Recent Advances in Pharmaceutical Analysis Using Transmission Raman Spectroscopy

Apr 01, 2017
Volume 32, Issue 4, pg 37–43

 

This article reviews recent advances in the application of transmission Raman spectroscopy (TRS) to pharmaceutical analysis. The TRS technique overcomes subsampling limitations of conventional Raman spectroscopy and enables rapid noninvasive volumetric analysis of intact pharmaceutical tablets and capsules in a quantitative manner with relevance to quality and process control applications. Although only recently introduced to this area, this technique’s uptake and breadth of applications are rapidly growing, with regulatory approvals for its use in quality control of manufactured pharmaceutical products recently being granted.


In pharmaceutical process and quality control it is often highly desirable to perform rapid quantitative volumetric analysis of tablet and capsule content by intact and noninvasive means without any sample preparation. In this area, transmission near-infrared (NIR) absorption spectroscopy established itself in the last two decades as an effective analytical tool. However, its wider spread is being restricted in part by its limited chemical specificity. This restriction can be lifted in a number of applications by a more recently introduced alternative technique, transmission Raman spectroscopy (TRS). TRS has been demonstrated to provide much higher chemical specificity, which is beneficial particularly in situations involving more-complex formulations. This type of measurement can be accomplished typically with acquisition times of seconds.  The ability of the technique to provide well-defined Raman bands easily attributable to individual sample components also aids data interpretation and facilitates means for clearer communication of results to regulatory bodies. Additional advantages of TRS include the ability to probe samples in the presence of water, making the technique also suitable for analyzing slurries and suspensions in aqueous environments (1). On the other hand, the technique can suffer from the inability to probe samples exhibiting excessively high levels of fluorescence; however, this issue can be mitigated by using long laser excitation wavelengths—for example, 830 nm—to minimize the likelihood of electronically exciting fluorophores, which could potentially give rise to excessively high baseline noise within the region of the Raman spectra. Highly absorbing samples at the laser excitation wavelength or Raman emission spectral region (typically ~800–1000 nm) are also inaccessible by this technique although this situation is rather uncommon in drug-product analysis. Excessive heating and thermal damage (burning) could potentially occur with absorbing samples but this heating can be mitigated by using an expanded laser illumination beam (for example 4–8 mm diameter).

Conventional Raman spectroscopy is traditionally performed in backscattering geometry where the Raman signal is collected from within the laser-illuminated zone on the sample’s surface. This configuration is unable to probe the entire volume of pharmaceutical tablets or capsules, limiting its use to characterizing near-surface regions of sample around the illumination–collection zone (2,3) (so called subsampling), in the absence of coatings or capsules. The advent of TRS, a variant of Raman spectroscopy, where the sample is illuminated over an extended zone on one side and signal collected from the other (see Figure 1), has radically reduced this issue because laser photons propagate via diffusion through the entire body of the sample to be detected on the other side, and, as such, the generated Raman signal conveys information on its volumetric content, although some comparatively small bias toward the center of the tablet and away from sample edges remains (4–6), similar to transmission near-infrared absorption spectroscopy (7). 

Figure 1: Schematic diagram of conventional and transmission Raman spectroscopy.

 

The transmission Raman concept was demonstrated in the very early days of Raman spectroscopy (8), but until relatively recently its properties and benefits for the noninvasive probing of the volumetric content of pharmaceutical samples highlighted in the study of Matousek and Parker (4) followed by other pioneering work (9–11) had not been recognized and exploited in pharmaceutical analysis. The fundamental concept, related methodologies, and early developments of the TRS technique for pharmaceutical analysis is covered in detail in an earlier review (1). Here we focus primarily on more-recent advances of the technique and its applications in pharmaceutical analysis over the past 5 years.

