ML20093J094
| ML20093J094 | |
| Person / Time | |
|---|---|
| Site: | Beaver Valley |
| Issue date: | 03/26/1984 |
| From: | STONE & WEBSTER ENGINEERING CORP. |
| To: | |
| Shared Package | |
| ML19269A632 | List: |
| References | |
| CSD-84-144, NUDOCS 8410160493 | |
| Download: ML20093J094 (20) | |
Text
-
'r' o
CSD-84-144 )
TESTING OF DISSOLVED OXYGEN ANALYZERS FOR POST-ACCIDENT ANALYSIS APPLICATION (Revised ihrch 26,1984) a.
4.
i l
8410160493 841010 , , ,
DR ADOCK 05000 1
,a - --
- - , ,,,.n,. .-,,-,.,,..,----..,---,--,--,-na-.,-,.- ,,,.e,-,.. - - , , , , , - ,_n.nn.,,. n,.,,,..nn.,..- e. ,-n-,- , - . --
g
~
CSD-84-144 SUMARY AND CONCLUSIONS This report gives the results of testing of the Orbisphere Model 2606 Dissolved Oxygen Analyzer with the Orbisphere Model 2110 High Sensitivity Oxygen Detector for post-accident analysis applications.
The L&N Model 7931 Dissolved Oxygen Receiver and Probe and the YSI Model 54 Dissolved Oxygen Analyzer were used as reference analyzers.
These testing results demonstrate the following:
e The Orbisphere Model . 2606- Analyzer /Model ~ 2110 Probe is ^
suitable for post-accident ~ dissolved oxygen analyses. The testing was conducted with demineralized water, PWR simulated reactor coolant, and PWR simulated sump water with air, oxygen, nitrogen, and hydrogen gas blankets to simulate accident conditions.
~
e The L&N Model 7931 Analyzer is not suitable for post-accident dissolved oxygen analyses. Hydrogen causes a negative interference, resultirg in erroneous readings. However the L&N model 7931 Analyzer is well suited for PWR feedwater applications.
e The YSI Model 54 Analyzer is not sufficiently accurate to cover dissolved oxygen levels below about 0.1 ppm, although this testing program does not indicate that post-accident
~
matrix environment affects instrument response.
Accordingly, NUS recomends that the Orbisphere Model 2606-2 Analyzer /Model 2110-2 Probe be used in post-accident applications.
These testing results conf.irm that the Orbisphere meets the Reg. -
Guide 1.97 (Revision 2) range of 0 to 20 ppm and accuracy requirements specified in NUREG 0737 (Section II.B.3, Evaluation Criteria ,
Guidelines).
l l
ENGIME N OR R TION 90 l7lhI
UNACCEPTA9tt A
. .<,, c. . < E,te.
O .PFf0V!3 45 RfVI!!P
] nevitwtp J. O. NO. _
r SPEC.NO M k okl64
$NGew y
i 1 .
e - ,- ,.- . ,,, .-.e ,, - . , . ,.,..,-en - . , - - - , - . - - - --,--,,-,---,,..,,w.-,.m.-------,n-,--,,e-. -,.,.,e------..,, , , . . . - - , ,a_n,, n-,-,
D -
. CSD-84-144 l
~
MATERIALS USED IN TESTING l
Dissolved Oxygen Analyzers The Orbisphere Model 2606 0xygen Indicating Instrument is a digital readout meter with the following scales:
(a) 0-20 ppm (b) 0-2 ppm (c) 0-200 ppb (d) 0-20 ppb Range selection is made automatically by an internal autoranging circuit. Temperature is measured by a thermistor located within the sensor. The sensor is housed in the Model 2950. Flow Chamber.
The flow rate must be controlled between 50 and 250 mL/ min.
The Model 2110 Sensor must be calibrated because the sensitivity of the instrument depends on the tension in the membrane mounted over the detector. The detector consists of a gold cathode and silver anode. This type of cell is known as a Clark polarographic cell.
The Orbisphere Model 2606-2 0xygen Indicating Instrument and Model 2110-2 Sensor is marketed for post-accident applications. This instrument and detector are identical to the ones tested except-I for replacement of plastic with stainless steel in the probe housing.
The 'L&N Model 7931 Dissolved Oxygen Receiver and Probe is an analog readout meter with the following scales:
(a) 0-20 ppm (b) 0-2 ppm (c) 0-200 ppb The probe is easily calibrated in air. When the probe is placed in a sample stream, oxygen diffuses through a membrane and is reduced at the cathode:
02 + 4H++4e-- 2H 2O An equal amount of oxygen is generated at the anode:
2H2 O W 02 + 4H+ + 48' The diffusion continues until the oxygen tension on both sides of the membrane is equal. The electrical circuitry is arranged such that the current necessary to maintain the equilibrium is converted to read out the dissolved oxygen level. No net oxygen, acid, or water is consumed.
