Science1 publisher2 min readPublished
Two labs agree on a quarter-electron charge in gallium arsenide's ν = 1/2 quantum Hall state
Labs at EPFL and the Weizmann Institute each measured quasiparticles with a quarter of an electron's charge in the ν = 1/2 quantum Hall state. The agreement settles the charge of a state thought to be non-Abelian, though the exchange behavior a topological qubit would need was not tested.
The Scientist · Science desk
What happened
- Teams led by Mitali Banerjee at EPFL, Moty Heiblum at the Weizmann Institute and Mansour Shayegan at Princeton reported on the ν = 1/2 quantum Hall state in Physical Review Letters.
- Electrons were confined in a 70-nanometer-wide gallium arsenide layer, and each device had a narrow constriction, a quantum point contact, used to measure shot noise.
- Two nearly identical devices, one at EPFL and one at Weizmann, were first checked on known states carrying a full electron charge and two-thirds of one.
- At ν = 1/2 both devices gave about a quarter of an electron's charge, 0.250 plus or minus 0.013 in one and 0.249 plus or minus 0.018 in the other.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- precedent Banerjee calls two-lab agreement on a fractional charge a first for the field, so later claims about even-denominator states will be expected to meet the same replication standard.
- constraint Charge is one property of the state; the case for fault-tolerant qubits at ν = 1/2 still depends on how the quasiparticles behave when exchanged, which this work did not test.
- decision With the charge agreed across two labs, groups working on ν = 1/2 can spend their next experiments on exchange statistics instead of re-measuring charge.
The first thing to check in a shot-noise charge claim is the device. Here there are two point contacts of one design, and the second sits in a different lab [6]. The method works out charge from randomness. Carriers cross the constriction in stop-and-start bursts, and the size of the resulting electrical noise shows how much charge each one moves [4]. The teams etched the constriction with their own method and put metal gates around it to set how much current gets through [5]. Because both devices were first run on states with known charges, a quarter-charge answer is hard to blame on a miscalibrated setup [6].
The two central values differ by 0.001 of an electron charge. That is less than a tenth of the smaller error bar [1]. Each error bar is roughly 5 to 7 percent of the quarter charge being measured [2]. "For the first time in the history of this field, two different groups (Weizmann and EPFL) have measured the same values of fractional charge," Banerjee said [1].
The thing this doesn't tell you is whether the particles are non-Abelian. Some theories predict that even-denominator states could host non-Abelian anyons, whose quantum properties make them candidates for fault-tolerant topological quantum computers [8]. Banerjee said the ν = 1/2 state "survives up to a few kelvins, and is thought to be only the second known state in gallium arsenide to possess special non-Abelian properties that can eventually enable a topological quantum computer" [2]. In her description, the computing payoff comes from exchanging the particles. "By simply moving such particles around each other, we can create error-free quantum computers," she said [3].
The experiment as reported measured charge through noise. The account describes no experiment in which quasiparticles were moved around one another [2][4]. I think the work supports a narrower result that is still useful: a quarter charge at ν = 1/2 that does not rest on one lab's device [1]. Whether this is the non-Abelian state the theories describe needs a separate measurement. The paper's title, Observation of e/4 Charge at ν = 1/2 in a Wide GaAs Quantum Well, claims only the charge [2].
What to watch
- An experiment that moves ν = 1/2 quasiparticles around one another, the operation Banerjee says a topological computer would rely on.
- A third lab, or a different well width or constriction design, reproducing e/4 at ν = 1/2.
- Whether the state's reported survival up to a few kelvins holds in devices built to manipulate quasiparticles, not just measure their noise.