Microstructural Evolution of AA5154 Layers Intermixed with Mo Powder during Electron Beam Wire-Feed Additive Manufacturing (EBAM)

Abstract

AA5154 aluminum alloy wall was built using EBAM where the wall’s top layers were alloyed by depositing and then remelting a Mo powder-bed with simultaneous transfer of aluminum alloy from the AA5154 wire. The powder-beds with different concentrations of Mo such as 0.3, 0.6, 0.9 and 1.2 g/layer were used to obtain composite AA5154/Mo samples. All samples were characterized by inhomogeneous structures composed of as-deposited AA5154 matrix with coarse unreacted Mo articles and intermetallic compounds (IMC) such as Al₁₂Mo, Al₅Mo, Al₈Mo₃, Al₁₈Mg₃Mo₂ which formed in the vicinity of these Mo particles. The IMC content increased with the Mo powder-bed concentrations. The AA5154 matrix grains away from the Mo particles contained Al-Fe grain boundary precipitates. Mo-rich regions in the 0.3, 0.6, 0.9 and 1.2 g/layer Mo samples had maximum microhardness at the level of 2300, 2600, 11,500 and 9000 GPa, respectively. Sliding pin-on-steel disk test showed that wear of A5154/Mo composite reduced as compared to that of as-deposited AA5154 due to composite structure, higher microhardness as a well as tribooxidation of Al/Mo IMCs and generation of mechanically mixed layers containing low shear strength Mo₈O₂₃ and Al₂(MoO₄)₃ oxides.

1. Introduction

The development of new high-strength and temperature-resistant metallic materials is an urgent task faced by specialists working in space, aircraft, and automobile industries. Nowadays, nickel-based, high-temperature strength alloys are widely applied for fabricating gas and steam turbine blades or exhaustive gas heat shields [1]. However, their working temperatures are less than those of refractory metal aluminides [2,3,4,5]. Molibdenum aluminides are attracting more and more attention of the researchers since they possess a combination of high melting temperature and thermal stability with high mechanical characteristics [6,7,8]. Their resistance to oxidizing is also very high because of the formation of passivation Al₂O₃ film [9].

Due to its low solubility in aluminum, molybdenum forms a number of hard intermetallic Al_xMo_y compounds (IMCs) that might have been used for reinforcing the aluminum alloy matrix if properly dispersed and distributed in it. Such a composite structure can be used for improving both mechanical and functional characteristics of the alloy.

Despite there has always been some constant interest in obtaining and studying the Al_xMo_y IMCs, their mechanical characteristics are still understudied and there are discrepancies in the experimental results reported by different researchers. For instance, Yihan Bian et al. [10] showed that MoAl₁₂ not only had a high melting point but also retained its strength under high-temperature loading. It was reported also [11] that MoAl₁₂, MoAl₅, MoAl₄, Mo₃Al₈ and Mo₃Al IMCs possess thermodynamic and dynamic stabilities, which are depended on the Mo content so that Mo₃Al₈ is the most stable IMC as compared to other ones, and bulk modulus of Al-Mo alloys tends to grow with the Mo content [11]. Mingliang Wang et al. [7] established that the calculated bulk elasticity and shear moduli of MoAl₅ were 115.7 GPa and 95.0 GPa, respectively. Eumann M. et al. [6] demonstrated that Mo₃Al had a yield stress of 1600 MPa and high brittleness. Rikiya Nino et al. [8] showed that coarse lamellar Mo₃Al–Mo₃Al₈ served to suppress crack nucleation in Vickers indentation at loads as high as 9.8 N, while fine lamellar structures fractured during indentation at 2.94 N load. Such a result seems to be surprising since common opinion is that the finer the structure the higher is the fracture toughness.

The processes used for synthesizing the Al_xMo_y IMCs mainly are as follows: laser powder-bed fusion [12], hot dipping of Mo rods into a commercial Al-12 wt.% Si alloy melt [13], stir casting [14,15], reactive hot-pressing [16], magnetron sputtering [17], friction stir processing [18,19], etc.

