top of page

Stefan Chen

Mikro Mesa

* All members of the platform can watch the entire presentation.

 

Please register to become a member.

Stefan Chen | Mikro Mesa: Why is the industry moving away from flip-chip to vertical structures for sub-10 micron microLEDs?

00:01:16 - 00:02:03

Other snippets from this talk

Summary of the clip:

Why is the industry moving away from flip-chip to vertical structures for sub-10 micron microLEDs?

As microLED display pixels shrink below 10 micrometers, traditional flip-chip architectures face severe performance and manufacturing bottlenecks. Flip-chip designs suffer from pronounced current crowding, complex electrode alignment requirements, and declining light extraction efficiency at ultra-small dimensions.

Stefan Chen of Mikro Mesa argues that a vertical chip structure is the only viable pathway for true mass-production microLED displays. By utilizing a vertical architecture, manufacturers can achieve highly uniform current density across the emitter, eliminate current crowding, and simplify the backplane electrical contact routing.

Furthermore, the vertical layout significantly enhances light extraction efficiency and provides a wider alignment tolerance during mass transfer. This geometric advantage is critical when handling 3-micrometer emitters, enabling higher-throughput manufacturing without sacrificing electro-optical performance.

In this short video, you can learn:
* The physical limits of flip-chip architectures at sub-10 micrometer dimensions
* Why vertical structures solve the current crowding phenomenon in micro-emitters
* How vertical designs improve light extraction efficiency and alignment tolerances during transfer

📋 **Clip Abstract** This clip explains why vertical microLED structures outperform traditional flip-chip configurations at sub-10 micrometer scales. Stefan Chen highlights how vertical current distribution prevents crowding and eases the high-precision alignment bottlenecks of mass transfer.

🔗 Link in comments 👇

#VerticalMicroLED, #CurrentCrowding, #MassTransfer, #Sub10Micron, #MicroLEDDisplays, #ARDisplays

This is a highlight of the presentation:

​

Mini- & Micro-LED Displays 2023: Markets, Manufacturing Innovations, Applications, Promising Start-ups

​

​

TechBlick Platform | Online

Organised By:

TechBlick

​

More Highlights from the same talk.

08:01 - 10:13

Why does a smaller µLED current of 1.9 µA need only 1.4 V TFT VDS while 45 µA needs 6.8 V?

Why does a smaller µLED current of 1.9 µA need only 1.4 V TFT VDS while 45 µA needs 6.8 V?

The driving TFT regulates the microLED current, so the relevant SPICE output is drain current ID versus VDS. An oxide TFT is simulated with mobility 20, a 10 µm channel and an 8 µm channel length. The operating VDS is taken where the characteristic has just moved from the linear region into saturation, marked by the knee; at that point the drain current is ideally equal to the microLED current.

For the two microLED sizes, the required currents are 1.9 µA and 45 µA. Reading the same simulated ID–VDS figure gives VDS values of 1.4 V and 6.8 V respectively. The smaller microLED therefore has the lowered VDS. The clip states plainly that VDS is proportional to R, and the example links the microLED current level to the TFT's saturation-knee operating point rather than an arbitrary supply voltage.

The comparison turns the simulation into a design check: pick the microLED current, locate that current on the oxide TFT's ID–VDS curve, and read the VDS at the linear-to-saturation transition. The resulting 1.4 V and 6.8 V values show how device size changes the TFT bias point. The clip then says the equations will illustrate this further, pointing to resistance as the link between microLED size, current and VDS.

In this short video, you can learn:
* The TFT's VDS is selected at the knee where its ID–VDS characteristic leaves the linear region and enters saturation.
* MicroLED currents of 1.9 µA and 45 µA correspond to read VDS values of 1.4 V and 6.8 V.
* The smaller microLED has the lower VDS, and the clip states that VDS is proportional to R.

📋 **Clip Abstract** The clip walks through a SPICE simulation of an oxide TFT used to regulate microLED drive current, plotting drain current against VDS. For microLED currents of 1.9 µA and 45 µA, the selected saturation-knee VDS values are 1.4 V and 6.8 V, showing the smaller device's lower VDS and the proportionality to resistance.

About the speaker:
* Speaker: Stefan Chen
* Company: Mikro Mesa
* Event: MicroLED Connect 2024
* Location: High Tech Campus, Eindhoven, Netherlands

Join the most event MicroLED, AR/VR display and Optical I/O Connect event of the year
Learn more https://www.microledconnect.com/

#OxideTFT, #SPICESimulation, #DrainCurrentVDS, #SaturationKnee, #MicroLEDDisplays, #ThinFilmTransistors

15:55 - 17:43

What does the 5 µm µLED power advantage mean at watch-panel resolution?

What does the 5 µm µLED power advantage mean at watch-panel resolution?

Per pixel, the micro LED draws 9.5 µW or 20.4 µW depending on the case presented. Scaling that to a watch panel of 502 by 410 pixels at a 3.6 V forward voltage gives total panel power of 1.9 and 4.2 respectively. Two caveats are stated: one loss term is not included, and the numbers are computed at peak brightness, so real consumer average power will be lower.

