Superconducting qubits and resonators with quality factors exceeding 107 are of great interest for quantum information processing applications. The improvement of present devices necessarily involves the consideration of participation ratios, which budget the influence of each physical component in the total energy decay rate. Experiments on compact resonators in which participation ratios were varied has demonstrated the validity of this method, yielding a two-fold improvement in quality factor. Similar experiments on compact transmon qubit devices led to a three-fold improvement over previous transmons, validating the method of participation ratios for qubits as well. Through the use of a 3D cavity, a further minimization of the participation of surface components combined with the removal of unnecessary components, produced an additional ten-fold increase in coherence times. Finally, the fluxonium qubit was redesigned in a similar minimalist environment with an improved superinductance, thus combining the advantages of the 3D architecture with the natural insensitivity to dissipation of the fluxonium, resulting in another tenfold increase in relaxation times. This large increase in relaxation and coherence times enables experiments that were previously impossible, thus preparing the field of quantum information to advance on other fronts.
Superconducting qubits and resonators with quality factors exceeding 107 are of great interest for quantum information processing applications. The improvement of present devices necessarily involves the consideration of participation ratios, which budget the influence of each physical component in the total energy decay rate. Experiments on compact resonators in which participation ratios were varied has demonstrated the validity of this method, yielding a two-fold improvement in quality factor. Similar experiments on compact transmon qubit devices led to a three-fold improvement over previous transmons, validating the method of participation ratios for qubits as well. Through the use of a 3D cavity, a further minimization of the participation of surface components combined with the removal of unnecessary components, produced an additional ten-fold increase in coherence times. Finally, the fluxonium qubit was redesigned in a similar minimalist environment with an improved superinductance, thus combining the advantages of the 3D architecture with the natural insensitivity to dissipation of the fluxonium, resulting in another tenfold increase in relaxation times. This large increase in relaxation and coherence times enables experiments that were previously impossible, thus preparing the field of quantum information to advance on other fronts.
Abstract : This is the final report for the IARPA CSQ project at Yale University, on the coherence limits of superconducting charge qubits. During this project, we performed the proposed work investigating the various sources of decoherence and their origins in different materials, using the participation ratio paradigm that we developed and disseminated in the community. The original project goal was to increase by an order of magnitude the coherence time of the state-of-the-art qubits (transmons and fluxonium designs) from a few microseconds to greater than 20 microseconds. By project end, we realized qubit and cavity devices with times in excess of a millisecond, or an improvement by almost three orders of magnitude. These results, especially improved 3D designs, have been widely adopted and successfully reproduced by many groups across the superconducting qubit community. At program end, we have an existence proof that qubits satisfying the quantum error correction threshold are possible with existing materials and fabrication techniques, leading to a growing belief that large-scale quantum computation with solid-state devices will eventually be realized.
We report superconducting fluxonium qubits with coherence times largely limited by energy relaxation and reproducibly satisfying T 2 > 100 μs (T 2 > 400 μs in one device).Moreover, given the state-of-the-art values of the surface loss tangent and the 1=f flux-noise amplitude, the coherence time can be further improved beyond 1 ms.Our results violate a common viewpoint that the number of Josephson junctions in a superconducting circuit-over 10 2 here-must be minimized for best qubit coherence.We outline how the unique to fluxonium combination of long coherence time and large anharmonicity can benefit both gate-based and adiabatic quantum computing.
