Pharmaceutical Applications of Raman Spectroscopy

Introduction

Raman spectroscopy is becoming one of the most popular analytical measurement tools for pharmaceutical applications ranging from verification of raw materials to process monitoring of drug production to quality control of products. Similar to an infrared spectrum, a Raman spectrum consists of a wavelength distribution of peaks corresponding to molecular vibrations specific to the sample being analyzed (see Figure 1B). Chemicals, such as drugs, can be identified by the frequency and quantified by the intensity of the peaks. In practice, a laser is focused into the sample, the inelastic scattered radiation (Raman) is optically collected and directed into a spectrometer, which provides wavelength separation, and a detector converts photon energy to electrical signal intensity. An attractive advantage to this technique is that samples do not have to be extracted or prepared, and the laser can simply be aimed at a sample to perform chemical measurements, which can often be accomplished in a minute or less.

A Look Back

The ease of using Raman spectroscopy has been made possible through numerous technological innovations since its prediction and discovery in 1928 [1-3]. The first major breakthrough came with the development of the laser, which provided considerably more photons to generate Raman scattered photons and therefore improved sensitivity [4]. In the mid-1970s, array detectors were beginning to replace photomultiplier tubes that were used with scanning spectrometers [5]. These, as well as the 2-dimensional charge coupled device (CCD) detectors that soon followed, reduced the measurement time from hours to minutes. These detectors also allowed replacing strip-chart recorders with XY plotters. In the 1980s, fiber optic probes were introduced [6, 7], which allowed the first process measurements [8]. Interferometer-based systems that employed 1064 nm lasers were also introduced during this time period [9]. Because very few chemicals have electronic absorptions at this longer wavelength, the generation of fluorescence that could obscure the Raman signal was virtually eliminated. This decade also saw the rapid development of the personal computer, which quickly became an important tool to analyze the spectra once collected. The 1990s introduced two new optical elements that simplified Raman spectrometer optical designs. Notch filters eliminated the need for large or multi-stage spectrometers to physically separate the excitation laser Rayleigh scattering from the Raman scattering [10], and sharp optical cut-off filters allowed the design and use of 180° backscattering single ended probes. In addition, diode lasers that were developed for the telecommunication industry were introduced. These power-efficient and stable lasers quickly replaced the traditional large Ar⁺, Kr⁺, and Nd:YAG lasers [11]. The combination of diode lasers, optics, CCD detectors, and laptop computers led to the first truly portable systems, with 785 nm laser excitation as the most popular [12]. The stability of these systems also allowed the development of the first dedicated process systems, while the size allowed development of Raman microscope systems [13]. During the present century, the miniaturization of Raman spectrometers continued as briefcase systems [12, 14], were replaced by handheld systems [15, 16]. Computing power also allowed the application of statistics to large spectra data sets, and during the last 25 years, chemometrics has become an important part of Raman spectral analysis [14, 17].

Spectrometer Designs

As indicated, these technological advances have led to a variety of spectrometer designs, each with advantages that may be important for a particular application. For example, interferometer-based Raman spectrometers off er high resolution and an invariant x-axis, along with fluorescence-free spectra. A stable x-axis allows monitoring long chemical reactions without constant calibration and performing complex chemometrics without fear of model failure. Dispersive-based Raman spectrometers off er high sensitivity and simple optical designs. The use of 785 nm lasers allow using Si detectors, which are much more efficient than the InGaAs detectors that are required when using 1064 nm lasers. The increased sensitivity allows measuring spectra in seconds not minutes, while the optical designs allow for light-weight, handheld spectrometers. Recently, several companies introduced dispersive-based Raman spectrometers that employ 1064 nm lasers and InGaAs array detectors. These spectrometers take advantage of both designs, in that the longer excitation virtually eliminates fluorescence interference, while the optics allow the production of handheld systems.

During the past decade, Raman spectroscopy has been applied to a wide range of pharmaceutical applications [18-21]. These include optimizing polymorph (or crystalline structure) analysis [22, 23], the synthesis of new drugs, gaining process understanding through real-time monitoring of reactors, crystallizers, and blenders [24-26], verifying raw materials [27] and assuring product quality [28]. Pharmaceutical companies are often reluctant to publish their results as it could weaken their business position, so representative examples for several of these applications, as well as two emerging applications, verifying product authenticity and product shelflife are provided here.

