Hydrogen Permeation in FeCrAl APMT Alloy for Accident Tolerant Fuel Cladding

Even though there was no loss of human life as a consequence of this event, the explosion of the accumulated hydrogen in the external buildings caused concern about the safe use of nuclear energy to generate electricity. After the accident, the global nuclear materials community acknowledged the need to develop and implement more accident tolerant materials. One of the most developed ATF concepts is to use monolithic iron-chromium-aluminum (FeCrAl) alloys to replace zirconium alloys for the cladding of the fuel.^2-3  Ferritic FeCrAl alloys are suitable cladding materials because they have a low coefficient of thermal expansion, high thermal conductivity, good mechanical properties at normal operation temperatures, and outstanding resistance to attack by steam at a loss of coolant accident conditions (T > 1,000°C). The two less desirable attributes of FeCrAl alloys are their higher than zirconium capture of thermal neutrons and higher than zirconium permeation of tritium.^1-4 

The proposed use of FeCrAl for cladding of light water reactor fuels is being currently investigated in the international materials community. However, there are two attributes of FeCrAl that may not make them ideal for cladding application: (1) a higher neutron absorption cross section than the current zirconium alloys; and (2) the possibility of a higher than current tritium release into the coolant.^2,5-6  Regarding the parasitic absorption of neutrons, the FeCrAl cladding wall was made 0.3 mm thick or approximately half the thickness of the current Zr alloy to make it cost neutral. The wall thickness reduction is possible due to the superior mechanical properties of FeCrAl compared to zirconium alloys at temperatures in the order of 400°C.⁷  Also, it is likely that the tritium release into the coolant will decrease as a function of time because of the development of oxide films on the surface of the cladding, i.e., alumina from the fuel cavity side and chromia from the coolant side. Both of these oxides would make tritium less likely to migrate from the fuel cavity to the coolant (Figure 1).^2,6 

The Electric Power Research Institute (EPRI) reported that when austenitic stainless steel claddings were used for power generation the amount of tritium in the coolant water was approximately 10 times higher than when zirconium claddings were used.⁸  As FeCrAl are ferritic (bcc) in nature, the diffusion of tritium through the cladding wall into the coolant could be even higher than when the austenitic (fcc) iron-based material was used.⁶  It is known that hydrogen intake into iron alloys decreases in the presence of oxide films on the surface of the iron-based alloys.^9-11 

Recently, tritium permeation studies were conducted on FeCrAl oxide dispersion strengthened (ODS) alloys.^12-13  Sakamoto, et al., studied the permeation of hydrogen as a function of the temperature on noncorroded FeCrAl specimen and on an FeCrAl ODS specimen immersed for 30 d in water with 8 ppm of dissolved oxygen at 290°C. They reported a modest decrease, by a factor of less than 10, in the tritium permeation through the preoxidized specimen as compared to the nonoxidized FeCrAl ODS specimen.¹³ 

One of the common failure mechanisms of the current cladding of zirconium alloys is the embrittlement by hydrides. Atomic hydrogen can form on the surface of the zirconium cladding and diffuse into the cladding wall and react with the metal to form stable zirconium hydrides. The hydrides tend to develop near the outer diameter (OD) of the zirconium alloy fuel rod due to the temperature gradient across the cladding wall. Hydrogen is less soluble in the zirconium alloy matrix near the cooler OD and may react with zirconium to form hydrides. By using FeCrAl alloys for the cladding, the hydride in the cladding issue is eliminated because none of the elements in FeCrAl (Fe, Cr, Al, or Mo) react with hydrogen (or tritium) to precipitate stable metal hydrides in the manner that Zr does.¹⁴ 

The literature contains limited data related to hydrogen or tritium permeation through ferritic alloys of FeCr and of some FeCrAl, but not advanced powder metallurgy tubing (APMT).⁶  The purpose of the current research was to use the Devanathan-Stachurski cell (captured in standards ASTM G148 and ISO 17081)^15-17  to measure the permeation of hydrogen at 30°C through clean (freshly polished) FeCrAl APMT alloy. The use of this technique, where atomic hydrogen is generated on the charging side by applying a cathodic potential, has never been reported in the literature for FeCrAl alloy.

