As Raman spectroscopy becomes a more sophisticated and widely used research tool, it is also penetrating industry, where manufacturers are harnessing it for process monitoring and quality control. This tutorial examines how Raman is well-suited to tackle two of the challenges faced by industry: mixtures and isomers. We look at recent examples in the literature addressing real industry applications and illustrate the solutions using some simple Raman experiments. First, we demonstrate nondestructive measurement of the concentration of a specific component in a mixture (ethanol), even when the mixture is complex and its other components are unknown. Second, we show how Raman spectroscopy is well-suited to distinguish the different isomers of a compound, showing both alcohol and xylene isomer spectra as examples.

Raman spectroscopy has come a long way since the Raman effect was first observed in 1928. For decades it was an esoteric technique, one that was complicated to use because of the need for a powerful, coherent source and sensitive detectors. The last two decades have seen the field come to maturity, thanks in large part to lasers with improved stability, power, size, and cost, as well as more sensitive detectors. Better filtering technology enables steeper and deeper blocking of the laser and fluorescence background and chemometrics software extracts more information from the generated spectra. Furthermore, the market now offers "build your own" modular systems in which the components are designed to be easily interfaced, giving users greater flexibility to customize or alter their system as required.   

When turnkey operation is necessary, application-specific systems designed for industry address chemical identification, pharmaceutical quality control, process monitoring, and first responder applications with novice-friendly features like trigger operation and integrated display screens. Though not quite a household name, Raman spectroscopy has become a mature spectroscopic technique in its application to industry. 

Nonetheless, Raman spectroscopy still retains an air of mystery for inexperienced users. Just how does one coax these spectra with hundreds of peaks to reveal their secrets? This article looks at some of the complex problems being tackled by Raman spectroscopy in research and industry and illustrates the method of analysis via spectra from chemicals possessing much simpler Raman signatures. 

It's easy enough to grasp how Raman spectroscopy can be used to identify a specific chemical — a spectrum is generated where peaks correspond to specific vibrational and rotational modes within the molecule. Each chemical has its own unique fingerprint, which can be cross-referenced with a library of known Raman spectra for identification. When dealing with a mixture, however, peaks quickly begin to overlap and spectra become more convoluted.   

In some cases, the Raman spectra of the mixture's constituents are quite simple, and it is only the conditions of measurement that are complex. Raman spectroscopy has been evaluated for real-time monitoring of the quality of liquid oxygen in the delivery line during ground testing of rocket engines, in which high temperature and pressure limit the performance of commercially available sensors. Tiwari and colleagues (1) developed an all-optical cryogenic fluid sensor to measure the concentration of liquid nitrogen in liquid oxygen, working with the 2331 cm

  peak of liquid nitrogen and the 1556 cm

  peak of liquid oxygen. Their work compared performance using both standard 532-nm lasers and a more cost-effective 670-nm diode laser. They found the sensor to be very sensitive at low concentrations, capable of measuring liquid nitrogen percentages as low as 1% in a sample mixture for a 532-nm laser and ~20% for a 670-nm diode laser (limited by the need for better filtering of background light). 

At other times, the constituent to be measured is a relatively simple molecule, but the mixture is complex. Anand and colleagues (2) demonstrated that it is possible to measure the proportion of methanol by volume in gasoline mixtures using Raman spectroscopy. Methanol is favored as an alternative transport fuel because it is relatively inexpensive, is nonpetroleum based, burns cleanly, and can be stored as a liquid at room temperature. It can be used in a modified version of the standard combustion engine, with an 85:15 methanol–gasoline mix used most often. Anand and colleagues used a 785-nm excitation laser, observing that the peak height at 729 cm

  decreased linearly as the proportion of methanol by volume increased.   

To illustrate how a single component in a mixture can be quantified, we prepared standard solutions of ethanol (Aldrich Chemicals) in deionized water ranging from 0–100% in glass vials. Raman spectra were taken using a modular system from Ocean Optics, comprising a QE65000-Raman spectrometer (range 780–930 nm, optical resolution ~0.4 nm FWHM, and 90% quantum efficiency), SpectraSuite operating software, an InPhotonics 785-nm probe (RIP-Probe-785), and a 785-nm stabilized laser. All measurements were taken using the same acquisition parameters: an 8-s integration time, 1 scan to average, and 0 box car smoothing. We used the peak height at 880 cm

  to generate a second-order polynomial calibration curve against concentrations for the standard solutions (Figure 1).  

 Figure 1: Calibration curve for ethanol in water (inset shows 880 cm-1 peaks used to generate the curve).			

Also, a Raman spectrum was taken for some discount-store hand sanitizer using the same acquisition parameters, with the peak height at 880 cm

  used to determine the ethanol content from the calibration curve. Hand sanitizers contain up to a dozen inactive ingredients, such as carbomer, triethanolamine, propylene glycol, tocopherol acetate, dyes, and scents. The resulting spectrum is more complicated than ethanol in water, but the 880-cm

  peak remains very clear (Figure 2). Though most brand-name sanitizers contain 60% or more ethanol, as recommended by the FDA, the discount sanitizer was found to have only ~45% ethanol. 

Figure 2: Raman spectra of pure ethanol and an unknown hand sanitizer.			

Peak height is often sufficient for quantitative analysis, but using area under the peak can improve accuracy.  When there is significant interference between characteristic peaks, peak ratios are sometimes used for analysis, or partial least squares analysis is applied through commercially available chemometric software packages such as GRAMS AI from Thermo Scientific and Analyze IQ Lab from Analyze IQ Limited.   

Sometimes extensive data processing is needed to achieve the desired accuracy, as in the case of McGoverin and colleagues (3), who had some success in predicting the solid fat content and triacylglycerols of anhydrous milk fat, the most commercially important separated fat component of milk. The study used a 752-nm excitation laser with a spectrometer for detection. They used a combination of software to perform a number of corrections, including smoothing, correction for baseline, standard normal variate (to remove scatter), Norris derivative smoothing, and area normalization. They used additional analysis methods to determine which calibration treatments yielded the best quality prediction models, working with multiple bands within the spectra.  

Figure 3: 1-Butanol and 2-butanol isomers.			

Raman spectroscopy also has the ability to look beyond chemical structure in quantification. Fourier transform (FT)-Raman has been used to measure the degree of crystallinity of cellulose, which is important because it affects physical, mechanical, and chemical properties. Tensile strength, dimensional stability, and density all increase with crystallinity, but chemical reactivity and swelling decrease. Though most often measured by wide-angle X-ray scattering, researchers Agarwal and colleagues (4) found FT-Raman to be a more accurate method for determining the degree of crystallinity in cellulose. FT-Raman is preferred over conventional Raman in this case because of its ability to analyze samples that would typically fluoresce.   

