i-ruu�_LiuAq

NUREG/CR-6029PNL-8594Vol. 1

_ Aging Assessment ofNuclear Air-Treatment SystemHEPA Filters and Adsorbers

Phase I*. 7 f

Prepared byW. K. Winegardner

Pacific Northwest LaboratoryOperated byBattelle Memorial Institute

Prepared forU.S. Nuclear Regulatory Commission

R A 44'at

AVAILABILITY NOTICE

Availability od Reference Materkals Cited In NRC Pubicaftons

Most documents cited h NRC publications win be available from one of the following sources:

1 The NRC Pubtio Document Room. 2120 L Street, NW. Lower Level, Washington. DC 20555-0001

2 The Superintendent of Documents, U.S. Government Printing Office. Malt Stop SSOP, Washington,DC 20402-9328

3. The National Technical Information Service. Springfield, VA 22151

Although the isting that follows represents the majority of documents cited In NRC pubicatlons, It Is notIntended to be exhaustive.

Referenced documents available for Inspection and copying for a fee from the NRC Publo Document RoomInclude NRC correspondence and internal NRC memoranda; NRC Office of Inspection and Enforcementbulletins. circulars, Information notices, Inspection and Investigation notices; Licensee Event Reports; ven-dor reports and correspondence: Commission papers; and applicant and licensee documents and corre-spondence.

The following documents In the NUREG series are available for purchase from the GPO Sales Program:formal NRC staff and contractor reports. NRC-sponsored conference proceedings, and NRC booklets andbrochures. Also available are Regulatory Guides, NRC regulations In the Code of Federal Regulations, andNuclear Regulatory Commission Issuances.

Documents available from the National Technical information Service Include NUREG series reports andtechnical reports prepared by other federal agencies and reports prepared by the Atomic Energy Commis-slon, forerunner agency to the Nuclear Regulatory Commission.

Documents available from public and special technical libraries Include al open literature Items, such asbooks, journal and periodical articles, and transactions. Federal Register notices, federal and state legisla-tlon, and congressional reports can usually be obtained from these libraries.

Documents such as theses, dissertations, forelgn reports and translations, and non-NRC conference pro-ceedings are available for purchase from the organization sponsoring the publication cited.

Single copies of NRC draft reports are available free, to the extent of supply, upon written request to theOffice of Information Resources Management. Distribution Section, U.S. Nuclear Regulatory Comrisslon.Washington, DC 20555-0001.

Copies of industry codes and standards used In a substantive manner In the NRC regulatory process aremaintained at the NRC Library, 7920 Norfolk Avenue. Bethesda, Maryland, and are available there for refer-ence use by the public. Codes and standards are usually copyrighted and may be purchased from theoriginating organization or. If they are American National Standards, from the American National StandardsInstitute, 1430 Broadway, Now York, NY 10018.

DISCLAIMER NOTICE

This report was prepared as an account of work sponsored by an agency of the United States Government.Neitherthe United States Government norany agencythereof, orany of their employees, makes any warranty,expresed or implied, or assumes any legal liability of responsibility for any third party's use, or the results ofsuch use, of any Information, apparatus, product or process disclosed in this report, or represents that its useby such third party would not Infringe privately owned rights.

NUREG/CR-6029PNL-8594Vol. 1RX

Aging Assessment ofNuclear Air-Treatment SystemHEPA Filters and Adsorbers

Phase I

Manuscript Completed: March 1993Date Published: August 1993

Prepared byW. K. Winegardner

Pacific Northwest LaboratoryRichland, WA 99352

Prepared forDivision of EngineeringOffice of Nuclear Regulatory ResearchU.S. Nuclear Regulatory CommissionWashington, DC 20555-0001NRC FIN B2911

Abstract

A Phase I aging assessment of high-efficiency particulate air (HEPA) filters and activated carbon gas adsorption units(adsorbers) was performed by the Pacific Northwest Laboratory (PNL) as part of the US. Nuclear Regulatory Com-mission's (NRC) Nuclear Plant Aging Research (NPAR) Program. Information concerning design features; failureexperience; aging mechanisms, effects, and stressors; and surveillance and monitoring methods for these key air-treatment system components was compiled. Over 1100 failures, or 12 percent of the filter installations, were reportedas part of a Department of Energy (DOE) survey. Investigators from other national laboratories have suggested thataging effects could have contributed to over 80 percent of these failures. Jensile strength tests on aged filter mediaspecimens Indicated a decrease in strength. Filter aging mechanisms range from those associated with particle loadingto reactions that alter properties of sealants and gaskets. Low radioiodine decontamination factors associated with theThree Mile Island (TMI) accident were attributed to the premature aging of the carbon in the adsorbers. Mechanismsthat can lead to impaired adsorber performance include oxidation as well as the loss of potentially available active sitesas a result of the adsorption of pollutants. Stressors include heat, moisture, radiation, and airborne particles andcontaminants.

iii NUREG/CR-6029

Contents

Abstract ............................................................... iii

Summary ............................................................... vii

Acknowledgments ............................................................... ix

Nomenclature ..................................................................................... xi

1 Introduction ................................................................................... 1.1

2 Background .................................................................................. 2.1

3 Component Design and Construction .............................................................. 3.1

3.1 HEPA Filters . .............................................................................. 3.1

3.2 Adsorbers .................................................................................. 33

3.3 Supporting Documents . ...................................................................... 3.4

4 Aging Mechanisms, Effects, and Stressors .......................................................... 4.1

4.1 HEPA Filters ............................................................................... 41

4.2 Adsorbers .................................................................................. 4.2

5 Failure Experience . .............................................................................. 5.1

5.1 Licensee Event Reports ....................................... 5.1

5.2 DOE Site Survey ....................................... 5.1

5.3 TMI Accident ..... ............. ...................... 5.2

6 Inspection, Surveillance, and Monitoring Methods .......................... ............. 6.1

6.1 HEPA Filters ........................................ 6.1

6.2 Adsorbers ........................................ 6.2

7 Conclusions ................................ 7.1

7.1 Normal Operating and Error-Induced Conditions ................... .................... 7.1

v NUREG/CR-6029

Contents

7.2 Reactor Accident Conditions ............................ 7.1

73 Conclusions and Recommendations ............................ 7.4

8 References ................................ &

NUREG/CR-6029 vi

Summary

Nuclear air-treatment systems are one of a variety of systems that provide a safe environment for plant personnel andequipment during normal operations as well as retain radionuclides during accidents. High-efficiency particulate air(HEPA) filters and activated carbon gas adsorption units (adsorbers), designed to capture radioactive particulate andgaseous contaminants, are the key components of these systems. This report presents the results of a Phase I agingassessment for these two key components. Information concerning design features and construction; failureexperience; aging mechanisms, effects, and related stressors (the agents or stimuli that can result in degradation); andinspection, surveillance, and monitoring methods (ISMM) are included in the report.

Information is summarized from the standards, codes, and specifications that combine to provide the basic physical,chemical, test, and performance criteria for the glass fiber filter media (paper) and impregnated activated carbon aswell as for assembled components. Failure experience, based on analyses of over 60,000 licensee event reports (LERs)and a survey of Department of Energy (DOE) sites, is summarized and, where possible, related to aging (Moeller 1979,Moeller and Sun 1983, Jacox 1989, Sommer and Otermat 1992, and Carbaugh 1983). Although less than one percentof the LERs appeared to be related to filters and adsorbers, several instances of the premature aging of carbon, fromerror-induced conditions, were reported. Over 1100 failures, or a total of 12 percent of the filter applications, werereported as part of the DOE survey (Carbaugh 1983). Investigators from other national laboratories have suggestedthat the effects of aging could have contributed to over 80 percent of these filter failures (Johnson et al. 1989). Thesesame investigators reported results of an experimental evaluation of the tensile breaking strength of aged filter mediaspecimens that revealed that 42 percent of the samples did not meet the specifications for test specimens obtainedfrom new material. A significant loss in water repellency from values specified for new media was also measured.

Low radioiodine decontamination factors associated with the Three Mile Island (IMI) accident were attributed to thepremature aging of the carbon in the adsorbers. Inspection, surveillance, and monitoring methods have beenestablished to observe filter pressure drop buildup, check both HEPA filters and adsorbers for bypass, and determinethe retention effectiveness of aged carbon. Exemptions to TMI technical specifications postponed the latter,surveillance tests that could have revealed the extent of carbon aging.

Aging mechanisms associated with filters range from particle loading of the media to corrosion of metal members, andphysicochemical reactions that alter properties of sealants, gaskets, and water repellents. Filter media may embrittlefrom prolonged exposure to air. In the case of adsorbers, it has long been established that aging mechanisms lead toimpaired performance in terms of the capture of volatile iodine species. The deterioration in performance is caused byoxidation as well as the competitive loading of other airborne constituents. Many airborne constituents, includingmoisture, can readily react or be adsorbed by carbon beds reducing the number of active "sites" that otherwise would beavailable for radioiodine adsorption. Stressors include heat, humidity, radiation, and airborne contaminants andpollution.

The Phase I study found that the HEPA filters and adsorbers are considered to have a long service life, especially thefilters. Thus, if a severe accident happens it is likely to occur at a time when these two final confinement barriers havebeen in use for an extended period, even years. Even with existing ISMM, the aged, and possibly degraded compo-nents, could fail to provide the controlled environment needed by personnel to ensure safe shutdown following theaccident, or be the weak link that allows the release of radionuclides to the environment. Further, the Phase Iassessment has revealed the need for an improved definition of accident conditions and the possible need foradditional information to comprehensively evaluate the performance of aged components under such conditions. Theexpansion of the Phase I aging assessment to other components of nuclear air treatment systems, namely demisters,

vii NUREG/CR-6029

Summary

heaters, coarse filters, fans or blowers, and dampers, is recommended. Several of these components, although notdesigned to retain radionuclides, could mitigate the impact of conditions that threaten to cause failure of aged HEPAfilters and adsorbers during accidents.

NUREG/CR-6029 viviii

Acknowledgments

The author would like to thank Dr. John (Jack) J. Burns Jr., Office of Nuclear Regulatory Research, Division ofEngineering, for his programmatic guidance and careful review of this effort.

The author would especially like to acknowledge the efforts of PNL staff member Elizabeth (Liza) J. Eschbach in thecollection and cataloging of many of the references that formed the basis of the assessment. The author would alsolike to thank Liza (PNL) and Melvin W. First, Sc.D., Harvard Air Cleaning Laboratory, for their review of thisdocument and valuable comments. Finally, the author would like to acknowledge the following PNL staff for theiroutstanding efforts in assisting with the preparation of this report: Susan (Sue) J. Arey, Norma J. Reed, and Sharon KLoverne.

ix NUREG/CR-6029

Nomenclature

ACS air cleaning system

ANSI American National Standards Institute

ASME American Society of Mechanical Engineers

ASTM American Society for lbsting and Materials

CFR Code of Federal Regulations

Cs! cesium iodide

DBA design-basis accident

DF decontamination factor

DOD Department of Defense

DOE Department of Energy

DOP dioctyl phthalate

EPRI Electric Power Research Institute

ESF engineered safety feature

GDC general design criteria

HEPA high-efficiency particulate air

ISMM inspection, surveillance, and monitoring methods

KI potassium iodide

LER licensee event report

LOCA loss-of-coolant accident

MEK methyl ethyl ketone

MIL military standard

NEA Nuclear Energy Agency

NPAR Nuclear Plant Aging Research

xi xi ~~~~~~NUREG/CR-6029

Nomenclature

NRC Nuclear Regulatory Commission, U.S.

OECD Organization for Economic Cooperation and Development

PNL Pacific Northwest Laboratory

RG regulatory guide

RH relative humidity

SGTS standby gas treatment system

SSC systems, structures, and components

STCP Source RTrm Code Package

TEDA triethylenediamine

TMI Three Mile Island

USAEC U.S. Atomic Energy Commission

NUREG/CR-6029 xdii

1 Introduction

This report presents the Phase I aging assessment of thehigh-efficiency particulate air (HEPA) filters and gasadsorption units (adsorbers) of nuclear air-treatment orcleaning and ventilation systems. The study was per-formed by Pacific Northwest Laboratory (PNL)(a) forthe U.S. Nuclear Regulatory Commission (NRC).Information concerning the design and construction ofthe two air-treatment system components is providedincluding summaries of related sections of the NRCregulatory guides (RG), American Society of Mechani-cal Engineers (ASME) standards and codes, and militaryand American Society for Testing and Materials(ASTM) specifications. These documents combine toprovide basic physical, chemical, test, and performancestandards for filter media and impregnated activatedcarbon as well as for assembled components.

