Nuclear Fuel, Cladding, and the “Discovery” of Zirconium

Abstract

Zirconium cladding played an important role in both the Fukushima and TMI accidents. As nuclear reactors increased in size and power, metallic cladding materials were needed to protect fuel elements. Initially, aluminum and stainless steel were used, but neither of these materials is ideal for use in the high temperature, high radiation environments of nuclear reactors. Zirconium was identified as a potential “miracle” cladding material because it maintains its characteristics at high temperatures and is very corrosion resistant. At 0.185 b, zirconium also has a very low neutron absorbing cross-section. Unfortunately, zirconium is not perfect. At very high temperatures, zirconium reacts with steam to generate heat and to produce hydrogen gas. Both of these attributes can have major negative impacts during nuclear emergencies.

Introduction

Almost all nuclear power plants rely on slow-moving “thermal” neutrons to produce fission inside the reactor vessel. The geometry of a reactor’s design, and the materials used to build it are chosen to ensure that few neutrons are lost and sufficient numbers of neutrons are slowed to maintain reactor criticality. To continue fission, it is imperative that a balance in neutron population be achieved where the number of neutrons produced equals the neutrons used, plus the neutrons lost from the system. If too many neutrons escape or are absorbed by nonfuel materials inside the core, the reactor cannot be kept critical and the power level will drop rapidly. Fermi’s original Chicago Pile 1 achieved this balance by using very pure graphite as the moderating material, pure, unenriched uranium metal (0.7 % U²³⁵) as the fuel, and by assembling all the material into a generally spherical shape (Rhodes, 1986). Pure graphite has an extremely low probability of absorbing neutrons. A spherical shape inherently generates the lowest surface area to volume ratio, and thereby minimizes neutron escape from the core.

Nuclear reactors have evolved since the Chicago Pile 1. In both BWR and PWR reactors, water acts as the neutron moderator and coolant media. For fuel, instead of pure, metallic uranium, BWR and PWR types now rely on ceramic fuel pellets. To configure the fuel as pellets, fabricators convert uranium into a ceramic oxide, UO₂. Producing pellets from this ceramic allows them to remain intact at temperatures up to 5000 °F, instead of the 2070 °F melting point of uranium metal (Lamarsh and Baratta, 2001). Because the fuel pellet is the first defensive mechanism to retain highly radioactive fission products, the higher melting point of the ceramic allows for easier and safer integration of the fuel into an electricity producing power plant. By retaining fission fragments within the fuel pellet, radioactive materials are prevented from being released and plant safety is increased.

Within the reactor, the cylindrical pellets are arranged into long assemblies of fuel elements, normally referred to as fuel rods. An unfortunate side effect of using uranium oxide is that additional atoms (oxygen) are added into the reactor core. These additional atoms absorb some neutrons, and make it more difficult to reach and sustain criticality. Nuclear engineers have overcome the increased neutron losses inherent with uranium oxide by enriching the fuel. The concentration of U²³⁵ is increased from the natural concentration of 0.7 % up to levels of 2–4 %. Enriching fuel with U²³⁵ means that more U²³⁵ atoms are present that can undergo fission when struck by a proper (low, thermal) energy neutron. Although U²³⁸ does not undergo fission, it can be transmuted into a fuel source by absorbing a neutron. If this occurs, instead of fissioning, the U²³⁸ will eventually decay into Pu²³⁹, a radioactive isotope that can undergo fission in a similar manner to U²³⁵. Because U²³⁸ comprises 96–98 % of the fuel, it will regularly absorb a neutron and produce Pu²³⁹ (Knief, 2008).

Both BWR and PWR reactors rely on enriched fuel, but the high cost of enriching uranium limits how high a concentration of U²³⁵ can economically be achieved. The US military often prepares enrichments of over 90 % U²³⁵ (Chunyan, 2001), but these defense agencies are not bound by the same financial constraints as commercial plants. Military users also tend to be focused on solving different engineering problems than civilian users. For example, the US Navy seeks to operate shipboard reactors for 10–30 years without refueling (Anon., 2013).

Both the design and operation of commercial nuclear power plants are planned with the safety of the public foremost in mind. Therefore, plants rely on multiple barriers to protect the public from the highly radioactive fuel and the multitude of radioactively “hot” materials created in the core. While the fuel itself is obviously radioactive, normally nonradioactive materials can also become hot by neutron absorption and gamma irradiation while inside the core.

