Influence of Milling Atmosphere on the Controlled Formation of Ultrafine Dispersoids in Al-Based MMCs

Properties of compacts made from aluminium powder, milled under different atmospheres, were evaluated. The duration of all the milling processes was 10 h, although different atmospheres were tested: vacuum, confined ammonia, and vacuum combined with a short-time ammonia gas flow (5 min). Milled powders were consolidated by cold uniaxial pressing and vacuum sintering.The nature and content of the second phases change with the milling atmosphere, allowing the modification of the mechanical properties of the compacts. Results showed that hardness and tensile strength were highly dependent on the milling atmosphere. Milling carried out in vacuum with a short-time ammonia gas flow notably improved compacts’ properties, as compared with 10 h vacuum milling. Hardness increased from 96 to 150 HB, and ultimate tensile strength rose from 247 to 476 MPa.


Introduction
Weight saving is a main aim in the automotive and aerospace industries.Increasing the use of lightweight alloys (e.g., aluminium) leads to a significant reduction in the overall weight, fuel consumption, and pollution of automobiles and aircrafts [1][2][3][4].Aluminium has a very low density (2.7 g/cm 3 ), high specific strength, excellent low temperature performance, good formability, and exceptional corrosion resistivity [5,6].Despite their interesting properties, commercial heat-treated aluminium alloys suffer a remarkable degradation of properties at high temperature.Traditionally, the strengthening of aluminium alloys has been obtained by precipitation hardening, resulting in a homogeneous dispersion of fine intermetallic particles [7].However, these particles are not able to avoid the degradation of mechanical properties at high temperatures [8], since their metallic nature promotes a rapid grain growth and an overaging of the alloy.To overcome these problems, aluminium alloys can be reinforced by hard ceramic particles to produce aluminium matrix composites (AMCs).
AMCs are low-weight and high-performance materials that have attracted considerable attention in many advanced applications in railway transport systems, the aerospace industry, and in armoured vehicles [9,10].Owing to their excellent physical and mechanical properties, AMCs have the potential to replace conventional materials in the automobile and aircraft industries.The properties of AMCs can be tailored by changing the fraction, size, and type of reinforcing particles [11][12][13][14]; it is possible to add oxides (Al 2 O 3 or Y 2 O 3 ), carbides (SiC or TiC), nitrides (AlN), and borides (TiB 2 or B 4 C), among many others [15][16][17][18].
Obtaining AMCs with excellent properties requires the successful incorporation of ceramic particles with a good bonding with the aluminium matrix.Several methods in solid and liquid states have been developed for the dispersion of fine particles [19][20][21][22].In liquid methods, the entire aluminium matrix is melted, and the ceramic particles are gradually introduced into the molten aluminium by stirring before casting.These methods are widely used due to their simplicity, although the resulting distribution is generally inhomogeneous.Conversely, the solid-state route ensures a uniform distribution of reinforcement particles in the matrix.In this regard, the powder metallurgy (PM) process allows aluminium powder and reinforcement particles to be mixed at room temperature.In the powder metallurgy field, aluminium has traditionally been reinforced with directly-added ceramic particles [23][24][25][26][27].However, another way to strengthen aluminium alloys by a homogeneous distribution of fine particles has been developed.Thus, it is known that mechanical alloying (MA) allows solid-gas reactions to take place between powder and milling atmosphere at room temperature, or at least at much lower temperatures than are normally required [28][29][30].
The aim of this work is to develop a new aluminium base material with better mechanical properties than commercial alloys.To attain this objective, as-received aluminium powder (AR Al) was milled in a combined atmosphere of vacuum and a short-time ammonia gas flow.This combination was used in order to achieve precise control over the dispersion of fine refractory formed particles (mainly nitrides).After milling, aluminium matrix composites (AMCs) were produced by uniaxial cold pressing and vacuum sintering