API Quantification 

Early proof-of-concept TRS quantification studies (9–11), using less challenging formulations, were followed quickly by more-advanced applications involving more-complex formulations and exploiting a wider range of sample types, which expanded the boundaries of the technique’s capability and widened its application space (1). In this area, Lee and colleagues (12) demonstrated that TRS can be successfully applied to the analysis of pharmaceutical formulations in capsules of different colors benefiting from reduced fluorescence background interference from the capsules themselves (13) compared to conventional back-scattering measurements. The team demonstrated that a single calibration model developed from samples held in glass vials could be used to determine active pharmaceutical ingredient (API) concentrations of formulations contained in capsules of different colors, thus avoiding the laborious step of constructing individual models for each capsule color. This benefit was shown by collecting TRS spectra of binary mixtures of ambroxol and lactose in a glass vial and developing a partial least squares (PLS) model for the determination of ambroxol concentration on this set. This model was then directly applied to determining ambroxol concentrations of samples contained in capsules of four colors (blue, green, white, and yellow). Although the prediction performance was slightly degraded when the samples were placed in blue or green capsules because of the presence of residual fluorescence, accurate determination of ambroxol was generally achieved in all cases. The prediction accuracy was also investigated when the thickness of the capsule was varied.

Griffen and colleagues (14) demonstrated that TRS is a viable tool for content uniformity testing of all the constituents of a complex formulation consisting of five components (three APIs and two excipients). The nominal concentration of individual components in this study ranged from 1 to 85% (w/w). The calibration set consisted of 40 tablets and the developed PLS model then successfully predicted all the components in a set of 10 validation tablets covering five sample points. A single PLS model for all components and five individual models each optimized for one component performed similarly and has been used to demonstrate that specificity and robustness of prediction can be achieved through using a robust multifactor orthogonal design-of-experiments approach for calibration samples. The ability to determine multiple analyte concentrations in a single measurement highlights the potential of TRS for assay and content uniformity testing.

To enhance TRS signals and permit further reduction of acquisition times, Griffen and colleagues (15) employed a photon-beam enhancing element (16) before the laser illumination zone to prevent loss of laser radiation from the sample. The study used a five-component formulation. The photon enhancing element permitted improved speed of acquisition by an order of magnitude compared with conventional TRS measurements. The three APIs and two excipients were used with nominal concentrations ranging between 0.4 and 89%. Acquisition times as short as 0.01 s per tablet were reached with acceptable performance for all the sample components. Results suggest that even faster sampling speeds could be achieved for components with stronger Raman scattering cross sections or with higher laser powers. This major improvement in speed of quantification opens exciting prospects for high-throughput TRS in-line analysis in quality control applications within a batch or continuous manufacturing process.

Peeters and colleagues (17) attempted to apply TRS and NIR spectroscopy to the prediction of physical properties of pharmaceutical tablets (as opposed to chemical characterization). Granules were produced on a continuous line by varying granulation parameters. Tableting process parameters were adjusted to obtain uniform tablet weight and thickness. PLS regression was used to correlate spectral information to tablet physical properties (friability, tensile strength, porosity, and disintegration time), but, in this study, no predictive models could be established, which indicated the insensitivity of the methods to these physical properties of sample. This result is in contrast with some earlier studies where such physical properties could be detected in TRS and other data (17). This difference is tentatively attributed to the fact that the previous studies used tablets of varying thickness, whereas in this study the tablet thickness was kept constant (4 mm), suggesting that perhaps tablet thickness variations could be responsible for the earlier observed sensitivity to physical properties of tablets. This hypothesis was not confirmed. Principal component analysis (PCA) was effectively used to distinguish theophylline concentrations and hydration levels and multiple linear regression (MLR) analysis provided insight about how granulation parameters affect granule and tablet properties. All the spectroscopic methods revealed a similar prediction performance with an RMSEP around 1%.

Li and colleagues (18) described the development and validation of a TRS method using the ICH-Q2 guidance as a template. Niacinamide content in tablet cores was determined in this study. The resultant model statistics were evaluated along with the linearity, accuracy, precision and robustness. Method specificity was demonstrated by accurate determination of niacinamide in the presence of niacin (an expected related substance). The method was demonstrated as fit for purpose with added benefit of very short analysis times (~2.5 s per tablet). The resulting method was used for routine content uniformity analysis of single dosage units in a stability study.