Fission of U²³⁵ will produce two smaller atoms, but the exact chemical makeup of those fission products has its own variability and probability. Regardless, these two smaller atoms are not very mobile and are generally trapped within the ceramic fuel pellet (Fig. 10.1).

Fig. 10.1

Typical UO₂ Ceramic Fuel Pellets

The ceramic fuel pellet represents the first barrier to the release of radioactivity from a nuclear power plant (see also Chap. 7). To keep the water coolant from touching the radioactive fuel, a thin cladding of metal is wrapped around the fuel pellet. The metallic cladding provides a second barrier to the release of radiation. By keeping the fuel pellets from touching the coolant, the vast majority of the fuel’s radiation stays within the ceramic pellets. The cladding also prevents the coolant from attacking or interacting with the fuel, thereby maintaining the fuel’s integrity (US Nuclear Regulatory Commission, n.d.).

The metal used for cladding has several constraints placed on it if it is to be used in a commercial nuclear power reactor. First, it must be thermally conductive to allow the heat of fission to leave the fuel pellet and efficiently interact with the coolant. This allows the fuel pellet to remain well below its melting point, and thus prevent the release of radioactive fission products. Second, the cladding metal must be noncorrosive so that it will remain intact during the fuel cycle while constantly being immersed in the coolant. This means that the cladding cannot oxidize so that it will retain its strength inside the high radiation field of the reactor core. Third, the cladding must not be a strong absorber of neutrons. If the cladding absorbed a significant amount of neutrons, fuel enrichment would need to be increased to overcome this parasitic loss. As fuel enrichment costs are the strongest driver in nuclear fuel costs, increasing enrichment significantly affects the final nuclear fuel costs. Fourth, the cladding must have sufficient strength to stay intact and provide structural support during normal operation, especially at the elevated temperatures found in BWR, PWR, and RBMK plant operations (~600 °F coolant temperature) (Lamarsh and Baratta, 2001). Figure 10.2 shows the temperature difference between a typical PWR fuel pellet and the water of the Reactor Coolant System.

Fig. 10.2

Fuel Pellet, Cladding, and water coolant temperatures (US Nuclear Regulatory Commission, n.d.)

Development and use of cladding evolved as the nuclear power industry developed. Fermi’s Chicago Pile 1 did not rely on cladding. It ran at a low enough power that the natural circulation of air and conduction through the graphite moderating blocks was sufficient to keep the fuel intact. Because the room air convected through the low-power core at very low velocities, there was very little chance of radioactivity leaving the fuel slugs.

As the nuclear industry advanced, reactor cores increased in physical size, core power, complexity, and location. Fermi and his team next moved their reactors from within the city limits of Chicago to a rural location west of the city; a site that would become the Argonne National Lab after the war. The Chicago Pile 2 (CP-2) was the first reactor built at the Argonne site, and the second overall reactor. This was essentially a recreation of the initial Chicago Pile 1, and like the CP-1, the CP-2 operated at a low enough power level that natural air cooling was sufficient to cool the reactor (Anon., unknown). Therefore, no cladding was incorporated into the fuel elements.

The third reactor built by the Argonne team, the CP-3, was a new design that came online in May 1944. It was built as a prototype plutonium production plant for the wartime bomb effort. This reactor used liquid heavy water (D₂O) as both a coolant and a moderator. In the course of events, the CP-3 ended up not being needed. The large X-10 pile, a plutonium production graphite moderated prototype reactor built at Oak Ridge, Tennessee, and the plutonium production reactors built in Hanford Washington used aluminum clad fuel elements. However, the CP-3 did represent a fundamental shift in fuel usage, because it was the first reactor to incorporate aluminum cladding (Anon., unknown).

Aluminum was considered an acceptable cladding material because it has a low neutron absorption probability, sufficient strength to retain fission products, and adequate corrosion resistance for the fuel’s expected lifetime. While aluminum will oxidize (i.e., corrode) in high temperature water, the CP-3 core had sufficient cooling flow that the temperatures were never expected to exceed ~100 °F (Anon., unknown). In addition, the wartime exigencies allowed the core designers leeway in incorporating aluminum as a fuel clad material. Any corrosion of the cladding was considered acceptable during this time period, as the emphasis was more on learning how to operate a reactor and in producing plutonium for a bomb, than in trying to prevent the potential release of radioactive fission products from the failure of fuel cladding.

