Annotation of the results obtained in 2022
The second year of work on the project was devoted to the completion of a number of tasks of the first year on a thorough search and analysis of systems, the selection of parameters and approaches for using specialized software, the calculation of the necessary energy and mechanical characteristics of molecular crystals - all with the aim of developing methods and approaches for determining and predicting organic crystals, capable of significant plastic deformation. More specifically, work has been completed on systems of pyrazinamide, 4-bromophenyl 4-benzoate, 6-chloro 2,4-dinitroaniline, and other substances whose polymorphs have different mechanical properties. Using the methods of molecular mechanics, the energies of crystal lattices of polymorphic modifications, the energies of individual layers, and the energies of layer association were calculated to determine the previously proposed anisotropy of crystals. It has been shown that neither visual analysis of layered structures nor the anisotropy of molecular interactions within and between layers can explain or predict the ability of a crystal to undergo significant plastic deformation. Next, a transition was made to the methods of the density functional theory for various calculations, including both the layer association energy and a number of macroscopic characteristics. An optimal method for calculating the association energy of layers with and without optimization of atomic positions was proposed - all calculations have excellent convergence with the methods of molecular mechanics. In the development of the approaches proposed in the literature, the use of which is limited, we proposed to study the energy barriers of sliding layers of crystals both by the methods of molecular mechanics and the density functional theory. Two models have been proposed - isolated layers and a "ladder" model for periodic calculations - to evaluate such barriers (estimation "from above" and "from below" by different approaches within the methods). It has been shown that typical values for plastic crystals are lower than for brittle forms. The transition to macroscopic characteristics turned out to be quite laborious and required the study of the phase stability of polymorphic modifications, which was successfully carried out (https://www.nsu.ru/n/media/news/nauka/uchenye-ngu-izuchili-otnositelnuyu-stabilnost-polimorfnykh-modifikatsiy-pirazinamida-v-shirokom-diap/). The flexible alpha shape has been shown to be more stable than the brittle delta shape over a wide temperature range. The mechanical properties calculated from the tensors also differed significantly and predicted well the directions of possible bending. However, the anisotropy index, as well as a number of "mechanical" moduli, did not make it possible to distinguish a brittle form from a plastic one. Thus, the completion of work in the second year of the project implementation made it possible to conclude that the model for determining bending crystals presented in the literature has a number of limitations, primarily related to its qualitative rather than quantitative criteria. We propose to use a set of approaches that make it possible to establish and predict unusual mechanical properties of molecular crystals in the future by calculating layer association energies, slip energy barriers, and also by studying and predicting crystal morphology. The results have been published in the leading specialized journal on crystallography and crystal engineering, Crystals, and generalizing materials are planned for publication in the near future.

 

