Influence of Si and Ge Microalloying Additions and Plastic Deformation on Precipitation Processes in Al-Cu Based Alloys

The combined effect of micro-alloying with Si and Ge and/or plastic deformation prior to ageing at 160 ̊C on age hardening has been studied in an Al-2 at% Cu alloy. The results obtained indicate that the hardness response is faster and the peak hardness is higher when plastic deformation and micro-alloying are applied together than performing each procedure individually. Different amounts of deformation, ranging from 0% to 30% have been utilized. An optimum deformation degree for the hardening response has been established around 8% for the Al-Cu-Si-Ge alloy. Characterization by transmission electron microscopy (TEM) shows that the peak hardness is due to a complex microstructure that contained θ ′′ disc shaped precipitates, rod-shaped Si-Ge precipitates and θ ′ plates that were heterogeneously nucleated on the Si-Ge particles. Pre-deformation has been found to stimulate the growth of the θ ′ plates due to enhanced diffusion along dislocation cores. Increasing deformation reduces the influence of the Si-Ge precipitates on heterogeneous nucleation, leading to reduced peak hardness and faster over-ageing.


Introduction
The Al-Cu system has been extensively investigated as a model system for precipitation hardening.In fact, the phenomenon of precipitation hardening has been first discovered in an Al-Cu based alloy [1].In order to improve the hardening evolution in heat treatable alloys, several methods have been devised.One method involves the addition of small amounts of alloying elements, known as micro-alloying elements [2]- [15].Another method is to apply a plastic deformation before artificial ageing takes place [4] [16] [17].Both methods result in faster hardening kinetics and higher hardness values by modifying the precipitation process.
In Al-Cu based alloys, different kinds of micro-alloying additions have been studied.Trace additions of Sn, Cd and In to Al-Cu have been shown to enhance the hardening rate and to increase peak hardness [2] [4] [5].In ternary Al-Cu-Mg alloys, the effects of minor Ag additions have been demonstrated to stimulate the precipitation of different metastable phases, depending on the Cu:Mg ratio [6]- [12].The addition of Si and Ge in equal proportions has also been proved to increase the hardening rate during artificial ageing, together with an increase in peak hardness [14] [15].It has been observed that diamond cubic Si-Ge precipitates provide nucleation sites for metastable θ ′ phase, leading to a high density of such precipitates at peak hardness.
Plastic deformation prior to artificial ageing has been able to increase peak hardness in binary Al-Cu alloys as well [4].In this case, the increase in hardening rate has been related to faster diffusion paths along the dislocations introduced by plastic deformation.Furthermore, dislocations also serve as heterogeneous nucleation sites for precipitates, thus enhancing the nucleation rate [4].In the commercial 2024 alloy, plastic deformation prior to artificial ageing has been found to increase the hardening rate and peak hardness.Microstructural characterization has shown that dislocations provide nucleation sites for the S phase, resulting in a homogeneous distribution of S phase precipitates [16] [17].
However, in Al-Cu alloys, the combination of microalloying additions of Cd and plastic deformation prior to artificial ageing has been found to produce worse results than for each method applied individually [18].
On the contrary, in Al-Cu with minor additions of Si and Ge, the combination of microalloying and plastic deformation prior to artificial ageing at 190˚C led to an acceleration of the ageing response [19].
The main goal of the present study is to further analyze the combined effects of microalloying with Si and Ge and plastic deformation in an Al-2 at% Cu alloy.
For this purpose, different degrees of deformation prior to artificial ageing at 160˚C have been applied.The hardness evolution during artificial ageing is studied and microstructural characterization by means of transmission electron microscopy (TEM) is also carried out after specific ageing times.In addition, to help understanding the processes involved, the simpler related systems (Al-Cu and Al-Si-Ge) are analyzed.