2
,,- _v ,---,----.-y. , , . _ , ,..,,,.,,.,,.c,,_-,,,v,.m...,s,. -....y,_7_,., ,,_,,,,y-.--,.,,,y, o_,_--,,-rw- ' , ww-y--=--**-*-w-w--+--+--veev r"*--* -
CSD-84-144
~
The YSI Model 54 Analyzer is a Clark polarographic type instrument I with two ranges:
(a) 0-10 ppm (b) 0-20 ppm This is a field type instrument and was previously installed in the testing apparatus.
TESTING APPARATUS Figure 1 illustrates the testing apparatus. Pumps ci.colate liquid
. to either the analyzers or sample tap and to the top of the apparatus.
The liquid is discharged into a four-inch diameter glass column to obtain equilibrium with the gas above the liquid. The gas blanket can be changed by placing the system under vacuum followed by displacement with air, nitrogen, oxygen, or hydrogen as appropriate for a test. Flow can be throttled to the analyzer loop as necessary.
O 3
O
..= e. e
- "e-----ee--+w-we- & g . ypwe-w w.r-<--y,,.,wwar q-g=&= - - -- '*-e.w ,g w g ag.,gg ee ep s- e -wmes - -weur, e -
M e---+M-wg *- -
f . ..
CSD-84-144
~
p,g,_[ ,, f gqq DATE M E; /fM s CLIENT N Mo,8M ~ FILE NO- 9384 o/ BY O AIOh -
SUBJECT <a O N A'd i
- 4 Checked By Qure[r Dissolve 2 Owygen A udyzerTe.eug Apparnlus u
[/k . N . ,
l 1 'a d6 t
h 6 4 ish 4 g, c- __
,, Flowde
~
\
ILM = -' .
Ghisphere mon y i 2 I I o Po obe.
T i -
Vac .- N Ho&
T a b _ ,
Ls M M.M g '1931 P4
% $= N
'( , .
l l z
CSD-84-144 CHEMICAL REFEREE TESTS The Winkler dissolved oxygen analysis was the referee chemical method for dissolved oxygen levels above approximately 0.1 ppm. This method is based on the reaction of dissolved oxygen reacting with an equivalent amount of dispersed divalent manganous hydroxide to fonn manganese (IV) exyhydroxide. Manganous sulfate solution is first added to a sample collected in a 300 mL BOD bottle. A potassium hydroxide / potassium iodide reagent is next added. After the manganese (II) hydroxide / manganese (IV) oxyhydroxide has settled, sulfuric acid is added. Free iodine is liberated in direct proportion to the original dissolved oxygen level. The iodine is titrated with standardized sodium thiosulfate solution, using Thyodene indicator.
The estimated precision is 10.05 ppm, and the estimated accuracy is approximately 10.1 ppm for oxygen levels above approximately 0.3 ppm.
The CHEMetrics comparator method is based on breaking a sealed tube, containing reagents under vacuum, in a flowing sample stream. The sample is forced into the tube by atmospheric pressure. The color developed is compared to standards supplied with the kit to determine the dissolved oxygen level. Kits and comparator standards are as follows:
CHEMetrics Kit Rang Standards for Comparison
(
I Model 0-100 0-100 ppb 0, 10. 20, 30, 40, 60, 80, 100 ppb Model 0-1 0-1 ppm 0.0, 0.5, 0.1, 0.2, 0.3, 0.4, 0.6, 0.8, 1.0 ppm Model 0-12 0-12 ppm 1, 2, 3, 4, 5, 6, 8, 10, 12 ppm Samples showing colors between two comparator standards are estimated as the average of the two standards.
TESTING METHOD The testing protocol consisted of calibrating the instruments and l then testing the following matrix conditions:
l Matrix Composition l
A Demineralized water with air / nitrogen / oxygen blanket to vary oxygen levels.
i l
5
~
t
CSD-84-144
~
Matrix Composition B Demineralized water with hydrogen blanket -
to vary oxygen levels.
C PWR reactor coolant with simulated fission products with air blanket.
D PWR reactor coolant with simulated fission products with hydrogen blanket.
E PWR sump water with simulated fission pro-ducts with hydrogen blanket.