Drug Development

Once a potentially new drug is identified, the method to synthesize the drug is developed. Raman spectroscopy is ideal for monitoring reactant, intermediate and product concentrations, determining pathways, kinetics, mechanisms, end-points, and yields for a variety of reaction types, such as Diels-Alder, Fischer esterification, Grignard, and hydrogenation. Fischer esterification is used in the synthesis of many drugs, such as benzocaine, but this reaction is more often used as an intermediate step to protect carboxylic acid groups, while functionalizing phenyl rings. As an example of this important pharmaceutical reaction, the esterification of benzoic acid was performed to produce methyl benzoate (Figure 1A). The reactant and product have unique spectra, with peaks at 780 and 817 cm^-1, respectively, which are ideal for real-time monitoring (Figure 1B). A fiber optic coupled emersion probe inserted into a 3-neck flask was used to collect spectra every 45 seconds using 0.5W of 1064 nm laser excitation. By fitting the change in the normalized peak intensities as a funtion of time (dLn/dt), the reaction rate, rate constant, and yield can easily be determined; while optimizing the effects of reactor temperature, catalyst concentration and type can be used to determine the activation energy and reaction endpoint, as well as optimize yield (Figure 1C).

Figure 1 - A) Esterification of Benzoic Acid, B) Raman spectra of Benzoic Acid and Benzoate (500 mW of 1064 nm LASER excitation), and C) Plot and fit of their respective normalized peak intensities as a funtion of time (dIn/dt). 90% yield was achieved at 60°C in 70 minutes.

Crystallization, often the final step in drug synthesis, is used to separate a drug from a solvent matrix so that it is suitable for final form manufacturing. This process must also be optimized, not only to maximize separation, but, in many cases, to ensure that the correct polymorph is formed. For some drugs, various polymorphs exist with dramatically different solubilities that affect bioavailability, an important consideration in defining dosage. Process conditions, such as temperature, mixing rate, and concentrations can affect crystalline formation kinetics, and which polymorph dominates. Again, Raman spectroscopy is well-suited for understanding and optimizing process conditions. In fact, drug synthesis and crystallization are often carried out in a single batch reactor. Figure 2A shows the Raman spectra for the initial reactants and the final crystalline product (proprietary). In this case, the product has a unique Raman peak at 1150 cm^-1, suitable for monitoring formation, while the shift in the peak ar 1234 to 1240 cm^-1 can be used to monitor crystallization. Although these spectral changes can be used, a chemometrics model that correlates these properties to the entire Raman spectrum allows monitoring both the synthesis and crystallization process with high accuracy and precision (green trace, Figure 2B). As previously stated, FT-Raman analyzers provide the necessary x-axis stability for reactions that take hours, however, temperature controlled, temperature immune optics, or inclusion of an x-axis reference in dispersive-based Raman analyzers can also be used.

Figure 2 - A) Raman spectra of the reactant mixture and the crystallized product (500 mW of 1064 nm LASER excitation, 5-min acquisition). The peak height is used to monitor product formation, while the shifts are used to follow chrystallization. B) Plot of the 1150 cm^-1 peak height and a chemometrics correlation to both product formation and crystallization (onset initiated by a temperature drop).

Drug Quality

Quality-by-design in drug manufacturing begins with verification of the purity of raw materials and ends with the quality of the product. The latter requires assurance that the product, as a gel cap, tablet, etc., contains the correct amount of active, its polymorph (if appropriate), excipient and other additives (e.g. dyes). Raman spectroscopy has been used to monitor mixing in blenders [25], as well as to inspect individual products before shipment. An Excedrin^® tablet is composed of three APIs; aspirin, acetaminophen, and caffeine, ideally at 44, 44, and 12%, respectively. The composition can be determined by fitting a measured Raman spectrum with the spectra of the pure APIs and normalizing the result to 100%, as shown in Figure 3A and 3B. Tablets are non-uniform and a single point measurement can produce misleading results. For example, measurement of a single 300 micron diameter spot indicated a composition of 45% aspirin, 22% acetaminophen, and 33% caffeine. To overcome this inaccurate result, several approaches have been developed [29-34]. These include mapping the sample using a raster or pattern approach, spinning the tablet, employing a large spot size or transmission Raman. It is worth noting that the entire pill does not have to be mapped. Figure 3C shows the concentrations for 8 points forming a circle on the pill, while Figure 3D shows the running average for three 8-point concentric circles. As demonstrated, the concentrations tend to stabilize at 43% aspirin, 45% acetaminophen, and 12% caffeine after ~20 measurements.

