A point-of-use plasma reactor chemically destroys airborne acids and organics inside tools, storage, and transit, reaching 7N purity in 16 minutes.
VOCs from ambient to <1 ppb in a single 30-minute run.
Measured by onboard PID, not modeled.
SEMICON West 2026 Startup Pitch
Finalist, Fab Infrastructure
Every process step passes inspection. Yield still slips. The damage is happening in the 20 to 90 minutes a lot spends staged, queued, or in transit, where nothing is measuring the air.
Cleanroom class tells you about particles. It tells you nothing about the parts-per-billion of acids, amines, and organics condensing on an exposed 3nm surface.
UHP N₂ dilutes contamination. It doesn't remove it. You're paying to push the problem around.
And you don't find out they've broken through until a haze excursion tells you.
Talos mounts at the point of exposure: a load port, stocker, or tool minienvironment. Onboard PID begins sampling the volume immediately.
A live baseline of what your wafers are actually breathing.
The plasma core chemically degrades acids, bases, and organics in the airstream. Nothing is trapped; there is no media to saturate.
Ambient to 7N in 16 minutes, below 1 ppb before the run ends.
The unit modulates to maintain setpoint continuously, replacing UHP N₂ purge in that volume.
A purge line item that goes to near zero.
Contamination levels stream to your MES or FDC as a numeric signal, tagged by lot and location.
A defect root-cause layer you've never had.
Additional modules chain into adjacent tools and, with CleanPipe, into sealed transport between them.
A wafer that never touches outside air from first step to last.
Pilot fabs recovered 0.6 to 1.4 points of yield on advanced-node lots by eliminating queue-time haze and corrosion defects.
One pilot replaced 340 SCFH of continuous purge on a single tool cluster, saving roughly $190K annually.
Before: contamination inferred quarterly from witness wafers. After: ppb readings every 10 seconds, per enclosure, in your FDC.
We'd been blaming the litho tool for a year. Talos showed us the defects were happening in the stocker. That alone paid for the pilot.
First AMC system where ‘clean’ is a number I can put on a dashboard.
Off-purge on a tool cluster in week two. Finance noticed before I did.
You have particle control and passive chemical filtration. Neither destroys contaminants, neither measures ppb in real time, and neither covers the time wafers spend between tools. Talos runs alongside what you have and fills the gap.
Talos treats the airstream, not the wafer. Reaction products are stable, benign species vented downstream, with onboard sensing confirming nothing new is introduced. Pilots are instrumented so you verify this in your own environment before anything scales.
Pilots mount on existing enclosures with no process-tool modification and no interruption. Typical install is a single shift.
Yes. That's why every engagement is a pilot, not a purchase. You measure, we publish the trace, and it scales only when the numbers earn it.
Standard FOUP staging, stockers, load ports, and tool minienvironments via modular adapters. Data exports to MES and FDC via SECS/GEM or a simple REST/CSV feed.
Fully instrumented pilots. No long-term commitment, no process-tool modification.