J. Phys. Chem. B 2006, 110, 15127-15133
15127
Molecular Structure of a D-homoandrostanyl Steroid Derivative: Single Crystal and Powder Diffraction Analyses P. Martinetto* Laboratoire de Cristallographie, UPR 5031, CNRS, 25 aVenue des Martyrs, BP166, 30842 Grenoble, France
P. Terech, A. Grand, and R. Ramasseul CNRS UMR 5819, CEA-Grenoble, DRFMC-SPrAM, 17, rue des Martyrs, 38054 Grenoble, France
E. Dooryhe´ e and M. Anne Laboratoire de Cristallographie, UPR 5031, CNRS, 25 aVenue des Martyrs, BP166, 30842 Grenoble, France ReceiVed: February 28, 2006; In Final Form: June 8, 2006
The knowledge of the structure of a molecular crystal is frequently a prerequisite for the understanding of its solid state properties. Even though single-crystal diffractometry is the method of choice when it comes to crystal structure determination, methods using powder diffraction data become more and more competitive. There has been much recent interest in the development of a new generation of “direct-space” approaches that are particularly suited for molecular crystals. The crystallographic structure of a steroid derivative molecule (17,17-di-n-propyl-17a-aza-D-homo-5R-androstan-3β-ol) was obtained in two independent ways: from a single crystal by laboratory X-rays and from a polycrystalline powder by high-resolution synchrotron powder diffraction. The molecule crystallizes in the orthorhombic space group P212121 (a ) 6.5346, b ) 17.6006 and c ) 19.6978 Å). Hydrogen bonds form infinite chains of molecules parallel to the c axis.
Introduction Recent methods have been developed to solve increasingly complicated organic molecular structures from powders. Powder diffraction data are reproduced by modeling the structure in direct space.1-3 These methods are characterized by direct handling of molecular fragments within the unit cell. They do not require the extraction of intensity data for individual reflections from the powder diffraction pattern, as needed in traditional approaches developed for single crystals. Position, orientation, and conformation of these fragments are then varied to generate “trial” crystal structures until optimum agreement between calculated and experimental powder diffraction patterns is achieved.4-7 The steroid derivative compound in the present work is a typical example of a complex molecular organic crystal, whose structure can be solved independently by powder and single-crystal methods. We show that the ab initio crystal method and the Monte Carlo rigid-body powder method both converge toward a reliable solution. In addition to this general interest of crystallographic determination using different routes of crystallization (single-crystal versus polycrystalline), the present structural description is also crucial for the investigation of some nanostructured materials. In particular, we report here the molecular structure of a special steroid molecule that exhibits spectacular properties of an efficient organogelator of saturated alkanes. This steroid derivative dissolved in cyclohexane or various saturated hydrocarbons does not precipitate but is known to preferably form gels. Aggregation of the molecules gives very long and chiral filaments, thoroughly studied in the past.8,9 The present structural * Corresponding author. E-mail:
[email protected]. Telephone: +33 4 76 88 74 14; +33 4 76 88 10 38.
Figure 1. Conventional structural formula of STNH.
study concerns the steroid recrystallized in nongelling conditions and should constitute a reference state of the system in the context of self-assembled fibrillar networks as found in molecular gels.10 Experimental Section Preparation. The amino alcohol steroid 17,17-di-n-propyl17a-aza-D-homo-5R-androstan-3β-ol (C25H45NO) with a sixmembered D ring (Figure 1) was prepared according to ref 11. The molecule is named STNH in the following. Single crystals were obtained from a diethyl ether solution and a microcrystalline powder from a heptane/diethyl ether solution, avoiding gelling conditions. Data Collection, Unit Cell Determination, and Space Group Assignment. X-ray Single-Crystal Data. The data collection was performed on a single crystal (ca. 0.2 × 0.2 × 0.25 mm3) using an automatic ENRAF-NONIUS CAD4 diffractometer with CuKR radiation. The cell dimensions a )
10.1021/jp0612681 CCC: $33.50 © 2006 American Chemical Society Published on Web 07/14/2006
15128 J. Phys. Chem. B, Vol. 110, No. 31, 2006
Martinetto et al.
