International Journal of Pharmaceutics

EXperimental and topological determination of the pressure-temperature phase diagram of racemic etifoXine, a pharmaceutical ingredient with anxiolytic properties

M. Barrioa, H. Allouchib, J.-Ll. Tamarita, R. Céolina,c, L. Berthon-Cédilled, I.B. Rietvelde,f,⁎
a Grup de Caracterització de Materials, Departament de Física and Barcelona Research Center in Multiscale Science and Engineering, Universitat Politècnica de Catalunya,
EEBE, Campus Diagonal-Besòs, Av. Eduard Maristany 10-14, 08019 Barcelona, Catalunya, Spain
b EA SIMBA, Synthèse et Isolement de Molécules BioActives, Laboratoire de Chimie Physique, Faculté de Pharmacie, 31, Avenue Monge, 37200 Tours, France
c LETIAM, EA7357, IUT Orsay, Université Paris Sud, rue Noetzlin, 91405 Orsay Cedex, France
d Biocodex, Centre de Recherche, ZAC de Mercière, Chemin d’Armancourt, 60200 Compiègne, France
e Normandie Université, Laboratoire SMS – EA 3233, Université de Rouen, F 76821 Mont Saint Aignan, France
f Faculté de Pharmacie, Université Paris Descartes, USPC, 4 avenue de l’observatoire, 75006 Paris, France
A R T I C L E I N F O

Keywords:
Active pharmaceutical ingredient Thermodynamic properties Crystallographic properties
Solid state Phase diagram Melting Vitreous state EtifoXine

A B S T R A C T

Information about the solid-state properties of etifoXine has been lacking, even if the active pharmaceutical ingredient has been used for its anxiolytic properties for decennia. The crystal structure of the racemic com- pound possesses a monoclinic space group P21/n with cell parameters a = 8.489(2) Å, b = 17.674(2) Å, c = 20.883(3) Å, β = 98.860(10)° and a unit-cell volume of 3095.8(9) Å3 at 293 K. The unit cell contains 8 molecules, while 2 independent molecules with different conformations are present in the asymmetric unit. The density of the crystal is 1.291 g/cm3 and its melting point was found at 362.6 ± 0.3 K with a melting enthalpy of 85.6 ± 3.0 J g−1. Its thermal expansion in the liquid and the solid state and the change in volume on melting and between the vitreous state and the crystalline solid have been studied. The results confirm the tendency of small organic molecules to increase about 11% in volume on melting, while the volume difference between the glass and the crystal at the glass transition temperature is about half this value at 6%. These values can be used in the construction of phase diagrams in the case that the experimental data for a given system is incomplete.

1. Introduction

The active pharmaceutical ingredient (API) etifoXine (Fig. 1) was developed about 60 years ago and the racemic compound is marketed for its anxiolytic properties (Kuch et al., 1968; Choi and Kim, 2015; Besnier and Blin, 2008). Its anxiolytic effect is comparable with that of some benzodiazepine drugs such as lorazepam, but it has less side ef- fects (Choi and Kim, 2015; Girard and Liu, 2008; Nguyen et al., 2006). EtifoXine is prescribed in France and in some other countries, con- sidering this drug as a safe alternative to benzodiazepine drugs. Al- though it has been marketed for years, structural and thermodynamic solid-state studies remain inadequate, even if its unit cell has previously been published in the CSD (Paulus et al., 1976a, 1976b). Considering that generic APIs are now often sourced from producers located in countries outside of the European Union, structural and thermodynamic data are the first necessary step to verify identity and purity. Moreover, within the European Union, the REACH regulation requires the

registration of the physicochemical properties of chemical compounds such as APIs (EC Regulation, 1907/2006). Last, improvements in for- mulations may be driven by modern processing techniques such as 3D printing. Crystal size and shape and melting temperatures of APIs will be of importance to optimize processing parameters for such novel techniques (Tagami et al., 2019; Goyanes et al., 2019). The main pur- pose of the present study was to provide this data and the results are presented hereafter.
Besides the API-specific use of the properties determined in this paper, the data also contribute to statistical information that can be used to study the general solid-state behavior of small organic mole- cules, which make up a large part of the present-day marketed oral drugs. In this light, the thermal expansion of the solid and of the liquid of etifoXine will be compared to that of other APIs. Moreover, the ex- perimental thermal expansion will be used to determine the change in volume of the compound on melting and the difference in volume be- tween the crystalline solid and the vitreous state. This kind of data

Corresponding author at: Normandie Université, Laboratoire SMS – EA 3233, Université de Rouen, F 76821 Mont Saint Aignan, France.
E-mail address: [email protected] (I.B. Rietveld).

https://doi.org/10.1016/j.ijpharm.2019.118812

Received 19 August 2019; Received in revised form 14 October 2019; Accepted 16 October 2019
0378-5173/©2019ElsevierB.V.Allrightsreserved.

