A Million-Crystal Chip Shows How Solar-Cell Semiconductors Split Into Cores and Shells
Northwestern researchers mapped perovskite-like crystals on a 2-centimeter chip, explaining why adding bromine can first raise, then sharply cut, a material's bandgap.

A chip just 2 centimeters across can now hold roughly a million microscopic crystals, each a slightly different recipe. Researchers at Northwestern University used the device to show how some semiconductor mixtures separate into distinct cores and shells, changing the light they absorb.
The work, published in Advanced Materials, comes from Olivier Chevalier, Chad Mirkin and colleagues. It targets perovskites and related materials called perovskitoids, which are promising for solar cells, LEDs and light detectors. Their ingredients and internal crystal structures determine how they absorb and emit light, so finding the best recipe has traditionally meant making and testing samples one at a time.
The team built what it calls a "megalibrary" of one million compositionally distinct, positionally encoded microcrystals. They patterned a removable polymer coating on glass with a million microscopic wells, then sprayed different precursor solutions onto the chip's four corners. Where the solutions overlapped, composition changed gradually from place to place. The recipes combined cesium with varying proportions of bismuth, antimony, bromine and iodine, and four chips covered a broad range of mixtures.
Heating the templates produced crystals about 10 micrometers wide. Peeling away the polymer left an orderly array in which each position corresponds to a different mixture. The crystals were big enough to supply material for both light-absorption measurements and X-ray diffraction, which reveals atomic arrangements. The team mapped hundreds of positions and linked structure to bandgap, the energy threshold for exciting electrons in a semiconductor. A smaller bandgap lets a material absorb lower-energy light.
In familiar lead-halide perovskites, swapping iodine for bromine generally raises the bandgap. Here, adding bromine could first increase it, then sharply reduce it, before it rose again. Earlier work had seen odd bandgap changes and hints that these mixtures separated. The new study shows how that separation is organized inside individual crystals. Chemical mapping and electron diffraction revealed cores enriched in bromine surrounded by shells enriched in iodine. In the core, atomic building blocks joined into layers, while in the shell they formed isolated pairs.
The layered core confines electrons less tightly, lowering the energy needed to excite them. That explains how it can have a smaller bandgap despite containing more bromine. Growth observations supported a two-stage picture in which bromine-rich material crystallized quickly to form the core, followed by slower growth of an iodine-rich shell. Calculations indicated that separation becomes more favorable as antimony content rises.
The maps also exposed a practical trade-off. The composition with the lowest measured bandgap contained two structural phases, while other recipes offered a single phase. In a 10-day test in humid air, unprotected iodine-rich and antimony-rich regions degraded first, while bromine-rich regions held up. A protective polymer coating greatly reduced deterioration. Mixing bismuth and antimony, the researchers write, offers "a means to balance functionality and stability."
The study did not test a solar cell. Its value is as a map: a favorable absorption reading can be dominated by one part of a crystal and hide a second structure with different properties. Mirkin and co-author Chris Wolverton disclosed financial interests in Mattiq, a company that could benefit from the research.





