I. INTRODUCTION
Curcumin, (1E,4Z,6E)-5-hydroxy-1,7-bis(4-hydroxy-3-methoxyphenyl)hepta-1,4,6-trien-3-one, is the principal active component of the spice turmeric, which is a common food additive, pigment, and ingredient for traditional medicines in Asia (Jayaprakasha et al., Reference Jayaprakasha, Jagan, Rao and Sakariah2005). Considerable recent attention has been directed to curcumin and structural analogues for their medicinal potential, including anti-inflammatory (Liang et al., Reference Liang, Yang, Zhou, Shao, Huang, Xiao, Huang and Li2009), anti-malarial (Mishra et al., Reference Mishra, Karmodiya, Surolia and Surolia2008), and anti-cancer properties (Mehta et al., Reference Mehta, Pantazis, McQueen and Aggarwal1997; Adams et al., Reference Adams, Ferstl, Davis, Herold, Kurtkaya, Camalier, Hollingshead, Kaur, Sausville, Rickles, Snyder, Liotta and Shoji2004), among others.
The structure of curcumin was originally solved by Tonnesen et al. (Reference Tonnesen, Karlsen and Mostad1982) using single-crystal X-ray data, but some uncertainty surrounded the enol H atom, which was assigned two atomic sites with the occupancy split between them. A more recent single-crystal study by Parimita et al. (Reference Parimita, Ramshankar, Suresh and Guru Row2007) established a single enol H site nearly symmetrically centered between the O2 and O3 atoms. Figure 1 displays two illustrations of the curcumin molecule, including the atom labels, which are identical to the site assignments previously defined by Parimita et al. (Reference Parimita, Ramshankar, Suresh and Guru Row2007).
Figure 1. (Colour online) Two 2D diagrams illustrating the molecular structure of curcumin (C21H20O6) and the assigned atom labeling.
Despite its usage and investigation for a variety of applications, for many years only one experimental powder diffraction pattern appeared for curcumin in the Powder Diffraction File (PDF 00-009-0816), which is an un-indexed, low precision entry. A new entry 00-063-0943 has been included as a private communication, based on data collected at 100 K and the structure of Form 1 from Sanphui et al. (Reference Sanphui, Goud, Khandavilli, Bhanoth and Nangia2011). Additional patterns 02-075-3596 and 02-093-7329 are included in the PDF-4 Organics calculated from the crystal structures of Ishigami et al. (Reference Ishigami, Goto, Masuda, Takizawa and Suzuki1999) and Suo et al. (Reference Suo, Huang, Weng, He, Li, Li and Hong2006), respectively. This paper examines the Rietveld refinement of curcumin using rigid body refinement with the Rietveld package GSAS/EXPGUI and provides a comprehensive reflection list for phase identification.
II. EXPERIMENTAL
Curcumin obtained from Sigma-Aldrich (product number C7727) was loaded as-supplied, with no grinding, into a 0.5 mm ID Kapton capillary, which was sealed at both ends with a Loctite adhesive.
Synchrotron powder diffraction (PXRD) patterns were collected using a Canadian Macromolecular Crystallography Facility beamline (CMCF-BM or 08B1-1) at the Canadian Light Source (CLS). 08B1-1 is a bending magnet beamline with an Si (111) double-crystal monochromator. Two-dimensional (2D) data were obtained using a Rayonix MX300HE detector with an active area of 300 mm × 300 mm. The patterns were collected at an energy of 18 keV (λ = 0.688 80 Å) and capillary–detector distance of 350 mm.
The 2D PXRD patterns were calibrated and integrated using the GSASII software package (Toby and Von Dreele, Reference Toby and Von Dreele2013). The sample–detector distance, detector centering and tilt were calibrated using a lanthanum hexaboride (LaB6) standard reference material (NIST SRM 660a LaB6) and the calibration parameters were applied to all patterns. After calibration, the 2D patterns were integrated to obtain standard 1D powder diffraction patterns. A pattern from an empty Kapton capillary (collected using the same conditions) was subtracted from the sample data during integration. The integrated LaB6 pattern was used to obtain the instrument resolution of the beamline for the refinement of the curcumin sample.
