2,2,2-Trifluoro-N-(1a,2,7,7a-tetra­hydro­naphtho­[2,3-b]oxiren-3-yl)acetamide by X-ray powder diffraction

The title compound, C₁₂H₁₀F₃NO₂, an important precursor in the preparation of benzovesamicol analogues for the diagnosis of Alzheimer's disease, was prepared by the epoxidation of 5,8-dihydro­naphthalen-1-amine using 3-chloro­per­oxy­benzoic acid. The structure was determined by X-ray powder diffraction, multinuclear NMR spectroscopy and FT–IR spectroscopy. A pair of mol­ecules form inter­molecular N—H...O hydrogen bonds, involving the amino and oxirene groups, to produce a dimer.

S1. Comment

2,2,2-Trifluoro-N-(1a,2,7,7a-tetrahydronaphtho[2,3-b]oxiren-3-yl)acetamide, (I), is an important precursor for the preparation of benzovesamicol analogues (Zea-Ponce et al., 2005; Jung et al., 1990; Mulholland & Jung 1992; Mulholland et al., 1993; Nicolas et al., 2007), which are used in cholinergic nerve imaging for the diagnosis of Alzheimer's disease and in other research areas (Mazère et al., 2008). This compound is a useful synthetic intermediate and has the advantage of high reactivity at the epoxide ring for introducing additional functionality as part of a more extensive molecular transformation.

The title compound, (I), was prepared as presented in the scheme. The synthesis started with the protection of the amine function of 5,8-dihydronaphthalen-1-amine, (II), by trifluroacetic anhydride to obtain the corresponding amide, (III), in 99% yield. The next step involved the epoxidation of the alkene, (III), by reaction with m-CPBA (m-chloroperbenzoic acid) to produce the corresponding epoxide, (I), in 78% yield. The major byproduct of the alkene epoxidation reaction, however, is the carboxylic acid precursor of the epoxide. The reaction proceeds via a concerted mechanism, as proposed by Bartlett (1950), in which the alkene and the electrophilic peroxy oxygen coordinate with a concomitant expulsion of the carboxylic acid and release of the epoxide (Smith, 2002). The transition state of this reaction has been characterized and is usually represented as shown in Fig. 1.

We employed laboratory powder X-ray diffraction data to solve and refine the crystal structure of (I). This compound crystallizes as a fine white powder and it was not possible to isolate a sample of sufficient size and quality for a single-crystal analysis. This is an 18-atom (non-H) problem, which requires careful measurement and interpretation of the in-house data in order to optimize the quality of the results. Over the past 10 years, the crystal structures of a number of compounds of pharmaceutical interest have been determined by X-ray powder data, essentially as a last resort in the absence of single crystals of sufficient quality (Chan et al., 1999; Shankland et al., 2004; Chernyshev et al., 2003; Kiang et al., 2003; Rukiah et al., 2004; van der Lee et al., 2005).

