Inversion twinning in a second polymorph of the hydro­chloride salt of the recreational drug ethyl­one

A second polymorph of the hydro­chloride salt of the recreational drug ethyl­one, C₁₂H₁₆NO₃⁺·Cl⁻, is reported [systematic name: (±)-2-ethyl­ammonio-1-(3,4-methyl­ene­dioxy­phen­yl)propane-1-one chloride]. This polymorph, denoted form (A), appears in crystallizations performed above 308 K. The originally reported form (B) [Wood et al. (2015). Acta Cryst. C71, 32–38] crystallizes preferentially at room temperature. The conformations of the cations in the two forms differ by a 180° rotation about the C—C bond linking the side chain to the aromatic ring. Hydrogen bonding links the cations and anions in both forms into similar extended chains in which any one chain contains only a single enanti­omer of the chiral cation, but the packing of the ions is different. In form (A), the aromatic rings of adjacent chains inter­leave, but pack equally well if neighbouring chains contain the same or opposite enanti­omorph of the cation. The consequence of this is then near perfect inversion twinning in the structure. In form (B), neighbouring chains are always inverted, leading to a centrosymmetric space group. The question as to why the polymorphs crystallize at slightly different temperatures has been examined by density functional theory (DFT) and lattice energy calculations and a consideration of packing compactness. The free energy (ΔG) of the crystal lattice for polymorph (A) lies some 52 kJ mol⁻¹ above that of polymorph (B).

1. Introduction

Ethyl­one [also called 3,4-methyl­enedi­oxy-N-ethyl­cathinone or (±)-1-(1,3-benzodioxol-5-yl)-2-(ethyl­amino)­propan-1-one], (I), is controlled as an amphetamine analogue under the Controlled Drugs and Substances Act in Canada. Ethyl­one was patented in 1996 as an anti­depressant (Jacob & Shulgin, 1996) and some analytical data were published shortly thereafter in an effort to rapidly identify this compound should it appear in the underground drug market (Dal Cason, 1997). However, two different polymorphic forms of the hydro­chloride salt of ethyl­one, herein labelled (A) and (B), were discovered when seized exhibits of ethyl­one hydro­chloride inter­cepted at the Canadian border were found to have different spectroscopic data (FT–IR, FT–Raman and powder X-ray diffraction) compared with those from a synthesized reference standard (Maheux et al., 2015). We have found that different methods of preparation of ethyl­one hydro­chloride at room temperature produced large block-shaped crystals of form (B), small crystals of form (A) that were not suitable for crystallography, or a mixture of both. Polymorph (A) appears as very small fine needle crystals. A typical large specimen among these small crystals measured 0.5 × 0.5 × 15.0 µm. If (A) is recrystallized at a temperature above 308 K, then polymorph (A) persists. After many recrystallization attempts, it was found that if solid (A) was left in contact with a saturated solution of the compound in a 50:50 v/v water–methanol mixture [essentially damp crystals, since (A) is very soluble] and this mixture was left in a temperature cycler for a period of more than four weeks, then ultimately a few needle-shaped crystals up to 0.3 mm long and of a quality just suitable for an X-ray crystal structure determination were obtained. We report here the crystal structure of (A) at 160 K, together with a comparison with the structure of (B) at 100 K, which has been reported recently (Wood et al., 2015), although we have also determined the structure of (B) at 160 K, at room temperature and at 313 K and found no phase change across this temperature range. We also consider reasons why there are two polymorphs and suggest why they might form at different temperatures.

2. Experimental

2.1. Synthesis and crystallization

Samples of synthesized and seized ethyl­one hydro­chloride were supplied by the Canada Border Services Agency. A small sample of polymorph (A) was dissolved in a methanol–water mixture (50:50 v/v) and allowed to evaporate slowly in a temperature cycler that raised and lowered the temperature of the solution over the range 308–311 K, with each complete cycle (308–311–308 K) lasting about 40 min. When very little liquid was left, the vial was sealed tightly and the temperature cycling continued for four weeks. The final result contained many very fine needles, but among the lumps a few larger needle-shaped crystals were found, the largest of these were approximately 0.04 × 0.05 × 0.30 mm and were (just) suitable for X-ray crystal structure determination.