Additive manufacturing (AM) is an alternative to traditionally used methods of fabricating materials and articles. Small-scale AM machines equipped with laser, plasma arc, or electron beam energy sources are produced commercially to enable production of machine components directly at place of use from powders or wires. AM has potential for fabricating figurine-shaped components from in-situ formed metal matrix composites, functionally graded or polymetallic materials as depending on the source materials used. Progress achieved in the field of additive manufacturing and materials science allowed using the AM in machine building, aircraft or space applications industries. One of the most important examples is using the AM for fabricating a jet engine collector for Airbus [20].

Using the AM methods for in-situ preparation of metallic matrix composites or functionally graded materials requires resolving problems with respect to materials compatibility, reactive diffusion zone characteristics, improving as-cast structure characteristics, avoiding porosity, shrinkage pores, etc. Some of the above problems may be resolved with properly selecting the AM method.

Additive manufacturing methods used for making metallic components include wire arc additive manufacturing (WAAM) [21,22,23,24,25], selective laser melting (SLM) [26,27,28] and electron beam additive manufacturing (EBAM) [29,30,31,32]. These methods have their own advantages and disadvantages but the majority of experiments on making aluminum alloy matrix composites were carried out using these three.

For instance, Gasper et al. [33] used a combined powder/wire feed direct energy deposition for in-situ preparation of titanium aluminides. SLM on AlSi10Mg/SiC allowed obtaining AlSi10Mg/SiC composites [33] characterized by reduced friction and wear. Intermixing TiC particles with aluminum alloy during the WAAM resulted in grain refining and improved mechanical characteristics of the composites obtained [34,35]. Applicability of inoculating particles of oxides, carbides, metallic for improving the microstructure and mechanical characteristics of alloys was discussed in a number of review articles as follows: [36,37,38,39].

A combination of powder-bed and wire-feed EBAM processes allowed fabricating boron carbide/aluminum bronze composite [40]. Such a process can also be used for preparation of an Al/Mo composite reinforced by in-situ reactive diffusion formed IMCs. Such an approach was never used earlier in this system and can be of interest from the viewpoint of materials science.

The objective of this work is, therefore, to prepare Al/Mo composite layers using EBAM and study their microstructural and mechanical characteristics depending on the Mo concentration.

2. Materials and Methods

Electron beam deposition was carried out using an aluminum alloy of ⌀1.2 mm wire of composition, wt.%: 3.8 ± 0.4 Mg, 0.2 ± 0.01 Fe, 0.1 ± 0.02 Mn, 0.1 ± 0.01 Ti, 0.02 ± 0.003 Cu, 0.01 ± 0.002 Zn, and Al—balance, that corresponded to that of AA5154. A molybdenum powder bed was formed by spreading 98.7% purity Mo particles on the as-deposited AA51514. The Mo powder was composed of deagglomerated ~2 µm particles and 3–12 μm agglomerates.

EBAM parameters used for deposition of first eight AA5154 layers on a AA5154 substrate (Figure 1a, pos. 3), are shown in

Table 1. The additive process parameters.

Acceleration Voltage, kV	Beam Current, mA	Wire Feed Rate, mm/min	Table Speed, mm/min
30	40	1590	380

. These parameters were the same as those used earlier for growing the quality and defectless AA5356 walls [41].

On depositing these 8 layers, a powder feeder (Figure 1b, pos. 4) was used to form a molybdenum powder bed of the total weight of 0.3 g on the as-deposited metal. The next step was remelting the powder bed with simultaneous deposition of AA5154 from the wire. Such a cycle was then repeated twice so that the total number of the remelted powder beds was 3. Similar procedures were carried out for powder beds amounting to 0.6, 0.9, and 1.2 g/per a layer. The resulting walls had the dimensions as follows: 60 × 20 × 30 mm³.

The distribution of Mo in the as-deposited samples was studied using an X-ray tomograph Cheetah X-ray Inspection System EVO (YXLON International GmbH, Hamburg, Cermany). Metallographic examination of samples was carried out using optical microscope Altami Met 1S (LCC, Altami, St. Petersburg, Russia) and scanning electron microscope Tescan MIRA 3 LMU (TESCAN ORSAY HOLDING, Brno, Czech Republic), equipped with Oxford Instruments Ultim Max 40 EDS detector (Oxford Instruments, High Wycombe, UK). TEM examinations were performed using TEM instrument JEOL-2100 (JEOL Ltd., Akishima, Japan) on foils thinned by a focused ion beam.

