Regulatory Guide 5.58

Considerations for Establishing Traceability of Special Nuclear Material Accounting Measurements

ML13350A228
Person / Time
Issue date:	11/30/1978
Revision:	0
From:	NRC/OSD
To:
References
RG-5.058
Download: ML13350A228 (9)

C." Clt RIC. U.S. NUCLEAR REGULATORY COMMISSION November 19780*REGULATORY GUIDE," OFFICE OF STANDARDS DEVELOPMENTREGULATORY GUIDE 5.58CONSIDERATIONS FOR ESTABLISHING TRACEABILITY OF SPECIALNUCLEAR MATERIAL ACCOUNTING MEASUREMENTSA. INTRODUCTION assigned value3 is known relative to national stand-ards or nationally accepted measurement systems.Part 70, -Domestic Licensing of Special NuclearMaterial," of Title 10 of the Code of Federal Regu- This guide presents conditions and procedural ap-lations requires that for approval to possess and use proaches acceptable to the NRC staff for establishingmore than one effective kilogram of special nuclear and maintaining traceability of SNM cpntrol andmaterial (SNM)' the licensee must provide proper accounting measurements. No speci ethods willphysical security and an adequate material control be presented herein since the methio o- to be usedand accounting system. Section 70.51, "Material Bal- for any given measurement musThe tai-' d to theance, Inventory, and Records Requirements,- re- needs and peculiarities of t proc's matequires licensees to calculate material unaccounted for rial, reference standarW in .enon, and cir-(MUF) and the limit of error of the MUF value cumstances. Rationat "a p analytical fac-(LEMUF) following each physical inventory and to tors will be pr te copsp ration as to theircompare the LEMUF with prescribed standards. Sec- applicability to .ea. at hand.tion 70.58, "Fundamental Nuclear Material Con-trols," requires licensees to maintain a program for CUSSIONthe continuing determination of systematic and ran- o W i] , lddom measurement errors and for maintaining controlO of such errors within prescribed limits. Section , "asurements for control and accounting are70.57, "Measurement Control Program for Spe'% ed on a great variety of material types andNuclear Materials Control and Accounting,C o c entrattons, with a diversity of measurement pro-vides criteria for establishing and an j cedures, by a large number of licensees at all theacceptable measurement and control sys various industrial, research and development, andImplicit in the criteria stated in § 57I the academic facilities involved. A way of linking allrequirement of traceability of all SNM and these measurements and their uncertainties to thereqountingrementof traceablityo N nall San- NMS is necessary to achieve valid overall accounta-accounting measurements tTo this end all measurement systems must bement System (NMS) b t'kns of reference stand- b il e with end, all measurementards. Traceability me .' Ibility to relate indi- compatible with the NMS, and all measurementvidual measure,,s.. s to ational standards or results must be traceable to the appropriate nationalnationally acc ted m s ,' ent systems through an (primary) reference standards or Primary CertifiedunbrokeniRai f con arisons, and reference stand- Reference Materials (PCRMs). To obtain this neces-rdm-* device, or inssary compatibility for any given SNM measurementtask, secondary (intermediate, working) reference__ _ standards or Secondary Certified Reference Materials'For definitlns, see paragraphs 70.4(m) and (t) of 10 CFR (SCRMs) appropriate for each SNM type and meas-Part 70. urement system are nearly always required. Table I'The listed regulations do not apply to special nuclear defines the various types of reference materials.materials involved in the operation of a nuclear reactor, in wastedisposal operations. or as scaled sources. See paragraphs 3The term "value" includes instrumental response and other70.51(e). 70.57(b), and 70.58(a) of 10 CFR Pan 70. pertinent factors.USNRC REGULATORY GUIDES Comments should be sent to the Secretary of the Commission. U.S. NuclearRegulatory Commission, Washington. D.C. 205l66, Attention: Docketing andRegulatory Guides are Issued to describe and make available to the public Service Branch.methods acceptable to the NRC stil" of implementing specific parts of theCommission's regulations, to delineate techniques used by the staff in ev"lu- The guides are Issued in the following ten brand divisions:sting problems or postulated accidents. or to provide guidance toapplicants. Regulatory Guldes are not substitutes for regulations. and corn- 1. Power Reactors 6. Productspliance With them Is not requked. Methods and solutions different from those 2. Research and Test Reactors 7. Transportationset Out in the guides win be acceptable If they pwovide a basis for the findings 3. Fuels and Materials Facititles 8. Occupational Healthtequisite to the Issuance or continuance of a permit or license by the 4. Environmental and Siting 9. Antitrust end Financial ReviewCommission. 