Did Physicists Reverse Time? The 2019 IBM Quantum Computer Experiment

Facebook ↗WhatsApp ↗

Originally published 2019-07-27.

Quantum-State Reversal on an IBM Computer in 2019

Pete Linforth/Pixabay

In 2019, a team of physicists from Russia and the United States used one of IBM's public quantum computers to make a small quantum system run "backwards" and return to its starting state. With two qubits, the system went back to where it began in about 85% of runs; with three qubits, in about 49%. The work, published in Scientific Reports by Gordey Lesovik, Valerii Vinokur and colleagues, was widely reported as physicists "reversing time". What they actually reversed was the evolution of a carefully prepared quantum state inside a computer, not time in the wider world.

That is still a remarkable result. It tells us something real about why time seems to flow in one direction, and why undoing it is so hard. Here is what the experiment did, what it did not do, and why it matters.

The arrow of time in everyday life

Most of the basic equations of physics work equally well forwards and backwards. If you filmed two billiard balls colliding and played the clip in reverse, nothing would look wrong. Yet our daily experience is one-way. A hot drink cools down in a cold room and never warms itself back up. An egg breaks easily and never reassembles. We remember yesterday, not tomorrow.

Physicists call this one-way character the "arrow of time". It is linked to the second law of thermodynamics, which says that in an isolated system, entropy (roughly, a measure of disorder or of how spread out energy is) tends to increase. The second law is statistical rather than absolute: a reversal is not strictly forbidden, it is just so overwhelmingly unlikely that we never see it. That is also why no one has ever built a working perpetual motion machine.

The puzzle is how the time-symmetric laws of individual particles add up to a time-asymmetric world. The 2019 experiment tackled that question at the smallest possible scale.

Why a single electron already has an arrow of time

In quantum mechanics, a particle such as an electron is not described by an exact position and speed. Instead it has a wave function, governed by the Schrödinger equation, which gives the probabilities of finding it in different places with different momenta.

If you pin an electron down in a small region and then let it go, its wave function spreads out over time. The longer you wait, the wider the range of places it might be found. Think of rolling a ball across a pool table in the dark: a moment after you let go, you know roughly where it is, but soon it could be almost anywhere on the table.

In principle, the spreading could reverse on its own and the wave could shrink back to its starting point. The researchers estimated how likely that is. Their answer was that, even if nature made a billion attempts every second, you would not expect to see it happen once during the age of the universe. So even a lone particle, left alone, behaves as if time has a direction.

How the IBM experiment worked

The team, from the Moscow Institute of Physics and Technology and the US Department of Energy's Argonne National Laboratory, used IBM's five-qubit "ibmqx4" processor, which was available over the cloud. Qubits are the basic units of a quantum computer; unlike ordinary bits, they can be in combinations of 0 and 1 at once.

The experiment had four stages:

  1. Prepare an ordered state. The qubits started in a simple, well-defined state, the equivalent of the ball held in your hand.
  2. Let it evolve. A program made the state spread into a more complex pattern, modelling an electron scattering off an impurity.
  3. Apply the reversal. A special operation, mathematically a complex conjugation of the state, flipped the direction of its evolution. This is the step that does not happen spontaneously in nature.
  4. Let it evolve again. Running the same evolution once more now drove the system back towards where it started, like a scattered set of pool balls rolling back into their triangle.

Each configuration was run 8,192 times. With two qubits, the system returned to its original state in 85.3% of runs. With three qubits, the success rate fell to 49.1%. In a perfect machine it would have returned every time.

Why the results were not perfect

The shortfall came from the hardware, not from a flaw in the idea. The authors pointed to three main problems: the limited coherence time of the qubits (how long they can hold quantum information before it leaks into the environment), errors in the gates that link two qubits, and errors when reading out the final result. A three-qubit run took about 7.5 microseconds, comparable to the qubits' coherence time, so the longer circuit suffered more.

The drop from two to three qubits also illustrates the main conclusion. The more complex the state, the harder it is to reverse. The difficulty grows with the size of the space the system has explored, which is exactly why reversing something as complex as a cooling cup of coffee is hopeless in practice.

Myths and misunderstandings

Headlines about this study often went further than the science. A few corrections help:

  • No one travelled in time. The computer's clock kept ticking forwards. What was reversed was the evolution of a modelled quantum state.
  • The second law was not broken. The reversal needed an external controller, the quantum computer and its operators, acting as a "supersystem". Energy and order were supplied from outside, which is consistent with thermodynamics.
  • It does not bring time machines closer. The authors argue the opposite: a reversal of this kind needs a carefully engineered external system that is extremely unlikely to arise on its own.
  • About the "ten billion pool tables" line. Popular coverage illustrated the odds with ten billion electron-sized pool tables watched every second. The paper's own estimate is phrased differently, using an attempt rate of around a billion per second, but the takeaway is the same: spontaneous reversal is practically never observed.

Why the research matters

Beyond the philosophical questions, the work has a practical angle. The authors suggested that time reversal could be used to test quantum programs. If you run a calculation and then reverse it, the computer should return exactly to its simple starting state. If it does not, something went wrong. That kind of check could be far cheaper than measuring the full state of a large quantum computer, which becomes exponentially costly as machines grow.

Since 2019, quantum hardware has improved considerably, with longer coherence times and lower error rates, and error correction has become one of the field's central goals. Experiments like this one help researchers measure how far a device is from ideal behaviour.

The study also adds to a deeper conversation. It shows that irreversibility is not just a feature of large, messy systems but appears even for a single particle, and that the effort needed to reverse a process grows with its complexity. That gives a concrete, testable picture of why our world runs one way, even if its underlying laws do not insist on it.

Sources: Lesovik et al., "Arrow of time and its reversal on the IBM quantum computer", Scientific Reports 9, 4396 (2019); ScienceAlert, Mike McRae; header image: Pete Linforth/Pixabay.

COMMENTS