I. INTRODUCTION
Imipenem monohydrate, C12H17N3O4S·H2O [6-(1-hydroxyethyl)-3-(2-(iminiomethylamino)ethylthio)-7-oxo-1-azabicyclo(3.2.0)hept-2-ene-2-carboxylate monohydrate] (Figure 1), has been used with cilastatin sodium as an antimicrobial injection. Imipenem monohydrate, such as penicillin, belongs to the β-lactam antibiotic class and is often used as a last resort in place of other antibiotics with similar pharmaceutical functions. The chemical structure of imipenem monohydrate, unlike penicillin and many other β-lactam antibiotics, has no sulfur in the five-membered ring fused to the β-lactam ring, and contains a double bond in this same five-membered ring. Despite the restricted use of β-lactam antibiotics as a result of the discovery of β-lactam-resistant bacteria, imipenem monohydrate has a broad application against a variety of bacteria.
Structural formula of imipenem monohydrate.
Rietveld refinement of the imipenem monohydrate powder data confirmed an orthorhombic crystal system and a P212121 space group. The molecular structure (Figure 1) includes a number of electronegative atoms, such as oxygen, nitrogen, and sulfur, in the forms of carboxyl, hydroxyl, imino, amino, and sulfide groups. The crystal structure also includes water, which contributes to intramolecular and intermolecular hydrogen bonding. This hydrogen-bonding network contributes greatly to the crystal packing arrangement. There have been few studies of this compound and only one concerning a single-crystal form (Ratcliffe et al., Reference Ratcliffe, Wildonger, Michele, Douglas, Hajdu, Goegelman, Springer and Hirshfield1989). However, its experimental powder data have not been explored. The PDF-4/Organics 2012 Database (ICDD, Reference Kabekkodu2011) contains one powder pattern calculated from the single-crystal data (Needham et al., Reference Needham and Faber2003). Here, we report our findings on the imipenem monohydrate experimental X-ray powder diffraction pattern.
II. EXPERIMENTAL
Imipenem monohydrate, purchased from United States Pharmacopeia (98–101% purity), was gently ground with an agate mortar and pestle. The powder was dusted on to a PANalytical PW1817/32 zero-background plate (obliquely cut silicon crystal) in a PANalytical PW1813/32 plate holder. The X-ray powder diffraction data were collected on a PANalytical X'Pert PRO system equipped with a copper X-ray source tube and an X'Celerator Detector. The scan range was 7–60° 2θ with a step size of 0.0167° 2θ. Two 0.02-rad Soller slits (for both incident and diffracted beams) were used to minimize the axial divergence aberration in the diffraction pattern. A 0.020-mm nickel filter was used to absorb CuKβ radiation and a 0.125° antiscatter slit was used to reduce background. The experimental conditions are listed in Table I. The powder pattern is presented in Figure 2. This procedure with a zero-background holder gave a very good texture index of 1.03 after refinement, indicating very little preferred orientation.
X-ray powder diffraction pattern for imipenem monohydrate. a = 8.2534(3) Å, b = 11.1293(4) Å, c = 15.4609(6) Å, R p = 2.68%, WR p = 4.03%, and χ2 = 3.883.
XRD data collection conditions for imipenem monohydrate.
III. RESULTS AND DISCUSSION
Analysis of the powder diffraction data was accomplished using PANalytical HighScore Plus (Needham et al., Reference Needham, Faber and Fawcett2006). Background subtraction and Kα2 removal were performed before determining the location of the diffraction peaks. The extensive overlapping strong peaks below 35° 2θ makes peak search a challenging task. HighScore Plus (PANalytical, 2011) was used to find peaks. SQLAids (Sagnella, Reference Sagnella2011) was used to evaluate the accuracy of these peak positions relative to the proposed orthorhombic cell. Results are shown in Table II. Rietveld refinement of the powder data was performed using GSAS (Von Dreele and Larson, Reference Von Dreele and Larson2001). The refinement parameters used were Chebyshev polynomial background, scaling, sample displacement, peak asymmetry, peak shape profiles (Gaussian and Lorentzian), U iso temperature factors, unit-cell parameters, and atomic coordinates. The results are shown in Figure 2, with final refinement statistics of R p = 2.68%, wR p = 4.03%, and χ 2 = 3.883 (Figure 2). The refined unit-cell parameters are a = 8.2534(3), b = 11.1293(4), and c = 15.4609(6) Å. These vary from the reported values [a = 8.268(3), b = 11.140(6), and c = 15.452(9) Å] of single-crystal data in the PDF-4/Organics 2012 (Ratcliffe et al., Reference Ratcliffe, Wildonger, Michele, Douglas, Hajdu, Goegelman, Springer and Hirshfield1989) by −0.18, −0.14, and 0.06%, respectively. One value of U iso was used for refinement. The resulting refined U iso value is 0.0893. The refined atomic coordinates are listed in Table III.
XRD data for imipenem monohydrate.
λ = 1.5406 Å and I/I max are based on peak height
Table IV lists three possible hydrogen bonds in addition to the water hydrogen bonds (Figure 3). Hydrogen bonds are shown as dotted lines in Figure 3. Two are intermolecular (N3/O2 and N3/O3) and one is intramolecular (O3/S1). Although the O3/S1 bonding distance is large and the sulfur atom has relatively weak electronegativity, the intramolecular proximity of the sulfur atom and hydroxyl hydrogen atom, and the resulting six-membered ring configuration makes its hydrogen bonding a very strong possibility. The water is hydrogen bonded to four molecules and the hydrogen-bonding length of water (Table IV) is similar to those of the reported values (Ratcliffe et al., Reference Ratcliffe, Wildonger, Michele, Douglas, Hajdu, Goegelman, Springer and Hirshfield1989). The O5/O2 and O5/O3 water hydrogen bonds are responsible for the head to tail packing of the fused rings of two adjacent molecules. Also, the O5/N2 and O5/O4 hydrogen bonds are orthogonal to the O5/O2 and O5/O3 bonds. The O5/O2 and O5/O3 hydrogen bonds are responsible for the arrangement of fused rings in a paralleled stacking fashion. The tetrahedral-like hydrogen bonding of the water molecule causes the stacking of fused rings in alternating fashion and hence forms a three-dimensional hydrogen-bonding network. The moderately long side chain with single bonds has a moderate amount of rotational freedom, which contributes to the conformational difference. The experimental unit cell is smaller than that derived from single-crystal data. The dusted powder samples on the zero-background plate minimized the preferred orientation effect. The thin layer of the experimental sample reduced the transparency aberration of the compound consisting of light elements. Finally, the zero-background sample holder minimized the background interference of the data analysis. The combination of all the advantages of the zero-background method for the imipenem monohydrate powder diffraction experiment enabled the successful resolution of overlapping peaks at low diffraction angles. This conclusion is supported by the more successful Rietveld refinement analysis as compared to other explored methods.
(Color online) Crystal structure of imipenem monohydrate.
Hydrogen-bonding lengths in imipenem monohydrate.
ACKNOWLEDGMENTS
The XRD data reported in our study were collected from an X-ray powder diffractometer supplied by PANalytical USA during ICDD clinics. We thank the generosity of PANalytical USA for making the instrument available for us to conduct experiments.
