This thesis has investigated the chemical evolution of the inner regions (r ≤ 10 AU) of a modelled protoplanetary disk surrounding a low-mass T Tauri star; a phase our own solar system underwent some 4.5–4.6 billion years ago. A 1+1- dimensional physical model of a radially-accreting protoplanetary disk was combined with a chemical model consisting of gas-phase reactions extracted from the RATE95 UMIST database for Astrochemistry, and to this, a network of gas-grain interaction and deuterated reactions were added. Influenced by the knowledge that radionuclides may have been abnormally abundant in our early solar system compared with the interstellar medium, and that the energy expelled from their decay is sufficient to ionize molecules, a 1-dimensional simulation along the disk midplane was performed, comparing the chemistry with and without radionuclides, as a function of the radionuclide ionization rate. The molecules C4H2, HC3N, C3H, HCN, CH4, C2H2 and N2 were found to be particularly sensitive. Of these, HCN and C2H2 have already been detected in protoplanetary disks. Motivated by observations which suggest that T Tauri systems vary from faster to slower accretion rates, the chemical distributions of two disks with stellar accretion rates of 10−7 M⊙ yr−1 and 10−8 M⊙ yr−1 were compared. Allowing the mass accretion rate (and thus physical conditions) to vary in time, starting from 10−7 M⊙ yr−1, and evolving to a mass accretion rate of 10−8 M⊙ yr−1, the molecules CN, HCN, H2CO and NH3 were found to be particularly sensitive when compared to a standalone 10−8 M⊙ yr−1 simulation. With the use of a 1-D CASA LTE algorithm for ALMA, the sensitivities of HCN and H2CO were transcribed into integrated intensity differences, as a function of emissivity and optical depth. The largest differences were associated with the largest feasible transitions of HCN and H2CO; J = 8-7 and JK = 10010-909 respectively, but could not be converted into potentially observable integrated fluxes due to the restrictions of this model. Both molecules were found to trace different regions of the disk. Using the stellar accretion rate-to-age linear relation evaluated from observations, the calculation times for the 10−7 M⊙ yr−1 and 10−8 M⊙ yr−1 were re-evaluated, of which CN and NH3 emerged as the most sensitive molecules. Thus, CN and NH3 may be a possible tracer of calculation time, and disk age.

In recent years, thousands of exoplanets have been discovered around a variety of stellar hosts. The disks of gas and dust surrounding young stars are the location and source of material for planet formation. The properties of these protoplanetary disks therefore directly affect the planetary systems that may form. However, the details of the planet formation process are still unclear. In this dissertation, I constrain planet formation mechanisms by measuring the properties of protoplanetary disks, focusing on mass, dust grain growth, and dust settling. I use physically-motivated models and an Artificial Neural Network along with a Markov Chain Monte Carlo (MCMC) fitting procedure to obtain these and other disk properties. This dissertation compiles the largest sample to date of consistently-modeled protoplanetary disks, probing how disk properties vary with host mass and age. The occurrence of planetary companions increases as stellar mass decreases. Thus, brown dwarfs (BDs), with smaller masses than pre-main-sequence stars, may commonly host planets. Studying properties of BD disks and comparing them to pre- main-sequence star disks is therefore important for constraining their planet-forming potential. I present spectral energy distribution (SED) models for BD and pre-main- sequence star disks in four star-forming regions. The SEDs consist of archival photometry data spanning optical to millimeter wavelengths. I model the BD disk SEDs using physically-motivated radiative transfer code; pre-main-sequence star SEDs are modeled using a newly-developed MCMC fitting procedure that allows for a more complete analysis of the disk properties. I compare disk masses and dust settling in these two disk categories to gauge how host mass affects these properties. Typical disk lifetimes are a few tens of millions of years; planet formation likely occurs within the first few million years or less. Comparing how disk properties vary between star-forming regions of different ages can help pinpoint the timeline for planet formation. I present SED models for BDs in four star-forming regions and pre-main-sequence stars in eleven star-forming regions. I obtain the disk masses, dust grain sizes, and amount of dust settling in the disks and discuss the differences and similarities of these properties across regions of varying age.

