MC² Market & Competitive ConvergencePrefiltration ConsiderationsPresented at: "Effective Technological Methods Concerning Pharmaceutical Filtrations for the Production Specialist", March 3-5, 1987, Adam's Mark Hotel, Philadelphia, Pennsylvania USA Sponsored by: Society of Manufacturing Engineers, Educational Clinics Department
Introduction Prefilters are often used to extend the life of more expensive final filters. Additionally, they play a roll in increasing the particle removal efficiency of the filtration system. For many applications, prefiltration is an option and its cost must be justified. Whether or not prefiltration is beneficial depends upon the morphology and size distribution of the particles. The benefits derived from prefiltration must outweigh its inherent costs: additional housings and filters, added maintenance, and increased system differential pressure. The selection of prefilters is usually a cut-and-try iterative process involving the consideration of many factors. As with final filters, the efficiency of prefilters is dependent on fluid and particle characteristics. However, for non-membrane filters (i.e., fiber-based filters), particle removal efficiency is more time dependent: Their efficiency can change dramatically with time. Non-membrane filter ratings are not as well defined as their membrane counterparts and a clearer understanding of how they are rated can help to narrow down the filter choices. Dirt Holding Capacity Dirt holding capacity is a measure of particles which have challenged the filter over the useful life of the filter, which usually occurs when its differential pressure rises sharply. Note that nothing was mentioned about how many particles are retained by the filter. A filter with a large dirt holding capacity could have a very low particle removal efficiency. Clearly a better definition is required; however, this is the prevailing method used to report dirt holding capacity. Close attention must be given to all of the filter's specifications and conditions under which they were determined. The key factors that affect the measured dirt holding capacity are the particle characteristics and flow rate. The more closely the particle size distribution matches the filter medium's pore size distribution, the faster the pores will plug. In addition, soft deformable particles block pores to a higher degree than similarly sized hard particles. Numerous studies with various media, particles, and fluids have shown that dirt holding capacity increases as the flow rate is decreased. For this reason, the flow rate should always be stated whenever the dirt holding capacity is specified. Increases in effective filtration surface area while maintaining the same flow rate can achieve dramatic increases in dirt holding capacity. This is due to the difference in flow density, the flow rate per unit area. Increases in filter surface area decreases the flow density and increased dirt holding capacity results. The relationship of dirt holding capacity to filter surface area, flow rate and particle characteristics is expressed by the following formula:
Under theoretically optimized conditions, when the flow rate remains the same and the filter surface area doubles, the dirt holding capacity quadruples. This catalytic effect is due to particle characteristics. n approaches 2 when: To illustrate the effect this has when n approaches 2, the following theoretical
examples are given:
In this case, C1, which had a useful life of 8 hours and during that time, had a dirt holding capacity of 100g, is replaced by C2, which has the same medium, the flow rate is unchanged, and the surface area is doubled. Now, the service life and dirt holding capacity is increased fourfold. The flow density for C1 is 16 L/min/m2, and 8 L/min/m2 for C2.
In this case, C filters a batch in four hours. If the same flow rate is maintained, C2 will filter the batch in the same time; however, the DP is halved. The increased dirt holding capacity derived from choosing a filtration system with a larger surface area can be expressed by the following formula.