2.2. Refinement

Crystal data, data collection and structure refinement details are summarized in Table 1. All H atoms were placed in geometrically idealized positions and constrained to ride on their parent atoms, with N—H = 0.91 Å and C—H = 0.95 (aromatic), 0.98 (meth­yl), 0.99 (methyl­ene) and 1.00 Å (methine), and with U_iso(H) = 1.5U_eq(C) for the methyl groups and 1.2U_eq(C,N) otherwise. Initial refinement of the completed structure model yielded an absolute structure parameter of 0.497 (12), based on the quotients method (Parsons et al., 2013), which indicated the presence of an inversion twin. For the final refinements, the TWIN/BASF instructions were included in the SHELXL2014 instruction file (Sheldrick, 2015), so as to include the contribution of both twin components to the structure-factor calculations during the least-squares optimization, and the major twin fraction refined to 0.50 (5). This procedure is important when an inversion twin has been detected, because, in the absence of these instructions, the absolute structure parameter calculated by SHELXL2014 is only done post-refinement, without including the contribution from the inverse model in the least-squares calculations. When the absolute structure parameter deviates significantly from zero and its standard uncertainty is sufficiently small for the value to be meaningful, failure to include TWIN/BASF in the refinement can lead to bias in the final model. In this case, refinement without TWIN/BASF led to R[F² > 2σ(F²)] = 0.069, compared with the lower value of 0.067 associated with the reported refinement results for which these instructions were included.

Table 1
Experimental details

Crystal data
Chemical formula	C12H16NO3^+·Cl^−
Mr	257.71
Crystal system, space group	Orthorhombic, P212121
Temperature (K)	160
a, b, c (Å)	6.90225 (16), 7.13000 (16), 25.4692 (5)
V (Å^3)	1253.42 (5)
Z	4
Radiation type	Cu Kα
μ (mm^−1)	2.69
Crystal size (mm)	0.30 × 0.06 × 0.03
Data collection
Diffractometer	Oxford Diffraction SuperNova (dual radiation) diffractometer
Absorption correction	Gaussian (CrysAlis PRO; Agilent, 2014)
Tmin, Tmax	0.707, 0.923
No. of measured, independent and observed [I > 2σ(I)] reflections	7202, 2421, 2286
Rint	0.032
(sin θ/λ)max (Å^−1)	0.626
Refinement
R[F^2 > 2σ(F^2)], wR(F^2), S	0.067, 0.180, 1.11
No. of reflections	2421
No. of parameters	157
H-atom treatment	H-atom parameters constrained
Δρmax, Δρmin (e Å^−3)	0.74, −0.34
Absolute structure	Refined as an inversion twin using 929 Friedel pairs
Absolute structure parameter	0.50 (5)
Computer programs: CrysAlis PRO (Agilent, 2014), SHELXS2014 (Sheldrick, 2008), ORTEPII (Johnson, 1976), Mercury (Macrae et al., 2008), CrystalStructure (Rigaku, 2007), SHELXL2014 (Sheldrick, 2015) and PLATON (Spek, 2015).

