ML17341B652

Evaluation of GE & Sec Chemical Procedures for Post-Accident Analysis of Reactor Coolant Samples

ML17341B652
Person / Time
Site:	Fermi
Issue date:	11/30/1981
From:	IDAHO NATIONAL ENGINEERING & ENVIRONMENTAL LABORATORY
To:	Office of Nuclear Reactor Regulation
Shared Package
ML17341B653	List: ML17341B652
References
RTR-NUREG-0737, RTR-NUREG-737, TASK-2.B.3, TASK-TM TAC-44449, TAC-44478, NUDOCS 8204070694
Download: ML17341B652 (58)

Text

'444 P/ALUATiOtf OF CE Ait0 SiC C I HL'CAL PROC"--GL'RES FOR POSTACCiDc.."IT ANALYSIS OF REACTOR COOLANT)T SAR)PL S

tlovember i981 Prepar ed by Exxon HucIear 'Idaho

Comoany, Inc.

idaho ifaiiona1 Enaineerina L&cra".ory idaho Falls, idaho 3340

=or The HucIear Regulator y Commission

1.0 SMMRY AHD CONCLUSIONS

.CQHTEHTS 2 s0 BAC<GROUHD

~

~

o o

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

3.0 REgUIREiMENTS ANO c;VALUATION CRITERIA FOR THE CH.!IC"L ANALYSIS OF REACTOR COOLANT SAMPLES...............

6 3.1 Requirements.....................

~

~

c 3.2 Evaluation Criter ia....................

7 4e0 EIALUATIOH OF CiiEi1ICAL PROC DURES FOR ANALYSI OF POSTACCIDENT REACTOR COOLANT SAii!PLES.....................

14 4.1 SEC and GE Sample Collection, Recommended Analysis Methodology, and Chemical Procedure Evaluation Pracram

.. I<

4.2 Qemical Procedure Oescr ptions, Advantages/Oisadvantaaes and Evaluation Suamaries

. 'a 4.2.3.

Boron A

4.2.I..> nalysis ?rocedures Fluoraborate Selec ive Icn Electrode (FSIE)

Curcumin Specrropho.cmetric Plasma Spe

.roscopy Boranametry OigiChem Analyzer o= i'!anua1 i4farn Tito at icns

~

~

~

o

~

~

~

~

~

~

Ian C'nrcmatograchy

{.'C)

Carminic Acid Spectrophotcmetry Conductivi y of Boron Salutions Summary and Analysis -,ar Boron

.Analysis Pracedures 4.2.i.2 4.2.7..3 4.2.1.4 4.2.1.5 4.2.3,. 6 4.2.1.7 4.2.1..8 4.2.'1.9

~

~

i.ol

~

~

~

0 I7

~

~

'e 0

~

~

V 20

~

t Zv Zz

~

~

tl

~

~

~

o Z>

~

~

V 4.2.2 Chloride Analysi s Nethads 4.2.2.1 Ion Chromatography

( IC) 4.2.2.2 Specir-c Ion Elec.rade (SIE}

4.2.2.3 Turbidimetric, Color.me ric, Titrimetr and Spectrophotcmetric 4.2.2.4 Conductivity or Chloride Sa lutians 4.2.2.5 Senaary and Conclusions or Chloride Analysi s Procedures sci

~,

32

~

~

VI

~

~

3v4 0C

~

~

VV

~

~

V6 37 4.2.3 Dissolved Hydroaen and Oxygen 4.2.3.1 Gas Chromatography (GC) - Hydrogen Analysss 4.2.3.2 Gas Chramatagraphy, Yellow Springs Ana Dissolved Oxygen Analysis 4,2.3.3 Evaluation Summary of Dissolved Oxycen and Hydrocen Analysis Iyz

~ i Vl er

~

~

~3 0

~

V

~

~

0 37 y ~

~

'I j ~

4.2.4 Conductivity and pH 4.2.4.1 Conduc ivity 4 'r4o2 pH

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

4.2.4.3 Summary o$ Conduc ivity. and pH Anal Methods V

~

~

~

~

~

~

~

~

ysi s

~

~

~

<6

<6 48

~

~

w 3 5.0 R "-~i" IC"S 4a isla I l4i4

~

o

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

V O 2.

TABLES Stations with Proposed usage oT ScC and Gc. Postaccidenz Sampling System Sunxnary of Requirements for Postacciden Chemical Analys Reactor Coolant Samples

~

~

V is of

~

~

~

~

~

8 SEC and Gc'R ac.or Coolant Analysis Methodology

~

~

~

~

~

chemical Pnaiysis Procedures Considered by ScC and G="

~

~

~

~

~

V ~

,"-eatures of ?roposed Analytical Procedures

,=ea-.ures of Procosed Analytical ?rccedures o>>

d'or

"-oron Ch !or'ce

~

~

~

~

~

t l,4

~l i

I

.'4

'4

0 EYALUATION OF SEC AN0 GE ANALYTICALCHEMICAL PROCEOURES FOR POSTACCIDENT A!NLYSES OF REACTOR COOLANT SAR~LES

.. 3..0 SUMNRY AdD CONCLUSIONS Sumtary As a result of the Three Mile Is Regulatory Conmission (NRC) required to implement, by January 1,

1982, the reactor coo lant samples following an have proposed the use of postaccldent olied by Sent",y Equipment Corporation (CE).

land Unit 2 incident, the Nuclear license s of nuclear pow r plants capability to collect and analy"e accident.

A number of Iic nse s

sampling and analysis sys ems sup-(SEC) or 'ene. al Elec.ric Company Under a technical assi stance con ract to the

'fRC Exxon Nuclear

.dano

Company, Inc.

(ENICQ) evaluated the sam@le co:e ion and chemical analysis proc dures associated with the two sys's.

,n ob'-ive

he evaluation was to determine applicab',e procedures and to ident-',-.y

he most approoriat me hod.

one s udy irvolved a r view of the NRC requirements, the es abIishment of review

c. iteria, and the evaluac on o= the proposed analysis methods and tes.

da-a aga'nst

.he re.uire.:. nt" and evaluation cri.aria.

one most appropriate methods selected by - E.'CHICO

.To

.the r qu i ed chemical analysis of postaccident reac or cool'ant samples ar shown b-low.

Detailed descriptions, advantages, disadvantag s,

and/or d f',-

ciencies of the selec ed procedures are summarized in se tion 4.Z.

Also in section 4.2. is the same rinformation for other prcc dures proposed by SEC and GE.

I is worthy of note tnat a number of the otner procedures proposed are al so appropriate, as indicated; included below are only those de med most appropriate.

0 i.

Boron - Fluoroborate Selective Ion Electrode 2.

Chloride - Ion chromatography 3;

Dissolved Hydrooen - Gas Chroma.cgraphy 4.

Dissolved Oxygen - Oxygen Probe 5.

Conductivity - C"nduc ivity Cell 6.

pH - pH?robe

Although EHlCQ did not conduct tas.s to evaluate the suitability of any of the procedures; in EHICO's judgement, the laboratory tests par-fowmd by SEC and G"- are sufficient to provide a high degreo of assurance of the suitabi lity of the selected and the noted alternate procedures for analysis of accident reactor coolant samples.

For suitability tasting of additional analytical procedures, Ei!CO recommends that standard test matrix samples ba utiliz d to demons rat Qaiz. acceptability..

Standard matrix solutions similar to test solutions employed by SEC are recommended as they contain the most significant cora degradation products in cancantrations equal o or greater than those projected from an accident with a

Regulatory Guide 1.3 or 1.4 source arm.

Test solutions used by SEC consider

he ef acts of chemi-cals which might be added to the reac or coolant following an accid rt.

For chemical procedures that are to be used

=or the ana lys s of undi lutad reactor coo lant

samples, the following standard as-marr ix containing nonradioact-;ve species is recsrmandad.

Constituient Concentration' ocm)

Cs

+2

~ +3

. Ca Cl 8

Li HQ+

3 w+

+4 K

2"0 jo 5

l,o 2000 2

150 20

For chemical proc dures that are to be used or analysis o

dilu ed reac or coolant

samples, es" ing should be performed with a

standard matrix diluted by a volume equal to the dilution to be used in the proce-dure to be tested.

It is also recommended that the procedures and as-sociated instrumentation be tested in an induced gama radia.ion field which will yield a total absorbed dose of 10 rads per gram of re c-or coo 1 ant.

(

lm

\\ ~ ~

~ ~

2.0 SAC(GROUHO Fram studies of. the incident a

Thre Hile Island Uni 2 (TNI-2),

the need for improvement or the capability of licensees of nuclear power plants to determine plant conditions in a more timely manner was identi-fied.

Subsequently, the HRC

issued, for implementation by the licensees, specific requirem(ents in several areas for improvement of the capability.

In addition ta the development and implementation of the upgraded capabilities, the requirem4nts specified that the license s

should pr par and have available documentation of the capabilities ror a post-implementation evaluation or campl,iance.

Exxon Huc lear Idaho

Company, Inc.

(EHICO) was contr c.ed by

.h HRC's Oivis-'.on( of Licensing to provide technic I assis-.

nc

=or the eval-uation of the post-implementation documentation n

a nurber or ar as.

One area was "Pastaccident S~~pling Caaabili y",

Item II.3.3 of HUR-:G-07 7.

It perta ns to the ability of the licensees

.to obtain reac:ar coolant and containment atmos pnere samples and ta aralyze the samales

-or selected r adionucl ides and chemical spe ies und r acc-'dent carditions.

In order to fac iIitate the eva luat; on of the pos-- imp lemenzat ion documentation, EHICO was reques.ed to evaluate

.the applicao.'ity or the chemical and radiological analysis capabilities associated with two pas accident sampling sys ems proposed for use by several.

power plants.

(Table 1).

Tne two system vendors are Sentry Equipment Corporation (SEC) and C neral Elec ric Company (GE).

The initial plan called for EHICO to eva luate the SEC system only and to per.arm the evaluation in two *phases.

As ft was believed that current technoloay was suitable for radiolagical

analysis, the two

'phases were to be a brief summary report on the chemical analysis proce-dures and a more detailed report on both th chemical and radiological analysis prac dures.

However, due to manpower 'shor age at EHICO and an HRC request to 'incorporate the Gc.

system inta the evaluation, an alter-nate approach was td(en.

The alternate approach is ta: ') evaIuat and prepar a detailed report or th~nemicaI procedures or both

he SEC

TABLE 1 STATIONS MITH PROPOSEO USAGE OF SEC AIRO GE POSTACCIOEHT SANPLI IG SYSTEH KHERAL ELECTRIC SENTRY Brunswick 1/2 Hine Nile Point 1

Fi "patrick Oys er Creek Pilgrim 1 Ouane Arnold Honticel 1 o P ach Bot om 2/3 Oresden 1/2 quad Cities 1/2 Zion 1/2 Brogans Ferry 1/2/3 Salem 1

Xewaune Indian Point 2

Surry 1/2 North Anna 1/2 Palisades ard GE system and 2) evaluate and document

.he radiochemicai ana>ys's oroc dur s associat d with both sampling sys.ems Iat r.

.;ne d =" i d evaluation OT tiie chemIci al analysis procedures the top.c o

hl s

report, munich is limited to the analysis o7 reac or coolant Sa.ples.

(

~

I I,.

• >>'t i >>I

3.0 REgUIREK'lTS ANO P/ALUATION CRITERIA FOR iriE CHEHICAL AjiALYSIS QF REACTOR COOLANT SAMPLES 3.1-Reauirements To provide information fo" the assessment of care integrity, shut-down neutron adsarber concentration, and reactor coolant corrosion poten-tial; Iic nsees or applicants for Iic nses of nuclear power plan.s are required to es ablish a capabili y for the timely collection and chemical analysis of reactor coolant samples under accident conditions.

Per NURH-0737 the requir d chemical analyses for reac or coolant samples are boron (PMR only), chloride, and either tatal dissolved gases or hy-drogen; the measurement of dissolved oxygen is recommended in NUREG-0737 but not required.

Per Regulatory Guide 1.97 the me surement of dis-10 solved oxygen, pH, and boron in all plants is required.

NUREG-0737 also specified hat the analysis cauld be pe. farmed by omploying a cambinat'ion of pressurized/unpressurized, diluted/undi luted graa samples or in I ine monitoring methods.

However, for analyses pe. formed by inline me-hads, a capability ta collect backup gr ab samoles and to provide prccedures for their analysis is required.

