Monoclinic and orthorhombic forms of ( RS )-( E )-4-[2-(4-chlorobenzylidene)hydrazinyl]-6,11-dimethyl-6,11-dihydro-5 H -benzo[ b ]pyrimido[5,4- f ]azepine: synthesis, concomitant polymorphism and supramolecular assembly mediated by C—H (cid:2) (cid:2) (cid:2) N, C—H (cid:2) (cid:2) (cid:2) p (arene) and C—Cl (cid:2) (cid:2) (cid:2) p (arene) interactions

The title compound, C 21 H 20 ClN 5 , has been synthesized in two steps from ( RS )-4-chloro-6,11-dimethyl-6,11-dihydro-5 H -benzo[ b ]pyrimido[5,4- f ]azepine and characterized by 1 H and 13 C NMR spectroscopy and by high-resolution mass spectrometry. Crystallization from hexane–ethyl acetate yields approximately equal quantities of a monoclinic polymorph in the space group Cc , (I), and an orthorhombic polymorph in the space group Pna 2 1 , (II). The molecules in polymorphs (I) and (II) show small differences in their molecular conformations, particularly in the shape of the azepine ring and the orientation of the chlorophenyl substituent. The molecules in polymorph (I) are linked by C— H (cid:2) (cid:2) (cid:2) N and C—H (cid:2) (cid:2) (cid:2) (cid:2) (arene) hydrogen bonds to form sheets, which are linked into a three-dimensional framework structure by C—Cl (cid:2) (cid:2) (cid:2) (cid:2) (arene) interactions. There are no C—Cl (cid:2) (cid:2) (cid:2) (cid:2) (arene) interactions between the molecules in polymorph (II) and the supramolecular assembly takes the form of sheets built from C—H (cid:2) (cid:2) (cid:2) N and C—H (cid:2) (cid:2) (cid:2) (cid:2) (arene) hydrogen bonds. Comparisons are made with some related structures.


Introduction
Pyrimidines and fused pyrimidines, particularly those compounds containing a pyrimidine unit fused to another heterocyclic ring, constitute a privileged class of structural motifs which are of great interest in the design and discovery of novel compounds exhibiting a wide range of biological and medicinal activities (Costantino & Barlocco, 2006;Gillespie et al., 2008;Kumar et al., 2009;Guetzoyan et al., 2010;Le Corre et al., 2010;Baraldi et al., 2012;DeNinno et al., 2012;Giles et al., 2012;Awadallah et al., 2013;Gangjee et al., 2013). In addition, hydrazones, the condensation products of either hydrazine or an arylhydrazine with carbonyl compounds, also represent valuable intermediates which have been extensively used in medicinal chemistry as building blocks for novel compounds having significant pharmacological properties, including anticancer (Islam et al., 2017), antifungal and antioxidant (Kauthale et al., 2017), antimycobacterial (Angelova et al., 2017), insecticidal (Yang et al., 2016) and leishmanicidal and trypanocidal (Coa et al., 2015) activities, as well as action against the hepatitis C virus (Ş enkardeş et al., 2016).

Refinement
Crystal data, data collection and structure refinement details are summarized in Table 1. All H atoms were located in difference maps. H atoms bonded to C atoms were subsequently treated as riding atoms in geometrically idealized positions, with C-H = 0.95 (aromatic and heteroaromatic), 0.98 (CH 3 ), 0.99 (CH 2 ) or 1.00 Å (aliphatic C-H), and with U iso (H) = kU eq (C), where k = 1.5 for the methyl groups which were allowed to rotate but not to tilt, and 1.2 for all other H atoms. For the H atoms bonded to N atoms, the atomic coordinates were refined with U iso (H) = 1.2U eq (N), giving the N-H distances shown in Table 2. In the refinement of polymorph (II), one bad outlier reflection, i.e. 730, was omitted from the data set prior to the final refinements. For both polymorphs, the correct orientation of the structure with respect to the directions of the polar axes was determined using the Flack x parameter (Flack, 1983), calculated using quotients of the type [(I + ) À (I À )]/[(I + ) + (I À )] (Parsons et al., 2013). For monoclinic polymorph (I), the use of 1264 such quotients gave x = 0.011 (11) and for orthorhombic polymorph (II), the use of 1075 quotients gave x = 0.048 (9). The corresponding values of the Hooft y parameter (Hooft et al., 2008) were y = 0.007 (12) for (I) and y = 0.049 (10) for (II).

