JQI Seminar October 28, 2019 - Mete Attature

JQI Seminar October 28, 2019 - Mete Attature

SUBTITLE'S INFO:

Language: English

Type: Robot

Number of phrases: 1415

Number of words: 9927

Number of symbols: 44639

DOWNLOAD SUBTITLES:

DOWNLOAD AUDIO AND VIDEO:

SUBTITLES:

Subtitles generated by robot
00:06
thank you very much I think last time I was here was four years ago it's always a pleasure to be here and also have the advantage a chance to explain that what we've been doing in the last 40 years so that's that's quite nice I hope you'll find it interesting as well I'll start with I will talk about is Gwen mentioned I'll talk about so States being photon quantum interfaces so basically the sole
00:42
estate versions of what people have been doing for many years on atomic physics we like to use those thank you for the tool boxes and then implement two physical systems that we think we can control and if we can't we that's new physics then we try to find out why and that's the whole storyline of this of this view before I get to the details of it I like to highlight that one question and I think this is this question and that is the concept of correlations so
01:12
in physics and you see why we're trying to do all these experiments it boils down to the concept of creating correlations would like to create quantum correlations between things alright say things the scientist obviously these two red blobs is the first thing that comes to mind I want to correlate two red blobs and what I find and please do let me know if there's there's there's a there's an answer to this question but what I find is that you can't get away from the concept of proximity when you're creating correlations somehow there is a characteristic length scale involved to
01:45
describe interactions and then when you describe the interactions you may or may not be able to create correlations between these say the internal degrees of freedom of these red blobs right so basically any kind of correlations you're building up in any physical system or any number of particles there is a characteristic proximity or length scale involved so proximity as a concept is fundamental correlations except one example that I know so please do let me know if you have other examples the one example I know that kind of motto motivates this idea of light
02:15
coupling is the fact that if I have this red hue around the red block that's the that signifies that it radiates it emits photons all right so if you can if these two red blobs can exchange energy directly then I'm no longer limited by proximity but how far this light or the photon can travel all right so think of it as let's do an example let's put two cavity mirrors here and here all right at that point if I want to create correlations between the two red blobs through white in mediating
02:47
this interaction then I'd no longer care about the exact distance between the two objects what I care about is that the two red blobs set at a good overlap with the mode defined by the light exactly thank you very much this was pre-planned it was and indeed the the oddity of
03:20
quantum physics shows up to give the the third option which is great and that is look they don't even need to interact directly with each other they don't need to sit anywhere a particular in fact the only thing we require is that emission for example is one one way of doing this light emission from each of those gets to in a third party that does a measurement it's a again an oddity of quantum physics you make a measurement to make sure you don't find out what the answer is but in return if you make sure that no one else does either so if you've
03:53
managed to do this there are protocols where actually you can realize correlations do this the universal sign of correlations between these two red blocks right that's very interesting now from a purely from a physics perspective this is nice because it completely removes this element of proximity is a requirement on the system it opens up the field from a practical perspective this is this lies it at the heart of four they say one of the key features of building a perhaps infeasible quantum network architecture where you don't
04:22
have to rely on one by one lining up quantum systems with individual axis the overhead of that you need a third party perhaps a puppet master if you will all all this required is that there's line of sight between the puppet master and the individual systems and they don't need to have line of sight so that's quite powerful let's say less overhead suggest the less overhead of control to create pairwise correlations in it all right so that's interesting and this is obviously what what we find
04:53
quite exciting as well and it goes back to a 1999 proposal it's very nice and the idea is and this is nice to give this talk here because it was a key place when it was realized but trapped ions the idea is that you try both of those red blobs now we get to see the inside of the red blobs because we want to talk about what is being correlated if the if we have a three level system and tonic system the spin projection can be a proportion or or dependent on photon being generated or not so by
05:24
having a probabilistic drive on both of those below saturation you can generate a photon and if you make sure that you detect the photon but not know where it came from the idea is this quantum duration realization and you end up creating correlations between these two probabilities of which one emitted the photon and therefore flipped its bit so spin up down and down up are two combinations that are not distinguishable alright so in realize this has been realized with trapped ions in that is as I said this is the place to talk about
05:54
it in the case of solar space systems we want to get to that experiment would like to do those kind of experiments to see if it is possible to treat a solar system in this three level it's relatively simple format and if it if it is possible can we actually how well does it perform how quickly can we generate entanglement how good is that quality of entanglement between the systems and so on so over the years for that kind of a goal people have been you know community has been working quite hard but these are the three areas I
06:26
would actually classify the SOL state quantum optics people you're either a spin physics person you've been doing a lot of spin physics to figure out how to how to control and and keep it decoupled from a whole bunch of noise or you're a photon person you're looking at equality of two single photons that come out of the system or maybe entangled photons and trying to characterize the purity the cleanliness of the of the optical systems or you are actually a cavity QED type person
