ML031530397

File from s Long Computer Regarding Boric Acid Corrosion of Vessel Pressure Boundary

ML031530397
Person / Time
Site:	Davis Besse
Issue date:	10/15/2002
From:	Long S - No Known Affiliation
To:	Office of Nuclear Reactor Regulation
References
FOIA/PA-2003-0018
Download: ML031530397 (19)

Text

rile IrurIn %WeveLong s computer aated 1U/15/02 0 8:27am named "Degradation of Vessel Head Report.doc" Boric Acid Corrosion of Vessel Pressure Boundary

1.0 BACKGROUND

In early March, 2002, during performance of inspections in response to the NRC Bulletin 2001-01, "Circumferential Cracking of Reactor Pressure Vessel Head Penetration Nozzles", significant degradation of the RPV top head base metal was discovered at the Davis-Besse (D-B) nuclear power plant between nozzles #3 and #11, and some minor corrosion at Nozzle #2. Downhill of Nozzle #3, a roughly triangular cavity with dimensions of about five inches width, seven inches length and completely through the low-alloy steel head thickness (about seven inches), had been created.

Between forty and sixty cubic inches of metal had corroded and been flushed from the cavity, leaving only a layer of cladding about 0.3 inches thick, with an exposed surface area of between 20 and 25 square inches. Although primary water stress corrosion cracking (PWSCC) of Alloy 600 CRDM nozzles and wastage of carbon and low-alloy steels by boric acid corrosion are known degradation mechanisms, and have been observed at other nuclear plants, the extent of corrosion wastage from the Davis-Besse RPV head had not been anticipated.

In the other instances of through wall cracking of CRDM nozzles that have occurred the total leakage from the crack into the annulus appears to have been very low. The Davis-Bessie experience demonstrates that this is not always the case. It is important to understand the conditions that can result in this aggressive attack. The critical issue is why the leaking CRDM Nozzle #3 at Davis-Besse progressed to high leak rates resulting in substantial RPV head wastage while at several other plants of similar design, leaks have remained small and damage inconsequential.

2.0 PURPOSE This report has been compiled in order to serve several purposes:

(a) To collect the literature references and other reporting citations of information on corrosion of reactor vessel head components (low-alloy steel, Alloy 600 and 308/309 cladding stainlesses)

(b) To tabulate the description and quantitative results of corrosion experiments that bear a reasonable similitude to the Davis-Besse vessel head degradation scenario, and to distill these into a range of rates, for the various phases of the degradation, that seem reasonable, given the time, temperature and geometry of the situation (c) To identify the research that would be helpful in further elucidation of the processes that have taken place at Davis-Besse, and, (d) To present a summary of the research being undertaken by the Materials Engineering Branch to support public safety and regulatory concerns about the potential for other incidents of boric acid-produced wastage of low-alloy steel vessel heads.

3.0

SUMMARY

OF ROOT CAUSE FINDINGS ON CAVITY DEVELOPMENT About a month-and-a-half following discovery of the cavity, the licensee issued a root cause report describing the background and events leading up to the discovery, and describing a host of "contributing causes" that resulted in the CRDM leaks and the ensuing vessel head corrosion. The Davis Besse root cause analysis report suggests the following scenario for the degradation of RPV head base metal:

1. Crack initiation and growth to through wall: The report postulates that a crack initiated in Nozzle #3 at around 1990 (+/-3 y) due to PWSCC. The crack grew to a through wall crack that penetrated above the J-groove weld in the 1994 to 1996 time period. At this stage, the report v

Boric Acid Corrosion of Vessel Pressure Boundary hypothesizes that the extent of through wall cracking was very limited and the reactor coolant system (RCS) leakage would have been extremely small.

2. Minor weepage / latency period: As the crack grew, leakage would have entered the annular region between the Alloy 600 nozzle and low-alloy steel RPV head. With addition of moist boric acid from the newly developed crack into the bi-metallic annulus, various corrosion and concentration processes, including galvanic attack, are possible. The report proposes that these corrosion processes would open the annular gap, although it could alternatively be argued that corrosion products and insoluble precipitation products like iron metaborate or nickel iron borate could plug the gap and reduce the leakage to very low levels. At this stage, a low level of leakage from the annulus could manifest itself as the classic "popcorn" crust of boric acid deposits. In contrast to other plants with leaking nozzles, the boric acid deposits on top of the D-B RPV head accruing from leakage from CRDM flange joints, could have acted as an "incubator", wherein leaking borated water is retained under the deposits. The identity of the boric acid species within the annular enclave is speculative; it could have ranged from aqueous, concentrated solutions of boric acid to molten mixtures of boric acid and boric oxide. The oxygen content of the solution is presumably small, due to the limited access through the annular gap, coupled with the probable egress of superheated steam through the same gap, and an uphill pressure gradient.

3. Late latency period: As the crack continued to grow, the root cause report assumes that the annular gap increased in width, and that because the growth in annulus width occured over about half of the circumference of the CRDM, the annulus flow area increased faster than the crack flow area.

The root cause report ignores any potential plugging of the annulus by corrosion products and insoluble precipitates, thus ensuring that the primary flow resistance would have been due to the dimensions of the crack, and not due to any restriction offered by the annulus geometry. Under these conditions, oxygen may have entered the annulus. If that happened, wastage rates would have been expected to increase dramatically.

