This thesis describes our work on optimizing quantum control and scaling up quantum computers. We demonstrate the application of automatic differentiation in quantum optimal control, which allows us to specify advanced optimization criteria and incorporate them in the optimization process of quantum control with ease. These criteria enable more intricate control on the evolution path, suppression of departures from the truncated model subspace, as well as minimization of the physical time needed to perform high-fidelity state preparation and unitary gates. We subsequently present our experimental efforts in efficiently scaling up a quantum computer. The innovation of quantum random access memory (qRAM) architecture with multimode circuit QED establishes a promising path towards expanding the information processing power of a quantum device, without introducing control-resources overhead. Using a single Josephson junction transmon circuit serving as the central processor, we demonstrate universal operations on a nine-qubit random access memory. We further present our experimental results in establishing bidirectional and deterministic photonic communication between two remote superconducting multimode processors, connected through a one-meter long coaxial cable. This device demonstrates a prototype of setting up a scalable and distributed quantum computing cluster that could solve a harder computational problem.
This thesis describes our work on optimizing quantum control and scaling up quantum computers. We demonstrate the application of automatic differentiation in quantum optimal control, which allows us to specify advanced optimization criteria and incorporate them in the optimization process of quantum control with ease. These criteria enable more intricate control on the evolution path, suppression of departures from the truncated model subspace, as well as minimization of the physical time needed to perform high-fidelity state preparation and unitary gates. We subsequently present our experimental efforts in efficiently scaling up a quantum computer. The innovation of quantum random access memory (qRAM) architecture with multimode circuit QED establishes a promising path towards expanding the information processing power of a quantum device, without introducing control-resources overhead. Using a single Josephson junction transmon circuit serving as the central processor, we demonstrate universal operations on a nine-qubit random access memory. We further present our experimental results in establishing bidirectional and deterministic photonic communication between two remote superconducting multimode processors, connected through a one-meter long coaxial cable. This device demonstrates a prototype of setting up a scalable and distributed quantum computing cluster that could solve a harder computational problem.
This thesis mainly summarizes two experiments that relate to building a quantum computer out of superconducting transmon qubits. Transmon qubits have emerged as one of the foremost solid state qubits, realizing processors with more than ten qubits and demonstrating small scale quantum algorithms as well as quantum error correction schemes. Right now, there is a race between different academic and industry research groups to scale up transmon qubit processors. The first experiment was a demonstration of qubit control by selective broadcasting in order to reduce the scaling of expensive electronics with the number of qubits for individual single-qubit control. We demonstrated that we can bring two transmon qubits to the same frequency (combining fabrication accuracy and in-situ fine tuning) and use the same hardware to control both, routing the pulses with a nanosecond-timescale vector switch matrix. Despite the compromises required by this technique, we show a scalable path to single qubit control beyond the threshold required for quantum error correction. In benchmarking, we take into account gate leakage due to the fact that transmons are fundamentally multi-level systems. In the second experiment we establish entanglement between two transmon qubits on different chips. We use an entanglement by measurement scheme and demonstrate that we can overcome minor fabrication imperfections by shaping our measurement pulses. Ultimately, performance is mainly limited by photon loss between the chips and up to the amplification chain. This entanglement mediated by traveling photons could be used to make a distributed transmon processor where computations are spread across several chip modules. This modularity could enable connectivities that cannot be realized on chip and ease fabrication requirements, as modules could be individually fabricated and selected. Thus, both of these experiments fit into the larger effort to converge on the hardware, control equipment and architecture of a future large-scale transmon quantum computer. Other experiments I contributed to are summarized in the conclusion chapter to show the diverse physics that can be studied in cQED experiments.
