Last year, in April, I wrote this blog about the Loophole Free Bell Test (before reading this one, better read that one first). I made a prediction I would have to write an exciting but also rather incomprehensible press release later that year. Well, I didn’t. Science is really tough when you are on the forefront of things, simply because nobody has done it before. You’re on your own figuring it out. Proving that Einstein was wrong about Quantum Mechanics is hard. And sometimes it even seems as if  Einstein’s spirit is actively trying to sabotage the experiment.

When I went to see researcher Bas Hensen to catch up, I met him at the service desk of the Faculty of Applied Science. He couldn’t enter his office because all the entry rights on his access card got mysteriously erased. Did I hear a faint chuckle in the background?

Bas is experiment leader for the Loophole Free Bell Test and works with a handful of colleagues on it, under supervision of professor Ronald Hanson. For their experiment, they need two ‘qubits’. Qubits are quantum bits, which can be either 1 or 0 just like normal bits, but also 1 and 0 at the same time (and in fact anything in between. Bring in the imaginary numbers). Qubits are pretty much beyond our imagination, but things get really interesting if you entangle two qubits: this means that if you look at one and you find it tol be ‘1’, automatically and instantaneously, the other qubit will become ‘0’, even if it is at the other end of the universe. Explaining that in ‘Einstein-physics’ would require some sort of faster-than-light communication, strictly forbidden in classic science where the speed of light is the ultimate maximum for both matter and information. It is bizarre, but so far (all) experimental results say it really happens the way quantum mechanics predicts.

During last century, many renowned scientists refused to accept that. For  decades, they accepted Einstein’s (et al) ‘Hidden Variable Theory‘: there must be some mechanism, so far unknown,  that presets the outcome. Due to some unknown science, one qubit already knows from the start it is going to be a 1 and the other a 0. No spooky science, no need for communications faster than light. Case closed.

However, in 1964, Scientist John Bell came up with a very clever experimental setup, the Bell Test. He proved that in a Bell Test, no form of unknown science can produce a correlation higher than 75%. But in Bell Tests, scientist measure correlations up to 85%, which only Quantum Mechanics can explain. (Although the debate is not settled yet, as I wrote in the previous blog: there are still loopholes open. Hence the fame of the prestigious Loophole Free Bell Test).

Back to Bas (who in the meantime has managed to use his charms to trick the secretary into giving him her card to regain access to his office).

So Bas and his colleagues need two qubits to entangle, over a distance of 1.3 km. The type of qubit he uses is pretty mindboggling in itself: a carbon atom that isn’t actually there. Bas builds qubits in tiny diamonds, 3×3 millimeter wide and 1 millimeter thick. (On a side note, for ladies on the Delft Campus: Bas is probably the only male bicycling regularly over the Delft University campus with a diamond in his backpack ). These diamonds are not natural, they are produced with a technique called CVD (Chemical Vapor Deposition). They are the purest diamonds on Earth, but still not perfect: in the carbon matrix, there are still a few nitrogen atoms present. Sometimes, next to such a nitrogen atom,there is a hole in the structure where a carbon atom should be. And that combination is actually what Bas needs: one nitrogen atom, and a neighboring missing carbon.

These holes, called ‘Nitrogen Vacancy centres’ or NV centres , can act as a trap for 5 or 6 electrons, which together can be used as a qubit. Shoot a certain laser pulse towards it, and the ‘NV centre’ will brightly flash a red photon back at you, like a lighthouse in the sea of carbon. The trapped electrons together have a ‘spin’, an electromagnetic property that can be measured. It can be 1 (up) or 0 (down). And it can be both at the same time.

However, making one working qubit from an NV-centre is hard.  And finding two to work together really is a hell of a job.

First: Bas and his colleagues need to get their hands on some of these little diamonds. As finding good candidates for qubits is hard, they need several diamonds to work with. They named the first pair they received ‘111-1’ and ‘111-2’, but soon they discovered that this was cumbersome in oral discussions. (Try saying out loud ‘ the qubit pair 111-2-17 / 111-2-13′ and you see why).  So the trio next coming in were named Frodo, Sam and Pippin, and the duo afterwards Hansel and Gretel.

