Review: titanium–titanium boride composites

Abstract

Titanium in situ reinforced with titanium boride (TiB_w) has garnered significant attention as one of the most promising titanium composites today. A significant number of processing approaches have been reported on the fabrication of these unique composites. This brief review provides an update on recent activities within the area focusing largely on research conducted over the past 5 years. Approaches discussed include bulk deformation processes such as rolling, 2D forging, equal channel angular pressing, (laser and non-laser-based) additive manufacturing, and new microstructural designs implemented for this composite system. Moreover, the potential of this composite as a biomaterial is also discussed.

Introduction

The interest in titanium (Ti)–titanium boride (TiB_w) composites has been significant over the past 5 years. It is well known that Ti inherently suffers from poor wear resistance and relatively low stiffness, and hence the incorporation of reinforcements in the form of particulates or whiskers has been successful in eliminating such disadvantages. One of the most successful reinforcement for titanium is in-situ-synthesized single-crystal TiB_w, to produce Ti–TiB_w composites. These lightweight composites provide advantages that include enhanced stiffness, hardness, strength, and wear resistance over unreinforced Ti [1]. The incorporation of TiB_w also provides for a chemically and mechanically compatible reinforcement with the Ti matrix. Applications/potential applications for Ti composites range from automotive and aerospace to sporting goods and even biomedical [2]. Recently, interesting approaches have been published on Ti–TiB_w composites, including selective laser melting [3], Ti-based hybrid composites [4], forging [5], extrusion [6], equal channel angular pressing [7], hot rolling [8], spark plasma sintering [9], various designs for spatial addition of TiB_w [10] and work that considers Ti–TiB_w as a biomaterial [11, 12].

Although very valuable earlier reviews have been published, recent reviews on titanium composites include the selective laser melting of Ti and Ti composites for biomedical applications [13], and a 2008 review on the processing, microstructure, and properties relations in particulate and whisker-reinforced titanium [14]. In 2007, the present author published a review with a specific focus on the processing and properties of Ti–TiB_w composites. Due to the important and significant research published recently on the topic, the present review provides a brief update on Ti–TiB_w composites, focusing largely on work published during the past 5 years. The review does not cover all the papers published but considers select papers to exemplify the recent directions in Ti–TiB_w research. The paper specifically tackles new research related to bulk deformation processing, interesting microstructural arrangements, additive manufacturing, and the consideration of Ti–TiB_w as a biomaterial.

Background

Powder Metallurgy Processing and TiB_w Aspect Ratio Effects

One of the fundamental methods for forming titanium–titanium boride composites is through the process of powder metallurgy [15]. This type of processing offers the advantages of net shape processing, minimal material waste, and microstructural control [16]. In a recent comparative study, Selvakumar et al. [17] investigated the microstructure and properties of Ti–TiB_w produced through spark plasma sintering (SPS), versus hot isostatic pressing (HIP) and vacuum sintering. Only the spark plasma-sintered products had predominantly converted the TiB₂ to fine TiB_w whiskers. This is not surprising as electric current had been previously shown to have an intrinsic effect that increases the diffusion and phase transformation kinetics [10, 18]. On the other hand, the morphology of TiB in Ti–TiB samples produced by HIP was spherical, while platelike agglomerates as well as fine whiskers were found in vacuum-sintered materials, together with residual porosity and TiB₂. These different microstructures consequently resulted in varying mechanical responses. For example, spark plasma-sintered materials mostly exhibited enhanced properties compared with hot isostatically pressed or vacuum-sintered products. An increase in TiB_w content also result in an increase in elastic modulus, hardness, contact stiffness, and creep resistance while decreasing the fracture toughness, which is also supported by other work [15]. The increase in residual titanium content typically results in better fracture toughness. In an interesting study, SPS above and below the β transus has also been shown to result in marked differences in the ductile to brittle transition temperature of Ti–TiB_w composites [19], where composites sintered at 1000 °C exhibited almost a 100 °C increase in the ductile to brittle transition temperature, from 400–500 to 500–600 °C (using the toughness values from the area under the stress strain curve at different temperatures). Although residual TiB₂ was retained following SPS at temperatures as high as 1000 °C, it was the interface between the TiB_w and Ti that was affected by the initial sintering temperature. Here, sintering in the β phase field followed by cooling through the β-α phase transformation temperature resulted in incompatible crystallographic orientation relations between the α matrix and the TiB_w, which promoted increased local strain and dislocation density. However, these were not expected to occur for materials sintered in the α phase field that did not have to undergo a phase transformation. It should be noted that the β transus depends on the Ti alloy composition and is usually determined through differential scanning calorimetry.

