High-heat flux Cu-0.8Y alloys investigated by positron annihilation spectroscopy

Contents lists available at ScienceDirect Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jalcom High-heat flux Cu-0.8Y alloys investigated by positron annihilation spectroscopy R. Domínguez-Reyes a,⁎ , M.A. Monge a , B. Galiana a , Y. Ortega b , A. Muñoz a , G. Carro-Sevillano c a Departamento de Física. Universidad Carlos III de Madrid, 28911 Leganés, Madrid, Spain b Departamento de Física de Materiales. Facultad de Ciencias Físicas. Universidad Complutense de Madrid, 28040 Madrid, Spain c Centro Nacional de Investigaciones Metalúrgicas (CENIM-CSIC), Departamento de Metalurgia Física, Avd. Gregorio del Amo 8, 28040 Madrid, Spain article info Article history: Received 20 September 2021 Received in revised form 24 November 2021 Accepted 22 December 2021 Available online 24 December 2021 Keywords: Copper alloys Oxide dispersion strengthened Positron annihilation High-heat flux materials Equal channel angular pressing Reinforced materials abstract This work studies the thermal stability of the microstructure and the evolution of the defects of two highheat flux Cu-0.8 wt%Y alloys fabricated following two alternative powder metallurgy routes. One batch was produced by direct hot isostatic pressing (HIP) consolidation of Cu-0.8 wt%Y pre-alloyed atomized powders while an additional ball milling processing step was introduced before HIP sintering for the second alloy. The stability and recovery characteristics of the vacancy type defects in these alloys in the as-produced state and after processing by severe equal channel angular pressing to achieve a refine microstructure have been investigated by positron lifetime and coincidence Doppler broadening measurements in samples subjected to isochronal annealing from room temperature to 900 °C. Microhardness measurements and electron transmission microscopy analysis have also been performed to support the results obtained from the positron annihilation spectroscopy analysis techniques. The recovery curves of the positron lifetime and S-W plots show a recovery stage in agreement with the recovery stage V for Cu. However, a full recovery is not accomplished, and a stage that reverts the previous recovery takes place after annealing above ~600 °C, that leads to the formation of very stable defects at temperatures up to 900 °C, identified as vacancy aggregates and nanocavities. The characteristic shape of the coincidence Doppler broadening indicates that the dispersed Y-O particles in the Cu matrix appear to be responsible for stabilizing the vacancy aggregates and nanocavities for temperatures above 600–700 °C. © 2021 The Author(s). Published by Elsevier B.V. CC_BY_NC_ND_4.0 1. Introduction Dispersion strengthened (DS) and precipitation hardened (PH) Cu alloys are primary candidate materials for heat sink applications in the plasma facing components (PFCs) of a near-term DEMO fusion reactor due to their high thermal conductivity, enhanced mechanical strength and irradiation resistance [1–4]. They have to be thermally stable and radiation, creep and fatigue resistant at elevated temperatures, under the severe working conditions required for DEMO; such as heavy thermal loads near the upper working temperature of the materials, repeated stepwise changes in temperature and heavy high-energy particle irradiation [5]. PH CuCrZr and DS Cu-Al 2 O 3 (GlidCop) are presently the baseline high-heat flux (HHF) materials in the water-cooled PFCs designs for the near-term DEMO as there exists an important database that has allowed to predict their performances under the ITER operation conditions [6,7]. However, PH CuCrZr at temperatures above 400 °C is susceptible of thermal softening due to over ageing and suffers irradiation embrittlement at temperatures below 200 °C [1,2,7,8]. Moreover, DS Cu-Al 2 O 3 and PH CuCrZr exhibit significant thermal creep at temperatures above ~ 350 °C for applied stresses above 60 MPa [8,9]. These HHF materials having critical limits in their working temperatures, i.e. 200 – 350 °C, would lower the overall efficiency of the reactor [10]. The formation of stable stacking faults, voids or bubbles in Cu-based materials is very dependent on multiple factors: alloying elements, presence of reinforcing particles, the irradiation temperature or the presence of thermal cycling, between others [3,4,11–14]. For example, stacking fault tetrahedra (STF) are commonly observed in pure copper, while the presence of dispersed particles, the irradiation at high temperatures or the thermal cycling tend to suppress its formation [3,13]. To develop new high-strength and heat-resistant Cu alloys with a more favorable working temperature window, DS Cu-0.8 wt%Y alloys have been produced by different powder metallurgy routes and their microstructural, thermal and mechanical characteristics have been https://doi.org/10.1016/j.jallcom.2021.163430 0925-8388/© 2021 The Author(s). Published by Elsevier B.V. CC_BY_NC_ND_4.0 ]]]] ]]]]]] ⁎ Corresponding author. E-mail address: [email protected] (R. Domínguez-Reyes). Journal of Alloys and Compounds 900 (2022) 163430 investigated [15–17]. The