Ip all

cases, th collec.ian of grA samples for analysis and the inline analysis must be able to be performed wi h or uithout the operation of an auxiliary reactar coolant sys em, e.o.,

letdown.

Mith the exc ption of the chloride analysis, the time allot ed or sampling and on site analysis of the samples is three (3) hours or less.

. Time allotted for the chloride analysis, which can be performed offsi e, depends on the type of reac ar coolant wat r and the numb r of barrie. s between the reactor coolant water and the primary containment system.

For plants midi seawater ar brackish reac.ar cao'lant water or with single-barrier, primar y cao lant containment

system, ch loride analyses are r quired within twenty-four.

(24) hours.

Far other plants the chlo-ride analysis is required within ninety-six (96) hours.

In addition, the licensees or applicants are required to consijder the radiological hazards associated with the sample colle ion.and analy-12 ses.

The assumatians of a

Regulatory Guide 1.3 or 1.4 source 13 term and radiation exposur limits" of five (5) rem to the whole body 6

y h

Hi or seventy-five (75) rem to the extremities of any individual are to be used in system design and selec ion of chemical analysis methods.

Las, the license s

or applicants are to provide provi sions

=or restric ing bac"ground radiation levels in the chemical analysis facility and for insuring the validity and accuracy of the sample analyses.

Tnese provisions include such things as samole shielding, adequate ventilation air and fil ration, proper sample disposal, sample line purging, roduc-tion of plate-out in sample lines, etc.

Tne requirem nts for post accident chemical analysis of reac.or coolant'~les are presented in Table 2.

~.2

="valua ion Criteria Tne objective of the present evaluation of poten-iai methods

-,or

.he chemical analysis of reactor coolant sanoles under "os: ';a..ad

=-cci-dent conditions is to determine appl icable pr".cedur es and:o identify the most appropriate procedure for each of the required ana Iyses.

Hany fac.ors were cons',dered in.he evalua.ion of the pro,"ose';,eth-ods.

Gbviously, c"mpliance with the requirements of sensitirity, accura-cy, r ang, analys",'s rre, radiological dose limita ions, and samole c"I-Iec.ion methods wer e eva luat d.

This included c

mp r sicns o=

.he advantages (lower radio loaical exposures) and di sadvan:ages (r duced sensitivity and accuracy}

of uti I izing di luted or very

-smaI I reactor coo lant samples ver sus larger undi lu ed reactor coo lant

.samples.

It also involved an estimation of the significanc o

chemic I

and r dio-Iogically<<induced inter erences.

Other fac ors which were considered are the c"mp'Iexity of ".he procedures and the applicability of the tech-nique to both accident and normal condition usage.

Due to the unavai labi Iity of informa.ion in a

nun er of ins.ances, fac ors not cons-',dered were specific design fea ur s of th two sarpling systems.

ExampIes are sampling locations, shield ng, sample I ine purg-

ing, sample validity, ventilation, etc.

%I TABLE 2 SUGARY OF REQUIREMENTS FOR POSTACCIDENT CHEMICAL At(ALYSIS OF R~c'CTOR COOLANT SAt":PLES

~

~

I

~

Analysis Capabi 1 ity Units Boron (Wn)

Chlor ide

( ppn)

Total scc/kg Dissolved Gases(-)

(Cl Hydrogen scc/k g Reauirement Range Accuracy(6)

Per cent Units 0-1000(>>

0<000(<).

0-20 0-2000 0-2000

+5 if >1000,

+10 if >0.5

+10 if >50

+10 i

>"0

+50 if <1000

&.05 if <0.5 ~ if <50

~5 if <50 Sampl igg method(j)

Inlinc Grab Samale Optional Reqiured Optional Required Qotional Required Qotional Reouired Analysis Location Onsite Off si.e Samole Col 'iect,ion and Analysis

~

Tine (hours)

Required Optional Optional (B)

Required Required 0ptional Optional Optional 2<(B) 96 Radio logical Exposure Limits (9)

(9)

(9) fg)

Notes')

A pressurized reactor coolant sample is not requirj;d ii the dissolved gases can be determined with an unpressur ized sample.

2)

The measurement of conductivity is not required in NUREG-0737 or Reg.

Cuide 1.97, Revision 2; houever, methods ta measur canduc ivity are proposed by Si'C and GE.

Accordingly, the measurenent of conductivity has be n includ d in th.is study.

TABLE 2 (Continued)

SUi IVORY OF RE/VIREi~cHTS FOR POSTACCI DEHT CHEa iICAL AiinLYSIS OF REACTOR COOLANT SAMPLES Aaalys-'s C'pability Unfts Oxygen (pram) pH un>vs Conductivity(2) pS/cm ll k

0 Re uirement Range Accur acy(>)

Per cent Units Sampl iso Method(")

In 1 ine Grab Samole Analys Location Onsit Offsite ca plo Co 1 1 eczi on and 'Ana lysis Time (hcurs)

Radi o 1 oci ca 1 Exposure I

'>r m

~ ml 4N 0-20

+'0 if )0.5

+0.05 if <0.5 Optional Requir d

Required Optional (a) 1-13 Not applicable

+ 0.3 if 5

>pH<9

+ 0.5 if 5

<pH>9 Optional Requir d

Requir ed Optional

~

3 (9.)

1-1000(5) 20 iNot Aopl icab 1 e Opt ona1

~~e'Q 3'eg~

~

'ptional

.(91

'(otes:

.(Ccntinued)

3) Boiling Mater Re c ors

4) Pressurized Mater Reactors

5) The required range

=or measurement of conduc ivi'ty was tak n

from refer nc 14.

i',

1

6) The designation of p rcent accur acy as

+

5 if

>1000 indicates the required accuracy i s

+

5 percent if the required measurement is gr ater than

~000 units.

The designation cf units accuracy as <<'0 if <1000 indicates that the required accuracy is

+ 50 measurement uni:s i= the required measurement is less than 1000 units.

lne required accuracies were taken from reference 7.

7) Analysis may be performed with either grab sampling or inline aanitoring methods.

However, 'or in line analysis methods the capability to collec and analyz backup grab samples is required.

The capabi 1ity to col lect and analyze at least cne sample per day for seven (7) days following the onset of the accident and at leas.

one samoie per week un.il th ac-cident no longer exis s is also required.

8) For nucle r power plants wnich utiiiz saawa r cr br 'ckish water as a

source of reactor coolant water cr chico have single-barr ie.,

re ctor coolant cGn7ainment sys em, the chloride analysis must be performed within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />; fcr other nuclear power plants, the required chlcr ide analysis time is 96 hour0.00111 days <br />0.0267 hours <br />1.587302e-4 weeks <br />3.6528e-5 months <br />'.

lne chloride analysis may be per-ormeo cffsite 9} The radiation exposures tc any individua 1 invo 1 veo in the cc]lection and analysis of reactor ecol n" samples under accident conditions may not exceed

"- r~a to the whole bcdy or 75 rem to the extremi ies.

r ~

10

~ '

~

The evaluation critiaria used in'the study were:

Analysis Time - As the time r quired for sample collection was not specified by SiC or G=", it was assumed sample coilec ion could be performed within one hour.

Accordingly, an. upper limit of two hours was allotted for sample analysis; chemical analysis proca"ures which required two hours or less for analysis were sa+ isfactory.

P..

Sensitivity,

Range, Accuracy - Chemical proc dures wh ch encom-

'assed the en ire measurement range with the required accuracy were considered adequate.

To cover the full range of measure-ment as

required, sample dilution me+hods wer consider d

satisfactory.

Radiological "xposure Limits, Samola Size - ?adio!ogical ex"o-s

~

sure to any individual is Iimtad to r~m to the whol>>

Body and 75 rem to tha extremit es during the collection and analysis of reactor coolant sano!as.

Under tho assumotion of Regulatory Guides. 1.3 or 1.4 releases of fission products to the raac."r

coolant, calculated dose rates from reac.or coo Iant samolas 14 are nominally '40 R/h/o at 10 cm with a

one hour decay.

<ain-t nanc of ra'dio!ogicaI axcsuras wi.hin ac ptable limits r-quiras the usage cf safety factors such as: shielding, dis-nc, exposure ti-z, sample di lution, very smal I undi lut d sama le,'nd/or inlinc moni+oring.

Tne chemical analysis proc dures, including di sso ived

gases, proposed by GE and SEC make use of inline monitor ing, very small undiluted

samples, and remote dilutions of the initial reac-or coolant samp1e..

The diluted reactor coolant samples are used for subsequent "hands-on" analysis.

With the axc p.ion of.he subsequent hands-on

analysis,

.his s.udy did not evaluat the radiologicaI hazards associated wi.h the above methods.

It was assumed that'dequa.e shielding and/or remote operation wou!d minimize radiological exposures to personnel.

In regard to tha subsequent analysis of dilu.ed r ac-"r coolant samples; on!y estimates of radiolooic I exposures could be made as they 1'

not only a function of the amount of reac or coolant in the

sample, but also depend on the techniques of the analyst and the design of the analytical facility.

0 The method which was established to lessen exposure is to limit the amount of reac or coo Iant in the sample taken for analysis to 0." ml.

The basis for this criterion is the knowledge that doses to the extremities will be the limiting factor for hands on chemical p".cedures..

For

example, calcu Iated exposures to the e"4remities,, using the above value of L40 R/h/g of r ac:or coolant at 10 cm, will exc ed the 75 rem Iimi by a factor or almost two for a two hour exposur to a L. ml sample.

It would require approximately five hours of continuous exposure to ex-c ed the dose limits =or a O.L ml sample.

it is reaIized the't the limitation to a O.L ml reac.or coolant sample size is c"n-servative as exposure time will, in reality, "e less than two hours and techniques to reduce the exposures will probabIy be employed.

However, to aIIow a suf, icient margin of sarety, 0.

ml reactor coolant sample was consi1de.ed an acceptable size sample in this szdy.

In a final evaluation or ace pt bIe sam-ple sizes, larger samples may be permissible, but all factors

, mus be considered.

Complexity, Routine/Accident Usage Two othe.

cr iteria which 1

were used are the compIexity of performance of the proc dures and the applicability of th procedures to both routine and accident condition usage.

The proc dures were assigned

low, medium, or high levels of c".mple"ity based primarily on the nurrbet and nature of mani pul ations invo lved in the procedure.

Procedures with applicability to both routine and accident con-ditions were consider d nore'atisfactory

.han procedures ap-plicable to accident conditions as the use of nonroutine proce-dures can create confusion and cause errors under accident conditions.

5.

&emicaI and Radiologically-Induced inter=erences

- Tne rele se of large quanities of both radioactive and nonradioac ive fis-sion produc.s will r sul in high radiaticn fields end chemiczl-19

I l

II 1

4 ly significant levels o

various ionic species in the reactor coolant.

Ho h the radiation and ionic sp cies can interfere wi'th the accuracy of chemical procedures used o analyze reactor coolant samples.

In the selection of an appropriate chemical

- analysis

method, these matrix effec s should be considered.

In this s.udy a

chemical proc dure was consider d unsatisfactory if the interfer nc s cause the accuracy of the procedure to

~wceed the required limits.

7ne evaluation included a

review of avai lable test data and prof ssicnal judgements based on past experienc s of personnel involved in the review.

4 i>>>>

l

4.0 EVALUATION OF CHEMICAL PROCUOURES FOR ANALYSIS OF POSTACCIDENT REACTOR COOLANT SAMPLES

\\

'I

', ~

hj In the evaluation of the applicability of chemical procedures

=or analysis of postaccident reac or coolant sampi s, ENICO s.udied the chem-is.ry of the procedures, compared their capabilities with NUREG-0737 requirements and the estab iished evaluation

criteria, and ranked the proc dures in order of appropriat ness.

Some of the proc dures ar simi-lar to ones used at the Idaho National Enaine ring Laboratory (INEL);

this experienc added to the data base.

Presented below in Section 4.1 is a summary of the sample coIIec ion and chmical analysis procedur g

oroposed by SEC and GE and a "ener I

outline of the testing program conduc-d by SEC and CE.

nis is followed by a presentation of ENICO's evaluation of the proc dures.

IncIuded are brief descriptions of the procedur methodo ioay and "he dvan aaes and/cr disadvantages of each procedure.

Las.,

the overai I eva luat ion of.

the individual proc dures are surmnarized, or a aiven type analysis.

II'hV 4.3.

