Article pubs.acs.org/JPCA
Investigating Tautomeric Polymorphism in Crystalline Anthranilic Acid Using Terahertz Spectroscopy and Solid-State Density Functional Theory Sean P. Delaney, Ewelina M. Witko, Tiffany M. Smith, and Timothy M. Korter* Department of Chemistry, Syracuse University, 1-014 Center for Science and Technology, Syracuse, New York 13244-4100, United States S Supporting Information *
ABSTRACT: Terahertz spectroscopy is sensitive to the interactions between molecules in the solid-state and recently has emerged as a new analytical tool for investigating polymorphism. Here, this technique is applied for the first time to the phenomenon of tautomeric polymorphism where the crystal structures of anthranilic acid (2-aminobenzoic acid) have been investigated. Three polymorphs of anthranilic acid (denoted Forms I, II and III) were studied using terahertz spectroscopy and the vibrational modes and relative polymorph stabilities analyzed using solid-state density functional theory calculations augmented with London dispersion force corrections. Form I consists of both neutral and zwitterionic molecules and was found to be the most stable polymorph as compared to Forms II and III (both containing only neutral molecules). The simulations suggest that a balance between steric interactions and electrostatic forces is responsible for the favoring of the mixed neutral/ zwitterion solid over the all neutral or all zwitterion crystalline arrangements.
1. INTRODUCTION The identification and differentiation of polymorphic molecular solids is a topic of great chemical interest because polymorphism directly affects the physical properties of substances, such as melting point and solubility. This is particularly relevant for the pharmaceutical industry where such characteristics influence drug stability and bioavailability. There are many different origins of polymorphism, all ultimately leading to the existence of various stable crystal forms for a given compound. At the simplest level, polymorphism is the formation of multiple crystalline solids resulting from different solid-state arrangements of conformationally rigid molecules.1 A more detailed view of polymorphism includes configurational polymorphism, where the molecules themselves are modified between forms, and include geometric isomers, diastereomers, enantiomers, tautomers, chiral structures, and zwitterions.1−4 There has been some debate over whether configurational polymorphs are indeed polymorphs in the traditional sense because they are, in theory, different molecules.1−4 The solidstate transformation between a zwitterionic molecule and its neutral, nonzwitterionic counterpart is known as desmotropy or © 2012 American Chemical Society
tautomeric polymorphism. A zwitterionic molecule can be identified by the charge distribution within itself, which consists of localized positive and negative charges (not including dipolar partial charges) on different parts of the same molecule that cannot be compensated for by a double-bond rearrangement.1 This modification of charge distribution (often via proton transfer) leads to the zwitterionic and neutral species having different properties, specifically different crystalline packing arrangements, and thus being polymorphs of each other. Anthranilic acid (2-aminobenzoic acid) exhibits three crystalline polymorphs (denoted Forms I, II, and III), but only one of these contains anthranilic acid molecules with zwitterionic character. The most fascinating detail of these solids is that Form I (Figure 1) has both neutral and zwitterionic molecules incorporated into the crystal structure, while Forms II and III (Figure 2) contain exclusively neutral molecules.5−9 The unusual structure of Form I has been Received: April 24, 2012 Revised: July 11, 2012 Published: July 11, 2012 8051
dx.doi.org/10.1021/jp303961w | J. Phys. Chem. A 2012, 116, 8051−8057
The Journal of Physical Chemistry A
Article
Figure 1. Anthranilic acid Form I asymmetric unit cell with atomic labels (a) and unit cell molecular packing down the a-axis (b).
Figure 2. Anthranilic acid Forms II and III asymmetric unit cells with atomic labels (a), unit cell molecular packing down the c-axis for Form II (b), and unit cell molecular packing down the a-axis for Form III (c).
verified by neutron diffraction studies.10 Forms II and III are clearly considered polymorphs because they fit the traditional view of polymorphs. The rationale behind labeling Form I a polymorph alongside Forms II and III is rooted in the definition of tautomeric polymorphism, where Form I differs from Forms II and III by only a single proton transfer. In this study, the low-frequency vibrations of all three polymorphs of anthranilic acid were investigated using terahertz (THz) spectroscopy over the range of 10−95 cm−1. Terahertz spectroscopy is becoming an increasingly utilized tool in the detection, identification, and characterization of diverse molecular samples. This includes exploration into intermolecular and intramolecular interactions within proteins, DNA, and other biological materials,11−13 as well as identification and detection of narcotics14−16 and explosives.17−20 Within the pharmaceutical industry, THz spectroscopy has been used in numerous ways. Some applications include the analysis of drugcoating materials,21,22 the study of liquid-crystalline and amorphous pharmaceutical compounds,23,24 examination of the differences between hydration states of a compound,25 and identification of polymorphs.26,27 More recently, terahertz spectroscopy and solid-state theory have been combined to predict the previously unknown crystal structures of complex organic solids.28 While THz investigations have yielded analytical information and new insights into molecular solids, the physical understanding of low-frequency THz vibrations is difficult to achieve because these vibrations result from the crystalline assembly (lattice vibrations) rather than from localized functional group
motions as is observed in the mid-infrared. In order to better understand these structures and vibrations, solid-state density functional theory (DFT) can be used to simulate the periodic structure of the molecular solid and the vibrations that arise from both intramolecular and intermolecular motions of the molecules in the crystal. Solid-state DFT also provides information concerning the relative energies of different polymorphs, which is particularly relevant in anthranilic acid, where the nature of the intermolecular forces changes between polymorphs.
