Categories: Perturbation, Physics, Quantum mechanics.

Time-independent perturbation theory

Time-independent perturbation theory, also known as stationary state perturbation theory, is a specific application of perturbation theory to the time-independent Schrödinger equation in quantum physics, for Hamiltonians of the following form:

H^=H^0+λH^1\begin{aligned} \hat{H} = \hat{H}_0 + \lambda \hat{H}_1 \end{aligned}

Where H^0\hat{H}_0 is a Hamiltonian for which the time-independent Schrödinger equation has a known solution, and H^1\hat{H}_1 is a small perturbing Hamiltonian. The eigenenergies EnE_n and eigenstates ψn\Ket{\psi_n} of the composite problem are expanded in the perturbation “bookkeeping” parameter λ\lambda:

ψn=ψn(0)+λψn(1)+λ2ψn(2)+...En=En(0)+λEn(1)+λ2En(2)+...\begin{aligned} \Ket{\psi_n} &= \ket{\psi_n^{(0)}} + \lambda \ket{\psi_n^{(1)}} + \lambda^2 \ket{\psi_n^{(2)}} + ... \\ E_n &= E_n^{(0)} + \lambda E_n^{(1)} + \lambda^2 E_n^{(2)} + ... \end{aligned}

Where En(1)E_n^{(1)} and ψn(1)\ket{\psi_n^{(1)}} are called the first-order corrections, and so on for higher orders. We insert this into the Schrödinger equation:

H^ψn=H^0ψn(0)+λ(H^1ψn(0)+H^0ψn(1))+λ2(H^1ψn(1)+H^0ψn(2))+...Enψn=En(0)ψn(0)+λ(En(1)ψn(0)+En(0)ψn(1))+λ2(En(2)ψn(0)+En(1)ψn(1)+En(0)ψn(2))+...\begin{aligned} \hat{H} \Ket{\psi_n} &= \hat{H}_0 \ket{\psi_n^{(0)}} + \lambda \big( \hat{H}_1 \ket{\psi_n^{(0)}} + \hat{H}_0 \ket{\psi_n^{(1)}} \big) \\ &\qquad + \lambda^2 \big( \hat{H}_1 \ket{\psi_n^{(1)}} + \hat{H}_0 \ket{\psi_n^{(2)}} \big) + ... \\ E_n \Ket{\psi_n} &= E_n^{(0)} \ket{\psi_n^{(0)}} + \lambda \big( E_n^{(1)} \ket{\psi_n^{(0)}} + E_n^{(0)} \ket{\psi_n^{(1)}} \big) \\ &\qquad + \lambda^2 \big( E_n^{(2)} \ket{\psi_n^{(0)}} + E_n^{(1)} \ket{\psi_n^{(1)}} + E_n^{(0)} \ket{\psi_n^{(2)}} \big) + ... \end{aligned}

If we collect the terms according to the order of λ\lambda, we arrive at the following endless series of equations, of which in practice only the first three are typically used:

H^0ψn(0)=En(0)ψn(0)H^1ψn(0)+H^0ψn(1)=En(1)ψn(0)+En(0)ψn(1)H^1ψn(1)+H^0ψn(2)=En(2)ψn(0)+En(1)ψn(1)+En(0)ψn(2)...=...\begin{aligned} \hat{H}_0 \ket{\psi_n^{(0)}} &= E_n^{(0)} \ket{\psi_n^{(0)}} \\ \hat{H}_1 \ket{\psi_n^{(0)}} + \hat{H}_0 \ket{\psi_n^{(1)}} &= E_n^{(1)} \ket{\psi_n^{(0)}} + E_n^{(0)} \ket{\psi_n^{(1)}} \\ \hat{H}_1 \ket{\psi_n^{(1)}} + \hat{H}_0 \ket{\psi_n^{(2)}} &= E_n^{(2)} \ket{\psi_n^{(0)}} + E_n^{(1)} \ket{\psi_n^{(1)}} + E_n^{(0)} \ket{\psi_n^{(2)}} \\ ... &= ... \end{aligned}

The first equation is the unperturbed problem, which we assume has already been solved, with eigenvalues En(0)=εnE_n^{(0)} = \varepsilon_n and eigenvectors ψn(0)=n\ket{\psi_n^{(0)}} = \Ket{n}:

H^0n=εnn\begin{aligned} \hat{H}_0 \Ket{n} = \varepsilon_n \Ket{n} \end{aligned}

The approach to solving the other two equations varies depending on whether this H^0\hat{H}_0 has a degenerate spectrum or not.

