Keywords: Electrets, Air Filtration, Heating, Ventilation,and Air Conditioning, Field Studies, 2-PropanolConditioning.

AbstractAir filters produced with charged, or electret, media for

HVAC filtration applications have gained significant marketshare and acceptance over the past few years. While these fil-ters provide the advantage of high initial efficiency and lowpressure drop, there are concerns about their ability to main-tain efficiency in service. Furthermore, there is a tendency toclassify all electret media in the same general category with-out any consideration of media structure, fiber size, andcharging technique. Current research suggests that a varietyof factors influence the loss of efficiency in use includinghumidity, exposure to certain chemicals, aging, temperature,and etc. While this is true of some charged media, the effectof environmental factors on filtration performance is highlydependent upon the media technology itself. This paper pro-vides an overview of current electret media types detailingmedia structure and charging techniques. Fundamentalimpact of environmental factors on filtration performance ispresented along with field studies detailing in-use perfor-mance of filterers manufactured with charged media.

INTRODUCTIONElectrets and Electrostatically Charged Filtration MediaCharging Techniques (Electret Forming Methods)

An electret is broadly defined as a dielectric material, whichexhibits an external electric field in the absence of an appliedfield (1). Many researchers have likened this definition withthat of a magnet, noting that an electret is the electrical equiv-alent of a magnet. The charging techniques used in the for-mation of electrets vary widely, and are dependent on thephysical form of the dielectric (i.e. film, fiber, foam, or non-woven), and the type of electret that one is attempting to

make. In order to fully describe and understand the differentcharging methods it is important first to briefly talk about thedifferent types of electrets that make up this important class ofmaterials.

Types of ElectretsElectrets can be divided into two distinct categories of

materials. Space charge electrets are those that possess aninjected or imbedded charge within the dielectric (1). Dipolarelectrets, as their name suggests, are formed by the orienta-tion of dipoles (i.e. polar groups) within the dielectric (1).Space charge electrets are typically formed by techniques thatdeposit or inject charge directly into the dielectric material.Dipolar electrets are formed, or polarized, by the applicationof an electric field to the material either at ambient tempera-ture, or by heating the material to a higher temperature whileapplying an external electric field and then cooling to somelower temperature while the external field is maintained. Indipolar electrets, the reorientation of electrical polarizationcan only be achieved at temperatures where the dipoles aremobile. For most polymers of interest this occurs above theglass transition temperature. Dipolar electrets can also beformed by charge injection techniques wherein the electricfield due to the imbedded charge causes dipole reorientation.

Electret Charging (Forming) TechniquesAll electret-charging techniques are limited by internal and

external electrical breakdown. The occurrence of internalbreakdown depends on the dielectric strength of the materialbeing charged. Most polymers have dielectric strengths onthe order of a few MV cm-1, thus polymers are able to sustaincharge densities of a few tenths of a µC cm-2 without danger ofbreakdown (2). The occurrence of external breakdown is verydifferent. Many electret-charging techniques rely on externalbreakdown to generate charge carriers, which can depositcharge on the dielectric. Corona charging, for example, is

Electret Media For HVACFiltration ApplicationsBy David L. Myers and B. Dean Arnold, Nonwoven Fabrics Research and Development, Kimberly-Clark Corporation, 1400 Holcomb Bridge Rd., Roswell, GA 30076

ORIGINAL PAPER/PEER-REVIEWED

43 INJ Winter 2003

based on the application of an electrical potential between twoelectrodes sufficient to create a corona discharge in the spacebetween the electrodes, which is referred to as the gap. Themajority of the energy dissipated in a corona discharge goes toexcitation of the gas present in the gap and is responsible forthe characteristic blue-indigo glow of the corona. Charge car-riers are generated in the gas phase by a complex series ofelectron-molecule and ion-molecule collisions. The interac-tion of these charge carriers with the dielectric results in thedeposition of charge at the surface.

Triboelectric Charging. Triboelectric charging is associatedwith the transfer of charge due to frictional contact betweendissimilar dielectric materials. Triboelectric charging is com-posed of kinetic and equilibrium components. The kineticcomponent arises from the energy dissipated when two dis-similar materials are rubbed together. This leads to frictionalheating at the area of contact, and thus charge transferbetween the materials. The equilibrium component is alsoknown as contact electrification. This arises from static con-tacts between different dielectric materials, leading to a trans-fer of charge. It is widely believed that contact electrificationis affected by the relative electron affinities of the two contact-ing materials and differences in their work functions (3-6).Both of these quantities are related to how tightly bound elec-trons are to their respective nuclei. Many researchers havecompiled extensive data organized into a triboelectric series(7-9). The triboelectric series is arranged from materials thatare electron accepting to those that are donating. The ratio-nale being that one can choose materials with electron transferproperties that are complimentary.

Triboelectric charging requires intimate contact between thetwo surfaces being charged. The magnitude of charge trans-fer between the contacting surfaces is dependent on a numberof variables, including: area of contact, force of contact, sur-face contamination, and environmental conditions such astemperature and relative humidity. Triboelectric charging hasnot been extensively used as an electret charging method duein large part to a lack of accurate reproducibility stemmingfrom the variables cited above (2).

Triboelectric charging has been used in the manufacturingof filtration media; however, there are limited numbers ofpolymer combinations that lend themselves to contact electri-fication. The earliest example of an electrostatic filter is theHansen resin-wool filter (9). A combination of wool fibers andresin particles were carded together resulting in a filter feltwhere the resin particles were negatively charged and thewool fibers, positively charged. A more recent example ofcontact charging of polymers is polypropylene andmodacrylic. Modacrylic is a copolymer of acrylonitrile andvinyl chloride monomers. In needled felts and carded web fil-tration media, this combination of polymers in the form of sta-ple fibers come in intimate contact during the manufacturingprocess. The polypropylene filaments acquire a net positivecharge and the modacrylic a net negative (9). One difficultyencountered with this approach to making an electricallycharged filter medium is the fiber finish present on both thesematerials. While the finishes provide a lubricious surface for

processing, they interfere with the transfer of charge betweenthe materials.

