Experimental and Theoretical Study of Ultra-Hard AlMgB₁₄-TiB₂ Composites: Structure, Hardness and Self-Lubricity

Abstract

It is known that the presence of oxygen phases in hard materials leads to an undesirable decrease in the mechanical properties. In materials based on AlMgB₁₄, the main oxygen impurity is spinel MgAl₂O₄; it significantly reduces the hardness of AlMgB₁₄ and its formation during sintering is inevitable. In this work, the ultra-hard spark plasma sintered (SPSed) AlMgB₁₄-TiB₂ composite material was fabricated from the AlMgB₁₄-TiB₂ precursor obtained by self-propagating high-temperature synthesis (SHS). Due to the high synthesis temperatures, the main oxygen phase in the obtained composite was Al₄B₂O₉ instead of spinel MgAl₂O₄. It was found that the obtained composite has excellent mechanical properties. The maximum hardness of the sample is 44.1 GPa. The presence of oxygen in the form of the Al₄B₂O₉ phase led to unexpected results: the friction coefficient of the obtained AlMgB₁₄-TiB₂ composite under dry conditions against the Al₂O₃ counter-specimen is approximately four times lower than the friction coefficient of pure ceramic AlMgB₁₄ (0.18 against 0.7, respectively). Based on the observed results, it was found that the Al₄B₂O₉ particles formed during the SHS are responsible for the low friction coefficient. The quantum chemical calculations showed that the elastic moduli of Al₄B₂O₉ are significantly smaller than the elastic moduli of AlMgB₁₄ and TiB₂. Thus, during sliding, Al₄B₂O₉ particles are squeezed out onto the composite surface, form the lubricating layer and reduce the friction coefficient.

1. Introduction

Currently, AlMgB₁₄-based materials (BAMs) are of great interest due to their unique physical and mechanical properties. BAMs have an extremely low friction coefficient (COF, ~0.08–0.02) [1,2,3], a high hardness, with value reaching 32 GPa [4], a coefficient of thermal expansion (CTE) close to steel (11.7 × 10⁻⁶ K⁻¹) [5], and a high thermal and chemical stability. Although AlMgB₁₄ single crystals were first obtained in 1970 by Matkovich and Economy [6], polycrystalline materials based on AlMgB₁₄ have been actively studied since 2000 in the Ames Laboratory (USA) [2,3,4]. First, the scientists of [4] found that the addition of TiB₂ increases the hardness of the AlMgB₁₄-TiB₂ composite to 40–46 GPa, and then the studies of the tribological characteristics showed that the friction coefficient of the AlMgB₁₄ coating in dry conditions can reach 0.05 [3], while the friction coefficient of the AlMgB₁₄-TiB₂ coating in a water-glycol lubricant can reach values of 0.02 [1]. Due to the unique combination of these properties, AlMgB₁₄ has attracted great attention as a wear-resistant material in the friction units of critical machine parts.

Tribological analysis showed that the extremely low friction coefficient in dry conditions and water-glycol lubricants is associated with the release of B₂O₃ and B(OH)₃ particles [1,2,3] during the friction. Studies of the tribological characteristics of AlMgB₁₄-TiB₂ composites in dry conditions [7] revealed that the COF of AlMgB₁₄-30 wt% TiB₂ at room temperature varies in the range from 0.45 to 0.65 against the SiC counterface and decreases significantly to 0.2 at 800 °C. The authors of [7] attribute this effect to the formation of TiO₂ and B₂O₃ oxygen compounds on the contact surfaces of materials. Studies of the tribological characteristics of AlMgB₁₄-Si composites [8] have shown that the friction coefficient under dry conditions varies from 0.19 to 0.28 when sliding against the 316 L counterface. A low COF is associated with the release of Si particles onto the worn surface. A similar effect was observed in [9]: studies have shown that one of the reasons for the low friction coefficient (COF, 0.12) of Cu/AlMgB₁₄ composites is that AlMgB₁₄ particles are squeezed out of a rigid Cu matrix onto a worn surface under the action of external stress.

