Raman Scattering and the Rotational/Vibrational Raman Effect

Raman scattering takes the quantity that governs how a molecule responds to light — the polarizability $\alpha$ in the induced dipole $\mathbf{p}=\alpha\mathbf{E}$ — and asks what happens when the molecule is also vibrating or tumbling as the field drives it. The preceding molecular articles built the two ladders this effect climbs: the rigid-rotor levels $F(J)=BJ(J+1)$ and the harmonic-oscillator levels spaced by $\omega_v$ . Raman scattering is what you see when a scattered photon exchanges exactly one rung with those ladders, emerging shifted by that quantum. Because that shift equals an internal energy spacing and is independent of the laser colour, a Raman spectrometer is a vibrational and rotational spectrometer that works in the visible instead of the infrared or microwave. Three facts organize everything below. First, the selection rule is a changing polarizability, not a changing dipole: a mode is Raman-active when $(\partial\alpha/\partial Q)_0\neq0$ , complementary to the infrared rule, which is why $\mathrm{N_2}$ , $\mathrm{O_2}$ and $\mathrm{H_2}$ — silent in the infrared — shine in Raman. Second, because the polarizability is a rank-2 tensor, rotational Raman obeys $\Delta J=0,\pm2$ , a comb spaced by $4B$ rather than the $2B$ of ordinary rotation. Third, the process is a weak two-photon event through a virtual state, formalized by the Kramers–Heisenberg–Dirac polarizability, and it flares up only when the laser is tuned near a real electronic level.

Key takeaways

Raman scattering is inelastic scattering of light by molecules: the scattered photon leaves shifted by a molecular quantum, $\omega_s=\omega_0\mp\omega_{\mathrm{mol}}$ , Stokes redshifted and anti-Stokes blueshifted.
A mode is Raman-active when it modulates the polarizability, $(\partial\alpha/\partial Q)_0\neq0$ — complementary to infrared activity, which needs a changing dipole.
Vibrational shifts equal the vibrational quantum ( $\mathrm{N_2}$ at $2331\ \mathrm{cm^{-1}}$ ); pure rotational Raman obeys $\Delta J=0,\pm2$ , giving lines spaced by $4B$ with a $12B$ -wide gap straddling the laser.
The anti-Stokes/Stokes ratio is fixed by the Boltzmann population and an $\omega^4$ factor — a calibration-free, non-contact thermometer.
Scattering is weak — cross-section $\sim10^{-30}\ \mathrm{cm^2/sr}$ , roughly $10^{-3}$ of Rayleigh — so lasers, sensitive detectors, or resonance and coherent (CARS, SRS) enhancement are essential.
Key idea: a mode is Raman-active if it modulates the molecular polarizability, $(\partial\alpha/\partial Q)_0\neq0$ ; the incident field then acquires sidebands at $\omega_0\pm\omega_{\mathrm{mol}}$ , which are the Stokes and anti-Stokes lines.

The core insight

A vibrational or rotational mode is Raman-active if it modulates the molecular polarizability, $(\partial\alpha/\partial Q)_0\neq0$ ; the incident field then acquires sidebands at $\omega_0\pm\omega_{\mathrm{mol}}$ , which are the Stokes and anti-Stokes lines.

Where this lives today

Modern Raman spectroscopy is a laser instrument: a continuous-wave laser (532, 633, or 785 nm), a holographic notch filter to kill the Rayleigh line, a grating spectrograph, and a cooled CCD turn the faint sidebands into a routine molecular fingerprint. Confocal Raman microscopes map composition at sub-micron resolution with no labels or sample preparation, so pharmaceutical firms verify polymorphs in a tablet, geologists identify minerals in a thin section, and the Perseverance rover's SuperCam and SHERLOC read Martian mineralogy and organics from the surface itself. Where the spontaneous signal is too weak, three tricks multiply it: resonance Raman tunes the laser onto an electronic band of a chromophore (a mainstay for heme proteins and photosynthetic pigments); SERS uses plasmonic fields at a nanostructured metal surface to lift the signal by $10^{8}$ or more, reaching single molecules; and coherent CARS and stimulated Raman drive the vibration with two lasers whose difference matches $\omega_v$ , enabling label-free imaging of living cells. The same stimulated effect is a photonics workhorse in another guise: in silica fibres it shifts pump energy down by the glass phonon band (peak gain near $13\ \mathrm{THz}$ ), and fibre Raman amplifiers use it to extend the gain bandwidth of long-haul optical links.

