Uncovering the role of nanoscale Si particles on the thermal stability of a lamellar-nanostructured Al–1%Si alloy

This study investigates particle governed thermal stability in lamellar-nanostructured Al–1.0%Si using in-situ transmission electron microscopy and post-mortem observations. Microstructural coarsening, dominated by Y-junction motion, is correlated with dispersed Si nanoparticles. Si particles within lamellae efficiently hinder dislocation movement during deformation, fostering a configuration with Si particles along incidental dislocation boundaries (IDBs). This particle–IDB configuration significantly impedes Y-junction motion, retarding lamellar coarsening. The enhanced pinning force from particle–IDB synergy, combined with direct pinning by Si particles, contributes to improved thermal stability in lamellar-nanostructured Al–1.0%Si. GRAPHICAL ABSTRACT IMPACT STATEMENT We uncover the mechanisms by which dispersed nanoparticles govern Y-junction motion during coarsening of lamellar nanostructures, in particular the synergetic pinning from particle-decorated interconnecting dislocation boundaries.


Introduction
Nanostructured metallic materials have a multitude of desirable properties unattainable for materials with larger grain sizes [1][2][3][4].However, such nanoscale structures come with a high density of crystallographic defects that make them thermally unstable at elevated temperatures, and sometimes even at room temperature, thereby leading to the loss of the exceptional properties inherent to the nanostructure through extensive structural coarsening [4,5].As a result, improving the thermal stability of nanostructures is of great importance from both academic and engineering standpoint.
In recent years, a number of approaches have been attempted to achieve this goal, and extensive studies including experiments, modeling, and simulations have been conducted [5][6][7][8][9].The key issue is to reduce the driving force for grain coarsening either by diminishing the free energy of grain boundaries or by kinetic pinning of grain boundary migration.Except for some special cases that rely on the microstructural architectures composed of low energy boundaries like twin boundaries or low angle boundaries to reduce grain boundary energy [6,10], the common approach to enhance the thermal stability of nanostructured metallic materials is to reduce the energy and mobility of boundaries by adding alloying elements [11,12].The solute atoms can segregate to grain boundaries, lowering the excess boundary energy and improving the thermal stability.This strategy has been used in various Al alloy systems such as Al-Mg, Al-Mg-Y, Al-Cu, Al-Cu-Sc, and Al-Cu-Mg [10,[13][14][15][16].The alloying elements may also form second phase particles, which can exert a local pinning force, thus retarding grain boundary migration and decelerating grain coarsening kinetics [17][18][19].A successful example is the significant improvement of thermal stability of ultrafine-grained Al-2%Fe alloy by introducing nanoparticles, as reported by Duchaussoy et al. recently [19].
For lamellar nanostructures, which are commonly obtained by high strain deformation, the dominating mechanism for uniform coarsening is Y-junction motion [20][21][22].A Y-junction in the lamellar nanostructure is a linear feature formed by three intersecting lamellar boundaries.Y-junction motion occurs along the direction parallel to the lamellar boundaries forming the junction [20][21][22], which differs from the grain boundary migration along the boundary curvature direction.The Y-junction motion removes a lamella and coarsens the two neighboring lamellae, a coarsening process different from the grain boundary migration.Consequently, the mechanisms by which dispersed nanoparticles govern Y-junction motion during the coarsening of lamellar nanostructures also differ from the Zener pinning mechanism during normal grain coarsening.Our previous study by in situ observation on annealing of lamellarstructured Al-0.3%Cu showed that Al 2 Cu particles precipitated dynamically along boundaries can stabilize the fine microstructures efficiently.Particularly, the combination of particles and incidental dislocation boundaries (IDBs) between lamellar boundaries can provide strong pinning effects to Y-junction motion [22].
Hence, in this study, a lamellar-structured aluminum Al-1.0%Si alloy with fine dispersed Si particles was designed to further reveal the mechanisms of nanoparticle mediated thermal stability and investigate potential routes for further enhancing the thermal stability of lamellar nanostructures.Unlike the Al 2 Cu particles precipitated dynamically during annealing in the lamellarstructured Al-0.3%Cu alloy, the Si particles were preexisting in the Al matrix before high strain cold rolling, so their pinning effect will be more consistent during the annealing process.