Concurrent with advancements at Argonne, the Oak Ridge, Tennessee X-10 pile became operational in November 1943 (Rosenthal, 2009). The sole purpose of the X-10 pile was to demonstrate the large scale viability of converting U²³⁸ into Pu²³⁹ in the high neutron field of a nuclear reactor. This large reactor (power output 1 MW) used air cooling just like the original CP-1. As the reactor was being designed, the engineers recognized that its fuel would require a cladding. The uranium in the X-10 pile would be in the core for long periods of time and exposed to large velocities of cooling air forcibly circulated through the core. The higher power level, longer fuel cycle duration, and higher air velocities all dictated the use of a fuel cladding. Because no water was incorporated into the reactor, the designers had little concern about corrosion and relied on aluminum to protect the individual fuel slugs. The aluminum cladding also provided a means to capture the derived plutonium by allowing it to be retained within each individual fuel slug.

The X-10 pile provided a major step forward in the World War II atomic bomb program. It demonstrated the ability of graphite moderated cores to produce significant quantities of plutonium. The first batch of plutonium from the X-10 was only 500 mg, but this still represented the largest mass of plutonium yet generated (Rosenthal, 2009). While the X-10 reactor demonstrated the proof of concept, the Hanford reactors would be the work horses of plutonium generation. They would create the 6 kg of Pu needed for the weapon that was eventually dropped on Nagasaki (Rosenthal, 2009).

The first Hanford Reactor B began operating in 1944 (it was shut down in 1968, and is now a National Historic Landmark) (National Park Service, 2007). This large graphite moderated reactor generated 250 MW of energy, over two hundred times greater than the X-10 pile. The large increase in power, coupled with new knowledge about the absorption of neutrons, led the designers to use water cooling for the B reactor, and all subsequent Hanford reactors. Water has a much greater capacity for absorbing heat than air, making it a generally preferred heat transfer media. While water has ~1000 times greater neutron absorption probability than air, the discovery of increased neutron production from fission allowed for water to be introduced into the core to help maintain criticality. The Hanford Reactor B represented an enormous increase in power compared to all previous reactors. Its sole purpose, however, was to convert U²³⁸ into Pu²³⁹ for atomic weapon’s development. The energy created in the Hanford reactor during the war was considered superfluous, and was simply dumped as thermal waste into the Columbia River (Gosling, 2010).

While the X-10 pile in Oak Ridge could generate up to 1 MW of power, the 250× increase in the Hanford reactor power represented the next step in reactor development. In fact, the facilities at Hanford were such an advancement that they were no longer referred to as piles; instead they were known as nuclear reactors. Aluminum cladding was used in both the X-10 and the first Hanford B reactor. Although the Hanford reactors were expected to reach temperatures high enough to corrode the aluminum cladding, the fuel elements were only expected to spend a few weeks in the reactor. The duration of exposure was limited to this short time frame because longer stays would actually start “burning” the Pu²³⁹ (“burning” is when the Pu²³⁹ absorbs another neutron and converts to non-fissile Pu²⁴⁰). Once the fuel elements had completed their stay in the core, they were pushed out of the reactor and dropped into a water bath.

Aluminum worked as a convenient transitional fuel cladding, but finding an appropriate cladding that could withstand long-term exposure in a nuclear reactor was very important. Aluminum was not an ideal choice. It did have a high thermal conductivity that allowed the fission energy to be easily conducted from the fuel elements, but it also had higher than desired affinity to absorb neutrons (Kaufman, 1962). Additionally, aluminum was not very corrosion resistant, and its strength was reduced at the elevated temperatures found inside the core of a nuclear reactor. Thus, aluminum cladding could work within some limited settings, but it was a poor choice when fuel elements needed to operate for long periods inside the core and were subjected to high temperatures, or were in contact with water at temperatures only slightly above normal room temperature.

Despite its drawbacks, aluminum remained the cladding material of choice as the initial batch of research reactors became operational in the post-war period between 1945 and 1952. During the same period, a second cladding material, stainless steel, also began to be used. Compared to aluminum, stainless steel had vastly superior corrosion resistance properties, especially when used in reactors with aggressive coolants like sodium. While stainless steel offered improved corrosion resistance versus aluminum, it did have two significant drawbacks. First, it was about 10× lower in thermal conductivity than aluminum, meaning that coolant temperatures would have to operate lower for the same fuel temperature. Second, and perhaps more importantly, stainless steel also had about a 10× increase in neutron cross-section, making it much more likely to absorb a neutron than aluminum (Kaufman, 1962).