Experimental
Alloys with the following nominal compositions were melted in an arc furnace: Al-0.5Si-0.5Ge,Al-2Cu, Al-2Cu-0.5Si-0.5Ge(at%).All of these alloys were ho-J.Minerals and Materials Characterization and Engineering mogenized at 520˚C under an Argon atmosphere.
Three specimens (diameter = 6 mm and length = 2 cm) from each alloy were machined.Two cylinders were subjected to either 8% or 15% deformation in compression, using an INSTRON 5563 Mechanical Testing Machine.The remaining cylinder was not deformed.Slices that were 1 mm in thickness were cut from each cylinder and were artificially aged at 160˚C for different ageing times, t a .For the Al-Cu-Si-Ge alloy a fourth specimen that had been pre-deformed to 30% by cold rolling was also studied.
The evolution of Vickers Hardness (H v ) as a function of t a was obtained for each degree of deformation; a Mitutoyo Hardness Testing Machine with a load of 100 g was used to make the measurements.
Samples for TEM characterization were prepared by mechanical grinding followed by double jet electropolishing with an electrolyte based on Methanol and a mixture of H 2 SO 4 , HF and Glycerol.The microstructure was analyzed after specific ageing conditions that were of interest using a Philips CM200 UT microscope with a LaB 6 filament and a FEI Tecnai F20 microscope with a Schottky field emission gun, both operated at 200 KeV.In order to determine precipitate densities, the local thickness was measured using convergent-beam electron diffraction (CBED) in a two-beam condition with g = 200 Al as reported by Kelly [20].

Hardness Measurements
Hardness evolution as a function of ageing time is displayed in Figure 1 where the behavior of the pre-deformed and non-deformed specimens is compared for each alloy.
For the ternary Al-Si-Ge alloy (Figure 1(a)) the hardness evolution curve has been uniformly shifted for all ageing times to higher values as a result of pre-deformation.The shift for 15% deformation was around two or three times larger than the shift observed for a deformation of 8%.The ageing time at which peak hardness occurred was 20 h for all cases, independent of the degree of pre-deformation.
For the binary Al-Cu alloy (Figure 1(b)), the annealing time required to reach the peak aged (PA) condition was significantly reduced.In the case of the undeformed alloy the maximum was observed for t a ~ 60 h and it remained constant for longer ageing times while for the pre-deformed alloys (either 8% or 15%) the peak was achieved after only 20 h ageing.However, the peak hardness was unaffected by pre-deformation.The main results extracted from Figure 1 (i.e.PA hardness, H v , and ageing time, t a ) are summarized in Table 1 for each alloy and for each pre-deformation degree ≤15%.The results for the 30% pre-deformed Al-Cu-Si-Ge alloy (H v = 160 ± 5 and t a = 3 h) were not included in Table 1 as they were almost coincident with those from the 15% pre-deformed Al-Cu-Si-Ge specimen.After 15% pre-deformation, the hardening response was even faster, reaching peak hardness at an ageing time of 3 h, but the peak hardness was the same as that obtained in the non-deformed condition.Increasing the pre-deformation to a degree of 30% resulted in an acceleration of the overageing stage and a reduction of the peak hardness.These results suggest that the optimum pre-deformation that produces maximum peak hardness is around 8% or could be even lower.In order to understand the observed hardness behavior, the microstructure of the Al-Cu and Al-Cu-Si-Ge alloys after certain ageing times at 160˚C, with and without pre-deformation of 8% has been investigated with TEM observations.Figure 2 shows dark field (DF) images of the Al-Cu alloy in the undeformed and after 8% pre-deformation conditions, after 20 h ageing.For the 8% pre-deformed condition this corresponds to the PA condition.

