 move my mask. Yeah, so we are at the, it's Olivier Pousseau, and he will tell us about fast, high fidelity cupid readouts using cross-curve coupling and CQED. Yeah, so first of all, I would like to thank the organizer to let me have the opportunity to present this experimental talk. And this work has been performed in Grenoble in our group, and it is mainly the work of two PhD students that I am going to present to you today, the work of Rémi Dassonville and Vladimir Milchalkov. And this work has been made in a strong collaboration also on the theoretical part with Thomas Ramos and Juan Jose Garcia-Ripol. But before starting on this subject, I would like only to flash you an interesting setup that we are starting to work, which is a completely different topic, but it's only one slide that I would like to show you. It's a, yes. So we are starting also to work on superconducting hybrid nanowire system based on germanium. And what it is interesting in my point of view in this system that we are starting in Grenoble is that we are able to work on a sample, sorry, you see that it's in completely monocrystalline structures where the aluminium, I don't see if you see here the atomic plane. So this is a transmission electronic microscope. So you see the atomic plane of aluminium. You see here the atom of germanium. And you see also the interface which is completely flat here. And this is along, this is the direction of nanowires. And in this kind of system, it was the thesis of Jovian de La Force. We have performed some preliminary experiments both in this germanium silicon core shell where the physics were more driven by quantum conductance in nanowires, but also in a germanium, interesting germanium system where we can control here the dot in some sense by up to 20 nanometers. And here we see some transition from a single old quantum dot system to a Josephson effect when we change the gate. So there is a lot of interesting things to do and I would like only to mention this topic that I find very nice and rich in terms of physics. But today I would like to concentrate only on this quantum measurement system. And what I would like to present you is that typically this is a scenario to perform measurements. So you have an artificial atoms or an atoms and usually you are going to measure the quantum state of the system by applying some microwave photon inside the cavity and by measuring the transmission you would get information about the quantum states. But of course, this physics of measurements is driven by the microscopic coupling between the artificial atom and the cavity. And this is the interesting part that I would like to discuss more in detail today with you. So the standard coupling I would be fast in this part is what is called the transverse coupling and this comes from an electrical interaction between the atoms and the cavity. And this produce the transverse coupling and the well-known genes coming in the Hamiltonians. And in this limit, when the qubit frequency is far detuned to the resonator you can make a strong approximation which is a dispersive limit which is called dispersive limit. And in this case, the Hamiltonians start to be very nice for the readout because you have this approximate cross-care coupling term which are going to change the frequency of the cavity depending on the qubit state. So this is the way that the people are measuring qubit in most of the superconducting qubit states. Or whether this is due to an approximation and this produce lots of drawbacks in some sense in these measurements. First of all, the coupling strength of this cross-care is an optimization between a far detuning but not too far because if not, this parameter will be too weak. Also an important thing that people do not insist enough. Of course, here the qubit which is described in this Hamiltonian is no more the initial qubit. It's the qubit which is dressed by the photon states of the cavity. And this means that your qubit here because of the stress of the photon states on the cavity have some relaxation channel which produce the so-called the partial effect of relaxation. And last point is this approximation is only valid when the photon number inside the resonator is small and this is typically a limit that we don't want to do when we perform measurements because we want to have a large signal. So for these different reasons, we were thinking to find an alternative coupling which changed on the majority. So typically all the people are working on this coupling that I showed you before but since 10 years typically there was some suggestion to use longitudinal coupling instead of a transverse and this was a proposal of a group of for example, Alexander Blay in Sherbrooke where they propose this kind of coupling and you see now the coupling is directly the coupling of electric field of the cavity with sigma z of a qubit. And in a group what we have proposed is to have a coupling which will be a cross-care non-perturbative cross-care so it's not the same origin that what I just speak before is directly related to a specific microscopic design of a circuit which produce this kind of coupling. So I am going to speak about this kind of this last coupling and I will show you the experimental results that we got in this coupling. So to obtain this coupling we use a transmole molecule and the transmole molecule is simply two transmons which are inductively coupled. So when you do that you can easily demonstrate that the quantum dynamics in this simple circuit is given by two modes, a symmetric mode which will give you exactly the description of the transmole qubits and there is an anti-symmetric mode which are going to be used for the readout. And what it is very important here is it's here which appear the coupling is that the coupling between these two modes so the qubit and the readout is given by this product of coarsiness. And if you do the expansion for small phase in this case you see that the coupling is given by this product terms which give you directly cross-care coupling which is different origin that the dispersive cross-care coupling. And the interest of this cross-care coupling is the amplitude is no more dependent on problem of detuning is related to the molecule parameters. And for example you could have very strong coupling in these terms. And also now this derivation show that you are no more qubit which is modified dressed in some sense by the photon of a cavity. So it will have also impact on the relaxation process. So of course here I show you internal modes of a molecule so we could not really apply the measurements. To do that we need to couple this internal mode of a readout to an external mode. And to realize that we introduce our molecule inside the cavity and we make an alignment in order to not break the symmetry. And in order that the electric field of the cavity will be transverse to the polarization of a qubit. By this way we avoid any transverse coupling between the qubit and the cavity. And it will be aligned to the electric field of a readout mode. So by this way we have a strong coupling. So this hybridization between the readout mode and the cavity will authorize to perform now really measurements because of this transverse coupling. So in the following we are going to see what will be the property. And this is in some sense our motivation in Grenoble is to understand this new kind of aminternials which describe a new kind of measurement system which is different by the Genscoming style aminternials. So to perform the experiments we need to inject microwave inside the cavity and we need to amplify this microwave which is usually low signal. And for that we use a Josephson parametric amplifier that has been developed in Grenoble in the