Opening the book…
\( \let\bibref\eqref %make-command_bibref-equivalent-to-eqref\)Let us use the time-propagator in a model calculation that we will refer to often. It is common to reduce or map quantum problems onto a two-level (or two-state) system. We will pick the most important states for our problem and find strategies for discarding or simplifying the influence of the remaining degrees of freedom. The two-level system is particularly common for spectroscopy in which light acts to drive transitions between a ground and excited state, and in the magnetic resonance spectroscopy for the \(\pm\) states of a spin \(\frac{1}{2}\) system. It is often used in describing an experimentally observed sub-system immersed in a continuum representative of condensed phases.
The two-level system consists of two unperturbed or "zero-order" states, \(\left|\varphi_{a}\right\rangle\) and \(\left|\varphi_{b}\right\rangle\), with energies \(\varepsilon_{a}\) and \(\varepsilon_{b}\), which are described by a zero-order Hamiltonian \(H_{0}\):
\[\begin{aligned}
\hat{H}_{0} &=\left|\varphi_{a}\right\rangle \varepsilon_{a}\left\langle\varphi_{a}\right| + \left| \varphi_{b}\right\rangle \varepsilon_{b}\left\langle\varphi_{b}\right| \\[3pt] &=\left(\begin{array}{rl} \varepsilon_{a} \! & 0 \\ 0 & \! \varepsilon_{b} \end{array}\right)
\end{aligned}\]
These states interact through a potential or coupling \(V\) of the form
\[\begin{aligned}
\hat{V} &=\left|\varphi_{a}\right\rangle V_{a b}\left\langle\varphi_{b}\right|+\left| \varphi_{b}\right\rangle V_{b a}\left\langle\varphi_{a}\right| \\[3pt]
&=\left(\begin{array}{rl} 0 \, & \! \! V_{a b} \\ V_{b a} \! \! \! & 0 \end{array}\right)
\end{aligned}\]
The full Hamiltonian for the two coupled states is
\[\begin{aligned}
\hat{H} &=\hat{H}_{0}+\hat{V} \\
&=\left(\begin{array}{rl} \varepsilon_{a} & \! \! V_{a b} \\ V_{b a} \! \! & \varepsilon_{b} \end{array}\right)
\end{aligned}\]
Since the Hamiltonian is Hermitian, we can also express the coupling matrix elements as
\[\begin{gather} V_{a b} =V_{b a}^{*}=V e^{-\mathrm{i} \phi} \\ \hat{H} =\left(\begin{array}{cc} \varepsilon_{a} & \! \! V e^{-\mathrm{i} \phi} \\ V e^{+\mathrm{i} \phi} \! \! & \varepsilon_{b} \end{array}\right) \end{gather}\]
This Hamiltonian could be used to describe a variety of problems involving two interacting quantum states, such as a dipole-dipole interaction between molecules, a mechanical interaction (or anharmonic coupling) between bond vibrations in a molecule, the action of an external light field to couple to quantum states, or the influence of the surroundings on the energy levels of a molecule.\(\bibref{Cite1} %in-text-citation\) For intermolecular interactions, this form of the Hamiltonian is sometimes referred to as the site basis to indicate that there are two particles or quantum states, the sites, interacting with one another via \(V\).
We now ask: If we prepare the system in state \(\left|\varphi_{a}\right\rangle\), what is the time-dependent probability of observing it in \(\left|\varphi_{b}\right\rangle\)? The coupling mixes the zero-order states, resulting in two eigenstates of \(\hat{H},\left|\varphi_{+}\right\rangle\)and \(\left|\varphi_{-}\right\rangle\), with corresponding energy eigenvalues \(\varepsilon_{+}\)and \(\varepsilon_{-}\). Since \(\left|\varphi_{a}\right\rangle\) and \(\left|\varphi_{b}\right\rangle\) are not eigenstates of \(\hat H\), a system prepared in one of these zero-order states will exchange population (the probability density) with the other. To describe this, our time-propagation will be performed in the eigenbasis using eq. (2.2.9), and we will need to find the transformation between these basis sets.
