Two-Level Systems

A two-level system is a quantum system whose relevant Hilbert space is two-dimensional. It is the canonical finite-dimensional model for superposition, level splitting, coherent oscillation, avoided crossings, spin-$1/2$ dynamics, and qubit language.

Choose an orthonormal basis

$$\{\lvert1\rangle,\lvert2\rangle\}.$$

A general pure state is

$$\lvert\psi\rangle =c_1\lvert1\rangle+c_2\lvert2\rangle, \qquad \lvert c_1\rvert^2+\lvert c_2\rvert^2=1.$$

The basis states need not be energy eigenstates. Many mistakes in two-level physics come from confusing the displayed basis with the Hamiltonian’s eigenbasis.

Why Two Levels Are Useful

Two-level models appear when only two states are relevant over the timescale and energy range of interest. The rest of the Hilbert space is ignored because it is far off resonance, weakly coupled, or intentionally projected out.

Common examples include:

• left and right localized states in a double well;
• spin up and spin down for a spin-$1/2$ particle;
• two selected atomic or molecular levels;
• a tight-binding dimer with one particle on two sites;
• the computational states of an idealized qubit.

A two-level model is therefore an approximation unless the physical Hilbert space is truly two-dimensional. Its quality depends on how well other states decouple.

General Hamiltonian

In a chosen basis, a closed two-level Hamiltonian is a Hermitian two-by-two matrix:

$$H= \begin{pmatrix} E_1 & \Delta\\ \Delta^* & E_2 \end{pmatrix}.$$

The diagonal entries are the energies of the basis states only when the coupling $\Delta$ is absent. The off-diagonal entry $\Delta$ mixes the two basis states.

If $\Delta=0$, then $\lvert1\rangle$ and $\lvert2\rangle$ are stationary energy eigenstates. If $\Delta\ne0$, the energy eigenstates are superpositions of the basis states.

The detailed diagonalization and dynamics are developed in Two-State Hamiltonians.

Energy Splitting

A common energy shift does not change transition probabilities. It is often useful to separate the Hamiltonian into an average energy plus a traceless part:

$$H=c_0I+\mathbf b\cdot\boldsymbol\sigma.$$

Here $\boldsymbol\sigma=(\sigma_x,\sigma_y,\sigma_z)$ is the vector of Pauli matrices. The two energies are

$$E_\pm=c_0\pm\lvert\mathbf b\rvert.$$

The splitting is

$$\Delta E=E_+-E_-=2\lvert\mathbf b\rvert.$$

The vector $\mathbf b$ is often called an effective field. For actual spin in a magnetic field it is directly related to the magnetic interaction; for other two-level systems it is a compact mathematical parametrization.

Coherent Oscillation

The simplest coupled two-level Hamiltonian is

$$H= \begin{pmatrix} 0 & K\\ K & 0 \end{pmatrix}, \qquad K\gt 0.$$

The energy eigenstates are

$$\lvert+\rangle =\frac{1}{\sqrt2} \left(\lvert1\rangle+\lvert2\rangle\right), \qquad \lvert-\rangle =\frac{1}{\sqrt2} \left(\lvert1\rangle-\lvert2\rangle\right),$$

with energies $+K$ and $-K$. If the system begins in $\lvert1\rangle$, then

$$P_2(t) =\sin^2\left(\frac{Kt}{\hbar}\right).$$

The system oscillates between the basis states because the initial localized basis state is not an energy eigenstate. This is the same finite-dimensional mechanism behind double-well tunneling oscillations, Rabi oscillations, and tight-binding dimers, though the physical interpretations differ.

Basis Choice

The same state can be described in many bases. For example, the energy basis diagonalizes $H$, while a localized basis may make physical preparation or measurement simpler.

Changing basis does not change the physics. It changes which coefficients and matrix entries carry the information. The invariant content includes:

• the energy splitting;
• transition probabilities between specified preparation and measurement states;
• relative phases accumulated under time evolution;
• expectation values of physical observables.

The two-level system is small enough that this lesson is visible immediately, which is why it is such a useful model.