Based on the given data, it can be concluded that the low friction coefficient of various composites AlMgB₁₄-X (X = TiB₂, Si, Cu) corresponds to the release of certain particles on the surface of the material during friction. These particles lubricate the surface, reducing the COF and wear, respectively. They are released either in the rigid metal matrix [8,9], or when using certain lubricants [1,2,3], or when the surface is heated [7]. At the same time, the question of the mechanism of the self-lubrication of bulk composites AlMgB₁₄-TiB₂ in dry conditions at room temperature remains open.

In this work, the AlMgB₁₄-TiB₂ precursor was produced using a cost-effective technology of self-propagating high-temperature synthesis (SHS) [10]. Thereafter, the bulk AlMgB₁₄-TiB₂ composite was fabricated using the spark plasma sintering (SPS) method. The mechanism of the self-lubricity, phase composition, structure and properties of the obtained AlMgB₁₄-TiB₂ composite were studied.

2. Materials and Methods

2.1. Materials Preparation

To obtain the AlMgB₁₄-TiB₂ precursor, a simple and low-cost technology of SHS was used. (Ti + 2B) exothermic mixture was chosen as the donor component. For this, 69 wt% of titanium was mixed with 31 wt% of boron. The mixture of an intermetallic alloy Al₁₂Mg₁₇ (the process of obtaining the Al₁₂Mg₁₇ powder is described in detail in [11] and amorphous boron powder) was used as an acceptor mixture. For this, Al₁₂Mg₁₇ powder was mixed with B powder in the ratio of 25 wt% Al₁₂Mg₁₇–75 wt% B. The characteristics of the raw precursors are given in

Table 1. Characteristics of the raw powders.

Raw Precursors	Particle Size, μm	Purity, %
Ti	140	≥99.2
Amorphous B	0.6	≥98.7
Al12Mg17	15	≥99.2

. The obtained powders were mixed in the ratio of 50 wt% (Al₁₂Mg₁₇: B)–50 wt% (Ti + 2B). The mixing was carried out in ethanol. After drying in a vacuum furnace at 200 °C for 4 h, a sample was prepared from the obtained mixture, which was then synthesized by the SHS method. The process of obtaining AlMgB₁₄-TiB₂ precursors by SHS is described in detail in [10].

In the second set, the produced AlMgB₁₄-TiB₂ precursor was consolidated and sintered by spark plasma sintering (DR. SINTER model SPS-625 Spark Plasma Sintering System, SPS SYNTEX INC. Ltd., Tokyo, Japan) in a 12.8 mm graphite die in an argon atmosphere. The sintering temperature was 1500 °C (the heating rate was 50 °C/min without holding time). The pressure was 70 MPa. The diameter and thickness of the obtained composite were ~12.5 mm and 3 mm, respectively. After sintering, the sample was taken out from the die, and the surface of the sample was polished with diamond pastes (the maximum surface roughness was 0.5 μm).

2.2. Measurements

The combustion temperature was measured using a tungsten-rhenium thermocouple (Promelectronica, Yekaterinburg, Russia). The morphology, elemental composition and grain size of the AlMgB₁₄-TiB₂ precursor were observed by scanning electron microscopy (JEOL JSM-6490, JEOL, Tokyo, Japan) at BEC (backscattered electron) and MIX (backscattered electron + secondary electron) modes equipped with an energy dispersive spectroscopy (EDX). To determine the amount of oxygen in the initial powders, a LECO ONH (St. Joseph, MI, USA) analyser was used. X-ray structural studies of the AlMgB₁₄-TiB₂ composite material were carried out on a Shimadzu 7000 (Kyoto, Japan) diffractometer. The phase composition was refined using the Rietveld method with the Diffrac.EVA program and Powder Diffraction File database. The theoretical XRD pattern was obtained by combining three reference XRD patterns of AlMgB₁₄, TiB₂ and Al₄B₂O₉ phases. The CASTEP program code [12,13] was used to calculate the free energies (E) of the reference and refined crystal lattices within the framework of the density functional theory (DFT) using the generalized-gradient approximation (GGA). Ultrasoft pseudopotentials were used, and the cutoff energy was 500 eV. The density of the AlMgB₁₄-TiB₂ composite was calculated by Archimedes’ principle in distilled water. The nanohardness was measured using a benchtop nanoindentation system (CSM Instruments— Peseux, Switzerland). The load was 300 mN, and the holding time was 10 s. The determination of the nanohardness was carried out by the methods of Oliver and Pharr [14]. The Berkovich indenter was used in the work. The obtained nanohardness results were converted into Vickers microhardness. The friction coefficient was measured on a pin-on-disk tribometer (TRIBOTester, TRIBOtechnic, San Francisco, CA, USA). Al₂O₃ ball was used as a counter-specimen. The normal load was 5 N, and the speed was 25 mm/s with the test duration of 1500 s. The tests were performed under normal conditions at room temperature in air without using a lubricant coating. The topographic and energy dispersive (EDX) analysis of the near-surface regions of the sample after the sliding friction was performed on a Tescan Mira 3 (TESCAN, Brno, Czech Republic) scanning microscope.