Another method of charging which appears to rely heavilyon triboelectric effects is hydrocharging (10). Hydrocharginginvolves the use of a high pressure water jet directed througha nonwoven filter media in order to impart charge. It isreported that quite good filtration media can be made usingthe hydrocharging technique (10). A drawback of thehydrocharging technique is the drying of the charged filters.If the drying is done at an elevated temperature, thermalcharge carries are liberated within the dielectric leading tocharge dissipation. If the web is insufficiently dried, theresidual moisture can lead to fungal or bacterial contamina-tion of the filter media during storage.

Thermal Charging. Thermal methods of charging electretsall involve the application of an electric field to a dielectricmaterial at elevated temperature and subsequent coolingwhile the field is maintained. The early work of Eguchi (11,12)on carnauba wax electrets involved the melting of the wax inan electric field followed by rapid cooling with field stillapplied. In electrets prepared from polar dielectrics, wherethe electrodes are vapor deposited directly on the dielectricmaterial, the thermal charging gives rise to dipole orientation.However, the use of external electrodes results in air gaps atthe dielectric/electrode interface that can lead to very compli-cated charging phenomena. Electrets made using this charg-ing method are called thermoelectrets.

Isothermal Charging (Corona Poling and Spark DischargeMethods). Isothermal charging is a somewhat misleading termin that many charging techniques can be carried out at constanttemperature. However, isothermal charging is widely accept-ed as referring to corona poling and spark discharge (1).Corona poling, as its name implies, involves the application ofa corona discharge to the implantation of charge in a dielectricmaterial. The most significant advantage of corona poling isthe speed of charging and the ease with which the corona isgenerated. Corona discharge relies on the breakdown charac-teristics of the gas present in the gap between a pair of elec-trodes. Many electrode geometries have been described (13),however, all rely on the generation of charge carriers in the gasphase for charge deposition on the dielectric. The majority ofthe energy dissipated within the corona discharge goes to theexcitation of the gas. Collisions leading to energy loss result incharge carriers having energies of only a few tens of eV. Thesecharge carriers deposit charge on the dielectric surface atdepths of only a few nanometers (1 x 10-9 m). Over time, chargetrapped at the surface can move into the bulk and become re-trapped at depths of several microns. This occurs only if suffi-cient charge carriers exist within the dielectric. A more exhaus-tive description of corona discharge and charge deposition canbe found in references 13-15 and reference 1, respectively.

Corona poling can be applied to the charging of any dielec-tric material independent of physical form (i.e. film, fiber, orfoam). Corona poling leads to the trapping of charge in ener-getically stable surface traps. The main disadvantage of coro-na poling is the non-uniform deposition of charge comparedto that obtained by thermal charging techniques. Filtration

44 INJ Winter 2003

media charged by corona poling likely benefit from this charg-ing heterogeneity because it creates a non-uniform charge dis-tribution within the media structure. This leads to very largeinhomogeneous electrical fields within the media structureand increases the particle capture properties of the filter.

Spark discharge methods are similar to corona poling withthe exception that an additional dielectric material of lesserresistance is introduced between the charging electrodes andthe material being charged. The second dielectric acts a seriesresistor and prevents spark discharges from the electrodesfrom degrading the material being charged. Spark dischargeis typically not used in the large-scale production of electretmaterials. The reader is referred to the following referencesfor further information on spark discharge charging methods(2 and references therein).

Electron and Ion Beam Charging. Electron and ion beamshave been used for the charging of film electrets, but typicallynot used for charging nonwoven or fibrous filtration media.Charge implantation with low energy electron or ion beamsrelies on the generation of a secondary electron cascade as aresult of scattering of the primary beam within the bulk of thedielectric. Low energy secondary electrons and the sloweddown primaries become trapped within the dielectric yieldingan electret state, which depending on the material can have avery high stability. High-energy electron and ion beams (i.e.ionizing radiation) do not work well for electret chargingbecause of the chemical damage caused to most dielectricmaterials as a result of radiation exposure. The damage leadsto induced conductivity that destabilizes the implantedcharge leading to recombination of positive and negative cen-ters (2).

The Photoelectret Process. The photoelectret process is men-tioned here for completeness of this brief review. This processis used in the charging of photoactive materials (photocon-ductors, amorphous semiconductors, and the like). Theprocess involves the application of an external electric fieldduring exposure of the photoactive dielectric to light of theappropriate wavelength. Light absorption gives rise to chargecarriers that are separated spatially by the applied field (2).This is not used in the charging of nonwoven or fibrous filtra-tion media.

Electrostatic Filtration MediaElectrostatic filtration media encompass a broad class of

materials that are capable of capturing and retaining airborneparticulates through electrostatic interactions. The electrosta-tic interactions are the direct result of the electric fields gener-ated by the presence of either uncompensated space charge ororiented dipoles present within the media structure. Manytimes electrostatic filtration media are referred to as "chargedmedia". It is important to note, that the charge implantationoccurring in the formation of an electret gives rise to stronginhomogeneous electric fields within the bulk of a filter mediastructure. It is the presence of these inhomogeneous electricfields that gives rise to the enhanced filtration properties ofelectret charged media. Inhomogeneous electric fields giverise to field gradients, which are necessary to cause the move-

ment of particles within the filter structure. This is especiallyimportant for the attraction and capture of non-charged parti-cles.

The capture of aerosolized particulates by electret chargedfiltration media relies on both Coulombic and dielec-trophoretic forces. Aerosols are characterized by a distribu-tion of neutral and charged particles. The distribution of neu-tral versus charged particles is governed by a Boltzmann dis-tribution, which is a function of temperature and aerosol com-position (16). Neutral particles are captured by electret filtermedia via polarization effects induced by the inhomogeneouselectric field within the media. Dielectrophoresis describesthe motion of particles with either induced polarization or realcharge within the media structure. A field gradient is neededto induce forces in neutral particles and cause their movementwithin the structure until capture. Charged aerosol particlesare captured by strong Coulombic forces acting on the parti-cles as they move in close proximity to the fibers that make upthe porous filter structure.

Despite the common feature of all being “charged” filtrationmedia, not all electret charged media are the same in terms ofstructure, charge distribution, stability, and magnitude of theinternal electric fields. The earliest electrostatic filter, theHansen resin/wool filter mentioned above, was charged byvirtue of the frictional effects that occurred between woolfibers and pine resin particles (9). Its filtration properties like-ly benefited from the irregular shape of the resin particles; thecharge on these particles would have given rise to very com-plex electric fields within the filter structure.

