TRACEABILITY

A uranyl. nitrate solution for use as a WorkinG Calibration and Test Hatertal (r:cn-t) was characterized, using a statistlcaHy designed procedure 1 to document traceability to National Bureau of Standards Reference Material (S~~960). A Reference· Cnlibra~ion and Test Hateri<il (!'.CTI-!) \~as prepnred frot:J SRH-960 uraniun metal to approximate the acid and uraniun concentration-of the WCTM. Thi~ solution was used in the characteriz~tion procedure. Details of pr~paring, handling, and packaging these solutions are covered. 1\;o outside laboratories, each having r.1easuret1ent· experti.se .using a different analytj.cal method, •;ere selected to measure both solutions according to the procedure for characterizing the WCTM. Two different methods were also used for the in-house characteriza tion work. All analytical results were tested for statistic~! agree m~nt before the WCIM concentration and limit 6f error values were calculated. A concentration value was determined \dth a relative limi.t of error (P.LE) of approximately 0.03% \~hich •;as better than the target RLE of 0.08%. The


DISCLAIMER
Portions of this document may be illegible in electronic image products.Images are produced from the best available original document. ..

••' . . (:
-~-This report was prepared as an account of work sponsored by the United States Government.Neither the United States nor the Depart• ment of Energy/Nuclear Regulatory Commission, nor any of their em• ployees, nor any of their contractors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness or use• fulness of any information, apparatus, product or process disclosure, or represents that its use would not infringe privately owned rights.Co!0~--??oSd-'J.. PORTIONS OF T~"'!C) r.llf:PORT ARE ILLEGIBLE.ff has be~n nr:m:!-:i-:r'~ from the best available COi11' to permit the broadest possible avail• ability. •

