Kai Keller | Notion Systems: Can an inkjet-printed hardmask survive 30 minutes in HF/HNO3 with Cpk 5?
23:13 - 24:43
Other snippets from this talk
Summary of the clip:
Can an inkjet-printed hardmask survive 30 minutes in HF/HNO3 with Cpk 5?
For power semiconductors, the required resolution can align with inkjet because structures are considerably larger than typical chips, IC substrates or ICs on wafers. The application is direct digital printing of the hardmask, where the resolution requirement is defined by those larger power-device structures. The open question is whether that printed mask can survive the etch.
The process must sustain a very long exothermic etch: 20 minutes, 30 minutes, in HF HNO3, described as really harsh environments. After etch, the hardmask must survive solvent or alkaline strip. The reported result is very high reliability and yields due to the digital process.
For this application, the client measured Cpk values. Typical requirements sit around 1.6 for a three-sigma process. The measured Cpk for the inkjet process was five or sixβ5.0, not 1.5βso it would be a 5S or 6S process. That is the reported Cpk result.
In this short video, you can learn:
* Inkjet resolution is sufficient for power-semiconductor structures that are considerably larger than typical chips, IC substrates or ICs on wafers.
* The printed hardmask process sustains 20β30 minute exothermic etches in HF HNO3 and solvent or alkaline strip.
* The client measured Cpk of five or six, versus typical requirements around 1.6 for three sigma.
π **Clip Abstract** Inkjet can directly print hardmasks for power semiconductors where structures are larger than typical chips, IC substrates or ICs on wafers. The process survives 20β30 minute HF HNO3 etches and solvent or alkaline strip, with client-measured Cpk of five or six versus typical 1.6 requirements.
About the speaker:
* Speaker: Kai Keller
* Company: Notion Systems
* Event: Electronics RESHAPED 2026 | Berlin
* Location: Estrel, Berlin
USA: The Future of Electronics RESHAPED | USA
π
Save the date: 9 and 10 June 2027
π Mountain View, California (Silicon Valley)
π Info: https://www.techblick.com/electronicsreshapedusa
Europe
The Future of Electronics RESHAPED | Berlin
π
Save the date: 20 and 21 October 2027
π Berlin, Germany
π Info: https://www.techblick.com/electronicsreshaped
#InkjetHardmask, #PowerSemiconductor, #HFHNO3Etch, #Cpk5, #PrintedElectronics, #SemiconductorManufacturing
This is a highlight of the presentation:
Inkjet Printing of Functional Materials: Understanding the new tool in the box of Semiconductor Manufacturing
More Highlights from the same talk.
12:24 - 13:19
Why can't printheads with Β±10% drop variation meet display MURA specs of Β±1%?
Why can't printheads with Β±10% drop variation meet display MURA specs of Β±1%?
Display requirements are unforgiving: pixel-to-pixel volume variations must stay below plus/minus 1% to avoid visible defects such as MURA. Any deviation at that level becomes something the eye can detect on the display. That is the tolerance the printing process has to work within, and it is set by what the viewer can see.
Against that, typical printhead manufacturers' spec sheets state drop-to-drop variations that can easily reach plus/minus 10%. So there is a mismatch β a tenfold gap between what the display demands and what the printhead datasheet offers. The head alone cannot deliver the spec, so the process has to compensate for it.
The answer is drop volume tuning β deliberately adjusting the drop volume so the printed result lands inside the display tolerance. That tuning is one item in what the speaker calls the NO MURA toolbox. Printing RGB pixels onto a display then becomes a question of actively managing that variation instead of accepting the printhead's raw specification.
In this short video, you can learn:
* Display MURA specs call for pixel-to-pixel volume variation below plus/minus 1%.
* Typical printhead spec sheets allow drop-to-drop variation of plus/minus 10%, a tenfold mismatch.
* Drop volume tuning, one item in the NO MURA toolbox, is the route to closing that gap when printing RGB pixels.
π **Clip Abstract** The clip sets a display tolerance of Β±1% pixel-to-pixel volume variation against printhead specs of Β±10% drop-to-drop variation, and names the mismatch directly. Drop volume tuning, part of the NO MURA toolbox, is presented as the way to close it when printing RGB pixels.
About the speaker:
* Speaker: Kai Keller
* Company: Notion Systems
* Event: Electronics RESHAPED 2026 | Berlin
* Location: Estrel, Berlin
USA: The Future of Electronics RESHAPED | USA
π
Save the date: 9 and 10 June 2027
π Mountain View, California (Silicon Valley)
π Info: https://www.techblick.com/electronicsreshapedusa
Europe
The Future of Electronics RESHAPED | Berlin
π
Save the date: 20 and 21 October 2027
π Berlin, Germany
π Info: https://www.techblick.com/electronicsreshaped
#DropVolumeTuning, #NoMuraToolbox, #PixelVolumeVariation, #PrintheadDropVariation, #InkjetPrinting, #DisplayManufacturing
18:49 - 19:55
Can inkjet print standard semiconductor resists to the same critical dimension while cutting spin-coating waste by 90%?
Can inkjet print standard semiconductor resists to the same critical dimension while cutting spin-coating waste by 90%?
Replacing spin coating is one very prominent application, and it becomes especially interesting as substrates get larger. The approach is simply to print standard resists taken directly from semiconductor manufacturing, rather than formulating something bespoke for the deposition tool. That substrate-size scaling argument is what moves the discussion from a lab curiosity toward a genuine manufacturing option.
