ML021000294

Corrosion Damage to RPV Head at the Davis-Besse Nuclear Generating Station

ML021000294
Person / Time
Site:	Davis Besse
Issue date:	03/29/2002
From:	Macdonald D Pennsylvania State Univ, University Park, PA
To:	Sheron B NRC/NRR/ADPT
References
Download: ML021000294 (15)

Text

PENNSTATE qW Center for Electrochemical Science and Technology The Pennsylvania State University 201 Steidle Building University Park, PA 16802-5006 Dr. Brian Sheron Associate Director for Project Licensing and Technical Analysis Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, DC 20555 (814) 863-7772 Fax: (814) 863-4718 March 29, 2002 Re: Corrosion damage to RPV head at the Davis-Besse Nuclear Generating Station.

Dear Dr. Sheron,

I have read with considerable interest the NRC and newspaper accounts on the corrosion damage to the reactor pressure vessel head at the Davis-Besse Nuclear Generating Station. Having worked on defining corrosion processes in nuclear plants for the past thirty years, starting as an employee of Atomic Energy of Canada Ltd in the early 1970s, I offer below an explanation of why the damage occurred. My explanation is undoubtedly incomplete in detail, but I believe that the general facts are correct.

A number of years ago, under EPRI sponsorship, and using unique experimental techniques that were developed in my laboratory, my colleagues and I studied the properties of very concentrated boric acid/sodium hydroxide systems at temperatures up to 300 TC.

(I am reluctant to call them "solutions", because 100 m B(OH)3 is essentially wet boric acid powder).

The property that we measured was the pH and, much to our surprise, we found that concentrated boric acid/NaOH systems are capable of generating very low pH values, as you will see fromthe attached publication. Indeed, pH values as low as 2 were found in some systems (see Figures 9 to 11). These highly acidic conditions are conducive to both general corrosion of carbon steel and stress corrosion cracking of Alloy 600 (so-called "acid cracking"). Accordingly, the scenario that I postulate is that a mechanism exists (possibly evaporation in a crevice at the head penetration) that results in the precipitation of boric acid, which, because it is wet, produces highly acidic local conditions, giving rise to accelerated attack. Given the observation of boric acid deposits in the damaged area, I believe that the above scenario is essentially correct.

Indeed, ever since performing the work that is described in the attached paper in the late 1980s/early 1990s, I have often wondered if this little-appreciated property (low pH) of ultra-concentrated boric acid might become an issue in the operation of PWRs. This work has been largely overlooked, but it would seem that my concerns were not unfounded.

I hope that the above material and the attached paper are of interest in your deliberations on the cause of damage at the Davis-Besse reactor and the implications that it raises for other PWRs. Should you wish to discuss this work further, please contact me by telephone at (814) 863-7772 or by e-mail at ddm2@psu.edu at your convenience.

Sincerely, Digby D. Macdonald, Director Professor of Materials Science and Engineering College of Earth and Mineral Sciences 0O An Equal Opportunity University

Proceedings of the Fourth International Symposium on Environmental Degradation of Materials in Nuclear Power Systems Water Reactors Proceedings of thp Fourth International Symposium on Environmental Degf*-dation of Materials in Nuclear Power

.Systems-Water Rejectors sponsored by the TMS/ASM Nuclear Metallurgy Committee, ANS Materials Science and Technology Division, and NA:CE"Unit Committee T-2A on Nuclear Systems and held August,.6-10, 1989, in Jekyll Island, Georgia Edited by Daniel Cubicciotti Electric Power Research Institute P.O. Box 10412 Palo Alto, CA 94303

