Can we change the past? It would seem to be obvious that this is not possible - once an event happens it is part of history and cannot be altered in anyway. Even in the most speculative theories on time travel there are usually strict restrictions on what is possible. 

But today's article is not on time travel, but rather it is on the mysterious quantum effect known as a quantum eraser, and how experiments in this field have suggested that on small scales, the future really can affect the past.

Consider this simple layout of mirrors and detectors, which guides a beam of photons. (The actual experiments that were carried out over the past twenty years are much more complicated, but this simple example demonstrates the underlying physics just as well)



A single photon from a laser is emitted into the system, depicted here by a yellow line. A special crystal is used to convert the single photon into two photons, each with half the original frequency and with complementary polarizations (usually this is done by adding a polarizer to one of the two paths). The photons are then reflected from a pair of mirrors so that the converge at a single point. At that point (shown here as a dark blue line) there is a beam splitter that will randomly reflect the photon or transmit it. Photon detectors are then placed at the end of the beam paths, and because of the random nature of the beam splitter, photons on either path can arrive at either detector. 

And now the weirdness of quantum mechanics begins...

If instead of splitting the original photon it is allowed to randomly travel on one of the two paths, then the photon detectors will detect an interference pattern. This is the famous double slit experiment in a slightly different form. The photon effectively splits into two virtual particles and travels along both paths at the same time. The two paths interfere with each other and create this interference pattern at the end.

Continuing that same single photon experiment, a detector could be placed along one of the two paths. Then each photon will either travel that path or it won't - there is no more option of a virtual particle traveling both paths. And according to quantum mechanics (and experiments) no matter how you set up that detector, it will destroy the interference pattern. Knowing which path it takes changes the properties of the photon.

The two entangled photon experiment described above is an intriguing mixture of the two options. Because the two photons are entangled, they should still demonstrate some form of interference. But the two photons have different polarizations, and as such adding a polarizing filter in front of the photon detectors would allow an experimenter to know which photon was on which path. That potential knowledge is present, even if the experimenter does not actually measure it or use it in anyway.

When this experiment was actually conducted, it was discovered that the existence of this information is equivalent to adding a detector along one path. Even if it is never used, this information destroys the interference of the two paths that the photons can follow.

What happens if we decide to erase this information completely? There are ways to 'de-polarize' the photons so that they both carry the same polarization. All information on the path taken by each photon has been erased. And so what do the detectors measure? The interference pattern has returned!

And that is the reason this experiment is fascinating. In the double slit experiment, the photon behaves as either a particle or a wave depending on whether its position is measured. But the decision happens before or during the photons travel through the experiment. Once the decision is made, it stays the same throughout the rest of the experiment.

In the quantum eraser experiment, that decision is made after the experiment is complete. Some of the recent measurements have been able to effectively make the photon into a particle or a wave tens of nanoseconds after it has already been received by the detectors. That might not seem like a long time, but it still shows that in the quantum world a decision made in the present can alter the past. (It should be noted that there are some counter-arguments in the research community that suggest explanations of this effect without this retro-causality, but they are not proven either).

And so although this is a single photon having its properties changed nanoseconds in the past, it is still experimental evidence that sometimes the future can change the past. And it shows that quantum mechanics still has some amazing surprises for us!