We start by searching for the eigenvalues of the Hamiltonian. To help simplify our notation, we define the mean energy and energy splitting between the uncoupled states as
\[\begin{align} &E=\dfrac{\varepsilon_{a}+\varepsilon_{b}}{2} \\ &\Delta=\dfrac{\varepsilon_{a}-\varepsilon_{b}}{2} \end{align}\]
We can then obtain the eigenvalues of the coupled system by solving the secular equation
\[\operatorname{det}(H-\lambda I)=0\]
giving
\[\begin{align} \varepsilon_{\pm}=E \pm \Omega \end{align}\]
with
\[\begin{align} \Omega=\sqrt{\Delta^{2}+V^{2}} \end{align}\]
To summarize, the explicit form of the energy eigenvalues of a coupled two-level system in terms of the zero-order energies and coupling is
\[\varepsilon_{\pm}=\dfrac{\varepsilon_{a}+\varepsilon_{b}}{2} \pm \dfrac{1}{2} \sqrt{\left(\varepsilon_{a}-\varepsilon_{b}\right)^{2}+4\left|V_{a b}\right|^{2}}\]
This is a common equation that will repeatedly arise in discussing quantum interactions.
To determine the eigenvectors of the coupled system \(\left|\varphi_{\pm}\right\rangle\), it proves to be a great simplification to define a mixing angle \(\theta\) that describes the relative magnitude of the coupling relative to the zero-order energy splitting through
\[\tan 2 \theta=\dfrac{V}{\Delta}\]
We see that the mixing angle adopts values such that \(0 \leq \theta \leq \dfrac {\pi} {4}\). Also, we note that
\[\begin{align} &\sin 2 \theta=V / \Omega \\[3pt] &\cos 2 \theta=\Delta / \Omega \end{align}\]
In this representation the Hamiltonian eq. (2.3.5) becomes
\[\hat{H}=E \mathbf{I}+\Delta\left(\begin{array}{cc} 1 & \! \! \tan (2 \theta) e^{-\mathrm{i} \phi} \\ \tan (2 \theta) e^{+\mathrm{i} \phi} \! \! & -1 \end{array}\right)\]
and we can express the eigenvalues as
\[\varepsilon_{\pm}=E \pm \Delta \sec 2 \theta\]
Next, we want to find \(S\), the transformation matrix that diagonalizes the Hamiltonian (i.e. \(S^{-1} \hat{H} S\) ) and transforms the coefficients of the wavefunction from the zero-order basis to the eigenbasis. The eigenstates can be expanded in the zero-order basis in the form
\[\left|\varphi_{\pm}\right\rangle=c_{a}\left|\varphi_{a}\right\rangle+c_{b}\left|\varphi_{b}\right\rangle\]
So that the transformation can be expressed in matrix form as
\[\left(\begin{array}{l} \varphi_{+} \\ \varphi_{-} \end{array}\right)=S\left(\begin{array}{l} \varphi_{a} \\ \varphi_{b} \end{array}\right)\]
To find \(S\), we substitute eq. (2.3.17) into the Schrödinger equation \(\hat{H}\left|\varphi_{\pm}\right\rangle=\varepsilon_{\pm}\left|\varphi_{\pm}\right\rangle\) to find
\[S=\left(\begin{array}{cc} \cos \theta \ e^{-\mathrm{i} \phi / 2} & \sin \theta \ e^{\mathrm{i} \phi / 2} \\ -\sin \theta \ e^{-\mathrm{i} \phi / 2} & \cos \theta \ e^{\mathrm{i} \phi / 2} \end{array}\right)\]
Note that \(S\) is unitary since \(S^{\dagger}=S^{-1}\) and \(\left(S^{T}\right)^{*}=S^{-1}\). Note also that \(S\) takes the form of a rotation matrix in Hilbert-space, and \(S^{-1}\) reverses the rotation of \(S\). The rotation is determined by the mixing angle, which encodes the projection of the eigenbasis and zero-order basis on one another.
