United States Nuclear Regulatory Commission - Protecting People and the Environment

Overview of Reactor Pressure Boundary Integrity Issues

On this page:

Generic Activities on Alloy-600 Cracking

Alloy 600 is used to fabricate various parts in nuclear power plants, including reactor vessel top head penetrations for control rod drive mechanism (CRDMs), control drive element mechanisms (CEDMs), in-core instruments (ICIs) and thermocouples, reactor vessel bottom head bottom mounted instruments (BMIs), pressurizer heater sleeves, and various other instrumentation ports. Related weld materials Alloy 82 and Alloy 182 are used to join these Alloy 600 parts to the ferritic steel components and also as a bi-metallic weld joining ferritic base materials to austenitic stainless steel base materials. Alloy 600 and its associated weld filler metals were originally used because of expectations of resistance to service-induced cracking. However, parts fabricated from these materials have demonstrated a susceptibility to primary water stress corrosion cracking (PWSCC), also referred to as low potential stress corrosion cracking (LPSCC).

In the United States, PWSCC of Alloy 600 became an issue following a leakage event of a pressurizer heater sleeve nozzle at Calvert Cliffs Unit 2 in 1989. Other instances of leakage in pressurizer instrument nozzles were identified in both domestic and foreign PWRs, as described in Information Notice 90-10.

The first indication of cracking in upper head Alloy 600 penetrations was identified in France at Bugey Unit 3 in 1991 during the ten-year primary system hydrostatic test. The leakage was from an axial flaw that had initiated on the nozzle inside surface near the elevation of the J-groove weld. Several other partial depth axial cracks were identified at a similar elevation in this nozzle. Failure analysis confirmed that the cracking was due to PWSCC.

In the United States, the NRC and the industry initiated activities to assess the safety significance of VHP nozzle cracking. An action plan was implemented by the NRC staff in 1991 to address PWSCC of Alloy 600 VHPs at all U.S. PWRs. This action plan included a review of safety assessments submitted by the PWR Owners Groups, the development of VHP mock-ups by the Electric Power Research Institute (EPRI), the qualification of inspectors on the VHP mock-ups by EPRI, the review of proposed generic acceptance criteria from the Nuclear Utility Management and Resource Council (NUMARC) [now the Nuclear Energy Institute (NEI)], and VHP inspections. As part of this action plan, the NRC staff met with each of the owners groups separately and with the entire industry throgh NUMARC/NEI. After reviewing the industry's safety assessments and examining the overseas inspection findings, the NRC staff concluded in a safety evaluation dated November 19, 1993, that VHP nozzle cracking was not an immediate safety concern. The bases for this conclusion were that if PWSCC occurred at VHP nozzles (1) the cracks would be predominately axial in orientation, (2) the cracks would result in detectable leakage before catastrophic failure, and (3) the leakage would be detected during visual examinations performed as part of surveillance walkdown inspections before significant damage to the reactor vessel closure head would occur.

The first U.S. inspection of VHPs took place in the spring of 1994 at the Point Beach Nuclear Generating Station, and no indications were detected in any of its 49 CRDM penetrations. The eddy current inspection at the Oconee Nuclear Station (ONS) in the fall of 1994 revealed 20 indications in one penetration. Ultrasonic testing (UT) did not reveal the depth of these indications because they were shallow. UT cannot accurately size defects that are less than one mil deep (0.03 mm). These indications may be associated with the original fabrication and may not grow; however, they will be reexamined during the next refueling outage. A limited examination of eight in-core instrumentation penetrations conducted at the Palisades plant found no cracking. An examination of the CRDM penetrations at the D. C. Cook plant in the fall of 1994 revealed three clustered indications in one penetration. The indications were 46 mm (1.81 in.), 16 mm (0.63 in.), and 6 to 8 mm (0.24 to 0.31 in.) in length, and the deepest flaw was 6.8 mm (0.27 in.) deep. The tip of the 46-mm (1.81 in.) flaw was just below the J-groove weld. Virginia Electric and Power Company inspected North Anna Unit 1 during its spring 1996 refueling outage. Some high-stress areas (e.g., upper and lower hillsides) were examined on each outer ring CRDM penetrations and no indications were observed using eddy current testing. During this time, each of the vendors was developing a susceptibility model for VHP nozzles based on a number of factors, including operating temperature, years of power operation, method of fabrication of the VHP, microstructure of the VHP, and the location of the VHP on the head. Each time a plant's VHPs are inspected, the inspection results are incorporated into the model.