To date, property improvements at low TiB_w volume fractions have not been targeted to a significant level. In fact, this is where there could be an opportunity to produce TiB_w with significantly larger aspect ratios than possible at the higher TiB_w contents, where the excessive formation of TiB impedes each others’ continued growth lengthwise. Interestingly, recently as-cast Ti–TiB_w composites with 8.3% TiB reported an aspect ratio of 32:1 [20], while aspect ratios as low as 10:1 have been reported for 30 vol% TiB content [21]. Moreover, previous attempts made to further improve on properties of Ti–TiB_w composites by increasing TiB_w volume fraction have largely led to a catastrophic reduction in fracture toughness and ductility due to whisker interlocking [15]. One of the particles used to form TiB_w in situ is TiB₂ via a reaction with Ti. Very recent work has shown that the aspect ratio of the in-situ-formed TiB_w can increase with a reduction in the initial TiB₂ particle size when it is below 1 μm [22]. The study was conducted for a 1 vol% TiB_w final loading (randomly orientated), with TiB_w aspect ratios up to 58:1 (and 100 nm TiB_w diameter) being reported for an initial TiB₂ particle size as low as 500 nm. However, when the vol% loading of TiB_w was increased to even just 5 vol% (with 3 μm initial TiB₂ particle size), it resulted in a reduction in TiB_w aspect ratio to 13:1. It is logical to assume that there may be opportunities for the synthesis of TiB_w with even greater aspect ratios at low nano-TiB₂ loadings if nanoscale (< 100 nm) TiB_w were somehow aligned (i.e., limiting TiB_w–TiB_w contact), assuming an abundant boron source (the subject of ongoing work by the present author). Previous work by the present author shows fine TiB_w with an average diameter of a few hundred nanometers [23], produced from Ti–TiB₂ composite particles where TiB₂ had been refined through the mechanical milling process (Fig. 1 [23]).

Figure 1

TiB whisker growth from high volume fraction TiB_w–Ti particle in dual matrix composites [23]

It is well known that the elastic modulus of TiB_w is superior to that of Ti, 371 GPa [24] and 105 GPa [25], respectively. Consequently, the increase in TiB_w content has been shown to further result in an increase in the elastic modulus [26, 27] of Ti–TiB_w composites. Moreover, simply the alignment of whiskers can also further result in an increase in the elastic modulus when measured along the direction of alignment [28].

Through extrusion of as-cast Ti–TiB_w composites, Dubey et al. [29] were able to align TiB whiskers (3.1 vol%, with mean aspect ratio ~ 5:1) and produce composites with  % elongation to fractures (~ 9%) and fracture toughness values ~ 50 MPa m^1/2. Noting that bulk deformation processing is known to fracture TiB_w leading to low TiB_w aspect ratios. The reduction in fracture toughness with increase in TiB_w volume fraction has also been reported [30]. It should, however, be noted that ductility and fracture toughness are also dependent upon microstructure, post-processing heat treatment, and composition. The following modified shear lag model [29, 31] can be used to predict the increase in elastic modulus due to the addition of TiB_w.

$$ \frac{{E_{\text{Ti - TiB}} }}{{E_{\text{Ti}} }} = \left( {1 - V_{\text{TiB}} } \right) + V_{\text{TiB}} \left( {\frac{{E_{\text{TiB}} }}{{E_{\text{Ti}} }}} \right)\left[ {1 - \frac{\tanh \left( X \right)}{X}} \right], $$

(1)

where X is given by:

$$ X = \frac{l}{{d\sqrt {\left( {1 + v_{\text{Ti}} } \right)\left( {\frac{{E_{\text{TiB}} }}{{E_{\text{Ti}} }}} \right)\ln \left( {V_{\text{TiB}} } \right)^{ - 1/2} } }}. $$

(2)

E is the respective elastic modulus, v is the Poisson’s ratio, V is the TiB volume fraction, and l/d is the TiB_w aspect ratio. On the other hand, the yield strength has been predicted using the following equation [29, 32, 33]:

$$ \frac{{\sigma_{{y{\text{Ti}} - {\text{TiB}}}} }}{{\sigma_{{y{\text{Ti}}}} }} = \frac{{V_{\text{TiB}} }}{2}\left( {2 + \frac{l}{d}} \right) + \left( {1 - V_{\text{TiB}} } \right), $$

(3)

where σ_y = 0.2% offset yield strength.

It is clear from the above equations that the TiB_w aspect ratio can have a significant influence on both the elastic modulus and the yield strength of Ti–TiB_w composites. Moreover, the processing of in situ nanoscale TiB_w reinforcements may also have some added benefits in terms of activating other strengthening mechanisms. For example, George et al. [34] showed that for metal–carbon nanotube (CNT) (i.e., nano-reinforcement) composites, the shear lag model also predicts the influence of the CNTs on the elastic modulus, while the Orowan looping model predicts well the yield strength increase due to CNT addition. In the case of nano-TiB_w (assuming an aspect ratio of 58:1), the shear lag model suggests an increase of ~ 38% over the elastic modulus of Ti for a 10 vol% TiB_w loading, which is higher than typical reported values, e.g., 20% [26]. So the high aspect ratio of TiB_w can have a significant influence on stiffness. Also, only a 5 vol% loading of nano-TiB_w should result in approximately similar stiffness gain as a 10 vol% TiB_w-reinforced Ti with a low TiB_w aspect ratio (5:1). Equation 3 also seems to suggest that if the aspect ratio of nano-TiB_w was maintained even at ~ 58:1, a volume loading of only 5% TiB_w will result in a composite with a strength ~ 2.5 times that of titanium. Of course other strengthening mechanisms need to be considered, for example the matrix contribution to strength through microstructural design. Hence, more research needs to be directed at nanoscale, high-aspect-ratio aligned TiB_w reinforcements to reap potential benefits of unprecedented strength and stiffness in combination with high toughness.

