Thoughts About ICORS 2010 and Where Raman Spectroscopy Is Headed

What I see as being really interesting is how the applications and instrumentation developments are driving each other. Raman, of course, is mainly a spectroscopy of vibrational transitions, and as such, is loaded with information-content. For a variety of reasons, IR absorption, in particular FT-IR, is more mature as an analytical technique than Raman, but with the introduction of the holographic notch filters in 1990 (1), air-cooled lasers, and low-noise CCD detectors, Raman instrumentation has become smaller, less expensive, and easier to use, so its successful implementation in many disciplines has created new applications. This in turn is pushing the instrumentation to do more.

Fran Adar

Microscopy is an area where Raman has had a potential impact since its introduction in 1973 (2). I say potential because as long as the instrumentation was large with utility-intensive lasers and little computational power, its use was limited. But now that proof-of-principle has been shown for many applications, and the instrumentation is more user friendly, people want higher sensitivity and higher spatial resolution, and these are two of the areas that I want to discuss.

Applications

We see a lot of activity in the study of carbon nanomaterials such as nanotubes and graphene (single hexagonal carbon sheets). There is no space to review all the carbon talks and posters, but we want to mention at least, the plenary talk by Andrea C. Ferrari (3) as well as talks by Mildred Dresselhaus (4), Marcos Pimenta (5), and Janina Maultzsch (6). These materials have fascinating resonance. Raman properties following from their unique electronic properties and because of their high electronic mobility are of interest in integrating into electronic circuitry. Raman studies will aid in selecting the particular species and engineering their use.

I discussed surfaced-enhanced Raman scattering (SERS) in a column almost two years ago. The topic was selected because of its potential to provide sensitivity to things that cannot be detected or studied presently. Because I want to discuss the sensitivity issue in this column, I will start with a review of some of the SERS work reported at ICORS 2010. John Lombardi (7) reviewed the various theories of SERS, and discussed how the mechanisms determining these effects can be differentiated. In one of the plenary talks, Martin Moskovits (8) reviewed the status of robust SERS platforms that can enable reliable application of this technology to reproducible analysis. In his opinion, this technology is robust enough for reliable commercial application of Raman. If he is correct, specialized Raman products will be used for routine analytical chemistry that can be applied to clinical chemistry, industrial analysis, homeland security, and so forth.

Using a "lab-on-a-chip" to study chemical reactions, separations, analysis, and bio-organisms has always seemed to have potential. For example, Anne Marz (9), working with Juergen Popp, demonstrated sensitivity and robustness for detecting specific molecules in blood assays. In one case, they looked for an antibiotic, and in the second case, they looked for a metabolite of an immunosuppressant agent that is myelotoxic. They used a sum vector machine (SVM) to evaluate the accuracy of their enzyme activity test. The goal was to provide a means to adjust the drug dosage within a range where it would be active, but not generate the toxic side product. Mischa Bonn (10)and Marcus Motzkus (11) have been developing micro-CARS for the lab-on-a-chip application, as well as some other applications, and their innovations will be discussed in the instrumentation section of this column.

For me, the greatest excitement lies in the developments in the biological and clinical sciences. My postgraduate career started in the Johnson Foundation at the University of Pennsylvania, which was the Department of Biochemistry and Biophysics. When I left academia for the commercial life, the hope was that Raman could play a role in these areas. But, at the time (1978), I looked at the complexity of these systems and felt that even the microscope offered little hope of being able to unravel complex biochemical and physiological states. I am glad that 30 years has changed this. Let me try to select some areas in which Raman is having an impact. There is a lot to choose from, as many researchers who have been active in these areas gave talks during the conference. In addition, I should report that there was a tribute to Michael Feld, a pioneer in this field, who passed away a few months ago.

