44
Part One · Quantum mechanics

Time-independent perturbation theory

Almost nothing in quantum mechanics can be solved exactly. The hydrogen atom can; helium cannot. The harmonic oscillator can; a real chemical bond cannot. Yet the magnetic strength of the electron has been calculated and measured to agree to about one part in a trillion. The tool that made that possible is perturbation theory: start from a problem you can solve, and add the rest a little at a time.

On the plate: an oscillator whose walls have been made slightly steeper by an extra λx4\lambda x^4.

First order

Split the energy operator into a solved part and a small extra, H=H0+λVH = H_0 + \lambda V. Expand the energies and states in powers of λ\lambda and match powers. The first correction to an energy is just the average of the extra term in the unperturbed state:

En(1)=⟨n∣V∣n⟩E_n^{(1)} = \langle n|V|n\rangle

For the oscillator with V=x4V = x^4, the ground state gives ⟨x4⟩=34\langle x^4\rangle = \tfrac34 in natural units, so E0≈12+34λE_0 \approx \tfrac12 + \tfrac34\lambda. Set λ=\lambda = 0.10. The plate shows the new potential and compares the exact ground energy, found numerically, with this estimate. Established

Second order

The state itself also changes, by mixing in other states, and that feeds back into the energy at the next order:

En(2)=∑m≠n∣⟨m∣V∣n⟩∣2En−EmE_n^{(2)} = \sum_{m\ne n}\frac{|\langle m|V|n\rangle|^2}{E_n - E_m}

Two rules follow at once. The ground state’s second-order shift is always downward, since every denominator is negative. And states close in energy push each other apart: level repulsion. For x4x^4 the sum gives −218λ2-\tfrac{21}{8}\lambda^2. On the plate, the exact energy is white, first order dashed, second order amber. For small λ\lambda second order hugs the exact curve; as λ\lambda grows it peels away.

When levels share an energy

If two unperturbed states have the same energy, the second-order formula divides by zero. The cure is to first find the right combinations: diagonalize VV within the degenerate group. Hydrogen’s n=2n = 2 level holds 2s and three 2p states. An electric field E\mathcal E along zz connects only 2s and 2p02p_0, and their combinations shift by ±3ea0E\pm3ea_0\mathcal E: a shift linear in the field, unusual because ordinary atoms have no permanent dipole. Raise the field to 5.0 MV/m; the other two 2p states stay put.

Measuring electric fields inside proteins

The ground state has no degeneracy, so its first-order Stark shift is zero, and the leading effect is second order: ΔE=−12αE2\Delta E = -\tfrac12\alpha\mathcal E^2, with the polarizability α\alpha built from the sum over states.

Molecular vibrations shift too, and that has become a practical tool. A carbonyl or nitrile group’s stretching frequency moves linearly with the electric field along its bond: the vibrational Stark effect. Steven Boxer’s group calibrated these probes and used them to measure the electric fields that enzymes apply to their substrates, which reach the order of 100 megavolts per centimetre, and found that stronger fields go with faster catalysis. Established

Fine structure, derived

Hydrogen’s fine structure from Chapter 38 is three perturbations added together: the relativistic correction to kinetic energy, spin–orbit coupling, and the Darwin term for s states. Each is of order α2\alpha^2 times the binding energy, and together they give

ΔEnj=−13.6 eVn2⋅α2n2(nj+12−34)\Delta E_{nj} = -\frac{13.6\ \text{eV}}{n^2}\cdot\frac{\alpha^2}{n^2}\left(\frac{n}{j+\frac12} - \frac34\right)

The energy depends only on nn and jj, so 2s1/2 and 2p1/2 stay together, exactly as Dirac’s equation says, until the Lamb shift separates them. The plate shows the n=2n = 2 and n=3n = 3 levels magnified about a hundred thousand times.

A series that never converges

Here is the surprise. Carl Bender and Tai Tsun Wu computed the x4x^4 series to high order in 1969: its coefficients are 34,−218,33316,−241.3,3581,…\tfrac34, -\tfrac{21}{8}, \tfrac{333}{16}, -241.3, 3581, \dots, growing like a factorial. The series diverges for every λ≠0\lambda \ne 0. Freeman Dyson saw the reason in 1952 for QED: if the series converged for small positive λ\lambda, it would also converge for small negative λ\lambda, where the potential turns over and no stable ground state exists.

The plate adds the terms one by one at your λ\lambda. The partial sums close in on the exact value, then swing away. The series is asymptotic: stop at the smallest term and you get an excellent answer; carry on and you get nonsense. QED’s series is believed to be the same, but its terms keep shrinking for about 137 orders, far more than anyone can compute. Established

What you now have

First and second-order perturbation theory, level repulsion, degenerate perturbation theory and the linear Stark effect, polarizability and vibrational Stark probes, fine structure as a sum of perturbations, and asymptotic series.

· · ·

Three questions to carry with you.

Why does second order always lower the ground state?tap
Every term has a positive numerator and a negative denominator E₀ − E_m.
Why does hydrogen n = 2 have a linear Stark effect?tap
2s and 2p₀ are degenerate; the field mixes them into states with permanent dipoles ±3ea₀.
Why can’t the x⁴ series converge?tap
It would then also converge for small negative λ, where the system has no ground state.

And one to work out. With H=12p2+12x2+λx4H = \tfrac12p^2 + \tfrac12x^2 + \lambda x^4 and λ=0.2\lambda = 0.2, what is the first-order shift of the ground-state energy?

E⁽¹⁾ =

Sources: J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, ch. 5. C. M. Bender and T. T. Wu, Phys. Rev. 184, 1231 (1969). F. J. Dyson, Phys. Rev. 85, 631 (1952). S. D. Fried, S. Bagchi and S. G. Boxer, Science 346, 1510 (2014). X. Fan et al., Phys. Rev. Lett. 130, 071801 (2023) (electron magnetic moment). Checked by code: exact E₀(λ = 0.1) = 0.559142 by diagonalization; partial sums 0.575, 0.549, 0.570, 0.545, 0.581, 0.517; λx gives ½ − λ²/2 exactly; n = 2 fine-structure splitting 45.3 µeV = 10.95 GHz.

Next · Chapter 45
Time-dependent perturbation theory and Fermi's golden rule

Quantum evolution is perfectly reversible. So why does an excited atom decay at a steady rate?

Turn the page