Million Crystal Chips Reveal Core Shell Semiconductor Structures
The bandgap bottomed at 2.11 electron volts near 30 percent bromine, below every parent compound, because the crystals split into cores and shells as they grew.
The recipe didn't predict the property. The architecture inside each crystal did.
Northwestern researchers grew one million crystals on a two-centimeter chip and found bromine-rich cores wrapped in iodine-rich shells.
Olivier Chevalier, Chad Mirkin and colleagues at Northwestern University patterned a parylene C template with about a million wells at 20-micrometer spacing on glass chips two centimeters square. They sprayed precursor solutions from the four corners, so the bromine-to-iodine ratio varied along one axis and the antimony-to-bismuth ratio along the other, then heated the chips to grow crystals about 10 micrometers wide and one micrometer thick. Four chips covered the range. The material is a lead-free family called quaternary perovskitoids, built from cesium, antimony, bismuth, bromine and iodine. The work is published in Advanced Materials.
Crystal size is what makes the measurement possible. Earlier arrays used nanocrystals, which carry only a few nanograms of material per square millimeter, too little for X-ray diffraction. These microcrystals carry up to 0.3 micrograms per square millimeter. That loading let the team run diffraction, photoelectron spectroscopy and light absorption on the same chip, and tie each position's crystal structure to its bandgap, the energy threshold a semiconductor needs to absorb light.
The bandgap refused to follow the recipe. In lead-halide perovskites, swapping iodine for bromine widens the gap in a steady line. In these crystals, the gap first rose to about 2.4 electron volts as bromine increased, then fell to about 2.15, and bottomed at 2.11 electron volts near 30 percent bromine. That minimum sits below every parent compound at the corners of the chip.
Chemical mapping and electron diffraction explain the dip. The mixed crystals don't stay mixed. Bromine-rich material crystallizes fast and forms a core of connected two-dimensional sheets. Iodine-rich material grows more slowly around it as a shell of isolated zero-dimensional pairs. Electrons are confined less tightly in the sheets, so the core absorbs at lower energy and drags the measured gap down even while the crystal's overall bromine content rises. The separation grows stronger as the antimony share increases.
The narrowest gap also isn't the toughest material. In a 10-day test in humid air, iodine-rich and antimony-rich regions degraded first, bromine-rich regions held up, and a protective polymer coating cut the damage sharply. The 2.11 electron volt minimum belongs to a two-phase crystal. A single-phase, bromide-rich composition measured 2.21 electron volts, with stability the mixed crystals didn't show.
The limits belong in the record. The team built no solar cell and no LED, so device performance is untested, and the maps average signals from neighboring crystals rather than reading all one million individually. Two authors, Mirkin and Chris Wolverton, disclosed financial interests in Mattiq, a company that could benefit from the research, and Northwestern disclosed related intellectual property.
The conversation.
Loading comments…
A little more perspective.
Reading is free. Verified members can comment and reply for $2 a month.