Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userMachine Learning of Interstellar Chemical Inventories
140
0
0.0
(
0
)
Added by
Kin Long Kelvin Lee
Publication date
2021
fields
Physics
and research's language is
English
Ask ChatGPT about the research
No Arabic abstract
The characterization of interstellar chemical inventories provides valuable insight into the chemical and physical processes in astrophysical sources. The discovery of new interstellar molecules becomes increasingly difficult as the number of viable species grows combinatorially, even when considering only the most thermodynamically stable. In this work, we present a novel approach for understanding and modeling interstellar chemical inventories by combining methodologies from cheminformatics and machine learning. Using multidimensional vector representations of molecules obtained through unsupervised machine learning, we show that identification of candidates for astrochemical study can be achieved through quantitative measures of chemical similarity in this vector space, highlighting molecules that are most similar to those already known in the interstellar medium. Furthermore, we show that simple, supervised learning regressors are capable of reproducing the abundances of entire chemical inventories, and predict the abundance of not yet seen molecules. As a proof-of-concept, we have developed and applied this discovery pipeline to the chemical inventory of a well-known dark molecular cloud, the Taurus Molecular Cloud 1 (TMC-1); one of the most chemically rich regions of space known to date. In this paper, we discuss the implications and new insights machine learning explorations of chemical space can provide in astrochemistry.
rate research
Read More
In recent years, phosphorus monoxide (PO) -- an important molecule for prebiotic chemistry -- has been detected in star-forming regions and in the comet 67P/Churyumov-Gerasimenko. These studies have revealed that, in the interstellar medium, PO is systematically the most abundant P-bearing species, with abundances that are $sim$1-3 times greater than those derived for phosphorus nitride (PN), the second most abundant P-containing molecule. The reason why PO is more abundant than PN remains still unclear. Experimental studies with phosphorus in the gas phase are not available, probably because of the difficulties in dealing with its compounds. Therefore, the reactivity of atomic phosphorus needs to be investigated using reliable computational tools. To this end, state-of-the-art quantum-chemical computations have been employed to evaluate accurate reaction rates and branching ratios for the P + OH $rightarrow$ PO + H and P + H$_2$O $rightarrow$ PO + H$_2$ reactions in the framework of a master equation approach based on ab-initio transition state theory. The hypothesis that OH and H${_2}$O can be potential oxidizing agents of atomic phosphorus is based on the ubiquitous presence of H${_2}$O in the ISM. Its destruction then produces OH, which is another very abundant species. While the reaction of atomic phosphorus in its gound state with water is not a relevant source of PO because of emerged energy barriers, the P + OH reaction represents an important formation route of PO in the interstellar medium. Our kinetic results show that this reaction follow an Arrhenius behavior, and thus its rate coefficients alpha=2.28$times$10$^{-10}$ cm${^3}$ molecule$^{-1}$ s$^{-1}$, beta=0.16 and gamma=0.37 K increase by increasing the temperature.
Interstellar complex organic molecules (iCOMs) are assumed to be mainly formed on dust-grain surfaces. However, neutral gas-phase reactions in the interstellar medium (ISM) can play an important role. In this paper, by investigating the reaction between aldehydes and the cyano radical, we show that both formaldehyde (CH$_2$O) and acetaldehyde (CH$_3$CHO) can lead to the formation of formyl cyanide (HCOCN). Owing to accurate quantum-chemical computations followed by rate constant evaluations, we have been able to suggest and validate an effective mechanism for the formation of HCOCN, one of the molecules observed in the ISM. Quite interestingly, the mechanism starting from CH$_2$O is very effective at low temperature, while that involving CH$_3$CHO becomes more efficient at temperatures above 200 K.
The chemical pathways linking the small organic molecules commonly observed in molecular clouds to the large, complex, polycyclic species long-suspected to be carriers of the ubiquitous unidentified infrared emission bands remain unclear. To investigate whether the formation of mono- and poly-cyclic molecules observed in cold cores could form via the bottom-up reaction of ubiquitous carbon-chain species with, e.g. atomic hydrogen, a search is made for possible intermediates in data taken as part of the GOTHAM (GBT Observations of TMC-1 Hunting for Aromatic Molecules) project. Markov-Chain Monte Carlo (MCMC) Source Models were run to obtain column densities and excitation temperatures. Astrochemical models were run to examine possible formation routes, including a novel grain-surface pathway involving the hydrogenation of C$_6$N and HC$_6$N, as well as purely gas-phase reactions between C$_3$N and both propyne (CH$_3$CCH) and allene (CH$_2$CCH$_2$), as well as via the reaction CN + H$_2$CCCHCCH. We report the first detection of cyanoacetyleneallene (H$_2$CCCHC$_3$N) in space toward the TMC-1 cold cloud using the Robert C. Byrd 100 m Green Bank Telescope (GBT). Cyanoacetyleneallene may represent an intermediate between less-saturated carbon-chains, such as the cyanopolyynes, that are characteristic of cold cores and the more recently-discovered cyclic species like cyanocyclopentadiene. Results from our models show that the gas-phase allene-based formation route in particular produces abundances of H$_2$CCCHC$_3$N that match the column density of $2times10^{11}$ cm$^{-2}$ obtained from the MCMC Source Model, and that the grain-surface route yields large abundances on ices that could potentially be important as precursors for cyclic molecules.
In this work we investigate the effects of ion accretion and size-dependent dust temperatures on the abundances of both gas-phase and grain-surface species. While past work has assumed a constant areal density for icy species, we show that this assumption is invalid and the chemical differentiation over grain sizes are significant. We use a gas-grain chemical code to numerically demonstrate this in two typical interstellar conditions: dark cloud (DC) and cold neutral medium (CNM). It is shown that, although the grain size distribution variation (but with the total grain surface area unchanged) has little effect on the gas-phase abundances, it can alter the abundances of some surface species by factors up to $sim2-4$ orders of magnitude. The areal densities of ice species are larger on smaller grains in the DC model as the consequence of ion accretion. However, the surface areal density evolution tracks are more complex in the CNM model due to the combined effects of ion accretion and dust temperature variation. The surface areal density differences between the smallest ($sim 0.01mu$m) and the biggest ($sim 0.2mu$m) grains can reach $sim$1 and $sim$5 orders of magnitude in the DC and CNM models, respectively.
In this work, we discuss use of machine learning techniques for rapid prediction of detonation properties including explosive energy, detonation velocity, and detonation pressure. Further, analysis is applied to individual molecules in order to explore the contribution of bonding motifs to these properties. Feature descriptors evaluated include Morgan fingerprints, E-state vectors, a custom sum over bonds descriptor, and coulomb matrices. Algorithms discussed include kernel ridge regression, least absolute shrinkage and selection operator (LASSO) regression, Gaussian process regression, and the multi-layer perceptron (a neural network). Effects of regularization, kernel selection, network parameters, and dimensionality reduction are discussed. We determine that even when using a small training set, non-linear regression methods may create models within a useful error tolerance for screening of materials.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
