- Email: [email protected]
Evolution of the radiation-induced defect structure in 316 type stainless steel after post-irradiation annealing
Accepted Manuscript Evolution of the radiation-induced defect structure in 316 type stainless steel after post-irradiation annealing W. Van Renterghem, M. Konstantinović, M. Vankeerberghen PII: DOI: Reference:
S0022-3115(14)00237-2 http://dx.doi.org/10.1016/j.jnucmat.2014.04.024 NUMA 48099
To appear in:
Journal of Nuclear Materials
Received Date: Accepted Date:
11 September 2013 21 April 2014
Please cite this article as: W. Van Renterghem, M. Konstantinović, M. Vankeerberghen, Evolution of the radiationinduced defect structure in 316 type stainless steel after post-irradiation annealing, Journal of Nuclear Materials (2014), doi: http://dx.doi.org/10.1016/j.jnucmat.2014.04.024
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Evolution of the radiation-induced defect structure in 316 type stainless steel after post-irradiation annealing
W. Van Renterghem*, M. Konstantinović, M. Vankeerberghen SCK•CEN, the Belgian nuclear research centre, Boeretang 200, 2400 Mol, Belgium
* Corresponding author Last name:
Van Renterghem
First name
Wouter
Address:
SCK•CEN NMS-MNA Boeretang 200 2400 Mol, Belgium
e-mail:
[email protected]
tel.:
+ 32 14 33 30 98
fax:
+ 32 14 32 12 16
e-mail other authors Milan Konstantinović: [email protected] Marc Vankeerberghen: [email protected]
Keywords: Transmission electron microscopy, irradiation assisted stress corrosion cracking, postirradiation annealing, stainless steel
Abstract The thermal stability of Frank loops, black dots, cavities and γ' precipitates in an irradiated 316 stainless steel was studied by transmission electron microscopy. The samples were retrieved from a thimble tube irradiated at around 320°C up to 80 dpa in a commercial nuclear power reactor, and thermally annealed, varying both annealing temperature and time. With increasing annealing temperature the density of all defects gradually decreased, resulting in the complete removal of Frank loops at 550°C. In contrast to other defects, the density of the γ' precipitates sharply decreased with increasing annealing time, which indicates that the dissolution of the γ' precipitates is governed by the iron diffusion length.
Highlights •
The effect of post-irradiation annealing on the microstructure of CW316 type stainless
steel irradiated to 80 dpa was investigated with transmission electron microscopy •
All Frank loops were removed after annealing at 550°C. They unfault and grow to large
perfect dislocation loops. •
The cavity density decreases after annealing at 550°C, but they were not completely
removed. •
The removal of Frank loops and cavities is controlled by the annealing temperature,
while the dissolution of γ' precipitates is controlled by the iron diffusion length.
1. Introduction Irradiation assisted stress corrosion cracking (IASCC) is an important effect limiting the integrity of light water reactor (LWR) core components, especially with respect to long-term operations [1]. In stainless steels, IASCC occurs as an interplay of various degradation processes due to the material being exposed to elevated temperature, stress, intense neutron bombardment and corrosion [2-9]. The complexity of IASCC is illustrated by the fact that in boiling water reactors (BWR) the IASCC occurs after about 1 dpa of accumulated damage, while in pressurized water reactors (PWR) it occurs after 4 dpa. Evidently, both changes of the water environment and the microstructure of the irradiated alloy [3] are influencing IASCC, but their individual contributions and the importance of their mutual interplay are not yet fully understood.
In the framework of the life extension of nuclear reactors to 60 years and beyond, mitigation strategies against IASCC are considered. They are important for the safety as well as for maintenance cost reductions, even though IASCC may not be a lifetime limiting factor because most stainless steel components are replaceable. Hereto, the water chemistry could be optimized by adding Zn to reduce IASCC and Li to increase pH [10]. Alloy compositions could be adjusted to reduce susceptibility to IASCC. It is considered beneficial [1] to have high Ni and Cr contents, low Si content, absence of brittle oxide and nitride inclusions, high coincident site lattice fraction of grain boundaries, low connectivity of high-angle grain boundaries and grain boundary coverage by chromium.
As an alternative, one could consider to recover the initial mechanical properties by thermal annealing. Post-irradiation annealing (PIA) experiments have shown that both hardening and radiation-induced segregation (RIS) are reduced, where the radiation-induced hardening appears to recover faster than RIS [11]. On the other hand, annealing at high temperatures will promote helium bubble growth, specifically at grain boundaries and line dislocations [12,13]. Helium is produced as a result of (n,α) interactions. It is practically insoluble in metals and it will cluster and form bubbles at the radiation induced defects. Upon annealing at high temperature, the radiation induced defects are removed and He will form bubbles at locations of low free energy, like line dislocations or grain boundaries. Bubbles grow by migration and coalescence [14]. In 304-type stainless steel, bubbles become visible after annealing at 650°C and reach a size of 1-4 nm [12]. PIA does not have a general healing effect for BWR conditions as the material is thermally sensitized during annealing, leading to cracking in high potential water. On the contrary, for PWR conditions, a clear decrease in IASCC was reported, and the decrease occurred even before hardness or RIS were significantly recovered [11,15]. The full recovery of IASCC was only reached when hardening as well as silicon segregation was significantly reduced.
The effects of PIA on the microstructure, as well as the importance of different annealing parameters, in particular the mutual influence of annealing temperature and time on the defect stability in highly irradiated 316 stainless steels, have not yet been studied in detail. Previously, experiments were performed on 304 [15,16] and 316 [12,17] type stainless steels irradiated to moderate doses of 25 dpa. For this paper, the effect of thermal annealing on 316 CW type stainless steel irradiated in a commercial reactor to 80 dpa was studied. Annealing conditions were selected to have a significant influence on the defect structure and to verify if the iron diffusion length is the optimal parameter to describe the defect dissolution.
2. Experimental The specimens which were used for the annealing experiments were retrieved from an irradiated thimble tube, extracted from the Tihange 2 PWR from position H11 in the reactor core. The tube was irradiated under normal PWR operation conditions. The maximal dose was 80 dpa and the irradiation temperature was around 320°C [18]. The tube is fabricated from a 316CW type stainless steel and the composition is given in Table 1. A tube segment of 10 mm in length was selected for this study from a part of the tube that was irradiated to 80 dpa. Four equal samples were fabricated and annealed under four different thermal conditions in a furnace installed in a hot cell. The furnace provides the possibility of temperature variation in the range from room temperature up to 1000 °C with a temperature stability in the whole temperature range of the order of 1-2 °C. Annealing was performed under vacuum, with a pressure of about 10-5 mBar. The oven was pre-heated to the annealing temperature before inserting the sample. In this way the sample heating time was reduced, ensuring an isothermal anneal.
(Insert Table 1 here)
The annealing parameters are summarized in Table 2. In PIA experiments on stainless steels it is common to combine the effect of the annealing temperature and time in a single parameter, the iron diffusion length. It is assumed that all recovery processes are driven by diffusion processes. The iron diffusion length is calculated using the formula:
with t the annealing time (seconds) and D the iron self-diffusion coefficient (m²/s) given by:
where k is the Boltzmann constant, T the annealing temperature (K) and Q the migration energy. Values are taken from Busby et al. [11]: Q = 1.29 eV and D0 = 1.54 × 10-6 m²/s.
(Insert Table 2 here)
After annealing, a thin slice was cut from the quarter tube and mechanically polished using SiC paper to a thickness of about 0.1 mm. A 1 mm disc was punched out of the polished slice and glued on a platinum 0.6 mm aperture grid with Struers Tripod wax. Finally, the samples were electrochemically polished until perforation by dual jet polishing in an electrolyte consisting of 5% perchloric acid and 95% methanol at -20°C and applying a voltage of 38 V.
The specimens were investigated in a JEOL 3010 transmission electron microscope operating at 300 kV. Conventional bright field, dark field and weak beam imaging techniques were applied. Frank loops were visualised with the relrod technique and cavities are revealed in under- and over-focused images. The local thickness of the sample was determined with convergent beam electron diffraction (CBED) [19].
3. Results 3.1 As-irradiated material The defect structure of the thimble tube after irradiation to 80 dpa is shown in the different images of Figure 1. The typical defects, observed in a stainless steel material irradiated to more than 1 dpa at 300°C [20], were found in this sample. Namely, the Frank loops (F), black
dots and small perfect dislocation loops (SL), are clearly visible in Figure 1a. The total defect density was found to be (10.0±0.5)×1022/m³.
(insert Figure 1 here)
Figure 1b is a dark field image obtained by using the relrod diffraction intensity which is induced by the Frank loops. In the inset, the corresponding diffraction pattern is shown. The white circle indicates the part of the diffraction pattern that was selected for the dark field image. It only includes the relrod diffraction intensity. Examination of several relrod images allows the estimation of the total amount of Frank loops and the density was found to be about (7.2±0.3)×1022/m³.
Apart from the Frank loops, round features were observed in Figure 1b as well, which can be attributed to precipitates. The properties of these precipitates agree with previous observations of γ' precipitates (Ni3Si) in neutron irradiated 304 and 316 types of stainless steel [21]. Ni3Si has a fcc crystal structure with a lattice parameter of 0.3505 nm, which is slightly smaller than the lattice parameter of AISI 316 steel. The diffraction patterns are very similar but because the lattice parameter is slightly lower, the locations of the diffraction spots are slightly further away from the transmitted beam than those from the steel. There is no orientation relation with the stainless steel and therefore part of the precipitates generate diffraction spots near the relrods and are visible in the relrod image. Since not all precipitates are visible under the applied imaging conditions, the measured precipitate density of about 0.5×1022/m³ is probably an underestimation of the actual density.
