Introduction

The β→α allotropic transformation in pure zirconium and its dilute alloys, the possibility of the displacive β→ω transformation in alloys containing β-stabilizing elements to a suitable level, and the clustering tendency in the β phase leading to a phase separation are responsible for the variety of transformations in the Zr-Nb system. The various possible sequences of phase transformations in alloys of different compositions are (1) martensitic transformation from the β (bcc) to the α (hcp) phase in alloys containing 0 to 7 wt pct Nb; (2) the β to ω (hexagonal) transformation in alloys containing 7 to 17 wt pct Nb during β quenching (athermal ω) or isothermal holding (ω precipitation); (3) the precipitation of α from the supersaturated β phase; (4) phase separation, β→β I (Zr rich) + β II (Nb rich), which is initiated by a spinodal decomposition in a suitable temperature and composition regime; and (5) a monotectoid reaction, β I → α + β II (at 883 K and 20 wt pct Nb).[1] Menon et al.[2] have shown that due to a large miscibility gap in the β phase, the rate of monotectoid decomposition of the β I phase to β II phase depends strongly on the heat-treatment temperature. In general, the β I phase is retained to a large fraction in the samples annealed at temperatures higher than the monotectoid temperature (T m ). Annealing at temperatures below T m , on the other hand, results in β precipitates with composition varying between β I and β II. However, at T«T m , a long duration heat treatment results only in formation of the equilibrium β II precipitates.

Recently, Zr-1Nb and Zr-1Nb-1Sn-0.1Fe alloys have drawn considerable interest because of their application in nuclear reactors as cladding material.[3] The microstructure of the Zr-1Nb alloy is different from those observed in the other commonly used zirconium-based alloys. Like Zr-2.5Nb, the Zr-1Nb alloy also has a two-phase microstructure consisting of the α and the β phases. However, in these alloys, the shape, size, and size distributions of the β phase are significantly different. In Zr-1Nb alloy, the β phase is distributed as fine spherical particles within the α matrix. Zircaloys also have a microstructure similar to the Zr-1Nb alloy, but they essentially have a single-phase microstructure in which intermetallic precipitates are distributed within the α matrix. In Zr-1Nb alloy, therefore, the size, size distribution, and composition of the β phase are expected to vary with aging time and temperature.

In quaternary Zr-Nb-Sn-Fe systems, where Sn and Fe are present in small quantities, the phase transformation behavior is not expected to vary drastically from the dilute binary Zr-Nb alloys. However, the presence of Sn and Fe is expected to influence the precipitation behavior because Sn and Fe are known to form various stable and metastable precipitates in the α-Zr.[4–9] In the Zr-Fe system, the intermetallic phases on the Zr-rich side are Zr3Fe (orthorhombic) and Zr2Fe (tetragonal). The Zr3Fe phase has been known to form as a metastable phase also.[8,9] In the Zr-Sn system, the intermetallic compounds of interest are the ones having compositions Zr4Sn and Zr5Sn3. These two phases have tetragonal and hexagonal structures, respectively. In addition, because Fe and Nb are β stabilizers and Sn is an α stabilizer, the simultaneous presence of these elements is also expected to change the precipitation kinetics. Recent studies have indeed demonstrated the presence of various precipitates in the quaternary Zr-Nb-Sn-Fe alloys. Efforts have also been made to identify the precipitates and the phase transitions causing the formation of these precipitates.[10,11,12] For example, Nikulina et al.[13] have shown the presence of Zr4Sn and Zr5Sn3 phases in the Zr-1Nb-1Sn-0.4Fe alloy.

In order to obtain a product with specified properties on a reproducible basis, a detailed study on the evolution of microstructure is essential. Detailed microstructural studies, dealing with the mechanism of phase and precipitate evolution and recrystallization, in binary Zr-1Nb and quaternary Zr-1Nb-1Sn-0.1Fe alloys, have not been reported too often in the international literature. On the contrary, these studies are essential for the better understanding of the microstructures in these systems and also for the better use of these alloys as cladding material with optimum in-reactor behavior. The β quenching heat treatment allows obtaining a material devoid of any pre-existing precipitates depending on the alloy chemistry and quench rate and thus can be used as a starting material to study precipitate evolution on tempering. Recrystallization studies are equally important as cold work followed by annealing constitutes a very vital step in the fabrication schedule of cladding tubes. The phase distribution generated by the preceding process controls the corrosion and mechanical behavior of the cladding tubes,[14–20] and thus, it is necessary to optimize the annealing parameters (temperature/ time) to generate the desired phase distribution.

In the present study, the aim was therefore

  1. (1)

    to study the β quenched and tempered microstructures in binary Zr-1Nb and quaternary Zr-1Nb-1Sn-0.1Fe alloys, and

  2. (2)

    to study the recrystallization behavior of these alloys and identify the various precipitates.

Experimental Procedures

The material selected for experimentation was hot-extruded rod for both alloys. The experimentation schedule was divided into two parts. First, the hot-extruded specimens were subjected to β quenching heat treatment by soaking the samples in the single β phase field at a temperature of 1273 K for 0.5 hours followed by quenching in water. The β-quenched samples were then tempered with different time-temperature combinations. In the second part of the study, the hot-extruded specimens were subjected to cold deformation by rolling. The deformation was carried out to the extent at which edge cracking of the samples had started. In the case of Zr-1Nb alloy, 90 pct deformation was given, and in the case of Zr-1Nb-1Sn-0.1Fe alloy, 83 pct deformation was given. These deformed samples were then annealed at temperatures of 853 K (below monotectoid temperature) and 903 K (above monotectoid temperature) for durations of 1 and 4 hours.

