Raman Spectroscopy (Raman)

The calculated energy difference between the incident light (known) and the Stokes/Anti-Stokes photons (measured during analysis) is equal to the energy required to excite a molecule to a higher vibrational energy level.  Because the vibrational modes of functional groups tend to be unique, measuring their energy in a sample provides valuable information on the types of molecules present.  Within each molecule there are multiple vibrational modes with the actual number present dependent on a number of factors including orientation of the atoms and bonds, the atomic mass of the atoms, bond order and H bonding.  The number of vibrational modes in a molecule can be calculated by:

​

      a.  Nonlinear Molecules:  3N-6 possible variations (where N = number of atoms)

      b.  Linear Molecules:  3N-5 possible variations 

​

Thus, as molecules become larger the number of signals present in the Raman spectra increase accordingly.  However, it must be stressed that not all of the vibrational modes will be 'Raman Active' meaning that only a portion will cause Raman scattering.   

​

It should be noted, that the process occurring in Raman spectroscopy is very different from that in Infrared Spectroscopy (e.g. FTIR) which is a type of absorption spectroscopy.  IR involves the absorption of infrared radiation at wavelengths corresponding to the transition of a molecule going from a ground to an excited Vibration State (see Figure 1) and the energy of that transition is characteristic to specific types of chemical bonds.  

​

In Raman Spectroscopy, by separating the scattered light by wavelength using a spectrophotometer one sees the Rayleigh scattered light equal to the wavelength of the incident light strongly detected while Raman scattered light is detected on both sides of this signal.  Raman scattered light detected on the shorter wavelength side relative to Rayleigh scattered light are called anti-Stokes lines, and those detected on the longer wavelength side are called Stokes line.  As discussed earlier, Stokes lines have considerably higher intensity than the Anti-Stokes lines which is why they are used for analysis.  In practice, filters are commonly used to remove much of the Rayleigh and Anti-Stokes lines. 

​

Fluorescence is often a concern in Raman spectroscopy as emitted fluorescence has similar energies to the Raman scattered photons.  Since the fluorescence efficiency is very high (typically 1-10%) relative to Raman scattering (<0.01%), it can easily swamp the Raman signal if it occurs.  Fluorescence occurs when the energy of the excitation photon gets close to the transition energy between two electronic states of a molecule.  If the excitation photon does not provide sufficient energy to the molecule, the required transition to generate fluorescence will not take place.  Fluorescence tends to be particularly problematic for highly colored and darker samples, dyes and chemical bonds containing elemental N.

​

Data

In Raman Spectroscopy, the various scattered wavelengths from a sample are measured.   (Energy and wavelength are related by the following equation:  E = hc/λ, where h is Planck’s Constant, c is the speed of light and λ is the wavelength.)  Each scattered wavelength is converted to wavenumber (1/wavelength) and the difference between this value and the wave number of incident light is calculated which corresponds to the vibrational modes present within the sample.  Raman spectrum plots this wavenumber difference (i.e. the molecular vibration information) on the horizontal axis and the intensity, which is related to the strength of the vibrational mode and the amount of the species present, on the vertical axis.

While data from pure compounds can be relatively simple, that of mixtures or larger molecules (e.g. proteins) may contain hundreds and even thousands of vibrational modes resulting in extremely complicated spectra.  

​

The ability to focus a laser to a small spot makes high resolution mapping/imaging with Raman possible.  Imaging Raman most commonly involves scanning the laser over the sample surface (typically by moving the sample stage) and acquiring spectra at every pixel.  The time required to obtain such images can be quite large depending on a number of factors including size of the image area, number of pixels and the acquisition time per pixel.  It has applications in a wide variety of fields.  It is commonly used to look at distribution of components and determine size of features.  It can also be used to look at crystallinity changes and phase across a sample as well as stress/strain distribution.  Many Raman instruments also have spatial filters that allow for confocal capabilities.  Confocal Raman provides spatial filtering not only laterally (i.e. the X-Y direction) but also vertically (i.e. depth or Z direction).  This capability allows the analysis of buried features if they can be observed and the laser focused onto them. 

   

Instrument

A Raman instrument consists of a laser that is shined onto a sample.  A microscope is used to identify the areas of interest for analysis.  The scattered light is collected (with lenses), filtered (notch or band pass filter) to remove the Rayleigh Scattered light and stray light, then sent onto a diffraction grating to spread the light out over the spectral range and finally to a detector (most commonly a CCD or electron multiplied CCD) to count the photons at each wavelength.

