Faulty Rods

Why are nuclear fuel rods failing? An engineer thinks a faulty manufacturing process he helped design may be to blame. But the Nuclear Regulatory Commission isn’t concerned, and the suspect parts are still being shipped, with potentially dangerous results.

MotherJones Reporter’s Notes

Subjects of investigation:
Teledyne Wah Chang Albany (TWCA); General Electric and other fuel rod suppliers; the Nuclear Regulatory Commission (NRC).
Principal allegation:
Using a process that engineers have testified is flawed, TWCA manufactures casings for nuclear fuel rods. The rods are installed in reactors all over the globe (see map). Nationally, rod failures have been increasing for two years. The NRC knows this but has done little to protect the public.
Principal source:
Chris Hall, a former TWCA engineer who blew the whistle on a questionable process used to manufacture fuel rod casings. Documents obtained from a lawsuit involving Hall, together with other expert testimony, corroborate the main charges.
Public implications:
The rods are a potential safety hazard to millions of people through the release of radioactive material into the environment. Some experts believe that corroded rods could exacerbate other system failures and contribute to a core meltdown, resulting in substantial fatalities and the contamination of food, air, and water.

In the cores of our nation’s 108 licensed nuclear power plants, fuel rods–the first level of protection against the release of deadly radioactive material–are failing in increasing numbers. The Nuclear Regulatory Commission knows this, and has been told one reason why it might be happening, but has done little to prevent potentially flawed fuel rod casings from being shipped to plants here and abroad.

How a situation with such deadly possibilities could continue uncorrected is a tale involving a corporation whose engineers question its manufacturing process, a whistle-blower who has been largely ignored, and a federal agency that critics say has abdicated its regulatory responsibility to the industry it is supposed to oversee.

The story begins with the fuel rods themselves, and what it means when they fail. Fuel rods are zirconium alloy tubes that contain the radioactive uranium in reactor cores. If they fail, radiation is next contained by the coolant system. The third and final level of defense is the actual containment building.

Dr. Michio Kaku, a professor of nuclear physics at City University of New York, likens the danger of a failed fuel rod to that of a car’s fuel line. “If the fuel line of a car ruptures, that in itself won’t cause an accident. But if there is an accident, the ruptured fuel line could cause an explosion and make it fatal pretty easily.”

Consider the 1979 disaster at Three Mile Island. Failed fuel rods didn’t cause the near-meltdown–a loss of cooling water did. But once the cooling water dropped, more than half the fuel rods melted, and radioactive material escaped through a valve in the cooling system. Luckily, the containment building held, and engineers controlled the reaction about one hour before total meltdown. Despite massive releases of radiation, a full-scale disaster was averted.

But we may not always be so lucky, since not all containment vessels are as beefy as Three Mile Island’s. In fact, in the mid-1980s the NRC estimated the nation’s risk of a severe meltdown to be 45 percent over a 20-year period. One reason? Some reactor containment buildings are of a type the NRC described in 1985 as likely to fail “within the first few hours following core melt.”

Given the NRC’s assessment, one would think a notice they sent to nuclear plant operators last fall would have caused more alarm. NRC notice #93-82 warned plant operators about a spate of fuel rod failures and specifically cited seven reactors where such failures had been observed. What the NRC didn’t mention in its quiet bulletin was that two of those reactors were of the type it called susceptible to containment building failure.

Nuclear industry spin doctors are quick to point out that fuel rod failures happened frequently until the mid-1970s, when they peaked, and that they have never caused a meltdown. But they won’t tell you that recent fuel rod failures (which the NRC says have been increasing for about two years) could be more dangerous than those in the past: since today’s manufacturing methods supposedly produce stronger rods, plant operators are running the rods longer and harder than ever before.

They also won’t tell you that even if no meltdown occurs, failing fuel rods pose other potentially lethal hazards. A 1990 study by the Massachusetts Department of Public Health shows that between 1972 and 1979, failed fuel rods at Boston Edison’s Pilgrim plant regularly released radioactivity into the atmosphere. The study noted an associative connection between radioactive releases and the fact that the adult leukemia rate within a 10-mile radius of the plant was four times that of outlying areas.

Despite the danger, the NRC bulletin was largely ignored by the media. And the agency has taken few steps to correct the pattern of fuel rod failures. Larry Phillips, co-author of the NRC notice, says that in Ohio’s Perry plant–the facility studied most closely–the failures were “definitely due to a flaw at some point in the quality check program.” Yet the NRC required “no specific action or written response,” asking only that reactor operators be alert.

If Phillips or anyone else in the NRC would heed the warnings of Chris Hall, a mechanical engineer in Albany, Ore., they might find a flaw in quality control pretty quickly. For more than three years, Hall has tried to tell the NRC that a manufacturing process used by his former employer, Teledyne Wah Chang Albany, could help explain the fuel rod failures.