Specimens for transmission electron microscopy (TEM) were prepared by twin jet electropolishing in a Fischione twin jet electropolishing unit using an electrolyte consisting of a solution of 80 pct methanol and 20 pct perchloric acid at a temperature of 223 K and using a voltage of 25 V. The TEM studies were carried out using a JEOLFootnote 1 2000 FX microscope. The high-resolution electron microscopy (HREM) studies were carried out using a JEOL 3010 microscope. Energy dispersive spectroscopic (EDS) analysis was carried out to estimate the composition of the precipitates.

Mechanical properties were determined as a function of annealing temperature and time. Tensile tests were performed using flat tensile specimens in an Instron tensile testing machine at a strain rate of 0.04/min.

Results

β-Quenched Microstructure

Zr-1Nb alloy

Figure1(a) shows a representative as-quenched microstructure. The presence of lath microstructure, a product of martensitic phase transformation, α′ (hcp), can be clearly noticed. The absence of the β phase in the as-quenched microstructure even after examining several samples confirmed complete transformation of the β into α′ upon quenching. This, in turn, suggested the martensite finish temperature (M f ) to be above room temperature. Crystallographic analysis and contrast experiments on these laths revealed that all the laths in a region belonged to a single-martensite variant. In most regions, the arrangement of laths was nearly parallel within any given colony and many such colonies were found to be present within each prior β grain.

Fig. 1
figure 1figure 1

(a) Bright-field TEM micrograph showing the lath martensitic microstructure in β-quenched Zr-1Nb alloy. (b) Dark-field TEM micrograph showing mutually twin-related two variant grouping of a nearly parallel stack of laths obtained by ( \( 10\overline{1} \overline{1} \))M reflection. (c) A typical twin-related SAED pattern obtained from the region 1–2 lath-lath interface showing { \( 1\overline{1} 01 \)} α type Twinning. (d) Bright-field TEM micrograph showing the solid block morphology of the three variant lath martensite grouping in Zr-1Nb alloy quenched from the β-phase field. (e) Combined diffraction and dark-field analyses of the 1–2–3 lath grouping, where the dark-field images of the three variants and the twin-related SAED patterns obtained from the 1–2, 2–3, and 1–3 lath-lath interfaces are shown. The SAED patterns resembled { \( 1\overline{1} 01 \)} α type twinning. (f) The single surface analysis for determining the traces of habit planes of lath martensite in Zr- 1Nb alloy.

Full size image

In some regions, the lath colonies consisted of two or three lath martensite variants, where these variants were found to be mutually twin related. An example of such mutually twin-related two-variant lath grouping is shown in Figure 1(b), where the laths marked as 1 and 2 belonged to the two different twin-related martensite variants. The selected area electron diffraction (SAED) pattern (Figure 1(c)) obtained from interfaces 1–2 showed twinning on an { \( 1\overline{1} 01 \)} α plane.

Figure 1(d) shows a bright-field TEM micrograph in which the multivariant lath martensite group having “solid block” morphology could be noticed. In this morphology, three different variants of the martensite were identified. Combined diffraction and contrast analyses of these laths in the regions 1-2-3 are shown in Figure 1(e). As may be noticed, these variants of martensites are mutually twin related to each other. Due to the absence of the parent β phase, single surface trace analysis was carried out to determine the habit plane of each martensite variant in single, two, and three lath groupings. The habit planes were determined assuming the interfaces between the martensite variants to be the traces of the habit planes and keeping in mind the Burgers’ orientation relationship for the bcc ↔ hcp transformation.[21] The habit plane traces for all three variants were observed to lie close to {334} β type poles (Figure 1(f)).

Zr-1Nb-1Sn-0.1Fe alloy

Figure 2 shows a typical as-quenched microstructure of the quaternary alloy comprising lath martensite. Arrangement of these laths was similar to the arrangement of laths in the binary alloy. However, upon comparing the microstructures of the two alloys, few differences have been noticed, which are summarized as follows.

  1. (1)

    Interfaces of the laths observed in Zr-1Nb-1Sn-0.1Fe were much straighter than the laths observed in the binary alloy.

  2. (2)

    Number densities of the laths in a given packet were much higher in the quaternary alloy as compared to the binary Zr-1Nb alloy.

  3. (3)

    Laths in Zr-1Nb-1Sn-0.1Fe could be observed in two to three distinct size groups, suggesting that the formation of the lath occurred over a range of temperatures. This, in turn, suggested that the difference between M s and M f is higher in this alloy as compared to the binary alloy.

  4. (4)

    In the Zr-1Nb-1Sn-0.1Fe alloy, twin-related two to three variant combinations of martensitic laths were not observed.

Fig. 2
figure 2

Bright-field TEM micrograph showing the lath martensite morphology obtained in the quenched Zr-1Nb-1Sn-0.1Fe alloy.

Full size image

Tempered Microstructure

The β-quenched samples of binary and quaternary alloys were subjected to tempering treatment in the temperature range of 773 to 923 K for a time period of 0.5 to 24 hours. On tempering, substantial modification in the microstructure was noticed in terms of the lath dimensions, morphology, and precipitation.