4. Deep annulus corrosive attack: In the scenario envisaged in the root cause report, continued widening of the annular gap would cause the velocity of flow out of the annulus, and the differential-pressure, to decrease, allowing greater penetration of oxygen and increased corrosion rates. The root cause report suggests that corrosion is likely to be greater in the vicinity of the crack because leakage through the crack would maintain a fresh supply of new reactive oxidizing ions in the boundary layer near the corroding metallic surface.

5. Boric acid corrosion: With high leakage rates, the annulus filled with an increasing amount of moist steam, mostly (about 80%) flashing as it exited. Heat transfer from the surrounding metal was no longer sufficient to immediately vaporize the portion of leakage that did not flash. The metal surface temperature was suppressed by the cooling effect of the large heat flux required to vaporize the leaking coolant. This effect allowed greater area to be wetted underneath the accumulations of boric acid. As the crack grew, and the leak rate from the crack increased, the corroding annulus began to fill up with a saturated boric acid solution. Because the wetted area would be the result of liquid flow, it would be expected to be predominantly downhill from the nozzle. This would result in high corrosion rates and wastage of RPV head material on the downhill side of the nozzle.

Although it is not possible at present to establish the exact progression of mechanisms that led to the observed RPV head wastage, the degradation modes on the two extremes of the overall progression may be described with reasonable confidence. At extremely small leak rates (106 to 10-5 gpm),

observed in most of the leaking CRDM nozzles, the leaking flow completely vaporizes to steam immediately downstream from the principal flashing location. This results in a dry annulus and no loss of material. The other extreme is associated with the classic boric acid corrosion mechanism caused by liquid boric acid solution concentrated through boiling and enhanced by oxygen directly available Page 2 of 19

Boric Acid Corrosion of Vessel Pressure Boundary from the ambient atmosphere. The extent of the boiling heat trarisfer associated with the relatively high leak rate of Nozzle #3 was sufficient to cool the head enough to allow liquid solution to cover the walls of the cavity. It is clear that relatively high leakage rates from CRDM cracks are necessary for such catastrophic corrosion.

The root cause report is incomplete in many regards, partially because much of the data necessary to support the hypotheses simply doesn't exist. Wastage of low alloy steel in molten boric acid species, or in concentrated, aqueous solutions is not well-described or quantified in the literature, and especially not under the temperature, flow or stirring rates, and concentration of species that may have been attained on the Davis-Besse head. The electrochemical potentials of the alloys and aqueous solutions involved are not known. Crack initiation times may have been short, and the stress-corrosion crack growth rate for the Alloy 182 in the J-groove weld and the Alloy 600 in the CRDM nozzles may have been atypically high, due perhaps to the thermomechanical processing of these materials. In short, when this issue was publicized on March 8, 2002, there were many more questions than the licensee, the industry or the NRC had answers.

4.0 RELEVANT REFERENCES A list of references has been appended to this report. The list is not intended to be exhaustive; the EPRI Guidebook - Revision 1 contains a more expanded selection of references that were available through its publication in late 2001. EPRI is considering a proposal for a second revision to the guidebook, incorporating the information that has emerged during the dispositioning of the Davis-Besse cavity.

5.0 PRIOR BORIC ACID CORROSION EVENTS Boric acid solution from small leaks onto the reactor vessel head during full power operation will have the water flash to steam, leaving behind white, popcorn-like boric acid crystals. Being readily soluable, these boric acid crystals are relatively easy to remove from the head during the next refueling outage. Any boric acid crystals that are not removed from the head prior to returning to full power operation can be converted to boric oxide crystals above a temperature of 302°C:

2 H2 BO 3 -* B 2 03 + 3H 2 0 The final condition of the mixture of boric acid and boric oxide is site specific, depending on the relative quantities of each compound, the amount of flow of additional boric acid solution, the rate of evaporation, the porosity caused by escaping steam, and the presence of impurities such as iron oxide. Any boric acid that is not converted to the oxide may become a viscous fluid when heated above 365°F and will conform to the surrounding geometry under the influence of gravity. Molten boric acid can contain 8 to 14% water and can be highly corrosive under some conditions.

Discussions between the NRC staff and the Brookhaven National Laboratory staff have revealed that the boric acid/boric oxide mixture can vitrify if concentrated sufficiently and if held at a sufficiently high temperature for an extended period of time.

Boric acid induced corrosion wastage of carbon steel and low alloy steel is a phenomenon that has been observed in power plants for at least 30 years. In the majority of cases of boric acid corrosion in nuclear power plants, boric acid leaked from flanged joints and resulted in damage to threaded fasteners. These cases involved primarily valves and pumps, and also manway closure studs. In a small number of cases, however, leakage at reactor pressure vessel head penetrations led directly to corrosive attack of the vessel head. In none of these pre-2002 cases did the corrosion wastage progress to a point at which the structural integrity of the vessel head was threatened. Several Page 3 of 19

Boric Acid Corrosion of Vessel Pressure Boundary relevant instances of boric acid corrosion attack that have been reported are summarized in the paragraphs to follow.

