Notoriously, the theory of quantum mechanics reveals a fundamental weirdness in the way the world works. Commonsense notions at the very heart of our everyday perceptions of reality turn out to be violated: contradictory alternatives can coexist, such as an object following two different paths at the same time; objects do not simultaneously have precise positions and velocities; and the properties of objects and events we observe can be subject to an ineradicable randomness that has nothing to do with the imperfection of our tools or our eyesight.

Gone is the reliable world in which atoms and other particles travel around like well-behaved billiard balls on the green baize of reality. Instead they behave (sometimes) like waves, becoming dispersed over a region and capable of crisscrossing to form interference patterns.

Yet all this strangeness still seems remote from ordinary life. Quantum effects are most evident when tiny systems are involved, such as electrons held within the confines of an atom. You might know in the abstract that quantum phenomena underlie most modern technologies and that various quantum oddities can be demonstrated in laboratories, but the only way to see them in the home is on science shows on television. Right? Not quite.

We will show you how to set up an experiment that illustrates what is known as quantum erasure. This effect involves one of the oddest features of quantum mechanics--the ability to take actions that change our basic interpretation of what happened in past events.

Before we explain what we mean by that and outline the experiment itself, we do have to emphasize one caveat in the interest of truth in advertising. The light patterns that you will see if you conduct the experiment successfully can be accounted for by considering the light to be a classical wave, with no quantum mechanics involved. So in that respect the experiment is a cheat and falls short of fully demonstrating the quantum nature of the effect.

Nevertheless, the individual photons that make up the light wave are indeed doing the full quantum dance with all its weirdness intact, although you could only truly prove that by sending the photons through the apparatus and detecting them one at a time. Such a procedure, unfortunately, remains beyond the average home experimenter. Still, by observing the patterns in your experiment and by thinking about what they mean in terms of the individual photons, you can get a firsthand glimpse of the bizarre quantum world.

If you want to go straight to the home experiment, it is detailed in the sidebar. The discussion that follows here delves into the science of quantum erasers in general. This explanation will help you understand what the do-it-yourself eraser demonstrates, but you might want to come back to it after seeing what that specific kind of eraser does.

What a Quantum Eraser Erases
One of the strange features of quantum mechanics is that the behavior that something exhibits can depend on what we try to find out about it. Thus, an electron can behave like a particle or like a wave, depending on which experimental setup we subject it to. For example, in some situations particlelike behavior emerges if we ascertain the specific trajectory that an electron has followed and wavelike behavior transpires if we do not.

A standard demonstration of this duality relies on what is called a two-slit experiment (your do-it-yourself quantum eraser is similar to this experiment in that it involves two pathways, but not two slits). A source emits particles, such as electrons, toward a screen that has two slits they can pass through. The particles ultimately arrive at a second screen where each one produces a spot. Where each particle lands is to some extent random and unpredictable, but as thousands of them accumulate, the spots build up into a definite, predictable pattern. When the conditions are right for the particles to behave as waves, the result is an interference pattern--in this case a series of fuzzy bars, called fringes, where most of the particles land, with very few hitting the gaps between them.

The particles will generate the interference pattern only if each particle could have traveled through either of the two slits, and there is no way of ascertaining which slit each one passed through. The two pathways are then said to be indistinguishable and each particle acts as if it actually traveled through both slits. According to the modern understanding of quantum mechanics, interference occurs when indistinguishable alternatives are combined in this way.

When two or more alternatives coexist, the situation is called a superposition. Erwin Schrödinger highlighted the oddity of quantum superpositions in 1935, when he proposed his now infamous concept of a cat that is simultaneously alive and dead, sealed inside a hermetic box where it cannot be observed. When quantum interference happens, something in the experiment is like a kind of Schrödinger's cat. But instead of being alive and dead at the same time, the cat may be walking by a tree, passing on both sides of it simultaneously.

