As many people involved in Atomic Deposition (ALD) might know, area-selective ALD is a hot topic at the moment.
Papers, workshops, and blog posts are numerous (e.g., the recent atomic limits blog post or posts on BALD engineering) and they explain well what exciting new options there are to go towards area-selective growth. In a sense area-selective ALD utilizes effects that historically have annoyed ALD users, namely that due to the chemical of ALD, nucleation can depend on the initial . In general, work in the ALD was undertaken to minimize this , for instance plasma ALD generally has negligible delay, but for area-selective ALD, nucleation delay effects have been cleverly enlarged instead.

Interestingly, even though plasma ALD generally has no nucleation delay, it can still be used for area-selective ALD, as my university colleagues in Eindhoven have recently shown (http://pubs.acs.org/doi/10.1021/acsnano.7b04701). Key aspect is to re-apply any inhibitor used, after the plasma exposure and this can be done in a three step ABC ALD process as seen below.

One question that came to my mind is what kind of ALD system would people want to study area-selective ALD, and I believe our systems have quite some features that could be beneficial. For instance:

• A wide range of precursors and other chemicals can be dosed. Gas pod allows wide range of MFC controlled gases.
• Inhibitor molecules can be used (e.g., flow NH₃ or CO during the precursor dose).
• H radicals as inhibitor: use H plasma (or another plasma) before the precursor dose to inhibit growth on certain surfaces
• Fluorine plasma: CF_x or F as inhibitor, run such a plasma before the precursor dose, or do area-selective ALD by doing an etch step in the same chamber every few cycles (note O₂plasma can etch Ru for instance and H₂ plasma can etch ZnO)
• SAMS dosing for growth inhibition
• Clustering with ICP-CVD chamber (a-Si:H growth), sputter chamber or atomic layer etching (ALE) chamber
• Substrate Biasing for in situ surface modification or etching

Our tools in general also provide excellent control, for instance they have:

• Tuneable dose which can be used such that you have nucleation on one surface but not on the other by using fast ALD valves and fast automated control. Also MFC controlled gases can be dosed precisely with ALD valves.
• system cleanliness due to turbo and loadlock to keep inhibition intact for a long time.
• Plasma capability to condition or clean chamber and “reset” the
• Growth monitoring with in-situ diagnostics: spectroscopic ellipsometry, mass spectrometry, and optical emission

The ideas of clustering and combining with etching which have been proposed for area-selective ALD lead me again to the idea of atomic scale processing, which I discussed in a blog post before (The toolbox for Atomic Scale Processing). I can imagine that by combining area-selective ALD and other processes you could grow novel metamaterials and unique structures. For instance: by growing on selectively exposed Cu, or only effectively grow on sidewalls of structures by periodically etching planar deposition away (i.e., topographically selective ALD). To this end combinations with direction ion exposure could be a benefit, either in a separate plasma etcher or in a FlexAL with the substrate biasing feature. Overall I am expecting a lot of new exciting discoveries and new structures and materials to be able to be made in a controlled fashion.

Author: Dr Harm Knoops