The mechanism is geometric. VDS scales with the voltage drop, and that drop is proportional to the lateral size of the micro LED. The same logic applies to the driving TFT: the voltage dropped across it is also a loss. Hence the rule offered — the smaller the ratio of VDS to VDD, the better, because that ratio is what sets the power consumption.

Two challenges remain for the smaller device: cost and power. At 5 µm the micro LED is argued to be cost competitive, to need lower driving current, to show a lower voltage drop and therefore to consume less power. That is why the 5 µm die is the one being pushed, rather than the larger pixel that inflates the VDS term.

In this short video, you can learn:
* Per-pixel micro LED power of 9.5 µW and 20.4 µW, scaled to a 502×410 watch panel at 3.6 V, yields total panel power of 1.9 and 4.2 at peak brightness.
* VDS is proportional to the micro LED's lateral size and to the voltage drop, and the driving TFT's own drop counts as loss, so a smaller VDS/VDD ratio means lower power.
* The 5 µm micro LED is presented as cost competitive, with lower driving current, lower voltage drop and lower power consumption.

📋 **Clip Abstract** Per-pixel micro LED power of 9.5 µW and 20.4 µW is projected onto a 502×410 watch panel at 3.6 V to give total panel power of 1.9 and 4.2 at peak brightness. Because VDS scales with lateral die size and with the voltage drop — including the driving TFT's own drop — moving to a 5 µm emitter is argued to cut both cost and power.

About the speaker:
* Speaker: Stefan Chen
* Company: Mikro Mesa
* Event: MicroLED Connect 2024
* Location: High Tech Campus, Eindhoven, Netherlands

Join the most event MicroLED, AR/VR display and Optical I/O Connect event of the year
Learn more https://www.microledconnect.com/

#MicroLEDPixelPower, #VoltageDropScaling, #DrivingTFTLoss, #FiveMicrometerMicroLED, #MicroLEDDisplays, #DisplayPowerBudget

13:23 - 15:39

Does a larger µLED really save power once PWM duty cycle is included?

Does a larger µLED really save power once PWM duty cycle is included?

Comparing two sets of micro-LEDs at the same brightness gives 9.5 and 468 microwatts respectively. The 468 µW figure is meaningless on its own, because PWM is not considered in it. The two sets are driven under completely different duty cycles, so the raw supply power cannot be used as a like-for-like comparison between a 5 µm emitter and a 15 by 30 µm one.

The explanation is in the equal-brightness chart: the same rectangular bar of the same block area represents the same brightness. If the 5 µm LED runs at 100% duty cycle, the green 15 by 30 µm bar must be shortened in time, giving a 4.3% duty cycle. The larger emitter's current is 23 times higher, so the bar is compressed accordingly.

Recalculating power with the PWM duty cycle applied brings the 15 by 30 µm device down to 20.4 microwatts. That is still higher than 9.5 µW. The residual gap comes from the voltage drop: the voltage drop on the regulator driving the larger micro-LED is much higher than when driving the 5 µm one, so duty-cycle correction alone does not make the bigger emitter cheaper in power.

In this short video, you can learn:
* Raw micro-LED supply power figures are not comparable unless the PWM duty cycle is quoted alongside them.
* At equal brightness a 15 by 30 µm emitter needs only a 4.3% duty cycle against 100% for a 5 µm device.
* Correcting for PWM brings the larger device to 20.4 µW, still above 9.5 µW because of higher regulator voltage drop.

📋 **Clip Abstract** Two micro-LED sets delivering identical brightness draw 9.5 and 468 microwatts, but the larger number ignores PWM duty cycle. Applying a 4.3% duty cycle to the 15 by 30 µm device cuts it to 20.4 µW, yet higher regulator voltage drop keeps it above the 5 µm reference.

About the speaker:
* Speaker: Stefan Chen
* Company: Mikro Mesa
* Event: MicroLED Connect 2024
* Location: High Tech Campus, Eindhoven, Netherlands

Join the most event MicroLED, AR/VR display and Optical I/O Connect event of the year
Learn more https://www.microledconnect.com/

#PWMDutyCycle, #RegulatorVoltageDrop, #MicroLEDCurrentDensity, #EqualBrightnessComparison, #MicroLEDDisplays, #MobileDisplayPower


Other talks by Stefan Chen


More highlights from Stefan Chen

More Snippets
CONTACT US

KGH Concepts GmbH

Mergenthalerallee 73-75, 65760, Eschborn

+49 17661704139

admin@techblick.com

TechBlick is owned and operated by KGH Concepts GmbH

Registration number HRB 121362

VAT number: DE 337022439

  • LinkedIn
  • YouTube

Sign up for our newsletter to receive updates on our latest speakers and events AND to receive analyst-written summaries of the key talks and happenings in our events.

Thanks for submitting!

© 2026 by KGH Concepts GmbH

bottom of page