Low temperature controllable synthesis of transition metal and rare earth borides mediated by molten salt

Principle and advantages

MSS is essentially a solvent-based method, which utilises a relatively low melting (and water-soluble) salt or salt mixture to form a liquid medium in which solid reactants can partially dissolve, which enhances the mobility of reaction species, improves their mixing, and increases their contact, on an atomic/molecular scale, facilitating the individual intermediate reactions and the overall reaction, and reducing the overall synthesis temperature and/or time. Moreover, the molten salt medium also assists in the formation of phase pure products with homogeneous powder characteristics and controlled morphology (shape/size). Such a technique is commonly used to synthesise a variety of oxides and non-oxides (borides, carbides, nitrides, alloys, etc.).^47–59 Figure 3 illustrates schematically the main steps involved in a typical MSS process, ⁵¹ including, combining raw materials and salt(s) in appropriate ratios, firing the powder mix in a reaction vessel (e.g. a crucible) to a target temperature above the melting/eutectic point of the salt(s) and holding at the temperature for a requisite time period, water-leaching repeatedly any residual salt(s) in the reacted mass after cooling to room temperature, and collecting and drying the final product. The salt(s) can be readily recycled via vaporising the water in the salt filtrate.

OPEN IN VIEWER

Figure 3.

Schematic of a typical MSS process, highlighting the main individual steps involved (after Ref. ⁵¹ ).

Apart from the considerable reduction in the synthesis temperature and/or time, there are several other advantages with MSS. It is easy to operate and readily scalable, and generally does not need to use any specialty reaction apparatus or vessels. Furthermore, in principle any types/forms of raw material precursors, including relatively inexpensive oxide-based raw materials, can be used. In addition, high quality product powders can be prepared, and more interestingly, their morphology (e.g. spheroidal, platelet- or lath/needle-shaped) and size (from nanoscale to microscale to macroscale) could be controlled/tailored. Benefiting from all of these, the overall production process is more economical than a conventional synthesis one.^47–59

Mechanisms

As well documented in the literature^47–59 depending on solubility values of the two reactants in molten salt under synthesis conditions, MSS could be dominated by different mechanisms. A ‘template-growth’ mechanism is dominant when one of the reactants has a much higher solubility than the other, whereas a ‘dissolution and precipitation’ mechanism is dominant when both reactants have relatively high solubility.

In the case of ‘template-growth’ mechanism dominated MSS, the reactant with a much higher solubility dissolves preferentially in the molten salt. The dissolved reactant species arrive at the surface of the much less-soluble reactant via rapid diffusion through the molten salt medium, and subsequently react with the latter to form an in-situ product phase. Initially the reaction is fast, but gradually becomes slower with the buildup of a product barrier layer on the less-soluble reactant since the reactant species then need to diffuse through it to continue the reaction. The final product particles from such a reaction process often largely retain morphologies and sizes of the much less-soluble reactant particles, which essentially act as the templates. Differently from this, in the case of a ‘dissolution-precipitation’ mechanism dominated MSS process, both reactants dissolve simultaneously in the salt. Upon oversaturation of the molten salt with the dissolved reactant species, product particles will precipitate from it, and generally exhibit different morphologies and sizes from those of either of the original reactants. In such a reaction process, mixing and diffusion of reactant species and their reactions are facilitated by the liquid salt medium, and more importantly, there is no product barrier layer formed on either of the reactant particles, leading to rapid reaction and significantly reduced synthesis temperature and/or time.

It should be noted, however, that the two reaction mechanisms described above could function simultaneously, although one of them is always more dominant than the other.

Conventional and molten salt mediated elemental reaction synthesis

Among the conventional techniques for metal boride synthesis, the ‘ERS’ is the most straightforward, which uses directly elemental metal (M) and amorphous or crystalline boron (B) powders as the starting materials (Reaction (1)).

M + xB = M B x

(1)

The negative Gibbs energy values calculated by using the Factsage package (Figure 4), ⁶⁰ along with the relevant thermodynamic data and phase diagrams reported in the literature^{1,6,11,23,60–70} suggest that the reaction between a transition metal or rare earth metal, and elemental boron is thermodynamically favourable, at all the temperatures, which is the thermodynamic basis behind various synthesis routes, including the conventional high-temperature and high pressure arc-melting,^71,72 high temperature solid-state reaction,^73,74 as well as long time mechanical alloying.^74–76 Given the extremely exothermic nature of Reaction (1),^1,6,60–70 the so-called self-propagating high-temperature synthesis (SHS) was also used.^77,78 These synthesis approaches suffer from various disadvantages, for example requirement of high temperature/pressure and/or long processing time and/or high agglomeration of product boride. Despite these, ERS overall does exhibit several advantages. For example, it is easy to control the stoichiometry of the product boride by simply controlling the original metal/boron ratio, and the product phase is generally of high purity (strongly depending on the purity of the two raw materials). Furthermore, the target boride is the only product from Reaction (1), that is no by-product is formed, avoiding otherwise required acid-leaching.

OPEN IN VIEWER

Figure 4.

The standard Gibbs free energy corresponding to Reaction (1) (per mole of metal) as a function of temperature and boride type.

As well as these ‘intrinsic’ advantages of ERS, additional advantages are exhibited when an appropriate type of salt is used to form a liquid reaction medium (referred to as ‘molten salt mediated-elemental reaction synthesis (MSM-ERS)’) (Figure 2). First, the synthesis temperature and/or time can be reduced thanks to the accelerating effects of molten salt medium; second, the resultant boride product powders are generally of good quality; and third, by using different types/compositions of salts, solubilities of the reactants can be adjusted, enabling a good control in the synthesis mechanism and thus in the product's morphology and size (Section ‘Mechanisms’).

By using MSM-ERS, several transition metal borides, including ZrB₂, HfB₂, TiB₂, NbB₂ and VB₂ were synthesised.^79–91 In these syntheses, Na₂B₄O₇ was more commonly used to form the required molten salt medium, probably owing to its low melting point and reasonably high solubility of B in it. Table 3 summarises synthesis conditions and characteristics of raw materials and boride powders resultant from MSM-ERS,^79–91 revealing the relatively low synthesis temperatures (750–850 °C) in all the cases, and the high effectiveness of Na₂B₄O₇ in facilitating the overall synthesis process.

OPEN IN VIEWER

Table 3.

List of common transition metal borides resultant from MSM-ERS.

Boride	Metal powder (size)	Boron (size)	Boride shape/size	Salt	Synthesis condition	Reference
TiB2	10–15 μm	10–20 μm	Spherical: 70–75 nm	Na2B4O7	750 °C/10 h or 850 °C/5 h	79
	1–10 μm	10–20 μm	70–75 nm	Na2B4O7	785 °C/10 h or 815 °C/5 h	81
	10–20 μm	10–20 μm	325 nm	Na2B4O7	850 °C/5 h	81
	3 μm (TiH2)	0.95 μm	60 nm	NaCl–KCl	1000 °C	90
ZrB2	10–15 μm	10–20 μm	90–95 nm	Na2B4O7	750–800 °C/10 h	79
	∼10 μm	10–15 μm	60–80 nm	Na2B4O7	750 °C/10 h or 850 °C/5 h	82
HfB2	10–15 μm	10–20 μm	50–55 nm	Na2B4O7	750 °C/15 h or 800–850 °C/10 h	79,80
NbB2	10–15 μm	10–20 μm	65 nm	Na2B4O7	800 °C/32 h	83,84
			64 nm	KCl		83,84
			67 nm	KBr		83
VB2	10–15 μm	10–20 μm	90 nm/95 nm	Na2B4O7	800 °C/32 h	85,86
				KCl
WB2	1 μm	3 μm	Hexagonal: 0.5–2.5 μm	NaCl–KCl	1000 °C/3 h	89
ReB2	–	–	Nanosheet	NaCl–KCl	1000 °C/1 h	91
FeB	∼45 μm	d50 ≤ 0.90 μm	Nanorod:30–40 nm in diameter and 100–150 nm in length	NaCl–KCl	1000 °C/1 h	87
Fe2B	∼45 μm	d50 ≤ 0.90 μm	Platelet: 20–40 nm thick	NaCl–KCl	1000 °C/1 h	87
	∼5 μm	∼5 μm	Nanoflake: 20 nm thick			88

Halide salts such as NaCl, KCl, KBr and their eutectic assemblies were also used to form the required liquid media for MEM-ERS of some borides. As seen from Table 3, in these salt media, transition metal borides still could be synthesised at a relatively low temperature (≤1000 °C), and in most cases nanosized product particles were formed, indicating that metal halide salts were also effective in promoting the synthesis processes.

A ‘common’ mechanism for the MSM-ERS in Na₂B₄O₇ was suggested as follows^79–86: B atoms first change to B²⁺ ions via interacting with B³⁺ ions from the molten Na₂B₄O₇ according to Reaction (2):

B + 2 B 3 + ( in salt melt ) = 3 B 2 + ( in salt melt )

(2)

3 B 2 + ( in salt melt ) + M ⇀ M B 2 + B 3 + ( in salt melt )

(3)

Apart from the salt type, the relative salt content (indicated by the salt/reactant ratio) affects the synthesis process and the final product, indicating the importance of using appropriate amounts of salt(s). The use of too little or too much salt may lead to less positive effect on the reaction extent and/or the final product. If the salt used is not sufficient, it would not be able to cover well the whole reaction system. On the other hand, if too excessive salt is used, the concentrations of reaction species in it would decrease, and their diffusion distances would increase. In both cases, the reaction rate would generally decrease. Nevertheless, the relative salt content does not seem to have any evident effect on the reaction mechanism. ⁸⁹

Despite its many advantages in the boride synthesis, the MSM-ERS route suffers several disadvantages, one of which is its unfeasibility for commercial production due to using expensive elemental material precursors. This also might be one of the main reasons for the lack of work on MSM-ERS of rare earth metal borides (Table 3), given that the elemental rare earth metal particles are even more expensive.