Zr-1Nb alloy

Extensive TEM investigation of the microstructure of samples, which were tempered at 773 K for durations less than 24 hours, failed to show the presence of any precipitate, though changes in the dimension of the laths were noticed (increase in lath width and decrease in length). Tempering beyond 24 hours at this temperature, however, showed precipitation. For example, in samples tempered for 24 hours, the precipitates were very small (∼10 nm) and were heterogeneously distributed at the lath boundaries. Figure 3(a) shows a TEM micrograph from such a sample, where the thicker lath boundary is essentially due to the preferential precipitation at the lath boundaries. Electron diffraction analysis confirmed that the precipitates were of the β phase.

Fig. 3
figure 3

The TEM micrographs showing the tempered microstructures of the Zr-1Nb alloy: (a) 773 K/24 h, (b) through (d) 823 K/24 h, (e) 923 K/0.5 h, (f) 923 K/2 h, and (g) 923 K/5 h.

Full size image

Samples, tempered at 823 K, showed precipitation even when the duration of tempering was as low as 10 hours. In this case, precipitates were observed at the boundaries as well as within the grains. With an increase in time, the precipitate size increased. Arrangement of the precipitates, however, remained similar. Figures 3(b) through (d) are micrographs from samples that were tempered at 823 K for 24 hours. There are a few points to be noticed from these micrographs. First, formation of precipitates was in narrow bands (typical thickness is ∼50 nm) and these bands appeared to align along some specific crystallographic direction (as could be noticed from the bright-field–dark-field combination in Figures 3(b) and (c), respectively). In some cases, formation of such a band is not noticed, but the alignment of the precipitates was still maintained (Figure 3(d)). Second, when a dark-field image (Figure 3(c)) was taken from the \( (\overline{1} 01) \) reflection ([111] zone) pertaining to the β precipitate, nearly all of the precipitates came in contrast together, suggesting that these precipitates maintain an orientation relationship with the matrix α phase. The orientation relationship turned out to be the Burgers’ orientation relationship for the bcc ↔ hcp transformation.[21]

The samples heat treated at 873 K or above showed substantial precipitation even at very short durations of tempering. Increasing the tempering time at these temperatures led to (1) substantial changes in the dimensions of the lath structures leading to equiaxed recrystallized grains of the α phase, and (2) coarsening of the precipitates as well as reduction in the number density of the precipitates. Bright-field TEM micrographs of the samples tempered at 923 K for 0.5, 2, and 5 hours are shown in Figures 3(e) through (g), respectively, to highlight the aforementioned behavior. From Figure 3(e), it could be clearly seen that upon tempering at 923 K for 0.5 hour, migration of the lath-lath interfaces has started. As may be noticed, the curved interface of a recrystallized grain is sweeping into the neighboring laths. Decoration of discontinuous β precipitates at the lath boundaries could be observed from the micrograph. It may also be noted that precipitates are nearly absent within the pre-existing laths.

Tempering at 923 K for 2 hours (Figure 3(f)) resulted in almost complete dissolution of the lath boundaries. The signature of initiation of lath coarsening was observed. A regular array of elongated β precipitates along the prior lath-lath interface could be seen with very few precipitates within the laths as well. Unlike tempering at lower temperatures, here, precipitates were not found to align along any crystallographic direction. Tempering for 5 hours (Figure 3(g)) led to the substantial coarsening of the lath structure and the precipitates. The volume fraction of precipitates within the laths was also found to be higher than at the lath boundaries. The TEM-EDS composition analyses of the β precipitates could not be performed in the samples tempered at 773 and 823 K, mainly due to the very small size of the precipitates. The relatively larger size (∼250 nm) of the β precipitates in the sample tempered at 923 K/5 h facilitated the estimation of the average composition, which was found to be Zr-8 wt pct Nb.

Zr-1Nb-1Sn-0.1Fe alloy

In contrast to the binary Zr-1Nb alloy, the quaternary alloy is expected to show a different precipitation behavior due to the presence of Sn and Fe. Unlike the binary alloy, precipitation of the second phase in the quaternary alloy was observed at the lath boundaries even in samples tempered at 823 K for 0.5 hours. Tempering for longer durations led to changes in the morphology of the lath structure as well as an increase in volume fraction of the precipitates. Diffraction and EDS analyses in TEM ascertained that the majority of these precipitates were of the β phase. Representative microstructures of samples tempered at 823 K for 5 hours and 24 hours are shown in Figures 4(a) and (b), respectively. At 883 K, a short duration tempering (3 hours) resulted in the disappearance of the lath-lath interfaces. Regularly spaced arrays of precipitates lying along prior lath-lath boundaries were noticed (Figure 4(c)). Some precipitates appeared to coarsen as the elongated lenticular morphology of these precipitates was observed. In this case also, almost all the precipitates were β with average Nb content of about 5 wt pct. A longer duration treatment (such as 6 hours) at 883 K further coarsened the precipitates (Figure 4(d)). However, in this case, besides the normal β type precipitates, a very small fraction of the precipitates were found to contain Fe. The average composition of β precipitates and Fe bearing precipitates were Zr-6 wt pct Nb and Zr-23 wt pct Nb-7 wt pct Fe, respectively. It could be noticed that the Fe bearing precipitates have higher Nb content and the stoichiometry (Zr/(Nb + Fe) ratio of 2) indicated them to be close to Zr2(Nb,Fe) phase. In order to confirm the identity of the Fe bearing precipitates, SAED analyses have been carried out. An SAED pattern is shown in Figure 4(e), where the reflections marked by arrows correspond to the Zr-Nb-Fe precipitate. The pattern formed by the precipitate reflections could be indexed close to the Zr2Fe phase. The Zr2Fe phase has a body-centered-tetragonal structure with lattice parameters of a = 0.636 nm and c = 0.582 nm.[5] The small angular deviations noticed during the indexing of the pattern with respect to the Zr2Fe phase could be attributed to the presence of Nb in the precipitates, which is expected to modify the lattice parameters. The convergent beam electron diffraction technique could not be employed to determine the exact lattice parameters and to reconfirm the crystal structure due to the small size of the precipitates. The tempered microstructure at 923 K, as shown in the Figure 4(f), exhibited a fully recrystallized microstructure, where precipitates were uniformly distributed within the recrystallized grains. The morphology of the precipitates varied from spherical to cuboidal to lenticular shape. Here again, the precipitates were found to be β type only, with average Nb content of 5 wt pct.