Modern electret charged media differ from the Hansen filterin the sense that most are made from combinations of fila-ments. The filaments can have different cross-sectional area,shape, and composition. In addition, the filaments can have afinite length as in stable fibers for carded web media or the fil-aments can be continuous as in continuous filament meltspunmedia. Staple fiber filaments are typically mechanicallycrimped to improve how the filaments entangle during thecarding process. Staple fiber media can be charged using anyof the techniques discussed above. However, corona polingand triboelectric charging is the most common. In some casesthe staple fiber filaments are made by fibrillation of a film,which has previously been electret charged (18,19). In thiscase, the final charge distribution in a finished filter mediadepends on the arrangement of charged filaments in the webstructure.

Triboelectric Charge in Filtration Media. Triboelectric charg-ing of filtration media is a direct consequence of the intimatecontact between fibers within the filter media structure.Virtually all triboelectrically charged filtration media consistof two or more polymers, chosen for their characteristic chargetransfer properties. As discussed above, the biggest drawbackto triboelectrically charged media is the inconsistent mannerin which charge is deposited within the media structure (2).The carding process yields a filter web with directionallydependent properties due to orientation of fibers in themachine direction. This necessitates the lapping of multiplewebs and a process called needling in order to produce a

45 INJ Winter 2003

densely packed randomized web structure.The charge transfer mechanisms proposed for triboelectric

charging involve surface states on the polymers (20-22). Theinjection of electrons or holes leads to the development of neg-atively and positively charged domains, respectively.Williams et. al (22) examined triboelectrically charged poly-mers using x-ray photoelectron spectroscopy (XPS). Thistechnique provides detailed chemical information about thesurface of the material analyzed. Through the use of very thincoatings on a variety of polymer surfaces, they determinedthat the penetration depth for triboelectric charge was on theorder of a few nanometers. Surface charge states are suscep-tible to ion pair absorption from the surrounding air, leadingto charge compensation. In addition, surface contaminationby lubricants used in the manufacturing process can give riseeither to surface conductivity, which destabilizes trappedcharge, or charge screening. Surface conduction leads toeither charge compensation or recombination; each of theseprocesses decreases the macroscopic electric field due to thetrapped charges, and in turn, degrades filtration properties.

Fibrillated Films. Among the earliest examples of electro-statically charged filter media were fibrillated films. Theprocess for making a fibrillated charged film is quite simple.A non-metalized film is electret charged using one of thecharging techniques described above. The most commonmethod of film charging is corona poling. The charged film isslit and stretched to form filaments with a ribbon-like struc-ture and dimensions on the order of tens of microns. The rib-bon-like filaments are then carded into a filter web. The charg-ing of film electrets has been very extensively studied. Filmelectrets were first used as components of electret micro-phones (23). The first example of a fibrillated electret film forair filtration came in the mid-1970’s with the invention of”Filtrete” by van Turnhout et. al. (18). The electret charging ofa film leads to the deposition of positive and negative chargeon opposite sides of the film structure. The fibrillation andcarding processes serve to produce a somewhat randomarrangement of fibrils within the finished filter web. The ran-domness of the structure is limited, however, by the webforming process. This in turn limits the degree of anisotropyin the distribution of charged domains within the filter webstructure.

Fine Particle Filtration. The most significant benefit of elec-tret charged filtration media is the ability to remove verysmall, aerosolized particles while maintaining low-pressuredrop through the filtering medium. Fine particle filtration isdefined as the removal of aerosolized particles below 1micron in diameter. Sub-micron particles are much smallerthan the void spaces present in most commercial electretmedia, yet due to the electrostatic forces within the mediastructure, they are removed with high efficiency. Small parti-cles can also be problematic for electret filters. The effect ofsmall particle contamination was recognized as early as thepioneering work of van Turnhout and coworkers (17). Theyfound that the accumulation of dust in electrostatic filters pre-pared from fibrillated films degraded their filtration proper-

ties due to screening of the electrostatic forces (17). In fact,several studies of the dust loading and filtration performanceof electrostatically charged media clearly showed that thedeposition of sub-micron (<1 mm) particles leads to a pro-gressive rise in filter penetration to dust loadings whereinmechanical entrapment begins to improve filter performance(24).

Conditioning of Electret Media: “Charge versus Structure.”Concerns over the fine particle contamination of electrostaticmedia have spawned a number of initiatives aimed at "remov-ing" the charge from the media in order to study mechanicalfiltration properties (25). It is widely held that one can divorcethe charge of an electret media from its structure, thus allow-ing one to study the mechanical filtration properties of the fil-ter media independent of the electrostatic properties. The"conditioning steps or protocols" have called for the immer-sion of charged media in 2-propanol prior to dust collectionefficiency testing (25). Another, calls for exposure of themedia to the exhaust gas of a diesel internal combustionengine (25). The argument being made is that by "removingthe charge" an accurate assessment can be made of themechanical filtration properties of the media.

The proliferation of electrostatically charged filtrationmedia in both the commercial/industrial and consumer filtra-tion markets has been accompanied by several misconcep-tions regarding this important class of media. Chief amongthese is the tendency toward grouping all electrostaticallycharged media together as if all are the same. In this paper wehighlight the filtration performance characteristics of continu-ous filament meltspun (CFM) media, which have been electri-cally charged to create a unique combination of structure andelectrostatic charge. We present field study data collected forboth electrically charged and non-charged media of the samestructure, which illustrates the underlying mechanical filtra-tion properties of the CFM media structure. In addition, bothaccelerated aging and long-term aging data will be presentedin support of the stability of the electrostatic charge on themedia. Finally, a critical evaluation is presented of the effectsof the proposed 2-propanol conditioning step on the physicalproperties of polyolefin based filtration media.

EXPERIMENTALMaterials. Continuous filament meltspun filtration media

were manufactured in accordance with the teachings of the fol-lowing United States patents: 4,340,563 to Appel et. al.,3,692,618 to Dorschner et. al., 3,802,817 to Matsuki et. al.,3,338,992 and 3,341,394 to Kinney, 3,502,763 to Hartman,3,542,615 to Dobo et. al., and 5,382,400 and 5,795,926 to Pike et.al.. The electret charging of CFM media was accomplishedaccording to the teachings of Tsai et. al. in United States patent5,401,446. Polyethylene and polypropylene used in the manu-facture of CFM media and film materials were obtained fromcommercial sources within the United States. All other reagentgrade chemicals were obtained from Sigma-Aldrich, Inc.

