Monotropic Transition Mechanism of m-Hydroxybenzoic Acid

Article pubs.acs.org/crystal

Monotropic Transition Mechanism of m‑Hydroxybenzoic Acid Investigated by Temperature-Resolved Second Harmonic Generation Published as part of a Crystal Growth and Design virtual special issue of selected papers presented at the 10th International Workshop on the Crystal Growth of Organic Materials (CGOM10) Simon Clevers, Florent Simon, Morgane Sanselme, Valerie Dupray,* and Gerard Coquerel PRES Normandie, Crystallogenesis Unit, SMS, EA 3233 Université de Rouen, F-76821 Mont-Saint-Aignan Cedex, France S Supporting Information *

ABSTRACT: Temperature-resolved second harmonic generation (TR-SHG) and SHG microscopy were used to study under normal pressure the solid−solid transition mechanism occurring between the two monotropically related polymorphic forms (metastable Pna21 and stable P21/n) of 3-hydroxybenzoic acid (MHBA). The activation energy Ea (as a measure of the barrier energy) of the irreversible transition was determined via isothermal TR-SHG (137−144 kJ·mol−1). It fits well with that determined from differential scanning calorimetry (139 kJ·mol−1). Regarding the two crystal structures, optical microscopy observations, and kinetics parameters from TRSHG, a destructive/reconstructive mechanism is proposed for the solid−solid transition. The present study clearly demonstrates that TR-SHG is a relevant and accurate technique for monitoring solid− solid phase transitions.

■

INTRODUCTION In material science and, in particular, in pharmaceuticals, molecules crystallizing in different crystal structures are frequently encountered. Pharmaceutical compounds can exist in various solid phases (polymorphs, hydrates, solvates, cocrystals, host−guest associations, salts and hybrids of all sorts) and are often administered in crystalline powder forms.1 Controlling polymorphism and assessing the structural purity2 are crucial for the production of polymorphs as they have different physicochemical and physical characteristics (solubility, hardness, compressibility, density, melting point, dissolution rate, etc.). When metastable polymorphs are involved, a spontaneous transition toward the most stable polymorph can occur. Many factors can induce or affect polymorphic transitions such as temperature, humidity, mechanical stress, compression, hydrostatic pressure, etc. Crystal packing and hydrogen bonding may also directly influence solid−solid transitions if structural filiations exist between the polymorphic forms. The presence of crystal defects and mosaicity3 may also change the kinetics of the solid−solid transition process. As a consequence, understanding the mechanism and kinetics of solid−solid transitions together with the establishment of the relative stability of polymorphs is nowadays a necessity. For this purpose, differential scanning calorimetry (DSC) and temperature resolved powder X-ray diffraction (TR-XRPD) (sometimes combined with FT-IR, Raman, optical microscopy, or SEM) are commonly used.4 However, because of their detection thresholds, these techniques present difficulties to © 2013 American Chemical Society

detect structural changes at a level below 1% wt. TR-XRPD certainly constitutes the most direct way to identify the presence of a polymorph and to monitor phase transitions but suffers limitations as long as the mass fraction remains below 5% in the physical mixture and if the kinetics of the phenomenon involved is too fast. DSC is also widely used to reveal phase transitions in solids with a limit of detection reported at 1% wt.5 In a particularly favorable case, a detection limit down to 0.1% has been reported.6 Among the nonlinear optics phenomena, second harmonic generation (SHG) can be a solution to overcome these limits of detection with a threshold at the ppm level.7 SHG occurs in crystal with non-centrosymmetric space groups. It was formerly used as a rapid, reliable, and very sensitive test of noncentrosymmetry for crystalline compounds with applications for structure assignments.8 Recently, it was proposed as an efficient technique for prescreening of conglomerate forming systems.9 SHG is all the more interesting because further data can be obtained by performing SHG measurements versus temperature. Temperature-resolved SHG (TR-SHG) has already been employed in inorganic and seldom in organic compounds and organic−inorganic hybrids to monitor solid− solid transitions between enantiotropically related compounds and order−disorder phase transitions.10−12 SHG was also used to Received: May 7, 2013 Revised: July 4, 2013 Published: July 9, 2013 3697