results have demonstrated the feasibility of strengthening Cu by dispersion of Y-rich particles maintaining a high thermal conductivity. These alloys as heat sink materials to be joined to PFCs could be exposed at temperatures as high as ~900 °C during the joining operation changing their microstructure and degrading the mechanical properties. However, materials produced through a powder metallurgical route usually have a higher number density of defects than the observed for similar materials produced by traditional techniques such as casting. The presence of a significant concentration of light impurities, point defects, interstitial-vacancy complexes or voids are commonly observed and have a strong effect in the properties of the material; for example, the presence of vacancy aggregates and voids induce material hardening, loss of ductility and embrittlement [18,19]. The objective of the present study is to investigate the presence and stability of the open volume defects in the studied DS Cu-Y alloys, their stability with the temperature and the interaction of these defects with the dispersoids, impurities and grain boundaries. Isochronal annealing and positron annihilation experiments have been accomplished in the DS-0.8Y alloy processed under different conditions, supported by transmission electron microscopy (TEM) observations in selected samples, to investigate the initial observed recovery up to intermediate temperatures and subsequent formation of voids on increasing temperature. 2. Experimental procedure DS Cu-Y alloys were produced using powder with a nominal composition Cu-0.8 wt%Y and particle size < 120 µm produced by vacuum induction melting and subsequent atomization in high purity Ar gas. The purity of the initial raw Cu and Y materials was 99.9% and 99.8%, respectively. The atomized powder was consolidated by HIP and the obtained alloy is labeled as CuY. To homogenize and refine the initial microstructure of the atomized powder, a second ingot was produced by milling the initial powder in a high energy planetary ball mill during 40 h at 100 rpm under an Ar atmosphere. In this case, the alloy resulting after HIP consolidation was labeled as CuYM. The milling media were Cr-steel balls with a ball/powder ratio of 5:1. For both materials, the HIP consolidation was performed for 2 h at 1123 K and 179 MPa. Previously, the powder was encapsulated into a steel can that was tightly sealed after degassing at 573 K for 24 h in vacuum. Billets of CuY and CuYM with dimensions of 10 × 10 × 70 were deformed by equal channel angular pressing (ECAP) at 623 K via B C route, i.e. the billet is rotated + 90° around its longitudinal axis between successive passes. The CuY billet was submitted to 8 passes and the CuYM one to 4 passes, and were denoted as CuYE and CuYME, respectively. Density measurements were carried out in a He ultrapycnometer. The oxygen content was determined in an infrared absorption spectrometer. The microstructure of the alloys was analyzed by ultra-high-resolution scanning electron using a TENEO (FEI) microscope. The processing details and the microstructural, mechanical and thermal properties of these alloys have been reported elsewhere [15–17]. The Positron Annihilation Spectroscopy (PAS) experiments were performed on samples of four materials after being isochronal annealed for 20 min up to 900 °C in a vacuum of < 10 −5 mbar. Complementary experiments were also performed on samples of pure Cu in different conditions. Positron lifetime measurements were carried out in a fast-fast coincidence spectrometer with a time resolution of 230 ps (FWHM) using a 22 Na source sealed in kapton sandwiched between a pair of identical samples. Positron lifetime spectra with a total count number > 10 6 were recorded and analyzed using the PATFIT-88 [20]. The instrumental time resolution and contribution of the positron source to the lifetime spectra were determined from measurements on reference samples of annealed pure Fe and Si single crystals. The source contribution was assessed in a component of 382 ps with an intensity of 13.6% and another very weak of 1.2 ns and 0.2%. The coincidence Doppler broadening (CDB) measurements were done using two HP Ge detectors in timing coincidence, placed face to face with the samples at the half-way point between the detectors. The resulting CDB spectrum was the cumulative spectrum of 24 CDB spectra without evidence of electronic shift, each of them with a count number > 10 6 in a 512 × 512 coincidence matrix. The cumulative spectra had 1 × 10 7 counts in the strip centered on the matrix diagonal for the energy range 2m o c 2 -1.6 keV < E 1 + E 2 < 2m o c 2 + 1.6 keV; E 1 and E 2 stand for the energies of the pair of annihilating photons, m o the electron rest mass and c the light speed. The spectra were binned from 40 to 512 bins with a bin width of 2.5 × 10 −3 m o c to decrease the statistical fluctuations of the data in the high momentum region. The spectra were normalized, and the intensity at a given photon momentum divided by the corresponding counts in the CDB spectrum of a reference sample of pure Cu to highlight the differences between the spectra. 