Fig. 1. Chemical formula of racemic etifoXine. C17H17ClN2O, M = 300.786 g mol−1. IUPAC name: 6-chloro-N-ethyl-4-methyl-4-phenyl-3,1-ben- zoXazin-2-amine. The stereogenic center is indicated by an asterisk.

appears to be relatively constant among small organic molecules and that knowledge is important because the data can therefore be used to construct phase diagrams of APIs such as volume-against-temperature representative of the Helmholtz free energy and the better-known pressure-against-temperature representative of the Gibbs free energy. Thus, even if very little is known about the specific volume of a given API, because it has not been determined or cannot be determined due to decomposition, it will still be possible to determine its phase behavior and with it the stability hierarchy between different solid phases of an API.
2. Materials and methods
2.1. Etifoxine sample
A sample of racemic etifoXine powder of medicinal grade (D4223) with a purity of 99.94% was kindly provided by Biocodex, France. It was used as such, while for single crystal X-ray diffraction studies sui- table crystals were obtained by slow evaporation from n-hexane solu- tions at room temperature (see Fig. 2).
2.2. Single crystal X-ray diffraction
X-ray diffraction intensities were collected at room temperature up to θ = 65° (θmax) with an Enraf-Nonius CAD4 diffractometer using Cu- Kα radiation (λ = 1.54178 Å) and equipped with a graphite mono- chromator (Enraf-Nonius, 1994). Unit cell determination, data

Fig. 2. Optical microscopy photograph of single crystals of racemic etifoXine, obtained by slow evaporation of a solution in n-hexane at room temperature. The biggest crystals measure 1–3 mm in length.

collection, and data reduction were carried out with the CAD4 EXpress Enraf-Nonius programs package and XCAD4 (Harms and Wocadlo, 1995). All data were corrected for Lorentz polarization effects. The structure was solved by direct methods using the SHELXS-97 package with which most non-hydrogen atoms were located (Sheldrick, 1997, 2008). The remaining atoms were located after successive Fourier synthesis runs. The atomic parameters were refined by a least-squares method on F2 with SHELXL-97 (Sheldrick, 1997, 2008). The co- ordinates of all non-hydrogen atoms were refined with anisotropic thermal parameters. Hydrogen atoms were placed at calculated posi- tions generated according to stereochemistry and refined using a riding model in SHELXL-97.
2.3. High-resolution X-ray powder diffraction
XRPD measurements were performed with a vertically mounted
INEL cylindrical position-sensitive detector (CPS-120) using the Debye–Scherrer geometry and transmission mode. Monochromatic Cu- Kα1 (λ = 1.54056 Å) radiation was selected by means of an asymme- trically focusing incident-beam curved quartz monochromator. Measurements as a function of temperature were carried out using a liquid nitrogen 700 series Cryostream Cooler from OXford Cryosystems. Cubic Na2Ca3Al2F4 was used for external calibration. The PEAKOC application from DIFFRACTINEL software was used for the calibration as well as for the peak position determinations after pseudo-Voigt fit- tings and lattice parameters were refined by way of the least-squares
option of the FullProf suite (Rodriguez-Carvajal, 1993; Rodriguez- Carvajal et al., 2005).
Specimens were introduced in a Lindemann capillary (0.5-mm diameter) and rotate perpendicularly to the X-ray beam during the experiments to improve the averaging of the crystallite orientations. Before each isothermal data acquisition, the specimen was allowed to equilibrate for about 10 min, and each acquisition time was no less than 1 h. The heating rate in-between data collection was 1.33 K min−1. Patterns were recorded on heating in the temperature range from 100 K up to the melting point.
2.4. Differential scanning calorimetry (DSC)
Temperature (onset) and heat of fusion were obtained with a Q100 thermal analyzer from TA Instruments at a 10 K min−1 heating rate. The analyzer was calibrated using the melting point of indium purity 99.999% purchased from Aldrich (Spain) (Tfus = 429.75 K and ΔfusH = 28.45 J g−1). The specimens were weighed using a micro- balance sensitive to 0.01 mg and sealed in aluminum pans.
2.5. Densitometry of the melt as a function of temperature
Liquid density as a function of temperature was measured with a DMA-5000 Density Meter from Anton-Paar. A melted specimen was introduced in the apparatus equilibrated at a temperature above the temperature of fusion. Data were obtained at isothermal steps while slowly cooling in the temperature range from 363 to 323 K. Dry air and bi-distilled water were used as calibration standards in the temperature range. The temperature was controlled at ± 1 mK and measurements were performed when temperature fluctuations were smaller than ± 0.5 mK.
2.6. High-pressure differential thermal analysis (HP-DTA)
HP-DTA measurements have been carried out at a heating rate of 2 K min−1 using an in-house constructed high-pressure differential thermal analyzer similar to Würflinger’s apparatus and operating in the 298–473 K and 0–250 MPa ranges (Würflinger, 1975). To determine the melting temperature as a function of pressure and to ascertain that in- pan volumes were free from residual air, specimens were miXed with an

