Scientists observe ‘negative time’ for the first time in a quantum experiment

A photon enters a cloud of atoms and emerges on the other side. When physicists calculate how long the atoms remained excited because of that transmitted photon, the answer falls below zero.

University of Toronto researchers have turned that apparently impossible result into a laboratory measurement. Their experiment found a negative weak value for the time transmitted photons left atoms excited. It does not show light traveling backward through time, but it does show that a negative delay can govern a measurable physical interaction.

The work was released in September 2024 by Daniela Angulo, Kyle Thompson, Vida-Michelle Nixon, Andy Jiao, Howard M. Wiseman and Aephraim M. Steinberg. Most worked through the University of Toronto, while Wiseman contributed from Griffith University in Australia.

The experiment drew worldwide attention because it appeared to give physical meaning to a quantity often treated mainly as wave reshaping. On April 13, 2026, the work passed peer review and appeared in Physical Review Letters.

Light usually slows when it travels through matter. Near an atomic resonance, however, different frequency components of a pulse can be delayed by different amounts. When those components recombine, interference can reshape the pulse so its peak exits earlier than expected.

Physicists describe this shift using group delay. Under some conditions, the calculated delay is negative, meaning the outgoing pulse peak appears before that of a comparable pulse traveling without the medium.

That does not mean information moved faster than light. The leading edge of the pulse still respects causality. The apparent advance comes from the way the medium changes the pulse’s shape, rather than a photon outrunning light traveling through a vacuum.

Many researchers therefore viewed negative group delay as a mathematical description of the outgoing waveform, not evidence that anything inside the material experienced a negative duration. Steinberg’s team tested whether the number could still predict a separate physical effect inside an atomic cloud.

The researchers used a cold cloud of rubidium-85 atoms. Weak signal pulses traveled through the cloud while a separate, off-resonant probe beam crossed the same region.

As a signal photon interacted with the atoms, part of its energy could temporarily exist as a collective atomic excitation. The excitation changed the cloud’s optical properties and shifted the phase of the probe beam through an interaction known as the cross-Kerr effect.

The phase shift provided an indirect record of how strongly and for how long the atoms were excited. Instead of stopping a photon to inspect what it was doing, the team measured the trace its interaction left on the separate probe beam.

The researchers then selected trials in which a signal photon was detected after passing through the cloud without scattering away. Comparing the probe’s response in selected and unselected trials allowed them to isolate the atomic excitation associated specifically with transmitted photons.

When the phase response was integrated over time, the result followed the photon’s group delay. Under some experimental conditions, both values were positive. Under others, both became negative.

The researchers measured excitation times ranging from minus 0.82, with an uncertainty of 0.31, to plus 0.54, with an uncertainty of 0.28, relative to the experiment’s reference excitation time. The negative result appeared most clearly for the narrowest-bandwidth pulses.

The key point was not that an atom literally remained excited for less than zero seconds. The measurable effect being used as a clock reversed sign. That sign appeared in another beam, giving the negative delay a consequence beyond the position of a reshaped pulse peak.

The experiment did not track one photon along a classical path. It used a weak measurement, which extracts only a small amount of information during each trial while limiting how much the measurement disturbs the quantum system.

Because each individual measurement supplies little information, the process must be repeated many times. The researchers then combined the results to determine the average effect produced by the transmitted photons.

They also used postselection. Rather than including every photon sent toward the cloud, they retained only trials that ended with a particular outcome: a photon being successfully transmitted through the atoms.

Combining weak measurement with postselection produces a weak value. Unlike conventional measurement results, weak values can fall outside the ordinary range of possible outcomes. Under the right interference conditions, they can become unusually large or negative.

The value is not arbitrary. It predicts the average shift recorded by the measuring device, which in this case was the phase change of the probe beam. The negative result describes the complete sequence of preparation, interaction and final selection, not an ordinary duration experienced by one atom.

This distinction explains why the work generated skepticism alongside public excitement. Some descriptions suggested photons had left the atoms before entering them. The actual claim is narrower but still unusual: a conditional average describing atomic excitation became negative, and another optical field recorded its sign.

The negative-time experiment built on a 2022 study published in PRX Quantum. Steinberg’s group had already measured how long atoms remained excited because of photons that passed through an atomic cloud without ultimately being absorbed.