X-ray diffraction (XRD) phase analysis was carried out by means of an X-ray diffractometer DRON-7 (Innovation Center “Burevestnik”, St. Petersburg, Russia) using CoKα radiation at 0.05° increments and an exposure time of 25 s/step that allowed detecting phases such as Al and Mo as well as their intermetallic compounds (IMCs). Grazing-incidence X-ray diffraction (GIAXD) at beam incidence angle of 10° was applied for studying phases formed in a thin subsurface layer by friction against the steel disk. The XRD peak identification and treatment was performed using Crystal Impact software “Match!” (version 3.9, Crystal Impact, Bonn, Germany). The phase contents were obtained from the XRD data according to Equation (1), where V_phase is the phase percentage, I_phase is the phase’s integral phase peak intensity –, and I_phase is the integral intensity of any other reflection.

$$V_{phase}=\left(\frac{\mathsf{\Sigma }(I_{phase})}{\mathsf{\Sigma }\left(I_{summ}\right)}\right)\times 100\%$$

(1)

Microhardness of the phases was measured using a Duramin5 (Struers A/S, Ballerup, Denmark) machine. A tensile test machine UTS-110M-100 (Testsystems, Ivanovo, Russia) and tribometer TRIBOtechnic (Tribotechnic, Clichy, France) were used for determining the sample strength and tribological behavior, respectively.

Pin-on-disk sliding was applied for measuring the friction and wear characteristics of samples. Samples in the form of 3 × 3 × 10 mm³ pins were cut off the AA5154/Mo zone, while Ø35 mm, 5 mm thickness counterbody disks were machined from 40 × 13 quenched steel. Rotation rate, normal load and sliding path length were 94 RPM, 15 N and 5600 m, respectively.

3. Results

Macrostructural cross section views of as-deposited AA5154 (zone 1) and AA5154/Mo (zone 2) demonstrate many semicircular fusion boundaries left after each track deposition according to the strategy selected, as shown in Figure 2.

The microstructure of as-deposited AA5154 (zone 1) is the same for all samples and represented by almost non-etched α-Al dendrites with numerous interdendrite precipitates. Zone 2 of all samples demonstrates both bright and dark-etched areas that allow suggesting inhomogeneous distribution of Mo powder in them (Figure 2). Coarse light particles and discontinuities can be observed in zone 2 of all samples, which could result from sintering the Mo powder agglomerates and diffusion reaction with Al.

Cross-section views of each deposition track show semicircular dark-etched bands with intermetallic compounds formed during the solidification of the melted pool.

X-ray tomography images in Figure 3 allow observing distribution of Mo in the cross section views of as-grown AA5154/Mo samples. The dark contrasting particles can be seen owing to their higher density as compared to the as-deposited A5154 layers and therefore reveal distribution of both unreacted Mo and Al-Mo IMCs formed from the Mo powder-bed. Samples obtained with 0.3 g/layer content of Mo show the most uniform dark-contrast phase distribution over the cross section view.

Typical needle-like IMC microstructures were found near the unreacted Mo-rich particles (Figure 4a,b) that precipitated from the solid solution during solidification. Areas darker than those of unreacted Mo ones can be observed in Figure 4b that were formed by dissolution of Mo particle in aluminum alloy α-Al with ensuing precipitation of Al/Mo IMCs.

3.1. Phase Detection with XRD

XRD patterns were obtained from cross section views of the AA5154/Mo samples (Figure 5a) that allowed detecting peaks of α-Al and unreacted Mo as well as those of Al/Mo IMCs identified as Al₁₂Mo, Al₅Mo, Al₈Mo₃, Al₁₈Mg₃Mo₂ (Figure 5a). Total amount of IMCs increased with the total Mo content (Figure 5b). The unreacted Mo content is reduced starting from 0.6 g/layer load when more IMCs are formed.

