Quantum simulation promises to model materials that are intractable for classical computers — systems where many quantum particles interact simultaneously. But the platforms have been small: tens of qubits, maybe hundreds. The gap between simulation size and the scale at which interesting physics happens (phase transitions, metal-insulator transitions, emergent order) has been wide enough to limit practical results.
Published in Nature, Michelle Simmons and colleagues at UNSW Sydney built a two-dimensional array of approximately 15,000 individually controllable quantum dots — each defined by phosphorus atoms embedded in silicon with sub-nanometer precision. They call it Quantum Twins. The array is roughly 1,500 times larger than previous attempts with this atom-based approach.
Using this platform, they observed a metal-insulator transition on a 2D square lattice — the point where a material switches from conducting to insulating as the ratio of electron-electron interaction to tunneling changes. The transition was visible in the array's transport properties, controlled by precisely tuning two independent parameters: on-site interaction strength and inter-dot tunneling.
The structural insight is about the relationship between precision and scale. The individual quantum dots must be positioned with sub-nanometer accuracy — a single misplaced atom changes the physics of its local site. But the phenomenon being studied (the phase transition) only appears at thousands of sites. Neither precision alone nor scale alone is sufficient. The breakthrough required both simultaneously: atomic-level control at every site, maintained across 15,000 sites.
This is a manufacturing achievement that enabled a physics achievement. The phase transition was always predicted to exist in the mathematics. Making it visible required building the material that hosts it, one atom at a time, fifteen thousand times.