The initial calibration consisted of using air-saturated water in a flowing system and adjusting meter responses to the value obtained by the Winkler analysis. However, the YSI could not be adjusted to referee analyses. Since the Orbisphere read 2.58 ppm relative to 1.91 ppm by the Winkler analysis during the initial testing on February 28, 1984, all instruments were then air-calibrated. The Orbisphere still gave nigh readings. The flowmeter was not properly calibrated and the flow rates were in excess of 250 mL/ min. as recommended by the manufacturer. The flowmeter was then calibrated and the flow rate was adjusted to 124 mL/ min. for the remainder of the test. On February 29, 1984, the flow rate through all probes was set at 124 mL/ min. The Orbisphere and L&N probes were calibrated
(
relative to Winkler analysis results. Thus, The the first valid L&N instru;nent Orbisphere data were obtained on February 29.
is not flow sensitive and thus, valid data were collected on February
- 28. All testing results are sumarized in Table 1.
Approximately three liters of simulated PWR reactor coolant and three liters of simulated PWR sump water were tested in the apparatus illustrated in Figure 1. This was insufficient sample for full testing. The reactor coolant contained approximately 2000 ppm boron as boric acid at a pH of approximately 5.5, whereas the sump sample contained boric acid at the same level but sufficient caustic to increase the pH to about 9.3. Sampling was compromised because of insufficient sample. The CHEMetrics method was also affected by the high pH sump water as shown in Table 1. The Winkler method was slightly modified as noted in Table 1 because of the matrix pH.
Additional testing on simulated PWR reactor coolant was conducted l
on March 21-22, 1984 and on simulated PWR sump water on March 23, l 1984. All test data are sumarized in Table 1.
i 6
i
s CS0-84-144 TABLE 1 SUleMRY 0F TESTING RESul.TS pga Chemical Analysis
. Model 2006 lbdel 7931 Date Matrix Cover Gas Winkler CHEMetrics Orbisphere L818 2/28/84 Demin. Water Alr-N2 ----- 0.010 1 1]l 0.007
0.25 I l I 0.28 O.289 ----- 1;1h 0.26 0.145 -----
I,;l 0.14
0.04 i
,, 1 0.041 0.067 0.1 I J 0.081 0.231 0.2 i l'l 0.230 0.752 - ----- I'll 0.70 2.33 ----- i 1;l 2.80 Air 7.30 ----- 1,ll 1 8.70 2/29/84 Demin. Water Air 6.67 -----
6.65(2) 6.67(2) 0.617 ----- 0.611 0.565 ,
, Air-N2 0.060
0.060 0.056 0.148 0.15 0.141 0.140 0.703 0.7 0.734 0.66 6.78 7. 6.87 6.70
) Air 12.65 ----- 12.95 12.90 Ai r-02 3.12 l 3/1/84 Demin. Water Air-N2 3.22 ----- 3.11 8.63 ----- 8.43 8.70 Air-02 16.16 17.0 8 17.6 -----
16.0 ----- 15.25 16.1 l 18.70 19.6 19.7 -----
l12 5.73 ----- 5.37 0.140(6) l 1.52 ----- 1.497 0.034(6) l H2 air 7.13 ----- 7.05 7.10
) 6.93 6.90(3) 3/2/84 PWR Reactor Coolant Air 7.17 -----
l 2.06 0.080 Matrix H2 2.20 -----
0.45 -----
.0.370(6) 0.016 (4)(6)
0.10(5) 0.051(6) 0.006 (6) 1 l
l 4
l 7 1
1 _ . _ _ _ _ _ _ _ _
i
. CSD-84-144 c TABLE I (Cont)
SUlelRRY OF TESTIllG RESULTS pse Chestcal Analysis Model 2606 Ilodel 7931 Date Matrix Cover Eas Winkler CHElletrics Orbisphere L811 3/3/84 PWR Sump H2 ila;l0.35 ----- ----- -----
Matrix I,b,10.32 ----- 0.293 0.014 )
0.13 ----- 0.112 0.008 l 0.15(5) 0.039(6) 0.006 ll (a)0.48 ----- ----- -----
(b)0.44 0.8(5) 0.402 0.017(6) 3/21/84 PWR Reactor Coolant Air 7.44 -----
7.44 7.45 Matrix Air + O2 19.78 ----- 19.79 19.90 3/22/84 15.78 ----- 15.72 16.60 11.15 -----
11.07 11.00 Air + H2 1.53 ----- 1.44 -----
H2 0.49 ----- 0.463 -----
H2 0.21 ----- 0.184 -----
Air + H2 0.73 ----- 0.700 -----
H2 0.365 -----
0.336 -----
3/23/84 PWR Sump Air 7.46 ----- /.49 7.49 Matrix Air + O2 16.16 -----
15.35 16.40 18.90 ----- 18.59 19.10 H2 10.04 ----- 10.08 -----
H2 4.32 ----- 4.31 -----
1.39 -----
1.40 -----
0.878 -----
0.873 - - - - - -
0.608 -----
0.599 -----
O.395 -----
0.380 -----
0.267 -----
0.259 -----
j j 8 t
. CSD-84-144 7 TABLE 1 (Cont)
St# MARY OF TESTING RESULTS NOTES:
(1) Flow rate was greater than 300 mL/ min. - Orbisphere and YSI are flow sensitive. The Orbisphere should be set between 50 and 250 mL/ min.