Filtration Testing. Flat sheet samples of filter media wastested using either a TSI 8110 Automated Filtration Tester or a

46 INJ Winter 2003

TSI 3160 Fractional Filtration Efficiency Tester (TSI, Inc.Minneapolis, MN). The TSI 8110 was operated at a flow rateof 32 L/min using a sodium chloride challenge aerosol. TheTSI 3160 was also operated at a flow rate of 32L/min, howev-er, a dioctylphthalate aerosol was used to challenge media.The TSI 3160 uses a particle size clarifier to size select the chal-lenge particles based on electrophoretic mobility. Captureefficiency was determined by counting particles upstreamand downstream of the media being tested.

ASHRAE 52.2 filtration testing was performed at LMSTechnologies, Inc.(Bloomington, MN) in compliance withguidelines set forth ANSI/ASHRAE 52.2 Standard.

Dynamic Mechanical Analysis. Dynamic mechanical analy-sis of polymer films was performed on a Rheometrics, Inc.RSA-2 Solids Analyzer. Samples were cooled to with liquidnitrogen -50ºC, and then heated at 1ºC/min. All films sam-ples were characterized at constant strain amplitude and a fre-quency of 1 Hz.

RESULTS AND DISCUSSIONThe results presented below represent a combination of lab-

oratory and field studies of the filtration, physical, and elec-trical properties of continuous filament meltspun media.Comparisons are drawn between purely mechanical filtrationmedia and electrostatically charged or electret filtrationmedia.HVAC Field and Environmental Studies

HVAC Field StudiesThe in-service filtration performance

of heating, ventilation, and air condi-tioning (HVAC) media was measuredfor purely mechanical filtration media, anon-electret CFM media, and an electretCFM media. The purely mechanicalmedia was a 2.6 oz. yd-2 (osy) cotton-polyester HVAC media chosen to accu-rately represent the in-use performanceof this class of filter. The electret CFMmedia was a 2.0 osy bicomponent poly-olefin filter media charged by the coronapoling method. The non-electret CFMmedia was identical in structure to theelectret CFM media with the exceptionthat the media was not charged. All thefilter media were evaluated as pleat fil-ters wherein the pleating characteristicswere the same for each material. Samplefilters prepared with each media typewere subjected to operating environ-ments that utilized 100% outdoor air,100% indoor air, and a mixture of recir-culated (indoor) and makeup (outdoor)air.

The objective of this field study was tocompare the filtration performance of

purely mechanical filters to that of an electret CFM pleat filterunder the same conditions. In addition, the underlyingmechanical filtration property of the CFM media was evaluat-ed and compared to the purely mechanical filter media.Samples of pleated filters were exposed to a variety of envi-ronments where these media find continuous use for bothcommercial/industrial and residential applications. Thestudy protocol called for the removal of filters at regular timeintervals of up to 90 days in-service. Thus, a filter would beremoved from the HVAC system being used in the study aftera period of days (e.g. 30 days) and sent to an independent fil-tration-testing laboratory for ASHRAE 52.2 characterization.It is important to note that once a filter was removed and eval-uated for filtration efficiency, it was not returned to the fieldtest site. Rather, a new filter was placed in the field and theexposure clock restarted at zero time. As a result of this test-ing protocol, small filter-to-filter variations and the potentialfor disrupting the dust cake all contribute to variability in thefiltration performance data. However, by using new filters foreach time interval tested, used filters did not have to beshipped back and forth from the test sites or repeatedly han-dled during the study as this would have likely increased thevariability of the results to an unacceptable level.

100% Outdoor Air Environment. Pleated filters wereinstalled in the pre-filter bank of a two- stage filtration systemservicing a hospital located in the Chicago, Illinois area.Filters remained in the unit for a total time of two months.Filters were removed from the system a various time intervalsto characterize the filtration performance as media became

47 INJ Winter 2003

Figure 1PLEATED FILTER FIELD STUDY CONDUCTED AT A CHICAGO

AREA HOSPITAL UTILIZING 100% OUTDOOR AIR. FILTRATIONEFFICIENCY IS THE AVERAGE EFFICIENCY PARTICLE SIZE

RANGE OF 3-10 MICRONS FOR A 20”X20”X2” FILTER AT 492 FPM

loaded during actual use.The actual filtration performance

(Figure 1) shows that the filtration effi-ciency (average efficiency for 3-10micron particles, E3 value for ASHRAE52.2) of the electret CFM media decreas-es during the first 15 days of the test. Thedecay in efficiency can be attributed tothe capture of fine particle contaminantsfrom the outdoor air as it passes throughthe media. Notably, the decrease in effi-ciency during this time period was <20%relative to the efficiency of a virgin filter.In addition, as the dust cake accumulateswithin the electret CFM media, the filtra-tion efficiency recovers to a level only ca.10% below that of the new filter. Initially,the purely mechanical cotton-polyestermedia had a filtration efficiency that wasca. 40% below that of the electret CFMmedia. While the efficiency of themechanical media increased steadilyover the course of the test, it neverexceeded the filtration performance ofthe electret media.

The most striking finding of this studyis that the initial efficiency of the cotton-polyester media is poor as compared tothe electret media. This means that thenew cotton-polyester media removes fewer contaminant par-ticles when it is placed in service as compared to the removalefficiency of the electret media. Less particulate removaltranslates to more contaminant particles in the air down-stream of the filter. A second finding, not unexpected, is thatthe efficiency of the electret CFM media decreased during ini-tial exposure to outside air. This behavior is entirely consis-tent with the observation made in numerous studies (17,24).However, at no time during the field study did the filtrationefficiency of the electret media deteriorate to the low levelcharacteristic of the cotton-polyester media.

