The Quantum Adiabatic Theorem: Statement and Conditions

From Classical Invariants to Quantum Tracking

Before diving into the quantum world, let's build some intuition from a classical system you're familiar with: the simple harmonic oscillator. Imagine a pendulum whose length is being slowly changed. The energy of the pendulum is not conserved because we are doing work on it. However, if the length changes "adiabatically" (i.e., very slowly compared to the pendulum's oscillation period), a different quantity remains almost constant: the ratio of the energy to the frequency, $E/\omega$. This quantity is called an adiabatic invariant. Geometrically, this invariant is proportional to the area enclosed by the oscillator's trajectory in phase space.

This classical idea has a powerful quantum analogue. In quantum mechanics, the role of an adiabatic invariant is played by the quantum numbers that label the eigenstates. The central idea of the quantum adiabatic theorem is that if a system starts in a particular eigenstate, and the Hamiltonian governing the system changes sufficiently slowly, the system will remain in the corresponding instantaneous eigenstate throughout the evolution. It "tracks" the eigenstate as the Hamiltonian parameters change.

Let's make this more concrete with the canonical example: a spin-1/2 particle in a magnetic field $\mathbf{B}(t)$. The Hamiltonian is $H(t)=-\mu \cdot \mathbf{B}(t)$, and its instantaneous eigenstates are spin-up and spin-down along the direction of $\mathbf{B}(t)$. If we start with the spin aligned with the field (the ground state) and then slowly rotate the direction of $\mathbf{B}(t)$, the spin vector will precess around the field direction and follow it.

The Adiabatic Ansatz and Its Phases

How do we describe this evolving state mathematically? A system's evolution is always governed by the time-dependent Schrödinger equation:

At any given instant $t$, we can solve for the stationary states of the Hamiltonian at that moment. These are the instantaneous eigenstates $|{\psi }_n(t)⟩$ and instantaneous eigenvalues $E_n(t)$:

Crucially, these instantaneous eigenstates are not solutions to the time-dependent Schrödinger equation. If we simply substitute $|{\psi }_n(t)⟩$ into the equation, the time derivative $d/dt$ acts on $|{\psi }_n(t)⟩$, which is generally not zero.

The adiabatic theorem tells us that if the system starts in $|\Psi (0)⟩=|{\psi }_n(0)⟩$ and evolves adiabatically, the state at a later time $t$ will be approximately the instantaneous eigenstate $|{\psi }_n(t)⟩$, but modified by a phase factor. The correct approximate solution, or "adiabatic ansatz," takes the form:

This expression contains two distinct phase factors that are essential to get right.

1. The Dynamical Phase, ${\theta }_n(t)$: This is a direct generalization of the familiar time-evolution phase $e^{-iEt/\hslash }$ for a static Hamiltonian.

   $${\theta }_n(t)=-\frac{1}{\hslash }{\int }_0^t{E}_n(t^{\prime })d{t}^{\prime }$$

2. The Geometric (or Berry) Phase, ${\gamma }_n(t)$: This is a more subtle phase that arises from the changing direction of the state vector $|{\psi }_n(t)⟩$ in Hilbert space.

   $${\gamma }_n(t)=i{\int }_0^t⟨{\psi }_n(t^{\prime })|\frac{d}{d{t}^{\prime }}|{\psi }_n(t^{\prime })⟩d{t}^{\prime }$$

   This term is called a "geometric" phase because, as we will see in the next lesson, it depends only on the path the system's parameters trace out, not on how fast that path is traversed.

For now, the key is to recognize that the adiabatically evolving state is the instantaneous eigenstate multiplied by these two specific phase factors. The following video provides a formal statement of the theorem and these definitions.

The Conditions for Validity

The word "slowly" is doing a lot of work in our qualitative statement of the theorem. The central task for a theorist using this approximation is to know precisely when it is justified. There are two critical, intertwined conditions.

1. The Energy Gap

The most intuitive requirement is that the energy level of interest must remain separated from all other energy levels by a distinct energy gap. If two levels $E_n(t)$ and $E_k(t)$ were to cross or become degenerate at some point in time, it would take infinitesimally little energy to cause a transition between them. The adiabatic approximation breaks down completely at such a crossing. Therefore, a necessary condition is that for the state $n$ we are tracking, the gap $\Delta E_k(t)=|E_n(t)-E_k(t)|$ must be strictly greater than zero for all other states $k$ and for all time $t$ during the evolution.

The plots below visualize the consequence of this. In the adiabatic limit (large gap, slow change), the system's population smoothly follows the evolution of the initial eigenstate. When the conditions are not met, the population can oscillate wildly between different instantaneous eigenstates, signaling the failure of the approximation.

2. The Rate of Change vs. The Gap

Having an energy gap is necessary, but not sufficient. The change in the Hamiltonian must be slow in a way that is quantified by the size of that gap. To understand why, we can ask: what causes transitions away from the adiabatic path?

The culprit is the term $⟨{\psi }_k(t)|\frac{d}{dt}|{\psi }_n(t)⟩$, which couples the state $n$ to all other states $k$. A more illuminating form for this coupling can be derived by using the instantaneous eigenstate equation.

This single equation provides the quantitative heart of the adiabatic condition. For the approximation to be valid, the probability of transitioning to any other state must be small. Since probabilities go as the square of this amplitude, we need this term to be small for all $k$. This leads to a practical rule of thumb for the theorist:

The rate of change of the Hamiltonian must be much smaller than the square of the energy gap.

More formally, a common statement of the condition is:

This inequality beautifully combines the two conditions. To keep the system on the adiabatic path, you need the numerator—the matrix element of the Hamiltonian's rate of change—to be small, and the denominator—the energy gap squared—to be large. The most dangerous points in the evolution are where the gap $\Delta E$ is at its minimum; this is where the Hamiltonian must change most slowly.

Conclusion

You now have the essential components to understand and apply the quantum adiabatic theorem. It is a powerful approximation that allows us to understand the behavior of complex systems under slowly varying external conditions, a scenario ubiquitous in condensed matter and atomic physics.

Key Takeaways:

• The Adiabatic Theorem: A quantum system that starts in a non-degenerate eigenstate of a Hamiltonian $H(t)$ will remain in the corresponding instantaneous eigenstate of that Hamiltonian, up to a phase, provided $H(t)$ changes sufficiently slowly.
• The Adiabatic State: The evolving state acquires both a dynamical phase, related to the integrated energy, and a geometric phase, related to the path taken in parameter space.
• Conditions for Validity: The approximation holds if two conditions are met throughout the evolution:
  1. There is a non-zero energy gap between the state of interest and all other states.
  2. The rate of change of the Hamiltonian ($\dot{H}$) is sufficiently slow compared to the timescale set by the energy gap, as quantified by $\hslash |⟨{\psi }_k|\dot{H}|{\psi }_n⟩|\ll (E_n-E_k{)}^2$.

In our next lesson, we will shift our focus entirely to the geometric phase, $\gamma (t)$. We will unpack its mysterious definition, discover its deep connection to geometry and topology, and learn how to compute it as the celebrated Berry phase.