Issue



Cleanroom particle sources, Part I: Gas to particle conversions


11/01/2003







By Robert P. Donovan

Not many aerosol particles in the make-up air drawn into a cleanroom from the outside ambient penetrate through the cleanroom's HELPA/ULPA filters. Yet the ambient air within a cleanroom is never free of aerosol particles. So, where do they come from?

Click here to enlarge image

In early cleanrooms, human operators were probably the primary source of most of these internally generated aerosol particles. With improved personnel protective gowns and dress protocol, as well as increased automation, particulate contamination from people has been dramatically reduced.

Processes and process equipment are now thought to be the two dominating sources of cleanroom aerosol particles. Over the next two issues, I'd like to discuss two general categories of equipment/process sources sometimes present in cleanroom operations. This month, I'll cover gas-to-particle conversions.

A gas-to-particle conversion can be a simple chemical reaction among vapors that results in a solid or liquid aerosol product. Such products are labeled secondary aerosol particles, as opposed to primary aerosol particles that are directly emitted into the environment without any gaseous precursors. Secondary aerosol particles often form by homogeneous nucleation, meaning no condensation nuclei involved. In the outdoor air, they can create significant atmospheric conditions, such as smog and haze. Similar interactions can occur in the cleanroom when, for example, acid and base vapors from adjacent wet benches intermix.

I recall generating challenge aerosol particles by introducing dilute HCl gas from a commercial cylinder into a transparent feed line containing a flow of dilute ammonia gas, also from a cylinder. I was following a published procedure recommended for generating a challenge aerosol.

Upon initiating the gas flows, such a dense cloud of ammonium chloride formed that I immediately shut down, installed a canopy over the entire apparatus ,and replumbed the gas flows to discharge outdoors. Even then, our environmental health and safety people approved the setup only grudgingly; and this was after I assured them that the arrangement was temporary, and that the test series would be short-lived.

Later, however, I realized that, in many fabs, wet benches often operate with open, heated, acid baths in close proximity to heated alkaline baths. Such an arrangement would seem to reproduce the conditions necessary for the interactions that previously had so alarmed me and that non-semiconductor people had actually recommended as a procedure for efficiently generating aerosol particles at high rates. Even without acid/base particle generation, however, other opportunities for gas-to-particle conversions in cleanroom operations remain. Often, these sources are in the wafer processing equipment along with product wafers.

For example, particle concentration, as measured by a condensation particle counter, is much higher in the gas exiting a gas cylinder through a critical orifice when the upstream gas pressure is high (say 2,000 psi) than when it is lower (say 200 psi).

While condensation of trace organic vapors in the cylinder gas can explain this observation for gases having positive Joule-Thompson coefficients (gas temperature decreases as gas pressure decreases), such as nitrogen and argon, the observation of similar behavior in helium, which has a negative Joule-Thompson coefficient (helium gas temperature increases as pressure decreases) implies that other physical chemical phenomena are also involved.

The data reported in Reference 1 and 2 show that simply placing a chemical purifier in the gas line upstream of the expansion orifice dramatically reduces the downstream particle concentration in both helium and nitrogen/argon gases, even at high-feed pressures. On the other hand, particle filtration upstream of the expansion valve has no effect on the downstream particle concentration, clearly indicating a gas-to-particle conversion amechanism during expansion.

Reference 3 reports similar gas-to-particle conversions in pumping down vacuum stations from atmospheric pressure. Reducing the pressure cools the atmospheric gases in the chamber, causing the well-known visible condensation of water vapor.

When other impurities, such as SO2 and H2O2 are also present and incorporated into the condensed water droplets, the subsequent water re-evaporation leaves behind small, non-volatile residue particles—most likely composed primarily of sulfuric acid. The size of these aerosol particles makes them most easily detected using a condensation particle counter.

ROBERT P. DONOVAN is a process engineer assigned to the Sandia National Laboratories and a monthly columnist for CleanRooms magazine. He can be reached at: [email protected]

References:

  1. Periasamy, R. et al, "Particles in High-Pressure Cylinder Gases," Aerosol Sci & Technol. 12 (3), 1990, pp. 762–776
  2. Wen, H. Y., G. Kasper and D. Montgomery, "Nucleation of Trace Amounts of Condensible Vapors in an Expanding Gas Jet," J. Aerosol Sci. 18 (1), 1988, pp. 153–156
  3. Ye, Y, B. Y. H. Liu and D. H. Y. Pui, "Condensation-Induced Particle Formation during Vacuum Pump Down," J. Electrochem. Soc. 140 (5), May 1993, pp. 1463–1468