dx.doi.org/10.1021/cg400712s | Cryst. Growth Des. 2013, 13, 3697−3704

Crystal Growth & Design

Article

Figure 1. Experimental temperature-resolved second harmonic generation apparatus constituted of Nd:YAG Q-switched laser operating at 1064 nm. integrator allowed an average spectrum (spectral range 490−590 nm) with a resolution of 0.1 nm to be recorded over 5 s (50 pulses). To avoid problems related to the sublimation of the samples, the heating-stage was opened but only during the SHG measurements so it had no significant influence on the temperature regulation. According to the Kurtz and Perry SHG powder method,8 SHG signal intensities were compared to the signal of a reference compound (α-quartz powder 45 μm average size). Single Crystal X-ray Diffraction (SC-XRD). The crystal structures were determined from single crystal diffraction on a SMART APEX diffractometer (with Mo Kα radiation: λ = 0.71073 Å). The cell parameters and the orientation matrix of the crystal were preliminarily determined by using SMART software. Data integration and global cell refinement were performed with SAINT Software. Intensities were corrected for Lorentz polarization and decay and absorption effects (SAINT and SADABS Softwares) and reduced to FO. The structures were solved by direct methods (SHELX-XS), and anisotropic displacement parameters were refined for all non-hydrogen atoms using SHELX-XL available within the WinGX package. All hydrogen atoms were located by Fourier-difference synthesis and fixed geometrically according to their environment with a predefined isotropic thermal factor. Differential Scanning Calorimetry (DSC). Thermal analyses of the solids were conducted on DSC 204 F1 Netzsch equipped with an Intracooler. Solid samples (mass of ca. 15 mg with maximum deviation of 0.05 mg) were placed in a 25 μL close aluminum crucible. The atmosphere of the analyses was regulated by helium flux (40 mL·min−1), and heat runs were conducted at different constant heating rates. The data treatment was performed with Netzsch-TA Proteus Software v 4.8.4. Hot-Stage Microscopy. Hot-stage Microscopy observations were performed with the Linkam THMS-600 hot-stage and a polarized light Nikon Eclipse LV100 microscope equipped with a video system. Set-up allowed isothermal or non-isothermal treatments of MHBA single crystals. Multiphoton Microscopy. A laser scanning microscope LSM 710 NLO Zeiss (Iena, Germany) was used. Excitation was provided by a CHAMELEON femtosecond titanium-sapphire laser (Coherent, Santa Clara, USA) set at 860 nm. Samples were put in a glass bottom box and were imaged with a 63×, 1.4 NA oil (or 20x, 0.8 NA) objective lens.

follow in situ the crystallization of a noncentrosymmetric phase from a supersaturated solution,13−15 to quantify crystalline phases arising from amorphous materials,4 or to study the relaxation of polymers or glasses above glass transitions.16,17 In this study, we demonstrate that TR-SHG can be a valuable tool to study solid−solid transitions provided one of these polymorphs crystallizes in a non-centrosymmetric space group. This is exemplified with the solid−solid transition between metastable and stable m-hydroxybenzoic acid (hereafter MHBA) polymorphs.

■

MATERIALS AND METHODS

3-Hydroxybenzoic acid (MHBA) was purchased from Acros Organic (CAS registry number 99-06-9) with a chemical purity better than 99% (produced in 2011). Preparation of Pure Polymorphic Forms. StabMHBA: Stable MHBA was obtained from commercial MHBA (ComMHBA) annealed at 145 °C for 1 h. MetMHBA: Metastable MHBA was obtained by evaporating, under normal pressure and temperature, an initially undersaturated solution of ComMHBA in acetone. MHBA large crystals of good optical quality (several micrometers size) were obtained by cooling of a saturated solution of ComMHBA in acetone from 45 to 20 °C with a rate of 1 K·h−1. The solution was kept at 20 °C for 24 h before drying. Temperature-Resolved Second Harmonic Generation (TRSHG). Figure 1 shows the experimental setup used for the TR-SHG measurements. A Nd:YAG Q-switched laser (Quantel) operating at 1.06 μm was used to deliver up to 360 mJ pulses of 5 ns duration with a repetition rate of 10 Hz. An energy adjustment device made up of two polarizers (P) and a half-wave plate (λ/2) allowed the incident energy to be varied from 0 to ca. 200 mJ per pulse. A RG1000 filter was used after the energy adjustment device to remove light from laser flash lamps. The samples (a few milligrams of powder in a crucible) were placed in a computer-controlled heating−cooling stage (Linkam THMS-600) and were irradiated with the laser beam (4 mm in diameter). The signal generated by the sample (diffused green light 532 nm) was collected into an optical fiber (500 μm of core diameter) and directed onto the entrance slit of a spectrometer (Ocean Optics). A boxcar 3698



Article

Emitted signal of SHG was collected with a bandpass filter (420−440 nm). Z stacks images were acquired with 0.8 μm Z step. Curve Fitting. Data were fitted to specific expressions by application of the Levenberg−Marquardt (least-squares method) algorithm to minimize χ2 using the program Gnuplot.