(2) Once the flowmeter was calibrated, the flow rate was set to 124 mL/ min. for all subsequent tests. YSI could not be calibrated against Winkler.
(3) BecausE of boric acid, 4 mL of KI/KOH reagent and 2.5 mL of 1:1 H2SO4 were used for the Winkler analysis versus 2 mL of each.
(4) Insufficient sample for proper purge.
(5) CHEMets are apparently affected by matrix.
( (6) Data was not used in statistical evaluation.
t i
l l
l .
9
. c, - s CSD-84-144 e EVALilATION OF TESTING RESULTS Statistical Evaluation Table 2 sunmarizes valid testing results for the Orbisphere and LAN analyzers. Data for the Orbisphere, which were obtained when the instrument was not calibrated properly, have not been included in Table 2. Table 3 gives the results of a statistical evaluation of the Table 2 data. The first evaluation determines the equation for "best fit" straight line:
Y = mX + b where "Y" is the analyzer reading, "X" is the chemical analysis referee result, "m" is the slope and "b" is the intercept. Since the NRC gives separate accuracy statements for 0 to 0.5 ppm and from 0.5 to 20 ppm, the statistical evaluation also considers these ranges. Table 3 also gives the mean difference between the chemical analyses and the corresponding analyzer readings. The mean difference is given in ppm for the O to 0.5 ppm range and in percent for the 0.5 to 20 ppm range. The predicted range for the population mean difference can be used to determine if the instruments meet NRC accuracy requirements. As shown in Table 3, NRC accuracy requirements the Orbisphere. The requirement for the O to 0.5 ppm are rangemet is bg_ 0.05 ppm, whereas this testing shows an upper limit of f 0.035 ppm. The accuracy requirement for the 0.5-20 ppm range is i 10%, whereas this testing shows an upper limit of 3.60%.
The L&N analyzer meets NRC accuracy requirements for samples not containing hydrogen, but full matrix testing with hydrogen shows that this analyzer cannot be used in post-accident or normal reactor coolant applications. However, the L&N analyzer should provide good service for PWR feedwater applications Figures' 2 and 3 show the data and line of "best fit" for the Orbisphere for the two ranges. It can be seen in each figure that the data are very close to the line of "best fit" indicating a high correlation coefficient. It can also be observed that the probe is not biased by any of the three matrices in which it was tested.
1 l -
10 i
- . . _ _ . ~ . _ _ _ . _ . . _ _ _ _ , , _ . _ _ _ - . , _ _ _ , . _ . . ~ . _ , _ _ _ _ _ . _ . . _ . _ . - _ - _ _ _ _ _ _ . _ _ . _ _
' ~
^
.. .. *CSD-84-144 TABLE 2
~
SUMARY OF DATA USED IN STATISTICAL EVALUATION ppm Dissolved Oxygen Instrument Chemical Absolute Analyzer Matrix (1) Range Reading Test- Difference Pl!'_
Model 2606 'A 0-0.5 0.056 0.060 0.004 Orbisphere A 0.141 0.150 0.009 0 0.184 0.210 0.026 D 0.336 0.365 0.029 D 0.463 0.490 0.027 E 0.112 0.130 0.018 E 0.259 C.267 0.008 E 0.293 0.33(2) 0.037 E 0.380 0.395 0.015 E 0.402 0.460 0.058 I_
A 0.5-20 0.611 0.62 1.47 4 / A 0.734 0.70 4.63 A 3.11 3.22 3.54 A 6.65 6.67 0.30
-A 6.87 6.78 1.31 "
A 7.05 7.13 1.13 A 8.43 8.63 2.37 A 12.95 12.70 1.93 -
A 15.25 16.00 4.92 A 16.16 17.60 8.91 A 18.70 19.70 5.35 B 1.497 1.52 1.54 8 5.37 5.73 6.70 C 6.93 7 17
. 3.46 C 7.44 7.44 0 C 11.07 11.15 0.72 C 15.72 15.78 0.38 C 19.79 19.78 0.05 D 0.700 0.73 4.29 D 1.44 1.53 6.25 D 2.06 2.20 6.80 t
E 0.599 0.608 1.50 E 0.873 0.878 0.57 E 1.40 1.39 0.71 E 4.31 4.32 0.23 E 10.08 10.04 0.40 t
11 9.s.. - - - + . - -- - ..-. ,-, w.e. .+g g,+%wy4-9# .-y. -, , . ~ep.m- ,
5,,-.w.-, ,,.,_-,,_y,w-,,,n,g,y.,w,.ywy, r--* wg ,-y-y-
~
CSD-84-144
[
TABLE 2 (Cont.)
