Hacker News new | past | comments | ask | show | jobs | submit login
Chinese scientists test quantum entanglement over unprecedented distance (scientificamerican.com)
138 points by javascript-this on June 16, 2017 | hide | past | web | favorite | 78 comments

A staple of sci-fi has been a pair of boxes with quantum entangled particles inside allowing FTL communications once the boxes travel (sub-FTL) to their destination.

By my understanding that's impossible, although I always wondered if you could coordinate FTL, by detecting when a particle was no longer entangled.

Can someone with the background comment?

Caveat: I only have a B.Science degree in Engineering Physics (Colorado, USA) and hobbyist-type research in PhD-level physics.

I've always found it easier to explain quantum entanglement thusly:

Imagine you have 2 cubes that alternate glowing between purple and orange. And they stay in one color for a random amount of time (100s of milliseconds). Holding both in your hands, you would just have randomly color-shifting cubes. But a neat trick of these cubes is that when you close your eyes, hit them together, and quickly put them into separate closed boxes, you know that three things will happen:

1. they will be locked into one color as long as the box stays closed (no more alternating)

2. each cube will be the opposite color from each other (when one is purple, the other is orange, and vice versa)

3. when you open the box and observe the cube, it will stay in the current color for 1 second, and then resume its randomness once again.

So you take one of the boxes containing a cube and ship it to your friend on the Moon (greater than 1 light-second away). When she gets the box, you call her on your phone and ask her to wait. Then you open your box and see that your cube is purple. And then the cube starts blinking again because the entanglement is broken on your end. So you wait a few seconds and say to her "open your box and I bet you the cube is orange". She does so and confirms the cube was glowing orange before resuming the random blinking. Now the entanglement is broken on her end, but you "knew" that by having the purple cube in your possession that hers would be orange.

Using this system, there is no way to exploit your knowledge of what the colors will be once observed (so sorry, no FTL communication here). The particles no longer influence each other after they are entangled at the start. It's not an active process, but rather a very specific type of setup. But it still bothers physicists (most famously Einstein), that you could "know" something without direct proof.

And people in general hope that this un-intuitive phenomenon will open up whole new worlds of sci-fi tech. But no, as the article explained: "Nothing we knew suggested this goal was unachievable. The significance of this news is not that it was unexpected or that it overturns anything previously believed, but simply that it’s a satisfying culmination of years of hard work."

If you had shipped the other cube-box to the opposite side of the galaxy, thousands of years after you were long dead, someone would open it and still find an orange cube.

The "spookiness" is more subtle than that. You could argue that the cubes have decided on their color as soon as you close the box. Why doesn't this argument work for real quantum systems?

In reality, you're not measuring colors of boxes, but spins of particles. You generate two particles with opposite spins along a known axis. If you measure the spins along that same axis, you always get the same result: one is spinning "clockwise", the other "counterclockwise".

On the other hand, if you measure one of the spins along a different axis, you'll randomly get "clockwise" or "counterclockwise" according to the angle between the original spin axis and the measurement axis. Measuring at a right angle to the original axis gives you random results, which is also what you'd expect. But what about the angles in between? Classical statistics says the correlation should vary linearly with the angle, but the actual results have a cosine factor (https://en.wikipedia.org/wiki/File:Bell.svg). This cosine factor can't be explained by any kind of classical statistical randomness. That's the "spooky" part.

But Bell's equality means that either the system is non-local or non-deterministic.

So either it's determined when you close the box, but it's spooky action at a distance. Or it's not determined on box closure (no spooky action) but it has "random-ness".

Is this fair? The real trick is science found a way to create a pair of of particles with predictable relatable spin values (though not known until observed) and the ability isolate those particles from interference.

The emphasis always seems to be on the "spookiness" of the observation, but the real trick is isolation and "entanglement" (pairing of values) in the first place.

Furthermore, nothing particularly real occurs at the time of observation, there's no way to tell that an observation of either of the particles has occurred, thus they aren't "tangled" as much encoded, isolated and separated.

[I'm not a physicist] I like your analogy for color entanglement. However, it doesn't show the "weirdness" of superposition (the balls don't have a concrete value until one of the boxes are opened, at which point, this triggers the determination of the open box's color value), nor the "spookiness" of action-at-a-distance (opening one box instantly influences the state of the ball in the other box -- which now must be the opposite color when its box is opened).

Intuitively, someone might think: well, the ball colors were "decided" while the two balls were next to each other (ballA will be orange, ballB will be purple), and the information about the color is attached to the balls, so the balls know what color to be when they are separated and later opened (they have "hidden variables" indicating the assigned color)....however, the reality (provable statistically by Bell's Theorem) is that the balls do not carry this color information, and instead the act of opening the box, randomly sets the color of BallA and instantly affects the color that BallB will have.