Paul Carey (12) gave a talk whose title ("Raman Revolution in Structural Biology") in a way describes the Raman potential for impact. He makes the point that Raman bridges the gap between structure, as determined by X-ray diffraction, and functionality at the molecular level. He used the study of RNA polymerase as an example. For this enzyme, Murakami had already developed a method to capture three-enzyme intermediates. Making use of spectra of microcrystals, Raman difference methods, as developed by Bob Callendar, and isotopically labeled nucleotides, he was able to characterize the enzymatic process and then, using that information, follow Mg binding to aspartate.

Juergen Popp (13) showed that spectra of single cells and micro-organisms can be recorded, and noted that microfluidics can be used for cell sorting (it takes 10 s to collect a spectrum), and he has produced an 85% correct recognition rate. In the case of bacteria, quickly identifying an infectious strain means that the correct antibiotic can be administered, avoiding sepsis. This is a good example illustrating how Raman can have payoff. People have been measuring and classifying bacteria with their spectra for more than 10 years. Combining these measurements with lab-on-a-chip cell separation and sorting can make this a clinically applicable technology.

James Chan (14) delivered another talk on the analysis of single cells. Using a laser to trap the cell and suspend it in solution ("laser tweezers"), the signal from the glass support or container is eliminated by confocal optics. This, of course, can be combined with microfluidics. He mentioned many interesting uses of the measurement. At the end, he discussed a method to scan the laser beam to produce multiple optical traps that could then be multiplexed onto the spectrograph slit. His goals are to produce 50 traps in a line and to make a two-dimensional microwell device for a cellular array.

Tatyana Chernenko (15), working with Max Diem, discussed their analysis of cells, tissues, and drug delivery devices. She discussed two multivariate analysis techniques used to analyze their spectra. Hierarchal cluster analysis (HCA) searches for similarities, producing a dendogram and a pseudocolor map, and reveals biochemical components that cannot otherwise be detected. Vertex component analysis (VCA) searches for dissimilarities in the data set, identifying pure component spectra, and has been successful in localizing drug nanoparticles in cells. Application examples included characterizing an embryonic colony to determine when an embryo can be implanted, and characterizing phospholipid-based nanoparticles used for drug delivery of insoluble molecules that are engineered to bypass lysosomal degradation so that the drug can be delivered to the target tissue or organelle.

Rick Mendelsohn (16) currently is studying skin pharmacology. Principle component analysis of a depth profile of the skin has revealed five factor loadings that are identified with the stratum corneum, the epidermis, the dermis, a lipid inclusion composed of ordered lipids with cholesterol, and nuclei. He studied the penetration of a drug that has to be delivered in the form of a "pro-drug" and its conversion to the active form. He detected the drug in the solvated state and in the crystalline form, a piece of information that the drug company would probably be interested to know.

Mike Morris (17) has been studying bone structure and bone disorders in vitro for quite a while. Using spectral interpretation, he is able to characterize the molecular origins of disease states, and with polarization analysis, proposes the most likely orientation of the structures in bone and tissue. A major hurdle to such work is the inherent light scattering of tissue, which, one would think, would degrade the spatial resolution of the measurement. To work around this limitation, he has collaborated with Pavel Matousek (18) on spatially offset Raman spectroscopy (SORS). The essential point here is that if you want to probe below the surface of a turbid material, you can collect the light at a distance from the point of illumination, and this distance is related to the depth from which the signal is generated.

Michel Pezolet (19) has spent years studying spider silk. There are seven types of glands that deliver seven types of silk, each with different mechanical properties as determined by stress/strain curves. Some fibers are highly extensible, some are not. Orientation is documented clearly in the polarization behavior of the spectra. These properties can be correlated with the amino acid composition and the Amide I band which provides information on the conformation of the amide linkage ( β sheet versus helical structures versus amorphous phase). Understanding the structure of this unique material can have commercial ramifications, but it also is useful to note that Mike Morris cited Pezolet's methodology in developing the protocol for interpretation of his collagen and mineral spectra.