Without degeneracy

We start by assuming that there is no degeneracy, in other words, each εn\varepsilon_n corresponds to one n\Ket{n}. At order λ1\lambda^1, we rewrite the equation as follows:

(H^1En(1))n+(H^0εn)ψn(1)=0\begin{aligned} (\hat{H}_1 - E_n^{(1)}) \Ket{n} + (\hat{H}_0 - \varepsilon_n) \ket{\psi_n^{(1)}} = 0 \end{aligned}

Since n\Ket{n} form a complete basis, we can express ψn(1)\ket{\psi_n^{(1)}} in terms of them:

ψn(1)=mncmm\begin{aligned} \ket{\psi_n^{(1)}} = \sum_{m \neq n} c_m \Ket{m} \end{aligned}

Importantly, nn has been removed from the summation to prevent dividing by zero later. We are allowed to do this, because ψn(1)cnn\ket{\psi_n^{(1)}} - c_n \Ket{n} also satisfies the order-λ1\lambda^1 equation for any value of cnc_n, as demonstrated here:

(H^1En(1))n+(H^0εn)ψn(1)(εnεn)cnn=0\begin{aligned} (\hat{H}_1 - E_n^{(1)}) \Ket{n} + (\hat{H}_0 - \varepsilon_n) \ket{\psi_n^{(1)}} - (\varepsilon_n - \varepsilon_n) c_n \Ket{n} = 0 \end{aligned}

Where we used H^0n=εnn\hat{H}_0 \Ket{n} = \varepsilon_n \Ket{n}. We insert the series form of ψn(1)\ket{\psi_n^{(1)}} into the λ1\lambda^1-equation:

(H^1En(1))n+mncm(εmεn)m=0\begin{aligned} (\hat{H}_1 - E_n^{(1)}) \Ket{n} + \sum_{m \neq n} c_m (\varepsilon_m - \varepsilon_n) \Ket{m} = 0 \end{aligned}

We then put an arbitrary basis vector k\Bra{k} in front of this equation to get:

kH^1nEn(1)k|n+mncm(εmεn)k|m=0\begin{aligned} \matrixel{k}{\hat{H}_1}{n} - E_n^{(1)} \Inprod{k}{n} + \sum_{m \neq n} c_m (\varepsilon_m - \varepsilon_n) \Inprod{k}{m} = 0 \end{aligned}

Suppose that k=nk = n. Since n\Ket{n} form an orthonormal basis, we end up with:

En(1)=nH^1n\begin{aligned} \boxed{ E_n^{(1)} = \matrixel{n}{\hat{H}_1}{n} } \end{aligned}

In other words, the first-order energy correction En(1)E_n^{(1)} is the expectation value of the perturbation H^1\hat{H}_1 for the unperturbed state n\Ket{n}.

Suppose now that knk \neq n, then only one term of the summation survives, and we are left with the following equation, which tells us clc_l:

kH^1n+ck(εkεn)=0\begin{aligned} \matrixel{k}{\hat{H}_1}{n} + c_k (\varepsilon_k - \varepsilon_n) = 0 \end{aligned}

We isolate this result for ckc_k and insert it into the series form of ψn(1)\ket{\psi_n^{(1)}} to get the full first-order correction to the wave function:

ψn(1)=mnmH^1nεnεmm\begin{aligned} \boxed{ \ket{\psi_n^{(1)}} = \sum_{m \neq n} \frac{\matrixel{m}{\hat{H}_1}{n}}{\varepsilon_n - \varepsilon_m} \Ket{m} } \end{aligned}

Here it is clear why this is only valid in the non-degenerate case: otherwise we would divide by zero in the denominator.