Bulk deformation processing (BDP)

A number of recent studies have investigated the BDP of Ti–TiB_w and Ti–TiB_w hybrid composites through forging [5], extrusion [35, 36], equal channel angular pressing [7], and hot rolling [37]. Whisker alignment, aspect ratio, and the titanium matrix microstructure can very much be modified through BDP, which in turn can have significant effects on the mechanical behavior of the composite. For example, Zhang et al. [4] hot-rolled (within the α + β phase field) in situ cast Ti–6Al–2.5Sn–4Zr–0.7Mo–0.3Si/(TiB_w + TiC_p) to increasing levels of rolling reductions up to 85%, reporting TiB_w fracture and overall shortening of TiB_w length with improved whisker distribution within the microstructure with increasing rolling reductions. Moreover, both tensile strength and ductility were also found to improve with increased rolling reduction compared with the as-cast material. Similar results were also reported for hot-rolled Ti–6Al–4V matrix reinforced with TiC and TiB_w (following initial consolidation using spark plasma sintering) [38], specifically for TiB_w alignment. In BDP, TiB_w fracture is typically reported, but sometimes to varying extents, due to the level of plastic deformation experienced and stresses involved. As mentioned earlier, TiB_w fracturing was reported by Zhang et al. [4] during hot rolling (within the α+β phase field) of 2.5 vol% TiB + TiC)/Ti–6Al–2.5Sn4Zr–0.7Mo–0.3Si (and also by others during extrusion [39]), while relatively less fracturing of TiB_w was reported by Tabrizi et al. [38] within a Ti–6Al–4V matrix for a TiB + TiC hybrid composite that was first SPS’d followed by rolling to comparable reductions. One possible reason for the difference in behavior may be the composition of the matrix and its flow characteristics during hot rolling. With such considerable deformation through hot rolling, not only the reinforcement is affected (in terms of alignment and possible fracturing) but also the matrix. For example, primary α can be refined with increasing rolling deformation, with tensile testing failure involving first the fracture of TiB_w followed by matrix ductile fracture [4]. Recrystallization of primary α has been reported [38]; similarly, dynamic recrystallization has also been reported in other bulk deformation processes such as extrusion applied to hybrid (TiB + TiC + La₂O₃)/Ti–6Al–4V composites [6]. Interestingly, Zhang et al. [40] reported differences in the fracture mode of hot-pressed Ti–TiB_w depending on the type of test and volume fraction of Ti. For example, radial cracks emanating from the corners of Vickers’ indents tended to be transgranular across TiB_w and Ti, and did not appear to depend on Ti content. However, compression testing and flexural testing change from intergranular fracture along TiB/Ti phase boundaries to transgranular fracture above a Ti content of 20%. Moreover, there was tendency for longer TiB_w to experience more transgranular fracture as compared to shorter ones. So the state of stress and Ti content may need to be considered as well when investigating fracture behavior. Interestingly, the alignment of TiB_w has been reported to improve with increase in rolling reduction and in turn improved the creep resistance of Ti (similar to composition to IMI 834)/(TiB_w–La₂O₃) composites [37] (similar results have also been reported for cast and subsequently 2D-forged near α Ti–TiB composites [5]). However, also the microstructural refinement of α plates at higher rolling reductions promoted a grain sliding mechanism that lowered the creep resistance. Hence, in creep deformation there is interplay between a number of competing factors, influenced by the existence of TiB_w (and their content [41]) and the scale of the matrix microstructure. The TiB_w fracturing has a direct negative impact on the properties, since it is aligned reinforcements with higher aspect ratios that are known to improve strength, elastic modulus, and, as mentioned above, creep resistance [5].

Although several studies applied BDP to as-cast materials, some work considered sintering prior to hot extrusion. Such research is important as in hot deformation processing the starting microstructure can affect the final matrix microstructure following BDP. Zhang et al. [36] sintered mechanically milled Ti–6Al–4V powder and TiB_w at 1150 °C for 1.5 h before extrusion. Such a sintering schedule was not enough to convert all the TiB₂ to TiB_w, and instead considerable Ti₃B₄ was formed. Extrusion (which is typically a fast process), however, provided better conditions for conversion to TiB_w due to the high level of deformation and pressures used, and good contact between the matrix and reinforcement precursors. The authors interestingly examined the extrudate at different locations along the extrusion (including material left behind inside the container, i.e., the discard) and hence could study the evolution of Ti–TiB_w during extrusion by spatial analysis of microstructure within the extrudate and discard. Similarly, hot extrusion has also been shown to promote a rapid formation of TiB_w from boron addition to titanium, which has been shown to form even before the material exits the extrusion chamber [35].