3.2. Microstructure of AA5154/Mo Composite

The zone 2 microstructure of samples intermixed with 0.3 wt.% Mo shows large 0.5–1 mm sized particles of Mo with about 5 at.% Al (Figure 6a,b,f,

Table 2. EDS analysis of the AA5154/Mo with 0.3 g/layer was obtained by using SEM (Figure 6b,d).

Spectrum	Concentration of Elements, at.%						Calculated Phase
	Al	Mo	Mg	Ti	Fe	Mn
1	5	95	-	-	-	-	Mo(Al)
2	69.3	30.7	-	-	-	-	Al8Mo3
3	35.6	64.4	-	-	-	-	Al8Mo3
4	82.2	17.8	-	-	-	-	Al5Mo
5	66.2	33.8	-	-	-	-	Al8Mo3
6	89.6	8.8	0.4	0.2	0.1	0.4	Al12Mo
7	87.7	10.1	0.4	0.2	0.5	-	Al12Mo
8	87.8	10.9	0.4	0.2	-	0.2	Al12Mo
9	98.9	0.3	0.7	0.1	-	-	α-Al

, spectrum 1). Transition areas can be seen between these Mo-rich particles and Al matrices that are filled with intermetallic dendritic structures composed of ~30 at.% Mo and ~70 at.% Al (Figure 6b). The atomic concentration Al/Mo ratio is close to that of the Al₈Mo₃ IMC that, according to Al-Mo diagram [42], exists within a wide concentration range 27–71 at.% Al and 27–77 at.% Mo.

Smaller particles (Figure 6b, Table 2, spectrums 3–5) contain more aluminum and less molybdenum so that their concentration ratio allows identifying them as Al₈Mo₃ and Al₅Mo. In addition to transition areas filled with dendritic IMCs, there are sharp interfaces between Mo-rich and α-Al-rich areas (Figure 6e,f). Almost homogeneously distributed needle-like and faceted 2–5 μm IMCs can be observed in the α-Al areas (Figure 6c,d). EDS shows that these precipitates contain 87–89 at.% Al, 8–11 at.% Mo and Mg, Mn, Ti, Fe as impurities (Figure 6d, Table 2, spectrums 6–8).

It should be noted that some Al-Mo precipitates in zone 2, α-Al, are characterized by the presence of Mg (Figure 7d), which allows suggesting the formation of Al₁₈Mg₃Mo₂ IMCs. The existence of these IMCs can be supported by TEM images and corresponding SAED patterns in Figure 7g as well as EDS (Figure 7,

Table 3. EDS analysis of the AA5154/Mo with 0.3 g/layer was obtained by using TEM (Figure 7g–j).

Spectrum	Concentration of Elements, at.%						Calculated Phase
	Al	Mo	Mg	Ti	Fe	Mn
1	76.8	8.3	14.9	-	-	-	Al18Mg3Mo2
2	87.6	-	0.8	1.4	9.9	0.3	Fe-containing phase
3	84.1	-	0.6	1.2	13.8	0.3	Fe-containing phase
4	98.2	-	1.8	-	-	-	α-Al
5	86.8	4.5	8.1	-	0.5	0.1	Al18Mg3Mo2
6	83.9	4.0	7.8	-	3.9	0.2	Fe-containing phase
7	98.5	0.4	1.1	-	-	0.3	α-Al

, spectrum 1). These faceted Al₁₈Mg₃Mo₂ particles are about 0.8 μm in size and localize on the α-Al grain boundaries together with the Fe-rich ones (Figure 7g–j, Table 3 spectrums 2, 3, 5, 6). The total concentration of Mo in the α-Al is not higher than 0.4 at.% (Figure 6d, Table 2, spectrum 9, Figure 7h, Table 3, spectrums 4, 7). The Fe-rich grain boundary precipitates are inherent in the Al-Mg aluminum alloys obtained using EBAM [26].

The layers deposited with 0.6 g/layer Mo show higher amounts of both ~1 μm and <100 nm Mo-rich particles as well as Al-Mo IMCs (Figure 8) in accordance to Figure 6b. EDS spectra 1 and 2 (

Table 4. EDS analysis of the AA5154/Mo with 0.6 g/layer was obtained by using SEM (Figure 8b–d).