3. Results and discussion

Ethyl­one, (I), is constructed from an aromatic planar 1,3-benzodioxole [1,2-(methyl­enedi­oxy)benzene] unit with a C(O)C(CH₃)NHCH₂CH₃ side chain at the 4-position of the benzene ring. In the hydro­chloride salt of ethyl­one, the N atom of the free base is protonated (see Scheme). Polymorph (A) of the hydro­chloride salt crystallizes in the space group P2₁2₁2₁ with one cation and a chloride anion in the asymmetric unit (Fig. 1). The cation contains a chiral C atom (C8), yet the compound is a racemate and in the chosen crystal crystallizes as a perfect inversion twin. The two NH₂⁺ H atoms hydrogen bond to two symmetry-related Cl⁻ ions, with H⋯Cl distances of 2.24 and 2.30 Å (Table 2). This links the ammonium groups of the cations to the anions in an alternating sequence into a simple zigzag chain that propagates parallel to the [010] direction and can be described by a graph-set motif of C₂¹(4) (Bernstein et al., 1995). The direction of the hydrogen-bonded chains corresponds to the needle axis of the crystal (b axis). The alkyl groups of the cations are directed towards the core of the chain, while the planar aromatic groups of adjacent cations within the chain are disposed alternately on opposite sides of the chain core (Fig. 1b) and, in the crystal, these planar groups lie almost exactly on and parallel to the (100) plane. Fig. 2 shows a view of the crystal packing projected down [010] parallel to the chain axis, so, for example, the bottom left shows the chain spiralling along a 2₁ screw axis going into the page. Adjacent chains along the [100] direction are simply repeats by a unit-cell translation. If one such sequence of parallel chains is considered, then the aromatic groups slot neatly between those from the neighbouring equivalent sequences on either side in the [001] direction, although they are offset in the [010] direction to preclude the existence of π–π stacking inter­actions. This is shown by the central vertical stack of aromatic groups in Fig. 2. This inter­leaving provides an explanation for both the formation of the inversion twin and the difficulty in growing a crystal of any size. In the crystal, each hydrogen-bonded chain is a unique enanti­opure unit. Yet the next chain, formed in the crystal by the inter­leaving of the aromatic planar groups (left-to-right in Fig. 2), has no guiding requirement other than that all cations in that chain be the same enanti­omorph; the packing is such that a chain need not contain the same enanti­omorph as that in an adjacent chain. However, the chains above and below (top and bottom left in Fig. 2) will probably need to be composed of the same enanti­omorph, thus the crystal will form enanti­opure layers lying parallel to (001), but with each layer able to be composed of either one of the two enanti­omorphs. This random packing of the layers of the two enanti­omorphs will produce an inversion twin. Moreover, the very precise requirement within any one chain and layer of having a single enanti­omorph within that chain and layer, while a cation of either enanti­omorph can slip into any slot, makes construction of a layer a process potentially littered with errors, which lead to defects that have to be corrected before the crystal can grow to any size. If the errors were not corrected, the structure would necessarily be disordered, which is not observed with the crystal used for the measurements.

Table 2
Hydrogen-bond geometry (Å, °)

D—H⋯A	D—H	H⋯A	D⋯A	D—H⋯A
N1—H1⋯Cl1	0.91	2.24	3.149 (4)	174
N1—H2⋯Cl1^i	0.91	2.30	3.134 (5)	153
Symmetry code: (i) .

Figure 1 (a) The asymmetric unit of polymorph (A), showing the atom-labelling scheme and one of the hydrogen bonds linking the ions (dashed line). Displacement ellipsoids are drawn at the 50% probability level. (b) A hydrogen-bonded chain (yellow bonds) within the structure of polymorph (A).

Figure 2 Packing in the unit cell of polymorph (A), viewed down the hydrogen-bonded chains (down [010]). Aromatic groups lie approximately parallel to the (100) plane; those from chains along the 21 screw axes at z = ¼, ¾ inter­leave at z = ½, although they are offset in the [010] direction.