Figure 3 - A) Raman spectra of the components in Excedrin^®: Aspirin, Acetaminophen, and Caffeine. B) Overlay of a measured Raman spectrum for a spot on an Excedrin^® tablet (500 mW of 1064 nm LASER excitation, 5-min acquisition), a spectrum reconstructed from the components. C) Intensites for Aspirin, Acetaminophen, and Caffeine at 8 spots on an Excedrin^® tablet normaized to total 100%. D) Plot of running average of normalized intenisties for Aspirin, Acetaminophen, and Caffeine for 24 spots. The most innacurate spot was chosen as the starting spot to accentuate the capability of this approach.

Product Authentication

Counterfeit drugs have become a significant problem during the past decade. The availability of such drugs has been made possible largely through purchases from fraudulent websites. The World Health Organization defines a counterfeit medicine as “one which is deliberately and fraudulently mislabeled with respect to identity and source” [35]. Counterfeit drugs range from those employing incorrect ingredients, no actives (e.g. sugar pills), or insufficient actives. The latter are the most challenging since a simple compositional analysis may pass the sample as the genuine product. One of the most commonly copied drugs is Pfizer’s Viagra^®, used to treat erectile dysfunction. Authentic and counterfeit products have been examined by many analytical methods, including infrared, near-infrared, and Raman spectroscopies [36, 37]. A key consideration for Raman analysis is the blue dye, indigo carmine aluminum lake, used in the coating. Dye coatings are often used along with size and shape to provide product uniqueness for easy recognition by the pharmacist and consumer alike. In the case of Viagra^®, Raman spectra measured using 785 nm laser excitation generates fluorescence that obscures the low wavenumber end of the Raman spectrum. This is unfortunate, as this region is important to the differentiation of authentic from counterfeit Viagra^®. This interference can be overcome by using 1064 nm excitation as shown in Figure 4A. However, this comes at cost. The first measurement required 10 seconds, while the latter, which used an interferometer-based spectrometer, required 25 minutes. Fortunately, the introduction of dispersive-based Raman spectrometers using the longer laser wavelength provides an ideal compromise. Spectra with high signalto- noise ratios can be obtained in 20 seconds as shown in Figure 4B. The counterfeit drug spectra show additional peaks at 381 and 438 cm^-1 that can be easily used to differentiate counterfeits from originals. To further illustrate this capability, principal component analysis was used to separate authentic Viagra^® from the counterfeits (3 measurements per tablet). It is also worth noting that the ratio of the Viagra^® active ingredient, sildenafil citrate to the titanium dioxide peak intensities, e.g. 1234 to 639 cm^-1, respectively, indicates that one of the counterfeits only contains ~50 mg of the active as opposed to the 100 mg indicated.

Figure 4 - Raman spectra of Viagra^® using 785 nm (top spectrum) and 1064 nm excitation (black). Conditions: 300 mW, 10-sec accumulation, and 500 mW and 25- min accumulation, respectively. Both at 8 cm^-1 resolution. B) Raman spectra of two counterfeit and authentic (bottom spectrum) Viagra^® samples Conditions: 500 mW at 1064 nm, 20-sec accumulation, 12-20 cm^-1 resolution. C) Photograph of two counterfeits and an authentic Viagra^® tablet. D) Principal Component plot illustrating ability to differentiate authentic from counterfeit Viagra^®.

