Gas phase filtration for cleanrooms
The major types of gas phase filters used to control low-level gas contaminants are compared.
By Fred Dwiggins, Hoechst Celanase Corp.
Designers of all types of cleanrooms — particularly those serving the microelectronics, pharmaceutical and similar sensitive manufacturing industries — will increasingly have to contend with the problem of airborne molecular contamination (AMC). Aside from considerations about contamination and its impact on product quality, it is clear that growing regulatory concerns for personnel health will increase.
This article is concerned with molecules in gas or vapor phase — in effect, anything other than oxygen and nitrogen in the air — and removing them to improve air quality in cleanroom applications. Some of these gases are detrimental to the quality of products such as electronic parts and pharmaceuticals. The focus of this article is to effectively use gas phase filters to remove AMCs from make-up air as well as recirculating air systems.
This article compares the major types of gas phase filters used to control both low level gas contaminants and the spikes of limited duration at slightly higher concentrations typically found in indoor air quality situations of relevance to cleanroom activities.
Adsorption
Adsorption is commonly used to remove low level AMCs from air streams. The sorbent employed should be a porous material in order to provide a large amount of surface area, most of which will be on the interior. In physical adsorption, the gas molecules are physically attracted into and held on the surface of the adsorbent. In chemical adsorption, or chemisorption, gas molecules are physically attracted into the sorbent, but are then chemically converted into other substances.
Commonly used gas phase filters activated carbon for adsorption because of high specific surface area (up to 1,500 m2 per gram). It is an excellent physical adsorbent for many gases by itself. It can also be impregnated with many materials for chemisorption of specific degrees.
Carbon particle size is critical in its performance as an adsorption medium, with smaller particles permitting more active surface areas to be exposed to the challenge, which improve the adsorption performance. However, smaller particles are more difficult to immobilize in an air stream and can restrict air flow when densely packed. Larger particles are often more easily used, but because of spacing between the particles, are subject to channeling and media by-pass. In addition, accessing the cores of the larger particles is more difficult, thus making the total particle less efficient.
Residence or contact time of the air and contaminants within the media also plays an important role. Greater residence time increases the likelihood that the contaminant is adsorbed. This is typically achieved by increasing the media or filter surface area and thus decreasing the velocity through the media; or by increasing the thickness of the media, which results in increased pressure drop.
Filter types
Three major types of gas phase filters exist: carbon slurry coated foams or nonwovens (CSC materials), loose carbon trays (LCT), and carbon-loaded nonwovens (CLN). In addition, owing to differences in technology and manufacturing, there are great variations in performance within each class.
As their name implies, CSC products are typically made by coating an open foam or nonwoven with a slurry of powdered carbon in an adhesive to bond the carbon to the substrate. These products are mostly used as flat panels or shallow pleats, and have very low pressure drop. However, the amount of carbon incorporated is minimal, and most of the carbon surface is blocked by adhesive. Thus, the adsorption performance is poor. These products are typically found in 1- to 2-inch deep filters.
LCT products are usually in 1- to 3-inch deep trays filled with larger granulate particles. If packed tightly, the pressure drop can be very high. Thus, they are usually found in V-bank configurations 12 inches or more deep to increase the surface area, resulting in reduced pressure drop and increased residence time. This requires a large space and the pressure drop can still be relatively high. Adsorption preference can be quite good due to the volume of material available.
Unlike the other types of carbon filters, these products typically require a post filter due to dusting, which increases the overall pressure drop and cost of the system.
CLN products are a hybrid filter solution in which smaller granular carbon is physically bound in a pleatable nonwoven media. The carbon particles are smaller than those used in the LCT products, which exposes a greater amount of surface area. Little of this surface is blocked as in the CSC products. The carbon is more or less uniformly dispersed and held in place. This promotes good gas contact, reduces channeling, lowers the pressure drop, and reduces or eliminates dusting during changeout and use. High pleatability of the media allows greater surface areas with the resulting lower pressure drop and increase residence time. These can be found in 1- to 12-inch pleated filters.
