Raman Spectroscopy

What Raman spectroscopy measures

Shine a single-color laser at a sample and watch the light that bounces back. The overwhelming majority of photons scatter elastically — they leave with exactly the energy they arrived with. This is Rayleigh scattering, the same process that makes the sky blue, and it carries no information about molecular structure. But buried in that flood, roughly one photon in ten million does something stranger: it exchanges a small, precisely quantized packet of energy with the molecule and emerges a slightly different color. That tiny, telltale shift is the Raman effect, and reading it is what Raman spectroscopy does.

The shifted photons come in two flavors. If the photon gives up energy to set the molecule vibrating, it emerges lower in energy — a red-shift called a Stokes line. If the photon takes energy from a molecule that was already vibrating, it emerges higher in energy — a blue-shift called an anti-Stokes line. The energy traded is always one vibrational quantum, so the magnitude of the shift is identical for Stokes and anti-Stokes; only the sign differs. Crucially, the shift does not depend on the laser wavelength. Use a green 532 nm laser or a near-infrared 785 nm laser, and a carbon–carbon stretch still appears at the same Raman shift — about 1600 cm⁻¹ from the laser line. That absolute, source-independent shift is what makes the spectrum a fingerprint.

The mechanism: a virtual state and a deforming electron cloud

Quantum-mechanically, the incoming photon momentarily lifts the molecule to a short-lived virtual state — not a real electronic level, just a transient distortion of the electron cloud that lasts on the order of femtoseconds. The molecule then relaxes by re-emitting a photon. If it returns to the same vibrational level it started from, the scattered photon has the same energy (Rayleigh). If it lands one vibrational level higher, the photon is poorer by that quantum (Stokes). If it started excited and ends in the ground state, the photon is richer (anti-Stokes).

The deciding property is polarizability — how easily the molecule’s electron cloud deforms in the oscillating electric field of light. A vibration is Raman-active only if it changes the polarizability as the atoms move. This is why symmetric, non-polar bonds shine in Raman: the symmetric stretch of CO₂, the C=C double bond, the S–S disulfide bridge in proteins, and the rigid carbon lattice of diamond all strongly modulate polarizability even though they have little or no changing dipole. It is also why water is nearly invisible in Raman — its O–H stretch is a weak Raman scatterer — making the technique uniquely suited to aqueous solutions, living cells, and biological tissue where IR would be swamped by water absorption.

The Raman cross-section is genuinely tiny, around 10⁻³⁰ cm² per molecule, some twelve to fourteen orders of magnitude smaller than a typical fluorescence cross-section. Two strategies fight this weakness. Resonance Raman tunes the laser near a real electronic absorption of the molecule, boosting specific modes by factors of 10³–10⁶. Surface-Enhanced Raman Spectroscopy (SERS) exploits plasmonic gold or silver nanostructures whose hot spots amplify the local field; because the Raman signal scales roughly as the fourth power of that field, enhancements of 10⁶–10¹¹ are achievable — enough, in the best hot spots, to detect a single molecule.

Reading a Raman spectrum

A Raman spectrum plots scattered intensity (vertical) against Raman shift in cm⁻¹ (horizontal), conventionally with the laser line at zero on the right and increasing shift to the left. Two regions matter. The fingerprint region below about 1500 cm⁻¹ is dense with skeletal bending and stretching modes unique to each compound — it is what lets Raman tell apart two molecules with the same functional groups. The functional-group region from roughly 1500–4000 cm⁻¹ hosts characteristic stretches: C=O near 1700 cm⁻¹, C=C near 1600 cm⁻¹, C≡N near 2200 cm⁻¹, and C–H near 2900 cm⁻¹.

Carbon materials give the cleanest illustration. Graphite and graphene show a strong G band (~1580 cm⁻¹) from in-plane sp² stretching and a 2D band (~2700 cm⁻¹) whose shape distinguishes single-layer from multilayer graphene; a D band (~1350 cm⁻¹) appears only when defects break the symmetry, so the D/G intensity ratio is a direct, quantitative defect gauge used across the entire nanocarbon industry.