So if both people synchronize the time to open the box (that has some time relativity problems), so that BoxA is opened a fraction of a second before BoxB, then BoxA's color will influence BoxB's color (seen a fraction of a second later), but that will have happened faster than the speed of light would allow if BoxA was sharing its color information with BoxB.

> the balls don't have a concrete value until one of the boxes are opened,

> opening one box instantly influences the state of the ball in the other box

How do you know that without opening the box?

This analogy is confusing because it is incorrect in some fundamental ways. You aren't just opening a box and seeing one or the other color. The measurement itself determines a role in the outcome that you couldn't predict in advance.

It is more like you get to choose to measure only one of the red, green or blue channels. If you send a message containing which channel and the measurement, the other person can do the same measurement and find the complementary value. Without the channel and the original measurement the other person just sees random behavior.

The spooky part is that I don't choose the color channel to measure until after the entangled balls are separated

Bohr and Heisenberg interpret it that way.


And Schrödinger's cat never told us what really happened inside the box :)


i fail to see whats mysterious about knowing something without direct proof. You could put a literal orange cube in a box, ship it across the galaxy, and you'd "know" anyone opening the box would be an orange cube. You don't need to be present at the unboxing to know that with certainty. How is this different? Am I missing something?

Allow me to give a more (but not perfectly) accurate analogy.

I give you a magic piece of paper with a two binary digits (00/01/10/11) on it. I also keep such a paper for myself. If we both start by looking at our first digits then those will mysteriosly match, but if we then look at our second digits they will have just a 50% chance of matching. However, if we both start by looking at the last digits, then those will match, and our first digits will have a 50% chance of matching. If you and I start by looking at different digits, then the digits we read are just standard random bits.

The cool thing about this is that our magic paper will also work instantly even if we are lightyears apart. However, if you decide to look at the last digit and you see a 1, then that tells you nothing about me or which digits (if any) I have looked at, so I can't send you a message with this paper.

It sounds like an interesting, if very expensive, way to get around primes for key exchange, in the world where prime factorization is much more efficient. Need a random number, look a bun of shared digits in a predefined order. You now have a shared random number that's perfectly secure, as long as you shipped out your shared state in advance.

With quantum entanglement, what happens to one particle prior to measurement affects the other even if the effect occurred after they were parted. That's where it gets really spooky.

To belabour the colours and boxes analogy:

We have 2 balls, if one is orange the other is always purple and vice versa.

Now these balls have an extra property. If you shake them vigorously they may or may not change colour randomly.

So, we put them in boxes without looking and take them far away from each other. Now, some time after they are far apart and before you look in the box, you shake one of the boxes. You then look.

In QM, if the ball you observe is orange(purple) then you know the other is purple(orange). But the other ball can't have known ahead of time that a) you were going to shake your one and b) that it changed colour (or didn't). That's why it's spooky.

What is the equivalent to "shaking" in QM? I'm trying to figure out why it matters whether you shake the box or not (i.e. could the shaking simply not have an effect if you're not looking?)

The shaking represents any sort of action that doesn't perform a measurement. It was the first thing I thought of that could affect the ball without opening the box.

It could be that the shaking didn't have an effect. But equally, it could have had an effect. The spooky bit is that the other ball 'knows' whether it did or not in a non-local way.

But isn't it possible that performing "shaking" might not affect any "ball" when it's "inside of a box"?

Sounds like they never were two separate entities to begin with or never parted location. Which is spooky too, in a different way.

When I first started to be interested in this area, it occurred to me that it might be something like a 'stretchy' spacetime i.e. the particles are actually locally connected but appear unconnected if you use a static spacetime as a reference as we tend to. I guess something analogous to the idea that someone moving close to c experiences time different from someone in a slow moving reference frame.

Unfortunately, I have neither the experience nor ability to progress that thought much further so, if someone wants to go off and get a Nobel from it, feel free...

>>Sounds like they never were two separate entities to begin with or never parted location. Which is spooky too, in a different way.

Sounds interesting. I wouldn't be surprised if it turns out that these two particles are really just a single particle in a different dimension that we don't know it exists yet.

> Sounds like they never were two separate entities to begin with

“Separate entities” may just be an artifact of intelligences trying to divide sense data about the universe into chunks they are capable of processing.

Experimentally, when they entangle photons, they just pick two random photons.

If two random photons used in an experiment weren't ever really separate entities to being with, even before the experiment ever started, then one should probably conclude that /all/ photons are not really separate entities.