Microstructural Characterization
Dark field images in Figure 2 have been obtained in 2-beam diffraction condition with the reflection g = 200 Al , its direction is indicated in the bottom-left corner in each micrograph.The specimen was oriented close to the [011] Al matrix Zone Axis (ZA).In Figure 2(a) and Figure 2(b) weak reflections about midway between the transmitted (000) and the diffracted (200 Al ) beams were used to highlight one of the three variants of the θ ′ or θ ′′ precipitates.In this condition the imaged plate shape precipitates were parallel to the electron beam so they are displayed "edge-on".In Figure 2(c) and Figure 2(d) dark field images from one of the two other variants of the θ ′ or θ ′′ precipitates, inclined to the beam direction were obtained using (112) θ ′′ / θ ′ type reflections, by tilting slightly away from the [011] Al matrix Zone Axis (ZA).While for the non-deformed condition only θ ′′ precipitates were observed (Figure 2(a) and Figure 2(c)), in the 8% pre-deformed condition larger θ ′ phase precipitates were observed together with a high density of smaller θ ′′ phase precipitates.In Figure 2(b) the edge-on θ ′′ precipitates are shorter and thinner than the θ ′ precipitates.In Figure 2(d) the inclined θ ′′ precipitates show weaker contrast than the larger θ ′ phase precipitates.
Results of precipitate size and density measured on micrographs like those shown in Figure 2 are summarized in Table 2.The lengths (  ) correspond to the average diameter of the disc-shaped θ ′′ or plate-shaped θ ′ precipitates, respectively.The average lengths are reported accompanied by their error ( ∆ ) and, in brackets, by their Standard Deviation (SD) to have an idea of the precipitate sizes dispersion.The error, ∆ , was considered to be the SD divided by N , where N is the number of measured precipitates.
To obtain the precipitate densities (ρ) the local thickness was measured using CBED patterns in a two-beam condition with g = 200 Al as reported by Kelly et al. [20].
The density of the θ ′ precipitates in the pre-deformed specimen was esti- mated from the proportion between the θ ′′ and the θ ′ precipitates observed in pictures like those in Figure 2(b) and Figure 2(d).The amount of the θ ′ precipitates was accounted to be around 15% of the amount of the θ ′′ precipi- tates.
For the Al-Cu-Si-Ge alloy, DF images in Figure 3 2. Effect of pre-deformation on the diameter and density of θ ′′ and θ ′ preci- pitates for the Al-Cu system after 20 h ageing at 160˚C.The obtained average lengths (  ) and the densities (ρ) for each precipitate type are shown.The lengths are reported together with their error and, in brackets, their standard deviation.

Al-Cu
No Def 8% Def that are highlighted using the (002) θ ′ or (002) θ ′′ reflections at ~(1/2).g, while in Figure 3(c) and Figure 3(d) only inclined θ ′ precipitates are observed by selecting the (112) θ ′ reflection.It's worth noting (as well as for Fig. 2d) that although both θ ′ and θ ′′ precipitates were present, the contrast from θ ′′ was very weak and these precipitates were not easily seen in this condition.
Si-Ge precipitates in the Al-Cu-Si-Ge alloy formed with a rod-shaped morphology oriented parallel to the <001> matrix directions, with the A few Si-Ge precipitates are indicated with white and short arrows in Figure 3(e) (non-deformed specimen) and Figure 3(f) (8% pre-deformed specimen) BF images.
Measured precipitate sizes and densities of the θ ′′ , θ ′ and Si-Ge precipi- tates in the Al-Cu-Si-Ge alloy (with and without pre-deformation) in peak aged condition are given in Table 3.The length (  ) of the θ ′ precipitates corres- ponds to their average projected plate diameter, determined in 2-beam condition with the 200 reflection strongly excited (Figure 3(a) and Figure 3(b)).In this condition, the θ ′′ phase precipitates were differentiated from those of the θ ′ phase due to their much smaller thickness.In the same way as for the Table 2, the average lengths in Table 3 are reported accompanied by their error and, in brackets, by their Standard Deviation.Considering that a large amount of precipitates was measured (usually between 100 and 400 precipitates in each case), the main source for error in ρ was attributed to the error in the local foil thickness determination.Thus, the uncertainty in all the measured density values (Δρ) was estimated to be 10%.

Underaged Microstructure in Al-Cu-Si-Ge
To analyze why differences in peak aged microstructure were generated between the un-deformed and the 8% pre-deformed Al-Cu-Si-Ge, specimens in the underaged condition were also studied.Figure 4(a) and Figure 4(b) show high resolution (HR) images to illustrate the microstructure at a very early stage of ageing (t a = 3 h), for both non-deformed and pre-deformed samples.In both cases, a very high density of θ ′′ precipitates was observed, together with occa- sional Si-Ge and θ ′ precipitates.For the non-deformed specimen, Figure 4(a) displays a Si-Ge precipitate imaged "end-on", that is, with the long direction of the rod parallel to the electron beam.A detailed view of this same Si-Ge precipitate is shown in Figure 4(c).In addition, a detailed view of a θ ′ "edge-on" precipitate can be seen in Figure 4(e).
Table 3.Effect of pre-deformation on the length and density of θ ′′ and θ′ precipitates for Al-Cu-Si-Ge in PA condition (t a = 20 h).The obtained average lengths (  ) and the densities (ρ) for each precipitate type are shown.The lengths are reported together with their error and, in brackets, their Standard Deviation.
At this short ageing time (t a = 3 h), precipitates from θ ′ phase are rarely observed in the non-deformed specimen (Figure 4(e)) although they are often observed after pre-deformation.Furthermore, in this latter case θ ′ precipitates appeared more developed (Figure 4(b), Figure 4(f)) and more frequently, but not always, attached to a Si-Ge bar.
After a longer ageing time (t a = 10 h), θ ′ precipitates are abundant in both un-deformed and deformed conditions.The DF images shown in Figure 5 reflection from (112) θ ′ planes has been used to highlight θ ′ precipitates in- clined with respect to the electron beam.Some notches can be observed in θ ′ plates (indicated with yellow arrows) that evidence heterogeneous nucleation of this θ ′ phase on previously formed Si-Ge precipitates [14].It can be clearly seen that θ ′ precipitates are more developed in the pre-deformed sample.

