Can we control how fast quantum interference disappears?

Where the Rock left off

In the first article, a slab of silicon that nothing read destroyed the interference fringes of electrons flying past it, and a slab of gold barely touched them. The difference was the record. The silicon kept a trace of the electron’s path as a little heat; the gold kept almost none. A detector, it turned out, is any body that keeps a record of the path, observer or no observer. Physicists call the record-keeping surroundings the bath, and the fading of the fringes decoherence.

That raises the next question. Does the fading happen all at once, like a switch, or does it have a rate? And if it has a rate, what sets it? The answer decides whether decoherence is a fact of nature with a fixed size, or something the laboratory can measure and control.

It can be controlled. Think of each experiment below as a valve on the bath. Open it, and the fringes fade faster. Close it, and they last longer.

Turn the pressure

In 2003 Hornberger and colleagues in Vienna sent C₇₀ molecules, cages of seventy carbon atoms, through an interferometer and let a little gas into the chamber. The more gas, the fainter the fringes. The contrast did not drop in a step; it fell smoothly with the pressure, and it fell as collision theory predicted. Each gas molecule that bounced off a C₇₀ carried off a trace of where it had been. More gas, more traces, faster fading.

A year later the same group turned a different valve. Hackermüller and colleagues heated the molecules before they flew. A hot molecule glows in the infrared, and each photon it gives off is a small record of its position. The hotter the molecules, the brighter the glow, the faster the fringes faded, again by a rate the theory could calculate from the temperature.

Two valves, two different kinds of bath, one behavior: the record forms at a rate, and the rate follows the valve.

Watch it happen

The clearest view came from Paris. In 1996 Brune and colleagues trapped a few microwave photons between two superconducting mirrors and prepared the field in a state with two components, oscillating in two different phases at once. Then they watched the two components lose their coherence by sending atoms through the cavity one at a time and reading what each atom reported.

Two things stood out. The coherence did not vanish in an instant; it decayed progressively, over a measurable time. And the decay was faster the farther apart the two phases were. That is exactly what the theory predicts when one coupling does two jobs: the slow leak of photons through the mirrors both damps the field and carries the record of which phase it is in. Open the leak and both quicken together.

In 2008 Deléglise and colleagues in the same laboratory filmed the process. They reconstructed the field’s state frame by frame as it decohered and timed the fringes’ decay: 17 ± 3 milliseconds measured, against 22 predicted. A rate, measured with error bars, from a bath whose strength was known in advance.

One photon at a time

How small can a valve be? In 1994 Pfau and colleagues showed that one notch is enough. Helium atoms passed through an interferometer, and a single spontaneously emitted photon, one, was enough to destroy their spatial coherence. One record carrier, one lost pattern.

Kokorowski and colleagues went further in 2001 and counted. They scattered photons off atoms in an interferometer and plotted the fringe contrast against the separation of the paths. Near one scattered photon the contrast showed revivals: at certain separations the fringes came back. As the number of photons grew, the revivals vanished. The pattern made sense only one way. What destroys the fringes is not how many times the atom was hit. It is whether the scattered light could tell the two paths apart. Distinguishability, not interaction count.

A detector with a volume knob

Buks and colleagues built the valve into a detector. In 1998 they made an electron interferometer in a semiconductor and placed, beside one of its arms, a quantum point contact: a narrow channel whose electrical resistance changes when an electron passes nearby. Current through the channel is a which-path detector; it responds to the electron’s presence.

Then they turned the current up. The fringes faded, and the more current flowed, the more they faded, continuously. The detector had a volume knob. At low volume it kept a faint record and the fringes dimmed a little; at high volume it kept a strong record and they were gone. There is no threshold at which a detector “counts”; there is a record of adjustable strength, and the fringes follow it.

Baths made to order

By 2000 the valves could be built from scratch. Myatt, Turchette and colleagues in Boulder trapped single ions, set their motion into two components, and surrounded them with reservoirs they engineered themselves. Two were made of random classical fields applied to the trap: one shook the ion’s amplitude, the other its phase. The third was made of laser cooling and worked at zero temperature; the photons it scattered carried the record away. Each had its own rate law, and the strength of each could be set by hand. For the amplitude reservoir the fading grew as the square of the separation between the two components, as the theory says.