99.999% pure Cu annealed at 900 °C and Y 2 O 3 single crystals were used as reference samples. To determine the experimental error bars of CDB curves, it has been considered the mean deviation between independently obtained CDB curves for the same samples, since it is representative of the consistency of the measurements. The fraction of positrons annihilated with lowand high-momentum electrons was also estimated using the S and W line shape parameters. The S parameter was calculated as the ratio of the count number in the energy window of 1.946 keV centered at 511 keV ( p m c0 7(10 ) L30 ) to the count total in the annihilation peak. The W parameter was calculated as the fraction of counts in the energy intervals 515.66 – 519.72 keV and 502.67–506.74 keV ( p m c16 32(10 ) L30 ). Samples at selected annealing temperatures have been also characterized by means of TEM to observe the open volume defects. For that purpose, the TEM images have been achieved focused and out of focus. Samples were examined in a Philips Tecnai F20 operated at 200 kV. Cross-sectional samples were prepared for electron transparency using a focused ion beam (FIB), protecting the sample surface through pre-deposition of a Pt layer, and milling with Ga + ions close to normal incidence. 3. Results 3.1. General description of the microstructure The four studied materials present a similar fully dense microstructure consisting of a copper matrix of equiaxed grains reinforced with a dispersion micrometric and sub-micrometric particles of Y rich oxides located at the grain boundaries and inside the grains. The main differences between them are: i. The mean grain size of CuY and CuYM after HIP consolidation is micrometric, (3.1 ± 1.6) µm and (2.4 ± 1.6) µm respectively, while after ECAP processing a reduction of the mean grain size of (0.5 ± 0.2) μm for CuYE and CuYME is achieved. ii. As a result of the milling process prior to HIP consolidation, the CuYM has a more homogeneous and a finer distribution of the reinforcing yttrium-enriched particles than CuY. The ECAP processing produced two materials with similar microstructure characterized by a refinement of the grains and reinforcing particles. The observed Y-O particles are very stable with the temperature and no significant changes on their size or composition have been observed in all the range of the annealing temperatures. iii. The presence of a significant number of twins in the CuYM is observed. However, the ECAP processing drastically reduced the density of twins. R. Domínguez-Reyes, M.A. Monge, B. Galiana et al. Journal of Alloys and Compounds 900 (2022) 163430 2 iv. Cavities or microvoids were not observed in any of the materials after their production process. v. ECAP B C processing route is the best for achieving grain refinement [21,22]. The specific geometry of the ECAP channel used in this work introduces a homogeneous shear strain per pass of 2.03 that results in an equivalent strain of ε = 1.17 per pass [21,22]. vi. CuY and CuYM do not exhibit significant changes in their hardness after the annealing treatments; however, CuYE and CuYME exhibit an important reduction of the hardness for temperatures higher than 300 °C and decrease to similar values to CuY and CuYM for temperatures ≥ 600 °C, respectively (see Fig. 1). A detailed description of the microstructure and mechanical properties of these materials after consolidation by HIP and different thermo-mechanical treatments, including ECAP, has been already published [15–17]. As reference, Fig. 2 presents the main characteristics of the CuYM microstructure. A bright-field TEM micrograph showing a general view of the as-HIP material, consisting of equiaxed grains, is depicted in Fig. 2(a). Fig. 2(b) and the subfigures (b 1 ), (b 2 ) and (b 3 ) correspond to a STEM image and the corresponding EDS elemental maps of the Y-O reinforcing particles. The presence of dislocation tangles is observed in some grains, as it is shown in the subfigure 1(a). The elemental composition of the different phases present in the CuYM was established by numerous EDX measurements. As can be observed in the two examples shown in Fig. 2(c), the reinforcing particles are composed of Y and O while the matrix is pure Cu. 3.2. Positron lifetime measurements The lifetime spectra are characterized by a single lifetime value, τ, the evolution of which with annealing temperature is summarized in Fig. 3. Neither of the measured lifetime spectra could be consistently fitted to a sum of two or more lifetime components even if a longlived lifetime value is constrained to any of the expected positron lifetime for defects in Cu (see Table 1). The unmilled CuY in the as-HIP condition exhibits a pronounced recovery stage of the measured mean positron lifetime starting at ≈ 200 °C and finishing at ≅ 600 °C, with a reduction of the measured positron lifetime from a value of 172 ps to a minimum of 133 ps (see Fig. 3(a)). This recovery is followed by a stage that reverts it that finishes after annealing at T ≥ 700 °C, reaching a value of 157 ps that remains constant up to the maximum annealing temperature of 900 °C. As shown in Fig. 3(a), the