TABLE 1: Crystal Data, Parameters of Data Collection, and Details of Structure Refinements of STNH (single crystal and powder data)a single crystal
powder
formula molar mass (g mol-1) space group
C25H45NO 375.62 P212121
a (Å) b (Å) c (Å) V (Å3) Z wavelength (Å) crystal size (mm) density (calcd) (g cm-3) data collection scan range scan mode scan rate data collected no. unique reflns no. reflns in refinement no. parameters for least squares
6.525(2) 17.557(1) 19.664(1) 2253 4 1.542 (Cu KR) 0.2 × 0.2 × 0.25 1.107 Enraf Nonius CAD4 2° < 2θ < 70° ω scan 0.5°‚min-1 ( h, ( k, ( l 2483 1591 241
structure solution structure refinement RF RwF
MULTAN 16 XFLSN 17 0.06 0.047
a
a (Å) b (Å) c (Å) V (Å3) Z wavelength (Å) glass capillary diameter (mm) density (calcd) (g cm-3) data collection 2θ step (deg) 2θ interval (deg) no. of measured reflections data/parameters structure solution structure refinement χ2 Rwp (background corrected) RB
6.5346(1) 17.6006(1) 19.6978(1) 2265 4 0.856 1 1.101 ESRF-BM16 powder diffractometer 0.004 0° < 2θ < 42° 797 797/10 FOX23 Fullprof.2k 15 10.7 0.18 0.104
The values of all the standard deviations correspond to the numerical refinement uncertainties.
6.525(2), b ) 17.557(1), and c ) 19.664(1) Å were determined by least-squares refinement of the angular position of the 25 reflections (2θ < 40°). The space group P212121 was found by using the systematic extinctions (h00, 0k0, and 00l absent for h, k, or l odd). The intensities of 2483 independent reflections (2° < 2θ < 70°) were collected in the θ/2θ scan mode (scan rate ) 0.5°‚min-1). They were corrected for Lorentz and polarization factors but not for absorption. The stability of the crystal was monitored by measuring the intensities of three particular reflections after every 100 measurements and 3600 s of exposure time. No significant variation in the intensities was observed during data acquisition. Crystal data, together with details of the diffraction experiment and subsequent calculations, are listed in Table 1. Synchrotron Powder Diffraction Data. The powder was packed into a 1 mm glass capillary and mounted in the DebyeScherrer mode on the ESRF-BM16 high-resolution diffractometer.12 The 2θ arm of the diffractometer was equipped with nine parallel analyzer Ge crystals and photon scintillation counters, which were continuously 2θ scanned around the capillary. The normal diffraction pattern is recomposed after binning the individual detector counts in 0.004° 2θ steps, summing over the nine channels and normalizing against the incident beam intensity. The wavelength was calibrated as λ ) 0.856 Å by measuring the 2θ peak positions of SRM640 silicon (NIST standard) reflections.13 All the relevant details concerning the data collection are listed in Table 1. We first used the powder method to solve the structure independently of the results from the single-crystal analysis. The powder X-ray diffraction pattern was indexed by the program DICVOL14 and the cell parameters determined: a ) 6.5346(1), b ) 17.6006(1), and c ) 19.6978(1) Å (the differences observed between the powder and single-crystal cell parameters are explained by the less-accurate determination in the case of single-crystal data). Intensities were then extracted from the powder diffraction pattern using the Le Bail method of full-pattern cell constrained refinements [program FullProf.2k, ref 15]. Systematic absences are consistent with space group
