data, provides a complementary view of the full story. Hence, in the future it will be crucial to interpret these two datasets within a single model framework to progress in our understanding of how planets are growing in these disks and how they acquire their final elemental compositions.
Acknowledgements
This work is based on observations made with the NASA/ESA/CSA James Webb Space Telescope. The data were obtained from the Mikulski Archive for Space Telescopes at the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS 5-03127 for JWST. These observations are associated with the European MIRI GTO program MINDS, program #1282 (PI: Th. Henning). The author would like to thank the entire MINDS team for their efforts in data reduction and analysis that made it possible to provide this first overview.
References
Anderson, D.E., Blake, G.A., Cleeves, L.I. et al. 2021, ApJ 909, 55.
Antonellini, S., Kamp, I., Riviere-Marichalar, P. et al. 2015, A&A 582, A105.
Antonellini, A., Kamp, I., Waters, L.B.F.M. 2023, A&A 672, A92.
Arabhavi, A.M., Kamp, I., Henning, Th. & MINDS team 2024, Science 384, 1086.
Bae, J. Isella, A., Zhu, Z., Martin, R., Okuzumi, S., Suriano, S. 2022, PPVII, Eds. Shu-ichiro Inutsuka, Yuri Aikawa, Takayuki Muto, Kengo Tomida, and Motohide Tamura.
Bethell, T., Bergin, E.A. 2009, Science 326, 1675B.
Bosman, A.D., Bruderer, S., van Dishoeck, E.F. 2017, A&A 601, A36.
Bosman, A., Bergin, E.A., Calahan, J.K., Duval, S.E. 2022a, ApJL 933, L26.
Bosman, A., Bergin, E.A., Calahan, J.K., Duval, S.E. 2022b, ApJL 933, L40.
Carr, J.S., Najita, J. 2008, Science 319, 1504.
Cazaux, S., Tielens, A.G.G.M. 2010, ApJ 715, 698.
Delahaye, T., Armante, R., Scott, N.A. et al. 2021, Journal of Molecular Spectroscopy 380, 111510.
Drazkowska, J, Bitsch, B., Lambrechts, M. et al. 2022, PPVII, Eds. Shu-ichiro Inutsuka, Yuri Aikawa, Takayuki Muto, Kengo Tomida, and Motohide Tamura.
Gordon, I.E., Rothman, L.S., Hargreaves, R.J. et al. 2022, J. Quant. Spectr. Rad. Transf. 277, 107949.
Greenwood, A., Kamp, I., Waters, L.B.F.M., Woitke, P., Thi, W.-F. 2019, A&A 626, A6.
Henning, Th., Kamp, I., Samland, M. & MINDS team 2024, PASP 136, 054302.
Juhasz, A., Henning, Th., Bouwman, J., Dullemond, C.P., Pascucci, I., Apai, D. 2009, ApJ 695, 1024.
Kamp, I., Henning, Th., Arabhavi, A.M. & MINDS team 2023, Faraday Discussion 245, 112.
Kanwar, J., Kamp, I., Woitke, P., Rab, Ch., Thi, W.-F., Min, M. 2024a, A&A 681, A22.
Kanwar, J., Kamp, I., Jang, H., et al. 2024b, A&A submitted.
Kauefer, T., Min, M.,Woitke, P., Kamp, I. Arabhavi, A. 2024, A&A 687, A209
Krijt, S., Bosman, A., Zhang, K, Schwarz, K.R., Ciesla, F.J., Bergin, E.A. 2020, ApJ 899, 134.
Kruijer, T.S., Touboul, M., Fischer-Gödde, M, Bermingham, K.R., Walker, R.J., Kleine, T. 2014, Science 344, 6188.
Lammer, H., Brasser, R., Johansen, A., Scherf, M., Leitzinger, M. 2021, Space Sci Rev 217, 7.
Mah, J., Bitsch, B., Pascucci, I., Henning, Th. 2023, A&A 677, L7.
Mezger, K., Schönbächler, M., Bouvier, A. 2020, Space Sci Rev 216, 27.
Miotello, A., Kamp, I., Birnstiel, T., Cleeves, L.I., Kataoka, A. 2022, PPVII, Eds. Shu-ichiro Inutsuka, Yuri Aikawa, Takayuki Muto, Kengo Tomida, and Motohide Tamura.
Najita, J.R., Adamkovics, M., Glassgold, A.E. 2011, ApJ 743, 147.
Pascucci, I., Apai, D., Luhman, K, et al. 2009, ApJ 696, 143.
Pascucci, I., Herczeg, G., Carr, J.S., Bruderer, S. 2013, ApJ 779, 178.
Pontoppidan, K.M., Salyk, C., Blake, G.A. et al. 2010, ApJ 720, 887.
Salyk, C., Pontoppidan, K.M., Blake, G.A. et al. 2008, ApJ 676, L49.
Tabone, B., Bettoni, G., van Dishoeck, E.F. & MINDS team 2023, Nature Astronomy 7, 805.
Temmink, M., van Dishoeck, E.F., Grant, S. et al. 2024, A&A in press.
van Dishoeck, E.F., Grant, S., Tabone, B. et al. 2023, Faraday Discussion 245, 52.
Vlasblom, M., van Dishoeck, E.F., Tabone, B., Bruderer, S. 2024, A&A 682, A91.
Woitke, P., Min, M., Thi, W.-F. et al. 2018, A&A 618, A57.
Woods, P., Willacy, K. 2007, ApJ 655, L49.
Xie, C., Pascucci, I., Long, F. et al. 2023, ApJ 959, L25.