This thesis explores novel strategies for advancing superconducting quantum computing through a series of experimental investigations and device innovations. Grounded in the development of circuit quantum electrodynamics, the work addresses the challenges posed by the limitations of coherent control, coherence time and scaling of superconducting qubits. Focusing on transmon qubits, the thesis first reviews the foundational principles of circuit quantum electrodynamics and the critical role of the transmon in achieving robust qubit performance. Three experimental studies form the core of this work. The first study demonstrates a universal quantum gate set for strongly coupled transmons. By exploiting the residual ZZ interaction, the work introduces a novel two-axis gate protocol for single-qubit operations and implements both CZ and CNOT gates with high fidelities, highlighting the potential for fast, low-error quantum operations in the strong coupling regime. The second study investigates a dual-rail qubit architecture using parametrically coupled transmons, showcasing a hardware-efficient approach to quantum error correction. This experiment converts single-photon loss errors into detectable erasures and leverages mid-circuit detection to enhance qubit coherence, thereby advancing fault-tolerant quantum computing schemes in the NISQ era. The final study focuses on the development of merged-element transmons (MET) and their FinMET variants. By integrating the Josephson junction and shunt capacitor into a single trilayer device, these innovations achieve a dramatic reduction in device footprint while addressing loss channels through advanced materials engineering, ultimately paving the way for scalable, high-coherence quantum circuits. Collectively, these contributions demonstrate significant progress in quantum gate design, error mitigation, and qubit architecture, offering promising routes toward the realization of largescale, fault-tolerant quantum computers
The thesis investigates three challenges in superconducting quantum computing: efficient quantum gate compilation, scalable microwave interfacing of quantum processors (QPUs), and electrical annealing of Josephson junctions.The first part addresses an inefficiency in contemporary two-qubit gate compilation. Universal quantum computation requires only a single entangling two-qubit gate alongside single-qubit rotations, but this minimal gate set requires up to three additional two-qubit gates for every unsupported gate. I show how access to multiple inequivalent two-qubit gate classes overcomes this inefficiency. I experimentally realised an unsupported SWAP operation using one parametric CZ and one parametric iSWAP gate on a circuit-QED architecture. A Kraus-Cirac analysis shows that this combination belongs to the SWAP equivalency class, the fundamental reason for the reduction in two-qubit gates. The experiment demonstrates a rare case of simultaneous CZ and iSWAP support in the same quantum circuit, illustrating how access to multiple inequivalent gate classes can reduce compilation overhead on real hardware.The second part explores direct digital synthesis of microwave signals for interfacing with quantum processors. Conventional control uses heterodyne upconversion with IQ mixers, requiring compensation of LO leakage and IQ skew. I investigated a mixerless architecture based on high-speed DAC/ADC systems to generate and analyse the required microwave signals. CZ and iSWAP gate experiments subsequently verify that the platform is equivalent to conventional QPU setups.The third part investigates post-fabrication electrical tuning of Josephson junctions, in which voltage pulses modify junction resistance to correct fabrication-induced transmon-frequency variations. The speed of resistance change was found to depend exponentially on pulse amplitude. A novel stepwise protocol achieved a 270% resistance increase in an 11-month-old 11.5 kΩ junction, corresponding to a ~2 GHz reduction in transmon frequency, the largest reported selective resistance modification to date. Simulations show that resistance, charging-energy, and superconducting-gap uncertainties contribute equally to the uncertainty in the final qubit frequency. The results support a picture where electrical manipulation modifies the energy landscape, accelerating natural junction ageing.The thesis demonstrates that investigating gate-level compilation strategies, fully digital QPU control, and post-fabrication tuning, helps advance the understanding of how to approach scalable superconducting quantum processors.
As quantum computing transitions into the Noisy Intermediate-Scale Quantum (NISQ) era, scaling monolithic quantum processors remains a significant challenge due to decoherence and operational noise. This dissertation proposes enhancing the connectivity of distributed quantum systems as a scalable solution, focusing on three major research tasks: architectural co-design, operation scheduling optimization, and network topology design. First, to address the physical qubit limits of single-chip processors, we investigate the co-design of network topology and logical qubit allocation for Distributed Quantum Computing (DQC). By modeling qubit interaction as a graph partition problem, we propose a joint optimization framework to minimize inter-QPU communication overhead. Our results demonstrate that the proposed linearized formulation, TACO-L, significantly outperforms traditional two-stage methods in executing complex quantum circuits. Second, we address the challenge of maintaining high-fidelity long-distance entanglements through the optimization of Swapping and Purification Schemes (SPS). We prove that in Binary systems, any optimal swapping and purification tree (OSPT) must follow a Restricted SPT (RT) pattern, where purification is integrated at the link level prior to swapping. Furthermore, we introduce the gradient-based GradTree algorithm, which enables an efficient search for low-cost, high-fidelity scheduling in complex noise environments. Finally, this dissertation formulates the first general Topology Design Problem (TDP) for quantum networks, jointly optimizing quantum memory allocation, optical channel deployment, and entanglement scheduling. By analyzing flow-based and path-based formulations, we prove the optimality of the Relaxed Complete Swapping Tree (RCST) in minimizing the expected cost of entanglement distribution. Collectively, these contributions provide a comprehensive theoretical and algorithmic framework for building scalable, cost-effective global quantum internet architectures.