SUl0ERY OF DATA USED IN STATISTICAL EVALUATION ppm Dissolved Oxygen Instrtment Chemical Absolute Annlyzer Matrix (1) Range Reading Test Difference 1
Modal 2606 F 0.5'-20 7.49 7.46 0.40 Orbisphere F 15.35 16.16 5.28 F 18.59 18.90 1.67
. .228 P.odel 7931 A 0-0.5 0.007 0.01 0.003 L&N A 0.041 0.04 0.001 A 0.060 0.06 0 A 0.081 0.07 0.011 A 0.140 0.15 0.010 A 0.140 0.15 0.010 A 0.230 0.23 0
( A 0.260 0.29 0.030 A 0.280 0.25 0.030 1
A 0.5-20 0.565 0.62 9.73 A 0.66 0.70 6.06 A 0.70 0.75 7.14 A 2.80 2.33 16.79 A 3.12 3.22 3.21 A 6.67 6.67 0 A 7.70 6.78 1.19 A 7.10 7.13 0.42 A 8.70 7.30 16.09 A 8.70 8.63 0.80 A 12.90 12.70 1.55 A 16.10 16.00 0.62 A 17.00 17.60 3.53 A 19.60 19.70 0.51 C 6.90 7.17 3.91 6
12
._r.. -~----,--...,._r,-,,,-_,_.,,.4 -, .,_,,.,.7,,--,,m-_.,,,-.,.-.-__.__~..._r.,__.,._.,.,_ ,,~.---._,,m,_,,,-__,,,..w.--ey , - - _ , _
>: l CSD-84-144
,. TABLE 2 (Cont.)
L d
SUMARY OF DATA USED IN STATISTICAL EVALUATION NOTES (1) Matrix Composition A Demineralized water with air / nitrogen / oxygen blanket B Demineralized water with hydrogen blanket C PWR reactor coolant with air blanket D PWR reactor coolant with hydrogen blanket E PWR sump water with hydrogen blanket F PWR sump water with air / oxygen blanket (2) Average of 0.346 and 0.322 ppm
, t
(-
e e
[
13
" -~ - ~ -- w - +,----*-ev
--+wr% , w..-w,-- ----g-- -- --yvg-- ,,,,r_e-g* y ,--
m, 3-(
~
I t
CSD-84-144 ,
TABLE 3 STATISTICALEVALUATIONOFTESTIIESULTS(I)
No. of Std. Dev. -
Data Hean of Hean Predicted Range for i Analyrer Range a(2) b(2) . Points (n) Dffference(3) Difference (4) Population Hean Difference (5) l 'O.0231+ 0.0162 x 2.262.,
l l Orbisphere 0-0.5 ppe 0.9263 -0.0020 10 0.0231 ppa 10.0162 ppe 0.0115-0.0347 ppe Model 2606 2.6486.+ 2.5004 x 2.048 =
. 6 l 0.5-20 ppe 0.9683 0.0714 29 2.64861 12.50045 1.6977-3.59951 0.01061 0 0119 x 2. M ,
1 h
I L&M Model 0-0.5 ppm 0.9767 0.0020 9 0.0106 ppa 10.0119 ppe 0.0015 - 0.0197 ppu
, 7931 (6)- .
l 4.77001 5 5111 x 2.145 ,
! 0.5-20 ppe 0.9874 0.1594 15 4.77005 15.51111 1.6178 - 7.82221 1
l noits:
(1) Table 2 data used (3) Mean difference. Y . [(X-Y)g (2) Y - Analyzer reading, ppe (4) Std. Der. Mean Diff., SX -Y . -(X-Y)g] 2 X = Chemical analysis, ppe n-1 m = Slope I (5) Predicted range for population mean difference = X i SX-Y x t p
i g=agtercept,ppe
= +b 1
1 t = 951 probability factor for a specific number of degrees of i I freedom, n-1 (6) Ma.trix conditons with hydrogen have not been included, 14 f
. , . .e
< O Is 2
C wW T C W 7.
N _
g --- _----
,d W E L
@ --~!
d g -
- _,m,
-2 t/1 n - "
I U -
75 L.
.* t ,
,m 3 3 ;-
n
.- 9. E
.- i M' ..m -
m g
t -
N
==
2E__
- O
- _r: - , '.
zw--
~- N O W= w g
. M ,m '
, 2 M aj i?, g J1 sy= (pg w -
M g
$s. r S h M in 1 - :'
,- J
-4 o L o - - - E g N
ag; e i i >
O 5 Qw g -'m W W,
[ me .
w .M
- u. .
m
. a
!. ..Im Q E
, c
- .- C
~
M G ',
.' - w x
. h c; a - .c, 4
',,', e i
f C **'a ag; '
.i g .~,.
h #b.
