Very close I⋯As and I⋯Sb interactions in trimethylpnictogen-pentafluoroiodobenzene cocrystals

Abstract

The cocrystals (CH₃)₃As·C₆F₅I (1) and (CH₃)₃Sb·C₆F₅I (2) were generated in situ from equimolar mixtures of their components. 1 and 2 show very close I⋯As and I⋯Sb directional intermolecular interactions. They are 0.5 and 0.7 Å shorter than the sums of van der Waals radii, respectively, and are the shortest C–I⋯As and C–I⋯Sb halogen bonds of this type found for experimentally characterized molecular (co)crystals. Comparisons of the packing motifs and contacts in 1 and 2 with those in (CH₃)₃As (3), (CH₃)₃Sb (4) and C₆F₅I (5) illustrate the occurrence and hierarchy of the specific interactions. The heteromolecular components in 1 and 2 are assembled by I⋯As, I⋯Sb and F⋯H interactions. There are no significant intermolecular As⋯As contacts in 3, but Sb⋯Sb interactions in 4. Molecules in 5 are mainly associated by I⋯F, F⋯F and F⋯C contacts. The intermolecular interactions observed in the (co)crystals correspond to the calculated electrostatic potentials.

---

Introduction

Chemical reactions carried out in research laboratories and chemical industry use the transformations of substrates that react with each other under specific conditions. Chemical transformations are in most cases performed in the liquid phase and involve breaking of chemical bonds and formation of new bonds. The obtained products are separated and purified, often using crystallization techniques. Cocrystallization has a lot in common with the broadly defined chemical transformations and syntheses. However, during cocrystal formation typical chemical bonds are neither broken nor formed, but rather the process involves the creation and balancing of subtle and specific attraction forces, i.e. typically weak intermolecular interactions formed between the cocrystal components. The products of cocrystallization show, similar to products of chemical reactions, unique physicochemical properties. The possible predictability of properties, on the basis of structural behaviour of the cocrystal components, as one of the benefit of cocrystallization should be mentioned. Therefore, cocrystals are not only interesting in terms of the molecular view of their growth, self-assembly, molecular recognition or polymorphism, but also from the point of view of (co)crystal engineering, i.e. the design of new materials including pharmaceuticals, with dedicated physicochemical properties.¹ Structural chemistry abounds in a number of relatively recently discovered and described intermolecular interactions. So, not only hydrogen or halogen bonds, but also e.g. chalcogen, pnictogen, tetrel and regium bonds have been defined as attraction forces.²

Cocrystals can be regarded as the products of processes involving the formation of heteromolecular interactions. One of the best models for the investigation of weak intermolecular interactions are cocrystals composed of simple and neutral molecules. Such components can form only a limited number of well-defined interactions. We have recently reported cocrystals of halogenated hydrocarbons, which are gases or liquids under ambient conditions, with benzene. These studies indicate that, although the molecular components form relatively weak heteromolecular interactions, they are strong enough to support the formation of cocrystal assemblies under the given thermodynamic conditions.³

Herein, we report the formation of the cocrystals (CH₃)₃As·C₆F₅I (1) and (CH₃)₃Sb·C₆F₅I (2) by the supramolecular reactions of the Lewis bases (CH₃)₃As (3) and (CH₃)₃Sb (4) with the halogen-bond donor Lewis acid C₆F₅I (5) and the identification of 1 and 2 by single-crystal X-ray diffraction. The structures of both pure components 3 and 4 were also investigated for comparison. Hirshfeld surface analysis and quantum-chemical calculations support the experimental study.

Results and discussion

Single-crystal samples of 1 and 2 were grown in situ at low temperatures in sealed capillaries from cryogenically loaded equimolar mixtures of the components. Their single-crystal structures were determined at several temperatures between 210 and 100 K. X-ray diffraction revealed that 1 and 2 crystallize in the monoclinic space groups C2/c and I2/a, respectively, with one (CH₃)₃As or (CH₃)₃Sb as well as one C₆F₅I molecule in the asymmetric units. There was no phase transition or symmetry change observed down to 100 K. Both cocrystals experience a similar contraction with decreasing temperature: the unit-cell volumes at 100 K are 3.2 and 3.1% less than those at 220 and 210 K for 1 and 2, respectively. All unit-cell parameters decrease upon cooling, except the unit-cell length c and angle β for 2, which increase. The largest decrease of 2.5% was found for the smallest length a in 1 (Tables S1 and S2, ESI†).