To top of page

Generic Letter 97-01

On April 1, 1997, the NRC issued Generic Letter 97-01 to request PWR licensees submit descriptions of their programs for inspecting CRDM and other VHP nozzle penetrations. The industry used a histogram grouping of plants, in combination with completed inspections and planned inspections as its approach for managing this issue. The plant grouping used probabilistic crack initiation and growth models to estimate the amount of time remaining (in effective full power years, EFPYs) until the plant reached a limiting condition for a reference plant. This limiting condition was the time for the plant to reach the same probability of having a crack 75% through-wall as D.C. Cook Unit 2 had at the time a 6.5 mm deep crack was identified in 1994. These models included differences in operating time and temperature, water chemistry environment, surface stress, component geometry, material yield strength and microstructure, and fabrication practices (amount of cold work during machining) between the subject plant and the reference plant in determining a plant's susceptibility.

Inspections continued into the fall of 2000, with no significant adverse results. The most significant crack identified in these inspections was a 6.8 mm (0.27 in.) deep crack found at D.C. Cook Unit 2. This flaw was repaired by a process that involved partial removal (by grinding) and overlay weld to isolate the remnant of the original flaw from the environment. Three plants identified small "craze cracks," generally found as cluster of shallow, less than 0.2 mm deep (0.008 in.) and axially oriented. At Millstone Unit 2, one nozzle with seven such indications in a single nozzle were removed by flapper wheel grinding to a depth of 0.8 mm (0.032 in.).

Worldwide, inspection activities were finding PWSCC in VHP nozzles, and in some cases RPV heads were being replaced. Common characteristics of these findings were the flaws originating in the nozzle base material and located on the inside surface of the nozzles.

To top of page

Hot Leg Cracking at the V.C. Summer Nuclear Station

On October 7, 2000, during a normally scheduled containment inspection after entering a refueling outage, the licensee for V.C. Summer Nuclear Power Station identified a circumferential indication of PWSCC in the first weld between the reactor vessel nozzle and the "A" loop RCS hot leg piping. Additionally, the licensee found more than 200 pounds of boric acid crystals on the containment floor. In response to this discovery, the NRC formed a Special Inspection Team, which identified potentially generic issues involving limitations of required nondestructive examinations. See Virgil C. Summer Nuclear Power Station for additional detail.

To top of page

Impetus for Bulletin 2001-01

In the fall of 2000, the inspection findings in RPV head penetrations became more significant. At ONS-1 that fall, boron deposits were identified on the RPV head at one CRDM nozzle and at five (of the eight) thermocouple nozzles (one of only two plants with small diameter thermocouple nozzles). Contrary to expectations, the boron deposits were very small (less than 1 in. total volume). Analysis of the CRDM nozzle identified an axial-radial PWSCC crack that initiated in the J-groove weld and propagated part way into the outer diameter surface of the nozzle. The crack in the J-groove weld was arrested when it encountered the RPV head base material, consistent with expectations.

In February 2001, ONS-3 identified nine nozzles with leaks (again small deposits). Additional inspections, including ultrasonic, eddy current and liquid penetrant examinations, identified numerous part- and through-wall axial cracks, generally initiated on the outer diameter surface of the nozzles below the J-groove weld. During the repair of these nozzles, two of the nozzles were found to have through-wall circumferential crack extending 165 around the nozzle, although the cracks were not through-wall for their entire circumferential extent. These cracks were identified as having initiated on the nozzle outer diameter surface. The findings at ONS-3 were the subject of NRC Information Notice 2001-05, issued on April 30, 2001.

In March 2001, Arkansas Nuclear One Unit 1 (ANO-1) identified boron deposits on a single CRDM nozzle. Examination of this nozzle identified an axial part-thoguh wall crack that initiated on the nozzle outer diameter surface below the J-groove weld and propagated to a distance 33 mm (1.3 in.) above the J-groove weld.

