Case Study 2: Selection Rules — Why Some Transitions Are Forbidden

Overview

When an atom absorbs or emits a photon, it transitions from one quantum state to another. But not every transition is possible. Certain changes in quantum numbers are forbidden — not by any explicit prohibition, but because the relevant matrix elements vanish identically. These selection rules are among the most powerful consequences of the Wigner-Eckart theorem and the angular momentum coupling formalism developed in this chapter.

This case study explores selection rules from three angles: their derivation from symmetry, their experimental manifestations in atomic spectroscopy, and their occasional violation — because in quantum mechanics, "forbidden" does not always mean "impossible."

Part 1: The Origin of Selection Rules

Transition Rates and Matrix Elements

From Fermi's golden rule (Chapter 21), the rate of a radiative transition between states $|i\rangle$ and $|f\rangle$ is proportional to the squared matrix element:

$$\Gamma_{i \to f} \propto |\langle f | \hat{H}_{\text{int}} | i \rangle|^2$$

where $\hat{H}_{\text{int}}$ is the interaction Hamiltonian between the atom and the electromagnetic field. If this matrix element is exactly zero, the transition does not occur (at least at that order of approximation). The conditions under which it vanishes are the selection rules.

The Multipole Expansion

The atom-field interaction can be expanded in powers of $(a/\lambda)$, where $a \sim a_0$ is the atomic size and $\lambda$ is the photon wavelength. For optical transitions, $a/\lambda \sim 10^{-3}$, so successive terms in the expansion are suppressed by factors of $\sim 10^{-3}$:

$$\hat{H}_{\text{int}} = \hat{H}_{\text{E1}} + \hat{H}_{\text{M1}} + \hat{H}_{\text{E2}} + \hat{H}_{\text{M2}} + \hat{H}_{\text{E3}} + \cdots$$

Each term is an irreducible tensor operator of definite rank, and the Wigner-Eckart theorem immediately gives the selection rules for each:

Part 2: Electric Dipole Selection Rules in Detail

Derivation from the Wigner-Eckart Theorem

The electric dipole Hamiltonian involves the operator $\hat{\mathbf{d}} = e\hat{\mathbf{r}}$, which is a rank-1 irreducible tensor operator. By the Wigner-Eckart theorem:

$$\langle n', \ell', j', m_j' | \hat{d}_q | n, \ell, j, m_j\rangle = \frac{\langle n', \ell', j' \| d \| n, \ell, j \rangle}{\sqrt{2j'+1}} \langle j, m_j; 1, q | j', m_j'\rangle$$

The CG coefficient immediately gives:

Rule 1: $\Delta m_j = 0, \pm 1$. This corresponds to the three possible values of $q = 0, \pm 1$ in the spherical components of the dipole operator.

• $q = 0$ ($\Delta m_j = 0$): $\pi$ polarization (linear, parallel to quantization axis)
• $q = +1$ ($\Delta m_j = +1$): $\sigma^+$ polarization (left circular)
• $q = -1$ ($\Delta m_j = -1$): $\sigma^-$ polarization (right circular)

Rule 2: $\Delta j = 0, \pm 1$ (but $j = 0 \to j' = 0$ is forbidden). This follows from the triangle rule for the CG coefficient with $k = 1$.

The $j = 0 \to j' = 0$ case is special: the only CG coefficient $\langle 0, 0; 1, q | 0, 0\rangle$ involves $M' = q$, but $j' = 0$ requires $M' = 0$, so only $q = 0$ survives. And $\langle 0, 0; 1, 0 | 0, 0\rangle = 0$ because the triangle rule gives $|0 - 1| = 1 > 0 = j'$. So the transition is forbidden for all polarizations.

Parity Selection Rule

The position operator $\hat{\mathbf{r}}$ has odd parity: $\hat{P}\hat{\mathbf{r}}\hat{P}^{-1} = -\hat{\mathbf{r}}$. Therefore:

$$\langle f | \hat{\mathbf{d}} | i \rangle = \langle f | e\hat{\mathbf{r}} | i \rangle = 0 \quad \text{if } |i\rangle \text{ and } |f\rangle \text{ have the same parity}$$

Since the parity of hydrogen states is $(-1)^\ell$, this gives:

Rule 3: $\Delta\ell = \pm 1$ (parity must change).

Combined, the E1 selection rules are: $$\boxed{\Delta\ell = \pm 1, \quad \Delta j = 0, \pm 1 \;\;(j = 0 \leftrightarrow j' = 0 \text{ forbidden}), \quad \Delta m_j = 0, \pm 1}$$

What About $\Delta n$?