Bilayer Tablets

Zhang and colleagues (19) applied TRS to the quantification of bilayer tablets. This area has so far received only cursory attention and represents an unusually challenging situation. A variety of tablet configurations were examined in this study. A complex spectral response was observed and was effectively modeled using a modified Schrader, Kubelka-Munk model in which both the Raman photon generation factor and photon losses were accounted for. Coupling the results of these studies together yields a comprehensive approach for modeling multicomponent bilayer tablets. The addition of a photon-beam enhancer on the bottom (illumination) surface allowed for a selective over-enhancement of the bottom layer, which aided the analysis of thin layers or coatings.

Quantification of Polymorphs in Pharmaceutical Formulations

Polymorphs are very important in pharmaceutical manufacturing because they determine physiological dissolution rates and their control is therefore often critical. In vibrational spectroscopy, polymorphic forms are best represented by low-wavenumber vibrational (phonon) modes accessible directly by Raman or terahertz spectroscopy. Following an early demonstration of TRS capability in this area by Aina and colleagues (20) to quantify polymorphic content of a binary pharmaceutical formulation, more studies have emerged in this area, setting a clear and important niche for TRS not deliverable by high performance liquid chromatography (HPLC) (where a dissolution step destroys such information) and accessible only partially by NIR, which cannot access phonon modes directly but can sense polymorphs by leveraging changes in overtone and combination bands if such changes are sufficiently pronounced. X-ray diffraction (XRD) is a workhorse here but it also suffers from limited sensitivity and inability to probe the entire sample depth effectively. The availability of the low-wavenumber region simultaneously with the fingerprint region in TRS provides a distinct advantage to TRS over terahertz spectroscopy where only low-wavenumber bands are detected, which therefore enables more straightforward assignment and interpretation of data. Other alternative methods, such as nuclear magnetic resonance (NMR) or differential scanning calorimetry, suffer from limited sensitivity, need sample preparation, or require long data acquisition times. 

The first demonstration of TRS for the analysis for polymorphic systems was performed on a binary mixture by Aina and colleagues (20). The study was followed by McGoverin’s (21) research performing the quantification of polymorphs in more-complex dosage forms. The efficacy of TRS measurements for the prediction of polymorph content was evaluated using a ranitidine hydrochloride test system. Four groups of ranitidine hydrochloride-based samples were prepared: three containing form I and II ranitidine hydrochloride and microcrystalline cellulose (spanning the ranges 0–10%, 90–100%, and 0–100% form I fraction of total ranitidine hydrochloride), and a fourth group consisting of a form I ranitidine hydrochloride (0–10%)-spiked commercial formulation. Transmission and conventional Raman spectroscopic measurements were recorded from both capsules and tablets of the four sample groups. Prediction models for polymorph and total ranitidine hydrochloride content were more accurate for the tablet than for the capsule systems. TRS was found to be superior to conventional backscattering Raman spectroscopy for the prediction of polymorph and total ranitidine hydrochloride content. 

Hennigan and Ryder (22) also carried out a study in which they generated a model tablet system with two excipients at 10% API concentration. The API was a mixture of the FII and FIII polymorphs of piracetam. The formulation was characterized using TRS, conventional backscattering Raman spectroscopy, and NIR spectroscopy. The team demonstrated that it is possible to detect FII polymorph contamination in these model tablets with limits of detection (LODs) of 0.6 and 0.7%, respectively, with respect to the total tablet weight (or ~6–7% of the API content). The TRS method was shown to be the superior method because of its higher speed of analysis (~6 s per sample), better sampling statistics, and the availability of sharper, more-resolved bands in the Raman spectra, which enable easier interpretation of the spectral data. An additional benefit highlighted was the direct access of TRS to the low-frequency (phonon) wavenumber region.