inert perfluorinated liquid (HT270 Galden®, from Bioblock Scientifics, Illkirch, France, which is guaranteed stable and inert up to 270 °C) as a pressure-transmitting medium, and the miXtures were sealed into cy- lindrical tin pans. To verify that the perfluorinated liquid was chemi- cally inactive and did not interfere with the melting temperature of
etifoXine, preliminary DSC measurements were carried out with a

Table 2
Atom coordinates and equivalent isotropic displacement parameters U(eq) for racemic etifoXine.a

Atom Identificationb
X (×104) y (×104) z (×104) U(eq)c (Å2 × 103)
Cl1_1 6465(1) −3857(1) −2(1) 70(1)

Galden®-etifoXine miXture on a Q100 analyzer of TA instruments

O1_1 10100(2) −992(1) 1252(1) 46(1)

without applied pressure. The DSC curves were equivalent to the ones without the perfluorinated liquid (See Fig. S4).

2.7. Thermogravimetric analysis (TGA)
Thermogravimetric measurements were performed by means of a Q50 system from TA Instruments (New Castle, DE, USA) under nitrogen fluX from room temperature to 473 K. Heating rates of 10 K·min−1 and sample masses of ca. 10 mg were used.

3. Results
3.1. Crystal structure
Racemic etifoXine was found to be monoclinic, space group P21/n (n°14), with unit-cell parameters a = 8.489(2) Å, b = 17.674(2) Å, c = 20.88(3) Å, β = 98.86(1)°, and Vcell = 3095.8(9) Å3 at 293 K con-
taining Z = 8 molecules per unit-cell and Z′ = 2 molecules in the asymmetric unit. The unit cell is in accordance with the unit-cell data published previously (Paulus et al., 1976a, 1976b). Crystal- and struc- ture-refinement data have been compiled in Table 1 and non-H atom coordinates have been listed in Table 2. Other data (fractional co- ordinates for H-atoms, bond lengths, bond angles, anisotropic dis- placement parameters, and torsion angles) can be found in Tables S1–S5 of the Supplementary Information. The atom labelling is shown in Fig. 3. The Cambridge Crystallographic Data Centre deposit (CCDC 1943706) contains the supplementary crystallographic data for this paper. It can be obtained free of charge from the CCDC via www.ccdc. cam.ac.uk/data_request/cif.
Table 1
Crystal- and structure-refinement data for racemic etifoXine.aa esd’s in parentheses.

Identification code/Empirical formula

etifoXine/C17H17ClN2O

b The last digit of the atom numbering (1 or 2) corresponds to either of

Formula weight 300.786 g mol−1
Temperature 293(2) K
Wavelength 1.5418 Å
Crystal system (Space group) Monoclinic (P 21/n)
Unit-cell dimensions a = 8.489(2) Å, b = 17.674(2) Å,
c = 20.883(3) Å, β = 98.860(10)°
Cell volume 3095.8(9) Å3

molecules 1 or 2, respectively.
c U(eq) is defined as one third of the trace of the orthogonalized Uij tensor.