There is no selection rule on $n$. The principal quantum number does not carry angular momentum information, so it is not constrained by the Wigner-Eckart theorem. The transition probability does depend on $n$ through the radial integral (the reduced matrix element), but no value of $\Delta n$ is forbidden by symmetry.

Part 3: Spectroscopic Consequences

The Hydrogen Spectrum

The Balmer series of hydrogen (transitions to $n = 2$) illustrates the selection rules beautifully:

• $3s \to 2p$: $\Delta\ell = +1$ ✓. Allowed E1 transition.
• $3p \to 2s$: $\Delta\ell = -1$ ✓. Allowed E1 transition.
• $3d \to 2p$: $\Delta\ell = -1$ ✓. Allowed E1 transition.
• $3d \to 2s$: $\Delta\ell = -2$ ✗. Forbidden as E1. (Allowed as E2 — much weaker.)
• $3s \to 2s$: $\Delta\ell = 0$ ✗. Forbidden as E1. (No parity change.)
• $3p \to 2p$: $\Delta\ell = 0$ ✗. Forbidden as E1. (No parity change.)

The hydrogen spectrum shows only those lines corresponding to allowed E1 transitions. The "missing" lines — transitions that violate $\Delta\ell = \pm 1$ — can sometimes be observed as very faint features arising from higher-order multipole processes.

The Grotrian Diagram

A Grotrian diagram is an energy level diagram with allowed E1 transitions shown as lines connecting levels. The selection rule $\Delta\ell = \pm 1$ means transitions only connect adjacent columns (where columns represent $\ell = 0, 1, 2, 3, \ldots$). No transitions connect $s$-states to $s$-states, or $d$-states to $s$-states.

The Zeeman Effect

In a magnetic field, each $j$ level splits into $2j + 1$ sublevels labeled by $m_j$. The selection rule $\Delta m_j = 0, \pm 1$ determines which transitions among these sublevels are allowed:

• $\Delta m_j = 0$ ($\pi$ transitions): Linearly polarized parallel to $\mathbf{B}$. These are observed perpendicular to the field direction.
• $\Delta m_j = +1$ ($\sigma^+$) and $\Delta m_j = -1$ ($\sigma^-$): Circularly polarized. These are observed along the field direction.

The anomalous Zeeman effect (the general case with spin-orbit coupling) produces a complicated pattern of lines because the Lande g-factor differs between upper and lower levels. The normal Zeeman effect (the special case where spin plays no role, or where both levels have the same $g_j$) produces the simple Lorentz triplet.

Part 4: "Forbidden" Transitions in Practice

Metastable States

A state that cannot decay by E1 radiation is called metastable. Its lifetime is dramatically longer than normal excited states:

The hydrogen $2s_{1/2}$ state: The only E1-allowed transition from $2s$ is to $1s$. But $2s \to 1s$ requires $\Delta\ell = 0$, which violates the parity selection rule. So the $2s$ state cannot decay by E1 radiation.

It can decay by: - Two-photon emission (a second-order process): $\tau \approx 0.14\,\text{s}$ - M1 transition (magnetic dipole): extremely slow, negligible

For comparison, the $2p$ state decays to $1s$ by E1 with $\tau \approx 1.6 \times 10^{-9}\,\text{s}$ — a factor of $10^8$ faster!

📊 By the Numbers: Metastable state lifetimes compared to normal E1 decay:

Transition	Type	Lifetime
$\text{H}(2p \to 1s)$	E1	$1.6 \times 10^{-9}\,\text{s}$
$\text{H}(2s \to 1s)$	Two-photon	$0.14\,\text{s}$
$\text{He}(2\,^3S_1 \to 1\,^1S_0)$	Spin-forbidden	$7870\,\text{s}$
$\text{O\,III}(^1D_2 \to \,^3P_2)$	M1	$38\,\text{s}$
$\text{O\,III}(^1S_0 \to \,^1D_2)$	E2	$0.7\,\text{s}$

Nebular Lines: Forbidden Lines in Astrophysics

Some of the most important spectral lines in astrophysics are "forbidden" — they violate E1 selection rules and proceed through M1 or E2 processes. In the low-density environment of interstellar nebulae (where densities are $\sim 10^3\,\text{cm}^{-3}$, compared to $\sim 10^{19}\,\text{cm}^{-3}$ in a gas at atmospheric pressure), collisional de-excitation is rare, and atoms in metastable states have time to radiate through these slow processes.