Aging mechanisms and effects are discussed in conjunc-tion with stressors, the agents, or stimuli that can resultin degradation. Failure experience, based on analyses oflicensee event reports (LERs) and a survey of UnitedStates Department of Energy (DOE) sites, is summar-ized and, where possible, related to aging. The discus-sion of failure experience also includes exhaust-airtreatment experience associated with the Three MileIsland (TMI) accident. Although TMI treatment sys-tems provided a significant final barrier to radioactivityrelease, lower than expected radioiodine removal wasattributed to "...poor performance from aged charcoalfilters' (Wilhelm and Deuber 1991).

Inspection, surveillance, and monitoring methods(ISMM) are reviewed. These methods are used to estab-lish the condition of the filters and adsorbers once thecomponents have been placed in operation. Surveil-lance tests, which are a series of tests periodicallyperformed to monitor component condition and

demonstrate the current ability to remove fine particlesand iodine and Iodine compounds, are described alongwith summaries of related sections of NRC regulatoryguides and ASME standards.

Definitions related to aging that are used in the reportwere obtained or derived from the results of a study toprovide uniform common terminology for addressingaging of systems, structures, and components (SSC) innuclear power plants (EPRI 1992). Aging from normalconditions assumes properly-designed, fabricated,installed, operated, and maintained SSC and is definedas the "general process in which characteristics of [the]SSC gradually change with time or use." It should benoted, however, because of the variety of normal condi-tions and stressors encountered, when consideringHEPA filters and adsorbers, the use of gradually todescribe change may not always appropriately reflect therate at which aging can occur. Normal aging degrada-tion is often the result of exposure to airborne contami-nants and concentrations and properties fluctuate. Fur-ther, contaminants and pollutants vary from site-to-siteand can result from maintenance and testing and possi-bly even standby or shutdown as well as normal opera-tion. Filters and adsorbers are also used in several sys-tems and large air flows can be involved. Therefore,general rules concerning design service life associatedwith even normal conditions are essentially impossibleto develop. A service life of several years is not uncom-mon, but significantly shorter periods may be involved,despite the fact that protective components, namely pre-filters, demisters, and air heaters, are used to extendusable life. Error-induced conditions leading to earlyaging are also discussed in the report. Although wear-out, including reduced retention efficiency, from error-induced aging degradation can be viewed as prematureaging, the root cause of failure in this case is humanerror, not aging.

(")Operated for the U.S. Department of Energy by Battelle MemorialInstitute under Contract DE-AC06-76RL0 1830.

1.1 NUREG/CR-6029

2 Background

This Phase I aging assessment of HEPA filters and gasadsorption units (adsorbers) is part of an evaluation ofthe effects of aging on engineered safety feature (ESF)systems, one of the groups of systems of current interestin the Nuclear Plant Aging Research (NPAR) Program(USNRC 1991). The identification of potential safety-related aging issues, coupled with the need to avoidduplication, were the key considerations in identifyingcandidate ESF systems for initial study. Air-treatmentor cleaning and ventilation systems were ultimatelyselected because failure of these systems can impactboth plant and public safety. System components can bethe last barrier in preventing the release of radioactivityto the public following an accident including that associ-ated with the airborne iodine and cesium radionuclidesthat can provide a substantial contribution to total dose.Further, air-treatment systems are needed to ensure safeshutdown of the plant or to allow equipment to beserviced. Satisfactory performance of certain air-treatment systems is essential to ensure control roomhabitability. Air-treatment systems and components arealso used to provide a safe and/or controlled environ-ment for personnel and equipment during normaloperations. Operability of some safety-related equip-ment is dependent upon particular systems to removeheat from the rooms where these components areinstalled.

As shown in Figure 2.1, air-treatment systems consist ofsome or all of the following components: moisture sep-arators or demisters, electrical heaters, prefilters, HEPAfilters, adsorbers, postfilters, fans or blowers, ductwork,and dampers and valves. Not shown are the instrumen-tation and equipment to sample and monitor systemperformance. The first step in the Phase I study was toidentify a boundary to isolate the components thatwould be assessed. The fundamental components interms of providing plant and public safety are thosedesigned to capture radioactive gaseous and particulatecontaminants, namely the adsorbers and HEPA filters,respectively. One or both are installed in nuclear air-treatment systems. The particles could be radioactivechemical compounds or otherwise inert airborne mate-rial contaminated by radioactive species. Gasses ofprimary interest include the elemental and organicforms of radioiodine. Activated carbon is used toremove the gaseous or volatile forms of iodine and isusually impregnated with other chemicals to enhancethe removal of the organic species. HEPA filters andactivated carbon adsorption units were selected as thefocus for initial Phase I assessment not only becausethey are essential for safety, but in the case of theadsorbers, it has long been established, and was evi-denced again during the TMI accident, that there areaging mechanisms that can lead to impaired perform-ance (Burchsted et al. 1976, Wilhelm and Deuber 1991).

2.1 NUREG/CR-6029

Background

-- E-N~~~~~~~~~~IM D X

A P H pI I I Na - I X

NAMNAMNM

EeuS rHeart

Pemier

Poor

dn�nH

K�HEPA Flter

Adsofb

Fan-drive Assy

[i Dmper

3SS92D9S

FIgure 2.1. Common Air-TReatment System Configuration

NUREGICR-6029 2.2

3 Component Design and Construction

Particle filtration and gas (vapor) adsorption units (ad-sorbers) are included in nuclear air-treatment systemsand are designed to remove radioactive materials. Thematerials, airborne particles or volatile species, may besuspended in or exist as gas phase constituents of recirc-ulating aerosols, gaseous effluents, or accidental re-leases. The adsorbers, activated carbon beds, can effec-tively remove elemental radioiodine and are oftenimpregnated with other chemicals to enhance the reten-tion of organic species. The fine particle filters, HEPAfilters, may also be significantly involved in radioiodineremoval because iodine can be in the form of solidcesium iodide (Csl). Design and construction of the twocomponents are discussed below.

3.1 HEPA Filters

HEPA filters clean or treat aerosols by separating sus-pended particles from an essentially atmospheric pres-sure gas stream. The filter media is made from a mix-ture of glass fibers and is in the form of filter 'paper."Particles are collected by interception, impaction, andBrownian diffusion. The continuous gas phase passesthrough virtually unchanged. HEPA filters are definedas a "high-efficiency particulate air filter having a fibrousmedium with a particle removal efficiency of at least99.97% for 0.3 gm particles of dioctyl plithalate [DOP](ASME 1989a).

An excellent review of the construction and service char-acteristics of HEPA filters has been prepared by First(1991). This reference also includes a brief history ofthe development of filters for the nuclear industry andincludes additional details concerning filtration mechan-isms. Much of the information presented in this sectionwas either abstracted or taken directly from First or ref-erenced military specifications. As indicated by First,'Filters constructed with paper pleated the full depth ofthe rigid outer frame and with adjacent pleats held apartby full-depth corrugated separators are the most widelyused." In other words, as shown in Figure 3.1, the glassfiber filter paper is in the form of a continuous sheetpleated vertically over the separators.

Figure 3.1. HEPA Filter

Physical and chemical filter medium/paper standards arecontained in Military Specification MIL-F-51079D, "Fil-ter Medium, Fire-Resistant, High-Efficiency," and in-clude those for air flow resistance, particle penetration,and tensile strength (U.S. Department of Defense[DOD] 1985). The latter, tensile strength criteria,includes requirements after exposure of the filter paperto heated air, water immersion, and after exposure togamma irradiation. Standards for water repellency priorto and after exposure to gamma irradiation and mildewresistance are also included. The penetration criteria(not to exceed 0.03% for 0.3 /m DOP particles) is dir-ectly related to the minimum efficiency used in theabove definition. The penetration criteria is associatedwith tests specifying a flow rate of 0.032 m3/min and anexposed test area of 100 cm2. Maximum pressure droprelated to the preceding test specifications (minimumvelocity of 320 cm/min) is 0.39 kPa (40 mm of water).The minimum average tensile breaking strength speci-fied for media specimens (N/n or lb/in. of width in thecross direction), initially, after exposure to heated air at371°( ± 28°C (700TF ± 50TF) for 5 min, after being

3.1 NUREG/CR-6029

Component Design

soaked for 15 min, and after radiation exposure are 350or 2.0 (438 or 2.5 in the machine direction), 105 or 0.6,175 or 1.0 (wet) and 175 or 1.0, respectively. Themilitary specification also includes the requirement thatthe combustible material in the filter medium shall notexceed 7 wt%. First (1991) notes that the latter 'permitsincorporation of . .. organic matter, divided between(1) a latex addition to give the paper strength andresistance to cracking at the bends [where the sheet ispleated] and (2) a water repellent to protect the paperagainst wetting from deposition of liquid droplets,should they be present.' Specifications for the filtermedium, derived from MIL-F-51079, are also presentedin the ASME Code on Nuclear Air and Gas Iteatment,ASME/American National Standards Institute (ANSI)AG-i- 1988 Section FC, 'HEPA Filters' (ASME/ANSI1988).

In addition to the medium, HEPA filter components in-dude the frames, gaskets, face guards, and the corru-gated separators that are used to separate adjacentsheets of filter paper. Specifications for these compo-nents are presented in MIL-F-51068F, 'Filters, Particu-late (High-Efficiency Fire Resistant),* along withcriteria for the adhesives, sealants, and paints used toassemble and finish the units (U.S. Department ofDefense 1986). A rigid outer frame or housing is usedto fully enclose the paper-separator assembly on foursides, and provide the connections to adjacent ductwork.Specifications for frame material include those for ply-wood, wood particle board, laminated veneer lumber,aluminum alloys, cold-rolled steel sheet, chrome steelsheet, and stainless steel A flame spread classificationis given for plywood. First (1991) notes that 'face gas-kets are usually constructed from flat strips of expandedclosed cell neoprene sponge rubber... cork and otherrubber formulations are also permitted [a proprietaryfluid seal design is also discussed].' It is specified thatthe corrugated separators be made from aluminumalloys or epoxy-coated aluminum. The specificationrequires that "adhesives used to splice the medium,fasten the gasket to filter face, and sealants or adhesivesused to seal the filter pack to the frame shall be self-extinguishing after direct contact with an openflame ...

Performance specifications for assembled standard-sizedfilters (nominal rated capacity ranging from 0.71 to57 m3/min or 25 to 2000 ft3/min) are included InMIL-F-51068F for DOP particle penetration (hogreater than 0.03% at specified flow rates) and resis-tance to air flow, pressure, heated air, and environmen-tal exposure. Maximum pressure drop specified at ratedair flow ranges from 0.25 to 032 kPa (1.0 to 1.3 in. ofwater), depending on size. The pressure test for 61 by61 cm units involves assuring that the assembled unit,after conditioning for not less than 24 hours in a cham-ber at 350C (-95*F) and at a relative humidity (RH) of95%, will then withstand a specified air flow (needed toproduce a 2.49 + 0.05 kPa or 10± 0.2 in. water pressuredifferential across filter) and water spray environmentwithout rupture of the filter medium. Evaluation ofenvironmental performance involves cyclic exposure fora total of 9 weeks to specified conditions termed arctic(-540C or -650F), desert (45*C or 1600F and 10% RH)and tropical (450C or 1130F and 88% RH). Particlepenetration criteria also include minimum values for thewet filter following the pressure test, after exposure tospecified flow rates of heated air (371 ± 28TC or 700±500F for at least 5 min), and after the environmentalexposure cycles.

Materials, design, and fabrication standards are also in-cluded in Section FC of the ASME Code on Nuclear Airand Gas 'Teatment (ASME/ANSI 1988). Three types ofHEPA filters are addressed: folded filter media withcorrugated separator/supports (described above); min-ipleat medium with glass fiber or noncombustible threadseparators (made from a series of flat panels of pleatedfilter media assembled in a V-shape, with adjacentpleats separated and supported by ribbons of glass fiberor noncombustible threads glued to the media, maxi-mum pleat height of 20 mm); and continuous corrugatedfilter media folded without separators. Steel and stain-less steel sheet frames are emphasized. A sequence ofqualification tests is described that includes evaluationof resistance to air flow, particle penetration, resistanceto rough handling, resistance to pressure, resistance toheated air, and spot flame and radiation resistance andinvolves a specified number of filters. No attempt wasmade to define all differences between the above

UREG/CR-6029 3.2

Component Design

military specifications and Section FC, but It was notedthat combustible material in the filter media is limitedto 5 wt% by the latter.