Before running the permeation tests, the electrochemical behavior of APMT was characterized in relevant electrolytes. Four types of electrochemical tests were conducted: (a) linear polarization resistance (LPR) tests, which involved scanning the potential from −20 mV below the open-circuit potential (OCP) to +20 mV above OCP at a rate of 0.167 mV/s; (b) electrochemical impedance spectroscopy (EIS) by applying 10 mV (rms) sine wave with respect to OCP and measuring the impedance of the system from 100 kHz to 10 mHz; (c) cyclic potentiodynamic polarization (CPP) tests, which consisted of scanning the potential at a rate of 0.167 mV/s from −150 mV_OCP until reaching a current density of 1 mA/cm² and then reversing the scan direction until reaching −100 mV_OCP; and (d) potentiostatic test by applying a constant potential of +300 mV_SCE or +400 mV_SCE. Prior to conducting any of the polarization measurements, the OCP was monitored for 24 h while test conditions stabilized.

The electrochemical evaluations were performed under deaerated conditions (high-purity N₂) at 30°C in the following four electrolytes: (1) 3.5% NaCl; (2) 0.1 N NaOH; (3) 1 N H₂SO₄; and (4) 0.5 M Na₂SO₄. These electrolytes were selected because they spanned a wide pH range and involved typical anions such as Cl⁻ and SO₄^2–. However, these electrolytes are not intended to represent the water chemistry in contact with the fuel rods in a reactor. A conventional three-electrode setup was used for conducting the measurements. A saturated calomel electrode (SCE) was used as a reference electrode, platinized-niobium served as a counter electrode, and the APMT specimen was the working electrode. The specimens were wet-polished with 600-grit silicon carbide paper within 1 h of starting the experiments. The exposed area of the specimens was approximately 1 cm². The density of APMT is 7.25 g/cm³ and its equivalent weight is 22.7 (unitless).

Hydrogen diffusion through APMT specimens was determined using the Devanathan and Stachurski¹⁵  cell and according to the ASTM G148 and ISO 17081 standards.^16-17  The permeation setup consisted of a thin APMT specimen (membrane) that was clamped between a hydrogen charging cell and an oxidation cell to create a hydrogen entry and a hydrogen exit side, respectively. The charging cell was filled with 3.5% NaCl and a cathodic potential was applied to the entry side of the specimen to generate hydrogen. The oxidation cell was filled with 0.5 M Na₂SO₄ solution and an anodic potential was applied to the exit side of the specimen to detect hydrogen. Atomic hydrogen (H) was generated on the cathodic side and it diffused through the specimen to the anodic side where it was oxidized to H⁺, which produced an oxidation current. This oxidation current is proportional to the amount of flux of hydrogen that permeates through the specimen.

The APMT surface that was exposed to the detection side was not electroplated with palladium, which is a common practice to reduce the background current and enhance oxidation kinetics.¹⁸  Instead, the specimens were tested with a bare surface finish condition (600-grit), which still allowed for very low background currents to be achieved based on the careful selection of the oxidation solution. The exposed area of the APMT specimens was approximately 1 cm² on each side of the permeation cell.

Based on the CPP data that was obtained on APMT, it was decided that 3.5% NaCl solution could be used in the charging cell because there would be no concern of crevice corrosion under the applied cathodic potential (−1 V_SCE). However, 3.5% NaCl solution would be a poor choice for the oxidation cell because of the threat of crevice corrosion under an oxidizing potential of +300 mV_SCE. In the available standards for hydrogen permeation testing, it is recommended that an alkaline solution, such as 0.1 N NaOH, be used in the oxidation cell because it is conducive to passivating low-alloy steels. But for APMT alloy, the 0.1 N NaOH solution did not appear to be a good option for the oxidizing solution because of the higher “passive” current densities that were observed (Figure 4). Ultimately, the 0.5 M Na₂SO₄ solution was considered the most suitable electrolyte for the oxidation cell because it led to low passive currents over a wide potential range.

The hydrogen permeation tests were performed on APMT specimens at 30°C and under deaerated conditions. The specimen side in the charging cell was exposed to 3.5% NaCl solution and was cycled between cathodic (−1 V_SCE) and OCP conditions. The specimen side in the oxidation cell was exposed to 0.5 M Na₂SO₄ and was maintained at a potential of +0.4 V_SCE. The specimens were bare APMT sheets of three different thicknesses (0.44 mm, 1.10 mm, and 1.99 mm). Three transients of hydrogen charging and discharging were conducted for each specimen thickness.