Superconducting quantum computing is a solid state quantum computing technology based on Josephson junction circuits. It becomes one of the most promising candidates for realizing practical quantum computer/simulator, because of its scalability, easy to coupling and control, compatibility to semiconductor technologies, etc. As the complicated electromagnetic environment in solid state metal and insulator, superconducting quantum bits (qubits) suffered from shot coherence time since its early development. In this review, we described some recently developed superconducting qubit structures, including Transmon/Xmon, Fluxonium and C-shunt flux qubit. Although these new qubit structures have very different circuit parameters and energy scales, they all find their way to effectively suppress environmental noises. As a result, their coherence time were improved significantly. Transmon, a transmission-line shunted plasma oscillation qubit, was now among the most popular structures for building a multi-qubit superconducting quantum circuit. It introduced a relatively large shunt capacitor to remarkably reduce the charge dispersion. Although the non-linearity is also decreased, by operating in a properly chosen EJ/Ec, transmon can greatly reduce the charge noise while only sacrificing a small amount of non-linearity. Xmon can be considered as a modified version of transmon. It is designed and optimized for multi-qubit operation and measurement with separate X/Y and Z control lines for each qubit, and a readout bus (a transmission line or a low-Q bandpass filter) for simultaneous readout of all the qubits. Fluxonium provide another solution for suppressing the charge noise in Cooper-pair box (CPB). It introduces a large shunt inductance to filter low-frequency offset charge in the small capacitor of the phase-slip junction (the small junction). To keep the qubit operating in the charge region, the shunt inductor should be large enough to keep the impedance larger than the superconducting impedance quantum RQ~1 kΩ. As a result, the inductance can only be provided by a large series of Josephson junctions. Fluxonium keeps the advantage of high anharmonicity of CPB. In addition, it introduces another degree of freedom: external flux Φext. Quasiparticles (electrons or holes) tunneling at the small junction add coherently to produce a ″1+cosφ″ term. As a result, when Φext= Φ0/2, dissipation caused by quasiparticle tunneling is suppressed and the energy relaxation time (T1) increases drastically. C-shunt flux qubit (CSFQ) decreases α and EJ to suppress flux noise in the qubit. While the charge noise becomes the dominant noise source, a large shunt capacitor is added to suppress the charge noise, similar to that in a transmon qubit. It is worth to notice that in CSFQ, the α decreases to below but near 1/2. As a result, the double-well potential no longer exists. The energy structure become similar to transmon, with the anharmonicity be opposite. CSFQ can get a much improved coherence time, while its anharmonicity can be larger than that of transmon. Those new superconducting qubit structures provide different solutions to improve the coherence time. Although it looks much different in circuit models and parameters, they have ideas in common, and may act as guides to other novel designs with longer coherence time. We can summarize as below. (1) The qubit circuit should be free of small superconducting islands. Charge fluctuation in small islands is a major source of decoherence. As a result, increase the area of the islands, that is, increase the capacitance becomes a straight forward way to suppress the charge noise. In Fluxonium, the small junction is connected with a large series of large junctions that there is also no small islands in the whole loop. (2) Remove superconducting loops in the qubit circuit, or minimize the loop size and loop current to suppress flux noise. (3) Bias the phase difference of the small junction to π, where dissipation caused by quasiparticle tunneling can be suppressed coherently. (4) Decrease the coupling bandwidth of qubit to the environment. In a circuit-QED readout scheme, a high-Q readout resonator that dispersively coupled to the qubit protects the qubit from exposed to the electromagnetic environment of the readout circuit. A ″Purcell filter″ can also be added to suppress such decoherence further. In conclusion, we overviewed some qubit engineering techniques or ideas to improve the coherence time, which is a very important figure for building practical quantum computer/simulator. We hope that someone can combine and optimize these ideas, while interact with developments in materials and micro-fabrication technologies, to develop novel new generation superconducting qubit structures in the near future.
While quantum computation has the potential to revolutionize the scientific community, to date no architecture has been developed which offers the necessary combination of high coherence times and massive scalability. Superconducting flux qubits satisfy the second requirement well but to date useful devices are limited to coherence times of typically only a few &mus. In this dissertation we examine the possibilities of improving the coherence performance of the flux qubit to the levels required for fault-tolerant quantum computation. We find that coherence times for many devices are limited by photon-induced quasiparticles and mitigation of these quasiparticles increases coherence times by more than a factor of two. Beyond this, however, we find little improvement in flux qubit performance compared to prior results. Despite improved fabrication techniques and varied device designs we find flux qubit coherence times are still typically below 5 &mus. Furthermore, wide device-to-device variations are observed which prevent effective scaling of the flux qubit to quantum information circuits.Based on the proposal by You, et al. we develop of a capacitively-shunted version of the flux qubit called the C-shunt flux qubit. With the addition of a capacitive shunt across the small junction of the flux qubit we are able to reduce the amplitude sensitivity to both charge and flux noise by more than a factor of three. The result is a predicted ten-fold enhancement in the coherence times compared to the unshunted flux qubit. At the same time we preserve much of the anharmonicity of the flux qubit resulting in a device with coherence times comparable to modern transmons but with a factor of four better anharmonicity and more flexible coupling configurations.By using a high-quality MBE aluminum shunt process on an annealed sapphire substrate coupled with a more conventional electron-beam-evaporated aluminum Josephson junction process we fabricate hybrid C-shunt flux qubits. We also incorporate a new bridgeless junction technique which offers more robust and repeatable fabrication processing. This technique enables improved accuracy of the junction sizing, a trait which will become essential as we move to multi-qubit experiments. T1 and T2 times for these C-shunt devices reach in excess of 40 &mus at degeneracy, achieving the predicted ten-fold improvement based on the reduced noise sensitivity. At the same time we see decreased device-to-device variation with typical T1 times exceeding 20&ndash25 &mus at all bias points. Combined with the high anharmonicity and standard planar fabrication techniques these coherence times make the C-shunt flux qubit a competitive candidate for fault-tolerant quantum computation.