​

Lasers

Lasers are used as the light source because they have a narrow energy beam (continuous wavelength sources produce much broader signals affecting spectral resolution) and they produce a lot of photons which is necessary since the Raman process is very inefficient.  While a wide variety of lasers are used (from the UV to IR) those in the visible wavelength range are most common although lasers with near IR wavelengths are also frequently used.  Lower wavelength lasers have higher energy (i.e. produces more signal) and shallower information depths but can cause more fluorescence (thus, swamping the scatter signal) while high wavelength lasers have less energy, greater sampling depths and are less likely to induce fluorescence but the Raman efficiency (i.e. signal) drops and more heat absorption occurs (possibly altering the sample).  Some of the lasers commonly used on Raman instruments are:

​

1. 455nm (Violet)

2. 488nm (Cyan, Ar Laser)

3. 514nm (Green, Ar Laser)

4. 532nm (Green, Standard Diode Laser)

5. 633nm (Red, HeNe Laser)

6. 785nm (Near IR laser)

7. 1064nm (Near IR laser, Common for FT Raman)

 

Generally, one attempts to use as high of an energy laser (i.e. short wavelength) as possible to increase Raman efficiency without causing undue fluorescence issues.  For instance, the 532nm laser has an excitation efficiency that is 4.7 and 16 times greater than the 785nm and 1064nm lasers, respectively.  It should be noted that penetration depths are also dependent on the type of laser used as well as the material being analyzed.  The penetration depth decreases as the wavelength decreases (i.e. energy increases).  For instance the penetration depth of Si using a 633nm laser is 3um but only ~0.5um if a 488nm laser is used.  Typically, Raman systems are equipped with multiple lasers to allow the user to select a laser best suited for a particular application.

​

Filters

Two types of filters are generally used to remove unwanted light in the Raman instrument.  One type is the Notch (or Interference) Filter which just removes the intense Rayleigh Scattering by cutting off the entire spectral range from ~80-120cm-1 on both sides of the intense laser line.  This type of filter tends to be expensive and doesn’t allow detection of low frequency Raman modes below ~100cm-1.  The second type is the Band Pass Filter which removes BOTH Rayleigh and Anti-Stokes lines.  These tend to be considerably cheaper and are more common since Anti-Stokes lines are often not used for most applications. 

​

A third way of removing unwanted light is through the use of multiple dispersion stages.  In this set-up two or three spectrometers are used to filter out the unwanted light negating the need for Notch and Band Pass Filters.  This type of set-up allows frequencies as low as 3-5cm-1 to be efficiently detected which is not possible using standard filters.

​

Dispersion Element

A diffraction grating is most commonly used to separate the scattered light by ‘spreading’ it out over the spectral range.  Different gratings can be used to achieve high resolution (light dispersed over a large/wide range) and low resolution (light dispersed over a narrow range). 

​

Detectors

A variety of detectors are used in Raman instruments with CCD (Charge Coupled Device) and Electron Multiplied CCD (EMCCD) being most common.  The CCD is a Si based array/multichannel detector that works by its individual pixels or elements interacting with light which causes a charge to build up.  The intensity of the charge is related to the brightness of the light and the length of its interaction with the respective pixel.  The charge is read and related back to the light impinging on the detector pixels.  Each different wavelength of light that is diffracted from the dispersion element impinges upon a different pixel.  Thus, the detector is both quite sensitive and capable of detecting the entire spectrum simultaneously.  The EMCDD is a specialized CCD designed so that a low number of photons can be enhanced resulting an increased signal level.  These detectors are more expensive but are useful for high-speed imaging purposes where many spectra must be collected very quickly.

      

Specialized Raman Techniques

UV Raman

As mentioned earlier, lasers in the visible wavelength and near IR range are most often utilized in Raman spectroscopy.  Nevertheless, several manufacturers have instruments that can be equipped with UV lasers (244-364) as well.  In theory, UV Raman functions the same as conventional (i.e. normal) Raman.  However, in practice instruments utilizing UV lasers must employ modified mirror coatings, microscopic objectives, diffraction gratings and CCDs catered for the higher energies of the UV lasers.  There are a number of advantages of using a UV laser as the light source.  It can provide increased sensitivities as well as shifting the appearance of fluorescence signals so that they no longer overlap with the Raman peaks.  Additionally, for some systems the penetration depth is very shallow (nm range) allowing Raman to analyze thin films such as those commonly found in semiconductor devices.  However, UV lasers are typically larger and more expensive and samples are more apt to be damaged by their higher energies.     