Otherwise, every quantum experiment we've ever done has just /happened/ to pick two photons that were actually one entity, unlike most other photons.

Too build on this thought - is it possible that these quantum properties aren't affected by locality? Which is obviously true given the explanation above... I guess I'm asking if there are other intuitions that can be made if locality is explicitly considered unrelated.

If I understand right, this starts to touch on Bell's Theorem.

The description as is means that the situation is identical if there's "spooky action at a distance" or if there's none and you just don't know what's in each box until you open one (and therefore can guarantee what's in the other). There are predictions from this however that when tested just don't work. There is something else going on. I would go further but I'm sure my explanations would contain huge inaccuracies that probably would cause more problems with understanding what's happening so will link you to a couple of things:



I'm sure there are others here who are well versed in this who can chip in.

That this example has very little to do with entanglement, which is not merely "knowing without direct proof". We do that everyday.

The same way, the issue with Shroedigger's cat is not that "the cat is in a single state, but we just don't know whether that's dead or alive unless we open the box" (as many laymen try to explain it).

That's just "I don't know yet" and has nothing to do with the issue Shroedigger tried to highlight or quantum mechanics (not to mention that this we encounter everyday and we can replicate perfectly with say, a box and a person that throws a dice and based on that releases or not a gas in the box with a cat -- that's not what the paradox is about).

For entanglement, see this:


> 1. they will be locked into one color

No, this is wrong! As someone else commented, this is just not knowing something. The weirdness of the quantum entanglement comes from the fact that the behavior of each cube is experimentally correctly described by equations based on the cubes continuing to change color randomly.

I probably should have clarified that my analogy was coming from a perspective of super-deterministic[1] compatibilism[2]. In the absence of knowing the truth, I feel that solution is the best model / explanation. It strips away the 'spookiness' and many other problems, though it admittedly has its own issues with testability and proof.

I do not believe in the idea that particles can super-luminally affect each other during the moment of observation. I think that is a failing of our measurement systems and ability to describe what's going on. Just as we often fail to accurately convey the Schrodinger's Cat thought experiment [3].

But for the record, if a different perspective were definitively proven to be true, I would of course switch to it, as any scientist should. But given that there are multiple avenues of investigation open, I'm exploring super-deterministic compatibilism.

[1] https://en.wikipedia.org/wiki/Superdeterminism

[2] https://en.wikipedia.org/wiki/Compatibilism

[3] http://www.smbc-comics.com/?id=2524

Sounds a little like Quanta's balloon analogy:


Damn, all of these explanations are overly complex! The most basic reason it doesn't work is because measuring your particle breaks the entanglement. So you would know that your partner has the opposite spin (or whatever) but you can't bootstrap any sort of communication system with that.

Not overly complex, but overly simple. I wish people talking about physics learn for the last time that analogies are a horrible way to explain a complex concept that at best adds an unnecessary translation layer increasing cognitive load and at worst hand waves away the most critical parts.

They just get too excited : )

>FTL, by detecting when a particle was no longer entangled

The trouble is you can't detect that. You can just make a measurement on your particle and get some result. That's the same whether it's entangled or not.

> A staple of sci-fi has been a pair of boxes with quantum entangled particles inside allowing FTL communications once the boxes travel (sub-FTL) to their destination.

Of course this is complete nonsense. It's useful for plotting stories, but seems to have imprinted on a large fraction of readers a total falsehood. Do you know where the idea started? I think the first time I saw it was in Orson Scott Card's novel "Ender's Game" (1985).

Looks like it may have started in 1966 with a story by Ursula K. Le Guin. She called the device an ansible, which is what it was also called in OSC's books.


According to that Wikipedia page, Le Guin's ansible might not be relying on entanglement.

It "doesn't involve radio waves, or any form of energy. The principle it works on, the constant of simultaneity, is analogous in some ways to gravity ... One point has to be fixed, on a planet of certain mass, but the other end is portable."

About Card's ansible, it says, "Card's description of the ansible's functions in Xenocide involves a fictional subatomic particle, the philote. In the "Enderverse", the two quarks inside a pi meson can be separated by an arbitrary distance while remaining connected by "philotic rays"."

So perhaps Card's "Xenocide" (1991) is where the entanglement => FTL communication myth began?

From the book The Dispossessed.

It's a great novel, by the way.

It's about as likely as reading a byte from an initialized /dev/urandom and measuring its entropy to see if someone removed a hardware RNG.

I'd guess it's impossible without measuring both particles and communication to compare the results.

A few weeks ago there was another announcement about the advancement in arresting the evolution of a quantum system


In this way it seems to me communication may be possible

A sort of sending bytes stead bits

I thought quantum entanglement is the idea that two quantum particles separated by distance will behave the same at the same time regardless od distance.