3. Results

3.1. Combustion Temperature and Morphology of the AlMgB₁₄-TiB₂ Precursor

The results of measuring the temperature (the heat pattern is given in the Supplementary Materials) showed that during the exothermic reaction (Ti + 2B), the temperature increases to 1980 °C, and titanium diboride grains are formed. The heat released during the reaction is spent on the reaction in the Al₁₂Mg₁₇:B acceptor mixture with the formation of AlMgB₁₄ [15]. According to the presented SEM images (Figure 1), the produced AlMgB₁₄-TiB₂ precursor consists of particles with an average size of 1 μm. EDX analysis (given in Supplementary Materials) showed that Ti and B elements were found in the light areas, which corresponds to the TiB₂ phase. The B, Al and Mg elements were found in the dark areas, which corresponds to the AlMgB₁₄ phase. The titanium diboride particles (light particles) also form agglomerates. The formation of TiB₂ agglomerates is due to the fact that the heat released during the reaction (Ti + 2B) also leads to the fusion of TiB₂ grains [16,17].

3.2. Phase Composition of the AlMgB₁₄-TiB₂ Composite

The results of the XRD analysis of the spark plasma sintered AlMgB₁₄-TiB₂ composite are shown in Figure 2. The results of the X-ray structural analysis are shown in

Table 2. Structural parameters of the detected phases.

Phase	State	a, Å	b, Å	c, Å	α = β	γ	V, Å3	E, eV
TiB2	Reference	3.009	3.009	3.262	90.00	120.00	25.578	−210.90
	Refined	3.033	3.033	3.240	90.00	120.00	25.816	−210.89
AlMgB14	Reference	5.848	8.115	10.313	90.00	90.00	489.419	−7149.38
	Refined	5.690	8.050	10.357	90.00	90.00	474.448	−7148.67
Al4B2O9	Reference	14.792	15.028	5.534	90.00	90.96	1230.15	−34471.7
	Refined	14.460	15.036	5.475	90.00	85.51	1152.61	−34469.3

(a, b, c, α, β, γ are the structural parameters, V is the volume of the crystal lattices and E is the free energy of crystal lattices of the detected phases). The analysis of the contributions to the integral intensity of the individual phases (Table 2, Figure 2) showed that the main phases are TiB₂, AlMgB₁₄, and Al₄B₂O₉.

The experimental XRD pattern of the obtained composite is well approximated by the calculated integral intensity (Figure 2a). The refined parameters of the main phases slightly differ from the reference. According to the results obtained using the light element analyzer LECO ONH, the main source of oxygen for the formation of the Al₄B₂O₉ oxide phase is boron powder (the oxygen content is 1.1 wt%).

3.3. Microstructure of the AlMgB₁₄-TiB₂ Composite

The microstructure of the fracture surface and the near-surface region of the spark plasma-sintered AlMgB₁₄-TiB₂ composite are shown in Figure 3. The structure of the obtained composite is not uniform. According to the EDX results, the Ti and B elements were found in the light areas (should be TiB₂ phase). In the dark areas, the B, Al and Mg elements were found in the ratio corresponding to the AlMgB₁₄ phase. AlMgB₁₄ grains have an average size of 3–5 μm, while the TiB₂ grains have an average size of 1–3 μm. The structure of the composite also contains large agglomerates of TiB₂, which are formed from agglomerates in the AlMgB₁₄-TiB₂ precursor [18].