P212121 and density considerations suggest that there are four molecules in the asymmetric unit. Structure Solution and Refinement Single-Crystal Data. The structure was determined by direct methods using the MULTAN program.16 The first normalized intensity E-map, based on 220 reflections for which E > 1.5, revealed the position of all non-hydrogen atoms. The refinement of the atomic positions was performed with a least-squares method using the XFLSN program.17 Successive Fourier difference maps localized the hydrogen atoms. Non-hydrogen atoms were assigned with anisotropic temperature factors, whereas hydrogen atoms were assumed to have isotropic thermal motions B (B being fixed ) 5 Å2). Both non-hydrogen and hydrogen atom coordinates were refined at the final step. The final refinements taking into account 1591 reflections with F0 > 2.5 σ(F0) of 241 parameters reached the values R ) 0.060 and Rw ) 0.047, respectively. Powder Diffraction Data. Because a molecular location method was used to solve the STNH crystal structure, the choice of the starting molecular conformational geometry, as used later in the structure solution and the Rietveld refinement processes, deserves careful attention. The Cambridge Structural Database (CSD) was consulted with the purpose of obtaining the geometries of analogous molecular fragments.18,19 The A, B, and C ring system (Figure 1) was found in epiandrosterone (5R androstan-3β-ol-17-one, refcode ANDRON20 and the fragment corresponding to the D cycle was found in the MPYRAN10 entry (17a-methyl-3β-pyrrolidinyl-17a-aza-D-homo-5R-androstane21). These fragments were assembled and completed with the side groups attached to the D ring using the Materials Studio Vizualizer program [version 3.2.0.0, ref 22] in order to obtain the molecular geometry for STNH. The structure was then solved by the global optimization of a structural model in direct space based on a Monte Carlo search using the simulated annealing algorithm (in parallel tempering mode), as implemented in the FOX program [version 1.6.0.2,
Structure of a D-homoandrostanyl Steroid Derivative ref 23]. The complete STNH molecule (excluding hydrogen atoms) was taken as the basic structural configuration. The system comprising the four rings A, B, C, and D was considered as a rigid unit and the side groups attached to the D ring as flexible units, with four rotatable bonds (C17-C20, C20-C21, C17-C23, and C23-C24). After about 60 million cycles, the agreement factor Rwp was near 0.22, and the corresponding configuration was assumed to constitute a relevant starting structural model in terms of crystal packing. In particular, there were no unrealistic close contacts and the hydrogens bonds were arranged into a realistic network. The best structure solution generated in the simulated annealing calculation was hence taken as the starting model for the rigid-body Rietveld refinement using the program FullProf.2k [version 2.60, ref 15]. In the final Rietveld refinement, there were 18 adjustable parameters: 10 profile parameters (one zero point, three cell parameters, and six parameters for the pseudoVoigt profile function convoluted with the axial divergence asymmetry function) and eight structural parameters (one scale factor, one global isotropic temperature factor, and six parameters for position and orientation of the molecule within the cell). Finally, hydrogen atoms were added to the molecule in positions consistent with standard geometries. The final agreement factors are Rwp ) 0.180, χ2 ) 10.7, RB ) 0.104. The details of the structure refinements are summarized in Table 2, and the Rietveld plot is shown in Figure 2.
J. Phys. Chem. B, Vol. 110, No. 31, 2006 15129 TABLE 2: Non-Hydrogen Atom Fractional Coordinates (×10 000 Å) and Temperature Factors (Å2) for STNH Single Crystala C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14
Results and Discussion
C15