Such methods considered the extrusion of blended powder compacts [35]. Here, the use of pre-alloyed powder can be viewed as a low-cost alternative to using just elemental powders. Specifically, boron was added to two powder compositions, in one, titanium powder (hydride–dehydride) was mixed with Al–40V master alloy powder, while the other was also hydride–dehydride Ti6–Al–4V, which were both then compacted and extruded at 1300 °C. The mere chemistry of the starting powder had a pronounced effect on the morphology and size of the TiB_w formed, where the reaction between boron and Ti–6Al–4V resulted in fine TiB_w with no porosity observed, while the reaction of boron with the Ti/Al–40V mixture resulted in coarse TiB_w with lower aspect ratio and TiB cluster formation. This was suggested to be due to a limitation in the homogenous growth of TiB_w within the pre-alloyed Ti–6Al–4V powder as opposed to the pure Ti. Also, other explanations for the formation of coarse TiB with the Ti/Al-40V/B mixture are a competition between the rate of TiB_w growth and nucleation where the rate of growth is more dominant, and the inhomogeneous distribution of boron within the powder mix, with the oxide-coated Al–40V acting as a diffusion barrier [35]. The presence of the whiskers was also found to affect the matrix microstructure, where regions that did not contain whiskers displayed alpha colonies/platelets, while the matrix around the whisker-populated regions displayed more of equiaxed alpha phase. Extrusion was, however, again accompanied by whisker fracturing (as in rolling [4]) and the generation of pores and poor reinforcement/matrix interfacial bonding, specifically in the coarse TiB_w-reinforced titanium, which promoted lower strength and poor ductility in this case [35]. In another study [6], as-cast TiB + TiC + La₂0₃ hybrid composite (with a low total reinforcement volume fraction, i.e., < 2 vol%) was subjected not only to hot forging secondary processing, but subsequent to that, also hot extrusion (i.e., tertiary processing). While hot forging resulted in randomly oriented TiB_w, with obvious cluster formation and a basket weave matrix microstructure, following extrusion, clusters were eliminated and two different types of microstructures were observed. In highly reinforcement-rich regions, dynamic recrystallization resulted again in equiaxed grains, while in other regions basket weave microstructures were observed but finer than those found in the forgings. This dynamic recrystallization was found to be more pronounced as the overall volume fraction of reinforcements increased from 0.5 to 2.0 vol%. Although the extruded material displayed slightly lower strength values than the forgings, the ductility improved significantly. A similar tertiary-type processing sequence was adopted which entailed casting, followed by forging to remove casting defects and then severe plastic deformation via equal channel angular pressing (ECAP) at temperatures ranging from 600 to 900 °C of (TiB_w + La₂O₃)/Ti–6Al–4V composites [7]. Such severe plastic deformation expectedly gave rise to significant fracturing of TiB_w, which was found to increase with an increase in ECAP temperature, to the extent where ECAP at 900 °C resulted in an aspect ratio reduction down to 2.1. In fact, fracturing was observed in this case not only to occur across the whisker but also along its length. Ultrafine microstructures are typically produced in ECAP, and in this case ultrafine microstructures were partially observed, and since the extrudates were quenched immediately following extrusion, favorable conditions were there to limit conditions for grain growth. Such ultrafine microstructures involved cell structures on the order of 500 nm for materials ECAP’d at 600 and 700 °C resulting from dislocation pileup and entanglement, with evidence of dynamic recrystallization resulting in grain sizes on the order of 100 nm–500 nm occurring when ECAP was conducted at 800 and 900 °C. Moreover, grain refinement in relation to the α lamella phase increased with an increase in ECAP temperature and hence reduced down to less than 1.0 μm thickness from 2.1 μm. So although microstructure refinement is an advantage, it may seem that the TiB_w fracturing may initially detract from the appeal of ECAP. It would be interesting though to examine the effect of ECAP above the β transus on whisker fracturing.

While whisker fracturing and a reduction in the TiB_w aspect ratio of the whiskers have been a main feature of extrusion and rolling, recently 2D forging [20] has been found to not only be able to align TiB_w but also result in a relatively higher TiB_w aspect ratios compared with traditional BDP. As shown in Fig. 2, the application of the forging stroke in two different directions (2D) sequentially as can occur in industrial practice resulted in predominantly aligned whiskers with aspect ratios around 20:1 [5]. One of the interesting issues is the initial boron content, i.e., whether it is at/below or above the TiB eutectic composition (which depends on the composition of the titanium matrix). For a hypereutectic alloy, the formation of primary TiB phase during melting and casting results in coarse TiB_w and sometimes TiB plates. These have been shown to fracture upon forging and result in poor mechanical properties. Moreover, forging within the β regime was found helpful in allowing less TiB_w fracturing for near-eutectic TiB compositions. This is attributed to the significantly superior strength of TiB_w compared with the lower flow stress of the titanium matrix at that temperature [20, 42]. Multi-direction forging (Fig. 3, based on [43]) was also applied to the processing of Ti-(TiB_w/Y₂O₃) composites [44], resulting in fine dynamically recrystallized matrix grain size on the order of ~ 2.6 μm. Very high room temperature tensile strength was reported (~ 1.47 GPa). Calculations of the relative contributions of strengthening methods revealed that TiB_w contributed to strength mainly through load transfer, while Y₂O₃ submicron particles contributed to strength enhancement predominantly through thermal expansion mismatch and to a lesser extent Orowan strengthening, while grain size reduction (Hall–Petch effect) also had a contributing effect to strengthening. However, high-aspect-ratio TiB_w were not reported as in the case of 2D forging.