The discovery of quantum mechanics in the 20th century forms the basis of many of the technologies that define our lives today. The ability to engineer and manipulate individual quantum systems -- even create artificial atoms -- promises a similar revolutionary leap in technology. A quantum technology of particular interest is quantum computing, which has the potential to solve problems that are intractable for classical computers, opening up new domains of computation. Meanwhile, an attractive approach to creating engineered quantum systems is circuit quantum electrodynamics (cQED). Where research initially focused on understanding the physics of cQED devices, focus has shifted to building systems capable of performing useful computations. However, this remains extremely challenging, in part due to the inherently fragile nature of the individual quantum bits, but also due to difficulties in controlling and scaling up these systems. This thesis focuses on the control aspects of building an extensible full-stack quantum computer based on superconducting transmon qubits. We define the demonstration of quantum fault-tolerance as our target application to give focus to our efforts. The QuSurf architecture for a full-stack quantum computer presented in this thesis is designed with this application in mind. We provide a detailed study of the error sources present in this system and give an overview of the relevant characterization techniques. In the second part of this thesis, we address several key challenges in the control of a quantum computer. To realize high-fidelity coherence limited gates, we present a novel tuneup protocol that achieves a tenfold speedup over the state-of-the-art. This is realized by eliminating the need for qubit initialization. We demonstrate this protocol by calibrating single-qubit gates to a coherence limited Clifford fidelity of 99.9% in one minute. Performing repeated parity checks, as is required for quantum error correction, requires reusing qubits quickly after they have been measured. By introducing a numerically optimized depletion pulse we are able to speeds up the depletion of measurement photons in a readout resonator without having to rely on specific symmetry conditions. Using this technique speed up photon depletion by more than six inverse resonator linewidths, reducing the error rate in an emulated ancilla parity check by a factor 75. Flux-pulsing based two-qubit gates are the fastest two-qubit gates. However, they are also very technically demanding. The key challenge in performing these gates is addressing the distortions that control signals experience as they traverse various electrical components. We have developed Cryoscope (short for cryogenic oscilloscope) to characterize and correct these distortions. Cryoscope is an in-situ technique that uses the qubit to sample control pulses of arbitrary shape. Even when correcting distortions to within sim0.1 two-qubit gates are history-dependent due to the long timescale upon which some of these distortions act. We have invented Net-Zero, a new type of flux-pulsing based two-qubit gate, to address this problem. It makes use of a symmetry condition of the transmon to have net-zero integral, making the gate resilient to long-timescale distortions. The gate suppresses leakage out of the computational subspace to 0.1% by making use of leakage interference and has a built-in echo effect that enhances the coherence of the gate, achieving a two-qubit gate fidelity of 99.1%. Custom software is required to perform the physics experiments needed to build and operate a quantum computer. PycQED is an open-source software framework we have developed for this purpose. We discuss the design choices and concepts of PycQED before turning our focus to characterization and calibration. Here we introduced dependency graphs as a useful abstraction and system emulation as an essential development tool for automating the characterization and calibration process. We conclude the thesis by reflecting on the limitations of our architecture and providing an outlook on the grand challenges of building a useful kilo-qubit sized quantum computer. We define these challenges as The Application Problem, The Fabrication Problem, and The Calibration Problem.
Abstract Distributed quantum computing (DQC) combines the computing power of multiple networked quantum processing modules, ideally enabling the execution of large quantum circuits without compromising performance or qubit connectivity 1,2 . Photonic networks are well suited as a versatile and reconfigurable interconnect layer for DQC; remote entanglement shared between matter qubits across the network enables all-to-all logical connectivity through quantum gate teleportation (QGT) 3,4 . For a scalable DQC architecture, the QGT implementation must be deterministic and repeatable; until now, no demonstration has satisfied these requirements. Here we experimentally demonstrate the distribution of quantum computations between two photonically interconnected trapped-ion modules. The modules, separated by about two metres, each contain dedicated network and circuit qubits. By using heralded remote entanglement between the network qubits, we deterministically teleport a controlled-Z (CZ) gate between two circuit qubits in separate modules, achieving 86% fidelity. We then execute Grover’s search algorithm 5 —to our knowledge, the first implementation of a distributed quantum algorithm comprising several non-local two-qubit gates—and measure a 71% success rate. Furthermore, we implement distributed iSWAP and SWAP circuits, compiled with two and three instances of QGT, respectively, demonstrating the ability to distribute arbitrary two-qubit operations 6 . As photons can be interfaced with a variety of systems, the versatile DQC architecture demonstrated here provides a viable pathway towards large-scale quantum computing for a range of physical platforms.