SEC and GE Sample Collection.

Rec mmended An=lysis Method" loav and Chemical Procedure Evaluation Program Pathods or anal jsis of pos accident r

ac or coolant samoles pl o-19 posed by SEC and GE include inline mani orina and laboratory anyiysis of grab samples.

For inlinc monitoring, sample s re ms ar diverted either continuously or intermittently throuah inlinc sensors.

For laboratory. analysis, the reac-or coo Ianna arab samples are di luted inline before transfer to the laboratory or direc ly to an analytical instrument.

Either diluted liquid ar dissolved gas arab samples c n be obtained.

To'btain liquid samples, 1:1.00 dilutions of 0.1 ml reac.or coo lant samples are typically performed; larger initia I di lutions or secondary dilutions o-the initial dilution c n also be performed.

To'btain a dissolved gas grab sample; thirty (30) to seventy (70) milli-liter s of oressurized reac or coolant are isolated, 0he sample is depres-

surized, and the dissolved gases are puraed into a

gas hoidina chamber with an inert aas.

One milliliter ar laraer aiiquots of the diluted sample are analyzed following dilut',on wi h the inert aas to a

known pressure.

e

The chemical proc dures associated with the proposed methods are ei her conven ional or modificaticns of conventional chemical analysis procedur s.

A summary of the methodoloay, ircludina chemical anaIysis procedures, recommerded by SEC and suggested by CE is present d in TabIe 3.

Ho included in Tab.ie 3

are other methods detailed by SEC and GE,'his information is included in sac.ion 4.2.

TABLE 3 SEC AHO GE REACTOR COOLANT ANALYSES METHOOOLOGY ANALYSIS Boron SYSTEM VEi'IOCR G-SAMPLE 7(PE Grab Grab ANALYSIS F00 Fluoroborate Elec=rode Spec.rophotometr 'c (carminic acid)

Chloride SEC GE Inline, ar b

Ion Chromotoaraphy Turbid imetric Qi ssolved Hydroaen SEc

'.nline, grab Gas Chrcmotcarapny Gas Chromotcgra"hy Oi sso Ived SEC Oxygen GE Inline Crab.

YSI Oxyaen Analyzer Gas. Chromotograpny SEC GE Inlinc Grab pH prooe pH pamr Corduc iv "y SEC Inline Inlinc Conduc.ivi ty Ce I I Conduc" ivity Ce i I

4

,I

~

,44 I:,1 4

t:.,

\\

Kg 1

I 4 '4 I

4 ~

Chemical procedures recommended by SEC are the result of a develop-ment and tes ing program conducted by Nuclear Uti'lity Servic s

(IIUS) for SEC.

In the study.the recatnitended methads and'everal other chemical analysis methods were eva'luated in the laboratory to identify chemic:I interferences due to sample

matrices, to detemi inc operational charac-teristics of instrumentation, and to measure the sensitivities, rang s,

'nd accuracies of methods.

Employed in the study we. e simulated post-accident reactor caolant test samples.

They contained; in addition to the chemical species of inter est, high-yield, stable fission produc.s and appropriate concentrations of chemical additives anticipated ia be present in the reactor caolant fallowing an accident.

Ine s.udy did not include actual measurements of possible effects of high radiation fields on the procedures; however, it did include the results of a

survey of personnel wi h prior experienc in the analysis or samples with high radiation fields and a literature r eview of erfec"s of hicn irradiation on different materials.

Chemical procedur s

'sugges ed by GE, except those cainciden al

o he SEC tes d methods, are not the result of detai Ied laboratory : s.-

ing.

The only testing of the pracedures is rela".<<

to the effects of high irradiation of the sampIes.

The suggested proc dures were selected primariIy an he basis of simplicity, s.ability and availability of r-

agents, minimum radiation exposure, and likelihood or c.using can.amina-'ian pr oblens.

~]

'>'.j Idj

,~P, j

Ij 4.2 Chemical Procedure Oescriotions.

Advantages/Oisadvantaaes and Evaluation Suanari es In the selec.ion, recommendation, and/or sugges ion or chemica1 pracedures far'analysis of pas accident reactor coolant, SEC and GE con-sidered a total of'wenty-seven (27) pracedur s.

Tne chemical procedures considered by the two vendors ar shawn in Tab le 4; also noted in who table are known pracedures in use at the IllEL.

As many of the procedures are similar or identical,. they have been grouped

together, as appropri-

ate, in EHICD's evaluation of the proc dur s.

Presented in order below are he evalua ions of the boron, chloride, hydrooen, oxygen pH and con-duc ivity measur m nt proc dur s.

44 p

';C

'4p, Afl44 4 444~~

HgM ~P g 446

}

I TABLE 4 CHEHICAL AlQLYSIS ?ROC DURES COHSIDEREO BY SEC AHO GE 0

CHEH}CAL A)JALYSIS

~oroa Chl or",'da Hydrogen 0"ygan r}Lt

~ ~ ~

. - Conductivi.y VENDOR m

SEC SEC SEC SEC

SEC, GK SEC SEC SEC SEC SEC

SEC, C"-

CE GE, SEC GE SEC Gc GK SEC

SEC, GE PROCEDURE 7tuaranaret se'[ective iao eject".aae Curcumin Spectrophotometric Plasma Spec roscopy Boronmetry*

Oigi Chem Analyzer Hannitol Titration Carminic acid Spactrophotometric Ion chrcmotography>>

Manual s<arnitol Titrat cn>>

Conductivity of Bol Qn Solutions Ion Chromatography>>

Selective Ion Elec.rode.

bfecuric Nitr to Titration Thiocyanat Spec-ropnoto'etric Silver Chloride Colormetric Conduc.ivity o} Chloride Solut.cns Gas Chr orna.tography Gas Chromatography YSI O~yoen Probe pH Paper Conductivity pH Probe" Conduc".iv,i y C

11>>

5 Indicates a procedure that is in use at the idaho 'Chemical Proc ss ng Plant (ICPP) or the Loss of Fluid Test FaciIity (LOFT) at he it(EL.

4.2 1 r2oron Anal sis Procedures 4.2.1.1 Fluoroborate Selective ion Electrode (FSIE).

the FSIE chemical analysis procedure, the. boron content o=

a sample is determined by the measurwnent of the concentration of the.etrafluorcbo-.

rata ion.

In addition to the sensing elec ",ode, which contains a

mem-brana with a selective tetra luoroborate icn exchanger, a single 'unction reference electrode (KCe/saturated AgCe) and a

conventional millivolt meter with a relative millivoltmode are requir d.

}~

17

The procedur requires pr cise laboratory techni-ques; care must be exercised to add the reagents to th standards or samples in sequence and to perform measurements at prescribed times.

In the analysis proc dure a standard and a sampIe are analyzed simultanecus-Iy.

Initially, 1.0.ml of saturated sodium fluoride is added to 5.0 ml of the standard, and then 0.5 ml 10 N sulfuric acid is added (the sod',um fluoride and sulfuric acid converts boric acid to the t trafluorobor te ion).

With the addition of the acid to the standard, a timer is started; fiv minutes later the same reagents are added to 5.0 ml of a previously diluted sample.

At eight minutes the electrodes are inserted into the standard solution which is being s irred; at ten minutes the millivolt response is adjusted to correspond to a

specific value on a

pre-established calibration curve.

The millivolt response

=or the samole is recorded

.at fifteen minutes and related

.o the ppm) boron from -he cal ibra icn

'curve.

To minimize radiologic I hazards 1.0 ml samples and

standards can be analyzed by the aoove proc dure with the use o-cor-respondingly less sodium fluoride and sulfuric acid.

In addit on, the anaIysis can be performed by using only Oi3 ml of the original 5.0 ml or 1.0 ml of sample taken for analysis.

Tne.. analysis using 0.3 ml is perfcrn:ed statical ly in rricrodishes.

There are two types of calibraticn curves.

One

-,or the 5.0 ml and/or 1.0 ml samples analyzed by immersion of the electrcdes into a stirring solution, and one for the 0.3 ml samples analy" d by im-mersion of the electr odes in the microdishes.

Tne caI ibrat ion curves are es ablished using the same techniques employed fcr the samples; the calibration curves are valid only for the pair of elec:rodes used to establish them.

Calibration curves are es imated o

be valid fcr six months;

however, frequent use of the electrodes shortens their Iife.

Accordingly, routine checks of the calibration curves ar recommended to maintain their currentness.

Approximat Iy a total of one hour is re-

quired, to generate new calibration curves for both large and small samples.

h j

Htmercus Iabcratc. y tests were carried out wi.h simulated postaccident matrix samples tc identi y chemical interfarances to the FSIE proc dura.

Ho sample matr ix effe s wer observed when. the proc dure described above was followed.

The advantages of the proc dur ar its wide ma-surement range and accuracy, the small sample sizes required,:he Iac!< of chemical int r a. ancas, its adaptability to routine and accident condi-tion usage, and th short analy'sis time required.

'0 The main disadvantage of the procedur is its r la-tive complexity, which will nec ssit ta well-trained analysts and rra-quant usage of the proc dure bv he analys s in cr der to retain their familiarity with it.

Another I imitation of:he prcc dur, und r

tha assumption of a minimum initial sample dilution of 1.:~00, is th inabili-

"y oi the procedure to measure boron ',avels in highly radioac='.ve caac:cr coolant below fifty ppm.

However,;n CHICO's opinion, this is not a

serious limitation as under accident conditions the ccncarzra:;on cf boron in the reactor coolant should be much higner than fifty ppm;

and, if it isn', conf-;rmation hat boron levels ara fifty ppm or aocve suf icient. infcrmation to determine the ne d fcr suosaquent corr ctive actions The FSIE anal vs i s procedure has nc been

-.as-.ad with high radiation field samples;

however, CHICO does not believe. ir-radiation assccia ad with hiahly radioactive

.sar.plas will sicAi, ic ntly altar th applicability of the method.

4.2.1.2 Cur cumin Scectrochotometric.

one curcumin spectrc-photmetric boron analysis method is based on the me surament of a red-cc Iored

product, rosocyanine, formad by the react ion of boron and curcumin.

To par form the measurement the 1.0 ml di luted sample and standards, which ara analyz d ccncurrantly with the

sampIes, are mixed with 4.0 ml of curcumin; evaporated to dryness; di sso lved in 95 pare nt isoprcpyl aIcchol tc a total volume of 25 ml; ard transferred o

a I.O cm spec+rophotcmeter call.

In the spectrcphotometar, a

Ba sch and Lomb Spec rcnic 20 or equivalent, the percent transmi

.ance of:.".

s r.:pie and

standard are measured at 540 nm.

A calibration chart is pr pared from the standards, and the concentration of the boron in the sample is det r-mined fram the calibration chart.

The curcumin spectrophotometric proc dure was labo-ratory tes ed by S~C/HUS.

In addition to sample matr'ix erfe t studies using samples containing selected nonradioactive fission products and chemicals anticipated ta be present af.er an accident, expe. iments wer e performed to optimize the precision,

accuracy, and required analysis time.

)ne advantages of the procedure ar its wide mea-surement

range, its accuracy, the small sample size r=-"uir d, he lack of chemical interferences, its utility under accident and rau.:ne condi-

tions, and its relative simolicity.

The disadvantages of the procedur are the long analysis t-;mes required and the necessity

.o generate cali-bration curves at he same time sample analyses are pe. fored.

Tne later is considered

'a disadvantace as a significant amount of time cauld be'asted if a satisfactory calibration can not be obtained the firs-

". me.

Another limita ion of the procedure is the inability to measure leveis of baron levels in reactor wa.er below

w acm.

'.-'.awever, as noted

above, BICO does not cafis ider this a major limitation as required cor-rective action.can be made bayed on the knowledge of boron conc ntra-icns of twenty ppa or more.

The effects af high radia ion ields on the'roc-dure'ave not been'etermined.

In E,"tICO's judgement, the accuracy or sensitivity of the procedure would not be compromised; but this reeds to be can. irmed. before the procedure is used.

4.2.1.3 Plasma Saectroscaa The ana lysi s of boron by plasma spectroscopy is achieved by vaporization of the sample in a plasma jet and analysis of the atomic emission spectra which is generat d.

The boron resanance wavelength of either 249.7 or 249.8 nm is used.

Readout of the unknown is comoared to standards.

Five millitars of a diluted r eac ar coolant sample is required.

