Factories depend on gas delivery systems to provide accurate and repeatable gas mixtures to assure tool process quality and productivity. Valves, Mass Flow Controllers (MFCs) and gas fittings - key components of these systems - are subject to reactive gases and water vapor that lead to their degradation, intermittent faults and eventual failure. With no ability to monitor their own performance in the delivery of compositionally-correct gas mixtures, these components place product yield and tool productivity at risk as they age.

Verionix sensors provide "real-time" views into the performance of gas delivery systems, identifying changes in not only gas purity but component performance as well. With the ability to detect minute changes in gas composition at more than 10 samples per second, Verionix sensors are ideal for characterizing both steady-state and transient gas flows that feed these critical processes.
 

Problem

  • Atomic Layer Deposition (ALD) tools use rapidly-pulsing doses of precursor gas to deposit films in small multiples of single-atom thickness.
  • ALD gas metering valves cannot detect random faults in their own performance.
  • Traditional gas composition sensors have neither the pressure compatibility (~ 300 mTorr) nor the time responsiveness necessary to monitor gas flow and resulting dose control.

Process Environment and Setup

  • Verionix sensor configured to sense changes in ALD metal precursor (TMA) and oxidant (water vapor) concurrently, and in real time.
  • Average process pressure level of 300 mTorr.
  • Verionix Vx-2000 series sensor mounted directly on the process chamber using a KF-25 flange.
  • Dose cycle repeats every 3 seconds. Verionix sensor samples composition at 10 Hertz.

Critical Verionix Data

  • Real-time composition data Verionix data detected an intermittent fault of the ALD valve that pulsed metal precursor gas (TMA) into the chamber.

Results

  • Faulty pulse flow controller for the ALD metal precursor replaced eliminating yield losses as the result of its random behavior.

 

Home  |  Privacy Policy  |  Legal Statement