Fig. 4
figure 4

(a), (b), (c), (d), and (f) Bright-field TEM micrographs showing the microstructures of the Zr-1Nb-1Sn-0.1Fe alloy tempered at 823 K/5 h, 823 K/24 h, 883 K/3 h, 883 K/6 h, and 923 K/6 h, respectively. In the SAED pattern, shown in (e), the Zr-Nb-Fe precipitate reflections marked by the arrow could be indexed in terms of the Zr2Fe phase with small angular deviations (zone axis = [ \( \overline{1} 13 \)]).

Full size image

Recrystallization Studies

Microstructure of Zr-1Nb alloy

A higher magnification bright-field TEM micrograph of the starting material in extruded condition (transverse section) is shown in Figure 5(a). Fine spherical β precipitates in the size range of 40 to 60 nm were present uniformly within the equiaxed α grains.

Fig. 5
figure 5

Bright-field TEM micrographs of binary alloy (a) as-extruded, (b) cold worked, (c) annealed (853 K/1 h), (d) annealed (853 K/4 h), (e) annealed (903 K/1 h), and (f) annealed (903 K/4 h).

Full size image

A typical cold-worked microstructure representing the deformed starting material is shown in Figure 5(b). High dislocation density arising out of the heavy deformation led to the formation of dislocation tangles, as could be seen from the figure.

Annealing of the cold-rolled sample at 853 K for 1 hour led to the complete recrystallization of the cold work microstructure (Figure 5(c)). Fine β precipitates were found distributed nonuniformly within α grains in the form of bands. The TEM-EDS analysis of the relatively large size β precipitates showed a composition close to Zr-69 wt pct Nb. The sample annealed at 853 K for 4 hours exhibited a microstructure very close to that of 1 hour. However, coarsening of the β precipitates was noticed. In addition, evidence of fresh precipitation was also noticed (Figure 5(d)).

Annealing at 903 K for 1 hour led to grain growth in some of the regions (Figure 5(e)). The volume fraction of the β precipitates had increased substantially. Bimodal size distribution of the precipitates could be seen, where big precipitates were found to coexist with very fine precipitates, indicating the simultaneous occurrence of two phenomena: nucleation of fresh precipitates and coarsening of existing precipitates. Again, the precipitates were distributed in a nonuniform manner in the form of bands within the grains. Coarsening of the α grains and β precipitates was noticed in the sample annealed for 4 hours at 903 K. The β precipitates had coarsened at the expense of the smaller precipitates, leading to a decrease in their number density (Figure 5(f)). The TEM-EDS analysis indicated β composition of Zr-45 wt pct Nb.

In order to examine the nature of the interface between the α matrix and the β precipitate and also the internal structure of the β precipitate, HREM examination of representative specimens of the recrystallized binary alloy was carried out. The HREM micrograph in Figure 6 shows a spherical β precipitate present in the α matrix. A careful examination of the α/β interface showed, in general, good lattice matching between the two phases. However, to accommodate the change in curvature of the precipitate, lattice mismatch in some patches was also noticed. This kind of α/β interface normally obeys Burgers’ orientation relationship,[21] which provides a low energy configuration. The good lattice matching also indicated the coherency of the interface between the α matrix and the β precipitate.

Fig. 6
figure 6

HREM micrograph showing the internal structure of the β precipitate and the nature of the α/β interface in recrystallized Zr-1Nb alloy.

Full size image

Microstructure of Zr-1Nb-1Sn-0.1Fe alloy

Microstructure of the transverse section of the hot-extruded quaternary alloy used as the starting material for this study is shown in Figure 7(a). The nearly equiaxed distribution of the α grains represented a typical dynamically recrystallized microstructure. Cold working of this starting material generated a microstructure very similar to that of the cold-worked microstructure of the binary alloy.

Fig. 7
figure 7

Bright-field TEM micrographs of quaternary alloy (a) as-extruded, (b) annealed (853 K/1 h), (c) annealed (853 K/4 h), (d) annealed (903 K/1 h), and (e) annealed (903 K/4 h).

