2,2′-(Disulfanediyl)dibenzoic acid N,N-dimethylformamide monosolvate: crystal structure, Hirshfeld surface analysis and computational study

In the title 1:1 solvate, 2,2′-dithiodibenzoic acid (DTBA):dimethylformamide (DMF), the DTBA molecule is twisted [C—S—S—C = −88.57 (6)°]. Four-molecule aggregates are formed in the crystal via DTBA-O—H⋯O(DMF) and DTBA-O—H⋯O(DTBA) hydrogen bonding. These are connected in three-dimensions by benzene-C—H⋯O(DTBA), DTBA-C=O⋯π(benzene) and benzene-C—H⋯π(benzene) interactions.


Structural commentary
The asymmetric unit of (I) comprises a molecule of dithiodibenzoic acid (DTBA) and dimethylformaide (DMF), each in a general position, Fig. 1. The crystals were obtained from the recrystallization of 2-mercaptobenzoic acid from a benzene/ DMF (5 ml/1 ml v/v) solution indicating the acid oxidized to DTBA during crystallization. The observed disparity in the C-O bond lengths in the carboxylic acid residues [C1-O1,O2 = 1.3177 (15) & 1.2216 (15) Å and C14-O3,O4 = 1.3184 (14) & 1.2295 (14) Å ] confirms the location of the acidic H atoms on the O1 and O3 atoms, respectively. A characteristic twisted conformation is evidenced in the C3-S1-S2-C8 torsion angle of À88.57 (6) . The dihedral angle between the benzene rings is 87.71 (3) , consistent with an orthogonal disposition. The C1-carboxylic acid group is almost co-planar with the (C2-C7) benzene ring to which it is connected with the dihedral angle between the least-squares planes being 1.03 (19) . By contrast, a small twist is noted for the C14-carboxylic acid residue where the comparable dihedral angle is 7.4 (2) . Intramolecular hypervalent S O interactions (Nakanishi et al., 2007) are indicated as the carbonyl-O2 and O4 atoms are orientated towards the disulfide-S1 and S2 atoms, respectively, with the S1Á Á ÁO2 and S2Á Á ÁO4 separations being 2.6140 (9) and 2.6827 (9) Å , respectively.

Supramolecular features
The key feature of the supramolecular aggregation in the crystal of (I) is the formation of hydrogen bonds between the DTBA-hydroxyl-O1 and the DMF-O5 atoms, as indicated in Fig. 1 and detailed in Table 1, along with hydrogen bonds between centrosymmetrically related C14-carboxylic acid groups associating via an eight-membered {Á Á ÁOHCO} 2 homosynthon. The result is the four-molecule aggregate shown in Fig. 2(a). For the DTBAÁ Á ÁDMF interaction, further stabilization is realized through a DMF-C15-HÁ Á Á O2(carbonyl) contact, Table 1, to close a seven-membered {Á Á ÁHOCOÁ Á ÁHCO} heterosynthon. This cooperativity accounts for the near co-planar relationship between the C1carboxylic acid group and the non-H atoms of the DMF molecule (r.m.s. deviation = 0.0125 Å ) as seen in the dihedral angle of 10.21 (19) between the two residues. The four-molecule aggregates are linked into supramolecular chains via benzene-C7-HÁ Á ÁO(hydroxyl) interactions occurring between centrosymmetrically related molecules. The chains are connected by parallel C OÁ Á Á(benzene) interactions as detailed in Fig. 2(b) and Table 1. The resulting supramolecular layer is parallel to (011), Fig. 2(c), with connections between them leading to a three-dimensional architecture being benzene-C11-HÁ Á Á(benzene), Fig. 2(d).
Crystal (I) was also subjected to the calculation of solventaccessible void space through Mercury (Macrae et al., 2020) with a probing radius of 1.2 Å within an approximate grid spacing of 0.3 Å . It was found that the DMF solvent molecules occupy about 25.4% or equivalent to 220.8 Å 3 of the unit-cell volume, whereas the remaining 74.6% or equivalent to The molecular structures of the constituents of (I) showing the atomlabelling scheme and displacement ellipsoids at the 70% probability level. The dashed line indicates a hydrogen bond. Table 1 Hydrogen-bond geometry (Å , ).