2. EXPERIMENTAL SECTION Anthranilic acid was purchased from Sigma Aldrich (≥98% purity), and the solid-state structure was verified by both powder and single-crystal X-ray diffraction to correspond to Form I. Form II was obtained by fast recrystallization of anthranilic acid from methanol, and Form III was obtained from sublimation as well as from the melt, as described previously.5 The polymorphic purities of both Forms II and III were also verified by powder and single-crystal X-ray diffraction. For THz sample preparation, each polymorph was mixed with a powdered polytetrafluoroethylene (PTFE) matrix at a concentration that was dependent on the polymorphs. Form I samples used 0.8%, Form II 1.6%, and Form III 1.6% by mass to ensure that measurements were made within an optimal absorption range. The mixtures were then pulverized using a stainless steel grinder/mixer (Dentsply Rinn 3110-3A) to minimize particle size and reduce both Mie scattering and crystal anisotropy.29 The samples were pulverized for less time than is required to 8052
dx.doi.org/10.1021/jp303961w | J. Phys. Chem. A 2012, 116, 8051−8057
The Journal of Physical Chemistry A
Article
change forms,5 but the effects of pressure on crystalline polymorphs of other substances have been reported.3 It should be noted that in this study, no pressure-induced changes were observed. Approximately 0.55 g of the sample mixtures were pressed, at 2000 psi, into pellets with diameters of 13 mm and thicknesses of 2.0 mm. Pure PTFE was pressed into a pellet of equal mass to be used as a blank reference. The experimental spectra were obtained using a time domain pulsed THz spectrometer based on an amplified Ti:Sapphire femtosecond laser system. Zinc telluride crystals were used for both generation of THz radiation by optical rectification and detection by free-space electro-optic sampling.30,31 A detailed description of the THz spectrometer has been reported elsewhere.15 The samples and blank for measurement were held under vacuum in a cryostat with data acquired at both 293 and 78 K. Samples and blanks were scanned 32 times for each individual data set over a time window of 32 ps consisting of 3200 data points, which were then symmetrically zero-padded to a total of 6000 data points. The ratio of the Fouriertransformed data sets of the sample and blank resulted in a THz spectrum over the range of 10−95 cm−1 with a spectral resolution of approximately 1.0 cm−1. Each data set was replicated four times at both temperatures and then averaged to obtain the final spectra reported here.
Table 1. Cryogenic (100 K) X-ray Diffraction Unit Cell Dimensions for Anthranilic Acid Forms I, II, and III Form I empirical formula f.w. crystal system space group a (Å) b (Å) c (Å) α (deg) β (deg) γ (deg) V (Å3) Z μ (mm−1) T (K) λ (Å) R1 wR2
Form II
Form III
C6H4(NH2)COOH 137.14 orthorhombic orthorhombic monoclinic Pna21 Pbca P21/c 9.2734(15) 11.6322(14) 6.4900(4) 10.7851(19) 7.0969(8) 15.1290(10) 12.695(2) 15.8282(19) 6.9900(5) 90 90 90 90 90 112.5430(10) 90 90 90 1269.7(4) 1306.7(3) 633.89(7) 8 8 4 0.107 0.103 0.107 90 0.71073 0.71073 0.71073 0.0707 0.0514 0.0332 0.0994 0.0736 0.0781
Normal-mode frequencies and infrared intensities were then calculated for the optimized structures. The frequency of each normal mode was calculated within the harmonic approximation by numerical differentiation of the analytical gradient of the potential energy with respect to atomic position.37 The infrared intensities for each normal mode were calculated from the dipole moment derivatives (dμ/dQ) determined using the Berry phase technique of calculating Born charges as polarization differences between equilibrium and distorted geometries.32,40 Mode descriptions and assignments were made by visual inspection of atomic displacements for each normal mode.
