4-Aminobenzoic acid 4-methylpyridine/4-methylpyridinium 4-aminobenzoate 0.58/0.42: a redetermination from the original data

The title structure, 4-aminobenzoic acid 4-methylpyridine/4-methylpyridinium 4-aminobenzoate 0.58/0.42, has been redetermined from the data published by Kumar et al. (2015). Acta Cryst. E71, o125-o126. The motivation for this redetermination follows from negligence of important features of the difference electron-density maps as well as from wrongly applied constraints, which significantly affect the structural model. The corrections affect mainly the positions of the H atoms involved in the hydrogen bonds (centered on the primary amine group for which the H atom turned out to be disordered over two positions about the centre of an N⋯H⋯O hydrogen bond) and the methyl group, which is disordered and has now been remodelled.


Chemical context
Crystal structures that contain hydroxyl, secondary and primary amine groups are sometimes determined incorrectly because of an assumed geometry of these groups from which the applied constraints or restraints were inferred. In such cases, the correct geometry is missed as it is not verified by inspection of the difference electron-density maps. Thus a considerable number of structures could have been determined more correctly -cf. Figs. 1 and 2 in Fá bry et al. (2014). The inclusion of such structures causes bias in crystallographic databases such as the Cambridge Crystallographic Database (CSD; Groom et al., 2016).
In the course of recalculation of suspect structures that were retrieved from the CSD, the structure determination of the title structure by Kumar et al. (2015), CSD refcode WOYPEH, became a candidate for a checking recalculation. The reason was that the primary amine group centered on N1 was constrained to be coplanar to the attached phenyl group with distances N1-H1a and N1-H1b constrained to be equal to ISSN 2056-9890 0.86 Å with U iso (H primary/secondary amine ) = 1.2U eq (N primary/ secondary amine ).
The hydroxyl hydrogen atom H1 was also suspect because the O-H bond length was reported to be restrained to the value 0.82 Å [the estimated standard deviation/elasticity (Mü ller et al., 2006) was not given in the original article]. However, the distance reported by Kumar et al. (2015) is 0.836 (10) Å , which indicated that the bridging hydrogen atom might have been situated towards the centre of the pertinent O1Á Á ÁN2 hydrogen bond. Recalculation with JANA2006 (Petříček et al., 2014) revealed hydrogen atom H1 to be disordered over two positions about the centre of the O1Á Á ÁN1 hydrogen bond in almost equal proportions, 0.58(7) (H1x) and 0.42(7) (H1y) (Figs. 1 and 2). This is different from the situation reported in the original article (Kumar et al., 2015). Moreover, inspection of the difference electron-density maps View of the constituent molecules of the title structure (top: the original determination (Kumar et al., 2015); bottom: present redetermination). Displacement ellipsoids are depicted at the 50% probability level.

Figure 2
A section of the difference electron-density map for the redetermined title structure without the atoms H1x and H1y. A build-up of the electron density between the atom O1 (red) and N2 iii (blue) [symmetry codes: (iii) x + 1, y, z + 1] is shown; the larger and the smaller peaks correspond to the electron density of 0.12 and 0.11 e A À3 , respectively. These peaks were assigned to the respective positions of H1x and H1y iii . The positive and negative electron densities are indicated by continuous and dashed lines, respectively. The increment of the electron density between neighbouring contours is 0.01 e Å À3 . Atom C7 is indicated by a gray circle. Table 1 Hydrogen bonds (Å , ) in the redetermined structure as well as in the determintion by Kumar et al. (2015). Some of the atoms in the original article were transformed. Symmetry codes: (i) x À 1, y, z À 1; (ii) x À 1, Ày, z À 1 2 ; (iii) x + 1, y, z + 1.

Figure 3
A section of the difference electron-density map for the redetermined title structure without the methyl H atoms. The positions of both methylhydrogen triplets are indicated by yellow circles of a different hue. The positive and negative electron densities are indicated by continuous and dashed lines, respectively. The increment of electron density between the neighbouring contours is 0.01 e Å À3 .

