Skip to content
Tutorials Center
My learning

Topics

Lesson 36 of 39 · Entanglement, Measurement and Decoherence

Decoherence and the Quantum-to-Classical Transition

Simulate a qubit dephasing through entanglement with an environment, see how coherence decays with the environment's size, and learn what decoherence does and does not explain.

Why don't we see tables in superpositions of two places? Why does a quantum computer lose its information so quickly? Both questions have the same answer: decoherence. No object is perfectly isolated. A system interacts with air molecules, photons and the vibrations of its surroundings. Those interactions entangle it with its environment, and as the previous lesson showed, a subsystem of an entangled state is a mixed state. The interference that made the system quantum is carried off into correlations that nobody can measure in practice.

Decoherence is a well-understood consequence of ordinary quantum mechanics, not an extra postulate. It explains very well why interference terms vanish for large objects and why certain states, the pointer states, are stable. It does not by itself explain why one particular outcome occurs; we return to that at the end.

A solvable model: pure dephasing

Take a central qubit S coupled to N environment qubits (a "spin bath"). The interaction conserves σ_z of the central qubit:

Text
H = (1/2) σ_z^S ⊗ Σ_k g_k σ_z^{(k)}        (ħ = 1)

Prepare the central qubit in (|0⟩ + |1⟩)/√2 and each environment qubit in |+⟩ = (|0⟩ + |1⟩)/√2. Because H is diagonal in the σ_z basis, the central qubit's populations |0⟩, |1⟩ never change. But the environment evolves differently depending on whether the qubit is |0⟩ (σ_z = +1) or |1⟩ (σ_z = -1): it rotates one way or the other. The joint state at time t is

Text
(|0⟩|E₀(t)⟩ + |1⟩|E₁(t)⟩)/√2

and tracing out the environment, the off-diagonal element (the coherence) of the central qubit is

Text
ρ₀₁(t) = ½ ⟨E₁(t)|E₀(t)⟩ = ½ Π_k cos(g_k t)

Each environment qubit contributes a factor cos(g_k t) of magnitude at most one. A product of many such factors becomes tiny quickly. The coherence of the central qubit decays toward zero, even though the total system evolves unitarily and nothing is lost: the "missing" information is in the correlations between S and the environment.

The first program checks this formula against a brute-force simulation of the complete state vector with N = 4.

Python Central-spin dephasing: brute force versus the cosine product

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Bigger environments, faster decoherence

With more environment qubits there are more factors in the product. For random couplings, Π|cos(g_k t)| falls roughly like a Gaussian exp(-t²/τ²) in time, with τ shrinking as 1/√N. Moreover, the coherence can recur if all factors happen to come back near 1, but for a large environment of different couplings the recurrence time grows enormously (like 2^N or worse), so for any practical purpose the coherence is gone for good. For a macroscopic object, the number of environmental degrees of freedom that "see" it (air molecules, thermal photons, internal vibrations) is astronomical.

The next program shows the effect of the environment size, and then estimates how fast decoherence is for a moving particle using a formula due to Zurek: for a particle of mass m in a thermal environment at temperature T, the decoherence time for a superposition of two positions separated by Δx is shorter than the relaxation (damping) time τ_R by the factor

Text
τ_D / τ_R = (λ_T / Δx)²,        λ_T = ħ / √(2 m k_B T)

where λ_T is the thermal de Broglie wavelength. This is a model result (for a harmonic oscillator weakly coupled to a thermal bath in the high-temperature limit) and should be read as an order-of-magnitude statement; it shows why the ratio becomes absurdly small for large m and Δx.

Python Coherence versus environment size, and Zurek's decoherence-time ratio

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

What decoherence achieves

  • Why interference disappears. For large, warm, interacting objects the off-diagonal terms of the density matrix vanish in a fraction of a second or far less. Superpositions of macroscopically distinct states (a "cat") decohere so fast that they cannot be observed in practice.
  • Pointer states. The environment selects a preferred basis, the states that are least disturbed by the interaction (here the σ_z eigenstates). Position is usually the pointer basis for macroscopic objects because most interactions depend on position. This explains why the world looks "classical" in position rather than in some strange superposition basis (Zurek calls this einselection).
  • Technology. Decoherence time sets the limit of quantum computers and sensors. Superconducting qubits typically keep coherence for tens to hundreds of microseconds (the best devices reach around a millisecond); trapped-ion qubits can stay coherent for seconds or longer. Error correction, shielding and cooling fight decoherence.
  • Quantum biology and chemistry tools use the same models to estimate how long coherence can last in warm, wet environments. These are open research questions.

What decoherence does not do

Decoherence turns the pure entangled state into a reduced density matrix that looks like a classical statistical mixture, with probabilities ½ and ½. But it describes the ensemble (or the reduced state): the density matrix is ½|0⟩⟨0| + ½|1⟩⟨1|, and that is also the mixed state you would get if you hadn't known which one was prepared. The global state is still a superposition of both branches. So decoherence by itself does not select one outcome. To get a single outcome from the many in a given run, you still need an interpretation: collapse, branches (many-worlds), hidden positions (Bohmian), or an epistemic view. Decoherence is a bridge from quantum to classical appearances, widely accepted; whether it solves the measurement problem is disputed.

Check your understanding

Quick check · 4 questions

of 4 answered You scored / 4 · perfect!

  1. 1. In the pure-dephasing model, what changes with time for the central qubit?

    H commutes with σ_z of the central qubit, so populations are fixed; entanglement with the environment destroys the coherence between |0⟩ and |1⟩.
  2. 2. Why does the information 'lost' by decoherence not violate unitarity?

    The joint system evolves unitarily; tracing out the environment hides the phase relations, which now live in the entangled state.
  3. 3. What are 'pointer states'?

    Einselection: the interaction picks out a basis of states that are not entangled further by the environment, such as position eigenstates for macroscopic objects.
  4. 4. Which statement is correct about decoherence and the measurement problem?

    Decoherence leaves the global state as a superposition; selecting one outcome requires an interpretation.

Summary

  • Interactions entangle a system with its environment; tracing out the environment leaves a mixed state with decaying coherences.
  • In the central-spin model ρ₀₁(t) = ½ Π cos(g_k t): more environment qubits mean faster, practically irreversible decay.
  • Decoherence picks out pointer states (einselection) and explains why macroscopic superpositions are never seen and why quantum devices need isolation.
  • Zurek's estimate τ_D/τ_R = (λ_T/Δx)² shows why decoherence is extremely fast for big, warm objects.
  • Decoherence explains the appearance of classical mixtures but not, by itself, why one outcome occurs.

Next: Project: Variational Versus Exact Energies for an Anharmonic Oscillator