In April 2001, ONS-2 identified boron deposits on four CRDM nozzles. Eddy current examinations of these nozzles identified cluster of shallow axial indications on the nozzle inside surfaces, ranging in depth from 0.35 to 0.8 mm (0.014 to 0.032 in.) and in length from 23 to 79 mm (0.9 to 3.1 in.). Ultrasonic examination of these nozzles identified numerous axial flaws on the nozzle outer diameter surfaces, including one circumferential crack above the J-groove weld. The latter was reported as 32 mm (1.25 in.) long and 1.8 mm (0.07 in.) deep. Leakage from these nozzles was identified as originating from the outer diameter surface cracks that propagated along the weld to nozzle interface from below the J-groove weld to above the weld.

To top of page

Bulletin 2001-01

The Nuclear Regulatory Commission issued Bulletin 2001-01, "Circumferential Cracking of Reactor Pressure Vessel Head Penetration Nozzles," dated August 3, 2001, seeking information from all PWR nuclear power licensees regarding the structural integrity of reactor pressure vessel head penetrations.

The focus of this Bulletin was the safety issue of circumferential cracking in VHP nozzles, with a goal of providing assurance that no such issues existed in plants. Because of the time frame involved in the development of a circumferential crack that could be subject to nozzle ejection, visual inspections of the RPV head outer surface, where the nozzle intersected the RPV head, were considered at that point to be an adequate inspection.

To top of page

Bulletin 2002-01

In response to the findings at Davis-Besse, the NRC issued Bulletin 2002-01 on March 18, 2002. The focus of this Bulletin was to assess licensee inspections and other information that could provide a basis for conclusions on the condition of the RPV head. The Bulletin also addressed boric acid corrosion of other parts of the reactor coolant system.

Following issuance of this Bulletin, the spring 2002 inspection findings were relatively quiet. The exception was an inspection at Millstone Unit 2, which identified three nozzles (no leaks) requiring repair, with axial outer diameter surface cracks that extended from below the J-groove into the weld zone. This finding is significant because this plant had the lowest susceptibility of any plant that had identified cracking.

To top of page

Bulletin 2002-02

Since the initial findings of circumferential cracking at ONS-3, the nuclear industry was working to develop inspection recommendations (and justification for the recommendations) that would provide effective management of the issue. This effort was continually challenged by new findings, e.g., Davis-Besse upper head wastage, and the industry did not have a proposal available for consideration by the summer of 2002. To address cracking and wastage on the upper RPV head, the NRC issued Bulletin 2002-02 in August 2002. This Bulletin provided a description of a comprehensive inspection program that addressed a combination of visual and non-visual examinations on a graded approach consistent with a variety of plant susceptibilities to PWSCC. This Bulletin used a parameter referred to as effective degradation years (EDY) to characterize plant susceptibility to PWSCC. Calculation of this parameter requires information on the RPV head operating temperature(s) and the operating time (i.e., effective full power years, EFPY) at each operating temperature. These data are used to integrate the effects of operating temperature, normalized to 316 C (600 F).

Notable inspection findings were prevalent during the fall 2002 outages. North Anna Unit 2 identified two leaking nozzles. One of these leaks was from a nozzle that had received a weld over-lay repair at the previous outage. Failure of the of the repair was attributed to the weld over-lay repair not completely covering the original Alloy 182 weld butter, with cracking then occurring in the original weld at the periphery of the repair weld. Surface examinations of the J-groove welds identified more than half of the welds with cracks. Ultrasonic testing of the nozzle base material identified twenty nozzles with axial indications. Several nozzles were identified with circumferential cracks on the nozzle outer diameter surface within the zone of the J-groove weld, just below the root of the weld. With the myriad of repairs necessary due to these findings, this plant became the first U.S. plant to install a new RPV head using Alloy 690 nozzle base material and Alloy 52 and 152 welds.

At ANO-1, a leak was identified from the nozzle that had been repaired in the spring of 2001. The failure of the repair was attributed to the weld over-lay repair not completely covering the original J-groove weld, similar to the North Anna Unit 2 finding.

At Sequoyah Unit 2, minor head corrosion was identified from a boron leak located above the RPV head. In particular, the licensee identified a leak from a valve in the reactor vessel level instrument system (RVLIS). Leaking coolant impacted the RPV head insulation below the valve, fell through a seam in the insulation and onto the RPV head. After the RPV head was cleaned up, a corrosion area was identified with dimensions 127 mm (5 in.) long and 8 mm (5/16 in.) wide, with a maximum depth of 3 mm (1/8 in.).