Developments of New Raman Methods and Instrumental Innovations

The Raman microscope, as conceived and developed in the early 1970s (2), took advantage of the potential spatial resolution, with the diffraction limit expected to be, and then confirmed, at approximately a half wavelength of the laser light. During the ensuing 25 years, the equipment was used to study polymers, ceramics, semiconductors, catalysts, glasses, works of art, and on and on. Even though performing analysis with 0.5-µm spatial resolution was of interest, there was always the nagging question as to whether a smaller spot could be achieved to study smaller features. In the case of polymers, there is the issue of the size of the particles of one polymer dispersed in a second continuous phase, in the case of semiconductors, there is the issue of lithographic feature size become smaller as the scale of integration became larger, in the case of oxides (ceramics, catalysts, works of art) identification of interfaces and surfaces that could be of different phase compared to the major phase.

Today we know that "near field" optics can provide spatial resolution better than the diffraction limit. In this case, the light is presented to a sample through an aperture, and because the sample is in close proximity, the aperture size defines the spatial resolution. One way to create such an aperture is with a fiber-optic whose tip is tapered to a submicrometer-sized aperture. These fibers typically were coated with metal to prevent photon leakage because the tapering destroys the confinement properties of the fiber. Raman was attempted through such fibers, but the transmission was so low that the technology proved to be impractical.

As an alternative a technique called tip-enhanced Raman scattering (TERS) has been developed. This is an apertureless technique in which a gold or silver coating or particle is deposited on an atomic force microscopy (AFM) tip. When the tip is in close proximity to a sample, it functions as a SERS particle to enhance the signal. The AFM system is integrated with a Raman microscope, and the difference between the signal with the tip in close proximity to the sample, and then withdrawn, is calculated and is equal to an enhanced signal from the small volume around the particle, whose size determines the spatial resolution.

One of the Tuesday afternoon sessions was entitled "Tip-Enhanced and Near Field Raman I." The opening talk was by Aaron Lewis (20) of the Hebrew University in Jerusalem, who, having been involved in the field since the 1970s, was able to review the history. His implementation on current instruments is to use a tapered glass fiber with a gold particle on the tip. The tip approaches the sample at something like a 60° angle so that the standard optical scheme of a Raman microscope can be used. This configuration has the additional advantages of enabling the electric vector of the laser to have a component along the particle axis, further enhancing the Raman effect, and also avoiding generation of a signal from silicon in AFM tips manufactured from this material.

There are other ways to implement TERS. The easiest is to use an inverted microscope for those cases in which the sample is transparent. The AFM system can be mounted on top of the microscope, with optical access to the sample from below. But for opaque samples, an upright microscope usually is used, and it is the axis of the laser and Raman light that is diverted. Talks applying this technology to the fields mentioned at the beginning of this column were given by well known leaders in the field: Lukas Novotny (21), Satoshi Kawata (22), Francois Lagugne-Labarthet (23), Volker Deckert (24), and Bruno Pettinger (25), and illustrate the goal to probe semiconductors and biomedical systems well below the diffraction limit. As an example of the importance of these developments, Pettinger discusses single-molecule detection using TERS and Novotny showed images of carbon nanotubes recorded with 15-nm resolution and 50 ms/pixel.