3.2 Adsorbers

As indicated above, adsorbers are used to remove gas-eous radioiodine species from effluents. Various stan-dards and specifications permit the use of any adsorbentmedium that has been demonstrated to be equal to orbetter than activated carbon in terms of radioiodine re-tention. However, only activated carbon is discussedbecause it Is essentially the only material used and sup-porting testing procedures and acceptance criteria arespecific to its use.

Gasses are directed through tightly-packed beds of acti-vated carbon granules. The carbon or charcoal is prep-ared by controlled heating in a steam environment to re-move volatile organic material. The heating generates amaterial with a large surface-to-mass ratio and internalsurfaces or sites where adsorption of iodine moleculescan take place. Ile active media is usually . ... activatedcharcoal that is derived from either coal or coconutshells. Although elemental radioiodine is retained effi-ciently by activated carbon alone..., the charcoal isoften impregnated with additional chemicals to improveits retention of organic species (e g., methyl iodide).Typical impregnants include potassium iodide (KI), tri-ethylenediamine (IEDA), or derivatives of TEDA.Mixtures of these compounds are also used. ... With KIas the impregnant, iodine retention is mainly due toisotopic exchange [nonradioactive iodine in the impreg-nant substitutes for the radioactive iodine of the organicspecies and then passes through the bed]. .. WithTEDA or its derivatives as the impregnant, the primaryretention reaction is due to the formation of quaternarysalts. Even in this case, however, isotopic exchange is animportant factor ... .t (Wilhelm and Deuber 1991).

Standards for the adsorbent media are included in the"Code on Nuclear Air and Gas ikeatment," ASME/ANSI AG-1-1988, Section FF, Adsorbent Media(ASME/ANSI 1988). In addition to specifications forphysical properties, Section FP lists requirements fortesting the adsorbent to verify retention performance. Afree-flowing granular substance is specified. Activation

and impregnation processes are left to the discretion ofthe manufacturer. Physical tests specified after impreg-nation include those for apparent density, particle sizedistribution, ignition temperature, and hardness. Qual-ification tests to establish the suitability of a grade ortype of impregnated activated carbon require verifica-tion that elemental iodine removal efficiency is not lessthan 99.9% at 300C and 95% RH. Corresponding effi-ciency values for methyl iodide are 99.0% at WC and95% RH and 98% at 1300C and 95% RH. Further,batch tests at the time of manufacture, to assure that thematerial being used has the same characteristics as thematerial qualified in testing, require demonstration ofminimum elemental iodine and methyl iodine retentionefficiencies of 99.5% (at 180IC) and 97% (at 30C and95% RH), respectively. Definitions and terms includedin Section FP include those for 'batch," 'grade or type,"and "qualification test."

ASTM standards are referenced in Section FF ofAG-1-1988 that describe methods for testing the adsorb-ent with radioactive material to verify retention char-acteristics as well as to evaluate physical properties suchas density and particle size distribution. The former,ASTM D 3803-89, 'Standard lbst Method for Nuclear-Grade Activated Carbon,' is ... a very stringent proce-dure for establishing the capability of... activated car-bon to remove radio-labeled methyl iodide from air andgas streams' (ASTM 1989). As will be discussed later inthe report, the test method is used to quantify the extentof degradation of aged carbon as well as to qualify newcarbons. Physical properties and performance specifica-tions for new, virgin impregnated activated carbon forremoving gaseous radioiodine are specified in ASTM D4069-90, 'Standard Specification for Impregnated Acti-vated Carbon Used to Remove Gaseous Radio-lodinesfrom Gas Streams" (ASTM 1990). Test methods forevaluating the various physical properties are includedas referenced documents.

Specifications for materials, design, fabrication, and theinspection and testing of two types of adsorbers are pre-sented in Sections FD, "Iype II Adsorber Cells," and FE,"Type III Adsorbers,' of ASME/ANSI AG-1-1988(ASMEIANSI 1988). 'lype 304 or 304L series stainlesssteels are specified for materials that contact the ad-sorbent. General design of the Type II adsorber unit isbased on a single cell consisting of 50.8-mm (2-in.) thick

3.3 NUREG/CR-6029

Component Design

minimum beds in a modular tray-type arrangement. Aminimum residence time of 0.25 sec. and maximumpressure drop of 0.31 kPa (1.25 in. water) at the ratedcapacity of 9.43 m3/min (333 ft3/mnin) is specified. Type11 design is based on single or multiple permanent unitswhere carbon is filled and emptied in place (sometimescalled deep bed or gasketless, see ASME/ANSIAG-1-1988 Section FE for additional details).

3.3 Supporting Documents

In summary, basic qualifications for the filter mediumand assembled HEPA particulate filters are contained intwo military specifications (U.S. Department of Defense1985, 1986). Primary standards for adsorbent media, as-sembled adsorbers, and related testing methods are pre-sented in Sections FD, FE, and FF of ASME/ANSI codeAG-1 and two ASTM specifications (ASMEJANSI 1988;ASTM 1989, 1990). Numerous documents supplementand/or support the preceding five references, includingNRC and other ASME and ASTM publications. One ofthe most closely related NRC publications, RG 1.52,'Design, Testing, and Maintenance Criteria for PostAccident Engineered-Safety-Feature AtmosphereCleanup System Air Filtration and Adsorption Units ofLight-'Water-Cooled Nuclear Power Plants," is currentlybeing revised (USNRC 1978; Bellamy 1991). This guidepresents methods for implementing the appropriategeneral design criteria (GDC) described in Appendix Aof 10 CFR 50 (U.S. Code of Federal Regulations 1992)and as suggested by the title of the guide, *... appliesonly to post accident engineered-safety-feature atmos-phere cleanup systems designed to mitigate the conse-quences of postulated accidents ... land) does not applyto atmosphere cleanup systems designed to collect air-borne radioactive materials during normal plant opera-tion, including anticipated operational occurrences.'

Regulatory Guide 1.140, "Design, Testing, and Mainten-ance Criteria for Normal Ventilation Exhaust SystemAir Filtration and Adsorption Units of Ught-Water-Cooled Nuclear Power Plants' provides guidance for thelatter, normal systems (USNRC 1979).

Regulatory Guide 1.52 references two ANSI/ASME stan-dards for the design and testing of engineered safetyfeature (ESE) system components including HEPA fil-ters and adsorbers. The latest revisions of these twostandards have been published by the ASME as Ameri-can National Standards ASME N509-1989, 'NuclearPower Plant Air-Cleaning Units and Components," andASME N510-1989, "Ibsting of Nuclear Air 'TeatmentSystems' (ASME 1989 ab). These two ASME stan-dards, in addition to being connected to the regulatoryguides, are also at least the initial link to the basic fivereferences that contain various construction, material,test, and qualification specifications for new HEPAfilters and adsorbers. Specifically, requirements ofASME N509-1989 include the use of the two militaryspecifications and the three sections of the nuclear airand gas treatment code (the above two ASTM standardsare referenced in Section FF of code, ASME/ANSIAG-1). There is no intent to imply that all of theinformation presented or referenced in the latest revi-sions of the two ASME standards has been evaluated bythe NRC in terms of applicability or acceptability.Further, the use of the term requirement in conjunctionwith the various standards, codes, and specificationsdoes not necessarily mean that a particular requirementhas been accepted by the NRC. Finally, it should benoted that some of the information included in the twoASME standards and NRC regulatory guides will be dis-cussed in greater detail later in the report in conjunctionwith surveillance tests that are used to monitor the con-dition of installed filters and adsorbers.

NUREG/CR-6029 3.4

4 Aging Mechanisms, Effects, and Stressors

This section combines information concerning agingmechanisms and effects and the various stressors thatare involved. As will be detailed below, stressors, theagents or stimuli that can contribute to aging (includingthe direct degradation in performance in terms of radio-nuclide retention) include heat, humidity, steam, radia-tion, and airborne contaminants and pollutants.

4.1 HEPA Filters

The obvious effect of service associated with HEPA fil-ter use is the increased pressure drop (resistance to airflow) through the filter media arising from particle re-tenuon (loading). ms ndicated in Section 3.0, new orclean filters are designed to provide a resistance ofabout 0.25 kPa. First (1991) notes that "a dust pickup ofapproximately 1 kgllOOOm3/h of design filtrationcapacity represents a resistance increase of 0.25 kPafrom a new condition. In a relatively clean environ-ment, filters may be used for several years before re-placement is required. Dust loading, along with heatand radiation, also have the potential to reduce the ef-fectiveness of the organic materials used to strengthenthe filter medium and provide water repellency.

High moisture content is another stressor that can leadto increased pressure drop as well as reduced filter me-dium strength. Ricketts, Ruedinger, and Wilhelm

nI>ote that 'Moisture .nduced .... [deterioration]in filter performance and filter structure failure resultprincipally from the presence of liquid water in the fiberstructure of the filter medium." These authors discussincorporation of liquid water into the filter medium bysorption, condensation, or droplet filtration (intercep-tion) and also summarize a large number of other litera-ture sources that deal with moisture effects on filter me-dia and construction materials and assembled units. Inconclusion, they note that * ... water in the filtermedium leads to an increase in differential pressure andto. . . [deterioration] in filter pack stability and in filtermedium performance characteristics, especially thetensile strength. The mechanical load on the filter isthus increased at the same time that the structuralstrength is decreased. The end result is that filter

structural failure can occur for unacceptably low valuesof &P, even at design flow rates....' (Ricketts,Ruedinger, and Wilhelm 1987). In reporting the resultsof an experimental program these same authors notethat 'the 21 filters tested at 500C, constant volumetricair flow rate, and ambient pressure demonstrated thatthe differential pressure of new clean filters increasedsignificantly only above 95% RH, rose up to 0.5 to 2 kPaat 100% RH and reached values between 6 and 9 kPaduring filtration of liquid-water aerosols ... .(Ruedinger, Ricketts, and Wilhelm 1985). Additionalinformation concerning the effects of moisture, empha-sizing immediate failure rather than gradual deteriora-tion and including details related to dust-loaded filters,is presented later in this report.

processs that cancnge the- ph sca ca -acteristics of HEPA filter components other than themedia include corrosion of metal members and physic-cMenucai reactions that alter the properties of sealants,gaskets, ano water repellents. Metal components thatcould be affected include the frames and corrugated sep-arators. Stressors associated with physicochemical reac-tions potentially affecting face gaskets and the adhesivesand sealants that are used to splice the medium, fastenthe gaskets to the filter face, and seal the filter pack tothe frame, include heat and radiation.

Johnson et al. (1989) have reported results of tests toquantitatively evaluate the effect of aging on HEPA fil-ters. In one group of tests aged, flat, sheet media sam-ples (obtained by dismantling filters with 13 to 14 yearsof service) were tested using the equipment and proce-dures of MIL-F-5Y1079 (Department of Defense 1985),the specification discussed earlier that includes the phys-ical and chemical criteria for the media used in new fil-ters. In summary, 42% of the samples failed tensilestrength tests; i.e., the samples did not meet specifica-tions for new media. Johnson et al. also note that 'dueto the brittleness of the bend area of the pleat, a sampleculd'notbe obtained for testing. There is no doubt,

Towfver,t hat tareao eagedHEPA filter m eia

4.1 NUREG/CR-6029

Aging

represents the weakest part and it should have evenlower tensile strength value.' The bench-scale tests alsorevealed a signican loss or water repellency. All sam-ples failed when the top or dirty side of sheets weretested while four of the seven samples failed to meet therequirements for water repellency when the bottom orclean side was examined. As anticipated, due to service,71 % of the samples failed the pressure drop test. How-ever, all samples passed the DOP efficiency tests (metthe 99.7% requirement for new material).

In addition to the above bench-scale tests, several of theaged filters were exposed in a wind tunnel to generatepulses equal to those associated with the standard NRCRegion I design basis tornado, and using a shock tube,to shock wave overpressures. Six filters, in service for15 to 19 years, and two filters, in service for 14 years,were subjected to the tornado pulse and additive shockoverpressure tests, respectively. Johnson et al. reportthat *the average breaking pressure of the six aged filterswas 1.38 ± 0.95 psi 19.51 + 6.6 kPal. The comparablebreaking pressure of a new unused filter is 2.89 psi119.9 kPal The average breaking pressure of the twoaged filters utilizing small incremental pressureincreases is 1.8 psi 112.4 kPal. The comparable breakingpressure of a new unused HEPA filter is 2.5 psi117.2 kPal...." In addition to these 52% and 28%decreases in breaking pressure, a large increase in com-plete filter pack blow-out was reported for the aged fil-ters during the simulated tornado tests.