It was the development of the nuclear reactor for navy propulsion that led to the use of zirconium and zirconium alloys as a fuel cladding material. This program, headed by Captain, and eventually Admiral, Rickover demonstrated the viability of using fission energy to propel navy ships and submarines.

Aluminum and stainless steel could not be used for the US Navy nuclear propulsion program. The PWR reactor that was being designed for the Nautilus would need to encapsulate fuel elements for years, and would have to operate with water coolant temperatures well in excess of 500 °F (R. G. Hewlett and Duncan, 1974). Aluminum had only operated in reactor cores for short operating periods (weeks) and due to corrosion concerns could only be exposed to water at low temperature (<250 °F). Stainless steel could operate at higher temperatures, but its neutron absorption was higher than desired, so using stainless steel cladding would cause a reactor to have a relatively short core life. The quest was on for a new material (R. G. Hewlett and Duncan, 1974).

In the early stages of the nuclear industry, zirconium had been discounted as a potential cladding material. It was expensive to produce, with costs around $600 per pound, and the quantity initially available was miniscule. In 1943, there were only 2 oz present in the entire US. By the end of the WWII, the situation began to change. The Foote Mineral Co. had produced 21 lb of zirconium for investigation of its fundamental properties as a corrosion resistant material for use in the airplane industry. The increased production caused the cost to drop, but it still was prohibitively expensive at ~$300 per pound (R. G. Hewlett and Duncan, 1974). More zirconium was produced after the war, and the price continued to drop. With the falling prices, interest in the material for the nuclear industry expanded (Enghag, 2004).

Zirconium had long been attractive to nuclear engineers because it possessed high strength, which was retained at high temperatures, and resisted corrosion even at elevated temperatures. These material characteristics could be very useful in a nuclear reactor, especially as they increased in power level and anticipated the use of water as a coolant. As attractive as zirconium was, however, scientists originally believed that its neutron absorption cross-section (the probability that it would absorb a neutron) was too high for use in cladding.

Based on the data available in 1947 the neutron absorption cross-section for zirconium was no better than stainless steel. Iron, the principal element in stainless steel, had a microscopic cross-section of 2.55 b (Lamarsh and Baratta, 2001). The barn had been adopted by early physicists as a measure of area for the incredibly small values at the atomic levels. Each barn represented 10⁻²⁴ cm². In a rare show of humor, the term “barn” arose from the expression of “hitting the broad side of a barn.” Zirconium had been found to have a cross-section nearly identical to iron, 2.5 b, thus seeming to offer little advantage.

Tantalized by the prospect of zirconium’s good physical characteristics in high temperature environments, researchers began to reinvestigate this material. As they did so, they realized that their initial neutron absorption cross-section calculations were inaccurate. In natural settings, zirconium is always found in association with hafnium, and the two metals share an extremely close affinity. The first methods used to purify zirconium ore were not sophisticated enough to remove the associated hafnium. Therefore, the early material samples used to measure zirconium’s neutron absorption were contaminated by the presence of small amounts of hafnium (~1 % by mass) (Krishnan, 1981). This small level of contamination did not affect the physical properties of the zirconium, but it greatly affected the samples cross-sectional absorption. Pure hafnium has a neutron absorption cross-section of 102 b, much larger than pure zirconium, and large enough to alter the reported absorption rate of neutrons on the impure zirconium samples (R. G. Hewlett and Duncan, 1974).

Further studies by researchers at MIT and ORNL rechecked the absorption cross-section of pure zirconium samples. They found the cross-section to be nearly a factor of 5× lower than initially reported. Its actual value was calculated as being between 0.4 and 0.5 b, well below the level of stainless steel. This new level meant that zirconium was significantly better than stainless steel at allowing neutrons to pass through it without being absorbed. Even more recent data has zirconium listed with a neutron absorption cross-section of 0.185, over a tenfold reduction in absorption cross-section over the initially listed value (El-Wakil, 1962).