Figure 1c contains an out-of-focus image of the edge of the sample. In this image, cavities appear as white spots and a few examples are indicated with a letter V. The cavities can be either gas filled bubbles or voids. In this paper, we do not differentiate those with or without gas and in the remaining of the text we refer to them as cavities. Both have been reported to occur in stainless steel irradiated to high doses at a temperature above 300°C [20,22]. The cavity density in the as-irradiated material was measured to be (12.7±0.5)×1022/m³.
3.2 Annealing at 500°C for 6h The general radiation-induced defect structure after annealing at 500°C for 6h is shown in Figure 2. Similarly to the as-irradiated sample, the same classes of defects were found. In Figure 2a one example of a Frank loop (F), black dot (BLD), precipitate (P) and perfect dislocation loop (SL) are indicated. However, unlike the as-irradiated material, larger dislocation structures are observed, which appear to be grown dislocation loops. An example is indicated by LL in Figure 2a. No stacking fault contrast can be observed in the interior of the dislocation loop, which means that it is not a Frank loop, but a perfect dislocation loop. The total defect density, without the large loops and cavities, measured at three different locations, was about (4.9±0.5)×1022/m³.
(insert Figure 2 here)
Further, Figure 2b shows a typical dark field image using the relrod diffraction intensity induced by Frank loops. It is clear that some Frank loops remain after annealing, but their amount has significantly reduced. An average density of (1.4±0.5)×1022/m³ was measured, which is a factor of 6 lower with respect to the as-irradiated sample.
Besides the Frank loops, round contrast induced by γ' precipitates was found in Figure 2b as well. The precipitate density after annealing was found to be 0.45×1022/m3, indicating that annealing at 500°C for 6h had little or no effect on the γ' precipitates.
The presence of cavities was revealed in the out-of-focus image of Figure 2c, where they appear as white spots. The cavity density was measured to be (11.1±0.5)×1022/m³, which is comparable to the cavity density in the as-irradiated material. It suggests that annealing at 500°C for 6h did not affect the cavities.
The densities of each type of defect and their average sizes are summarized in Table 3 and Table 4, respectively. The precision on the average size, given in the table, is a measure for the width of the size distribution. Large values simply indicate that there is a significant variation in the defect size. The large dislocation loops which are the result of the annealing are difficult to quantify and were not included in the table. Compared to the density and sizes in the as-irradiated material, this anneal mainly affected the Frank loops. The density has reduced with a factor of 6 and also the average size was significantly smaller. It suggests that mainly the larger Frank loops have grown and unfaulted, whilst the smaller loops remain in the material.
3.3 Annealing at 550°C for 6h The microstructure after annealing at 550°C for 6h is shown in Figure 3. Similar to the sample annealed at 500°C for 6h a significant amount of small defects are still present. Mainly black dots and perfect dislocation loops were found. The total concentration of small defects is slightly lower than in the previous sample: a defect density of 2.8±0.8×1022/m³ was measured, not including cavities. It should be noted that perfect dislocation loops in thin TEM foils can
migrate towards the surface which may result in an underestimation of the defect density [23]. However, no migration of loops was observed during the TEM investigation. Further, it was found that several loops have grown to form large dislocation loops. The growth of the loops is affected by the small defects. Several examples can be observed in Figure 3a, where the small defects are interacting and pinning the large loops. Because of this pinning, the growth of the loops is tempered and their circular shape becomes more irregular.
(Insert Figure 3 here)
To characterize the type of defect, it was attempted to visualize the presence of Frank loops by using the relrod contrast method. Figure 3b is recorded at approximately the same location in the sample as Figure 3a, but after tilting the sample to an orientation close to the [0-11] zone axis. This orientation is optimal to reveal the relrods, but as can be seen in the diffraction pattern in the inset, no clear relrods are present. Moreover, the dark field image of Figure 3b was obtained when inserting an objective aperture at the positions indicated by the white circle in the diffraction pattern. In this way, any contribution from the matrix material to the image intensity is excluded and if Frank loops are present, they should appear as thin bright lines. The intensity in both dark field images is very low and no linear contrast features can be recognized, which means that no Frank loops are present anymore after annealing at 550°C for 6h.
Similarly, it was verified if there are still γ' precipitates (Ni3Si) present after these annealing conditions. In the previous anneal, precipitates appeared in the dark field images when selecting the relrod diffraction intensity. In Figure 3b no clear precipitates can be observed. Moreover, no additional weak diffraction spots were found in the diffraction patterns. These
observations would suggest that also the precipitates have all dissolved during annealing at 550°C for 6h. However, in the bright field image of Figure 3a a few defects were observed that show a moiré fringe pattern. A few of these defects are indicated by the letter "m". A moiré pattern in a bright field images is usually the result of the overlap of two crystal lattices with a slightly different lattice parameter or orientation. Even though, the TEM investigation is not as clear on this point, indications are found that a small amount of precipitates remains after annealing at 550°C for 6h.
The presence of cavities was verified in the out-of-focus bright field image of Figure 3c. Compared to the as-irradiated and the 500°C, 6h anneal, cavities are still present, but the density is lower after annealing at 550°C, for 6h. A density of 5.1±0.5×1022/m³ was measured, which is about half the density of the previous annealing (500°C,6h).
The defect densities and average sizes were measured and the results are given in Table 3 and Table 4, respectively. Again, the large unfaulted loops were not taken into account in the quantification. Furthermore, no Frank loops were observed anymore. The values for the precipitates are based on the observation of the moiré fringes as discussed above. Whereas the size of the defects is comparable to the annealing at 500°C for 6h, the defect density and in particular the cavity density is reduced.
3.4 Annealing at 500 °C for 19.5 h For the third annealing condition, a temperature of 500°C was maintained, but the annealing time was extended to 19.5 hours. Applying the parameters used by Busby et al. [11], the iron diffusion length is kept constant as in the case of thermal annealing at 550°C, for 6h. Therefore, if the defect evolution is controlled by the iron diffusion length, the results of the
TEM analysis should be the same as in the previous paragraph. However, this is found not to be the case.
Figure 4 shows a few typical images of the defect structure. Similar to the two previously annealed samples (500°C, 6h and 550°C, 6h), it was observed that many small loops have grown to become large perfect dislocation loops. However, at the same time, a large number of small defects remain. The total density of small defects was measured to be 4.7±0.5×1022/m³, not including cavities. This defect density is higher than in the sample annealed at 550°C for 6h and comparable to the density in the sample annealed at 500°C for 6h.
(Insert Figure 4 here)
Related to the type of defects, it was found that all types of defects detected in the sample annealed at 500°C for 6h were also found here. In Figure 4a a few examples are indicated of a large, perfect dislocation loop (LL), a small dislocation loop (SL), a black dot (BLD) and a Frank loop (F). A more clear evidence of the presence of Frank loops is given in Figure 4b, where the relrod imaging technique was used. Under these conditions, a number of parallel line segments appear in the dark field image, which can be attributed to Frank loops. The amount of Frank loops was 1.8±0.5×1022/m³. This value is of the same order as in the sample annealed at 500°C for 6h. The longer annealing time did not influence the Frank loop density.
Apart from the line segments a few round features can be recognized. Based on previous results, it is reasonable to assume that this contrast is induced by small γ'-precipitates, which were also detected in the sample annealed at 500°C for 6h. As mentioned already, it is not
possible to visualize all precipitates in a single dark field image, so it is difficult to measure their total amount. However, by comparing to the precipitate number obtained from the image in Figure 2b, which was recorded under similar conditions, the amount of precipitates is found to be clearly different. In this sample, annealed at 500°C for 19.5h, an amount of 0.15×1022/m³ of precipitates were measured. This value is a factor of 3 lower than in the sample annealed at 500°C for 6h and of the same order as the sample annealed at 550°C for 6h.
The presence of cavities was evaluated in out-of-focus bright field images as for example shown in Figure 4c. The cavity density was measured to be 10.9±0.5×1022/m³, which is of the same order as the amount of cavities after annealing at 500°C for 6h.
The defect densities of each type of defect and their average sizes are summarized in Table 3 and Table 4, respectively. The precipitate density was determined from the relrod images which also show the Frank loops. Even though it is not possible to visualize all precipitates in a single image, it was found that the precipitate density is a factor of 3 lower than in the sample annealed for 6h. Because the conditions under which these images were recorded are similar, the reduction of the amount of observable precipitates cannot be due to the imaging conditions only. Therefore it can be stated that the reduction of the amount of precipitates is significant and due to the prolonged annealing time. The defect densities of the Frank loops, all other loops and cavities are similar to the value after 6h annealing at 500°C.
Also, the average defect sizes are comparable with the sample annealed for 6h. Some Frank loops have grown to sizes above 30 nm, which slightly increases the average value and size distribution. The unfaulted loops, not including the large loops, are slightly larger. However, the size of the black dots and the cavities were not affected and, because these defects occur
predominantly, also the overall defect size was not significantly affect by the prolonged annealing time.
3.5 Annealing at 600°C for 6h The annealing of the fourth sample was performed at 600°C for 6h. These parameters give an iron diffusion length of 0.0344 cm. The general defect structure is shown in Figure 5 and is comparable to the previously annealed samples. A significant amount of small black dots was observed. Furthermore, it was found that the initially formed Frank loops have unfaulted to form perfect dislocation loops. Most of these loops have grown to large dislocation loops which interact with other dislocation loops and are pinned by the remaining black dots and small dislocation loops. The total density of small defects was measured to be 1.3±0.3×1022/m³, not including cavities. This density is the lowest of all four anneals and confirms the trend that an increased annealing temperature reduces the amount of small radiation defects.
(Insert Figure 5 here)
In Figure 5b, the TEM picture, obtained under relrod imaging conditions is presented. The corresponding diffraction pattern in the inset shows that the sample is oriented close to the [110] zone axis, which is optimal to get relrod diffraction intensity. No Frank loops were found. The complete removal of the Frank loops was expected as they were already removed after annealing at 550°C.