The details of the solution and refinement of (I) merit a brief comment. We shall refer to the three stages of crystal structure determination from diffraction data: (i) unit-cell determination and space-group assignment, (ii) structure solution and (iii) structure refinement. Structure solution, which aims to obtain an initial approximation to the structure, using the unit cell and space group determined in the first stage, tends to be more arduous with powder data. The final step, refinement, is commonly carried out using the Rietveld method. Structure solution for (I) was initially attempted using the traditional approach (direct methods) with the program EXPO2004 (Altomare et al., 2004); all of our attempts failed. We then used Monte Carlo simulated annealing (parallel tempering algorithm) to solve the structure in direct space using the program FOX (Favre-Nicolin & Černý, 2002). FOX solves structures by altering the positions, orientations and conformations of the molecule(s) in the unit cell, maintaining the space-group symmetry, until a good match is obtained between the calculated and observed intensities. One molecule of (I) was introduced randomly in the monoclinic cell calculated by Le Bail refinement. (H atoms were ignored during structure solution.) During the parallel tempering calculations, the molecule of (I) was permitted to vary by translation, rotation around its centre of mass and variation of its torsion angles. Because the molecule has three independent torsion angles, there are nine degrees of freedom for determination of the starting model by FOX. The model found by this program was introduced into the program GSAS (Larson & Von Dreele, 2004), interfaced by EXPGUI (Toby, 2001) for Rietveld refinements. All non-H bond distances were restrained to their expected values. Before the final refinement, H atoms of CH and CH₂ groups were introduced geometrically and the H atom of the amide was located in a Fourier difference synthesis. The H atoms were refined as riding atoms. The final refinement cycles were performed using isotropic displacement parameters for C, N and O atoms and anisotropic displacement parameters for F. No restraints were used for these displacement parameters. It was necessary to use anisotropic displacement parameters for F to obtain good agreement between the calculated and observed profile at the final stage of the refinement. Of course, the data were of sufficient quality to permit free refinement of these parameters. Intensities were corrected for absorption effects with a function for a flat plate sample in transmission geometry (function number 5 in GSAS) with a µ.d value of 0.2517, determined experimentally. The preferred orientation was modelled using a spherical harmonic description as per Von Dreele (1997) with 20 coefficients. The use of the preferred orientation correction leads to better molecular geometry with better agreement factors.

Compound (I) crystallizes with one molecule in the asymmetric unit in space group P2₁/c (Fig. 2). All bond lengths and bond angles in (I) are in their normal ranges (Allen et al., 1987). The epoxide group, which is the most interesting function in (I), has normal geometry (Table 1). The phenyl ring plus C7 and C10 are coplanar [maximum deviation is 0.035 (11) for C5, when the latter is not used in the calculation of the plane]. The non-F atoms of the acetamide function lie in a plane tilted 62.6 (6)° from the phenyl plane. In the extended structure (Fig. 3) the molecules are connected by van der Waals interactions and hydrogen bonds. Intermolecular N—H···O hydrogen bonds (Table 2) link two independent molecules in a self-recognition pattern to form a dimer (Fig. 3).

The results of the structure analysis by powder diffraction can be correlated with the spectroscopic data obtained for this compound. The IR spectrum of (I) gave the expected band at 3440 cm^-1 which corresponds with the NH group. The characteristic absorption bands of the COCF₃ group appeared at 1715 and 1170 cm^-1 for the CO and CF₃ groups, respectively. The epoxide bands appeared in 3000 cm^-1 (stretch) and 800 cm^-1 (bending), which fall within their normal ranges. The ¹H-NMR spectrum of (I) gave multiplets at 2.95 and 3.40 p.p.m., which are attributed to the two CH₂ and two CH groups. The three aromatic protons (attached to C1, C6, C5, Fig. 2) appeared as a multiplet at 7.17 p.p.m.. A singlet at 8.18 p.p.m. corresponds to the amide group. The ¹³C-NMR of (I) showed 12 singlets corresponding to 12 different carbon environments in the molecule at their expected chemical shifts.

Related literature

For related literature, see: Allen et al. (1987); Altomare et al. (2004); Bartlett (1950); Boultif & Louër (2004); Chan et al. (1999); Chernyshev et al. (2003); Favre-Nicolin & Černý (2002); Finger et al. (1994); Jung et al. (1990); Kiang et al. (2003); Larson & Von Dreele (2004); van der Lee et al. (2005); Mazère et al. (2008); Mulholland & Jung (1992); Mulholland et al. (1993); Nicolas et al. (2007); Roisnel & Rodriguez-Carvajal (2001); Rukiah et al. (2004); Shankland et al. (2004); Smith (2002); Thompson et al. (1987); Toby (2001); Von Dreele (1997); Zea-Ponce et al. (2005).

S2. Experimental

All chemical reagents and solvents were of commercial quality and were used as received. IR spectra were recorded on an FTIR-JASCO 300E spectrometer. NMR spectra were recorded on a Bruker Biospin 400 spectrometer (400 MHz for ¹H, 100 MHz for ¹³C). Chemical shifts (d) were expressed in p.p.m. relative to TMS [definition?] as an internal standard. The melting point was determined in a Stuart SMP3 melting point apparatus. Microanalysis was performed using a EURO EA analyzer.