5. Materials and Ptanr Protection 10. GeneralComments and suggestions for Improvements In these guides are encouraged Requests for single copies of Issued guides whlich may be reproduced) or forat at times, and guides win be revised. as aMoprioate, to accommodate com- placement on an automatic distributlon list for single copies of future guidesmerts and to reflect new Information or experlence. Hower, comments on in specific divisions should be made In writing to the U.S. Nuclear Regulatorythis guide, if received within about two months after Its issuance, will be Commission, Washington, D.C. 20O6, Attention: ODrector. Division ofparticularly useful In evaluating the need for an early revision. Technical Information and Document Contro Traceability is a property of the overall measure-ment, including all Certified Reference Materials(CRMs), instruments, procedures, measurement con-ditions, techniques, and calculations employed. Eachcomponent of a measurement contributes to the un-certainty of the measurement result relative to theNMS. The NMS itself comprises a number of com-ponents, including Standard Reference Materials(SRMs) or PCRMs, national laboratories, calibrationfacilities, and standard-writing groups. If the NMS isviewed as an entity capable of making measurementswithout error, traceability can be defined as the abil-ity to relate any measurement made by a local station(e.g., licensee) to the "correct" value as measuredby the NMS. If it were possible for the NMS to makemeasurements on the same item or material as thelocal station, this relationship, and hencetraceability, could be directly obtained. Since theNMS is largely an intangible reference system, not afunctioning entity, such direct comparisons are notordinarily possible, and alternative means forachieving traceability must be employed. This neces-sary linkage of measurement results and their uncer-tainties to the NMS can be achieved by:a. Periodic measurements by the licensee of SRMsor PCRMs whose assigned values and uncertaintieshave been certified by the National Bureau of Stand-ards (NBS). These measurements may include inter-national reference materials whose assigned valueshave been approved and accepted by the NBS. Thisoption applies only if the materials to be measuredhave a substantially identical effect upon the meas-urement process as do the reference materials (RMs)or if the difference is relatively small and easilycorrectable by means of the known effects *of allinterfering parameters. Also, of course, the meas-urement of the RMs must be performed in a manneridentical to that employed for the SNM measurements(see Section B.3.1 of this guide).Table 1TYPES OF REFERENCE MATERIALSDefinitionRM Type andAbbreviationExamplesReference Material (RM)Certified Reference Material (CRM)Primary Certified ReferenceMaterial (PCRM)Secondary Certified ReferenceMaterial (SCRM)Working Reference Material (WRM)A general term that is recom-mended as a substitute for thatwhich previously has been re-ferred to as a standard or standardmaterial.A general term for any PCRM orSCRM or these materials as agroup.A stable material characterized,certified, and distributed by anational or international standardsbody.An RM characterized againstPCRMs, usually by several lab-oratories. Unlike PCRMs,SCRMs can be typical, somewhatless stable materials.An RM derived from CRMs orcharacterized against CRMs, usedto monitor measurement methods,to calibrate and test methods andequipment, and to train and testpersonnel.Any or all of the materials listedbelow.Any PCRM or SCRM or thesematerials as a group. See ex-amples below.Standard Reference Materials ofthe National Bureau of Standards(NBS SRMs) and Standard Mate-rials of the International AtomicEnergy Agency (IAEA) bearingthe IAEA classification, S.Reference Materials availablefrom New Brunswick Laboratory(NBL) or from IAEA. Thosefrom the latter bear the IAEAclassification, R.Process stream materials and anyRM prepared according to thisand related reports.5.58-2 b. Periodic measurements of well-characterizedprocess materials or synthesized artifacts that havebeen shown to be substantially stable and either (a)homogeneous or (b) having small variability ofknown limits. The uncertainties (relative to the NMS)associated with the values assigned to such processmaterials or artifacts are obtained by direct or indirectcomparisons with PCRMs or NBS SRMs.c. Periodic submission of samples for comparativemeasurement by a recognized facility having estab-lished traceability in the measurement involved,employing one or both of the above procedures, andinvolving only samples not subject to change in theirmeasured values during storage or transit.