After the 550°C, 6h anneal, not all precipitates were annealed. However, in this sample, no traces of small γ' precipitates were found anymore. In the relrod images, clear precipitate contrast could not be observed, nor was a moiré pattern found in the bright field images.
Of particular interest is the effect of the higher annealing temperature on the cavities. Figure 5c shows a high magnification bright field image in out-of-focus condition. The cavity density was measured to be 5.1±0.2×1022/m³, which is of the same order as the amount of cavities after annealing at 550°C for 6h.
The defect densities and average sizes were measured and the results are summarized in Table 3 and Table 4. No Frank loops or precipitates were observed anymore and only the black dots, unfaulted loops and cavities could be measured. Compared to the sample annealed at 550°C for 6h, the density of the black dots and unfaulted loops is lower, but the cavity density is comparable. The average sizes of the defects are comparable.
4. Discussion The annealing conditions were selected to have three different temperatures and three different iron diffusion lengths. In the literature, the calculated iron diffusion length is used as a single parameter to compare different annealing conditions, even though it was not verified in detail if this parameter is suitable to describe the evolution of the microstructure. Moreover, at least two different parameters for calculation of the iron diffusion length can be found in the literature. Both use the same equations, but they use different values for the activation energy and iron self-diffusion coefficient (D0). Fukuya et al. [17] use a high activation energy of 2.95 eV and a larger D0 of 4.90×10-5 m²/s. Because of the high activation energy, the iron diffusion length is mainly determined by the annealing temperature. Busby et
al. [11], which was used in this report, use a much lower activation energy of 1.29 eV. Applying these parameters, the temperature is still an important factor in the calculation of the iron diffusion length, but, as shown in anneals two and three, it is possible to reach the same diffusion length at a lower annealing temperature within a reasonable annealing time.
To evaluate the effect of the annealing temperature and iron diffusion length, all defect densities and average sizes are summarized in Table 3 and Table 4. Figure 6 shows the evolution of the defect density as a function of the annealing temperature (Figure 6a) or the iron diffusion length (Figure 6b).
(Insert Table 3, Table 4 and Figure 6 here)
For all four annealing conditions that were applied for this investigation, a significant effect on the microstructure was stated. In stainless steel irradiated to high doses, a high density of Frank type loops, black dots, γ' precipitates and cavities are formed. After annealing, large dislocation loops were found in all samples, while the Frank loops, γ' precipitates and cavities were removed. This evolution of the defect structure is beneficial for the mitigation of IASCC. In the following paragraphs, a process is proposed for the evolution of the defects during annealing.
4.1 Frank loops In this work we did not characterize the nature of the loops, but according to the literature we can safely assume that neutron irradiation created both vacancy and interstitial types [24,25] under the present irradiation conditions. The growth mechanism is not yet elucidated, and it might be very different for different types of loops and sizes. For example, it is known that at
higher temperatures the vacancy clusters become unstable and start to emit vacancies to the structure [26]. These vacancies will diffuse towards the Interstitial type Frank loops and annihilate the interstitials. These Frank loops will shrink and eventually disappear. On the other hand, Frank loops of vacancy type will start to grow. Alternatively, a recent study shows that interstitial loops in pure iron can grow during annealing at about 550 – 600 °C presumably by vacancy emission [27]. With increasing loop size, also the stacking fault area within the loop grows and it becomes energetically favourable to unfault and to form perfect dislocation loops. Unfaulting of a Frank loops occurs most probably by the nucleation of a Shockley partial within the loop, induced by thermal activation [28,29]. We find that the loops, once they are of certain size, mutually interact and form large dislocation loops. During growth, the loop can interact with remaining small defects like the black dots and be pinned. These pinning effects explain why the large loops have an irregular shape.
All annealing conditions applied in this study induce a decrease of the number of Frank loops and annealing at temperatures of 550°C or higher even completely removed all Frank loops. Compared to the results in the AISI 304 steel [15], a higher annealing temperature is required in 316CW steel to obtain the same effect. In the 304 steel, all Frank loops were removed after annealing at 500°C, while it was shown here that annealing needed to be performed at 550°C in 316CW steel.
The removal of the Frank loops is controlled by the annealing temperature and not by the iron diffusion length. The samples annealed at 500°C for 19.5h and at 550°C for 6h have the same iron diffusion length, but the first sample still contains Frank loops, while the second sample does not. Moreover, the number of Frank loops after annealing at 500°C for 19.5h is comparable to the number after annealing at 500°C for 6h.
This evolution is visually represented in the graphs in Figure 6. In the graph of Figure 6a, there are two data points at an annealing temperature of 500°C which agree within the experimental error. In the graph of Figure 6b, there are two data points at an iron diffusion length of 0.00205 cm. Here the difference in the measured defect density is larger than the experimental error. It means that the removal of the Frank loops is controlled by the annealing temperature and not by the iron diffusion length. This observation supports the loop growth mechanism based on vacancies [26]. Moreover, our finding is in good agreement molecular dynamics calculation [30,31] which show that dissolution of He-V clusters occurs at about 750K (477°C) and consistent with annealing behaviour experimentally observed for cavities, as discussed in the next paragraph.
4.2 Cavities In a similar way, the graphs of Figure 6 show that also the evolution of the cavities is rather controlled by the annealing temperature than by the iron diffusion length. Annealing at 500°C results in a slight reduction of the cavity density compared to the as-irradiated sample. The reduction in density is still moderate. It is important to note that the longer annealing time did not influence the cavity density. The average size of the cavities is not affected by the annealing. All obtained sizes are equal within the precision of the measurement.
When annealing at 550°C, a significant decrease of the cavity density was stated. Only halve of the cavities remain. This observation shows again that also for the cavities, the annealing temperature is the controlling factor and not the iron diffusion length.
However, annealing at 600°C did not further decrease the cavity density. This may be related to differences in thermal stability of different cavities and in particular between He filled
bubbles and voids. Molecular dynamics simulations and first principles calculations [30,31] showed that the binding energy of vacancies in clusters increase with increasing number of He impurity atoms. Moreover, it is suggested that HeV clusters dissociate around 750K (477°C) while higher stability clusters dissociate at 800 – 1000K (527 - 727°C) [31]. Cole and Allen [32] noted that during annealing of high temperature irradiated 304 type stainless steel, part of the voids shrunk while other voids were not affected. They also observed the creation of new, small bubbles during annealing, which they related to the trapping of vacancies at small He clusters. Applying these statements to the present experiments, suggests that the voids emitted vacancies during annealing and were annihilated completely at 550°C, while the cavities that remained are He filled gas bubbles. The removal of the voids is responsible for the reduction of the cavity density.
The evolution of the cavities is in agreement with the evolution of the Frank loops. At 500°C only the vacancy clusters and part of the voids were annealed, which did not generate sufficient vacancies to remove all Frank loops. More vacancies became available at higher annealing temperatures, which initiated the complete removal of all Frank loops.
4.3 Precipitates The effect of annealing on the precipitates is extremely important and interesting. Previous studies showed that after the post-irradiation annealing at 500°C of AISI 304 the precipitates are no longer visible by TEM [15]. Subsequent atom probe tomography on the same samples [16] indicated that precipitates do not dissolve completely during annealing, but that the Ni and Si slowly diffuse giving a broadening of the precipitate and a gradual decrease of the concentration of the Ni and Si. The combination of TEM and APT results suggested that the precipitate does not dissolve completely, since the precipitate remains in some form of Ni-Si
cluster but without a full crystallographic structure. These clusters no longer provide diffraction contrast in the TEM images. For these defects, diffusion plays an important role in their evolution, so the iron diffusion length is considered as the relevant parameter. Additionally, the formation mechanism of γ' precipitates is governed by radiation induced segregation of Ni and Si towards sinks [20], so it is reasonable to expect that dissolution is controlled by diffusion of iron and other depleted elements towards the precipitates.
The results obtained in this study confirm that the iron diffusion length correctly describes the evolution of the precipitates. Annealing at 500°C for 6h reduces the amount of precipitates to about one halve. A longer annealing time further reduces the precipitate density. Moreover, it was found that in the sample annealed at 550°C for 6h, which has the same iron diffusion length as the 500°C, 19.5h annealed sample, also has a similar precipitate density. Annealing at 600°C for 6h, which has the longest iron diffusion length, completely removed the precipitates.
These findings are also represented in Figure 6. A significant difference in precipitate density after the two anneals at 500°C but with different annealing times is clearly observed. If the two annealing conditions with the same iron diffusion length are compared, similar densities were measured.
5. Conclusions 316CW stainless steel samples, obtained from a thimble tube irradiated to 80 dpa in the Tihange 2 reactor, were annealed after irradiation, applying four different conditions. The annealing times and temperatures were selected to have three different annealing temperatures and three different iron diffusion lengths. All four samples were investigated with TEM to
reveal the changes in microstructure as a result of the annealing. In this work the thermal stability of various types of defects created by high dose neutron irradiation was discussed.
It was found in all samples that part of the radiation-induced Frank loops grew, were unfaulted and formed large perfect dislocation loops. The growing loops are pinned by the remaining radiation-induced defects. After annealing at 500°C, a reduced number of Frank loops were still present, regardless of the annealing time of 6h or 19.5h. Annealing at 550°C or higher completely removed all Frank loops. The required annealing temperature to remove defects is about 50°C higher than in AISI304 steel.
The cavity density and size was hardly affected by the annealing at 500°C. However, when annealing at 550°C, the density was halved. Annealing at 600°C did not further reduce the cavity density. For both the Frank loops and the cavities, the results show that the annealing temperature is the qualifying parameter, while the iron diffusion length does not give an accurate description.