Results and discussion
The synthesis of the title compound started with precursor (A). (RS)-4-Chloro-6,11-dimethyl-6,11-dihydro-5H-benzo[b]pyrimido[5,4-f]azepine (see Scheme 1), whose synthesis (Acosta-Quintero et al., 2015) and structure  were reported several years ago. Reaction of this precursor with hydrazine hydrate gives the hydrazino intermediate (B), by means of a nucleophilic substitution reaction, and, in a second reaction step, intermediate (B) undergoes an acid-catalysed condensation reaction with 4-chlorobenzaldehyde to give the title compound. The composition of the title  Table 1 Experimental details. For all structures: C 21 H 20 ClN 5 , M r = 377.87, Z = 4. Experiments were carried out at 100 K with Cu K radiation using a Bruker D8 Venture. Absorption was corrected for by multi-scan methods (SADABS; Bruker, 2015). Refinement was on 250 parameters. H atoms were treated by a mixture of independent and constrained refinement. Computer programs: APEX2 (Bruker, 2014), SAINT (Bruker, 2013), SIR92 (Altomare et al., 1994), SHELXL2014 (Sheldrick, 2015) and PLATON (Spek, 2009). compound was confirmed by high-resolution mass spectrometry and its constitution was fully established from the 1 H and 13 C NMR spectra (x2.1 above). Crystallization of the title compound from hexane-ethyl acetate (50:50 v/v) yielded two types of crystal, in approximately equal abundance, having monoclinic space group Cc for polymorph (I) and orthorhombic space group Pna2 1 for polymorph (II) (Figs. 1 and 2). These two forms are thus concomitant polymorphs (Bernstein et al., 1999;Bernstein, 2011) and their formation in approximately equal quantities requires both that their energies are very similar and that their rates of crystallization are very similar.
Although the two polymorphs crystallize in unit cells having similar volumes (Table 1) and each containing four molecules, there is no obvious relationship between the unit-cell dimensions of the two forms, regardless of whether the monoclinic polymorph is considered in the conventional Cc setting of space group No. 9 or in the alternative Ia setting; in particular, the unit-cell repeat vector b in the monoclinic form has no match in the orthorhombic form. The orthorhombic form has the higher density (Table 1) and for all of the hydrogen-bond types which the two polymorphs exhibit in common, the donor-acceptor (DÁ Á ÁA) distances are always shorter in the orthorhombic form (Table 2), thus accounting, at least in part, for the higher density of this form and suggesting that this polymorph is the more stable (Nelyubina et al., 2010).
There is a stereogenic centre at atom C6 (Figs. 1 and 2) and, for each polymorph, the reference molecule was selected to be that having the R configuration at atom C6; the space groups confirm that the each polymorph has crystallized as a racemic mixture. The conformations of the molecules in the two forms differ slightly (Table 3); firstly, the orientation of the chlorinated ring relative to the adjacent pyrimidine ring differs, as  Table 3 Selected geometric parameters (Å , ) for (I) and (II).
'Dihedral 1' represents the dihedral angle between the two rings fused to the azepine ring and 'dihedral 2' represents the dihedral angle between the pyrimidine ring and the chlorinated C431-C436 ring. Ring-puckering angles are calculated for the atom sequence N11-C10A-C6A-C6-C5-C4A-C11A.

Figure 1
The molecular structure of monoclinic polymorph (I), showing the atomlabelling scheme. Displacement ellipsoids are drawn at the 50% probability level.

Figure 2
The molecular structure of orthorhombic polymorph (II), showing the atom-labelling scheme. Displacement ellipsoids are drawn at the 50% probability level.
indicated in particular by the N42-C43-C431-C432 torsion angles and by the dihedral angles between the planes of the pyrimidine and chlorinated aryl rings. Secondly, the dihedral angles between the two rings fused to the central azepine unit differ by ca 15 ; this reflects the different conformations of the azepine ring in the two polymorphs. In monoclinic polymorph (I), the azepine ring adopts a conformation intermediate between the boat and twist-boat forms, whereas in orthorhombic polymorph (II), the conformation is dominated by the twist-boat form, with minor influences from the chair and twist-chair forms (Boessenkool & Boeyens, 1980;Evans & Boeyens, 1989). The differences in the molecular conformations may perhaps be traceable to the differences in the directionspecific intermolecular interactions involving the aryl rings, as discussed in detail below. Thus, the fused C6A/C7-C10/C10A aryl ring acts as a hydrogen-bond acceptor in polymorph (I) but not in polymorph (II). Similarly, the chlorinated C431-C436 aryl ring is involved in a C-HÁ Á Á(arene) hydrogen bond in polymorph (II), although not in polymorph (I). On the other hand, this ring is involved in a C-ClÁ Á Á(arene) interaction in polymorph (I), although not in polymorph (II). Interactions of this type have been shown (Imai et al., 2008) to be attractive and to be primarily electrostatic in origin.
The supramolecular assembly in monoclinic polymorph (I) is three-dimensional (3D), mediated by C-HÁ Á ÁN and C-HÁ Á Á(arene) hydrogen bonds and C-ClÁ Á Á(arene) interactions (Tables 2 and 4), but the formation of the 3D framework structure is readily analyzed in terms of simple substructures of lower dimensionality (Ferguson et al., 1998a,b;Gregson et al., 2000). Molecules of (I) which are related by the c-glide plane at y = 1 2 are linked by the combined action of the two hydrogen bonds having atom N3 as the acceptor to form a C(4)C(6)[R 1 2 (6)] chain of rings (Etter, 1990;Etter et al., 1990;Bernstein et al., 1995) running parallel to the [001] direction (Fig. 3). The formation of this chain is weakly reinforced by a rather long hydrogen bond having atom N42 as the acceptor. A second substructure can be identified in which molecules, also related to one another by the c-glide plane at y = 1 2 , are linked by a C-HÁ Á Á(arene) hydrogen bond to form a chain running parallel to the  , (x, Ày + 1, z À 1 2 ), (x, y, z + 1) and (x, y, z À 1), respectively. Table 4 Geometric parameters (Å , ) for the C-ClÁ Á Á interaction in monoclinic polymorph (I).