A number of cases of boric acid corrosion of manway closure studs occurred in the late 1970's and early 1980's. Probably the most severe bolting material wastage corrosion occurred at Fort Calhoun in 1980. Three of the four reactor coolant pumps (RCP) at the plant exhibited leakage at the pump cover I case interface. Each cover was secured with 16 studs, 3.5 inches in diameter, which were made of ASTM A 193 Grade B7 material with chrome plated threads. Several studs on each pump were affected, with the worst having lost 1.1 inches from its original 3.5 inch diameter.

In 1986, wastage was found at Arkansas Nuclear One Unit 1, affecting carbon steel components on a high pressure injection nozzle and reactor coolant system piping. The coolant leak that led to the corrosion occurred at a high pressure injection (HPI) manual isolation valve, and apparently continued for about six months. The maximum depth of attack on the bottom exterior surface of the HPI nozzle was 0.5 inch, while the adjacent RCS piping experienced 0.25 inch of wastage. The stainless steel cladding at the affected location was confirmed to be intact.

At the North Anna plant, wastage was found in carbon steel reactor coolant charging pump casings, associated with cracks in stainless steel cladding. Damage was found over the period 1989-1994.

The maximum depth of attack reported was about 0.5 inch.

Reactor pressure vessel head wastage was observed in the US at Turkey Point Unit 4 in 1987. A conoseal joint in an instrumentation port column assembly experienced leakage that resulted in borated coolant dripping onto the vessel head. Although the leakage was detected at an outage in August, 1986, a determination was made that the risk of significant corrosion wastage was low. At an outage in March, 1987, however, a large quantity (about 500 pounds) of boric acid crystals was found to be covering part of the vessel head. The leakage location was near the edge of the vessel head, and the most affected components were the head flange and several flange studs and nuts. A curved

("boomerang-shaped") depression was found in the vessel head, with dimensions of 8.5 inches long x 1.25 inches wide x 0.25 inch deep. Six sets of head flange studs, nuts and washers were replaced.

At the Salem Unit 2 reactor, boric acid crystals were found on a seam in the ventilation cowling surrounding the reactor head area in August, 1987. A seal weld at a thermocouple threaded connection (conoseal) had allowed reactor coolant leakage. Corrosion damage to the head consisted of nine corrosion pits, 1 to 3 inches in diameter and 0.36 to 0.40 inch deep.

In 1994, Calvert Cliffs Unit 1 produced three nuts on an incore instrumentation flange that were corroded by boric acid leaking past a flange gasket, and at Three Mile Island, maintenance personnel found four corroded studs (out of eight total) holding the pressurizer spray valve bonnet gasket in place.

Some overseas plants have also experienced inconsequential boric acid wastage of reactor vessel heads. In 1970, the Swiss reactor Beznau Unit 1 experienced a canopy seal leak at a weld defect. A large quantity of boric acid was observed on the upper head. After cleaning, a crescent-shaped area of attack was noted adjacent to a CRDM penetration. The corroded area was 50 mm wide (radially from the penetration) and 40 mm deep (in the vessel head thickness direction). After cleaning, inspection and stress analysis, the head was returned to service without repair.

In 1996, the French reactor Bugey 3 experienced a leak from an improperly assembled bolted flange at an air vent line. Another French unit, Tricastin 4, also experienced a leaking canopy seal, in 1998.

In both of these cases, although significant amounts of boric acid were removed from the vessel heads, the maximum depth of corrosive attack was only a few millimeters, and required no repairs.

Page 4 of 19

Boric Acid Corrosion of Vessel Pressure Boundary 6.0

SUMMARY

OF NRC AND INDUSTRY ACTIONS CONCERNING BORIC ACID CORROSION The NRC issued two Bulletins in 1982 related to boric acid corrosion of carbon and low-alloy steels.

The first, chronologically, was Notice 82-06 to describe the corrosion of the steam generator manway closure studs at Maine Yankee. The second chronologically was Bulletin 82-02, which notified licensees about corrosion of coolant pump closure studs at Ft. Calhoun, reiterated the information about SG manway closure studs at Maine Yankee, and requested specific actions of the licensee regarding the inspection and maintenance of closure studs that might be affected by boric acid leakage. As a response to the corrosion incident at ANO-1, the NRC issued Information Notice 86-108, that was updated with Supplement No. 1 in 1987, describing the Turkey Point Unit 4 observations, Supplement No. 2 later in 1987, describing the issue at Salem 2, and Supplement No.

3 in 1995, describing the incidents at Calvert Cliffs - 1 and Arkansas Nuclear - 1. In 1988, the NRC issued Generic Letter 88-05 describing the concern over the corrosion of carbon steel pressure boundary components by boric acid. The Letter required all licensees to develop a program of leak identification, inspection, engineering evaluations and corrective actions "to address the corrosive effects of reactor coolant system leakage at less than technical specification limits".