The physical picture

Picture the molecule as a deformable cloud of charge with a certain "squeezability" — its polarizability $\alpha$ . The field pushes the electrons back and forth, building an induced dipole $\mathbf{p}=\alpha\mathbf{E}$ that re-radiates. If $\alpha$ were fixed, the dipole would radiate only at $\omega_0$ : Rayleigh scattering, elastic and colour-preserving. But let the nuclei stretch or the molecule tumble, and the squeezability becomes a function of time, $\alpha(t)$ . A carrier at $\omega_0$ times an envelope at $\omega_{\mathrm{mol}}$ is amplitude modulation, and modulation spills energy into sidebands at $\omega_0\pm\omega_{\mathrm{mol}}$ — the Raman lines, Stokes on the red side where the molecule kept a quantum, anti-Stokes on the blue side where it gave one up.

Figure

Energy-level view of Raman scattering. A photon is taken up to a short-lived virtual state and re-emitted. Rayleigh returns elastically; Stokes leaves the molecule one vibrational quantum higher so the photon is redshifted; anti-Stokes starts from an already-excited level and leaves the photon blueshifted. The rung spacing is the N₂ vibrational quantum ω_v ≈ 2331 cm⁻¹.

Two features of that picture already do real work. Anti-Stokes must start from an already-excited molecule, so its brightness relative to Stokes tracks the thermal population of the upper level — the basis of the calibration-free thermometer. And rotational Raman skips by two units, $\Delta J=\pm2$ , because the polarizability is a rank-2 (quadrupole-like) tensor: it returns to itself after a half-turn, coupling states two quanta of angular momentum apart, unlike the rank-1 dipole that drives $\Delta J=\pm1$ in absorption. The second figure shows the fingerprint this leaves: a comb of pure rotational lines with a wide, clean gap around the laser.

Figure

Pure rotational Raman spectrum of N₂ from a single laser line at shift zero. The rank-2 polarizability enforces ΔJ = ±2, so lines appear only at shifts 6B, 10B, 14B, … on each side. With B ≈ 2.0 cm⁻¹ they are spaced by 4B ≈ 8 cm⁻¹, and the first lines sit 6B ≈ 12 cm⁻¹ out, leaving a clean 12B ≈ 24 cm⁻¹ gap across the laser.

The whole effect is faint because it is second order: the photon must reach a virtual state and come back, an intrinsically small two-photon amplitude. Only when $\omega_0$ nears a real electronic transition does the virtual state become nearly real, the denominator collapse, and the cross-section climb a million-fold — resonance Raman.

The core equations

The starting point is the induced dipole: a field polarizes the molecule and the oscillating dipole radiates. For a monochromatic drive, this is the source of all scattered light, elastic or not.

\mathbf{p}(t)=\alpha\,\mathbf{E}(t),\qquad E(t)=E_0\cos\omega_0 t \qquad (1)

where: $\mathbf{p}$ — induced electric dipole moment; $\alpha$ — molecular polarizability, a symmetric rank-2 tensor in general; $E_0,\ \omega_0$ — amplitude and angular frequency of the incident field. The field drives the charge and the dipole re-radiates; if $\alpha$ is constant the radiation stays at $\omega_0$ . The Raman physics enters when the polarizability varies with a nuclear coordinate. Expanding $\alpha$ to first order in $Q$ and letting $Q$ oscillate harmonically is the entire content of the classical selection rule.