The structure of polymorph (A) described here was determined at 160 K. The structure has also been determined at room temperature where, apart from the expected differences caused by the rise in temperature, it is exactly the same as the low-temperature structure and there is no indication of a phase change to polymorph (B). Similarly, the structure of polymorph (B) shows no indication of a phase change to polymorph (A) across the temperature range from 100 (Wood et al., 2015) to 313 K, as mentioned above. Polymorph (B) crystallizes in the centrosymmetric space group P2₁/c with thus both enanti­omers present as a perfect racemate. The hydrogen-bonding scheme in (B), not described in detail by Wood et al. (2015), also produces a simple single chain sequence with the C₂¹(4) motif (Fig. 3a), where the aromatic groups again alternate on opposite sides of the chain. Here, however, the cations are much more tightly concentrated so that the aromatic groups on each side of the chain are now tightly packed together (Figs. 3a and 3b) and there is no possibility of inter­leaving with an adjacent chain. Since the chain is of a zigzag nature and propagates along a 2₁ screw axis, each chain, as expected, contains only one of the two enanti­omorphs, while the chain related by the inversion centre contains the other enanti­omorph.

Figure 3 (a) A hydrogen-bonded chain (yellow bonds) within the structure of polymorph (B) (Wood et al., 2015). (b) Close overlap of the aromatic rings within the hydrogen-bonded chain of polymorph (B).

Once the cations in each polymorph are examined, the reason for the two polymorphs is immediately obvious. The carbonyl group in polymorph (A) is oriented in the opposite direction with respect to its orientation in polymorph (B) relative to the fused-ring system. Essentially, the entire side chain is rotated by approximately 180° about the C1—C7 bond linking the side chain to the ring system (Fig. 4). Thus, each polymorph contains a completely different conformer. Given the considerable mol­ecular reorganization required to change from one conformation to the other, it is not surprising that no phase change is observed in the solid state when moving from the preferred crystallization temperature of one polymorph to that of the other.

Figure 4 An overlay of the cations in polymorphs (A) and (B), showing the different conformation of the side chain. For clarity, the image of polymorph (A) is displaced slightly upwards.

An examination of Figs. 1(b) and 3(b) shows a considerable difference between the arrangement of adjacent aromatic groups within the hydrogen-bonded chain of each of the two polymorphs. This is perhaps the result of π-stacking of the aromatic groups of adjacent cations for (B), which is not available for (A). In (A), the aromatic groups of the cations pack side-by-side along b, with very poor overlap with the groups in the cations above and below, while in the crystal of (B), pairs of aromatic groups from adjacent chains manage a reasonable overlap of the benzene rings at a centroid–centroid distance of 3.6174 (12) Å and a slippage of 1.15 Å [based on the data of Wood et al. (2015)].

The remaining question is why polymorph (A) is formed at temperatures above 308 K, when (B) is preferred at, or below, room temperature (293 K). To examine this question we need to consider (a) the relative solvation energies for the two conformers of the ammonium cations, plus the solvation energy of the chloride ion, (b) the activation energy barrier for conversion of one conformer of the solvated ammonium cations into the other conformer, (c) the activation energy for the change from solution state to solid state, (d) the concentration of the two conformers when crystallization occurs, (e) the difference in the lattice energies of the two polymorphs, and (f) effects of spontaneous nucleation (Fig. 5).

Figure 5 Energy changes from a crystal of polymorph (A) to a crystal of polymorph (B) through a solution in methanol/water (ts indicates the transition states).

Given the complexity of the methanol–water system used for crystallization, we did not investigate variable (a), other than to recognize that the solvation energies of the two conformers will be different. It has been noted that solvation of alkyl­ammonium ions is different from that of the ammonium ion and is influenced by electrostatic nonlocal inter­actions involving the alkyl groups (Vallet & Masella, 2015).

Variable (b) was estimated using density functional theory (DFT) computations (GAUSSIAN09; Frisch et al., 2010) using the wB97XD/6-311+g(d) level functional (Chai & Head-Gordon, 2008) and basis set. Solvation was modelled using the self-consistent reaction field polarizable continuum model (SCRF-PCM; Tomasi et al., 2005) with the solvent set to the dielectric constants for either methanol or water. All structures were fully optimized and a frequency analysis was done also, which gave energy minima having no imaginary frequencies and transition states with one. The comparable gas-phase computations were made for comparison. Results giving the free energies (ΔG) are given in Table 3. Thus, the transition state energy relative to the two conformers appears to be slightly higher in water and their solvation energies also reflect the differences in solvent polarity.