The outdoor air exposure study was extended to include asecond hospital in the Memphis, Tennessee area. Filters wereproduced using non-electret CFM, electret CFM, and the cot-ton-polyester media described above. The objective of thispart of the study was to characterize the mechanical filtrationproperties of the CFM media basic web structure independentof the electret effect. Again, comparative data were collectedfrom both the electret CFM media and the purely mechanicalcotton-polyester media.

Filtration efficiency data was obtained over a 90-day expo-sure period. As shown in Figure 2, the electret CFM mediadrops in efficiency over the first 10-15 days in-service, andthen regains efficiency as it loads with dust. The non-electretCFM media had an initial efficiency intermediate betweenthat of the electret and cotton-polyester media. Its efficiencyincreases over the course of the test as the structure steadilybuilds mechanical efficiency during dust loading. The same

behavior was observed for the cotton-polyester media; how-ever, its initial efficiency was approximately half that of theelectret media, and it was in service for approximately onemonth prior to achieving parity performance to the either ofthe CFM media types tested.

Notably, if one compares the results presented in Figures 1and 2, while the trends in performance are the same, theabsolute results are different. In the first field study the cot-ton-polyester media never achieved parity performance withthe electret CFM media. Whereas, in the second the filtrationproperties of the three media types tested became indistin-guishable after about one month. This likely reflects differ-ences in the outdoor air quality at the two test sites.Regardless, the initial filtration performance of the cotton-polyester media fell well below that of the electret CFM mediaand this condition persisted for upwards of one month afterthe filters were placed in service.

100% Indoor Air Environment. Pleated filters made usingelectret CFM media and mechanical cotton/polyester mediawere tested at residential locations in the Atlanta, GA. area.Filters were removed for ASHRAE 52.2 testing at regular timeintervals up to 90 days. The ASHRAE 52.2 dust removal effi-ciency of these filters is shown in Figure 3. The initial efficien-cy of the electret CFM media was greater than twice that of thecotton/polyester media. Initially, the electret CFM media dis-played a small increase in efficiency after 10 days in-use.Examination of the filter indicated a large percentage of thecontaminant present was in the form of small fiber fragments.

48 INJ Winter 2003

Figure 2PLEATED FILTER FIELD STUDY CONDUCTED AT A MEMPHIS

AREA HOSPITAL SITE UTILIZING 100% OUTDOOR AIR. FILTRATION EFFICIENCY IS THE AVERAGE EFFICIENCY

OVER THE PARTICLE SIZE RANGE OF 3-10 MICRONS FOR A16”X25”X2” FILTER AT 295 FPM FACE VELOCITY

It is believed that the presence of thesefragments altered the mechanical filtra-tion properties of the electret

CFM media. Independent of this ini-tial increase in efficiency, the electretCFM media displays a monotonicincrease in efficiency with time in-ser-vice. Notably, the cotton/polyestermedia also displays a steady increase inefficiency with time in-service, indica-tive of increasing mechanical efficiency.However, the cotton/polyester mediareaches a maximum efficiency of ca. 60%after one month, and does not achieveparity performance with the electretCFM media.

Mixture of Outdoor and Indoor Air.The final field study presented hereinvolves the exposure of pleated filtersto a mixture of outdoor (fresh) air andindoor (recycled) air. The study was per-formed in an office building within theAtlanta, GA. metropolitan area. Pleatedfilters made using electret CFM and cot-ton/polyester mechanical media wereinstalled in the building HVAC systemand tested according to the same proto-col as outlined above for a period of upto 80 days.

The dust removal efficiency of thesepleated filters after up to 80 days of expo-sure is shown in Figure 4. Notably, thecotton/polyester mechanical mediareaches a maximum efficiency after ca. 15days in-use. The efficiency remainsunchanged for the duration of the study.The electret CFM media initially has anefficiency more that twice that of the cot-ton/polyester media and maintainsroughly a 2X advantage over the dura-tion of the study. This data indicates thatexposure of electret CFM filtration mediato a mixture of outdoor and indoor airdoes not significantly degrade its filtra-tion efficiency over the lifetime of the fil-ter.

Effects of Humidity on FiltrationPropertiesElectret CFM media was tested in accor-dance with the ASHRAE 52.2 protocol asa function of increasing relative humidityin the challenge aerosol stream. Flatsheet media samples were tested for frac-tional filtration efficiency at normal con-ditions (70∞F (21∞C) and 50% rh) andmedia face velocity of 110 fpm. The same

49 INJ Winter 2003

Figure 3PLEATED FILTER FIELD STUDY CONDUCTED AT RESIDENTIAL

LOCATIONS IN THE METROPOLITAN ATLANTA AREA UTILIZ-ING 100% INDOOR AIR. FILTRATION EFFICIENCY IS THE AVER-

AGE EFFICIENCY OVER THE PARTICLE SIZE RANGE OF 3-10MICRONS FOR A 20”X200”X1” FILTER AT 295 FPM FACE VELOCITY

Figure 4PLEATED FILTER FIELD STUDY CONDUCTED AT OFFICE

BUILDING IN THE METROPOLITAN ATLANTA AREA UTILIZINGA MIXTURE OF OUTDOOR AND INDOOR AIR. FILTRATION

EFFICIENCY IS THE AVERAGE EFFICIENCY OVER THE PARTI-CLE SIZE RANGE OF 3-10 MICRONS FOR A 24”X24”X4” FILTER

AT 492 FPM FACE VELOCITY

media sample was then conditioned at80% rh , and then 90% rh and was retest-ed at these conditions. The fractionalfiltration efficiency at each humiditylevel is shown in Figure 5. Notably, thefiltration efficiency was unaffected byrelative humidity. In addition, while theincrease in relative humidity alters thedistribution of charged versus neutralparticles in the challenge aerosol (16),and water molecules adsorb to theaerosolized particulates, neither of thesephenomena had any effect on the abilityof the electret CFM media to captureand remove the contaminant from theair stream.

Effects of Long Term Storageon Filtration Properties

A number of concerns have beenraised over the long-term stability ofelectret CFM media. Concern hasfocused on a loss of efficiency due to adecay of the electret charge presentwithin the media structure. To addressthis issue two studies were conductedwhich focused on the effects of long-term storage on filtration properties.The first study involved storage of elec-tret CFM media in a non-temperaturecontrolled warehouse located in theSoutheastern United States. The secondstudy involved the accelerated aging ofelectret CFM media at 1300F for a periodof six weeks.

