We present the graph-based molecule software MOLASSEMBLER for building organic and inorganic molecules. MOLASSEMBLER provides algorithms for the construction of molecules built from any set of elements from the periodic table. In particular, polynuclear transition-metal complexes and clusters can be considered. Structural information is encoded as a graph. Stereocenter configurations are interpretable from Cartesian coordinates into an abstract index of permutation for an extensible set of polyhedral shapes. Substituents are distinguished through a ranking algorithm. Graph and stereocenter representations are freely modifiable, and the chiral state is propagated where possible through incurred ranking changes. Conformers are generated with full stereoisomer control by four spatial dimension Distance Geometry with a refinement error function including dihedral terms. Molecules are comparable by an extended graph isomorphism, and their representation is canonicalizeable. MOLASSEMBLER is written in C++ and provides Python bindings.

An activated fragment which is structurally unstable when considered isolated can be stabilized through binding to a suitable molecular environment; for instance, to a transition-metal fragment. The metal fragment may be designed in a shell-wise build-up of a surrounding molecular environment. However, adding more and more atoms in a consecutive fashion soon leads to a combinatorial explosion of structures, which is unfeasible to handle without automation. Here, we present a fully automated and parallelized framework that constructs such an embedding environment atom-wise. Molecular realizations of such an environment are constructed based on simple heuristic rules intended to screen a sufficiently large portion of the possible compound space and are then subsequently optimized by electronic structure methods. (Constrained-optimized) structures are then evaluated with respect to a scoring function, for which we choose here the concept of gradient-driven molecule construction. This concept searches for structure modifications that reduce the forces on all atoms. We develop and analyze our approach at the example of CO activation by reproducing a known compound and mapping out possible alternative structures and their effect on the stabilization of a (bent) CO ligand. For all generated structures, the nuclear gradient on the activated fragment and its coordination energy are evaluated to steer the design process. © 2017 Wiley Periodicals, Inc.

The computation of reaction selectivity represents an appealing complementary route to experi- mental studies and a powerful mean to refine catalyst design strategies. Accurately establishing the selectivity of reactions facilitated...

The computation of reaction selectivity represents an appealing complementary route to experimental studies and a powerful mean to refine catalyst design strategies. Accurately establishing the selectivity of reactions facilitated by molecular catalysts, however, remains a challenging task for computational chemistry. The small free energy differences that lead to large variations in the enantiomeric ratio represent particularly tricky quantities to predict with sufficient accuracy to be helpful for prioritizing experi- ments. Further complicating this problem is the fact that standard approaches typically consider only one or a handful of conformers identified through human intuition as pars pro toto of the conformational space. Obviously, this assumption can potentially lead to dramatic failures should key energetic low-lying structures be missed. Here, we in- troduce a multi-level computational pipeline built upon the graph-based Molassembler library that combines conformer generation and tailored functionalization to facilitate high-throughput mechanistic investigations of chemical reactions. The capabilities of this approach are validated by examining a Rh(III) catalyzed asymmetric C-H activa- tion reaction and assessing the limitations associated with the underlying ligand design model. Specifically, the presence of remarkably flexible chiral Cp ligands, which induce the experimentally observed high level of selectivity, present a rich configurational landscape where multiple unexpected conformations contribute to the reported enan- tiomeric ratios (er). Using Molassembler, we show that considering about 20 transition state conformations per catalysts, which are generated with little human intervention and are not tied to “back of the envelope” models, accurately reproduces experimental er values with limited computational expense.

No abstract