Table 3
Relative free energies for polymorphs (A) and (B), as well as the rotational transition state TS(AB) in the gas phase [298 K, 1 atm (1 atm = 101 325 Pa)] and applying the self-consistent reaction field polarizable continuum model (SCRF-PCM) for methanol and water (298 K, 1 atm)^a

Computation	ΔG (kJ mol^−1)
(A), gas phase	0
(AB), gas phase	34
(B), gas phase	35
(A), MeOH	0
(AB), MeOH	31
(B), MeOH	14
(A), HOH	0
(AB), HOH	30
(B), HOH	12
Note: (a) computations done at 288 and 318 K yielded essentially the same relative energies.

Variables (c) and (d) were deemed to be not qu­anti­fiable, but variable (e) could be computed. This was done using the CP2K computation suite for condensed matter (Hutter et al., 2014) at the PBE+D3(TZV2PX) level. The computations were done within periodic boundary conditions with four cations and four anions per unit cell (132 atoms), with pseudo­potentials for all atoms and 376 valence electrons per unit cell. These computations showed that the free energy (ΔG) of the crystal lattice for polymorph (A) was some 52 kJ mol⁻¹ greater than that of polymorph (B). Finally, variable (f) was also considered not to be qu­anti­fiable.

These energy values in conjunction with the recrystallization observations suggest a plausible explanation for the formation of the two polymorphs. At the higher recrystallization temperature in the methanol–water system, the single cation with the anion of the (A) conformer, which is more stable in solution than that of the (B) conformer, dominates. At this higher temperature, with the moderate activation energy barrier for the (B) conformer to change to (A), the (A) polymorph is formed. As the temperature falls, the cations and anions start to aggregate and at this point the more compact aggregate (compare Figs. 1b and 3b) favours the more stable crystal structure, which is (B), and the (B) polymorph is formed.

Supporting information

Computing details

Data collection: CrysAlis PRO (Agilent, 2014); cell refinement: CrysAlis PRO (Agilent, 2014); data reduction: CrysAlis PRO (Agilent, 2014); program(s) used to solve structure: SHELXS2014 (Sheldrick, 2008); program(s) used to refine structure: SHELXL2014 (Sheldrick, 2015); molecular graphics: ORTEPII (Johnson, 1976), Mercury (Macrae et al., 2008) and CrystalStructure (Rigaku, 2007); software used to prepare material for publication: SHELXL2014 (Sheldrick, 2015) and PLATON (Spek, 2015).

[1-(1,3-Benzodioxol-5-yl)-1-oxopropan-2-yl]ethanaminium chloride

Crystal data

C12H16NO3^+·Cl^−	Dx = 1.366 Mg m^−3
Mr = 257.71	Cu Kα radiation, λ = 1.54184 Å
Orthorhombic, P212121	Cell parameters from 3179 reflections
a = 6.90225 (16) Å	θ = 6.4–73.0°
b = 7.13000 (16) Å	µ = 2.69 mm^−1
c = 25.4692 (5) Å	T = 160 K
V = 1253.42 (5) Å^3	Needle, pale yellow
Z = 4	0.30 × 0.06 × 0.03 mm
F(000) = 544

Data collection

Oxford Diffraction SuperNova (dual radiation) diffractometer	2421 independent reflections
Radiation source: SuperNova (Cu) X-ray Source	2286 reflections with I > 2σ(I)
Mirror monochromator	Rint = 0.032
Detector resolution: 10.3801 pixels mm^-1	θmax = 74.8°, θmin = 3.5°
ω scans	h = −8→8
Absorption correction: gaussian (CrysAlisPro; Agilent, 2014)	k = −8→8
Tmin = 0.707, Tmax = 0.923	l = −31→31
7202 measured reflections