mean lifetime values obtained for the isochronal annealing process of the CuYM exhibits a similar behavior with the curve, shifted to higher values. The initial τ value before recovering at 200 °C is roughly 180 ps, decreasing to a minimum of 168 ps at 600 °C to increase to a steady value of ≈ 178 ps for temperatures higher than 700 °C. Fig. 1. Microhardness measurements obtained at RT as a function of the isochronal annealing temperature. Fig. 2. Main characteristics of the microstructure of the CuYM. (a) BF-TEM image showing a general view of the equiaxed grain microstructure of CuYM. The sub-figure (a) reveals the presence of dislocations tangles. (b) STEM image and EDS elemental maps of the marked region showing the Y-O reinforcing particles: (b1) HAADF-STEM image, (b2) O-K and (b3) Y-K. (c) EDX measurements at the points indicated by arrows in (b) revealed that the reinforcing particles are composed of Y and O, and the matrix is pure Cu. White arrow points to numerous nanometric Y-rich particles present in the Cu matrix. R. Domínguez-Reyes, M.A. Monge, B. Galiana et al. Journal of Alloys and Compounds 900 (2022) 163430 3 For both materials, the evolution of the values with the annealing temperature after ECAP processing is shown in Fig. 3(b) and presents a similar overall shape to the observed for the unprocessed materials CuY and CuYM with a lower variation range of τ. The CuYE initiates the recovery at higher temperature than the CuY, at ≈ 350 °C, and also finishes at a higher temperature, ≈ 700 °C. After a short process where the recovery is reverted, the value of τ increases to a constant value of ≈ 162 ps for temperatures ≥ 800 °C, similar to the observed one for CuY at high temperatures. The samples of the CuYM material in the as-HIP or ECAP processed states have very similar values in all the temperature range of the isochronal annealing treatments (see Fig. 3). The only significant differences are higher values of τ for all the annealing temperatures and a lower slope at the state after reverting the recovery on increasing temperature. 3.3. CDB measurements The ratio CDB curves of the materials and their evolution with isochronal annealing are shown in Fig. 4 and Fig. 5 for selected annealing temperatures. As reference, the ratio CBD curves of a high purity samples of Y 2 O 3 (using 99.99% purity nanopowder compacted before and after vacuum sintering at 1400 °C) and a well annealed elemental Y (99.98% purity at 1000 °C) are also depicted in Fig. 6. Both reference CDB curves present the higher values at low momentum, p0 L of all measured CDB spectra and a characteristic structure at the intermediate momentum region, p m c5 30(10 ), L30 showing the Y 2 O 3 distinctive peaks that are clear revealed in CDB Y2O3 /CDB Y normalized spectrum (see Fig. 6). Its characteristic shape provides the characteristic signature of positron annihilation with Y in copper or oxygen anions in Y 2 O 3 . The trapping and subsequent annihilation of positrons with this Y-O oxides or Y aggregates is expected due to the very high positron affinity of Y and oxygen anions in the Y-O oxides [23,24], which are the highest compared to Cu or any other elements present in the studied materials including light impurities like C, or Cr and Fe from the grinding media [25]. The CDB Y2O3 /CDB Y ratio curve is especially interesting because yttria has been poorly studied with positrons techniques and presents a complex crystalline structure, having a less ionic character than other similar oxides, which may result in a CDB spectrum with a significantly different shape that the reported for other widely studied oxides, such as Al 2 O 3 , MgO or SiO 2 [26–29] and references therein. It highlights the characteristics of the momentum distribution of the oxygen anions core electrons of the Y oxide, that are significantly different to the found for other similar oxides. As it is observed in Fig. 4(a) for CuY, the low momentum peak at p0 L decreases on increasing temperature up to the annealing temperature of 600 °C where the minimum value is reached, with the corresponding increase of the CDB curves at the intermediate momentum region. After annealing at 700 °C, the low momentum peak increases, with a corresponding decrease in the intermediate momentum region, remaining the CDB curves similar for higher annealing temperatures (Fig. 4(b)). A similar but less pronounced behavior is observed for CuYM in Fig. 5(a) and (b). Finally, the CDB curves for the ECAP processed materials exhibit the same overall shape and annealing behavior than the one corresponding to their respective materials in the as-HIP condition (see Fig. 4(c) and (d), and Fig. 5(c) and (d) for the CuYE and CuYME, respectively). Figs. 7 and 8 show the correlation plots of the W values versus S calculated for the different annealing temperatures for all the materials. Relative values for W and S parameters were used (instead of the directly calculated values) for a better comparison. Well annealed 99.999% pure Cu was used as reference, so its W and S parameter values were used to normalize all the studied sample parameters. The value obtained from heavily