The results related to the thermal stability of the γ' precipitates confirm that the iron diffusion length correctly describes their evolution. It was found that a similar precipitation density exists after annealing conditions having the same iron diffusion length. This is found to be in clear contrast to the behaviour of all other defects such as Frank loops, cavities and black dots. Annealing at 600°C, which has the longest iron diffusion length, completely removed the precipitates.
The evolution of the defect structure during post-irradiation annealing and, in particular, the removal of a large fraction of the radiation induced defects is beneficial for the mitigation of
IASCC. Therefore PIA is relevant for long term operation of internal components and nuclear power plant life time extension.
Acknowledgements This work was financially supported by the PERFORM60 project, as a part of the 7th EURATOM FRAMEWORK PROGRAMME 2007-2013 (FP7-232612-PERFORM60). The authors would like to thank GDF-Suez Tractebel for delivery and use of the highly irradiated thimble tube, Kris Kaers for his indispensable help with sample preparation and Dmitry Terentyev for the useful discussions.
References [1] S.J. Zinkle, G.S. Was., Acta Materialia 61 (2013) 735-758 [2] P.M. Scott, J. Nucl. Mat. 211 (1994) 101-122. [3] P.L. Andresen, F.P. Ford, S.M. Murphy, J.M. Perks, in: Proceedings of 4th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, NACE International, Houston, 1990, pp. 1-83–1-121. [4] G.S. Was, P.L. Andresen, Corrosion, 61 (2007) 19-45. [5] G.S. Was, J.T. Busby, Phil. Mag. 85 (2005) 443-465. [6] T. Onchi, K. Dohi, N. Soneda, M. Navas, M.L. Castaño, J. Nucl. Mater. 340 (2005) 219236. [7] L. Fournier, M. Savoie, D. Delafosse, J. Nucl. Mater. 366 (2007) 187-197. [8] Z. Jiao, J.T. Busby, G.S. Was, J. Nucl. Mater. 361 (2007) 218-227. [9] Z. Jiao, G.S. Was, J. Nucl. Mater. 382 (2008) 203-209.
[10] R. Yang, B. Cheng, J. Deshon, K. Edsinger and O. Ozer, J. Nucl. Sci. & Tech. 43, 9 (2006) 951-959. [11] J. T. Busby, G. S. Was, E. A. Kenik, J. Nucl. Mater. 302 (2002) 20-40. [12] Y. Ishiyama, M . Kodama, N. Yokota, K. Asano, T. Kato, K. Fukuya, J. Nucl. Mater. 239 (1996) 90-94. [13] F. Carsughi, H. Ullmaier, H. Trinkaus, W. Kesternich, V. Zell, J. Nucl. Mater 212-215 (1994) 336-340. [14] V. Zell, H. Trinkaus, H. Schroeder, J. Nucl. Mater. 212-215 (1994) 320-324 [15] W. Van Renterghem, A. Al Mazouzi, S. Van Dyck, J. Nucl. Mater. 413 (2011) 95-102. [16] T. Toyama, Y. Nozawa, W. Van Renterghem, Y. Matsukawa, M. Hatakeyama, Y. Nagai, A. Al Mazouzi, S. Van Dyck, J. Nucl. Mater. 418 (2011) 62-68. [17] K. Fukuya, M. Nakano, K. Fujii, T. Torimaru, Y. Kitsunai, J. Nucl. Sci. Tech. 41 (2004) 1218-1227. [18] R. Gerard, GDF Suez, private communication [19] P.M. Kelly, P.A. Jostons, R.G. Blake, J.G. Napier, Phys. Stat. Sol. A31 (1975) 771-780. [20] S.M. Bruemmer, E.P. Simonen, P.M. Scott, P.L. Andresen, G.S. Was, J.L. Nelson, J. Nucl. Mater 274 (1999) 299-314. [21] P.J. Maziasz, J. Nucl Mater. 205 (1993) 118-145. [22] F.A. Garner, I. Porollo, Yu. V. Konobeev, O.P. Maksimkin, Proceedings of the 12th International Conference on Environmental Degradation of Materials in Nuclear Power System – Water Reactors, TMS, (2005) 439-448. [23] M.R. Gilbert, Z. Yao, M.A. Kirk, M.L. Jenkins, S.L. Dudarov, J. Nucl. Mater. 386-388 (2009) 36-40. [24] D.J. Edwards, E.P. Simonen, S.M. Bruemmer, J. Nucl. Mater. 317 (2003) 13-31 [25] R. Stoenescu, R. Schäublin, D. Gavillet, N. Baluc, J. Nucl. Mater. 360 (2007) 186-195. [26] S.J. Zinkle, P.J. Maziasz, R.E. Stoller, J. Nucl. Mater. 206 (1993) 266-286.
[27] S. Moll, T. Jourdan, H. Lefaix-Jeuland, Phys. Rev. Lett 111,1 (2013) 015503. [28] E.H. Lee, M.H. Yoo, T.S. Byun, J.D. Hunn, K. Frarrell, L.K. Mansur, Acta Mater. 49 (2001) 3277-3287. [29] A. Bakaev, D. Terentyev, X. He, E.E. Zhurkin, Nucl. Instr. And Meth. in Phys. Res. B 303 (2013) 33-36 [30] K. Morishita, R. Sugano, B.D. Wirth, T. Diaz de la Rubia, Nucl. Instr. and Meth. In Phys. Res. B, 202 (2003) 76-81. [31] C.-C. Fu, F. Willaime, J. Nucl. Mater. 367-370 (2007) 244-250. [32] J.I. Cole, T.R. Allen, J; Nucl. Mater. 283-287 (2000) 329-333.
Figure Captions Figure 1. a) Bright field image of the as-irradiated thimble tube, containing Frank loops (F), black dots (BLD) and small loops (SL). b) Relrod image of the Frank loops (F) and a fraction of the γ' precipitates (P). The diffraction pattern and the position of the objective aperture (white circle) are given in the inset. c) Out-of-focus image to visualize cavities (V).
Figure 2. a) Bright field image of the defect structure after annealing at 500°C for 6h. Examples of precipitates (P), Frank loops (F), black dots (BLD), small loops (SL) and large loops (LL) are indicated. b) Relrod image of the remaining Frank loops and γ' precipitates. The inset shows the diffraction pattern and the white circle indicates the location of the objective aperture. c) Out-of-focus images to visualize cavities (V).
Figure 3. a) Bright field image of the defect structure after annealing at 550°C for 6h. The moiré fringes (m) are induced by precipitates. b) Dark field image using relrod contrast. The inset shows the diffraction pattern and the white circle indicates the location of the objective aperture. c) Out-of-focus image to visualize cavities (V).
Figure 4. a) Bright field image of the defect structure after annealing at 500°C for 19.5h, containing Frank loops (F), black dots (BLD), small loops (SL) and large loops (LL). b) Relrod image of the remaining Frank loops (F) and a fraction of the γ' precipitates (P). The inset shows the diffraction pattern and the white circle indicates the location of the objective aperture. c) Out-of-focus image to visualize cavities (V).
Figure 5. a) Bright field image of the defect structure after annealing at 600°C for 6h. Examples of Large dislocation loops (LL), small unfaulted loops (SL) and black dots are indicated. b) Dark field image recorded under relrod conditions.
The inset shows the
diffraction pattern and the white circle indicates the position of the objective aperture. c) Outof-focus image revealing the presence of cavities (V).
Figure 6. Evolution of the defect density of the Frank loops, cavities, all loops and all defects(left Y-axis) and the precipitates (right Y-axis) a) as a function of the annealing temperature and b) as a function of the iron diffusion length. (in colour on web, grey scale in printed copy)
Tables Table 1. Composition of the 316CW stainless steel. C Si Mn P S
Cr
Ni
Mo
Co
0.044
17.40
12.8
2.68
0.07
0.53
1.79
0.022
0.009
Table 2. Annealing parameters of the different specimens. Segment anneal T (°C) anneal time (h)
Fe diffusion length (cm)
quarter 1
500
6
0.00114
quarter 2
550
6
0.00204
quarter 3
500
19.5
0.00205
quarter 4
600
6
0.00344
Table 3. Defect densities of all defects before and after annealing. Loops and precipitates Density All loops (1022/m³) precip. Frank BLD unf.1) & precip.
cavities
All
As irr.
0.5 2)
7.2 ± 0.3
-
-
10.0 ± 0.5
12.7 ± 0.5
22.7 ± 0.7
500°C, 6h
0.45 2)
1.4±0.5
- 3)
- 3)
4.9±0.5
11.1 ± 0.5
16.0 ± 0.7
550°C, 6h
0.16 2)
-
2.8±0.8
0.56±0.2
3.5±0.9
5.1 ± 0.5
8.6 ± 1.0
500°C, 19.5h
0.15 2)
1.8±0.5
- 3)
- 3)
4.7±0.5
10.9 ± 0.5
15.6 ± 0.7
600°C, 6h
-
-
1.1±0.5
0.2±0.1
1.3±0.5
5.1 ± 0.3
6.4 ± 0.6
1) 2)
The unfaulted loops do not include the large loops formed during annealing. underestimation as discussed above
3)
not measured individually
Table 4. Average sizes of all defects before and after annealing Loops and precipitates Size (nm)
cavities
All
All loops precip.
Frank
BLD
unf. & precip.
As-irr.
