38 (1994) 13S-145
Effect of hydrogen gas on the mechanical properties of a zirconium alloy J.-H. Huang
and C.-S. Ho
Department of Nuclear Engineering National Tsing Hua University Hsinchu (Taiwan, R.O. C.)
(Received February 24, 1993; accepted December 22,1993)
Abstract To study the mechanical properties of Zircaloy-4 in hydrogen gas, notched Zircaloy-4 plate specimens were tested in a hydrogen gas environment of various pressures up to 2020 kPa at 25, 100 and 200°C. The results showed that, for all test temperatures, the notch tensile strength of Zircaloy-4 was slightly increased with hydrogen pressure until 101 kPa and then decreased with increasing hydrogen pressure. A ductile-brittle transition of reduction of area at 25°C was found when the alloy was tested at hydrogen pressure between 101 and 1010 kPa. From the fractographic finding, the ductile-brittle transition was closely related to the precipitation of brittle hydrides. The ductile-brittle transition became Lessdistinct at 100°C and above. This was attributed to the improved ductility of zirconium matrix with increasing temperature.
Introduction Zirconium alloys are generally used as fuel element cladding and an in-core structure component in light water reactors. It is well recognized that the mechanical properties of zirconium alloys can be degraded by the presence of hydrides. Due to their technical significance, the embrittlement of zirconium alloys by excessive hydrogen uptake and subsequent precipitation of hydrides have been extensively studied [l-3]. Since hydrogen gas may be present inside fuel elements or from corrosion of zirconium alloys in water, it is crucial to assess the gaseous hydrogen embrittlement on zirconium alloys. However, the effect of hydrogen gas on the mechanical properties of zirconium alloys is less well known. Only two literatures are found on the subject of gaseous hydrogen-induced cracking of zirconium alloys. Nelson and Wachob  carried out slow crack growth experiments on annealed Zircaloy-4 in gaseous hydrogen environments. They found that the crack growth rate increases with increasing hydrogen pressure and temperature. Coleman and Cox  performed gaseous hydrogen-induced crack propagation experiments on Zr-2.5Nb and Zircaloy-2. They pointed out that zirconium alloys can crack in hydrogen gas at Elsevier Science S.A. SSDI 0254-0584(94)01360-s
velocities much higher than those associated with delayed hydride cracking at the same temperature. Both studies investigated the crack growth behavior in hydrogen gas at low pressures, less than 100 kPa, and at temperatures below 80°C. To understand the embrittlement of zirconium alloy in hydrogen gas, in addition to crack growth behavior, the study on the change of mechanical properties due to hydrogen is important. However, there has been no previous work on this subject. In the present investigation, to assess the effect of hydrogen gas on the mechanical properties of Zircaloy4, notch tensile tests were performed on the uncharged Zircaloy-4 specimens at various temperatures up to 200°C in hydrogen gas environments with pressures up to 2020 kPa. An interesting phenomenon of hydrided zirconium alloys reported previously is the room temperature ductile-brittle transition between certain hydrogen contents. Lin et aE.  observed the ductile-brittle transition on hydrided Zircaloy-4 alloys at hydrogen contents between 400 and 800 ppm. Recently, Bai et al. [7-91 also found that Zircaloy-4 undergoes a ductile-brittle transition at room temperature when the hydrogen content in the specimen is higher than some critical value depending on the microstructure and the hydride morphology.
J.-H. Huang, C.-S. Ho / Materials
Chemistry and Physics 38 (1994) 138-145
Their definition of “ductile-brittle transition” is the abrupt change of the value of reduction of area, which is somewhat different from the conventional definition determined from impact tests. In this paper we followed the definition of Lin and Bai. These observations of ductile-brittle transitions were performed using smooth tensile specimens in standard tensile tests. Since notched specimens provide a triaxial state of stress, which will enhance brittle fracture, they are more suitable to be used to study the ductile-brittle transition of metals than smooth tensile specimens. In a previous research, Huang and Huang [lo] conducted tensile tests on notched specimens of hydrided Zircaloy-4. They found that, due to notch effect, the ductile-brittle transition is at a lower hydrogen content (30-140 ppm) than those reported by Lin et al. and Bai et al.. They further examined the effect of temperature and found that the ductile-brittle transition becomes less distinct at 100°C and above. Although the ductile-brittle transition for hydrided zirconium alloys has been well recognized, no previous study has been done on Zircaloy-4 alloy to determine whether the ductile-brittle transition is existed when tested in a hydrogen gas environment. The purpose of this study is to determine the existence of ductile-brittle transition for Zircaloy-4 tested in hydrogen gas.