Hirshfeld surface analysis
To better comprehend the supramolecular features of (I), it was subjected to Hirshfeld surface analysis through Crystal Explorer 17 (Turner et al., 2017) using the established methods (Tan et al., 2019). Several close contacts with distances shorter than the sum of van der Waals radii (Spackman & Jayatilaka, 2009) are manifested by red spots of varying intensities on the Hirshfeld surface calculated over d norm in Fig. 4 A perspective view of the solvent-accessible voids in the crystal of (I), calculated after removal of the DMF solvent molecules within 2 Â 2 Â 1 unit-cells.

Table 2
A summary d norm contact distances (adjusted to neutron values) for interactions present in the crystal of (I) as computed through a Hirshfeld surface analysis.

Contact
Distance AEvdW a Á|(d norm À AEvdW)| Symmetry operation 3.40 0.03 Àx, 2 À y, Àz H1OÁ Á ÁO5(carbonyl) and hydroxy-O3-H3OÁ Á ÁO4(carbonyl) hydrogen bonds with the corresponding d norm contact distances being 1.62 and 1.64 Å , respectively, i.e. significantly shorter by almost 1 Å compared to the sum of the van der Waals radii of 2.61 Å (adjusted to neutron values), Table 2. Red spots of moderate intensity are observed for DMF-C15-H15Á Á ÁO2(carbonyl) contact with a distance of 2.29 Å , while spots with weak to diminutive intensities are observed for other close contacts which mainly involve the aromatic rings and carboxylic groups of DTBA as well as the carbonyl group of DMF.
Of particular interest among all close contacts present in (I) is a O3Á Á ÁC14 interaction, which is included within an apparentinteraction formed between the C8-C13 benzene ring and a quasi--system defined by O3-H3OÁ Á ÁO4 hydrogen bonds between a DTBA dimer, i.e. the eightmembered {Á Á ÁO4-C14-O3-H3O} 2 ring system. A similar observation is also noted for the C1Á Á ÁC15 contact which is encapsulated within an apparent (C2-C7)Á Á Áquasi-(O2-C1-O1-H1OÁ Á ÁO5-C15-H15) interaction. The separation between the ring centroids of the aforementionedcontacts are 3.65 and 3.49 Å , respectively. The stacking arrangement between the relevant aromatic and quasiaromatic rings is supported by shape complementarity as revealed by the concave (red) and convex (blue) regions in the shape index, Fig The electrostatic potential property was mapped onto the Hirshfeld surface using the DFT-B3LYP/6-31G(d,p) approach to verify the nature of the contacts present in (I). The electrostatic charges for the points of contacts between each Hatom donor and acceptor are collated in Table 3. The results show that those interactions involving H-donors and Oacceptors are electrostatic in nature owing to the relatively great charge disparity between interacting atoms, with the greatest disparity being observed for the H1OÁ Á ÁO5 followed by H3OÁ Á ÁO4 interactions which is consistent with their corresponding short contact distances. By contrast, for the HÁ Á ÁC and CÁ Á ÁO interactions relatively smaller charge disparity is noted indicating weaker attractions between the participating atoms,. The exception is found for the CÁ Á ÁC contacts which exhibit positive electrostatic charge for both donor and acceptor atoms signifying the dispersive nature of the contacts.
The quantification of the corresponding close contacts on the Hirshfeld surface through fingerprint plot analysis for overall (I) and its individual components, Fig. 6 Table 3 Electrostatic potential charge (V ESP ) for each hydrogen-atom donor and acceptor in (I) participating in a close contact identified through the Hirshfeld surface analysis.  H1OÁ Á ÁO5, O3-H3OÁ Á ÁO4 and C15-H15Á Á ÁO2, and for the HÁ Á ÁC/CÁ Á ÁH contacts, to C5-H5Á Á ÁC11 and C11-H11Á Á ÁC6, while the peaks for HÁ Á ÁS/ SÁ Á ÁH exhibit a d i + d e contact distance of $2.92 Å , which is slightly shorter than the sum of the van der Waals radii ( P vdW radii) of 2.89 Å , Fig. 6(e). Further delineation of HÁ Á ÁO/OÁ Á ÁH, HÁ Á ÁC/CÁ Á ÁH and HÁ Á ÁS/SÁ Á ÁH shows that those heterogeneous contacts are more inclined towards (internal)-XÁ Á ÁH-(external) in DTBA, while the opposite is true for DMF indicating the complementary H-bond accepting and donating nature of DTBA and DMF, respectively. The inclination is more towards (internal)-XÁ Á ÁH-(external) for (I) which reflects the relatively small exposed surface for the DMF molecule and limited hydrogenbond donating role in the overall molecular packing.