5,8-Dihydronaphthalen-1-amine, (II) (25 g, 177 mmol), was dissolved in benzene (84 ml) and cooled to 273 K. Trifluoroacetic anhydride (24.7 ml) was added slowly due to the exothermicity of the reaction. The ammonium salt began to precipitate immediately, but the reaction solution was homogenous upon complete addition of the anhydride. The solution was kept at 273 K for 1 h, after which benzene and trifluoroacetic acid were removed under reduced pressure. More benzene was added, then evaporated in order to aid the removal of trifluoroacetic acid. A beige powder of (III) was obtained without purification in 99% yield (41.5 g). The amide, (III), was dissolved in Et₂O (150 ml), then 3-chloroperoxybenzoic acid (70.66 g, 0.204 mmol, 50–55% pure) was added. The solution was maintained at ~283 K throughout the addition and then left stirring at room temperature for 5 h. The epoxide, (I), was collected by filtration and washed with Et₂O, to give 34.7 g (78%) of (I); (m.p. 453 K). Analysis, found for C₁₂H₁₀F₃NO₂: C 56.36, H 4.14, N 4.88%; calculated: C 56.04, H 3.92, N 5.45%.

S3. Refinement

The X-ray powder diffraction pattern was obtained with a Stoe Stadi P diffractometer using monochromatic Cu Kα₁ radiation (λ = 1.5406 Å) selected with an incident beam curved-crystal germanium Ge(111) monochromator, using Stoe transmission geometry (horizontal set-up) with a linear position-sensitive detector (PSD). The powder was ground and loaded between two Mylar foils and fixed in the sample holder with a mask of suitable internal diameter (7.0 mm). Data were collected at room temperature and pressure over the angular range 8–100° (2θ) with a step length of the PSD of 0.5° (2θ) and a counting time of 240 s per step.

The first 20 lines of the powder pattern were indexed with the program DICVOL04 (Boultif & Louër, 2004), using an absolute error of 0.02° (2θ) on the peak positions, on the basis of a monoclinic solution with the unit-cell dimensions a = 15.994 (3), b = 8.8084 (17), c = 8.0584 (14) Å, β = 99.45 (18)°, V = 1119.87 ų with good figures of merit [M₂₀ = 44.5, F₂₀ = 115.5(0.0047, 37)]. The space group was identified as P2₁/a using the program Check Group interfaced by WinPLOTR (Roisnel & Rodriguez-Carvajal, 2001). The parameters a and c were interchanged to give the standard space-group setting, P2₁/c. The number of molecules per unit cell was estimated to be Z = 4, giving Z' = 1. The program FOX was employed for structure solution. The powder pattern was truncated to 45° in 2θ (Cu Kα₁ ), corresponding to a real-space resolution of 2.0 Å.The model found by this program was introduced in the program GSAS (Larson & Von Dreele, 2004), interfaced by EXPGUI (Toby, 2001) for Rietveld refinements.

The background was refined using a shifted Chebyshev polynomial with 20 coefficients. The Thompson–Cox–Hastings (Thompson et al., 1987) pseudo-Voigt profile function was used with an axial divergence asymmetry correction (Finger et al., 1994). The two asymmetry parameters of this function S/L and D/L were both fixed at 0.0215 during the Rietveld refinement. The observed and calculated powder patterns are shown in Fig. 4.

Computing details

Data collection: WinXPOW (Stoe & Cie, 1999); cell refinement: GSAS (Larson & Von Dreele, 2004); data reduction: WinXPOW (Stoe & Cie, 1999); program(s) used to solve structure: (Favre-Nicolin & Černý, 2002); program(s) used to refine structure: GSAS (Larson & Von Dreele, 2004); molecular graphics: ORTEP-3 (Farrugia, 1997); software used to prepare material for publication: publCIF (Westrip, 2010).