4.2 Adsorbers

Exposure to air containing stressors, namely moistureand contaminants or pollutants, can slowly and continu-ously degrade ete performane of gas rbers withime. ihis aginigalso termed weatvierhng, is inherent

use of the nature of the material used, i.e., one thathas been 'activated' to dramatically increase surface tomass ratio and provide countless reaction sites, coupledwith the essentially ubiquitous nature of the contami-nants. Many airborne constituents, including moisture,can readily react or be adsorbed by carbon beds reducingthe number of active "sites" that otherwise would beavailable for radioiodine adsorption. During normaloperation, charcoal may be exposed to air flows contain-ing contaminants including volatile organic solvents,

sulfur dioxide, nitrogen oxides, and carbon monoxide.Because of system flow rates, even traces of pollutantscan have a significant, cumulative effect. Oxidation, aswell as competitive loading, can impair bed perform-ance, including decreasing the efficiency of the impreg-fian. Because airborne constituents can vary dramati-carlywith time and from site-to-site, it is essentiallyimpossible to provide criteria concerning the useful lifeof impregnated activated carbon.

Aging degradation or weathering of iodine adsorbers asa result of exposure to various stressors has been treatedextensively in the literature, particularly in terms ofmethyl iodide retention efficiency. Results of numerousexperimental investigations involving the latter com-pound have been reported. Two summary reviews,which also include numerous pertinent references, havebeen prepared by Wilhelm and Deuber (1991) and agroup of Nuclear Energy Agency (NEA) experts[Organization for Economic Cooperation and Develop-ment (OECD) 19841. Briefly, studies have revealed thatincreasing temperature and humidity can be agents thatcause aging and aging effects. Deitz (1978) notes that0... there is strong evidence that the interaction of watervapor and charcoal is a significant factor in the degrada-tion of the charcoals when the RH is 70% and greater.The laboratory air mixtures studied [in this investigationof methyl iodide penetration] were water vapor, watervapor and sulfur dioxide, water vapor and ozone, andwater vapor and carbon monoxide....'

Wren and Moore (1991a) in an investigation of the ad-sorption and desorption behavior of contaminants stud-ied the effect of dry and humid conditions on adsorptionof NO2, SO2, methyl ethyl ketone (MEK), and NH3.Their conclusions concerning the latter, humid condi-tions include the observations that '. . . adsorbed waterincreases the adsorption rate and capacity of TEDAcharcoal for NO2, while it does not significantly changethose for SO2. However, it appears that SO2 is adsorbedas H2SO4 on the wet charcoal. Adsorbed water slightlyreduces the adsorption capacity of the charcoal forMEK, but does not affect the adsorption of NH3.'These same authors, reporting on the effect of contam-inants on CH31 removal efficiency, noted 'he efficiencyof TEDA charcoal is degraded most by NO2 and SO2,NH3 has a negligible effect, and MEK produces a milddegradation. The degree of degradation parallels the

NUREG/CR-6029 4.2

Aging

contaminant's ability to be chemisorbed.... Nitrogendioxide adsorbed under dry conditions is more effectivein degrading the CH31 removal efficiency of the charcoalthan when adsorbed under humid conditions. On theother hand, a completely opposite result is observed forSO2. The MEK contaminant behaves similarly to SO2,but the effect of humidity was less significant than forSO2. Ammonia has no effect on the efficiency of thecharcoal regardless of humidity" (Wren and Moore1991b).

Amine (TEDA or derivatives to reduce volatility) im-pregnated carbons age less rapidly than those where KIis used or that are unimpregnated (OECD 1984). In aninvestigation of temperature and humidity on the agingof TEDA impregnated charcoals, Billinge andBroadbent (1989) note that "... KI impregnated char-coals ... rapidly lose efficiency when exposed to highhumidity air land) this deterioration was mainly causedby oxidative aging of the carbon surface... Amines area class of compounds well-known for their anti-oxidantproperties so TEDA carbon may age more slowly thanKI material because the rate of formation of organo-oxygen surface groups is reduced; a second possibility isthat the basic properties of the amine neutralize anysurface acidity formed...."

The effect of the accumulation of organic contaminantson the aging process was investi ated be Hyder (1289).As in the case of the above investigation involvingMEK, results 1. . . show the effect of organics on carbonperformance [in terms of methyl iodide] is limited ...Some inference is possible regarding the mechanism bywhich sorbed organics affect methyl iodide retention. Inthese used carbon samples, all of the impregnant TEDAhas been lost by evaporation or reaction. Methyl iodideretention depends on the availability of the potassiumiodide impregnant for exchange with the methyl iodide.Other aging processes chemically alter this impregnantor remove it, but the sorbed organics do neither...."

As indicated above, much of the literature concerningaging mechanisms and effects is related to methyl iodidepenetration. It is noted in an older but valuable and

comprehensive review of air cleaning systems (ACS)(Burchted, Kahn, and Fuller 1976) that "the loss incapacity for elemental iodine is much slower than thatfor methyl iodide. Beds exposed continuously to flowingair at one installation showed adequate remainingcapacity for elemental iodine after 4 years of service. Atother installations, however, exposure of beds to paintand solvent fumes reduced capacity to the point thatefficiency fell to unacceptable levels in only a fewmonths...." Adequate long-term performance in termsof removing elemental iodine from plant ventilationexhaust air under normal operating conditions is alsoreported in Pelletier et al. (1978). The associated filtersystems were used continuously for treating effluentsfrom the auxiliary building. Average efficiency over a3-year period was 99.23%.

Relatively rapid deterioration of stainless steel compo-nents can result from contact with wet carbon. Severepitting resulting from galvanic corrosion has been re-ported (LIening 1991). Obviously, carbon should be re-moved immediately if it becomes wet. The revised NRCRG 1.52 . . . will clearly indicate that wetting ofimpregnated activated carbon should be avoidedbecause it establishes conditions for rapid chemical cor-rosion between the impregnants and stainless steelstructural material supporting the charcoal beds"(Bellamy 1991).

Finally, it should be emphasized that the possibility ofimpaired performance in terms of radiolodine retentionas a result of the aging or weathering of impregnatedactivated carbon has been a concern almost from thestart of the nuclear industries' existence. In fact,concern about the useful life of carbon could be con-sidered one of the first, if not the first, aging issue. Aswill be detailed later in the report, the issue is addressedby a surveillance test, a series of tests that areperiodically conducted to define the remaining radio-iodine retention capacity of used or exposed carbon.However, as will also be discussed, additional considera-tion of aging issues is believed to be warranted for bothHEPA filters and adsorbers, especially in relation toaccident environments.

4.3 NUREG/CR-6029

5 Failure Experience

This section describes failure experience that has beenreported for HEPA filters and adsorbers. Where possi-ble, the experience is discussed in terms of aging effectsand degradation and in conjunction with specific stres-sors. Information related to both normal operating andaccident conditions is presented. Analyses of LERs sub-mitted by commercial nuclear power plant operators arereported along with results of a survey of DOE sites toobtain information on the number of and reasons forHEPA filter changeouts and failures. Information toevaluate air-cleaning system performance during theTMI accident is also summarized.

5.1 Licensee Event Reports

An unusually large amount of potentially useful failuredata were found in the literature. Analyses of LERscovering the periods from January 1, 1975 throughJune 30, 1978, 1978-1981,1985-1987, and 1988-1991, toidentify those pertaining to air-monitoring, air-treatment, and ventilation systems, have been preparedby Moeller (1979), Moeller and Sun (1983), Jacox(1989), and Sommer and Otermat (1992). In summary,15% of 43,500 reports were associated with the subjectsystems while an estimated, roughly 100, or about 2% ofthe system reports, appeared to be associated withHEPA filters and carbon adsorbers. No relationship toaging was reported.

Moeller and Kotra (1985) and Moeller and Sun (1987)reviewed LERs for the periods 1981 through 1983 and1984 through 1986, respectively, to isolate those pertain-ing to control room habitability. Of the approximately20,800 reports, 4% were related to this focus. Duringthe 1981 through 1983 period, 28 of the LERs wereassociated with impaired filter or adsorber efficiency.Moeller and Kotra note that 'The majority of LERs re-lated to HEPA filters and charcoal adsorbers reflect thatthe performance of such units degrades with time andthat routine tests reveal that their efficiencies do notmeet technical specifications. The normal licensee re-sponse is to replace the units. More serious occur-rences, however, include the fouling of charcoaladsorbers by welding and/or paint fumes and the

occasional wetting of filters or adsorbers by drainagewater or the inadvertent actuation of their associatedfire protection deluge systems. At one plant, paintfumes contaminated the control room emergency char-coal filters so severely that their measured removalefficiencies were significantly degraded. Apparently, thecontrol room emergency ventilation system was in a re-circulation mode during painting in an adjacent room.Backflow through component drain lines into filterpackages necessitated their replacement at one plant...Unintentional actuation of fire suppression systems wasreported at three... plants... [that] rendered the filtersand adsorbers inoperable and they had to be replaced."

Moeller and Sun note that "a total of 30 LERs related tocharcoal adsorber problems were filed between 1984 and1986. Many of these events resulted from tests havingrevealed that the charcoal adsorber bank associated withan emergency ventilation system was not able to removethe required more than 99% of the halogenated hydro-carbon refrigerant test gas as stipulated by the TechnicalSpecifications [test is discussed later in the report inSection 6.2]. In most cases, the reduction in adsorptioncapability was thought to be due to moisture havinggained access to the charcoal. Compounding the prob-lem was the fact that round robin tests had shown thatcommercial charcoal testing laboratories, both in theU.S. and abroad, lacked the capabilities of determiningthe adsorption capacity of such charcoals on an accurateand reliable basis.'

5.2 DOE Site Survey

Using a survey, Carbaugh (1983) collected data on thenumbers of and reasons for HEPA filter changeouts andfailures at DOE sites for the years 1977 through 1979.A total of 1105 filter failures, or 12% of the 9154 filterapplications, was reported. The ratio of filters changedto those failed, 6.2 to 1, suggested that most filters arechanged out prior to failure. Carbaugh notes that "thelargest majority (63%) of filter changeouts were attri-buted to high differential pressure (0P) across the filter,indicative of filter plugging.... The vast majority offilters were reported as [having been] exposed to no

5.1 NUREG/CR-6029

Failure Experience

distinguishing environmental characteristics (i.e., theyfiltered essentially clean, dry, air environments similarto those that might be found in typical building ventila-tion systems or in systems with good pre-HEPA treat-ment features).... The majority of [the 11051 failuresoccurred for unknown or unreported reasons.... Theincidences of frame failures, gasket or seal failures andfilter media raptures were approximately equal; eachconstituted between 5% and 6% of all filter failures.When frame failure was identified, essentially all failedframes (58 out of 65) were wood. No observation ofsteel frame failure was reported... calculating theratios of seal-type failures to seal-type applications[3920 gaskets and 151 fluid seals] can lead one to con-clude that little difference may exist between gasket andfluid seal failure rates.... Al media ruptures occurredin wood-frame filters, and almost all (48) had aluminumseparators and polyurethane foam sealants. [Media toframe] Sealant failure was not identified as a significantfilter failure mode (3 occurrences reported out of 1105).... Ratios of filters failed to filter applications in whichthese failures were experienced are several times higherfor HF acid and high moisture environments than forenvironments having no distinguishing characteristics,or the average of all single environment exposures[exposure of filters to a single significant environmentalfactor, e.g., solvent, high moisture, high dust]...."

Johnson et al. (1989) explain the results of the abovesurvey in the following manner: "If one examines thecategories reported for filter failure, it is evident that theeffects of aging could contribute to 81% of these fail-ures, except for the 19% resulting from handling or in-stallation damage." Johnson et at. also provide twoother references that include qualitative informationthat reveal indications of aging: a technique to dismantleand separate HEPA filter components, based on unfold-ing the pack and rerolling the media, worked well onnew filters but became impractical on used filters be-cause of the "weakness of the filter paper"; and as part ofa report concerning the in-service aging effects of un-used and used cleanroom, chemical, and radioactivecontaminated HEPA filters it was noted that "aluminumspacer deterioration was observed in all used filters ex-amined with the most significant levels observed whenhigh humidities and acid gases were present."

5.3 TMI Accident

As suggested earlier in the report, selection of adsorbersfor Phase I study was reinforced by literature referencesrevealing that the low radioiodine decontamination fac-tors (DF) associated with the TMI accident were attri-buted to the premature aging of charcoal. Before pre-senting details related to aging issues, it should beemphasized that building ventilation systems played animportant role in decreasing the radionuclide release as-sociated with this incident and in fact, prevented the re-lease of most of the radioiodine. The estimated quantityof I-131 released was 5.6 x 10-9 Bq or 15 Ci (in contrastto about 9.3 x 104 Bq or 2.5 million Ci of noble gas)(Rogovin and Frampton 1980).