Expanding the viscosity window is the enabling advantage here. Because the tool can handle standard industry resists, they can be applied in a different way instead of being re-engineered. Anyone familiar with spin coating knows roughly 90% of the material is thrown away. If the material is cheap that may not matter, but as resist costs rise, discarding 90% becomes expensive.
The claimed outcome is parity, not compromise: the same final results and the same final critical dimension, achieved while applying standard lacquers from semiconductor manufacturing. That is the commercial argument, because a deposition change only survives if the printed feature matches the spun reference. Waste reduction counts for little if the dimension shifts, so equivalent critical dimension is the metric that matters.
In this short video, you can learn:
* Spin coating is identified as one of the most prominent processes that inkjet printing can replace, with the case strengthening as substrate sizes increase.
* Spin coating discards roughly 90% of the material, which is tolerable for cheap resists but becomes costly as material prices rise.
* Printing standard semiconductor resists is claimed to deliver the same final results and the same final critical dimension as the incumbent process.
π **Clip Abstract** Printing standard semiconductor resists by inkjet is positioned as a route to replacing spin coating, particularly as substrates grow larger. The expanded viscosity window lets industry-standard resists be applied differently while reaching the same final critical dimension, avoiding the roughly 90% material waste typical of spin coating.
About the speaker:
* Speaker: Kai Keller
* Company: Notion Systems
* Event: Electronics RESHAPED 2026 | Berlin
* Location: Estrel, Berlin
USA: The Future of Electronics RESHAPED | USA
π
Save the date: 9 and 10 June 2027
π Mountain View, California (Silicon Valley)
π Info: https://www.techblick.com/electronicsreshapedusa
Europe
The Future of Electronics RESHAPED | Berlin
π
Save the date: 20 and 21 October 2027
π Berlin, Germany
π Info: https://www.techblick.com/electronicsreshaped
#InkjetResistPrinting, #SpinCoatingReplacement, #HighViscosityInkjet, #CriticalDimension, #PrintedElectronics, #SemiconductorManufacturing
15:00 - 16:12
How do you tune 3 pL drops to Β±1% volume across a production printhead?
How do you tune 3 pL drops to Β±1% volume across a production printhead?
Drop volume tuning is one tool in the box. Measuring drop volume over the number of nozzles for a given ink and printhead can easily look like the first curve. For high-accuracy, high-resolution displays, the target was a three picoliter droplet, shown at zero.
Using electronics and process development called drop volume tuning, the variation in droplet volumes was reduced. It is the same curve and the same nozzles, but with much lower variation in drop volume. You cannot put it to zero, but the reduction was considerable.
The result was getting into a range to achieve plus/minus 1% volume accuracy in the drops. The target was a three picoliter droplet for high-accuracy, high-resolution displays, and the tuned curve reduced variation across the nozzles.
In this short video, you can learn:
* Drop volume tuning reduced droplet-volume variation across the same nozzles and the same measured curve.
* The application target was a three picoliter droplet for high-accuracy, high-resolution displays.
* The tuned process achieved plus/minus 1% volume accuracy in the drops, though not zero variation.
π **Clip Abstract** Drop volume tuning uses electronics and process development to reduce droplet-volume variation for a given ink and printhead. For a three picoliter target aimed at high-accuracy, high-resolution displays, the result was plus/minus 1% volume accuracy in the drops.
About the speaker:
* Speaker: Kai Keller
* Company: Notion Systems
* Event: Electronics RESHAPED 2026 | Berlin
* Location: Estrel, Berlin
USA: The Future of Electronics RESHAPED | USA
π
Save the date: 9 and 10 June 2027
π Mountain View, California (Silicon Valley)
π Info: https://www.techblick.com/electronicsreshapedusa
Europe
The Future of Electronics RESHAPED | Berlin
π
Save the date: 20 and 21 October 2027
π Berlin, Germany
π Info: https://www.techblick.com/electronicsreshaped
#DropVolumeTuning, #PicoliterDroplets, #PrintheadNozzles, #VolumeAccuracy, #PrintedElectronics, #SemiconductorManufacturing
Other talks by Kai Keller
- It's not nozzle count alone that matters: n.jet EHD powered by Scrona Β· The Future of Electronics RESHAPED 2025
- Advancing Additive Processes into Electronics Mass Production - Breaking Boundaries One at a Time Β· The Future of Electronics RESHAPED 2024 Berlin
- Making an Impact with Inkjet. Additive process in Electronics Production. The Future of Additive Manufacturing Β· Innovations Festival: Printed, Flexible, Hybrid, 3D, Wearable, & Textile Electronics 2023
- Replacing subtractive with additive processes in electronics production - making an impact with inkjet. Β· Digital & 3D Additive Manufacturing of Electronics, Sensors, Photovoltaics, Displays, etc 2023
More highlights from Kai Keller
- βΆ How can additive manufacturing disrupt traditional electronics manufacturing? Β· The Future of Electronics RESHAPED 2025
- βΆ What fundamentally limits the resolution of inkjet printing? Β· The Future of Electronics RESHAPED 2025
- βΆ What is the core innovation that makes the Trona EHD technology stable? Β· The Future of Electronics RESHAPED 2025
- βΆ How is industrial inkjet like a "digital slot-die coater" capable of firing 100,000 droplets per second? Β· The Future of Electronics RESHAPED 2024
- βΆ Your PCB solder mask has the same thickness everywhere, but should it? Β· The Future of Electronics RESHAPED 2024
- βΆ How do you prevent a single clogged nozzle from ruining production in a 24/7 automated factory? Β· The Future of Electronics RESHAPED 2024