Higher Deformation Degrees (15% Def.) and the Effect on Microstructure
Figure 6 shows the microstructural evolution in Al-Cu-Si-Ge after 15% pre-deformation, for three different ageing conditions: underaging (t a = 1 h), peak aged (t a = 3 h) and overaged (t a = 100 h).
The HR images were obtained in 2-beam condition using (200) reflections  In the under-aged condition (Figure 6(a)), HR imaging allows the observation of a Si-Ge precipitate (enclosed by an orange dotted line) which is associated to an inclined precipitate from θ ′ phase (inside a red dotted line).Two other θ ′ precipitates which are not attached to Si-Ge precipitates are shown inside red dotted lines, one of them (perpendicularly oriented with respect to the Si-Ge rod) is oriented "edge-on".Moreover, some smaller "edge-on" θ ′′ preci- pitates are seen parallel to this θ ′ plate.In Figure 6  In Figure 6(c) and Figure 6(e) the precipitates are observed "edge-on" because they are highlighted using the reflection at ∼(1/2).g, while in Figure 6(d) and Figure 6(f) the precipitates are inclined with respect to the electron beam and they are highlighted by selecting the reflection from the (112) θ ′′ / θ ′ planes.

Discussion
For the Al-Si-Ge system the hardness evolution after pre-deformation shows a uniform shift towards higher values for all ageing times.This shift larger for higher degrees of deformation.However, this shift is not significant.The peak hardness after 15% pre-deformation barely reaches the value of the initial hardness in the Al-Cu alloy at very short ageing times (Figure 1(a) and Figure 1(b)).
The fact that the same shift is obtained for all ageing times indicates that this shift can be attributed to work hardening and that the dislocations introduced by pre-deformation did not alter the nucleation and growth processes.It also suggests that the additional dislocations introduced by the previous deformation did not provide additional nucleation sites for the precipitation process, since this would alter the precipitation kinetics and thus modify the hardness evolution.
The main effect of pre-deformation in the Al-Cu system is a significant reduction in the ageing time to reach peak hardness, while the maximum H v value is unaffected.The TEM observations show that pre-deformation enhances the precipitation of the θ ′ phase, as previously reported by Silcock [4].Whereas without pre-deformation the microstructure at peak hardness in Al-Cu (t a = 100 h) contained only θ ′′ precipitates, after pre-deformation the peak hardness microstructure contained sparsely distributed θ ′ precipitates among a dense distribution of θ ′′ precipitates.
Microalloying additions of Si and Ge to Al-Cu exhibit a marked increase in peak hardness values compared to the binary Al-Cu alloy, confirming previously reported results for artificial ageing at 190˚C [14] [15] [19].For treatment at 160˚C, the time to reach peak ageing has been longer and the maximum hardness obtained is higher when compared to the values measured after ageing at 190˚C (20 h as compared to 3 h, and H v = 170 as compared to H v = 158 for ageing at 160˚C and 190˚C, respectively).This is the expected behavior for precipitation processes and can be explained by an increased undercooling at 160˚C, that produces a higher density of precipitates due to a larger driving force, but slower precipitation kinetics due to the lower temperature.
TEM characterization evidences that at 160˚C the microstructure in the peak aged condition is complex and included small disc shaped θ ′′ precipitates pa- rallel to the {100} matrix planes, rod-shaped Si-Ge precipitates, parallel to the <100> matrix directions and plate shaped θ ′ precipitates parallel to the {100} matrix planes.The rod-shaped morphology of the Si-Ge precipitates in the aluminum matrix is characteristic of Cu containing alloys and differs from that of the ternary Al-Si-Ge alloy where equiaxed precipitates are dominant, as previously described by Mitlin et al. [15] and by Mukhopadhyay [21].These Si-Ge precipitates are effective nucleation sites for heterogeneous preci-pitation of θ ′ phase, thus reducing the ageing time to reach peak hardness and achieving a higher hardness.However, the microstructure observed at peak ageing at 160˚C differs from that observed at peak ageing when aging is performed at 190˚C [15], where only Si-Ge and θ ′ precipitates are observed, but no θ ′′ precipitates.At this ageing temperature (190˚C), θ ′′ precipitates are formed during the early stages of ageing but dissolved completely in favor of the θ ′ precipitates once the peak hardness ageing time is reached.The high density of θ ′′ precipitates observed in the peak aged condition at 160˚C suggests that such precipitates contribute significantly to peak hardness.This conclusion is supported by comparing the precipitate densities measured at peak ageing in 8% pre-deformed material after aging at 160˚C with those reported in an alloy of similar composition but aged at 190˚C [19].After aging at 190˚C, the microstructure contains only Si-Ge precipitates and θ ′ plates with similar densities to those observed in the present work, but peak hardness is significantly lower (H v = 160) than after ageing at 160˚C (H v = 190).The difference in peak hardness can then be attributed to the additional θ ′′ precipitates observed after ageing at 160˚C.An important result in this work is that for the microalloyed Al-Cu-Si-Ge alloy, plastic deformation prior to ageing at 160˚C results in a faster hardening response and further increase in peak hardness, if the pre-deformation is below 15%. Figure 1(c) shows that a maximum peak hardness value of H v = 190 is obtained for 8% pre-deformation.For 15% pre-deformation peak hardness is the same as that for the undeformed alloy (H v = 170), but it is reached after 3 h, which is about six times faster than without pre-deformation.Further increasing pre-deformation to 30% prior to ageing results in a reduction of peak hardness.Thus, pre-deformation of about 8% is considered as an optimum to maximize peak hardness.