One milliliter of the sample and assaciated radioactivity is campletely vaporized and released; the other 0 20

'0

,our milliliters are collected in a waste container as condensed spray droplets.

Tne required analysis time is ifteen to thirty minutes.

No specific laboratory t s" ing details were pro-vided; however, it was indicat d that, limit d tes s

were per=or,.ed on simulated reac.or matrix solutions with satisfac.ory r producibility and accuracy.

Tne measurement range associated with the procedur also wa not provided; however, a lower detection.limit of less than 1

ppn boron is remrt d.

Mith this sensitivity and appropriate sample dilution, it appears the measurement r ange would be sufficient to cover the measure-ment range r quired.

Tne advantages of the prccedur are its appar nt simplici.y, time required

=or analysis, and small sample sizes.

"0 Tne disadvantages o-,

the procedure are "he lac!< of suff',cient laboratory. testing and the radioac iv i y l ale ses associated arith it. It is assumed that appropriate design modifications could be

ncorporat d to circumvent this latter deficiency;. but the design mus-include features to colle t all he radioactive r leases, not mer lv to contain them in a fumehood, as is done with th exis ing design.

H'gh radiation fields will not a-,=,ect the applicability of he proc cu.

4.2.1.4 Horonometrv.

The analysis of boron by boronometry is based an the at enuation of a collimat d neutron beam by a solution of boron between the sourc

'of neutrons and the detec:or.;ne neutr.n coun. rate.frcm the detector tube is conver:ed directly. to boron concen-tration on the readout electrometer or pulse counter.

Californium-2HZ or plutonium-beryllium ar typically used as sources of neutrons.

-"oron trifluoride (BF ) tubes or fission chaWers are:wo types of detec=ors.

A1though SF3 tubes have been reported to operate sati sfactor ily ir, galena-r ay fields up to 100 R/hr, later boronmeters use, ission chambers as they are virtually insensitive to gambia-r ay fields.

Boroncmeters typically employ relative ly large volume

samples, 1-2 liters or more.

Accordinoly, massive shielding o<

the sample station and separation of the sample station and r'eadout in-strumenation is required for accident'ondition usage.

As the de ec ors are 'sensitive to other sources of neutrons, location of the detectors within the plant should be considered, and the detec ors should be lo-cated away from these sources.

The sensitivitj of boronometers is on the order of 1 ppa boron with a useful range of 5000 ppn or more.

Calibration of boronometers can be performed s ati-cally or by flowing standards with a variety of boron core ntraiions past the detector.

Although no laboratory

.es ing was perfor-...e" on he e.fec s of sample matrices, no chemical interferenc s are anticipai d.

The advantages of bor on an alys i s by boronor;,etry are the continuous readout o< the boron conc ntration, the wide measurement rarge with or without sample dilutions, the applicability o;

he method

.to routine and accident,

use, and exis ence of proven boronomete. s.

0 The disadvantage of-the m thod is he use of Iarge volume samples,.

which could create maintenanc problems should a failure occur during an accident.

However, the impact of such an occurrence could be minimiz d as a

backup boron analysis capability us ng grao samples is required for inline monitoring methods.

As noted

above, high.radiation fields will not affect the performance of boron analysis performed by boronometry.

4.2.1.5 OioiChem Analyzer of Manual Mannitol Ti.rations.

The procedure for boron analysis using either the OigiChem Analyze.

or manual titrimetry methods is, in principle, the same.

Mannitol is added to the sample to form a

boron mannitol complex; hydrochlor ic acid is added to initiaIly adjust the pH of the solution to 4.4; and th sample 22

0

~%l is titrated to the end point

{pH 8.5) wi h sodium hydroxide.

Tne boron content of the, sample is derived fr cm the volume of scdium hydroxid titrant used and,comparision to standards data.

The difference in the two procedures is obvicus.

one employs hands-on techniques and the other employs remot analysis.

The r ate analysis is p rfcrmed. automatically with the OigiChem Ana-lyzer. It makes use of a microprocessor

=or sample and reagent dispens-ing solution mixing, and canc ntration measurements.

Tne analyzer auta-matica11y calculates the boron content and outputs it on a

computer-campatible tape.

Analysis by the analyz r can be performed ccntiruously, semicontinuously, or in the batch'made.

Separation of the sen'sing ele-ment and the readout device is required to eIiminate radiation effects on the sys em electronics; the sensor and electronics can be separated by at least twenty-f'.ve feet without degradation of the signal.~ne analy-si s

.ines ar seven minut s with t he automat ic ana lyzer

=-nd twenty minutes for the hands-on methods.

'0 A total of two hundred microarams of boron is re-quired for analysis with either the automatic or manual proced re.

nc.-

cardingly, the required sampIe siz s

depend on the conc ntraticn in:he sample.

For examale, under the assumption hat O.l ml of rg c.ar coolant is an upper limit for the reactor caolant sample

size, the ini ial con-centration of boron in the reac or coolant would have to be two thousand ppn or greater to provide sufficient boron for analysis.

Tne two thou-sahd pram repr sents the lower limit or detection

=or 0.1 ml samples and.,

as a result, precludes the usage of he hands-on mannitol titration pro-cwture usag on accident condition samples.

Hcwever, it. does not pr-elude

.he use of the OigiChem Analyzer for accident conditions as large.

samples can be collected and analyzed remotely.

In fact, the OigiChem Analyz r has be n laboratory tested on standards, with and without the presenc of paten 'ial inter-ferenc s;

accurate, pr ci se, interferenc free resul ts were obtained.

The measurement range of the prccedure for a 4.0 ml sample is 50-o000 ppn boron, which cauld be ex.ended dcwnward by the use of larger samples.

The advantage of the automatic mannitol titration procedure is its relative simplicity, remote operational charac e, is ics, utility under routine and accident conditions, and wide measurement'ange.

ine only apparent disadvantage of the procedur is the potential maintenance di'fficulty which might occur during replacem nt of sensing elements under accident condi ions or rapid repair of the microproc ssor.

However, as backup capabilities to analyze boron samples are r equired for inline sample

methods, the DigiChem analyzer should me t all measurement requirements.

The effects of high radiation fields have not be n

tes ed.

EHICO fe Is that th ef,ects probably will not be sign ficar,-';

however,

.his should be confirmed.

4.2.1.6 jon Chromatooraoh (lC).

Pn ion chromatocracn oper-ates on the principle of selective retention and elution of ionic s,"ec',es on and from ion excnange media.

It basically consists QT a

52paratol column and eluent,.

a supcressor

column, a conduct-me-r c

de-. ctor, and a

readout device.

To perform an analysis for anions, such as borates or chlorides, the sample is fir passed throuch the separator column -.an anion exchange medium which retains the anions and r plac s

them with another anion from he exchange medium.

The retained anions ar then selectively removed fr'cm the separator column with th

eluen, normal ly a dilute salt solution, and passed through the suppr ssor column.

jn the suppressor column -

a cation exchange medium - the anions ar converted to their acid forms which pass unretarded to.he conduc.imetric detector.

The conductivity of these dilate acid solu ions is a;unc.ion of the anion concentrations in the sample.

The time between sample injection and the appearanc of conductivity peak for a particular anion depends on the sample siz, the physical size of the columns, the types of exchange m dia, and

-he types, concentrations, and f1ow rate of the eluent.

As a r sul "if=er-ent anions in a single sample can be separated and analyz d

by prope.

selec ion o parameters..

In the development of an ion chromatographic proc-dure for the analysis o= boron and/or chloride; SFC/NUS s udied various ccmbinations of eluents, separator

columns, suppressor

columns, and sam-ple injection loop sizes.

Initial testing resulted in a

method

~hich used a scdium tetrabora.

eluent and was applicable for chloride analysis of pos accid nt reactor coolant samples

{c Se tian 4.2.2.~).

However, the analysis of boric acid solutions wi'th. the procedure shcwed inconsis-tent r esults.

Additional. developnent and'es ing by Oianex, the manufacturer of the ion chramotcgraph

used, resulted in a procedure 'or the simultaneous analysis of boron and chloride using a single sample.

In the test proaram a madi7ied Oionex madel '0;cn Chramataaraph was used.

Tne modifications included two 4 x ZGQ mm seaar=.

tor columns, a

3 x 250 mm suppressor column, a:wenty cm {0.043 mi.'am-ple injection

loop, and a

sadium caroonat

/sodium

~ydraxide/manni:oi e luent.

An additcnal requirement identified was the need af a ca,cn pre-column ta r move exc ss base and conve. t borates to boric acid "rior to loading highly basic samples into the inje tion 1ccp.

Wi"h

~

wen:y-five percent pump stmke, the necessary times for the baron and chloride paks to.appear following injec ion to the samp1ing laop are.respec-.;ve-ly "-o and 9-10 minutes.

To consistently obtain satis=actary

results, geri-odic washing and/or reaen r ation af the suppressar and pr -co lumns is necessary.

The pre-column requires regener ation after the ana lysi s af every two to three samples cantaining 0.4 M sodium hydroxide.

The re-quired frequency or washing and reaeneration o=

the supressar columns was not stated.

However, based on the frequency noted in the ini:al chloride analysis development work, estimat, d =r quency,or r generation is every 7our hours o7 continuous operation.

one ne d,or this is indi-cated by an erratic baseline on the readout device.

Tne required Fre-quency for washing the suppr ssor is anc daily or priar ta each regener ation.

If colutnn washing and regeneration ar not required, the analysis time is forty minutes.

If column washing and regeneration are required prior to analysis, the sample analysis time is approximately two. hours.

Neither.

case includes sys.em ca 1 ibration

time, wnich i s fif.een minutes.

The IC procedure for simulianeous chloride and boron analysis has be n laboratory tested using simulated pos>accident reac.or coolan't

samples, stable fission products,

caustic, cooling water impur i-

ties, and normal reac or coolant chemical additives.

No sample matrix effects were observed within the specified measurement range.

ine advantages of the proc dure ar its adaotability to remote operation, the large chloride measurem n-range,

.'he simplicity of operation, small sample sizes, potentially shori sample analysis

ime, and he Iac!< of chemic I inierferenc s.

tne disadvantages of a sufficient me surement r nge for boron, basic samples and the ne d

or co Iumn might lead to Iong anaIysis tim s.

the procedure ar the Iac!< of the need of a

pre-column or washes and r egeneratior which Tne effects of large irrad ations associat d with highly radioactive samples have not be n evaluat d.

Ho~ever, based on a

literature study of radiation effects on the components of the IC and on limited laboratory tests used to determine the effects of 0-200 ppn hy-drogen peroxide in

samples, no radiological effe -s are antic i oated.

The literature showed that cation resins begin to degrad at approxi-8 mately 10 rads and that the elec.ronic ccmponents are resis-ant to

~1 ex posure well above 10 'ads.

Both level s ar wel I above

.hose anticipat d to be encountered by the iC during analysis of samples, 4.2.1.7 Carminic Acid Spectrophotometry.

Two procedures wer presented for boron analysis with carminic acid, one by ScC and one by Gi.

The on presented by Gc, was detailed; it was developed by HACH 20 21 Chemical Company and closely fo I laws an ASTi<1 procedure.

The 25

5 It 5 I

.i' procedure presented by SEC was only an outline.

Sine both methods

were, similar and the HACH procedure had a slightly larger measurement

rance, only the HACH proc dure is discussed.

The HACH procedure is very simple.

Firs.

the carminic acid in pr weighed tablet form is add d to

?5 ml of sulfuric acid and mixed; then, 35 ml of the prepared solution is added to 2.0 mi of the

sample, blank, and/or s andard.

After the developnent of the color, 20-30 minutes, 25 ml of the solution(s) is transferred to spectro-photometric ce11s and the percent transmit ance is measured at 605 nm with a

8ausch and Lo& Spectronic 20 spectropnotmeter, or equivalent.

The measurement range is 0-L5 ppa boron without sample diiuticn and 0-several thousand pcm boron with samp'ie dilution.

The total analysis time is approxima ely 40 minutes.

The procedure has not been tes ea

-,or pcs-..ccicen-.

reac.or coolant sample chemical matrix effects; it nas be n -es:ed

=or effec.s of high sample radiation fields.

At the maxirr.

m antic'.pat d

source

term, 8 x 10 rad/h for a 0.1 mi reactor cooIant sample diluted to 25.0 'ml'he effec s of irradiation should be equivalent to no,".,or than 5

ppa boron.

7nis would result in negligible error when camp red to Ieveis of borcn in postaccident samples.

