Jong-Jan (JJ) Lee | eLux Inc: Why does mass-transfer randomization of microLEDs increase display costs by 30%, and how does fluidic assembly solve it?
00:01:05 - 00:02:45
Other snippets from this talk
Summary of the clip:
Why does mass-transfer randomization of microLEDs increase display costs by 30%, and how does fluidic assembly solve it?
Traditional sequential mass-transfer processes like mechanical pick-and-place or laser-assisted transfer suffer from systematic signature defects, commonly known as "stamp mura." To eliminate these visible spatial patterns, manufacturers must implement complex randomization algorithms during transfer, which typically escalates total display manufacturing costs by up to 30%. Fluidic self-assembly (FSA) naturally bypasses this limitation.
Because FSA relies on suspended microLEDs randomly moving and settling in a fluid medium, the physical distribution of LEDs from different wafer zones onto the display panel is inherently randomized. This natural mixing relaxes the strict, costly requirements for extremely tight wavelength binning on the epitaxy wafer. Even with a wavelength variation of 10 nanometers, FSA-assembled panels exhibit exceptional visual uniformity without localized mura.
Furthermore, sequential methods face scaling limitations as display panel size increases, requiring hundreds of hours to transfer millions of dies for larger displays. FSA acts as a massively parallel transfer technology, enabling simultaneous assembly across large-format glass substrates such as Generation 6 or larger. This transition from sequential mechanical alignment to fluidic parallel settling significantly lowers capital equipment constraints.
In this short video, you can learn:
* The cost penalties of die randomization in sequential transfer versus the natural randomization of fluidic self-assembly.
* How fluidic self-assembly relaxes epitaxial MOCVD wavelength uniformity requirements to 10 nm while maintaining a mura-free display.
* The scalability limits of sequential mass transfer on large-format substrates compared to parallel fluidic processing.
š **Clip Abstract** Fluidic self-assembly offers a natural, mura-free randomization process that eliminates stamp-induced optical defects without the 30% cost penalty associated with mechanical randomization. By shifting from sequential laser or stamp transfer to fluidic parallel settling, display makers can scale production to Gen 6 substrates while utilizing wider wavelength binning.
š Link in comments š
#FluidicSelfAssembly, #MassTransfer, #StampMura, #WavelengthBinning, #MicroLEDDisplays, #DisplayManufacturing
This is a highlight of the presentation:
More Highlights from the same talk.
00:05:31 - 00:07:05
How does your choice of mass-transfer technology limit your microLED wafer yield by up to 8x?
How does your choice of mass-transfer technology limit your microLED wafer yield by up to 8x?
The commercial viability of microLED displays is heavily dictated by wafer utilization, which is often limited not by the physical size of the die, but by the minimum spacing required between devices on the donor wafer. Traditional transfer techniques, such as stamp-based pick-and-place or laser-assisted transfer, require relatively large kerf or street spacing (typically 15 microns or more) to accommodate the mechanical limits of the transfer tool or laser spot. This limitation severely restricts the total number of dies that can be packed onto a single wafer.
In contrast, fluidic self-assembly (FSA) decouples wafer dicing and extraction from the tool's mechanical alignment tolerances. Because the microLEDs are harvested directly into a fluid solvent, the die spacing on the growth wafer is limited only by the precision of the photolithography and semiconductor etching processes. This allows street widths to be reduced to as little as 1 to 5 microns at the wafer level.
The financial implications of this spacing difference are stark. For example, on a standard six-inch wafer utilizing 18x36 micron microLEDs with a 15-micron mechanical stamp spacing, yield is limited to approximately 6 million LEDs. Utilizing FSA with a 30-micron circular die and a 5-micron chemical etch spacing increases the die count to 50 million LEDs per wafer, dramatically dropping the cost per die.
In this short video, you can learn:
* How mechanical transfer tools impose strict die-spacing limits on wafers, capping maximum device density.
* The photolithographic and wet/dry etch limitations that govern microLED wafer utilization in fluidic self-assembly.
* A direct yield comparison showing an 8-fold increase in dies per 6-inch wafer by optimizing kerf spacing for fluidic harvesting.
š **Clip Abstract** Mechanical stamp and laser transfer methods constrain microLED wafer density by requiring wide die-to-die spacing for tool clearance. Fluidic self-assembly overcomes this by harvesting microLEDs directly from wafer-level etch steps with spacing as narrow as 1 to 5 microns, raising 6-inch wafer yields from 6 million to 50 million dies.
š Link in comments š
#FluidicSelfAssembly, #MicroLEDMassTransfer, #KerfSpacing, #WaferYield, #MicroLEDDisplays, #AdvancedPackaging
00:10:00 - 00:11:40
Can gravity and fluid dynamics alone guarantee 100% correct orientation of asymmetric microLEDs in substrate wells?
Can gravity and fluid dynamics alone guarantee 100% correct orientation of asymmetric microLEDs in substrate wells?
One of the fundamental challenges in fluidic self-assembly is ensuring that microscopic, asymmetric dies settle into their designated substrate wells with the correct orientation. To achieve this without active electrical steering, eLux utilizes a passive orientation mechanism based on gravity and fluid drag. The microLEDs are fabricated as circular discs with an integrated post on the side opposite the contact electrodes, creating an asymmetric physical structure.
During the fluidic assembly process, the microLED suspension flows across the panel surface. If a die settles with its post facing downward against the panel surface, the fluid velocity gradient exerts a high drag force on the protruding body of the die, making this orientation unstable. The fluid flow naturally flips the microLED over so that the post points upward into the stream, which is a highly stable orientation that allows the die to glide smoothly along the substrate surface.
Once the properly oriented die slides over an empty well, gravity pulls it into place, matching its concentric center and edge contacts with the matching electrodes patterned inside the well. Subsequent thermal processing triggers a eutectic bonding reaction between the microLED and substrate electrodes. Cross-sectional scanning electron microscopy (SEM) confirms that this method yields clean, highly uniform, and mechanically stable electrical interfaces.
In this short video, you can learn:
* The asymmetric physical design of microLEDs featuring integrated orientation posts to harness fluid drag.
* Fluid dynamics gradients that automatically flip incorrectly oriented microLED dies into stable settling states.
* The registration of concentric electrodes inside substrate wells and their subsequent eutectic bonding.
š **Clip Abstract** The eLux fluidic transfer process relies on disk-shaped microLEDs featuring an asymmetric post that interacts with fluid drag to automatically correct die orientation. Once flipped into a stable state, gravity drops the microLEDs into substrate wells where concentric contacts align perfectly for eutectic bonding.
š Link in comments š
#FluidicSelfAssembly, #AsymmetricMicroLEDs, #FluidDragOrientation, #EutecticBonding, #MicroLEDDisplays, #MassTransfer