3.2. The specific volume of racemic etifoxine as a function of temperature
An experimental and a calculated X-ray powder diffraction pattern are shown in Fig. S1 of the supplementary materials for comparison. It
can be seen that the two patterns are virtually the same, although some
differences in peak intensities indicate the presence of preferred or-

Density (calculated) 1.291 g/cm
Absorption coefficient 2.178 mm−1
F(0 0 0) 1264
Crystal size 0.550 × 0.200 × 0.100 mm3
Theta range for data collection 3.292 to 64.944°.
Limiting indices 0 ≤ h ≤ 9, 0 ≤ k ≤ 20, −24 ≤ l ≤ 24 Reflections collected/independent 5239/5239 [R(int) = 0.0162] Completeness to theta = 64.944° 99.9%
Absorption correction Psi-scan
Max. and min. transmission 0.9971 and 0.8015

ientation. High resolution X-ray diffraction patterns have been collected from 100 to 340 K to determine the lattice parameters of the unit cell as a function of temperature by least-squares refinement. The results have been compiled in Table S6 in the supplementary materials. The specific volumes (v/cm3 g−1) have been fitted to the following quadratic equation as a function of temperature (T/K):
vs = 0.74375(73) + 4.38(71) × 10 5T + 2.24(16) × 10 7T2 (r2 = 0.9995)

Refinement method
Full-matriX least-squares on F2(1)

Data/restraints/parameters 5239/0/382
Goodness-of-fit on F2 1.131
Final R indices [I > 2sigma(I)] R1 = 0.0488, wR2 = 0.1272 R indices (all data) R1 = 0.0589, wR2 = 0.1347
EXtinction coefficient 0.0088(4)
Largest diff. peak and hole 0.3304 and −0.348 e.Å−3

The specific volume of molten etifoXine from experimental density measurements have been reported in Table S7 of the supplementary materials and have been fitted as a function of the temperature to the following equation:observed in the HP-DTA experiments. The onset temperature increases with the pressure presented in Fig. 5a and the corresponding pressur- e–temperature phase diagram is depicted in Fig. 5b. The onset tem- peratures (T/K) of the melting peaks and the corresponding pressures (P/MPa) listed in Table S9 in the supplementary materials have been fitted to the quadratic equation:
Fig. 3. Molecular structure of R-etifoXine with atom identification. The hy- drogen atoms are labelled with the same numbers as the atoms to which they are linked. When more than one H is linked to a C-atom, these H-atoms are labelled with supplementary letters A, B and C. The presented molecule is 2(R) in the racemate and the same labelling has been applied to molecule 1(R). The stereogenic center is atom C1.

3.3. Calorimetric behavior
Specimens of the as-received sample were subjected to differential scanning calorimetry in closed pans. Heating from 213 K led to a single endothermic effect ascribed to a melting transition with a mean onset at
362.6 ± 0.3 K. The mean value of the melting enthalpy was found to be 85.6 ± 3.0 J g−1 (25.75 ± 0.90 kJ mol−1). On reheating after quick or slow cooling of the melt to 213 K, a midpoint glass transition was observed at 297.2 ± 0.8 K. Data from 16 experiments have been

The centrosymmetric unit-cell of racemic etifoXine contains two independent molecules numbered 1 and 2 whose 3D projection is shown in Fig. 6. The two independent molecules mainly differ in the orientations of the phenyl rings φ1 with respect to the heterocycles (see Fig. 7): the values of the dihedral angles C2-C1-C9-C14 are 53.95° and
−37.88° in molecules 1 and 2, respectively. Weak intramolecular in- teractions prevent the rotation of the two NHeCH2-CH3 chains around the N2-C4 bonds: N2-H2_1···O1_1 is 2.250 Å, and N2eH2_2···O1_2 equals
2.239 Å, as well as the rotation of the two phenyl rings φ1 around the C1eC9 bonds: C10eH10_1···O1_1 is 2.404 Å and C10eH10_2···O1_2 equals 2.454 Å. The molecules are held together by N1···HeN2 hy- drogen bonds that form infinite chains parallel to the a axis and to the ac plane (see supplementary materials Fig. S2). These interactions, in addition to the weak CeH···O interactions have been compiled in Table S10 in the supplementary materials.
4.2. Thermal expansion of the solid and the liquid
Although the thermal expansion of the solid is best fitted by a quadratic equation (Eq. (1)), it can be approXimated by a linear ex- pression:

compiled in Table S8 of the supplementary materials. Fig. 4 contains
vs (T ) = 0.7343(11) + 0.0001426(46)T(r2 = 0.9887).(4)
typical DSC curves, one for the crystalline powder and one for the glass. No decomposition has been observed during the calorimetric mea- surements; thermogravimetric analysis demonstrates that weight loss of the sample only sets in at about 40°above the melting point (see Fig. S3).