06:57
looking into complex geometries cavities waveguides in some cases how to do any nice matching for one system for the next each of these require expertise and over the years if you look at soil steps quantum optics papers they almost always end with and therefore we achieve the amazing results in this channel and if we can combine it with the other two it'd be great wouldn't it you know you paved the way to and then you point that other people that done this and that and same thing here if you've done a nice spin physics you point and say but it's
07:30
relatively straightforward to combine the two advancements and proceed forward and that's actually finally we're about in that region now we're actually doing those right but it's only in the last few years that all of these have matured enough you now talk about coupling there all right there's one more thing I'll add here local register that is at the side of the spin if you want to define it as a qubit you do want additional spins or additional qubits to talk to for the next stage of what how useful this could be but I'll leave that as a
08:03
tag and now refer to it later on all right so when I say solid state what do I mean well I mean much more than this but at least this is a family of physical systems that spin photon links have been seen and are being studied actively some earlier than others and somewhere new comes today the type of dog is old friends in you this is the old friend corner dot and the new friend is 2d materials I'll tell you a bit about both I'll spend more time on quantum dots because I know more about
08:36
it about 2d materials we hardly know anything about it so it'll be very easy to explain how little we know but hopefully you'll appreciate that I won't talk about all these other stuff the nitrogen makes the center is actually the the Europe's choice of system to pursue core networks from the flagship projects this is a big massive project right now to realize the quantum network based on domestic defects in life in diamond but it's nitrogen based
09:06
and all I'm gonna say about this is look out for the next couple of years there's a lot of great work coming out for silicon in fact all of group for color centers and diamond with both spin and optical qualities significantly superior to the Russians a spin significant experience has been matching the quality of the so a lot is going to be done so keep an eye on this field in this particular topic but I'll start with talking about corn bats all right so
09:39
what do I mean by quantum dots these are semiconductors indium arsenide is a direct bandgap materials were going on here as well on this system it's basically a small island of about 20 nanometers to 3 nanometers little pancakes trapped inside a larger band gap material gallium arsenite all right so it's material based confinement for exit tones or electrons and pull separately that is our starting point and then we put them in electrical contact so we make devices out of them so we can control some of these
10:11
properties we have an endo player and electron gas if you will with a tunnel barrier and then I shot your window to be able to apply a voltage in black what this provides is a a comfortable region of around 200 millivolts 300 millivolts you can easily apply in the lab of charging plateaus so we can load up we can keep the charging state of the quantum dot as empty like you can see here the reservoir does not energetic it's not favored for an electron to hop in so nothing happens the con rod is empty this is a great set up to create
10:42
the energy levels needed for a cascade decay from a by x1 we generate entangled photons but if you just change your voltage a bit then you're in the regime where one and only one electron is captured in the quantum dot and then you have a different completely different energy level scheme that's something that looks like this where the ground state spin up and spin down are the spin projections of that extra excess electron the excited state is putting one more exit on into the or one exit on into the system addition to the electrons so that is determined by these two the closest
11:13
energy levels in the x-axis so it's a four level system with selection rules but these selection rules are also broken to give us than the skin as well right so it's our four level system that we work with everything I'll talk about today is not going to be on an empty column that but on a charge on so that's our ground state spin essentially the ground state meaningful represents if you want the qubit this is what it looks like so when I remember when I said you're either good at spin or photon or fabricating very
11:45
nice structures this is our fabrication this is a half ball length glued on a piece of semiconductor right that's as soon as you put it on top you're done that's your navigation but the dots themselves the quantum dots are so nice that we typically get around few million photons per second all right so there's a very nice system you look through this lens you see these titanium windows the very thin Schottky window controlling and if you zoom in you see the
12:14
individual dots and you that's our so I promise that it's a four level system to work with so even the photon generation part keeping in mind that we want to do this puppetmaster kind of experiment that was initially done with ions and we want to generate photons using the three level system that's the original idea the idea is you drive you prepare the system into the state you optically drive it and then photon is emitted having scattered quietly around one Photon and you also flip the spin at the same time so both on and spin flip comes at a
12:47
joint when you do it first the first thing first thing we need to do is check the purity of the photon generated this way and if you for that usually you do it a simple g2 type experiment for two detectors and check the probability that two photons are emitted from this otherwise system and it turns out it's relatively at this point I think we're more limited by dark rounds contributing to zero time delay than anything else so we're getting around 10 to minus 4 that
13:19
when swaras you move excess generation if you will so it's a very clean system to begin all right these are done it for Kelvin the wavelength I'm talking about here is around 969 the next thing you do is you take that set up and then you put one more Christ that next to it with an optical path and that is your again the magical element in Cana optics the beam splitter by combining the the light emission from both what deterministic so you guarantee that one light from each one photon from
13:49