Materials and methods
A high-purity Al-1.0%Si alloy was prepared for this study, fabricated from 99.999% pure aluminum and highly pure Si (1.0 wt.%).An as-cast ingot was hotisostatic pressed at 200 °C to obtain a thick plate with a uniform grain structure and evenly dispersed Si particles.A plate with a thickness of 50 mm was subsequently rolled to a von Mises strain of 4.5 (98% reduction in thickness) in multiple passes to obtain a nanoscale lamellar structure.Note that the rolled sheet was quenched into liquid nitrogen cooled alcohol after each rolling pass to reduce the influence of sample heating on the deformation microstructure.To study the thermal stability of the lamellar nanostructure, several bulk samples cut from the cold-rolled plate were subjected to isochronal annealing for 1 h at temperatures in the range of 100-250 °C in a tube furnace.
The microstructures in these samples were examined in a JEOL JEM-2100 transmission electron microscope (TEM).All TEM foils were mechanically polished down to ∼ 100 µm in thickness and then twin-jet electropolished using a solution of 90% ethanol and 10% perchloric acid at -20 °C.In-situ heating experiments by a double-tilt heating holder (JEOL EM-31670SHTH) in the TEM were also performed to investigate the effects of Si particles on coarsening of the lamellar structure during recovery.For this experiment a thin foil sample was heated up to 250 °C over a time period of ∼ 13 min.Microstructural changes during heating were recorded by a TVIPS FastScan camera with a maximum frame rate of 12 fps at 1024 × 1024 pixel resolution.The observations were focused on the periods while the temperature was increasing, as most of the microstructural evolution occurred in these heating-up periods [23].All observations were conducted at a layer corresponding to the center of the cold rolled sheet.

Results and discussion
The microstructure of the Al-1.0%Sisample, subjected to 98% cold rolling, is illustrated in Figure 1.This figure, specifically Figure 1(a), displays a representative lamellar structure captured through a bright-field TEM image covering an area of approximately 200 µm 2 .The lamellae are constituted by lamellar boundaries (LBs) with an average spacing of 230 nm that are aligned parallel to the rolling plane and interconnecting incidental dislocation boundaries (IDBs) with an average boundary spacing of 760 nm that are aligned perpendicular to the rolling plane.Noteworthy features of this structure include Yjunctions, each constituted by the intersection of three LBs, and H-junction pairs, which consist of two LBs with an intervening IDB.These features are exemplified in Figure 1(a) and schematically represented in Figure 1(b).
The lamellar structure, beyond encompassing LBs, IDBs, Y-junctions, and H-junctions, also incorporates Si particles with a mean diameter of 15 nm.It is pertinent to mention that Si nanoparticles are discernible both at lamellar boundaries and in the interior of lamellae.This is exemplified by the black dotted and solid arrows in Figure 1(d).Predominantly, Si nanoparticles within lamellae are localized at IDBs, culminating in a common particle-IDB configuration, as denoted by the red arrows in Figure 1 this configuration can be observed in the magnified TEM image in Figure 1(d).Quantitative characterization reveals that this particle-decorated IDB configuration constitutes approximately 60% of all IDBs, stabilizing corresponding H-junctions, as shown in Figure 1(e).It is hypothesized that Si nanoparticles inside lamellae serve as obstacles to migrating dislocations, subsequently becoming sites for dislocation storage during plastic deformation.This results in the prevalent observation of IDBs adorned with Si nanoparticles.Such configurations are predominantly observed proximate to Yjunctions, indicating their potent pinning effect on moving Y-junctions during deformation.