Spectrum	Concentration of Elements, at.%						Calculated Phase
	Al	Mo	Mg	Ti	Fe	Mn
1	2.5	97.5	-	-	-	-	Mo(Al)
2	2.5	97.5	-	-	-	-	Mo(Al)
3	83.8	16.7	0.3	0.1	0.1	-	Al12Mo
4	82.1	17.6	0.2	0.2	0.1	-	Al12Mo
5	87.3	11.8	0.2	0.3	0.1	0.3	Al12Mo
6	86.7	12.6	-	0.3	0.1	0.3	Al12Mo
7	86.9	12.4	-	0.3	0.1	0.3	Al12Mo
8	92.5	6.7	0.1	0.1	0.2	0.4	α-Al(Mo)
9	93.2	6.5	-	0.1	0.2	-	α-Al(Mo)
10	86.8	13	-	0.1	0.1	-	Al12Mo
11	99.7	0.2	0.3	-	-	-	α-Al
12	99.4	0.2	0.3	0.1	-	-	α-Al

) show that Mo-rich particles contain up to 2.3 at.% Al and are encompassed by IMC needle-like and faceted precipitates (Figure 8a,b). The α-Al grains contain mostly faceted IMCs that correspond to Al₁₂Mo according to EDS data (Figure 8b–d, Table 4, spectrums 3–7, 10).

TEM shows that these precipitates are really Al₁₂Mo (Figure 9a–e,g). Another finding was that irregularly shaped IMCs were found in the α-Al regions which contained up to ~7 at.% Mo and <1 at% of impurities such as Fe, Ti, Mn, Mg (Table 4, spectrums 8, 9). The concentration of Mg in the α-Al regions is at the level of 0.3 at.% (Table 4, spectrums 11, 12). According to the TEM data there are also precipitates such as Al₁₈Mg₃Mo₂ and as well as the grain boundary precipitates containing up to 11–15 at.% Fe (Figure 9f,h).

For samples obtained with Mo loads of 0.9 and 1.2 g/layer, it becomes observable that the volume content of Mo-rich particles is reduced (Figure 10 and Figure 11) with simultaneous increasing of the areas occupied by Al-Mo IMCs (Figure 10d and Figure 11c,e). Mo-rich particles containing 1.6–7.6 at.% Al were detected (Figure 10d and Figure 11c,

Table 5. EDS analysis of the AA5154/Mo with 0.9 g/layer was obtained by using SEM (Figure 10d,f).

Spectrum	Concentration of Elements, at.%						Calculated Phase
	Al	Mo	Mg	Ti	Fe	Mn
1	7.6	92.4	-	-	-	-	Mo(Al)
2	49.5	50.5	-	-	-	-	Al8Mo3
3	61.9	38.1	-	-	-	-	Al8Mo3
4	72.6	27.4	-	-	-	-	Al17Mo4
5	86.7	12.1	-	-	-	-	Al12Mo
6	92.6	5.4	0.3	0.1	0.3	0.6	Al12Mo
7	87.8	10.0	0.5	0.3	0.1	0.4	Al12Mo
8	98.5		1.5	-	-	-	α-Al
9	98.2	0.1	1.6	-	-	0.1	α-Al

,

Table 6. EDS analysis of the AA5154/Mo with 1.2 g/layer was obtained by using SEM (Figure 11e,f).

Spectrum	Concentration of Elements, at.%		Calculated Phase
	Al	Mo
1	1.6	98.4	Mo(Al)
2	52.3	47.7	Al8Mo3
3	61.9	38.1	Al8Mo3
4	1.4	98.6	Mo(Al)
5	46.2	53.8	Al8Mo3
6	71.6	28.4	Al17Mo4
7	52.3	47.7	Al8Mo3
8	81.8	18.2	Al5Mo
9	43.1	56.9	Al8Mo3
10	65.5	33.5	Al8Mo3

and Table spectrum 1) that were encompassed with the light-grey Al₈Mo₃ shells (~46–62 at.% Al, ~38–54 at.% Mo) (Figure 10d and Figure 11c, Table 5, Table 6 and Table spectrums 2, 3). According to the Al-Mo diagram, this phase may exist in a wide concentration range so that increasing Al concentration by 20 at.% would not change the stoichiometry but only provide darker contrast in SEM BSE images. Al₁₇Mo₄ and Al₁₂Mo shells are formed in regions closer to α-Al ones (Figure 10d, Table 5, spectrums 4, 5).