Experimental Procedures The specimen was made from commercial Zircaloy4 purchased in plate form. The material had a composition of Sn 1.391, Fe 0.193 and Cr 0.103 wt.% and impurity (in parts per million) of 1410 0 and less than 5 H, the balance being Zr. All specimens were used in the as-received condition, annealed at 760°C for one hour and air cooled. The microstructure contained equiaxed grains roughly 10 p in diameter. The room temperature yield strength, ultimate tensile strength and elongation were 435 MPa, 467 MPa and 23%, respectively. Sheet specimens with dimensions 75 mm x 25 mm x 1.6 mm were cut transverse to the rolling direction. A notch was made by a saw cut (about 7 mm deep, notch radius about 0.125 mm) in the middle of one edge of each sheet specimen. Before testing, the specimens were polished with abrasive paper to 400 grit and then ultrasonically cleaned in water. The specimen was then sealed in an autoclave. The autoclave system was evacuated down to 6.67 Pa using a mechanical pump, then flushed with high purity argon gas (99.9995%) to ambient pressure, and evacuated again. The autoclave was then at least twice flushed with hydrogen to ambient pressure and evacuated. Finally, hydrogen was admitted up to experimental pressure.
The hydrogen pressures used in this study were 101, 1010 and 2020 kPa. High purity hydrogen with H2 > 99.9995% was used. Notch tensile tests were conducted on a screw-driven constant extension rate test (CERT) machine at a crosshead speed 3 x 10.’ m/s at 25,100 and 200°C. The time for a single test lasted from seven hours to twelve hours, depending on hydrogen pressure or experimental temperature. In general, as the hydrogen pressure or experimental temperature increased, the time for a single test became shorter. The desired temperature was controlled to -+l“C. The baseline properties were measured by testing specimens at the same crosshead speed and temperatures in an argon environment at a pressure of 101 kPa, which was later referred as “0 kPa Hz”. All the specimens were tested with the loading axis perpendicular to the rolling direction. For each condition, at least two specimens were tested and the average value was reported. After testing, the fracture surface was examined by a scanning electron microscope (SEM). The coverage percentage of brittle area on the fracture surface was measured by point counting method [ll]. A 616-point (22x28) grid was used to measure the fraction of brittle area on an SEM fractograph with a magnification of 500x. Each reported value was the average of at least 10 measurements of different positions on the fracture surface. Phase analysis was carried out by means of X-ray diffraction with CuKa X-radiation. The bulk hydrogen uptake of the specimen after testing was measured by inert-gas fusion method using a STROLEIN H-Mat 251 hydrogen determinator. The detailed hydrogen measurement process is described elsewhere .
Results Mechanical Tests
Three temperatures and three hydrogen pressures were selected to be the experimental conditions for notch tensile tests. The correlation between notch tensile strength (NTS) and hydrogen pressure is shown in Fig. 1. At a constant temperature, the values of NTS slightly increased as the hydrogen pressure increased from 0 to 101 kPa and then decreased with increasing hydrogen pressure. As expected, NTS decreased as temperature increased. Fig. 2 depicts the reduction of area (RA) vs. environmental hydrogen pressure. On the curve for 25”C, a large decrease of the value of reduction of area, which was defined as ductile-brittle transition, was found at the hydrogen pressure between 101 and 1010 kPa. As the temperature increased to 100°C and above, the ductile-brittle transition became less distinct, and the values of RA at the three experimental hydrogen
J.-H. Huang, C.-S. Ho / Maieriab Chemistryand Physics38 (1994) 138-145
Q ,“: _J
Fig. 1. The notch tensife strength of Zircaloy -4 vs. environmental hydrogen pressures at three temperatures.
Fig. 3. The loss of reduction of area (with respect to specimen tested in argon) of Zircaloy -4 vs. environmental hydrogen pressures at three temperatures.
value of reduction of area tested in argon.
