Zhanxiang Zhao | Oxford Instruments: Can photoresist alone mask GaN mesa and isolation etching without losing profile control or sidewall smoothness?
06:02 - 07:44
Other snippets from this talk
Summary of the clip:
Can photoresist alone mask GaN mesa and isolation etching without losing profile control or sidewall smoothness?
Mesa etching for GaN was demonstrated using a photoresist mask alone, so no oxide layer is needed and the step sequence stays simple. The process delivers both vertical and slow-slope profiles where each is required, with smooth sidewalls and smooth bottoms. Bias is controlled very carefully during the etch to limit damage to the mesa.
Isolation etching is a considerably deeper etch, so the process is designed with a rate high enough to give good throughput, which lowers the overall cost. The profile can again be adjusted, and both the mesa and isolation processes were developed together, with requirements that are slightly different for each.
Both dry etch processes were developed to run with either a photoresist mask or an oxide mask. The requirements of the two differ slightly, but a single development covers both, which keeps the flow simple. The sidewall can also be etched further to remove any roughness issue after the main etch.
In this short video, you can learn:
* Photoresist alone can mask GaN mesa etching, taking the oxide layer out of the step sequence.
* Bias is tuned carefully during the mesa etch to limit damage while holding vertical or slow-slope profiles with smooth sidewalls and bottoms.
* The isolation etch is deep, so a high rate is chosen for throughput and lower overall cost, with the profile still adjustable.
π **Clip Abstract** GaN mesa and isolation etches were developed for both photoresist and oxide masks, with the mesa process demonstrated on photoresist alone. Results show selectable vertical or slow-slope profiles, smooth sidewalls and bottoms, careful bias control to limit mesa damage, and a high-rate deep isolation etch that keeps throughput up and cost down.
About the speaker:
* Speaker: Zhanxiang Zhao
* Company: Oxford Instruments
* 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/
#GaNMesaEtching, #PhotoresistMasking, #SlowSlopeProfile, #DeepIsolationEtch, #MicroLEDFabrication, #DryEtchProcess
This is a highlight of the presentation:
Dry etching technique for microLED applications
More Highlights from the same talk.
08:01 - 09:24
Why does AlInGaP demand a hot etch, and what does that cost in mask choice and throughput?
Why does AlInGaP demand a hot etch, and what does that cost in mask choice and throughput?
AlInGaP is different because it contains a high amount of indium, and indium is not very volatile from a chemistry viewpoint. That is why the material needs a hot etch process to etch it nicely. The result shown for this method achieved a change size below two micrometres, with a smooth surface and sidewall.
Because the process is hot and must deliver a high etch rate for throughput, mask choice is constrained: only an oxide mask can be used. The PP 100 cover 300 cannot handle a hot process, so the system is still used to etch this material. Profile and sidewall smoothness still need control.
For the AlInGaP isolation etch, the feature is quite deep, so a high rate is needed to deliver high throughput. The hot process again limits the mask to oxide. Besides MEA and isolation for AlInGaP, another example also used in microLED reinforces that profile control and sidewall smoothness are key etch outputs.
In this short video, you can learn:
* Indium in AlInGaP is not very volatile, so a hot etch is required to etch the material nicely.
* The hot process forces an oxide mask and demands high etch rate for throughput.
* The demonstrated etch achieved change size below two micrometres with smooth surface and sidewall.
π **Clip Abstract** AlInGaP needs a hot dry etch because its high indium content is not volatile, and the presented method achieved change size below two micrometres with smooth surfaces and sidewalls. The hot process constrains the mask to oxide and requires high etch rate for throughput, while profile and sidewall smoothness remain the control targets for isolation and other microLED etches.
About the speaker:
* Speaker: Zhanxiang Zhao
* Company: Oxford Instruments
* 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/
#AlInGaP, #DryEtching, #OxideMask, #SidewallSmoothness, #MicroLED, #CompoundSemiconductors
12:39 - 14:45
Hexagonal pits and spontaneous nanocolumns after GaN etch: process artifact or epi-quality signature?
Hexagonal pits and spontaneous nanocolumns after GaN etch: process artifact or epi-quality signature?
Asked whether the hexagonal repeat pits are inherent to the etch or to the material, the answer is that both are possible. The first requirement is that the epi is grown very nicely. The same recipe was applied to epi of different quality: when epi quality improved, most of the pits disappeared. So epi quality dominates the starting surface.
The recipe also matters. Pits can be reduced by focusing on the RF powers used and by adjusting the gas ratios, giving a second knob to control them. That means pit density is not fixed by the epi alone; the etch conditions shift the result. Process and material contributions are therefore separable, and both should be tuned together.
Spontaneous nanocolumns instead of pits appear where parts of the surface are not etched at all. The cause is residues that were probably not noticed before etching; they behave like a micro-mask, blocking the etch locally. Inspecting pre-etch samples is therefore critical, and the surface must be verified very nice and clean beforehand.
In this short video, you can learn:
* The same etch recipe applied to better-quality epi removed most of the hexagonal pits, so epi quality is a primary driver.
* Pit density can also be tuned on the tool side through the RF powers used and by adjusting the gas ratios.
* Unetched nanocolumns come from residues acting as a micro-mask, making pre-etch surface inspection critical.
π **Clip Abstract** Hexagonal pits after GaN etching trace back to both epi quality and etch recipe: improved epi removed most pits, while RF powers and gas ratios provide a second control knob. Spontaneous nanocolumns instead come from residues micro-masking the surface, so pre-etch cleanliness must be verified.
About the speaker:
* Speaker: Zhanxiang Zhao
* Company: Oxford Instruments
* 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/
#GaNEtching, #HexagonalPits, #ResidueMicroMasking, #NanocolumnDefects, #MicroLEDFabrication, #PlasmaEtching
About this talk
β Dry etching technique for microLED applications Β· Zhanxiang Zhao Β· MicroLED Connect 2024