("Round-robin" sample exchanges between facilitiescan be useful in confirming or denying compatibilityof results, but such exchanges do not of themselvesconstitute the establishment or maintenance oftraceability.)Valid assignment of an uncertainty value to anymeasurement result demands a thorough knowledgeof all the observed or assigned uncertainties in themeasurement system, including an understanding ofthe nature of the sources of these uncertainties, notjust a statistical measure of their existence. It is notsufficient, for example. to derive a root-mean-squarevalue for a succession of observed or assigned un-certainties (CRM, instrumental, and procedural) forwhich standard deviation values have been calculatedby statistical methods for random events. To do soinvolves assumptions as to the randomness of thesevariances that may not be at all valid. The variancesmay, in faci. be due to a combination of systematicerrors that appear to be randomly distributed over thelong run but that are not at all random in theiroccurrence for a given analyst employing a givencombination of standards, tools, and instrumentalcomponents. Thus, it is necessary to derive the un-certainty value of a measurement from methods thatalso involve a summation of the nonrandom (sys-tematic) uncertainties, not from the mathematics ofrandom events alone. The valid determination of theuncertainty of a measurement relative to the N MS.and thus of the degree of traceability, is not arigorous procedure hut is the result of sound judg-mtent based on thorough knowledge aul understand-ing of all factors involved.Obviously, the sources of systematic error can bereduced if the Working Reference Materials (WRMs)are included at least once in every series of relatedmeasurements by a given analyst and combination oftools, instruments, and conditions. The calibrationand correlation factors so obtained cannot be applieduncritically to successive measurements. It also fol-lows that the applicability of any given RM to aseries of measurements of process material should beexamined critically both periodically and with everychange or hint of change in the measurement charac-teristics of the process material.It is doubtful that the WRMs can ever be exactrepresentations of the material under measurement inany given instance, even for highly controlled proc-ess materials, such as formed fuel pieces or uniformpowdered oxide, shown to be substantially uniform inboth composition and measurement-affecting physicalcharacteristics (e.g., density or shape for nonde-structive assay (NDA) measurements). However, inmost cases RMs that yield measurement uncertaintieswithin the selected limits for the material in questioncan be achieved. Obviously, the errors resulting frommismatch of the RM with the measure-d material willbe largest in heterogeneous matter such as wastematerials, but in these cases the SNM concentrationsnormally will be low and the allowable limits ofuncertainty correspondingly less stringent.The important truth being stressed here is thateveryn measurement ntist be considered, in all as-pects, as an individual determination subject to errorfrom a variety of sources, none of which mayuv besafely ignored. The all-too-natural tendency to treatsuccessive measurements as routine must be rigor-ously avoided. Physical RMs, in particular, tend tobe mistakenly accepted as true and unvarying: butthey may well be subject to changes in effectivevalue (measured response), as well as unrepresenta-tive of the samples. unless wisely, selected andcarefully handled.The characteristics required of CRMs include:a. Sufficiently small and known uncertainties inthe assigned values. (Normally, the uncertainties ofthe CRMs will contribute only a small fraction of thetotal uncertainty of the measurement.)b. Predictability in the response produced in themeasurement process. (Ideally, the measurementprocess will respond to the reference materials in thesame way as to the item or material to be measured.If there is a difference in measurement response tothe measured parameter arising from othermeasurement-affecting factors, these effects must beknown and quantifiable.)c. Adequate stability with respect to allmeasurement-affecting characteristics of the stand-ard. (This is necessary to avoid systematic errors dueto changes in such properties as density. concentra-tion, shape, and distribution.)d. Availability in quantities adequate for the in-tended applications.It cannot be assumed that RMs will always remainwholly stable as seen by the measurement systememployed, that working RMs will forever remainrepresentative of the measured materiel for whichthey were prepared or selected, or that the measuredmaterial itself will remain unchanged in its measure-05.58-3 ment characteristics. Therefore, it is essential thatthese RMs, as well as the measurement instrumenta-tion and procedures, be subject to a program ofcontinuing confirmation of traceability. Many of thefactors involved in such a program are discussed inReference I.L2. Mass and Volume MeasurementsThe national systems of mass and volume meas-*urements are so well established that RMs meetingthe above criteria arc readily available. Where neces-sary, the licensee can use the RMs to calibrateWRMs that more closely match the characteristicsof the measured material in terms of mass, shape, anddensity in the case of mass measurements or are moreeasily adapted to the calibration