■

RESULTS AND DISCUSSION Measurement of Activation Energy (Ea) by Means of TR-SHG. MHBA (Figure 2) is an intermediate in the production of pharmaceuticals, plasticizer, germicides, and preservatives.18

Figure 3. Isothermal SHG measurements on pure MetMHBA for 85 °C (circle), 115 °C (square), 130 °C (triangle), and 145 °C (plus symbol). Dash-lines are fit curves obtained by the least-squares method.

transition). It can be noticed that for the lower temperatures (85 and 115 °C) the curves are stepped. This feature will be discussed later. The evolution of the volume fraction x of the stable form can be described by the Avrami equation:21−24 Figure 2. ORTEP view of MHBA molecule with labeled atoms. Both polymorphs comprise the same conformer.

n

(2) x = 1 − e−k(T )t with t the time of reaction, n the Avrami exponent, and k(T) the Avrami constant. Avrami constant k(T) depends on temperature and can be expressed with an Arrhenius law as follows:

Two unsolvated polymorphic forms of MHBA (a metastable form and a stable form) are reported in the literature.19 The metastable form of MHBA crystallizes as an orthorhombic structure in the non-centrosymmetric space group Pna21. Thus, it exhibits SHG activity. The stable form crystallizes as a monoclinic structure in the centrosymmetric space group P21/n (therefore SHG inactive). Under normal pressure, a known irreversible solid−solid transition (from the metastable form toward the stable form) occurs associated with an exothermic event.20 The two polymorphic forms are in a monotropic relationship (the non-centrosymmetric form is metastable at all temperatures). In a previous study,7 we demonstrated that SHG can be used with relevancy to detect a small amount of metastable MHBA in a matrix of stable MHBA. A common approach to determine the activation energy, Ea, is to track the conversion of the metastable form (toward the stable form) by means of isothermal experiments at various temperatures. Isothermal transformation of MetMHBA was then studied by monitoring the time dependence of the SHG signal for four different temperatures: 85, 115, 130, and 145 °C. The evolution of the transformed fraction x versus time is plotted in Figure 3 (x is defined as the volume fraction of stable form into the sample). Assuming that the SHG intensity depends linearly on x,7 it is calculated as follows: x=1−

k(T ) = A e−Ea / RT

Constant A is a pre-exponential factor, Ea is the activation energy of the transition, R is the universal gas constant, and T is the temperature. A standard method to calculate Ea is to use the natural logarithms of eq 2, which leads to the following equation: ln[− ln(1 − x)] = ln[k(T )] + n ln(t )

(4)

A linear regression is then applied to the plot of ln[− ln(1 − x)] versus ln(t) to obtain n as the slope and ln[k(T)] as the intercept. A natural logarithm of ln[k(T)] leads to ln[k(T )] = ln(A) −

Ea RT

(5)

Finally, Ea is obtained from the slope of the plot of ln[k(T)] versus 1000/T (Figure 4). Usually, the extreme experimental values are not taken into account (because of logarithm divergence). An alternative to this method is to proceed to a direct fit of the experimental data using the least-squares method, with the Avrami equation set in the program and all parameters let free for the fit. In this last case, the whole range of data is used. These two different approaches lead to the results summarized in Table 1. The classical method leads to an Ea equal to 137 kJ·mol−1 with a regression coefficient R2 of 0.963. With the direct fit method, activation energy equals 144 kJ·mol−1 with a regression coefficient R2 of 0.962. Results from the direct fit of experimental data are in good agreement with results obtained by linear regression from the plot of ln[−ln(1 − x)] versus ln(t). The deviation between the two methods is more pronounced at 85 and 115 °C than at 130

ISHG(T ) 0 ISHG

(3)

(1)

I0SHG

is the initial SHG intensity (at room temperature) and ISHG(T) is the SHG intensity measured at temperature T. Isothermal curves depicted in Figure 3 exhibit a classical “S” shape for transformation profile (for each isothermal temperature, the origin corresponds to the beginning of the solid−solid 3699