The sensitivity of TRS was further enhanced in a study by Griffen and colleagues (23), who performed a proof of concept study using a commercial TRS instrument with an excitation wavelength of 830 nm, demonstrating the application of TRS to the noninvasive and nondestructive quantification of low levels (0.62–1.32% w/w) of an active pharmaceutical ingredient’s polymorphic forms in a pharmaceutical formulation. PLS calibration models were validated with independent validation samples resulting in root mean square error of prediction (RMSEP) values of 0.03–0.05% w/w and a limit of detection of 0.1–0.2% w/w. The study also demonstrated the ability of TRS to quantify all tablet constituents in a single measurement. The team also performed TRS analysis on degraded stability samples for which transformation between polymorphic forms was observed while excipient levels remained constant. Additionally, the authors demonstrated a dramatically enhanced collection speed for TRS measurements by deploying a beam enhancing element that permitted comparable prediction performance at 60 times faster rates (for example, 0.2 s per measurement) than in standard mode.

Vigh and colleagues (24) applied TRS to the estimation of degraded drug percentage, residual drug crystallinity, and glass-transition temperature in the context of melt-extrusion. Tight correlation was shown to exist between the results obtained by confocal Raman mapping and TRS. The investigation of the relationship between process parameters, residual drug crystallinity, and degradation was performed using statistical tools and a factorial experimental design defining 54 different circumstances for the preparation of solid dispersions. Drug content, temperature, and residence time were found to have a significant and considerable effect on the examined factors. By forming physically stable homogeneous dispersions, the originally very slow dissolution of the components was improved, making 3-min release possible in acidic medium.

Determination of Crystalline Content in Amorphous Formulations

Kumar and colleagues (25) investigated the applicability of TRS to characterizing the residual crystallinity in an amorphous material. Amorphous materials are sometime used to enhance oral bioavailability of poorly water-soluble drugs. However, amorphous forms are thermodynamically driven to crystallize to the less soluble forms during processing or storage. Sensitive quantitation of crystallinity is therefore critical in process control and in monitoring the stability and bioavailability of amorphous drug product. The study compared TRS with X-ray powder diffraction and solid-state NMR spectroscopy (ssNMR) approaches to quantification of low levels of crystalline material in an amorphous spray-dried dispersion with a moderate 20% drug load. TRS was demonstrated as a viable alternative for detecting and quantifying the crystalline form of an API in an amorphous solid dispersion. It was shown that TRS has better sensitivity compared to the other techniques with an added benefit of faster acquisition times.

Cocrystals

The high chemical specificity of Raman spectroscopy renders the technique amenable to characterizing more-complex forms of APIs and formulations such as cocrystals. TRS was demonstrated and compared with established technique to be viable to detect cocrystals in a study by Elbagerma and colleagues (26). The study confirmed that TRS is an effective TRS tool to evaluate cocrystal formation through interaction of their components. Burley and colleagues (27) further employed TRS to analyze model formulations comprising tableted cocrystals. The ability of TRS to differentiate between formulations on the basis of both drug loading and drug chemistry (cocrystal versus separate components) was confirmed. It was concluded that TRS allows for fast, automated, unsupervised classification of realistic cocrystal tablet formulations, both in terms of drug loading and in terms of whether two APIs are included as a cocrystal or as separate components. The study also warned that mere visual inspection of data to identify variance and overall data structure could be misleading, and in the concerned study such cursory inspection, in effect, misidentified API chemistry (cocrystal versus “separate components”) as the main variance in the data. On the other hand, rigorous quantitative numerical analysis showed that API content in fact leads to the largest variations between datasets. 

Suspensions

Shin and colleagues (28) demonstrated the ability of TRS to assess suspensions with minimal influence from internal particle settling. Because particle concentrations at given points throughout the sample can differ in a partially settled suspension sample, the acquisition of Raman spectra representative of the entire sample composition is critically important for accurate quantitative analysis. The proposed scheme used axially irradiated laser radiation (TRS) in the same or opposite direction of settling, thus allowing laser photons to migrate through the settling-induced particle-density gradient formed in the suspension and to widely interact with particles regardless of their settled locations. As such TRS was expected to be more representative of the overall suspension composition, even with partial settling, compared with conventional localized Raman measurement. The study showed that settling did not significantly degrade the accuracy of the concentration determination, thereby indicating effective acquisition of settling-tolerant TRS spectra. 