1)
. p -
4 Q 's
- W '. .[4 1 o -
, > -' . ' - S a - _
o
- m .1'.
- l. .M i !' Q ,',._
) ,g 1..
@ (
~
I w m. m'Y,,y. ,
,x O
. -m i @
- i' T O
.$. $b
- 2 ..
2 m
& E
'-.-W i M
g a m e m N
'O h d b O f
hi!W;,$o ' ---us -
- m. o
. 'CSD-84-144 4 FIGURE 3 Orbisphere Model 2606 Dissolved Oxygen ,
Analyzer Re.sponse: 0.5 to 20 ppe
, .}.
N n f d
! .i/l
/ fI
- 18 .
r f , litil l;fm =
. II .
llil; l L 1:h -ni i:al il (11 %
i, d-7 16 "
o,*
, a 14 .
/
, e e a 12
' /
1 Ip / .
, j- 10 g g
i, L ,
i- 8 f ii-- hijil iiiip;i 1 ;jiijl~
~
, toi.pj q r (p ggpj y .
o Nit 11 si <:ce (oo ini 4 g
,1 6 Nil ; ii.p hit.i- g ' H,l-
, '. 6
, i f .
4 }
/
f 2 A i
'#q M
I i 0 i
0 2 4 li ;!, 10 12" "14 Ils 18 20 Chemical Analysis PPM Dissolved Oxygen IA
- ?
E CSD-84-144 Other Considerations Affecting Results
(~
When changing oxygen levels from high (e.g., several ppm) to low (e.g., less than 100 ppb), the analyzers may require three hours or longer for a stable response. Because of the limited time available for this testing and of the concern to simulate power plant conditions, the stabilization period generally did not exceed two hours except for two occasions in .which the testing apparatus was allowed to remain operating overnight.
Another consideration is the sampling technique used for low oxygen levels when testing pWR matrix solutions. Normally, the BOD bottle is allowed to overflow at least three bottle volumes. The mouth of the bottle is normally maintained under liquid during sampling
.and white reagents are being added. Sufficient matrix solutions
-were not available to allow this technique. However, Table 1 data show that chemical results in these few cases are only slightly high.
When calibrating the Orbisphere for post-accident applications, the following is recomended.:
- a. The probe should be maintained in water at least several hours prior to use.
N' b. Air-saturated water should be recirculated through the probe assembly for 30 to 60 minutes. The calibration control knob should be adjusted to either the Winkler chemical analysis results or to the air-saturated valw for the particular temperature l
i of the water. ,
- c. The sample should be circulated through the. probe assembly for at least 60 minutes. This may be longer if oxygen levels are low.
?
e I
17 l l
i l
=.... - - -- -
CSD-84-144 !
I i
i EFFECTS OF RADIATION ON ORBISPHERE PROBE.
The attachment is a technical note published by Orbisphere which indicates that the Model 2110 sensor is expected to operate for about one year in a radiation field of 6.104 Rads / hour before the electrolyte and membrane must be changed. This assumes that the membrane is made of Tefzel and the electrolyte is 6M KOH solution.
l 5
i I
r 1
l 18
-, . . , - . - - - . , - . , - - - , . . . ...,y-, _ . . . . _ . . - , , - . - , . - - - - . . , -
,-,,-,,,,m,-..,. - , . , ,,c_ .._ . ~ ~ , . . , _ .,. ._
f: -
~'
nrbi m h n e d m.. -W. m. t s a Dnnsson of Ortssphere Corporabon 287 Lackawanna Avenw.W Patersort N J OLtN Ptum. 201785 8280 -
c
~R [L. R -. _
q _ .!,-
fA D (M iP if ,j i s %, ... l j ."t / 3 !'
,..(c") 1 * (*
yA] y l/ atp (i4::.M *',
a:; ..h) t lt : t .
}
4 .
V U Ui H@ii O.? O Jd -
~
- Measurements of Traces of Dissolved Oxygen in Radioactive Water by l. M. Hal. with the Orbisphere System 2713 Analyzer l
The problem often arises, at nuclear reactor or Gas evolutlen high eneray accelerator installations, to deter.
- mitte trace oxygen levels dissolved in radioac. Another consideration with fluorinated materials tive water. These oxygen levels can be as low such as Kel.F, is the nefarious effect of gases.
as several (ug/ litre), and their determination notably hydrofluoric acid, resulting from chem.
l therefore demands highly sophisticated, sense. ical decomposition of the plastic caused by the I ~ tive instrumentation. The Orbisphere System radiation.