<<g

~ ~

The advantages of the procedure are the smai I sa'oie siz s required, the riide measurement

range, the adaotability to routine and accident conditions, and the simplicity.

The disadvantage is lack of labcra-'ory t sting with pastaccident,chemical matrix samples.

'V

<<:5 5

}~,<<p

(-'-

4.2.1.8 Conduc.ivitv of 8orcn Solutions.

A CE specification requires the Standby Liquid ControI System (SLCS) at 8~BR's to be,illed with a solution of borax and baric acid at a ratio of 1.028.

Cc oroposed

hat, in the event the SLCS were ac uat d, the borcn, conc ntration in the'reactor coolant could be estimated from conductivity.

6="tes ed the I

\\ v<<IPl5 4 + i

~

27

hypothesis with a 1.028 borax to boric acid solution by varying the boron conc ntrat-'on be+we n 5.4 and 201 ppn boron.

The calibration curve was linear betwe n.l0.8 and 201 ppn boron.

This suggests that, with sample dilution, the boron concentration of reactor coolant can be determined by conductimetric measurements.

However, EHICO believes that under accident condition there are too many other variables which could affec the conductivity of the reactor coolant and cause erroneous measurement.

Accordingly, the approach is not considered to be applicable for measurement of boron conc ntration in. reactor coolant.-

V 4.2.1.9 Summar and Conclusions for 2oron Analvsis procedures.

The results of ENRICO' valuation of potential chemical analysis procedures and methods for pos.accident reac-tor coolant sample boron analysis are surrnarized in Table

=.

Include are the measurement

ranges, sensitivities, accur acies, analysis t-'r es, sample

sizes, and analysis methods.

Also noted are the complexi:y of the procedures and tPe existence, based on actuaI tes.ing and/or profes-sional judgements, of known or anticipated chemical or radiological in-ter erenc s.

Finally, the applicabili y of the proc dure to routine nd accident condition use is indicated.

As all but one of he procedur s

met or exc ded the criteria for required sample

size, radiological exposures, measure-ment range and

accuracy, and analysis tines; the selection

'and ranking of the proc dures in order of applicability wer e based to a degre cn the complexity of the procedure and the labor atory testing which had been performed.

if two procedures had similar complexities or amounts of laboratory testing; other

factors, like time of analysis~

were con-sidered.

InIine analysis procedures were ranked lower than grab sampling procedures with similar qualifications as the capabi Iity to analyze backup grab samples is r quired fo" inline methods.

Last, anticipat d

maintenanc problems or potential contamination wer co'nsidered.

4 ~ ~

V P$

TABLE 5 FEATURES OF PROPOSED AHALYTICALPROCEDURES FOR BOROH bhthod Feature RAHGc, {ppn)

Oirec Analysis With l.:100 Dilution Mith 1:1000 Oilution

'Aith Other Oi lutions

>1:100 (2)

Accuracy (~)

{B in pptt)

Sample and/or Analysis Method?

online Grab Analytical Bac! "00 Ho Yes Ho 1.0-5.0 0.01-0.05 20 Medium Ho Yes Unknown Ho Ho Yes Yes Curcumin Spectophotometric 0,2 - 2.0 20-200 200-2000 20-6000 Ho Yes Ho 1.0 0.01 "120 Medium t/o Yes Unknown No Ho Yes Yes Plasma Spectroscopy 0-<1.0 0-<100 0-<1000 0-6000

+ZO(3) nlfo 0.25

-C. 006 "0

Low Unknown Limited Ho Unknown llo Ho Yes es 20

TABLE 5 (Continued)

FUTURES OF PROPOSED ANALYTICALPROCEDURES FOR BORON i&thod Feature e

=-

Carminic Acid Spectrophotome ric Fermi to 1 Titrimetry (Manual or Dioi-Chem. Analyzer)

Boronomet r Ion Chromatography

RAHt" (pea)

Dirac Analysis Aith 1:100 Dilution Mith 1:1000 Dilution Mith Other Dilutions

>1:100 (2)

Accuracy (~)

8 in ppn)

(Sample and/or

-,'nalysis Nthod?

Inl inc, Grab'nalytical Backup Required?

Samo le Co 1 lec.ion i>ethod Sample Analysis %thod Sample Siz (ml)

'iluted Ana1,vobis Sample Actual RC(>)

Ana lysis Time (min)

P'rocadure Complexity i

Chemical Intarferencas?

Tested Anticipated Radiological Eff cts?

Tested Anticipat d I

l 1 Application Routine Accident 0-10.0 10-100

,100-1000 0-6000

+15(3)

No Yes Ho 2.0

'.02

~0(6)

Unknown No No No Yes Yas Yes 50-6000 Not appropriate due to lack of sensitivity Yes No

Yes, For Inline Avai lab le Not Spec ied e.3(5) 1-2 No Yes Unknown No Ho Yes(8)

Yes O-5OOO(1) 0-500,000 0-5,000,000 0-6000 Yas Ho

Yes,

."-or In 1 inc Avai lab le Hot Scecified 1000-2000 10-20 continuous Unknown Ho Ho Ho yes Yes Yes Ko-6000 Hot appropriate due to lack of sensitivity Yas Ho 1lo 0.04 O.O~(5)

-'. 0-120 ( I )

%dium Ho Yes Unknown No Ho (o(g)

No Hot s:

t J jj 1

H

~

~')

The range of measuramants usino naut. on adsorpt'on is based on boron densitcmeters used at the Idaho National Engine ring L~>orator y.

"'0 t

=

l

'8 ~

2) 3)

Mith dilutions great r than 1:100 the upper '.limit of the measureman rance can be extended to ten-af-thousands af.

p~~.

Ho~ever an upper limit of 6000 ppn is noted as measur-ments above 6000 pea are not required.

In the procedure presented the uncertainty of the method was not included; based on professional judgement the urc rtainty.

has be n estimated at +20 pere nt.

The actual volume of, reac.or caoIant used in the analysis was determined frcm a 100.-fold dilutian of 0.3. ml of r ac ar caolant and the volume of diluted sample r quired

=or the ana lysis.

6) 7)

Oue to.a lack of sensitivi y far boron, typical sample dilu-tion of 1 00 of. Q.l ml reac or coolant samples is not aporo-priate.

Consequently, boron analysis of grab samples can not be made with the procedure.

However, the procedure nas suf-ficient sensitivity o analyze chloride in diluted rab sa--

ples (sa sec.ians 4.2.2.1 and 4.2.2.=).

Two p'ocedures were presented for boron analys; s with carminic acid, one by Sr":C and one by GE.

me araiysis specified by G" and SEC were 40 ard 90 m nutas, res"ec=-'.vely.

The difference in tir,es is the number of

.-.ir.utes l equi ".~d far cooling fo llowing carminic acid addition and

-.or ca icr deve loarent.

As CE had's ed the procedur and ScC hac no,

40 minu+es is assumed to be corre t.

The actual borcn analysis time is for+y (46) m'rue s.

How-ever during continucus operation a column wash/rege..eration/

equilibration cycle is required every four hours.

Acc r".in".-

ly, an analysis cauld require aaproximat ly twa haul s.

The manual mannitol tftrimetry s apprapiate

-,or

. ou+ ne usa only as the method lacks sensit-'vity to anaIyze small reactor coolant samples; the manual me.hod is cor,.only used at O'AR's under normal conditions.

r ne Oiai Chem Analyzer methcd i s applicaale "to routine or ac"ident ccndition usaae as the method uses r emote analysis of larger reac-or c"o lant samples.

9) The ion chromatagraph c

pracedur i s not aopropri ate r"r routine or accident candition usaae due ta insuf ic.'ent sen-sitivity. if 'the lower detec ion of GGO ppa boron were dem'd ta be sufficiently sensitive, the procedure would appropriate for accidert condition use only.

Listed in order of appropriat ness is the result of ENICQ'S evaluation of the boron analysis procedures:

1.

Fluoroborate Electrode 2.

OigiChem Analyzer Mannitol Titrimetry 3.

Curcumin Spec rophatometric 4.

8oronometer 5.

Carm'inic Acid Spectrophotometric 6.

Plasma Spectroscopy 7.

Ion.Chromatography It should be emphasiz d that the order of rankirg is based on presently available information only.

With additional

". s--

ing De order could change.

For example, with conf;rmation that there are no chemical interferences to the carminic acid spec.rophotcmetric method, it wauld be ranked at or near the top due to ease of use.

Like-wise, modifications to the plasma spec.roscopy instrument, wnich would insur containnent of volatiliz' radioactivity, would improve i

s rat-ing.

L-s, confirmation of the exis ence or nonexis.e..ce of radiiolcgic I

interfer nces could alter the order of ranking.

4.2.2 Chloride Analysis Methods 4.2.2.1 Ion Chramatoaraah v

IC j.

Oescribed in Sect ion 4.2.1.6 was an ion chromatographic pracedure

=or the simultaneous analy-sis of boron and chloride.

Included in the description were the columns; sample sizes,

eluent, and operational character is ics required for satis-factory analysis of boron and chloride in a single sample.

The measurement

range, accuracy, sample

size, and analysis tine for chloride analysis with the procedure are resp c ively 0.1-ZCGO

ppn,

+

10 percent, 0.04 ml of undiluted reac.or

coolant, and 40-120 minutes.

Tne procedure, which has be n

Iaboratcry tes ed, i s applicable for routine and accident condition use.

It can also be used as an inline monitor or =or analysis of grab samples.

32

The advantages oi the proc dure are Ne measurement range for chloride, normal and accident

usage, small sample siz, the

'ack of chemical in". rfarencas, remote operabili y, simplicity of op r-tion, and potentially short analysis tim.

Tne disadvantages of the procedure are the lack of a sufficien't measurement range for boron, the required column washes/

regenerations, wnich increase the analysis tim s, and the need of a Gl column for basic samples.

Another unknown is the lack of data on the potential effec.s o 'highly radioactive samples.

Sentry Equi paint Corporation al -o developed ard tasted another ion chroilaiographic procaduro'or chloride analys..'s.

procedure can not be used for boron analysis; hcwever, it is ve.y similar to the boron~hlor de analysis described previcusly.

Tna procedure uses a

3 x 250 mm separator

column, a

~ " 250 m suporessor

column, a

socium tatraborata

eluent, and a 0.04 ml Sample,.

The procaaure does not use a

pre-column.

To obtain sat siactory results the columns mus".

be washed and r egeneratad.

Mashing f"equency regenerat-'on.

is onc da-i ly or prior to

=- h Regeneration frequency is ore every four hours contiruous operation.

A nigh erratic basaIine, a change in the cf the appearance of the chloride peak (normally six minutes),

and/or a

change in the peak height for the s andard indicate a

need

=or.

rageneraticn.

Tne tetrabora.e EC proc d re has be n

~=. d in tNe laboratory with simulated samples of fissicn products and chemical addi-tives.

Spec-'aI laboratory t sts were performed to det rmine the e=f cps of morpholine, hydrazine, arii;onia, and natural and synthet c oils.

The only effec observed was due to oils, which caused a progressive 0-:0 percent increase in the chloride response and a maiory efiac

Hcwever, as the memory effec c n be eliminated with column washing and resene.

tion and as the increase in chloride peak height is associat a

with longer elution times, the effect is not considered signif'.c nt as i. c n

be detac.ad and corr ct d.

33

O'Moratory tests were also performed with the te>>

traborate IC proc dure to determine its ability to measure fluoride and iodide.

The data indicated

~hat the fluoride elution time was I:5

-minutes and that measurement of f1uoride is possible down to 25 ppa

(+ 0 percent) in the pr sence of fewer than 100 ppn boron.

Attempts to mea-sur e fluoride in the presence of higher concentrations of boron were unsuccessful due to peak overlap.

The iodide measurements indicate that iodide could not be detected at low concentrations (0.5 pram),

and at high concentrations"(up-to 100 ppn) small responses were observed.

lne iodide. data indicates that iodide will not interfere with the tetraborate IC chloride analysis method.

The*advantages and disadvantages of the tetrabora IC procedure are essentially the same as the ones pr sented above ror the boron-chlor.'de IC procedure.

4.2.2.2 Soecif ic Ion Electrode SIE).

inc procodure

-.or chloride'analysis by SIF is verv simple and rapid.

The pH of the solu-tion is ad-:us ed to Z~

and the SIF and a reference electrode are im-mersed in the solution and the millivolt response is related to the chlo-ride concentration.

