The first paper, “A New Approach for Fabricating High-Performance MicroLED Displays” demonstrates how to remove several processing steps altogether when manufacturing full colour microLED-based displays. Instead of successively singulating blue, green and red microLEDs from their native wafers in order to transfer them onto a TFT backplane or to stack them with CMOS drivers themselves singulated and placed on a backplane (four transfer steps or more), the researchers first bonded a blue LED epiwafer onto a full wafer of CMOS driving circuits. They then fabricated the InGaN blue LEDs directly on the bonded epiwafer and then created the RGB pixels at wafer level, depositing adequate conversion phosphors on the blue microLEDs to obtain the green and red emitters.
The result is a full RGB microLED wafer from which they can singulate tens of thousands of RGB microLEDs each already stacked and connected with the underlying CMOS driving circuit. What’s more, the full connectivity provided at wafer level makes it much easier to test the individual RGB-LED-on-CMOS dies just after their fabrication and before their transfer.
These all-in-one RGB microLEDs can then be transferred in one single step onto a simpler backplane merely consisting of a passive grid of conductive lines and columns, eliminating the need for a TFT backplane and its inherent driving and size limitations (low carrier mobilities compared to CMOS and large feature sizes).
“This new process, in the proof-of-concept stage, paves the way to commercial, high-performance microLED displays,” explained François Templier, CEA-Leti’s strategic marketing manager for photonic devices and first author of the paper.
Indeed, one of the biggest challenges to support high resolution microLED displays is to improve the performance of the driving electronics, which require more power and faster electronics to power millions of pixels in a fixed-frame time. Today’s TFTs used in active matrix substrates cannot provide the necessary current and speed supported by microLEDs.
With this approach, the receiving backplane could even be fabricated with non-lithographic techniques such as ink-jet or stencil printing, opening the field to display sizes in excess of 100-inches on almost any kind of material such as glass, plastic, metal; rigid or flexible.
The second paper titled “MicroLED Displays based on Transfer with Microtubes Interconnections” gives the solution for a solder-free interconnect mechanism that would be cost effective to implement on the receiving substrate, regardless of the pixel pitch envisaged during transfer.
As an example, the researchers envisage that the small pitch microLEDs initially fabricated at wafer level could be transferred on a substrate at a much larger pixel pitch, ranging from 50 to 500μm. This means the emitting area would leave plenty of free space within each pixel, which OEMs could use to offer more display transparency or to integrate other display functionalities such as optical sensing.
Here the authors propose to secure both electrical and mechanical connections in one press-fitting operation at room temperature, using microtube interconnections they demonstrated in a previous paper. The microtubes are grown directly on the receiving substrate, consisting of micro-cylinders built over a rigid Tungsten Silicide supporting structure. Here, the MicroLEDs are fabricated upside down, light being emitted towards the epitaxial substrate. And for the purpose of the press-fit connection, both N and P contact pads are created on the top side (the N contact being routed through the active to the top side) so the relatively soft and deformable pads end up being pressed onto the hard microtubes, forming a solid electrical and mechanical connection.