The results presented in this work clearly explain how the incorporation of Si and Ge microalloying elements combined with prior plastic deformation dramatically alters the hardening response with respect to that of the binary Al-Cu alloy.In Al-Cu, a peak hardness of H v = 134 is obtained after 100 h ageing at 160˚C whereas the combination of microalloying with Si and Ge and 8% pre-deformation increases peak hardness to H v = 190 and reduces the time to peak ageing by a factor of 5, to 20 h.For 15% pre-deformation, the time to obtain peak hardness is reduced to 3 h, i.e. by a factor of about 30 with a peak hardness of H v = 170, which is considerably higher than that of the binary alloy.
TEM characterization of the material deformed to 8% demonstrated that at peak ageing the microstructure is qualitatively similar to that without pre-deformation, containing θ ′′ disc shaped particles, rod-shaped Si-Ge precipitates and θ ′ plate shaped precipitates.However, as can be observed in Table 3, the PA microstructure in the Al-Cu-Si-Ge alloy after 8% pre-deformation displays larger diameter plates of θ ′ phase with a similar density when compared to the un- deformed material, whereas the size and density of Si-Ge and θ ′′ precipitates is almost unchanged by pre-deformation.This indicates that the increase in hardness of the material deformed to 8% can be attributed to the increased size of θ ′ precipitates.
It is worth noting that the largest difference in hardness between 8% pre-deformed and undeformed Al-Cu-Si-Ge specimens, is observed during the early stages of ageing.Microstructural characterization after 3 hours (Figure 4) presents a dense distribution of precipitates in both materials while the densities of Si-Ge and θ ′ precipitates are around 20 times lower than that of the θ ′′ precipitates.Detailed analysis of the microstructures reveal that the Si-Ge/ θ ′ precipitates are usually larger in the 8% pre-deformed specimen (l Si−Ge > 5 nm, l θ ′ > 10 nm) than in the undeformed material (l Si−Ge < 5 nm, l θ ′ < 10 nm).The under-aged microstructure after 10 hours ageing clearly shows that the θ ′ precipitates are much more developed in the 8% pre-deformed material compared to the undeformed material (Figure 5).In both alloys heterogeneous precipitation of θ ′ precipitates on Si-Ge precipitates is observed.Based on these results, it's possible to relate the faster development of θ ′ precipitates to the dislocations introduced by pre-deformation that provide fast diffusion paths for solute atoms.In summary, the faster increase of hardness in the material that has been pre-deformed to 8% is associated to the accelerated growth of θ ′ pre- cipitates caused by enhanced solute diffusion along dislocation cores.
In the Al-Cu-Si-Ge alloy that has been pre-deformed to 15%, peak hardness is reached after ageing for 3 hours.The peak-aged microstructure qualitatively resembles that of the Al-Cu alloy pre-deformed to 8%, where a very high density of θ ′′ precipitates is observed among a low density of large θ ′ plates.However, the marked difference in peak hardness (H v = 170 in Al-Cu-Si-Ge compared to H v = 130 in Al-Cu) is associated with the contribution of heterogeneous nucleation of θ ′ plates at Si-Ge precipitates in the quaternary alloy.However, θ ′ plates that are not associated with Si-Ge precipitates are also observed indicating that as pre-deformation is increased, a fraction of θ ′ plates form by heteroge- neous nucleation at dislocations.The faster evolution of the θ ′ plates in pre-deformed Al-Cu-Si-Ge as compared to the undeformed alloys is accounted for by enhanced diffusion along dislocation cores.In the overaged condition, θ ′ plates grow at the expense of θ ′′ precipitates.The faster hardening kinetics and higher hardness values of the microalloyed and pre-deformed material as compared to binary Al-Cu, observed after pre-deformation to either 8% or 15%, are then related to the combination of heterogeneous nucleation of the θ ′ phase on Si-Ge precipitates and enhanced diffusion along dislocation cores.
For 30% pre-deformation, peak hardness occurs after very short aging times, but to a lower hardness value.Since θ ′ plates are observed and they are not associated to Si-Ge precipitates in the 15% pre-deformed alloy, it is suggested that as pre-deformation is increased, a progressively larger fraction of θ ′ plates nucleate on dislocations.The distribution of θ ′ plates nucleated on disloca- tions is less efficient in increasing hardness, as borne out by the results for the binary Al-Cu alloy.Finally, in the Al-Cu-Si-Ge alloy it has been observed that the over-aged condition is reached earlier as the degree of pre-deformation increases, indicating that deformation promotes coarsening.This effect can again be related to faster diffusion along dislocation cores.