Filter comparison
Figure 1 shows a pressure drop comparison for representative media in commercially available 24- ¥ 24-inch filters. The two CSC products are nominally 2-inch pleated panels. The two LCT products have eight 1-inch trays in a V-bank configuration which is 12 inches deep. The two CLN products are nominally 12-inch deep pleated filters. As expected, the pressure drop of these products falls between CSC and LCT products. However, CLN product “E” has a significantly lower pressure drop at higher air flow rates typical for the intended applications. (Note that LCT product “D” has a low pressure drop due to its “partial bypass” construction.)
Adsorption performance testing
Difference in adsorption performance can be determined by dynamic adsorption testing at operating condition. Briefly, this test exposes the filter to flowing air spiked with test gases at a controlled temperature and humidity, and monitors the inlet and outlet concentrations over time. Currently, the industry is looking for better ways to measure and compare filter performance and responses to various challenges. Dynamic testing at model end-use conditions gives information on both efficiency and capacity over time. Today, about a handful of companies are capable of testing filter construction under dynamic conditions. ASHRAE has a research program and committee looking at developing a test to measure the efficiency and capacity of full-scale filters to remove a variety of common contaminants in the parts per billion range.
A typical breakthrough curve for a constant rate challenge is shown in Figure 2. The key information from this curve is the initial breakthrough and the total adsorption capacity 100 percent breakthrough. Preferably, the initial breakthrough would be 0 percent for some time as shown. The total adsorption capacity, which is the area above the curve, can be used to predict filter service life given a known, constant challenge and the desired level of removal. Note that low initial breakthrough also suggests good recovery from spikes and intermittent exposure. This can be confirmed by testing with intermittent challenges, and may better simulate actual end-use conditions. A typical breakthrough curve for an intermittent challenge is shown in Figure 3.
Dynamic adsorption test results
Figures 4 and 5 show typical breakthrough curves for the different filter media against the two challenges when exposed at a constant rate. Note that all of the media contain unimpregnated carbons. This allows a more direct comparison of the results as impregnates could change the performance dramatically.
For toluene, the CSC products and LCT product “D” show almost immediate and total breakthrough and very low capacity. It is doubtful they would be of much use even of low or short challenges. LCT product “C” shows no initial breakthrough, and a high capacity until 100 percent breakthrough. CLN product “E” also shows no breakthrough initially, and a comparatively high capacity until 100 percent breakthrough. CLN product “F” performance is worse than “E” or LCT product “C,” but better than all of the others.
For an intermittent challenge, these curves suggest that the CSC products would have essentially immediate breakthrough and no recovery. Similar performance would be expected from LCT product “D” and CLN product “F.” The other LCT and CLN products would be expected to have little or no breakthrough and recover well for multiple intermittent challenges. Such performance has been observed in lab testing.
For H2S, all products show high initial breakthrough and low capacity except for CLN product “E”, which has minimal initial breakthrough and exceptional capacity. CLN product “E” would be the only one expected to show some performance and recovery for intermittent challenge, which again has been demonstrated in the lab.
Pressure drop and dynamic adsorption performance are two key parameters to consider when evaluating gas phase filters for IAQ applications.
From the above it is clear that CLN product “E” would perform well for the cleanroom applications intended. It can remove a broad range of continuous low-level contaminants, or intermittent peak challenges of modest frequency and severity, with good performance over time. This, coupled with the other advantages of low pressure drop, no dusting, and ease of handling make it an excellent choice for these applications. In addition, when considering removal of ammonia from the atmosphere, these filters have a service life of hundreds of days, while for boron compound removal they have a service life of hundreds of thousands of hours.
Summary
An effort has been made to shed some light on the complexities of airborne molecular contamination in indoor air, and the various solutions available. In order to arrive at an appropriate measured solution to the AMC problem, one must understand both the challenges of the application plus the capability of the solution. Testing the filters at realistic conditions and determining the challenges in the field are key to that understanding.
Carlton Frederick Dwiggins, applications manager in the Air Quality Filtration Group at Hoechst Celanese Corp. (Charlotte, NC), has been with the company for the past 15 years. A member of ASHRAE and the Institute of Environmental Sciences, Dwiggins holds a master of science degree in chemical engineering from North Carolina State University.
 |
 |
 |
 |