Conventional and molten salt mediated borothermal reduction synthesis

To avoid using directly elemental metal (M) as a starting raw material (as in ERS), its metal oxide (MO_x) can be used alternatively, along with elemental B. Such a ‘BRS’ process can be expressed as:

M O x + B → M B v + B O z

(4)

Except in the case of LaB₆, the standard Gibbs free energy values corresponding to the formation of all the common borides shown in Figure 5 (via Reaction (4)) are negative, even at a very low temperature, which is similar to in the case of ERS (Figure 4), suggesting that their formation processes are also thermodynamically favourable at a very low temperature. However, given its use of a high melting metal oxide precursor and solid-state reaction nature, the conventional BRS can only be completed at a higher synthesis temperature for a longer time than in the case of ERS. For example, to synthesise phase pure ZrB₂ via the conventional BRS route, a temperature as high as 1700–1750 °C was required. ⁹⁴ This, along with the use of much excessive amount of expensive elemental B, makes the conventional BRS and the resultant boride product quite expensive, although BRS is considered a good method for industrial scale production of alkali metal, transition metal, and rare-earth borides.^98–100 In the case of LaB₆, although the standard Gibbs free energy values are positive, it can still be prepared via Reaction (4) by reducing partial pressures of boron oxide by-products via operation at a relatively high temperature in vacuum. This was demonstrated by Berrada et al. and Zheng et al. who synthesised LaB₆ via borothermal reduction of La₂O₃ in vacuum at as high as 1800 °C, and 1650 °C (for 6 h), respectively.^101,102 By using submicron-sized raw materials, the synthesis temperature in vacuum could be reduced to 1250 °C, but a much longer soaking time (8 h) was required. ¹⁰³

OPEN IN VIEWER

Figure 5.

The standard Gibbs free energy corresponding to Reaction (4) (per mole of metal oxide) as a function of temperature and boride type.

To overcome the above-mentioned drawbacks of conventional BRS, a molten salt medium can be similarly introduced to mediate it (referred to as ‘molten salt mediated BRS (MSM-BRS)’) (Figure 2).

Surprisingly, to the knowledge of the authors, no work has been reported on using Na₂B₄O₇ as a reaction medium for MSM-BRS, differently from in the case of MSM-ERS described above and shown in Table 3. Instead, chloride salts, including LiCl, NaCl, KCl, MgCl₂ and their combined salts were often used. Table 4^91,104–113 summarises the studies to date on MSM-BRS of transition metal borides (mainly diborides), showing that the synthesis temperature was in the range of 800–1100 °C, which overall was slightly higher than that required by MSM-ERS discussed above, but still significantly lower than that required by the conventional solid-state BRS process without being mediated by a molten salt.

OPEN IN VIEWER

Table 4.

List of common transition metal borides resultant from MSM-BRS.

Boride	Metal oxide (size)	Boron (size)	Boride shape and size	Salt	Synthesis condition	Reference
TiB2	–	<5 μm	Nanosheet	MgCl2	1100 °C/4 h	104
	100–300 nm	<3 μm	40 nm	NaCl–KCl	900 °C/1 h	105
	–	–	Nanosheet			91
ZrB2	–	–	Particles: 100–500 nm	NaCl	950 °C/3 h	106
	50–100 nm	<3 μm	Nanoplates: lateral sizes of 2–6 μm and thicknesses of 50–200 nm	NaCl–KCl	1000 °C/1 h	107
	–	–	Nanosheet			91
HfB2	1–3 μm	<3 μm	Irregular polyhedral: 155 nm		1100 °C/1 h	108
	–	–	Nanosheet		1000 °C/1 h	91
NbB2	–	d50 ≤ 0.90 μm	32–81 nm		800–1000 °C/1 h	111
	–	–	Nano-needle		900 °C/4 h	91
VB2	0.5–2 μm	Few μm	100 nm	NaCl		112
	–	–	Nanosheet	NaCl–KCl		91
W2B5		d50 ≤ 0.90μm	Agglomerates comprising nanosized particles		1100 °C/1 h	113
TaB2	–	d50 ≤ 0.90μm	Nanoflower		800–1000 °C/1 h	109
	–	–	Nanoneedle		900 °C/1 h	91
CrB2	–	d50 ≤ 0.90μm	104 nm			110
	–	–	Nanosheet			91
MnB2	–	–	Nanoneedle
MoB2	–	–	Nanoparticle
RuB2	–	–
WB2	–	–			1000 °C/1 h

In some cases of conventional BRS (without using a salt medium), B reduces the metal oxide precursor to various extents, leading to the formation of one or more borides with undesirable stoichiometries,^114,115 especially under the conditions of insufficiently high firing temperature and/or insufficiently long soaking time. However, this can be avoided in most cases of MSM-BRS, since the synthesis reaction could be completed at much milder conditions thanks to the facilitation from the molten salt.

The salt type in the MSM-BRS showed some effects on the reaction extent and the synthesis temperature/time. For example, in the MSM-BRS of TiB₂, the synthesis temperature in the case of using MgCl₂ was 1100 °C, ¹⁰⁴ but it was reduced to 900 or 950 °C, upon replacing MgCl₂ with the NaCl–KCl eutectic salt.^91,105 In the MSM-BRS of ZrB₂, the synthesis temperature in NaCl was 950 °C, ¹⁰⁶ but it was increased slightly to 1000 °C when the NaCl–KCl eutectic salt was used instead.^91,107 The salt type was also found to have some effect on morphologies/sizes of the final boride particles. For example, ZrB₂ nanoparticles were formed in NaCl, ¹⁰⁶ whereas ZrB₂ nanoplatelets/nanosheets were formed in NaCl–KCl.^91,107 These results also indirectly reflected the effect of salt type on the synthesis mechanism. As seen from Table 4, in the most cases, the boride products showed different morphologies and/or sizes from the original reactant particles, seemingly suggesting the dominance of ‘dissolution-precipitation’ mechanism. Unfortunately, due to the limited relevant data, especially the solubility data of the reactants in the salts under test conditions, it is difficult to theoretically predict and precisely identify the reaction mechanism, as well as the morphologies of the final product(s). This is further compounded by the co-mediating effect from the by-product B₂O₃ which, once formed, could interact with the original chloride salt(s), additionally contributing to the formation of the liquid medium and changing the properties of the initial salt medium.

As in the MSM-ERS discussed above (Section ‘Conventional and molten salt mediated elemental reaction synthesis’), the relative salt content (salt/reactant ratio) also showed some effect on the reaction extent. It appeared that there existed an optimal range of salt/reactant ratio, beyond which the accelerating effect of the salt on the synthesis reaction might be weakened. Figure 6, as an example, shows the effect of salt/reactant weight ratio on the reaction extent and phase formation in the MSM-BRS of W₂B₅ in NaCl–KCl. ¹¹³ When no salt was used (salt/reactant ratio = 0), the reaction product contained lots of intermediate WB and metal W (Figure 6(a)). However, when 5 times salt (with respect to the total amount of reactants, i.e. salt/reactant = 5:1) was used, W₂B₅ was formed as the primary phase, and only minor WO₃ and WB were detected (Figure 6(b)), indicating the beneficial effect of the salt on the synthesis reaction. When the salt/reactant ratio was increased to 10:1, only the target product phase of W₂B₅ was detected, and neither intermediate phase nor unreacted reactant was detected (Figure 6(c)). However, when the salt/reactant ratio was further increased to 15:1, apart from the primary product phase, some intermediate WB appeared again (Figure 6(d)). The above results suggested that the use of inappropriate amounts of molten salt might result in less positive effect on the boride formation. The reasons behind this are believed to be the same as those explained in the case of MSM-ERS above. Therefore, to obtain phase pure boride product, it is of vital importance to control or optimise the relative amount of the salt.

OPEN IN VIEWER

Figure 6.

X-ray diffraction (XRD) patterns of W₂B₅ samples resultant from MSM-BRS in NaCl/KCl at 1100 °C, with the salt/reactant weight ratio of (a) 0, (b) 5:1, (c) 10:1 and (d) 15:1, respectively. ¹¹³

The salt/reactant ratio also affected morphology/size of the final boride product. A typical example is with the MSM-BRS of VB₂ in NaCl. ¹¹² As shown in Figure 7, uniform and smooth VB₂ nanorods were formed via the conventional BRS (i.e. without using a salt) at 1000 °C. When a small amount of molten salt (the salt/reactant mass ratio = 0.25:1) was used, the rods became more irregular and coarser. When the salt/reactant ratio was further increased to 0.75:1, instead of nanorods, approximately spherical particles with a diameter of about 100 nm were formed. Therefore, by adjusting the relative amount of molten salt, boride products with different morphologies/sizes could be formed, offering a good strategy for the morphology/size control.

OPEN IN VIEWER

Figure 7.

High resolution (HR) scanning electron microscope (SEM) images of VB₂ powders resultant from MSM-BRS at 1000°C with different NaCl/reactant mass ratios: (a–c) 0:1 (i.e. no salt), (d–f) 0.25:1 and (g–i) 0.75:1. ¹¹²

Interestingly, even for a given salt and a given B source, morphologies and sizes of the boride product powders still could vary, depending on the metal type. As presented and compared in Figure 8, after MSM-BRS in NaCl–KCl at 900 or 1000 °C for 1 h, TiB₂, ZrB₂, HfB₂, VB₂, CrB₂ and ReB₂ nanosheets, NbB₂, TaB₂ and MnB₂ needles, and MoB₂, WB₂ and RuB₂ nanoparticles were formed, showing the ‘intrinsic’ effects of metal ions on the boride formation. ⁹¹

OPEN IN VIEWER

Figure 8.