each system arrives it in splitter you look at interference of those probabilities and again it's a absence of coincidences at time T equals zero kind of picture and then when you do this experiment you're at 93 percent visibility moment of visibility in those two systems now it's important to know two things one there is no cavity - to compensate for any spectral wanderings which is a typical sort of state problem
14:20
well actually it's a it's a typical any system problem if your your your ions are good enough if there's electrical noise you will see it right that's that's exactly what we're seeing our electrical noise is nearby to our excellent sensation same physics applies but basically broadening the transition by cavities helps erase this detrimental effect in this case there are no cavities so even without that we're at 93% the second thing that is interesting which is already a hint that we're going
14:51
to have some issues with the ground state coherence is that if I take this the wave packet of the photon and and and we sort of filter out the last two hundred picoseconds you remove it the time filter then the visibility goes up to ninety seven already just by removing two hundred picoseconds out of one nanosecond roughly all right and if you do the same thing to the beginning of the photon it actually goes below ninety three so you can tell that what's happening actually is the beginning of the wave packet is clean for both photons but then it kind of dies off and
15:22
quality goes away and this is a usually it's seen if the ground state the ground state coherence is not really strong so the long lasting and we'll see see what that the other thing I wanted to highlight here is that we have two separate Chrysler's in in the lab two separate samples two separate controls two separate quantum dots the usual leftover knowledge from many years ago is that all quantum dots are all separate they're different from each other so they you know you cannot get them to be
15:53
the same it takes us about roughly one hour if you have the right samples it takes us about one hour to find two matching quantum dots enterprise that so this electrical control is quite powerful in that sense and magnetic fine-tuning of the Zeeman splitting spur both is also possible that's right so
16:26
thank you when you do that what happens is we should flush that that allows you to characterize the noise profile on your indistinguishability is a function of time the delay time and what we see is that the beginning is very nice ninety-eight ninety-nine percent and then over some time scale which is again related to ground state coherence it actually drops down to around 90 ninety-three percent about 95 percent that's it but if you want to entangle two different spins you are ultimately
16:57
limited to obviously so so that system actually the experiment where you use two separate quantum dots is the same as using one dot experiment in terms of timescales doing it at infinity separation because there there's no correlation between the reservoirs that cause delays all right so I'm going to flash is like a sales pitch flash few experiments that have selected randomly this is obviously not the exhaustive list but some key experiments and using the quality of the photons quantum dots are able to provide ignoring everything else ignoring the spin part of the story
17:29
and in fact this for example this one the multiphoton was on something there's an archive paper now from again triumphs charring group we need to replace that this is ten photon boson sampling this is quite a competitive field right now one that's are indeed young systems where you can that kind of large states now that said we now take off my sales hat I put on my let's say admission to the problems of
18:02
the system the proper scientist hat and we want to let the photos I to the story we want to control the spin right the spin has to be also good quality if you want to do something with it well let's see what happens a number of experiments have tried doing a good old magnetic coupling of the ground space plating there are two challenges for this one is if your reading is read out as optical then you need to split this significantly and that corresponds to many gigahertz usually for magnetic fields where your your system is clean we're talking about 1020 ego Hertz
18:33
splitting so applying at 20 gigahertz microwave field is not easy but in addition when people did try to do something similar or lower fields the challenge is that they found the odium are line the the resonance line magnetic resonance line but it was kind of everything every time it was shifting to another point it was a bit random it was distributed but at each try it showed up a different place which makes it very difficult to say I'm driving the system on resonance or I'm driving system with some be tuning I don't know what I'm drawing the system because it could be anyway so magnetic coupling actually didn't
19:06
really work for combat systems so another approach again borrowed from from quantum optics was employed the idea is to far detuned a laser pulse laser strong pulse laser and actually create effectively this kind of coupling between the two spits this is great it's a broadband approach it doesn't depend on how far the tune you are essentially in terms of being able to do the rotation it relates to the intensity of the light so in three Pico seconds you're able to do a Rabi flip essentially this is spin Ravi curves extracted from optical readout once you
19:37
apply this pulse right so we do things optically in the sense you make use of the electric dipole strength of these systems and then you say okay well what is the ground state coherence and you do PI over two wait a bit PI over two the Ramsey experiment and you find out that your T to start your spin coherence it's crap Vega that's one point five or two nanoseconds this is terrible now do you understand why we had this issue with the filtering time filtering of the photon and we could see the effect because spin
20:07
coherence so the ground state coherence dives off on the order of two nanoseconds the photon itself is one nanosecond long so already by the time the photon is is finished we've already lost some coherence and that's that effect going from 99 to 93 or even lower all right so normally you could say well that's not a problem we can do dynamic decoupling right you can make and hence if this is some recoverable noise problem we should have much longer spin coherence and if you do spin echo indeed it does give us basically a limit of
20:37