Results. Powder Versus Single-Crystal. The non-hydrogen atom coordinates are listed in Table 2 for single-crystal data and powder diffraction data, respectively. The crystal structure found by the powder diffraction method is identical to that obtained by the single-crystal analysis within a root-mean-square deviation of 0.120 Å for the 27 non-hydrogen atoms. The location of the C17 atom shows the largest deviation (0.271 Å), probably due to the fact that the MPYRAN10 molecule found in the CSD database and used to build the modified D ring of the STNH molecule bears a methyl group on N17a and no propyl group. It is likely that the agreement between the single-crystal and powder structures can be improved, when it is possible, by refining the atom coordinates from the powder data. However, the rigid-body Rietveld refinements of the structure calculated from the powder data already yield a reliable structural model. It helps understand most aspects of the crystal structure such as the details of the molecular packing and the identification of the major intermolecular interactions. Figure 3 shows the unicity of the structures determined by single-crystal and powder methods. Discussion. In the following, we discuss the structural details from the single-crystal results. Structural and conformational data are reported in Table 2 (non-hydrogen atom coordinates), Table 3 (hydrogen atom coordinates), and Table 4 (bond lengths and bond angles). Conformational Analysis. The overall molecular conformation is shown in Figure 4 and may be described in terms of the dihedral angles between the pairs of planes containing the A-B, B-C, and C-D rings, respectively, equal to 171.34°, 2.35° and 9.83°. The twist of the steroid molecule along its main axis may be described relative to the orientation of the angular methyl groups (i.e., angle between C19-C10 and C13-C18) and is equal to -0.52°. The four rings of the steroid skeleton are transconnected (they also are in naturally occurring steroids24) and the three rings A, B, and C exhibit highly symmetrical chair
C16 C17 C18 C19 C20 C21 C22 C23 C24 C25 N O a
x
y
z
B
5132(10) 4964(4) 5293(10) 5155(3) 7045(9) 6997(2) 6788(9) 6969(3) 6634(9) 6598(3) 6626(11) 6649(5) 6745(9) 6888(5) 5076(8) 5330(3) 5110(8) 5133(3) 4866(9) 4842(2) 3530(9) 3409(4) 3788(9) 3853(5) 3680(0) 3835(3) 5413(9) 5350(3) 5720(9) 5746(3) 6223(9) 6251(4) 4576(9) 4716(3) 1535(8) 1572(2) 2733(9) 2752(3) 2609(9) 2637(3) 2797(10) 2666(10) 824(11) 705(11) 5518(9) 5714(3) 4136(9) 3745(9) 5028(11) 4883(4) 4219(6) 4220(3) 7070(6) 7146(2)
-463(3) -471(2) -721(3) -738(1) -317(3) -326(2) 536(3) 492(2) 784(3) 733(1) 1645(3) 1679(1) 1888(3) 1849(1) 1497(3) 1373(1) 628(2) 507(1) 394(3) 359(1) 230(2) 168(1) 470(2) 367(1) 1345(0) 1244(1) 1684(3) 1651(1) 2530(3) 2403(1) 2638(3) 2709(2) 2300(3) 2440(2) 1627(3) 1564(2) 634(3) 651(2) 2800(3) 2930(3) 3570(3) 3683(4) 4029(3) 4045(4) 2242(3) 2214(2) 1894(3) 1759(5) 1975(3) 2040(2) 1493(2) 1612(2) -546(2) -560(2)
4841(2) 4774(1) 4099(2) 4035(1) 3725(2) 3762(1) 3774(2) 3787(1) 4526(2) 4521(1) 4603(2) 4616(1) 5342(2) 5385(1) 5782(2) 5782(1) 5683(2) 5650(1) 4922(2) 4857(1) 6148(2) 6115(1) 6907(2) 6898(1) 6981(0) 6945(1) 6546(2) 6492(1) 6701(2) 6710(1) 7459(2) 7489(2) 7938(2) 8026(1) 6784(2) 6853(1) 4641(2) 4669(1) 7945(2) 7890(2) 8319(3) 8371(4) 8259(3) 8146(4) 8655(2) 8756(1) 9193(2) 9159(2) 9912(2) 9864(2) 7717(1) 7758(1) 3027(1) 3086(1)
3.74 4.92 3.87 4.92 3.78 4.92 3.44 4.92 3.37 4.92 4.05 4.92 3.49 4.92 2.66 4.92 2.38 4.92 2.66 4.92 2.79 4.92 2.9 4.92 2.61 4.92 2.45 4.92 3.25 4.92 3.02 4.92 2.79 4.92 3.13 4.92 3.47 4.92 3.28 4.92 4.37 4.92 5.46 4.92 3.17 4.92 3.63 4.92 5.1 4.92 2.59 4.92 4.37 4.92
The values for STNH powder are in bold.
conformations with small asymmetry parameters ∆Cs and ∆C2 25 (the average ∆C over 18 values is 4.6°). The modified N-bearing D ring is in a distorted chair conformation with ∆Cs ) 20.52° and ∆C2 ) 19.19°. In rings A, B, and C, the average ring C-C bond length is 1.538 Å and the average of the 18 C-C-C bond angles is 110.9° (Table 4). These values are in agreement with those expected: 1.532 Å and 111.0°, respectively (these are the average values over 24 structures extracted from ref 25). This confirms that STNH is a relatively rigid molecule. In the D ring, the average C-N bond length is 1.506 Å, compared to the expected value of 1.477 Å (average value of the six modified N-bearing ring D structures extracted from the
15130 J. Phys. Chem. B, Vol. 110, No. 31, 2006
Martinetto et al.