Recent Advances of the TRS Technique

Following earlier investigations of basic properties of the technique in both the lateral and depth dimensions (29,30), the underpinning TRS technology and understanding of the technique itself have been advancing rapidly. Considerable efforts have been devoted to advances such as boosting signals in Raman spectroscopy, which in general are weaker in TRS geometry than in conventional backscattering Raman configuration. Earlier and subsequent investigations showed that the photon diode concept, which consists of an optical filter placed over an illumination zone and returning laser and Raman photons escaping from the turbid sample back into it, can boost Raman signals in TRS by an order of magnitude (15,16,23). Subsequently, an alternative concept using hemispherical mirrors placed over the illumination zone was demonstrated and investigated by Pelletier (31). This approach was shown to be capable of boosting a TRS signal from a representative commercial pharmaceutical tablet by a factor of 40. These approaches promise to further enhance the potential of TRS by shortening acquisition time and enhancing measurement sensitivity (15,23).

Other methods for enhancing signals include increasing the ability of the Raman detection system to collect more Raman signal from the sample. Because the diffuse TRS signal is spread over a wide area of sample, its collection is highly challenging and in practical scenarios only a small fraction of signal available on sample surface can be collected. The limit is typically set by the signal collection gathering capability of spectrographs (etendue), which represents a bottleneck. Strange and colleagues (32) investigated the use of alternative spectrometer approach based around spatial heterodyne Raman spectrometry (SHRS), which enables them to  to achieve potentially much higher etendue values. The SHRS detection concept was applied to measuring TRS spectra of ibuprofen yielding only 2.4 times lower TRS intensity than in backscattering mode. The throughput of the SHRS system was about eight times higher than that of an f/1.8 dispersive spectrometer. However, the signal-to-noise ratio (S/N) was still two times lower for the SHRS than the f/1.8 dispersive spectrometer, apparently because of high levels of stray light. As such, there is the possibility for further improvement of the SHRS technology with a potential to outperform conventional dispersive systems in terms of S/N. Another notable attractive feature of this technology is its potential for miniaturization compared with dispersive spectrometer technology. 

Oelkrug and colleagues (33) studied the origins and dependencies of TRS (XT) and conventional backscattering Raman (XR) signals both theoretically and experimentally as a function of sample thickness, absorption, and scattering. The study showed that for nonabsorbing layers, the Raman reflection and transmission intensities rise steadily with the layer thickness, starting for very thin layers with the ratio XT/XR = 1 and for thick layers, approaching a lower limit of XT/XR = 0.5. In stratified systems, Raman transmission allows deep probing even of small quantities in buried layers. In double layers, the information is independent from the side of the measurements. In triple layers simulating coated tablets, the information of XT originates mainly from the center of the bulk material whereas XR highlights the irradiated boundary region. However, if the stratified sample is measured in a Raman reflection setup in front of a white diffusely reflecting surface, the study concludes that it is possible to monitor the whole depth of a multiple scattering sample with equal statistical weight. In other words, the approach enables transmission-like measurements from reflectance setups.

Sparen and colleagues (34) investigated the dependence of the accuracy of quantification of TRS signals on matrix properties and compared it with that of NIR spectroscopy. In this work, matrix effects in transmission NIR and Raman spectroscopy were systematically investigated for a solid pharmaceutical formulation varying the factors particle size of the drug substance, particle size of the filler, compression force, and content of drug substance. Principal component analysis of NIR and Raman spectra showed that the drug substance content and particle size, the particle size of the filler, and the compression force affected both NIR and Raman spectra. All factors varied in the experimental design influenced the prediction of the drug substance content to some extent, both for NIR and Raman spectroscopy, with the particle size of the filler having the largest effect. When all matrix variations were included in the multivariate calibrations, however, good predictions of all types of tablets were obtained, for both NIR and Raman spectroscopy. The prediction error using transmission Raman spectroscopy was about 30% lower than that obtained with transmission NIR spectroscopy.