As a result of this concern, the PWR Owners' Groups embarked individually on experimental and analytical programs to understand better the corrosion processes, rates and safety issues resulting from such corrosion. The more applicable of these results are described in the next section. As information from these individual efforts came to light, coupled with the knowledge that a CRDM crack had developed at Bugey-2, the owners groups coordinated their efforts, with the assistance of NUMARC (now NEI - The Nuclear Energy Institute). As the studies were concluded, each of the Owners' Groups issued a report describing the safety significance (or generally, the lack of safety significance as it was perceived at the time) of boric acid corrosion as a degradation threat for pressure boundary components. Generally, the position taken was that leaks would be small and that the annulus might well block projection of the boric acid to the exterior of the vessel. The owners' groups included a calculational exercise showing that even if boric acid did produce wastage to the head, the ASME code margins would not be threatened by the ensuing reduction in thickness, or development of a cavity near a CRDM.

7.0

SUMMARY

OF LABORATORY EXPERIMENTS TO EVALUATE BORIC ACID CORROSION This section contains information taken from the EPRI Boric Acid Corrosion Guidebook that appears to have the most similitude to the early and late phases of the Davis-Besse cavity development. A tabulation of these experiments is given in Table 1 that is appended to this report.

The best experimental representation available of the early phases of the Davis-Besse head cavity formation would be the high temperature CRDM mockup tests labeled EPRI-6 in the Boric Acid Corrosion Guidebook, Rev 1 (BACG). These experiments exhibit a corrosion rate of up to 1 to 3 in./yr, for the various test variables chosen. However, the test duration (50 days) and range of flow rates may not be extensive enough to properly apply these results to the Davis-Besse cavity formation. Test L describes the effects of borated steam impingement on carbon/low alloy steel material. Maximum corrosion rates exceed 11 in./yr for a flow rate of 0.1 gpm., with the corrosion rate dropping as the target is moved farther from the steam source. Critical in this consideration is that the annulus be open enough to admit oxygen, and that the escaping steam be at a low enough velocity to allow that to happen. All experiments with deaerated solutions produced very low wastage rates.

It is hypothesized in EPRI's PWR RPV Upper Head Inspection Plan (MRP-75) that once the RCS coolant leak rate exceeded 0.1 g.p.m., the rate of leakage onto the head exceeded the rate of evaporation of the leaking coolant at atmospheric pressure (Ref. MRP-75, App. C). Physically, the superheated water squirts out of the nozzle leak at a temperature of about 600 0F. A large fraction of Page 5 of 19

Boric Acid Corrosion of Vessel Pressure Boundary that water is converted quickly to steam, but the consequent removal of the heat of vaporization from the escaping flow of water cools the small fraction left to a sub-boiling temperature. This process would allow a concentrated boric acid solution at approximately 200-212 F (concentrated further through the evaporation and possibly by rewetting of dry boric acid left on the head) to pool and flow down the head away from the leaking nozzle, while corroding away the head in a top-down fashion until exposure of the stainless steel cladding. Experiments H, G, F and A in the BACG show that carbon/low alloy steels exposed to aerated, concentrated, boric acid solutions exhibit corrosion rates in the short term of up to -7 in/yr, depending on the level of concentration of the boric acid solution.

More dilute solutions produced wastage rates of 0.64 to 1.32 inches per year. As these experiments ran longer, the corrosion rates decreased by factors of 2 to 4. That decrease was presumed to be due to depletion of the boric acid species. That would not be applicable of the Davis-Besse cavity contents, which were continuously refreshed from the leaking nozzle. There is no explanation offered for the data that shows the temperature dependence of the wastage inverts for a concentration of 2500 ppm boron, for which corrosion rates increased with the temperature.

Several other experimental results listed in the BACG may relate well to the Davis-Besse head degradation phenomenon. Tests G, , and J (Table c) describe corrosion rates associated with leaking boric acid dripping onto high temperature carbon/low alloy steel material. The rates range from near 0.0 at elevated temperature, approaching 5 in/yr. as the temperature dropped to 200°F -

300°F, or the concentration increased to >20,000 ppm. This experiment corresponded geometrically to leakage dripping from CRDM flanges.

Concentrated solutions of boric acid can also attack stainless steels, although at a much slower rate than carbon steels. The following table is transcribed from information found in Reference 14 Table 2. Corrosion rates of 304 stainless steel in concentrated boric acid solutions.

Solution Concentration () Temperature (F) Corrosion Rate (inches per year) 30% 302 0.025 50% 302 0.045 70% 302 0.35 and 0.245 2.5% 175 <1 x 1 0 4 in/year Although two variables are changing (concentration and temperature), it appears likely that for stainless steels, increasing temperature leads to increasing corrosion rates.

In the case of the Davis-Besse cavity development, it does not appear that the stainless cladding was corroded by its exposure to the boric acid solution. The average thickness of the exposed clad is 0.256 in., with a maximum measurement of 0.314 in. Measurements taken from a micrograph of cladding intact on another portion of the vessel head near the cavity show an average thickness of 0.267 in. with a maximum of 0.300 in. This lack of corrosion of the stainless cladding suggests that the temperature of the corroding liquid was low, and is commensurate with the thinking that the solution temperature in the Davis-Besse cavity may have been about 200°F.

Combining these observations, and using the conclusion from MRP-75 that the Davis-Besse cavity contents existed at a temperature of about 200°F (i.e., the leak rate was > 0.1 gpm), leads to the serious possibility that corrosion rates could have been quite high (5 to 7 inches per year) at all the phases of the cavity development, once the annular gap had opened sufficiently to admit oxygen.

