next up previous contents index
Next: Degenerate Perturbation Theory for Up: Perturbation Theory for Fine Previous: Perturbation Theory for Fine   Contents   Index

Degenerate Perturbation Theory

Assume a \bgroup\color{col1}$ d$\egroup-fold degenerate energy level \bgroup\color{col1}$ E$\egroup with \bgroup\color{col1}$ d$\egroup degenerate eigenstates of \bgroup\color{col1}$ \hat{H}_0$\egroup
$\displaystyle \vert 1\rangle, \vert 2\rangle, \dots, \vert d\rangle,\quad \hat{H}_0\vert i\rangle = E\vert i\rangle.$     (3.16)

The perturbation \bgroup\color{col1}$ \hat{H}_1$\egroup leads to new eigenfunctions
$\displaystyle x_1 \vert 1\rangle + x_2 \vert 2\rangle + ...+ x_d \vert d\rangle...
...quad
{\bf d}^T \equiv = (\vert 1\rangle, \vert 2\rangle, \dots, \vert d\rangle)$     (3.17)

where the notation \bgroup\color{col1}$ {\bf x} \cdot {\bf d}^T$\egroup is just an abbreviation using the coefficient vector \bgroup\color{col1}$ {\bf x}$\egroup and the vector of the degenerate states \bgroup\color{col1}$ {\bf d}^T$\egroup. The coefficient vectors \bgroup\color{col1}$ {\bf x}$\egroup are then determined from the matrix eigenvalue equation
$\displaystyle \underline{\underline{H}} {\bf x} = E' {\bf x},\quad \underline{\underline{H}}_{ij} \equiv
\langle i \vert \hat{H}_1 \vert j \rangle$     (3.18)

with the Hermitian \bgroup\color{col1}$ d$\egroup times \bgroup\color{col1}$ d$\egroup matrix \bgroup\color{col1}$ \underline{\underline{H}}$\egroup of the matrix elements of the perturbation \bgroup\color{col1}$ \hat{H}_1$\egroup in the sub-space of the degenerate eigenstates \bgroup\color{col1}$ \vert i\rangle$\egroup of \bgroup\color{col1}$ \hat{H}_0$\egroup.

The solutions for \bgroup\color{col1}$ E'$\egroup are determined from \bgroup\color{col1}$ \det \left( \underline{\underline{H}} - E' \underline{\underline{1}}\right)=0 $\egroup or

$\displaystyle \left\vert \begin{array}{cccc}
\langle 1 \vert \hat{H}_1 \vert 1\...
.... & \langle d \vert \hat{H}_1 \vert d \rangle - E'\\
\end{array}\right\vert=0,$     (3.19)

which is an algebraic equation with \bgroup\color{col1}$ d$\egroup real solutions \bgroup\color{col1}$ E'_i$\egroup, \bgroup\color{col1}$ i=1,...,d$\egroup. Correspondingly, one obtains \bgroup\color{col1}$ d$\egroup coefficient vectors \bgroup\color{col1}$ {\bf x}_i$\egroup leading to \bgroup\color{col1}$ d$\egroup new linear combinations \bgroup\color{col1}$ {\bf x}_i \cdot {\bf d}^T$\egroup, \bgroup\color{col1}$ i=1,...,d$\egroup, of states within the \bgroup\color{col1}$ d$\egroup-dimensional subspace spanned by \bgroup\color{col1}$ \vert 1\rangle, \vert 2\rangle, \dots, \vert d\rangle$\egroup.

Exercise 1: Revise if necessary Gasiorowicz [3] cp. 11.2, plus the corresponding math background: eigenvalues, eigenvalue equations, vector spaces, matrices etc. !

Exercise 2: Revise degenerate perturbation theory by applying it to the 2-level system \bgroup\color{col1}$ {\mathcal H}_{\rm TLS}$\egroup from section II.2 for the case \bgroup\color{col1}$ \varepsilon=\varepsilon_L-\varepsilon_R=0$\egroup. How good is first order perturbation theory in this case?


next up previous contents index
Next: Degenerate Perturbation Theory for Up: Perturbation Theory for Fine Previous: Perturbation Theory for Fine   Contents   Index
Tobias Brandes 2005-04-26