deformed cold-work 99.999% pure Cu is shown as reference. These plots are useful for revealing the presence of different positron traps or changes in the chemical surrounding of the positron annihilation sites during isochronal annealing. In the case of positron annihilation taking place in two positron states, i.e. in a free state and in a trapped state in a single type of defect, or otherwise in two trapped state in two different types of defect, it is known that the measured line shape parameters S and W are given by Fig. 3. Positron annihilation spectroscopy measured lifetimes as a function of annealing temperature for: (a) CuY and CuYM after HIP consolidation, (b) CuYE and CuYME after ECAP processing. Table 1 Experimental, Exp , and theoretical, Th , values of the positron in pure Cu and Cu defects. The presented values correspond to results that have been confirmed by several works. Some theoretical predictions derived from different framework calculations are also presented. References to some representative literature works are included. The Th values range from divancancies up to aggregates of 40 vacancies (*). Exp (ps) Reference Th (ps) Reference Pure Cu 106 122 [32–34] 108 113 [34,35] Dislocation 122 [34] monovacancy 158 166 [32,34] 166 [34,35] Dislocation + vacancy 166 176 [36] 179 [34,37] Divacancy 180 190 [34,36,38] Vacancy aggregates/ voids 340 500 [38,39] 180 400 (*) [34] R. Domínguez-Reyes, M.A. Monge, B. Galiana et al. Journal of Alloys and Compounds 900 (2022) 163430 4 ( ) ( ) S f S f S W f W f W 1 1 2122 2122 = + = + (2) where f 2 is the fraction of annihilations in the state 2, Si stands for the line shape parameters representative of the annihilations with the low momentum valence electrons for each state, and W i the corresponding with the high momentum electrons. From Eq. 2 the following correlation is found W RS W RS( ) 2 2 = + + (3) where R is a constant defect-specific parameter independent of the number density of positron traps given by [30]. RW W S S W W S S 1 2 2 1 2 2 = = (4) Then, if the W-S plot exhibits a linear dependence, either a single type of defect acts as positron traps or there exists positron trapping saturation in two effective defects, or positron traps with similar positron affinity, through the recovery and isochronal annealing. In most of the annealing conditions, positron trapping saturation at open volume defects appears to be the case for either alloy in the asHIP condition or after being ECAP deformed. Nevertheless, the apparent linear dependence of the W-S plots may be misleading and requires a careful analysis of the variations of consecutive points in the W-S plots with their isochronal annealing temperature, as will be discussed below. 3.4. TEM measurements The presence of nanovoids was studied by bright-field TEM using a through-focal technique, being the under/over focus distance of 0.5 µm. The observations were carried out on samples CuY in the as-HIP state, and in CuYE samples after the annealing treatment at 600 °C and 900 °C. The presence of nanocavities in the CuY and CuYE was not observed after HIP consolidation. Fig. 11 gathers the set of measurements to different focal lengths. In the sample annealed at 600 °C, the observed nanovoids are mainly located on grain boundaries; while in the annealed sample at 900 °C they appear both randomly dispersed in the matrix and on the grain boundaries. 4. Discussion The interpretation of the positron lifetime and CDB experiments in the studied materials is not straightforward due to the impossibility to resolve the positron lifetime spectra into several components irrespective of the temperature of the annealing treatment or the initial microstructure of the materials (two-component fit provided non-reliable results). This suggests the presence of a mixed set of annihilation sites with similar characteristic positron lifetime that compete as positron traps. The trapping saturation at a unique type of defect is discarded due to the observed changes of the measured positron lifetime and the S-W plots with the temperature [31] (see Figs. 3 and 7), therefore the single lifetime component will also be considered as an average lifetime, since it includes the annihilation in several types of positron traps and several components were not separated. The variation observed in the S-W plots discards positron trapping saturation in a single positron traps because the value of both parameters is an intrinsic characteristic of the type of defect. The single lifetime τ of CuY and CuYM in the as-HIP condition remains almost unchanged to τ ~ (170 ± 3) and (180 ± 4) ps, respectively, from room temperature up to ~200 °C, indicating the Fig. 4. CDB ratio curves measured at selected temperatures of the isochronal annealing treatment (a) for CuY from 25 °C to 600 °C and (b) 600 °C to 900 °C; (c) for CuYE from 25 °C to 600 °C and (d) 600 °C to 900 °C. Studied Cu samples CDB ratio curves are normalized to a well annealed high purity Cu sample. For comparison, the CDB normalized ratio curve of a highly purity Y 2 O 3 sample is shown. Error bars have been included only for the Y 2 O 3 curve for clarity purposes. Equivalent