The complex build algorithm to set up starting structures of lanthanoid complexes with stereochemical control for molecular modeling

When handling metallic centers of higher coordination numbers, one is commonly deluded with the presumption that any assembled metal complex geometry (including a crystallographic one) is good enough as a starting structure for computational chemistry calculations; all oblivious to the fact that such a structure is nothing short of just one out of several, sometimes dozens, or even thousands of other stereoisomers. Moreover, coordination chirality, so frequently present in complexes of higher coordination numbers, is another often overlooked property, rarely recognized as such. The Complex Build algorithm advanced in this article has been designed with the purpose of generating starting structures for molecular modeling calculations with full stereochemical control, including stereoisomer complete identification and coordination chirality recognition. Besides being in the chosen correct stereochemistry, the ligands are positioned by the Complex Build algorithm in a very unobstructed and unclogged manner, so that their degrees of freedom do not hinder or even choke one another, something that would otherwise tend to lead to negative force constants after further geometry optimizations by more advanced computational model chemistries. The Complex Build algorithm has been conceived for any metallic center, but at present is targeting primarily lanthanoids whose coordination numbers range mostly from 5 to 12 and often lead to a combinatorial explosion of stereoisomers.


Introduction
The figures in this Supplementary Information were made for 14 different coordination compounds, each one for a different lanthanoid, from lanthanum to lutetium, with the sole exception of the exceedingly rare and radioactive element promethium. The compounds were also chosen with the intention of spanning a variety of chemically different ligands, ranging from rigid to very flexible ones -types that constitute a true challenge to the Complex Build algorithm. For this study, we chose to use the RM1 Hamiltonian 1 , available in the software MOPAC 2016 2 as a feasible and accurate enough 3 quantum mechanical model to further optimize both the crystallographic and Complex Build structures, used as starting geometries.
For each of the compounds, we show, on the top left, the crystallographic structure from the Cambridge Crystallographic Data Center, and, on the bottom left, the structure obtained after its RM1 optimization.
On the top right, we show the structure of the same compound, using the Complex Build algorithm at the same stereoisomer configuration as that of its corresponding crystallographic one; and, on the bottom right, similarly, the structure obtained after its optimization using the RM1 Hamiltonian.
Each of the figures also displays its corresponding coordination polyhedron shape and point group symbols.
The figure captions also contain further comments on the specifics of each case, highlighting a comparison between the results after using, either the crystallographic, or the Complex Build assembled structures, as starting geometries for further model chemistries calculations.
More information on the Complex Build algorithm, including software download and video tutorials, can be found in https://complexbuild.sparkle.pro.br/.      The pyridine derivative is neutral and each of the three -diketonates has a charge of -1, that cancel the +3 charge in the samarium trivalent cation, all leading to a neutral complex. Coordination stereochemistry, and even the overall structural orientation of ligands is evidently preserved in all cases, meaning that the Complex Build algorithm was able to construct a structure very similar to the crystal structure, and that both of these structures were already very close to local minima.  Each of the three identical ligands has a charge of -1, leading to a neutral complex. Once again, it is evident that the Complex Build managed to construct the same stereoisomer as the crystal structure and that both of these starting structures led to very similar optimized structures.   Each of the three -diketonates has a charge of -1, that cancel the charge of the central holmium trivalent cation. Since the water, the last ligand, is neutral, so is the complex. Whereas most ligands have remained in their original relative positions after the RM1 optimization, the -diketonate shown in the lower part of the images has changed its coordination orientation in the optimized crystallographic structure. This change led to a different stereoisomer, something that did not happen for the structure obtained from the Complex Build algorithm, used as a starting geometry. These large groups generate images in which the Erbium ion is depicted slightly off the center, something that makes a comparison among the structures a less straightforward affair. We chose to depict this particular orientation, even though it is not immediately obvious that the larger ligands are monodentates. The little circle in front of the erbium atom is the oxygen atom of a coordinated water. Since the two larger ligands and the water are neutral, and each nitrate has a -1 charge, then the complex is neutral. Each of the three -diketonates has a charge of -1, cancelling the +3 charge in the Europium cation. The remaining ligands are neutral, leading to a neutral complex. It is interesting to note the left-right inversion of the tilt in the bidentate ligand in the back after the RM1 optimization in the ComplexBuild structure. This inversion changes the stereoisomer of the optimized ComplexBuild structure from the original starting structure to its enantiomeric pair, given by the permutation [1 6 4 2 8 5 3 7]. Indeed, in some cases such as this one, minute variations in the starting geometry are enough to veer the optimization path towards either one. } found in CSD by the entry RENXIR. Since each 2,2-bypiridine is neutral, the complex has a +3 charge. Despite the slight differences in the starting structures, they both converged to the exact same final optimized structure, which was closer to the starting geometry obtained from the Complex Build algorithm. Every ligand is neutral, leading to a complex with a net +3 electric charge. Once more, the greatest source of difference between the four structures are the rotations around wheel angles since these are all large monodentate ligands. Nevertheless, both final optimized structures preserved the coordination stereochemistry despite displaying distinct geometric features.