In April 1990 Hall, a 13-year veteran of TWCA, approached company president Al Riesen and ethics officer Jim Ferguson with concerns about the way the company was producing the zirconium alloy tubes that are later made into fuel rods. Nine months later, after getting no response, he filed a complaint with the NRC. He was fired in July 1991.

In September 1991 Hall sued the company and several of its officers for $260 million for wrongful discharge, intentional infliction of emotional distress, fraud, and racketeering. Hall’s lawsuit charged that after he contacted the NRC, company employees harassed and humiliated him and that the company was guilty of “a cover-up of . . . shoddy goods” that pose “a risk to the life and health of millions of people around the world.”

Just over two years later TWCA and Hall settled out of court. Under the terms of the deal, Hall can’t say how much he was paid but is free to tell what he knows about TWCA’s zirconium production process. As an engineer who helped design the process, Hall knows a lot, and what he is saying is frightening.

Hall’s former employer, Teledyne Wah Chang Albany, is one of the four largest producers of zirconium in the world. Zirconium is used in showerheads and most recently in designer buttons, but 90 percent of what is manufactured worldwide feeds the nuclear industry. For years TWCA was the main supplier of zirconium alloy tubes to companies such as General Electric, which made the tubes into either water rods (used to slow nuclear reactions in reactor cores) or fuel rods.

Although nuclear reactors in the United States are each slightly different, they come in two basic varieties: GE’s boiling-water reactor, and a pressurized-water reactor made chiefly by Westinghouse. While zirconium alloy rods fuel both, the rods in GE’s reactors are subject to a particular type of failure that simply doesn’t occur in pressurized-water reactors–a rustlike scabbing called nodular corrosion. Like rust, the corrosion can ultimately create pinhole leaks in the tubes. This releases radioactive material into the coolant water and eventually into the environment.

In 1985 the zirconium industry was in a worldwide slump: 100 nuclear power plants, ordered during the 1960s and 1970s, were canceled after Three Mile Island cooled people’s enthusiasm for nuclear power. Around that time, customers like GE, which had experienced an excessive number of corrosion-induced failures, demanded a better manufacturing process. With the zirconium industry operating at 60 percent capacity, TWCA managers were in “a panic,” as one engineer stated under oath, to find an improved process to manufacture corrosion-resistant tubes for boiling-water reactors.


The nodules can cause pinholes in the tubes, allowing the nuclear material to escape into the coolant water and then into the environment.

Fuel rods weakened by nodular corrosion may be more susceptible to accelerated meltdown in case of other system failures–like a loss of coolant water surrounding the rods. If a meltdown occurred in the United States, it could result in massive loss of life and property. Experts say that the effects could be more devastating than those wrought by the 1986 disaster at Chernobyl in the former Soviet Union.

In 1987 the company brought on line the Beta Quench Tower Mark I, a machine intended to provide the sought-after corrosion resistance. Project leader Loren Humphrey said in a 1993 deposition that the Mark I was built as “fast as the humans that were there at that time could put it together.”

According to Chris Hall, the Mark I was a fairly simple device. “We heated [zirconium alloy] up to an elevated temperature, at which it could go through a beta phase change,” he explains. “And then we dropped it into water.”

Metallurgical principles suggest that the phase-changed zirconium alloy and any products made from it would have increased resistance to nodular corrosion–just what boiling-water reactors needed.

But from the very beginning, Humphrey and at least three other engineers were dissatisfied. “I designed the beta quench tower,” Humphrey wrote in a 1993 statement to the Department of Labor. “I was the one who raised the problems regarding this thing.”

Among other concerns, they questioned whether operators would be able to center the zirconium alloy in the Mark I’s heating coil consistently. If it is not centered, then the metal may not heat uniformly and go through the complete phase change. Chris Hall compares the difficulty of getting evenly heated zirconium out of a beta quench tower to that of getting a uniformly baked potato out of a noncarousel microwave oven: You almost always get overcooked and undercooked sections.

Hall believes, and a company metallurgist testified in a deposition, that not only may improperly beta-quenched zirconium alloy remain vulnerable to nodular corrosion, it may even be more susceptible to other types of corrosion; in other words, the process designed to improve corrosion resistance could actually degrade it.

A second generation of the beta quench machine, the Mark II, joined the Mark I in 1988, and later replaced it. But Loren Humphrey, Chris Hall, and other engineers remained concerned.

In response to their concerns, TWCA organized a team to analyze the beta quench process. In a February 1991 memo, the team informed TWCA management that the equipment used to center the zirconium alloy needed “immediate attention.” (Four months later, team members could not guarantee the problem had been fixed.) In April, the same team warned that “the coil and the support structure [of the beta quench tower] get a lot of physical abuse and routinely are moved out of the original position [which] affects heating.” Meanwhile, beta-quenched tubes were still being shipped to fuel rod manufacturers worldwide.

