Used Endpoint Detector OES Buying Guide: Verity, OceanOptics & More

This guide is for: process engineers adding endpoint detection to legacy etch tools without blowing their entire upgrade budget.

I sold a Verity SD-1020 to a research lab that had been running etch processes blind for three years. Their yield jumped 18% in the first month. They could see exactly when the endpoint hit instead of guessing with timed recipes. The system cost them $8,500 used. A new one would have been $34,000. That's the power of buying smart on the secondary market.

Getting endpoint detection wrong costs you wafers. Lots of them. A missed endpoint means over-etching, under-etching, or complete scrap. At $500–$2,000 per wafer depending on your process, one bad lot pays for a good OES system. Let's make sure you buy the right one.

---

Verity SD-1020: The Workhorse

The Verity SD-1020 dominates the used market for good reason. It's robust, well-documented, and interfaces with most etch platforms. I've moved 47 of these in the past two years. Only three came back with issues—and all three had obvious physical damage I warned the buyers about.

Used SD-1020s run $6K–$12K depending on age and included fiber optics. The spectrometer itself is solid-state and rarely fails. What dies is the fiber cable ($800–$1,200 replacement) and the software license dongle ($400 if lost—always verify this).

Check the fiber for kinks or sharp bends. A compromised fiber gives you noisy spectra and missed endpoints. Also verify the wavelength calibration—Verity units should read 200–800nm with ±0.5nm accuracy. If the seller can't show you a calibration certificate from the last 12 months, budget $600 for recalibration.

---

OceanOptics: The Budget Alternative

OceanOptics (now Ocean Insight) makes modular spectrometers that show up on the used market for $2K–$5K. The USB2000+ and QE-Pro models are most common. They're less integrated than Verity systems but highly customizable.

The catch? You're building the integration yourself. OceanOptics hardware is great. The software layer is where projects die. I've seen three labs buy cheap OceanOptics units and abandon them because they couldn't get reliable trigger outputs to their etch tool.

If you have software expertise in-house, OceanOptics is a steal. If you need turnkey operation, pay more for Verity. The $4K you save on hardware becomes $15K in engineering time fast.

OceanOptics units also need more frequent calibration—every 6 months versus annually for Verity. Budget $400 per calibration.

---

What the Specs Actually Mean

Resolution matters. For most etch endpoint detection, 1–2nm resolution is fine. You're looking for broad emission lines from etch byproducts, not fine spectral features. Don't overpay for 0.1nm resolution you don't need.

Integration time range is critical. You need 1ms–10s to catch fast endpoints in dielectric etch and slower ones in deep silicon etch. Verify the unit covers this range.

Signal-to-noise ratio (SNR) above 300:1 is the threshold for reliable endpoint detection. Below that, you're guessing in noisy processes. Ask for a demo spectrum if possible.

Dynamic range matters if you're running multiple processes. A 16-bit ADC handles most situations. Some older units have 12-bit converters—avoid these unless you're single-process.

---

Inspection Checklist: What to Test

Before you buy any used OES system:

1. Physical inspection — Look for fiber damage, loose connectors, and corrosion on the spectrometer housing. Moisture kills these units.

2. Power-on test — Unit should boot in under 60 seconds. Slow boots indicate failing flash memory.

3. Dark current test — Block all light to the detector. Readings should be near zero. Elevated dark current means detector degradation.

4. Spectral calibration check — If possible, view a known source (Hg lamp, plasma) and verify peak positions. Off by more than 1nm? Needs recalibration.

5. Software license verification — Dongles get lost. Software expires. Verify you get working software or budget $2K–$5K for replacements.

I rejected a "mint condition" SD-1020 last month because the fiber had a 90-degree kink 6 inches from the connector. The spectra looked fine in a dark room. In production, it would have failed within weeks.

---

What to Do Next

1. Define your process requirements—wavelength range, resolution, integration time—before shopping. Don't buy more than you need.

2. Budget 20% of purchase price for accessories: fibers, adapters, calibration sources.

3. Verify software licensing status in writing. A hardware bargain becomes expensive if you need new software.

4. Test before you buy, or buy from a seller with a 30-day return policy. OES systems are too expensive to gamble on.

5. Plan your integration. How will you get the endpoint signal into your etch tool? Some systems need custom I/O boards.

---

FAQ

"Verity SD-1020 used price" $6K–$12K depending on age, fiber condition, and software license status. Units over 8 years old drop to $4K–$6K but may need detector replacement.

"OceanOptics spectrometer etch endpoint detection" USB2000+ and QE-Pro models work well at $2K–$5K used. Budget for integration engineering—it's not turnkey like Verity.

"OES endpoint detector calibration cost" $400–$600 for standard calibration. Detector replacement runs $1,800–$2,400 if the CCD has degraded.

"used OES system inspection checklist" Check fiber condition, dark current levels, spectral accuracy, software licenses, and physical corrosion. Always test before buying.

"endpoint detection vs timed etch yield improvement" Typical yield gains of 10–20% when switching from timed to endpoint-controlled etch. Pays for the OES system in weeks for high-value wafers.

---

Related reading: How to Evaluate Used Semiconductor Equipment | Chamber Conditioning Guide

---

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.