\alpha(Q)=\alpha_0+\left(\frac{\partial\alpha}{\partial Q}\right)_0 Q+\cdots,\qquad Q(t)=Q_0\cos\omega_v t \qquad (2)

where: $Q$ — normal (vibrational) coordinate; $\alpha_0$ — equilibrium polarizability; $(\partial\alpha/\partial Q)_0$ — polarizability derivative, the switch that must be nonzero for Raman activity; $\omega_v$ — vibrational angular frequency. Only the derivative term carries mode information. Feeding (2) into (1) and using a product-to-sum identity — done in full below — leaves three radiated frequencies. The shift law, for a vibrational or a rotational quantum, reads:

\omega_s=\omega_0\mp\omega_{\mathrm{mol}} \qquad (3)

where: $\omega_s$ — scattered photon frequency; $\omega_{\mathrm{mol}}$ — the exchanged rotational or vibrational quantum; the minus sign gives Stokes (molecule gains energy), the plus sign anti-Stokes (molecule loses energy). The shift, not the absolute colour, carries the spectroscopy, so any laser works. The activity condition is worth setting beside its infrared counterpart. Their complementarity is the most useful organizing fact in vibrational spectroscopy, so both rules are shown together.

\left(\frac{\partial\alpha}{\partial Q}\right)_0\neq0\ \ (\text{Raman}); \qquad \left(\frac{\partial\mu}{\partial Q}\right)_0\neq0\ \ (\text{IR}) \qquad (4)

where: $\mu$ — the electric dipole moment relevant to infrared absorption. Infrared listens for a wagging dipole, Raman for a breathing polarizability; a symmetric stretch hides the dipole but flaunts the polarizability. For rotations, the tensor rank fixes the ladder step at two. The rigid-rotor terms then place every line — the equation a rotational-Raman lidar actually fits.

\Delta J=0,\pm2;\qquad \Delta\tilde\nu_{J\to J+2}=\mp B(4J+6),\quad J=0,1,2,\dots \qquad (5)

where: $\tilde\nu$ — wavenumber shift from the exciting line; $B=\hbar/(4\pi cI)$ — rotational constant; $I$ — moment of inertia; $J$ — rotational quantum number. The first line sits $6B$ from the laser and successive lines step out by $4B$ , so the spacing measures $B$ and hence the bond length. The quantum amplitude behind all of this is the Kramers–Heisenberg–Dirac transition polarizability. It is a sum over virtual states that makes both the two-photon character and the resonance enhancement explicit.

\alpha_{fi}=\sum_n\left[\frac{\langle f|\mu_\rho|n\rangle\langle n|\mu_\sigma|i\rangle}{\omega_{ni}-\omega_0-\mathrm{i}\gamma_n}+\frac{\langle f|\mu_\sigma|n\rangle\langle n|\mu_\rho|i\rangle}{\omega_{nf}+\omega_0}\right] \qquad (6)

where: $|i\rangle,|f\rangle$ — initial and final states; $|n\rangle$ — intermediate states summed over; $\mu_\rho,\mu_\sigma$ — Cartesian dipole components; $\omega_{ni}$ — transition frequency; $\gamma_n$ — linewidth of state $n$ . Raman is absorption-then-emission through every virtual level at once; when $\omega_0\to\omega_{ni}$ one denominator collapses — resonance Raman. Finally, the brightness of the lines. The scattered power carries the universal $\omega^4$ factor of dipole radiation, and the anti-Stokes/Stokes ratio folds in the population of the upper level.

\frac{\mathrm{d}\sigma}{\mathrm{d}\Omega}\propto\omega_s^{4}\,|\alpha_{fi}|^{2},\qquad \frac{I_{\mathrm{aS}}}{I_{\mathrm{S}}}=\left(\frac{\omega_0+\omega_{\mathrm{mol}}}{\omega_0-\omega_{\mathrm{mol}}}\right)^{4}\exp\!\left(-\frac{\hbar\omega_{\mathrm{mol}}}{k_BT}\right) \qquad (7)

where: $\sigma$ — differential cross-section, typically $\sim10^{-30}\ \mathrm{cm^2/sr}$ ; $k_B$ — Boltzmann constant; $T$ — absolute temperature. Bluer scattering radiates more strongly (the $\omega^4$ ), but anti-Stokes is throttled by the small excited-state population, so Stokes dominates at room temperature.