Fig. 2. The reduction of area of Zircaloy -4 vs. environmental hydrogen pressures at three temperatures.
pressures were not quite different from those obtained in argon. If loss of RA (with respect to specimen tested in argon) is plotted against hydrogen pressures, as shown in Fig. 3, the ductile-brittle transition at room temperature is more apparent. The loss of RA is calculated by the equation given below. loss of RA (Pm) = [RA(Ar)-RA(P&]/
where RA(P& is the value of reduction of area at the experimental hydrogen pressure, and RA(Ar) is the
SEM fractographs from a series of specimens tested at progressivefy higher hydrogen pressures at 25°C are shown in Fig. 4. Figs. 4 (a) and 4 (b) depict the fracture surface morphologies for the specimens tested in argon and at 101 kPa hydrogen pressure, respectively. The fracture surface showed typical microvoid coalescence. No apparent difference was observed in the fracture surface for these two conditions. As the hydrogen pressure increased to 1010 kPa (Fig. 4 (c)), some isolated brittle regions surrounded by ductile tear ridges appeared on the fracture surface, and a selected area from Fig, 4 (c) is magnified in Fig. 4 (e). From Fig. 4 (e), it can be seen that inside the brittle region the fracture surface morphology was quite different from that of the surrounding ductile region. Many secondary cracks appeared on the relatively flat fracture surface and no microvoid coalescence can be found inside this region As the hydrogen pressure further increased to 2020 kPa, the coverage of the isolated brittle regions became larger (Fig. 4 (d)). Fig. 4 (f) shows the enlargement of one of the brittle areas formed at 2020 kPa HZ, in which the fracture mode is similar to that found in Fig. 4 (e). The observed brittle phase was identified as &hydride by X-ray diffraction. The X-ray diffraction pattern was shown in Fig. 5. In Fig. 5 (a), the (111) peak of &hydride was found to be very close to the (1010) peak of the zirconium matrix. A more distinct peak for
J.-H. Huang, C.-S. Ho i Materials Chem&y
and Physics 38 (1994)
ICI Fig. 4.
SEM fractographs of Zircaloy-4 tested at 25°C: (a) in argon ^. enlargement of one of the brittle areas III (c): and (t) enlargement
Hz; (c) at 1010 kPa
of one of the brittle
areas in (d)
Hz: (d) at 2020 kPa
J.-H. Huang, C.-S. Ho / Materials
Chemistry and Physics 38 (1994) 138-145
the (111) peak of &hydride can be found in Fig. 5 (b). As the temperature increased to lOO”C, the fracture surface morphologies for the specimens tested in argon and at 101 kPa H2 were similar to those at 25°C mostly microvoid coalescence without isolated brittle areas. As the hydrogen pressure increased to 1010 and above, the coverage of isolated brittle area on the fracture surface was apparently increased with increasing hydrogen pressure. Fig. 6 (a) shows the fracture surface morphologies for the specimens tested at 2020 kPa. Isolated brittle areas, which was relatively flat and contained many secondary cracks, were observed on the fracture surfaces. The enlargement of the brittle areas is shown in Fig. 6 (b). It can be seen that the fracture surfaces consisted of stepped and highly fissured blocks of hydride, which clearly indicated that these areas were much brittler than the surrounding area. The fracture surface morphologies for the specimens tested at 200°C were similar to those at lOO”C, except that the coverage of the isolated brittle areas was increased as the hydrogen pressure reached 1010 kPa and above. Fig. 7 (a) shows the fracture surface morphology
TABLE 1. Coverage Percentage Surface 0 kPa Hz* ZB(degree)
X-ray diffraction pattern of Zircaloy -4 after testing at 25°C at 2020 kPa Hz: (a) for 28 between 20” and 80”; and (b) for 20 between 25” and 45”.
25°C 1OQ”C 200°C
of Brittle Area on the Fracture
101 kPa H2 1010 kPa He 2020 kPa H2
5% 7% 9%
8% 14% 21%
* System was pumped down to 6.67 Pa and argon (99.9995%) was admitted into the system up to 101 kPa.
(b) Fig. 6. SEM fractographs of Zircaloy -4 tested at 100°C: (a) at 2020 kPa Hz; and (b) enlargement of one of the brittle areas in (a)
J.-H. Huang, C.-S. Ho I Materials Chemkty
Fig. 7. SEM fractographs of Zircaloy -4 tested at 200°C: (a) at 2020 kPa Hz; and (b) enlargement of one of the brittle areas in (a).
for the specimen tested at 2020 kPa. Fig. 7 (b) shows the enlargement of one of the brittle areas in Fig. 7 (a). As shown in Fig. 7 (b), the stepped and fissured feature was remained on the brittle region, which indicated that the ductility of the brittle area did not increase with temperature. Table 1 summarizes the coverage percentage of brittle area on the fracture surface. The brittle areas increased their coverage on the fracture surface as temperature or hydrogen gas pressure increased.