Conclusions
The combined effect of microalloying with Si and Ge additions and pre-deformation on the hardening response at 160˚C in Al-Cu has been thoroughly studied.effects of pre-deformation on the hardening response in binary Al-Cu and ternary Al-Si-Ge have been also investigated for comparison.
The main conclusions are as follows: 1) In Al-Si-Ge, pre-deformation results in a small shift of hardness to higher values, but does not alter the precipitation kinetics of Si-Ge precipitates.
2) In Al-Cu, prior deformation promotes the precipitation of the θ ′ phase, resulting in a shorter time to achieve peak hardness, but does not increase the peak hardness value.
3) In Al-Cu-Si-Ge, ageing at 160˚C without pre-deformation results in higher peak hardness values and much shorter times to achieve peak hardness as in Al-Cu.Compared to previous results at 190˚C, peak hardness is higher but the time to reach peak hardness is increased.TEM characterization shows a complex microstructure that includes a very high density of disc-shaped θ ′′ precipitates and lower densities of rod-shaped Si-Ge particles and plate-shaped θ ′ precipi- tates.The θ ′′ precipitates contribute significantly to peak hardness.4) Pre-deformation to 8% prior to ageing at 160˚C in microalloyed Al-Cu-Si-Ge further accelerated the hardening response and increased peak hardness.Higher amounts of pre-deformation result in reduced peak hardness and accelerate the onset to the overaged condition.
5) An optimum degree of pre-deformation that maximizes the hardening response is estimated to be about 8%.
6) TEM characterization of 8% pre-deformed Al-Cu-Si-Ge in the peak aged condition shows a qualitatively similar microstructure as that of the undeformed alloy.
Precipitate densities are similar, but the θ ′ precipitates are larger.This effect is attributed to enhanced diffusion along dislocation cores.Overageing was associated with the coarsening of θ ′ phase plates at the expense of θ ′′ precipitates.7) In the under-aged condition, the microstructure of both pre-deformed and undeformed specimens is dominated by a very high density of θ ′′ precipitates and the incipient formation of Si-Ge and θ ′ phase particles.Pre-deformation resulted in faster growth of the θ ′ phase and Si-Ge precipitates, particularly during the early stages of hardening, and is related to enhanced diffusion along dislocation cores.