Now, let’s examine the eigenstates and eigenvalues of the two-level system in two limits:
1. Weak coupling refers to the case \(|\mathrm{V} / \Delta| \ll 1\), or \(\theta \approx 0\). The eigenstate \(\left|\varphi_{+}\right\rangle\) corresponds to \(\left|\varphi_{a}\right\rangle\) weakly perturbed by the \(V_{a b}\) interaction. Similarly, \(\left|\varphi_{-}\right\rangle\) corresponds closely to \(\left|\varphi_{b}\right\rangle\). Stated another way, one finds that as \(\theta \rightarrow 0\), the character of \(\left|\varphi_{+}\right\rangle \rightarrow\left|\varphi_{a}\right\rangle\) and \(\left|\varphi_{-}\right\rangle \rightarrow\left|\varphi_{b}\right\rangle\). Expanding the square root in eq. (2.3.10), we find—not surprisingly—that the eigenstate energies are those we calculate from first-order perturbation theory:
\[\varepsilon_{\pm}=\varepsilon_{a}+\dfrac{\left|V_{a b}\right|^{2}}{\varepsilon_{a}-\varepsilon_{b}} \quad \text { and } \quad \varepsilon_{b}-\dfrac{\left|V_{a b}\right|^{2}}{\varepsilon_{a}-\varepsilon_{b}}\]
2. Strong coupling. When \(|V / \Delta| \gg 1\), the mixing angle \(\theta=\pi / 4\). In this case, the original \((a, b)\) basis states have become so strongly mixed that they are indistinguishable. The eigenstates are symmetric and antisymmetric combinations:
\[\left|\varphi_{\pm}\right\rangle=\dfrac{1}{\sqrt{2}}\left(\left|\varphi_{b}\right\rangle \pm\left|\varphi_{a}\right\rangle\right)\]
Note from eq. (2.3.19) that the sign of \(V\) dictates whether \(\left|\varphi_{+}\right\rangle\) or \(\left|\varphi_{-}\right\rangle\) corresponds to the symmetric or antisymmetric eigenstate. For negative couplings such that \(-V \gg \Delta \), the correspondence in eq (2.3.21) is inverted to \(\mp\), and \( \theta=-\pi / 4\). For the energy eigenvalues in the strong coupling limit, we find a quadratic dependence on splitting: \(\varepsilon_{\pm} \approx E \pm\left(V+\Delta^{2} / 2 V\right)\).
We can schematically represent the full range of energy eigenvalues with the diagram in Figure 2.3.3. Here we plot \(\varepsilon_{\pm}\)as a function of the splitting \(\Delta\) for a fixed coupling \(V\).
This diagram illustrates an avoided crossing effect. The strong coupling limit is equivalent to a degeneracy point (\(\Delta \approx 0\)) between the states \(\left|\varphi_{a}\right\rangle\) and \(\left|\varphi_{b}\right\rangle\). The eigenstates completely mix the unperturbed states yet remain split by the strength of interaction \(2 V\). Of course, this is not a real space representation. We will return to the discussion of avoided crossings when we describe potential energy surfaces and the adiabatic approximation, where the dependence of \(V\) and \(\Delta\) on position must be considered.
Now we can turn to describing the dynamics of the two-level system when we prepare the system initially in \(\left|\varphi_{a}\right\rangle\).
\[|\psi(0)\rangle=\left|\varphi_{a}\right\rangle\]
The time-propagator can be evaluated in the eigenbasis using eq. (2.2.9)
\[U(t) \ = \ \left|\varphi_{+}\right\rangle \! e^{-\mathrm{i} \varepsilon_{+} t / \hbar} \!\left\langle\varphi_{+}\right| + \left| \varphi_{-}\right\rangle \! e^{-\mathrm{i} \varepsilon_{-} t / \hbar} \! \left\langle\varphi_{-}\right|\]
Since \(\varphi_{a}\) and \(\varphi_{b}\) are not eigenstates, preparing the system in state \(\varphi_{a}\) creates a superposition of \(\left|\varphi_{+}\right\rangle\)and \(\left|\varphi_{-}\right\rangle\), which will lead to time evolution. With the help of the matrix elements of \(S\) in eq. (2.3.19) that transform the wavefunction from the zero-order \((a, b)\) basis to the eigenbasis \((\pm)\), we can use eq. (2.3.23) to evaluate the time-dependent amplitudes of initial and final states
\[\begin{align}
& \begin{aligned}
c_{a}(t) & =\left\langle\varphi_{a}|U(t)| \varphi_{a}\right\rangle \\[3pt]
& = e^{-\mathrm{i} E t / \hbar}[\cos (\Omega t / \hbar)-\mathrm{i} \cos( 2 \theta) \sin (\Omega t / \hbar)]
\end{aligned} \\[10pt]
& \begin{aligned}
c_{b}(t) &=\left\langle\varphi_{b}|U(t)| \varphi_{a}\right\rangle \\[3pt]