Addition inspections since then have provided further instances of axial and circumferential indications being found, attributed to PWSCC of Alloy 82/182 welds. The most significant indications were discovered in 2012 at North Anna in 2012. During a planned pre-emptive full structural weld overlay of a reactor vessel hot leg nozzle to safe end weld, two leaks were discovered. Further examinations revealed five flaws with the deepest estimated at 80% through-wall.

To top of page


The Davis-Besse refueling outage began on February 16, 2002. The licensee planned to perform a visual inspection of the outer surface RPV head looking for signs of boron deposits, and ultrasonic inspection of all CRDM nozzles. The inspection identified five nozzles with indications, including three with through-wall cracks, and the licensee decided to repair all five nozzles. During machining to facilitate repair of nozzle #3, the equipment rotated and was removed from the head. Upon removal, the licensee found that the nozzle had tipped, with the CRDM flange (located above the head) contacting the flange of an adjacent CRDM. The licensee cleaned the surface of the RPV head and found a large cavity adjacent to nozzle #3, where the RPV head base material had been corroded down to the stainless steel cladding. Subsequent investigation revealed an additional much smaller degraded area near nozzle #2, located within the wall thickness (no cladding was exposed).

After the initial finding of the cavity at Davis-Besse, the NRC issued Information Notice 2002-11, "Recent Experience with Degradation of Reactor Pressure Vessel Head," on March 12, 2002. After some of the evidence began to be accumulated regarding secondary indications of a serious ferritic corrosion event, the NRC issued Information Notice 2002-13, "Possible Indicators of Ongoing Reactor Pressure Vessel Head Degradation," on April 4, 2002.

In 2004 an unused RPV head from the cancelled Midland Unit 2 with 82/182 welds was placed in service at Davis-Besse, while a new RPV with Alloy 52/152 welds was fabricated, and scheduled to replace the Midland RPV head. Bare metal visual examinations performed in 2005, 2006, and again in 2008 with no indications of leakage being identified. On March 12, 2010, the licensee identified indications during Volumetric examinations of multiple CRDM nozzles. Additionally, bare metal visual examinations on the RPV head revealed boric acid deposits indicative of leakage. Further examinations using ultrasonic, liquid dye penetrant, visual, and eddy current testing revealed that a total of 24 CRDM nozzles needed repair of either nozzle or j-groove weld, two of those nozzles were confirmed to have pressure boundary leakage. It should be noted that unlike the previous event at Davis-Besse, this time there was no evidence of excessive wastage of the RPV head as a result of the leakage.

NRC Confirmatory Action Letter, dated June 23, 2010, allowed the licensee to operate with the repaired replacement RPV head until October 1, 2011. During the outage in fall of 2011 a new RPV head with Alloy 52/152 welds was installed at Davis-Besse.

To top of page

Order EA-03-009

The NRC issued Order EA-03-009 to all PWR licensees on February 11, 2003. This Order provided specific inspection requirements for all PWR plants. The Order required that plants evaluate their susceptibility to PWSCC using a formula for effective degradation years, EDY. The Order then provided specific inspection requirements based upon the EDY level of the plant. The Order provided requirements for plants with EDY greater than 12 or have experienced PWSCC. These plants were required to perform a bare metal visual examination and a non-visual examination every refueling outage. Moderate susceptibility plants (those with EDY from 8 to 12) were required to perform either bare metal visual or non-visual examination every outage, alternating the two methods each RFO. Low susceptibility plants (with EDY less than 8) were required to perform a bare metal visual examination by their second refueling outage after issuance of the Order and every third refueling outage or five years thereafter. In addition, low susceptibility plants were required to perform non-visual examination by February 11, 2008, and then repeat every fourth refueling outage or seven years thereafter.

The non-visual examinations described in the Order were ultrasonic examination or surface examination. The ultrasonic examination covered from the bottom of the nozzle to 2 inches above the J-groove weld, and included an assessment to determine if leakage has occurred in the interference fit zone of the nozzles. The scope of the surface examination included the surface of the J-groove weld, the outer diameter surface of the VHP nozzle base metal, and the inside surface of the VHP nozzle to a point 2 inches above the J-groove weld.