Despite the above, the terms "somewhat dismal results'and 'rather unsatisfactory" have been used to describeexhaust air filtering experience during the TMI accident(OECD 1984; Wilhelm and Deuber 1991). In summary,it was determined that the filtering systems installed atthe time of the accident provided a DF of 9.5 for all spe-cies of radioiodine (corresponds to a retention efficiencyof 89.5%), and the radioiodine releases were higher byabout a factor of five than they would have been if therehad been no system deterioration. The preceding valuesand reasons for the degradation are detailed in Rogovinand Frampton (1980) and discussed by Bellamy (1981)in a paper that was prepared to summarize pertinent(related to air-cleaning technology) efforts and recom-mendations of various investigative groups. Briefly, sys-tem degradation and the fact that radioiodine retentionwas associated with an efficiency of 89.5% was ulti-mately attributed to pre-accident aging of the impreg-nated activated carbon. For a number of reasons venti-lation air was passed through exhaust systems from thetime of carbon installation to just shortly after the acci-dent. This approximately 1-year time frame includedperiods of exposure of the carbon beds to fumes frompainting and cleanup efforts. Furthermore, exemptionsto technical specifications postponed surveillance teststhat could have revealed the extent of aging. (Tests aredescribed in the next section and involve the periodiclaboratory evaluation of exposed carbon samples.) Theexemptions also permitted the use of carbon with amethyl iodide removal efficiency of 96.97%.

NUREG/CR-6029 5.2

Failure Experience

Carbon removed from four filter trains following theaccident, along with data on various iodine species ob-tained from an air sample, suggest removal efficienciesranging from 49.1% to 75.6% and 98.5% to 99.9% formethyl iodide and elemental iodine, respectively. Theseefficiencies are based on calculated releases conserva-tively assuming 95% RH and percentages of volatilespecies of 40% and 35% for organic and elementaliodine, respectively. The efficiency range is attributed tounbalanced ventilation flows. It should be noted that itwas estimated that the highest RH that existed insidethe auxiliary building during the accident was 80%.

Review indicated that releases of radioactive material inparticulate form were negligible. Postaccident effortsincluded changeout of HEPA filters as well as carbon.As indicated in Rogovin and Frampton (1980) TheseHEPAs were visually examined before changeout andwere intact and in satisfactory condition, but were dam-aged during changeout.... Unfortunately, no usedHEPA filters or sections of filter media were retainedfor analysis." Again, the highest RH estimated for theauxiliary building was only 80%.

5.3 NUREG/CR-6029

6 Inspection, Surveillance, and Monitoring Methods

This section describes the inspection, surveillance, andmonitoring methods (ISMM) that are used to establishthe condition of HEPA filters and adsorbers once thecomponents have been placed in operation. As will bedetailed below, the use of surveillance tests, series oftests periodically performed to monitor component con-dition and demonstrate the current ability to removefine particles and iodine and iodine compounds, is em-phasized. Surveillance tests include in-place leak testsand visual inspections to evaluate filter banks and ad-sorbers in terms of component damage and bypass. Alsoincluded, and specifically related to aging concerns, arelaboratory tests of used or aged carbon samples to deter-mine remaining radioiodine adsorption capacity. In ad-dition to the surveillance tests, HEPA filter pressuredrop is continuously monitored (alarmed andindicated).

6.1 HEPA Filters

Instrumentation provisions, including alarms, for air-cleaning systems and components are presented inASME N509-1989 (ASME 1989a). High differentialpressure alarms are specified for local as well as remote-manned control panel locations for the HEPA filters ofESF air-cleaning systems. Pressure drop indication isspecified for the local location. For non-ESF units, dif-ferential pressure indication and high differential pres-sure alarms are recommended only for the local loca-tion. Particle loading along with the moisture contentof the air most adversely influence the rate and extent offilter pressure drop increase. As indicated earlier, mostfilters are changed out prior to failure, primarily as a theresult of an indication of a high differential pressureacross the filter.

Criteria for surveillance tests involving installed systemsand the various components are presented in ASMEN510-1989 (ASME 1989b). As in the case of new com-ponents, other documents, including NRC RG 1.52 andRG 1.140, supplement and support this basic reference(USNRC 1978,1979). Visual inspection is recom-mended before each test series and specific guidance isprovided for HEPA filters and adsorbers. Air flow

distribution tests for both HEPA filter and adsorberbanks are described. 'Testing to verify uniform mixing inthe air stream approaching the HEPA filter bank oradsorber stage is a prerequisite for conducting surveil-lance leak testing of the installed filter bank. The rec-ommended frequency for this "air-aerosol mixing uni-formity test" includes performance after completion ofinitial construction and after major system modificationor repair (acceptance tests). The method is based on theintroduction of DOP aerosol and '. . . concentrationreadings ... taken across a plane parallel to, and a shortdistance upstream of, the HEPA filter bank or adsorberstage (ASME 1989b).

Surveillance leak testing of installed HEPA filters isspecified because gradual deterioration and leaks coulddevelop under standby as well as service conditions.This test is also based on a DOP challenge aerosol intro-duced upstream of the filters. Concentrations upstreamand downstream are then measured. Recommended fre-quency given in Tible 1 of ASME N510-1989 includesacceptance testing, after each HEPA filter replacementand at least once each operating cycle (ASME 1989b).Nuclear Regulatory Commission RG 1.52 states that'HEPA filters should be tested in place (1) initially,(2) at least once per 18 months thereafter, and (3) fol-lowing painting, fire, or chemical release in any ventila-tion zone communicating with the system to confirm apenetration [ratio of downstream to upstream concen-tration in percent] of less than 0.05% at rated flow...Filters that fail to satisfy this condition should bereplaced....' (USNRC 1978). Nuclear RegulatoryCommission RG 1.140 contains essentially the sameinformation (USNRC 1979). It is recognized that thisoriginal guidance should be changed and as indicatedearlier, revisions to RG 1.52 are planned. Specifically,Bellamy (1991) notes that guidance in the revised ver-sion of the RG will be supplemented such that the18-month criteria will include"... or once per refuelingoutage. The requirement for testing after painting, fire,or chemical release will include additional guidance toindicate that this testing need be done only if communi-cation with the system occurred in such a manner thatthe HEPA filters or carbon adsorbers could become ad-versely affected by the fumes, chemicals, or foreign

6.1 NUREG/CR-6029

Inspection

materials. Testing will also specifically be required(1) after each partial or complete replacement of aHEPA filter bank...." (Bellamy 1991).

The surveillance leak test of the installed filters shouldnot be confused with the efficiency tests for new individ-ual filters that were described earlier in the report. Apolydisperse DOP aerosol is used for the former, in-place test while a 0.3,um monodisperse aerosol is usedfor the latter. It should be noted that both RG 1.52 andRG 1.140 state that a filtration system satisfying theabove 0.05% maximum penetration requirement'. . . can be considered to warrant a 99% removal effi-ciency for particulates.. .' (USNRC 1978, 1979).

6.2 Adsorbers

The information concerning visual inspection and theair flow distribution and aerosol mixing tests discussedbriefly in the previous section also pertains to adsorbers.In the case of the installed adsorber bank, the surveil-lance leak test is based on a halide challenge gas. Theslightly adsorbable and readily desorbed halide (haloge-nated hydrocarbon refrigerant, fluorocarbon) gas is in-jected upstream of the adsorber bank. As in the case ofthe filter test, concentration is then measured upstreamand downstream of the bank and percent of penetrationdetermined from the ratio of downstream to upstreamvalues. The recommended test frequency given inASME N510 (ASME 1989b) is the same as that for thefilters and includes acceptance testing, after each ad-sorber replacement, and at least once each operating cy-cle. Regulatory Guides 1.52 and 1.140 require testinginitially, at intervals of 18 months thereafter, followingremoval of samples for laboratory tests (if the integrityof the adsorber section is affected), and following paint-ing, fire, or chemical release in any ventilation zonecommunicating with the system (USNRC 1978, 1979).Both RG also indicate that bypass leakage through theadsorber should be less than 0.05%.

The above planned revisions to RG 1.52 to supplementthe 18-month test frequency and require testing only ifcommunication occurred in a manner that componentscould become adversely affected by fumes, etc., will alsoinfluence adsorber testing. In addition, 'testing will alsospecifically be required (1) after each partial or

complete replacement ... of a carbon adsorber in anadsorber section or bank, (2) following detection of, orevidence of, penetration or intrusion of water or otherforeign material into any portion of an ESF atmospherecleanup system, and (3) for adsorber banks followingremoval of an adsorber sample for laboratory testing ifthe integrity of the adsorber section is affected'(Bellamy 1991).

As in the case of the polydisperse DOP test of filters, thein-place halide challenge test is a leak rather than an ef-ficiency test. For adsorbers, however, the leak test issupplemented with laboratory tests of aged or used car-bon samples to determine system efficiency and remain-ing capacity for methyl iodide. It is stated in ASMEN509 that 'sufficient test canisters or other means of ob-taining samples ... of used adsorbent shall be installedin the adsorber system to provide a representative deter-mination of the response of the adsorbent to the serviceenvironment over the predicted life of the adsorbent.lest canisters shall be installed in a location where theywill be exposed to the same air flow conditions as theadsorbent in the system, shall have the same adsorbentbed-depth as the adsorbent in the system, and shall befilled with representative adsorbent from the same batchof adsorbent as that of the system' (ASME 1989a).Recommendations concerning the number of canistersto be provided based on the expected frequency of op-eration are also included in ASTM N509. Details con-cerning the design basis for samplers and the generaltypes are presented in Appendix A of ASTM N509,'Sampling of Installed Adsorbents for SurveillanceTesting.'

Recommended frequency for the above laboratory ad-sorbent tests, presented in Tbble 1 of ASME N510,includes acceptance tests, before each adsorber replace-ment, and at least once each operating cycle. Supple-mental footnotes state that 'adsorbents must be testedbefore installation or replacement to establish effi-ciency. Samples for laboratory testing should be takenbefore the routine in-place testing of the installed sys-tem to verify the condition of the adsorbent. . . Labora-tory tests shall be made to confirm performance atintervals not exceeding 720 hr of system operation or forany system immediately following inadvertent exposureto solvents, paints, or other organic fumes or vapors thatcould degrade performance of the adsorbent. The

NUREG/CR-6029 6.2

Inspection

720 hr requirement may be modified based on labora-tory test histoiy (ASME 1989b). As in the case of thequalification of new carbons, the test method and lab-oratory apparatus described in ASTM D3803 are speci-fied for use in the quantitative evaluation of the reten-tion properties of the aged carbon samples (ASTM1989). Information to improve repeatability and accur-acy is provided in Appendix B of ASME N510, 'Addi-tional Guidance for use of ASTM D3803, 1979.0

Frequency guidelines listed in NRC RG 1.52 and RG1.140 for the surveillance test involving adsorbent

sampling and laboratory testing are essentially the sameas those recommended above. Used carbon decontami-nation efficiencies are specified for both ESF andnormal systems for a variety of air-filtration system loca-tions, carbon bed depths, and relative humidities. Forexample, a 95% decontamination efficiency is specifiedfor both elemental iodine and organic iodide, for a50.8-mm (2-in.) bed depth, an ESF system designed tooperate outside the primary containment, and whereRH is controlled to 70% (USNRC 1978).

6.3 NUREG/CR-6029

7 Conclusions

This section presents conclusions and describes con-cerns originating from the study. Opinions and recom-mendations concerning the need for expanded or morecomprehensive assessments are also presented. Con-siderations related to the need for additional studiesinclude the safety significance of failure modes andexperience, evaluation of ISMM, and the range of condi-tions that HEPA filters and adsorbers may experience.Safety significance was emphasized as part of the back-ground information provided earlier in this report. Inreview, component failure can impact both plant andpublic safety. Satisfactory performance is essential toensure control room habitability, and HEPA filters andadsorbers can be the last barrier in preventing therelease of radioactivity. Monitoring and surveillancetests have been established to observe HEPA filter pres-sure drop buildup, check both HEPA filters andadsorbers for bypass or pathways through which air canescape treatment, and determine the retention effective-ness of aged carbon. Conditions range from those asso-ciated with normal operating environments to thoseestimated for reactor accidents. Further development ofthe preceding considerations, in terms of normal, error-induced, and accident conditions, follow. The need forthe expansion of the Phase I interim aging assessment toother air-treatment system components is alsodiscussed.

