Powder study of N-[ 2-( 4-hydroxy-2-oxo-2 , 3-dihydro-1 , 3-benzothiazol-7-yl ) ethyl ]-3-[ 2-( 2-naphthalen-1-ylethoxy ) ethyl-sulfonyl ] propylaminium benzoate

# 2004 International Union of Crystallography Printed in Great Britain ± all rights reserved The crystal structure of the title compound, C26H31N2O5S2 + C7O2H5 ÿ, also known as AR-C69457CC, was solved by simulated annealing from laboratory X-ray powder diffraction data collected at room temperature to 2.1 AÊ resolution. Subsequent Rietveld re®nement yielded an Rwp of 0.038 and site-occupancy factors for the disordered anion components of 0.5.

# 2004 International Union of Crystallography Printed in Great Britain ± all rights reserved The crystal structure of the title compound, C 26 H 31 N 2 O 5 S 2 + Á-C 7 O 2 H 5 À , also known as AR-C69457CC, was solved by simulated annealing from laboratory X-ray powder diffraction data collected at room temperature to 2.1 A Ê resolution. Subsequent Rietveld re®nement yielded an R wp of 0.038 and site-occupancy factors for the disordered anion components of 0.5.

Comment
The title compound, (I), was synthesized by AstraZeneca during the development of a potential treatment for chronic obstructive pulmonary disease. The crystal structure of (I) was solved as part of a wider investigation into the application of simulated annealing to the problem of solving pharmaceutical crystal structures from laboratory X-ray powder diffraction data (Docherty, 2004). The hydrogen bonding and ring interactions in (I) are summarized in Fig. 3. Hydrogen bond`a' [O1Á Á ÁN2 = 2.82 (6) A Ê ] links two cations to form a centrosymmetric dimer, within which the heterocyclic rings make face-to-face contact (R1Á Á ÁR1 H in Fig. 3) and the carbonyl O atom makes a close approach to the centroid of benzene ring R2 H [O1Á Á Ácentroid = 3.54 (3) A Ê and C1ÐO1Á Á Ácentroid = 95 (3) ]. The heterocyclic ring also engages in face-to-face contact with the C2±C7 benzene ring (Fig. 3, top right, R1Á Á ÁR2a and R2Á Á ÁR1a). The donor±acceptor distances for the three cation±anion hydrogen bonds`b' to`d' fall in the range 2.38 (12)±2.51 (13) A Ê and the hydrogen-bonding scheme is preserved on switching between the two halfoccupancy anion sites. The naphthalene rings engage with each other in offset face-to-face interactions (Fig. 3, bottom right) and pack, along with the benzoate phenyl ring, to form a hydrophobic layer in the ab plane.

Experimental
A polycrystalline sample of the title compound was recrystallized from acetonitrile solution by slow evaporation at room temperature. Data were collected from a sample in a rotating 0.7 mm borosilicate glass capillary using a variable count time scheme (Hill & Madsen, 2002). The diffraction pattern indexed to a triclinic cell [F(20) = 124.5, M(20) = 33.5; DICVOL91 (Boultif & Louer, 1991)] and space group P1 was assigned from volume considerations and a lack of systematic absences. The data set was background subtracted and truncated to 42 2 for Pawley ®tting (Pawley, 1981; 1 2 Pawley = 2.7) and the structure solved using the simulated annealing (SA) global optimization procedure, described previously (David et al., 1998), that is now implemented in the DASH computer program (David et al., 2001). The SA structure solution involved the optimization of two fragments (the cation with 13 torsion angles plus the anion) totaling 26 degrees of freedom. The best SA solution had a favourable 1 2 SA / 1 2 Pawley ratio of 5.7 and a chemically sensible packing arrangement, but suffered from a signi®cant mis®t to the data, even at modest 2 angles. Rerunning the SA with the cation ®xed in its previously determined position and optimizing the positions and orientations of two 50% occupancy anions halved the 1 2 SA /1 2 Pawley ratio to 2.9 and signi®cantly improved the ®t at lower 2 angles. The solved structure was then re®ned against the full data set (5±69 2) using a restrained Rietveld method (Rietveld, 1969) as implemented in TOPAS (Coelho, 2003), with the R wp falling from 0.064 to 0.038 during the re®nement. All cation atomic positions (including H atoms) were re®ned, subject to a series of restraints on bond lengths, angles and, where appropriate, planarity. The distance and angle restraints were based on a geometric analysis of ®ve cations in four crystal structures (Docherty, 2004) closely related to the title compound, namely (a) 2-(4-hydroxy-2-oxo-2,3-dihydro-1,3-benzothiazol-7-yl)ethylammonium chloride, (b) the monohydrate of (a), (c) N-[2-(4hydroxy-2-oxo-2,3-dihydro-1,3-benzothiazol-7-yl)ethyl]-3-[2-(2-(4methylphenyl)ethoxy)ethylsulfamoyl]propylaminium besilate besilate and (d) the tosilate analogue of (c). This was supplemented by a geometric analysis of naphthalene rings using the knowledge base, MOGUL (Bruno et al., 2004). The half-occupancy anions could not be re®ned reliably using the strategy just described and were therefore re®ned as rigid bodies. A March±Dollase correction of intensities for preferred orientation (Dollase, 1986)     The atomic arrangement in (I), showing the anion disordered over two half-occupancy sites. Isotropic displacement spheres are shown at the 50% probability level.

Figure 3
The hydrogen-bonding and ring interactions in (I), calculated and illustrated using PLATON (Spek, 2003;program version 280604).
the re®ned value of the preferred orientation coef®cient along the [001] direction was 1.13 (1).