Figures

	Fig. 1. Molecular structure of (I) with atomic numbering. Displacement ellipsoids are shown at the 50% probability level; only the F atoms were refined anisotropically.
	Fig. 2. Packing of (I) viewed down the a axis, showing N—H······O hydrogen bonds as dashed lines. H atoms not involved in the motifs have been omitted for clarity. [Author: is the origin in the correct place?]
	Fig. 3. Final observed (points), calculated (line) and difference profiles for the Rietvled refinement of (I).

2,2,2-Trifluoro-N-(1a,2,7,7a-tetrahydronaphtho[2,3-b]oxiren- 3-yl)acetamide

Crystal data

C12H10F3NO2	F(000) = 528.0
Mr = 257.21	Dx = 1.525 Mg m^−3
Monoclinic, P21/c	Cu Kα1 radiation, λ = 1.54060 Å
Hall symbol: -p 2ybc	µ = 1.20 mm^−1
a = 8.0594 (5) Å	T = 298 K
b = 8.8122 (5) Å	Particle morphology: fine powder visual estimate
c = 15.9964 (9) Å	White
β = 99.4514 (7)°	flat sheet, 7 × 7 mm
V = 1120.66 (11) Å^3	Specimen preparation: Prepared at 298 K and 101.3 kPa
Z = 4

Data collection

Stoe transmission STADI P diffractometer	Data collection mode: transmission
Radiation source: sealed X-ray tube, C-Tech	Scan method: step
Ge 111 monochromator	Absorption correction: for a cylinder mounted on the ϕ axis As implemented and documented in GSAS (Larson & Von Dreele, 2004)
Specimen mounting: powder extended between two Mylar foils	Tmin = 0.522, Tmax = 0.545

Refinement

Least-squares matrix: full	Profile function: CW Profile function number 4 with 21 terms Pseudovoigt profile coefficients as parameterized in Thompson et al. (1987). Asymmetry correction of Finger et al. (1994). Microstrain broadening by Stephens (1999). #1(GU) = 0.000 #2(GV) = 0.000 #3(GW) = 11.342 #4(GP) = 0.000 #5(LX) = 0.084 #6(ptec) = 0.00 #7(trns) = 4.80 #8(shft) = -0.2090 #9(sfec) = 0.00 #10(S/L) = 0.0215 #11(H/L) = 0.0215 #12(eta) = 0.7115 #13(S400 ) = 6.6E-02 #14(S040 ) = 3.1E-02 #15(S004 ) = 2.0E-03 #16(S220 ) = 1.5E-02 #17(S202 ) = 9.1E-03 #18(S022 ) = 1.1E-02 #19(S301 ) = 1.9E-03 #20(S103 ) = 4.1E-03 #21(S121 ) = 1.9E-03 Peak tails are ignored where the intensity is below 0.0010 times the peak Aniso. broadening axis 0.0 0.0 1.0
Rp = 0.022	246 parameters
Rwp = 0.030	54 restraints
Rexp = 0.024	H-atom parameters constrained
R(F^2) = 0.04069	(Δ/σ)max < 0.001
χ^2 = 1.613	Background function: GSAS Background function number 1 with 20 terms. Shifted Chebyshev function of 1st kind 1: 1376.94 2: -1226.22 3: 677.625 4: -236.386 5: 46.5456 6: 18.0385 7: -56.8405 8: 40.5980 9: 23.1853 10: -47.8058 11: 26.7618 12: -3.89844 13: -18.4559 14: 12.1618 15: -0.366313 16: 1.98861 17: -4.92608 18: 5.25575 19: -6.53693 20: 3.00978
? data points	Preferred orientation correction: Spherical harmonics function
Excluded region(s): none

Special details

Experimental. The sample was ground lightly in a mortar, loaded between two Mylar foils and fixed in the sample holder with a mask of 7.0 mm internal diameter.

Refinement. Coordinates of the non H-atoms are refined with geometric soft constraints on bond lengths.Values of these constraints are average values of bond lengths for similar chemical functions listed in the paper of Allen et al. (1987). The standard deviation was 0.005 with FACTR = 1.