Greater numbers of needle-like and faceted Al₁₂Mo IMCs are formed in samples charged with 0.9 g/layer Mo powder (Figure 10). Additionally, these particles are larger than those formed in samples with 0.3 and 0.6 g/layer.

Similar structures were observed in samples obtained with even higher concentration of Mo (1.2 g/layer) (Figure 11c,e, Table 6). Along with the Mo-rich multi-shell Al-Mo regions, there are large regions composed of eutectic-like Al₈Mo₃ structures (Figure 11f, Table 6). Coarse platelet- or needle-like Al₅Mo precipitates can be found near these IMC shells (Figure 11e, Table 6). The α-Al regions contain particles with up to 2 at.% of Mo as well as faceted Al₁₂Mo ones (Table 5, spectrums 8, 9).

Partial sedimentation of Mo particles on the melted pool bottom was observed during the deposition of AA5154 layer on the Mo powder bed so that the zone 1/zone 2 IMC boundary became clearly seen in Figure 11a. This IMC boundary contains 75.1–97.7 at.% Mo, 2.3–20 at.% Al, 11.1–14.42 at.% Mg and 0.7–0.9 at.% Fe and shows some cracking.

TEM images show the presence of Al₁₂Mo IMCs in samples obtained with Mo powder concentrations 0.9 and 1.2 g/layer (Figure 12a,b). Their composition coincides with those obtained from EDS on SEM images (Table 5 and Table 6). Grain boundary Fe-rich particles were also identified in the TEM images (Figure 12c).

3.3. Microhardness of AA5154/Mo

Microhardness number vertical profiles were obtained on the cross-section views, as shown in Figure 13a–d, at a distance of 100 μm between the lines. The microhardness number distribution along the corresponding lines revealed sharp changes at the matrix/IMC interfaces. Mo-rich regions in the 0.3, 0.6, 0.9 and 1.2 g/layer Mo samples had maximum microhardness at the level of 2300, 2600, 11,500 and 9000 GPa, respectively (Figure 13a–d). The microhardness of these Mo-rich regions in the last two samples was much greater than those in the first and second ones (Figure 13c,d). Such a result may be explained by indenting those wide Al-Mo IMC shells formed around the Mo-rich particles. The microhardness of α-Al solid solution in as-deposited AA5154 (zone 1) was higher than that of zone 2 with IMCs due to depletion by Mg which diffused into Al₁₈Mg₃Mo₂ IMCs.

3.4. Sliding Wear A5154/Mo

All samples had their coefficient of friction (CoF) vs. time dependencies as shown in Figure 14. All CoFs mean values lie within the 0.45–0.55 range including that of the base AA5154. Somewhat lower CoF values are demonstrated by samples obtained with the maximal 1.2 g/layer input of Mo powder.

The maximum wear was detected in sliding the base alloy against the steel disk (Figure 15). The in situ formed Al/Mo IMCs allowed reducing the wear so that its minimum value was shown by samples obtained with the maximal 1.2 g/layer input of Mo powder.

The worn surface EDS analysis shows the presence of Mo-rich areas as well as Fe-rich areas in addition to those containing aluminum. Intense tribooxidation of all samples occurred in all samples during sliding wear against a steel disk.

Mo-rich areas (Figure 16) are generated by deformation, adhesion transfer and smearing of Mo-rich IMCs during sliding, while Fe-rich ones appeared because of reverse transfer from the steel counterbody (Figure 17).