Figure 2. Rietveld plot of STNH (Rwp ) 0.18, χ2 ) 10.7). Observed (dots) and calculated (solid line) synchrotron powder diffraction patterns (λ ) 0.856 Å) are shown with the normalized difference curve below. The vertical ticks indicate the reflection positions consistent with the P212121 symmetry.
Figure 3. An overlay of the molecules (excluding the hydrogen atoms) within the crystallographic unit cell. Molecular positioning by the directspace approach is in gray thick line. Molecular positioning from single-crystal method is in black thin line. Projections of the molecules and the unit cell onto the (100), (001), and (010) planes, respectively.
CSD database) and the C-N-C bond angle is equal to 118.41° (117.23° from the CSD database). As expected, the hydroxyl group is β-oriented and occupies an equatorial position with respect to the modified steroid nucleus. The two propyl groups have R-equatorial and β-axial orientations, respectively. Intermolecular Interactions. The STNH molecule crystallizes in the P212121 crystallographic space group, as in 55% of steroids and 51% of androstanes.25 The packing arrangement in STNH structure is 212, which means two molecules thick, one wide, and two long in the orthorhombic cell. This is the predominant pattern observed for steroids crystallizing with the orthorhombic symmetry. Infinite chains of STNH molecules connected by hydrogen bonds in head-to-tail configurations
extend parallel to the c axis with adjacent molecules related by the 21 screw operation (Figure 5a). Hydrogen bonds involve the hydroxyl group (ring A) and the amine group (ring D) of adjacent molecules: N-O, 3.000 Å, O-H, 1.023 Å, N-H(OH), 2.027 Å, O-H‚‚‚N, 157.81°. These zigzag chains form stacks along the b axis and chains, that are translationally equivalent in the a direction, interleave with each other (Figure 5b). The collective hydrogen bonding process in the different states of the STNH compound, (solution, gel, xerogel, and solid) has been previously characterized by infrared absorption spectroscopy.26 The IR spectrum of the polycrystalline solid exhibits a broad band at a wave number ν ) 3500 cm-1, confirming the involvement of H bonds. Hydroxylated compounds in the solid state commonly exhibit an IR stretching frequency signature in
Structure of a D-homoandrostanyl Steroid Derivative
J. Phys. Chem. B, Vol. 110, No. 31, 2006 15131
TABLE 3: Hydrogen Atom Fractional Coordinates (×10 000) for STNH from Single Crystal Data H1C1 H2C1 H1C2 H2C2 HC3 H1C4 H2C4 HC5 H1C6 H2C6 H1C7 H2C7 HC8 HC9 H1C11 H2C11 H1C12 H2C12 HC14 H1C15 H2C15 H1C16 H2C16 H1C18 H2C18 H3C18 H1C19 H2C19 H3C19 H1C20 H2C20 H1C21 H2C21 H1C22 H2C22 H3C22 H1C23 H2C23 H1C24 H2C24 H1C25 H2C25 H3C25 HN HO
TABLE 4: Bond Lengths and Bond Angles for STNH from Single Crystal Data
x
y
z
6200 3580 5520 3910 8560 8060 5300 8060 5170 7910 8260 6630 3570 6640 3730 1980 2500 5220 6730 4330 7030 6420 7700 1020 400 1160 1650 2750 2210 1370 2130 3160 4050 -200 -150 1100 5980 6960 2620 3920 3970 5410 6410 2770 8090
-690 -720 -1330 -600 -600 810 670 580 1860 1890 1700 2500 1720 330 -390 380 220 260 1390 2840 2750 3250 2350 2170 1240 1410 590 1230 220 2470 2910 3460 3900 4020 3870 4600 2820 1900 2190 1280 1790 2580 1610 1310 -930
5130 5010 4080 3800 3970 3530 3520 4770 4390 4320 5550 5390 5630 5840 6100 5970 7200 7090 6640 6590 6400 7570 7570 6830 7000 6280 5100 4470 4260 8170 7420 8850 8080 7750 8700 8430 8830 8630 9170 9070 10300 10000 9950 7970 2830