In the first study, newly manufacturedelectret CFM media were tested for frac-tional efficiency in accordance withASHRAE 52.2. The media were thenpackaged and stored in warehouseinventory for one and two years andthen retested. Figure 6 shows that thereis no significant difference in the effi-ciency of newly produced media andmedia aged for one year under ware-house storage conditions. A smalldecrease in efficiency was measuredform media aged for two years underthe same conditions. This decrease islimited to a narrow range of the smallestparticle sizes tested (0.3 to 0.9 µm). Atall other particles sizes tested the CFMmedia aged for two years in unchangedin performance relative to when it wasinitially manufactured and tested.

The second study involved the acceler-ated aging of electret CFM media at

50 INJ Winter 2003

Figure 5FRACTIONAL FILTRATION EFFICIENCY OF ELECTRET CFMMEDIA TESTED AS A FLAT SHEET, WHERE THE RELATIVE

HUMIDITY OF THE CHALLENGE AEROSOL WAS VARIED FROM50% TO 90% AT A TEMPERATURE OF 700F (210C)

Figure 6FRACTIONAL FILTRATION EFFICIENCY OF ELECTRET CFM

MEDIA AS MANUFACTURED AND FOLLOWING STORAGE IN ANON-TEMPERATURE CONTROLLED WAREHOUSE IN THESOUTHEASTERN UNITED STATES FOR UP TO TWO YEARS

1300F for up to six weeks. For this study, flat sheet media sam-ples were cut to 8” squares and were wrapped in aluminumfoil. The aluminum foil was sealed to form an air-tight pouchso that only the effect of temperature on filtration efficiencycould be studied. The media was tested using a TSI Model3160 Fractional Efficiency Tester at a flow rate of 32 L min-1

and a particle size of 0.03 mm using a dioctylphthalate chal-lenge aerosol. The data shown in Figure 7 indicate no signifi-cant change in filtration efficiency over the six-week timeperiod. Thus, even under very aggressive aging conditionsthe electret CFM media retains very high fine particle filtra-tion efficiency. As a figure of merit, the filtration efficiency ofnon-electret CFM media under the conditions of this test isless than 2%. This further supports thefact that the electret effect is very resis-tant to age related decay under condi-tions relevant to its storage and use.

Conditioning Treatments for Electrostatic Filtration Media

It has been proposed (25) that in orderto examine the mechanical filtrationproperties of electrostatically charged orenhanced filtration media, the media bedipped or washed in a 2-propanol bathand dried prior to testing. The 2-propanol wash is intended to remove theelectrostatic treatment yielding a non-

charged media for testing. In the studypresented here, we examined the effectof 2-propanol on the filtration proper-ties of a variety of electrostaticallycharged filtration media. In addition,we examined the mechanical propertiesof the respective polymers followingimmersion in 2-propanol liquid atambient temperature (ca. 75∞F).

The filtration properties of filter mediabefore and after 2-propanol exposurewere measured using a TSI Model 8110Filtration Tester with a sodium chloridechallenge aerosol. For liquid exposure,samples of media were immersed in 2-propanol for 10 min followed by air-drying for 12 h and oven drying at 800C(the boiling point of 2-propanol) for 0.5h. The data are shown in Table 1 below.

The data in Table 1 present a directcomparison between samples of mediatested prior to electret charging, afterelectret charging, and after electret

charging and immersion in 2-propanol.Notably, electret charging of each of themedia types leads to a marked improve-ment in filtration efficiency. In addition,while immersing the media in 2-propanol clearly leads to deterioration in

filtration efficiency; it does not restore the media to its originaluncharged state. In order to fully understand the effect of 2-propanol on the filtration media, films were prepared fromthe polymers used in each media. The films were immersedin liquid 2-propanol and the solvent uptake was examinedgravimetrically. In addition, dynamic mechanical analysiswas enlisted to characterize the physical properties of thepolymers before and after 2-propanol immersion.

Solvent Swelling: The intrusion of small molecules intopolymers is a well-known phenomenon (26-28). In general,small molecule penetrants diffuse into a polymeric materialand cause a phenomenon referred to as swelling. Swellingoccurs when the small molecule penetrant disrupts the van

51 INJ Winter 2003

Figure 7ACCELERATED AGING OF ELECTRET CFM MEDIA AT 1300F

(540C). FILTRATION EFFICIENCY AT 0.03 µm PARTICLE SIZE AND32L/MIN FLOWRATE

Table 12-PROPANOL EXPOSURE OF ELECTROSTATICALLY

CHARGED FILTRATION MEDIA

Filter Media Non-Electret Electret 2-Propanol Washed(% Efficiency) (%Efficiency) (%Efficiency)

2.0 osy PE/PPCFM Media 5.0 + 1.8 50.6 + 3.0 10.3 + 1.61.25 osy PPSpunbond Media 9.5 + 1.2 57.5 + 1.7 17.7 + 2.51.0 osy PPMeltblown Media 58.1 + 1.1 93.8 + 0.6 83.5 + 1.6

der Waals forces that normallydominate the interactions betweenadjacent polymer chains. In somesystems this leads to a change in thevolume of the sample, and hencethe term swelling. More important-ly, swelling disrupts and destabi-lizes the polymer microstructureand in the limit of a perfectlyswollen material the chains becomecompletely solvated and free tomove in a manner similar to theirmovement in a melt. One of the keyindicators of swelling is the glass-rubber transition temperature. Theso-called glass transition tempera-ture is the temperature where thethermal energy of adjacent chains ishigh enough that long rangeBrownian motion occurs. Thus,shifts in the glass transition temper-ature either up or down in value areindicative of changes in the poly-mer microstructure. Changes in theglass transition temperature can bemeasured very precisely using atechnique called dynamic mechani-cal analysis (DMA). Several excellent resources are availableon the subject of DMA (29-31), therefore only a brief overviewis provided here.