error bars should be considered for the rest of the curves. R. Domínguez-Reyes, M.A. Monge, B. Galiana et al. Journal of Alloys and Compounds 900 (2022) 163430 5 presence of stable defects in that range of temperatures due to its high value compared with the expected for defect-free Cu. Table 1 summarizes the characteristic values of the positron lifetime for pure copper and its defects that have been confirmed by several works in the literature (see references in the table caption). Values of τ of 170–180 ps are representative either of competitive positron trapping and annihilation in different types of defects such as vacancy aggregates, nanovoids and vacancy complexes attached to dislocation, or at the Y-O particles and grain boundaries. Fig. 9 shows the temperature dependence of the S value for CuY in comparison with the values obtained for well annealed highpurity Cu (defect-free), and heavily cold-deformed high-purity Cu (thickness reduction of 50%). The S values in the temperature range of 25–200 °C are similar to cold-deformed pure Copper, where it has been well established in the literature that vacancy-like defects, vacancy agglomerates and vacancy-dislocations complexes are responsible of the positron annihilation (see [40,41] and references therein). This result agrees with the observed for the positron lifetime that exhibits the same dependence as S parameter over the whole temperature range (both compared in Fig. 9). The decrease of the positron lifetime to its minimum value of 133 ps is caused by a recovery stage in CuY that initiates at temperatures around ~300 °C and finishes at 600 °C (see Fig. 3). The onset and end temperature of the recovery process is shifted to higher values for CuYE, being the minimum lifetime of 145 ps reached at a temperature 50 °C higher than the observed for CuY. Fig. 9 also shows the recovery of the CuY, with the expected decrease of the S values on increasing temperature up to values near of the defect-free Cu, which can be attributed to the dissociation and annealing of vacancy aggregates in the Cu matrix. This recovery stage, that can be clearly followed in the plot of the (S, W) values that follows the same linear trend with temperature that the recovery of cold deformed high purity Cu to a thickness reduction of 50% (see Fig. 7(a) and (b)). These results seem to indicate that the annealing vacancy-related defects in these materials behave similarly to the well-studied recovery of irradiated or heavily deformed Cu. The annealing behavior of point defects and dislocations in pure copper and single-phase Cu alloys progresses gradually through five recovery stages [38,42–45]. The initial stages start at low temperature being stages III and IV corresponding to the recombination of migrating vacancies with interstitials, and the formation of new vacancies released by the dissociation of clusters. Stage V ends with Fig. 5. CDB ratio curves measured at selected temperatures of the isochronal annealing treatment (a) for CuYM from 25 °C to 600 °C and (b) 600 °C to 900 °C; (c) for CuYME from 25 °C to 600 °C and (d) 600 °C to 900 °C. Studied Cu samples CDB ratio curves are normalized to a well annealed high purity Cu sample. For comparison, the CDB normalized ratio curve of a highly purity Y 2 O 3 sample is shown. Error bars have been included only for the Y 2 O 3 curve for clarity purposes. Equivalent error bars should be considered for the rest of the curves. Fig. 6. Ratio CDB curves of Y2O3 and pure Y, divided by the reference annealed pure Cu. The ratio CDB curve of Y2O3 to pure Y is also depicted and shows the predominant characteristics of the CDB spectra associated with core electrons of the oxygen anions of the Y oxide. R. Domínguez-Reyes, M.A. Monge, B. Galiana et al. Journal of Alloys and Compounds 900 (2022) 163430 6 the full recovery of the microstructure at ≈ 450 °C when Cu has been subjected to extreme plastic deformation [46–49]. Even though recovery and recrystallization are competitive phenomena, at higher temperatures the changes in the microstructure are caused by grain growth. In the studied Cu reinforced materials, the presence of light impurities and precipitates, such as C or O, and Y-O particles can increase the temperature of the recovery processes [50]. However, CuY and CuYM have been submitted to a high sintering temperature of 850 °C, and therefore recovery should have been completed in the materials prior to the isochronal annealing treatments where the measurements have been performed. The presence of these vacancy-type defects in the as-HIP materials is due to two main reasons. CuY and CuYM were quenched from the sintering temperature to 350 °C in a short time of ≈ 15 min (cooling from that temperature to 50 °C inside the furnace took a time of ≈ 60 min), retaining vacancy-like defects produced from the high number of thermal activated vacancies at the sintering temperature. Also, the mismatch between the thermal expansion coefficients of the steel container and the reinforced Cu, 14.4 × 10 −6 K −1 for the steel, and (18.5 ± 0.1)× 10 −6 K −1 and (17.7 ± 0.2)× 10 −6 K −1 for the CuY and CuYM respectively, can produce