Economics of Prefiltration Whether or not prefiltration is economically justified will be primarily dependent on the characteristics of the particles being filtered, which in turn determine the power factor discussed above. In actual practice, the larger filter in our example would have an expected life 2.5 to 3.0 times longer. Two examples are offered below to illustrate the affect various particles can have on filtration economics under more realistic conditions. Example 1 During trial filtration runs, it was determined that a 25 cm, 0.2µm membrane cartridge costing $100.00 could remove 34 particle units (arbitrary units) from a given fluid system. The particle characteristics are such that the power factor approaches 2; i.e., particles uniformly sized, noncompressible and cake filtration efficiency does not exceed the medium's Since the batch we will filter contains 100 particle units, a 50 cm version of the membrane final filter is used. The differential pressure equals 0.25 kg/cm . Because n approaches 2, the 50 cm filter has a dirt holding capacity of 100 units, enough to complete the batch. The filter cost $200.00. In the next experiment, a prefilter costing $50.00 and capable of removing 80% of the particle units is placed upstream of the 25 cm 0.2µm membrane cartridge. The prefilter pressure differential is 0.1 kg/cm² and 0.5 kg/cm² for the final filter, for a system pressure drop of 0.6 kg/cm². The same batch is filtered at the same flow rate and the final filter is challenged by 20 units, 80 units having been removed by the prefilter. The total filter cost is $150.00. Although we have saved $50.00 in filter expense by the use of a prefilter, we have incurred the expense of an additional housing which will have to be cleaned and maintained. Also, the cost of operating the filtration system has increased significantly with the addition of the prefilter because the differential pressure more than doubled. Under these conditions, it would probably be more economical to invest in larger surface area final filters. Example 2 In this case the batch is different, consisting of bacteria and many organic particles. During trial filtration runs, it was determined that a 25 cm 0.2µm membrane cartridge costing $100.00 could remove 34 particle units. The particle characteristics are such that the power factor approaches 1; i.e., particles are of varying sizes, and compressible, and cake filtration efficiency exceeds the medium's. It is determined that the batch, containing 100 particle units, can be filtered using a 75 cm version of the 0.2µm membrane final filter. The filter costs $300.00 and has a differential pressure of 0.25 kg/cm² In the next trial run, a prefilter costing $50.00 and capable of removing 80% of the particle units is placed upstream of the 25 cm 0.2µm membrane cartridge. The prefilter differential pressure is 0.1 kg/cm² and 0.5 kg/cm² for the final filter, for a system pressure drop of 0.6 kg/cm² The same batch is filtered and the final filter is challenged by only 20 units, 80 units having been removed by the prefilter. The total filter cost is $150.00. In this case, the particles' characteristics caused n to approach 1 and the economics of prefiltration changed accordingly, $300.00 without prefiltration versus $150.00 with prefiltration. Although we have the additional expense associated with the prefilter housing, we can start to see the substantial savings that are possible with prefiltration use. The Impact on System Performance Because the final filter's particle removal efficiency and life is heavily dependent on particle characteristics, the addition of or changes in prefiltration can benefit or detract from system performance. Our experience has shown that it is possible to actually shorten the life of the final filter with the addition of prefiltration. In this case the prefilter is retaining the particles in the size range which contributes to final filter efficiency and life by partially blocking pores and cake development. The number of particles challenging the final filter also affects performance. The presence of a high concentration of particles, for instance, could lead to flocculation which effectively increases particle size and retention efficiency. In this case, the addition of prefiltration could reduce retention. Of course, the degree of attraction the particles have for each other will be dependent on the chemistry of particles in a given fluid. However, in most cases, prefiltration lowers particle levels to the extent that the system's filtration efficiency increases. Since the probability of particles passing through the final filter medium is dependent on the number of upstream particles, the reduction in particle burden with prefiltration can lower the risk of contamination downstream or the final filter. Prefilter Types Integral prefiltration layers, which are available with many polymeric membrane final filters, can, under certain conditions, extend filter life and, under all conditions, change final filter performance. Polymeric membranes used as prefilters have the advantage of high particle removal efficiency for particles equal to and larger than their ratings. However, membranes have very low dirt holding capacities when compared to non-membrane fiber-based cartridges. Fiber-based cartridges can be classified into 3 distinct medium types. These cartridges usually have dirt-holding capacities which far exceed the membrane filters'. • Needled Fiber. Fibers are intertwined through a needling process. It has the advantage of not having any binders. However, fibers can slough off. Prefilter Particle Removal Ratings While the removal ratings for final membrane filters are well defined, there is no universally accepted standard method for rating nonmembrane filters, although there are some guidelines and practices in the literature. Most non-membrane filters can not approach the particle removal efficiency of a membrane-type filter. Perhaps a very efficient microfiber filter could, when challenged with Pseudomonas diminuta, achieve a log reduction value (LRV) of 2 (an LRV of 2 means that one organism out of every 102 bacteria put on the filter appears in the filtrate). This is compared to "sterilizing" 0.2µm membrane filters that can achieve an LRV of 7. However, if we judged the non-membrane filters solely on LRVs, their true benefits would be overlooked. We must go beyond LRVs and assess their dirt holding capacity and particle removal efficiencies over their useful life. Unlike their membrane counterparts, the particle removal efficiencies can change quite dramatically with time. For this reason, the filters are rated on their ability to retain particles over their service life. See figures 1 and 2 . A well accepted test method involves challenging the filter with a constant influent at a uniform flow rate (aqueous dispersion of A.C. Test Dust for example). The filter is exposed to any given particle only once (single pass). Samples, taken upstream and downstream of the filter at timed intervals throughout the filter's life, are analyzed on a multichannel particle counter (usually 16 channels). See figure 3 . The counts in each channel are time averaged and the results presented as a Beta Ratio (ß) and % Cumulative Removal Efficiency using the following formulas:
For Example If over the useful life of the filter, the time averaged upstream counts for particles 1µm and larger equaled 100,000 and the downstream particle counts for particles in the range equaled 10 then:
Therefore, over the service life of the cartridge, on the average, one particle out of every 10,000 challenging the filter 1µm and larger passes through the filter or, inversely, 9,999 particles out of every 10,000 are retained for a removal efficiency of 99.99%.