Article

for the stable MHBA. This value fits with the enthalpy reported by Nordstrom et al (35.9 kJ·mol−1)20 and Perlovich et al 34.9 kJ·mol−1.25 Note that a lower value of 26.2 kJ·mol−1 was also reported.26 In addition to the endothermic peak corresponding to the fusion of the stable MHBA, all thermograms (except for 1 K/min heating rate) exhibit a weak exothermic event at ca. 160 °C (see enlargement of the DSC traces in Figure 5). It corresponds to the solid−solid transition of the metastable (Pna21) form toward the stable form (P21/n). The exothermic transformation heat presents an average value of 0.57 kJ·mol−1, which is in good agreement with the value of 0.51 kJ·mol−1 reported by Nordstrom et al.20 The peak temperature Tp of the solid−solid transition increases with the heating rate. The dependence of the peak temperature Tp on the heating rate φ can be used to determine the Ea of the solid−solid transition on the basis of the Kissinger relation written as

Figure 4. ln(k(T)) versus 1000/T plots for isothermal transformation of MetMHBA to StabMHBA. Ea from the slope is 144 kJ·mol−1 for the direct method and 137 kJ·mol−1 for the classical method.

⎛ AR ⎞ E ln(φ) = − a + ln⎜ ⎟ 2 RTp Tp ⎝ Ea ⎠

and 145 °C. This is certainly due to the steps observed on the isothermal curves, which limits the fit accuracy. These steps or “plateau effect” are probably related to the crystal size distribution of the powder sample. Assuming that every crystal contains an even distribution of defects concentration and/or identical mosaicity, the larger the crystal, the higher the thermal energy (or time) required to initiate the transition. Consequently, the phase transition occurs progressively by domains or by particle size. Therefore, the “plateau effect” is more pronounced at lower temperature. For higher isothermal temperatures, this size effect is no more visible due to the increase of the conversion rate. Indeed, the kinetics of the phase transition increases exponentially with temperature (the calculated Avrami constant k(T) at 145 °C is 3 orders of magnitude higher than at 85 °C). In order to evaluate the accuracy of TR-SHG measurements for the determination of the activation energy, results obtained by DSC and TR-SHG were compared. In the particular case of MHBA, differential isothermal calorimetric measurements cannot be used because of the weakness of the exothermic event associated with the transformation of the metastable form into the stable form. Indeed, the limit of detection of the metastable form by DSC has been evaluated to 17 wt % (for a 16 K/min heating rate).7 Therefore, only dynamic DSC measurements are possible. The solid−solid transition of MetMHBA toward StabMHBA was studied at five different heating rates (φ): 1 K·min−1, 2 K·min−1, 4 K·min−1, 8 K·min−1, and 16 K·min−1 from 35 to 220 °C, and thermograms are presented in Figure 5. The corresponding thermal data are listed in Table 2. The average onset fusion temperature Tonset is 202.04 °C. Data m processing leads to an average melting enthalpy of 32.7 kJ·mol−1

(6)

where R is the universal gas constant and A is a constant. Figure 6 shows the plot of ln(φ)/T2p versus 1000/Tp. The slope of the resulting straight line (linear regression with regression coefficient R2 = 0.993) gives a calculated activation energy of 139 kJ·mol−1. Ea, determined by isothermal TR-SHG measurements using Avrami model (range 137−144 kJ·mol−1 depending of the data processing used), is in good agreement with values found by DSC using Kissinger and Ozawa methods (139 kJ·mol−1). However, in order to avoid any bias related to a step effect, it is suitable to limit the dispersion in the crystal size distribution. Another point to consider is the possibility of a local heating induced by the laser beam, which could modify kinetics by energy input. In the present case, this effect was limited by choosing the minimum of incident energy in order to keep the better SHG signal/noise ratio. During isothermal experiments, the laser always strikes the same part of the powder. SHG measurements at other parts of the powder were performed, and values found were similar (maximum deviation of 0.5%). Thus, energy brought by the laser has no significant effect on the global result of the experiments. Kinetic Stability of MetMHBA. An accurate monitoring of the evolution of transformed fraction x by TR-XRPD is difficult because of analysis duration and the low vapor pressure of MHBA. DSC cannot be used with relevancy in this case due to the low detection threshold of metMHBA (e.g., 17 wt %). In contrast, a time temperature transformation diagram (TTT) giving the relative stability of metastable form versus temperature and necessary time to complete the phase transition of the MHBA polymorphic system can be established from isothermal SHG measurements (Figure 7).

