STORY OF THE MONTH

All I want for Christmas is QKD

Dec 2024

Alessandro Marcomini

Dear Santa,

Finally my favourite time of the year has come. I love the magic of Christmas, the smell of hot chocolate while spending time with my family, the lights, the snow and of course all the presents!

This year I believe I behaved very well, and I am really looking forward to unbox all the wonderful presents you are going to leave under our tree. As usual, I am writing you my wishes, and I really hope you can find and bring all these toys to me!!

In the past months, I have been very scared. The amount of crimes in the web has increased a lot, and new quantum computers have been announced, meaning that soon I will probably not be able anymore to buy things online, to use social media or to even chat with my friends or use the internet without bad people being around the corner, reading all our messages and personal data. Therefore, for this Christmas, I strongly wish o be able to communicate safely with my family and all my friends. About this, I recently read online an interview by a funny PhD student which taught me how the solution might be given by a new fantastic technology, which is called quantum key distribution (QKD)! I was very curious about solutions to make the internet quantum-safe, so I subscribed to the QSI newsletter and I made up my mind: this year, for Christmas, all I want is the material to build my own QKD setup!!!

Unfortunately I am still very young, so I do not know all the technicalities about QKD is actually built… since you are magic I am sure you will find the best possible devices to give me the perfect QKD setup, but in any case I will tell you a bit of what I could learn. I am a very curious little researcher myself!

So, dear Santa, there are two main ways of doing QKD. The first one is called “discrete-variable QKD”. It is done by employing single particles of light (we call them “photons”) to encode the key values. Physicists typically choose one feature of such particles and assign labels to the different cases that can be observed. For example, if the chosen feature is polarisation, they could say that the horizontal polarisation in called “0”, and the vertical polarisation is called “1”. Another example could be the feature being the time bin: people agree on a very specific time in which the particle should arrive at destination, and then if it arrives slightly early they label the particle “0”, while if it is a little late they label it “1”. These features that can be labelled “0” or “1” are called “degrees of freedom” and they are the secret carrier of information that nobody can hack! I am pretty sure polarisation and time-of-arrival are the most popular choices for encoding, although I heard someone talking about “phase-encoding” as well. Unfortunately, when I tried to ask what is the “phase” and how we could observe it, all quantum scientists became quite thoughtful and they spent the rest of the day discussing among themselves, so I did not get a clear answer…

Anyways, apparently to adopt discrete-variable QKD the tricky part is to produce the single photons. Scientists so far have been using atoms that naturally emit single photons of a specific wavelength (they call these “quantum-dots”), but they are very difficult to control and they emit only so many photons per second. Alternatively, people often use lasers emitting very little light: these are much more precise in the rate of emission, but they are not guaranteed to emit only a single photon per time.

The second approach is called “continuous-variable QKD”. Its name is due to the fact that instead of single photons, here scientists can use a stronger beam of light to encode information. The way this is done is very complicated, but they told me that security is guaranteed by a quantum thing called “Heisenberg uncertainty principle”… this is very cool, as it basically means that in the quantum world everybody is always at least a tiny bit ignorant! I feel much better now when I don’t know everything at school… It seems that when using continuous-variable QKD, we could adopt standard devices like normal lasers. This is much cheaper and reliable! However, extracting information in this case is much more complex, and therefore continuous-variable QKD is only performing better when used for communication over short distances. Hence, my first wish is a reliable and fast single-photon gun to be used as a source for discrete-variable QKD, so I will be able to send you secret Christmas letters all the way to the north pole!

When it comes to protocols, there are so many out there. Most of them follow a scheme that people call “prepare-and-measure”. This means that to share the key one person has to prepare and send the light, while the other has to measure it. This requires very efficient detectors. Unfortunately, detectors happen to be Achille’s heel of QKD, due to the fact that they receive, by definition, all the light sent through the channel, which makes them more vulnerable. Moreover, the performances of single-photon detectors are not so good. As a result, clever scientists have been proposing new solutions which are not relying on the measurement devices, which can be left in the hands of the bad people. This scheme is called “measurement-device-independent” QKD, but I like it better “MDI-QKD”. It comes in different styles, and the most promising one is named “twin-field”, I like that name as well! In any case, these MDI-QKD schemes have a great advantage: they do not require to have perfect detectors, and actually the devices themselves can be completely uncharacterised and unsafe! Still, it comes with a cost, as the performances for this schemes are, in general, worse than for prepare-and-measure schemes. Therefore, since you are magic, I want you to bring me the best, perfect, super-performant, 100% efficiency, perfectly secure, loophole-free, single-photon detector ever!!! Or even better, bring please two of them, but try to make them very very identical, otherwise it could be a problem!