By Rob Herman

In 1995, the U.S. Environmental Protection Agency (USEPA) and the Centers for Disease Control and Prevention (CDC) issued a joint announcement on how the public can protect themselves against accidental incursions of Cryptosporidium in public water supplies.

This announcement was spurred by the 1993 Milwaukee Cryptosporidium outbreak that occurred in the city’s water supply, sickening hundreds of thousands and killing 104 people. In the announcement, the USEPA and CDC recommended boiling water as the primary method of disinfection during a period when a public water supply is compromised. But they realized some of the public desired additional protection in their homes in case of contamination of their water supply. The additional technologies recommended were reverse osmosis (RO) systems, NSF-certified cyst reduction systems under ANSI/NSF Standard 53, or absolute 1-micron filters. The recommendation to use RO and NSF-certified systems was very clear and easy for a consumer to grasp and practice. The problem came with the recommendation to use absolute 1-micron filters, because there was no agreement within the industry exactly what an absolute 1-micron filter was. Even today, its meaning is controversial.

Going mechanical
Mechanical filtration refers to a process where particles are removed from the water by a physical barrier. A simple analogy is a window screen that keeps flies out of the house. The screen has holes that allow air to pass freely but keep anything larger than the holes outside, this is referred to as sieving. Some types of membrane filters use sieving exclusively, but most filters use a combination of sieving and entrapment to remove particles from water. Entrapment occurs when the particles pass into the body of the filter and are captured. Some membrane filters and all depth filters use both sieving and entrapment. Depth filters consist of a matrix of non-woven fibers or granular material where the water passage through the matrix is very tortuous (i.e., long and twisted) and the particles are captured primarily by impaction and adsorption on the filter material.

The most important characteristic of a mechanical filter is its efficiency in removing various size particles. For membrane filters that use sieving exclusively, if each hole (pore) is the exact same shape and dimensions, then we would only need to state the size of the holes—anything smaller than the hole passes through and anything larger is retained. Again, most mechanical filters are not just sieves, they use both sieving and entrapment so the effective pore sizes vary across the filter. It would not be unusual to have a filter that removes 90 percent of particles larger than 1 micron that only removes 95 percent of particles larger than 5 microns. This does not begin to discuss the issue of varying particle characteristics, such as shape, charge and flexibility, which in the case of cysts and oocysts, for instance, can allow them to pass through a pore that’s actually smaller than its average diameter.

A uniform approach
At the time the EPA/CDC recommendation was made, the point-of-use/point-of-entry (POU/POE) industry didn’t have a uniform definition of an absolute rated filter. In fact manufacturers used two descriptions for filter ratings, absolute or nominal. The general idea was that absolute filters removed more particles than nominal rated filters. Each manufacturer had its own definition of absolute and nominal so users had a difficult time comparing filters and determining which filter met their need. To lessen the confusion the Water Quality Association (WQA) proposed definitions in 1998 for absolute and nominal filter ratings.

“Absolute Filter Rating: Filter rating meaning that 99.9 percent—or essentially all—of the particles larger than a specified micron rating will be trapped on or within the filter.”

“Nominal Filter Rating: Filter rating indicating the approximate size particle, the majority of which will not pass through the filter. It’s generally interpreted as meaning that 85 percent of the particles of the size equal to the nominal filter rating will be retained by the filter.”

These definitions helped with some of the confusion, especially since the nominal filter rating reflected the ANSI/NSF Standard 42 protocol for particulate reduction that requires a minimum of 85 percent reduction at the particle size claimed. The wording of this test in ANSI/NSF Standard 42 has been updated to reflect that it’s a “nominal” particulate reduction protocol. The absolute definition has not been as fortunate since there’s still no standard that specifies a test to determine the absolute rating of a mechanical filter.

Actual reduction
The nominal particle reduction test in ANSI/NSF Standard 42 does not provide a means to establish an absolute rating. The maximum filtration that the test can verify is 99 percent reduction of a particle size range. This falls short of the 99.9 percent reduction required for an absolute rating. It cannot be assumed that if a mechanical filter reaches 99 percent reduction of 0.5-to-1.0 micron particles under ANSI/NSF Standard 42 this filter would also remove 99.95 percent of 3-micron particles as required in ANSI/NSF Standard 53 for cyst reduction. The creation of a definition of the absolute filter rating helped reduce confusion, but it didn’t solve the problem of determining when a filter actually achieved the 99.9 percent reduction and under what conditions.

The U.S. Food and Drug Administration (FDA) and the Health Industry Manufacturers Association (HIMA) had been using since 1982 a basic protocol to establish mechanical filtration efficiency for health related uses in clinical and research environments. This protocol challenged the filter with a large number of selected organisms and if any of the organisms passed through the filter, then the filter failed the test. Each type of organism used had a specific micron size that was associated with it and the filter then could claim that size. The difficulty with this type of testing is that it was designed for filters being used in a controlled situation. Using the same filters in a home did not guarantee the same results. A protocol was needed to validate that a filter would remove a minimum of 99.9 percent of 1-micron particles over its useful life in a home. This task was begun by NSF’s Drinking Water Treatment Unit (DWTU) Joint Committee and staff in 1999.

Early on in the standard development process it was recognized that an absolute 1-micron filter would be treated by the public as a health protection device. So, mechanical filtration took two paths of protocol development—one was to develop a protocol for determining the absolute rating of a filter under ANSI/NSF Standard 42 and a second was to evaluate mechanical filters for microbiological treatment. These tasks were assigned by the DWTU Joint Committee to two task groups, the Aesthetic Mechanical Filtration Task Group and Mechanical Filtration Disinfection Task Group. Since the primary concern is to evaluate health protection capabilities of mechanical filtration, the second group was given priority. This group recognized that mechanical filtration could be an acceptable method for the reduction of cysts and bacteria. The addition of bacteria to the test protocol required some additional evaluations, however. Cysts in water can be treated as though they’re simply particles as they don’t reproduce or change from their cyst state in this environment. Bacteria on the other hand can grow and reproduce in water and could actually grow through the filter to eventually contaminant the product water.

These issues and others are currently being addressed by this task group and should be balloted for inclusion into the standard by 2001. In the meantime adequate protection from cysts may be obtained through the other EPA/CDC recommendations which include boiling water for three minutes, reverse osmosis systems or NSF International certified systems for cyst reduction.

About the author
Rob Herman is technical manager of the NSF International’s Drinking Water Treatment Unit Program and has been with NSF since 1985. He can be reached at (800) 673-7275, (734) 769-0109 (fax) or email:

“The test for cyst reduction currently allows the selection of one of three options, a silica test dust using a particle counter (the old method), 3-micron microspheres with optical counting, or live Cryptosporidium. The silica test dust method will be phased out by March 2004 leaving microspheres and live cysts as the approved test methods.”


Comments are closed.