5.4±1.4 8.9 ± 4.0
-
-
8.7±3.8
1.7 ± 0.4
4.8±1.7
500°C, 6h
5.0±1.2
6.9±3.1
4.3±1.4
9.9±3.0
5.7±2.1
1.9 ±0.4
3.1±1.0
550°C, 6h
6.4±1.2
-
4.9±1.7
11.2±3.8
5.9±3.0
1.7±0.3
3.4±0.7
500°C, 19h
5.7±1.2
8.2±6.8
4.1±1.1
12.4±4.4
6.3±4.7
1.8 ±0.3
3.2±1.6
600°C, 6h
-
-
4.3±1.1
9.8±3.4
5.1±1.4
1.7±0.3
2.4±0.4
S0022-3115(14)00237-2 http://dx.doi.org/10.1016/j.jnucmat.2014.04.024 NUMA 48099
To appear in:
Journal of Nuclear Materials
Received Date: Accepted Date:
11 September 2013 21 April 2014
Please cite this article as: W. Van Renterghem, M. Konstantinović, M. Vankeerberghen, Evolution of the radiationinduced defect structure in 316 type stainless steel after post-irradiation annealing, Journal of Nuclear Materials (2014), doi: http://dx.doi.org/10.1016/j.jnucmat.2014.04.024
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Evolution of the radiation-induced defect structure in 316 type stainless steel after post-irradiation annealing
W. Van Renterghem*, M. Konstantinović, M. Vankeerberghen SCK•CEN, the Belgian nuclear research centre, Boeretang 200, 2400 Mol, Belgium
* Corresponding author Last name:
Van Renterghem
First name
Wouter
Address:
SCK•CEN NMS-MNA Boeretang 200 2400 Mol, Belgium
e-mail:
[email protected]
tel.:
+ 32 14 33 30 98
fax:
+ 32 14 32 12 16
e-mail other authors Milan Konstantinović: [email protected] Marc Vankeerberghen: [email protected]
Keywords: Transmission electron microscopy, irradiation assisted stress corrosion cracking, postirradiation annealing, stainless steel
Abstract The thermal stability of Frank loops, black dots, cavities and γ' precipitates in an irradiated 316 stainless steel was studied by transmission electron microscopy. The samples were retrieved from a thimble tube irradiated at around 320°C up to 80 dpa in a commercial nuclear power reactor, and thermally annealed, varying both annealing temperature and time. With increasing annealing temperature the density of all defects gradually decreased, resulting in the complete removal of Frank loops at 550°C. In contrast to other defects, the density of the γ' precipitates sharply decreased with increasing annealing time, which indicates that the dissolution of the γ' precipitates is governed by the iron diffusion length.
Highlights •
The effect of post-irradiation annealing on the microstructure of CW316 type stainless
steel irradiated to 80 dpa was investigated with transmission electron microscopy •
All Frank loops were removed after annealing at 550°C. They unfault and grow to large
perfect dislocation loops. •
The cavity density decreases after annealing at 550°C, but they were not completely
removed. •
The removal of Frank loops and cavities is controlled by the annealing temperature,
while the dissolution of γ' precipitates is controlled by the iron diffusion length.
1. Introduction Irradiation assisted stress corrosion cracking (IASCC) is an important effect limiting the integrity of light water reactor (LWR) core components, especially with respect to long-term operations [1]. In stainless steels, IASCC occurs as an interplay of various degradation processes due to the material being exposed to elevated temperature, stress, intense neutron bombardment and corrosion [2-9]. The complexity of IASCC is illustrated by the fact that in boiling water reactors (BWR) the IASCC occurs after about 1 dpa of accumulated damage, while in pressurized water reactors (PWR) it occurs after 4 dpa. Evidently, both changes of the water environment and the microstructure of the irradiated alloy [3] are influencing IASCC, but their individual contributions and the importance of their mutual interplay are not yet fully understood.
In the framework of the life extension of nuclear reactors to 60 years and beyond, mitigation strategies against IASCC are considered. They are important for the safety as well as for maintenance cost reductions, even though IASCC may not be a lifetime limiting factor because most stainless steel components are replaceable. Hereto, the water chemistry could be optimized by adding Zn to reduce IASCC and Li to increase pH [10]. Alloy compositions could be adjusted to reduce susceptibility to IASCC. It is considered beneficial [1] to have high Ni and Cr contents, low Si content, absence of brittle oxide and nitride inclusions, high coincident site lattice fraction of grain boundaries, low connectivity of high-angle grain boundaries and grain boundary coverage by chromium.
As an alternative, one could consider to recover the initial mechanical properties by thermal annealing. Post-irradiation annealing (PIA) experiments have shown that both hardening and radiation-induced segregation (RIS) are reduced, where the radiation-induced hardening appears to recover faster than RIS [11]. On the other hand, annealing at high temperatures will promote helium bubble growth, specifically at grain boundaries and line dislocations [12,13]. Helium is produced as a result of (n,α) interactions. It is practically insoluble in metals and it will cluster and form bubbles at the radiation induced defects. Upon annealing at high temperature, the radiation induced defects are removed and He will form bubbles at locations of low free energy, like line dislocations or grain boundaries. Bubbles grow by migration and coalescence [14]. In 304-type stainless steel, bubbles become visible after annealing at 650°C and reach a size of 1-4 nm [12]. PIA does not have a general healing effect for BWR conditions as the material is thermally sensitized during annealing, leading to cracking in high potential water. On the contrary, for PWR conditions, a clear decrease in IASCC was reported, and the decrease occurred even before hardness or RIS were significantly recovered [11,15]. The full recovery of IASCC was only reached when hardening as well as silicon segregation was significantly reduced.
The effects of PIA on the microstructure, as well as the importance of different annealing parameters, in particular the mutual influence of annealing temperature and time on the defect stability in highly irradiated 316 stainless steels, have not yet been studied in detail. Previously, experiments were performed on 304 [15,16] and 316 [12,17] type stainless steels irradiated to moderate doses of 25 dpa. For this paper, the effect of thermal annealing on 316 CW type stainless steel irradiated in a commercial reactor to 80 dpa was studied. Annealing conditions were selected to have a significant influence on the defect structure and to verify if the iron diffusion length is the optimal parameter to describe the defect dissolution.
2. Experimental The specimens which were used for the annealing experiments were retrieved from an irradiated thimble tube, extracted from the Tihange 2 PWR from position H11 in the reactor core. The tube was irradiated under normal PWR operation conditions. The maximal dose was 80 dpa and the irradiation temperature was around 320°C [18]. The tube is fabricated from a 316CW type stainless steel and the composition is given in Table 1. A tube segment of 10 mm in length was selected for this study from a part of the tube that was irradiated to 80 dpa. Four equal samples were fabricated and annealed under four different thermal conditions in a furnace installed in a hot cell. The furnace provides the possibility of temperature variation in the range from room temperature up to 1000 °C with a temperature stability in the whole temperature range of the order of 1-2 °C. Annealing was performed under vacuum, with a pressure of about 10-5 mBar. The oven was pre-heated to the annealing temperature before inserting the sample. In this way the sample heating time was reduced, ensuring an isothermal anneal.
(Insert Table 1 here)
The annealing parameters are summarized in Table 2. In PIA experiments on stainless steels it is common to combine the effect of the annealing temperature and time in a single parameter, the iron diffusion length. It is assumed that all recovery processes are driven by diffusion processes. The iron diffusion length is calculated using the formula:
with t the annealing time (seconds) and D the iron self-diffusion coefficient (m²/s) given by:
where k is the Boltzmann constant, T the annealing temperature (K) and Q the migration energy. Values are taken from Busby et al. [11]: Q = 1.29 eV and D0 = 1.54 × 10-6 m²/s.
(Insert Table 2 here)
After annealing, a thin slice was cut from the quarter tube and mechanically polished using SiC paper to a thickness of about 0.1 mm. A 1 mm disc was punched out of the polished slice and glued on a platinum 0.6 mm aperture grid with Struers Tripod wax. Finally, the samples were electrochemically polished until perforation by dual jet polishing in an electrolyte consisting of 5% perchloric acid and 95% methanol at -20°C and applying a voltage of 38 V.
The specimens were investigated in a JEOL 3010 transmission electron microscope operating at 300 kV. Conventional bright field, dark field and weak beam imaging techniques were applied. Frank loops were visualised with the relrod technique and cavities are revealed in under- and over-focused images. The local thickness of the sample was determined with convergent beam electron diffraction (CBED) [19].
3. Results 3.1 As-irradiated material The defect structure of the thimble tube after irradiation to 80 dpa is shown in the different images of Figure 1. The typical defects, observed in a stainless steel material irradiated to more than 1 dpa at 300°C [20], were found in this sample. Namely, the Frank loops (F), black
dots and small perfect dislocation loops (SL), are clearly visible in Figure 1a. The total defect density was found to be (10.0±0.5)×1022/m³.
(insert Figure 1 here)
Figure 1b is a dark field image obtained by using the relrod diffraction intensity which is induced by the Frank loops. In the inset, the corresponding diffraction pattern is shown. The white circle indicates the part of the diffraction pattern that was selected for the dark field image. It only includes the relrod diffraction intensity. Examination of several relrod images allows the estimation of the total amount of Frank loops and the density was found to be about (7.2±0.3)×1022/m³.
Apart from the Frank loops, round features were observed in Figure 1b as well, which can be attributed to precipitates. The properties of these precipitates agree with previous observations of γ' precipitates (Ni3Si) in neutron irradiated 304 and 316 types of stainless steel [21]. Ni3Si has a fcc crystal structure with a lattice parameter of 0.3505 nm, which is slightly smaller than the lattice parameter of AISI 316 steel. The diffraction patterns are very similar but because the lattice parameter is slightly lower, the locations of the diffraction spots are slightly further away from the transmitted beam than those from the steel. There is no orientation relation with the stainless steel and therefore part of the precipitates generate diffraction spots near the relrods and are visible in the relrod image. Since not all precipitates are visible under the applied imaging conditions, the measured precipitate density of about 0.5×1022/m³ is probably an underestimation of the actual density.
Figure 1c contains an out-of-focus image of the edge of the sample. In this image, cavities appear as white spots and a few examples are indicated with a letter V. The cavities can be either gas filled bubbles or voids. In this paper, we do not differentiate those with or without gas and in the remaining of the text we refer to them as cavities. Both have been reported to occur in stainless steel irradiated to high doses at a temperature above 300°C [20,22]. The cavity density in the as-irradiated material was measured to be (12.7±0.5)×1022/m³.