The investigative s~udies performed by SEC/NUS em-ployed a Craphic Controls Ultra-Sensitive Sol',d S ate Chloride "-lectrode (Yodel PHI 9I100) and a Graphic Controls double-junction referenc elec-trode (No.

GC 54473).

In he proc dure i.0 ml of nitric acid was added to 100 ml of sample to adjust the pH.

The measurement range determ ned with standard chloride solutions was 0.01 o 35,000 ppn chloride.

'>lith the above measurement r ange, the S IF.

i s. ap-plicable to routine use only as approximat ly 10.0 ml of reac:or coolant sample, diluted to the 100 ml sample analysis

size, would be required to detect 0.1 ppn chloride.

Furthermore, the method suf ers from inter-

, er enc of other halogens.

The intorferenc prob lem possib ly c

n be solved by a combination of selective oxidation and solvent extrac ions;

however, at present the SIE is not applicable to postac ident chloride analysis due to the relatively large samole size required.

34

II g 'I Conc ivab ly, the Sie.

could be adapt d

tc remote

-operation;

but, as

noted, the chemical interferenc problem must be solved.

Overall, the method is not a good candidate.

l 4

S 7.

PS C r

~ S ~i1 4.2.2.3 Turbidimetric. Colorimetric. Titrine ric and Saectrochotm tr ic.

General Elec ric and SEC/HUS evaluated or suog s.

d a

nuprber of o.her candida e procedures for cnlo-ride analysis.

AI'I are basically hands-on methods;

however, one (titri-.

metry) could be

. adapt d

to remote inlinc ana lys is.

There has be n

limited or no laboratory testing of the procedures by S"-C/HUS or GE in regard.

to their applicability ta analysis of r actor ceo lant samples with potential fission praduc or chemical inter1.erences.

However, based on the judgement cf personnel at ICP.

Ewho have prior ex

e. ience w1th tie same problems on simi.ar procedures, iit is antic ipat d that iodid s and/

or other halogens will int rfere with all the procedur s

presented in

.his section.

Futhermore, due to the relatively large s ze reac.cr cool-ant samoles required or analysis, 2-=0 ml, use of the procedures

-.cr hands-on analysis is prohibited under accident condit"iong Acc rdi gly, the pracadures are not applicab le to analysis a

postaccicent samppes without fur her testing, modification, and development, cr withcu. rema.

use For informational

'pUrposes, each pr cc dur i s brierly outlined below.

'PE EES ELj Turbidimetric and Co Iorimetric Tne turbidimetric and co lorimetric procedures ar ver y similar.

Six drops of concentrated n1 ric acid ar added to:he

savpIe, 2 ml for colorimetric and 25 ml rar turbidimetr"'c; the pere n:

transmittance is recorded; seven drops of L

i'l si Iver ni:rat are add c; and the prec nt transmittance is recorded again.

The differ nce betwe n

the two recorded measurements is related to 'the canc ntration of chloride by the use of calibration s.andards.

For turbidimetry a

HACH Turbidi-meter. or equ valent is recopnttended, and for colorimetry a

Coleman i'~epho-CoIcrimeter, or equivalent 1s reccmm fld d.

p SE, E

E E

E

~

'l S '

I&

Soectrochotometric The spectrophotometric procedure presented is also

.simple and commonly used for chloride analysis.

It.involves the mixing'f

~0 ml of ferric amonium sulfate

solution, 5.0 ml of mercuric thio-cyanate methanol solution, and 25 ml of sample.

Tnis is,ollowed by the measurement of the percent transmittanc at 4o3 nm in a

10 cm spec.ro-photometric cell.

0 Titrimetr The titrimetry method is based on he formation of a

meraary

complex, diphenylcarbozone-brompheno I

b lue, and mecurous chloride.

The end-point color developnent occurs whenoarcurous ions are tn excess of the chloride.

In the procedure 25 ml of sample,

~-2 ml of diphenyl-carbo"one, and a

few drops of the bromphenol blue indica.or are mixed.

Tnis is followed by the addition of mecuric nitrate.

ine quanity of mercuric nitrate added is a func.ion of the chloride concentration.

4.2.2.4 Conductivity of Chloride 5o lutions..=or a

di lute solution of an. ionic species the specific conduc.anc, K,

in A/c.-., is given by:

K=10hC

~here A is the equivalent conduc ivi y and C

is the concentration of the ionic species in solution in electrolytic equi-valents.

Mhen the conduc ivity of a solution is due to s'everal ionic-

species, the specific conductance of the so lu.ion can be expressed as the sumation of the conductanc s of each of the separate ionic spe ies:

K ~

LO

$ (A..C )

where X

and C

are respectively the limiting equivalent ionic conductance and concentration of the individual species in solution.

Values, which are available in handbooks, of the equivalent conductance of different ionic species can be used to c IcuIate the con-ductivity or, aIternateIy, the conc ntration of the ionic species pro-vided the ionic species conc ntrations are known or the conduc-ivity of the solution is.known.

'I%

The proposed procedure u&lizes the above technique for estimation of upper limits of chloride concentration in. postaccident reactor.coolant samples.

CHICO agrees such a technique is applicable for estimation of upper limits of chloride or other ionic sp cies in solu-tion, hut does not believe the.echnique me ts the intent of the NRC r-quirement for chloride analysi s.

For

example, chloride concentrations calculated from the conductivity of postaccident soIu ion will, in all probabili y, be in excess of the 0.1 ppn limitation due o the presence of fission products, high radiation fields, and/or other ch micaIs.

As a result, corrective actions wiII be'aken or, more I ikely, accurate analysis of chloride concentrations will be made.

Initial accur te de-terminaticns will preclude undue concern and/or unnec ssary actions.

4.2.2.8 Suamar'nd Conclusion of Chlor;de 'nalvs s

Procedures.

At present the. e is cnly are apolicabi me"hod

=or chloride analysis of postacc'dent chloride analysis:

ion chromatography.

The other procedures evaluated.

are not appropriate due

.o tne large sample s iz s required and known or antic-.'ca ed chemic I

in.erfer nces to the pocedures.

The results and features of the oroce-dur s evaluat d are shown in Table 6.

The chem'ical procedures have roz een r:-need

.n order of applicability.

Of the methods rot presently app1 icab le, the specific ion elec. rode and the titrimetry methods aop r to have:he most potential due to adaptability to t emote use i ~e, reduc ion OF radiological exposures.

Their use, however, will depend on eIimination of chemical inter

e. ences, such as othe.

halogens.

Limi. d invest-'gative work was performed by SFC/HUS to eliminate the chemical interferences.

Their technique, which ENRICO believes has good potential, was selec.ive oxidation - solvent extract'.on.

Consequently, wi.h additional test-'ng and develop"ent one or more of these procedur s

could be adapted For postaccident use.

Specific procedures proposed in the Future will r e-quire evaluation as they become available.

4.2.3 Oissolved Hvdroa n

and Oxygen 4.2.3.1 Gas Chromatoaranhv GC) - Hvdrca n ".nalvsis.

A gas chromatograpn consists of a

sample injection

loop, a

hrcmatogr.phic 37

column containing a media such as charcoal or molecular sieves, a thermal conductivity call, and a meter-readout device.

The thermal'onductivity cell, or detector, has two, matched hot wire filaments.

Two streams of carrier gas, e.g.,

argon, are supplied to the GC fram a

coo+on sourco.

One, stream flows directly past one of the filaments; the other stre m

flows through the GC column then to the second hot wire filament.

In the absenc of a

sample, the two fi laments reach thermal equi Iibrium (cons ant resistance) and no detector output is observed.

Upon injection of a sample into the GC column, non~quilibrium between the two filaments created due to the different thermal conductivi ties of the gases eluted from the GC column to the sample stream filament.

one thermal conductivity imbalance generates a detector output.

As th'e different constituents of a sample are eluded from the GC column at different and specific times, the observed detector outputs can be attributed to the individual component of the gas sample.

The magnitude of the outputs are proportional to he conc ntratians or the dif erent gases in the sample.

guantification of the conc n rations is achieved by compari sion of the detector out~t of s amp Ies and standards.

The GC suggested by GE is a 3asel ine Model

>036, or equivalent.

Ine Model 1030 is a microprocessor cont. o Tied instrumen".

with thermal conductivity detectors.

It is equipped with a

gas condi-

tioner, an automatic retention-time indicator, and thermal conduc.ivity pe&~

integrator.

The sugg sted GC column is ten fe t -of 1./S to 3/15 inch tubing with

""A molecular sieves.

The carrier gas (helium) flowrat and pressure are 30 cm /minute and 15-30 ps ig; the suggested co lumn

.emperature was 30-50 C.

Although a Fisher ModeI 1200 Gas Chromatograph was used in the ScC/NUS developrent and testing

program, ScC/NUS also sug-gests Base line GC, or equivalent, for plant app I ications due to i ts larger measurement range.

Specific GC columns and operational paramete.

s were not given by SEC/NUS.

It is assumed he specifications will be similar to those not d by Gi.

Many combinations of columns, carr',er gas flowrates, and temperatures have been used success=ully in the past.

4 o

)

TABLE 6 FEATURES OF PRQPOSFD ANALYTICALPROCEDURES FQR C.'".'LORIOE Athod Feature Ion Chromatogr aphy Specific Ion Electrode TU3 bid ii7.etr 1 c Sample and/ar Ana'lysi s

)method Inlinc Gr b Analyt al Bac.<u p Required?

ample Col lection ivotho d ample Analysi s )vathcd SamDie Size (m) i)

Oiluted Analysis Sample

'c.ual RC Analysis Time (min)

Procedure Complexity Chemical Inter-,erences Tested Anticipated Radialagical Effec s?

Tested Anticipated Application Routine Accident RANGE (ppn)

( I 'i)eat An)1ySi)

Mith 1:100 Dilution Over al 1 Rang Accuracy (,.)

(Cl in pram) i-i i

0.1 - 100 10 - 10,QOO 0.1 - 10,000

+15(1)

Yes Yes Yes, For Inl inc Avai lab le Hat Speci i led W.Q"(6) v'O. Qg 40 120(8)

Medium No Yes Unkrawn No Ho Yes Yes 0.010 - 36,000 Nat Applicab'ie due to lac!<

af sensitivity

+ 20

.es VIII

~

4

Yes, For rn>>ne(')

Avai 1 ab le

.'lot Spe if 2'0(7)

Unknown Limit d Yes Unknown iso No Yes No

0(2)

Ho Yes bio 20 f7}

4ilkriawn No Yes VeI Yes(g)

Yes No 0.02 - '0 Nor. App', icab le due ta lac!<

of sensitivi"y

TABLE 6 (Continued)

FUTURES OF PROPOSED ANALYTICALPROC=-DURES FOR CHLORIDE o or>metric RANGE (ppn)

Oir ec Analysis Mith 1:100 Dilution Over all Range Accuracy (")

(Cl in ppa)

Sample and/or Analysis Method Inlinc Gr ab Analytical Backup R quired?

Sample Co Ilection bhthod Sample Analysis Method 0.04 - 10 Not Applicable due to lack of sensitivity

+ 25(2)

No Yes 0.1 - 10 Ho App1 icab le due to lack of sensitivity

+ 20(3)

Yes Yes

Yes, Fqr lnlinc< 5)

Avai'Iab Ie Hot Specified 0.02 -10.0 Hot Applicable due to lack of sensitivity

+ 20(3)

No Yes

'lo Sample Size (ml) 12 Qi luted Analysi s Sample '.4(")

Actual RC 100(8) 50 255(I )

Analysis T-me (min)

Procedure CcmpIexity QemicaI 1nterferences?

Tested Anticipated Radiological Effects?

Tested Anticipated Application Routine Accident 30 Unknown No Yes Unknown No Yes Yes No 20 Low Unknown No Yes.

Unknown No Ho Ho i'lo 20 Low Unknown Ho.

i'lo Unknown i'lo Yes Yes i'lo Notes:

1)

The accuracy of the IC measurements is

+ 15" in the 0.1 to 1.0 ppn chloride range and is

+

25Ã for higher concentra-tions.

By calibration at higher concentrations,:he ac ur cy can be maintained at + 15".

-".0

2)

The uncertainties

~era stimatad from calibration curve cata presented in the associated documentation.

3)

The unc rtainties are based on professional judgement.

4) 'he SLi method could be adapted

-or inline usa.