Some of the mixtures under scrutiny in industry contain two or more isomers of a given chemical. Even though isomers share the same molecular formula, their properties can be very different depending on the type of isomer. Structural isomers differ in arrangement of their functional groups, such as degree of branching of hydrocarbons or the positions of functional groups on a chain or ring. Their chemical properties tend to be quite different. Stereoisomers share the same bond formation, but differ in how one or more atoms or functional groups are positioned geometrically in space; diastereomers possess one or more portions that appear "flipped" or rotated relative to the other, while enantiomers are full mirror images of one another. Stereoisomers share the same physical properties and behave very similarly in chemical reactions, unless reacting with other stereoisomers or enzymes, giving them very different biological effects.  

Figure 4: Structures of p-, o-, and m-xylene isomers.			

Separating isomers is a challenge that appears regularly in industry and can require highly specialized techniques. Chromatography, absorption, and crystallization all can be used, but become more difficult when the physical properties of the isomers are similar. Any improvement in efficiency or selectivity can mean significant cost savings, driving the development of new separation techniques.   

We find one such example in the gasoline industry, where benzene content has been decreased, increasing the need for catalytic isomerization to convert straight-chain hydrocarbons to branched hydrocarbons. Branched hydrocarbons burn more efficiently and have a higher octane number, but are difficult to separate from the residual linear hydrocarbons. Molecular sieving is typically used, but requires large amounts of zeolite. The use of novel sorbents such as silicalite to exploit subtle entropy effects may prove more efficient (5), but requires real-time monitoring.   

Another example comes from the textiles industry, where dinitrotoluene is a key component in making polyurethanes used in materials like insulation, soundproofing, linings for clothing and sleeping bags, cushions, and spandex. A mixture of the isomers 2,4- and 2,6-dinitrotoluene is used because of the difficulty of separation, despite the expectation that polyurethanes synthesized from pure 2,6-dinitrotoluene would have a higher glass transition temperature and greater stability (6). If it could be efficiently isolated, the reaction speed with a single isomer would also be more consistent. 

Raman spectroscopy is well-suited to serve as a diagnostic tool for these types of separation challenges, both for development and in-process. It is rapid, noninvasive, moves easily between solids and liquids, and deals well with mixtures. Its ability to instantly probe the vibrational and rotational modes of any molecule with a single incident wavelength makes it extremely versatile. To better understand how Raman would be used in these applications, let's look at the spectra of some simple compounds occurring as isomers: butanol and xylene.   

Butanol occurs as 1-butanol, a naturally occurring minor product of fermentation and an intermediate in the production of butyl acetate (a flavorant and industrial solvent), and as 2-butanol, produced on a large scale as a precursor to the industrial solvent methyl ethyl ketone (found in many domestic cleaning agents and paint removers). 2-Butanol is chiral, found as an equal mix of the two stereoisomers.   

Xylene occurs as three isomers: meta-, ortho-, and para-xylene, depending on where the methyl groups are attached to the benzene ring. Technical grade xylene is typically 45–55% 

-xylene, with the remainder consisting of ethyl benzene. Xylene is used as a solvent in many industries, from leather and printing to rubber and plastic. It is also used for cleaning steel, as paint thinner, and can be found in gasoline and airplane fuel. Knowledge of the relative proportion of the three isomers is important in the manufacturing of polymer compounds.  

Because of its potentially adverse health effects, xylene falls into the volatile organic compound (VOC) category. Regulatory agencies monitor the levels of xylene in drinking water, in the workplace and in the environment. Because only 

-xylene are man-made), a technique that can discriminate among the three isomers is useful.   

Samples of each chemical isomer were prepared in glass vials (Aldrich Chemicals) and placed in sequence into a closed container into which the laser probe was directed. Measurements were made using the same modular Raman system from Ocean Optics described previously, but using an 8-s integration time.   

The spectra acquired for butanol are shown in Figure 5. Some peaks are shared between the two isomers, including the typical CH

  (7). Alcohols have distinct skeletal vibrations that are caused by hydrogen bonding. The addition or deletion of groups causes changes in the geometry and distribution of electrons, altering the Raman spectra. This is seen in the range 700–1200 cm

  because of differences in the C-C and C-O stretching modes of the two isomers. The broadened peak at 775–785 cm

  for 2-butane is consistent with a wider range of vibrational frequencies calculated for the different conformers relative to the more consistent vibration at 842–844 cm

Figure 5: Raman spectra of 1- and 2-butanol.			

Spectra acquired for xylene are shown in Figure 6. Although all three isomers of the xylene share peaks that are caused by ring modes in the 1400–1600 cm

  range, with a unique C-H bending band for 

  (7). By using partial least-squares regression analysis, Cooper and colleagues (9) demonstrated that the concentration of an individual xylene isomer can be measured to within ±0.15% using a low-cost dispersive spectrometer. More recently, coherent anti-Stokes Raman scattering (CARS) has been used by Li and colleagues (10) to measure concentrations as low as 1% in binary mixtures of 

Figure 6: Raman spectra of p-, m-, and o-xylene.			

When the isomerism is more subtle, such as in stereoisomers, multiple peaks and peak ratios may be needed for identification, as demonstrated by Zheng and colleagues (11) in their studies of reversibly photoswitchable azobenzene. Photoswitchable molecules can act as molecular machines. Azobenzene is a popular choice for study because of its simple molecular structure and spectra. It alternates between the cis and trans isomers upon exposure to UV and blue light, respectively. Zheng and colleagues had the idea to combine the single-molecule sensitivity of surface enhanced Raman scattering (SERS) with a controlled nanoscale environment of self-assembled monolayers of alkanethiol molecules on gold. They matched the excitation wavelength to the resonant wavelength of the Au substrates, using a 632.8-nm He-Ne laser with a Raman system for detection. They then took SERS spectra of the tethered azobenzene while alternating exposure of the sample to UV (365 nm) and blue (450 nm) light to trigger photoisomerization of the azobenzene between the cis and trans forms of the molecule (Figure 7). 

Figure 7: Isomerization of azobenzene between cis and trans forms. Image adapted from http://en.wikipedia.org/wiki/File:Azobenzene_isomerization.png.			

As a benchmark, they compared their data to theoretical calculations of the isomers' Raman spectra, generated via B3LYP/6-31G* calculations (12) using a NWChem program package. Reversible cis

trans photoisomerization was observed by monitoring peak area ratios as a function of time as the sample was exposed to UV and then blue light in turn, showing good agreement with simulation. Their work has applicability to all-optical molecular devices. 

Raman spectroscopy is even finding its way into medical applications and biotechnology. Cysteamine exhibits rotational isomerism and is used in the treatment of radiation sickness as well as other disorders of cystine excretion. Goto and colleagues (13) used SERS to probe the molecule while exerting a magnetic pulling force to control the rotational isomerization equilibrium between the trans and gauche modes. They used a 514.5-nm laser for excitation in combination with a Raman microscope system. Relative peak intensities for the C-S stretching modes for the two isomers clearly showed that the exerted magnetic force was able to shift the molecule from the gauche  to the trans conformation by exerting a stretching force.   

