Building a SiC Fab with Used Equipment: The Complete Equipment Guide

This guide is for: Engineering teams at compound semiconductor startups building 150mm or 200mm SiC/GaN fabs on a budget.

---

Every SiC startup team has the same conversation at some point: "Can we just buy silicon equipment from the secondary market?" The answer is mostly yes — but with enough caveats that getting it wrong costs you 6–18 months of process development time and $500K+ you didn't budget for.

This guide is written from the perspective of someone who has sourced SiC equipment from the used market many times. Here's what actually transfers from silicon, what needs modification, and what you should never try to make work.

---

The 150mm vs. 200mm Decision: Make It Now

Before you buy a single tool, decide your wafer size and commit. This shapes every purchasing decision you'll make for the next five years.

The 150mm reality in 2026: 150mm SiC wafers from Wolfspeed, Coherent, and SICC have been production-proven for years. 150mm silicon tools are abundant on the secondary market and relatively inexpensive. The problem: that's changing fast. SiC demand is pulling 150mm tools toward other SiC fabs, and the supply of 150mm silicon tools in good condition is contracting as leading-edge silicon fabs have been on 300mm for over a decade.

The 200mm case in 2026: 200mm SiC wafers are now commercially available from Wolfspeed (Gen3), Resonac, and others. Yields are production-viable for many device types. 200mm silicon tools are more expensive and slightly harder to source than 150mm, but the forward compatibility makes a strong argument for starting at 200mm if your device roadmap supports it. 200mm SiC tools from OEMs carry 14–20 month lead times.

Bottom line: If you're building a fab that will be in production past 2028, 200mm is the right bet. If you need to be in production in 18 months on a startup budget, 150mm used tools may be your only realistic path.

The rest of this guide covers both, noting key differences where they exist.

---

What Silicon Tools Work for SiC

Dry Etch: Strong Yes

Dry etch is the most straightforward category. Silicon etch tools run chemistry that is mostly compatible with SiC process requirements, and the mechanical requirements (vacuum, RF, temperature control) are similar.

Tools that work:

• AMAT Centura with DPS or MxP+ chambers: The workhorse of used SiC etch sourcing. Centura mainframes ($150K–$230K for a running 200mm platform) with DPS chambers handle SiC etch reasonably well with SF6/O2 or Cl2/HBr chemistry. The MxP+ chamber is particularly useful for SiC MESA etch applications. Check chamber condition carefully — SiC etch is harder on chamber walls than silicon, so chamber replacement cost ($25K–$45K per chamber) needs to be in your budget.

• Lam Research 2300 Versys and 9400 series: Lam's 200mm platforms handle high-power SiC etch applications well. The 2300 Versys in particular is used in production SiC fabs. Used pricing: $160K–$210K for a running 200mm unit. Lam 9400 is still available but getting scarcer.

• TEL Tactras (200mm) and Exelan (150mm): Less commonly sourced for SiC but functional. Parts availability for Tactras is moderate.

What to check before buying a silicon etch tool for SiC:

• Chamber material. Anodized aluminum chambers are acceptable; silicon-coated chambers may need strip and re-coat for SiC chemistry compatibility.
• RF generator condition. SiC etch runs higher power density than most silicon processes. An RF generator that was marginal on silicon will fail faster on SiC. Budget for generator inspection and refurb. (See Advanced Energy Pinnacle Plus and APEX 3013 RF Generator for pricing on common SiC etch tool generators.)
• Gas delivery system. Confirm the tool has the gas sticks you need — SF6, O2, Cl2, HBr at minimum. Adding gas lines after the fact is possible but adds $8K–$20K to the bring-up cost.

---

CVD / Epitaxy: Selective Yes

This is where the silicon/SiC split gets complicated. Standard silicon CVD tools are not SiC epitaxy tools — SiC epi requires growth temperatures of 1500–1650°C, well above what most silicon CVD systems are designed for. You will not grow your active SiC device layers in a used silicon CVD tool.

What silicon CVD tools can do in a SiC fab:

• Oxide deposition: TEL Alpha-8, ASM Eagle, Novellus Concept One for TEOS, HTO, or thermal oxide. These are all straightforward transfers. A TEL Alpha-8 in 200mm configuration runs $120K–$185K used.

• Polysilicon deposition: Useful for gate electrodes and certain device structures. Standard LPCVD poly furnaces work fine.

• SiN deposition: PECVD SiN for passivation. Standard silicon tools (Novellus, AMAT Producer, TEL) work. Watch for contamination if the tool previously ran boron or phosphorus-doped films.

• Metal CVD (TiN, W): AMAT Endura platforms for TiN barrier and W plug deposition are direct transfers. 200mm Endura with TiN/Ti or W chamber: $180K–$260K used.