Product Shelf Life

Drug formulations include additives and coatings to minimize degradation of the active ingredient due to heat, moisture and radiation, and thereby maximize product shelf-life. The shelf-life, or more specifically, the expiration date, is based on the time that a drug maintains greater than 90% potency. Most drugs fall into two categories, those that maintain potency for at least one year, and those that maintain potency for greater than two years from the time of manufacture. While most drug degradation products are benign and simply ineffective, acetaminophen, one of the most popular and effective drugs for pain relief, degrades into a poison, p-aminophenol, which can cause liver damage [38]. Acetaminophen is the number one over-the-counter drug associated with accidental overdose and death [39]. Although unproven, some of these accidents may be due to use of expired drugs. The primary method used today to determine drug degradation is high performance liquid chromatography (HPLC) [40]. The advantages of Raman spectroscopy, (minimal or no sample preparation, non-destructive analysis and speed), suggest that it may also find value for this important application. The Raman spectra of acetaminophen and p-aminophenol are unique (Figure 5A), and spectral analysis of prepared mixtures, as performed for Excedrin^® above, allowed identifying p-aminophenol with an error of less than 1%.

Figure 5 - A) Raman spectra of acetaminophen (red) and p-aminophenol (black). B) Raman spectra of pure epinephrine (USP grade, top spectrum) and injectable epinephrine product (0.1% in saline). C) SERS of epinephrine product (blue spectrum) compared to prepared 0.1% epinephrine (red), and prepared 0.1% nor- epinephrine (black). Conditions for A and B: 500 mW 1064 nm and 10-min accumulation at 8 cm^-1 resolution, Conditions for C: 80 mW of 785 nm, 10 sec.

While the analysis of drug degradation may be straightforward for a product such as Tylenol^®, composed of 75% acetaminophen, it is more difficult for low concentration products, such as injectables that contain less than a percent active. In fact, the active ingredient is undetectable in a Raman spectrum of a commercial 0.1% epinephrine product (Figure 5B). Surface-enhanced Raman spectroscopy (SERS) can be used to amplify the signal by as much as 6-orders of magnitude [41, 42]. The SERS effect has several requirements [43]. First, the materials used must have conducting electrons, such as copper, gold and silver. Second, the size of the material must be similar in magnitude to the wavelength of light used in measurements so that the electromagnetic wave, i.e. light, can cause the electrons to oscillate forming a plasmon field. Third, the drug of interest must be within the plasmon field so that its Raman scattering can be generated by both the plasmon and electromagnetic fields. Since the intensity of Raman scattering is a function of the field strength squared, then the signal intensity is amplified to the fourth power. Also, certain relationships (e.g. orientation) between the electron distribution for a given vibrational mode and the plasmon field must be met. The last requirement, often overlooked, determines the extent that a vibrational mode is enhanced. In fact, the Raman and SER spectra for epinephrine are quite different (Figure 5B and C). The SER spectrum is dominated by the vibrations that include nitrogen, such as the CCN bending and secondary amine modes at 485 and 1663 cm^-1, while the Raman spectrum is dominated by the aromatic ring modes at 779 and 955 cm^-1. This is not surprising, since epinephrine is expected to interact with the silver particles through the nitrogen lone pair. More importantly, the degradation product, nor-epinephrine, in which hydrogen replaces a methyl group, produces a SER spectrum that is sufficiently different for identification and quantitation. Specifically, the CCN bending and secondary amine modes shift to 446 and 1656 cm^-1.

Conclusion

Raman spectrometers will continue to decrease in size, integrate into other measurement apparatus, and provide incredible analyses through chemometrics. I hope this short review provides some insight into the capabilities of Raman spectroscopy for pharmaceutical applications.