While TERS addresses the goal for higher spatial resolution, it does not achieve high sensitivity for rapid measurements. The sensitivity to vibrational spectroscopy is being addressed by developments in microcoherent anti-Stokes Raman scattering (CARS) and stimulated Raman scattering (SRS). Hideako Kano (26) discussed a new CARS method that uses a white light laser for broadband Stokes to enable the simultaneous detection of the spectrum out to 3500 cm^-1. The white laser is produced in a photonic crystal fiber. But CARS usually is subject to distortions due to the nonresonant background that Kano and his group have been able to suppress using a maximum entropy method (MEMS). Mischa Bonn (10) is also applying CARS to study chemistry in confined regions: microfluidic devices, heterogeneous catalysts, and small lipid droplets. He has applied the MEMS algorithm after isolating the resonant CARS signal by using phase separation; the algorithm is discussed at amolf.memcars.nl. Marcus Motzkus (11) is using picosecond and femtosecond lasers to produce CARS and nonlinear Raman spectra. In his multiplexed CARS experiment, he uses one femtosecond laser to eliminate the requirement for synchronization and thus enables "single-beam CARS." The width of the laser pulse determines the spectral coverage of the CARS output. He uses coherent control to restore spectral resolution, and a liquid crystal mask at the Fourier plane of the laser pulse to recover the spectral information. His goal is to study lipids in the CH region which will require 10-fs pulses and will be implemented on the next generation of his equipment. He uses polarization to reduce the nonresonant background, and heterodyne detection for low concentration species. At the end of the conference Sunney Xie (27), another well-known pioneer of micro CARS, delivered a plenary talk on stimulated Raman microscopy that also offers label-free imaging of biological cells and organisms. He actually commented that he now believes that micro SRM has more potential than micro CARS, and this idea was explored more fully at the satellite meeting mentioned earlier. And on the opening day, Richard Mathies (28) described his use of femtosecond lasers to excite stimulated Raman spectra and study the underlying physics and physical chemistry of the photochemical processes.

Pavel Matousek has developed techniques for examining highly turbid materials (18), in which microscopy runs into the largest difficulties. Originally, he did time-resolved measurements (picosecond time scale) to determine from how deep in the sample the light was coming, but later realized that time could be converted to spatial separation, resulting in the acronym SORS. He is now using an Axicon component to excite the sample in an annulus and collect on axis. This scheme has the advantage of diffusing the light on a laser-sensitive material. He has also demonstrated the advantage of transmission Raman (TR) to provide spectra that can quantify bulk compositions. This already has been commercialized for QA and QC of pharmaceutical tablets.

One more instrumentation development deserves reporting. A new type of filter based on a volume Bragg grating has been engineered as a notch filter and demonstrated to enable the acquisition of spectra to better than 5 cm^-1 from the exciting line (29). These filters are available for lasers at 632.8, 532, 514.5, and 488.0 nm. The capability is of interest for recording longitudinal acoustic modes (LAMs) in polymers and folded acoustic modes in layered semiconductors. Their performance rivals that of the best triple spectrographs, except that they of course are not tunable.

Chemometrics or Multivariate Analysis

My final remarks deal with the use of chemometrics for the processing of Raman data. Because of the complexity of the Raman spectra (which is why they are so information-rich) it would be impossible to produce images of most materials without the ability to include information from the entire spectrum. This requires sophisticated software based upon statistics (collect enough data to average out the noise) and matrix algebra (one axis for the spectrum and a second axis for the varying quantity, be it spatial position, time, concentration, and so forth). Discussion of multivariate analysis is far beyond what can be done here, but I do want to say how important it will be for many of the developing Raman fields to reach their potential. There are some commercial software packages available, or you can write your own using a mathematics package. Be forewarned, learning this field is not for the faint-hearted.

Conclusion

Being a Raman spectroscopist used to be for people who had patience and didn't mind tending their instruments in the dark. With the simplification of the equipment has come new capabilities that match the applications that were always potentially of interest. It's exciting for me to witness these changes, and to participate in the evolution in some small way.

Fran Adar is the Worldwide Raman Applications Manager for Horiba Jobin Yvon (Edison, NJ). She can be reached by email at fran.adar@horiba.com

References

(1) M Carrabba, K.M. Spencer, C. Rich, D. Rauh, Appl. Spectrosc. 44, 1558–1561 (1990).

(2) M. Delhaye and P. Dhamelincourt, J. Raman Spectrosc. 3, 33–43 (1975).

(3) Raman Spectroscopy of Graphene: State of the Art, ICORS 2010, Boston, Massachusetts.

