Exploring Machine Learning Tools for the Prediction of the Stability of New Togni-type Reagents.

In the context of the prediction of the (in-)stability of chemical compounds using machine learning tools, we are often confronted with a basic issue: Whereas much information is available on stable (existing) compounds, little is known about compounds that might well exist, but that have not yet been successfully synthesized, or compounds that are inherently unstable (kinetically and thermodynamically). In the search for Togni-type reagents, many of them kinetically instable, the stability of the prospects can be assessed based on the transition state for the conversion to their non-hypervalent inactive isomer. In earlier work, we determined the barriers of conversion for over one-hundred reagents, still not enough information to train a tool such as a vector support machine. Here, instead, we focus on the early intermediate structures expressed along the isomerization pathway, i.e. transition state searches are replaced by finding (local) minima. Based on an array of 382 Togni-type reagents whose behaviour was known in advance, we show that it is possible to have the machine predict the intermediate form expressed. The approach introduced here can be used to make predictions on the stability and possibly also the reactivity of Togni-type reagents in general.

Why do the Togni reagent and some of its derivatives exist in the high-energy hypervalent iodine form? New insight into the origins of their kinetic stability.

In a recent study published in ChemComm, H. F. Schaefer and coworkers showed that the Togni trifluoromethyltation reagent and some of its derivatives appear in a high-energy hypervalent form. The (kinetic) stability of these reagents is granted by the five-membered ring of their benziodoxole-based scaffold, which prevents isomerization to the (inactive) acyclic ether form. Whereas the thermodynamic stability of these reagents can be predicted on the basis of the trans influence of the electrophilic substituent, no such descriptor was found for their kinetic stability. In this study, we explore an array of Togni-type reagents, and show that the barrier to isomerization can be predicted based on the bond length between the iodine atom and the electrophilic substituent. For compounds, where this correlation does not hold, we have a reliable indication that the structure of the transition state is at variance with those in the series.

Code interoperability and standard data formats in quantum chemistry and quantum dynamics: The Q5/D5Cost data model.

Code interoperability and the search for domain-specific standard data formats represent critical issues in many areas of computational science. The advent of novel computing infrastructures such as computational grids and clouds make these issues even more urgent. The design and implementation of a common data format for quantum chemistry (QC) and quantum dynamics (QD) computer programs is discussed with reference to the research performed in the course of two Collaboration in Science and Technology Actions. The specific data models adopted, Q5Cost and D5Cost, are shown to work for a number of interoperating codes, regardless of the type and amount of information (small or large datasets) to be exchanged. The codes are either interfaced directly, or transfer data by means of wrappers; both types of data exchange are supported by the Q5/D5Cost library. Further, the exchange of data between QC and QD codes is addressed. As a proof of concept, the H + H2 reaction is discussed. The proposed scheme is shown to provide an excellent basis for cooperative code development, even across domain boundaries. Moreover, the scheme presented is found to be useful also as a production tool in the grid distributed computing environment.

Impact of donor-acceptor functionalization on the properties of linearly π-conjugated oligomers: establishing quantitative relationships for the substituent and substituent cooperative effect based on quantum chemical calculations.

To understand better the impact of donor-acceptor substitution on the properties of linearly π-conjugated compounds, we performed a computational study on a series of variably substituted trans-polyacetylenes, polyynes, and polythiophenes. The focus of this work is on how rapidly the impact of a given substituent or a given combination of substituents vanishes along the π-conjugated chain. The response of the structural (bond-length alternation, rotational barrier) and molecular properties ((hyper)polarizability, chemical shift) to substitution is analyzed using different protocols, including a superposition model for the evaluation of the cooperative effect of substituents in homo- and heterosubstituted oligomers. With the exception of the (hyper)polarizability, the impact of donor-acceptor substitution is found to vanish following an exponential. The rate of decay of the substituent impact is found to be characteristic for each backbone, whereas the choice of substituent determines the absolute value of the respective property. The combination of substituents is shown to determine whether the substituent cooperative effect on a property is of an enhancing or damping nature. The rate of decay of the cooperative effect on most properties, including the (hyper)polarizability, is also found to follow an exponential law.

On the effect of electron correlation on the static second hyperpolarizability of π conjugated oligomer chains.

In this article, we report on the ab initio calculation of the static longitudinal second hyperpolarizability (γ) of π conjugated unsaturated oligomer chains using polyacetylene and polyyne as model compounds. The common observation is that the electron correlation enhances γ in these systems. The present study reveals that for extended chain lengths the opposite appears to be true: Electron correlation may have a damping effect on this property. For double-zeta basis sets, a negative contribution from electron correlation to γ is found within the range of chain lengths investigated. For triple-zeta basis sets, the same behavior must be anticipated at larger chain lengths based on extrapolation schemes. The analysis of the excitation energies and transition moments shows that transition moments between excited states as predicted by the Hartree-Fock and coupled cluster methods have a different response to chain length extension. There also are indications that higher order correlation effects will enhance γ.

Endocyclic cleavage in glycosides with 2,3-trans cyclic protecting groups.

An endocyclic pathway is proposed as a reaction mechanism for the anomerization from the ß (1,2-trans) to the α (1,2-cis) configuration observed in glycosides carrying 2,3-trans cyclic protecting groups. This reaction occurs in the presence of a weak Lewis or Brønsted acid, while endocyclic cleavage (endocleavage) in typical glycosides was observed only when mediated by protic media or strong Lewis acids. To rationalize the behavior of this class of compounds, the reaction mechanism and the promoting factors of the endocleavage are investigated using quantum-mechanical (QM) calculations and experimental studies. We examine anomerization reactions of thioglycosides carrying 2,3-trans cyclic protecting groups, employing boron trifluoride etherate (BF(3)·OEt(2)) as a Lewis acid. The estimated theoretical reactivity, based on a simple model to predict transition state (TS) energies from the strain caused by the fused rings, is very close to the TS energies calculated by the TS search along the C1-C2 bond rotation after the endo C-O bond breaking. Excellent agreement is found between the predicted TS energies and the experimental reactivity ranking. The series of calculations and experiments strongly supports the predominance of the endocyclic rather than the exocyclic mechanism. Furthermore, these investigations suggest that the inner strain is the primary factor enhancing the endocleavage reaction. The effect of the cyclic protecting group in restricting the pyranoside ring to a (4)C(1) conformation, extensively discussed in conjunction with the stereoelectronic effect theory, is shown to be a secondary factor.