Introduction

Real-time monitoring of water (H₂O), oxygen (O₂) and particles is essential to maintain process integrity and to rapidly detect excursions. This article covers detection technologies, placement strategies and operational best practices.

Why online monitoring matters

 • Initial commissioning tests can pass, but contamination can occur during operation (valve wear, leaks, adsorption/desorption cycles). Without online monitoring, issues are detected late — often after product quality is impacted.

 • Key monitored parameters: H₂O, O₂, and particulate concentration/size distribution.

Water (H₂O) monitoring

Techniques:

 • Cavity Ring-Down Spectroscopy (CRDS): ultra-sensitive, ppb → ppt capable; highly suitable for semiconductor POU monitoring.

 • Tuned Diode Laser Absorption Spectroscopy (TDLAS): another optical option for low-ppt H₂O detection.

 • Electrochemical sensors: cost-effective for higher concentration ranges (ppm).

Placement & practice:

 • Install CRDS/TDLAS analyzers at or immediately upstream of the POU for fastest fault isolation.

 • Use heated sample lines and appropriate flow conditioning to avoid condensation and sample loss.

Oxygen (O₂) monitoring

Techniques:

 • Zirconia (ZrO₂) sensors and optical methods are common for low-ppm to sub-ppm O₂ detection.

 • For ppb detection, high-sensitivity optical analyzers or trace O₂ analyzers should be used.

Placement & practice:

 • POU or critical branch monitoring enables quick detection of ingress or permeation events.

 • Calibrate instruments against certified standards on a defined schedule.

Particle monitoring

Techniques:

 • Laser particle counters detect and count particles down to 0.05–0.1 µm depending on the instrument.

 • Use in-line or extractive sampling depending on installation constraints.

Placement & practice:

 • In wafer fabs and clean labs, POU particle monitoring is common. Combine continuous counters with periodic filter sampling for trend validation.

Data handling & alerts

 • Integrate analyzers into a data historian/EMS so trends can be trended and alarms issued automatically.

 • Implement threshold alarms with tiered responses (e.g., automated purge, manual inspection, shutoff) depending on severity.

Practical example

An in-line particle and H₂O spike was detected by POU monitors; the event trace pointed to a specific bypass valve that had degraded seats. Replacing the valve and performing targeted purge corrected the issue before product impact.

Conclusion

Online monitoring is not optional for UHP systems — it’s an operational necessity. Deploy sensitive H₂O and O₂ analyzers and laser particle counters at POU points, integrate data for trend analysis, and act quickly on alarms to protect process yield.