3.4. Thermal behavior under pressure
A single endothermic peak of the fusion of racemic etifoXine was

Fig. 4. Differential scanning calorimetry curves of racemic monoclinic etifoXine obtained at 10 K min−1. (a) First heating of the crystalline commercial product, Tm: melting temperature, (b) second heating after quenching the melt to 213 K in which a glass transition (Tg: glass transition temperature) can be observed.

An average expansivity of the solid αv,s of 1.94 × 10−4 K−1 can be obtained by rewriting Eq. (4) in the form of vs(T) = v0 × (1 + αv,s T). This solid expansivity is very close to the mean value for molecular organic solids of 2.21 × 10−4 K−1 that has been reported previously (Gavezzotti, 2013; Céolin and Rietveld, 2015; Rietveld and Céolin, 2015).
The expansivity of liquid etifoXine is found to be 0.97 × 10−3 K−1
obtained from Eq. (2). This is smaller than the mean value of
1.20 ± 0.25 × 10−3 K−1 found previously for the expansivity of molten active pharmaceutical ingredients (Céolin and Rietveld, 2015; Rietveld and Céolin, 2015); however, it still falls within the range of experimental uncertainty of 0.25 × 10−3 K−1.
4.3. Volume change on melting and volume difference at the glass transition
It has been previously reported that, in the case of molecular or- ganic solids, the change in the specific volume on melting is about 11 ± 3% (Céolin and Rietveld, 2015; Rietveld and Céolin, 2015; Barrio et al., 2017), i.e. slightly smaller than the 12% value found previously by Goodman et al. (2004) Calculating the specific volumes of the solid and of the liquid using Eqs. (1) and (2) at Tfus = 362.6 K, it is found that the volume change on melting is in the present case 0.0709 cm3 g−1, leading to the value of 1.090 ± 0.010 for the vliq/vs ratio at Tfus, in fair agreement with the above-mentioned value.
At the glass transition, the metastable undercooled melt rigidifies,
i.e. becomes as viscous as the crystalline solid. The glass transition temperature has been found at 297.2 K (10 K min−1). The specific vo- lume of the melt at Tg can be calculated with Eq. (2) and is found to be 0.8196 cm3 g−1, while that of the crystalline solid at this temperature is 0.7766 cm3 g−1 from Eq. (1). Thus, the ratio vglass/vs at Tg equals
Fig. 5. (a) Melting peaks of racemic etifoXine at various pressures and (b) the pressure-temperature phase diagram.

Fig. 6. Monoclinic unit-cell of racemic etifoXine with the two independent molecules 1 (red) and 2 (blue). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 7. Superposition showing the difference in conformation of molecules 1 (red) and 2 (blue). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

1.055 ± 0.010, i.e. the volume difference between the crystalline and the amorphous phase is approXimately half that of the difference at the melting point. With this information, one can draw the temperature – (specific) volume diagram, shown in Fig. 8, which is a partial projection

Fig. 8. Specific volumes of solid (S) and liquid (L) etifoXine as a function of temperature with extrapolation to the Kauzmann temperature Tk = 204.6 K using Eqs. (1) and (2).