about ten microseconds for quantum dots so we have electron spin coherence of about ten microseconds to work with it is modest but it is not nanoseconds anymore it's useful the problem is this if I want to store quantum information some sort of a quantum state within the spin then I do have that time available because I can do some operations if I want to generate a photon by flipping spin in a coherent manner as I was hoping then I don't have that control of the company I am stuck with this ticket so
21:09
the quality of the spin and photon entanglement in the first place I have to wait for that photon to be generated that takes about a nanosecond and I can't do anything to change that so basically I'm stuck so that is a fundamental problem even if you can prepare enhance your spin coherence later by decoupling techniques the initial step is going to get an over a coefficient that limits alright but nevertheless this to the experiment this is the coordinate system that we cry
21:41
it's now exactly the same experiment and basically we prepare the spins in the same state we do this week probabilistic drive and it mostly won't emit a photon neither of them will emit a photon sometimes the one of Google limited that's the subset we're interested in sometimes you get two photons being emitted and that's an error systematic error and our experiments and then you choose a basis to make a measurement and you see the spins are correlated with the photon that you see so it essentially becomes a three photon
22:12
correlation measure and this there is not this poor semiconductor quantum dots were roughly at that level 63% entanglement of spins enterprise of course that basis is better than X basis expects is the most vulnerable one for the coherence but the average gives us right now this is a what I think is a reasonable chart distribution of all physical systems and all they all reported experiments doing
22:43
exactly that measurement there are others that have done direct experiments like REM phase results and undress wharfs results on superconducting systems I'm excluding those I'm interested in the third option where you do a measurement independently let third party a measurement based entanglement and there are two schemes where you can do it in detail that the single proton scheme is the one I talked about which gives you better rate but it has a cost of fidelity if you if you're able to sacrifice your rate by taking one more proton to correct for systematic errors
23:14
then increase your and you see this distribution now this is Integra generation rate as a function or I should say Bell State the Delta is a function of n time generation right these are all these papers that have done this experiment quantum dots sit right there so that's three percent is so you see what's happening the this rate and time regeneration rate is directly related to the rate at which you can generate a photon how how strong
23:47
the how bright the source is and then how quickly can you prepare the system to emit again do you need to load do you need to cool down you need to do something else or is it all ready to go after the first photon being fired in case of Carmen as it is ready to go that's why it gives us an end it's bright with only 10 percent collection efficiency of the photons this is this it allows us to be all the way up here but the fidelity is low describe a bit about one thing we can do with this of course is you realize that this whole
24:18
thing is actually a massive interferometer in the lab this is a big prize that is another big cryostat this is actually a table this is one big table so these are two separate optical tables to Christ that's in a third table that hasn't affected so you're trying to stabilize their lap frequency face stabilize the lab and if you don't you don't see entangle of course because your face which directly maps into this one among other things is going to be completely washed out you see no correlations if you don't stabilize this pace if you do then this
24:49
error few degrees in our case is going to map directly as loss of a reduction of the visibility that we're trying to test so there's going to be a bit of loss coming from this technical and of course if you're able to control this face then you can also control where it's going to sit so you can actually choose what that face is going to be and it's a replica basically off of a previous experiment with ions where you can actually choose the measurement projects you to different states we can
25:20
essentially create or for Bell States as your starting point yes in our case it was intentional because if we would have to do would actually have to do a lot more building up to make sure that we have two independent paths running the experiment it's just easier we had the two cries that's it was just so much easier to have two fibers coming back to the same setup that we're interested in the measurement setup and having two separate cries destitute each one because we also use as I said magnetic
25:55
field tuning to make sure all right so there's so many exercise all right if we didn't have that you know crap nanoseconds t2 star but instead infinite wire is in the ground state this experiment would have been technically up here changing nothing else alright so we would have been up there if spin was better than it is and if the face stability was actually a hundred percent let's say you know ideal that we would have been up here so it would be one of those would be actually quite an attractive system if we had better
26:28
technical prowess in the lab and also infinite spin quartz two things that you can and and second a second issue is if we actually use current ongoing techniques this is my line at the end of the sent a per this paves the way to you know X&Y if we now take all our results and just add the collection efficiency is due to current cavity QED points we should be at Magnus which is nice because that's the rate at which you're
26:59
generating attainment faster than losing the entanglement so this is quite nice so let me talk talk through some of these basically these are the three things you need to basically improve your collection efficiency you need to improve the the spin quality and the third thing we need to do actually is remember that plus local register the extra spins we need to have extra spins so if costs are going to be interesting then you need to have some sort of a spin reservoir that you can control so
27:29
the first part as I said I won't do it but this is quite a advanced field by now we should expect about ten to thirty fold increase in our photo collection efficiency based on these results these are from work this is brightness versus how good photons are it's indistinguishable to measure if you were this is where quantum dots are tend to talk to my hours was down here somewhere all right and I'll highlight this one word quite a recent work a lot has been going on in the last few years on