Derivation AVibrational sidebands from a modulated polarizability

The goal is to show, with nothing more than trigonometry, why a Raman spectrum has exactly three frequencies and why one of them can vanish. We take the induced dipole $p=\alpha E$ , insert a polarizability modulated by a harmonic vibration, and watch a single product of cosines split into a carrier plus two symmetric sidebands. The payoff is the classical Raman selection rule — the sidebands live or die on a single derivative, exactly complementary to the infrared rule.

Step 1. Keep the polarizability to first order in the normal coordinate, $\alpha(Q)=\alpha_0+(\partial\alpha/\partial Q)_0\,Q$ , and let the nuclei oscillate as $Q(t)=Q_0\cos\omega_v t$ . Why: the displacement is small, so a linear expansion captures the fundamental band, and the harmonic motion is the oscillator's ground-to-first-excited step.

Step 2. Form the dipole $p=\alpha(Q)\,E_0\cos\omega_0 t=\alpha_0 E_0\cos\omega_0 t+(\partial\alpha/\partial Q)_0 Q_0 E_0\cos\omega_v t\cos\omega_0 t$ . Why: the constant part rides the carrier untouched; the modulated part multiplies two oscillations together, and that product must be unpacked.

Step 3. Apply the product-to-sum identity $\cos\omega_v t\cos\omega_0 t=\tfrac12[\cos(\omega_0-\omega_v)t+\cos(\omega_0+\omega_v)t]$ . Why: a product of two cosines equals a sum of cosines at the sum and difference frequencies — the mathematical heart of every modulation sideband.

Step 4. Collect the result, $p(t)=\alpha_0 E_0\cos\omega_0 t+\tfrac12(\partial\alpha/\partial Q)_0 Q_0 E_0[\cos(\omega_0-\omega_v)t+\cos(\omega_0+\omega_v)t]$ : Rayleigh at $\omega_0$ , Stokes at $\omega_0-\omega_v$ , anti-Stokes at $\omega_0+\omega_v$ . Why: equal sideband amplitudes is the classical, temperature-blind prediction that the quantum populations later correct.

Step 5. Read off the activity criterion: both sidebands carry the factor $(\partial\alpha/\partial Q)_0$ , so a polarizability that does not change with the coordinate gives no Raman lines. Why: this is the rule $(\partial\alpha/\partial Q)_0\neq0$ , and it is why $\mathrm{N_2}$ — no dipole to wag but a bond whose polarizability breathes — is Raman-strong and infrared-silent.

Go deeper · Derivation B · 1 minRotational Raman line spacing of 4B

Derivation BRotational Raman line spacing of 4B

The goal here is to derive the fingerprint of pure rotational Raman — a comb spaced by $4B$ with a conspicuous $12B$ -wide gap centred on the laser — from the rigid-rotor ladder and the $\Delta J=\pm2$ rule the rank-2 polarizability enforces. That spacing is a direct ruler for $B$ , hence the moment of inertia, hence the bond length; measuring the comb of atmospheric $\mathrm{N_2}$ or $\mathrm{O_2}$ is how a rotational Raman lidar turns scattered light into a vertical temperature profile.

Step 1. Write the rigid-rotor term values in wavenumbers, $F(J)=BJ(J+1)$ . Why: this is the energy ladder the scattered photon adds to or subtracts from, and in $\mathrm{cm^{-1}}$ the answers come out directly as measurable Raman shifts.

Step 2. For the Stokes (S) branch the molecule climbs two rungs, $J\to J+2$ , so the shift is $\Delta F=F(J+2)-F(J)=B[(J+2)(J+3)-J(J+1)]=B(4J+6)$ . Why: expanding and cancelling leaves a clean linear function of $J$ , which makes the line pattern a simple arithmetic comb.