Figure 1 (
Figure 1(c) shows the combined effect of microalloying and pre-deformation on Al-Cu.The non-deformed Al-Cu alloy with minor additions of Si and Ge reached peak hardness after 20 h ageing, a considerably shorter time than for the binary alloy; in addition the peak hardness (H v = 170) was considerably higher.After

Figure 2 .
Figure 2. DF images in Al-Cu, ta = 20 h: with no deformation (left column) and with 8% pre-deformation (right column).All images were acquired in two-beam condition with the g = (200) Al reflection (described by a thick arrow in the left bottom corner within each image), close to ZA = [011] Al .(a) (b) Edge-on θ ′ or θ ′′ precipitates (only one of the three variants), obtained from weak reflections at ~(1/2).g; (c) (d) Inclined θ ′′ precipi- tates and also some θ ′ precipitates in Figure 2(d) (one of the two other variants) are highlighted by tilting slightly away from the ZA = [011]Al and using the reflections from (112) θ ′′ / θ ′ planes.

Figure 3 .
Figure 3. Microstructure in Al-Cu-Si-Ge without pre-deformation (left) and after 8% pre-deformation (right), in PA (t a = 20 h).All images were obtained in 2-beam condition with g = (200) Al near to the [011] Al Zone Axis.Vector g is indicated in the bottom left corner of every image.(a) (b) Edge-on θ ′ precipitates, highlighted by selecting the weak reflection placed at ∼(1/2).g (only one of the three variants); (c) (d) Inclined precipitates (one of the two other variants) are highlighted by tilting slightly away from the ZA = [011] Al and using a reflection from (112)θ ′ planes; (e) (f) BF images showing bar type Si-Ge precipitates with fringed Moirée pattern (one of three variants).White short arrows are pointing at some of these Si-Ge bar precipitates.

4 Figure 4 .
Figure 4. HR images of Al-Cu-Si-Ge alloy without pre-deformation (left) and after 8% deformed (right), in an early underaged condition, t a = 3 h.Images in the un-deformed alloy were obtained with the electron beam parallel to ZA = [001] Al while in the case of the pre-deformed specimen the electron beam was oriented parallel to [011] Al Zone Axis.(a) (b) General view, FFTs are calculated over the whole images and displayed within the top right corner.(c) (d) Detailed view of Si-Ge precipitates: the precipitate in (c) is oriented "end on" and is the same one enclosed in dotted lines in (a).The precipitate in (d) has its longest dimension lying parallel to [100] Al direction thus Moirée contrast can be observed.(e) (f) Detailed view of θ′ precipitates imaged "edge on", the precipitate in (f) corresponds to the area enclosed in dotted lines in (b).All the sub-images in (c)-(f) have a width of 10 nm and are displayed together with their FFT (these FFTs are calculated only in the small region were the precipitate is shown).Each FFT has a width of 18.5 nm −1 .
(a) and Figure 5(b), have been obtained in 2-beam condition with the g = 200 Al reflection from Al matrix at the Bragg condition, being close to the ZA = [011] Al .A

Figure 5 .
Figure 5. Microstructure of Al-Cu-Si-Ge alloy in a slightly underaged condition, t a = 10 h: (a) without pre-deformation, (b) after 8% pre-deformation.In both cases DF imaging was performed in a two-beam condition with (200) Al reflection close to ZA = [011] Al .The inclined θ ′ plates are highlighted using a reflection from (112)θ ′ planes.The yellow arrows are pointing out some notches on the θ ′ precipitates which indicate heteroge- neous precipitation of the θ ′ phase on the previously formed Si-Ge precipitates.
(b), every θ ′ precipitate in Figure 6(a) is highlighted by Fourier filtering.The image was obtained by calculating the Inverse Fast Fourier Transform (IFFT), using the spots circled in the corresponding FFT (Figure 6(b) inset).

Figure 6 (
Figure 6(c) and Figure 6(d) are DF images showing the peak aged microstructure (t a = 3 h) where many small θ ′′ precipitates are observed together with some large θ ′ precipitates.For the overaged condition (t a = 100 h), Fig- ure 6(e) and Figure 6(f) show a reduced amount of θ ′′ precipitates and some very large, sparsely distributed θ ′ precipitates.

Table 1 .
PA hardness and ageing time for non-deformed and pre-deformed specimens of each alloy.