It is of interest to build realistic charge distribution models of biological macromolecules. For this purpose, there are computationally efficient approaches based on transferable building blocks. Transferable quantities can be electron density parameters of atoms or of functional groups, or localized orbitals giving access to molecular charge distributions [1]. The first case is at the basis of libraries of transferable multipolar pseudoatoms built either from X-ray diffraction experiment [2], or from single point quantum calculations [3,4]. Electron density parameters transferred to molecular structures from these libraries are however either averaged, or issued from gas-phase quantum calculations. They are therefore practically devoid of any intermolecular effects due to the non-covalently bonded environment. These effects should be accounted for, especially in protein-ligand complexes.

To compensate this drawback, we implemented in the MoProViewer software methods designed to account for intermolecular dipolar induction in a transferred multipolar electron distribution [5]. For this purpose, atomic anisotropic polarizabilities have been added to the definition of transferable multipolar pseudoatoms, as defined in the ELMAM2 library.

The construction of this database of polarizabilities associated to ELMAM2 transferable pseudoatoms will be described, and comparisons of the resulting polarization energies against a theoretical reference will be presented. Finally, application examples on protein ligand complexes will be discussed.

[1] Meyer B, Guillot B, Ruiz-Lopez M & Genoni A (2016). J. Chem. Theory Comput, 12, 1052.

[2] Domagala S, Fournier B, Liebschner D, Guillot B & Jelsch C (2012). Acta Cryst. A68, 337.

[3] Kumar P, Gruza B, Bojarowski S.A, Dominiak P.M. (2019). Acta Cryst. A75, 398.

[4] Dittrich B, Hübschle CB, Pröpper K, Dietrich F, Stolper T & Holstein JJ (2013). Acta Cryst. B69, 91.

[5] Leduc T, Aubert E, Espinosa E, Jelsch C & Guillot B (2019) J. Phys. Chem. A, 123, 7156.

Single-molecule magnet (SMM) is the generic name given to a broad class of molecules, which exhibit an energy barrier to magnetization reversal. In simpler terms, SMMs have that special trait that once they have become magnetized by an external magnetic field, the induced magnetic moment (which we, for simplicity, could call spin up or spin down) resists reorientation. For that reason, such fascinating molecules are envisaged to act as molecular bits, or quantum bits, qubits. The origin of this effect is magnetic anisotropy, i.e. the different magnetic response to an external field (quantified by the magnetic susceptibility) depending on the relative orientation of field and molecule. Magnetic anisotropy splits the magnetic substates, and the reason for this is the presence of unquenched orbital angular momentum. Thus, at the very core, to be able to develop novel SMMs we need to understand how to control the electronic ground state of a complex. This has followed two paths, depending on whether the electron-carrier is a 3d or 4f element.

For 4f-based SMMs, a widespread approach has aimed at developing complementary ligand fields relative to the valence electron density shape of the most magnetic Mj-state of the 4f-ion in question. However simple and unvalidated by experiment, this approach has been fantastically useful. Recently, we showed how the experimental electron density from X-ray diffraction could reveal hitherto unseen details in the electronic structure of a Dy-based SMM, thus elucidating the mechanism[1]. For 3d-systems, the ligand field is much stronger and the approach is thus different. The magnetic anisotropy is enhanced in distorted tetrahedral complexes of Co^II, as has recently been shown[2-4].

Herein, I will show how a combination of high-resolution synchrotron X-ray diffraction (XRD) and polarized neutron diffraction (PND) can be used to quantify the magnetic anisotropy in [CoX₂tmtu₂] (X=Cl, Br, tmtu = tetramethylthiourea). The XRD data provides a multipole model of the electron density, while the PND provides the full magnetic susceptibility tensor. The experimental results are supported by ab initio calculations.

Figure 1. ORTEP drawing of the Cl-complex studied here based on 20 K synchrotron data, showing 90% ellipsoids.

[1] Gao, C., Genoni, A., Gao, S., Jiang, S., Soncini, A. & Overgaard, J. (2020). Nat. Chem. 12, 213.

[2] Vaidya, S., Shukla, P., Tripathi, S., Rivière, E., Mallah, T., Rajaraman, G. & Shanmugam, M. (2018). Inorg. Chem. 57, 3371.

[3] Rechkemmer, Y., Breitgoff, F. D., van der Meer, M., Atanasov, M., Hakl, M., Orlita, M., Neugebauer, P., Neese, F., Sarkar, B. & van Slageren, J. (2016). Nat. Commun. 7, 10467.

[4] Damgaard‐Møller, E., Krause, L., Tolborg, K., Macetti, G., Genoni, A. & Overgaard, J. (2020). Angew. Chem. Int. Ed. 59, 21203.