Dynamic mechanical analysis is a method of quantitativelymeasuring the mechanical behavior of a material over a rangeof temperatures. Polymers can be examined over a broadrange of temperatures. Polypropylene is commonly exam-ined from ca. -2000C up to its melting point at ca. 1650C. Theexperiment measures the relationship between the stress andstrain during the forced oscillatory deformation of the poly-mer specimen. DMA allows one to examine the relationshipbetween polymer microstructure and physical properties atsmall strain amplitude. The experiment yields the dynamicstorage and dynamic loss moduli, and the loss tangent or tan-gent of the phase angle over the range of temperatures mea-sured. The tangent of the phase angle is very importantbecause it is not affected by the sample or test geometry (29).The tangent of the phase angle (Tan d) is very sensitive tomechanical transitions that occur in the polymer as a functionof temperature. The phase angle for a perfectly elastic mater-ial is 0∞ (i.e. the stress and strain are in phase). For perfectlyviscous fluids, the stress and strain are exactly 900 out-of-phase. Polymeric materials typically lie between these twoextremes. Tan δ is also the ratio of the loss modulus to thestorage modulus. In the present study, DMA was used toexamine the polyethylene and polypropylene films followingimmersion of the films in 2-propanol for 60 minutes at ambi-ent temperature.

Gravimetric Swelling. Film (2 mm thickness) prepared fromthe same type of polypropylene and polyethylene used in the

manufacture of the electret CMA media was immersed in 2-propanol for periods of up to 60 min. Samples were removedat regular intervals, allowed to air dry, and then were weighedto the nearest 0.01 mg. Swelling is normally accompanied byan increase in mass of the polymer due to intrusion of the sol-vent molecules. Gravimetric swelling data are presented inFigure 8. The weight percent solvent was defined as the quo-tient of the mass of solvent to the mass of polymer. The massof solvent was calculated from the initial mass of polymer lessits mass after immersion in 2-propanol. The extent of 2-propanol uptake by either the polyethylene or polypropylenefilms was small. Typically, about 0.05 wt% was the limit interms of relative mass increase after immersion. Notably, theweight percent solvent varied up and down during the testand, in the case of polyethylene, the sample decreased in massfollowing immersion in 2-propanol.

The gravimetric swelling of polyethylene and polypropy-lene films clearly indicates that 2-propanol behaves as a smallmolecule penetrant for both polymers. This was surprisinggiven that fact that aliphatic alcohols are not considered goodsolvents for polyolefins (32). Perhaps the most noteworthyfinding was that several film samples suffered a net mass lossas a result of immersion in 2-propanol. This was due to theremoval of low molecular weight material from each of thefilms tested. Infrared spectroscopic analysis of residueobtained by extraction with 2-propanol indicated that thematerial removed was low molecular weight polymer.

Dynamic Mechanical Analysis. To further investigate theeffects of 2-propanol immersion, films of polypropylene wereexamined by DMA . Polyethylene films were also examined

52 INJ Winter 2003

Figure 8GRAVIMETRIC SWELLING OF POLYPROPYLENE AND

POLYETHYLENE FILMS DURING IMMERSION IN 2-PROPANOL

by DMA, however, because the glass transi-tion temperature (Tg) of polyethyleneoccurs at very low temperature, it is diffi-cult to measure accurately, and in somecases to observe. Polypropylene, on theother hand, gives rise to a peak in Tan δ dueto the glass transition over the temperaturerange of -200C to about 100C. The exact posi-tion of the peak is dependent on a numberof variables, however; if these are kept con-stant then the shift in Tg can be measuredaccurately.

Two types of polypropylene film samplewere analyzed by DMA following immer-sion in 2-propanol. The first was a com-pression molded film which was allowed tocool slowly to form a spherulitic crystalline morphology. Thesecond was an extrusion cast film, which was oriented andquenched to yield a row-nucleated crystalline morphology.These two samples were chosen in order to simulate two pos-sible extremes of polypropylene microstructure that mightoccur in filtration media. The Tg measured before and after 2-propanol immersion is given in Table 2.

In both cases, very significant increases in Tg result fromimmersion of the film samples in 2-propanol. The differencein the Tg between the two control samples (i.e. prior to immer-sion) is due to effects of orientation within the amorphousphase of the polypropylene. The increase in Tg observed fol-lowing immersion in 2-propanol is due to a shift in the mole-cular weight distribution caused by dissolution of low molec-ular weight polymer (33). The loss of low molecular weightpolymer observed in the gravimetric swelling experimentsand confirmed by both infrared analysis and DMA, clearlyshow that 2-propanol is not passive in its effect on the poly-mers used to manufacture the electret CFM media.

A final experiment was conducted in an attempt to removeall traces of 2-propanol from film samples following immer-sion in the solvent. Films were immersed in 2-propanol for 60min, dried and placed in a vacuum oven for several days.Films samples were then analyzed by headspace gas chro-matography. Headspace GC allows one to detect traceamounts of volatile materials present in a solid. In all cases, 2-propanol was still present in the polymers even after severaldays’ storage in vacuo.

The combination of gravimetric swelling, dynamic mechan-ical analysis, and headspace GC clearly show that immersionof polyolefin filtration media in 2-propanol is not a passive"conditioning treatment". The 2-propanol causes swellingand dissolution of low molecular weight polymer resulting ina material with different mechanical properties. In addition,analyses of the filtration properties of a variety of electretmedia clearly indicate that the 2-propanol immersion does notremove the electret effect. The changes observed in filtrationproperties observed following 2-propanol immersion are dueto penetration of the solvent into the polymer matrix. Thiscauses irreversible changes to the polymer, which results in adecay in filtration properties. It is possible that the 2-propanol

gives rise to conductive pathways in the bulk polymer lead-ing to charge recombination or compensation. However, thefact that the filtration properties of media that has beenimmersed in solvent are not the same as those of non-electretmedia invalidates the 2-propanol immersion as a conditioningstep.

ConclusionA brief review of electret charging techniques provided the

background for a discussion of different types of electrostaticfiltration media. The comprehensive field studies of pleatedfilters made using electret CFM media and purely mechanicalcotton/polyester media clearly demonstrates the overallsuperiority of the electret CFM media over the lifespan of thefilter. While small particle contaminants were observed tocause decay in the efficiency early in the lifetime of electretCFM media, the building of mechanical efficiency in these fil-ters compensates for this loss and yields a filtering mediumwith better initial efficiency and better long term filtration effi-ciency as compared to cotton/polyester media of equivalentbasis weight. In addition, it was shown that non-electret CFMmedia also provide better initial efficiency as compared to cot-ton/polyester media and maintain parity performance to thesame over lifetime of the filter. The combination of structureand electret charge, which exists in the electret CFM mediayields a more efficient filter overall.

