A delayed choice Mach-Zehnder experiment

Backwards in time effects?

When the photon passes the double slit or the first beamsplitter in the Mach-Zehnder, there is still a very small amount of time before it hits the detector. Depending on the configuration of your experiment you will see interference fringes or not. If you configure the experiment in such a way that you can gain information on the path of the photon, the interference fringes dissappear. If you configure it back so that the information about the path is not measured, the interference fringes reappear. Physicist John Wheeler asked the following question: ‘What will happen if you change the configuration from no path detection – interference fringes – to path detection – no interference fringes – between the moment the photon made its choice which path it would go and the moment of detection? Will it change its behavior according to the configuration on detection? That would be a backwards in time effect!’

Reality Doesn’t Exist Until We Measure It, Quantum Experiment Confirms

Photons are lightning quick. If you want to change the set-up of your experiment in the time between the moment the photon makes his choice of path or slit and the moment of detection you have to think smart. In 2015 Australian physicists used slow moving helium atoms instead of photons because moving atoms show also wave behavior and interference.

The Australian experimenters created an atomic analog of the optical Mach-Zehnder interferometer. They replaced the photons by cold helium atoms and the mirrors by laser pulses hitting the atom in such a way that it changed its course to the detector, which is an analog of a photon reflecting on a mirror. The phase change on mirror reflection of a photon was also cleverly imitated by the laser pulse. In de figure below their set-up is depicted on the right (b) so it can be compared with the optical Mach-Zehnder depicted on the left (a). The atom travels downward and is detected on the DLD detector at [0> or [1>.

Australian delayed choice experiment with cold atoms. It is an atomic version of the optical Mach-Zehnder. Left the optical Mach-Zehnder, right the atomic version.

The π/2 laser pulses – the violet arrows – are the atomic analogs of a photon hitting a beam splitter. 50% of the atoms that are hit thus do change their course, which is the atomic analog of reflection. The other 50% will continue their course, which is the atomic analog of a photon passing through the beam splitter.

The π laser pulses – the golden arrows – are the atomic analogs of full mirrors. All atoms are ‘reflected’.

Because π/2 laser pulses are the analog of a beam splitter, not applying them is the analog of removing the last beam splitter. Firing them or not is controlled by a quantum random number generator (QRNG). This means that the travelling atom cannot ‘know’ beforehand if the last beam splitter will be there.

  • If the last beam splitter is there it should show interference, which means that the atoms will always end up on [0>. Interference means that the atom shows wave behavior and is not a traveling particle.
  • If the last beam splitter is not there the atom won’t show interference and the atoms will end up in a fifty-fifty division on [0> and [1>. In that case the atoms do not show wave behavior and behave as a traveling particles.

To be or not to be

The decision to travel as a wave or as a particle has to be made at the first beam splitter, however, this decision depends on the last beam splitter being there – particle – or not being there – wave.

The outcome of the Australian experiment is that the behavior of the atom – interference or not – is always in line the ‘presence’ of the last beam splitter. See figure below. The red curve shows the hits when the last beam splitter was ‘present’. The hit rate shown by the red curve depends on an extra added phase shift. This is clearly unquestionably the case. The blue line shows the hit rate when the last beam splitter was not ‘present’. An extra added phase shift – Δφ – has clearly no influence on the hit rate, both detectors are hit evenly. There is no interference.

Result of experiment with cold atoms. When no phase difference is added and the chosen path is unknown all atoms will arrive at detector [0>. When the chosen path is known atoms arrive 50/50 at detectors [0> and [1> independently of an added phase difference.

How do we explain this outcome? We have these options:

  • The atom knows the status of the last beam splitter in the near future and adapts its behavior on that knowledge.
  • The atom changes its behavior backwards in time depending on the presence of the last beam splitter.
  • The atom, its wave or particle behavior and its detection become manifest at the moment of detection. It does not exist as a particle before measurement but only as a possibility wave.

The last option is – although contrary to our impression of a world of permanent objects – the least unbelievable in my opinion. However, if this was just one single stand-alone experiment, we could perhaps still ignore it despite the the scientific and academic status of the experimenters. Luckily, there are other experiments showing the same results.

Proceed your exploration here: The two-photon quantum eraser delayed choice experiment – 2007