 

Manufacturing and Regulatory Perspective

Villaumié and colleagues (35) described the application of TRS to manufacturing at Actavis UK Ltd. The company undertook a project to modernize how samples were tested in the QC laboratory and to make significant improvements to the supply chain and speed up the process of product supply to the market. After the researchers reviewed the technologies that can nondestructively test solid oral medicinal products, they identified TRS as the most chemically specific analytical option that avoids subsampling issues from the tablets or capsules; the technique eliminates sample standard preparation time, with routine sample analysis in just minutes, and has an added benefit of eliminating environmental waste. The article describes how approval for the use of TRS as an alternative content uniformity (CU) test was acquired from UK’s Medicines and Healthcare Products Regulatory Agency (MHRA). By removing dependence on HPLC, an expensive and time-consuming “workhorse” technique, the company was able to save cost and resources from the QC laboratory and eliminate environmental waste. The study details the method development process—from initial feasibility studies, calibration, and validation to routine use of the model—and describes the regulatory framework for their successful method approval. The company is actively developing additional TRS methods for different solid-dose products with the aim of reducing laboratory costs and time for release of products to the market (36).

Conclusions

The recent advent of TRS for pharmaceutical analysis heralds a new era in intact rapid volumetric analysis of pharmaceutical products in process and quality control. A number of advanced diverse applications have been demonstrated in recent years, including API content uniformity testing of complex formulations, polymorph quantification, and detection and excipient characterization. Beneficial characteristics include experimental simplicity, ease of data interpretation, and the ability to use existing multivariate data analysis tools. Several new regulatory approvals granted recently indicate the initiation of the process of the translation of TRS technology from pharmaceutical R&D laboratories to manufacturing environment.