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
F1	0.9390 (12)	1.1210 (8)	0.6623 (7)	0.08775
F2	1.0605 (12)	0.9365 (10)	0.7380 (7)	0.1202
F3	0.8279 (12)	1.0211 (11)	0.7592 (8)	0.12805
O1	0.1462 (9)	0.8324 (7)	0.5334 (5)	0.047 (3)*
C1	0.6974 (7)	0.8135 (12)	0.4250 (4)	0.036 (5)*
C2	0.6396 (10)	0.8323 (10)	0.4996 (5)	0.036 (4)*
C3	0.4955 (11)	0.7596 (8)	0.5211 (4)	0.029 (4)*
C4	0.3996 (10)	0.6719 (10)	0.4555 (4)	0.030 (4)*
C5	0.4494 (8)	0.6625 (11)	0.3759 (4)	0.047 (5)*
C6	0.5976 (9)	0.7335 (11)	0.3604 (4)	0.038 (4)*
C7	0.4448 (5)	0.7755 (5)	0.6065 (3)	0.041 (5)*
C8	0.2602 (5)	0.7581 (8)	0.6011 (3)	0.051 (5)*
C9	0.1603 (6)	0.6682 (6)	0.5340 (4)	0.041 (5)*
C10	0.2381 (5)	0.5959 (5)	0.4654 (3)	0.039 (5)*
N1	0.7401 (11)	0.9191 (5)	0.5670 (4)	0.027 (3)*
C11	0.8183 (17)	0.8624 (10)	0.6405 (7)	0.057 (5)*
C12	0.9131 (14)	0.9857 (11)	0.6966 (7)	0.063 (7)*
O2	0.8148 (13)	0.7335 (9)	0.6642 (7)	0.060 (4)*
H1	0.81749	0.8314	0.42226	0.075*
H5	0.37644	0.60945	0.32914	0.075*
H6	0.6239	0.73698	0.30214	0.075*
H7a	0.50136	0.69831	0.64402	0.075*
H7b	0.47813	0.87495	0.62927	0.075*
H8	0.20211	0.77819	0.64915	0.075*
H9	0.04886	0.64023	0.54604	0.075*
H10a	0.16035	0.60307	0.41234	0.075*
H10b	0.26004	0.48952	0.47905	0.075*
H1n1	0.73919	1.01745	0.56146	0.075*

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
F1	0.113 (10)	0.047 (6)	0.090 (12)	−0.036 (5)	−0.022 (8)	0.014 (6)
F2	0.104 (10)	0.086 (8)	0.148 (16)	0.026 (6)	−0.047 (9)	−0.059 (7)
F3	0.165 (13)	0.110 (9)	0.126 (13)	−0.053 (7)	0.073 (10)	−0.071 (9)

Geometric parameters (Å, º)