Cocrystals 1 and 2 can be considered to consist of alternating molecular layers composed of their molecular components (Fig. 1). At 100 K, the C₆F₅I molecules in both cocrystals are basically planar with expected intramolecular dimensions.^3b,4 The average E–C distances (E = As or Sb) within the (CH₃)₃E units in 1 and 2 at 100 K are 1.945 and 2.134 Å, whereas the C–E–C angles are 98.5 and 95.5°, respectively (Tables S5 and S6, ESI†). The difference between the E–C bond lengths of 1 and 2 corresponds to the difference in the covalent radii of As and Sb of ca. 0.2 Å.⁵ The r_a-structures of the free molecules (CH₃)₃E (E = As, Sb) determined by gas electron diffraction show E–C distances of 1.968(3) and 2.163(3) Å and C–E–C angles of 96.1(5) and 94.1(5)°, respectively.⁶ This means that cocrystal formation with 5, leads to a shortening of the E–C distances by 0.023 and 0.029 Å, and a widening of the C–E–C angles by 2.4 and 1.4° for 1 and 2, respectively. Although these changes are close to the experimental errors, they are reasonable in terms of the valence shell electron pair repulsion (VSEPR) model.⁷ Both the above-mentioned intramolecular bond lengths and angles, for (CH₃)₃As and (CH₃)₃Sb in 1 and 2, respectively, are comparable to previous results obtained for analogous cocrystals.⁸

Fig. 1 The structures of 1 (a) and 2 (b) at 100 K in views along the a (1) and b axes (2). The intermolecular voids accessible to a probing sphere of radius 0.65 Å and grid spacing of 0.6 Å are indicated in yellow. The void volumes are 2.0% (50.03 ų, for 1) and 5.8% (154.04 ų, for 2).⁹ The dotted red lines indicate the I⋯As, I⋯Sb, F⋯F, F⋯C and F⋯H intermolecular contacts. Displacement ellipsoids are plotted at the 25% probability level.

For the purpose of comparing and better understanding the structural properties and interaction behaviour in 1 and 2, single crystals of the trimethylpnictogen components (CH₃)₃As (3) and (CH₃)₃Sb (4) were grown in situ at low temperatures. Their structures were determined at several temperatures between 170 and 100 K. 3 and 4 show orthorhombic and monoclinic symmetry in the space groups Cmc2₁ and P2₁/n, respectively. The asymmetric units contain half a (CH₃)₃As and one (CH₃)₃Sb molecule. 4 demonstrates a similar overall contraction with decreasing temperature as 1 and 2, whereas the contraction of 3 is weaker. The unit-cell volumes at 100 K for 3 and 4 are 1.3 and 2.8% less than those at 150 and 170 K, respectively. All unit-cell lengths, except the slight increase of the β angle for 4, shrink with decreasing temperature. The largest decrease of 1.2% was noticed for the smallest unit-cell parameter a in 4 (Tables S3 and S4, ESI†). Although 3 and 4 show different crystal structures, both can be regarded being composed of molecular layers (Fig. 2). As in 1 and 2, the difference in As–C (av. 1.963 Å) and Sb–C (av. 2.151 Å) bond lengths is ca. 0.2 Å at 100 K.⁵ The average C–As–C and C–Sb–C angles are 96.7° and 94.6°, respectively (Tables S7 and S8, ESI†). The values for the E–C distances and C–E–C angles are similar to those for the free molecules determined by gas electron diffraction⁶ (see above) and independently confirm the longer distances and wider angles in 1 and 2.

Fig. 2 The structures of 3 (a) and 4 (b), at 100 K, along a axes. The intermolecular voids accessible to a probing sphere of radius 0.65 Å and grid spacing of 0.6 Å are indicated in yellow. The void volume is: 1.8% (9.64 ų, for 3) and 1.1% (6.52 ų, for 4).⁹ The dotted red lines indicate the Sb⋯Sb intermolecular contacts in 4. Displacement ellipsoids are plotted at the 25% probability level.

