Two state system

A two state system is the simplest to exhibit quantum behavior. For such a system, we can write the Hamiltonian as a 2 by 2 matrix in some basis {1,2}\{ \ket 1, \ket 2 \}

H=(E1ΔΔE2). \begin{align*} H = \begin{pmatrix} E_1 & \Delta \\ \Delta\conj & E_2 \end{pmatrix}. \end{align*}

Here 1H1=E1 \braket{1 | H | 1} = E_1 is the energy of being in state 1\ket 1, E2E_2 is the energy of being in state 2\ket 2, and Δ\Delta is the energy associated with transitioning between the two states (i.e. tunneling). In some basis, HH will be diagonal, but this is often not the most natural basis to write it in.

To find the eigenstates and diagonalize HH we can take a similar approach to what we did with spin. We defined the three Pauli matrices σxyz\sigma_{xyz} and the spin operators SxyzS_{xyz}, and then used these to define a spin operator Sn=nxSx+nySy+nzSzS_{\vec n} = n_x S_x + n_y S_y + n_z S_z in some arbitrary direction n=(nx,ny,nz)\vec n = (n_x, n_y, n_z). We found the eigenvalues were ±2\pm \frac\hbar2 and found the eigenvectors.

Let’s write HH in this form

H=(g0+g3g1ig2g1+ig2g0g3)=g01^+g1σ1+g2σ2+g3σ3=g01^+gσ=g01^+gnσ. \begin{align*} H &= \begin{pmatrix} g_0 + g_3 & g_1 - ig_2 \\ g_1 + ig_2 & g_0 - g_3 \end{pmatrix} = g_0 \hat 1 + g_1 \sigma_1 + g_2 \sigma_2 + g_3 \sigma_3 \\ &= g_0 \hat 1 + \vec g \cdot \vec \sigma \\ &= g_0 \hat 1 + g \, \vec n \cdot \vec \sigma. \end{align*}

We can write the gng_n constants in terms of E1,E2E_1,E_2 and Δ\Delta

g0=12(E1+E2)g3=12(E1E2)g1=Re[Δ]g2=Im[Δ]g=Δ2+14(E1E2)2 \begin{align*} g_0 &= \frac{1}{2} (E_1 + E_2) &\quad g_3 &= \frac12 (E_1 - E_2) \\ g_1 &= \re[\Delta] &\quad g_2 &= -\im[\Delta] \end{align*} \\ g = \sqrt{|\Delta|^2 + \frac14 \big(E_1 - E_2\big)^2}

The last term of the new expression gnσg \, \vec n \cdot \vec \sigma is like the spin operator SnS_{\vec n}, but with a gg instead of 2\frac\hbar2. We can use the same eigenvectors and eigenvalues we already found for SnS_n here. The g01^g_0 \hat 1 term is just the identity, and so will add g0g_0 to each eigenvalue without affecting the eigenvector.

We write the eigenstates as

H±=(g0±g)±+=(1+n3n1+in2)=(n1+in21+n3). \begin{align*} H \ket{\pm} &= \big(g_0 \pm g \big) \ket \pm \\ \ket{+} &= \mat{1 + n_3 \\ n_1 + in_2} \\ \ket{-} &= \mat{-n_1 + in_2 \\ 1 + n_3}. \end{align*}

The eigenvalues g0±gg_0 \pm g represent a ground state and an excited state. When Δ=0\Delta = 0, g1g_1 and g2g_2 disappear, and the eigenvectors become (1,0)(1,0) and (0,1)(0,1) as expected, with eigenvalues E1E_1 and E2E_2.