Raman spectroscopy has many applications within the pharmaceutical industry. It can be used to identify polymorphs, for example, and to analyze active pharmaceutical ingredient (API) forms and their distribution within formulated products. But what is it, and how can it be applied in practice?

In contrast to standard infrared (IR) spectroscopy, which identifies the specific frequencies of radiation that are absorbed by a sample, Raman spectroscopy studies the way light is scattered by molecules. As a laser beam passes through the sample, much of the light passes through and scatters with its energy unchanged; this is known as Rayleigh scattering.

However, some of its photons collide with the molecules and lose energy, in a phenomenon known as a Stokes shift. Others may pick up energy from excited molecules and emerge with a higher energy level, or an anti-Stokes shift. In Raman spectroscopy, the light that emerges is collected, and that which is scattered without changing energy is filtered out. What remains provides a unique spectral pattern for that individual molecule. This fingerprint can be used to identify the molecule by comparing the pattern to a known reference.

Raman spectroscopy complements IR spectroscopy, as it tends to give strong peaks at wavelengths where they are weak in IR spectra, and vice versa as shown in Figure 1, Raman can often give good results if there are no functional groups in a molecule that give a meaningful IR fingerprint.

There are several different spectroscopic techniques based on the Raman principle which can be applied to pharmaceutical products. Fourier transform Raman spectroscopy, for example, uses a back-scattering technique, with the radiation reflected back off the sample onto the detector rather than passing through the sample. It is suitable for both blends and ground-up tablets, and gives a representative spectrum of the entire sample.

Transmission Raman spectrometry, meanwhile, often gives better results when sampling solids than a conventional backscatter Raman technique as the radiation passes through the sample analyzing a much larger volume. As the technique is non-evasive and non-destructive, it can be used for the direct analysis of batches of hundreds of whole tablets or capsules that can be scanned in minutes, and can quantify both the API (down to less than 1% drug loading) and the excipient in a single measurement using appropriately developed partial least-squares calibration models.

A third option, Raman microscopy, is particularly suitable for more detailed analytical work. In this technique, the incident light is passed through an objective lens, and focused onto a very small spot. This allows resolution down to fractions of a micron to be achieved. The distribution of components within a sample can be determined in this way, and importantly, the laser can be focused on specific areas of concern. This may be to determine the presence and identification of a suspected contaminant, particle or other unexpected feature, and as such, Raman microscopy is much more sensitive than techniques used for the analysis of a material’s bulk properties.

Pinpointing polymorphs
As an example, Raman microscopy can be used to characterize amorphous materials produced by hot melt extrusion. In this example, dispersions of carbamazepine at 15, 30 and 45% by weight in a polymer carrier were examined. From an initial study using scanning electron microscopy (SEM), it was clear that more surface features appeared as drug loading increased. At 30%, long, needle-like filamentous particles were clear on the surface of the extrudate; and at 45%, while these were still visible, plate-like particles had also formed. These features needed to be determined and identified, as well as whether they may affect the product’s performance.

Raman microscopy is the ideal technique for this, and Figure 2 shows the results of the analysis of the different structures observed on the 45% dispersion’s surface. The traces in the top right of the figure give the full spectral range of the plate and ribbon structures, and also the solid solution. Expanding the right-hand sides of the spectra, it becomes clear that the peaks are broader when the carbamazepine is dispersed as a solid solution within the polymer, as might be expected for an amorphous material. Although the peaks for the plates and ribbons are in similar positions, they are narrower and slightly shifted from those for the solid solution. This slight shift in the position indicates that two different polymorphs had indeed formed on the surface of the extrudate.


A similar study, combining differential scanning calorimetry and Raman spectroscopy, has been reported.¹ This captured the Raman spectra of crystalline carbamazepine across increasing temperatures. At 30°C, the spectral pattern corresponds to form 3 of the API, which corresponds with the pattern observed for the plate form in the extrudate. Once the temperature reached 175 °C, it was consistent with form 1 of the API, matching that seen for the ribbons in the extrudate. It is clear that the crystalline API changes from polymorph form 3 to form 1 as the temperature rises.

Combining these data, the analysis shows that at a drug loading of above 30%, the API recrystallizes as either form 3 or form 1 within the dispersion, rather than remaining in the desired amorphous form. The drug loading of the hot melt extruded formulation was, therefore, adjusted to 20% to prevent this undesirable recrystallization.

This highlights the importance of asking the right questions if particles are observed on the surface of an extrudate. Is it a particle of API that has crystallized? Is it a contaminant? Is it merely a roughening of the surface? Raman microscopy is able to answer all of these questions, and therefore direct formulation decisions.

Investigating pellet microstructure
When formulating a drug, it is important to get the dissolution profile correct; and when a generic product is being developed, then the dissolution profile must match that of the originator. When that originator formulation has a specific dissolution-enhancing component, a generic developer needs to show that they have managed to replicate that modified profile in their version of the product.