strain in the sintered materials at cooling [51]. CuYE and CuYME follow the same behavior, with a larger positron lifetime and varying within a narrower range (See Fig. 3(b)). Fig. 7. Relative S-W plots for CuY, (a)-(b), and CuYM, (c)-(d), where the annealing temperatures are shown in the ranges 25 °C to 600 °C and 600 °C to 900 °C. All values are relative to pure annealed Cu. For reference, the values for high purity deformed Cu (cold-worked 50% thickness reduction) are shown, also, the relative S-W pair for CuYME cold-worked. Fig. 8. Relative S-W plots for (a) CuYE and (b) CuYME where the annealing temperatures are shown. All values are relative to pure annealed Cu. For reference, the values for high purity deformed Cu (cold-worked 50% thickness reduction) are shown, also, the relative S-W pair for CuYME cold-worked. R. Domínguez-Reyes, M.A. Monge, B. Galiana et al. Journal of Alloys and Compounds 900 (2022) 163430 7 Severe plastic deformation processing, such as ECAP, produces a remarkable transformation in the microstructure of these alloys, giving rise to a high density of dislocations that dynamically reorganize to form low and high angle boundaries and vacancy-like defects [52,53]. This should modify the observed positron lifetime and CDB spectra. However, dislocations act as shallow positron traps and interact with positrons in a two-stage process. Once the positron is weakly localized at the dislocation core, it diffuses along the dislocation line and annihilates with deep traps, mainly vacancy-like defects anchored to the dislocation elastic field [54–56]. This dislocation-vacancy complex increases the positron lifetime by 15–34 ps respect the expected for isolated defects [57–59]. Also, grain boundaries and triple-grain points are defective regions that behave as vacancy-like complex traps. Furthermore, the studied materials have a high density of Y-O particles that are expected to exhibit a high positron affinity. CuYE and CuYME exhibit a finegrained microstructure with a grain mean size ~500 nm, allowing a significant fraction of positron to reach the grain boundaries and annihilate with the vacancy-like defects, nanovoids or particles located there. The fraction of positrons that interact with grain boundaries is strongly dependent on the grain size and becomes significant when it is of the order of magnitude of the positron diffusion length, ~300 nm for defect-free Cu [60,61]. This submicrometric grained microstructure and defects produce at ECAP processing, causes the overall increase of the positron lifetime observed for CuYE and CuYME compared with the τ values of their unprocessed CuY and CuYM counterparts, which have a larger micrometric mean grain size. Before the recovery stage there is only a difference of ~10 ps in the τ values for ECAP materials because also as-HIP materials retain a high density of defects that hinder the effect of the grain boundaries, producing trapping saturation. As the recovery progresses and the minimum τ value is reached the difference increases to ps40 CuYE CuY and 25 CuYM CuYME ps for each material, that also correlates with the initial larger grain size and lower τ values before ECAP processing. The difference in the number of ECAP processing steps should not have a significant impact in the measured positron lifetimes because even though the density of defects strongly depends on the number of ECAP passes, saturation is achieved after between the 4th-5th pass [62,63]. It is common to observe a decrease of the density of induced defects for a higher number of passes due to dynamic recovery process activated after achieving a certain degree of strain in the material, which has been also observed in Cu and other severe plastic deformed materials processed with techniques that allow accumulative processing such as HPT (High-Pressure Torsion) [64]. The behavior of τ correlates with the changes observed in the relative S-W plots, where there is a slight overall shift of the S parameter to higher values after ECAP processing, indicating that the fraction of positron annihilating with low-momentum electrons increases, and therefore points out to trapping and annihilation with open volume defects (see Figs. 7 and 8). Fig. 9 showed that for CuY the behavior of the S parameter as a function of temperature matched the positron lifetime behavior shown in Fig. 3(a). As S-W plots seem to indicate that all the alloys present the same kind of openvolume defects, it might be the case that the similarity between S parameter and positron lifetime that CuY presents, is also exhibited by the rest of the studied samples. Fig. 10 contains the evolution of S parameter as a function of the annealing temperature for the four studied samples. It is evident that for the non-processed samples (Fig. 10 (a)) the S parameter replicates the distribution over temperature that positron lifetime shown in Fig. 3(a). For ECAP processed alloys, the similarity is less clear but it is still present. CuYE sample, exhibits a minimum in the 600 – 700 °C region in both S parameter and positron lifetime distributions, followed by an increase; and CuYME shows the minimum slightly shifted