In a laboratory at ETH Zurich, physicist Yiwen Chu and her team have shown that a quantum chip can hold its working memory not as electromagnetic fields but as tiny, inaudible vibrations inside microscopic resonators. Borrowing the old and proven division between a processor and its separate RAM, the design asks quantum machines to organize themselves the way ordinary computers learned to think long ago. The result, published in Science, is an early step rather than a finished machine, yet it marks a moment when the quiet structure of memory itself becomes part of the story of building compute
ETH Zurich Builds Quantum Chip That Stores Memory as Vibrations, Like a Guitar String
Memory held not in fields, but in vibration
Mark
So what actually changed here? Quantum memory already exists.
Mimi
What changed is the medium. Instead of electromagnetic fields, the memory is mechanical vibration in tiny resonators, and a superconducting qubit reads and rewrites it.
Mark
Why does the CPU and RAM comparison matter?
Mimi
The team separated the roles. The qubit processes and controls, the memory holds information temporarily, much like a normal computer.
Luke
Careful with the advantages, though. Smaller, more vibration patterns, longer storage: the reporting presents these as potential benefits. What's confirmed is the first demonstration that resonators and qubits can compute together.
Mark
And the tests?
Mimi
The quantum Fourier transform and period finding. Both need several quantum states coordinated precisely.
Luke
The source gives no figures, though: no error rates, no storage times, no number of resonators. So we can't say how it compares with electromagnetic memory.
Mark
What about the claim it could do arbitrary computations?
Mimi
It's stated as in principle, that the architecture has the fundamental operations needed.
Luke
Right, in principle. Scaling while keeping precision is the open question, and nobody has shown that yet.
Der Puls
- Quantum computers are constrained by bulky electromagnetic memory that makes them hard to grow into larger machines.
- The ETH Zurich team has replaced that memory with mechanical vibrations in microscopic resonators, a disruption to the usual design.
- By splitting a superconducting qubit processor from a separate quantum memory, the team is copying the logic of classical CPU and RAM architecture.
- The chip has passed tests with the quantum Fourier transform and period finding, showing it can support real computation.
- The architecture is landing as a promising experimental proof, but scaling it to larger, reliable machines remains the open challenge.
In a laboratory at ETH Zurich, physicist Yiwen Chu and her team have shown that a quantum chip can hold its working memory not as electromagnetic fields but as tiny, inaudible vibrations inside microscopic resonators. Borrowing the old and proven division between a processor and its separate RAM, the design asks quantum machines to organize themselves the way ordinary computers learned to think long ago. The result, published in Science, is an early step rather than a finished machine, yet it marks a moment when the quiet structure of memory itself becomes part of the story of building computers that can do what classical ones cannot.
At ETH Zurich, physicist Yiwen Chu and her team have built a quantum chip whose working memory lives in vibration. The vibrations sit at frequencies far beyond human hearing, held inside a tiny chip rather than an instrument. The work appears in the journal Science.
The design follows the logic of an ordinary computer. A conventional machine separates the processor that calculates from the working memory, or RAM, where information waits, and that arrangement shapes how well it runs. Chu's group applied the same split to quantum hardware: a superconducting qubit serves as processor, while a separate quantum memory holds information mid-calculation for retrieval and modification.
What sets the memory apart is its medium. Most systems store quantum information electromagnetically, which is well understood and precise but takes up considerable space and makes larger machines hard to build. Mechanical resonators offer a smaller footprint, room for many distinct vibration patterns, and quantum states that last longer before fading.
The team tested the architecture with the quantum Fourier transform and period finding, both demanding precise coordination of several quantum states. Igor Kladaric, a doctoral student and co-author, notes that the Fourier transform underpins many quantum algorithms. Passing these tests showed the system could store, control and keep coherent multiple states at once, which is evidence of real computation rather than mere storage.
Still, this remains an experimental chip. The next task is scaling to larger systems while keeping lab-level precision, and practical quantum computers for wide scientific and industrial use remain some distance away.
At ETH Zurich, physicist Yiwen Chu and her team have built a quantum chip in which working memory is held as vibration. The vibrations sit at frequencies far above anything a human ear could pick up, and they happen inside a tiny chip rather than on an instrument. The work is described in the journal Science.
The design borrows from the layout of an ordinary digital computer. In a conventional machine, the central processing unit does the calculating while information waits in separate working memory, known as RAM. How those parts are arranged shapes how efficiently the machine runs. Chu's group applied the same split to a quantum system. A superconducting qubit acts as the processor and control unit, and a separate quantum memory temporarily holds the information needed mid-calculation, so the processor can retrieve and modify it.
The difference is in how that memory is stored. Chu explains that today's systems usually store information electromagnetically, whereas in her team's working memory it is held as mechanical vibrations. Microscopic structures called mechanical resonators carry the information in particular patterns of vibration. When the qubit needs a stored state, it accesses that pattern, performs the required operation, and returns the modified information to memory.
Many existing quantum computers pair superconducting qubits with electromagnetic memory. Researchers understand those technologies well, and combined they give very accurate control over quantum states, including reading, changing and manipulating what is stored. The drawback is size: electromagnetic memory can take up considerable space, which makes it hard to grow experimental devices into the larger machines researchers and industry want.
Mechanical resonators address several of those limits. They are substantially smaller, allowing more compact designs. They can support numerous distinct vibration patterns, so they can in principle hold more information at once. And mechanical vibrations can preserve quantum states longer before the information fades, which matters for calculations that need data to stay available across many processing steps.
The Science paper is the first experimental demonstration that mechanical resonators can work with superconducting qubits to carry out quantum computations. To test the architecture, the team ran two core techniques: the quantum Fourier transform and period finding. Both require coordinating several quantum states with considerable precision. Igor Kladaric, a doctoral student in Chu's group and co-author, explains that the Fourier transform is a fundamental procedure for many quantum algorithms, and that the period-finding algorithm they implemented showed how it can be used.
Passing those tests means the system stored, controlled and kept coherent relationships among multiple quantum states at once, evidence that it can support real computation and not just serve as a memory device. Chu says the interaction between processor and memory is a crucial foundation for making quantum computers a reliable way to do calculations conventional machines cannot. The architecture can, in principle, perform all the fundamental operations needed for arbitrary quantum computation, the essentials of a programmable, general-purpose machine.
This is still an experimental chip. The next challenge is scaling to larger systems with more processing power while keeping the precision shown in the lab, and the team is working on how to develop the approach further. Whether mechanical quantum memory stays effective as components and calculations multiply will decide how far it goes. Practical quantum computers for wide scientific and industrial use remain some distance off.
Bemerkenswerte Zitate
In our quantum working memory, information is not stored electromagnetically, as is usually the case today, but in the form of mechanical vibrations.— Yiwen Chu, ETH Zurich
The Quantum Fourier Transform is a fundamental procedure for many quantum algorithms; the period-finding algorithm showed how it can be used.— Igor Kladaric, doctoral student and co-author (paraphrased)
Möchten Sie die ganze Geschichte?
Das Original lesen bei EUROPE SAYS ↗