Full size image

In this case as well, annealing of the cold-worked samples at 853 K for 1 hour led to complete recrystallization of the cold work microstructure (Figure 7(b)). The precipitates were found to be associated with strain contrast, suggesting partial recovery of the precipitates, and were nonuniformly distributed within the grain as well as on the grain boundary. Annealing at 853 K for 4 hours exhibited a fully recrystallized microstructure. Evidence of grain coarsening could also be noticed. The absence of strain field around the precipitates suggested complete recovery of the precipitates. The volume fraction of the β precipitates inside the grains was found to increase with the annealing time. Precipitates at the grain boundary and within the grain seem to have coarsened with annealing time (Figure 7(c)). The TEM-EDS analyses of the sample annealed at 853 K for 1 hour have shown that the intragranular precipitates were of two types: β type with average Nb content of 9 wt pct and Zr-Nb-Fe type with the average composition of Zr-12 wt pct Nb-3.5 wt pct Fe. In contrast, the grain boundary precipitates were only Zr-Nb-Fe type with average composition of Zr-13.5 wt pct Nb-4 wt pct Fe. In the case of the sample annealed at 853 K for 4 hours, both the intragranular and intergranular precipitates were Zr-Nb-Fe type with average compositions of Zr-22 wt pct Nb-5 wt pct Fe and Zr-28 wt pct Nb-6 wt pct Fe, respectively. However, TEM-EDS analysis has failed to show the presence of any β precipitate in this sample. In the case of 1-hour annealing, the Zr-Nb-Fe precipitates (both intra and intergranular) could be inferred as a Zr4(Nb,Fe) phase, whereas, in case of 4-hour annealing, the precipitates (again intra and intergranular) turned out to be a Zr2(Nb,Fe) phase. The preceding inference was drawn based on the compositions of the precipitates only.

Annealing at 903 K for 1 hour showed a microstructure very similar to that of the sample annealed at 853 K for 4 hours (Figure 7(d)). However, in the case of the former, a substantial distribution of very fine precipitates was noticed within the grains. Annealing for 4 hours at 903 K resulted in a grain-coarsened microstructure. The β precipitate size has increased and their number density has decreased, keeping the volume fraction nearly the same (Figure 7(e)).

The HREM studies have shown a similar type of interface between the α matrix and β precipitates in the quaternary alloy, as was observed in the case of binary alloy. However, the nature of the interface between the Zr-Nb-Fe type precipitates and the α matrix was different from the α/β interface. This interface was observed to be completely incoherent, as can be seen from Figure 8. Defects such as faulting could also be noticed inside the precipitate, which was not seen in the case of β precipitates. These faults were aligned in many directions and were of extremely fine dimensions. The presence of faults indicated the intermetallic nature of these precipitates. The nature of these faults could not be ascertained conclusively due to their fine dimensions.

Fig. 8
figure 8

HREM micrograph showing the internal structure of the Zr-Nb-Fe–type precipitate in recrystallized Zr-1Nb-1Sn-0.1Fe alloy.

Full size image

Mechanical property evaluation

In order to understand the effect of annealing on the properties of these two alloys, a systematic evaluation of the mechanical properties as a function of annealing temperature and time were carried out. An ambient temperature tensile test was performed on the as-received, cold-worked, and annealed samples. From the engineering stress–engineering strain plots, yield strength, ultimate tensile strength, uniform elongation, and total elongation have been calculated. The data have been presented in tabular form in Table I. Yield strength and elongation values showed that the quaternary alloy is harder in comparison to the binary alloy with comparable ductility. Cold working increased the yield strength with the drop in ductility in both the alloys. However, in the case of the cold-worked quaternary alloy, the ultimate tensile strength increased significantly (732 MPa) with a considerable drop in the uniform elongation (3.8 pct). In both alloys, annealing of cold-worked material for 1 hour at 853 K resulted in (1) lowering of the yield strength close to the level of the as-received material and (2) improvement of ductility. These findings suggested complete recrystallization of the cold-worked structure, which is well supported by microstructural observations, as reported in the earlier sections. In the case of the quaternary alloy, the yield strength of the sample annealed at 853 K for 1 hour was found to be higher in comparison to the other annealed samples. However, in the case of the binary alloy, the sample, which was annealed at 903 K for 1 hour, showed little increase in yield strength.

Table I. Ambient Temperature Tensile Test Properties of Binary and Quaternary Alloys
Full size table

Discussion

Martensite Formation

In several Zr-based alloy systems, the morphology and substructure of martensites have been investigated in detail. The phenomenological theory of martensite crystallography (PTMC) has been used successfully to predict the substructure of the martensite crystals.[22–31] The morphology and substructure of the martensite have been found to be essentially governed by factors such as alloy composition and self-accommodation tendency.[23,25,32] In the present study, the M s value has been roughly estimated for the Zr-1Nb alloy. In the estimation of the M s temperature, first, the T 0 temperature, at which the free energies of the α and the β phases are equal, has been calculated using the regular solution thermodynamics model. The G-HSER, the Gibbs energy relative to the enthalpy of the “standard element reference,” i.e., the reference phase for the element at 298.15 K, and the interaction parameter data have been used for the preceding calculation.[33] The T 0 temperature was found out to be 1084 K. Again, from Kaufman’s work,[34] it is known that the M s temperature for zirconium base alloys is usually about 50 K lower than the T 0 temperature. Thus, the M s temperature is expected to be 1034 K. The observation of predominantly lath martensite in the β-quenched samples is justified because the M s temperature associated with it is quite high (∼1034 K). This observation was found to be consistent with that of Srivastava et al.[35] The observation of the {334}C type habit plane of the lath martensite was consistent with the earlier observations made in Zr-Nb alloys and other alloys.[22–25]

The observation of packets in some regions consisting of two-variant and three-variant combinations of the mutually twin-related martensite laths suggested the possibility of self-accommodation. The self-accommodation effect has also been proposed in some other Zr-Nb alloys.[22,24,35] The shape strain associated with the martensite crystal nucleation can be minimized by simultaneous nucleation of a group of nonparallel martensite plates or when one plate sympathetically nucleates another, which has been suggested as the motivation for self-accommodation. In this process, the strain fields of individual martensite plates interact with each other in such a way that the total transformation strain energy becomes minimized. Because the M s temperature for this alloy is high, the shape strain could be easily accommodated in the matrix. However, a small volume fraction of the parent phase, which was transformed in the last stages at lower temperatures, was responsible for exhibiting such two- and three-martensite variant groups. This is because, at lower temperature, the matrix became stronger and resistive forces increased, promoting the formation of self-accommodating martensite lath groups.

