Â鶹ÒùÔº

The operation of quantum computers, systems that process information leveraging quantum mechanical effects, relies on the implementation of quantum logic gates. These are essentially operations that manipulate qubits, units of information that can exist in a superposition of states and can become entangled.

A type of quantum logic gate that enables the entanglement between qubits is a so-called two-qubit gate. Notably, most existing schemes for generating these gates force qubits outside of the conditions or parameters in which they can best store information and are easier to control.

Researchers at the Beijing Academy of Quantum Information Sciences (BAQIS) and Tsinghua University recently introduced a new universal scheme to implement two-qubit gates in superconducting quantum processors. This scheme, outlined in a paper in Nature Â鶹ÒùÔºics, was found to reliably enable the generation of entanglement between qubits in superconductor-based quantum computers.

"The original motivation for our study is rather straightforward," Jianxin Chen, co-senior author of the paper, told Â鶹ÒùÔº. "Textbooks establish that all quantum operations correspond to unitary matrices; yet, historically, the CNOT gate has long been treated as the de facto standard for implementing such operations.

"This framing becomes particularly relevant when considering that over the past few years, the community has identified a far broader range of implementable two-qubit gates beyond conventional options, including the square root of iSWAP, fSim gates, fractional gates, and various others."

Quantum physics and quantum computing theories suggest that quantum operations can also take the form of grids known as unitary matrices. As part of their study, Chen and his colleagues set out to identify types of unitary matrices can be natively implemented with high-fidelity, thus acting as 'quantum instructions'.

"In the context of classical computation, system performance is restricted by 'instruction set architecture'," said Chen. "We anticipate that if we can further expand the high-fidelity implementation of other native gates, we will be able to push forward the system performance."

In their paper, Chen and his colleagues introduce a unified control scheme for the implementation of any possible two-qubit gate. In mathematical terms, their scheme supports all operations in the special unitary group SU(4), which is the full set of valid two-qubit transformations, sometimes requiring only tiny adjustments with simple one-qubit gates.

To control qubits, the team's newly introduced scheme combines two distinct approaches. Firstly, it allows the qubits to interact directly, which is known as exchange interaction. Secondly, it steers the qubits using microwave signals, in ways that are compatible with existing superconducting hardware systems for quantum computing.

"Our scheme has several important advantages," explained Dr. Yan. "First, it is truly universal, meaning that the same hardware configuration and control framework can implement any desired two-qubit gate, providing maximal flexibility in quantum operations.

"More crucially, unlike conventional approaches that rely on problematic |11⟩–|20⟩ transitions (as in standard CZ gates), our method operates exclusively upon the |01⟩–|10⟩ transition. This avoids the risk of leakage into higher energy states, which has been one of the most persistent challenges for quantum error correction protocols."

The recent work by Chen and his colleagues demonstrates that it is possible for an individual scheme to implement all two-qubit gates with high fidelity. In the future, their approach could contribute to the advancement of quantum algorithms and superconducting quantum computers.

"While the relevance of this result to quantum error correction may not be immediately obvious, that this new perspective also holds value for addressing defects in the surface code and easing hardware requirements for the qLDPC code," added Chen.

"We aim to re-examine the boundaries of quantum advantageous by this fresh perspective, and to explore more efficient physical implementation of error correction codes. Another important direction will involve developing more efficient calibration schemes tailored to this work."

© 2025 Science X Network