Page 6 of 19

Boric Acid Corrosion of Vessel Pressure Boundary 8.0 DEFICIENCIES IN THE EXISTING BODY OF RESEARCH RESULTS Certainly the research that has been published was conceived and conducted with a specific objective in each case. However, the Davis-Besse issue revealed that no previous experiment had captured all of the important aspects of the process. In particular, there appears to be no information on the electrochemical properties of boric acid solutions in contact with the alloys involved, nor on the galvanic interactions of those alloys. Close observations of the Davis-Besse cavity show undercutting of the low-alloy steel right at the clad interface. This geometry could have'been set up by galvanic coupling between the clad and the low-alloy steel. There is no information on the effects of stirring or rapid flow on corrosion rates; it is likely that a leak rate of 0.1 gpm would provide considerable turbulence to a cavity full of solution. Most of the experiments were conducted for a specific concentration at a specific temperature, and there is little useful information about the effect of temperature on corrosion rates or electrochemical potentials.

9.0

SUMMARY

OF RESEARCH BEING INITIATED BY NRC/RES In August, 2002, RES/DET/MEB funded a program with Argonne National Laboratory titled

'Degradation of Reactor Pressure Vessel Boundary Components in Concentrated Boric Acid Solutions". This program consists of several tasks, summarized below:

Task 1: Crack Initiation and Growth Rates of Alloys 600 and 182 Removed from Davis-Besse Nozzles and J-weld The objective of this task is to conduct stress-corrosion crack initiation and crack growth rate tests in simulated PWR coolant of CRDM and J-weld alloys removed from the head of the Davis-Besse plant.

The initiation tests will show whether cracks could have formed with little or no incubation period, and give some information about the shape of the cracks, given the particular material characteristics of superficial work hardening, residual stress and grain size. The crack growth rate tests will demonstrate how the materials used in the construction of the original Davis-Besse head compare with the existing SCC data base.

Task 2: Development of an Integrated Crack Growth Rate and Inspection Frequency Determination Model The objective of this task is to invoke a probabilistic approach to the development of a calculational model leading to prediction of appropriate inspection intervals for vessel penetrations. The probabilistic model will incorporate uncertainties in factors, such as: (a) sizing of cracks as determined through non-destructive inspections; (b) probability of detection; (c) variation in crack growth rates, due to microstructural and environmental conditions; (d) variations in stress intensity factor, due principally to residual stresses; (e) variations in leakage rates; and ( structural integrity evaluations.

Task 3: Corrosion of Reactor Steels in Concentrated Boric Acid Solutions The objective of this task is to measure the wastage rates of A533B pressure vessel steel and 308SS cladding steel in flowing and quiescent boric acid solutions of varying concentrations and at various temperatures. Additionally, corrosion tests will be conducted in molten boric acid and boric oxide mixtures at two temperatures, to be determined, and under pressure and humidity conditions that provide chemical compound stability for the molten species.

Task 4: Measurement of Electrochemical Potential (Corrosion Potential) of Alloy 600, Alloy 182 and A533B in Concentrated Boric Acid Solutions Page 7 of 19

Boric Acid Corrosion of Vessel Pressure Boundary The objective of this task is to measure the electrochemical corrosion potential of the materials found in the head and nozzles of the Davis Besse reactor, under the same set of temperature, solution concentration and flow/no flow conditions as will be explored in Task 2.

Task 5: Technical and Organization Assistance in the Conduct of an International Workshop on Nickel-Base Alloy Corrosion Cracking Issues The objective of this task is to assist NRC staff in securing a venue, appropriate staffing, audiovisual assistance and program development for an international workshop on nickel-base alloy corrosion cracking issues.

Task 6 (Optional): Construction of a "Mock-Up" of the Davis -Besse CRDM Leak and Wastage Process The objective of this task is to develop and test a simulated leak into a crevice and to monitor the degree of plugging of the annulus openings, as well as the progress of the corrosion of the low-alloy steel, and to determine the dependence of the leak path resistance on the plugging and corrosion of the annular cavity. A second phase of the experiment is to simulate a large cavity, such as that found in the Davis-Besse head, to establish temperature and solution concentration and aeration conditions believed to maximize corrosion rates of low-alloy steel, and to determine the linear rates of wastage and nominal growth characteristics of this cavity.

9.0

SUMMARY

The major points of importance in this report may be summarized as follows:

(a) There have been many, many instances of corrosion of low-alloy steels that have occurred on the reactor pressure boundary. The phenomenon has been well-known to the industry and the NRC for over thirty years; (b) The dissemination of information through NRC Bulletins and Information Notices, and through the publication of the EPRI Boric Acid Corrosion Guidebook assured that the requisite information about the process and consequences of boric acid corrosion was readily available; (c) The experimental evidence to support high corrosion rates has been available. However, the most relevant of the data consists of just one or two data points, and there is no information about the dependence of corrosion rates on temperature, concentration, and other helpful variables. The lack of comprehensive information allows the data to be easily overlooked, or discarded as not relevant.

(d) The program recently initiated at Argonne National Laboratory seeks to fill the gaps, and extend the database over a span of applicable variables.