The large decrease in neutron absorption cross-section for zirconium led Admiral Rickover to vigorously pursue using zirconium in the Submarine Thermal Reactor (STR) being designed by Argonne National Lab and Westinghouse. He realized that zirconium cladding would produce a smaller reactor that could operate for longer periods of time between refueling. These attributes would be significant advantages for the reactor that would first be operated as a land-based prototype and then installed in the Nautilus.

Zirconium and its alloys (primarily Zirc-2 and Zirc-4) appeared to be a panacea for the naval and commercial nuclear industry. It was a corrosion resistant metal that retained its strength at the high temperatures found inside nuclear reactors. It did not embrittle after being exposed to the high neutron flux inside reactor cores. It also absorbed fewer neutrons compared to its chief competitor stainless steel. Some analyses even indicated that reactors using zirconium fuel cladding could operate with 1 % lower U²³⁵ enrichment than stainless steel, an enormous cost savings given the exorbitant price of gaseous diffusion fuel enrichment (Whitmarsh, 1962). Finally, due to the increased demand from the naval and commercial nuclear industries, new zirconium processing technologies were adopted and the cost for zirconium metal dropped throughout the 1950s.

Both the naval and commercial nuclear industries adopted zirconium alloys as their preferred cladding material. There was, however, one small problem with the new miracle material zirconium and its alloys. At elevated temperatures (~1000 °C for zirconium alloys) the metals react with water to give off heat and to produce hydrogen gas. The zirconium water (steam) reaction can occur at temperatures well below the melting point of the metal (1852 °C for pure zirconium) (Knief, 2008). The equation below shows the fundamental reaction of zirconium with water, its reaction products, and heat of reaction.

Zr + 2H₂O → ZrO₂ + 2H₂ 140 kcal/mol zirconium

In an accident situation, the large energy release of this oxidation reaction coupled with hydrogen gas generation could lead to additional damage and reactor failures. The released energy can function as a positive feedback mechanism and exacerbate a critical situation. The hydrogen gas is not only very difficult to handle, but it is also highly explosive. The risk of hydrogen explosions becomes a very real threat in any nuclear accident involving zirconium cladding (Baker Jr and Just, 1962).

It should be noted that stainless steel also undergoes a very similar reaction with steam to produce hydrogen and release energy. The stoichiometry of the stainless steel reaction with water is identical to the reaction above, except that Zr is replaced with Fe. However, the energy released from a stainless steel and steam reaction is about 10× lower than the zirconium reaction and hence generally less dangerous (Knief, 2008).

The decision by Argonne, Westinghouse, and Rickover to use zirconium as the cladding material for the upcoming Nautilus reactor core appears to have been the result of a neutron’s versus hydrogen calculation. While the designers were aware of the high temperature metal-steam reaction, and the potential for hydrogen gas production, the new data demonstrating the low neutron absorption capability of the pure zirconium proved more attractive. By designing the naval core with zirconium, engineers would have a higher core power per unit reactor volume—not a trivial matter when faced with the very tight quarters inherent in submarines. It appears that once the US Navy adopted zirconium and zirconium alloys for use in the PWR that powered the Idaho Mark I prototype, the Mark II Nautilus powerplant, and eventually dozens of follow-on ship reactors, the commercial world simply followed suit (R. G. Hewlett and Duncan, 1974).

Despite the potential dangers of zirconium, hydrogen gas generation was seen as a low-probability risk. Emergency systems would be incorporated into all reactor designs so it was thought that any accident situation where hydrogen gas would be produced would be extremely rare. Unfortunately, in the real-world setting, reactor operators in Pennsylvania, the Ukraine, and Japan would demonstrate that these conditions were not as rare as first believed.

Zirconium has now been used as the predominant cladding material for nuclear fuels for decades. It has yielded positive results in day-to-day reactor operations, but has also contributed significantly to several of the world’s major nuclear disasters. Given its drawbacks, research is underway to find alternate cladding materials. The US Department of Energy has a program underway to investigate accident-tolerant fuels (Carmack, 2013). This program is investigating materials that will not generate hydrogen, or otherwise exacerbate reactor conditions, during an accident. Certain ceramic materials may provide a technical solution. These substances provide similar low neutron absorption features and strong corrosion resistance and strength under the high temperature and high radiation fields encountered within a reactor. However, this research program is not yet mature, and investigations are ongoing (Fig. 10.3).

Fig. 10.3

PWR Fuel Assembly using zirconium alloy cladding and support structures