In that experiment, an off-resonant probe laser monitored changes in the cloud’s index of refraction. Direct photon detection allowed the researchers to separate the effects of photons that were transmitted from those that were scattered.

The results showed that transmitted photons could leave a measurable excitation history even though the atoms did not permanently absorb them. That finding challenged the intuitive assumption that only scattered or absorbed photons contribute to the time atoms spend excited.

The later experiment used the same general measurement strategy but pushed the system into conditions where theory predicted the group delay would cross below zero.

Across different pulse durations and atomic-cloud densities, the measured excitation times followed the predicted group delays. That agreement suggested the negative result was not simply an artifact caused by tracking the maximum of a distorted outgoing pulse.

A theoretical analysis published in APL Quantum in September 2025 gave the experiment a broader framework. Thompson and his colleagues treated atomic excitation as a form of quantum dwell time, which measures how long a particle’s energy occupies a particular region or state during an interaction.

Their calculations showed that the excitation time associated with transmitted photons equals the spectrally averaged group delay, including when that delay is negative.

The situation differs for photons scattered out of the forward-moving beam. For those photons, the predicted excitation time contains both the group delay and an additional, always-positive quantity called the Wigner scattering delay.

The researchers also developed a simplified model showing how a negative dwell time can emerge from quantum interference.

A transmitted photon can be described through multiple possible histories. In one history, it crosses the cloud while producing little atomic excitation. In another, part of its energy is stored collectively among the atoms before returning to forward-moving light.

Quantum mechanics combines the probability amplitudes associated with those histories. The amplitudes can reinforce one another or partially cancel.

After the experiment keeps only trials ending with a transmitted photon, destructive interference can make the weakly measured contribution associated with atomic excitation appear with a negative sign.

The model does not require energy to remain inside an atom for less than zero seconds. Instead, it shows how a conditional quantum average can become negative when different possible histories interfere.

The title of the Physical Review Letters paper "Experimental Observation of Negative Weak Values for the Time Atoms Spend in the Excited State as a Photon Is Transmitted" replaced the preprint’s provocative wording about a photon spending a negative amount of time inside an atom cloud with the more precise phrase “negative weak values.”

That wording matters. The experiment did not overturn relativity, reverse cause and effect or create a route to time travel. It showed that a negative weak value corresponds to a measurable phase response in another optical field.

Physicists continue to debate how weak values should be interpreted. Some view them as providing information about a quantum system between its preparation and final measurement. Others treat them primarily as a way of describing conditional measurement statistics.

The experiment does not settle that philosophical disagreement. It strengthens the case that, regardless of interpretation, the negative value predicts an observable laboratory effect rather than serving only as an abstract mathematical result.

In June 2026, Jiao, Nixon, Thompson and Steinberg reported another experiment extending the weak-value framework beyond the original negative-time question.

Single-photon optical nonlinearities face a difficult trade-off. Strong interaction with atoms favors light that is narrowly tuned to an atomic resonance, which generally requires a long pulse. High peak intensity, however, favors a short pulse.

Ordinarily, a photon cannot be both sharply defined in frequency and tightly confined in time because of the time-energy uncertainty relationship.

The team prepared narrowband photons near resonance, transmitted them through a cold atomic cloud and then postselected photons detected within a narrow time window. This preserved the narrow frequency range of the prepared state while localizing the selected photons in time.

The researchers measured a peak cross-phase shift six times larger than that produced by comparable Gaussian pulses without postselection, with an uncertainty of one. The result was recorded at an optical depth of about 2.4 and qualitatively matched the 2025 weak-value theory of atomic excitation.

That follow-up remains a preprint and has not completed journal peer review. Still, it shows how the original experiment has developed into a wider investigation of how preparation, interference and postselection determine the effects produced by individual photons.

The story is no longer only about a pulse that seems to arrive early. The measurements now connect negative group delay to atomic excitation, quantum dwell time and stronger photon-induced phase shifts.

What first looked like an impossible clock reading has become a reproducible sign of how competing quantum histories interfere.

The original story "Scientists observe ‘negative time’ for the first time in a quantum experiment" is published in The Brighter Side of News.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.