Transmission electron microscope (TEM) and HR-TEM (inset) images of 12 metal diborides synthesised via MSM-BRS (except ReB₂ which was synthesised via MSM-ERS), revealing the formation of (a) TiB₂, (b) ZrB₂, (c) HfB₂, (d) VB₂, (g) CrB₂ and (k) ReB₂ nanosheets, (e) NbB₂, (f) TaB₂ and (j) MnB₂ nanowires, and (h) MoB₂, (i) WB₂ and (l) RuB₂ nanoparticles. ⁹¹

Instead of using a metal oxide precursor, a metal chloride also could be used as a metal source for MSM-BRS of metal boride. As reported by Liu and Gong, ¹¹⁶ β-MoB₂ nanoparticles were successfully synthesised in the LiCl-KCl eutectic salt, using MoCl₃ as the Mo source, at a relatively low temperature of 850 °C. However, given the even higher cost of transition metal chlorides, and/or their hygroscopic nature, such a synthesis route should be even less feasible for large scale productions, than in the case of using an oxide precursor counterpart discussed above.

Conventional and molten salt mediated metallothermic reduction synthesis

To avoid using both expensive elemental metal and boron as starting precursors (as used in ERS), their less expensive oxide counterparts (MO_x) can be used instead. However, a more reactive metal (M’) needs to be used as a reductant. In principle, any reactive metal (e.g. Na, K, Mg or Ca) can be used as a reductant, however, given the issues such as cost, stability in the atmosphere, reducing ability, as well as by-product elimination, Mg is the most commonly used. Such an MRS process can be expressed as:

M O x + B 2 O 3 + M ′ → M B v + M ′ O z

(5)

OPEN IN VIEWER

Figure 9.

The standard Gibbs free energy values corresponding to various magnesiothermal reduction reactions (per mole of metal oxide) as a function of temperature.

Differently from that indicated by the overall Reaction (5), the actual MSM-MRS process is rather complicated, involving several individual reaction steps. The metal reductant (M’) reduces the transition oxide precursor and B₂O₃ (or B₂O₃ containing precursors such as H₃BO₃ and Na₂B₄O₇) to respectively elemental transition metal and boron (Reactions (6) and (7)), both of which further react with each other, forming the desirable boride (Reaction (8)). In addition, once B is formed, it could also act as a reducing agent, reacting directly with the unreacted transition metal oxide to form additionally the target boride (Reaction (4)). So, MSM-MRS is essentially achieved via combined MSM-ERS (Reaction (8)) and MSM-BRS (Reaction (4)) discussed earlier.

M O x + M ′ → M + M ′ O z

(6)

B 2 O 3 + M ′ → B + M ′ O z

(7)

M + B → M B v

(8)

Another issue concerning the synthesis process is evaporation/vaporisation losses of B₂O₃ and Mg, since the former has a high vapour pressure at a relatively high firing temperature, ¹²⁶ and the latter has a relatively low boiling point of 1090 °C. ¹²⁷ Such evaporation/vaporisation losses lead to incompletion of the synthesis reaction, and the formation of boride(s) with undesirable stoichiometries and/or intermediate/impurity phases in the final reacted samples. These are generally avoided by using excess Mg and/or B₂O₃. For example, in the MSM-MRS of ZrB₂ in MgCl₂ at 1200 °C, 20 wt-% extra Mg was needed to complete the synthesis reaction, along with 30 wt-% extra Na₂B₄O₇ (B₂O₃). ⁵⁶ In the syntheses of LaB₆ and TiB₂ in MgCl₂ at a lower temperature (800–1000 °C), 20 mol-% extra Mg was still used, but only the stoichiometric amount of B₂O₃ was required.^55,58 If the temperature was sufficiently low, neither extra Mg nor extra B₂O₃ would be required. A typical example was with the MSM-MRS of NbB₂ in MgCl₂–NaCl at 650 °C, despite the long reaction time (10 h), the stoichiometric amounts of Mg and B₂O₃ (H₃BO₃) were used. ¹²⁸ Apart from firing temperature and time, several other factors affect the extra amounts of Mg and B₂O₃ required for a specific synthesis process, including soaking time at a high firing temperature, raw material particles, and salt type and amount. Nevertheless, firing temperature and time almost always played the major roles.

Table 5 lists the transition metal and rare-earth borides synthesised via the MSM-MRS route.^{54–58,124,125,128} The synthesis temperature in the case of MSM-MRS was overall higher than in the cases of MSM-ERS and MSM-BRS. The salts used were mainly chlorides, such as LiCl, KCl, NaCl, anhydrous MgCl₂ and their combinations, although occasionally other halides such as NaF were also used. Among the chloride salts, MgCl₂ has been found to show the best accelerating-effect on the MRS using Mg as a reductant, owing to mainly the relatively high solubility of Mg in it. ¹²⁹ The two synthesis mechanisms were identified in different specific cases, but the ‘dissolution-precipitation’ mechanism overall appeared to be more dominant than the ‘template-growth’ mechanism. In the MSM-MRS of TiB₂ in MgCl₂, ⁵⁵ three different types of TiO₂ precursors were used along with B₂O₃ as the main raw materials, that is TiO₂ I (powder, <25 nm), TiO₂ II (nanowires, ∼10 μm long and 10 nm in diameter), and TiO₂ III (powder, ∼300 nm). Despite such big differences in morphologies/sizes of TiO₂ raw materials, TiB₂ product particles in the three cases (formed at 1000 °C for 4 h) appeared to have similar morphology/size (100–200 nm particles), demonstrating that the morphology and size of original TiO₂ precursor had little effect on the reaction extent and morphology/size of TiB₂ product particles (Figure 10), suggesting the dominance of the ‘dissolution-precipitation’ mechanism in the overall synthesis process. Such a mechanism was also identified in the MSM-MRS of HfB₂ and LaB₆.^57,58

OPEN IN VIEWER

Figure 10.

MSM-MRS of TiB₂ by using different TiO₂ sources in MgCl₂ at 1000 °C for 4 h, indicating the dominance of the ‘dissolution-precipitation’ mechanism in the reaction process. ⁵⁵

OPEN IN VIEWER

Table 5.

List of transition metal borides resultant from MSM-MRS.

Boride	Metal oxide (shape/size)	Boron source	Boride shape and size	Salt	Synthesis condition	Reference
TiB2	TiO2 Nanoparticles: <25 nm or ∼300 nm Nanowires: ∼10 μm long and 10 nm in diameter	B2O3	Nanoparticles:100–200 nm	MgCl2, NaCl, KCl	1000 °C/4 h	55
	TiO2: 5 nm		Nanoparticle: ∼8 nm	LiCl/KCl	700 °C/0 h	54
ZrB2	ZrO2: <500 nm	Na2B4O7	Particles:300–400 nm	MgCl2, NaCl, KCl	1200 °C/3 h	56
	ZrO2: ∼85 μm	B4C/20 μm	Particles: 30–300 nm	NaCl–KCl	1200 °C/3 h	125
HfB2	HfO2: 1–3 μm	B2O3	Submicron	MgCl2, NaCl, KCl	1100 °C/1 h	57
NbB2	Nb2O5	Boric acid	Nanopartciles: 30/49 nm	MgCl2–NaCl	650 °C/10 h	128
LaB6	La2O3: ∼1 μm	B2O3	∼200 nm/∼100 nm	MgCl2, NaCl, KCl	1000 °C/4 h or 900 °C/5 h	58
WB/W2B	WO3	B2O3, Na2B4O7, KBF4	Mixture with W	NaF–NaCl	800–900 °C/1 h	124

On the other hand, in some other cases, for example in the ZrB₂ synthesis in MgCl₂, the other mechanism, the ‘template-growth’ mechanism, was found to be dominant. ⁵⁶ As shown in Figure 11, ZrB₂ product particles largely retained the shape and size of original ZrO₂ particles, suggesting that the latter had acted as the template. This is believed to be related to the much lower solubility of ZrO₂ and Zr than B in the salt at the synthesis temperature (1200 °C). This was different from the ‘dissolution-precipitation’ mechanism that was implied in the case of MSM-ERS (Table 3). The reason behind was not clear, but it should be related to the differences in solubility of the reactants in the two different salt media at two different temperatures. Another example is with the MSM-MRS of TiB₂ in LiCl–KCl at 700 °C. ⁵⁴ A ‘template growth’ mechanism was implied, due to probably the low solubility of TiO₂/Ti in the salt at such a low temperature, which was different from that in the MSM-MRS of TiB₂ in MgCl₂ at a much higher temperature (1000 °C), ⁵⁵ in which case, a ‘dissolution-precipitation’ mechanism was proposed, as discussed above. In a ‘template-growth’ mechanism dominated synthesis process, since the transition metal oxide acts as the template, reducing its particle size would be beneficial to the acceleration of the reaction, and the reduction of synthesis temperature and time. This could explain the above-mentioned low synthesis temperature (700 °C). ⁵⁴ Due to such a low synthesis temperature and short reaction time (no holding at the target temperature), both Mg and B₂O₃ did not experience significant evaporation/vaporisation losses, so the stoichiometric amounts of them were used. ⁵⁴

OPEN IN VIEWER

Figure 11.

SEM images of (a) the starting ZrO₂ powder, and (b) ZrB₂ resulted from firing batch powder with 20 wt-% excess Mg and 30 wt-% excess Na₂B₄O₇ at 1200°C for 3 h in MgCl₂.