F1—C12	1.343 (13)	C5—H5	0.989
F2—C12	1.334 (15)	C6—H6	0.989
F3—C12	1.340 (16)	C7—C8	1.485 (4)
O1—C8	1.455 (5)	C7—H7a	0.970
O1—C9	1.451 (5)	C7—H7b	0.970
C1—C2	1.360 (6)	C8—C9	1.463 (5)
C1—C6	1.393 (11)	C8—H8	0.980
C1—H1	0.989	C9—C10	1.494 (4)
C2—C3	1.417 (6)	C9—H9	0.981
C2—N1	1.454 (6)	C10—H10a	0.970
C3—C4	1.424 (5)	C10—H10b	0.972
C3—C7	1.495 (5)	N1—C11	1.336 (13)
C4—C5	1.400 (6)	N1—H1n1	0.871
C4—C10	1.495 (5)	C11—C12	1.532 (15)
C5—C6	1.406 (11)	C11—O2	1.199 (12)
C8—O1—C9	60.4 (3)	C7—C8—C9	121.6 (4)
C2—C1—C6	118.3 (5)	C7—C8—H8	122.6
C2—C1—H1	119.9	C9—C8—H8	113.4
C6—C1—H1	119.8	O1—C9—C8	59.9 (3)
C1—C2—C3	124.7 (6)	O1—C9—C10	117.5 (5)
C1—C2—N1	119.0 (6)	O1—C9—H9	100.3
C3—C2—N1	115.9 (7)	C8—C9—C10	121.5 (4)
C2—C3—C4	115.7 (6)	C8—C9—H9	113.8
C2—C3—C7	122.4 (6)	C10—C9—H9	123.0
C4—C3—C7	121.8 (6)	C4—C10—C9	111.4 (4)
C3—C4—C5	120.0 (5)	C4—C10—H10a	109.1
C3—C4—C10	122.8 (6)	C4—C10—H10b	109.1
C5—C4—C10	117.0 (4)	C9—C10—H10a	109.0
C4—C5—C6	120.9 (4)	C9—C10—H10b	109.2
C4—C5—H5	119.4	H10a—C10—H10b	109.1
C6—C5—H5	119.6	C2—N1—C11	125.4 (6)
C1—C6—C5	119.9 (4)	C2—N1—H1n1	117.0
C1—C6—H6	119.8	C11—N1—H1n1	117.1
C5—C6—H6	119.9	N1—C11—C12	111.6 (7)
C3—C7—C8	111.0 (5)	N1—C11—O2	126.9 (10)
C3—C7—H7a	109.2	C12—C11—O2	121.4 (9)
C3—C7—H7b	109.2	F1—C12—F2	107.8 (11)
C8—C7—H7a	109.1	F1—C12—F3	103.4 (10)
C8—C7—H7b	109.2	F1—C12—C11	119.1 (9)
H7a—C7—H7b	109.1	F2—C12—F3	103.3 (10)
O1—C8—C7	120.0 (6)	F2—C12—C11	112.2 (9)
O1—C8—C9	59.7 (3)	F3—C12—C11	109.6 (10)
O1—C8—H8	100.3
C11—N1—C2—C1	−111.8 (12)	C3—C4—C5—C6	2.7 (13)
C11—N1—C2—C3	61.6 (13)	C3—C4—C10—C9	21.6 (9)
C2—N1—C11—O2	−3 (2)	C5—C4—C10—C9	−154.1 (7)
C2—N1—C11—C12	179.8 (9)	C4—C5—C6—C1	0.5 (13)
C6—C1—C2—C3	8.9 (14)	C3—C7—C8—C9	27.0 (7)
C2—C1—C6—C5	−6.0 (13)	C3—C7—C8—O1	−43.7 (7)
C6—C1—C2—N1	−178.4 (8)	C7—C8—C9—O1	−108.7 (6)
C1—C2—C3—C4	−5.6 (12)	O1—C8—C9—C10	105.8 (6)
C1—C2—C3—C7	175.4 (7)	C7—C8—C9—C10	−3.0 (8)
N1—C2—C3—C4	−178.6 (7)	C8—C9—C10—C4	−21.4 (7)
N1—C2—C3—C7	2.4 (11)	O1—C9—C10—C4	48.5 (7)
C7—C3—C4—C5	178.7 (7)	O2—C11—C12—F1	168.2 (12)
C2—C3—C4—C5	−0.3 (12)	O2—C11—C12—F2	40.9 (17)
C2—C3—C4—C10	−176.0 (7)	O2—C11—C12—F3	−73.2 (15)
C4—C3—C7—C8	−27.2 (9)	N1—C11—C12—F1	−14.0 (16)
C2—C3—C7—C8	151.7 (7)	N1—C11—C12—F2	−141.3 (11)
C7—C3—C4—C10	3.0 (11)	N1—C11—C12—F3	104.6 (12)
C10—C4—C5—C6	178.6 (7)

Hydrogen-bond geometry (Å, º)

D—H···A	D—H	H···A	D···A	D—H···A
N1—H1N1···O1^i	0.87	2.32	2.949 (9)	130

Symmetry code: (i) −x+1, −y+2, −z+1.