Figure 2

Schematic of 2D forging [5]

Figure 3

Schematic example for multi-direction forging based on [43]

Alternative elemental/compound designs

Although a significant amount of work has been published on conventional methods of constituent mixing whether via powder metallurgy processing or through casting, a few studies considered alternative designs [10, 23, 45, 46]. For example, Wang et al. [46] considered the effect of heat treatment and solution treatment on the geometric distribution of the reinforcements within Ti–TiB_w-based composites. Here, a near-α titanium alloy (Ti–5.8Al–4.0Sn–3.5Zr–1Mo–0.85Nd–0.4Si) and TiB₂ were reactively hot-pressed to form Ti–TiB_w composites. This resulted in the formation of networks of TiB_w surrounding titanium particles aided by the initial powder mixing procedure where the boron source was embedded at/on the titanium powder surfaces. Such TiB_w network-based microstructures are reported to nearly maximize the elastic modulus and improve fracture behavior (similar TiB_w networks within Ti–6Al–4V have also been recently produced through hot isostatic pressing [47], and high-temperature oxidation resistance has been reported in another study to be inferior to the unreinforced material due to the increased boundaries generated due to the presence of TiB_w [48]). The matrix consisted of coarse lamellar or near-equiaxed α with intergranular β phase. Although no silicides were observed within the β phase, nanometric α₂ (Ti₃Al) was found within the α phase (partly due to the excess alpha stabilizers in the composition). Two post-processing treatments were investigated: one is solution heat treatment within the β regime and the other thermal aging at 700 °C for 5 h. The thermal aging had the effect of increasing the precipitation of α₂ and further precipitation of silicides within the β phase, with also preferred precipitation at the TiB_w boundary, which acted as nucleation sites. On the other hand, the solution treatment within the β phase had the effect of dissolving the precipitates into solution and the subsequent water quenching promoted the formation of martensite, α’. Both the formation of martensite for the solution-treated composite and the precipitates for the thermally aged composites had the effect of increasing tensile strength compared with the as-reactively hot-pressed composite, while the ductility was only improved for the solution-treated composite. In another study, such TiB_w networks became coarser with increased volume fraction of TiB_w [47]. More interestingly, the TiB_w networks limited the growth of α/β lamella beyond the TiB_w 3D network boundaries resulting in thicker and shorter lamella as compared with TiB_w-free titanium. The growth of TiB_w into the titanium surrounded by the TiB_w 3D networks has, however, been observed in a number of studies [46, 47, 49, 50]. Figure 4 (extracted from [49]) shows exaggerated nano-TiB_w-woven interior structures resulting from the ingrowth of TiB_w inside the interior of the TiB_w 3D networks. The ball-on-plate wear resistance was found to improve with increase in TiB_w volume fraction, but also two wear mechanisms were reported, one involving chemical reaction of elements with oxygen to form various oxides, which was facilitated by the heat generated during sliding, while the other involves the action of mechanical friction [47]. In recent work by the present author, Ti–TiB_wdual matrix composites were fabricated through pressureless sintering as well as spark plasma sintering [10, 51]. In this microstructural design instead of distributing the precursor TiB₂ particles homogenously throughout the titanium, TiB_w were selectively grown in localized regions of the microstructure. This was accomplished by embedding the TiB₂ particles inside titanium powders through mechanical milling, thus producing mechanically activated Ti–TiB₂ ‘composite’ powder. These powders were then mixed with pure titanium powder, to generate a microstructure (following high-temperature consolidation) that contains regions rich in TiB_w (fully converted following SPS), and surrounded by regions of predominantly ductile titanium. The microstructure is shown schematically in Fig. 5 [10] with micrographs in Fig. 6 [10]. Such microstructures had been previously used for WC–Co among others using different processing approaches [52] and also recently by the present author for Al–carbon nanotube composites [53]. Figure 7 [10] shows how such a microstructural design can produce composites with the same ‘overall TiB_w volume fraction’ but of varying hardness, due to different sizes of the Ti–TiB_w regions, and therefore different inter-(Ti-TiB) composite particle spacing. These different size regions were obtained by classifying the initially mechanically milled Ti–TiB₂ powder into different size ranges and then mixing them with titanium powder followed by SPS to produce Ti–TiB_w dual matrix composites. In other words, the mechanical properties of Ti–TiB_w composites may therefore become tailorable. In fact, the 3D TiB_w network design and the dual matrix design can be viewed as one (3D TiB_w networks) where the TiB_w-rich networks encapsulate the more ductile Ti phase, while the other (dual matrix composites) titanium surrounds the TiB_w-rich Ti phase.

Figure 4

3D networks of TiB_w and cross-network-woven TiB_w (extracted from [49])

Figure 5

Schematic of dual matrix Ti–TiB_w composites microstructure with TiB_w–Ti regions surrounded by Ti outer matrix [10]

Figure 6

Micrographs showing spark plasma-sintered dual matrix Ti–TiB_w composites having the same overall volume fraction but with different composite particle sizes [10]

Figure 7

Effect of classified Ti–TiB_w particles size on the hardness of Ti–TiB_w dual matrix composites (all with same overall volume fraction of reinforcements) [10]