Page 8 of 19

Boric Acid Corrosion of Vessel Pressure Boundary Table a. Corrosion Mechanism #1: Leakagqe into Annulus Experimen BACG Constants Variables Corrosion Applicability to Davis-Besse t Name Test Rate (in/yr)

Referenc e

These tests show what could occur in very early Immersion

• Deaerated stages of leakage, where oxygen cannot penetrate in Low

• 3000 ppm B the depth of the annular crevice and there is little Oxygen B

• 590 F 0.0002 pressure drop across the crack such that the solution solution does not vaporize and concentrate. (Does not account for galvanic corrosion)

Deaerated, concentrated boric acid solution does not exhibit very large corrosion rates, though they are more than two orders of magnitude greater than dilute, deaerated boric acid solutions. This set of Immersion a Deaerated experiments demonstrate would occur in the annular in Low

• 20,000 ppm

• 15 days 0.050 crevice should evaporative concentration occur with Oxygen, EPRI-4 B

• 30 days 0.036 little or no ingress of oxygen due to steam exiting the Concentrat

• 180 F

• 60 days crevice. Decreasing corrosion rates with test time for ed 0.027 these experiments is due to corrosion product Solutions formation on the surface of the samples, which may have occurred during formation of the early stages of the Davis-Besse corrosion cavity assuming the annulus did plug, and depending on the ability for the corrosion products to be swept off by escaping RCS

Boric Acid Corrosion of Vessel Pressure Boundary coolant.

1 - 1* t +

• No insulation Leakage into the annular region surrounding the

• Upwards

• 0.01 gpm penetrations could produce significant corrosion v~ {

orientation

• Insulation rates, assuming oxygen ingress is sufficient to

• 2000 ppm B

• 0.01 gpm support the wastage process. Lower flow rates

• 600°F 1.75 tested in these experiments show larger corrosion Leakage

• Annulus gap:

• Insulation rates, presumably due to the greater ability for into EPRI-6 0.005 in a 0.10gpm oxygen to diffuse against the flow of steam/corrosion Annulus 1.23 products exiting the annular region. The added 450 angled presence of insulation also tends to lower the surface (of corrosion rate, and while Davis-Besse's mirror simulated head insulation is not attached to the head, the presence of material) dried boric acid left on the head due to flange leakage could contribute to this effect.

,A L 10

Boric Acid Corrosion of Vessel Pressure Boundary Table lb. Corrosion Mechanism #2: Steam Impingement BACG Experimen Test Corrosion Constants Variables Applicability to Davis-Besse t Name Referenc Rate (in/yr) e

• 0.05 gpm flow

• target 0.5 6.4 - 9.4 These tests provide another example of how in away borated steam can produce high corrosion rates as

• 1000 ppm B

• target 2 in the saturated steam locally cools the target metal Steam aa Impgeame

• 600 F steam away 1.33 - 1.49 surface and concentrates boric acid. Increasing Impingeme L 65 Ftre Lnt onto n 650 F target a 0.10 gpm distance from the steam source decreases the nt onto

• 0.10 gpm Plates plate and observed corrosion rates, which would suggest that Plates ~~~~~~~~flow flow capillary

• target 0.5 9.28 - 11.1 cavity growth rates due to steam escaping from the

. ae0 leaking nozzle would decrease as the cavity became In awaylagr a target 2 in larger.

away 1.61 - 3.90

Boric Acid Corrosion of Vessel Pressure Boundary Table c. Corrosion Mechanism #3: Concentrated, Aerated Boric Acid Corrosion Experiment BACG Constants Variables Corrosion Applicability to Davis-Besse Name Test Rate (in/yr)

Referenc e

Immersion

• Aerated

• 79,000 ppm Experiment H characterizes the corrosion of carbon in Aerated

• 220 F B steel subject to a pool of concentrated boric acid 7.25 BA solution at 220°F - simulating the situation

• Solution not

• 4 hr test Solution hypothesized in MRP-75. Although Exp. H was short replenished in duration, the authors suggest that rates would

• 26,200 ppm

• Vapor B 1.32 remain high as long as the concentration of the condensed at solution continued to be replenished.

H

• 6 hr test 0.30 mouth of test flasks to

• 27 hr test prevent loss Concentrated BA solutions expected to be present of solution

• 22,000 ppm due to evaporation/concentration of RCS leakage in B contact with carbon/low alloy steels show potentially 0.64 high corrosion rates. The decrease in corrosion rates

• 24 hr test 0.28 observed in experiment H is potentially due to the

• 96 hr test solution not being replenished and the oxidizing agents being used up, a situation not expected to be

• Aerated replicated by continuously leaking nozzles.