Step 3. For the anti-Stokes (O) branch, $J\to J-2$ , giving $F(J)-F(J-2)=B(4J-2)$ ; the lowest such line, from $J=2$ , is again $6B$ . Why: the S and O branches mirror about the laser, so the first line on each side sits the same $6B$ out.

Step 4. Take the spacing between consecutive S lines, $B[4(J+1)+6]-B(4J+6)=4B$ . Why: adjacent lines differ by a constant $4B$ regardless of $J$ , so one measured spacing delivers $B$ — the ruler that reads the bond length.

Step 5. Locate the central gap: the first Stokes and anti-Stokes lines lie $6B$ on either side of the laser, a clear region of width $12B$ . Why: with $B\approx2.0\ \mathrm{cm^{-1}}$ for $\mathrm{N_2}$ this gap is about $24\ \mathrm{cm^{-1}}$ — a signature that says pure rotational Raman at a glance.

Limiting cases

Case	Result	Meaning
High temperature ( $k_BT\gg\hbar\omega_{\mathrm{mol}}$ )	$I_{\mathrm{aS}}/I_{\mathrm{S}}\to(\omega_{\mathrm{aS}}/\omega_{\mathrm{S}})^4$	populations equalize, so anti-Stokes approaches Stokes up to the frequency factor
Low temperature ( $k_BT\ll\hbar\omega_{\mathrm{mol}}$ )	$I_{\mathrm{aS}}/I_{\mathrm{S}}\to0$	the excited level is empty; anti-Stokes vanishes, only Stokes survives
Resonance ( $\omega_0\to\omega_{ni}$ )	$\alpha_{fi}\sim\dfrac{\mu_{fn}\mu_{ni}}{\omega_{ni}-\omega_0-\mathrm{i}\gamma_n}$	the KHD denominator collapses, enhancing $\sigma$ by up to $\sim10^6$ (resonance Raman)
Non-resonant, transparent	$\dfrac{\mathrm{d}\sigma}{\mathrm{d}\Omega}\propto\omega_s^4\|\alpha_{fi}\|^2$	ordinary spontaneous Raman: weak, $\sim10^{-30}\ \mathrm{cm^2/sr}$ , needs a laser
Spherical top / isotropic mode	$\Delta\alpha=0$	vanishing anisotropy gives no rotational Raman, only a polarized $Q$ -type contribution

Common trapRaman activity is not the same as infrared activity, and rotational Raman is ΔJ = ±2, not ±1

The commonest error is to import the infrared rule wholesale. Infrared needs a changing dipole, $(\partial\mu/\partial Q)_0\neq0$ ; Raman needs a changing polarizability, $(\partial\alpha/\partial Q)_0\neq0$ — different modes. In a centrosymmetric molecule the mutual exclusion rule is exact: no mode is both infrared- and Raman-active, because the dipole (odd under inversion) and polarizability (even) select opposite parities. So $\mathrm{N_2}$ is Raman-strong and infrared-silent, while the antisymmetric stretch of $\mathrm{CO_2}$ is the reverse. The second trap is the rotational step: because the polarizability is rank-2, pure rotational Raman obeys $\Delta J=\pm2$ , so lines are spaced by $4B$ , not the $2B$ of a rotational absorption spectrum — quote $2B$ and every bond length is off by a factor of two.

Worked examples

1 · Pure rotational Raman spectrum of N₂ (a real measurement)

An argon-ion laser at $514.5\ \mathrm{nm}$ excites nitrogen gas. Using the $\mathrm{N_2}$ bond length $r=1.098\ \text{Å}$ and $^{14}\mathrm{N}$ mass $14.003\ \mathrm{u}$ , find the rotational constant, the first Stokes rotational shift, and its absolute wavelength.