Finally, our review would not be complete without an example of quantifying isomers using Raman. P. Novak and colleagues (14) used FT-Raman spectroscopy for in-situ monitoring during the synthesis of entacapone, a drug used in the treatment of Parkinson's disease. Entacapone exists as two isomers, E (trans-) and Z (cis-), with the drug form consisting of 99.5% E-isomer, the more pharmaceutically active of the two. They used the technique to compare reaction speed in different solvents, analyzing their data using principal component analysis.   

Raman spectroscopy has moved well beyond basic identification of compounds in its applicability to industry and now tackles more complex problems like quantification in mixtures and isomers with ease. The increasing availability of both turnkey and high quality modular systems allows even inexperienced users to harness this powerful technique.  

, is senior applications scientist with Ocean Optics in Dunedin, Florida. 

 is an applications laboratory specialist for Ocean Optics. 

  is principal at Mapleseed, LLC in Dunedin, Florida. Direct correspondence to 

 (1) V.S. Tiwari, R.R. Kalluru, F.Y. Yueh, J.P. Singh, W. St. Cyr, and S.K. Khijwania, 

 (2) K. Anand, V. Asundi, and R. Vasudeva, 

 (3) C.M. McGoverin, A.S.S. Clark, S.E. Holroyd, and K.C. Gordon, 

 (4) U.P. Agarwal, R.S. Reiner, and S.A. Ralph, 

Process for Separating Isomers of Dinitrotoluene, 

 Infrared Spectral Interpretation: A Systematic Approach

 (CRC Press, Boca Raton, Florida, 1999). 

 (8) H. Hageman, J. Mareda, C. Chiancone, and H. Bill, 

Articles

This tutorial demonstrates that Raman is well-suited to tackle two challenges faced by industry: complex mixtures and isomers.

Raman Spectroscopy as a Quantitative Tool for Industry

There is a continuing need to develop new techniques for rapid and specific identification of bacterial pathogens in human body fluids especially given the increasing prevalence of drug-resistant strains. Efforts to develop a surface-enhanced Raman spectroscopy (SERS)-based approach, which encompasses sample preparation, SERS substrates, portable Raman microscopy instrumentation, and novel identification software, are described. The progress made in each of these areas in our laboratory is summarized and illustrated by a spiked infectious sample for urinary tract infection diagnostics. SERS bacterial spectra exhibit both enhanced sensitivity and specificity, allowing the development of an easy-to-use, portable, optical platform for pathogen detection and identification. SERS of bacterial cells is shown to offer reproducible molecular spectroscopic signatures for analytical applications in clinical diagnostics. The approach is a new tool for studying biochemical activity in real time at the outer layers of these organisms. 

The ability to rapidly detect and identify bacterial cells in human body fluids for a relatively low cost in point-of-care settings is a continuing need for health care providers worldwide. 

, the presence of bacteria in blood, results from severe infections at sites in the body, surgical wounds, or contaminated implanted devices and may lead to the potentially fatal condition of sepsis. Because of an ~40% mortality rate, sepsis ranks 13th among the causes of overall death in the United States (1–3). Urinary tract infections (UTI), evidenced by bacteria in urine, are among the most common types of infections in humans. Approximately 50% of all women will have at least one UTI in their lifetime, and a sizable percentage of those will suffer from chronic UTIs (4,5). In the United States alone, UTIs are responsible for more than 7 million doctor office visits and more than 1 million hospital admissions at a cost of approximately $1 billion per year. Sepsis and UTI are caused by a variety of bacterial species and strains. Furthermore, the increasing proliferation of drug-resistant strains places an increasing premium on identifying these microbes with strain specificity in a useful time frame for narrow spectrum antibiotic drug prescription (6). In addition, the need for rapid bacterial identification methods for food and water safety applications has been illustrated by the several high-profile incidents reported nationwide over the past few years (7).  

Traditional methods of bacterial identification are phenotypic-based approaches that require a cell growth period and are consequentially slow (24–48 h or longer). Also, distinguishing closely related strains may be difficult via traditional methods and it is not a point-of-care technique. The best current methods are molecular diagnostic approaches that utilize specific primers or probes for particular gene targets, such as "real time" polymerase chain reaction methods (PCR) (8), which are increasingly finding use in clinical settings. If no culturing is required, PCR time frames are typically in the 2–6 h framework. However, PCR is not without some limitations or liabilities, including sample contamination, infectious mixture resolution, the need for required primer sets, speed, cost, and point-of-care capabilities (9). Hence, at the very least, there is a continuing need for competing or orthogonal bacterial identification methods. 

For the past few years we have been developing an optical approach for rapid, sensitive, and specific bacterial diagnostics based on surface-enhanced Raman spectroscopy (SERS) (10–14). SERS is a well-known spontaneous light scattering technique, discovered ~35 years ago, that results in the 10

  effective enhancement of the Raman scattering intensity of some molecular vibrational modes of molecules that are close (≤~5 nm) to nanostructured metal surfaces (15,16). This effect is predominantly attributed to the plasmonically enhanced local electric fields that become concentrated near nanosized structures that are coincident with the surface plasmon resonances of these nanomaterials. These plasmon resonances are usually in the near-infrared (NIR) to visible range for the most commonly employed metals, silver (Ag) or gold (Au). Aside from the advantages of speed, ease-of-use, and potential cost introduced by the development of an optical approach for bacterial diagnostics, a SERS-based platform also offers the advantage of portability because only a few milliwatts of laser power are required for acquisition of high-quality SERS spectra. This portability is central for the application of this SERS platform to a wide variety of applications in addition to clinical bacterial diagnostics, such as forensics, food and water safety testing, and in situ studies of works of art and cultural heritage materials. 

The total SERS pathogen detection platform may be viewed as consisting of four key components, described below. 

 Although often overlooked, sample preparation is a key ingredient for the effective use of novel pathogen detection methods. We are developing a two-stage approach for bacterial enrichment from human blood (12,17). Detecting bacteria in blood is extremely challenging because red blood cell concentrations are ~10

 /mL in human blood, and other particles, such as white blood cells and platelets, are present at ~10

 /mL levels. In contrast, medically relevant concentrations corresponding to bacteremic presentations are in the 10

 /mL range, and the surface area of a red blood cell is about an order of magnitude larger than that of a typical bacterial cell. However, through selective lysis, centrifugation, and micro-evaporation sufficient enrichments of ~10

  can be achieved, allowing for the acquisition of SERS spectra in 100-nL volumes starting with 10 mL of infected blood. Sensitivities down to 10

 /mL have already been achieved in a bench top procedure and an automated device is currently being developed in our laboratory (12). The required sample preparation procedures for UTI detection is simpler because the concentration of other cells in urine is normally low and the clinical levels of bacterial infection concentrations are much higher in urine (≥10

 /mL) than blood, as described further below. 

 Most of our SERS data have been acquired from bacterial cells placed on a substrate produced by a gold-ion doped sol-gel procedure (10). These substrates yield strongly enhanced, reproducible SERS signals for vegetative bacterial cells with 785-nm excitation. As shown in Figure 1, these substrates have clusters of 80-nm Au nanoparticles physiadsorbed to the outer surface of a SiO

 matrix. Further improvements in SERS chip performance will only enhance the value of this technique. We have collaborated on the design and production of several engineered substrates (18–20), however, the sol-gel Au nanoparticle "chips" have proven to provide the best signal and noise for bacterial identification purposes of those we have tested so far. 

