Structure, Morphology and Optical Properties of Chiral N-(4-X-phenyl)-N-[1(S)-1-phenylethyl]thiourea, X= Cl, Br, and NO2

Three new enantiopure aryl-thioureas have been synthesized, N-(4-X-phenyl)-N-[1(S)-1-phenylethyl]thiourea, X= Cl, Br, and NO2 (compounds 1-3, respectively). Large single crystals of up to 0.5 cm3 were grown from methanol/ethanol solutions. Molecular structures were derived from X-ray diffraction studies and the crystal morphology was compared to calculations employing the Bravais-Friedel, Donnay-Harker model. Molecular packing was further studied with Hirshfeld surface calculations. Semi-empirical classical model calculations of refractive indices, optical rotation and the electro-optic effect were performed with OPTACT on the basis of experimentally determined refractive indices. Compound 3 (space group P 1 (No. 1)) was estimated to possess a large electro-optic coefficient r333 of approximately 30 pm/V, whereas 1 and 2 (space Group P 21 (No. 4) exhibit much smaller effects.


Introduction
Physical features of crystals are governed by their composition and symmetries (Neumann principle: a physical feature possesses at least the symmetry of the material of consideration). In the absence of inversion symmetry, several non-linear optical properties can exist whereas the presence of OPEN ACCESS inversion symmetry prohibits optical rotation or the linear electro-optic effect for example [1]. Within the context of our goal to predict non-linear optical properties from the data provided in molecular crystal structures obtained via X-ray diffraction we were studying closely related chiral compounds for the impact of small differences in chemical composition on the structure-feature relationship.
Crystals will grow in a way that minimizes as much as possible the sum of dipoles of all molecules inside a single unit cell. The same molecule may arrange in different ways to cancel out a net-dipole. Examples of polymorphs of the same chemical compound usually show a minimized dipole charge of the crystal [2]. Geometrical strategies to minimize a total dipole include dimers of a right handed and a left handed molecule of the same chemical composition, related to each other through inversion symmetry. But inversion symmetry inside the unit cell is not possible in case of enantiopure molecules which causes difficulties for some molecules to grow into larger crystals. A residual dipole charge of the crystal, in addition, may cause individuals to grow in form of twins where the dipoles of the twin components cancel out.
Here we study enantiopure compounds synthesized from the reaction of aryl isothiocyanates with enantiopure α-S-methylbenzylamine (aMBA), The products of aMBA when linked to isothiocynates will also be enantiopure. With some odds against our project to obtain useful crystals for the aforementioned reasons, thioureas are also known for the -H-N-C=S functional group to dimerize via the R 2 2 (8) ring formation, {…H-N-C=S} 2 [3], structuring the packing of the molecule, minimizing the dipoles, and thereby enhancing crystal growth. While aryl isothiocyanates have been studied in biological [4] and medicinal applications [5,6], structural or optical properties, and potential optical applications of chiral thioureas remain mostly unexplored. The point groups of compounds presented here are compatible with the aforementioned features that require the absence of inversion symmetry. Scheme 1. The compounds of this report.

Synthesis and basic properties
Chemical reactions were based on the general procedure described elsewhere [25]. A stochiometric mixture of isothiocyanate and α-(S)-methylbenzylamine was gently stirred in excess of ethanol at 0 °C for 2 hours. The resulting solid product was filtered out and dissolved in 80% ethanol 20% methanol and left for slow evaporation to crystallize. Crystals of compound 3 are very soft and bend easily. Attempts to cut the crystals reveal cleavage on ( ⎯110). Crystals of 1 cleave on (010).
Morphology predictions with BFDH are closer to the experimental observations in crystals with fewer directional bonds. BFDH predictions for compounds 1 and 2 deviate from the indexed shape significantly which indicates strong directional bonds, whereas compound 3 adds H...O 2 -N hydrogen bonds which dilute the directional character of the N..H-S formation and we observe a far better fit of the predicted morphology (cf. Figure 2). All faces of the indexed twin crystal are observed in the BFDH model, predicting (001) and ( ⎯100) faces to be dominant in 3 and (010) as well as (0 ⎯10) largest in 1 and 2. If strong polarity is added to the calculation for 3 by which the central distances of (h k l) are extended and those of ( ⎯h⎯k⎯l ) are shortened as a simple approach to simulate the symmetry independence of Friedel pairs, the model can be made to exhibit exclusively the observed faces. Several individuals of compound 3 were twinned after a twofold twin axis along [110] which eliminates any residual dipole perpendicular to that direction ( Figure 1). Compound 1 grew frequently in the shape of conglomerates. 2 developed into non-twinned specimen.