Step 1. Reduced mass and moment of inertia. For the homonuclear molecule $\mu=m_N/2=1.163\times10^{-26}\ \mathrm{kg}$ , so $I=\mu r^2=1.163\times10^{-26}\times(1.098\times10^{-10})^2=1.401\times10^{-46}\ \mathrm{kg\,m^2}$ .

Step 2. Rotational constant. $B=\dfrac{h}{8\pi^2 cI}=\dfrac{6.626\times10^{-34}}{8\pi^2\,(2.998\times10^{10})\,(1.401\times10^{-46})}=\mathbf{2.00\ cm^{-1}}$ , matching the tabulated $B_0=1.99\ \mathrm{cm^{-1}}$ .

Step 3. First Stokes line. The lowest S-branch line, $J=0\to2$ , sits at shift $6B=\mathbf{12.0\ cm^{-1}}$ , and consecutive lines are spaced by $4B=\mathbf{8.0\ cm^{-1}}$ .

Step 4. Absolute wavelength. The laser wavenumber is $\tilde\nu_0=10^7/514.5=19436\ \mathrm{cm^{-1}}$ ; the Stokes line sits at $19436-12.0=19424\ \mathrm{cm^{-1}}$ , so $\lambda=10^7/19424=\mathbf{514.82\ nm}$ . The whole comb is squeezed into a fraction of a nanometre around the laser — hence the need for a high-resolution spectrometer and strong laser-line rejection.

2 · Stokes / anti-Stokes thermometry of a hot gas

The $\mathrm{N_2}$ vibrational Raman line lies at $\tilde\nu=2331\ \mathrm{cm^{-1}}$ . A frequency-doubled Nd:YAG laser at $532\ \mathrm{nm}$ ( $\tilde\nu_0=18797\ \mathrm{cm^{-1}}$ ) probes a flame. Predict the anti-Stokes/Stokes ratio at $2000\ \mathrm{K}$ , and compare with room temperature.

Step 1. Boltzmann factor. With $c_2=hc/k_B=1.4388\ \mathrm{cm\,K}$ , the exponent is $c_2\tilde\nu/T=1.4388\times2331/2000=1.677$ , so $e^{-1.677}=0.187$ .

Step 2. Frequency factor. $\left(\dfrac{\tilde\nu_0+\tilde\nu}{\tilde\nu_0-\tilde\nu}\right)^4=\left(\dfrac{21128}{16466}\right)^4=(1.283)^4=2.71$ .

Step 3. Ratio at 2000 K. $I_{\mathrm{aS}}/I_{\mathrm{S}}=2.71\times0.187=\mathbf{0.51}$ — anti-Stokes is half as bright as Stokes, easily measured, so inverting the formula returns the flame temperature.

Step 4. Room temperature for contrast. At $300\ \mathrm{K}$ the exponent balloons to $1.4388\times2331/300=11.18$ , giving $e^{-11.18}=1.4\times10^{-5}$ and a ratio of $2.71\times1.4\times10^{-5}=\mathbf{3.8\times10^{-5}}$ — essentially invisible. This is why vibrational anti-Stokes thermometry needs a hot source, while near room temperature one uses the low-lying rotational lines, whose small $\hbar\omega_{\mathrm{mol}}$ keeps both branches comparably bright.

Physics of the portrait

Raman scattering bears the name of the person who first saw it — someone who spent years chasing a faint glow with sunlight and coloured filters before a mercury lamp revealed the shifted line.