As for the boron source, B₂O₃, Na₂B₄O₇ and several other boron-containing phases such as KBF₄ and B₄C, were used. Nevertheless, the first two were used much more commonly, as seen from Table 5. Given their low melting points, before their reduction, they could interact with the original salt(s) used, forming together the salt medium, increasing the solubility of B, and facilitating the synthesis reaction. Consequently, the synthesis mechanism was determined by the solubility of transition metal and its oxide precursor in the newly formed molten salt, as seen in the MSM-MRS of TiB₂ at a relatively high temperature (‘dissolution-precipitation’ mechanism due to high solubility of Ti and TiO₂) ⁵⁵ and ZrB₂ (‘template-growth’ mechanism due to low solubility of Zr and ZrO₂), ⁵⁶ as discussed above. B₄C and KBF₄ as boron sources did not show any advantages over B₂O₃ and Na₂B₄O₇.^124,125 For example, in the synthesis of ZrB₂ using 20 wt-% extra B₄C as a boron source, even very fine raw materials were used along with 50 wt-% extra Mg, the synthesis temperature remained as high as 1200 °C (though still much lower than required by the conventional route without using a molten salt). ¹²⁵ On the other hand, when KBF₄ was used as a boron source, some intermediate F-containing phases were formed, some of which are not eco-friendly and/or not water-soluble (so difficult to eliminate). ¹²⁴

Si was also used as a reductant for the MSM-MRS of borides. Li et al.^130,131 synthesised sheet-like ZrB₂ (0.7 μm in average size) via silicothermal reduction of ZrO₂ and Na₂B₄O₇ in molten sodium orthosilicate (Na₄SiO₄) at as low as 1000 °C (for 3 h). Both Na₂B₄O₇ and Na₄SiO₄ acted together as the molten salt medium and partially took part in the reactions as reactants. Along with the ZrB₂ product, two by-products Na₂Si₂O₅ and SiO₂ were formed from the silicothermal reduction reaction (Reaction (9)). The latter is not water soluble but could be converted into water soluble Na₂Si₂O₅ via reacting with the Na₄SiO₄ (Reaction (10)) or Na₂O (Reaction (11)) added to the original powder batch.

2 Zr O 2 + 5 Si + N a 2 B 4 O 7 = 2 Zr B 2 + N a 2 S i 2 O 5 + 3 Si O 2

(9)

3 SiO 2 + N a 4 Si O 4 = 2 N a 2 S i 2 O 5

(10)

2 Si O 2 + N a 2 O = N a 2 S i 2 O 5

(11)

Si + Zr O 2 = Zr + Si O 2

(12)

10 B + 3 Zr O 2 = 3 Zr B 2 + 2 B 2 O 3

(13)

In addition to Mg and Si, Al could be another good reductant for MSM-MRS. By using it as a reductant, Velashjerdia et al. synthesised ZrB₂ platelets in a KCl–NaCl or NaF–KF eutectic salt at a much lowered temperature (800 °C). The process was found to be controlled by the ‘dissolution-precipitation’ mechanism. ¹³³ Instead of oxide precursors, complex F-containing salts, KBF₄ and K₂ZrF₆, were used as the B and Zr sources, respectively. A few intermediate phases such as KAlF₄, K₃AlF₆ and Zr₃Zr were formed, which were difficult to remove from the final product powder, since they are not water soluble and acid leachable. ¹³³ Recently, Zhang's group have synthesised boride-alumina and boride–magnesioaluminate spinel composite powders,^134,135 via MSM-MRS using Al, further illustrating its feasibility as a reductant in the boride syntheses (see Section ‘Molten salt synthesis of metal boride-based composite powder’ below).

As described above, in a typical MRS process, both metal- and boron-containing compounds (in particular their oxides) are generally used as starting precursors. But in some MSM-MRS work, only one compound precursor was used together with another elemental precursor, as done by Wang et al. in the magnesiothermal reduction synthesis of TiB₂ particles in NaCl–KCl via Reactions (14) and (15), respectively.^136,137 They found that the onset formation temperature (800 °C) and the optimal synthesis temperature (1000 °C) were the same in both cases. The only difference was in the particle size of the final product. The particle size of TiB₂ (30–100 nm) generated from Reaction (14) overall was slightly smaller than that (40–200 nm) from Reaction (15).

Ti O 2 + 2 Mg + 2 B = Ti B 2 + 2 MgO

(14)

B 2 O 3 + Mg + Ti = Ti B 2 + 3 MgO

(15)

Ti O 2 + 5 / 4 Mg B 2 = Ti B 2 + 5 / 4 MgO + 1 / 4 B 2 O 3

(16)

T a 2 O 5 + 11 / 4 Mg B 2 = 2 Ta B 2 + 11 / 4 MgO + 3 / 4 B 2 O 3

(17)

In addition, Ma et al. prepared CrB nanorods of 10–30 nm in diameter and up to 1.5 μm in length, using CrCl₃ and elemental B precursors along with a Na reductant, in anhydrous aluminium trichloride at 650 °C (Reaction (18)). ¹⁴⁰ Although the synthesis temperature was very low, more expensive raw materials were used.

CrC l 3 + B + 3 Na = CrB + 3 NaCl

(18)

Conventional and molten salt mediated carbo- and boro/carbo-thermal reduction synthesis

Carbon (C) can be used as a reductant to replace the metal reductant used in MRS, to synthesise a metal boride. Such a ‘carbothermal reduction synthesis (CRS)’ process can be expressed as:

M O x + B 2 O 3 + C → M B v + CO

(19)

B₄C also can be used to replace B₂O₃ or both B₂O₃ and C in Reaction (19) to synthesise metal borides (referred to as ‘Boro/carbothermic reduction synthesis (BCRS)’), as indicated by the following Reactions (20) and (21).

M O x + B 4 C + C → M B y + CO ( g )

(20)

M O x + B 4 C → M B y + CO ( g )

(21)

A good strategy for reducing the temperatures required for the conventional CRS and BCRS is to use fine raw material precursors prepared via a ‘wet chemical synthesis’ process, for example a sol–gel process. For example, Ji et al. and Li et al. synthesised nanosized ZrB₂ at 1450–1550 °C for 2 h via CRS, using the precursors prepared via sol–gel routes,^160–162 and Zhang et al. synthesised TiB₂ via CRS at 1450 °C. ¹⁶³ Nevertheless, such synthesis temperatures remain too high.

To further reduce the synthesis temperature, a molten salt medium can be similarly introduced to mediate the CRS and BCRS (respectively referred to as ‘molten salt mediated CRS (MSM-CRS) and BCRS (MSM-BCRS)’) (Figure 2). Song et al. fabricated hexagonal TiB₂ microplatelets of 3-8μm in diameter and 200-500 nm in thickness, via MSM-CRS (Figure 12). ¹⁶⁴ The starting precursor mixture with Ti/B/C molar ratios of 1:4:6 was prepared via sol–gel processing using C₁₆H₃₆O₄Ti, Na₂B₄O₇·10H₂O and C₁₂H₂₂O₁₁ as the main starting precursors. Some in-situ formed NaCl acted as the molten salt medium for the synthesis reaction. The carbothermal reduction reaction started at 1100 °C, became significant at 1200 °C, and completed at 1300 °C. Although very fine raw material precursors were used, and the synthesis reaction was facilitated by the molten NaCl, the reaction temperature remained much higher than in the cases of using other MSS routes discussed above (Sections ‘Conventional and molten salt mediated elemental reaction synthesis’ to ‘Conventional and molten salt mediated metallothermic reduction synthesis’), although it was already much lower than in the case of using the conventional CRS method without molten salt mediation as mentioned above.

OPEN IN VIEWER

Figure 12.

SEM images of the TiB₂ products obtained via molten NaCl mediated CRS at 1300 °C for 2 h.

Similarly, according to Song et al., ¹⁶⁵ ZrB₂ rods with a diameter of 1–2 μm and an aspect ratio of 3–10 were synthesised using a precursor mixture prepared via sol-gel processing, along with in-situ formed NaCl. Despite the very fine sizes of ZrO₂ and B₂O₃ in-situ formed from pyrolysis of the precursor mixture, the synthesis temperature remained as high as 1400 °C. As expected, significant vaporisation loss of B₂O₃ occurred, so, as much as 5 times of stoichiometric amounts of B₂O₃ were used. At such a high temperature, the NaCl is also believed to have experienced considerable vaporisation loss. So, it can be considered that the NaCl only had functioned for a short period as the reaction medium before it was completely vaporised. This also might be the case in the MSM-CRS of TiB₂ platelets stated above. In both cases, the significant vaporisation loss of NaCl also might be partially responsible for the still relatively high synthesis temperature.

Conventional and molten salt mediated alkaline borohydride reduction synthesis

Transition metal and rare earth borides were also produced via reaction of a corresponding metal halide (in particular, a chloride) or metal oxide, with an alkaline borohydride (e.g. NaBH₄).^166–176 The former acts as a metal source, and the latter acts as both a boron source and a reductant. The latter is used to replace a reactive metal reductant, to avoid its reactions with oxygen- and moisture-containing species, which otherwise could lead to some side reactions.^166–176 Typically, the synthesis process using a metal floride/chloride and NaBH₄ can be expressed as follows:

M X x + NaB H 4 → M B y + NaX + HX + H 2 ( X = F , Cl )

(20)

By introducing a molten salt to mediate ABRS (referred to as ‘molten salt mediated-ABRS (MSM-ABRS)’ (Figure 2)), a range of metal borides could be synthesised at the atmospheric pressure for a much shorter time, although the use of a salt in this case does not seem to lead to further reduction in the synthesis temperature.^175,176 Figure 13 illustrates a typical MSM-ABRS process, along with types and morphologies of metal borides synthesised.^175,176

OPEN IN VIEWER

Figure 13.

MSM-ABRS of various metal borides with different compositions/stoichiometries and morphologies.

