International
Tables for
Crystallography
Volume H
Powder diffraction
Edited by C. J. Gilmore, J. A. Kaduk and H. Schenk

International Tables for Crystallography (2018). Vol. H, ch. 3.8, pp. 339-340

Section 3.8.8. Using spectroscopic data

C. J. Gilmore,a G. Barra and W. Donga*

aDepartment of Chemistry, University of Glasgow, University Avenue, Glasgow, G12 8QQ, UK
Correspondence e-mail:  chris@chem.gla.ac.uk

3.8.8. Using spectroscopic data

| top | pdf |

There is no reason why the methodology described in this chapter cannot be used for other 1D data sets, e.g. Raman, IR, NMR and near-IR spectroscopies, although different data pre-processing is usually required. Raman spectroscopy is well suited to high-throughput screening: good-quality spectra can be collected in a few minutes, and sample preparation is straightforward and flexible, although the resulting spectra are not always as distinct as the PXRD equivalents (Mehrens et al., 2005[link]; Boccaleri et al., 2007[link]).

As an example we show the results of cluster analysis carried out on samples of carbamazepine, cimetidene, furosemide, mefenamic acid, phenilbutazone and sulfamerazine using Raman spectroscopy. A total of 74 samples were measured on a LabRam HR-800/HTS-Multiwell spectrometer at room temperature, equipped with a backscattering light path system of a light-emitting diode laser (785 nm, 300 mW) as an excitation source and an air-cooled charge-coupled device detector. A 20-fold superlong working distance objective lens was used to collect the backscattered light. The spectra were acquired with 5.84 cm−1 spectral width and at least 30 s exposure (Kojima et al., 2006[link]). The spectra had backgrounds subtracted but no other corrections were carried out.

The initial clustering is shown in Fig. 3.8.13[link](a) with the default cut level in the dendrogram. There are six clusters: labelling from the left-hand side, the red are three polymorphs of carbamazepine; the orange are cimetidene; the green cluster contains three polymorphs of furosemide; the light blue contains three polymorphs of mefenamic acid; the dark blue contains phenilbutazone; and finally the purple cluster contains sulfamerazine. The MMDS plot gives a complementary visualization of the data that supports the clustering.

[Figure 3.8.13]

Figure 3.8.13 | top | pdf |

(a) The dendrogram generated from 74 Raman spectra without background corrections applied. Labelling from the left-hand side, the red samples are carbamazepine, the orange are cimetidene, the green are two forms of furosemide, the light blue is mefenamic acid, the dark blue is phenilbutazone and the purple at the right-hand side is sulfamerazine. (b) The MMDS plot. The sphere colours are taken from the dendrogram. This representation shows clearly discrete clusters in correspondence with those generated by the dendrogram.

It is also possible to use derivative data in place of the original spectra for clustering. The results of this for the 74 Raman spectra without initial background subtraction followed by the generation of first-derivative data are shown in Fig. 3.8.14[link]. The clusters are well defined but now the carbamazepine data have split into two clusters. These correspond to forms I and III of carbamazepine, although the differences in the Raman spectra for these three species are small (O'Brien et al., 2004[link]). At the same time, both furosemide and mefenamic acid are each split into two groups. This is probably the best description of the data in terms of clustering and cluster membership corresponding to the chemical differences in the samples. The dendrogram also has the feature that the tie bars between samples are higher, i.e. the similarities are lower, reflecting the fact that the use of first derivatives accentuates small differences in the data.

[Figure 3.8.14]

Figure 3.8.14 | top | pdf |

Clustering the 74 Raman spectra without background corrections applied using first-derivative data. (a) The dendrogram. Labelling from the left-hand side, the red and orange entries are carbamazepine; the green are cimetidene; the light blue and dark blue are two forms of furosemide; the purple are sulfamerazine; the brown are phenilbutazone and the right-hand light and dark green are two forms of mefenamic acid. (b) The MMDS plot. The clusters are well defined but the orange and red (both carbamazepine) are very close to each other.

It is interesting to note that, in general, PXRD works less well with derivative data. The reason for this is not clear, but possibly the presence of partial overlapping peaks and the associated issues of peak shape are partly responsible.

References

Boccaleri, E., Carniato, F., Croce, G., Viterbo, D., van Beek, W., Emerich, H. & Milanesio, M. (2007). In situ simultaneous Raman/high-resolution X-ray powder diffraction study of transformations occurring in materials at non-ambient conditions. J. Appl. Cryst. 40, 684–693.Google Scholar
Kojima, T., Onoue, S., Murase, N., Katoh, F., Mano, T. & Matsuda, Y. (2006). Crystalline form information from multiwell plate salt screening by use of Raman microscopy. Pharm. Res. 23, 806–812.Google Scholar
Mehrens, S. M., Kale, U. J. & Qu, X. (2005). Statistical analysis of differences in the Raman spectra of polymorphs. J. Pharm. Sci. 94, 1354–1367.Google Scholar
O'Brien, L. E., Timmins, P., Williams, A. C. & York, P. (2004). Use of in situ FT-Raman spectroscopy to study the kinetics of the transformation of carbamazepine polymorphs. J. Pharm. Biomed. Anal. 36, 335–340.Google Scholar








































to end of page
to top of page