7.1 Normal Operating and Error-Induced Conditions

Normal operating conditions are characterized as rang-ing from the exposure of components to dry air contain-ing trace atmospheric contaminants or stressors (e.g.,sulfur dioxide and nitric oxides) to relatively humid airand even occasionally to potentially polluted air flowsfrom routine activities such as maintenance and testing.Because of large air flows, even traces of pollutants orcontaminants can have a significant cumulative effect onadsorber performance. Two examples of error-induced,degraded adsorber retention efficiency from relativelyhigh contaminant levels (i.e., from the exposure tofumes from cleanup and/or painting) are reported inSections 5.1 and 5.3. The surveillance test involving the

laboratory quantification of the extent of degradation ofused carbon samples is specifically designed to revealaging or premature aging as a result of the normal orerror-induced exposure of adsorbers to contaminants orpollutants.

fteatment of large volumes of air containing small par-ticle concentrations can, of course, result in a gradualbut continuous increase in pressure drop through filters.Pressure drop or resistance to air flow across HEPA fil-ter banks is monitored. First (1991) notes that'although HEPA filters are qualified to maintain theirintegrity and filtering efficiency up to a minimum resist-ance of 2.5 kPa, they are seldom operated up to thatresistance level because of fan and fan motor limita-tions. An increase to 1.0 kPa before change is commonfor nuclear installations to maintain a large reservecapacity at all times in readiness for extended service lifein the event of an emergency' (First 1991). The surveil-lance test involving the leak test of HEPA filter banksshould detect defects in the filter media or unit arisingfrom normal service conditions and that result in bypass,including those that could result from the aging of gas-ket material. However, neither pressure drop monitor-ing nor the in-place surveillance test using a polydis-perse DOP aerosol will provide indications of aging interms of reduced filter medium strength. The latter, thelack of indication of reduced strength, may be importantwhen considering reactor accident conditions.

7.2 Reactor Accident Conditions

Nuclear Regulatory Commission regulations require at-mospheric cleanup systems to mitigate the consequencesof postulated accidents by removing radioactive mate-rials released during the accident from containment ves-sel or building atmospheres. General Design Criterion(GDC) 41, 'Containment Atmosphere Cleanup* of Ap-pendix A, 'General Design Criteria for Nuclear PowerPlants," to fltle 10, Part 50 of the Code of Federal Reg-ulations states that 'systems to control fission prod-ucts .. which may be released into the reactor contain-ment shall be provided as necessary to reduce, consis-tent with the functioning of other associated systems,

7.1 NUREG/CR-6029

Conclusions

the concentration and quality of fission products re-leased to the environment following postulated acci-dents ... I General Design Criterion 61, Ehel Storageand Handling and Radioactivity Control" states that thefuel storage and handling, radioactive waste, and othersystems which may contain radioactivity shall be de-signed to assure adequate safety under normal andpostulated accident conditions. These systems shall bedesigned ... with appropriate containment, confine-ment, and filtering systems, ... ' General Design Criter-ion 19, "Control Room' requires that' ... adequateradiation protection shall be provided to permit accessand occupancy of the control room under accident con-ditions ... . (10 CFR 501992).

Nuclear Regulatory Commission RG 1.52 presentsacceptable methods for implementing the above (andother) GDCs and includes guidance that aging as well aspostulated accident conditions should be consideredduring component design. As indicated earlier, this RGapplies '.. . to postaccident engineered-safety-feature at-mosphere cleanup systems designed to mitigate the con-sequences of postulated accidents. It addresses the ESFatmosphere cleanup system, including the various com-ponents and ductwork, in the postulated DBA [design-basis accident) environment.... systems that must oper-ate under postulated DBA conditions inside the primarycontainment. . . (i.e., recirculating systems) aredesignated as... [primary). ESF systems required tooperate under conditions that are generally less severe...(i.e., recirculating or once-through systems) aredesignated as ... [secondary]. Secondary systemstypically include the standby gas treatment system[SGTS] and emergency ACS for the fuel handling build-ing, control room, and shield building.... Unless theapplicable engineered-safety-feature atmospherecleanup system operates continuously during all timesthat a DBA can be postulated to occur, the systemshould be automatically activated upon the occurrenceof a DBA.. . (USNRC 1978). Information is includedin the RG in recognition of the fact that environmentalconditions preceding a postulated DBA may contributeto aging and degraded performance of filters andadsorbers. It is specifically noted that aging needs to beconsidered during design and operation. Furthermore,remembering that the guide was developed for accidentmitigation systems, it is also indicated that all compo-nents need to be designed for reliable performance inpotentially hostile environments.

Typical DBA environmental conditions, including pres-sure surge, maximum pressure, maximum temperatureand RH of the influent, average radiation levels for air-borne materials and iodine buildup on the adsorber, andairborne iodine concentrations for elemental iodine andfor methyl iodide and particulate iodine, are tabulatedin RG 1.52 for the primary and secondary ESF systemsdescribed in the preceding paragraph (USNRC 1978).The radiation levels are based on U.S. Atomic EnergyCommission (USAEC) RG 1.3 and RG 1.4, which pro-vide acceptable assumptions for use in evaluating theradiological consequences of one of the postulated acci-dents for boiling- and pressurized-water reactors (i.e.,the design basis loss-of-coolant accident [LOCA]). Inboth of these regulations, it is stated that "twenty-fivepercent of the equilibrium radioactive iodine inventorydeveloped from maximum full power operation of thecore should be assumed to be immediately available forleakage from the primary reactor containment Ninety-one percent of this 25 percent is to be assumed to be inthe form of elemental iodine, 5 percent of this 25 per-cent in the form of particulate iodine, and 4 percent ofthis 25 percent in the form of organic iodides (USAEC1974a,b).

One of the two major concerns arising from this Phase Istudy is that the conditions developed for the above de-sign basis assessments may not realistically representthose under which HEPA filters and adsorbers, includ-ing aged components, may operate. Design-basis acci-dent evaluations are based on non-mechanistic hypothe-tical events. Pasedag, Blond, and Jankowski (1981) notethat ... the Design-Basis Accidents (DBAs) are a set ofaccidents which have been chosen to envelope the an-ticipated worst credible conditions in what was per-ceived to be a very conservative manner. Thus theseaccidents are not representative of expected or realisticconditions but have been judged to bound any credibleaccident... The DBA-LOCA cannot be expected to bepredictive of any specific accident situation.' Over thepast 10 to 15 years, a large amount of information hasbeen developed that may permit significantly better esti-mates of the conditions, including radiation levels, tem-peratures, humidities, and the chemical and physicalproperties of the materials that aged components mightencounter during severe accidents. This informationwas developed as part of, or in support of, probabilisticrisk studies and extensive efforts to reassess sourceterms (characterization of radionuclide releases from

NUREGICR-6029 7.2

Conclusions

reactor accidents) and has included efforts to provideimproved analytical models to describe thermal-hydraulic behavior and mass transport. The latterincludes detailed models that take Into account thechemical and physical processes associated with the re-lease of fission products from the degraded core and thesubsequent transport and retention of material throughand in the reactor vessel, coolant systems, containment,and exterior compartments.

A system of computer codes, the Source Tbrm CodePackage (STCP), has been developed for analyzing spec-ific postulated accident situations and calculating thequantity, timing, and characteristics of the release ofradioactive material to the environment following theincident (Gieseke et al. 1986). This code system or themore recently developed MELCOR computer code forsource term and risk assessment analyses could be usedto provide significantly more realistic insights into theimpact of an accident on filters and adsorbers (Summerset al. 1991). For example, rather than using assump-tions, information could be developed from calculationsthat would be extremely useful in estimating the rangeand combination of accident conditions that may be en-countered, e.g., in providing best-estimates of the likeli-hood for large particle concentrations and/or moistureconditions that could saturate and overwhelm HEPA fil-ters. An improved definition of accident conditions mayalso reveal that aging concerns associated with the filtersand adsorbers of 'normal' treatment systems may equalthose of ESF systems and that the designations of pri-mary and secondary for the latter may be trivial in termsof accident considerations. One of the findings resultingfrom investigations into the TMI accident was the factthat". . . a nonengineered safety feature filter system de-signed for normal operation only, i.e., the auxiliarybuilding exhaust ventilation filtration system, greatly re-duced the quantity of radioiodine release to the environ-ment. .. The safety grade versus nonsafety grade desig-nation [for the two air treatment systems operating atthe time of the accident] was meaningless...." (Rogovinand Frampton 1980).

It should also be noted that in a recent study Beahm,Weber, and Kress (1991) used the STCP to address theissue of the chemical form of iodine in the coolant sys-tem and in containment. Using calculated data fromseven severe accident sequences, it was found that 'inmost of the calculations for the seven sequences, iodine

entering containment from the reactor coolant systemwas almost entirely in the form of CsI with very smallcontributions of I or HI The study did not address ulti-mate disposition. Iodine could, of course, subsequentlybe converted to the volatile form, e.g., if the CsI dis-solved in low pH water, was ultimately converted toorganic form, or decomposed after being captured onfilters. However, the study suggests the perhaps unex-pected conclusion that, during accidents, intact HEPAfilters could be significantly involved in iodine retention.Furthermore, past design basis evaluations based on theassumption that iodine is primarily in the volatile formmay neither be correct nor conservative when evaluatingair-cleaning system performance. L1l-defined accidentconditions even where a DBA philosophy is applied" aredescribed as a major problem by the group of NEA ex-perts. As further support for the possibility of noncon-servative assumptions, this group also notes that *. . . theDBA may not represent the most severe conditionsunder which the ACS has to operate" (OECD 1984).

The second major concern arising from the Phase Istudy is that even if details concerning accident condi-tions were available, there may be insufficient informa-tion to develop reliable predictions concerning the per-formance of aged components in such potentiallyextreme environments. In 1984, the group of NEAexperts characterized the extent of knowledge that isavailable to evaluate aged HEPA filter and adsorberperformance during accidents as non-existent and forthe most part, poor, respectively. The group of expertsstated that "...the available information on HEPA filtersrelates primarily to new filters and virtually no data existfor the normal or accident condition behavior of agedfilters. Many ACS HEPA filters remain in place formore than 5 years. It is unrealistic to expect that acci-dents will occur only with freshly installed HEPA filterbanks" (OECD 1984). When considering aged HEPAfilters in conjunction with accident environments, thegroup described the extent of knowledge as non-existentfor the following parameters: pressure differential,vibration, high humidity/free water, chemicals, radiation,temperature, and loading capacity. When appraising theextent of knowledge for aged adsorbers, pressure dif-ferential was identified as good; vibration as non-existent; radiation as fair, and high humidity/free water,chemicals (weathering), temperature, and loadingcapacity as poor.

7.3 NUREG/CR-6029

Conclusions

The above relative gloomy picture has brightened some-what in recent years, presumably as the result of theemphasis on severe accidents. Information has been ob-tained to at least partially fill in some of the gaps associ-ated with the above parameters. As indicated earlier,above saturation, new or clean filters are vulnera todeterioration and failure. Obviously, high humidity airflows can jeopardize dirrtice iloaded filters. Ekem6s -an M&ch__amsms of the failure of dfust-loaded fil-ters under high humidity conditions are discussed inRicketts, Ruedinger, and Wilhelm (1987) and inRuedinger, Ricketts, and Wilhelm (1985). The struc-tural limits, in terms of AP at failure, are tabulated in thelatter reference for HEPA filters tested under high airhumidities at 1700 m3ih and 500C. Norman (1987) re-ports results from the exposure of filter media to steam.The media were dried subsequent to exposure and eval-uated with respect to penetration and tensile strength.IncreasedPettion (using DOP aeross)was ob-svedvef the media to high humiditiesandcelevated temperature. Tensile strength decreased toa6ut 40% of its original value. Aspects of the increasein pressure drop of 1HEPA filters under fog conditionswas also investigated by Ricketts, Ruedinger, andWilhelm (1990). 'Airstreams with between 80% and100% RH primarily threaten dust-loaded filters. Atconditions above saturation, new clean filters alsobecome subject to rapid deteriorations in performance."Penetration of water through filters by seepage wascharacterized as 'rather rapid." Such penetration sug-gests that water soluble salts containing radionuclidescould penetrate even intact filters. The performance ofdownstream adsorption units could also be affected.