Experimental Procedure
The starting material (AR Al) was atomized elemental aluminium powder (AS 61, Eckart) with a purity level higher than 99.7% and a mean particle size of 80.5 µm.A high-energy attritor ball-mill was used to process the aluminium powder for 10 h.Different atmospheres were tested: vacuum (V), confined ammonia (A), and vacuum followed by 5 min ammonia gas flow (AF-5).In this last case, ammonia was incorporated after 2 h of milling, and then millings continued in vacuum until reaching 10 h.The ammonia gas purity was higher than 99.96% (Air Liquide).The water-cooled stainless steel vessel had a capacity of 1400 cm 3 .Etilen bis-stereamide (EBS) micropowder organic wax (3 wt %) was used as a process control agent (PCA) during milling.Mill charge contained 72 g of powder and 3600 g of balls (charge ratio 50:1).All millings were carried out with a rotor speed of 500 rpm, at room temperature, for a total time of 10 h.
After powder milling, granulometry and morphology were analysed by laser diffraction (Mastersizer 2000, Malvern, Worcestershire, UK) and scanning electron microscopy (SEM XL-30, Philips, Amsterdam, The Netherlands), respectively.Powders were consolidated by cold uniaxial pressing (850 MPa, 2 min hold time, using EBS wax as die-wall lubricant) and vacuum (5 Pa) sintering (650 • C, 1 h), followed by furnace cooling.The sintering ramp used was: 10 • C/min up to 600 • C, followed by 5 • C/min up to 645 • C, and finally, 1 • C/min up to 650 • C, which was kept for 1 h.Both cylindrical (diameter: 12 mm; mass: ca. 4 g) and "dog-bone" shape [31] compacts (32 mm × 4 mm × 4 mm) were produced, and 15 samples of each geometry were tested for the three materials studied.X-ray diffraction analysis (XRD, Bruker D8 Advance, CuKα radiation) was carried out on powders, and also on compacts, to identify the phases formed both before and during sintering.The microstructure of the sintered compacts were also studied by transmission electron microscopy (TEM Talos F200, FEI, Hillsboro, OR, USA).Compacts' density, hardness (Emco-M4U-025), and tensile strength and elongation (Instron 5505) were evaluated.

