Rick Smith | Pixel-Flo: How does slot-die coating and photolithographic trapping enable high-precision microLED fluidic self-assembly without expensive alignment stages?
04:00 - 05:59
Other snippets from this talk
Summary of the clip:
How does slot-die coating and photolithographic trapping enable high-precision microLED fluidic self-assembly without expensive alignment stages?
The approach is a continuous-flow fluidic self-assembly built on slot-die coating. Standard epi wafers are processed into dies with standard fabrication, then bulk harvested into an ink: a suspension of microLEDs delivered to the substrate. Slot-die coating was chosen because it scales, it is recognizable to the display industry, and it scales up to large area well.
Assembly begins by patterning an array of traps on the surface using standard photolithography. The precision comes from the same photolithography tools used for TFT substrates: that alignment accuracy is borrowed, so the mass-transfer tool needs no high-precision alignment stages. Traps sit on the display glass substrate, and ink is then pumped through the slot die.
Pumping forms a bead on the surface with a meniscus at the trailing edge. Sweeping the ink drags the meniscus across dies not interacting with transfer sites; dies pile up in that trailing edge, raising the interaction rate with traps. When a die enters a trap, the meniscus deforms around it and pins it so it cannot come out. A benchtop proof of concept assembled five-micron dies at 25-micron pitch.
In this short video, you can learn:
* Slot-die coating is used because it scales to large areas and is recognizable to the display industry.
* Photolithographic traps borrow TFT-substrate alignment accuracy, removing the need for high-precision alignment stages in the mass-transfer tool.
* Dies accumulate at the meniscus trailing edge, and the deforming meniscus pins captured dies so they stay in place.
📋 **Clip Abstract** A continuous-flow fluidic self-assembly process uses slot-die coating to deliver a microLED ink across a display glass substrate. Traps patterned by standard photolithography capture dies with TFT-tool alignment precision, and a benchtop proof of concept assembled five-micron dies at 25-micron pitch.
About the speaker:
* Speaker: Rick Smith
* Company: Pixel-Flo
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#SlotDieCoating, #FluidicSelfAssembly, #MicroLEDMassTransfer, #PhotolithographicTraps, #MicroLEDDisplays, #DisplayManufacturing
This is a highlight of the presentation:
A scalability breakthrough for microLED mass transfer using a slot-die coating approach
More Highlights from the same talk.
01:09 - 02:53
Why can't wafer-scale stamp and laser transfer deliver affordable microLEDs on large glass panels?
Why can't wafer-scale stamp and laser transfer deliver affordable microLEDs on large glass panels?
Mass transfer is framed as assembling hundreds of millions of dies at scale, but the bigger problem is a mismatch between the semiconductor and display industries. One industry measures substrates in inches; the other measures substrates in meters. Large panel display fabs make many displays at once on huge sheets of glass. A transfer solution must therefore be a display-industry solution, not merely a semiconductor die-assembly solution.
LCD dominates manufacturing because its manufacturing technology is scalable. That scalability has driven a huge reduction in LCD manufacturing costs through economies of scale. If every process works on large generation glass, including generation 10.5, unit costs fall. OLED has found that harder: pushing scale has reached some generation 8.6 fabs, but it has been a difficult journey to push OLED manufacturing technologies to large scale. That is the problem with microLED.
For microLED, solutions are needed all the way through SAC, mass transfer, and everything else that scales to large panels. That is what will drive costs down, make microLED significantly more affordable, and unlock its advantages. Typical stamp and laser-based transfer processes are limited by wafer size, whether picking up from an epi wafer or a carrier. The mismatch remains: wafer-scale transfer cannot match meter-scale glass.
In this short video, you can learn:
* The core mismatch is that semiconductor substrates are measured in inches while large-panel display substrates are measured in meters.
* LCD manufacturing costs fell because processes scaled to large generation glass, such as generation 10.5, whereas OLED has struggled toward generation 8.6.
* Stamp and laser-based microLED transfer remain limited by wafer size, whether the source is an epi wafer or a carrier.
📋 **Clip Abstract** MicroLED mass transfer is constrained less by assembling hundreds of millions of dies than by a mismatch between inch-scale semiconductor substrates and meter-scale display glass. LCD's scalable large-generation-glass manufacturing drove unit costs down, while OLED's push to generation 8.6 has been difficult, and stamp and laser transfer remain wafer-size-limited.