The salts attempted for MSM-ABRS included mainly, LiCl–KCl, LiCl–NaCl, NaCl–KCl and KBr, but the first salt assembly was used most, due to its much lower eutectic point (353 °C only). According to Reaction (20), some sodium halide, for example NaCl (when a metal chloride is used), is formed as a by-product, which could contribute additionally to the formation of the reaction medium (if the firing temperature is above its melting point or the eutectic point of its mixture with the original salt used), further reducing the eutectic point of the salt system and increasing its total volume.^177,178

The metal halide (chlorides mainly) starting precursors used in the reaction process are generally more sensitive to moisture than their oxide counterparts. Also, many of them have a very low boiling point, for example NbCl₅ (248 °C), TiCl₄ (136.4 °C) and ZrCl₄ (331 °C),^179–181 so when they are used, the synthesis reactions often have to be carried out in a closed vessel such as a quartz ampoule or an autoclave. Despite these drawbacks, the MSM-ABRS process does have several merits. It is straightforward and universally applicable to the synthesis of a wide range of borides and requires a relatively low temperature. For example, in the MSM-ABRS of LaB₆, CeB₆ and CoB, the synthesis temperature was as low as 700–750 °C.^182–184 Furthermore, very fine product particles from several nanometers to 100 nm, can be prepared, and the size can be tailored/controlled by varying the metal/boron ratio.^175,176 In addition, by adjusting the metal/boron ratio and/or temperature, borides with different stoichiometries and morphologies could be formed. For example, in the MSM-ABRS of Sm boride, flower-like clusters composed of SmB₄ nanorods with an average diameter of 60 nm were formed, when the ratio of SmCl₄/NaHB₄ was 1:3, whereas 50–100 nm SmB₆ nanoparticles were formed, when the ratio was decreased to 1:10 (Figure 14). ¹⁸⁵ Moreover, in the MSM-ABRS of CoB, at a given firing condition (900 °C for 4 h), the increase in the CoCl₂/NaHB₄ ratio led to evident change in the morphology of product CoB. On the other hand, at a given CoCl₂/NaHB₄ ratio, with increasing the temperature the morphology of the product also varied (Figure 15). ¹⁸⁶

OPEN IN VIEWER

Figure 14.

SEM images of (a–c) SmB₄ and (d–f) SmB₆ resultant from MSM-ABRS in LiCl–KCl eutectic salt at 750 °C for 4 h.

OPEN IN VIEWER

Figure 15.

Schematic illustration of morphology evolution of CoB resultant from MSM-ABRS with the variation of the cobalt chloride/sodium borohydride ratio and firing temperature.

Instead of metal chloride, a metal floride was also occasionally used for MSM-ABRS of a metal boride. Wang et al. synthesised approximately 50 nm CeB₄ particles in a LiCl–KCl eutectic salt at 750 °C for 3 h, using CeF₃ as the Ce source, ¹⁸⁷ and Liu et al. synthesised 15–30 nm CeB₆ particles at 900 °C for 4 h, using the same raw materials and salt. ¹⁸⁸ In the former, the CeF₃/NaHB₄ molar ratio (1:3.2) used was much lower than that (1:6 to 1:10) in the latter, demonstrating again the effect of metal/B ratio on the boride's stoichiometry, as mentioned above. Zhang et al. further synthesised ∼100 nm GdB₆ particles using GdF₃ and NaBH₄ (in the molar ratio of 1:12) as raw materials, in the same salt under the identical firing conditions. ¹⁸⁹ In addition, Liu and Gong prepared SmB₄ nanoflowers and SmB₆ cubes in the same salt, by using SmF₃ as the Sm source. The former was formed after 4 h firing at 850 °C by using a Sm/B molar ratio of 1:3, and the latter formed under the identical firing conditions by using a much lower Sm/B ratio of 1:10, indicating again that by adjusting the metal/B ratio, the boride's stoichiometry and morphology could be tailored.¹⁸⁵ Nevertheless, it has been found that the type of halide ion showed little effect on the synthesis process and morphology of the boride products.

As for the synthesis mechanism in MSM-ABRS, the work on it was rather limited. Given the relatively high solubility of SmF₃/SmCl₃ in molten LiCl–KCl, a ‘dissolution-precipitation’ mechanism was proposed for the MSM-ABRS of SmB₄ and SmB₆ stated above¹⁸⁵_, although the solubility of B derived from NaHB₄ in the molten salt is unknown.

Metal boride-oxide composite powder

In an MRS process discussed above, the metal reductant used (e.g. Mg, Si and Al) is oxidised, forming its corresponding oxide (e.g. MgO, SiO₂ and Al₂O₃). So, the final product from the reaction is a hybrid mixture of primary product boride and by-product oxide. Owing to its high quality and good properties inherited from its two constituent phases, such a hybrid powder can be used directly as the main raw material to prepare a corresponding bulk composite with homogeneous microstructure and excellent properties, avoiding several problems which otherwise would occur, including wasting the by-product oxide, requirement of acid leaching, and difficulty in dispersion. In the case of using Al as a reductant, a range of high-quality metal boride–alumina composite powders can be prepared via MSM-MRS (Reaction (21)):

M O x + B 2 O 3 + Al → M B y + A l 2 O 3

(21)

OPEN IN VIEWER

Figure 16.

SEM images (a, b) and corresponding EDS (c, d) of the TiB₂–Al₂O₃ composite powder resultant from aluminothermal reduction in NaCl at 1150 °C for 4 h.

Interestingly, if MgCl₂ is used to form the salt medium, once Al₂O₃ is formed via Reaction (21), it can further react with the MgCl₂ to form another important ceramic material, magnesioaluminate spinel (MgAl₂O₄) (Reaction (22)). In this way, a range of metal boride–MgAl₂O₄ composite powders also can be readily prepared.

4 A l 2 O 3 + 3 MgC l 2 = 3 MgA l 2 O 4 + 2 AlC l 3

(22)

Metal boride-carbide composite powder

A transition metal boride-carbide composite powder, in the form of core-shell structured hybrid, also can be fabricated via in-situ reaction of a transition metal with a carbide core in an appropriate salt. As illustrated by Wang et al. and Chen et al., by reacting Ti or Hf with B₄C particles in the NaCl–KCl eutectic salt at a relatively low temperature of 1000 °C, TiB₂- or HfB₂-coated B₄C particles were successfully formed (Reactions (23) and (24)) (Figure 17).^190–192

3 M + B 4 C = MC + 2 M B 2 ( M = Ti or Hf )

(23)

2 MC + B 4 C = 2 M B 2 + 3 C

(24)

OPEN IN VIEWER

Figure 17.

Schematic of the MSS process used to prepare B₄C–HfB₂ composite powder (after Ref. ¹⁹⁰ ).

The ‘template-growth’ mechanism clearly was dominant in these cases. B₄C hardly, whereas Ti/Hf partially, dissolved in the molten salt. So, the latter arrived at the surface of the former via diffusion through the molten salt medium and then reacted in-situ to form a boride-based shell.^190–192 The thickness of the boride-based shell was readily controllable by controlling the relative amount of the metal used, which was very useful in tuning the compositions of boride and B₄C in the final bulk composites fabricated by using such novel structured hybrid powders, and thus their properties (see Section ‘Preliminary applications of molten salt synthesised boride powders’ below).

Ternary metal borides

Apart from binary borides containing one metal element, ternary borides containing two metal elements, especially those containing a transition metal and Al or Zn (the so-called ‘MAB phases’) have recently attracted great research interest, because of their unique layered structures and superior properties.^193–200 Several techniques have been attempted to synthesise both bulk- and powder-formed ternary borides.^193,197,198 Due to the involvement of two metal elements, more intermediate phases (binary and ternary phases) might be formed, making the synthesis process more complicated, so, a relatively high temperature and/or a very long time is required to ensure the reaction completion and high product purity.^197,198

The ‘post-processing of arc-melted ingots’ was initially used for the synthesis of ternary boride. In the preparation of Fe₂AlB₂, a mixture of elemental reagents, or binary boride and large amounts of excess Al (c.a. 50–200 mol.-%), was arc-melted, followed by crushing and reannealing for up to 14 days at 800–1000 °C.^193,197 On annealing, the formation of ternary boride increased at the expense of the binary borides. The use of large amounts of extra Al not only compensated for its evaporation during the arc melting but also contributed to the stabilisation of large amounts of binary Al-rich intermetallics. After elimination of these intermetallics via preferential HCl leaching from the annealed product, nearly phase-pure target ternary boride was obtained.^193,197 Nevertheless, the disadvantages of this technique are obvious, including requirement of high temperature, energy- and time-consuming, use of large amounts of excess Al, and involvement of extensive acid-leaching. Another relatively easier method is to heat a powder mixture of binary boride, metal Al, and B in an inert atmosphere or vacuum at a relatively low temperature (<1200 °C) for a relatively short period (<15 h).^193,198 Unfortunately, with this technique, a pre-synthesised binary boride needs to be used. Moreover, it is difficult to form high quality ternary boride powders, since some intermediate/impurity phases (e.g. binary borides) almost always remain in the final product powder. To overcome the disadvantages of these two conventional methods, the MSM-ERS method discussed earlier (Section ‘‘Conventional and molten salt mediated elemental reaction synthesis’’) has been attempted to prepare a ternary boride recently by the present authors. ²⁰¹

As presented in Figure 18, phase pure MoAlB product particles were successfully prepared via MSM-ERS in NaCl after 6 h at 1000 °C. ²⁰¹ They exhibited three different shapes/sizes: spherical grains (1–3 µm), plate-like particles (<5 µm in diameter) and columnar crystals with lengths up to 20 µm and diameters up to 5 µm, resultant from different intermediate reactions. The synthesis condition (1000 °C for 6 h) was much milder than in the cases of using the two conventional techniques stated above, although the use of excess Al and post-firing acid leaching was still required. ²⁰¹

OPEN IN VIEWER

Figure 18.