Performance limitations and efficiency of and releasesfrom HEPA filters at high temperatures were the topicsof three of the papers of Session 3B, "Response ofHEPA Filters to Physical Stress,' of the 20th DOE/NRCNuclear Air Cleaning Conference (Kratzke 1989). Ef-fects of temperature on HEPA filter media was again atopic for discussion at the most recent, 22nd, conferencewhere" .. . observed changes in strength and paper stiff-ness ... [were) explained in terms of alterations to thebinder due to thermal degradation" (Hamblin andGoodchild 1992).

A comprehensive study of the performance of TEDAand KI impregnated charcoals under 'reactor accidentconditions" is reported in Wren et al. (1989), "The

efficiency of these charcoals in removing CH3I and I2was studied as a function of temperature (250 to 800C),RH (O to 90% RH), radiation field (0 to 2kGy.hr-) withand without H2 , contaminants (SO2, NO2, NH3, andMEK) and other factors." Deitz (1985) reports that"based on the experimental results, it appears thatreactor-grade activated charcoal can satisfy the stipula-tion in RG 1.52 that requires successful trapping ofradiolodine during a DBAwith radiation levels up to[107 Gyl io1 rads." In fact, it was found that exposure tolevels of 167 to 169 regenerates the iodine isotope ex-change capacity. Paraphrasing Deitz, where the trap-ping of methyl iodide-131 is concerned, radiationimprovement rather than damage is observed.

7.3 Conclusions and Recommendations

The Phase I study found that the HEPA filters and ad-sorbers are considered to have a long service life,especially the filters. Thus, if a severe accident happensit is likely to occur at a time when these two final con-finement barriers have been in use for an extendedperiod, even years. Even with existing inspection, moni-toring and monitoring methods, aged, and possiblydegraded components, could fail to provide the radia-tion protection needed for safe shutdown or be the weaklink that allows the release of radionuclides to theenvironment Furthermore, the assessment has revealedthe need for an improved definition of accident condi-tions and the possible need for additional informationto evaluate the performance of aged components undersuch conditions. It is recognized that this improveddefinition of accident conditions is outside the scope ofthe NPAR program. It is also recognized that compre-hensive assessments limited solely to the developmentof information for performance evaluation could bemade. However, for such studies to be definitive, know-ledge concerning challenges must ultimately be ob-tained, as will be discussed below.

Limited information concerning the range of challengesexpected for filters and adsorbers during accidents isalready available as a result of source term reassessmentand reactor risk studies (Silberberg et al. 1986; USNRC1990). Additional insights concerning accident condi-tions should also become available as a result of recentregulatory efforts to provide improved source term

NUREG/CR-6029 7.4

Conclusions

definitions. As part of several regulatory activities toincorporate severe accident insights into the safetyassessment of future plants, the NRC has issued aproposed revision of the reactor accident source terms.The proposed revision is in terms of fission productcomposition, magnitude, timing, and iodine chemicalform, for release into containment (Soffer 1992). Utili-zation of these revised source terms should provideimproved estimates of accident conditions and the asso-ciated challenges to filters and adsorbers. For example,evaluation of iodine chemistry during the portion of thetransient that follows release into containment will beneeded, resulting in better estimates of whether iodineremains primarily in the particulate form or is convertedto a volatile species before it reaches filters andadsorbers.

As previously indicated, even if details concerning acci-dent conditions were available, there may be insufficientinformation to develop reliable predictions concerningthe performance of aged components. Quantification ofthe structural limits and failure mechanism of aged fil-ters in moist airflows and of the particle loadings thatwould cause failure of reduced strength and/or particle-laden filter media are examples of potential informationneeds. Statistically-designed engineering scale experi-ments, involving aged components and utilizing theimproved estimates of the ranges and combinations ofchallenges, may ultimately be justified. It is also possi-ble that improved estimates of conditions could negatethe need for further work, because either moderate orextremely severe stressors are predicted. In the case ofmoderate stressors, for example, additional studyappears unwarranted if near normal operating condi-tions are identified. In the case of severe conditions,HEPA filters are simply not sized to handle the massiveloading of non-radioactive particles that conceivablycould result from molten core-concrete interactions.

An improved understanding of severe accident condi-tions could also further emphasize the importance ofother air-treatment system components. For example,along with HEPA filters, greater emphasis would beplaced on coarse or roughing pre-filters and evendemisters if it were determined that high particle con-centrations were involved. As a result, expansion of thePhase I interim aging assessment to other componentsof nuclear air-treatment systems, namely demisters,heaters, coarse filters, fans or blowers, and dampers isrecommended. Several of these components, althoughnot designed to retain radionuclides, could mitigate theImpact of conditions that threaten to cause failure ofaged HEPA filters and adsorbers during accidents. Theaging of these protective components is, therefore,also of concern. The information that is developed forfans can be used in the assessment of other ESF systemsspecifically installed to ensure public and plant safety.In addition to air cleaning, fans or blowers are includedin heating, ventilation, and air conditioning and contain-ment cooling and pressure suppression systems.

Finally, it should be emphasized that information usedto estimate the performance of aged air-treatment sys-tem components can have a significant impact on thecalculated consequences of postulated accidents. This isillustrated by the analyses of a postulated loss of decayheat removal accident at Browns Ferry. The SGTS fail-ure model used, listed as the area of greatest uncertaintywith respect to informal sensitivity analyses, includesbest estimates concerning the filter loading that willcause failure, the type of failure, and the functioning ofthe adsorber under the projected accident environment.It is noted that "the SOTS failure model assumes primeimportance because the SGTS is the last barrier to theatmosphere in this accident sequence' (Wichner et al.1984).

7.5 NUREG/CR-6029

8 References

ASME,ASME Standard N509-1989; Nuclear PowerPlantAir-Cleaning Units and Components, The Amer-ican Society of Mechanical Engineers, New York, 1989a.

ASME,ASME Standard N510-1989, Testing ofNuclearAir Treatment Systems, American Society of MechanicalEngineers, New York, 1989b.

ASMEIANSI,ASME CodeASMEUNSIAG-1-1988;Code on NuclearAir and Gas Treatment, The AmericanSociety of Mechanical Engineers, New York, 1988.

ASTMASTMStandardD3803-89; Standard TestMethodfor Nuclear-Grade Activated Carbon, AmericanSociety for Testing and Materials, Philadelphia,Pennsylvania, 1989.

ASTM, ASTM Standard D4069-90, Standard Specifica-tion for Impregnated Activated Carbon Used to RemoveGaseous Radio-Iodinesfrom Gas Streams, AmericanSociety for Testing and Materials, Philadelphia,Pennsylvania, 1990.

Beahm, E. C., C. F Weber, and 1: S. Kress, IodineChemical Forms in LWR Severe Accidents, NUREG/CR-5732, U.S. Nuclear Regulatory Commission, Wash-ington, D.C., 1991.

Bellamy, R. R, Investigations Into the Air Cleaning As-pects of the Three Mile Island Accident." In 16th DOENuclearAir Cleaning Conference: Proceedings, Vol. 1, Ed.M. W First, pp. 1427-1441, CONF-801038, HarvardUniversity, Boston, Massachusetts, 1981.

Bellamy, R. R, 'USNRC Regulatory Guidance for Engi-neered Safety Feature Air Cleaning Systems.' In Proc-eeding of the 21st DOE/NRC Nuclear Air Cleaning Con-ference, Vol. 1, ed. M. W First, pp. 626-633. NUREG/CP-01 16, U.S. Nuclear Regulatory Commission, Wash-ington, D.C., 1991.

Billinge, B.H.M., and D. Broadbent, 'he Effects ofTbmperature and Humidity on the Aging of TEDA Im-pregnated Charcoals.' In Proceedings of the 20th DOEINRC NuclearAir Cleaning Conference, VoL 1, ed. M. W.First, pp. 572-588. NUREG/CP-0098, U.S. NuclearRegulatory Commission, Washington, D.C., 1989.

Burchsted, C. A., J. E. Kahn, and A. B. FuIller, NuclearAir Cleaning Handbook. ERDA 76-32, Oak Ridge Na-tional Laboratory, Oak Ridge, lbunessee, 1976.

Carbaugh, E. X, 'Nuclear Air Cleaning The Need for aChange in Emphasis.' In Proceedings of the Fourth DOEEnvironmentalProtection Information Meeting. CONF-821215, U.S. Department of Energy, Washington, D.C.,1983.

Deitz, V. R.,Effects of Weathering on Impregnated Char-coal Performance. NUREG/CR-0025, U.S. NuclearRegulatory Commission, Washington, D.C., 1978.

Deitz, V R., Charcoal Performance UnderAccident Con-ditions in Light-Water Reactors. NUREG/CR-3990, U.S.Nuclear Regulatory Commission, Washington, D.C.,1985.

EPRI, Nuclear Power Plant Common Aging Terminolog,EPRI TR-100844. Prepared by MPR Associates, Inc.,Washington D.C. for Electric Power Research Institute,Palo Alto, California, 1992.

First, M. W, 'Removal of Airborne Particles from Rad-ioactive Aerosols.' Chapter 2 in -Reatment of GaseousEffluents at Nuclear Facilities, eds. W.R.A Goossens,(. G. Eichholz, and D. W. Tbdder, pp. 13-66. HarwoodAcademic Publishers, New York, 1991.

Gieseke, J. A, P. Cybulskis, H. Jordan, K W Lee, P. M.Schumacher, L A. Curtis, R. 0. Wooton, S. F. Quayle,and V Kogan, Source Term Code Package, A User'sGuide (Mod 1). NUREG/CR-4587, U.S. Nuclear Reg-ulatory Commission, Washington, D.C., 1986.

8.1 NUREG/CR-6029

References

Hamblin, C and P. J. Goodchild, The Effects of Tem-perature on HEPA Filtera Media. Presented at the 22ndDOEINRC Nuclear Air Cleaning and tReatment Con-ference, August 24-27,1992, Denver, Colorado, 1992.

Hyder, M. L, "Effect of Organic Material on CarbonAging." InProceedings of the 20th DOE/NRCNuclearAir Cleaning Conference, Vol. 1, ed. M. W. First, pp. 560-570. NUREG/CP-0098, U.S. Nuclear Regulatory Com-mission, Washington, D.C., 1989.

Jacox, J. W., 'Review of Nuclear Air fteatment SystemRelated License Event Reports for Period 1985-1987."In Proceedings of the 28th DOEINRC NuclearAir Clean-ing Conference, VoL 1, ed. M. W. First, pp. 386-407.NUREG/CP-0098, U.S. Nuclear Regulatory Commis-sion, Washington, D.C., 1989.

Johnson, J. ., D. G. Beason, P. R. Smith, and W. S.Gregory, 'he Effect of Age on the Structural Integrityof HEPA Filters.' In Proceedings of the 20th DOEINRCNuclearAir Cleaning Conference, Vol. 1, ed. M. W. First,pp. 366-380. NUREG/CP-0098, U.S. Nuclear Regula-tory Commission, Washington, D.C., 1989.

Kratzke, R. T, 'Opening Comments of Session Chair-man Kratzke. In Proceedings of the 20th DOE/NRCNu-clearAir Cleaning Conference, Vol. 1, ed. M. W. First,pp. 247. NUREG/CP-0098, U.S. Nuclear RegulatoryCommission, Washington, D.C, 1989.

Liening, E. L, "Solving Materials Problems in CarbonBed Absorbers." Presented at Corrosion 91, The NACEAnnual Conference and Corrosion Show. Paper No. 165,National Association of Corrosion Engineers (NACE),Houston, Ibxas, 1991.

Moeller, D. W, 'Failures in Air-Monitoring, Air-Cleaning, and Ventilation Systems in Commercial Nu-clear Power Plants (Jan. 1, 1975-June 30, 1978).' Nu-clear Safety, Vol. 20, No. 2, 1979.

Moeller, D. W., and L-S. C. Sun, *Failures in Air-Cleaning, Air-Monitoring, and Ventilation Systems inCommercial Nuclear Power Plants, 1978-1981.' NuclearSafety, Vol.24, No. 3,1983.

Moeller, D. W., and J. P. Kotra, 'Commentary onNuclear Power Plant Control Room Habitability Includ-ing a Review of Related LERs (1981-1983).' In 18thDOE NuclearAirborne Waste Management andAirCleaning Conference: Proceedings, Vol.1, ed. MW. First,pp. 145-161. CONF-840806, U.S. Department of En-ergy, Washington, D.C., 1985.

Moeller, D. W., and L-S. C. Sun, 'Review of LERs Per-taining to Control Room Habitability.' Nuclear PlantJournal, November-December, 1987.