Measurement of pH in High Temperature Concentrated Solutions Simulating PWR Steam Generator Crevice Environments Herking Song Digby D. Macdonald SRI International SRI International 333 Ravenswood Avenue 333 Ravenswood Avenue Menlo Park, CA 94025 Menlo Park, CA 94025 Carl Shoemaker Peter Paine Electric Power Research Institute Electric Power Research Institute Palo Alto, CA 94303 Palo Alto, CA 94303 Abstract High-temperature sensors [PtIH2 and Hg/HgO/ZrO2(Y2O3)] have been used to measure pH as a function of temperature for 1 m NaOH, 1 m NaOH + 1 m Na2SO4, 1 m NaOH + 1 m Na2SO4 +

x m B(OH)3 (x = 2, 5, 20, 50, 100), and 1 m NaOH + 1 m Na2SO4+ 1 m NaC1 at temperatures between 1250 and 3000C. The ceramic membrane sensor was found to provide more consistent data than the hydrogen electfode. After a detailed analysis of the errors associated with the isothermal liquid juncti6n0otential of the reference electrode, the activity of water, and the partial pressure of hydrogen in;the pressure vessel, and allowing for possible errors in the calculated pH values, we conclude that EPRI's MULTEQ chemistry code is capable of providing reasonable estimates of the pH of high-temperature aqueous solutions of the type that exist in PWR steam generators. The concentrated boric acid solutions were calculated (and observed) to yield acidic solutions with pH <3 for x > 50 m.

Introduction Concentrated electrolyte solutions are known to form in crevices under high heat transfer conditions and in other dry-out regions of nuclear steam generators, and these solutions pose a potential corrosion problem when the activity of hydrogen ion is high. A multicomponent system of particular interest is B(OH)3/NaOH/NaCI/ H2SO4/-I20, because the components either are present in the bulk water as impurities (e.g., Cl-, SO4 2') or are added for corrosion control (e.g.,

B(OH)3 to control denting). Although the acidity of such systems can be estimated by using sophisticated computer codes (e.g., EPRIs MULTEQ), considerable uncertainty exists about the accuracy of the calculated pH because of poorly known activity coefficients and other factors.

Accordingly, direct experimental determination of the pH would greatly contribute to validating the codes for application to concentrated, multicomponent systems.

In this study, we used Pt/H2 and yttria-stabilized zinconia membrane (YSZ) electrodes to measure the pH of high-temperature/high-pressure concentrated aqueous solutions (Table 1) at temperatures from 1250 to 3000C. The cell and electrodes were designed to allow precise pH measurements to be made under typical PWR steam generator conditions, and we assessed the accuracy of the pH measurements by a comprehensive error analysis. The measured values were compared with data calculated using MULTEQ. The reasonable agreement between the measured and calculated values (ApH < +/-1) shows that the apparatus developed in this study is capable of yielding useful pH measurements in concentrated, high-temperature/high-pressure solutions and demonstrates the capability of MULTEQ for estimating the pH of simulated crevice environments.

12-77

Experimental Two kinds of sensors were used in this study to measure the pH of the concentrated solutions listed in Table 1: Pt/H2 hydrogen electrodes and Hg/HgO/ZrO2(Y2O3) ceramic membrane electrodes, with the potentials of both being measured against a Ag/AgCl, 0.1 m KCl external pressure balanced reference electrode (EPBRE). These electrodes have been used extensively in the past in our laboratory to measure the pH of dilute, high-temperature aqueous solutions (1-5),

but to our knowledge, this study is the first attempt to measure the pH of concentrated solutions at elevated temperatures.

pH Electrodes Pt Bead Electrode. A platinum bead hydrogen electrode was used to sense the pH of a solution at a given hydrogen fugacity (partial pressure). The hydrogen fugacity in the pressure vessel was established using a Pd/25%Ag hydrogen diffuser with the hydrogen fugacity being taken as that on the dry side of the diffuser (Figure 1). In some cases, the diffuser itself was used as the pH sensor, as in our previous work (6,7) and, more recently, that of Nagy and Yonco (8).

Yttria-Stabilized Zirconia (YSZ) Electrode. Our previous work (1-5) on measuring the pH of high-temperature, dilute solutions with Hg/HgO/ZrO2(Y2O3) electrodes initially indicated that this sensor experiences a "boric acid error" [compare the alkaline error in glass electrodes (9)] and hence might not be suitable for measuring the pH of concentrated boric acid solutions. However, difficulties early in this program with the hydrogen diffuser (including cracking of the pressure vessel) led us to reconsider the ceramic membrane sensor, and ultimately most of the data reported here were obtained with this sensor.