Figure 1: A scanning electron micrograph image of the Au nanoparticleâcovered SERS substrate produced by an Au-ion doped sol-gel process. Clusters of 1â15 ~80-nm Au particles are evident on the surface of the SiO2 SERS chip. Adapted from reference 10 with permission.			

A portable SERS signal acquisition instrument:

 Although several handheld Raman devices with nearly point-and-click capabilities are commercially available, true microscopic capabilities are necessary for this SERS application. Thus, we designed and built several prototype high-performance (cooled CCD array, ~5-cm

  spectral resolution, piezo-driven sample positioning stages, and submicrometer imaging resolution) portable Raman microscopes for this purpose. This unique instrument allows positioning of the 785-nm Raman excitation on a specific cell or sample region on the SERS substrate for Raman analysis. Imaging also provides additional morphology information that can help with identification procedures. Our most recent version of the Raman microscope was built in collaboration with BioTools, Inc.   

 As described further below, we have developed a novel multivariate data analysis technique that provides accurate distinction between the SERS spectra of closely related bacterial species and strains when combined with a previously developed reference library. 

Following multiple cycles of centrifugation and washing with distilled water, bacterial cells are resuspended in water and placed on the SERS substrate for signal acquisition. The maximum scattering intensity of the SERS spectrum on a per-bacterium (not per-molecule) basis is between 10

  times larger than the observed maximum signal for the normal Raman spectrum of the same bacterium (10). This per-bacterium signal enhancement factor enables signal acquisition close to the single-cell level, although usually SERS spectra are acquired from about 10–20 bacterial cells in the sample region. This level of signal enhancement allows the development of a portable Raman instrument because excellent signal-to-noise scattering can be achieved with low incident laser power at 785 nm. Aside from a measure of sensitivity, the capability of obtaining spectra at the single-cell level offers the possibility of resolving infectious mixtures by selectively analyzing different members of the bacterial population. 

Interestingly, there is no correspondence between the vibrational bands observed in the normal Raman spectrum and those in the SERS spectrum (10,13). Because of the distance dependence of the SERS enhancement mechanism, only those molecules at the outer layers or very close to the cell wall can contribute to the bacterial SERS spectrum. In contrast, all the molecules in the Raman illuminated region contribute to the normal Raman spectrum and that volume is dominated by the cell cytoplasm. Hence, the SERS and normal Raman spectrum of a given bacterial cell type are from totally different parts of the cell and have virtually no features in common.   

SERS spectra of eight different laboratory-cultured bacterial species from five genera are shown in Figure 2. These spectra are acquired with 10 s of ~2 mW of 785-nm incident laser power and result from the Raman excitation of 10–20 cells, typically. Spectra with excellent signal-to-noise ratios are obtained on our SERS substrates. As is evident in this figure, each species exhibits a unique SERS signature. Although the spectra have been arranged top to bottom by phylogenic proximity, only moderate correlation with this lineage is observed. Furthermore, no clear distinction between Gram-positive and Gram-negative bacteria is apparent. An under-appreciated aspect of bacterial SERS spectra is that, in addition to the enhanced sensitivity that SERS provides, the spectral differences between SERS spectra of different species or strains is greater in SERS than for normal Raman (13). In other words, SERS results in enhanced specificity as well as enhanced sensitivity for bacterial samples. 

Figure 2: SERS spectra of eight bacterial species obtained on Au aggregateâcoated SiO2 chips (Figure 1). An incident laser power of ~2 mW and a data accumulation time of  10 s were used to obtain these spectra. Spectra are offset vertically for display purposes and top-to-bottom ordered according to their phylogenetic relationship.			

To exploit these unique SERS signatures for bacterial diagnostics, we have developed a principal components analysis (PCA)-based methodology for matching the acquired SERS spectrum to a known species or strain in a previously compiled library (13). After Fourier filtering the SERS spectrum, a bar-code is generated by using the sign of the second derivative of the spectrum — that is, whether the spectrum has curvature up or curvature down at each wavenumber — to assign a "one" or a "zero" as a function of scattered frequency. An example of a SERS-based bar-code is shown in Figure 3 for a SERS spectrum of 

. These bar-codes, or arrays of ones and zeroes, color-coded as black (curvature down) or white (curvature up) per wavenumber interval in this figure, are the inputs for PCA and related standard analyses, such as discriminant function analysis (DFA) or hieracrchical cluster analysis (HCA). The bar-code methodology results in significantly enhanced diagnostic specificity and is the tool used for quantitative measures of reproducibility.  

Figure 3: SERS spectrum of B. cereus. The corresponding second-derivative "bar-code" spectrum is given below. The bar-code is determined only by the sign of the second derivative (curvature up = white, curvature down = black). These bar-codes are the inputs for the principal component analysis (PCA) used to identify unknown bacteria with SERS, demonstrated in Figures 4 and 5. 			

As an example of the effectiveness of this bar-code methodology and how it is used for bacterial diagnostics, two PCA plots (PC2 vs. PC3) are shown in Figure 4. The left-hand panel results when six SERS spectra of four closely related 

Ames) are the input vectors to a standard PCA procedure. Each spectrum corresponds to a "dot" in this plot, which has been color-coded to indicate the organism it represents; the rings correspond to 2D standard deviations. The tightness of the clusters for a given species or strain characterizes the reproducibility of the spectra and the distance between the color-coded clusters represents the specificity offered by these measurements. For example, 

 cannot be separated in this treatment (Figure 4, left). However, when the input to the PCA treatment is the bar-code spectra, all species or strains are well-separated, and much tighter clusters are obtained, as seen in the PCA plot on the right side in Figure 4. The contrast between these two panels illustrates the enhanced specificity that results from this bar-coding procedure and is key to the use of this methodology for bacterial diagnostics (13). 

Figure 4: Comparison of PCA plots (PC3 vs. PC2) for SERS spectra of four closely related Bacillus organisms (B. thuringiensis, B. cereus, B. anthracis Sterne, B. anthracis Ames). When the observed SERS spectra are the inputs to the PCA, the contour plot shown in the left-hand panel results. By contrast, when the input to the PCA are the bar-code representations of the observed SERS spectra, the contour plot shown in the right-hand panel results. The enhancement in specificity resulting from the bar-code spectra is clearly evident by this comparison. Adapted from reference 13 with permission.			

Following acquisition of the SERS spectrum of the unknown sample, the spectrum is converted to its corresponding bar-code and PCA analysis is performed with groups of bar-code SERS spectra of known bacterial identity selected from a previously developed library to identify the unknown bacteria. The species or strain group that the unknown clusters with determines the identity of the unknown. The identification method is only limited by the library size, which can be continuously and readily expanded, and is not dependent upon specific primer sets or antibody specificity. This approach is quite general and, in principle, could be expanded to other pathogens such as fungi or viral particles. 

