Dimethyl 4,4′-dihydroxy-3,3′-{[(3aRS,7aRS)-2,3,3a,4,5,6,7,7a-octahydro-1H-1,3-benzimidazole-1,3-diyl]bis(methylene)}dibenzoate

The title compound, C25H30N2O6, has the imidazolidine ring in an envelope conformation. There are two intramolecular O—H⋯N hydrogen-bond interactions with graph-set motif S(6). The cyclohexane ring adopts a slightly distorted chair conformation. One methyl carboxylate substituent forms a dihedral angle of 12.00 (5)° with the plane of the benzene ring, while the other methyl carboxylate group is almost coplanar, making a dihedral angle of 2.26 (9)°. In the crystal, pairs of intermolecular C—H⋯O hydrogen bonds form racemic dimers, corresponding to an R 2 2(18) graph-set motif. Further weak C—H⋯O interactions generate a chain running along the c axis.

The title compound, C 25 H 30 N 2 O 6 , has the imidazolidine ring in an envelope conformation. There are two intramolecular O-HÁ Á ÁN hydrogen-bond interactions with graph-set motif S(6). The cyclohexane ring adopts a slightly distorted chair conformation. One methyl carboxylate substituent forms a dihedral angle of 12.00 (5) with the plane of the benzene ring, while the other methyl carboxylate group is almost coplanar, making a dihedral angle of 2.26 (9) . In the crystal, pairs of intermolecular C-HÁ Á ÁO hydrogen bonds form racemic dimers, corresponding to an R 2 2 (18) graph-set motif. Further weak C-HÁ Á ÁO interactions generate a chain running along the c axis.

Comment
The synthesis of the title compound (I) represents an expansion of our previous work exploring the substituent effects on the solid state structures of di-Mannich bases (Rivera et al., 2011a,b,c). In the title compound (I), C 25 H 30 N 2 O 6 , the heterocyclc ring has a envelope conformation on C7 (Q(2) = 0.4439 (15) Å, φ = 296.59 (19)°). It is surprising, however, that of the 12 fused 1,3-disubstituted-(3aR,7aR/3aS,7aS)-2,3,3a,4,5,6,7,7a-octahydro-1H-1,3-benzimidazole of this type studied to date by us, including the title compound, only two has an envelope conformation on aminalic carbon (NCH 2 N); the other all have a twisted conformation on C-C bond joint to both rings. The molecular conformation is stabilized by two intramolecular O-H···N hydrogen-bond interaction with set graph motif S(6) (Bernstein,et al. 1995). The cyclohexane ring adopt a slightly distorted chair conformation (Cremer & Pople, 1975) with puckering parameters Q, θ and φ of 0.5834 (16) Å, 5.35 (16)°, 300.8 (17)°. In the molecule of the title compound ( Fig. 1), bond lengths and angles are generally within normal ranges and comparable with those observed in related compounds (Rivera et al., 2011a,b,c). Whereas the first carbonyl group is coplanar with the phenyl ring [torsion angle C10-C11-C15-O1 of 1.3 (2)°], the second carbonyl group is slightly twisted out of the plane of the aromatic ring. The torsion angle C19-C20-C24-O4 amounts to -10.1 (2)°. Therefore, the differences in orientation likely arise as a result of steric considerations with respect to the crystal packing.
The crystal packing is dominated by non-conventional C-H···O hydrogen bond interactions, Table 1
The solvent was evaporated under reduced pressure until a sticky residue appeared. The product was purified by chromatography on a silica column, and subjected to gradient elution with benzene:ethyl acetate (yield 20%, m.p. = 449-450 K).
Single crystals of racemic (I) were grown from a chloroform: methanol solution by slow evaporation of the solvent at room temperature over a period of about 2 weeks.

Refinement
All hydrogen atoms were discernible in difference Fourier maps and could be refined to reasonable geometry. According to common practice H atoms bonded to C atoms were kept in ideal positions with C-H distance 0.96 Å during the refinement.
The methyl H atoms were allowed to rotate freely about the adjacent C-C bonds. The hydroxyl H atoms were found in difference Fourier maps and their coordinates were refined freely. All H atoms were refined with displacement coefficients U iso (H) set to 1.5Ueq(C, O) for methyl and hydroxyl groups and to to 1.2Ueq(C) for the CH-and CH 2 -groups. Fig. 1. A view of (I) with the numbering scheme. Displacement ellipsoids are drawn at the 50% probability level.
Refinement. The refinement was carried out against all reflections. The conventional R-factor is always based on F. The goodness of fit as well as the weighted R-factor are based on F and F 2 for refinement carried out on F and F 2 , respectively. The threshold expression is used only for calculating R-factors etc. and it is not relevant to the choice of reflections for refinement.
The program used for refinement, Jana2006, uses the weighting scheme based on the experimental expectations, see _refine_ls_weighting_details, that does not force S to be one. Therefore the values of S are usually larger than the ones from the SHELX program.