Table 1. Fit Parameters of Isothermal SHG Measurements at 85 °C, 115 °C, 130 °C, and 145 °C direct fit of experimental data using least-squares method temperature (°C) 85 115 130 145

Avrami exponent n 1.399 ± 0.003 1.223 ± 0.047 1.004 ± 0.006 1.070 ± 0.002

k(T) −6

rms −8

3.64 × 10 ± 9.063 × 10 4.37 × 10−5 ± 1.304 × 10−6 7.06 × 10−4 ± 2.86 × 10−5 3.55 × 10−3 ± 3.559 × 10−5

1.324 2.412 1.874 3.109 3700

linear regression of the ln[−ln(1 − x)] versus ln(t) plot Avrami exponent n 1.354 ± 0.036 1.205 ± 0.038 1.025 ± 0.043 1.068 ± 0.024

R2

k(T) −6

−7

4.64 × 10 ± 1.151 × 10 5.12 × 10−5 ± 1.581 × 10−5 5.77 × 10−4 ± 2.234 × 10−5 3.97 × 10−3 ± 8.659 × 10−5

0.991 0.988 0.977 0.997



Article

Figure 5. DSC curves for 1, 2, 4, 8, and 16 K·min−1 heating rates. The inlet is a zoom of DSC scan from 130 to 195 °C starting from pure MetMHBA. Weak exothermic event is visible and attributed to the transformation of the metastable form. The peak temperature, Tp, increases with the heating rate.

Table 2. Thermal Data for metMHBA for Five Different Heating Rates (Average of Two Set of Measurements) solid−solid transition heating rate φ (K/min) 1 2 4 8 16

Tp (°C) 150.0 ± 0.3 156.4 ± 0.7 163.1 ± 0.6 172.4 ± 0.8

fusion

enthalpy of transformation (kJ/mol)(exothermic)

Tonset m

(°C)

201.8 ± 0.4 201.9 ± 0.3 202 ± 0.8 202 ± 0.3 202.5 ± 1

0.592 ± 0.118 0.567 ± 0.108 0.600 ± 0.131 0.523 ± 0.110

Figure 6. Kissinger plot of ln(φ)/T2p versus 1000/Tp. The activation energy Ea is 139 kJ·mol−1. The same value is obtained by application of the Ozawa model − data not shown.

Tpeak m

(°C)

203.5 ± 0.3 204 ± 0.5 205.1 ± 0.3 207 ± 0.8 208.3 ± 1.4

melting enthalpy (kJ/mol) 31.19 ± 1.75 32.93 ± 1.05 33.04 ± 0.44 32.62 ± 0.39 33.70 ± 0.71

Figure 7. TTT diagram of MHBA metastable form for 85−145 °C temperature range; three domains can be determined: (a) pure stable form; (b) mixture of stable and metastable forms; (c) pure metastable form. Lines represent the converted fraction of MetMHBA.

Figure 7 shows that the kinetic stability of MetMHBA decreases with temperature and that kinetics of the solid−solid transition becomes predominant from 85 °C to higher temperatures where the metastable phase could survives only for a few minutes. Of course, the higher the temperature, the shorter the necessary period of time to complete the transition

toward the stable form: more than 2 days are necessary to complete the solid−solid transition at 85 °C instead of 20 min at 145 °C, and MetMHBA still exhibits a high SHG signal after a storage of several months at room temperature. Therefore, 3701