Structural commentary
On the other hand, the tendency for a hydrogen atom to be situated just between the donor and acceptor atoms has been observed for strong hydrogen bonds, especially of the type OÁ Á ÁHÁ Á ÁO ). Such bonds tend to occur in the structures where the difference ÁpK a = pK a (base) À pK a (acid) is close to 0 . The difference ÁpK a is correlated with the occurrence of structures where the base and acid components are not ionized, thus forming a co-crystal (ÁpK a < 0), or ionized, forming a salt (ÁpK a > 3; Childs et al., 2007). It is difficult to predict the form in which the acid and the base are present for 0 < ÁpK a < 3.
In the case of the title structure, pK a of 4-methylpyridine and of 4-aminobenzoic acid are equal to 5.99 (CRC Handbook of Chemistry and Physics, 2009) and 2.38 (Kortü m et al., 1961), respectively. Thus ÁpK a = 3.61 for the title structure, which means that the salt form should be slightly more probable for the present structure.
The primary amine group centered on N1 was originally constrained to be coplanar with the attached phenyl ring while the N1-H1a and N1-H1b distances were both constrained to 0.86 Å .
The difference electron-density map in the plane of the methyl hydrogen atoms that were excluded from the structure for the sake of this checking calculation (Fig. 4) shows that the methyl group can be better modelled by a disorder over two positions with equal occupancies. The disordered positions of the methyl group are related by a rotation of 60.19 (5) about the C10-C13 bond. Table 1, which also compares the values of the hydrogenbond pattern in the title and the original structures (Kumar et al., 2015), emphasizes the importance of a careful examination of the difference electron-density maps during structure 1510 Jan Fábry 0.58(C 6 H 7 NÁC 7 H 7 NO 2 )Á0.42(C 6 H 8 N + ÁC 7 H 6 NO 2 À ) Acta Cryst. A section of the title structure. Symmetry codes (i): Àx + 1, y, z À 1; (ii): Àx + 1, Ày, z À 1 2 ; (iii): x + 1, y, z + 1; (iv): Àx + 2, Ày, z À 3 2 (v): x À 2, y, z À 1. Applied colours for the atoms: grey -C and H, blue -N, O -red; applied colours for the bonds: black -covalent bonds, dashed orange -hydrogen bonds.
determinations. It serves as an example of the bias that is caused by unsubstantiated constraints of the primary amine groups as well as by constraints or restraints imposed on the hydroxyl groups.

Database survey
The structure determination by Kumar et al. (2015) is included in the Cambridge Structural Database (Groom et al., 2016) under refcode WOYPEH.

Synthesis and crystallization
The preparation of the title crystals was described by Kumar et al. (2015). Table 2 lists the details regarding the crystal data, data collection and the refinement [some pieces of information were taken from the downloaded CIF of the original article by Kumar et al. (2015)]. The refinement was carried out on the data for which the 826 Friedel pairs were not merged. Since the structure is composed of light atoms only and the applied radiation was Mo K the absolute structure could not be determined. All hydrogen atoms were discernible in the difference electron-density map. The aryl hydrogens were constrained by the constraints C aryl -H aryl = 0.93 Å and U iso (H aryl ) = 1.2U eq (C aryl ). The positional parameters of the primary amine hydrogen atoms H1a and H1b were refined freely while their displacement parameters were constrained by U iso (H N1 ) = 1.2U eq (N1).

Refinement
The positional parameters of the bridging hydrogen atoms, H1x and H1y, were determined from difference electrondensity maps (Fig. 2) and fixed in the subsequent refinement. Their isotropic displacement parameters were set equal and their occupational parameters were refined under the condition that the sum of their occupancies was equal to 1.
The electron density in the plane of the methyl hydrogen atoms, which was centered on atom C13, was found to be quite smeared (Fig. 3). It was modelled by a disorder over two positions with equal occupancies. The rotation between both triplets of the methyl hydrogen atoms is 60.19 (5) . In order to account for this model, dummy atoms C10a and C13a, both with occupancies equal to 0, were introduced into the structure; their atomic parameters were otherwise constrained to be equal to those of atoms C10 and C13, respectively. The methyl hydrogen atoms were constrained by distance constraints C methyl -H methyl = 0.96 Å with U iso (H methyl ) = 1.5U eq (C methyl ).
It is worthwhile mentioning that the recalculation of the original model with JANA2006 (Petříček et al., 2014) in order to reproduce the original constraints and restraints converged with difficulty {Á[last step of the parameter(i)]/(i) < 0.6}. The indicators of the refinement of such a model were substantially higher: R obs = 0.0503, Rw obs = 0.1035, R all = 0.0930, Rw all = 0.1119. The condition for the observed diffractions was I/(I) > 3, cf. Table 2 for indicators of the refinement for the redetermined structure.