(a) Low and (b) high magnification SEM images of MoAlB particles resultant from MSM-ERS in NaCl at 1000 °C for 6 h and the subsequent acid leaching. ²⁰¹

The MSM-ERS approach was similarly used to synthesise Cr₂AlB₂ in a KCl–NaCl eutectic salt. Like in the MoAlB synthesis, several intermediate steps were involved in the synthesis process, leading to the formation of several intermediate phases such as Cr₅Al₈ and Cr₉Al₁₇. ²⁰² Despite this, nearly phase pure Cr₂AlB₂ (with minor Cr₅Al₈, CrB and Al₂O₃) was obtained after only 0.5 h at 1100°C, although large amounts of excess Al (100 mol% extra) were also used to compensate for the evaporation loss of Al. ²⁰²

Boride solid solutions/high entropy borides

The molten salt mediated synthesis techniques reviewed above (Sections ‘Conventional and molten salt mediated elemental reaction synthesis’ to ‘Conventional and molten salt mediated alkaline borohydride reduction synthesis’) can be readily extended/modified to synthesise a range of single-phase solid solution powders comprising two or more borides, at a relatively low temperature, provided that their formation is thermodynamically favourable. Such boride solid solutions are the ‘new’ members of solid-solution family, which have been attracting a great deal of attention because of their much better properties and performance than their constituent borides.^203,204

Under MSS conditions, whether two or more borides form a homogeneous single-phase solid solution, or a physical mixture, can be predicted according to the criteria established based on thermodynamics.^204–206

Based on the mixing Gibbs free energy value and the lattice size difference determined by the first principles calculations, Wen et al. ²⁰⁵ evaluated theoretically the possibility of formation of (Hf_1/3Zr_1/3Ti_1/3)B₂ (HZTB) and (Ta_1/3Nb_1/3Ti_1/3)B₂ (TNTB) boride solid solutions. The latter was predicted to be more likely to form than the former. This prediction was verified experimentally via MSM-BRS in NaCl–KCl at 1100 °C for 1 h (using 30% excessive B). As predicted, high purity single phase TNTB nanorods of 20–40 nm in diameter and 100–200 nm in length were successfully formed, whereas no single phase HZTB solid solution was formed under identical conditions.

The same group further established a ‘new’ set of criteria for the formation of a boride solid solution, by combining the first principles calculations with machine learning. ²⁰⁶ A series of (M _x N_1−x)B₂ (where M, N = Hf, Zr, Ta, Nb, Ti) solid solutions were employed as working tools. Four important parameters, atomic size difference (δ), mixing enthalpy at 0 K and 0 Pa (ΔH^0K_mix), doping condition (φ), and valence electron concentration (VEC) were correlated, and their effects on the formation ability of (M_xN_1−x)B₂ solid solutions were found to be in the following descending order: δ > ΔH^0K_mix> φ > VEC. In addition, a new parameter, β, which has linear relationships with these four factors, was further introduced, to evaluate the formation ability of (M_xN_1−x)B₂ solid solution. When β > 0, a single-phase (M_xN_1−x)B₂ solid solution will form, otherwise, a physical mixture of multiphase will form. ²⁰⁶ This work provided a theoretical tool for the discovery and design of various boride solid-solutions as well as other types of solid-solutions (e.g. metal, oxide and carbide solid solutions) with novel properties. To validate the ‘new’ criteria, MSM-BRS was carried out using two oxide precursors. The results showed that a series of solid solution comprising two borides which met the criteria were successfully formed after 1 h in NaCl/KCl at 1150 °C. ²⁰⁶

By using MSM-BRS in NaCl–KCl, a few other nanosized boride solid solutions were also prepared, including Ti_0.2Zr_0.8B₂, ²⁰⁷ Ni _x Fe_1−xB (x = 0.25, 0.5, 0.75) ²⁰⁸ and Hf _x Zr_1−xB₂ (x = 0.2, 0.5, 0.8), ²⁰⁹ at 1050–1100 °C for 1 h, 800 °C for 8 h, and 1150 °C for 1 h, respectively. In these cases, the products exhibited high composition uniformity and good single hexagonal structure.

In addition to MSM-BRS, in principle, the other MSM synthesis approaches reviewed above could be used similarly to synthesise boride solutions. For example, by using the MSM-ABRS technique, La_1-xCe _x B₆ (x = 0.2, 0.4, 0.6, 0.8) was successfully prepared from LaCl₃, CeCl₃ and NaBH₄, in a LiCl–KCl eutectic salt at 800°C for 1 h. ²¹⁰ During the reaction process, LaCl₃ and CeCl₃ reacted with NaBH₄, forming LaB₆ and CeB₆, respectively (Reactions (25) and (26)).

LaC l 3 + 6 NaB H 4 = La B 6 + 3 NaCl + 3 Na + 12 H 2

(25)

CeC l 3 + 6 NaB H 4 = Ce B 6 + 3 NaCl + 3 Na + 12 H 2

(26)

Based on the calculated lattice size difference (δ) and the Gibbs free energy of the solid solution formation, Liu et al. ²¹¹ predicted the favourable formation of (Ta_0.2Nb_0.2Ti_0.2W_0.2Mo_0.2)B₂ high-entropy diboride (HEB-1), and experimentally synthesised it in KCl at 1150 °C for 30 min via the MSM-BRS route. As-prepared HEB-1 showed an interesting chrysanthemum-like morphology composed of numerous boride nanorods of 20–30 nm in diameter and 100–200 nm in length (Figure 19(a) and (b)). The boride nanorods possessed a single-crystalline hexagonal structure, and high compositional uniformity from nanoscale to microscale, as revealed by EDS (Figure 19(c)). ²¹¹

OPEN IN VIEWER

Figure 19.

(a) Low-magnification and (b) high-magnification SEM images of HEB-1 resultant from MSM-BRS in KCl at 1150 °C for 30 min and (c) the EDS elemental mapping from the highlighted area in (a).

On the other hand, by MSM-MRS in MgCl₂, Ye et al. successfully synthesised single phase nanosized powders of four different types of high entropy boride ceramics, (Zr_0.25Ta_0.25Nb_0.25Ti_0.25)B₂, (Hf_0.25Ta_0.25Nb_0.25Ti_0.25)B₂, (Hf_0.25Zr_0.25Nb_0.25Ti_0.25)B₂ and (Ta_0.25Nb_0.25Ti_0.25Cr_0.25)B₂ from oxide precursors, at 1000 °C for only 30 min. These nanopowders had particle sizes of 28–56 nm, showed good compositional homogeneity and contained low level of oxygen impurity. ²¹² Recently, Gao et al. synthesised (Hf_0.2Ti_0.2Mo_0.2Ta_0.2Nb_0.2)B₂ nanopowder via MSM-BRS in KCl at 1300 °C for 6 h, and predicted its properties such as hardness and thermal conductivity, based on the density functional theory (DFT) calculations. ²¹³

While in many cases single-phase high entropy boride powder could be synthesised directly at a relatively low temperature via one of the molten salt mediated routes discussed above, in some cases, a mixture of multiple phases was obtained. Even so, owing to its good quality such as homogeneous structure and composition, high surface area and high reactivity/sinterablity, it could still be readily converted into the corresponding single phase high entropy boride powder or bulk ceramic, via post-synthesis heat-treatment or densification, under appropriate conditions, as illustrated by Guo et al. ²¹⁴ Despite the fact that the product powder resultant from MSM-BRS in NaCl–KCl at 1100 °C comprised two boride phases, (Hf,Zr)B₂ and (Ta,Nb,Ti)B₂, after 10 min spark plasma sintering (SPS) at 1900 °C and 50 MPa, the originally desired single-phase bulk (Zr_0.2Hf_0.2Ta_0.2Nb_0.2Ti_0.2)B₂ high-entropy ceramic was still obtained. ²¹⁴

To better understand the current situation in this specific area, Table 6 outlines the main works to date on MSS of boride solid solutions.

OPEN IN VIEWER

Table 6.

Boride solid solutions synthesised via MSS.

Boride solution	Metal source	Boron source	Boride shape and size	MSS route	Salt and firing condition	Reference
(Ta1/3Nb1/3Ti1/3)B2	Ta2O5 and Nb2O5: 1–3 μm; TiO2 powders: <300 nm	B: <3 μm	Clusters of nanorods with diameter of 20-40 nm and length of 100-200 nm	MSM-BRS	NaCl–KCl, 1100 °C/1 h	205
M x N1−xB2 (M, N = Hf, Zr, Ta, Nb, Ti)	HfO2, Nb2O5 and Ta2O5: 1–3 μm ZrO2 and TiO2: 100–300 nm	B: 1–3 μm	Clusters of nanorods with diameter of 20–30 nm and length of 100–200 nm		NaCl/KCl, 1150 °C/1 h	206
(Hf x Zr1−x)B2 (x = 0.2, 0.5, 0.8)	HfO2 powder: 1–3 μm ZrO2 powder: <300 nm	B: < 3 μm	Irregular polyhedrons: 170–185 nm			209
Ti0.2Zr0.8B2	TiO2 and ZrO2：100–300 nm	B: < 3 μm	Particles of 86 nm		NaCl–KCl, 1050 °C/1 h	207
Ni x Fe1−xB	Oxides	B	Nanosheets		NaCl–KCl, 800 °C/8 h	208
La1-xCe x B6 (x = 0.2, 0.4, 0.6, 0.8)	LaCl3, CeCl3	NaBH4	Cubic particles of 90 nm	MSM-ABRS	LiCl–KCl, 800 °C/1 h	210
(Ta0.2Nb0.2Ti0.2W0.2Mo0.2)B2	Nb2O5 and Ta2O5 powders: 1–3 μm TiO2, WO3, and MoO3 powders: 100–300 nm	B: <3 μm	Chrysanthemum-like clusters of nanorods with diameters of 20–30 nm and lengths of 100–200 nm	MSM-BRS	KCl, 1423 K/30 min	211
(Zr0.25Ta0.25Nb0.25Ti0.25)B2, (Hf0.25Ta0.25Nb0.25Ti0.25)B2, (Hf0.25Zr0.25Ta0.25Nb0.25)B2, (Ta0.25Nb0.25Ti0.25Cr0.25)B2,	Metal oxides	B2O3	Nanoparticles: 28–56 nm	MSM-MRS using Mg	MgCl2, 1000 °C/30 min	212
(Hf0.2Ti0.2Mo0.2Ta0.2Nb0.2)B2	Metal oxides	B	Nanoparticles: 35 nm	MSM-BRS	KCl, 1300 °C/6 h	213
(Zr0.2Hf0.2Ta0.2Nb0.2Ti0.2)B2	50–600 nm	B: ∼1 μm	Mixture of two borides	MSM-BRS	NaCl–KCl, 1100 °C/1 h	214

Microwave-assisted molten salt synthesis

Given its distinct advantages such as energy saving, high efficiency heating (>400 °C min⁻¹), and substantial reduction in processing time and temperature, microwave heating has been used in many industrial processes as an alternative heating route. ^215–218 It also has been introduced by the corresponding author's group to an MSS process to replace the conventional furnace heating, leading to the development of an improved MSS technique, referred to as ‘microwave-assisted molten salt synthesis (MA-MSS)’.^219–225

Table 7 summarises the studies on MA-MSS of metal boride-based powders.^219–225 As shown, by using microwave heating combined with MSM-BCRS, phase pure ZrB₂ nanorods, TiB₂ microplatelets and TaB₂ nanoflowers were successfully prepared in the NaCl-KCl eutectic salt after only 20 min at 1100–1200 °C. Such a condition was milder than that in the MSM-CRS/BCRS and even MSM-MRS using conventional heating (Sections ‘Conventional and molten salt mediated borothermal reduction synthesis’ and ‘Conventional and molten salt mediated carbo- and boro/carbo-thermal reduction synthesis’), and further milder than in the conventional CRS/BCTR. Under the conditions listed in Table 7, the reaction extent in the case of either without using a molten salt or without using microwave heating was much lower, demonstrating the synergistic effect from the combination of molten salt with microwave heating.^219–225

OPEN IN VIEWER

Table 7.