G

• 200 F

• 43,750 ppm B 4.8 Aerated solutions with lower boron concentrations G u 2000 F~~~~~~~~~~~~~~~~~~~~~~

than those expected to be produced from solution

• Aerated

• 4000 ppm B evaporation/concentration produce lower corrosion

• 212°F 0.11 - 0.12 rates, as do increasing and decreasing temperatures F from the BA solution boiling point (200-212 F). In

• 352 °F 0.04 - 0.05 Test Reference F, addition of LiOH to the solution did

• 600°F 0.02 - 0.03

Boric Acid Corrosion of Vessel Pressure Boundar

• 4000 ppm + not affect the corrosion rates. I LiOH to pH 3

• 212°F 0.11 - 0.13

• 352°F 0.046 -

0.054 I +

• Aerated

• 500 F 0.24

• 2500 ppm B

• 140 F 0.015 A

• 100 F 0.007

• 70 F 0.002 13

Boric Acid Corrosion of Vessel Pressure Boundary Table c. Corrosion Mechanism #3: Concentrated, Aerated Boric Acid Corrosion (continued)

Experiment BACG Constants Variables Corrosion Applicability to Davis-Besse Name Test Rate (in/yr)

Referenc e

V 2000 ppm B

• horizontal For dripping onto hot carbon/low alloy steel surfaces solution target from CRDM flanges or cracked nozzles, corrosion

• 180 F

• 0.01 gpm rates approach 5 in/yr. Tilting the target surface does solution temp 1.7 not significantly decrease the corrosion rate, as

• 0.10gpm shown in experiment EPRI-5. For the EPRI-5 tests,

• solution 1.9 minimal corrosion occurred at the drip point.

EPRI-5 dripped on Maximum corrosion rates occurred at narrow troughs 600 °F target

• target with coincident with boric acid flow lines, with trough plates 450 tilt depths decreasing as distance from drip point

• 0.01 gpm increases, which is consistent with results observed

• 0.10 gpm 1.5 in test 1. The two flow rates tested in EPRI-5 here Dripping showed little difference in corrosion rates, though it onto 1.2 4 1~~~~ I1 would be expected that at very small drip rates the Heated solution evaporation rate would be fast enough to Surfaces produce minimal corrosion rates.

Test reference G shows that increasing the boron

• 26,250 ppm B 4.2 - 4.8 concentration in solution increases the corrosion rate G 210 F when the solution is dripped onto a heated surface.

Tests I and J describe tests of solution dripping onto pipe surfaces of various temperatures. Again the corrosion rates approach 4-5 in/yr depending on the variables selected during the experiment.

14

Boric Acid Corrosion of Vessel Pressure Boundary Table c. Corrosion Mechanism #3: Concentrated, Aerated Boric Acid Corrosion (continued)

• 30° Angled 36 in. trough target

• 200 F solution Dripping

• 300°F

• top of trough 4.9 Tests I and J describe tests of solution dripping onto onto trough pipe surfaces of various temperatures. Again the I a mid length 4.4 corrosion rates approach 4-5 in/yr depending on the Heated

• 2600 ppm B

• bottom 2.8 variables selected during the experiment.

Surfaces

. 1.7 ppm Li

• 0.03 gpm flow dripped at top of trough 15

Boric Acid Corrosion of Vessel Pressure Boundary

• 30 Angle

• 2400 ppm B pipe + 1.6 ppm Li

• 200°F

• 325 °F solution

• 300 °F 4.0

• 0.03 gpm

• 280°F 3.5 flow 1.9

• 2280 ppm B

• 3 drip points + 1.5 ppm Li at 3 different temps on the

• 300 °F pipe

• 500 °F J 4.6

• 530 F 1.3

• 2420 ppm B

+ 1.3 ppm Li 0.3

• 330-430 OF

• 556 °F 1.9

• 554-593 OF 1.5 0.0 16

Boric Acid Corrosion of Vessel Pressure Boundary LIST OF REFERENCES FOR BORIC ACID/OXIDE

1. W.J. Shack, "Technical Evaluation Report on the Environment in the Crevice Between a Leaking CRDM Nozzle and the Reactor Vessel Head", DRAFT Report, 2002.

2. "Crack Growth of Alloy 182 Weld Metal in PWR Environments", PWRMRP-21, Electric Power Research Institute, Palo Alto, CA, 2000.

3. "PWR Materials Reliability Program Response to NRC Bulletin 2001-01", MRP-48, Electric Power Research Institute, Palo Alto, CA, 2001.

4. F.A Cotton, G. Wilkinson, Advanced Inorganic Chemistry, 4 th Edition, John Wiley & Sons, New York, 1980.

5. N.T. Moisidis, M.D. Ratiu, "Pump and Valve Fastener Serviceability in PWR Nuclear Facilities", Journal of Pressure Vessel Technology, Vol. 118, Transactions of the ASME, February, 1996, pp. 38-41.

6. Boric Acid Corrosion Guidebook, Revision 1: Managinq Boric Acid Corrosion Issues at PWR Power Stations, Electric Power Research Institute, Palo Alto, CA, 2001.

7. CRC Handbook of Chemistry and Physics, 82nd Edition, ed. David R. Lide, CRC Press, New York, 2001, pp. 4 4-47.

8. N. Moisidis, M. Popescu, M. Ratiu, "Protect Nuclear Plant Fasteners from Boric Acid Corrosion," Power Engineering, March, 1992, pp. 38-41.

9. Davis-Besse Nuclear Power Station NRC Augmented Inspection Team: "Degradation of the Reactor Pressure Vessel Head", Report No. 50-346/02-03(DRS), USNRC, Washington, DC, 2002.

10. Davis-Besse Nuclear Power Station Root Cause Analysis Report for CR2002-0891:

"Significant Degradation of the Reactor Pressure Vessel Head", First Energy, April 15, 2002.