Origins · 1 minC. V. Raman and how this was discovered

C. V. Raman

1888–1970 · molecular light scattering and the inelastic Raman effect

Chandrasekhara Venkata Raman was drawn to light scattering by a question that nagged him on a 1921 sea voyage: why is the deep sea blue, and not merely a mirror of the sky? Rayleigh had blamed reflected skylight; Raman showed instead that water molecules scatter light on their own account, a pursuit that led him from the colour of the ocean to that of scattered light itself. Through the mid-1920s he and his students at Calcutta pushed sunlight through liquid after liquid, hunting a "feeble fluorescence" that survived crossed filters — scattered light whose colour differed from the incident beam. On 28 February 1928, using a mercury arc and a spectrograph, Raman and K. S. Krishnan resolved the shifted lines and reported "A new type of secondary radiation" in Nature; the same weeks saw Landsberg and Mandelstam observe the effect in quartz in Moscow, while the prediction had come earlier from Adolf Smekal (1923) and the Kramers–Heisenberg dispersion theory (1925). Raman received the 1930 Nobel Prize, and once lasers arrived his effect became one of the most versatile probes in chemistry and physics. Its deeper legacy is a shift of viewpoint: scattered light carries a fingerprint of the scatterer's internal motions, written not in the dipole but in the polarizability.

History

1923

— Smekal predicts frequency-shifted (inelastic) light scattering theoretically

1928

— Raman and Krishnan observe the effect in liquids; Landsberg and Mandelstam see it in quartz

1930

— Raman is awarded the Nobel Prize in Physics

1934

— Placzek formulates the polarizability theory of the Raman effect

1962

— Woodbury and Ng discover stimulated Raman scattering in a laser cavity

Going further

Texts. Loudon, The Quantum Theory of Light, 3rd ed., Ch. 8 (light scattering, Kramers–Heisenberg) · Demtröder, Laser Spectroscopy, Vol. 1 (Raman spectroscopy) · Hecht, Optics, 5th ed., §13.4 · Saleh and Teich, Fundamentals of Photonics, 3rd ed. · Long, The Raman Effect (Wiley, 2002), the standard monograph.

Provenance. The classical modulation derivation follows Hecht §13.4; the Kramers–Heisenberg–Dirac polarizability and Stokes/anti-Stokes ratio follow Loudon Ch. 8; the rotational $4B$ / $6B$ structure and $\Delta J=\pm2$ rule follow Demtröder and Long. A verifier should recompute: (1) the product-to-sum sidebands at $\omega_0\pm\omega_v$ ; (2) $F(J+2)-F(J)=B(4J+6)$ and spacing $4B$ ; (3) the intensity ratio with $\omega^4$ and Boltzmann factors; (4) the $\mathrm{N_2}$ numbers — $B=2.00\ \mathrm{cm^{-1}}$ from $I=\mu r^2$ , first Stokes rotational line at $514.82\ \mathrm{nm}$ under $514.5\ \mathrm{nm}$ excitation, and $I_{\mathrm{aS}}/I_{\mathrm{S}}=0.51$ at $2000\ \mathrm{K}$ for the $2331\ \mathrm{cm^{-1}}$ vibration. Anchor values ( $\mathrm{N_2}$ $2331\ \mathrm{cm^{-1}}$ , $\mathrm{O_2}$ $1556\ \mathrm{cm^{-1}}$ , silica gain peak near $13\ \mathrm{THz}$ , $\sigma\sim10^{-30}\ \mathrm{cm^2/sr}$ ) are textbook standard; note the earlier migrated draft mislabelled the $\mathrm{N_2}$ shift as $1580\ \mathrm{cm^{-1}}$ , which is the $\mathrm{O_2}$ value, corrected here.

Prerequisites. The Rigid Rotor Model · The Harmonic Oscillator Model of Vibration · Selection Rules for Molecular Transitions · Rotational Spectra and the Rotational Constant. Leads to. Vibration–Rotation Spectra and P, Q, R Branches · Anharmonicity and the Morse Potential · Electronic Transitions in Molecules.

Ask the tutor: Want to see how the same polarizability-tensor rule that gives $\Delta J=\pm2$ also carves the O, Q, and S branches out of a full vibration–rotation Raman band?

What is the difference between Rayleigh, Stokes, and anti-Stokes scattering?

