Traditional computers operate on the basis of binary systems. The fact that silicon transistors are either on or off—representing two distinct states—forms the building block of the digital world. Quantum computers continue this concept; their fundamental unit of information, the “qubit,” collapses into only one of two possible states, like a classical bit, at the moment of measurement.
However, quantum physics is not limited solely to binary systems. By nature, quantum systems can access many more states. For example, consider an electron: It can exist at different energy levels around the atomic nucleus. In quantum computers, qubits are usually created by selecting the lowest two of these energy levels. However, in theory, it is possible to use more levels beyond these two states.
The Era of “Qudits” Offers More In a new study published in the journal Nature, researchers announced a new experimental method in which quantum information units can harbor not just two, but three or four states. These multi-state systems are generally known as “qudits” (quantum digits). Those with three states are called “qutrits,” and those with four are called “ququarts.”
The researchers’ greatest achievement in this study was the successful application of an error correction method in these multi-level quantum systems for the first time. This development could open the door to the possibility of processing more information using less hardware.
On the other hand, there are several reasons why qudit systems have not become widespread. First, much existing quantum hardware is designed to handle only two states. When more energy levels are added, the differences between these levels shrink, making them harder to distinguish. Additionally, working with multi-level systems may require a completely different programming model compared to qubits.
Despite this, systems like qutrits and ququarts could be a solution to the capacity problems experienced in quantum hardware. Today, the largest quantum computer manufacturers struggle to produce enough qubits and connect them to perform meaningful calculations. If more information can be stored in fewer physical systems, this could lead to reaching quantum supremacy sooner.
Error Correction in Qudits The system used in the new study relies on the transmon, a currently common piece of quantum hardware. This superconducting structure acts as a quantum bit connected to a microwave resonator. However, in this experiment, the transmon was integrated with an additional microwave cavity to make it capable of carrying more modes.
When a sufficient number of photons are sent into this cavity, interference patterns form between the photons. These patterns represent different energy modes, and each can be used as an information state. More modes mean more information. However, at the same time, the risk of photon loss increases, and error rates rise.
Researchers successfully applied error correction algorithms to reduce error rates by creating qutrit and ququart structures in this system. This demonstrated that a step previously possible only with qubits can also be achieved in more complex systems.
New Horizons for Quantum Memory In research on quantum information units that go beyond qubits, ensuring the stability of these systems is as important as developing multi-state quantum systems. While these systems, which researchers call qudits, show promise for the future of quantum computing, they also bring certain technical challenges.
In the new experiments, the technology playing a key role in the stability of these systems stands out as the transmon and its associated microwave cavity. Typically, a transmon is used to control the quantum state of the cavity and read this state when necessary. However, in this study, scientists used the transmon not just for data reading, but for weak measurement, a much more delicate process.
Weak measurements offer clues as to whether the system’s quantum state has changed, rather than disrupting it. In other words, while it doesn’t say exactly what the state in the resonator is, it can detect whether an error has occurred in the system. By performing these measurements in series, researchers revealed not only the presence of an error but also its nature and how it could be corrected. This error correction process was optimized to ensure system stability. Interestingly, the researchers did not design this control mechanism directly based on theoretical models. Instead, they identified all variables effective in controlling the system and optimized them using reinforcement learning. The ultimate goal was to enable the quantum state to be preserved for a longer time—in other words, to convince the system to act like a memory, even if temporarily.
In the experiments, the system was operated sequentially as a qubit, qutrit, and ququart. For each, the duration the system remained stable was measured—both with error correction enabled and disabled.
The results were quite striking: As the move was made from qubit to qutrit, and then to ququart—meaning as the system harbored more information—the life of the quantum memory shortened. However, when error correction was activated, some of these performance losses were compensated. For example: An error-corrected qutrit was able to remain stable as long as an error-free qubit. An error-corrected ququart performed better than an error-free qutrit. In every case, system life was extended by approximately 1.8 times with error correction.
Of course, these experiments were conducted on a single device for now, without establishing connections with other qudits or performing real calculations. However, considering that previous studies on qubits started on a similarly small scale, it would not be wrong to say that such proofs of concept are critical steps for future technologies.
While computational complexity remains a significant barrier, given the two fundamental problems facing quantum systems today—low qubit count and high error rate—developing an approach that offers a solution to at least one of these is noteworthy progress.
You Might Also Like;