Both cocrystals, 1 and 2, show similar patterns of voids and interactions schemes⁹ that are dominated by the heteromolecular I⋯As and I⋯Sb, but also by weaker F⋯H contacts (Fig. 1, Tables S9 and S10, ESI†). In contrast, within the limits of van der Waals radii,¹⁰3 does not show any pronounced intermolecular interactions (at 100 K the shortest are H⋯H distances of 2.62 Å, while the closest As⋯As separations are 5.172(1) Å), whereas the molecules in 4 are joined together only by Sb⋯Sb contacts (Fig. 2, Table S11, ESI†). Within the intermolecular distances the largest temperature-dependent differences were noted for the F⋯C and F⋯F contacts in 1 and 2, respectively; they become shorter both by ca. 0.07(1) Å upon cooling to 100 K. Simultaneously, the I⋯As and I⋯Sb distances contract by 0.063(1) and 0.051(1) Å, reaching 3.3044(3) and 3.4951(4) Å, respectively. These are the shortest C–I⋯E (E = As or Sb) halogen bonds of this type for experimentally studied molecular (co)crystals. On the basis of their geometrical parameters, the intermolecular interactions in cocrystals 1 and 2 can be classified as distorted electrostatically preferred type II halogen bonds for I⋯E, and type I dispersive forces for F⋯H contacts.^2b,11

Among the properties of interactions that indicate their stabilizing role in (co)crystal structures are energy and related geometry as well as frequency of appearance. The geometry and relative distribution of interactions in 1 and 2 along with those in 3 and 4 were clarified and compared using the Hirshfeld surface analysis.¹² This method covers all possible contacts including those with separations above the van der Waals radii regime¹⁰ and allows identifying particular types of contacts together with their impact on (co)crystal packing.

The Hirshfeld surfaces were independently generated for the molecular components in 1 and 2 as well as for the molecules in 3 and 4. The red regions, on the Hirshfeld surfaces, correspond to I⋯As and F⋯F/C contacts in 1, and I⋯Sb and F⋯F interactions in 2 (Fig. 3). The red zone marked on the Hirshfeld surface of (CH₃)₃Sb in 4 relates to the Sb⋯Sb contacts (Fig. 4).

Fig. 3 Hirshfeld surfaces (generated separately for (CH₃)₃As, (CH₃)₃Sb and C₆F₅I molecules) mapped with d_norm (−0.1000–1.0000), at 100 K, for: 1 (a) and 2 (b). The dotted red lines indicate the I⋯As and I⋯Sb intermolecular interactions. Overlapping surfaces are marked in red, touching in white, and separated in blue.

Fig. 4 Hirshfeld surfaces for (CH₃)₃As and (CH₃)₃Sb molecules, mapped with d_norm(−0.1000–1.0000), at 100 K, for: 3 (a) and 4 (b), respectively. Overlapping surfaces are marked in red, touching in white, and separated in blue.

Considering all possible contacts of the individual molecular (co)crystal components in the structures 1–4, it is obvious that the interactions involving H atoms should mainly contribute to the Hirshfeld surface area (Fig. 5 and S1, ESI†). The interactions in 3 and 4 are characterized by only three types of contacts: E⋯E, E⋯H and H⋯H (E = As for 3 and E = Sb for 4). The conversions 3 → 1 and 4 → 2 by formal incorporation of C₆F₅I molecules leads to a complete change in the interaction patterns. In both 1 and 2, the ‘new’ F⋯H, I⋯H and C⋯H contacts constitute ca. 51% of all cohesion forces. In addition, and most intriguingly, very short I⋯As and I⋯Sb interactions are introduced; their contribution is about 6%.

Fig. 5 Distribution of contacts, on the basis of their relative contributions to the Hirshfeld surface area, for (CH₃)₃E molecules, in: 1–4 at 100 K. E = As for 1 and 3, and E = Sb for 2 and 4.

A similar inspection of the formal conversions 5 → 1 and 5 → 2 by incorporation of (CH₃)₃E molecules into the known polymorphic forms 5-I⁴ and 5-II^3b of C₆F₅I (5) leads to related observations (Fig. 6 and S2, ESI†). The dominating I⋯F, F⋯F and F⋯C contacts with the total average contribution of ca. 88% in 5-I and 5-II are reduced to an average ca. 43% in both 1 and 2. In turn, the input of above-mentioned ‘new’ F⋯H, I⋯H and C⋯H contacts constitute at least 51% of all cohesion forces in 1 and 2. Also, both cocrystals contain the important and strong contributions of the I⋯As and I⋯Sb interactions of ca. 5% on average.

Fig. 6 Distribution of contacts, on the basis of their relative contributions to the Hirshfeld surface area, for C₆F₅I molecules in 1 and 2 and two polymorphic forms of C₆F₅I 5-I⁴ at 150 K and 5-II^3b at 100 K. E = As for 1, and E = Sb for 2.