Results and Discussion
The densification ability of milled powders under different atmospheres was firstly studied.It is well known in the powder metallurgy field that densification ability is strongly influenced by different factors: powder hardness, particle morphology, mean particle size, as well as their size distribution.The last two factors were determined from granulometric characterization.Because of the relatively long milling (10 h), all milled powders exhibited a similar granulometric distribution, with a mean particle size around 20 µm (Figure 1). the relatively long milling (10 h), all milled powders exhibited a similar granulometric distribution, with a mean particle size around 20 μm (Figure 1).On the other hand, particle morphology was studied by SE-SEM (secondary electrons SEM).Results show that all milled powders tended to an equiaxial morphology.Figure 2 collects SEM micrographs corresponding to vacuum (V) milled powders and powders milled in confined NH3 (A), that of the AF-5 powder being very similar.The combination of a reduced particle size together with an equiaxial geometry suggests that a high relative density can be achieved.In this way, a compressibility test (Figure 3a) was carried out in order to measure the relative green density versus the applied compaction pressure, as well as to indirectly estimate the hardness of the powders; the lower the relative density, the higher the powder hardness.Results show that densification ability is almost the same for powders milled in vacuum or in ammonia (confined or in flow).Thus, all powders reach relative densities above 90% for compaction pressures of about 850 MPa or higher.Since all milling experiments were carried out for the same total time of ten hours, the strain hardening of the aluminium powders should be the same in all of them.Therefore, the presence of ammonia gas during the milling process only provoked a slight hardening of aluminium powders with respect to those that are mechanically alloyed in vacuum.This small increase in hardening is due to nitrogen incorporation into the aluminium lattice in solid solution [32].On the other hand, particle morphology was studied by SE-SEM (secondary electrons SEM).Results show that all milled powders tended to an equiaxial morphology.Figure 2 collects SEM micrographs corresponding to vacuum (V) milled powders and powders milled in confined NH 3 (A), that of the AF-5 powder being very similar.The combination of a reduced particle size together with an equiaxial geometry suggests that a high relative density can be achieved.the relatively long milling (10 h), all milled powders exhibited a similar granulometric distribution, with a mean particle size around 20 μm (Figure 1).On the other hand, particle morphology was studied by SE-SEM (secondary electrons SEM).Results show that all milled powders tended to an equiaxial morphology.Figure 2 collects SEM micrographs corresponding to vacuum (V) milled powders and powders milled in confined NH3 (A), that of the AF-5 powder being very similar.The combination of a reduced particle size together with an equiaxial geometry suggests that a high relative density can be achieved.In this way, a compressibility test (Figure 3a) was carried out in order to measure the relative green density versus the applied compaction pressure, as well as to indirectly estimate the hardness of the powders; the lower the relative density, the higher the powder hardness.Results show that densification ability is almost the same for powders milled in vacuum or in ammonia (confined or in flow).Thus, all powders reach relative densities above 90% for compaction pressures of about 850 MPa or higher.Since all milling experiments were carried out for the same total time of ten hours, the strain hardening of the aluminium powders should be the same in all of them.Therefore, the presence of ammonia gas during the milling process only provoked a slight hardening of aluminium powders with respect to those that are mechanically alloyed in vacuum.This small increase in hardening is due to nitrogen incorporation into the aluminium lattice in solid solution [32].In this way, a compressibility test (Figure 3a) was carried out in order to measure the relative green density versus the applied compaction pressure, as well as to indirectly estimate the hardness of the powders; the lower the relative density, the higher the powder hardness.Results show that densification ability is almost the same for powders milled in vacuum or in ammonia (confined or in flow).Thus, all powders reach relative densities above 90% for compaction pressures of about 850 MPa or higher.Since all milling experiments were carried out for the same total time of ten hours, the strain hardening of the aluminium powders should be the same in all of them.Therefore, the presence of ammonia gas during the milling process only provoked a slight hardening of aluminium powders with respect to those that are mechanically alloyed in vacuum.This small increase in hardening is due to nitrogen incorporation into the aluminium lattice in solid solution [32].On the other hand, Figure 3b shows the relative density of the different compacts, both after being pressed (850 MPa) and after being sintered (650 °C, 1 h).As can be observed, the relative density of green compacts is almost the same (90%), which is congruent with the compressibility curves.After sintering, the relative density of V compacts decreased, reaching a value of 95%.On the contrary, a relative density close to 100% was reached in sintered compacts from powders milled under ammonia gas (AF-5 and A).Therefore, the ability to densify is improved by the presence of ammonia gas.Thus, a high level of porosity (10%) is eliminated during sintering in compacts from powders milled in ammonia, compared vacuum milled powder compacts.
Obviously, the effect of ammonia gas on the milled powder must also affect to the mechanical behaviour of compacts made from them.In this way, tensile behaviour (ultimate tensile strength, UTS, and elongation, E) and hardness (HB) were evaluated on "dog-bone" and cylindrical shape compacts, respectively.