References

1. Smekal A “Zur Quantentheorie der Dispersion”, Die Naturwissenschaften, 11, 873-875 (1923).
2. Raman CV, KS Krishnan, “A New Type of Secondary Radiation”, Nature, 121, 501-502 (1928).
3. Landsherg GS, LI Mandelstam, “New phenomenon in scattering of light (preliminary report)”, J Russ Phys-Chem Soc, 60, 335 (1928).
4. Porto SP, DL Wood “Ruby optical maser as a Raman source,” J Opt Soc Am, 52, 251-252 (1962).
5. Woodruff WH, S Farquharson “Time-Resolved Resonance Raman Spectroscopy of Hemoglobin Derivatives: Heme Structure Changes in 7 Nanoseconds,” Science, 201, 831- 833 (1978).
6. McCreery RL, M Fleischmann, P Hendra, “Fiber optic probe for remote Raman spectrometry”, Anal Chem, 56, 146-148 (1983).
7. McLachlan RD, GL Jewett, JC Evens, “Fiber-optic probe for sensitive Raman analysis” US Patent 4,573,761 (1986).
8. Farquharson S, SF Simpson, “Applications of Fiber Optic Raman Spectroscopy to Chemical Processes,” Proc SPIE, 1796, 272-285 (1992).
9. Hirschfeld T, B Chase, “FT-Raman Spectroscopy: Development and Justification”, Appl Spectrosc, 40, 133-137 (1986).
10. Yang B, M Morris, H Owen, “Holgraphic notch filters for low-wavenumber Stokes and anti-Stokes Raman spectroscopy, Appl Spectrosc, 45, 1533-1536 (1991).
11. Angel SM, M Carrabba, TF Cooney “The utilization of diode lasers for Raman spectroscopy”, Spectrochim Acta A, 51, 1779-1799 (1995).
12. Christesen S, B Maciver, L Procell, D Sorrick, M Carrabba, J Bello, “Nonintrusive analysis of chemical agent identifi cation sets using a portable fi ber-optic Raman spectrometer” Appl Spectrosc, 53, 850-855 (1999).
13. Gendrin C, Y Roggo, C Collet,” Pharmaceutical applications of vibrational chemical imaging and chemometrics: A review, J Pharma Biomed Anal, 48, 533-553 (2008).
14. Smith W, S Farquharson, “A portable fuel Analyzer”, Proc SPIE, 6377, 63770E (2006).
15. Carron K, MA Watson, SA Buttler “Raman Spectrometer”, US Patent Application 2005/0248758, (2005).
16. Santillan JD, CD Brown, W Jalenak, “Advances in Raman spectroscopy for explosive identifi cation in aviation security, Proc SPIE, 6540, 1-8 (2007).
17. Williams KPJ, RE Aries, DJ Cutler, DP Lidiard “Determination of gas oil cetane number and cetane index using near-infrared Fourier-transform Raman spectroscopy”Anal Chem, 62 (23), 2553- 2556 (1990).
18. Cutmore EA, PW Skett, “Application of Fourier-transform Raman spectroscopy to a range of compounds of pharmaceutical interest”, Spectrochim Acta, 49A, 809-818 (1993).
19. Petty CJ, DE Bugay, WP Findlay, C Rodreigez, “Applications of FT-Raman spectroscopy in the pharmaceutical industry” Spectroscopy, 11, 41-45 (1996).
20. Hendra PJ, “Fourier-transform Raman spectroscopy in pharmaceutical analysis and research”, Amer Lab, December, 17-24 (1996).
21. Frank C, “Review of pharmaceutical applications of Raman spectroscopy” Chapter 6 in Analytical Applications of Raman Spectroscopy, MJ Pellitier, Ed, Blackwell Science, Oxford, UK (1999).
22. Deeley CM, RA Spragg, TL Threlfall, “A comparison of Fourier-transform infrared and nearinfrared Fourier-transform Raman spectroscopy for quantitative measurements: an application to polymorphism”, Spectrochim Acta, 47A, 1217-1223 (1991).
23. Neville GA, HD Beckstead, HF Shurvell, “Utility of Fourier-transform Raman and Fourier-transform infrared diffuse relectance spectroscopy for differentiation of polymorphic spironolactone samples”, J Pharmaceut Sci, 81, 1141-1146 (1992).
24. McGill CA, A Nordon, DLittlejohn, “Comparison of In-Line NIR, Raman and UV-Visible Spectrometries, and At-Line NMR Spectrometry for the Monitoring of an Esterifi cation Reaction”, Analyst, 127, 287-292 (2002).
25. Vergote GJ, TRM De Beer, C Vervaet, JP Remon, W Baeyens, N Diericx, F Verpoort,”In-Line Monitoring of a Pharmaceutical Blending Process Using FT-Raman Spectroscopy”, Eur J Pharma Sci, 21, 479-485 (2004).
