Michael Haslinger | Profactor: Why do fully populated 200 mm AR waveguide masters reach €100,000, and how do working stamps protect them?
18:29 - 19:56
Other snippets from this talk
Summary of the clip:
Why do fully populated 200 mm AR waveguide masters reach €100,000, and how do working stamps protect them?
Mastering is time-consuming and very expensive, and masters can cost up to €100,000 and more when you want a fully populated wafer on 200 millimetres. Almost everything can be used to create a copy of a structure, but the master itself is the costly asset. Because of that, the next step in the nanoimprint process is to create a working stamp from the master, a first copy that carries the pattern for later imprinting.
The purpose of the working stamp is to give you a controlled process and to avoid damaging your master. You create the copy, the master stays on the shelf and is not used anymore, and you use the working stamp for the nanoimprinting later on. This separates the expensive, hard-to-replace original from the repeated production runs.
Stamp material selection matters: soft and hard stamps are available, and you have to select the correct one. For high resolution you need harder materials; for bigger structures you can use softer materials. Mechanical durability is important because you want a long stamp lifetime, and sometimes you need an anti-sticking layer. You also need to know what the material is capable of in the process. For substrates, the key requirement is high surface quality.
In this short video, you can learn:
* A fully populated 200 mm master can cost up to €100,000 and more, because mastering is time-consuming and expensive.
* The working stamp is the first copy, used for imprinting so that the master stays on the shelf undamaged.
* High-resolution stamps need harder materials while bigger structures allow softer stamps, with mechanical durability and anti-sticking layers governing stamp lifetime.
📋 **Clip Abstract** Mastering a fully populated 200 mm wafer can cost €100,000 and more, so the nanoimprint flow takes a first copy, the working stamp, and keeps the master on the shelf. Stamp material choice, from hard stamps for high resolution to soft stamps for larger structures, plus durability and anti-sticking layers, determines lifetime and process fit.,
About the speaker:
* Speaker: Michael Haslinger
* Company: Profactor
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#NanoimprintLithography, #WorkingStamp, #AntiStickingLayer, #HardStamps, #Nanofabrication, #AROptics
This is a highlight of the presentation:
Nanoimprinting Technology: Fundamentals and Applications Including AR and Microlenses
More Highlights from the same talk.
11:36 - 12:28
Why can nanoimprint span single-digit nanometres to square metres per second when e-beam and EUV cannot?
Why can nanoimprint span single-digit nanometres to square metres per second when e-beam and EUV cannot?
Nanoimprint covers a huge span of throughput and resolution. The same family of process can reach down to single-digit nanometer resolution and up to square meters per second throughput. That combination is why it is described here as the most powerful technique among the compared options. No other presented method pairs that resolution floor with that throughput ceiling, because multi-beam electron beam lithography and extreme UV carry their own advantages but remain throughput-limited.
The catch is that nanoimprint is a replication technique. It does not make the pattern itself; it copies the pattern and replicates it. That distinction matters because the pattern origin must come from somewhere else. The mastering step therefore sits upstream of every nanoimprint run, and that step depends on the other lithography methods rather than on nanoimprint itself.
Multi-beam electron beam lithography and extreme UV are named as the other examples. They each have their advantages, but in terms of throughput they are limited. For that reason the talk calls them mastering technologies: they are needed, or used, to fabricate masters for the nanoimprint process. Nanoimprint then takes over as the replication step, copying those masters across the resolution and throughput span it uniquely covers.
In this short video, you can learn:
* Nanoimprint spans single-digit nanometer resolution up to square meters per second throughput within a very similar process.
* Nanoimprint is a replication technique: it copies an existing pattern rather than making the pattern itself.
* Multi-beam electron beam lithography and extreme UV are mastering technologies, needed to fabricate masters for nanoimprint and limited in throughput.
📋 **Clip Abstract** Nanoimprint's reach runs from single-digit nanometer resolution to square meters per second throughput, using a very similar process, which is why it is called the most powerful of the compared techniques. It remains a replication method, so multi-beam electron beam lithography and extreme UV serve as throughput-limited mastering technologies that fabricate the masters it copies.
About the speaker:
* Speaker: Michael Haslinger
* Company: Profactor
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#NanoimprintLithography, #SingleDigitNanometer, #SquareMetersPerSecond, #MultiBeamElectronBeamLithography, #Nanofabrication, #ExtremeUVLithography
27:38 - 29:29
Why can you never fully eliminate the residual layer under an AR grating?