X-ray diffraction studies
Crystals measuring 0.2 to 0.5 mm 3 were mounted on glass capillaries with paratone oil. Diffraction patterns were collected at 130 K on a Nonius Kappa CCD FR590 single crystal X-ray diffractometer, Mo-radiation. Crystal-to-detector distances were 30 mm and exposure times ranged from 20 to 100 seconds per degree for all sets. The scan width was 1-2 o . The data was integrated and scaled using hkl-SCALEPACK [31]. This program applies a multiplicative correction factor (S) to the observed intensities (I) and has the following form: S is calculated from the scale and the B factor determined for each frame and is then applied to I to give the corrected intensity (I corr ). Solution by direct methods (SIR97 [32,33] produced nearly complete heavy atom phasing models consistent with the proposed structures. The structures were completed by difference Fourier synthesis with SHELXL97 [34,35]. . Scattering factors are from Waasmair and Kirfel [36]. All hydrogen atoms were located using a riding model. All non-hydrogen atoms were refined anisotropically by full-matrix least-squares. Table 1 summarizes the data collection details. ORTEP drawings of the structures for compounds 1-3 were rendered with ORTEP3 for Windows (see below, Figure 2) [37].

Refractive indices
As light passes from one transparent medium to another, it changes speed and the direction bends as a result of refraction. The degree to which this happens depends on the refractive index of the media along the light polarization, and the angle of incidence. Using a micron scale microscope and the "three height method" [38] which is derived from Snell's law of refraction, we were able to determine some of the refractive indices in compound 3 directly in the visible spectral range. The refractive indices are computed approximately from the ratio of distances measured from top to bottom of a crystal plate and from the top to the bottom seen through the (001) crystal section.
The values and directions are approximately 1.61(2) along the a-axis and 1.49(2) perpendicular to that along b* in the projection of the crystal face.
Conoscopic images suggest that these two are to be extended by a third, bigger value. The material was too soft to prepare a prism or a plate thin enough to determine the third value quantitatively. The eigenmode of the largest refractive index is approximately parallel to the direction of the sulfur atom to the nitro group.

Thioamide R 2 2 (8) {…H-N-C=S} 2 dimer
All structures studied here organize via the thioamide {…H-N-C=S} 2 dimer formation (Scheme 2). Hydrogen bonds are most intuitively studied with Hirshfeld surfaces (HSs, Figure 2, Table 2) which are designed to smoothly enclose almost all available space around a molecule [39][40][41][42] Points r on a HSs are found where the sum of all spherically approximated electron densities of a molecule at r is twice the contribution of all atoms in the unit cell at point r. The surface can then be color-coded relative to the distances from points r to the nearest atoms. When choosing a color scheme where red indicates short distances and blue large ones, hydrogen bonds or other close contacts of a molecule to a neighboring one are easily spotted. Figure 2 shows HSs of compounds 1-3 in orientation similar to scheme 2. Clearly visible are two 'red circles' in close vicinity of sulfur and the corresponding hydrogen bonding nitrogen in each of the shown cases.
HSs can be studied further to establish fingerprint plots where the distance d i to a point r on the surface to a close atom inside the surface is plotted against the related distance, d e to an outside atom. Figure 2c presents such fingerprint plots for the compounds 1-3. N-H..S hydrogen bonds are contributing to the two outside spikes, H-C bonds occupy the diagonal and two more spikes related to O..H-C hydrogen bonds are found just inside the R 2 2 (8) formation spikes in 3.