the form of an intense and broad absorption band representative of intermolecular H bonds.27 There is no simple relation between the shape and intensity of the characteristic band with the configuration of the H bonds. Nevertheless, its frequency shift with respect to the position of the absorption band for the nonH-bonded species is a good indicator of the strength of the H bonds.28 Here, ∆ν ∼ 3600-3500 cm-1 corresponds to a rather weak strength of the O-HsN oscillators. The present X-ray analysis shows that H bonds develop between the hydroxyl group (ring A) and amine group (ring D) of adjacent molecules with a mean distance d ) 3.05 Å. The observed ∆ν, d couple of parameters is in the range of the experimental correlations of frequencies and distances already reported in a number of IR studies.27 The present study identifies the development of an H bond sequence along the c axis of the crystalline STNH. Previous investigations26,29 have also shown the implication of H bonds in the fibrillar aggregates forming the solidlike component of the STNH organogels. The molecular design of a molecule so as to be an efficient molecular gelator (hydro or organo) cannot yet be rationally achieved, and most of the discoveries of new gelators are made by serendipity. Also, the molecule-solvent interaction is an important parameter that governs the competi-
atom no. 1
atom no. 2
dist. [Å]
atom no. 3
dist. [Å]
angle [deg]
C1 C2 C3
C10 C1 O O C4 C3 C6 C6 C4 C7 C6 C9 C9 C7 C8 C8 C11 C1 C1 C1 C5 C5 C19 C9 C13 N N N C18 C18 C14 C15 C15 C13 C14 C15 N N N C23 C23 C16 C21 C22 C24 C23 C17
1.523 1.531 1.43 1.43 1.51 1.51 1.519 1.519 1.545 1.516 1.516 1.538 1.538 1.551 1.538 1.538 1.545 1.523 1.523 1.523 1.551 1.551 1.556 1.545 1.545 1.512 1.512 1.512 1.534 1.534 1.538 1.529 1.529 1.538 1.529 1.538 1.5 1.5 1.5 1.541 1.541 1.547 1.544 1.523 1.518 1.518 1.5
C2 C3 C4 C2 C2 C5 C4 C10 C10 C5 C8 C7 C14 C14 C11 C10 C10 C5 C19 C9 C19 C9 C9 C12 C11 C18 C14 C12 C14 C12 C12 C13 C8 C8 C16 C17 C23 C16 C20 C16 C20 C20 C17 C20 C17 C25 C13
1.531 1.533 1.51 1.533 1.533 1.545 1.545 1.551 1.551 1.519 1.551 1.551 1.553 1.553 1.545 1.56 1.56 1.551 1.556 1.56 1.556 1.56 1.56 1.56 1.56 1.534 1.538 1.545 1.538 1.545 1.545 1.538 1.553 1.553 1.538 1.547 1.541 1.547 1.555 1.547 1.555 1.555 1.555 1.544 1.541 1.536 1.512
113.56 111.8 109.95 109.8 110.21 110.36 112.09 112.75 113.86 112.04 111.97 110.99 109.27 110.28 111.35 112.4 112.36 107.39 109.42 110.42 111.56 106.82 111.14 111.84 110.72 113.48 107.15 104.51 113.92 109.8 107.4 111.18 114.61 110.61 109.93 113.23 105.34 107.08 114.05 107.77 111 111.23 115.66 111.05 115.28 112.28 118.41
C4 C5 C6 C7 C8 C9 C10
C11 C12 C13
C14 C15 C16 C17
C20 C21 C23 C24 N
tion between solubilization and crystallization, two processes between which the gelation phenomenon fluctuates.30 It is known31 that some gelator molecules exhibit a crystalline arrangement within the fibers of the gel similar to that in the monocrystalline solid phase. However, while if different morphs are available, the molecular ordering may choose a configuration that has to be understood. In any case, the gelator molecules need to be dissolved at “high” temperatures in the appropriate solvent and then cooled to an instable state driven by the supersaturation degree. In the objective of deciphering the
Figure 4. Overall molecular conformation of STNH.