Selected diagnostic Raman bands

Bond / mode	Approx. Raman shift (cm⁻¹)	What it reports
S–S (disulfide)	500–550	Protein crosslinks; strong in Raman, weak in IR
Diamond C–C lattice	1332	Sharp single line; sp³ carbon, diamond identification
Graphitic D band	~1350	Defect / disorder in sp² carbon
C=C alkene	1600–1680	Unsaturation; symmetric, polarizable
Graphitic G band	~1580	In-plane sp² stretch; baseline for D/G ratio
C≡N / C≡C	2100–2260	Triple bonds; quiet spectral window
C–H stretch	2800–3000	Aliphatic vs aromatic hydrogen

Raman versus IR — complementary, not competing

Raman and infrared spectroscopy both interrogate molecular vibrations, yet they obey opposite selection rules. IR absorption requires a vibration that changes the dipole moment; Raman requires one that changes the polarizability. For any molecule with a center of inversion symmetry, the mutual exclusion rule guarantees that a vibration cannot be active in both: modes that are Raman-active are IR-silent and vice versa. The practical upshot is that the two techniques see different bonds, and analysts routinely run them together for a complete vibrational picture.

Raman vs. infrared (IR) spectroscopy

Property	Raman	Infrared (IR)
Physical process	Inelastic scattering of light	Direct absorption of IR photons
Selection rule	Change in polarizability	Change in dipole moment
Strongest for	Symmetric, non-polar bonds (C=C, S–S, C–C)	Polar bonds (O–H, C=O, N–H)
Water interference	Very low — ideal for aqueous samples	Severe — water absorbs strongly
Sample prep	Minimal; through glass/plastic possible	Often needs thin films or special cells
Spatial resolution	Sub-micron with confocal microscope	Diffraction-limited, longer wavelength
Main nuisance	Fluorescence background	Atmospheric CO₂ / H₂O bands

Inside the instrument

A modern Raman system is conceptually simple but optically demanding. A laser — commonly 532, 633, 785, or 1064 nm — illuminates the sample. Longer wavelengths cost signal (Raman intensity scales with the fourth power of frequency) but dramatically reduce fluorescence, the chief enemy of clean spectra, because fewer molecules absorb the redder light. The scattered light is collected and passed through a notch or edge filter that rejects the Rayleigh line by a factor of 10⁶ or more — without it, the unshifted glare would bury the Raman signal entirely. A diffraction grating disperses the remaining light onto a cooled CCD or InGaAs array, where each pixel reads one wavenumber.

Confocal Raman microscopes add a pinhole that rejects out-of-focus light, giving lateral resolution below one micron and the ability to map chemistry point by point across a surface. Tip-Enhanced Raman Spectroscopy (TERS) pushes further, combining SERS with a scanning-probe tip to break the diffraction limit and reach nanometer-scale chemical imaging.

Where it earns its keep

• Pharmaceuticals. Distinguishes polymorphs (the same molecule packed into different crystal forms with different solubility and efficacy), verifies identity through sealed bottles and blister packs, and supports process analytical technology on the production line.
• Carbon nanomaterials. The D, G, and 2D bands quantify graphene layer count and defect density, and identify single-walled nanotube diameter via radial breathing modes near 100–300 cm⁻¹.
• Art and heritage. Non-destructive pigment identification — distinguishing, say, genuine ultramarine from synthetic substitutes — helps date and authenticate paintings without taking a sample.
• Biomedicine. Because water is quiet, Raman images live cells and tissue; it is being developed to flag tumor margins during surgery and to fingerprint pathogens.
• Forensics and security. Handheld units identify explosives, narcotics, and hazardous chemicals in the field, often through transparent containers.
• Geology and gemology. Mineral and gemstone identification, and detection of inclusions; the 1332 cm⁻¹ diamond line separates real diamond from look-alikes.

Limits and pitfalls

• Fluorescence swamping. A single fluorescent impurity can emit a background millions of times stronger than the Raman signal; mitigated with longer-wavelength lasers, time-gated detection, or shifted-excitation subtraction.
• Sample heating. Focusing a laser onto a dark or absorbing sample can scorch or transform it; power must be managed carefully.
• Inherent weakness. Without resonance or surface enhancement, trace detection is hard — the unenhanced effect simply scatters too few photons.
• Calibration. Wavenumber and intensity axes drift; standards such as silicon (520.7 cm⁻¹) and known reference materials are used to keep spectra comparable across instruments.