This analysis demonstrates that both the nature of molecular components and the distribution of particular interactions are the factors that are responsible for cocrystal formation. It also clearly shows the real possibility of a supramolecular engineering of dedicated interactions established by certain combinations of donor- and acceptor-type functions. Note that the contact distances, like in 3, do not have to be rigorously limited by the sums of the van der Waals radii of the relevant atoms¹³ as well as that the intermolecular distances are shortened upon decreasing the temperature. Moreover, the observed molecular aggregations in all studied compounds correlate with the calculated electrostatic potentials and polarizability of molecules (Fig. S3, ESI†).

To understand the nature of binding in the studied systems we applied an analysis using the local energy decomposition (LED) method.¹⁴ For this, we investigated the isolated molecules of 1–5 at the DLPNO-CCSD(T)/def2-QZVPP//PBE0-D3BJ/def2-QZVPP level of theory. This approximation was chosen after a calibration procedure using two model systems, H₃As⋯ICH₃ and H₃Sb⋯ICH₃. This procedure and other details of calculations are provided in the ESI.† The LED results are summarized in Table 1. Here we focus on the properties of the title systems in comparison to the other calculated molecular model systems including H₃N⋯ClF, HF⋯HF and CH₄⋯CH₄, which represent dominating halogen, electrostatic and dispersion-type interactions, respectively. The two calibration systems H₃As⋯ICH₃ and H₃Sb⋯ICH₃ are also included for comparison.

Table 1 Interaction energies (kcal mol⁻¹) of investigated molecular systems in LED analysis^a

	ΔEtot^b	ΔEtot		ΔEint		ΔE^Cint/ΔEint,^g %	E ^C disp/ΔEint,^h %	E ^C CT/ΔE^Cint,^i %	E ^C (T)/ΔE^Cint,^j %
		ΔEint^c	ΔEgeo-prep^d	ΔE^HFint^e	ΔE^Cint^f
a Calculations are at the DLPNO-CCSD(T)/def2-QZVPP//PBE0-D3BJ/def2-QZVPP level. b Total binding energy. c Total interaction energy. d Geometry preparation energy (deformation energy). e Interaction energy at the HF level. f Total correlation part of the interaction energy. g Contribution of the correlation into the total interaction energy. h Contribution of the dispersion interaction into the total interaction energy. i Contribution of the charge transfer part of the correlation into the total correlation interaction energy. j Contribution of triple excitations into the total correlation interaction energy.
(CH3)3As⋯IC6F5	−4.96	−5.59	0.63	0.83	−6.42	115	95	152	28
(CH3)3Sb⋯IC6F5	−4.12	−4.51	0.39	1.61	−6.12	136	106	129	27
H3As⋯ICH3	−1.52	−1.51	0.00	1.54	−3.06	202	157	75	19
H3Sb⋯ICH3	−1.57	−1.58	0.01	1.75	−3.33	211	159	78	19
H3N⋯ClF	−9.35	−10.84	1.49	−5.00	−5.84	54	55	380	49
HF⋯HF	−4.58	−4.65	0.07	−3.82	−0.83	18	23	809	36
CH4⋯CH4	−0.47	−0.47	0.00	0.37	−0.84	179	173	13	15

---

The calculated systems are stabilized due to interactions between their components and, on the other hand, destabilized because of the geometrical deformation of their components during adduct formation. The latter is especially notable for (CH₃)₃E⋯IC₆F₅ (here and below E = As or Sb) and H₃N⋯ClF. The total interaction energy is decomposed into Hartree–Fock (HF) and correlation (C) contributions. At the HF level the interaction in both (CH₃)₃E⋯IC₆F₅ systems is repulsive and the overall stabilization is solely due to electron correlation. This is also true for H₃E⋯ICH₃, in which the relative contribution of correlation in the total interaction energy is even larger (up to about 200% of the total interaction energy). H₃N⋯ClF shows a different picture: it is already stable at the HF level and additionally stabilized to a similar extent by correlation effects. The smallest influence of the electron correlation on the binding energy is in HF⋯HF.

The energy attributed to dispersion interactions between the components can be separated from the correlation part of the interaction energy. Table 1 lists these values relative to the total interaction energy; this allows a simplified characterization of the interaction type. Values larger than 100% clearly indicate dispersion-driven interactions, values between 50 and 100% can be related to interactions with balanced contributions of dispersive and non-dispersive nature, while values below 50% are specific for cases with primarily non-dispersive types of interactions. In (CH₃)₃E⋯IC₆F₅ these values are 95 and 106% for E = As or Sb, respectively. This shows importance of dispersion interactions in the stabilization of these systems.