The molecular origins of the vibrational bands usually observed in the SERS spectra of vegetative bacterial cells are attributed to the diversity of lipoproteins, proteins, lipids, and other cell-wall components at the outer membrane layer of these bacterial cells (21–23). The distance dependence of the SERS enhancement mechanism supports such a supposition. Despite this, we have discovered that nearly all of the observed vibrational signatures are due to small metabolites that are concentrated at the outer region of bacterial cells (24). In particular, the largest features in the SERS spectra are generally products in the purine degradation pathway. The large range and unique reproducible signatures found for the SERS bacterial spectra result from the large number of enzymatic variations in these metabolic pathways and the resulting different relative concentration of the many small molecules found in this near extracellular region. Thus, SERS offers not only reproducible molecular signatures for analytical detection and identification applications as shown here, but also a new tool for studying biochemical activity in real time at the outer layers of these organisms. 

Although the bulk of our effort has been directed toward the very challenging problem of bacteremia diagnostics, efforts in our laboratory have recently expanded to develop this SERS-based approach for UTI diagnostics. With relatively routine centrifugation-based techniques, bacterial concentration enrichments of 10

  can be achieved in urine samples that have been spiked at the clinically relevant concentration of 10

  cfu/mL. Figure 5 shows the PCA plot for three different organisms known to be causative agents of UTI. These bacteria were spiked into human urine at 10

  cfu/mL concentration, enriched through a centrifugation procedure, and then placed on the SERS substrate for data acquisition. Distinct clusters are readily obtained for the bacterial cells after recovery from urine indicating that species-specific SERS spectral signatures can be identified by this SERS bar-code procedure. This prototype, benchtop UTI test took about 30 min to perform, and the SERS based analysis illustrates the potential value of this promising methodology. 

Figure 5: A PCA plot resulting from a SERS bar-code cluster analysis for three bacteria (E. coli, K. pneumoniae, and S. saprophyticus) that were recovered from spiked urine samples at medically relevant concentrations (105/mL).			

The elements and basis for a SERS-based platform for bacterial diagnostics, which includes a sample preparation device, a portable Raman microscope, SERS substrates, and software for bacterial identification, have been described. This novel optical approach is a potentially transformative diagnostic procedure for the rapid, reliable identification of bacterial species or strains in patients with infectious disease symptoms. In addition, SERS provides an effective tool for learning about the chemical activity of molecules at the outer layers of these organisms and, in particular, is a new method for use in the recently established field of metabolomics. 

 are with the Department of Chemistry and the Photonics Center at Boston University in Boston, Massachusetts. 

 is is with the Fraunhofer USA Center for Manufacturing Innovation in Brookline, Massachusetts and the Department of Biomedical Engineering at Boston University. 

 is with Brigham and Women's Hospital at Harvard Medical School in Boston, Massachusetts. 

 is with the Department of Biomedical Engineering, the Department of Mechanical Engineering and the Photonics Center at Boston University. Direct correspondence to 

 (2) D.C. Angus, W.T. Linde-Zwirble, J. Lidicker, G. Clermont, J. Carcillo, and M.R. Pinsky, 

 (3) J.M. O'Brien, Jr., N.A. Ali, S.K. Aberegg, and E. Abraham, 

 (4) S.M. Schappert and E.A. Rechtsteiner, 

 (National Center for Health Statistics, Hyattsville, Maryland, 2008). 

, M.S. Litwin and C.S. Saigal, Eds. (Department of Health and Human Services, Public Health Service, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Washington, D.C, 2007), pp. 587–619. 

 (6) X. Yu, M. Susa, J. Weile, C. Knabbe, R.D. Schmid, and T.T. Bachmann,

http://en.wikipedia.org/wiki/List_of_foodborne_illness_outbreaks_in_the_United_States

 (9) C.E. Corless, M. Guiver, R. Borrow, V. Edwards-Jones, E.B. Kaczmarski, and A.J. Fox, 

(10) W.R. Premasiri, D.T. Moir, M.S. Klempner, N. Krieger, G. Jones II, and L.D. Ziegler, 

(11) W.R. Premasiri, D.T. Moir, and L.D. Ziegler, 

(12) W.R. Premasiri, D.T. Moir, M.S. Klempner, and L.D. Ziegler, in 

New Approaches in Biomedical Spectroscopy

, K. Kneipp, R. Aroca, H. Kneipp, and E. Wentrup-Byrne, Eds. (Oxford University Press, New York, New York, 2007), p. 164. 

(13) I.S. Patel, W.R. Premasiri, D.T. Moir, and L.D. Ziegler,

(14) W.R. Premasiri, Y. Gebregziabher, and L.D. Ziegler,

SERS of bacterial cells can be a useful technique for clinical diagnostics, as illustrated by the analysis of a human urine sample spiked with urinary tract infection bacteria.

Rapid Bacterial Diagnostics Via Surface-Enhanced Raman Microscopy

Shifted-excitation Raman difference spectroscopy (SERDS) has been demonstrated to be effective in situations where high fluorescence prevents effective use of Raman spectroscopy. Here, we describe a novel way to apply SERDS using pairs of high-power laser diodes with a fixed-wavelength separation. This approach was used to carry out quantitative Raman analysis of liquid mixtures in the presence of high levels of fluorescence as well as qualitative analysis of several naturally fluorescing compounds. Also, the use of a simple spectrum reconstruction algorithm is compared with the conventional method of removing the fluorescence background via numerical background fitting.

There is growing interest in the use of Raman spectroscopy in a vast scope of applications in various industries, including pharmaceuticals (1–5), petrochemicals (6–8) and law enforcement (9–13). This growth is driven in part by increased offerings of portable and affordable Raman instruments, which are powered by compact, high-performance wavelength-stabilized laser diodes. However, as the scope of applications for Raman spectroscopy broadens, the challenge of sample fluorescence arises more frequently. Besides the natural fluorescence that many substances possess, samples can also become contaminated with fluorescent compounds in the field (9,10). An example is that of fluorescent agents such as caffeine and flour regularly used to cut illicit street drugs (10). 

To combat sample fluorescence, laser excitation at longer wavelengths has long been used (14–16). Indeed, if it were not for sample fluorescence, nearly all Raman spectroscopy could be performed with the same short-wavelength laser, especially because blue and violet diode lasers are becoming more and more powerful and affordable. For substances with weak short-infrared (IR) fluorescence, 785-nm and 830-nm laser excitation have been an acceptable compromise. For samples exhibiting stronger short-IR fluorescence, one must resort to 1064-nm excitation. 