28:01
coupling having open cavity structures where all this gated systems are present plus you have a cavity incorporated to it and Richard were burning in Basel achieved cooperatives 150 in ladies work and this claim the claim is that the 99.7% of the protons are coupling into a mode that they collect so this is quite promising that you can actually work with such systems we don't want to work in a strong coupling but we want to be
28:32
able to extract the photon salts so instead of over 10% this is this is quite attractive alright I'll talk about number two that's the main thing I can do because I I do less photons more more spin and the source of this error and my promise now is to give you infinite spin choirs all right that's what I'll promise so the idea is that you have quantum dot actually has around 50,000 nuclear spins each one of these has a spin indium gallium and arsenic all have
29:02
spins and also isotopes so it's actually quite a noisy place for the electron to be in the electron wave function is as big as 50 of these of 50,000 of these guys so electron is in contact not even dipolar interaction contact interaction with 50000 slowly wandering nuclear spins which are at thermal equilibrium at 4 Kelvin so so that's a problem and the fact that this contact interaction means that this hyperfine interaction is very strong actually not weak background noise but
29:32
quite a strong coupling and these fluctuations are seen by the electron as overhauser field and the electron spin itself is seen as my cue this very interesting physics problems that were so say about a decade ago quantum that's a understand what's happening now I don't care in this work so far I don't care about the interesting physics of nuclear spins I just want them out out of there I don't want them right so on this I go to silicon type systems where there is no spins the only thing I can do trying to salvage
30:03
here is to reduce their fluctuations okay one way to do that is to polarize them so if we polarize all the nuclear spins fluctuations will go away it's the fluctuations time-dependent fluctuations slow wanderings that caused the problem good and people did that so they can do it by dynamic nuclear spin polarization they can optically drive the electron which leads to this coupling polarizing nuclear spins and H if 70 percent you could spin polarization which is amazing right you have this 50 thousand nuclear spin spins ensemble with 70 percent
30:34
polarization it turns out it's also useless because fluctuations due to seventy thousand do to 70 percent polarization is roughly the same as not only when you get to 95 percent or so fluctuations are reduced significantly that you actually see the effect so it's great but useless and in turn if you do this polarization you end up with something like 5 to 7 Tesla extra magnetic field due to this overhauser field so the electron is split as if
31:05
you're operating at 7 Tesla magnetic field but you certainly don't need it makes the so it actually is the is the worst idea but there's there's an option B and that's what I want to talk about today all right so to understand where I'm gonna go with this my main goal is not to polarize them but to quiet them down alright think of it as a Zamboni I want to I want a is a moaning machine to quiet down the fluctuations on the so the idea is until now spin control we're here control was not optically yes but
31:36
it was done with a broadband field was a ultrashort pulse and it didn't actually rely on particular Zeeman splitting of these guys any Zeeman splitting would give us the same kind of rotation we wanted to change that because we want to have access to this this internal structure so if I represent the spin-up and spin-down instead of saying just blurry lines I actually should write it in this format that you have the electron spin down manifold and then it just been up manifold they determined including multiple lines each
32:07
corresponding to a single nuclear spin out of the right that's what the actual levels should look like and of course that thermal equilibrium you know these are occupied hundreds of these levels are occupied because nuclear spins are spread over so what we've done is we've used the same concept but with narrowband lasers and we mix microwaves with narrowband lasers effectively giving us direct coupling so this transition so we can actually probe where that spin resonances
32:37
but still using it optically so it actually works as fast so if I take the that energy level and rotate it it kind of hours our system now looks like this the electron pair system looks like so this is the electron up manifold is the electron down manifold again and each of these is one nuclear spin flip the Zeeman energy due to seen by the electron due to one nuclear spin flip absolutely absolutely do this extends that way this excess that way there should be about $50,000 and it's the state when I say where is my state it
33:10
actually is spread over a few hundred of these now on top of this this what I call here Omega is the two photon detuning that we select that is resonant with one particular net nuclear spin polarization whatever it might be let's say zero all right that means this is in this picture I go back here this is that splitting and I'm resonant with a particular I set all right but that means that one nuclear spin flip I'm no longer resin now this is slightly different than harmonic systems
33:41
because I set is the field that the electron will feel due to nuclear spin essentially there's a magnetic field over also field so if I change I set the net feel the electron sees and therefore it just Azima splitting is based on that so I actually we this is actually an honor harmonic system in the sense that each I said will have a corresponding splitting for the electron that we can we can target so in a way this does not drive all of them so I cannot do a
34:12
nuclear spin not flipping direct electron spin flipping transition regardless of I said I can only do for a given license all right are you with me on this this is key because that gives us the unharmed that we'll be using later but due to the quadrupolar coupling there's presents quantum knots in the presence of strain so this is not isolated quantum dots there aren't the strong strain strain allows that transition to be available normally wouldn't be but allows that to be possible because this is possible now
34:43