What you cannot do with silicon CVD tools:

• SiC epitaxy (the core active layer). This requires a dedicated SiC epi reactor from Wolfspeed GTAT, LPE, AIXTRON, or equivalent. New SiC epi reactors are $800K–$2M+ and have 12–24 month lead times. Used SiC epi reactors are rare and expensive when they appear. This is typically the largest capital item in a SiC fab.

---

Ion Implant: Yes, With Caveats

SiC requires nitrogen (n-type) and aluminum (p-type) implants. Standard silicon implant tools handle nitrogen fine. Aluminum implant is trickier — aluminum is not commonly implanted in silicon fabs, so many used silicon implant tools haven't been configured for it.

Best tools for SiC implant:

• Axcelis Purion H and Purion XE: The Purion platform is genuinely SiC-capable and widely used in production SiC fabs. A used Axcelis Purion H in 150mm or 200mm configuration is $300K–$450K depending on configuration and energy range. These are not cheap, but they're the right tool for the job.

• Varian HVHC (High Voltage High Current): 200mm capable, handles the high-dose aluminum implant SiC p-body structures require. Used: $250K–$380K. Parts are available but getting harder to find from Axcelis (which acquired Varian's implant business).

• Axcelis GSD VHE: Good for medium-current applications including nitrogen channel implants. Used: $200K–$300K.

What to watch for: SiC implant typically runs at elevated substrate temperature (300–600°C) to reduce damage. Not all used silicon implant tools have high-temperature substrate capability. Confirm before purchase — adding high-temperature capability to a tool that doesn't have it is expensive and not always feasible.

---

Photolithography: Direct Transfer

i-line steppers work for SiC device lithography. SiC device geometries for power devices are typically 1–5 micron features, well within i-line capability.

Tools that work:

• Canon 2000i/2500i series: Readily available, $85K–$150K for a running unit with track. Good parts availability.

• Nikon NSR-2205i14E and similar: Similar price range. Nikon has a larger installed base at mature SiC fabs.

• ASML PAS 5500/200 series: More expensive ($160K–$250K), but highest i-line performance for tight CD control.

For contact lithography (common for prototyping and R&D), Karl Suss MA6/MA8 contact aligners are available for $15K–$40K and handle SiC wafers without issue.

Track systems (TEL Mark 8, SVG 8600 series): Available $40K–$80K. Critical for consistent coat and develop — don't try to run an isolated stepper without a matched track.

---

CMP: Yes With Process Development Required

Standard oxide CMP on SiC is more difficult than on silicon because SiC is significantly harder (Mohs 9.5 vs. silicon's 7). Silicon CMP tools work — the mechanical platform transfers — but your slurry chemistry and pad selection will need substantial development.

Tools that work:

• AMAT Mirra: The 200mm Mirra is the most commonly sourced CMP platform. Used: $90K–$150K. Parts availability is decent.

• Speedfam 22B / 32B: Lower cost ($45K–$85K), lower throughput. More commonly used in R&D and pilot production.

Process considerations: SiC CMP removal rates are 3–10x lower than silicon oxide. Plan for longer cycle times and higher slurry consumption. Some silicon CMP tools have platen speed and pressure limitations that become a bottleneck on SiC. Verify the tool's pressure range — you'll want to run 5–8 psi, which most Mirra platforms support.

---

Inspection and Metrology: Direct Transfer

Most inspection and metrology tools transfer directly.

KLA Tencor Surfscan 6200/6220: Excellent for unpatterned SiC wafer inspection pre- and post-process. Used: $35K–$65K. Highly recommended as a first metrology purchase.

KLA Tencor P-series profilometers (P-10, P-16): Essential for SiC MESA and trench depth measurement. Used: $20K–$45K.

KLA 2XXX patterned wafer inspection: If you can find one, buy it. Used units are scarce and prices have appreciated significantly — $300K–$500K+ for a working unit. Most early-stage SiC fabs substitute with less capable tools and accept the limitations.

Therma-Wave TP300/500 series (optical metrology): Useful for implant dose monitoring. Used: $30K–$65K.

Four-point probe (KLA RS75, Prometrix): Essential for SiC substrate resistivity and implanted layer sheet resistance. Available for $15K–$35K.

---

Anneal Furnaces: Yes, For Post-Implant Anneal

SiC requires post-implant activation anneal at 1600–1800°C. This is not achievable in standard silicon diffusion furnaces (which top out at 1200°C).