3.2 Annealing at 500°C for 6h The general radiation-induced defect structure after annealing at 500°C for 6h is shown in Figure 2. Similarly to the as-irradiated sample, the same classes of defects were found. In Figure 2a one example of a Frank loop (F), black dot (BLD), precipitate (P) and perfect dislocation loop (SL) are indicated. However, unlike the as-irradiated material, larger dislocation structures are observed, which appear to be grown dislocation loops. An example is indicated by LL in Figure 2a. No stacking fault contrast can be observed in the interior of the dislocation loop, which means that it is not a Frank loop, but a perfect dislocation loop. The total defect density, without the large loops and cavities, measured at three different locations, was about (4.9±0.5)×1022/m³.
(insert Figure 2 here)
Further, Figure 2b shows a typical dark field image using the relrod diffraction intensity induced by Frank loops. It is clear that some Frank loops remain after annealing, but their amount has significantly reduced. An average density of (1.4±0.5)×1022/m³ was measured, which is a factor of 6 lower with respect to the as-irradiated sample.
Besides the Frank loops, round contrast induced by γ' precipitates was found in Figure 2b as well. The precipitate density after annealing was found to be 0.45×1022/m3, indicating that annealing at 500°C for 6h had little or no effect on the γ' precipitates.
The presence of cavities was revealed in the out-of-focus image of Figure 2c, where they appear as white spots. The cavity density was measured to be (11.1±0.5)×1022/m³, which is comparable to the cavity density in the as-irradiated material. It suggests that annealing at 500°C for 6h did not affect the cavities.
The densities of each type of defect and their average sizes are summarized in Table 3 and Table 4, respectively. The precision on the average size, given in the table, is a measure for the width of the size distribution. Large values simply indicate that there is a significant variation in the defect size. The large dislocation loops which are the result of the annealing are difficult to quantify and were not included in the table. Compared to the density and sizes in the as-irradiated material, this anneal mainly affected the Frank loops. The density has reduced with a factor of 6 and also the average size was significantly smaller. It suggests that mainly the larger Frank loops have grown and unfaulted, whilst the smaller loops remain in the material.
3.3 Annealing at 550°C for 6h The microstructure after annealing at 550°C for 6h is shown in Figure 3. Similar to the sample annealed at 500°C for 6h a significant amount of small defects are still present. Mainly black dots and perfect dislocation loops were found. The total concentration of small defects is slightly lower than in the previous sample: a defect density of 2.8±0.8×1022/m³ was measured, not including cavities. It should be noted that perfect dislocation loops in thin TEM foils can
migrate towards the surface which may result in an underestimation of the defect density [23]. However, no migration of loops was observed during the TEM investigation. Further, it was found that several loops have grown to form large dislocation loops. The growth of the loops is affected by the small defects. Several examples can be observed in Figure 3a, where the small defects are interacting and pinning the large loops. Because of this pinning, the growth of the loops is tempered and their circular shape becomes more irregular.
(Insert Figure 3 here)
To characterize the type of defect, it was attempted to visualize the presence of Frank loops by using the relrod contrast method. Figure 3b is recorded at approximately the same location in the sample as Figure 3a, but after tilting the sample to an orientation close to the [0-11] zone axis. This orientation is optimal to reveal the relrods, but as can be seen in the diffraction pattern in the inset, no clear relrods are present. Moreover, the dark field image of Figure 3b was obtained when inserting an objective aperture at the positions indicated by the white circle in the diffraction pattern. In this way, any contribution from the matrix material to the image intensity is excluded and if Frank loops are present, they should appear as thin bright lines. The intensity in both dark field images is very low and no linear contrast features can be recognized, which means that no Frank loops are present anymore after annealing at 550°C for 6h.
Similarly, it was verified if there are still γ' precipitates (Ni3Si) present after these annealing conditions. In the previous anneal, precipitates appeared in the dark field images when selecting the relrod diffraction intensity. In Figure 3b no clear precipitates can be observed. Moreover, no additional weak diffraction spots were found in the diffraction patterns. These
observations would suggest that also the precipitates have all dissolved during annealing at 550°C for 6h. However, in the bright field image of Figure 3a a few defects were observed that show a moiré fringe pattern. A few of these defects are indicated by the letter "m". A moiré pattern in a bright field images is usually the result of the overlap of two crystal lattices with a slightly different lattice parameter or orientation. Even though, the TEM investigation is not as clear on this point, indications are found that a small amount of precipitates remains after annealing at 550°C for 6h.
The presence of cavities was verified in the out-of-focus bright field image of Figure 3c. Compared to the as-irradiated and the 500°C, 6h anneal, cavities are still present, but the density is lower after annealing at 550°C, for 6h. A density of 5.1±0.5×1022/m³ was measured, which is about half the density of the previous annealing (500°C,6h).
The defect densities and average sizes were measured and the results are given in Table 3 and Table 4, respectively. Again, the large unfaulted loops were not taken into account in the quantification. Furthermore, no Frank loops were observed anymore. The values for the precipitates are based on the observation of the moiré fringes as discussed above. Whereas the size of the defects is comparable to the annealing at 500°C for 6h, the defect density and in particular the cavity density is reduced.
3.4 Annealing at 500 °C for 19.5 h For the third annealing condition, a temperature of 500°C was maintained, but the annealing time was extended to 19.5 hours. Applying the parameters used by Busby et al. [11], the iron diffusion length is kept constant as in the case of thermal annealing at 550°C, for 6h. Therefore, if the defect evolution is controlled by the iron diffusion length, the results of the
TEM analysis should be the same as in the previous paragraph. However, this is found not to be the case.
Figure 4 shows a few typical images of the defect structure. Similar to the two previously annealed samples (500°C, 6h and 550°C, 6h), it was observed that many small loops have grown to become large perfect dislocation loops. However, at the same time, a large number of small defects remain. The total density of small defects was measured to be 4.7±0.5×1022/m³, not including cavities. This defect density is higher than in the sample annealed at 550°C for 6h and comparable to the density in the sample annealed at 500°C for 6h.
(Insert Figure 4 here)
Related to the type of defects, it was found that all types of defects detected in the sample annealed at 500°C for 6h were also found here. In Figure 4a a few examples are indicated of a large, perfect dislocation loop (LL), a small dislocation loop (SL), a black dot (BLD) and a Frank loop (F). A more clear evidence of the presence of Frank loops is given in Figure 4b, where the relrod imaging technique was used. Under these conditions, a number of parallel line segments appear in the dark field image, which can be attributed to Frank loops. The amount of Frank loops was 1.8±0.5×1022/m³. This value is of the same order as in the sample annealed at 500°C for 6h. The longer annealing time did not influence the Frank loop density.
Apart from the line segments a few round features can be recognized. Based on previous results, it is reasonable to assume that this contrast is induced by small γ'-precipitates, which were also detected in the sample annealed at 500°C for 6h. As mentioned already, it is not
possible to visualize all precipitates in a single dark field image, so it is difficult to measure their total amount. However, by comparing to the precipitate number obtained from the image in Figure 2b, which was recorded under similar conditions, the amount of precipitates is found to be clearly different. In this sample, annealed at 500°C for 19.5h, an amount of 0.15×1022/m³ of precipitates were measured. This value is a factor of 3 lower than in the sample annealed at 500°C for 6h and of the same order as the sample annealed at 550°C for 6h.
The presence of cavities was evaluated in out-of-focus bright field images as for example shown in Figure 4c. The cavity density was measured to be 10.9±0.5×1022/m³, which is of the same order as the amount of cavities after annealing at 500°C for 6h.
The defect densities of each type of defect and their average sizes are summarized in Table 3 and Table 4, respectively. The precipitate density was determined from the relrod images which also show the Frank loops. Even though it is not possible to visualize all precipitates in a single image, it was found that the precipitate density is a factor of 3 lower than in the sample annealed for 6h. Because the conditions under which these images were recorded are similar, the reduction of the amount of observable precipitates cannot be due to the imaging conditions only. Therefore it can be stated that the reduction of the amount of precipitates is significant and due to the prolonged annealing time. The defect densities of the Frank loops, all other loops and cavities are similar to the value after 6h annealing at 500°C.
Also, the average defect sizes are comparable with the sample annealed for 6h. Some Frank loops have grown to sizes above 30 nm, which slightly increases the average value and size distribution. The unfaulted loops, not including the large loops, are slightly larger. However, the size of the black dots and the cavities were not affected and, because these defects occur
predominantly, also the overall defect size was not significantly affect by the prolonged annealing time.
3.5 Annealing at 600°C for 6h The annealing of the fourth sample was performed at 600°C for 6h. These parameters give an iron diffusion length of 0.0344 cm. The general defect structure is shown in Figure 5 and is comparable to the previously annealed samples. A significant amount of small black dots was observed. Furthermore, it was found that the initially formed Frank loops have unfaulted to form perfect dislocation loops. Most of these loops have grown to large dislocation loops which interact with other dislocation loops and are pinned by the remaining black dots and small dislocation loops. The total density of small defects was measured to be 1.3±0.3×1022/m³, not including cavities. This density is the lowest of all four anneals and confirms the trend that an increased annealing temperature reduces the amount of small radiation defects.
(Insert Figure 5 here)
In Figure 5b, the TEM picture, obtained under relrod imaging conditions is presented. The corresponding diffraction pattern in the inset shows that the sample is oriented close to the [110] zone axis, which is optimal to get relrod diffraction intensity. No Frank loops were found. The complete removal of the Frank loops was expected as they were already removed after annealing at 550°C.