The following three features will be discussed. (1) The NTS was first slightly increased with hydrogen
and Physics 38 (1994) 138-145
pressure until 101 kPa and then decreased with increasing hydrogen pressure (Fig. 1). (2) There was a ductiIe-bottle transition at 25°C between 101 kPa Hz and 1010 kPa Hz, which became less distinct at 100°C and above (Figs. 2 and 3). (3) Isolated brittle areas were observed on the fracture surface as hydrogen pressure increased from 101 kPa to 1010 kPa and above (Fig. 4). The trend of NTS with respect to hydrogen pressure was similar to that reported by Lin et al.  on hydrided Zircaloy-4. They found that the UTS of hydrided Zircaloy-4 is first increased with hydrogen content until 400 ppm and then decreased with hydrogen content. Lin et at. attributed the strengthening effect below 400 pprn H to dispersion hardening by hydride precipitates. In the present study, for the specimens tested at 101 kPa HZ, there may be some small hydrides precipitated, which can not be resolved by SEM fractography. These small hydride precipitates may cause dispersion hardening and account for the slightly increase of NTS as hydrogen pressure from 0 to 101 kPa. As the hydrogen pressure further increased, from the fractographic findings (Figs. 4 and 6) the coverage of brittle hydride increased with hydrogen pressure. This may be due to the hydrogen flux diffused into the specimen was increased with hydrogen pressure. The increasing coverage of brittle hydride led to the reduction of the area to sustain the load and thereby the NTS was decreased. Fig. 1 also showed that the curves for the three temperatures were nearly parallel, which indicated that the effect of environmental hydrogen gas on NTS was almost the same at the three experimental temperatures and was not enhanced by the increase of temperature. The room temperature ductile-brittle transition of the reduction of area on hydrided Zircaloy-4 has been reported in a few previous studies [&lo]. Huang and Huang [lo] attributed this transition to the precipitation of brittle hydrides which leads to the decrease of fracture toughness. In this study, a ductile-brittle transition on the reduction of area at 25°C was found and was accompanied with the appearance of the isolated brittle regions on the fracture surface. The brittle regions, which mostly consist of hydride, were found to be parallel to the fracture surface. This was due to the stress effect which caused the hydride preferably precipitated on the plane perpendicular to the tensile axis . We believed that the brittle areas were caused by the crack of hydride precipitates during tensile testing. From Figs. 4 (b) and 4 (c), the fracture surface at 25°C exhibited ductile behavior, mostly microvoid coalescence, but the reduction of area was decreased in a large extent between 101 kPa H2 and 1010 kPa Hz. The presence of dimple like fracture surface suggested that the matrix had plastic flowed before fracture. Comparison of the fracture surface of Fig. 4 (c) with that of
J.-H. Huang, C.-S. Ho I Materials Chemistry and Physics 38 (1994) 138-145
Fig. 4 (b) showed that the fracture mode was similar and suggested that the matrix controlled the failure in both cases. The difference between these two conditions was the appearance of the isolated brittle region for the specimen tested at 1010 kPa HZ. In this case, failure occurred by ductile parting of the matrix between the isolated brittle region. Therefore, the change of reduction of area can be attributed to a reduction of the load bearing area of the matrix. Compared with the previous study, the coverage percentage of hydride on the fracture was different. Nelson and Wachob  found that after the slow crack growth test in environmental hydrogen, the entire fracture surface of the Zircaloy-4 specimen was covered with a lom-thick layer, which was identified as ZrH by X-ray diffraction. In this study, only some isolated hydrided regions, instead of the entire fracture surface, were formed during the tensile testing. The possible reason of this discrepancy is the contamination of the specimen surface by the impurity gases. Despite the high purity hydrogen (>99.9995%) was used, the initial evacuation by mechanical pump down to about 7 Pa may be not sufficient to eliminate the residual oxygen or other impurity gases. These impurity gases, especially the residual oxygen, may