Microsoft-backed startup Lace raises $40m series A

[Daniel Nenni]

Lace, a Norway-headquartered chipmaking equipment startup backed by Microsoft, has raised $40 million in a series A round to develop lithography tools that use a helium atom beam instead of light.

The round was led by Atomico and included Microsoft’s venture arm, M12, Linse Capital, the Spanish Society for Technological Transformation, and Nysnø, while Lace did not disclose its valuation.

Chipmakers such as TSMC and Intel use ASML’s light-based lithography systems, and Lace’s approach could print far smaller features by using a beam about 0.1 nanometer wide.

CEO Bodil Holst told Reuters the method could enable designs about 10 times smaller than today’s tools, and Imec’s John Petersen said the concept could push feature sizes down by an order of magnitude.

Lace has prototypes and aims for a test tool in a pilot fab around 2029, and presented an invited paper at a lithography summit in February.

Lace’s technology stems from a decade of European research

Lace’s atom-beam approach grew out of FabouLACE, a European Union-funded effort to build a mask-less patterning method for a 2-nanometer process.
A related program, NanoLACE, received €3.36 million in funding from 2019 through 2024.

• The European Commission authorized Lace to bring the technology to market by 2031, while imec, a Belgium-based semiconductor R&D institute, will monitor and verify performance.
• 
  
• Public support ties Lace to a wider European push to advance semiconductor manufacturing technology, beyond venture capital funding alone.
• 
  
• The search for ASML alternatives extends beyond atom beams
• Lace’s funding fits into a wider hunt for next-step lithography options outside ASML’s extreme ultraviolet (EUV) systems.
• Laser-produced plasma (LPP) EUV sources waste much of their input power, and one retired researcher estimated the overall EUV-LPP electro-optical conversion efficiency may fall below 0.1%.
• 
  
• U.S. startups xLight and Inversion Semiconductor are building particle-accelerator-based light sources, which they say can deliver more output with better energy use 1.
• Research groups in Japan are working on free-electron lasers that could be 10 to 100 times more efficient, adding another path in a global effort to ease lithography bottlenecks.

https://www.techinasia.com/news/microsoft-backed-startup-lace-raises-40m-series-a

 

[user nl]

For those interested in EUVL and Sources, online or go to UC Berkeley, 8-11 June 2026:

https://euvlitho.com/

 

[Fred Chen]

I read about this in the last couple of days. Some key details are missing.

First, how are these atoms produced? Plasma or ion beam?

Second, how are they guided to the wafer? Direct-beam or through a mask? A focused beam will be low throughput. Can they have multiple beams which can be deflected (just like for electron beams)? If a mask is used, won't it get sputtered?

Finally, maybe most importantly, how do they deposit their energy? Direct sputtering or etching? Or chemical reactions in the resist? If it's still chemistry, excess energy (e.g., >10 eV) can still be ionizing and lead to (electron) blur on top of the chemical activation blur.

A diffraction-based image from a remote mask is also expected to be lower contrast than one which is in much closer proximity. Again, mask wear is reminiscent of nanoimprint.

Going back to stochastics, as long as there is some dose (doesn't have to be photons), there's an opportunity for stochastic variations.

 

[user nl]

I read about this in the last couple of days. Some key details are missing.

Here you can find some info with references (e.g. Google Scholar) to the CEO of Lace Technologies, Bodil Holst
https://en.wikipedia.org/wiki/Bodil_Holst

Here a link to the most recent project report on the EU-funded FET that was concluding end of 2024, not many detailed results/publications are listed though. You can click on the various links on the results link:

Mask Based Lithography for Fast, Large Scale Pattern Generation with Nanometer Resolution | Nanolace | Project | Results | H2020 | CORDIS | European Commission

Deliverables, publications, datasets, software, exploitable results

cordis.europa.eu

 

[U235]

I read about this in the last couple of days. Some key details are missing.

The incident helium has been previously excited to a high metastable state (20 eV internal energy, 1000s lifetime) by DC discharge, but has low kinetic energy (perhaps cooled).

Patterning results from the interaction and decay of the exited atomic state, rather than by collision momentum transfer.

It is mask based.

The claim is a localised energy transfer in the resist, but low sputtering.