3.4. Hardness and Density of the AlMgB₁₄-TiB₂ Composite

According to the nanohardness measurement results, in the areas corresponding mainly to AlMgB₁₄, the average Vickers microhardness is 32.5 GPa. In the regions corresponding to TiB₂, the average Vickers microhardness hardness is 33.1 GPa. In the AlMgB₁₄-TiB₂ interlayer, the average hardness of the sample is 37.4 GPa with a maximum detected hardness value of 44.1 GPa. The obtained results are similar to the results of other authors (

Table 3. Comparison of hardness values of AlMgB₁₄-based materials.

Composition	Hardness Value, GPa	Reference
AlMgB14 + 50 wt% TiB2	40–46	[4]
AlMgB14 + 30 wt% Si	27	[8]
AlMgB14-5Al	22.4	[21]
AlMgB14	26.7	[22]
AlMgB14 + 50 wt% TiB2	37.4	This work

) [4,7,19,20]. The density of the obtained sample is 3.271 g/cm³ at a theoretical density of 3.29 g/cm³ (pure 50 wt% of TiB₂ + 50 wt% of AlMgB₁₄ composite).

3.5. Tribological Behavior of the AlMgB₁₄-TiB₂ Composite

Figure 4 shows the curve of the change in the friction coefficient of the obtained AlMgB₁₄-TiB₂ composite. To compare this result, we also measured the friction coefficient of the AlMgB₁₄ and 30 wt% TiB₂ samples obtained in our previous works [10,23]. It was found that the friction curve of the obtained AlMgB₁₄-TiB₂ sample has a downward trend, while the curves of the AlMgB₁₄ and AlMgB₁₄-TiB₂ samples, on the contrary, have an upward trend. Moreover, the change in the values of the friction coefficient of the obtained AlMgB₁₄-TiB₂ composite has an atypical character in comparison with other samples: as can be seen from Figure 4, two extrema were found in the range of the COF values of 0.35 and 0.4, respectively. After reaching the value of 0.4, the friction coefficient sharply decreases to a value of 0.18, increases to 0.22 and decreases linearly to 0.18. At the same time, according to the XRD results, in contrast to the AlMgB₁₄ [23] and AlMgB₁₄–30 wt% TiB₂ [10] samples, the Al₄B₂O₉ phase was found in the obtained composite. A similar result was observed in [7], where due to the plastic deformation and elevated temperature TiO₂ particles formed a lubrication layer. Thus, it can be assumed that during friction, the friction coefficient increases until some particles (presumably, Al₄B₂O₉) appear on the surface of the sample and form a lubricating coating. After that, a decrease in the friction coefficient is observed.

4. Discussion

4.1. Formation of the Al₄B₂O₉ Phase

According to the XRD results (Table 2), the Al₄B₂O₉ phase was found in the obtained AlMgB₁₄-TiB₂ composite. It is known that the optimum temperature for obtaining AlMgB₁₄-based materials is 1400 °C [4]. In work [24], a mechanism for the decomposition of AlMgB₁₄ during spark plasma sintering was proposed. It was found that, upon the local overheating of the powder mixture (1470 °C) during spark plasma sintering, AlMgB₁₄ decomposes and the decomposition products react with MgAl₂O₄ spinel, forming the Al₄B₂O₉ borate. In this work, the overheating of the powder mixture was also observed both during the preparation of the AlMgB₁₄-TiB₂ precursor (the synthesis temperature was 1980 °C) and during the fabrication of the AlMgB₁₄-TiB₂ composite (the sintering temperature was 1500 °C). It should also be noted that the main impurity phase MgAl₂O₄ [4] was not detected in the AlMgB₁₄-TiB₂ composite. Thus, it is confirmed that AlMgB₁₄ decomposes and reacts with MgAl₂O₄, forming aluminum borate Al₄B₂O₉, during SHS and SPS due to local overheating.

4.2. Structure and Phase Composition of the AlMgB₁₄-TiB₂ Composite

The crystalline phases of TiB₂ and AlMgB₁₄ are precipitated in the composite material. Indeed, quantitative phase analysis (Table 2, Figure 2b) showed that the contribution of the intensity of the TiB₂ lattice (Figure 2b, mark 4) is dominant in the crystalline part of the integrated intensity (Figure 2a, mark 2). For these phases, complete structural information was found (Table 2). An increase in the lattice parameters was found, but the volume of the AlMgB₁₄ crystal lattice decreased relative to the reference, which is associated with a change in the shape of the lattice.