Computational chemistry
The program NCIPLOT (Johnson et al., 2010)     The strength of each close contact between all pairwise molecules in (I) was quantified through the calculation of the interaction energies using Crystal Explorer 17 (Turner et al., 2017). As expected, the conventional hydroxy-O3-H3OÁ Á ÁO4(carbonyl) hydrogen bond, leading to the eightmembered homosynthon as well as the seven-membered heterosynthon formed between hydroxy-O1-H1OÁ Á ÁO5(carbonyl) and DMF-C15-H15Á Á ÁO2(carbonyl) exhibit the greatest interaction energies (E int ) of À69.8 and À58.9 kJ mol À1 , respectively. These are relatively stronger than the other supplementary contacts in (I), in which the corresponding energy terms, viz. electrostatic (E ele ), polarization (E pol ), dispersion (E dis ), exchange-repulsion (E rep ) together with the total energy are collated in Table 4.

Figure 8
The energy frameworks for (I) viewed along the a axis, showing the (a) electrostatic force, (b) dispersion force and (c) total energy diagram. The cylindrical radius is proportional to the relative strength of the corresponding energies and they were adjusted to the same scale factor of 100 with a cut-off value of 8 kJ mol À1 within a 2 Â 2 Â 2 unit cells.

Comparison of (I) with the di-DMF solvate
The crystal structure of DTBAÁ2DMF (II) is also known, being reported four times (XEBDEO: Cai et al., 2006;XEBDEO01: Ma et al., 2013;AYIVAH: Baruah, 2016;CUNJUT: Tan & Tiekink, 2020). The key feature of the molecular packing of (II) is that each carboxylic acid residue of the DTBA acid molecule, which lacks crystallographic symmetry, is hydrogen bonded to a DMF molecule to form a three-molecule aggregate. For comparison purposes, (II) (CUNJUT: Tan & Tiekink, 2020), which was evaluated under similar experimental conditions as (I), was also subjected to molecular packing and contact distribution studies. The calculation of the solvent accessible void space using the parameters as mentioned previously shows that the inclusion of additional DMF molecules in the unit-cell is almost directly proportional to the occupied volume by the solvent molecule, i.e. occupied unit-cell volume = 220.8 Å 3 = 25.4% for (I) and 526.4 Å 3 and 47.5% for (II). An analysis of the molecular packing similarity between (I) and (II) demonstrates that although the crystal solvates contain DTBA molecule in common, the inclusion of additional DMF results results in a significant deviation in the molecular packing as evidenced in Fig. 9. Here, only two out of 15 molecules in the cluster of molecules being studied are overlapped (within 20% geometric tolerance), with the r.m.s. deviation of the molecular packing being 0.337 Å .
In term of contact distribution on the Hirshfeld surface for the corresponding individual DTBA molecules and overall (I) and (II), it is noted there are no great disparities in the percentage contributions to the calculated surfaces, Fig. 10.

Database survey
As mentioned in the Chemical Context, DTBA is usually generated during co-crystallization experiments with 2-mercaptobenzoic acid (2-MBA), implying oxidation of the latter. In addition to oxidation of 2-MBA, other crystallization outcomes have been observed during recent experiments suggesting chemical reactions are occurring. A less common outcome of crystallization experiments with 2-MBA was the sulfur extrusion product, 2,2 0 -thiodibenzoic acid (Gorobet et al., 2018), obtained during attempts to react 2-MBA with copper(I) chloride in the presence of two equivalents of triphenylphosphane (Tan & Tiekink, 2018). In a series of experiments with the isomeric Schiff bases, N,N-bis[(pyridinen-yl)methylene]cyclohexane-1,4-diamine, for n = 2, 3 and 4 (Lai et al., 2006), very different products have been characterized from comparable reaction conditions. Referring to Chemical diagrams for (III) and (IV).

Synthesis and crystallization
The

Refinement
Crystal data, data collection and structure refinement details are summarized in  (Rigaku OD, 2018); cell refinement: CrysAlis PRO (Rigaku OD, 2018); data reduction: CrysAlis PRO (Rigaku OD, 2018); program(s) used to solve structure: SHELXS (Sheldrick, 2015a); program(s) used to refine structure: SHELXL2017/1 (Sheldrick, 2015b); molecular graphics: ORTEP-3 for Windows (Farrugia, 2012), DIAMOND (Brandenburg, 2006); software used to prepare material for publication: publCIF (Westrip, 2010). Special details Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.