Model calculations
Properties of purely dispersive chiro-optical properties in crystals may be obtained using the dipoledipole interaction theory [43][44][45][46][47][48] (Figure 3), which has been applied with success to ionic crystals using OPTACT [49]. Empirical electronic polarizability volumes of the different elements in a structure need to be modelled until the refractive indices calculated with the dipole-dipole model are close to the experimentally derived ones. Starting values for electronic polarizabilities, derived from refractive indices of many different compounds, are tabulated elsewhere [50]. Appendix 1 sketches the theory underlying the OPTACT program used here to calculate refractive indices and optical rotation.
While progress has been made in the application of quantum mechanics to the calculation of optical rotation in molecules [51,52], in crystals we chose above more reliable theory that embodies long range interactions and accommodates the periodicity of the crystal lattice. To calculate the higher order optical properties, classical polarizability theory was employed whereby an external electric field displaces the nuclei from the centers of the surrounding electron clouds in proportion to the atomic polarizabilities [53]. Here, the external electric field shifts the atomic nuclei of the k th atom by distance x approximated by its electronic polarizability (Figure 4) [54]: where e is the charge of an electron, ε o is the permittivity of free space and E j loc is the local electric field connected with the external field E ext via the effective relative dielectric constant ε' in the direction of the external field, where a spherical depolarization field (Lorentz-depolarization) is assumed: (3) Figure 3. Illustration of the dipole-dipole interaction theory (image from [38], reproduced by permission of The Royal Society of Chemistry). When an initial wave with polarization E passes the first atom positioned at (x,0,0) in a Cartesian reference system (z//k, k is the wave vector), a dipole field is created which induces in a second atom at (0,y,z) a secondary dipole field. This field oscillates in a direction different to E for y, z ≠ 0. The interference of all induced waves with the initial wave (calculated via an Ewald sum) on passing through the crystal leads to a rotation of E when the atoms adopt a chiral arrangement.
Atoms with small electronic polarizabilities are less affected by a static electric field than heavy elements [55]. Although it seems to be inappropriate to use electronic instead of ionic polarizabilities, it is observed that if the sum of the molar polarizabilities is increased only by about a factor of three, the dielectric constant tends already to diverge to infinity, as is easily seen from the Clausius-Mossotti relation. Thus we can expect only a relatively small error by this approximation as long as the dielectric constant remains small.

Figure 4
The shift of the nuclei relative to the electron clouds by an external electric field E (image from [38], reproduced by permission of The Royal Society of Chemistry). The radius of the clouds indicates the size of the polarizability. The external field induces a small change in the value of the electronic polarizability volumes, but this effect is neglected. Furthermore, in the hard-spheres approximation, the electronic clouds of different atoms are not free to move relative to each other. In a classical picture at optical frequencies, the atom's nucleus is the center of the light-wave induced vibration of the shell.
The aforementioned OPTACT program was used to calculate the optical relative dielectric constant ε ij with and without the applied field, and with polarization tensor [56] a ij =ε ij -1 , linear electro-optic effects at constant strain r ijk are calculated from: All tensors were calculated assuming no symmetry at all. Therefore, it was satisfying to discover that the resulting tensors conform to Neumann's law.
When choosing electronic polarizabilities to calculate refractive indices and optical rotation we took the published values [50] first, but varying those of O, N, C and H for fitting the experimental refractive indices in 3. The best set found for elements S, Br, Cl, O, N, C, H was 5.3, 4.4, 3, 2.5, 0.005, 0.09, 0.8 Å 3 , respectively. Table 3 summarizes the extrapolated results [57].

Estimation of optical features
The measured refractive indices and directions in the projection of the (001) crystal face in compound 3 correspond roughly to calculated refractive indices n β and n α , see Table 3. Table 3. Calculation of refractive indices, components of corresponding direction cosines (e o x , e o y , e o z ), optical rotation and normalized static electro-optic effect at constant strain. Realistic r-coefficients may be obtained from calculated r ijk via r' ijp = ⅓(2+ε pq ) r ijq , ε pq static dielectric constants. Tensor components and directions are related to the physical reference system {x, y, z} via convention y // b*, z // c, x = y x z; a, b, c crystallographic axes.   In compound 2, the extinction angle of the largest refractive index in the (010) plane was determined to 11 degrees counter clockwise towards [001] when looking towards the b-axis. The calculated angle was 12.18 o , in excellent agreement with the observation. Figure 5. Molecules in similar orientation, representation surface for optical rotation, morphology of typical specimen, and the electro-optic effect, shown to scale to enhance comparison between compounds.
Only weak similarities are observed between the calculated optical rotation of the three substances (Table 3, Figure 5). The coefficient ρ 22 of optical rotation along the y-axis of the physical reference system is similar of size and negative in 1, 2 and the equivalent direction in 3 mainly normal to the aryl ring planes. Because there is no symmetry restriction in 3, we should expect the OR tensor to be different to that of 1 and 2. Evident is also the effect of the polarizability of the halides in the large differences between 1 and 2 even so their atomic structures are almost identical.
Compound 3 promises to be a photonic material of interest as it exhibits large calculated non-linear optical coefficients, is stable at ambient conditions and crystallizes well. It is evident that the orientation of the molecules and the anisotropy of the optical rotation and the electro-optic effect are coupled. For instance, when the molecules of 1 -3 are aligned in a similar fashion as in Figure 5, the representation surfaces of the electro-optic effect are also similar in shape and orientation. Assuming a moderate relative dielectric coefficient of about 10, compound 3 could reach a r' 333 coefficient of 30 pm/V which compares to other non-linear compounds of interest like LiNBO 3 (r' 333 = -30.9 pm/V) or KH 2 PO 4 (r' 123 =10.3 pm/V) [58].