Chemistry · Lesson
Selection rules, transition intensities and symmetry
Ask which interaction connects two states before naming a transition allowed or forbidden. Use symmetry and transition integrals within their stated approximations, and keep strength separate from a measured peak height.
Back to the Chemistry mapBy the end, you can
- Distinguish an energy match from a nonzero transition matrix element.
- Compare box, harmonic, IR and Raman selection conditions.
- Apply model-limited symmetry and spin statements without treating weakly enabled transitions as impossible.
A transition requires the relevant coupling
In a weak-field one-photon electric-dipole model, the transition matrix element is M_fi = ⟨f|μ̂|i⟩, where μ̂ is the electric-dipole operator. An energy-matched transition can still have zero M_fi. Within that interaction model, the strength contains |M_fi|². Population, frequency factors, orientation, line profile and detection also affect the measured signal.
The electric-dipole approximation assumes the spatial light-field variation is small across the molecule. Magnetic dipole, higher multipole and other interactions have different operators and rules. A transition forbidden in one approximation can gain strength through other couplings, symmetry breaking or state mixing.
Do not copy a rule from one Hamiltonian to another
For the one-dimensional infinite-well position operator, off-diagonal ⟨m|x|n⟩ is nonzero when m+n is odd, equivalently when Δn is odd. Thus 1→4 can be dipole allowed even though it changes n by 3; 1→3 has a zero integral in this model. Signs of off-diagonal matrix elements depend on chosen state phases, while their squared magnitudes do not.
For the exact harmonic oscillator with a dipole expansion linear in its displacement, the relevant coordinate matrix elements instead give Δv = ±1. Anharmonic state mixing or nonlinear dipole dependence can enable other bands. These two rules belong to different model wavefunctions and coupling operators.
| Box transition | Δn | ⟨final|x|initial⟩ / L | Relative squared matrix element |
|---|---|---|---|
| 1 → 2 | 1 | −0.18012655 | 1 |
| 1 → 3 | 2 | 0 | 0 |
| 1 → 4 | 3 | −0.01441012 | 0.00640 |
Original exact 1D infinite-well position integrals using normalized sine states. Matrix elements are reported divided by L. The final column compares only squared matrix-element magnitudes to the 1→2 value; it is not a complete absorption-intensity ratio. The sign is phase-convention dependent.
A dipole derivative and a polarizability derivative test different activity
Ordinary vibrational IR activity requires a nonzero change of molecular dipole along the normal coordinate. First-order vibrational Raman activity requires the appropriate polarizability derivative. A mode can therefore appear through one interaction and be absent through another. For centrosymmetric molecules under these standard approximations, the mutual-exclusion symmetry result separates IR-active and Raman-active modes.
In Stokes Raman scattering, the scattered photon has lower frequency and the molecule gains excitation. Anti-Stokes scattering produces a higher-frequency photon from an initially excited molecule. Populations, scattering frequency factors and instrument response matter for observed intensity ratios; a Boltzmann population ratio alone is not the complete raw signal ratio.
An approximation-limited zero is useful information
Symmetry can force a transition integral to vanish when the combined state and operator symmetries cannot produce the required invariant. The answer depends on the actual molecular symmetry, polarization and interaction, not only a label attached to a state. Degenerate modes can share a frequency, and overlapping bands can hide distinct transitions.
A spin-independent electric-dipole approximation preserves spin multiplicity. Spin–orbit coupling, including spin-vibronic mechanisms involving it, can enable nominally spin-forbidden transitions. Vibronic mixing can also relax spatial-symmetry rules; spin-free vibronic mixing alone does not remove the spin-multiplicity rule. Franck–Condon vibrational overlaps add another factor to an electronic transition, within the Born–Oppenheimer and Condon assumptions rather than as a universal exact intensity law.
Worked examples
Compare squared matrix elements under the same model
The supplied box integrals are x₁₂/L≈−0.18012655 and x₁₄/L≈−0.01441012. Find their squared-magnitude ratio.
- Divide the two unrounded analytic magnitudes; their ratio is 0.0800.
- Square that ratio: 0.0800² = 0.00640.
- A phase change can reverse a matrix-element sign without changing its square.
- A complete observed intensity comparison would also require frequency, population and measurement information.
The 1→4 squared matrix element is 0.00640 times the 1→2 value. It is nonzero despite Δn=3.
Convert a Raman shift with its correct sign
Use a 532 nm vacuum laser wavelength and a stipulated positive vibrational shift of 1.00 × 10³ cm⁻¹. Find the Stokes and anti-Stokes vacuum wavelengths.
- The incident wavenumber is 10⁷/532 ≈ 18796.99248 cm⁻¹.
- Stokes: subtract1000 cm⁻¹, then take the reciprocal length.
- Anti-Stokes: add1000 cm⁻¹, then take the reciprocal length.
- The corresponding wavelengths are approximately 561.893 and 505.127 nm.
Stokes λ≈ 562 nm; anti-Stokes λ≈ 505 nm. Subtracting a wavenumber does not mean subtracting the same numerical value from wavelength.
Try it yourself
In the exact box position-operator model, is 1→4 forbidden merely because Δn is 3?
Show the answer and reasoning
No.
Its odd Δn gives a nonzero off-diagonal position integral.
What Δv does the harmonic linear-displacement electric-dipole model permit?
Show the answer and reasoning
±1.
The rule is model-specific and differs from the box rule.
Which derivative controls ordinary first-order vibrational Raman activity?
Show the answer and reasoning
A polarizability derivative along the normal coordinate.
The corresponding IR condition instead uses the dipole derivative.
Does ordinary IR/Raman mutual exclusion apply to every molecule?
Show the answer and reasoning
No.
It is a centrosymmetric-molecule result under the stated interaction approximations.
What to carry forward
- Choose the coupling operator before using a selection rule.
- An energy match alone does not supply a transition strength.
- IR and Raman test different molecular responses.
- Read forbidden labels within their symmetry and interaction assumptions.
References
Teaching resources and reference material supporting this lesson’s explanations and examples.
Massachusetts Institute of Technology3 sources
- 5.61 Lecture 19: Spectroscopy: Probing Molecules with LightTroy Van Voorhis (original Fall 2013 notes, explicit PDF footnote); Robert W. Field (2017 course instructor)Source · PDF
- 5.61 Lecture 8: Quantum Mechanical Harmonic OscillatorRobert W. Field (course instructor; note-specific contributor identified where visible)Source · PDF
- 5.61 Lecture 34: Electronic Spectroscopy and PhotochemistryRobert W. Field (course instructor; note-specific contributor identified where visible)Source · PDF
- Advanced lab course physical chemistry: Raman scatteringInstitutional teaching material; individual author not identified on retrieved pageSource
- CHMY361 Lecture 39: Raman selection rulesPatrik R. Callis (course-owner attribution; PDF has no explicit author byline)Source · PDF
- Quantum Mechanics, §§2.1.2 and 2.2David TongSource
- 2022 CODATA recommended constants: complete ASCII tableNIST Physical Measurement Laboratory; CODATA adjustmentSource
- Modern Analytical Chemistry, §§10C.2–10C.5 and vibrational-mode paragraphDavid HarveySource · PDF