All three are light scattered by the same molecule, distinguished only by the energy the photon exchanges. Rayleigh is elastic: the photon leaves at $\omega_0$ , having polarized the molecule and let it relax. Stokes leaves the molecule one quantum higher, so the photon departs redshifted to $\omega_0-\omega_{\mathrm{mol}}$ . Anti-Stokes starts from an already-excited molecule and carries away its quantum, so the photon is blueshifted to $\omega_0+\omega_{\mathrm{mol}}$ . Rayleigh is by far the brightest; the Raman sidebands are typically a thousand times fainter, and anti-Stokes is fainter still because it needs a pre-excited molecule.

Why does the shift not depend on the colour of the exciting laser?

Because the shift is set by the molecule, not the light. The scattered frequency is $\omega_s=\omega_0\pm\omega_{\mathrm{mol}}$ , so the difference from the laser, $\omega_s-\omega_0=\pm\omega_{\mathrm{mol}}$ , is always exactly one internal quantum regardless of $\omega_0$ . This is why Raman spectra are plotted against the shift in $\mathrm{cm^{-1}}$ : a bond vibrating at $2331\ \mathrm{cm^{-1}}$ shows a line $2331\ \mathrm{cm^{-1}}$ from the laser whether you excite at 532 or 785 nm. It also frees the experimenter to choose the laser for other reasons — dodging fluorescence, hitting a resonance, matching a detector — without changing the spectroscopy.

How can Raman probe vibrations using visible light?

The visible photon is not absorbed by the vibration — its energy is far too large. Instead it is scattered, and the vibrational information rides in the tiny frequency shift it picks up. A $532\ \mathrm{nm}$ photon carries about $18800\ \mathrm{cm^{-1}}$ ; a vibration is only a percent or so of that, say $2331\ \mathrm{cm^{-1}}$ for $\mathrm{N_2}$ . Raman reads the vibration as the small offset between incident and scattered photon, effectively doing infrared spectroscopy with visible optics, silicon detectors, and glass lenses. The payoff is large: water is nearly transparent in the visible but opaque across much of the infrared, which is why Raman excels at aqueous and biological samples.

What is the mutual exclusion rule, and when does it apply?

In a molecule with a centre of symmetry ( $\mathrm{CO_2}$ , $\mathrm{N_2}$ , benzene), no vibrational mode can be simultaneously infrared- and Raman-active. The reason is parity: every mode is either gerade or ungerade under inversion. The dipole is ungerade, so infrared activity needs an ungerade mode; the polarizability is gerade, so Raman activity needs a gerade mode. A mode cannot be both, so the two spectra are complementary and non-overlapping. For molecules without a centre of symmetry the rule fails and a mode can appear in both — which makes the rule itself a sensitive test of whether a molecule is centrosymmetric.

Why is spontaneous Raman so weak, and how do CARS and SERS help?

Raman is a second-order, two-photon process: the light must reach a virtual state and re-radiate, an intrinsically small amplitude, so only about one in $10^{6}$ to $10^{7}$ incident photons scatters inelastically and the cross-section is a mere $\sim10^{-30}\ \mathrm{cm^2/sr}$ . Three strategies fight this. Resonance Raman tunes the laser onto a real electronic transition, collapsing a Kramers–Heisenberg denominator and lifting the signal by up to $10^{6}$ . CARS and stimulated Raman drive the vibration coherently with two lasers whose beat matches $\omega_v$ . SERS exploits plasmonic fields at nanostructured metal surfaces to boost signals by $10^{8}$ or more, reaching single molecules.

Can the Stokes/anti-Stokes ratio really measure temperature with no calibration?

Yes, and that is its appeal. Because both lines come from the same mode with essentially the same polarizability matrix element, every instrumental constant — laser power, collection efficiency, cross-section — cancels in their ratio. What survives is a known frequency factor and the Boltzmann exponential $e^{-\hbar\omega_{\mathrm{mol}}/k_BT}$ , so the ratio inverts directly to absolute temperature: a non-contact, self-calibrating thermometer for flames and plasmas. Two cautions: a careful treatment adds the rotational degeneracy the two-level formula omits, and the mode must be low enough in frequency that anti-Stokes is measurable — vibrational lines for hot sources, rotational lines near room temperature.