5)

The titrimetry procedure could be used as the inline method by amploymant of a technique similar to the OigiChem Analyzer method for boron analysis.

5)

Tne ion chromatogr aph ic proc dure uses smal I

(~

0. 4 ml )

undiluted r ac or coolant samples.

7)

Dua to insufficient sensit ivity, smaller reactor coolant samples are inappropriate fcr.hase me.hods.

s)

The titrimetry procedure has suffici nr. sansit-'vity to mea-sure 0.3.

ppn chloride; no~ever,:"0 ml c-reac or coolant is required.

Tne methcd is now in usa at LQ"'. a Li'ic,L.

Limi ed r adiolog'.cal effec

'es.ing was "e. formed oy 6:-

or.

the turbidimetric procedure.

At the ax

~ um anTic.oated dose rata, 8 " '0- rad/h in a di'u ad

"-".-I samo',e (Q.'

i Iutad to 2""ml),

an ecu iva Ianna res con sa of

~. ="

"wn chloride was calcu',atad from measuremenz casa.

The sample collection procedures fe dissolved gases.

proposed by ScC/NUS and g are similar. The GE procedure involves the iso-lation of 70 ml of pressurized r actor'.water, the depressurization of the sample into a

20 ml gas holding container, and the trans erral of ali-quots "from the gas holding container to 15 ml septum bottles.

one I5 ml septum bot Ies are transferred to the laboratory =or GC analysis and/ or further di lution.

In the labor atory, gas tight syringes are used o

take 1.0 ml aliquots from the septum bottles for injection into the GC.

The procedur4 employs Henry's Law and a tracer

gas, which is injec.ed into the sample prior to depressurization, to det rmine sample yield.

M The sample collection procedure of SEC/HUS involves the isolation of a 30 ml pressurized reactor water

sampIe, depressuriza-tion of the sample,. the quantitative trans,erral of the dissolved gases into a 300 ml gas holding cylinder via an ar gon gas

purge, and the pres-surization o

the 300 ml cylinder to atmospheric pressure with.he ar"on purge gas.

From the 300 ml gas holding cylinder small sampi s,

0.25 or.

1.0 ml are injec ed remotely into the GC or analysis.

Following collec.ion of -he dissolved gas samoIes the time requir d,for GC analysis is less than ten minutes.

4 4

It M

The measurement range reported by SEC/HUS is based on extensive laboratory studies and is applicable for 25-2:,GCO ppn hy-

'rogen or a 1.0 ml.dissqlved gas sample.

The dilutions.associated with the sample collection procedure and the 30 ml sample used

-or depres-3 sur ization creat a range of 0.5 - 2000 cn of dissolved hydrogen p r kilogram of re ctor coolant.

ine ac uracy of the measurements is

+

10 percent.

C~neral Electric did not, report a measurement range;

however, an estimate of the lower Iimi of detection was mentioned.

Their es imate of the lower detection limit, based on limited Iabora ory

studies, i's 0.1 volume percent or '000 ppn for a 1.0 ml dissolved gas sample.

EilICO believes this detection limit is a -actor of ten or more hig4 and that the actual de ection limit will be similar to the one

-.. a-sured by SEC/tiU~e.,

100 ppm or lower.

If such a detect'on limit is ver ified by GE, EHICQ estimates the measurement r ange of the GE gas d2

chromatocraph method for dissolved hydroaen in reactor water will be

&.1-2000 cm per kilogram of water.'ne estimate is based on the relative siz of reactor water samples taken for analysIs and the rela-tive volu1r s of the gas holding cylinders of the ScC/NUS and G~

sample collec ion systems.

4

~

0 Tne advantages of the GC iviethods procosed by the two vendors hre suffIcient measurerent

ranges, the application to rou-tine and accident

usaae, the simpIicity of operation, and the select1ve measurement of hydrogen, not total gases.

Tne advantag s of the S"-C/i(US method over the GE method are the ext nsiveness of laboratory

es ing per,ormed by SEC/HUS and the remote analysis capabili y of the ScC/iiUS 14 sys.em.

The lat er advantaae is quite sianificant as calculat d

dose rates due to noble gases associated ui.h the dissolved gases in 4

unit, volumes of reactor water are in exc ss of iO R/h at one centi-meter.

As a resul the dose rates associated wi.h the GC samo'.es, ven with di lution, are potenti al ly, a few R/h and wi I I require more caution

, or hands-on ana1ysi s than remote analysi s.

The disadvantage oi tile GC method in cenera I

'. s r clat d to maintenance of the ins rument; however, this is not considered sjanificant as CC's are ceneraIIy very Ceendable.

of th methcd is the lac!< of Iabcratory tests on radf ation fie let on ihe procedure;

however, there effects-Another I imitatior the eT ects of hicn are no ant ic. pated 4.2.3.2 Gas Qromatocrachv, Ye I low Sorinas Anal vzer Oissolved Ox aen Anal vsi s.

As described in the previous sec ion, di= erent constituents in a

gas sample are sear. t d in a

GC column due to.heir charac er istic diffusion rates throuch medium sucN as

charcoal, molecular

sieve, etc.

As a r suit GC lends itself to the simultaneous determination of oxygen and hydrocen;rcm th analysis of a single sample.

C ne. al elec ric proposed to use this

". chnique

=or described above.'peci fi I )

~ T

~ 'll,l

~~

I P.+of disso lved oxyaen ana lysis, i.e.,

simultaneous measurement of hydroaen and oxyaen in a single sample.

Tne sample c"llec.ion pro

dur, instru-mentation, and associa.

d equipnent proposed ar identical to the ones

~ ~

c measurement'anges were not provided bv C-".

43

ENRICO' estimate of the measurement range i s

~1 to 400 ppn in the reactor coolant.

The basis of the es imate are the relative thermal conductivi ties (detector responses) of oxygen and hydrogen'nd

.he hydrogen measurement range estimated for the C"-

system in the above section.

This estimated oxygen is'inadequa e for postacci-dent analysis of reac or coolant samples as the sensitivity of the pro-cedure is insufficient to measure below 1

ppm dissolved oxygen.

However, before the GC procedure is precluded

=rom pos accident applic tion,'t.

should be experimentally verified that the sensitivity of the GC method is inadequat Sentry Equi peen Cor por ation proposed an in 1 ine monitor =or postacc',d nt determinations of disso.ived oxygen in r actor coolant.

The ins rument se lected and Iaboratory '.ed was a 'e I low Springs Instrument (YSI)

Model 54 Oxygen AnaIyzer.

The sensing

probe, which contains a

semipermeable

membrane, is remotely located from

".he meter and output device.

The probe holder was redesigned to minimize fluid volume and associated radiation exposure.

Cai ioration of

".he system is achieved with an oxygen saturated demineralized water source.

The ac ual oxygen content.

Of the s andard solu ion is determined fr m the temperatures of the water and a solubility chart relating dissolved oxygen to xater temperature.

Laboratory tests verified that :here we. o no inter-ferences due to hydrogen in.solution or variations in samole flowraze.

One problem observed during the tests was a pin hole in one of the probe membranes.

inis resulted in erratic r suits.

Replac ment o"

the

-.em-brane correc.ed the problem.

Laboratory testing also verified that the accuracy

(+ 5X) was sufficient to measure 0.1 ppa dissolved oxyg n.

The me sure-ment range was linear betwe n 0.1-7.85 ppn oxygen.

Concentrations above 7.85 ppa oxygen were not laboratory tested.

It is anticipated the mea-surement range will be valid up to ZO ppn oxygen;

however, his ne ds verification.

Provided the measurement range can be demons razed to be 0.1-ZO

ppm, the YSI Analyzer is appl icab le to postacci dent

applicatiqn's-The ktodal 54 Analyzer lacks sufficient sensitivity for A

routine usa.

SEC/BUS proposed a

Rexnord Analyzer for routine usa (sensitivity-ppb) or alternately a Yodel 5o YSI Analyzer with reported highe~ sensitivity.

me rcutine condition monitor will be ins.al lad in parallel with tha'ccident condi ion monitor.

r ss

'I

%s

'4 t

5 The advan ages of the YSE oxygen monitors are the rema-.a operabi li y, simplicity, and potentially adequat ma sur.,ent range.

The disadvantage of the sys ams is the

time, i-4

hours, required for the system to reach equi I ibrium aft the internal porticns of the sensing probe are exposed tc air.

Based on a review of t.'re I-'tarature on the effac s of irradiation cn the ccmccnents of

'.he sensing

prcbe, no radiological effects ar antic'pat d.

me

~avi.-.

m dose anticipat d to the di,ferant ma arials of cons-ructicn in.ha "r"ba is

~0" rads; th minimum doses causing damage to

'he ma a i ls was racrtad at 10 rads.

0 4.2.3.3 Evaluation Sunmar'f gi sso lved 0" vaan and:~vdr-.can analvsis.

Tice gas chroma.ographic eethods procosed by SIC/hgg

-;o" dis-soived hydrogen analysis is a

plpic ba1e to postac-ident sar,"."le analyst it has suTTiciant sensitivity,

accuracy, and range of measuraman-.

{0.5-2GGQ cm hydrogen per kilogram of raac-.'or

coolant,

= i0..).

ine 3

rr asurement range associated with the Ci'method needs o

be verified.

After completing collection of the di sso'.ved gas

sample, the anal ysi s I

time is.an minutes or less.

ma SFC/HUS sample handling has an advan-tage over the K proc dura due to its remo-a mode of sanvle handling.

P~autions shculd be taken when manually handling the dissolved gas samoles due to the potentially high radiological fields.

The GC method is applicabla to routine condit ons also.

Radiological inter faranc s to the GC method are not anticipated.

t The GC method proposed by analysis appears to lack sutficient sansitivi y of Icw

{<0.1 ppm) concantT ation of cxygen.

to demonstrate th capability of the method

.o

~ation, the GC method or oxygen analysis is accident sa,pie analysis; GF

=or dissolved oxygen for required rraasuramant

'without Tur her tasting m asura tha lcw conc n-not app11cab le to post

The inline oxygen monitor proposed by SiC is appli-cable to postacci dent analys i s provided information i s avai Iab le to verify its abiIity to measure dissolved oxygen over the entire rang of O.l-~c ppm.

At present the measurement

. range has been demonstrated to be va'lid betwe n 0.1 and 7.85 ppm only.

In EHICO's opinion, the proven measurement range is sufficient as the intent is to measure the absence of oxygen, not nec s-sariTy the presenc If HRC does not

agre, additional laboratory studies ne d to be performed to extend the measur ment r ange.

To meet a11 NUREG-0737 requirements, license s wnich use inIine method for analysis must have a

backup capability to ootain grab samples and to perform analysis performed by the in

< ine;-,onizcr.

Oissolved oxygen analysis by hands-on tachiques will requir dilut d, pressurized samples or techniques to collect tne gases 1,'oxygen)

~

m

~ 'iquid sample.

Conventional

methods, eg., Minkler, of hands-cn anaIys s

can not be used due to the large sample sizes required and/or a

lack af sensitivity.

Alt rnative methods must be identif-;ed.

4;2.4 Conductivit and oH 4.2..4.5 Conductivitv.

Soth Si=.C/>>US and GE pl Qpose he use of inline monitors for measurement of condisctivity.

iheir proposed con-ductivity meters have measurement ranges of 0-F00

.2/cm and 0-CO gS/ca, respectively.

The proposed probes have conductivi y c Ils 'with I

O.l c3 cell constants; they are located remote to the meters.

in inline probe tested by C=

was a standard, commercially available

one, and:he probe tes ed by NUS used a mdified probe holder designed to minimize fIuid voIume..

Tne ac.ual cell volumes were not specified.

The ac"uracy associated wi> the measurements was not specified 'either;

however, high accuracy for O-Z gS/cm and decreasing accuracy for higher conductivi-ties ms noted.

laboratory tests were performed by both Gc, and SiC/

HUS.

The G= tes s

involved measurements of the conduc ivity of wat flowing first through a

conduc ivity eel I under irradi at ion and then through a second in-ser ies conductivity cell not under irradiation..ne irradiation fields were var ied betwe n

L.3 x

~0 rads/h to 9.3 x '0 46

'P rads/h.

CE also made static (no flow} measurements with the above ar-

. rangam nt.

Finally, GE per-.crmad ccnductivi y measurements on a

10 ppn

'chIor ide solution with (9.8 x

10 rads/h) and withou irradiation and wiD and withcut flow'through the calls.

one SEC/HUS laboratory test na was limited to establishing the operability with a fIowing sample s.r m,

the ef ec.s air bubbles in the air s ream, and the effects of hydrogen peroxide on conductivity measuremants.