After the 550°C, 6h anneal, not all precipitates were annealed. However, in this sample, no traces of small γ' precipitates were found anymore. In the relrod images, clear precipitate contrast could not be observed, nor was a moiré pattern found in the bright field images.
Of particular interest is the effect of the higher annealing temperature on the cavities. Figure 5c shows a high magnification bright field image in out-of-focus condition. The cavity density was measured to be 5.1±0.2×1022/m³, which is of the same order as the amount of cavities after annealing at 550°C for 6h.
The defect densities and average sizes were measured and the results are summarized in Table 3 and Table 4. No Frank loops or precipitates were observed anymore and only the black dots, unfaulted loops and cavities could be measured. Compared to the sample annealed at 550°C for 6h, the density of the black dots and unfaulted loops is lower, but the cavity density is comparable. The average sizes of the defects are comparable.
4. Discussion The annealing conditions were selected to have three different temperatures and three different iron diffusion lengths. In the literature, the calculated iron diffusion length is used as a single parameter to compare different annealing conditions, even though it was not verified in detail if this parameter is suitable to describe the evolution of the microstructure. Moreover, at least two different parameters for calculation of the iron diffusion length can be found in the literature. Both use the same equations, but they use different values for the activation energy and iron self-diffusion coefficient (D0). Fukuya et al. [17] use a high activation energy of 2.95 eV and a larger D0 of 4.90×10-5 m²/s. Because of the high activation energy, the iron diffusion length is mainly determined by the annealing temperature. Busby et
al. [11], which was used in this report, use a much lower activation energy of 1.29 eV. Applying these parameters, the temperature is still an important factor in the calculation of the iron diffusion length, but, as shown in anneals two and three, it is possible to reach the same diffusion length at a lower annealing temperature within a reasonable annealing time.
To evaluate the effect of the annealing temperature and iron diffusion length, all defect densities and average sizes are summarized in Table 3 and Table 4. Figure 6 shows the evolution of the defect density as a function of the annealing temperature (Figure 6a) or the iron diffusion length (Figure 6b).
(Insert Table 3, Table 4 and Figure 6 here)
For all four annealing conditions that were applied for this investigation, a significant effect on the microstructure was stated. In stainless steel irradiated to high doses, a high density of Frank type loops, black dots, γ' precipitates and cavities are formed. After annealing, large dislocation loops were found in all samples, while the Frank loops, γ' precipitates and cavities were removed. This evolution of the defect structure is beneficial for the mitigation of IASCC. In the following paragraphs, a process is proposed for the evolution of the defects during annealing.
4.1 Frank loops In this work we did not characterize the nature of the loops, but according to the literature we can safely assume that neutron irradiation created both vacancy and interstitial types [24,25] under the present irradiation conditions. The growth mechanism is not yet elucidated, and it might be very different for different types of loops and sizes. For example, it is known that at
higher temperatures the vacancy clusters become unstable and start to emit vacancies to the structure [26]. These vacancies will diffuse towards the Interstitial type Frank loops and annihilate the interstitials. These Frank loops will shrink and eventually disappear. On the other hand, Frank loops of vacancy type will start to grow. Alternatively, a recent study shows that interstitial loops in pure iron can grow during annealing at about 550 – 600 °C presumably by vacancy emission [27]. With increasing loop size, also the stacking fault area within the loop grows and it becomes energetically favourable to unfault and to form perfect dislocation loops. Unfaulting of a Frank loops occurs most probably by the nucleation of a Shockley partial within the loop, induced by thermal activation [28,29]. We find that the loops, once they are of certain size, mutually interact and form large dislocation loops. During growth, the loop can interact with remaining small defects like the black dots and be pinned. These pinning effects explain why the large loops have an irregular shape.
All annealing conditions applied in this study induce a decrease of the number of Frank loops and annealing at temperatures of 550°C or higher even completely removed all Frank loops. Compared to the results in the AISI 304 steel [15], a higher annealing temperature is required in 316CW steel to obtain the same effect. In the 304 steel, all Frank loops were removed after annealing at 500°C, while it was shown here that annealing needed to be performed at 550°C in 316CW steel.
The removal of the Frank loops is controlled by the annealing temperature and not by the iron diffusion length. The samples annealed at 500°C for 19.5h and at 550°C for 6h have the same iron diffusion length, but the first sample still contains Frank loops, while the second sample does not. Moreover, the number of Frank loops after annealing at 500°C for 19.5h is comparable to the number after annealing at 500°C for 6h.
This evolution is visually represented in the graphs in Figure 6. In the graph of Figure 6a, there are two data points at an annealing temperature of 500°C which agree within the experimental error. In the graph of Figure 6b, there are two data points at an iron diffusion length of 0.00205 cm. Here the difference in the measured defect density is larger than the experimental error. It means that the removal of the Frank loops is controlled by the annealing temperature and not by the iron diffusion length. This observation supports the loop growth mechanism based on vacancies [26]. Moreover, our finding is in good agreement molecular dynamics calculation [30,31] which show that dissolution of He-V clusters occurs at about 750K (477°C) and consistent with annealing behaviour experimentally observed for cavities, as discussed in the next paragraph.
4.2 Cavities In a similar way, the graphs of Figure 6 show that also the evolution of the cavities is rather controlled by the annealing temperature than by the iron diffusion length. Annealing at 500°C results in a slight reduction of the cavity density compared to the as-irradiated sample. The reduction in density is still moderate. It is important to note that the longer annealing time did not influence the cavity density. The average size of the cavities is not affected by the annealing. All obtained sizes are equal within the precision of the measurement.
When annealing at 550°C, a significant decrease of the cavity density was stated. Only halve of the cavities remain. This observation shows again that also for the cavities, the annealing temperature is the controlling factor and not the iron diffusion length.
However, annealing at 600°C did not further decrease the cavity density. This may be related to differences in thermal stability of different cavities and in particular between He filled
bubbles and voids. Molecular dynamics simulations and first principles calculations [30,31] showed that the binding energy of vacancies in clusters increase with increasing number of He impurity atoms. Moreover, it is suggested that HeV clusters dissociate around 750K (477°C) while higher stability clusters dissociate at 800 – 1000K (527 - 727°C) [31]. Cole and Allen [32] noted that during annealing of high temperature irradiated 304 type stainless steel, part of the voids shrunk while other voids were not affected. They also observed the creation of new, small bubbles during annealing, which they related to the trapping of vacancies at small He clusters. Applying these statements to the present experiments, suggests that the voids emitted vacancies during annealing and were annihilated completely at 550°C, while the cavities that remained are He filled gas bubbles. The removal of the voids is responsible for the reduction of the cavity density.
The evolution of the cavities is in agreement with the evolution of the Frank loops. At 500°C only the vacancy clusters and part of the voids were annealed, which did not generate sufficient vacancies to remove all Frank loops. More vacancies became available at higher annealing temperatures, which initiated the complete removal of all Frank loops.
4.3 Precipitates The effect of annealing on the precipitates is extremely important and interesting. Previous studies showed that after the post-irradiation annealing at 500°C of AISI 304 the precipitates are no longer visible by TEM [15]. Subsequent atom probe tomography on the same samples [16] indicated that precipitates do not dissolve completely during annealing, but that the Ni and Si slowly diffuse giving a broadening of the precipitate and a gradual decrease of the concentration of the Ni and Si. The combination of TEM and APT results suggested that the precipitate does not dissolve completely, since the precipitate remains in some form of Ni-Si
cluster but without a full crystallographic structure. These clusters no longer provide diffraction contrast in the TEM images. For these defects, diffusion plays an important role in their evolution, so the iron diffusion length is considered as the relevant parameter. Additionally, the formation mechanism of γ' precipitates is governed by radiation induced segregation of Ni and Si towards sinks [20], so it is reasonable to expect that dissolution is controlled by diffusion of iron and other depleted elements towards the precipitates.
The results obtained in this study confirm that the iron diffusion length correctly describes the evolution of the precipitates. Annealing at 500°C for 6h reduces the amount of precipitates to about one halve. A longer annealing time further reduces the precipitate density. Moreover, it was found that in the sample annealed at 550°C for 6h, which has the same iron diffusion length as the 500°C, 19.5h annealed sample, also has a similar precipitate density. Annealing at 600°C for 6h, which has the longest iron diffusion length, completely removed the precipitates.
These findings are also represented in Figure 6. A significant difference in precipitate density after the two anneals at 500°C but with different annealing times is clearly observed. If the two annealing conditions with the same iron diffusion length are compared, similar densities were measured.
5. Conclusions 316CW stainless steel samples, obtained from a thimble tube irradiated to 80 dpa in the Tihange 2 reactor, were annealed after irradiation, applying four different conditions. The annealing times and temperatures were selected to have three different annealing temperatures and three different iron diffusion lengths. All four samples were investigated with TEM to
reveal the changes in microstructure as a result of the annealing. In this work the thermal stability of various types of defects created by high dose neutron irradiation was discussed.
It was found in all samples that part of the radiation-induced Frank loops grew, were unfaulted and formed large perfect dislocation loops. The growing loops are pinned by the remaining radiation-induced defects. After annealing at 500°C, a reduced number of Frank loops were still present, regardless of the annealing time of 6h or 19.5h. Annealing at 550°C or higher completely removed all Frank loops. The required annealing temperature to remove defects is about 50°C higher than in AISI304 steel.
The cavity density and size was hardly affected by the annealing at 500°C. However, when annealing at 550°C, the density was halved. Annealing at 600°C did not further reduce the cavity density. For both the Frank loops and the cavities, the results show that the annealing temperature is the qualifying parameter, while the iron diffusion length does not give an accurate description.
The results related to the thermal stability of the γ' precipitates confirm that the iron diffusion length correctly describes their evolution. It was found that a similar precipitation density exists after annealing conditions having the same iron diffusion length. This is found to be in clear contrast to the behaviour of all other defects such as Frank loops, cavities and black dots. Annealing at 600°C, which has the longest iron diffusion length, completely removed the precipitates.