contaminate the fresh fracture surface during crack propagation, and thereby impede the occlusion of hydrogen. As a result, only some isolated areas can be hydrided by the environmental hydrogen gas. Hydrogen content measurement was performed on the specimens tested in hydrogen gas. The bulk hydrogen picked up by the specimens during tensile testing was very small (less than 30 ppm even after testing at 200°C and 2020 kPa Hz). This may also be attributed to the reference vacuum of the environment is poor, and the surface contamination of the specimen inhibits hydrogen uptaking. The ductile-brittle transition on RA became less distinct at 100°C and above. This may be due to the increasing ductility from hydrides or zirconium matrix. From Table 1, the coverage percentage of brittle hydride was increased with temperature. This can be related to the hydrogen flux diffused into the specimen. The hydrogen flux is proportional to the diffusivity of hydrogen in Zircaloy-4 which is increased with temperature; as a result, the amount of hydride formed on the fracture surface is increased with temperature. From SEM fractography, as shown in Figs. 6 and7, the hydride cracking areas exhibited brittle behavior, even at 200°C. Simpson and Cann  reported that the fracture toughness of zirconium hydride does not increase in a large extent from 25°C to 300°C. They found that even at 300°C the fracture toughness of zirconium hydride is extremely low, only 3 MPadm. From the SEM results on hydrided area at 200°C (Fig. 7 (b)), the brittle regions
remained stepped and friable morphology. This agreed with the result of very low fracture toughness of zirconium hydride found by Simpson and Cann. The coverage of brittle hydride increased with increasing temperature, while the RA values showed only slightly different from those tested in argon; apparently, the reduction of area was not controlled by the coverage percentage of hydride precipitation. Although the load bearing area of the matrix was decreased with increasing coverage of brittle areas, the increasing ductility of matrix with temperature may compensate the loss of reduction of load bearing area. Therefore, the improved ductility with increasing temperature, which led to the disappearance of ductile-brittle transition, was mainly from the improved ductility of zirconium matrix with increasing temperature.
Conclusions For all test temperatures, the notch tensile strength of Zircaloy-4 was slightly increased with hydrogen pressure until 101 kPa and then decreased with increasing hydrogen pressure. A ductile-brittle transition of reduction of area was found to occur at 25°C when Zircaloy-4 was tested at hydrogen pressures between 101 and 1010 kPa. The ductile-brittle transition was closely related to the precipitation of brittle hydrides. The ductile-brittle transition became less distinct as temperature reached 100°C and above. This was attributed to the improved ductility of zirconium matrix.
Acknowledgments This research was funded by the National Science Council of the Republic of China under Contract NSC 80-0413-E007-24.
References C.E. Coleman and D. Hardie, J. Less-Common Met., 2 (1966) 168. C.E. Ells, J. NucLMater., 28 (1968) 129. D.O. Northwood and U. Losasih, Int.Met.Rev., 28 (1983) 92. H.B. Nelson and H.F. Wachob, Stress Corrosion Cracking of Zircaloys, Electric Power Research Institute. Report EPRI-NP717, 1978, Section 6. 5. C.E. Coleman and B. Cox, in D.G. Franklin and R.B. Adamson (eds.), Zirconium in the NuclearIndustry: Sixth InrernationalSymposium, ASTM-STP-824, (1984) 675. 6. S.-C. Lin, M. Hamasaki and Y.-D. Chuang, Nucl. Sci. Eng., 72 (1979) 251. 7. J.B. Bai, C. Prioul, J. Pelchat and F. Barcelo, Proc. of Inc. Conf ANYENS, Avignon, France, April 21-24, 1991, p. 223. 1. 2. 3. 4.
J.-H. Huang, C.-S. Ho / Materials Chemby
8. J.B. Bai, C. Prioul, S. Lansiart Mater., 2.5 (1991) 2559.
and D. Francois.
Scri. Melall. ef.
9. J.B. Bai and D. Francois, J. Nucl. Muter., 187 (1992) 186. 10. J.-H Huang and S-P. Huang, J. Nucl.Mater., 208 (1994) 166. 11.G.F. Van der Voort, Metallography, Principles and Practice, McGraw-Hill,
NY. (1984) 426.
and Physics 38 (1994) 138-145
12. J.-H. Huang and S.-P. Huang, A161 (1993) 241.
145 Materials Science und Engineering,
13. M.R. Louthan, Jr. and R.P. Marshall. J. Nucl. Mater., 9 (1963) 170. 14. L.A. Simpson and CD. Cann. J. Nucl. Mater., 87 (1979) 3 03