Discovery and Formulation - Emerging Raman Applications and Techniques in Biomedical and Pharmaceutical Fields - page 217

introduced by manufacturing processes. In some instances fluorescence may be
removed by photo-bleaching the sample (irradiating it with laser light prior to
analysis) [33]. Fluorescence may also be significantly reduced or even removed
by changing the excitation wavelength of the laser. The fluorescence emission
of aromatic molecules generally occurs in the 300-700 nm region; therefore,
fluorescence may often be removed by moving from a 633 nm to a 785 nm or
1064 nm excitation wavelength. The main disadvantage associated with mov-
ing to longer wavelength excitation is that the Raman intensity is inversely
proportional to the fourth power of the laser wavelength [34]. The Raman sig-
nal therefore decreases rapidly with increasing wavelength [32]. When using
FT-Raman spectroscopy (1064 nm lasers) the Raman signal strength is typ-
ically substantially lower due to both the wavelength dependence of Raman
scattering cross section and often lower performance of FT-Raman systems
compared with their dispersive counterparts typically used in the visible spec-
tral range. Hence the laser power must be often dramatically increased in or-
der to obtain good spectra. Thus, while fluorescence is rarely a problem when
using FT-Raman, sample burning and sample heating are common. This is
often overcome by using a spinning sample holder to disperse the laser energy
[24, 26].
Other sources of error, particularly in quantitative Raman analysis, include
laser self-absorption effects leading to attenuation of some spectral bands. Sim-
ilarly diffuse reflectance of the laser light, which is dependent on the particle
size of the formulation components, may increase or decrease the collection
volume. However, normalisation techniques can be used to overcome some of
these effects [35].
Another diculty is sub-sampling [23]. It is dicult to measure a statisti-
cally representative amount of sample when the focus of the laser beam is a
few tenths of a millimetre. This problem may be addressed by using a bulk or
wide-area sampling probe, by rotating the sample or by translating the sample
combined with multiple replicates [17]. Similarly, sample inhomogeneity can
be a major source of error when using Raman spectroscopy to analyse mix-
tures. A number of different approaches have been used to ensure complete
sample mixing including vibrating the samples and shaking the samples in a
ball mill without the steel balls [27].
In summary, the analyst has a number of options in deciding which
Raman method to use for their specific analysis. As a general rule of thumb
FT-Raman spectroscopy should be used for bulk samples (ca. > 1-10 mg sam-
ple). A dispersive Raman microscope system is generally the best option for
individual particles or small areas of larger samples.
9.4 API Form Behaviour and Relationships
Having identified and characterised all known polymorphs, hydrates and sol-
vates of an API it is important to determine the relationship between all
known forms. Conversion of one form to another can be caused by common