In this study, it was observed that the martensite morphology and substructure did not show any remarkable difference with the addition of Sn and Fe in the Zr-1Nb alloy. Oh et al.[36] studied the effect of Fe, V, Sn, and Mn addition on the characteristics of the martensitic transformation in Zr-0.8Sn alloy. In this study, they have reported lath martensites both in binary Zr-0.8Sn and ternary Zr-0.8Sn-0.4Fe alloys, which indicated that even with the addition of Fe, which is a β stabilizer, the M s temperature of Zr-0.8Sn alloy has not decreased substantially. In the present study, the mutually opposite effect due to the addition of β stabilizer (0.1 wt pct Fe) and α stabilizer (1 wt pct Sn) in Zr-1Nb alloy probably did not change the M s temperature substantially, and this was evident from the presence of lath martensites in the microstructure of β-quenched Zr-1Nb-1Sn-0.1Fe alloy. However, addition of Sn and Fe has provided substitutional solid solution strengthening to the α′ matrix, resulting in straighter lath interfaces and finer lath morphology.

Tempering of Martensite

According to the Zr-Nb binary phase diagram,[4] tempering of the β-quenched Zr-1Nb alloy below the monotectoid temperature (i.e., 883 K) would transform the metastable martensite α′ phase to the equilibrium α + β II phase. However, the nucleation of a β precipitate in the α′ matrix of composition Zr-1 wt pct Nb involves (1) segregation of Nb atoms and (2) transformation of crystal structure from hcp to bcc.[2] A typical G-X plot (where X is represented in an atomic fraction of Nb) calculated for the Zr-Nb alloy at 773 K is shown in Figure 9. It can be seen from the figure that the occurrence of step (1) before step (2) will involve a very large free energy barrier, as indicated in the figure by ΔG α (0.01, Xn), which is essentially due to the sharp rise in the free energy of the α phase with increasing Nb content. On the other hand, formation of β nuclei having a Nb concentration close to but higher than Xn = 14 wt pct and a subsequent Nb enrichment of these nuclei will involve a much smaller free energy barrier. The composition Xn for different tempering temperatures has been calculated from the respective G-X plots and presented in Table II. The value Xn = 14 wt pct Nb at 773 K explains the absence of any precipitation reaction when martensite was tempered even for a longer time (less than 24 hours). This is because nucleation of the β phase (being a diffusion-controlled phenomenon) even with 14 wt pct Nb would involve long-range diffusion of Nb atoms. The diffusivity of Nb in α-Zr at 773 K is 8.18 × 10−19 m2 s−1,[37] which shows that the typical diffusion length ( \( x = {\sqrt {Dt} } \)) of Nb atoms would be of the order of 0.3 μm in 24 hours. Such a low value of diffusion length would make the kinetics of the reaction extremely slow even in the presence of a strong thermodynamic drive. Tempering at slightly higher temperature, 823 K (where Xn = 12 wt pct Nb), would have slightly faster kinetics (2.84 × 10−18 m2 s−1[37]), which made the precipitation reaction possible when α′ was tempered for longer duration (more than 10 hours). The observation of precipitates to align along certain specific crystallographic directions in the matrix could be explained from the phenomenon of polygonization, which occurs during the recrystallization of the lath structure. Because the lath martensites inherently contain high dislocation densities, which would undergo a recovery process by polygonization and hence form a low energy configuration in the form of low-angle boundaries, further diffusion through these dislocations would be higher by several orders of magnitude, and therefore, they act as preferential sites for the nucleation of the β precipitates. The formation of precipitates in such a way would acquire the arrangement of the dislocations.

Fig. 9
figure 9

Free energy–composition diagrams for the α and β phases in the Zr-Nb system at 773 K. The inset figure depicts that the only common tangent that could be drawn between the α and β phases is between the α and β II phases, i.e., α′ → α + β II on tempering α′ at 773 K. The magnified G-X plot shows the alternative route for β nucleation with Nb content << β II and also involving a much smaller free energy barrier.

Full size image
Table II. Minimum Nb Content Required for β Precipitate Nucleation as a Function of Tempering Temperature
Full size table

The tempering of α′ phase above 873 K is expected to show precipitation reaction even for short duration tempering. This was confirmed from the observation of β precipitates in all the samples tempered at 873 to 923 K for 0.5 hours. Tempering near the monotectoid temperature (883 K) would require a much lesser extent of Nb enrichment to form β phase (as indicated in Table II). In addition to the preceding factor, at a higher temperature, kinetics also would be more favorable. Good agreement between the predicted composition (Zr-7 wt pct Nb) and experimentally observed composition (Zr-8 wt pct Nb) of the β phase could be seen in the sample tempered at 923 K. A similar tempering behavior has been reported by other workers in dilute Zr-Nb alloys with varying Nb concentration.[38–41]