 

[user nl]

The incident helium has been previously excited to a high metastable state (20 eV internal energy, 1000s lifetime) by DC discharge, but has low kinetic energy (perhaps cooled).

Patterning results from the interaction and decay of the exited atomic state, rather than by collision momentum transfer.

It is mask based.

The claim is a localised energy transfer in the resist, but low sputtering.

Suppose the metastable He-beam goes with 1500 m/s and has a density of say 10^14 / cm^3 at the msk. At the mask surface, where they hit the mask outside the patterened nanoholes, the He* atoms will be quenched by Penning ionization. Will this not result in catastrofic mask damage from heating. The monolayer mask is in vacuum, so can only deposit energy by radiation?

 

[Fred Chen]

20 eV internal energy could ionize, but the number of electrons wouldn't be as many as EUV. I wonder if this reduces the sensitivity significantly?

 

[user nl]

From Gemini:

20 eV internal energy could ionize, but the number of electrons wouldn't be as many as EUV. I wonder if this reduces the sensitivity significantly?

I mean absorption of the metastable He energy by the nanohole patterned single layer MASK.

Gemini estimates:

1. Secondary Electron Emission (SEE) from $He^*$​

When a metastable helium atom ($2^3S$ state) hits the solid part of the mask, it doesn't just bounce off; it "quenches." Because the internal potential energy of the $He^*$ atom (19.8 eV) is significantly higher than the work function of materials like h-BN (~5–6 eV), it triggers a process called Penning Ionization or Auger Neutralization.

• Mechanism: The excited electron in the helium atom drops to a lower energy state, transferring its energy to an electron in the mask material.
• Emission: This electron is ejected from the surface as a secondary electron.
• Yield: For 2D materials, the secondary electron yield ($\gamma$) is typically 0.05 to 0.2 electrons per incident atom.

2. Calculating the Power Load​

Using your parameters ($n = 10^{14} \text{ /cm}^3$, $v = 1500 \text{ m/s}$):

Step A: Particle Flux ($\Phi$)

$$\Phi = n \cdot v = (10^{14} \text{ atoms/cm}^3) \times (1.5 \times 10^5 \text{ cm/s}) = 1.5 \times 10^{19} \text{ atoms / (cm}^2 \cdot \text{s)}$$
In nanometer terms, this is $150,000$ atoms / ($nm^2 \cdot s$).

Step B: Energy Flux (Power Density)

Each quenched atom deposits its internal potential energy ($19.8 \text{ eV}$) into the mask.

$$P = \Phi \times E_{potential} = (1.5 \times 10^{19}) \times (19.8 \text{ eV}) \times (1.602 \times 10^{-19} \text{ J/eV})$$
$$P \approx \mathbf{47.6 \text{ W/cm}^2}$$

3. Will the Mask Survive?​

A load of $47.6 \text{ W/cm}^2$ is exceptionally high for a single atomic layer. Whether it survives depends on how it manages this heat.

The "Thermal Death" Risk​

• No Bulk Sink: Unlike a TSMC EUV mask, which is 6.35 mm thick and made of LTEM, your mask is a single layer of atoms (~0.34 nm thick). It has almost zero thermal mass to soak up energy.
• Suspended Geometry: In a nanohole mask, the atoms are blocked by the solid monolayer, while others pass through nanoholes whose effective size is reduced by dispersion-force "exclusion zones". The energy from the blocked atoms must travel laterally through the monolayer to reach the support frame.
• Radiation Cooling: At these power levels, the mask would need to reach temperatures over 1,500°C to radiate away $48 \text{ W/cm}^2$ of heat. Graphene and h-BN can technically survive such temperatures in a vacuum, but they become highly susceptible to "tearing" or structural failure.

The "Stochastic" Damage​

• Secondary Electron Scouring: The flux of secondary electrons (calculated at ~$\gamma \times \Phi$) will create a constant "micro-current" through the monolayer. Over time, this can lead to electron-stimulated desorption, where atoms are literally knocked out of the lattice, causing the nanoholes to "grow" or merge, destroying the pattern.

 

[U235]

I don't think the masks have to be a monolayer.

5+ nm thick SiN membranes from what I can see.