of volume-measurement equipment.Specific procedures for the use of mass and volumeRMs for the calibration of measurement processesand equipment arc given in the corresponding ANSIstandards (Refs. 2 and 3). Factors likely to affectuncertainty levels in inventory measurements of massand volume are discussed in other regulatory guides(Refs. 4, 5, and 6).3. Chemical Assay and Isotopic MeasurementsMethods for chemical analysis and isotopic meas-urement often are subject to systematic errors causedby the presence of interfering impurities, gross dif-ferences in t(ie concentrations of the measured com-ponent(s) or of measurement-affecting matrix mate-rials, and other compositional factors. Traceability inthese measurements can be obtained only if sucheffects are recognized and either are eliminated byadjustment of the RM (or sample) composition or, insome cases, are compensated for by secondary meas-urements of the measurement-affecting variable com-ponent(s) and corresponding correction of the meas-ured SNM value. The latter procedure involves addi-tional sources of uncertainty and therefore should beemployed only if it has a substantial economic ortime advantage, if the interferences or biasing effectsare small and limited in range, if the correctedmethod is reliable, and if the correction itself isverifiable and is regularly verified.3.1 National Standards -Uses and LimitationsNBS SRMs generally are not recommended foruse directly as WRMs, not only because of cost andrequired quantities but also because of differences incomposition (or isotopic ratios) compared to theprocess materials to be measured. NBS SRMs aremore often used to prepare synthesized intermediateRMs of composition and form matching the process'Regulatory guides under development on measurement con-trol progrdms for SNM accounting and on considerations fordetermining the systematic error and the random error of SNMaccounting measurements will also discuss the factors involvedin a program of continuing confirmation of traceability.material or to evaluate (and give traceability to)non-NBS but substantially identical material fromwhich matching WRMs are then prepared. This isnecessary because of both the wide diversity of proc-ess materials encountered and the very small numberand variety of SNM SRMs available. These inter-mediate RMs may be used directly as WRMs, ifappropriate, or may be reserved for less frequent usein the calibration of suitable synthetic or process-material WRMs of like characteristics, as well as forverifying instrumental response factors and other as-pects of the measurement system. However, eachlevel of subsidiary RMs adds another level of uncer-tainty to the overall uncertainty of the SNM meas-urement.SRMs can also be used to "spike" process sam-ples or WRMs to determine or verify the measurabil-ity of incremental changes at the working SNM level.However, because of possible "threshold" or "zeroerror" effects and/or nonlinearity or irregularity ofmeasurement response with concentration, this pro-cess does not of itself establish traceability.3.2 Working Reference MaterialsWRMs that closely match the effective compo-sition of process material, or a series of such WRMsthat encompass the full range of variation therein,serve as the traceability link in most chemical analy-sis and isotopic measurements. The WRMs derivetraceability through calibration relative to eitherSRMs or, more often, synthesized intermediateCRMs containing either SRMs or other materialevaluated relative to the SRM (see Section B:3.1 ofthis guide).The characteristics required of a WRM are that itbe chemically similar to the material to be measured(including interfering substances), that it be suffi-ciently stable to have a useful lifetime, and that ithave sufficiently low uncertainty in its assigned valueto meet the requirements of the measurement methodsand of the accountability limits of error.WRMs can be prepared (a) from process mate-rials characteristic of the material to be measured or(b) by synthesis using known quantities of pureSNM. The former method offers the advantage thatthe WRM will include all the properties that canaffect the measurement such as impurities, SNMconcentration level, and chemical and physical form;it suffers from the disadvantage that the assignedvalue is determined by analyses subject to uncertain-ties that must be ascertained. The latter method in-volves preparations using standard reference material(not usually economical unless small amounts areused) or SCRMs (see Section B.3.1) with the appro-priate combination of other materials to simulate thematerial to be measured. The advantages of the lattermethod include more accurate knowledge of the SNMI5.58-4 content and better control of other variables such asthe amount of impurities and the matrix composition.The chief disadvantage is that the synthesized RMmay not possess all the subtle measurement-affectingcharacteristics of the process material. Moreover, thepreparation of synthesized WRMs may be