YSZ sensors were constructed by filling about one inch of a closed end ZrO2 (9% Y203) tube (5 in. long, 1/4 in. OD and 5/32 in. ID) with a ground mixture of Hg/HgO (Figure 2). A mild steel wire was embedded in this mixture to serve as an electrical contact, and the tube was backfilled with high-temperature epoxy to form a tight seal. The upper part of the sensor was sheathed in heat-shrink polytetrafluoroethylene (PTFE) and then introduced into a standard CONAX fitting.

External Pressure Balanced Reference Electrode (EPBRE). The potentials of the platinum bead electrode and the Hg/HgO/ZrO2(Y203) sensor were measured against a Ag/AgC1, 0.1 m KCl external pressure balanced reference electrode of the type shown in Figure 2 (10). The potential of the reference electrode was corrected to the standard hydrogen scale using the data of Macdonald et al. (10).

Measurement Cell All experiments reported in this work were performed in a 75-mL Type 316 stainless steel pressure vessel equipped with ports for the introduction of the pH sensors and reference electrode, a thermocouple, and a stainless steel capillary tube to a pressure gauge (Figure 4). The vessel was heated with two 900-W thin-cast (WATLOW) heaters controlled by a microprocessor controller (WATLOW Series 910).

12-78

Potential Measurements All potentials were measured using a Keithley 617 electrometer having an input impedance of 2 x 101*

. Measurements were continuously monitored by a Macintosh microcomputer through an IEEE 488 interface.

Solutions All solutions were prepared by weight using doubly distilled water and AR grade chemicals.

Procedure All measurements were performed at 250C intervals within the temperature range 1250 to 3000C.

The potentials of the two pH sensors were continuously monitored at each temperature for 30 minutes or until they had become stable. Generally, the potential remained constant to within + 5 mV over the monitoring period. At the two highest boric acid concentrations, the solutions formed hard glasses when they cooled; these glasses had to be chipped from the pressure vessel. Removal of the glass invariably resulted in destruction of the YSZ sensor and the reference electrode.

Results and Discussion...

Reference Electrodes The potential of the platinum bead hydrogen electrode and that of the Hg/HgO/ZrO2(Y203) pH sensor measured against the EPBRE were corrected to the standard hydrogen scale (SHE) using the empirical relation derived by Macdonald et al. (3,10).

ESHE(T) = BOBS + 0.286637 - 1.003217 x 10-3 AT + 0.0174478 x 10-5 AT 2

- 0.3030048 x 10-8 AT 3 (1) where AT = T - 298.15 and T is the Kelvin temperature. This correction is accurate to about + 10 mV at 2750C but is considered (3,10) somewhat more accurate at lower temperatures L 5 mV at 1000 C).

Because the solutions employed in this study were concentrated, significant isothermal liquid junction potentials were expected to exist across the junction between the internal solution of the reference electrode and the test environment. Therefore, it was necessary to estimate the liquid junction potential so that the measured potential could be accurately referenced to the standard hydrogen scale. As in previous work (4), we estimated the isothermal junction potential using Henderson's equation Ej = -Y-Di[(mi)2 - (mi)l]

(2) 12-79

where the summation is performed over all charged species in the system and subscripts 1 and 2 designate the two sides of the junction. The Dj coefficients are given in terms of the molal concentrations (mj) and equivalent conductances (Xj) by Dj = RTI zI j / (zjrFz zj mj(3) where zj is the ionic charge and mj designates the mean concentration between the two compartments.

Because we did not know the exact composition of the system (although we knew the stoichiometric concentrations), we adopted the following approximations to calculate the isothermal liquid junction potential:

1.

All salts are completely dissociated.

2.

The concentration of borate [B(OH)4] is equal to the stoichiometric concentration of NaOH (1 m).

3.

The concentration of hydrogen ion is equal to 10"pH where the pH is determined iteratively to converge on -log(mH+).

4.

The concentration of OH-is equal to 10 "(pKw-pH) where pKw = -log(Kw) and Kw is the ionic product of water.

Equivalent conductances were not available for B(OH)Y, but experimental data are available for the other species from the work of Quist and Marshall (11). Accordingly, we adopted the following values for Xi:

)H+= -2759.6378 + 17.5151 T - 0.028435 T2 + 1.56794x1005 T3 (4)

XOH- = -929.116 + 3.3085 T - 0.003754 T2 - 7.326785 x 10-6 T 3 (5)

XNa+ = -85.971104 - 1.82398 T + 0.00726322 T2 - 5.1394 x 10-6 T3 (6)

?,HSO 4 = -226.5844 - 2.7298 T + 0.009087 T2 - 6.4037 x 10-6 T3 (7)

XSO2- = -40.616 + 1.4136 T + 0.0076865 T2 -1.3204 x 10-5 T3 (8)

= -150.66099 - 0.493813 T + 0.00554 T2 - 4.396 x 10-6 T3 (9) where T is the Kelvin temperature. The equivalent conductance of B(OH)4 was taken to be equal to that for HSO4 XB(OH)4 = - 226.5844 - 2.7289 T + 0.009097 T2 - 6.4037 x 10-6 T 3 (10) 12-80

Isothermal liquid junction potentials calculated using the data listed above and equation (2) are summarized in Table 2.

Hydrogen Electrode The potential-determining reaction for the hydrogen electrode can be written as H+ + e- <=> 1/2 H2 (11) which yields a Nernst potential of ESHE-Ej = -(2.303RT/2F) log (PH2) - (2.303RT/F)pH (12) where PH2 is the partial pressure (fugacity) of hydrogen. Rearrangement of equation (12) therefore yields pH: 'F(Ei-EsBE)/2.303 RT - 1/2 log (PH2)

(13) where EJ is the liquid junction potential [equation (2)].

YSZ Ceramic Membrane Electrode As in our previous work (3), we write the Nernst equation for the Hg/HgO/ZrO2(Y203) pH sensor in the form ESHE - Ej = Eo - [2.303RT/(2F)]log aH20 - (2.303RT/F)pH (14) where the standard potential for the Hg/HgO internal reference is given by EHg/HgO = 1.0540 - 14.177 x 10-4 T + 9.193 x 10-5 T ln(T/298.15)

+ 3.5x10- 8 T2 -1.7996/T (15) where T is the Kelvin temperature and aH20 is the activity of water. The activity of water is defined as the ratio of the partial pressure of water above the solution to that above pure water at the same temperature. However, we lacked experimental data for aH,so we used values calculated from EPRIrs chemistry code, MULTEQ (Table 3). The accurac~ya these data is unknown.

pH = (F/2.303RT) (EHg/HgO - ESHE + Ej ) - 1/2 log aH2 0 4 (16) 12-81

In calculating the pH using either equation (13) or equation (16), we initialized the iterative procedure by first setting EJ = 0. This value was then used to estimate the liquid junction potential from equation (2), which ini turn was used to calculate a new value of pH. The iterative procedure was repeated4until convergence was achieved to 0.001 units in pH.

Comparison of Measured and Calculated pH Values Figures 5 through 12 summarize the pH data measured in this work and the values calculated by EPRI using MULTEQ. We estimate that the data measured with the Pt/H2 electrodes are probably accurate to +/-0.3 at 125 0C and to +/--0.45 at 3000 C, and those measured with the YSZ electrode are accurate to +/-0.35 and +/-0.5, respectively, at the same temperatures, depending upon the concentration. These uncertainties were established by differentiating equations (13) and (16) to yield SpH = (F/2.303RT) BE + (1/2PH2)8PH2 (17)

SpH = (F/2.303RT) SE + (1/2 aH20 )SaH20 (18) where 5X is the estimated error in parameter X. Typically, the error in the measured potential is

+/-5 mV, but when this value is corrected to the standard hydrogen scale, the error increases to

+/-15 mV; the difference is due to the uncertainty (+/-10 mV) in the thermal liquid junction potential.

The uncertainty in EHgHg.O has been estimated (3) from the thermodynamic data employed to be

+/-5mV. Finally, the error in Ei is difficult to estimate with precision because we do not know the accuracies of the Xi values or now how applicable Henderson's equation is to concentrated solutions of the type employed here. Nevertheless, we estimated BEJ to be on the order of +/-5 mV for the dilute solutions and possibly as high as +/-10 mV for the more concentrated solutions.

Accordingly, we estimated SE to vary between +/-15 mV and +/-25 mV for the Pt/H2 electrodes and between +/-20 mV and +/-30 mV for the YSZ electrodes, with the higheirvalues corresponding to the most concentrated solutions.

The uncertainty 8PH2 is estimated to be +/-10% of the partial pressure, because it was difficult to establish equilibrium across the hydrogen diffuser. Accordingly, we set PBP-2/}-H2 -0.1. A similar uncertainty is estimated for BaH 1aH1O, although we do not have any experimental data for an independent check of this estiMare'. Fod the most concentrated solutions, the uncertainty in the value for 8a%20/aH20 could be considerably higher than 0.1; it could approach +/-0.25 (SaH20

= +/-0. 1, aH20 ~-0.2) for solutions with boric acid concentrations above 10 m.

The data for the two highest boric acid concentrations, particularly for the 100 m B(OH)3 system, are badly scattered. In neither case is the boric acid completely soluble at ambient temperature, but available data (12) show that the solubility increases rapidly as the temperature is raised. However, we had no way to determine whether the solid was completely dissolved at the highest experimental temperature, so it is possible that the system remained in two phases over the entire temperature range. To minimize the effect of incomplete boric acid dissolution, we performed one experiment in which we first heated the solution to 3000C and then measured the pH as the pressure vessel cooled (Figure 11). Good agreement was obtained between the measured and calculated pH at temperatures for 1750C < T < 2750C, but at lower temperatures, the measured pH diverged rapidly in the positive direction from the calculated values. This behavior is consistent with the formation of a glass, which we represent by the reactions 12-82

OH x B(OH)3 -- (-O-B-)x + xH20 (19)

B(OH)4

-+ B(OH)3 + OH (20)

Thus, reaction of the mono boric acid species to form the glass causes the borate/boric acid equilibrium, reaction (19), to shift to the right, which releases hydroxide ion and increases the pH.

We have chosen to write reaction (19) in terms of formation of a "metaborate" glass. The glass that forms is more likely to consist of a mixture of borate chains and sheets formed, by six membered rings. In any event, polymerization of the free boric acid entities will result in the release of hydroxide as described by reaction (20) and hence in the observed increase in pI-L Poor agreement was observed between the pH calculated for the 5 m B(OH)3 solution and that measured with the YSZ electrode, although good agreement was obtained with one of two runs performed with the hydrogen electrode (Figure 8). However, the three runs performed using the ceramic membrane pH sensor gave consistent results that diverged by no more than +/-0.2 pH units but were as much as 2 pH units more positive (alkaline) than the calculated values. We do not currently have an explanitidn for this observation.

Comparison of the measdied and calculated data in Figures 5 through 12 indicates that the YSZ ceramic membrane electrode provides more consistent pH values than does the Pt/H2 electrode.

The large scatter observed in the data obtained with the hydrogen electrode is attributed to the difficulty in establishing an equilibrium partial pressure of hydrogen in the pressure vessel by means of the Pd/25%Ag diffuser. The main problem appears to be the difficulty in preventing hydrogen escaping from the vessel, probably by diffusion through seals and possibly through the vessel walls. Accordingly, the pH values calculated from the measured voltage and the assumed partial pressure of hydrogen using equation (13) could be greatly in error. On the other hand, the pH values calculated from the YSZ sensor are highly susceptible to uncertainty in E- (as is the hydrogen electrode), so the level of agreement between calculated and measured pH values found in this work is probably as good as can be expected. We should note that the calculated values themselves could also contain significant error because of uncertainties in equilibrium constants, poor definition of ion-pairing phenomena, and uncertainties in the activity coefficients of ionic and neutral species. Finally, both the calculated and experimental data show that the pH of the concentrated boric acid solution is quite low (pH < 3 for x > 50), so B(OH)3 concentrated into crevices by boiling in operating steam generators is expected to form highly acidic and possibly aggressive environments.

Summary and Conclusions Pt/H2 and Hg/HgO/ZrO2(Y203) ceramic membrane electrodes have been used to measure pH versus temperature data for 1 m NaOH, 1 m NaOH + 1 m Na2SO4, 1 m NaOH + 1 m Na2SO4 +

x m B(OH)3 (x = 2, 5, 20, 50, 100), and 1 m NaOH+I m Na2SO4+ 1 m NaCi at temperatures between 1250 and 3000C. These pH values were compared with data calculated using EPRrs chemistry code, MULTEQ. Allowing for the considerable uncertainties that exist in the isothermal liquid junction potential, the pressure of hydrogen in the pressure vessel (established by using a Pd/25%Ag diffuser), and the activity of water, particularly for the more concentrated solutions, the measured and calculated pH values agree reasonably well. Those measured using the YSZ 12-83

electrode were more consistent than those obtained using the hydrogen electrode. Even though considerable uncertainties may exist in the calculated values, because of errors in measured dissociation constants and calculated activity coefficients, and because of the poor definition of ion pairing phenomena, the generally reasonable agreement observed in this study between the calculated andiexperimental pH values greatly increases our confidencein the ability of MULTEQ to model the chemistry of high-temperature concentrated solutions of the type that exist in crevices in pressurized water reactor (PWR) steam generators. The concentrated boric acid solutions were calculated and observed to yield pH < 3 for x > 50.

Acknowledgement The authors gratefully acknowledge the assistance and advice of Mr. Robert Emerson and Dr.

Samson Hettiarachchi in the experimental part of this program. We also gratefully acknowledge the financial support of this work by the Electric Power Research Institute, Palo Alto, California.

References

1.

S. Hettiarachchi, P. Kedzierzawski, and D. D. Macdonald, J. Electrochem.

Soc., 132, 1866 (1985).

2.

S. Hettiarachchi and D. D. Macdonald, J. Electrochem. Soc., 131, 2206 (1984).

3.

D. D. Macdonald, S. Hettiarachchi, and S J. Lenhart, J. Solution Chem.,

17, 719 (1988).

4.

D. D. Macdonald, P. Butler, and D. Owen, Can. J. Chem.,-,1, 2590 (1973).

5.

D. D. Macdonald and D. Owen, Can. J. Chem., 51, 27471(1973).

S6.

D. D. Macdonald, P. R. Wentrcek, and A. C. Scott, J. Electrochem. Soc.,

127, 1745 (1980).

7.

D. D. Macdonald, C. Scott, and P. R. Wentrcek, unpublished data, Project RP1168-1, Electric Power Research Institute, Palo Alto, CA, 1978/1979.

8.

Z. Nagy and R. M. Yonco, J. Electrochem. Soc., 133, 2232 (1986).

9.

Hubert T.S. Britton, Hydrogen Ions, Vol. 1, Chap. 7 (D. Van Nostrand Company, Inc., London, 1955).

10.

D. D. Macdonald, A. C. Scott, and P. R. Wentrcek, J. Electrochem.

Soc., 126, 908 (1979).

11.

A. S. Quist and W. L. Marshall, J. Phys. Chem., 69, 2984, (1965).

12.

H. Stephen and T. Stephen, editors, Solubility, Vol. 1 (Macmillan, New York, 1963).

12-84

Table 1 TEST SOLUTIONS Solution No.

Comnositdon No. of Tests 1

I m NaOH 2

2....

1 m NaOH + I m Na2SO4 2

3.

1 m NaOH + I m Na2SO4 + 2 m B(OH)3 2

4 1 m NaOH + I m Na2SO4 + 5 m B(OH)3 5

5 1 m NaOH + I m Na2SO4 + 20 m B(OH)3 3

6 lm NaOH +lm Na2SO4 + 50 m B(OH)3 1

7 1 mNaOH + I mNa2SO4 + 100m B(OH)3 2

8 I m NaOH + I m Na2SO4 + I m NaC 2

Table 2 ESTIMATED LIQUID JUNCTION POTETALS (mY)

Temperature (OC S* uQ 125 150 175 200 225 250 275 300 1

15.6 15.8 15.6 15.2 14.8 14.2 13.3 12.3 2

10.7 10.3 9.7 9.0 8.1 7.0 5.6 3.8 3

9.4 9.5 9.6 9.7 9.6 9.4 9.2 8.8 4

8.5 9.0 9.5 10.0 10.5 10.9 11.1 11.2 5

7.5 8.4 9.4 10.4 11.5 12.5 13.6 14.7 6

7.2 8.2 9.3 10.5 11.8 13.0 14.3 15.6 7

7.0 8.2 9.3 10.6 11.9 13.2 14.6 16.0 8

7.1 6.8 6.4 6.0 5.4 4.7 3.9 2.9 Table 3 ACTIVITY OF WATER Temperature (OQ Solution No. 125 150 175 200 225 250 275 300 1

0.968 0.968 0.968 0.970 0.971 0.972 0.973 0.975 2

0.930 0.932 0.934 0.936 0.939 0.942 0.946 0.952 3

0.930 0.931 0.933 0.936 0.939 0.943 0.949 0.958 4

0.894 0.894 0.895 0.895 0.895 0.895 0.896 0.898 5

0.670 0.667 0.663 0.657 0.650 0.642 0.633 0.625 6

0.375 0.368 0.360 0.349 0.336 0.322 0.307 0.394 7

0.142 0.135 0.128 0.119 0.109 0.099 0.089 0.081 8

0.894 0.897 0.900 0.904 0.908 0.913 0.919 0.928 Rearrangement of equation (14) therefore yields the pH as 12-85

Pd/25%Ag Tube (0.005 In. wall)

Packed will T

Sand

.05 In.

tI; Silver Braze Tube 1.0 In.

Ferrule H2 Fligm 1.

Design of Pdf25%A Hydrogen Diffuser Inside End Chloridized Silver Wire (Ag/AgCI)

-Teflon Insulating Sheath

-Swagelok Fitting Cu Tubing for Air or Water Cooling o

Ag/AgCI Element Teflon Tubing 316 SS Tubing Sur the Electrolyte-Co Teflon Tube

-Swagelok Fitting 0.1 M KCI Solution Teflon Tube Teflon Insert

.-"1 Heat Shrink Teflon ZrO2 Plug RA-M-20583-13 Figure 2.

Yttria-Stabilized Zirconia pH Sensor Assembly rounding ntaining H2 PI Wire or YSZ Electrode RA-6117-2 R"-x20583-1 I A Figure 3.

External Pressure-Balanced Reference Electrode Assembly Figure 4.

Schematic of test cell showing electrode configuration for measuring the pH of concentrated solutions at elevated temperatures.

12-86 h

r



14 13 12 11 10 9

8 7

100 150 200 250 300 TEMPERATURE ("C) 14 13 12 11 pH 10 9

8 7

350 400 RA-6 117-3 Figure 5.

Plots of pH versus temperature for 1 m NaOH solution.

(1)-a: Measured using Hg/HgO/ZrO2(Y20 3 ) electrode.

(1)-b: Calculated using MULTEO.

100 150 200 250 300 TEMPERATURE (-C) 350 400 RA-6117-4 Figure 6.

Plots of pH versus temperature for I m NaOH +

1 m Na2SO4 solution.

(2)-a: Measured using HgHgO/ZrO 2(Y203) electrode.

(2)-b: Calculated using MULTEG.

14 13 12 11 pH 10 9

8 7

100 150 200 250 300 TEMPERATURE (-C) 350 8

7 6

pH 5 4

34 400 RA-6117-5 Figure 7.

Plots of pH versus temperature for 1 m NaOH +

1 m Na2SO4 + 2 m B (OH)3 solution.

(3)-a: Measured using Hg/HgO/ZrO2(Y203) electrode.

(3)-b: Calculated using MULTEG.

7 6

5 pH 4 3

2 1

100 150 200 250 300 TEMPERATURE (OCI 100 i

I 1

150 200 250 300 350 400 TEMPERATURE (-0)

RA-6117-6 Figure 8.

Plots of pH versus temperature for 1 m NaOH +

1 m Na 2 SO4 + 5 m B(OH)3 solution.

(4)-a: Measured using HgHgO/'Z0 2 (Y20 3) electrode.

(4)-b: Calculated using MULTEC.

(4)-c: Measured using Pt/H2 electrode.

7 6

5 4

pH 3

2 1

0 350 400 RA-6117-7 Figure 9.

Plots of pH versus temperature for 1 m NaOH +

1 m Na 2 SO4 + 20 m B(OH)3 solution.

(5)-a: Measured using Hg/HgfZrO 2(Y20 3) electrode.