Next, to find the second-order energy correction En(2)E_n^{(2)}, we take the corresponding equation and put n\Bra{n} in front of it:

nH^1ψn(1)+nH^0ψn(2)=En(2)n|n+En(1)nψn(1)+εnnψn(2)\begin{aligned} \matrixel{n}{\hat{H}_1}{\psi_n^{(1)}} + \matrixel{n}{\hat{H}_0}{\psi_n^{(2)}} &= E_n^{(2)} \Inprod{n}{n} + E_n^{(1)} \inprod{n}{\psi_n^{(1)}} + \varepsilon_n \inprod{n}{\psi_n^{(2)}} \end{aligned}

Because H^0\hat{H}_0 is Hermitian, we know that nH^0ψn(2)=εnnψn(2)\matrixel{n}{\hat{H}_0}{\psi_n^{(2)}} = \varepsilon_n \inprod{n}{\psi_n^{(2)}}, i.e. we apply it to the bra, which lets us eliminate two terms. Also, since n\Ket{n} is normalized, we find:

En(2)=nH^1ψn(1)En(1)nψn(1)\begin{aligned} E_n^{(2)} = \matrixel{n}{\hat{H}_1}{\psi_n^{(1)}} - E_n^{(1)} \inprod{n}{\psi_n^{(1)}} \end{aligned}

We explicitly removed the n\Ket{n}-dependence of ψn(1)\ket{\psi_n^{(1)}}, so the last term is zero. By simply inserting our result for ψn(1)\ket{\psi_n^{(1)}}, we thus arrive at:

En(2)=mnmH^1n2εnεm\begin{aligned} \boxed{ E_n^{(2)} = \sum_{m \neq n} \frac{\big| \matrixel{m}{\hat{H}_1}{n} \big|^2}{\varepsilon_n - \varepsilon_m} } \end{aligned}

In practice, it is not particulary useful to calculate more corrections.

With degeneracy

If εn\varepsilon_n is DD-fold degenerate, then its eigenstate could be any vector n,d\Ket{n, d} from the corresponding DD-dimensional eigenspace:

H^0n=εnnwheren=d=1Dcdn,d\begin{aligned} \hat{H}_0 \Ket{n} = \varepsilon_n \Ket{n} \quad \mathrm{where} \quad \Ket{n} = \sum_{d = 1}^{D} c_{d} \Ket{n, d} \end{aligned}

In general, adding the perturbation H^1\hat{H}_1 will lift the degeneracy, meaning the perturbed states will be non-degenerate. In the limit λ0\lambda \to 0, these DD perturbed states change into DD orthogonal states which are all valid n\Ket{n}.

However, the n\Ket{n} that they converge to are not arbitrary: only certain unperturbed eigenstates are “good” states. Without H^1\hat{H}_1, this distinction is irrelevant, but in the perturbed case it will turn out to be important.

For now, we write n,d\Ket{n, d} to refer to any orthonormal set of vectors in the eigenspace of εn\varepsilon_n (not necessarily the “good” ones), and n\Ket{n} to denote any linear combination of these. We then take the equation at order λ1\lambda^1 and prepend an arbitrary eigenspace basis vector n,δ\Bra{n, \delta}:

n,δH^1n+n,δH^0ψn(1)=En(1)n,δ|n+εnn,δψn(1)\begin{aligned} \matrixel{n, \delta}{\hat{H}_1}{n} + \matrixel{n, \delta}{\hat{H}_0}{\psi_n^{(1)}} &= E_n^{(1)} \Inprod{n, \delta}{n} + \varepsilon_n \inprod{n, \delta}{\psi_n^{(1)}} \end{aligned}

Since H^0\hat{H}_0 is Hermitian, we use the same trick as before to reduce the problem to:

n,δH^1n=En(1)n,δ|n\begin{aligned} \matrixel{n, \delta}{\hat{H}_1}{n} &= E_n^{(1)} \Inprod{n, \delta}{n} \end{aligned}

We express n\Ket{n} as a linear combination of the eigenbasis vectors n,d\Ket{n, d} to get:

d=1Dcdn,δH^1n,d=En(1)d=1Dcdn,δ|n,d=cδEn(1)\begin{aligned} \sum_{d = 1}^{D} c_d \matrixel{n, \delta}{\hat{H}_1}{n, d} = E_n^{(1)} \sum_{d = 1}^{D} c_d \Inprod{n, \delta}{n, d} = c_{\delta} E_n^{(1)} \end{aligned}