Why is sending photons over distances a verification of quantum entanglement?

Can someone elaborate on what this is?

It's more subtle than that. In this case with photons it's that the correlation between a photon going through a polarizer at point A and it's entangled pair going through a polarizer at point B is a function of the angle between the polarizers.

The odd thing is it depends on the relative angle of the polarizers at the time the photons go through even if they are far apart. So if you use the analogy of God rolling a dice to determine if the photon goes through or not then He has to know the angle of the other polarizer in a faster than light way even if us mortals can not know that.

Another factor to consider is that relativity showed that whether two events at different locations are simultaneous depends on the observer's reference frame. See (https://en.wikipedia.org/wiki/Relativity_of_simultaneity). So you can't even say which measurement of the state is the one that collapses the waveform. Observers in different reference frames would disagree.

I too am wondering about this. I thought entanglement couldn't be used for communication because observing both particles simultaneously will reveal that they have opposite spins. It's a conformation after observation, based on what the quantum superposition collapses into once you observe the state. But the actual state is also unknown until you observe it.

I'm not a physicist, but I'm sure there're be one on HN that will tell us what this study really means.

Not a physicist either, but as I understand it, entanglement is not useful for communication on its own, but allows for tamper-proof key exchange.

Basically, you generate a pair of entangled particles. Keep one of the pair and send the other away. Now both you and your partner randomly select a basis to measure the particle in.

After a few repetitions, you tell each other (over a classic channel) how you measured and discard measurements made in different bases. The remaining results will be strongly correlated, depending on the quality of the entanglement. By comparing a small part, you can find out exactly how much.

If the correlation is lower than expected, you might have some dust in your measurement apparatus, or an eavesdropper has intercepted a few quanta. If the correlation is good enough, you now both have a value that is secret to anyone who did not measure the particles.

The secret value can then be used as a key to encrypt your communication over the classic channel.

If this description is not satisfying, maybe the Wikipedia article explains it better: https://en.wikipedia.org/wiki/Quantum_key_distribution

Isn't that no better than a one-time pad if you need a classic channel?

Quantum key-exchange solves the distribution problem for one-time pads. And a one-time pad is the best security you could possibly get, so I'm not sure what you mean by "no better", since that's a given.

If you need a classical channel doesn't that just create another distribution problem though?

And yeah, but im considering one time pad as good as the security of its key in transfer

I am a physicist.

I only glossed the article, but from what I gathered, this didn't prove anything about QM, rather it was a technical achievement. One of the biggest problems with entangled pair distribution is actually separating the particles over some distance, because when the entangled particles interact with something ("observed") their state collapses and the entanglement may end.

This is what the achievement was, they managed to not destroy entanglement of a pair over a very long distance.

Well, if there is one, you will only know for sure after they reply. It's a conformation after observation too.

Probably it is similar to what has been done here:


Information can't travel faster than the speed of light. You have to actually take the photons and move them away from one another, then check one which will modify both.

They are breaking a record, that's all.

Wikipedia says:

"If Bohmian Mechanics as an non-local hidden variables interpretation of quantum mechanics is accurate it should allow quantum computers to implement a search of an N-item database at most in O ( N 3 ) {\displaystyle O({\sqrt[{3}]{N}})} steps. This is slightly faster than the O ( N ) O({\sqrt {N}}) steps taken by Grover's algorithm using standard interpretations of quantum mechanics. Neither search method will allow quantum computers to solve NP-Complete problems in polynomial time.[4]"

I thought there was no way to figure out which interpretation of QM is correct. Can someone explain?

It's just a wrong statement. Follow the reference to the paper and see:

> [...] we show that if we could examine the entire history of a hidden variable, then we could efficiently solve [...]

Bohmian mechanics doesn't give you access to the entire history of a hidden variable.

But in principle could we access the history?


You can't just keep measuring the position of the particle and keep a list. Recording quantum information entangles the recorder and the system under test, which affects how interference plays out, which affects the later parts of the path in a way that makes your path-computer not work.

How? Think a bit harder. It's pretty obvious once you get it. John Titor explained it to me last year.

I'm just kidding. Yeah, what you said makes sense. How do we know though that IN PRINCIPLE we can't figure out which QM interprtation is correct? Because we are sure there is no additional data/math that can be developed?

I'm pretty sure you are just mistaken in that assertion. There are experiments we could do that would validate it ... but it allows for FTL communications and thus breaks causality. It might be like string theory in that there are extensions which you can't rule out....