​

SERS

Surface-Enhanced Raman Spectroscopy (SERS) utilizes the fact that molecules adsorbed on certain metal surfaces results in Raman signals that are 5-6 orders of magnitude higher than the same molecules found in bulk.  This signal intensity increase is thought to result from either enhancement in the polarizability of the molecules or the electrical field.   The most commonly used metal substrates for SERS are electrochemically etched silver electrodes as well as silver and gold colloids with average particle size below 20nm.  One disadvantage of SERS is that the spectra are often more difficult to interpret.  This difficulty arises from the fact that many signals that are weak and unobservable in normal Raman appear as intense signals in SERS.  Additionally, SERS allows the detection of contaminants not visible in regular Raman.  Another problem is that some signals that are intense in regular Raman do not appear in SERS because of chemical interaction with the metal substrate.  However, the use of SERRS (Surface Enhanced Resonance Raman Spectroscopy) was developed to overcome some of these limitations.  SERRS spectra often closely resemble regular Resonance Raman spectra making them easier to interpret. 

 

TERS

The ultimate goal of Tip Enhanced Raman Scattering (TERS) is to obtain Raman images on the nm scale, something not possible with conventional Raman where spatial resolutions are typically only ~0.5-1um due to diffraction limitations.  Improved resolution is achieved by combing Raman Spectroscopy with AFM.  An AFM with a probe modified with SERS active metals or nanoparticles is synchronized to scan within the same location as the Raman laser beam.  Signal originates both from the laser beam and from the immediate vicinity of the tip.  Because the SERS enhancement can be as much as 14 to 15 orders of magnitude higher than the signal from the laser illuminated region, the signal from the tip region can dominate.  When this occurs, the vast majority of the signal detected originates from an area that is dependent on the size of the tip.  Since AFM tips typically have dimensions of <100nm, spatial resolutions of this range are possible.  However, TERS success is difficult and not guaranteed due in part to the smaller volume of molecules being sampled compared to conventional Raman as well as the problems associated with SERS.  Thus, TERS is most often relegated more to R&D applications.

​

Resonance Raman

Resonance Raman is another method of gaining signal enhancement over the relatively poor sensitivities of conventional Raman.  Such enhancement is accomplished in this case by choosing an excitation wavelength that overlaps (or is very close to) an electronic transition (see Figure 1).  Such wavelengths are typically in the range of UV-Visible absorption.  The resonance effect occurs as the excitation laser frequency crosses frequencies of the electronic excited states and resonates with them.  Raman bands originating from electronic states between those states are enhanced by 3-5 orders of magnitude.  However, not all bands resonate meaning that Resonant Raman spectra typically have fewer signals than normal Raman.  Additionally, since the excitation corresponds with UV-Vis absorption, fluorescence can be considerably more problematic than with normal Raman.  Nevertheless, some materials exhibit signal enhancement without significant fluorescence.  Resonance Raman most commonly finds success with molecules containing chromophores (e.g. conjugated double bonds) such as some environmental pollutants which are typically present in the ppm-ppb levels.  Conventional Raman is unable to detect these pollutants at these levels.  Because of the need to choose specific excitation wavelengths that are close to the electronic transitions of the molecules of interest tunable lasers are most frequently used.  Obviously, Resonance Raman is not well-suited for the analysis of unknowns but rather for the detection of specific species.     

​

Comparison

FTIR is a complementary measurement technique to Raman Spectroscopy. Both measure the energy required to change the vibrational state of a molecule but the mechanism and selection rules differ.  This is due in large part because FTIR is based on absorption of light while Raman measures the inelastic scattering of light.  However, the spectral positions of their signals for specific functional groups are the same.  Typically, functional groups that give intense signals via Raman Spectroscopy produce weak FTIR signals and vice versa.  This is because Raman detects vibrations where the polarizability of the molecule changes (i.e. a dipole moment is induced; the electron cloud of the molecule undergoes positional change) while FTIR detects vibrations that result in a net change in dipole moment of the molecule.  So for a centrosymmetric molecule asymmetrical stretching and bending are IR Active but Raman Inactive while symmetrical stretching and bending are IR Inactive but Raman Active.  For molecules without a center of symmetry, each vibrational mode may be IR Active, Raman Active, both or neither.  Thus, these two techniques provide complementary information. 

 

The relatively cheaper and easier to use instrumentation tends to make FTIR the technique of choice for the general analysis of organic materials.  However, Raman tends to be a much better analysis tool for inorganic compounds.  This is because it can easily analyze aqueous solutions (which many inorganics are soluble), it can detect metal-ligand bonds which occur in the range of 100-700cm-1 (a region difficult to study by FTIR), and can provide information on composition, structure and stability of coordination compounds.  Raman is typically good at looking at metal oxides but not for looking at pure metals or alloys.  Raman also is often used for the analysis of elemental C as its different forms give unique Raman spectral signatures.   The much better spatial resolution makes Raman Spectroscopy a better choice for any sample in which a small area needs analysis.  Imaging by Raman also provides better resolution than FTIR imaging.