15132 J. Phys. Chem. B, Vol. 110, No. 31, 2006
Martinetto et al.
Figure 5. Molecular packing in STNH. Hydrogen bonds are indicated by broken lines. The content of one supercell (2a, 2b, 2c) is projected onto (a) the (100) plane and (b) the (010) plane.
structural requirements or evolutions of the condensed states of a low mass molecule to be an efficient gelator, the knowledge of the accurate structure of a reference morph of crystals obtained in nongelling condition is an invaluable prerequisite. Such a favorable context exists due to the easy obtaining of STNH single crystals, and that is not a common situation with gelators (see for instance the organogels derived from fatty acids).32 Future works will be dedicated to the elucidation of the structural relations between the present reference crystalline morph and the solidlike aggregates forming various STNH organogel networks. Conclusions We have solved the crystal structure of the STNH molecule using single-crystal X-ray diffraction and powder synchrotron diffraction data independently. The structure given by the single crystal is more reliable, but we show that direct-space methods using powder diffraction data give satisfactory results concerning the overall conformation of the molecule, the description of the intermolecular interactions, and the crystal packing. STNH crystallizes in the orthorhombic space group P212121, frequently encountered in steroids. The crystalline cohesion is due to the hydrogen bonds that form extended zigzag chains parallel to the c axis and stacked along the b axis. The present demonstration is promising for resolving the structure at an atomic resolution of compounds that cannot be obtained as single crystals. In particular, it is frequent that a derivative exhibiting a good gelating ability cannot be easily obtained as single
crystals suitable for X-ray diffraction analyses. For instance, the fatty acid derivative, 12-hydroxy stearic acid, is known to be a very efficient organogelator, but single crystals cannot be obtained, which does not facilitate the structural investigation.32 In this context, STNH is a nonfrequent exception that will be used in the future to draw structural relationships from the crystal and various solidlike aggregates in the gel networks. The present investigation is a prerequisite for the investigation of the structural features in solidlike self-assembled networks found in the STNH organogels. STNH steroid is a small molecule compared to polymers, and its corresponding molecular gels can be considered in several aspects as model gels in the field. In particular, the steroid organogels in cyclohexane exhibit a remarkable time stability, a high optical transparency, and highly monodisperse cross-sections of the chiral 1d fibrillar supramolecular aggregates. STNH itself can be obtained as single crystals from nongelling conditions. The investigation of the structures in the gel will use the present data on the crystalline STNH as a reference framework. Acknowledgment. Dr. V. Rodriguez is thanked for fruitful discussions and preliminary tests. References and Notes (1) Kariuki, B. M.; Psallidas, K.; Harris, K. D. M.; Johnston, R. L.; Lancaster, R. W.; Staniford, S. E.; Cooper, S. M. Chem. Commun. 1999, 1677-1678. (2) Pagola, S.; Stephens, P. W.; Bohle, D. S.; Kosar, A. D.; Madsen, S. K. Nature 2000, 404, 307-310.