The single-crystal structure model of Parimita et al. (Reference Parimita, Ramshankar, Suresh and Guru Row2007) was used as a starting point for the refinement. The structure was analyzed with the Mogul 1.6 module of the Cambridge Structural Database (Allen, Reference Allen2002) in order to prepare appropriate restraints for the distances and angles associated with the carbon and oxygen atoms.
Rigid body Rietveld refinement was performed with the GSAS/EXPGUI program (Toby, Reference Toby2001; Larson and Von Dreele, Reference Larson and Von Dreele2004). The implementation of rigid body refinements in GSAS and EXPGUI have been described previously in the literature (Dinnebier, Reference Dinnebier1999; Lake and Toby, Reference Lake and Toby2011). Rigid bodies were created for each phenyl group based on the single-crystal atomic coordinates, including the hydrogen atoms, using the centroid of the ring as the rigid body origin. The isotropic displacement parameters (U iso) of the carbon and oxygen atoms were given the initial values from the single crystal model. Hydrogen atoms were constrained to U iso values of 1.3 times the nearest carbon or oxygen neighbor. The positions of the hydrogen atoms were unrefined but periodically optimized with the Mercury 3.3 module of the Cambridge Structural Database (Allen, Reference Allen2002). The global weight factor for the restraints was initially set at 100, but lowered in steps to a value of 60 for the final refinement. The restraints contributed 13.8% to the final χ 2. The background was refined using a combination of an orthogonal Chebyschev polynomial (three terms) and Debye scattering function (six terms).
A density functional geometry optimization (fixed experimental unit cell) was carried out using CRYSTAL09 (Dovesi et al., Reference Dovesi, Orlando, Civalleri, Roetti, Saunders and Zicovich-Wilson2005). The basis sets for the H, C, and O atoms were those of Gatti et al. (Reference Gatti, Saunders and Roetti1994). The calculation used 8 k-points and the B3LYP functional.
III. RESULTS AND DISCUSSION
A 2D PXRD pattern for curcumin is illustrated in Figure 2, which exhibits mild graininess, suggesting a portion of the crystallite size distribution is larger than optimal and the crystallite statistics deviate from those of an ideal powder. Patterns were also collected after grinding the powder in an attempt to eliminate the graininess; however, the patterns obtained from the ground powder suggest that the crystallinity decreases rapidly with grinding. Approximately 4 min grinding nearly completely eliminated the crystalline reflections, as illustrated by Figure 3, and while grinding for 2 min clearly decreased the crystallinity, it did not completely eliminate the graininess. This type of material is well suited to the current Debye–Scherrer geometry with a large area detector, where large portions of the Debye rings can be integrated to obtain accurate intensity estimates. Although the 2D CCD detector limits the resolution, it is superior for obtaining reasonable intensity estimates compared with point detector or multi-detector/analyzer setups which are designed to optimize resolution. This can be crucial for some molecular compounds which cannot be mechanically ground without significantly reducing their crystallinity.
Figure 2. (Colour online) A 2D pattern obtained from curcumin, illustrating the slight graininess of the as-supplied powder.
Figure 3. (Colour online) A comparison of the PXRD patterns for curcumin with no grinding (top), ~2 min grinding (middle), and ~4 min grinding (bottom). The patterns were obtained at 18 keV (0.688 80 Å).
The final Rietveld refinement is illustrated in Figure 4, whereas the refined atomic coordinates are shown in Table I. Selected bond lengths and angles are provided in Table II for the single-crystal refinement (Parimita et al., Reference Parimita, Ramshankar, Suresh and Guru Row2007), GSAS PXRD refinement and the density functional theory (DFT) calculations for comparison. The bond lengths are generally comparable between the single-crystal and PXRD refinements, but there are some differences. The root-mean-square (RMS) difference between the heavy atom positions and those of Parimita et al. (Reference Parimita, Ramshankar, Suresh and Guru Row2007) is 0.110 Å, and the RMS difference between the experimental and DFT structures is 0.176 Å. In particular, the C12–C13 and C13–C14 bond differ significantly, with the DFT calculated bond lengths falling between the values in both cases. The PXRD refined isotropic displacement parameters do not suggest significant issues with the refinement, and generally show marginal differences from the single-crystal values.