For SiC post-implant anneal, you need dedicated equipment:

• Taicang NAURA, Centrotherm SiCAnneal, AXT / RTP dedicated systems: High-temperature SiC anneal systems run $200K–$600K new. Used units are rare — most SiC fabs that have them keep them running.

Standard silicon furnaces are still useful in a SiC fab for oxide growth, CVD, and contact anneal steps that don't require extreme temperatures. A 4-tube Thermco or Tystar furnace at $40K–$70K is worth having for these steps.

---

Red Flags: Silicon Tools That Won't Work for SiC

1. Tools with aluminum chamber components not rated for halogen chemistry. Chlorine etch of SiC is aggressive. Uncoated aluminum chamber walls will corrode. Inspect before you buy.

2. CVD furnaces with hot wall contamination from boron or phosphorus. If you're using these tools for SiC gate oxide or passivation layers, contamination from silicon production history will destroy your device performance. Get full process history before purchase.

3. Implant tools without high-temperature chuck capability for SiC-temperature implant. Buying a tool that can't run hot implant means buying a second tool to handle p-body implant.

4. CMP tools with worn platens or damaged platen drives. SiC's hardness accelerates platen wear. A used CMP tool with marginal platen flatness that ran fine on silicon will struggle on SiC. Ask for recent platen metrology data.

5. Any tool with "process memory" contamination. Silicon fabs running heavy metals for specialty devices can leave contamination in systems that will ruin SiC device yields. If you don't have full process history, assume the worst and plan accordingly.

---

Typical Budget: Used Equipment Baseline for a 150mm SiC Pilot Line

| Tool | Recommended Used Model | Estimated Cost | |---|---|---| | SiC Epi Reactor | LPE PE106 or AIXTRON G5 (used, rare) | $800K–$1.5M | | Etch mainframe | AMAT Centura w/ 2x DPS | $175K–$230K | | Ion implant | Axcelis Purion H or Varian HVHC | $280K–$400K | | i-line stepper + track | Canon 2000i + TEL Mark 8 | $120K–$200K | | CVD (oxide/poly/SiN) | TEL Alpha-8 LPCVD | $120K–$180K | | Metal CVD | AMAT Endura Ti/TiN | $150K–$220K | | CMP | AMAT Mirra | $100K–$150K | | Post-implant anneal | Centrotherm SiCAnneal (used or new) | $300K–$500K | | Inspection (unpatterned) | KLA Surfscan 6200 | $35K–$55K | | Metrology (profilometer) | KLA P-16 | $25K–$40K | | Wet bench / cleaning | FSI / CFM | $25K–$50K | | Total baseline | | ~$2.1M–$3.5M |

This is equipment only. Add 25–40% for installation, PM, process bring-up, and initial consumables.

---

FAQ

Q: Can we use a silicon CVD tool for SiC gate oxide?
A: Yes, for thermal oxide growth you can use a standard vertical furnace with oxidation capability. SiC oxidation rates are about 10x slower than silicon at the same temperature, so plan for longer cycle times. The tool itself doesn't need modification — your process recipe and furnace temperature uniformity matter more.

Q: Is the Axcelis Purion really necessary, or can we start with a lower-cost Varian implanter?
A: You can start with a Varian E500 or HVHC for nitrogen implant, but if your device requires high-dose aluminum p-body implant, the Varian's limitations become apparent quickly. The Purion H's ability to run aluminum at the doses SiC requires (>1E15 cm-2) without excessive beam contamination makes it worth the premium for production. For prototyping, a lower-cost Varian can work.

Q: What's the biggest sourcing mistake SiC teams make?
A: Buying etch tools without confirming gas delivery compatibility, then discovering the tool needs $15K–$25K in gas stick additions and modifications after delivery. The second-biggest is underestimating CMP process development time — figure 3–6 months to develop and qualify a SiC CMP process on a used silicon tool.

Q: Should we buy 150mm tools if 200mm SiC wafers are now available?
A: Only if you need to be in production quickly and your device roadmap is committed to 150mm. 150mm SiC tools are still available but the window is closing — prices are up 20–30% from 2023 and availability is declining. If your device requires features not yet proven on 200mm SiC wafers, 150mm makes sense for now. Otherwise, building for 200mm is the stronger position.

Q: How do we handle process contamination risk from silicon fab history?
A: Get full process history documentation before purchase. For tools where history is unknown, budget for a deep clean and contamination verification protocol — typically surface analysis (TXRF or VPD-ICP-MS) on blanket wafers run through the tool. This costs $2K–$8K in lab fees and is worth every dollar before you stake your device yield on a tool with unknown history.

Related Parts

• AMAT Electrostatic Chuck
• Lam 2300 Versys Parts

---

Need pricing or availability? Request a quote and our team will respond within 24 hours.