(c). A detailed example of
In-situ TEM annealing experiments allow direct observation of microstructural evolution, visualizing the effects of Si nanoparticles on thermal stability.Figure 2 presents a typical example of coarsening of lamellar structure during annealing where migrations of four Y-junctions with medium/high misorientation angles (marked as Y1, Y2, Y3, and Y4) were tracked in real-time.Microstructural coarsening is caused by the shortening or disappearing of lamellae through Y-junction motion along the direction parallel to lamellar boundaries, as revealed in Figure 2(a-c).Considerable impacts on structural coarsening from Si particles can be seen in the interplay between moving Y-junctions and particles/particledecorated IDBs.During heating, the four traced Yjunctions underwent different migration processes due to different microstructural environments encountered.
In the context of migration of the Y-junction labeled Y1, as illustrated in Figure 3, the Y-junction was initially pinned by a configuration comprising a dislocation and a Si particle (refer to Figure 3(a1)).Upon heating the foil for approximately four minutes, the Y-junction commenced its movement to the left along the RD.Subsequently, the Y-junction was pinned by Si particles that were originally present at the Y-junction lamellar boundaries (see Figure 3(a2)).The Si particles along its path operated primarily as obstacles during further heating, delaying the Y-junction from moving.It is worth noting that in addition to the Si particles observed at lamellar boundaries, those situated within shortening lamellae can also exert a pinning force upon encountering Y-junctions.This pinning phenomenon, which might go unnoticed during static annealing, supplements the findings presented in the lamellar-nanostructured Al-0.3%Cu study [22], thereby enriching our understanding of particle-mediated thermal stability mechanisms.
Si particles can indirectly contribute to the pinning of moving Y-junctions by generating particle-IDB configurations in addition to directly pinning Y-junction motion.An illustrative example of particle indirect pinning of Y-junction motion is presented in Figure 3(b1-b4).The Y-junction (labeled Y2), indicated by an orange arrow, was initially pinned by a configuration of Si particles synergized with a dislocation boundary on its upward side prior to annealing (see Figure 3(b1)).As the annealing temperature increased, the Y-junction began to slowly move to the right (see Figure 3(b2)).During extended annealing, a span of 100 s was required to fully unpin from this configuration (as shown in Figure 3(b3)), indicating a strong pinning effect collectively exerted by the dislocation boundary and the Si particles.This pinning effect is typically enhanced by both an increase in the misorientation angle of the attached IDB and the thermal stability of the particles [19,22,24].Upon breaking through this formidable pinning, the Y-junction readily migrated and then halted in proximity to a new IDB (see Figure 3(b4)).
Multiple pinning effects on Y-junction motion were clearly seen in the case of Y3 migration (see Figure 3(c1)).At the early stage of heating, four dislocations attached to the upper receding lamellar boundary and several particles located at the lower receding lamellar boundary provided collectively dragging forces to hinder its movement (see Figure 3(c2)).As the annealing temperature increased, the pinning effects were mainly imposed by the encountered particles at the two receding lamellar boundaries, resulting in a jerky migration (see Figure 3(c3)).In the following heating time, Y3 stopped at the position decorated with particles inside the shortening lamella (see Figure 3(c4)).Additionally, an IDB decorated with two particles on the upper side of Y3 may also give a pinning effect.Taken altogether, throughout its migration, the Y-junction was exposed to various pinning effects from dislocations, particles, the synergy of particles and dislocations, and particle-IDB configurations.It is important to note that Y4, which was pinned originally by a particle-decorated IDB, remained stable throughout the heating process (as shown in Figure 2), demonstrating a potent pinning effect induced collectively by IDBs and particles.Furthermore, it can be seen that Si particles are unchanged during heating, stabilizing IDBs and maintaining the pinning effect.