TWCA public affairs coordinator Jim Denham believes the company adequately investigated Hall’s charges. “We spent a good number of man-hours on it,” he says. “The basic finding was that the material we shipped met specification. Mr. Hall’s concerns were unfounded.”

But Hall and other engineers are still worried. They point out that it’s important that the beta quench process work, because the nuclear industry is pushing the new fuel rods harder. “In going to these new and improved tubes for nodular corrosion,” says Hall, “the presumption was that they could operate in a harsher environment. So operators started running reactors at a higher temperature. Well, that puts more stress on already frail tubes.”

“Overheating leads to embrittlement,” says Mary Olson of the Nuclear Information and Resource Service, a Washington, D.C., anti-nuclear group. “An overheated fuel rod is exactly like a superheated cast-iron skillet that cracks with the small bump that normally wouldn’t even dent it. A fuel rod has a higher possibility of failure when it runs longer and harder in a reactor. If you have a defective rod to begin with, it’s more likely to fail.”


The zirconium alloy is heated in the coil to undergo a BETA PHASE change, which increases corrosion resistance, then dropped in the bath for cooling. That’s what the beta quench tower (pictured at left) does. Critics say that it is virtually impossible to ensure that the zirconium alloy is evenly heated. So it may not undergo the beta phase change. After the zirconium alloy is extruded into tubes, the uranium is stacked inside. In reactors, the rods can develop NODULES and other types of corrosion, which threaten their integrity.

Despite the alleged problems with the beta quench procedure, TWCA’s manufacturing process might reliably produce acceptable fuel rods if the company’s quality control was rigorous enough to eliminate all insufficiently heated zirconium alloy. As TWCA’s Jim Denham says, his company ensures the quality of its product in two primary ways: through process control (which its own engineers question) and through sampling–but some say TWCA’s sampling is inadequate.

According to several sworn statements, TWCA quality assurance engineers usually test one end of three random tubes out of each batch of approximately 120 tubes (or 1.25 percent, given that each tube has two ends) for nodular corrosion resistance. But critics say that the stakes are too high to test such a small sample. “We’re not just talking about buildings here. They’re nuclear plants,” says Dr. Michio Kaku. “It’s ridiculous to only inspect 1 percent of all rods.”

Asked under oath if a sample of 1.25 percent is sufficient to control the quality of a batch of tubes, TWCA statistician Dr. Jerry Wille waffled. “I certainly did not make that determination, and I don’t know whether anyone else did or not,” he testified.

TWCA’s quality control process actually exceeds U.S. standards, which only require checking two samples in each lot. But those standards are set very close to home: TWCA’s manager of process analysis, Jack Tosdale, chairs the American Society for Testing and Materials committee that approves industry standards for zirconium.

The potential risk in using such a small sample is illustrated by Dr. Wille’s account of an incident that took place in October 1989. After testing six ends of a batch of zirconium tubes for nodular corrosion resistance, engineers found one tube to be improperly beta-quenched. Dr. Wille, testifying in a 1993 sworn deposition, called the discovery “very fortunate.” TWCA then tested all the tubes in the batch and ultimately rejected 17.

Although one company metallurgist admitted in a deposition that TWCA’s failure to give tubes sufficient corrosion resistance could affect its customers’ attempts to build reliable rods, TWCA defends its quality control process by saying that it is customer driven. TWCA says its customers are ultimately responsible for the behavior of fuel rods in reactors. “We are metallurgists. We’re not nuclear engineers,” says Jim Denham. “We’re responsible for manufacturing a product that meets customer specifications.” Denham says that because TWCA has no control over processes used by manufacturers to turn tubes into fuel rods, the company is “not in the business of guaranteeing the performance of what our material gets made into in its ultimate configuration.”

The NRC agrees that the customer, not the manufacturer, is responsible. NRC investigator Larry Phillips acknowledges that manufacturing flaws caused the fuel rod failures at Ohio’s Perry plant, a GE-supplied reactor. But he says wherever those flaws originated, the responsibility for detecting them lies with GE.

For its part, General Electric attributes the fuel rod failures at Perry to either fretting caused by debris in the coolant system (a finding the reactor operators dispute since they say there was no evidence of debris), or to undetected manufacturing flaws. Spokesperson Lynn Wallis says GE will not discuss its quality control procedures. Wallis also refuses to discuss GE’s suppliers and will not say whether any recent rod failures have involved zirconium alloy tubing from TWCA. Although GE canceled a TWCA contract for zirconium alloy tubes in the late 1980s, it is still probable that TWCA-manufactured rods, given their shelf and reactor life, are on line in GE reactors. Wallis, however, will not say whether those rods are still in use.

So where, in this saga, do we find the Nuclear Regulatory Commission, the federal agency charged with limiting the risks that reactors pose to public health and safety?