Refinement

Refinement on F^2	Hydrogen site location: inferred from neighbouring sites
Least-squares matrix: full	H-atom parameters constrained
R[F^2 > 2σ(F^2)] = 0.067	w = 1/[σ^2(Fo^2) + (0.1079P)^2 + 1.0836P] where P = (Fo^2 + 2Fc^2)/3
wR(F^2) = 0.180	(Δ/σ)max = 0.001
S = 1.11	Δρmax = 0.74 e Å^−3
2421 reflections	Δρmin = −0.34 e Å^−3
157 parameters	Absolute structure: Refined as an inversion twin using 929 Friedel pairs
0 restraints	Absolute structure parameter: 0.50 (5)

Special details

Experimental. Solvent used: 50:50 v.v. MeOH-H2O Cooling Device: Oxford Instruments Cryojet XL Crystal mount: on a glass fibre Frames collected: 2184 Seconds exposure per frame: 5.0–75.0 Degrees rotation per frame: 0.5 Crystal-detector distance (mm): 55.0 Client: Stan Cameron Sample code: ethlone hydrochloride (L1407)

Geometry. All e.s.d.'s (except the e.s.d. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell e.s.d.'s are taken into account individually in the estimation of e.s.d.'s in distances, angles and torsion angles; correlations between e.s.d.'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell e.s.d.'s is used for estimating e.s.d.'s involving l.s. planes.

Refinement. Refined as a 2-component inversion twin using 929 Friedel pairs

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

	x	y	z	Uiso*/Ueq
Cl1	−0.0103 (2)	0.35002 (17)	0.80071 (4)	0.0414 (4)
O1	0.1371 (7)	0.1702 (6)	0.58409 (15)	0.0469 (10)
O2	0.0680 (8)	0.4897 (6)	0.40324 (15)	0.0494 (11)
O3	0.0217 (6)	0.8065 (5)	0.41849 (13)	0.0424 (9)
N1	0.1330 (7)	0.2704 (6)	0.68548 (16)	0.0335 (9)
H1	0.0832	0.2945	0.7179	0.040*
H2	0.1048	0.1491	0.6774	0.040*
C1	0.0643 (8)	0.4676 (8)	0.54772 (19)	0.0337 (11)
C2	0.0773 (9)	0.3968 (8)	0.4959 (2)	0.0393 (12)
H21	0.0992	0.2676	0.4889	0.047*
C3	0.0565 (8)	0.5253 (8)	0.45632 (19)	0.0360 (12)
C4	0.0302 (7)	0.7116 (7)	0.46539 (19)	0.0333 (11)
C5	0.0214 (7)	0.7849 (7)	0.51485 (18)	0.0319 (10)
H5	0.0059	0.9156	0.5206	0.038*
C6	0.0361 (7)	0.6590 (7)	0.55644 (18)	0.0323 (10)
H6	0.0268	0.7042	0.5914	0.039*
C7	0.0879 (8)	0.3315 (8)	0.59123 (19)	0.0343 (10)
C8	0.0363 (8)	0.3947 (7)	0.64665 (17)	0.0311 (11)
H8	0.0811	0.5267	0.6520	0.037*
C9	−0.1808 (9)	0.3853 (9)	0.6540 (2)	0.0436 (14)
H91	−0.2247	0.2557	0.6493	0.065*
H92	−0.2442	0.4660	0.6280	0.065*
H93	−0.2141	0.4281	0.6894	0.065*
C10	0.3466 (8)	0.2923 (8)	0.6877 (2)	0.0389 (12)
H101	0.3792	0.4244	0.6957	0.047*
H102	0.4031	0.2608	0.6530	0.047*
C11	0.4327 (9)	0.1655 (9)	0.7293 (2)	0.0441 (13)
H111	0.3945	0.0355	0.7224	0.066*
H112	0.3850	0.2036	0.7640	0.066*
H113	0.5742	0.1755	0.7285	0.066*
C12	0.0225 (9)	0.6636 (8)	0.37870 (19)	0.0437 (12)
H121	−0.1062	0.6562	0.3617	0.052*
H122	0.1203	0.6934	0.3515	0.052*