(4) Raman Spectroscopy of Carbon Nanotubes, ICORS 2010, Boston, Massachusetts.

(5) Resonance Raman Spectroscopy in Carbon Nanostructures, ICORS 2010, Boston, Massachusetts.

(6) Raman Spectroscopy of Carbon Nanostructures, ICORS 2010, Boston, Massachusetts.

(7) A Unified Theory of Surface Enhanced Raman Scattering, ICORS 2010, Boston, Massachusetts.

(8) Transforming SERS into a Dependable Platform for Ultra-Sensitive Molecular Sensing, ICORS 2010, Boston, Massachusetts.

(9) Towards an Analytical Tool Based on Lab-on-a-Chip-SERS (LOC-SERS) for Detection of Drugs in Complex Matrices, ICORS 2010, Boston, Massachusetts.

(10) Quantitative CARS – Chemistry in Confinement, ICORS 2010, Boston, Massachusetts.

(11) Nonlinear Raman Spectroscopy with Shaped Femtosecond Laser Pulses, ICORS 2010, Boston, Massachusetts.

(12) Raman Revolution in Structural Biology, ICORS 2010, Boston, Massachusetts.

(13) Raman Spectroscopic Characterization of Single Cells, ICORS 2010, Boston, Massachusetts.

(14) Laser Tweezers – Raman Spectroscopic Analysis of Single Cells and Their Dynamics, ICORS 2010, Boston, Massachusetts.

(15) Non-Invasive Raman Spectroscopic Imaging in Biology and Pharmaceutical Sciences (title modified from the published title Non-Invasive Imaging of Modified Liposomal Pharmaceuticla nanocarriers by Raman Microscopy), ICORS 2010, Boston.

(16) Vibrational Spectroscopy and Micro Spectroscopic Imaging – Applications to Skin Pharmacology and Wound Healing, ICORS 2010, Boston, Massachusetts.

(17) Non-Invasive and Minimally Invasive Raman Spectroscopic Diagnostics for MusculoSkeletal Tissue Disorders, ICORS 2010, Boston, MA.

(18) Spatially Offset Raman Spectroscopy — Emerging Concepts and Applications, ICORS 2010, Boston, Massachusetts.

(19) Structure Function Relationships in Spider Silk, ICORS 2010, Boston.

(19) Probe and Instrument Development for Tip-Enhanced Raman Scattering and Shadow Near-Field Scanning Optical Microscopy, ICORS 2010, Boston, Massachusetts.

(20) Near Field Raman Microscopy and Spectroscopy of Carbon Nanotubes, ICORS 2010, Boston, Massachusetts.

(21) Tip-pressurized Near Field Raman Microscopy: A Breakthrough towards Molecular Resolution, ICORS 2010, Boston, Massachusetts.

(22) Raman Spectroscopy of Single Semiconductor Nanowires: from Confocal Microscopy to TERS, ICORS 2010, Boston, Massachusetts.

(23) Tip-enhanced Raman Scattering Sensitive, Label-free, Nanoscale , and Imaging and Characterization of Caveolae with TERS during Stimulated Wound Healing, ICORS 2010, Boston, Massachusetts.

(24) Single Molecule Surface and Tip Enhanced Raman Spectroscopy, ICORS 2010, Boston, Massachusetts.

(25) CARS Molecular Fingerprinting using a White Light Source, ICORS 2010, Boston, Massachusetts.

(26) Stimulated Raman Scattering Microscopy, ICORS 2010, Boston, Massachusetts.

(27) Femtosecond Stimulated Raman Spectroscopy, ICORS 2010, Boston, Massachusetts.

(28) TP113 Alexandra Rapaport et.al, Very Low Frequency Stokes and Anti-Stokes Raman Spectra Accessible with a Single Multichannel Spectrograph and Volume Bragg Grating Optical Filters, ICORS 2010, Boston, Massachusetts.