According to Dr. Kaku, the NRC relies on a “see no evil, hear no evil” approach. Rather than policing the industry, the agency allows power plants and their suppliers to oversee themselves.

A 1990 General Accounting Office report concluded that this policy historically resulted in the installation of substandard and even counterfeit parts in nuclear reactors. The report stated, for example, that after finding “problems with 12 utilities’ quality assurance programs” (out of 13 inspected), the NRC simply concluded that substandard products were an industrywide problem and, as such, weren’t the fault of individual utilities. Instead of stepping up enforcement, the NRC then suspended investigations for more than a year. The GAO report blasted the NRC for “deferring its regulatory responsibility” at a time when “an increasing number of commercial-grade products” used in nuclear reactors were of questionable quality.

In a December 1993 report, Public Citizen’s Critical Mass Energy Project analyzed internal nuclear industry documents. The watchdog group revealed discrepancies between the industry’s own evaluations of its operating problems and the NRC’s kinder, gentler report card. Internal industry documents highlighted problems at 86 commercial reactors: the NRC failed to address 185 of the 463 problems, and directly contradicted another 115.

When Chris Hall filed his 1991 complaint, the NRC investigated the TWCA plant for the first time in the history of its operation. Hall pointed inspectors to the records concerning the 17 faulty zirconium alloy tubes. The NRC investigation report dismissed the incident’s significance because those particular tubes were destined to become water rods rather than fuel rods. But Hall says that the beta quenching for water rod tubes and fuel rod tubes is identical. He says that if the process fails on water rod tubes, it is just as likely to fail on fuel rod tubes.

The NRC report does note that a quality control process that checks so few samples is insufficient to guarantee the product, but NRC officials still seem unconcerned. “We can’t do 100 percent coverage,” says Leif Norrholm, chief of the NRC’s vendor inspection branch. “If 10,000 pieces from one lot are made under common conditions, [a sample of] one should be sufficient.”

While investigating TWCA, the NRC inspectors did not, for lack of time, “verify accuracy and repeatability of [zirconium] positioning.” Despite the evidence–a string of internal memos, a company whistle-blower, a batch of failed rods, and their own admission in the report that “an inconsistent process would affect final product quality”–investigators observed fewer than a dozen beta quench runs, or one-tenth of a normal batch.

“What we’re interested in is the end-product quality, not the manufacture of the product,” says Uldis Potapovs, an NRC division chief who oversaw the TWCA inspection. As for the flawed batch of 17 zirconium alloy tubes, Potapovs says, “Even if they had gone out, it wouldn’t have mattered [because they were water rods].” Later, he adds, “They weren’t for domestic consumption, so they weren’t an NRC concern.”

NRC investigators also accepted TWCA’s assurances that the Beta Quench Mark II would be replaced with “a different design.” The NRC report on TWCA is dated Nov. 27, 1991. No replacement has yet been brought on line.

The NRC usually justifies its regulatory style by insisting that fears of a nuclear disaster are overblown. And, of course, it’s conceivable that no disaster will ever occur. But the possibility of a major release of radioactivity is not to be taken lightly.

“We do not fully understand the risks of nuclear power, and we should not be fearful of saying so,” former NRC commissioner James Asselstine wrote in 1986. “The operating experience of our existing plants and the industry’s failure to heed the lessons of experience indicate . . . that we can expect to see another serious accident in this country during the next 20 years.”

In the end, the NRC scarcely slapped TWCA’s hand. TWCA still uses the beta quench tower and still distributes fuel and water tubes worldwide. Their customers continue to fill TWCA-manufactured tubes with radioactive pellets and sell them to Mexico, Spain, Japan, Germany, and other countries. “You can pretty much assume that if these guys are in the business of manufacturing fuel, we supply them on some level,” says TWCA spokesperson Denham.

If NRC standards seem lax, it’s hard to glean much comfort from those of foreign regulatory agencies. Japan’s nuclear regulatory branch promotes nuclear energy and tolerates publicity stunts such as Pluto Boy, a cartoon figure who insists that plutonium has never been shown to cause cancer in humans. In Germany, allegations of cover-ups–perpetrated by safety inspectors reportedly told “not to find too many faults”–of some 130 cracks in one reactor’s cooling system offer no more solace. More disturbing, of all the countries receiving TWCA-generated rods, Japan and Germany are among the most technologically reliable. If safety concerns plague them, what is happening in countries like Mexico, which typically have less stringent quality controls?

“Fuel rod failure is like a time bomb,” Dr. Michio Kaku says. “Short term, we might have nothing, no problems. Longer term, who knows? We’re guinea pigs, to test whether we can run reactors that haven’t been manufactured successfully. And sure, guinea pigs do survive. But not always.”

Ashley Craddock is a fellow at Mother Jones.