by actually driving this you're also driving a bit of that so that is all Fela okay so you can flip a nuclear spin that way however this is also allowed because you're also populating the system but this is stronger because it is closer to resonance than the other one the anomalous he kicks in here and the surprising thing is if you do it on the other side you see that they're both pushing the nuclear systems towards where your laser is I set in this case is purely determined by you you set your two proton be tuning in the lab for the
35:14
laser and then let the cyst just drive the system for a while as you do this more and more of these guys getting pushed this way and more of these guys getting pushed that way so you're in a way do you see that you're sweeping the thermal distribution that you had to a narrower distribution by optically Drive but not at the extremes but not at the extremes of the so we're not being
35:44
that's right extends out that's right exactly the key point is the the ratio of the two arrows the two directions if you are able to sustain the fact that this is stronger than this rate when you go far out there as well that's actually crucial you still need to sustain that so you can bring all those outliers in as well otherwise you end up reducing a
36:23
narrow band and then the rest stays the same you're absolutely right so it does seem like it is actually drawing in significant portion of this and I'll show you so essentially there's a feedback mechanism to the thermal equilibrium you end up with a non thermal distribution that's right so that ratio the the the difference between the two directions is getting less and less and less an advantage is that some so you have finite range over which you can match I can spoil it basically our simulation suggests that we're cooling thirty five thousand as opposed to fifty thousand
36:57
roughly that we expect so that's and that works it's also trying to exactly that's coming next so you didn't realize that we organized this talk together actually anyway so the key point I want to highlight in this this kind of drive
37:31
is the spectral selectivity that's what we're gonna rely so essentially I'm gonna present optical electron spin resonance that's the power of this technique that allows us to probe in here what actually happens in these levels and yeah I can we probe so when we do this experiment so essentially the idea is if we are genuinely cooling the nuclei into a lower state then a smaller distribution then this is kind of what it's going to look like the spread is less and indeed we see this so this is
38:02
the ESR optically oh sorry experiment two photon detuning that is swept over time with very low power so they were not we're trying not to disturb the system were only probing what the distribution looks like and indeed this is what electron spin resonance is when you do a pooja bit preparing and when you do it well then it looks much nervous so this is exactly the along the lines that we we were thinking and now what it means is that the this line would suggest now fifteen megahertz married and I will not say that this is the temperature of the system because not a thermal equilibrium here for the
38:32
temperature but if I were to quarter temperature it will correspond to with that's on the order of five on the micro gap right so but I won't call it that exactly it's not it's not thermalized into this temperature and then I can quote it but if I were to take the width of the distribution that we're saying without looking at the actual physical functional form of it then it looks like you would correspond to a normalized at 500 just to get an idea of the fluctuations all right so then the whole point was to give you infinite spin
39:15
coherence let's do that this is ramsey and of course we get the the few nanosecond spin coherence that is the problem now if I include this preparation stage I talked about to quiet down the nuclear spins and then as soon as I do that we probe to see what T 2 star were able to get for the electron and that becomes around 49 seconds so the electron is now is living without any protection is living in this quieted spin system that is long term that's the Zamboni I'm
39:47
talking about you do the Zamboni and then the ice is nice and smooth for a while now given that I'm only interested in T to start because of the photon generation and given that it takes about a nanosecond to generate the photon infinity and 39 and seconds are the same their effects are the same so there you go I promised you infinite infinite means there's no point in freezing this more we won't win or the quality of the spin for the interface that's already
40:16
enough but that comes this question so this funny non thermal distribution how long does it last how long can I preserve this t2 started I was talking about it turns out it's actually quite reasonably long it's about 50 milliseconds before t2 start drops back to its original value or the one over U values is 50 milliseconds so this funny nuclear spin profile actively leaves quite a long time so indeed I can prepare nuclear spins and then if I can address other parts of it it's perhaps possible the milliseconds is long and
40:49
then but there's one other feature so if I look at this I did to tell you this you know this width is 15 megahertz but it's single nuclear Zeeman flip this difference is 25 minutes so actually we didn't just quite down a bit we are in the regime where I we should be able to resolve this transition and that transition the first time and in second fact this is the five end result is you so we should be able to see the side vent that we didn't weren't anticipating so then what we do is we coherently
41:22
drive the system and then we drive and then sweep the s our line to see you what lit up which states lit up and indeed we do start seeing some features coming up and if I do a line cut and this is what the uncle system looks like and this is what the cool system when we drive for a while that looks like so you're able to see the original SR line you see the first one this is one nuclear spin flip out of fifty thousand two nuclear spin flips and right now normally there's a difference between
41:52
just driving the electron or driving the electron and and the election and the nuclear spin flip at the same time as a second order process normally these guys should be all the way down here if I am flipping one electron or one nuclear spin right second order process it's very unlikely so it should be down here so but the difference is I'm not actually exciting one is then one person stadium is not going and waving what we have actually is a spin wave so the
42:24
