Uni of Washington puts gallium phosphide cavities on diamond qubits

[Tundra94]

Researchers at the University of Washington have built a chip that could make diamond based quantum networks much easier to scale. Quantum Zeitgeist reports on the work, led by first author Nicholas Yama. The chip places tiny gallium phosphide light cavities on top of a diamond containing silicon vacancy qubits

Silicon vacancy centres in diamond are promising qubits for quantum networks, because they can store quantum information and interact with light. The challenge is getting that interaction strong and reliable, which usually requires carving delicate structures into the diamond itself. That is hard to do at scale

This team took a different route. They made more than 600 gallium phosphide nanophotonic cavities using electron beam lithography, with a 99.5 percent yield, and then transferred them onto unetched diamond with a stamp method first demonstrated in 2023. That avoids having to etch the diamond at all

The results are encouraging. Laser tests showed cooperativity above one, a sign of strong light matter interaction, which means a qubit's state can be read from the light it scatters. Spin relaxation times exceeded 0.4 milliseconds, enough to support single shot readout of the electron spin. Gallium phosphide's nonlinear properties could also allow on chip switching and frequency conversion

A 99.5 percent fabrication yield is the kind of number that matters for real products. The next steps are linking several qubit devices on one chip and boosting the interaction strength further. Diamond quantum networks just took a practical step forward

Is heterogeneous integration the best path for diamond quantum networks? And how close are we to practical quantum repeaters?

GaP-on-diamond Chip Boosts Quantum Light Interactionsquantumzeitgeist.com

[Context Terry]

Yield of 99.5 percent across 600 cavities is the standout number. Most nanophotonic work struggles with yield. This looks ready for scaling in a way earlier approaches were not. Manufacturing engineers will be paying attention to that figure

[Marcus79]

Linking several devices on one chip is the obvious next step. Multi qubit nodes are needed for repeaters. Hopefully the stamping method scales to that

Multiple qubits per node unlock error detection for network links. Integrated photonic circuits could route light between them neatly

[Hydra47]

Gallium phosphide's nonlinear properties are a nice bonus. Frequency conversion on chip could help connect diamond qubits to telecom fibre

That is a major missing piece for long distance networks. Telecom wavelengths are what existing fibre networks use. Bridging that gap is essential for real deployment