substan-tially more costly than the analysis of WRMs pre-pared fronh process material. Detailed procedures forpreparing plutonium and uranium WRMs are de-scribed in NRC reports (Refs. 7 and 8).The primary concern in the use of a WRM toestablish traceability in SNM measurements is thevalidity of the assigned value and its uncertain.y.Considerable care is necessary to ensure that theWRMs are prepared with a minimal increase in theuncertainty of the assigned value above that of theSRM upon which the WRM value is based. If theassigned value of a WRM is to be determined byanalysis, the use of more than one method of analysisis necessary to enhance confidence in the validity ofthe assigned value. The methods should respond dif-ferently to impurities and to other compositionalvariations. If the WRM has been synthesized fromstandard reference material or from intermediate ref-erence material, the composition and SNM contentcan be verified by subsequent analyses.The composition of a WRM can change withtime, e.g., changes in oxidation state, crystallineform, hydration, or adsorption. These changes andtheir effects on measurement are minimized by ap-propriate packaging and proper storage conditions.Additional assurance is attained by distributing pre-measured amounts of the material into individualpackets at the time of preparation, and these packetscan be appropriately sized so that the entire packet isused for a single calibration or test. Even among suchsubsamples there may be variability in SNM content,and this variability must he taken into account indetermining the uncertainty of the assigned value.3.3 Standard Laboratories and Sample Inter-changeTraceability of chemical assay and isotopic anal-ysis values also may be obtainable through compara-tive analyses of identical samples under parallel con-ditions. A comparative-measurement program maytake either or both of two forms:a. Periodic submission of process samples foranalysis by a recognized facility having demonstratedtraceability in the desired measurement.b. Interfacility interchange and measurement ofwell-characterized and representative materials withvalues assigned by a facility having demonstratedtraceability in the measurement.Round-robin programs in which representativesamples are analyzed by a number of laboratories donot establish traceability but can only indicate inter-laboratory agreement or differences, unless tracea-bility of one or more of the samples in a set has beenestablished as above.4. Nondestructive AssayNondestructive assay (NDA) measurementmethods are those that leave the measured materialunchanged (e.g., gamma emission methods) or withno significant change (e.g., neutron activation) rela-tive to its corresponding unmeasured state (Ref. i).NDA offers the advantages that the same RM or thesame sample can be measured repeatedly and yieldsvaluable data on sysiem uncertainties not otherwiseobtained, that the nmeasurcinent made does not con-sume process material ,Ad that measurements can bemade more frequently or in greater number, usuallyat a lesser unit cost than destructive chemicalmethods. These advantages often yield better processand inventory control and enhanced statistical signifi-cance in the measurement data. However, likechemical analytical methods, NDA methods havemany sources of interferences that may affect theiraccuracy and reliability.In nearly all NDA methods.' the integrity andtraceability of the measurements depend on the va-lidity of the RMs by which the NDA system iscalibrated. Calibrations generally are based onWRMs that are or are intended to be well-characterized and representative of the process mate-rial or items to be measured. While the matching ofRMs to process itoms, and consequent valid tracea-bility, is not difficult to achieve for homogeneousmaterials of substantially constant composition (e.g..alloys) having fixed size and shape (e.g.. machinedpieces), such ideal conditions are not obtained formost SNM measurements. Many of the materials anditems encountered are nonhomogeneous, noncon-forming in distribution, size, or shape, and highlyvariable in type of material and composition. In orderto ensure traceability of the measurement results tothe NMS, variations in the physical characteristicsand composition of process items and in their effectsupon the response of the NDA measurement systemmust be evaluated and carefully considered in theselection or design of WRMs and measurement pro-cedures (Refs. 9 and 10).WRMs usually (a) are prepared from process ma-terials that have been characterized by measurementmethods whose uncertainties have been ascertainedrelative to the NMS (i.e., are traceable) or (b) areartifacts synthesized from well-characterized mate-rials to replicate the process material." However,calibration of the NDA method by means of such-Absolute calorimetry of SNM of known chemical andisotopic composition is an exception.'The advantages stated for similarly derived WRMs (seeSection 8.3.2) also apply here.5.58-5 RMs does not