The evolution of the defect structure during post-irradiation annealing and, in particular, the removal of a large fraction of the radiation induced defects is beneficial for the mitigation of
IASCC. Therefore PIA is relevant for long term operation of internal components and nuclear power plant life time extension.
Acknowledgements This work was financially supported by the PERFORM60 project, as a part of the 7th EURATOM FRAMEWORK PROGRAMME 2007-2013 (FP7-232612-PERFORM60). The authors would like to thank GDF-Suez Tractebel for delivery and use of the highly irradiated thimble tube, Kris Kaers for his indispensable help with sample preparation and Dmitry Terentyev for the useful discussions.
References [1] S.J. Zinkle, G.S. Was., Acta Materialia 61 (2013) 735-758 [2] P.M. Scott, J. Nucl. Mat. 211 (1994) 101-122. [3] P.L. Andresen, F.P. Ford, S.M. Murphy, J.M. Perks, in: Proceedings of 4th International Symposium on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors, NACE International, Houston, 1990, pp. 1-83–1-121. [4] G.S. Was, P.L. Andresen, Corrosion, 61 (2007) 19-45. [5] G.S. Was, J.T. Busby, Phil. Mag. 85 (2005) 443-465. [6] T. Onchi, K. Dohi, N. Soneda, M. Navas, M.L. Castaño, J. Nucl. Mater. 340 (2005) 219236. [7] L. Fournier, M. Savoie, D. Delafosse, J. Nucl. Mater. 366 (2007) 187-197. [8] Z. Jiao, J.T. Busby, G.S. Was, J. Nucl. Mater. 361 (2007) 218-227. [9] Z. Jiao, G.S. Was, J. Nucl. Mater. 382 (2008) 203-209.
[10] R. Yang, B. Cheng, J. Deshon, K. Edsinger and O. Ozer, J. Nucl. Sci. & Tech. 43, 9 (2006) 951-959. [11] J. T. Busby, G. S. Was, E. A. Kenik, J. Nucl. Mater. 302 (2002) 20-40. [12] Y. Ishiyama, M . Kodama, N. Yokota, K. Asano, T. Kato, K. Fukuya, J. Nucl. Mater. 239 (1996) 90-94. [13] F. Carsughi, H. Ullmaier, H. Trinkaus, W. Kesternich, V. Zell, J. Nucl. Mater 212-215 (1994) 336-340. [14] V. Zell, H. Trinkaus, H. Schroeder, J. Nucl. Mater. 212-215 (1994) 320-324 [15] W. Van Renterghem, A. Al Mazouzi, S. Van Dyck, J. Nucl. Mater. 413 (2011) 95-102. [16] T. Toyama, Y. Nozawa, W. Van Renterghem, Y. Matsukawa, M. Hatakeyama, Y. Nagai, A. Al Mazouzi, S. Van Dyck, J. Nucl. Mater. 418 (2011) 62-68. [17] K. Fukuya, M. Nakano, K. Fujii, T. Torimaru, Y. Kitsunai, J. Nucl. Sci. Tech. 41 (2004) 1218-1227. [18] R. Gerard, GDF Suez, private communication [19] P.M. Kelly, P.A. Jostons, R.G. Blake, J.G. Napier, Phys. Stat. Sol. A31 (1975) 771-780. [20] S.M. Bruemmer, E.P. Simonen, P.M. Scott, P.L. Andresen, G.S. Was, J.L. Nelson, J. Nucl. Mater 274 (1999) 299-314. [21] P.J. Maziasz, J. Nucl Mater. 205 (1993) 118-145. [22] F.A. Garner, I. Porollo, Yu. V. Konobeev, O.P. Maksimkin, Proceedings of the 12th International Conference on Environmental Degradation of Materials in Nuclear Power System – Water Reactors, TMS, (2005) 439-448. [23] M.R. Gilbert, Z. Yao, M.A. Kirk, M.L. Jenkins, S.L. Dudarov, J. Nucl. Mater. 386-388 (2009) 36-40. [24] D.J. Edwards, E.P. Simonen, S.M. Bruemmer, J. Nucl. Mater. 317 (2003) 13-31 [25] R. Stoenescu, R. Schäublin, D. Gavillet, N. Baluc, J. Nucl. Mater. 360 (2007) 186-195. [26] S.J. Zinkle, P.J. Maziasz, R.E. Stoller, J. Nucl. Mater. 206 (1993) 266-286.
[27] S. Moll, T. Jourdan, H. Lefaix-Jeuland, Phys. Rev. Lett 111,1 (2013) 015503. [28] E.H. Lee, M.H. Yoo, T.S. Byun, J.D. Hunn, K. Frarrell, L.K. Mansur, Acta Mater. 49 (2001) 3277-3287. [29] A. Bakaev, D. Terentyev, X. He, E.E. Zhurkin, Nucl. Instr. And Meth. in Phys. Res. B 303 (2013) 33-36 [30] K. Morishita, R. Sugano, B.D. Wirth, T. Diaz de la Rubia, Nucl. Instr. and Meth. In Phys. Res. B, 202 (2003) 76-81. [31] C.-C. Fu, F. Willaime, J. Nucl. Mater. 367-370 (2007) 244-250. [32] J.I. Cole, T.R. Allen, J; Nucl. Mater. 283-287 (2000) 329-333.
Figure Captions Figure 1. a) Bright field image of the as-irradiated thimble tube, containing Frank loops (F), black dots (BLD) and small loops (SL). b) Relrod image of the Frank loops (F) and a fraction of the γ' precipitates (P). The diffraction pattern and the position of the objective aperture (white circle) are given in the inset. c) Out-of-focus image to visualize cavities (V).
Figure 2. a) Bright field image of the defect structure after annealing at 500°C for 6h. Examples of precipitates (P), Frank loops (F), black dots (BLD), small loops (SL) and large loops (LL) are indicated. b) Relrod image of the remaining Frank loops and γ' precipitates. The inset shows the diffraction pattern and the white circle indicates the location of the objective aperture. c) Out-of-focus images to visualize cavities (V).
Figure 3. a) Bright field image of the defect structure after annealing at 550°C for 6h. The moiré fringes (m) are induced by precipitates. b) Dark field image using relrod contrast. The inset shows the diffraction pattern and the white circle indicates the location of the objective aperture. c) Out-of-focus image to visualize cavities (V).
Figure 4. a) Bright field image of the defect structure after annealing at 500°C for 19.5h, containing Frank loops (F), black dots (BLD), small loops (SL) and large loops (LL). b) Relrod image of the remaining Frank loops (F) and a fraction of the γ' precipitates (P). The inset shows the diffraction pattern and the white circle indicates the location of the objective aperture. c) Out-of-focus image to visualize cavities (V).
Figure 5. a) Bright field image of the defect structure after annealing at 600°C for 6h. Examples of Large dislocation loops (LL), small unfaulted loops (SL) and black dots are indicated. b) Dark field image recorded under relrod conditions.
The inset shows the
diffraction pattern and the white circle indicates the position of the objective aperture. c) Outof-focus image revealing the presence of cavities (V).
Figure 6. Evolution of the defect density of the Frank loops, cavities, all loops and all defects(left Y-axis) and the precipitates (right Y-axis) a) as a function of the annealing temperature and b) as a function of the iron diffusion length. (in colour on web, grey scale in printed copy)
Tables Table 1. Composition of the 316CW stainless steel. C Si Mn P S
Cr
Ni
Mo
Co
0.044
17.40
12.8
2.68
0.07
0.53
1.79
0.022
0.009
Table 2. Annealing parameters of the different specimens. Segment anneal T (°C) anneal time (h)
Fe diffusion length (cm)
quarter 1
500
6
0.00114
quarter 2
550
6
0.00204
quarter 3
500
19.5
0.00205
quarter 4
600
6
0.00344
Table 3. Defect densities of all defects before and after annealing. Loops and precipitates Density All loops (1022/m³) precip. Frank BLD unf.1) & precip.
cavities
All
As irr.
0.5 2)
7.2 ± 0.3
-
-
10.0 ± 0.5
12.7 ± 0.5
22.7 ± 0.7
500°C, 6h
0.45 2)
1.4±0.5
- 3)
- 3)
4.9±0.5
11.1 ± 0.5
16.0 ± 0.7
550°C, 6h
0.16 2)
-
2.8±0.8
0.56±0.2
3.5±0.9
5.1 ± 0.5
8.6 ± 1.0
500°C, 19.5h
0.15 2)
1.8±0.5
- 3)
- 3)
4.7±0.5
10.9 ± 0.5
15.6 ± 0.7
600°C, 6h
-
-
1.1±0.5
0.2±0.1
1.3±0.5
5.1 ± 0.3
6.4 ± 0.6
1) 2)
The unfaulted loops do not include the large loops formed during annealing. underestimation as discussed above
3)
not measured individually
Table 4. Average sizes of all defects before and after annealing Loops and precipitates Size (nm)
cavities
All
All loops precip.
Frank
BLD
unf. & precip.
As-irr.
5.4±1.4 8.9 ± 4.0
-
-
8.7±3.8
1.7 ± 0.4
4.8±1.7
500°C, 6h
5.0±1.2
6.9±3.1
4.3±1.4
9.9±3.0
5.7±2.1
1.9 ±0.4
3.1±1.0
550°C, 6h
6.4±1.2
-
4.9±1.7
11.2±3.8
5.9±3.0
1.7±0.3
3.4±0.7
500°C, 19h
5.7±1.2
8.2±6.8
4.1±1.1
12.4±4.4
6.3±4.7
1.8 ±0.3
3.2±1.6
600°C, 6h
-
-
4.3±1.1
9.8±3.4
5.1±1.4
1.7±0.3
2.4±0.4