X-ray charge density is the most powerful experimental method to study interatomic and intermolecular interactions, such as two-electron multicentric (2e/mc) covalent bonding [1-3]. However, it is limited to high-quality crystals and good enough data can be collected only at low temperature and ambient pressure. In order to gain more information on behaviour of novel 2e/mc interactions, a broader range of conditions (temperatures and pressures) are required. These are normally limited to resolutions of 0.8 Å or lower and are thus unsuitable for multipolar refinement and study of charge density.

If good high-resolution diffraction data are not available, charge density can be obtained using transferrable multipoles from optimal data set [4]. Thus, multipoles obtained by multipolar refinement of high-resolution data can be transferred to lower-resolution variable-temperature (VT) and high pressure (HP) diffraction data, allowing us to study charge density at a broad range of conditions. We have tested this method in study of 2e/mc bonding in 4-cyano-N-methylpyridinium salt of 5,6-dichloro-2,3-dicyanosemiquinone radical anion ([4-CN-N-MePy]⁺[DDQ]^-), which we have recently studied by VT and HP X-ray diffraction [5] and by X-ray charge density [6]. Multipolar parameters obtained by a multipolar refinement of high-resolution data measured at 100 K [6] were thus transferred to lower-resolution VT and HP data; the results and their validity are discussed. Since 2e/mc is an intermolecular interaction, which involves a non-localised electron pair, its electron density is low; so its study is less reliable than that of stronger intramolecular covalent bonding. Therefore, our transferred-multipole models must satisfy the following three criteria to be considered valid:

(i) overall reduction of disagreement R-factors and residual density compared to regular spherical refinement;

(ii) electron densities should follow a clearly defined trend;

(iii) experimentally obtained electron densities should be in a good agreement with theoretical ones.

[1] Kertesz, M. (2018). Chem. Eur. J., 25, 400-416.

[2] Molčanov, K. & Kojić-Prodić, B. (2019). IUCrJ, 6, 156-166.

[3] Molčanov, K.; Milašinović, V. & Kojić-Prodić, B. (2019). Cryst. Growth Des., 19, 5967-5980.

[4] Domagała, S.; B. Fournier, D. Liebschner, B. Guillot, Jelsch, C. (2012). Acta Cryst. A., A68, 337-351.

[5] Bogdanov, N. E.; Milašinović, V.; Zahkarov, B.; Boldyreva, E. V.; Molčanov, K. (2020). Acta Cryst. B., B76, manuscript XK5067, in print.

[6] Milašinović, V.; Krawczuk, A.; Kojić-Prodić, B.; Molčanov, K. (2020). Manuscript in preparation.

Keywords: charge density; high pressure; variable temperature; transferrable multipoles; two-electron multicentric bonding

This work was funded by the Croatian Science Foundation, grant no. IP-2019-04-4674.

Non-covalent interactions uniquely define the structure of macromolecules. Therefore, a thorough analysis of the non-covalent interaction network is crucial to gain insights into functions and dynamics of macromolecules.

A strategy that is able to detect non-covalent interactions for a large variety of molecules is the Non-Covalent Interactions (NCI) method [1,2], a technique simultaneously based on the electron densities and the reduced density gradients of the molecules under exam. Unfortunately, accurate molecular electron densities can be obtained through traditional quantum chemistry computations at a feasible computational cost only for small to medium-sized systems, whereas these calculations become impractical for larger molecules. Therefore, until now, for NCI analyses on large systems one had to resort to the promolecular density approximation, where the electron density of the investigated molecule is described as a sum of independent and spherically averaged atomic densities. These promolecular densities lack accuracy, and although they might lead to visually similar results when compared to those obtained from fully quantum mechanical calculations, the underlying electron density is known to be incorrect. Hence, the analysis of the non-covalent interactions is also biased.

To overcome the previous shortcoming, one should exploit techniques that allow to rapidly obtain accurate and reliable electron densities for macromolecules. In this context, one possibility is represented by the recently constructed database of extremely localized molecular orbitals (ELMOs) [3-5]. In fact, ELMOs are orbitals strictly localized on small molecular fragments, i.e. atoms, bonds or functional groups [3]. Due to this strict localization, they are easily transferable from one molecule to another, provided that the subunits on which they are localized have the same chemical environment in the starting and final systems [3,4]. By exploiting this intrinsic transferability, a databank of ELMOs has been constructed [5]. It currently contains orbitals associated with all the fragments for the twenty natural amino acids and allows rapid and reliable reconstructions of wavefunctions and electron densities of very large biomolecules.