In other work, Ti/Ti–TiB_w composites were produced in the form of laminates by hot pressing stacked alternating layers of Ti and Ti–TiB₂ powders [54]. The microstructures were vastly different following hot pressing. Within the Ti layer, coarse Widmanstatten microstructures were generated, while within the converted Ti + TiB_w, a high population of 3D networks of whiskers resulted in the formation of fine equiaxed grains between the whiskers and coarser grains within the interior of the networks. Such an arrangement had a profound effect on the mechanical behavior with true stresses and ductility being superior as compared with the control (i.e., just hot-pressed Ti powder). Moreover, a non-catastrophic failure (i.e., not failing abruptly) was observed in the laminated composite as opposed to pure Ti. The hot-pressed pure Ti showed inhomogeneous shear band distributions believed to be caused by the coarse Widmanstatten microstructure, whereas the Ti–TiB_w layer promoted more of a homogeneous strain distribution. Such improved ductility over unreinforced Ti was observed only for a 5 vol% TiB_w-loaded Ti–TiB_w layers in these laminates. In a separate study [50] on Ti/Ti–TiB_w laminates, increase in TiB content within these Ti–TiB layers was investigated, which showed increases in strength and stiffness but a decline in ductility, with the lowest being observed for 12 vol% TiB_w content (the maximum volume fraction investigated in that study). The increase in TiB_w content had profound effects not only on the microstructures observed but also on the mechanical properties and fracture behavior. For example, at low TiB_w vol% needle-shaped TiB_w with ‘relatively’ larger effective aspect ratios (14.85:1) were reported as compared with 12.06:1 for 12 vol% counterpart, probably due to the TiB interlocking. The TiB morphology was found to change also with an increase in TiB volume fraction, even resulting in clawlike morphologies at the highest TiB_w additions. This is attributed to the stagnation of the growth of TiB_w emerging from the TiB₂ particles due to its contact with neighboring TiB_w, and the growth of other TiB_w from other parts of the TiB₂ particle. Such growth mechanisms have also previously been reported by Sahay et al. [21] for larger volume fractions of TiB. However, when we take into account the geometric 3D network or necklace-like distribution of TiB_w within the Ti–TiB_w layer microstructures, the volume fractions are expected to be ‘locally’ much higher than 12 vol%, resulting in these kind of morphologies as observed previously. The main effect of TiB_w volume fraction on the tensile and fracture behavior of the laminates was related to both work hardening rate and a transition in fracture behavior. For example, the work hardening rate increased from 5 to 12 vol% TiB_w addition, while the fracture behavior changed from ductile (5 vol% TiB_w) to more brittle for the 12 vol% TiB_w. Similar observation of a ductile to brittle behavior was also noted when the thickness of the Ti layer was reduced in size [55].

Chaudhari and Bauri [1] investigated carbide coating, in which they SPS’d powder compacts of Ti, Al, and KBF₄ salt (a novel cheap boron source) powders with graphite foils sandwiching the powder compact. Here, graphite was used as a carbon source to promote the formation of hard TiC layers that sandwich Ti–4Al–2Fe/TiB_w, resulting in wear-resistant surface layers. Not only has Ti–TiB been the main focus of the synthesis as a stand-alone composite, but they have also been used to enhance the properties of intermetallics. For example, Qin et al. [56] reaction-annealed alternating foils of Ti/TiB_w (cut from hot-pressed samples) and Al to produce laminates with alternating layers of Ti₃Al and Ti/TiB_w, resulting in improved strength and fracture toughness in addition to better ductility than just the Ti₃Al counterpart.