Schrödinger's cat ceases to be in a superposition as soon as we look inside its box: we always see it to be either alive or dead, not both (although some interpretations of quantum mechanics have it that we become in a superposition of having seen a dead or a live cat). If a spotlight is shining near the tree, we see the quantum cat go one way or the other. Similarly, we can add a measurement tool to watch each particle as it passes the slits. One could imagine having a light shining on the slits so that as each particle comes through we can see a flash of light scatter from where the particle went. The flash makes the two alternative pathways distinguishable, which destroys the superposition, and the particles arrive at the final screen not in a pattern of fringes but in one featureless blob. Experiments analogous to this scenario have been conducted, and, as predicted by quantum mechanics, no interference pattern builds up.

We need not actually "do the looking." We do not have to detect the light flashes and ascertain which way each particle went. It suffices that the information is available in the flashes and could have been observed in that way.

Now we finally get to the quantum eraser. The eraser is something that can erase the information indicating which path each particle has followed, thereby restoring the indistinguishability of the alternatives and restoring interference.

How might an eraser do that? Imagine that the "flash of light" that scatters from each particle is a single photon. For the photon to reveal the "which path?" information of the particle, it must be possible (even if only in principle) to tell which slit the photon came from. That means we must be able to measure the position of where each photon scattered accurately enough to tell the slits apart. Heisenberg's uncertainty principle, however, tells us that if we instead measure the momentum of each photon with great accuracy, then the photons' positions become less well defined. So if we pass the photons through a lens that makes their momentum information available, the information about their positions is erased. When that happens, the two paths the particles can follow are again indistinguishable and interference is restored.

We have omitted one last tricky detail, but we will come back to that. First, stop and think a bit more about what is happening in the erasing process we just described, because that is where the weirdness lies. When we detect the position where one of the photons scattered, we learn which slit its corresponding particle went through, which means the particle did go through one slit or the other, not both. If we instead detect the photon's momentum, however, we cannot know which slit the particle went through. What is more, when we do many momentum measurements and see an interference pattern, we infer that in those cases the particles went through both slits (interference would be impossible otherwise).

In other words, the answer to the question, "Did the particle go through one slit or both slits?" depends on what we do with its corresponding photon long after the particle has gone through. It is almost as if our actions with the photons influence what has happened in past events. We can find out which slit the particle went through, or with our quantum eraser we can delete that information from the universe.

Strangest of all, we can decide which measurement to make after the particle has passed through the slits--we can have the apparatus for both alternative measurements in place, with a switch that we flick one way or the other just before each photon arrives. Physicists call this variation a delayed-choice experiment, an idea introduced by John A. Wheeler of the University of Texas at Austin in 1978 that extends a scenario that Niels Bohr and Albert Einstein used in their arguments about quantum mechanics and the nature of reality in 1935.

At this point, some particularly clever readers will be worrying about a fundamental problem that seems to undermine what we have just described: Why can't we delay the choice of our photon measurement until after we have seen if the particles form an interference pattern? We could, in fact, arrange to do just that by having the final screen not too far from the slits and the photon detector much farther away. So what would happen if we saw the particles form fringes but then chose to do photon position measurements that should prevent such fringes from forming? Wouldn't we have created a paradox? Surely we would not expect the already registered interference pattern to vanish! Similar reasoning suggests we could use the delayed-choice effect to transmit messages instantaneously over arbitrary distances (thereby circumventing the speed of light).

That tricky detail that we omitted earlier is what saves the day: to see the interference of the particles after applying the quantum eraser, we first have to divide them into two groups and observe the groups separately. One group will display the original pattern of fringes; the other will display the inverse of that pattern, with particles landing on what were originally the dark bands and avoiding the places where the bright fringes were. The two groups combined fill in all the gaps, hiding the interference.

The paradox is avoided because we need data from the photon measurement to know which group each particle belongs to. Thus, we cannot observe the fringes until after we have done the photon measurements, because only then do we know how to split the particles into groups. In the home experiment, dividing particles into groups is done for you automatically because one group gets blocked by a polarizing filter, and you can therefore see the interference pattern of the group that gets through with your own eyes. In the final step you can see the interference patterns of the two groups right next to each other.

From a practical standpoint, the inability to send messages faster than the speed of light and create a paradox is perhaps disappointing, but physicists and logicians consider it to be a very good feature.