Testing of Ultrafiltration Systems
By Rick Andrew
Scope of ultrafiltration
Figure 1: Filtration spectrum
Ultrafiltration falls between nano and microfiltration in the filtration spectrum (Figure 1). UF systems can separate liquids from particles in the size range from about 25 nanometers to about 1,000 nanometers. This size range correlates roughly with the size range for viruses, and with particles with a molecular weight of roughly 2,500 to 250,000. The most common configuration for POU/POE ultrafiltration membranes is in the form of hollow fibers, with flow from inside the hollow fiber through the membrane to the outside of the hollow fiber. To form a module, the hollow fibers themselves are embedded in potting material at both ends of a cylindrical module housing. Often, small, relatively inexpensive POU ultrafiltration modules are configured with dead-end flow, meaning that there is no backwashing or forward flushing of accumulated filtered particles from within the hollow fibers. With this design, eventually, pressure drop across the module increases as the hollow fibers become clogged with particles. Typical POE UF systems include pressure sensors, electronic controls, and valves to facilitate backwashing, forward flushing, or both. This increased complexity of technology and design (as compared to dead-end flow) allows for increased life of hollow fiber modules because filtered particles are removed frequently enough that they do not cause pressure drop to occur.
The question often arises as to whether ultrafiltration systems can be evaluated to the protocols in NSF/ANSI 58 Reverse osmosis drinking water treatment systems. Although this standard does address membrane technology, the scope of NSF/ANSI 58 is strictly limited to RO systems only; ultrafiltration is definitely not reverse osmosis. RO incorporates a reject stream because of the crossflow design. As described above, however, UF modules use either dead-end design, or backwashing and/or forward flushing, not crossflow. Because the scope of NSF/ANSI 58 is limited to RO only, the test methods all assume a crossflow arrangement, which would not be applicable to ultrafiltration. Further, NSF/ANSI 58 requires all systems to reduce 75 percent of TDS as 750-mg/L sodium chloride in water. Considering where UF falls in the filtration spectrum, this level of ionic rejection is not achieved, so these systems would not meet this requirement of NSF/ANSI 58. With the inapplicability of NSF/ANSI 58 in mind, UF systems are tested per NSF/ANSI 42 Drinking water treatment units – Aesthetic effects and/or NSF/ANSI 53 Drinking water treatment units – Health effects. Because ultrafiltration systems provide mechanical filtration, or physical separation of particles from water, the claims under NSF/ANSI 42 and 53 that are potentially applicable to UF systems are the mechanical filtration claims (Figure 2). Ultrafiltration systems do not have any absorptive or adsorptive capabilities; they do not perform metals reduction or organic chemical reduction functions. There are some systems on the market that do incorporate activated carbon or other media along with the hollow fiber ultrafiltration technology to achieve additional performance, and these systems can be certified for additional claims under NSF/ANSI 42 and/or NSF/ANSI 53.
Mechanical reduction testing of hollow fiber ultrafiltration systems is conducted to the protocols described in the applicable standard. Test dust is used to build pressure drop and reduce flow rates through the system. The specification of the test dust particle size distribution is different for different tests and appropriate to the type of test being conducted. For nominal particulate reduction and turbidity reduction tests, test dust is used as the actual contaminant challenge test particle. A laser particle counter is used to count particles in the appropriate size range in influent and effluent samples, with an 85-percent reduction required. A nephelometric turbidimeter is used to measure turbidity in the influent and effluent samples, with a reduction in turbidity from 11 NTU in the influent to <0.5 NTU in the effluent samples required for turbidity reduction testing. For asbestos and cyst reduction testing, test dust is used only to build pressure drop and reduce system flow rate. Test-specific particles are introduced at sample points to serve as the contaminant challenge test particle used to establish system performance. Asbestos reduction requires 99-percent reduction of a mixture of both anthophyllite and chrysotile asbestos fibers, with analysis by either transmission electron microscopy or x-ray diffraction. Cyst reduction requires 99.95-percent reduction of either three-micron polystyrene microspheres or live Cryptosporidium parvum oocsyts. Analysis is by epifluorescent microscopy. All of these tests require analysis of multiple sets of influent and effluent samples to establish that sufficient percent reduction of the influent concentration has been achieved.
In order to achieve certification to NSF/ANSI 42 and/or 53, hollow fiber UF systems must conform to the material safety, structural integrity and product literature requirements of the standards, in addition to meeting the requirements of at least one contaminant reduction test. Material safety is established by conducting a formulation review of all materials in contact with drinking water, and then extraction testing of the whole product to assess the potential for any harmful compounds to leach from the water contact materials in the system into drinking water. Structural integrity is confirmed through a cyclic test, with repeated cycling between zero and 150 psig, as well as a separate test of elevated hydrostatic pressure for 15 minutes. Product literature must include a permanent system data plate, an installation and operation manual, a performance data sheet and replacement element packaging, if applicable. This literature must include specific information intended to help consumers better understand the benefits, requirements, and limitations of the products.
New technology, accurate and reliable test methods
POU/POE hollow fiber UF systems have become popular over the last few years. One might wonder about the effectiveness of this newer technology, or the accuracy of the methods of evaluation. Fortunately, the methods for determining material safety, structural integrity, and contaminant reduction performance are well-established methods that have been used very successfully for many years to evaluate other POU/POE mechanical filtration technologies. Manufacturers, retailers, dealers, and consumers can rest assured that systems conforming to NSF/ANSI 42 and 53 are safe, durable, and perform as advertised, despite the relatively recent emergence of this product category.
About the author
Rick Andrew is the Operations Manager of the NSF Drinking Water Treatment Units Program for certification of POE and POU systems and components. Prior to joining NSF, his previous experience was in the area of analytical and environmental chemistry consulting. Andrew has a Bachelor’s Degree in chemistry and an MBA from the University of Michigan. He can be reached at (800) NSF-MARK or email: Andrew@nsf.org.