The Order provided explicit inspection requirements for repaired nozzles and welds, and makes no distinction for heads fabricated from Alloy 600 or Alloy 690.

In addition to the susceptibility based inspections of the RPV head surface and VHP nozzles, the Order required that all licensees perform visual inspections to identify boric acid leaks from components above the RPV head, with follow-up actions including inspection of potentially-affected RPV head areas and VHP nozzles should any leaks be identified.

The Order also provided means for licensees to request relaxation from its requirements upon demonstration of good cause. As of January 2004, twenty-four plants had made specific requests for relaxation. These requests related to limitations in inspection accessablility and technology.

The NRC revised certain inspection aspects of the original NRC Order EA-03-009 with respect to bare metal visual inspections, penetration nozzle inspection coverage, flexibility in combination of non-destructive examination methods, flaw evaluation, and requirements for plants which have replaced their reactor pressure vessel head. The First Revised NRC Order EA-03-009 Rev. 01 was isssued on February 13, 2004.

To top of page

Bottom-Mounted Instrument Cracking

With the focus of attention on PWSCC of Alloy 600 on the upper RPV head and possible boric acid corrosion of ferritic components throughout the reactor coolant system, visual examinations of other applications of Alloy 600 have increased in their thoroughness and effectiveness. One area that was not anticipated to provide short-term PWSCC concerns was the RPV lower head BMIs, due to the cold-leg operating temperature of the RPV lower head. However, in the spring of 2003, the licensee for the South Texas Project Unit 1 (STP-1) identified apparent boron deposits on the lower RPV head near two BMIs. Characterization of all of the BMI nozzles at STP-1 identified PWSCC in these two nozzles, and no PWSCC in any other nozzle. The operating temperature of the STP-1 lower head was ∼ 294 C (561 F), and the calculated EDY was less than three (3). The NRC issued Bulletin 2003-02 to obtain information on licensee inspection activities and inspection plans for the RPV lower head.  Up until October of 2013, no other plants had identified white residue on the lower head due to leakage of a BMI nozzle.  Frequently boron traced to refueling seal leakage or other sources above the RPV lower head, and no other plant has identified PWSCC in the BMIs.  However, on October 6, 2013, during a scheduled visual examination of the BMI nozzles at Palo Verde Nuclear Generating Station Unit 3, the licensee identified white residue around the BMI nozzle. The visual inspections were performed in accordance with the American Society for Mechanical Engineers (ASME) Code Case N-722-1.  Licensee's engineering personnel determined that the white residue was most likely boric acid resulting from pressure boundary leakage.  Non-destructive examinations confirmed that axial cracking at the nozzle weld was responsible for the leakage. Additional information on the event is available on the NRC website under Agencywide Documents Access and Management System Accession No. ML14177A533.

Based on the available operating experience, current requirements in 10 CFR50.55a(g)(6)(ii)(E) "Augmented ISI requirements: Reactor coolant pressure boundary visual inspections," requires that all licensees of pressurized water reactors augment their inservice inspection program with ASME Code Case N-722-1, "Visual Examinations for PWR Pressure Retaining Welds in Class 1 Components Fabricated with Alloy 600/82/182 Materials Section XI, Division 1," subject to the conditions specified in paragraphs (g)(6)(ii)(E)(2) through (4) of [10 CFR 50.55a].  Consequently, licensee with PWR units with Alloy 600/82/182 BMI instrument penetrations are required to perform visual examinations of these components every other refueling outage.

To top of page


Operating experience, both domestic and foreign, has demonstrated that Alloy 82/182/600 materials connected to a PWR's pressurizer may be particularly susceptible to PWSCC. Since the late 1980's, approximately 50 Alloy 600 pressurizer heater sleeves at Combustion Engineering-designed (CE-designed) facilities in the United States have shown evidence of RCPB leakage which has been attributed to PWSCC. The most recent events of this type occurred at Millstone, Unit 2, and Waterford, Unit 3, in October 2003. All available evidence from finite element modeling studies and limited nondestructive evaluation (NDE) has suggested that these leakage events were the result of axially-oriented PWSCC of the pressure boundary portion of these heater sleeves. However, recent NDE results from Palo Verde, Unit 2, on heater sleeves which had not shown evidence of leakage have demonstrated that circumferentially-oriented PWSCC can occur in the non-pressure boundary portion (i.e., above the J-groove attachment weld) of these components.