Additive manufacturing

A significant number of studies have investigated the laser-based additive manufacturing of Ti–TiB_w composites. For a detailed treatment, the reader is referred to the review by Attar et al. [57]. Hu et al. [58] investigated the effect of laser power on the developed microstructures of Ti–TiB_w composites using a laser deposition-additive manufacturing process (Fig. 8 [58, 59]). They found that only at high laser power (200 W) was it possible to obtain 3D quasi-continuous networks (3DQCN) of TiB_w due to enhanced flow, liquid capillary force, and Marangoni convection. As the power decreases, the microstructure gradually changes from having partial development of these networks to no development at the lowest power (125 W). The other notable change in the microstructure is the redistribution of boron originally at the Ti powder surfaces post-milling. At low laser power, this boron is not redistributed/mixed efficiently following melting, resulting in higher volume fraction of B locally, which promotes hypereutectic microstructures including coarse primary TiB formation. The increasing formation of these 3DQCNs with applied laser power leads to a gradual increase in strength and toughness (area under the stress–strain curve) and also microhardness [59]. Such change in laser input power did not have any appreciable effect on the mechanical properties of TiB_w-free Ti. In other words, the effect was largely focused on reorganization of TiB_w into 3D networks. Pore formation can also be an issue in SLM of Ti–TiB_w composites. In one study [60], particle shape had a pronounced effect on the pore content of the final selective laser-melted product. Milled Ti and TiB₂ powders after 2 h were spherical with TiB₂ largely being positioned on the surface of Ti powder particles, while after 4 h of milling these now composite particles were irregular in shape. The irregular powder shape results in poor flow characteristics and hence can lead to poor characteristics of the deposited layer of powder mixture, i.e., poor apparent density, which in turn negatively affects the final product following laser processing. As such, the 2-h-milled spherical particles resulted in 99.5% dense Ti–TiB_w product, while the 4-h-milled irregular particles resulted in 95.1%, and these in turn resulted in inferior mechanical properties for irregular powders (883 MPa and 5.5% compressive ultimate strength and strain as opposed to 1421 MPa and 17.8% for spherical particles). Attar et al. [61] further provided a comparative study for the microstructure and properties of Ti–TiB_w produced by SLM, sintering, and casting. Here, the starting powder in all cases was milled Ti–TiB₂ powder that had a near-spherical shape. For sintering, the milled powders were compacted at different pressures and subsequently sintered at 1100 °C for 3 h in argon followed by water quenching. Although only Ti and TiB_w were present after sintering, the compacts were porous with porosity decreasing from 36 to 25% as the initial powder compaction pressure increased from 450 to 700 MPa. Such porous microstructures have also been previously observed by the present author through pressureless sintering Ti–TiB_w/Ti dual matrix composites; however, in that case 3-h sintering at 1100 °C and 1200 °C resulted in lower levels of porosity (due to excess outer matrix titanium) and some unconverted TiB₂ (possible due to an initial higher volume fraction of TiB₂ locally) [10]. In comparison with the sintered material, SLM and cast structures boasted dense products, with as-cast materials generating coarser TiB_w in comparison with SLM products due to the different cooling rates experienced for both methods. In fact, even colonies on the nanoscale were found in the SLM products, presumably as a result of the rapid cooling. The overall microstructures had a profound effect on the mechanical properties where SLM and as-cast products exhibited generally similar moduli and strengths, while the porous sintered products displayed much lower elastic modulus and strength (mainly as a consequence of the porosity, which also affected the failure mechanism) [61]. Some work recently considered non-laser-based additive manufacturing approaches. For example, Sheydaeian and Toyserkani [62] considered an approach that combines binder jet additive manufacturing with extrusion technology where a high resin-based TiB₂ is applied at predetermined locations between successive Ti layers followed by pressureless sintering, to produce what they named Ti–TiB_w periodic composites. Although TiB_w formation was reported, the porosity was significant, even with sintering at 1400 °C. The approach is interesting though and may find useful applications. Another non-laser additive manufacturing approach involved the wire arc additive manufacturing of Ti–TiB_w composites, specifically of Ti–6Al–4V with trace of B addition, which generated 0.9 vol% TiB_w and less (~ 1 μm diameter and 10 μm long on average) [63]. Such TiB_w levels are not expected to intrinsically result in significant improvements in mechanical properties; however, their presence can act as nucleation sites for the formation of homogenous equiaxed α, with compressive true strain at failure increasing to around 40% with boron addition of 0.13% following heat treatment.

Figure 8

Schematics of setup and laser–material interaction in laser deposition–additive manufacturing process [58, 59]

Ti–TiB_w as a biomaterial

Within the orthopedic industry, normally titanium is not looked upon as being a material suitable for articulating applications, largely due to its poor wear resistance. Hence, other materials such as ceramics and more wear-resistant alternatives had been at the forefront in these applications [64, 65]. Understandably, composite and ceramic coatings have been suggested to enable Ti to compete in such applications, for example, as femoral heads [66, 67] and recent studies had considered Ti composites as a biomaterial [68, 69]. As pointed out by Ege et al. [67], there are very few papers that considered the mechanical and tribological properties of bio-inert ceramic coatings on Ti–6Al–4V (amounting to only 19 papers over the period 1990–2017), with nitrides receiving the most significant share of attention (47%). However, since the emergence of titanium composites (including Ti–TiB_w and Ti–TiC) has shown significant wear resistance enhancements over the wear resistance of Ti, it may now be possible to consider these composites in such articulating applications. Due to the reported cytotoxicity of vanadium and aluminum in Ti–6Al–4V, other alloys such as β titanium alloys have been developed to avoid such effects. These alloys typically have elastic moduli values (~ 55–60 GPa) closer to bone (10–40 GPa) than Ti [66] and hence reduce the issue of stress shielding. Examples of these alloys are Ti–Nb–Zr–Ta (TNZT) and Ti–Nb–Ta–Zr (TNTZ) depending on the relative amounts of Zr and Ta. Although the incorporation of titanium borides within these materials may increase further the modulus (but now starting from a lower matrix modulus compared with other Ti alloys), they provide greatly needed wear resistance. Hence, some compromise may need to be achieved, through microstructural design, coating approaches, or other novel approaches. The work on Ti–TiB_w as biomaterials is, however, comparatively still in its infancy.