Structure of a D-homoandrostanyl Steroid Derivative (3) Nowell, H.; Attfield, J. P.; Cole, J. C.; Cox, P. J.; Shankland, K.; Maginn, S. J.; Motherwell, W. D. S. New J. Chem. 2002, 26, 469-472. (4) Harris, K. D. M.; Tremayne, M.; Kariuki, B. M. Angew. Chem., Int. Ed. 2001, 40, 1626-1651. (5) Shankland, K.; David, W. I. F. In Structure Determination from Powder Diffraction Data; David, W. I. F., Shankland, K., McCusker L. B., Baerlocher, Ch., Eds.; Oxford University Press: New York, 2002; 252282. (6) Harris, K. D. M. Curr. Opin. Solid State Mater. Sci. 2002, 6, 125130. (7) Harris, K. D. M.; Cheung, E. Y. Chem. Soc. ReV. 2004, 33, 526538. (8) Terech, P.; Volino, F.; Ramasseul, R. J. Phys. France 1985, 46, 895-903. (9) Wade, R. H.; Terech, P.; Hewat, E. A.; Ramasseul, R.; Volino, F. J. Colloid Interface Sci. 1986, 11, 442-451. (10) Weiss, R. G.; Terech, P. Molecular Gels: Materials with SelfAssembled Fibrillar Networks; Springer: Dordrecht, The Netherlands, 2006. (11) Martin-Borret, O.; Ramasseul, R.; Rassat, A. Bull. Soc. Chim. Fr. 1979, II, 401-408. (12) Fitch, A. N. Mater. Sci. Forum 1996, 228-231, 219-221. (13) Hubbard, C. R.; Swanson, H. E.; Mauer, F. A. J. Appl. Crystallogr. 1975, 8, 45-48. (14) Boultif, A.; Loue¨r, D. J. Appl. Crystallogr. 1991, 24, 987-993. (15) Rodrı´guez-Carjaval, J. Program Fullprof.2k, version 2.60; Laboratoire Le´on Brillouin, (CEA-CNRS): Gif-sur-Yvette, France, 2004. (16) Main, P.; Fiske, S. J.; Hull, S. E.; Lessinger, L.; Germain, G.; Declercq, J. P.; Woolfson, M. M. MULTAN 80: A System of Computer Programs for the Automatic Solution of Crystal Structures from X-ray Diffraction Data; Universities of York and Louvain: York, England, and Louvain, Belgium, 1980. (17) Busing, W. R.; Martin, K. O.; Levy, H. A. Report ORNL 59-4-37; Oak Ridge National Laboratory; Oak Ridge, TN, 1971.
J. Phys. Chem. B, Vol. 110, No. 31, 2006 15133 (18) Allen F. H. Acta Crystallogr., Sect. B 2002, 58, 380-388. (19) Bruno, I. J.; Cole, J. C.; Edgington, P. R.; Kessler, M.; Macrae, C. F.; McCabe, P.; Pearson, J.; Taylor, R. Acta Crystallogr., Sect. B 2002, 58, 389-397. (20) Weeks, C. M.; Cooper, A.; Norton, D. A. Acta Crystallogr., Sect. B 1971, 27, 1562-1572. (21) Husain, J.; Tickle, I. J.; Palmer, R. A.; Singh, H.; Bhardwaj, T. R.; Dharam, P. Acta Crystallogr., Sect. B 1982, 38, 130-135. (22) Materials Studio Modeling; Accelrys Software Inc.: San Diego, CA, 2005. (23) Favre-Nicolin, V.; C ˇ erny´, R. J. Appl. Crystallogr. 2002, 35, 734743. (24) Klyne, W. The Chemistry of the Steroids; Wiley: New York, 1960. (25) Duax, W. L.; Norton, D. A. Atlas of Steroid Structure; Plenum: New York, 1975. (26) Terech P. Mol. Cryst. Liq. Cryst. 1989, 166, 29-41. (27) Novak, A. Hydrogen Bonding in Solids. Correlation of Spectroscopic and Crystallographic Data. In Structure and Bonding; Dunitz, J. D., Hemmerich, P., Holm, R. H., Ibers, J. A., Jorgensen, C. K., Neilands, J. B., Reinen, D., Williams, R. J. P., Eds.; Springer-Verlag: Berlin, 1974; p 177. (28) Bratos, S.; Lascombe, J.; Novak, A. Stretching Band of Hydrogen Bonded Systems in Condensed Phases. In Molecular Interactions; Ratajczak, H., Orville-Thomas, W. J., Eds.; Wiley: Chichester, U.K., 1980; Vol. 1. (29) Terech, P. J. Colloid Interface Sci. 1985, 107, 244-255. (30) Estroff, L. A.; Leiserowitz, L.; Addadi, L.; Weiner, S.; Hamilton, A. D. AdV. Mater. 2003, 15, 38-42. (31) Ostuni, E.; Kamaras, P.; Weiss, R. G. Angew. Chem., Int. Ed. Engl. 1996, 35, 1324-1326. (32) Terech, P.; Rodriguez, V.; Barnes, J.; McKenna, G. B. Langmuir 1994, 10, 3406-3418.