of F(v,T), the Helmholtz function of the system, on a T-v plane, one of the four thermodynamically equivalent energy functions of the system. vglass/vs ratios have been found for other cases, which have been compiled in Table 3.
4.4. Slope of the solid-liquid equilibrium in the pressure-temperature diagram
Using Eq. (3), it is found that the slope of the melting equilibrium of racemic etifoXine at P = 0 MPa equals 3.29 ± 0.06 MPa K−1 (Fig. 5b). This slope can also be determined through the Clapeyron equation dP/ dT = ΔfusH/(Tfus Δv) in which ΔfusH is the melting enthalpy (85.60 J g−1), Tfus the melting temperature (362.6 K), and Δv the vo- lume change on melting (0.07094 cm3 g−1). With these experimental values, the value of dP/dT is found to be 3.33 ± 0.12 MPa K−1, i.e. extremely close to the direct experimental value of 3.29 MPa K−1, in- dicating that at least in the case of etifoXine, the Clapeyron equation is a good alternative for direct measurements to obtain the response of the system to pressure.
4.5. Isobaric thermal expansion tensor for the solid
The anisotropy of the intermolecular interactions can be accounted for by the isobaric thermal expansion tensor, which shows the influence of these interactions in the crystal on heating (Salud et al., 1998). The eigenvalues and the eigenvectors of the tensor specify the strongest and the weakest directions for the intermolecular interactions, commonly

Table 3
EXperimental volume changes for molecular compounds and pharmaceuticals at their melting point and glass transition temperature.

compound vL/vI at Tfus Tfus/K vL/vI at Tg Tg /K Tg/Tfus Reference
paracetamol 1.15 442.3 1.07 298 0.673 (Espeau et al., 2005)
prilocaine 1.13 311.5 1.07 218 0.700 (Rietveld et al., 2013)
rimonabant 1.11 429.2 1.04 350 0.820 (Perrin et al., 2013)
biclotymol 1.13 400.5 1.05 294 0.734 (Ceolin et al., 2008)
ternidazole 1.11 333.0 1.07 235 0.706 (Mahé et al., 2011)
morniflumate 1.12 348.1 1.06 249 0.715 (Barrio et al., 2017)
etifoXine 1.09 362.4 1.055 297.2 0.820 This work
Mean 1.12(2) 1.06(1)

independent molecules (Z′ = 2) with different conformations, which
implies a possibility for etifoXine to exhibit conformational poly-

Fig. 9. αi eigenvalues of the thermal-expansion tensor as a function of tem- perature for the P21/n monoclinic phase of etifoXine with the α2 eigenvector parallel to the two-fold crystallographic axis b, as evidenced in the tensor re- presentation at 140 K in the inset (full length scale of the αi eigenvectors cor- responds to 10−4 K−1).

referred to as “hard” and “soft” directions, respectively (Salud et al., 1998).
The lattice parameters as a function of temperature have been de- termined (see supplementary materials, Table S6) and fitted using the least-squares method. The coefficients of the related polynomial equa- tions have been compiled in Table S11 in the supplementary materials, together with the reliability factor, defined as R = Σ(yoi − yci)2/y 2, where yoi and yci are the measured and calculated lattice constants, respectively.
The procedure to determine the eigenvalues and eigenvectors of the second-rank αij thermal expansion tensor has been detailed previously (Salud et al., 1998). For monoclinic symmetry, the tensor is fully de- fined by three eigenvalues αi (i = 1,2,3) and by an angle between the direction of one of the eigenvectors (α1 in the present case) and the crystallographic axis a, the α2 direction being parallel to the two-fold axis b. Fig. 9 depicts the variation of the eigenvalues of the thermal- expansion tensor as a function of temperature as well as the 3D-tensor within the orthogonal frame of eigenvectors together with the crystal- lographic axes at 140 K. The tensor is rather anisotropic and this ani- sotropy increases with temperature. The thermal expansion within the ab plane is the highest (“soft” plane) while the lowest deformation (α3
direction) corresponding to a “hard direction” is found within the ac plane as a consequence of the intermolecular hydrogen bonding (N2eH···N1).

5. Concluding remarks
The crystal structure of racemic etifoXine has been solved at room temperature and thermal expansion studies in the range from 100 K to 340 K as well as differential scanning calorimetry did not reveal any crystalline polymorphism. However the crystal structure contains two