Oxide phases such as Fe₃O₄, non-stoichiometric Mo₈O₂₃ and mixed Al₂(MoO₄)₃ have been detected on the worn surfaces of samples by means of GIAXD (Figure 18) to the exclusion of samples with an initial Mo concentration of 0.6 g/layer where no Al₂(MoO₄)₃ was present. Such a finding may be the result of selective wear of the tribooxidized aluminum matrix (Figure 19a). Semi-quantitative analysis of GIAXD peaks of Fe₃O₄, Mo₈O₂₃ and Al₂(MoO₄)₃ oxides showed that their amount increased from 15 to 24 vol.% in a 10–15 μm thickness subsurface layer for samples with initial Mo concentrations of 0.3–1.2 g/layer.

It can be observed from Figure 18 that free molybdenum was found only on the worn surface of the 0.9 g/layer sample, which also showed the maximum wear among all the AA5154/Mo samples (Figure 15). The worn surface of this sample demonstrates large, damaged light-grey areas formed by the removal of a mechanically mixed layer (Figure 19b).

It looks obvious that coarse unoxidized Mo particles underlying the mechanically mixed oxide layer make their way to the surface, thus becoming easily detected with the GIAXD. Fewer damaged areas are observed on the worn surfaces of all other AA5154 g/layer samples which are covered by mechanically mixed layer areas composed of Al/Mo IMCs tribooxidation products such as Mo₈O₂₃ and Al₂(MoO₄)₃.

The slight increase in wear for samples with initial concentrations of Mo 0.3, 0.6 and 0.9 g/layer can be explained by the insufficient amount of Mo₈O₂₃ and Al₂(MoO₄)₃ oxides formed but an increasing amount of brittle Al/Mo IMCs (Figure 5b) [43].

4. Discussion

Molybdenum has low diffusion factor, rather low solubility in aluminum and is capable of forming a row of hard Al_xMo_y IMCs. The results of this work allowed showing that changing the amount of Mo charged into the aluminum matrix at a step as low as 0.3 g resulted in a notable increase in the volume fraction of IMCs in the aluminum alloy matrix. Both Mo-rich and Al-rich IMCs were formed depending on the distribution of the Mo powder particles in the melted pool. The latter ones, such as Al₁₂Mo and Al₁₈Mg₃Mo_2, were detected using XRD, EDS and partially TEM. Let us note that many IMCs were pulled out of the matrix during sample preparation procedures and therefore had never been analyzed. The morphology of Al₁₂Mo IMCs changed depending on the concentration differences and solidification front atomic morphology [44] that is determined according to the expression as follows:

$$\alpha =\frac{\Delta S_f}{R}$$

(2)

where α is the fusion entropy, ΔS_f is the specific fusion entropy per mole, R is the universal gas constant. The unfaceted crystals grow at α < 2, while values higher than α > 2 facilitate the formation of crystalline faces [44].

The IMC phase formation is determined by Gibbs energy change during the diffusion reaction between Al and Mo as follows: ($\Delta G^0$) [45]:

$$\Delta G^0=\Delta H^0-T\Delta S^0$$

(3)

where $\Delta H^0$ is the change in enthalpy; T is the reaction temperature; $\Delta S^0$ is the change in entropy that can be neglected in this solid state case to allow rewriting Equation (3) as follows: $\Delta G^0$ ≈ $\Delta H^0$.

For the Al-Mo system, the probability of an IMC formation by diffusion reaction can be explained using several adopted models [45,46]. Pretorius R. et al. [45] proposed a model for determining the effective heat of formation ($\Delta H_{}^{\prime }$) on the binary metal/aluminum phase interface:

$$\Delta H_i^{\prime }=\Delta H_i^0·\frac{C_e}{C_c}$$

(4)

where $\Delta H_i^{\prime }$ is the effective heat of formation of the i-phase, C_e is the effective concentration of the limiting element at the interface, and C_c is the concentration of the limiting element in the compound. Assuming that $\Delta G^0$ ≈ $\Delta H^0$, the effective Gibbs energy change in the case of forming the i-phase in the Al/Mo layer may be determined as follows:

$$\Delta G_i^{\prime }=\Delta G_i^0·\frac{C_e}{C_c}$$

(5)

Table 7. Effective Gibbs free energy changes calculated for all Al/Mo phases in AA5154/Mo at 2073 K.