Both methods can be subject to sources of error in the approximation of D_eff. For example, an unstable or inconsistent baseline current can lead to poor estimations of t_b in the break-through method while the inability of a system to achieve “steady-state” permeation current (J_ss) often confounds the application of the t_lag method. In the literature, it seems that errors are more frequently encountered in the t_lag method because there are a number of systems that can lead to surface films such as metal sulfides, corrosion layers, etc., which manifest as peaks in the permeation current rather than steady-state flux. The formation of voids in a material can also interfere with the attainment of steady permeation current. In this work, these issues were not encountered in the permeation data, which allowed for both methods to be utilized without any special caveats for the reported D_eff values. The calculated values of D_eff for the three transients that were obtained for each membrane thickness are summarized in Table 1. The D_eff values were estimated using both the t_b and t_lag methods, which are shown in separate columns in the table. Both methods for calculating D_eff yielded the same average value of 2.8 × 10⁻⁸ cm²/s, with only small differences in the standard deviation.

This is the first time that the Devanathan-Stachurski cell has been used to measure effective hydrogen diffusion coefficients through APMT alloy. Current findings show that the effective diffusion of hydrogen through ferritic APMT at 30°C is 2.8 × 10⁻⁸ cm²/s (Table 1), which is approximately three orders of magnitude lower than the D_eff values reported for pure iron (8 to 9.5 × 10⁻⁵ cm²/s).^24-26  The presence of alloying elements in APMT (Cr, Al, and Mo) would decrease the movement of hydrogen through the metal.⁶  The reduction in the effective diffusion of H in steel-containing alloying elements compared to pure iron is an established fact.⁶  The steady-state flux of hydrogen through APMT in Figure 10 is for cathodically generated atomic hydrogen on a clean surface. It has been shown many times that the presence of surface oxides on the charging side may limit the amount of atomic hydrogen that is generated and absorbed by the metal, which would affect migration to the other side of the membrane.⁶  Figure 6 shows that the membrane specimen did not show obvious oxide formation on the discharge or oxidation side of the membrane when the sodium sulfate electrolyte was used. Even though the current findings are not intended to actually embody the behavior of a nuclear fuel rod, it is a simplistic analogy. The APMT specimen that was used in this study is a representation of the cladding barrier between the fuel cavity (where the atomic hydrogen may be charged into the cladding wall) and the water side (where the permeated atomic hydrogen through the cladding wall would be oxidized to water). A recent review article⁶  showed that the permeation rates of hydrogen through FeCrAl such as APMT is between those of ferritic FeCr steels, which have faster diffusivities, and those of austenitic FeCrNi steels, which have slower diffusivities compared to APMT. As stated in the introduction, one of the concerns of using FeCrAl alloys such as APMT for the cladding of the fuel is the risk of an increase in the tritium concentration in the water. This risk of tritium increase in the coolant will be reduced once the cladding develops an aluminum oxide in the fuel cavity (charging side) and a chromium oxide on the water side. It has been reported that oxides can reduce the flux rates of hydrogen by a factor of 1,000.⁹  Thermodynamic calculations have been conducted which show that the dissociation of uranium dioxide in the fuel cavity may generate enough partial pressure of oxygen to allow for the formation of alumina on the inner diameter of the cladding. These predictions still need to be corroborated by performing post irradiation examinations on FeCrAl rods that are currently under irradiation in test and commercial reactors.

• The Devanathan-Stachurski cell technique, which was the foundation for the ASTM G148 and ISO 17081 methods, was used to measure the effective hydrogen diffusion coefficient through APMT at 30°C.

• For evaluating APMT specimens, it was determined that the most suitable electrolyte to use in the oxidation/detection cell was 0.5 M Na₂SO₄ and not the traditionally used NaOH electrolyte.

• The average D_eff for APMT was found to be 2.8 × 10⁻⁸ cm²/s for the tested conditions.

• A linear relationship was found between the hydrogen flux and the inverse of the APMT specimen thickness, which indicates that the D_eff values that were determined corresponded to bulk diffusion of hydrogen.

The authors would like to express their gratitude to Dr. Feng Gui (DNV, USA) for his help in analyzing the permeation data, Steve J. Buresh at GE Research for the thermal-mechanical processing of the APMT material, and Dr. Shenyan Huang (GE Research) for performing the metallographic studies of APMT membranes. This material is based on work supported by the Department of Energy [National Nuclear Security Administration] under Award No. DE-NE0009047. The financial support of Global Nuclear Fuel and GE Research is gratefully acknowledged. This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of the authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.