morphism as previously defined by Bernstein and Hagler (1978) How conformational polymorphism can cause problems for the pharmaceu- tical industry can be read in the paper “Disappearing polymorphs re- visited” by Bucar et al. in which different conformations are at the basis of a court case (ranitidine hydrochloride), a formulation disaster (ri- tonavir) and another case in which a more stable form crystallizes out in a marketed drug formulation (rotigotine) to name a few (Céolin and Rietveld, 2015; Bucar et al., 2015; Bauer et al., 2001; Chaudhuri, 2008; Rietveld and Ceolin, 2015).
It has been shown that the thermal expansion of etifoXine is very similar to that of small organic molecules in general, for the solid as well as for the liquid. As far as the volume change on melting is con- cerned, it has been found that, again in the present case, it agrees with the average value of 11% within an error of less than 3%. In addition, the volume difference between the crystal and the metastable liquid at the temperature of the glass transition is found to be about 6%, in close accordance with previously found data. Further case studies are still needed before one can conclude that these volume changes for mole- cular organic compounds are more or less independent of the chemical composition and thus ‘statistically’ and possibly in a deeper sense uni- versal.
Funding
This work was supported by the MINECO, Project No. FIS2017- 82625-P, and AGAUR, DGU Project No. 2017SGR-42.
Declaration of Competing Interest
The authors declare the following financial interests/personal re- lationships which may be considered as potential competing interests:
L. Berthon-Cédille compound 3k is employee of Biocodex. The other researchers de- clare no conflict of interest.
Appendix A. Supplementary material
Supplementary data to this article can be found online at https:// doi.org/10.1016/j.ijpharm.2019.118812.

References
Barrio, M., Tamarit, J.L., Ceolin, R., Robert, B., Guechot, C., Teulon, J.M., Rietveld, I.B., 2017. EXperimental and topological determination of the pressure temperature phase
diagram of morniflumate, a pharmaceutical ingredient with anti-inflammatory properties. J. Chem. Thermodyn. 112, 308–313.
Bauer, J., Spanton, S., Henry, R., Quick, J., Dziki, W., Porter, W., Morris, J., 2001.
Ritonavir: an extraordinary example of conformational polymorphism. Pharm. Res. 18 (6), 859–866.
Bernstein, J., Hagler, A.T., 1978. Conformational polymorphism. The influence of crystal- structure on molecular-conformation. J. Am. Chem. Soc. 100 (3), 673–681.
Besnier, N., Blin, O., 2008. ÉtifoXine: études cliniques récentes. L’encéphale 34 (suppl 1), S9–S14.
Bucar, D.K., Lancaster, R.W., Bernstein, J., 2015. Disappearing polymorphs revisited.
Angew. Chem. Int. Ed. Engl. 54 (24), 6972–6993.
Céolin, R., Rietveld, I.B., 2015. The topological pressure-temperature phase diagram of ritonavir, an extraordinary case of crystalline dimorphism. Ann. Pharm. Fr. 73 (1),
22–30.
Ceolin, R., Tamarit, J.L., Barrio, M., Lopez, D.O., Nicolai, B., Veglio, N., Perrin, M.A.,
Espeau, P., 2008. Overall monotropic behavior of a metastable phase of biclotymol, 2,2′-methylenebis(4-chloro-3-methyl-isopropylphenol), inferred from experimental
and topological construction of the related P-T state diagram. J. Pharm. Sci. 97 (9), 3927–3941.
Chaudhuri, K.R., 2008. Crystallisation within transdermal rotigotine patch: is there cause
for concern? EXpert. Opin. Drug Del. 5 (11), 1169–1171.
Choi, Y.M., Kim, K.H., 2015. EtifoXine for pain patients with anxiety. Korean J. Pain 28 (1), 4–10.
Enraf-Nonius. 1994. CAD4 EXpress Software, Enraf Nonius: Delft, The Netherlands.
Espeau, P., Céolin, R., Tamarit, J.L., Perrin, M.A., Gauchi, J.P., Leveiller, F., 2005.
Polymorphism of paracetamol: Relative stabilities of the monoclinic and orthor- hombic phases inferred from topological pressure-temperature and temperature-vo-
lume phase diagrams. J. Pharm. Sci. 94 (3), 524–539.
Gavezzotti, A., 2013. Molecular Aggregation. Structure Analysis and Molecular Simulation of Crystals and Liquids. OXford University Press, OXford, UK, pp. 448.
Girard, C., Liu, S., Cadepond, F., Adams, D., LacroiX, C., Verleye, M., Gillardin, J.M.,
Baulieu, E.-E., Schumacher, M., Schweizer-Groyer, G., 2008. EtifoXine improves
peripheral nerve regeneration and functional recovery. PNAS 105 (51), 20505–20510.
Goodman, B.T., Wilding, W.V., Oscarson, J.L., Rowley, R.L., 2004. A note on the re- lationship between organic solid density and liquid density at the triple point. J.
Chem. Eng. Data 49 (6), 1512–1514.
Goyanes, A., Allahham, N., Trenfield, S.J., Stoyanov, E., Gaisford, S., Basit, A.W., 2019. Direct powder extrusion 3D printing: FABRICATION of drug products using a novel
single-step process. Int. J. Pharm. 567, 118471.
Harms, K., Wocadlo, S., 1995. XCAD-CAD4 Data Reduction. University of Marburg, Marburg.
Kuch, H., Seidl, G., Hoffmann, I., Soden, B. 1968. 3,1-benzothiazines and 3,1-benzoX-
azines.
Mahé, N., Perrin, M., Barrio, M., Nicolaï, B., Rietveld, I., Tamarit, J., Céolin, R., 2011. Solid-state studies of the triclinic (Z ‘=2) antiprotozoal drug ternidazole. J. Pharm.
Sci. 100 (6), 2258–2266.
Nguyen, N., Fakra, E., Pradel, V., Jouve, E., Alquier, C., Le Guern, M.-E., Micallef, J., Blin, O., 2006. Efficacy of etifoXine compared to lorazepam monotherapy in the treatment of patients with adjustment disorders with anxiety: a double-blind controlled study in