automatically establish continuingtraceability of all process item measurement resultsobtained by that method. The effects of small varia-tions in the materials being assayed may lead tobiased results even when the WRM and the materialunder assay were obtained from nominally the sameprocess material. It therefore may be necessary either(a) to establishtraceability of process item measure-ment results by comparing the NDA measurement.results with those obtained by means of a reliablealternative measurement system of known traceabil-ity, e.g.. by total dissolution and chemical analysis(see Section B.4.1) or (b) to establish adequatesample characterization to permit the selection of asimilarly characterized WRM for method calibration(see Section B.4.2).4.1 Traceability Assay by a Second MethodAny NDA method would be of little practicaluse if every measurement also required a confirma-tory analysis. However, in cases in which there are anumber of items or. material samples of establishedsimilar characteristics, it is practical to establishtraceability for a series of measurements by means oftraceable second-method evaluations of an appro-priate proportion of randomly selected samples. If thecorrelation between the two methods is then found tobe consistent, tracedbility is established for all NDAmeasurements on that lot of SNM and on other highlysimilar material.For nominally uniform process or productionmaterial of which multiple subsamples can be ob-tained from a gross sample. the uniformity can bededuced from the distribution of the NDA measure-ment data. For thus characterized material, traceabil-ity can be established for all subsamples that ap-proximate the mean7 from the separate traceablesecond-method analysis of a few of the subsamples.Other like subsamples can then be selected as trace-able WRMs whose assigned values are related to theseparately analyzed subsamples through their re-spective NDA measurement results.For subsample populations exhibiting a range ofNDA values, especially where a destructivesecond-method analysis is used, the "twinning"method of sample selection may be employed. In thismethod, pairs of subsamples are matched by theirNDA measurement values, and the matches are con-firmed by NDA reruns. One member of each pair isevaluated by the traceable second-method analysis-the other member of that pair is then assigned the.value determined for its twin and may serve thereaf-ter as a traceable WRM for the measurement of thatprocess material by that NDA method.ISubsamples whose measured values markedly deviate fromthe mean (i.e.. "flyers") are not used for second-method* analysis or for WRMs.4.2 Characterization by a Second MethodIf the process items or materials being measuredare subject to non-SNM variations that affect theSNM measurement, it may be possible to employ oneor more additional methods of analysis to measurethese variations and thus to characterize process ma-terials in terms of such analysis results. If the sec-ondary analyses also are by an NDA method, theymay often be performed routinely with the SNMmeasurements. In many cases, the results of sec-ondary analyses may be used to derive simple cor-rections to the SNM measurement results. Correctionalso may be obtained and traceability preserved bythe judicious modification of RMs so as to incorpo-rate the same variable factors, i.e., so that they canproduce the same relative effects in the SNM andnon-SNM measurements as do the process vari-able(s).Alternatively, it may be advantageous to prepareWRMs that span the normal range of variability ofthe measurement-affecting non-SNM parameter(s)(and also the SNM-concept range, if appropriate).These standards can then be characterized on thebasis of their non-SNM measurement results or ofsome function(s) of SNM and non-SNM measurementresults and can be assigned a corresponding"characteristic figure. ' If this procedure can be car-ried out with adequate sensitivity and specificity rel-ative to the interfering factors, and within acceptablelimits of uncertainty, the process material can beroutinely characterized in like manner and the appro-priate WRM selected on the basis of such characteri-zation.5. Continuing Traceability AssuranceInitial or occasional demonstration that a laboratoryhas made measurements compatible with the NMS isnot sufficient to support a claim of traceability.Measurement processes are by their nature dynamic.They are vulnerable to small changes in the skill andcare with which they are performed. Deterioration inthe reliability of their measurement results can becaused by (a) changes in personnel performance, (b)deterioration in or the development of defects inRMs, instrumentation, or other devices, or (c) varia-tion in the environmental conditions under which themeasurements are performed. The techniques dis-cussed in preceding sections ensure traceability onlyif they are used *within a continuing program ofmeasurement control.C. REGULATORY POSITIONThe measurement control. program used by