Hacker News Reader

Comment by lebubule | Hacker News Reader

Recent experiments have put the claim to test

Overview https://www.nature.com/articles/nature.2016.20282

The actual work https://www.nature.com/articles/ncomms12172

> Here we report that humans can detect a single-photon incident on the cornea with a probability significantly above chance.

View on HN

> To optimize this ratio in practice, we used the multi-pixel sensor of an electron multiplying charge-coupled device (EMCCD) camera as our idler detector (Fig. 1a). As the EMCCD can detect multiple photons simultaneously, it allowed us to identify and reject, that is, post-select, all events other than those where a single-photon pair was generated with a higher efficiency than with more traditional single-photon avalanche diodes (SPAD)

How is it possible to both detect a photon and then allow it to travel to the human eye? Wouldn't detection require absorption of the photon?

"SPDC is a quantum optical technique in which correlated pairs of photons (called signal and idler) are produced probabilistically from a higher energetic pump photon in a non-linear crystal following energy and momentum conservation16,17 (Fig. 1a). By detecting one of the photons (idler) and sending the other (signal) to the observer’s eye..." I suppose conservation of momentum allows them to guarantee pairs, in which case a photon in one direction guarantees a photon in the opposite direction.

> How is it possible to both detect a photon and then allow it to travel to the human eye?

Leonard Susskind explained it like this in one of his lectures (they are on YouTube, he's an excellent explainer):

From the moment the photon is emitted, to the moment it's detected, the photon exists in entanglement with all the intermediary things it "touched". Only at the final location it's "absorbed" (with a probability). At the intermediary locations the probability ended on the low side so it passed through.

There is a process called spontaneous parametric downconversion (SPDC), where if you shine laser of frequency f at a crystal, it will with small probability p emit two photons of frequency f/2 (energy conservation in play here).

If done correctly, the outgoing laser light and the two photons all travel in different directions and so can be separated and further directed using mirrors or optical fiber cables.

Because the process is non-deterministic, what we usually do is direct one of the photon beams towards a "heralding" [1] detector, while the other is directed towards the optical setup where we need a single photon [2]. If at a given moment a photon pair is produced, then the heralding detector will click; which tells us that is also a photon currently in our optical setup.

Finally, there is a ~p^2 probability that two photon-pairs will be produced at the same time by this process (and p^3 etc). To eliminate this possibility, in this experiment their heralding detector can detect how many photons landed on it any given moment. So if they see 2 or more photons in their heralding detector, then they discard this run, because now there are multiple photons heading towards the human eye.

[1] Herald as in the guy who announced that the King was approaching.

[2] In this case, towards the human eye.

I drew a schematic using https://asciiflow.com/

                                        +---------------------+
                                        | Heralding detector  |
                                        |                     |
                                        +---------------------+
                                       /
                                      /
                       +----------+  /  Photon 1
                       |          | /
    Laser beam ------> | Crystal  | ------------> Outgoing laser beam
                       |          | \
                       +----------+  \ Photon 2
                                      \
                                       \
                                        To human eye

Interesting study, though if you look at the results 'significantly above chance' looks to be 0.60 +/- .05, which is indeed better than 0.50 but not what anyone could call reliable detection.

I wonder if trained owls could detect single photons, or if their night vision is based on just having much larger lenses that collect more light?

It seems that all rods in retinas are activated by single-photon-absorption, it's just about how many have to be activated to generate a neural signal.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4629483/

Owl eyes have really cool adaptations (see https://abcbirds.org/blog/owl-eyes/). If we can somewhat detect single photons, it seems very likely owls could do it much more repeatably.

It wasn’t clear to me whether that 0.60 accounted for the fact (mentioned in the overview) that only 10% of photons reaching the eye actually make it through.

"Significantly" in statistics does not mean "reliably", but "confidently".

Would be fun to see if it’s possible to set up the experiment as a double slit.

Would humans sense an interference pattern? Is the back of an eye a sufficient observer? :)

> Is the back of an eye a sufficient observer? :)

It definitely is. The retina "measures" photon positions, which is why you see images. Observation is just interaction, no need to consider whether A can observe B. If they interact, some kind of observation takes place.

When the photon passes through the lens and the watery liquid part, its path is bent by refraction.

Does that process also count as an "interaction", or is that different to what happens when the photon hits the retina?

It does count as interaction. The systems are different.

Yep! Any interaction counts. If something interacts with a photon, and that interaction can only happen if it goes through one of the two slits (i.e., it tells you "which way" the photon went), then the interference pattern will disappear.

What escapes my understanding with the whole "interaction is observation" thing is that I don't get how anything could ever not be interacting with a whole lot of other stuff. Gravity and EM fields are everywhere. Even if photons are somehow immune to that (which, they're not, because gravity can re-direct photons) it's my understanding that we can see the same interference patterns with particle streams of ordinary matter, and I can't for the life of me figure out how those could ever not be interacting with basically everything remotely nearby, including the entire test apparatus.

> Gravity and EM fields are everywhere.

And now you understand why "quantum gravity" is such a big question in physics right now! We don't understand it all. I actually don't know anything about how EM fields affect superposition, perhaps someone else can chime in.

The feeling of asking a question and realizing you've reached the edge of human understanding is very fun

The short answer (as I understand it) is that decoherence is not binary. The gravitational field may be able to partially resolve the position, for example, and so you only get slight reduction in interference.

https://physics.stackexchange.com/questions/34993/reversing-...

Yup, everything interacts with everything close by all the time.

The important thing is by how much, and what sort of interference patterns can this produce.

As it turns out, interference is quite hard to produce randomly because two fields only produce wavering patterns when their frequency and other parameters are almost equal.

So yes, the ball you just threw to your friend is actually spread out over a whole region, that spread is about 10^-34 m so it impact is not visible at all.

When you look at the flickering pattern a laser pointer makes on a rough surface, that flickering pattern is similar in nature to the double slit interference, and you are seeing it with your eyes.

Don't forget that the interference pattern is a statistical one, you need to average over a number of photons to "see" it emerge.

Reply on news.ycombinator.com