Noam Shapiro | InZiv: Think you can just calibrate your fast PL measurements to predict EL performance? This data suggests otherwise.
00:08:31 - 00:10:01
Other snippets from this talk
Summary of the clip:
Think you can just calibrate your fast PL measurements to predict EL performance? This data suggests otherwise.
A common strategy to leverage the speed of photoluminescence (PL) is to attempt to create a correlation model that maps PL results to the more accurate electroluminescence (EL) data. However, this clip presents compelling evidence from a university study that demonstrates the fundamental inconsistency of PL, making such a correlation unreliable. The study analyzed two microLED chips produced under the exact same epi-wafer and fabrication recipe.
When measuring the peak emission wavelength, the EL results for both Chip A and Chip B were virtually identical, showing process consistency as expected. In stark contrast, the PL measurements for the same two chips showed a significant shift in peak wavelength relative to each other. This demonstrates that subtle, inevitable process variations can affect PL and EL results differently, breaking any simple correlation.
The core issue is the lack of a consistent pattern or predictable offset between PL and EL measurements across different chips, even from the same batch. Because the PL data shifts unpredictably while the EL data remains stable, it's impossible to establish a reliable calibration factor. This invalidates the use of PL as a fast proxy for EL, as the results cannot be trusted to reflect true device performance for binning or quality control.
In this short video, you can learn:
* Why creating a simple correlation between PL and EL data is not feasible.
* Data showing how two identical microLEDs can have stable EL wavelengths but shifting PL wavelengths.
* The impact of process variations on the consistency of PL measurements, making them unreliable for prediction.
๐ **Clip Abstract** This clip debunks the idea of using a simple calibration to make fast PL testing predict accurate EL results. Data shows that even on identically produced microLEDs, PL measurements are inconsistent and do not correlate reliably with stable EL performance, making PL an untrustworthy proxy.
๐ Link in comments ๐
#MicroLEDs, #Photoluminescence, #Electroluminescence, #DeviceCharacterization, #ARdisplays, #DisplayTechnology
This is a highlight of the presentation:
Unleashing microLEDโs Future: The Power of Electroluminescence Testing
MicroLEDs, AR/VR Displays, Micro-Optics 2025: Innovations, Start-Ups, Market Trends
Online | TechBlick platform
Organised By:
TechBlick
MicroLED Connect
More Highlights from the same talk.
00:02:45 - 00:05:13
Why does photoluminescence fail to guarantee functional microLED displays?
Why does photoluminescence fail to guarantee functional microLED displays?
Standard microLED inspection relies heavily on photoluminescence (PL), which excites the active region optically to measure emission properties. However, PL misses critical electrical defects because it does not replicate the final device operating conditions. To capture true performance, manufacturers must employ electroluminescence (EL), which excites the junctions via current injection.
Electroluminescence serves as the gold standard because it provides direct electrical feedback and accurately forecasts how the microLED will perform in the target display backplane. Unfortunately, implementing standard EL on a wafer scale introduces two massive issues: physical contact damage to micro-scale pads and extremely slow probing speeds.
With conventional EL probers testing at a sluggish rate of just one LED per second, completing an inspection run on a single dense microLED wafer would take roughly two entire months. This throughput bottleneck has forced the display industry to rely on suboptimal testing methods, leading to poor yields and skyrocketing manufacturing costs.
In this short video, you can learn:
* The fundamental thermodynamic and physical differences between PL and EL testing
* Why electrical excitation is required to predict final display performance
* The crippling throughput bottlenecks and mechanical damage risks of conventional probe testing
๐ **Clip Abstract** This clip explores the critical differences between photoluminescence and electroluminescence in microLED inspection. It highlights why standard electrical probing is too slow and mechanically risky for volume manufacturing, despite being the only accurate way to predict device performance.
#Electroluminescence, #MicroLEDInspection, #WaferLevelTesting, #ElectricalDefectDetection, #MicroLEDDisplays, #DisplayMetrology
00:07:30 - 00:09:53
Can we scan 9 million microLEDs electrically without touching or destroying the wafer?
Can we scan 9 million microLEDs electrically without touching or destroying the wafer?
Overcoming the speed and contact damage limitations of standard electroluminescence is the Holy Grail of microLED manufacturing. A high-throughput, non-damaging EL wafer inspection system can scan up to 6 million chips per hour. This reduces a full wafer scan time from a commercially non-viable two months down to just one hour.
This rapid testing capability functions by safely exciting and evaluating 1,600 microLEDs every single second. The resulting dataset provides display engineers with critical parameter mapping, including go/no-go functionality, individual pixel emission intensity, precise dominant wavelength, and luminance measurements.
Most importantly, the platform extracts full current-voltage (IV) curves and electrical properties without physical damage to either vertical or flip-chip architectures. Armed with these comprehensive IV characteristics, manufacturers can establish precise pass-fail thresholds to filter out defective dies before mass transfer.
In this short video, you can learn:
* How high-throughput EL testing achieves scanning speeds of 1,600 chips per second
* The critical parameters captured, including dominant wavelength, luminance, and IV curves
* How non-damaging inspection protects delicate contact pads while enabling automated mass production
๐ **Clip Abstract** This clip introduces high-throughput, non-damaging electroluminescence inspection capable of testing 6 million microLEDs per hour. It details how capturing rapid optical and electrical data helps manufacturers secure high transfer yields and lower production costs.
#NonContactElectroluminescence, #MicroLEDMetrology, #WaferLevelInspection, #MassTransferYield, #MicroLEDDisplays, #DisplayManufacturing
00:05:13 - 00:07:09
How many "good" microLEDs are actually dead silent under electrical current?
How many "good" microLEDs are actually dead silent under electrical current?
Relying solely on optical photoluminescence (PL) creates a massive yield trap due to false positives. PL can cause defective microLEDs to glow under external optical excitation, suggesting they are functional when they actually suffer from electrical open circuits or leakage. When these chips proceed to transfer, they fail to illuminate, driving down panel yields.
Scalability compounds this inspection challenge as the industry transitions from legacy LED architectures to micro-scale devices. A typical traditional LED wafer holds around 9,000 chips, whereas a single microLED wafer can contain up to 9 million chips, scaling the inspection overhead by three orders of magnitude.
Furthermore, diverse contact layouts (vertical vs. lateral stacks) and varying substrate sizes (from 4-inch to 12-inch wafers) make uniform testing difficult. Conventional mechanical contact probing at this density leads to severe micro-abrasions and contact pad destruction, visible under atomic force microscopy (AFM).
In this short video, you can learn:
* The phenomenon of PL false positives and how they escape to final assembly
* The massive scaling challenge of testing 9 million microLEDs per wafer
* How physical contact probing causes irreversible damage to micro-contact pads
๐ **Clip Abstract** This segment highlights how optical inspection methods hide critical electrical defects, resulting in costly false positives during microLED production. It also addresses the structural variations and physical damage risks of trying to scale traditional contact probing to 9 million chips per wafer.
#PhotoluminescenceFalsePositives, #MicroLEDMetrology, #ContactProbingDamage, #ElectroOpticalTesting, #MicroLEDDisplays, #ARMicrodisplays