Physical Interpretations

The same two-dimensional mathematics supports different physical stories:

• In a double well, basis states may be left-localized and right-localized wavefunctions.
• For spin-$1/2$, basis states may be eigenstates of one spin component.
• For an atom, basis states may be two selected internal energy levels.
• For a qubit, basis states are computational states selected by a control architecture.
• For a dimer, basis states may be localized site orbitals.

Do not assume that all two-level systems are qubits. A qubit is a controlled, measurable, sufficiently coherent two-level system used for quantum information tasks.

Common Mistakes

• Treating the chosen basis as automatically diagonalizing the Hamiltonian.
• Forgetting normalization, especially after changing basis.
• Ignoring the common energy shift $c_0I$ when it is physically irrelevant.
• Calling every two-level model a qubit.
• Forgetting that a two-level approximation can fail when other states become resonant or strongly coupled.
• Confusing energy splitting with off-diagonal coupling; they agree only in special symmetric cases.

Where This Is Used

• Two-State Hamiltonians gives the general diagonalization and dynamics.
• Pauli-Matrix Hamiltonians develops the $c_0I+\mathbf b\cdot\boldsymbol\sigma$ form.
• Bloch Sphere: Wave-Mechanics Perspective visualizes pure two-level states after global phase is removed.
• Coupled Wells and Avoided Crossings connects the two-level Hamiltonian to localized wave-mechanics states.
• Tight-Binding Dimer gives the two-site lattice version of the same two-state algebra.
• Landau–Zener Problem: First Encounter introduces swept avoided crossings.
• Rabi Oscillations: First Encounter introduces near-resonantly driven population oscillations.
• Two-Level Atom shows how a physical multilevel atom is reduced to the driven two-state model and how leakage, detuning, and frame conventions are audited.
• Bits, Qubits, Qudits, and Modes distinguishes an abstract qubit, a selected physical two-level subspace, and a protected logical encoding.
• Spin-1/2 as a Canonical System: First Encounter translates the two-level algebra into the minimal spin-$1/2$ magnetic-field model.
• Double-Well Potential provides a wave-mechanics source of two-state tunneling.
• Spin-Half Hilbert Space gives the canonical spin example.
• Pauli Matrices gives the two-by-two matrix basis.
• Two-Level System Model Card is the quick reference entry.

References

• J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, 3rd ed., Cambridge University Press, 2020.
• R. Shankar, Principles of Quantum Mechanics, 2nd ed., Springer, 1994.
• D. J. Griffiths and D. F. Schroeter, Introduction to Quantum Mechanics, 3rd ed., Cambridge University Press, 2018.
• M. A. Nielsen and I. L. Chuang, Quantum Computation and Quantum Information, Cambridge University Press, 2010.

Exercises

1. Normalize the state $\lvert\psi\rangle=A(\lvert1\rangle+i\lvert2\rangle)$.

Solution

The norm is

$$\langle\psi\vert\psi\rangle =\lvert A\rvert^2 \left( \langle1\vert1\rangle+\langle2\vert2\rangle \right) =2\lvert A\rvert^2.$$

Normalization requires $2\lvert A\rvert^2=1$, so one may choose

$$A=\frac{1}{\sqrt2}.$$

2. Diagonalize $H=\begin{pmatrix}0&K\\K&0\end{pmatrix}$.

Solution

The symmetric vector $(1,1)/\sqrt2$ has eigenvalue $K$, and the antisymmetric vector $(1,-1)/\sqrt2$ has eigenvalue $-K$. Thus

$$\lvert+\rangle =\frac{1}{\sqrt2} \left(\lvert1\rangle+\lvert2\rangle\right), \qquad E_+=K,$$

and

$$\lvert-\rangle =\frac{1}{\sqrt2} \left(\lvert1\rangle-\lvert2\rangle\right), \qquad E_-=-K.$$

3. Explain why a two-level approximation can break down under a strong drive.

Solution

A two-level approximation assumes all other states remain dynamically irrelevant. A strong drive can couple the chosen two states to additional levels, shift energies, or make off-resonant transitions significant. Once other states acquire appreciable amplitude, the two-dimensional model no longer gives a closed description.