Norman, B, "The Effects of High Relative Humiditieson HEPA Filter Media.' In Proceedings of the 19th DOENRC NuclearAir Cleaning Conference, Vol. 1, ed. KW.First, pp. 300-31& NUREG CP-0086, U.S. NuclearRegulatory Commission, Washington, D.C, 1987.

OECD,Air Cleaning in Accident Situations, Report toCSNI by an NEA Group of Experts. Organization forEconomic Cooperation and Development, Paris,France, 1984.

Pasedag, W F., R. M. Blond, and M. W. Jankowski, Reg-ulatory Impact of Nuclear ReactorAccident Source TermAssumptions, For Comment. NUREG-0771, U.S. Nu-clear Regulatory Commission, Washington, D.C., 1981.

Pelletier, C. A, E. D. Barefoot, R. N. Hemphill, andJ. P. Frederickson, Long-Term Performance of CharcoalAbsorbers RemovingRadioiodine in Ventilation ExhaustAir. NP-534, Science Applications, Inc., Rockville,Maryland, 1978.

Ricketts, C I., V. Ruedinger, and J. G. Wilhelm,'HEPA-Filter Behavior Under High Humidity Air-flows.' In Proceedings of the 19th DOE NRC NuclearAirCleaning Conference, Vol. 1, ed. M.W First, pp. 319-350.NUREG CP-0086, U.S. Nuclear Regulatory Commis-sion, Washington, D.C., 1987.

Ricketts, C I., V. Ruedinger, and J. G. Wilhelm,"Aspects of the Increase in Pressure Drop of High-Efficiency Particulate Air Filters Under FogConditions." Nuclear Technology 90:50-65, 1990.

NUREG/CR-6029 8.2

References

Rogovin, M., and 0. T. Frampton, Jr., Volume II, Part 2,Three Mile Island, A Report to the Commissioners and tothe Public, Nuclear Regulatory Commission Special In-quiy Group. U.S. Nuclear Regulatory Commission,Washington D.C., 1980.

Ruedinger, V, C. L Ricketts, and J. G. Wilhelm, Limitsof HEPA-Filter Application Under High-HumidityConditions.' In 18th DOE NuclearAirborne Waste Man-agement and Air Cleaning Conference: Proceedings,Vol. 2, ed. MW. First, pp. 1058-1096. CONF-840806,U.S. Department of Energy, Washington, D.C., 1985.

Silberberg, M., J. A. Mitchell, R. 0. Meyer, and C. P.Ryder, Reassessment of the Technical Bases for Estimat-ing Source Terms. NUREG-0956, U.S. Nuclear Regula-tory Commission, Washington, D.C., 1986.

Soffer L., Revision of ReactorAccident Source Terms andImplications for NuclearAir CleaningRequirements.Presented at the 22nd DOE/NRC Nuclear Air Cleaningand 7i-eatment Conference, August 24-27, 1992, Denver,Colorado, 1992.

Sommer, R. R. and J. E. Otermat,A Review of LicenseeEvent Reports Relating to NuclearAir 7Teatment Systemsand Heating Ventilating, Air Conditioning Systems FiledDuring the Period 1988-1991. Presented at the 22ndDOEINRC Nuclear Air Cleaning and Mfeatment Con-ference, August 24-27, 1992, Denver, Colorado, 1992.

Summers, R. M., R. K. Cole Jr., E. A. Boucheron, M. K.Carmel, S. E. Dingman, and J. E. Kelly, MELCOR 1.8.0:A Computer Code for Nuclear Reactor Severe AccidentSource Term and RiskAssessmentAnalyses.NUREG/CR-5531, U.S. Nuclear Regulatory Commis-sion, Washington, D.C., 1991.

USAEC, USAEC Regulatory Guide 1.3;AssumptionsUsed For Evaluating the Potential Radiological Conse-quences of a Loss-of-CoolantAccidentfor Boiling WaterReactors. U.S. Atomic Energy Commission, Washing-ton, D.C., 1974a.

USAEC, USAEC Regulatory Guide 1.4; AssumptionsUsed for Evaluating the Potential Radiological Conse-quences of a Loss-of-CoolantAccident for PressurizedWater Reactors. U.S. Atomic Energy Commission,Washington, D.C., 1974b.

U.S. Department of Defense, Military Specification MIL-F-51079D; Filter Medium, Fire-Resistant, High-Efficiency.Washington, D.C., 1985.

U.S. Department of Defense, Military Specification MIL-F-51068F; Filters, Particulate (High-Efficiency Fire Resis-tant). Washington, D.C., 1986.

USNRC, USNRC Regulatory Guide 1.52, Design, Testing,and Maintenance Criteria for PostAccident Engineered-Safety-Feature Atmosphere Cleanup System Air Filtrationand Adsorption Units of Light-Water-Cooled NuclearPower Plants. U.S. Nuclear Regulatory Commission,Washington, D.C., 197&

USNRC, USNRC Regulatory Guide 1. 140; Design, Test-ing and Maintenance Criteria for Normal Ventilation Er-haust System Air Filtration and Adsorption Units of Light-Water-Cooled Nuclear Power Plants. U.S. NuclearRegulatory Commission, Washington, D.C., 1979.

USNRC, Severe Accident Risks: An Assessment for FiveUS. NuclearPowerPlants. NUREG-1150, U.S. NuclearRegulatory Commission, Washington, D.C., 1990.

USNRC, NuclearPlantAgingResearch. NUREG 1144,Rev. 2, U.S. Nuclear Regulatory Commission, Washing-ton, D.C., 1991.

Wichner, R.P., C. F Weber, S. A. Hodge, E. C. Beahm,and A. L. Wright, Noble Gas, Iodine, and Cesium Rans-port in a Postulated Loss of Decay Heat RemovalAcci-dent at Browns Feny. NUREG/CR-3617, Oak RidgeNational Laboratory, Oak Ridge, Tennessee, 1984.

Wilhelm, J. G., and H. Deuber, "The Removal of Short-Lived Iodine Isotopes." Chapter 3 in Treatment ofGaseous Effluents at Nuclear Facilities, eds. WR.A.Goossens, G. G. Eichholz, and D. W Tedder, pp. 67-101. Harwood Academic Publishers, New York, 1991.

8.3 NUREG/CR-6029

References

Wren, J. C., C J. Moore, A. C Vikis, and R. J. Fluke, AStudy of the Performance of Charcoal Filters UnderPost-LOCA Conditions. In Poceedings of the 20thDOE/NRC NuclearAir Cleaning Conference, Vol. 2, ed.M.W. First, pp. 786-799. NUREG/CP-O098, U.S. Nu-clear Regulatory Commission, Washington, D.C., 1989.

Wren, J. C, and C J. Moore, The Effect of Weatheringon Charcoal Filter Performance: L [be Adsorption andDesorption Behavior of Contaminants. Nuclear Tech-nology 94:242-251, 1991a.

Wren, J. C., and C J. Moore, The Effect of Weatheringon Charcoal Filter Performance: IL The Effect of Con-taminants on the CH3I Removal Efficiency of TEDACharcoal. Nuclear Technology 94:252-261, 1991b.

10 CFR 50, U.S. Nuclear Regulatory Commission,'Domestic Licensing of Production and UtilizationFacilities." U.S. Code of Federal Regulations, 1992.

NUREG/CR-6029 8.4

NUREG/CR-6029PNL-8594

Distribution

No. ofCopies

No. ofCopies

ONSITEOFFSITE

R. R. BellamyUS. Nuclear Regulatory CommissionRGN-1Washington, DC 20555

20 J. J. BurnsUS. Nuclear Regulatory CommissionOffice of Nuclear Regulatory ResearchMS: NS217BWashington, DC 20555

D. A. CasadaOak Ridge National LaboratoryP.O. Box 2009Oak Ridge, TN 37831-8038

M. W. FirstHarvard University655 Huntington AvenueBoston, MA 02115-9957

R. J. LofaroBrookhaven National LaboratoryBuilding 130Upton, Long Island, New York 11973

H. L MaglebyEG&G Idaho, Inc.P.O. Box 1625Idaho Falls, ID 83415-2406

2 DOE Richland Field Office

D. C. Langstaff (2)

47 Pacific Northwest Laboratory

R. P. Allen (10)D. E. BlahnikS. H. BushE. H. CarbaughM. E. CunninghamM. K DrostE. 1. EschbachJ. L EthridgeK R. HoopingarnerD. B. JarrellA. B. Johnson Jr.L. D. Kannberg1. S. LevyM. W. 1UgotkeR. L. MoffittW. C. MorganJ. V. Ramsdell Jr.W. K Winegardner (15)Publishing CoordinationTechnical Information (5)

M. VaginsU.S. Nuclear Regulatory CommissionOffice of Nuclear Regulatory ResearchMS: NS217BWashington, DC 20555

J. P. VoraU.S. Nuclear Regulatory CommissionOffice of Nuclear Regulatory ResearchMS: NS217BWashington, DC 20555

Distr.1

NRC FORM 335 U.S. NUCLEAR REGULATORY COMMISSION 1. REPORT NUMBER2-891 (A112ned by NRC. Add Vol., Supp., Rev..NRCM 112 end Addendumn Numbei, It anyJ13201,3202 BIBLIOGRAPHIC DATA SHEET NUREG/CR-6029

{See instnxctins on the reverse)PNL-8594

2. TITLE AND SUBTITLE Vol. 1

Aging Assessment of Nuclear Air-Treatment SystemHEPA Filters and Adsorbers 3. DATE REPORT PUBLISHED

MONTH YEAR

Phase I August 19934. FIN OR GRANT NUMBER

B29115. AUTHOR (S) 6. TYPE OF REPORT

Technical7. PERIOD COVERED 7Inclusiw Dvta

W. K. Winegardner

8. PERFORMING ORGANIZATION - NAME AND ADDRESS litfNRC. provide DVion, Offi, or Resion, U.S Nuckar Aegutatory Cernmiuioa lingt dcf :N addre itcoracor, providenuae and Hilin .dd s&)re

Pacific Northwest LaboratoryRichland, WA 99352

S. SPONSORING ORGANIZATION -NAME AND ADDRESS hRC. oVe -- ~a a if onrrctor, Provdr JAVC Div, Offier r epgon, t& t.S.aruguiatoqy ConCnin.- ailn g .fdd4=1s

Division of EngineeringOffice of Nuclear Regulatory ResearchU.S. Nuclear Regulatory CommissionWashington, DC 20555-0001

10. SUPPLEMENTARY NOTES

11. ABSTRACT Vo" wordserka

A Phase I aging assessment of high-efficiency particulate air (HEPA) filters and activated carbon gas adsorption units(adsorbers) was performed by the Pacific Northwest Laboratory (PNL) as part of the U.S. Nuclear Regulatory Com-mission's (NRC) Nuclear Plant Aging Research (NPAR) Program. Information concerning design features; failureexperience; aging mechanisms, effects, and stressors; and surveillance and monitoring methods for these key air-treatment system components was compiled. Over 1100 failures, or 12 percent of the filter installations, were reportedas part of a Department of Energy (DOE) survey. Investigators from other national laboratories have suggested thataging effects could have contributed to over 80 percent of these failures. Tensile strength tests on aged filter mediaspecimens indicated a decrease in strength. Filter aging mechanisms range from those associated with particle loadingto reactions that alter properties of sealants and gaskets. Low radioiodine decontamination factors associated with theThree Mile Island (TMI) accident were attributed to the premature aging of the carbon in the adsorbers. Mechanismsthat can lead to impaired adsorber performance include oxidation as well as the loss of potentially available active sitesas a result of the adsorption of pollutants. Stressors include heat, moisture, radiation, and airborne particles andcontaminants.

1Z KEY WORDS/DESCR.PTORS 1t aw*d erphrs iWwat Ossswsea In 1, h em ,por.j 13. AVAOLABILITY STATEMENT

IEPA filters UnlimitedPhase I aging 14.SECURITYCLASIFICATION

adsorbers (his page)

air-treatment system Unclassifieddesign feature (TsM Reprt)

failure experience Unclassifiedaging mechanisms 15. NUMBER OF PAGES

stressors16. PRICE

NRC FORM 335(2891

Federal Recycling Program

NUREG/CR-6029, Vol. 1 AGING ASSESSMENT OF NUCLEAR AIR-TREATMENT SYSTEMHEPA FILTERS AND ADSORBERS

AUGUST 1993

UNITED STATESNUCLEAR REGULATORY COMMISSION

WASHINGTON, D.C. 20555-0001

FIRST CLASS MAILPOSTAGE AND FEES PAID

USNRCPERMIT NO. G-67

OFFICIAL BUSINESSPENALTY FOR PRIVATE USE, $300