Degradation attributed to PWSCC has also been observed in the pressurizer heater bundles used in B&W-designed PWRs. The B&W-designed heater bundle employs a diaphragm plate manufactured from Alloy 600 and seal welded with Alloy 82/182, with structural support for the diaphragm plate being provided by a low alloy steel strongback which is bolted to the pressurizer shell. Most recently, in October 2003, pressure boundary leakage through a cracked diaphragm plate was observed at Three Mile Island, Unit 1 (TMI-1). The cracking in the TMI-1 diaphragm plate was attributed to PWSCC in the heat affected zone of the seal weld. Boric acid corrosion of the low alloy steel strongback was also observed to have resulted from the leakage.

Small diameter Alloy 82/182 instrument line penetrations have also shown evidence of PWSCC at many PWR facilities since the 1980's. For example, in October 2003, the Crystal River, Unit 3, licensee reported RCPB leakage from three pressurizer upper level instrument tap nozzles, which are exposed to the steam space in the pressurizer. The leakage was attributed to PWSCC of Alloy 82/182/600 material from which the connections were constructed.

Finally, inspections conducted in September 2003 at Tsuruga, Unit 2, in Japan demonstrated that larger diameter, butt welded lines connected to the steam space of the pressurizer may also be susceptible to PWSCC. Evidence of boron deposits on the surface of a pressurizer relief valve nozzle (inside diameter 130 mm, or approximately 5 inches) lead to the discovery of five axially-oriented flaws in the Inconel alloy weld material used in the fabrication of the nozzle-to-safe end weld. Subsequent NDE performed on a safety valve nozzle of similar diameter resulted in the discovery of two additional flaws in its nozzle-to-safe end weld. Fractographic analysis of the flaw surfaces confirmed PWSCC as the mechanism for flaw initiation and growth.

Extensive operational experience with PWSCC in Alloy 82/182/600 materials used in the fabrication of pressurizer penetrations and steam space piping connections is not surprising. The initiation and growth of PWSCC flaws is known to be strongly dependent on the temperature of the primary system water to which the Alloy 82/182/600 materials are exposed. Given the fact that at the pressurizer the reactor coolant system environment attains a temperature of about 650 F (343 C), PWSCC should be expected to occur in these materials and locations, therefore an effective degradation management program is warranted.

To top of page


Due to the safety significance of RPV Upper Head Issues NRC incorporated inspection requirements into 10 CFR 50.55a to ensure the integrity of the RPV head and VHP nozzles.  This was accomplished with the cooperation of the cognizant American Society for Mechanical Engineers (ASME Code) Section XI committee, through the publication of ASME Code Case N-729, "Alternative Examination Requirements for PWR Reactor Vessel Upper Heads With Nozzles Having Pressure-Retaining Partial Penetration Welds Section XI, Division 1."  Current regulation in 10 CFR 50.55a(g)(6)(ii)(D), "Augmented ISI Requirements:  Reactor vessel head inspections," requires that all licensees of pressurized water reactors augment their inservice inspection program with ASME Code Case N-729-4, subject to the conditions specified in paragraphs (g)(6)(ii)(D)(2) through (4) of [10 CFR 50.55a].

Regarding the future operating experience of Alloy 600 in PWRs, it is reasonable to expect that additional instances of PWSCC will occur for parts, components and joints fabricated from Alloy 600 and weld filler metals Alloys 82 and 182.  Many plants have now replacement RPV heads with Alloy 690 base metals and Alloy 52 and 152 for weld filler metals, which are less susceptible to PWSCC.

In addition, operating experience has shown that the surface stress improvement by peening is a viable mitigation to minimize PWSCC initiation. Several plants have considered implementing peening or surface stress improvement as a method to mitigate the PWSCC of nickel-based dissimilar metal butt or J-groove welds in the reactor coolant pressure boundary. EPRI MRP-335, Revision 3-A "Topical Report for Primary Water Stress Corrosion Cracking Mitigation by Surface Stress Improvement" documents the technical basis for peening application which includes peening processes, performance criteria, analyses, and alternative inspection requirements.

To top of page

Page Last Reviewed/Updated Thursday, May 14, 2020