Bohmian Mechanics is a pretty garbage theory, it was literally invented as a way to show that the math could be done in a certain way. It's only been of interest to fringe theorists and in pop-science venues because it preserves notions of discrete/continuous space in exchange for violating causality. It might be easier to imagine time-travel than throw away human notions of physical reality, but the universe doesn't care about our puny brains : )

Right, well I can say the same about your intuition regarding causality and locality. Personally, I never thought that the locality requirement or speed of light limit was an absolute necessity. After all, you get spooky action at a distance either way. In fact, I would submit that avoiding pilot wave theory doesn't rid you of the possibility of non-locality. It might still be the case and now you have even more baggage.

Frankly, it's much easier for me to imagine a particle traveling along a pilot wave, and non-local events, then to imagine the world splitting into an infinite number of possibilities at every instance and everything being entangled with everything else in infinite to the third power set of possibilities.

By the way, is this an accurate overview? https://youtu.be/rbRVnC92sMs

> Personally, I never thought that the locality requirement or speed of light limit was an absolute necessity.

The only thing weirder than wave/particle dualism is violating causality.

> Frankly, it's much easier for me to imagine a particle traveling along a pilot wave, and non-local events...

Yup, that's why pilot wave theory is a hot topic in pop-science. You can totally imagine Marty McFly traveling back in time, that doesn't mean it would work.

I'm now entering tricky territory as I usually walk through this stuff with an actual physicist, but I don't have one one hand. So with the disclaimer that the following thought experiment is subtly incorrect [0].

We conceptualize things as either continuous or discrete. Take a solid block of matter and across the surface it is continuous. Break the matter in half and you now have two discrete units. If you keep breaking the matter in half, the conceptualization of a continuous/discrete has to break down.

So when you get rid of discrete space, is it really that weird when two "particles" seemingly "interact" when separated by ... space? Now trash the idea that time has discrete units and suddenly it's not that weird when particles "interact" across what we perceive to be discrete units of time.

The discrete/continuous paradigm is an artificial mental model that we impose based on our sensory inputs. That model is incorrect and you have to replace it with really complex metaphors derived from observing the behavior of these particles in experiments. From all the weird stuff we see in Quantum mechanics, we've never experienced FTL communication ... which is why Pilot Wave theory isn't taken seriously.

> then to imagine the world splitting into an infinite number of possibilities at every instance

The multiverse theory is a bullshit cop-out.

> By the way, is this an accurate overview? https://youtu.be/rbRVnC92sMs

Probably not, I stopped when they said that a particle travels all paths simultaneously. That's not correct, a particle does not exist in all places at the same time.

0: I'm pretty sure that the conceptual metaphors above still make use of global hidden variables....

Try physics.stackoverflow.com.

As a child I dreamt of having two futuristic walkie-talkies sharing several pairs of entangled particles making perfect instant communication across the galaxy possible. Later my physics teacher crushed my dreams by telling me it's far from as easy as that and probably impossible due to Heisenberg's ... (no clue what it's called in English). It's pretty awesome to hear that something similar is actively being worked on and seems possible, although being a huge challenge.

You can't communicate (FTL) with entangled particles unless fundamental assumptions in quantum mechanics are wrong.

Your dreams are still crushed. This research is about encrypting normal-speed communication. Instant communication is still impossible.

Heisenberg's uncertainty principle


Note that while particles can be entangled, the special/useful thing is the postponement of further entanglement (with the rest of the world).

Useful for what? Every time I see someone propose a use for QE someone else comes along and assures us it won't work.

The interesting part of "quantum communications" is the impact on security and surveillance. Does "quantum communications" even needs switches and routers? Backdoors don't seem to be possible...

Quantum communication works best if you have a direct connection, because the entanglement is easily destroyed if the particles interact with anything. This is the whole point, you can't intercept it without completely degrading the connection. If you try to redirect your particles through a network of switches and routers, be prepared for lots of dropped packets.

What if quantum communication is a backdoor? Wireless, undetectable bugs!

So if quantum entanglement does not enable FTL communication, what is it good for? How do you keep getting entangled particles from one place to another?

Could you use this for untraceable communication with hidden non-traceable submerged subs? That would be a military advantage.

No because they are only using entanglement to ensure no-one has intercepted the particle. This way you can send your symmetric keys and know they weren't intercepted. This is a real logistics issue for governments, who still physically move symmetric keys between locations.

As long as you have a fiber between them..

Thanks to your comment I am now daydreaming of a neutrino radio, with a design inspired from a Fusor!

we can use quantum tech to break the GFW firewall

Imagine being able to explore distant planets with robots and VR googles in real time!

Quantum entanglement obeys the speed limits.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact