OES Endpoint Detection for Used Plasma Etch Systems 2026

This guide is for: Process engineers and fab managers evaluating used plasma etch systems with questionable or broken OES endpoint detection.

Is a Broken OES Fixable? A Broker's Unfiltered Take

Last month, a client bought a Lam 2300 Flex Etch for $450k, excited about its 300mm capability. Three weeks later, they called me in a panic: the OES endpoint detection was dead. They'd already spent $80k on parts, and the system was idling. This isn't rare. In my 15 years brokering used etch tools, ~30% of systems with OES issues are sold with hidden or downplayed failures.

Here's why you should care: A broken OES can cost you $150k-$300k in lost throughput if you're lucky. If you're unlucky, it turns your etcher into a paperweight. Let's break down your options when evaluating a used plasma etch system with a broken OES endpoint.

---

What Even Is OES Endpoint Detection?

Optical Emission Spectroscopy (OES) monitors light emitted by plasma during etching. It detects shifts in spectral lines to signal when a layer is done etching-critical for sub-10nm nodes. Without it, you're guessing when to stop the process.

In older systems like the Applied Materials P5000, OES is often the first component to fail due to UV sensor degradation or contaminated optics. In 2026, even more systems are hitting this age threshold.

---

Fix It? Replace It? Or Live With It?

Let's compare real-world scenarios using three common models.

1. Fix the OES Sensor

• Cost: $25k-$75k (parts + labor).
• Feasibility: High if the sensor is physically intact but degraded. For example, the AE Apex 3013 RF Generator often pairs with replaceable OES modules.
• Risk: If the issue is in the optics (e.g., scum buildup from years of use), cleaning might cost $10k but save the system.

2. Replace the Entire OES Stack

• Cost: $150k-$250k.
• Feasibility: Only worth it if the system is otherwise pristine and you need full endpoint capability. The Lam 2300 Flex Etch can take a used OES stack from a similar model, but sourcing one is tough-expect to bid against competitors.

3. Work Around It

• Cost: $0-$50k (for manual monitoring upgrades).
• Feasibility: If you're running non-critical processes (e.g., back-end-of-line steps), you can use timed etching or profilometry. But this limits your system's value and increases defect risk.

---

The Hidden Costs of "Fixable" OES Failures

Let's talk numbers. In 2026, used etch systems with broken OES typically sell for 15-25% less than sticker price. But here's the catch:

• Labor is 50% of the cost. Even if you buy a $30k sensor, the 40-hour install (at $150/hour) adds $6k.
• Optics are fragile. A contaminated window might need replacement ($8k for a P5000), not just cleaning.
• Software mismatches. Older OES sensors might not communicate with upgraded controllers, requiring firmware hacks.

I've seen systems where the OES "fix" ended up costing 30% of the system's value. Ask yourself: Is that cheaper than buying a system with a working OES? Often not.

---

What to Check During Inspection

If you're evaluating a used etcher, here's what to demand:

1. Sensor Health: Request a dark current test. A healthy UV sensor should read <5nA; anything above 20nA is trouble.
2. Optics Condition: Look for scum or etch byproducts on the quartz window. If it's cloudy, factor in $5k-$10k for cleaning.
3. Software Logs: Check for "endpoint detection error" codes in the last 12 months. Recurring issues = red flags.
4. RF Generator Sync: The AE Apex 3013 RF Generator and OES must handshake properly. Test this with a dummy load.

---

Action Steps When You Find a Broken OES

1. Negotiate Harder: