Presentation on theme: "The Electrochemistry of Stress Corrosion Cracking in Water-Cooled Reactor Coolant Circuits Digby D. Macdonald Center for Electrochemical Science and Technology."— Presentation transcript:
The Electrochemistry of Stress Corrosion Cracking in Water-Cooled Reactor Coolant Circuits Digby D. Macdonald Center for Electrochemical Science and Technology Pennsylvania State University University Park, PA Presented at OLI Systems Inc Morris Plains, NJ October 23, 2007
Background Localized corrosion is characterized by non-homogeneous, quasi- stationary anodic and cathodic partial current densities on a surface. The anodic and cathodic regions are spatially separated and are coupled according to the Differential Aeration Hypothesis (U. R. Evans, circa 1927). The anode occurs in that region of the system that has minimal access to the cathodic depolarizer (e.g., O 2 ), whereas the cathode occurs in that region that has the greatest access to the depolarizer. In order to maintain charge balance a positive current flows through the solution from the local anode to the local cathode, where it is consumed by the reduction of the depolarizer. In the author’s opinion, the key to understanding and predicting stress corrosion cracking lies in developing an understanding of the origin and properties of the coupling current. These principles have been applied to predicting the accumulation of damage due to Intergranular Stress Corrosion Cracking (IGSCC) in sensitized Type 304 SS in BWR primary coolant circuits.
Figure 1. Coupling of crack internal and external environments.
The Coupling Current Recognizes the important (and perhaps) dominant role played by the external surfaces in SCC. Can be readily measured with great accuracy. The noise contained in the coupling current enables direct interrogation of the processes that occur at the crack tip. To the authors knowledge, this is the only method that has been devised for direct interrogation. Has provided unprecedented information of crack tip dynamics in a variety of systems. Difficult to understand why the technique is not more widely used.
E tip Z tip R in ZCZC R crevice ECEC R soln CRACK TIP CRACK FLANK CRACK MOUTH Crack Equivalent Electrical Circuit
Figure 2. Schematic of specimen used for measuring the coupling current. Note that the specimen, except for the crack, is electrically isolated from the environment by a PTFE coating so as to inhibit oxygen reduction from occurring on the surface. The electron current generated at the crack tip is then forced to flow through the ZRA to the external cathodes where oxygen reduction does occur.
Figure 3. Coupling current and stress intensity versus time for Type 304 SS in simulated BWR coolant at 250 o C. The specimen was equipped with one platinized nickel side cathode (see Figure 2).
Figure 4. Typical form of the noise in the coupling current for Type 304 SS with two Type 304 SS side cathodes in simulated BWR coolant at 250 o C and at a stress intensity of 27.5 MPa m. Stainless steel cathodes.
Frequency of micro-fracture events versus stress intensity for IGSCC in sensitized Type 304 SS in water at 250 o C. K ISCC
Proposed fracture mechanism, in which the crack advances in a series of brittle, micro fracture events at the crack front.
Brittle Micro Fracture Event Size f = Frequency of brittle micro fracture event (~ 2 s-1) B = Width of specimen (1.27 cm) r = Radius of brittle micro fracture event. ~ 2-3μm Given a grain size of 20 – 50 μm, there should be roughly 4 to 10 events in each package, which is in good agreement with experiment. The remarkable conclusion is that, in this case, the crack advances fracture event-by-fracture event with minimal overlap between events.
Coupling Current Assume that the current from any given micro fracture event decays as The total current due to all of the micro fracture events that have occurred from zero time until the time at which the crack length is L is given by To a good approximation, the coupling current can be simplified to Setting L = cm, B = 1.27 cm, r = 3 x cm, and i 0 and t f equal to 1-10 A/cm 2 and 0.01 s (Ford, et.al.), respectively, the coupling current is calculated to be 68 – 683 μA, which is also in good agreement with experiment.
The Critical Potential A critical potential exists below which IGSCC does not occur to any significant extent. The critical potential is a function of a number of variables: - Degree of sensitization - pH - Temperature - Solution composition - Strain rate - Crack length For ECP < E IGSCC, failure occurs by mechanical overload, except in acidic solutions where cracking occurs by a hydrogen mechanism.
Figure 1: Measured and calculated (via the CEFM) crack growth rates for sensitized Type 304 SS in high temperature aqueous solutions as a function of ECP and conductivity. The citations refer to references in the original source . 1.5 mm/yr 0.06 mm/yr
Figure 3: Calculated crack growth rate and coupling current for IGSCC in sensitized (DOS = 15 C/cm 2 ) Type 304 SS in dilute NaCl solution (0.135 ppm Na) at 288 o C as a function of potential of the steel at the external surfaces remote from the crack mouth and as modeled using the Coupled Environment Fracture Model [3-5]. K I = 27 MPa m, crack length = 0.5 cm, crack width = 1.0 cm, crack mouth opening displacement = 5x10 -4 cm, solution flow velocity = 100 cm/s, hydrodynamic diameter = 50 cm, 25 = S/cm, 288 = 6.89 S/cm, and pH 288 = 5.89.
Figure 5: Relationship between the crack growth rate and coupling current for IGSCC in sensitized Type 304 SS in oxygenated (7.6 ppm, 2.38 x 10-4 m) sodium chloride (50 ppm, 8.62 x 10-4 m) solution at 250 oC.
Impact of Crack Length Necessary to differentiate between mechanical crack length (MCL) and electrochemical crack length (ECL). The former is the distance between the crack tip and the loading line, while the latter is the shortest distance between the crack tip and the external surface. CEFM predicts that the CGR decreases as the crack length increases because of the increase in the IR potential drop down the crack. CEFM also predicts that the critical potential shifts in the positive direction with increasing crack length. Predictions difficult to test using fracture mechanics specimen, because the ECL is constant. Predictions are confirmed by indirect evidence.
Application to Practical Systems Boiling Water Nuclear Reactors. IGSCC in weld sensitized Type 304 SS in the primary coolant circuit. Problem has cost the utilities hundreds of millions to billions of dollars through lost production. Damage can now be predicted using the CEFM. Fourteen reactors modeled to date.
Where and When Does IGSCC Occur n SS Heat Transport Components – Ex-Vessel – In-Vessel n Three Factors Together – High Stress – An Aggressive Environment – A Susceptible Material
Flow path of the primary coolant circuit of a BWR having external coolant pumps (GE Model VI).
Connections n How all the topics fit together
Methodology Radiolysis of Water in BWR Heat Transport Circuit –Chemical/ionic composition of bulk water Electrochemical Corrosion Potential –Charge/mass transfer at metal/water interface Crack Growth Rates in Components –Electrochemical Dissolution Of Metal
Radiolysis of Water in the Heat Transport Circuit Balance of Species: LOSS + GAIN = 0 –Radiolytic Yield –Chemical Kinetics (Reactants & Products) –Boiling (Liquid and Gas Exchange) –Convection
Electrochemical Corrosion Potential (ECP) R/O Reactions at Metal/Solution Interface Charge and Mass Transfer Conservation of Charge: Net Current is Zero Mixed Potential, ECP
ALERT Code Developed from an earlier code DAMAGE- PREDICTOR that predicted only state points properties. ALERT predicts both state point properties and damage integrated over a specified Corrosion Evolutionary Path (CEP). ALERT and its companion code for BWRs with internal pumps, REMAIN, are the only codes in existence, to the author’s knowledge, that predict integrated damage. Codes currently lack an adequate model for crack nucleation.
Two Types of Calculations State points defined for a given set of conditions - [O 2 ], [H 2 O 2 ], [H 2 ], T, pH, flow velocity, stress intensity factor, crack length, etc. Provides a “snap-shot” in time. IGSCC damage (crack length) for a specific point in the coolant circuit integrated along the corrosion evolutionary path.
“Double blind” test of ECP prediction at Leibstadt in Switzerland
Corrosion Potential – Leibstadt, Switzerland
Crack Growth Rate – Leibstadt, Switzerland
Case Study GE Model 6 BWR. Path to future state specified. Normal Water Chemistry / Hydrogen Water Chemistry. Catalytic Coatings or Inhibiting Dielectric Films (e.g., ZrO 2 ). Calculates Radiolysis Product Concentrations, ECP, Crack Growth Rate, and Accumulated Damage over the Plant Evolutionary Path. Modeled a crack on the inside of the core shroud adjacent to the H-3 weld near the top. ALERT takes into account the dependence of crack growth rate on crack length. First and only instance where a prediction of this type has been made by anyone.
Input File: State Point Data Core_Channel Core_Bypass Upper_Plenum Mixing_Plenum Upper_Downcomer Lower_Downcomer Recirculation Jet_Pump Bottom_of_the_Lower_Plenum Top_of_the_Lower_Plenum E E E E E E E E E E E E E E E E E E E E E E
Output: Chemistry, ECP, CGR
Crack Growth Rate
Corrosion Evolutionary Path
Plant Evolutionary Path, cont’d
Accumulated Damage, Core Shroud
Comparison with plant data – BWR in Taiwan
What Can Be Done To Reduce IGSCC Damage in BWRs? Enhance the rate of the hydrogen electrode reaction that occurs on the external surface (Noble Metal Chemical Addition ® – introduced by GE to enhance HWC). Inhibit redox reactions (e.g., O 2 and H 2 O 2 reduction) on external surface. Use more resistant alloy (Type 304K SS instead of Type 304 SS) – applicable to new plants. Catalytic decomposition of H 2 O 2 – shown by ALERT to help but not to be totally effective. Ultra low conductivity operation – the required conductivity probably cannot be achieved in an operating plant.
Figure 4: ECP variation in the heat transport circuit of a BWR employing general catalysis (exchange current density multiplier 10 4 ) under NWC ([H 2 ] FW =0.0 ppm) and HWC ([H 2 ] FW =1.0 ppm) conditions. Note that the multiplier employed in the calculation is considered to be extreme and probably could not be achieved with deposited noble metal coatings. The value is used for illustrative purposes only.
Figure 5: Crack growth rate variation in the heat transport circuit of a BWR employing general catalysis (exchange current density multiplier 10 4 ) under NWC ([H 2 ] FW =0.0 ppm) and HWC ([H 2 ] FW =1.0 ppm) conditions. Note that the multiplier employed in the calculation is considered to be extreme and probably could not be achieved with deposited noble metal coatings. The value is used for illustrative purposes only.
Figure 6: ECP variation in the heat transport circuit of a BWR employing general inhibition (exchange current density multiplier ) under NWC ([H 2 ] FW =0.0 ppm) and HWC ([H 2 ] FW =1.0 ppm) conditions. Note that the multiplier employed in the calculation is arbitrarily chosen and represents a highly resistive dielectric film.
Figure 7: Crack growth rate variation in the heat transport circuit of a BWR employing general inhibition (exchange current density multiplier ) under NWC ([H 2 ] FW =0.0 ppm) and HWC ([H 2 ] FW =1.0 ppm). Note that the multiplier employed in the calculation is arbitrarily chosen and represents a highly resistive dielectric film.
Figure 8: Calculated ECP and IGSCC crack growth rate for sensitized Type 304 SS according to the MPM and CEFM, respectively, corresponding to the experimental conditions summarized in Figure 9. Other model parameters are given in Ref. .
Figure 9: Inhibition of IGSCC in Type 304SS by a dielectric ZrO 2 coating on the specimen external surfaces. The conductivity of the solution was 220 μS/cm at 25 o C and the solution was saturated with O 2 at ambient temperature ([O 2 ] = 40 ppm). K I = 25 MPa√m, T = 288 o C. The crack growth rate for the uncoated specimen is 4x10 -7 cm/s while that for the coated specimen is < 2x10 -8 cm/s.
SUMMARY AND CONCLUSIONS Intergranular Stress Corrosion Cracking (IGSCC) is primarily electrochemical in nature as are the most effective mitigation measures. The coupling current provides a wealth of information concerning events that occur at the crack tip. The critical potential for IGSCC in sensitized Type 304 SS in high temperature aqueous solutions can be accounted for in terms of the dependence of the coupling current on the potential at the external surface remote from the crack mouth. The critical potential depends on the crack length, with E IGSCC becoming more positive with increasing length. All cracks must eventually die when E IGSCC > ECP.
Summary and Conclusions (cont.) Shifting the ECP to a sufficiently negative value reduces the coupling current and hence the IGSCC growth rate to the extent that it becomes negligible, signifying the critical potential. The coupling current and IGSCC growth rate may be reduced by decreasing the charge transfer impedance at the external surface provided that the molar ratio of H 2 to O 2 is greater than 2 (as achieved in hydrogen water chemistry), by increasing the charge transfer impedance at the external surface regardless of the molar ratio of H 2 to O 2, or by increasing the resistivity of the external environment (lowering the conductivity). These strategies are a direct manifestation of the fact that stress corrosion cracking falls within the framework of the differential aeration hypothesis, which requires strong coupling between the crack internal and external environments via the coupling current.
Summary and Conclusions (cont.) The CEFM has been combined with the radiolysis model, RADIOCHEM, and with the Mixed Potential Model, MPM, to produce powerful BWR simulation codes, DAMAGE- PREDICTOR, ALERT, and REMAIN. These codes predict coolant chemistry, ECP, and state point crack growth rate for any location in the coolant circuit and, in the case of ALERT and REMAIN, also predict accumulated damage along the Corrosion Evolutionary Path (CEP). Comparison of the predictions with plant data from the Leibstadt BWR (Switzerland) and from a BWR in Taiwan reveals excellent agreement. An even more advanced BWR code is under development, We have developed a somewhat similar code for predicting the electrochemistry of Pressurized Water Reactor (PWR) primary coolant circuits.
ACKNOWLEDGEMTS The author gratefully acknowledges the contributions made to this work by the following: Professor Mirna Urquidi-Macdonald, Pennsylvania State University. Dr. George R. Engelhardt, OLI Systems Inc and Adjunct Professor, Pennsylvania State University. Dr. T.-K. Yeh, National Tsing-Hua University, Taiwan, former student. Dr. Iouri Balachov, SRI International, Menlo Park, CA. Amit Jian, former MS student at Penn State. HanSang Kim, Penn State, current Ph.D. student. Department of Energy and SRI International for financial support.