List of boride-based materials synthesised via MA-MSS in NaCl–KCl.

Boride-based system	Raw materials	Boride/carbide shape and size	Synthesis condition	Reference
TiB2	TiO2, B4C, amorphous C	Microplatelet: 4.2 µm in plane direction and 0.6 µm in thickness	1150 °C/1 h or 1200 °C/20 min	219
ZrB2	ZrO2, B4C, amorphous C	Nanorod: 40–80 nm in diameter, aspect ratio >10	1200 °C/20 min	220
TaB2	Ta2O5, B4C	Nanoflower composed of nanorods of 410 nm in diameter and 4.16 µm in length		221
(Hf x Zr1−x)B2 (x = 0.25, 0.50, 0.75)	HfO2, ZrO2, B4C and amorphous C	Micro-rod: length of 27.8–41.6 μm and high aspect ratio of 10.1–17.2	1100 °C/20 min	222
ZrB2–SiC	ZrO2, SiO2, B4C, amorphous C	ZrB2 nanorod and SiC nanosheet	1200 °C/20 min	223
	ZrSiO4, B4C, C	ZrB2 angular platelet several µm	1250 °C/20 min	224
HfB2–SiC	HfO2, SiO2, B4C, C	HfB2 nanorod of 5 µm in length and 300 nm in diameter and SiC particles of 300 nm		225

Apart from a single metal boride, a metal boride solid solution or composite powder comprising a boride and another component such as a carbide can be prepared similarly under milder conditions by using MA-MSS. As shown in Table 7, (Hf _x Zr_1−x)B₂ microrods were synthesised from oxide precursors, B₄C and amorphous carbon, in NaCl–KCl after only 20 min at as low as 1100 °C. ²²² On the other hand, ZrB₂–SiC composite powders comprising platelet-like ZrB₂ and SiC were successfully synthesised from ZrSiO₄, B₄C and amorphous C in the same salt after only 20 min at 1200 °C. ²²⁴ In addition, HfB₂-SiC powders were similarly synthesised in the same salt after 20 min at 1250 °C, from HfO₂, SiO₂, B₄C and amorphous C. ²²⁵ These synthesis conditions were also much milder than in the cases of conventional CRS and BCRS (a few hours at >1500 °C)^226,227 and even MSM-CRS/BCRS using the conventional heating method (Section ‘Metal boride–carbide composite powder’).

Interestingly, MA-MSS not only reduces the synthesis temperature and/or time, but also enables the formation of anisotropic boride crystals (Table 7 and Figure 20), especially in the cases involved with ZrO₂, C, SiC and B₄C which are good susceptors for microwave heating.^218,228 In all the MA-MSS cases shown in Table 7, the boride phases showed different morphologies and/or sizes from those of the raw materials precursors (reactants), indicating that the ‘dissolution-precipitation’ mechanism dominated their formation processes. The reason was not clear, although it could be that the presence of microwave susceptors such as ZrO₂, C, SiC and B₄C increased local temperature of the salt surrounding them, increasing the solubilities of the reactant species (especially originally less-soluble ones) in it. However, this still needs to be clarified further by future work.

OPEN IN VIEWER

Figure 20.

Microstructural morphologies of (a, b) ZrB₂, ²²⁰ (c, d) TiB₂ ²¹⁹ and (e, f) TaB₂ ²²¹ particles resultant from MA-MSS.

Molten salt shielded synthesis

In all the MSS studies reviewed above, an inert atmosphere, for example Ar, or vacuum, was used to protect non-oxide raw materials (such as metal, boron, carbide and carbon) and boride products from oxidation, making the synthesis processes complex and expensive, limiting their use on an industrial scale. Given the good shielding effect of molten salt itself, the requirement of a protective atmosphere or vacuum could be avoided via slightly modifying the conventional MSS processes discussed above.

In MSS, once the salt melts, it will cover the reaction system, isolating it effectively from the contact with the oxidising atmosphere. So, the possible oxidation mainly occurs within the temperature range from the onset oxidation temperature of a non-oxide reactant to the melting (eutectic) point of the salt. Clearly, if this range is narrow and/or the heating rate is sufficiently high, then the oxidation extent before the salt melts will be limited. If this is the case, then no protective atmosphere would be required, as demonstrated by Ma et al. who synthesised ZrB₂ nanosheets in KCl in air at 800 °C, from Zr(OH)₄ and NaBH₄, and did not find any obvious oxidation and impurity phase formation. ²²⁹ On the other hand, if this range is broad, and/or the heating rate is low, then there will be a potential risk of oxidation before the salt melts. This could be addressed by encapsulating a reaction sample with a solid salt layer via simply pressing. Given its high ductility at room temperature and high bulk density (>95% dense) after cold-pressing, KBr is typically used to form the gas-tight layer around a reaction sample before it is placed in a medium-forming salt bed for further heating (Figure 21-I). In this way the sample can be protected from oxidation before the salt melts. Once the salt melts, it will continue to protect the whole reaction system from oxidation (Figure 21-II). After the reaction and furnace cooling to room temperature, the product can be collected by following the general practice, that is water leaching and filtration/separation (Figure 21-III and IV). ²³⁰ To illustrate the feasibility of such a ‘molten salt shielded synthesis (MS3)’ technique, several nonoxides including MoAlB, as working tools, were synthesised. ²³⁰

OPEN IN VIEWER

Figure 21.

Schematic of a typical MS3 process. ²³⁰

Following this pioneer work, several ternary boride (MAB) phases, including Fe₂AlB₂, Mn₂AlB₂ and MoAlB, were further synthesised via the MS3 route at 1000 °C. ²³¹ The first two were obtained in KBr by using a nominal starting composition T₂Al_1.2B₂ (T = Fe, Mn), and the third one was synthesised in NaCl with a nominal composition of MoAl_1.2B₂. In addition, instead of using single salt KBr and NaCl, a KCl–NaCl eutectic salt mixture was used as the alternative shielding salt. ²³² Because of its low eutectic point (657 °C), it not only protected more effectively the oxidation prone reactants from oxidation, but also enabled the completion of synthesis reactions at a lower temperature for a shorter soaking time. By using such a modified MS3 method, WAlB and Mn₂AlB₂ phases with reasonably high purity were synthesised at about 900 and 1000 °C, respectively, for only 2 h. ²³²

The advantages of MS3 are obvious, including energy-saving, cost reduction, and ease of scale up. However, it also suffers from some disadvantages. For example, it requires preparation of salt-encapsulating samples, and in many cases, it would not be possible to protect the reaction system completely from the oxidation, because even after the salt melts, oxygen can still diffuse through the molten salt medium, potentially oxidising the non-oxide phases in it. The actual oxidation extent will be dependent on several factors such as reactivity and onset oxidation temperature of non-oxide phase, the diffusion rate of oxygen in the molten salt, salt type and amount, reaction temperature and holding time, and heating and cooling rates, all of which need to be investigated further.

Waste-water treatment

Some boride powders synthesised by molten salt mediated methods were tested as absorbents for waste-water treatment.^187–189 50 nm CeB₄ and 15–30 nm CeB₆ particles formed via MSM-ABRS (Section ‘Conventional and molten salt mediated alkaline borohydride reduction synthesis’) showed remarkable performance in the removal of azo-dyes such as Congo red (CR), PS, and MO from their aqueous solutions.^187,188 For example, in the case of using CeB₄ for CR removal, the adsorption ratio reached 94.1–97.8% for solutions with initial concentrations of 100–500 mg L⁻¹, and the maximum removal capacity reached 491.8 mg g⁻¹. ¹⁸⁷ And in the case of using CeB₆, the maximum removal capacity reached as high as 775.6 mg g⁻¹. Such a good performance was attributed to its high adsorption ability, and additional high photocatalytic activity in the CR degradation. ¹⁸⁸ On the other hand, 100 nm cubic GdB₆ nanoparticles synthesised via MSM-ABRS also showed an excellent performance in methyl blue (MB) removal. Under their catalysis, MB was reduced to colourless leucomethylene (LMB). An adsorption capacity of 3108.8 mg g⁻¹, along with an adsorption ratio of 77.7%, was achieved. ¹⁸⁹ It is worthy to note that the above boride absorbents also demonstrated excellent recyclability/reusability in the azo-dyes removal. For example, the performance of GdB₆ particles remained unchanged even after six adsorption-desorption cycles. ¹⁸⁹