Publications

---

Annotation of the results obtained in 2021
This work is devoted to the development of computational methods and approaches for determining and predicting molecular crystals capable of significant plastic deformation. Considering that this phenomenon is quite rare and does not exceed a few hundredths of a percent of all structures presented in the CCDC database, the key step was the selection of systems for research and testing of the hypotheses developed in the project. Also in the first year it was extremely important to choose the parameters of the calculations and select specific programs for subsequent calculations. Moreover, it is necessary to check the convergence of the methods themselves in order to subsequently make the transition from more resource-intensive methods to simpler and faster ones. At the end of the work carried out in the first year, it was necessary to obtain the first energy characteristics for the systems under study in order to obtain primary data for the study of individual systems and put forward the first assumptions on the conditions for the mechanical bending of organic crystals. It was this plan (in full accordance with the application) that was implemented in the first year of the work. At the first stage, a literature search was carried out by keywords, which were collected from the metadata of the ten most significant articles, in our opinion, devoted to the phenomenon of plastic deformation of molecular crystals. This procedure was automated by writing a simple Python script. According to the keywords found, articles were searched in the Scopus database. The analysis of the obtained works revealed that most of them are irrelevant in terms of objects of study (organometallic, inorganic and composite systems), in terms of effect (elastic, thermo- or photomechanical bending), or research methods (insufficient confirmation of the effect). However, after also analyzing the relevant articles, it was noticed that most of them refer to the same works, namely the works of C.M. Reddy [1-3]. Thus, we selected works and built a citation graph of articles that refer to the above works of Indian colleagues. So, 243 jobs were received, to which a few more were added from the keyword search. Already a new pool of work has been analyzed and potential systems (several dozen) have been identified for research. Obviously, the task of studying molecular crystals capable of significant plastic deformation in general fits into N.S. Kurnakov's triad "composition - structure - property", which in our case was simplified to "structure - property" by choosing polymorphic modifications the same substances. In other words, we have limited the list of systems to those where the conditional form I bends, and the conditional form II is brittle, despite the fact that the substance does not change. Thus, the factor of differences between the molecules themselves was eliminated when comparing different systems. In addition, it was necessary to determine other (technical) limitations that may arise during the calculations. To do this, we turned to the experience of our studies of bending crystals. L-leucinium hydrogen maleate crystals were chosen as such a system. This system is a good benchmark for testing the complexity of calculations by various methods and the corresponding limitations that must be avoided in the early stages of work. The system consists of 24 molecules in a complete cell (6 in an independent part), has charged molecules (molecular salt). These limitations lead to difficulties when using the specialized software proposed in the application. Technical details are given in the main report. Using computational methods, we were able not only to determine the required parameters by the number of atoms in the unit cell and charges (more precisely, their absence, i.e., systems that are not salts and zwitterions), but also the experimental behavior of this system at low temperatures and high temperatures. pressures [4]. As a result, taking into account technical limitations, 8 out of 20 systems were selected, which were subsequently reduced to 5 for research and initial testing of our hypotheses. The names of the systems in this case are not given due to the fact that these works have not yet been prepared for publication. Further, various programs and their parameters were tested to carry out full-fledged calculations. It was shown that the optimal choice of software for DFT calculations is the VASP program (license), GGA PBE functional, with an energy cutoff (Ecutoff) of 550 eV and a dispersion correction GD3BJ. The grid of k-points was chosen individually for the systems. The selection of all conditions was performed "from scratch" without the use of empirical rules in order to prove applicability for specific systems. The convergence criteria are the energy difference of no more than 2 meV per atom in an elementary cell, the difference between the calculated cell volume and the experimental one is no more than 3%. For molecular mechanics methods, the CrystalExplorer21 program was chosen (due to the more modern DFT energy parameterization in comparison with AA-CLP and PIXEL) for which the cluster radii were selected to calculate the energy of the crystal lattice. For the MM methods, it was shown that the energies converge at different radii, not exceeding 30 A. This value can be recommended for all systems that do not contain significant charges on atoms. The energies of crystal lattices were also calculated by the MM and DFT methods, where it was shown that the methods converge well and the error does not exceed 3–4 kJ/mol, which is about 3%. This proves that for our objects we can use both methods on the fitted parameters. Thus, the task of testing programs and their parameters for subsequent calculations was completed to the required (declared) extent. Having selected the systems and calculation parameters, we proceeded to carry out separate calculations in order to evaluate the models proposed in the literature and obtain quantitative characteristics of the interactions in our systems. A script was written that allows you to convert structures to other bases (changing the unit cell), the use of which facilitates further manipulations with individual layers in crystalline structures (in individual structures). For one system (all polymorphic modifications, both "flexible" and "brittle"), the energies of individual layers isolated from the crystallographic analysis, the energies of interactions of these layers with the rest of the crystal structure were calculated. It was shown that the main interaction occurs with neighboring layers, and the model can often be simplified to two layers, which greatly simplifies the calculations. Considering that one form of the chosen system is "bending" and the other "brittle" and extremely similar energies within layer and interlayer interactions, it was shown that the conditions previously formulated by C.M. Reddy are not sufficient to describe and predict the plastic deformation properties of organic crystals. Recall that, within the framework of this model, it is assumed that the crystal structure at the molecular level must meet the following conditions: it must be formed by layers (layered structure), the bonds within the layer must be sufficiently strong (qualitatively), the bonds between the layers must be relatively weak (qualitatively). Moreover, as an “advance” of the work schedule, the displacement energies of one layer relative to another were calculated, where the energies do not exceed several kJ/mol for a bending crystal shape. For all displacements, displacement energies were constructed in two directions, that is, potential energy surfaces, which show that the least energy-consuming ways of layer displacements were chosen. Similar calculations are carried out for the second selected system. All calculations with layer shifts were carried out using the molecular mechanics method in the CE21 program. Based on the data obtained, an assumption was made about the need for low slip energies (and the corresponding verification algorithm) in addition to the criteria put forward by C.M. Reddy. A more detailed interpretation will be given in the next reporting period. Additionally, macroscopic parameters (modulus of elasticity and compressibility) are calculated using the DFT method - these are long-term calculations that are done according to the scheme proposed in [5]. The data obtained is being prepared for publication. In this part of the report, specific systems are not named (given in the main report) due to the fact that these results have not yet been published. Thus, we can say with confidence that all the planned work was completed (and in some part even over-fulfilled, as in the case of sliding layers and partly by “unscheduled” publication). This greatly facilitates further research in the second year and, to a certain extent, guarantees the implementation of the project in view of the fact that now the performance of the selected methods has been shown, the parameters have been selected, and a number of assumptions have been put forward for individual interactions that can be verified by other methods and on the following systems. [1] C.M. Reddy, K.A. Padmanabhan, G.R. Desiraju, Structure−Property Correlations in Bending and Brittle Organic Crystals, Cryst. Growth Des. 6 (2006) 2720–2731. https://doi.org/10.1021/cg060398w. [2] C.M. Reddy, R.C. Gundakaram, S. Basavoju, M.T. Kirchner, K.A. Padmanabhan, G.R. Desiraju, Structural basis for bending of organic crystals, Chem. Commun. 1 (2005) 3945. https://doi.org/10.1039/b505103g. [3] C.M. Reddy, S. Basavoju, G.R. Desiraju, Sorting of polymorphs based on mechanical properties. Trimorphs of 6-chloro-2,4-dinitroaniline, Chem. Commun. (2005) 2439. https://doi.org/10.1039/b500712g. [4] K.D. Skakunova, D.A. Rychkov, Low Temperature and High-Pressure Study of Bending L-Leucinium Hydrogen Maleate Crystals, Crystals. 11 (2021) 1575. https://doi.org/10.3390/cryst11121575. [5] Y. V. Matveychuk, E. V. Bartashevich, V.G. Tsirelson, How the H-Bond Layout Determines Mechanical Properties of Crystalline Amino Acid Hydrogen Maleates, Cryst. Growth Des. 18 (2018) 3366–3375. https://doi.org/10.1021/acs.cgd.8b00067.

 

Publications

---