Article

H3A).29 The structure of the stable form is based on the usual carboxylic acid dimers (intermolecular hydrogen bonds between opposite carboxylic groups). Each dimer is linked to two other dimers by two H-bonds (O3H3A---O3). The structure of the metastable form consists of molecular ribbons spreading along c stacked through π−π interactions along b. Indeed, MHBA molecule creates intermolecular hydrogen bond network with four other molecules using the −OH and −COOH groups that leads to ribbons. Consecutive ribbons along b establish π−π interactions (phenyl rings are 3.4 Å and parallel), and this stacking gives rise to molecular layers in bc. The similarity between crystal lattice volumes of both polymorphs and the efficient crystal packing of the metastable form associated with high Ea of the solid−solid transition can explain the kinetic stability of the metastable phase. There is no structural affiliation between the hydrogen bond networks of the two polymorphic forms. This excludes the possibility of a smooth and continuous transition via a concerted movement of molecules. Thus, it is likely that the irreversible solid−solid transition from the metastable to the stable form proceeds through a destructive/reconstructive mechanism. To confirm a destructive/reconstructive mechanism, hot-stage polarized light microscopy can be used. This is a relevant technique to gather information on solid−solid transition mechanisms. Therefore, isothermal temperature-resolved microscopy was performed on large metastable single crystals (a few millimeters) of MHBA at 85, 115, 130, and 145 °C (the sublimation induced by an isolated crystal is limited and allows the heating stage to be kept closed during the whole experiment). At 85 °C, a crystallization front became visible and started to move after an annealing of 20 h. Then it propagated through the single crystal, which kept its shape integrity in agreement with the very close volumes of the two forms. Solid−solid transition was completed within 2 days and 14 h in good agreement with the value obtain by isothermal treatment of the powder (see Supporting Information). For higher isothermal temperatures, observations were similar, but the time required to complete the solid−solid transition decreased dramatically with increasing temperature. Displacement of a crystallization front is another argument in favor of the destructive/reconstructive solid phase transition mechanism.30 Since the metastable phase is non-centrosymmetric, SHG microscopy is particularly appropriate to differentiate the polymorphs and to image the crystallization front. Therefore, SHG microscopy was performed at room temperature on a single crystal of MetMHBA annealed at 100 °C for 20 min. A 3D reconstruction of the metastable part of the crystal is shown in Figure 10. Results show that nucleation of the stable form (white part) occurred at a different location in the metastable crystal (green part) and was followed by the growth of the stable form. The step shape observed on the x = f(time) curves (Figure 3) is also consistent with a destructive/reconstructive mechanism as described by Mnyukh.30 Another fact that supports this mechanism is the value of the Avrami exponent, n, close to 1 (1.070−1.399 range) that corresponds, in metal and alloy compounds, to solid−solid transition initiated at a grain boundary.31 This could be correlated, in the molecular crystal, to the nucleation on crystal defects and/or on boundaries of mosaic blocks.

according to the exponential dependency of k(t) with the temperature, a complete conversion of the metastable form could take several years at room temperature (a SHG signal, corresponding to ca. 1% wt of the metastable form, was detected in a commercial sample produced in 2011; see Supporting Information). If the study of the metastable equilibrium is of particular interest regarding the relative stability of the phase, it does not provide information about the molecular reorganization of the solid along the transition. In order to gather information about the transition mechanism, and the possible similarities between the two polymorphic forms such as hydrogen bonds network, packing, cell parameters), the study of the crystalline structure of both phases was performed. Solid−solid Transition Mechanism. The two unsolvated polymorphic forms of MHBA are referenced in Crystal Structure Database as BIDLOP01 (metastable Pna21 form) and BIDLOP (stable P21/n form). Crystallographic data from CSD are reported in Table 3. Single crystal X-ray diffraction (SC-XRD) Table 3. Crystal Data and Refinement Parameters Obtained from CCSD and SC-XRD for Stable and Metastable 3Hydroxybenzoic Acid ref code crystal system space group a b c β (°) volume (Å3) Z analysis temperature (K) wR (F2) measured data unique data (I > 2σ) no. of restraints/ parameters goodness of fit density

BIDLOP

BILOP01

this work CCDC 933474

monoclinic P21/n 5.493(2) 4.943(0) 23.022(1) 92.421(1) 624.536 4 283−303

orthorhombic Pna21 20.075(5) 3.751(2) 8.291(2) 90 624.323 4 283−303

orthorhombic Pna21 20.085(2) 3.7591(4) 8.293(1) 90 626.1(1) 4 293(2)