In addition, it is worth to discuss the contributions of charge transfer (CT), relative to the total correlation interaction energy. Note, in LED analysis only those CT components are calculated, which originate from the correlation part of the interaction. Our calculations show that for (CH₃)₃E⋯IC₆F₅ both, dispersion and charge transfer effects, are important. The latter is larger for the system with arsenic (152%) than that with antimony (129%). In the smaller H₃E⋯ICH₃ systems, the relative importance of these electronic effects is different. The dispersion contribution is significantly larger, while the CT is less pronounced. In H₃N⋯ClF the contribution of dispersion is relatively small, 55% of the total interaction energy, whereas that of charge transfer is large, 380% of total correlation part of the interaction energy. The extreme example of this type of interaction is provided by the dimer HF⋯HF, showing very weak dispersion and very strong CT components. The obverse extreme example is the methane dimer, in which dispersion plays the most important role (173%), whereas the charge transfer contribution is very small (13%).

We also calculated the relative contributions of triple excitations within the correlation part of the interaction energy. They are the largest in the cases with dominant CT interactions, up to 49% in H₃N⋯ClF, and far less important for systems with strong dispersion-type component. This can be taken into account in calculations of properties for other molecular systems.

I⋯As and I⋯Sb interactions on the basis of 6-OI-fulvene with pnictogen electron donors have been recently investigated theoretically by others.¹⁵ In this work the MP2 method was used for geometry optimizations and calculations of stabilization energies. It is known that this method is not suitable for an accurate description of systems with significant contribution of dispersion.¹⁶ Thus, the reported stabilization energies for the aforementioned I⋯As and I⋯Sb systems are likely to be overestimated. An approach to energy decomposition these systems was undertaken as part of an NBO analysis. We note, that in MP2 calculations the NBO analysis is performed for the HF part. As we demonstrate, this is the destabilizing part of the total energy, at least for 1 and 2. Thus, the analysis of the HF energy does give the understanding of the stabilization effects.

It is worth paying attention to the fact that the I⋯As interaction of 3.3044(3) Å, in 1 at 100 K, is the shortest one of this type found for any experimentally characterized molecular (co)crystal. A search (crystal structure data without errors) of the Cambridge Structural Database^9,17 for the ‘I⋯As’ structural motif revealed only four compounds in which the contact distances between As and I are shorter than in 1, but they are either between As and isolated iodide ions¹⁸ or constitute a secondary bond within an [As₂I₈]²⁻ anion.¹⁹ Also, a similar search of the structural fragment ‘I⋯Sb’ for contact distances that are shorter than that found in 2 at 3.4951(4) Å, resulted mainly in Sb–I secondary bonds in compounds showing polymeric {[SbI₅]²⁻}_n chains,²⁰ interactions between Sb and isolated iodide ions²¹ or in I⋯Sb contacts in dialkylcarbamodithioato-iodo-antimony(III) complexes.²² There are two molecular crystals showing shorter interactions, one of them is described as a (CH₃)₃Sb·SbI₂CH₃ adduct with I⋯Sb distances of 3.400 and 3.479 Å,²³ whereas the second one is (CH₃)SbI₂ with antimony-involving halogen bonds of 3.397 and 3.469 Å.²⁴ The distorted type II I⋯Sb halogen bonds, similar to that in 2, were only found between (CH₃)SbI₂ molecules. The existence of intermolecular C–I⋯As and C–I⋯Sb metal-involving halogen bonds was described only for two systems, that are similar to 1 and 2.^9,17 These are the products of cocrystallization of 1,3,5-trifluoro-2,4,6-triiodobenzene (tftib) with triphenylarsine and -stibine ((tftib)(AsPh₃) and (tftib)-(SbPh₃), respectively). In contrast to 1 and 2 the interactions in (tftib)(AsPh₃) and (tftib)(SbPh₃) are less linear and their separation distances I⋯As and I⋯Sb of 3.4211(3) and 3.5747(3) Å, respectively, are significantly longer (Fig. S5, ESI†).⁸

Conclusions

Investigating small and structurally similar molecular systems with a limited number of well-defined interaction types, we have shown a successful approach to supramolecular reactions and (co)crystal engineering – an engineering of the very close, directional interactions that control the organization of cocrystals. Both, (CH₃)₃E·C₆F₅I (E = As: 1 and E = Sb: 2), show very close I⋯As and I⋯Sb intermolecular interactions, that represent directional and the shortest C–I⋯As and C–I⋯Sb halogen bonds of this type for any experimentally characterized molecular (co)crystal. The analysis of the overall picture of interactions in 1 and 2, as well as in their molecular components 3–5 shows the significance of certain short heteromolecular contacts, mainly charge transfer (CT) and dispersion interactions, that can be considered as the main cohesion forces holding the cocrystal components together and determining their relative geometric and packing arrangements. This observation is supported by the Hirshfeld surface analysis and quantum-chemical calculations.