3. THEORETICAL The simulations in this work were performed using the CRYSTAL09 software package32 utilizing the Perdew−Burke− Ernzerhof (PBE)33 generalized gradient approximation (GGA) density functional in combination with the atom-centered Gaussian-type 6-31G(d,p)34 and 6-311G(2d,2p)35 basis sets. The total energy convergence criteria were ΔE < 10−8 hartree for geometry optimizations and ΔE < 10−11 hartree for frequency calculations. All structural optimizations were performed without constraints on atomic positions or unit cell dimensions, other than those imposed by space group symmetry, and were begun using starting structures obtained by experimental X-ray diffraction measurements. A shrinking factor of 6 (64 k points in the irreducible Brillouin zone) was determined after sampling and monitoring of the total energy convergence as a function of the k-point count in reciprocal space according to the Pack−Monkhorst method.36 Truncation tolerances for Coulomb and HF exchange integrals were defined as 10 −8 , 10 −8 , 10 −8, 10 −8, and 10 −16 hartree (TOLINTEG command32,37). The radial and angular distributions for DFT integration were defined by a pruned (75, 974) grid. These computational parameters were used consistently for all three polymorphs. One limitation of solid-state DFT (utilizing typical functionals) is the proper representation of weak dispersion interactions. In this study, the solid-state DFT approach was supplemented with corrections for London-type dispersion forces, which are important to incorporate because they directly influence the solid-state structures of the molecules. Londontype dispersion interactions were added using a semiempirical correction proposed by Grimme38 and then later modified for the CRYSTAL program by Civalleri et al.39 Global scaling factors for the dispersion correction (s6 value) of 0.4 and 0.6 were used for the 6-31G(d,p) and 6-311G(2d,2p) basis sets, respectively, and were obtained through comparison of the calculated unit cell volumes and the experimental 100 K X-ray data (Table 1).
4. RESULTS AND DISCUSSION 4.1. Terahertz Spectroscopy. The 78 and 293 K THz spectra from 10 to 95 cm−1 of Forms I, II, and III of anthranilic acid are shown in Figure 3, and the observed vibrational frequencies are listed in Table 2. The low-temperature spectra exhibited sharpened spectral features due to the decrease in the populated vibrational states of the molecules in the solid state, making the 78 K spectra essential for the proper assignment of vibrational modes. All of the 78 K spectral features for the three forms of anthranilic acid exhibit a slight shift to higher energy versus that at room temperature, a common event in the THz spectroscopy of cooled solids.15,26,41 In the THz spectrum of Form I, seven total absorptions are clearly observed, six revealed at 53.9, 57.2, 63.9, 72.2, 80.0, and 87.2 cm−1, with one additional feature seen as a shoulder at 76.7 cm−1. Forms II and III reveal fewer spectral absorptions as compared to Form I, likely due to the reduction in asymmetric unit complexity in Forms II and III (Z′ = 1) versus that in Form I (Z′ = 2). The 78 K spectrum of Form II has frequencies centered at 32.2, 67.8, 70.0, and 77.2 cm−1. Form III also has four vibrational features in this spectral region but centered at 45.0, 56.1, 80.5, and 92.8 cm−1. The observed differences in the THz spectra between the three polymorphs are readily apparent and serve to uniquely identify the different forms. This can be attributed to variations in the crystal packing arrangements, as well as to changes in the neutral and zwitterionic character of the molecules. In order to better understand the chemical origins of the numerous 8053
dx.doi.org/10.1021/jp303961w | J. Phys. Chem. A 2012, 116, 8051−8057
The Journal of Physical Chemistry A
Article
4.2. Solid-State DFT Simulations. 4.2.1. Structural Analysis. Two different basis sets (6-31G(d,p) and 6311G(2d,2p)) were used with the PBE functional in the solid-state simulations to gauge the importance of basis set size. The structure of each polymorph was first fully optimized with each combination, and the optimized structures were then compared to the cryogenic X-ray diffraction data through root mean-squared deviations (RMSDs) of bond lengths, bond angles, and dihedral angles (Table 3) to determine the best theoretical method for subsequent simulation of THz spectra. The RMSDs represented here exclude all hydrogen atoms, and the calculation details can be found in the Supporting Information (Tables S1−S3). The quality of the unit cell parameters, including the unit cell volume, was also evaluated, and the results are listed in Table S4 of the Supporting Information. The optimization completed using PBE/6-31G(d,p) was determined to be of reasonable quality but was clearly the underperforming of the two as revealed by the RMSD values. On average, use of the 6-311G(2d,2p) basis set resulted in a 25% reduction in the RMSDs as compared to those with 6-31G(d,p). The quality of the structural reproduction by PBE/6-311G(2d,2p) led to its use in all of the THz spectral simulations that follow. 4.2.2. Vibrational Analysis. The simulated THz spectra of Forms I, II, and III are displayed in Figures 4−6, respectively. Figure 3. Terahertz spectra of anthranilic acid in polymorphic Form I (a), Form II (b), and Form III (c). The 293 K data are shown in gray, and the 78 K data are shown in black. Note the large changes in extinction coefficients between polymorphs.
Table 2. Observed Terahertz Vibrational Frequencies (cm−1) of Anthranilic Acid Forms I, II, and III at 293 and 78 K and the Approximate Correlations between the Two Temperatures Form I
Form II
Form III
293 K
78 K
293 K
78 K
293 K
78 K
53.2 55.6 63.2 69.4 − 77.2 84.4 91.7 95.0
53.9 57.2 63.9 72.2 76.7 80.0 87.2 98.9 −
− 62.2 − − 88.8 92.2
33.2 67.8 70.0 77.2 87.2 −
41.7 51.1 74.4 84.4 87.2 −
45 56.1 80.5 − − 92.8
Figure 4. Overlay of the 78 K experimental terahertz spectrum of anthranilic acid Form I (black) and the spectrum simulated by PBE/6311G(2d,2p) (gray).
The calculated vibrational modes (