The antifungal-medicine itraconazole is a good example of this situation. Janssen’s originator Sporanox dosage forms are manufactured by spray-coating a layer of the API in hydroxypropyl methylcellulose onto a sugar/starch core to enhance the solubility of the drug in the digestive tract. This is then sealed with the application of a polyethylene glycol coating. The drug is now available in generic form, yet comparisons of older generic versions with the original showed there are substantial differences in the rates of dissolution, and also the amount that dissolves in the first hour, between the different products in in vitro studies.

The cause of this dissimilarity was investigated using SEM and Raman spectroscopy. In the Raman spectrum of the Sporanox formulation, as shown in Figure 3, the peaks are approximately in the same position as they are for the crystalline reference, but a slight broadening of the peaks is observed because of its amorphous nature.


Pellets of the Sporanox formulation and also one of the generic forms were cross-sectioned, and the layers analyzed with both SEM and Raman spectroscopy. As expected, Sporanox exhibited three layers, and the drug-polymer layer was both uniform and smooth. This indicates that there was homogeneous drug-polymer distribution within the pellet, and there were no obvious defects. This was confirmed by Raman spectroscopy, looking specifically at the intensity of the peak at a wavelength of 1613 cm-1 over a 100 µm square area. This showed that there was a very uniform distribution, with only minimal spots where the drug content was slightly different.

However, when analyzing the generic products, it was very different. In one, it was immediately noticeable that the drug-polymer layer appeared very different from the originator in the SEM scan, with pores evident throughout, which were also seen in the Raman study. It could be assumed that parameters such as spray rates and temperature were sub-optimal during its manufacture, creating those pores. However, these pores alone would not explain the poor dissolution seen in the generics.

Looking more closely at the Raman spectrum, there were areas within the pores where there was a much stronger signal for the API at that 1613 cm-1 peak. The regions within these pores were further investigated, along with the area around them. As can be seen in Figure 4, the spectrum from within the pore (at the top) had a much better-defined peak structure, similar to that of the crystalline API. In the area between the pores, the peaks were broadened and there was evidence of a slight peak shift. This suggests that the API was recrystallizing between the pores, leading to the slower dissolution rate. This could also be explained by sub-optimal spraying parameters.


A better taste?
Slower dissolution rates are not the only problem with drug products that might be caused by ineffective spray-coating. If the coating is designed to prevent the API from being released in the mouth—particularly important for pediatric formulations—if the coating is ineffective, then an unpleasant-tasting API may be subjected to the patient.

As an example, a coated multi-particulate pediatric formulation of a bitter compound was studied to see if the coating was uniform and intact, and whether there was any migration of the API through the coating, which would render the taste-masking ineffective. Initially, optical microscopy was used to create an image of the edge of a sectioned pellet (approximately 600–700 µm in diameter) and to identify the different components of the pellet: the core, the API and the polymer coating.

Six distinct spectra were identified with Raman microscopy, as shown in Figure 5. At the top, there is a clear signal for the API. The second trace is the core material, microcrystalline cellulose (HCC), and then next comes the taste-masking coating polymer, Eudragit. Talc is also present as an anti-caking agent.


There were also spectra for some areas where the signal was mixed between the API and the binding polymer, hydroxypropyl methylcellulose (HPMC), and between the API and the Eudragit.

There are, of course, limits to how small the area of sample the laser can be focused on are, depending on the equipment used. In this example, it could focus down to approximately 1 µm, and within that spot, all molecules are exposed to the laser, giving emission signals for each component. It is therefore unsurprising that at the interface, both components are seen.

Raman mapping was then used to determine the distribution of components across the pellet layers. This is shown in Figure 6, and was created from 8400 individual spectra, clearly showing the layers. At the top in blue is the MCC core, followed by the API in red and the API–binder mix in magenta. These are about evenly distributed. A thin cyan layer below this is seen at the interface between the Eudragit and API layers, indicating that they are in very close contact with each other. If the coating is ineffective, then there may be poor adhesion, and gaps will be in evidence. The final layer is the mixture between coating polymer and anticaking agent.


In a separate batch of coated pellets, they stuck to each other during an accelerated stability study and Raman microscopy was used to investigate what was occurring. The analysis showed that rather than having the same even distribution of talc as seen in Figure 6, it was all deposited close to the interface of the API. When the spray-coating process was reviewed, it was found that the feed solution had not been adequately stirred throughout, allowing the talc to settle at the bottom rather than being applied correctly.

Raman mapping was an important tool in determining the homogeneity of the layers, and how the components were distributed within them. It also gave a valuable insight into the microstructure of the coated pellets, revealing an issue within the manufacturing process that affected quality. 

References

1. Simultaneous DSC‐Raman Analysis of a Pharmaceutical Polymorph, courtesy TA Instruments: http://www.tainstruments.com/pdf/Simultaneous_DSC-Raman_Analysis_of_a_Polymorphic_Transition.pdf