2713 Analyzer was designed expressly for the type of application, and has earned an excellent An intearated dose of 10e Reds produces about reputation for reliable. trouble free service in 3 ml of HF at normal temperature and pressure fossil fueled power plants. Before the system per gram of the polymer. At a dose rate of can be implemented for monitoring radioactive 6.104 Rad /hr. the Kel.F pieces in the Orbis.
water, however. it must be verified that the ra. phere sensor weighina 18 g would therefore diation level to which the oxygen sensor is sub, produce 1.5.10 8 moles HF per hour.
iacted does not exceed certain desian limits. Some of this acid vapor is washed away by the This note discusses the effects of radiation up.
on the various components of the sensor. pre. flowing sample water and is therefore of no l sents limiting integrated doses for structural concern to the operation of the sensor.
parts and determines servicing frecuency for
(, electrolyte and membrane changes as a func. That fraction of the vapor which dissolves in A tion of dose. For illustrative purposes, lifetime the electrclyte in the interior of the sensor is calculatioris are performed with assumed aver. without effect at first, since it is neutralized by age radiation leve's of los Rad /h over long per. potassium hydroxide:
iods, and servicing frecuencies are deduced for sustained doses of b.i0s Rad /h. KOH + HF --* KF + H O Because oxygen measurements must be per.
formed on cooled, depressurized water in a Once the alkali has been removed by this flow circuit outside a reactor core the rad,a. i means, however the accumulation of acid in tion as assumed to consist mainly of I (elec-the electrolyte will lead to a diminished sensi.
tromagnetic) and p (electronic) radiations. tivity of the system towards oxygen. At the
$0"'e',*n$dtr Structure! Inte@gmaterials ,; on trolyte. these considerations suggest that elec.
) of K In "'"eel c.
The following table lists the materials used in construction of the Orbisphere probe. trolyte renewal will be reovired sometime after a minimum period of about 30 days, depend.
Material Type of Limiting dose Lifetime at ina upon the rate of diffusion of the vapor out stress Reds 104 R/hr of the solid, and upon the fraction of the vapor which enters the sensor. A most probable fro.
Gold Shear 1,0'. 10s years cuency of servicing due to this cause of about Silver ,, ,, ,, once every 2 months seerns likely. This frecuen.
Bronze , e 316 Sta.m.
,J can KOH beindecreased the sensor.byAtincreasing 6 M. for example, the cuantity a
(5,s psteel "
lifetime of one year should be realizable.
M 0s 28 wh The metals are of no concern,in respect of the effects of radiation upon their strength, since a The Membrane - ~ ~ ~ -
very large inteorated a noticeable dose difference to is reavired their physical to makepro. The membrane of an oxygen sensor is a critical porties. item. in that it determines the accuracy of the measurement.
Plastic materials in general are more subject to
! afteration by radiation and Kel.F in particular Any chances of chemical composition, density
( suffers a rapid deterioration in its rnechanical or crystalknity induced by radiaticn. might in.
i strength when subjected to an integrated dose fluence the oxygen permeability of the menp l
in excess of 2.10e Rads, it is an ideal material brane. and henco cause a drift in the sensitivi.
I from the point of view of other recuisites, suen tv of the detector and a progressively increas.
I as chemical resistance and machineability. in ing error of measurement. Another relevant the present application. Radiation damaged consideration is that the membrane is subjected
(
l parts could feasibly be replaced by new ones. to tensile stress, and it is this property which
~
provided the sensor did not acquire a danger. usually is most snarply affected by ionizing
. ous level of radioactivity, radiations.
19 _
l
.Q.*j _
11 in other apolications a teflon 9 (registered Secondiv, the membrane which encloses the I trademark of Dupont) PFA membrane is em- sensor is permeable to the gas. and hence per. I ployed in the 2713 System. This material is too mits its continuous escape to the flowing sam-
-ouickly deteriorated. however, by radiation. A ple water. In the caso of % mil telzel mem.
preferred substitute is leftelt (registered trade. brane. for c ample the escape rate is about mark of Dupont) which shares the strenoth and 30 pl per hour. per cma of membrane, and per i chemical stability of teflon. but can withstand one atmosphere difference of hydro 0en partial t an inteorated dose of 10* Rad. thus permittino pressure across the membrance. This is suffi- !
e a lifetime of 14 months at the standard dose ciently high to ensure that excess hydrogen rate of 10* R/h. pressures do not build up inside the sensor.
I As was shown above, hydrocen peroxide could The Electrolyte be formed at the rate of 4.10a moles /hr. when 6.10* Rads /hr of radiation is completely ab.