References

  1. K. Buckley and P. Matousek, J. Pharm. Biomed. Anal. 55, 645–652 (2011).
  2. H. Wang, C.K. Mann, and T.J. Vickers, Appl. Spectrosc. 56, 1538–1544 (2002).
  3. J. Johansson, S. Pettersson, and S. Folestad, J. Pharm. Biomed. Anal. 39, 510–516 (2005).
  4. P. Matousek and A.W. Parker, Appl. Spectrosc. 60, 1353–1357 (2006).
  5. N. Townshend, D. Littlejohn, A. Nordon, M. Myrick, J. Andrews, and P. Dallin, Anal. Chem. 11, 4671–4676 (2009).
  6. J. Johansson, O. Svensson, S. Folestad, A. Sparen, and M. Claybourn, “Transmission Raman Spectroscopy for Robust Tablet Assessment,” FACSS Conference Proceedings, Abstract 300 (Louisville, Kentucky, 2009), p. 131.
  7. N. Kellichan, A. Nordon, P. Matousek, D. Littlejohn, and G. McGeorge, Appl. Spectrosc. 68, 383–387 (2014).
  8. B. Schrader and G. Bergmann, Z. Anal. Chem. Freseniu. 225, 230–247 (1967).
  9. J. Johansson, A. Sparen, O. Svensson, S. Folestad, and M. Claybourn, Appl. Spectrosc. 61, 1211–1218 (2007).
  10. C. Eliasson, N.A. Macleod, L.C. Jayes, F.C. Clarke, S.V. Hammond, M.R. Smith, and P. Matousek, J. Pharm. Biomed. Anal. 47, 221–229 (2008).
  11. M. Hargreaves, N.A. Macleod, M.R. Smith, D. Andrews, S.V. Hammond, and P. Matousek J. Pharm. Biomed. Anal. 54, 463–468 (2010).
  12. Y. Lee, J. Kim, S. Lee, Y.-A. Woo, and H. Chung, Talanta 89, 109–116 (2012).
  13. P. Matousek and A.W. Parker, J. Raman Spectrosc. 38, 563–567 (2007).
  14. J. Griffen, A.W. Owen, and P. Matousek, J. Pharm. Biomed. Anal. 115, 277–282 (2015).
  15. J. Griffen, A.W. Owen, and P. Matousek, Analyst 140, 107–112 (2015).
  16. P. Matousek, Appl. Spectrosc. 61, 845–854 (2007).
  17. E. Peeters, A.F. Tavares da Silva, M. Toiviainen, J. Van Renterghem, J. Vercruysse, M. Juuti, J.A. Lopes, T. De Beer, C. Vervaet, and J.-P. Remon,  Asian J. Pharm. Sci. 11, 547–558 (2016).
  18. Y. Li, B. Igne, J.K. Drennen, and C.A. Anderson  Int. J. Pharm. 498, 318–325 (2016).
  19. Y. Zhang and G. McGeorge, J. Pharm. Innov. 3, 269–280 (2015).
  20. A. Aina, M.D. Hargreaves, P. Matousek, and J.C. Burley, Analyst 135, 2328–2333 (2010).
  21. C. McGoverin, M.D. Hargreaves, P. Matousek, and K.C. Gordon, J. Raman Spectrosc. 43, 280–285 (2012).
  22. M.C. Hennigan and A.G. Ryder, J. Pharm. Biomed. Anal. 72, 163–171 (2013).
  23. J.A. Griffen, A.W. Owen, J. Burley, V. Taresco, and P. Matousek, J. Pharm. Biomed. Anal. 128, 35–45 (2016).
  24. T. Vigh, G. Drávavölgyi, P.L. Sóti, H. Pataki, T. Igricz, I. Wagner, B. Vajna, J. Madarász, G. Marosi, and Z.K. Nagy,  J. Pharm. Biomed. Anal. 98, 166–177 (2014).
  25. A. Kumar, L. Joseph, J. Griffen, W. Jenny, T. Chi, Y.J. Hau, M. Bloomfield, P. Matousek, and L. Wigman,  Am. Pharmacetical Rev. 19(1),(2016).
  26. M.A. Elbagerma, H.G. M. Edwards, T. Munshi, M.D. Hargreaves, P. Matousek, and I.J. Scowen, Cryst Growth Des. 10, 2360–2371 (2010).
  27. J.C. Burley, A. Alkhalil, M. Bloomfield, and P. Matousek, Analyst 137, 3052-3057 (2012).
  28. K. Shin, P.K. Duy, S. Park, Y.-A. Woo, and H. Chung, Analyst 139, 2813–2822 (2014).
  29. N. Everall, P. Matousek, N. Macleod, K.L. Ronayne, and I.P. Clark, Appl. Spectrosc. 64, 52–60 (2010).
  30. N. Everall, I. Priestnall, P. Dallin, J. Andrews, I. Lewis, K. Davis, H. Owen, and M.W. George, Appl. Spectrosc. 64, 476–484 (2010).
  31. M.J. Pelletier, Appl. Spectrosc. 67, 829–840 (2013).
  32. K.A. Strange, K.C. Paul, and S.M. Angel, Appl. Spectrosc. Published online 3702816654156 (2016).
  33. D. Oelkrug, E. Ostertag, and R.W. Kessler, Anal. Bioanal. Chem. 405, 3367–3379 (2013).
  34. A. Sparén, M. Hartman, M. Fransson, J. Johansson, and O. Svensson, Appl. Spectrosc. 69, 580–589 (2015).
  35. J. Villaumie and H. Jeffreys, Eur. Pharm. Rev. 20, 41–45 (2015)
  36. Actavis UK Describes their Regulatory Approval Process for Content Uniformity Testing using Transmission Raman Spectroscopy https://www.cobaltlight.com/news/Actavis-UK-Describes-their-Regulatory-Approval-Process-for-Content-Uniformity-Testing-using-Transmission-Raman-Spectroscopy

Julia A. Griffen, Andrew W. Owen, and Darren Andrews are with Cobalt Light Systems Ltd. Pavel Matousek is with Cobalt Light Systems Ltd. and the Central Laser Facility at STFC Rutherford Appleton Laboratory on the Harwell Campus (UK). Please direct correspondence to: [email protected]

 

native1_300x100
lorem ipsum