Electrocatalysts

Some transition metal borides ‘unexpectedly’ showed excellent catalytic activity and could be potentially used for rechargeable battery and electrolytic hydrogen generation. Thanks to the synergy from their high catalytic activity and electronic conductivity and novel porous sheet structure, CoB nanosheets prepared via MSM-ABRS (Figure 15) acted as an excellent cathode catalyst for high performance rechargeable Na–O₂ batteries, as indicated by a low charge overpotential, good rate capability, high specific capacity (11482 mAh g⁻¹) and long cycle life (74 cycles). Such transition metal borides were also considered to be applicable to other important areas such as energy storage and conversion. ¹⁸⁶ Furthermore, Ni _x Fe_1−xB particles prepared via MSM-BRS (Table 6) were identified as a robust bifunctional electrocatalyst for overall water splitting, that is for both the hydrogen evolution reaction (HER) and the oxygen evolution reaction (OER). ²⁰⁸ A current density of 10 mA cm⁻² was achieved at overpotentials of 63.5 and 282 mV for the HER and the OER (in 1 M KOH electrolyte), respectively. DFT calculations reveal that Ni _x Fe_1−xB has a lower absolute value of H* adsorption free energy than FeB and NiB, which is responsible for the improved HER activity. When two identical Ni _x Fe_1−xB electrodes were combined to form an alkaline electrolyzer, a current density of 10 mA cm⁻² was achieved at a small cell voltage of 1.57 V for overall water splitting, and the stability could be maintained beyond 20 h, which was close to that of a Pt/C//RuO₂ electrolyzer. ²⁰⁸ Similarly, Mo–Ru–B electrocatalysts synthesised via MSM-BRS from MoCl₅, B, and RuCl₃·xH₂O in molten KCl–NaCl at 900 °C for 2 h exhibited high catalytic activity in HER in an alkaline, acidic, or neutral solution, along with excellent stability in several types of electrolytes. ²³⁸

On the other hand, to address the intrinsic low electrocatalytic activity of β-MoB₂ for the HER (due to the presence of puckered B layers), β-MoB₂ nanosheets with a BET surface area of 48 m² g⁻¹ were formed at 850 °C via MSM-BRS ¹¹⁶ (Section ‘Conventional and molten salt mediated borothermal reduction synthesis’), which exhibited a good HER performance in an acidic medium with an overpotential of 187 mV, a current density of 10 mA cm⁻² and a Tafel slope of 49.3 mV per decade. Such a performance was much better than that of β-MoB₂ in other forms, and even close to that of α-MoB₂ which was considered as the best molybdenum boride electrocatalyst for HER. The much improved HER activity of the β-MoB₂ nanosheets was mainly attributed to their novel morphology, exposing simultaneously both less- and more-active sites, counteracting the effect of less active puckered phosphorene-like B layers in the conventional β-MoB₂. ¹¹⁶

Twelve other types of TMDB powders obtained via MSM-BRS/ERS (Figure 8) were also tested and compared as electrocatalysts for the HER. ⁹¹ ZrB₂, HfB₂ and TaB₂ showed relatively poor catalytic performances, TiB₂ and NbB₂ showed moderate catalytic activities, and the others showed desirable catalytic activities. Among these borides, RuB₂ showed the highest catalytic activity, achieving 10 mA cm⁻² at an overpotential of 35 mV, which was close to that of 20% Pt/C in an acidic solution, and WB₂ performed the best among the non-precious metal diborides, achieving 10 mA cm⁻² at an overpotential of 203 mV. The performance of the latter could be further enhanced via doping with Ni.

In addition, layer structured Mo₂AlB₂ prepared via etching MoAlB in KCl–NaCl by using CuCl₂ as the etching agent performed much better than MoAlB in the HER catalysis in an alkaline solution, ²³⁹ thanks to its enhanced active surface area and improved charge transfer.

Microwave absorbent

It remains a great challenge to develop high performance microwave absorbers used in some harsh environments. By using the MSS approaches discussed in the previous sections, some borides with novel morphologies could be synthesised for such a purpose. The porous microspheres comprising Fe₂B nanoflakes prepared via MSM-ERS. ⁸⁸ showed a magnetic loss factor larger than the dielectric loss factor at 2–18 GHz, owing to their shape anisotropy and multiple magnetic resonances. Consequently, their composite with paraffin showed a good microwave absorption performance. For 1.8 and 1.9 mm thick layered absorbers, the effective absorption band, in which the absolute value of the reflection loss exceeded 10 dB, reached 4.4 GHz, covering 13.6–18 GHz and 12.8–17.2 GHz, respectively. In another study by Feng et al., TiB₂ microplatelets decorated with multi-layered TiC nanosheets (TiB₂/TiC hybrids) were in-situ fabricated via an MSM-CRS route (Figure 23). ²⁴⁰ With increasing the reflective interface, such a hybrid exhibited improved microwave absorption performance with a minimum reflection loss value of −52.5 dB (13.1 GHz, 5.2 mm). The reasonable dielectric loss, higher dipole moments and interfacial effect, along with the 3D multilayer network established by TiC nanosheets, conferred the high microwave absorption capacity on the hybrid.

OPEN IN VIEWER

Figure 23.

Scheme of microwave absorption mechanism of the TiB₂/TiC hybrids prepared via MSM-CRS in NaCl.

Bulk metal boride-based ceramics with improved microstructure and properties

Given their high quality, metal boride powders formed via an MSS route could be used directly to make bulk metal boride monoliths or boride-based composites with improved microstructures and properties, avoiding or alleviating the issues such as prolonged mixing, milling and dispersion, as required conventionally.

Phase pure LaB₆ nanoparticles of approximately 94.7 nm prepared via MSM-ABRS at 800 °C for 1 h were used to fabricate a bulk LaB₆ ceramic via 30 min hot pressing under 1800 °C and 50 MPa.²⁴¹ Due to the high dispersivity and high surface area of the original LaB₆ nanoparticles, the bulk LaB₆ with a relative density as high as 98% was obtained. The onset sintering temperature and full densification temperature were 1050 and 1700 °C, respectively, which were respectively about 550 and 300 °C lower than in the case of using coarser commercial powders. The bulk LaB₆ ceramic also exhibited a low work function of 2.87 eV and an excellent thermionic emission performance with the saturation emission current density at 1500 and 1600 °C of 37.4 and 44.3 A cm⁻², respectively. ²⁴¹ Similarly, a bulk high entropy ceramic of (Zr_0.2Hf_0.2Ta_0.2Nb_0.2Ti_0.2)B₂ was prepared via SPS at 1900 °C by using the highly active precursor prepared via MSM-BRS (Table 6), ²¹⁴ which showed improved mechanical properties, such as a Vickers hardness of 32.5 GPa, and fracture toughness of 5 MPa m^−1/2.

On the other hand, two types of metal boride-based composite powders, physical mixture of borides, and core-shell structured hybrid powders, resultant from in-situ MSS discussed above, could also be used directly to make the corresponding bulk composites with much improved microstructure and properties.

ZrB₂–SiC composite powders comprising rod-like ZrB₂ and sheet-like SiC prepared via MA-MSS (Table 7) were used directly to prepare the corresponding bulk ceramics via SPS. The resultant composite showed a high density and excellent mechanical properties including Vickers hardness of 24.5 GPa and fracture toughness of 4.8 MPa m^−1/2. ²²³

By using molten salt synthesised HfB₂ particles covered with an amorphous boron-rich shell as the starting precursor, Gouget et al. fabricated, via SPS, a bulk nanocomposite in which high electrical conductivity (‘metallic’) HfB₂ nanocrystals of controlled amounts were distributed homogeneously in a low thermal conductivity amorphous boron-rich matrix, potentially achieving a ‘trade-off’ between enhanced electrical conductivity and low thermal conductivity in the boron-rich matrix. ²⁴² By further oxidising the boron-rich shell covering HfB₂ particles, and high temperature and high pressure treatment (densification), while the particle size of HfB₂ was retained, the original oxidised amorphous boron-rich layer could be recrystallised, forming a matrix composed of nanosized β-HfB₂O₅. ²⁴³ This approach paves an avenue for the design and development of a novel composite with nanoscaled heterostructures possessing various advanced mechanical, electrical, thermal and/or optical properties.

A similar strategy was also used to fabricate B₄C-metal boride functional composites, aiming to overcome/alleviate several problems with monolith B₄C, including mechanical property degradation under high external pressure or shear stress, high electrical resistance and low fracture toughness. Based on the ‘template-growth’ mechanism of MSS (Section ‘Mechanisms’), HfB₂- or TiB₂-coated B₄C hybrid particles were prepared via in-situ reaction of Hf or Ti with B₄C surface in a molten salt,^190–192 as mentioned earlier. Such core-shell structured hybrid particles were used directly to make HfB₂ or TiB₂ bonded B₄C composites via SPS.

Both HfB₂- and TiB₂-bonded B₄C composites exhibited enhanced properties. The former possessed much enhanced mechanical properties (sliding friction decreased by ∼36%, and the average hardness increased from 512.1 to 2151.9 HV) and wear resistance (increased by ∼57%), compared to monolith B₄C. The interconnected 3D boride network effectively restrained the plastic deformation in B₄C grains, avoiding premature property attenuation. ¹⁹⁰ In the case of 29.8 vol.-% TiB₂-bonded B₄C, an elastic modulus of 515.6 GPa, Vickers hardness of 32.1 GPa, fracture toughness of 4.38 MPa m^1/2, electrical conductivity of 4.06 × 10⁵ S m⁻¹, and thermal conductivity of 33 W m⁻¹ K⁻¹, were achieved, ¹⁹² most of which were better than those reported for the composites with similar compositions but prepared by using raw material precursors from other sources. Such a composite also exhibited some good functional properties. Due to the high thermal conductivity of the 3D network of TiB₂, the final composite showed an improved thermal conductivity. It also exhibited a better resistance to disintegration under irradiation and in the extreme environment of a nuclear reactor than pure B₄C pellets. In addition, it has a great potential for the application as neutron-absorbing pellets in nuclear reactors.