Obliviously, a nitrogen atom is very small. And if you have found one hidden in a matrix of carbon atoms, finding BACK something so tiny in a relatively huge diamond is very time-consuming. So firstly, Bas’ colleagues Hannes and Machiel create a navigation grid over the diamond: a net of tiny gold dots, a miniscule grid-map on nanoscale, just to be able to navigate. Bas is looking for NV centres about  3 micrometer under the surface, with their spin axis oriented up/down. Although the diamond is very pure he will find normally about 20 somewhere around the middle of the diamond. All potential qubits for the experiment.

Next step: bring in the focused ion beam. It’s the chisel of every quantum physicist, allowing to craft form into the diamond on an incredibly small scale. Using the focused ion beam, they cut out the diamond, leaving half a sphere around the qubit, a little globe acting as a lens. This lens makes sure the single photon the qubit emits is captured with high probability, looks neat and uniform, and travels smoothly through the special fiber optics between the buildings. These photons from both ends have to meet up in the middle and create the entanglement.

Science is hard work. All created qubit candidates have to be tested one by one, to see if they are any good. Sadly, crafting little 3 micrometer domes is difficult (you can craft a dozen across one human hair, just to put things in perspective). When the shape is irregular, or when the NV centre isn’t spot on in the middle, it won’t work. Also, isotopes are trouble: 1% of the carbon in diamond is Carbon-13, an isotope of the common Carbon-12. And Carbon-13 has a ‘spin’ of its own, creating noise for the spin of the qubit itself. When there is Carbon-13 too close to the qubit candidate, it is rendered useless. Months later, Bas has himself about 10 good candidates of qubits for the experiment.

Job done?

No.

To be able to work, the two qubits he needs, must send out a photon on exactly the same wavelength. In fact, all NV qubits emit red light, but just ‘some’ red (600-650 nm) isn’t good enough: they need to have a wavelength that is the same at three digits: so for instance both 637.202 nanometer (and not 637,201 and 637.203 nm). He can tune them a little – about 0.1 nm), but not much. So he needs to find two candidates which are emitting ‘red’ photons with very similar wavelengths. In other words, he needs to find two Brandarissen, and not one Brandaris and one Lange Jaap (to stay on the lighthouse theme).

After three months of painstaking labour, Bas has two candidates working together properly: Hans-1 and Pippin-3. Which we should now rename ‘Alice’ and ‘Bob’, as is the consensus in quantum mechanics. So, a walk in the park from here?

No again.

In November, actual photons were flowing through the cables from both the Reactor Institute and the Faculty of Applied Science. But there was no entanglement. Nothing at all. Was Einstein right after all? It took them a month to figure out that the problem was in the lasers used to excite the qubits. These are extremely precise instruments, but their frequency is still fluctuating a tiny bit. In itself, a little fluctuation isn’t a big problem, but as they learned it IS if the lasers are not synchronized. So the team came up with a system to ‘couple’ the lasers to make sure they always worked at the same frequency at the same time. And YES: Just before Christmas, they managed to create entanglement over 1.3 km – which has never been done before. The loophole free bell test was ready to start! Hurray.

Bas went on a well-deserved Christmas holiday.

When he came back on 5 January, ‘Bob’ was dead.

Normally, when Bas arrives in the lab, the first thing he needs to do is to find back the qubit. The scale he works on is so ludicrously small that the laser he uses will not stay pointed on target overnight, so he needs to rescan a small area to find back the qubit. It takes a few minutes but isn’t very difficult, because the qubit , once excited by the laser, stands out like a lighthouse.

On 5 January, it didn’t. He scanned and rescanned, but no red photon came from the qubit. It took Bas a day to check all equipment to rule out any other possibilities and conclude: the qubit is dead. It’s gone. How? We will never know. Although unlikely, it could be an atom slid into the vacancy. Also, it could be possible that there are now not 5 or 6 electrons trapped, but 4 or 7, rendering the NV useless as qubit.

This is uncharted territory. They tried to ‘laser in’ an extra photons (or out), and ‘baking’ the diamond on 1300 degrees C to see if that would clear the qubit again. ‘Bob’ did not rise from his grave.

For Bas it is now back to the drawing board. He needs to find two other qubits that can work together. Luckily, he got much better at fine-tuning them, making it easier to find a pair that can work together.

Einstein still stands, metaphorically. Probably not for long. As Bas said, while wrestling with a door that wouldn’t allow him into his office: ‘as a scientist, when really absurd things start to happen, you know you are very close to a result’. I’ll keep you posted.

Michel