&=\mathrm{i} \sin (2 \theta) e^{-\mathrm{i} E t / \hbar} \sin (\Omega t / \hbar)
\end{aligned}
\end{align}\]
So, what is the probability that it is found in state \(\left|\varphi_{b}\right\rangle\) at time \(t\)? With the help of eq. (2.3.13), we obtain
\[\begin{align} \begin{aligned}
P_{b a}(t) &=\left|c_{b}(t)\right|^{2} \\[3pt]
&=\dfrac{V^{2}}{V^{2}+\Delta^{2}} \sin ^{2} \Omega_{R} t
\end{aligned} \end{align}\]
where
\[\begin{align} \Omega_{R} =\dfrac{1}{\hbar} \sqrt{\Delta^{2}+V^{2}} \end{align}\]
The probability amplitude \(P_{a b}\), or population, oscillates between the \(\varphi_{a}\) and \(\varphi_{b}\) states, on a timescale given by the Rabi Frequency, \(\Omega_{R}\), as seen in Figure 2.3.4. Both the maximum value of \(P_{b a}\) and \(\Omega_{R}\) depend on the strength of coupling. Notice for the weak coupling limit the eigenstates resemble the zero-order states, and the time-dependence disappears. In the strong coupling limit \((V \gg \Delta)\), amplitude is exchanged completely between the zero-order states at a rate given by the coupling: \(\Omega_{R} \rightarrow V / \hbar\). Even in this limit it takes a finite amount of time for amplitude to move between states. To get \(P_{b a}=1\) requires a time \(t= \frac{\pi} { 2 \Omega_{R}}= \frac {\hbar \pi} { 2 V} \). We can determine the probability that the system is still found in \(|a\rangle\) after time \(t\) is \(P_{a a}\), in much the same manner; however, with the expression for \(P_{a b}\) in hand, it is easiest to recognize that it can be quickly determined from \(P_{a a}=1-P_{a b}\)
Note that as an operator, \(\hat{U}\) can be expressed as a matrix, and eq. (2.3.24) and eq. (2.3.25) are matrix elements of the time-evolution operator in the zero-order basis, \(\hat{\mathbf{U}}_{a, b}\). We obtain \(\hat{\mathbf{U}}_{a, b}\) using a transformation from the eigenbasis:
\[\hat{\mathbf{U}}_{a, b}=\mathbf{S}^{-1} \hat{\mathbf{U}} \mathbf{S}\]
The full time-propagator for the two-level system in the matrix representation is
\[ \begin{align} \hat{\mathbf{U}}_{a, b}(t) = e^{-\mathrm{i} E t / \hbar}\left(\begin{array}{cc}
\cos \Omega_{R} t-\mathrm{i} (\tfrac {\Delta}{\Omega}) \sin \Omega_{R} t \!\! & -\mathrm{i} (\tfrac {V}{\Omega}) \sin \Omega_{R} t \\[5pt]
-\mathrm{i} (\tfrac {V}{\Omega}) \sin \Omega_{R} t & \!\! \cos \Omega_{R} t+ \mathrm{i}(\tfrac {\Delta}{\Omega}) \sin \Omega_{R} t \end{array}\right) \end{align}\]
and the time-evolution of the wavefunctions are calculated from
\[\boldsymbol{\Psi}(t)=\hat{\mathbf{U}}_{a, b} \boldsymbol{\Psi}(0) \qquad \text { with } \boldsymbol{\Psi}(0)=\left(\begin{array}{c} c_{a} \\ c_{b} \end{array}\right)\]
More generally, in the matrix representation the time propagator is an exponential operator \(\hat{\mathbf{U}}=\exp (-\mathrm{i} \hat{\mathbf{H}} t / \hbar)\), defined through its series expansion in products of Hamiltonian matrices. Evaluating the time-dependence in the eigenbasis, although not required, is far easier than the zero-order basis. The exponential of a diagonal matrix is readily obtained by exponentiating the elements, so if the Hamiltonian can be diagonalized as \(\hat{\Lambda}=S^{-1} \hat{H} S\) then \(e^{\hat{H}}=S e^{\hat{\Lambda}} S^{-1}\).
References
| \[\tag {1} \label {Cite1}\] |
When used in magnetic resonance, quantum optics, and atomic spectroscopy it is common to represent the Energies and interactions in spin \(1 / 2\) systems using the Pauli matrices \[\sigma_{x}=\left(\begin{array}{ll} 0 & 1 \\ 1 & 0 \end{array}\right) \quad \sigma_{y}=\left(\begin{array}{cc} 0 & -\mathrm{i} \\ \mathrm{i} & 0 \end{array}\right) \quad \sigma_{z}=\left(\begin{array}{cc} 1 & 0 \\ 0 & -1 \end{array}\right)\nonumber\] For which our two-level system Hamiltonian is written \(\hat{H}=E \bar{I}+\Delta \sigma_{z}+V \sigma_{x}\). |
This page titled 2.3: Two-Level System was last modified on Thu, 04 Jun 2026 01:16:17 GMT and is shared under a CC BY-NC-SA 4.0 license and was authored, remixed, and/or curated by Andrei Tokmakoff via source content that was edited to the style and standards of the LibreTexts platform.
The book could not be loaded just now.