Complete information about the structure makes it possible to elucidate from the first principles the stability of the lattices of the main phases of the AlMgB₁₄-TiB₂ composite. The details of the calculations are given in Section 4.3. The estimates from the first principles showed that the lattice energies of the main phases (TiB₂, AlMgB₁₄ and Al₄B₂O₉) are negative; therefore, they are stable (Table 2). The Al₄B₂O₉ supercell is also stable to the lattice separation of the Al₂O₃ and B₂O₃ phases (Section 4.3). Indeed, its binding energy is |ΔEst| ≈ 2163 eV. The analysis of the morphology of the fracture surface (Figure 3a–c) showed that TiB₂ agglomerates (light areas) are characterized by a clear separation of interfaces between crystallites of various shapes and sizes with an arbitrary direction, while AlMgB₁₄ (dark areas), on the contrary, are characterized by a blurred surface relief (Figure 3c).

4.3. Mechanism of Self-Lubricity of the AlMgB₁₄-TiB₂ Composite

An analysis of the literature data [7,8,9] shows that the nature of the change in the friction coefficient under the dry conditions of the obtained AlMgB₁₄-TiB₂ composite is not typical for AlMgB₁₄ materials and AlMgB₁₄-TiB₂ composite materials (Figure 4). Thus, of particular interest are studies aimed at identifying the mechanism for reducing the friction coefficient in the obtained composite. Since the Al₄B₂O₉ lattice is a supercell with the chemical formula {2(Al₂O₃) × (B₂O₃)} (that is, it contains B₂O₃ particles in its compound, which are responsible for the lubrication of AlMgB₁₄-TiB₂ composite materials [2]), it is assumed that the dry friction process is significantly influenced by B₂O₃ particles, as well as the morphology of AlMgB₁₄.

To reveal the mechanism of self-lubrication of the obtained AlMgB₁₄-TiB₂ composite, the stability of the Al₄B₂O₉ phase was calculated. Since Al₄B₂O₉ is a supercell with the chemical formula {2(Al₂O₃) × (B₂O₃)} (Figure 5), its stability ΔEst can be estimated by Formula (1):

ΔE_st = ΔE_Al4B2O9 − (xΔE_Al2O3 + yΔE_B2O3)

(1)

where ΔEst < 0 and x and y are the relative contents of the Al₂O₃ and B₂O₃ phases in the Al₄B₂O₉ supercell. ΔE_Al4B2O9 is the supercell energy. The lattice energies of the Al₂O₃ and B₂O₃ phases were calculated from the first principles and were ΔE_Al2O3 = −8590.47 eV and ΔE_B2O3 = −4243.67 eV. The lattice energy of Al₄B₂O₉ is given in Table 2, and the coefficients are: x = 16/6, y = 8/3. The estimates showed that ΔEst ≈ −2163 eV, demonstrating that the Al₄B₂O₉ lattice is substantially stable.

In the next stage, quantum-chemical calculations of the elastic moduli of the main phases were performed using the method [25,26]. The results are given in

Table 4. Elastic moduli (GPa) of the main phases of the obtained material.

Phase	Young’s Modulus	Young’s Modulus (Experimental Value) *	Bulk Modulus	Shear Modulus
TiB2	491.8	480.1	245.9	210.8
AlMgB14	424.7		237.3	182.9
Al4B2O9	206.2		132.0	83.2

*—From nanohardness measurements.

. The obtained results are in a good agreement with the experimentally observed value. The elastic lattice moduli of Al₄B₂O₉ are significantly smaller than those of TiB₂ and AlMgB₁₄. It should also be noted that the elastic moduli of the AlMgB₁₄ are less than the corresponding values of TiB₂; AlMgB₁₄ can be a “viscous” base of the composite during the dry friction. Thus, it is assumed that the Al₄B₂O₉ particles are squeezed out of the rigid TiB₂ matrix and viscous AlMgB₁₄ matrix onto the composite surface and lubricate it, which leads to a decrease in the friction coefficient.

To find particles of the Al₄B₂O₉ composition on the surface and thus prove this hypothesis, the elemental composition of the EDX method was determined in the track region. Figure 6 shows a fragment of the topographic surface of the AlMgB₁₄-TiB₂ composite in the track region after measuring the friction coefficient.