However, using 1064-nm excitation presents a number of issues for portable field instrumentation. First, it is well known that the Raman signal diminishes as the fourth power of laser wavelength, which means that it is 16 times weaker with 1064-nm excitation as it is with commonly used 532-nm excitation. At the same time, it is not always possible to compensate for a decreased Raman scattering cross-section by increasing laser power, because a stronger laser may damage the sample. As a result, collection times that are nearly an order of magnitude longer would be necessary to collect as much Raman signal when using a 1064-nm laser. Furthermore, when a laser of this wavelength is used for conventional Raman spectroscopy, the Raman signal appears at wavelengths well outside the sensitivity range of silicon detectors as a result of Stokes shift, which makes it necessary to use an InGaAs

detector instead of a silicon charge-coupled device (CCD). This presents an additional problem, because the dark noise of an InGaAs array is more than an order of magnitude worse than that of the Si CCDs. As a result, InGaAs detector arrays for dispersive Raman instruments need to be cooled to achieve satisfactory noise levels, and even then they don't reach the performance of CCDs operating at room temperature without cooling. This situation presents an obvious problem for portable, battery-operated Raman instrumentation, because it means more power consumption, a larger battery, and shorter battery life, not to mention that it increases the cost and complexity of the instrument. Another common approach to performing Raman analysis with 1064-nm excitation is Fourier transform Raman (FT-Raman). This approach simplifies detection by using a single, discrete cooled InGaAs or Ge detector, but at the same time significantly increases the complexity of the optical system, because it requires a stable Fabri-Perot interferometer to perform the analysis. FT-Raman systems tend to be larger and more expensive and consume more power than a dispersive Raman instrument that uses an uncooled CCD detector would. 

For these reasons, the use of shifted-excitation Raman difference spectroscopy (SERDS) has long been considered attractive, effectively expanding the use of conventional dispersive Raman instrumentation equipped with efficient CCD detectors to the classes of samples that exhibit fluorescence. Historically, the SERDS method has been used in laboratories employing tunable-wavelength lasers. However, this type of laser source presents a significant challenge for portable Raman systems. Generally, these lasers are much more costly than simple and efficient wavelength-stabilized laser diodes. In addition, there are challenges in ensuring the exact wavelength control over these lasers required to perform accurate qualitative and especially quantitative Raman analysis. To ensure accurate analysis, the wavelength must be monitored constantly. 

We have studied a comparatively simple and practical approach to performing SERDS analysis using two affordable wavelength-stabilized laser diodes operating at slightly offset wavelengths. The lasers are stabilized by use of volume Bragg gratings (VBGs) and are compact and efficient, making them well suited for portable battery-operated Raman instrumentation. 

The experiments were performed using an LS-2 dual-laser SERDS laser source (PD-LD, Inc.). The lasers operated at the 784.5-nm and 785.5-nm wavelengths. The wavelength separation was selected to correspond to the approximate line width of the Raman lines of the substances under study. Note that the exact wavelength separation is not significant for the accuracy and practicality of SERDS; however, wavelength separation that is much smaller than the width of the Raman bands would result in increased noise in the SERDS spectra.  

The LS-2 laser source delivers the output of either one of the lasers to the output port through a fiber-optic switch. A fiber-optic Raman probe was attached to the output port of the laser source and thus delivered light to the samples under study. The same probe collected the Raman signal in the back-reflection geometry, together with fluorescence and the Rayleigh scattering. The filters installed in the fiber-optic probe suppressed the Rayleigh scattering so that the collected Raman scattering signal and the fluorescence were delivered to the input port of the spectrometer via the collection fiber in the probe.  

For these experiments, we used a compact spectrometer with an uncooled CCD array (model USB2000+, Ocean Optics). The spectrometer had a 50-µm slit and approximately 10-cm

  resolution. For SERDS analysis, the spectra were collected sequentially, first with laser 1 and then with laser 2. The lasers ran continuously during the course of the experiment to ensure best stability of the output power. The wavelengths of the lasers were stable to <5 pm over the course of the day. The samples under study were in either liquid or solid form. A majority of the samples (usually liquids) were placed in small glass vials and illuminated through the bottom of the vial using the fiber-optic probe. 

Figure 1: Normalized raw Raman spectra of R6G dye in water solution collected at two different laser excitation wavelengths.			

After collection, the spectra were stored for further processing. The spectra were corrected for the dark background and the relative pixel response (the white light correction) and normalized. After that, the difference spectra were calculated for SERDS analysis. For conventional Raman analysis, the spectra were fitted with a polynomial (a fourth- or sixth-order polynomial was used) and the polynomial background was then subtracted from the original spectrum. In addition, a numerical derivative of the original spectrum was sometimes calculated for performing some of the quantitative Raman analysis. 

Figure 2: SERDS difference spectrum of the R6G dye in water solution. The spectrum was obtained with the data shown in Figure 1.			

Both SERDS spectra and the Raman spectra processed by conventional methods were compared with the library spectra for identification. Note that in no case was any a priori knowledge used for either identification or quantitative Raman analysis. Once identified, the spectra were also processed for quantitative determination of the relative concentration of the two major identified constituents, as is discussed below in more detail. 

To illustrate what collecting Raman spectra at the two SERDS wavelengths achieved, we show the spectra of R6G dye dissolved in water (Figure 1). As one can easily see, Raman peaks shift together with the laser wavelength, whereas the fluorescence background does not. Figure 2 shows the difference spectrum produced from the spectra shown in Figure 1. It is easy to see that the spectrum is very clean and free from any residual features of the fluorescence. Note also that SERDS spectrum automatically removes any etalon effects on the CCD. 

Figure 3: Raw, unnormalized spectra of a binary mixture of methanol and ethanol with strong fluorescence background from the R6G dye. The spectra shown were collected at the two SERDS fixed excitation wavelengths.			

Given that SERDS spectra are free from any background, have zero integral over the spectral window, and have a nearly perfect derivative shape, they are extremely convenient for performing Raman analysis, as they are nearly orthogonal for the majority of the substances of interest for Raman analysis. To illustrate this point, we performed quantitative Raman analysis on binary mixtures of methanol, ethanol, isopropanol, and acetone in the presence of a strong fluorescence background. In this experiment, the obtained SERDS spectrum was compared with library SERDS spectra, the components of the mixtures were identified, and then the relative concentration of the two most likely components in the mixture was determined. No multivariate calibration methods (such as partial least squares regression) were used. Thus, the system was not trained on the known mixtures; all samples were treated as completely unknown. 

Figure 4: Prediction of the relative concentrations of the two alcohols, methanol and ethanol, obtained in the presence of a strong fluorescence background from the R6G dye. The amount of fluorescence in the samples was the same as is shown in Figure 3. The purple squares show prediction based on SERDS analysis. The blue diamonds show prediction based on a conventional numerical processing method (calculating the derivative numerically).			