In order to further evaluate the pinning effects of Si particles on Y-junction motion, the migrated distances of the above four Y-junctions at the corresponding times were measured during in-situ heating.Figure 4 shows the evolution of migration distance as a function of heating time, with the slopes of the plotted curves reflecting the migration velocity of Y-junction.It can be seen that the migration curves of mobile Y-junctions exhibit a regular pattern with variations in migration distance.At the early stage of heating, Y-junctions stay stable due to the low annealing temperatures.These moved Y-junctions have different incubation times (as indicated by arrows in Figure 4), and the minimum temperature that drives them to start moving is different.Moreover, there are also differences in the time required for these Y-junctions to completely escape from the pinning configurations created initially by deformation.A larger pinning effect from particle-IDB configurations may be implied by the longer incubation and unpinning times for Y2, as well as by the immobility of Y4 throughout the entire heating period.The most striking feature of the distance-time curves is a significant decrease in the migration velocity of Y-junctions after encountering particles or particle-IDB configurations and an abrupt increase in the velocity after passing through the pinning sites, revealing a central  The role of Si particles in improving the resistance to thermally induced coarsening of lamellar structure has been also demonstrated by post-mortem characterization.Figure 5(a-c) shows uniform coarsening of the lamellar structure without a significant change in the lamellar morphology after annealing at temperatures from 100 °C to 200 °C.Such a coarsening pattern is consistent with the Y-junction motion mechanism [25].In line with the deformation microstructure, all of the annealed microstructures feature Si particles that are uniformly distributed at lamellar boundaries and inside lamellae, IDBs, and particle-IDB configurations, as shown in Figure 5(d-f).It is worth stressing that the remaining Y-junctions after annealing are typically with Si particles and/or close to particle-IDB configurations, indicating a direct pinning from Si particles and/or a synergic pinning from particle-IDB configurations.
Here we define a grain size retention [26] (d 0 /d f , where d 0 is the average initial grain size and d f is the average grain size after annealing, so this ratio represents roughly the remaining fraction of stored energy) and plot the retention of average lamellar boundary spacing as a function of temperature, as shown in Figure 5(g).A significantly lower coarsening rate is found in the Al-1%Si alloy with evenly dispersed particles compared to other nanostructured Al alloys without intragranular particles [22,27,28].Furthermore, compared to the lamellar-nanostructured Al-0.3%Cu alloy, where recrystallization was observed after annealing at 175 °C for 1 h, the recrystallization in Al-1%Si alloy is notably delayed (recrystallized grains are observed after annealing at 225 °C for 1 h).Such a remarkable thermal stability of the Al-1%Si alloy indicates a strong particle-related resistance to the coarsening of lamellar structures.In addition, it is seen that these particles are quite stable, only becoming slightly coarser during annealing.This is because almost all Si elements are present as particles due to the extremely low solid solubility of Si in the Al matrix.These pre-existing Si particles aid in the formation of IDBs through accumulating dislocations during deformation.The density of such particle-IDB configurations, as well as IDBs without particle decoration, decreases during annealing as shown in Figure 5(h).The removal of particle-IDB configurations is a natural consequence of the loss of the lamellae housing them.Additionally, it can be seen that, particularly at annealing temperatures below 150 °C, the area density of Y-junctions decreases only slightly during annealing, indicating maintenance of the lamellar structure.
Several concurrent pinning mechanisms such as dislocation pinning, IDB pinning, particle pinning and particle-IDB pinning act in concert to improve the Here the detailed interaction is neglected and only the initial and final states are considered.The pinning effect of particles is defined as the case where Si nanoparticles lying at lamellar boundaries pin Y-junction motion directly, whereas the pinning effect of particle-IDB configurations is defined as the case where Si nanoparticles lying at IDBs pin Y-junction motion indirectly.Evidently, the pinning from particle-IDB configurations is dominant in hindering the coarsening of lamellar structures.Structural self-stabilization by pinnings of dislocations, IDBs, and LBs, inherently exists in lamellar structures, but it is usually too weak to stabilize the lamellar nanostructures, as reflected by the reported poor thermal stability of lamellar-nanostructured pure metals [29,30].It is reasonable that the enhanced resistance to lamellar structure coarsening is mainly attributed to particle-related pinnings of Y-junction motion.During Y-junction migration, the encountered Si particles, both located at receding lamellar boundaries and in the interior of shortening lamellae, can act as obstacles to directly retard the moving Y-junctions.Additionally, Si particles can participate in the pinning of moving Y-junctions in an indirect manner-a particle-decorated IDB configuration.Such a pinning has a synergetic effect from particles and IDBs.Taken altogether thermally stable particles can provide a persistent pinning force-both directly and indirectly during annealing of the lamellar structures.With higher number densities of Si nanoparticles, which can create denser particle groups and particle-IDB configurations, further enhancement of the thermal durability of lamellar structures may be achieved.