Recently, a preliminary evaluation of the biocompatibility of Ti–TiB_w dual matrix composites was conducted [12]. The composite was shown not to be cytotoxic with no observed abnormal morphologies in fibroblast cells. Additionally, they were found to have a hemolytic index similar to that of CP-Ti and even marginally lower than Ti–6Al–4V, as shown in Fig. 9. Similar positive results were reported for TiB–TiN-reinforced Ti–6Al–4V hybrid composite coatings, where the composite coatings were shown to be non-toxic, with similar biocompatibility (and interaction between material surface and cells) to that of CP-Ti [70]. Targeting dental applications, the culturing of gingival fibroblasts and osteoblasts on the surface of CP-Ti plasma surface alloyed with B (which formed TiB reinforcements at the surface) again showed promise as a biomaterial [71]. In another study, with the addition of only 0.1 wt% B to Ti–6Al–4V the response of osteoblasts was not significantly altered compared with just Ti–6Al–4V [72]. Majumdar et al. [73], however, reported that the addition of 0.5 wt% boron (i.e., higher B content) to Ti–13Zr–13Nb resulted in lower protein surface adsorption and cell response compared with boron-free Ti–13Zr–13Nb, which may possibly be related to differences in surface chemistry of this particular alloy composition. Similar results were reported by the same group for Ti–35–5.7Ta–7.2Zr–0.5B, where cell attachment was better for the boron-free alloy [74], yet both were also still better than the control. In another study [68], SLM’d Ti–TiB_w composites were found to display enhanced corrosion resistance at body temperature in simulated body fluid (Hank’s solution) as opposed to CP-Ti. Here, the TiB_w or unconverted TiB₂ acted as micro-cathodes with the Ti matrix acting as the anode, thus promoting the dissolution of Ti which in turn resulted in the formation of a protective layer for Ti–TiB composites. Similar results were reported by Bahl et al. [72] where TiB_w was also reported to act as cathodes, while both β and α Ti act as the anodes resulting in their dissolution, leading to a lower passivation potential which could lead to spontaneous passivation in oxidizing conditions. Interestingly though, the alignment of the TiB_w relative to the test surface was found to affect corrosion behavior. Here, TiB_w aligned perpendicular to the surface resulted in less planar area of exposed TiB_w, hence resulting in a lowering of galvanic effect in comparison to TiB_w that are aligned parallel to the surface. From the point of view of corrosion, Sivakumar et al. [75] had previously investigated the corrosion behavior of boride-coated CP-Ti in Ringer’s solution for implant applications. Here, CP-Ti was borided at varying temperatures and times to produce TiB₂/Ti₃B₄/TiB coatings that were less than 30 μm thick. The outer layer (which was comparatively small in thickness) was TiB₂, which possessed a hardness close to 3000 HV. The boriding had an effect of reducing the pitting corrosion resistance (but was still within an acceptable range) in Ringer’s solution. Here, the corrosion mechanism involved the dissolution of TiB₂. Hence, the spatial incorporation of TiB₂ (i.e., whether as a coat or particulate with a matrix) is important.

Figure 9

Optical images of fibroblasts (high and low magnifications) a,b for Ti–TiB_w composite together with three controls (negative, medium, and positive), c normalized hemolytic index [12]

As mentioned, although Ti alloys are largely considered in orthopedics for the femoral stem, their reinforcement with borides may present other possibilities in wear or articulating applications, for example as a femoral head for which good wear properties are required. In a recent study, the laser processing of TNZT reinforced with borides has led to primary borides within the microstructures with varying compositions throughout its cross section, where different levels of Ti, Ta, Zr, and Nb were present [66]. The oxidation of the composite produced a rutile layer which significantly lowered the coefficient of friction and wear when the mating surface was a harder material (i.e., Si₃N₄ ball), which was believed to protect the material against boride pullout and third-body wear. This was not the case in the absence of oxidation, where the boride-reinforced composite resulted in worse wear characteristics than the Ti–6Al–4V ELI (extra low interstitial), due to boride pullout and third-body wear. Against a softer material (i.e., SS440C stainless steel balls) without the rutile oxide layer, the coefficient of friction and wear characteristics of the boride-reinforced material were superior to those of Ti–6Al–4V ELI; hence, the hardness of the mating surface will certainly have an effect on the wear characteristics. For Ti–6Al–4V reinforced with TiB and TiN, superior (approximately an order of magnitude better) wear resistance in simulated body fluid has been reported in comparison with CoCrMo alloy (an alloy used in these articulating applications) [76]. Although the consideration of Ti–TiB_w in articulating orthopedic applications is still evolving, results are generally promising. Table 1 summarizes some reported mechanical properties of Ti–TiB composites.

Table 1 Summary of selected reported mechanical properties of Ti and Ti–TiB_w composites (the table is an expansion on previously reported tables [15, 77]), where the current table covers additional reported work within the past 5 years

Summary

The research into Ti–TiB_w composites has been significant over the past 5 years, providing advancements within the field. The application of 2D forging has provided means to extend the aspect ratio of TiB_w to higher levels than usually possible in bulk deformation processing. Such enhancements may have significant advantages in terms of improved load transfer and strength. Similarly for general processing, the effect initial TiB₂ particle size can also have a similar effect of increasing aspect ratios so far as high as 58:1. The generation of 3D TiB_w networks within Ti–TiB composites provides interesting combinations of properties, as it affects not only the TiB distributions but also the internal Ti matrix microstructure. However, initial indications on their oxidation resistance suggest more work is needed in this regard. Other notable microstructural designs are the dual matrix composite design where TiB_w are grown only in selected regions of the microstructure always being surrounded by predominantly unreinforced titanium. Such microstructural designs applied to Ti–TiB_w provide the potential of not only improved toughness at the same TiB_w volume fraction but also improved wear resistance and the possibility of tailorable properties for the same overall volume fraction of reinforcements. Ti/TiB laminates have also shown interesting results, and the application of additive manufacturing to Ti–TiB_w is another important area with significant potential. The consideration of Ti–TiB_w as a biomaterial has also been outlined and should receive more attention.