Atomic displacement parameters (Ų)

	U^11	U^22	U^33	U^12	U^13	U^23
Cl1	0.0673 (8)	0.0288 (6)	0.0282 (5)	0.0097 (6)	0.0067 (6)	0.0024 (4)
O1	0.075 (3)	0.033 (2)	0.0327 (18)	0.004 (2)	0.0023 (18)	−0.0044 (16)
O2	0.080 (3)	0.043 (2)	0.0249 (17)	−0.005 (2)	0.0033 (18)	−0.0025 (16)
O3	0.059 (2)	0.042 (2)	0.0267 (16)	−0.006 (2)	0.0019 (16)	0.0035 (14)
N1	0.046 (2)	0.027 (2)	0.0273 (18)	0.0010 (18)	0.0011 (17)	−0.0015 (16)
C1	0.037 (2)	0.036 (3)	0.028 (2)	−0.005 (2)	0.0015 (19)	−0.001 (2)
C2	0.051 (3)	0.037 (3)	0.030 (2)	−0.006 (2)	0.003 (2)	−0.003 (2)
C3	0.040 (3)	0.042 (3)	0.026 (2)	−0.006 (2)	0.0033 (19)	−0.002 (2)
C4	0.033 (2)	0.037 (3)	0.030 (2)	−0.008 (2)	0.0016 (18)	0.0057 (19)
C5	0.033 (2)	0.029 (2)	0.034 (2)	−0.002 (2)	0.003 (2)	0.0005 (18)
C6	0.035 (2)	0.034 (2)	0.028 (2)	−0.004 (2)	0.0006 (17)	−0.0041 (19)
C7	0.043 (2)	0.032 (3)	0.028 (2)	−0.005 (2)	−0.0032 (19)	−0.001 (2)
C8	0.046 (3)	0.021 (2)	0.026 (2)	−0.003 (2)	−0.0004 (19)	−0.0007 (16)
C9	0.048 (3)	0.050 (4)	0.033 (3)	0.000 (3)	0.004 (2)	−0.001 (2)
C10	0.043 (3)	0.040 (3)	0.033 (2)	−0.005 (2)	0.001 (2)	0.002 (2)
C11	0.050 (3)	0.044 (3)	0.038 (3)	−0.005 (3)	−0.004 (2)	0.009 (2)
C12	0.059 (3)	0.044 (3)	0.027 (2)	−0.005 (3)	0.000 (2)	0.000 (2)

Geometric parameters (Å, º)