ESTABLISHING THE TRACEABILITY OF A URANYL NITRATE SOLUTION TO A STANDARD REFERENCE MATERIAL
C. H. Jackson J. P. Clark For Presentation At The American Nuclear Society Meeting May 15 -l7, 1978 1\;o outside laboratories, each having r.1easuret1ent• experti.se.using a different analytj.calmethod, •;ere selected to measure both solutions according to the procedure for characterizing the WCTM.Two different methods were also used for the in-house characterization work.All analytical results were tested for statistic~!agree-m~nt before the WCIM concentration and limit 6f error values were calculated.A concentration value was determined \dth a relative limi.t of error (P.LE) of approximately 0.03% \~hich •;as better than the target RLE of 0.08%.The use of this \;orl:ing l!'.ated.al.eliminates the expense of using SIDts to fulfill traceability requirements for uranium meas~re ments on this type materinl.Several years' supply of uranyl nitrate solut:i.onwith NBS tr.acf:abiU.tyv:as produced.The cost of this material ''as .lessthan 10% of an equal quantity of SRH-960 uranium metal.
be convenient to u~e, nnd 5. be economical.
The cost of t•:ES SR.':s prohi.bit.f;thci.rusc where large qu.1nti.t:l.es of ¥.'ork:l.ngstandard nrc re~uir.edfor routin~ analytical measur~Mcnt operations.Therefore, the second method was used in preparing a large supply of uranyl nitrate WCTM.The overall cost of procurement, .preparation,and chnracterizntion ~as considerably less than.the cost of prc~aring an equal q~antity of the standard from SRM-960.By characterizing, the desired reliability was obtained and J:ns traccabi.lity~:..1s established.
PP.f.J>ARATIO~l OF CALIBRATIO~ AND TEST HATERIALS T1.•o uranyl nitrate solutions \•~ere prepared as calibration and test materials as sh01vn in Figure 1.A large quantity of uranyl nitrate solution was prepared for cl1aracterizing and est3blisl1ing tracc:ahility to the standard reference material.This quantity would meet the future needs for a working calibration and test material.After.the specifications of the 'l-10r.kingstandard had been de tend.nccl,calcula ti.on s were Tilade to• deterrni.nehoH much uranjum and acid \Wuld he needed fo1: the RCTN.Then a small qu<!ntity of •uranyl nitrate solution was prepare~ for use as a re[erence standard to calibrate the measurement systems used in the characterization of• .the\•.'or1:ing standard.A mechanical ampul fi.llcr was used to rapidly aliq110t solution tn the ampuls, which ~o.•erc irr.r:!cd:latE:ly fla~ie scaled \J:Ith a coc::nercial atr.pul.se;.ler to prevent cveporat:l.on . .Special precautions were taken wh~n trnnsferr:lng the solution from the large poly bottles .to the ampuls.Care wa 5 taken to prevent d r i rrlnr.sol uti. on :!.~ the :~mpuJ.•necb; v1hich could dry when the n~ck was sealed and affe~t the solution concentratloh.
The packaginr.provided J.CO larce ampuls containing nhout-130 grams of ura~ium which ls equivalent to approxi.matel.y five units of: SK~!-960 uranium meta 1. • Hany smaller anpuls 1-1ere also filled.for characterizinG tl1e working standard in section 5.3.The desired limit of error is related to the end use of the I•~CT>! and is a fu:1ction of the accuracy and precision of the plant material to which the WCT1 applies.A relative li~it of error (RLE) of•les~ than 1 h of the.0.2S%RLE established for the plant natedal measurements \•.'as 'selected.The number of samples analyzed determines if satfsf?.ctory measurement precisions are attained.In clwracterizing the 1-:cn: by t\.JO methods, the limit of error is a function of the precisions of the two analytical methods used.
The results of the outside laboratories analyses are given in Appendix I ~ith the sta-t~stical trentt:Jcnts U.1;tcd <Jbove.\•.11C:n the \•:Ct:•l JT,can results of the• two lahoratori.es.weretested using the t-test of equality of me2ns, the difference in pc~ulation means was significant.• InvestJgation revealed that Laboratory II's analyscs.ofthe VCTM had hettcr precision th~n th~ hi~toricnl precision of th~ laboratory.Therefore, a pooled standard •deviation was calculated from historical data and characterization data and used instead of tl1e stan-• dard deviation of the \•:C'D-! and RC1}l results.This uas pernissible since precision data previously obtained for a nethod used for similar sample analyscs.canbe pooled with c~rrent •data to compute a better est~~te of the standard devintion with a• larger number of degrees of freedom.
The results. of the outside laboratories appeared to be positively biased from the makeup value and in-house analyses of tl1e reference ~tandnid.• However, no significant difference \•las observed between the laboratories' values for the working star~dard nfte1: correcting for the di f:fcrence bet\•1een the observed values and the reference value of the TI.CTH.These results confirmed the in-bouse assir;ned value of thr, \•lCT:1.
•'S the actual data used in c~lculating the concentration and associated standard deviation values for tl1e .1978RCTH.Several 500 ml ampuls of the HCT:•: \,'ere opened and six aliquots of each r;tandard \.;ere analyzed by two mctlwds as done pieviously.The analytical measurements and subsequent ~tatisticnl treatment of this work.aregiven in Appendix II.These analyses conHmed the nnl:.eupvalue of the 1973 RC'f:-1 :111d the value previously asstgne.d to the \•:CnL The ne\,' assigned value of the h'CTr•l ,,,a•s \•Ji.thin the conf:i.dcnceinterval established in the previous \•:otk.

Table I .
lations.1976 RCTI1 ~akeup value and as~ociated standard deviation calcu~ • • Ta'!>le III.. •Characterization data sur!l:riary generated hy three dif~ ferett laboratories, using ~wo diffcirent analyticnl chemistry methods, in two different year~, vhile concurrently analyzing.one of two different n.cnr' s.Each Dean has been corrected for the apparc~nt: bias observed in the anc:lysis of the l~CTH.
Data CoJ.Jected in ConfirmaUon l•.'ork From Outs)dc Lnborntori.c:s -.The Mean:(M) and Standard Deviation (S) of each set of re~ults were calculated using the Calct!latcd F ratios for both methods arc compared to tabulated F values at the 95~{ confidence level.