 |
by Robert P. Donovan
My May 2000 CleanRooms column used the non-volatile residue monitor (NRM) as an example of how one discipline can gain performance advantages by adopting technology from a distinctly different technology.
I'd like to continue this story but in another vein—the story of how the seemingly obvious performance advantages of the NRM have gone largely unappreciated in the semiconductor ultrapure water (UPW) business. This is, perhaps the application one would think most likely to appreciate the improved sensitivity and on-line compatibility offered by the NRM and, hence, the application in which it would prove most beneficial and popular.
 |
Part of the problem seems to be that the NRM measures a property not included in the traditional set of specifications for UPW.* It measures all the nonvolatile residue (NVR) in a liquid sample without regard to the residue's composition, source or other properties and expresses the result in terms of an equivalent concentration of potassium chloride in an aqueous solution. Thus, one gets a quantitative measure of water quality but no clue as to how to lower a currently measured value of the water's NVR. The composition of the measured NVR remains unidentified and can include a large number of possibilities. The manufacturer1,2 lists the following potential contributors to NVR: dissolved silica, particles, colloidal silica, organics and both ionic and non-ionic impurities.
Consider each of these species.
Dissolved silica: While the NRM measures dissolved silica, its results haven't always correlated well with those of conventional, on-line silica analyzers using the molybdenum blue chemical reaction. Figure 1, however, shows that, under well-controlled conditions, the sensitivity of the NRM is superior to that of the conventional silica analyzer. Nonetheless, the conventional silica monitor, a significantly lower cost analyzer, continues to be the relied-on analyzer for this contaminant.
Particles: The NRM certainly counts particles—after all, its detector is a condensation particle counter—but it's hard to conceive of a water sample in which the particle concentration would be sufficiently high to dominate NVR. The concentration of aspirated water droplets is on the order of 1012 per cm3 so that the particle concentration would have to be at least 1010 per cm3 in order to have any chance of being detected.
Light-scattering particle counters are better suited for analyzing UPW samples.
Colloidal silica: This term refers to polymeric aggregates of silica atoms, generally too small to be counted by light-scattering analyzers but not too small to be counted by the CPC of the NRM. Indeed this capability of the NRM is one of its unique strengths. As was true for particles, the number density of the colloids must be near that of the aspirated water droplets, but a solution that is 1 ppb by weight silica corresponds to about 1010 particles /cm3 of 5 nm colloidal silica particles.2 No other analyzer, off- or online, can match this sensitivity for measuring colloidal silica.
Organics: Some organics are highly volatile and others, less so. The highly volatile species disappear with the water vapor in the drying section of the NRM. The nonvolatile organics remain behind and are counted. A third class of organics—those whose volatility depends on the drying temperature of the NRM, which can be adjusted between room temperature and 125 degrees Celsius—offers the intriguing possibility of being able to speciate organic contributions by noting the temperature signature of a contaminant. This potential capability has yet to prove of practical significance.
Ionic and non-ionic impurities: Detecting and measuring these contaminants is a real strength of the NRM. However, low cost resistivity cells also perform the measurement of conductive ions with high sensitivity. The NRM, however, also measures nonconductive salts, unlike a resisitivity cell.
Thus, in view of the foregoing, the NRM remains primarily a solution looking for a problem from the semiconductor UPW viewpoint. Its ability to detect colloidal silica online may be its best performance claim at present. Given the NRM's clever design, sensitivity and online capability, it seems it should play a more prominent role in UPW technology.
Robert P. Donovan is a process engineer assigned to the Sandia National Laboratories as a contract employee by L & M Technologies Inc., Albuquerque, NM. His Sandia project work is developing technology for recycling spent rinse waters from semiconductor wet benches.
References
- Blackford, D. B., “The Measurement of Nonvolatile Residue in High-Purity Water,” J. Process Analytical Chemistry, vol. IV, Nos. 3,4, Winter 1998/1999, pp 92-98.
- Wilbowo, J., F. Shadman and D. Blackford, “Measuring and Removing Dissolved and Colloidal Silica in Ultrapure Water,” MICRO 15 (5), May 1997, pp. 41-42, 44, 46 – 50.
Acknowledgement. I thank Dr. David Blackford, Fluid Measurement Technologies Inc., for his critique of this column and CT. Associates for Figure 1, made available by courtesy of FSI International.
___________________
*ASTM D 5127-99 (Standard Guide for Ultra Pure Water Used in the Electronics and Semiconductor Industry) now lists “residue after evaporation” as a required UPW parameter and references D 5544 (Standard Test Method for On-Line Measurement of Residue After Evaporation of High-Purity Water), based on the NRM, for its measurement.