Figure 4
Part of the crystal structure of polymorph (I), showing the formation of a hydrogen-bonded chain running parallel to the [203] direction. Hydrogen bonds are shown as dashed lines and, for the sake of clarity, H atoms not involved in the motif shown have been omitted. The Cl atoms marked with an asterisk (*), a hash (#), a dollar sign ($) or an ampersand (&) are at the symmetry positions (x À 1, Ày + 1, z À 3 (Fig. 4). The combination of these two chain motifs generates a complex sheet lying parallel to (010) and occupying the domain 1 4 < y < 3 4 (Fig. 5). A second sheet of this type, related to the first by the C-centring operation, lies in the domain À 1 4 < y < 1 4 , and adjacent sheets are linked by the C-ClÁ Á Á(arene) interaction (Table 4; cf. Imai et al., 2008). This interaction links molecules related by the n-glide plane at y = 1 4 to form a chain running parallel to the [101] direction (Fig. 6), where alternate molecules lie in different hydrogen-bonded sheets; propagation of this interaction by the space group thus links all the sheets into a single 3D framework structure.
There are no C-ClÁ Á Á(arene) interactions in the structure of orthorhombic polymorph (II), so that the two-dimensional (2D) supramolecular structure is built solely from hydrogen bonds, one each of the N-HÁ Á ÁN, C-HÁ Á ÁN and C-HÁ Á Á(arene) types ( Table 2). As in polymorph (I), the two hydrogen bonds having atom N3 as the acceptor form a C(4)C(6)[R 1 2 (6)] chain of rings, this time comprising molecules related by the a-glide plane at y = 1 4 and running parallel to the [100] direction (Fig. 7). The C-HÁ Á Á(arene) hydrogen bond links molecules which are related by the n-glide plane at x = 1 4 to form a chain running parallel to the [011] direction (Fig. 8). The combination of chain motifs along [100] and [011] generates a rather complex sheet lying parallel to (011), but there are no direction-specific interactions between adjacent sheets. Hence, the supramolecular assembly in the orthorhombic polymorph (II) is 2D, in contrast to the 3D assembly in monoclinic polymorph (I).
It is of interest briefly to compare the supramolecular assemblies in polymorphs (I) and (II) with those in some related compounds. In the structure of the precursor compound (A) (see Scheme 1), there are no hydrogen bonds of any kind, but instead inversion-related pairs molecules are linked into centrosymmetric dimers by astacking interaction involving pyrimidine rings ; this structure also contains a C-ClÁ Á Á(pyrimidine) contact, again between inversion-related pairs of molecules. Compound (III) (see Scheme 2) is a hydrolysis product of the precursor (A) and inversion-related pairs molecules are linked by N-HÁ Á ÁO hydrogen bonds to form centrosymmetric R 2 2 (8) dimers which

Figure 7
Part of the crystal structure of orthorhombic polymorph (II), showing the formation of a hydrogen-bonded chain of rings running parallel to the [100] direction. Hydrogen bonds are shown as dashed lines and, for the sake of clarity, H atoms not involved in the motif shown have been omitted. The Cl atoms marked with an asterisk (*), a hash (#), a dollar sign ($) or an ampersand (&) are at the symmetry positions (x À 1 2 , Ày + 1 2 , z), (x + 1 2 , Ày + 1 2 , z), (x À 1, y, z) and (x + 1, y, z), respectively. are themselves linked by C-HÁ Á ÁN hydrogen bonds to form sheets containing equal numbers of R 2 2 (8) and R 6 6 (24) rings (Acosta . In the structure of the N-methyl-N-phenylamino derivative, (IV), the only directionspecific interaction between the molecules is a C-HÁ Á Á(pyrimidine) hydrogen bond, which links the molecules into simple chains (Acosta Quintero et al., 2018). Finally, we note the tetracyclic analogues (V) and (VI) (Acosta Quintero, Burgos et al., 2016) of the precursor compound (A). In (V), molecules are linked into chains by a C-HÁ Á ÁN hydrogen bond, and inversion-related pairs of chains are linked by astacking interaction; by contrast, the molecules of (VI) are linked into sheets by a combination of a C-HÁ Á Á(pyrimidine) hydrogen bond and two independent C-ClÁ Á Á interactions. For both structures, data collection: APEX2 (Bruker, 2014); cell refinement: SAINT (Bruker, 2013); data reduction: SAINT (Bruker, 2013); program(s) used to solve structure: SIR92 (Altomare et al., 1994); program(s) used to refine structure: SHELXL2014 (Sheldrick, 2015); molecular graphics: PLATON (Spek, 2009); software used to prepare material for publication: SHELXL2014 (Sheldrick, 2015) and PLATON (Spek, 2009). 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.

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å 2 )
x y z U iso */U eq N1 0.5994 ( (9) 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.