Compound	Composition	$\mathbf{\Delta }{\mathit{G}}_{\mathit{i}}^{},\mathbf{J}/\mathbf{mol}$	$\mathbf{\Delta }{\mathit{G}}_{\mathit{i}}^{} \mathbf{for}2073\mathbf{K},\mathbf{kJ}/\mathbf{mol}$	$\mathbf{\Delta }{\mathit{G}}_{0.3}^{\prime },\mathbf{kJ}/\mathbf{mol}$	$\mathbf{\Delta }{\mathit{G}}_{0.6}^{\prime },\mathbf{kJ}/\mathbf{mol}$	$\mathbf{\Delta }{\mathit{G}}_{0.9}^{\prime },\mathbf{kJ}/\mathbf{mol}$	$\mathbf{\Delta }{\mathit{G}}_{1.2}^{\prime },\mathbf{kJ}/\mathbf{mol}$
Al12Mo	Al0.923Mo0.077	−10,700  +  1.865 T	–6.8338	–5.97	–3.70	–5.72	–6.83
Al5Mo	Al0.833 Mo0.167	−23,000  +  4.381 T	–13.9182	–6.02	–6.41	–6.41	–5.88
Al17Mo4	Al0.810 Mo0.190	−27,200  +  4.8 T	–16.8350	–6.62	–6.62	–4.84	–4.67
Al8Mo3	Al0.728 Mo0.272	−35,470  +  7.12 T	–20.7102	–3.71	–3.93	–4.19	–2.96
AlMo3	Al0.25Mo0.75	−25,250  + 3 T	–19.031	–1.95	–1.95	–1.95	–1.95

contains the effective Gibbs free energy changes calculated for all Al/Mo phases.

It follows from Table 7 that all Al_xMo_y phases have negative values of the Gibbs free energy changes and therefore can be formed in our experimental conditions. The minimal values belong to the Al-rich ones detected in the experiment. Al-Mo solid solutions may also form in the Al-rich areas

Inhomogeneous distribution of Mo particles and brittle Al/Mo IMCs, as well as the presence of defects such as voids, resulted in non-uniform distribution of the microhardness numbers. The friction and wear behavior of samples showed reduced friction and wear for all the AA5154/Mo samples as compared to that of AA5154. The IMCs allowed increasing the worn subsurface hardness and generating mechanically mixed layers (MML) consisting of Fe₃O_4, Mo₈O₂₃ and Al₂(MoO₄)₃ oxides, which formed due to tribooxidation of Al_xMo_y IMCs and adhesive transfer of iron from the steel counterbody. These oxides may be related to easy shear compounds that provide a lubrication effect and improve the tribological characteristics of composites [47]. It is also known that [48] introducing Mo into ceramic coatings served not only for their enhanced hardness and wear resistance, but also reduced friction due to the formation of MoO₃ on the worn surfaces.

However, there was another factor that served for some minor increase in wear for samples containing from 0.3 to 0.9 g/layer of Mo. This factor was the bearing ability of the MML depending on the number of brittle IMCs in the subsurface of the samples, which were pulled out and abraded the worn surface. However, this did not occur with the 1.2 g/layer sample where the maximal amount of self-lubrication oxides was formed to reduce the abrasion intensity.

5. Conclusions

Combination of wire-feed and powder-bed electron beam additive manufacturing was used for intermixing Mo powder into melted AA5154. Different amounts of Mo powder were used to form the powder-beds that were remelted together with previously deposited AA5154 layers to result in inhomogeneous distribution of coarse Mo particles in the AA5154 matrix. Intermetallic Al/Mo compounds were found near these particles in the form of needle-like precipitates. The intermixed zone structure was composed of AA5154 matrix with unreacted Mo particles and Al/Mo intermetallic precipitates whose amount depended on the amount of Mo powder introduced. The resulting structures possessed high microhardness and reduced sliding wear. The worn surface analysis showed the presence of mechanically mixed layers composed of wear particles and Fe₃O₄, Mo₈O₂₃ and Al₂(MoO₄)₃ oxides which could be the reason for reduced wear of the composite materials as compared to that of the as-deposited AA5154.