Experimental

All chemicals were used as received from commercial sources. The single crystal samples of all (CH₃)₃As·C₆F₅I (1), (CH₃)₃Sb·C₆F₅I (2), (CH₃)₃As (3) and (CH₃)₃Sb (4) were obtained in the in situ low-temperature (co)crystallization experiments. A procedure of the in situ low-temperature (co)-crystallization technique was earlier reported and described in detail.^3,25 The samples of 3 and 4 as well as 1 and 2 – equimolar mixtures of 3 and 4 with C₆F₅I (5), respectively, were cryogenically loaded, using the Schlenk technique, into thin-walled glass capillaries, ca. 0.3 mm in diameter. The liquid samples were cooled on the diffractometer, in a nitrogen gas stream of an Oxford Cryosystems cooler. The samples initially froze as milky polycrystalline materials, which were then slowly warmed close to their melting points to reduce the number of (co)crystal seeds. Then, by manual local warming, monitored under a microscope mounted on the diffractometer, everything but a few crystal seeds was melted followed by slow cooling (rate of 1–5 K h⁻¹). This procedure was repeated until a single seed crystal could be selected and grown. First data sets were collected at the highest possible temperatures restricted by the melting points²⁶ and stability of the studied samples, then the samples were successively cooled and the X-ray scattering intensity data were collected at one or two low-temperature points and finally at 100.0(1) K. There was no indication of any phase change/transition, from the collected diffraction patterns for all single (co)crystals 1–4.

The low-temperature X-ray diffraction intensity data were gathered on a SuperNova diffractometer, with MoKα radiation (λ = 0.71073 Å). The reflections were measured using the ω-scan technique. All data were accounted for the Lorentz, polarization and sample absorption effects. The CrysAlisPro program was used for the data collection, unit-cell determination and data reduction processes.²⁷ The structures were solved by direct methods and refined with SHELX.²⁸3 was refined as an inversion twin. All the non-hydrogen atoms, in the studied (co)crystals, were refined with anisotropic displacement parameters. All hydrogen atoms were located in the subsequent difference Fourier maps and then the riding model was applied (in the case of 3 the geometrical restrains were also used). The hydrogen atom displacement parameters were taken with coefficient being 1.5 times larger than the respective equivalent isotropic parameters of their carrier carbon atoms.

The selected crystal data and structure determination summary are listed in Tables S1–S4 in the ESI.† The bond lengths, angles and the shortest intermolecular contacts are gathered in Tables S5–S11, ESI.† The intermolecular contacts were compared using the Hirshfeld surface analysis provided by CrystalExplorer17.¹² The structure drawings were prepared using Mercury.⁹ The details on quantum-chemical calculations are provided as a supplementary material in ESI.†

Author contributions

M. Bujak: (co)crystallization, structure determination, XRD data analysis, quantum-chemical calculations and analysis, writing, reviewing and editing the manuscript. H.-G. Stammler: (co)crystallization, structure determination. Yu. V. Vishnevskiy: quantum-chemical calculations and analysis, writing. N. W. Mitzel: sample preparation, writing, reviewing and editing the manuscript.

Conflicts of interest

The authors declare no conflicts of interest.

Acknowledgements

We gratefully acknowledge financial support from Deutsche Forschungsgemeinschaft (DFG, grant MI477/28-2, project no. 271386299 and grant VI 713/1-3). We are also grateful to HPC facilities at the Universität zu Köln for providing computational time and programs. We thank T. Glodde (Universität Bielefeld) for laboratory assistance.