whole stadium takes part collectively so we start one one way of going around the stadium that's this one a second way going around this video independent and that's this one so it's a collectively enhanced process that's why we don't see there's a tiny blip down here even though a second order we actually see it my stroke if we check that the factor that we expect so what it means is that we're able to start these spin waves
42:56
inside a little corner we the whole point about quantum dot work for many years was to make it more and more and more as simple as you can imagine the four levels or the two levels whatever you want it to do and isolate from everything else but there's more we managed to zoom in we we managed to open up to all of a sudden access coherent access to fifty thousand or twenty five thousand thirty five thousand nuclear spins so what we want to do next is drive this transition prepared the system into this few state
43:27
distribution we're not able to do it only on single one yet it's still distributed a bit but once we get there we want to coherently drive one of the side bands and see if we see some coherent exchange between the electron and the and so we chose this one to be far from these guys because we were I'm hoping to be driving everything actually nothing is very cleverly but when we do that we do indeed see signs of taurine oscillations the first signatures of
43:57
exchange going on humble and single electronic you're talking about the difference between yes it's interesting actually it's partly time data acquisition time so part of it is technical it's more a reference point a reference measurement have been the same quality to be able to resolve this partly the other part is you are actually driving so there's even though I'm operating at low power to make sure there were only probing system
44:45
we're not only probing system we're still cooling so it's the same mechanism so by we have to do this quickly to avoid actually deforming the profile but otherwise there's no reason for the visa I think their their effect is probably the same the number of atoms that take part in creating this width is the same number of atoms contributing to this is very probably less but not not much more than not more than factor two all right
45:25
so the latest from the lab now is when we look closely into this remember this is a composite system right these are not 35,000 nuclear spin 1/2 particles this is we have Indian with 9 half we have gallium and arsenic with 3 halves and they're isotopes as well so it's a pretty messy system so first of all it's amazing that we can actually resolve it in first place it should go blurred out but if we zoom in and do a bit of careful measurement then we're we're starting to see now more systematic this is really hot from the oven so I'm not gonna say too much about it until we're
45:58
confident but it looks like this we can repeat these measurements many times and identify these two particular species so in looking to the subset of species in this hello looking alright so basically the directions would like to go but this is twofold one is the advantage of having this nearby deterministic spin so not deterministic probabilistic the
46:28
spin where you die put do dipolar coupling is that it's there as long as you find it from say in these centers there are a few nuclear spins around you find where they are and they become your resource in the case of say rare earth or ensemble the quantum memories you have to convert your information from your spin to a photon and photon gets absorbed into spin waves it's very similar concept is what we have here directly but then there's this photon loss and photon absorption rate coefficients coming in this gives us a
46:59
chance if we can manage to get the quality to be reasonable it gives us a chance to actually combine the nice size of both the collectiveness of an ensemble absorbing or storing the information with the the nature of control to each one and journalistic aspect of it without an interface direct coupling without interface and of course you can do a whole bunch of other stuff but we need to get to that this is more a motivation for us then then the next steps that we're doing right now i'll pestis there are many questions that we need to answer our passes for now for the it's time what I'll do is for the
47:30
next it's something like less than 10 minutes right 5 minutes are you I'm gonna switch completely now all that is fine I'm gonna switch to the topic I don't know as I promised it's not gonna take long but it's entertaining it's these 2d materials alright and there you know Glan get correct me but I think with 3 5 for example you work very hard to produce very good quality samples and hetero structures all that and then and
48:02
then you go home and then you dream of this you dream of being able to make atomic layer precision of any arbitrary material on top of the next one if you want to or any reason of any property from this this chart that this table that you have you can have super conducting layer plus the insulating layer plus a semiconductor layer plus the semi metal a magnetic one in a non-magnetic one any combination you want to create both actual working hetero structures for a particular purpose or emergent new structures new
48:34
and new features through the yellowing that comes from this kind of control so there's a lot too not too slow in this in this whole business there's so many opportunities but also there are lot of technical challenges in material quality in all these but nevertheless it's it's a very hot topic and there's this one issue of nature nanotechnology that basically had four papers back-to-back talking about confined emission from these laying materials these transition metal like alkaloids so this is monolayer basically
49:05
it's a three atom atom complex WSC - thanks by Selma for example and these four papers plus Rudy's paper and opteka oh five papers and they're all reported more less the same thing saying there are local spots that emit light that shows an T bunching so these are some confinement centers and no one knows what they are except I mean you can see in the title that they're suffering it's suffering to label what it is so it's vague and we didn't do this first but I
49:39
was attached as postdoc Nick was my postdoc cherian was my PhD student I had nothing to do with mark but you know the paper is great so we said okay well we'll look into it too because an interesting material and so the way worked was we when the student who was very good at fabrication and just could not see these quantum dots these these localized emissions and we another student was let's say efficient at the time in and out of the fabrication lab
50:09
