Researchers from the Hong Kong University of Science and Technology (Guangzhou), led by Prof. Yunda Wang, have developed a novel microLED transfer process, based on a dynamically programmable transfer head that uses localized heating to control a polymer's stickiness.
The researchers say that this new tool's ability to selectively handle a diverse range of geometries addresses a critical bottleneck in building complex microsystems. The researchers proved their system could selectively sort and transfer functioning 45-by-25-micrometer microLEDs to form custom layouts without degrading their performance.
As part of the research, the researchers successfully transferred semiconductor chips, 90-nanometer-thick copper films, and perfectly spherical 50-micrometer polystyrene beads. The components were placed with extreme precision, showing a positional drift of less than 0.7 micrometers and a rotational error under 0.04 radians.
To build the system, Wang and his colleagues formulated a specific polymer that undergoes a sharp physical transition, changing from a rigid plastic to a rubbery state at exactly 44 °C. They coated a 30-micrometer-thick layer of this polymer over an array of individually controllable microheaters.
During the transfer process, the stamp presses against an array of components. The team activates specific heaters, melting targeted 50-micrometer spots of the polymer in about 60 milliseconds so it conforms to the selected chips. As the polymer naturally cools down in roughly 40 milliseconds, it hardens and physically locks onto the chips. To drop the components at their new destination, the heaters are triggered again, softening the polymer to release its grip. This temperature-driven mechanism provides a pickup-to-release adhesion strength ratio higher than 190 to 1.
The team is now investigating ways to scale up the microheater array. This poses a challenge because densely packed heaters can cause thermal crosstalk, where heat bleeds into adjacent pixels. To overcome this, the researchers plan to use thinner polymer layers and introduce active-matrix driving circuits — similar to the architecture used in commercial flat-screen televisions — to manage massive arrays without requiring overwhelmingly complex wiring.