Let us now interpret the summation terms as matrix elements Mδ,dM_{\delta, d}:

Mδ,d=n,δH^1n,d\begin{aligned} M_{\delta, d} = \matrixel{n, \delta}{\hat{H}_1}{n, d} \end{aligned}

By varying the value of δ\delta from 11 to DD, we end up with equations of the form:

[M1,1M1,DMD,1MD,D][c1cD]=En(1)[c1cD]\begin{aligned} \begin{bmatrix} M_{1, 1} & \cdots & M_{1, D} \\ \vdots & \ddots & \vdots \\ M_{D, 1} & \cdots & M_{D, D} \end{bmatrix} \begin{bmatrix} c_1 \\ \vdots \\ c_D \end{bmatrix} = E_n^{(1)} \begin{bmatrix} c_1 \\ \vdots \\ c_D \end{bmatrix} \end{aligned}

This is an eigenvalue problem for En(1)E_n^{(1)}, where cdc_d are the components of the eigenvectors which represent the “good” states. After solving this, let n,g\Ket{n, g} be the resulting “good” states. Then, as long as En(1)E_n^{(1)} is a non-degenerate eigenvalue of MM:

En,g(1)=n,gH^1n,g\begin{aligned} \boxed{ E_{n, g}^{(1)} = \matrixel{n, g}{\hat{H}_1}{n, g} } \end{aligned}

Which is the same as in the non-degenerate case! Even better, the first-order wave function correction is also unchanged:

ψn,g(1)=m(n,g)mH^1n,gεnεmm\begin{aligned} \boxed{ \ket{\psi_{n,g}^{(1)}} = \sum_{m \neq (n, g)} \frac{\matrixel{m}{\hat{H}_1}{n, g}}{\varepsilon_n - \varepsilon_m} \Ket{m} } \end{aligned}

This works because the matrix MM is diagonal in the n,g\Ket{n, g}-basis, such that when m\Ket{m} is any vector n,γ\Ket{n, \gamma} in the n\Ket{n}-eigenspace (except for n,g\Ket{n,g}, which is explicitly excluded), then the corresponding numerator n,γH^1n,g=Mγ,g=0\matrixel{n, \gamma}{\hat{H}_1}{n, g} = M_{\gamma, g} = 0, so the term does not contribute.

If any of the eigenvalues En(1)E_n^{(1)} of MM are degenerate, then there is still information missing about the components cdc_d of the “good” states, in which case we must find them some other way.

Such an alternative way of determining these “good” states is also of interest even if there is no degeneracy in MM, since such a shortcut would allow us to use the formulae from non-degenerate perturbation theory straight away.

The trick is to find a Hermitian operator L^\hat{L} (usually using symmetries of the system) which commutes with both H^0\hat{H}_0 and H^1\hat{H}_1:

[L^,H^0]=[L^,H^1]=0\begin{aligned} \comm{\hat{L}}{\hat{H}_0} = \comm{\hat{L}}{\hat{H}_1} = 0 \end{aligned}

So that it shares its eigenstates with H^0\hat{H}_0 (and H^1\hat{H}_1), meaning all the vectors of the DD-dimensional n\Ket{n}-eigenspace are also eigenvectors of L^\hat{L}.

The crucial part, however, is that L^\hat{L} must be chosen such that n,d1\Ket{n, d_1} and n,d2\Ket{n, d_2} have distinct eigenvalues 12\ell_1 \neq \ell_2 for d1d2d_1 \neq d_2:

L^n,d1=1n,d1L^n,d2=2n,d2\begin{aligned} \hat{L} \Ket{n, d_1} = \ell_1 \Ket{n, d_1} \qquad \hat{L} \Ket{n, d_2} = \ell_2 \Ket{n, d_2} \end{aligned}

When this condition holds for any orthogonal choice of n,d1\Ket{n, d_1} and n,d2\Ket{n, d_2}, then these specific eigenvectors of L^\hat{L} are the “good states”, for any valid choice of L^\hat{L}.


  1. D.J. Griffiths, D.F. Schroeter, Introduction to quantum mechanics, 3rd edition, Cambridge.