Figure 4. (Colour online) A plot of the final Rietveld refinement obtained for the curcumin sample using GSAS (χ 2 = 1.54).
Table I. The refined crystal structure of curcumin obtained from the GSAS refinement (χ 2 = 1.54, R p = 0.0322, and R wp = 0.0428). The refined lattice parameters obtained are a = 12.6967(1) Å, b = 7.198 52(3) Å, c = 19.9533(2) Å, and β = 95.1241(6)°. All atom positions, in space group P2/n (#13), correspond to 4 g Wyckoff sites with occupancies of 1.
Table II. A comparison of selected bond distances and angles obtained from the literature single-crystal refinement (Parimita et al., Reference Parimita, Ramshankar, Suresh and Guru Row2007), the PXRD refinement, and the DFT calculations.
Hydrogen bonds obtained from the DFT calculations are shown in Table III, with the D···A values from the GSAS refinement provided for comparison. The shorter distances (below 3 Å) vary little between the DFT and PXRD results, generally less than 0.05 Å; however, larger differences can be seen between some of the longer distances. Much of the impetus behind the single-crystal study of Parimita et al. (Reference Parimita, Ramshankar, Suresh and Guru Row2007) was an accurate determination of the position of the enol H atom, H23. The DFT calculations suggest that H23 is most closely associated with O2, whereas the single-crystal results place the H23 atom nearly symmetrically between O2 and O3 [at distances of 1.26(5) and 1.28(5) Å, respectively].
Table III. Curcumin hydrogen bonds obtained from the DFT calculations, with PXRD values for the D···A distances given underneath the DFT values.
aThe hydrogen bond energies were calculated from the overlap populations using the correlations of Rammohan and Kaduk (Reference Rammohan and Kaduk2014).
The final reflection list for curcumin obtained from the GSAS refinement is provided in Table IV. To process the data, adjacent reflections with relative integrated intensities larger than 0.2% and closer than 0.02°2θ were summed as multiple reflections and assigned a weighted average reflection position. The final reflection list in Table IV contains all reflections with relative integrated intensities greater than 0.5%.
Table IV. The reflection list obtained for curcumin from the Rietveld refinement, including integrated intensities ≥0.6%, after summing reflections closer than 0.02° as multiple reflections and using a weighted average reflection position.
This curcumin pattern explains well the crystalline peaks in the pattern of commercial ground turmeric (Figure 5). The sample apparently contains a trace of quartz. The peaks of pattern 00-063-0943 (measured at 100 K) are significantly shifted from the experimental peaks and those of the room-temperature calculated patterns 02-075-3596 and 02-093-7329. Anisotropic thermal expansion between 100 and 300 K means that patterns calculated from low-temperature crystal structures are not as useful for phase identification as might be expected, as the peaks are shifted significantly from their positions at ambient conditions. This high-quality powder pattern measured at ambient conditions should prove useful for identification of curcurmin using normal powder diffraction techniques.
Figure 5. (Colour online) The PXRD pattern obtained from commercial ground turmeric.
ACKNOWLEDGEMENTS
Thanks to Dr. Pawel Grochulski for helpful discussions which improved this manuscript. Research described in this paper was performed using beamline 08B1-1 at the Canadian Light Source, which is supported by the Canadian Foundation for Innovation, the Natural Sciences and Engineering Research Council of Canada, the National Research Council Canada, the Canadian Institutes of Health Research, the Government of Saskatchewan, Western Economic Diversification Canada, and the University of Saskatchewan.
SUPPLEMENTARY DATA
The supplementary material for this article can be found at http://www.journals.cambridge.org/PDJ