very quickly producing samples and miraculously having many of these limiters so the two people doing more or less the same recipe the recipe is borrowed we are foreign to the whole field so we borrowed the recipe and we executed it so two people made very different end up with very different results turns out the person that went in and out very quickly it was a bit sloppy and in terms of removing all the fabrication errors and and leftover junk essentially alright and when we looked at the at the AFM we could actually see
50:40
these these bombs pendel bombs left over under the otherwise flat material so we said why don't we make it so that the other student does the fabrication from now on so it's clean but we introduced the bombs ourselves deterministic so this in collaboration with Marco in Harvard Marco launcher basically this is silicon nano pillars we edged out nano pillars and then you put the material on top instead of trying to make it flat you actually create like teepee cats periodic TB cats so the idea is if local
51:12
bending or strain is the cause of confinement potential emerging that captures X bones then we should see the top of these pillars line up as quantum sources if it works right that was more gut feeling and this is what it looks like these are this one nano filler this is a material on top spreading and we're looking for features here what you see is that it's actually stretched down to the to the base for hold on so it takes to be a stretch material actually alright and this is this is a mono layer and a bilayer of this material in
51:42
microscopy so you see where these these Peaks are and then if you I mean if you look at the top of these pillars yet to the see quantum emission narrow lines on each one this show on indeed each of each one of these pillars does generate a quantum imager I still don't know what it is but basically it suggests that it's not defect based but rather confinement based due to local space and what is nice is that basically we have about 96 97 percent yield this is even better now so every pillar more or less
52:15
has a quantum imager so you can I can guarantee you economy which is to be present and in fact half of those will have only one text and pillar height reduces the spectral one so these are for the first time actually better than naturally found one so this is like a nice in this experiment is like four or five microns just to make sure that's closest so you
52:44
have a modulating potential there's one paper if you're interested in that talks about how confinement potential can rise over a small spatial by bad mixing I won't go into that but I'll highlight two two experiments very quickly one is you can functionalize the pillory it doesn't have to be just silicon oxide if you put diamond you can actually talk about perhaps a coupling spins to each other or you can talk about not one layer but multi layers and then you can
53:16
talk about indirect X right and that's the one I'll very briefly show the latest results so very excited about it I'll pass these oh this is what the Nano diamond by the way this is to show that if you just put we don't even make pillars you just sprinkle some nano diamonds on the surface and put your layer materials on top of every nano environment you get a quantum scale nitride same thing I'll highlight this work function up one strap and this is if there's anyone here was we had to work with plasmonics and had to put a quantum emitter in the hot spot of a
53:48
plasmonics structure your paper most likely ended up worth saying we observe this much coupling it could have been that but positioning that the emitter at the two nanometer window is very difficult and therefore we didn't get it so I probably ended up like that this is one example where this goes away Stephon actually decided to put gold nanotubes on two layered materials otherwise flat there is no emitter or anything but as soon as you put gold it actually creates tension or stress I should say at the four corners and then
54:18
it actually produces a quantum light emission from there they become the confinement potentials this is irreversible if you remove course the cube it goes away so I don't know any other quantity that depends on you you push a button your finger essentially you create a localization and get quantum light out if you move your finger it will follow you it's aligned if you remove your finger it goes away so I don't know any other example of this anyway she is able to see 46 million photons per second from this main material as auntie bunch anyway
54:52
what I'll show is one final thing and that is what we want in these systems is actually long lived next door these are typically living on the order of a nanosecond I actually this is quite relevant to the dipolar gas work as well can we actually have EXA tones in layered materials or in any solid state system that actually has much longer interaction than the simple single particle picture and the idea for solar state systems in realizing this is a polar molecule concept is to create the X or not in the same specially confined
55:23
location but separate the electron in the hole so the simplest thing you can do with a layer multi-layer system so the whole sets here electron sits there for example and if you look at the the band valence and conduction bands of these layered materials we pick WS 2 and W ec2 and I'm looking at that transition from here to here as the lowest energy system possible as the track so I'm looking at that system where the electoral sits in one material no es 2 all citizen wc2 and
55:55
there's a permanent dipole built in place permanent that pole means that we'll have a very long lifetime and now I talk about how these dipoles will interact with each other so this is our structure two of those overlapping together bottom line is this picture we're having we're seeing like that this didn't show because of Mac this is my microseconds what we're able to see is actually it's unfortunate but it's a few microseconds of lifetime for the exit
56:27
box this is the longest we've ever seen in these materials of excellence dipolar expose living for a long time this is on pillars again so every pillar has these X terms and they're confined they're long lived but what's interesting is that if you go away from the pillars even just the substrate flat region you're getting dipolar gas essentially we see that they're attracting in a possible way to each other and with a lifetimes of so it gives us a whole new range of time scale to probe interacting
56:58
expose in the absence of any polarity on picture or cavities it all right I'll pass this this is implicit the quantum dots in these systems as well just finish off by saying thank you very much for your attention and you join us if you're interested in [Applause]

DOWNLOAD SUBTITLES: