Case Study

Baratron Etch Manometer reduces zero drift

Baratron Etch Manometer reduces zero drift
Capacitance manometers measure absolute pressure in a process chamber by sensing and measuring extremely minute displacements of a thin diaphragm. The zero setting of the manometer will drift under process conditions which deposit films or particles on the diaphragm. In some etch systems, the by-products of the etch reaction can condense as a solid film and react with residual air or water vapor to produce particles. Re-zeroing the manometer costs valuable production time and reduces tool availability. N/A's Baratron Etch Manometer eliminates manometer drift in deposition- and particle-prone semiconductor processes.


Film deposits formed on a capacitance manometer housing and diaphragm during aluminum etching. The film is thickest near the vacuum port, opposite the center of the diaphragm. An analysis of the film indicated that it was mainly aluminum fluoride.

The new manometer design minimizes drift by significantly reducing deposits on the diaphragm. A channel machined into the lower vacuum housing of the manometer acts as a baffle between the vacuum chamber and the capacitance manometer diaphragm. The aspect ratio of the radial gas flow path is greater than 40:1. Molecules entering the sensor collide with boundary surfaces thousands (theoretically ~402) of times before reaching the diaphragm, increasing the likelihood that they will deposit on the channel walls. The channel protects the diaphragm from deposits while allowing the gas molecules to reach the diaphragm for accurate pressure measurement.


Design of deposition baffle in etch manometer.

Particle contamination is another cause of process-related zero drift. In aluminum etch systems, for example, the aluminum chloride etch by-products can react with residual water or air to form particles of aluminum oxychloride and aluminum hydroxide. MKS Instruments' patented particle trap uses a two-stage approach to prevent particles from contaminating the diaphragm without affecting vacuum conductance or reducing response time.


Capacitance manometer with two-stage particle trap (U.S. Patent 5,811,685) Stage one excludes large particles over 500 µm. Stage two captures particles from 250-500 µm, collecting them within a ring-shaped "well" to keep them from affecting conductance. Smaller particles (less than 250 µm) can enter the space between the lower vacuum housing and the diaphragm and do not inhibit the pressure measurement.

Particulate contamination is not exclusive to etch processes. All new Baratron 600B Series capacitance manometers now include this feature.

Results – Case study
Beta testing of the new Baratron Etch Manometer was conducted at Vitesse Semiconductor (Colorado Springs, CO). Prior to the installation of the Baratron Etch Manometer, Vitesse had significant problems with drift while performing an aluminum etch process with a TCP Lam 9600 metal etch system.

Mike Bouvy, a Vitesse equipment engineer, made the following observations in February of 1999: "As we run product, we see the drift occur … if the chamber sits idle, there is no drift. For a continuous 24-hour period of running product, we see a drift of about 1 mTorr/day. Typically, for every two days of running product, we have to re-calibrate the manometers."

Baratron Etch Manometer beta units were installed on March 30,1999. Six days later, this is what Mike Bouvy had to say: "I checked the logs, and there is no indication that the base pressures had drifted up on the daily checks over the weekend and no indication of being auto-zeroed. This has never happened before, and things are looking great with both styles of these manometers! Might be time for champagne!"

On Nov. 2, 1999, after six months with the new etch manometers, Bouvy said: "The new manometers have proven to be very reliable with virtually no drift over six months."


Performance of a conventional capacitance manometer compared with an Etch Manometer over 90 RF-hours of operation on an aluminum etch chamber.

A simple cost/benefit analysis based on taking 10 minutes to re-zero every 24 hours, with downtime cost of $10K to $100K per hour, shows that eliminating downtime for re-zeroing translates to savings of $30K to $300K per month. In addition, the pressure measurements are more reliable.

Because of the small volume between the baffle and diaphragm, the time constant for gas to flow within the sensor is less than 10 milliseconds. The partial pressure of process by-product species is typically less than 10-7 Torr during the time that they are deposited. Their contribution to the total pressure being measured is insignificant, and trapping these molecules will not affect the pressure reading.

Source: MKS Instruments

Edited by Katherine Derbyshire
Managing Editor, Semiconductor Online

Baratron is a registered trademark and Etch Manometer is a trademark of MKS Instruments, Andover, MA, USA.