Notes and references

1. (a) C. Aakeröy, Acta Crystallogr., Sect. B: Struct. Sci., Cryst. Eng. Mater., 2015, 71, 387–391; (b) S. Cherukuvada, R. Kaur and T. N. Guru Row, CrystEngComm, 2016, 18, 8528–8555; (c) M. Karimi-Jafari, L. Padrela, G. M. Walker and D. M. Croker, Cryst. Growth Des., 2018, 18, 6370–6387; (d) L. Sun, W. Zhu, F. Yang, B. Li, X. Ren, X. Zhang and W. Hu, Phys. Chem. Chem. Phys., 2018, 20, 6009–6023; (e) O. N. Kavanagh, D. M. Croker, G. M. Walker and M. J. Zaworotko, Drug Discovery Today, 2019, 24, 796–804.
2. (a) G. R. Desiraju and T. Steiner, The Weak Hydrogen Bond: In Structural Chemistry and Biology, Oxford University Press, Oxford and New York, 1999; (b) G. Cavallo, P. Metrangolo, R. Milani, T. Pilati, A. Priimagi, G. Resnati and G. Terraneo, Chem. Rev., 2016, 116, 2478–2601; (c) A. S. Mahadevi and G. N. Sastry, Chem. Rev., 2016, 116, 2775–2825; (d) W. Wang, B. Ji and Y. Zhang, J. Phys. Chem. A, 2009, 113, 8132–8135; (e) S. Scheiner, Acc. Chem. Res., 2013, 46, 280–288; (f) A. Bauzá, T. J. Mooibroek and A. Frontera, Angew. Chem., Int. Ed., 2013, 52, 12317–12321; (g) G. Sánchez-Sanz, C. Trujillo, I. Alkorta and J. Elguero, ChemPhysChem, 2019, 20, 1572–1580.
3. (a) M. Bujak, H.-G. Stammler, S. Blomeyer and N. W. Mitzel, Chem. Commun., 2019, 55, 175–178; (b) M. Bujak, H.-G. Stammler and N. W. Mitzel, Cryst. Growth Des., 2020, 20, 3217–3223.
4. H. J. Frohn, S. Görg, G. Henkel and M. Läge, Z. Anorg. Allg. Chem., 1995, 621, 1251–1256.
5. P. Pyykkö and M. Atsumi, Chem. – Eur. J., 2009, 15, 186–197.
6. R. Blom, A. Haaland and R. Seip, Acta Chem. Scand., 1983, 37, 595–599.
7. (a) R. J. Gillespie and R. S. Nyholm, Q. Rev., Chem. Soc., 1957, 11, 339–380; (b) R. J. Gillespie, Coord. Chem. Rev., 2008, 252, 1315–1327.
8. K. Lisac, F. Topić, M. Arhangelskis, S. Cepić, P. A. Julien, C. W. Nickels, A. J. Morris, T. Friščić and D. Cinčić, Nat. Commun., 2019, 10, 61.
9. C. F. Macrae, I. Sovago, S. J. Cottrell, P. T. A. Galek, P. McCabe, E. Pidcock, M. Platings, G. P. Shields, J. S. Stevens, M. Towler and P. A. Wood, J. Appl. Crystallogr., 2020, 53, 226–235.
10. (a) A. Bondi, J. Phys. Chem., 1964, 68, 441–451; (b) S. S. Batsanov, Inorg. Mater., 2001, 37, 871–885; (c) S.-Z. Hu, Z.-H. Zhou, Z.-X. Xie and B. E. Robertson, Z. Kristallogr., 2014, 229, 517–523.
11. (a) G. R. Desiraju and R. Parthasarathy, J. Am. Chem. Soc., 1989, 111, 8725–8726; (b) A. Mukherjee, S. Tothadi and G. R. Desiraju, Acc. Chem. Res., 2014, 47, 2514–2524.
12. (a) M. J. Turner, J. J. McKinnon, S. K. Wolff, D. J. Grimwood, P. R. Spackman, D. Jayatalika and M. A. Spackman, Crystal Explorer17, University of Western Australia, Perth (Australia), 2017, http://hirshfeldsurface.net; (b) P. R. Spackman, M. J. Turner, J. J. McKinnon, S. K. Wolff, D. J. Grimwood, D. Jayatilaka and M. A. Spackman, J. Appl. Crystallogr., 2021, 54, 1006–1011; (c) J. J. McKinnon, M. A. Spackman and A. S. Mitchell, Acta Crystallogr., Sect. B: Struct. Sci., 2004, 60, 627–668; (d) J. J. McKinnon, D. Jayatilaka and M. A. Spackman, Chem. Commun., 2007, 3814–3816; (e) M. A. Spackman and D. Jayatilaka, CrystEngComm, 2009, 11, 19–32.
13. (a) I. Dance, New J. Chem., 2003, 27, 22–27; (b) M. Bujak, M. Podsiadło and A. Katrusiak, CrystEngComm, 2011, 13, 396–398.
14. W. B. Schneider, G. Bistoni, M. Sparta, M. Saitow, C. Riplinger, A. A. Auer and F. Neese, J. Chem. Theory Comput., 2016, 12, 4778–4792.