0.055

0.047

0.0319 4530/1283 1224 1/93

1.469

1.469

1.052 1.465

was performed for MetMHBA in order to detect a potential disorder in the crystal packing. The corresponding crystallographic data (reported in Table 3) are in good agreement with the data from CSD (with a slightly better reliability factor) and did not give any argument in favor of a disordered structure. Unfortunately, single crystal of StabMHBA with sufficient good quality to perform SC-XRD could not be isolated due to simultaneous crystallization of stable and metastable polymorphs.27,28 The crystal structures of the two polymorphs are represented in Figures 8 and 9. The hydrogen bonds involving the carboxylic moieties, i.e., the formation of dimers are represented in dashed blue lines; the hydrogen bonds involving the hydroxyl moieties and linking the dimers along the c axis are represented in dashed green lines. These interactions build F faces parallel to the bc plan. The corresponding H-bonds are listed in Table 4. In both polymorphs, the molecule of MHBA adopts a lowenergy stable conformer form as shown in Figures 8 and 9 (coplanarity of the phenyl ring (C2, C3, C4, C5, C6, C7) and the carboxyl group, and C1O2 carboxyl directed toward the O3−

■

CONCLUSION This study underlines the interest of the TR-SHG method for the investigation of solid−solid transitions. In particular, results 3702



Article

Figure 8. Packing of the stable form of MHBA. In dashed blue lines: hydrogen bonds involving carboxylic moieties and in dashed green lines: hydrogen bonds involving the hydroxyl moieties. Top view: Projection along a: dimers are linked along the c axis. Bottom view: projection along b.

Figure 9. Packing of the metastable form: On the left: view along the b axis (hydrogen bonds are represented in dashed blue lines). Top right: view along the c axis: consecutive ribbons are stacked through π−π interactions (d ∼ 3.4 Å, dashed green lines). Bottom right: two consecutive molecular ribbons stacked through π−π interactions.

Table 4. Hydrogen Bond Parameters of Both Polymorphsa structure

interaction

d(H···A) (Å)

d(D···A) (Å)

angle (deg)

P21/n

O1H1−O2 O3H3A−O3 C5H5−O2 O1H1−O3 C4H4−O1

1.473 1.976 2.626 1.874 2.673

2.617 2.796 3.455 2.687 3.601

172.16 164.12 148.16 170.86 176.10

Pna21 a

demonstrate that TR-SHG can be considered as an interesting alternative to the usual methods (DSC, XRPD, RAMAN) for the determination of the activation energy and the kinetic stability of MetMHBA with the establishment of a TTT diagram obtained from isothermal TR-SHG measurements. This kind of diagram could be helpful for the determination of storage conditions of metastable polymorphs. TR-SHG, optical and SHG microscopy experiments highlighted the presence of a crystallization front consistent with the crystal structure analyses that reveal no structural filiation between the polymorphs. These observations

D = donor and A = acceptor.

3703



■

Article

ACKNOWLEDGMENTS Plateforme Imagerie Cellulaire et Tissulaire IBISA de l’université de Reims Champagne Ardenne. Dr. Christine TERRYN is greatly acknowledged for her contribution to this study.

■

Figure 10. SHG microscopy of a MHBA metastable single crystal in transition toward the stable form. Green volume corresponds to the metastable form, and white volume corresponds to the stable form.

led us to propose a destructive/reconstructive mechanism for the irreversible MHBA solid−solid transition. It is worth mentioning that SHG is observed only in noncentrosymmetric space groups (24.1% of CSD),32 which limits the application of this technique to transitions involving at least one non-centrosymmetric phase. Nonetheless, this class includes all chiral molecules, and due to the high sensibility and the rapidity of SHG measurements, TR-SHG and SHG microscopy seem particularly adequate for the study of their solid−solid transitions. Moreover, in the favorable case, SHG analyses can detect only a few ppm of non-centrosymmetric crystal and consequently allows detection of the formation of a new crystalline phase before usual methods (i.e., in StabMHBA matrix, the detection of the percentage of metastable form is 17% by DSC, 1% by XRPD and only 2 ppm by SHG).7 This aspect is important when possible conversion of the metastable form occurs during heating.

■

ASSOCIATED CONTENT

S Supporting Information *

X-ray crystallographic information files (CIFs) for metastable MHBA and XRPD of metastable, stable, and commercial compounds. Results of Osawa model and of linear fit are presented. Optical microscopy movies of isothermal treatment at 85 °C (Movie 1) and of a single crystal of MetMHBA (Movie 2) undergoing a solid−solid transition are presented (ZIP file). This material is available free of charge via the Internet at http://pubs. acs.org.