The Orbisphere probe contains about 2 milli. sorbed in 2ml of water. Various "back reac.
liters of an aqueous electrolyte (IMKCl. pH : sons" such as
+
14). Water is subject to radiolysis in the pre.
sence of radiation with production of hydrocen H + H O ~ 2H O peroxide and hydrogen oas. These products and are both of concern though for different rea. 2H2&O ~+ 2H 2O' + 0 &
sons; hydrogen peroxide could be responsible for an erroneously high oxygen measurement. - tration serve to to limit the hydrogen a steady state value, peroxide typically concen.
of the and hydrogen gas, if it were evolved at a rate order of 10 s molar in pure water.
exoeding the rate of permeation of the cas through the membrane to the exterior, could it must be recognized, however. that most'of cause the build-up of pressure in the sensor this hydrogen peroxide has no influence upon agam interfering with the accuracy of oxygen measurements with the Orbischere sensor since measurement. it cannot cain access to the cathode.
Fortunately. due to certain unieve desian fee. Peroxide which is formed in the reservoir of the tures of the Orbisphere sensor, the rate of sensor is "traoped" at the rad:olysis of water is not sufficient to pose a which surrounds the cathode. guard electrode
- problem in practice. The maximum amount of Th.is electrode reduces the' hydrogen peroxide enerav absorbed by 2 mi of water irradiated to hydroside ions accordmg to the equation with TL10d Rad /hr is:
H22O + 2e = 2 OH-6.10+ (Rads) . 6.242.10m (eV_) .1.fa ) . 2 (ml)
(hr) (Rad g). (ml) and these hydroxide ions are totally inactive at the oxygen sensing cathode.
= 7.5.100 eV/hr There remain for consideration only those per.
hence the maximum number of water molecules oxide molecules formed in the thin layer of decomposed per hout is: solution sandwiched between the membrane and the cathode. The maximum rate of forma.
tion of these is :
, C 7.5.101 (eV) . 0.065 (mole _culeJ 4 . (hr) (eV) %.6.10* . 6.242.10's .1.W. (0.316): .10 * . 0.065
= 4.85.10n molecules = 8.10 ' moles
. (64233u).:
hr ~hT = 6.3.10 in moles /hr if the only reaction of sinnificance were as- Here 0.316 cm is the radius of the cathode and 10
- cm is a typical thickness of the said elec.
sumed to be": ._ trolyte layer.
- (in practice half the yield. of the yield of excited hydrogen water molecules, is lessdue thanat the if allcathode, of this hydrogen it produces peroxide a current is reduced of to recombination of free radicals)
- 2. 6.3.10 n . 96500 = 0.34 nA 3600 2H2 O ** H + g H O2g This current is to be comoared with the sensi.
then the maximum rate of evolut. ion of hydro
- tivity of the Orbisphere sensor towards oxygen, gen would be: namely 2 nA/ ppb at 20*C.
.3 6.10 * (mas) .22400 (ml _ ) The hydrocen oeroxide produced by radiofysis-2 hr (mole) at the 6.10* Rad /hr rate, therefore causes a
'"*""'"'**""'"" '" "O '
= 0.009 ml/hr Conclusions in practice. water .is highly transparent to I ra.
diation (penetration creath ~ 30cm). so that ,in summary the materials used in the Orbi.
ont.y a small fraction of the order of 5% of the sphere 2110 sensor are expected to withstand raciation leads to,radiolysis in the 1, cm path an inteorated radiation dose of 10* Rads with.
length in the Orbischere sensor. This implies out structural failure. Tetzel is a preferred mem-a hydrogen evolution rate of only 0.45 pl/hr. brane material. Radiolysis of water does not constitute a prob'em because, of uniaue fea-
<> This hydrogon dissolves in the acueous elec- tures of the Orbisphere desion. namely the trolyte, but its concentration does not increase
. without lirmit due to two limiting phenomena. f,},aunc f e gu rd r g' e actrod . At Firstly, various back reactions Rads /hr continuous operating periods of at least one month are to be expected betw og:H A + H 2 O 2 ~ 2H ,2O electrolyte and membrane changes. and a m,een s.
naturally lead to a steady state limiting con. nor chance of electrolyte composition will per-centration of the hydrogen gas even in a com. mit this period to be extended to about one pletely closed system. year.
20 e e e-
,-*-e,- ww- ,_, - , - - - __-,, -, -,., - , - , . _ ,-,-m._,,-w,,w,,-w y-,_,-e.,,,,,,n__y., m.--+-,,-,,.,m.,... -9,., _ . , , , , ,- ,m.__ ym- -+,y,- ,,,,%,,,y. ,7 , g._,,