(5)-b: Calculated using MULTEQ.

(5)-c: Measured using Pt/H2 electrode.

100 150 200 250 t300 TEMPERATURE (o0) 350 400 RA-6117-8 Figure 10.

Plots of pH versus temperature for 1 m NaOH +

I m Na2SO4 + 50 m B(OH) 3 solution.

(6)-a: Measured using Hg/HgOfZrO 2(Y203) electrode.

(6)-b: Calculated using MULTEO.

12-87 pH E).

(3)-a

-*.i (3)-b e~(4)-a

• (4)-b

• (4)-C o (6)-a (6)-b N

I 4

pH 7

6 5

4 3

2 1

0 100 150 200 250 300 TEMPERATURE (-C) 350 400 RA-6117-9 Figure 11.

Plots of pH versus temperature for I m NaOH +

1 m Na2SO4 + 100 m B(OH)3 solution.

(7)-a: Measured using Ho/HgO/ZrO2(Y203) electrode.

(7)-b: Calculated using MULTEO.

DISCUSSION Presenting Author: D. Macdonald Questioner: P. Andresen, GE, R&D Question/Comments: In most of your solutions above -10m, the activity of water plays an important role. Can you comment on (1) the possible circular error in your use of MULTEQ calculated activity of water and (2) loss of water to the vapor space or through the seals, etc. of your autoclave.

Reply: (1) You are quite correct. However, the water activities predicted by MULTEQ are typical of concentrated solutions so that any error is within the limits of uncertainties claimed.

(2) We minimized the vapor space as much as possible to minimize the error. Furthermore, water is strongly bound to borate (as indicated by the water activity) so that the error from this source is probably not too large.

DISCUSSION Presenting Author: D. Macdonald Questioner: J. Atkinson, CEGB-CERL Question/Comments:

Can the yttria stablized zirconia electrode be manufactured in a miniature form suitable for insertion into cracks or crevices.

Reply:

We have tried to make miniature electrodes but with little success so far. I believe that the development of such electrodes should be an activity of high priority.

14 13 12 11 pH 10 9

8 7

100 150 200 250 300 TEMPERATURE (°C) 350 400 RA-6117-10 Figure 12.

Plots of pH versus temperature for I m NaOH +

I m Na2SO4 + 1 m NaCI solution.

(8)-a: Measured using HgH-gO/ZrO2(Y2Oa) electrode.

(8)-b: Calculated using MULTEQ.

DISCUSSION Presenting Author: P. Paine Questioner: T. Beineke, Combustion Engineering, Inc.

Question/Comment:

How many cycles of concentration iwere used in the MULTEQ calculations?

Reply:

Since thQJiboratory system is a water solid system no-concentration factors were used in these calcdIlati6ns.

Dr. Beineke correctly implies that With the boric acid volatility from a crevice, MULTEQ or any other computer code would be hard pressed at the present to calculate the SRI measured pHs.

DISCUSSION Presenting Author: D. Macdonald Questioner: P. Gonzalez, Ontario Hydro Question/Comments:

The-I3 B03 *H that was measured was quite low, 2-5. How could this be related with the apparent inhibiting effect of H3 BO3 on the denting process?

Reply:

I believe that inhibition of steel in concentrate boric acid is due to the formation of a passivating compound between iron and borate, although to my knowledge this has never been confirmed.

12-88 a (8)-a

• (8)-b

DISCUSSION Presenting Author: P. Paine Questioner: J. Gorman,. Dominion Engineering, Inc.

Question/Comment:

What concentrations of boric acid can be expected to occur in steam generator crevices?

Reply: Equilibrium boric acid concentrations will depend on crevice pH (preserice of NaOH and KOH as well as acidic species), boric acid volatility, reactions to produce precipitated Na and Fe borates, bulk water boric acid concentrations feeding the crevices and locally available crevice superheat (T, - T,). We don't expect a uniform concentration of boric acid to form in the crevice, but measurements have been made of solid borate compounds by several techniques indicating that solubility was exceeded on a local basis. Corrosion rates have46en reduced in laboratory tests to levels that imply. a pH reduction of 3 pH units ijt the presence of concentrated NaOH.

/

2 12-89