to lower temperatures. This correlates with the CDB curves that present a higher fraction of positrons annihilating with low-momentum electrons, i.e p m c3 10 L 30 × in the processed materials (Figs. 4 and 5). The CDB curves of CuY show the decrease of the CDB curve values at low momentum as the annealing temperature increases and progress the recovery of the Cu matrix with the annealing out of vacancies and vacancy clusters (see Fig. 6). This effect is hindered in the CDB as the Fig. 9. S parameter as a function of the annealing temperature for CuY compared to measured positron lifetime. The values for high purity Cu defect-free after annealing and heavily deformed (cold-worked 50% thickness reduction), are shown for reference. Fig. 10. S parameter as a function of the annealing temperature for (a) CuY and CuYM, and (b) CuYE and CuYME. R. Domínguez-Reyes, M.A. Monge, B. Galiana et al. Journal of Alloys and Compounds 900 (2022) 163430 8 grain size of the materials decrease due to the signal coming from positrons annihilating at grain boundary defects, such as vacancy clusters, nanocavities or Y-O particles. The CDB curves of all the studied reinforced Cu materials exhibits an overall similar structure in the high-momentum region, p m c10 10 L 30 × (see Figs. 4 and 5) to the observed in Fig. 6 for the Y 2 O 3 , and in particular to the Y 2 O 3 /Y ratio curve, disclosing the effect of the reinforcing Y-O oxides dispersion on the positron trapping characteristics [65,66]. Fig. 4 shows an increase in the high moment region of the CBD curves with increasing annealing temperature 300–600 °C for CuY, accompanied by the development of a structure similar to that observed for the Y 2 O 3 /Y ratio curve (see Fig. 6). The result indicates that positrons annihilate with oxygen core electrons and that there is an increase of the fraction of positrons trapped in defects whose chemical environment is rich in oxygen as recovery proceeds. CuYM, CuYE and CuYME exhibit a similar behavior, although this trend is partially masked by the signal coming from other positron traps (see Figs. 4 and 5). Numerous works have been published showing that positron annihilation spectroscopy is capable of characterizing the features associated to the presence of Y 2 O 3 and Y oxides in ODS steels [67–70]. However, there is still a certain degree of controversy in the interpretation of some results because the positron can annihilate in the Y 2 O 3 particles, or similar oxides, or in defects associated with these particles such as nanovoids or gas bubbles at the matrix/particle interface [67–70]. The characteristic positron lifetime for Y 2 O 3 is not clearly established due to the extreme difficulty of obtaining good reference pure Y 2 O 3 samples and the fact that this oxide has a high density of intrinsic structural lattice defects [65]. As previously appointed, measurements performed in our laboratory of sintered high purity nanopowder compacts of Y 2 O 3 gave a positron lifetime value of (229 ± 2) ps, that agrees with the single result of (239 ± 6) ps that has been published in the literature to our knowledge [71]. These values are higher than the measured lifetime of our samples, indicating that there is no positron trapping saturation at the reinforcing Y-O particles. Moreover, Y-O particles are extremely stable and do not undergo significant changes in the temperature range of the isochronous treatment. Therefore, the results obtained for the reinforced Cu materials are consistent with the presence of stable vacancy-aggregates or nanocavities after the recovery stage associated with Y-O particles. These defects have to be at the interface of the Cu/Y-O particles or associated with impurities or particles at the grain boundaries since they are not stable at that temperature in the pure Cu matrix [46–49]. Also, this is consistent with results obtained from positron studies performed in reinforced steels with a dispersion of Y 2 O 3 nanoparticles [67–69,72]. If the only surviving positron trapping sites after recovery were the Y-O particles there would only have been two possibilities that are not observed: i) saturation by positron trapping, in which case recovery would not have been observed, or ii) the positron lifetime spectra would have been resolved in two components: one corresponding to the bulk and the other to the reinforcing particles. The possibility that a fraction of annihilations occurs at the matrix of the Y-O and contribute to the mixed set of annihilation sites that gives a single unresolved positron lifetime cannot be excluded. The most startling change observed in Fig. 3 is the increase of the lifetime τ after the recovery stage, to reach an almost constant value for temperatures 100 °C above that at which the minimum τ value Fig. 11. TEM through-focal images of in CuY samples: after annealing at 600 °C ((a)-Under focused, (b)-Focused, (c)-Over focused) and after annealing at 900 °C (°C ((d)-Under focused, (e)-Focused, (f)-Over focused)). A detailed image is included for the white squared in images. R. Domínguez-Reyes, M.A. Monge, B. Galiana et al. Journal of Alloys and Compounds 900 (2022) 163430 9