To compare the results obtained by SERDS with what conventional Raman analysis would predict, the fluorescence background was removed via numerical derivatization and then an analysis similar to the SERDS analysis was carried out. The Raman spectra were collected for a series of solutions with different concentrations of methanol in the presence of ethanol. All solutions contained a large amount of R6G dye to simulate strong interference from fluorescence. The raw Raman spectra of one of these solutions is shown in Figure 3. Figure 4 shows the predicted concentrations of methanol in the mixtures obtained via SERDS and the conventional method of numerical derivatization. As one can see, although both methods give reasonably accurate predictions of the methanol concentration when it is present in large proportions, the picture is quite different when a lower proportion of the methanol is present in the mixture. The conventional method starts to lose accuracy at a concentration of about 20–25% and then fails completely at concentrations lower than 10%. SERDS, on the other hand, predicts accurate concentrations even at 5% methanol content in the mixture. Note that in these experiments the ratio of Raman signal to fluorescence background was about 1:200, so that Raman peaks were barely observable over the fluorescence. 

Figure 5: Raw Raman spectrum of dark Jamaican rum. Strong fluorescence masks any sign of the ethanol Raman peaks.			

The above example of quantitative Raman analysis clearly illustrates the power of the SERDS methodology as compared with conventional methods. In addition, SERDS also can be useful for simpler tasks of identification of a compound that is naturally fluorescent or is masked by fluorescence from an unwanted additive. In such situations, identification should also be carried out using the differential SERDS spectra, without reconstruction. Quite often, however, there is a need for a human operator to take a look at the obtained spectra for additional control. In those cases, the derivative shape of the SERDS spectrum is undesirable, because it is not easily recognizable by a human. To assist in this task, reconstruction of the proper Raman spectrum from the differential SERDS spectrum must be carried out. This can be easily accomplished using a simple interpolation + integration algorithm described in an article by Matousek and colleagues (17). All the reconstructed spectra that appear in this article have been obtained using that algorithm. 

Figure 6: SERDS difference spectrum of the rum obtained from the raw Raman spectrum shown in Figure 5. The derivative shape of the Raman spectrum of ethanol is very clearly defined.			

To illustrate the power of SERDS for this type of application, we analyzed strongly fluorescent Jamaican dark rum, which is often used as a benchmark in Raman analysis of fluorescent samples. Figure 5 shows a raw Raman spectrum of the dark rum. As can be easily noticed, essentially no Raman features are observable in this spectrum, justifying its use as a benchmark substance. Figure 6 shows the SERDS spectrum of the rum, where derivative-shape Raman features of ethanol are easily identifiable. The reconstructed spectrum of the dark rum is shown in Figure 7 in comparison with the conventional Raman spectrum obtained via polynomial baseline subtraction. There are no Raman peaks observable in the conventional spectrum, whereas a trained eye can easily recognize the Raman features of ethanol in the spectrum reconstructed from SERDS. 

Figure 7: Raman spectra of the dark rum: reconstruction based on SERDS data shown in Figure 6 (red line), and the conventional Raman spectrum obtained with polynomial baseline fitting (blue line).			

Besides rum, which is used as a benchmark material but presents little interest for most other purposes, there are a number of substances that must be quantitatively and qualitatively analyzed for legitimate purposes. As an example, Figure 8 shows normalized Raman spectra of a regular olive oil and an extra virgin olive oil. The extra virgin olive oil is clearly more fluorescent and could be distinguished by just that, but analysis of the Raman features provides much more useful information. The Raman spectra of both oils reconstructed from SERDS are shown in Figure 9. The differences in Raman features between the two oils are subtle, and proper multivariate analysis is required for this classification task. 

Figure 8: Normalized conventional Raman spectra of two olive oils: regular oil (blue line) and extra virgin olive oil (red line).			

The results presented above demonstrate the advantages of SERDS over conventional Raman methods. These advantages stem from the accurate elimination of the fluorescence contribution, as opposed to approximating it by some smoothly varying polynomial. Another advantage of the method is its simple and accurate elimination of the stationary noise in the signal in the process of spectra subtraction. The only source of noise that is not eliminated is the shot noise associated with fluorescence. Nevertheless, it is easy to reduce shot noise by averaging over a larger number of collected spectra. Note that because exposure times need to be shortened to avoid saturation of the CCD when the fluorescence is strong, increasing the number of averages may not even result in much longer times for completing the analysis. It is instructive to observe that an increase in the number of averaged spectra does not lead to better results in the case of the conventional method of baseline fitting for fluorescence removal. 

Figure 9: Raman spectra of the two olive oils reconstructed from SERDS difference spectra. Blue line shows regular olive oil and the red line the extra virgin olive oil.			

We have shown that SERDS analysis can be easily and successfully carried out using two VBG-stabilized laser diodes operating at two slightly offset fixed wavelengths. The method has been shown to provide accurate measurement of the concentration of a minor constituent in a binary liquid mixture of alcohols at 5% of the minor component in the presence of a strongly fluorescence background that exceeded the Raman signal by a ratio of approximately 200:1. Any conventional Raman analysis carried out under the same conditions with the same excitation wavelength would fail at a concentration approximately three times larger of the same component in the mixture. The analysis was performed without any a priori knowledge about the substances in the sample; no prior training was performed on the system using any of the multivariate analysis methods. 

To illustrate the power of the approach compared with the conventional method of the polynomial background subtraction, we have analyzed dark rum and several other naturally fluorescent compounds. From this analysis it is evident that the conventional method cannot recognize or extract the spectrum of ethanol from the rum regardless of the number of averages taken on the original spectra. By contrast, SERDS gives a clear identification of ethanol and allows a reasonably clean reconstruction of the ethanol spectrum, which improves proportionately to the number of averages. 

Therefore, the SERDS method performed in the fashion described here is a practical and viable solution for Raman analysis in situations where a strongly fluorescent background is present. The solution is especially suitable for portable Raman systems where size, cost, and power consumption are of primary importance. It allows quick, effective, and affordable Raman analysis in the majority of real-life situations, thus obviating the need for more expensive, bulky, and power-hungry 1064-nm Raman systems. 

B. Appiah, S. Dolgy, V.S. Ban, E.D. Melnik

 are with PD-LD Inc. in Pennington, New Jersey. B. Appiah is also with Princeton University, in Princeton, New Jersey. Direct correspondence to 

  (1) C. Gendrin, Y. Roggo, and C. Collet, 

 (5) D. Pratiwi, J.P. Fawcett, and K.C. Gordon, T. Rades, 

 (6) W.M. Chung, Q. Wang, U. Sezerman, and R.H. Clarke, 

 (7) J.B. Cooper, P.E. Flecher, T.M. Vess, and W.T. Welch, 

 (9) R.J. Stokes, K. Faulds, and W.E. Smith, 

 (10) R.E. Littleford, P. Matousek, M. Towrie, A.W. Parker, G. Dent, R.J. Lacey, and W.E. Smith, 

 (11) E. Horvath, J. Mink, and J. Kristof, 

 (12) K. Faulds, W.E. Smith, D. Graham, and R.J. Lacey, 

SERDS analysis can be easily and successfully carried out using two high-power laser diodes with a fixed-wavelength separation in the presence of high fluorescence.