About the speaker:
* Speaker: Rick Smith
* Company: Pixel-Flo
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#MicroLEDMassTransfer, #LargeGenerationGlass, #StampLaserTransfer, #WaferSizeLimit, #DisplayManufacturing, #MicroLEDDisplays
09:36 - 11:17
Is the notorious EQE roll-off still the limiting factor for single-digit die sizes in microLEDs?
Is the notorious EQE roll-off still the limiting factor for single-digit die sizes in microLEDs?
Now, small die sizes unlock potential cost reductions, but they bring challenges. There are EQE concerns, and we must get enough light out of these devices. Most importantly, we must handle these dies in a mass transfer process and the rest of the manufacturing stack. So smaller die is not without its challenges.
We are used to seeing EQE roll-off with die size. This is well understood: reducing die size for different colors of LEDs reduces overall device EQE. But there has been significant progress. We have seen great reports at this meeting and papers in the last few months showing surface treatment and neutral beam etching keep EQEs high, even at small single-digit die sizes. This is not the same as five or 10 years ago.
So this EQE problem is not the limiting factor. Handling these dies at scale is. That is where continuous flow comes in. With our proof of concept, we have unlocked five-micron dies. We saw increasing EQEs with smaller and smaller dies, including Meta's talk yesterday, and great numbers from phosphide-based red micro-LEDs, right?
In this short video, you can learn:
* Reducing die size for different colors of LEDs reduces overall device EQE, but surface treatment and neutral beam etching can keep EQEs high even at single-digit die sizes.
* The EQE roll-off is not the limiting factor; handling dies at scale is.
* Continuous flow has unlocked five-micron dies in a proof of concept.
📋 **Clip Abstract** The talk addresses EQE concerns for small microLED dies, noting that surface treatment and neutral beam etching have raised EQEs even at single-digit sizes. It argues that EQE is no longer the limiting factor and that continuous flow has unlocked five-micron dies for scalable handling.
About the speaker:
* Speaker: Rick Smith
* Company: Pixel-Flo
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#EQERollOff, #NeutralBeamEtching, #SurfaceTreatment, #FiveMicronDies, #MicroLED, #MassTransfer
07:57 - 09:37
How does shrinking microLED dies from 20 microns to 5 microns and bulk lift-off cut bill of materials?
How does shrinking microLED dies from 20 microns to 5 microns and bulk lift-off cut bill of materials?
Moving to small die halves die height and width, giving four times as many dies on the wafer if treated as pure area scaling at fixed wafer cost. Going from 20 microns down to five microns provides a significant multiplier in dies per wafer at the same wafer cost, significantly unlocking bill of materials reduction. The requirement is that these dies can be handled.
Die count is not only about die size. To make some transfer steps parallel, an integer multiple of die separation on the epi wafer is needed for the eventual display substrate, which can lead to a larger separation in some transfer technologies. With laser transfer, using high energy flux to blast dies off an epi wafer needs good separation, creating pressure on the street widths required to separate dies for laser lift-off.
In this approach, bulk lifting off these dies from the wafer lets more of these small dies be crammed in with smaller street widths. Those street widths are limited only by the fabrication limits of the process rather than by the removal from the substrate. That difference delivers a significant packing gain.
In this short video, you can learn:
* Halving microLED die height and width can put four times as many dies on a wafer at fixed wafer cost.
* Laser lift-off requires wider separation streets because high energy flux blasts dies off the epi wafer.
* Bulk lift-off removes that constraint, limiting street widths only by fabrication limits and delivering a significant packing gain.
📋 **Clip Abstract** Scaling microLED dies from 20 microns to five microns multiplies dies per wafer at the same wafer cost, significantly unlocking bill of materials reduction. Bulk lift-off bypasses laser lift-off street-width pressure, so packing is limited by fabrication limits and yields a significant packing gain.
About the speaker:
* Speaker: Rick Smith
* Company: Pixel-Flo
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#MicroLEDDieScaling, #LaserLiftOffStreetWidth, #BulkLiftOff, #DiesPerWafer, #MicroLEDMassTransfer, #DisplayManufacturing