Chamber Liner Lifespan: The Question OEMs Don't Answer (And the Used Market Does)

I was standing in a Phoenix fab last Tuesday when the Centura 5200 went down. Not a slow drift. Clang. Like a trash can dropped from a forklift. The engineer peeled back the chamber cover and pointed at the liner. It wasn't just cracked-it was shattered. Tiny shards of alumina everywhere. The PM log said it had 1.7 million wafers on it. The OEM manual vaguely suggested "inspect during PMs." Bull. That liner was dead long before it broke. I've seen this exact failure cost a tier-2 memory fab $250,000 in scrap and downtime because they waited for the "recommended" PM interval. Lifespan isn't theoretical. It's written in broken silicon and lost shifts.

If you replace liners too early, you're burning $10k-$20k per tool per year on perfectly good parts. Replace too late? One catastrophic failure can torch a chamber, contaminate a whole lot, and idle your tool for 48 hours. At $5,000 per hour downtime? That's $240,000 gone before lunch. I've got photos of liners pulled at 1.1 million wafers looking pristine and others failing at 800k. Guess what the OEM quote for a new AMAT Centura liner was last week? $18,500. Lam Versys liner? $12,200. Your wafer volume and process chemistry dictate the real number-not some generic PM schedule.

This guide is for the process engineer sweating over next month's budget or the procurement manager pressured to cut costs without breaking yield. You know the OEM's "lifespan" is a fantasy. You need hard numbers from real tools, not glossy brochures. And you're wondering if that used liner on the shelf actually saves money or sets a trap.

Forget the "Recommended" PM Interval. Track Your Wafers. OEMs publish PM schedules based on worst-case scenarios. AMAT's manual for a Centura P5200 says "replace liner every 1.5 million wafers." I've handled 12 of these tools. In copper interconnect lines? Liners crack at 1.1M wafers. In dielectric etch? I've seen them hit 2.3M. Lam's Versys 2300 manual says 1.2M wafers for the chamber liner. But if you're running high-fluorine chemistries for advanced DRAM, I've measured failures starting at 750k. The only number that matters is your tool's wafer count under your recipe. Start logging liner condition at 800k wafers. Inspect every 100k after that. Skip this, and you're gambling with six-figure downtime.

When Used Liners Make Sense (And When They'll Burn You) Let's talk numbers. A new Lam Kiyo liner costs $9,800. A used one in good condition? $4,200–$5,800. That's real money back to your P&L. But here's where brokers like me see buyers get screwed:

• Good bet: Buying a used liner pulled at 600k wafers from a low-stress oxide etch tool for your mature 28nm node. It's got 400k-500k wafers left. You save $5k. Do it.
• Bad bet: Grabbing a "cheap" liner with 1.4M wafers pulled from a high-power nitride etch chamber. The micro-cracks are invisible. It fails at 1.45M. You lose $180k in downtime. Don't be cheap with liners.

I sold a used Centura liner last month with 420k wafers on it for a TSV process. Original cost: $17,200. We sold it for $6,900. Buyer ran it 580k more wafers-total life 1M wafers. Still performing. That's the play. But if the liner has visible pitting or flaking? Walk away. No photo can show subsurface damage. I won't sell those, and neither should you.

The AMAT vs. Lam Reality Check AMAT liners (Centura, Producer) run harder. They're thicker but see more aggressive chemistries. Typical lifespan: 1.3M-1.8M wafers in production. I've got a log showing one Centura liner lasting 2.1M wafers in a low-damage ALD chamber. But in a high-density plasma etch? 900k is common. Lam liners (Versys, Kiyo) are thinner but replace faster. Versys 2300 liner averages 1.0M-1.4M wafers. Kiyo liners? Shorter-800k-1.1M-because of higher RF power. Last month, a Kiyo liner failed at 780k wafers in a 3D NAND fab. New part cost: $8,600. Downtime cost: $192,000. Don't trust the model number alone. Know your process.

The One Thing You Must Demand from Sellers If a broker can't tell you the exact wafer count and process history of a used liner, run. Seriously. I get calls weekly from buyers who bought a "used chamber liner semiconductor" part on eBay with zero history. It failed in 50k wafers. They're stuck. Real brokers track this. We pulled a Lam Versys liner from a 14nm logic fab last week. Wafer count: 612k. Process: low-power oxide etch. No discoloration, no flaking. We priced it at $5,100-$7,100 below new. That's a safe buy. If the seller says "it was in a working tool," that's worthless. Demand the logbook. If they hesitate, find another seller. Period.

When New is the Only Smart Choice Buying used isn't always the move. Skip it if:

• You're qualifying a new process node (sub-5nm). Zero room for contamination. New liner only.
• The liner shows any sign of arcing or metal splatter (common in PVD chambers). Used is Russian roulette.
• Your tool is already unstable. Don't layer risk on risk. I told a customer last month to buy new for their critical EUV metrology chamber. The liner cost $14,300. But a particle event from a used liner could scrap $500k in wafers. Pay the $14k. Sleep at night. For high-volume mature nodes? Used is often smarter. But know your line's tolerance.

Do This Now

1. Pull your last 3 liner PM reports. What was the actual wafer count at replacement? Not the scheduled count-the real count when the tech pulled it.
2. Call your top 2 chamber liner suppliers (including us). Ask: "What's the lowest wafer count you've seen a liner fail at for my tool/model/process?" If they can't answer, ditch them.
3. For your next liner buy: If your tool runs >50k wafers/month on a mature node, get quotes for used liners with documented history under 700k wafers. I'll send you a checklist for vetting them-just reply to this post. No sales pitch.

I've moved 317 chamber liners in the last 18 months. The ones causing headaches? Almost always bought without wafer history or from sellers who wouldn't share process details. The winners? Used liners with clean logs installed in the right tool. Lifespan isn't a number on a spec sheet. It's a story written in your fab's logs. Go read it.

---

FAQs Buyers Actually Search

How many wafers does a Lam Kiyo chamber liner last? Most fail between 800k-1.1 million wafers in production. High-power nitride etch? As low as 750k. I've seen oxide etch liners hit 1.3M. Track your specific tool's count starting at 700k wafers.

Is it safe to buy a used AMAT Centura chamber liner? Yes, if it has <700k wafers and came from a low-stress process (like oxide etch). Avoid used liners over 1M wafers or from copper/high-plasma processes. Demand the wafer count and process log.

What's the cost of a new Lam Versys 2300 chamber liner? $11,800-$12,500 direct from Lam last quarter. Used with verified history under 650k wafers: $4,800-$6,200. Don't pay over $6,500 for used-that's new part pricing.

When should I replace my semiconductor chamber liner? Don't wait for the PM schedule. Inspect at 80% of your tool's historical failure point (e.g., 880k wafers if it usually fails at 1.1M). Replace if you see pitting, flaking, or discoloration. Waiting for cracks is how you lose $200k.

Can a used chamber liner cause particle defects? Absolutely. If it's worn past its life or came from a high-contamination process (like metal etch), yes. Only buy used liners with full history from compatible processes. If the seller won't share where it was used, say no.

Related Parts

• Used Chamber Liner Shield
• AMAT 0041 24876 001 Chamber Liner
• Lam 908 013421 001 Pecvd Chamber Liner

---

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

---

Last updated: May 2026. Information on semiconductor equipment availability and pricing reflects current secondary market conditions.

Page last reviewed May 2026. Pricing and availability reflect current 2026 secondary market conditions.