Figure 4 shows an increase of UTS values from 302 MPa, corresponding to V compacts, up to 476 and 489 MPa for AF-5 and A compacts, respectively.Despite a slightly higher UTS being reached with confined ammonia (A), the results corresponding to elongation show that it is more effective to incorporate a short-time ammonia gas flow during milling in vacuum.In any case, both UTS and elongation values are higher when milling is carried out in the presence of ammonia gas (AF-5 and A).This can be due to two factors: firstly, the low porosity achieved after sintering (Figure 3b) in compacts from powders milled in ammonia reduces tension concentrators, so that UTS is improved whereas elongation is kept or even increases; and on the other hand, the possible formation of second phases in the aluminium matrix as result of solid solution hardening from NH3 gas may have a remarkable influence on mechanical properties.Effectively, TEM images corresponding to V and AF-5 sintered compacts (Figure 5) show the presence of nanosized phases finely dispersed in the matrix.However, results indicate that the strengthening is more efficient when compacts are made from powder milled under ammonia gas (AF-5 and A) compared to compacts powder milled in vacuum (V).On the other hand, Figure 3b shows the relative density of the different compacts, both after being pressed (850 MPa) and after being sintered (650 • C, 1 h).As can be observed, the relative density of green compacts is almost the same (90%), which is congruent with the compressibility curves.After sintering, the relative density of V compacts decreased, reaching a value of 95%.On the contrary, a relative density close to 100% was reached in sintered compacts from powders milled under ammonia gas (AF-5 and A).Therefore, the ability to densify is improved by the presence of ammonia gas.Thus, a high level of porosity (10%) is eliminated during sintering in compacts from powders milled in ammonia, compared vacuum milled powder compacts.
Obviously, the effect of ammonia gas on the milled powder must also affect to the mechanical behaviour of compacts made from them.In this way, tensile behaviour (ultimate tensile strength, UTS, and elongation, E) and hardness (HB) were evaluated on "dog-bone" and cylindrical shape compacts, respectively.
Figure 4 shows an increase of UTS values from 302 MPa, corresponding to V compacts, up to 476 and 489 MPa for AF-5 and A compacts, respectively.Despite a slightly higher UTS being reached with confined ammonia (A), the results corresponding to elongation show that it is more effective to incorporate a short-time ammonia gas flow during milling in vacuum.In any case, both UTS and elongation values are higher when milling is carried out in the presence of ammonia gas (AF-5 and A).This can be due to two factors: firstly, the low porosity achieved after sintering (Figure 3b) in compacts from powders milled in ammonia reduces tension concentrators, so that UTS is improved whereas elongation is kept or even increases; and on the other hand, the possible formation of second phases in the aluminium matrix as result of solid solution hardening from NH 3 gas may have a remarkable influence on mechanical properties.Effectively, TEM images corresponding to V and AF-5 sintered compacts (Figure 5) show the presence of nanosized phases finely dispersed in the matrix.However, results indicate that the strengthening is more efficient when compacts are made from powder milled under ammonia gas (AF-5 and A) compared to compacts powder milled in vacuum (V). Figure 6 shows that reflections corresponding to aluminium carbide (Al4C3) and aluminium oxide (Al2O3) were detected in all sintered samples.However, reflections corresponding to nitrogenrich second phases (Al4CON and Al5O6N) were also observed in compacts powder milled with ammonia gas (AF-5 and A).During the milling process, aluminium particles react with the carbon and oxygen coming from the organic wax used as PCA, as well as with the nitrogen arising from the ammonia gas dissociation, with only aluminium peaks being detected in as-milled powders because a solid solution is formed.In this way, the Al4C3, Al3CON, and Al5O6N phases are formed during the subsequent sintering step [32].On the other hand, the presence of aluminium oxide (Al2O3) comes from the oxide layers covering the surface of aluminium particles, while hydrogen is lost during the process.Figure 6 shows that reflections corresponding to aluminium carbide (Al4C3) and aluminium oxide (Al2O3) were detected in all sintered samples.However, reflections corresponding to nitrogenrich second phases (Al4CON and Al5O6N) were also observed in compacts powder milled with ammonia gas (AF-5 and A).During the milling process, aluminium particles react with the carbon and oxygen coming from the organic wax used as PCA, as well as with the nitrogen arising from the ammonia gas dissociation, with only aluminium peaks being detected in as-milled powders because a solid solution is formed.In this way, the Al4C3, Al3CON, and Al5O6N phases are formed during the subsequent sintering step [32].On the other hand, the presence of aluminium oxide (Al2O3) comes from the oxide layers covering the surface of aluminium particles, while hydrogen is lost during the process.Figure 6 shows that reflections corresponding to aluminium carbide (Al 4 C 3 ) and aluminium oxide (Al 2 O 3 ) were detected in all sintered samples.However, reflections corresponding to nitrogen-rich second phases (Al 4 CON and Al 5 O 6 N) were also observed in compacts powder milled with ammonia gas (AF-5 and A).During the milling process, aluminium particles react with the carbon and oxygen coming from the organic wax used as PCA, as well as with the nitrogen arising from the ammonia gas dissociation, with only aluminium peaks being detected in as-milled powders because a solid solution is formed.In this way, the Al 4 C 3 , Al 3 CON, and Al 5 O 6 N phases are formed during the subsequent sintering step [32].On the other hand, the presence of aluminium oxide (Al 2 O 3 ) comes from the oxide layers covering the surface of aluminium particles, while hydrogen is lost during the process.