In addition, evidence has been presented which refutes theuse of "conditioning treatments" which decouple electrostaticeffects from mechanical efficiency. The 2-propanol immersionwas clearly shown to cause irreversible changes to electrosta-tically charged media, which are related to dissolution of lowmolecular weight polymer by the solvent. Once exposed, the2-propanol is at best difficult and at worst impossible toremove from the polymers used in manufacturing electretCFM media. Thus, rather than being a passive treatment, the2-propanol immersion yields material with vastly differentmechanical and filtration properties, which do not reflectthose of non-electret charged media of the same structure.

AcknowledgementsThe authors gratefully acknowledge the contributions of

53 INJ Winter 2003

Table 2GLASS TRANSITION TEMPERATURE FOR

POLYPROPYLENE FILMSFOLLOWING IMMERSION IN 2-PROPANOL

Film Sample Glass Transition Temperature (0C)Compression Molded Control -1.20 + 0.03Compression Molded

Following 2-propanol ImmersionExtrusion Cast Control 4.8 + 0.01Extrusion Cast

Following 2-propanol Immersion 7.8 + 1.3

Whitney Durham, Burt Kobylivker, and Thuy Ahn Le forgravimetric swelling experiments, Hristo Hristov for DMAcharacterization of polymer films, and Richard Borders forhead space gas chromatography. Finally, we would like tothank Kimberly-Clark Corporation for the funding of thisresearch and the permission to publish the results.

References1. Sessler, G. M. In Electrets; Sessler, G. M., Ed.; Laplacian

Press: Morgan Hill, California, 1998; Vol. 1, Chapter 1.2. Sessler, G. M. In Electrets; Sessler, G. M., Ed.; Laplacian

Press: Morgan Hill, California, 1998; Vol. 1, Chapter 2.3. Duke, C. B. “Charge States in Polymers: Application to

Triboelectricity; Physiochemical Aspects of Polymer Surfaces,”International Symposium, 1983, p. 463.

4. Williams, M. W. J. Macromol. Sci.-Rev. Macromol. Chem.1976, C14 (2), pp 251-265.

5. Sviridenok, A. I.;Klimovich, A. F.; Kestelman, V. N.Proceedings of the 11thInternational Conference onComposite Materials, 1977; Vol. III, pp 744-753.

6. Hays, D. A.; Watson, P. K. In Second InternationalConference on Electrophotography; White, D. R., Ed.;Photographic Science and Engineering, Washington, DC,1973, pp 108-112.

7. Henniker, J. Nature (London) 1962, 196(3), 474.8. Hersh, S. P. and Montgomery, D. J. Textile Research Journal

1955, 25(4), 279.9. Smith, P. A.; East, G. C.; Brown, R. C.; Wake, D. J.

Electrostat. 1988, 21, 81-89.10. Agadjivand, S. A.; Jones, M. E.; Meyer, D. E. PCT

International Application WO 9505501, 1995.11. Eguchi, M. Proc. Phys. Math. Soc. 1921, 3, 2-4.12. Eguchi, M. Proc. Phys. Math. Soc. 1921, 3, 169-76.13. Loeb, Leonard B. Electrical Coronas: Their Basic

Physical Mechanisms; University of California Press, Berkleyand Los Angeles, 1965, Chapters 1 and 2.

14. Chang, J-S.; Lawless, P. A.; Yamamoto, T. IEEE Trans. OnPlasma Sci., 1991, 19(60), 1152-66.

15. Stearns, R. G. J. Appl. Phys. 1989, 66(7), 2899-2913.16. Hinds, W. C. “Aerosol Technology: Properties, Behavior,

and Measurement of Airborne Particles,” 2nd ed., John Wiley& Sons, Inc., New York, 1999, Chapter 15.

17. van Turnhout, J.; Adamse, J. W. C.; Hoeneveld, W. J. J.Electrostat. 1980, 8, 369-379.

18. Klasse, P. T. A.; van Turnhout, J. U.S. Patent 4,588,537,1986.

19. van Turnhout, J. U.S. Patent 3,998,916, 1976.20. Morris, W. T. Plastics & Polymers 1970, 38(133), 41-45.21. Salaneck, W. R.; Paton, A.; Clark, D. T. J. Appl. Phys. 1976,

47(1), 144-147.22. Williams, M. W.; Au Clair, C. J.; Ceasar, G. P.; Short, J. M.

Conf. Rec. – IAS Annu. Meet. (IEEE) 1975, 436-41.23. Sessler, G. M.; West, J. E. J. Acoust. Soc. Am. 1962, 34,

1787.24. Walsh, D. C.; Stenhouse, J. I. T. Aerosol Sci. and Tech. 1998,

29, 419-432.25. Lehtimäki, M. “Development of Test Methods for

Electret Filters,” Nordtest Project, VTT ManufacturingTechnology, Tampere, Finland, 1995.

26. Vrentas, J. S.; Vrentas, C. M. J. Polym. Sci.: Part B: Polym.Phys. 1992, 30, 1005-1011.

27. Rehage, G.; Ernst, O.; Fuhrmann, J. Discussions of theFaraday Society, 1970, No. 49, 208-221.

28. Gao, P.; Mackley, M. R. Proc. of the Royal Society, SeriesA 1994, 444(1921), 267-428.

29. Thermal Characterization of Polymeric Materials, SecondEdition; Turi, E. A., Ed.; Academic Press, New York, NewYork, 1996.

30. Datta, A; Zhou, J. J.; Yuan, J.; Monisera, A J. Appl. Med.Polym. 1998, 2(2), 54-59.

31. Seferis, J. C.; McCullough, R. L. Applied PolymerSymposium 1975, No. 2, 205-228.

32. Polymer Handbook, Brandrup, J.; Immergut, E. H.;Grulke, E. A. Eds.; John Wiley & Sons, New York, New York,1999, pp VII/497 – 545.

33. Billmeyer, F. W. Textbook of Polymer Science; John Wiley &Sons, New York, New York, 1984, Chapters 11 and 12. — INJ

54 INJ Winter 2003