The mass fraction of the elements in the corresponding regions is given in

Table 5. Spectra of elements in selected places on the surface of the AlMgB₁₄-TiB₂ composite.

Element	Mark 1	Mark 2	Mark 3	Mark 4	Mark 5	Mark 6
	wt%
B	84.4	93.1	66.7	49.0	55.8	43.9
Al	8.8	8.5	8.7	17.7	15.0	20.9
Mg	6.1	6.0	5.9	9.3	7.5	9.7
Ti	0.4	2.1	0.4	0.4	0.3	0.3
O	−	0.2	18.3	23.5	21.3	25.1
Estimated phase	AlMgB14	AlMgB14/TiB2	Al4B2O9	Al4B2O9	Al4B2O9	Al4B2O9

. The results of the estimates of the molar fraction indicate that in spectra 3, 4, 5 and 6 (Figure 6), the areas of destruction of the Al₄B₂O₉ particles as a result of the sliding friction with the release of the B₂O₃ and Al₂O₃ compounds of a variable composition are found.

Thus, Al₄B₂O₉ particles are distributed on the surface of a sample consisting of a rigid TiB₂ matrix with a “viscous” base of the AlMgB₁₄ (spectra 1 and 2) phase.

As previously reported in the Introduction, the effect of squeezing out particles onto the surface of AlMgB₁₄-based materials was observed in other studies. A comparison of the friction coefficients is shown in

Table 6. Comparison of friction coefficients in dry conditions of the AlMgB₁₄-based materials.

Composition	Friction Coefficient	Reason of Low COF	Reference
AlMgB14	0.38	−	[11]
AlMgB14-TiB2	0.45−0.65	−	[7]
AlMgB14-TiB2	0.12 (800 °C)	TiO2 particles	[7]
AlMgB14-Si	0.19−0.28	316L counter-specimen Si particles	[8]
AlMgB14-Cu	0.2	AlMgB14 particles	[9]
AlMgB14-TiB2	0.18	Al4B2O9 particles	This work

. In [27,28], it was found that in the Al₂O₃-TiB₂ composites, during friction at elevated temperatures in the air, an oxide layer is formed, which lubricates the sample. In turn, the formation of the oxide layer is due to the plastic deformation of the tribo-contact region and the fact that the critical shear modulus of the oxide layer is much less than the shear modulus of the substrate. In this work, similar results were obtained: a sharp decrease in the COF is associated with the release of B₂O₃ and Al₂O₃ particles of a variable composition from the rigid TiB₂ matrix with the “viscous” base of the AlMgB₁₄ phase onto the composite surface during friction. These particles form a lubricating tribolayer and reduce the friction coefficient. In other words, the effect of the self-lubrication of the surface of the composite material AlMgB₁₄-TiB₂ with Al₄B₂O₉ particles is observed.

Thus, the Al₄B₂O₉ additive can be used as a lubricant in AlMgB₁₄-TiB₂ composite materials under dry friction conditions. It should be noted that according to the presented results, Al₄B₂O₉ does not negatively affect the hardness of the AlMgB₁₄-TiB₂ composite system. In the future, of particular interest is the study of AlMgB₁₄-TiB₂ composite materials that use prereacted Al₄B₂O₉ as the initial precursor.

5. Conclusions

• The AlMgB₁₄-TiB₂ composite was spark plasma sintered from the SHS precursor.
• The main phases in the obtained composite are TiB₂, AlMgB₁₄ and Al₄B₂O₉.
• The maximum hardness of the composite is 44.1 GPa at a density of 3.271 g/cm³.
• The presence of oxygen in the form of the Al₄B₂O₉ phase led to unexpectable results: the friction coefficient of the obtained AlMgB₁₄-TiB₂ composite in dry conditions against the Al₂O₃ counter-specimen is 0.18. It is approximately four times lower than the friction coefficient of pure ceramic AlMgB₁₄ (COF, 0.7).
• It was found that Al₄B₂O₉ particles are responsible for the low friction coefficient: the elastic moduli of Al₄B₂O₉ are significantly smaller than the elastic moduli of AlMgB₁₄ and TiB₂. Thus, during the friction, due to the plastic deformation of the tribo-contact region, Al₄B₂O₉ particles are squeezed out onto the composite surface, form the lubricating layer and reduce the friction coefficient.