Why can you never fully eliminate the residual layer under an AR grating?
In nanoimprint for augmented reality, the stack starts with a coating of known layer thickness. After patterning, that thickness separates into the pattern height and a residual layer underneath. AR needs the residual layer as thin as possible. The required initial thickness is not set by structure height alone: the stamp fill factor also matters. By calculating the initial layer thickness from both, the process can be optimized to leave very thin residual layers.
Eliminating the residual layer completely is almost never possible. As the stamp presses into the polymer resist, the residual layer becomes thinner. The thinner it gets, the harder it becomes to push out the material, and there are limits; it also takes longer. That is why a truly residual-layer-free imprint is rarely achieved, and the residual layer must be handled in some way or another.
If the imprint process is run correctly and the fill factor is known, very thin residual layers are achievable: the example given is residual layers well below 20 nanometers. A short break switch then opens up the substrate underneath, allowing the reactive ion etching process to begin. So the residual layer is not eliminated at imprint; it is controlled to a thin, known thickness and then opened for subsequent etching.
In this short video, you can learn:
* Residual layer thickness under an AR grating depends on initial coating thickness, structure height, and stamp fill factor.
* The thinner the residual layer, the harder it is to push out resist, so a completely residual-free imprint is almost never possible.
* Correct imprint process with known fill factor can achieve residual layers well below 20 nm, then a short break switch opens the substrate for reactive ion etching.
📋 **Clip Abstract** Nanoimprinting leaves a residual layer beneath the patterned resist, and AR applications require it to be as thin as possible. By calculating the initial coating thickness from structure height and stamp fill factor, the imprint can be optimized to residual layers well below 20 nanometers, though a fully residual-free imprint is almost never possible because pushing out the last resist becomes harder and slower.
About the speaker:
* Speaker: Michael Haslinger
* Company: Profactor
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#Nanoimprint, #ResidualLayer, #StampFillFactor, #ReactiveIonEtching, #AugmentedReality, #Nanofabrication
22:53 - 25:05
Can inkjet coating beat spin coating when a stamp mixes many fill factors on one substrate?
Can inkjet coating beat spin coating when a stamp mixes many fill factors on one substrate?
Spin coating deposits material, spins the substrate, and layer thickness depends on rotation speed. It enables large areas, very homogeneous layers, and multiple layers on top of each other. Inkjet deposition became available in recent years and is very interesting for augmented reality. It offers variable layer thicknesses and very little material waste; spin coating wastes most material, while inkjet uses only the material needed. Inkjet also supports multi-layers.
Variable layer thickness matters with a stamp or mold having different fill factors, with different pattern sizes next to each other. Start with one homogeneous layer; after imprinting, different fill factors produce different layer thicknesses. Inkjet deposition addresses this by depositing less amount at such areas, ending with a very homogeneous layer. This is the mechanism for matching varied imprint patterns on one substrate.
An example uses high refractive index materials on silicon. The nice color shows that it is very homogeneous. Here there is just 75 nanometers difference in layer thickness. You would see layer thickness variations just by the color. With inkjet, it is possible to have very homogeneous layers as well. That demonstrates inkjet’s ability to reduce fill-factor-induced thickness non-uniformity for high-index layers on silicon.
In this short video, you can learn:
* Spin coating couples layer thickness to rotation speed and wastes most of the deposited material.
* Inkjet deposition enables variable layer thicknesses, minimal material waste, and multi-layers for augmented reality.
* A stamp with different fill factors creates thickness differences after imprinting, and inkjet compensates by depositing less material at those areas.
📋 **Clip Abstract** Spin coating gives large-area homogeneous layers with thickness set by rotation speed, but wastes most material. Inkjet deposition offers variable layer thicknesses, low waste, and multi-layers, and can compensate for different fill factors in a stamp by depositing less material, yielding homogeneous high-refractive-index layers on silicon with only 75 nanometers variation.
About the speaker:
* Speaker: Michael Haslinger
* Company: Profactor
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#InkjetDeposition, #VariableLayerThickness, #FillFactorImprinting, #HighRefractiveIndexSilicon, #AugmentedRealityOptics, #NanoimprintLithography