A Decades Old Puzzle Gets a Map, a Step Toward Better Superconductors One Day

Think what a wire costs us. Electricity loses energy to resistance on its way to a house, a hospital or a train. Some materials, cooled down, carry current with no resistance at all, and copper-based compounds called the cuprates do it at surprisingly warm temperatures. If anyone understood why, better superconductors for power lines, magnets and electronics would be a more realistic aim. Nobody does. The cuprates were found decades ago and physicists are still arguing.

In the middle of that argument sits a strange state called the pseudogap: a kind of metal that seems to be missing some of its low-energy electrons. A team led by Markus Greiner at Harvard, with a coauthor now at JILA in Colorado, has built that state out of atoms and mapped where it appears. Their results were published in Nature on August 5, 2026.

The quantum machine here is not a general-purpose computer. It is an analog quantum simulator: a controlled cloud of ultracold atoms that obeys the same physics as the model it stands in for, so the experimenters can simply measure what happens.

The model in question is the Hubbard model, physics’ famously simple sketch of these materials. Particles hop around a grid and repel each other when they share a site. You can write it on a napkin, yet it is very hard to solve. For the low temperatures where the pseudogap lives, the team’s preprint states that a standard, numerically exact classical method, determinant quantum Monte Carlo, is “infeasible due to the fermion sign problem.” That is the knot this experiment set out to work around.

So instead of computing the model, they built it. According to the preprint, the team trapped ultracold fermionic lithium atoms in a two-dimensional optical lattice made of laser light, with a uniform region of roughly 300 sites, and photographed them site by site with a quantum gas microscope. The atoms play the role of electrons. By tuning how strongly they repel and how many holes are in the grid (the doping), the team could walk across the model’s phase diagram directly. There is something quietly satisfying about that: a hard equation answered by arranging atoms on a grid of light and photographing them.

What made it possible was cold. A several-fold drop in the temperatures the team can reach let them see the crossover from a normal metal to a pseudogapped one. As the system cooled, its compressibility (how easily you can squeeze more particles in) developed a maximum at intermediate doping, and following that maximum traced a line in the phase diagram separating an underdoped metal from an overdoped one at strong interactions.

Then they shook the lattice and listened. In the underdoped regime the low-energy response went missing, and the loss was strongest in regions of momentum space known as the antinodes. That is the pseudogap’s signature. Using it, the team mapped a pseudogap phase diagram across interaction strength and doping, and the paper says the results experimentally establish and characterize the pseudogap metal in the model.

It is fair to note that theory was not standing still. Approximate methods, such as extensions of dynamical mean-field theory, had predicted some of these features. What the experiment adds is a direct test of them in a regime where the exact method runs out. The authors sum up the larger point in their abstract:

this work demonstrates the utility of quantum simulation in addressing frontier problems in correlated electron physics. Kendrick, Kale et al., Nature (2026)

Be clear about the distance. This is not a measurement inside an actual cuprate, and a pseudogap in the Hubbard model does not by itself explain how high-temperature superconductivity arises. Better superconducting wire is a long road, and this is one careful step along it.

The people whose work changes today are the theorists. Some have spent entire careers on this model, and instead of arguing only with each other they now have a measured map to check their ideas against. That is how an argument this old gets settled, one honest measurement at a time, and the paper already points at the next piece: possible connections with charge order, which the authors leave for future work.