In a similar way to UTS values, hardness is also expected to increase because of the incorporation of ammonia gas to the milling atmosphere.As observed in Figure 7, hardness values as high as 150 HB were obtained when ammonia gas was involved in mechanical alloying (AF-5 and A), being lower than 100 HB when milling was carried out only in vacuum.Furthermore, the hardness improvement was more effective in the case where a short-time ammonia gas flow was used, instead of a confined ammonia atmosphere.It has been proven that the ability to incorporate nitrogen to the aluminium lattice is much higher when milling starts in vacuum [33].This high reactivity is also revealed by the fact that AF-5 and A compacts reach almost the same UTS and hardness in spite of the considerable difference in time under ammonia gas.Moreover, ammonia flow implies a continuous supply of renewed gas, contrary to the confined ammonia atmosphere.Thus, compact hardness can be modified by controlling the atmosphere during mechanical alloying.In a similar way to UTS values, hardness is also expected to increase because of the incorporation of ammonia gas to the milling atmosphere.As observed in Figure 7, hardness values as high as 150 HB were obtained when ammonia gas was involved in mechanical alloying (AF-5 and A), being lower than 100 HB when milling was carried out only in vacuum.Furthermore, the hardness improvement was more effective in the case where a short-time ammonia gas flow was used, instead of a confined ammonia atmosphere.It has been proven that the ability to incorporate nitrogen to the aluminium lattice is much higher when milling starts in vacuum [33].This high reactivity is also revealed by the fact that AF-5 and A compacts reach almost the same UTS and hardness in spite of the considerable difference in time under ammonia gas.Moreover, ammonia flow implies a continuous supply of renewed gas, contrary to the confined ammonia atmosphere.Thus, compact hardness can be modified by controlling the atmosphere during mechanical alloying.As discussed above, the hardness of a PM sintered compact is influenced by the compact porosity and the hardness of the constituent particles.Compacts prepared with AF-5 and A powders reached a relative density near 100% after being sintered.Therefore, the hardness of such compacts is only due to the hardness of the powder particles.The aforementioned strain hardening is the same for all millings, so hardness and UTS differences must be produced by the presence of hard phases dispersed in the compacts, which are formed during sintering.Regarding the latter, XRD patterns were fitted via Rietveld refinement [34] to quantify the phases formed after sintering.An increase in the weight percentage of dispersoids from 12% up to 20% when ammonia gas was used during  In a similar way to UTS values, hardness is also expected to increase because of the incorporation of ammonia gas to the milling atmosphere.As observed in Figure 7, hardness values as high as 150 HB were obtained when ammonia gas was involved in mechanical alloying (AF-5 and A), being lower than 100 HB when milling was carried out only in vacuum.Furthermore, the hardness improvement was more effective in the case where a short-time ammonia gas flow was used, instead of a confined ammonia atmosphere.It has been proven that the ability to incorporate nitrogen to the aluminium lattice is much higher when milling starts in vacuum [33].This high reactivity is also revealed by the fact that AF-5 and A compacts reach almost the same UTS and hardness in spite of the considerable difference in time under ammonia gas.Moreover, ammonia flow implies a continuous supply of renewed gas, contrary to the confined ammonia atmosphere.Thus, compact hardness can be modified by controlling the atmosphere during mechanical alloying.As discussed above, the hardness of a PM sintered compact is influenced by the compact porosity and the hardness of the constituent particles.Compacts prepared with AF-5 and A powders reached a relative density near 100% after being sintered.Therefore, the hardness of such compacts is only due to the hardness of the powder particles.The aforementioned strain hardening is the same for all millings, so hardness and UTS differences must be produced by the presence of hard phases dispersed in the compacts, which are formed during sintering.Regarding the latter, XRD patterns were fitted via Rietveld refinement [34] to quantify the phases formed after sintering.An increase in the weight percentage of dispersoids from 12% up to 20% when ammonia gas was used during As discussed above, the hardness of a PM sintered compact is influenced by the compact porosity and the hardness of the constituent particles.Compacts prepared with AF-5 and A powders reached a relative density near 100% after being sintered.Therefore, the hardness of such compacts is only due to the hardness of the powder particles.The aforementioned strain hardening is the same for all millings, so hardness and UTS differences must be produced by the presence of hard phases dispersed in the compacts, which are formed during sintering.Regarding the latter, XRD patterns were fitted via Rietveld refinement [34] to quantify the phases formed after sintering.An increase in the weight percentage of dispersoids from 12% up to 20% when ammonia gas was used during milling was observed (Figure 7).Furthermore, the amount of oxynitride and, mainly, oxycarbonitride aluminium (Al 5 O 6 N and Al 3 CON, respectively) increased with the amount of time that ammonia gas was incorporated to the milling process, while that of aluminium carbide (Al 4 C 3 ) decreased compared to vacuum milling (V).Likewise, the increase of the weight percentage of Al 3 CON was much faster if ammonia gas was flowed rather than under confined conditions.It can be observed that only five minutes of ammonia flow allowed a weight percentage of second phases as high as in confined conditions during ten hours to be achieved.
Therefore, ammonia gas allows the formation of significant amounts of nanocrystalline dispersoids, having the previously described remarkable influence on the mechanical properties.Thereby, a better combination of mechanical properties is achieved by using a short-time ammonia flow (AF-5) instead of confined conditions (A).In this regard, it is important to note that the mechanical properties reached are higher than those of the widely-used commercial alloys 2024-T4 (469 MPa and 120 HB) and 6061-T6 (310 MPa and 95 HB) at room temperature [35].That is, when ammonia gas is incorporated as short-time flow instead of as confined atmosphere, the mechanical properties are similar, despite the fact that time under such gas is shorter.

