EUV Lpp Source

Every EUV scanner in production today generates its 13.5nm light the same way: by firing a high-power CO2 drive laser at a stream of free-falling molten tin droplets until each one ignites into a plasma hot enough to emit extreme ultraviolet light, a method called laser-produced plasma, or LPP, and ASML's source supplier Cymer is the only organization building and qualifying this hardware at production volume. A tin-droplet generator ejects roughly 50,000 droplets of molten tin per second, each only tens of microns across, falling through a vacuum vessel at high speed; hitting a droplet that small, that fast, with enough laser energy to turn it into a 13.5nm-emitting plasma, and doing it 50,000 times a second with the timing precision to catch every droplet, is the entire engineering problem the EUV source represents. Nothing about this is a simple point-and-fire laser; the drive laser itself is a CO2 laser chosen specifically because its longer infrared wavelength couples efficiently into tin vapor and plasma, and the targeting system has to track and hit a falling droplet stream with micron-level precision on a microsecond timescale, which is why Cymer's LPP source modules, not the scanner's optics, are usually the first place a fab looks when EUV tool uptime or power output underperforms.

The actual droplet-to-plasma conversion happens in two deliberate laser pulses, not one, and that two-pulse scheme is what makes tin-droplet LPP sources practical rather than wasteful. A low-energy pre-pulse strikes each droplet first, not to ionize it but to deform and flatten it into a thin, wide disk, dramatically increasing the surface area the main pulse will have to work with; a short delay later, the high-energy main CO2 pulse strikes that flattened disk, and because the target area is now so much larger than the original spherical droplet, the main pulse couples into it far more efficiently than it would into an unmodified droplet, converting a much larger share of the tin into usable 13.5nm-emitting plasma instead of wasting laser energy heating tin that never reaches emission temperature. This pre-pulse/main-pulse dual-pulse scheme is the single biggest reason LPP conversion efficiency climbed from the low single digits in early EUV source generations to roughly 5 percent today, and that 5 percent figure is doing an enormous amount of work upstream of every other EUV limitation already documented in this batch: it is the ceiling on how much of the drive laser's input power ever becomes 13.5nm light in the first place, before a single photon reaches the euv-scanner-reflective-mirror-optics-stack row's chain of roughly 10 to 11 Mo/Si mirrors, each keeping only 65 to 70 percent of what hits it.

Tin droplet generator ejects molten Sn stream
~50,000 droplets/second, tens of microns diameter, free-falling in vacuum
Dual-pulse CO2 drive laser
Pre-pulse strikes droplet
Low-energy pulse flattens the sphere into a thin, wide disk; does not ionize
Main pulse strikes flattened disk
High-energy CO2 pulse couples into enlarged surface area far more efficiently
Tin ionizes into EUV-emitting plasma
Droplet-to-plasma conversion efficiency ~5%; this is the hard ceiling on source power
13.5nm light collected by collector mirror
Grazing-incidence collector gathers emitted EUV toward the illumination optics; debris from the plasma is the collector's main degradation threat
Light enters illumination/projection mirror path
Feeds directly into euv-scanner-reflective-mirror-optics-stack's ~10-11 mirror chain

The plasma-emission step in that flow is also the dirtiest one, and it is the direct upstream cause of a contamination problem this batch has already documented from the downstream side. Every tin-droplet ignition that produces 13.5nm light also throws off tin debris, fast ions, and neutral atoms that travel outward from the plasma in every direction, including straight back toward the grazing-incidence collector mirror positioned to gather the emitted EUV; that same debris is what slowly coats and degrades the collector mirror's reflective surface over the source's operating life, which is the specific contamination pathway the old collector-mirror reference material covers from the mirror side and this row covers from the emission side. A fab cannot eliminate the debris without eliminating the plasma that makes the light, so the real engineering response has been debris mitigation around the source, hydrogen gas flow, magnetic debris deflection, and the collector mirror's own sacrificial coating layers, treating a known byproduct of tin-droplet LPP as a maintenance cost to be managed rather than a defect to be designed away, since any collector-mirror reflectivity lost to debris is power the scanner never gets back, stacking directly onto the ~5 percent conversion-efficiency loss already taken at the plasma itself.

Early EUV source generationsCurrent production LPP sources
Drive laser schemeSingle-pulse or early dual-pulse CO2, less refined droplet targetingMature pre-pulse/main-pulse dual-pulse CO2 scheme, micron-precision droplet tracking
Droplet-to-plasma conversion efficiencyLow single-digit percent~5 percent, the current production benchmark
Source output powerWell under 100W at intermediate focus, insufficient for high-volume throughput250W in mainstream production EUV; 500W+ targeted and demonstrated for High-NA EUV
Collector mirror debris impactSame debris mechanism, but lower source power meant lower absolute debris fluxHigher power means higher debris flux; mitigation (gas flow, magnetic deflection, sacrificial coatings) now load-bearing, not optional
SupplierCymer under early ASML integration, still maturing the dual-pulse approachCymer/ASML production-qualified LPP source modules, the sole supplier at this scale
LPP Tin-Droplet EUV Source: Dual-Pulse Conversion Sequence Three stages of one droplet's trip from molten sphere to 13.5nm-emitting plasma 1. FREE-FALLING DROPLET Sn molten tin sphere tens of microns wide ~50,000/second 2. PRE-PULSE FLATTENS DROPLET low-energy pre-pulse disk-shaped target larger surface area 3. MAIN PULSE IGNITES PLASMA high-energy main CO2 pulse 13.5nm plasma emits EUV + debris, ions, neutrals debris travels outward toward collector mirror Conversion Efficiency: Why ~5% Is the Hard Ceiling 5% input laser power that becomes usable 13.5nm EUV light remaining 95% lost to heat, non-EUV emission, debris kinetic energy This ~5% ceiling is set before a single photon reaches the Mo/Si mirror chain in euv-scanner-reflective-mirror-optics-stack Source Power Roadmap: 250W Production to 500W+ High-NA 250W mainstream production EUV source power Intermediate incremental power increases to hold throughput as dose rises 500W+ targeted/demonstrated for High-NA EUV throughput needs Every watt of source power gained must still survive the mirror chain's cumulative reflectivity loss downstream

Every number in this row is a cause sitting one step upstream of two effects this batch already covers from the downstream side. The roughly 5 percent droplet-to-plasma conversion efficiency, achieved through the pre-pulse/main-pulse dual-pulse CO2 scheme, is the hard ceiling that determines how much 13.5nm power a source like Cymer's can hand off at all, and that ceiling is exactly why euv-scanner-reflective-mirror-optics-stack's chain of roughly 10 to 11 Mo/Si mirrors at 65 to 70 percent reflectivity each is such an unforgiving multiplier instead of a tolerable one: there was never much power to lose in the first place. The same ignition event that makes the light also throws tin debris back at the collector mirror, the contamination pathway that row's predecessor material documents from the mirror side, meaning the source is the actual origin point of a degradation problem usually described only in terms of its optical symptom. And the industry's answer to both constraints at once, pushing source power from 250W in mainstream production toward 500W and beyond for High-NA EUV, is also the reason mask contamination risk keeps rising in step, since more source power means more plasma debris reaching the collector and more EUV dose reaching the mask, which is precisely why euv-pellicle-process-and-requirement's few-millimeter standoff membrane has become a production necessity rather than an optional add-on as this source technology has scaled up.

Take EUV LPP source further

Ask the copilot about this term, or have our engineers assess it against your process.