O1—C7	1.213 (7)	C5—H5	0.9500
O2—C3	1.378 (6)	C6—H6	0.9500
O2—C12	1.424 (7)	C7—C8	1.524 (7)
O3—C4	1.374 (6)	C8—C9	1.511 (8)
O3—C12	1.437 (6)	C8—H8	1.0000
N1—C10	1.484 (7)	C9—H91	0.9800
N1—C8	1.486 (6)	C9—H92	0.9800
N1—H1	0.9100	C9—H93	0.9800
N1—H2	0.9100	C10—C11	1.515 (8)
C1—C6	1.396 (7)	C10—H101	0.9900
C1—C2	1.417 (7)	C10—H102	0.9900
C1—C7	1.482 (7)	C11—H111	0.9800
C2—C3	1.369 (8)	C11—H112	0.9800
C2—H21	0.9500	C11—H113	0.9800
C3—C4	1.360 (8)	C12—H121	0.9900
C4—C5	1.365 (7)	C12—H122	0.9900
C5—C6	1.392 (7)
C3—O2—C12	104.9 (4)	N1—C8—C7	109.6 (4)
C4—O3—C12	105.3 (4)	C9—C8—C7	109.5 (4)
C10—N1—C8	114.1 (4)	N1—C8—H8	109.4
C10—N1—H1	108.7	C9—C8—H8	109.4
C8—N1—H1	108.7	C7—C8—H8	109.4
C10—N1—H2	108.7	C8—C9—H91	109.5
C8—N1—H2	108.7	C8—C9—H92	109.5
H1—N1—H2	107.6	H91—C9—H92	109.5
C6—C1—C2	120.4 (5)	C8—C9—H93	109.5
C6—C1—C7	122.5 (4)	H91—C9—H93	109.5
C2—C1—C7	117.2 (5)	H92—C9—H93	109.5
C3—C2—C1	116.1 (5)	N1—C10—C11	110.7 (4)
C3—C2—H21	121.9	N1—C10—H101	109.5
C1—C2—H21	121.9	C11—C10—H101	109.5
C4—C3—C2	122.9 (5)	N1—C10—H102	109.5
C4—C3—O2	110.7 (5)	C11—C10—H102	109.5
C2—C3—O2	126.3 (5)	H101—C10—H102	108.1
C3—C4—C5	122.4 (5)	C10—C11—H111	109.5
C3—C4—O3	109.8 (4)	C10—C11—H112	109.5
C5—C4—O3	127.7 (5)	H111—C11—H112	109.5
C4—C5—C6	116.9 (5)	C10—C11—H113	109.5
C4—C5—H5	121.6	H111—C11—H113	109.5
C6—C5—H5	121.6	H112—C11—H113	109.5
C5—C6—C1	121.3 (4)	O2—C12—O3	108.0 (4)
C5—C6—H6	119.4	O2—C12—H121	110.1
C1—C6—H6	119.4	O3—C12—H121	110.1
O1—C7—C1	122.7 (5)	O2—C12—H122	110.1
O1—C7—C8	119.0 (5)	O3—C12—H122	110.1
C1—C7—C8	118.2 (5)	H121—C12—H122	108.4
N1—C8—C9	109.7 (4)
C6—C1—C2—C3	−1.3 (8)	C2—C1—C6—C5	−0.4 (8)
C7—C1—C2—C3	−179.5 (5)	C7—C1—C6—C5	177.6 (5)
C1—C2—C3—C4	1.8 (8)	C6—C1—C7—O1	−170.2 (5)
C1—C2—C3—O2	178.5 (5)	C2—C1—C7—O1	7.9 (8)
C12—O2—C3—C4	−7.2 (6)	C6—C1—C7—C8	13.5 (7)
C12—O2—C3—C2	175.8 (6)	C2—C1—C7—C8	−168.3 (5)
C2—C3—C4—C5	−0.4 (9)	C10—N1—C8—C9	−170.5 (4)
O2—C3—C4—C5	−177.5 (5)	C10—N1—C8—C7	69.3 (5)
C2—C3—C4—O3	177.6 (5)	O1—C7—C8—N1	23.9 (7)
O2—C3—C4—O3	0.5 (7)	C1—C7—C8—N1	−159.8 (4)
C12—O3—C4—C3	6.4 (6)	O1—C7—C8—C9	−96.4 (6)
C12—O3—C4—C5	−175.8 (5)	C1—C7—C8—C9	79.9 (6)
C3—C4—C5—C6	−1.4 (8)	C8—N1—C10—C11	178.2 (4)
O3—C4—C5—C6	−179.1 (5)	C3—O2—C12—O3	10.9 (6)
C4—C5—C6—C1	1.8 (7)	C4—O3—C12—O2	−10.7 (6)

Hydrogen-bond geometry (Å, º)

D—H···A	D—H	H···A	D···A	D—H···A
N1—H1···Cl1	0.91	2.24	3.149 (4)	174
N1—H2···Cl1^i	0.91	2.30	3.134 (5)	153

Symmetry code: (i) −x, y−1/2, −z+3/2.