Legacy Node Equipment Guide: What to Buy for 180nm–350nm Fabs in 2026

This guide is for: a fab manager or procurement lead building or expanding capacity for legacy node production (180nm–350nm) who is discovering that the "cheap used equipment" assumption was wrong.

The dynamic shifted around 2022 and hasn't reversed. Automotive chips, industrial controllers, power management ICs — these all run on 180nm–350nm process nodes and they are not moving to advanced nodes. Ever. The fabs that made them are being retired anyway. The used tool market for legacy lithography, implant, and etch is tight. Prices are up 40–60% from 2019 in some categories.

Get your sourcing wrong and you'll either overpay, buy tools you can't service, or wait 12 months for a part that hasn't been manufactured since 2015.

Why This Market Moved

Three things happened simultaneously. CHIPS Act funding created domestic greenfield fab projects that needed legacy node capability. Automotive electrification drove massive 180nm–250nm demand (power MOSFETs, gate drivers, microcontrollers). And the legacy fab retirement wave — tools from 1990s fabs finally hitting end-of-life — removed inventory from the market faster than new decommissioning is adding it back.

The shortage isn't evenly distributed. 350nm litho tools are easier to find than 250nm. Diffusion is fine. Implant, specifically high-energy and high-current for well formation and deep junction work, is genuinely scarce.

Lithography: The Real Constraint

For 350nm work, GCA 6500/6300 g-line steppers and Canon i-line systems (i4/i6) are serviceable and reasonably available at $80K–$200K. Parts are a concern — GCA is essentially unsupported, so source a parts donor simultaneously.

For 250nm–350nm with tighter CD control, ASML PAS 5500/60 and 5500/100 systems are the workhorses. Budget $300K–$700K for a working system. The 5500/200 and later variants with better CD uniformity run $400K–$900K. These are complex machines — laser source condition matters enormously, and a KrF laser rebuild is $50K–$80K.

For true 180nm production, you need i-line with aggressive OPC or early DUV (248nm). ASML 5500/900 or Canon EX series: $600K–$1.5M. These are no longer commodity tools.

Don't buy a stepper without a printed wafer measurement. Stage metrology, alignment, and CD uniformity need to be verified before payment. A sticker that says "250nm capable" means nothing if the lens is delaminated.

Etch: Mostly Available, Watch RF

For 350nm poly and oxide etch: LAM TCP 9600/9400 series, AMAT MXP+, TEL Unity. $80K–$200K each. Parts availability is reasonable — TCP 9600 magnets and match networks still trade.

For 180nm–250nm: you want better plasma uniformity. Lam 2300 Kiyo is overkill and expensive. A Lam TCP 9400 in good condition with updated RF runs fine for this geometry. AMAT DPS (Decoupled Plasma Source) is another option at $120K–$250K.

Metal etch (aluminum stack, barrier): AMAT MxP+ or HDP runs fine. Budget $80K–$180K.

The RF generator situation deserves a separate check. Advanced Energy Pinnacle+ and MKS Astron generators for legacy Lam systems are still available used at $4K–$12K but increasingly sourced through refurbishers. Budget a spare.

Diffusion and Oxidation: Not the Bottleneck

4-stack horizontal furnaces (Thermco OMEGA, SVG Thermco, Tystar) are plentiful. $40K–$120K for a working 4-tube system. Quartz ware is the ongoing cost — budget $8K–$20K/year per stack.

Gate oxide for 350nm isn't demanding. For 180nm gate oxide (6nm–8nm range), you need clean furnace tubes and good process control. The equipment is available; the process recipes need careful transfer.

RTP for dopant activation: AMAT RTP 610 or Mattson Heatpulse are both fine and run $50K–$130K.

Ion Implant: Source Early

This is where teams get stuck. Varian 350D (medium current) handles well implant, PLDD, VT adjust — $120K–$280K. Available. High-current implant (Varian VIISta 900/810, Axcelis Optima) for source/drain and buried layer: $200K–$500K, and availability is spotty. Start sourcing implant 9–12 months before you need it.

Don't buy implant without verifying beam current spec and end-station integrity. A contaminated end station is a 6-month delay. Get the implant history log.

What's Not Worth Buying Used for Legacy Nodes

Photomasks for old geometries aren't a used equipment issue — you'll need new reticles for any serious production. Budget $5K–$30K per layer depending on complexity.

Process gas systems and delivery hardware: buy new. A gas panel failure on a running tool costs more in downtime than the hardware savings.

Planning Your Build Timeline

Equipment sourcing to qualified process: 12–18 months minimum. Tools need PM on arrival, installation, process baseline, and qualification lots before you run production wafers.

Build in at least one 6-month slip. Every legacy fab build I've seen runs late — not because of equipment delays alone, but because facility prep (utilities, HVAC, gas lines) competes with equipment installation and someone always sequences it wrong.

FAQ

What does it cost to build a 180nm fab? A minimum viable 180nm production fab (single metal layer, basic front-end) runs $20M–$50M for equipment. Full process coverage with redundancy is $60M–$120M. These are equipment-only numbers — facility and infrastructure add 40–60% more.

Is 350nm equipment still available used? Yes, but the best systems are going fast. 4-stack furnaces and g-line steppers are plentiful. Spare parts for older GCA steppers are increasingly hard to source. Buy sooner rather than later.

What's the hardest legacy node equipment to find in 2026? High-current ion implant (Varian VIISta, Axcelis Optima) and DUV lithography (ASML 5500/900 and equivalents) for 180nm work. Budget 9–12 months to source these.

Can I run 180nm with used ASML i-line? Possibly, with good OPC. The ASML PAS 5500/100 series has been used for 250nm production successfully. True 180nm with tight CD control needs DUV or very aggressive process development on i-line.

Where do I find legacy node equipment? Broker networks with fab-closure relationships are better than auctions for this class of equipment. Auctions get traffic; brokers get you tools with process history documentation.

What service support exists for legacy tools? Third-party service is active for most 180nm–350nm era tools. Companies like FSI, Entegris (for parts), and regional service providers cover Lam TCP, AMAT Centura/P5000/MXP+, and ASML PAS 5500. Pure GCA support is essentially independent/aftermarket only.