thelicensee should include provisions to ensure that in-dividual measurement results are traceable to thenational measurement system (NMS). RMs used toestablish traceability of measurement results to theNMS should have assigned values whose uncertain-ties are known relative to the NMS. To meet thiscondition, the licensee should maintain a continuing05.58-6 program for calibrating each measurement process.using RMs that meet the criteria in the followingparagraphs.1. Reference Materials1.1 The National Bureau of StandardsDevices, instruments, and materials calibrated orapproved by the NBS are acceptable RMs' for calilbrating either methods or WRMs. However, it is veryimportant that the licensee be able to demonstrate thatthe RMs are stable under the conditions for whichthey are used, that their validity has not been com-promised, and that they meet the accuracy require-ments of the intended applications.1.2 Secondary Certified Reference and Work-ing Reference MaterialsLower-order SCRMs or WRMs that have beenproduced by the licensee or by a commercial supplierare acceptable provided their uncertainties relative toPCRMs are known.A statement of uncertainty should be assigned toeach RM based on an evaluation of the uncertaintiesof the calibration process. The statement should con-tain both the standard deviation and the estimatedbounds of the systematic errors associated with theassigned value.1.2. / RAfs for Chemical arnd Isotopic Analyses.WRMs used for calibrating chemical assay andisotopic measurements may be prepared from stand-ard reference materials (SRMs) supplied by NBS orfrom other well-characterized materials available tothe industry. Such WRMs should be prepared underconditions that ensure high reliability and should bepackaged and stored in a way that eliminates anypotential for degradation of the WRM.The assigned values of WRMs prepared fromprocess materials should be determined by analysis,using two different methods whenever possible. Asufficient number of analyses should be done by bothmethods to allow a reliable estimate of the compo-nents of random variation that affect the measure-ment. If two methods are not available, as may be thecase for isotopic analysis, it is recommended that averification analysis be obtained from another lab-oratory.If WRMs are prepared from NBS SRMs or otherPCRMs, they should be analyzed to verify that themakeup value is correct. i.e., that no mistakes havebeen made in their preparation. For this verification,at least five samples should be analyzed, using themost reliable method available. Should the analyticalresults differ significantly from the makeup value."International RMs and reference material such as IAEA RN~sare included, if accepted by NBS.the WRM should not be used. Typical statistical andanalytical procedures acceptable to the NRC staff forpreparing WRMs are found in References 7 and 8.Storage and packaging of WRMs should followprocedures designed to minimize any changes likelyto affect the validity of the assigned values. When-ever practical, the WRM should be divided into smallmeasured quantities at the time of preparation, and thequantities should be of appropriate size so that eachentir, unit is used for a single calibration or calibra-tion test.1.2.2 Nondestructive Assay. RMs for NDA shouldbe prepared from well-characterized materials whoseSNM contents have been measured by methods thathave been calibrated with CRMs or from syntheticmaterials of known SNM content. The NDA RMsshould closely resemble in all key characteristics theprocess items to be measured by the system. Sincedestructive measurements ordinarily cannot be madeon NDA RMs in order to verify makeup, as requiredfor WRMs for chemical assay and isotopic analyses,RMs should be prepared in sets of at least three.using procedures that guard against errors common toall members of the set. The consistency of the NDAsystem response to all the RMs in the set provides abasis for judging the validity of the set of RMs. Ifone or more of the RMs in the set differs significantlyfrom the expected response. no RMs from that setshould be used. Statistical tests for this comparisoncan be found in References 7 and 8.The design and fabrication of the RMs shouldtake into account the measurement process parame-ters affecting the response of the system (Ref. 1).including:a. SNM content,b. Isotopic content,c. Matrix material,d. Density,e. Container material and dimensions,f. Self-absorption effects, andg. Absorption and moderation effects.Studies should be carried out in sufficient detailto identify the process item characteristics and thevariations of the characteristics that can cause sys-tematic error. The results of the studies should beused to establish reasonable bounds for the systematicerrors.NDA systems whose uncertainties relative to theNMS cannot be satisfactorily established directlythrough the calibration process should be tested bycomparative analysis. This test should be done byperiodically analyzing randomly selected