Conclusions
Aluminium compacts made from powders mechanically alloyed in vacuum, confined ammonia, or vacuum in combination with a short-time ammonia gas flow were characterized.
The following conclusions are drawn: (1) Ammonia gas improves the ability of powders to densify during sintering, reaching a density near 100%.
(2) Compacts from powder milled in vacuum are only reinforced by Al 4 C 3 and Al 2 O 3 .However, milling with ammonia gas allows the formation of significant amounts of other nanocrystalline dispersoids (Al 4 CON and Al 5 O 6 N), having a high influence on the improved mechanical properties of sintered compacts.
(3) A remarkable improvement on mechanical properties is achieved when a mechanical alloying process is carried out with ammonia, both confined and with a short-time flow.Furthermore, it has been noted that a short-time ammonia gas flow is more efficient than a confined atmosphere, since a deeper control of the mechanical properties may be achieved.
(4) The use of an ammonia gas flow of 5 min allows hardness and UTS values as high as with a confined ammonia atmosphere to be reached (150 HB and 476 MPa), despite the fact that ammonia gas consumption is much lower (0.3 L for AF-5 versus 1.85 L for A).Furthermore, under flowing gas, a greater elongation is achieved.

Figure 3 .
Figure 3. (a) Compressibility curves of mechanically alloyed aluminium powders and (b) relative density after being pressed and sintered.

Figure 3 .
Figure 3. (a) Compressibility curves of mechanically alloyed aluminium powders and (b) relative density after being pressed and sintered.

Figure 6 .
Figure 6.X-ray diffraction patterns of mechanically alloyed Al powders after sintering.

Figure 7 .
Figure 7. Hardness and weight percentage of the second phases formed in the different sintered samples.

Figure 6 .
Figure 6.X-ray diffraction patterns of mechanically alloyed Al powders after sintering.

Figure 6 .
Figure 6.X-ray diffraction patterns of mechanically alloyed Al powders after sintering.

Figure 7 .
Figure 7. Hardness and weight percentage of the second phases formed in the different sintered samples.

Figure 7 .
Figure 7. Hardness and weight percentage of the second phases formed in the different sintered samples.