11. U. Gurbuz, N. Bulutcu, "A New Process to Produce Granular Boric Oxide by High Temperature Dehydration of Boric Acid in a Fluidized Bed," Transactions of the Institute of Chemical Engineers, 74A, 133, 1996.

12. "Processing of Nuclear Power Plant Waste Streams Containing Boric Acid", International Atomic Energy Agency.

13..Various Proceedings of EPRI Workshops on PWSCC of Alloy 600 in PWR's. Available from the EPRI website - Nuclear Power/Materials Reliability Program/Orderable Items.

14. J. P. Polar, A Guide to Corrosion Resistance, published by Climax Molybdenum Company, and distributed by the International Molydenum Association, Pittsburg, PA. fourth printing, 1981.

NUREG's

1. C.J. Czajkowski, "Survey of Boric Acid Corrosion of Carbon Steel Components in Nuclear Plants", NUREG/CR-5576, USNRC, Washington, DC, 1990.

Boric Acid Corrosion of Vessel Pressure Boundary

2. V.N. Shah, A.G. Ware, A.M. Porter, "Assessment of Pressurized Water Reactor Control Rod Drive Mechanism Nozzle Cracking", NUREG/CR-6245, USNRC, Washington, DC, 1994.

Requested Through Interlibrary Loan

1. A.S. Myerson, Handbook of Industrial Crystallization, Butterworth-Heinemann, Boston, 1993.

2. E.L. Muetterties, The Chemistry of Boron and Its Compounds, Wiley, 1967.

3. P. Nelson, G.W. Campbell, Boron, Metallo-boron Compounds, and Boranes, R.M.

Adams,ed., Interscience Publishers, New York, 1964.

EPRI Reports Requested from Library

1. "Degradation and Failure of Bolting in Nuclear Power Plants," EPRI-NP-5769, Vol. 1 and 2, April 1988.

2. "Boric Acid Corrosion of Carbon and Low Alloy Steel Pressure-Boundary Components in PWRs," EPRI-NP-5985.

NRC Generic Communications

1. Information Notice 80-27: "Degradation of Reactor Coolant Pump Studs", USNRC, Office of Nuclear Reactor Regulation, June 11, 1980.

2. Information Notice 82-06: "Failure of Steam Generator Primary Side Manway Closure Studs", USNRC, Office of Nuclear Reactor Regulation, March 12,1982.

3. IE Bulletin 82-02: "Degradation of Threaded Fasteners in the Reactor Coolant Pressure Boundary of PWR Plants", USNRC, Office of Nuclear Reactor Regulation, June 2,1982.

4. Information Notice 86-108: "Degradation of Reactor Coolant System Pressure Boundary Resulting from Boric Acid Corrosion", USNRC, Office of Nuclear Reactor Regulation, December 29, 1986.

Supplement 1, April 20,1987.

Supplement 2, November 19, 1987.

Supplement 3, January 5,1995.

5. Generic Letter 88-05: "Boric Acid Corrosion of Carbon Steel Reactor Pressure Boundary Components in PWR Plants", USNRC, Office of Nuclear Reactor Regulation, March 17,1988.

6. Information Notice 90-10: "Primary Water Stress Corrosion Cracking (PWSCC) of Inconel 600, USNRC", Office of Nuclear Reactor Regulation, February 23,1990.

7. Information Notice 94-63: "Boric Acid Corrosion of Charging Pump Casing Caused By Cladding Cracks", USNRC, Office of Nuclear Reactor Regulation, August 30, 1994.

18

Boric Acid Corrosion of Vessel Pressure Boundary

8. Information Notice 96-1 1: "Ingress of Demineralizer Resins Increases Potential for Stress Corrosion Cracking of Control Rod Drive Mechanism Penetrations", USNRC, Office of Nuclear Reactor Regulation, February 14, 1996.

9. Generic Letter 97-01: "Degradation of Control Rod Drive Mechanism Nozzle and Other Vessel Closure Head Penetrations", USNRC, Office of Nuclear Reactor Regulation, April 1, 1997.

10. Information Notice 01-05: "Through-wall Circumferential Cracking of Reactor Pressure Vessel Head Control Rod Drive Mechanism Penetration Nozzles at Oconee Nuclear Station",

Unit 3, USNRC, Office of Nuclear Reactor Regulation, April 30, 2001.

11. Bulletin 01-01: "Circumferential Cracking of Reactor Pressure Vessel Head Penetration Nozzles", USNRC, Office of Nuclear Reactor Regulation, August 3, 2001.

12. Information Notice 02-11: "Recent Experience with Degradation of Reactor Pressure Vessel Head", USNRC, Office of Nuclear Reactor Regulation, March 12, 2002.

13. Bulletin 02-01: "Reactor Pressure Vessel Head Degradation and Reactor Coolant Pressure Boundary Integrity", USNRC, Office of Nuclear Reactor Regulation, March 18, 2002.

14. Information Notice 02-13: "Possible Indicators of Ongoing Reactor Pressure Vessel Head Degradation", USNRC, Office of Nuclear Reactor Regulation, April 4, 2002.

15. Bulletin 02-02: "Reactor Pressure Vessel Head and Vessel Head Penetration Nozzle Inspection Programs", USNRC, Office of Nuclear Reactor Regulation, August 9, 2002.

19