group with Nicola and Nukaplana. And so we are going to use this amplifier to perform our measurements. And typically we are going to apply the readout frequency which is in resonance to the cavity resonator when the qubit is in the excited state. So typically we will have a large signal when the excited state, the qubit is in the excited state and low signal when the qubit is in the grand state. So if you do measurements in this case you obtain in the IQ plane the response of a microwave measurement when the qubit is ever in G or E states but to be more precise we perform histogram here where we prepare the qubit ever in the grand state so it will be the blue curves. And in this case you see that we have a Gaussian the main Gaussian is centered close to zero so it corresponds to low microwave amplitude signal the width is related to the noise of the amplification chain and we have some small excitations of the qubit showing the second Gaussian. And of course we can prepare the qubit in the excited state and in this case we have a main Gaussian which is at higher signal because it's the condition of our measurements and we have higher Gaussian showing the relaxation between the excited state to the grand state. So if we make the analysis of the error we have a fidelity on this first generation readout of about 97% which was quite good and the main error was coming from the relaxation to the excited state to the grand state. So it was the end of the thesis of Rimi Dassolville and so Vladimir Mishalkov decided to optimize the circuit parameter keeping the same idea but changing completely the sample and he developed this new geometry which was inspired by the group of Alexi Ostinov and he used this cylindrical symmetry in some sense to reduce the residual transverse coupling between the qubit and the cavity. So in these new generations you can see that again we can have this histogram analysis but you see for example from the grand states the grand states now is many a single Gaussian and you have only a residue of error here and typically it means that it's important to say that here in this measurement we perform last in the previous one pre-selection measurements. So we remove, we do two measurements the first one to see what is the state of the qubit and so for example from the grand states we remove all the measurements we have shown that the grand state is in fact not in the grand state but in the excited states. So it's a way to prepare or to select only the experiments where the qubit is in the grand state. So when we do that we have an error which is very small typically 0.1% of error of preparation on the grand states and when we do the excited experiments so we have this second Gaussian which is a lower second Gaussian because the relaxation time on this qubit was much more improved it was typically 20 microseconds in this second sample and so in this case we got fidelity of readout which is close to the state of the art of readout fidelity and also we were interested to probe the quantum non-demolition performance of a readout. So to do that Vladimir has performed two successive readout measurements so the first one is to initiate the qubit in the good states so this is more for the preparation but he performed two readout, two successive readout and he plot the results of the readout in this direction and for the second readout it is in this direction so in some simple analysis if the point are in the diagonal it means that the two measurements are given the same results so the measurement is not destroying the states and if you have off-diagonal it means that you are starting to produce some destruction your measurements start to have some destruction on the qubit state so you see when the qubit is prepared on the grand state then all the points are along the diagonal square and when we are in the excited states again it's mainly on the diagonal but you see here that there is much more points indicating that in this case that there is a small scenario that the qubit are prepared in the excited states and before the second readout measurement he has time to relax but nevertheless in this case we get a typically a very good quantum emission performance with close to 99% we can do also continuous measurements now and this was on the first generation typically and here it's typically a quantum trajectory where each point is related to an integration time of a signal of 30 nanoseconds so here you see in a single trajectory of the qubit, the evolution of the qubits in the time domain with a resolution time of 30 nanoseconds so from this kind of measurement we can also by making some analysis of all the different histograms, statistics we can obtain a decay which is close to the decay that we have measured by a different way showing again that the quantum non-immunitions of a measurement is quite weak using this cross-carcoupling so I need to compare this was experiment performed in Yale group with a longitudinal coupling and they get also quantum dimension very good in this kind of coupling also well now I think it's time to conclude well we can compare to other project but maybe it's time is over now I guess and so in summary I will say that we have developed a new kind of readout which is completely different from the cross-carcoupling so we are no more based by a description using Jens Cummings Hamiltonians but we need to develop a new modellization using this particular coupling which is different so in this sense the theory is not yet well developed and what we have demonstrated in the experiments that we are able to show very good fidelity using these new techniques and also quite good performance and quantum non-demunitions so thank you very much for your attention Thanks Questions Sivan Thank you for the very interesting talk so I have a question it seems to me that still your coupling has like is related to some kind of expansion like this is not like the sigma z times a plus a dagger is still an expansion right so how would it change so would it change your analysis of the QNDness if you would incorporate like high order terms For sure in terms of QNDness for example the question is for example when we change the photon number inside the cavity even our coupling are going to deviate to the simple description of sigma z cross-carcoupling because of this photon when we consider larger photon signals inside the cavity so but typically if you see the perturbation on photos inside the cavity what is funny is in the Jens Cummings descriptions already in the ground first excited states if you are empty of photo in the cavity you are already inducing relaxations and in our case due to this coupling the system is completely non-relaxation channel for when it is in the ground set when the cavity has zero or one photon the system start to be more complicated when you put more one photon in the cavity but it's interesting to probe yes okay thank you other questions thank you it's a comment so Olivier is it really honest to quote such numbers numbers when your numbers are actually much better because here you tell the readout is 99 point something fidelity but it's just due to the relaxation of your qubit your readout is and your queendiness of your readout are better that the numbers you could yes but in some sense you wouldn't know the source of a relaxation so if the source of relaxation is coming from the measurements it's good to include if it's a different source but you can estimate and you know that it doesn't so you present numbers that are pessimistic your system is better yes so thank you Denis thank you I paid him for these questions very nice comment other nice comments okay then I think everybody is hungry so we want to thank the speaker once more