The calculation of the physical conditions near the inner rim of a protoplanetary disk using the new computational model Pisco is described. Diagnostic plots illustrate solutions for disk structure, radiation field, chemical composition, and heating and cooling of the disk in a steady-state approximation for both disks with unsettled dust and with settled dust. Disks with unsettled dust are found to have hotter gas temperatures above the disk photosphere and a more pronounced temperature inversion at the disk photosphere. Recommendations are made for the development of Pisco. Pisco has the potential to explore what observed molecular emission can imply about disk structure. / text

京都大学 / 0048 / 新制・課程博士 / 博士(理学) / 甲第18790号 / 理博第4048号 / 新制||理||1582(附属図書館) / 31741 / 京都大学大学院理学研究科物理学・宇宙物理学専攻 / (主査)教授 中村 卓史, 教授 鶴 剛, 教授 田中 貴浩 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DGAM

We present optical and near-infrared high-contrast images of the transitional disk HD 100546 taken with the Magellan Adaptive Optics system (MagAO) and the Gemini Planet Imager (GPI). GPI data include both polarized intensity and total intensity imagery, and MagAO data are taken in Simultaneous Differential Imaging mode at Ha. The new GPI H-band total intensity data represent a significant enhancement in sensitivity and field rotation compared to previous data sets and enable a detailed exploration of substructure in the disk. The data are processed with a variety of differential imaging techniques (polarized, angular, reference, and simultaneous differential imaging) in an attempt to identify the disk structures that are most consistent across wavelengths, processing techniques, and algorithmic parameters. The inner disk cavity at 15 au is clearly resolved in multiple data sets, as are a variety of spiral features. While the cavity and spiral structures are identified at levels significantly distinct from the neighboring regions of the disk under several algorithms and with a range of algorithmic parameters, emission at the location of HD 100546 "c" varies from point-like under aggressive algorithmic parameters to a smooth continuous structure with conservative parameters, and is consistent with disk emission. Features identified in the HD 100546 disk bear qualitative similarity to computational models of a moderately inclined two-armed spiral disk, where projection effects and wrapping of the spiral arms around the star result in a number of truncated spiral features in forward-modeled images.

The number of extrasolar planet discoveries has increased dramatically over the last 15 years. Nearly 700 exoplanets have currently been observed through a variety of observation techniques. Most of the currently documented exoplanets differ greatly from the planets in our own Solar System, with various combinations of eccentric orbits, short orbital periods, and masses many times that of Jupiter. More recently, planets belonging to a new class of `distant gas giants' have also been discovered with orbits of 30 to 100 times that of Jupiter. The wide variety of different planet formation outcomes stem from a complex interplay between gravitational interactions, hydrodynamic interactions and competitive accretion among the planets that is not yet fully understood. Simulations performed using a series of modifications to an existing, widely used hydrodynamic code (FARGO) are presented. The main goal is to develop a more rigorous and robust gas accretion scheme that is valid and consistent for the ranges of exolanetary gas giant masses, eccentricities and semimajor axes that have been observed to better understand the mechanisms involved in their formation. The resulting scheme is a more robust and accurate prescription for gas accretion onto planetary cores in a manner that is mostly resolution independent and valid over a large range of masses (less than an Earth mass to multiple Jupiter masses). The modified scheme accounts for multiple, competing, dynamic accretion mechanisms (including atmospheric effects) and their associated time scales between an arbitrary number of protoplanets. This updated accretion scheme provides a means for exploring the entire formation process of gas giants out of a variety of initial conditions in a self-consistent manner. The modifications made to the code as well as simulation results will be discussed and explored.

La zone morte, région laminaire confinée au coeur des disques protoplanétaires dont la turbulence de l'écoulement à petite échelle explique l'accrétion de matière sur l'étoile en formation, semble être un lieu propice à la formation planétaire. En effet, au bord interne de la zone morte la différence d'accrétion entraîne le développement d'une sur-densité capable de piéger les grains de poussière qui dérivent vers l'étoile. L'écoulement à cet endroit est de plus potentiellement instable. Le cas échéant, il s'organise en structures tourbillonnaires appelées ''vortex'' qui collectent efficacement la poussière. La position du bord interne est toutefois très incertaine et dépend en particulier de la thermodynamique du modèle de disque considéré. Récemment, le déplacement du bord interne a été envisagé pour expliquer la variabilité de l'accrétion des étoiles jeunes. Cette thèse aborde le problème posé par l'influence de la thermodynamique sur la dynamique du bord interne de la zone morte. Des simulations MHD qui incluent le couplage entre les processus thermodynamiques avec la dynamique de l'écoulement ont tout d'abord permis de confirmer le comportement dynamique du bord interne ainsi que de réaliser la mesure inédite de sa vitesse typique de déplacement. La comparaison de ces résultats avec les prédictions données par un modèle de champ moyen a révélé le rôle du transport d'énergie par des ondes excitées au bord interne de la zone morte. Ces simulations présentent de plus un phénomène nouveau: les vortex formés à l'interface suivent un cycle de formation-migration-destruction. Cette découverte est susceptible de modifier notre vision du scénario de formation planétaire. En résumé, cette thèse met en évidence le fait que les processus thermodynamiques sont au coeur du fonctionnement de la région du bord interne de la zone morte dans les disques protoplanétaires. / The dead zone, a quiescent region enclosed in the turbulent flow of a protoplanetary disk, seems to be a promising site for planet formation. Indeed, the development of a density maximum at the dead zone inner edge, that has the property to trap the infalling dust, is a natural outcome of the accretion mismatch at this interface. Moreover, the flow here may be unstable and organize itself into vortical structures that efficiently collect dust grains. The inner edge location is however loosely constrained. In particular, it depends on the thermodynamical prescriptions of the disk model that is considered. It has been recently proposed that the inner edge is not static and that the variations of young stars accretion luminosity are the signature of this interface displacements. This thesis address the question of the impact of the gas thermodynamics onto its dynamics around the dead zone inner edge. MHD simulations including the complex interplay between thermodynamical processes and the dynamics confirmed the dynamical behaviour of the inner edge. A first measure of the interface velocity has been realised. This result has been compared to the predictions of a mean field model. It revealed the crucial role of the energy transport by density waves excited at the interface. These simulations also exhibit a new intriguing phenomenon: vortices forming at the interface follow a cycle of formation-migration-destruction. This vortex cycle may compromise the formation of planetesimals at the inner edge. This thesis claims that thermodynamical processes are at the heart of how the region around the dead zone inner edge in protoplanetary disks works.

We show how physics-informed neural networks can be used to solve compressible \NS equations in protoplanetary disks. While young planets form in protoplanetary disks, because of the limitation of current techniques, direct observations of them are challenging. So instead, existing methods infer the presence and properties of planets from the disk structures created by disk-planet interactions. Hydrodynamic and radiative transfer simulations play essential roles in this process. Currently, the lack of computer resources for these expensive simulations has become one of the field's main bottlenecks. To solve this problem, we explore the possibility of using physics-informed neural networks, a machine learning method that trains neural networks using physical laws, to substitute the simulations. We identify three main bottlenecks that prevent the physics-informed neural networks from achieving this goal, which we overcome by hard-constraining initial conditions, scaling outputs and balancing gradients. With these improvements, we reduce the relative L2 errors of predicted solutions by 97% ~ 99\% compared to the vanilla PINNs on solving compressible NS equations in protoplanetary disks. / Graduate / 2022-12-10

京都大学 / 0048 / 新制・課程博士 / 博士(理学) / 甲第20917号 / 理博第4369号 / 新制||理||1627(附属図書館) / 京都大学大学院理学研究科物理学・宇宙物理学専攻 / (主査)教授 嶺重 慎, 准教授 前田 啓一, 教授 太田 耕司 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DFAM

Les étoiles de type solaire se forment par l'effondrement d'un nuage moléculaire, durant lequel la matière s'organise autour de l'étoile en formation sous la forme d'un disque, appelé disque protoplanétaire. Dans ce disque se forment les planètes, comètes et autres objets du système stellaire. La nature de ces objets peut donc avoir un lien avec l'histoire de la matière du disque.J'ai étudié l'évolution chimique et physique de cette matière, du nuage au disque, à l'aide du code de chimie gaz-grain Nautilus.Une étude de sensibilité à divers paramètres du modèle (comme les abondances élémentaires et les paramètres de chimie de surface) a été réalisée. Notamment, la mise à jour des constantes de vitesse et des rapports de branchement des réactions de notre réseau chimique s'est avérée influente sur de nombreux points, comme les abondances de certaines espèces chimiques, et la sensibilité du modèle à ses autres paramètres.Plusieurs modèles physiques d'effondrement ont également été considérés. L'approche la plus complexe et la plus consistante a été d'interfacer notre code de chimie avec le code radiatif magnétohydrodynamique de formation stellaire RAMSES, pour modéliser en trois dimensions l'évolution physique et chimique de la formation d'un jeune disque. Notre étude a démontré que le disque garde une trace de l'histoire passée de la matière, et sa composition chimique est donc sensible aux conditions initiales. / Low mass stars, like our Sun, are born from the collapse of a molecular cloud. The matter falls in the center of the cloud, creating a protoplanetary disk surrounding a protostar. Planets and other solar system bodies will be formed in the disk.The chemical composition of the interstellar matter and its evolution during the formation of the disk are important to better understand the formation process of these objects.I studied the chemical and physical evolution of this matter, from the cloud to the disk, using the chemical gas-grain code Nautilus.A sensitivity study to some parameters of the code (such as elemental abundances and parameters of grain surface chemistry) has been done. More particularly, the updates of rate coefficients and branching ratios of the reactions of our chemical network showed their importance, such as on the abundances of some chemical species, and on the code sensitivity to others parameters.Several physical models of collapsing dense core have also been considered. The more complex and solid approach has been to interface our chemical code with the radiation-magneto-hydrodynamic model of stellar formation RAMSES, in order to model in three dimensions the physical and chemical evolution of a young disk formation. Our study showed that the disk keeps imprints of the past history of the matter, and so its chemical composition is sensitive to the initial conditions.