processitems with the NDA system in question and byanother method with known uncertainty. The verifi-cation analysis can be done on samples obtained afterreduction of the entire item to a homogeneous form.In some cases, verification analysis by small-sample5.58-7 NDA- or by other NDA methods may be acceptable ifthe uncertainties of the verification method areknown relative to the NMS.2. Measurement AssuranceTable 2RECERTIFICATION OR REPLACEMENTINTERVALS FOR CRMsThe traceability of each measurement process tothe NMS should be maintained by a continuingprogram of measurement assurance. This programshould include planned periodic verifications of theassigned values of all RMs used for calibrations.2.1 VerifIcition of CalibrationsA formal program fixing the frequency at whichcalibrations and calibration checks are performedshould be established. The required frequencies arestrongly dependent on system sability and should bedetermined for each case by using historical perform-ance experience. Current performance of the meas-urement system based on measurement control pro-gram data may signal the need for more frequentverifications. Also, the effects of cL'angcs in processparameters such as composition of material or mate-rial flows should De evaluated when they occur todetermine the need for new calibrations.WRMs that are subject to deterioration should berecertified or replaced on a predetermined schedule.The frequency of recertification or replacementshould be based on performance history. If theintegrity of an RM is in doubt, it must be discardedor recalibrated.2.2 Recertification or Replacement of CRMsObjects, instruments, or materials calibrated byNBS or other authoritative laboratories and used asCRMs by the licensee should be monitored byintercomparisons with other CRMs to establish theircontinued validity. In any case, the values should beredetermined periodically according to Table 2.Test Objects and DevicesMassLengthVolumetric ProversThermometers andThermocouplesCalorimetric StandardsCertified Reference MaterialsPlutonium Metal(after unpacking)U 30s (after unpacking)Maximum Periods5 yr5 yr5 yr3 yr2 yr3 moI yr2.3 Interlaboratory Exchange ProgramsThe licensee should participate in interlaboratoryexchange programs when such programs are relevantto the types of measurements performed in hislaboratory. The data obtained through this participa-tion and other comparative measurement data (suchas shipper-receiver differences and'inventory verifi-cation analyses) should be used to substantiate theuncertainty statements of his measurements.When significant deviations in the rcsults of thecomparative measurements occur, indicating lack ofconsistency in measurements, the licensee shouldconduct an investigation. The investigation shouldidentify the cause of the inconsistency and, if thecause is within his organization, the licensee shouldinitiate corrective actions to remove the inconsis-tency. The investigation may involve a reevaluationof the measurement process and the CRMs to locatesources of bias or systematic error or a reevaluationof the measurement errors to determine if the stateduncertainties are correct.3. RecordsThe licensee should retain all records relevant tothe uncertainty of each measurement process for 5years. The records should include documents orcertificates of CRMs, the measurement and statisticaldata used for assigning values to WRMs, and thecalibration procedures used in preparing the WRMs..58-8r~-2f: .-. --

REFERENCES1. Regulatory Guide 5.11, "Nondestructive Assayof Special Nuclear Material Contained in Scrapand Waste" (1973).2. ANSI Standard N15.18, "Mass CalibrationTechniques for Nuclear Material Control,"American National Standards Institute, 1430Broadway, New York, New York (1975).3. ANSI Standard N15.19, "Volume CalibrationTechniques for Nuclear Material Control,"American National Standards Institute, 1430Broadway, New York, New York (1975).4. Regulatory Guide 5.25, "Design Considerationsfor Minimizing Residual Holdup of Special Nu-clear Material in Equipment for Wet ProcessOperations" (1974).5. Regulatory Guide 5.42, "Design Considerationsfor Minimizing Residual Holdup of Special Nu-clear Material in Equipment for Dry ProcessOperations" (1975).6. Regulatory Guide 5.48, "Design Consid-erations-Systems for Measuring the Mass ofLiquids" (1975).7. G. C. Swanson, S. F. Marsh, J. E. Rein, G. L.Tietjen, R. K. Zeigler, and G. R. Waterbury,"Preparation of Working Calibration and TestMaterials-Plutonium Nitrate Solution," NRCreport NUREG-01 18 (1977).8. S. S. Yamamura, F. W. Spraktes, J. M. BaldwinR. L. Hand, R. P. Lash, and J. P. Clark,"Preparation of Working Calibration and TestMaterials: Uranyl Nitrate Solution," NRC reportNUREG-0253 (1977).9. ANSI Standard N15.20, "Guide to CalibratingNondestructive Assay Systems," American Na-tional Standards Institute, 1430 Broadway, NewYork, New York (1975).10. Regulatory Guide 5.53, "Qualification, Calibra-tion, and Error Estimation Methods for Nondes-tructive Assay" (1975).0.5.58-9