The two noise reservoirs are the exception that marks the rule. They dimmed the fringes in the average over many runs, and by the right rate law. But nothing kept a record. The injected noise carried away no trace of the ion’s state; it scrambled the phase differently in each run. That is the fake valve. The laser-cooling reservoir is the real one: its scattered photons leave with the record.

The laboratory has a test that tells the two apart, and it is old. In 1950 Hahn showed that when spins in a magnet dephase because each feels a slightly different field, a single pulse reverses the dephasing, and the signal comes back: the spin echo. Dephasing that an echo can undo was never a record. Nothing left; nothing had to come back. A real bath keeps the record and carries it away, and no pulse recalls it.

One rule

Put the valves side by side and one rule appears. The fringes fade faster for a stronger coupling to the bath, for a hotter bath, and, up to a point, for a wider gap between the paths. The rule holds for gas, for thermal glow, for a leaking cavity, for a detector’s current and for reservoirs built by hand.

The bath’s exactly solvable model, quantum Brownian motion, writes the rule in one line. For small separations in a hot bath, the decoherence rate is the damping rate multiplied by half the square of the separation, measured in units of the bath’s thermal wavelength. In plain words: the same friction that slows the particle down sets how fast its two paths are told apart, scaled by how far apart they are. At a given gap and temperature, the decoherence cannot be dialled without dialling the damping. Brune’s cavity showed exactly that coupling: the leak that damps the field is the leak that decoheres it.

Nothing in the rule is a free constant. Every quantity in it, the coupling, the temperature, the gap, is something the experimenter sets or measures. That is the point of the Valves. The Rock found the source of the decoherence. The Valves measure it, and show that its rate is physics, not a parameter.

What ACT makes of it

Decoherence theory predicts every rate above, and the measurements bear it out. What decoherence does not say is why, at the end of each run, one thing happens: one dot, one click, one reading. Anchored Causality Theory (ACT) adds that step, and the Valves tell it exactly where to put the rate.

In ACT the electron is a wave from start to finish, a stable wave-mode structure of one field, never a particle on a path. The bath keeps a record that tells the alternatives apart. When the record has passed beyond recall, so that nothing the system will actually undergo can restore the interference, the record is released, and one alternative becomes an event: a mark in the one actual history. Events happen at a rate, and ACT does not invent one. The rate is read off the same coupling that draws the decoherence curve. Where the bath only reads the path and does not disturb it, the two curves are one: the fading of the fringes and the arrival of events are the same curve, measured in Paris, Vienna and Boulder.

The fake valve marks the line ACT draws. Its event law counts only phase that can no longer be recovered. Classical noise that records nothing, and a dephasing that an echo undoes, make no events. A quantum bath, whose record has left, does. The distinction is not a convenience; it is what the spin echo and the random-field reservoir measure.

Where to go next

1. K. Hornberger et al., Phys. Rev. Lett. 90, 160401 (2003).
2. L. Hackermüller et al., Nature 427, 711 (2004).
3. M. Brune et al., Phys. Rev. Lett. 77, 4887 (1996).
4. S. Deléglise et al., Nature 455, 510 (2008).
5. T. Pfau, S. Spälter, Ch. Kurtsiefer, C. R. Ekstrom & J. Mlynek, Phys. Rev. Lett. 73, 1223 (1994).
6. D. A. Kokorowski et al., Phys. Rev. Lett. 86, 2191 (2001).
7. E. Buks, R. Schuster, M. Heiblum, D. Mahalu & V. Umansky, Nature 391, 871 (1998).
8. C. J. Myatt et al., Nature 403, 269 (2000); Q. A. Turchette et al., Phys. Rev. A 62, 053807 (2000).
9. E. L. Hahn, Phys. Rev. 80, 580 (1950).
10. A. O. Caldeira & A. J. Leggett, Physica A 121, 587 (1983).
11. K. Sonderegger, Anchored Causality Theory: A Record-History Ontology for Open Quantum Systems, preprint, version 3.6 (2026), doi:10.5281/zenodo.23811463.