Vitaly Zubialevich | Tyndall National Institute, University College Cork: Which four micro-LED parameters decide display versus interconnect viability, and why does shrinking the chip destroy efficiency?
01:08 - 03:08
Other snippets from this talk
Summary of the clip:
Which four micro-LED parameters decide display versus interconnect viability, and why does shrinking the chip destroy efficiency?
Four parameters govern whether a micro-LED suits displays or data links. Efficiency sets battery time in micro-displays. Energy per bit defines optical interconnect performance. Emission directionality decides what fraction of emitted light the system usefully utilises. Switching speed is already sufficient for micro-displays but very important for data transmission, and even slow-and-wide need not be particularly slow. Highest achievable brightness matters for micro-displays fighting daylight background, yet is less critical for interconnects.
Brightness belongs to the material systems being worked on, not to the particular designs proposed, so the other parameters carry more relevance to architecture. Meanwhile micro-LEDs lose efficiency as size shrinks, and the push is toward micron-scale devices or below. The infamous culprit is surface recombination at the chip edges: the smaller the emitter, the larger the share of its volume that edge represents.
Traditional remedies for edge recombination are surface treatment, passivation, wet edge, regrowth, or some combination of the above. A more radical solution is to adopt architectures involving core-shell structures, which avoid the surface in the first place, so the problem should be mitigated. That reframes the design question: can geometry, rather than post-processing, decide how much light escapes usefully and how fast the device switches?
In this short video, you can learn:
* Efficiency sets battery time in micro-displays, while energy per bit is the metric that matters for optical interconnects.
* Emission directionality determines what fraction of emitted light a system can usefully utilise, and switching speed is already sufficient for micro-displays but vital for data transmission.
* Shrinking micro-LEDs to micron scale or below sacrifices efficiency to surface recombination at the edges, which core-shell architectures aim to avoid entirely.
📋 **Clip Abstract** Micro-LED viability splits along four parameters: efficiency, energy per bit, emission directionality and switching speed, with brightness a material-system property that matters mostly for displays fighting daylight. Efficiency collapses as chips shrink toward micron size because of edge surface recombination, driving interest in core-shell architectures that never expose a surface.
About the speaker:
* Speaker: Vitaly Zubialevich
* Companies: Tyndall National Institute, University College Cork
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#MicroLED, #SurfaceRecombination, #CoreShell, #EmissionDirectionality, #OpticalInterconnects, #MicroDisplays
This is a highlight of the presentation:
A Micro-LED Architecture for Enhanced Emission Directionality in Displays and Optical Interconnects
More Highlights from the same talk.
06:28 - 08:50
How do you build a self-aligned nanocolumn micro-LED with a built-in pyramidal micro-reflector, step by step?
How do you build a self-aligned nanocolumn micro-LED with a built-in pyramidal micro-reflector, step by step?
The Gainable architecture centres on a nano-column of one conductivity type, here N-type, with passivation of the floor at its feet. A self-centred cap, called an umbrella, sits atop. The active region is organised on the nano-column sidewalls. A pyramidal, slanted P-type shell—in this case—acts as a built-in micro-reflector. Shown as a 3D schematic in III-nitrides, the structure is also realised fully grown.
Fabrication starts by defining micro-LED positions. A top-down etch into gallium nitride forms the columns; this defines very homogeneous tops. A wet etch removes damaged material. AlGaN is overgrown, then removed everywhere except the sidewalls. A gallium nitride reflow creates the nano-umbrella, requiring undercut, followed by sidewall verticalisation via wet etch. The umbrella then passivates everything except the sidewalls. The device is ready for active-region growth and micro-reflector formation.
The devices were fabricated. However, with the available resources, individual addressing of this pixel was not possible, so microscopic devices were made instead. This means the demonstration used microscopic devices rather than an individually addressable array. The architecture still includes the self-centred umbrella cap, sidewall active region, and pyramidal P-type shell that provides the built-in micro-reflector.
In this short video, you can learn:
* The Gainable micro-LED uses a nano-column, floor passivation, a self-centred umbrella cap, sidewall active region, and a pyramidal P-type shell as a built-in micro-reflector.
* Fabrication involves top-down gallium nitride etching, wet etch damage removal, AlGaN overgrowth and sidewall removal, gallium nitride reflow for undercut, sidewall verticalisation, and umbrella-based passivation.
* Devices were fabricated, but resource limits prevented individual pixel addressing, so microscopic devices were produced instead.
📋 **Clip Abstract** A micro-LED architecture called Gainable uses a self-aligned nano-column with an umbrella cap, sidewall active region, and a pyramidal P-type shell acting as a built-in micro-reflector. Fabrication combines top-down etching, wet etch cleaning, AlGaN overgrowth, gallium nitride reflow for undercut, and passivation, though resource constraints led to microscopic devices rather than individually addressable pixels.
About the speaker:
* Speaker: Vitaly Zubialevich
* Companies: Tyndall National Institute, University College Cork
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#GalliumNitride, #AlGaNOvergrowth, #NanoUmbrella, #PyramidalMicroreflector, #MicroLED, #OpticalInterconnects
09:52 - 11:31
A 28° slanted facet instead of 45°: how much directional extraction gain survives the geometry penalty?
A 28° slanted facet instead of 45°: how much directional extraction gain survives the geometry penalty?
The P-type work completely failed, and the active region efficiency is probably not great, but the structure is still evaluated for light extraction and directionality. In a planar structure, the escape route is vertically up and vertically down, two solid angles, assuming the top contact is fully reflective. Adding 45° slanted facets changes the escape route to a larger belt-shaped solid angle, and the advantage can be quite large.
For 45° facets, calculating alpha four degrees, which corresponds to plus or minus 10° outside the structure, gives an advantage of 90 times. In the actual case, the angle is not 45° but 28°, and the escape route splits into two belt-like regions. Recalculating everything still gives an advantage around 25 times. That is the directional extraction gain reported for this geometry.
There is a price, however. Only the upper part of the nano-column can benefit from the sidewalls. Light emitted at a right angle will never reach the sidewall, so only the upper part benefits. The speaker also notes uncertainty about whether the laser works, while the upper nano-column remains the region that benefits.
In this short video, you can learn:
* Planar escape routes are only vertically up and vertically down, with the top contact assumed fully reflective.
* A 45° slanted facet creates a belt-shaped escape solid angle and gives a 90 times advantage for alpha four degrees, i.e. ±10° outside the structure.
* At the real 28° facet angle the escape route splits into two belt-like regions and the recalculated advantage is around 25 times, but only the upper part of the nano-column benefits from the sidewalls.
📋 **Clip Abstract** The P-type approach completely failed and active-region efficiency is probably not great, but the architecture is assessed for light extraction and directionality. For a 45° facet the calculated advantage is 90 times for ±10° outside the structure, while the actual 28° facet splits the escape route into two belt-like regions and still yields around 25 times, with only the upper nano-column benefiting from the sidewalls.
About the speaker:
* Speaker: Vitaly Zubialevich
* Companies: Tyndall National Institute, University College Cork
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#SlantedFacet, #NanoColumn, #LightExtraction, #DirectionalEmission, #MicroLED, #OpticalInterconnects