A quantum computer will potentially solve far-reaching problems which are currently intractable on any classical computer. Many technological obstacles have prevented the realization of a quantum computer, the main obstacle being decoherence, which is the loss of quantum information. Decoherence arises from the undesired interaction between qubits and their environment. Isolated qubits have better coherence but are more difficult to control. Superconducting qubits are a promising platform since their macroscopic size allows for easy control and coupling to other qubits. While the coherence of superconducting qubits has substantially improved over the past two decades, further improvements in coherence are required.We have repeatedly and reliably increased the coherence times of superconducting qubits. Currently decoherence in these devices is dominated by coupling to material defects. These defects are present in the dielectrics used to fabricate these devices or introduced during fabrication. Using simpler resonators as a testbed, we individually isolate, characterize, and then improve each step of the more complicated fabrication of superconducting qubits. We increased the quality factor of resonators by a factor of four by first identifying the surfaces and interfaces as a major source of loss and then by optimizing the substrate preparation. Furthermore, we measure and subsequently mitigate additional defect loss, which is dependent on the position of ground plane holes used to limit the loss from magnetic vortices. Implementing these improvements led to an increase of our qubit coherence times by more than an order of magnitude.The progress made in coherence while maintaining a high degree of connectivity and controllability has been directly used in more complex circuits. One such device is a fully connected three qubit ring with both tunable qubit frequencies and adjustable qubit-qubit couplings. The considerable level of control allows us to generate the quantum dynamics resulting from a collision between Na and He atoms by encoding the problem into a time-dependent multi-qubit interaction. We compute a scattering cross section by measuring the probability of exciting the Na atom over a wide range of collision parameters. The scattering cross-section shows that there is a velocity where the excitation probability is maximized, in agreement with scattering experiments.
Bosonic codes in superconducting resonators are a hardware-efficient avenue for quantum error correction and benefit from the inherent bias toward relaxation errors provided by long-lived cavities compared to typical superconducting qubits. The required coupling to an ancillary control qubit, however, can negate these benefits by inducing highly detrimental effects such as excess decoherence and undesired nonlinearities. This is exemplified by the typically employed transmon qubit that imposes limits on cavity operations due to its finite lifetime as well as the inevitably induced self-Kerr nonlinearity. An important question is thus whether a cavity-qubit coupling can be realized that offers readout and control capabilities without spoiling the cavity. Here, motivated by its long lifetime and design flexibility of its Hamiltonian, we experimentally investigate the fluxonium as a control qubit for superconducting cavities. We couple a fluxonium qubit to a superconducting resonator in the strong-dispersive regime and use it to measure the coherence and flux-tunable inherited nonlinearities of the resonator. We then demonstrate quantum control of the oscillator by preparing and characterizing resonator Fock states and their superpositions, with fidelities limited by resonator decay in our planar prototype device. Finally, we show that the fluxonium can realize cavity-couplings that eliminate undesirable cavity nonlinearities and reach regimes of suppressed self-Kerr that are inaccessible to transmon-based devices. Our results thus demonstrate the potential of the fluxonium as a high-performance bosonic control qubit for superconducting cavities that can outperform existing transmon-based architectures.