The green line of the Orion Nebula at $\lambda = 5007\,\text{\AA}$ — one of the most iconic emission lines in astronomy — is a forbidden transition of doubly ionized oxygen (O III):

$$[\text{O\,III}]: ^1D_2 \to \,^3P_2 \quad (\lambda = 5007\,\text{\AA}, \text{ M1 transition})$$

The notation $[\text{O\,III}]$ with square brackets is the standard astronomical convention for a forbidden line.

🔵 Historical Note: When the green nebular lines were first observed in the 19th century, they could not be identified with any known element. The hypothetical element "nebulium" was proposed to explain them. In 1927, Ira Bowen showed that the lines were forbidden transitions of known elements (oxygen and nitrogen) that are unobservable in the laboratory because collisional de-excitation quenches them at terrestrial densities. This was a triumph of quantum mechanical selection rule theory applied to astrophysics.

Intercombination Lines

Transitions between states of different multiplicity (e.g., singlet $\to$ triplet in a two-electron atom) are forbidden in strict L-S coupling because the E1 operator does not act on spin. In reality, spin-orbit coupling mixes states of different $S$, and "spin-forbidden" transitions become weakly allowed.

Example: The $2\,^3P_1 \to 1\,^1S_0$ transition in helium (a triplet-to-singlet transition) is forbidden in strict L-S coupling but actually occurs with a lifetime of $\sim 10^{-7}\,\text{s}$ (much longer than the allowed singlet-singlet $2\,^1P_1 \to 1\,^1S_0$ at $\sim 10^{-9}\,\text{s}$, but much shorter than truly forbidden transitions).

For heavy atoms, spin-orbit mixing is strong, and intercombination lines can be quite prominent. The mercury $6\,^3P_1 \to 6\,^1S_0$ line at $253.7\,\text{nm}$ (used in fluorescent lamps and mercury vapor lamps) is an intercombination line that is intense enough for practical applications.

Part 5: Selection Rules Beyond Atoms

Nuclear Gamma Transitions

Selection rules apply to nuclear electromagnetic transitions just as they do to atomic ones. The key difference is that nuclear levels can have high spin (up to $j \sim 30$ or more in deformed nuclei), and higher-order multipoles are more important because the nuclear size is a larger fraction of the photon wavelength.

The selection rules for nuclear gamma decay follow exactly the same Wigner-Eckart theorem framework: - E$\lambda$ transitions: $\Delta j = \lambda$, parity change $(-1)^\lambda$ - M$\lambda$ transitions: $\Delta j = \lambda$, parity change $(-1)^{\lambda+1}$

Nuclear isomers are metastable nuclear states with very long lifetimes — sometimes hours, years, or even billions of years. They occur when the only available decay path requires a very high multipole order (large $\lambda$), which makes the transition rate extremely slow.

Molecular Transitions

In molecules, additional selection rules arise from the vibrational and rotational degrees of freedom:

• Rotational transitions: $\Delta J = \pm 1$ (for linear molecules with a permanent dipole moment)
• Vibrational transitions: $\Delta v = \pm 1$ (harmonic approximation; overtones $\Delta v = \pm 2, \pm 3$ are weakly allowed by anharmonicity)
• Electronic-rotational: Additional rules involving molecular symmetry point groups

The angular momentum coupling formalism of this chapter extends naturally to molecular systems, with the complication that the nuclear framework breaks spherical symmetry, and Hund's coupling cases (a, b, c, d) replace L-S and j-j coupling.

Discussion Questions

1. The $2s_{1/2}$ state of hydrogen is metastable, with a lifetime of 0.14 seconds (seven orders of magnitude longer than $2p$). Yet in a hydrogen discharge tube, we observe bright Lyman-$\alpha$ emission from $2p \to 1s$. Why don't atoms "pile up" in the metastable $2s$ state and quench the discharge?

2. The green nebular line of [O III] at 5007 Angstroms was attributed to "nebulium" for over 60 years before Bowen identified it in 1927. What prevented laboratory identification earlier? Could a modern experiment detect this line in the lab?

3. If E1 selection rules were somehow "turned off" — if all E1 matrix elements were zero — how would the world be different? Consider: atomic lifetimes, the appearance of the night sky, the efficiency of lasers, and the transparency of materials.

4. Selection rules in nuclear physics allow isomeric states with lifetimes exceeding billions of years (e.g., $^{180m}\text{Ta}$ with $t_{1/2} > 10^{15}\,\text{yr}$). These nuclei are "stuck" in excited states. What does this tell us about the relationship between energy and stability in quantum systems?

5. In what sense are "forbidden" transitions not really forbidden? If the selection rule says a matrix element is zero, how can the transition happen at all? Discuss the role of higher-order processes, configuration mixing, and the limitations of the approximations underlying the selection rules.