26. Starbuck C, A Spartalis, L Wai, J Wang, P Fernandez, CM Lindemann, GX Zhou, ZH Ge; Process Optimization of a Complex Pharmaceutical Polymorphic System Via In Situ Raman Spectroscopy; Cryst Growth Des, 2, 515-522 (2002).
27. Lewis IR, “Process Raman Spectroscopy”, In: Handbook of Raman Spectroscopy: From the Research Laboratory to the Process Line; Lewis IR, HGM Edwards, Eds, Marcel Dekker, New York, 2001
28. Kim J, J Noh, H Chung, YA Woo, MS Kemper, Y Lee, “Direct, non-destructive quantitative measurement of an active pharmaceutical ingredient in an intact capsule formulation using Raman spectroscopy”, Anal Chim Acta, 598, 280-285 (2007).
29. Breitenbach J, W Schrof, J Neumann, “Confocal Raman-spectroscopy: Analytical approach to solid dispersions and mapping of drugs”, Pharm Res, 16, 1109-1113 (1999).
30. 30. Schlucker S, MD Schaeberle, SW Huffman, IW Levin, “Raman Microspectroscopy: A Comparison of Point, Line, and Wide-Field Imaging Methodologies”, Anal Chem, 75, 4312-4318 (2003).
31. Sasic S, DA Clark, “Defi ning a strategy for chemical imaging of industrial pharmaceutical samples on Raman line-mapping and global illumination instruments”, Appl Spectrosc, 60, 494-502 (2006).
32. El Hagrasy AS, S-Y Chang, D Desai, S Kiang, “Raman spectroscopy for the determination of coating uniformity of tablets: assessment of product quality and coating pan mixing efficiency during scale-up”, J Pharm Innov, Sept/Oct, 37-42 (2006).
33. Matousek P, AW Parker, “Non-invasive probing of pharmaceutical capsules using transmission Raman spectroscopy”, J Raman Spectrosc, 38, 563 (2007).
34. Adar F, E Lee, A Whitley, “Transmission Raman Offers Improved Quantitation of Pharmaceutical Solids”, Spectroscopy (2009).
35. World Health Organization, “62nd World Health Assembly item 12.9: Counterfeit medical products”, April 2009.
36. Roggo Y1, Degardin K, Margot P. “Identifi cation of pharmaceutical tablets by Raman spectroscopy and chemometrics”, Talanta, 81,988-95 (2010).
37. Sacré PY, E Deconinck, L Saerens, T De Beer, P Courselle, R Vancauwenberghe, P Chiap, J Crommen, JO De Beer “Detection of counterfeit Viagra^® by Raman microspectroscopy imaging and multivariate analysis”, J Pharm Biomed Anal, 56, 454-61 (2011).
38. Song H, TS Chen, “P-aminophenol-induced liver toxicity: tentative evidence for a role of acetaminophen”, J Biochem Mol Toxicol, 15, 34-40 (2001).
39. Nourjah P1, Ahmad SR, Karwoski C, Willy M, “Estimates of acetaminophen (Paracetomal)- associated overdoses in the United States”, Pharmacoepidemiol Drug Saf. 2006 Jun;15(6):398-405.
40. Sornchaithawatwong C, S Vorrarat and P Nunthanavanit, “Simultaneous determination of paracetamol and its main degradation product in generic paracetamol tablets using reversephase HPLC”, J Health Res, 24, 103-106 (2010).
41. Jeanmaire DL and RP Van Duyne, “Surface Raman spectroelectrochemistry: Part I. Heterocyclic, aromatic, and aliphatic amines adsorbed on the anodized silver electrode”, J. Electroanalytical Chem., 84, 1-20 (1977).
42. Weaver MJ, S Farquharson, MA Tadayyoni, “Surface-enhancement factors for Raman scattering at silver electrodes. Role of adsorbate-surface interactions and electrode structure”, J Chem Phys, 82, 4867-4874 (1985).
43. Pettinger BJ “Light scattering by adsorbates at Ag particles”, Chem Phys, 85, 7442-74

Dr. Stuart Farquharson  published his first paper with William Woodruff in Science in 1978 that described one of the first dispersive Raman systems. His first SERS paper was published in JACS in 1983 with Michael Weaver and Nobel Laureate Henry Taube. Dr. Farquharson ran a Process Analytical Group for nearly twenty years and in 2001, he founded a company that manufactures Raman analyzers and has worked as the President and CEO ever since.