The coupling of the NCI technique with the ELMO database gave rise to the new NCI-ELMO method [6] that was successfully applied to analyse a variety of non-covalent interactions in polypeptides and proteins. Test calculations showed that qualitative results obtained with the NCI-ELMO technique are very similar to the ones based on fully quantum chemical calculations, but definitely better than those resulting from the promolecular-NCI approach. In this presentation, the previously mentioned qualitative results [6] will be discussed. Additionally, we will illustrate how the new NCI-ELMO technique has been recently extended to quantify non-covalent interactions. Other than applications to protein-ligand interactions, we will show the results of benchmark calculations on smaller systems (e.g., simple molecular dimers) to highlight the differences between the NCI-ELMO and promolecular-NCI approaches also at a quantitative level.

[1] Johnson, E. R.; Keinan, S., Mori-Sanchez, P., Contreras-García, J., Cohen, A. J. & Yang, W. (2010). J. Am. Chem. Soc. 132, 6498.

[2] Contreras-García, J., Johnson, E., Keinan, S., Chaudret, R., Piquemal, J.-P., Beratan, D. & Yang, W. (2011). J. Chem. Theory Comput. 7, 625.

[3] Meyer, B., Guillot, B., Ruiz-Lopez, M. F. & Genoni, A. (2016). J. Chem. Theory Comput. 12, 1052.

[4] Meyer, B., Guillot, B., Ruiz-Lopez, M. F., Jelsch, C. & Genoni, A. (2016). J. Chem. Theory Comput. 12, 1068.

[5] Meyer, B. & Genoni, A. (2018). J. Phys. Chem. A 122, 8965. [6] Arias Olivares, D., Wieduwilt, E. K., Contreras-García, J. & Genoni, A. (2019). J. Chem. Theory Comput. 15, 6456.

Eutectics are well-known multi-component systems used in various day-to-day applications. However, they are enigmatic in terms of structural organization (interactions and packing, the two prime features of a crystalline entity), despite having a long history. At the microstructural level, they are phase-separated (multi-phasic) solid solutions i.e. they are heterogeneous crystalline materials composed of homogeneous (single-phase) but multiple solid solutions [1]. This phase heterogeneity in structural integrity is what makes them complex-to-understand materials. Although research has been done in understanding the eutectic structural organization particularly in inorganic systems using advanced techniques such as atomic pair distribution function (PDF) analysis, X-ray microtomography, and atomic force microscopy (AFM), no comprehension of eutectic microstructural integrity was achieved [2]. Furthermore, the structural and microstructural arrangement of organic eutectic systems has not been addressed so far in the literature [3,4]. This complexity in organic eutectic systems is augmented by several aspects such as 1) the constituents are primarily C, H, N and O which makes them soft materials, 2) atomic number contrast essential to image the microstructure is lacking, 3) frequent existence of polymorphism, 4) occurrence in lower structural symmetry. In this regard, one can transfer the knowledge of inorganic eutectics to organic eutectics or can verify the organic eutectics with competent experimental techniques in search of an improvised understanding from the molecular perspective. Here, we manage to solve the microstructural features of organic eutectics through in-situ variable temperature (VT) PXRD experiments, DSC experiments with multiple heating and cooling cycles, in-situ VT Raman spectroscopic studies, gas-phase energy calculations using Gaussian09 and electron microscopy imaging technique on a series of systems. We observe for the first time, the evolution of eutectic systems through the formation of multi-domain eutectic particles at higher temperatures. The eutectic particles melt altogether near the melting point of the eutectic system as showed in DSC experiments, via thermal energy induced heteromolecular interaction through the domain boundaries as confirmed from VT-Raman studies.

A key feature of metal-organic frameworks (MOFs) is their ability to capture, transport, and release guest molecules. The nature, quality, and quantity of the associated absorption depend on pore size and volume, surface area, solvent, and in particular the host-guest intermolecular interactions.

Various methods for the analysis of intermolecular interactions have been described in the literature and were applied to study e.g. chemical reactivity, catalysis, biomolecular interactions, or organic electronics. However, the application of such methods to host-guest interactions in MOFs is still scarce. For this reason, we computed periodic and finite wavefunctions for well-chosen MOF-guest systems and tested these tools [1]. This includes the interaction energy, its decomposition with different energy decomposition schemes, investigation of the electron density with Bader’s quantum theory of atoms in molecules, the non-covalent interaction index [2], or the density overlap regions indicator [3]. This analysis contributes to the understanding of host-guest interactions, with the ultimate goal of rationally designing MOFs for targeted applications.

[1] Ernst, Michelle; Gryn'ova, Ganna (2021): Strength and Nature of Host-Guest Interactions in Metal-Organic Frameworks from a Quantum-Chemical Perspective. ChemRxiv. Preprint. https://doi.org/10.26434/chemrxiv.14363024.v1

[2] Johnson, E. R., Keinan, S., Mori-Sánchez, P., Contreras-García, J., Cohen, A. J., & Yang, W. (2010). J. Am. Chem. Soc. 132, 6498.

[3] De Silva, P., & Corminboeuf, C. (2014). J. Chem. Theory Comput. 10, 3745