The tempering behavior of the quaternary alloy was different from that of the binary alloy in the context of precipitate evolution and precipitation kinetics. Mainly, two distinct types of precipitates have been seen: Zr-Nb β type (in all the tempered samples) and Zr-Nb-Fe type (in some samples only). As could be seen from Section III, the experimentally observed composition of the β precipitates at different temperatures did not match with what has been predicted from the thermodynamic analyses of the Zr-Nb binary system. This might be because of the presence of Sn and Fe in the quaternary alloy. The reason for the nonobservation of Zr-Nb-Fe type precipitates in some of the samples may be due to the fact that their volume fraction and size were quite less, and a longer duration heat treatment would have been helpful in generating a substantial volume fraction and reasonable size of these precipitates suitable for identification. A similar situation has been encountered in the case of the sample tempered at 883 K where the identification of the Zr-Nb-Fe type precipitates was only possible after tempering for 6 hours. The high Nb content of the Fe bearing precipitates can be explained from the fact that both are β stabilizing elements and have a tendency to remain associated together. In general, the precipitation kinetics was found to be much faster in the case of the quaternary alloy.

Recrystallization Studies

In the case of the binary alloy, the resultant microstructures after annealing above or below the monotectoid temperature were different mainly with respect to the precipitate composition. Menon et al.[2] carried out a detailed thermodynamic analysis of the monotectoid reaction β I → α + βII. They have shown that there is a large miscibility gap, which makes the preceding transformation sluggish and, hence, requires a large driving force for β I → α + β II transformation. On heat treating the Zr-1Nb alloy at 903 K, the β phase becomes enriched with Nb to the extent of about 18 wt pct, and while cooling below the monotectoid temperature, the Nb enrichment of the β phase should exceed 85 wt pct per the equilibrium phase diagram.[4] However, TEM-EDS analyses have shown the β precipitates to be Zr rich with compositions ranging from 25.6 wt pct Nb to 45 wt pct Nb, thus explaining the sluggish nature of the transformation, as mentioned previously. Moreover, the observation of composition range in the precipitates suggested (1) the continuous nucleation of β phase with composition close to β I (20 wt pct Nb) during the cooling process and (2) the enrichment of the existing β precipitates with Nb. The preferential nucleation of metastable β I is due to the smaller activation energy requirement in comparison to the β II phase. The continuous nucleation process is also supported by the bimodal precipitate size distribution, where large and small precipitates were distributed inside the grains in a nonuniform manner. Precipitation of particles could be seen at the grain boundary also. Annealing at temperatures below T m (as in the present case, 853 K), unlike the previous case, did not show a large composition range of the β precipitate. Instead, these precipitates were much richer in Nb (average ∼69 wt pct Nb). This suggested that the driving force to nucleate β II phase was substantially larger at this temperature. Further enrichment of these precipitates was not expected due to slower kinetics at lower annealing temperature. Because of this reason, precipitates in this case were more or less the same size. With regard to the extent of recrystallization, annealing below the monotectoid temperature led to the complete recrystallization of the cold-worked microstructure.

In the case of the quaternary alloy, annealing up to 1 hour at 853 K led to complete recrystallization of the cold-worked microstructure. Both intragranular and intergranular precipitates were observed in all the annealed samples. Furthermore, comparing the results, it could be concluded that (1) the intergranular precipitates are richer in both Nb and Fe than the intragranular precipitates; (2) the higher the Nb content of the precipitate, the higher is the Fe content or vice versa; and (3) like tempering, here also annealing for a longer duration promoted the formation of the Zr-Nb-Fe type of precipitates, as was evident by the presence of only Zr-Nb-Fe type intragranular precipitates in the 853 K/4 h sample compared to both β and Zr-Nb-Fe type intragranular precipitates in the 853 K/1 h sample. Moreover, longer duration annealing resulted in the enrichment of the Zr-Nb-Fe type precipitates with Nb and Fe. The aforementioned phenomena could be explained taking the example of annealing at 853 K. The fact that the diffusivity of Nb atoms (5.59 × 10−18 m2 s−1 [37] at 853 K) in the α matrix is faster than that of the Fe atoms (1.25 × 10−18 m2 s−1 [37] at 853 K) resulted in the formation of mainly β precipitates during the early stages of annealing. An increase in the annealing time increased the diffusion length of both the Nb and Fe atoms, thus enabling the Fe atoms to reach a Zr-Nb atomic cluster and form Zr-Nb-Fe type precipitate. The increase in the diffusion length with time was also responsible for the formation of only Zr-Nb-Fe type precipitates and also their enrichment with Nb and Fe (at relatively longer tempering/annealing time). The high Nb and Fe content of the intergranular precipitates in comparison with the intragranular precipitates could be explained by the easy diffusion path provided by the grain boundaries to Nb and Fe atoms. In a recent TEM study carried out on ZIRLO,Footnote 2 it has been found that the β II phase and a Zr-Nb-Fe phase are the two main precipitating second phases. The Zr-Nb-Fe phase has a hexagonal structure with lattice parameters of a = 0.53 nm and c = 0.875 nm.[42] Other workers have reported a hexagonal Zr(Nb,Fe)2 phase with a Zr/(Nb + Fe) ratio of 0.5.[43] However, in the present study, the Zr-Nb-Fe precipitates obtained could be inferred as Zr4(Nb,Fe) phase (at shorter annealing time) and Zr2(Nb,Fe) phase (at longer annealing time). The absence of Zr-Sn precipitates in Zr-1Nb-1Sn-0.1Fe alloy could be due to the higher solubility of Sn in the α phase because Sn is a well-known α stabilizer and also due to the low amount of Sn in the alloy (1 wt pct). In Zircaloys also, for example, Zircaloy-2, even with a comparatively higher percentage of Sn (∼1.5 wt pct), Zr-Sn precipitates could not be observed.[44]