15. M. Hou, Q. Li and S. Scheiner, ChemPhysChem, 2019, 20, 1978–1984.
16. (a) A. Tkatchenko, R. A. DiStasio Jr., M. Head-Gordon and M. Scheffler, J. Chem. Phys., 2009, 131, 094106; (b) S. M. Cybulski and M. L. Lytle, J. Chem. Phys., 2007, 127, 141102; (c) P. Jurečka, J. Šponer, J. Černý and P. Hobza, Phys. Chem. Chem. Phys., 2006, 8, 1985–1993.
17. (a) C. R. Groom, I. J. Bruno, M. P. Lightfoot and S. C. Ward, Acta Crystallogr., Sect. B: Struct. Sci., Cryst. Eng. Mater., 2016, 72, 171–179; (b) I. J. Bruno, J. C. Cole, P. R. Edgington, M. Kessler, C. F. Macrae, P. McCabe, J. Pearson and R. Taylor, Acta Crystallogr., Sect. B: Struct. Sci., 2002, 58, 389–397.
18. (a) G. Beurskens, P. T. Beurskens, J. H. Noordik, J. Willemse and J. A. Cras, Recl.: J. R. Neth. Chem. Soc., 1979, 98, 416–419; (b) O. Blacque, H. Brunner, M. M. Kubicki, F. Leis, D. Lucas, Y. Mugnier, B. Nuber and J. Wachter, Chem. – Eur. J., 2001, 7, 1342–1349; (c) J. W. Dube and P. J. Ragogna, Chem. – Eur. J., 2013, 19, 11768–11775.
19. J. Janczak and G. J. Perpétuo, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 2006, 62, m45–m47.
20. N. Louvain, N. Mercier and F. Boucher, Inorg. Chem., 2009, 48, 879–888.
21. J. W. Wielandt, S. Petrie, N. L. Kilah, A. C. Willis, R. D. Dewhurst, F. Belaj, A. Orthaber, R. Stranger and S. B. Wild, Aust. J. Chem., 2016, 69, 524–532.
22. O. S. Urgut, I. I. Ozturk, C. N. Banti, N. Kourkoumelis, M. Manoli, A. J. Tasiopoulos and S. K. Hadjikakou, Mater. Sci. Eng., C, 2016, 58, 396–408.
23. H. J. Breunig, M. Denker and K. H. Ebert, J. Chem. Soc., Chem. Commun., 1994, 875–876.
24. H. J. Breunig, K. H. Ebert, S. Gülec, M. Dräger, D. B. Sowerby, M. J. Begley and U. Behrens, J. Organomet. Chem., 1992, 427, 39–48.
25. (a) M. A. Viswamitra and K. K. Kannan, Nature, 1966, 209, 1016–1017; (b) D. Brodalla, D. Mootz, R. Boese and W. Osswald, J. Appl. Crystallogr., 1985, 18, 316–319; (c) A. D. Bond, Chem. Commun., 2003, 250–251; (d) A. E. Goeta and J. A. K. Howard, Chem. Soc. Rev., 2004, 33, 490–500; (e) M. T. Kirchner, D. Bläser and R. Boese, Chem. – Eur. J., 2010, 16, 2131–2146; (f) M. Bujak and A. Katrusiak, CrystEngComm, 2010, 12, 1263–1268; (g) R. Boese, Z. Kristallogr., 2014, 229, 595–601; (h) D. Li, J. Schwabedissen, H.-G. Stammler, N. W. Mitzel, H. Willner and X. Zeng, Phys. Chem. Chem. Phys., 2016, 18, 26245–26253; (i) D. Dey, S. K. Nayak and D. Chopra, Intermolecular Interactions in In situ Cryocrystallized Compounds, in Understanding Intermolecular Interactions in the Solid State – Approaches and Techniques, Monographs in Supramolecular Chemistry No. 26, ed. D. Chopra, The Royal Society of Chemistry, London, 2019.
26. CRC Handbook of Chemistry and Physics, ed. D. R. Lide, Taylor and Francis, Boca Raton, FL, 87^th Edition, 2007.
27. CrysAlisPro, Rigaku Oxford Diffraction, version 1.171.40.56a, Rigaku Corporation, The Woodlands TX (USA), 2019.
28. (a) G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2008, 64, 112–122; (b) G. M. Sheldrick, Acta Crystallogr., Sect. C: Struct. Chem., 2015, 71, 3–8.

---

Footnote

† Electronic supplementary information (ESI) available: Crystallographic details and quantum-chemical calculations. CCDC 2057063–2057076. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/d1ce01268a

---