While this test is useful in comparing filters, caution crust be used when comparing data. For instance, there is no correlation to single-pass methods and methods where the effluent is looped back to the filter (multipass method). Also, attention must be given to the flow rate conditions of the test because filter efficiency can be very sensitive to flow rates, with lower flow rates resulting in possibly significant increases in efficiency. See Figure 4 . In addition, some filters are rated for particles larger than a specified particle size; a throw-back to hydraulic fluid filter ratings. How much larger is left up to the user to decide. There are certain fiber-based filters whose high efficiency in the submicron range rules out over-life testing with A.C. Test Dust. For these cartridges, monosized latex challenge data is often provided in addition to bacteria retention test results. The rating of filters as "absolute", more than anything else, illustrates the scope of the prefilter rating problem. Studies indicate that the term absolute! is not a precise term to describe the expected performance or a 0.2µm membrane final filter. So when this rating is used to classify a non-membrane prefilter for use with a "sterilizing" 0.2µm membrane, it is all the more ridiculous. Prefilter Quality Many of the same quality levels that are demanded for final filters should be sought in prefilters. If appropriate, supporting documentation concerning chemical compatibility, extractables, media migration (sloughing), biological safety, purity, sterilization procedures and retention should be demanded. Some manufacturers are now supplying prefilter validation guides, which previously were only available for final filters. If the filter is capable of integrity testing by bubble point, the manufacturer should have a statistical in-process sampling plan to ensure filter-to-filter consistency. Just as the final filter's integrity is correlated to actual retention testing, so should the prefilter. Conclusions The economical benefits of prefiltration use is dependent on particle uniformity and morphology along with the efficiency of the filter cake. The more the particles vary, compress, and form a cake more efficient than the medium, the more beneficial prefiltration will be. Prefilters impact on the final filter's particle removal efficiency and life, for better or worse. Not unlike any filtration application, prefiltration performance will depend on particle and fluid characteristics and how they interact with the filter medium. Conducting a complete suspension analysis is the first step in identifying the requirements of the filtration system. Selecting the appropriate prefilter cartridge is a complex process involving trial and error. Fortunately, there are a wide variety of cartridges available to serve as prefilters. This wide selection can also add to the complexity because of their wide variations in performance and the lack of any standardized testing. A single-pass over-life test using time-averaged particle counts is an appropriate test for narrowing down the selection of certain nonmembrane prefilters. Bibliography (1) ASTM Practice for determining the Performance of a Filter Medium Employing a Single-Pass, Constant-Pressure, Liquid Test, F 795-82, 1985. (2) ASTM Practice for Determining the Performance of a Filter Medium Employing a Single-Pass, Constant-Rate, Liquid Test, F 796-82, 1985. (3) ASTM Practice for Determining the Performance of a Filter Medium Employing a Multi-pass, Constant-Rate, Liquid Test, F 797-82, 1985. (4) Bensch, P. E. , "Dirt capacity: the misused selection factor," Hydraulics & Pneumatics, November, 1985, pp 55-59. (5) International Organization for Standardization, "Hydraulic Fluid Power - Filters - Multi-pass method for evaluating filtration performance," ref. no. ISO 4572-1981 (E). (6) Johnson, P.R. and Meltzer, T.H., "Comments on organism challenge levels in sterilizing filter efficiency testing," Pharmaceutical Technology, 3; November, 1979, pp. 66-70, 110. (7) Trasen, B., "Designing process microfiltration systems," Pharmaceutical
Technology, November, 1981, pp. 62-64, 66-67, 69.
|
||||||||||||||||||||||||||||||||||||||||||||||