The mechanical behavior of the two alloys in response to the annealing treatment can be explained on the basis of their microstructural features, as observed in this study. The nature of precipitates is quite different in the two alloys, as seen earlier. The TEM-EDS study has shown that in the case of binary alloy, the precipitates are only soft solid solution β type, whereas, in the case of quaternary alloy, Zr4(Nb,Fe)/Zr2(Nb,Fe) type precipitates were present in addition to the β type precipitates. These Zr-Nb-Fe precipitates are expected to be harder because of their intermetallic nature (as confirmed by the presence of faults in the HREM image). These hard precipitates, being sufficiently strong, are not expected to be sheared by dislocations moving through the matrix. These fine nondeforming precipitates not only increase the yield strength but also affect the work-hardening behavior. Strong precipitates increase the initial rate of work hardening. The additional work hardening caused by the precipitates arises mainly for the following reasons: (1) in the early stages of straining, plastic flow in the matrix causes a build up of elastic strain in the precipitates, thereby generating internal stresses, which oppose continued flow;[45,46] and (2) rapid buildup of dislocation density in the presence of hard precipitates, thereby restricting slip processes in the matrix. However, this increased rate of hardening is not sustained at high strains mainly because of the activation of the slip processes in the surrounding matrix in preference to the loading of the precipitates,[47] and as a result, the work-hardening rate falls off after a low strain, thereby leading to a decrease in the uniform elongation and early onset of necking. Often, the reduction in uniform elongation is approximately proportional to the increase in strength in the dispersion-strengthened systems, and this behavior has been noticed in the cold-worked sample of the quaternary alloy, where the presence of high levels of dislocations introduced during the cold working operation in conjunction with the presence of hard precipitates has led to a significant increase in tensile strength with a substantial drop in uniform elongation. Beyond the onset of necking, the precipitates have hardly any role to play in the work-hardening behavior, and hence, the necking elongation has been found to be nearly the same with some small variations in almost all the samples. In both the alloys, recovery of the stresses introduced during cold working takes place in the matrix α phase as well as in the precipitates during annealing. The partial recovery of the elastic strain present in the precipitates was responsible for the observed high yield strength when the cold-worked quaternary alloy was annealed for 1 hour at 853 K. This was reflected in the strain contrast surrounding the precipitates. Though the β precipitates are deformable and are sheared by moving dislocations, a coherent nature, substantial volume fraction, and small size could lead to strengthening. The microstructural observations (as made by TEM and HREM studies) and higher yield strength of the binary alloy sample annealed at 903 K for 1 hour supported the preceding fact. Apart from the hard precipitates, the presence of Sn also led to some degree of solid solution strengthening of the matrix α phase in the quaternary alloy. Hence, the hardness of the matrix α phase and the fine dispersion of nondeformable hard precipitates were responsible for the observation of a generally higher strength of the quaternary alloy.

Summary

In this article, the evolution of microstructure and precipitation behavior of two important fuel cladding alloys, viz. binary Zr-1Nb and quaternary Zr-1Nb-1Sn-0.1Fe, have been examined after different thermomechanical treatments, which included β quenching followed by tempering and cold working followed by annealing. The precipitate composition, morphology, and volume fraction depended strongly on the composition of the alloy and tempering/annealing parameters (temperature and time). An attempt has also been made to correlate the annealed microstructures with the measured mechanical properties. The extent of recrystallization of the matrix α phase and the nature and volume fraction of the precipitate phase have been found to be important parameters in deciding the mechanical properties. Some of the important observations are summarized as follows.

  1. 1.

    The β-quenched microstructure was predominantly lath martensitic type with dislocated substucture. In Zr-1Nb alloy, apart from the single variant, two and three variant grouping of mutually twin-related laths were also seen, which suggested the possibility of self-accommodation.

  2. 2.

    In Zr-1Nb alloy formation of metastable β precipitates on tempering has been explained on the basis of computed free energy–composition plots. In general, the kinetics of tempering was found to be faster in the case of quaternary alloy.

  3. 3.

    In the case of Zr-1Nb alloy, the role of the monotectoid reaction in the Zr-Nb phase diagram on the evolution of β precipitates could be clearly noticed, where the samples annealed above and below the monotectoid temperature (883 K) exhibited substantially different β precipitate compositions. The Zr-rich β precipitates have been found when annealing was carried out above 883 K, while Nb-rich β precipitates were seen on annealing below 883 K. The desired microstructure for cladding application is one in which the precipitation of the Nb-rich β phase is complete or nearly so, because it significantly improves the corrosion resistance. Hence, from the present study, it can be suggested that annealing below 883 K (such as 853 K) for a longer duration (>4 hours) would be helpful in generating a microstructure in which the majority of the β precipitates are Nb rich.

  4. 4.

    In quaternary alloy, both β precipitates and Zr-Nb-Fe–type precipitates were observed in the tempered and recrystallized samples. Holding for a longer duration not only favored the formation of Zr-Nb-Fe–type precipitates but also led to their enrichment with Nb and Fe. The HREM examination revealed the intermetallic nature of the Zr-Nb-Fe precipitates.

  5. 5.

    Microstructural observations together with mechanical property data concluded that in both the binary and quaternary alloys annealing at 853 K for 1 hour led to the complete recrystallization of the cold-worked samples.

  6. 6.

    In general, the strength of the quaternary alloy was found to be higher than the binary alloy with comparable ductility.