Conclusions
To conclude, through combing post-mortem and in-situ TEM observations, we have uncovered the role of dispersed nanoparticles in stabilizing lamellarnanostructured Al-1%Si alloy against recovery coarsening.Encountered Si particles, both at lamellar boundaries and inside lamellar grains, can directly pin Y-junctions during their motion.Furthermore, particles can interact with interconnecting dislocation boundaries to generate a synergic pinning of Y-junction motion.These particlerelated pinning mechanisms improve the thermal stability of lamellar-nanostructured Al-1.0%Si.

Figure 1 .
Figure 1.Microstructure of the as-deformed Al-1%Si alloy.(a) An overview TEM micrograph shows the lamellar structure in the longitudinal section, where examples of Y-junctions and H-junctions are pointed by an orange arrow and a red arrow, respectively.(b) An illustration of a Y-junction formed by three lamellar boundaries (LBs) and three H-junction pairs each formed by two LBs and an interconnecting incidental dislocation boundary (IDB) between them.(c) A bright-field TEM micrograph shows Si particles and particle-IDB configurations (marked by red arrows) with particles lying on IDBs.(d) A higher magnification TEM micrograph shows an example of the configuration with particles lying on IDB, corresponding to the area enclosed by the red dash box in (c).(e) Area densities (number per unit area) of triple junctions in the deformation microstructure.

Figure 2 .
Figure 2. A global evolution of the lamellar structure in the same position of the TEM foil during in-situ annealing at the target temperature of 250 °C where four Y-junctions marked as Y1, Y2, Y3, and Y4 are traced and shaded.(a) Before heating; (b) After heating; (c) The corresponding sketch where light grey indicates areas swept by Y-junction motion.

Figure 3 .
Figure 3. In-situ observation of moving Y-junctions and their interactions with Si particles and/or particle-IDB configurations in lamellarnanostructured Al-1.0%Si.(a1-a3) Direct pinning of Y-junction motion by Si particles; (b1-b4) Synergic pinning of Y-junction motion by particle-IDB configurations; (c1-c4) Multiple pinning effects on Y-junction motion.TEM micrographs and corresponding sketches are presented together.The annealing times are given on each micrograph.

Figure 4 .
Figure 4. Evolution of migration distances of the four Y-junctions as a function of heating time.

Figure 5 .
Figure 5. Microstructural evolution in the Al-1%Si samples isochronally annealed for 1 h at (a) 100 °C, (b) 150 °C, (c) 200 °C.(d-f) High magnification TEM micrographs correspond to the areas enclosed by red dash boxes in (a-c), respectively.Y-junctions and corresponding attached particle-IDB configurations are pointed out by orange and red arrows.(g) Retentions (d0/df, where d0 is the average initial grain size and df is the average grain size after annealing) of average lamellar boundary spacing and Si particle diameter as a function of annealing temperature.(h, i) Variations of area density of (h) triple junctions and (i) Y-junctions pinned by IDB, particle, and particle-IDB configuration during annealing.