|q⟩ Bad Qubits

intermediate · Physics · Time-Independent Perturbation Theory

Degenerate Perturbation Theory

Why the simple formulas break

The first- and second-order formulas both contain energy denominators En(0)Em(0)E_n^{(0)} - E_m^{(0)}. If two unperturbed states share the same energy — a degeneracy — that denominator is zero and the expressions diverge. The failure is not physical; it is a sign that we picked the wrong starting basis.

When a level is gg-fold degenerate, any linear combination of its eigenstates is also an eigenstate of H^(0)\hat{H}^{(0)} with the same energy. The perturbation selects a particular set of combinations as the correct zeroth-order states — the ones that diagonalize H^\hat{H}' inside the degenerate subspace. Ordinary perturbation theory assumed we already knew them; degenerate theory finds them.

The recipe

Let {a,b,}\{|a\rangle, |b\rangle, \dots\} span the degenerate subspace with common energy E(0)E^{(0)}. Build the matrix of the perturbation restricted to that subspace:

Wij=iH^j,i,jdegenerate subspace.W_{ij} = \langle i | \hat{H}' | j \rangle, \qquad i, j \in \text{degenerate subspace}.

Then:

  1. Diagonalize WW. Its eigenvalues are the first-order energy corrections Ek(1)E^{(1)}_k; the degeneracy is (generically) lifted.
  2. The eigenvectors of WW are the good zeroth-order states — the correct combinations to use if you continue to higher order.

For a two-dimensional subspace with

W=(WaaWabWabWbb),W = \begin{pmatrix} W_{aa} & W_{ab} \\ W_{ab}^* & W_{bb} \end{pmatrix},

the eigenvalues are

E±(1)=Waa+Wbb2±(WaaWbb2)2+Wab2.E^{(1)}_\pm = \frac{W_{aa} + W_{bb}}{2} \pm \sqrt{\left(\frac{W_{aa} - W_{bb}}{2}\right)^2 + |W_{ab}|^2}.

The two levels, originally coincident, split by E+(1)E(1)=2(Δ/2)2+Wab2E^{(1)}_+ - E^{(1)}_- = 2\sqrt{(\Delta/2)^2 + |W_{ab}|^2} where Δ=WaaWbb\Delta = W_{aa} - W_{bb}.

A worked split

Suppose a level at E(0)=5E^{(0)} = 5 is doubly degenerate and the perturbation couples the two states symmetrically:

W=(00.20.20).W = \begin{pmatrix} 0 & 0.2 \\ 0.2 & 0 \end{pmatrix}.

Here Waa=Wbb=0W_{aa} = W_{bb} = 0 and Wab=0.2W_{ab} = 0.2, so E±(1)=±0.2E^{(1)}_\pm = \pm 0.2. The level splits into two: 5+0.2=5.25 + 0.2 = 5.2 and 50.2=4.85 - 0.2 = 4.8. The good states are the symmetric and antisymmetric combinations 12(a±b)\tfrac{1}{\sqrt 2}(|a\rangle \pm |b\rangle), the eigenvectors of WW.

Try it

For the degenerate level above with W=(00.20.20)W = \begin{pmatrix} 0 & 0.2 \\ 0.2 & 0 \end{pmatrix}, compute the larger first-order eigenvalue E+(1)E^{(1)}_+ and return it as a number.

Run your code to see the quantum state.

Sign in on the full site to ask questions and join the discussion.