general practice. Hum. Psychopharmacol. Clin. EXp. 21, 139–149.
Paulus, E.F., Bartl, H., Ruggeberg, K., 1976a. CSD entry ECMPBX. Eur. Cryst. Meeting 442.
Paulus, E.F., Bartl, H., Ruggeberg, K., 1976b,. CSD Entry ECMPBX01. Eur. Cryst. Meeting
442.
Perrin, M.A., Bauer, M., Barrio, M., Tamarit, J.L., Ceolin, R., Rietveld, I.B., 2013.
Rimonabant dimorphism and its pressure-temperature phase diagram: a delicate case of overall monotropic behavior. J. Pharm. Sci. 102 (7), 2311–2321.
Regulation (EC) N° 1907/2006 of the European Parliament and of the Council of 18
December 2006, concerning the Registration, Evaluation, Authorization and Restriction of Chemicals (REACH). 2006.
Rietveld, I.B., Ceolin, R., 2015. Rotigotine: unexpected polymorphism with predictable
overall monotropic behavior. J. Pharm. Sci. 104 (12), 4117–4122.
Rietveld, I.B., Céolin, R., 2015. Phenomenology of crystalline polymorphism: overal
monotropic behavior of the cardiotonic agent FK664 forms A and B. J. Therm. Anal. Calorim. 120 (2), 1079–1087.
Rietveld, I.B., Perrin, M.-A., Toscani, S., Barrio, M., Nicolaï, B., Tamarit, J.-L., Céolin, R., 2013. Liquid-liquid miscibility gaps in drug-water binary systems: crystal structure
and thermodynamic properties of prilocaine and the temperature-composition phase
diagram of the prilocaine-water system. Mol. Pharmaceut. 10 (4), 1332–1339.
Rodriguez-Carvajal, J., 1993. Recent advances in magnetic structure determination by neutron powder diffraction. Physica B 192, 55–69.
Rodriguez-Carvajal, J., Roisnel, T., Gonzales-Platas, J. 2005. Full-Prof suite version 2005.
Laboratoire Léon Brillouin, CEA-CNRS, CEN Saclay, France.
Salud, J., Barrio, M., Lopez, D.O., Tamarit, J.L., Alcobe, X., 1998. Anisotropy of inter- molecular interactions from the study of the thermal-expansion tensor. J. Appl.
Crystallogr. 31, 748–757.
Sheldrick, G.M., 1997. SHELXL97: Program for Crystal Structure Refinement. Universität Göttingen, Göttingen, Germany.
Sheldrick, G.M., 2008. A short history of SHELX. Acta Crystallogr. A 64, 112–122.
Tagami, T., Hayashi, N., Sakai, N., Ozeki, T., 2019. 3D printing of unique water-soluble
polymer-based suppository shell for controlled drug release. Int. J. Pharm. 568, 118494.
Würflinger, A., 1975. Differential thermal-analysis under high-pressure IV. Low-tem- perature DTA of solid-solid and solid-liquid transitions of several hydrocarbons up to 3 kbar. Ber. Bunsen-Ges. Phys. Chem. 79 (12), 1195–1201.