Scientific Challenge

A solid-state spin works as a qubit only while it holds quantum phase, and what erases that phase is the fluctuating magnetic environment around it. The dominant strategy in the field is to implant a defect into a clean host and fight the environment afterward — costly to place deterministically, and limited by the host one starts with. We take a different position: a spin centre generated by the host’s own defect chemistry is a fundamentally better design space, because the same chemistry that creates the spin also tunes the disorder that limits it. Where a vacancy or a reduced site forms the spin centre, its concentration and geometry are not fixed properties of the host but variables — and they set the spacing, the mutual exchange, and the hyperfine environment that together determine coherence. The question is then concrete: given a chemically generated spin centre, what does the geometry of the surrounding disorder, not merely its amount, do to the couplings that set the coherence time?

Our Approach

Answering that requires the full chain from electronic structure to a coherence time, and we build it along the same logic the group uses for disordered magnets — structure, to couplings, to dynamics — applied here to the spin bath. First-principles calculations, with the on-site interaction fixed from first principles rather than fitted, localise the defect spin correctly and return its magnetic ground state; all-electron reconstruction of the spin density at the nucleus yields the hyperfine tensors that quantify how strongly the central spin feels its environment; and those tensors are the direct input to coherence-time modelling through the cluster-correlation expansion. Because disorder is the object of study, the calculations vary one thing at a time — clustered against dispersed defect arrangements at fixed concentration to isolate geometry, concentration at fixed geometry to isolate crowding. Oxygen-reduced cerium dioxide, where removing a lattice oxygen creates a shielded rare-earth spin centre by intrinsic stoichiometry, is the clean realisation we use, but the framing is general: a spin centre the host chemistry generates and the defect geometry tunes.

Findings

We have computed the complete hyperfine and magnetic characterisation that any first-principles coherence-time calculation requires — the input that, in this field, is usually assumed rather than computed:

• A concentration scale where the single-site picture fails. The isotropic hyperfine coupling is nearly constant and only weakly geometry-dependent while the spin centres are dilute, then collapses by roughly a factor of three once they are crowded — locating the precise concentration at which neighbouring spin densities overlap and the clean single-centre description breaks down.
• Direct structural control over inter-spin exchange. At fixed concentration, the magnetic ground state switches between ferromagnetic and antiferromagnetic order purely as a function of how the defects are arranged — demonstrating that the geometry of disorder, not just its amount, governs the exchange between spin centres, and that this exchange is a knob the host chemistry hands us.
• A spectroscopic fingerprint of the disorder. The anisotropy of the weak ligand hyperfine couplings carries a distinctive, geometry-dependent signature, suggesting that distinct defect-cluster arrangements could be told apart in measurement — a route to reading the disorder that otherwise stays hidden.

These establish, on a real chemically generated host, that we can compute how concentration and geometry reshape the couplings that govern coherence. It is deliberately the input to a coherence time, not yet the coherence time itself.

Outlook

The programme is to close the chain. With the hyperfine and exchange couplings in hand, the next step propagates them through the cluster-correlation expansion to a predicted coherence time — and, further, to coherence as an explicit function of disorder geometry, so that the arrangement of defects becomes a design parameter for the coherence time rather than an uncontrolled property of the sample. Almost no group runs this entire chain — chemically generated spin centre, first-principles couplings, disorder-resolved coherence — under one roof. Building it is the work; the characterisation above is the evidence that its hardest computational link is already in place.

• ← Go back to Research