A Novel Method for Practical Implementation of the Shifted Excitation Raman Difference Spectroscopy (SERDS)

Micro-Raman spectroscopy is often performed to spatially distinguish materials of different chemical composition in solid samples. However, there are many instances in which the spatial heterogeneity consists not of chemical composition, but of solid-state atomic structure within a material of otherwise uniform composition. Diamond films prepared by chemical vapor deposition (CVD) on silicon wafers present just such a case in which strain, crystallinity, and defects vary spatially. An object can appear to be spectrally uniform if one is not aware of how the degree of crystallinity, strain, or state (amorphous or crystalline) of the matter manifests itself spectroscopically. To access the full benefits of micro-Raman spectroscopy, we show how the spatial heterogeneity of solid-state structure and the orientation dependence of strain in a nominal single crystal of CVD diamond can be characterized through the application of Raman polarization selection rules. 

The fabrication of diamond films by chemical vapor deposition (CVD) has been under development for many years now for applications including hardened surfaces, optical windows, and electrodes. The solid-state structure of these films has a direct impact on the material properties and, therefore, figures of merit of the final product. Crystallinity, single or poly, the number of defects, strain, and even the crystal face can affect performance. Furthermore, the uniformity of such films can affect the overall utility of the final product. Therefore, analytical methods are needed that can characterize the spatially varying structure of CVD diamond films. 

To that end, we used the application of Raman polarization selection rules for the purpose of simply differentiating single from polycrystalline diamond and to probe the extent of lattice defects, and thereby the degree of disorder, present in a crystal. 

 Figure 1: Calculated results for a P/O micro-Raman backscattering experiment for a diamond (100) face.			

The first figure displays the theoretical results for polarization/orientation (P/O) Raman backscattering from a diamond crystal with the laser beam incident on the (100) face (1). The Raman signal is calculated for configurations with the polarization analyzer parallel and perpendicular as the crystal is rotated in the plane. The experimentally obtained P/O micro-Raman diagram generated by plotting the peak area of the 1333-cm

  Raman band from a Drukker diamond (serving as our reference) is shown in Figure 2. 

Figure 2: Experimental results for the P/O micro-Raman backscattering experiment with a Drukker diamond (100) face.			

The diamond (100) experimental P/O plot consists of two sinusoidal functions 45° out of phase and the plot form is in excellent agreement with the calculated diagram of Figure 1. However, note that calculations predict a maximum response at 0° for the perpendicular configuration, whereas the parallel configuration is near maximum at 0° for the experimentally obtained data. This can be understood in terms of the crystal orientation of the calculation compared to the sample angle with respect to a laboratory reference frame. The approximate 50° phase shift between the experimentally obtained and calculated diamond (100) P/O plots is understood as a difference in the crystallographic starting orientations for data collection. This illustrates an application of P/O micro-Raman spectroscopy for the analysis of texture in natural materials or fabricated devices. P/O plot phase shifts can be used to quantitate the degree of crystallographic orientation or alignment of different features on a micrometer scale relative to a sample orientation and coordinate system in the laboratory reference frame. 

Figure 3: Reflected light image of CVD diamond from which P/O micro-Raman spectra were obtained.			

The CVD films consisted of diamond clusters whose edge lengths ranged from ~5 µm to 15 µm. The faces of the diamonds are square and triangular and presumably correspond to the (100) and (111) faces, respectively. We performed P/O micro-Raman spectroscopy on the isolated square faces to demonstrate the utility of this technique for the characterization of CVD diamond crystallinity. A reflected light image of one such square face from which P/O micro-Raman spectra were obtained is shown in Figure 3. 

Figure 4: P/O micro-Raman spectra of a (100) face of CVD diamond the light collection analyzer parallel and perpendicular to the incident polarization.			

We obtained P/O micro-Raman spectra of diamond films prepared by chemical vapor deposition using a 1.6-mW, 488.0-nm laser beam focused to a 0.6-µm diameter with an Olympus MS Plan 100 (0.95 NA) objective. Raman scattered light was collected using the same microscope objective (backscattering configuration). An analyzer was placed in front of the spectrometer entrance slit and configured either parallel or perpendicular to the incident polarization. Individual spectra were collected at 5° rotational increments for each analyzer polarization configuration as the sample was rotated within the focal plane. The sensitivities of the microscope and spectrometer to the two, orthogonal polarizations were characterized against a depolarized white light source. Accordingly, corrections were made to the intensities of P/O micro-Raman spectra obtained at the corresponding polarizations. 

Figure 5: Reflected light image of CVD diamond showing the assignment of crystallographic faces based upon the phase of the P/O micro-Raman spectral diagram.			

The P/O micro-Raman spectra (Figure 4) indicate that the diamond is a single crystal that contains substantial defects. The P/O micro-Raman spectra manifest the expected sinusoidal response for a single-crystal diamond rotated in a (100) plane. However, the failure of the signal to vanish at the minima and the presence of an amorphous carbon band at ~1560 cm

  invariant with orientation reveal considerable defects in what is nominally a single crystal. The phase of the experimentally obtained P/O diagram with respect to the calculated diagram in Figure 1 allows crystallographic assignment of the faces as shown in Figure 5. 

Figure 6: P/O micro-Raman spectra separated by 5 rotational degrees. The spectra are fitted for strain-free diamond (~1334 cm-1) and a strain component.			

In addition to the disorder manifest in the spectra, orientationally dependent strains can be identified through band fitting. The strain-free Raman band of diamond appears at 1333 cm

 . Previous work by other authors has demonstrated the crystallographic orientational variation of strain in diamond films and the use of polarized Raman spectroscopy to detect it (2–4). Figure 6 shows two micro-Raman spectra from the perpendicular polarization set (Figure 4) separated by 5 rotational degrees. Band fitting reveals that the relative strain-to-strain-free components vary with the direction of phonon propagation through the crystal. Indeed, the three-dimensional variation of the strain in this crystal contributes significantly to the deviation of the experimentally obtained P/O diagram from one calculated for an ideal crystal. 			

We have demonstrated through the application of Raman polarization selection rules that the orientation of a crystal can be determined from the phase shift of the experimental P/O micro-Raman diagram relative to that calculated for a particular crystal face. Furthermore, the efficacy of the P/O micro-Raman method for identification of disorder components or defects and the detection of directionally dependent compressive strain components in the diamond has been demonstrated.  

 is the manager of Raman applications at Horiba Scientific, in Edison, New Jersey. Direct correspondence to 

(2) S.-A. Stuart, S. Prawer, and P.S. Weiser, 

(3) T. Gries, L. Vandenbulcke, P. Simon, and A. Canizares, 

(4) T. Gries, L. Vandenbulcke, P. Simon, and A. Canizares, 

The orientation of a crystal can be determined from the phase shift of the experimental P/O micro-Raman diagram relative to that calculated for a particular crystal face.

Special Issues-06-01-2012

Raman Spectroscopy as a Quantitative Tool for Industry

Rapid Bacterial Diagnostics Via Surface-Enhanced Raman Microscopy

A Novel Method for Practical Implementation of the Shifted Excitation Raman Difference Spectroscopy (SERDS)

Application of Raman Polarization Selection Rules: Heterogeneous Solid State Structure