SEC/t)US also conducted a

I itare-ture review for

. potential radiation effects to the components o

tha sensing prcbe.

)~

The results or CE tests on the two in-series cells indicated that the call under irradiation and the one not under irradi-tion gave the same results and that the conduc v'.ty cf the soluticn

',ncraasad frcm 0.1

~P/cm to 0.65 gZ/cn as tha irr ad ation n:ans-ty was increased.

ine causa, or the increase in ccnductirity is unknown; however, i" obviously is du to the generation or an unknown c"nduc.iva species.

The hypothesis that the unknc.vn.species is hydrcoen

-e. Ox.'"e is not supported by the chlcride solution tests pa. formed by GE and:he hydrogen paroxid t sts perfor...ed by SEC/NUS; i

the conduct'.vasty did not change iith tha addition of chlor de -

added to decrease "ha cena-ratad species

- or wi.h the addition of hydrogen peroxide dirac;iy to the flowing stream.

p

~

Further results of the SECINUS tas s

shcw that:ha monitor is applicable to a flowing sample stream and that tha pr sane o, air bubbIes at five pere nt or the water volume dces no alter the accuracy of tha me suraments.

The literature review indicates that the resistance or tha probe component" c radiation exposure axc ds

.ha an icipated radiation does by a fac.or of one hundred or more.

The advantacas of.he mathcd ar its utility under accident and normal condition, its r sistanca zo radiation

damace, the r mote operational

mode, and its simplicity.

Tnere are no apparent disadvan.ages even though the conduc ivity of water solutions increased wi.h incr asing doses.

Tnis cbsarvat:on only impl as t!iat the mcnitcr was 47 r diat cn

>pnrg t'ng

proper ly as its response incr'eased with an increase in the conductivity of the solution.

A backup method for measurement of conductivity'f grab

'samples was not noted.

However, there are commercial ly available portable conductivity meters which are appropriat for this purpose.

4.2.4.2 pH.

To investigate methods of determining pH under post'accident conditions, SEC/HUS evaluated an industrial grade inline pH probe and a sealed, permanently-filled reference electrode.

The vertical probe holder, modified to minimize fluid volume, prevents.

entrapment of air bubbles.

A double 0-ring seal is used to pr vent leakage.

The probe can be calibrated in place by injection of buf er solu ions (pH 7

and

10) into the sample loop.

The probe output is recorded on an indus rial grade meter mounted in a remotely located instrument panel.

Testing of the pH monitor was per-,armed to deter-mine its applic bility to a flowing sample s ream and to evaluate the effects of air bubbles in.he liquid.

Oata indicat d:he pH monitor is not affected by var'.a ions in flow or by the presenc of air bubbles.

The optir-um ope. ating temperature r ange of:he pH probe is 75-90 r, the maximum temperature and pressure are 1ZG F and 100 psig.

With constant control of the pH probe at a given t ape. ature within the. optimum oper ational

range, the accuracy and measurement range wi11 comply with HRC requirements, 1-13 0.3 pH units.

As inline pH moni.ors have be n used reliably for a

number, of years in the chemical and nucIear industries, there ar no apparent disadvantages.

The advantages are the application to accident and routine use, the remote oper ability, the simplicity, and the suffi-cient measurement range and accuracy.

To fulfill al I t/UREG-0737 requirements,

however, a

backup capabi1ity to measure the pH of grab sample must be provided.

Mith the prop sed grab sample collec ion sys

ems, this will r quir H

mew~urements of diluted reactor coolant samples.

B(ICO anticipates that pH's measured in diluted'samples c nnot be used to accurately det rmine

the actual pH of the reactor water.

For example, based on a 1:iCQ dilu-tion of a 0.1 ml r actor coolant'ample witii deionized water (pH 7.0),

the estimated pH of the r actor coolant watei determined,rom he analy-sis o

the diluted sample could be in error by 2-7 pH units.

Tnis tastes into 'accoun+

the results oi dilution only and not the presenc Gf 0 her cons ituents which can affec the pH.

Accordingly, E,"CHICO does not r-ccrnnend the use of di luted grab samples for measurement oi reactor coolant pH.

Cenera1 Electric Company suggest d

the use of pH paper for measurement of pH in reac.or

water, In conjunc ion with this

idea, a mies of laboratory tests were performed to determ-ine

.he ef-fects of high irradiations on the accuracy of the method.

in pH paper was iozrersed in solutions wi h pH 3.8 and 10.0 and:rradia ed

=or.

n minutes (1.5 x

10 rads) in one study and one minute (I 6 x '0".

5 rads) in another study.

one colors of the so lutions wer c "mp Iete ly des:royed in the ten minute test ard sicnific ntly altered in "he one minute iMC (0 5 pH unit shiit)

To compensate for. this effec, CE the pmcalure be modif ied to de rease the exposures to s 'gested pH pa"er.

.ne suggestion was to moisten the paper wi.h a

drop "r t'i(o of samo le ins ead cf total inversion of the pH paper in" the sample solution.

ine proposed ~di ied procedure was not demonstrated to be succ ssful.

ENRICO does not believe the pH pa er me+hod is sat-isiactory at present due to the irradiation effec.s observed.

Its future apoIicabi'hty will depend on the additional i sting and he developmen of. a tecmique o collec" a small, undiluted rea.or coolant samole ard to perform the me surement in a radiologically safe manner.

Ceneral Electr',c also suggest d tha the conduct-vity of a solution is a potentiaI method to ensure that t e pH Gf reactor volant is within cer ain acc p able

ranaes, i.e.

5.6 o 8.5.

EHICO does not belieie this t chnique meets tte

'IRC intent as the con-duc ivi.y of the reac+or coolant can possioly vary over a

large ranae under accident condit'.ons and cannot be used as an indication of "H wi h he r qu;r d accuracy.

4.2.4.3 Summar o,

Conductivit and oH Anal vsi s Yzthods.

The method. proposed by SFC/HUS and Gc. is applicable or measurement of the conductivity of reactor coolant water.

It includes an inline con-ductivity ce11.with a remote readou met r.

A backup capability to mea-sure the conductivity of grab samples was not noted by SEC or Cc,.

This backup capabi1ity can be provided with corrmerciaIIy available, portable conductivity met rs for analysis of grab samples.

Alternately, a backup capability, with NRC's concurrenc, would be a second, independent inline monitor which could be put into servic upon the failur of the first monitor.

The monitors are applicable to accident and normal conditions.

i

~

~

II Although there was an increase in conductivity of test solutions with an increase in radiation exposures, the effect was not due to monitor component failure.

It was a result of an increase or the conductivity 'of the test solution.

I is unknown if the increase in conduc ivity of the test solutions was inherent to he exrimental c"n-ditions or whethe".'ne should anticipate the generation of a conduc-'.ve species in postaccident re'actor coolan water.

\\

At present the only proven methods which has satis-factory accuracy and is applicable to measurement of pH unoer acciden.

condi ions are inline monitors.

The use of pH paper is not applic ole due o inaccuracies caused by hign radiation fields.

one

'pH paper method may be applicable with further testing;

however, the method wi 11 require the developnent of techniques, for the remote-addition of small re c-.or coolant samples to the pH paper and for th rema e comoarison of the pH paper colors with standards.

The pH analysis of dilu ed re c or c"olant samoles is not reconmended due to the potential inaccuracies of the measurements.

The. only alternative for a backup analysis capability is the use of two independent inline monitors; one in servic and on in standby.

The pH inline monitors are applic ble to both acci-

, dent and routine use.

There are no. anticipated radiologic I effects.

50

~ ~

~ ~

~

~ A

5 I

5.0 REFERENCES

TMI-2 Lessons Learned Task Fore Status Report and Shor -Term Recom-mendations, HUREG-0578, July 1979.

Ei senhut, 0.

G.,

Letter to AlI 'perating September 1979.

Nuclear P Iants, 13

3. Canton, H.

R.,

Latter to all Opera ing Nucle r Power P iants, 30 October 1979.

,'(RC Action Plan Developed as a Resul of TNI-2 Accident, iNUREG-0660, May ~oSO, revised August 1960.

5; Eisenhut, O.

C.,

Let r to All License s of Operating P iants and Applicants fcr Oper aCing Lic nsas and Ho iders of Construction

Permits, 5 Sept~Her 1980.

6. Clarification of 7i)E Ac.icn Plan Requirements, NUREG-0737, Nova..ber 19SO.

7.

10.

ar.

McCracken, C., Telecopy to R.

Huchton, Initiation of Work on Proac 83 (Post Accia nt Sampling),

23 Aoril 1981.

Priviata Co~i, nications, C.

McCrac!<en to R. Huch.on, 13 Aug s~,

1961.

Eme I, 8-A., Latter to C.

E.

Gi Irrcre, "Prcoosad cwork Schedule",

MAE-98-81, 14 May 1981.

lns rument tion fcr Light4at r-Cooled Nuclear Piants to Assess-PIanz and Environs Conditions Ouring and Following an Ac'ident, Regulatory Guide 1.97, Revision 2, Oec mer 1980.

Assumptions Used for Evaluating the Potential Radiological Consaauen-cas of a Loss of Coolant Accident fcr Boiling 'Rater Reactors, Regu-latory Cuida 1.3, Revision 2, June 1974.

12. Assumpt ions Used for Evaluating the Potenti al Radio logical Ccnsa-quencas of a Loss of Coolant Accident for Pressurized Mater Reactors, Regulatory Guide 1.4, Revision 2, June 1974.

13.

14.

GOC 19, Appendix A 10 ~ Part 50.

Helmhciz, H. R., Post Accident Sample Station Activity Source Te.ms, Caner al Electric Nuclear Engine ring Oivision decurrent QRF 000-.".,

March 19Sl.

15.

OeveIopaent of Procedures and Analysis M thods, or Post Acciden.

Reactor Coolant..

(nuclear Utili y Sarvic Corporation (SEC),

capri I 19S1.

'0

16. Technical Oescription and General information, General

=Ia

=". ic

Company, EEOC-24889, Section 1, data unknown.

17. Operation, C~neral E1ectric

'Company, t)EDC-24889, Sec ion 6,

date unknown.

18. &emical/Radiochemical Procedures, General Electric

Company, ORF C00-3, Section 7 and Appendix A, date unknown.

19.

BMR Cene. ic Pos -Accident Samp1 ing System Oesi gn Reouir ements, Caner al Electric Company, C~474-SP-1, Revision 1, April, 1~80.

20.

Boron Carmine Me hod, HACH Chemical

Company, APHA Standard

Method, 14th Ed., p. 290, I975.

2l. Standard Test Method for Boron Analysis in Mater, Rothod A - Carminic Acid Co1orimetric Method, ASTH 03082-74.

22. Perry, R.

H.

and

Chilton, C.H.,

Editors, Chemical Engineers'andbook, Fif.h Edition, bfcGraw-Hill'Book Company,

p. 3-215, 1973.

TADL At K

IN I LI5

~ALE <<Oi E

E 1&it ALP

.FOU.EK Rttv(:Iw tine I'age g

~ p Hethod/Analysis Titrimetric/Chloride Radiological Effects Anticipated Unknown Oasis Chemistry of the procedure and SEC leterature review of radiological effects on pil probes.

Spectrophotometric/

Chloride Unknown Yes Chemistry of tire procedure.

Gas Chromatography/

llydrogen Gas Chroma tograph/Oxygen II Yellow Spingsjhnalyzer/

Oxygen I~

I Conductivi ty Heter/

Conductivity pli Paper/pli pli Probe/pll Unknown Unknown Unknown lio Yes tlo fio Ho t/o f(ature of measurement method and pvior use of GC to to analyze velatively high activity samples

(~0.5R)

Nature of measurement method and prior use of GC to analyze relatively high activity samples

(~0.5R)

SEC litevature review of the radiological effects on the matevials of construction nf the sample probe.

Laboratory tests by GE, the SEC literature review of tire radiological effects on materials of probe con-struction and the SEC personnel interviews of in-dividuals with prior experience with similar measurements.

Laboratory tests performed by GE.

Literature review by SEC of the radiological effects on the mater ials of probe construction, SEC personnel in-terviews with individuals with priov experience with similar measurements, and common use of pll pvnhes in high radiation fields'.

<<liH'%

~ "

T