■

REFERENCES

(1) Bernstein, J. Polymorphism in Molecular Crystals; Clarendon Press/ International Union of Crystallography: Oxford, 2002; Vol. 14. (2) Coquerel, G. Chem. Eng. Process. 2006, 45, 857−862. (3) Martins, D.; Sanselme, M.; Houssin, O.; Dupray, V.; Petit, M.; Pasquier, D.; Diolez, C.; Coquerel, G. CrystEngComm 2012, 14, 2507− 2519. (4) Wanapun, D.; Kestur, U. S.; Kissick, D. J.; Simpson, G. J.; Taylor, L. S. Anal. Chem. 2010, 82, 5425−5432. (5) McGregor, C.; Saunders, M. H.; Buckton, G.; Saklatvala, R. D. Thermochim. Acta 2004, 417, 231−237. (6) Coste, S.; Schneider, J.-M.; Petit, M.-N.; Coquerel, G. Cryst. Growth Des. 2004, 4, 1237−1244. (7) Clevers, S.; Simon, F.; Dupray, V.; Coquerel, G. J. Therm. Anal. Calorim. 2012, 112, 271−277. (8) Kurtz, S. K.; Perry, T. T. J. Appl. Phys. 1968, 39, 3798−3813. (9) Galland, A.; Dupray, V.; Berton, B.; Morin-Grognet, S.; Sanselme, M.; Atmani, H.; Coquerel, G. Cryst. Growth Des. 2009, 9, 2713−2718. (10) Dougherty, J. P.; Kurtz, S. K. J. Appl. Crystallogr. 1976, 9, 145− 158. (11) Bi, W.; Louvain, N.; Mercier, N.; Luc, J.; Rau, I.; Kajzar, F.; Sahraoui, B. Adv. Mater. 2008, 20, 1013−1017. (12) Smilowitz, L.; Henson, B. F.; Romero, J. J. J. Phys. Chem. A 2009, 113, 9650−9657. (13) LeCaptain, D. J.; Berglund, K. A. J. Cryst. Growth 1999, 203, 564− 569. (14) Kissick, D. J.; Wanapun, D.; Simpson, G. J. Annu. Rev. Anal. Chem. 2011, 4, 419−437. (15) Hall, V. J.; Simpson, G. J. J. Am. Chem. Soc. 2010, 132, 13598− 13599. (16) Dhinojwala, A.; Wong, G. K.; Torkelson, J. M. J. Chem. Phys. 1994, 100, 6046. (17) Liu, L.; Xu, L.; Hou, Z.; Xu, Z.; Chen, J.; Wang, W.; Li, F. Phys. Lett. A 1999, 262, 206−211. (18) Kirk-Othmer Encyclopedia of Chemical Technology; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2000. (19) Gridunova, G. V.; Furmunova, N. G.; Struchkov, Y. T.; Ezhkova, Z. I.; Grigoreva, L. P.; Chayanov, B. A. Kristallografiya 1982, 27, 267. (20) Nordström, F. L.; Rasmuson, Å. C. Eur. J. Pharmaceut. Sci. 2006, 28, 377−384. (21) Avrami, M. J. Chem. Phys. 1939, 7, 1103−1112. (22) Avrami, M. J. Chem. Phys. 1940, 8, 212−224. (23) Avrami, M. J. Chem. Phys. 1941, 9, 177. (24) Fanfoni, M.; Tomellini, M. Nouv Cim D 1998, 20, 1171−1182. (25) Perlovich, G. L.; Volkova, T. V.; Bauer-Brandl, A. J. Pharm. Sci. 2006, 95, 1448−1458. (26) Sabbah, R.; Le, T. H. D. Can. J. Chem. 1993, 71, 1378−1383. (27) Nordstrom, F. L.; Svärd, M.; Malmberg, B.; Rasmuson, Å. C. Cryst. Growth Des. 2012, 12, 4340−4348. (28) Svard, M.; Rasmuson, Å. C. Cryst. Growth Des. 2013, 13, 1140− 1152. (29) Aarset, K.; Page, E. M.; Rice, D. A. J. Phys. Chem. A 2008, 112, 10040−10045. (30) Mnyukh, Y. Fundamentals of Solid-State Phase Transitions, Ferromagnetism and Ferroelectricity; Authorhouse: United States, 2001. (31) Christian, J. The Theory of Transformations in Metals and Alloys: Part I + II; Elsevier: Amsterdam, 2002. (32) Allen, F. H. Acta Crystallogr. B 2002, 58, 380−388.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest. 3704


Monotropic Transition Mechanism of m-Hydroxybenzoic Acid

Recommend Documents